diff options
Diffstat (limited to 'drivers/gpu')
546 files changed, 19020 insertions, 7031 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h index da3d02154fa6..8ebc5f1eb4c0 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h @@ -85,8 +85,12 @@ extern int amdgpu_vm_debug; extern int amdgpu_sched_jobs; extern int amdgpu_sched_hw_submission; extern int amdgpu_powerplay; +extern int amdgpu_powercontainment; extern unsigned amdgpu_pcie_gen_cap; extern unsigned amdgpu_pcie_lane_cap; +extern unsigned amdgpu_cg_mask; +extern unsigned amdgpu_pg_mask; +extern char *amdgpu_disable_cu; #define AMDGPU_WAIT_IDLE_TIMEOUT_IN_MS 3000 #define AMDGPU_MAX_USEC_TIMEOUT 100000 /* 100 ms */ @@ -183,6 +187,10 @@ int amdgpu_set_clockgating_state(struct amdgpu_device *adev, int amdgpu_set_powergating_state(struct amdgpu_device *adev, enum amd_ip_block_type block_type, enum amd_powergating_state state); +int amdgpu_wait_for_idle(struct amdgpu_device *adev, + enum amd_ip_block_type block_type); +bool amdgpu_is_idle(struct amdgpu_device *adev, + enum amd_ip_block_type block_type); struct amdgpu_ip_block_version { enum amd_ip_block_type type; @@ -298,13 +306,16 @@ struct amdgpu_ring_funcs { uint32_t oa_base, uint32_t oa_size); /* testing functions */ int (*test_ring)(struct amdgpu_ring *ring); - int (*test_ib)(struct amdgpu_ring *ring); + int (*test_ib)(struct amdgpu_ring *ring, long timeout); /* insert NOP packets */ void (*insert_nop)(struct amdgpu_ring *ring, uint32_t count); /* pad the indirect buffer to the necessary number of dw */ void (*pad_ib)(struct amdgpu_ring *ring, struct amdgpu_ib *ib); unsigned (*init_cond_exec)(struct amdgpu_ring *ring); void (*patch_cond_exec)(struct amdgpu_ring *ring, unsigned offset); + /* note usage for clock and power gating */ + void (*begin_use)(struct amdgpu_ring *ring); + void (*end_use)(struct amdgpu_ring *ring); }; /* @@ -594,11 +605,9 @@ int amdgpu_sync_resv(struct amdgpu_device *adev, struct amdgpu_sync *sync, struct reservation_object *resv, void *owner); -bool amdgpu_sync_is_idle(struct amdgpu_sync *sync); -int amdgpu_sync_cycle_fences(struct amdgpu_sync *dst, struct amdgpu_sync *src, - struct fence *fence); +struct fence *amdgpu_sync_peek_fence(struct amdgpu_sync *sync, + struct amdgpu_ring *ring); struct fence *amdgpu_sync_get_fence(struct amdgpu_sync *sync); -int amdgpu_sync_wait(struct amdgpu_sync *sync); void amdgpu_sync_free(struct amdgpu_sync *sync); int amdgpu_sync_init(void); void amdgpu_sync_fini(void); @@ -754,12 +763,11 @@ int amdgpu_job_alloc(struct amdgpu_device *adev, unsigned num_ibs, int amdgpu_job_alloc_with_ib(struct amdgpu_device *adev, unsigned size, struct amdgpu_job **job); +void amdgpu_job_free_resources(struct amdgpu_job *job); void amdgpu_job_free(struct amdgpu_job *job); -void amdgpu_job_free_func(struct kref *refcount); int amdgpu_job_submit(struct amdgpu_job *job, struct amdgpu_ring *ring, struct amd_sched_entity *entity, void *owner, struct fence **f); -void amdgpu_job_timeout_func(struct work_struct *work); struct amdgpu_ring { struct amdgpu_device *adev; @@ -767,12 +775,9 @@ struct amdgpu_ring { struct amdgpu_fence_driver fence_drv; struct amd_gpu_scheduler sched; - spinlock_t fence_lock; struct amdgpu_bo *ring_obj; volatile uint32_t *ring; unsigned rptr_offs; - u64 next_rptr_gpu_addr; - volatile u32 *next_rptr_cpu_addr; unsigned wptr; unsigned wptr_old; unsigned ring_size; @@ -791,14 +796,16 @@ struct amdgpu_ring { u32 doorbell_index; bool use_doorbell; unsigned wptr_offs; - unsigned next_rptr_offs; unsigned fence_offs; uint64_t current_ctx; enum amdgpu_ring_type type; char name[16]; unsigned cond_exe_offs; - u64 cond_exe_gpu_addr; - volatile u32 *cond_exe_cpu_addr; + u64 cond_exe_gpu_addr; + volatile u32 *cond_exe_cpu_addr; +#if defined(CONFIG_DEBUG_FS) + struct dentry *ent; +#endif }; /* @@ -861,6 +868,7 @@ struct amdgpu_vm { struct amdgpu_bo *page_directory; unsigned max_pde_used; struct fence *page_directory_fence; + uint64_t last_eviction_counter; /* array of page tables, one for each page directory entry */ struct amdgpu_vm_pt *page_tables; @@ -883,13 +891,14 @@ struct amdgpu_vm_id { struct fence *first; struct amdgpu_sync active; struct fence *last_flush; - struct amdgpu_ring *last_user; atomic64_t owner; uint64_t pd_gpu_addr; /* last flushed PD/PT update */ struct fence *flushed_updates; + uint32_t current_gpu_reset_count; + uint32_t gds_base; uint32_t gds_size; uint32_t gws_base; @@ -905,6 +914,10 @@ struct amdgpu_vm_manager { struct list_head ids_lru; struct amdgpu_vm_id ids[AMDGPU_NUM_VM]; + /* Handling of VM fences */ + u64 fence_context; + unsigned seqno[AMDGPU_MAX_RINGS]; + uint32_t max_pfn; /* vram base address for page table entry */ u64 vram_base_offset; @@ -926,17 +939,14 @@ void amdgpu_vm_fini(struct amdgpu_device *adev, struct amdgpu_vm *vm); void amdgpu_vm_get_pd_bo(struct amdgpu_vm *vm, struct list_head *validated, struct amdgpu_bo_list_entry *entry); -void amdgpu_vm_get_pt_bos(struct amdgpu_vm *vm, struct list_head *duplicates); +void amdgpu_vm_get_pt_bos(struct amdgpu_device *adev, struct amdgpu_vm *vm, + struct list_head *duplicates); void amdgpu_vm_move_pt_bos_in_lru(struct amdgpu_device *adev, struct amdgpu_vm *vm); int amdgpu_vm_grab_id(struct amdgpu_vm *vm, struct amdgpu_ring *ring, struct amdgpu_sync *sync, struct fence *fence, - unsigned *vm_id, uint64_t *vm_pd_addr); -int amdgpu_vm_flush(struct amdgpu_ring *ring, - unsigned vm_id, uint64_t pd_addr, - uint32_t gds_base, uint32_t gds_size, - uint32_t gws_base, uint32_t gws_size, - uint32_t oa_base, uint32_t oa_size); + struct amdgpu_job *job); +int amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job); void amdgpu_vm_reset_id(struct amdgpu_device *adev, unsigned vm_id); uint64_t amdgpu_vm_map_gart(const dma_addr_t *pages_addr, uint64_t addr); int amdgpu_vm_update_page_directory(struct amdgpu_device *adev, @@ -1142,6 +1152,12 @@ struct amdgpu_cu_info { uint32_t bitmap[4][4]; }; +struct amdgpu_gfx_funcs { + /* get the gpu clock counter */ + uint64_t (*get_gpu_clock_counter)(struct amdgpu_device *adev); + void (*select_se_sh)(struct amdgpu_device *adev, u32 se_num, u32 sh_num, u32 instance); +}; + struct amdgpu_gfx { struct mutex gpu_clock_mutex; struct amdgpu_gca_config config; @@ -1178,6 +1194,7 @@ struct amdgpu_gfx { /* ce ram size*/ unsigned ce_ram_size; struct amdgpu_cu_info cu_info; + const struct amdgpu_gfx_funcs *funcs; }; int amdgpu_ib_get(struct amdgpu_device *adev, struct amdgpu_vm *vm, @@ -1195,10 +1212,6 @@ void amdgpu_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count); void amdgpu_ring_generic_pad_ib(struct amdgpu_ring *ring, struct amdgpu_ib *ib); void amdgpu_ring_commit(struct amdgpu_ring *ring); void amdgpu_ring_undo(struct amdgpu_ring *ring); -unsigned amdgpu_ring_backup(struct amdgpu_ring *ring, - uint32_t **data); -int amdgpu_ring_restore(struct amdgpu_ring *ring, - unsigned size, uint32_t *data); int amdgpu_ring_init(struct amdgpu_device *adev, struct amdgpu_ring *ring, unsigned ring_size, u32 nop, u32 align_mask, struct amdgpu_irq_src *irq_src, unsigned irq_type, @@ -1250,6 +1263,7 @@ struct amdgpu_job { uint32_t num_ibs; void *owner; uint64_t ctx; + bool vm_needs_flush; unsigned vm_id; uint64_t vm_pd_addr; uint32_t gds_base, gds_size; @@ -1257,8 +1271,7 @@ struct amdgpu_job { uint32_t oa_base, oa_size; /* user fence handling */ - struct amdgpu_bo *uf_bo; - uint32_t uf_offset; + uint64_t uf_addr; uint64_t uf_sequence; }; @@ -1560,6 +1573,12 @@ struct amdgpu_dpm_funcs { u32 (*get_fan_control_mode)(struct amdgpu_device *adev); int (*set_fan_speed_percent)(struct amdgpu_device *adev, u32 speed); int (*get_fan_speed_percent)(struct amdgpu_device *adev, u32 *speed); + int (*force_clock_level)(struct amdgpu_device *adev, enum pp_clock_type type, uint32_t mask); + int (*print_clock_levels)(struct amdgpu_device *adev, enum pp_clock_type type, char *buf); + int (*get_sclk_od)(struct amdgpu_device *adev); + int (*set_sclk_od)(struct amdgpu_device *adev, uint32_t value); + int (*get_mclk_od)(struct amdgpu_device *adev); + int (*set_mclk_od)(struct amdgpu_device *adev, uint32_t value); }; struct amdgpu_dpm { @@ -1662,6 +1681,7 @@ struct amdgpu_uvd { struct amdgpu_ring ring; struct amdgpu_irq_src irq; bool address_64_bit; + bool use_ctx_buf; struct amd_sched_entity entity; }; @@ -1683,6 +1703,7 @@ struct amdgpu_vce { struct drm_file *filp[AMDGPU_MAX_VCE_HANDLES]; uint32_t img_size[AMDGPU_MAX_VCE_HANDLES]; struct delayed_work idle_work; + struct mutex idle_mutex; const struct firmware *fw; /* VCE firmware */ struct amdgpu_ring ring[AMDGPU_MAX_VCE_RINGS]; struct amdgpu_irq_src irq; @@ -1767,6 +1788,8 @@ int amdgpu_debugfs_init(struct drm_minor *minor); void amdgpu_debugfs_cleanup(struct drm_minor *minor); #endif +int amdgpu_debugfs_firmware_init(struct amdgpu_device *adev); + /* * amdgpu smumgr functions */ @@ -1811,15 +1834,13 @@ struct amdgpu_asic_funcs { u32 sh_num, u32 reg_offset, u32 *value); void (*set_vga_state)(struct amdgpu_device *adev, bool state); int (*reset)(struct amdgpu_device *adev); - /* wait for mc_idle */ - int (*wait_for_mc_idle)(struct amdgpu_device *adev); /* get the reference clock */ u32 (*get_xclk)(struct amdgpu_device *adev); - /* get the gpu clock counter */ - uint64_t (*get_gpu_clock_counter)(struct amdgpu_device *adev); /* MM block clocks */ int (*set_uvd_clocks)(struct amdgpu_device *adev, u32 vclk, u32 dclk); int (*set_vce_clocks)(struct amdgpu_device *adev, u32 evclk, u32 ecclk); + /* query virtual capabilities */ + u32 (*get_virtual_caps)(struct amdgpu_device *adev); }; /* @@ -1914,8 +1935,12 @@ void amdgpu_cgs_destroy_device(struct cgs_device *cgs_device); /* GPU virtualization */ +#define AMDGPU_VIRT_CAPS_SRIOV_EN (1 << 0) +#define AMDGPU_VIRT_CAPS_IS_VF (1 << 1) struct amdgpu_virtualization { bool supports_sr_iov; + bool is_virtual; + u32 caps; }; /* @@ -1997,6 +2022,10 @@ struct amdgpu_device { spinlock_t didt_idx_lock; amdgpu_rreg_t didt_rreg; amdgpu_wreg_t didt_wreg; + /* protects concurrent gc_cac register access */ + spinlock_t gc_cac_idx_lock; + amdgpu_rreg_t gc_cac_rreg; + amdgpu_wreg_t gc_cac_wreg; /* protects concurrent ENDPOINT (audio) register access */ spinlock_t audio_endpt_idx_lock; amdgpu_block_rreg_t audio_endpt_rreg; @@ -2022,6 +2051,7 @@ struct amdgpu_device { atomic64_t vram_vis_usage; atomic64_t gtt_usage; atomic64_t num_bytes_moved; + atomic64_t num_evictions; atomic_t gpu_reset_counter; /* display */ @@ -2125,6 +2155,8 @@ void amdgpu_mm_wdoorbell(struct amdgpu_device *adev, u32 index, u32 v); #define WREG32_UVD_CTX(reg, v) adev->uvd_ctx_wreg(adev, (reg), (v)) #define RREG32_DIDT(reg) adev->didt_rreg(adev, (reg)) #define WREG32_DIDT(reg, v) adev->didt_wreg(adev, (reg), (v)) +#define RREG32_GC_CAC(reg) adev->gc_cac_rreg(adev, (reg)) +#define WREG32_GC_CAC(reg, v) adev->gc_cac_wreg(adev, (reg), (v)) #define RREG32_AUDIO_ENDPT(block, reg) adev->audio_endpt_rreg(adev, (block), (reg)) #define WREG32_AUDIO_ENDPT(block, reg, v) adev->audio_endpt_wreg(adev, (block), (reg), (v)) #define WREG32_P(reg, val, mask) \ @@ -2200,11 +2232,10 @@ amdgpu_get_sdma_instance(struct amdgpu_ring *ring) */ #define amdgpu_asic_set_vga_state(adev, state) (adev)->asic_funcs->set_vga_state((adev), (state)) #define amdgpu_asic_reset(adev) (adev)->asic_funcs->reset((adev)) -#define amdgpu_asic_wait_for_mc_idle(adev) (adev)->asic_funcs->wait_for_mc_idle((adev)) #define amdgpu_asic_get_xclk(adev) (adev)->asic_funcs->get_xclk((adev)) #define amdgpu_asic_set_uvd_clocks(adev, v, d) (adev)->asic_funcs->set_uvd_clocks((adev), (v), (d)) #define amdgpu_asic_set_vce_clocks(adev, ev, ec) (adev)->asic_funcs->set_vce_clocks((adev), (ev), (ec)) -#define amdgpu_asic_get_gpu_clock_counter(adev) (adev)->asic_funcs->get_gpu_clock_counter((adev)) +#define amdgpu_asic_get_virtual_caps(adev) ((adev)->asic_funcs->get_virtual_caps((adev))) #define amdgpu_asic_read_disabled_bios(adev) (adev)->asic_funcs->read_disabled_bios((adev)) #define amdgpu_asic_read_bios_from_rom(adev, b, l) (adev)->asic_funcs->read_bios_from_rom((adev), (b), (l)) #define amdgpu_asic_read_register(adev, se, sh, offset, v)((adev)->asic_funcs->read_register((adev), (se), (sh), (offset), (v))) @@ -2215,7 +2246,7 @@ amdgpu_get_sdma_instance(struct amdgpu_ring *ring) #define amdgpu_vm_set_pte_pde(adev, ib, pe, addr, count, incr, flags) ((adev)->vm_manager.vm_pte_funcs->set_pte_pde((ib), (pe), (addr), (count), (incr), (flags))) #define amdgpu_ring_parse_cs(r, p, ib) ((r)->funcs->parse_cs((p), (ib))) #define amdgpu_ring_test_ring(r) (r)->funcs->test_ring((r)) -#define amdgpu_ring_test_ib(r) (r)->funcs->test_ib((r)) +#define amdgpu_ring_test_ib(r, t) (r)->funcs->test_ib((r), (t)) #define amdgpu_ring_get_rptr(r) (r)->funcs->get_rptr((r)) #define amdgpu_ring_get_wptr(r) (r)->funcs->get_wptr((r)) #define amdgpu_ring_set_wptr(r) (r)->funcs->set_wptr((r)) @@ -2257,6 +2288,8 @@ amdgpu_get_sdma_instance(struct amdgpu_ring *ring) #define amdgpu_dpm_print_power_state(adev, ps) (adev)->pm.funcs->print_power_state((adev), (ps)) #define amdgpu_dpm_vblank_too_short(adev) (adev)->pm.funcs->vblank_too_short((adev)) #define amdgpu_dpm_enable_bapm(adev, e) (adev)->pm.funcs->enable_bapm((adev), (e)) +#define amdgpu_gfx_get_gpu_clock_counter(adev) (adev)->gfx.funcs->get_gpu_clock_counter((adev)) +#define amdgpu_gfx_select_se_sh(adev, se, sh, instance) (adev)->gfx.funcs->select_se_sh((adev), (se), (sh), (instance)) #define amdgpu_dpm_get_temperature(adev) \ ((adev)->pp_enabled ? \ @@ -2335,6 +2368,18 @@ amdgpu_get_sdma_instance(struct amdgpu_ring *ring) #define amdgpu_dpm_force_clock_level(adev, type, level) \ (adev)->powerplay.pp_funcs->force_clock_level((adev)->powerplay.pp_handle, type, level) +#define amdgpu_dpm_get_sclk_od(adev) \ + (adev)->powerplay.pp_funcs->get_sclk_od((adev)->powerplay.pp_handle) + +#define amdgpu_dpm_set_sclk_od(adev, value) \ + (adev)->powerplay.pp_funcs->set_sclk_od((adev)->powerplay.pp_handle, value) + +#define amdgpu_dpm_get_mclk_od(adev) \ + ((adev)->powerplay.pp_funcs->get_mclk_od((adev)->powerplay.pp_handle)) + +#define amdgpu_dpm_set_mclk_od(adev, value) \ + ((adev)->powerplay.pp_funcs->set_mclk_od((adev)->powerplay.pp_handle, value)) + #define amdgpu_dpm_dispatch_task(adev, event_id, input, output) \ (adev)->powerplay.pp_funcs->dispatch_tasks((adev)->powerplay.pp_handle, (event_id), (input), (output)) @@ -2376,9 +2421,13 @@ bool amdgpu_device_is_px(struct drm_device *dev); #if defined(CONFIG_VGA_SWITCHEROO) void amdgpu_register_atpx_handler(void); void amdgpu_unregister_atpx_handler(void); +bool amdgpu_has_atpx_dgpu_power_cntl(void); +bool amdgpu_is_atpx_hybrid(void); #else static inline void amdgpu_register_atpx_handler(void) {} static inline void amdgpu_unregister_atpx_handler(void) {} +static inline bool amdgpu_has_atpx_dgpu_power_cntl(void) { return false; } +static inline bool amdgpu_is_atpx_hybrid(void) { return false; } #endif /* diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c index 32809f749903..d080d0807a5b 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c @@ -240,8 +240,8 @@ uint64_t get_gpu_clock_counter(struct kgd_dev *kgd) { struct amdgpu_device *rdev = (struct amdgpu_device *)kgd; - if (rdev->asic_funcs->get_gpu_clock_counter) - return rdev->asic_funcs->get_gpu_clock_counter(rdev); + if (rdev->gfx.funcs->get_gpu_clock_counter) + return rdev->gfx.funcs->get_gpu_clock_counter(rdev); return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c index 9df1bcb35bf0..983175363b06 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c @@ -551,28 +551,19 @@ int amdgpu_atombios_get_clock_info(struct amdgpu_device *adev) le16_to_cpu(firmware_info->info.usReferenceClock); ppll->reference_div = 0; - if (crev < 2) - ppll->pll_out_min = - le16_to_cpu(firmware_info->info.usMinPixelClockPLL_Output); - else - ppll->pll_out_min = - le32_to_cpu(firmware_info->info_12.ulMinPixelClockPLL_Output); + ppll->pll_out_min = + le32_to_cpu(firmware_info->info_12.ulMinPixelClockPLL_Output); ppll->pll_out_max = le32_to_cpu(firmware_info->info.ulMaxPixelClockPLL_Output); - if (crev >= 4) { - ppll->lcd_pll_out_min = - le16_to_cpu(firmware_info->info_14.usLcdMinPixelClockPLL_Output) * 100; - if (ppll->lcd_pll_out_min == 0) - ppll->lcd_pll_out_min = ppll->pll_out_min; - ppll->lcd_pll_out_max = - le16_to_cpu(firmware_info->info_14.usLcdMaxPixelClockPLL_Output) * 100; - if (ppll->lcd_pll_out_max == 0) - ppll->lcd_pll_out_max = ppll->pll_out_max; - } else { + ppll->lcd_pll_out_min = + le16_to_cpu(firmware_info->info_14.usLcdMinPixelClockPLL_Output) * 100; + if (ppll->lcd_pll_out_min == 0) ppll->lcd_pll_out_min = ppll->pll_out_min; + ppll->lcd_pll_out_max = + le16_to_cpu(firmware_info->info_14.usLcdMaxPixelClockPLL_Output) * 100; + if (ppll->lcd_pll_out_max == 0) ppll->lcd_pll_out_max = ppll->pll_out_max; - } if (ppll->pll_out_min == 0) ppll->pll_out_min = 64800; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atpx_handler.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_atpx_handler.c index 35a1248aaa77..49de92600074 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atpx_handler.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atpx_handler.c @@ -10,6 +10,7 @@ #include <linux/slab.h> #include <linux/acpi.h> #include <linux/pci.h> +#include <linux/delay.h> #include "amd_acpi.h" @@ -27,6 +28,7 @@ struct amdgpu_atpx_functions { struct amdgpu_atpx { acpi_handle handle; struct amdgpu_atpx_functions functions; + bool is_hybrid; }; static struct amdgpu_atpx_priv { @@ -63,6 +65,14 @@ bool amdgpu_has_atpx(void) { return amdgpu_atpx_priv.atpx_detected; } +bool amdgpu_has_atpx_dgpu_power_cntl(void) { + return amdgpu_atpx_priv.atpx.functions.power_cntl; +} + +bool amdgpu_is_atpx_hybrid(void) { + return amdgpu_atpx_priv.atpx.is_hybrid; +} + /** * amdgpu_atpx_call - call an ATPX method * @@ -142,18 +152,12 @@ static void amdgpu_atpx_parse_functions(struct amdgpu_atpx_functions *f, u32 mas */ static int amdgpu_atpx_validate(struct amdgpu_atpx *atpx) { - /* make sure required functions are enabled */ - /* dGPU power control is required */ - if (atpx->functions.power_cntl == false) { - printk("ATPX dGPU power cntl not present, forcing\n"); - atpx->functions.power_cntl = true; - } + u32 valid_bits = 0; if (atpx->functions.px_params) { union acpi_object *info; struct atpx_px_params output; size_t size; - u32 valid_bits; info = amdgpu_atpx_call(atpx->handle, ATPX_FUNCTION_GET_PX_PARAMETERS, NULL); if (!info) @@ -172,19 +176,43 @@ static int amdgpu_atpx_validate(struct amdgpu_atpx *atpx) memcpy(&output, info->buffer.pointer, size); valid_bits = output.flags & output.valid_flags; - /* if separate mux flag is set, mux controls are required */ - if (valid_bits & ATPX_SEPARATE_MUX_FOR_I2C) { - atpx->functions.i2c_mux_cntl = true; - atpx->functions.disp_mux_cntl = true; - } - /* if any outputs are muxed, mux controls are required */ - if (valid_bits & (ATPX_CRT1_RGB_SIGNAL_MUXED | - ATPX_TV_SIGNAL_MUXED | - ATPX_DFP_SIGNAL_MUXED)) - atpx->functions.disp_mux_cntl = true; kfree(info); } + + /* if separate mux flag is set, mux controls are required */ + if (valid_bits & ATPX_SEPARATE_MUX_FOR_I2C) { + atpx->functions.i2c_mux_cntl = true; + atpx->functions.disp_mux_cntl = true; + } + /* if any outputs are muxed, mux controls are required */ + if (valid_bits & (ATPX_CRT1_RGB_SIGNAL_MUXED | + ATPX_TV_SIGNAL_MUXED | + ATPX_DFP_SIGNAL_MUXED)) + atpx->functions.disp_mux_cntl = true; + + + /* some bioses set these bits rather than flagging power_cntl as supported */ + if (valid_bits & (ATPX_DYNAMIC_PX_SUPPORTED | + ATPX_DYNAMIC_DGPU_POWER_OFF_SUPPORTED)) + atpx->functions.power_cntl = true; + + atpx->is_hybrid = false; + if (valid_bits & ATPX_MS_HYBRID_GFX_SUPPORTED) { + printk("ATPX Hybrid Graphics\n"); +#if 1 + /* This is a temporary hack until the D3 cold support + * makes it upstream. The ATPX power_control method seems + * to still work on even if the system should be using + * the new standardized hybrid D3 cold ACPI interface. + */ + atpx->functions.power_cntl = true; +#else + atpx->functions.power_cntl = false; +#endif + atpx->is_hybrid = true; + } + return 0; } @@ -259,6 +287,10 @@ static int amdgpu_atpx_set_discrete_state(struct amdgpu_atpx *atpx, u8 state) if (!info) return -EIO; kfree(info); + + /* 200ms delay is required after off */ + if (state == 0) + msleep(200); } return 0; } @@ -507,7 +539,6 @@ static int amdgpu_atpx_get_client_id(struct pci_dev *pdev) static const struct vga_switcheroo_handler amdgpu_atpx_handler = { .switchto = amdgpu_atpx_switchto, .power_state = amdgpu_atpx_power_state, - .init = amdgpu_atpx_init, .get_client_id = amdgpu_atpx_get_client_id, }; @@ -542,6 +573,7 @@ static bool amdgpu_atpx_detect(void) printk(KERN_INFO "vga_switcheroo: detected switching method %s handle\n", acpi_method_name); amdgpu_atpx_priv.atpx_detected = true; + amdgpu_atpx_init(); return true; } return false; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_bios.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_bios.c index 99ca75baa47d..2b6afe123f3d 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_bios.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_bios.c @@ -352,22 +352,22 @@ bool amdgpu_get_bios(struct amdgpu_device *adev) uint16_t tmp, bios_header_start; r = amdgpu_atrm_get_bios(adev); - if (r == false) + if (!r) r = amdgpu_acpi_vfct_bios(adev); - if (r == false) + if (!r) r = igp_read_bios_from_vram(adev); - if (r == false) + if (!r) r = amdgpu_read_bios(adev); - if (r == false) { + if (!r) { r = amdgpu_read_bios_from_rom(adev); } - if (r == false) { + if (!r) { r = amdgpu_read_disabled_bios(adev); } - if (r == false) { + if (!r) { r = amdgpu_read_platform_bios(adev); } - if (r == false || adev->bios == NULL) { + if (!r || adev->bios == NULL) { DRM_ERROR("Unable to locate a BIOS ROM\n"); adev->bios = NULL; return false; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c index 823bf5e0b0c8..651115dcce12 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c @@ -94,6 +94,7 @@ static int amdgpu_bo_list_set(struct amdgpu_device *adev, unsigned last_entry = 0, first_userptr = num_entries; unsigned i; int r; + unsigned long total_size = 0; array = drm_malloc_ab(num_entries, sizeof(struct amdgpu_bo_list_entry)); if (!array) @@ -140,6 +141,7 @@ static int amdgpu_bo_list_set(struct amdgpu_device *adev, if (entry->robj->prefered_domains == AMDGPU_GEM_DOMAIN_OA) oa_obj = entry->robj; + total_size += amdgpu_bo_size(entry->robj); trace_amdgpu_bo_list_set(list, entry->robj); } @@ -155,6 +157,7 @@ static int amdgpu_bo_list_set(struct amdgpu_device *adev, list->array = array; list->num_entries = num_entries; + trace_amdgpu_cs_bo_status(list->num_entries, total_size); return 0; error_free: diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c index 199f76baf22c..bc0440f7a31d 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c @@ -312,6 +312,8 @@ static uint32_t amdgpu_cgs_read_ind_register(struct cgs_device *cgs_device, return RREG32_UVD_CTX(index); case CGS_IND_REG__DIDT: return RREG32_DIDT(index); + case CGS_IND_REG_GC_CAC: + return RREG32_GC_CAC(index); case CGS_IND_REG__AUDIO_ENDPT: DRM_ERROR("audio endpt register access not implemented.\n"); return 0; @@ -336,6 +338,8 @@ static void amdgpu_cgs_write_ind_register(struct cgs_device *cgs_device, return WREG32_UVD_CTX(index, value); case CGS_IND_REG__DIDT: return WREG32_DIDT(index, value); + case CGS_IND_REG_GC_CAC: + return WREG32_GC_CAC(index, value); case CGS_IND_REG__AUDIO_ENDPT: DRM_ERROR("audio endpt register access not implemented.\n"); return; @@ -696,6 +700,17 @@ static uint32_t fw_type_convert(struct cgs_device *cgs_device, uint32_t fw_type) return result; } +static int amdgpu_cgs_rel_firmware(struct cgs_device *cgs_device, enum cgs_ucode_id type) +{ + CGS_FUNC_ADEV; + if ((CGS_UCODE_ID_SMU == type) || (CGS_UCODE_ID_SMU_SK == type)) { + release_firmware(adev->pm.fw); + return 0; + } + /* cannot release other firmware because they are not created by cgs */ + return -EINVAL; +} + static int amdgpu_cgs_get_firmware_info(struct cgs_device *cgs_device, enum cgs_ucode_id type, struct cgs_firmware_info *info) @@ -737,6 +752,9 @@ static int amdgpu_cgs_get_firmware_info(struct cgs_device *cgs_device, if (!adev->pm.fw) { switch (adev->asic_type) { + case CHIP_TOPAZ: + strcpy(fw_name, "amdgpu/topaz_smc.bin"); + break; case CHIP_TONGA: strcpy(fw_name, "amdgpu/tonga_smc.bin"); break; @@ -776,6 +794,7 @@ static int amdgpu_cgs_get_firmware_info(struct cgs_device *cgs_device, } hdr = (const struct smc_firmware_header_v1_0 *) adev->pm.fw->data; + amdgpu_ucode_print_smc_hdr(&hdr->header); adev->pm.fw_version = le32_to_cpu(hdr->header.ucode_version); ucode_size = le32_to_cpu(hdr->header.ucode_size_bytes); ucode_start_address = le32_to_cpu(hdr->ucode_start_addr); @@ -784,13 +803,14 @@ static int amdgpu_cgs_get_firmware_info(struct cgs_device *cgs_device, info->version = adev->pm.fw_version; info->image_size = ucode_size; + info->ucode_start_address = ucode_start_address; info->kptr = (void *)src; } return 0; } static int amdgpu_cgs_query_system_info(struct cgs_device *cgs_device, - struct cgs_system_info *sys_info) + struct cgs_system_info *sys_info) { CGS_FUNC_ADEV; @@ -810,6 +830,12 @@ static int amdgpu_cgs_query_system_info(struct cgs_device *cgs_device, case CGS_SYSTEM_INFO_PCIE_MLW: sys_info->value = adev->pm.pcie_mlw_mask; break; + case CGS_SYSTEM_INFO_PCIE_DEV: + sys_info->value = adev->pdev->device; + break; + case CGS_SYSTEM_INFO_PCIE_REV: + sys_info->value = adev->pdev->revision; + break; case CGS_SYSTEM_INFO_CG_FLAGS: sys_info->value = adev->cg_flags; break; @@ -819,6 +845,9 @@ static int amdgpu_cgs_query_system_info(struct cgs_device *cgs_device, case CGS_SYSTEM_INFO_GFX_CU_INFO: sys_info->value = adev->gfx.cu_info.number; break; + case CGS_SYSTEM_INFO_GFX_SE_INFO: + sys_info->value = adev->gfx.config.max_shader_engines; + break; default: return -ENODEV; } @@ -892,14 +921,12 @@ static int amdgpu_cgs_acpi_eval_object(struct cgs_device *cgs_device, acpi_handle handle; struct acpi_object_list input; struct acpi_buffer output = { ACPI_ALLOCATE_BUFFER, NULL }; - union acpi_object *params = NULL; - union acpi_object *obj = NULL; + union acpi_object *params, *obj; uint8_t name[5] = {'\0'}; - struct cgs_acpi_method_argument *argument = NULL; + struct cgs_acpi_method_argument *argument; uint32_t i, count; acpi_status status; int result; - uint32_t func_no = 0xFFFFFFFF; handle = ACPI_HANDLE(&adev->pdev->dev); if (!handle) @@ -916,7 +943,6 @@ static int amdgpu_cgs_acpi_eval_object(struct cgs_device *cgs_device, if (info->pinput_argument == NULL) return -EINVAL; argument = info->pinput_argument; - func_no = argument->value; for (i = 0; i < info->input_count; i++) { if (((argument->type == ACPI_TYPE_STRING) || (argument->type == ACPI_TYPE_BUFFER)) && @@ -961,11 +987,11 @@ static int amdgpu_cgs_acpi_eval_object(struct cgs_device *cgs_device, params->integer.value = argument->value; break; case ACPI_TYPE_STRING: - params->string.length = argument->method_length; + params->string.length = argument->data_length; params->string.pointer = argument->pointer; break; case ACPI_TYPE_BUFFER: - params->buffer.length = argument->method_length; + params->buffer.length = argument->data_length; params->buffer.pointer = argument->pointer; break; default: @@ -985,7 +1011,7 @@ static int amdgpu_cgs_acpi_eval_object(struct cgs_device *cgs_device, if (ACPI_FAILURE(status)) { result = -EIO; - goto error; + goto free_input; } /* return the output info */ @@ -995,7 +1021,7 @@ static int amdgpu_cgs_acpi_eval_object(struct cgs_device *cgs_device, if ((obj->type != ACPI_TYPE_PACKAGE) || (obj->package.count != count)) { result = -EIO; - goto error; + goto free_obj; } params = obj->package.elements; } else @@ -1003,13 +1029,13 @@ static int amdgpu_cgs_acpi_eval_object(struct cgs_device *cgs_device, if (params == NULL) { result = -EIO; - goto error; + goto free_obj; } for (i = 0; i < count; i++) { if (argument->type != params->type) { result = -EIO; - goto error; + goto free_obj; } switch (params->type) { case ACPI_TYPE_INTEGER: @@ -1019,7 +1045,7 @@ static int amdgpu_cgs_acpi_eval_object(struct cgs_device *cgs_device, if ((params->string.length != argument->data_length) || (params->string.pointer == NULL)) { result = -EIO; - goto error; + goto free_obj; } strncpy(argument->pointer, params->string.pointer, @@ -1028,7 +1054,7 @@ static int amdgpu_cgs_acpi_eval_object(struct cgs_device *cgs_device, case ACPI_TYPE_BUFFER: if (params->buffer.pointer == NULL) { result = -EIO; - goto error; + goto free_obj; } memcpy(argument->pointer, params->buffer.pointer, @@ -1041,9 +1067,10 @@ static int amdgpu_cgs_acpi_eval_object(struct cgs_device *cgs_device, params++; } -error: - if (obj != NULL) - kfree(obj); + result = 0; +free_obj: + kfree(obj); +free_input: kfree((void *)input.pointer); return result; } @@ -1055,7 +1082,7 @@ static int amdgpu_cgs_acpi_eval_object(struct cgs_device *cgs_device, } #endif -int amdgpu_cgs_call_acpi_method(struct cgs_device *cgs_device, +static int amdgpu_cgs_call_acpi_method(struct cgs_device *cgs_device, uint32_t acpi_method, uint32_t acpi_function, void *pinput, void *poutput, @@ -1068,17 +1095,14 @@ int amdgpu_cgs_call_acpi_method(struct cgs_device *cgs_device, struct cgs_acpi_method_info info = {0}; acpi_input[0].type = CGS_ACPI_TYPE_INTEGER; - acpi_input[0].method_length = sizeof(uint32_t); acpi_input[0].data_length = sizeof(uint32_t); acpi_input[0].value = acpi_function; acpi_input[1].type = CGS_ACPI_TYPE_BUFFER; - acpi_input[1].method_length = CGS_ACPI_MAX_BUFFER_SIZE; acpi_input[1].data_length = input_size; acpi_input[1].pointer = pinput; acpi_output.type = CGS_ACPI_TYPE_BUFFER; - acpi_output.method_length = CGS_ACPI_MAX_BUFFER_SIZE; acpi_output.data_length = output_size; acpi_output.pointer = poutput; @@ -1125,6 +1149,7 @@ static const struct cgs_ops amdgpu_cgs_ops = { amdgpu_cgs_pm_query_clock_limits, amdgpu_cgs_set_camera_voltages, amdgpu_cgs_get_firmware_info, + amdgpu_cgs_rel_firmware, amdgpu_cgs_set_powergating_state, amdgpu_cgs_set_clockgating_state, amdgpu_cgs_get_active_displays_info, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c index cb07da41152b..ff0b55a65ca3 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c @@ -1690,7 +1690,6 @@ amdgpu_connector_add(struct amdgpu_device *adev, DRM_MODE_SCALE_NONE); /* no HPD on analog connectors */ amdgpu_connector->hpd.hpd = AMDGPU_HPD_NONE; - connector->polled = DRM_CONNECTOR_POLL_CONNECT; connector->interlace_allowed = true; connector->doublescan_allowed = true; break; @@ -1893,8 +1892,10 @@ amdgpu_connector_add(struct amdgpu_device *adev, } if (amdgpu_connector->hpd.hpd == AMDGPU_HPD_NONE) { - if (i2c_bus->valid) - connector->polled = DRM_CONNECTOR_POLL_CONNECT; + if (i2c_bus->valid) { + connector->polled = DRM_CONNECTOR_POLL_CONNECT | + DRM_CONNECTOR_POLL_DISCONNECT; + } } else connector->polled = DRM_CONNECTOR_POLL_HPD; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c index 9bc8f1d99733..0307ff5887c5 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c @@ -216,11 +216,8 @@ int amdgpu_cs_parser_init(struct amdgpu_cs_parser *p, void *data) if (ret) goto free_all_kdata; - if (p->uf_entry.robj) { - p->job->uf_bo = amdgpu_bo_ref(p->uf_entry.robj); - p->job->uf_offset = uf_offset; - } - + if (p->uf_entry.robj) + p->job->uf_addr = uf_offset; kfree(chunk_array); return 0; @@ -459,7 +456,7 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p, list_splice(&need_pages, &p->validated); } - amdgpu_vm_get_pt_bos(&fpriv->vm, &duplicates); + amdgpu_vm_get_pt_bos(p->adev, &fpriv->vm, &duplicates); p->bytes_moved_threshold = amdgpu_cs_get_threshold_for_moves(p->adev); p->bytes_moved = 0; @@ -472,6 +469,9 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p, if (r) goto error_validate; + fpriv->vm.last_eviction_counter = + atomic64_read(&p->adev->num_evictions); + if (p->bo_list) { struct amdgpu_bo *gds = p->bo_list->gds_obj; struct amdgpu_bo *gws = p->bo_list->gws_obj; @@ -499,6 +499,9 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p, } } + if (p->uf_entry.robj) + p->job->uf_addr += amdgpu_bo_gpu_offset(p->uf_entry.robj); + error_validate: if (r) { amdgpu_vm_move_pt_bos_in_lru(p->adev, &fpriv->vm); @@ -653,18 +656,21 @@ static int amdgpu_cs_ib_vm_chunk(struct amdgpu_device *adev, /* Only for UVD/VCE VM emulation */ if (ring->funcs->parse_cs) { + p->job->vm = NULL; for (i = 0; i < p->job->num_ibs; i++) { r = amdgpu_ring_parse_cs(ring, p, i); if (r) return r; } - } + } else { + p->job->vm_pd_addr = amdgpu_bo_gpu_offset(vm->page_directory); - r = amdgpu_bo_vm_update_pte(p, vm); - if (!r) - amdgpu_cs_sync_rings(p); + r = amdgpu_bo_vm_update_pte(p, vm); + if (r) + return r; + } - return r; + return amdgpu_cs_sync_rings(p); } static int amdgpu_cs_handle_lockup(struct amdgpu_device *adev, int r) @@ -761,7 +767,7 @@ static int amdgpu_cs_ib_fill(struct amdgpu_device *adev, } /* UVD & VCE fw doesn't support user fences */ - if (parser->job->uf_bo && ( + if (parser->job->uf_addr && ( parser->job->ring->type == AMDGPU_RING_TYPE_UVD || parser->job->ring->type == AMDGPU_RING_TYPE_VCE)) return -EINVAL; @@ -830,17 +836,13 @@ static int amdgpu_cs_submit(struct amdgpu_cs_parser *p, { struct amdgpu_ring *ring = p->job->ring; struct amd_sched_entity *entity = &p->ctx->rings[ring->idx].entity; - struct fence *fence; struct amdgpu_job *job; int r; job = p->job; p->job = NULL; - r = amd_sched_job_init(&job->base, &ring->sched, - entity, amdgpu_job_timeout_func, - amdgpu_job_free_func, - p->filp, &fence); + r = amd_sched_job_init(&job->base, &ring->sched, entity, p->filp); if (r) { amdgpu_job_free(job); return r; @@ -848,9 +850,10 @@ static int amdgpu_cs_submit(struct amdgpu_cs_parser *p, job->owner = p->filp; job->ctx = entity->fence_context; - p->fence = fence_get(fence); - cs->out.handle = amdgpu_ctx_add_fence(p->ctx, ring, fence); + p->fence = fence_get(&job->base.s_fence->finished); + cs->out.handle = amdgpu_ctx_add_fence(p->ctx, ring, p->fence); job->uf_sequence = cs->out.handle; + amdgpu_job_free_resources(job); trace_amdgpu_cs_ioctl(job); amd_sched_entity_push_job(&job->base); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index bb8b149786d7..df7ab2458e50 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -25,6 +25,7 @@ * Alex Deucher * Jerome Glisse */ +#include <linux/kthread.h> #include <linux/console.h> #include <linux/slab.h> #include <linux/debugfs.h> @@ -35,6 +36,7 @@ #include <linux/vga_switcheroo.h> #include <linux/efi.h> #include "amdgpu.h" +#include "amdgpu_trace.h" #include "amdgpu_i2c.h" #include "atom.h" #include "amdgpu_atombios.h" @@ -79,24 +81,27 @@ bool amdgpu_device_is_px(struct drm_device *dev) uint32_t amdgpu_mm_rreg(struct amdgpu_device *adev, uint32_t reg, bool always_indirect) { + uint32_t ret; + if ((reg * 4) < adev->rmmio_size && !always_indirect) - return readl(((void __iomem *)adev->rmmio) + (reg * 4)); + ret = readl(((void __iomem *)adev->rmmio) + (reg * 4)); else { unsigned long flags; - uint32_t ret; spin_lock_irqsave(&adev->mmio_idx_lock, flags); writel((reg * 4), ((void __iomem *)adev->rmmio) + (mmMM_INDEX * 4)); ret = readl(((void __iomem *)adev->rmmio) + (mmMM_DATA * 4)); spin_unlock_irqrestore(&adev->mmio_idx_lock, flags); - - return ret; } + trace_amdgpu_mm_rreg(adev->pdev->device, reg, ret); + return ret; } void amdgpu_mm_wreg(struct amdgpu_device *adev, uint32_t reg, uint32_t v, bool always_indirect) { + trace_amdgpu_mm_wreg(adev->pdev->device, reg, v); + if ((reg * 4) < adev->rmmio_size && !always_indirect) writel(v, ((void __iomem *)adev->rmmio) + (reg * 4)); else { @@ -827,8 +832,10 @@ static uint32_t cail_ioreg_read(struct card_info *info, uint32_t reg) */ static void amdgpu_atombios_fini(struct amdgpu_device *adev) { - if (adev->mode_info.atom_context) + if (adev->mode_info.atom_context) { kfree(adev->mode_info.atom_context->scratch); + kfree(adev->mode_info.atom_context->iio); + } kfree(adev->mode_info.atom_context); adev->mode_info.atom_context = NULL; kfree(adev->mode_info.atom_card_info); @@ -1068,11 +1075,14 @@ int amdgpu_set_clockgating_state(struct amdgpu_device *adev, int i, r = 0; for (i = 0; i < adev->num_ip_blocks; i++) { + if (!adev->ip_block_status[i].valid) + continue; if (adev->ip_blocks[i].type == block_type) { r = adev->ip_blocks[i].funcs->set_clockgating_state((void *)adev, state); if (r) return r; + break; } } return r; @@ -1085,16 +1095,53 @@ int amdgpu_set_powergating_state(struct amdgpu_device *adev, int i, r = 0; for (i = 0; i < adev->num_ip_blocks; i++) { + if (!adev->ip_block_status[i].valid) + continue; if (adev->ip_blocks[i].type == block_type) { r = adev->ip_blocks[i].funcs->set_powergating_state((void *)adev, state); if (r) return r; + break; } } return r; } +int amdgpu_wait_for_idle(struct amdgpu_device *adev, + enum amd_ip_block_type block_type) +{ + int i, r; + + for (i = 0; i < adev->num_ip_blocks; i++) { + if (!adev->ip_block_status[i].valid) + continue; + if (adev->ip_blocks[i].type == block_type) { + r = adev->ip_blocks[i].funcs->wait_for_idle((void *)adev); + if (r) + return r; + break; + } + } + return 0; + +} + +bool amdgpu_is_idle(struct amdgpu_device *adev, + enum amd_ip_block_type block_type) +{ + int i; + + for (i = 0; i < adev->num_ip_blocks; i++) { + if (!adev->ip_block_status[i].valid) + continue; + if (adev->ip_blocks[i].type == block_type) + return adev->ip_blocks[i].funcs->is_idle((void *)adev); + } + return true; + +} + const struct amdgpu_ip_block_version * amdgpu_get_ip_block( struct amdgpu_device *adev, enum amd_ip_block_type type) @@ -1207,6 +1254,9 @@ static int amdgpu_early_init(struct amdgpu_device *adev) } } + adev->cg_flags &= amdgpu_cg_mask; + adev->pg_flags &= amdgpu_pg_mask; + return 0; } @@ -1325,6 +1375,11 @@ static int amdgpu_fini(struct amdgpu_device *adev) adev->ip_block_status[i].valid = false; } + for (i = adev->num_ip_blocks - 1; i >= 0; i--) { + if (adev->ip_blocks[i].funcs->late_fini) + adev->ip_blocks[i].funcs->late_fini((void *)adev); + } + return 0; } @@ -1378,6 +1433,15 @@ static int amdgpu_resume(struct amdgpu_device *adev) return 0; } +static bool amdgpu_device_is_virtual(void) +{ +#ifdef CONFIG_X86 + return boot_cpu_has(X86_FEATURE_HYPERVISOR); +#else + return false; +#endif +} + /** * amdgpu_device_init - initialize the driver * @@ -1424,9 +1488,12 @@ int amdgpu_device_init(struct amdgpu_device *adev, adev->uvd_ctx_wreg = &amdgpu_invalid_wreg; adev->didt_rreg = &amdgpu_invalid_rreg; adev->didt_wreg = &amdgpu_invalid_wreg; + adev->gc_cac_rreg = &amdgpu_invalid_rreg; + adev->gc_cac_wreg = &amdgpu_invalid_wreg; adev->audio_endpt_rreg = &amdgpu_block_invalid_rreg; adev->audio_endpt_wreg = &amdgpu_block_invalid_wreg; + DRM_INFO("initializing kernel modesetting (%s 0x%04X:0x%04X 0x%04X:0x%04X 0x%02X).\n", amdgpu_asic_name[adev->asic_type], pdev->vendor, pdev->device, pdev->subsystem_vendor, pdev->subsystem_device, pdev->revision); @@ -1451,6 +1518,7 @@ int amdgpu_device_init(struct amdgpu_device *adev, spin_lock_init(&adev->pcie_idx_lock); spin_lock_init(&adev->uvd_ctx_idx_lock); spin_lock_init(&adev->didt_idx_lock); + spin_lock_init(&adev->gc_cac_idx_lock); spin_lock_init(&adev->audio_endpt_idx_lock); adev->rmmio_base = pci_resource_start(adev->pdev, 5); @@ -1495,29 +1563,38 @@ int amdgpu_device_init(struct amdgpu_device *adev, vga_switcheroo_init_domain_pm_ops(adev->dev, &adev->vga_pm_domain); /* Read BIOS */ - if (!amdgpu_get_bios(adev)) - return -EINVAL; + if (!amdgpu_get_bios(adev)) { + r = -EINVAL; + goto failed; + } /* Must be an ATOMBIOS */ if (!adev->is_atom_bios) { dev_err(adev->dev, "Expecting atombios for GPU\n"); - return -EINVAL; + r = -EINVAL; + goto failed; } r = amdgpu_atombios_init(adev); if (r) { dev_err(adev->dev, "amdgpu_atombios_init failed\n"); - return r; + goto failed; } /* See if the asic supports SR-IOV */ adev->virtualization.supports_sr_iov = amdgpu_atombios_has_gpu_virtualization_table(adev); + /* Check if we are executing in a virtualized environment */ + adev->virtualization.is_virtual = amdgpu_device_is_virtual(); + adev->virtualization.caps = amdgpu_asic_get_virtual_caps(adev); + /* Post card if necessary */ if (!amdgpu_card_posted(adev) || - adev->virtualization.supports_sr_iov) { + (adev->virtualization.is_virtual && + !(adev->virtualization.caps & AMDGPU_VIRT_CAPS_SRIOV_EN))) { if (!adev->bios) { dev_err(adev->dev, "Card not posted and no BIOS - ignoring\n"); - return -EINVAL; + r = -EINVAL; + goto failed; } DRM_INFO("GPU not posted. posting now...\n"); amdgpu_atom_asic_init(adev->mode_info.atom_context); @@ -1527,7 +1604,7 @@ int amdgpu_device_init(struct amdgpu_device *adev, r = amdgpu_atombios_get_clock_info(adev); if (r) { dev_err(adev->dev, "amdgpu_atombios_get_clock_info failed\n"); - return r; + goto failed; } /* init i2c buses */ amdgpu_atombios_i2c_init(adev); @@ -1536,7 +1613,7 @@ int amdgpu_device_init(struct amdgpu_device *adev, r = amdgpu_fence_driver_init(adev); if (r) { dev_err(adev->dev, "amdgpu_fence_driver_init failed\n"); - return r; + goto failed; } /* init the mode config */ @@ -1546,7 +1623,7 @@ int amdgpu_device_init(struct amdgpu_device *adev, if (r) { dev_err(adev->dev, "amdgpu_init failed\n"); amdgpu_fini(adev); - return r; + goto failed; } adev->accel_working = true; @@ -1556,7 +1633,7 @@ int amdgpu_device_init(struct amdgpu_device *adev, r = amdgpu_ib_pool_init(adev); if (r) { dev_err(adev->dev, "IB initialization failed (%d).\n", r); - return r; + goto failed; } r = amdgpu_ib_ring_tests(adev); @@ -1573,6 +1650,12 @@ int amdgpu_device_init(struct amdgpu_device *adev, DRM_ERROR("registering register debugfs failed (%d).\n", r); } + r = amdgpu_debugfs_firmware_init(adev); + if (r) { + DRM_ERROR("registering firmware debugfs failed (%d).\n", r); + return r; + } + if ((amdgpu_testing & 1)) { if (adev->accel_working) amdgpu_test_moves(adev); @@ -1598,10 +1681,15 @@ int amdgpu_device_init(struct amdgpu_device *adev, r = amdgpu_late_init(adev); if (r) { dev_err(adev->dev, "amdgpu_late_init failed\n"); - return r; + goto failed; } return 0; + +failed: + if (runtime) + vga_switcheroo_fini_domain_pm_ops(adev->dev); + return r; } static void amdgpu_debugfs_remove_files(struct amdgpu_device *adev); @@ -1624,6 +1712,7 @@ void amdgpu_device_fini(struct amdgpu_device *adev) amdgpu_bo_evict_vram(adev); amdgpu_ib_pool_fini(adev); amdgpu_fence_driver_fini(adev); + drm_crtc_force_disable_all(adev->ddev); amdgpu_fbdev_fini(adev); r = amdgpu_fini(adev); kfree(adev->ip_block_status); @@ -1635,6 +1724,8 @@ void amdgpu_device_fini(struct amdgpu_device *adev) kfree(adev->bios); adev->bios = NULL; vga_switcheroo_unregister_client(adev->pdev); + if (adev->flags & AMD_IS_PX) + vga_switcheroo_fini_domain_pm_ops(adev->dev); vga_client_register(adev->pdev, NULL, NULL, NULL); if (adev->rio_mem) pci_iounmap(adev->pdev, adev->rio_mem); @@ -1820,7 +1911,23 @@ int amdgpu_resume_kms(struct drm_device *dev, bool resume, bool fbcon) } drm_kms_helper_poll_enable(dev); + + /* + * Most of the connector probing functions try to acquire runtime pm + * refs to ensure that the GPU is powered on when connector polling is + * performed. Since we're calling this from a runtime PM callback, + * trying to acquire rpm refs will cause us to deadlock. + * + * Since we're guaranteed to be holding the rpm lock, it's safe to + * temporarily disable the rpm helpers so this doesn't deadlock us. + */ +#ifdef CONFIG_PM + dev->dev->power.disable_depth++; +#endif drm_helper_hpd_irq_event(dev); +#ifdef CONFIG_PM + dev->dev->power.disable_depth--; +#endif if (fbcon) { amdgpu_fbdev_set_suspend(adev, 0); @@ -1840,11 +1947,6 @@ int amdgpu_resume_kms(struct drm_device *dev, bool resume, bool fbcon) */ int amdgpu_gpu_reset(struct amdgpu_device *adev) { - unsigned ring_sizes[AMDGPU_MAX_RINGS]; - uint32_t *ring_data[AMDGPU_MAX_RINGS]; - - bool saved = false; - int i, r; int resched; @@ -1853,22 +1955,30 @@ int amdgpu_gpu_reset(struct amdgpu_device *adev) /* block TTM */ resched = ttm_bo_lock_delayed_workqueue(&adev->mman.bdev); - r = amdgpu_suspend(adev); - + /* block scheduler */ for (i = 0; i < AMDGPU_MAX_RINGS; ++i) { struct amdgpu_ring *ring = adev->rings[i]; + if (!ring) continue; - - ring_sizes[i] = amdgpu_ring_backup(ring, &ring_data[i]); - if (ring_sizes[i]) { - saved = true; - dev_info(adev->dev, "Saved %d dwords of commands " - "on ring %d.\n", ring_sizes[i], i); - } + kthread_park(ring->sched.thread); + amd_sched_hw_job_reset(&ring->sched); } + /* after all hw jobs are reset, hw fence is meaningless, so force_completion */ + amdgpu_fence_driver_force_completion(adev); + + /* save scratch */ + amdgpu_atombios_scratch_regs_save(adev); + r = amdgpu_suspend(adev); retry: + /* Disable fb access */ + if (adev->mode_info.num_crtc) { + struct amdgpu_mode_mc_save save; + amdgpu_display_stop_mc_access(adev, &save); + amdgpu_wait_for_idle(adev, AMD_IP_BLOCK_TYPE_GMC); + } + r = amdgpu_asic_reset(adev); /* post card */ amdgpu_atom_asic_init(adev->mode_info.atom_context); @@ -1877,32 +1987,29 @@ retry: dev_info(adev->dev, "GPU reset succeeded, trying to resume\n"); r = amdgpu_resume(adev); } - + /* restore scratch */ + amdgpu_atombios_scratch_regs_restore(adev); if (!r) { + r = amdgpu_ib_ring_tests(adev); + if (r) { + dev_err(adev->dev, "ib ring test failed (%d).\n", r); + r = amdgpu_suspend(adev); + goto retry; + } + for (i = 0; i < AMDGPU_MAX_RINGS; ++i) { struct amdgpu_ring *ring = adev->rings[i]; if (!ring) continue; - - amdgpu_ring_restore(ring, ring_sizes[i], ring_data[i]); - ring_sizes[i] = 0; - ring_data[i] = NULL; - } - - r = amdgpu_ib_ring_tests(adev); - if (r) { - dev_err(adev->dev, "ib ring test failed (%d).\n", r); - if (saved) { - saved = false; - r = amdgpu_suspend(adev); - goto retry; - } + amd_sched_job_recovery(&ring->sched); + kthread_unpark(ring->sched.thread); } } else { - amdgpu_fence_driver_force_completion(adev); + dev_err(adev->dev, "asic resume failed (%d).\n", r); for (i = 0; i < AMDGPU_MAX_RINGS; ++i) { - if (adev->rings[i]) - kfree(ring_data[i]); + if (adev->rings[i]) { + kthread_unpark(adev->rings[i]->sched.thread); + } } } @@ -1913,13 +2020,11 @@ retry: /* bad news, how to tell it to userspace ? */ dev_info(adev->dev, "GPU reset failed\n"); } + amdgpu_irq_gpu_reset_resume_helper(adev); return r; } -#define AMDGPU_DEFAULT_PCIE_GEN_MASK 0x30007 /* gen: chipset 1/2, asic 1/2/3 */ -#define AMDGPU_DEFAULT_PCIE_MLW_MASK 0x2f0000 /* 1/2/4/8/16 lanes */ - void amdgpu_get_pcie_info(struct amdgpu_device *adev) { u32 mask; @@ -2073,20 +2178,43 @@ static ssize_t amdgpu_debugfs_regs_read(struct file *f, char __user *buf, struct amdgpu_device *adev = f->f_inode->i_private; ssize_t result = 0; int r; + bool use_bank; + unsigned instance_bank, sh_bank, se_bank; if (size & 0x3 || *pos & 0x3) return -EINVAL; + if (*pos & (1ULL << 62)) { + se_bank = (*pos >> 24) & 0x3FF; + sh_bank = (*pos >> 34) & 0x3FF; + instance_bank = (*pos >> 44) & 0x3FF; + use_bank = 1; + *pos &= 0xFFFFFF; + } else { + use_bank = 0; + } + + if (use_bank) { + if (sh_bank >= adev->gfx.config.max_sh_per_se || + se_bank >= adev->gfx.config.max_shader_engines) + return -EINVAL; + mutex_lock(&adev->grbm_idx_mutex); + amdgpu_gfx_select_se_sh(adev, se_bank, + sh_bank, instance_bank); + } + while (size) { uint32_t value; if (*pos > adev->rmmio_size) - return result; + goto end; value = RREG32(*pos >> 2); r = put_user(value, (uint32_t *)buf); - if (r) - return r; + if (r) { + result = r; + goto end; + } result += 4; buf += 4; @@ -2094,6 +2222,12 @@ static ssize_t amdgpu_debugfs_regs_read(struct file *f, char __user *buf, size -= 4; } +end: + if (use_bank) { + amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); + mutex_unlock(&adev->grbm_idx_mutex); + } + return result; } @@ -2293,6 +2427,68 @@ static ssize_t amdgpu_debugfs_regs_smc_write(struct file *f, const char __user * return result; } +static ssize_t amdgpu_debugfs_gca_config_read(struct file *f, char __user *buf, + size_t size, loff_t *pos) +{ + struct amdgpu_device *adev = f->f_inode->i_private; + ssize_t result = 0; + int r; + uint32_t *config, no_regs = 0; + + if (size & 0x3 || *pos & 0x3) + return -EINVAL; + + config = kmalloc(256 * sizeof(*config), GFP_KERNEL); + if (!config) + return -ENOMEM; + + /* version, increment each time something is added */ + config[no_regs++] = 0; + config[no_regs++] = adev->gfx.config.max_shader_engines; + config[no_regs++] = adev->gfx.config.max_tile_pipes; + config[no_regs++] = adev->gfx.config.max_cu_per_sh; + config[no_regs++] = adev->gfx.config.max_sh_per_se; + config[no_regs++] = adev->gfx.config.max_backends_per_se; + config[no_regs++] = adev->gfx.config.max_texture_channel_caches; + config[no_regs++] = adev->gfx.config.max_gprs; + config[no_regs++] = adev->gfx.config.max_gs_threads; + config[no_regs++] = adev->gfx.config.max_hw_contexts; + config[no_regs++] = adev->gfx.config.sc_prim_fifo_size_frontend; + config[no_regs++] = adev->gfx.config.sc_prim_fifo_size_backend; + config[no_regs++] = adev->gfx.config.sc_hiz_tile_fifo_size; + config[no_regs++] = adev->gfx.config.sc_earlyz_tile_fifo_size; + config[no_regs++] = adev->gfx.config.num_tile_pipes; + config[no_regs++] = adev->gfx.config.backend_enable_mask; + config[no_regs++] = adev->gfx.config.mem_max_burst_length_bytes; + config[no_regs++] = adev->gfx.config.mem_row_size_in_kb; + config[no_regs++] = adev->gfx.config.shader_engine_tile_size; + config[no_regs++] = adev->gfx.config.num_gpus; + config[no_regs++] = adev->gfx.config.multi_gpu_tile_size; + config[no_regs++] = adev->gfx.config.mc_arb_ramcfg; + config[no_regs++] = adev->gfx.config.gb_addr_config; + config[no_regs++] = adev->gfx.config.num_rbs; + + while (size && (*pos < no_regs * 4)) { + uint32_t value; + + value = config[*pos >> 2]; + r = put_user(value, (uint32_t *)buf); + if (r) { + kfree(config); + return r; + } + + result += 4; + buf += 4; + *pos += 4; + size -= 4; + } + + kfree(config); + return result; +} + + static const struct file_operations amdgpu_debugfs_regs_fops = { .owner = THIS_MODULE, .read = amdgpu_debugfs_regs_read, @@ -2318,11 +2514,18 @@ static const struct file_operations amdgpu_debugfs_regs_smc_fops = { .llseek = default_llseek }; +static const struct file_operations amdgpu_debugfs_gca_config_fops = { + .owner = THIS_MODULE, + .read = amdgpu_debugfs_gca_config_read, + .llseek = default_llseek +}; + static const struct file_operations *debugfs_regs[] = { &amdgpu_debugfs_regs_fops, &amdgpu_debugfs_regs_didt_fops, &amdgpu_debugfs_regs_pcie_fops, &amdgpu_debugfs_regs_smc_fops, + &amdgpu_debugfs_gca_config_fops, }; static const char *debugfs_regs_names[] = { @@ -2330,6 +2533,7 @@ static const char *debugfs_regs_names[] = { "amdgpu_regs_didt", "amdgpu_regs_pcie", "amdgpu_regs_smc", + "amdgpu_gca_config", }; static int amdgpu_debugfs_regs_init(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c index a6eecf6f9065..76f96028313d 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c @@ -122,7 +122,7 @@ static void amdgpu_flip_work_func(struct work_struct *__work) spin_unlock_irqrestore(&crtc->dev->event_lock, flags); usleep_range(min_udelay, 2 * min_udelay); spin_lock_irqsave(&crtc->dev->event_lock, flags); - }; + } if (!repcnt) DRM_DEBUG_DRIVER("Delay problem on crtc %d: min_udelay %d, " @@ -220,19 +220,17 @@ int amdgpu_crtc_page_flip(struct drm_crtc *crtc, r = amdgpu_bo_pin_restricted(new_rbo, AMDGPU_GEM_DOMAIN_VRAM, 0, 0, &base); if (unlikely(r != 0)) { - amdgpu_bo_unreserve(new_rbo); r = -EINVAL; DRM_ERROR("failed to pin new rbo buffer before flip\n"); - goto cleanup; + goto unreserve; } r = reservation_object_get_fences_rcu(new_rbo->tbo.resv, &work->excl, &work->shared_count, &work->shared); if (unlikely(r != 0)) { - amdgpu_bo_unreserve(new_rbo); DRM_ERROR("failed to get fences for buffer\n"); - goto cleanup; + goto unpin; } amdgpu_bo_get_tiling_flags(new_rbo, &tiling_flags); @@ -275,9 +273,11 @@ pflip_cleanup: DRM_ERROR("failed to reserve new rbo in error path\n"); goto cleanup; } +unpin: if (unlikely(amdgpu_bo_unpin(new_rbo) != 0)) { DRM_ERROR("failed to unpin new rbo in error path\n"); } +unreserve: amdgpu_bo_unreserve(new_rbo); cleanup: @@ -516,9 +516,7 @@ static void amdgpu_user_framebuffer_destroy(struct drm_framebuffer *fb) { struct amdgpu_framebuffer *amdgpu_fb = to_amdgpu_framebuffer(fb); - if (amdgpu_fb->obj) { - drm_gem_object_unreference_unlocked(amdgpu_fb->obj); - } + drm_gem_object_unreference_unlocked(amdgpu_fb->obj); drm_framebuffer_cleanup(fb); kfree(amdgpu_fb); } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c index f888c015f76c..9aa533cf4ad1 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c @@ -52,9 +52,10 @@ * - 3.1.0 - allow reading more status registers (GRBM, SRBM, SDMA, CP) * - 3.2.0 - GFX8: Uses EOP_TC_WB_ACTION_EN, so UMDs don't have to do the same * at the end of IBs. + * - 3.3.0 - Add VM support for UVD on supported hardware. */ #define KMS_DRIVER_MAJOR 3 -#define KMS_DRIVER_MINOR 2 +#define KMS_DRIVER_MINOR 3 #define KMS_DRIVER_PATCHLEVEL 0 int amdgpu_vram_limit = 0; @@ -82,8 +83,12 @@ int amdgpu_exp_hw_support = 0; int amdgpu_sched_jobs = 32; int amdgpu_sched_hw_submission = 2; int amdgpu_powerplay = -1; +int amdgpu_powercontainment = 1; unsigned amdgpu_pcie_gen_cap = 0; unsigned amdgpu_pcie_lane_cap = 0; +unsigned amdgpu_cg_mask = 0xffffffff; +unsigned amdgpu_pg_mask = 0xffffffff; +char *amdgpu_disable_cu = NULL; MODULE_PARM_DESC(vramlimit, "Restrict VRAM for testing, in megabytes"); module_param_named(vramlimit, amdgpu_vram_limit, int, 0600); @@ -160,6 +165,9 @@ module_param_named(sched_hw_submission, amdgpu_sched_hw_submission, int, 0444); #ifdef CONFIG_DRM_AMD_POWERPLAY MODULE_PARM_DESC(powerplay, "Powerplay component (1 = enable, 0 = disable, -1 = auto (default))"); module_param_named(powerplay, amdgpu_powerplay, int, 0444); + +MODULE_PARM_DESC(powercontainment, "Power Containment (1 = enable (default), 0 = disable)"); +module_param_named(powercontainment, amdgpu_powercontainment, int, 0444); #endif MODULE_PARM_DESC(pcie_gen_cap, "PCIE Gen Caps (0: autodetect (default))"); @@ -168,6 +176,15 @@ module_param_named(pcie_gen_cap, amdgpu_pcie_gen_cap, uint, 0444); MODULE_PARM_DESC(pcie_lane_cap, "PCIE Lane Caps (0: autodetect (default))"); module_param_named(pcie_lane_cap, amdgpu_pcie_lane_cap, uint, 0444); +MODULE_PARM_DESC(cg_mask, "Clockgating flags mask (0 = disable clock gating)"); +module_param_named(cg_mask, amdgpu_cg_mask, uint, 0444); + +MODULE_PARM_DESC(pg_mask, "Powergating flags mask (0 = disable power gating)"); +module_param_named(pg_mask, amdgpu_pg_mask, uint, 0444); + +MODULE_PARM_DESC(disable_cu, "Disable CUs (se.sh.cu,...)"); +module_param_named(disable_cu, amdgpu_disable_cu, charp, 0444); + static const struct pci_device_id pciidlist[] = { #ifdef CONFIG_DRM_AMDGPU_CIK /* Kaveri */ @@ -413,7 +430,10 @@ static int amdgpu_pmops_runtime_suspend(struct device *dev) pci_save_state(pdev); pci_disable_device(pdev); pci_ignore_hotplug(pdev); - pci_set_power_state(pdev, PCI_D3cold); + if (amdgpu_is_atpx_hybrid()) + pci_set_power_state(pdev, PCI_D3cold); + else if (!amdgpu_has_atpx_dgpu_power_cntl()) + pci_set_power_state(pdev, PCI_D3hot); drm_dev->switch_power_state = DRM_SWITCH_POWER_DYNAMIC_OFF; return 0; @@ -430,7 +450,9 @@ static int amdgpu_pmops_runtime_resume(struct device *dev) drm_dev->switch_power_state = DRM_SWITCH_POWER_CHANGING; - pci_set_power_state(pdev, PCI_D0); + if (amdgpu_is_atpx_hybrid() || + !amdgpu_has_atpx_dgpu_power_cntl()) + pci_set_power_state(pdev, PCI_D0); pci_restore_state(pdev); ret = pci_enable_device(pdev); if (ret) @@ -515,7 +537,7 @@ static struct drm_driver kms_driver = { .driver_features = DRIVER_USE_AGP | DRIVER_HAVE_IRQ | DRIVER_IRQ_SHARED | DRIVER_GEM | - DRIVER_PRIME | DRIVER_RENDER, + DRIVER_PRIME | DRIVER_RENDER | DRIVER_MODESET, .dev_priv_size = 0, .load = amdgpu_driver_load_kms, .open = amdgpu_driver_open_kms, @@ -590,7 +612,6 @@ static int __init amdgpu_init(void) DRM_INFO("amdgpu kernel modesetting enabled.\n"); driver = &kms_driver; pdriver = &amdgpu_kms_pci_driver; - driver->driver_features |= DRIVER_MODESET; driver->num_ioctls = amdgpu_max_kms_ioctl; amdgpu_register_atpx_handler(); /* let modprobe override vga console setting */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c index d1558768cfb7..0b109aebfec6 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c @@ -204,16 +204,25 @@ void amdgpu_fence_process(struct amdgpu_ring *ring) if (seq != ring->fence_drv.sync_seq) amdgpu_fence_schedule_fallback(ring); - while (last_seq != seq) { + if (unlikely(seq == last_seq)) + return; + + last_seq &= drv->num_fences_mask; + seq &= drv->num_fences_mask; + + do { struct fence *fence, **ptr; - ptr = &drv->fences[++last_seq & drv->num_fences_mask]; + ++last_seq; + last_seq &= drv->num_fences_mask; + ptr = &drv->fences[last_seq]; /* There is always exactly one thread signaling this fence slot */ fence = rcu_dereference_protected(*ptr, 1); RCU_INIT_POINTER(*ptr, NULL); - BUG_ON(!fence); + if (!fence) + continue; r = fence_signal(fence); if (!r) @@ -222,7 +231,7 @@ void amdgpu_fence_process(struct amdgpu_ring *ring) BUG(); fence_put(fence); - } + } while (last_seq != seq); } /** diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c index 8fab6486064f..88fbed2389c0 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c @@ -503,7 +503,7 @@ static void amdgpu_gem_va_update_vm(struct amdgpu_device *adev, if (r) goto error_print; - amdgpu_vm_get_pt_bos(bo_va->vm, &duplicates); + amdgpu_vm_get_pt_bos(adev, bo_va->vm, &duplicates); list_for_each_entry(entry, &list, head) { domain = amdgpu_mem_type_to_domain(entry->bo->mem.mem_type); /* if anything is swapped out don't swap it in here, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c index 9f95da4f0536..a074edd95c70 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c @@ -70,3 +70,47 @@ void amdgpu_gfx_scratch_free(struct amdgpu_device *adev, uint32_t reg) } } } + +/** + * amdgpu_gfx_parse_disable_cu - Parse the disable_cu module parameter + * + * @mask: array in which the per-shader array disable masks will be stored + * @max_se: number of SEs + * @max_sh: number of SHs + * + * The bitmask of CUs to be disabled in the shader array determined by se and + * sh is stored in mask[se * max_sh + sh]. + */ +void amdgpu_gfx_parse_disable_cu(unsigned *mask, unsigned max_se, unsigned max_sh) +{ + unsigned se, sh, cu; + const char *p; + + memset(mask, 0, sizeof(*mask) * max_se * max_sh); + + if (!amdgpu_disable_cu || !*amdgpu_disable_cu) + return; + + p = amdgpu_disable_cu; + for (;;) { + char *next; + int ret = sscanf(p, "%u.%u.%u", &se, &sh, &cu); + if (ret < 3) { + DRM_ERROR("amdgpu: could not parse disable_cu\n"); + return; + } + + if (se < max_se && sh < max_sh && cu < 16) { + DRM_INFO("amdgpu: disabling CU %u.%u.%u\n", se, sh, cu); + mask[se * max_sh + sh] |= 1u << cu; + } else { + DRM_ERROR("amdgpu: disable_cu %u.%u.%u is out of range\n", + se, sh, cu); + } + + next = strchr(p, ','); + if (!next) + break; + p = next + 1; + } +} diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h index dc06cbda7be6..51321e154c09 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h @@ -27,4 +27,6 @@ int amdgpu_gfx_scratch_get(struct amdgpu_device *adev, uint32_t *reg); void amdgpu_gfx_scratch_free(struct amdgpu_device *adev, uint32_t reg); +unsigned amdgpu_gfx_parse_disable_cu(unsigned *mask, unsigned max_se, unsigned max_sh); + #endif diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c index 34e35423b78e..a31d7ef3032c 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c @@ -33,6 +33,8 @@ #include "amdgpu.h" #include "atom.h" +#define AMDGPU_IB_TEST_TIMEOUT msecs_to_jiffies(1000) + /* * IB * IBs (Indirect Buffers) and areas of GPU accessible memory where @@ -122,7 +124,6 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs, bool skip_preamble, need_ctx_switch; unsigned patch_offset = ~0; struct amdgpu_vm *vm; - struct fence *hwf; uint64_t ctx; unsigned i; @@ -160,10 +161,7 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs, patch_offset = amdgpu_ring_init_cond_exec(ring); if (vm) { - r = amdgpu_vm_flush(ring, job->vm_id, job->vm_pd_addr, - job->gds_base, job->gds_size, - job->gws_base, job->gws_size, - job->oa_base, job->oa_size); + r = amdgpu_vm_flush(ring, job); if (r) { amdgpu_ring_undo(ring); return r; @@ -193,7 +191,7 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs, if (ring->funcs->emit_hdp_invalidate) amdgpu_ring_emit_hdp_invalidate(ring); - r = amdgpu_fence_emit(ring, &hwf); + r = amdgpu_fence_emit(ring, f); if (r) { dev_err(adev->dev, "failed to emit fence (%d)\n", r); if (job && job->vm_id) @@ -203,17 +201,11 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs, } /* wrap the last IB with fence */ - if (job && job->uf_bo) { - uint64_t addr = amdgpu_bo_gpu_offset(job->uf_bo); - - addr += job->uf_offset; - amdgpu_ring_emit_fence(ring, addr, job->uf_sequence, + if (job && job->uf_addr) { + amdgpu_ring_emit_fence(ring, job->uf_addr, job->uf_sequence, AMDGPU_FENCE_FLAG_64BIT); } - if (f) - *f = fence_get(hwf); - if (patch_offset != ~0 && ring->funcs->patch_cond_exec) amdgpu_ring_patch_cond_exec(ring, patch_offset); @@ -296,7 +288,7 @@ int amdgpu_ib_ring_tests(struct amdgpu_device *adev) if (!ring || !ring->ready) continue; - r = amdgpu_ring_test_ib(ring); + r = amdgpu_ring_test_ib(ring, AMDGPU_IB_TEST_TIMEOUT); if (r) { ring->ready = false; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c index 835a3fa8d8df..278708f5a744 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c @@ -383,6 +383,18 @@ int amdgpu_irq_update(struct amdgpu_device *adev, return r; } +void amdgpu_irq_gpu_reset_resume_helper(struct amdgpu_device *adev) +{ + int i, j; + for (i = 0; i < AMDGPU_MAX_IRQ_SRC_ID; i++) { + struct amdgpu_irq_src *src = adev->irq.sources[i]; + if (!src) + continue; + for (j = 0; j < src->num_types; j++) + amdgpu_irq_update(adev, src, j); + } +} + /** * amdgpu_irq_get - enable interrupt * diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.h index e124b59f39c1..7ef09352e534 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.h @@ -94,6 +94,7 @@ int amdgpu_irq_put(struct amdgpu_device *adev, struct amdgpu_irq_src *src, unsigned type); bool amdgpu_irq_enabled(struct amdgpu_device *adev, struct amdgpu_irq_src *src, unsigned type); +void amdgpu_irq_gpu_reset_resume_helper(struct amdgpu_device *adev); int amdgpu_irq_add_domain(struct amdgpu_device *adev); void amdgpu_irq_remove_domain(struct amdgpu_device *adev); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c index f0dafa514fe4..6674d40eb3ab 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c @@ -28,21 +28,15 @@ #include "amdgpu.h" #include "amdgpu_trace.h" -static void amdgpu_job_free_handler(struct work_struct *ws) +static void amdgpu_job_timedout(struct amd_sched_job *s_job) { - struct amdgpu_job *job = container_of(ws, struct amdgpu_job, base.work_free_job); - amd_sched_job_put(&job->base); -} + struct amdgpu_job *job = container_of(s_job, struct amdgpu_job, base); -void amdgpu_job_timeout_func(struct work_struct *work) -{ - struct amdgpu_job *job = container_of(work, struct amdgpu_job, base.work_tdr.work); DRM_ERROR("ring %s timeout, last signaled seq=%u, last emitted seq=%u\n", - job->base.sched->name, - (uint32_t)atomic_read(&job->ring->fence_drv.last_seq), - job->ring->fence_drv.sync_seq); - - amd_sched_job_put(&job->base); + job->base.sched->name, + atomic_read(&job->ring->fence_drv.last_seq), + job->ring->fence_drv.sync_seq); + amdgpu_gpu_reset(job->adev); } int amdgpu_job_alloc(struct amdgpu_device *adev, unsigned num_ibs, @@ -63,7 +57,6 @@ int amdgpu_job_alloc(struct amdgpu_device *adev, unsigned num_ibs, (*job)->vm = vm; (*job)->ibs = (void *)&(*job)[1]; (*job)->num_ibs = num_ibs; - INIT_WORK(&(*job)->base.work_free_job, amdgpu_job_free_handler); amdgpu_sync_create(&(*job)->sync); @@ -86,27 +79,33 @@ int amdgpu_job_alloc_with_ib(struct amdgpu_device *adev, unsigned size, return r; } -void amdgpu_job_free(struct amdgpu_job *job) +void amdgpu_job_free_resources(struct amdgpu_job *job) { - unsigned i; struct fence *f; + unsigned i; + /* use sched fence if available */ - f = (job->base.s_fence)? &job->base.s_fence->base : job->fence; + f = job->base.s_fence ? &job->base.s_fence->finished : job->fence; for (i = 0; i < job->num_ibs; ++i) - amdgpu_sa_bo_free(job->adev, &job->ibs[i].sa_bo, f); - fence_put(job->fence); + amdgpu_ib_free(job->adev, &job->ibs[i], f); +} - amdgpu_bo_unref(&job->uf_bo); - amdgpu_sync_free(&job->sync); +void amdgpu_job_free_cb(struct amd_sched_job *s_job) +{ + struct amdgpu_job *job = container_of(s_job, struct amdgpu_job, base); - if (!job->base.use_sched) - kfree(job); + fence_put(job->fence); + amdgpu_sync_free(&job->sync); + kfree(job); } -void amdgpu_job_free_func(struct kref *refcount) +void amdgpu_job_free(struct amdgpu_job *job) { - struct amdgpu_job *job = container_of(refcount, struct amdgpu_job, base.refcount); + amdgpu_job_free_resources(job); + + fence_put(job->fence); + amdgpu_sync_free(&job->sync); kfree(job); } @@ -114,22 +113,20 @@ int amdgpu_job_submit(struct amdgpu_job *job, struct amdgpu_ring *ring, struct amd_sched_entity *entity, void *owner, struct fence **f) { - struct fence *fence; int r; job->ring = ring; if (!f) return -EINVAL; - r = amd_sched_job_init(&job->base, &ring->sched, - entity, amdgpu_job_timeout_func, - amdgpu_job_free_func, owner, &fence); + r = amd_sched_job_init(&job->base, &ring->sched, entity, owner); if (r) return r; job->owner = owner; job->ctx = entity->fence_context; - *f = fence_get(fence); + *f = fence_get(&job->base.s_fence->finished); + amdgpu_job_free_resources(job); amd_sched_entity_push_job(&job->base); return 0; @@ -147,8 +144,8 @@ static struct fence *amdgpu_job_dependency(struct amd_sched_job *sched_job) int r; r = amdgpu_vm_grab_id(vm, ring, &job->sync, - &job->base.s_fence->base, - &job->vm_id, &job->vm_pd_addr); + &job->base.s_fence->finished, + job); if (r) DRM_ERROR("Error getting VM ID (%d)\n", r); @@ -170,29 +167,24 @@ static struct fence *amdgpu_job_run(struct amd_sched_job *sched_job) } job = to_amdgpu_job(sched_job); - r = amdgpu_sync_wait(&job->sync); - if (r) { - DRM_ERROR("failed to sync wait (%d)\n", r); - return NULL; - } + BUG_ON(amdgpu_sync_peek_fence(&job->sync, NULL)); trace_amdgpu_sched_run_job(job); r = amdgpu_ib_schedule(job->ring, job->num_ibs, job->ibs, job->sync.last_vm_update, job, &fence); - if (r) { + if (r) DRM_ERROR("Error scheduling IBs (%d)\n", r); - goto err; - } -err: - job->fence = fence; - amdgpu_job_free(job); + /* if gpu reset, hw fence will be replaced here */ + fence_put(job->fence); + job->fence = fence_get(fence); + amdgpu_job_free_resources(job); return fence; } const struct amd_sched_backend_ops amdgpu_sched_ops = { .dependency = amdgpu_job_dependency, .run_job = amdgpu_job_run, - .begin_job = amd_sched_job_begin, - .finish_job = amd_sched_job_finish, + .timedout_job = amdgpu_job_timedout, + .free_job = amdgpu_job_free_cb }; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c index 40a23704a981..d942654a1de0 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c @@ -60,7 +60,10 @@ int amdgpu_driver_unload_kms(struct drm_device *dev) if (adev->rmmio == NULL) goto done_free; - pm_runtime_get_sync(dev->dev); + if (amdgpu_device_is_px(dev)) { + pm_runtime_get_sync(dev->dev); + pm_runtime_forbid(dev->dev); + } amdgpu_amdkfd_device_fini(adev); @@ -135,13 +138,75 @@ int amdgpu_driver_load_kms(struct drm_device *dev, unsigned long flags) } out: - if (r) + if (r) { + /* balance pm_runtime_get_sync in amdgpu_driver_unload_kms */ + if (adev->rmmio && amdgpu_device_is_px(dev)) + pm_runtime_put_noidle(dev->dev); amdgpu_driver_unload_kms(dev); - + } return r; } +static int amdgpu_firmware_info(struct drm_amdgpu_info_firmware *fw_info, + struct drm_amdgpu_query_fw *query_fw, + struct amdgpu_device *adev) +{ + switch (query_fw->fw_type) { + case AMDGPU_INFO_FW_VCE: + fw_info->ver = adev->vce.fw_version; + fw_info->feature = adev->vce.fb_version; + break; + case AMDGPU_INFO_FW_UVD: + fw_info->ver = adev->uvd.fw_version; + fw_info->feature = 0; + break; + case AMDGPU_INFO_FW_GMC: + fw_info->ver = adev->mc.fw_version; + fw_info->feature = 0; + break; + case AMDGPU_INFO_FW_GFX_ME: + fw_info->ver = adev->gfx.me_fw_version; + fw_info->feature = adev->gfx.me_feature_version; + break; + case AMDGPU_INFO_FW_GFX_PFP: + fw_info->ver = adev->gfx.pfp_fw_version; + fw_info->feature = adev->gfx.pfp_feature_version; + break; + case AMDGPU_INFO_FW_GFX_CE: + fw_info->ver = adev->gfx.ce_fw_version; + fw_info->feature = adev->gfx.ce_feature_version; + break; + case AMDGPU_INFO_FW_GFX_RLC: + fw_info->ver = adev->gfx.rlc_fw_version; + fw_info->feature = adev->gfx.rlc_feature_version; + break; + case AMDGPU_INFO_FW_GFX_MEC: + if (query_fw->index == 0) { + fw_info->ver = adev->gfx.mec_fw_version; + fw_info->feature = adev->gfx.mec_feature_version; + } else if (query_fw->index == 1) { + fw_info->ver = adev->gfx.mec2_fw_version; + fw_info->feature = adev->gfx.mec2_feature_version; + } else + return -EINVAL; + break; + case AMDGPU_INFO_FW_SMC: + fw_info->ver = adev->pm.fw_version; + fw_info->feature = 0; + break; + case AMDGPU_INFO_FW_SDMA: + if (query_fw->index >= adev->sdma.num_instances) + return -EINVAL; + fw_info->ver = adev->sdma.instance[query_fw->index].fw_version; + fw_info->feature = adev->sdma.instance[query_fw->index].feature_version; + break; + default: + return -EINVAL; + } + return 0; +} + /* * Userspace get information ioctl */ @@ -288,67 +353,20 @@ static int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file return copy_to_user(out, &count, min(size, 4u)) ? -EFAULT : 0; } case AMDGPU_INFO_TIMESTAMP: - ui64 = amdgpu_asic_get_gpu_clock_counter(adev); + ui64 = amdgpu_gfx_get_gpu_clock_counter(adev); return copy_to_user(out, &ui64, min(size, 8u)) ? -EFAULT : 0; case AMDGPU_INFO_FW_VERSION: { struct drm_amdgpu_info_firmware fw_info; + int ret; /* We only support one instance of each IP block right now. */ if (info->query_fw.ip_instance != 0) return -EINVAL; - switch (info->query_fw.fw_type) { - case AMDGPU_INFO_FW_VCE: - fw_info.ver = adev->vce.fw_version; - fw_info.feature = adev->vce.fb_version; - break; - case AMDGPU_INFO_FW_UVD: - fw_info.ver = adev->uvd.fw_version; - fw_info.feature = 0; - break; - case AMDGPU_INFO_FW_GMC: - fw_info.ver = adev->mc.fw_version; - fw_info.feature = 0; - break; - case AMDGPU_INFO_FW_GFX_ME: - fw_info.ver = adev->gfx.me_fw_version; - fw_info.feature = adev->gfx.me_feature_version; - break; - case AMDGPU_INFO_FW_GFX_PFP: - fw_info.ver = adev->gfx.pfp_fw_version; - fw_info.feature = adev->gfx.pfp_feature_version; - break; - case AMDGPU_INFO_FW_GFX_CE: - fw_info.ver = adev->gfx.ce_fw_version; - fw_info.feature = adev->gfx.ce_feature_version; - break; - case AMDGPU_INFO_FW_GFX_RLC: - fw_info.ver = adev->gfx.rlc_fw_version; - fw_info.feature = adev->gfx.rlc_feature_version; - break; - case AMDGPU_INFO_FW_GFX_MEC: - if (info->query_fw.index == 0) { - fw_info.ver = adev->gfx.mec_fw_version; - fw_info.feature = adev->gfx.mec_feature_version; - } else if (info->query_fw.index == 1) { - fw_info.ver = adev->gfx.mec2_fw_version; - fw_info.feature = adev->gfx.mec2_feature_version; - } else - return -EINVAL; - break; - case AMDGPU_INFO_FW_SMC: - fw_info.ver = adev->pm.fw_version; - fw_info.feature = 0; - break; - case AMDGPU_INFO_FW_SDMA: - if (info->query_fw.index >= adev->sdma.num_instances) - return -EINVAL; - fw_info.ver = adev->sdma.instance[info->query_fw.index].fw_version; - fw_info.feature = adev->sdma.instance[info->query_fw.index].feature_version; - break; - default: - return -EINVAL; - } + ret = amdgpu_firmware_info(&fw_info, &info->query_fw, adev); + if (ret) + return ret; + return copy_to_user(out, &fw_info, min((size_t)size, sizeof(fw_info))) ? -EFAULT : 0; } @@ -447,7 +465,8 @@ static int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file dev_info.max_memory_clock = adev->pm.default_mclk * 10; } dev_info.enabled_rb_pipes_mask = adev->gfx.config.backend_enable_mask; - dev_info.num_rb_pipes = adev->gfx.config.num_rbs; + dev_info.num_rb_pipes = adev->gfx.config.max_backends_per_se * + adev->gfx.config.max_shader_engines; dev_info.num_hw_gfx_contexts = adev->gfx.config.max_hw_contexts; dev_info._pad = 0; dev_info.ids_flags = 0; @@ -565,6 +584,9 @@ void amdgpu_driver_postclose_kms(struct drm_device *dev, amdgpu_ctx_mgr_fini(&fpriv->ctx_mgr); + amdgpu_uvd_free_handles(adev, file_priv); + amdgpu_vce_free_handles(adev, file_priv); + amdgpu_vm_fini(adev, &fpriv->vm); idr_for_each_entry(&fpriv->bo_list_handles, list, handle) @@ -589,10 +611,6 @@ void amdgpu_driver_postclose_kms(struct drm_device *dev, void amdgpu_driver_preclose_kms(struct drm_device *dev, struct drm_file *file_priv) { - struct amdgpu_device *adev = dev->dev_private; - - amdgpu_uvd_free_handles(adev, file_priv); - amdgpu_vce_free_handles(adev, file_priv); } /* @@ -755,3 +773,130 @@ const struct drm_ioctl_desc amdgpu_ioctls_kms[] = { DRM_IOCTL_DEF_DRV(AMDGPU_GEM_USERPTR, amdgpu_gem_userptr_ioctl, DRM_AUTH|DRM_RENDER_ALLOW), }; const int amdgpu_max_kms_ioctl = ARRAY_SIZE(amdgpu_ioctls_kms); + +/* + * Debugfs info + */ +#if defined(CONFIG_DEBUG_FS) + +static int amdgpu_debugfs_firmware_info(struct seq_file *m, void *data) +{ + struct drm_info_node *node = (struct drm_info_node *) m->private; + struct drm_device *dev = node->minor->dev; + struct amdgpu_device *adev = dev->dev_private; + struct drm_amdgpu_info_firmware fw_info; + struct drm_amdgpu_query_fw query_fw; + int ret, i; + + /* VCE */ + query_fw.fw_type = AMDGPU_INFO_FW_VCE; + ret = amdgpu_firmware_info(&fw_info, &query_fw, adev); + if (ret) + return ret; + seq_printf(m, "VCE feature version: %u, firmware version: 0x%08x\n", + fw_info.feature, fw_info.ver); + + /* UVD */ + query_fw.fw_type = AMDGPU_INFO_FW_UVD; + ret = amdgpu_firmware_info(&fw_info, &query_fw, adev); + if (ret) + return ret; + seq_printf(m, "UVD feature version: %u, firmware version: 0x%08x\n", + fw_info.feature, fw_info.ver); + + /* GMC */ + query_fw.fw_type = AMDGPU_INFO_FW_GMC; + ret = amdgpu_firmware_info(&fw_info, &query_fw, adev); + if (ret) + return ret; + seq_printf(m, "MC feature version: %u, firmware version: 0x%08x\n", + fw_info.feature, fw_info.ver); + + /* ME */ + query_fw.fw_type = AMDGPU_INFO_FW_GFX_ME; + ret = amdgpu_firmware_info(&fw_info, &query_fw, adev); + if (ret) + return ret; + seq_printf(m, "ME feature version: %u, firmware version: 0x%08x\n", + fw_info.feature, fw_info.ver); + + /* PFP */ + query_fw.fw_type = AMDGPU_INFO_FW_GFX_PFP; + ret = amdgpu_firmware_info(&fw_info, &query_fw, adev); + if (ret) + return ret; + seq_printf(m, "PFP feature version: %u, firmware version: 0x%08x\n", + fw_info.feature, fw_info.ver); + + /* CE */ + query_fw.fw_type = AMDGPU_INFO_FW_GFX_CE; + ret = amdgpu_firmware_info(&fw_info, &query_fw, adev); + if (ret) + return ret; + seq_printf(m, "CE feature version: %u, firmware version: 0x%08x\n", + fw_info.feature, fw_info.ver); + + /* RLC */ + query_fw.fw_type = AMDGPU_INFO_FW_GFX_RLC; + ret = amdgpu_firmware_info(&fw_info, &query_fw, adev); + if (ret) + return ret; + seq_printf(m, "RLC feature version: %u, firmware version: 0x%08x\n", + fw_info.feature, fw_info.ver); + + /* MEC */ + query_fw.fw_type = AMDGPU_INFO_FW_GFX_MEC; + query_fw.index = 0; + ret = amdgpu_firmware_info(&fw_info, &query_fw, adev); + if (ret) + return ret; + seq_printf(m, "MEC feature version: %u, firmware version: 0x%08x\n", + fw_info.feature, fw_info.ver); + + /* MEC2 */ + if (adev->asic_type == CHIP_KAVERI || + (adev->asic_type > CHIP_TOPAZ && adev->asic_type != CHIP_STONEY)) { + query_fw.index = 1; + ret = amdgpu_firmware_info(&fw_info, &query_fw, adev); + if (ret) + return ret; + seq_printf(m, "MEC2 feature version: %u, firmware version: 0x%08x\n", + fw_info.feature, fw_info.ver); + } + + /* SMC */ + query_fw.fw_type = AMDGPU_INFO_FW_SMC; + ret = amdgpu_firmware_info(&fw_info, &query_fw, adev); + if (ret) + return ret; + seq_printf(m, "SMC feature version: %u, firmware version: 0x%08x\n", + fw_info.feature, fw_info.ver); + + /* SDMA */ + query_fw.fw_type = AMDGPU_INFO_FW_SDMA; + for (i = 0; i < adev->sdma.num_instances; i++) { + query_fw.index = i; + ret = amdgpu_firmware_info(&fw_info, &query_fw, adev); + if (ret) + return ret; + seq_printf(m, "SDMA%d feature version: %u, firmware version: 0x%08x\n", + i, fw_info.feature, fw_info.ver); + } + + return 0; +} + +static const struct drm_info_list amdgpu_firmware_info_list[] = { + {"amdgpu_firmware_info", amdgpu_debugfs_firmware_info, 0, NULL}, +}; +#endif + +int amdgpu_debugfs_firmware_init(struct amdgpu_device *adev) +{ +#if defined(CONFIG_DEBUG_FS) + return amdgpu_debugfs_add_files(adev, amdgpu_firmware_info_list, + ARRAY_SIZE(amdgpu_firmware_info_list)); +#else + return 0; +#endif +} diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c index 7ecea83ce453..6f0873c75a25 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c @@ -589,6 +589,7 @@ void amdgpu_bo_move_notify(struct ttm_buffer_object *bo, struct ttm_mem_reg *new_mem) { struct amdgpu_bo *rbo; + struct ttm_mem_reg *old_mem = &bo->mem; if (!amdgpu_ttm_bo_is_amdgpu_bo(bo)) return; @@ -602,6 +603,8 @@ void amdgpu_bo_move_notify(struct ttm_buffer_object *bo, /* move_notify is called before move happens */ amdgpu_update_memory_usage(rbo->adev, &bo->mem, new_mem); + + trace_amdgpu_ttm_bo_move(rbo, new_mem->mem_type, old_mem->mem_type); } int amdgpu_bo_fault_reserve_notify(struct ttm_buffer_object *bo) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c index 589b36e8c5cf..ff63b88b0ffa 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c @@ -270,30 +270,28 @@ static ssize_t amdgpu_set_pp_force_state(struct device *dev, struct drm_device *ddev = dev_get_drvdata(dev); struct amdgpu_device *adev = ddev->dev_private; enum amd_pm_state_type state = 0; - long idx; + unsigned long idx; int ret; if (strlen(buf) == 1) adev->pp_force_state_enabled = false; - else { - ret = kstrtol(buf, 0, &idx); + else if (adev->pp_enabled) { + struct pp_states_info data; - if (ret) { + ret = kstrtoul(buf, 0, &idx); + if (ret || idx >= ARRAY_SIZE(data.states)) { count = -EINVAL; goto fail; } - if (adev->pp_enabled) { - struct pp_states_info data; - amdgpu_dpm_get_pp_num_states(adev, &data); - state = data.states[idx]; - /* only set user selected power states */ - if (state != POWER_STATE_TYPE_INTERNAL_BOOT && - state != POWER_STATE_TYPE_DEFAULT) { - amdgpu_dpm_dispatch_task(adev, - AMD_PP_EVENT_ENABLE_USER_STATE, &state, NULL); - adev->pp_force_state_enabled = true; - } + amdgpu_dpm_get_pp_num_states(adev, &data); + state = data.states[idx]; + /* only set user selected power states */ + if (state != POWER_STATE_TYPE_INTERNAL_BOOT && + state != POWER_STATE_TYPE_DEFAULT) { + amdgpu_dpm_dispatch_task(adev, + AMD_PP_EVENT_ENABLE_USER_STATE, &state, NULL); + adev->pp_force_state_enabled = true; } } fail: @@ -349,6 +347,8 @@ static ssize_t amdgpu_get_pp_dpm_sclk(struct device *dev, if (adev->pp_enabled) size = amdgpu_dpm_print_clock_levels(adev, PP_SCLK, buf); + else if (adev->pm.funcs->print_clock_levels) + size = adev->pm.funcs->print_clock_levels(adev, PP_SCLK, buf); return size; } @@ -365,7 +365,9 @@ static ssize_t amdgpu_set_pp_dpm_sclk(struct device *dev, uint32_t i, mask = 0; char sub_str[2]; - for (i = 0; i < strlen(buf) - 1; i++) { + for (i = 0; i < strlen(buf); i++) { + if (*(buf + i) == '\n') + continue; sub_str[0] = *(buf + i); sub_str[1] = '\0'; ret = kstrtol(sub_str, 0, &level); @@ -379,6 +381,8 @@ static ssize_t amdgpu_set_pp_dpm_sclk(struct device *dev, if (adev->pp_enabled) amdgpu_dpm_force_clock_level(adev, PP_SCLK, mask); + else if (adev->pm.funcs->force_clock_level) + adev->pm.funcs->force_clock_level(adev, PP_SCLK, mask); fail: return count; } @@ -393,6 +397,8 @@ static ssize_t amdgpu_get_pp_dpm_mclk(struct device *dev, if (adev->pp_enabled) size = amdgpu_dpm_print_clock_levels(adev, PP_MCLK, buf); + else if (adev->pm.funcs->print_clock_levels) + size = adev->pm.funcs->print_clock_levels(adev, PP_MCLK, buf); return size; } @@ -409,7 +415,9 @@ static ssize_t amdgpu_set_pp_dpm_mclk(struct device *dev, uint32_t i, mask = 0; char sub_str[2]; - for (i = 0; i < strlen(buf) - 1; i++) { + for (i = 0; i < strlen(buf); i++) { + if (*(buf + i) == '\n') + continue; sub_str[0] = *(buf + i); sub_str[1] = '\0'; ret = kstrtol(sub_str, 0, &level); @@ -423,6 +431,8 @@ static ssize_t amdgpu_set_pp_dpm_mclk(struct device *dev, if (adev->pp_enabled) amdgpu_dpm_force_clock_level(adev, PP_MCLK, mask); + else if (adev->pm.funcs->force_clock_level) + adev->pm.funcs->force_clock_level(adev, PP_MCLK, mask); fail: return count; } @@ -437,6 +447,8 @@ static ssize_t amdgpu_get_pp_dpm_pcie(struct device *dev, if (adev->pp_enabled) size = amdgpu_dpm_print_clock_levels(adev, PP_PCIE, buf); + else if (adev->pm.funcs->print_clock_levels) + size = adev->pm.funcs->print_clock_levels(adev, PP_PCIE, buf); return size; } @@ -453,7 +465,9 @@ static ssize_t amdgpu_set_pp_dpm_pcie(struct device *dev, uint32_t i, mask = 0; char sub_str[2]; - for (i = 0; i < strlen(buf) - 1; i++) { + for (i = 0; i < strlen(buf); i++) { + if (*(buf + i) == '\n') + continue; sub_str[0] = *(buf + i); sub_str[1] = '\0'; ret = kstrtol(sub_str, 0, &level); @@ -467,6 +481,100 @@ static ssize_t amdgpu_set_pp_dpm_pcie(struct device *dev, if (adev->pp_enabled) amdgpu_dpm_force_clock_level(adev, PP_PCIE, mask); + else if (adev->pm.funcs->force_clock_level) + adev->pm.funcs->force_clock_level(adev, PP_PCIE, mask); +fail: + return count; +} + +static ssize_t amdgpu_get_pp_sclk_od(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + struct drm_device *ddev = dev_get_drvdata(dev); + struct amdgpu_device *adev = ddev->dev_private; + uint32_t value = 0; + + if (adev->pp_enabled) + value = amdgpu_dpm_get_sclk_od(adev); + else if (adev->pm.funcs->get_sclk_od) + value = adev->pm.funcs->get_sclk_od(adev); + + return snprintf(buf, PAGE_SIZE, "%d\n", value); +} + +static ssize_t amdgpu_set_pp_sclk_od(struct device *dev, + struct device_attribute *attr, + const char *buf, + size_t count) +{ + struct drm_device *ddev = dev_get_drvdata(dev); + struct amdgpu_device *adev = ddev->dev_private; + int ret; + long int value; + + ret = kstrtol(buf, 0, &value); + + if (ret) { + count = -EINVAL; + goto fail; + } + + if (adev->pp_enabled) { + amdgpu_dpm_set_sclk_od(adev, (uint32_t)value); + amdgpu_dpm_dispatch_task(adev, AMD_PP_EVENT_READJUST_POWER_STATE, NULL, NULL); + } else if (adev->pm.funcs->set_sclk_od) { + adev->pm.funcs->set_sclk_od(adev, (uint32_t)value); + adev->pm.dpm.current_ps = adev->pm.dpm.boot_ps; + amdgpu_pm_compute_clocks(adev); + } + +fail: + return count; +} + +static ssize_t amdgpu_get_pp_mclk_od(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + struct drm_device *ddev = dev_get_drvdata(dev); + struct amdgpu_device *adev = ddev->dev_private; + uint32_t value = 0; + + if (adev->pp_enabled) + value = amdgpu_dpm_get_mclk_od(adev); + else if (adev->pm.funcs->get_mclk_od) + value = adev->pm.funcs->get_mclk_od(adev); + + return snprintf(buf, PAGE_SIZE, "%d\n", value); +} + +static ssize_t amdgpu_set_pp_mclk_od(struct device *dev, + struct device_attribute *attr, + const char *buf, + size_t count) +{ + struct drm_device *ddev = dev_get_drvdata(dev); + struct amdgpu_device *adev = ddev->dev_private; + int ret; + long int value; + + ret = kstrtol(buf, 0, &value); + + if (ret) { + count = -EINVAL; + goto fail; + } + + if (adev->pp_enabled) { + amdgpu_dpm_set_mclk_od(adev, (uint32_t)value); + amdgpu_dpm_dispatch_task(adev, AMD_PP_EVENT_READJUST_POWER_STATE, NULL, NULL); + } else if (adev->pm.funcs->set_mclk_od) { + adev->pm.funcs->set_mclk_od(adev, (uint32_t)value); + adev->pm.dpm.current_ps = adev->pm.dpm.boot_ps; + amdgpu_pm_compute_clocks(adev); + } + fail: return count; } @@ -492,6 +600,12 @@ static DEVICE_ATTR(pp_dpm_mclk, S_IRUGO | S_IWUSR, static DEVICE_ATTR(pp_dpm_pcie, S_IRUGO | S_IWUSR, amdgpu_get_pp_dpm_pcie, amdgpu_set_pp_dpm_pcie); +static DEVICE_ATTR(pp_sclk_od, S_IRUGO | S_IWUSR, + amdgpu_get_pp_sclk_od, + amdgpu_set_pp_sclk_od); +static DEVICE_ATTR(pp_mclk_od, S_IRUGO | S_IWUSR, + amdgpu_get_pp_mclk_od, + amdgpu_set_pp_mclk_od); static ssize_t amdgpu_hwmon_show_temp(struct device *dev, struct device_attribute *attr, @@ -1110,22 +1224,34 @@ int amdgpu_pm_sysfs_init(struct amdgpu_device *adev) DRM_ERROR("failed to create device file pp_table\n"); return ret; } - ret = device_create_file(adev->dev, &dev_attr_pp_dpm_sclk); - if (ret) { - DRM_ERROR("failed to create device file pp_dpm_sclk\n"); - return ret; - } - ret = device_create_file(adev->dev, &dev_attr_pp_dpm_mclk); - if (ret) { - DRM_ERROR("failed to create device file pp_dpm_mclk\n"); - return ret; - } - ret = device_create_file(adev->dev, &dev_attr_pp_dpm_pcie); - if (ret) { - DRM_ERROR("failed to create device file pp_dpm_pcie\n"); - return ret; - } } + + ret = device_create_file(adev->dev, &dev_attr_pp_dpm_sclk); + if (ret) { + DRM_ERROR("failed to create device file pp_dpm_sclk\n"); + return ret; + } + ret = device_create_file(adev->dev, &dev_attr_pp_dpm_mclk); + if (ret) { + DRM_ERROR("failed to create device file pp_dpm_mclk\n"); + return ret; + } + ret = device_create_file(adev->dev, &dev_attr_pp_dpm_pcie); + if (ret) { + DRM_ERROR("failed to create device file pp_dpm_pcie\n"); + return ret; + } + ret = device_create_file(adev->dev, &dev_attr_pp_sclk_od); + if (ret) { + DRM_ERROR("failed to create device file pp_sclk_od\n"); + return ret; + } + ret = device_create_file(adev->dev, &dev_attr_pp_mclk_od); + if (ret) { + DRM_ERROR("failed to create device file pp_mclk_od\n"); + return ret; + } + ret = amdgpu_debugfs_pm_init(adev); if (ret) { DRM_ERROR("Failed to register debugfs file for dpm!\n"); @@ -1148,10 +1274,12 @@ void amdgpu_pm_sysfs_fini(struct amdgpu_device *adev) device_remove_file(adev->dev, &dev_attr_pp_cur_state); device_remove_file(adev->dev, &dev_attr_pp_force_state); device_remove_file(adev->dev, &dev_attr_pp_table); - device_remove_file(adev->dev, &dev_attr_pp_dpm_sclk); - device_remove_file(adev->dev, &dev_attr_pp_dpm_mclk); - device_remove_file(adev->dev, &dev_attr_pp_dpm_pcie); } + device_remove_file(adev->dev, &dev_attr_pp_dpm_sclk); + device_remove_file(adev->dev, &dev_attr_pp_dpm_mclk); + device_remove_file(adev->dev, &dev_attr_pp_dpm_pcie); + device_remove_file(adev->dev, &dev_attr_pp_sclk_od); + device_remove_file(adev->dev, &dev_attr_pp_mclk_od); } void amdgpu_pm_compute_clocks(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_powerplay.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_powerplay.c index 6bd961fb43dc..c5738a22b690 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_powerplay.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_powerplay.c @@ -52,6 +52,7 @@ static int amdgpu_powerplay_init(struct amdgpu_device *adev) pp_init->chip_family = adev->family; pp_init->chip_id = adev->asic_type; pp_init->device = amdgpu_cgs_create_device(adev); + pp_init->powercontainment_enabled = amdgpu_powercontainment; ret = amd_powerplay_init(pp_init, amd_pp); kfree(pp_init); @@ -183,13 +184,6 @@ static int amdgpu_pp_sw_fini(void *handle) if (ret) return ret; -#ifdef CONFIG_DRM_AMD_POWERPLAY - if (adev->pp_enabled) { - amdgpu_pm_sysfs_fini(adev); - amd_powerplay_fini(adev->powerplay.pp_handle); - } -#endif - return ret; } @@ -223,6 +217,22 @@ static int amdgpu_pp_hw_fini(void *handle) return ret; } +static void amdgpu_pp_late_fini(void *handle) +{ +#ifdef CONFIG_DRM_AMD_POWERPLAY + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + + if (adev->pp_enabled) { + amdgpu_pm_sysfs_fini(adev); + amd_powerplay_fini(adev->powerplay.pp_handle); + } + + if (adev->powerplay.ip_funcs->late_fini) + adev->powerplay.ip_funcs->late_fini( + adev->powerplay.pp_handle); +#endif +} + static int amdgpu_pp_suspend(void *handle) { int ret = 0; @@ -311,6 +321,7 @@ const struct amd_ip_funcs amdgpu_pp_ip_funcs = { .sw_fini = amdgpu_pp_sw_fini, .hw_init = amdgpu_pp_hw_init, .hw_fini = amdgpu_pp_hw_fini, + .late_fini = amdgpu_pp_late_fini, .suspend = amdgpu_pp_suspend, .resume = amdgpu_pp_resume, .is_idle = amdgpu_pp_is_idle, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c index 3b02272db678..85aeb0a804bb 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c @@ -28,6 +28,7 @@ */ #include <linux/seq_file.h> #include <linux/slab.h> +#include <linux/debugfs.h> #include <drm/drmP.h> #include <drm/amdgpu_drm.h> #include "amdgpu.h" @@ -48,6 +49,7 @@ */ static int amdgpu_debugfs_ring_init(struct amdgpu_device *adev, struct amdgpu_ring *ring); +static void amdgpu_debugfs_ring_fini(struct amdgpu_ring *ring); /** * amdgpu_ring_alloc - allocate space on the ring buffer @@ -73,6 +75,10 @@ int amdgpu_ring_alloc(struct amdgpu_ring *ring, unsigned ndw) ring->count_dw = ndw; ring->wptr_old = ring->wptr; + + if (ring->funcs->begin_use) + ring->funcs->begin_use(ring); + return 0; } @@ -125,6 +131,9 @@ void amdgpu_ring_commit(struct amdgpu_ring *ring) mb(); amdgpu_ring_set_wptr(ring); + + if (ring->funcs->end_use) + ring->funcs->end_use(ring); } /** @@ -137,78 +146,9 @@ void amdgpu_ring_commit(struct amdgpu_ring *ring) void amdgpu_ring_undo(struct amdgpu_ring *ring) { ring->wptr = ring->wptr_old; -} - -/** - * amdgpu_ring_backup - Back up the content of a ring - * - * @ring: the ring we want to back up - * - * Saves all unprocessed commits from a ring, returns the number of dwords saved. - */ -unsigned amdgpu_ring_backup(struct amdgpu_ring *ring, - uint32_t **data) -{ - unsigned size, ptr, i; - - *data = NULL; - - if (ring->ring_obj == NULL) - return 0; - - /* it doesn't make sense to save anything if all fences are signaled */ - if (!amdgpu_fence_count_emitted(ring)) - return 0; - - ptr = le32_to_cpu(*ring->next_rptr_cpu_addr); - - size = ring->wptr + (ring->ring_size / 4); - size -= ptr; - size &= ring->ptr_mask; - if (size == 0) - return 0; - - /* and then save the content of the ring */ - *data = kmalloc_array(size, sizeof(uint32_t), GFP_KERNEL); - if (!*data) - return 0; - for (i = 0; i < size; ++i) { - (*data)[i] = ring->ring[ptr++]; - ptr &= ring->ptr_mask; - } - - return size; -} - -/** - * amdgpu_ring_restore - append saved commands to the ring again - * - * @ring: ring to append commands to - * @size: number of dwords we want to write - * @data: saved commands - * - * Allocates space on the ring and restore the previously saved commands. - */ -int amdgpu_ring_restore(struct amdgpu_ring *ring, - unsigned size, uint32_t *data) -{ - int i, r; - - if (!size || !data) - return 0; - - /* restore the saved ring content */ - r = amdgpu_ring_alloc(ring, size); - if (r) - return r; - - for (i = 0; i < size; ++i) { - amdgpu_ring_write(ring, data[i]); - } - amdgpu_ring_commit(ring); - kfree(data); - return 0; + if (ring->funcs->end_use) + ring->funcs->end_use(ring); } /** @@ -260,14 +200,6 @@ int amdgpu_ring_init(struct amdgpu_device *adev, struct amdgpu_ring *ring, return r; } - r = amdgpu_wb_get(adev, &ring->next_rptr_offs); - if (r) { - dev_err(adev->dev, "(%d) ring next_rptr wb alloc failed\n", r); - return r; - } - ring->next_rptr_gpu_addr = adev->wb.gpu_addr + ring->next_rptr_offs * 4; - ring->next_rptr_cpu_addr = &adev->wb.wb[ring->next_rptr_offs]; - r = amdgpu_wb_get(adev, &ring->cond_exe_offs); if (r) { dev_err(adev->dev, "(%d) ring cond_exec_polling wb alloc failed\n", r); @@ -276,7 +208,6 @@ int amdgpu_ring_init(struct amdgpu_device *adev, struct amdgpu_ring *ring, ring->cond_exe_gpu_addr = adev->wb.gpu_addr + (ring->cond_exe_offs * 4); ring->cond_exe_cpu_addr = &adev->wb.wb[ring->cond_exe_offs]; - spin_lock_init(&ring->fence_lock); r = amdgpu_fence_driver_start_ring(ring, irq_src, irq_type); if (r) { dev_err(adev->dev, "failed initializing fences (%d).\n", r); @@ -310,6 +241,9 @@ int amdgpu_ring_init(struct amdgpu_device *adev, struct amdgpu_ring *ring, } r = amdgpu_bo_kmap(ring->ring_obj, (void **)&ring->ring); + + memset((void *)ring->ring, 0, ring->ring_size); + amdgpu_bo_unreserve(ring->ring_obj); if (r) { dev_err(adev->dev, "(%d) ring map failed\n", r); @@ -343,10 +277,10 @@ void amdgpu_ring_fini(struct amdgpu_ring *ring) ring->ring = NULL; ring->ring_obj = NULL; + amdgpu_wb_free(ring->adev, ring->cond_exe_offs); amdgpu_wb_free(ring->adev, ring->fence_offs); amdgpu_wb_free(ring->adev, ring->rptr_offs); amdgpu_wb_free(ring->adev, ring->wptr_offs); - amdgpu_wb_free(ring->adev, ring->next_rptr_offs); if (ring_obj) { r = amdgpu_bo_reserve(ring_obj, false); @@ -357,6 +291,7 @@ void amdgpu_ring_fini(struct amdgpu_ring *ring) } amdgpu_bo_unref(&ring_obj); } + amdgpu_debugfs_ring_fini(ring); } /* @@ -364,57 +299,62 @@ void amdgpu_ring_fini(struct amdgpu_ring *ring) */ #if defined(CONFIG_DEBUG_FS) -static int amdgpu_debugfs_ring_info(struct seq_file *m, void *data) +/* Layout of file is 12 bytes consisting of + * - rptr + * - wptr + * - driver's copy of wptr + * + * followed by n-words of ring data + */ +static ssize_t amdgpu_debugfs_ring_read(struct file *f, char __user *buf, + size_t size, loff_t *pos) { - struct drm_info_node *node = (struct drm_info_node *) m->private; - struct drm_device *dev = node->minor->dev; - struct amdgpu_device *adev = dev->dev_private; - int roffset = (unsigned long)node->info_ent->data; - struct amdgpu_ring *ring = (void *)(((uint8_t*)adev) + roffset); - uint32_t rptr, wptr, rptr_next; - unsigned i; - - wptr = amdgpu_ring_get_wptr(ring); - seq_printf(m, "wptr: 0x%08x [%5d]\n", wptr, wptr); - - rptr = amdgpu_ring_get_rptr(ring); - rptr_next = le32_to_cpu(*ring->next_rptr_cpu_addr); - - seq_printf(m, "rptr: 0x%08x [%5d]\n", rptr, rptr); - - seq_printf(m, "driver's copy of the wptr: 0x%08x [%5d]\n", - ring->wptr, ring->wptr); - - if (!ring->ready) - return 0; - - /* print 8 dw before current rptr as often it's the last executed - * packet that is the root issue - */ - i = (rptr + ring->ptr_mask + 1 - 32) & ring->ptr_mask; - while (i != rptr) { - seq_printf(m, "r[%5d]=0x%08x", i, ring->ring[i]); - if (i == rptr) - seq_puts(m, " *"); - if (i == rptr_next) - seq_puts(m, " #"); - seq_puts(m, "\n"); - i = (i + 1) & ring->ptr_mask; + struct amdgpu_ring *ring = (struct amdgpu_ring*)f->f_inode->i_private; + int r, i; + uint32_t value, result, early[3]; + + if (*pos & 3 || size & 3) + return -EINVAL; + + result = 0; + + if (*pos < 12) { + early[0] = amdgpu_ring_get_rptr(ring); + early[1] = amdgpu_ring_get_wptr(ring); + early[2] = ring->wptr; + for (i = *pos / 4; i < 3 && size; i++) { + r = put_user(early[i], (uint32_t *)buf); + if (r) + return r; + buf += 4; + result += 4; + size -= 4; + *pos += 4; + } } - while (i != wptr) { - seq_printf(m, "r[%5d]=0x%08x", i, ring->ring[i]); - if (i == rptr) - seq_puts(m, " *"); - if (i == rptr_next) - seq_puts(m, " #"); - seq_puts(m, "\n"); - i = (i + 1) & ring->ptr_mask; + + while (size) { + if (*pos >= (ring->ring_size + 12)) + return result; + + value = ring->ring[(*pos - 12)/4]; + r = put_user(value, (uint32_t*)buf); + if (r) + return r; + buf += 4; + result += 4; + size -= 4; + *pos += 4; } - return 0; + + return result; } -static struct drm_info_list amdgpu_debugfs_ring_info_list[AMDGPU_MAX_RINGS]; -static char amdgpu_debugfs_ring_names[AMDGPU_MAX_RINGS][32]; +static const struct file_operations amdgpu_debugfs_ring_fops = { + .owner = THIS_MODULE, + .read = amdgpu_debugfs_ring_read, + .llseek = default_llseek +}; #endif @@ -422,28 +362,27 @@ static int amdgpu_debugfs_ring_init(struct amdgpu_device *adev, struct amdgpu_ring *ring) { #if defined(CONFIG_DEBUG_FS) - unsigned offset = (uint8_t*)ring - (uint8_t*)adev; - unsigned i; - struct drm_info_list *info; - char *name; - - for (i = 0; i < ARRAY_SIZE(amdgpu_debugfs_ring_info_list); ++i) { - info = &amdgpu_debugfs_ring_info_list[i]; - if (!info->data) - break; - } + struct drm_minor *minor = adev->ddev->primary; + struct dentry *ent, *root = minor->debugfs_root; + char name[32]; - if (i == ARRAY_SIZE(amdgpu_debugfs_ring_info_list)) - return -ENOSPC; - - name = &amdgpu_debugfs_ring_names[i][0]; sprintf(name, "amdgpu_ring_%s", ring->name); - info->name = name; - info->show = amdgpu_debugfs_ring_info; - info->driver_features = 0; - info->data = (void*)(uintptr_t)offset; - return amdgpu_debugfs_add_files(adev, info, 1); + ent = debugfs_create_file(name, + S_IFREG | S_IRUGO, root, + ring, &amdgpu_debugfs_ring_fops); + if (IS_ERR(ent)) + return PTR_ERR(ent); + + i_size_write(ent->d_inode, ring->ring_size + 12); + ring->ent = ent; #endif return 0; } + +static void amdgpu_debugfs_ring_fini(struct amdgpu_ring *ring) +{ +#if defined(CONFIG_DEBUG_FS) + debugfs_remove(ring->ent); +#endif +} diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sa.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_sa.c index b16366c2b4a0..d8af37a845f4 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sa.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sa.c @@ -115,6 +115,7 @@ int amdgpu_sa_bo_manager_start(struct amdgpu_device *adev, return r; } r = amdgpu_bo_kmap(sa_manager->bo, &sa_manager->cpu_ptr); + memset(sa_manager->cpu_ptr, 0, sa_manager->size); amdgpu_bo_unreserve(sa_manager->bo); return r; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c index 34a92808bbd4..5c8d3022fb87 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c @@ -223,13 +223,16 @@ int amdgpu_sync_resv(struct amdgpu_device *adev, } /** - * amdgpu_sync_is_idle - test if all fences are signaled + * amdgpu_sync_peek_fence - get the next fence not signaled yet * * @sync: the sync object + * @ring: optional ring to use for test * - * Returns true if all fences in the sync object are signaled. + * Returns the next fence not signaled yet without removing it from the sync + * object. */ -bool amdgpu_sync_is_idle(struct amdgpu_sync *sync) +struct fence *amdgpu_sync_peek_fence(struct amdgpu_sync *sync, + struct amdgpu_ring *ring) { struct amdgpu_sync_entry *e; struct hlist_node *tmp; @@ -237,6 +240,19 @@ bool amdgpu_sync_is_idle(struct amdgpu_sync *sync) hash_for_each_safe(sync->fences, i, tmp, e, node) { struct fence *f = e->fence; + struct amd_sched_fence *s_fence = to_amd_sched_fence(f); + + if (ring && s_fence) { + /* For fences from the same ring it is sufficient + * when they are scheduled. + */ + if (s_fence->sched == &ring->sched) { + if (fence_is_signaled(&s_fence->scheduled)) + continue; + + return &s_fence->scheduled; + } + } if (fence_is_signaled(f)) { hash_del(&e->node); @@ -245,58 +261,19 @@ bool amdgpu_sync_is_idle(struct amdgpu_sync *sync) continue; } - return false; + return f; } - return true; + return NULL; } /** - * amdgpu_sync_cycle_fences - move fences from one sync object into another + * amdgpu_sync_get_fence - get the next fence from the sync object * - * @dst: the destination sync object - * @src: the source sync object - * @fence: fence to add to source + * @sync: sync object to use * - * Remove all fences from source and put them into destination and add - * fence as new one into source. + * Get and removes the next fence from the sync object not signaled yet. */ -int amdgpu_sync_cycle_fences(struct amdgpu_sync *dst, struct amdgpu_sync *src, - struct fence *fence) -{ - struct amdgpu_sync_entry *e, *newone; - struct hlist_node *tmp; - int i; - - /* Allocate the new entry before moving the old ones */ - newone = kmem_cache_alloc(amdgpu_sync_slab, GFP_KERNEL); - if (!newone) - return -ENOMEM; - - hash_for_each_safe(src->fences, i, tmp, e, node) { - struct fence *f = e->fence; - - hash_del(&e->node); - if (fence_is_signaled(f)) { - fence_put(f); - kmem_cache_free(amdgpu_sync_slab, e); - continue; - } - - if (amdgpu_sync_add_later(dst, f)) { - kmem_cache_free(amdgpu_sync_slab, e); - continue; - } - - hash_add(dst->fences, &e->node, f->context); - } - - hash_add(src->fences, &newone->node, fence->context); - newone->fence = fence_get(fence); - - return 0; -} - struct fence *amdgpu_sync_get_fence(struct amdgpu_sync *sync) { struct amdgpu_sync_entry *e; @@ -319,25 +296,6 @@ struct fence *amdgpu_sync_get_fence(struct amdgpu_sync *sync) return NULL; } -int amdgpu_sync_wait(struct amdgpu_sync *sync) -{ - struct amdgpu_sync_entry *e; - struct hlist_node *tmp; - int i, r; - - hash_for_each_safe(sync->fences, i, tmp, e, node) { - r = fence_wait(e->fence, false); - if (r) - return r; - - hash_del(&e->node); - fence_put(e->fence); - kmem_cache_free(amdgpu_sync_slab, e); - } - - return 0; -} - /** * amdgpu_sync_free - free the sync object * diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h index 26a5f4acf584..0d8d65eb46cd 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h @@ -11,19 +11,68 @@ #define TRACE_SYSTEM amdgpu #define TRACE_INCLUDE_FILE amdgpu_trace +TRACE_EVENT(amdgpu_mm_rreg, + TP_PROTO(unsigned did, uint32_t reg, uint32_t value), + TP_ARGS(did, reg, value), + TP_STRUCT__entry( + __field(unsigned, did) + __field(uint32_t, reg) + __field(uint32_t, value) + ), + TP_fast_assign( + __entry->did = did; + __entry->reg = reg; + __entry->value = value; + ), + TP_printk("0x%04lx, 0x%04lx, 0x%08lx", + (unsigned long)__entry->did, + (unsigned long)__entry->reg, + (unsigned long)__entry->value) +); + +TRACE_EVENT(amdgpu_mm_wreg, + TP_PROTO(unsigned did, uint32_t reg, uint32_t value), + TP_ARGS(did, reg, value), + TP_STRUCT__entry( + __field(unsigned, did) + __field(uint32_t, reg) + __field(uint32_t, value) + ), + TP_fast_assign( + __entry->did = did; + __entry->reg = reg; + __entry->value = value; + ), + TP_printk("0x%04lx, 0x%04lx, 0x%08lx", + (unsigned long)__entry->did, + (unsigned long)__entry->reg, + (unsigned long)__entry->value) +); + TRACE_EVENT(amdgpu_bo_create, TP_PROTO(struct amdgpu_bo *bo), TP_ARGS(bo), TP_STRUCT__entry( __field(struct amdgpu_bo *, bo) __field(u32, pages) + __field(u32, type) + __field(u32, prefer) + __field(u32, allow) + __field(u32, visible) ), TP_fast_assign( __entry->bo = bo; __entry->pages = bo->tbo.num_pages; + __entry->type = bo->tbo.mem.mem_type; + __entry->prefer = bo->prefered_domains; + __entry->allow = bo->allowed_domains; + __entry->visible = bo->flags; ), - TP_printk("bo=%p, pages=%u", __entry->bo, __entry->pages) + + TP_printk("bo=%p,pages=%u,type=%d,prefered=%d,allowed=%d,visible=%d", + __entry->bo, __entry->pages, __entry->type, + __entry->prefer, __entry->allow, __entry->visible) ); TRACE_EVENT(amdgpu_cs, @@ -64,7 +113,7 @@ TRACE_EVENT(amdgpu_cs_ioctl, __entry->adev = job->adev; __entry->sched_job = &job->base; __entry->ib = job->ibs; - __entry->fence = &job->base.s_fence->base; + __entry->fence = &job->base.s_fence->finished; __entry->ring_name = job->ring->name; __entry->num_ibs = job->num_ibs; ), @@ -89,7 +138,7 @@ TRACE_EVENT(amdgpu_sched_run_job, __entry->adev = job->adev; __entry->sched_job = &job->base; __entry->ib = job->ibs; - __entry->fence = &job->base.s_fence->base; + __entry->fence = &job->base.s_fence->finished; __entry->ring_name = job->ring->name; __entry->num_ibs = job->num_ibs; ), @@ -100,24 +149,26 @@ TRACE_EVENT(amdgpu_sched_run_job, TRACE_EVENT(amdgpu_vm_grab_id, - TP_PROTO(struct amdgpu_vm *vm, int ring, unsigned vmid, - uint64_t pd_addr), - TP_ARGS(vm, ring, vmid, pd_addr), + TP_PROTO(struct amdgpu_vm *vm, int ring, struct amdgpu_job *job), + TP_ARGS(vm, ring, job), TP_STRUCT__entry( __field(struct amdgpu_vm *, vm) __field(u32, ring) __field(u32, vmid) __field(u64, pd_addr) + __field(u32, needs_flush) ), TP_fast_assign( __entry->vm = vm; __entry->ring = ring; - __entry->vmid = vmid; - __entry->pd_addr = pd_addr; + __entry->vmid = job->vm_id; + __entry->pd_addr = job->vm_pd_addr; + __entry->needs_flush = job->vm_needs_flush; ), - TP_printk("vm=%p, ring=%u, id=%u, pd_addr=%010Lx", __entry->vm, - __entry->ring, __entry->vmid, __entry->pd_addr) + TP_printk("vm=%p, ring=%u, id=%u, pd_addr=%010Lx needs_flush=%u", + __entry->vm, __entry->ring, __entry->vmid, + __entry->pd_addr, __entry->needs_flush) ); TRACE_EVENT(amdgpu_vm_bo_map, @@ -244,13 +295,55 @@ TRACE_EVENT(amdgpu_bo_list_set, TP_STRUCT__entry( __field(struct amdgpu_bo_list *, list) __field(struct amdgpu_bo *, bo) + __field(u64, bo_size) ), TP_fast_assign( __entry->list = list; __entry->bo = bo; + __entry->bo_size = amdgpu_bo_size(bo); ), - TP_printk("list=%p, bo=%p", __entry->list, __entry->bo) + TP_printk("list=%p, bo=%p, bo_size = %Ld", + __entry->list, + __entry->bo, + __entry->bo_size) +); + +TRACE_EVENT(amdgpu_cs_bo_status, + TP_PROTO(uint64_t total_bo, uint64_t total_size), + TP_ARGS(total_bo, total_size), + TP_STRUCT__entry( + __field(u64, total_bo) + __field(u64, total_size) + ), + + TP_fast_assign( + __entry->total_bo = total_bo; + __entry->total_size = total_size; + ), + TP_printk("total bo size = %Ld, total bo count = %Ld", + __entry->total_bo, __entry->total_size) +); + +TRACE_EVENT(amdgpu_ttm_bo_move, + TP_PROTO(struct amdgpu_bo* bo, uint32_t new_placement, uint32_t old_placement), + TP_ARGS(bo, new_placement, old_placement), + TP_STRUCT__entry( + __field(struct amdgpu_bo *, bo) + __field(u64, bo_size) + __field(u32, new_placement) + __field(u32, old_placement) + ), + + TP_fast_assign( + __entry->bo = bo; + __entry->bo_size = amdgpu_bo_size(bo); + __entry->new_placement = new_placement; + __entry->old_placement = old_placement; + ), + TP_printk("bo=%p from:%d to %d with size = %Ld", + __entry->bo, __entry->old_placement, + __entry->new_placement, __entry->bo_size) ); #endif diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c index 3b9053af4762..b7742e62972a 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c @@ -286,9 +286,10 @@ static int amdgpu_move_blit(struct ttm_buffer_object *bo, r = amdgpu_copy_buffer(ring, old_start, new_start, new_mem->num_pages * PAGE_SIZE, /* bytes */ bo->resv, &fence); - /* FIXME: handle copy error */ - r = ttm_bo_move_accel_cleanup(bo, fence, - evict, no_wait_gpu, new_mem); + if (r) + return r; + + r = ttm_bo_pipeline_move(bo, fence, evict, new_mem); fence_put(fence); return r; } @@ -396,6 +397,11 @@ static int amdgpu_bo_move(struct ttm_buffer_object *bo, return -EINVAL; adev = amdgpu_get_adev(bo->bdev); + + /* remember the eviction */ + if (evict) + atomic64_inc(&adev->num_evictions); + if (old_mem->mem_type == TTM_PL_SYSTEM && bo->ttm == NULL) { amdgpu_move_null(bo, new_mem); return 0; @@ -429,7 +435,8 @@ static int amdgpu_bo_move(struct ttm_buffer_object *bo, if (r) { memcpy: - r = ttm_bo_move_memcpy(bo, evict, no_wait_gpu, new_mem); + r = ttm_bo_move_memcpy(bo, evict, interruptible, + no_wait_gpu, new_mem); if (r) { return r; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c index 01abfc21b4a2..b11f4e8868d7 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c @@ -40,9 +40,16 @@ #include "uvd/uvd_4_2_d.h" /* 1 second timeout */ -#define UVD_IDLE_TIMEOUT_MS 1000 +#define UVD_IDLE_TIMEOUT msecs_to_jiffies(1000) + +/* Firmware versions for VI */ +#define FW_1_65_10 ((1 << 24) | (65 << 16) | (10 << 8)) +#define FW_1_87_11 ((1 << 24) | (87 << 16) | (11 << 8)) +#define FW_1_87_12 ((1 << 24) | (87 << 16) | (12 << 8)) +#define FW_1_37_15 ((1 << 24) | (37 << 16) | (15 << 8)) + /* Polaris10/11 firmware version */ -#define FW_1_66_16 ((1 << 24) | (66 << 16) | (16 << 8)) +#define FW_1_66_16 ((1 << 24) | (66 << 16) | (16 << 8)) /* Firmware Names */ #ifdef CONFIG_DRM_AMDGPU_CIK @@ -92,7 +99,6 @@ MODULE_FIRMWARE(FIRMWARE_STONEY); MODULE_FIRMWARE(FIRMWARE_POLARIS10); MODULE_FIRMWARE(FIRMWARE_POLARIS11); -static void amdgpu_uvd_note_usage(struct amdgpu_device *adev); static void amdgpu_uvd_idle_work_handler(struct work_struct *work); int amdgpu_uvd_sw_init(struct amdgpu_device *adev) @@ -246,6 +252,23 @@ int amdgpu_uvd_sw_init(struct amdgpu_device *adev) if (!amdgpu_ip_block_version_cmp(adev, AMD_IP_BLOCK_TYPE_UVD, 5, 0)) adev->uvd.address_64_bit = true; + switch (adev->asic_type) { + case CHIP_TONGA: + adev->uvd.use_ctx_buf = adev->uvd.fw_version >= FW_1_65_10; + break; + case CHIP_CARRIZO: + adev->uvd.use_ctx_buf = adev->uvd.fw_version >= FW_1_87_11; + break; + case CHIP_FIJI: + adev->uvd.use_ctx_buf = adev->uvd.fw_version >= FW_1_87_12; + break; + case CHIP_STONEY: + adev->uvd.use_ctx_buf = adev->uvd.fw_version >= FW_1_37_15; + break; + default: + adev->uvd.use_ctx_buf = adev->asic_type >= CHIP_POLARIS10; + } + return 0; } @@ -253,19 +276,20 @@ int amdgpu_uvd_sw_fini(struct amdgpu_device *adev) { int r; - if (adev->uvd.vcpu_bo == NULL) - return 0; + kfree(adev->uvd.saved_bo); amd_sched_entity_fini(&adev->uvd.ring.sched, &adev->uvd.entity); - r = amdgpu_bo_reserve(adev->uvd.vcpu_bo, false); - if (!r) { - amdgpu_bo_kunmap(adev->uvd.vcpu_bo); - amdgpu_bo_unpin(adev->uvd.vcpu_bo); - amdgpu_bo_unreserve(adev->uvd.vcpu_bo); - } + if (adev->uvd.vcpu_bo) { + r = amdgpu_bo_reserve(adev->uvd.vcpu_bo, false); + if (!r) { + amdgpu_bo_kunmap(adev->uvd.vcpu_bo); + amdgpu_bo_unpin(adev->uvd.vcpu_bo); + amdgpu_bo_unreserve(adev->uvd.vcpu_bo); + } - amdgpu_bo_unref(&adev->uvd.vcpu_bo); + amdgpu_bo_unref(&adev->uvd.vcpu_bo); + } amdgpu_ring_fini(&adev->uvd.ring); @@ -345,8 +369,6 @@ void amdgpu_uvd_free_handles(struct amdgpu_device *adev, struct drm_file *filp) if (handle != 0 && adev->uvd.filp[i] == filp) { struct fence *fence; - amdgpu_uvd_note_usage(adev); - r = amdgpu_uvd_get_destroy_msg(ring, handle, false, &fence); if (r) { @@ -437,7 +459,7 @@ static int amdgpu_uvd_cs_msg_decode(struct amdgpu_device *adev, uint32_t *msg, unsigned fs_in_mb = width_in_mb * height_in_mb; unsigned image_size, tmp, min_dpb_size, num_dpb_buffer; - unsigned min_ctx_size = 0; + unsigned min_ctx_size = ~0; image_size = width * height; image_size += image_size / 2; @@ -556,7 +578,7 @@ static int amdgpu_uvd_cs_msg_decode(struct amdgpu_device *adev, uint32_t *msg, /* reference picture buffer */ min_dpb_size = image_size * num_dpb_buffer; - if (adev->asic_type < CHIP_POLARIS10){ + if (!adev->uvd.use_ctx_buf){ /* macroblock context buffer */ min_dpb_size += width_in_mb * height_in_mb * num_dpb_buffer * 192; @@ -661,7 +683,7 @@ static int amdgpu_uvd_cs_msg(struct amdgpu_uvd_cs_ctx *ctx, } DRM_ERROR("No more free UVD handles!\n"); - return -EINVAL; + return -ENOSPC; case 1: /* it's a decode msg, calc buffer sizes */ @@ -912,8 +934,6 @@ int amdgpu_uvd_ring_parse_cs(struct amdgpu_cs_parser *parser, uint32_t ib_idx) return -EINVAL; } - amdgpu_uvd_note_usage(ctx.parser->adev); - return 0; } @@ -967,7 +987,7 @@ static int amdgpu_uvd_send_msg(struct amdgpu_ring *ring, struct amdgpu_bo *bo, if (direct) { r = amdgpu_ib_schedule(ring, 1, ib, NULL, NULL, &f); - job->fence = f; + job->fence = fence_get(f); if (r) goto err_free; @@ -1109,16 +1129,14 @@ static void amdgpu_uvd_idle_work_handler(struct work_struct *work) amdgpu_asic_set_uvd_clocks(adev, 0, 0); } } else { - schedule_delayed_work(&adev->uvd.idle_work, - msecs_to_jiffies(UVD_IDLE_TIMEOUT_MS)); + schedule_delayed_work(&adev->uvd.idle_work, UVD_IDLE_TIMEOUT); } } -static void amdgpu_uvd_note_usage(struct amdgpu_device *adev) +void amdgpu_uvd_ring_begin_use(struct amdgpu_ring *ring) { + struct amdgpu_device *adev = ring->adev; bool set_clocks = !cancel_delayed_work_sync(&adev->uvd.idle_work); - set_clocks &= schedule_delayed_work(&adev->uvd.idle_work, - msecs_to_jiffies(UVD_IDLE_TIMEOUT_MS)); if (set_clocks) { if (adev->pm.dpm_enabled) { @@ -1128,3 +1146,48 @@ static void amdgpu_uvd_note_usage(struct amdgpu_device *adev) } } } + +void amdgpu_uvd_ring_end_use(struct amdgpu_ring *ring) +{ + schedule_delayed_work(&ring->adev->uvd.idle_work, UVD_IDLE_TIMEOUT); +} + +/** + * amdgpu_uvd_ring_test_ib - test ib execution + * + * @ring: amdgpu_ring pointer + * + * Test if we can successfully execute an IB + */ +int amdgpu_uvd_ring_test_ib(struct amdgpu_ring *ring, long timeout) +{ + struct fence *fence; + long r; + + r = amdgpu_uvd_get_create_msg(ring, 1, NULL); + if (r) { + DRM_ERROR("amdgpu: failed to get create msg (%ld).\n", r); + goto error; + } + + r = amdgpu_uvd_get_destroy_msg(ring, 1, true, &fence); + if (r) { + DRM_ERROR("amdgpu: failed to get destroy ib (%ld).\n", r); + goto error; + } + + r = fence_wait_timeout(fence, false, timeout); + if (r == 0) { + DRM_ERROR("amdgpu: IB test timed out.\n"); + r = -ETIMEDOUT; + } else if (r < 0) { + DRM_ERROR("amdgpu: fence wait failed (%ld).\n", r); + } else { + DRM_INFO("ib test on ring %d succeeded\n", ring->idx); + r = 0; + } + +error: + fence_put(fence); + return r; +} diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.h index 9a3b449081a7..c850009602d1 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.h @@ -35,5 +35,8 @@ int amdgpu_uvd_get_destroy_msg(struct amdgpu_ring *ring, uint32_t handle, void amdgpu_uvd_free_handles(struct amdgpu_device *adev, struct drm_file *filp); int amdgpu_uvd_ring_parse_cs(struct amdgpu_cs_parser *parser, uint32_t ib_idx); +void amdgpu_uvd_ring_begin_use(struct amdgpu_ring *ring); +void amdgpu_uvd_ring_end_use(struct amdgpu_ring *ring); +int amdgpu_uvd_ring_test_ib(struct amdgpu_ring *ring, long timeout); #endif diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c index 875626a2eccb..05865ce35351 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c @@ -36,7 +36,7 @@ #include "cikd.h" /* 1 second timeout */ -#define VCE_IDLE_TIMEOUT_MS 1000 +#define VCE_IDLE_TIMEOUT msecs_to_jiffies(1000) /* Firmware Names */ #ifdef CONFIG_DRM_AMDGPU_CIK @@ -85,8 +85,6 @@ int amdgpu_vce_sw_init(struct amdgpu_device *adev, unsigned long size) unsigned ucode_version, version_major, version_minor, binary_id; int i, r; - INIT_DELAYED_WORK(&adev->vce.idle_work, amdgpu_vce_idle_work_handler); - switch (adev->asic_type) { #ifdef CONFIG_DRM_AMDGPU_CIK case CHIP_BONAIRE: @@ -197,6 +195,9 @@ int amdgpu_vce_sw_init(struct amdgpu_device *adev, unsigned long size) adev->vce.filp[i] = NULL; } + INIT_DELAYED_WORK(&adev->vce.idle_work, amdgpu_vce_idle_work_handler); + mutex_init(&adev->vce.idle_mutex); + return 0; } @@ -220,6 +221,7 @@ int amdgpu_vce_sw_fini(struct amdgpu_device *adev) amdgpu_ring_fini(&adev->vce.ring[1]); release_firmware(adev->vce.fw); + mutex_destroy(&adev->vce.idle_mutex); return 0; } @@ -310,37 +312,44 @@ static void amdgpu_vce_idle_work_handler(struct work_struct *work) amdgpu_asic_set_vce_clocks(adev, 0, 0); } } else { - schedule_delayed_work(&adev->vce.idle_work, - msecs_to_jiffies(VCE_IDLE_TIMEOUT_MS)); + schedule_delayed_work(&adev->vce.idle_work, VCE_IDLE_TIMEOUT); } } /** - * amdgpu_vce_note_usage - power up VCE + * amdgpu_vce_ring_begin_use - power up VCE * - * @adev: amdgpu_device pointer + * @ring: amdgpu ring * * Make sure VCE is powerd up when we want to use it */ -static void amdgpu_vce_note_usage(struct amdgpu_device *adev) +void amdgpu_vce_ring_begin_use(struct amdgpu_ring *ring) { - bool streams_changed = false; - bool set_clocks = !cancel_delayed_work_sync(&adev->vce.idle_work); - set_clocks &= schedule_delayed_work(&adev->vce.idle_work, - msecs_to_jiffies(VCE_IDLE_TIMEOUT_MS)); - - if (adev->pm.dpm_enabled) { - /* XXX figure out if the streams changed */ - streams_changed = false; - } + struct amdgpu_device *adev = ring->adev; + bool set_clocks; - if (set_clocks || streams_changed) { + mutex_lock(&adev->vce.idle_mutex); + set_clocks = !cancel_delayed_work_sync(&adev->vce.idle_work); + if (set_clocks) { if (adev->pm.dpm_enabled) { amdgpu_dpm_enable_vce(adev, true); } else { amdgpu_asic_set_vce_clocks(adev, 53300, 40000); } } + mutex_unlock(&adev->vce.idle_mutex); +} + +/** + * amdgpu_vce_ring_end_use - power VCE down + * + * @ring: amdgpu ring + * + * Schedule work to power VCE down again + */ +void amdgpu_vce_ring_end_use(struct amdgpu_ring *ring) +{ + schedule_delayed_work(&ring->adev->vce.idle_work, VCE_IDLE_TIMEOUT); } /** @@ -357,11 +366,10 @@ void amdgpu_vce_free_handles(struct amdgpu_device *adev, struct drm_file *filp) int i, r; for (i = 0; i < AMDGPU_MAX_VCE_HANDLES; ++i) { uint32_t handle = atomic_read(&adev->vce.handles[i]); + if (!handle || adev->vce.filp[i] != filp) continue; - amdgpu_vce_note_usage(adev); - r = amdgpu_vce_get_destroy_msg(ring, handle, false, NULL); if (r) DRM_ERROR("Error destroying VCE handle (%d)!\n", r); @@ -437,7 +445,7 @@ int amdgpu_vce_get_create_msg(struct amdgpu_ring *ring, uint32_t handle, ib->ptr[i] = 0x0; r = amdgpu_ib_schedule(ring, 1, ib, NULL, NULL, &f); - job->fence = f; + job->fence = fence_get(f); if (r) goto err; @@ -469,7 +477,6 @@ int amdgpu_vce_get_destroy_msg(struct amdgpu_ring *ring, uint32_t handle, struct amdgpu_job *job; struct amdgpu_ib *ib; struct fence *f = NULL; - uint64_t dummy; int i, r; r = amdgpu_job_alloc_with_ib(ring->adev, ib_size_dw * 4, &job); @@ -477,7 +484,6 @@ int amdgpu_vce_get_destroy_msg(struct amdgpu_ring *ring, uint32_t handle, return r; ib = &job->ibs[0]; - dummy = ib->gpu_addr + 1024; /* stitch together an VCE destroy msg */ ib->length_dw = 0; @@ -485,11 +491,14 @@ int amdgpu_vce_get_destroy_msg(struct amdgpu_ring *ring, uint32_t handle, ib->ptr[ib->length_dw++] = 0x00000001; /* session cmd */ ib->ptr[ib->length_dw++] = handle; - ib->ptr[ib->length_dw++] = 0x00000014; /* len */ - ib->ptr[ib->length_dw++] = 0x05000005; /* feedback buffer */ - ib->ptr[ib->length_dw++] = upper_32_bits(dummy); - ib->ptr[ib->length_dw++] = dummy; - ib->ptr[ib->length_dw++] = 0x00000001; + ib->ptr[ib->length_dw++] = 0x00000020; /* len */ + ib->ptr[ib->length_dw++] = 0x00000002; /* task info */ + ib->ptr[ib->length_dw++] = 0xffffffff; /* next task info, set to 0xffffffff if no */ + ib->ptr[ib->length_dw++] = 0x00000001; /* destroy session */ + ib->ptr[ib->length_dw++] = 0x00000000; + ib->ptr[ib->length_dw++] = 0x00000000; + ib->ptr[ib->length_dw++] = 0xffffffff; /* feedback is not needed, set to 0xffffffff and firmware will not output feedback */ + ib->ptr[ib->length_dw++] = 0x00000000; ib->ptr[ib->length_dw++] = 0x00000008; /* len */ ib->ptr[ib->length_dw++] = 0x02000001; /* destroy cmd */ @@ -499,7 +508,7 @@ int amdgpu_vce_get_destroy_msg(struct amdgpu_ring *ring, uint32_t handle, if (direct) { r = amdgpu_ib_schedule(ring, 1, ib, NULL, NULL, &f); - job->fence = f; + job->fence = fence_get(f); if (r) goto err; @@ -580,12 +589,10 @@ static int amdgpu_vce_cs_reloc(struct amdgpu_cs_parser *p, uint32_t ib_idx, * we we don't have another free session index. */ static int amdgpu_vce_validate_handle(struct amdgpu_cs_parser *p, - uint32_t handle, bool *allocated) + uint32_t handle, uint32_t *allocated) { unsigned i; - *allocated = false; - /* validate the handle */ for (i = 0; i < AMDGPU_MAX_VCE_HANDLES; ++i) { if (atomic_read(&p->adev->vce.handles[i]) == handle) { @@ -602,7 +609,7 @@ static int amdgpu_vce_validate_handle(struct amdgpu_cs_parser *p, if (!atomic_cmpxchg(&p->adev->vce.handles[i], 0, handle)) { p->adev->vce.filp[i] = p->filp; p->adev->vce.img_size[i] = 0; - *allocated = true; + *allocated |= 1 << i; return i; } } @@ -622,15 +629,13 @@ int amdgpu_vce_ring_parse_cs(struct amdgpu_cs_parser *p, uint32_t ib_idx) struct amdgpu_ib *ib = &p->job->ibs[ib_idx]; unsigned fb_idx = 0, bs_idx = 0; int session_idx = -1; - bool destroyed = false; - bool created = false; - bool allocated = false; + uint32_t destroyed = 0; + uint32_t created = 0; + uint32_t allocated = 0; uint32_t tmp, handle = 0; uint32_t *size = &tmp; int i, r = 0, idx = 0; - amdgpu_vce_note_usage(p->adev); - while (idx < ib->length_dw) { uint32_t len = amdgpu_get_ib_value(p, ib_idx, idx); uint32_t cmd = amdgpu_get_ib_value(p, ib_idx, idx + 1); @@ -641,30 +646,30 @@ int amdgpu_vce_ring_parse_cs(struct amdgpu_cs_parser *p, uint32_t ib_idx) goto out; } - if (destroyed) { - DRM_ERROR("No other command allowed after destroy!\n"); - r = -EINVAL; - goto out; - } - switch (cmd) { - case 0x00000001: // session + case 0x00000001: /* session */ handle = amdgpu_get_ib_value(p, ib_idx, idx + 2); session_idx = amdgpu_vce_validate_handle(p, handle, &allocated); - if (session_idx < 0) - return session_idx; + if (session_idx < 0) { + r = session_idx; + goto out; + } size = &p->adev->vce.img_size[session_idx]; break; - case 0x00000002: // task info + case 0x00000002: /* task info */ fb_idx = amdgpu_get_ib_value(p, ib_idx, idx + 6); bs_idx = amdgpu_get_ib_value(p, ib_idx, idx + 7); break; - case 0x01000001: // create - created = true; - if (!allocated) { + case 0x01000001: /* create */ + created |= 1 << session_idx; + if (destroyed & (1 << session_idx)) { + destroyed &= ~(1 << session_idx); + allocated |= 1 << session_idx; + + } else if (!(allocated & (1 << session_idx))) { DRM_ERROR("Handle already in use!\n"); r = -EINVAL; goto out; @@ -675,16 +680,16 @@ int amdgpu_vce_ring_parse_cs(struct amdgpu_cs_parser *p, uint32_t ib_idx) 8 * 3 / 2; break; - case 0x04000001: // config extension - case 0x04000002: // pic control - case 0x04000005: // rate control - case 0x04000007: // motion estimation - case 0x04000008: // rdo - case 0x04000009: // vui - case 0x05000002: // auxiliary buffer + case 0x04000001: /* config extension */ + case 0x04000002: /* pic control */ + case 0x04000005: /* rate control */ + case 0x04000007: /* motion estimation */ + case 0x04000008: /* rdo */ + case 0x04000009: /* vui */ + case 0x05000002: /* auxiliary buffer */ break; - case 0x03000001: // encode + case 0x03000001: /* encode */ r = amdgpu_vce_cs_reloc(p, ib_idx, idx + 10, idx + 9, *size, 0); if (r) @@ -696,18 +701,18 @@ int amdgpu_vce_ring_parse_cs(struct amdgpu_cs_parser *p, uint32_t ib_idx) goto out; break; - case 0x02000001: // destroy - destroyed = true; + case 0x02000001: /* destroy */ + destroyed |= 1 << session_idx; break; - case 0x05000001: // context buffer + case 0x05000001: /* context buffer */ r = amdgpu_vce_cs_reloc(p, ib_idx, idx + 3, idx + 2, *size * 2, 0); if (r) goto out; break; - case 0x05000004: // video bitstream buffer + case 0x05000004: /* video bitstream buffer */ tmp = amdgpu_get_ib_value(p, ib_idx, idx + 4); r = amdgpu_vce_cs_reloc(p, ib_idx, idx + 3, idx + 2, tmp, bs_idx); @@ -715,7 +720,7 @@ int amdgpu_vce_ring_parse_cs(struct amdgpu_cs_parser *p, uint32_t ib_idx) goto out; break; - case 0x05000005: // feedback buffer + case 0x05000005: /* feedback buffer */ r = amdgpu_vce_cs_reloc(p, ib_idx, idx + 3, idx + 2, 4096, fb_idx); if (r) @@ -737,21 +742,24 @@ int amdgpu_vce_ring_parse_cs(struct amdgpu_cs_parser *p, uint32_t ib_idx) idx += len / 4; } - if (allocated && !created) { + if (allocated & ~created) { DRM_ERROR("New session without create command!\n"); r = -ENOENT; } out: - if ((!r && destroyed) || (r && allocated)) { - /* - * IB contains a destroy msg or we have allocated an - * handle and got an error, anyway free the handle - */ - for (i = 0; i < AMDGPU_MAX_VCE_HANDLES; ++i) - atomic_cmpxchg(&p->adev->vce.handles[i], handle, 0); + if (!r) { + /* No error, free all destroyed handle slots */ + tmp = destroyed; + } else { + /* Error during parsing, free all allocated handle slots */ + tmp = allocated; } + for (i = 0; i < AMDGPU_MAX_VCE_HANDLES; ++i) + if (tmp & (1 << i)) + atomic_set(&p->adev->vce.handles[i], 0); + return r; } @@ -837,10 +845,10 @@ int amdgpu_vce_ring_test_ring(struct amdgpu_ring *ring) * @ring: the engine to test on * */ -int amdgpu_vce_ring_test_ib(struct amdgpu_ring *ring) +int amdgpu_vce_ring_test_ib(struct amdgpu_ring *ring, long timeout) { struct fence *fence = NULL; - int r; + long r; /* skip vce ring1 ib test for now, since it's not reliable */ if (ring == &ring->adev->vce.ring[1]) @@ -848,21 +856,25 @@ int amdgpu_vce_ring_test_ib(struct amdgpu_ring *ring) r = amdgpu_vce_get_create_msg(ring, 1, NULL); if (r) { - DRM_ERROR("amdgpu: failed to get create msg (%d).\n", r); + DRM_ERROR("amdgpu: failed to get create msg (%ld).\n", r); goto error; } r = amdgpu_vce_get_destroy_msg(ring, 1, true, &fence); if (r) { - DRM_ERROR("amdgpu: failed to get destroy ib (%d).\n", r); + DRM_ERROR("amdgpu: failed to get destroy ib (%ld).\n", r); goto error; } - r = fence_wait(fence, false); - if (r) { - DRM_ERROR("amdgpu: fence wait failed (%d).\n", r); + r = fence_wait_timeout(fence, false, timeout); + if (r == 0) { + DRM_ERROR("amdgpu: IB test timed out.\n"); + r = -ETIMEDOUT; + } else if (r < 0) { + DRM_ERROR("amdgpu: fence wait failed (%ld).\n", r); } else { DRM_INFO("ib test on ring %d succeeded\n", ring->idx); + r = 0; } error: fence_put(fence); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.h index f40cf761c66f..63f83d0d985c 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.h @@ -39,6 +39,8 @@ void amdgpu_vce_ring_emit_ib(struct amdgpu_ring *ring, struct amdgpu_ib *ib, void amdgpu_vce_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 seq, unsigned flags); int amdgpu_vce_ring_test_ring(struct amdgpu_ring *ring); -int amdgpu_vce_ring_test_ib(struct amdgpu_ring *ring); +int amdgpu_vce_ring_test_ib(struct amdgpu_ring *ring, long timeout); +void amdgpu_vce_ring_begin_use(struct amdgpu_ring *ring); +void amdgpu_vce_ring_end_use(struct amdgpu_ring *ring); #endif diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c index 9f36ed30ba11..8e642fc48df4 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c @@ -25,6 +25,7 @@ * Alex Deucher * Jerome Glisse */ +#include <linux/fence-array.h> #include <drm/drmP.h> #include <drm/amdgpu_drm.h> #include "amdgpu.h" @@ -114,16 +115,26 @@ void amdgpu_vm_get_pd_bo(struct amdgpu_vm *vm, /** * amdgpu_vm_get_bos - add the vm BOs to a duplicates list * + * @adev: amdgpu device pointer * @vm: vm providing the BOs * @duplicates: head of duplicates list * * Add the page directory to the BO duplicates list * for command submission. */ -void amdgpu_vm_get_pt_bos(struct amdgpu_vm *vm, struct list_head *duplicates) +void amdgpu_vm_get_pt_bos(struct amdgpu_device *adev, struct amdgpu_vm *vm, + struct list_head *duplicates) { + uint64_t num_evictions; unsigned i; + /* We only need to validate the page tables + * if they aren't already valid. + */ + num_evictions = atomic64_read(&adev->num_evictions); + if (num_evictions == vm->last_eviction_counter) + return; + /* add the vm page table to the list */ for (i = 0; i <= vm->max_pde_used; ++i) { struct amdgpu_bo_list_entry *entry = &vm->page_tables[i].entry; @@ -162,6 +173,13 @@ void amdgpu_vm_move_pt_bos_in_lru(struct amdgpu_device *adev, spin_unlock(&glob->lru_lock); } +static bool amdgpu_vm_is_gpu_reset(struct amdgpu_device *adev, + struct amdgpu_vm_id *id) +{ + return id->current_gpu_reset_count != + atomic_read(&adev->gpu_reset_counter) ? true : false; +} + /** * amdgpu_vm_grab_id - allocate the next free VMID * @@ -174,18 +192,67 @@ void amdgpu_vm_move_pt_bos_in_lru(struct amdgpu_device *adev, */ int amdgpu_vm_grab_id(struct amdgpu_vm *vm, struct amdgpu_ring *ring, struct amdgpu_sync *sync, struct fence *fence, - unsigned *vm_id, uint64_t *vm_pd_addr) + struct amdgpu_job *job) { - uint64_t pd_addr = amdgpu_bo_gpu_offset(vm->page_directory); struct amdgpu_device *adev = ring->adev; + uint64_t fence_context = adev->fence_context + ring->idx; struct fence *updates = sync->last_vm_update; - struct amdgpu_vm_id *id; - unsigned i = ring->idx; - int r; + struct amdgpu_vm_id *id, *idle; + struct fence **fences; + unsigned i; + int r = 0; + + fences = kmalloc_array(sizeof(void *), adev->vm_manager.num_ids, + GFP_KERNEL); + if (!fences) + return -ENOMEM; mutex_lock(&adev->vm_manager.lock); + /* Check if we have an idle VMID */ + i = 0; + list_for_each_entry(idle, &adev->vm_manager.ids_lru, list) { + fences[i] = amdgpu_sync_peek_fence(&idle->active, ring); + if (!fences[i]) + break; + ++i; + } + + /* If we can't find a idle VMID to use, wait till one becomes available */ + if (&idle->list == &adev->vm_manager.ids_lru) { + u64 fence_context = adev->vm_manager.fence_context + ring->idx; + unsigned seqno = ++adev->vm_manager.seqno[ring->idx]; + struct fence_array *array; + unsigned j; + + for (j = 0; j < i; ++j) + fence_get(fences[j]); + + array = fence_array_create(i, fences, fence_context, + seqno, true); + if (!array) { + for (j = 0; j < i; ++j) + fence_put(fences[j]); + kfree(fences); + r = -ENOMEM; + goto error; + } + + + r = amdgpu_sync_fence(ring->adev, sync, &array->base); + fence_put(&array->base); + if (r) + goto error; + + mutex_unlock(&adev->vm_manager.lock); + return 0; + + } + kfree(fences); + + job->vm_needs_flush = true; /* Check if we can use a VMID already assigned to this VM */ + i = ring->idx; do { struct fence *flushed; @@ -196,67 +263,52 @@ int amdgpu_vm_grab_id(struct amdgpu_vm *vm, struct amdgpu_ring *ring, /* Check all the prerequisites to using this VMID */ if (!id) continue; + if (amdgpu_vm_is_gpu_reset(adev, id)) + continue; if (atomic64_read(&id->owner) != vm->client_id) continue; - if (pd_addr != id->pd_gpu_addr) + if (job->vm_pd_addr != id->pd_gpu_addr) continue; - if (id->last_user != ring && - (!id->last_flush || !fence_is_signaled(id->last_flush))) + if (!id->last_flush) continue; - flushed = id->flushed_updates; - if (updates && (!flushed || fence_is_later(updates, flushed))) + if (id->last_flush->context != fence_context && + !fence_is_signaled(id->last_flush)) continue; - /* Good we can use this VMID */ - if (id->last_user == ring) { - r = amdgpu_sync_fence(ring->adev, sync, - id->first); - if (r) - goto error; - } + flushed = id->flushed_updates; + if (updates && + (!flushed || fence_is_later(updates, flushed))) + continue; - /* And remember this submission as user of the VMID */ + /* Good we can use this VMID. Remember this submission as + * user of the VMID. + */ r = amdgpu_sync_fence(ring->adev, &id->active, fence); if (r) goto error; + id->current_gpu_reset_count = atomic_read(&adev->gpu_reset_counter); list_move_tail(&id->list, &adev->vm_manager.ids_lru); vm->ids[ring->idx] = id; - *vm_id = id - adev->vm_manager.ids; - *vm_pd_addr = AMDGPU_VM_NO_FLUSH; - trace_amdgpu_vm_grab_id(vm, ring->idx, *vm_id, *vm_pd_addr); + job->vm_id = id - adev->vm_manager.ids; + job->vm_needs_flush = false; + trace_amdgpu_vm_grab_id(vm, ring->idx, job); mutex_unlock(&adev->vm_manager.lock); return 0; } while (i != ring->idx); - id = list_first_entry(&adev->vm_manager.ids_lru, - struct amdgpu_vm_id, - list); - - if (!amdgpu_sync_is_idle(&id->active)) { - struct list_head *head = &adev->vm_manager.ids_lru; - struct amdgpu_vm_id *tmp; + /* Still no ID to use? Then use the idle one found earlier */ + id = idle; - list_for_each_entry_safe(id, tmp, &adev->vm_manager.ids_lru, - list) { - if (amdgpu_sync_is_idle(&id->active)) { - list_move(&id->list, head); - head = &id->list; - } - } - id = list_first_entry(&adev->vm_manager.ids_lru, - struct amdgpu_vm_id, - list); - } - - r = amdgpu_sync_cycle_fences(sync, &id->active, fence); + /* Remember this submission as user of the VMID */ + r = amdgpu_sync_fence(ring->adev, &id->active, fence); if (r) goto error; @@ -269,22 +321,46 @@ int amdgpu_vm_grab_id(struct amdgpu_vm *vm, struct amdgpu_ring *ring, fence_put(id->flushed_updates); id->flushed_updates = fence_get(updates); - id->pd_gpu_addr = pd_addr; - + id->pd_gpu_addr = job->vm_pd_addr; + id->current_gpu_reset_count = atomic_read(&adev->gpu_reset_counter); list_move_tail(&id->list, &adev->vm_manager.ids_lru); - id->last_user = ring; atomic64_set(&id->owner, vm->client_id); vm->ids[ring->idx] = id; - *vm_id = id - adev->vm_manager.ids; - *vm_pd_addr = pd_addr; - trace_amdgpu_vm_grab_id(vm, ring->idx, *vm_id, *vm_pd_addr); + job->vm_id = id - adev->vm_manager.ids; + trace_amdgpu_vm_grab_id(vm, ring->idx, job); error: mutex_unlock(&adev->vm_manager.lock); return r; } +static bool amdgpu_vm_ring_has_compute_vm_bug(struct amdgpu_ring *ring) +{ + struct amdgpu_device *adev = ring->adev; + const struct amdgpu_ip_block_version *ip_block; + + if (ring->type != AMDGPU_RING_TYPE_COMPUTE) + /* only compute rings */ + return false; + + ip_block = amdgpu_get_ip_block(adev, AMD_IP_BLOCK_TYPE_GFX); + if (!ip_block) + return false; + + if (ip_block->major <= 7) { + /* gfx7 has no workaround */ + return true; + } else if (ip_block->major == 8) { + if (adev->gfx.mec_fw_version >= 673) + /* gfx8 is fixed in MEC firmware 673 */ + return false; + else + return true; + } + return false; +} + /** * amdgpu_vm_flush - hardware flush the vm * @@ -294,59 +370,52 @@ error: * * Emit a VM flush when it is necessary. */ -int amdgpu_vm_flush(struct amdgpu_ring *ring, - unsigned vm_id, uint64_t pd_addr, - uint32_t gds_base, uint32_t gds_size, - uint32_t gws_base, uint32_t gws_size, - uint32_t oa_base, uint32_t oa_size) +int amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job) { struct amdgpu_device *adev = ring->adev; - struct amdgpu_vm_id *id = &adev->vm_manager.ids[vm_id]; + struct amdgpu_vm_id *id = &adev->vm_manager.ids[job->vm_id]; bool gds_switch_needed = ring->funcs->emit_gds_switch && ( - id->gds_base != gds_base || - id->gds_size != gds_size || - id->gws_base != gws_base || - id->gws_size != gws_size || - id->oa_base != oa_base || - id->oa_size != oa_size); + id->gds_base != job->gds_base || + id->gds_size != job->gds_size || + id->gws_base != job->gws_base || + id->gws_size != job->gws_size || + id->oa_base != job->oa_base || + id->oa_size != job->oa_size); int r; if (ring->funcs->emit_pipeline_sync && ( - pd_addr != AMDGPU_VM_NO_FLUSH || gds_switch_needed || - ring->type == AMDGPU_RING_TYPE_COMPUTE)) + job->vm_needs_flush || gds_switch_needed || + amdgpu_vm_ring_has_compute_vm_bug(ring))) amdgpu_ring_emit_pipeline_sync(ring); - if (ring->funcs->emit_vm_flush && - pd_addr != AMDGPU_VM_NO_FLUSH) { + if (ring->funcs->emit_vm_flush && (job->vm_needs_flush || + amdgpu_vm_is_gpu_reset(adev, id))) { struct fence *fence; - trace_amdgpu_vm_flush(pd_addr, ring->idx, vm_id); - amdgpu_ring_emit_vm_flush(ring, vm_id, pd_addr); + trace_amdgpu_vm_flush(job->vm_pd_addr, ring->idx, job->vm_id); + amdgpu_ring_emit_vm_flush(ring, job->vm_id, job->vm_pd_addr); + + r = amdgpu_fence_emit(ring, &fence); + if (r) + return r; mutex_lock(&adev->vm_manager.lock); - if ((id->pd_gpu_addr == pd_addr) && (id->last_user == ring)) { - r = amdgpu_fence_emit(ring, &fence); - if (r) { - mutex_unlock(&adev->vm_manager.lock); - return r; - } - fence_put(id->last_flush); - id->last_flush = fence; - } + fence_put(id->last_flush); + id->last_flush = fence; mutex_unlock(&adev->vm_manager.lock); } if (gds_switch_needed) { - id->gds_base = gds_base; - id->gds_size = gds_size; - id->gws_base = gws_base; - id->gws_size = gws_size; - id->oa_base = oa_base; - id->oa_size = oa_size; - amdgpu_ring_emit_gds_switch(ring, vm_id, - gds_base, gds_size, - gws_base, gws_size, - oa_base, oa_size); + id->gds_base = job->gds_base; + id->gds_size = job->gds_size; + id->gws_base = job->gws_base; + id->gws_size = job->gws_size; + id->oa_base = job->oa_base; + id->oa_size = job->oa_size; + amdgpu_ring_emit_gds_switch(ring, job->vm_id, + job->gds_base, job->gds_size, + job->gws_base, job->gws_size, + job->oa_base, job->oa_size); } return 0; @@ -723,7 +792,7 @@ static void amdgpu_vm_frag_ptes(struct amdgpu_device *adev, * @vm: requested vm * @start: start of GPU address range * @end: end of GPU address range - * @dst: destination address to map to + * @dst: destination address to map to, the next dst inside the function * @flags: mapping flags * * Update the page tables in the range @start - @end. @@ -737,49 +806,75 @@ static void amdgpu_vm_update_ptes(struct amdgpu_device *adev, { const uint64_t mask = AMDGPU_VM_PTE_COUNT - 1; - uint64_t last_pe_start = ~0, last_pe_end = ~0, last_dst = ~0; - uint64_t addr; + uint64_t cur_pe_start, cur_pe_end, cur_dst; + uint64_t addr; /* next GPU address to be updated */ + uint64_t pt_idx; + struct amdgpu_bo *pt; + unsigned nptes; /* next number of ptes to be updated */ + uint64_t next_pe_start; + + /* initialize the variables */ + addr = start; + pt_idx = addr >> amdgpu_vm_block_size; + pt = vm->page_tables[pt_idx].entry.robj; + + if ((addr & ~mask) == (end & ~mask)) + nptes = end - addr; + else + nptes = AMDGPU_VM_PTE_COUNT - (addr & mask); + + cur_pe_start = amdgpu_bo_gpu_offset(pt); + cur_pe_start += (addr & mask) * 8; + cur_pe_end = cur_pe_start + 8 * nptes; + cur_dst = dst; + + /* for next ptb*/ + addr += nptes; + dst += nptes * AMDGPU_GPU_PAGE_SIZE; /* walk over the address space and update the page tables */ - for (addr = start; addr < end; ) { - uint64_t pt_idx = addr >> amdgpu_vm_block_size; - struct amdgpu_bo *pt = vm->page_tables[pt_idx].entry.robj; - unsigned nptes; - uint64_t pe_start; + while (addr < end) { + pt_idx = addr >> amdgpu_vm_block_size; + pt = vm->page_tables[pt_idx].entry.robj; if ((addr & ~mask) == (end & ~mask)) nptes = end - addr; else nptes = AMDGPU_VM_PTE_COUNT - (addr & mask); - pe_start = amdgpu_bo_gpu_offset(pt); - pe_start += (addr & mask) * 8; - - if (last_pe_end != pe_start) { + next_pe_start = amdgpu_bo_gpu_offset(pt); + next_pe_start += (addr & mask) * 8; + if (cur_pe_end == next_pe_start) { + /* The next ptb is consecutive to current ptb. + * Don't call amdgpu_vm_frag_ptes now. + * Will update two ptbs together in future. + */ + cur_pe_end += 8 * nptes; + } else { amdgpu_vm_frag_ptes(adev, vm_update_params, - last_pe_start, last_pe_end, - last_dst, flags); + cur_pe_start, cur_pe_end, + cur_dst, flags); - last_pe_start = pe_start; - last_pe_end = pe_start + 8 * nptes; - last_dst = dst; - } else { - last_pe_end += 8 * nptes; + cur_pe_start = next_pe_start; + cur_pe_end = next_pe_start + 8 * nptes; + cur_dst = dst; } + /* for next ptb*/ addr += nptes; dst += nptes * AMDGPU_GPU_PAGE_SIZE; } - amdgpu_vm_frag_ptes(adev, vm_update_params, last_pe_start, - last_pe_end, last_dst, flags); + amdgpu_vm_frag_ptes(adev, vm_update_params, cur_pe_start, + cur_pe_end, cur_dst, flags); } /** * amdgpu_vm_bo_update_mapping - update a mapping in the vm page table * * @adev: amdgpu_device pointer + * @exclusive: fence we need to sync to * @src: address where to copy page table entries from * @pages_addr: DMA addresses to use for mapping * @vm: requested vm @@ -793,6 +888,7 @@ static void amdgpu_vm_update_ptes(struct amdgpu_device *adev, * Returns 0 for success, -EINVAL for failure. */ static int amdgpu_vm_bo_update_mapping(struct amdgpu_device *adev, + struct fence *exclusive, uint64_t src, dma_addr_t *pages_addr, struct amdgpu_vm *vm, @@ -853,6 +949,10 @@ static int amdgpu_vm_bo_update_mapping(struct amdgpu_device *adev, vm_update_params.ib = &job->ibs[0]; + r = amdgpu_sync_fence(adev, &job->sync, exclusive); + if (r) + goto error_free; + r = amdgpu_sync_resv(adev, &job->sync, vm->page_directory->tbo.resv, owner); if (r) @@ -889,6 +989,7 @@ error_free: * amdgpu_vm_bo_split_mapping - split a mapping into smaller chunks * * @adev: amdgpu_device pointer + * @exclusive: fence we need to sync to * @gtt_flags: flags as they are used for GTT * @pages_addr: DMA addresses to use for mapping * @vm: requested vm @@ -902,6 +1003,7 @@ error_free: * Returns 0 for success, -EINVAL for failure. */ static int amdgpu_vm_bo_split_mapping(struct amdgpu_device *adev, + struct fence *exclusive, uint32_t gtt_flags, dma_addr_t *pages_addr, struct amdgpu_vm *vm, @@ -932,7 +1034,8 @@ static int amdgpu_vm_bo_split_mapping(struct amdgpu_device *adev, addr += mapping->offset; if (!pages_addr || src) - return amdgpu_vm_bo_update_mapping(adev, src, pages_addr, vm, + return amdgpu_vm_bo_update_mapping(adev, exclusive, + src, pages_addr, vm, start, mapping->it.last, flags, addr, fence); @@ -940,7 +1043,8 @@ static int amdgpu_vm_bo_split_mapping(struct amdgpu_device *adev, uint64_t last; last = min((uint64_t)mapping->it.last, start + max_size - 1); - r = amdgpu_vm_bo_update_mapping(adev, src, pages_addr, vm, + r = amdgpu_vm_bo_update_mapping(adev, exclusive, + src, pages_addr, vm, start, last, flags, addr, fence); if (r) @@ -973,6 +1077,7 @@ int amdgpu_vm_bo_update(struct amdgpu_device *adev, struct amdgpu_bo_va_mapping *mapping; dma_addr_t *pages_addr = NULL; uint32_t gtt_flags, flags; + struct fence *exclusive; uint64_t addr; int r; @@ -994,8 +1099,11 @@ int amdgpu_vm_bo_update(struct amdgpu_device *adev, default: break; } + + exclusive = reservation_object_get_excl(bo_va->bo->tbo.resv); } else { addr = 0; + exclusive = NULL; } flags = amdgpu_ttm_tt_pte_flags(adev, bo_va->bo->tbo.ttm, mem); @@ -1007,7 +1115,8 @@ int amdgpu_vm_bo_update(struct amdgpu_device *adev, spin_unlock(&vm->status_lock); list_for_each_entry(mapping, &bo_va->invalids, list) { - r = amdgpu_vm_bo_split_mapping(adev, gtt_flags, pages_addr, vm, + r = amdgpu_vm_bo_split_mapping(adev, exclusive, + gtt_flags, pages_addr, vm, mapping, flags, addr, &bo_va->last_pt_update); if (r) @@ -1054,7 +1163,7 @@ int amdgpu_vm_clear_freed(struct amdgpu_device *adev, struct amdgpu_bo_va_mapping, list); list_del(&mapping->list); - r = amdgpu_vm_bo_split_mapping(adev, 0, NULL, vm, mapping, + r = amdgpu_vm_bo_split_mapping(adev, NULL, 0, NULL, vm, mapping, 0, 0, NULL); kfree(mapping); if (r) @@ -1445,6 +1554,7 @@ int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm) amdgpu_bo_unreserve(vm->page_directory); if (r) goto error_free_page_directory; + vm->last_eviction_counter = atomic64_read(&adev->num_evictions); return 0; @@ -1516,6 +1626,10 @@ void amdgpu_vm_manager_init(struct amdgpu_device *adev) &adev->vm_manager.ids_lru); } + adev->vm_manager.fence_context = fence_context_alloc(AMDGPU_MAX_RINGS); + for (i = 0; i < AMDGPU_MAX_RINGS; ++i) + adev->vm_manager.seqno[i] = 0; + atomic_set(&adev->vm_manager.vm_pte_next_ring, 0); atomic64_set(&adev->vm_manager.client_counter, 0); } diff --git a/drivers/gpu/drm/amd/amdgpu/atombios_encoders.c b/drivers/gpu/drm/amd/amdgpu/atombios_encoders.c index 48b6bd671cda..c32eca26155c 100644 --- a/drivers/gpu/drm/amd/amdgpu/atombios_encoders.c +++ b/drivers/gpu/drm/amd/amdgpu/atombios_encoders.c @@ -98,6 +98,7 @@ amdgpu_atombios_encoder_set_backlight_level(struct amdgpu_encoder *amdgpu_encode case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA: case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1: case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2: + case ENCODER_OBJECT_ID_INTERNAL_UNIPHY3: if (dig->backlight_level == 0) amdgpu_atombios_encoder_setup_dig_transmitter(encoder, ATOM_TRANSMITTER_ACTION_LCD_BLOFF, 0, 0); diff --git a/drivers/gpu/drm/amd/amdgpu/atombios_i2c.c b/drivers/gpu/drm/amd/amdgpu/atombios_i2c.c index 13cdb01e9b45..bc56c8a181e6 100644 --- a/drivers/gpu/drm/amd/amdgpu/atombios_i2c.c +++ b/drivers/gpu/drm/amd/amdgpu/atombios_i2c.c @@ -156,3 +156,18 @@ u32 amdgpu_atombios_i2c_func(struct i2c_adapter *adap) return I2C_FUNC_I2C | I2C_FUNC_SMBUS_EMUL; } +void amdgpu_atombios_i2c_channel_trans(struct amdgpu_device* adev, u8 slave_addr, u8 line_number, u8 offset, u8 data) +{ + PROCESS_I2C_CHANNEL_TRANSACTION_PS_ALLOCATION args; + int index = GetIndexIntoMasterTable(COMMAND, ProcessI2cChannelTransaction); + + args.ucRegIndex = offset; + args.lpI2CDataOut = data; + args.ucFlag = 1; + args.ucI2CSpeed = TARGET_HW_I2C_CLOCK; + args.ucTransBytes = 1; + args.ucSlaveAddr = slave_addr; + args.ucLineNumber = line_number; + + amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args); +} diff --git a/drivers/gpu/drm/amd/amdgpu/atombios_i2c.h b/drivers/gpu/drm/amd/amdgpu/atombios_i2c.h index d6128d9de56e..251aaf41f65d 100644 --- a/drivers/gpu/drm/amd/amdgpu/atombios_i2c.h +++ b/drivers/gpu/drm/amd/amdgpu/atombios_i2c.h @@ -27,5 +27,7 @@ int amdgpu_atombios_i2c_xfer(struct i2c_adapter *i2c_adap, struct i2c_msg *msgs, int num); u32 amdgpu_atombios_i2c_func(struct i2c_adapter *adap); +void amdgpu_atombios_i2c_channel_trans(struct amdgpu_device* adev, + u8 slave_addr, u8 line_number, u8 offset, u8 data); #endif diff --git a/drivers/gpu/drm/amd/amdgpu/ci_dpm.c b/drivers/gpu/drm/amd/amdgpu/ci_dpm.c index ea407db1fbcf..e2f0e5d58d5c 100644 --- a/drivers/gpu/drm/amd/amdgpu/ci_dpm.c +++ b/drivers/gpu/drm/amd/amdgpu/ci_dpm.c @@ -50,7 +50,9 @@ #include "gmc/gmc_7_1_sh_mask.h" MODULE_FIRMWARE("radeon/bonaire_smc.bin"); +MODULE_FIRMWARE("radeon/bonaire_k_smc.bin"); MODULE_FIRMWARE("radeon/hawaii_smc.bin"); +MODULE_FIRMWARE("radeon/hawaii_k_smc.bin"); #define MC_CG_ARB_FREQ_F0 0x0a #define MC_CG_ARB_FREQ_F1 0x0b @@ -84,12 +86,14 @@ static const struct ci_pt_defaults defaults_bonaire_xt = { 0x17C, 0x172, 0x180, 0x1BC, 0x1B3, 0x1BD, 0x206, 0x200, 0x203, 0x25D, 0x25A, 0x255, 0x2C3, 0x2C5, 0x2B4 } }; +#if 0 static const struct ci_pt_defaults defaults_bonaire_pro = { 1, 0xF, 0xFD, 0x19, 5, 45, 0, 0x65062, { 0x8C, 0x23F, 0x244, 0xA6, 0x83, 0x85, 0x86, 0x86, 0x83, 0xDB, 0xDB, 0xDA, 0x67, 0x60, 0x5F }, { 0x187, 0x193, 0x193, 0x1C7, 0x1D1, 0x1D1, 0x210, 0x219, 0x219, 0x266, 0x26C, 0x26C, 0x2C9, 0x2CB, 0x2CB } }; +#endif static const struct ci_pt_defaults defaults_saturn_xt = { @@ -98,12 +102,14 @@ static const struct ci_pt_defaults defaults_saturn_xt = { 0x187, 0x187, 0x187, 0x1C7, 0x1C7, 0x1C7, 0x210, 0x210, 0x210, 0x266, 0x266, 0x266, 0x2C9, 0x2C9, 0x2C9 } }; +#if 0 static const struct ci_pt_defaults defaults_saturn_pro = { 1, 0xF, 0xFD, 0x19, 5, 55, 0, 0x30000, { 0x96, 0x21D, 0x23B, 0xA1, 0x85, 0x87, 0x83, 0x84, 0x81, 0xE6, 0xE6, 0xE6, 0x71, 0x6A, 0x6A }, { 0x193, 0x19E, 0x19E, 0x1D2, 0x1DC, 0x1DC, 0x21A, 0x223, 0x223, 0x26E, 0x27E, 0x274, 0x2CF, 0x2D2, 0x2D2 } }; +#endif static const struct ci_pt_config_reg didt_config_ci[] = { @@ -736,19 +742,19 @@ static int ci_enable_didt(struct amdgpu_device *adev, bool enable) if (pi->caps_sq_ramping || pi->caps_db_ramping || pi->caps_td_ramping || pi->caps_tcp_ramping) { - gfx_v7_0_enter_rlc_safe_mode(adev); + adev->gfx.rlc.funcs->enter_safe_mode(adev); if (enable) { ret = ci_program_pt_config_registers(adev, didt_config_ci); if (ret) { - gfx_v7_0_exit_rlc_safe_mode(adev); + adev->gfx.rlc.funcs->exit_safe_mode(adev); return ret; } } ci_do_enable_didt(adev, enable); - gfx_v7_0_exit_rlc_safe_mode(adev); + adev->gfx.rlc.funcs->exit_safe_mode(adev); } return 0; @@ -3030,7 +3036,7 @@ static int ci_populate_single_memory_level(struct amdgpu_device *adev, if (pi->mclk_stutter_mode_threshold && (memory_clock <= pi->mclk_stutter_mode_threshold) && - (pi->uvd_enabled == false) && + (!pi->uvd_enabled) && (RREG32(mmDPG_PIPE_STUTTER_CONTROL) & DPG_PIPE_STUTTER_CONTROL__STUTTER_ENABLE_MASK) && (adev->pm.dpm.new_active_crtc_count <= 2)) memory_level->StutterEnable = true; @@ -3636,6 +3642,10 @@ static int ci_setup_default_dpm_tables(struct amdgpu_device *adev) ci_setup_default_pcie_tables(adev); + /* save a copy of the default DPM table */ + memcpy(&(pi->golden_dpm_table), &(pi->dpm_table), + sizeof(struct ci_dpm_table)); + return 0; } @@ -5754,10 +5764,18 @@ static int ci_dpm_init_microcode(struct amdgpu_device *adev) switch (adev->asic_type) { case CHIP_BONAIRE: - chip_name = "bonaire"; + if ((adev->pdev->revision == 0x80) || + (adev->pdev->revision == 0x81) || + (adev->pdev->device == 0x665f)) + chip_name = "bonaire_k"; + else + chip_name = "bonaire"; break; case CHIP_HAWAII: - chip_name = "hawaii"; + if (adev->pdev->revision == 0x80) + chip_name = "hawaii_k"; + else + chip_name = "hawaii"; break; case CHIP_KAVERI: case CHIP_KABINI: @@ -6221,6 +6239,9 @@ static int ci_dpm_sw_fini(void *handle) ci_dpm_fini(adev); mutex_unlock(&adev->pm.mutex); + release_firmware(adev->pm.fw); + adev->pm.fw = NULL; + return 0; } @@ -6401,6 +6422,186 @@ static int ci_dpm_set_powergating_state(void *handle, return 0; } +static int ci_dpm_print_clock_levels(struct amdgpu_device *adev, + enum pp_clock_type type, char *buf) +{ + struct ci_power_info *pi = ci_get_pi(adev); + struct ci_single_dpm_table *sclk_table = &pi->dpm_table.sclk_table; + struct ci_single_dpm_table *mclk_table = &pi->dpm_table.mclk_table; + struct ci_single_dpm_table *pcie_table = &pi->dpm_table.pcie_speed_table; + + int i, now, size = 0; + uint32_t clock, pcie_speed; + + switch (type) { + case PP_SCLK: + amdgpu_ci_send_msg_to_smc(adev, PPSMC_MSG_API_GetSclkFrequency); + clock = RREG32(mmSMC_MSG_ARG_0); + + for (i = 0; i < sclk_table->count; i++) { + if (clock > sclk_table->dpm_levels[i].value) + continue; + break; + } + now = i; + + for (i = 0; i < sclk_table->count; i++) + size += sprintf(buf + size, "%d: %uMhz %s\n", + i, sclk_table->dpm_levels[i].value / 100, + (i == now) ? "*" : ""); + break; + case PP_MCLK: + amdgpu_ci_send_msg_to_smc(adev, PPSMC_MSG_API_GetMclkFrequency); + clock = RREG32(mmSMC_MSG_ARG_0); + + for (i = 0; i < mclk_table->count; i++) { + if (clock > mclk_table->dpm_levels[i].value) + continue; + break; + } + now = i; + + for (i = 0; i < mclk_table->count; i++) + size += sprintf(buf + size, "%d: %uMhz %s\n", + i, mclk_table->dpm_levels[i].value / 100, + (i == now) ? "*" : ""); + break; + case PP_PCIE: + pcie_speed = ci_get_current_pcie_speed(adev); + for (i = 0; i < pcie_table->count; i++) { + if (pcie_speed != pcie_table->dpm_levels[i].value) + continue; + break; + } + now = i; + + for (i = 0; i < pcie_table->count; i++) + size += sprintf(buf + size, "%d: %s %s\n", i, + (pcie_table->dpm_levels[i].value == 0) ? "2.5GB, x1" : + (pcie_table->dpm_levels[i].value == 1) ? "5.0GB, x16" : + (pcie_table->dpm_levels[i].value == 2) ? "8.0GB, x16" : "", + (i == now) ? "*" : ""); + break; + default: + break; + } + + return size; +} + +static int ci_dpm_force_clock_level(struct amdgpu_device *adev, + enum pp_clock_type type, uint32_t mask) +{ + struct ci_power_info *pi = ci_get_pi(adev); + + if (adev->pm.dpm.forced_level + != AMDGPU_DPM_FORCED_LEVEL_MANUAL) + return -EINVAL; + + switch (type) { + case PP_SCLK: + if (!pi->sclk_dpm_key_disabled) + amdgpu_ci_send_msg_to_smc_with_parameter(adev, + PPSMC_MSG_SCLKDPM_SetEnabledMask, + pi->dpm_level_enable_mask.sclk_dpm_enable_mask & mask); + break; + + case PP_MCLK: + if (!pi->mclk_dpm_key_disabled) + amdgpu_ci_send_msg_to_smc_with_parameter(adev, + PPSMC_MSG_MCLKDPM_SetEnabledMask, + pi->dpm_level_enable_mask.mclk_dpm_enable_mask & mask); + break; + + case PP_PCIE: + { + uint32_t tmp = mask & pi->dpm_level_enable_mask.pcie_dpm_enable_mask; + uint32_t level = 0; + + while (tmp >>= 1) + level++; + + if (!pi->pcie_dpm_key_disabled) + amdgpu_ci_send_msg_to_smc_with_parameter(adev, + PPSMC_MSG_PCIeDPM_ForceLevel, + level); + break; + } + default: + break; + } + + return 0; +} + +static int ci_dpm_get_sclk_od(struct amdgpu_device *adev) +{ + struct ci_power_info *pi = ci_get_pi(adev); + struct ci_single_dpm_table *sclk_table = &(pi->dpm_table.sclk_table); + struct ci_single_dpm_table *golden_sclk_table = + &(pi->golden_dpm_table.sclk_table); + int value; + + value = (sclk_table->dpm_levels[sclk_table->count - 1].value - + golden_sclk_table->dpm_levels[golden_sclk_table->count - 1].value) * + 100 / + golden_sclk_table->dpm_levels[golden_sclk_table->count - 1].value; + + return value; +} + +static int ci_dpm_set_sclk_od(struct amdgpu_device *adev, uint32_t value) +{ + struct ci_power_info *pi = ci_get_pi(adev); + struct ci_ps *ps = ci_get_ps(adev->pm.dpm.requested_ps); + struct ci_single_dpm_table *golden_sclk_table = + &(pi->golden_dpm_table.sclk_table); + + if (value > 20) + value = 20; + + ps->performance_levels[ps->performance_level_count - 1].sclk = + golden_sclk_table->dpm_levels[golden_sclk_table->count - 1].value * + value / 100 + + golden_sclk_table->dpm_levels[golden_sclk_table->count - 1].value; + + return 0; +} + +static int ci_dpm_get_mclk_od(struct amdgpu_device *adev) +{ + struct ci_power_info *pi = ci_get_pi(adev); + struct ci_single_dpm_table *mclk_table = &(pi->dpm_table.mclk_table); + struct ci_single_dpm_table *golden_mclk_table = + &(pi->golden_dpm_table.mclk_table); + int value; + + value = (mclk_table->dpm_levels[mclk_table->count - 1].value - + golden_mclk_table->dpm_levels[golden_mclk_table->count - 1].value) * + 100 / + golden_mclk_table->dpm_levels[golden_mclk_table->count - 1].value; + + return value; +} + +static int ci_dpm_set_mclk_od(struct amdgpu_device *adev, uint32_t value) +{ + struct ci_power_info *pi = ci_get_pi(adev); + struct ci_ps *ps = ci_get_ps(adev->pm.dpm.requested_ps); + struct ci_single_dpm_table *golden_mclk_table = + &(pi->golden_dpm_table.mclk_table); + + if (value > 20) + value = 20; + + ps->performance_levels[ps->performance_level_count - 1].mclk = + golden_mclk_table->dpm_levels[golden_mclk_table->count - 1].value * + value / 100 + + golden_mclk_table->dpm_levels[golden_mclk_table->count - 1].value; + + return 0; +} + const struct amd_ip_funcs ci_dpm_ip_funcs = { .name = "ci_dpm", .early_init = ci_dpm_early_init, @@ -6435,6 +6636,12 @@ static const struct amdgpu_dpm_funcs ci_dpm_funcs = { .get_fan_control_mode = &ci_dpm_get_fan_control_mode, .set_fan_speed_percent = &ci_dpm_set_fan_speed_percent, .get_fan_speed_percent = &ci_dpm_get_fan_speed_percent, + .print_clock_levels = ci_dpm_print_clock_levels, + .force_clock_level = ci_dpm_force_clock_level, + .get_sclk_od = ci_dpm_get_sclk_od, + .set_sclk_od = ci_dpm_set_sclk_od, + .get_mclk_od = ci_dpm_get_mclk_od, + .set_mclk_od = ci_dpm_set_mclk_od, }; static void ci_dpm_set_dpm_funcs(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/amdgpu/ci_dpm.h b/drivers/gpu/drm/amd/amdgpu/ci_dpm.h index faccc30c93bf..91be2996ae7c 100644 --- a/drivers/gpu/drm/amd/amdgpu/ci_dpm.h +++ b/drivers/gpu/drm/amd/amdgpu/ci_dpm.h @@ -193,6 +193,7 @@ struct ci_pt_defaults { struct ci_power_info { struct ci_dpm_table dpm_table; + struct ci_dpm_table golden_dpm_table; u32 voltage_control; u32 mvdd_control; u32 vddci_control; diff --git a/drivers/gpu/drm/amd/amdgpu/cik.c b/drivers/gpu/drm/amd/amdgpu/cik.c index 07bc795a4ca9..4efc901f658c 100644 --- a/drivers/gpu/drm/amd/amdgpu/cik.c +++ b/drivers/gpu/drm/amd/amdgpu/cik.c @@ -879,7 +879,7 @@ static void cik_vga_set_state(struct amdgpu_device *adev, bool state) uint32_t tmp; tmp = RREG32(mmCONFIG_CNTL); - if (state == false) + if (!state) tmp |= CONFIG_CNTL__VGA_DIS_MASK; else tmp &= ~CONFIG_CNTL__VGA_DIS_MASK; @@ -962,6 +962,12 @@ static bool cik_read_bios_from_rom(struct amdgpu_device *adev, return true; } +static u32 cik_get_virtual_caps(struct amdgpu_device *adev) +{ + /* CIK does not support SR-IOV */ + return 0; +} + static const struct amdgpu_allowed_register_entry cik_allowed_read_registers[] = { {mmGRBM_STATUS, false}, {mmGB_ADDR_CONFIG, false}, @@ -1029,12 +1035,12 @@ static uint32_t cik_read_indexed_register(struct amdgpu_device *adev, mutex_lock(&adev->grbm_idx_mutex); if (se_num != 0xffffffff || sh_num != 0xffffffff) - gfx_v7_0_select_se_sh(adev, se_num, sh_num); + amdgpu_gfx_select_se_sh(adev, se_num, sh_num, 0xffffffff); val = RREG32(reg_offset); if (se_num != 0xffffffff || sh_num != 0xffffffff) - gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff); + amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); mutex_unlock(&adev->grbm_idx_mutex); return val; } @@ -1152,10 +1158,11 @@ static void kv_restore_regs_for_reset(struct amdgpu_device *adev, WREG32(mmGMCON_RENG_EXECUTE, save->gmcon_reng_execute); } -static void cik_gpu_pci_config_reset(struct amdgpu_device *adev) +static int cik_gpu_pci_config_reset(struct amdgpu_device *adev) { struct kv_reset_save_regs kv_save = { 0 }; u32 i; + int r = -EINVAL; dev_info(adev->dev, "GPU pci config reset\n"); @@ -1171,14 +1178,20 @@ static void cik_gpu_pci_config_reset(struct amdgpu_device *adev) /* wait for asic to come out of reset */ for (i = 0; i < adev->usec_timeout; i++) { - if (RREG32(mmCONFIG_MEMSIZE) != 0xffffffff) + if (RREG32(mmCONFIG_MEMSIZE) != 0xffffffff) { + /* enable BM */ + pci_set_master(adev->pdev); + r = 0; break; + } udelay(1); } /* does asic init need to be run first??? */ if (adev->flags & AMD_IS_APU) kv_restore_regs_for_reset(adev, &kv_save); + + return r; } static void cik_set_bios_scratch_engine_hung(struct amdgpu_device *adev, bool hung) @@ -1204,13 +1217,14 @@ static void cik_set_bios_scratch_engine_hung(struct amdgpu_device *adev, bool hu */ static int cik_asic_reset(struct amdgpu_device *adev) { + int r; cik_set_bios_scratch_engine_hung(adev, true); - cik_gpu_pci_config_reset(adev); + r = cik_gpu_pci_config_reset(adev); cik_set_bios_scratch_engine_hung(adev, false); - return 0; + return r; } static int cik_set_uvd_clock(struct amdgpu_device *adev, u32 clock, @@ -2007,9 +2021,7 @@ static const struct amdgpu_asic_funcs cik_asic_funcs = .get_xclk = &cik_get_xclk, .set_uvd_clocks = &cik_set_uvd_clocks, .set_vce_clocks = &cik_set_vce_clocks, - /* these should be moved to their own ip modules */ - .get_gpu_clock_counter = &gfx_v7_0_get_gpu_clock_counter, - .wait_for_mc_idle = &gmc_v7_0_mc_wait_for_idle, + .get_virtual_caps = &cik_get_virtual_caps, }; static int cik_common_early_init(void *handle) diff --git a/drivers/gpu/drm/amd/amdgpu/cik_sdma.c b/drivers/gpu/drm/amd/amdgpu/cik_sdma.c index 518dca43b133..ee6466912497 100644 --- a/drivers/gpu/drm/amd/amdgpu/cik_sdma.c +++ b/drivers/gpu/drm/amd/amdgpu/cik_sdma.c @@ -66,6 +66,16 @@ MODULE_FIRMWARE("radeon/mullins_sdma1.bin"); u32 amdgpu_cik_gpu_check_soft_reset(struct amdgpu_device *adev); + +static void cik_sdma_free_microcode(struct amdgpu_device *adev) +{ + int i; + for (i = 0; i < adev->sdma.num_instances; i++) { + release_firmware(adev->sdma.instance[i].fw); + adev->sdma.instance[i].fw = NULL; + } +} + /* * sDMA - System DMA * Starting with CIK, the GPU has new asynchronous @@ -214,17 +224,6 @@ static void cik_sdma_ring_emit_ib(struct amdgpu_ring *ring, unsigned vm_id, bool ctx_switch) { u32 extra_bits = vm_id & 0xf; - u32 next_rptr = ring->wptr + 5; - - while ((next_rptr & 7) != 4) - next_rptr++; - - next_rptr += 4; - amdgpu_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_WRITE, SDMA_WRITE_SUB_OPCODE_LINEAR, 0)); - amdgpu_ring_write(ring, ring->next_rptr_gpu_addr & 0xfffffffc); - amdgpu_ring_write(ring, upper_32_bits(ring->next_rptr_gpu_addr) & 0xffffffff); - amdgpu_ring_write(ring, 1); /* number of DWs to follow */ - amdgpu_ring_write(ring, next_rptr); /* IB packet must end on a 8 DW boundary */ cik_sdma_ring_insert_nop(ring, (12 - (ring->wptr & 7)) % 8); @@ -355,7 +354,7 @@ static void cik_sdma_enable(struct amdgpu_device *adev, bool enable) u32 me_cntl; int i; - if (enable == false) { + if (!enable) { cik_sdma_gfx_stop(adev); cik_sdma_rlc_stop(adev); } @@ -419,6 +418,8 @@ static int cik_sdma_gfx_resume(struct amdgpu_device *adev) /* Initialize the ring buffer's read and write pointers */ WREG32(mmSDMA0_GFX_RB_RPTR + sdma_offsets[i], 0); WREG32(mmSDMA0_GFX_RB_WPTR + sdma_offsets[i], 0); + WREG32(mmSDMA0_GFX_IB_RPTR + sdma_offsets[i], 0); + WREG32(mmSDMA0_GFX_IB_OFFSET + sdma_offsets[i], 0); /* set the wb address whether it's enabled or not */ WREG32(mmSDMA0_GFX_RB_RPTR_ADDR_HI + sdma_offsets[i], @@ -446,7 +447,12 @@ static int cik_sdma_gfx_resume(struct amdgpu_device *adev) WREG32(mmSDMA0_GFX_IB_CNTL + sdma_offsets[i], ib_cntl); ring->ready = true; + } + + cik_sdma_enable(adev, true); + for (i = 0; i < adev->sdma.num_instances; i++) { + ring = &adev->sdma.instance[i].ring; r = amdgpu_ring_test_ring(ring); if (r) { ring->ready = false; @@ -529,8 +535,8 @@ static int cik_sdma_start(struct amdgpu_device *adev) if (r) return r; - /* unhalt the MEs */ - cik_sdma_enable(adev, true); + /* halt the engine before programing */ + cik_sdma_enable(adev, false); /* start the gfx rings and rlc compute queues */ r = cik_sdma_gfx_resume(adev); @@ -611,20 +617,19 @@ static int cik_sdma_ring_test_ring(struct amdgpu_ring *ring) * Test a simple IB in the DMA ring (CIK). * Returns 0 on success, error on failure. */ -static int cik_sdma_ring_test_ib(struct amdgpu_ring *ring) +static int cik_sdma_ring_test_ib(struct amdgpu_ring *ring, long timeout) { struct amdgpu_device *adev = ring->adev; struct amdgpu_ib ib; struct fence *f = NULL; - unsigned i; unsigned index; - int r; u32 tmp = 0; u64 gpu_addr; + long r; r = amdgpu_wb_get(adev, &index); if (r) { - dev_err(adev->dev, "(%d) failed to allocate wb slot\n", r); + dev_err(adev->dev, "(%ld) failed to allocate wb slot\n", r); return r; } @@ -634,11 +639,12 @@ static int cik_sdma_ring_test_ib(struct amdgpu_ring *ring) memset(&ib, 0, sizeof(ib)); r = amdgpu_ib_get(adev, NULL, 256, &ib); if (r) { - DRM_ERROR("amdgpu: failed to get ib (%d).\n", r); + DRM_ERROR("amdgpu: failed to get ib (%ld).\n", r); goto err0; } - ib.ptr[0] = SDMA_PACKET(SDMA_OPCODE_WRITE, SDMA_WRITE_SUB_OPCODE_LINEAR, 0); + ib.ptr[0] = SDMA_PACKET(SDMA_OPCODE_WRITE, + SDMA_WRITE_SUB_OPCODE_LINEAR, 0); ib.ptr[1] = lower_32_bits(gpu_addr); ib.ptr[2] = upper_32_bits(gpu_addr); ib.ptr[3] = 1; @@ -648,28 +654,25 @@ static int cik_sdma_ring_test_ib(struct amdgpu_ring *ring) if (r) goto err1; - r = fence_wait(f, false); - if (r) { - DRM_ERROR("amdgpu: fence wait failed (%d).\n", r); + r = fence_wait_timeout(f, false, timeout); + if (r == 0) { + DRM_ERROR("amdgpu: IB test timed out\n"); + r = -ETIMEDOUT; goto err1; - } - for (i = 0; i < adev->usec_timeout; i++) { - tmp = le32_to_cpu(adev->wb.wb[index]); - if (tmp == 0xDEADBEEF) - break; - DRM_UDELAY(1); - } - if (i < adev->usec_timeout) { - DRM_INFO("ib test on ring %d succeeded in %u usecs\n", - ring->idx, i); + } else if (r < 0) { + DRM_ERROR("amdgpu: fence wait failed (%ld).\n", r); goto err1; + } + tmp = le32_to_cpu(adev->wb.wb[index]); + if (tmp == 0xDEADBEEF) { + DRM_INFO("ib test on ring %d succeeded\n", ring->idx); + r = 0; } else { DRM_ERROR("amdgpu: ib test failed (0x%08X)\n", tmp); r = -EINVAL; } err1: - fence_put(f); amdgpu_ib_free(adev, &ib, NULL); fence_put(f); err0: @@ -998,6 +1001,7 @@ static int cik_sdma_sw_fini(void *handle) for (i = 0; i < adev->sdma.num_instances; i++) amdgpu_ring_fini(&adev->sdma.instance[i].ring); + cik_sdma_free_microcode(adev); return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/cz_dpm.c b/drivers/gpu/drm/amd/amdgpu/cz_dpm.c index 933e425a8154..2a11413ed54a 100644 --- a/drivers/gpu/drm/amd/amdgpu/cz_dpm.c +++ b/drivers/gpu/drm/amd/amdgpu/cz_dpm.c @@ -425,7 +425,7 @@ static int cz_dpm_init(struct amdgpu_device *adev) pi->mgcg_cgtt_local1 = 0x0; pi->clock_slow_down_step = 25000; pi->skip_clock_slow_down = 1; - pi->enable_nb_ps_policy = 0; + pi->enable_nb_ps_policy = false; pi->caps_power_containment = true; pi->caps_cac = true; pi->didt_enabled = false; @@ -2219,6 +2219,7 @@ static void cz_dpm_powergate_vce(struct amdgpu_device *adev, bool gate) } } } else { /*pi->caps_vce_pg*/ + pi->vce_power_gated = gate; cz_update_vce_dpm(adev); cz_enable_vce_dpm(adev, !gate); } diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c index c90408bc0fde..d4bf133908b1 100644 --- a/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c +++ b/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c @@ -307,11 +307,10 @@ static void dce_v11_0_page_flip(struct amdgpu_device *adev, struct amdgpu_crtc *amdgpu_crtc = adev->mode_info.crtcs[crtc_id]; u32 tmp; - /* flip at hsync for async, default is vsync */ - /* use UPDATE_IMMEDIATE_EN instead for async? */ + /* flip immediate for async, default is vsync */ tmp = RREG32(mmGRPH_FLIP_CONTROL + amdgpu_crtc->crtc_offset); tmp = REG_SET_FIELD(tmp, GRPH_FLIP_CONTROL, - GRPH_SURFACE_UPDATE_H_RETRACE_EN, async ? 1 : 0); + GRPH_SURFACE_UPDATE_IMMEDIATE_EN, async ? 1 : 0); WREG32(mmGRPH_FLIP_CONTROL + amdgpu_crtc->crtc_offset, tmp); /* update the scanout addresses */ WREG32(mmGRPH_PRIMARY_SURFACE_ADDRESS_HIGH + amdgpu_crtc->crtc_offset, diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c index 300ff4aab0fd..4fdfab1e9200 100644 --- a/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c +++ b/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c @@ -526,36 +526,16 @@ static void dce_v8_0_stop_mc_access(struct amdgpu_device *adev, crtc_enabled = REG_GET_FIELD(RREG32(mmCRTC_CONTROL + crtc_offsets[i]), CRTC_CONTROL, CRTC_MASTER_EN); if (crtc_enabled) { -#if 0 - u32 frame_count; - int j; - +#if 1 save->crtc_enabled[i] = true; tmp = RREG32(mmCRTC_BLANK_CONTROL + crtc_offsets[i]); if (REG_GET_FIELD(tmp, CRTC_BLANK_CONTROL, CRTC_BLANK_DATA_EN) == 0) { - amdgpu_display_vblank_wait(adev, i); - WREG32(mmCRTC_UPDATE_LOCK + crtc_offsets[i], 1); + /*it is correct only for RGB ; black is 0*/ + WREG32(mmCRTC_BLANK_DATA_COLOR + crtc_offsets[i], 0); tmp = REG_SET_FIELD(tmp, CRTC_BLANK_CONTROL, CRTC_BLANK_DATA_EN, 1); WREG32(mmCRTC_BLANK_CONTROL + crtc_offsets[i], tmp); - WREG32(mmCRTC_UPDATE_LOCK + crtc_offsets[i], 0); - } - /* wait for the next frame */ - frame_count = amdgpu_display_vblank_get_counter(adev, i); - for (j = 0; j < adev->usec_timeout; j++) { - if (amdgpu_display_vblank_get_counter(adev, i) != frame_count) - break; - udelay(1); - } - tmp = RREG32(mmGRPH_UPDATE + crtc_offsets[i]); - if (REG_GET_FIELD(tmp, GRPH_UPDATE, GRPH_UPDATE_LOCK) == 0) { - tmp = REG_SET_FIELD(tmp, GRPH_UPDATE, GRPH_UPDATE_LOCK, 1); - WREG32(mmGRPH_UPDATE + crtc_offsets[i], tmp); - } - tmp = RREG32(mmMASTER_UPDATE_LOCK + crtc_offsets[i]); - if (REG_GET_FIELD(tmp, MASTER_UPDATE_LOCK, MASTER_UPDATE_LOCK) == 0) { - tmp = REG_SET_FIELD(tmp, MASTER_UPDATE_LOCK, MASTER_UPDATE_LOCK, 1); - WREG32(mmMASTER_UPDATE_LOCK + crtc_offsets[i], tmp); } + mdelay(20); #else /* XXX this is a hack to avoid strange behavior with EFI on certain systems */ WREG32(mmCRTC_UPDATE_LOCK + crtc_offsets[i], 1); @@ -575,55 +555,22 @@ static void dce_v8_0_stop_mc_access(struct amdgpu_device *adev, static void dce_v8_0_resume_mc_access(struct amdgpu_device *adev, struct amdgpu_mode_mc_save *save) { - u32 tmp, frame_count; - int i, j; + u32 tmp; + int i; /* update crtc base addresses */ for (i = 0; i < adev->mode_info.num_crtc; i++) { WREG32(mmGRPH_PRIMARY_SURFACE_ADDRESS_HIGH + crtc_offsets[i], upper_32_bits(adev->mc.vram_start)); - WREG32(mmGRPH_SECONDARY_SURFACE_ADDRESS_HIGH + crtc_offsets[i], - upper_32_bits(adev->mc.vram_start)); WREG32(mmGRPH_PRIMARY_SURFACE_ADDRESS + crtc_offsets[i], (u32)adev->mc.vram_start); - WREG32(mmGRPH_SECONDARY_SURFACE_ADDRESS + crtc_offsets[i], - (u32)adev->mc.vram_start); if (save->crtc_enabled[i]) { - tmp = RREG32(mmMASTER_UPDATE_MODE + crtc_offsets[i]); - if (REG_GET_FIELD(tmp, MASTER_UPDATE_MODE, MASTER_UPDATE_MODE) != 3) { - tmp = REG_SET_FIELD(tmp, MASTER_UPDATE_MODE, MASTER_UPDATE_MODE, 3); - WREG32(mmMASTER_UPDATE_MODE + crtc_offsets[i], tmp); - } - tmp = RREG32(mmGRPH_UPDATE + crtc_offsets[i]); - if (REG_GET_FIELD(tmp, GRPH_UPDATE, GRPH_UPDATE_LOCK)) { - tmp = REG_SET_FIELD(tmp, GRPH_UPDATE, GRPH_UPDATE_LOCK, 0); - WREG32(mmGRPH_UPDATE + crtc_offsets[i], tmp); - } - tmp = RREG32(mmMASTER_UPDATE_LOCK + crtc_offsets[i]); - if (REG_GET_FIELD(tmp, MASTER_UPDATE_LOCK, MASTER_UPDATE_LOCK)) { - tmp = REG_SET_FIELD(tmp, MASTER_UPDATE_LOCK, MASTER_UPDATE_LOCK, 0); - WREG32(mmMASTER_UPDATE_LOCK + crtc_offsets[i], tmp); - } - for (j = 0; j < adev->usec_timeout; j++) { - tmp = RREG32(mmGRPH_UPDATE + crtc_offsets[i]); - if (REG_GET_FIELD(tmp, GRPH_UPDATE, GRPH_SURFACE_UPDATE_PENDING) == 0) - break; - udelay(1); - } tmp = RREG32(mmCRTC_BLANK_CONTROL + crtc_offsets[i]); tmp = REG_SET_FIELD(tmp, CRTC_BLANK_CONTROL, CRTC_BLANK_DATA_EN, 0); - WREG32(mmCRTC_UPDATE_LOCK + crtc_offsets[i], 1); WREG32(mmCRTC_BLANK_CONTROL + crtc_offsets[i], tmp); - WREG32(mmCRTC_UPDATE_LOCK + crtc_offsets[i], 0); - /* wait for the next frame */ - frame_count = amdgpu_display_vblank_get_counter(adev, i); - for (j = 0; j < adev->usec_timeout; j++) { - if (amdgpu_display_vblank_get_counter(adev, i) != frame_count) - break; - udelay(1); - } } + mdelay(20); } WREG32(mmVGA_MEMORY_BASE_ADDRESS_HIGH, upper_32_bits(adev->mc.vram_start)); diff --git a/drivers/gpu/drm/amd/amdgpu/fiji_dpm.c b/drivers/gpu/drm/amd/amdgpu/fiji_dpm.c index 245cabf06575..ed03b75175d4 100644 --- a/drivers/gpu/drm/amd/amdgpu/fiji_dpm.c +++ b/drivers/gpu/drm/amd/amdgpu/fiji_dpm.c @@ -72,6 +72,11 @@ static int fiji_dpm_sw_init(void *handle) static int fiji_dpm_sw_fini(void *handle) { + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + + release_firmware(adev->pm.fw); + adev->pm.fw = NULL; + return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/fiji_smc.c b/drivers/gpu/drm/amd/amdgpu/fiji_smc.c index b336c918d6a7..b3e19ba4c57f 100644 --- a/drivers/gpu/drm/amd/amdgpu/fiji_smc.c +++ b/drivers/gpu/drm/amd/amdgpu/fiji_smc.c @@ -173,7 +173,7 @@ static int fiji_send_msg_to_smc(struct amdgpu_device *adev, PPSMC_Msg msg) { if (!fiji_is_smc_ram_running(adev)) { - return -EINVAL;; + return -EINVAL; } if (wait_smu_response(adev)) { diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c index 7f18a53ab53a..d869d058ef24 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c @@ -991,6 +991,22 @@ out: return err; } +static void gfx_v7_0_free_microcode(struct amdgpu_device *adev) +{ + release_firmware(adev->gfx.pfp_fw); + adev->gfx.pfp_fw = NULL; + release_firmware(adev->gfx.me_fw); + adev->gfx.me_fw = NULL; + release_firmware(adev->gfx.ce_fw); + adev->gfx.ce_fw = NULL; + release_firmware(adev->gfx.mec_fw); + adev->gfx.mec_fw = NULL; + release_firmware(adev->gfx.mec2_fw); + adev->gfx.mec2_fw = NULL; + release_firmware(adev->gfx.rlc_fw); + adev->gfx.rlc_fw = NULL; +} + /** * gfx_v7_0_tiling_mode_table_init - init the hw tiling table * @@ -1567,9 +1583,15 @@ static void gfx_v7_0_tiling_mode_table_init(struct amdgpu_device *adev) * registers are instanced per SE or SH. 0xffffffff means * broadcast to all SEs or SHs (CIK). */ -void gfx_v7_0_select_se_sh(struct amdgpu_device *adev, u32 se_num, u32 sh_num) +static void gfx_v7_0_select_se_sh(struct amdgpu_device *adev, + u32 se_num, u32 sh_num, u32 instance) { - u32 data = GRBM_GFX_INDEX__INSTANCE_BROADCAST_WRITES_MASK; + u32 data; + + if (instance == 0xffffffff) + data = REG_SET_FIELD(0, GRBM_GFX_INDEX, INSTANCE_BROADCAST_WRITES, 1); + else + data = REG_SET_FIELD(0, GRBM_GFX_INDEX, INSTANCE_INDEX, instance); if ((se_num == 0xffffffff) && (sh_num == 0xffffffff)) data |= GRBM_GFX_INDEX__SH_BROADCAST_WRITES_MASK | @@ -1643,13 +1665,13 @@ static void gfx_v7_0_setup_rb(struct amdgpu_device *adev) mutex_lock(&adev->grbm_idx_mutex); for (i = 0; i < adev->gfx.config.max_shader_engines; i++) { for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) { - gfx_v7_0_select_se_sh(adev, i, j); + gfx_v7_0_select_se_sh(adev, i, j, 0xffffffff); data = gfx_v7_0_get_rb_active_bitmap(adev); active_rbs |= data << ((i * adev->gfx.config.max_sh_per_se + j) * rb_bitmap_width_per_sh); } } - gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff); + gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); mutex_unlock(&adev->grbm_idx_mutex); adev->gfx.config.backend_enable_mask = active_rbs; @@ -1730,7 +1752,7 @@ static void gfx_v7_0_gpu_init(struct amdgpu_device *adev) * making sure that the following register writes will be broadcasted * to all the shaders */ - gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff); + gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); /* XXX SH_MEM regs */ /* where to put LDS, scratch, GPUVM in FSA64 space */ @@ -2034,17 +2056,6 @@ static void gfx_v7_0_ring_emit_ib_gfx(struct amdgpu_ring *ring, unsigned vm_id, bool ctx_switch) { u32 header, control = 0; - u32 next_rptr = ring->wptr + 5; - - if (ctx_switch) - next_rptr += 2; - - next_rptr += 4; - amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3)); - amdgpu_ring_write(ring, WRITE_DATA_DST_SEL(5) | WR_CONFIRM); - amdgpu_ring_write(ring, ring->next_rptr_gpu_addr & 0xfffffffc); - amdgpu_ring_write(ring, upper_32_bits(ring->next_rptr_gpu_addr) & 0xffffffff); - amdgpu_ring_write(ring, next_rptr); /* insert SWITCH_BUFFER packet before first IB in the ring frame */ if (ctx_switch) { @@ -2073,22 +2084,9 @@ static void gfx_v7_0_ring_emit_ib_compute(struct amdgpu_ring *ring, struct amdgpu_ib *ib, unsigned vm_id, bool ctx_switch) { - u32 header, control = 0; - u32 next_rptr = ring->wptr + 5; + u32 control = INDIRECT_BUFFER_VALID | ib->length_dw | (vm_id << 24); - control |= INDIRECT_BUFFER_VALID; - next_rptr += 4; - amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3)); - amdgpu_ring_write(ring, WRITE_DATA_DST_SEL(5) | WR_CONFIRM); - amdgpu_ring_write(ring, ring->next_rptr_gpu_addr & 0xfffffffc); - amdgpu_ring_write(ring, upper_32_bits(ring->next_rptr_gpu_addr) & 0xffffffff); - amdgpu_ring_write(ring, next_rptr); - - header = PACKET3(PACKET3_INDIRECT_BUFFER, 2); - - control |= ib->length_dw | (vm_id << 24); - - amdgpu_ring_write(ring, header); + amdgpu_ring_write(ring, PACKET3(PACKET3_INDIRECT_BUFFER, 2)); amdgpu_ring_write(ring, #ifdef __BIG_ENDIAN (2 << 0) | @@ -2107,26 +2105,25 @@ static void gfx_v7_0_ring_emit_ib_compute(struct amdgpu_ring *ring, * Provides a basic gfx ring test to verify that IBs are working. * Returns 0 on success, error on failure. */ -static int gfx_v7_0_ring_test_ib(struct amdgpu_ring *ring) +static int gfx_v7_0_ring_test_ib(struct amdgpu_ring *ring, long timeout) { struct amdgpu_device *adev = ring->adev; struct amdgpu_ib ib; struct fence *f = NULL; uint32_t scratch; uint32_t tmp = 0; - unsigned i; - int r; + long r; r = amdgpu_gfx_scratch_get(adev, &scratch); if (r) { - DRM_ERROR("amdgpu: failed to get scratch reg (%d).\n", r); + DRM_ERROR("amdgpu: failed to get scratch reg (%ld).\n", r); return r; } WREG32(scratch, 0xCAFEDEAD); memset(&ib, 0, sizeof(ib)); r = amdgpu_ib_get(adev, NULL, 256, &ib); if (r) { - DRM_ERROR("amdgpu: failed to get ib (%d).\n", r); + DRM_ERROR("amdgpu: failed to get ib (%ld).\n", r); goto err1; } ib.ptr[0] = PACKET3(PACKET3_SET_UCONFIG_REG, 1); @@ -2138,21 +2135,19 @@ static int gfx_v7_0_ring_test_ib(struct amdgpu_ring *ring) if (r) goto err2; - r = fence_wait(f, false); - if (r) { - DRM_ERROR("amdgpu: fence wait failed (%d).\n", r); + r = fence_wait_timeout(f, false, timeout); + if (r == 0) { + DRM_ERROR("amdgpu: IB test timed out\n"); + r = -ETIMEDOUT; goto err2; - } - for (i = 0; i < adev->usec_timeout; i++) { - tmp = RREG32(scratch); - if (tmp == 0xDEADBEEF) - break; - DRM_UDELAY(1); - } - if (i < adev->usec_timeout) { - DRM_INFO("ib test on ring %d succeeded in %u usecs\n", - ring->idx, i); + } else if (r < 0) { + DRM_ERROR("amdgpu: fence wait failed (%ld).\n", r); goto err2; + } + tmp = RREG32(scratch); + if (tmp == 0xDEADBEEF) { + DRM_INFO("ib test on ring %d succeeded\n", ring->idx); + r = 0; } else { DRM_ERROR("amdgpu: ib test failed (scratch(0x%04X)=0x%08X)\n", scratch, tmp); @@ -2160,7 +2155,6 @@ static int gfx_v7_0_ring_test_ib(struct amdgpu_ring *ring) } err2: - fence_put(f); amdgpu_ib_free(adev, &ib, NULL); fence_put(f); err1: @@ -3205,7 +3199,8 @@ static int gfx_v7_0_rlc_init(struct amdgpu_device *adev) } } adev->gfx.rlc.cs_data = ci_cs_data; - adev->gfx.rlc.cp_table_size = CP_ME_TABLE_SIZE * 5 * 4; + adev->gfx.rlc.cp_table_size = ALIGN(CP_ME_TABLE_SIZE * 5 * 4, 2048); /* CP JT */ + adev->gfx.rlc.cp_table_size += 64 * 1024; /* GDS */ src_ptr = adev->gfx.rlc.reg_list; dws = adev->gfx.rlc.reg_list_size; @@ -3363,7 +3358,7 @@ static void gfx_v7_0_wait_for_rlc_serdes(struct amdgpu_device *adev) mutex_lock(&adev->grbm_idx_mutex); for (i = 0; i < adev->gfx.config.max_shader_engines; i++) { for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) { - gfx_v7_0_select_se_sh(adev, i, j); + gfx_v7_0_select_se_sh(adev, i, j, 0xffffffff); for (k = 0; k < adev->usec_timeout; k++) { if (RREG32(mmRLC_SERDES_CU_MASTER_BUSY) == 0) break; @@ -3371,7 +3366,7 @@ static void gfx_v7_0_wait_for_rlc_serdes(struct amdgpu_device *adev) } } } - gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff); + gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); mutex_unlock(&adev->grbm_idx_mutex); mask = RLC_SERDES_NONCU_MASTER_BUSY__SE_MASTER_BUSY_MASK | @@ -3418,7 +3413,7 @@ static u32 gfx_v7_0_halt_rlc(struct amdgpu_device *adev) return orig; } -void gfx_v7_0_enter_rlc_safe_mode(struct amdgpu_device *adev) +static void gfx_v7_0_enter_rlc_safe_mode(struct amdgpu_device *adev) { u32 tmp, i, mask; @@ -3440,7 +3435,7 @@ void gfx_v7_0_enter_rlc_safe_mode(struct amdgpu_device *adev) } } -void gfx_v7_0_exit_rlc_safe_mode(struct amdgpu_device *adev) +static void gfx_v7_0_exit_rlc_safe_mode(struct amdgpu_device *adev) { u32 tmp; @@ -3455,7 +3450,7 @@ void gfx_v7_0_exit_rlc_safe_mode(struct amdgpu_device *adev) * * Halt the RLC ME (MicroEngine) (CIK). */ -void gfx_v7_0_rlc_stop(struct amdgpu_device *adev) +static void gfx_v7_0_rlc_stop(struct amdgpu_device *adev) { WREG32(mmRLC_CNTL, 0); @@ -3531,7 +3526,7 @@ static int gfx_v7_0_rlc_resume(struct amdgpu_device *adev) WREG32(mmRLC_LB_CNTR_MAX, 0x00008000); mutex_lock(&adev->grbm_idx_mutex); - gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff); + gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); WREG32(mmRLC_LB_INIT_CU_MASK, 0xffffffff); WREG32(mmRLC_LB_PARAMS, 0x00600408); WREG32(mmRLC_LB_CNTL, 0x80000004); @@ -3571,7 +3566,7 @@ static void gfx_v7_0_enable_cgcg(struct amdgpu_device *adev, bool enable) tmp = gfx_v7_0_halt_rlc(adev); mutex_lock(&adev->grbm_idx_mutex); - gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff); + gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); WREG32(mmRLC_SERDES_WR_CU_MASTER_MASK, 0xffffffff); WREG32(mmRLC_SERDES_WR_NONCU_MASTER_MASK, 0xffffffff); tmp2 = RLC_SERDES_WR_CTRL__BPM_ADDR_MASK | @@ -3622,7 +3617,7 @@ static void gfx_v7_0_enable_mgcg(struct amdgpu_device *adev, bool enable) tmp = gfx_v7_0_halt_rlc(adev); mutex_lock(&adev->grbm_idx_mutex); - gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff); + gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); WREG32(mmRLC_SERDES_WR_CU_MASTER_MASK, 0xffffffff); WREG32(mmRLC_SERDES_WR_NONCU_MASTER_MASK, 0xffffffff); data = RLC_SERDES_WR_CTRL__BPM_ADDR_MASK | @@ -3673,7 +3668,7 @@ static void gfx_v7_0_enable_mgcg(struct amdgpu_device *adev, bool enable) tmp = gfx_v7_0_halt_rlc(adev); mutex_lock(&adev->grbm_idx_mutex); - gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff); + gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); WREG32(mmRLC_SERDES_WR_CU_MASTER_MASK, 0xffffffff); WREG32(mmRLC_SERDES_WR_NONCU_MASTER_MASK, 0xffffffff); data = RLC_SERDES_WR_CTRL__BPM_ADDR_MASK | RLC_SERDES_WR_CTRL__MGCG_OVERRIDE_1_MASK; @@ -3851,6 +3846,20 @@ static void gfx_v7_0_enable_gfx_cgpg(struct amdgpu_device *adev, } } +static void gfx_v7_0_set_user_cu_inactive_bitmap(struct amdgpu_device *adev, + u32 bitmap) +{ + u32 data; + + if (!bitmap) + return; + + data = bitmap << GC_USER_SHADER_ARRAY_CONFIG__INACTIVE_CUS__SHIFT; + data &= GC_USER_SHADER_ARRAY_CONFIG__INACTIVE_CUS_MASK; + + WREG32(mmGC_USER_SHADER_ARRAY_CONFIG, data); +} + static u32 gfx_v7_0_get_cu_active_bitmap(struct amdgpu_device *adev) { u32 data, mask; @@ -4107,7 +4116,7 @@ static void gfx_v7_0_fini_pg(struct amdgpu_device *adev) * Fetches a GPU clock counter snapshot (SI). * Returns the 64 bit clock counter snapshot. */ -uint64_t gfx_v7_0_get_gpu_clock_counter(struct amdgpu_device *adev) +static uint64_t gfx_v7_0_get_gpu_clock_counter(struct amdgpu_device *adev) { uint64_t clock; @@ -4167,12 +4176,24 @@ static void gfx_v7_0_ring_emit_gds_switch(struct amdgpu_ring *ring, amdgpu_ring_write(ring, (1 << (oa_size + oa_base)) - (1 << oa_base)); } +static const struct amdgpu_gfx_funcs gfx_v7_0_gfx_funcs = { + .get_gpu_clock_counter = &gfx_v7_0_get_gpu_clock_counter, + .select_se_sh = &gfx_v7_0_select_se_sh, +}; + +static const struct amdgpu_rlc_funcs gfx_v7_0_rlc_funcs = { + .enter_safe_mode = gfx_v7_0_enter_rlc_safe_mode, + .exit_safe_mode = gfx_v7_0_exit_rlc_safe_mode +}; + static int gfx_v7_0_early_init(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; adev->gfx.num_gfx_rings = GFX7_NUM_GFX_RINGS; adev->gfx.num_compute_rings = GFX7_NUM_COMPUTE_RINGS; + adev->gfx.funcs = &gfx_v7_0_gfx_funcs; + adev->gfx.rlc.funcs = &gfx_v7_0_rlc_funcs; gfx_v7_0_set_ring_funcs(adev); gfx_v7_0_set_irq_funcs(adev); gfx_v7_0_set_gds_init(adev); @@ -4489,6 +4510,7 @@ static int gfx_v7_0_sw_fini(void *handle) gfx_v7_0_cp_compute_fini(adev); gfx_v7_0_rlc_fini(adev); gfx_v7_0_mec_fini(adev); + gfx_v7_0_free_microcode(adev); return 0; } @@ -4816,7 +4838,7 @@ static int gfx_v7_0_eop_irq(struct amdgpu_device *adev, case 2: for (i = 0; i < adev->gfx.num_compute_rings; i++) { ring = &adev->gfx.compute_ring[i]; - if ((ring->me == me_id) & (ring->pipe == pipe_id)) + if ((ring->me == me_id) && (ring->pipe == pipe_id)) amdgpu_fence_process(ring); } break; @@ -5015,16 +5037,22 @@ static void gfx_v7_0_get_cu_info(struct amdgpu_device *adev) int i, j, k, counter, active_cu_number = 0; u32 mask, bitmap, ao_bitmap, ao_cu_mask = 0; struct amdgpu_cu_info *cu_info = &adev->gfx.cu_info; + unsigned disable_masks[4 * 2]; memset(cu_info, 0, sizeof(*cu_info)); + amdgpu_gfx_parse_disable_cu(disable_masks, 4, 2); + mutex_lock(&adev->grbm_idx_mutex); for (i = 0; i < adev->gfx.config.max_shader_engines; i++) { for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) { mask = 1; ao_bitmap = 0; counter = 0; - gfx_v7_0_select_se_sh(adev, i, j); + gfx_v7_0_select_se_sh(adev, i, j, 0xffffffff); + if (i < 4 && j < 2) + gfx_v7_0_set_user_cu_inactive_bitmap( + adev, disable_masks[i * 2 + j]); bitmap = gfx_v7_0_get_cu_active_bitmap(adev); cu_info->bitmap[i][j] = bitmap; @@ -5040,7 +5068,7 @@ static void gfx_v7_0_get_cu_info(struct amdgpu_device *adev) ao_cu_mask |= (ao_bitmap << (i * 16 + j * 8)); } } - gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff); + gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); mutex_unlock(&adev->grbm_idx_mutex); cu_info->number = active_cu_number; diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.h b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.h index e747aa935c88..94e3ea147c26 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.h +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.h @@ -26,11 +26,4 @@ extern const struct amd_ip_funcs gfx_v7_0_ip_funcs; -/* XXX these shouldn't be exported */ -void gfx_v7_0_enter_rlc_safe_mode(struct amdgpu_device *adev); -void gfx_v7_0_exit_rlc_safe_mode(struct amdgpu_device *adev); -void gfx_v7_0_rlc_stop(struct amdgpu_device *adev); -uint64_t gfx_v7_0_get_gpu_clock_counter(struct amdgpu_device *adev); -void gfx_v7_0_select_se_sh(struct amdgpu_device *adev, u32 se_num, u32 sh_num); - #endif diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c index f19bab68fd83..bff8668e9e6d 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c @@ -28,6 +28,7 @@ #include "vid.h" #include "amdgpu_ucode.h" #include "amdgpu_atombios.h" +#include "atombios_i2c.h" #include "clearstate_vi.h" #include "gmc/gmc_8_2_d.h" @@ -47,6 +48,8 @@ #include "dce/dce_10_0_d.h" #include "dce/dce_10_0_sh_mask.h" +#include "smu/smu_7_1_3_d.h" + #define GFX8_NUM_GFX_RINGS 1 #define GFX8_NUM_COMPUTE_RINGS 8 @@ -282,6 +285,7 @@ static const u32 golden_settings_polaris11_a11[] = mmTCP_ADDR_CONFIG, 0x000003ff, 0x000000f3, mmTCP_CHAN_STEER_HI, 0xffffffff, 0x00000000, mmTCP_CHAN_STEER_LO, 0xffffffff, 0x00003210, + mmVGT_RESET_DEBUG, 0x00000004, 0x00000004, }; static const u32 polaris11_golden_common_all[] = @@ -297,7 +301,8 @@ static const u32 polaris11_golden_common_all[] = static const u32 golden_settings_polaris10_a11[] = { mmATC_MISC_CG, 0x000c0fc0, 0x000c0200, - mmCB_HW_CONTROL, 0xfffdf3cf, 0x00006208, + mmCB_HW_CONTROL, 0xfffdf3cf, 0x00007208, + mmCB_HW_CONTROL_2, 0, 0x0f000000, mmCB_HW_CONTROL_3, 0x000001ff, 0x00000040, mmDB_DEBUG2, 0xf00fffff, 0x00000400, mmPA_SC_ENHANCE, 0xffffffff, 0x20000001, @@ -311,6 +316,7 @@ static const u32 golden_settings_polaris10_a11[] = mmTCC_CTRL, 0x00100000, 0xf31fff7f, mmTCP_ADDR_CONFIG, 0x000003ff, 0x000000f7, mmTCP_CHAN_STEER_HI, 0xffffffff, 0x00000000, + mmVGT_RESET_DEBUG, 0x00000004, 0x00000004, }; static const u32 polaris10_golden_common_all[] = @@ -692,6 +698,11 @@ static void gfx_v8_0_init_golden_registers(struct amdgpu_device *adev) amdgpu_program_register_sequence(adev, polaris10_golden_common_all, (const u32)ARRAY_SIZE(polaris10_golden_common_all)); + WREG32_SMC(ixCG_ACLK_CNTL, 0x0000001C); + if (adev->pdev->revision == 0xc7) { + amdgpu_atombios_i2c_channel_trans(adev, 0x10, 0x96, 0x1E, 0xDD); + amdgpu_atombios_i2c_channel_trans(adev, 0x10, 0x96, 0x1F, 0xD0); + } break; case CHIP_CARRIZO: amdgpu_program_register_sequence(adev, @@ -776,26 +787,25 @@ static int gfx_v8_0_ring_test_ring(struct amdgpu_ring *ring) return r; } -static int gfx_v8_0_ring_test_ib(struct amdgpu_ring *ring) +static int gfx_v8_0_ring_test_ib(struct amdgpu_ring *ring, long timeout) { struct amdgpu_device *adev = ring->adev; struct amdgpu_ib ib; struct fence *f = NULL; uint32_t scratch; uint32_t tmp = 0; - unsigned i; - int r; + long r; r = amdgpu_gfx_scratch_get(adev, &scratch); if (r) { - DRM_ERROR("amdgpu: failed to get scratch reg (%d).\n", r); + DRM_ERROR("amdgpu: failed to get scratch reg (%ld).\n", r); return r; } WREG32(scratch, 0xCAFEDEAD); memset(&ib, 0, sizeof(ib)); r = amdgpu_ib_get(adev, NULL, 256, &ib); if (r) { - DRM_ERROR("amdgpu: failed to get ib (%d).\n", r); + DRM_ERROR("amdgpu: failed to get ib (%ld).\n", r); goto err1; } ib.ptr[0] = PACKET3(PACKET3_SET_UCONFIG_REG, 1); @@ -807,28 +817,25 @@ static int gfx_v8_0_ring_test_ib(struct amdgpu_ring *ring) if (r) goto err2; - r = fence_wait(f, false); - if (r) { - DRM_ERROR("amdgpu: fence wait failed (%d).\n", r); + r = fence_wait_timeout(f, false, timeout); + if (r == 0) { + DRM_ERROR("amdgpu: IB test timed out.\n"); + r = -ETIMEDOUT; goto err2; - } - for (i = 0; i < adev->usec_timeout; i++) { - tmp = RREG32(scratch); - if (tmp == 0xDEADBEEF) - break; - DRM_UDELAY(1); - } - if (i < adev->usec_timeout) { - DRM_INFO("ib test on ring %d succeeded in %u usecs\n", - ring->idx, i); + } else if (r < 0) { + DRM_ERROR("amdgpu: fence wait failed (%ld).\n", r); goto err2; + } + tmp = RREG32(scratch); + if (tmp == 0xDEADBEEF) { + DRM_INFO("ib test on ring %d succeeded\n", ring->idx); + r = 0; } else { DRM_ERROR("amdgpu: ib test failed (scratch(0x%04X)=0x%08X)\n", scratch, tmp); r = -EINVAL; } err2: - fence_put(f); amdgpu_ib_free(adev, &ib, NULL); fence_put(f); err1: @@ -836,6 +843,26 @@ err1: return r; } + +static void gfx_v8_0_free_microcode(struct amdgpu_device *adev) { + release_firmware(adev->gfx.pfp_fw); + adev->gfx.pfp_fw = NULL; + release_firmware(adev->gfx.me_fw); + adev->gfx.me_fw = NULL; + release_firmware(adev->gfx.ce_fw); + adev->gfx.ce_fw = NULL; + release_firmware(adev->gfx.rlc_fw); + adev->gfx.rlc_fw = NULL; + release_firmware(adev->gfx.mec_fw); + adev->gfx.mec_fw = NULL; + if ((adev->asic_type != CHIP_STONEY) && + (adev->asic_type != CHIP_TOPAZ)) + release_firmware(adev->gfx.mec2_fw); + adev->gfx.mec2_fw = NULL; + + kfree(adev->gfx.rlc.register_list_format); +} + static int gfx_v8_0_init_microcode(struct amdgpu_device *adev) { const char *chip_name; @@ -1129,6 +1156,71 @@ static void gfx_v8_0_get_csb_buffer(struct amdgpu_device *adev, buffer[count++] = cpu_to_le32(0); } +static void cz_init_cp_jump_table(struct amdgpu_device *adev) +{ + const __le32 *fw_data; + volatile u32 *dst_ptr; + int me, i, max_me = 4; + u32 bo_offset = 0; + u32 table_offset, table_size; + + if (adev->asic_type == CHIP_CARRIZO) + max_me = 5; + + /* write the cp table buffer */ + dst_ptr = adev->gfx.rlc.cp_table_ptr; + for (me = 0; me < max_me; me++) { + if (me == 0) { + const struct gfx_firmware_header_v1_0 *hdr = + (const struct gfx_firmware_header_v1_0 *)adev->gfx.ce_fw->data; + fw_data = (const __le32 *) + (adev->gfx.ce_fw->data + + le32_to_cpu(hdr->header.ucode_array_offset_bytes)); + table_offset = le32_to_cpu(hdr->jt_offset); + table_size = le32_to_cpu(hdr->jt_size); + } else if (me == 1) { + const struct gfx_firmware_header_v1_0 *hdr = + (const struct gfx_firmware_header_v1_0 *)adev->gfx.pfp_fw->data; + fw_data = (const __le32 *) + (adev->gfx.pfp_fw->data + + le32_to_cpu(hdr->header.ucode_array_offset_bytes)); + table_offset = le32_to_cpu(hdr->jt_offset); + table_size = le32_to_cpu(hdr->jt_size); + } else if (me == 2) { + const struct gfx_firmware_header_v1_0 *hdr = + (const struct gfx_firmware_header_v1_0 *)adev->gfx.me_fw->data; + fw_data = (const __le32 *) + (adev->gfx.me_fw->data + + le32_to_cpu(hdr->header.ucode_array_offset_bytes)); + table_offset = le32_to_cpu(hdr->jt_offset); + table_size = le32_to_cpu(hdr->jt_size); + } else if (me == 3) { + const struct gfx_firmware_header_v1_0 *hdr = + (const struct gfx_firmware_header_v1_0 *)adev->gfx.mec_fw->data; + fw_data = (const __le32 *) + (adev->gfx.mec_fw->data + + le32_to_cpu(hdr->header.ucode_array_offset_bytes)); + table_offset = le32_to_cpu(hdr->jt_offset); + table_size = le32_to_cpu(hdr->jt_size); + } else if (me == 4) { + const struct gfx_firmware_header_v1_0 *hdr = + (const struct gfx_firmware_header_v1_0 *)adev->gfx.mec2_fw->data; + fw_data = (const __le32 *) + (adev->gfx.mec2_fw->data + + le32_to_cpu(hdr->header.ucode_array_offset_bytes)); + table_offset = le32_to_cpu(hdr->jt_offset); + table_size = le32_to_cpu(hdr->jt_size); + } + + for (i = 0; i < table_size; i ++) { + dst_ptr[bo_offset + i] = + cpu_to_le32(le32_to_cpu(fw_data[table_offset + i])); + } + + bo_offset += table_size; + } +} + static void gfx_v8_0_rlc_fini(struct amdgpu_device *adev) { int r; @@ -1144,6 +1236,18 @@ static void gfx_v8_0_rlc_fini(struct amdgpu_device *adev) amdgpu_bo_unref(&adev->gfx.rlc.clear_state_obj); adev->gfx.rlc.clear_state_obj = NULL; } + + /* jump table block */ + if (adev->gfx.rlc.cp_table_obj) { + r = amdgpu_bo_reserve(adev->gfx.rlc.cp_table_obj, false); + if (unlikely(r != 0)) + dev_warn(adev->dev, "(%d) reserve RLC cp table bo failed\n", r); + amdgpu_bo_unpin(adev->gfx.rlc.cp_table_obj); + amdgpu_bo_unreserve(adev->gfx.rlc.cp_table_obj); + + amdgpu_bo_unref(&adev->gfx.rlc.cp_table_obj); + adev->gfx.rlc.cp_table_obj = NULL; + } } static int gfx_v8_0_rlc_init(struct amdgpu_device *adev) @@ -1200,6 +1304,46 @@ static int gfx_v8_0_rlc_init(struct amdgpu_device *adev) amdgpu_bo_unreserve(adev->gfx.rlc.clear_state_obj); } + if ((adev->asic_type == CHIP_CARRIZO) || + (adev->asic_type == CHIP_STONEY)) { + adev->gfx.rlc.cp_table_size = ALIGN(96 * 5 * 4, 2048) + (64 * 1024); /* JT + GDS */ + if (adev->gfx.rlc.cp_table_obj == NULL) { + r = amdgpu_bo_create(adev, adev->gfx.rlc.cp_table_size, PAGE_SIZE, true, + AMDGPU_GEM_DOMAIN_VRAM, + AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED, + NULL, NULL, + &adev->gfx.rlc.cp_table_obj); + if (r) { + dev_warn(adev->dev, "(%d) create RLC cp table bo failed\n", r); + return r; + } + } + + r = amdgpu_bo_reserve(adev->gfx.rlc.cp_table_obj, false); + if (unlikely(r != 0)) { + dev_warn(adev->dev, "(%d) reserve RLC cp table bo failed\n", r); + return r; + } + r = amdgpu_bo_pin(adev->gfx.rlc.cp_table_obj, AMDGPU_GEM_DOMAIN_VRAM, + &adev->gfx.rlc.cp_table_gpu_addr); + if (r) { + amdgpu_bo_unreserve(adev->gfx.rlc.cp_table_obj); + dev_warn(adev->dev, "(%d) pin RLC cp_table bo failed\n", r); + return r; + } + r = amdgpu_bo_kmap(adev->gfx.rlc.cp_table_obj, (void **)&adev->gfx.rlc.cp_table_ptr); + if (r) { + dev_warn(adev->dev, "(%d) map RLC cp table bo failed\n", r); + return r; + } + + cz_init_cp_jump_table(adev); + + amdgpu_bo_kunmap(adev->gfx.rlc.cp_table_obj); + amdgpu_bo_unreserve(adev->gfx.rlc.cp_table_obj); + + } + return 0; } @@ -1581,7 +1725,6 @@ static int gfx_v8_0_do_edc_gpr_workarounds(struct amdgpu_device *adev) RREG32(sec_ded_counter_registers[i]); fail: - fence_put(f); amdgpu_ib_free(adev, &ib, NULL); fence_put(f); @@ -1983,7 +2126,7 @@ static int gfx_v8_0_sw_fini(void *handle) gfx_v8_0_rlc_fini(adev); - kfree(adev->gfx.rlc.register_list_format); + gfx_v8_0_free_microcode(adev); return 0; } @@ -3308,9 +3451,15 @@ static void gfx_v8_0_tiling_mode_table_init(struct amdgpu_device *adev) } } -void gfx_v8_0_select_se_sh(struct amdgpu_device *adev, u32 se_num, u32 sh_num) +static void gfx_v8_0_select_se_sh(struct amdgpu_device *adev, + u32 se_num, u32 sh_num, u32 instance) { - u32 data = REG_SET_FIELD(0, GRBM_GFX_INDEX, INSTANCE_BROADCAST_WRITES, 1); + u32 data; + + if (instance == 0xffffffff) + data = REG_SET_FIELD(0, GRBM_GFX_INDEX, INSTANCE_BROADCAST_WRITES, 1); + else + data = REG_SET_FIELD(0, GRBM_GFX_INDEX, INSTANCE_INDEX, instance); if ((se_num == 0xffffffff) && (sh_num == 0xffffffff)) { data = REG_SET_FIELD(data, GRBM_GFX_INDEX, SH_BROADCAST_WRITES, 1); @@ -3360,13 +3509,13 @@ static void gfx_v8_0_setup_rb(struct amdgpu_device *adev) mutex_lock(&adev->grbm_idx_mutex); for (i = 0; i < adev->gfx.config.max_shader_engines; i++) { for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) { - gfx_v8_0_select_se_sh(adev, i, j); + gfx_v8_0_select_se_sh(adev, i, j, 0xffffffff); data = gfx_v8_0_get_rb_active_bitmap(adev); active_rbs |= data << ((i * adev->gfx.config.max_sh_per_se + j) * rb_bitmap_width_per_sh); } } - gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff); + gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); mutex_unlock(&adev->grbm_idx_mutex); adev->gfx.config.backend_enable_mask = active_rbs; @@ -3470,7 +3619,7 @@ static void gfx_v8_0_gpu_init(struct amdgpu_device *adev) * making sure that the following register writes will be broadcasted * to all the shaders */ - gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff); + gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); WREG32(mmPA_SC_FIFO_SIZE, (adev->gfx.config.sc_prim_fifo_size_frontend << @@ -3493,7 +3642,7 @@ static void gfx_v8_0_wait_for_rlc_serdes(struct amdgpu_device *adev) mutex_lock(&adev->grbm_idx_mutex); for (i = 0; i < adev->gfx.config.max_shader_engines; i++) { for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) { - gfx_v8_0_select_se_sh(adev, i, j); + gfx_v8_0_select_se_sh(adev, i, j, 0xffffffff); for (k = 0; k < adev->usec_timeout; k++) { if (RREG32(mmRLC_SERDES_CU_MASTER_BUSY) == 0) break; @@ -3501,7 +3650,7 @@ static void gfx_v8_0_wait_for_rlc_serdes(struct amdgpu_device *adev) } } } - gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff); + gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); mutex_unlock(&adev->grbm_idx_mutex); mask = RLC_SERDES_NONCU_MASTER_BUSY__SE_MASTER_BUSY_MASK | @@ -3662,13 +3811,13 @@ static void gfx_v8_0_enable_save_restore_machine(struct amdgpu_device *adev) WREG32(mmRLC_SRM_CNTL, data); } -static void polaris11_init_power_gating(struct amdgpu_device *adev) +static void gfx_v8_0_init_power_gating(struct amdgpu_device *adev) { uint32_t data; if (adev->pg_flags & (AMD_PG_SUPPORT_GFX_PG | - AMD_PG_SUPPORT_GFX_SMG | - AMD_PG_SUPPORT_GFX_DMG)) { + AMD_PG_SUPPORT_GFX_SMG | + AMD_PG_SUPPORT_GFX_DMG)) { data = RREG32(mmCP_RB_WPTR_POLL_CNTL); data &= ~CP_RB_WPTR_POLL_CNTL__IDLE_POLL_COUNT_MASK; data |= (0x60 << CP_RB_WPTR_POLL_CNTL__IDLE_POLL_COUNT__SHIFT); @@ -3693,6 +3842,53 @@ static void polaris11_init_power_gating(struct amdgpu_device *adev) } } +static void cz_enable_sck_slow_down_on_power_up(struct amdgpu_device *adev, + bool enable) +{ + u32 data, orig; + + orig = data = RREG32(mmRLC_PG_CNTL); + + if (enable) + data |= RLC_PG_CNTL__SMU_CLK_SLOWDOWN_ON_PU_ENABLE_MASK; + else + data &= ~RLC_PG_CNTL__SMU_CLK_SLOWDOWN_ON_PU_ENABLE_MASK; + + if (orig != data) + WREG32(mmRLC_PG_CNTL, data); +} + +static void cz_enable_sck_slow_down_on_power_down(struct amdgpu_device *adev, + bool enable) +{ + u32 data, orig; + + orig = data = RREG32(mmRLC_PG_CNTL); + + if (enable) + data |= RLC_PG_CNTL__SMU_CLK_SLOWDOWN_ON_PD_ENABLE_MASK; + else + data &= ~RLC_PG_CNTL__SMU_CLK_SLOWDOWN_ON_PD_ENABLE_MASK; + + if (orig != data) + WREG32(mmRLC_PG_CNTL, data); +} + +static void cz_enable_cp_power_gating(struct amdgpu_device *adev, bool enable) +{ + u32 data, orig; + + orig = data = RREG32(mmRLC_PG_CNTL); + + if (enable) + data &= ~RLC_PG_CNTL__CP_PG_DISABLE_MASK; + else + data |= RLC_PG_CNTL__CP_PG_DISABLE_MASK; + + if (orig != data) + WREG32(mmRLC_PG_CNTL, data); +} + static void gfx_v8_0_init_pg(struct amdgpu_device *adev) { if (adev->pg_flags & (AMD_PG_SUPPORT_GFX_PG | @@ -3705,8 +3901,25 @@ static void gfx_v8_0_init_pg(struct amdgpu_device *adev) gfx_v8_0_init_save_restore_list(adev); gfx_v8_0_enable_save_restore_machine(adev); - if (adev->asic_type == CHIP_POLARIS11) - polaris11_init_power_gating(adev); + if ((adev->asic_type == CHIP_CARRIZO) || + (adev->asic_type == CHIP_STONEY)) { + WREG32(mmRLC_JUMP_TABLE_RESTORE, adev->gfx.rlc.cp_table_gpu_addr >> 8); + gfx_v8_0_init_power_gating(adev); + WREG32(mmRLC_PG_ALWAYS_ON_CU_MASK, adev->gfx.cu_info.ao_cu_mask); + if (adev->pg_flags & AMD_PG_SUPPORT_RLC_SMU_HS) { + cz_enable_sck_slow_down_on_power_up(adev, true); + cz_enable_sck_slow_down_on_power_down(adev, true); + } else { + cz_enable_sck_slow_down_on_power_up(adev, false); + cz_enable_sck_slow_down_on_power_down(adev, false); + } + if (adev->pg_flags & AMD_PG_SUPPORT_CP) + cz_enable_cp_power_gating(adev, true); + else + cz_enable_cp_power_gating(adev, false); + } else if (adev->asic_type == CHIP_POLARIS11) { + gfx_v8_0_init_power_gating(adev); + } } } @@ -3974,11 +4187,15 @@ static int gfx_v8_0_cp_gfx_start(struct amdgpu_device *adev) amdgpu_ring_write(ring, 0x3a00161a); amdgpu_ring_write(ring, 0x0000002e); break; - case CHIP_TOPAZ: case CHIP_CARRIZO: amdgpu_ring_write(ring, 0x00000002); amdgpu_ring_write(ring, 0x00000000); break; + case CHIP_TOPAZ: + amdgpu_ring_write(ring, adev->gfx.config.num_rbs == 1 ? + 0x00000000 : 0x00000002); + amdgpu_ring_write(ring, 0x00000000); + break; case CHIP_STONEY: amdgpu_ring_write(ring, 0x00000000); amdgpu_ring_write(ring, 0x00000000); @@ -4941,7 +5158,7 @@ static int gfx_v8_0_soft_reset(void *handle) * Fetches a GPU clock counter snapshot. * Returns the 64 bit clock counter snapshot. */ -uint64_t gfx_v8_0_get_gpu_clock_counter(struct amdgpu_device *adev) +static uint64_t gfx_v8_0_get_gpu_clock_counter(struct amdgpu_device *adev) { uint64_t clock; @@ -5001,12 +5218,18 @@ static void gfx_v8_0_ring_emit_gds_switch(struct amdgpu_ring *ring, amdgpu_ring_write(ring, (1 << (oa_size + oa_base)) - (1 << oa_base)); } +static const struct amdgpu_gfx_funcs gfx_v8_0_gfx_funcs = { + .get_gpu_clock_counter = &gfx_v8_0_get_gpu_clock_counter, + .select_se_sh = &gfx_v8_0_select_se_sh, +}; + static int gfx_v8_0_early_init(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; adev->gfx.num_gfx_rings = GFX8_NUM_GFX_RINGS; adev->gfx.num_compute_rings = GFX8_NUM_COMPUTE_RINGS; + adev->gfx.funcs = &gfx_v8_0_gfx_funcs; gfx_v8_0_set_ring_funcs(adev); gfx_v8_0_set_irq_funcs(adev); gfx_v8_0_set_gds_init(adev); @@ -5039,51 +5262,43 @@ static int gfx_v8_0_late_init(void *handle) return 0; } -static void polaris11_enable_gfx_static_mg_power_gating(struct amdgpu_device *adev, - bool enable) +static void gfx_v8_0_enable_gfx_static_mg_power_gating(struct amdgpu_device *adev, + bool enable) { uint32_t data, temp; - /* Send msg to SMU via Powerplay */ - amdgpu_set_powergating_state(adev, - AMD_IP_BLOCK_TYPE_SMC, - enable ? AMD_PG_STATE_GATE : AMD_PG_STATE_UNGATE); + if (adev->asic_type == CHIP_POLARIS11) + /* Send msg to SMU via Powerplay */ + amdgpu_set_powergating_state(adev, + AMD_IP_BLOCK_TYPE_SMC, + enable ? + AMD_PG_STATE_GATE : AMD_PG_STATE_UNGATE); - if (enable) { - /* Enable static MGPG */ - temp = data = RREG32(mmRLC_PG_CNTL); + temp = data = RREG32(mmRLC_PG_CNTL); + /* Enable static MGPG */ + if (enable) data |= RLC_PG_CNTL__STATIC_PER_CU_PG_ENABLE_MASK; - - if (temp != data) - WREG32(mmRLC_PG_CNTL, data); - } else { - temp = data = RREG32(mmRLC_PG_CNTL); + else data &= ~RLC_PG_CNTL__STATIC_PER_CU_PG_ENABLE_MASK; - if (temp != data) - WREG32(mmRLC_PG_CNTL, data); - } + if (temp != data) + WREG32(mmRLC_PG_CNTL, data); } -static void polaris11_enable_gfx_dynamic_mg_power_gating(struct amdgpu_device *adev, - bool enable) +static void gfx_v8_0_enable_gfx_dynamic_mg_power_gating(struct amdgpu_device *adev, + bool enable) { uint32_t data, temp; - if (enable) { - /* Enable dynamic MGPG */ - temp = data = RREG32(mmRLC_PG_CNTL); + temp = data = RREG32(mmRLC_PG_CNTL); + /* Enable dynamic MGPG */ + if (enable) data |= RLC_PG_CNTL__DYN_PER_CU_PG_ENABLE_MASK; - - if (temp != data) - WREG32(mmRLC_PG_CNTL, data); - } else { - temp = data = RREG32(mmRLC_PG_CNTL); + else data &= ~RLC_PG_CNTL__DYN_PER_CU_PG_ENABLE_MASK; - if (temp != data) - WREG32(mmRLC_PG_CNTL, data); - } + if (temp != data) + WREG32(mmRLC_PG_CNTL, data); } static void polaris11_enable_gfx_quick_mg_power_gating(struct amdgpu_device *adev, @@ -5091,19 +5306,63 @@ static void polaris11_enable_gfx_quick_mg_power_gating(struct amdgpu_device *ade { uint32_t data, temp; - if (enable) { - /* Enable quick PG */ - temp = data = RREG32(mmRLC_PG_CNTL); - data |= 0x100000; + temp = data = RREG32(mmRLC_PG_CNTL); + /* Enable quick PG */ + if (enable) + data |= RLC_PG_CNTL__QUICK_PG_ENABLE_MASK; + else + data &= ~RLC_PG_CNTL__QUICK_PG_ENABLE_MASK; - if (temp != data) - WREG32(mmRLC_PG_CNTL, data); - } else { - temp = data = RREG32(mmRLC_PG_CNTL); - data &= ~0x100000; + if (temp != data) + WREG32(mmRLC_PG_CNTL, data); +} - if (temp != data) - WREG32(mmRLC_PG_CNTL, data); +static void cz_enable_gfx_cg_power_gating(struct amdgpu_device *adev, + bool enable) +{ + u32 data, orig; + + orig = data = RREG32(mmRLC_PG_CNTL); + + if (enable) + data |= RLC_PG_CNTL__GFX_POWER_GATING_ENABLE_MASK; + else + data &= ~RLC_PG_CNTL__GFX_POWER_GATING_ENABLE_MASK; + + if (orig != data) + WREG32(mmRLC_PG_CNTL, data); +} + +static void cz_enable_gfx_pipeline_power_gating(struct amdgpu_device *adev, + bool enable) +{ + u32 data, orig; + + orig = data = RREG32(mmRLC_PG_CNTL); + + if (enable) + data |= RLC_PG_CNTL__GFX_PIPELINE_PG_ENABLE_MASK; + else + data &= ~RLC_PG_CNTL__GFX_PIPELINE_PG_ENABLE_MASK; + + if (orig != data) + WREG32(mmRLC_PG_CNTL, data); + + /* Read any GFX register to wake up GFX. */ + if (!enable) + data = RREG32(mmDB_RENDER_CONTROL); +} + +static void cz_update_gfx_cg_power_gating(struct amdgpu_device *adev, + bool enable) +{ + if ((adev->pg_flags & AMD_PG_SUPPORT_GFX_PG) && enable) { + cz_enable_gfx_cg_power_gating(adev, true); + if (adev->pg_flags & AMD_PG_SUPPORT_GFX_PIPELINE) + cz_enable_gfx_pipeline_power_gating(adev, true); + } else { + cz_enable_gfx_cg_power_gating(adev, false); + cz_enable_gfx_pipeline_power_gating(adev, false); } } @@ -5111,21 +5370,42 @@ static int gfx_v8_0_set_powergating_state(void *handle, enum amd_powergating_state state) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; + bool enable = (state == AMD_PG_STATE_GATE) ? true : false; if (!(adev->pg_flags & AMD_PG_SUPPORT_GFX_PG)) return 0; switch (adev->asic_type) { + case CHIP_CARRIZO: + case CHIP_STONEY: + if (adev->pg_flags & AMD_PG_SUPPORT_GFX_PG) + cz_update_gfx_cg_power_gating(adev, enable); + + if ((adev->pg_flags & AMD_PG_SUPPORT_GFX_SMG) && enable) + gfx_v8_0_enable_gfx_static_mg_power_gating(adev, true); + else + gfx_v8_0_enable_gfx_static_mg_power_gating(adev, false); + + if ((adev->pg_flags & AMD_PG_SUPPORT_GFX_DMG) && enable) + gfx_v8_0_enable_gfx_dynamic_mg_power_gating(adev, true); + else + gfx_v8_0_enable_gfx_dynamic_mg_power_gating(adev, false); + break; case CHIP_POLARIS11: - if (adev->pg_flags & AMD_PG_SUPPORT_GFX_SMG) - polaris11_enable_gfx_static_mg_power_gating(adev, - state == AMD_PG_STATE_GATE ? true : false); - else if (adev->pg_flags & AMD_PG_SUPPORT_GFX_DMG) - polaris11_enable_gfx_dynamic_mg_power_gating(adev, - state == AMD_PG_STATE_GATE ? true : false); + if ((adev->pg_flags & AMD_PG_SUPPORT_GFX_SMG) && enable) + gfx_v8_0_enable_gfx_static_mg_power_gating(adev, true); else - polaris11_enable_gfx_quick_mg_power_gating(adev, - state == AMD_PG_STATE_GATE ? true : false); + gfx_v8_0_enable_gfx_static_mg_power_gating(adev, false); + + if ((adev->pg_flags & AMD_PG_SUPPORT_GFX_DMG) && enable) + gfx_v8_0_enable_gfx_dynamic_mg_power_gating(adev, true); + else + gfx_v8_0_enable_gfx_dynamic_mg_power_gating(adev, false); + + if ((adev->pg_flags & AMD_PG_SUPPORT_GFX_QUICK_MG) && enable) + polaris11_enable_gfx_quick_mg_power_gating(adev, true); + else + polaris11_enable_gfx_quick_mg_power_gating(adev, false); break; default: break; @@ -5139,7 +5419,7 @@ static void gfx_v8_0_send_serdes_cmd(struct amdgpu_device *adev, { uint32_t data; - gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff); + gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); WREG32(mmRLC_SERDES_WR_CU_MASTER_MASK, 0xffffffff); WREG32(mmRLC_SERDES_WR_NONCU_MASTER_MASK, 0xffffffff); @@ -5527,6 +5807,8 @@ static void gfx_v8_0_update_coarse_grain_clock_gating(struct amdgpu_device *adev WREG32(mmRLC_CGCG_CGLS_CTRL, data); } + gfx_v8_0_wait_for_rlc_serdes(adev); + adev->gfx.rlc.funcs->exit_safe_mode(adev); } static int gfx_v8_0_update_gfx_clock_gating(struct amdgpu_device *adev, @@ -5652,17 +5934,6 @@ static void gfx_v8_0_ring_emit_ib_gfx(struct amdgpu_ring *ring, unsigned vm_id, bool ctx_switch) { u32 header, control = 0; - u32 next_rptr = ring->wptr + 5; - - if (ctx_switch) - next_rptr += 2; - - next_rptr += 4; - amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3)); - amdgpu_ring_write(ring, WRITE_DATA_DST_SEL(5) | WR_CONFIRM); - amdgpu_ring_write(ring, ring->next_rptr_gpu_addr & 0xfffffffc); - amdgpu_ring_write(ring, upper_32_bits(ring->next_rptr_gpu_addr) & 0xffffffff); - amdgpu_ring_write(ring, next_rptr); /* insert SWITCH_BUFFER packet before first IB in the ring frame */ if (ctx_switch) { @@ -5691,23 +5962,9 @@ static void gfx_v8_0_ring_emit_ib_compute(struct amdgpu_ring *ring, struct amdgpu_ib *ib, unsigned vm_id, bool ctx_switch) { - u32 header, control = 0; - u32 next_rptr = ring->wptr + 5; - - control |= INDIRECT_BUFFER_VALID; + u32 control = INDIRECT_BUFFER_VALID | ib->length_dw | (vm_id << 24); - next_rptr += 4; - amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3)); - amdgpu_ring_write(ring, WRITE_DATA_DST_SEL(5) | WR_CONFIRM); - amdgpu_ring_write(ring, ring->next_rptr_gpu_addr & 0xfffffffc); - amdgpu_ring_write(ring, upper_32_bits(ring->next_rptr_gpu_addr) & 0xffffffff); - amdgpu_ring_write(ring, next_rptr); - - header = PACKET3(PACKET3_INDIRECT_BUFFER, 2); - - control |= ib->length_dw | (vm_id << 24); - - amdgpu_ring_write(ring, header); + amdgpu_ring_write(ring, PACKET3(PACKET3_INDIRECT_BUFFER, 2)); amdgpu_ring_write(ring, #ifdef __BIG_ENDIAN (2 << 0) | @@ -6160,9 +6417,9 @@ static void gfx_v8_0_set_rlc_funcs(struct amdgpu_device *adev) { switch (adev->asic_type) { case CHIP_TOPAZ: - case CHIP_STONEY: adev->gfx.rlc.funcs = &iceland_rlc_funcs; break; + case CHIP_STONEY: case CHIP_CARRIZO: adev->gfx.rlc.funcs = &cz_rlc_funcs; break; @@ -6200,6 +6457,20 @@ static void gfx_v8_0_set_gds_init(struct amdgpu_device *adev) } } +static void gfx_v8_0_set_user_cu_inactive_bitmap(struct amdgpu_device *adev, + u32 bitmap) +{ + u32 data; + + if (!bitmap) + return; + + data = bitmap << GC_USER_SHADER_ARRAY_CONFIG__INACTIVE_CUS__SHIFT; + data &= GC_USER_SHADER_ARRAY_CONFIG__INACTIVE_CUS_MASK; + + WREG32(mmGC_USER_SHADER_ARRAY_CONFIG, data); +} + static u32 gfx_v8_0_get_cu_active_bitmap(struct amdgpu_device *adev) { u32 data, mask; @@ -6220,16 +6491,22 @@ static void gfx_v8_0_get_cu_info(struct amdgpu_device *adev) int i, j, k, counter, active_cu_number = 0; u32 mask, bitmap, ao_bitmap, ao_cu_mask = 0; struct amdgpu_cu_info *cu_info = &adev->gfx.cu_info; + unsigned disable_masks[4 * 2]; memset(cu_info, 0, sizeof(*cu_info)); + amdgpu_gfx_parse_disable_cu(disable_masks, 4, 2); + mutex_lock(&adev->grbm_idx_mutex); for (i = 0; i < adev->gfx.config.max_shader_engines; i++) { for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) { mask = 1; ao_bitmap = 0; counter = 0; - gfx_v8_0_select_se_sh(adev, i, j); + gfx_v8_0_select_se_sh(adev, i, j, 0xffffffff); + if (i < 4 && j < 2) + gfx_v8_0_set_user_cu_inactive_bitmap( + adev, disable_masks[i * 2 + j]); bitmap = gfx_v8_0_get_cu_active_bitmap(adev); cu_info->bitmap[i][j] = bitmap; @@ -6245,7 +6522,7 @@ static void gfx_v8_0_get_cu_info(struct amdgpu_device *adev) ao_cu_mask |= (ao_bitmap << (i * 16 + j * 8)); } } - gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff); + gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); mutex_unlock(&adev->grbm_idx_mutex); cu_info->number = active_cu_number; diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.h b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.h index 16a49f53a2fa..bc82c794312c 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.h +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.h @@ -26,7 +26,6 @@ extern const struct amd_ip_funcs gfx_v8_0_ip_funcs; -uint64_t gfx_v8_0_get_gpu_clock_counter(struct amdgpu_device *adev); void gfx_v8_0_select_se_sh(struct amdgpu_device *adev, u32 se_num, u32 sh_num); #endif diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c index 1feb6439cb0b..d24a82bd0c7a 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c @@ -39,6 +39,7 @@ static void gmc_v7_0_set_gart_funcs(struct amdgpu_device *adev); static void gmc_v7_0_set_irq_funcs(struct amdgpu_device *adev); +static int gmc_v7_0_wait_for_idle(void *handle); MODULE_FIRMWARE("radeon/bonaire_mc.bin"); MODULE_FIRMWARE("radeon/hawaii_mc.bin"); @@ -73,39 +74,15 @@ static void gmc_v7_0_init_golden_registers(struct amdgpu_device *adev) } } -/** - * gmc7_mc_wait_for_idle - wait for MC idle callback. - * - * @adev: amdgpu_device pointer - * - * Wait for the MC (memory controller) to be idle. - * (evergreen+). - * Returns 0 if the MC is idle, -1 if not. - */ -int gmc_v7_0_mc_wait_for_idle(struct amdgpu_device *adev) -{ - unsigned i; - u32 tmp; - - for (i = 0; i < adev->usec_timeout; i++) { - /* read MC_STATUS */ - tmp = RREG32(mmSRBM_STATUS) & 0x1F00; - if (!tmp) - return 0; - udelay(1); - } - return -1; -} - -void gmc_v7_0_mc_stop(struct amdgpu_device *adev, - struct amdgpu_mode_mc_save *save) +static void gmc_v7_0_mc_stop(struct amdgpu_device *adev, + struct amdgpu_mode_mc_save *save) { u32 blackout; if (adev->mode_info.num_crtc) amdgpu_display_stop_mc_access(adev, save); - amdgpu_asic_wait_for_mc_idle(adev); + gmc_v7_0_wait_for_idle((void *)adev); blackout = RREG32(mmMC_SHARED_BLACKOUT_CNTL); if (REG_GET_FIELD(blackout, MC_SHARED_BLACKOUT_CNTL, BLACKOUT_MODE) != 1) { @@ -120,8 +97,8 @@ void gmc_v7_0_mc_stop(struct amdgpu_device *adev, udelay(100); } -void gmc_v7_0_mc_resume(struct amdgpu_device *adev, - struct amdgpu_mode_mc_save *save) +static void gmc_v7_0_mc_resume(struct amdgpu_device *adev, + struct amdgpu_mode_mc_save *save) { u32 tmp; @@ -311,7 +288,7 @@ static void gmc_v7_0_mc_program(struct amdgpu_device *adev) amdgpu_display_set_vga_render_state(adev, false); gmc_v7_0_mc_stop(adev, &save); - if (amdgpu_asic_wait_for_mc_idle(adev)) { + if (gmc_v7_0_wait_for_idle((void *)adev)) { dev_warn(adev->dev, "Wait for MC idle timedout !\n"); } /* Update configuration */ @@ -331,7 +308,7 @@ static void gmc_v7_0_mc_program(struct amdgpu_device *adev) WREG32(mmMC_VM_AGP_BASE, 0); WREG32(mmMC_VM_AGP_TOP, 0x0FFFFFFF); WREG32(mmMC_VM_AGP_BOT, 0x0FFFFFFF); - if (amdgpu_asic_wait_for_mc_idle(adev)) { + if (gmc_v7_0_wait_for_idle((void *)adev)) { dev_warn(adev->dev, "Wait for MC idle timedout !\n"); } gmc_v7_0_mc_resume(adev, &save); @@ -1137,7 +1114,7 @@ static int gmc_v7_0_soft_reset(void *handle) if (srbm_soft_reset) { gmc_v7_0_mc_stop(adev, &save); - if (gmc_v7_0_wait_for_idle(adev)) { + if (gmc_v7_0_wait_for_idle((void *)adev)) { dev_warn(adev->dev, "Wait for GMC idle timed out !\n"); } diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.h b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.h index 36fcbbc46ada..0b386b5d2f7a 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.h +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.h @@ -26,11 +26,4 @@ extern const struct amd_ip_funcs gmc_v7_0_ip_funcs; -/* XXX these shouldn't be exported */ -void gmc_v7_0_mc_stop(struct amdgpu_device *adev, - struct amdgpu_mode_mc_save *save); -void gmc_v7_0_mc_resume(struct amdgpu_device *adev, - struct amdgpu_mode_mc_save *save); -int gmc_v7_0_mc_wait_for_idle(struct amdgpu_device *adev); - #endif diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c index 9945d5bbf1fe..717359d3ba8c 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c @@ -41,6 +41,7 @@ static void gmc_v8_0_set_gart_funcs(struct amdgpu_device *adev); static void gmc_v8_0_set_irq_funcs(struct amdgpu_device *adev); +static int gmc_v8_0_wait_for_idle(void *handle); MODULE_FIRMWARE("amdgpu/tonga_mc.bin"); MODULE_FIRMWARE("amdgpu/polaris11_mc.bin"); @@ -147,44 +148,15 @@ static void gmc_v8_0_init_golden_registers(struct amdgpu_device *adev) } } -/** - * gmc8_mc_wait_for_idle - wait for MC idle callback. - * - * @adev: amdgpu_device pointer - * - * Wait for the MC (memory controller) to be idle. - * (evergreen+). - * Returns 0 if the MC is idle, -1 if not. - */ -int gmc_v8_0_mc_wait_for_idle(struct amdgpu_device *adev) -{ - unsigned i; - u32 tmp; - - for (i = 0; i < adev->usec_timeout; i++) { - /* read MC_STATUS */ - tmp = RREG32(mmSRBM_STATUS) & (SRBM_STATUS__VMC_BUSY_MASK | - SRBM_STATUS__MCB_BUSY_MASK | - SRBM_STATUS__MCB_NON_DISPLAY_BUSY_MASK | - SRBM_STATUS__MCC_BUSY_MASK | - SRBM_STATUS__MCD_BUSY_MASK | - SRBM_STATUS__VMC1_BUSY_MASK); - if (!tmp) - return 0; - udelay(1); - } - return -1; -} - -void gmc_v8_0_mc_stop(struct amdgpu_device *adev, - struct amdgpu_mode_mc_save *save) +static void gmc_v8_0_mc_stop(struct amdgpu_device *adev, + struct amdgpu_mode_mc_save *save) { u32 blackout; if (adev->mode_info.num_crtc) amdgpu_display_stop_mc_access(adev, save); - amdgpu_asic_wait_for_mc_idle(adev); + gmc_v8_0_wait_for_idle(adev); blackout = RREG32(mmMC_SHARED_BLACKOUT_CNTL); if (REG_GET_FIELD(blackout, MC_SHARED_BLACKOUT_CNTL, BLACKOUT_MODE) != 1) { @@ -199,8 +171,8 @@ void gmc_v8_0_mc_stop(struct amdgpu_device *adev, udelay(100); } -void gmc_v8_0_mc_resume(struct amdgpu_device *adev, - struct amdgpu_mode_mc_save *save) +static void gmc_v8_0_mc_resume(struct amdgpu_device *adev, + struct amdgpu_mode_mc_save *save) { u32 tmp; @@ -393,7 +365,7 @@ static void gmc_v8_0_mc_program(struct amdgpu_device *adev) amdgpu_display_set_vga_render_state(adev, false); gmc_v8_0_mc_stop(adev, &save); - if (amdgpu_asic_wait_for_mc_idle(adev)) { + if (gmc_v8_0_wait_for_idle((void *)adev)) { dev_warn(adev->dev, "Wait for MC idle timedout !\n"); } /* Update configuration */ @@ -413,7 +385,7 @@ static void gmc_v8_0_mc_program(struct amdgpu_device *adev) WREG32(mmMC_VM_AGP_BASE, 0); WREG32(mmMC_VM_AGP_TOP, 0x0FFFFFFF); WREG32(mmMC_VM_AGP_BOT, 0x0FFFFFFF); - if (amdgpu_asic_wait_for_mc_idle(adev)) { + if (gmc_v8_0_wait_for_idle((void *)adev)) { dev_warn(adev->dev, "Wait for MC idle timedout !\n"); } gmc_v8_0_mc_resume(adev, &save); @@ -1140,7 +1112,7 @@ static int gmc_v8_0_soft_reset(void *handle) if (srbm_soft_reset) { gmc_v8_0_mc_stop(adev, &save); - if (gmc_v8_0_wait_for_idle(adev)) { + if (gmc_v8_0_wait_for_idle((void *)adev)) { dev_warn(adev->dev, "Wait for GMC idle timed out !\n"); } diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.h b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.h index 973436086b38..fc5001a8119d 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.h +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.h @@ -26,11 +26,4 @@ extern const struct amd_ip_funcs gmc_v8_0_ip_funcs; -/* XXX these shouldn't be exported */ -void gmc_v8_0_mc_stop(struct amdgpu_device *adev, - struct amdgpu_mode_mc_save *save); -void gmc_v8_0_mc_resume(struct amdgpu_device *adev, - struct amdgpu_mode_mc_save *save); -int gmc_v8_0_mc_wait_for_idle(struct amdgpu_device *adev); - #endif diff --git a/drivers/gpu/drm/amd/amdgpu/iceland_dpm.c b/drivers/gpu/drm/amd/amdgpu/iceland_dpm.c index 460bc8ad37e6..2f078ad6095c 100644 --- a/drivers/gpu/drm/amd/amdgpu/iceland_dpm.c +++ b/drivers/gpu/drm/amd/amdgpu/iceland_dpm.c @@ -24,7 +24,7 @@ #include <linux/firmware.h> #include "drmP.h" #include "amdgpu.h" -#include "iceland_smumgr.h" +#include "iceland_smum.h" MODULE_FIRMWARE("amdgpu/topaz_smc.bin"); @@ -72,6 +72,11 @@ static int iceland_dpm_sw_init(void *handle) static int iceland_dpm_sw_fini(void *handle) { + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + + release_firmware(adev->pm.fw); + adev->pm.fw = NULL; + return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/iceland_smc.c b/drivers/gpu/drm/amd/amdgpu/iceland_smc.c index 52ee08193295..211839913728 100644 --- a/drivers/gpu/drm/amd/amdgpu/iceland_smc.c +++ b/drivers/gpu/drm/amd/amdgpu/iceland_smc.c @@ -25,7 +25,7 @@ #include "drmP.h" #include "amdgpu.h" #include "ppsmc.h" -#include "iceland_smumgr.h" +#include "iceland_smum.h" #include "smu_ucode_xfer_vi.h" #include "amdgpu_ucode.h" @@ -211,7 +211,7 @@ static int iceland_send_msg_to_smc_without_waiting(struct amdgpu_device *adev, PPSMC_Msg msg) { if (!iceland_is_smc_ram_running(adev)) - return -EINVAL;; + return -EINVAL; if (wait_smu_response(adev)) { DRM_ERROR("Failed to send previous message\n"); diff --git a/drivers/gpu/drm/amd/amdgpu/iceland_smumgr.h b/drivers/gpu/drm/amd/amdgpu/iceland_smum.h index 1e0769e110fa..5983e3150cc5 100644 --- a/drivers/gpu/drm/amd/amdgpu/iceland_smumgr.h +++ b/drivers/gpu/drm/amd/amdgpu/iceland_smum.h @@ -21,8 +21,8 @@ * */ -#ifndef ICELAND_SMUMGR_H -#define ICELAND_SMUMGR_H +#ifndef ICELAND_SMUM_H +#define ICELAND_SMUM_H #include "ppsmc.h" diff --git a/drivers/gpu/drm/amd/amdgpu/kv_dpm.c b/drivers/gpu/drm/amd/amdgpu/kv_dpm.c index a789a863d677..a845e883f5fa 100644 --- a/drivers/gpu/drm/amd/amdgpu/kv_dpm.c +++ b/drivers/gpu/drm/amd/amdgpu/kv_dpm.c @@ -191,6 +191,7 @@ static void sumo_construct_vid_mapping_table(struct amdgpu_device *adev, vid_mapping_table->num_entries = i; } +#if 0 static const struct kv_lcac_config_values sx_local_cac_cfg_kv[] = { { 0, 4, 1 }, @@ -289,6 +290,7 @@ static const struct kv_lcac_config_reg cpl_cac_config_reg[] = { { 0xc0400d80, 0x003e0000, 17, 0x3fc00000, 22, 0x0001fffe, 1, 0x00000001, 0 } }; +#endif static const struct kv_pt_config_reg didt_config_kv[] = { @@ -507,19 +509,19 @@ static int kv_enable_didt(struct amdgpu_device *adev, bool enable) pi->caps_db_ramping || pi->caps_td_ramping || pi->caps_tcp_ramping) { - gfx_v7_0_enter_rlc_safe_mode(adev); + adev->gfx.rlc.funcs->enter_safe_mode(adev); if (enable) { ret = kv_program_pt_config_registers(adev, didt_config_kv); if (ret) { - gfx_v7_0_exit_rlc_safe_mode(adev); + adev->gfx.rlc.funcs->exit_safe_mode(adev); return ret; } } kv_do_enable_didt(adev, enable); - gfx_v7_0_exit_rlc_safe_mode(adev); + adev->gfx.rlc.funcs->exit_safe_mode(adev); } return 0; diff --git a/drivers/gpu/drm/amd/amdgpu/ppsmc.h b/drivers/gpu/drm/amd/amdgpu/ppsmc.h index 7837f2ecc357..8463245f424f 100644 --- a/drivers/gpu/drm/amd/amdgpu/ppsmc.h +++ b/drivers/gpu/drm/amd/amdgpu/ppsmc.h @@ -90,7 +90,9 @@ typedef uint8_t PPSMC_Result; #define PPSMC_StartFanControl ((uint8_t)0x5B) #define PPSMC_StopFanControl ((uint8_t)0x5C) #define PPSMC_MSG_NoDisplay ((uint8_t)0x5D) +#define PPSMC_NoDisplay ((uint8_t)0x5D) #define PPSMC_MSG_HasDisplay ((uint8_t)0x5E) +#define PPSMC_HasDisplay ((uint8_t)0x5E) #define PPSMC_MSG_UVDPowerOFF ((uint8_t)0x60) #define PPSMC_MSG_UVDPowerON ((uint8_t)0x61) #define PPSMC_MSG_EnableULV ((uint8_t)0x62) @@ -108,6 +110,7 @@ typedef uint8_t PPSMC_Result; #define PPSMC_MSG_DisableDTE ((uint8_t)0x88) #define PPSMC_MSG_ThrottleOVRDSCLKDS ((uint8_t)0x96) #define PPSMC_MSG_CancelThrottleOVRDSCLKDS ((uint8_t)0x97) +#define PPSMC_MSG_EnableACDCGPIOInterrupt ((uint16_t) 0x149) /* CI/KV/KB */ #define PPSMC_MSG_UVDDPM_SetEnabledMask ((uint16_t) 0x12D) @@ -161,6 +164,7 @@ typedef uint8_t PPSMC_Result; #define PPSMC_MSG_MASTER_DeepSleep_OFF ((uint16_t) 0x190) #define PPSMC_MSG_Remove_DC_Clamp ((uint16_t) 0x191) #define PPSMC_MSG_SetFanPwmMax ((uint16_t) 0x19A) +#define PPSMC_MSG_SetFanRpmMax ((uint16_t) 0x205) #define PPSMC_MSG_ENABLE_THERMAL_DPM ((uint16_t) 0x19C) #define PPSMC_MSG_DISABLE_THERMAL_DPM ((uint16_t) 0x19D) diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c b/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c index f4c3130d3fdb..1351c7e834a2 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c @@ -105,6 +105,15 @@ static void sdma_v2_4_init_golden_registers(struct amdgpu_device *adev) } } +static void sdma_v2_4_free_microcode(struct amdgpu_device *adev) +{ + int i; + for (i = 0; i < adev->sdma.num_instances; i++) { + release_firmware(adev->sdma.instance[i].fw); + adev->sdma.instance[i].fw = NULL; + } +} + /** * sdma_v2_4_init_microcode - load ucode images from disk * @@ -246,19 +255,6 @@ static void sdma_v2_4_ring_emit_ib(struct amdgpu_ring *ring, unsigned vm_id, bool ctx_switch) { u32 vmid = vm_id & 0xf; - u32 next_rptr = ring->wptr + 5; - - while ((next_rptr & 7) != 2) - next_rptr++; - - next_rptr += 6; - - amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_WRITE) | - SDMA_PKT_HEADER_SUB_OP(SDMA_SUBOP_WRITE_LINEAR)); - amdgpu_ring_write(ring, lower_32_bits(ring->next_rptr_gpu_addr) & 0xfffffffc); - amdgpu_ring_write(ring, upper_32_bits(ring->next_rptr_gpu_addr)); - amdgpu_ring_write(ring, SDMA_PKT_WRITE_UNTILED_DW_3_COUNT(1)); - amdgpu_ring_write(ring, next_rptr); /* IB packet must end on a 8 DW boundary */ sdma_v2_4_ring_insert_nop(ring, (10 - (ring->wptr & 7)) % 8); @@ -397,7 +393,7 @@ static void sdma_v2_4_enable(struct amdgpu_device *adev, bool enable) u32 f32_cntl; int i; - if (enable == false) { + if (!enable) { sdma_v2_4_gfx_stop(adev); sdma_v2_4_rlc_stop(adev); } @@ -461,6 +457,8 @@ static int sdma_v2_4_gfx_resume(struct amdgpu_device *adev) /* Initialize the ring buffer's read and write pointers */ WREG32(mmSDMA0_GFX_RB_RPTR + sdma_offsets[i], 0); WREG32(mmSDMA0_GFX_RB_WPTR + sdma_offsets[i], 0); + WREG32(mmSDMA0_GFX_IB_RPTR + sdma_offsets[i], 0); + WREG32(mmSDMA0_GFX_IB_OFFSET + sdma_offsets[i], 0); /* set the wb address whether it's enabled or not */ WREG32(mmSDMA0_GFX_RB_RPTR_ADDR_HI + sdma_offsets[i], @@ -489,7 +487,11 @@ static int sdma_v2_4_gfx_resume(struct amdgpu_device *adev) WREG32(mmSDMA0_GFX_IB_CNTL + sdma_offsets[i], ib_cntl); ring->ready = true; + } + sdma_v2_4_enable(adev, true); + for (i = 0; i < adev->sdma.num_instances; i++) { + ring = &adev->sdma.instance[i].ring; r = amdgpu_ring_test_ring(ring); if (r) { ring->ready = false; @@ -565,23 +567,25 @@ static int sdma_v2_4_start(struct amdgpu_device *adev) { int r; - if (!adev->firmware.smu_load) { - r = sdma_v2_4_load_microcode(adev); - if (r) - return r; - } else { - r = adev->smu.smumgr_funcs->check_fw_load_finish(adev, - AMDGPU_UCODE_ID_SDMA0); - if (r) - return -EINVAL; - r = adev->smu.smumgr_funcs->check_fw_load_finish(adev, - AMDGPU_UCODE_ID_SDMA1); - if (r) - return -EINVAL; + if (!adev->pp_enabled) { + if (!adev->firmware.smu_load) { + r = sdma_v2_4_load_microcode(adev); + if (r) + return r; + } else { + r = adev->smu.smumgr_funcs->check_fw_load_finish(adev, + AMDGPU_UCODE_ID_SDMA0); + if (r) + return -EINVAL; + r = adev->smu.smumgr_funcs->check_fw_load_finish(adev, + AMDGPU_UCODE_ID_SDMA1); + if (r) + return -EINVAL; + } } - /* unhalt the MEs */ - sdma_v2_4_enable(adev, true); + /* halt the engine before programing */ + sdma_v2_4_enable(adev, false); /* start the gfx rings and rlc compute queues */ r = sdma_v2_4_gfx_resume(adev); @@ -664,20 +668,19 @@ static int sdma_v2_4_ring_test_ring(struct amdgpu_ring *ring) * Test a simple IB in the DMA ring (VI). * Returns 0 on success, error on failure. */ -static int sdma_v2_4_ring_test_ib(struct amdgpu_ring *ring) +static int sdma_v2_4_ring_test_ib(struct amdgpu_ring *ring, long timeout) { struct amdgpu_device *adev = ring->adev; struct amdgpu_ib ib; struct fence *f = NULL; - unsigned i; unsigned index; - int r; u32 tmp = 0; u64 gpu_addr; + long r; r = amdgpu_wb_get(adev, &index); if (r) { - dev_err(adev->dev, "(%d) failed to allocate wb slot\n", r); + dev_err(adev->dev, "(%ld) failed to allocate wb slot\n", r); return r; } @@ -687,7 +690,7 @@ static int sdma_v2_4_ring_test_ib(struct amdgpu_ring *ring) memset(&ib, 0, sizeof(ib)); r = amdgpu_ib_get(adev, NULL, 256, &ib); if (r) { - DRM_ERROR("amdgpu: failed to get ib (%d).\n", r); + DRM_ERROR("amdgpu: failed to get ib (%ld).\n", r); goto err0; } @@ -706,28 +709,25 @@ static int sdma_v2_4_ring_test_ib(struct amdgpu_ring *ring) if (r) goto err1; - r = fence_wait(f, false); - if (r) { - DRM_ERROR("amdgpu: fence wait failed (%d).\n", r); + r = fence_wait_timeout(f, false, timeout); + if (r == 0) { + DRM_ERROR("amdgpu: IB test timed out\n"); + r = -ETIMEDOUT; goto err1; - } - for (i = 0; i < adev->usec_timeout; i++) { - tmp = le32_to_cpu(adev->wb.wb[index]); - if (tmp == 0xDEADBEEF) - break; - DRM_UDELAY(1); - } - if (i < adev->usec_timeout) { - DRM_INFO("ib test on ring %d succeeded in %u usecs\n", - ring->idx, i); + } else if (r) { + DRM_ERROR("amdgpu: fence wait failed (%ld).\n", r); goto err1; + } + tmp = le32_to_cpu(adev->wb.wb[index]); + if (tmp == 0xDEADBEEF) { + DRM_INFO("ib test on ring %d succeeded\n", ring->idx); + r = 0; } else { DRM_ERROR("amdgpu: ib test failed (0x%08X)\n", tmp); r = -EINVAL; } err1: - fence_put(f); amdgpu_ib_free(adev, &ib, NULL); fence_put(f); err0: @@ -1012,6 +1012,7 @@ static int sdma_v2_4_sw_fini(void *handle) for (i = 0; i < adev->sdma.num_instances; i++) amdgpu_ring_fini(&adev->sdma.instance[i].ring); + sdma_v2_4_free_microcode(adev); return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c index 31d99b0010f7..653ce5ed55ae 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c @@ -236,6 +236,15 @@ static void sdma_v3_0_init_golden_registers(struct amdgpu_device *adev) } } +static void sdma_v3_0_free_microcode(struct amdgpu_device *adev) +{ + int i; + for (i = 0; i < adev->sdma.num_instances; i++) { + release_firmware(adev->sdma.instance[i].fw); + adev->sdma.instance[i].fw = NULL; + } +} + /** * sdma_v3_0_init_microcode - load ucode images from disk * @@ -406,18 +415,6 @@ static void sdma_v3_0_ring_emit_ib(struct amdgpu_ring *ring, unsigned vm_id, bool ctx_switch) { u32 vmid = vm_id & 0xf; - u32 next_rptr = ring->wptr + 5; - - while ((next_rptr & 7) != 2) - next_rptr++; - next_rptr += 6; - - amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_WRITE) | - SDMA_PKT_HEADER_SUB_OP(SDMA_SUBOP_WRITE_LINEAR)); - amdgpu_ring_write(ring, lower_32_bits(ring->next_rptr_gpu_addr) & 0xfffffffc); - amdgpu_ring_write(ring, upper_32_bits(ring->next_rptr_gpu_addr)); - amdgpu_ring_write(ring, SDMA_PKT_WRITE_UNTILED_DW_3_COUNT(1)); - amdgpu_ring_write(ring, next_rptr); /* IB packet must end on a 8 DW boundary */ sdma_v3_0_ring_insert_nop(ring, (10 - (ring->wptr & 7)) % 8); @@ -607,7 +604,7 @@ static void sdma_v3_0_enable(struct amdgpu_device *adev, bool enable) u32 f32_cntl; int i; - if (enable == false) { + if (!enable) { sdma_v3_0_gfx_stop(adev); sdma_v3_0_rlc_stop(adev); } @@ -672,6 +669,8 @@ static int sdma_v3_0_gfx_resume(struct amdgpu_device *adev) /* Initialize the ring buffer's read and write pointers */ WREG32(mmSDMA0_GFX_RB_RPTR + sdma_offsets[i], 0); WREG32(mmSDMA0_GFX_RB_WPTR + sdma_offsets[i], 0); + WREG32(mmSDMA0_GFX_IB_RPTR + sdma_offsets[i], 0); + WREG32(mmSDMA0_GFX_IB_OFFSET + sdma_offsets[i], 0); /* set the wb address whether it's enabled or not */ WREG32(mmSDMA0_GFX_RB_RPTR_ADDR_HI + sdma_offsets[i], @@ -711,7 +710,15 @@ static int sdma_v3_0_gfx_resume(struct amdgpu_device *adev) WREG32(mmSDMA0_GFX_IB_CNTL + sdma_offsets[i], ib_cntl); ring->ready = true; + } + + /* unhalt the MEs */ + sdma_v3_0_enable(adev, true); + /* enable sdma ring preemption */ + sdma_v3_0_ctx_switch_enable(adev, true); + for (i = 0; i < adev->sdma.num_instances; i++) { + ring = &adev->sdma.instance[i].ring; r = amdgpu_ring_test_ring(ring); if (r) { ring->ready = false; @@ -804,10 +811,9 @@ static int sdma_v3_0_start(struct amdgpu_device *adev) } } - /* unhalt the MEs */ - sdma_v3_0_enable(adev, true); - /* enable sdma ring preemption */ - sdma_v3_0_ctx_switch_enable(adev, true); + /* disble sdma engine before programing it */ + sdma_v3_0_ctx_switch_enable(adev, false); + sdma_v3_0_enable(adev, false); /* start the gfx rings and rlc compute queues */ r = sdma_v3_0_gfx_resume(adev); @@ -890,20 +896,19 @@ static int sdma_v3_0_ring_test_ring(struct amdgpu_ring *ring) * Test a simple IB in the DMA ring (VI). * Returns 0 on success, error on failure. */ -static int sdma_v3_0_ring_test_ib(struct amdgpu_ring *ring) +static int sdma_v3_0_ring_test_ib(struct amdgpu_ring *ring, long timeout) { struct amdgpu_device *adev = ring->adev; struct amdgpu_ib ib; struct fence *f = NULL; - unsigned i; unsigned index; - int r; u32 tmp = 0; u64 gpu_addr; + long r; r = amdgpu_wb_get(adev, &index); if (r) { - dev_err(adev->dev, "(%d) failed to allocate wb slot\n", r); + dev_err(adev->dev, "(%ld) failed to allocate wb slot\n", r); return r; } @@ -913,7 +918,7 @@ static int sdma_v3_0_ring_test_ib(struct amdgpu_ring *ring) memset(&ib, 0, sizeof(ib)); r = amdgpu_ib_get(adev, NULL, 256, &ib); if (r) { - DRM_ERROR("amdgpu: failed to get ib (%d).\n", r); + DRM_ERROR("amdgpu: failed to get ib (%ld).\n", r); goto err0; } @@ -932,27 +937,24 @@ static int sdma_v3_0_ring_test_ib(struct amdgpu_ring *ring) if (r) goto err1; - r = fence_wait(f, false); - if (r) { - DRM_ERROR("amdgpu: fence wait failed (%d).\n", r); + r = fence_wait_timeout(f, false, timeout); + if (r == 0) { + DRM_ERROR("amdgpu: IB test timed out\n"); + r = -ETIMEDOUT; goto err1; - } - for (i = 0; i < adev->usec_timeout; i++) { - tmp = le32_to_cpu(adev->wb.wb[index]); - if (tmp == 0xDEADBEEF) - break; - DRM_UDELAY(1); - } - if (i < adev->usec_timeout) { - DRM_INFO("ib test on ring %d succeeded in %u usecs\n", - ring->idx, i); + } else if (r < 0) { + DRM_ERROR("amdgpu: fence wait failed (%ld).\n", r); goto err1; + } + tmp = le32_to_cpu(adev->wb.wb[index]); + if (tmp == 0xDEADBEEF) { + DRM_INFO("ib test on ring %d succeeded\n", ring->idx); + r = 0; } else { DRM_ERROR("amdgpu: ib test failed (0x%08X)\n", tmp); r = -EINVAL; } err1: - fence_put(f); amdgpu_ib_free(adev, &ib, NULL); fence_put(f); err0: @@ -1247,6 +1249,7 @@ static int sdma_v3_0_sw_fini(void *handle) for (i = 0; i < adev->sdma.num_instances; i++) amdgpu_ring_fini(&adev->sdma.instance[i].ring); + sdma_v3_0_free_microcode(adev); return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/tonga_dpm.c b/drivers/gpu/drm/amd/amdgpu/tonga_dpm.c index b7615cefcac4..f06f6f4dc3a8 100644 --- a/drivers/gpu/drm/amd/amdgpu/tonga_dpm.c +++ b/drivers/gpu/drm/amd/amdgpu/tonga_dpm.c @@ -71,6 +71,11 @@ static int tonga_dpm_sw_init(void *handle) static int tonga_dpm_sw_fini(void *handle) { + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + + release_firmware(adev->pm.fw); + adev->pm.fw = NULL; + return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/tonga_smc.c b/drivers/gpu/drm/amd/amdgpu/tonga_smc.c index 083893dd68c0..940de1836f8f 100644 --- a/drivers/gpu/drm/amd/amdgpu/tonga_smc.c +++ b/drivers/gpu/drm/amd/amdgpu/tonga_smc.c @@ -173,7 +173,7 @@ static int tonga_send_msg_to_smc(struct amdgpu_device *adev, PPSMC_Msg msg) { if (!tonga_is_smc_ram_running(adev)) { - return -EINVAL;; + return -EINVAL; } if (wait_smu_response(adev)) { diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c b/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c index f07551476a70..132e613ed674 100644 --- a/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c +++ b/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c @@ -34,6 +34,8 @@ #include "oss/oss_2_0_d.h" #include "oss/oss_2_0_sh_mask.h" +#include "bif/bif_4_1_d.h" + static void uvd_v4_2_mc_resume(struct amdgpu_device *adev); static void uvd_v4_2_init_cg(struct amdgpu_device *adev); static void uvd_v4_2_set_ring_funcs(struct amdgpu_device *adev); @@ -439,6 +441,32 @@ static void uvd_v4_2_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 seq } /** + * uvd_v4_2_ring_emit_hdp_flush - emit an hdp flush + * + * @ring: amdgpu_ring pointer + * + * Emits an hdp flush. + */ +static void uvd_v4_2_ring_emit_hdp_flush(struct amdgpu_ring *ring) +{ + amdgpu_ring_write(ring, PACKET0(mmHDP_MEM_COHERENCY_FLUSH_CNTL, 0)); + amdgpu_ring_write(ring, 0); +} + +/** + * uvd_v4_2_ring_hdp_invalidate - emit an hdp invalidate + * + * @ring: amdgpu_ring pointer + * + * Emits an hdp invalidate. + */ +static void uvd_v4_2_ring_emit_hdp_invalidate(struct amdgpu_ring *ring) +{ + amdgpu_ring_write(ring, PACKET0(mmHDP_DEBUG0, 0)); + amdgpu_ring_write(ring, 1); +} + +/** * uvd_v4_2_ring_test_ring - register write test * * @ring: amdgpu_ring pointer @@ -499,49 +527,6 @@ static void uvd_v4_2_ring_emit_ib(struct amdgpu_ring *ring, } /** - * uvd_v4_2_ring_test_ib - test ib execution - * - * @ring: amdgpu_ring pointer - * - * Test if we can successfully execute an IB - */ -static int uvd_v4_2_ring_test_ib(struct amdgpu_ring *ring) -{ - struct amdgpu_device *adev = ring->adev; - struct fence *fence = NULL; - int r; - - r = amdgpu_asic_set_uvd_clocks(adev, 53300, 40000); - if (r) { - DRM_ERROR("amdgpu: failed to raise UVD clocks (%d).\n", r); - return r; - } - - r = amdgpu_uvd_get_create_msg(ring, 1, NULL); - if (r) { - DRM_ERROR("amdgpu: failed to get create msg (%d).\n", r); - goto error; - } - - r = amdgpu_uvd_get_destroy_msg(ring, 1, true, &fence); - if (r) { - DRM_ERROR("amdgpu: failed to get destroy ib (%d).\n", r); - goto error; - } - - r = fence_wait(fence, false); - if (r) { - DRM_ERROR("amdgpu: fence wait failed (%d).\n", r); - goto error; - } - DRM_INFO("ib test on ring %d succeeded\n", ring->idx); -error: - fence_put(fence); - amdgpu_asic_set_uvd_clocks(adev, 0, 0); - return r; -} - -/** * uvd_v4_2_mc_resume - memory controller programming * * @adev: amdgpu_device pointer @@ -763,10 +748,14 @@ static const struct amdgpu_ring_funcs uvd_v4_2_ring_funcs = { .parse_cs = amdgpu_uvd_ring_parse_cs, .emit_ib = uvd_v4_2_ring_emit_ib, .emit_fence = uvd_v4_2_ring_emit_fence, + .emit_hdp_flush = uvd_v4_2_ring_emit_hdp_flush, + .emit_hdp_invalidate = uvd_v4_2_ring_emit_hdp_invalidate, .test_ring = uvd_v4_2_ring_test_ring, - .test_ib = uvd_v4_2_ring_test_ib, + .test_ib = amdgpu_uvd_ring_test_ib, .insert_nop = amdgpu_ring_insert_nop, .pad_ib = amdgpu_ring_generic_pad_ib, + .begin_use = amdgpu_uvd_ring_begin_use, + .end_use = amdgpu_uvd_ring_end_use, }; static void uvd_v4_2_set_ring_funcs(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c index e0a76a883d46..101de136ba63 100644 --- a/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c +++ b/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c @@ -31,6 +31,7 @@ #include "uvd/uvd_5_0_sh_mask.h" #include "oss/oss_2_0_d.h" #include "oss/oss_2_0_sh_mask.h" +#include "bif/bif_5_0_d.h" #include "vi.h" static void uvd_v5_0_set_ring_funcs(struct amdgpu_device *adev); @@ -489,6 +490,32 @@ static void uvd_v5_0_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 seq } /** + * uvd_v5_0_ring_emit_hdp_flush - emit an hdp flush + * + * @ring: amdgpu_ring pointer + * + * Emits an hdp flush. + */ +static void uvd_v5_0_ring_emit_hdp_flush(struct amdgpu_ring *ring) +{ + amdgpu_ring_write(ring, PACKET0(mmHDP_MEM_COHERENCY_FLUSH_CNTL, 0)); + amdgpu_ring_write(ring, 0); +} + +/** + * uvd_v5_0_ring_hdp_invalidate - emit an hdp invalidate + * + * @ring: amdgpu_ring pointer + * + * Emits an hdp invalidate. + */ +static void uvd_v5_0_ring_emit_hdp_invalidate(struct amdgpu_ring *ring) +{ + amdgpu_ring_write(ring, PACKET0(mmHDP_DEBUG0, 0)); + amdgpu_ring_write(ring, 1); +} + +/** * uvd_v5_0_ring_test_ring - register write test * * @ring: amdgpu_ring pointer @@ -550,49 +577,6 @@ static void uvd_v5_0_ring_emit_ib(struct amdgpu_ring *ring, amdgpu_ring_write(ring, ib->length_dw); } -/** - * uvd_v5_0_ring_test_ib - test ib execution - * - * @ring: amdgpu_ring pointer - * - * Test if we can successfully execute an IB - */ -static int uvd_v5_0_ring_test_ib(struct amdgpu_ring *ring) -{ - struct amdgpu_device *adev = ring->adev; - struct fence *fence = NULL; - int r; - - r = amdgpu_asic_set_uvd_clocks(adev, 53300, 40000); - if (r) { - DRM_ERROR("amdgpu: failed to raise UVD clocks (%d).\n", r); - return r; - } - - r = amdgpu_uvd_get_create_msg(ring, 1, NULL); - if (r) { - DRM_ERROR("amdgpu: failed to get create msg (%d).\n", r); - goto error; - } - - r = amdgpu_uvd_get_destroy_msg(ring, 1, true, &fence); - if (r) { - DRM_ERROR("amdgpu: failed to get destroy ib (%d).\n", r); - goto error; - } - - r = fence_wait(fence, false); - if (r) { - DRM_ERROR("amdgpu: fence wait failed (%d).\n", r); - goto error; - } - DRM_INFO("ib test on ring %d succeeded\n", ring->idx); -error: - fence_put(fence); - amdgpu_asic_set_uvd_clocks(adev, 0, 0); - return r; -} - static bool uvd_v5_0_is_idle(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; @@ -815,10 +799,14 @@ static const struct amdgpu_ring_funcs uvd_v5_0_ring_funcs = { .parse_cs = amdgpu_uvd_ring_parse_cs, .emit_ib = uvd_v5_0_ring_emit_ib, .emit_fence = uvd_v5_0_ring_emit_fence, + .emit_hdp_flush = uvd_v5_0_ring_emit_hdp_flush, + .emit_hdp_invalidate = uvd_v5_0_ring_emit_hdp_invalidate, .test_ring = uvd_v5_0_ring_test_ring, - .test_ib = uvd_v5_0_ring_test_ib, + .test_ib = amdgpu_uvd_ring_test_ib, .insert_nop = amdgpu_ring_insert_nop, .pad_ib = amdgpu_ring_generic_pad_ib, + .begin_use = amdgpu_uvd_ring_begin_use, + .end_use = amdgpu_uvd_ring_end_use, }; static void uvd_v5_0_set_ring_funcs(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c index c9929d665c01..7f21102bfb99 100644 --- a/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c +++ b/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c @@ -33,6 +33,8 @@ #include "oss/oss_2_0_sh_mask.h" #include "smu/smu_7_1_3_d.h" #include "smu/smu_7_1_3_sh_mask.h" +#include "bif/bif_5_1_d.h" +#include "gmc/gmc_8_1_d.h" #include "vi.h" static void uvd_v6_0_set_ring_funcs(struct amdgpu_device *adev); @@ -385,8 +387,8 @@ static int uvd_v6_0_start(struct amdgpu_device *adev) uint32_t mp_swap_cntl; int i, j, r; - /*disable DPG */ - WREG32_P(mmUVD_POWER_STATUS, 0, ~(1 << 2)); + /* disable DPG */ + WREG32_P(mmUVD_POWER_STATUS, 0, ~UVD_POWER_STATUS__UVD_PG_MODE_MASK); /* disable byte swapping */ lmi_swap_cntl = 0; @@ -405,17 +407,21 @@ static int uvd_v6_0_start(struct amdgpu_device *adev) } /* disable interupt */ - WREG32_P(mmUVD_MASTINT_EN, 0, ~(1 << 1)); + WREG32_P(mmUVD_MASTINT_EN, 0, ~UVD_MASTINT_EN__VCPU_EN_MASK); /* stall UMC and register bus before resetting VCPU */ - WREG32_P(mmUVD_LMI_CTRL2, 1 << 8, ~(1 << 8)); + WREG32_P(mmUVD_LMI_CTRL2, UVD_LMI_CTRL2__STALL_ARB_UMC_MASK, ~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK); mdelay(1); /* put LMI, VCPU, RBC etc... into reset */ - WREG32(mmUVD_SOFT_RESET, UVD_SOFT_RESET__LMI_SOFT_RESET_MASK | - UVD_SOFT_RESET__VCPU_SOFT_RESET_MASK | UVD_SOFT_RESET__LBSI_SOFT_RESET_MASK | - UVD_SOFT_RESET__RBC_SOFT_RESET_MASK | UVD_SOFT_RESET__CSM_SOFT_RESET_MASK | - UVD_SOFT_RESET__CXW_SOFT_RESET_MASK | UVD_SOFT_RESET__TAP_SOFT_RESET_MASK | + WREG32(mmUVD_SOFT_RESET, + UVD_SOFT_RESET__LMI_SOFT_RESET_MASK | + UVD_SOFT_RESET__VCPU_SOFT_RESET_MASK | + UVD_SOFT_RESET__LBSI_SOFT_RESET_MASK | + UVD_SOFT_RESET__RBC_SOFT_RESET_MASK | + UVD_SOFT_RESET__CSM_SOFT_RESET_MASK | + UVD_SOFT_RESET__CXW_SOFT_RESET_MASK | + UVD_SOFT_RESET__TAP_SOFT_RESET_MASK | UVD_SOFT_RESET__LMI_UMC_SOFT_RESET_MASK); mdelay(5); @@ -424,8 +430,13 @@ static int uvd_v6_0_start(struct amdgpu_device *adev) mdelay(5); /* initialize UVD memory controller */ - WREG32(mmUVD_LMI_CTRL, 0x40 | (1 << 8) | (1 << 13) | - (1 << 21) | (1 << 9) | (1 << 20)); + WREG32(mmUVD_LMI_CTRL, + (0x40 << UVD_LMI_CTRL__WRITE_CLEAN_TIMER__SHIFT) | + UVD_LMI_CTRL__WRITE_CLEAN_TIMER_EN_MASK | + UVD_LMI_CTRL__DATA_COHERENCY_EN_MASK | + UVD_LMI_CTRL__VCPU_DATA_COHERENCY_EN_MASK | + UVD_LMI_CTRL__REQ_MODE_MASK | + UVD_LMI_CTRL__DISABLE_ON_FWV_FAIL_MASK); #ifdef __BIG_ENDIAN /* swap (8 in 32) RB and IB */ @@ -447,10 +458,10 @@ static int uvd_v6_0_start(struct amdgpu_device *adev) mdelay(5); /* enable VCPU clock */ - WREG32(mmUVD_VCPU_CNTL, 1 << 9); + WREG32(mmUVD_VCPU_CNTL, UVD_VCPU_CNTL__CLK_EN_MASK); /* enable UMC */ - WREG32_P(mmUVD_LMI_CTRL2, 0, ~(1 << 8)); + WREG32_P(mmUVD_LMI_CTRL2, 0, ~UVD_LMI_CTRL2__STALL_ARB_UMC_MASK); /* boot up the VCPU */ WREG32(mmUVD_SOFT_RESET, 0); @@ -484,10 +495,12 @@ static int uvd_v6_0_start(struct amdgpu_device *adev) return r; } /* enable master interrupt */ - WREG32_P(mmUVD_MASTINT_EN, 3 << 1, ~(3 << 1)); + WREG32_P(mmUVD_MASTINT_EN, + (UVD_MASTINT_EN__VCPU_EN_MASK|UVD_MASTINT_EN__SYS_EN_MASK), + ~(UVD_MASTINT_EN__VCPU_EN_MASK|UVD_MASTINT_EN__SYS_EN_MASK)); /* clear the bit 4 of UVD_STATUS */ - WREG32_P(mmUVD_STATUS, 0, ~(2 << 1)); + WREG32_P(mmUVD_STATUS, 0, ~(2 << UVD_STATUS__VCPU_REPORT__SHIFT)); rb_bufsz = order_base_2(ring->ring_size); tmp = 0; @@ -581,6 +594,32 @@ static void uvd_v6_0_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 seq } /** + * uvd_v6_0_ring_emit_hdp_flush - emit an hdp flush + * + * @ring: amdgpu_ring pointer + * + * Emits an hdp flush. + */ +static void uvd_v6_0_ring_emit_hdp_flush(struct amdgpu_ring *ring) +{ + amdgpu_ring_write(ring, PACKET0(mmHDP_MEM_COHERENCY_FLUSH_CNTL, 0)); + amdgpu_ring_write(ring, 0); +} + +/** + * uvd_v6_0_ring_hdp_invalidate - emit an hdp invalidate + * + * @ring: amdgpu_ring pointer + * + * Emits an hdp invalidate. + */ +static void uvd_v6_0_ring_emit_hdp_invalidate(struct amdgpu_ring *ring) +{ + amdgpu_ring_write(ring, PACKET0(mmHDP_DEBUG0, 0)); + amdgpu_ring_write(ring, 1); +} + +/** * uvd_v6_0_ring_test_ring - register write test * * @ring: amdgpu_ring pointer @@ -634,6 +673,9 @@ static void uvd_v6_0_ring_emit_ib(struct amdgpu_ring *ring, struct amdgpu_ib *ib, unsigned vm_id, bool ctx_switch) { + amdgpu_ring_write(ring, PACKET0(mmUVD_LMI_RBC_IB_VMID, 0)); + amdgpu_ring_write(ring, vm_id); + amdgpu_ring_write(ring, PACKET0(mmUVD_LMI_RBC_IB_64BIT_BAR_LOW, 0)); amdgpu_ring_write(ring, lower_32_bits(ib->gpu_addr)); amdgpu_ring_write(ring, PACKET0(mmUVD_LMI_RBC_IB_64BIT_BAR_HIGH, 0)); @@ -642,39 +684,55 @@ static void uvd_v6_0_ring_emit_ib(struct amdgpu_ring *ring, amdgpu_ring_write(ring, ib->length_dw); } -/** - * uvd_v6_0_ring_test_ib - test ib execution - * - * @ring: amdgpu_ring pointer - * - * Test if we can successfully execute an IB - */ -static int uvd_v6_0_ring_test_ib(struct amdgpu_ring *ring) +static void uvd_v6_0_ring_emit_vm_flush(struct amdgpu_ring *ring, + unsigned vm_id, uint64_t pd_addr) { - struct fence *fence = NULL; - int r; + uint32_t reg; - r = amdgpu_uvd_get_create_msg(ring, 1, NULL); - if (r) { - DRM_ERROR("amdgpu: failed to get create msg (%d).\n", r); - goto error; - } + if (vm_id < 8) + reg = mmVM_CONTEXT0_PAGE_TABLE_BASE_ADDR + vm_id; + else + reg = mmVM_CONTEXT8_PAGE_TABLE_BASE_ADDR + vm_id - 8; - r = amdgpu_uvd_get_destroy_msg(ring, 1, true, &fence); - if (r) { - DRM_ERROR("amdgpu: failed to get destroy ib (%d).\n", r); - goto error; - } + amdgpu_ring_write(ring, PACKET0(mmUVD_GPCOM_VCPU_DATA0, 0)); + amdgpu_ring_write(ring, reg << 2); + amdgpu_ring_write(ring, PACKET0(mmUVD_GPCOM_VCPU_DATA1, 0)); + amdgpu_ring_write(ring, pd_addr >> 12); + amdgpu_ring_write(ring, PACKET0(mmUVD_GPCOM_VCPU_CMD, 0)); + amdgpu_ring_write(ring, 0x8); - r = fence_wait(fence, false); - if (r) { - DRM_ERROR("amdgpu: fence wait failed (%d).\n", r); - goto error; - } - DRM_INFO("ib test on ring %d succeeded\n", ring->idx); -error: - fence_put(fence); - return r; + amdgpu_ring_write(ring, PACKET0(mmUVD_GPCOM_VCPU_DATA0, 0)); + amdgpu_ring_write(ring, mmVM_INVALIDATE_REQUEST << 2); + amdgpu_ring_write(ring, PACKET0(mmUVD_GPCOM_VCPU_DATA1, 0)); + amdgpu_ring_write(ring, 1 << vm_id); + amdgpu_ring_write(ring, PACKET0(mmUVD_GPCOM_VCPU_CMD, 0)); + amdgpu_ring_write(ring, 0x8); + + amdgpu_ring_write(ring, PACKET0(mmUVD_GPCOM_VCPU_DATA0, 0)); + amdgpu_ring_write(ring, mmVM_INVALIDATE_REQUEST << 2); + amdgpu_ring_write(ring, PACKET0(mmUVD_GPCOM_VCPU_DATA1, 0)); + amdgpu_ring_write(ring, 0); + amdgpu_ring_write(ring, PACKET0(mmUVD_GP_SCRATCH8, 0)); + amdgpu_ring_write(ring, 1 << vm_id); /* mask */ + amdgpu_ring_write(ring, PACKET0(mmUVD_GPCOM_VCPU_CMD, 0)); + amdgpu_ring_write(ring, 0xC); +} + +static void uvd_v6_0_ring_emit_pipeline_sync(struct amdgpu_ring *ring) +{ + uint32_t seq = ring->fence_drv.sync_seq; + uint64_t addr = ring->fence_drv.gpu_addr; + + amdgpu_ring_write(ring, PACKET0(mmUVD_GPCOM_VCPU_DATA0, 0)); + amdgpu_ring_write(ring, lower_32_bits(addr)); + amdgpu_ring_write(ring, PACKET0(mmUVD_GPCOM_VCPU_DATA1, 0)); + amdgpu_ring_write(ring, upper_32_bits(addr)); + amdgpu_ring_write(ring, PACKET0(mmUVD_GP_SCRATCH8, 0)); + amdgpu_ring_write(ring, 0xffffffff); /* mask */ + amdgpu_ring_write(ring, PACKET0(mmUVD_GP_SCRATCH9, 0)); + amdgpu_ring_write(ring, seq); + amdgpu_ring_write(ring, PACKET0(mmUVD_GPCOM_VCPU_CMD, 0)); + amdgpu_ring_write(ring, 0xE); } static bool uvd_v6_0_is_idle(void *handle) @@ -847,7 +905,8 @@ static int uvd_v6_0_set_clockgating_state(void *handle, bool enable = (state == AMD_CG_STATE_GATE) ? true : false; static int curstate = -1; - if (adev->asic_type == CHIP_FIJI) + if (adev->asic_type == CHIP_FIJI || + adev->asic_type == CHIP_POLARIS10) uvd_v6_set_bypass_mode(adev, enable); if (!(adev->cg_flags & AMD_CG_SUPPORT_UVD_MGCG)) @@ -912,22 +971,51 @@ const struct amd_ip_funcs uvd_v6_0_ip_funcs = { .set_powergating_state = uvd_v6_0_set_powergating_state, }; -static const struct amdgpu_ring_funcs uvd_v6_0_ring_funcs = { +static const struct amdgpu_ring_funcs uvd_v6_0_ring_phys_funcs = { .get_rptr = uvd_v6_0_ring_get_rptr, .get_wptr = uvd_v6_0_ring_get_wptr, .set_wptr = uvd_v6_0_ring_set_wptr, .parse_cs = amdgpu_uvd_ring_parse_cs, .emit_ib = uvd_v6_0_ring_emit_ib, .emit_fence = uvd_v6_0_ring_emit_fence, + .emit_hdp_flush = uvd_v6_0_ring_emit_hdp_flush, + .emit_hdp_invalidate = uvd_v6_0_ring_emit_hdp_invalidate, + .test_ring = uvd_v6_0_ring_test_ring, + .test_ib = amdgpu_uvd_ring_test_ib, + .insert_nop = amdgpu_ring_insert_nop, + .pad_ib = amdgpu_ring_generic_pad_ib, + .begin_use = amdgpu_uvd_ring_begin_use, + .end_use = amdgpu_uvd_ring_end_use, +}; + +static const struct amdgpu_ring_funcs uvd_v6_0_ring_vm_funcs = { + .get_rptr = uvd_v6_0_ring_get_rptr, + .get_wptr = uvd_v6_0_ring_get_wptr, + .set_wptr = uvd_v6_0_ring_set_wptr, + .parse_cs = NULL, + .emit_ib = uvd_v6_0_ring_emit_ib, + .emit_fence = uvd_v6_0_ring_emit_fence, + .emit_vm_flush = uvd_v6_0_ring_emit_vm_flush, + .emit_pipeline_sync = uvd_v6_0_ring_emit_pipeline_sync, + .emit_hdp_flush = uvd_v6_0_ring_emit_hdp_flush, + .emit_hdp_invalidate = uvd_v6_0_ring_emit_hdp_invalidate, .test_ring = uvd_v6_0_ring_test_ring, - .test_ib = uvd_v6_0_ring_test_ib, + .test_ib = amdgpu_uvd_ring_test_ib, .insert_nop = amdgpu_ring_insert_nop, .pad_ib = amdgpu_ring_generic_pad_ib, + .begin_use = amdgpu_uvd_ring_begin_use, + .end_use = amdgpu_uvd_ring_end_use, }; static void uvd_v6_0_set_ring_funcs(struct amdgpu_device *adev) { - adev->uvd.ring.funcs = &uvd_v6_0_ring_funcs; + if (adev->asic_type >= CHIP_POLARIS10) { + adev->uvd.ring.funcs = &uvd_v6_0_ring_vm_funcs; + DRM_INFO("UVD is enabled in VM mode\n"); + } else { + adev->uvd.ring.funcs = &uvd_v6_0_ring_phys_funcs; + DRM_INFO("UVD is enabled in physical mode\n"); + } } static const struct amdgpu_irq_src_funcs uvd_v6_0_irq_funcs = { diff --git a/drivers/gpu/drm/amd/amdgpu/vce_v2_0.c b/drivers/gpu/drm/amd/amdgpu/vce_v2_0.c index 45d92aceb485..80a37a602181 100644 --- a/drivers/gpu/drm/amd/amdgpu/vce_v2_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vce_v2_0.c @@ -594,6 +594,8 @@ static const struct amdgpu_ring_funcs vce_v2_0_ring_funcs = { .test_ib = amdgpu_vce_ring_test_ib, .insert_nop = amdgpu_ring_insert_nop, .pad_ib = amdgpu_ring_generic_pad_ib, + .begin_use = amdgpu_vce_ring_begin_use, + .end_use = amdgpu_vce_ring_end_use, }; static void vce_v2_0_set_ring_funcs(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c index 30e8099e94c5..c271abffd8dd 100644 --- a/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c @@ -43,6 +43,7 @@ #define mmVCE_LMI_VCPU_CACHE_40BIT_BAR0 0x8616 #define mmVCE_LMI_VCPU_CACHE_40BIT_BAR1 0x8617 #define mmVCE_LMI_VCPU_CACHE_40BIT_BAR2 0x8618 +#define VCE_STATUS_VCPU_REPORT_FW_LOADED_MASK 0x02 #define VCE_V3_0_FW_SIZE (384 * 1024) #define VCE_V3_0_STACK_SIZE (64 * 1024) @@ -51,6 +52,7 @@ static void vce_v3_0_mc_resume(struct amdgpu_device *adev, int idx); static void vce_v3_0_set_ring_funcs(struct amdgpu_device *adev); static void vce_v3_0_set_irq_funcs(struct amdgpu_device *adev); +static int vce_v3_0_wait_for_idle(void *handle); /** * vce_v3_0_ring_get_rptr - get read pointer @@ -205,6 +207,32 @@ static void vce_v3_0_set_vce_sw_clock_gating(struct amdgpu_device *adev, vce_v3_0_override_vce_clock_gating(adev, false); } +static int vce_v3_0_firmware_loaded(struct amdgpu_device *adev) +{ + int i, j; + + for (i = 0; i < 10; ++i) { + for (j = 0; j < 100; ++j) { + uint32_t status = RREG32(mmVCE_STATUS); + + if (status & VCE_STATUS_VCPU_REPORT_FW_LOADED_MASK) + return 0; + mdelay(10); + } + + DRM_ERROR("VCE not responding, trying to reset the ECPU!!!\n"); + WREG32_P(mmVCE_SOFT_RESET, + VCE_SOFT_RESET__ECPU_SOFT_RESET_MASK, + ~VCE_SOFT_RESET__ECPU_SOFT_RESET_MASK); + mdelay(10); + WREG32_P(mmVCE_SOFT_RESET, 0, + ~VCE_SOFT_RESET__ECPU_SOFT_RESET_MASK); + mdelay(10); + } + + return -ETIMEDOUT; +} + /** * vce_v3_0_start - start VCE block * @@ -215,11 +243,24 @@ static void vce_v3_0_set_vce_sw_clock_gating(struct amdgpu_device *adev, static int vce_v3_0_start(struct amdgpu_device *adev) { struct amdgpu_ring *ring; - int idx, i, j, r; + int idx, r; + + ring = &adev->vce.ring[0]; + WREG32(mmVCE_RB_RPTR, ring->wptr); + WREG32(mmVCE_RB_WPTR, ring->wptr); + WREG32(mmVCE_RB_BASE_LO, ring->gpu_addr); + WREG32(mmVCE_RB_BASE_HI, upper_32_bits(ring->gpu_addr)); + WREG32(mmVCE_RB_SIZE, ring->ring_size / 4); + + ring = &adev->vce.ring[1]; + WREG32(mmVCE_RB_RPTR2, ring->wptr); + WREG32(mmVCE_RB_WPTR2, ring->wptr); + WREG32(mmVCE_RB_BASE_LO2, ring->gpu_addr); + WREG32(mmVCE_RB_BASE_HI2, upper_32_bits(ring->gpu_addr)); + WREG32(mmVCE_RB_SIZE2, ring->ring_size / 4); mutex_lock(&adev->grbm_idx_mutex); for (idx = 0; idx < 2; ++idx) { - if (adev->vce.harvest_config & (1 << idx)) continue; @@ -233,48 +274,24 @@ static int vce_v3_0_start(struct amdgpu_device *adev) vce_v3_0_mc_resume(adev, idx); - /* set BUSY flag */ - WREG32_P(mmVCE_STATUS, 1, ~1); + WREG32_P(mmVCE_STATUS, VCE_STATUS__JOB_BUSY_MASK, + ~VCE_STATUS__JOB_BUSY_MASK); + if (adev->asic_type >= CHIP_STONEY) WREG32_P(mmVCE_VCPU_CNTL, 1, ~0x200001); else WREG32_P(mmVCE_VCPU_CNTL, VCE_VCPU_CNTL__CLK_EN_MASK, ~VCE_VCPU_CNTL__CLK_EN_MASK); - WREG32_P(mmVCE_SOFT_RESET, - VCE_SOFT_RESET__ECPU_SOFT_RESET_MASK, - ~VCE_SOFT_RESET__ECPU_SOFT_RESET_MASK); - - mdelay(100); - WREG32_P(mmVCE_SOFT_RESET, 0, ~VCE_SOFT_RESET__ECPU_SOFT_RESET_MASK); - for (i = 0; i < 10; ++i) { - uint32_t status; - for (j = 0; j < 100; ++j) { - status = RREG32(mmVCE_STATUS); - if (status & 2) - break; - mdelay(10); - } - r = 0; - if (status & 2) - break; - - DRM_ERROR("VCE not responding, trying to reset the ECPU!!!\n"); - WREG32_P(mmVCE_SOFT_RESET, - VCE_SOFT_RESET__ECPU_SOFT_RESET_MASK, - ~VCE_SOFT_RESET__ECPU_SOFT_RESET_MASK); - mdelay(10); - WREG32_P(mmVCE_SOFT_RESET, 0, - ~VCE_SOFT_RESET__ECPU_SOFT_RESET_MASK); - mdelay(10); - r = -1; - } + mdelay(100); + + r = vce_v3_0_firmware_loaded(adev); /* clear BUSY flag */ - WREG32_P(mmVCE_STATUS, 0, ~1); + WREG32_P(mmVCE_STATUS, 0, ~VCE_STATUS__JOB_BUSY_MASK); /* Set Clock-Gating off */ if (adev->cg_flags & AMD_CG_SUPPORT_VCE_MGCG) @@ -290,19 +307,46 @@ static int vce_v3_0_start(struct amdgpu_device *adev) WREG32_P(mmGRBM_GFX_INDEX, 0, ~GRBM_GFX_INDEX__VCE_INSTANCE_MASK); mutex_unlock(&adev->grbm_idx_mutex); - ring = &adev->vce.ring[0]; - WREG32(mmVCE_RB_RPTR, ring->wptr); - WREG32(mmVCE_RB_WPTR, ring->wptr); - WREG32(mmVCE_RB_BASE_LO, ring->gpu_addr); - WREG32(mmVCE_RB_BASE_HI, upper_32_bits(ring->gpu_addr)); - WREG32(mmVCE_RB_SIZE, ring->ring_size / 4); + return 0; +} - ring = &adev->vce.ring[1]; - WREG32(mmVCE_RB_RPTR2, ring->wptr); - WREG32(mmVCE_RB_WPTR2, ring->wptr); - WREG32(mmVCE_RB_BASE_LO2, ring->gpu_addr); - WREG32(mmVCE_RB_BASE_HI2, upper_32_bits(ring->gpu_addr)); - WREG32(mmVCE_RB_SIZE2, ring->ring_size / 4); +static int vce_v3_0_stop(struct amdgpu_device *adev) +{ + int idx; + + mutex_lock(&adev->grbm_idx_mutex); + for (idx = 0; idx < 2; ++idx) { + if (adev->vce.harvest_config & (1 << idx)) + continue; + + if (idx == 0) + WREG32_P(mmGRBM_GFX_INDEX, 0, + ~GRBM_GFX_INDEX__VCE_INSTANCE_MASK); + else + WREG32_P(mmGRBM_GFX_INDEX, + GRBM_GFX_INDEX__VCE_INSTANCE_MASK, + ~GRBM_GFX_INDEX__VCE_INSTANCE_MASK); + + if (adev->asic_type >= CHIP_STONEY) + WREG32_P(mmVCE_VCPU_CNTL, 0, ~0x200001); + else + WREG32_P(mmVCE_VCPU_CNTL, 0, + ~VCE_VCPU_CNTL__CLK_EN_MASK); + /* hold on ECPU */ + WREG32_P(mmVCE_SOFT_RESET, + VCE_SOFT_RESET__ECPU_SOFT_RESET_MASK, + ~VCE_SOFT_RESET__ECPU_SOFT_RESET_MASK); + + /* clear BUSY flag */ + WREG32_P(mmVCE_STATUS, 0, ~VCE_STATUS__JOB_BUSY_MASK); + + /* Set Clock-Gating off */ + if (adev->cg_flags & AMD_CG_SUPPORT_VCE_MGCG) + vce_v3_0_set_vce_sw_clock_gating(adev, false); + } + + WREG32_P(mmGRBM_GFX_INDEX, 0, ~GRBM_GFX_INDEX__VCE_INSTANCE_MASK); + mutex_unlock(&adev->grbm_idx_mutex); return 0; } @@ -441,7 +485,14 @@ static int vce_v3_0_hw_init(void *handle) static int vce_v3_0_hw_fini(void *handle) { - return 0; + int r; + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + + r = vce_v3_0_wait_for_idle(handle); + if (r) + return r; + + return vce_v3_0_stop(adev); } static int vce_v3_0_suspend(void *handle) @@ -604,6 +655,18 @@ static int vce_v3_0_process_interrupt(struct amdgpu_device *adev, return 0; } +static void vce_v3_set_bypass_mode(struct amdgpu_device *adev, bool enable) +{ + u32 tmp = RREG32_SMC(ixGCK_DFS_BYPASS_CNTL); + + if (enable) + tmp |= GCK_DFS_BYPASS_CNTL__BYPASSECLK_MASK; + else + tmp &= ~GCK_DFS_BYPASS_CNTL__BYPASSECLK_MASK; + + WREG32_SMC(ixGCK_DFS_BYPASS_CNTL, tmp); +} + static int vce_v3_0_set_clockgating_state(void *handle, enum amd_clockgating_state state) { @@ -611,6 +674,9 @@ static int vce_v3_0_set_clockgating_state(void *handle, bool enable = (state == AMD_CG_STATE_GATE) ? true : false; int i; + if (adev->asic_type == CHIP_POLARIS10) + vce_v3_set_bypass_mode(adev, enable); + if (!(adev->cg_flags & AMD_CG_SUPPORT_VCE_MGCG)) return 0; @@ -701,6 +767,8 @@ static const struct amdgpu_ring_funcs vce_v3_0_ring_funcs = { .test_ib = amdgpu_vce_ring_test_ib, .insert_nop = amdgpu_ring_insert_nop, .pad_ib = amdgpu_ring_generic_pad_ib, + .begin_use = amdgpu_vce_ring_begin_use, + .end_use = amdgpu_vce_ring_end_use, }; static void vce_v3_0_set_ring_funcs(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/amdgpu/vi.c b/drivers/gpu/drm/amd/amdgpu/vi.c index 2c88d0b66cf3..03a31c53aec3 100644 --- a/drivers/gpu/drm/amd/amdgpu/vi.c +++ b/drivers/gpu/drm/amd/amdgpu/vi.c @@ -203,6 +203,29 @@ static void vi_didt_wreg(struct amdgpu_device *adev, u32 reg, u32 v) spin_unlock_irqrestore(&adev->didt_idx_lock, flags); } +static u32 vi_gc_cac_rreg(struct amdgpu_device *adev, u32 reg) +{ + unsigned long flags; + u32 r; + + spin_lock_irqsave(&adev->gc_cac_idx_lock, flags); + WREG32(mmGC_CAC_IND_INDEX, (reg)); + r = RREG32(mmGC_CAC_IND_DATA); + spin_unlock_irqrestore(&adev->gc_cac_idx_lock, flags); + return r; +} + +static void vi_gc_cac_wreg(struct amdgpu_device *adev, u32 reg, u32 v) +{ + unsigned long flags; + + spin_lock_irqsave(&adev->gc_cac_idx_lock, flags); + WREG32(mmGC_CAC_IND_INDEX, (reg)); + WREG32(mmGC_CAC_IND_DATA, (v)); + spin_unlock_irqrestore(&adev->gc_cac_idx_lock, flags); +} + + static const u32 tonga_mgcg_cgcg_init[] = { mmCGTT_DRM_CLK_CTRL0, 0xffffffff, 0x00600100, @@ -421,6 +444,20 @@ static bool vi_read_bios_from_rom(struct amdgpu_device *adev, return true; } +static u32 vi_get_virtual_caps(struct amdgpu_device *adev) +{ + u32 caps = 0; + u32 reg = RREG32(mmBIF_IOV_FUNC_IDENTIFIER); + + if (REG_GET_FIELD(reg, BIF_IOV_FUNC_IDENTIFIER, IOV_ENABLE)) + caps |= AMDGPU_VIRT_CAPS_SRIOV_EN; + + if (REG_GET_FIELD(reg, BIF_IOV_FUNC_IDENTIFIER, FUNC_IDENTIFIER)) + caps |= AMDGPU_VIRT_CAPS_IS_VF; + + return caps; +} + static const struct amdgpu_allowed_register_entry tonga_allowed_read_registers[] = { {mmGB_MACROTILE_MODE7, true}, }; @@ -519,12 +556,12 @@ static uint32_t vi_read_indexed_register(struct amdgpu_device *adev, u32 se_num, mutex_lock(&adev->grbm_idx_mutex); if (se_num != 0xffffffff || sh_num != 0xffffffff) - gfx_v8_0_select_se_sh(adev, se_num, sh_num); + amdgpu_gfx_select_se_sh(adev, se_num, sh_num, 0xffffffff); val = RREG32(reg_offset); if (se_num != 0xffffffff || sh_num != 0xffffffff) - gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff); + amdgpu_gfx_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff); mutex_unlock(&adev->grbm_idx_mutex); return val; } @@ -583,7 +620,7 @@ static int vi_read_register(struct amdgpu_device *adev, u32 se_num, return -EINVAL; } -static void vi_gpu_pci_config_reset(struct amdgpu_device *adev) +static int vi_gpu_pci_config_reset(struct amdgpu_device *adev) { u32 i; @@ -598,11 +635,14 @@ static void vi_gpu_pci_config_reset(struct amdgpu_device *adev) /* wait for asic to come out of reset */ for (i = 0; i < adev->usec_timeout; i++) { - if (RREG32(mmCONFIG_MEMSIZE) != 0xffffffff) - break; + if (RREG32(mmCONFIG_MEMSIZE) != 0xffffffff) { + /* enable BM */ + pci_set_master(adev->pdev); + return 0; + } udelay(1); } - + return -EINVAL; } static void vi_set_bios_scratch_engine_hung(struct amdgpu_device *adev, bool hung) @@ -628,13 +668,15 @@ static void vi_set_bios_scratch_engine_hung(struct amdgpu_device *adev, bool hun */ static int vi_asic_reset(struct amdgpu_device *adev) { + int r; + vi_set_bios_scratch_engine_hung(adev, true); - vi_gpu_pci_config_reset(adev); + r = vi_gpu_pci_config_reset(adev); vi_set_bios_scratch_engine_hung(adev, false); - return 0; + return r; } static int vi_set_uvd_clock(struct amdgpu_device *adev, u32 clock, @@ -1118,9 +1160,7 @@ static const struct amdgpu_asic_funcs vi_asic_funcs = .get_xclk = &vi_get_xclk, .set_uvd_clocks = &vi_set_uvd_clocks, .set_vce_clocks = &vi_set_vce_clocks, - /* these should be moved to their own ip modules */ - .get_gpu_clock_counter = &gfx_v8_0_get_gpu_clock_counter, - .wait_for_mc_idle = &gmc_v8_0_mc_wait_for_idle, + .get_virtual_caps = &vi_get_virtual_caps, }; static int vi_common_early_init(void *handle) @@ -1141,6 +1181,8 @@ static int vi_common_early_init(void *handle) adev->uvd_ctx_wreg = &vi_uvd_ctx_wreg; adev->didt_rreg = &vi_didt_rreg; adev->didt_wreg = &vi_didt_wreg; + adev->gc_cac_rreg = &vi_gc_cac_rreg; + adev->gc_cac_wreg = &vi_gc_cac_wreg; adev->asic_funcs = &vi_asic_funcs; @@ -1214,12 +1256,18 @@ static int vi_common_early_init(void *handle) adev->cg_flags = AMD_CG_SUPPORT_UVD_MGCG | AMD_CG_SUPPORT_GFX_MGCG | AMD_CG_SUPPORT_GFX_MGLS | + AMD_CG_SUPPORT_GFX_RLC_LS | + AMD_CG_SUPPORT_GFX_CP_LS | + AMD_CG_SUPPORT_GFX_CGTS | + AMD_CG_SUPPORT_GFX_MGLS | + AMD_CG_SUPPORT_GFX_CGTS_LS | + AMD_CG_SUPPORT_GFX_CGCG | + AMD_CG_SUPPORT_GFX_CGLS | AMD_CG_SUPPORT_BIF_LS | AMD_CG_SUPPORT_HDP_MGCG | AMD_CG_SUPPORT_HDP_LS | AMD_CG_SUPPORT_SDMA_MGCG | AMD_CG_SUPPORT_SDMA_LS; - adev->pg_flags = 0; adev->external_rev_id = adev->rev_id + 0x1; break; default: diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_process.c b/drivers/gpu/drm/amd/amdkfd/kfd_process.c index ac005796b71c..4f3849ac8c07 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_process.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_process.c @@ -63,13 +63,12 @@ static struct kfd_process *create_process(const struct task_struct *thread); void kfd_process_create_wq(void) { if (!kfd_process_wq) - kfd_process_wq = create_workqueue("kfd_process_wq"); + kfd_process_wq = alloc_workqueue("kfd_process_wq", 0, 0); } void kfd_process_destroy_wq(void) { if (kfd_process_wq) { - flush_workqueue(kfd_process_wq); destroy_workqueue(kfd_process_wq); kfd_process_wq = NULL; } @@ -242,13 +241,19 @@ static void kfd_process_notifier_release(struct mmu_notifier *mn, pqm_uninit(&p->pqm); /* Iterate over all process device data structure and check - * if we should reset all wavefronts */ - list_for_each_entry(pdd, &p->per_device_data, per_device_list) + * if we should delete debug managers and reset all wavefronts + */ + list_for_each_entry(pdd, &p->per_device_data, per_device_list) { + if ((pdd->dev->dbgmgr) && + (pdd->dev->dbgmgr->pasid == p->pasid)) + kfd_dbgmgr_destroy(pdd->dev->dbgmgr); + if (pdd->reset_wavefronts) { pr_warn("amdkfd: Resetting all wave fronts\n"); dbgdev_wave_reset_wavefronts(pdd->dev, p); pdd->reset_wavefronts = false; } + } mutex_unlock(&p->mutex); @@ -324,6 +329,7 @@ err_process_pqm_init: synchronize_rcu(); mmu_notifier_unregister_no_release(&process->mmu_notifier, process->mm); err_mmu_notifier: + mutex_destroy(&process->mutex); kfd_pasid_free(process->pasid); err_alloc_pasid: kfree(process->queues); @@ -404,42 +410,52 @@ void kfd_unbind_process_from_device(struct kfd_dev *dev, unsigned int pasid) idx = srcu_read_lock(&kfd_processes_srcu); + /* + * Look for the process that matches the pasid. If there is no such + * process, we either released it in amdkfd's own notifier, or there + * is a bug. Unfortunately, there is no way to tell... + */ hash_for_each_rcu(kfd_processes_table, i, p, kfd_processes) - if (p->pasid == pasid) - break; + if (p->pasid == pasid) { - srcu_read_unlock(&kfd_processes_srcu, idx); + srcu_read_unlock(&kfd_processes_srcu, idx); - BUG_ON(p->pasid != pasid); + pr_debug("Unbinding process %d from IOMMU\n", pasid); - mutex_lock(&p->mutex); + mutex_lock(&p->mutex); - if ((dev->dbgmgr) && (dev->dbgmgr->pasid == p->pasid)) - kfd_dbgmgr_destroy(dev->dbgmgr); + if ((dev->dbgmgr) && (dev->dbgmgr->pasid == p->pasid)) + kfd_dbgmgr_destroy(dev->dbgmgr); - pqm_uninit(&p->pqm); + pqm_uninit(&p->pqm); - pdd = kfd_get_process_device_data(dev, p); + pdd = kfd_get_process_device_data(dev, p); - if (!pdd) { - mutex_unlock(&p->mutex); - return; - } + if (!pdd) { + mutex_unlock(&p->mutex); + return; + } - if (pdd->reset_wavefronts) { - dbgdev_wave_reset_wavefronts(pdd->dev, p); - pdd->reset_wavefronts = false; - } + if (pdd->reset_wavefronts) { + dbgdev_wave_reset_wavefronts(pdd->dev, p); + pdd->reset_wavefronts = false; + } - /* - * Just mark pdd as unbound, because we still need it to call - * amd_iommu_unbind_pasid() in when the process exits. - * We don't call amd_iommu_unbind_pasid() here - * because the IOMMU called us. - */ - pdd->bound = false; + /* + * Just mark pdd as unbound, because we still need it + * to call amd_iommu_unbind_pasid() in when the + * process exits. + * We don't call amd_iommu_unbind_pasid() here + * because the IOMMU called us. + */ + pdd->bound = false; - mutex_unlock(&p->mutex); + mutex_unlock(&p->mutex); + + return; + } + + srcu_read_unlock(&kfd_processes_srcu, idx); } struct kfd_process_device *kfd_get_first_process_device_data(struct kfd_process *p) diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c index 74909e72a009..884c96f50c3d 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c @@ -666,7 +666,7 @@ static ssize_t node_show(struct kobject *kobj, struct attribute *attr, dev->node_props.simd_count); if (dev->mem_bank_count < dev->node_props.mem_banks_count) { - pr_warn("kfd: mem_banks_count truncated from %d to %d\n", + pr_info_once("kfd: mem_banks_count truncated from %d to %d\n", dev->node_props.mem_banks_count, dev->mem_bank_count); sysfs_show_32bit_prop(buffer, "mem_banks_count", diff --git a/drivers/gpu/drm/amd/include/amd_pcie.h b/drivers/gpu/drm/amd/include/amd_pcie.h index 7c2a916c1e63..5eb895fd98bf 100644 --- a/drivers/gpu/drm/amd/include/amd_pcie.h +++ b/drivers/gpu/drm/amd/include/amd_pcie.h @@ -37,6 +37,13 @@ #define CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_MASK 0x0000FFFF #define CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_SHIFT 0 +/* gen: chipset 1/2, asic 1/2/3 */ +#define AMDGPU_DEFAULT_PCIE_GEN_MASK (CAIL_PCIE_LINK_SPEED_SUPPORT_GEN1 \ + | CAIL_PCIE_LINK_SPEED_SUPPORT_GEN2 \ + | CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN1 \ + | CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN2 \ + | CAIL_ASIC_PCIE_LINK_SPEED_SUPPORT_GEN3) + /* Following flags shows PCIe lane width switch supported in driver which are decided by chipset and ASIC */ #define CAIL_PCIE_LINK_WIDTH_SUPPORT_X1 0x00010000 #define CAIL_PCIE_LINK_WIDTH_SUPPORT_X2 0x00020000 @@ -47,4 +54,11 @@ #define CAIL_PCIE_LINK_WIDTH_SUPPORT_X32 0x00400000 #define CAIL_PCIE_LINK_WIDTH_SUPPORT_SHIFT 16 +/* 1/2/4/8/16 lanes */ +#define AMDGPU_DEFAULT_PCIE_MLW_MASK (CAIL_PCIE_LINK_WIDTH_SUPPORT_X1 \ + | CAIL_PCIE_LINK_WIDTH_SUPPORT_X2 \ + | CAIL_PCIE_LINK_WIDTH_SUPPORT_X4 \ + | CAIL_PCIE_LINK_WIDTH_SUPPORT_X8 \ + | CAIL_PCIE_LINK_WIDTH_SUPPORT_X16) + #endif diff --git a/drivers/gpu/drm/amd/include/amd_shared.h b/drivers/gpu/drm/amd/include/amd_shared.h index 6080951d539d..a74a0d2ff1ca 100644 --- a/drivers/gpu/drm/amd/include/amd_shared.h +++ b/drivers/gpu/drm/amd/include/amd_shared.h @@ -26,15 +26,6 @@ #define AMD_MAX_USEC_TIMEOUT 100000 /* 100 ms */ /* -* Supported GPU families (aligned with amdgpu_drm.h) -*/ -#define AMD_FAMILY_UNKNOWN 0 -#define AMD_FAMILY_CI 120 /* Bonaire, Hawaii */ -#define AMD_FAMILY_KV 125 /* Kaveri, Kabini, Mullins */ -#define AMD_FAMILY_VI 130 /* Iceland, Tonga */ -#define AMD_FAMILY_CZ 135 /* Carrizo */ - -/* * Supported ASIC types */ enum amd_asic_type { @@ -120,6 +111,8 @@ enum amd_powergating_state { #define AMD_PG_SUPPORT_SDMA (1 << 8) #define AMD_PG_SUPPORT_ACP (1 << 9) #define AMD_PG_SUPPORT_SAMU (1 << 10) +#define AMD_PG_SUPPORT_GFX_QUICK_MG (1 << 11) +#define AMD_PG_SUPPORT_GFX_PIPELINE (1 << 12) enum amd_pm_state_type { /* not used for dpm */ @@ -157,6 +150,7 @@ struct amd_ip_funcs { int (*hw_init)(void *handle); /* tears down the hw state */ int (*hw_fini)(void *handle); + void (*late_fini)(void *handle); /* handles IP specific hw/sw changes for suspend */ int (*suspend)(void *handle); /* handles IP specific hw/sw changes for resume */ diff --git a/drivers/gpu/drm/amd/include/asic_reg/bif/bif_5_0_d.h b/drivers/gpu/drm/amd/include/asic_reg/bif/bif_5_0_d.h index 293329719bba..809759f7bb81 100644 --- a/drivers/gpu/drm/amd/include/asic_reg/bif/bif_5_0_d.h +++ b/drivers/gpu/drm/amd/include/asic_reg/bif/bif_5_0_d.h @@ -27,6 +27,7 @@ #define mmMM_INDEX 0x0 #define mmMM_INDEX_HI 0x6 #define mmMM_DATA 0x1 +#define mmCC_BIF_BX_STRAP2 0x152A #define mmBIF_MM_INDACCESS_CNTL 0x1500 #define mmBIF_DOORBELL_APER_EN 0x1501 #define mmBUS_CNTL 0x1508 diff --git a/drivers/gpu/drm/amd/include/asic_reg/gca/gfx_8_0_d.h b/drivers/gpu/drm/amd/include/asic_reg/gca/gfx_8_0_d.h index ebaf67bb1589..90ff7c8a6011 100644 --- a/drivers/gpu/drm/amd/include/asic_reg/gca/gfx_8_0_d.h +++ b/drivers/gpu/drm/amd/include/asic_reg/gca/gfx_8_0_d.h @@ -2823,4 +2823,7 @@ #define mmDC_EDC_CSINVOC_CNT 0x3192 #define mmDC_EDC_RESTORE_CNT 0x3193 +#define mmGC_CAC_IND_INDEX 0x129a +#define mmGC_CAC_IND_DATA 0x129b + #endif /* GFX_8_0_D_H */ diff --git a/drivers/gpu/drm/amd/include/asic_reg/gca/gfx_8_0_sh_mask.h b/drivers/gpu/drm/amd/include/asic_reg/gca/gfx_8_0_sh_mask.h index 7d722458d9f5..4070ca3a68eb 100644 --- a/drivers/gpu/drm/amd/include/asic_reg/gca/gfx_8_0_sh_mask.h +++ b/drivers/gpu/drm/amd/include/asic_reg/gca/gfx_8_0_sh_mask.h @@ -8730,8 +8730,6 @@ #define RLC_GPM_STAT__DYN_CU_POWERING_DOWN__SHIFT 0x10 #define RLC_GPM_STAT__ABORTED_PD_SEQUENCE_MASK 0x20000 #define RLC_GPM_STAT__ABORTED_PD_SEQUENCE__SHIFT 0x11 -#define RLC_GPM_STAT__RESERVED_MASK 0xfc0000 -#define RLC_GPM_STAT__RESERVED__SHIFT 0x12 #define RLC_GPM_STAT__PG_ERROR_STATUS_MASK 0xff000000 #define RLC_GPM_STAT__PG_ERROR_STATUS__SHIFT 0x18 #define RLC_GPU_CLOCK_32_RES_SEL__RES_SEL_MASK 0x3f @@ -8764,8 +8762,10 @@ #define RLC_PG_CNTL__SMU_CLK_SLOWDOWN_ON_PD_ENABLE__SHIFT 0x12 #define RLC_PG_CNTL__SMU_HANDSHAKE_ENABLE_MASK 0x80000 #define RLC_PG_CNTL__SMU_HANDSHAKE_ENABLE__SHIFT 0x13 -#define RLC_PG_CNTL__RESERVED1_MASK 0xf00000 -#define RLC_PG_CNTL__RESERVED1__SHIFT 0x14 +#define RLC_PG_CNTL__QUICK_PG_ENABLE_MASK 0x100000 +#define RLC_PG_CNTL__QUICK_PG_ENABLE__SHIFT 0x14 +#define RLC_PG_CNTL__RESERVED1_MASK 0xe00000 +#define RLC_PG_CNTL__RESERVED1__SHIFT 0x15 #define RLC_GPM_THREAD_PRIORITY__THREAD0_PRIORITY_MASK 0xff #define RLC_GPM_THREAD_PRIORITY__THREAD0_PRIORITY__SHIFT 0x0 #define RLC_GPM_THREAD_PRIORITY__THREAD1_PRIORITY_MASK 0xff00 @@ -9102,8 +9102,6 @@ #define RLC_GPM_LOG_CONT__CONT__SHIFT 0x0 #define RLC_PG_DELAY_3__CGCG_ACTIVE_BEFORE_CGPG_MASK 0xff #define RLC_PG_DELAY_3__CGCG_ACTIVE_BEFORE_CGPG__SHIFT 0x0 -#define RLC_PG_DELAY_3__RESERVED_MASK 0xffffff00 -#define RLC_PG_DELAY_3__RESERVED__SHIFT 0x8 #define RLC_GPM_INT_DISABLE_TH0__DISABLE_MASK 0xffffffff #define RLC_GPM_INT_DISABLE_TH0__DISABLE__SHIFT 0x0 #define RLC_GPM_INT_DISABLE_TH1__DISABLE_MASK 0xffffffff @@ -9124,14 +9122,8 @@ #define RLC_SRM_DEBUG_SELECT__RESERVED__SHIFT 0x8 #define RLC_SRM_DEBUG__DATA_MASK 0xffffffff #define RLC_SRM_DEBUG__DATA__SHIFT 0x0 -#define RLC_SRM_ARAM_ADDR__ADDR_MASK 0x3ff -#define RLC_SRM_ARAM_ADDR__ADDR__SHIFT 0x0 -#define RLC_SRM_ARAM_ADDR__RESERVED_MASK 0xfffffc00 -#define RLC_SRM_ARAM_ADDR__RESERVED__SHIFT 0xa #define RLC_SRM_ARAM_DATA__DATA_MASK 0xffffffff #define RLC_SRM_ARAM_DATA__DATA__SHIFT 0x0 -#define RLC_SRM_DRAM_ADDR__ADDR_MASK 0x3ff -#define RLC_SRM_DRAM_ADDR__ADDR__SHIFT 0x0 #define RLC_SRM_DRAM_ADDR__RESERVED_MASK 0xfffffc00 #define RLC_SRM_DRAM_ADDR__RESERVED__SHIFT 0xa #define RLC_SRM_DRAM_DATA__DATA_MASK 0xffffffff @@ -17946,8 +17938,6 @@ #define VGT_TESS_DISTRIBUTION__ACCUM_TRI__SHIFT 0x8 #define VGT_TESS_DISTRIBUTION__ACCUM_QUAD_MASK 0xff0000 #define VGT_TESS_DISTRIBUTION__ACCUM_QUAD__SHIFT 0x10 -#define VGT_TESS_DISTRIBUTION__DONUT_SPLIT_MASK 0xff000000 -#define VGT_TESS_DISTRIBUTION__DONUT_SPLIT__SHIFT 0x18 #define VGT_TF_RING_SIZE__SIZE_MASK 0xffff #define VGT_TF_RING_SIZE__SIZE__SHIFT 0x0 #define VGT_SYS_CONFIG__DUAL_CORE_EN_MASK 0x1 @@ -20502,8 +20492,6 @@ #define DIDT_SQ_CTRL0__DIDT_CTRL_RST__SHIFT 0x4 #define DIDT_SQ_CTRL0__DIDT_CLK_EN_OVERRIDE_MASK 0x20 #define DIDT_SQ_CTRL0__DIDT_CLK_EN_OVERRIDE__SHIFT 0x5 -#define DIDT_SQ_CTRL0__UNUSED_0_MASK 0xffffffc0 -#define DIDT_SQ_CTRL0__UNUSED_0__SHIFT 0x6 #define DIDT_SQ_CTRL1__MIN_POWER_MASK 0xffff #define DIDT_SQ_CTRL1__MIN_POWER__SHIFT 0x0 #define DIDT_SQ_CTRL1__MAX_POWER_MASK 0xffff0000 @@ -20558,8 +20546,6 @@ #define DIDT_DB_CTRL0__DIDT_CTRL_RST__SHIFT 0x4 #define DIDT_DB_CTRL0__DIDT_CLK_EN_OVERRIDE_MASK 0x20 #define DIDT_DB_CTRL0__DIDT_CLK_EN_OVERRIDE__SHIFT 0x5 -#define DIDT_DB_CTRL0__UNUSED_0_MASK 0xffffffc0 -#define DIDT_DB_CTRL0__UNUSED_0__SHIFT 0x6 #define DIDT_DB_CTRL1__MIN_POWER_MASK 0xffff #define DIDT_DB_CTRL1__MIN_POWER__SHIFT 0x0 #define DIDT_DB_CTRL1__MAX_POWER_MASK 0xffff0000 @@ -20614,8 +20600,6 @@ #define DIDT_TD_CTRL0__DIDT_CTRL_RST__SHIFT 0x4 #define DIDT_TD_CTRL0__DIDT_CLK_EN_OVERRIDE_MASK 0x20 #define DIDT_TD_CTRL0__DIDT_CLK_EN_OVERRIDE__SHIFT 0x5 -#define DIDT_TD_CTRL0__UNUSED_0_MASK 0xffffffc0 -#define DIDT_TD_CTRL0__UNUSED_0__SHIFT 0x6 #define DIDT_TD_CTRL1__MIN_POWER_MASK 0xffff #define DIDT_TD_CTRL1__MIN_POWER__SHIFT 0x0 #define DIDT_TD_CTRL1__MAX_POWER_MASK 0xffff0000 @@ -20670,8 +20654,6 @@ #define DIDT_TCP_CTRL0__DIDT_CTRL_RST__SHIFT 0x4 #define DIDT_TCP_CTRL0__DIDT_CLK_EN_OVERRIDE_MASK 0x20 #define DIDT_TCP_CTRL0__DIDT_CLK_EN_OVERRIDE__SHIFT 0x5 -#define DIDT_TCP_CTRL0__UNUSED_0_MASK 0xffffffc0 -#define DIDT_TCP_CTRL0__UNUSED_0__SHIFT 0x6 #define DIDT_TCP_CTRL1__MIN_POWER_MASK 0xffff #define DIDT_TCP_CTRL1__MIN_POWER__SHIFT 0x0 #define DIDT_TCP_CTRL1__MAX_POWER_MASK 0xffff0000 @@ -20726,8 +20708,6 @@ #define DIDT_DBR_CTRL0__DIDT_CTRL_RST__SHIFT 0x4 #define DIDT_DBR_CTRL0__DIDT_CLK_EN_OVERRIDE_MASK 0x20 #define DIDT_DBR_CTRL0__DIDT_CLK_EN_OVERRIDE__SHIFT 0x5 -#define DIDT_DBR_CTRL0__UNUSED_0_MASK 0xffffffc0 -#define DIDT_DBR_CTRL0__UNUSED_0__SHIFT 0x6 #define DIDT_DBR_CTRL1__MIN_POWER_MASK 0xffff #define DIDT_DBR_CTRL1__MIN_POWER__SHIFT 0x0 #define DIDT_DBR_CTRL1__MAX_POWER_MASK 0xffff0000 @@ -20773,4 +20753,84 @@ #define DIDT_DBR_WEIGHT8_11__WEIGHT11_MASK 0xff000000 #define DIDT_DBR_WEIGHT8_11__WEIGHT11__SHIFT 0x18 +#define DIDT_SQ_STALL_CTRL__DIDT_STALL_CTRL_ENABLE_MASK 0x00000001 +#define DIDT_SQ_STALL_CTRL__DIDT_STALL_CTRL_ENABLE__SHIFT 0x00000000 + +#define DIDT_SQ_STALL_CTRL__DIDT_STALL_DELAY_HI_MASK 0x0000007e +#define DIDT_SQ_STALL_CTRL__DIDT_STALL_DELAY_LO_MASK 0x00001f80L +#define DIDT_SQ_STALL_CTRL__DIDT_STALL_DELAY_HI__SHIFT 0x00000001 +#define DIDT_SQ_STALL_CTRL__DIDT_STALL_DELAY_LO__SHIFT 0x00000007 + +#define DIDT_SQ_STALL_CTRL__DIDT_HI_POWER_THRESHOLD_MASK 0x1fffe000L +#define DIDT_SQ_STALL_CTRL__DIDT_HI_POWER_THRESHOLD__SHIFT 0x0000000d + +#define DIDT_SQ_STALL_CTRL__UNUSED_0_MASK 0xe0000000L +#define DIDT_SQ_STALL_CTRL__UNUSED_0__SHIFT 0x0000001d + +#define DIDT_SQ_TUNING_CTRL__DIDT_TUNING_ENABLE_MASK 0x00000001L +#define DIDT_SQ_TUNING_CTRL__DIDT_TUNING_ENABLE__SHIFT 0x00000000 + +#define DIDT_SQ_TUNING_CTRL__MAX_POWER_DELTA_HI_MASK 0x00007ffeL +#define DIDT_SQ_TUNING_CTRL__MAX_POWER_DELTA_HI__SHIFT 0x00000001 +#define DIDT_SQ_TUNING_CTRL__MAX_POWER_DELTA_LO_MASK 0x1fff8000L +#define DIDT_SQ_TUNING_CTRL__MAX_POWER_DELTA_LO__SHIFT 0x0000000f + +#define DIDT_TD_STALL_CTRL__DIDT_STALL_CTRL_ENABLE_MASK 0x00000001L +#define DIDT_TD_STALL_CTRL__DIDT_STALL_CTRL_ENABLE__SHIFT 0x00000000 + +#define DIDT_TD_STALL_CTRL__DIDT_STALL_DELAY_HI_MASK 0x0000007eL +#define DIDT_TD_STALL_CTRL__DIDT_STALL_DELAY_LO_MASK 0x00001f80L +#define DIDT_TD_STALL_CTRL__DIDT_STALL_DELAY_HI__SHIFT 0x00000001 +#define DIDT_TD_STALL_CTRL__DIDT_STALL_DELAY_LO__SHIFT 0x00000007 + +#define DIDT_TD_STALL_CTRL__DIDT_HI_POWER_THRESHOLD_MASK 0x1fffe000L +#define DIDT_TD_STALL_CTRL__DIDT_HI_POWER_THRESHOLD__SHIFT 0x0000000d + +#define DIDT_SQ_CTRL0__DIDT_MAX_STALLS_ALLOWED_HI_MASK 0x00000fc0L +#define DIDT_SQ_CTRL0__DIDT_MAX_STALLS_ALLOWED_LO_MASK 0x0003f000L +#define DIDT_SQ_CTRL0__DIDT_MAX_STALLS_ALLOWED_HI__SHIFT 0x00000006 +#define DIDT_SQ_CTRL0__DIDT_MAX_STALLS_ALLOWED_LO__SHIFT 0x0000000c + +#define DIDT_TD_TUNING_CTRL__DIDT_TUNING_ENABLE_MASK 0x00000001L +#define DIDT_TD_TUNING_CTRL__MAX_POWER_DELTA_HI_MASK 0x00007ffeL +#define DIDT_TD_TUNING_CTRL__MAX_POWER_DELTA_LO_MASK 0x1fff8000L + +#define DIDT_TD_TUNING_CTRL__DIDT_TUNING_ENABLE__SHIFT 0x00000000 +#define DIDT_TD_TUNING_CTRL__MAX_POWER_DELTA_HI__SHIFT 0x00000001 +#define DIDT_TD_TUNING_CTRL__MAX_POWER_DELTA_LO__SHIFT 0x0000000f + +#define DIDT_TD_STALL_CTRL__UNUSED_0_MASK 0xe0000000L +#define DIDT_TD_STALL_CTRL__UNUSED_0__SHIFT 0x0000001d + +#define DIDT_TD_CTRL0__DIDT_MAX_STALLS_ALLOWED_HI_MASK 0x00000fc0L +#define DIDT_TD_CTRL0__DIDT_MAX_STALLS_ALLOWED_LO_MASK 0x0003f000L +#define DIDT_TD_CTRL0__DIDT_MAX_STALLS_ALLOWED_HI__SHIFT 0x00000006 +#define DIDT_TD_CTRL0__DIDT_MAX_STALLS_ALLOWED_LO__SHIFT 0x0000000c + +#define DIDT_TCP_STALL_CTRL__DIDT_STALL_CTRL_ENABLE_MASK 0x00000001L +#define DIDT_TCP_STALL_CTRL__DIDT_STALL_CTRL_ENABLE__SHIFT 0x00000000 + +#define DIDT_TCP_STALL_CTRL__DIDT_STALL_DELAY_HI_MASK 0x0000007eL +#define DIDT_TCP_STALL_CTRL__DIDT_STALL_DELAY_LO_MASK 0x00001f80L +#define DIDT_TCP_STALL_CTRL__DIDT_STALL_DELAY_HI__SHIFT 0x00000001 +#define DIDT_TCP_STALL_CTRL__DIDT_STALL_DELAY_LO__SHIFT 0x00000007 + +#define DIDT_TCP_STALL_CTRL__DIDT_HI_POWER_THRESHOLD_MASK 0x1fffe000L +#define DIDT_TCP_STALL_CTRL__DIDT_HI_POWER_THRESHOLD__SHIFT 0x0000000d + +#define DIDT_TCP_STALL_CTRL__UNUSED_0_MASK 0xe0000000L +#define DIDT_TCP_STALL_CTRL__UNUSED_0__SHIFT 0x0000001d + +#define DIDT_TCP_TUNING_CTRL__DIDT_TUNING_ENABLE_MASK 0x00000001L +#define DIDT_TCP_TUNING_CTRL__MAX_POWER_DELTA_HI_MASK 0x00007ffeL +#define DIDT_TCP_TUNING_CTRL__MAX_POWER_DELTA_LO_MASK 0x1fff8000L +#define DIDT_TCP_TUNING_CTRL__DIDT_TUNING_ENABLE__SHIFT 0x00000000 +#define DIDT_TCP_TUNING_CTRL__MAX_POWER_DELTA_HI__SHIFT 0x00000001 +#define DIDT_TCP_TUNING_CTRL__MAX_POWER_DELTA_LO__SHIFT 0x0000000f + +#define DIDT_TCP_CTRL0__DIDT_MAX_STALLS_ALLOWED_HI_MASK 0x00000fc0L +#define DIDT_TCP_CTRL0__DIDT_MAX_STALLS_ALLOWED_LO_MASK 0x0003f000L +#define DIDT_TCP_CTRL0__DIDT_MAX_STALLS_ALLOWED_HI__SHIFT 0x00000006 +#define DIDT_TCP_CTRL0__DIDT_MAX_STALLS_ALLOWED_LO__SHIFT 0x0000000c + #endif /* GFX_8_0_SH_MASK_H */ diff --git a/drivers/gpu/drm/amd/include/asic_reg/uvd/uvd_6_0_d.h b/drivers/gpu/drm/amd/include/asic_reg/uvd/uvd_6_0_d.h index 6f6fb34742d2..ec69869c55ff 100644 --- a/drivers/gpu/drm/amd/include/asic_reg/uvd/uvd_6_0_d.h +++ b/drivers/gpu/drm/amd/include/asic_reg/uvd/uvd_6_0_d.h @@ -111,6 +111,8 @@ #define mmUVD_MIF_RECON1_ADDR_CONFIG 0x39c5 #define ixUVD_MIF_SCLR_ADDR_CONFIG 0x4 #define mmUVD_JPEG_ADDR_CONFIG 0x3a1f +#define mmUVD_GP_SCRATCH8 0x3c0a +#define mmUVD_GP_SCRATCH9 0x3c0b #define mmUVD_GP_SCRATCH4 0x3d38 #endif /* UVD_6_0_D_H */ diff --git a/drivers/gpu/drm/amd/include/atombios.h b/drivers/gpu/drm/amd/include/atombios.h index 32f3e345de08..3493da5c8f0e 100644 --- a/drivers/gpu/drm/amd/include/atombios.h +++ b/drivers/gpu/drm/amd/include/atombios.h @@ -5538,6 +5538,78 @@ typedef struct _ATOM_ASIC_PROFILING_INFO_V3_5 ULONG ulReserved[12]; }ATOM_ASIC_PROFILING_INFO_V3_5; +/* for Polars10/11 AVFS parameters */ +typedef struct _ATOM_ASIC_PROFILING_INFO_V3_6 +{ + ATOM_COMMON_TABLE_HEADER asHeader; + ULONG ulMaxVddc; + ULONG ulMinVddc; + USHORT usLkgEuseIndex; + UCHAR ucLkgEfuseBitLSB; + UCHAR ucLkgEfuseLength; + ULONG ulLkgEncodeLn_MaxDivMin; + ULONG ulLkgEncodeMax; + ULONG ulLkgEncodeMin; + EFUSE_LINEAR_FUNC_PARAM sRoFuse; + ULONG ulEvvDefaultVddc; + ULONG ulEvvNoCalcVddc; + ULONG ulSpeed_Model; + ULONG ulSM_A0; + ULONG ulSM_A1; + ULONG ulSM_A2; + ULONG ulSM_A3; + ULONG ulSM_A4; + ULONG ulSM_A5; + ULONG ulSM_A6; + ULONG ulSM_A7; + UCHAR ucSM_A0_sign; + UCHAR ucSM_A1_sign; + UCHAR ucSM_A2_sign; + UCHAR ucSM_A3_sign; + UCHAR ucSM_A4_sign; + UCHAR ucSM_A5_sign; + UCHAR ucSM_A6_sign; + UCHAR ucSM_A7_sign; + ULONG ulMargin_RO_a; + ULONG ulMargin_RO_b; + ULONG ulMargin_RO_c; + ULONG ulMargin_fixed; + ULONG ulMargin_Fmax_mean; + ULONG ulMargin_plat_mean; + ULONG ulMargin_Fmax_sigma; + ULONG ulMargin_plat_sigma; + ULONG ulMargin_DC_sigma; + ULONG ulLoadLineSlop; + ULONG ulaTDClimitPerDPM[8]; + ULONG ulaNoCalcVddcPerDPM[8]; + ULONG ulAVFS_meanNsigma_Acontant0; + ULONG ulAVFS_meanNsigma_Acontant1; + ULONG ulAVFS_meanNsigma_Acontant2; + USHORT usAVFS_meanNsigma_DC_tol_sigma; + USHORT usAVFS_meanNsigma_Platform_mean; + USHORT usAVFS_meanNsigma_Platform_sigma; + ULONG ulGB_VDROOP_TABLE_CKSOFF_a0; + ULONG ulGB_VDROOP_TABLE_CKSOFF_a1; + ULONG ulGB_VDROOP_TABLE_CKSOFF_a2; + ULONG ulGB_VDROOP_TABLE_CKSON_a0; + ULONG ulGB_VDROOP_TABLE_CKSON_a1; + ULONG ulGB_VDROOP_TABLE_CKSON_a2; + ULONG ulAVFSGB_FUSE_TABLE_CKSOFF_m1; + USHORT usAVFSGB_FUSE_TABLE_CKSOFF_m2; + ULONG ulAVFSGB_FUSE_TABLE_CKSOFF_b; + ULONG ulAVFSGB_FUSE_TABLE_CKSON_m1; + USHORT usAVFSGB_FUSE_TABLE_CKSON_m2; + ULONG ulAVFSGB_FUSE_TABLE_CKSON_b; + USHORT usMaxVoltage_0_25mv; + UCHAR ucEnableGB_VDROOP_TABLE_CKSOFF; + UCHAR ucEnableGB_VDROOP_TABLE_CKSON; + UCHAR ucEnableGB_FUSE_TABLE_CKSOFF; + UCHAR ucEnableGB_FUSE_TABLE_CKSON; + USHORT usPSM_Age_ComFactor; + UCHAR ucEnableApplyAVFS_CKS_OFF_Voltage; + UCHAR ucReserved; +}ATOM_ASIC_PROFILING_INFO_V3_6; + typedef struct _ATOM_SCLK_FCW_RANGE_ENTRY_V1{ ULONG ulMaxSclkFreq; diff --git a/drivers/gpu/drm/amd/include/cgs_common.h b/drivers/gpu/drm/amd/include/cgs_common.h index a461e155a160..b86aba9d019f 100644 --- a/drivers/gpu/drm/amd/include/cgs_common.h +++ b/drivers/gpu/drm/amd/include/cgs_common.h @@ -49,6 +49,7 @@ enum cgs_ind_reg { CGS_IND_REG__SMC, CGS_IND_REG__UVD_CTX, CGS_IND_REG__DIDT, + CGS_IND_REG_GC_CAC, CGS_IND_REG__AUDIO_ENDPT }; @@ -112,20 +113,23 @@ enum cgs_system_info_id { CGS_SYSTEM_INFO_ADAPTER_BDF_ID = 1, CGS_SYSTEM_INFO_PCIE_GEN_INFO, CGS_SYSTEM_INFO_PCIE_MLW, + CGS_SYSTEM_INFO_PCIE_DEV, + CGS_SYSTEM_INFO_PCIE_REV, CGS_SYSTEM_INFO_CG_FLAGS, CGS_SYSTEM_INFO_PG_FLAGS, CGS_SYSTEM_INFO_GFX_CU_INFO, + CGS_SYSTEM_INFO_GFX_SE_INFO, CGS_SYSTEM_INFO_ID_MAXIMUM, }; struct cgs_system_info { - uint64_t size; - uint64_t info_id; + uint64_t size; + enum cgs_system_info_id info_id; union { - void *ptr; - uint64_t value; + void *ptr; + uint64_t value; }; - uint64_t padding[13]; + uint64_t padding[13]; }; /* @@ -158,6 +162,10 @@ struct cgs_firmware_info { uint16_t feature_version; uint32_t image_size; uint64_t mc_addr; + + /* only for smc firmware */ + uint32_t ucode_start_address; + void *kptr; }; @@ -189,7 +197,6 @@ typedef unsigned long cgs_handle_t; struct cgs_acpi_method_argument { uint32_t type; - uint32_t method_length; uint32_t data_length; union{ uint32_t value; @@ -581,6 +588,9 @@ typedef int (*cgs_get_firmware_info)(struct cgs_device *cgs_device, enum cgs_ucode_id type, struct cgs_firmware_info *info); +typedef int (*cgs_rel_firmware)(struct cgs_device *cgs_device, + enum cgs_ucode_id type); + typedef int(*cgs_set_powergating_state)(struct cgs_device *cgs_device, enum amd_ip_block_type block_type, enum amd_powergating_state state); @@ -645,6 +655,7 @@ struct cgs_ops { cgs_set_camera_voltages_t set_camera_voltages; /* Firmware Info */ cgs_get_firmware_info get_firmware_info; + cgs_rel_firmware rel_firmware; /* cg pg interface*/ cgs_set_powergating_state set_powergating_state; cgs_set_clockgating_state set_clockgating_state; @@ -738,6 +749,8 @@ struct cgs_device CGS_CALL(set_camera_voltages,dev,mask,voltages) #define cgs_get_firmware_info(dev, type, info) \ CGS_CALL(get_firmware_info, dev, type, info) +#define cgs_rel_firmware(dev, type) \ + CGS_CALL(rel_firmware, dev, type) #define cgs_set_powergating_state(dev, block_type, state) \ CGS_CALL(set_powergating_state, dev, block_type, state) #define cgs_set_clockgating_state(dev, block_type, state) \ diff --git a/drivers/gpu/drm/amd/powerplay/amd_powerplay.c b/drivers/gpu/drm/amd/powerplay/amd_powerplay.c index 8e345bfddb69..abbb658bdc1e 100644 --- a/drivers/gpu/drm/amd/powerplay/amd_powerplay.c +++ b/drivers/gpu/drm/amd/powerplay/amd_powerplay.c @@ -73,11 +73,14 @@ static int pp_sw_init(void *handle) ret = hwmgr->hwmgr_func->backend_init(hwmgr); if (ret) - goto err; + goto err1; pr_info("amdgpu: powerplay initialized\n"); return 0; +err1: + if (hwmgr->pptable_func->pptable_fini) + hwmgr->pptable_func->pptable_fini(hwmgr); err: pr_err("amdgpu: powerplay initialization failed\n"); return ret; @@ -100,6 +103,9 @@ static int pp_sw_fini(void *handle) if (hwmgr->hwmgr_func->backend_fini != NULL) ret = hwmgr->hwmgr_func->backend_fini(hwmgr); + if (hwmgr->pptable_func->pptable_fini) + hwmgr->pptable_func->pptable_fini(hwmgr); + return ret; } @@ -170,7 +176,7 @@ static int pp_hw_fini(void *handle) static bool pp_is_idle(void *handle) { - return 0; + return false; } static int pp_wait_for_idle(void *handle) @@ -530,6 +536,10 @@ int pp_dpm_dispatch_tasks(void *handle, enum amd_pp_event event_id, void *input, case AMD_PP_EVENT_COMPLETE_INIT: ret = pem_handle_event(pp_handle->eventmgr, event_id, &data); break; + case AMD_PP_EVENT_READJUST_POWER_STATE: + pp_handle->hwmgr->current_ps = pp_handle->hwmgr->boot_ps; + ret = pem_handle_event(pp_handle->eventmgr, event_id, &data); + break; default: break; } @@ -734,12 +744,12 @@ static int pp_dpm_get_pp_table(void *handle, char **table) PP_CHECK_HW(hwmgr); - if (hwmgr->hwmgr_func->get_pp_table == NULL) { - printk(KERN_INFO "%s was not implemented.\n", __func__); - return 0; - } + if (!hwmgr->soft_pp_table) + return -EINVAL; - return hwmgr->hwmgr_func->get_pp_table(hwmgr, table); + *table = (char *)hwmgr->soft_pp_table; + + return hwmgr->soft_pp_table_size; } static int pp_dpm_set_pp_table(void *handle, const char *buf, size_t size) @@ -753,12 +763,23 @@ static int pp_dpm_set_pp_table(void *handle, const char *buf, size_t size) PP_CHECK_HW(hwmgr); - if (hwmgr->hwmgr_func->set_pp_table == NULL) { - printk(KERN_INFO "%s was not implemented.\n", __func__); - return 0; + if (!hwmgr->hardcode_pp_table) { + hwmgr->hardcode_pp_table = + kzalloc(hwmgr->soft_pp_table_size, GFP_KERNEL); + + if (!hwmgr->hardcode_pp_table) + return -ENOMEM; + + /* to avoid powerplay crash when hardcode pptable is empty */ + memcpy(hwmgr->hardcode_pp_table, hwmgr->soft_pp_table, + hwmgr->soft_pp_table_size); } - return hwmgr->hwmgr_func->set_pp_table(hwmgr, buf, size); + memcpy(hwmgr->hardcode_pp_table, buf, size); + + hwmgr->soft_pp_table = hwmgr->hardcode_pp_table; + + return amd_powerplay_reset(handle); } static int pp_dpm_force_clock_level(void *handle, @@ -800,6 +821,82 @@ static int pp_dpm_print_clock_levels(void *handle, return hwmgr->hwmgr_func->print_clock_levels(hwmgr, type, buf); } +static int pp_dpm_get_sclk_od(void *handle) +{ + struct pp_hwmgr *hwmgr; + + if (!handle) + return -EINVAL; + + hwmgr = ((struct pp_instance *)handle)->hwmgr; + + PP_CHECK_HW(hwmgr); + + if (hwmgr->hwmgr_func->get_sclk_od == NULL) { + printk(KERN_INFO "%s was not implemented.\n", __func__); + return 0; + } + + return hwmgr->hwmgr_func->get_sclk_od(hwmgr); +} + +static int pp_dpm_set_sclk_od(void *handle, uint32_t value) +{ + struct pp_hwmgr *hwmgr; + + if (!handle) + return -EINVAL; + + hwmgr = ((struct pp_instance *)handle)->hwmgr; + + PP_CHECK_HW(hwmgr); + + if (hwmgr->hwmgr_func->set_sclk_od == NULL) { + printk(KERN_INFO "%s was not implemented.\n", __func__); + return 0; + } + + return hwmgr->hwmgr_func->set_sclk_od(hwmgr, value); +} + +static int pp_dpm_get_mclk_od(void *handle) +{ + struct pp_hwmgr *hwmgr; + + if (!handle) + return -EINVAL; + + hwmgr = ((struct pp_instance *)handle)->hwmgr; + + PP_CHECK_HW(hwmgr); + + if (hwmgr->hwmgr_func->get_mclk_od == NULL) { + printk(KERN_INFO "%s was not implemented.\n", __func__); + return 0; + } + + return hwmgr->hwmgr_func->get_mclk_od(hwmgr); +} + +static int pp_dpm_set_mclk_od(void *handle, uint32_t value) +{ + struct pp_hwmgr *hwmgr; + + if (!handle) + return -EINVAL; + + hwmgr = ((struct pp_instance *)handle)->hwmgr; + + PP_CHECK_HW(hwmgr); + + if (hwmgr->hwmgr_func->set_mclk_od == NULL) { + printk(KERN_INFO "%s was not implemented.\n", __func__); + return 0; + } + + return hwmgr->hwmgr_func->set_mclk_od(hwmgr, value); +} + const struct amd_powerplay_funcs pp_dpm_funcs = { .get_temperature = pp_dpm_get_temperature, .load_firmware = pp_dpm_load_fw, @@ -822,6 +919,10 @@ const struct amd_powerplay_funcs pp_dpm_funcs = { .set_pp_table = pp_dpm_set_pp_table, .force_clock_level = pp_dpm_force_clock_level, .print_clock_levels = pp_dpm_print_clock_levels, + .get_sclk_od = pp_dpm_get_sclk_od, + .set_sclk_od = pp_dpm_set_sclk_od, + .get_mclk_od = pp_dpm_get_mclk_od, + .set_mclk_od = pp_dpm_set_mclk_od, }; static int amd_pp_instance_init(struct amd_pp_init *pp_init, @@ -903,6 +1004,44 @@ int amd_powerplay_fini(void *handle) return 0; } +int amd_powerplay_reset(void *handle) +{ + struct pp_instance *instance = (struct pp_instance *)handle; + struct pp_eventmgr *eventmgr; + struct pem_event_data event_data = { {0} }; + int ret; + + if (instance == NULL) + return -EINVAL; + + eventmgr = instance->eventmgr; + if (!eventmgr || !eventmgr->pp_eventmgr_fini) + return -EINVAL; + + eventmgr->pp_eventmgr_fini(eventmgr); + + ret = pp_sw_fini(handle); + if (ret) + return ret; + + kfree(instance->hwmgr->ps); + + ret = pp_sw_init(handle); + if (ret) + return ret; + + hw_init_power_state_table(instance->hwmgr); + + if (eventmgr == NULL || eventmgr->pp_eventmgr_init == NULL) + return -EINVAL; + + ret = eventmgr->pp_eventmgr_init(eventmgr); + if (ret) + return ret; + + return pem_handle_event(eventmgr, AMD_PP_EVENT_COMPLETE_INIT, &event_data); +} + /* export this function to DAL */ int amd_powerplay_display_configuration_change(void *handle, diff --git a/drivers/gpu/drm/amd/powerplay/eventmgr/eventactionchains.c b/drivers/gpu/drm/amd/powerplay/eventmgr/eventactionchains.c index d6635cc4b0fc..635fc4b48184 100644 --- a/drivers/gpu/drm/amd/powerplay/eventmgr/eventactionchains.c +++ b/drivers/gpu/drm/amd/powerplay/eventmgr/eventactionchains.c @@ -30,7 +30,6 @@ static const pem_event_action * const initialize_event[] = { system_config_tasks, setup_asic_tasks, enable_dynamic_state_management_tasks, - enable_clock_power_gatings_tasks, get_2d_performance_state_tasks, set_performance_state_tasks, initialize_thermal_controller_tasks, @@ -140,7 +139,6 @@ static const pem_event_action * const resume_event[] = { setup_asic_tasks, enable_stutter_mode_tasks, /*must do this in boot state and before SMC is started */ enable_dynamic_state_management_tasks, - enable_clock_power_gatings_tasks, enable_disable_bapm_tasks, initialize_thermal_controller_tasks, get_2d_performance_state_tasks, diff --git a/drivers/gpu/drm/amd/powerplay/eventmgr/eventmgr.c b/drivers/gpu/drm/amd/powerplay/eventmgr/eventmgr.c index 46410e3c7349..fb88e4e5d625 100644 --- a/drivers/gpu/drm/amd/powerplay/eventmgr/eventmgr.c +++ b/drivers/gpu/drm/amd/powerplay/eventmgr/eventmgr.c @@ -58,9 +58,6 @@ static void pem_fini(struct pp_eventmgr *eventmgr) pem_unregister_interrupts(eventmgr); pem_handle_event(eventmgr, AMD_PP_EVENT_UNINITIALIZE, &event_data); - - if (eventmgr != NULL) - kfree(eventmgr); } int eventmgr_init(struct pp_instance *handle) diff --git a/drivers/gpu/drm/amd/powerplay/eventmgr/eventtasks.c b/drivers/gpu/drm/amd/powerplay/eventmgr/eventtasks.c index 5cd123472db4..b6f45fd01fa6 100644 --- a/drivers/gpu/drm/amd/powerplay/eventmgr/eventtasks.c +++ b/drivers/gpu/drm/amd/powerplay/eventmgr/eventtasks.c @@ -132,8 +132,7 @@ int pem_task_enable_dynamic_state_management(struct pp_eventmgr *eventmgr, struc int pem_task_disable_dynamic_state_management(struct pp_eventmgr *eventmgr, struct pem_event_data *event_data) { - /* TODO */ - return 0; + return phm_disable_dynamic_state_management(eventmgr->hwmgr); } int pem_task_enable_clock_power_gatings_tasks(struct pp_eventmgr *eventmgr, struct pem_event_data *event_data) diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/cz_clockpowergating.c b/drivers/gpu/drm/amd/powerplay/hwmgr/cz_clockpowergating.c index 436fc16dabb6..2028980f1ed4 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/cz_clockpowergating.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/cz_clockpowergating.c @@ -177,12 +177,12 @@ int cz_dpm_powergate_uvd(struct pp_hwmgr *hwmgr, bool bgate) cz_dpm_powerdown_uvd(hwmgr); } else { cz_dpm_powerup_uvd(hwmgr); - cgs_set_clockgating_state(hwmgr->device, - AMD_IP_BLOCK_TYPE_UVD, - AMD_PG_STATE_GATE); cgs_set_powergating_state(hwmgr->device, AMD_IP_BLOCK_TYPE_UVD, AMD_CG_STATE_UNGATE); + cgs_set_clockgating_state(hwmgr->device, + AMD_IP_BLOCK_TYPE_UVD, + AMD_PG_STATE_GATE); cz_dpm_update_uvd_dpm(hwmgr, false); } @@ -206,25 +206,26 @@ int cz_dpm_powergate_vce(struct pp_hwmgr *hwmgr, bool bgate) AMD_IP_BLOCK_TYPE_VCE, AMD_PG_STATE_GATE); cz_enable_disable_vce_dpm(hwmgr, false); - /* TODO: to figure out why vce can't be poweroff*/ + cz_dpm_powerdown_vce(hwmgr); cz_hwmgr->vce_power_gated = true; } else { cz_dpm_powerup_vce(hwmgr); cz_hwmgr->vce_power_gated = false; - cgs_set_clockgating_state( - hwmgr->device, - AMD_IP_BLOCK_TYPE_VCE, - AMD_PG_STATE_GATE); cgs_set_powergating_state( hwmgr->device, AMD_IP_BLOCK_TYPE_VCE, AMD_CG_STATE_UNGATE); + cgs_set_clockgating_state( + hwmgr->device, + AMD_IP_BLOCK_TYPE_VCE, + AMD_PG_STATE_GATE); cz_dpm_update_vce_dpm(hwmgr); cz_enable_disable_vce_dpm(hwmgr, true); return 0; } } } else { + cz_hwmgr->vce_power_gated = bgate; cz_dpm_update_vce_dpm(hwmgr); cz_enable_disable_vce_dpm(hwmgr, !bgate); return 0; diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/cz_hwmgr.c b/drivers/gpu/drm/amd/powerplay/hwmgr/cz_hwmgr.c index 1f14c477d15d..8cc0df9b534a 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/cz_hwmgr.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/cz_hwmgr.c @@ -1167,9 +1167,9 @@ static int cz_apply_state_adjust_rules(struct pp_hwmgr *hwmgr, cz_ps->action = cz_current_ps->action; - if ((force_high == false) && (cz_ps->action == FORCE_HIGH)) + if (!force_high && (cz_ps->action == FORCE_HIGH)) cz_ps->action = CANCEL_FORCE_HIGH; - else if ((force_high == true) && (cz_ps->action != FORCE_HIGH)) + else if (force_high && (cz_ps->action != FORCE_HIGH)) cz_ps->action = FORCE_HIGH; else cz_ps->action = DO_NOTHING; @@ -1180,6 +1180,13 @@ static int cz_apply_state_adjust_rules(struct pp_hwmgr *hwmgr, static int cz_hwmgr_backend_init(struct pp_hwmgr *hwmgr) { int result = 0; + struct cz_hwmgr *data; + + data = kzalloc(sizeof(struct cz_hwmgr), GFP_KERNEL); + if (data == NULL) + return -ENOMEM; + + hwmgr->backend = data; result = cz_initialize_dpm_defaults(hwmgr); if (result != 0) { @@ -1649,7 +1656,7 @@ static void cz_hw_print_display_cfg( struct cz_hwmgr *hw_data = (struct cz_hwmgr *)(hwmgr->backend); uint32_t data = 0; - if (hw_data->cc6_settings.cc6_setting_changed == true) { + if (hw_data->cc6_settings.cc6_setting_changed) { hw_data->cc6_settings.cc6_setting_changed = false; @@ -1909,15 +1916,7 @@ static const struct pp_hwmgr_func cz_hwmgr_funcs = { int cz_hwmgr_init(struct pp_hwmgr *hwmgr) { - struct cz_hwmgr *cz_hwmgr; - int ret = 0; - - cz_hwmgr = kzalloc(sizeof(struct cz_hwmgr), GFP_KERNEL); - if (cz_hwmgr == NULL) - return -ENOMEM; - - hwmgr->backend = cz_hwmgr; hwmgr->hwmgr_func = &cz_hwmgr_funcs; hwmgr->pptable_func = &pptable_funcs; - return ret; + return 0; } diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/fiji_clockpowergating.c b/drivers/gpu/drm/amd/powerplay/hwmgr/fiji_clockpowergating.c index e1b649bd5344..5afe82068b29 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/fiji_clockpowergating.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/fiji_clockpowergating.c @@ -56,7 +56,7 @@ int fiji_phm_powergate_uvd(struct pp_hwmgr *hwmgr, bool bgate) fiji_update_uvd_dpm(hwmgr, false); cgs_set_clockgating_state(hwmgr->device, AMD_IP_BLOCK_TYPE_UVD, - AMD_PG_STATE_UNGATE); + AMD_CG_STATE_UNGATE); } return 0; diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/fiji_hwmgr.c b/drivers/gpu/drm/amd/powerplay/hwmgr/fiji_hwmgr.c index 24a16e49b571..120a9e2c3152 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/fiji_hwmgr.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/fiji_hwmgr.c @@ -581,25 +581,24 @@ static int fiji_patch_boot_state(struct pp_hwmgr *hwmgr, static int fiji_hwmgr_backend_fini(struct pp_hwmgr *hwmgr) { - struct fiji_hwmgr *data = (struct fiji_hwmgr *)(hwmgr->backend); - - if (data->soft_pp_table) { - kfree(data->soft_pp_table); - data->soft_pp_table = NULL; - } - return phm_hwmgr_backend_fini(hwmgr); } static int fiji_hwmgr_backend_init(struct pp_hwmgr *hwmgr) { - struct fiji_hwmgr *data = (struct fiji_hwmgr *)(hwmgr->backend); + struct fiji_hwmgr *data; uint32_t i; struct phm_ppt_v1_information *table_info = (struct phm_ppt_v1_information *)(hwmgr->pptable); bool stay_in_boot; int result; + data = kzalloc(sizeof(struct fiji_hwmgr), GFP_KERNEL); + if (data == NULL) + return -ENOMEM; + + hwmgr->backend = data; + data->dll_default_on = false; data->sram_end = SMC_RAM_END; @@ -633,6 +632,8 @@ static int fiji_hwmgr_backend_init(struct pp_hwmgr *hwmgr) data->vddci_control = FIJI_VOLTAGE_CONTROL_NONE; data->mvdd_control = FIJI_VOLTAGE_CONTROL_NONE; + data->force_pcie_gen = PP_PCIEGenInvalid; + if (atomctrl_is_voltage_controled_by_gpio_v3(hwmgr, VOLTAGE_TYPE_VDDC, VOLTAGE_OBJ_SVID2)) data->voltage_control = FIJI_VOLTAGE_CONTROL_BY_SVID2; @@ -697,7 +698,7 @@ static int fiji_hwmgr_backend_init(struct pp_hwmgr *hwmgr) if (0 == result) { struct cgs_system_info sys_info = {0}; - data->is_tlu_enabled = 0; + data->is_tlu_enabled = false; hwmgr->platform_descriptor.hardwareActivityPerformanceLevels = FIJI_MAX_HARDWARE_POWERLEVELS; hwmgr->platform_descriptor.hardwarePerformanceLevels = 2; @@ -732,7 +733,7 @@ static int fiji_hwmgr_backend_init(struct pp_hwmgr *hwmgr) sys_info.info_id = CGS_SYSTEM_INFO_PCIE_GEN_INFO; result = cgs_query_system_info(hwmgr->device, &sys_info); if (result) - data->pcie_gen_cap = 0x30007; + data->pcie_gen_cap = AMDGPU_DEFAULT_PCIE_GEN_MASK; else data->pcie_gen_cap = (uint32_t)sys_info.value; if (data->pcie_gen_cap & CAIL_PCIE_LINK_SPEED_SUPPORT_GEN3) @@ -741,7 +742,7 @@ static int fiji_hwmgr_backend_init(struct pp_hwmgr *hwmgr) sys_info.info_id = CGS_SYSTEM_INFO_PCIE_MLW; result = cgs_query_system_info(hwmgr->device, &sys_info); if (result) - data->pcie_lane_cap = 0x2f0000; + data->pcie_lane_cap = AMDGPU_DEFAULT_PCIE_MLW_MASK; else data->pcie_lane_cap = (uint32_t)sys_info.value; } else { @@ -1234,6 +1235,34 @@ static int fiji_program_voting_clients(struct pp_hwmgr *hwmgr) return 0; } +static int fiji_clear_voting_clients(struct pp_hwmgr *hwmgr) +{ + /* Reset voting clients before disabling DPM */ + PHM_WRITE_INDIRECT_FIELD(hwmgr->device, CGS_IND_REG__SMC, + SCLK_PWRMGT_CNTL, RESET_SCLK_CNT, 1); + PHM_WRITE_INDIRECT_FIELD(hwmgr->device, CGS_IND_REG__SMC, + SCLK_PWRMGT_CNTL, RESET_BUSY_CNT, 1); + + cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC, + ixCG_FREQ_TRAN_VOTING_0, 0); + cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC, + ixCG_FREQ_TRAN_VOTING_1, 0); + cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC, + ixCG_FREQ_TRAN_VOTING_2, 0); + cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC, + ixCG_FREQ_TRAN_VOTING_3, 0); + cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC, + ixCG_FREQ_TRAN_VOTING_4, 0); + cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC, + ixCG_FREQ_TRAN_VOTING_5, 0); + cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC, + ixCG_FREQ_TRAN_VOTING_6, 0); + cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC, + ixCG_FREQ_TRAN_VOTING_7, 0); + + return 0; +} + /** * Get the location of various tables inside the FW image. * @@ -1361,6 +1390,17 @@ static int fiji_copy_and_switch_arb_sets(struct pp_hwmgr *hwmgr, } /** +* Call SMC to reset S0/S1 to S1 and Reset SMIO to initial value +* +* @param hwmgr the address of the powerplay hardware manager. +* @return if success then 0; +*/ +static int fiji_reset_to_default(struct pp_hwmgr *hwmgr) +{ + return smum_send_msg_to_smc(hwmgr->smumgr, PPSMC_MSG_ResetToDefaults); +} + +/** * Initial switch from ARB F0->F1 * * @param hwmgr the address of the powerplay hardware manager. @@ -1373,6 +1413,21 @@ static int fiji_initial_switch_from_arbf0_to_f1(struct pp_hwmgr *hwmgr) MC_CG_ARB_FREQ_F0, MC_CG_ARB_FREQ_F1); } +static int fiji_force_switch_to_arbf0(struct pp_hwmgr *hwmgr) +{ + uint32_t tmp; + + tmp = (cgs_read_ind_register(hwmgr->device, + CGS_IND_REG__SMC, ixSMC_SCRATCH9) & + 0x0000ff00) >> 8; + + if (tmp == MC_CG_ARB_FREQ_F0) + return 0; + + return fiji_copy_and_switch_arb_sets(hwmgr, + tmp, MC_CG_ARB_FREQ_F0); +} + static int fiji_reset_single_dpm_table(struct pp_hwmgr *hwmgr, struct fiji_single_dpm_table *dpm_table, uint32_t count) { @@ -1395,7 +1450,7 @@ static void fiji_setup_pcie_table_entry( { dpm_table->dpm_levels[index].value = pcie_gen; dpm_table->dpm_levels[index].param1 = pcie_lanes; - dpm_table->dpm_levels[index].enabled = 1; + dpm_table->dpm_levels[index].enabled = true; } static int fiji_setup_default_pcie_table(struct pp_hwmgr *hwmgr) @@ -1607,7 +1662,6 @@ static int fiji_populate_cac_table(struct pp_hwmgr *hwmgr, { uint32_t count; uint8_t index; - int result = 0; struct fiji_hwmgr *data = (struct fiji_hwmgr *)(hwmgr->backend); struct phm_ppt_v1_information *table_info = (struct phm_ppt_v1_information *)(hwmgr->pptable); @@ -1629,7 +1683,7 @@ static int fiji_populate_cac_table(struct pp_hwmgr *hwmgr, VOLTAGE_SCALE)) / 25); } - return result; + return 0; } /** @@ -1830,7 +1884,7 @@ static uint16_t fiji_find_closest_vddci(struct pp_hwmgr *hwmgr, uint16_t vddci) PP_ASSERT_WITH_CODE(false, "VDDCI is larger than max VDDCI in VDDCI Voltage Table!", - return vddci_table->entries[i].value); + return vddci_table->entries[i-1].value); } static int fiji_get_dependency_volt_by_clk(struct pp_hwmgr *hwmgr, @@ -3175,6 +3229,17 @@ static int fiji_enable_ulv(struct pp_hwmgr *hwmgr) return 0; } +static int fiji_disable_ulv(struct pp_hwmgr *hwmgr) +{ + struct fiji_hwmgr *data = (struct fiji_hwmgr *)(hwmgr->backend); + struct fiji_ulv_parm *ulv = &(data->ulv); + + if (ulv->ulv_supported) + return smum_send_msg_to_smc(hwmgr->smumgr, PPSMC_MSG_DisableULV); + + return 0; +} + static int fiji_enable_deep_sleep_master_switch(struct pp_hwmgr *hwmgr) { if (phm_cap_enabled(hwmgr->platform_descriptor.platformCaps, @@ -3195,6 +3260,21 @@ static int fiji_enable_deep_sleep_master_switch(struct pp_hwmgr *hwmgr) return 0; } +static int fiji_disable_deep_sleep_master_switch(struct pp_hwmgr *hwmgr) +{ + if (phm_cap_enabled(hwmgr->platform_descriptor.platformCaps, + PHM_PlatformCaps_SclkDeepSleep)) { + if (smum_send_msg_to_smc(hwmgr->smumgr, + PPSMC_MSG_MASTER_DeepSleep_OFF)) { + PP_ASSERT_WITH_CODE(false, + "Attempt to disable Master Deep Sleep switch failed!", + return -1); + } + } + + return 0; +} + static int fiji_enable_sclk_mclk_dpm(struct pp_hwmgr *hwmgr) { struct fiji_hwmgr *data = (struct fiji_hwmgr *)(hwmgr->backend); @@ -3355,6 +3435,70 @@ static int fiji_start_dpm(struct pp_hwmgr *hwmgr) return 0; } +static int fiji_disable_sclk_mclk_dpm(struct pp_hwmgr *hwmgr) +{ + struct fiji_hwmgr *data = (struct fiji_hwmgr *)(hwmgr->backend); + + /* disable SCLK dpm */ + if (!data->sclk_dpm_key_disabled) + PP_ASSERT_WITH_CODE( + (smum_send_msg_to_smc(hwmgr->smumgr, + PPSMC_MSG_DPM_Disable) == 0), + "Failed to disable SCLK DPM!", + return -1); + + /* disable MCLK dpm */ + if (!data->mclk_dpm_key_disabled) { + PP_ASSERT_WITH_CODE( + (smum_send_msg_to_smc_with_parameter(hwmgr->smumgr, + PPSMC_MSG_MCLKDPM_SetEnabledMask, 1) == 0), + "Failed to force MCLK DPM0!", + return -1); + + PP_ASSERT_WITH_CODE( + (smum_send_msg_to_smc(hwmgr->smumgr, + PPSMC_MSG_MCLKDPM_Disable) == 0), + "Failed to disable MCLK DPM!", + return -1); + } + + return 0; +} + +static int fiji_stop_dpm(struct pp_hwmgr *hwmgr) +{ + struct fiji_hwmgr *data = (struct fiji_hwmgr *)(hwmgr->backend); + + /* disable general power management */ + PHM_WRITE_INDIRECT_FIELD(hwmgr->device, CGS_IND_REG__SMC, GENERAL_PWRMGT, + GLOBAL_PWRMGT_EN, 0); + /* disable sclk deep sleep */ + PHM_WRITE_INDIRECT_FIELD(hwmgr->device, CGS_IND_REG__SMC, SCLK_PWRMGT_CNTL, + DYNAMIC_PM_EN, 0); + + /* disable PCIE dpm */ + if (!data->pcie_dpm_key_disabled) { + PP_ASSERT_WITH_CODE( + (smum_send_msg_to_smc(hwmgr->smumgr, + PPSMC_MSG_PCIeDPM_Disable) == 0), + "Failed to disable pcie DPM during DPM Stop Function!", + return -1); + } + + if (fiji_disable_sclk_mclk_dpm(hwmgr)) { + printk(KERN_ERR "Failed to disable Sclk DPM and Mclk DPM!"); + return -1; + } + + PP_ASSERT_WITH_CODE( + (smum_send_msg_to_smc(hwmgr->smumgr, + PPSMC_MSG_Voltage_Cntl_Disable) == 0), + "Failed to disable voltage DPM during DPM Stop Function!", + return -1); + + return 0; +} + static void fiji_set_dpm_event_sources(struct pp_hwmgr *hwmgr, uint32_t sources) { @@ -3413,6 +3557,23 @@ static int fiji_enable_thermal_auto_throttle(struct pp_hwmgr *hwmgr) return fiji_enable_auto_throttle_source(hwmgr, PHM_AutoThrottleSource_Thermal); } +static int fiji_disable_auto_throttle_source(struct pp_hwmgr *hwmgr, + PHM_AutoThrottleSource source) +{ + struct fiji_hwmgr *data = (struct fiji_hwmgr *)(hwmgr->backend); + + if (data->active_auto_throttle_sources & (1 << source)) { + data->active_auto_throttle_sources &= ~(1 << source); + fiji_set_dpm_event_sources(hwmgr, data->active_auto_throttle_sources); + } + return 0; +} + +static int fiji_disable_thermal_auto_throttle(struct pp_hwmgr *hwmgr) +{ + return fiji_disable_auto_throttle_source(hwmgr, PHM_AutoThrottleSource_Thermal); +} + static int fiji_enable_dpm_tasks(struct pp_hwmgr *hwmgr) { int tmp_result, result = 0; @@ -3527,6 +3688,64 @@ static int fiji_enable_dpm_tasks(struct pp_hwmgr *hwmgr) return result; } +static int fiji_disable_dpm_tasks(struct pp_hwmgr *hwmgr) +{ + int tmp_result, result = 0; + + tmp_result = (fiji_is_dpm_running(hwmgr)) ? 0 : -1; + PP_ASSERT_WITH_CODE(tmp_result == 0, + "DPM is not running right now, no need to disable DPM!", + return 0); + + if (phm_cap_enabled(hwmgr->platform_descriptor.platformCaps, + PHM_PlatformCaps_ThermalController)) + PHM_WRITE_INDIRECT_FIELD(hwmgr->device, CGS_IND_REG__SMC, + GENERAL_PWRMGT, THERMAL_PROTECTION_DIS, 1); + + tmp_result = fiji_disable_power_containment(hwmgr); + PP_ASSERT_WITH_CODE((tmp_result == 0), + "Failed to disable power containment!", result = tmp_result); + + tmp_result = fiji_disable_smc_cac(hwmgr); + PP_ASSERT_WITH_CODE((tmp_result == 0), + "Failed to disable SMC CAC!", result = tmp_result); + + PHM_WRITE_INDIRECT_FIELD(hwmgr->device, CGS_IND_REG__SMC, + CG_SPLL_SPREAD_SPECTRUM, SSEN, 0); + PHM_WRITE_INDIRECT_FIELD(hwmgr->device, CGS_IND_REG__SMC, + GENERAL_PWRMGT, DYN_SPREAD_SPECTRUM_EN, 0); + + tmp_result = fiji_disable_thermal_auto_throttle(hwmgr); + PP_ASSERT_WITH_CODE((tmp_result == 0), + "Failed to disable thermal auto throttle!", result = tmp_result); + + tmp_result = fiji_stop_dpm(hwmgr); + PP_ASSERT_WITH_CODE((tmp_result == 0), + "Failed to stop DPM!", result = tmp_result); + + tmp_result = fiji_disable_deep_sleep_master_switch(hwmgr); + PP_ASSERT_WITH_CODE((tmp_result == 0), + "Failed to disable deep sleep master switch!", result = tmp_result); + + tmp_result = fiji_disable_ulv(hwmgr); + PP_ASSERT_WITH_CODE((tmp_result == 0), + "Failed to disable ULV!", result = tmp_result); + + tmp_result = fiji_clear_voting_clients(hwmgr); + PP_ASSERT_WITH_CODE((tmp_result == 0), + "Failed to clear voting clients!", result = tmp_result); + + tmp_result = fiji_reset_to_default(hwmgr); + PP_ASSERT_WITH_CODE((tmp_result == 0), + "Failed to reset to default!", result = tmp_result); + + tmp_result = fiji_force_switch_to_arbf0(hwmgr); + PP_ASSERT_WITH_CODE((tmp_result == 0), + "Failed to force to switch arbf0!", result = tmp_result); + + return result; +} + static int fiji_force_dpm_highest(struct pp_hwmgr *hwmgr) { struct fiji_hwmgr *data = (struct fiji_hwmgr *)(hwmgr->backend); @@ -4169,8 +4388,9 @@ static int fiji_freeze_sclk_mclk_dpm(struct pp_hwmgr *hwmgr) if ((0 == data->sclk_dpm_key_disabled) && (data->need_update_smu7_dpm_table & (DPMTABLE_OD_UPDATE_SCLK + DPMTABLE_UPDATE_SCLK))) { - PP_ASSERT_WITH_CODE(true == fiji_is_dpm_running(hwmgr), - "Trying to freeze SCLK DPM when DPM is disabled",); + PP_ASSERT_WITH_CODE(fiji_is_dpm_running(hwmgr), + "Trying to freeze SCLK DPM when DPM is disabled", + ); PP_ASSERT_WITH_CODE(0 == smum_send_msg_to_smc(hwmgr->smumgr, PPSMC_MSG_SCLKDPM_FreezeLevel), "Failed to freeze SCLK DPM during FreezeSclkMclkDPM Function!", @@ -4180,8 +4400,9 @@ static int fiji_freeze_sclk_mclk_dpm(struct pp_hwmgr *hwmgr) if ((0 == data->mclk_dpm_key_disabled) && (data->need_update_smu7_dpm_table & DPMTABLE_OD_UPDATE_MCLK)) { - PP_ASSERT_WITH_CODE(true == fiji_is_dpm_running(hwmgr), - "Trying to freeze MCLK DPM when DPM is disabled",); + PP_ASSERT_WITH_CODE(fiji_is_dpm_running(hwmgr), + "Trying to freeze MCLK DPM when DPM is disabled", + ); PP_ASSERT_WITH_CODE(0 == smum_send_msg_to_smc(hwmgr->smumgr, PPSMC_MSG_MCLKDPM_FreezeLevel), "Failed to freeze MCLK DPM during FreezeSclkMclkDPM Function!", @@ -4351,7 +4572,6 @@ static int fiji_trim_single_dpm_states(struct pp_hwmgr *hwmgr, static int fiji_trim_dpm_states(struct pp_hwmgr *hwmgr, const struct fiji_power_state *fiji_ps) { - int result = 0; struct fiji_hwmgr *data = (struct fiji_hwmgr *)(hwmgr->backend); uint32_t high_limit_count; @@ -4371,7 +4591,7 @@ static int fiji_trim_dpm_states(struct pp_hwmgr *hwmgr, fiji_ps->performance_levels[0].memory_clock, fiji_ps->performance_levels[high_limit_count].memory_clock); - return result; + return 0; } static int fiji_generate_dpm_level_enable_mask( @@ -4630,8 +4850,9 @@ static int fiji_unfreeze_sclk_mclk_dpm(struct pp_hwmgr *hwmgr) (data->need_update_smu7_dpm_table & (DPMTABLE_OD_UPDATE_SCLK + DPMTABLE_UPDATE_SCLK))) { - PP_ASSERT_WITH_CODE(true == fiji_is_dpm_running(hwmgr), - "Trying to Unfreeze SCLK DPM when DPM is disabled",); + PP_ASSERT_WITH_CODE(fiji_is_dpm_running(hwmgr), + "Trying to Unfreeze SCLK DPM when DPM is disabled", + ); PP_ASSERT_WITH_CODE(0 == smum_send_msg_to_smc(hwmgr->smumgr, PPSMC_MSG_SCLKDPM_UnfreezeLevel), "Failed to unfreeze SCLK DPM during UnFreezeSclkMclkDPM Function!", @@ -4641,8 +4862,9 @@ static int fiji_unfreeze_sclk_mclk_dpm(struct pp_hwmgr *hwmgr) if ((0 == data->mclk_dpm_key_disabled) && (data->need_update_smu7_dpm_table & DPMTABLE_OD_UPDATE_MCLK)) { - PP_ASSERT_WITH_CODE(true == fiji_is_dpm_running(hwmgr), - "Trying to Unfreeze MCLK DPM when DPM is disabled",); + PP_ASSERT_WITH_CODE(fiji_is_dpm_running(hwmgr), + "Trying to Unfreeze MCLK DPM when DPM is disabled", + ); PP_ASSERT_WITH_CODE(0 == smum_send_msg_to_smc(hwmgr->smumgr, PPSMC_MSG_SCLKDPM_UnfreezeLevel), "Failed to unfreeze MCLK DPM during UnFreezeSclkMclkDPM Function!", @@ -5069,42 +5291,6 @@ static int fiji_get_fan_control_mode(struct pp_hwmgr *hwmgr) CG_FDO_CTRL2, FDO_PWM_MODE); } -static int fiji_get_pp_table(struct pp_hwmgr *hwmgr, char **table) -{ - struct fiji_hwmgr *data = (struct fiji_hwmgr *)(hwmgr->backend); - - if (!data->soft_pp_table) { - data->soft_pp_table = kmemdup(hwmgr->soft_pp_table, - hwmgr->soft_pp_table_size, - GFP_KERNEL); - if (!data->soft_pp_table) - return -ENOMEM; - } - - *table = (char *)&data->soft_pp_table; - - return hwmgr->soft_pp_table_size; -} - -static int fiji_set_pp_table(struct pp_hwmgr *hwmgr, const char *buf, size_t size) -{ - struct fiji_hwmgr *data = (struct fiji_hwmgr *)(hwmgr->backend); - - if (!data->soft_pp_table) { - data->soft_pp_table = kzalloc(hwmgr->soft_pp_table_size, GFP_KERNEL); - if (!data->soft_pp_table) - return -ENOMEM; - } - - memcpy(data->soft_pp_table, buf, size); - - hwmgr->soft_pp_table = data->soft_pp_table; - - /* TODO: re-init powerplay to implement modified pptable */ - - return 0; -} - static int fiji_force_clock_level(struct pp_hwmgr *hwmgr, enum pp_clock_type type, uint32_t mask) { @@ -5274,12 +5460,96 @@ bool fiji_check_smc_update_required_for_display_configuration(struct pp_hwmgr *h return is_update_required; } +static int fiji_get_sclk_od(struct pp_hwmgr *hwmgr) +{ + struct fiji_hwmgr *data = (struct fiji_hwmgr *)(hwmgr->backend); + struct fiji_single_dpm_table *sclk_table = &(data->dpm_table.sclk_table); + struct fiji_single_dpm_table *golden_sclk_table = + &(data->golden_dpm_table.sclk_table); + int value; + + value = (sclk_table->dpm_levels[sclk_table->count - 1].value - + golden_sclk_table->dpm_levels[golden_sclk_table->count - 1].value) * + 100 / + golden_sclk_table->dpm_levels[golden_sclk_table->count - 1].value; + + return value; +} + +static int fiji_set_sclk_od(struct pp_hwmgr *hwmgr, uint32_t value) +{ + struct fiji_hwmgr *data = (struct fiji_hwmgr *)(hwmgr->backend); + struct fiji_single_dpm_table *golden_sclk_table = + &(data->golden_dpm_table.sclk_table); + struct pp_power_state *ps; + struct fiji_power_state *fiji_ps; + + if (value > 20) + value = 20; + + ps = hwmgr->request_ps; + + if (ps == NULL) + return -EINVAL; + + fiji_ps = cast_phw_fiji_power_state(&ps->hardware); + + fiji_ps->performance_levels[fiji_ps->performance_level_count - 1].engine_clock = + golden_sclk_table->dpm_levels[golden_sclk_table->count - 1].value * + value / 100 + + golden_sclk_table->dpm_levels[golden_sclk_table->count - 1].value; + + return 0; +} + +static int fiji_get_mclk_od(struct pp_hwmgr *hwmgr) +{ + struct fiji_hwmgr *data = (struct fiji_hwmgr *)(hwmgr->backend); + struct fiji_single_dpm_table *mclk_table = &(data->dpm_table.mclk_table); + struct fiji_single_dpm_table *golden_mclk_table = + &(data->golden_dpm_table.mclk_table); + int value; + + value = (mclk_table->dpm_levels[mclk_table->count - 1].value - + golden_mclk_table->dpm_levels[golden_mclk_table->count - 1].value) * + 100 / + golden_mclk_table->dpm_levels[golden_mclk_table->count - 1].value; + + return value; +} + +static int fiji_set_mclk_od(struct pp_hwmgr *hwmgr, uint32_t value) +{ + struct fiji_hwmgr *data = (struct fiji_hwmgr *)(hwmgr->backend); + struct fiji_single_dpm_table *golden_mclk_table = + &(data->golden_dpm_table.mclk_table); + struct pp_power_state *ps; + struct fiji_power_state *fiji_ps; + + if (value > 20) + value = 20; + + ps = hwmgr->request_ps; + + if (ps == NULL) + return -EINVAL; + + fiji_ps = cast_phw_fiji_power_state(&ps->hardware); + + fiji_ps->performance_levels[fiji_ps->performance_level_count - 1].memory_clock = + golden_mclk_table->dpm_levels[golden_mclk_table->count - 1].value * + value / 100 + + golden_mclk_table->dpm_levels[golden_mclk_table->count - 1].value; + + return 0; +} static const struct pp_hwmgr_func fiji_hwmgr_funcs = { .backend_init = &fiji_hwmgr_backend_init, .backend_fini = &fiji_hwmgr_backend_fini, .asic_setup = &fiji_setup_asic_task, .dynamic_state_management_enable = &fiji_enable_dpm_tasks, + .dynamic_state_management_disable = &fiji_disable_dpm_tasks, .force_dpm_level = &fiji_dpm_force_dpm_level, .get_num_of_pp_table_entries = &tonga_get_number_of_powerplay_table_entries, .get_power_state_size = &fiji_get_power_state_size, @@ -5312,24 +5582,18 @@ static const struct pp_hwmgr_func fiji_hwmgr_funcs = { .get_fan_control_mode = fiji_get_fan_control_mode, .check_states_equal = fiji_check_states_equal, .check_smc_update_required_for_display_configuration = fiji_check_smc_update_required_for_display_configuration, - .get_pp_table = fiji_get_pp_table, - .set_pp_table = fiji_set_pp_table, .force_clock_level = fiji_force_clock_level, .print_clock_levels = fiji_print_clock_levels, + .get_sclk_od = fiji_get_sclk_od, + .set_sclk_od = fiji_set_sclk_od, + .get_mclk_od = fiji_get_mclk_od, + .set_mclk_od = fiji_set_mclk_od, }; int fiji_hwmgr_init(struct pp_hwmgr *hwmgr) { - struct fiji_hwmgr *data; - int ret = 0; - - data = kzalloc(sizeof(struct fiji_hwmgr), GFP_KERNEL); - if (data == NULL) - return -ENOMEM; - - hwmgr->backend = data; hwmgr->hwmgr_func = &fiji_hwmgr_funcs; hwmgr->pptable_func = &tonga_pptable_funcs; pp_fiji_thermal_initialize(hwmgr); - return ret; + return 0; } diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/fiji_hwmgr.h b/drivers/gpu/drm/amd/powerplay/hwmgr/fiji_hwmgr.h index 170edf5a772d..bf67c2a92c68 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/fiji_hwmgr.h +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/fiji_hwmgr.h @@ -302,9 +302,6 @@ struct fiji_hwmgr { bool pg_acp_init; bool frtc_enabled; bool frtc_status_changed; - - /* soft pptable for re-uploading into smu */ - void *soft_pp_table; }; /* To convert to Q8.8 format for firmware */ diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/fiji_powertune.c b/drivers/gpu/drm/amd/powerplay/hwmgr/fiji_powertune.c index db23a4068baf..44658451a8d2 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/fiji_powertune.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/fiji_powertune.c @@ -73,17 +73,18 @@ void fiji_initialize_power_tune_defaults(struct pp_hwmgr *hwmgr) if (!tmp) { phm_cap_set(hwmgr->platform_descriptor.platformCaps, - PHM_PlatformCaps_PowerContainment); - - phm_cap_set(hwmgr->platform_descriptor.platformCaps, PHM_PlatformCaps_CAC); fiji_hwmgr->fast_watermark_threshold = 100; - tmp = 1; - fiji_hwmgr->enable_dte_feature = tmp ? false : true; - fiji_hwmgr->enable_tdc_limit_feature = tmp ? true : false; - fiji_hwmgr->enable_pkg_pwr_tracking_feature = tmp ? true : false; + if (hwmgr->powercontainment_enabled) { + phm_cap_set(hwmgr->platform_descriptor.platformCaps, + PHM_PlatformCaps_PowerContainment); + tmp = 1; + fiji_hwmgr->enable_dte_feature = tmp ? false : true; + fiji_hwmgr->enable_tdc_limit_feature = tmp ? true : false; + fiji_hwmgr->enable_pkg_pwr_tracking_feature = tmp ? true : false; + } } } @@ -459,6 +460,23 @@ int fiji_enable_smc_cac(struct pp_hwmgr *hwmgr) return result; } +int fiji_disable_smc_cac(struct pp_hwmgr *hwmgr) +{ + struct fiji_hwmgr *data = (struct fiji_hwmgr *)(hwmgr->backend); + int result = 0; + + if (phm_cap_enabled(hwmgr->platform_descriptor.platformCaps, + PHM_PlatformCaps_CAC) && data->cac_enabled) { + int smc_result = smum_send_msg_to_smc(hwmgr->smumgr, + (uint16_t)(PPSMC_MSG_DisableCac)); + PP_ASSERT_WITH_CODE((smc_result == 0), + "Failed to disable CAC in SMC.", result = -1); + + data->cac_enabled = false; + } + return result; +} + int fiji_set_power_limit(struct pp_hwmgr *hwmgr, uint32_t n) { struct fiji_hwmgr *data = (struct fiji_hwmgr *)(hwmgr->backend); @@ -528,6 +546,48 @@ int fiji_enable_power_containment(struct pp_hwmgr *hwmgr) return result; } +int fiji_disable_power_containment(struct pp_hwmgr *hwmgr) +{ + struct fiji_hwmgr *data = (struct fiji_hwmgr *)(hwmgr->backend); + int result = 0; + + if (phm_cap_enabled(hwmgr->platform_descriptor.platformCaps, + PHM_PlatformCaps_PowerContainment) && + data->power_containment_features) { + int smc_result; + + if (data->power_containment_features & + POWERCONTAINMENT_FEATURE_TDCLimit) { + smc_result = smum_send_msg_to_smc(hwmgr->smumgr, + (uint16_t)(PPSMC_MSG_TDCLimitDisable)); + PP_ASSERT_WITH_CODE((smc_result == 0), + "Failed to disable TDCLimit in SMC.", + result = smc_result); + } + + if (data->power_containment_features & + POWERCONTAINMENT_FEATURE_DTE) { + smc_result = smum_send_msg_to_smc(hwmgr->smumgr, + (uint16_t)(PPSMC_MSG_DisableDTE)); + PP_ASSERT_WITH_CODE((smc_result == 0), + "Failed to disable DTE in SMC.", + result = smc_result); + } + + if (data->power_containment_features & + POWERCONTAINMENT_FEATURE_PkgPwrLimit) { + smc_result = smum_send_msg_to_smc(hwmgr->smumgr, + (uint16_t)(PPSMC_MSG_PkgPwrLimitDisable)); + PP_ASSERT_WITH_CODE((smc_result == 0), + "Failed to disable PkgPwrTracking in SMC.", + result = smc_result); + } + data->power_containment_features = 0; + } + + return result; +} + int fiji_power_control_set_level(struct pp_hwmgr *hwmgr) { struct phm_ppt_v1_information *table_info = diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/fiji_powertune.h b/drivers/gpu/drm/amd/powerplay/hwmgr/fiji_powertune.h index 55e58200f33a..fec772421733 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/fiji_powertune.h +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/fiji_powertune.h @@ -36,6 +36,19 @@ enum fiji_pt_config_reg_type { #define POWERCONTAINMENT_FEATURE_TDCLimit 0x00000002 #define POWERCONTAINMENT_FEATURE_PkgPwrLimit 0x00000004 +#define DIDT_SQ_CTRL0__UNUSED_0_MASK 0xffffffc0 +#define DIDT_SQ_CTRL0__UNUSED_0__SHIFT 0x6 +#define DIDT_TD_CTRL0__UNUSED_0_MASK 0xffffffc0 +#define DIDT_TD_CTRL0__UNUSED_0__SHIFT 0x6 +#define DIDT_TCP_CTRL0__UNUSED_0_MASK 0xffffffc0 +#define DIDT_TCP_CTRL0__UNUSED_0__SHIFT 0x6 +#define DIDT_SQ_TUNING_CTRL__UNUSED_0_MASK 0xe0000000 +#define DIDT_SQ_TUNING_CTRL__UNUSED_0__SHIFT 0x0000001d +#define DIDT_TD_TUNING_CTRL__UNUSED_0_MASK 0xe0000000 +#define DIDT_TD_TUNING_CTRL__UNUSED_0__SHIFT 0x0000001d +#define DIDT_TCP_TUNING_CTRL__UNUSED_0_MASK 0xe0000000 +#define DIDT_TCP_TUNING_CTRL__UNUSED_0__SHIFT 0x0000001d + struct fiji_pt_config_reg { uint32_t offset; uint32_t mask; @@ -58,7 +71,9 @@ void fiji_initialize_power_tune_defaults(struct pp_hwmgr *hwmgr); int fiji_populate_bapm_parameters_in_dpm_table(struct pp_hwmgr *hwmgr); int fiji_populate_pm_fuses(struct pp_hwmgr *hwmgr); int fiji_enable_smc_cac(struct pp_hwmgr *hwmgr); +int fiji_disable_smc_cac(struct pp_hwmgr *hwmgr); int fiji_enable_power_containment(struct pp_hwmgr *hwmgr); +int fiji_disable_power_containment(struct pp_hwmgr *hwmgr); int fiji_set_power_limit(struct pp_hwmgr *hwmgr, uint32_t n); int fiji_power_control_set_level(struct pp_hwmgr *hwmgr); diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/functiontables.c b/drivers/gpu/drm/amd/powerplay/hwmgr/functiontables.c index 7a705cee0cc2..a6abe81bc843 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/functiontables.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/functiontables.c @@ -59,8 +59,8 @@ int phm_dispatch_table(struct pp_hwmgr *hwmgr, struct phm_runtime_table_header *rt_table, void *input, void *output) { - int result = 0; - void *temp_storage = NULL; + int result; + void *temp_storage; if (hwmgr == NULL || rt_table == NULL) { printk(KERN_ERR "[ powerplay ] Invalid Parameter!\n"); @@ -73,12 +73,13 @@ int phm_dispatch_table(struct pp_hwmgr *hwmgr, printk(KERN_ERR "[ powerplay ] Could not allocate table temporary storage\n"); return -ENOMEM; } + } else { + temp_storage = NULL; } result = phm_run_table(hwmgr, rt_table, input, output, temp_storage); - if (NULL != temp_storage) - kfree(temp_storage); + kfree(temp_storage); return result; } diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/hardwaremanager.c b/drivers/gpu/drm/amd/powerplay/hwmgr/hardwaremanager.c index fa208ada6892..789f98ad2615 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/hardwaremanager.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/hardwaremanager.c @@ -154,6 +154,30 @@ int phm_enable_dynamic_state_management(struct pp_hwmgr *hwmgr) return ret; } +int phm_disable_dynamic_state_management(struct pp_hwmgr *hwmgr) +{ + int ret = -1; + bool enabled; + + PHM_FUNC_CHECK(hwmgr); + + if (phm_cap_enabled(hwmgr->platform_descriptor.platformCaps, + PHM_PlatformCaps_TablelessHardwareInterface)) { + if (hwmgr->hwmgr_func->dynamic_state_management_disable) + ret = hwmgr->hwmgr_func->dynamic_state_management_disable(hwmgr); + } else { + ret = phm_dispatch_table(hwmgr, + &(hwmgr->disable_dynamic_state_management), + NULL, NULL); + } + + enabled = ret == 0 ? false : true; + + cgs_notify_dpm_enabled(hwmgr->device, enabled); + + return ret; +} + int phm_force_dpm_levels(struct pp_hwmgr *hwmgr, enum amd_dpm_forced_level level) { PHM_FUNC_CHECK(hwmgr); @@ -306,11 +330,15 @@ int phm_store_dal_configuration_data(struct pp_hwmgr *hwmgr, { PHM_FUNC_CHECK(hwmgr); - if (hwmgr->hwmgr_func->store_cc6_data == NULL) + if (display_config == NULL) return -EINVAL; hwmgr->display_config = *display_config; - /* to do pass other display configuration in furture */ + + if (hwmgr->hwmgr_func->store_cc6_data == NULL) + return -EINVAL; + + /* TODO: pass other display configuration in the future */ if (hwmgr->hwmgr_func->store_cc6_data) hwmgr->hwmgr_func->store_cc6_data(hwmgr, diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/hwmgr.c b/drivers/gpu/drm/amd/powerplay/hwmgr/hwmgr.c index 1c48917da3cf..27e07624ac28 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/hwmgr.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/hwmgr.c @@ -24,6 +24,7 @@ #include <linux/types.h> #include <linux/kernel.h> #include <linux/slab.h> +#include <drm/amdgpu_drm.h> #include "cgs_common.h" #include "power_state.h" #include "hwmgr.h" @@ -58,12 +59,13 @@ int hwmgr_init(struct amd_pp_init *pp_init, struct pp_instance *handle) hwmgr->hw_revision = pp_init->rev_id; hwmgr->usec_timeout = AMD_MAX_USEC_TIMEOUT; hwmgr->power_source = PP_PowerSource_AC; + hwmgr->powercontainment_enabled = pp_init->powercontainment_enabled; switch (hwmgr->chip_family) { - case AMD_FAMILY_CZ: + case AMDGPU_FAMILY_CZ: cz_hwmgr_init(hwmgr); break; - case AMD_FAMILY_VI: + case AMDGPU_FAMILY_VI: switch (hwmgr->chip_id) { case CHIP_TONGA: tonga_hwmgr_init(hwmgr); @@ -93,6 +95,15 @@ int hwmgr_fini(struct pp_hwmgr *hwmgr) if (hwmgr == NULL || hwmgr->ps == NULL) return -EINVAL; + /* do hwmgr finish*/ + kfree(hwmgr->hardcode_pp_table); + + kfree(hwmgr->backend); + + kfree(hwmgr->start_thermal_controller.function_list); + + kfree(hwmgr->set_temperature_range.function_list); + kfree(hwmgr->ps); kfree(hwmgr); return 0; @@ -462,7 +473,7 @@ uint16_t phm_find_closest_vddci(struct pp_atomctrl_voltage_table *vddci_table, u PP_ASSERT_WITH_CODE(false, "VDDCI is larger than max VDDCI in VDDCI Voltage Table!", - return vddci_table->entries[i].value); + return vddci_table->entries[i-1].value); } int phm_find_boot_level(void *table, @@ -523,7 +534,7 @@ int phm_initializa_dynamic_state_adjustment_rule_settings(struct pp_hwmgr *hwmgr /* initialize vddc_dep_on_dal_pwrl table */ table_size = sizeof(uint32_t) + 4 * sizeof(struct phm_clock_voltage_dependency_record); - table_clk_vlt = (struct phm_clock_voltage_dependency_table *)kzalloc(table_size, GFP_KERNEL); + table_clk_vlt = kzalloc(table_size, GFP_KERNEL); if (NULL == table_clk_vlt) { printk(KERN_ERR "[ powerplay ] Can not allocate space for vddc_dep_on_dal_pwrl! \n"); diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/hwmgr_ppt.h b/drivers/gpu/drm/amd/powerplay/hwmgr/hwmgr_ppt.h index 347fef127ce9..2930a3355948 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/hwmgr_ppt.h +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/hwmgr_ppt.h @@ -39,6 +39,7 @@ struct phm_ppt_v1_clock_voltage_dependency_record { uint8_t phases; uint8_t cks_enable; uint8_t cks_voffset; + uint32_t sclk_offset; }; typedef struct phm_ppt_v1_clock_voltage_dependency_record phm_ppt_v1_clock_voltage_dependency_record; diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_clockpowergating.c b/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_clockpowergating.c index 8f142a74ad08..b5edb5105986 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_clockpowergating.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_clockpowergating.c @@ -106,11 +106,17 @@ int polaris10_phm_powergate_uvd(struct pp_hwmgr *hwmgr, bool bgate) data->uvd_power_gated = bgate; if (bgate) { + cgs_set_clockgating_state(hwmgr->device, + AMD_IP_BLOCK_TYPE_UVD, + AMD_CG_STATE_GATE); polaris10_update_uvd_dpm(hwmgr, true); polaris10_phm_powerdown_uvd(hwmgr); } else { polaris10_phm_powerup_uvd(hwmgr); polaris10_update_uvd_dpm(hwmgr, false); + cgs_set_clockgating_state(hwmgr->device, + AMD_IP_BLOCK_TYPE_UVD, + AMD_CG_STATE_UNGATE); } return 0; @@ -125,11 +131,19 @@ int polaris10_phm_powergate_vce(struct pp_hwmgr *hwmgr, bool bgate) data->vce_power_gated = bgate; - if (bgate) + if (bgate) { + cgs_set_clockgating_state(hwmgr->device, + AMD_IP_BLOCK_TYPE_VCE, + AMD_CG_STATE_GATE); + polaris10_update_vce_dpm(hwmgr, true); polaris10_phm_powerdown_vce(hwmgr); - else + } else { polaris10_phm_powerup_vce(hwmgr); - + polaris10_update_vce_dpm(hwmgr, false); + cgs_set_clockgating_state(hwmgr->device, + AMD_IP_BLOCK_TYPE_VCE, + AMD_CG_STATE_UNGATE); + } return 0; } diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_hwmgr.c b/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_hwmgr.c index aa6be033f21b..769636a0c5b5 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_hwmgr.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_hwmgr.c @@ -389,6 +389,34 @@ static int polaris10_program_voting_clients(struct pp_hwmgr *hwmgr) return 0; } +static int polaris10_clear_voting_clients(struct pp_hwmgr *hwmgr) +{ + /* Reset voting clients before disabling DPM */ + PHM_WRITE_INDIRECT_FIELD(hwmgr->device, CGS_IND_REG__SMC, + SCLK_PWRMGT_CNTL, RESET_SCLK_CNT, 1); + PHM_WRITE_INDIRECT_FIELD(hwmgr->device, CGS_IND_REG__SMC, + SCLK_PWRMGT_CNTL, RESET_BUSY_CNT, 1); + + cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC, + ixCG_FREQ_TRAN_VOTING_0, 0); + cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC, + ixCG_FREQ_TRAN_VOTING_1, 0); + cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC, + ixCG_FREQ_TRAN_VOTING_2, 0); + cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC, + ixCG_FREQ_TRAN_VOTING_3, 0); + cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC, + ixCG_FREQ_TRAN_VOTING_4, 0); + cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC, + ixCG_FREQ_TRAN_VOTING_5, 0); + cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC, + ixCG_FREQ_TRAN_VOTING_6, 0); + cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC, + ixCG_FREQ_TRAN_VOTING_7, 0); + + return 0; +} + /** * Get the location of various tables inside the FW image. * @@ -515,6 +543,11 @@ static int polaris10_copy_and_switch_arb_sets(struct pp_hwmgr *hwmgr, return 0; } +static int polaris10_reset_to_default(struct pp_hwmgr *hwmgr) +{ + return smum_send_msg_to_smc(hwmgr->smumgr, PPSMC_MSG_ResetToDefaults); +} + /** * Initial switch from ARB F0->F1 * @@ -528,6 +561,21 @@ static int polaris10_initial_switch_from_arbf0_to_f1(struct pp_hwmgr *hwmgr) MC_CG_ARB_FREQ_F0, MC_CG_ARB_FREQ_F1); } +static int polaris10_force_switch_to_arbf0(struct pp_hwmgr *hwmgr) +{ + uint32_t tmp; + + tmp = (cgs_read_ind_register(hwmgr->device, + CGS_IND_REG__SMC, ixSMC_SCRATCH9) & + 0x0000ff00) >> 8; + + if (tmp == MC_CG_ARB_FREQ_F0) + return 0; + + return polaris10_copy_and_switch_arb_sets(hwmgr, + tmp, MC_CG_ARB_FREQ_F0); +} + static int polaris10_setup_default_pcie_table(struct pp_hwmgr *hwmgr) { struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend); @@ -732,7 +780,7 @@ static int polaris10_populate_smc_mvdd_table(struct pp_hwmgr *hwmgr, table->Smio[level] |= data->mvdd_voltage_table.entries[level].smio_low; } - table->SmioMask2 = data->vddci_voltage_table.mask_low; + table->SmioMask2 = data->mvdd_voltage_table.mask_low; table->MvddLevelCount = (uint32_t) PP_HOST_TO_SMC_UL(count); } @@ -999,7 +1047,7 @@ static int polaris10_get_dependency_volt_by_clk(struct pp_hwmgr *hwmgr, vddci = phm_find_closest_vddci(&(data->vddci_voltage_table), (dep_table->entries[i].vddc - (uint16_t)data->vddc_vddci_delta)); - *voltage |= (vddci * VOLTAGE_SCALE) << VDDCI_SHIFT; + *voltage |= (vddci * VOLTAGE_SCALE) << VDDCI_SHIFT; } if (POLARIS10_VOLTAGE_CONTROL_NONE == data->mvdd_control) @@ -1296,7 +1344,6 @@ static int polaris10_populate_single_memory_level(struct pp_hwmgr *hwmgr, } mem_level->MclkFrequency = clock; - mem_level->StutterEnable = 0; mem_level->EnabledForThrottle = 1; mem_level->EnabledForActivity = 0; mem_level->UpHyst = 0; @@ -1304,7 +1351,6 @@ static int polaris10_populate_single_memory_level(struct pp_hwmgr *hwmgr, mem_level->VoltageDownHyst = 0; mem_level->ActivityLevel = (uint16_t)data->mclk_activity_target; mem_level->StutterEnable = false; - mem_level->DisplayWatermark = PPSMC_DISPLAY_WATERMARK_LOW; data->display_timing.num_existing_displays = info.display_count; @@ -1358,12 +1404,12 @@ static int polaris10_populate_all_memory_levels(struct pp_hwmgr *hwmgr) return result; } - /* in order to prevent MC activity from stutter mode to push DPM up. + /* In order to prevent MC activity from stutter mode to push DPM up, * the UVD change complements this by putting the MCLK in - * a higher state by default such that we are not effected by + * a higher state by default such that we are not affected by * up threshold or and MCLK DPM latency. */ - levels[0].ActivityLevel = (uint16_t)data->mclk_dpm0_activity_target; + levels[0].ActivityLevel = 0x1f; CONVERT_FROM_HOST_TO_SMC_US(levels[0].ActivityLevel); data->smc_state_table.MemoryDpmLevelCount = @@ -1424,22 +1470,19 @@ static int polaris10_populate_smc_acpi_level(struct pp_hwmgr *hwmgr, table->ACPILevel.Flags &= ~PPSMC_SWSTATE_FLAG_DC; - if (!data->sclk_dpm_key_disabled) { - /* Get MinVoltage and Frequency from DPM0, - * already converted to SMC_UL */ - sclk_frequency = data->dpm_table.sclk_table.dpm_levels[0].value; - result = polaris10_get_dependency_volt_by_clk(hwmgr, - table_info->vdd_dep_on_sclk, - table->ACPILevel.SclkFrequency, - &table->ACPILevel.MinVoltage, &mvdd); - PP_ASSERT_WITH_CODE((0 == result), - "Cannot find ACPI VDDC voltage value " - "in Clock Dependency Table", ); - } else { - sclk_frequency = data->vbios_boot_state.sclk_bootup_value; - table->ACPILevel.MinVoltage = - data->vbios_boot_state.vddc_bootup_value * VOLTAGE_SCALE; - } + + /* Get MinVoltage and Frequency from DPM0, + * already converted to SMC_UL */ + sclk_frequency = data->vbios_boot_state.sclk_bootup_value; + result = polaris10_get_dependency_volt_by_clk(hwmgr, + table_info->vdd_dep_on_sclk, + sclk_frequency, + &table->ACPILevel.MinVoltage, &mvdd); + PP_ASSERT_WITH_CODE((0 == result), + "Cannot find ACPI VDDC voltage value " + "in Clock Dependency Table", + ); + result = polaris10_calculate_sclk_params(hwmgr, sclk_frequency, &(table->ACPILevel.SclkSetting)); PP_ASSERT_WITH_CODE(result == 0, "Error retrieving Engine Clock dividers from VBIOS.", return result); @@ -1464,24 +1507,17 @@ static int polaris10_populate_smc_acpi_level(struct pp_hwmgr *hwmgr, CONVERT_FROM_HOST_TO_SMC_US(table->ACPILevel.SclkSetting.Fcw1_frac); CONVERT_FROM_HOST_TO_SMC_US(table->ACPILevel.SclkSetting.Sclk_ss_slew_rate); - if (!data->mclk_dpm_key_disabled) { - /* Get MinVoltage and Frequency from DPM0, already converted to SMC_UL */ - table->MemoryACPILevel.MclkFrequency = - data->dpm_table.mclk_table.dpm_levels[0].value; - result = polaris10_get_dependency_volt_by_clk(hwmgr, - table_info->vdd_dep_on_mclk, - table->MemoryACPILevel.MclkFrequency, - &table->MemoryACPILevel.MinVoltage, &mvdd); - PP_ASSERT_WITH_CODE((0 == result), - "Cannot find ACPI VDDCI voltage value " - "in Clock Dependency Table", - ); - } else { - table->MemoryACPILevel.MclkFrequency = - data->vbios_boot_state.mclk_bootup_value; - table->MemoryACPILevel.MinVoltage = - data->vbios_boot_state.vddci_bootup_value * VOLTAGE_SCALE; - } + + /* Get MinVoltage and Frequency from DPM0, already converted to SMC_UL */ + table->MemoryACPILevel.MclkFrequency = data->vbios_boot_state.mclk_bootup_value; + result = polaris10_get_dependency_volt_by_clk(hwmgr, + table_info->vdd_dep_on_mclk, + table->MemoryACPILevel.MclkFrequency, + &table->MemoryACPILevel.MinVoltage, &mvdd); + PP_ASSERT_WITH_CODE((0 == result), + "Cannot find ACPI VDDCI voltage value " + "in Clock Dependency Table", + ); us_mvdd = 0; if ((POLARIS10_VOLTAGE_CONTROL_NONE == data->mvdd_control) || @@ -1526,6 +1562,7 @@ static int polaris10_populate_smc_vce_level(struct pp_hwmgr *hwmgr, struct phm_ppt_v1_mm_clock_voltage_dependency_table *mm_table = table_info->mm_dep_table; struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend); + uint32_t vddci; table->VceLevelCount = (uint8_t)(mm_table->count); table->VceBootLevel = 0; @@ -1535,9 +1572,18 @@ static int polaris10_populate_smc_vce_level(struct pp_hwmgr *hwmgr, table->VceLevel[count].MinVoltage = 0; table->VceLevel[count].MinVoltage |= (mm_table->entries[count].vddc * VOLTAGE_SCALE) << VDDC_SHIFT; + + if (POLARIS10_VOLTAGE_CONTROL_BY_GPIO == data->vddci_control) + vddci = (uint32_t)phm_find_closest_vddci(&(data->vddci_voltage_table), + mm_table->entries[count].vddc - VDDC_VDDCI_DELTA); + else if (POLARIS10_VOLTAGE_CONTROL_BY_SVID2 == data->vddci_control) + vddci = mm_table->entries[count].vddc - VDDC_VDDCI_DELTA; + else + vddci = (data->vbios_boot_state.vddci_bootup_value * VOLTAGE_SCALE) << VDDCI_SHIFT; + + table->VceLevel[count].MinVoltage |= - ((mm_table->entries[count].vddc - data->vddc_vddci_delta) * - VOLTAGE_SCALE) << VDDCI_SHIFT; + (vddci * VOLTAGE_SCALE) << VDDCI_SHIFT; table->VceLevel[count].MinVoltage |= 1 << PHASES_SHIFT; /*retrieve divider value for VBIOS */ @@ -1566,6 +1612,7 @@ static int polaris10_populate_smc_samu_level(struct pp_hwmgr *hwmgr, struct phm_ppt_v1_mm_clock_voltage_dependency_table *mm_table = table_info->mm_dep_table; struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend); + uint32_t vddci; table->SamuBootLevel = 0; table->SamuLevelCount = (uint8_t)(mm_table->count); @@ -1576,8 +1623,16 @@ static int polaris10_populate_smc_samu_level(struct pp_hwmgr *hwmgr, table->SamuLevel[count].Frequency = mm_table->entries[count].samclock; table->SamuLevel[count].MinVoltage |= (mm_table->entries[count].vddc * VOLTAGE_SCALE) << VDDC_SHIFT; - table->SamuLevel[count].MinVoltage |= ((mm_table->entries[count].vddc - - data->vddc_vddci_delta) * VOLTAGE_SCALE) << VDDCI_SHIFT; + + if (POLARIS10_VOLTAGE_CONTROL_BY_GPIO == data->vddci_control) + vddci = (uint32_t)phm_find_closest_vddci(&(data->vddci_voltage_table), + mm_table->entries[count].vddc - VDDC_VDDCI_DELTA); + else if (POLARIS10_VOLTAGE_CONTROL_BY_SVID2 == data->vddci_control) + vddci = mm_table->entries[count].vddc - VDDC_VDDCI_DELTA; + else + vddci = (data->vbios_boot_state.vddci_bootup_value * VOLTAGE_SCALE) << VDDCI_SHIFT; + + table->SamuLevel[count].MinVoltage |= (vddci * VOLTAGE_SCALE) << VDDCI_SHIFT; table->SamuLevel[count].MinVoltage |= 1 << PHASES_SHIFT; /* retrieve divider value for VBIOS */ @@ -1660,6 +1715,7 @@ static int polaris10_populate_smc_uvd_level(struct pp_hwmgr *hwmgr, struct phm_ppt_v1_mm_clock_voltage_dependency_table *mm_table = table_info->mm_dep_table; struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend); + uint32_t vddci; table->UvdLevelCount = (uint8_t)(mm_table->count); table->UvdBootLevel = 0; @@ -1670,8 +1726,16 @@ static int polaris10_populate_smc_uvd_level(struct pp_hwmgr *hwmgr, table->UvdLevel[count].DclkFrequency = mm_table->entries[count].dclk; table->UvdLevel[count].MinVoltage |= (mm_table->entries[count].vddc * VOLTAGE_SCALE) << VDDC_SHIFT; - table->UvdLevel[count].MinVoltage |= ((mm_table->entries[count].vddc - - data->vddc_vddci_delta) * VOLTAGE_SCALE) << VDDCI_SHIFT; + + if (POLARIS10_VOLTAGE_CONTROL_BY_GPIO == data->vddci_control) + vddci = (uint32_t)phm_find_closest_vddci(&(data->vddci_voltage_table), + mm_table->entries[count].vddc - VDDC_VDDCI_DELTA); + else if (POLARIS10_VOLTAGE_CONTROL_BY_SVID2 == data->vddci_control) + vddci = mm_table->entries[count].vddc - VDDC_VDDCI_DELTA; + else + vddci = (data->vbios_boot_state.vddci_bootup_value * VOLTAGE_SCALE) << VDDCI_SHIFT; + + table->UvdLevel[count].MinVoltage |= (vddci * VOLTAGE_SCALE) << VDDCI_SHIFT; table->UvdLevel[count].MinVoltage |= 1 << PHASES_SHIFT; /* retrieve divider value for VBIOS */ @@ -1692,8 +1756,8 @@ static int polaris10_populate_smc_uvd_level(struct pp_hwmgr *hwmgr, CONVERT_FROM_HOST_TO_SMC_UL(table->UvdLevel[count].VclkFrequency); CONVERT_FROM_HOST_TO_SMC_UL(table->UvdLevel[count].DclkFrequency); CONVERT_FROM_HOST_TO_SMC_UL(table->UvdLevel[count].MinVoltage); - } + return result; } @@ -1761,12 +1825,9 @@ static int polaris10_populate_smc_initailial_state(struct pp_hwmgr *hwmgr) static int polaris10_populate_clock_stretcher_data_table(struct pp_hwmgr *hwmgr) { - uint32_t ro, efuse, efuse2, clock_freq, volt_without_cks, - volt_with_cks, value; - uint16_t clock_freq_u16; + uint32_t ro, efuse, volt_without_cks, volt_with_cks, value, max, min; struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend); - uint8_t type, i, j, cks_setting, stretch_amount, stretch_amount2, - volt_offset = 0; + uint8_t i, stretch_amount, volt_offset = 0; struct phm_ppt_v1_information *table_info = (struct phm_ppt_v1_information *)(hwmgr->pptable); struct phm_ppt_v1_clock_voltage_dependency_table *sclk_table = @@ -1778,56 +1839,47 @@ static int polaris10_populate_clock_stretcher_data_table(struct pp_hwmgr *hwmgr) * if the part is SS or FF. if RO >= 1660MHz, part is FF. */ efuse = cgs_read_ind_register(hwmgr->device, CGS_IND_REG__SMC, - ixSMU_EFUSE_0 + (146 * 4)); - efuse2 = cgs_read_ind_register(hwmgr->device, CGS_IND_REG__SMC, - ixSMU_EFUSE_0 + (148 * 4)); + ixSMU_EFUSE_0 + (67 * 4)); efuse &= 0xFF000000; efuse = efuse >> 24; - efuse2 &= 0xF; - if (efuse2 == 1) - ro = (2300 - 1350) * efuse / 255 + 1350; - else - ro = (2500 - 1000) * efuse / 255 + 1000; - - if (ro >= 1660) - type = 0; - else - type = 1; + if (hwmgr->chip_id == CHIP_POLARIS10) { + min = 1000; + max = 2300; + } else { + min = 1100; + max = 2100; + } - /* Populate Stretch amount */ - data->smc_state_table.ClockStretcherAmount = stretch_amount; + ro = efuse * (max -min)/255 + min; /* Populate Sclk_CKS_masterEn0_7 and Sclk_voltageOffset */ for (i = 0; i < sclk_table->count; i++) { data->smc_state_table.Sclk_CKS_masterEn0_7 |= sclk_table->entries[i].cks_enable << i; - volt_without_cks = (uint32_t)((14041 * - (sclk_table->entries[i].clk/100) / 10000 + 3571 + 75 - ro) * 1000 / - (4026 - (13924 * (sclk_table->entries[i].clk/100) / 10000))); - volt_with_cks = (uint32_t)((13946 * - (sclk_table->entries[i].clk/100) / 10000 + 3320 + 45 - ro) * 1000 / - (3664 - (11454 * (sclk_table->entries[i].clk/100) / 10000))); + if (hwmgr->chip_id == CHIP_POLARIS10) { + volt_without_cks = (uint32_t)((2753594000U + (sclk_table->entries[i].clk/100) * 136418 -(ro - 70) * 1000000) / \ + (2424180 - (sclk_table->entries[i].clk/100) * 1132925/1000)); + volt_with_cks = (uint32_t)((2797202000U + sclk_table->entries[i].clk/100 * 3232 - (ro - 65) * 1000000) / \ + (2522480 - sclk_table->entries[i].clk/100 * 115764/100)); + } else { + volt_without_cks = (uint32_t)((2416794800U + (sclk_table->entries[i].clk/100) * 1476925/10 -(ro - 50) * 1000000) / \ + (2625416 - (sclk_table->entries[i].clk/100) * (12586807/10000))); + volt_with_cks = (uint32_t)((2999656000U - sclk_table->entries[i].clk/100 * 392803 - (ro - 44) * 1000000) / \ + (3422454 - sclk_table->entries[i].clk/100 * (18886376/10000))); + } + if (volt_without_cks >= volt_with_cks) volt_offset = (uint8_t)(((volt_without_cks - volt_with_cks + - sclk_table->entries[i].cks_voffset) * 100 / 625) + 1); + sclk_table->entries[i].cks_voffset) * 100 + 624) / 625); + data->smc_state_table.Sclk_voltageOffset[i] = volt_offset; } - PHM_WRITE_INDIRECT_FIELD(hwmgr->device, CGS_IND_REG__SMC, PWR_CKS_ENABLE, - STRETCH_ENABLE, 0x0); - PHM_WRITE_INDIRECT_FIELD(hwmgr->device, CGS_IND_REG__SMC, PWR_CKS_ENABLE, - masterReset, 0x1); - /* PHM_WRITE_INDIRECT_FIELD(hwmgr->device, CGS_IND_REG__SMC, PWR_CKS_ENABLE, staticEnable, 0x1); */ - PHM_WRITE_INDIRECT_FIELD(hwmgr->device, CGS_IND_REG__SMC, PWR_CKS_ENABLE, - masterReset, 0x0); - + data->smc_state_table.LdoRefSel = (table_info->cac_dtp_table->ucCKS_LDO_REFSEL != 0) ? table_info->cac_dtp_table->ucCKS_LDO_REFSEL : 6; /* Populate CKS Lookup Table */ - if (stretch_amount == 1 || stretch_amount == 2 || stretch_amount == 5) - stretch_amount2 = 0; - else if (stretch_amount == 3 || stretch_amount == 4) - stretch_amount2 = 1; - else { + if (stretch_amount != 1 && stretch_amount != 2 && stretch_amount != 3 && + stretch_amount != 4 && stretch_amount != 5) { phm_cap_unset(hwmgr->platform_descriptor.platformCaps, PHM_PlatformCaps_ClockStretcher); PP_ASSERT_WITH_CODE(false, @@ -1835,69 +1887,6 @@ static int polaris10_populate_clock_stretcher_data_table(struct pp_hwmgr *hwmgr) return -EINVAL); } - value = cgs_read_ind_register(hwmgr->device, CGS_IND_REG__SMC, - ixPWR_CKS_CNTL); - value &= 0xFFC2FF87; - data->smc_state_table.CKS_LOOKUPTable.CKS_LOOKUPTableEntry[0].minFreq = - polaris10_clock_stretcher_lookup_table[stretch_amount2][0]; - data->smc_state_table.CKS_LOOKUPTable.CKS_LOOKUPTableEntry[0].maxFreq = - polaris10_clock_stretcher_lookup_table[stretch_amount2][1]; - clock_freq_u16 = (uint16_t)(PP_SMC_TO_HOST_UL(data->smc_state_table. - GraphicsLevel[data->smc_state_table.GraphicsDpmLevelCount - 1].SclkSetting.SclkFrequency) / 100); - if (polaris10_clock_stretcher_lookup_table[stretch_amount2][0] < clock_freq_u16 - && polaris10_clock_stretcher_lookup_table[stretch_amount2][1] > clock_freq_u16) { - /* Program PWR_CKS_CNTL. CKS_USE_FOR_LOW_FREQ */ - value |= (polaris10_clock_stretcher_lookup_table[stretch_amount2][3]) << 16; - /* Program PWR_CKS_CNTL. CKS_LDO_REFSEL */ - value |= (polaris10_clock_stretcher_lookup_table[stretch_amount2][2]) << 18; - /* Program PWR_CKS_CNTL. CKS_STRETCH_AMOUNT */ - value |= (polaris10_clock_stretch_amount_conversion - [polaris10_clock_stretcher_lookup_table[stretch_amount2][3]] - [stretch_amount]) << 3; - } - CONVERT_FROM_HOST_TO_SMC_US(data->smc_state_table.CKS_LOOKUPTable.CKS_LOOKUPTableEntry[0].minFreq); - CONVERT_FROM_HOST_TO_SMC_US(data->smc_state_table.CKS_LOOKUPTable.CKS_LOOKUPTableEntry[0].maxFreq); - data->smc_state_table.CKS_LOOKUPTable.CKS_LOOKUPTableEntry[0].setting = - polaris10_clock_stretcher_lookup_table[stretch_amount2][2] & 0x7F; - data->smc_state_table.CKS_LOOKUPTable.CKS_LOOKUPTableEntry[0].setting |= - (polaris10_clock_stretcher_lookup_table[stretch_amount2][3]) << 7; - - cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC, - ixPWR_CKS_CNTL, value); - - /* Populate DDT Lookup Table */ - for (i = 0; i < 4; i++) { - /* Assign the minimum and maximum VID stored - * in the last row of Clock Stretcher Voltage Table. - */ - data->smc_state_table.ClockStretcherDataTable.ClockStretcherDataTableEntry[i].minVID = - (uint8_t) polaris10_clock_stretcher_ddt_table[type][i][2]; - data->smc_state_table.ClockStretcherDataTable.ClockStretcherDataTableEntry[i].maxVID = - (uint8_t) polaris10_clock_stretcher_ddt_table[type][i][3]; - /* Loop through each SCLK and check the frequency - * to see if it lies within the frequency for clock stretcher. - */ - for (j = 0; j < data->smc_state_table.GraphicsDpmLevelCount; j++) { - cks_setting = 0; - clock_freq = PP_SMC_TO_HOST_UL( - data->smc_state_table.GraphicsLevel[j].SclkSetting.SclkFrequency); - /* Check the allowed frequency against the sclk level[j]. - * Sclk's endianness has already been converted, - * and it's in 10Khz unit, - * as opposed to Data table, which is in Mhz unit. - */ - if (clock_freq >= (polaris10_clock_stretcher_ddt_table[type][i][0]) * 100) { - cks_setting |= 0x2; - if (clock_freq < (polaris10_clock_stretcher_ddt_table[type][i][1]) * 100) - cks_setting |= 0x1; - } - data->smc_state_table.ClockStretcherDataTable.ClockStretcherDataTableEntry[i].setting - |= cks_setting << (j * 2); - } - CONVERT_FROM_HOST_TO_SMC_US( - data->smc_state_table.ClockStretcherDataTable.ClockStretcherDataTableEntry[i].setting); - } - value = cgs_read_ind_register(hwmgr->device, CGS_IND_REG__SMC, ixPWR_CKS_CNTL); value &= 0xFFFFFFFE; cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC, ixPWR_CKS_CNTL, value); @@ -1945,9 +1934,8 @@ static int polaris10_populate_vr_config(struct pp_hwmgr *hwmgr, if (POLARIS10_VOLTAGE_CONTROL_BY_SVID2 == data->mvdd_control) { config = VR_SVI2_PLANE_2; table->VRConfig |= (config << VRCONF_MVDD_SHIFT); - } else if (POLARIS10_VOLTAGE_CONTROL_BY_GPIO == data->mvdd_control) { - config = VR_SMIO_PATTERN_2; - table->VRConfig |= (config << VRCONF_MVDD_SHIFT); + cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC, data->soft_regs_start + + offsetof(SMU74_SoftRegisters, AllowMvddSwitch), 0x1); } else { config = VR_STATIC_VOLTAGE; table->VRConfig |= (config << VRCONF_MVDD_SHIFT); @@ -1956,6 +1944,90 @@ static int polaris10_populate_vr_config(struct pp_hwmgr *hwmgr, return 0; } + +int polaris10_populate_avfs_parameters(struct pp_hwmgr *hwmgr) +{ + struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend); + SMU74_Discrete_DpmTable *table = &(data->smc_state_table); + int result = 0; + struct pp_atom_ctrl__avfs_parameters avfs_params = {0}; + AVFS_meanNsigma_t AVFS_meanNsigma = { {0} }; + AVFS_Sclk_Offset_t AVFS_SclkOffset = { {0} }; + uint32_t tmp, i; + struct pp_smumgr *smumgr = hwmgr->smumgr; + struct polaris10_smumgr *smu_data = (struct polaris10_smumgr *)(smumgr->backend); + + struct phm_ppt_v1_information *table_info = + (struct phm_ppt_v1_information *)hwmgr->pptable; + struct phm_ppt_v1_clock_voltage_dependency_table *sclk_table = + table_info->vdd_dep_on_sclk; + + + if (smu_data->avfs.avfs_btc_status == AVFS_BTC_NOTSUPPORTED) + return result; + + result = atomctrl_get_avfs_information(hwmgr, &avfs_params); + + if (0 == result) { + table->BTCGB_VDROOP_TABLE[0].a0 = PP_HOST_TO_SMC_UL(avfs_params.ulGB_VDROOP_TABLE_CKSON_a0); + table->BTCGB_VDROOP_TABLE[0].a1 = PP_HOST_TO_SMC_UL(avfs_params.ulGB_VDROOP_TABLE_CKSON_a1); + table->BTCGB_VDROOP_TABLE[0].a2 = PP_HOST_TO_SMC_UL(avfs_params.ulGB_VDROOP_TABLE_CKSON_a2); + table->BTCGB_VDROOP_TABLE[1].a0 = PP_HOST_TO_SMC_UL(avfs_params.ulGB_VDROOP_TABLE_CKSOFF_a0); + table->BTCGB_VDROOP_TABLE[1].a1 = PP_HOST_TO_SMC_UL(avfs_params.ulGB_VDROOP_TABLE_CKSOFF_a1); + table->BTCGB_VDROOP_TABLE[1].a2 = PP_HOST_TO_SMC_UL(avfs_params.ulGB_VDROOP_TABLE_CKSOFF_a2); + table->AVFSGB_VDROOP_TABLE[0].m1 = PP_HOST_TO_SMC_UL(avfs_params.ulAVFSGB_FUSE_TABLE_CKSON_m1); + table->AVFSGB_VDROOP_TABLE[0].m2 = PP_HOST_TO_SMC_US(avfs_params.usAVFSGB_FUSE_TABLE_CKSON_m2); + table->AVFSGB_VDROOP_TABLE[0].b = PP_HOST_TO_SMC_UL(avfs_params.ulAVFSGB_FUSE_TABLE_CKSON_b); + table->AVFSGB_VDROOP_TABLE[0].m1_shift = 24; + table->AVFSGB_VDROOP_TABLE[0].m2_shift = 12; + table->AVFSGB_VDROOP_TABLE[1].m1 = PP_HOST_TO_SMC_UL(avfs_params.ulAVFSGB_FUSE_TABLE_CKSOFF_m1); + table->AVFSGB_VDROOP_TABLE[1].m2 = PP_HOST_TO_SMC_US(avfs_params.usAVFSGB_FUSE_TABLE_CKSOFF_m2); + table->AVFSGB_VDROOP_TABLE[1].b = PP_HOST_TO_SMC_UL(avfs_params.ulAVFSGB_FUSE_TABLE_CKSOFF_b); + table->AVFSGB_VDROOP_TABLE[1].m1_shift = 24; + table->AVFSGB_VDROOP_TABLE[1].m2_shift = 12; + table->MaxVoltage = PP_HOST_TO_SMC_US(avfs_params.usMaxVoltage_0_25mv); + AVFS_meanNsigma.Aconstant[0] = PP_HOST_TO_SMC_UL(avfs_params.ulAVFS_meanNsigma_Acontant0); + AVFS_meanNsigma.Aconstant[1] = PP_HOST_TO_SMC_UL(avfs_params.ulAVFS_meanNsigma_Acontant1); + AVFS_meanNsigma.Aconstant[2] = PP_HOST_TO_SMC_UL(avfs_params.ulAVFS_meanNsigma_Acontant2); + AVFS_meanNsigma.DC_tol_sigma = PP_HOST_TO_SMC_US(avfs_params.usAVFS_meanNsigma_DC_tol_sigma); + AVFS_meanNsigma.Platform_mean = PP_HOST_TO_SMC_US(avfs_params.usAVFS_meanNsigma_Platform_mean); + AVFS_meanNsigma.PSM_Age_CompFactor = PP_HOST_TO_SMC_US(avfs_params.usPSM_Age_ComFactor); + AVFS_meanNsigma.Platform_sigma = PP_HOST_TO_SMC_US(avfs_params.usAVFS_meanNsigma_Platform_sigma); + + for (i = 0; i < NUM_VFT_COLUMNS; i++) { + AVFS_meanNsigma.Static_Voltage_Offset[i] = (uint8_t)(sclk_table->entries[i].cks_voffset * 100 / 625); + AVFS_SclkOffset.Sclk_Offset[i] = PP_HOST_TO_SMC_US((uint16_t)(sclk_table->entries[i].sclk_offset) / 100); + } + + result = polaris10_read_smc_sram_dword(smumgr, + SMU7_FIRMWARE_HEADER_LOCATION + offsetof(SMU74_Firmware_Header, AvfsMeanNSigma), + &tmp, data->sram_end); + + polaris10_copy_bytes_to_smc(smumgr, + tmp, + (uint8_t *)&AVFS_meanNsigma, + sizeof(AVFS_meanNsigma_t), + data->sram_end); + + result = polaris10_read_smc_sram_dword(smumgr, + SMU7_FIRMWARE_HEADER_LOCATION + offsetof(SMU74_Firmware_Header, AvfsSclkOffsetTable), + &tmp, data->sram_end); + polaris10_copy_bytes_to_smc(smumgr, + tmp, + (uint8_t *)&AVFS_SclkOffset, + sizeof(AVFS_Sclk_Offset_t), + data->sram_end); + + data->avfs_vdroop_override_setting = (avfs_params.ucEnableGB_VDROOP_TABLE_CKSON << BTCGB0_Vdroop_Enable_SHIFT) | + (avfs_params.ucEnableGB_VDROOP_TABLE_CKSOFF << BTCGB1_Vdroop_Enable_SHIFT) | + (avfs_params.ucEnableGB_FUSE_TABLE_CKSON << AVFSGB0_Vdroop_Enable_SHIFT) | + (avfs_params.ucEnableGB_FUSE_TABLE_CKSOFF << AVFSGB1_Vdroop_Enable_SHIFT); + data->apply_avfs_cks_off_voltage = (avfs_params.ucEnableApplyAVFS_CKS_OFF_Voltage == 1) ? true : false; + } + return result; +} + + /** * Initializes the SMC table and uploads it * @@ -2056,6 +2128,10 @@ static int polaris10_init_smc_table(struct pp_hwmgr *hwmgr) "Failed to populate Clock Stretcher Data Table!", return result); } + + result = polaris10_populate_avfs_parameters(hwmgr); + PP_ASSERT_WITH_CODE(0 == result, "Failed to populate AVFS Parameters!", return result;); + table->CurrSclkPllRange = 0xff; table->GraphicsVoltageChangeEnable = 1; table->GraphicsThermThrottleEnable = 1; @@ -2229,6 +2305,17 @@ static int polaris10_enable_ulv(struct pp_hwmgr *hwmgr) return 0; } +static int polaris10_disable_ulv(struct pp_hwmgr *hwmgr) +{ + struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend); + struct polaris10_ulv_parm *ulv = &(data->ulv); + + if (ulv->ulv_supported) + return smum_send_msg_to_smc(hwmgr->smumgr, PPSMC_MSG_DisableULV); + + return 0; +} + static int polaris10_enable_deep_sleep_master_switch(struct pp_hwmgr *hwmgr) { if (phm_cap_enabled(hwmgr->platform_descriptor.platformCaps, @@ -2249,9 +2336,27 @@ static int polaris10_enable_deep_sleep_master_switch(struct pp_hwmgr *hwmgr) return 0; } +static int polaris10_disable_deep_sleep_master_switch(struct pp_hwmgr *hwmgr) +{ + if (phm_cap_enabled(hwmgr->platform_descriptor.platformCaps, + PHM_PlatformCaps_SclkDeepSleep)) { + if (smum_send_msg_to_smc(hwmgr->smumgr, + PPSMC_MSG_MASTER_DeepSleep_OFF)) { + PP_ASSERT_WITH_CODE(false, + "Attempt to disable Master Deep Sleep switch failed!", + return -1); + } + } + + return 0; +} + static int polaris10_enable_sclk_mclk_dpm(struct pp_hwmgr *hwmgr) { struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend); + uint32_t soft_register_value = 0; + uint32_t handshake_disables_offset = data->soft_regs_start + + offsetof(SMU74_SoftRegisters, HandshakeDisables); /* enable SCLK dpm */ if (!data->sclk_dpm_key_disabled) @@ -2262,6 +2367,12 @@ static int polaris10_enable_sclk_mclk_dpm(struct pp_hwmgr *hwmgr) /* enable MCLK dpm */ if (0 == data->mclk_dpm_key_disabled) { +/* Disable UVD - SMU handshake for MCLK. */ + soft_register_value = cgs_read_ind_register(hwmgr->device, + CGS_IND_REG__SMC, handshake_disables_offset); + soft_register_value |= SMU7_UVD_MCLK_HANDSHAKE_DISABLE; + cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC, + handshake_disables_offset, soft_register_value); PP_ASSERT_WITH_CODE( (0 == smum_send_msg_to_smc(hwmgr->smumgr, @@ -2269,7 +2380,6 @@ static int polaris10_enable_sclk_mclk_dpm(struct pp_hwmgr *hwmgr) "Failed to enable MCLK DPM during DPM Start Function!", return -1); - PHM_WRITE_FIELD(hwmgr->device, MC_SEQ_CNTL_3, CAC_EN, 0x1); cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC, ixLCAC_MC0_CNTL, 0x5); @@ -2338,6 +2448,58 @@ static int polaris10_start_dpm(struct pp_hwmgr *hwmgr) return 0; } +static int polaris10_disable_sclk_mclk_dpm(struct pp_hwmgr *hwmgr) +{ + struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend); + + /* disable SCLK dpm */ + if (!data->sclk_dpm_key_disabled) + PP_ASSERT_WITH_CODE( + (smum_send_msg_to_smc(hwmgr->smumgr, + PPSMC_MSG_DPM_Disable) == 0), + "Failed to disable SCLK DPM!", + return -1); + + /* disable MCLK dpm */ + if (!data->mclk_dpm_key_disabled) { + PP_ASSERT_WITH_CODE( + (smum_send_msg_to_smc(hwmgr->smumgr, + PPSMC_MSG_MCLKDPM_Disable) == 0), + "Failed to disable MCLK DPM!", + return -1); + } + + return 0; +} + +static int polaris10_stop_dpm(struct pp_hwmgr *hwmgr) +{ + struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend); + + /* disable general power management */ + PHM_WRITE_INDIRECT_FIELD(hwmgr->device, CGS_IND_REG__SMC, GENERAL_PWRMGT, + GLOBAL_PWRMGT_EN, 0); + /* disable sclk deep sleep */ + PHM_WRITE_INDIRECT_FIELD(hwmgr->device, CGS_IND_REG__SMC, SCLK_PWRMGT_CNTL, + DYNAMIC_PM_EN, 0); + + /* disable PCIE dpm */ + if (!data->pcie_dpm_key_disabled) { + PP_ASSERT_WITH_CODE( + (smum_send_msg_to_smc(hwmgr->smumgr, + PPSMC_MSG_PCIeDPM_Disable) == 0), + "Failed to disable pcie DPM during DPM Stop Function!", + return -1); + } + + if (polaris10_disable_sclk_mclk_dpm(hwmgr)) { + printk(KERN_ERR "Failed to disable Sclk DPM and Mclk DPM!"); + return -1; + } + + return 0; +} + static void polaris10_set_dpm_event_sources(struct pp_hwmgr *hwmgr, uint32_t sources) { bool protection; @@ -2395,6 +2557,23 @@ static int polaris10_enable_thermal_auto_throttle(struct pp_hwmgr *hwmgr) return polaris10_enable_auto_throttle_source(hwmgr, PHM_AutoThrottleSource_Thermal); } +static int polaris10_disable_auto_throttle_source(struct pp_hwmgr *hwmgr, + PHM_AutoThrottleSource source) +{ + struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend); + + if (data->active_auto_throttle_sources & (1 << source)) { + data->active_auto_throttle_sources &= ~(1 << source); + polaris10_set_dpm_event_sources(hwmgr, data->active_auto_throttle_sources); + } + return 0; +} + +static int polaris10_disable_thermal_auto_throttle(struct pp_hwmgr *hwmgr) +{ + return polaris10_disable_auto_throttle_source(hwmgr, PHM_AutoThrottleSource_Thermal); +} + int polaris10_pcie_performance_request(struct pp_hwmgr *hwmgr) { struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend); @@ -2471,6 +2650,8 @@ int polaris10_enable_dpm_tasks(struct pp_hwmgr *hwmgr) PP_ASSERT_WITH_CODE((0 == tmp_result), "Failed to enable VR hot GPIO interrupt!", result = tmp_result); + smum_send_msg_to_smc(hwmgr->smumgr, (PPSMC_Msg)PPSMC_HasDisplay); + tmp_result = polaris10_enable_sclk_control(hwmgr); PP_ASSERT_WITH_CODE((0 == tmp_result), "Failed to enable SCLK control!", result = tmp_result); @@ -2487,6 +2668,10 @@ int polaris10_enable_dpm_tasks(struct pp_hwmgr *hwmgr) PP_ASSERT_WITH_CODE((0 == tmp_result), "Failed to enable deep sleep master switch!", result = tmp_result); + tmp_result = polaris10_enable_didt_config(hwmgr); + PP_ASSERT_WITH_CODE((tmp_result == 0), + "Failed to enable deep sleep master switch!", result = tmp_result); + tmp_result = polaris10_start_dpm(hwmgr); PP_ASSERT_WITH_CODE((0 == tmp_result), "Failed to start DPM!", result = tmp_result); @@ -2516,8 +2701,60 @@ int polaris10_enable_dpm_tasks(struct pp_hwmgr *hwmgr) int polaris10_disable_dpm_tasks(struct pp_hwmgr *hwmgr) { + int tmp_result, result = 0; - return 0; + tmp_result = (polaris10_is_dpm_running(hwmgr)) ? 0 : -1; + PP_ASSERT_WITH_CODE(tmp_result == 0, + "DPM is not running right now, no need to disable DPM!", + return 0); + + if (phm_cap_enabled(hwmgr->platform_descriptor.platformCaps, + PHM_PlatformCaps_ThermalController)) + PHM_WRITE_INDIRECT_FIELD(hwmgr->device, CGS_IND_REG__SMC, + GENERAL_PWRMGT, THERMAL_PROTECTION_DIS, 1); + + tmp_result = polaris10_disable_power_containment(hwmgr); + PP_ASSERT_WITH_CODE((tmp_result == 0), + "Failed to disable power containment!", result = tmp_result); + + tmp_result = polaris10_disable_smc_cac(hwmgr); + PP_ASSERT_WITH_CODE((tmp_result == 0), + "Failed to disable SMC CAC!", result = tmp_result); + + PHM_WRITE_INDIRECT_FIELD(hwmgr->device, CGS_IND_REG__SMC, + CG_SPLL_SPREAD_SPECTRUM, SSEN, 0); + PHM_WRITE_INDIRECT_FIELD(hwmgr->device, CGS_IND_REG__SMC, + GENERAL_PWRMGT, DYN_SPREAD_SPECTRUM_EN, 0); + + tmp_result = polaris10_disable_thermal_auto_throttle(hwmgr); + PP_ASSERT_WITH_CODE((tmp_result == 0), + "Failed to disable thermal auto throttle!", result = tmp_result); + + tmp_result = polaris10_stop_dpm(hwmgr); + PP_ASSERT_WITH_CODE((tmp_result == 0), + "Failed to stop DPM!", result = tmp_result); + + tmp_result = polaris10_disable_deep_sleep_master_switch(hwmgr); + PP_ASSERT_WITH_CODE((tmp_result == 0), + "Failed to disable deep sleep master switch!", result = tmp_result); + + tmp_result = polaris10_disable_ulv(hwmgr); + PP_ASSERT_WITH_CODE((tmp_result == 0), + "Failed to disable ULV!", result = tmp_result); + + tmp_result = polaris10_clear_voting_clients(hwmgr); + PP_ASSERT_WITH_CODE((tmp_result == 0), + "Failed to clear voting clients!", result = tmp_result); + + tmp_result = polaris10_reset_to_default(hwmgr); + PP_ASSERT_WITH_CODE((tmp_result == 0), + "Failed to reset to default!", result = tmp_result); + + tmp_result = polaris10_force_switch_to_arbf0(hwmgr); + PP_ASSERT_WITH_CODE((tmp_result == 0), + "Failed to force to switch arbf0!", result = tmp_result); + + return result; } int polaris10_reset_asic_tasks(struct pp_hwmgr *hwmgr) @@ -2528,13 +2765,6 @@ int polaris10_reset_asic_tasks(struct pp_hwmgr *hwmgr) int polaris10_hwmgr_backend_fini(struct pp_hwmgr *hwmgr) { - struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend); - - if (data->soft_pp_table) { - kfree(data->soft_pp_table); - data->soft_pp_table = NULL; - } - return phm_hwmgr_backend_fini(hwmgr); } @@ -2581,17 +2811,22 @@ int polaris10_set_features_platform_caps(struct pp_hwmgr *hwmgr) PHM_PlatformCaps_DynamicUVDState); /* power tune caps Assume disabled */ - phm_cap_unset(hwmgr->platform_descriptor.platformCaps, + phm_cap_set(hwmgr->platform_descriptor.platformCaps, PHM_PlatformCaps_SQRamping); - phm_cap_unset(hwmgr->platform_descriptor.platformCaps, + phm_cap_set(hwmgr->platform_descriptor.platformCaps, PHM_PlatformCaps_DBRamping); - phm_cap_unset(hwmgr->platform_descriptor.platformCaps, + phm_cap_set(hwmgr->platform_descriptor.platformCaps, PHM_PlatformCaps_TDRamping); - phm_cap_unset(hwmgr->platform_descriptor.platformCaps, + phm_cap_set(hwmgr->platform_descriptor.platformCaps, PHM_PlatformCaps_TCPRamping); - phm_cap_set(hwmgr->platform_descriptor.platformCaps, - PHM_PlatformCaps_PowerContainment); + if (hwmgr->powercontainment_enabled) + phm_cap_set(hwmgr->platform_descriptor.platformCaps, + PHM_PlatformCaps_PowerContainment); + else + phm_cap_unset(hwmgr->platform_descriptor.platformCaps, + PHM_PlatformCaps_PowerContainment); + phm_cap_set(hwmgr->platform_descriptor.platformCaps, PHM_PlatformCaps_CAC); @@ -2606,6 +2841,7 @@ int polaris10_set_features_platform_caps(struct pp_hwmgr *hwmgr) phm_cap_set(hwmgr->platform_descriptor.platformCaps, PHM_PlatformCaps_FanSpeedInTableIsRPM); + if (hwmgr->chip_id == CHIP_POLARIS11) phm_cap_set(hwmgr->platform_descriptor.platformCaps, PHM_PlatformCaps_SPLLShutdownSupport); @@ -2638,7 +2874,7 @@ static int polaris10_get_evv_voltages(struct pp_hwmgr *hwmgr) { struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend); uint16_t vv_id; - uint16_t vddc = 0; + uint32_t vddc = 0; uint16_t i, j; uint32_t sclk = 0; struct phm_ppt_v1_information *table_info = @@ -2662,15 +2898,16 @@ static int polaris10_get_evv_voltages(struct pp_hwmgr *hwmgr) } } + if (atomctrl_get_voltage_evv_on_sclk_ai(hwmgr, + VOLTAGE_TYPE_VDDC, + sclk, vv_id, &vddc) != 0) { + printk(KERN_WARNING "failed to retrieving EVV voltage!\n"); + continue; + } - PP_ASSERT_WITH_CODE(0 == atomctrl_get_voltage_evv_on_sclk_ai(hwmgr, - VOLTAGE_TYPE_VDDC, sclk, vv_id, &vddc), - "Error retrieving EVV voltage value!", - continue); - - - /* need to make sure vddc is less than 2v or else, it could burn the ASIC. */ - PP_ASSERT_WITH_CODE((vddc < 2000 && vddc != 0), + /* need to make sure vddc is less than 2v or else, it could burn the ASIC. + * real voltage level in unit of 0.01mv */ + PP_ASSERT_WITH_CODE((vddc < 200000 && vddc != 0), "Invalid VDDC value", result = -EINVAL;); /* the voltage should not be zero nor equal to leakage ID */ @@ -2896,15 +3133,46 @@ static int polaris10_set_private_data_based_on_pptable(struct pp_hwmgr *hwmgr) return 0; } +int polaris10_patch_voltage_workaround(struct pp_hwmgr *hwmgr) +{ + struct phm_ppt_v1_information *table_info = + (struct phm_ppt_v1_information *)(hwmgr->pptable); + struct phm_ppt_v1_clock_voltage_dependency_table *dep_mclk_table = + table_info->vdd_dep_on_mclk; + struct phm_ppt_v1_voltage_lookup_table *lookup_table = + table_info->vddc_lookup_table; + uint32_t i; + + if (hwmgr->chip_id == CHIP_POLARIS10 && hwmgr->hw_revision == 0xC7) { + if (lookup_table->entries[dep_mclk_table->entries[dep_mclk_table->count-1].vddInd].us_vdd >= 1000) + return 0; + + for (i = 0; i < lookup_table->count; i++) { + if (lookup_table->entries[i].us_vdd < 0xff01 && lookup_table->entries[i].us_vdd >= 1000) { + dep_mclk_table->entries[dep_mclk_table->count-1].vddInd = (uint8_t) i; + return 0; + } + } + } + return 0; +} + + int polaris10_hwmgr_backend_init(struct pp_hwmgr *hwmgr) { - struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend); + struct polaris10_hwmgr *data; struct pp_atomctrl_gpio_pin_assignment gpio_pin_assignment; uint32_t temp_reg; int result; struct phm_ppt_v1_information *table_info = (struct phm_ppt_v1_information *)(hwmgr->pptable); + data = kzalloc(sizeof(struct polaris10_hwmgr), GFP_KERNEL); + if (data == NULL) + return -ENOMEM; + + hwmgr->backend = data; + data->dll_default_on = false; data->sram_end = SMC_RAM_END; data->mclk_dpm0_activity_target = 0xa; @@ -2938,6 +3206,11 @@ int polaris10_hwmgr_backend_init(struct pp_hwmgr *hwmgr) data->vddci_control = POLARIS10_VOLTAGE_CONTROL_NONE; data->mvdd_control = POLARIS10_VOLTAGE_CONTROL_NONE; + data->enable_tdc_limit_feature = true; + data->enable_pkg_pwr_tracking_feature = true; + data->force_pcie_gen = PP_PCIEGenInvalid; + data->mclk_stutter_mode_threshold = 40000; + if (atomctrl_is_voltage_controled_by_gpio_v3(hwmgr, VOLTAGE_TYPE_VDDC, VOLTAGE_OBJ_SVID2)) data->voltage_control = POLARIS10_VOLTAGE_CONTROL_BY_SVID2; @@ -2962,8 +3235,13 @@ int polaris10_hwmgr_backend_init(struct pp_hwmgr *hwmgr) data->vddci_control = POLARIS10_VOLTAGE_CONTROL_BY_SVID2; } + if (table_info->cac_dtp_table->usClockStretchAmount != 0) + phm_cap_set(hwmgr->platform_descriptor.platformCaps, + PHM_PlatformCaps_ClockStretcher); + polaris10_set_features_platform_caps(hwmgr); + polaris10_patch_voltage_workaround(hwmgr); polaris10_init_dpm_defaults(hwmgr); /* Get leakage voltage based on leakage ID. */ @@ -2983,7 +3261,7 @@ int polaris10_hwmgr_backend_init(struct pp_hwmgr *hwmgr) if (0 == result) { struct cgs_system_info sys_info = {0}; - data->is_tlu_enabled = 0; + data->is_tlu_enabled = false; hwmgr->platform_descriptor.hardwareActivityPerformanceLevels = POLARIS10_MAX_HARDWARE_POWERLEVELS; @@ -3068,7 +3346,7 @@ int polaris10_hwmgr_backend_init(struct pp_hwmgr *hwmgr) sys_info.info_id = CGS_SYSTEM_INFO_PCIE_GEN_INFO; result = cgs_query_system_info(hwmgr->device, &sys_info); if (result) - data->pcie_gen_cap = 0x30007; + data->pcie_gen_cap = AMDGPU_DEFAULT_PCIE_GEN_MASK; else data->pcie_gen_cap = (uint32_t)sys_info.value; if (data->pcie_gen_cap & CAIL_PCIE_LINK_SPEED_SUPPORT_GEN3) @@ -3077,7 +3355,7 @@ int polaris10_hwmgr_backend_init(struct pp_hwmgr *hwmgr) sys_info.info_id = CGS_SYSTEM_INFO_PCIE_MLW; result = cgs_query_system_info(hwmgr->device, &sys_info); if (result) - data->pcie_lane_cap = 0x2f0000; + data->pcie_lane_cap = AMDGPU_DEFAULT_PCIE_MLW_MASK; else data->pcie_lane_cap = (uint32_t)sys_info.value; @@ -3366,6 +3644,7 @@ static int polaris10_apply_state_adjust_rules(struct pp_hwmgr *hwmgr, hwmgr->platform_descriptor.platformCaps, PHM_PlatformCaps_DisableMclkSwitchingForFrameLock); + disable_mclk_switching = (1 < info.display_count) || disable_mclk_switching_for_frame_lock; @@ -3520,10 +3799,11 @@ static int polaris10_get_pp_table_entry_callback_func(struct pp_hwmgr *hwmgr, ATOM_Tonga_State *state_entry = (ATOM_Tonga_State *)state; ATOM_Tonga_POWERPLAYTABLE *powerplay_table = (ATOM_Tonga_POWERPLAYTABLE *)pp_table; - ATOM_Tonga_SCLK_Dependency_Table *sclk_dep_table = - (ATOM_Tonga_SCLK_Dependency_Table *) + PPTable_Generic_SubTable_Header *sclk_dep_table = + (PPTable_Generic_SubTable_Header *) (((unsigned long)powerplay_table) + le16_to_cpu(powerplay_table->usSclkDependencyTableOffset)); + ATOM_Tonga_MCLK_Dependency_Table *mclk_dep_table = (ATOM_Tonga_MCLK_Dependency_Table *) (((unsigned long)powerplay_table) + @@ -3575,7 +3855,11 @@ static int polaris10_get_pp_table_entry_callback_func(struct pp_hwmgr *hwmgr, /* Performance levels are arranged from low to high. */ performance_level->memory_clock = mclk_dep_table->entries [state_entry->ucMemoryClockIndexLow].ulMclk; - performance_level->engine_clock = sclk_dep_table->entries + if (sclk_dep_table->ucRevId == 0) + performance_level->engine_clock = ((ATOM_Tonga_SCLK_Dependency_Table *)sclk_dep_table)->entries + [state_entry->ucEngineClockIndexLow].ulSclk; + else if (sclk_dep_table->ucRevId == 1) + performance_level->engine_clock = ((ATOM_Polaris_SCLK_Dependency_Table *)sclk_dep_table)->entries [state_entry->ucEngineClockIndexLow].ulSclk; performance_level->pcie_gen = get_pcie_gen_support(data->pcie_gen_cap, state_entry->ucPCIEGenLow); @@ -3586,8 +3870,14 @@ static int polaris10_get_pp_table_entry_callback_func(struct pp_hwmgr *hwmgr, [polaris10_power_state->performance_level_count++]); performance_level->memory_clock = mclk_dep_table->entries [state_entry->ucMemoryClockIndexHigh].ulMclk; - performance_level->engine_clock = sclk_dep_table->entries + + if (sclk_dep_table->ucRevId == 0) + performance_level->engine_clock = ((ATOM_Tonga_SCLK_Dependency_Table *)sclk_dep_table)->entries [state_entry->ucEngineClockIndexHigh].ulSclk; + else if (sclk_dep_table->ucRevId == 1) + performance_level->engine_clock = ((ATOM_Polaris_SCLK_Dependency_Table *)sclk_dep_table)->entries + [state_entry->ucEngineClockIndexHigh].ulSclk; + performance_level->pcie_gen = get_pcie_gen_support(data->pcie_gen_cap, state_entry->ucPCIEGenHigh); performance_level->pcie_lane = get_pcie_lane_support(data->pcie_lane_cap, @@ -3645,7 +3935,6 @@ static int polaris10_get_pp_table_entry(struct pp_hwmgr *hwmgr, switch (state->classification.ui_label) { case PP_StateUILabel_Performance: data->use_pcie_performance_levels = true; - for (i = 0; i < ps->performance_level_count; i++) { if (data->pcie_gen_performance.max < ps->performance_levels[i].pcie_gen) @@ -3661,7 +3950,6 @@ static int polaris10_get_pp_table_entry(struct pp_hwmgr *hwmgr, ps->performance_levels[i].pcie_lane) data->pcie_lane_performance.max = ps->performance_levels[i].pcie_lane; - if (data->pcie_lane_performance.min > ps->performance_levels[i].pcie_lane) data->pcie_lane_performance.min = @@ -3861,8 +4149,8 @@ static int polaris10_freeze_sclk_mclk_dpm(struct pp_hwmgr *hwmgr) if ((0 == data->sclk_dpm_key_disabled) && (data->need_update_smu7_dpm_table & (DPMTABLE_OD_UPDATE_SCLK + DPMTABLE_UPDATE_SCLK))) { - PP_ASSERT_WITH_CODE(true == polaris10_is_dpm_running(hwmgr), - "Trying to freeze SCLK DPM when DPM is disabled", + PP_ASSERT_WITH_CODE(polaris10_is_dpm_running(hwmgr), + "Trying to freeze SCLK DPM when DPM is disabled", ); PP_ASSERT_WITH_CODE(0 == smum_send_msg_to_smc(hwmgr->smumgr, PPSMC_MSG_SCLKDPM_FreezeLevel), @@ -3873,8 +4161,8 @@ static int polaris10_freeze_sclk_mclk_dpm(struct pp_hwmgr *hwmgr) if ((0 == data->mclk_dpm_key_disabled) && (data->need_update_smu7_dpm_table & DPMTABLE_OD_UPDATE_MCLK)) { - PP_ASSERT_WITH_CODE(true == polaris10_is_dpm_running(hwmgr), - "Trying to freeze MCLK DPM when DPM is disabled", + PP_ASSERT_WITH_CODE(polaris10_is_dpm_running(hwmgr), + "Trying to freeze MCLK DPM when DPM is disabled", ); PP_ASSERT_WITH_CODE(0 == smum_send_msg_to_smc(hwmgr->smumgr, PPSMC_MSG_MCLKDPM_FreezeLevel), @@ -4034,7 +4322,6 @@ static int polaris10_trim_single_dpm_states(struct pp_hwmgr *hwmgr, static int polaris10_trim_dpm_states(struct pp_hwmgr *hwmgr, const struct polaris10_power_state *polaris10_ps) { - int result = 0; struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend); uint32_t high_limit_count; @@ -4054,7 +4341,7 @@ static int polaris10_trim_dpm_states(struct pp_hwmgr *hwmgr, polaris10_ps->performance_levels[0].memory_clock, polaris10_ps->performance_levels[high_limit_count].memory_clock); - return result; + return 0; } static int polaris10_generate_dpm_level_enable_mask( @@ -4137,25 +4424,20 @@ int polaris10_update_uvd_dpm(struct pp_hwmgr *hwmgr, bool bgate) return polaris10_enable_disable_uvd_dpm(hwmgr, !bgate); } -static int polaris10_update_vce_dpm(struct pp_hwmgr *hwmgr, const void *input) +int polaris10_update_vce_dpm(struct pp_hwmgr *hwmgr, bool bgate) { - const struct phm_set_power_state_input *states = - (const struct phm_set_power_state_input *)input; struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend); - const struct polaris10_power_state *polaris10_nps = - cast_const_phw_polaris10_power_state(states->pnew_state); - const struct polaris10_power_state *polaris10_cps = - cast_const_phw_polaris10_power_state(states->pcurrent_state); - uint32_t mm_boot_level_offset, mm_boot_level_value; struct phm_ppt_v1_information *table_info = (struct phm_ppt_v1_information *)(hwmgr->pptable); - if (polaris10_nps->vce_clks.evclk > 0 && - (polaris10_cps == NULL || polaris10_cps->vce_clks.evclk == 0)) { - - data->smc_state_table.VceBootLevel = + if (!bgate) { + if (phm_cap_enabled(hwmgr->platform_descriptor.platformCaps, + PHM_PlatformCaps_StablePState)) + data->smc_state_table.VceBootLevel = (uint8_t) (table_info->mm_dep_table->count - 1); + else + data->smc_state_table.VceBootLevel = 0; mm_boot_level_offset = data->dpm_table_start + offsetof(SMU74_Discrete_DpmTable, VceBootLevel); @@ -4168,18 +4450,14 @@ static int polaris10_update_vce_dpm(struct pp_hwmgr *hwmgr, const void *input) cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC, mm_boot_level_offset, mm_boot_level_value); - if (phm_cap_enabled(hwmgr->platform_descriptor.platformCaps, PHM_PlatformCaps_StablePState)) { + if (phm_cap_enabled(hwmgr->platform_descriptor.platformCaps, PHM_PlatformCaps_StablePState)) smum_send_msg_to_smc_with_parameter(hwmgr->smumgr, PPSMC_MSG_VCEDPM_SetEnabledMask, (uint32_t)1 << data->smc_state_table.VceBootLevel); - - polaris10_enable_disable_vce_dpm(hwmgr, true); - } else if (polaris10_nps->vce_clks.evclk == 0 && - polaris10_cps != NULL && - polaris10_cps->vce_clks.evclk > 0) - polaris10_enable_disable_vce_dpm(hwmgr, false); } + polaris10_enable_disable_vce_dpm(hwmgr, !bgate); + return 0; } @@ -4187,12 +4465,9 @@ int polaris10_update_samu_dpm(struct pp_hwmgr *hwmgr, bool bgate) { struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend); uint32_t mm_boot_level_offset, mm_boot_level_value; - struct phm_ppt_v1_information *table_info = - (struct phm_ppt_v1_information *)(hwmgr->pptable); if (!bgate) { - data->smc_state_table.SamuBootLevel = - (uint8_t) (table_info->mm_dep_table->count - 1); + data->smc_state_table.SamuBootLevel = 0; mm_boot_level_offset = data->dpm_table_start + offsetof(SMU74_Discrete_DpmTable, SamuBootLevel); mm_boot_level_offset /= 4; @@ -4267,8 +4542,8 @@ static int polaris10_unfreeze_sclk_mclk_dpm(struct pp_hwmgr *hwmgr) (data->need_update_smu7_dpm_table & (DPMTABLE_OD_UPDATE_SCLK + DPMTABLE_UPDATE_SCLK))) { - PP_ASSERT_WITH_CODE(true == polaris10_is_dpm_running(hwmgr), - "Trying to Unfreeze SCLK DPM when DPM is disabled", + PP_ASSERT_WITH_CODE(polaris10_is_dpm_running(hwmgr), + "Trying to Unfreeze SCLK DPM when DPM is disabled", ); PP_ASSERT_WITH_CODE(0 == smum_send_msg_to_smc(hwmgr->smumgr, PPSMC_MSG_SCLKDPM_UnfreezeLevel), @@ -4279,8 +4554,8 @@ static int polaris10_unfreeze_sclk_mclk_dpm(struct pp_hwmgr *hwmgr) if ((0 == data->mclk_dpm_key_disabled) && (data->need_update_smu7_dpm_table & DPMTABLE_OD_UPDATE_MCLK)) { - PP_ASSERT_WITH_CODE(true == polaris10_is_dpm_running(hwmgr), - "Trying to Unfreeze MCLK DPM when DPM is disabled", + PP_ASSERT_WITH_CODE(polaris10_is_dpm_running(hwmgr), + "Trying to Unfreeze MCLK DPM when DPM is disabled", ); PP_ASSERT_WITH_CODE(0 == smum_send_msg_to_smc(hwmgr->smumgr, PPSMC_MSG_SCLKDPM_UnfreezeLevel), @@ -4327,6 +4602,17 @@ static int polaris10_notify_link_speed_change_after_state_change( return 0; } +static int polaris10_notify_smc_display(struct pp_hwmgr *hwmgr) +{ + struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend); + + smum_send_msg_to_smc_with_parameter(hwmgr->smumgr, + (PPSMC_Msg)PPSMC_MSG_SetVBITimeout, data->frame_time_x2); + return (smum_send_msg_to_smc(hwmgr->smumgr, (PPSMC_Msg)PPSMC_HasDisplay) == 0) ? 0 : -EINVAL; +} + + + static int polaris10_set_power_state_tasks(struct pp_hwmgr *hwmgr, const void *input) { int tmp_result, result = 0; @@ -4360,11 +4646,6 @@ static int polaris10_set_power_state_tasks(struct pp_hwmgr *hwmgr, const void *i "Failed to generate DPM level enabled mask!", result = tmp_result); - tmp_result = polaris10_update_vce_dpm(hwmgr, input); - PP_ASSERT_WITH_CODE((0 == tmp_result), - "Failed to update VCE DPM!", - result = tmp_result); - tmp_result = polaris10_update_sclk_threshold(hwmgr); PP_ASSERT_WITH_CODE((0 == tmp_result), "Failed to update SCLK threshold!", @@ -4375,6 +4656,11 @@ static int polaris10_set_power_state_tasks(struct pp_hwmgr *hwmgr, const void *i "Failed to program memory timing parameters!", result = tmp_result); + tmp_result = polaris10_notify_smc_display(hwmgr); + PP_ASSERT_WITH_CODE((0 == tmp_result), + "Failed to notify smc display settings!", + result = tmp_result); + tmp_result = polaris10_unfreeze_sclk_mclk_dpm(hwmgr); PP_ASSERT_WITH_CODE((0 == tmp_result), "Failed to unfreeze SCLK MCLK DPM!", @@ -4409,6 +4695,7 @@ static int polaris10_set_max_fan_pwm_output(struct pp_hwmgr *hwmgr, uint16_t us_ PPSMC_MSG_SetFanPwmMax, us_max_fan_pwm); } + int polaris10_notify_smc_display_change(struct pp_hwmgr *hwmgr, bool has_display) { PPSMC_Msg msg = has_display ? (PPSMC_Msg)PPSMC_HasDisplay : (PPSMC_Msg)PPSMC_NoDisplay; @@ -4428,8 +4715,7 @@ int polaris10_notify_smc_display_config_after_ps_adjustment(struct pp_hwmgr *hwm if (num_active_displays > 1) /* to do && (pHwMgr->pPECI->displayConfiguration.bMultiMonitorInSync != TRUE)) */ polaris10_notify_smc_display_change(hwmgr, false); - else - polaris10_notify_smc_display_change(hwmgr, true); + return 0; } @@ -4470,6 +4756,8 @@ int polaris10_program_display_gap(struct pp_hwmgr *hwmgr) frame_time_in_us = 1000000 / refresh_rate; pre_vbi_time_in_us = frame_time_in_us - 200 - mode_info.vblank_time_us; + data->frame_time_x2 = frame_time_in_us * 2 / 100; + display_gap2 = pre_vbi_time_in_us * (ref_clock / 100); cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC, ixCG_DISPLAY_GAP_CNTL2, display_gap2); @@ -4478,7 +4766,6 @@ int polaris10_program_display_gap(struct pp_hwmgr *hwmgr) cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC, data->soft_regs_start + offsetof(SMU74_SoftRegisters, VBlankTimeout), (frame_time_in_us - pre_vbi_time_in_us)); - polaris10_notify_smc_display_change(hwmgr, num_active_displays != 0); return 0; } @@ -4591,7 +4878,7 @@ int polaris10_upload_mc_firmware(struct pp_hwmgr *hwmgr) return 0; } - data->need_long_memory_training = true; + data->need_long_memory_training = false; /* * PPMCME_FirmwareDescriptorEntry *pfd = NULL; @@ -4721,42 +5008,6 @@ int polaris10_setup_asic_task(struct pp_hwmgr *hwmgr) return result; } -static int polaris10_get_pp_table(struct pp_hwmgr *hwmgr, char **table) -{ - struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend); - - if (!data->soft_pp_table) { - data->soft_pp_table = kmemdup(hwmgr->soft_pp_table, - hwmgr->soft_pp_table_size, - GFP_KERNEL); - if (!data->soft_pp_table) - return -ENOMEM; - } - - *table = (char *)&data->soft_pp_table; - - return hwmgr->soft_pp_table_size; -} - -static int polaris10_set_pp_table(struct pp_hwmgr *hwmgr, const char *buf, size_t size) -{ - struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend); - - if (!data->soft_pp_table) { - data->soft_pp_table = kzalloc(hwmgr->soft_pp_table_size, GFP_KERNEL); - if (!data->soft_pp_table) - return -ENOMEM; - } - - memcpy(data->soft_pp_table, buf, size); - - hwmgr->soft_pp_table = data->soft_pp_table; - - /* TODO: re-init powerplay to implement modified pptable */ - - return 0; -} - static int polaris10_force_clock_level(struct pp_hwmgr *hwmgr, enum pp_clock_type type, uint32_t mask) { @@ -4899,6 +5150,89 @@ static int polaris10_get_fan_control_mode(struct pp_hwmgr *hwmgr) CG_FDO_CTRL2, FDO_PWM_MODE); } +static int polaris10_get_sclk_od(struct pp_hwmgr *hwmgr) +{ + struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend); + struct polaris10_single_dpm_table *sclk_table = &(data->dpm_table.sclk_table); + struct polaris10_single_dpm_table *golden_sclk_table = + &(data->golden_dpm_table.sclk_table); + int value; + + value = (sclk_table->dpm_levels[sclk_table->count - 1].value - + golden_sclk_table->dpm_levels[golden_sclk_table->count - 1].value) * + 100 / + golden_sclk_table->dpm_levels[golden_sclk_table->count - 1].value; + + return value; +} + +static int polaris10_set_sclk_od(struct pp_hwmgr *hwmgr, uint32_t value) +{ + struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend); + struct polaris10_single_dpm_table *golden_sclk_table = + &(data->golden_dpm_table.sclk_table); + struct pp_power_state *ps; + struct polaris10_power_state *polaris10_ps; + + if (value > 20) + value = 20; + + ps = hwmgr->request_ps; + + if (ps == NULL) + return -EINVAL; + + polaris10_ps = cast_phw_polaris10_power_state(&ps->hardware); + + polaris10_ps->performance_levels[polaris10_ps->performance_level_count - 1].engine_clock = + golden_sclk_table->dpm_levels[golden_sclk_table->count - 1].value * + value / 100 + + golden_sclk_table->dpm_levels[golden_sclk_table->count - 1].value; + + return 0; +} + +static int polaris10_get_mclk_od(struct pp_hwmgr *hwmgr) +{ + struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend); + struct polaris10_single_dpm_table *mclk_table = &(data->dpm_table.mclk_table); + struct polaris10_single_dpm_table *golden_mclk_table = + &(data->golden_dpm_table.mclk_table); + int value; + + value = (mclk_table->dpm_levels[mclk_table->count - 1].value - + golden_mclk_table->dpm_levels[golden_mclk_table->count - 1].value) * + 100 / + golden_mclk_table->dpm_levels[golden_mclk_table->count - 1].value; + + return value; +} + +static int polaris10_set_mclk_od(struct pp_hwmgr *hwmgr, uint32_t value) +{ + struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend); + struct polaris10_single_dpm_table *golden_mclk_table = + &(data->golden_dpm_table.mclk_table); + struct pp_power_state *ps; + struct polaris10_power_state *polaris10_ps; + + if (value > 20) + value = 20; + + ps = hwmgr->request_ps; + + if (ps == NULL) + return -EINVAL; + + polaris10_ps = cast_phw_polaris10_power_state(&ps->hardware); + + polaris10_ps->performance_levels[polaris10_ps->performance_level_count - 1].memory_clock = + golden_mclk_table->dpm_levels[golden_mclk_table->count - 1].value * + value / 100 + + golden_mclk_table->dpm_levels[golden_mclk_table->count - 1].value; + + return 0; +} static const struct pp_hwmgr_func polaris10_hwmgr_funcs = { .backend_init = &polaris10_hwmgr_backend_init, .backend_fini = &polaris10_hwmgr_backend_fini, @@ -4937,22 +5271,17 @@ static const struct pp_hwmgr_func polaris10_hwmgr_funcs = { .check_states_equal = polaris10_check_states_equal, .set_fan_control_mode = polaris10_set_fan_control_mode, .get_fan_control_mode = polaris10_get_fan_control_mode, - .get_pp_table = polaris10_get_pp_table, - .set_pp_table = polaris10_set_pp_table, .force_clock_level = polaris10_force_clock_level, .print_clock_levels = polaris10_print_clock_levels, .enable_per_cu_power_gating = polaris10_phm_enable_per_cu_power_gating, + .get_sclk_od = polaris10_get_sclk_od, + .set_sclk_od = polaris10_set_sclk_od, + .get_mclk_od = polaris10_get_mclk_od, + .set_mclk_od = polaris10_set_mclk_od, }; int polaris10_hwmgr_init(struct pp_hwmgr *hwmgr) { - struct polaris10_hwmgr *data; - - data = kzalloc (sizeof(struct polaris10_hwmgr), GFP_KERNEL); - if (data == NULL) - return -ENOMEM; - - hwmgr->backend = data; hwmgr->hwmgr_func = &polaris10_hwmgr_funcs; hwmgr->pptable_func = &tonga_pptable_funcs; pp_polaris10_thermal_initialize(hwmgr); diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_hwmgr.h b/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_hwmgr.h index beedf35cbfa6..33c33947e827 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_hwmgr.h +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_hwmgr.h @@ -309,9 +309,9 @@ struct polaris10_hwmgr { uint32_t up_hyst; uint32_t disable_dpm_mask; bool apply_optimized_settings; - - /* soft pptable for re-uploading into smu */ - void *soft_pp_table; + uint32_t avfs_vdroop_override_setting; + bool apply_avfs_cks_off_voltage; + uint32_t frame_time_x2; }; /* To convert to Q8.8 format for firmware */ @@ -352,6 +352,6 @@ int polaris10_hwmgr_init(struct pp_hwmgr *hwmgr); int polaris10_update_uvd_dpm(struct pp_hwmgr *hwmgr, bool bgate); int polaris10_update_samu_dpm(struct pp_hwmgr *hwmgr, bool bgate); int polaris10_enable_disable_vce_dpm(struct pp_hwmgr *hwmgr, bool enable); - +int polaris10_update_vce_dpm(struct pp_hwmgr *hwmgr, bool bgate); #endif diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_powertune.c b/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_powertune.c index 0b99ab3ba0c5..b9cb240a135d 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_powertune.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_powertune.c @@ -28,10 +28,360 @@ #include "polaris10_smumgr.h" #include "smu74_discrete.h" #include "pp_debug.h" +#include "gca/gfx_8_0_d.h" +#include "gca/gfx_8_0_sh_mask.h" +#include "oss/oss_3_0_sh_mask.h" #define VOLTAGE_SCALE 4 #define POWERTUNE_DEFAULT_SET_MAX 1 +uint32_t DIDTBlock_Info = SQ_IR_MASK | TCP_IR_MASK | TD_PCC_MASK; + +struct polaris10_pt_config_reg GCCACConfig_Polaris10[] = { +/* --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- + * Offset Mask Shift Value Type + * --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- + */ + { ixGC_CAC_CNTL, 0xFFFFFFFF, 0, 0x00060013, POLARIS10_CONFIGREG_GC_CAC_IND }, + { ixGC_CAC_CNTL, 0xFFFFFFFF, 0, 0x00860013, POLARIS10_CONFIGREG_GC_CAC_IND }, + { ixGC_CAC_CNTL, 0xFFFFFFFF, 0, 0x01060013, POLARIS10_CONFIGREG_GC_CAC_IND }, + { ixGC_CAC_CNTL, 0xFFFFFFFF, 0, 0x01860013, POLARIS10_CONFIGREG_GC_CAC_IND }, + { ixGC_CAC_CNTL, 0xFFFFFFFF, 0, 0x02060013, POLARIS10_CONFIGREG_GC_CAC_IND }, + { ixGC_CAC_CNTL, 0xFFFFFFFF, 0, 0x02860013, POLARIS10_CONFIGREG_GC_CAC_IND }, + { ixGC_CAC_CNTL, 0xFFFFFFFF, 0, 0x03060013, POLARIS10_CONFIGREG_GC_CAC_IND }, + { ixGC_CAC_CNTL, 0xFFFFFFFF, 0, 0x03860013, POLARIS10_CONFIGREG_GC_CAC_IND }, + { ixGC_CAC_CNTL, 0xFFFFFFFF, 0, 0x04060013, POLARIS10_CONFIGREG_GC_CAC_IND }, + + { ixGC_CAC_CNTL, 0xFFFFFFFF, 0, 0x000E0013, POLARIS10_CONFIGREG_GC_CAC_IND }, + { ixGC_CAC_CNTL, 0xFFFFFFFF, 0, 0x008E0013, POLARIS10_CONFIGREG_GC_CAC_IND }, + { ixGC_CAC_CNTL, 0xFFFFFFFF, 0, 0x010E0013, POLARIS10_CONFIGREG_GC_CAC_IND }, + { ixGC_CAC_CNTL, 0xFFFFFFFF, 0, 0x018E0013, POLARIS10_CONFIGREG_GC_CAC_IND }, + { ixGC_CAC_CNTL, 0xFFFFFFFF, 0, 0x020E0013, POLARIS10_CONFIGREG_GC_CAC_IND }, + + { ixGC_CAC_CNTL, 0xFFFFFFFF, 0, 0x00100013, POLARIS10_CONFIGREG_GC_CAC_IND }, + { ixGC_CAC_CNTL, 0xFFFFFFFF, 0, 0x00900013, POLARIS10_CONFIGREG_GC_CAC_IND }, + { ixGC_CAC_CNTL, 0xFFFFFFFF, 0, 0x01100013, POLARIS10_CONFIGREG_GC_CAC_IND }, + { ixGC_CAC_CNTL, 0xFFFFFFFF, 0, 0x01900013, POLARIS10_CONFIGREG_GC_CAC_IND }, + { ixGC_CAC_CNTL, 0xFFFFFFFF, 0, 0x02100013, POLARIS10_CONFIGREG_GC_CAC_IND }, + { ixGC_CAC_CNTL, 0xFFFFFFFF, 0, 0x02900013, POLARIS10_CONFIGREG_GC_CAC_IND }, + + { 0xFFFFFFFF } +}; + +struct polaris10_pt_config_reg GCCACConfig_Polaris11[] = { +/* --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- + * Offset Mask Shift Value Type + * --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- + */ + { ixGC_CAC_CNTL, 0xFFFFFFFF, 0, 0x00060011, POLARIS10_CONFIGREG_GC_CAC_IND }, + { ixGC_CAC_CNTL, 0xFFFFFFFF, 0, 0x00860011, POLARIS10_CONFIGREG_GC_CAC_IND }, + { ixGC_CAC_CNTL, 0xFFFFFFFF, 0, 0x01060011, POLARIS10_CONFIGREG_GC_CAC_IND }, + { ixGC_CAC_CNTL, 0xFFFFFFFF, 0, 0x01860011, POLARIS10_CONFIGREG_GC_CAC_IND }, + { ixGC_CAC_CNTL, 0xFFFFFFFF, 0, 0x02060011, POLARIS10_CONFIGREG_GC_CAC_IND }, + { ixGC_CAC_CNTL, 0xFFFFFFFF, 0, 0x02860011, POLARIS10_CONFIGREG_GC_CAC_IND }, + { ixGC_CAC_CNTL, 0xFFFFFFFF, 0, 0x03060011, POLARIS10_CONFIGREG_GC_CAC_IND }, + { ixGC_CAC_CNTL, 0xFFFFFFFF, 0, 0x03860011, POLARIS10_CONFIGREG_GC_CAC_IND }, + { ixGC_CAC_CNTL, 0xFFFFFFFF, 0, 0x04060011, POLARIS10_CONFIGREG_GC_CAC_IND }, + + { ixGC_CAC_CNTL, 0xFFFFFFFF, 0, 0x000E0011, POLARIS10_CONFIGREG_GC_CAC_IND }, + { ixGC_CAC_CNTL, 0xFFFFFFFF, 0, 0x008E0011, POLARIS10_CONFIGREG_GC_CAC_IND }, + { ixGC_CAC_CNTL, 0xFFFFFFFF, 0, 0x010E0011, POLARIS10_CONFIGREG_GC_CAC_IND }, + { ixGC_CAC_CNTL, 0xFFFFFFFF, 0, 0x018E0011, POLARIS10_CONFIGREG_GC_CAC_IND }, + { ixGC_CAC_CNTL, 0xFFFFFFFF, 0, 0x020E0011, POLARIS10_CONFIGREG_GC_CAC_IND }, + + { ixGC_CAC_CNTL, 0xFFFFFFFF, 0, 0x00100011, POLARIS10_CONFIGREG_GC_CAC_IND }, + { ixGC_CAC_CNTL, 0xFFFFFFFF, 0, 0x00900011, POLARIS10_CONFIGREG_GC_CAC_IND }, + { ixGC_CAC_CNTL, 0xFFFFFFFF, 0, 0x01100011, POLARIS10_CONFIGREG_GC_CAC_IND }, + { ixGC_CAC_CNTL, 0xFFFFFFFF, 0, 0x01900011, POLARIS10_CONFIGREG_GC_CAC_IND }, + { ixGC_CAC_CNTL, 0xFFFFFFFF, 0, 0x02100011, POLARIS10_CONFIGREG_GC_CAC_IND }, + { ixGC_CAC_CNTL, 0xFFFFFFFF, 0, 0x02900011, POLARIS10_CONFIGREG_GC_CAC_IND }, + + { 0xFFFFFFFF } +}; + +struct polaris10_pt_config_reg DIDTConfig_Polaris10[] = { +/* --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- + * Offset Mask Shift Value Type + * --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- + */ + { ixDIDT_SQ_WEIGHT0_3, DIDT_SQ_WEIGHT0_3__WEIGHT0_MASK, DIDT_SQ_WEIGHT0_3__WEIGHT0__SHIFT, 0x0073, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_WEIGHT0_3, DIDT_SQ_WEIGHT0_3__WEIGHT1_MASK, DIDT_SQ_WEIGHT0_3__WEIGHT1__SHIFT, 0x00ab, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_WEIGHT0_3, DIDT_SQ_WEIGHT0_3__WEIGHT2_MASK, DIDT_SQ_WEIGHT0_3__WEIGHT2__SHIFT, 0x0084, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_WEIGHT0_3, DIDT_SQ_WEIGHT0_3__WEIGHT3_MASK, DIDT_SQ_WEIGHT0_3__WEIGHT3__SHIFT, 0x005a, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_SQ_WEIGHT4_7, DIDT_SQ_WEIGHT4_7__WEIGHT4_MASK, DIDT_SQ_WEIGHT4_7__WEIGHT4__SHIFT, 0x0067, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_WEIGHT4_7, DIDT_SQ_WEIGHT4_7__WEIGHT5_MASK, DIDT_SQ_WEIGHT4_7__WEIGHT5__SHIFT, 0x0084, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_WEIGHT4_7, DIDT_SQ_WEIGHT4_7__WEIGHT6_MASK, DIDT_SQ_WEIGHT4_7__WEIGHT6__SHIFT, 0x0027, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_WEIGHT4_7, DIDT_SQ_WEIGHT4_7__WEIGHT7_MASK, DIDT_SQ_WEIGHT4_7__WEIGHT7__SHIFT, 0x0046, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_SQ_WEIGHT8_11, DIDT_SQ_WEIGHT8_11__WEIGHT8_MASK, DIDT_SQ_WEIGHT8_11__WEIGHT8__SHIFT, 0x00aa, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_WEIGHT8_11, DIDT_SQ_WEIGHT8_11__WEIGHT9_MASK, DIDT_SQ_WEIGHT8_11__WEIGHT9__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_WEIGHT8_11, DIDT_SQ_WEIGHT8_11__WEIGHT10_MASK, DIDT_SQ_WEIGHT8_11__WEIGHT10__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_WEIGHT8_11, DIDT_SQ_WEIGHT8_11__WEIGHT11_MASK, DIDT_SQ_WEIGHT8_11__WEIGHT11__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_SQ_CTRL1, DIDT_SQ_CTRL1__MIN_POWER_MASK, DIDT_SQ_CTRL1__MIN_POWER__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_CTRL1, DIDT_SQ_CTRL1__MAX_POWER_MASK, DIDT_SQ_CTRL1__MAX_POWER__SHIFT, 0xffff, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_SQ_CTRL_OCP, DIDT_SQ_CTRL_OCP__UNUSED_0_MASK, DIDT_SQ_CTRL_OCP__UNUSED_0__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_CTRL_OCP, DIDT_SQ_CTRL_OCP__OCP_MAX_POWER_MASK, DIDT_SQ_CTRL_OCP__OCP_MAX_POWER__SHIFT, 0xffff, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_SQ_CTRL2, DIDT_SQ_CTRL2__MAX_POWER_DELTA_MASK, DIDT_SQ_CTRL2__MAX_POWER_DELTA__SHIFT, 0x3853, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_CTRL2, DIDT_SQ_CTRL2__UNUSED_0_MASK, DIDT_SQ_CTRL2__UNUSED_0__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_CTRL2, DIDT_SQ_CTRL2__SHORT_TERM_INTERVAL_SIZE_MASK, DIDT_SQ_CTRL2__SHORT_TERM_INTERVAL_SIZE__SHIFT, 0x005a, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_CTRL2, DIDT_SQ_CTRL2__UNUSED_1_MASK, DIDT_SQ_CTRL2__UNUSED_1__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_CTRL2, DIDT_SQ_CTRL2__LONG_TERM_INTERVAL_RATIO_MASK, DIDT_SQ_CTRL2__LONG_TERM_INTERVAL_RATIO__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_CTRL2, DIDT_SQ_CTRL2__UNUSED_2_MASK, DIDT_SQ_CTRL2__UNUSED_2__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_SQ_STALL_CTRL, DIDT_SQ_STALL_CTRL__DIDT_STALL_CTRL_ENABLE_MASK, DIDT_SQ_STALL_CTRL__DIDT_STALL_CTRL_ENABLE__SHIFT, 0x0001, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_STALL_CTRL, DIDT_SQ_STALL_CTRL__DIDT_STALL_DELAY_HI_MASK, DIDT_SQ_STALL_CTRL__DIDT_STALL_DELAY_HI__SHIFT, 0x0001, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_STALL_CTRL, DIDT_SQ_STALL_CTRL__DIDT_STALL_DELAY_LO_MASK, DIDT_SQ_STALL_CTRL__DIDT_STALL_DELAY_LO__SHIFT, 0x0001, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_STALL_CTRL, DIDT_SQ_STALL_CTRL__DIDT_HI_POWER_THRESHOLD_MASK, DIDT_SQ_STALL_CTRL__DIDT_HI_POWER_THRESHOLD__SHIFT, 0x0ebb, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_STALL_CTRL, DIDT_SQ_STALL_CTRL__UNUSED_0_MASK, DIDT_SQ_STALL_CTRL__UNUSED_0__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_SQ_TUNING_CTRL, DIDT_SQ_TUNING_CTRL__DIDT_TUNING_ENABLE_MASK, DIDT_SQ_TUNING_CTRL__DIDT_TUNING_ENABLE__SHIFT, 0x0001, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_TUNING_CTRL, DIDT_SQ_TUNING_CTRL__MAX_POWER_DELTA_HI_MASK, DIDT_SQ_TUNING_CTRL__MAX_POWER_DELTA_HI__SHIFT, 0x3853, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_TUNING_CTRL, DIDT_SQ_TUNING_CTRL__MAX_POWER_DELTA_LO_MASK, DIDT_SQ_TUNING_CTRL__MAX_POWER_DELTA_LO__SHIFT, 0x3153, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_TUNING_CTRL, DIDT_SQ_TUNING_CTRL__UNUSED_0_MASK, DIDT_SQ_TUNING_CTRL__UNUSED_0__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_SQ_CTRL0, DIDT_SQ_CTRL0__DIDT_CTRL_EN_MASK, DIDT_SQ_CTRL0__DIDT_CTRL_EN__SHIFT, 0x0001, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_CTRL0, DIDT_SQ_CTRL0__USE_REF_CLOCK_MASK, DIDT_SQ_CTRL0__USE_REF_CLOCK__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_CTRL0, DIDT_SQ_CTRL0__PHASE_OFFSET_MASK, DIDT_SQ_CTRL0__PHASE_OFFSET__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_CTRL0, DIDT_SQ_CTRL0__DIDT_CTRL_RST_MASK, DIDT_SQ_CTRL0__DIDT_CTRL_RST__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_CTRL0, DIDT_SQ_CTRL0__DIDT_CLK_EN_OVERRIDE_MASK, DIDT_SQ_CTRL0__DIDT_CLK_EN_OVERRIDE__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_CTRL0, DIDT_SQ_CTRL0__DIDT_MAX_STALLS_ALLOWED_HI_MASK, DIDT_SQ_CTRL0__DIDT_MAX_STALLS_ALLOWED_HI__SHIFT, 0x0010, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_CTRL0, DIDT_SQ_CTRL0__DIDT_MAX_STALLS_ALLOWED_LO_MASK, DIDT_SQ_CTRL0__DIDT_MAX_STALLS_ALLOWED_LO__SHIFT, 0x0010, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_CTRL0, DIDT_SQ_CTRL0__UNUSED_0_MASK, DIDT_SQ_CTRL0__UNUSED_0__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_TD_WEIGHT0_3, DIDT_TD_WEIGHT0_3__WEIGHT0_MASK, DIDT_TD_WEIGHT0_3__WEIGHT0__SHIFT, 0x000a, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_WEIGHT0_3, DIDT_TD_WEIGHT0_3__WEIGHT1_MASK, DIDT_TD_WEIGHT0_3__WEIGHT1__SHIFT, 0x0010, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_WEIGHT0_3, DIDT_TD_WEIGHT0_3__WEIGHT2_MASK, DIDT_TD_WEIGHT0_3__WEIGHT2__SHIFT, 0x0017, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_WEIGHT0_3, DIDT_TD_WEIGHT0_3__WEIGHT3_MASK, DIDT_TD_WEIGHT0_3__WEIGHT3__SHIFT, 0x002f, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_TD_WEIGHT4_7, DIDT_TD_WEIGHT4_7__WEIGHT4_MASK, DIDT_TD_WEIGHT4_7__WEIGHT4__SHIFT, 0x0046, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_WEIGHT4_7, DIDT_TD_WEIGHT4_7__WEIGHT5_MASK, DIDT_TD_WEIGHT4_7__WEIGHT5__SHIFT, 0x005d, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_WEIGHT4_7, DIDT_TD_WEIGHT4_7__WEIGHT6_MASK, DIDT_TD_WEIGHT4_7__WEIGHT6__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_WEIGHT4_7, DIDT_TD_WEIGHT4_7__WEIGHT7_MASK, DIDT_TD_WEIGHT4_7__WEIGHT7__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_TD_CTRL1, DIDT_TD_CTRL1__MIN_POWER_MASK, DIDT_TD_CTRL1__MIN_POWER__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_CTRL1, DIDT_TD_CTRL1__MAX_POWER_MASK, DIDT_TD_CTRL1__MAX_POWER__SHIFT, 0xffff, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_TD_CTRL_OCP, DIDT_TD_CTRL_OCP__UNUSED_0_MASK, DIDT_TD_CTRL_OCP__UNUSED_0__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_CTRL_OCP, DIDT_TD_CTRL_OCP__OCP_MAX_POWER_MASK, DIDT_TD_CTRL_OCP__OCP_MAX_POWER__SHIFT, 0x00ff, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_TD_CTRL2, DIDT_TD_CTRL2__MAX_POWER_DELTA_MASK, DIDT_TD_CTRL2__MAX_POWER_DELTA__SHIFT, 0x3fff, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_CTRL2, DIDT_TD_CTRL2__UNUSED_0_MASK, DIDT_TD_CTRL2__UNUSED_0__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_CTRL2, DIDT_TD_CTRL2__SHORT_TERM_INTERVAL_SIZE_MASK, DIDT_TD_CTRL2__SHORT_TERM_INTERVAL_SIZE__SHIFT, 0x000f, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_CTRL2, DIDT_TD_CTRL2__UNUSED_1_MASK, DIDT_TD_CTRL2__UNUSED_1__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_CTRL2, DIDT_TD_CTRL2__LONG_TERM_INTERVAL_RATIO_MASK, DIDT_TD_CTRL2__LONG_TERM_INTERVAL_RATIO__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_CTRL2, DIDT_TD_CTRL2__UNUSED_2_MASK, DIDT_TD_CTRL2__UNUSED_2__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_TD_STALL_CTRL, DIDT_TD_STALL_CTRL__DIDT_STALL_CTRL_ENABLE_MASK, DIDT_TD_STALL_CTRL__DIDT_STALL_CTRL_ENABLE__SHIFT, 0x0001, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_STALL_CTRL, DIDT_TD_STALL_CTRL__DIDT_STALL_DELAY_HI_MASK, DIDT_TD_STALL_CTRL__DIDT_STALL_DELAY_HI__SHIFT, 0x0001, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_STALL_CTRL, DIDT_TD_STALL_CTRL__DIDT_STALL_DELAY_LO_MASK, DIDT_TD_STALL_CTRL__DIDT_STALL_DELAY_LO__SHIFT, 0x0001, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_STALL_CTRL, DIDT_TD_STALL_CTRL__DIDT_HI_POWER_THRESHOLD_MASK, DIDT_TD_STALL_CTRL__DIDT_HI_POWER_THRESHOLD__SHIFT, 0x01aa, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_STALL_CTRL, DIDT_TD_STALL_CTRL__UNUSED_0_MASK, DIDT_TD_STALL_CTRL__UNUSED_0__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_TD_TUNING_CTRL, DIDT_TD_TUNING_CTRL__DIDT_TUNING_ENABLE_MASK, DIDT_TD_TUNING_CTRL__DIDT_TUNING_ENABLE__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_TUNING_CTRL, DIDT_TD_TUNING_CTRL__MAX_POWER_DELTA_HI_MASK, DIDT_TD_TUNING_CTRL__MAX_POWER_DELTA_HI__SHIFT, 0x0dde, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_TUNING_CTRL, DIDT_TD_TUNING_CTRL__MAX_POWER_DELTA_LO_MASK, DIDT_TD_TUNING_CTRL__MAX_POWER_DELTA_LO__SHIFT, 0x0dde, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_TUNING_CTRL, DIDT_TD_TUNING_CTRL__UNUSED_0_MASK, DIDT_TD_TUNING_CTRL__UNUSED_0__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_TD_CTRL0, DIDT_TD_CTRL0__DIDT_CTRL_EN_MASK, DIDT_TD_CTRL0__DIDT_CTRL_EN__SHIFT, 0x0001, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_CTRL0, DIDT_TD_CTRL0__USE_REF_CLOCK_MASK, DIDT_TD_CTRL0__USE_REF_CLOCK__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_CTRL0, DIDT_TD_CTRL0__PHASE_OFFSET_MASK, DIDT_TD_CTRL0__PHASE_OFFSET__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_CTRL0, DIDT_TD_CTRL0__DIDT_CTRL_RST_MASK, DIDT_TD_CTRL0__DIDT_CTRL_RST__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_CTRL0, DIDT_TD_CTRL0__DIDT_CLK_EN_OVERRIDE_MASK, DIDT_TD_CTRL0__DIDT_CLK_EN_OVERRIDE__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_CTRL0, DIDT_TD_CTRL0__DIDT_MAX_STALLS_ALLOWED_HI_MASK, DIDT_TD_CTRL0__DIDT_MAX_STALLS_ALLOWED_HI__SHIFT, 0x0009, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_CTRL0, DIDT_TD_CTRL0__DIDT_MAX_STALLS_ALLOWED_LO_MASK, DIDT_TD_CTRL0__DIDT_MAX_STALLS_ALLOWED_LO__SHIFT, 0x0009, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_CTRL0, DIDT_TD_CTRL0__UNUSED_0_MASK, DIDT_TD_CTRL0__UNUSED_0__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_TCP_WEIGHT0_3, DIDT_TCP_WEIGHT0_3__WEIGHT0_MASK, DIDT_TCP_WEIGHT0_3__WEIGHT0__SHIFT, 0x0004, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_WEIGHT0_3, DIDT_TCP_WEIGHT0_3__WEIGHT1_MASK, DIDT_TCP_WEIGHT0_3__WEIGHT1__SHIFT, 0x0037, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_WEIGHT0_3, DIDT_TCP_WEIGHT0_3__WEIGHT2_MASK, DIDT_TCP_WEIGHT0_3__WEIGHT2__SHIFT, 0x0001, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_WEIGHT0_3, DIDT_TCP_WEIGHT0_3__WEIGHT3_MASK, DIDT_TCP_WEIGHT0_3__WEIGHT3__SHIFT, 0x00ff, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_TCP_WEIGHT4_7, DIDT_TCP_WEIGHT4_7__WEIGHT4_MASK, DIDT_TCP_WEIGHT4_7__WEIGHT4__SHIFT, 0x0054, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_WEIGHT4_7, DIDT_TCP_WEIGHT4_7__WEIGHT5_MASK, DIDT_TCP_WEIGHT4_7__WEIGHT5__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_WEIGHT4_7, DIDT_TCP_WEIGHT4_7__WEIGHT6_MASK, DIDT_TCP_WEIGHT4_7__WEIGHT6__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_WEIGHT4_7, DIDT_TCP_WEIGHT4_7__WEIGHT7_MASK, DIDT_TCP_WEIGHT4_7__WEIGHT7__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_TCP_CTRL1, DIDT_TCP_CTRL1__MIN_POWER_MASK, DIDT_TCP_CTRL1__MIN_POWER__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_CTRL1, DIDT_TCP_CTRL1__MAX_POWER_MASK, DIDT_TCP_CTRL1__MAX_POWER__SHIFT, 0xffff, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_TCP_CTRL_OCP, DIDT_TCP_CTRL_OCP__UNUSED_0_MASK, DIDT_TCP_CTRL_OCP__UNUSED_0__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_CTRL_OCP, DIDT_TCP_CTRL_OCP__OCP_MAX_POWER_MASK, DIDT_TCP_CTRL_OCP__OCP_MAX_POWER__SHIFT, 0xffff, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_TCP_CTRL2, DIDT_TCP_CTRL2__MAX_POWER_DELTA_MASK, DIDT_TCP_CTRL2__MAX_POWER_DELTA__SHIFT, 0x3dde, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_CTRL2, DIDT_TCP_CTRL2__UNUSED_0_MASK, DIDT_TCP_CTRL2__UNUSED_0__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_CTRL2, DIDT_TCP_CTRL2__SHORT_TERM_INTERVAL_SIZE_MASK, DIDT_TCP_CTRL2__SHORT_TERM_INTERVAL_SIZE__SHIFT, 0x0032, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_CTRL2, DIDT_TCP_CTRL2__UNUSED_1_MASK, DIDT_TCP_CTRL2__UNUSED_1__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_CTRL2, DIDT_TCP_CTRL2__LONG_TERM_INTERVAL_RATIO_MASK, DIDT_TCP_CTRL2__LONG_TERM_INTERVAL_RATIO__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_CTRL2, DIDT_TCP_CTRL2__UNUSED_2_MASK, DIDT_TCP_CTRL2__UNUSED_2__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_TCP_STALL_CTRL, DIDT_TCP_STALL_CTRL__DIDT_STALL_CTRL_ENABLE_MASK, DIDT_TCP_STALL_CTRL__DIDT_STALL_CTRL_ENABLE__SHIFT, 0x0001, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_STALL_CTRL, DIDT_TCP_STALL_CTRL__DIDT_STALL_DELAY_HI_MASK, DIDT_TCP_STALL_CTRL__DIDT_STALL_DELAY_HI__SHIFT, 0x0001, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_STALL_CTRL, DIDT_TCP_STALL_CTRL__DIDT_STALL_DELAY_LO_MASK, DIDT_TCP_STALL_CTRL__DIDT_STALL_DELAY_LO__SHIFT, 0x0001, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_STALL_CTRL, DIDT_TCP_STALL_CTRL__DIDT_HI_POWER_THRESHOLD_MASK, DIDT_TCP_STALL_CTRL__DIDT_HI_POWER_THRESHOLD__SHIFT, 0x01aa, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_STALL_CTRL, DIDT_TCP_STALL_CTRL__UNUSED_0_MASK, DIDT_TCP_STALL_CTRL__UNUSED_0__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_TCP_TUNING_CTRL, DIDT_TCP_TUNING_CTRL__DIDT_TUNING_ENABLE_MASK, DIDT_TCP_TUNING_CTRL__DIDT_TUNING_ENABLE__SHIFT, 0x0001, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_TUNING_CTRL, DIDT_TCP_TUNING_CTRL__MAX_POWER_DELTA_HI_MASK, DIDT_TCP_TUNING_CTRL__MAX_POWER_DELTA_HI__SHIFT, 0x3dde, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_TUNING_CTRL, DIDT_TCP_TUNING_CTRL__MAX_POWER_DELTA_LO_MASK, DIDT_TCP_TUNING_CTRL__MAX_POWER_DELTA_LO__SHIFT, 0x3dde, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_TUNING_CTRL, DIDT_TCP_TUNING_CTRL__UNUSED_0_MASK, DIDT_TCP_TUNING_CTRL__UNUSED_0__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_TCP_CTRL0, DIDT_TCP_CTRL0__DIDT_CTRL_EN_MASK, DIDT_TCP_CTRL0__DIDT_CTRL_EN__SHIFT, 0x0001, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_CTRL0, DIDT_TCP_CTRL0__USE_REF_CLOCK_MASK, DIDT_TCP_CTRL0__USE_REF_CLOCK__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_CTRL0, DIDT_TCP_CTRL0__PHASE_OFFSET_MASK, DIDT_TCP_CTRL0__PHASE_OFFSET__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_CTRL0, DIDT_TCP_CTRL0__DIDT_CTRL_RST_MASK, DIDT_TCP_CTRL0__DIDT_CTRL_RST__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_CTRL0, DIDT_TCP_CTRL0__DIDT_CLK_EN_OVERRIDE_MASK, DIDT_TCP_CTRL0__DIDT_CLK_EN_OVERRIDE__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_CTRL0, DIDT_TCP_CTRL0__DIDT_MAX_STALLS_ALLOWED_HI_MASK, DIDT_TCP_CTRL0__DIDT_MAX_STALLS_ALLOWED_HI__SHIFT, 0x0010, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_CTRL0, DIDT_TCP_CTRL0__DIDT_MAX_STALLS_ALLOWED_LO_MASK, DIDT_TCP_CTRL0__DIDT_MAX_STALLS_ALLOWED_LO__SHIFT, 0x0010, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_CTRL0, DIDT_TCP_CTRL0__UNUSED_0_MASK, DIDT_TCP_CTRL0__UNUSED_0__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + + { 0xFFFFFFFF } +}; + +struct polaris10_pt_config_reg DIDTConfig_Polaris11[] = { +/* --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- + * Offset Mask Shift Value Type + * --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- + */ + { ixDIDT_SQ_WEIGHT0_3, DIDT_SQ_WEIGHT0_3__WEIGHT0_MASK, DIDT_SQ_WEIGHT0_3__WEIGHT0__SHIFT, 0x0073, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_WEIGHT0_3, DIDT_SQ_WEIGHT0_3__WEIGHT1_MASK, DIDT_SQ_WEIGHT0_3__WEIGHT1__SHIFT, 0x00ab, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_WEIGHT0_3, DIDT_SQ_WEIGHT0_3__WEIGHT2_MASK, DIDT_SQ_WEIGHT0_3__WEIGHT2__SHIFT, 0x0084, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_WEIGHT0_3, DIDT_SQ_WEIGHT0_3__WEIGHT3_MASK, DIDT_SQ_WEIGHT0_3__WEIGHT3__SHIFT, 0x005a, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_SQ_WEIGHT4_7, DIDT_SQ_WEIGHT4_7__WEIGHT4_MASK, DIDT_SQ_WEIGHT4_7__WEIGHT4__SHIFT, 0x0067, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_WEIGHT4_7, DIDT_SQ_WEIGHT4_7__WEIGHT5_MASK, DIDT_SQ_WEIGHT4_7__WEIGHT5__SHIFT, 0x0084, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_WEIGHT4_7, DIDT_SQ_WEIGHT4_7__WEIGHT6_MASK, DIDT_SQ_WEIGHT4_7__WEIGHT6__SHIFT, 0x0027, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_WEIGHT4_7, DIDT_SQ_WEIGHT4_7__WEIGHT7_MASK, DIDT_SQ_WEIGHT4_7__WEIGHT7__SHIFT, 0x0046, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_SQ_WEIGHT8_11, DIDT_SQ_WEIGHT8_11__WEIGHT8_MASK, DIDT_SQ_WEIGHT8_11__WEIGHT8__SHIFT, 0x00aa, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_WEIGHT8_11, DIDT_SQ_WEIGHT8_11__WEIGHT9_MASK, DIDT_SQ_WEIGHT8_11__WEIGHT9__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_WEIGHT8_11, DIDT_SQ_WEIGHT8_11__WEIGHT10_MASK, DIDT_SQ_WEIGHT8_11__WEIGHT10__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_WEIGHT8_11, DIDT_SQ_WEIGHT8_11__WEIGHT11_MASK, DIDT_SQ_WEIGHT8_11__WEIGHT11__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_SQ_CTRL1, DIDT_SQ_CTRL1__MIN_POWER_MASK, DIDT_SQ_CTRL1__MIN_POWER__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_CTRL1, DIDT_SQ_CTRL1__MAX_POWER_MASK, DIDT_SQ_CTRL1__MAX_POWER__SHIFT, 0xffff, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_SQ_CTRL_OCP, DIDT_SQ_CTRL_OCP__UNUSED_0_MASK, DIDT_SQ_CTRL_OCP__UNUSED_0__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_CTRL_OCP, DIDT_SQ_CTRL_OCP__OCP_MAX_POWER_MASK, DIDT_SQ_CTRL_OCP__OCP_MAX_POWER__SHIFT, 0xffff, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_SQ_CTRL2, DIDT_SQ_CTRL2__MAX_POWER_DELTA_MASK, DIDT_SQ_CTRL2__MAX_POWER_DELTA__SHIFT, 0x3853, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_CTRL2, DIDT_SQ_CTRL2__UNUSED_0_MASK, DIDT_SQ_CTRL2__UNUSED_0__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_CTRL2, DIDT_SQ_CTRL2__SHORT_TERM_INTERVAL_SIZE_MASK, DIDT_SQ_CTRL2__SHORT_TERM_INTERVAL_SIZE__SHIFT, 0x005a, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_CTRL2, DIDT_SQ_CTRL2__UNUSED_1_MASK, DIDT_SQ_CTRL2__UNUSED_1__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_CTRL2, DIDT_SQ_CTRL2__LONG_TERM_INTERVAL_RATIO_MASK, DIDT_SQ_CTRL2__LONG_TERM_INTERVAL_RATIO__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_CTRL2, DIDT_SQ_CTRL2__UNUSED_2_MASK, DIDT_SQ_CTRL2__UNUSED_2__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_SQ_STALL_CTRL, DIDT_SQ_STALL_CTRL__DIDT_STALL_CTRL_ENABLE_MASK, DIDT_SQ_STALL_CTRL__DIDT_STALL_CTRL_ENABLE__SHIFT, 0x0001, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_STALL_CTRL, DIDT_SQ_STALL_CTRL__DIDT_STALL_DELAY_HI_MASK, DIDT_SQ_STALL_CTRL__DIDT_STALL_DELAY_HI__SHIFT, 0x0001, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_STALL_CTRL, DIDT_SQ_STALL_CTRL__DIDT_STALL_DELAY_LO_MASK, DIDT_SQ_STALL_CTRL__DIDT_STALL_DELAY_LO__SHIFT, 0x0001, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_STALL_CTRL, DIDT_SQ_STALL_CTRL__DIDT_HI_POWER_THRESHOLD_MASK, DIDT_SQ_STALL_CTRL__DIDT_HI_POWER_THRESHOLD__SHIFT, 0x0ebb, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_STALL_CTRL, DIDT_SQ_STALL_CTRL__UNUSED_0_MASK, DIDT_SQ_STALL_CTRL__UNUSED_0__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_SQ_TUNING_CTRL, DIDT_SQ_TUNING_CTRL__DIDT_TUNING_ENABLE_MASK, DIDT_SQ_TUNING_CTRL__DIDT_TUNING_ENABLE__SHIFT, 0x0001, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_TUNING_CTRL, DIDT_SQ_TUNING_CTRL__MAX_POWER_DELTA_HI_MASK, DIDT_SQ_TUNING_CTRL__MAX_POWER_DELTA_HI__SHIFT, 0x3853, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_TUNING_CTRL, DIDT_SQ_TUNING_CTRL__MAX_POWER_DELTA_LO_MASK, DIDT_SQ_TUNING_CTRL__MAX_POWER_DELTA_LO__SHIFT, 0x3153, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_TUNING_CTRL, DIDT_SQ_TUNING_CTRL__UNUSED_0_MASK, DIDT_SQ_TUNING_CTRL__UNUSED_0__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_SQ_CTRL0, DIDT_SQ_CTRL0__DIDT_CTRL_EN_MASK, DIDT_SQ_CTRL0__DIDT_CTRL_EN__SHIFT, 0x0001, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_CTRL0, DIDT_SQ_CTRL0__USE_REF_CLOCK_MASK, DIDT_SQ_CTRL0__USE_REF_CLOCK__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_CTRL0, DIDT_SQ_CTRL0__PHASE_OFFSET_MASK, DIDT_SQ_CTRL0__PHASE_OFFSET__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_CTRL0, DIDT_SQ_CTRL0__DIDT_CTRL_RST_MASK, DIDT_SQ_CTRL0__DIDT_CTRL_RST__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_CTRL0, DIDT_SQ_CTRL0__DIDT_CLK_EN_OVERRIDE_MASK, DIDT_SQ_CTRL0__DIDT_CLK_EN_OVERRIDE__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_CTRL0, DIDT_SQ_CTRL0__DIDT_MAX_STALLS_ALLOWED_HI_MASK, DIDT_SQ_CTRL0__DIDT_MAX_STALLS_ALLOWED_HI__SHIFT, 0x0010, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_CTRL0, DIDT_SQ_CTRL0__DIDT_MAX_STALLS_ALLOWED_LO_MASK, DIDT_SQ_CTRL0__DIDT_MAX_STALLS_ALLOWED_LO__SHIFT, 0x0010, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_SQ_CTRL0, DIDT_SQ_CTRL0__UNUSED_0_MASK, DIDT_SQ_CTRL0__UNUSED_0__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_TD_WEIGHT0_3, DIDT_TD_WEIGHT0_3__WEIGHT0_MASK, DIDT_TD_WEIGHT0_3__WEIGHT0__SHIFT, 0x000a, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_WEIGHT0_3, DIDT_TD_WEIGHT0_3__WEIGHT1_MASK, DIDT_TD_WEIGHT0_3__WEIGHT1__SHIFT, 0x0010, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_WEIGHT0_3, DIDT_TD_WEIGHT0_3__WEIGHT2_MASK, DIDT_TD_WEIGHT0_3__WEIGHT2__SHIFT, 0x0017, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_WEIGHT0_3, DIDT_TD_WEIGHT0_3__WEIGHT3_MASK, DIDT_TD_WEIGHT0_3__WEIGHT3__SHIFT, 0x002f, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_TD_WEIGHT4_7, DIDT_TD_WEIGHT4_7__WEIGHT4_MASK, DIDT_TD_WEIGHT4_7__WEIGHT4__SHIFT, 0x0046, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_WEIGHT4_7, DIDT_TD_WEIGHT4_7__WEIGHT5_MASK, DIDT_TD_WEIGHT4_7__WEIGHT5__SHIFT, 0x005d, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_WEIGHT4_7, DIDT_TD_WEIGHT4_7__WEIGHT6_MASK, DIDT_TD_WEIGHT4_7__WEIGHT6__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_WEIGHT4_7, DIDT_TD_WEIGHT4_7__WEIGHT7_MASK, DIDT_TD_WEIGHT4_7__WEIGHT7__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_TD_CTRL1, DIDT_TD_CTRL1__MIN_POWER_MASK, DIDT_TD_CTRL1__MIN_POWER__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_CTRL1, DIDT_TD_CTRL1__MAX_POWER_MASK, DIDT_TD_CTRL1__MAX_POWER__SHIFT, 0xffff, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_TD_CTRL_OCP, DIDT_TD_CTRL_OCP__UNUSED_0_MASK, DIDT_TD_CTRL_OCP__UNUSED_0__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_CTRL_OCP, DIDT_TD_CTRL_OCP__OCP_MAX_POWER_MASK, DIDT_TD_CTRL_OCP__OCP_MAX_POWER__SHIFT, 0x00ff, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_TD_CTRL2, DIDT_TD_CTRL2__MAX_POWER_DELTA_MASK, DIDT_TD_CTRL2__MAX_POWER_DELTA__SHIFT, 0x3fff, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_CTRL2, DIDT_TD_CTRL2__UNUSED_0_MASK, DIDT_TD_CTRL2__UNUSED_0__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_CTRL2, DIDT_TD_CTRL2__SHORT_TERM_INTERVAL_SIZE_MASK, DIDT_TD_CTRL2__SHORT_TERM_INTERVAL_SIZE__SHIFT, 0x000f, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_CTRL2, DIDT_TD_CTRL2__UNUSED_1_MASK, DIDT_TD_CTRL2__UNUSED_1__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_CTRL2, DIDT_TD_CTRL2__LONG_TERM_INTERVAL_RATIO_MASK, DIDT_TD_CTRL2__LONG_TERM_INTERVAL_RATIO__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_CTRL2, DIDT_TD_CTRL2__UNUSED_2_MASK, DIDT_TD_CTRL2__UNUSED_2__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_TD_STALL_CTRL, DIDT_TD_STALL_CTRL__DIDT_STALL_CTRL_ENABLE_MASK, DIDT_TD_STALL_CTRL__DIDT_STALL_CTRL_ENABLE__SHIFT, 0x0001, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_STALL_CTRL, DIDT_TD_STALL_CTRL__DIDT_STALL_DELAY_HI_MASK, DIDT_TD_STALL_CTRL__DIDT_STALL_DELAY_HI__SHIFT, 0x0001, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_STALL_CTRL, DIDT_TD_STALL_CTRL__DIDT_STALL_DELAY_LO_MASK, DIDT_TD_STALL_CTRL__DIDT_STALL_DELAY_LO__SHIFT, 0x0001, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_STALL_CTRL, DIDT_TD_STALL_CTRL__DIDT_HI_POWER_THRESHOLD_MASK, DIDT_TD_STALL_CTRL__DIDT_HI_POWER_THRESHOLD__SHIFT, 0x01aa, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_STALL_CTRL, DIDT_TD_STALL_CTRL__UNUSED_0_MASK, DIDT_TD_STALL_CTRL__UNUSED_0__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_TD_TUNING_CTRL, DIDT_TD_TUNING_CTRL__DIDT_TUNING_ENABLE_MASK, DIDT_TD_TUNING_CTRL__DIDT_TUNING_ENABLE__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_TUNING_CTRL, DIDT_TD_TUNING_CTRL__MAX_POWER_DELTA_HI_MASK, DIDT_TD_TUNING_CTRL__MAX_POWER_DELTA_HI__SHIFT, 0x0dde, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_TUNING_CTRL, DIDT_TD_TUNING_CTRL__MAX_POWER_DELTA_LO_MASK, DIDT_TD_TUNING_CTRL__MAX_POWER_DELTA_LO__SHIFT, 0x0dde, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_TUNING_CTRL, DIDT_TD_TUNING_CTRL__UNUSED_0_MASK, DIDT_TD_TUNING_CTRL__UNUSED_0__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_TD_CTRL0, DIDT_TD_CTRL0__DIDT_CTRL_EN_MASK, DIDT_TD_CTRL0__DIDT_CTRL_EN__SHIFT, 0x0001, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_CTRL0, DIDT_TD_CTRL0__USE_REF_CLOCK_MASK, DIDT_TD_CTRL0__USE_REF_CLOCK__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_CTRL0, DIDT_TD_CTRL0__PHASE_OFFSET_MASK, DIDT_TD_CTRL0__PHASE_OFFSET__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_CTRL0, DIDT_TD_CTRL0__DIDT_CTRL_RST_MASK, DIDT_TD_CTRL0__DIDT_CTRL_RST__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_CTRL0, DIDT_TD_CTRL0__DIDT_CLK_EN_OVERRIDE_MASK, DIDT_TD_CTRL0__DIDT_CLK_EN_OVERRIDE__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_CTRL0, DIDT_TD_CTRL0__DIDT_MAX_STALLS_ALLOWED_HI_MASK, DIDT_TD_CTRL0__DIDT_MAX_STALLS_ALLOWED_HI__SHIFT, 0x0008, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_CTRL0, DIDT_TD_CTRL0__DIDT_MAX_STALLS_ALLOWED_LO_MASK, DIDT_TD_CTRL0__DIDT_MAX_STALLS_ALLOWED_LO__SHIFT, 0x0008, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TD_CTRL0, DIDT_TD_CTRL0__UNUSED_0_MASK, DIDT_TD_CTRL0__UNUSED_0__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_TCP_WEIGHT0_3, DIDT_TCP_WEIGHT0_3__WEIGHT0_MASK, DIDT_TCP_WEIGHT0_3__WEIGHT0__SHIFT, 0x0004, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_WEIGHT0_3, DIDT_TCP_WEIGHT0_3__WEIGHT1_MASK, DIDT_TCP_WEIGHT0_3__WEIGHT1__SHIFT, 0x0037, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_WEIGHT0_3, DIDT_TCP_WEIGHT0_3__WEIGHT2_MASK, DIDT_TCP_WEIGHT0_3__WEIGHT2__SHIFT, 0x0001, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_WEIGHT0_3, DIDT_TCP_WEIGHT0_3__WEIGHT3_MASK, DIDT_TCP_WEIGHT0_3__WEIGHT3__SHIFT, 0x00ff, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_TCP_WEIGHT4_7, DIDT_TCP_WEIGHT4_7__WEIGHT4_MASK, DIDT_TCP_WEIGHT4_7__WEIGHT4__SHIFT, 0x0054, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_WEIGHT4_7, DIDT_TCP_WEIGHT4_7__WEIGHT5_MASK, DIDT_TCP_WEIGHT4_7__WEIGHT5__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_WEIGHT4_7, DIDT_TCP_WEIGHT4_7__WEIGHT6_MASK, DIDT_TCP_WEIGHT4_7__WEIGHT6__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_WEIGHT4_7, DIDT_TCP_WEIGHT4_7__WEIGHT7_MASK, DIDT_TCP_WEIGHT4_7__WEIGHT7__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_TCP_CTRL1, DIDT_TCP_CTRL1__MIN_POWER_MASK, DIDT_TCP_CTRL1__MIN_POWER__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_CTRL1, DIDT_TCP_CTRL1__MAX_POWER_MASK, DIDT_TCP_CTRL1__MAX_POWER__SHIFT, 0xffff, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_TCP_CTRL_OCP, DIDT_TCP_CTRL_OCP__UNUSED_0_MASK, DIDT_TCP_CTRL_OCP__UNUSED_0__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_CTRL_OCP, DIDT_TCP_CTRL_OCP__OCP_MAX_POWER_MASK, DIDT_TCP_CTRL_OCP__OCP_MAX_POWER__SHIFT, 0xffff, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_TCP_CTRL2, DIDT_TCP_CTRL2__MAX_POWER_DELTA_MASK, DIDT_TCP_CTRL2__MAX_POWER_DELTA__SHIFT, 0x3dde, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_CTRL2, DIDT_TCP_CTRL2__UNUSED_0_MASK, DIDT_TCP_CTRL2__UNUSED_0__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_CTRL2, DIDT_TCP_CTRL2__SHORT_TERM_INTERVAL_SIZE_MASK, DIDT_TCP_CTRL2__SHORT_TERM_INTERVAL_SIZE__SHIFT, 0x0032, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_CTRL2, DIDT_TCP_CTRL2__UNUSED_1_MASK, DIDT_TCP_CTRL2__UNUSED_1__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_CTRL2, DIDT_TCP_CTRL2__LONG_TERM_INTERVAL_RATIO_MASK, DIDT_TCP_CTRL2__LONG_TERM_INTERVAL_RATIO__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_CTRL2, DIDT_TCP_CTRL2__UNUSED_2_MASK, DIDT_TCP_CTRL2__UNUSED_2__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_TCP_STALL_CTRL, DIDT_TCP_STALL_CTRL__DIDT_STALL_CTRL_ENABLE_MASK, DIDT_TCP_STALL_CTRL__DIDT_STALL_CTRL_ENABLE__SHIFT, 0x0001, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_STALL_CTRL, DIDT_TCP_STALL_CTRL__DIDT_STALL_DELAY_HI_MASK, DIDT_TCP_STALL_CTRL__DIDT_STALL_DELAY_HI__SHIFT, 0x0001, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_STALL_CTRL, DIDT_TCP_STALL_CTRL__DIDT_STALL_DELAY_LO_MASK, DIDT_TCP_STALL_CTRL__DIDT_STALL_DELAY_LO__SHIFT, 0x0001, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_STALL_CTRL, DIDT_TCP_STALL_CTRL__DIDT_HI_POWER_THRESHOLD_MASK, DIDT_TCP_STALL_CTRL__DIDT_HI_POWER_THRESHOLD__SHIFT, 0x01aa, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_STALL_CTRL, DIDT_TCP_STALL_CTRL__UNUSED_0_MASK, DIDT_TCP_STALL_CTRL__UNUSED_0__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_TCP_TUNING_CTRL, DIDT_TCP_TUNING_CTRL__DIDT_TUNING_ENABLE_MASK, DIDT_TCP_TUNING_CTRL__DIDT_TUNING_ENABLE__SHIFT, 0x0001, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_TUNING_CTRL, DIDT_TCP_TUNING_CTRL__MAX_POWER_DELTA_HI_MASK, DIDT_TCP_TUNING_CTRL__MAX_POWER_DELTA_HI__SHIFT, 0x3dde, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_TUNING_CTRL, DIDT_TCP_TUNING_CTRL__MAX_POWER_DELTA_LO_MASK, DIDT_TCP_TUNING_CTRL__MAX_POWER_DELTA_LO__SHIFT, 0x3dde, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_TUNING_CTRL, DIDT_TCP_TUNING_CTRL__UNUSED_0_MASK, DIDT_TCP_TUNING_CTRL__UNUSED_0__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + + { ixDIDT_TCP_CTRL0, DIDT_TCP_CTRL0__DIDT_CTRL_EN_MASK, DIDT_TCP_CTRL0__DIDT_CTRL_EN__SHIFT, 0x0001, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_CTRL0, DIDT_TCP_CTRL0__USE_REF_CLOCK_MASK, DIDT_TCP_CTRL0__USE_REF_CLOCK__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_CTRL0, DIDT_TCP_CTRL0__PHASE_OFFSET_MASK, DIDT_TCP_CTRL0__PHASE_OFFSET__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_CTRL0, DIDT_TCP_CTRL0__DIDT_CTRL_RST_MASK, DIDT_TCP_CTRL0__DIDT_CTRL_RST__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_CTRL0, DIDT_TCP_CTRL0__DIDT_CLK_EN_OVERRIDE_MASK, DIDT_TCP_CTRL0__DIDT_CLK_EN_OVERRIDE__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_CTRL0, DIDT_TCP_CTRL0__DIDT_MAX_STALLS_ALLOWED_HI_MASK, DIDT_TCP_CTRL0__DIDT_MAX_STALLS_ALLOWED_HI__SHIFT, 0x0010, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_CTRL0, DIDT_TCP_CTRL0__DIDT_MAX_STALLS_ALLOWED_LO_MASK, DIDT_TCP_CTRL0__DIDT_MAX_STALLS_ALLOWED_LO__SHIFT, 0x0010, POLARIS10_CONFIGREG_DIDT_IND }, + { ixDIDT_TCP_CTRL0, DIDT_TCP_CTRL0__UNUSED_0_MASK, DIDT_TCP_CTRL0__UNUSED_0__SHIFT, 0x0000, POLARIS10_CONFIGREG_DIDT_IND }, + { 0xFFFFFFFF } +}; + static const struct polaris10_pt_defaults polaris10_power_tune_data_set_array[POWERTUNE_DEFAULT_SET_MAX] = { /* sviLoadLIneEn, SviLoadLineVddC, TDC_VDDC_ThrottleReleaseLimitPerc, TDC_MAWt, * TdcWaterfallCtl, DTEAmbientTempBase, DisplayCac, BAPM_TEMP_GRADIENT */ @@ -209,6 +559,187 @@ static int polaris10_min_max_vgnb_lpml_id_from_bapm_vddc(struct pp_hwmgr *hwmgr) return 0; } +static int polaris10_enable_didt(struct pp_hwmgr *hwmgr, const bool enable) +{ + + uint32_t en = enable ? 1 : 0; + int32_t result = 0; + uint32_t data; + + if (phm_cap_enabled(hwmgr->platform_descriptor.platformCaps, PHM_PlatformCaps_SQRamping)) { + data = cgs_read_ind_register(hwmgr->device, CGS_IND_REG__DIDT, ixDIDT_SQ_CTRL0); + data &= ~DIDT_SQ_CTRL0__DIDT_CTRL_EN_MASK; + data |= ((en << DIDT_SQ_CTRL0__DIDT_CTRL_EN__SHIFT) & DIDT_SQ_CTRL0__DIDT_CTRL_EN_MASK); + cgs_write_ind_register(hwmgr->device, CGS_IND_REG__DIDT, ixDIDT_SQ_CTRL0, data); + DIDTBlock_Info &= ~SQ_Enable_MASK; + DIDTBlock_Info |= en << SQ_Enable_SHIFT; + } + + if (phm_cap_enabled(hwmgr->platform_descriptor.platformCaps, PHM_PlatformCaps_DBRamping)) { + data = cgs_read_ind_register(hwmgr->device, CGS_IND_REG__DIDT, ixDIDT_DB_CTRL0); + data &= ~DIDT_DB_CTRL0__DIDT_CTRL_EN_MASK; + data |= ((en << DIDT_DB_CTRL0__DIDT_CTRL_EN__SHIFT) & DIDT_DB_CTRL0__DIDT_CTRL_EN_MASK); + cgs_write_ind_register(hwmgr->device, CGS_IND_REG__DIDT, ixDIDT_DB_CTRL0, data); + DIDTBlock_Info &= ~DB_Enable_MASK; + DIDTBlock_Info |= en << DB_Enable_SHIFT; + } + + if (phm_cap_enabled(hwmgr->platform_descriptor.platformCaps, PHM_PlatformCaps_TDRamping)) { + data = cgs_read_ind_register(hwmgr->device, CGS_IND_REG__DIDT, ixDIDT_TD_CTRL0); + data &= ~DIDT_TD_CTRL0__DIDT_CTRL_EN_MASK; + data |= ((en << DIDT_TD_CTRL0__DIDT_CTRL_EN__SHIFT) & DIDT_TD_CTRL0__DIDT_CTRL_EN_MASK); + cgs_write_ind_register(hwmgr->device, CGS_IND_REG__DIDT, ixDIDT_TD_CTRL0, data); + DIDTBlock_Info &= ~TD_Enable_MASK; + DIDTBlock_Info |= en << TD_Enable_SHIFT; + } + + if (phm_cap_enabled(hwmgr->platform_descriptor.platformCaps, PHM_PlatformCaps_TCPRamping)) { + data = cgs_read_ind_register(hwmgr->device, CGS_IND_REG__DIDT, ixDIDT_TCP_CTRL0); + data &= ~DIDT_TCP_CTRL0__DIDT_CTRL_EN_MASK; + data |= ((en << DIDT_TCP_CTRL0__DIDT_CTRL_EN__SHIFT) & DIDT_TCP_CTRL0__DIDT_CTRL_EN_MASK); + cgs_write_ind_register(hwmgr->device, CGS_IND_REG__DIDT, ixDIDT_TCP_CTRL0, data); + DIDTBlock_Info &= ~TCP_Enable_MASK; + DIDTBlock_Info |= en << TCP_Enable_SHIFT; + } + + if (enable) + result = smum_send_msg_to_smc_with_parameter(hwmgr->smumgr, PPSMC_MSG_Didt_Block_Function, DIDTBlock_Info); + + return result; +} + +static int polaris10_program_pt_config_registers(struct pp_hwmgr *hwmgr, + struct polaris10_pt_config_reg *cac_config_regs) +{ + struct polaris10_pt_config_reg *config_regs = cac_config_regs; + uint32_t cache = 0; + uint32_t data = 0; + + PP_ASSERT_WITH_CODE((config_regs != NULL), "Invalid config register table.", return -EINVAL); + + while (config_regs->offset != 0xFFFFFFFF) { + if (config_regs->type == POLARIS10_CONFIGREG_CACHE) + cache |= ((config_regs->value << config_regs->shift) & config_regs->mask); + else { + switch (config_regs->type) { + case POLARIS10_CONFIGREG_SMC_IND: + data = cgs_read_ind_register(hwmgr->device, CGS_IND_REG__SMC, config_regs->offset); + break; + + case POLARIS10_CONFIGREG_DIDT_IND: + data = cgs_read_ind_register(hwmgr->device, CGS_IND_REG__DIDT, config_regs->offset); + break; + + case POLARIS10_CONFIGREG_GC_CAC_IND: + data = cgs_read_ind_register(hwmgr->device, CGS_IND_REG_GC_CAC, config_regs->offset); + break; + + default: + data = cgs_read_register(hwmgr->device, config_regs->offset); + break; + } + + data &= ~config_regs->mask; + data |= ((config_regs->value << config_regs->shift) & config_regs->mask); + data |= cache; + + switch (config_regs->type) { + case POLARIS10_CONFIGREG_SMC_IND: + cgs_write_ind_register(hwmgr->device, CGS_IND_REG__SMC, config_regs->offset, data); + break; + + case POLARIS10_CONFIGREG_DIDT_IND: + cgs_write_ind_register(hwmgr->device, CGS_IND_REG__DIDT, config_regs->offset, data); + break; + + case POLARIS10_CONFIGREG_GC_CAC_IND: + cgs_write_ind_register(hwmgr->device, CGS_IND_REG_GC_CAC, config_regs->offset, data); + break; + + default: + cgs_write_register(hwmgr->device, config_regs->offset, data); + break; + } + cache = 0; + } + + config_regs++; + } + + return 0; +} + +int polaris10_enable_didt_config(struct pp_hwmgr *hwmgr) +{ + int result; + uint32_t num_se = 0; + uint32_t count, value, value2; + struct cgs_system_info sys_info = {0}; + + sys_info.size = sizeof(struct cgs_system_info); + sys_info.info_id = CGS_SYSTEM_INFO_GFX_SE_INFO; + result = cgs_query_system_info(hwmgr->device, &sys_info); + + + if (result == 0) + num_se = sys_info.value; + + if (phm_cap_enabled(hwmgr->platform_descriptor.platformCaps, PHM_PlatformCaps_SQRamping) || + phm_cap_enabled(hwmgr->platform_descriptor.platformCaps, PHM_PlatformCaps_DBRamping) || + phm_cap_enabled(hwmgr->platform_descriptor.platformCaps, PHM_PlatformCaps_TDRamping) || + phm_cap_enabled(hwmgr->platform_descriptor.platformCaps, PHM_PlatformCaps_TCPRamping)) { + + /* TO DO Pre DIDT disable clock gating */ + value = 0; + value2 = cgs_read_register(hwmgr->device, mmGRBM_GFX_INDEX); + for (count = 0; count < num_se; count++) { + value = SYS_GRBM_GFX_INDEX_DATA__INSTANCE_BROADCAST_WRITES_MASK + | SYS_GRBM_GFX_INDEX_DATA__SH_BROADCAST_WRITES_MASK + | (count << SYS_GRBM_GFX_INDEX_DATA__SE_INDEX__SHIFT); + cgs_write_register(hwmgr->device, mmGRBM_GFX_INDEX, value); + + if (hwmgr->chip_id == CHIP_POLARIS10) { + result = polaris10_program_pt_config_registers(hwmgr, GCCACConfig_Polaris10); + PP_ASSERT_WITH_CODE((result == 0), "DIDT Config failed.", return result); + result = polaris10_program_pt_config_registers(hwmgr, DIDTConfig_Polaris10); + PP_ASSERT_WITH_CODE((result == 0), "DIDT Config failed.", return result); + } else if (hwmgr->chip_id == CHIP_POLARIS11) { + result = polaris10_program_pt_config_registers(hwmgr, GCCACConfig_Polaris11); + PP_ASSERT_WITH_CODE((result == 0), "DIDT Config failed.", return result); + result = polaris10_program_pt_config_registers(hwmgr, DIDTConfig_Polaris11); + PP_ASSERT_WITH_CODE((result == 0), "DIDT Config failed.", return result); + } + } + cgs_write_register(hwmgr->device, mmGRBM_GFX_INDEX, value2); + + result = polaris10_enable_didt(hwmgr, true); + PP_ASSERT_WITH_CODE((result == 0), "EnableDiDt failed.", return result); + + /* TO DO Post DIDT enable clock gating */ + } + + return 0; +} + +int polaris10_disable_didt_config(struct pp_hwmgr *hwmgr) +{ + int result; + + if (phm_cap_enabled(hwmgr->platform_descriptor.platformCaps, PHM_PlatformCaps_SQRamping) || + phm_cap_enabled(hwmgr->platform_descriptor.platformCaps, PHM_PlatformCaps_DBRamping) || + phm_cap_enabled(hwmgr->platform_descriptor.platformCaps, PHM_PlatformCaps_TDRamping) || + phm_cap_enabled(hwmgr->platform_descriptor.platformCaps, PHM_PlatformCaps_TCPRamping)) { + /* TO DO Pre DIDT disable clock gating */ + + result = polaris10_enable_didt(hwmgr, false); + PP_ASSERT_WITH_CODE((result == 0), "Post DIDT enable clock gating failed.", return result); + /* TO DO Post DIDT enable clock gating */ + } + + return 0; +} + + static int polaris10_populate_bapm_vddc_base_leakage_sidd(struct pp_hwmgr *hwmgr) { struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend); @@ -286,7 +817,7 @@ int polaris10_populate_pm_fuses(struct pp_hwmgr *hwmgr) if (polaris10_copy_bytes_to_smc(hwmgr->smumgr, pm_fuse_table_offset, (uint8_t *)&data->power_tune_table, - sizeof(struct SMU74_Discrete_PmFuses), data->sram_end)) + (sizeof(struct SMU74_Discrete_PmFuses) - 92), data->sram_end)) PP_ASSERT_WITH_CODE(false, "Attempt to download PmFuseTable Failed!", return -EINVAL); @@ -312,6 +843,23 @@ int polaris10_enable_smc_cac(struct pp_hwmgr *hwmgr) return result; } +int polaris10_disable_smc_cac(struct pp_hwmgr *hwmgr) +{ + struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend); + int result = 0; + + if (phm_cap_enabled(hwmgr->platform_descriptor.platformCaps, + PHM_PlatformCaps_CAC) && data->cac_enabled) { + int smc_result = smum_send_msg_to_smc(hwmgr->smumgr, + (uint16_t)(PPSMC_MSG_DisableCac)); + PP_ASSERT_WITH_CODE((smc_result == 0), + "Failed to disable CAC in SMC.", result = -1); + + data->cac_enabled = false; + } + return result; +} + int polaris10_set_power_limit(struct pp_hwmgr *hwmgr, uint32_t n) { struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend); @@ -373,6 +921,48 @@ int polaris10_enable_power_containment(struct pp_hwmgr *hwmgr) return result; } +int polaris10_disable_power_containment(struct pp_hwmgr *hwmgr) +{ + struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend); + int result = 0; + + if (phm_cap_enabled(hwmgr->platform_descriptor.platformCaps, + PHM_PlatformCaps_PowerContainment) && + data->power_containment_features) { + int smc_result; + + if (data->power_containment_features & + POWERCONTAINMENT_FEATURE_TDCLimit) { + smc_result = smum_send_msg_to_smc(hwmgr->smumgr, + (uint16_t)(PPSMC_MSG_TDCLimitDisable)); + PP_ASSERT_WITH_CODE((smc_result == 0), + "Failed to disable TDCLimit in SMC.", + result = smc_result); + } + + if (data->power_containment_features & + POWERCONTAINMENT_FEATURE_DTE) { + smc_result = smum_send_msg_to_smc(hwmgr->smumgr, + (uint16_t)(PPSMC_MSG_DisableDTE)); + PP_ASSERT_WITH_CODE((smc_result == 0), + "Failed to disable DTE in SMC.", + result = smc_result); + } + + if (data->power_containment_features & + POWERCONTAINMENT_FEATURE_PkgPwrLimit) { + smc_result = smum_send_msg_to_smc(hwmgr->smumgr, + (uint16_t)(PPSMC_MSG_PkgPwrLimitDisable)); + PP_ASSERT_WITH_CODE((smc_result == 0), + "Failed to disable PkgPwrTracking in SMC.", + result = smc_result); + } + data->power_containment_features = 0; + } + + return result; +} + int polaris10_power_control_set_level(struct pp_hwmgr *hwmgr) { struct phm_ppt_v1_information *table_info = diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_powertune.h b/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_powertune.h index 68bc1cb6d40c..bc78e28f010d 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_powertune.h +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_powertune.h @@ -27,15 +27,37 @@ enum polaris10_pt_config_reg_type { POLARIS10_CONFIGREG_MMR = 0, POLARIS10_CONFIGREG_SMC_IND, POLARIS10_CONFIGREG_DIDT_IND, + POLARIS10_CONFIGREG_GC_CAC_IND, POLARIS10_CONFIGREG_CACHE, POLARIS10_CONFIGREG_MAX }; +#define DIDT_SQ_CTRL0__UNUSED_0_MASK 0xfffc0000 +#define DIDT_SQ_CTRL0__UNUSED_0__SHIFT 0x12 +#define DIDT_TD_CTRL0__UNUSED_0_MASK 0xfffc0000 +#define DIDT_TD_CTRL0__UNUSED_0__SHIFT 0x12 +#define DIDT_TCP_CTRL0__UNUSED_0_MASK 0xfffc0000 +#define DIDT_TCP_CTRL0__UNUSED_0__SHIFT 0x12 +#define DIDT_SQ_TUNING_CTRL__UNUSED_0_MASK 0xc0000000 +#define DIDT_SQ_TUNING_CTRL__UNUSED_0__SHIFT 0x0000001e +#define DIDT_TD_TUNING_CTRL__UNUSED_0_MASK 0xc0000000 +#define DIDT_TD_TUNING_CTRL__UNUSED_0__SHIFT 0x0000001e +#define DIDT_TCP_TUNING_CTRL__UNUSED_0_MASK 0xc0000000 +#define DIDT_TCP_TUNING_CTRL__UNUSED_0__SHIFT 0x0000001e + /* PowerContainment Features */ #define POWERCONTAINMENT_FEATURE_DTE 0x00000001 #define POWERCONTAINMENT_FEATURE_TDCLimit 0x00000002 #define POWERCONTAINMENT_FEATURE_PkgPwrLimit 0x00000004 +#define ixGC_CAC_CNTL 0x0000 +#define ixDIDT_SQ_STALL_CTRL 0x0004 +#define ixDIDT_SQ_TUNING_CTRL 0x0005 +#define ixDIDT_TD_STALL_CTRL 0x0044 +#define ixDIDT_TD_TUNING_CTRL 0x0045 +#define ixDIDT_TCP_STALL_CTRL 0x0064 +#define ixDIDT_TCP_TUNING_CTRL 0x0065 + struct polaris10_pt_config_reg { uint32_t offset; uint32_t mask; @@ -62,9 +84,11 @@ void polaris10_initialize_power_tune_defaults(struct pp_hwmgr *hwmgr); int polaris10_populate_bapm_parameters_in_dpm_table(struct pp_hwmgr *hwmgr); int polaris10_populate_pm_fuses(struct pp_hwmgr *hwmgr); int polaris10_enable_smc_cac(struct pp_hwmgr *hwmgr); +int polaris10_disable_smc_cac(struct pp_hwmgr *hwmgr); int polaris10_enable_power_containment(struct pp_hwmgr *hwmgr); +int polaris10_disable_power_containment(struct pp_hwmgr *hwmgr); int polaris10_set_power_limit(struct pp_hwmgr *hwmgr, uint32_t n); int polaris10_power_control_set_level(struct pp_hwmgr *hwmgr); - +int polaris10_enable_didt_config(struct pp_hwmgr *hwmgr); #endif /* POLARIS10_POWERTUNE_H */ diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_thermal.c b/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_thermal.c index aba167f7d167..b206632d4650 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_thermal.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_thermal.c @@ -625,10 +625,14 @@ static int tf_polaris10_thermal_avfs_enable(struct pp_hwmgr *hwmgr, int ret; struct pp_smumgr *smumgr = (struct pp_smumgr *)(hwmgr->smumgr); struct polaris10_smumgr *smu_data = (struct polaris10_smumgr *)(smumgr->backend); + struct polaris10_hwmgr *data = (struct polaris10_hwmgr *)(hwmgr->backend); - if (smu_data->avfs.avfs_btc_status != AVFS_BTC_ENABLEAVFS) + if (smu_data->avfs.avfs_btc_status == AVFS_BTC_NOTSUPPORTED) return 0; + ret = smum_send_msg_to_smc_with_parameter(hwmgr->smumgr, + PPSMC_MSG_SetGBDroopSettings, data->avfs_vdroop_override_setting); + ret = (smum_send_msg_to_smc(smumgr, PPSMC_MSG_EnableAvfs) == 0) ? 0 : -1; diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/pp_acpi.c b/drivers/gpu/drm/amd/powerplay/hwmgr/pp_acpi.c index 58742e0d1492..1944d289f846 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/pp_acpi.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/pp_acpi.c @@ -44,6 +44,20 @@ bool acpi_atcs_functions_supported(void *device, uint32_t index) return result == 0 ? (output_buf.function_bits & (1 << (index - 1))) != 0 : false; } +bool acpi_atcs_notify_pcie_device_ready(void *device) +{ + int32_t temp_buffer = 1; + + return cgs_call_acpi_method(device, CGS_ACPI_METHOD_ATCS, + ATCS_FUNCTION_PCIE_DEVICE_READY_NOTIFICATION, + &temp_buffer, + NULL, + 0, + sizeof(temp_buffer), + 0); +} + + int acpi_pcie_perf_request(void *device, uint8_t perf_req, bool advertise) { struct atcs_pref_req_input atcs_input; @@ -52,7 +66,7 @@ int acpi_pcie_perf_request(void *device, uint8_t perf_req, bool advertise) int result; struct cgs_system_info info = {0}; - if (!acpi_atcs_functions_supported(device, ATCS_FUNCTION_PCIE_PERFORMANCE_REQUEST)) + if (acpi_atcs_notify_pcie_device_ready(device)) return -EINVAL; info.size = sizeof(struct cgs_system_info); @@ -77,7 +91,7 @@ int acpi_pcie_perf_request(void *device, uint8_t perf_req, bool advertise) ATCS_FUNCTION_PCIE_PERFORMANCE_REQUEST, &atcs_input, &atcs_output, - 0, + 1, sizeof(atcs_input), sizeof(atcs_output)); if (result != 0) diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/ppatomctrl.c b/drivers/gpu/drm/amd/powerplay/hwmgr/ppatomctrl.c index da9f5f1b6dc2..26f3e30d0fef 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/ppatomctrl.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/ppatomctrl.c @@ -179,13 +179,12 @@ int atomctrl_set_engine_dram_timings_rv770( /* They are both in 10KHz Units. */ engine_clock_parameters.ulTargetEngineClock = - (uint32_t) engine_clock & SET_CLOCK_FREQ_MASK; - engine_clock_parameters.ulTargetEngineClock |= - (COMPUTE_ENGINE_PLL_PARAM << 24); + cpu_to_le32((engine_clock & SET_CLOCK_FREQ_MASK) | + ((COMPUTE_ENGINE_PLL_PARAM << 24))); /* in 10 khz units.*/ engine_clock_parameters.sReserved.ulClock = - (uint32_t) memory_clock & SET_CLOCK_FREQ_MASK; + cpu_to_le32(memory_clock & SET_CLOCK_FREQ_MASK); return cgs_atom_exec_cmd_table(hwmgr->device, GetIndexIntoMasterTable(COMMAND, DynamicMemorySettings), &engine_clock_parameters); @@ -252,7 +251,7 @@ int atomctrl_get_memory_pll_dividers_si( COMPUTE_MEMORY_CLOCK_PARAM_PARAMETERS_V2_1 mpll_parameters; int result; - mpll_parameters.ulClock = (uint32_t) clock_value; + mpll_parameters.ulClock = cpu_to_le32(clock_value); mpll_parameters.ucInputFlag = (uint8_t)((strobe_mode) ? 1 : 0); result = cgs_atom_exec_cmd_table @@ -262,9 +261,9 @@ int atomctrl_get_memory_pll_dividers_si( if (0 == result) { mpll_param->mpll_fb_divider.clk_frac = - mpll_parameters.ulFbDiv.usFbDivFrac; + le16_to_cpu(mpll_parameters.ulFbDiv.usFbDivFrac); mpll_param->mpll_fb_divider.cl_kf = - mpll_parameters.ulFbDiv.usFbDiv; + le16_to_cpu(mpll_parameters.ulFbDiv.usFbDiv); mpll_param->mpll_post_divider = (uint32_t)mpll_parameters.ucPostDiv; mpll_param->vco_mode = @@ -300,7 +299,7 @@ int atomctrl_get_memory_pll_dividers_vi(struct pp_hwmgr *hwmgr, COMPUTE_MEMORY_CLOCK_PARAM_PARAMETERS_V2_2 mpll_parameters; int result; - mpll_parameters.ulClock.ulClock = (uint32_t)clock_value; + mpll_parameters.ulClock.ulClock = cpu_to_le32(clock_value); result = cgs_atom_exec_cmd_table(hwmgr->device, GetIndexIntoMasterTable(COMMAND, ComputeMemoryClockParam), @@ -320,7 +319,7 @@ int atomctrl_get_engine_pll_dividers_kong(struct pp_hwmgr *hwmgr, COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_V4 pll_parameters; int result; - pll_parameters.ulClock = clock_value; + pll_parameters.ulClock = cpu_to_le32(clock_value); result = cgs_atom_exec_cmd_table (hwmgr->device, @@ -329,7 +328,7 @@ int atomctrl_get_engine_pll_dividers_kong(struct pp_hwmgr *hwmgr, if (0 == result) { dividers->pll_post_divider = pll_parameters.ucPostDiv; - dividers->real_clock = pll_parameters.ulClock; + dividers->real_clock = le32_to_cpu(pll_parameters.ulClock); } return result; @@ -343,7 +342,7 @@ int atomctrl_get_engine_pll_dividers_vi( COMPUTE_GPU_CLOCK_OUTPUT_PARAMETERS_V1_6 pll_patameters; int result; - pll_patameters.ulClock.ulClock = clock_value; + pll_patameters.ulClock.ulClock = cpu_to_le32(clock_value); pll_patameters.ulClock.ucPostDiv = COMPUTE_GPUCLK_INPUT_FLAG_SCLK; result = cgs_atom_exec_cmd_table @@ -355,12 +354,12 @@ int atomctrl_get_engine_pll_dividers_vi( dividers->pll_post_divider = pll_patameters.ulClock.ucPostDiv; dividers->real_clock = - pll_patameters.ulClock.ulClock; + le32_to_cpu(pll_patameters.ulClock.ulClock); dividers->ul_fb_div.ul_fb_div_frac = - pll_patameters.ulFbDiv.usFbDivFrac; + le16_to_cpu(pll_patameters.ulFbDiv.usFbDivFrac); dividers->ul_fb_div.ul_fb_div = - pll_patameters.ulFbDiv.usFbDiv; + le16_to_cpu(pll_patameters.ulFbDiv.usFbDiv); dividers->uc_pll_ref_div = pll_patameters.ucPllRefDiv; @@ -380,7 +379,7 @@ int atomctrl_get_engine_pll_dividers_ai(struct pp_hwmgr *hwmgr, COMPUTE_GPU_CLOCK_OUTPUT_PARAMETERS_V1_7 pll_patameters; int result; - pll_patameters.ulClock.ulClock = clock_value; + pll_patameters.ulClock.ulClock = cpu_to_le32(clock_value); pll_patameters.ulClock.ucPostDiv = COMPUTE_GPUCLK_INPUT_FLAG_SCLK; result = cgs_atom_exec_cmd_table @@ -412,7 +411,7 @@ int atomctrl_get_dfs_pll_dividers_vi( COMPUTE_GPU_CLOCK_OUTPUT_PARAMETERS_V1_6 pll_patameters; int result; - pll_patameters.ulClock.ulClock = clock_value; + pll_patameters.ulClock.ulClock = cpu_to_le32(clock_value); pll_patameters.ulClock.ucPostDiv = COMPUTE_GPUCLK_INPUT_FLAG_DEFAULT_GPUCLK; @@ -425,12 +424,12 @@ int atomctrl_get_dfs_pll_dividers_vi( dividers->pll_post_divider = pll_patameters.ulClock.ucPostDiv; dividers->real_clock = - pll_patameters.ulClock.ulClock; + le32_to_cpu(pll_patameters.ulClock.ulClock); dividers->ul_fb_div.ul_fb_div_frac = - pll_patameters.ulFbDiv.usFbDivFrac; + le16_to_cpu(pll_patameters.ulFbDiv.usFbDivFrac); dividers->ul_fb_div.ul_fb_div = - pll_patameters.ulFbDiv.usFbDiv; + le16_to_cpu(pll_patameters.ulFbDiv.usFbDiv); dividers->uc_pll_ref_div = pll_patameters.ucPllRefDiv; @@ -519,13 +518,13 @@ int atomctrl_get_voltage_table_v3( for (i = 0; i < voltage_object->asGpioVoltageObj.ucGpioEntryNum; i++) { voltage_table->entries[i].value = - voltage_object->asGpioVoltageObj.asVolGpioLut[i].usVoltageValue; + le16_to_cpu(voltage_object->asGpioVoltageObj.asVolGpioLut[i].usVoltageValue); voltage_table->entries[i].smio_low = - voltage_object->asGpioVoltageObj.asVolGpioLut[i].ulVoltageId; + le32_to_cpu(voltage_object->asGpioVoltageObj.asVolGpioLut[i].ulVoltageId); } voltage_table->mask_low = - voltage_object->asGpioVoltageObj.ulGpioMaskVal; + le32_to_cpu(voltage_object->asGpioVoltageObj.ulGpioMaskVal); voltage_table->count = voltage_object->asGpioVoltageObj.ucGpioEntryNum; voltage_table->phase_delay = @@ -552,13 +551,13 @@ static bool atomctrl_lookup_gpio_pin( pin_assignment->ucGpioPinBitShift; gpio_pin_assignment->us_gpio_pin_aindex = le16_to_cpu(pin_assignment->usGpioPin_AIndex); - return false; + return true; } offset += offsetof(ATOM_GPIO_PIN_ASSIGNMENT, ucGPIO_ID) + 1; } - return true; + return false; } /** @@ -592,12 +591,12 @@ bool atomctrl_get_pp_assign_pin( const uint32_t pinId, pp_atomctrl_gpio_pin_assignment *gpio_pin_assignment) { - bool bRet = 0; + bool bRet = false; ATOM_GPIO_PIN_LUT *gpio_lookup_table = get_gpio_lookup_table(hwmgr->device); PP_ASSERT_WITH_CODE((NULL != gpio_lookup_table), - "Could not find GPIO lookup Table in BIOS.", return -1); + "Could not find GPIO lookup Table in BIOS.", return false); bRet = atomctrl_lookup_gpio_pin(gpio_lookup_table, pinId, gpio_pin_assignment); @@ -650,8 +649,8 @@ int atomctrl_calculate_voltage_evv_on_sclk( return -1; if (getASICProfilingInfo->asHeader.ucTableFormatRevision < 3 || - (getASICProfilingInfo->asHeader.ucTableFormatRevision == 3 && - getASICProfilingInfo->asHeader.ucTableContentRevision < 4)) + (getASICProfilingInfo->asHeader.ucTableFormatRevision == 3 && + getASICProfilingInfo->asHeader.ucTableContentRevision < 4)) return -1; /*----------------------------------------------------------- @@ -662,37 +661,37 @@ int atomctrl_calculate_voltage_evv_on_sclk( switch (dpm_level) { case 1: - fPowerDPMx = Convert_ULONG_ToFraction(getASICProfilingInfo->usPowerDpm1); - fDerateTDP = GetScaledFraction(getASICProfilingInfo->ulTdpDerateDPM1, 1000); + fPowerDPMx = Convert_ULONG_ToFraction(le16_to_cpu(getASICProfilingInfo->usPowerDpm1)); + fDerateTDP = GetScaledFraction(le32_to_cpu(getASICProfilingInfo->ulTdpDerateDPM1), 1000); break; case 2: - fPowerDPMx = Convert_ULONG_ToFraction(getASICProfilingInfo->usPowerDpm2); - fDerateTDP = GetScaledFraction(getASICProfilingInfo->ulTdpDerateDPM2, 1000); + fPowerDPMx = Convert_ULONG_ToFraction(le16_to_cpu(getASICProfilingInfo->usPowerDpm2)); + fDerateTDP = GetScaledFraction(le32_to_cpu(getASICProfilingInfo->ulTdpDerateDPM2), 1000); break; case 3: - fPowerDPMx = Convert_ULONG_ToFraction(getASICProfilingInfo->usPowerDpm3); - fDerateTDP = GetScaledFraction(getASICProfilingInfo->ulTdpDerateDPM3, 1000); + fPowerDPMx = Convert_ULONG_ToFraction(le16_to_cpu(getASICProfilingInfo->usPowerDpm3)); + fDerateTDP = GetScaledFraction(le32_to_cpu(getASICProfilingInfo->ulTdpDerateDPM3), 1000); break; case 4: - fPowerDPMx = Convert_ULONG_ToFraction(getASICProfilingInfo->usPowerDpm4); - fDerateTDP = GetScaledFraction(getASICProfilingInfo->ulTdpDerateDPM4, 1000); + fPowerDPMx = Convert_ULONG_ToFraction(le16_to_cpu(getASICProfilingInfo->usPowerDpm4)); + fDerateTDP = GetScaledFraction(le32_to_cpu(getASICProfilingInfo->ulTdpDerateDPM4), 1000); break; case 5: - fPowerDPMx = Convert_ULONG_ToFraction(getASICProfilingInfo->usPowerDpm5); - fDerateTDP = GetScaledFraction(getASICProfilingInfo->ulTdpDerateDPM5, 1000); + fPowerDPMx = Convert_ULONG_ToFraction(le16_to_cpu(getASICProfilingInfo->usPowerDpm5)); + fDerateTDP = GetScaledFraction(le32_to_cpu(getASICProfilingInfo->ulTdpDerateDPM5), 1000); break; case 6: - fPowerDPMx = Convert_ULONG_ToFraction(getASICProfilingInfo->usPowerDpm6); - fDerateTDP = GetScaledFraction(getASICProfilingInfo->ulTdpDerateDPM6, 1000); + fPowerDPMx = Convert_ULONG_ToFraction(le16_to_cpu(getASICProfilingInfo->usPowerDpm6)); + fDerateTDP = GetScaledFraction(le32_to_cpu(getASICProfilingInfo->ulTdpDerateDPM6), 1000); break; case 7: - fPowerDPMx = Convert_ULONG_ToFraction(getASICProfilingInfo->usPowerDpm7); - fDerateTDP = GetScaledFraction(getASICProfilingInfo->ulTdpDerateDPM7, 1000); + fPowerDPMx = Convert_ULONG_ToFraction(le16_to_cpu(getASICProfilingInfo->usPowerDpm7)); + fDerateTDP = GetScaledFraction(le32_to_cpu(getASICProfilingInfo->ulTdpDerateDPM7), 1000); break; default: printk(KERN_ERR "DPM Level not supported\n"); fPowerDPMx = Convert_ULONG_ToFraction(1); - fDerateTDP = GetScaledFraction(getASICProfilingInfo->ulTdpDerateDPM0, 1000); + fDerateTDP = GetScaledFraction(le32_to_cpu(getASICProfilingInfo->ulTdpDerateDPM0), 1000); } /*------------------------- @@ -716,9 +715,9 @@ int atomctrl_calculate_voltage_evv_on_sclk( return result; /* Finally, the actual fuse value */ - ul_RO_fused = sOutput_FuseValues.ulEfuseValue; - fMin = GetScaledFraction(sRO_fuse.ulEfuseMin, 1); - fRange = GetScaledFraction(sRO_fuse.ulEfuseEncodeRange, 1); + ul_RO_fused = le32_to_cpu(sOutput_FuseValues.ulEfuseValue); + fMin = GetScaledFraction(le32_to_cpu(sRO_fuse.ulEfuseMin), 1); + fRange = GetScaledFraction(le32_to_cpu(sRO_fuse.ulEfuseEncodeRange), 1); fRO_fused = fDecodeLinearFuse(ul_RO_fused, fMin, fRange, sRO_fuse.ucEfuseLength); sCACm_fuse = getASICProfilingInfo->sCACm; @@ -736,9 +735,9 @@ int atomctrl_calculate_voltage_evv_on_sclk( if (result) return result; - ul_CACm_fused = sOutput_FuseValues.ulEfuseValue; - fMin = GetScaledFraction(sCACm_fuse.ulEfuseMin, 1000); - fRange = GetScaledFraction(sCACm_fuse.ulEfuseEncodeRange, 1000); + ul_CACm_fused = le32_to_cpu(sOutput_FuseValues.ulEfuseValue); + fMin = GetScaledFraction(le32_to_cpu(sCACm_fuse.ulEfuseMin), 1000); + fRange = GetScaledFraction(le32_to_cpu(sCACm_fuse.ulEfuseEncodeRange), 1000); fCACm_fused = fDecodeLinearFuse(ul_CACm_fused, fMin, fRange, sCACm_fuse.ucEfuseLength); @@ -756,9 +755,9 @@ int atomctrl_calculate_voltage_evv_on_sclk( if (result) return result; - ul_CACb_fused = sOutput_FuseValues.ulEfuseValue; - fMin = GetScaledFraction(sCACb_fuse.ulEfuseMin, 1000); - fRange = GetScaledFraction(sCACb_fuse.ulEfuseEncodeRange, 1000); + ul_CACb_fused = le32_to_cpu(sOutput_FuseValues.ulEfuseValue); + fMin = GetScaledFraction(le32_to_cpu(sCACb_fuse.ulEfuseMin), 1000); + fRange = GetScaledFraction(le32_to_cpu(sCACb_fuse.ulEfuseEncodeRange), 1000); fCACb_fused = fDecodeLinearFuse(ul_CACb_fused, fMin, fRange, sCACb_fuse.ucEfuseLength); @@ -777,9 +776,9 @@ int atomctrl_calculate_voltage_evv_on_sclk( if (result) return result; - ul_Kt_Beta_fused = sOutput_FuseValues.ulEfuseValue; - fAverage = GetScaledFraction(sKt_Beta_fuse.ulEfuseEncodeAverage, 1000); - fRange = GetScaledFraction(sKt_Beta_fuse.ulEfuseEncodeRange, 1000); + ul_Kt_Beta_fused = le32_to_cpu(sOutput_FuseValues.ulEfuseValue); + fAverage = GetScaledFraction(le32_to_cpu(sKt_Beta_fuse.ulEfuseEncodeAverage), 1000); + fRange = GetScaledFraction(le32_to_cpu(sKt_Beta_fuse.ulEfuseEncodeRange), 1000); fKt_Beta_fused = fDecodeLogisticFuse(ul_Kt_Beta_fused, fAverage, fRange, sKt_Beta_fuse.ucEfuseLength); @@ -798,9 +797,9 @@ int atomctrl_calculate_voltage_evv_on_sclk( if (result) return result; - ul_Kv_m_fused = sOutput_FuseValues.ulEfuseValue; - fAverage = GetScaledFraction(sKv_m_fuse.ulEfuseEncodeAverage, 1000); - fRange = GetScaledFraction((sKv_m_fuse.ulEfuseEncodeRange & 0x7fffffff), 1000); + ul_Kv_m_fused = le32_to_cpu(sOutput_FuseValues.ulEfuseValue); + fAverage = GetScaledFraction(le32_to_cpu(sKv_m_fuse.ulEfuseEncodeAverage), 1000); + fRange = GetScaledFraction((le32_to_cpu(sKv_m_fuse.ulEfuseEncodeRange) & 0x7fffffff), 1000); fRange = fMultiply(fRange, ConvertToFraction(-1)); fKv_m_fused = fDecodeLogisticFuse(ul_Kv_m_fused, @@ -820,9 +819,9 @@ int atomctrl_calculate_voltage_evv_on_sclk( if (result) return result; - ul_Kv_b_fused = sOutput_FuseValues.ulEfuseValue; - fAverage = GetScaledFraction(sKv_b_fuse.ulEfuseEncodeAverage, 1000); - fRange = GetScaledFraction(sKv_b_fuse.ulEfuseEncodeRange, 1000); + ul_Kv_b_fused = le32_to_cpu(sOutput_FuseValues.ulEfuseValue); + fAverage = GetScaledFraction(le32_to_cpu(sKv_b_fuse.ulEfuseEncodeAverage), 1000); + fRange = GetScaledFraction(le32_to_cpu(sKv_b_fuse.ulEfuseEncodeRange), 1000); fKv_b_fused = fDecodeLogisticFuse(ul_Kv_b_fused, fAverage, fRange, sKv_b_fuse.ucEfuseLength); @@ -851,9 +850,9 @@ int atomctrl_calculate_voltage_evv_on_sclk( if (result) return result; - ul_FT_Lkg_V0NORM = sOutput_FuseValues.ulEfuseValue; - fLn_MaxDivMin = GetScaledFraction(getASICProfilingInfo->ulLkgEncodeLn_MaxDivMin, 10000); - fMin = GetScaledFraction(getASICProfilingInfo->ulLkgEncodeMin, 10000); + ul_FT_Lkg_V0NORM = le32_to_cpu(sOutput_FuseValues.ulEfuseValue); + fLn_MaxDivMin = GetScaledFraction(le32_to_cpu(getASICProfilingInfo->ulLkgEncodeLn_MaxDivMin), 10000); + fMin = GetScaledFraction(le32_to_cpu(getASICProfilingInfo->ulLkgEncodeMin), 10000); fFT_Lkg_V0NORM = fDecodeLeakageID(ul_FT_Lkg_V0NORM, fLn_MaxDivMin, fMin, getASICProfilingInfo->ucLkgEfuseLength); @@ -863,40 +862,40 @@ int atomctrl_calculate_voltage_evv_on_sclk( * PART 2 - Grabbing all required values *------------------------------------------- */ - fSM_A0 = fMultiply(GetScaledFraction(getASICProfilingInfo->ulSM_A0, 1000000), + fSM_A0 = fMultiply(GetScaledFraction(le32_to_cpu(getASICProfilingInfo->ulSM_A0), 1000000), ConvertToFraction(uPow(-1, getASICProfilingInfo->ucSM_A0_sign))); - fSM_A1 = fMultiply(GetScaledFraction(getASICProfilingInfo->ulSM_A1, 1000000), + fSM_A1 = fMultiply(GetScaledFraction(le32_to_cpu(getASICProfilingInfo->ulSM_A1), 1000000), ConvertToFraction(uPow(-1, getASICProfilingInfo->ucSM_A1_sign))); - fSM_A2 = fMultiply(GetScaledFraction(getASICProfilingInfo->ulSM_A2, 100000), + fSM_A2 = fMultiply(GetScaledFraction(le32_to_cpu(getASICProfilingInfo->ulSM_A2), 100000), ConvertToFraction(uPow(-1, getASICProfilingInfo->ucSM_A2_sign))); - fSM_A3 = fMultiply(GetScaledFraction(getASICProfilingInfo->ulSM_A3, 1000000), + fSM_A3 = fMultiply(GetScaledFraction(le32_to_cpu(getASICProfilingInfo->ulSM_A3), 1000000), ConvertToFraction(uPow(-1, getASICProfilingInfo->ucSM_A3_sign))); - fSM_A4 = fMultiply(GetScaledFraction(getASICProfilingInfo->ulSM_A4, 1000000), + fSM_A4 = fMultiply(GetScaledFraction(le32_to_cpu(getASICProfilingInfo->ulSM_A4), 1000000), ConvertToFraction(uPow(-1, getASICProfilingInfo->ucSM_A4_sign))); - fSM_A5 = fMultiply(GetScaledFraction(getASICProfilingInfo->ulSM_A5, 1000), + fSM_A5 = fMultiply(GetScaledFraction(le32_to_cpu(getASICProfilingInfo->ulSM_A5), 1000), ConvertToFraction(uPow(-1, getASICProfilingInfo->ucSM_A5_sign))); - fSM_A6 = fMultiply(GetScaledFraction(getASICProfilingInfo->ulSM_A6, 1000), + fSM_A6 = fMultiply(GetScaledFraction(le32_to_cpu(getASICProfilingInfo->ulSM_A6), 1000), ConvertToFraction(uPow(-1, getASICProfilingInfo->ucSM_A6_sign))); - fSM_A7 = fMultiply(GetScaledFraction(getASICProfilingInfo->ulSM_A7, 1000), + fSM_A7 = fMultiply(GetScaledFraction(le32_to_cpu(getASICProfilingInfo->ulSM_A7), 1000), ConvertToFraction(uPow(-1, getASICProfilingInfo->ucSM_A7_sign))); - fMargin_RO_a = ConvertToFraction(getASICProfilingInfo->ulMargin_RO_a); - fMargin_RO_b = ConvertToFraction(getASICProfilingInfo->ulMargin_RO_b); - fMargin_RO_c = ConvertToFraction(getASICProfilingInfo->ulMargin_RO_c); + fMargin_RO_a = ConvertToFraction(le32_to_cpu(getASICProfilingInfo->ulMargin_RO_a)); + fMargin_RO_b = ConvertToFraction(le32_to_cpu(getASICProfilingInfo->ulMargin_RO_b)); + fMargin_RO_c = ConvertToFraction(le32_to_cpu(getASICProfilingInfo->ulMargin_RO_c)); - fMargin_fixed = ConvertToFraction(getASICProfilingInfo->ulMargin_fixed); + fMargin_fixed = ConvertToFraction(le32_to_cpu(getASICProfilingInfo->ulMargin_fixed)); fMargin_FMAX_mean = GetScaledFraction( - getASICProfilingInfo->ulMargin_Fmax_mean, 10000); + le32_to_cpu(getASICProfilingInfo->ulMargin_Fmax_mean), 10000); fMargin_Plat_mean = GetScaledFraction( - getASICProfilingInfo->ulMargin_plat_mean, 10000); + le32_to_cpu(getASICProfilingInfo->ulMargin_plat_mean), 10000); fMargin_FMAX_sigma = GetScaledFraction( - getASICProfilingInfo->ulMargin_Fmax_sigma, 10000); + le32_to_cpu(getASICProfilingInfo->ulMargin_Fmax_sigma), 10000); fMargin_Plat_sigma = GetScaledFraction( - getASICProfilingInfo->ulMargin_plat_sigma, 10000); + le32_to_cpu(getASICProfilingInfo->ulMargin_plat_sigma), 10000); fMargin_DC_sigma = GetScaledFraction( - getASICProfilingInfo->ulMargin_DC_sigma, 100); + le32_to_cpu(getASICProfilingInfo->ulMargin_DC_sigma), 100); fMargin_DC_sigma = fDivide(fMargin_DC_sigma, ConvertToFraction(1000)); fCACm_fused = fDivide(fCACm_fused, ConvertToFraction(100)); @@ -908,14 +907,14 @@ int atomctrl_calculate_voltage_evv_on_sclk( fSclk = GetScaledFraction(sclk, 100); fV_max = fDivide(GetScaledFraction( - getASICProfilingInfo->ulMaxVddc, 1000), ConvertToFraction(4)); - fT_prod = GetScaledFraction(getASICProfilingInfo->ulBoardCoreTemp, 10); - fLKG_Factor = GetScaledFraction(getASICProfilingInfo->ulEvvLkgFactor, 100); - fT_FT = GetScaledFraction(getASICProfilingInfo->ulLeakageTemp, 10); + le32_to_cpu(getASICProfilingInfo->ulMaxVddc), 1000), ConvertToFraction(4)); + fT_prod = GetScaledFraction(le32_to_cpu(getASICProfilingInfo->ulBoardCoreTemp), 10); + fLKG_Factor = GetScaledFraction(le32_to_cpu(getASICProfilingInfo->ulEvvLkgFactor), 100); + fT_FT = GetScaledFraction(le32_to_cpu(getASICProfilingInfo->ulLeakageTemp), 10); fV_FT = fDivide(GetScaledFraction( - getASICProfilingInfo->ulLeakageVoltage, 1000), ConvertToFraction(4)); + le32_to_cpu(getASICProfilingInfo->ulLeakageVoltage), 1000), ConvertToFraction(4)); fV_min = fDivide(GetScaledFraction( - getASICProfilingInfo->ulMinVddc, 1000), ConvertToFraction(4)); + le32_to_cpu(getASICProfilingInfo->ulMinVddc), 1000), ConvertToFraction(4)); /*----------------------- * PART 3 @@ -925,7 +924,7 @@ int atomctrl_calculate_voltage_evv_on_sclk( fA_Term = fAdd(fMargin_RO_a, fAdd(fMultiply(fSM_A4, fSclk), fSM_A5)); fB_Term = fAdd(fAdd(fMultiply(fSM_A2, fSclk), fSM_A6), fMargin_RO_b); fC_Term = fAdd(fMargin_RO_c, - fAdd(fMultiply(fSM_A0,fLkg_FT), + fAdd(fMultiply(fSM_A0, fLkg_FT), fAdd(fMultiply(fSM_A1, fMultiply(fLkg_FT, fSclk)), fAdd(fMultiply(fSM_A3, fSclk), fSubtract(fSM_A7, fRO_fused))))); @@ -1063,9 +1062,9 @@ int atomctrl_get_voltage_evv_on_sclk( get_voltage_info_param_space.ucVoltageMode = ATOM_GET_VOLTAGE_EVV_VOLTAGE; get_voltage_info_param_space.usVoltageLevel = - virtual_voltage_Id; + cpu_to_le16(virtual_voltage_Id); get_voltage_info_param_space.ulSCLKFreq = - sclk; + cpu_to_le32(sclk); result = cgs_atom_exec_cmd_table(hwmgr->device, GetIndexIntoMasterTable(COMMAND, GetVoltageInfo), @@ -1074,8 +1073,54 @@ int atomctrl_get_voltage_evv_on_sclk( if (0 != result) return result; - *voltage = ((GET_EVV_VOLTAGE_INFO_OUTPUT_PARAMETER_V1_2 *) - (&get_voltage_info_param_space))->usVoltageLevel; + *voltage = le16_to_cpu(((GET_EVV_VOLTAGE_INFO_OUTPUT_PARAMETER_V1_2 *) + (&get_voltage_info_param_space))->usVoltageLevel); + + return result; +} + +/** + * atomctrl_get_voltage_evv gets voltage via call to ATOM COMMAND table. + * @param hwmgr input: pointer to hwManager + * @param virtual_voltage_id input: voltage id which match per voltage DPM state: 0xff01, 0xff02.. 0xff08 + * @param voltage output: real voltage level in unit of mv + */ +int atomctrl_get_voltage_evv(struct pp_hwmgr *hwmgr, + uint16_t virtual_voltage_id, + uint16_t *voltage) +{ + int result; + int entry_id; + GET_VOLTAGE_INFO_INPUT_PARAMETER_V1_2 get_voltage_info_param_space; + + /* search for leakage voltage ID 0xff01 ~ 0xff08 and sckl */ + for (entry_id = 0; entry_id < hwmgr->dyn_state.vddc_dependency_on_sclk->count; entry_id++) { + if (hwmgr->dyn_state.vddc_dependency_on_sclk->entries[entry_id].v == virtual_voltage_id) { + /* found */ + break; + } + } + + PP_ASSERT_WITH_CODE(entry_id < hwmgr->dyn_state.vddc_dependency_on_sclk->count, + "Can't find requested voltage id in vddc_dependency_on_sclk table!", + return -EINVAL; + ); + + get_voltage_info_param_space.ucVoltageType = VOLTAGE_TYPE_VDDC; + get_voltage_info_param_space.ucVoltageMode = ATOM_GET_VOLTAGE_EVV_VOLTAGE; + get_voltage_info_param_space.usVoltageLevel = virtual_voltage_id; + get_voltage_info_param_space.ulSCLKFreq = + cpu_to_le32(hwmgr->dyn_state.vddc_dependency_on_sclk->entries[entry_id].clk); + + result = cgs_atom_exec_cmd_table(hwmgr->device, + GetIndexIntoMasterTable(COMMAND, GetVoltageInfo), + &get_voltage_info_param_space); + + if (0 != result) + return result; + + *voltage = le16_to_cpu(((GET_EVV_VOLTAGE_INFO_OUTPUT_PARAMETER_V1_2 *) + (&get_voltage_info_param_space))->usVoltageLevel); return result; } @@ -1165,8 +1210,8 @@ static int asic_internal_ss_get_ss_asignment(struct pp_hwmgr *hwmgr, if (entry_found) { ssEntry->speed_spectrum_percentage = - ssInfo->usSpreadSpectrumPercentage; - ssEntry->speed_spectrum_rate = ssInfo->usSpreadRateInKhz; + le16_to_cpu(ssInfo->usSpreadSpectrumPercentage); + ssEntry->speed_spectrum_rate = le16_to_cpu(ssInfo->usSpreadRateInKhz); if (((GET_DATA_TABLE_MAJOR_REVISION(table) == 2) && (GET_DATA_TABLE_MINOR_REVISION(table) >= 2)) || @@ -1222,7 +1267,7 @@ int atomctrl_read_efuse(void *device, uint16_t start_index, int result; READ_EFUSE_VALUE_PARAMETER efuse_param; - efuse_param.sEfuse.usEfuseIndex = (start_index / 32) * 4; + efuse_param.sEfuse.usEfuseIndex = cpu_to_le16((start_index / 32) * 4); efuse_param.sEfuse.ucBitShift = (uint8_t) (start_index - ((start_index / 32) * 32)); efuse_param.sEfuse.ucBitLength = (uint8_t) @@ -1232,19 +1277,21 @@ int atomctrl_read_efuse(void *device, uint16_t start_index, GetIndexIntoMasterTable(COMMAND, ReadEfuseValue), &efuse_param); if (!result) - *efuse = efuse_param.ulEfuseValue & mask; + *efuse = le32_to_cpu(efuse_param.ulEfuseValue) & mask; return result; } int atomctrl_set_ac_timing_ai(struct pp_hwmgr *hwmgr, uint32_t memory_clock, - uint8_t level) + uint8_t level) { DYNAMICE_MEMORY_SETTINGS_PARAMETER_V2_1 memory_clock_parameters; int result; - memory_clock_parameters.asDPMMCReg.ulClock.ulClockFreq = memory_clock & SET_CLOCK_FREQ_MASK; - memory_clock_parameters.asDPMMCReg.ulClock.ulComputeClockFlag = ADJUST_MC_SETTING_PARAM; + memory_clock_parameters.asDPMMCReg.ulClock.ulClockFreq = + memory_clock & SET_CLOCK_FREQ_MASK; + memory_clock_parameters.asDPMMCReg.ulClock.ulComputeClockFlag = + ADJUST_MC_SETTING_PARAM; memory_clock_parameters.asDPMMCReg.ucMclkDPMState = level; result = cgs_atom_exec_cmd_table @@ -1256,7 +1303,7 @@ int atomctrl_set_ac_timing_ai(struct pp_hwmgr *hwmgr, uint32_t memory_clock, } int atomctrl_get_voltage_evv_on_sclk_ai(struct pp_hwmgr *hwmgr, uint8_t voltage_type, - uint32_t sclk, uint16_t virtual_voltage_Id, uint16_t *voltage) + uint32_t sclk, uint16_t virtual_voltage_Id, uint32_t *voltage) { int result; @@ -1264,8 +1311,8 @@ int atomctrl_get_voltage_evv_on_sclk_ai(struct pp_hwmgr *hwmgr, uint8_t voltage_ get_voltage_info_param_space.ucVoltageType = voltage_type; get_voltage_info_param_space.ucVoltageMode = ATOM_GET_VOLTAGE_EVV_VOLTAGE; - get_voltage_info_param_space.usVoltageLevel = virtual_voltage_Id; - get_voltage_info_param_space.ulSCLKFreq = sclk; + get_voltage_info_param_space.usVoltageLevel = cpu_to_le16(virtual_voltage_Id); + get_voltage_info_param_space.ulSCLKFreq = cpu_to_le32(sclk); result = cgs_atom_exec_cmd_table(hwmgr->device, GetIndexIntoMasterTable(COMMAND, GetVoltageInfo), @@ -1274,7 +1321,7 @@ int atomctrl_get_voltage_evv_on_sclk_ai(struct pp_hwmgr *hwmgr, uint8_t voltage_ if (0 != result) return result; - *voltage = get_voltage_info_param_space.usVoltageLevel; + *voltage = le32_to_cpu(((GET_EVV_VOLTAGE_INFO_OUTPUT_PARAMETER_V1_3 *)(&get_voltage_info_param_space))->ulVoltageLevel); return result; } @@ -1295,10 +1342,57 @@ int atomctrl_get_smc_sclk_range_table(struct pp_hwmgr *hwmgr, struct pp_atom_ctr for (i = 0; i < psmu_info->ucSclkEntryNum; i++) { table->entry[i].ucVco_setting = psmu_info->asSclkFcwRangeEntry[i].ucVco_setting; table->entry[i].ucPostdiv = psmu_info->asSclkFcwRangeEntry[i].ucPostdiv; - table->entry[i].usFcw_pcc = psmu_info->asSclkFcwRangeEntry[i].ucFcw_pcc; - table->entry[i].usFcw_trans_upper = psmu_info->asSclkFcwRangeEntry[i].ucFcw_trans_upper; - table->entry[i].usRcw_trans_lower = psmu_info->asSclkFcwRangeEntry[i].ucRcw_trans_lower; + table->entry[i].usFcw_pcc = + le16_to_cpu(psmu_info->asSclkFcwRangeEntry[i].ucFcw_pcc); + table->entry[i].usFcw_trans_upper = + le16_to_cpu(psmu_info->asSclkFcwRangeEntry[i].ucFcw_trans_upper); + table->entry[i].usRcw_trans_lower = + le16_to_cpu(psmu_info->asSclkFcwRangeEntry[i].ucRcw_trans_lower); } return 0; } + +int atomctrl_get_avfs_information(struct pp_hwmgr *hwmgr, + struct pp_atom_ctrl__avfs_parameters *param) +{ + ATOM_ASIC_PROFILING_INFO_V3_6 *profile = NULL; + + if (param == NULL) + return -EINVAL; + + profile = (ATOM_ASIC_PROFILING_INFO_V3_6 *) + cgs_atom_get_data_table(hwmgr->device, + GetIndexIntoMasterTable(DATA, ASIC_ProfilingInfo), + NULL, NULL, NULL); + if (!profile) + return -1; + + param->ulAVFS_meanNsigma_Acontant0 = le32_to_cpu(profile->ulAVFS_meanNsigma_Acontant0); + param->ulAVFS_meanNsigma_Acontant1 = le32_to_cpu(profile->ulAVFS_meanNsigma_Acontant1); + param->ulAVFS_meanNsigma_Acontant2 = le32_to_cpu(profile->ulAVFS_meanNsigma_Acontant2); + param->usAVFS_meanNsigma_DC_tol_sigma = le16_to_cpu(profile->usAVFS_meanNsigma_DC_tol_sigma); + param->usAVFS_meanNsigma_Platform_mean = le16_to_cpu(profile->usAVFS_meanNsigma_Platform_mean); + param->usAVFS_meanNsigma_Platform_sigma = le16_to_cpu(profile->usAVFS_meanNsigma_Platform_sigma); + param->ulGB_VDROOP_TABLE_CKSOFF_a0 = le32_to_cpu(profile->ulGB_VDROOP_TABLE_CKSOFF_a0); + param->ulGB_VDROOP_TABLE_CKSOFF_a1 = le32_to_cpu(profile->ulGB_VDROOP_TABLE_CKSOFF_a1); + param->ulGB_VDROOP_TABLE_CKSOFF_a2 = le32_to_cpu(profile->ulGB_VDROOP_TABLE_CKSOFF_a2); + param->ulGB_VDROOP_TABLE_CKSON_a0 = le32_to_cpu(profile->ulGB_VDROOP_TABLE_CKSON_a0); + param->ulGB_VDROOP_TABLE_CKSON_a1 = le32_to_cpu(profile->ulGB_VDROOP_TABLE_CKSON_a1); + param->ulGB_VDROOP_TABLE_CKSON_a2 = le32_to_cpu(profile->ulGB_VDROOP_TABLE_CKSON_a2); + param->ulAVFSGB_FUSE_TABLE_CKSOFF_m1 = le32_to_cpu(profile->ulAVFSGB_FUSE_TABLE_CKSOFF_m1); + param->usAVFSGB_FUSE_TABLE_CKSOFF_m2 = le16_to_cpu(profile->usAVFSGB_FUSE_TABLE_CKSOFF_m2); + param->ulAVFSGB_FUSE_TABLE_CKSOFF_b = le32_to_cpu(profile->ulAVFSGB_FUSE_TABLE_CKSOFF_b); + param->ulAVFSGB_FUSE_TABLE_CKSON_m1 = le32_to_cpu(profile->ulAVFSGB_FUSE_TABLE_CKSON_m1); + param->usAVFSGB_FUSE_TABLE_CKSON_m2 = le16_to_cpu(profile->usAVFSGB_FUSE_TABLE_CKSON_m2); + param->ulAVFSGB_FUSE_TABLE_CKSON_b = le32_to_cpu(profile->ulAVFSGB_FUSE_TABLE_CKSON_b); + param->usMaxVoltage_0_25mv = le16_to_cpu(profile->usMaxVoltage_0_25mv); + param->ucEnableGB_VDROOP_TABLE_CKSOFF = profile->ucEnableGB_VDROOP_TABLE_CKSOFF; + param->ucEnableGB_VDROOP_TABLE_CKSON = profile->ucEnableGB_VDROOP_TABLE_CKSON; + param->ucEnableGB_FUSE_TABLE_CKSOFF = profile->ucEnableGB_FUSE_TABLE_CKSOFF; + param->ucEnableGB_FUSE_TABLE_CKSON = profile->ucEnableGB_FUSE_TABLE_CKSON; + param->usPSM_Age_ComFactor = le16_to_cpu(profile->usPSM_Age_ComFactor); + param->ucEnableApplyAVFS_CKS_OFF_Voltage = profile->ucEnableApplyAVFS_CKS_OFF_Voltage; + + return 0; +} diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/ppatomctrl.h b/drivers/gpu/drm/amd/powerplay/hwmgr/ppatomctrl.h index d24ebb566905..fc898afce002 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/ppatomctrl.h +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/ppatomctrl.h @@ -250,8 +250,38 @@ struct pp_atomctrl_gpio_pin_assignment { }; typedef struct pp_atomctrl_gpio_pin_assignment pp_atomctrl_gpio_pin_assignment; +struct pp_atom_ctrl__avfs_parameters { + uint32_t ulAVFS_meanNsigma_Acontant0; + uint32_t ulAVFS_meanNsigma_Acontant1; + uint32_t ulAVFS_meanNsigma_Acontant2; + uint16_t usAVFS_meanNsigma_DC_tol_sigma; + uint16_t usAVFS_meanNsigma_Platform_mean; + uint16_t usAVFS_meanNsigma_Platform_sigma; + uint32_t ulGB_VDROOP_TABLE_CKSOFF_a0; + uint32_t ulGB_VDROOP_TABLE_CKSOFF_a1; + uint32_t ulGB_VDROOP_TABLE_CKSOFF_a2; + uint32_t ulGB_VDROOP_TABLE_CKSON_a0; + uint32_t ulGB_VDROOP_TABLE_CKSON_a1; + uint32_t ulGB_VDROOP_TABLE_CKSON_a2; + uint32_t ulAVFSGB_FUSE_TABLE_CKSOFF_m1; + uint16_t usAVFSGB_FUSE_TABLE_CKSOFF_m2; + uint32_t ulAVFSGB_FUSE_TABLE_CKSOFF_b; + uint32_t ulAVFSGB_FUSE_TABLE_CKSON_m1; + uint16_t usAVFSGB_FUSE_TABLE_CKSON_m2; + uint32_t ulAVFSGB_FUSE_TABLE_CKSON_b; + uint16_t usMaxVoltage_0_25mv; + uint8_t ucEnableGB_VDROOP_TABLE_CKSOFF; + uint8_t ucEnableGB_VDROOP_TABLE_CKSON; + uint8_t ucEnableGB_FUSE_TABLE_CKSOFF; + uint8_t ucEnableGB_FUSE_TABLE_CKSON; + uint16_t usPSM_Age_ComFactor; + uint8_t ucEnableApplyAVFS_CKS_OFF_Voltage; + uint8_t ucReserved; +}; + extern bool atomctrl_get_pp_assign_pin(struct pp_hwmgr *hwmgr, const uint32_t pinId, pp_atomctrl_gpio_pin_assignment *gpio_pin_assignment); extern int atomctrl_get_voltage_evv_on_sclk(struct pp_hwmgr *hwmgr, uint8_t voltage_type, uint32_t sclk, uint16_t virtual_voltage_Id, uint16_t *voltage); +extern int atomctrl_get_voltage_evv(struct pp_hwmgr *hwmgr, uint16_t virtual_voltage_id, uint16_t *voltage); extern uint32_t atomctrl_get_mpll_reference_clock(struct pp_hwmgr *hwmgr); extern int atomctrl_get_memory_clock_spread_spectrum(struct pp_hwmgr *hwmgr, const uint32_t memory_clock, pp_atomctrl_internal_ss_info *ssInfo); extern int atomctrl_get_engine_clock_spread_spectrum(struct pp_hwmgr *hwmgr, const uint32_t engine_clock, pp_atomctrl_internal_ss_info *ssInfo); @@ -276,7 +306,10 @@ extern int atomctrl_get_engine_pll_dividers_ai(struct pp_hwmgr *hwmgr, uint32_t extern int atomctrl_set_ac_timing_ai(struct pp_hwmgr *hwmgr, uint32_t memory_clock, uint8_t level); extern int atomctrl_get_voltage_evv_on_sclk_ai(struct pp_hwmgr *hwmgr, uint8_t voltage_type, - uint32_t sclk, uint16_t virtual_voltage_Id, uint16_t *voltage); + uint32_t sclk, uint16_t virtual_voltage_Id, uint32_t *voltage); extern int atomctrl_get_smc_sclk_range_table(struct pp_hwmgr *hwmgr, struct pp_atom_ctrl_sclk_range_table *table); + +extern int atomctrl_get_avfs_information(struct pp_hwmgr *hwmgr, struct pp_atom_ctrl__avfs_parameters *param); + #endif diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/ppevvmath.h b/drivers/gpu/drm/amd/powerplay/hwmgr/ppevvmath.h index 009bd5963ed8..8f50a038396c 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/ppevvmath.h +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/ppevvmath.h @@ -50,55 +50,45 @@ typedef union _fInt { * Function Declarations * ------------------------------------------------------------------------------- */ -fInt ConvertToFraction(int); /* Use this to convert an INT to a FINT */ -fInt Convert_ULONG_ToFraction(uint32_t); /* Use this to convert an uint32_t to a FINT */ -fInt GetScaledFraction(int, int); /* Use this to convert an INT to a FINT after scaling it by a factor */ -int ConvertBackToInteger(fInt); /* Convert a FINT back to an INT that is scaled by 1000 (i.e. last 3 digits are the decimal digits) */ - -fInt fNegate(fInt); /* Returns -1 * input fInt value */ -fInt fAdd (fInt, fInt); /* Returns the sum of two fInt numbers */ -fInt fSubtract (fInt A, fInt B); /* Returns A-B - Sometimes easier than Adding negative numbers */ -fInt fMultiply (fInt, fInt); /* Returns the product of two fInt numbers */ -fInt fDivide (fInt A, fInt B); /* Returns A/B */ -fInt fGetSquare(fInt); /* Returns the square of a fInt number */ -fInt fSqrt(fInt); /* Returns the Square Root of a fInt number */ - -int uAbs(int); /* Returns the Absolute value of the Int */ -fInt fAbs(fInt); /* Returns the Absolute value of the fInt */ -int uPow(int base, int exponent); /* Returns base^exponent an INT */ - -void SolveQuadracticEqn(fInt, fInt, fInt, fInt[]); /* Returns the 2 roots via the array */ -bool Equal(fInt, fInt); /* Returns true if two fInts are equal to each other */ -bool GreaterThan(fInt A, fInt B); /* Returns true if A > B */ - -fInt fExponential(fInt exponent); /* Can be used to calculate e^exponent */ -fInt fNaturalLog(fInt value); /* Can be used to calculate ln(value) */ +static fInt ConvertToFraction(int); /* Use this to convert an INT to a FINT */ +static fInt Convert_ULONG_ToFraction(uint32_t); /* Use this to convert an uint32_t to a FINT */ +static fInt GetScaledFraction(int, int); /* Use this to convert an INT to a FINT after scaling it by a factor */ +static int ConvertBackToInteger(fInt); /* Convert a FINT back to an INT that is scaled by 1000 (i.e. last 3 digits are the decimal digits) */ + +static fInt fNegate(fInt); /* Returns -1 * input fInt value */ +static fInt fAdd (fInt, fInt); /* Returns the sum of two fInt numbers */ +static fInt fSubtract (fInt A, fInt B); /* Returns A-B - Sometimes easier than Adding negative numbers */ +static fInt fMultiply (fInt, fInt); /* Returns the product of two fInt numbers */ +static fInt fDivide (fInt A, fInt B); /* Returns A/B */ +static fInt fGetSquare(fInt); /* Returns the square of a fInt number */ +static fInt fSqrt(fInt); /* Returns the Square Root of a fInt number */ + +static int uAbs(int); /* Returns the Absolute value of the Int */ +static int uPow(int base, int exponent); /* Returns base^exponent an INT */ + +static void SolveQuadracticEqn(fInt, fInt, fInt, fInt[]); /* Returns the 2 roots via the array */ +static bool Equal(fInt, fInt); /* Returns true if two fInts are equal to each other */ +static bool GreaterThan(fInt A, fInt B); /* Returns true if A > B */ + +static fInt fExponential(fInt exponent); /* Can be used to calculate e^exponent */ +static fInt fNaturalLog(fInt value); /* Can be used to calculate ln(value) */ /* Fuse decoding functions * ------------------------------------------------------------------------------------- */ -fInt fDecodeLinearFuse(uint32_t fuse_value, fInt f_min, fInt f_range, uint32_t bitlength); -fInt fDecodeLogisticFuse(uint32_t fuse_value, fInt f_average, fInt f_range, uint32_t bitlength); -fInt fDecodeLeakageID (uint32_t leakageID_fuse, fInt ln_max_div_min, fInt f_min, uint32_t bitlength); +static fInt fDecodeLinearFuse(uint32_t fuse_value, fInt f_min, fInt f_range, uint32_t bitlength); +static fInt fDecodeLogisticFuse(uint32_t fuse_value, fInt f_average, fInt f_range, uint32_t bitlength); +static fInt fDecodeLeakageID (uint32_t leakageID_fuse, fInt ln_max_div_min, fInt f_min, uint32_t bitlength); /* Internal Support Functions - Use these ONLY for testing or adding to internal functions * ------------------------------------------------------------------------------------- * Some of the following functions take two INTs as their input - This is unsafe for a variety of reasons. */ -fInt Add (int, int); /* Add two INTs and return Sum as FINT */ -fInt Multiply (int, int); /* Multiply two INTs and return Product as FINT */ -fInt Divide (int, int); /* You get the idea... */ -fInt fNegate(fInt); +static fInt Divide (int, int); /* Divide two INTs and return result as FINT */ +static fInt fNegate(fInt); -int uGetScaledDecimal (fInt); /* Internal function */ -int GetReal (fInt A); /* Internal function */ - -/* Future Additions and Incomplete Functions - * ------------------------------------------------------------------------------------- - */ -int GetRoundedValue(fInt); /* Incomplete function - Useful only when Precision is lacking */ - /* Let us say we have 2.126 but can only handle 2 decimal points. We could */ - /* either chop of 6 and keep 2.12 or use this function to get 2.13, which is more accurate */ +static int uGetScaledDecimal (fInt); /* Internal function */ +static int GetReal (fInt A); /* Internal function */ /* ------------------------------------------------------------------------------------- * TROUBLESHOOTING INFORMATION @@ -115,7 +105,7 @@ int GetRoundedValue(fInt); /* Incomplete function - Usef * START OF CODE * ------------------------------------------------------------------------------------- */ -fInt fExponential(fInt exponent) /*Can be used to calculate e^exponent*/ +static fInt fExponential(fInt exponent) /*Can be used to calculate e^exponent*/ { uint32_t i; bool bNegated = false; @@ -154,7 +144,7 @@ fInt fExponential(fInt exponent) /*Can be used to calculate e^exponent*/ return solution; } -fInt fNaturalLog(fInt value) +static fInt fNaturalLog(fInt value) { uint32_t i; fInt upper_bound = Divide(8, 1000); @@ -179,7 +169,7 @@ fInt fNaturalLog(fInt value) return (fAdd(solution, error_term)); } -fInt fDecodeLinearFuse(uint32_t fuse_value, fInt f_min, fInt f_range, uint32_t bitlength) +static fInt fDecodeLinearFuse(uint32_t fuse_value, fInt f_min, fInt f_range, uint32_t bitlength) { fInt f_fuse_value = Convert_ULONG_ToFraction(fuse_value); fInt f_bit_max_value = Convert_ULONG_ToFraction((uPow(2, bitlength)) - 1); @@ -194,7 +184,7 @@ fInt fDecodeLinearFuse(uint32_t fuse_value, fInt f_min, fInt f_range, uint32_t b } -fInt fDecodeLogisticFuse(uint32_t fuse_value, fInt f_average, fInt f_range, uint32_t bitlength) +static fInt fDecodeLogisticFuse(uint32_t fuse_value, fInt f_average, fInt f_range, uint32_t bitlength) { fInt f_fuse_value = Convert_ULONG_ToFraction(fuse_value); fInt f_bit_max_value = Convert_ULONG_ToFraction((uPow(2, bitlength)) - 1); @@ -212,7 +202,7 @@ fInt fDecodeLogisticFuse(uint32_t fuse_value, fInt f_average, fInt f_range, uint return f_decoded_value; } -fInt fDecodeLeakageID (uint32_t leakageID_fuse, fInt ln_max_div_min, fInt f_min, uint32_t bitlength) +static fInt fDecodeLeakageID (uint32_t leakageID_fuse, fInt ln_max_div_min, fInt f_min, uint32_t bitlength) { fInt fLeakage; fInt f_bit_max_value = Convert_ULONG_ToFraction((uPow(2, bitlength)) - 1); @@ -225,7 +215,7 @@ fInt fDecodeLeakageID (uint32_t leakageID_fuse, fInt ln_max_div_min, fInt f_min, return fLeakage; } -fInt ConvertToFraction(int X) /*Add all range checking here. Is it possible to make fInt a private declaration? */ +static fInt ConvertToFraction(int X) /*Add all range checking here. Is it possible to make fInt a private declaration? */ { fInt temp; @@ -237,13 +227,13 @@ fInt ConvertToFraction(int X) /*Add all range checking here. Is it possible to m return temp; } -fInt fNegate(fInt X) +static fInt fNegate(fInt X) { fInt CONSTANT_NEGONE = ConvertToFraction(-1); return (fMultiply(X, CONSTANT_NEGONE)); } -fInt Convert_ULONG_ToFraction(uint32_t X) +static fInt Convert_ULONG_ToFraction(uint32_t X) { fInt temp; @@ -255,7 +245,7 @@ fInt Convert_ULONG_ToFraction(uint32_t X) return temp; } -fInt GetScaledFraction(int X, int factor) +static fInt GetScaledFraction(int X, int factor) { int times_shifted, factor_shifted; bool bNEGATED; @@ -304,7 +294,7 @@ fInt GetScaledFraction(int X, int factor) } /* Addition using two fInts */ -fInt fAdd (fInt X, fInt Y) +static fInt fAdd (fInt X, fInt Y) { fInt Sum; @@ -314,7 +304,7 @@ fInt fAdd (fInt X, fInt Y) } /* Addition using two fInts */ -fInt fSubtract (fInt X, fInt Y) +static fInt fSubtract (fInt X, fInt Y) { fInt Difference; @@ -323,7 +313,7 @@ fInt fSubtract (fInt X, fInt Y) return Difference; } -bool Equal(fInt A, fInt B) +static bool Equal(fInt A, fInt B) { if (A.full == B.full) return true; @@ -331,7 +321,7 @@ bool Equal(fInt A, fInt B) return false; } -bool GreaterThan(fInt A, fInt B) +static bool GreaterThan(fInt A, fInt B) { if (A.full > B.full) return true; @@ -339,7 +329,7 @@ bool GreaterThan(fInt A, fInt B) return false; } -fInt fMultiply (fInt X, fInt Y) /* Uses 64-bit integers (int64_t) */ +static fInt fMultiply (fInt X, fInt Y) /* Uses 64-bit integers (int64_t) */ { fInt Product; int64_t tempProduct; @@ -363,7 +353,7 @@ fInt fMultiply (fInt X, fInt Y) /* Uses 64-bit integers (int64_t) */ return Product; } -fInt fDivide (fInt X, fInt Y) +static fInt fDivide (fInt X, fInt Y) { fInt fZERO, fQuotient; int64_t longlongX, longlongY; @@ -384,7 +374,7 @@ fInt fDivide (fInt X, fInt Y) return fQuotient; } -int ConvertBackToInteger (fInt A) /*THIS is the function that will be used to check with the Golden settings table*/ +static int ConvertBackToInteger (fInt A) /*THIS is the function that will be used to check with the Golden settings table*/ { fInt fullNumber, scaledDecimal, scaledReal; @@ -397,13 +387,13 @@ int ConvertBackToInteger (fInt A) /*THIS is the function that will be used to ch return fullNumber.full; } -fInt fGetSquare(fInt A) +static fInt fGetSquare(fInt A) { return fMultiply(A,A); } /* x_new = x_old - (x_old^2 - C) / (2 * x_old) */ -fInt fSqrt(fInt num) +static fInt fSqrt(fInt num) { fInt F_divide_Fprime, Fprime; fInt test; @@ -460,7 +450,7 @@ fInt fSqrt(fInt num) return (x_new); } -void SolveQuadracticEqn(fInt A, fInt B, fInt C, fInt Roots[]) +static void SolveQuadracticEqn(fInt A, fInt B, fInt C, fInt Roots[]) { fInt *pRoots = &Roots[0]; fInt temp, root_first, root_second; @@ -498,52 +488,13 @@ void SolveQuadracticEqn(fInt A, fInt B, fInt C, fInt Roots[]) * ----------------------------------------------------------------------------- */ -/* Addition using two normal ints - Temporary - Use only for testing purposes?. */ -fInt Add (int X, int Y) -{ - fInt A, B, Sum; - - A.full = (X << SHIFT_AMOUNT); - B.full = (Y << SHIFT_AMOUNT); - - Sum.full = A.full + B.full; - - return Sum; -} - /* Conversion Functions */ -int GetReal (fInt A) +static int GetReal (fInt A) { return (A.full >> SHIFT_AMOUNT); } -/* Temporarily Disabled */ -int GetRoundedValue(fInt A) /*For now, round the 3rd decimal place */ -{ - /* ROUNDING TEMPORARLY DISABLED - int temp = A.full; - int decimal_cutoff, decimal_mask = 0x000001FF; - decimal_cutoff = temp & decimal_mask; - if (decimal_cutoff > 0x147) { - temp += 673; - }*/ - - return ConvertBackToInteger(A)/10000; /*Temporary - in case this was used somewhere else */ -} - -fInt Multiply (int X, int Y) -{ - fInt A, B, Product; - - A.full = X << SHIFT_AMOUNT; - B.full = Y << SHIFT_AMOUNT; - - Product = fMultiply(A, B); - - return Product; -} - -fInt Divide (int X, int Y) +static fInt Divide (int X, int Y) { fInt A, B, Quotient; @@ -555,7 +506,7 @@ fInt Divide (int X, int Y) return Quotient; } -int uGetScaledDecimal (fInt A) /*Converts the fractional portion to whole integers - Costly function */ +static int uGetScaledDecimal (fInt A) /*Converts the fractional portion to whole integers - Costly function */ { int dec[PRECISION]; int i, scaledDecimal = 0, tmp = A.partial.decimal; @@ -570,7 +521,7 @@ int uGetScaledDecimal (fInt A) /*Converts the fractional portion to whole intege return scaledDecimal; } -int uPow(int base, int power) +static int uPow(int base, int power) { if (power == 0) return 1; @@ -578,15 +529,7 @@ int uPow(int base, int power) return (base)*uPow(base, power - 1); } -fInt fAbs(fInt A) -{ - if (A.partial.real < 0) - return (fMultiply(A, ConvertToFraction(-1))); - else - return A; -} - -int uAbs(int X) +static int uAbs(int X) { if (X < 0) return (X * -1); @@ -594,7 +537,7 @@ int uAbs(int X) return X; } -fInt fRoundUpByStepSize(fInt A, fInt fStepSize, bool error_term) +static fInt fRoundUpByStepSize(fInt A, fInt fStepSize, bool error_term) { fInt solution; diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/processpptables.c b/drivers/gpu/drm/amd/powerplay/hwmgr/processpptables.c index 2f1a14fe05b1..6c321b0d8a1e 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/processpptables.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/processpptables.c @@ -794,19 +794,35 @@ static const ATOM_PPLIB_STATE_V2 *get_state_entry_v2( static const ATOM_PPLIB_POWERPLAYTABLE *get_powerplay_table( struct pp_hwmgr *hwmgr) { - const void *table_addr = NULL; + const void *table_addr = hwmgr->soft_pp_table; uint8_t frev, crev; uint16_t size; - table_addr = cgs_atom_get_data_table(hwmgr->device, - GetIndexIntoMasterTable(DATA, PowerPlayInfo), - &size, &frev, &crev); + if (!table_addr) { + table_addr = cgs_atom_get_data_table(hwmgr->device, + GetIndexIntoMasterTable(DATA, PowerPlayInfo), + &size, &frev, &crev); - hwmgr->soft_pp_table = table_addr; + hwmgr->soft_pp_table = table_addr; + hwmgr->soft_pp_table_size = size; + } return (const ATOM_PPLIB_POWERPLAYTABLE *)table_addr; } +int pp_tables_get_response_times(struct pp_hwmgr *hwmgr, + uint32_t *vol_rep_time, uint32_t *bb_rep_time) +{ + const ATOM_PPLIB_POWERPLAYTABLE *powerplay_tab = get_powerplay_table(hwmgr); + + PP_ASSERT_WITH_CODE(NULL != powerplay_tab, + "Missing PowerPlay Table!", return -EINVAL); + + *vol_rep_time = (uint32_t)le16_to_cpu(powerplay_tab->usVoltageTime); + *bb_rep_time = (uint32_t)le16_to_cpu(powerplay_tab->usBackbiasTime); + + return 0; +} int pp_tables_get_num_of_entries(struct pp_hwmgr *hwmgr, unsigned long *num_of_entries) @@ -1499,7 +1515,7 @@ int get_number_of_vce_state_table_entries( const ATOM_PPLIB_VCE_State_Table *vce_table = get_vce_state_table(hwmgr, table); - if (vce_table > 0) + if (vce_table) return vce_table->numEntries; return 0; @@ -1589,11 +1605,6 @@ static int pp_tables_initialize(struct pp_hwmgr *hwmgr) static int pp_tables_uninitialize(struct pp_hwmgr *hwmgr) { - if (NULL != hwmgr->soft_pp_table) { - kfree(hwmgr->soft_pp_table); - hwmgr->soft_pp_table = NULL; - } - if (NULL != hwmgr->dyn_state.vddc_dependency_on_sclk) { kfree(hwmgr->dyn_state.vddc_dependency_on_sclk); hwmgr->dyn_state.vddc_dependency_on_sclk = NULL; diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/processpptables.h b/drivers/gpu/drm/amd/powerplay/hwmgr/processpptables.h index 30434802417e..baddaa75693b 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/processpptables.h +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/processpptables.h @@ -32,16 +32,19 @@ struct pp_hw_power_state; extern const struct pp_table_func pptable_funcs; typedef int (*pp_tables_hw_clock_info_callback)(struct pp_hwmgr *hwmgr, - struct pp_hw_power_state *hw_ps, - unsigned int index, - const void *clock_info); + struct pp_hw_power_state *hw_ps, + unsigned int index, + const void *clock_info); int pp_tables_get_num_of_entries(struct pp_hwmgr *hwmgr, - unsigned long *num_of_entries); + unsigned long *num_of_entries); int pp_tables_get_entry(struct pp_hwmgr *hwmgr, - unsigned long entry_index, - struct pp_power_state *ps, - pp_tables_hw_clock_info_callback func); + unsigned long entry_index, + struct pp_power_state *ps, + pp_tables_hw_clock_info_callback func); + +int pp_tables_get_response_times(struct pp_hwmgr *hwmgr, + uint32_t *vol_rep_time, uint32_t *bb_rep_time); #endif diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/tonga_hwmgr.c b/drivers/gpu/drm/amd/powerplay/hwmgr/tonga_hwmgr.c index 16fed487973b..c7dc111221c2 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/tonga_hwmgr.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/tonga_hwmgr.c @@ -571,7 +571,7 @@ int tonga_disable_sclk_mclk_dpm(struct pp_hwmgr *hwmgr) if (0 == data->sclk_dpm_key_disabled) { /* Checking if DPM is running. If we discover hang because of this, we should skip this message.*/ PP_ASSERT_WITH_CODE( - (0 == tonga_is_dpm_running(hwmgr)), + !tonga_is_dpm_running(hwmgr), "Trying to Disable SCLK DPM when DPM is disabled", return -1 ); @@ -587,7 +587,7 @@ int tonga_disable_sclk_mclk_dpm(struct pp_hwmgr *hwmgr) if (0 == data->mclk_dpm_key_disabled) { /* Checking if DPM is running. If we discover hang because of this, we should skip this message. */ PP_ASSERT_WITH_CODE( - (0 == tonga_is_dpm_running(hwmgr)), + !tonga_is_dpm_running(hwmgr), "Trying to Disable MCLK DPM when DPM is disabled", return -1 ); @@ -614,7 +614,7 @@ int tonga_stop_dpm(struct pp_hwmgr *hwmgr) if (0 == data->pcie_dpm_key_disabled) { /* Checking if DPM is running. If we discover hang because of this, we should skip this message.*/ PP_ASSERT_WITH_CODE( - (0 == tonga_is_dpm_running(hwmgr)), + !tonga_is_dpm_running(hwmgr), "Trying to Disable PCIE DPM when DPM is disabled", return -1 ); @@ -630,7 +630,7 @@ int tonga_stop_dpm(struct pp_hwmgr *hwmgr) /* Checking if DPM is running. If we discover hang because of this, we should skip this message.*/ PP_ASSERT_WITH_CODE( - (0 == tonga_is_dpm_running(hwmgr)), + !tonga_is_dpm_running(hwmgr), "Trying to Disable Voltage CNTL when DPM is disabled", return -1 ); @@ -688,8 +688,9 @@ int tonga_dpm_force_state(struct pp_hwmgr *hwmgr, uint32_t n) uint32_t level_mask = 1 << n; /* Checking if DPM is running. If we discover hang because of this, we should skip this message. */ - PP_ASSERT_WITH_CODE(0 == tonga_is_dpm_running(hwmgr), - "Trying to force SCLK when DPM is disabled", return -1;); + PP_ASSERT_WITH_CODE(!tonga_is_dpm_running(hwmgr), + "Trying to force SCLK when DPM is disabled", + return -1;); if (0 == data->sclk_dpm_key_disabled) return (0 == smum_send_msg_to_smc_with_parameter( hwmgr->smumgr, @@ -712,8 +713,9 @@ int tonga_dpm_force_state_mclk(struct pp_hwmgr *hwmgr, uint32_t n) uint32_t level_mask = 1 << n; /* Checking if DPM is running. If we discover hang because of this, we should skip this message. */ - PP_ASSERT_WITH_CODE(0 == tonga_is_dpm_running(hwmgr), - "Trying to Force MCLK when DPM is disabled", return -1;); + PP_ASSERT_WITH_CODE(!tonga_is_dpm_running(hwmgr), + "Trying to Force MCLK when DPM is disabled", + return -1;); if (0 == data->mclk_dpm_key_disabled) return (0 == smum_send_msg_to_smc_with_parameter( hwmgr->smumgr, @@ -735,8 +737,9 @@ int tonga_dpm_force_state_pcie(struct pp_hwmgr *hwmgr, uint32_t n) tonga_hwmgr *data = (tonga_hwmgr *)(hwmgr->backend); /* Checking if DPM is running. If we discover hang because of this, we should skip this message.*/ - PP_ASSERT_WITH_CODE(0 == tonga_is_dpm_running(hwmgr), - "Trying to Force PCIE level when DPM is disabled", return -1;); + PP_ASSERT_WITH_CODE(!tonga_is_dpm_running(hwmgr), + "Trying to Force PCIE level when DPM is disabled", + return -1;); if (0 == data->pcie_dpm_key_disabled) return (0 == smum_send_msg_to_smc_with_parameter( hwmgr->smumgr, @@ -774,7 +777,7 @@ int tonga_process_firmware_header(struct pp_hwmgr *hwmgr) uint32_t tmp; int result; - bool error = 0; + bool error = false; result = tonga_read_smc_sram_dword(hwmgr->smumgr, SMU72_FIRMWARE_HEADER_LOCATION + @@ -933,11 +936,11 @@ int tonga_init_power_gate_state(struct pp_hwmgr *hwmgr) { tonga_hwmgr *data = (tonga_hwmgr *)(hwmgr->backend); - data->uvd_power_gated = 0; - data->vce_power_gated = 0; - data->samu_power_gated = 0; - data->acp_power_gated = 0; - data->pg_acp_init = 1; + data->uvd_power_gated = false; + data->vce_power_gated = false; + data->samu_power_gated = false; + data->acp_power_gated = false; + data->pg_acp_init = true; return 0; } @@ -955,7 +958,7 @@ int tonga_check_for_dpm_running(struct pp_hwmgr *hwmgr) * because we may have test scenarios that need us intentionly disable SCLK/MCLK DPM, * whereas voltage control is a fundemental change that will not be disabled */ - return (0 == tonga_is_dpm_running(hwmgr) ? 0 : 1); + return (!tonga_is_dpm_running(hwmgr) ? 0 : 1); } /** @@ -968,7 +971,7 @@ int tonga_check_for_dpm_stopped(struct pp_hwmgr *hwmgr) { tonga_hwmgr *data = (tonga_hwmgr *)(hwmgr->backend); - if (0 != tonga_is_dpm_running(hwmgr)) { + if (tonga_is_dpm_running(hwmgr)) { /* If HW Virtualization is enabled, dpm_table_start will not have a valid value */ if (!data->dpm_table_start) { return 1; @@ -991,7 +994,7 @@ static int tonga_trim_voltage_table(struct pp_hwmgr *hwmgr, { uint32_t table_size, i, j; uint16_t vvalue; - bool bVoltageFound = 0; + bool bVoltageFound = false; pp_atomctrl_voltage_table *table; PP_ASSERT_WITH_CODE((NULL != voltage_table), "Voltage Table empty.", return -1;); @@ -1007,11 +1010,11 @@ static int tonga_trim_voltage_table(struct pp_hwmgr *hwmgr, for (i = 0; i < voltage_table->count; i++) { vvalue = voltage_table->entries[i].value; - bVoltageFound = 0; + bVoltageFound = false; for (j = 0; j < table->count; j++) { if (vvalue == table->entries[j].value) { - bVoltageFound = 1; + bVoltageFound = true; break; } } @@ -1302,7 +1305,7 @@ static int tonga_populate_smc_mvdd_table(struct pp_hwmgr *hwmgr, table->Smio[count] |= data->mvdd_voltage_table.entries[count].smio_low; } - table->SmioMask2 = data->vddci_voltage_table.mask_low; + table->SmioMask2 = data->mvdd_voltage_table.mask_low; CONVERT_FROM_HOST_TO_SMC_UL(table->MvddLevelCount); } @@ -1331,7 +1334,6 @@ static int tonga_populate_cac_tables(struct pp_hwmgr *hwmgr, { uint32_t count; uint8_t index; - int result = 0; tonga_hwmgr *data = (tonga_hwmgr *)(hwmgr->backend); struct phm_ppt_v1_information *pptable_info = (struct phm_ppt_v1_information *)(hwmgr->pptable); struct phm_ppt_v1_voltage_lookup_table *vddgfx_lookup_table = pptable_info->vddgfx_lookup_table; @@ -1378,7 +1380,7 @@ static int tonga_populate_cac_tables(struct pp_hwmgr *hwmgr, } } - return result; + return 0; } @@ -2042,7 +2044,7 @@ static int tonga_populate_single_memory_level( if ((data->mclk_stutter_mode_threshold != 0) && (memory_clock <= data->mclk_stutter_mode_threshold) && - (data->is_uvd_enabled == 0) + (!data->is_uvd_enabled) && (PHM_READ_FIELD(hwmgr->device, DPG_PIPE_STUTTER_CONTROL, STUTTER_ENABLE) & 0x1) && (data->display_timing.num_existing_displays <= 2) && (data->display_timing.num_existing_displays != 0)) @@ -2705,7 +2707,7 @@ static int tonga_reset_single_dpm_table( dpm_table->count = count; for (i = 0; i < MAX_REGULAR_DPM_NUMBER; i++) { - dpm_table->dpm_levels[i].enabled = 0; + dpm_table->dpm_levels[i].enabled = false; } return 0; @@ -2718,7 +2720,7 @@ static void tonga_setup_pcie_table_entry( { dpm_table->dpm_levels[index].value = pcie_gen; dpm_table->dpm_levels[index].param1 = pcie_lanes; - dpm_table->dpm_levels[index].enabled = 1; + dpm_table->dpm_levels[index].enabled = true; } static int tonga_setup_default_pcie_tables(struct pp_hwmgr *hwmgr) @@ -2828,7 +2830,7 @@ static int tonga_setup_default_dpm_tables(struct pp_hwmgr *hwmgr) allowed_vdd_sclk_table->entries[i].clk) { data->dpm_table.sclk_table.dpm_levels[data->dpm_table.sclk_table.count].value = allowed_vdd_sclk_table->entries[i].clk; - data->dpm_table.sclk_table.dpm_levels[data->dpm_table.sclk_table.count].enabled = 1; /*(i==0) ? 1 : 0; to do */ + data->dpm_table.sclk_table.dpm_levels[data->dpm_table.sclk_table.count].enabled = true; /*(i==0) ? 1 : 0; to do */ data->dpm_table.sclk_table.count++; } } @@ -2842,32 +2844,11 @@ static int tonga_setup_default_dpm_tables(struct pp_hwmgr *hwmgr) allowed_vdd_mclk_table->entries[i].clk) { data->dpm_table.mclk_table.dpm_levels[data->dpm_table.mclk_table.count].value = allowed_vdd_mclk_table->entries[i].clk; - data->dpm_table.mclk_table.dpm_levels[data->dpm_table.mclk_table.count].enabled = 1; /*(i==0) ? 1 : 0; */ + data->dpm_table.mclk_table.dpm_levels[data->dpm_table.mclk_table.count].enabled = true; /*(i==0) ? 1 : 0; */ data->dpm_table.mclk_table.count++; } } - /* Initialize Vddc DPM table based on allow Vddc values. And populate corresponding std values. */ - for (i = 0; i < allowed_vdd_sclk_table->count; i++) { - data->dpm_table.vddc_table.dpm_levels[i].value = allowed_vdd_mclk_table->entries[i].vddc; - /* tonga_hwmgr->dpm_table.VddcTable.dpm_levels[i].param1 = stdVoltageTable->entries[i].Leakage; */ - /* param1 is for corresponding std voltage */ - data->dpm_table.vddc_table.dpm_levels[i].enabled = 1; - } - data->dpm_table.vddc_table.count = allowed_vdd_sclk_table->count; - - if (NULL != allowed_vdd_mclk_table) { - /* Initialize Vddci DPM table based on allow Mclk values */ - for (i = 0; i < allowed_vdd_mclk_table->count; i++) { - data->dpm_table.vdd_ci_table.dpm_levels[i].value = allowed_vdd_mclk_table->entries[i].vddci; - data->dpm_table.vdd_ci_table.dpm_levels[i].enabled = 1; - data->dpm_table.mvdd_table.dpm_levels[i].value = allowed_vdd_mclk_table->entries[i].mvdd; - data->dpm_table.mvdd_table.dpm_levels[i].enabled = 1; - } - data->dpm_table.vdd_ci_table.count = allowed_vdd_mclk_table->count; - data->dpm_table.mvdd_table.count = allowed_vdd_mclk_table->count; - } - /* setup PCIE gen speed levels*/ tonga_setup_default_pcie_tables(hwmgr); @@ -3047,8 +3028,8 @@ int tonga_init_smc_table(struct pp_hwmgr *hwmgr) reg_value = 0; if ((0 == reg_value) && - (0 == atomctrl_get_pp_assign_pin(hwmgr, - VDDC_VRHOT_GPIO_PINID, &gpio_pin_assignment))) { + (atomctrl_get_pp_assign_pin(hwmgr, VDDC_VRHOT_GPIO_PINID, + &gpio_pin_assignment))) { table->VRHotGpio = gpio_pin_assignment.uc_gpio_pin_bit_shift; phm_cap_set(hwmgr->platform_descriptor.platformCaps, PHM_PlatformCaps_RegulatorHot); @@ -3061,8 +3042,8 @@ int tonga_init_smc_table(struct pp_hwmgr *hwmgr) /* ACDC Switch GPIO */ reg_value = 0; if ((0 == reg_value) && - (0 == atomctrl_get_pp_assign_pin(hwmgr, - PP_AC_DC_SWITCH_GPIO_PINID, &gpio_pin_assignment))) { + (atomctrl_get_pp_assign_pin(hwmgr, PP_AC_DC_SWITCH_GPIO_PINID, + &gpio_pin_assignment))) { table->AcDcGpio = gpio_pin_assignment.uc_gpio_pin_bit_shift; phm_cap_set(hwmgr->platform_descriptor.platformCaps, PHM_PlatformCaps_AutomaticDCTransition); @@ -3084,8 +3065,7 @@ int tonga_init_smc_table(struct pp_hwmgr *hwmgr) } reg_value = 0; - if ((0 == reg_value) && - (0 == atomctrl_get_pp_assign_pin(hwmgr, + if ((0 == reg_value) && (atomctrl_get_pp_assign_pin(hwmgr, THERMAL_INT_OUTPUT_GPIO_PINID, &gpio_pin_assignment))) { phm_cap_set(hwmgr->platform_descriptor.platformCaps, PHM_PlatformCaps_ThermalOutGPIO); @@ -3156,7 +3136,7 @@ int tonga_upload_dpm_level_enable_mask(struct pp_hwmgr *hwmgr) if (0 == data->sclk_dpm_key_disabled) { /* Checking if DPM is running. If we discover hang because of this, we should skip this message.*/ - if (0 != tonga_is_dpm_running(hwmgr)) + if (tonga_is_dpm_running(hwmgr)) printk(KERN_ERR "[ powerplay ] Trying to set Enable Mask when DPM is disabled \n"); if (0 != data->dpm_level_enable_mask.sclk_dpm_enable_mask) { @@ -3171,7 +3151,7 @@ int tonga_upload_dpm_level_enable_mask(struct pp_hwmgr *hwmgr) if (0 == data->mclk_dpm_key_disabled) { /* Checking if DPM is running. If we discover hang because of this, we should skip this message.*/ - if (0 != tonga_is_dpm_running(hwmgr)) + if (tonga_is_dpm_running(hwmgr)) printk(KERN_ERR "[ powerplay ] Trying to set Enable Mask when DPM is disabled \n"); if (0 != data->dpm_level_enable_mask.mclk_dpm_enable_mask) { @@ -3282,7 +3262,7 @@ int tonga_initializa_dynamic_state_adjustment_rule_settings(struct pp_hwmgr *hwm /* initialize vddc_dep_on_dal_pwrl table */ table_size = sizeof(uint32_t) + 4 * sizeof(struct phm_clock_voltage_dependency_record); - table_clk_vlt = (struct phm_clock_voltage_dependency_table *)kzalloc(table_size, GFP_KERNEL); + table_clk_vlt = kzalloc(table_size, GFP_KERNEL); if (NULL == table_clk_vlt) { printk(KERN_ERR "[ powerplay ] Can not allocate space for vddc_dep_on_dal_pwrl! \n"); @@ -3357,9 +3337,9 @@ int tonga_unforce_dpm_levels(struct pp_hwmgr *hwmgr) tonga_hwmgr *data = (tonga_hwmgr *)(hwmgr->backend); int result = 1; - PP_ASSERT_WITH_CODE (0 == tonga_is_dpm_running(hwmgr), - "Trying to Unforce DPM when DPM is disabled. Returning without sending SMC message.", - return result); + PP_ASSERT_WITH_CODE (!tonga_is_dpm_running(hwmgr), + "Trying to Unforce DPM when DPM is disabled. Returning without sending SMC message.", + return result); if (0 == data->pcie_dpm_key_disabled) { PP_ASSERT_WITH_CODE((0 == smum_send_msg_to_smc( @@ -3763,7 +3743,7 @@ uint8_t tonga_get_memory_modile_index(struct pp_hwmgr *hwmgr) bool tonga_check_s0_mc_reg_index(uint16_t inReg, uint16_t *outReg) { - bool result = 1; + bool result = true; switch (inReg) { case mmMC_SEQ_RAS_TIMING: @@ -3847,7 +3827,7 @@ bool tonga_check_s0_mc_reg_index(uint16_t inReg, uint16_t *outReg) break; default: - result = 0; + result = false; break; } @@ -4443,13 +4423,6 @@ int tonga_reset_asic_tasks(struct pp_hwmgr *hwmgr) int tonga_hwmgr_backend_fini(struct pp_hwmgr *hwmgr) { - struct tonga_hwmgr *data = (struct tonga_hwmgr *)(hwmgr->backend); - - if (data->soft_pp_table) { - kfree(data->soft_pp_table); - data->soft_pp_table = NULL; - } - return phm_hwmgr_backend_fini(hwmgr); } @@ -4463,7 +4436,7 @@ int tonga_hwmgr_backend_init(struct pp_hwmgr *hwmgr) { int result = 0; SMU72_Discrete_DpmTable *table = NULL; - tonga_hwmgr *data = (struct tonga_hwmgr *)(hwmgr->backend); + tonga_hwmgr *data; pp_atomctrl_gpio_pin_assignment gpio_pin_assignment; struct phm_ppt_v1_information *pptable_info = (struct phm_ppt_v1_information *)(hwmgr->pptable); phw_tonga_ulv_parm *ulv; @@ -4472,7 +4445,13 @@ int tonga_hwmgr_backend_init(struct pp_hwmgr *hwmgr) PP_ASSERT_WITH_CODE((NULL != hwmgr), "Invalid Parameter!", return -1;); - data->dll_defaule_on = 0; + data = kzalloc(sizeof(struct tonga_hwmgr), GFP_KERNEL); + if (data == NULL) + return -ENOMEM; + + hwmgr->backend = data; + + data->dll_defaule_on = false; data->sram_end = SMC_RAM_END; data->activity_target[0] = PPTONGA_TARGETACTIVITY_DFLT; @@ -4510,6 +4489,7 @@ int tonga_hwmgr_backend_init(struct pp_hwmgr *hwmgr) data->vdd_ci_control = TONGA_VOLTAGE_CONTROL_NONE; data->vdd_gfx_control = TONGA_VOLTAGE_CONTROL_NONE; data->mvdd_control = TONGA_VOLTAGE_CONTROL_NONE; + data->force_pcie_gen = PP_PCIEGenInvalid; if (atomctrl_is_voltage_controled_by_gpio_v3(hwmgr, VOLTAGE_TYPE_VDDC, VOLTAGE_OBJ_SVID2)) { @@ -4577,13 +4557,13 @@ int tonga_hwmgr_backend_init(struct pp_hwmgr *hwmgr) /* ULV Support*/ ulv = &(data->ulv); - ulv->ulv_supported = 0; + ulv->ulv_supported = false; /* Initalize Dynamic State Adjustment Rule Settings*/ result = tonga_initializa_dynamic_state_adjustment_rule_settings(hwmgr); if (result) printk(KERN_ERR "[ powerplay ] tonga_initializa_dynamic_state_adjustment_rule_settings failed!\n"); - data->uvd_enabled = 0; + data->uvd_enabled = false; table = &(data->smc_state_table); @@ -4591,7 +4571,7 @@ int tonga_hwmgr_backend_init(struct pp_hwmgr *hwmgr) * if ucGPIO_ID=VDDC_PCC_GPIO_PINID in GPIO_LUTable, * Peak Current Control feature is enabled and we should program PCC HW register */ - if (0 == atomctrl_get_pp_assign_pin(hwmgr, VDDC_PCC_GPIO_PINID, &gpio_pin_assignment)) { + if (atomctrl_get_pp_assign_pin(hwmgr, VDDC_PCC_GPIO_PINID, &gpio_pin_assignment)) { uint32_t temp_reg = cgs_read_ind_register(hwmgr->device, CGS_IND_REG__SMC, ixCNB_PWRMGT_CNTL); @@ -4630,7 +4610,7 @@ int tonga_hwmgr_backend_init(struct pp_hwmgr *hwmgr) phm_cap_set(hwmgr->platform_descriptor.platformCaps, PHM_PlatformCaps_SMU7); - data->vddc_phase_shed_control = 0; + data->vddc_phase_shed_control = false; phm_cap_unset(hwmgr->platform_descriptor.platformCaps, PHM_PlatformCaps_UVDPowerGating); @@ -4649,7 +4629,7 @@ int tonga_hwmgr_backend_init(struct pp_hwmgr *hwmgr) } if (0 == result) { - data->is_tlu_enabled = 0; + data->is_tlu_enabled = false; hwmgr->platform_descriptor.hardwareActivityPerformanceLevels = TONGA_MAX_HARDWARE_POWERLEVELS; hwmgr->platform_descriptor.hardwarePerformanceLevels = 2; @@ -4659,7 +4639,7 @@ int tonga_hwmgr_backend_init(struct pp_hwmgr *hwmgr) sys_info.info_id = CGS_SYSTEM_INFO_PCIE_GEN_INFO; result = cgs_query_system_info(hwmgr->device, &sys_info); if (result) - data->pcie_gen_cap = 0x30007; + data->pcie_gen_cap = AMDGPU_DEFAULT_PCIE_GEN_MASK; else data->pcie_gen_cap = (uint32_t)sys_info.value; if (data->pcie_gen_cap & CAIL_PCIE_LINK_SPEED_SUPPORT_GEN3) @@ -4668,7 +4648,7 @@ int tonga_hwmgr_backend_init(struct pp_hwmgr *hwmgr) sys_info.info_id = CGS_SYSTEM_INFO_PCIE_MLW; result = cgs_query_system_info(hwmgr->device, &sys_info); if (result) - data->pcie_lane_cap = 0x2f0000; + data->pcie_lane_cap = AMDGPU_DEFAULT_PCIE_MLW_MASK; else data->pcie_lane_cap = (uint32_t)sys_info.value; } else { @@ -5330,9 +5310,8 @@ static int tonga_freeze_sclk_mclk_dpm(struct pp_hwmgr *hwmgr) if ((0 == data->sclk_dpm_key_disabled) && (data->need_update_smu7_dpm_table & (DPMTABLE_OD_UPDATE_SCLK + DPMTABLE_UPDATE_SCLK))) { - PP_ASSERT_WITH_CODE( - 0 == tonga_is_dpm_running(hwmgr), - "Trying to freeze SCLK DPM when DPM is disabled", + PP_ASSERT_WITH_CODE(!tonga_is_dpm_running(hwmgr), + "Trying to freeze SCLK DPM when DPM is disabled", ); PP_ASSERT_WITH_CODE( 0 == smum_send_msg_to_smc(hwmgr->smumgr, @@ -5344,8 +5323,8 @@ static int tonga_freeze_sclk_mclk_dpm(struct pp_hwmgr *hwmgr) if ((0 == data->mclk_dpm_key_disabled) && (data->need_update_smu7_dpm_table & DPMTABLE_OD_UPDATE_MCLK)) { - PP_ASSERT_WITH_CODE(0 == tonga_is_dpm_running(hwmgr), - "Trying to freeze MCLK DPM when DPM is disabled", + PP_ASSERT_WITH_CODE(!tonga_is_dpm_running(hwmgr), + "Trying to freeze MCLK DPM when DPM is disabled", ); PP_ASSERT_WITH_CODE( 0 == smum_send_msg_to_smc(hwmgr->smumgr, @@ -5480,7 +5459,6 @@ static int tonga_trim_single_dpm_states(struct pp_hwmgr *hwmgr, static int tonga_trim_dpm_states(struct pp_hwmgr *hwmgr, const struct tonga_power_state *hw_state) { - int result = 0; struct tonga_hwmgr *data = (struct tonga_hwmgr *)(hwmgr->backend); uint32_t high_limit_count; @@ -5500,7 +5478,7 @@ static int tonga_trim_dpm_states(struct pp_hwmgr *hwmgr, const struct tonga_powe hw_state->performance_levels[0].memory_clock, hw_state->performance_levels[high_limit_count].memory_clock); - return result; + return 0; } static int tonga_generate_dpm_level_enable_mask(struct pp_hwmgr *hwmgr, const void *input) @@ -5647,8 +5625,8 @@ static int tonga_unfreeze_sclk_mclk_dpm(struct pp_hwmgr *hwmgr) (data->need_update_smu7_dpm_table & (DPMTABLE_OD_UPDATE_SCLK + DPMTABLE_UPDATE_SCLK))) { - PP_ASSERT_WITH_CODE(0 == tonga_is_dpm_running(hwmgr), - "Trying to Unfreeze SCLK DPM when DPM is disabled", + PP_ASSERT_WITH_CODE(!tonga_is_dpm_running(hwmgr), + "Trying to Unfreeze SCLK DPM when DPM is disabled", ); PP_ASSERT_WITH_CODE( 0 == smum_send_msg_to_smc(hwmgr->smumgr, @@ -5660,9 +5638,8 @@ static int tonga_unfreeze_sclk_mclk_dpm(struct pp_hwmgr *hwmgr) if ((0 == data->mclk_dpm_key_disabled) && (data->need_update_smu7_dpm_table & DPMTABLE_OD_UPDATE_MCLK)) { - PP_ASSERT_WITH_CODE( - 0 == tonga_is_dpm_running(hwmgr), - "Trying to Unfreeze MCLK DPM when DPM is disabled", + PP_ASSERT_WITH_CODE(!tonga_is_dpm_running(hwmgr), + "Trying to Unfreeze MCLK DPM when DPM is disabled", ); PP_ASSERT_WITH_CODE( 0 == smum_send_msg_to_smc(hwmgr->smumgr, @@ -6051,42 +6028,6 @@ static int tonga_get_fan_control_mode(struct pp_hwmgr *hwmgr) CG_FDO_CTRL2, FDO_PWM_MODE); } -static int tonga_get_pp_table(struct pp_hwmgr *hwmgr, char **table) -{ - struct tonga_hwmgr *data = (struct tonga_hwmgr *)(hwmgr->backend); - - if (!data->soft_pp_table) { - data->soft_pp_table = kmemdup(hwmgr->soft_pp_table, - hwmgr->soft_pp_table_size, - GFP_KERNEL); - if (!data->soft_pp_table) - return -ENOMEM; - } - - *table = (char *)&data->soft_pp_table; - - return hwmgr->soft_pp_table_size; -} - -static int tonga_set_pp_table(struct pp_hwmgr *hwmgr, const char *buf, size_t size) -{ - struct tonga_hwmgr *data = (struct tonga_hwmgr *)(hwmgr->backend); - - if (!data->soft_pp_table) { - data->soft_pp_table = kzalloc(hwmgr->soft_pp_table_size, GFP_KERNEL); - if (!data->soft_pp_table) - return -ENOMEM; - } - - memcpy(data->soft_pp_table, buf, size); - - hwmgr->soft_pp_table = data->soft_pp_table; - - /* TODO: re-init powerplay to implement modified pptable */ - - return 0; -} - static int tonga_force_clock_level(struct pp_hwmgr *hwmgr, enum pp_clock_type type, uint32_t mask) { @@ -6194,11 +6135,96 @@ static int tonga_print_clock_levels(struct pp_hwmgr *hwmgr, return size; } +static int tonga_get_sclk_od(struct pp_hwmgr *hwmgr) +{ + struct tonga_hwmgr *data = (struct tonga_hwmgr *)(hwmgr->backend); + struct tonga_single_dpm_table *sclk_table = &(data->dpm_table.sclk_table); + struct tonga_single_dpm_table *golden_sclk_table = + &(data->golden_dpm_table.sclk_table); + int value; + + value = (sclk_table->dpm_levels[sclk_table->count - 1].value - + golden_sclk_table->dpm_levels[golden_sclk_table->count - 1].value) * + 100 / + golden_sclk_table->dpm_levels[golden_sclk_table->count - 1].value; + + return value; +} + +static int tonga_set_sclk_od(struct pp_hwmgr *hwmgr, uint32_t value) +{ + struct tonga_hwmgr *data = (struct tonga_hwmgr *)(hwmgr->backend); + struct tonga_single_dpm_table *golden_sclk_table = + &(data->golden_dpm_table.sclk_table); + struct pp_power_state *ps; + struct tonga_power_state *tonga_ps; + + if (value > 20) + value = 20; + + ps = hwmgr->request_ps; + + if (ps == NULL) + return -EINVAL; + + tonga_ps = cast_phw_tonga_power_state(&ps->hardware); + + tonga_ps->performance_levels[tonga_ps->performance_level_count - 1].engine_clock = + golden_sclk_table->dpm_levels[golden_sclk_table->count - 1].value * + value / 100 + + golden_sclk_table->dpm_levels[golden_sclk_table->count - 1].value; + + return 0; +} + +static int tonga_get_mclk_od(struct pp_hwmgr *hwmgr) +{ + struct tonga_hwmgr *data = (struct tonga_hwmgr *)(hwmgr->backend); + struct tonga_single_dpm_table *mclk_table = &(data->dpm_table.mclk_table); + struct tonga_single_dpm_table *golden_mclk_table = + &(data->golden_dpm_table.mclk_table); + int value; + + value = (mclk_table->dpm_levels[mclk_table->count - 1].value - + golden_mclk_table->dpm_levels[golden_mclk_table->count - 1].value) * + 100 / + golden_mclk_table->dpm_levels[golden_mclk_table->count - 1].value; + + return value; +} + +static int tonga_set_mclk_od(struct pp_hwmgr *hwmgr, uint32_t value) +{ + struct tonga_hwmgr *data = (struct tonga_hwmgr *)(hwmgr->backend); + struct tonga_single_dpm_table *golden_mclk_table = + &(data->golden_dpm_table.mclk_table); + struct pp_power_state *ps; + struct tonga_power_state *tonga_ps; + + if (value > 20) + value = 20; + + ps = hwmgr->request_ps; + + if (ps == NULL) + return -EINVAL; + + tonga_ps = cast_phw_tonga_power_state(&ps->hardware); + + tonga_ps->performance_levels[tonga_ps->performance_level_count - 1].memory_clock = + golden_mclk_table->dpm_levels[golden_mclk_table->count - 1].value * + value / 100 + + golden_mclk_table->dpm_levels[golden_mclk_table->count - 1].value; + + return 0; +} + static const struct pp_hwmgr_func tonga_hwmgr_funcs = { .backend_init = &tonga_hwmgr_backend_init, .backend_fini = &tonga_hwmgr_backend_fini, .asic_setup = &tonga_setup_asic_task, .dynamic_state_management_enable = &tonga_enable_dpm_tasks, + .dynamic_state_management_disable = &tonga_disable_dpm_tasks, .apply_state_adjust_rules = tonga_apply_state_adjust_rules, .force_dpm_level = &tonga_force_dpm_level, .power_state_set = tonga_set_power_state_tasks, @@ -6232,22 +6258,16 @@ static const struct pp_hwmgr_func tonga_hwmgr_funcs = { .check_states_equal = tonga_check_states_equal, .set_fan_control_mode = tonga_set_fan_control_mode, .get_fan_control_mode = tonga_get_fan_control_mode, - .get_pp_table = tonga_get_pp_table, - .set_pp_table = tonga_set_pp_table, .force_clock_level = tonga_force_clock_level, .print_clock_levels = tonga_print_clock_levels, + .get_sclk_od = tonga_get_sclk_od, + .set_sclk_od = tonga_set_sclk_od, + .get_mclk_od = tonga_get_mclk_od, + .set_mclk_od = tonga_set_mclk_od, }; int tonga_hwmgr_init(struct pp_hwmgr *hwmgr) { - tonga_hwmgr *data; - - data = kzalloc (sizeof(tonga_hwmgr), GFP_KERNEL); - if (data == NULL) - return -ENOMEM; - memset(data, 0x00, sizeof(tonga_hwmgr)); - - hwmgr->backend = data; hwmgr->hwmgr_func = &tonga_hwmgr_funcs; hwmgr->pptable_func = &tonga_pptable_funcs; pp_tonga_thermal_initialize(hwmgr); diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/tonga_hwmgr.h b/drivers/gpu/drm/amd/powerplay/hwmgr/tonga_hwmgr.h index 573cd39fe78d..3961884bfa9b 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/tonga_hwmgr.h +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/tonga_hwmgr.h @@ -352,9 +352,6 @@ struct tonga_hwmgr { bool samu_power_gated; /* 1: gated, 0:not gated */ bool acp_power_gated; /* 1: gated, 0:not gated */ bool pg_acp_init; - - /* soft pptable for re-uploading into smu */ - void *soft_pp_table; }; typedef struct tonga_hwmgr tonga_hwmgr; diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/tonga_pptable.h b/drivers/gpu/drm/amd/powerplay/hwmgr/tonga_pptable.h index 1b44f4e9b8f5..f127198aafc4 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/tonga_pptable.h +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/tonga_pptable.h @@ -197,6 +197,22 @@ typedef struct _ATOM_Tonga_SCLK_Dependency_Table { ATOM_Tonga_SCLK_Dependency_Record entries[1]; /* Dynamically allocate entries. */ } ATOM_Tonga_SCLK_Dependency_Table; +typedef struct _ATOM_Polaris_SCLK_Dependency_Record { + UCHAR ucVddInd; /* Base voltage */ + USHORT usVddcOffset; /* Offset relative to base voltage */ + ULONG ulSclk; + USHORT usEdcCurrent; + UCHAR ucReliabilityTemperature; + UCHAR ucCKSVOffsetandDisable; /* Bits 0~6: Voltage offset for CKS, Bit 7: Disable/enable for the SCLK level. */ + ULONG ulSclkOffset; +} ATOM_Polaris_SCLK_Dependency_Record; + +typedef struct _ATOM_Polaris_SCLK_Dependency_Table { + UCHAR ucRevId; + UCHAR ucNumEntries; /* Number of entries. */ + ATOM_Polaris_SCLK_Dependency_Record entries[1]; /* Dynamically allocate entries. */ +} ATOM_Polaris_SCLK_Dependency_Table; + typedef struct _ATOM_Tonga_PCIE_Record { UCHAR ucPCIEGenSpeed; UCHAR usPCIELaneWidth; diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/tonga_processpptables.c b/drivers/gpu/drm/amd/powerplay/hwmgr/tonga_processpptables.c index 10e3630ee39d..cfb647f76cbe 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/tonga_processpptables.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/tonga_processpptables.c @@ -167,8 +167,7 @@ static int get_vddc_lookup_table( table_size = sizeof(uint32_t) + sizeof(phm_ppt_v1_voltage_lookup_record) * max_levels; - table = (phm_ppt_v1_voltage_lookup_table *) - kzalloc(table_size, GFP_KERNEL); + table = kzalloc(table_size, GFP_KERNEL); if (NULL == table) return -ENOMEM; @@ -302,7 +301,7 @@ static int init_dpm_2_parameters( (((unsigned long)powerplay_table) + le16_to_cpu(powerplay_table->usPPMTableOffset)); if (0 != powerplay_table->usPPMTableOffset) { - if (1 == get_platform_power_management_table(hwmgr, atom_ppm_table)) { + if (get_platform_power_management_table(hwmgr, atom_ppm_table) == 0) { phm_cap_set(hwmgr->platform_descriptor.platformCaps, PHM_PlatformCaps_EnablePlatformPowerManagement); } @@ -327,7 +326,7 @@ static int get_valid_clk( table_size = sizeof(uint32_t) + sizeof(uint32_t) * clk_volt_pp_table->count; - table = (struct phm_clock_array *)kzalloc(table_size, GFP_KERNEL); + table = kzalloc(table_size, GFP_KERNEL); if (NULL == table) return -ENOMEM; @@ -377,8 +376,7 @@ static int get_mclk_voltage_dependency_table( table_size = sizeof(uint32_t) + sizeof(phm_ppt_v1_clock_voltage_dependency_record) * mclk_dep_table->ucNumEntries; - mclk_table = (phm_ppt_v1_clock_voltage_dependency_table *) - kzalloc(table_size, GFP_KERNEL); + mclk_table = kzalloc(table_size, GFP_KERNEL); if (NULL == mclk_table) return -ENOMEM; @@ -408,41 +406,76 @@ static int get_mclk_voltage_dependency_table( static int get_sclk_voltage_dependency_table( struct pp_hwmgr *hwmgr, phm_ppt_v1_clock_voltage_dependency_table **pp_tonga_sclk_dep_table, - const ATOM_Tonga_SCLK_Dependency_Table * sclk_dep_table + const PPTable_Generic_SubTable_Header *sclk_dep_table ) { uint32_t table_size, i; phm_ppt_v1_clock_voltage_dependency_table *sclk_table; - PP_ASSERT_WITH_CODE((0 != sclk_dep_table->ucNumEntries), - "Invalid PowerPlay Table!", return -1); + if (sclk_dep_table->ucRevId < 1) { + const ATOM_Tonga_SCLK_Dependency_Table *tonga_table = + (ATOM_Tonga_SCLK_Dependency_Table *)sclk_dep_table; - table_size = sizeof(uint32_t) + sizeof(phm_ppt_v1_clock_voltage_dependency_record) - * sclk_dep_table->ucNumEntries; + PP_ASSERT_WITH_CODE((0 != tonga_table->ucNumEntries), + "Invalid PowerPlay Table!", return -1); - sclk_table = (phm_ppt_v1_clock_voltage_dependency_table *) - kzalloc(table_size, GFP_KERNEL); + table_size = sizeof(uint32_t) + sizeof(phm_ppt_v1_clock_voltage_dependency_record) + * tonga_table->ucNumEntries; - if (NULL == sclk_table) - return -ENOMEM; + sclk_table = kzalloc(table_size, GFP_KERNEL); - memset(sclk_table, 0x00, table_size); - - sclk_table->count = (uint32_t)sclk_dep_table->ucNumEntries; - - for (i = 0; i < sclk_dep_table->ucNumEntries; i++) { - sclk_table->entries[i].vddInd = - sclk_dep_table->entries[i].ucVddInd; - sclk_table->entries[i].vdd_offset = - sclk_dep_table->entries[i].usVddcOffset; - sclk_table->entries[i].clk = - sclk_dep_table->entries[i].ulSclk; - sclk_table->entries[i].cks_enable = - (((sclk_dep_table->entries[i].ucCKSVOffsetandDisable & 0x80) >> 7) == 0) ? 1 : 0; - sclk_table->entries[i].cks_voffset = - (sclk_dep_table->entries[i].ucCKSVOffsetandDisable & 0x7F); - } + if (NULL == sclk_table) + return -ENOMEM; + + memset(sclk_table, 0x00, table_size); + + sclk_table->count = (uint32_t)tonga_table->ucNumEntries; + + for (i = 0; i < tonga_table->ucNumEntries; i++) { + sclk_table->entries[i].vddInd = + tonga_table->entries[i].ucVddInd; + sclk_table->entries[i].vdd_offset = + tonga_table->entries[i].usVddcOffset; + sclk_table->entries[i].clk = + tonga_table->entries[i].ulSclk; + sclk_table->entries[i].cks_enable = + (((tonga_table->entries[i].ucCKSVOffsetandDisable & 0x80) >> 7) == 0) ? 1 : 0; + sclk_table->entries[i].cks_voffset = + (tonga_table->entries[i].ucCKSVOffsetandDisable & 0x7F); + } + } else { + const ATOM_Polaris_SCLK_Dependency_Table *polaris_table = + (ATOM_Polaris_SCLK_Dependency_Table *)sclk_dep_table; + + PP_ASSERT_WITH_CODE((0 != polaris_table->ucNumEntries), + "Invalid PowerPlay Table!", return -1); + + table_size = sizeof(uint32_t) + sizeof(phm_ppt_v1_clock_voltage_dependency_record) + * polaris_table->ucNumEntries; + sclk_table = kzalloc(table_size, GFP_KERNEL); + + if (NULL == sclk_table) + return -ENOMEM; + + memset(sclk_table, 0x00, table_size); + + sclk_table->count = (uint32_t)polaris_table->ucNumEntries; + + for (i = 0; i < polaris_table->ucNumEntries; i++) { + sclk_table->entries[i].vddInd = + polaris_table->entries[i].ucVddInd; + sclk_table->entries[i].vdd_offset = + polaris_table->entries[i].usVddcOffset; + sclk_table->entries[i].clk = + polaris_table->entries[i].ulSclk; + sclk_table->entries[i].cks_enable = + (((polaris_table->entries[i].ucCKSVOffsetandDisable & 0x80) >> 7) == 0) ? 1 : 0; + sclk_table->entries[i].cks_voffset = + (polaris_table->entries[i].ucCKSVOffsetandDisable & 0x7F); + sclk_table->entries[i].sclk_offset = polaris_table->entries[i].ulSclkOffset; + } + } *pp_tonga_sclk_dep_table = sclk_table; return 0; @@ -467,7 +500,7 @@ static int get_pcie_table( table_size = sizeof(uint32_t) + sizeof(phm_ppt_v1_pcie_record) * atom_pcie_table->ucNumEntries; - pcie_table = (phm_ppt_v1_pcie_table *)kzalloc(table_size, GFP_KERNEL); + pcie_table = kzalloc(table_size, GFP_KERNEL); if (pcie_table == NULL) return -ENOMEM; @@ -504,7 +537,7 @@ static int get_pcie_table( table_size = sizeof(uint32_t) + sizeof(phm_ppt_v1_pcie_record) * atom_pcie_table->ucNumEntries; - pcie_table = (phm_ppt_v1_pcie_table *)kzalloc(table_size, GFP_KERNEL); + pcie_table = kzalloc(table_size, GFP_KERNEL); if (pcie_table == NULL) return -ENOMEM; @@ -658,8 +691,7 @@ static int get_mm_clock_voltage_table( table_size = sizeof(uint32_t) + sizeof(phm_ppt_v1_mm_clock_voltage_dependency_record) * mm_dependency_table->ucNumEntries; - mm_table = (phm_ppt_v1_mm_clock_voltage_dependency_table *) - kzalloc(table_size, GFP_KERNEL); + mm_table = kzalloc(table_size, GFP_KERNEL); if (NULL == mm_table) return -ENOMEM; @@ -708,8 +740,8 @@ static int init_clock_voltage_dependency( const ATOM_Tonga_MCLK_Dependency_Table *mclk_dep_table = (const ATOM_Tonga_MCLK_Dependency_Table *)(((unsigned long) powerplay_table) + le16_to_cpu(powerplay_table->usMclkDependencyTableOffset)); - const ATOM_Tonga_SCLK_Dependency_Table *sclk_dep_table = - (const ATOM_Tonga_SCLK_Dependency_Table *)(((unsigned long) powerplay_table) + + const PPTable_Generic_SubTable_Header *sclk_dep_table = + (const PPTable_Generic_SubTable_Header *)(((unsigned long) powerplay_table) + le16_to_cpu(powerplay_table->usSclkDependencyTableOffset)); const ATOM_Tonga_Hard_Limit_Table *pHardLimits = (const ATOM_Tonga_Hard_Limit_Table *)(((unsigned long) powerplay_table) + @@ -1036,54 +1068,46 @@ int tonga_pp_tables_initialize(struct pp_hwmgr *hwmgr) int tonga_pp_tables_uninitialize(struct pp_hwmgr *hwmgr) { - int result = 0; struct phm_ppt_v1_information *pp_table_information = (struct phm_ppt_v1_information *)(hwmgr->pptable); - if (NULL != hwmgr->soft_pp_table) { - kfree(hwmgr->soft_pp_table); - hwmgr->soft_pp_table = NULL; - } - - if (NULL != pp_table_information->vdd_dep_on_sclk) - pp_table_information->vdd_dep_on_sclk = NULL; + kfree(pp_table_information->vdd_dep_on_sclk); + pp_table_information->vdd_dep_on_sclk = NULL; - if (NULL != pp_table_information->vdd_dep_on_mclk) - pp_table_information->vdd_dep_on_mclk = NULL; + kfree(pp_table_information->vdd_dep_on_mclk); + pp_table_information->vdd_dep_on_mclk = NULL; - if (NULL != pp_table_information->valid_mclk_values) - pp_table_information->valid_mclk_values = NULL; + kfree(pp_table_information->valid_mclk_values); + pp_table_information->valid_mclk_values = NULL; - if (NULL != pp_table_information->valid_sclk_values) - pp_table_information->valid_sclk_values = NULL; + kfree(pp_table_information->valid_sclk_values); + pp_table_information->valid_sclk_values = NULL; - if (NULL != pp_table_information->vddc_lookup_table) - pp_table_information->vddc_lookup_table = NULL; + kfree(pp_table_information->vddc_lookup_table); + pp_table_information->vddc_lookup_table = NULL; - if (NULL != pp_table_information->vddgfx_lookup_table) - pp_table_information->vddgfx_lookup_table = NULL; + kfree(pp_table_information->vddgfx_lookup_table); + pp_table_information->vddgfx_lookup_table = NULL; - if (NULL != pp_table_information->mm_dep_table) - pp_table_information->mm_dep_table = NULL; + kfree(pp_table_information->mm_dep_table); + pp_table_information->mm_dep_table = NULL; - if (NULL != pp_table_information->cac_dtp_table) - pp_table_information->cac_dtp_table = NULL; + kfree(pp_table_information->cac_dtp_table); + pp_table_information->cac_dtp_table = NULL; - if (NULL != hwmgr->dyn_state.cac_dtp_table) - hwmgr->dyn_state.cac_dtp_table = NULL; + kfree(hwmgr->dyn_state.cac_dtp_table); + hwmgr->dyn_state.cac_dtp_table = NULL; - if (NULL != pp_table_information->ppm_parameter_table) - pp_table_information->ppm_parameter_table = NULL; + kfree(pp_table_information->ppm_parameter_table); + pp_table_information->ppm_parameter_table = NULL; - if (NULL != pp_table_information->pcie_table) - pp_table_information->pcie_table = NULL; + kfree(pp_table_information->pcie_table); + pp_table_information->pcie_table = NULL; - if (NULL != hwmgr->pptable) { - kfree(hwmgr->pptable); - hwmgr->pptable = NULL; - } + kfree(hwmgr->pptable); + hwmgr->pptable = NULL; - return result; + return 0; } const struct pp_table_func tonga_pptable_funcs = { diff --git a/drivers/gpu/drm/amd/powerplay/inc/amd_powerplay.h b/drivers/gpu/drm/amd/powerplay/inc/amd_powerplay.h index 50b367d44307..b764c8c05ec8 100644 --- a/drivers/gpu/drm/amd/powerplay/inc/amd_powerplay.h +++ b/drivers/gpu/drm/amd/powerplay/inc/amd_powerplay.h @@ -132,6 +132,7 @@ struct amd_pp_init { uint32_t chip_family; uint32_t chip_id; uint32_t rev_id; + bool powercontainment_enabled; }; enum amd_pp_display_config_type{ AMD_PP_DisplayConfigType_None = 0, @@ -342,6 +343,10 @@ struct amd_powerplay_funcs { int (*set_pp_table)(void *handle, const char *buf, size_t size); int (*force_clock_level)(void *handle, enum pp_clock_type type, uint32_t mask); int (*print_clock_levels)(void *handle, enum pp_clock_type type, char *buf); + int (*get_sclk_od)(void *handle); + int (*set_sclk_od)(void *handle, uint32_t value); + int (*get_mclk_od)(void *handle); + int (*set_mclk_od)(void *handle, uint32_t value); }; struct amd_powerplay { @@ -355,6 +360,8 @@ int amd_powerplay_init(struct amd_pp_init *pp_init, int amd_powerplay_fini(void *handle); +int amd_powerplay_reset(void *handle); + int amd_powerplay_display_configuration_change(void *handle, const struct amd_pp_display_configuration *input); diff --git a/drivers/gpu/drm/amd/powerplay/inc/hardwaremanager.h b/drivers/gpu/drm/amd/powerplay/inc/hardwaremanager.h index 56f712c7d07a..962cb5385951 100644 --- a/drivers/gpu/drm/amd/powerplay/inc/hardwaremanager.h +++ b/drivers/gpu/drm/amd/powerplay/inc/hardwaremanager.h @@ -340,6 +340,7 @@ extern int phm_powergate_vce(struct pp_hwmgr *hwmgr, bool gate); extern int phm_powerdown_uvd(struct pp_hwmgr *hwmgr); extern int phm_setup_asic(struct pp_hwmgr *hwmgr); extern int phm_enable_dynamic_state_management(struct pp_hwmgr *hwmgr); +extern int phm_disable_dynamic_state_management(struct pp_hwmgr *hwmgr); extern void phm_init_dynamic_caps(struct pp_hwmgr *hwmgr); extern bool phm_is_hw_access_blocked(struct pp_hwmgr *hwmgr); extern int phm_block_hw_access(struct pp_hwmgr *hwmgr, bool block); diff --git a/drivers/gpu/drm/amd/powerplay/inc/hwmgr.h b/drivers/gpu/drm/amd/powerplay/inc/hwmgr.h index 28f571449495..bf0d2accf7bf 100644 --- a/drivers/gpu/drm/amd/powerplay/inc/hwmgr.h +++ b/drivers/gpu/drm/amd/powerplay/inc/hwmgr.h @@ -278,6 +278,8 @@ struct pp_hwmgr_func { int (*dynamic_state_management_enable)( struct pp_hwmgr *hw_mgr); + int (*dynamic_state_management_disable)( + struct pp_hwmgr *hw_mgr); int (*patch_boot_state)(struct pp_hwmgr *hwmgr, struct pp_hw_power_state *hw_ps); @@ -333,11 +335,13 @@ struct pp_hwmgr_func { int (*get_clock_by_type)(struct pp_hwmgr *hwmgr, enum amd_pp_clock_type type, struct amd_pp_clocks *clocks); int (*get_max_high_clocks)(struct pp_hwmgr *hwmgr, struct amd_pp_simple_clock_info *clocks); int (*power_off_asic)(struct pp_hwmgr *hwmgr); - int (*get_pp_table)(struct pp_hwmgr *hwmgr, char **table); - int (*set_pp_table)(struct pp_hwmgr *hwmgr, const char *buf, size_t size); int (*force_clock_level)(struct pp_hwmgr *hwmgr, enum pp_clock_type type, uint32_t mask); int (*print_clock_levels)(struct pp_hwmgr *hwmgr, enum pp_clock_type type, char *buf); int (*enable_per_cu_power_gating)(struct pp_hwmgr *hwmgr, bool enable); + int (*get_sclk_od)(struct pp_hwmgr *hwmgr); + int (*set_sclk_od)(struct pp_hwmgr *hwmgr, uint32_t value); + int (*get_mclk_od)(struct pp_hwmgr *hwmgr); + int (*set_mclk_od)(struct pp_hwmgr *hwmgr, uint32_t value); }; struct pp_table_func { @@ -411,6 +415,8 @@ struct phm_cac_tdp_table { uint8_t ucVr_I2C_Line; uint8_t ucPlx_I2C_address; uint8_t ucPlx_I2C_Line; + uint32_t usBoostPowerLimit; + uint8_t ucCKS_LDO_REFSEL; }; struct phm_ppm_table { @@ -578,6 +584,7 @@ struct pp_hwmgr { struct pp_smumgr *smumgr; const void *soft_pp_table; uint32_t soft_pp_table_size; + void *hardcode_pp_table; bool need_pp_table_upload; enum amd_dpm_forced_level dpm_level; bool block_hw_access; @@ -607,6 +614,7 @@ struct pp_hwmgr { uint32_t num_ps; struct pp_thermal_controller_info thermal_controller; bool fan_ctrl_is_in_default_mode; + bool powercontainment_enabled; uint32_t fan_ctrl_default_mode; uint32_t tmin; struct phm_microcode_version_info microcode_version_info; diff --git a/drivers/gpu/drm/amd/powerplay/inc/polaris10_ppsmc.h b/drivers/gpu/drm/amd/powerplay/inc/polaris10_ppsmc.h index 0c6a413eaa5b..b8f4b73c322e 100644 --- a/drivers/gpu/drm/amd/powerplay/inc/polaris10_ppsmc.h +++ b/drivers/gpu/drm/amd/powerplay/inc/polaris10_ppsmc.h @@ -27,6 +27,7 @@ #pragma pack(push, 1) +#define PPSMC_MSG_SetGBDroopSettings ((uint16_t) 0x305) #define PPSMC_SWSTATE_FLAG_DC 0x01 #define PPSMC_SWSTATE_FLAG_UVD 0x02 @@ -391,6 +392,8 @@ typedef uint16_t PPSMC_Result; #define PPSMC_MSG_SetGpuPllDfsForSclk ((uint16_t) 0x300) #define PPSMC_MSG_Didt_Block_Function ((uint16_t) 0x301) +#define PPSMC_MSG_SetVBITimeout ((uint16_t) 0x306) + #define PPSMC_MSG_SecureSRBMWrite ((uint16_t) 0x600) #define PPSMC_MSG_SecureSRBMRead ((uint16_t) 0x601) #define PPSMC_MSG_SetAddress ((uint16_t) 0x800) diff --git a/drivers/gpu/drm/amd/powerplay/inc/pp_acpi.h b/drivers/gpu/drm/amd/powerplay/inc/pp_acpi.h index 3bd5e69b9045..3df5de2cdab0 100644 --- a/drivers/gpu/drm/amd/powerplay/inc/pp_acpi.h +++ b/drivers/gpu/drm/amd/powerplay/inc/pp_acpi.h @@ -26,3 +26,4 @@ extern bool acpi_atcs_functions_supported(void *device, extern int acpi_pcie_perf_request(void *device, uint8_t perf_req, bool advertise); +extern bool acpi_atcs_notify_pcie_device_ready(void *device); diff --git a/drivers/gpu/drm/amd/powerplay/inc/smu74.h b/drivers/gpu/drm/amd/powerplay/inc/smu74.h index 1a12d85b8e97..fd10a9fa843d 100644 --- a/drivers/gpu/drm/amd/powerplay/inc/smu74.h +++ b/drivers/gpu/drm/amd/powerplay/inc/smu74.h @@ -34,6 +34,30 @@ #define SMU__NUM_LCLK_DPM_LEVELS 8 #define SMU__NUM_PCIE_DPM_LEVELS 8 +#define EXP_M1 35 +#define EXP_M2 92821 +#define EXP_B 66629747 + +#define EXP_M1_1 365 +#define EXP_M2_1 658700 +#define EXP_B_1 305506134 + +#define EXP_M1_2 189 +#define EXP_M2_2 379692 +#define EXP_B_2 194609469 + +#define EXP_M1_3 99 +#define EXP_M2_3 217915 +#define EXP_B_3 122255994 + +#define EXP_M1_4 51 +#define EXP_M2_4 122643 +#define EXP_B_4 74893384 + +#define EXP_M1_5 423 +#define EXP_M2_5 1103326 +#define EXP_B_5 728122621 + enum SID_OPTION { SID_OPTION_HI, SID_OPTION_LO, @@ -548,20 +572,20 @@ struct SMU74_Firmware_Header { uint32_t CacConfigTable; uint32_t CacStatusTable; - uint32_t mcRegisterTable; - uint32_t mcArbDramTimingTable; - - - uint32_t PmFuseTable; uint32_t Globals; uint32_t ClockStretcherTable; uint32_t VftTable; - uint32_t Reserved[21]; + uint32_t Reserved1; + uint32_t AvfsTable; + uint32_t AvfsCksOffGbvTable; + uint32_t AvfsMeanNSigma; + uint32_t AvfsSclkOffsetTable; + uint32_t Reserved[16]; uint32_t Signature; }; @@ -701,8 +725,6 @@ VR Config info is contained in dpmTable.VRConfig */ struct SMU_ClockStretcherDataTableEntry { uint8_t minVID; uint8_t maxVID; - - uint16_t setting; }; typedef struct SMU_ClockStretcherDataTableEntry SMU_ClockStretcherDataTableEntry; @@ -769,6 +791,43 @@ struct VFT_TABLE_t { typedef struct VFT_TABLE_t VFT_TABLE_t; +/* Total margin, root mean square of Fmax + DC + Platform */ +struct AVFS_Margin_t { + VFT_CELL_t Cell[NUM_VFT_COLUMNS]; +}; +typedef struct AVFS_Margin_t AVFS_Margin_t; + +#define BTCGB_VDROOP_TABLE_MAX_ENTRIES 2 +#define AVFSGB_VDROOP_TABLE_MAX_ENTRIES 2 + +struct GB_VDROOP_TABLE_t { + int32_t a0; + int32_t a1; + int32_t a2; + uint32_t spare; +}; +typedef struct GB_VDROOP_TABLE_t GB_VDROOP_TABLE_t; + +struct AVFS_CksOff_Gbv_t { + VFT_CELL_t Cell[NUM_VFT_COLUMNS]; +}; +typedef struct AVFS_CksOff_Gbv_t AVFS_CksOff_Gbv_t; + +struct AVFS_meanNsigma_t { + uint32_t Aconstant[3]; + uint16_t DC_tol_sigma; + uint16_t Platform_mean; + uint16_t Platform_sigma; + uint16_t PSM_Age_CompFactor; + uint8_t Static_Voltage_Offset[NUM_VFT_COLUMNS]; +}; +typedef struct AVFS_meanNsigma_t AVFS_meanNsigma_t; + +struct AVFS_Sclk_Offset_t { + uint16_t Sclk_Offset[8]; +}; +typedef struct AVFS_Sclk_Offset_t AVFS_Sclk_Offset_t; + #endif diff --git a/drivers/gpu/drm/amd/powerplay/inc/smu74_discrete.h b/drivers/gpu/drm/amd/powerplay/inc/smu74_discrete.h index 0dfe82336dc7..899d6d8108c2 100644 --- a/drivers/gpu/drm/amd/powerplay/inc/smu74_discrete.h +++ b/drivers/gpu/drm/amd/powerplay/inc/smu74_discrete.h @@ -223,6 +223,16 @@ struct SMU74_Discrete_StateInfo { typedef struct SMU74_Discrete_StateInfo SMU74_Discrete_StateInfo; +struct SMU_QuadraticCoeffs { + int32_t m1; + uint32_t b; + + int16_t m2; + uint8_t m1_shift; + uint8_t m2_shift; +}; +typedef struct SMU_QuadraticCoeffs SMU_QuadraticCoeffs; + struct SMU74_Discrete_DpmTable { SMU74_PIDController GraphicsPIDController; @@ -258,7 +268,15 @@ struct SMU74_Discrete_DpmTable { uint8_t ThermOutPolarity; uint8_t ThermOutMode; uint8_t BootPhases; - uint32_t Reserved[4]; + + uint8_t VRHotLevel; + uint8_t LdoRefSel; + uint8_t Reserved1[2]; + uint16_t FanStartTemperature; + uint16_t FanStopTemperature; + uint16_t MaxVoltage; + uint16_t Reserved2; + uint32_t Reserved[1]; SMU74_Discrete_GraphicsLevel GraphicsLevel[SMU74_MAX_LEVELS_GRAPHICS]; SMU74_Discrete_MemoryLevel MemoryACPILevel; @@ -347,6 +365,8 @@ struct SMU74_Discrete_DpmTable { uint32_t CurrSclkPllRange; sclkFcwRange_t SclkFcwRangeTable[NUM_SCLK_RANGE]; + GB_VDROOP_TABLE_t BTCGB_VDROOP_TABLE[BTCGB_VDROOP_TABLE_MAX_ENTRIES]; + SMU_QuadraticCoeffs AVFSGB_VDROOP_TABLE[AVFSGB_VDROOP_TABLE_MAX_ENTRIES]; }; typedef struct SMU74_Discrete_DpmTable SMU74_Discrete_DpmTable; @@ -550,16 +570,6 @@ struct SMU7_AcpiScoreboard { typedef struct SMU7_AcpiScoreboard SMU7_AcpiScoreboard; -struct SMU_QuadraticCoeffs { - int32_t m1; - uint32_t b; - - int16_t m2; - uint8_t m1_shift; - uint8_t m2_shift; -}; -typedef struct SMU_QuadraticCoeffs SMU_QuadraticCoeffs; - struct SMU74_Discrete_PmFuses { uint8_t BapmVddCVidHiSidd[8]; uint8_t BapmVddCVidLoSidd[8]; @@ -821,6 +831,17 @@ typedef struct SMU7_GfxCuPgScoreboard SMU7_GfxCuPgScoreboard; #define DB_PCC_SHIFT 26 #define DB_EDC_SHIFT 27 +#define BTCGB0_Vdroop_Enable_MASK 0x1 +#define BTCGB1_Vdroop_Enable_MASK 0x2 +#define AVFSGB0_Vdroop_Enable_MASK 0x4 +#define AVFSGB1_Vdroop_Enable_MASK 0x8 + +#define BTCGB0_Vdroop_Enable_SHIFT 0 +#define BTCGB1_Vdroop_Enable_SHIFT 1 +#define AVFSGB0_Vdroop_Enable_SHIFT 2 +#define AVFSGB1_Vdroop_Enable_SHIFT 3 + + #pragma pack(pop) diff --git a/drivers/gpu/drm/amd/powerplay/inc/smumgr.h b/drivers/gpu/drm/amd/powerplay/inc/smumgr.h index fc9e3d1dd409..3c235f0177cd 100644 --- a/drivers/gpu/drm/amd/powerplay/inc/smumgr.h +++ b/drivers/gpu/drm/amd/powerplay/inc/smumgr.h @@ -131,6 +131,12 @@ extern int smu_free_memory(void *device, void *handle); smum_wait_on_indirect_register(smumgr, \ mm##port##_INDEX, index, value, mask) +#define SMUM_WAIT_INDIRECT_REGISTER(smumgr, port, reg, value, mask) \ + SMUM_WAIT_INDIRECT_REGISTER_GIVEN_INDEX(smumgr, port, ix##reg, value, mask) + +#define SMUM_WAIT_INDIRECT_FIELD(smumgr, port, reg, field, fieldval) \ + SMUM_WAIT_INDIRECT_REGISTER(smumgr, port, reg, (fieldval) << SMUM_FIELD_SHIFT(reg, field), \ + SMUM_FIELD_MASK(reg, field) ) #define SMUM_WAIT_REGISTER_UNEQUAL_GIVEN_INDEX(smumgr, \ index, value, mask) \ @@ -158,6 +164,10 @@ extern int smu_free_memory(void *device, void *handle); (SMUM_FIELD_MASK(reg, field) & ((field_val) << \ SMUM_FIELD_SHIFT(reg, field)))) +#define SMUM_READ_INDIRECT_FIELD(device, port, reg, field) \ + SMUM_GET_FIELD(cgs_read_ind_register(device, port, ix##reg), \ + reg, field) + #define SMUM_WAIT_VFPF_INDIRECT_REGISTER_GIVEN_INDEX(smumgr, \ port, index, value, mask) \ smum_wait_on_indirect_register(smumgr, \ @@ -191,6 +201,13 @@ extern int smu_free_memory(void *device, void *handle); SMUM_SET_FIELD(cgs_read_ind_register(device, port, ix##reg), \ reg, field, fieldval)) + +#define SMUM_WRITE_INDIRECT_FIELD(device, port, reg, field, fieldval) \ + cgs_write_ind_register(device, port, ix##reg, \ + SMUM_SET_FIELD(cgs_read_ind_register(device, port, ix##reg), \ + reg, field, fieldval)) + + #define SMUM_WAIT_VFPF_INDIRECT_FIELD(smumgr, port, reg, field, fieldval) \ SMUM_WAIT_VFPF_INDIRECT_REGISTER(smumgr, port, reg, \ (fieldval) << SMUM_FIELD_SHIFT(reg, field), \ @@ -200,4 +217,16 @@ extern int smu_free_memory(void *device, void *handle); SMUM_WAIT_VFPF_INDIRECT_REGISTER_UNEQUAL(smumgr, port, reg, \ (fieldval) << SMUM_FIELD_SHIFT(reg, field), \ SMUM_FIELD_MASK(reg, field)) + +#define SMUM_WAIT_INDIRECT_REGISTER_UNEQUAL_GIVEN_INDEX(smumgr, port, index, value, mask) \ + smum_wait_for_indirect_register_unequal(smumgr, \ + mm##port##_INDEX, index, value, mask) + +#define SMUM_WAIT_INDIRECT_REGISTER_UNEQUAL(smumgr, port, reg, value, mask) \ + SMUM_WAIT_INDIRECT_REGISTER_UNEQUAL_GIVEN_INDEX(smumgr, port, ix##reg, value, mask) + +#define SMUM_WAIT_INDIRECT_FIELD_UNEQUAL(smumgr, port, reg, field, fieldval) \ + SMUM_WAIT_INDIRECT_REGISTER_UNEQUAL(smumgr, port, reg, (fieldval) << SMUM_FIELD_SHIFT(reg, field), \ + SMUM_FIELD_MASK(reg, field) ) + #endif diff --git a/drivers/gpu/drm/amd/powerplay/smumgr/fiji_smumgr.c b/drivers/gpu/drm/amd/powerplay/smumgr/fiji_smumgr.c index 673a75c74e18..8e52a2e82db5 100644 --- a/drivers/gpu/drm/amd/powerplay/smumgr/fiji_smumgr.c +++ b/drivers/gpu/drm/amd/powerplay/smumgr/fiji_smumgr.c @@ -1006,10 +1006,16 @@ static int fiji_smu_init(struct pp_smumgr *smumgr) static int fiji_smu_fini(struct pp_smumgr *smumgr) { + struct fiji_smumgr *priv = (struct fiji_smumgr *)(smumgr->backend); + + smu_free_memory(smumgr->device, (void *)priv->header_buffer.handle); + if (smumgr->backend) { kfree(smumgr->backend); smumgr->backend = NULL; } + + cgs_rel_firmware(smumgr->device, CGS_UCODE_ID_SMU); return 0; } diff --git a/drivers/gpu/drm/amd/powerplay/smumgr/polaris10_smumgr.c b/drivers/gpu/drm/amd/powerplay/smumgr/polaris10_smumgr.c index de618ead9db8..5dba7c509710 100644 --- a/drivers/gpu/drm/amd/powerplay/smumgr/polaris10_smumgr.c +++ b/drivers/gpu/drm/amd/powerplay/smumgr/polaris10_smumgr.c @@ -52,19 +52,18 @@ static const SMU74_Discrete_GraphicsLevel avfs_graphics_level_polaris10[8] = { /* Min pcie DeepSleep Activity CgSpll CgSpll CcPwr CcPwr Sclk Enabled Enabled Voltage Power */ /* Voltage, DpmLevel, DivId, Level, FuncCntl3, FuncCntl4, DynRm, DynRm1 Did, Padding,ForActivity, ForThrottle, UpHyst, DownHyst, DownHyst, Throttle */ - { 0x3c0fd047, 0x00, 0x03, 0x1e00, 0x00200410, 0x87020000, 0, 0, 0x16, 0, 0x01, 0x01, 0x00, 0x00, 0x00, 0x00, { 0x30750000, 0, 0, 0, 0, 0, 0, 0 } }, - { 0xa00fd047, 0x01, 0x04, 0x1e00, 0x00800510, 0x87020000, 0, 0, 0x16, 0, 0x01, 0x01, 0x00, 0x00, 0x00, 0x00, { 0x409c0000, 0, 0, 0, 0, 0, 0, 0 } }, - { 0x0410d047, 0x01, 0x00, 0x1e00, 0x00600410, 0x87020000, 0, 0, 0x0e, 0, 0x01, 0x01, 0x00, 0x00, 0x00, 0x00, { 0x50c30000, 0, 0, 0, 0, 0, 0, 0 } }, - { 0x6810d047, 0x01, 0x00, 0x1e00, 0x00800410, 0x87020000, 0, 0, 0x0c, 0, 0x01, 0x01, 0x00, 0x00, 0x00, 0x00, { 0x60ea0000, 0, 0, 0, 0, 0, 0, 0 } }, - { 0xcc10d047, 0x01, 0x00, 0x1e00, 0x00e00410, 0x87020000, 0, 0, 0x0c, 0, 0x01, 0x01, 0x00, 0x00, 0x00, 0x00, { 0xe8fd0000, 0, 0, 0, 0, 0, 0, 0 } }, - { 0x3011d047, 0x01, 0x00, 0x1e00, 0x00400510, 0x87020000, 0, 0, 0x0c, 0, 0x01, 0x01, 0x00, 0x00, 0x00, 0x00, { 0x70110100, 0, 0, 0, 0, 0, 0, 0 } }, - { 0x9411d047, 0x01, 0x00, 0x1e00, 0x00a00510, 0x87020000, 0, 0, 0x0c, 0, 0x01, 0x01, 0x00, 0x00, 0x00, 0x00, { 0xf8240100, 0, 0, 0, 0, 0, 0, 0 } }, - { 0xf811d047, 0x01, 0x00, 0x1e00, 0x00000610, 0x87020000, 0, 0, 0x0c, 0, 0x01, 0x01, 0x00, 0x00, 0x00, 0x00, { 0x80380100, 0, 0, 0, 0, 0, 0, 0 } } + { 0x100ea446, 0x00, 0x03, 0x3200, 0, 0, 0, 0, 0, 0, 0x01, 0x01, 0x0a, 0x00, 0x00, 0x00, { 0x30750000, 0x3000, 0, 0x2600, 0, 0, 0x0004, 0x8f02, 0xffff, 0x2f00, 0x300e, 0x2700 } }, + { 0x400ea446, 0x01, 0x04, 0x3200, 0, 0, 0, 0, 0, 0, 0x01, 0x01, 0x0a, 0x00, 0x00, 0x00, { 0x409c0000, 0x2000, 0, 0x1e00, 1, 1, 0x0004, 0x8300, 0xffff, 0x1f00, 0xcb5e, 0x1a00 } }, + { 0x740ea446, 0x01, 0x00, 0x3200, 0, 0, 0, 0, 0, 0, 0x01, 0x01, 0x0a, 0x00, 0x00, 0x00, { 0x50c30000, 0x2800, 0, 0x2000, 1, 1, 0x0004, 0x0c02, 0xffff, 0x2700, 0x6433, 0x2100 } }, + { 0xa40ea446, 0x01, 0x00, 0x3200, 0, 0, 0, 0, 0, 0, 0x01, 0x01, 0x0a, 0x00, 0x00, 0x00, { 0x60ea0000, 0x3000, 0, 0x2600, 1, 1, 0x0004, 0x8f02, 0xffff, 0x2f00, 0x300e, 0x2700 } }, + { 0xd80ea446, 0x01, 0x00, 0x3200, 0, 0, 0, 0, 0, 0, 0x01, 0x01, 0x0a, 0x00, 0x00, 0x00, { 0x70110100, 0x3800, 0, 0x2c00, 1, 1, 0x0004, 0x1203, 0xffff, 0x3600, 0xc9e2, 0x2e00 } }, + { 0x3c0fa446, 0x01, 0x00, 0x3200, 0, 0, 0, 0, 0, 0, 0x01, 0x01, 0x0a, 0x00, 0x00, 0x00, { 0x80380100, 0x2000, 0, 0x1e00, 2, 1, 0x0004, 0x8300, 0xffff, 0x1f00, 0xcb5e, 0x1a00 } }, + { 0x6c0fa446, 0x01, 0x00, 0x3200, 0, 0, 0, 0, 0, 0, 0x01, 0x01, 0x0a, 0x00, 0x00, 0x00, { 0x905f0100, 0x2400, 0, 0x1e00, 2, 1, 0x0004, 0x8901, 0xffff, 0x2300, 0x314c, 0x1d00 } }, + { 0xa00fa446, 0x01, 0x00, 0x3200, 0, 0, 0, 0, 0, 0, 0x01, 0x01, 0x0a, 0x00, 0x00, 0x00, { 0xa0860100, 0x2800, 0, 0x2000, 2, 1, 0x0004, 0x0c02, 0xffff, 0x2700, 0x6433, 0x2100 } } }; static const SMU74_Discrete_MemoryLevel avfs_memory_level_polaris10 = - {0x50140000, 0x50140000, 0x00320000, 0x00, 0x00, - 0x00, 0x10, 0x00, 0x00, 0x00, 0x00, 0x0000, 0x00, 0x00}; + {0x100ea446, 0, 0x30750000, 0x01, 0x01, 0x01, 0x00, 0x00, 0x64, 0x00, 0x00, 0x1f00, 0x00, 0x00}; /** * Set the address for reading/writing the SMC SRAM space. @@ -219,6 +218,18 @@ bool polaris10_is_smc_ram_running(struct pp_smumgr *smumgr) && (0x20100 <= cgs_read_ind_register(smumgr->device, CGS_IND_REG__SMC, ixSMC_PC_C))); } +static bool polaris10_is_hw_avfs_present(struct pp_smumgr *smumgr) +{ + uint32_t efuse; + + efuse = cgs_read_ind_register(smumgr->device, CGS_IND_REG__SMC, ixSMU_EFUSE_0 + (49*4)); + efuse &= 0x00000001; + if (efuse) + return true; + + return false; +} + /** * Send a message to the SMC, and wait for its response. * @@ -228,21 +239,27 @@ bool polaris10_is_smc_ram_running(struct pp_smumgr *smumgr) */ int polaris10_send_msg_to_smc(struct pp_smumgr *smumgr, uint16_t msg) { + int ret; + if (!polaris10_is_smc_ram_running(smumgr)) return -1; + SMUM_WAIT_FIELD_UNEQUAL(smumgr, SMC_RESP_0, SMC_RESP, 0); - if (1 != SMUM_READ_FIELD(smumgr->device, SMC_RESP_0, SMC_RESP)) - printk("Failed to send Previous Message.\n"); + ret = SMUM_READ_FIELD(smumgr->device, SMC_RESP_0, SMC_RESP); + if (ret != 1) + printk("\n failed to send pre message %x ret is %d \n", msg, ret); cgs_write_register(smumgr->device, mmSMC_MESSAGE_0, msg); SMUM_WAIT_FIELD_UNEQUAL(smumgr, SMC_RESP_0, SMC_RESP, 0); - if (1 != SMUM_READ_FIELD(smumgr->device, SMC_RESP_0, SMC_RESP)) - printk("Failed to send Message.\n"); + ret = SMUM_READ_FIELD(smumgr->device, SMC_RESP_0, SMC_RESP); + + if (ret != 1) + printk("\n failed to send message %x ret is %d \n", msg, ret); return 0; } @@ -469,6 +486,7 @@ int polaris10_smu_fini(struct pp_smumgr *smumgr) kfree(smumgr->backend); smumgr->backend = NULL; } + cgs_rel_firmware(smumgr->device, CGS_UCODE_ID_SMU); return 0; } @@ -952,6 +970,11 @@ static int polaris10_smu_init(struct pp_smumgr *smumgr) (cgs_handle_t)smu_data->smu_buffer.handle); return -1;); + if (polaris10_is_hw_avfs_present(smumgr)) + smu_data->avfs.avfs_btc_status = AVFS_BTC_BOOT; + else + smu_data->avfs.avfs_btc_status = AVFS_BTC_NOTSUPPORTED; + return 0; } diff --git a/drivers/gpu/drm/amd/powerplay/smumgr/smumgr.c b/drivers/gpu/drm/amd/powerplay/smumgr/smumgr.c index c483baf6b4fb..7723473e51a0 100644 --- a/drivers/gpu/drm/amd/powerplay/smumgr/smumgr.c +++ b/drivers/gpu/drm/amd/powerplay/smumgr/smumgr.c @@ -23,6 +23,7 @@ #include <linux/types.h> #include <linux/kernel.h> #include <linux/slab.h> +#include <drm/amdgpu_drm.h> #include "pp_instance.h" #include "smumgr.h" #include "cgs_common.h" @@ -52,10 +53,10 @@ int smum_init(struct amd_pp_init *pp_init, struct pp_instance *handle) handle->smu_mgr = smumgr; switch (smumgr->chip_family) { - case AMD_FAMILY_CZ: + case AMDGPU_FAMILY_CZ: cz_smum_init(smumgr); break; - case AMD_FAMILY_VI: + case AMDGPU_FAMILY_VI: switch (smumgr->chip_id) { case CHIP_TONGA: tonga_smum_init(smumgr); @@ -81,6 +82,7 @@ int smum_init(struct amd_pp_init *pp_init, struct pp_instance *handle) int smum_fini(struct pp_smumgr *smumgr) { + kfree(smumgr->device); kfree(smumgr); return 0; } diff --git a/drivers/gpu/drm/amd/powerplay/smumgr/tonga_smumgr.c b/drivers/gpu/drm/amd/powerplay/smumgr/tonga_smumgr.c index 32820b680d88..f42c536b3af1 100644 --- a/drivers/gpu/drm/amd/powerplay/smumgr/tonga_smumgr.c +++ b/drivers/gpu/drm/amd/powerplay/smumgr/tonga_smumgr.c @@ -328,10 +328,17 @@ int tonga_write_smc_sram_dword(struct pp_smumgr *smumgr, static int tonga_smu_fini(struct pp_smumgr *smumgr) { + struct tonga_smumgr *priv = (struct tonga_smumgr *)(smumgr->backend); + + smu_free_memory(smumgr->device, (void *)priv->smu_buffer.handle); + smu_free_memory(smumgr->device, (void *)priv->header_buffer.handle); + if (smumgr->backend != NULL) { kfree(smumgr->backend); smumgr->backend = NULL; } + + cgs_rel_firmware(smumgr->device, CGS_UCODE_ID_SMU); return 0; } @@ -472,7 +479,6 @@ static int tonga_request_smu_reload_fw(struct pp_smumgr *smumgr) struct tonga_smumgr *tonga_smu = (struct tonga_smumgr *)(smumgr->backend); uint16_t fw_to_load; - int result = 0; struct SMU_DRAMData_TOC *toc; /** * First time this gets called during SmuMgr init, @@ -556,7 +562,7 @@ static int tonga_request_smu_reload_fw(struct pp_smumgr *smumgr) smumgr, PPSMC_MSG_LoadUcodes, fw_to_load), "Fail to Request SMU Load uCode", return 0); - return result; + return 0; } static int tonga_request_smu_load_specific_fw(struct pp_smumgr *smumgr, diff --git a/drivers/gpu/drm/amd/scheduler/gpu_sched_trace.h b/drivers/gpu/drm/amd/scheduler/gpu_sched_trace.h index c89dc777768f..b961a1c6caf3 100644 --- a/drivers/gpu/drm/amd/scheduler/gpu_sched_trace.h +++ b/drivers/gpu/drm/amd/scheduler/gpu_sched_trace.h @@ -26,7 +26,7 @@ TRACE_EVENT(amd_sched_job, TP_fast_assign( __entry->entity = sched_job->s_entity; __entry->sched_job = sched_job; - __entry->fence = &sched_job->s_fence->base; + __entry->fence = &sched_job->s_fence->finished; __entry->name = sched_job->sched->name; __entry->job_count = kfifo_len( &sched_job->s_entity->job_queue) / sizeof(sched_job); @@ -46,7 +46,7 @@ TRACE_EVENT(amd_sched_process_job, ), TP_fast_assign( - __entry->fence = &fence->base; + __entry->fence = &fence->finished; ), TP_printk("fence=%p signaled", __entry->fence) ); diff --git a/drivers/gpu/drm/amd/scheduler/gpu_scheduler.c b/drivers/gpu/drm/amd/scheduler/gpu_scheduler.c index c16248cee779..ef312bb75fda 100644 --- a/drivers/gpu/drm/amd/scheduler/gpu_scheduler.c +++ b/drivers/gpu/drm/amd/scheduler/gpu_scheduler.c @@ -32,6 +32,7 @@ static bool amd_sched_entity_is_ready(struct amd_sched_entity *entity); static void amd_sched_wakeup(struct amd_gpu_scheduler *sched); +static void amd_sched_process_job(struct fence *f, struct fence_cb *cb); struct kmem_cache *sched_fence_slab; atomic_t sched_fence_slab_ref = ATOMIC_INIT(0); @@ -140,7 +141,7 @@ int amd_sched_entity_init(struct amd_gpu_scheduler *sched, return r; atomic_set(&entity->fence_seq, 0); - entity->fence_context = fence_context_alloc(1); + entity->fence_context = fence_context_alloc(2); return 0; } @@ -251,17 +252,21 @@ static bool amd_sched_entity_add_dependency_cb(struct amd_sched_entity *entity) s_fence = to_amd_sched_fence(fence); if (s_fence && s_fence->sched == sched) { - /* Fence is from the same scheduler */ - if (test_bit(AMD_SCHED_FENCE_SCHEDULED_BIT, &fence->flags)) { - /* Ignore it when it is already scheduled */ - fence_put(entity->dependency); - return false; - } - /* Wait for fence to be scheduled */ - entity->cb.func = amd_sched_entity_clear_dep; - list_add_tail(&entity->cb.node, &s_fence->scheduled_cb); - return true; + /* + * Fence is from the same scheduler, only need to wait for + * it to be scheduled + */ + fence = fence_get(&s_fence->scheduled); + fence_put(entity->dependency); + entity->dependency = fence; + if (!fence_add_callback(fence, &entity->cb, + amd_sched_entity_clear_dep)) + return true; + + /* Ignore it when it is already scheduled */ + fence_put(fence); + return false; } if (!fence_add_callback(entity->dependency, &entity->cb, @@ -319,46 +324,114 @@ static bool amd_sched_entity_in(struct amd_sched_job *sched_job) return added; } -static void amd_sched_free_job(struct fence *f, struct fence_cb *cb) { - struct amd_sched_job *job = container_of(cb, struct amd_sched_job, cb_free_job); - schedule_work(&job->work_free_job); -} - /* job_finish is called after hw fence signaled, and * the job had already been deleted from ring_mirror_list */ -void amd_sched_job_finish(struct amd_sched_job *s_job) +static void amd_sched_job_finish(struct work_struct *work) { - struct amd_sched_job *next; + struct amd_sched_job *s_job = container_of(work, struct amd_sched_job, + finish_work); struct amd_gpu_scheduler *sched = s_job->sched; + /* remove job from ring_mirror_list */ + spin_lock(&sched->job_list_lock); + list_del_init(&s_job->node); if (sched->timeout != MAX_SCHEDULE_TIMEOUT) { - if (cancel_delayed_work(&s_job->work_tdr)) - amd_sched_job_put(s_job); + struct amd_sched_job *next; + + spin_unlock(&sched->job_list_lock); + cancel_delayed_work_sync(&s_job->work_tdr); + spin_lock(&sched->job_list_lock); /* queue TDR for next job */ next = list_first_entry_or_null(&sched->ring_mirror_list, struct amd_sched_job, node); - if (next) { - INIT_DELAYED_WORK(&next->work_tdr, s_job->timeout_callback); - amd_sched_job_get(next); + if (next) schedule_delayed_work(&next->work_tdr, sched->timeout); - } } + spin_unlock(&sched->job_list_lock); + sched->ops->free_job(s_job); } -void amd_sched_job_begin(struct amd_sched_job *s_job) +static void amd_sched_job_finish_cb(struct fence *f, struct fence_cb *cb) +{ + struct amd_sched_job *job = container_of(cb, struct amd_sched_job, + finish_cb); + schedule_work(&job->finish_work); +} + +static void amd_sched_job_begin(struct amd_sched_job *s_job) { struct amd_gpu_scheduler *sched = s_job->sched; + spin_lock(&sched->job_list_lock); + list_add_tail(&s_job->node, &sched->ring_mirror_list); if (sched->timeout != MAX_SCHEDULE_TIMEOUT && - list_first_entry_or_null(&sched->ring_mirror_list, struct amd_sched_job, node) == s_job) - { - INIT_DELAYED_WORK(&s_job->work_tdr, s_job->timeout_callback); - amd_sched_job_get(s_job); + list_first_entry_or_null(&sched->ring_mirror_list, + struct amd_sched_job, node) == s_job) + schedule_delayed_work(&s_job->work_tdr, sched->timeout); + spin_unlock(&sched->job_list_lock); +} + +static void amd_sched_job_timedout(struct work_struct *work) +{ + struct amd_sched_job *job = container_of(work, struct amd_sched_job, + work_tdr.work); + + job->sched->ops->timedout_job(job); +} + +void amd_sched_hw_job_reset(struct amd_gpu_scheduler *sched) +{ + struct amd_sched_job *s_job; + + spin_lock(&sched->job_list_lock); + list_for_each_entry_reverse(s_job, &sched->ring_mirror_list, node) { + if (fence_remove_callback(s_job->s_fence->parent, &s_job->s_fence->cb)) { + fence_put(s_job->s_fence->parent); + s_job->s_fence->parent = NULL; + } + } + atomic_set(&sched->hw_rq_count, 0); + spin_unlock(&sched->job_list_lock); +} + +void amd_sched_job_recovery(struct amd_gpu_scheduler *sched) +{ + struct amd_sched_job *s_job, *tmp; + int r; + + spin_lock(&sched->job_list_lock); + s_job = list_first_entry_or_null(&sched->ring_mirror_list, + struct amd_sched_job, node); + if (s_job) schedule_delayed_work(&s_job->work_tdr, sched->timeout); + + list_for_each_entry_safe(s_job, tmp, &sched->ring_mirror_list, node) { + struct amd_sched_fence *s_fence = s_job->s_fence; + struct fence *fence; + + spin_unlock(&sched->job_list_lock); + fence = sched->ops->run_job(s_job); + atomic_inc(&sched->hw_rq_count); + if (fence) { + s_fence->parent = fence_get(fence); + r = fence_add_callback(fence, &s_fence->cb, + amd_sched_process_job); + if (r == -ENOENT) + amd_sched_process_job(fence, &s_fence->cb); + else if (r) + DRM_ERROR("fence add callback failed (%d)\n", + r); + fence_put(fence); + } else { + DRM_ERROR("Failed to run job!\n"); + amd_sched_process_job(NULL, &s_fence->cb); + } + spin_lock(&sched->job_list_lock); } + spin_unlock(&sched->job_list_lock); } /** @@ -372,36 +445,29 @@ void amd_sched_entity_push_job(struct amd_sched_job *sched_job) { struct amd_sched_entity *entity = sched_job->s_entity; - sched_job->use_sched = 1; - fence_add_callback(&sched_job->s_fence->base, - &sched_job->cb_free_job, amd_sched_free_job); trace_amd_sched_job(sched_job); + fence_add_callback(&sched_job->s_fence->finished, &sched_job->finish_cb, + amd_sched_job_finish_cb); wait_event(entity->sched->job_scheduled, amd_sched_entity_in(sched_job)); } /* init a sched_job with basic field */ int amd_sched_job_init(struct amd_sched_job *job, - struct amd_gpu_scheduler *sched, - struct amd_sched_entity *entity, - void (*timeout_cb)(struct work_struct *work), - void (*free_cb)(struct kref *refcount), - void *owner, struct fence **fence) + struct amd_gpu_scheduler *sched, + struct amd_sched_entity *entity, + void *owner) { - INIT_LIST_HEAD(&job->node); - kref_init(&job->refcount); job->sched = sched; job->s_entity = entity; job->s_fence = amd_sched_fence_create(entity, owner); if (!job->s_fence) return -ENOMEM; - job->s_fence->s_job = job; - job->timeout_callback = timeout_cb; - job->free_callback = free_cb; + INIT_WORK(&job->finish_work, amd_sched_job_finish); + INIT_LIST_HEAD(&job->node); + INIT_DELAYED_WORK(&job->work_tdr, amd_sched_job_timedout); - if (fence) - *fence = &job->s_fence->base; return 0; } @@ -450,23 +516,25 @@ static void amd_sched_process_job(struct fence *f, struct fence_cb *cb) struct amd_sched_fence *s_fence = container_of(cb, struct amd_sched_fence, cb); struct amd_gpu_scheduler *sched = s_fence->sched; - unsigned long flags; atomic_dec(&sched->hw_rq_count); - - /* remove job from ring_mirror_list */ - spin_lock_irqsave(&sched->job_list_lock, flags); - list_del_init(&s_fence->s_job->node); - sched->ops->finish_job(s_fence->s_job); - spin_unlock_irqrestore(&sched->job_list_lock, flags); - - amd_sched_fence_signal(s_fence); + amd_sched_fence_finished(s_fence); trace_amd_sched_process_job(s_fence); - fence_put(&s_fence->base); + fence_put(&s_fence->finished); wake_up_interruptible(&sched->wake_up_worker); } +static bool amd_sched_blocked(struct amd_gpu_scheduler *sched) +{ + if (kthread_should_park()) { + kthread_parkme(); + return true; + } + + return false; +} + static int amd_sched_main(void *param) { struct sched_param sparam = {.sched_priority = 1}; @@ -476,14 +544,15 @@ static int amd_sched_main(void *param) sched_setscheduler(current, SCHED_FIFO, &sparam); while (!kthread_should_stop()) { - struct amd_sched_entity *entity; + struct amd_sched_entity *entity = NULL; struct amd_sched_fence *s_fence; struct amd_sched_job *sched_job; struct fence *fence; wait_event_interruptible(sched->wake_up_worker, - (entity = amd_sched_select_entity(sched)) || - kthread_should_stop()); + (!amd_sched_blocked(sched) && + (entity = amd_sched_select_entity(sched))) || + kthread_should_stop()); if (!entity) continue; @@ -495,16 +564,19 @@ static int amd_sched_main(void *param) s_fence = sched_job->s_fence; atomic_inc(&sched->hw_rq_count); - amd_sched_job_pre_schedule(sched, sched_job); + amd_sched_job_begin(sched_job); + fence = sched->ops->run_job(sched_job); amd_sched_fence_scheduled(s_fence); if (fence) { + s_fence->parent = fence_get(fence); r = fence_add_callback(fence, &s_fence->cb, amd_sched_process_job); if (r == -ENOENT) amd_sched_process_job(fence, &s_fence->cb); else if (r) - DRM_ERROR("fence add callback failed (%d)\n", r); + DRM_ERROR("fence add callback failed (%d)\n", + r); fence_put(fence); } else { DRM_ERROR("Failed to run job!\n"); diff --git a/drivers/gpu/drm/amd/scheduler/gpu_scheduler.h b/drivers/gpu/drm/amd/scheduler/gpu_scheduler.h index 070095a9433c..7cbbbfb502ef 100644 --- a/drivers/gpu/drm/amd/scheduler/gpu_scheduler.h +++ b/drivers/gpu/drm/amd/scheduler/gpu_scheduler.h @@ -27,8 +27,6 @@ #include <linux/kfifo.h> #include <linux/fence.h> -#define AMD_SCHED_FENCE_SCHEDULED_BIT FENCE_FLAG_USER_BITS - struct amd_gpu_scheduler; struct amd_sched_rq; @@ -68,36 +66,34 @@ struct amd_sched_rq { }; struct amd_sched_fence { - struct fence base; + struct fence scheduled; + struct fence finished; struct fence_cb cb; - struct list_head scheduled_cb; + struct fence *parent; struct amd_gpu_scheduler *sched; spinlock_t lock; void *owner; - struct amd_sched_job *s_job; }; struct amd_sched_job { - struct kref refcount; struct amd_gpu_scheduler *sched; struct amd_sched_entity *s_entity; struct amd_sched_fence *s_fence; - bool use_sched; /* true if the job goes to scheduler */ - struct fence_cb cb_free_job; - struct work_struct work_free_job; - struct list_head node; - struct delayed_work work_tdr; - void (*timeout_callback) (struct work_struct *work); - void (*free_callback)(struct kref *refcount); + struct fence_cb finish_cb; + struct work_struct finish_work; + struct list_head node; + struct delayed_work work_tdr; }; -extern const struct fence_ops amd_sched_fence_ops; +extern const struct fence_ops amd_sched_fence_ops_scheduled; +extern const struct fence_ops amd_sched_fence_ops_finished; static inline struct amd_sched_fence *to_amd_sched_fence(struct fence *f) { - struct amd_sched_fence *__f = container_of(f, struct amd_sched_fence, base); + if (f->ops == &amd_sched_fence_ops_scheduled) + return container_of(f, struct amd_sched_fence, scheduled); - if (__f->base.ops == &amd_sched_fence_ops) - return __f; + if (f->ops == &amd_sched_fence_ops_finished) + return container_of(f, struct amd_sched_fence, finished); return NULL; } @@ -109,8 +105,8 @@ static inline struct amd_sched_fence *to_amd_sched_fence(struct fence *f) struct amd_sched_backend_ops { struct fence *(*dependency)(struct amd_sched_job *sched_job); struct fence *(*run_job)(struct amd_sched_job *sched_job); - void (*begin_job)(struct amd_sched_job *sched_job); - void (*finish_job)(struct amd_sched_job *sched_job); + void (*timedout_job)(struct amd_sched_job *sched_job); + void (*free_job)(struct amd_sched_job *sched_job); }; enum amd_sched_priority { @@ -152,25 +148,11 @@ void amd_sched_entity_push_job(struct amd_sched_job *sched_job); struct amd_sched_fence *amd_sched_fence_create( struct amd_sched_entity *s_entity, void *owner); void amd_sched_fence_scheduled(struct amd_sched_fence *fence); -void amd_sched_fence_signal(struct amd_sched_fence *fence); +void amd_sched_fence_finished(struct amd_sched_fence *fence); int amd_sched_job_init(struct amd_sched_job *job, - struct amd_gpu_scheduler *sched, - struct amd_sched_entity *entity, - void (*timeout_cb)(struct work_struct *work), - void (*free_cb)(struct kref* refcount), - void *owner, struct fence **fence); -void amd_sched_job_pre_schedule(struct amd_gpu_scheduler *sched , - struct amd_sched_job *s_job); -void amd_sched_job_finish(struct amd_sched_job *s_job); -void amd_sched_job_begin(struct amd_sched_job *s_job); -static inline void amd_sched_job_get(struct amd_sched_job *job) { - if (job) - kref_get(&job->refcount); -} - -static inline void amd_sched_job_put(struct amd_sched_job *job) { - if (job) - kref_put(&job->refcount, job->free_callback); -} - + struct amd_gpu_scheduler *sched, + struct amd_sched_entity *entity, + void *owner); +void amd_sched_hw_job_reset(struct amd_gpu_scheduler *sched); +void amd_sched_job_recovery(struct amd_gpu_scheduler *sched); #endif diff --git a/drivers/gpu/drm/amd/scheduler/sched_fence.c b/drivers/gpu/drm/amd/scheduler/sched_fence.c index 2a732c490375..6b63beaf7574 100644 --- a/drivers/gpu/drm/amd/scheduler/sched_fence.c +++ b/drivers/gpu/drm/amd/scheduler/sched_fence.c @@ -27,7 +27,8 @@ #include <drm/drmP.h> #include "gpu_scheduler.h" -struct amd_sched_fence *amd_sched_fence_create(struct amd_sched_entity *s_entity, void *owner) +struct amd_sched_fence *amd_sched_fence_create(struct amd_sched_entity *entity, + void *owner) { struct amd_sched_fence *fence = NULL; unsigned seq; @@ -36,46 +37,37 @@ struct amd_sched_fence *amd_sched_fence_create(struct amd_sched_entity *s_entity if (fence == NULL) return NULL; - INIT_LIST_HEAD(&fence->scheduled_cb); fence->owner = owner; - fence->sched = s_entity->sched; + fence->sched = entity->sched; spin_lock_init(&fence->lock); - seq = atomic_inc_return(&s_entity->fence_seq); - fence_init(&fence->base, &amd_sched_fence_ops, &fence->lock, - s_entity->fence_context, seq); + seq = atomic_inc_return(&entity->fence_seq); + fence_init(&fence->scheduled, &amd_sched_fence_ops_scheduled, + &fence->lock, entity->fence_context, seq); + fence_init(&fence->finished, &amd_sched_fence_ops_finished, + &fence->lock, entity->fence_context + 1, seq); return fence; } -void amd_sched_fence_signal(struct amd_sched_fence *fence) +void amd_sched_fence_scheduled(struct amd_sched_fence *fence) { - int ret = fence_signal(&fence->base); + int ret = fence_signal(&fence->scheduled); + if (!ret) - FENCE_TRACE(&fence->base, "signaled from irq context\n"); + FENCE_TRACE(&fence->scheduled, "signaled from irq context\n"); else - FENCE_TRACE(&fence->base, "was already signaled\n"); -} - -void amd_sched_job_pre_schedule(struct amd_gpu_scheduler *sched , - struct amd_sched_job *s_job) -{ - unsigned long flags; - spin_lock_irqsave(&sched->job_list_lock, flags); - list_add_tail(&s_job->node, &sched->ring_mirror_list); - sched->ops->begin_job(s_job); - spin_unlock_irqrestore(&sched->job_list_lock, flags); + FENCE_TRACE(&fence->scheduled, "was already signaled\n"); } -void amd_sched_fence_scheduled(struct amd_sched_fence *s_fence) +void amd_sched_fence_finished(struct amd_sched_fence *fence) { - struct fence_cb *cur, *tmp; + int ret = fence_signal(&fence->finished); - set_bit(AMD_SCHED_FENCE_SCHEDULED_BIT, &s_fence->base.flags); - list_for_each_entry_safe(cur, tmp, &s_fence->scheduled_cb, node) { - list_del_init(&cur->node); - cur->func(&s_fence->base, cur); - } + if (!ret) + FENCE_TRACE(&fence->finished, "signaled from irq context\n"); + else + FENCE_TRACE(&fence->finished, "was already signaled\n"); } static const char *amd_sched_fence_get_driver_name(struct fence *fence) @@ -105,6 +97,8 @@ static void amd_sched_fence_free(struct rcu_head *rcu) { struct fence *f = container_of(rcu, struct fence, rcu); struct amd_sched_fence *fence = to_amd_sched_fence(f); + + fence_put(fence->parent); kmem_cache_free(sched_fence_slab, fence); } @@ -116,16 +110,41 @@ static void amd_sched_fence_free(struct rcu_head *rcu) * This function is called when the reference count becomes zero. * It just RCU schedules freeing up the fence. */ -static void amd_sched_fence_release(struct fence *f) +static void amd_sched_fence_release_scheduled(struct fence *f) { - call_rcu(&f->rcu, amd_sched_fence_free); + struct amd_sched_fence *fence = to_amd_sched_fence(f); + + call_rcu(&fence->finished.rcu, amd_sched_fence_free); } -const struct fence_ops amd_sched_fence_ops = { +/** + * amd_sched_fence_release_scheduled - drop extra reference + * + * @f: fence + * + * Drop the extra reference from the scheduled fence to the base fence. + */ +static void amd_sched_fence_release_finished(struct fence *f) +{ + struct amd_sched_fence *fence = to_amd_sched_fence(f); + + fence_put(&fence->scheduled); +} + +const struct fence_ops amd_sched_fence_ops_scheduled = { + .get_driver_name = amd_sched_fence_get_driver_name, + .get_timeline_name = amd_sched_fence_get_timeline_name, + .enable_signaling = amd_sched_fence_enable_signaling, + .signaled = NULL, + .wait = fence_default_wait, + .release = amd_sched_fence_release_scheduled, +}; + +const struct fence_ops amd_sched_fence_ops_finished = { .get_driver_name = amd_sched_fence_get_driver_name, .get_timeline_name = amd_sched_fence_get_timeline_name, .enable_signaling = amd_sched_fence_enable_signaling, .signaled = NULL, .wait = fence_default_wait, - .release = amd_sched_fence_release, + .release = amd_sched_fence_release_finished, }; diff --git a/drivers/gpu/drm/arc/Kconfig b/drivers/gpu/drm/arc/Kconfig index f9a13b658fea..f47d88ba4fa5 100644 --- a/drivers/gpu/drm/arc/Kconfig +++ b/drivers/gpu/drm/arc/Kconfig @@ -2,7 +2,6 @@ config DRM_ARCPGU tristate "ARC PGU" depends on DRM && OF select DRM_KMS_CMA_HELPER - select DRM_KMS_FB_HELPER select DRM_KMS_HELPER help Choose this option if you have an ARC PGU controller. diff --git a/drivers/gpu/drm/arc/arcpgu_drv.c b/drivers/gpu/drm/arc/arcpgu_drv.c index ccbdadb108dc..6d4ff34737cb 100644 --- a/drivers/gpu/drm/arc/arcpgu_drv.c +++ b/drivers/gpu/drm/arc/arcpgu_drv.c @@ -28,8 +28,7 @@ static void arcpgu_fb_output_poll_changed(struct drm_device *dev) { struct arcpgu_drm_private *arcpgu = dev->dev_private; - if (arcpgu->fbdev) - drm_fbdev_cma_hotplug_event(arcpgu->fbdev); + drm_fbdev_cma_hotplug_event(arcpgu->fbdev); } static struct drm_mode_config_funcs arcpgu_drm_modecfg_funcs = { @@ -49,7 +48,7 @@ static void arcpgu_setup_mode_config(struct drm_device *drm) drm->mode_config.funcs = &arcpgu_drm_modecfg_funcs; } -int arcpgu_gem_mmap(struct file *filp, struct vm_area_struct *vma) +static int arcpgu_gem_mmap(struct file *filp, struct vm_area_struct *vma) { int ret; @@ -104,10 +103,8 @@ static int arcpgu_load(struct drm_device *drm) res = platform_get_resource(pdev, IORESOURCE_MEM, 0); arcpgu->regs = devm_ioremap_resource(&pdev->dev, res); - if (IS_ERR(arcpgu->regs)) { - dev_err(drm->dev, "Could not remap IO mem\n"); + if (IS_ERR(arcpgu->regs)) return PTR_ERR(arcpgu->regs); - } dev_info(drm->dev, "arc_pgu ID: 0x%x\n", arc_pgu_read(arcpgu, ARCPGU_REG_ID)); @@ -127,10 +124,11 @@ static int arcpgu_load(struct drm_device *drm) encoder_node = of_parse_phandle(drm->dev->of_node, "encoder-slave", 0); if (encoder_node) { ret = arcpgu_drm_hdmi_init(drm, encoder_node); + of_node_put(encoder_node); if (ret < 0) return ret; } else { - ret = arcpgu_drm_sim_init(drm, 0); + ret = arcpgu_drm_sim_init(drm, NULL); if (ret < 0) return ret; } @@ -151,7 +149,7 @@ static int arcpgu_load(struct drm_device *drm) return 0; } -int arcpgu_unload(struct drm_device *drm) +static int arcpgu_unload(struct drm_device *drm) { struct arcpgu_drm_private *arcpgu = drm->dev_private; diff --git a/drivers/gpu/drm/arm/Kconfig b/drivers/gpu/drm/arm/Kconfig index 1b2906568a48..9a18e1bd57b4 100644 --- a/drivers/gpu/drm/arm/Kconfig +++ b/drivers/gpu/drm/arm/Kconfig @@ -9,7 +9,6 @@ config DRM_HDLCD depends on COMMON_CLK select DRM_ARM select DRM_KMS_HELPER - select DRM_KMS_FB_HELPER select DRM_KMS_CMA_HELPER help Choose this option if you have an ARM High Definition Colour LCD diff --git a/drivers/gpu/drm/arm/hdlcd_drv.c b/drivers/gpu/drm/arm/hdlcd_drv.c index 74279be20b75..d83b46a30327 100644 --- a/drivers/gpu/drm/arm/hdlcd_drv.c +++ b/drivers/gpu/drm/arm/hdlcd_drv.c @@ -102,8 +102,7 @@ static void hdlcd_fb_output_poll_changed(struct drm_device *drm) { struct hdlcd_drm_private *hdlcd = drm->dev_private; - if (hdlcd->fbdev) - drm_fbdev_cma_hotplug_event(hdlcd->fbdev); + drm_fbdev_cma_hotplug_event(hdlcd->fbdev); } static const struct drm_mode_config_funcs hdlcd_mode_config_funcs = { diff --git a/drivers/gpu/drm/arm/malidp_drv.c b/drivers/gpu/drm/arm/malidp_drv.c index e5b44e92f8cf..82171d223f2d 100644 --- a/drivers/gpu/drm/arm/malidp_drv.c +++ b/drivers/gpu/drm/arm/malidp_drv.c @@ -257,6 +257,7 @@ static int malidp_bind(struct device *dev) { struct resource *res; struct drm_device *drm; + struct device_node *ep; struct malidp_drm *malidp; struct malidp_hw_device *hwdev; struct platform_device *pdev = to_platform_device(dev); @@ -284,10 +285,8 @@ static int malidp_bind(struct device *dev) res = platform_get_resource(pdev, IORESOURCE_MEM, 0); hwdev->regs = devm_ioremap_resource(dev, res); - if (IS_ERR(hwdev->regs)) { - DRM_ERROR("Failed to map control registers area\n"); + if (IS_ERR(hwdev->regs)) return PTR_ERR(hwdev->regs); - } hwdev->pclk = devm_clk_get(dev, "pclk"); if (IS_ERR(hwdev->pclk)) @@ -360,11 +359,14 @@ static int malidp_bind(struct device *dev) goto register_fail; /* Set the CRTC's port so that the encoder component can find it */ - malidp->crtc.port = of_graph_get_next_endpoint(dev->of_node, NULL); + ep = of_graph_get_next_endpoint(dev->of_node, NULL); + if (!ep) { + ret = -EINVAL; + goto port_fail; + } + malidp->crtc.port = of_get_next_parent(ep); ret = component_bind_all(dev, drm); - of_node_put(malidp->crtc.port); - if (ret) { DRM_ERROR("Failed to bind all components\n"); goto bind_fail; @@ -402,6 +404,9 @@ vblank_fail: irq_init_fail: component_unbind_all(dev, drm); bind_fail: + of_node_put(malidp->crtc.port); + malidp->crtc.port = NULL; +port_fail: drm_dev_unregister(drm); register_fail: malidp_de_planes_destroy(drm); @@ -435,6 +440,8 @@ static void malidp_unbind(struct device *dev) malidp_de_irq_fini(drm); drm_vblank_cleanup(drm); component_unbind_all(dev, drm); + of_node_put(malidp->crtc.port); + malidp->crtc.port = NULL; drm_dev_unregister(drm); malidp_de_planes_destroy(drm); drm_mode_config_cleanup(drm); diff --git a/drivers/gpu/drm/armada/Kconfig b/drivers/gpu/drm/armada/Kconfig index eb773e9af313..15f3ecfb16f1 100644 --- a/drivers/gpu/drm/armada/Kconfig +++ b/drivers/gpu/drm/armada/Kconfig @@ -1,11 +1,7 @@ config DRM_ARMADA tristate "DRM support for Marvell Armada SoCs" depends on DRM && HAVE_CLK && ARM - select FB_CFB_FILLRECT - select FB_CFB_COPYAREA - select FB_CFB_IMAGEBLIT select DRM_KMS_HELPER - select DRM_KMS_FB_HELPER help Support the "LCD" controllers found on the Marvell Armada 510 devices. There are two controllers on the device, each controller diff --git a/drivers/gpu/drm/armada/armada_crtc.c b/drivers/gpu/drm/armada/armada_crtc.c index 34405e4a5d36..2f58e9e2a59c 100644 --- a/drivers/gpu/drm/armada/armada_crtc.c +++ b/drivers/gpu/drm/armada/armada_crtc.c @@ -410,7 +410,7 @@ static void armada_drm_crtc_irq(struct armada_crtc *dcrtc, u32 stat) DRM_ERROR("graphics underflow on crtc %u\n", dcrtc->num); if (stat & VSYNC_IRQ) - drm_handle_vblank(dcrtc->crtc.dev, dcrtc->num); + drm_crtc_handle_vblank(&dcrtc->crtc); spin_lock(&dcrtc->irq_lock); ovl_plane = dcrtc->plane; diff --git a/drivers/gpu/drm/ast/Kconfig b/drivers/gpu/drm/ast/Kconfig index 8a784c460c89..15f6ce7acb2a 100644 --- a/drivers/gpu/drm/ast/Kconfig +++ b/drivers/gpu/drm/ast/Kconfig @@ -2,11 +2,7 @@ config DRM_AST tristate "AST server chips" depends on DRM && PCI select DRM_TTM - select FB_SYS_COPYAREA - select FB_SYS_FILLRECT - select FB_SYS_IMAGEBLIT select DRM_KMS_HELPER - select DRM_KMS_FB_HELPER select DRM_TTM help Say yes for experimental AST GPU driver. Do not enable diff --git a/drivers/gpu/drm/ast/ast_main.c b/drivers/gpu/drm/ast/ast_main.c index 7bc3aa6dda8c..904beaa932d0 100644 --- a/drivers/gpu/drm/ast/ast_main.c +++ b/drivers/gpu/drm/ast/ast_main.c @@ -295,9 +295,8 @@ static int ast_get_dram_info(struct drm_device *dev) static void ast_user_framebuffer_destroy(struct drm_framebuffer *fb) { struct ast_framebuffer *ast_fb = to_ast_framebuffer(fb); - if (ast_fb->obj) - drm_gem_object_unreference_unlocked(ast_fb->obj); + drm_gem_object_unreference_unlocked(ast_fb->obj); drm_framebuffer_cleanup(fb); kfree(fb); } diff --git a/drivers/gpu/drm/ast/ast_ttm.c b/drivers/gpu/drm/ast/ast_ttm.c index 59f2f93b6f84..b29a41218fc9 100644 --- a/drivers/gpu/drm/ast/ast_ttm.c +++ b/drivers/gpu/drm/ast/ast_ttm.c @@ -186,17 +186,6 @@ static void ast_ttm_io_mem_free(struct ttm_bo_device *bdev, struct ttm_mem_reg * { } -static int ast_bo_move(struct ttm_buffer_object *bo, - bool evict, bool interruptible, - bool no_wait_gpu, - struct ttm_mem_reg *new_mem) -{ - int r; - r = ttm_bo_move_memcpy(bo, evict, no_wait_gpu, new_mem); - return r; -} - - static void ast_ttm_backend_destroy(struct ttm_tt *tt) { ttm_tt_fini(tt); @@ -241,7 +230,7 @@ struct ttm_bo_driver ast_bo_driver = { .ttm_tt_unpopulate = ast_ttm_tt_unpopulate, .init_mem_type = ast_bo_init_mem_type, .evict_flags = ast_bo_evict_flags, - .move = ast_bo_move, + .move = NULL, .verify_access = ast_bo_verify_access, .io_mem_reserve = &ast_ttm_io_mem_reserve, .io_mem_free = &ast_ttm_io_mem_free, diff --git a/drivers/gpu/drm/atmel-hlcdc/Kconfig b/drivers/gpu/drm/atmel-hlcdc/Kconfig index 99b4f0698a30..32bcc4bad06a 100644 --- a/drivers/gpu/drm/atmel-hlcdc/Kconfig +++ b/drivers/gpu/drm/atmel-hlcdc/Kconfig @@ -3,7 +3,6 @@ config DRM_ATMEL_HLCDC depends on DRM && OF && COMMON_CLK && MFD_ATMEL_HLCDC && ARM select DRM_GEM_CMA_HELPER select DRM_KMS_HELPER - select DRM_KMS_FB_HELPER select DRM_KMS_CMA_HELPER select DRM_PANEL help diff --git a/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_crtc.c b/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_crtc.c index 613f6c99b76a..a978381ef95b 100644 --- a/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_crtc.c +++ b/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_crtc.c @@ -383,7 +383,7 @@ static void atmel_hlcdc_crtc_finish_page_flip(struct atmel_hlcdc_crtc *crtc) void atmel_hlcdc_crtc_irq(struct drm_crtc *c) { - drm_handle_vblank(c->dev, 0); + drm_crtc_handle_vblank(c); atmel_hlcdc_crtc_finish_page_flip(drm_crtc_to_atmel_hlcdc_crtc(c)); } diff --git a/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_output.c b/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_output.c index 473a475f27b1..6119b5085501 100644 --- a/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_output.c +++ b/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_output.c @@ -254,9 +254,10 @@ int atmel_hlcdc_create_outputs(struct drm_device *dev) if (!ret) ret = atmel_hlcdc_check_endpoint(dev, &ep); - of_node_put(ep_np); - if (ret) + if (ret) { + of_node_put(ep_np); return ret; + } } for_each_endpoint_of_node(dev->dev->of_node, ep_np) { @@ -264,9 +265,10 @@ int atmel_hlcdc_create_outputs(struct drm_device *dev) if (!ret) ret = atmel_hlcdc_attach_endpoint(dev, &ep); - of_node_put(ep_np); - if (ret) + if (ret) { + of_node_put(ep_np); return ret; + } } return 0; diff --git a/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_plane.c b/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_plane.c index aef3ca8a81fa..016c191221f3 100644 --- a/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_plane.c +++ b/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_plane.c @@ -339,6 +339,8 @@ atmel_hlcdc_plane_update_pos_and_size(struct atmel_hlcdc_plane *plane, atmel_hlcdc_layer_update_cfg(&plane->layer, 13, 0xffffffff, factor_reg); + } else { + atmel_hlcdc_layer_update_cfg(&plane->layer, 13, 0xffffffff, 0); } } diff --git a/drivers/gpu/drm/bochs/Kconfig b/drivers/gpu/drm/bochs/Kconfig index 5f8b0c2b9a44..f739763f47ce 100644 --- a/drivers/gpu/drm/bochs/Kconfig +++ b/drivers/gpu/drm/bochs/Kconfig @@ -2,10 +2,6 @@ config DRM_BOCHS tristate "DRM Support for bochs dispi vga interface (qemu stdvga)" depends on DRM && PCI select DRM_KMS_HELPER - select DRM_KMS_FB_HELPER - select FB_SYS_FILLRECT - select FB_SYS_COPYAREA - select FB_SYS_IMAGEBLIT select DRM_TTM help Choose this option for qemu. diff --git a/drivers/gpu/drm/bochs/bochs_mm.c b/drivers/gpu/drm/bochs/bochs_mm.c index 6cf912c45e48..5c5638a777a1 100644 --- a/drivers/gpu/drm/bochs/bochs_mm.c +++ b/drivers/gpu/drm/bochs/bochs_mm.c @@ -165,15 +165,6 @@ static void bochs_ttm_io_mem_free(struct ttm_bo_device *bdev, { } -static int bochs_bo_move(struct ttm_buffer_object *bo, - bool evict, bool interruptible, - bool no_wait_gpu, - struct ttm_mem_reg *new_mem) -{ - return ttm_bo_move_memcpy(bo, evict, no_wait_gpu, new_mem); -} - - static void bochs_ttm_backend_destroy(struct ttm_tt *tt) { ttm_tt_fini(tt); @@ -208,7 +199,7 @@ struct ttm_bo_driver bochs_bo_driver = { .ttm_tt_unpopulate = ttm_pool_unpopulate, .init_mem_type = bochs_bo_init_mem_type, .evict_flags = bochs_bo_evict_flags, - .move = bochs_bo_move, + .move = NULL, .verify_access = bochs_bo_verify_access, .io_mem_reserve = &bochs_ttm_io_mem_reserve, .io_mem_free = &bochs_ttm_io_mem_free, @@ -474,8 +465,8 @@ int bochs_dumb_mmap_offset(struct drm_file *file, struct drm_device *dev, static void bochs_user_framebuffer_destroy(struct drm_framebuffer *fb) { struct bochs_framebuffer *bochs_fb = to_bochs_framebuffer(fb); - if (bochs_fb->obj) - drm_gem_object_unreference_unlocked(bochs_fb->obj); + + drm_gem_object_unreference_unlocked(bochs_fb->obj); drm_framebuffer_cleanup(fb); kfree(fb); } diff --git a/drivers/gpu/drm/bridge/Kconfig b/drivers/gpu/drm/bridge/Kconfig index a141921445f4..b590e678052d 100644 --- a/drivers/gpu/drm/bridge/Kconfig +++ b/drivers/gpu/drm/bridge/Kconfig @@ -58,6 +58,17 @@ config DRM_SII902X ---help--- Silicon Image sii902x bridge chip driver. +config DRM_TOSHIBA_TC358767 + tristate "Toshiba TC358767 eDP bridge" + depends on OF + select DRM_KMS_HELPER + select REGMAP_I2C + select DRM_PANEL + ---help--- + Toshiba TC358767 eDP bridge chip driver. + source "drivers/gpu/drm/bridge/analogix/Kconfig" +source "drivers/gpu/drm/bridge/adv7511/Kconfig" + endmenu diff --git a/drivers/gpu/drm/bridge/Makefile b/drivers/gpu/drm/bridge/Makefile index bfec9f8cb9d2..efdb07e878f5 100644 --- a/drivers/gpu/drm/bridge/Makefile +++ b/drivers/gpu/drm/bridge/Makefile @@ -6,4 +6,6 @@ obj-$(CONFIG_DRM_DW_HDMI_AHB_AUDIO) += dw-hdmi-ahb-audio.o obj-$(CONFIG_DRM_NXP_PTN3460) += nxp-ptn3460.o obj-$(CONFIG_DRM_PARADE_PS8622) += parade-ps8622.o obj-$(CONFIG_DRM_SII902X) += sii902x.o +obj-$(CONFIG_DRM_TOSHIBA_TC358767) += tc358767.o obj-$(CONFIG_DRM_ANALOGIX_DP) += analogix/ +obj-$(CONFIG_DRM_I2C_ADV7511) += adv7511/ diff --git a/drivers/gpu/drm/bridge/adv7511/Kconfig b/drivers/gpu/drm/bridge/adv7511/Kconfig new file mode 100644 index 000000000000..d2b0499ab7d7 --- /dev/null +++ b/drivers/gpu/drm/bridge/adv7511/Kconfig @@ -0,0 +1,15 @@ +config DRM_I2C_ADV7511 + tristate "AV7511 encoder" + depends on OF + select DRM_KMS_HELPER + select REGMAP_I2C + help + Support for the Analog Device ADV7511(W) and ADV7513 HDMI encoders. + +config DRM_I2C_ADV7533 + bool "ADV7533 encoder" + depends on DRM_I2C_ADV7511 + select DRM_MIPI_DSI + default y + help + Support for the Analog Devices ADV7533 DSI to HDMI encoder. diff --git a/drivers/gpu/drm/bridge/adv7511/Makefile b/drivers/gpu/drm/bridge/adv7511/Makefile new file mode 100644 index 000000000000..9019327fff4c --- /dev/null +++ b/drivers/gpu/drm/bridge/adv7511/Makefile @@ -0,0 +1,3 @@ +adv7511-y := adv7511_drv.o +adv7511-$(CONFIG_DRM_I2C_ADV7533) += adv7533.o +obj-$(CONFIG_DRM_I2C_ADV7511) += adv7511.o diff --git a/drivers/gpu/drm/i2c/adv7511.h b/drivers/gpu/drm/bridge/adv7511/adv7511.h index 38515b30cedf..161c923d6162 100644 --- a/drivers/gpu/drm/i2c/adv7511.h +++ b/drivers/gpu/drm/bridge/adv7511/adv7511.h @@ -10,6 +10,11 @@ #define __DRM_I2C_ADV7511_H__ #include <linux/hdmi.h> +#include <linux/i2c.h> +#include <linux/regmap.h> + +#include <drm/drm_crtc_helper.h> +#include <drm/drm_mipi_dsi.h> #define ADV7511_REG_CHIP_REVISION 0x00 #define ADV7511_REG_N0 0x01 @@ -286,4 +291,102 @@ struct adv7511_video_config { struct hdmi_avi_infoframe avi_infoframe; }; +enum adv7511_type { + ADV7511, + ADV7533, +}; + +struct adv7511 { + struct i2c_client *i2c_main; + struct i2c_client *i2c_edid; + struct i2c_client *i2c_cec; + + struct regmap *regmap; + struct regmap *regmap_cec; + enum drm_connector_status status; + bool powered; + + struct drm_display_mode curr_mode; + + unsigned int f_tmds; + + unsigned int current_edid_segment; + uint8_t edid_buf[256]; + bool edid_read; + + wait_queue_head_t wq; + struct drm_bridge bridge; + struct drm_connector connector; + + bool embedded_sync; + enum adv7511_sync_polarity vsync_polarity; + enum adv7511_sync_polarity hsync_polarity; + bool rgb; + + struct edid *edid; + + struct gpio_desc *gpio_pd; + + /* ADV7533 DSI RX related params */ + struct device_node *host_node; + struct mipi_dsi_device *dsi; + u8 num_dsi_lanes; + bool use_timing_gen; + + enum adv7511_type type; +}; + +#ifdef CONFIG_DRM_I2C_ADV7533 +void adv7533_dsi_power_on(struct adv7511 *adv); +void adv7533_dsi_power_off(struct adv7511 *adv); +void adv7533_mode_set(struct adv7511 *adv, struct drm_display_mode *mode); +int adv7533_patch_registers(struct adv7511 *adv); +void adv7533_uninit_cec(struct adv7511 *adv); +int adv7533_init_cec(struct adv7511 *adv); +int adv7533_attach_dsi(struct adv7511 *adv); +void adv7533_detach_dsi(struct adv7511 *adv); +int adv7533_parse_dt(struct device_node *np, struct adv7511 *adv); +#else +static inline void adv7533_dsi_power_on(struct adv7511 *adv) +{ +} + +static inline void adv7533_dsi_power_off(struct adv7511 *adv) +{ +} + +static inline void adv7533_mode_set(struct adv7511 *adv, + struct drm_display_mode *mode) +{ +} + +static inline int adv7533_patch_registers(struct adv7511 *adv) +{ + return -ENODEV; +} + +static inline void adv7533_uninit_cec(struct adv7511 *adv) +{ +} + +static inline int adv7533_init_cec(struct adv7511 *adv) +{ + return -ENODEV; +} + +static inline int adv7533_attach_dsi(struct adv7511 *adv) +{ + return -ENODEV; +} + +static inline void adv7533_detach_dsi(struct adv7511 *adv) +{ +} + +static inline int adv7533_parse_dt(struct device_node *np, struct adv7511 *adv) +{ + return -ENODEV; +} +#endif + #endif /* __DRM_I2C_ADV7511_H__ */ diff --git a/drivers/gpu/drm/i2c/adv7511.c b/drivers/gpu/drm/bridge/adv7511/adv7511_drv.c index a02112ba1c3d..ec8fb2ed3275 100644 --- a/drivers/gpu/drm/i2c/adv7511.c +++ b/drivers/gpu/drm/bridge/adv7511/adv7511_drv.c @@ -8,51 +8,17 @@ #include <linux/device.h> #include <linux/gpio/consumer.h> -#include <linux/i2c.h> #include <linux/module.h> -#include <linux/regmap.h> +#include <linux/of_device.h> #include <linux/slab.h> #include <drm/drmP.h> -#include <drm/drm_crtc_helper.h> +#include <drm/drm_atomic.h> +#include <drm/drm_atomic_helper.h> #include <drm/drm_edid.h> -#include <drm/drm_encoder_slave.h> #include "adv7511.h" -struct adv7511 { - struct i2c_client *i2c_main; - struct i2c_client *i2c_edid; - - struct regmap *regmap; - struct regmap *packet_memory_regmap; - enum drm_connector_status status; - bool powered; - - unsigned int f_tmds; - - unsigned int current_edid_segment; - uint8_t edid_buf[256]; - bool edid_read; - - wait_queue_head_t wq; - struct drm_encoder *encoder; - - bool embedded_sync; - enum adv7511_sync_polarity vsync_polarity; - enum adv7511_sync_polarity hsync_polarity; - bool rgb; - - struct edid *edid; - - struct gpio_desc *gpio_pd; -}; - -static struct adv7511 *encoder_to_adv7511(struct drm_encoder *encoder) -{ - return to_encoder_slave(encoder)->slave_priv; -} - /* ADI recommended values for proper operation. */ static const struct reg_sequence adv7511_fixed_registers[] = { { 0x98, 0x03 }, @@ -394,6 +360,9 @@ static void adv7511_power_on(struct adv7511 *adv7511) */ regcache_sync(adv7511->regmap); + if (adv7511->type == ADV7533) + adv7533_dsi_power_on(adv7511); + adv7511->powered = true; } @@ -405,6 +374,9 @@ static void adv7511_power_off(struct adv7511 *adv7511) ADV7511_POWER_POWER_DOWN); regcache_mark_dirty(adv7511->regmap); + if (adv7511->type == ADV7533) + adv7533_dsi_power_off(adv7511); + adv7511->powered = false; } @@ -430,7 +402,7 @@ static bool adv7511_hpd(struct adv7511 *adv7511) return false; } -static int adv7511_irq_process(struct adv7511 *adv7511) +static int adv7511_irq_process(struct adv7511 *adv7511, bool process_hpd) { unsigned int irq0, irq1; int ret; @@ -446,8 +418,8 @@ static int adv7511_irq_process(struct adv7511 *adv7511) regmap_write(adv7511->regmap, ADV7511_REG_INT(0), irq0); regmap_write(adv7511->regmap, ADV7511_REG_INT(1), irq1); - if (irq0 & ADV7511_INT0_HPD && adv7511->encoder) - drm_helper_hpd_irq_event(adv7511->encoder->dev); + if (process_hpd && irq0 & ADV7511_INT0_HPD && adv7511->bridge.encoder) + drm_helper_hpd_irq_event(adv7511->connector.dev); if (irq0 & ADV7511_INT0_EDID_READY || irq1 & ADV7511_INT1_DDC_ERROR) { adv7511->edid_read = true; @@ -464,7 +436,7 @@ static irqreturn_t adv7511_irq_handler(int irq, void *devid) struct adv7511 *adv7511 = devid; int ret; - ret = adv7511_irq_process(adv7511); + ret = adv7511_irq_process(adv7511, true); return ret < 0 ? IRQ_NONE : IRQ_HANDLED; } @@ -481,7 +453,7 @@ static int adv7511_wait_for_edid(struct adv7511 *adv7511, int timeout) adv7511->edid_read, msecs_to_jiffies(timeout)); } else { for (; timeout > 0; timeout -= 25) { - ret = adv7511_irq_process(adv7511); + ret = adv7511_irq_process(adv7511, false); if (ret < 0) break; @@ -563,13 +535,12 @@ static int adv7511_get_edid_block(void *data, u8 *buf, unsigned int block, } /* ----------------------------------------------------------------------------- - * Encoder operations + * ADV75xx helpers */ -static int adv7511_get_modes(struct drm_encoder *encoder, +static int adv7511_get_modes(struct adv7511 *adv7511, struct drm_connector *connector) { - struct adv7511 *adv7511 = encoder_to_adv7511(encoder); struct edid *edid; unsigned int count; @@ -606,21 +577,9 @@ static int adv7511_get_modes(struct drm_encoder *encoder, return count; } -static void adv7511_encoder_dpms(struct drm_encoder *encoder, int mode) -{ - struct adv7511 *adv7511 = encoder_to_adv7511(encoder); - - if (mode == DRM_MODE_DPMS_ON) - adv7511_power_on(adv7511); - else - adv7511_power_off(adv7511); -} - static enum drm_connector_status -adv7511_encoder_detect(struct drm_encoder *encoder, - struct drm_connector *connector) +adv7511_detect(struct adv7511 *adv7511, struct drm_connector *connector) { - struct adv7511 *adv7511 = encoder_to_adv7511(encoder); enum drm_connector_status status; unsigned int val; bool hpd; @@ -644,7 +603,7 @@ adv7511_encoder_detect(struct drm_encoder *encoder, if (status == connector_status_connected && hpd && adv7511->powered) { regcache_mark_dirty(adv7511->regmap); adv7511_power_on(adv7511); - adv7511_get_modes(encoder, connector); + adv7511_get_modes(adv7511, connector); if (adv7511->status == connector_status_connected) status = connector_status_disconnected; } else { @@ -658,8 +617,8 @@ adv7511_encoder_detect(struct drm_encoder *encoder, return status; } -static int adv7511_encoder_mode_valid(struct drm_encoder *encoder, - struct drm_display_mode *mode) +static int adv7511_mode_valid(struct adv7511 *adv7511, + struct drm_display_mode *mode) { if (mode->clock > 165000) return MODE_CLOCK_HIGH; @@ -667,11 +626,10 @@ static int adv7511_encoder_mode_valid(struct drm_encoder *encoder, return MODE_OK; } -static void adv7511_encoder_mode_set(struct drm_encoder *encoder, - struct drm_display_mode *mode, - struct drm_display_mode *adj_mode) +static void adv7511_mode_set(struct adv7511 *adv7511, + struct drm_display_mode *mode, + struct drm_display_mode *adj_mode) { - struct adv7511 *adv7511 = encoder_to_adv7511(encoder); unsigned int low_refresh_rate; unsigned int hsync_polarity = 0; unsigned int vsync_polarity = 0; @@ -754,6 +712,11 @@ static void adv7511_encoder_mode_set(struct drm_encoder *encoder, regmap_update_bits(adv7511->regmap, 0x17, 0x60, (vsync_polarity << 6) | (hsync_polarity << 5)); + if (adv7511->type == ADV7533) + adv7533_mode_set(adv7511, adj_mode); + + drm_mode_copy(&adv7511->curr_mode, adj_mode); + /* * TODO Test first order 4:2:2 to 4:4:4 up conversion method, which is * supposed to give better results. @@ -762,12 +725,114 @@ static void adv7511_encoder_mode_set(struct drm_encoder *encoder, adv7511->f_tmds = mode->clock; } -static const struct drm_encoder_slave_funcs adv7511_encoder_funcs = { - .dpms = adv7511_encoder_dpms, - .mode_valid = adv7511_encoder_mode_valid, - .mode_set = adv7511_encoder_mode_set, - .detect = adv7511_encoder_detect, - .get_modes = adv7511_get_modes, +/* Connector funcs */ +static struct adv7511 *connector_to_adv7511(struct drm_connector *connector) +{ + return container_of(connector, struct adv7511, connector); +} + +static int adv7511_connector_get_modes(struct drm_connector *connector) +{ + struct adv7511 *adv = connector_to_adv7511(connector); + + return adv7511_get_modes(adv, connector); +} + +static enum drm_mode_status +adv7511_connector_mode_valid(struct drm_connector *connector, + struct drm_display_mode *mode) +{ + struct adv7511 *adv = connector_to_adv7511(connector); + + return adv7511_mode_valid(adv, mode); +} + +static struct drm_connector_helper_funcs adv7511_connector_helper_funcs = { + .get_modes = adv7511_connector_get_modes, + .mode_valid = adv7511_connector_mode_valid, +}; + +static enum drm_connector_status +adv7511_connector_detect(struct drm_connector *connector, bool force) +{ + struct adv7511 *adv = connector_to_adv7511(connector); + + return adv7511_detect(adv, connector); +} + +static struct drm_connector_funcs adv7511_connector_funcs = { + .dpms = drm_atomic_helper_connector_dpms, + .fill_modes = drm_helper_probe_single_connector_modes, + .detect = adv7511_connector_detect, + .destroy = drm_connector_cleanup, + .reset = drm_atomic_helper_connector_reset, + .atomic_duplicate_state = drm_atomic_helper_connector_duplicate_state, + .atomic_destroy_state = drm_atomic_helper_connector_destroy_state, +}; + +/* Bridge funcs */ +static struct adv7511 *bridge_to_adv7511(struct drm_bridge *bridge) +{ + return container_of(bridge, struct adv7511, bridge); +} + +static void adv7511_bridge_enable(struct drm_bridge *bridge) +{ + struct adv7511 *adv = bridge_to_adv7511(bridge); + + adv7511_power_on(adv); +} + +static void adv7511_bridge_disable(struct drm_bridge *bridge) +{ + struct adv7511 *adv = bridge_to_adv7511(bridge); + + adv7511_power_off(adv); +} + +static void adv7511_bridge_mode_set(struct drm_bridge *bridge, + struct drm_display_mode *mode, + struct drm_display_mode *adj_mode) +{ + struct adv7511 *adv = bridge_to_adv7511(bridge); + + adv7511_mode_set(adv, mode, adj_mode); +} + +static int adv7511_bridge_attach(struct drm_bridge *bridge) +{ + struct adv7511 *adv = bridge_to_adv7511(bridge); + int ret; + + if (!bridge->encoder) { + DRM_ERROR("Parent encoder object not found"); + return -ENODEV; + } + + adv->connector.polled = DRM_CONNECTOR_POLL_HPD; + + ret = drm_connector_init(bridge->dev, &adv->connector, + &adv7511_connector_funcs, + DRM_MODE_CONNECTOR_HDMIA); + if (ret) { + DRM_ERROR("Failed to initialize connector with drm\n"); + return ret; + } + drm_connector_helper_add(&adv->connector, + &adv7511_connector_helper_funcs); + drm_mode_connector_attach_encoder(&adv->connector, bridge->encoder); + + if (adv->type == ADV7533) + ret = adv7533_attach_dsi(adv); + + return ret; +} + +static struct drm_bridge_funcs adv7511_bridge_funcs = { + .enable = adv7511_bridge_enable, + .disable = adv7511_bridge_disable, + .mode_set = adv7511_bridge_mode_set, + .attach = adv7511_bridge_attach, }; /* ----------------------------------------------------------------------------- @@ -780,8 +845,6 @@ static int adv7511_parse_dt(struct device_node *np, const char *str; int ret; - memset(config, 0, sizeof(*config)); - of_property_read_u32(np, "adi,input-depth", &config->input_color_depth); if (config->input_color_depth != 8 && config->input_color_depth != 10 && config->input_color_depth != 12) @@ -881,7 +944,17 @@ static int adv7511_probe(struct i2c_client *i2c, const struct i2c_device_id *id) adv7511->powered = false; adv7511->status = connector_status_disconnected; - ret = adv7511_parse_dt(dev->of_node, &link_config); + if (dev->of_node) + adv7511->type = (enum adv7511_type)of_device_get_match_data(dev); + else + adv7511->type = id->driver_data; + + memset(&link_config, 0, sizeof(link_config)); + + if (adv7511->type == ADV7511) + ret = adv7511_parse_dt(dev->of_node, &link_config); + else + ret = adv7533_parse_dt(dev->of_node, adv7511); if (ret) return ret; @@ -907,8 +980,12 @@ static int adv7511_probe(struct i2c_client *i2c, const struct i2c_device_id *id) return ret; dev_dbg(dev, "Rev. %d\n", val); - ret = regmap_register_patch(adv7511->regmap, adv7511_fixed_registers, - ARRAY_SIZE(adv7511_fixed_registers)); + if (adv7511->type == ADV7511) + ret = regmap_register_patch(adv7511->regmap, + adv7511_fixed_registers, + ARRAY_SIZE(adv7511_fixed_registers)); + else + ret = adv7533_patch_registers(adv7511); if (ret) return ret; @@ -923,6 +1000,12 @@ static int adv7511_probe(struct i2c_client *i2c, const struct i2c_device_id *id) if (!adv7511->i2c_edid) return -ENOMEM; + if (adv7511->type == ADV7533) { + ret = adv7533_init_cec(adv7511); + if (ret) + goto err_i2c_unregister_edid; + } + if (i2c->irq) { init_waitqueue_head(&adv7511->wq); @@ -931,7 +1014,7 @@ static int adv7511_probe(struct i2c_client *i2c, const struct i2c_device_id *id) IRQF_ONESHOT, dev_name(dev), adv7511); if (ret) - goto err_i2c_unregister_device; + goto err_unregister_cec; } /* CEC is unused for now */ @@ -942,11 +1025,23 @@ static int adv7511_probe(struct i2c_client *i2c, const struct i2c_device_id *id) i2c_set_clientdata(i2c, adv7511); - adv7511_set_link_config(adv7511, &link_config); + if (adv7511->type == ADV7511) + adv7511_set_link_config(adv7511, &link_config); + + adv7511->bridge.funcs = &adv7511_bridge_funcs; + adv7511->bridge.of_node = dev->of_node; + + ret = drm_bridge_add(&adv7511->bridge); + if (ret) { + dev_err(dev, "failed to add adv7511 bridge\n"); + goto err_unregister_cec; + } return 0; -err_i2c_unregister_device: +err_unregister_cec: + adv7533_uninit_cec(adv7511); +err_i2c_unregister_edid: i2c_unregister_device(adv7511->i2c_edid); return ret; @@ -956,66 +1051,71 @@ static int adv7511_remove(struct i2c_client *i2c) { struct adv7511 *adv7511 = i2c_get_clientdata(i2c); - i2c_unregister_device(adv7511->i2c_edid); - - kfree(adv7511->edid); - - return 0; -} - -static int adv7511_encoder_init(struct i2c_client *i2c, struct drm_device *dev, - struct drm_encoder_slave *encoder) -{ + if (adv7511->type == ADV7533) { + adv7533_detach_dsi(adv7511); + adv7533_uninit_cec(adv7511); + } - struct adv7511 *adv7511 = i2c_get_clientdata(i2c); + drm_bridge_remove(&adv7511->bridge); - encoder->slave_priv = adv7511; - encoder->slave_funcs = &adv7511_encoder_funcs; + i2c_unregister_device(adv7511->i2c_edid); - adv7511->encoder = &encoder->base; + kfree(adv7511->edid); return 0; } static const struct i2c_device_id adv7511_i2c_ids[] = { - { "adv7511", 0 }, - { "adv7511w", 0 }, - { "adv7513", 0 }, + { "adv7511", ADV7511 }, + { "adv7511w", ADV7511 }, + { "adv7513", ADV7511 }, +#ifdef CONFIG_DRM_I2C_ADV7533 + { "adv7533", ADV7533 }, +#endif { } }; MODULE_DEVICE_TABLE(i2c, adv7511_i2c_ids); static const struct of_device_id adv7511_of_ids[] = { - { .compatible = "adi,adv7511", }, - { .compatible = "adi,adv7511w", }, - { .compatible = "adi,adv7513", }, + { .compatible = "adi,adv7511", .data = (void *)ADV7511 }, + { .compatible = "adi,adv7511w", .data = (void *)ADV7511 }, + { .compatible = "adi,adv7513", .data = (void *)ADV7511 }, +#ifdef CONFIG_DRM_I2C_ADV7533 + { .compatible = "adi,adv7533", .data = (void *)ADV7533 }, +#endif { } }; MODULE_DEVICE_TABLE(of, adv7511_of_ids); -static struct drm_i2c_encoder_driver adv7511_driver = { - .i2c_driver = { - .driver = { - .name = "adv7511", - .of_match_table = adv7511_of_ids, - }, - .id_table = adv7511_i2c_ids, - .probe = adv7511_probe, - .remove = adv7511_remove, - }, +static struct mipi_dsi_driver adv7533_dsi_driver = { + .driver.name = "adv7533", +}; - .encoder_init = adv7511_encoder_init, +static struct i2c_driver adv7511_driver = { + .driver = { + .name = "adv7511", + .of_match_table = adv7511_of_ids, + }, + .id_table = adv7511_i2c_ids, + .probe = adv7511_probe, + .remove = adv7511_remove, }; static int __init adv7511_init(void) { - return drm_i2c_encoder_register(THIS_MODULE, &adv7511_driver); + if (IS_ENABLED(CONFIG_DRM_MIPI_DSI)) + mipi_dsi_driver_register(&adv7533_dsi_driver); + + return i2c_add_driver(&adv7511_driver); } module_init(adv7511_init); static void __exit adv7511_exit(void) { - drm_i2c_encoder_unregister(&adv7511_driver); + i2c_del_driver(&adv7511_driver); + + if (IS_ENABLED(CONFIG_DRM_MIPI_DSI)) + mipi_dsi_driver_unregister(&adv7533_dsi_driver); } module_exit(adv7511_exit); diff --git a/drivers/gpu/drm/bridge/adv7511/adv7533.c b/drivers/gpu/drm/bridge/adv7511/adv7533.c new file mode 100644 index 000000000000..5eebd15899b1 --- /dev/null +++ b/drivers/gpu/drm/bridge/adv7511/adv7533.c @@ -0,0 +1,265 @@ +/* + * Copyright (c) 2016, The Linux Foundation. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 and + * only version 2 as published by the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + */ + +#include <linux/of_graph.h> + +#include "adv7511.h" + +static const struct reg_sequence adv7533_fixed_registers[] = { + { 0x16, 0x20 }, + { 0x9a, 0xe0 }, + { 0xba, 0x70 }, + { 0xde, 0x82 }, + { 0xe4, 0x40 }, + { 0xe5, 0x80 }, +}; + +static const struct reg_sequence adv7533_cec_fixed_registers[] = { + { 0x15, 0xd0 }, + { 0x17, 0xd0 }, + { 0x24, 0x20 }, + { 0x57, 0x11 }, +}; + +static const struct regmap_config adv7533_cec_regmap_config = { + .reg_bits = 8, + .val_bits = 8, + + .max_register = 0xff, + .cache_type = REGCACHE_RBTREE, +}; + +static void adv7511_dsi_config_timing_gen(struct adv7511 *adv) +{ + struct mipi_dsi_device *dsi = adv->dsi; + struct drm_display_mode *mode = &adv->curr_mode; + unsigned int hsw, hfp, hbp, vsw, vfp, vbp; + u8 clock_div_by_lanes[] = { 6, 4, 3 }; /* 2, 3, 4 lanes */ + + hsw = mode->hsync_end - mode->hsync_start; + hfp = mode->hsync_start - mode->hdisplay; + hbp = mode->htotal - mode->hsync_end; + vsw = mode->vsync_end - mode->vsync_start; + vfp = mode->vsync_start - mode->vdisplay; + vbp = mode->vtotal - mode->vsync_end; + + /* set pixel clock divider mode */ + regmap_write(adv->regmap_cec, 0x16, + clock_div_by_lanes[dsi->lanes - 2] << 3); + + /* horizontal porch params */ + regmap_write(adv->regmap_cec, 0x28, mode->htotal >> 4); + regmap_write(adv->regmap_cec, 0x29, (mode->htotal << 4) & 0xff); + regmap_write(adv->regmap_cec, 0x2a, hsw >> 4); + regmap_write(adv->regmap_cec, 0x2b, (hsw << 4) & 0xff); + regmap_write(adv->regmap_cec, 0x2c, hfp >> 4); + regmap_write(adv->regmap_cec, 0x2d, (hfp << 4) & 0xff); + regmap_write(adv->regmap_cec, 0x2e, hbp >> 4); + regmap_write(adv->regmap_cec, 0x2f, (hbp << 4) & 0xff); + + /* vertical porch params */ + regmap_write(adv->regmap_cec, 0x30, mode->vtotal >> 4); + regmap_write(adv->regmap_cec, 0x31, (mode->vtotal << 4) & 0xff); + regmap_write(adv->regmap_cec, 0x32, vsw >> 4); + regmap_write(adv->regmap_cec, 0x33, (vsw << 4) & 0xff); + regmap_write(adv->regmap_cec, 0x34, vfp >> 4); + regmap_write(adv->regmap_cec, 0x35, (vfp << 4) & 0xff); + regmap_write(adv->regmap_cec, 0x36, vbp >> 4); + regmap_write(adv->regmap_cec, 0x37, (vbp << 4) & 0xff); +} + +void adv7533_dsi_power_on(struct adv7511 *adv) +{ + struct mipi_dsi_device *dsi = adv->dsi; + + if (adv->use_timing_gen) + adv7511_dsi_config_timing_gen(adv); + + /* set number of dsi lanes */ + regmap_write(adv->regmap_cec, 0x1c, dsi->lanes << 4); + + if (adv->use_timing_gen) { + /* reset internal timing generator */ + regmap_write(adv->regmap_cec, 0x27, 0xcb); + regmap_write(adv->regmap_cec, 0x27, 0x8b); + regmap_write(adv->regmap_cec, 0x27, 0xcb); + } else { + /* disable internal timing generator */ + regmap_write(adv->regmap_cec, 0x27, 0x0b); + } + + /* enable hdmi */ + regmap_write(adv->regmap_cec, 0x03, 0x89); + /* disable test mode */ + regmap_write(adv->regmap_cec, 0x55, 0x00); + + regmap_register_patch(adv->regmap_cec, adv7533_cec_fixed_registers, + ARRAY_SIZE(adv7533_cec_fixed_registers)); +} + +void adv7533_dsi_power_off(struct adv7511 *adv) +{ + /* disable hdmi */ + regmap_write(adv->regmap_cec, 0x03, 0x0b); + /* disable internal timing generator */ + regmap_write(adv->regmap_cec, 0x27, 0x0b); +} + +void adv7533_mode_set(struct adv7511 *adv, struct drm_display_mode *mode) +{ + struct mipi_dsi_device *dsi = adv->dsi; + int lanes, ret; + + if (adv->num_dsi_lanes != 4) + return; + + if (mode->clock > 80000) + lanes = 4; + else + lanes = 3; + + if (lanes != dsi->lanes) { + mipi_dsi_detach(dsi); + dsi->lanes = lanes; + ret = mipi_dsi_attach(dsi); + if (ret) + dev_err(&dsi->dev, "failed to change host lanes\n"); + } +} + +int adv7533_patch_registers(struct adv7511 *adv) +{ + return regmap_register_patch(adv->regmap, + adv7533_fixed_registers, + ARRAY_SIZE(adv7533_fixed_registers)); +} + +void adv7533_uninit_cec(struct adv7511 *adv) +{ + i2c_unregister_device(adv->i2c_cec); +} + +static const int cec_i2c_addr = 0x78; + +int adv7533_init_cec(struct adv7511 *adv) +{ + int ret; + + adv->i2c_cec = i2c_new_dummy(adv->i2c_main->adapter, cec_i2c_addr >> 1); + if (!adv->i2c_cec) + return -ENOMEM; + + adv->regmap_cec = devm_regmap_init_i2c(adv->i2c_cec, + &adv7533_cec_regmap_config); + if (IS_ERR(adv->regmap_cec)) { + ret = PTR_ERR(adv->regmap_cec); + goto err; + } + + ret = regmap_register_patch(adv->regmap_cec, + adv7533_cec_fixed_registers, + ARRAY_SIZE(adv7533_cec_fixed_registers)); + if (ret) + goto err; + + return 0; +err: + adv7533_uninit_cec(adv); + return ret; +} + +int adv7533_attach_dsi(struct adv7511 *adv) +{ + struct device *dev = &adv->i2c_main->dev; + struct mipi_dsi_host *host; + struct mipi_dsi_device *dsi; + int ret = 0; + const struct mipi_dsi_device_info info = { .type = "adv7533", + .channel = 0, + .node = NULL, + }; + + host = of_find_mipi_dsi_host_by_node(adv->host_node); + if (!host) { + dev_err(dev, "failed to find dsi host\n"); + return -EPROBE_DEFER; + } + + dsi = mipi_dsi_device_register_full(host, &info); + if (IS_ERR(dsi)) { + dev_err(dev, "failed to create dsi device\n"); + ret = PTR_ERR(dsi); + goto err_dsi_device; + } + + adv->dsi = dsi; + + dsi->lanes = adv->num_dsi_lanes; + dsi->format = MIPI_DSI_FMT_RGB888; + dsi->mode_flags = MIPI_DSI_MODE_VIDEO | MIPI_DSI_MODE_VIDEO_SYNC_PULSE | + MIPI_DSI_MODE_EOT_PACKET | MIPI_DSI_MODE_VIDEO_HSE; + + ret = mipi_dsi_attach(dsi); + if (ret < 0) { + dev_err(dev, "failed to attach dsi to host\n"); + goto err_dsi_attach; + } + + return 0; + +err_dsi_attach: + mipi_dsi_device_unregister(dsi); +err_dsi_device: + return ret; +} + +void adv7533_detach_dsi(struct adv7511 *adv) +{ + mipi_dsi_detach(adv->dsi); + mipi_dsi_device_unregister(adv->dsi); +} + +int adv7533_parse_dt(struct device_node *np, struct adv7511 *adv) +{ + u32 num_lanes; + struct device_node *endpoint; + + of_property_read_u32(np, "adi,dsi-lanes", &num_lanes); + + if (num_lanes < 1 || num_lanes > 4) + return -EINVAL; + + adv->num_dsi_lanes = num_lanes; + + endpoint = of_graph_get_next_endpoint(np, NULL); + if (!endpoint) + return -ENODEV; + + adv->host_node = of_graph_get_remote_port_parent(endpoint); + if (!adv->host_node) { + of_node_put(endpoint); + return -ENODEV; + } + + of_node_put(endpoint); + of_node_put(adv->host_node); + + adv->use_timing_gen = !of_property_read_bool(np, + "adi,disable-timing-generator"); + + /* TODO: Check if these need to be parsed by DT or not */ + adv->rgb = true; + adv->embedded_sync = false; + + return 0; +} diff --git a/drivers/gpu/drm/bridge/analogix/analogix_dp_core.c b/drivers/gpu/drm/bridge/analogix/analogix_dp_core.c index 7699597070a1..32715daf73cb 100644 --- a/drivers/gpu/drm/bridge/analogix/analogix_dp_core.c +++ b/drivers/gpu/drm/bridge/analogix/analogix_dp_core.c @@ -938,7 +938,7 @@ int analogix_dp_get_modes(struct drm_connector *connector) num_modes += drm_panel_get_modes(dp->plat_data->panel); if (dp->plat_data->get_modes) - num_modes += dp->plat_data->get_modes(dp->plat_data); + num_modes += dp->plat_data->get_modes(dp->plat_data, connector); return num_modes; } @@ -1208,6 +1208,7 @@ static int analogix_dp_dt_parse_pdata(struct analogix_dp_device *dp) switch (dp->plat_data->dev_type) { case RK3288_DP: + case RK3399_EDP: /* * Like Rk3288 DisplayPort TRM indicate that "Main link * containing 4 physical lanes of 2.7/1.62 Gbps/lane". diff --git a/drivers/gpu/drm/bridge/analogix/analogix_dp_core.h b/drivers/gpu/drm/bridge/analogix/analogix_dp_core.h index f09275d40f70..b45638043ec4 100644 --- a/drivers/gpu/drm/bridge/analogix/analogix_dp_core.h +++ b/drivers/gpu/drm/bridge/analogix/analogix_dp_core.h @@ -127,10 +127,10 @@ enum analog_power_block { }; enum dp_irq_type { - DP_IRQ_TYPE_HP_CABLE_IN, - DP_IRQ_TYPE_HP_CABLE_OUT, - DP_IRQ_TYPE_HP_CHANGE, - DP_IRQ_TYPE_UNKNOWN, + DP_IRQ_TYPE_HP_CABLE_IN = BIT(0), + DP_IRQ_TYPE_HP_CABLE_OUT = BIT(1), + DP_IRQ_TYPE_HP_CHANGE = BIT(2), + DP_IRQ_TYPE_UNKNOWN = BIT(3), }; struct video_info { diff --git a/drivers/gpu/drm/bridge/analogix/analogix_dp_reg.c b/drivers/gpu/drm/bridge/analogix/analogix_dp_reg.c index 49205ef02be3..48030f0cf497 100644 --- a/drivers/gpu/drm/bridge/analogix/analogix_dp_reg.c +++ b/drivers/gpu/drm/bridge/analogix/analogix_dp_reg.c @@ -74,8 +74,12 @@ void analogix_dp_init_analog_param(struct analogix_dp_device *dp) reg = SEL_24M | TX_DVDD_BIT_1_0625V; writel(reg, dp->reg_base + ANALOGIX_DP_ANALOG_CTL_2); - if (dp->plat_data && (dp->plat_data->dev_type == RK3288_DP)) { - writel(REF_CLK_24M, dp->reg_base + ANALOGIX_DP_PLL_REG_1); + if (dp->plat_data && is_rockchip(dp->plat_data->dev_type)) { + reg = REF_CLK_24M; + if (dp->plat_data->dev_type == RK3288_DP) + reg ^= REF_CLK_MASK; + + writel(reg, dp->reg_base + ANALOGIX_DP_PLL_REG_1); writel(0x95, dp->reg_base + ANALOGIX_DP_PLL_REG_2); writel(0x40, dp->reg_base + ANALOGIX_DP_PLL_REG_3); writel(0x58, dp->reg_base + ANALOGIX_DP_PLL_REG_4); @@ -244,7 +248,7 @@ void analogix_dp_set_analog_power_down(struct analogix_dp_device *dp, u32 reg; u32 phy_pd_addr = ANALOGIX_DP_PHY_PD; - if (dp->plat_data && (dp->plat_data->dev_type == RK3288_DP)) + if (dp->plat_data && is_rockchip(dp->plat_data->dev_type)) phy_pd_addr = ANALOGIX_DP_PD; switch (block) { @@ -448,7 +452,7 @@ void analogix_dp_init_aux(struct analogix_dp_device *dp) analogix_dp_reset_aux(dp); /* Disable AUX transaction H/W retry */ - if (dp->plat_data && (dp->plat_data->dev_type == RK3288_DP)) + if (dp->plat_data && is_rockchip(dp->plat_data->dev_type)) reg = AUX_BIT_PERIOD_EXPECTED_DELAY(0) | AUX_HW_RETRY_COUNT_SEL(3) | AUX_HW_RETRY_INTERVAL_600_MICROSECONDS; diff --git a/drivers/gpu/drm/bridge/analogix/analogix_dp_reg.h b/drivers/gpu/drm/bridge/analogix/analogix_dp_reg.h index 337912b0aeab..cdcc6c5add5e 100644 --- a/drivers/gpu/drm/bridge/analogix/analogix_dp_reg.h +++ b/drivers/gpu/drm/bridge/analogix/analogix_dp_reg.h @@ -163,8 +163,9 @@ #define HSYNC_POLARITY_CFG (0x1 << 0) /* ANALOGIX_DP_PLL_REG_1 */ -#define REF_CLK_24M (0x1 << 1) -#define REF_CLK_27M (0x0 << 1) +#define REF_CLK_24M (0x1 << 0) +#define REF_CLK_27M (0x0 << 0) +#define REF_CLK_MASK (0x1 << 0) /* ANALOGIX_DP_LANE_MAP */ #define LANE3_MAP_LOGIC_LANE_0 (0x0 << 6) diff --git a/drivers/gpu/drm/bridge/dw-hdmi.c b/drivers/gpu/drm/bridge/dw-hdmi.c index 70b1f7d4270b..77ab47341658 100644 --- a/drivers/gpu/drm/bridge/dw-hdmi.c +++ b/drivers/gpu/drm/bridge/dw-hdmi.c @@ -1495,14 +1495,6 @@ static void dw_hdmi_connector_force(struct drm_connector *connector) } static const struct drm_connector_funcs dw_hdmi_connector_funcs = { - .dpms = drm_helper_connector_dpms, - .fill_modes = drm_helper_probe_single_connector_modes, - .detect = dw_hdmi_connector_detect, - .destroy = dw_hdmi_connector_destroy, - .force = dw_hdmi_connector_force, -}; - -static const struct drm_connector_funcs dw_hdmi_atomic_connector_funcs = { .dpms = drm_atomic_helper_connector_dpms, .fill_modes = drm_helper_probe_single_connector_modes, .detect = dw_hdmi_connector_detect, @@ -1634,14 +1626,9 @@ static int dw_hdmi_register(struct drm_device *drm, struct dw_hdmi *hdmi) drm_connector_helper_add(&hdmi->connector, &dw_hdmi_connector_helper_funcs); - if (drm_core_check_feature(drm, DRIVER_ATOMIC)) - drm_connector_init(drm, &hdmi->connector, - &dw_hdmi_atomic_connector_funcs, - DRM_MODE_CONNECTOR_HDMIA); - else - drm_connector_init(drm, &hdmi->connector, - &dw_hdmi_connector_funcs, - DRM_MODE_CONNECTOR_HDMIA); + drm_connector_init(drm, &hdmi->connector, + &dw_hdmi_connector_funcs, + DRM_MODE_CONNECTOR_HDMIA); drm_mode_connector_attach_encoder(&hdmi->connector, encoder); diff --git a/drivers/gpu/drm/bridge/parade-ps8622.c b/drivers/gpu/drm/bridge/parade-ps8622.c index 5cd8dd7e5904..583b8ce614e3 100644 --- a/drivers/gpu/drm/bridge/parade-ps8622.c +++ b/drivers/gpu/drm/bridge/parade-ps8622.c @@ -636,9 +636,7 @@ static int ps8622_remove(struct i2c_client *client) { struct ps8622_bridge *ps8622 = i2c_get_clientdata(client); - if (ps8622->bl) - backlight_device_unregister(ps8622->bl); - + backlight_device_unregister(ps8622->bl); drm_bridge_remove(&ps8622->bridge); return 0; diff --git a/drivers/gpu/drm/bridge/tc358767.c b/drivers/gpu/drm/bridge/tc358767.c new file mode 100644 index 000000000000..a09825d8c94a --- /dev/null +++ b/drivers/gpu/drm/bridge/tc358767.c @@ -0,0 +1,1413 @@ +/* + * tc358767 eDP bridge driver + * + * Copyright (C) 2016 CogentEmbedded Inc + * Author: Andrey Gusakov <andrey.gusakov@cogentembedded.com> + * + * Copyright (C) 2016 Pengutronix, Philipp Zabel <p.zabel@pengutronix.de> + * + * Initially based on: drivers/gpu/drm/i2c/tda998x_drv.c + * + * Copyright (C) 2012 Texas Instruments + * Author: Rob Clark <robdclark@gmail.com> + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + */ + +#include <linux/clk.h> +#include <linux/device.h> +#include <linux/gpio/consumer.h> +#include <linux/i2c.h> +#include <linux/kernel.h> +#include <linux/module.h> +#include <linux/regmap.h> +#include <linux/slab.h> + +#include <drm/drm_atomic_helper.h> +#include <drm/drm_crtc_helper.h> +#include <drm/drm_dp_helper.h> +#include <drm/drm_edid.h> +#include <drm/drm_of.h> +#include <drm/drm_panel.h> + +/* Registers */ + +/* Display Parallel Interface */ +#define DPIPXLFMT 0x0440 +#define VS_POL_ACTIVE_LOW (1 << 10) +#define HS_POL_ACTIVE_LOW (1 << 9) +#define DE_POL_ACTIVE_HIGH (0 << 8) +#define SUB_CFG_TYPE_CONFIG1 (0 << 2) /* LSB aligned */ +#define SUB_CFG_TYPE_CONFIG2 (1 << 2) /* Loosely Packed */ +#define SUB_CFG_TYPE_CONFIG3 (2 << 2) /* LSB aligned 8-bit */ +#define DPI_BPP_RGB888 (0 << 0) +#define DPI_BPP_RGB666 (1 << 0) +#define DPI_BPP_RGB565 (2 << 0) + +/* Video Path */ +#define VPCTRL0 0x0450 +#define OPXLFMT_RGB666 (0 << 8) +#define OPXLFMT_RGB888 (1 << 8) +#define FRMSYNC_DISABLED (0 << 4) /* Video Timing Gen Disabled */ +#define FRMSYNC_ENABLED (1 << 4) /* Video Timing Gen Enabled */ +#define MSF_DISABLED (0 << 0) /* Magic Square FRC disabled */ +#define MSF_ENABLED (1 << 0) /* Magic Square FRC enabled */ +#define HTIM01 0x0454 +#define HTIM02 0x0458 +#define VTIM01 0x045c +#define VTIM02 0x0460 +#define VFUEN0 0x0464 +#define VFUEN BIT(0) /* Video Frame Timing Upload */ + +/* System */ +#define TC_IDREG 0x0500 +#define SYSCTRL 0x0510 +#define DP0_AUDSRC_NO_INPUT (0 << 3) +#define DP0_AUDSRC_I2S_RX (1 << 3) +#define DP0_VIDSRC_NO_INPUT (0 << 0) +#define DP0_VIDSRC_DSI_RX (1 << 0) +#define DP0_VIDSRC_DPI_RX (2 << 0) +#define DP0_VIDSRC_COLOR_BAR (3 << 0) + +/* Control */ +#define DP0CTL 0x0600 +#define VID_MN_GEN BIT(6) /* Auto-generate M/N values */ +#define EF_EN BIT(5) /* Enable Enhanced Framing */ +#define VID_EN BIT(1) /* Video transmission enable */ +#define DP_EN BIT(0) /* Enable DPTX function */ + +/* Clocks */ +#define DP0_VIDMNGEN0 0x0610 +#define DP0_VIDMNGEN1 0x0614 +#define DP0_VMNGENSTATUS 0x0618 + +/* Main Channel */ +#define DP0_SECSAMPLE 0x0640 +#define DP0_VIDSYNCDELAY 0x0644 +#define DP0_TOTALVAL 0x0648 +#define DP0_STARTVAL 0x064c +#define DP0_ACTIVEVAL 0x0650 +#define DP0_SYNCVAL 0x0654 +#define DP0_MISC 0x0658 +#define TU_SIZE_RECOMMENDED (0x3f << 16) /* LSCLK cycles per TU */ +#define BPC_6 (0 << 5) +#define BPC_8 (1 << 5) + +/* AUX channel */ +#define DP0_AUXCFG0 0x0660 +#define DP0_AUXCFG1 0x0664 +#define AUX_RX_FILTER_EN BIT(16) + +#define DP0_AUXADDR 0x0668 +#define DP0_AUXWDATA(i) (0x066c + (i) * 4) +#define DP0_AUXRDATA(i) (0x067c + (i) * 4) +#define DP0_AUXSTATUS 0x068c +#define AUX_STATUS_MASK 0xf0 +#define AUX_STATUS_SHIFT 4 +#define AUX_TIMEOUT BIT(1) +#define AUX_BUSY BIT(0) +#define DP0_AUXI2CADR 0x0698 + +/* Link Training */ +#define DP0_SRCCTRL 0x06a0 +#define DP0_SRCCTRL_SCRMBLDIS BIT(13) +#define DP0_SRCCTRL_EN810B BIT(12) +#define DP0_SRCCTRL_NOTP (0 << 8) +#define DP0_SRCCTRL_TP1 (1 << 8) +#define DP0_SRCCTRL_TP2 (2 << 8) +#define DP0_SRCCTRL_LANESKEW BIT(7) +#define DP0_SRCCTRL_SSCG BIT(3) +#define DP0_SRCCTRL_LANES_1 (0 << 2) +#define DP0_SRCCTRL_LANES_2 (1 << 2) +#define DP0_SRCCTRL_BW27 (1 << 1) +#define DP0_SRCCTRL_BW162 (0 << 1) +#define DP0_SRCCTRL_AUTOCORRECT BIT(0) +#define DP0_LTSTAT 0x06d0 +#define LT_LOOPDONE BIT(13) +#define LT_STATUS_MASK (0x1f << 8) +#define LT_CHANNEL1_EQ_BITS (DP_CHANNEL_EQ_BITS << 4) +#define LT_INTERLANE_ALIGN_DONE BIT(3) +#define LT_CHANNEL0_EQ_BITS (DP_CHANNEL_EQ_BITS) +#define DP0_SNKLTCHGREQ 0x06d4 +#define DP0_LTLOOPCTRL 0x06d8 +#define DP0_SNKLTCTRL 0x06e4 + +/* PHY */ +#define DP_PHY_CTRL 0x0800 +#define DP_PHY_RST BIT(28) /* DP PHY Global Soft Reset */ +#define BGREN BIT(25) /* AUX PHY BGR Enable */ +#define PWR_SW_EN BIT(24) /* PHY Power Switch Enable */ +#define PHY_M1_RST BIT(12) /* Reset PHY1 Main Channel */ +#define PHY_RDY BIT(16) /* PHY Main Channels Ready */ +#define PHY_M0_RST BIT(8) /* Reset PHY0 Main Channel */ +#define PHY_A0_EN BIT(1) /* PHY Aux Channel0 Enable */ +#define PHY_M0_EN BIT(0) /* PHY Main Channel0 Enable */ + +/* PLL */ +#define DP0_PLLCTRL 0x0900 +#define DP1_PLLCTRL 0x0904 /* not defined in DS */ +#define PXL_PLLCTRL 0x0908 +#define PLLUPDATE BIT(2) +#define PLLBYP BIT(1) +#define PLLEN BIT(0) +#define PXL_PLLPARAM 0x0914 +#define IN_SEL_REFCLK (0 << 14) +#define SYS_PLLPARAM 0x0918 +#define REF_FREQ_38M4 (0 << 8) /* 38.4 MHz */ +#define REF_FREQ_19M2 (1 << 8) /* 19.2 MHz */ +#define REF_FREQ_26M (2 << 8) /* 26 MHz */ +#define REF_FREQ_13M (3 << 8) /* 13 MHz */ +#define SYSCLK_SEL_LSCLK (0 << 4) +#define LSCLK_DIV_1 (0 << 0) +#define LSCLK_DIV_2 (1 << 0) + +/* Test & Debug */ +#define TSTCTL 0x0a00 +#define PLL_DBG 0x0a04 + +static bool tc_test_pattern; +module_param_named(test, tc_test_pattern, bool, 0644); + +struct tc_edp_link { + struct drm_dp_link base; + u8 assr; + int scrambler_dis; + int spread; + int coding8b10b; + u8 swing; + u8 preemp; +}; + +struct tc_data { + struct device *dev; + struct regmap *regmap; + struct drm_dp_aux aux; + + struct drm_bridge bridge; + struct drm_connector connector; + struct drm_panel *panel; + + /* link settings */ + struct tc_edp_link link; + + /* display edid */ + struct edid *edid; + /* current mode */ + struct drm_display_mode *mode; + + u32 rev; + u8 assr; + + struct gpio_desc *sd_gpio; + struct gpio_desc *reset_gpio; + struct clk *refclk; +}; + +static inline struct tc_data *aux_to_tc(struct drm_dp_aux *a) +{ + return container_of(a, struct tc_data, aux); +} + +static inline struct tc_data *bridge_to_tc(struct drm_bridge *b) +{ + return container_of(b, struct tc_data, bridge); +} + +static inline struct tc_data *connector_to_tc(struct drm_connector *c) +{ + return container_of(c, struct tc_data, connector); +} + +/* Simple macros to avoid repeated error checks */ +#define tc_write(reg, var) \ + do { \ + ret = regmap_write(tc->regmap, reg, var); \ + if (ret) \ + goto err; \ + } while (0) +#define tc_read(reg, var) \ + do { \ + ret = regmap_read(tc->regmap, reg, var); \ + if (ret) \ + goto err; \ + } while (0) + +static inline int tc_poll_timeout(struct regmap *map, unsigned int addr, + unsigned int cond_mask, + unsigned int cond_value, + unsigned long sleep_us, u64 timeout_us) +{ + ktime_t timeout = ktime_add_us(ktime_get(), timeout_us); + unsigned int val; + int ret; + + for (;;) { + ret = regmap_read(map, addr, &val); + if (ret) + break; + if ((val & cond_mask) == cond_value) + break; + if (timeout_us && ktime_compare(ktime_get(), timeout) > 0) { + ret = regmap_read(map, addr, &val); + break; + } + if (sleep_us) + usleep_range((sleep_us >> 2) + 1, sleep_us); + } + return ret ?: (((val & cond_mask) == cond_value) ? 0 : -ETIMEDOUT); +} + +static int tc_aux_wait_busy(struct tc_data *tc, unsigned int timeout_ms) +{ + return tc_poll_timeout(tc->regmap, DP0_AUXSTATUS, AUX_BUSY, 0, + 1000, 1000 * timeout_ms); +} + +static int tc_aux_get_status(struct tc_data *tc, u8 *reply) +{ + int ret; + u32 value; + + ret = regmap_read(tc->regmap, DP0_AUXSTATUS, &value); + if (ret < 0) + return ret; + if (value & AUX_BUSY) { + if (value & AUX_TIMEOUT) { + dev_err(tc->dev, "i2c access timeout!\n"); + return -ETIMEDOUT; + } + return -EBUSY; + } + + *reply = (value & AUX_STATUS_MASK) >> AUX_STATUS_SHIFT; + return 0; +} + +static ssize_t tc_aux_transfer(struct drm_dp_aux *aux, + struct drm_dp_aux_msg *msg) +{ + struct tc_data *tc = aux_to_tc(aux); + size_t size = min_t(size_t, 8, msg->size); + u8 request = msg->request & ~DP_AUX_I2C_MOT; + u8 *buf = msg->buffer; + u32 tmp = 0; + int i = 0; + int ret; + + if (size == 0) + return 0; + + ret = tc_aux_wait_busy(tc, 100); + if (ret) + goto err; + + if (request == DP_AUX_I2C_WRITE || request == DP_AUX_NATIVE_WRITE) { + /* Store data */ + while (i < size) { + if (request == DP_AUX_NATIVE_WRITE) + tmp = tmp | (buf[i] << (8 * (i & 0x3))); + else + tmp = (tmp << 8) | buf[i]; + i++; + if (((i % 4) == 0) || (i == size)) { + tc_write(DP0_AUXWDATA(i >> 2), tmp); + tmp = 0; + } + } + } else if (request != DP_AUX_I2C_READ && + request != DP_AUX_NATIVE_READ) { + return -EINVAL; + } + + /* Store address */ + tc_write(DP0_AUXADDR, msg->address); + /* Start transfer */ + tc_write(DP0_AUXCFG0, ((size - 1) << 8) | request); + + ret = tc_aux_wait_busy(tc, 100); + if (ret) + goto err; + + ret = tc_aux_get_status(tc, &msg->reply); + if (ret) + goto err; + + if (request == DP_AUX_I2C_READ || request == DP_AUX_NATIVE_READ) { + /* Read data */ + while (i < size) { + if ((i % 4) == 0) + tc_read(DP0_AUXRDATA(i >> 2), &tmp); + buf[i] = tmp & 0xff; + tmp = tmp >> 8; + i++; + } + } + + return size; +err: + return ret; +} + +static const char * const training_pattern1_errors[] = { + "No errors", + "Aux write error", + "Aux read error", + "Max voltage reached error", + "Loop counter expired error", + "res", "res", "res" +}; + +static const char * const training_pattern2_errors[] = { + "No errors", + "Aux write error", + "Aux read error", + "Clock recovery failed error", + "Loop counter expired error", + "res", "res", "res" +}; + +static u32 tc_srcctrl(struct tc_data *tc) +{ + /* + * No training pattern, skew lane 1 data by two LSCLK cycles with + * respect to lane 0 data, AutoCorrect Mode = 0 + */ + u32 reg = DP0_SRCCTRL_NOTP | DP0_SRCCTRL_LANESKEW; + + if (tc->link.scrambler_dis) + reg |= DP0_SRCCTRL_SCRMBLDIS; /* Scrambler Disabled */ + if (tc->link.coding8b10b) + /* Enable 8/10B Encoder (TxData[19:16] not used) */ + reg |= DP0_SRCCTRL_EN810B; + if (tc->link.spread) + reg |= DP0_SRCCTRL_SSCG; /* Spread Spectrum Enable */ + if (tc->link.base.num_lanes == 2) + reg |= DP0_SRCCTRL_LANES_2; /* Two Main Channel Lanes */ + if (tc->link.base.rate != 162000) + reg |= DP0_SRCCTRL_BW27; /* 2.7 Gbps link */ + return reg; +} + +static void tc_wait_pll_lock(struct tc_data *tc) +{ + /* Wait for PLL to lock: up to 2.09 ms, depending on refclk */ + usleep_range(3000, 6000); +} + +static int tc_pxl_pll_en(struct tc_data *tc, u32 refclk, u32 pixelclock) +{ + int ret; + int i_pre, best_pre = 1; + int i_post, best_post = 1; + int div, best_div = 1; + int mul, best_mul = 1; + int delta, best_delta; + int ext_div[] = {1, 2, 3, 5, 7}; + int best_pixelclock = 0; + int vco_hi = 0; + + dev_dbg(tc->dev, "PLL: requested %d pixelclock, ref %d\n", pixelclock, + refclk); + best_delta = pixelclock; + /* Loop over all possible ext_divs, skipping invalid configurations */ + for (i_pre = 0; i_pre < ARRAY_SIZE(ext_div); i_pre++) { + /* + * refclk / ext_pre_div should be in the 1 to 200 MHz range. + * We don't allow any refclk > 200 MHz, only check lower bounds. + */ + if (refclk / ext_div[i_pre] < 1000000) + continue; + for (i_post = 0; i_post < ARRAY_SIZE(ext_div); i_post++) { + for (div = 1; div <= 16; div++) { + u32 clk; + u64 tmp; + + tmp = pixelclock * ext_div[i_pre] * + ext_div[i_post] * div; + do_div(tmp, refclk); + mul = tmp; + + /* Check limits */ + if ((mul < 1) || (mul > 128)) + continue; + + clk = (refclk / ext_div[i_pre] / div) * mul; + /* + * refclk * mul / (ext_pre_div * pre_div) + * should be in the 150 to 650 MHz range + */ + if ((clk > 650000000) || (clk < 150000000)) + continue; + + clk = clk / ext_div[i_post]; + delta = clk - pixelclock; + + if (abs(delta) < abs(best_delta)) { + best_pre = i_pre; + best_post = i_post; + best_div = div; + best_mul = mul; + best_delta = delta; + best_pixelclock = clk; + } + } + } + } + if (best_pixelclock == 0) { + dev_err(tc->dev, "Failed to calc clock for %d pixelclock\n", + pixelclock); + return -EINVAL; + } + + dev_dbg(tc->dev, "PLL: got %d, delta %d\n", best_pixelclock, + best_delta); + dev_dbg(tc->dev, "PLL: %d / %d / %d * %d / %d\n", refclk, + ext_div[best_pre], best_div, best_mul, ext_div[best_post]); + + /* if VCO >= 300 MHz */ + if (refclk / ext_div[best_pre] / best_div * best_mul >= 300000000) + vco_hi = 1; + /* see DS */ + if (best_div == 16) + best_div = 0; + if (best_mul == 128) + best_mul = 0; + + /* Power up PLL and switch to bypass */ + tc_write(PXL_PLLCTRL, PLLBYP | PLLEN); + + tc_write(PXL_PLLPARAM, + (vco_hi << 24) | /* For PLL VCO >= 300 MHz = 1 */ + (ext_div[best_pre] << 20) | /* External Pre-divider */ + (ext_div[best_post] << 16) | /* External Post-divider */ + IN_SEL_REFCLK | /* Use RefClk as PLL input */ + (best_div << 8) | /* Divider for PLL RefClk */ + (best_mul << 0)); /* Multiplier for PLL */ + + /* Force PLL parameter update and disable bypass */ + tc_write(PXL_PLLCTRL, PLLUPDATE | PLLEN); + + tc_wait_pll_lock(tc); + + return 0; +err: + return ret; +} + +static int tc_pxl_pll_dis(struct tc_data *tc) +{ + /* Enable PLL bypass, power down PLL */ + return regmap_write(tc->regmap, PXL_PLLCTRL, PLLBYP); +} + +static int tc_stream_clock_calc(struct tc_data *tc) +{ + int ret; + /* + * If the Stream clock and Link Symbol clock are + * asynchronous with each other, the value of M changes over + * time. This way of generating link clock and stream + * clock is called Asynchronous Clock mode. The value M + * must change while the value N stays constant. The + * value of N in this Asynchronous Clock mode must be set + * to 2^15 or 32,768. + * + * LSCLK = 1/10 of high speed link clock + * + * f_STRMCLK = M/N * f_LSCLK + * M/N = f_STRMCLK / f_LSCLK + * + */ + tc_write(DP0_VIDMNGEN1, 32768); + + return 0; +err: + return ret; +} + +static int tc_aux_link_setup(struct tc_data *tc) +{ + unsigned long rate; + u32 value; + int ret; + + rate = clk_get_rate(tc->refclk); + switch (rate) { + case 38400000: + value = REF_FREQ_38M4; + break; + case 26000000: + value = REF_FREQ_26M; + break; + case 19200000: + value = REF_FREQ_19M2; + break; + case 13000000: + value = REF_FREQ_13M; + break; + default: + dev_err(tc->dev, "Invalid refclk rate: %lu Hz\n", rate); + return -EINVAL; + } + + /* Setup DP-PHY / PLL */ + value |= SYSCLK_SEL_LSCLK | LSCLK_DIV_2; + tc_write(SYS_PLLPARAM, value); + + tc_write(DP_PHY_CTRL, BGREN | PWR_SW_EN | BIT(2) | PHY_A0_EN); + + /* + * Initially PLLs are in bypass. Force PLL parameter update, + * disable PLL bypass, enable PLL + */ + tc_write(DP0_PLLCTRL, PLLUPDATE | PLLEN); + tc_wait_pll_lock(tc); + + tc_write(DP1_PLLCTRL, PLLUPDATE | PLLEN); + tc_wait_pll_lock(tc); + + ret = tc_poll_timeout(tc->regmap, DP_PHY_CTRL, PHY_RDY, PHY_RDY, 1, + 1000); + if (ret == -ETIMEDOUT) { + dev_err(tc->dev, "Timeout waiting for PHY to become ready"); + return ret; + } else if (ret) + goto err; + + /* Setup AUX link */ + tc_write(DP0_AUXCFG1, AUX_RX_FILTER_EN | + (0x06 << 8) | /* Aux Bit Period Calculator Threshold */ + (0x3f << 0)); /* Aux Response Timeout Timer */ + + return 0; +err: + dev_err(tc->dev, "tc_aux_link_setup failed: %d\n", ret); + return ret; +} + +static int tc_get_display_props(struct tc_data *tc) +{ + int ret; + /* temp buffer */ + u8 tmp[8]; + + /* Read DP Rx Link Capability */ + ret = drm_dp_link_probe(&tc->aux, &tc->link.base); + if (ret < 0) + goto err_dpcd_read; + if ((tc->link.base.rate != 162000) && (tc->link.base.rate != 270000)) + goto err_dpcd_inval; + + ret = drm_dp_dpcd_readb(&tc->aux, DP_MAX_DOWNSPREAD, tmp); + if (ret < 0) + goto err_dpcd_read; + tc->link.spread = tmp[0] & BIT(0); /* 0.5% down spread */ + + ret = drm_dp_dpcd_readb(&tc->aux, DP_MAIN_LINK_CHANNEL_CODING, tmp); + if (ret < 0) + goto err_dpcd_read; + tc->link.coding8b10b = tmp[0] & BIT(0); + tc->link.scrambler_dis = 0; + /* read assr */ + ret = drm_dp_dpcd_readb(&tc->aux, DP_EDP_CONFIGURATION_SET, tmp); + if (ret < 0) + goto err_dpcd_read; + tc->link.assr = tmp[0] & DP_ALTERNATE_SCRAMBLER_RESET_ENABLE; + + dev_dbg(tc->dev, "DPCD rev: %d.%d, rate: %s, lanes: %d, framing: %s\n", + tc->link.base.revision >> 4, tc->link.base.revision & 0x0f, + (tc->link.base.rate == 162000) ? "1.62Gbps" : "2.7Gbps", + tc->link.base.num_lanes, + (tc->link.base.capabilities & DP_LINK_CAP_ENHANCED_FRAMING) ? + "enhanced" : "non-enhanced"); + dev_dbg(tc->dev, "ANSI 8B/10B: %d\n", tc->link.coding8b10b); + dev_dbg(tc->dev, "Display ASSR: %d, TC358767 ASSR: %d\n", + tc->link.assr, tc->assr); + + return 0; + +err_dpcd_read: + dev_err(tc->dev, "failed to read DPCD: %d\n", ret); + return ret; +err_dpcd_inval: + dev_err(tc->dev, "invalid DPCD\n"); + return -EINVAL; +} + +static int tc_set_video_mode(struct tc_data *tc, struct drm_display_mode *mode) +{ + int ret; + int vid_sync_dly; + int max_tu_symbol; + + int left_margin = mode->htotal - mode->hsync_end; + int right_margin = mode->hsync_start - mode->hdisplay; + int hsync_len = mode->hsync_end - mode->hsync_start; + int upper_margin = mode->vtotal - mode->vsync_end; + int lower_margin = mode->vsync_start - mode->vdisplay; + int vsync_len = mode->vsync_end - mode->vsync_start; + + dev_dbg(tc->dev, "set mode %dx%d\n", + mode->hdisplay, mode->vdisplay); + dev_dbg(tc->dev, "H margin %d,%d sync %d\n", + left_margin, right_margin, hsync_len); + dev_dbg(tc->dev, "V margin %d,%d sync %d\n", + upper_margin, lower_margin, vsync_len); + dev_dbg(tc->dev, "total: %dx%d\n", mode->htotal, mode->vtotal); + + + /* LCD Ctl Frame Size */ + tc_write(VPCTRL0, (0x40 << 20) /* VSDELAY */ | + OPXLFMT_RGB888 | FRMSYNC_DISABLED | MSF_DISABLED); + tc_write(HTIM01, (left_margin << 16) | /* H back porch */ + (hsync_len << 0)); /* Hsync */ + tc_write(HTIM02, (right_margin << 16) | /* H front porch */ + (mode->hdisplay << 0)); /* width */ + tc_write(VTIM01, (upper_margin << 16) | /* V back porch */ + (vsync_len << 0)); /* Vsync */ + tc_write(VTIM02, (lower_margin << 16) | /* V front porch */ + (mode->vdisplay << 0)); /* height */ + tc_write(VFUEN0, VFUEN); /* update settings */ + + /* Test pattern settings */ + tc_write(TSTCTL, + (120 << 24) | /* Red Color component value */ + (20 << 16) | /* Green Color component value */ + (99 << 8) | /* Blue Color component value */ + (1 << 4) | /* Enable I2C Filter */ + (2 << 0) | /* Color bar Mode */ + 0); + + /* DP Main Stream Attributes */ + vid_sync_dly = hsync_len + left_margin + mode->hdisplay; + tc_write(DP0_VIDSYNCDELAY, + (0x003e << 16) | /* thresh_dly */ + (vid_sync_dly << 0)); + + tc_write(DP0_TOTALVAL, (mode->vtotal << 16) | (mode->htotal)); + + tc_write(DP0_STARTVAL, + ((upper_margin + vsync_len) << 16) | + ((left_margin + hsync_len) << 0)); + + tc_write(DP0_ACTIVEVAL, (mode->vdisplay << 16) | (mode->hdisplay)); + + tc_write(DP0_SYNCVAL, (vsync_len << 16) | (hsync_len << 0)); + + tc_write(DPIPXLFMT, VS_POL_ACTIVE_LOW | HS_POL_ACTIVE_LOW | + DE_POL_ACTIVE_HIGH | SUB_CFG_TYPE_CONFIG1 | DPI_BPP_RGB888); + + /* + * Recommended maximum number of symbols transferred in a transfer unit: + * DIV_ROUND_UP((input active video bandwidth in bytes) * tu_size, + * (output active video bandwidth in bytes)) + * Must be less than tu_size. + */ + max_tu_symbol = TU_SIZE_RECOMMENDED - 1; + tc_write(DP0_MISC, (max_tu_symbol << 23) | TU_SIZE_RECOMMENDED | BPC_8); + + return 0; +err: + return ret; +} + +static int tc_link_training(struct tc_data *tc, int pattern) +{ + const char * const *errors; + u32 srcctrl = tc_srcctrl(tc) | DP0_SRCCTRL_SCRMBLDIS | + DP0_SRCCTRL_AUTOCORRECT; + int timeout; + int retry; + u32 value; + int ret; + + if (pattern == DP_TRAINING_PATTERN_1) { + srcctrl |= DP0_SRCCTRL_TP1; + errors = training_pattern1_errors; + } else { + srcctrl |= DP0_SRCCTRL_TP2; + errors = training_pattern2_errors; + } + + /* Set DPCD 0x102 for Training Part 1 or 2 */ + tc_write(DP0_SNKLTCTRL, DP_LINK_SCRAMBLING_DISABLE | pattern); + + tc_write(DP0_LTLOOPCTRL, + (0x0f << 28) | /* Defer Iteration Count */ + (0x0f << 24) | /* Loop Iteration Count */ + (0x0d << 0)); /* Loop Timer Delay */ + + retry = 5; + do { + /* Set DP0 Training Pattern */ + tc_write(DP0_SRCCTRL, srcctrl); + + /* Enable DP0 to start Link Training */ + tc_write(DP0CTL, DP_EN); + + /* wait */ + timeout = 1000; + do { + tc_read(DP0_LTSTAT, &value); + udelay(1); + } while ((!(value & LT_LOOPDONE)) && (--timeout)); + if (timeout == 0) { + dev_err(tc->dev, "Link training timeout!\n"); + } else { + int pattern = (value >> 11) & 0x3; + int error = (value >> 8) & 0x7; + + dev_dbg(tc->dev, + "Link training phase %d done after %d uS: %s\n", + pattern, 1000 - timeout, errors[error]); + if (pattern == DP_TRAINING_PATTERN_1 && error == 0) + break; + if (pattern == DP_TRAINING_PATTERN_2) { + value &= LT_CHANNEL1_EQ_BITS | + LT_INTERLANE_ALIGN_DONE | + LT_CHANNEL0_EQ_BITS; + /* in case of two lanes */ + if ((tc->link.base.num_lanes == 2) && + (value == (LT_CHANNEL1_EQ_BITS | + LT_INTERLANE_ALIGN_DONE | + LT_CHANNEL0_EQ_BITS))) + break; + /* in case of one line */ + if ((tc->link.base.num_lanes == 1) && + (value == (LT_INTERLANE_ALIGN_DONE | + LT_CHANNEL0_EQ_BITS))) + break; + } + } + /* restart */ + tc_write(DP0CTL, 0); + usleep_range(10, 20); + } while (--retry); + if (retry == 0) { + dev_err(tc->dev, "Failed to finish training phase %d\n", + pattern); + } + + return 0; +err: + return ret; +} + +static int tc_main_link_setup(struct tc_data *tc) +{ + struct drm_dp_aux *aux = &tc->aux; + struct device *dev = tc->dev; + unsigned int rate; + u32 dp_phy_ctrl; + int timeout; + bool aligned; + bool ready; + u32 value; + int ret; + u8 tmp[8]; + + /* display mode should be set at this point */ + if (!tc->mode) + return -EINVAL; + + /* from excel file - DP0_SrcCtrl */ + tc_write(DP0_SRCCTRL, DP0_SRCCTRL_SCRMBLDIS | DP0_SRCCTRL_EN810B | + DP0_SRCCTRL_LANESKEW | DP0_SRCCTRL_LANES_2 | + DP0_SRCCTRL_BW27 | DP0_SRCCTRL_AUTOCORRECT); + /* from excel file - DP1_SrcCtrl */ + tc_write(0x07a0, 0x00003083); + + rate = clk_get_rate(tc->refclk); + switch (rate) { + case 38400000: + value = REF_FREQ_38M4; + break; + case 26000000: + value = REF_FREQ_26M; + break; + case 19200000: + value = REF_FREQ_19M2; + break; + case 13000000: + value = REF_FREQ_13M; + break; + default: + return -EINVAL; + } + value |= SYSCLK_SEL_LSCLK | LSCLK_DIV_2; + tc_write(SYS_PLLPARAM, value); + /* Setup Main Link */ + dp_phy_ctrl = BGREN | PWR_SW_EN | BIT(2) | PHY_A0_EN | PHY_M0_EN; + tc_write(DP_PHY_CTRL, dp_phy_ctrl); + msleep(100); + + /* PLL setup */ + tc_write(DP0_PLLCTRL, PLLUPDATE | PLLEN); + tc_wait_pll_lock(tc); + + tc_write(DP1_PLLCTRL, PLLUPDATE | PLLEN); + tc_wait_pll_lock(tc); + + /* PXL PLL setup */ + if (tc_test_pattern) { + ret = tc_pxl_pll_en(tc, clk_get_rate(tc->refclk), + 1000 * tc->mode->clock); + if (ret) + goto err; + } + + /* Reset/Enable Main Links */ + dp_phy_ctrl |= DP_PHY_RST | PHY_M1_RST | PHY_M0_RST; + tc_write(DP_PHY_CTRL, dp_phy_ctrl); + usleep_range(100, 200); + dp_phy_ctrl &= ~(DP_PHY_RST | PHY_M1_RST | PHY_M0_RST); + tc_write(DP_PHY_CTRL, dp_phy_ctrl); + + timeout = 1000; + do { + tc_read(DP_PHY_CTRL, &value); + udelay(1); + } while ((!(value & PHY_RDY)) && (--timeout)); + + if (timeout == 0) { + dev_err(dev, "timeout waiting for phy become ready"); + return -ETIMEDOUT; + } + + /* Set misc: 8 bits per color */ + ret = regmap_update_bits(tc->regmap, DP0_MISC, BPC_8, BPC_8); + if (ret) + goto err; + + /* + * ASSR mode + * on TC358767 side ASSR configured through strap pin + * seems there is no way to change this setting from SW + * + * check is tc configured for same mode + */ + if (tc->assr != tc->link.assr) { + dev_dbg(dev, "Trying to set display to ASSR: %d\n", + tc->assr); + /* try to set ASSR on display side */ + tmp[0] = tc->assr; + ret = drm_dp_dpcd_writeb(aux, DP_EDP_CONFIGURATION_SET, tmp[0]); + if (ret < 0) + goto err_dpcd_read; + /* read back */ + ret = drm_dp_dpcd_readb(aux, DP_EDP_CONFIGURATION_SET, tmp); + if (ret < 0) + goto err_dpcd_read; + + if (tmp[0] != tc->assr) { + dev_warn(dev, "Failed to switch display ASSR to %d, falling back to unscrambled mode\n", + tc->assr); + /* trying with disabled scrambler */ + tc->link.scrambler_dis = 1; + } + } + + /* Setup Link & DPRx Config for Training */ + ret = drm_dp_link_configure(aux, &tc->link.base); + if (ret < 0) + goto err_dpcd_write; + + /* DOWNSPREAD_CTRL */ + tmp[0] = tc->link.spread ? DP_SPREAD_AMP_0_5 : 0x00; + /* MAIN_LINK_CHANNEL_CODING_SET */ + tmp[1] = tc->link.coding8b10b ? DP_SET_ANSI_8B10B : 0x00; + ret = drm_dp_dpcd_write(aux, DP_DOWNSPREAD_CTRL, tmp, 2); + if (ret < 0) + goto err_dpcd_write; + + ret = tc_link_training(tc, DP_TRAINING_PATTERN_1); + if (ret) + goto err; + + ret = tc_link_training(tc, DP_TRAINING_PATTERN_2); + if (ret) + goto err; + + /* Clear DPCD 0x102 */ + /* Note: Can Not use DP0_SNKLTCTRL (0x06E4) short cut */ + tmp[0] = tc->link.scrambler_dis ? DP_LINK_SCRAMBLING_DISABLE : 0x00; + ret = drm_dp_dpcd_writeb(aux, DP_TRAINING_PATTERN_SET, tmp[0]); + if (ret < 0) + goto err_dpcd_write; + + /* Clear Training Pattern, set AutoCorrect Mode = 1 */ + tc_write(DP0_SRCCTRL, tc_srcctrl(tc) | DP0_SRCCTRL_AUTOCORRECT); + + /* Wait */ + timeout = 100; + do { + udelay(1); + /* Read DPCD 0x202-0x207 */ + ret = drm_dp_dpcd_read_link_status(aux, tmp + 2); + if (ret < 0) + goto err_dpcd_read; + ready = (tmp[2] == ((DP_CHANNEL_EQ_BITS << 4) | /* Lane1 */ + DP_CHANNEL_EQ_BITS)); /* Lane0 */ + aligned = tmp[4] & DP_INTERLANE_ALIGN_DONE; + } while ((--timeout) && !(ready && aligned)); + + if (timeout == 0) { + /* Read DPCD 0x200-0x201 */ + ret = drm_dp_dpcd_read(aux, DP_SINK_COUNT, tmp, 2); + if (ret < 0) + goto err_dpcd_read; + dev_info(dev, "0x0200 SINK_COUNT: 0x%02x\n", tmp[0]); + dev_info(dev, "0x0201 DEVICE_SERVICE_IRQ_VECTOR: 0x%02x\n", + tmp[1]); + dev_info(dev, "0x0202 LANE0_1_STATUS: 0x%02x\n", tmp[2]); + dev_info(dev, "0x0204 LANE_ALIGN_STATUS_UPDATED: 0x%02x\n", + tmp[4]); + dev_info(dev, "0x0205 SINK_STATUS: 0x%02x\n", tmp[5]); + dev_info(dev, "0x0206 ADJUST_REQUEST_LANE0_1: 0x%02x\n", + tmp[6]); + + if (!ready) + dev_err(dev, "Lane0/1 not ready\n"); + if (!aligned) + dev_err(dev, "Lane0/1 not aligned\n"); + return -EAGAIN; + } + + ret = tc_set_video_mode(tc, tc->mode); + if (ret) + goto err; + + /* Set M/N */ + ret = tc_stream_clock_calc(tc); + if (ret) + goto err; + + return 0; +err_dpcd_read: + dev_err(tc->dev, "Failed to read DPCD: %d\n", ret); + return ret; +err_dpcd_write: + dev_err(tc->dev, "Failed to write DPCD: %d\n", ret); +err: + return ret; +} + +static int tc_main_link_stream(struct tc_data *tc, int state) +{ + int ret; + u32 value; + + dev_dbg(tc->dev, "stream: %d\n", state); + + if (state) { + value = VID_MN_GEN | DP_EN; + if (tc->link.base.capabilities & DP_LINK_CAP_ENHANCED_FRAMING) + value |= EF_EN; + tc_write(DP0CTL, value); + /* + * VID_EN assertion should be delayed by at least N * LSCLK + * cycles from the time VID_MN_GEN is enabled in order to + * generate stable values for VID_M. LSCLK is 270 MHz or + * 162 MHz, VID_N is set to 32768 in tc_stream_clock_calc(), + * so a delay of at least 203 us should suffice. + */ + usleep_range(500, 1000); + value |= VID_EN; + tc_write(DP0CTL, value); + /* Set input interface */ + value = DP0_AUDSRC_NO_INPUT; + if (tc_test_pattern) + value |= DP0_VIDSRC_COLOR_BAR; + else + value |= DP0_VIDSRC_DPI_RX; + tc_write(SYSCTRL, value); + } else { + tc_write(DP0CTL, 0); + } + + return 0; +err: + return ret; +} + +static enum drm_connector_status +tc_connector_detect(struct drm_connector *connector, bool force) +{ + return connector_status_connected; +} + +static void tc_bridge_pre_enable(struct drm_bridge *bridge) +{ + struct tc_data *tc = bridge_to_tc(bridge); + + drm_panel_prepare(tc->panel); +} + +static void tc_bridge_enable(struct drm_bridge *bridge) +{ + struct tc_data *tc = bridge_to_tc(bridge); + int ret; + + ret = tc_main_link_setup(tc); + if (ret < 0) { + dev_err(tc->dev, "main link setup error: %d\n", ret); + return; + } + + ret = tc_main_link_stream(tc, 1); + if (ret < 0) { + dev_err(tc->dev, "main link stream start error: %d\n", ret); + return; + } + + drm_panel_enable(tc->panel); +} + +static void tc_bridge_disable(struct drm_bridge *bridge) +{ + struct tc_data *tc = bridge_to_tc(bridge); + int ret; + + drm_panel_disable(tc->panel); + + ret = tc_main_link_stream(tc, 0); + if (ret < 0) + dev_err(tc->dev, "main link stream stop error: %d\n", ret); +} + +static void tc_bridge_post_disable(struct drm_bridge *bridge) +{ + struct tc_data *tc = bridge_to_tc(bridge); + + drm_panel_unprepare(tc->panel); +} + +static bool tc_bridge_mode_fixup(struct drm_bridge *bridge, + const struct drm_display_mode *mode, + struct drm_display_mode *adj) +{ + /* Fixup sync polarities, both hsync and vsync are active low */ + adj->flags = mode->flags; + adj->flags |= (DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC); + adj->flags &= ~(DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC); + + return true; +} + +static int tc_connector_mode_valid(struct drm_connector *connector, + struct drm_display_mode *mode) +{ + /* Accept any mode */ + return MODE_OK; +} + +static void tc_bridge_mode_set(struct drm_bridge *bridge, + struct drm_display_mode *mode, + struct drm_display_mode *adj) +{ + struct tc_data *tc = bridge_to_tc(bridge); + + tc->mode = mode; +} + +static int tc_connector_get_modes(struct drm_connector *connector) +{ + struct tc_data *tc = connector_to_tc(connector); + struct edid *edid; + unsigned int count; + + if (tc->panel && tc->panel->funcs && tc->panel->funcs->get_modes) { + count = tc->panel->funcs->get_modes(tc->panel); + if (count > 0) + return count; + } + + edid = drm_get_edid(connector, &tc->aux.ddc); + + kfree(tc->edid); + tc->edid = edid; + if (!edid) + return 0; + + drm_mode_connector_update_edid_property(connector, edid); + count = drm_add_edid_modes(connector, edid); + + return count; +} + +static void tc_connector_set_polling(struct tc_data *tc, + struct drm_connector *connector) +{ + /* TODO: add support for HPD */ + connector->polled = DRM_CONNECTOR_POLL_CONNECT | + DRM_CONNECTOR_POLL_DISCONNECT; +} + +static struct drm_encoder * +tc_connector_best_encoder(struct drm_connector *connector) +{ + struct tc_data *tc = connector_to_tc(connector); + + return tc->bridge.encoder; +} + +static const struct drm_connector_helper_funcs tc_connector_helper_funcs = { + .get_modes = tc_connector_get_modes, + .mode_valid = tc_connector_mode_valid, + .best_encoder = tc_connector_best_encoder, +}; + +static void tc_connector_destroy(struct drm_connector *connector) +{ + drm_connector_unregister(connector); + drm_connector_cleanup(connector); +} + +static const struct drm_connector_funcs tc_connector_funcs = { + .dpms = drm_atomic_helper_connector_dpms, + .fill_modes = drm_helper_probe_single_connector_modes, + .detect = tc_connector_detect, + .destroy = tc_connector_destroy, + .reset = drm_atomic_helper_connector_reset, + .atomic_duplicate_state = drm_atomic_helper_connector_duplicate_state, + .atomic_destroy_state = drm_atomic_helper_connector_destroy_state, +}; + +static int tc_bridge_attach(struct drm_bridge *bridge) +{ + u32 bus_format = MEDIA_BUS_FMT_RGB888_1X24; + struct tc_data *tc = bridge_to_tc(bridge); + struct drm_device *drm = bridge->dev; + int ret; + + /* Create eDP connector */ + drm_connector_helper_add(&tc->connector, &tc_connector_helper_funcs); + ret = drm_connector_init(drm, &tc->connector, &tc_connector_funcs, + DRM_MODE_CONNECTOR_eDP); + if (ret) + return ret; + + if (tc->panel) + drm_panel_attach(tc->panel, &tc->connector); + + drm_display_info_set_bus_formats(&tc->connector.display_info, + &bus_format, 1); + drm_mode_connector_attach_encoder(&tc->connector, tc->bridge.encoder); + + return 0; +} + +static const struct drm_bridge_funcs tc_bridge_funcs = { + .attach = tc_bridge_attach, + .mode_set = tc_bridge_mode_set, + .pre_enable = tc_bridge_pre_enable, + .enable = tc_bridge_enable, + .disable = tc_bridge_disable, + .post_disable = tc_bridge_post_disable, + .mode_fixup = tc_bridge_mode_fixup, +}; + +static bool tc_readable_reg(struct device *dev, unsigned int reg) +{ + return reg != SYSCTRL; +} + +static const struct regmap_range tc_volatile_ranges[] = { + regmap_reg_range(DP0_AUXWDATA(0), DP0_AUXSTATUS), + regmap_reg_range(DP0_LTSTAT, DP0_SNKLTCHGREQ), + regmap_reg_range(DP_PHY_CTRL, DP_PHY_CTRL), + regmap_reg_range(DP0_PLLCTRL, PXL_PLLCTRL), + regmap_reg_range(VFUEN0, VFUEN0), +}; + +static const struct regmap_access_table tc_volatile_table = { + .yes_ranges = tc_volatile_ranges, + .n_yes_ranges = ARRAY_SIZE(tc_volatile_ranges), +}; + +static bool tc_writeable_reg(struct device *dev, unsigned int reg) +{ + return (reg != TC_IDREG) && + (reg != DP0_LTSTAT) && + (reg != DP0_SNKLTCHGREQ); +} + +static const struct regmap_config tc_regmap_config = { + .name = "tc358767", + .reg_bits = 16, + .val_bits = 32, + .reg_stride = 4, + .max_register = PLL_DBG, + .cache_type = REGCACHE_RBTREE, + .readable_reg = tc_readable_reg, + .volatile_table = &tc_volatile_table, + .writeable_reg = tc_writeable_reg, + .reg_format_endian = REGMAP_ENDIAN_BIG, + .val_format_endian = REGMAP_ENDIAN_LITTLE, +}; + +static int tc_probe(struct i2c_client *client, const struct i2c_device_id *id) +{ + struct device *dev = &client->dev; + struct device_node *ep; + struct tc_data *tc; + int ret; + + tc = devm_kzalloc(dev, sizeof(*tc), GFP_KERNEL); + if (!tc) + return -ENOMEM; + + tc->dev = dev; + + /* port@2 is the output port */ + ep = of_graph_get_endpoint_by_regs(dev->of_node, 2, -1); + if (ep) { + struct device_node *remote; + + remote = of_graph_get_remote_port_parent(ep); + if (!remote) { + dev_warn(dev, "endpoint %s not connected\n", + ep->full_name); + of_node_put(ep); + return -ENODEV; + } + of_node_put(ep); + tc->panel = of_drm_find_panel(remote); + if (tc->panel) { + dev_dbg(dev, "found panel %s\n", remote->full_name); + } else { + dev_dbg(dev, "waiting for panel %s\n", + remote->full_name); + of_node_put(remote); + return -EPROBE_DEFER; + } + of_node_put(remote); + } + + /* Shut down GPIO is optional */ + tc->sd_gpio = devm_gpiod_get_optional(dev, "shutdown", GPIOD_OUT_HIGH); + if (IS_ERR(tc->sd_gpio)) + return PTR_ERR(tc->sd_gpio); + + if (tc->sd_gpio) { + gpiod_set_value_cansleep(tc->sd_gpio, 0); + usleep_range(5000, 10000); + } + + /* Reset GPIO is optional */ + tc->reset_gpio = devm_gpiod_get_optional(dev, "reset", GPIOD_OUT_LOW); + if (IS_ERR(tc->reset_gpio)) + return PTR_ERR(tc->reset_gpio); + + if (tc->reset_gpio) { + gpiod_set_value_cansleep(tc->reset_gpio, 1); + usleep_range(5000, 10000); + } + + tc->refclk = devm_clk_get(dev, "ref"); + if (IS_ERR(tc->refclk)) { + ret = PTR_ERR(tc->refclk); + dev_err(dev, "Failed to get refclk: %d\n", ret); + return ret; + } + + tc->regmap = devm_regmap_init_i2c(client, &tc_regmap_config); + if (IS_ERR(tc->regmap)) { + ret = PTR_ERR(tc->regmap); + dev_err(dev, "Failed to initialize regmap: %d\n", ret); + return ret; + } + + ret = regmap_read(tc->regmap, TC_IDREG, &tc->rev); + if (ret) { + dev_err(tc->dev, "can not read device ID: %d\n", ret); + return ret; + } + + if ((tc->rev != 0x6601) && (tc->rev != 0x6603)) { + dev_err(tc->dev, "invalid device ID: 0x%08x\n", tc->rev); + return -EINVAL; + } + + tc->assr = (tc->rev == 0x6601); /* Enable ASSR for eDP panels */ + + ret = tc_aux_link_setup(tc); + if (ret) + return ret; + + /* Register DP AUX channel */ + tc->aux.name = "TC358767 AUX i2c adapter"; + tc->aux.dev = tc->dev; + tc->aux.transfer = tc_aux_transfer; + ret = drm_dp_aux_register(&tc->aux); + if (ret) + return ret; + + ret = tc_get_display_props(tc); + if (ret) + goto err_unregister_aux; + + tc_connector_set_polling(tc, &tc->connector); + + tc->bridge.funcs = &tc_bridge_funcs; + tc->bridge.of_node = dev->of_node; + ret = drm_bridge_add(&tc->bridge); + if (ret) { + dev_err(dev, "Failed to add drm_bridge: %d\n", ret); + goto err_unregister_aux; + } + + i2c_set_clientdata(client, tc); + + return 0; +err_unregister_aux: + drm_dp_aux_unregister(&tc->aux); + return ret; +} + +static int tc_remove(struct i2c_client *client) +{ + struct tc_data *tc = i2c_get_clientdata(client); + + drm_bridge_remove(&tc->bridge); + drm_dp_aux_unregister(&tc->aux); + + tc_pxl_pll_dis(tc); + + return 0; +} + +static const struct i2c_device_id tc358767_i2c_ids[] = { + { "tc358767", 0 }, + { } +}; +MODULE_DEVICE_TABLE(i2c, tc358767_i2c_ids); + +static const struct of_device_id tc358767_of_ids[] = { + { .compatible = "toshiba,tc358767", }, + { } +}; +MODULE_DEVICE_TABLE(of, tc358767_of_ids); + +static struct i2c_driver tc358767_driver = { + .driver = { + .name = "tc358767", + .of_match_table = tc358767_of_ids, + }, + .id_table = tc358767_i2c_ids, + .probe = tc_probe, + .remove = tc_remove, +}; +module_i2c_driver(tc358767_driver); + +MODULE_AUTHOR("Andrey Gusakov <andrey.gusakov@cogentembedded.com>"); +MODULE_DESCRIPTION("tc358767 eDP encoder driver"); +MODULE_LICENSE("GPL"); diff --git a/drivers/gpu/drm/cirrus/Kconfig b/drivers/gpu/drm/cirrus/Kconfig index 9864559e5fb9..04b3c161dfae 100644 --- a/drivers/gpu/drm/cirrus/Kconfig +++ b/drivers/gpu/drm/cirrus/Kconfig @@ -1,11 +1,7 @@ config DRM_CIRRUS_QEMU tristate "Cirrus driver for QEMU emulated device" depends on DRM && PCI - select FB_SYS_FILLRECT - select FB_SYS_COPYAREA - select FB_SYS_IMAGEBLIT select DRM_KMS_HELPER - select DRM_KMS_FB_HELPER select DRM_TTM help This is a KMS driver for emulated cirrus device in qemu. diff --git a/drivers/gpu/drm/cirrus/cirrus_main.c b/drivers/gpu/drm/cirrus/cirrus_main.c index 32d32c5b7b17..80446e2d3ab6 100644 --- a/drivers/gpu/drm/cirrus/cirrus_main.c +++ b/drivers/gpu/drm/cirrus/cirrus_main.c @@ -17,8 +17,8 @@ static void cirrus_user_framebuffer_destroy(struct drm_framebuffer *fb) { struct cirrus_framebuffer *cirrus_fb = to_cirrus_framebuffer(fb); - if (cirrus_fb->obj) - drm_gem_object_unreference_unlocked(cirrus_fb->obj); + + drm_gem_object_unreference_unlocked(cirrus_fb->obj); drm_framebuffer_cleanup(fb); kfree(fb); } diff --git a/drivers/gpu/drm/cirrus/cirrus_ttm.c b/drivers/gpu/drm/cirrus/cirrus_ttm.c index 6768b7b1af32..1cc9ee607128 100644 --- a/drivers/gpu/drm/cirrus/cirrus_ttm.c +++ b/drivers/gpu/drm/cirrus/cirrus_ttm.c @@ -186,17 +186,6 @@ static void cirrus_ttm_io_mem_free(struct ttm_bo_device *bdev, struct ttm_mem_re { } -static int cirrus_bo_move(struct ttm_buffer_object *bo, - bool evict, bool interruptible, - bool no_wait_gpu, - struct ttm_mem_reg *new_mem) -{ - int r; - r = ttm_bo_move_memcpy(bo, evict, no_wait_gpu, new_mem); - return r; -} - - static void cirrus_ttm_backend_destroy(struct ttm_tt *tt) { ttm_tt_fini(tt); @@ -241,7 +230,7 @@ struct ttm_bo_driver cirrus_bo_driver = { .ttm_tt_unpopulate = cirrus_ttm_tt_unpopulate, .init_mem_type = cirrus_bo_init_mem_type, .evict_flags = cirrus_bo_evict_flags, - .move = cirrus_bo_move, + .move = NULL, .verify_access = cirrus_bo_verify_access, .io_mem_reserve = &cirrus_ttm_io_mem_reserve, .io_mem_free = &cirrus_ttm_io_mem_free, diff --git a/drivers/gpu/drm/drm_atomic.c b/drivers/gpu/drm/drm_atomic.c index d99ab2f6663f..8d2f111fa113 100644 --- a/drivers/gpu/drm/drm_atomic.c +++ b/drivers/gpu/drm/drm_atomic.c @@ -404,8 +404,7 @@ drm_atomic_replace_property_blob(struct drm_property_blob **blob, if (old_blob == new_blob) return; - if (old_blob) - drm_property_unreference_blob(old_blob); + drm_property_unreference_blob(old_blob); if (new_blob) drm_property_reference_blob(new_blob); *blob = new_blob; @@ -1299,14 +1298,39 @@ EXPORT_SYMBOL(drm_atomic_add_affected_planes); */ void drm_atomic_legacy_backoff(struct drm_atomic_state *state) { + struct drm_device *dev = state->dev; + unsigned crtc_mask = 0; + struct drm_crtc *crtc; int ret; + bool global = false; + + drm_for_each_crtc(crtc, dev) { + if (crtc->acquire_ctx != state->acquire_ctx) + continue; + + crtc_mask |= drm_crtc_mask(crtc); + crtc->acquire_ctx = NULL; + } + + if (WARN_ON(dev->mode_config.acquire_ctx == state->acquire_ctx)) { + global = true; + + dev->mode_config.acquire_ctx = NULL; + } retry: drm_modeset_backoff(state->acquire_ctx); - ret = drm_modeset_lock_all_ctx(state->dev, state->acquire_ctx); + ret = drm_modeset_lock_all_ctx(dev, state->acquire_ctx); if (ret) goto retry; + + drm_for_each_crtc(crtc, dev) + if (drm_crtc_mask(crtc) & crtc_mask) + crtc->acquire_ctx = state->acquire_ctx; + + if (global) + dev->mode_config.acquire_ctx = state->acquire_ctx; } EXPORT_SYMBOL(drm_atomic_legacy_backoff); diff --git a/drivers/gpu/drm/drm_cache.c b/drivers/gpu/drm/drm_cache.c index 059f7c39c582..a7916e5f8864 100644 --- a/drivers/gpu/drm/drm_cache.c +++ b/drivers/gpu/drm/drm_cache.c @@ -136,6 +136,7 @@ drm_clflush_virt_range(void *addr, unsigned long length) mb(); for (; addr < end; addr += size) clflushopt(addr); + clflushopt(end - 1); /* force serialisation */ mb(); return; } diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c index fd93e9c79d28..f1d9f0569d7f 100644 --- a/drivers/gpu/drm/drm_crtc.c +++ b/drivers/gpu/drm/drm_crtc.c @@ -396,6 +396,51 @@ void drm_mode_object_reference(struct drm_mode_object *obj) } EXPORT_SYMBOL(drm_mode_object_reference); +/** + * drm_crtc_force_disable - Forcibly turn off a CRTC + * @crtc: CRTC to turn off + * + * Returns: + * Zero on success, error code on failure. + */ +int drm_crtc_force_disable(struct drm_crtc *crtc) +{ + struct drm_mode_set set = { + .crtc = crtc, + }; + + return drm_mode_set_config_internal(&set); +} +EXPORT_SYMBOL(drm_crtc_force_disable); + +/** + * drm_crtc_force_disable_all - Forcibly turn off all enabled CRTCs + * @dev: DRM device whose CRTCs to turn off + * + * Drivers may want to call this on unload to ensure that all displays are + * unlit and the GPU is in a consistent, low power state. Takes modeset locks. + * + * Returns: + * Zero on success, error code on failure. + */ +int drm_crtc_force_disable_all(struct drm_device *dev) +{ + struct drm_crtc *crtc; + int ret = 0; + + drm_modeset_lock_all(dev); + drm_for_each_crtc(crtc, dev) + if (crtc->enabled) { + ret = drm_crtc_force_disable(crtc); + if (ret) + goto out; + } +out: + drm_modeset_unlock_all(dev); + return ret; +} +EXPORT_SYMBOL(drm_crtc_force_disable_all); + static void drm_framebuffer_free(struct kref *kref) { struct drm_framebuffer *fb = @@ -544,8 +589,6 @@ void drm_framebuffer_remove(struct drm_framebuffer *fb) struct drm_device *dev; struct drm_crtc *crtc; struct drm_plane *plane; - struct drm_mode_set set; - int ret; if (!fb) return; @@ -575,11 +618,7 @@ void drm_framebuffer_remove(struct drm_framebuffer *fb) drm_for_each_crtc(crtc, dev) { if (crtc->primary->fb == fb) { /* should turn off the crtc */ - memset(&set, 0, sizeof(struct drm_mode_set)); - set.crtc = crtc; - set.fb = NULL; - ret = drm_mode_set_config_internal(&set); - if (ret) + if (drm_crtc_force_disable(crtc)) DRM_ERROR("failed to reset crtc %p when fb was deleted\n", crtc); } } @@ -889,11 +928,11 @@ int drm_connector_init(struct drm_device *dev, connector->dev = dev; connector->funcs = funcs; - connector->connector_id = ida_simple_get(&config->connector_ida, 0, 0, GFP_KERNEL); - if (connector->connector_id < 0) { - ret = connector->connector_id; + ret = ida_simple_get(&config->connector_ida, 0, 0, GFP_KERNEL); + if (ret < 0) goto out_put; - } + connector->index = ret; + ret = 0; connector->connector_type = connector_type; connector->connector_type_id = @@ -941,7 +980,7 @@ out_put_type_id: ida_remove(connector_ida, connector->connector_type_id); out_put_id: if (ret) - ida_remove(&config->connector_ida, connector->connector_id); + ida_remove(&config->connector_ida, connector->index); out_put: if (ret) drm_mode_object_unregister(dev, &connector->base); @@ -985,7 +1024,7 @@ void drm_connector_cleanup(struct drm_connector *connector) connector->connector_type_id); ida_remove(&dev->mode_config.connector_ida, - connector->connector_id); + connector->index); kfree(connector->display_info.bus_formats); drm_mode_object_unregister(dev, &connector->base); @@ -1068,23 +1107,16 @@ void drm_connector_unregister(struct drm_connector *connector) } EXPORT_SYMBOL(drm_connector_unregister); -/** - * drm_connector_register_all - register all connectors - * @dev: drm device - * - * This function registers all connectors in sysfs and other places so that - * userspace can start to access them. drm_connector_register_all() is called - * automatically from drm_dev_register() to complete the device registration, - * if they don't call drm_connector_register() on each connector individually. - * - * When a device is unplugged and should be removed from userspace access, - * call drm_connector_unregister_all(), which is the inverse of this - * function. - * - * Returns: - * Zero on success, error code on failure. - */ -int drm_connector_register_all(struct drm_device *dev) +static void drm_connector_unregister_all(struct drm_device *dev) +{ + struct drm_connector *connector; + + /* FIXME: taking the mode config mutex ends up in a clash with sysfs */ + list_for_each_entry(connector, &dev->mode_config.connector_list, head) + drm_connector_unregister(connector); +} + +static int drm_connector_register_all(struct drm_device *dev) { struct drm_connector *connector; int ret; @@ -1106,27 +1138,6 @@ err: drm_connector_unregister_all(dev); return ret; } -EXPORT_SYMBOL(drm_connector_register_all); - -/** - * drm_connector_unregister_all - unregister connector userspace interfaces - * @dev: drm device - * - * This functions unregisters all connectors from sysfs and other places so - * that userspace can no longer access them. Drivers should call this as the - * first step tearing down the device instace, or when the underlying - * physical device disappeared (e.g. USB unplug), right before calling - * drm_dev_unregister(). - */ -void drm_connector_unregister_all(struct drm_device *dev) -{ - struct drm_connector *connector; - - /* FIXME: taking the mode config mutex ends up in a clash with sysfs */ - list_for_each_entry(connector, &dev->mode_config.connector_list, head) - drm_connector_unregister(connector); -} -EXPORT_SYMBOL(drm_connector_unregister_all); static int drm_encoder_register_all(struct drm_device *dev) { diff --git a/drivers/gpu/drm/drm_crtc_helper.c b/drivers/gpu/drm/drm_crtc_helper.c index d61591274ff6..604d3ef72ffa 100644 --- a/drivers/gpu/drm/drm_crtc_helper.c +++ b/drivers/gpu/drm/drm_crtc_helper.c @@ -531,11 +531,11 @@ drm_crtc_helper_disable(struct drm_crtc *crtc) int drm_crtc_helper_set_config(struct drm_mode_set *set) { struct drm_device *dev; - struct drm_crtc *new_crtc; - struct drm_encoder *save_encoders, *new_encoder, *encoder; + struct drm_crtc **save_encoder_crtcs, *new_crtc; + struct drm_encoder **save_connector_encoders, *new_encoder, *encoder; bool mode_changed = false; /* if true do a full mode set */ bool fb_changed = false; /* if true and !mode_changed just do a flip */ - struct drm_connector *save_connectors, *connector; + struct drm_connector *connector; int count = 0, ro, fail = 0; const struct drm_crtc_helper_funcs *crtc_funcs; struct drm_mode_set save_set; @@ -577,15 +577,15 @@ int drm_crtc_helper_set_config(struct drm_mode_set *set) * Allocate space for the backup of all (non-pointer) encoder and * connector data. */ - save_encoders = kzalloc(dev->mode_config.num_encoder * - sizeof(struct drm_encoder), GFP_KERNEL); - if (!save_encoders) + save_encoder_crtcs = kzalloc(dev->mode_config.num_encoder * + sizeof(struct drm_crtc *), GFP_KERNEL); + if (!save_encoder_crtcs) return -ENOMEM; - save_connectors = kzalloc(dev->mode_config.num_connector * - sizeof(struct drm_connector), GFP_KERNEL); - if (!save_connectors) { - kfree(save_encoders); + save_connector_encoders = kzalloc(dev->mode_config.num_connector * + sizeof(struct drm_encoder *), GFP_KERNEL); + if (!save_connector_encoders) { + kfree(save_encoder_crtcs); return -ENOMEM; } @@ -596,12 +596,12 @@ int drm_crtc_helper_set_config(struct drm_mode_set *set) */ count = 0; drm_for_each_encoder(encoder, dev) { - save_encoders[count++] = *encoder; + save_encoder_crtcs[count++] = encoder->crtc; } count = 0; drm_for_each_connector(connector, dev) { - save_connectors[count++] = *connector; + save_connector_encoders[count++] = connector->encoder; } save_set.crtc = set->crtc; @@ -634,8 +634,12 @@ int drm_crtc_helper_set_config(struct drm_mode_set *set) mode_changed = true; } - /* take a reference on all connectors in set */ + /* take a reference on all unbound connectors in set, reuse the + * already taken reference for bound connectors + */ for (ro = 0; ro < set->num_connectors; ro++) { + if (set->connectors[ro]->encoder) + continue; drm_connector_reference(set->connectors[ro]); } @@ -757,30 +761,28 @@ int drm_crtc_helper_set_config(struct drm_mode_set *set) } } - /* after fail drop reference on all connectors in save set */ - count = 0; - drm_for_each_connector(connector, dev) { - drm_connector_unreference(&save_connectors[count++]); - } - - kfree(save_connectors); - kfree(save_encoders); + kfree(save_connector_encoders); + kfree(save_encoder_crtcs); return 0; fail: /* Restore all previous data. */ count = 0; drm_for_each_encoder(encoder, dev) { - *encoder = save_encoders[count++]; + encoder->crtc = save_encoder_crtcs[count++]; } count = 0; drm_for_each_connector(connector, dev) { - *connector = save_connectors[count++]; + connector->encoder = save_connector_encoders[count++]; } - /* after fail drop reference on all connectors in set */ + /* after fail drop reference on all unbound connectors in set, let + * bound connectors keep their reference + */ for (ro = 0; ro < set->num_connectors; ro++) { + if (set->connectors[ro]->encoder) + continue; drm_connector_unreference(set->connectors[ro]); } @@ -790,8 +792,8 @@ fail: save_set.y, save_set.fb)) DRM_ERROR("failed to restore config after modeset failure\n"); - kfree(save_connectors); - kfree(save_encoders); + kfree(save_connector_encoders); + kfree(save_encoder_crtcs); return ret; } EXPORT_SYMBOL(drm_crtc_helper_set_config); diff --git a/drivers/gpu/drm/drm_dp_aux_dev.c b/drivers/gpu/drm/drm_dp_aux_dev.c index 3334baacf43d..734f86a345f6 100644 --- a/drivers/gpu/drm/drm_dp_aux_dev.c +++ b/drivers/gpu/drm/drm_dp_aux_dev.c @@ -355,8 +355,7 @@ int drm_dp_aux_dev_init(void) drm_dp_aux_dev_class = class_create(THIS_MODULE, "drm_dp_aux_dev"); if (IS_ERR(drm_dp_aux_dev_class)) { - res = PTR_ERR(drm_dp_aux_dev_class); - goto out; + return PTR_ERR(drm_dp_aux_dev_class); } drm_dp_aux_dev_class->dev_groups = drm_dp_aux_groups; diff --git a/drivers/gpu/drm/drm_dp_helper.c b/drivers/gpu/drm/drm_dp_helper.c index 091053e995e5..8f11b8741e42 100644 --- a/drivers/gpu/drm/drm_dp_helper.c +++ b/drivers/gpu/drm/drm_dp_helper.c @@ -203,7 +203,7 @@ static int drm_dp_dpcd_access(struct drm_dp_aux *aux, u8 request, ret = aux->transfer(aux, &msg); - if (ret > 0) { + if (ret >= 0) { native_reply = msg.reply & DP_AUX_NATIVE_REPLY_MASK; if (native_reply == DP_AUX_NATIVE_REPLY_ACK) { if (ret == size) diff --git a/drivers/gpu/drm/drm_dp_mst_topology.c b/drivers/gpu/drm/drm_dp_mst_topology.c index a13edf5de2d6..04e457117980 100644 --- a/drivers/gpu/drm/drm_dp_mst_topology.c +++ b/drivers/gpu/drm/drm_dp_mst_topology.c @@ -1493,11 +1493,8 @@ static void process_single_down_tx_qlock(struct drm_dp_mst_topology_mgr *mgr) WARN_ON(!mutex_is_locked(&mgr->qlock)); /* construct a chunk from the first msg in the tx_msg queue */ - if (list_empty(&mgr->tx_msg_downq)) { - mgr->tx_down_in_progress = false; + if (list_empty(&mgr->tx_msg_downq)) return; - } - mgr->tx_down_in_progress = true; txmsg = list_first_entry(&mgr->tx_msg_downq, struct drm_dp_sideband_msg_tx, next); ret = process_single_tx_qlock(mgr, txmsg, false); @@ -1512,10 +1509,6 @@ static void process_single_down_tx_qlock(struct drm_dp_mst_topology_mgr *mgr) txmsg->state = DRM_DP_SIDEBAND_TX_TIMEOUT; wake_up(&mgr->tx_waitq); } - if (list_empty(&mgr->tx_msg_downq)) { - mgr->tx_down_in_progress = false; - return; - } } /* called holding qlock */ @@ -1538,7 +1531,7 @@ static void drm_dp_queue_down_tx(struct drm_dp_mst_topology_mgr *mgr, { mutex_lock(&mgr->qlock); list_add_tail(&txmsg->next, &mgr->tx_msg_downq); - if (!mgr->tx_down_in_progress) + if (list_is_singular(&mgr->tx_msg_downq)) process_single_down_tx_qlock(mgr); mutex_unlock(&mgr->qlock); } @@ -2372,6 +2365,7 @@ EXPORT_SYMBOL(drm_dp_mst_hpd_irq); /** * drm_dp_mst_detect_port() - get connection status for an MST port + * @connector: DRM connector for this port * @mgr: manager for this port * @port: unverified pointer to a port * @@ -2887,7 +2881,7 @@ static void drm_dp_tx_work(struct work_struct *work) struct drm_dp_mst_topology_mgr *mgr = container_of(work, struct drm_dp_mst_topology_mgr, tx_work); mutex_lock(&mgr->qlock); - if (mgr->tx_down_in_progress) + if (!list_empty(&mgr->tx_msg_downq)) process_single_down_tx_qlock(mgr); mutex_unlock(&mgr->qlock); } @@ -2927,11 +2921,9 @@ static void drm_dp_destroy_connector_work(struct work_struct *work) drm_dp_port_teardown_pdt(port, port->pdt); if (!port->input && port->vcpi.vcpi > 0) { - if (mgr->mst_state) { - drm_dp_mst_reset_vcpi_slots(mgr, port); - drm_dp_update_payload_part1(mgr); - drm_dp_mst_put_payload_id(mgr, port->vcpi.vcpi); - } + drm_dp_mst_reset_vcpi_slots(mgr, port); + drm_dp_update_payload_part1(mgr); + drm_dp_mst_put_payload_id(mgr, port->vcpi.vcpi); } kref_put(&port->kref, drm_dp_free_mst_port); diff --git a/drivers/gpu/drm/drm_drv.c b/drivers/gpu/drm/drm_drv.c index aead9ffcbe29..be27ed36f56e 100644 --- a/drivers/gpu/drm/drm_drv.c +++ b/drivers/gpu/drm/drm_drv.c @@ -362,9 +362,7 @@ EXPORT_SYMBOL(drm_put_dev); void drm_unplug_dev(struct drm_device *dev) { /* for a USB device */ - drm_minor_unregister(dev, DRM_MINOR_LEGACY); - drm_minor_unregister(dev, DRM_MINOR_RENDER); - drm_minor_unregister(dev, DRM_MINOR_CONTROL); + drm_dev_unregister(dev); mutex_lock(&drm_global_mutex); diff --git a/drivers/gpu/drm/drm_fb_cma_helper.c b/drivers/gpu/drm/drm_fb_cma_helper.c index c0b0c718994a..1fd6eac1400c 100644 --- a/drivers/gpu/drm/drm_fb_cma_helper.c +++ b/drivers/gpu/drm/drm_fb_cma_helper.c @@ -596,3 +596,18 @@ void drm_fbdev_cma_hotplug_event(struct drm_fbdev_cma *fbdev_cma) drm_fb_helper_hotplug_event(&fbdev_cma->fb_helper); } EXPORT_SYMBOL_GPL(drm_fbdev_cma_hotplug_event); + +/** + * drm_fbdev_cma_set_suspend - wrapper around drm_fb_helper_set_suspend + * @fbdev_cma: The drm_fbdev_cma struct, may be NULL + * @state: desired state, zero to resume, non-zero to suspend + * + * Calls drm_fb_helper_set_suspend, which is a wrapper around + * fb_set_suspend implemented by fbdev core. + */ +void drm_fbdev_cma_set_suspend(struct drm_fbdev_cma *fbdev_cma, int state) +{ + if (fbdev_cma) + drm_fb_helper_set_suspend(&fbdev_cma->fb_helper, state); +} +EXPORT_SYMBOL(drm_fbdev_cma_set_suspend); diff --git a/drivers/gpu/drm/drm_ioctl.c b/drivers/gpu/drm/drm_ioctl.c index 1f84ff5f1bf8..33af4a5ddca1 100644 --- a/drivers/gpu/drm/drm_ioctl.c +++ b/drivers/gpu/drm/drm_ioctl.c @@ -648,7 +648,7 @@ long drm_ioctl(struct file *filp, int retcode = -EINVAL; char stack_kdata[128]; char *kdata = NULL; - unsigned int usize, asize, drv_size; + unsigned int in_size, out_size, drv_size, ksize; bool is_driver_ioctl; dev = file_priv->minor->dev; @@ -671,9 +671,12 @@ long drm_ioctl(struct file *filp, } drv_size = _IOC_SIZE(ioctl->cmd); - usize = _IOC_SIZE(cmd); - asize = max(usize, drv_size); - cmd = ioctl->cmd; + out_size = in_size = _IOC_SIZE(cmd); + if ((cmd & ioctl->cmd & IOC_IN) == 0) + in_size = 0; + if ((cmd & ioctl->cmd & IOC_OUT) == 0) + out_size = 0; + ksize = max(max(in_size, out_size), drv_size); DRM_DEBUG("pid=%d, dev=0x%lx, auth=%d, %s\n", task_pid_nr(current), @@ -693,30 +696,24 @@ long drm_ioctl(struct file *filp, if (unlikely(retcode)) goto err_i1; - if (cmd & (IOC_IN | IOC_OUT)) { - if (asize <= sizeof(stack_kdata)) { - kdata = stack_kdata; - } else { - kdata = kmalloc(asize, GFP_KERNEL); - if (!kdata) { - retcode = -ENOMEM; - goto err_i1; - } + if (ksize <= sizeof(stack_kdata)) { + kdata = stack_kdata; + } else { + kdata = kmalloc(ksize, GFP_KERNEL); + if (!kdata) { + retcode = -ENOMEM; + goto err_i1; } - if (asize > usize) - memset(kdata + usize, 0, asize - usize); } - if (cmd & IOC_IN) { - if (copy_from_user(kdata, (void __user *)arg, - usize) != 0) { - retcode = -EFAULT; - goto err_i1; - } - } else if (cmd & IOC_OUT) { - memset(kdata, 0, usize); + if (copy_from_user(kdata, (void __user *)arg, in_size) != 0) { + retcode = -EFAULT; + goto err_i1; } + if (ksize > in_size) + memset(kdata + in_size, 0, ksize - in_size); + /* Enforce sane locking for kms driver ioctls. Core ioctls are * too messy still. */ if ((drm_core_check_feature(dev, DRIVER_MODESET) && is_driver_ioctl) || @@ -728,11 +725,8 @@ long drm_ioctl(struct file *filp, mutex_unlock(&drm_global_mutex); } - if (cmd & IOC_OUT) { - if (copy_to_user((void __user *)arg, kdata, - usize) != 0) - retcode = -EFAULT; - } + if (copy_to_user((void __user *)arg, kdata, out_size) != 0) + retcode = -EFAULT; err_i1: if (!ioctl) @@ -759,7 +753,7 @@ EXPORT_SYMBOL(drm_ioctl); * shouldn't be used by any drivers. * * Returns: - * True if the @nr corresponds to a DRM core ioctl numer, false otherwise. + * True if the @nr corresponds to a DRM core ioctl number, false otherwise. */ bool drm_ioctl_flags(unsigned int nr, unsigned int *flags) { diff --git a/drivers/gpu/drm/drm_irq.c b/drivers/gpu/drm/drm_irq.c index 8ca3d2bf2bda..77f357b2c386 100644 --- a/drivers/gpu/drm/drm_irq.c +++ b/drivers/gpu/drm/drm_irq.c @@ -532,7 +532,7 @@ int drm_irq_uninstall(struct drm_device *dev) /* * Wake up any waiters so they don't hang. This is just to paper over - * isssues for UMS drivers which aren't in full control of their + * issues for UMS drivers which aren't in full control of their * vblank/irq handling. KMS drivers must ensure that vblanks are all * disabled when uninstalling the irq handler. */ @@ -594,7 +594,7 @@ int drm_control(struct drm_device *dev, void *data, return 0; if (drm_core_check_feature(dev, DRIVER_MODESET)) return 0; - /* UMS was only ever support on pci devices. */ + /* UMS was only ever supported on pci devices. */ if (WARN_ON(!dev->pdev)) return -EINVAL; @@ -945,8 +945,8 @@ EXPORT_SYMBOL(drm_crtc_vblank_count); * * This is the legacy version of drm_crtc_vblank_count_and_time(). */ -u32 drm_vblank_count_and_time(struct drm_device *dev, unsigned int pipe, - struct timeval *vblanktime) +static u32 drm_vblank_count_and_time(struct drm_device *dev, unsigned int pipe, + struct timeval *vblanktime) { struct drm_vblank_crtc *vblank = &dev->vblank[pipe]; u32 vblank_count; @@ -963,7 +963,6 @@ u32 drm_vblank_count_and_time(struct drm_device *dev, unsigned int pipe, return vblank_count; } -EXPORT_SYMBOL(drm_vblank_count_and_time); /** * drm_crtc_vblank_count_and_time - retrieve "cooked" vblank counter value @@ -975,8 +974,6 @@ EXPORT_SYMBOL(drm_vblank_count_and_time); * vblank events since the system was booted, including lost events due to * modesetting activity. Returns corresponding system timestamp of the time * of the vblank interval that corresponds to the current vblank counter value. - * - * This is the native KMS version of drm_vblank_count_and_time(). */ u32 drm_crtc_vblank_count_and_time(struct drm_crtc *crtc, struct timeval *vblanktime) @@ -1588,12 +1585,6 @@ static int drm_queue_vblank_event(struct drm_device *dev, unsigned int pipe, seq = drm_vblank_count_and_time(dev, pipe, &now); - if ((vblwait->request.type & _DRM_VBLANK_NEXTONMISS) && - (seq - vblwait->request.sequence) <= (1 << 23)) { - vblwait->request.sequence = seq + 1; - vblwait->reply.sequence = vblwait->request.sequence; - } - DRM_DEBUG("event on vblank count %d, current %d, crtc %u\n", vblwait->request.sequence, seq, pipe); @@ -1690,6 +1681,11 @@ int drm_wait_vblank(struct drm_device *dev, void *data, goto done; } + if ((flags & _DRM_VBLANK_NEXTONMISS) && + (seq - vblwait->request.sequence) <= (1 << 23)) { + vblwait->request.sequence = seq + 1; + } + if (flags & _DRM_VBLANK_EVENT) { /* must hold on to the vblank ref until the event fires * drm_vblank_put will be called asynchronously @@ -1697,14 +1693,8 @@ int drm_wait_vblank(struct drm_device *dev, void *data, return drm_queue_vblank_event(dev, pipe, vblwait, file_priv); } - if ((flags & _DRM_VBLANK_NEXTONMISS) && - (seq - vblwait->request.sequence) <= (1<<23)) { - vblwait->request.sequence = seq + 1; - } - DRM_DEBUG("waiting on vblank count %d, crtc %u\n", vblwait->request.sequence, pipe); - vblank->last_wait = vblwait->request.sequence; DRM_WAIT_ON(ret, vblank->queue, 3 * HZ, (((drm_vblank_count(dev, pipe) - vblwait->request.sequence) <= (1 << 23)) || diff --git a/drivers/gpu/drm/drm_memory.c b/drivers/gpu/drm/drm_memory.c index 87a8cb73366f..fc0ebd273ef8 100644 --- a/drivers/gpu/drm/drm_memory.c +++ b/drivers/gpu/drm/drm_memory.c @@ -44,7 +44,7 @@ # include <asm/agp.h> #else # ifdef __powerpc__ -# define PAGE_AGP __pgprot(_PAGE_KERNEL | _PAGE_NO_CACHE) +# define PAGE_AGP pgprot_noncached_wc(PAGE_KERNEL) # else # define PAGE_AGP PAGE_KERNEL # endif diff --git a/drivers/gpu/drm/drm_mipi_dsi.c b/drivers/gpu/drm/drm_mipi_dsi.c index 49311fc61d5d..af0d471ee246 100644 --- a/drivers/gpu/drm/drm_mipi_dsi.c +++ b/drivers/gpu/drm/drm_mipi_dsi.c @@ -999,17 +999,17 @@ int mipi_dsi_dcs_set_tear_on(struct mipi_dsi_device *dsi, EXPORT_SYMBOL(mipi_dsi_dcs_set_tear_on); /** - * mipi_dsi_set_tear_scanline() - turn on the display module's Tearing Effect - * output signal on the TE signal line when display module reaches line N - * defined by STS[n:0]. + * mipi_dsi_dcs_set_tear_scanline() - set the scanline to use as trigger for + * the Tearing Effect output signal of the display module * @dsi: DSI peripheral device - * @param: STS[10:0] + * @scanline: scanline to use as trigger + * * Return: 0 on success or a negative error code on failure */ -int mipi_dsi_set_tear_scanline(struct mipi_dsi_device *dsi, u16 param) +int mipi_dsi_dcs_set_tear_scanline(struct mipi_dsi_device *dsi, u16 scanline) { - u8 payload[3] = { MIPI_DCS_SET_TEAR_SCANLINE, param >> 8, - param & 0xff }; + u8 payload[3] = { MIPI_DCS_SET_TEAR_SCANLINE, scanline >> 8, + scanline & 0xff }; ssize_t err; err = mipi_dsi_generic_write(dsi, payload, sizeof(payload)); @@ -1018,7 +1018,7 @@ int mipi_dsi_set_tear_scanline(struct mipi_dsi_device *dsi, u16 param) return 0; } -EXPORT_SYMBOL(mipi_dsi_set_tear_scanline); +EXPORT_SYMBOL(mipi_dsi_dcs_set_tear_scanline); /** * mipi_dsi_dcs_set_pixel_format() - sets the pixel format for the RGB image diff --git a/drivers/gpu/drm/drm_scatter.c b/drivers/gpu/drm/drm_scatter.c index 4f0f3b36d537..bf70431073f6 100644 --- a/drivers/gpu/drm/drm_scatter.c +++ b/drivers/gpu/drm/drm_scatter.c @@ -41,7 +41,7 @@ static inline void *drm_vmalloc_dma(unsigned long size) { #if defined(__powerpc__) && defined(CONFIG_NOT_COHERENT_CACHE) - return __vmalloc(size, GFP_KERNEL, PAGE_KERNEL | _PAGE_NO_CACHE); + return __vmalloc(size, GFP_KERNEL, pgprot_noncached_wc(PAGE_KERNEL)); #else return vmalloc_32(size); #endif diff --git a/drivers/gpu/drm/drm_vm.c b/drivers/gpu/drm/drm_vm.c index 43ff44a2b8e7..caa4e4ca616d 100644 --- a/drivers/gpu/drm/drm_vm.c +++ b/drivers/gpu/drm/drm_vm.c @@ -80,7 +80,7 @@ static pgprot_t drm_dma_prot(uint32_t map_type, struct vm_area_struct *vma) pgprot_t tmp = vm_get_page_prot(vma->vm_flags); #if defined(__powerpc__) && defined(CONFIG_NOT_COHERENT_CACHE) - tmp |= _PAGE_NO_CACHE; + tmp = pgprot_noncached_wc(tmp); #endif return tmp; } @@ -593,7 +593,7 @@ static int drm_mmap_locked(struct file *filp, struct vm_area_struct *vma) * pages and mappings in fault() */ #if defined(__powerpc__) - pgprot_val(vma->vm_page_prot) |= _PAGE_NO_CACHE; + vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot); #endif vma->vm_ops = &drm_vm_ops; break; diff --git a/drivers/gpu/drm/etnaviv/etnaviv_drv.c b/drivers/gpu/drm/etnaviv/etnaviv_drv.c index 340d390306d8..ffd1b32caa8d 100644 --- a/drivers/gpu/drm/etnaviv/etnaviv_drv.c +++ b/drivers/gpu/drm/etnaviv/etnaviv_drv.c @@ -91,10 +91,8 @@ static void load_gpu(struct drm_device *dev) int ret; ret = etnaviv_gpu_init(g); - if (ret) { - dev_err(g->dev, "hw init failed: %d\n", ret); + if (ret) priv->gpu[i] = NULL; - } } } } diff --git a/drivers/gpu/drm/etnaviv/etnaviv_gem.c b/drivers/gpu/drm/etnaviv/etnaviv_gem.c index df9bcbab922f..56fb8637bb57 100644 --- a/drivers/gpu/drm/etnaviv/etnaviv_gem.c +++ b/drivers/gpu/drm/etnaviv/etnaviv_gem.c @@ -535,8 +535,7 @@ void etnaviv_gem_describe_objects(struct etnaviv_drm_private *priv, static void etnaviv_gem_shmem_release(struct etnaviv_gem_object *etnaviv_obj) { - if (etnaviv_obj->vaddr) - vunmap(etnaviv_obj->vaddr); + vunmap(etnaviv_obj->vaddr); put_pages(etnaviv_obj); } @@ -670,9 +669,7 @@ static struct drm_gem_object *__etnaviv_gem_new(struct drm_device *dev, return obj; fail: - if (obj) - drm_gem_object_unreference_unlocked(obj); - + drm_gem_object_unreference_unlocked(obj); return ERR_PTR(ret); } @@ -916,15 +913,12 @@ int etnaviv_gem_new_userptr(struct drm_device *dev, struct drm_file *file, get_task_struct(current); ret = etnaviv_gem_obj_add(dev, &etnaviv_obj->base); - if (ret) { - drm_gem_object_unreference_unlocked(&etnaviv_obj->base); - return ret; - } + if (ret) + goto unreference; ret = drm_gem_handle_create(file, &etnaviv_obj->base, handle); - +unreference: /* drop reference from allocate - handle holds it now */ drm_gem_object_unreference_unlocked(&etnaviv_obj->base); - return ret; } diff --git a/drivers/gpu/drm/etnaviv/etnaviv_gpu.c b/drivers/gpu/drm/etnaviv/etnaviv_gpu.c index ff6aa5dfb2d7..87ef34150d46 100644 --- a/drivers/gpu/drm/etnaviv/etnaviv_gpu.c +++ b/drivers/gpu/drm/etnaviv/etnaviv_gpu.c @@ -487,6 +487,47 @@ static int etnaviv_hw_reset(struct etnaviv_gpu *gpu) return 0; } +static void etnaviv_gpu_enable_mlcg(struct etnaviv_gpu *gpu) +{ + u32 pmc, ppc; + + /* enable clock gating */ + ppc = gpu_read(gpu, VIVS_PM_POWER_CONTROLS); + ppc |= VIVS_PM_POWER_CONTROLS_ENABLE_MODULE_CLOCK_GATING; + + /* Disable stall module clock gating for 4.3.0.1 and 4.3.0.2 revs */ + if (gpu->identity.revision == 0x4301 || + gpu->identity.revision == 0x4302) + ppc |= VIVS_PM_POWER_CONTROLS_DISABLE_STALL_MODULE_CLOCK_GATING; + + gpu_write(gpu, VIVS_PM_POWER_CONTROLS, ppc); + + pmc = gpu_read(gpu, VIVS_PM_MODULE_CONTROLS); + + /* Disable PA clock gating for GC400+ except for GC420 */ + if (gpu->identity.model >= chipModel_GC400 && + gpu->identity.model != chipModel_GC420) + pmc |= VIVS_PM_MODULE_CONTROLS_DISABLE_MODULE_CLOCK_GATING_PA; + + /* + * Disable PE clock gating on revs < 5.0.0.0 when HZ is + * present without a bug fix. + */ + if (gpu->identity.revision < 0x5000 && + gpu->identity.minor_features0 & chipMinorFeatures0_HZ && + !(gpu->identity.minor_features1 & + chipMinorFeatures1_DISABLE_PE_GATING)) + pmc |= VIVS_PM_MODULE_CONTROLS_DISABLE_MODULE_CLOCK_GATING_PE; + + if (gpu->identity.revision < 0x5422) + pmc |= BIT(15); /* Unknown bit */ + + pmc |= VIVS_PM_MODULE_CONTROLS_DISABLE_MODULE_CLOCK_GATING_RA_HZ; + pmc |= VIVS_PM_MODULE_CONTROLS_DISABLE_MODULE_CLOCK_GATING_RA_EZ; + + gpu_write(gpu, VIVS_PM_MODULE_CONTROLS, pmc); +} + static void etnaviv_gpu_hw_init(struct etnaviv_gpu *gpu) { u16 prefetch; @@ -506,6 +547,9 @@ static void etnaviv_gpu_hw_init(struct etnaviv_gpu *gpu) gpu_write(gpu, VIVS_MC_DEBUG_MEMORY, mc_memory_debug); } + /* enable module-level clock gating */ + etnaviv_gpu_enable_mlcg(gpu); + /* * Update GPU AXI cache atttribute to "cacheable, no allocate". * This is necessary to prevent the iMX6 SoC locking up. @@ -553,8 +597,10 @@ int etnaviv_gpu_init(struct etnaviv_gpu *gpu) bool mmuv2; ret = pm_runtime_get_sync(gpu->dev); - if (ret < 0) + if (ret < 0) { + dev_err(gpu->dev, "Failed to enable GPU power domain\n"); return ret; + } etnaviv_hw_identify(gpu); @@ -591,8 +637,10 @@ int etnaviv_gpu_init(struct etnaviv_gpu *gpu) } ret = etnaviv_hw_reset(gpu); - if (ret) + if (ret) { + dev_err(gpu->dev, "GPU reset failed\n"); goto fail; + } /* Setup IOMMU.. eventually we will (I think) do this once per context * and have separate page tables per context. For now, to keep things @@ -610,12 +658,14 @@ int etnaviv_gpu_init(struct etnaviv_gpu *gpu) } if (!iommu) { + dev_err(gpu->dev, "Failed to allocate GPU IOMMU domain\n"); ret = -ENOMEM; goto fail; } gpu->mmu = etnaviv_iommu_new(gpu, iommu, version); if (!gpu->mmu) { + dev_err(gpu->dev, "Failed to instantiate GPU IOMMU\n"); iommu_domain_free(iommu); ret = -ENOMEM; goto fail; diff --git a/drivers/gpu/drm/etnaviv/etnaviv_iommu.c b/drivers/gpu/drm/etnaviv/etnaviv_iommu.c index 522cfd447892..16353ee81651 100644 --- a/drivers/gpu/drm/etnaviv/etnaviv_iommu.c +++ b/drivers/gpu/drm/etnaviv/etnaviv_iommu.c @@ -225,6 +225,7 @@ struct iommu_domain *etnaviv_iommu_domain_alloc(struct etnaviv_gpu *gpu) etnaviv_domain->domain.type = __IOMMU_DOMAIN_PAGING; etnaviv_domain->domain.ops = &etnaviv_iommu_ops.ops; + etnaviv_domain->domain.pgsize_bitmap = SZ_4K; etnaviv_domain->domain.geometry.aperture_start = GPU_MEM_START; etnaviv_domain->domain.geometry.aperture_end = GPU_MEM_START + PT_ENTRIES * SZ_4K - 1; diff --git a/drivers/gpu/drm/etnaviv/state_hi.xml.h b/drivers/gpu/drm/etnaviv/state_hi.xml.h index 6a7de5f1454a..807a3d9e0dd5 100644 --- a/drivers/gpu/drm/etnaviv/state_hi.xml.h +++ b/drivers/gpu/drm/etnaviv/state_hi.xml.h @@ -218,6 +218,13 @@ Copyright (C) 2015 #define VIVS_PM_MODULE_CONTROLS_DISABLE_MODULE_CLOCK_GATING_FE 0x00000001 #define VIVS_PM_MODULE_CONTROLS_DISABLE_MODULE_CLOCK_GATING_DE 0x00000002 #define VIVS_PM_MODULE_CONTROLS_DISABLE_MODULE_CLOCK_GATING_PE 0x00000004 +#define VIVS_PM_MODULE_CONTROLS_DISABLE_MODULE_CLOCK_GATING_SH 0x00000008 +#define VIVS_PM_MODULE_CONTROLS_DISABLE_MODULE_CLOCK_GATING_PA 0x00000010 +#define VIVS_PM_MODULE_CONTROLS_DISABLE_MODULE_CLOCK_GATING_SE 0x00000020 +#define VIVS_PM_MODULE_CONTROLS_DISABLE_MODULE_CLOCK_GATING_RA 0x00000040 +#define VIVS_PM_MODULE_CONTROLS_DISABLE_MODULE_CLOCK_GATING_TX 0x00000080 +#define VIVS_PM_MODULE_CONTROLS_DISABLE_MODULE_CLOCK_GATING_RA_EZ 0x00010000 +#define VIVS_PM_MODULE_CONTROLS_DISABLE_MODULE_CLOCK_GATING_RA_HZ 0x00020000 #define VIVS_PM_MODULE_STATUS 0x00000108 #define VIVS_PM_MODULE_STATUS_MODULE_CLOCK_GATED_FE 0x00000001 diff --git a/drivers/gpu/drm/exynos/Kconfig b/drivers/gpu/drm/exynos/Kconfig index d814b3048ee5..83f61c513b7e 100644 --- a/drivers/gpu/drm/exynos/Kconfig +++ b/drivers/gpu/drm/exynos/Kconfig @@ -2,10 +2,6 @@ config DRM_EXYNOS tristate "DRM Support for Samsung SoC EXYNOS Series" depends on OF && DRM && (ARCH_S3C64XX || ARCH_EXYNOS || ARCH_MULTIPLATFORM) select DRM_KMS_HELPER - select DRM_KMS_FB_HELPER - select FB_CFB_FILLRECT - select FB_CFB_COPYAREA - select FB_CFB_IMAGEBLIT select VIDEOMODE_HELPERS help Choose this option if you have a Samsung SoC EXYNOS chipset. @@ -15,7 +11,7 @@ if DRM_EXYNOS config DRM_EXYNOS_IOMMU bool - depends on EXYNOS_IOMMU && ARM_DMA_USE_IOMMU + depends on EXYNOS_IOMMU default y comment "CRTCs" diff --git a/drivers/gpu/drm/exynos/exynos7_drm_decon.c b/drivers/gpu/drm/exynos/exynos7_drm_decon.c index f6223f907c15..7f9901b7777b 100644 --- a/drivers/gpu/drm/exynos/exynos7_drm_decon.c +++ b/drivers/gpu/drm/exynos/exynos7_drm_decon.c @@ -31,7 +31,6 @@ #include "exynos_drm_plane.h" #include "exynos_drm_drv.h" #include "exynos_drm_fb.h" -#include "exynos_drm_fbdev.h" #include "exynos_drm_iommu.h" /* diff --git a/drivers/gpu/drm/exynos/exynos_dp.c b/drivers/gpu/drm/exynos/exynos_dp.c index 468498e3fec1..4f0850585b8e 100644 --- a/drivers/gpu/drm/exynos/exynos_dp.c +++ b/drivers/gpu/drm/exynos/exynos_dp.c @@ -34,7 +34,7 @@ struct exynos_dp_device { struct drm_encoder encoder; - struct drm_connector connector; + struct drm_connector *connector; struct drm_bridge *ptn_bridge; struct drm_device *drm_dev; struct device *dev; @@ -67,10 +67,10 @@ static int exynos_dp_poweroff(struct analogix_dp_plat_data *plat_data) return exynos_dp_crtc_clock_enable(plat_data, false); } -static int exynos_dp_get_modes(struct analogix_dp_plat_data *plat_data) +static int exynos_dp_get_modes(struct analogix_dp_plat_data *plat_data, + struct drm_connector *connector) { struct exynos_dp_device *dp = to_dp(plat_data); - struct drm_connector *connector = &dp->connector; struct drm_display_mode *mode; int num_modes = 0; @@ -103,6 +103,7 @@ static int exynos_dp_bridge_attach(struct analogix_dp_plat_data *plat_data, int ret; drm_connector_register(connector); + dp->connector = connector; /* Pre-empt DP connector creation if there's a bridge */ if (dp->ptn_bridge) { diff --git a/drivers/gpu/drm/exynos/exynos_drm_core.c b/drivers/gpu/drm/exynos/exynos_drm_core.c index 011211e4167d..edbd98ff293e 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_core.c +++ b/drivers/gpu/drm/exynos/exynos_drm_core.c @@ -15,7 +15,6 @@ #include <drm/drmP.h> #include "exynos_drm_drv.h" #include "exynos_drm_crtc.h" -#include "exynos_drm_fbdev.h" static LIST_HEAD(exynos_drm_subdrv_list); diff --git a/drivers/gpu/drm/exynos/exynos_drm_drv.c b/drivers/gpu/drm/exynos/exynos_drm_drv.c index 13d28d4229e2..877d2efa28e2 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_drv.c +++ b/drivers/gpu/drm/exynos/exynos_drm_drv.c @@ -159,12 +159,7 @@ static int exynos_drm_load(struct drm_device *dev, unsigned long flags) DRM_INFO("Exynos DRM: using %s device for DMA mapping operations\n", dev_name(private->dma_dev)); - /* - * create mapping to manage iommu table and set a pointer to iommu - * mapping structure to iommu_mapping of private data. - * also this iommu_mapping can be used to check if iommu is supported - * or not. - */ + /* create common IOMMU mapping for all devices attached to Exynos DRM */ ret = drm_create_iommu_mapping(dev); if (ret < 0) { DRM_ERROR("failed to create iommu mapping.\n"); diff --git a/drivers/gpu/drm/exynos/exynos_drm_drv.h b/drivers/gpu/drm/exynos/exynos_drm_drv.h index cc33ec9296e7..b39d521f093d 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_drv.h +++ b/drivers/gpu/drm/exynos/exynos_drm_drv.h @@ -224,8 +224,6 @@ struct exynos_drm_private { struct drm_property *plane_zpos_property; struct device *dma_dev; - unsigned long da_start; - unsigned long da_space_size; void *mapping; unsigned int pipe; diff --git a/drivers/gpu/drm/exynos/exynos_drm_fbdev.c b/drivers/gpu/drm/exynos/exynos_drm_fbdev.c index 67dcd6831291..fb49443bfd32 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_fbdev.c +++ b/drivers/gpu/drm/exynos/exynos_drm_fbdev.c @@ -269,8 +269,7 @@ static void exynos_drm_fbdev_destroy(struct drm_device *dev, struct exynos_drm_gem *exynos_gem = exynos_fbd->exynos_gem; struct drm_framebuffer *fb; - if (exynos_gem->kvaddr) - vunmap(exynos_gem->kvaddr); + vunmap(exynos_gem->kvaddr); /* release drm framebuffer and real buffer */ if (fb_helper->fb && fb_helper->fb->funcs) { diff --git a/drivers/gpu/drm/exynos/exynos_drm_fimd.c b/drivers/gpu/drm/exynos/exynos_drm_fimd.c index 3efe1aa89416..d47216488985 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_fimd.c +++ b/drivers/gpu/drm/exynos/exynos_drm_fimd.c @@ -30,7 +30,6 @@ #include "exynos_drm_drv.h" #include "exynos_drm_fb.h" -#include "exynos_drm_fbdev.h" #include "exynos_drm_crtc.h" #include "exynos_drm_plane.h" #include "exynos_drm_iommu.h" @@ -120,7 +119,6 @@ static struct fimd_driver_data s3c64xx_fimd_driver_data = { .timing_base = 0x0, .has_clksel = 1, .has_limited_fmt = 1, - .has_hw_trigger = 1, }; static struct fimd_driver_data exynos3_fimd_driver_data = { @@ -171,14 +169,11 @@ static struct fimd_driver_data exynos5420_fimd_driver_data = { .lcdblk_vt_shift = 24, .lcdblk_bypass_shift = 15, .lcdblk_mic_bypass_shift = 11, - .trg_type = I80_HW_TRG, .has_shadowcon = 1, .has_vidoutcon = 1, .has_vtsel = 1, .has_mic_bypass = 1, .has_dp_clk = 1, - .has_hw_trigger = 1, - .has_trigger_per_te = 1, }; struct fimd_context { diff --git a/drivers/gpu/drm/exynos/exynos_drm_g2d.c b/drivers/gpu/drm/exynos/exynos_drm_g2d.c index 493552368295..8564c3da0d22 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_g2d.c +++ b/drivers/gpu/drm/exynos/exynos_drm_g2d.c @@ -48,13 +48,13 @@ /* registers for base address */ #define G2D_SRC_BASE_ADDR 0x0304 -#define G2D_SRC_STRIDE_REG 0x0308 +#define G2D_SRC_STRIDE 0x0308 #define G2D_SRC_COLOR_MODE 0x030C #define G2D_SRC_LEFT_TOP 0x0310 #define G2D_SRC_RIGHT_BOTTOM 0x0314 #define G2D_SRC_PLANE2_BASE_ADDR 0x0318 #define G2D_DST_BASE_ADDR 0x0404 -#define G2D_DST_STRIDE_REG 0x0408 +#define G2D_DST_STRIDE 0x0408 #define G2D_DST_COLOR_MODE 0x040C #define G2D_DST_LEFT_TOP 0x0410 #define G2D_DST_RIGHT_BOTTOM 0x0414 @@ -563,7 +563,7 @@ static enum g2d_reg_type g2d_get_reg_type(int reg_offset) switch (reg_offset) { case G2D_SRC_BASE_ADDR: - case G2D_SRC_STRIDE_REG: + case G2D_SRC_STRIDE: case G2D_SRC_COLOR_MODE: case G2D_SRC_LEFT_TOP: case G2D_SRC_RIGHT_BOTTOM: @@ -573,7 +573,7 @@ static enum g2d_reg_type g2d_get_reg_type(int reg_offset) reg_type = REG_TYPE_SRC_PLANE2; break; case G2D_DST_BASE_ADDR: - case G2D_DST_STRIDE_REG: + case G2D_DST_STRIDE: case G2D_DST_COLOR_MODE: case G2D_DST_LEFT_TOP: case G2D_DST_RIGHT_BOTTOM: @@ -968,8 +968,8 @@ static int g2d_check_reg_offset(struct device *dev, } else buf_info->types[reg_type] = BUF_TYPE_GEM; break; - case G2D_SRC_STRIDE_REG: - case G2D_DST_STRIDE_REG: + case G2D_SRC_STRIDE: + case G2D_DST_STRIDE: if (for_addr) goto err; diff --git a/drivers/gpu/drm/exynos/exynos_drm_iommu.c b/drivers/gpu/drm/exynos/exynos_drm_iommu.c index 7ca09ee19656..0f373702414e 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_iommu.c +++ b/drivers/gpu/drm/exynos/exynos_drm_iommu.c @@ -14,13 +14,27 @@ #include <linux/dma-mapping.h> #include <linux/iommu.h> -#include <linux/kref.h> - -#include <asm/dma-iommu.h> #include "exynos_drm_drv.h" #include "exynos_drm_iommu.h" +static inline int configure_dma_max_seg_size(struct device *dev) +{ + if (!dev->dma_parms) + dev->dma_parms = kzalloc(sizeof(*dev->dma_parms), GFP_KERNEL); + if (!dev->dma_parms) + return -ENOMEM; + + dma_set_max_seg_size(dev, DMA_BIT_MASK(32)); + return 0; +} + +static inline void clear_dma_max_seg_size(struct device *dev) +{ + kfree(dev->dma_parms); + dev->dma_parms = NULL; +} + /* * drm_create_iommu_mapping - create a mapping structure * @@ -28,38 +42,22 @@ */ int drm_create_iommu_mapping(struct drm_device *drm_dev) { - struct dma_iommu_mapping *mapping = NULL; struct exynos_drm_private *priv = drm_dev->dev_private; - if (!priv->da_start) - priv->da_start = EXYNOS_DEV_ADDR_START; - if (!priv->da_space_size) - priv->da_space_size = EXYNOS_DEV_ADDR_SIZE; - - mapping = arm_iommu_create_mapping(&platform_bus_type, priv->da_start, - priv->da_space_size); - - if (IS_ERR(mapping)) - return PTR_ERR(mapping); - - priv->mapping = mapping; - - return 0; + return __exynos_iommu_create_mapping(priv, EXYNOS_DEV_ADDR_START, + EXYNOS_DEV_ADDR_SIZE); } /* * drm_release_iommu_mapping - release iommu mapping structure * * @drm_dev: DRM device - * - * if mapping->kref becomes 0 then all things related to iommu mapping - * will be released */ void drm_release_iommu_mapping(struct drm_device *drm_dev) { struct exynos_drm_private *priv = drm_dev->dev_private; - arm_iommu_release_mapping(priv->mapping); + __exynos_iommu_release_mapping(priv); } /* @@ -77,25 +75,19 @@ int drm_iommu_attach_device(struct drm_device *drm_dev, struct exynos_drm_private *priv = drm_dev->dev_private; int ret; - if (!priv->mapping) - return 0; - - subdrv_dev->dma_parms = devm_kzalloc(subdrv_dev, - sizeof(*subdrv_dev->dma_parms), - GFP_KERNEL); - if (!subdrv_dev->dma_parms) - return -ENOMEM; - - dma_set_max_seg_size(subdrv_dev, 0xffffffffu); - - if (subdrv_dev->archdata.mapping) - arm_iommu_detach_device(subdrv_dev); + if (get_dma_ops(priv->dma_dev) != get_dma_ops(subdrv_dev)) { + DRM_ERROR("Device %s lacks support for IOMMU\n", + dev_name(subdrv_dev)); + return -EINVAL; + } - ret = arm_iommu_attach_device(subdrv_dev, priv->mapping); - if (ret < 0) { - DRM_DEBUG_KMS("failed iommu attach.\n"); + ret = configure_dma_max_seg_size(subdrv_dev); + if (ret) return ret; - } + + ret = __exynos_iommu_attach(priv, subdrv_dev); + if (ret) + clear_dma_max_seg_size(subdrv_dev); return 0; } @@ -113,10 +105,7 @@ void drm_iommu_detach_device(struct drm_device *drm_dev, struct device *subdrv_dev) { struct exynos_drm_private *priv = drm_dev->dev_private; - struct dma_iommu_mapping *mapping = priv->mapping; - - if (!mapping || !mapping->domain) - return; - arm_iommu_detach_device(subdrv_dev); + __exynos_iommu_detach(priv, subdrv_dev); + clear_dma_max_seg_size(subdrv_dev); } diff --git a/drivers/gpu/drm/exynos/exynos_drm_iommu.h b/drivers/gpu/drm/exynos/exynos_drm_iommu.h index 5ffebe02ee4d..c8de4913fdbe 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_iommu.h +++ b/drivers/gpu/drm/exynos/exynos_drm_iommu.h @@ -17,6 +17,97 @@ #ifdef CONFIG_DRM_EXYNOS_IOMMU +#if defined(CONFIG_ARM_DMA_USE_IOMMU) +#include <asm/dma-iommu.h> + +static inline int __exynos_iommu_create_mapping(struct exynos_drm_private *priv, + unsigned long start, unsigned long size) +{ + priv->mapping = arm_iommu_create_mapping(&platform_bus_type, start, + size); + return IS_ERR(priv->mapping); +} + +static inline void +__exynos_iommu_release_mapping(struct exynos_drm_private *priv) +{ + arm_iommu_release_mapping(priv->mapping); +} + +static inline int __exynos_iommu_attach(struct exynos_drm_private *priv, + struct device *dev) +{ + if (dev->archdata.mapping) + arm_iommu_detach_device(dev); + + return arm_iommu_attach_device(dev, priv->mapping); +} + +static inline void __exynos_iommu_detach(struct exynos_drm_private *priv, + struct device *dev) +{ + arm_iommu_detach_device(dev); +} + +#elif defined(CONFIG_IOMMU_DMA) +#include <linux/dma-iommu.h> + +static inline int __exynos_iommu_create_mapping(struct exynos_drm_private *priv, + unsigned long start, unsigned long size) +{ + struct iommu_domain *domain; + int ret; + + domain = iommu_domain_alloc(priv->dma_dev->bus); + if (!domain) + return -ENOMEM; + + ret = iommu_get_dma_cookie(domain); + if (ret) + goto free_domain; + + ret = iommu_dma_init_domain(domain, start, size); + if (ret) + goto put_cookie; + + priv->mapping = domain; + return 0; + +put_cookie: + iommu_put_dma_cookie(domain); +free_domain: + iommu_domain_free(domain); + return ret; +} + +static inline void __exynos_iommu_release_mapping(struct exynos_drm_private *priv) +{ + struct iommu_domain *domain = priv->mapping; + + iommu_put_dma_cookie(domain); + iommu_domain_free(domain); + priv->mapping = NULL; +} + +static inline int __exynos_iommu_attach(struct exynos_drm_private *priv, + struct device *dev) +{ + struct iommu_domain *domain = priv->mapping; + + return iommu_attach_device(domain, dev); +} + +static inline void __exynos_iommu_detach(struct exynos_drm_private *priv, + struct device *dev) +{ + struct iommu_domain *domain = priv->mapping; + + iommu_detach_device(domain, dev); +} +#else +#error Unsupported architecture and IOMMU/DMA-mapping glue code +#endif + int drm_create_iommu_mapping(struct drm_device *drm_dev); void drm_release_iommu_mapping(struct drm_device *drm_dev); diff --git a/drivers/gpu/drm/exynos/exynos_hdmi.c b/drivers/gpu/drm/exynos/exynos_hdmi.c index 1625d7c8a319..2275efe41acd 100644 --- a/drivers/gpu/drm/exynos/exynos_hdmi.c +++ b/drivers/gpu/drm/exynos/exynos_hdmi.c @@ -1820,6 +1820,7 @@ static int hdmi_probe(struct platform_device *pdev) DRM_ERROR("Failed to find ddc node in device tree\n"); return -ENODEV; } + of_node_put(dev->of_node); out_get_ddc_adpt: hdata->ddc_adpt = of_find_i2c_adapter_by_node(ddc_node); @@ -1838,6 +1839,7 @@ out_get_ddc_adpt: ret = -ENODEV; goto err_ddc; } + of_node_put(dev->of_node); out_get_phy_port: if (hdata->drv_data->is_apb_phy) { diff --git a/drivers/gpu/drm/fsl-dcu/Kconfig b/drivers/gpu/drm/fsl-dcu/Kconfig index b9c714de6e40..14a72c4c496d 100644 --- a/drivers/gpu/drm/fsl-dcu/Kconfig +++ b/drivers/gpu/drm/fsl-dcu/Kconfig @@ -5,12 +5,7 @@ config DRM_FSL_DCU select BACKLIGHT_LCD_SUPPORT select DRM_KMS_HELPER select DRM_KMS_CMA_HELPER - select DRM_KMS_FB_HELPER select DRM_PANEL - select FB_SYS_FILLRECT - select FB_SYS_COPYAREA - select FB_SYS_IMAGEBLIT - select FB_SYS_FOPS select REGMAP_MMIO select VIDEOMODE_HELPERS help diff --git a/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_crtc.c b/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_crtc.c index 706de3278f1c..3371635cd4d7 100644 --- a/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_crtc.c +++ b/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_crtc.c @@ -44,6 +44,8 @@ static void fsl_dcu_drm_disable_crtc(struct drm_crtc *crtc) struct drm_device *dev = crtc->dev; struct fsl_dcu_drm_device *fsl_dev = dev->dev_private; + drm_crtc_vblank_off(crtc); + regmap_update_bits(fsl_dev->regmap, DCU_DCU_MODE, DCU_MODE_DCU_MODE_MASK, DCU_MODE_DCU_MODE(DCU_MODE_OFF)); @@ -61,6 +63,8 @@ static void fsl_dcu_drm_crtc_enable(struct drm_crtc *crtc) DCU_MODE_DCU_MODE(DCU_MODE_NORMAL)); regmap_write(fsl_dev->regmap, DCU_UPDATE_MODE, DCU_UPDATE_MODE_READREG); + + drm_crtc_vblank_on(crtc); } static void fsl_dcu_drm_crtc_mode_set_nofb(struct drm_crtc *crtc) @@ -137,9 +141,10 @@ int fsl_dcu_drm_crtc_create(struct fsl_dcu_drm_device *fsl_dev) { struct drm_plane *primary; struct drm_crtc *crtc = &fsl_dev->crtc; - unsigned int i, j, reg_num; int ret; + fsl_dcu_drm_init_planes(fsl_dev->drm); + primary = fsl_dcu_drm_primary_create_plane(fsl_dev->drm); if (!primary) return -ENOMEM; @@ -153,19 +158,5 @@ int fsl_dcu_drm_crtc_create(struct fsl_dcu_drm_device *fsl_dev) drm_crtc_helper_add(crtc, &fsl_dcu_drm_crtc_helper_funcs); - if (!strcmp(fsl_dev->soc->name, "ls1021a")) - reg_num = LS1021A_LAYER_REG_NUM; - else - reg_num = VF610_LAYER_REG_NUM; - for (i = 0; i < fsl_dev->soc->total_layer; i++) { - for (j = 1; j <= reg_num; j++) - regmap_write(fsl_dev->regmap, DCU_CTRLDESCLN(i, j), 0); - } - regmap_update_bits(fsl_dev->regmap, DCU_DCU_MODE, - DCU_MODE_DCU_MODE_MASK, - DCU_MODE_DCU_MODE(DCU_MODE_OFF)); - regmap_write(fsl_dev->regmap, DCU_UPDATE_MODE, - DCU_UPDATE_MODE_READREG); - return 0; } diff --git a/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_drv.c b/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_drv.c index 33727d5d826a..7882387f9bff 100644 --- a/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_drv.c +++ b/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_drv.c @@ -11,6 +11,7 @@ #include <linux/clk.h> #include <linux/clk-provider.h> +#include <linux/console.h> #include <linux/io.h> #include <linux/mfd/syscon.h> #include <linux/mm.h> @@ -22,6 +23,7 @@ #include <linux/regmap.h> #include <drm/drmP.h> +#include <drm/drm_atomic_helper.h> #include <drm/drm_crtc_helper.h> #include <drm/drm_fb_cma_helper.h> #include <drm/drm_gem_cma_helper.h> @@ -42,7 +44,6 @@ static const struct regmap_config fsl_dcu_regmap_config = { .reg_bits = 32, .reg_stride = 4, .val_bits = 32, - .cache_type = REGCACHE_RBTREE, .volatile_reg = fsl_dcu_drm_is_volatile_reg, }; @@ -228,11 +229,26 @@ static int fsl_dcu_drm_pm_suspend(struct device *dev) if (!fsl_dev) return 0; + disable_irq(fsl_dev->irq); drm_kms_helper_poll_disable(fsl_dev->drm); - regcache_cache_only(fsl_dev->regmap, true); - regcache_mark_dirty(fsl_dev->regmap); - clk_disable(fsl_dev->clk); - clk_unprepare(fsl_dev->clk); + + console_lock(); + drm_fbdev_cma_set_suspend(fsl_dev->fbdev, 1); + console_unlock(); + + fsl_dev->state = drm_atomic_helper_suspend(fsl_dev->drm); + if (IS_ERR(fsl_dev->state)) { + console_lock(); + drm_fbdev_cma_set_suspend(fsl_dev->fbdev, 0); + console_unlock(); + + drm_kms_helper_poll_enable(fsl_dev->drm); + enable_irq(fsl_dev->irq); + return PTR_ERR(fsl_dev->state); + } + + clk_disable_unprepare(fsl_dev->pix_clk); + clk_disable_unprepare(fsl_dev->clk); return 0; } @@ -245,21 +261,27 @@ static int fsl_dcu_drm_pm_resume(struct device *dev) if (!fsl_dev) return 0; - ret = clk_enable(fsl_dev->clk); + ret = clk_prepare_enable(fsl_dev->clk); if (ret < 0) { dev_err(dev, "failed to enable dcu clk\n"); - clk_unprepare(fsl_dev->clk); return ret; } - ret = clk_prepare(fsl_dev->clk); + + ret = clk_prepare_enable(fsl_dev->pix_clk); if (ret < 0) { - dev_err(dev, "failed to prepare dcu clk\n"); + dev_err(dev, "failed to enable pix clk\n"); return ret; } + fsl_dcu_drm_init_planes(fsl_dev->drm); + drm_atomic_helper_resume(fsl_dev->drm, fsl_dev->state); + + console_lock(); + drm_fbdev_cma_set_suspend(fsl_dev->fbdev, 0); + console_unlock(); + drm_kms_helper_poll_enable(fsl_dev->drm); - regcache_cache_only(fsl_dev->regmap, false); - regcache_sync(fsl_dev->regmap); + enable_irq(fsl_dev->irq); return 0; } @@ -273,12 +295,14 @@ static const struct fsl_dcu_soc_data fsl_dcu_ls1021a_data = { .name = "ls1021a", .total_layer = 16, .max_layer = 4, + .layer_regs = LS1021A_LAYER_REG_NUM, }; static const struct fsl_dcu_soc_data fsl_dcu_vf610_data = { .name = "vf610", .total_layer = 64, .max_layer = 6, + .layer_regs = VF610_LAYER_REG_NUM, }; static const struct of_device_id fsl_dcu_of_match[] = { diff --git a/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_drv.h b/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_drv.h index c275f900ff23..3b371fe7491e 100644 --- a/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_drv.h +++ b/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_drv.h @@ -175,6 +175,7 @@ struct fsl_dcu_soc_data { unsigned int total_layer; /*max layer number DCU supported*/ unsigned int max_layer; + unsigned int layer_regs; }; struct fsl_dcu_drm_device { @@ -193,6 +194,7 @@ struct fsl_dcu_drm_device { struct drm_encoder encoder; struct fsl_dcu_drm_connector connector; const struct fsl_dcu_soc_data *soc; + struct drm_atomic_state *state; }; void fsl_dcu_fbdev_init(struct drm_device *dev); diff --git a/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_kms.c b/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_kms.c index a6e4cd591960..d9d6cc1c8e39 100644 --- a/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_kms.c +++ b/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_kms.c @@ -43,7 +43,7 @@ int fsl_dcu_drm_modeset_init(struct fsl_dcu_drm_device *fsl_dev) if (ret) goto err; - ret = fsl_dcu_drm_connector_create(fsl_dev, &fsl_dev->encoder); + ret = fsl_dcu_create_outputs(fsl_dev); if (ret) goto err; diff --git a/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_output.h b/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_output.h index 7093109fbc21..5a7b88e19e44 100644 --- a/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_output.h +++ b/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_output.h @@ -25,9 +25,8 @@ to_fsl_dcu_connector(struct drm_connector *con) : NULL; } -int fsl_dcu_drm_connector_create(struct fsl_dcu_drm_device *fsl_dev, - struct drm_encoder *encoder); int fsl_dcu_drm_encoder_create(struct fsl_dcu_drm_device *fsl_dev, struct drm_crtc *crtc); +int fsl_dcu_create_outputs(struct fsl_dcu_drm_device *fsl_dev); #endif /* __FSL_DCU_DRM_CONNECTOR_H__ */ diff --git a/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_plane.c b/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_plane.c index 274558b3b32b..e50467a0deb0 100644 --- a/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_plane.c +++ b/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_plane.c @@ -217,6 +217,22 @@ static const u32 fsl_dcu_drm_plane_formats[] = { DRM_FORMAT_YUV422, }; +void fsl_dcu_drm_init_planes(struct drm_device *dev) +{ + struct fsl_dcu_drm_device *fsl_dev = dev->dev_private; + int i, j; + + for (i = 0; i < fsl_dev->soc->total_layer; i++) { + for (j = 1; j <= fsl_dev->soc->layer_regs; j++) + regmap_write(fsl_dev->regmap, DCU_CTRLDESCLN(i, j), 0); + } + regmap_update_bits(fsl_dev->regmap, DCU_DCU_MODE, + DCU_MODE_DCU_MODE_MASK, + DCU_MODE_DCU_MODE(DCU_MODE_OFF)); + regmap_write(fsl_dev->regmap, DCU_UPDATE_MODE, + DCU_UPDATE_MODE_READREG); +} + struct drm_plane *fsl_dcu_drm_primary_create_plane(struct drm_device *dev) { struct drm_plane *primary; diff --git a/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_plane.h b/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_plane.h index d657f088d859..8ee45f813ee8 100644 --- a/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_plane.h +++ b/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_plane.h @@ -12,6 +12,7 @@ #ifndef __FSL_DCU_DRM_PLANE_H__ #define __FSL_DCU_DRM_PLANE_H__ +void fsl_dcu_drm_init_planes(struct drm_device *dev); struct drm_plane *fsl_dcu_drm_primary_create_plane(struct drm_device *dev); #endif /* __FSL_DCU_DRM_PLANE_H__ */ diff --git a/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_rgb.c b/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_rgb.c index 0b0989e503ea..26edcc899712 100644 --- a/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_rgb.c +++ b/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_rgb.c @@ -10,6 +10,7 @@ */ #include <linux/backlight.h> +#include <linux/of_graph.h> #include <drm/drmP.h> #include <drm/drm_atomic_helper.h> @@ -132,12 +133,12 @@ static const struct drm_connector_helper_funcs connector_helper_funcs = { .mode_valid = fsl_dcu_drm_connector_mode_valid, }; -int fsl_dcu_drm_connector_create(struct fsl_dcu_drm_device *fsl_dev, - struct drm_encoder *encoder) +static int fsl_dcu_attach_panel(struct fsl_dcu_drm_device *fsl_dev, + struct drm_panel *panel) { + struct drm_encoder *encoder = &fsl_dev->encoder; struct drm_connector *connector = &fsl_dev->connector.base; struct drm_mode_config *mode_config = &fsl_dev->drm->mode_config; - struct device_node *panel_node; int ret; fsl_dev->connector.encoder = encoder; @@ -161,21 +162,7 @@ int fsl_dcu_drm_connector_create(struct fsl_dcu_drm_device *fsl_dev, mode_config->dpms_property, DRM_MODE_DPMS_OFF); - panel_node = of_parse_phandle(fsl_dev->np, "fsl,panel", 0); - if (!panel_node) { - dev_err(fsl_dev->dev, "fsl,panel property not found\n"); - ret = -ENODEV; - goto err_sysfs; - } - - fsl_dev->connector.panel = of_drm_find_panel(panel_node); - if (!fsl_dev->connector.panel) { - ret = -EPROBE_DEFER; - goto err_panel; - } - of_node_put(panel_node); - - ret = drm_panel_attach(fsl_dev->connector.panel, connector); + ret = drm_panel_attach(panel, connector); if (ret) { dev_err(fsl_dev->dev, "failed to attach panel\n"); goto err_sysfs; @@ -183,11 +170,62 @@ int fsl_dcu_drm_connector_create(struct fsl_dcu_drm_device *fsl_dev, return 0; -err_panel: - of_node_put(panel_node); err_sysfs: drm_connector_unregister(connector); err_cleanup: drm_connector_cleanup(connector); return ret; } + +static int fsl_dcu_attach_endpoint(struct fsl_dcu_drm_device *fsl_dev, + const struct of_endpoint *ep) +{ + struct drm_bridge *bridge; + struct device_node *np; + + np = of_graph_get_remote_port_parent(ep->local_node); + + fsl_dev->connector.panel = of_drm_find_panel(np); + if (fsl_dev->connector.panel) { + of_node_put(np); + return fsl_dcu_attach_panel(fsl_dev, fsl_dev->connector.panel); + } + + bridge = of_drm_find_bridge(np); + of_node_put(np); + if (!bridge) + return -ENODEV; + + fsl_dev->encoder.bridge = bridge; + bridge->encoder = &fsl_dev->encoder; + + return drm_bridge_attach(fsl_dev->drm, bridge); +} + +int fsl_dcu_create_outputs(struct fsl_dcu_drm_device *fsl_dev) +{ + struct of_endpoint ep; + struct device_node *ep_node, *panel_node; + int ret; + + /* This is for backward compatibility */ + panel_node = of_parse_phandle(fsl_dev->np, "fsl,panel", 0); + if (panel_node) { + fsl_dev->connector.panel = of_drm_find_panel(panel_node); + of_node_put(panel_node); + if (!fsl_dev->connector.panel) + return -EPROBE_DEFER; + return fsl_dcu_attach_panel(fsl_dev, fsl_dev->connector.panel); + } + + ep_node = of_graph_get_next_endpoint(fsl_dev->np, NULL); + if (!ep_node) + return -ENODEV; + + ret = of_graph_parse_endpoint(ep_node, &ep); + of_node_put(ep_node); + if (ret) + return -ENODEV; + + return fsl_dcu_attach_endpoint(fsl_dev, &ep); +} diff --git a/drivers/gpu/drm/fsl-dcu/fsl_tcon.c b/drivers/gpu/drm/fsl-dcu/fsl_tcon.c index bbe34f1c0505..bca09ea24632 100644 --- a/drivers/gpu/drm/fsl-dcu/fsl_tcon.c +++ b/drivers/gpu/drm/fsl-dcu/fsl_tcon.c @@ -92,6 +92,7 @@ struct fsl_tcon *fsl_tcon_init(struct device *dev) goto err_node_put; } + of_node_put(np); clk_prepare_enable(tcon->ipg_clk); dev_info(dev, "Using TCON in bypass mode\n"); diff --git a/drivers/gpu/drm/gma500/Kconfig b/drivers/gpu/drm/gma500/Kconfig index 17f928ec84ea..8906d67494fc 100644 --- a/drivers/gpu/drm/gma500/Kconfig +++ b/drivers/gpu/drm/gma500/Kconfig @@ -1,11 +1,7 @@ config DRM_GMA500 tristate "Intel GMA5/600 KMS Framebuffer" depends on DRM && PCI && X86 - select FB_CFB_COPYAREA - select FB_CFB_FILLRECT - select FB_CFB_IMAGEBLIT select DRM_KMS_HELPER - select DRM_KMS_FB_HELPER select DRM_TTM # GMA500 depends on ACPI_VIDEO when ACPI is enabled, just like i915 select ACPI_VIDEO if ACPI diff --git a/drivers/gpu/drm/gma500/cdv_intel_hdmi.c b/drivers/gpu/drm/gma500/cdv_intel_hdmi.c index 28f9d90988ff..563f193fcfac 100644 --- a/drivers/gpu/drm/gma500/cdv_intel_hdmi.c +++ b/drivers/gpu/drm/gma500/cdv_intel_hdmi.c @@ -246,8 +246,7 @@ static void cdv_hdmi_destroy(struct drm_connector *connector) { struct gma_encoder *gma_encoder = gma_attached_encoder(connector); - if (gma_encoder->i2c_bus) - psb_intel_i2c_destroy(gma_encoder->i2c_bus); + psb_intel_i2c_destroy(gma_encoder->i2c_bus); drm_connector_unregister(connector); drm_connector_cleanup(connector); kfree(connector); diff --git a/drivers/gpu/drm/gma500/cdv_intel_lvds.c b/drivers/gpu/drm/gma500/cdv_intel_lvds.c index 813ef23a8054..38dc89083148 100644 --- a/drivers/gpu/drm/gma500/cdv_intel_lvds.c +++ b/drivers/gpu/drm/gma500/cdv_intel_lvds.c @@ -444,8 +444,7 @@ static void cdv_intel_lvds_destroy(struct drm_connector *connector) { struct gma_encoder *gma_encoder = gma_attached_encoder(connector); - if (gma_encoder->i2c_bus) - psb_intel_i2c_destroy(gma_encoder->i2c_bus); + psb_intel_i2c_destroy(gma_encoder->i2c_bus); drm_connector_unregister(connector); drm_connector_cleanup(connector); kfree(connector); @@ -780,12 +779,10 @@ out: failed_find: mutex_unlock(&dev->mode_config.mutex); printk(KERN_ERR "Failed find\n"); - if (gma_encoder->ddc_bus) - psb_intel_i2c_destroy(gma_encoder->ddc_bus); + psb_intel_i2c_destroy(gma_encoder->ddc_bus); failed_ddc: printk(KERN_ERR "Failed DDC\n"); - if (gma_encoder->i2c_bus) - psb_intel_i2c_destroy(gma_encoder->i2c_bus); + psb_intel_i2c_destroy(gma_encoder->i2c_bus); failed_blc_i2c: printk(KERN_ERR "Failed BLC\n"); drm_encoder_cleanup(encoder); diff --git a/drivers/gpu/drm/gma500/framebuffer.c b/drivers/gpu/drm/gma500/framebuffer.c index 7440bf90ac9c..0fcdce0817de 100644 --- a/drivers/gpu/drm/gma500/framebuffer.c +++ b/drivers/gpu/drm/gma500/framebuffer.c @@ -184,12 +184,6 @@ static int psbfb_mmap(struct fb_info *info, struct vm_area_struct *vma) return 0; } -static int psbfb_ioctl(struct fb_info *info, unsigned int cmd, - unsigned long arg) -{ - return -ENOTTY; -} - static struct fb_ops psbfb_ops = { .owner = THIS_MODULE, .fb_check_var = drm_fb_helper_check_var, @@ -201,7 +195,6 @@ static struct fb_ops psbfb_ops = { .fb_imageblit = drm_fb_helper_cfb_imageblit, .fb_mmap = psbfb_mmap, .fb_sync = psbfb_sync, - .fb_ioctl = psbfb_ioctl, }; static struct fb_ops psbfb_roll_ops = { @@ -215,7 +208,6 @@ static struct fb_ops psbfb_roll_ops = { .fb_imageblit = drm_fb_helper_cfb_imageblit, .fb_pan_display = psbfb_pan, .fb_mmap = psbfb_mmap, - .fb_ioctl = psbfb_ioctl, }; static struct fb_ops psbfb_unaccel_ops = { @@ -228,7 +220,6 @@ static struct fb_ops psbfb_unaccel_ops = { .fb_copyarea = drm_fb_helper_cfb_copyarea, .fb_imageblit = drm_fb_helper_cfb_imageblit, .fb_mmap = psbfb_mmap, - .fb_ioctl = psbfb_ioctl, }; /** diff --git a/drivers/gpu/drm/gma500/psb_drv.c b/drivers/gpu/drm/gma500/psb_drv.c index 82b8ce418b27..50eb944fb78a 100644 --- a/drivers/gpu/drm/gma500/psb_drv.c +++ b/drivers/gpu/drm/gma500/psb_drv.c @@ -210,10 +210,8 @@ static int psb_driver_unload(struct drm_device *dev) iounmap(dev_priv->aux_reg); dev_priv->aux_reg = NULL; } - if (dev_priv->aux_pdev) - pci_dev_put(dev_priv->aux_pdev); - if (dev_priv->lpc_pdev) - pci_dev_put(dev_priv->lpc_pdev); + pci_dev_put(dev_priv->aux_pdev); + pci_dev_put(dev_priv->lpc_pdev); /* Destroy VBT data */ psb_intel_destroy_bios(dev); diff --git a/drivers/gpu/drm/gma500/psb_intel_lvds.c b/drivers/gpu/drm/gma500/psb_intel_lvds.c index b1b93317d054..e55733ca46d2 100644 --- a/drivers/gpu/drm/gma500/psb_intel_lvds.c +++ b/drivers/gpu/drm/gma500/psb_intel_lvds.c @@ -561,8 +561,7 @@ void psb_intel_lvds_destroy(struct drm_connector *connector) struct gma_encoder *gma_encoder = gma_attached_encoder(connector); struct psb_intel_lvds_priv *lvds_priv = gma_encoder->dev_priv; - if (lvds_priv->ddc_bus) - psb_intel_i2c_destroy(lvds_priv->ddc_bus); + psb_intel_i2c_destroy(lvds_priv->ddc_bus); drm_connector_unregister(connector); drm_connector_cleanup(connector); kfree(connector); @@ -835,11 +834,9 @@ out: failed_find: mutex_unlock(&dev->mode_config.mutex); - if (lvds_priv->ddc_bus) - psb_intel_i2c_destroy(lvds_priv->ddc_bus); + psb_intel_i2c_destroy(lvds_priv->ddc_bus); failed_ddc: - if (lvds_priv->i2c_bus) - psb_intel_i2c_destroy(lvds_priv->i2c_bus); + psb_intel_i2c_destroy(lvds_priv->i2c_bus); failed_blc_i2c: drm_encoder_cleanup(encoder); drm_connector_cleanup(connector); diff --git a/drivers/gpu/drm/hisilicon/kirin/Kconfig b/drivers/gpu/drm/hisilicon/kirin/Kconfig index ea0df6115f7e..499f64405dac 100644 --- a/drivers/gpu/drm/hisilicon/kirin/Kconfig +++ b/drivers/gpu/drm/hisilicon/kirin/Kconfig @@ -4,6 +4,7 @@ config DRM_HISI_KIRIN select DRM_KMS_HELPER select DRM_GEM_CMA_HELPER select DRM_KMS_CMA_HELPER + select HISI_KIRIN_DW_DSI help Choose this option if you have a hisilicon Kirin chipsets(hi6220). If M is selected the module will be called kirin-drm. diff --git a/drivers/gpu/drm/hisilicon/kirin/kirin_drm_ade.c b/drivers/gpu/drm/hisilicon/kirin/kirin_drm_ade.c index ed76baad525f..c3707d47cd89 100644 --- a/drivers/gpu/drm/hisilicon/kirin/kirin_drm_ade.c +++ b/drivers/gpu/drm/hisilicon/kirin/kirin_drm_ade.c @@ -487,6 +487,7 @@ static void ade_crtc_enable(struct drm_crtc *crtc) ade_set_medianoc_qos(acrtc); ade_display_enable(acrtc); ade_dump_regs(ctx->base); + drm_crtc_vblank_on(crtc); acrtc->enable = true; } @@ -498,6 +499,7 @@ static void ade_crtc_disable(struct drm_crtc *crtc) if (!acrtc->enable) return; + drm_crtc_vblank_off(crtc); ade_power_down(ctx); acrtc->enable = false; } @@ -965,21 +967,21 @@ static int ade_dts_parse(struct platform_device *pdev, struct ade_hw_ctx *ctx) } ctx->ade_core_clk = devm_clk_get(dev, "clk_ade_core"); - if (!ctx->ade_core_clk) { + if (IS_ERR(ctx->ade_core_clk)) { DRM_ERROR("failed to parse clk ADE_CORE\n"); - return -ENODEV; + return PTR_ERR(ctx->ade_core_clk); } ctx->media_noc_clk = devm_clk_get(dev, "clk_codec_jpeg"); - if (!ctx->media_noc_clk) { + if (IS_ERR(ctx->media_noc_clk)) { DRM_ERROR("failed to parse clk CODEC_JPEG\n"); - return -ENODEV; + return PTR_ERR(ctx->media_noc_clk); } ctx->ade_pix_clk = devm_clk_get(dev, "clk_ade_pix"); - if (!ctx->ade_pix_clk) { + if (IS_ERR(ctx->ade_pix_clk)) { DRM_ERROR("failed to parse clk ADE_PIX\n"); - return -ENODEV; + return PTR_ERR(ctx->ade_pix_clk); } return 0; diff --git a/drivers/gpu/drm/i2c/Kconfig b/drivers/gpu/drm/i2c/Kconfig index 22c7ed63a001..4d341db462a2 100644 --- a/drivers/gpu/drm/i2c/Kconfig +++ b/drivers/gpu/drm/i2c/Kconfig @@ -1,12 +1,6 @@ menu "I2C encoder or helper chips" depends on DRM && DRM_KMS_HELPER && I2C -config DRM_I2C_ADV7511 - tristate "AV7511 encoder" - select REGMAP_I2C - help - Support for the Analog Device ADV7511(W) and ADV7513 HDMI encoders. - config DRM_I2C_CH7006 tristate "Chrontel ch7006 TV encoder" default m if DRM_NOUVEAU diff --git a/drivers/gpu/drm/i2c/Makefile b/drivers/gpu/drm/i2c/Makefile index 2c72eb584ab7..43aa33baebed 100644 --- a/drivers/gpu/drm/i2c/Makefile +++ b/drivers/gpu/drm/i2c/Makefile @@ -1,7 +1,5 @@ ccflags-y := -Iinclude/drm -obj-$(CONFIG_DRM_I2C_ADV7511) += adv7511.o - ch7006-y := ch7006_drv.o ch7006_mode.o obj-$(CONFIG_DRM_I2C_CH7006) += ch7006.o diff --git a/drivers/gpu/drm/i2c/ch7006_drv.c b/drivers/gpu/drm/i2c/ch7006_drv.c index 0594c45f7164..e9e8ae2ec06b 100644 --- a/drivers/gpu/drm/i2c/ch7006_drv.c +++ b/drivers/gpu/drm/i2c/ch7006_drv.c @@ -361,13 +361,8 @@ static int ch7006_encoder_set_property(struct drm_encoder *encoder, /* Disable the crtc to ensure a full modeset is * performed whenever it's turned on again. */ - if (crtc) { - struct drm_mode_set modeset = { - .crtc = crtc, - }; - - drm_mode_set_config_internal(&modeset); - } + if (crtc) + drm_crtc_force_disable(crtc); } return 0; diff --git a/drivers/gpu/drm/i915/intel_csr.c b/drivers/gpu/drm/i915/intel_csr.c index 1ea0e1f43397..fb27d187876c 100644 --- a/drivers/gpu/drm/i915/intel_csr.c +++ b/drivers/gpu/drm/i915/intel_csr.c @@ -34,15 +34,15 @@ * low-power state and comes back to normal. */ -#define I915_CSR_KBL "i915/kbl_dmc_ver1_01.bin" +#define I915_CSR_KBL "i915/kbl_dmc_ver1.bin" MODULE_FIRMWARE(I915_CSR_KBL); #define KBL_CSR_VERSION_REQUIRED CSR_VERSION(1, 1) -#define I915_CSR_SKL "i915/skl_dmc_ver1_26.bin" +#define I915_CSR_SKL "i915/skl_dmc_ver1.bin" MODULE_FIRMWARE(I915_CSR_SKL); -#define SKL_CSR_VERSION_REQUIRED CSR_VERSION(1, 26) +#define SKL_CSR_VERSION_REQUIRED CSR_VERSION(1, 23) -#define I915_CSR_BXT "i915/bxt_dmc_ver1_07.bin" +#define I915_CSR_BXT "i915/bxt_dmc_ver1.bin" MODULE_FIRMWARE(I915_CSR_BXT); #define BXT_CSR_VERSION_REQUIRED CSR_VERSION(1, 7) diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c index 361977fd1170..a07da548ff49 100644 --- a/drivers/gpu/drm/i915/intel_lrc.c +++ b/drivers/gpu/drm/i915/intel_lrc.c @@ -928,6 +928,7 @@ static inline int gen8_emit_flush_coherentl3_wa(struct intel_engine_cs *engine, uint32_t *batch, uint32_t index) { + struct drm_i915_private *dev_priv = engine->i915; uint32_t l3sqc4_flush = (0x40400000 | GEN8_LQSC_FLUSH_COHERENT_LINES); /* @@ -936,8 +937,8 @@ static inline int gen8_emit_flush_coherentl3_wa(struct intel_engine_cs *engine, * this batch updates GEN8_L3SQCREG4 with default value we need to * set this bit here to retain the WA during flush. */ - if (IS_SKL_REVID(engine->i915, 0, SKL_REVID_E0) || - IS_KBL_REVID(engine->i915, 0, KBL_REVID_E0)) + if (IS_SKL_REVID(dev_priv, 0, SKL_REVID_E0) || + IS_KBL_REVID(dev_priv, 0, KBL_REVID_E0)) l3sqc4_flush |= GEN8_LQSC_RO_PERF_DIS; wa_ctx_emit(batch, index, (MI_STORE_REGISTER_MEM_GEN8 | @@ -1077,11 +1078,12 @@ static int gen9_init_indirectctx_bb(struct intel_engine_cs *engine, uint32_t *offset) { int ret; + struct drm_i915_private *dev_priv = engine->i915; uint32_t index = wa_ctx_start(wa_ctx, *offset, CACHELINE_DWORDS); /* WaDisableCtxRestoreArbitration:skl,bxt */ - if (IS_SKL_REVID(engine->i915, 0, SKL_REVID_D0) || - IS_BXT_REVID(engine->i915, 0, BXT_REVID_A1)) + if (IS_SKL_REVID(dev_priv, 0, SKL_REVID_D0) || + IS_BXT_REVID(dev_priv, 0, BXT_REVID_A1)) wa_ctx_emit(batch, index, MI_ARB_ON_OFF | MI_ARB_DISABLE); /* WaFlushCoherentL3CacheLinesAtContextSwitch:skl,bxt */ @@ -1099,7 +1101,7 @@ static int gen9_init_indirectctx_bb(struct intel_engine_cs *engine, /* WaClearSlmSpaceAtContextSwitch:kbl */ /* Actual scratch location is at 128 bytes offset */ - if (IS_KBL_REVID(engine->i915, 0, KBL_REVID_A0)) { + if (IS_KBL_REVID(dev_priv, 0, KBL_REVID_A0)) { uint32_t scratch_addr = engine->scratch.gtt_offset + 2*CACHELINE_BYTES; diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index 948f24418eac..6bd352a8f30e 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -57,7 +57,7 @@ static void gen9_init_clock_gating(struct drm_device *dev) { - struct drm_i915_private *dev_priv = to_i915(dev); + struct drm_i915_private *dev_priv = dev->dev_private; /* See Bspec note for PSR2_CTL bit 31, Wa#828:skl,bxt,kbl */ I915_WRITE(CHICKEN_PAR1_1, @@ -7044,7 +7044,7 @@ static void gen8_set_l3sqc_credits(struct drm_i915_private *dev_priv, static void kabylake_init_clock_gating(struct drm_device *dev) { - struct drm_i915_private *dev_priv = to_i915(dev); + struct drm_i915_private *dev_priv = dev->dev_private; gen9_init_clock_gating(dev); @@ -7065,7 +7065,7 @@ static void kabylake_init_clock_gating(struct drm_device *dev) static void skylake_init_clock_gating(struct drm_device *dev) { - struct drm_i915_private *dev_priv = to_i915(dev); + struct drm_i915_private *dev_priv = dev->dev_private; gen9_init_clock_gating(dev); diff --git a/drivers/gpu/drm/imx/Kconfig b/drivers/gpu/drm/imx/Kconfig index a1844b50546c..f2c9ae822149 100644 --- a/drivers/gpu/drm/imx/Kconfig +++ b/drivers/gpu/drm/imx/Kconfig @@ -1,7 +1,6 @@ config DRM_IMX tristate "DRM Support for Freescale i.MX" select DRM_KMS_HELPER - select DRM_KMS_FB_HELPER select VIDEOMODE_HELPERS select DRM_GEM_CMA_HELPER select DRM_KMS_CMA_HELPER diff --git a/drivers/gpu/drm/imx/dw_hdmi-imx.c b/drivers/gpu/drm/imx/dw_hdmi-imx.c index a24631fdf4ad..359cd2765552 100644 --- a/drivers/gpu/drm/imx/dw_hdmi-imx.c +++ b/drivers/gpu/drm/imx/dw_hdmi-imx.c @@ -28,6 +28,11 @@ struct imx_hdmi { struct regmap *regmap; }; +static inline struct imx_hdmi *enc_to_imx_hdmi(struct drm_encoder *e) +{ + return container_of(e, struct imx_hdmi, encoder); +} + static const struct dw_hdmi_mpll_config imx_mpll_cfg[] = { { 45250000, { @@ -109,15 +114,9 @@ static void dw_hdmi_imx_encoder_disable(struct drm_encoder *encoder) { } -static void dw_hdmi_imx_encoder_mode_set(struct drm_encoder *encoder, - struct drm_display_mode *mode, - struct drm_display_mode *adj_mode) +static void dw_hdmi_imx_encoder_enable(struct drm_encoder *encoder) { -} - -static void dw_hdmi_imx_encoder_commit(struct drm_encoder *encoder) -{ - struct imx_hdmi *hdmi = container_of(encoder, struct imx_hdmi, encoder); + struct imx_hdmi *hdmi = enc_to_imx_hdmi(encoder); int mux = drm_of_encoder_active_port_id(hdmi->dev->of_node, encoder); regmap_update_bits(hdmi->regmap, IOMUXC_GPR3, @@ -125,16 +124,23 @@ static void dw_hdmi_imx_encoder_commit(struct drm_encoder *encoder) mux << IMX6Q_GPR3_HDMI_MUX_CTL_SHIFT); } -static void dw_hdmi_imx_encoder_prepare(struct drm_encoder *encoder) +static int dw_hdmi_imx_atomic_check(struct drm_encoder *encoder, + struct drm_crtc_state *crtc_state, + struct drm_connector_state *conn_state) { - imx_drm_set_bus_format(encoder, MEDIA_BUS_FMT_RGB888_1X24); + struct imx_crtc_state *imx_crtc_state = to_imx_crtc_state(crtc_state); + + imx_crtc_state->bus_format = MEDIA_BUS_FMT_RGB888_1X24; + imx_crtc_state->di_hsync_pin = 2; + imx_crtc_state->di_vsync_pin = 3; + + return 0; } static const struct drm_encoder_helper_funcs dw_hdmi_imx_encoder_helper_funcs = { - .mode_set = dw_hdmi_imx_encoder_mode_set, - .prepare = dw_hdmi_imx_encoder_prepare, - .commit = dw_hdmi_imx_encoder_commit, + .enable = dw_hdmi_imx_encoder_enable, .disable = dw_hdmi_imx_encoder_disable, + .atomic_check = dw_hdmi_imx_atomic_check, }; static const struct drm_encoder_funcs dw_hdmi_imx_encoder_funcs = { diff --git a/drivers/gpu/drm/imx/imx-drm-core.c b/drivers/gpu/drm/imx/imx-drm-core.c index 7746418a4c08..9f7dafce3a4c 100644 --- a/drivers/gpu/drm/imx/imx-drm-core.c +++ b/drivers/gpu/drm/imx/imx-drm-core.c @@ -15,10 +15,14 @@ */ #include <linux/component.h> #include <linux/device.h> +#include <linux/dma-buf.h> #include <linux/fb.h> #include <linux/module.h> #include <linux/platform_device.h> +#include <linux/reservation.h> #include <drm/drmP.h> +#include <drm/drm_atomic.h> +#include <drm/drm_atomic_helper.h> #include <drm/drm_fb_helper.h> #include <drm/drm_crtc_helper.h> #include <drm/drm_gem_cma_helper.h> @@ -41,6 +45,7 @@ struct imx_drm_device { struct imx_drm_crtc *crtc[MAX_CRTC]; unsigned int pipes; struct drm_fbdev_cma *fbhelper; + struct drm_atomic_state *state; }; struct imx_drm_crtc { @@ -85,45 +90,6 @@ static int imx_drm_driver_unload(struct drm_device *drm) return 0; } -static struct imx_drm_crtc *imx_drm_find_crtc(struct drm_crtc *crtc) -{ - struct imx_drm_device *imxdrm = crtc->dev->dev_private; - unsigned i; - - for (i = 0; i < MAX_CRTC; i++) - if (imxdrm->crtc[i] && imxdrm->crtc[i]->crtc == crtc) - return imxdrm->crtc[i]; - - return NULL; -} - -int imx_drm_set_bus_config(struct drm_encoder *encoder, u32 bus_format, - int hsync_pin, int vsync_pin, u32 bus_flags) -{ - struct imx_drm_crtc_helper_funcs *helper; - struct imx_drm_crtc *imx_crtc; - - imx_crtc = imx_drm_find_crtc(encoder->crtc); - if (!imx_crtc) - return -EINVAL; - - helper = &imx_crtc->imx_drm_helper_funcs; - if (helper->set_interface_pix_fmt) - return helper->set_interface_pix_fmt(encoder->crtc, - bus_format, hsync_pin, vsync_pin, - bus_flags); - return 0; -} -EXPORT_SYMBOL_GPL(imx_drm_set_bus_config); - -int imx_drm_set_bus_format(struct drm_encoder *encoder, u32 bus_format) -{ - return imx_drm_set_bus_config(encoder, bus_format, 2, 3, - DRM_BUS_FLAG_DE_HIGH | - DRM_BUS_FLAG_PIXDATA_NEGEDGE); -} -EXPORT_SYMBOL_GPL(imx_drm_set_bus_format); - int imx_drm_crtc_vblank_get(struct imx_drm_crtc *imx_drm_crtc) { return drm_crtc_vblank_get(imx_drm_crtc->crtc); @@ -208,6 +174,63 @@ static void imx_drm_output_poll_changed(struct drm_device *drm) static const struct drm_mode_config_funcs imx_drm_mode_config_funcs = { .fb_create = drm_fb_cma_create, .output_poll_changed = imx_drm_output_poll_changed, + .atomic_check = drm_atomic_helper_check, + .atomic_commit = drm_atomic_helper_commit, +}; + +static void imx_drm_atomic_commit_tail(struct drm_atomic_state *state) +{ + struct drm_device *dev = state->dev; + struct drm_crtc *crtc; + struct drm_crtc_state *crtc_state; + struct drm_plane_state *plane_state; + struct drm_gem_cma_object *cma_obj; + struct fence *excl; + unsigned shared_count; + struct fence **shared; + unsigned int i, j; + int ret; + + /* Wait for fences. */ + for_each_crtc_in_state(state, crtc, crtc_state, i) { + plane_state = crtc->primary->state; + if (plane_state->fb) { + cma_obj = drm_fb_cma_get_gem_obj(plane_state->fb, 0); + if (cma_obj->base.dma_buf) { + ret = reservation_object_get_fences_rcu( + cma_obj->base.dma_buf->resv, &excl, + &shared_count, &shared); + if (unlikely(ret)) + DRM_ERROR("failed to get fences " + "for buffer\n"); + + if (excl) { + fence_wait(excl, false); + fence_put(excl); + } + for (j = 0; j < shared_count; i++) { + fence_wait(shared[j], false); + fence_put(shared[j]); + } + } + } + } + + drm_atomic_helper_commit_modeset_disables(dev, state); + + drm_atomic_helper_commit_planes(dev, state, true); + + drm_atomic_helper_commit_modeset_enables(dev, state); + + drm_atomic_helper_commit_hw_done(state); + + drm_atomic_helper_wait_for_vblanks(dev, state); + + drm_atomic_helper_cleanup_planes(dev, state); +} + +static struct drm_mode_config_helper_funcs imx_drm_mode_config_helpers = { + .atomic_commit_tail = imx_drm_atomic_commit_tail, }; /* @@ -249,6 +272,7 @@ static int imx_drm_driver_load(struct drm_device *drm, unsigned long flags) drm->mode_config.max_width = 4096; drm->mode_config.max_height = 4096; drm->mode_config.funcs = &imx_drm_mode_config_funcs; + drm->mode_config.helper_private = &imx_drm_mode_config_helpers; drm_mode_config_init(drm); @@ -279,6 +303,8 @@ static int imx_drm_driver_load(struct drm_device *drm, unsigned long flags) } } + drm_mode_config_reset(drm); + /* * All components are now initialised, so setup the fb helper. * The fb helper takes copies of key hardware information, so the @@ -289,7 +315,6 @@ static int imx_drm_driver_load(struct drm_device *drm, unsigned long flags) dev_warn(drm->dev, "Invalid legacyfb_depth. Defaulting to 16bpp\n"); legacyfb_depth = 16; } - drm_helper_disable_unused_functions(drm); imxdrm->fbhelper = drm_fbdev_cma_init(drm, legacyfb_depth, drm->mode_config.num_crtc, MAX_CRTC); if (IS_ERR(imxdrm->fbhelper)) { @@ -403,7 +428,8 @@ static const struct drm_ioctl_desc imx_drm_ioctls[] = { }; static struct drm_driver imx_drm_driver = { - .driver_features = DRIVER_MODESET | DRIVER_GEM | DRIVER_PRIME, + .driver_features = DRIVER_MODESET | DRIVER_GEM | DRIVER_PRIME | + DRIVER_ATOMIC, .load = imx_drm_driver_load, .unload = imx_drm_driver_unload, .lastclose = imx_drm_driver_lastclose, @@ -491,6 +517,7 @@ static int imx_drm_platform_remove(struct platform_device *pdev) static int imx_drm_suspend(struct device *dev) { struct drm_device *drm_dev = dev_get_drvdata(dev); + struct imx_drm_device *imxdrm; /* The drm_dev is NULL before .load hook is called */ if (drm_dev == NULL) @@ -498,17 +525,26 @@ static int imx_drm_suspend(struct device *dev) drm_kms_helper_poll_disable(drm_dev); + imxdrm = drm_dev->dev_private; + imxdrm->state = drm_atomic_helper_suspend(drm_dev); + if (IS_ERR(imxdrm->state)) { + drm_kms_helper_poll_enable(drm_dev); + return PTR_ERR(imxdrm->state); + } + return 0; } static int imx_drm_resume(struct device *dev) { struct drm_device *drm_dev = dev_get_drvdata(dev); + struct imx_drm_device *imx_drm; if (drm_dev == NULL) return 0; - drm_helper_resume_force_mode(drm_dev); + imx_drm = drm_dev->dev_private; + drm_atomic_helper_resume(drm_dev, imx_drm->state); drm_kms_helper_poll_enable(drm_dev); return 0; diff --git a/drivers/gpu/drm/imx/imx-drm.h b/drivers/gpu/drm/imx/imx-drm.h index 74320a1723b7..07d33e45f90f 100644 --- a/drivers/gpu/drm/imx/imx-drm.h +++ b/drivers/gpu/drm/imx/imx-drm.h @@ -15,12 +15,22 @@ struct platform_device; unsigned int imx_drm_crtc_id(struct imx_drm_crtc *crtc); +struct imx_crtc_state { + struct drm_crtc_state base; + u32 bus_format; + u32 bus_flags; + int di_hsync_pin; + int di_vsync_pin; +}; + +static inline struct imx_crtc_state *to_imx_crtc_state(struct drm_crtc_state *s) +{ + return container_of(s, struct imx_crtc_state, base); +} + struct imx_drm_crtc_helper_funcs { int (*enable_vblank)(struct drm_crtc *crtc); void (*disable_vblank)(struct drm_crtc *crtc); - int (*set_interface_pix_fmt)(struct drm_crtc *crtc, - u32 bus_format, int hsync_pin, int vsync_pin, - u32 bus_flags); const struct drm_crtc_helper_funcs *crtc_helper_funcs; const struct drm_crtc_funcs *crtc_funcs; }; @@ -42,11 +52,6 @@ void imx_drm_mode_config_init(struct drm_device *drm); struct drm_gem_cma_object *imx_drm_fb_get_obj(struct drm_framebuffer *fb); -int imx_drm_set_bus_config(struct drm_encoder *encoder, u32 bus_format, - int hsync_pin, int vsync_pin, u32 bus_flags); -int imx_drm_set_bus_format(struct drm_encoder *encoder, - u32 bus_format); - int imx_drm_encoder_parse_of(struct drm_device *drm, struct drm_encoder *encoder, struct device_node *np); diff --git a/drivers/gpu/drm/imx/imx-ldb.c b/drivers/gpu/drm/imx/imx-ldb.c index beff793bb717..b03919ed60ba 100644 --- a/drivers/gpu/drm/imx/imx-ldb.c +++ b/drivers/gpu/drm/imx/imx-ldb.c @@ -17,6 +17,8 @@ #include <linux/clk.h> #include <linux/component.h> #include <drm/drmP.h> +#include <drm/drm_atomic.h> +#include <drm/drm_atomic_helper.h> #include <drm/drm_fb_helper.h> #include <drm/drm_crtc_helper.h> #include <drm/drm_of.h> @@ -49,9 +51,6 @@ #define LDB_DI1_VS_POL_ACT_LOW (1 << 10) #define LDB_BGREF_RMODE_INT (1 << 15) -#define con_to_imx_ldb_ch(x) container_of(x, struct imx_ldb_channel, connector) -#define enc_to_imx_ldb_ch(x) container_of(x, struct imx_ldb_channel, encoder) - struct imx_ldb; struct imx_ldb_channel { @@ -66,9 +65,19 @@ struct imx_ldb_channel { int edid_len; struct drm_display_mode mode; int mode_valid; - int bus_format; + u32 bus_format; }; +static inline struct imx_ldb_channel *con_to_imx_ldb_ch(struct drm_connector *c) +{ + return container_of(c, struct imx_ldb_channel, connector); +} + +static inline struct imx_ldb_channel *enc_to_imx_ldb_ch(struct drm_encoder *e) +{ + return container_of(e, struct imx_ldb_channel, encoder); +} + struct bus_mux { int reg; int shift; @@ -93,6 +102,32 @@ static enum drm_connector_status imx_ldb_connector_detect( return connector_status_connected; } +static void imx_ldb_ch_set_bus_format(struct imx_ldb_channel *imx_ldb_ch, + u32 bus_format) +{ + struct imx_ldb *ldb = imx_ldb_ch->ldb; + int dual = ldb->ldb_ctrl & LDB_SPLIT_MODE_EN; + + switch (bus_format) { + case MEDIA_BUS_FMT_RGB666_1X7X3_SPWG: + break; + case MEDIA_BUS_FMT_RGB888_1X7X4_SPWG: + if (imx_ldb_ch->chno == 0 || dual) + ldb->ldb_ctrl |= LDB_DATA_WIDTH_CH0_24; + if (imx_ldb_ch->chno == 1 || dual) + ldb->ldb_ctrl |= LDB_DATA_WIDTH_CH1_24; + break; + case MEDIA_BUS_FMT_RGB888_1X7X4_JEIDA: + if (imx_ldb_ch->chno == 0 || dual) + ldb->ldb_ctrl |= LDB_DATA_WIDTH_CH0_24 | + LDB_BIT_MAP_CH0_JEIDA; + if (imx_ldb_ch->chno == 1 || dual) + ldb->ldb_ctrl |= LDB_DATA_WIDTH_CH1_24 | + LDB_BIT_MAP_CH1_JEIDA; + break; + } +} + static int imx_ldb_connector_get_modes(struct drm_connector *connector) { struct imx_ldb_channel *imx_ldb_ch = con_to_imx_ldb_ch(connector); @@ -100,11 +135,7 @@ static int imx_ldb_connector_get_modes(struct drm_connector *connector) if (imx_ldb_ch->panel && imx_ldb_ch->panel->funcs && imx_ldb_ch->panel->funcs->get_modes) { - struct drm_display_info *di = &connector->display_info; - num_modes = imx_ldb_ch->panel->funcs->get_modes(imx_ldb_ch->panel); - if (!imx_ldb_ch->bus_format && di->num_bus_formats) - imx_ldb_ch->bus_format = di->bus_formats[0]; if (num_modes > 0) return num_modes; } @@ -141,10 +172,6 @@ static struct drm_encoder *imx_ldb_connector_best_encoder( return &imx_ldb_ch->encoder; } -static void imx_ldb_encoder_dpms(struct drm_encoder *encoder, int mode) -{ -} - static void imx_ldb_set_clock(struct imx_ldb *ldb, int mux, int chno, unsigned long serial_clk, unsigned long di_clk) { @@ -173,43 +200,7 @@ static void imx_ldb_set_clock(struct imx_ldb *ldb, int mux, int chno, chno); } -static void imx_ldb_encoder_prepare(struct drm_encoder *encoder) -{ - struct imx_ldb_channel *imx_ldb_ch = enc_to_imx_ldb_ch(encoder); - struct imx_ldb *ldb = imx_ldb_ch->ldb; - int dual = ldb->ldb_ctrl & LDB_SPLIT_MODE_EN; - u32 bus_format; - - switch (imx_ldb_ch->bus_format) { - default: - dev_warn(ldb->dev, - "could not determine data mapping, default to 18-bit \"spwg\"\n"); - /* fallthrough */ - case MEDIA_BUS_FMT_RGB666_1X7X3_SPWG: - bus_format = MEDIA_BUS_FMT_RGB666_1X18; - break; - case MEDIA_BUS_FMT_RGB888_1X7X4_SPWG: - bus_format = MEDIA_BUS_FMT_RGB888_1X24; - if (imx_ldb_ch->chno == 0 || dual) - ldb->ldb_ctrl |= LDB_DATA_WIDTH_CH0_24; - if (imx_ldb_ch->chno == 1 || dual) - ldb->ldb_ctrl |= LDB_DATA_WIDTH_CH1_24; - break; - case MEDIA_BUS_FMT_RGB888_1X7X4_JEIDA: - bus_format = MEDIA_BUS_FMT_RGB888_1X24; - if (imx_ldb_ch->chno == 0 || dual) - ldb->ldb_ctrl |= LDB_DATA_WIDTH_CH0_24 | - LDB_BIT_MAP_CH0_JEIDA; - if (imx_ldb_ch->chno == 1 || dual) - ldb->ldb_ctrl |= LDB_DATA_WIDTH_CH1_24 | - LDB_BIT_MAP_CH1_JEIDA; - break; - } - - imx_drm_set_bus_format(encoder, bus_format); -} - -static void imx_ldb_encoder_commit(struct drm_encoder *encoder) +static void imx_ldb_encoder_enable(struct drm_encoder *encoder) { struct imx_ldb_channel *imx_ldb_ch = enc_to_imx_ldb_ch(encoder); struct imx_ldb *ldb = imx_ldb_ch->ldb; @@ -219,8 +210,13 @@ static void imx_ldb_encoder_commit(struct drm_encoder *encoder) drm_panel_prepare(imx_ldb_ch->panel); if (dual) { + clk_set_parent(ldb->clk_sel[mux], ldb->clk[0]); + clk_set_parent(ldb->clk_sel[mux], ldb->clk[1]); + clk_prepare_enable(ldb->clk[0]); clk_prepare_enable(ldb->clk[1]); + } else { + clk_set_parent(ldb->clk_sel[mux], ldb->clk[imx_ldb_ch->chno]); } if (imx_ldb_ch == &ldb->channel[0] || dual) { @@ -265,6 +261,7 @@ static void imx_ldb_encoder_mode_set(struct drm_encoder *encoder, unsigned long serial_clk; unsigned long di_clk = mode->clock * 1000; int mux = drm_of_encoder_active_port_id(imx_ldb_ch->child, encoder); + u32 bus_format = imx_ldb_ch->bus_format; if (mode->clock > 170000) { dev_warn(ldb->dev, @@ -286,18 +283,33 @@ static void imx_ldb_encoder_mode_set(struct drm_encoder *encoder, } /* FIXME - assumes straight connections DI0 --> CH0, DI1 --> CH1 */ - if (imx_ldb_ch == &ldb->channel[0]) { + if (imx_ldb_ch == &ldb->channel[0] || dual) { if (mode->flags & DRM_MODE_FLAG_NVSYNC) ldb->ldb_ctrl |= LDB_DI0_VS_POL_ACT_LOW; else if (mode->flags & DRM_MODE_FLAG_PVSYNC) ldb->ldb_ctrl &= ~LDB_DI0_VS_POL_ACT_LOW; } - if (imx_ldb_ch == &ldb->channel[1]) { + if (imx_ldb_ch == &ldb->channel[1] || dual) { if (mode->flags & DRM_MODE_FLAG_NVSYNC) ldb->ldb_ctrl |= LDB_DI1_VS_POL_ACT_LOW; else if (mode->flags & DRM_MODE_FLAG_PVSYNC) ldb->ldb_ctrl &= ~LDB_DI1_VS_POL_ACT_LOW; } + + if (!bus_format) { + struct drm_connector *connector; + + drm_for_each_connector(connector, encoder->dev) { + struct drm_display_info *di = &connector->display_info; + + if (connector->encoder == encoder && + di->num_bus_formats) { + bus_format = di->bus_formats[0]; + break; + } + } + } + imx_ldb_ch_set_bus_format(imx_ldb_ch, bus_format); } static void imx_ldb_encoder_disable(struct drm_encoder *encoder) @@ -357,11 +369,45 @@ static void imx_ldb_encoder_disable(struct drm_encoder *encoder) drm_panel_unprepare(imx_ldb_ch->panel); } +static int imx_ldb_encoder_atomic_check(struct drm_encoder *encoder, + struct drm_crtc_state *crtc_state, + struct drm_connector_state *conn_state) +{ + struct imx_crtc_state *imx_crtc_state = to_imx_crtc_state(crtc_state); + struct imx_ldb_channel *imx_ldb_ch = enc_to_imx_ldb_ch(encoder); + struct drm_display_info *di = &conn_state->connector->display_info; + u32 bus_format = imx_ldb_ch->bus_format; + + /* Bus format description in DT overrides connector display info. */ + if (!bus_format && di->num_bus_formats) + bus_format = di->bus_formats[0]; + switch (bus_format) { + case MEDIA_BUS_FMT_RGB666_1X7X3_SPWG: + imx_crtc_state->bus_format = MEDIA_BUS_FMT_RGB666_1X18; + break; + case MEDIA_BUS_FMT_RGB888_1X7X4_SPWG: + case MEDIA_BUS_FMT_RGB888_1X7X4_JEIDA: + imx_crtc_state->bus_format = MEDIA_BUS_FMT_RGB888_1X24; + break; + default: + return -EINVAL; + } + + imx_crtc_state->di_hsync_pin = 2; + imx_crtc_state->di_vsync_pin = 3; + + return 0; +} + + static const struct drm_connector_funcs imx_ldb_connector_funcs = { - .dpms = drm_helper_connector_dpms, + .dpms = drm_atomic_helper_connector_dpms, .fill_modes = drm_helper_probe_single_connector_modes, .detect = imx_ldb_connector_detect, .destroy = imx_drm_connector_destroy, + .reset = drm_atomic_helper_connector_reset, + .atomic_duplicate_state = drm_atomic_helper_connector_duplicate_state, + .atomic_destroy_state = drm_atomic_helper_connector_destroy_state, }; static const struct drm_connector_helper_funcs imx_ldb_connector_helper_funcs = { @@ -374,11 +420,10 @@ static const struct drm_encoder_funcs imx_ldb_encoder_funcs = { }; static const struct drm_encoder_helper_funcs imx_ldb_encoder_helper_funcs = { - .dpms = imx_ldb_encoder_dpms, - .prepare = imx_ldb_encoder_prepare, - .commit = imx_ldb_encoder_commit, .mode_set = imx_ldb_encoder_mode_set, + .enable = imx_ldb_encoder_enable, .disable = imx_ldb_encoder_disable, + .atomic_check = imx_ldb_encoder_atomic_check, }; static int imx_ldb_get_clk(struct imx_ldb *ldb, int chno) @@ -400,10 +445,10 @@ static int imx_ldb_register(struct drm_device *drm, struct imx_ldb_channel *imx_ldb_ch) { struct imx_ldb *ldb = imx_ldb_ch->ldb; + struct drm_encoder *encoder = &imx_ldb_ch->encoder; int ret; - ret = imx_drm_encoder_parse_of(drm, &imx_ldb_ch->encoder, - imx_ldb_ch->child); + ret = imx_drm_encoder_parse_of(drm, encoder, imx_ldb_ch->child); if (ret) return ret; @@ -417,9 +462,8 @@ static int imx_ldb_register(struct drm_device *drm, return ret; } - drm_encoder_helper_add(&imx_ldb_ch->encoder, - &imx_ldb_encoder_helper_funcs); - drm_encoder_init(drm, &imx_ldb_ch->encoder, &imx_ldb_encoder_funcs, + drm_encoder_helper_add(encoder, &imx_ldb_encoder_helper_funcs); + drm_encoder_init(drm, encoder, &imx_ldb_encoder_funcs, DRM_MODE_ENCODER_LVDS, NULL); drm_connector_helper_add(&imx_ldb_ch->connector, @@ -427,11 +471,14 @@ static int imx_ldb_register(struct drm_device *drm, drm_connector_init(drm, &imx_ldb_ch->connector, &imx_ldb_connector_funcs, DRM_MODE_CONNECTOR_LVDS); - if (imx_ldb_ch->panel) - drm_panel_attach(imx_ldb_ch->panel, &imx_ldb_ch->connector); + if (imx_ldb_ch->panel) { + ret = drm_panel_attach(imx_ldb_ch->panel, + &imx_ldb_ch->connector); + if (ret) + return ret; + } - drm_mode_connector_attach_encoder(&imx_ldb_ch->connector, - &imx_ldb_ch->encoder); + drm_mode_connector_attach_encoder(&imx_ldb_ch->connector, encoder); return 0; } @@ -560,6 +607,7 @@ static int imx_ldb_bind(struct device *dev, struct device *master, void *data) struct imx_ldb_channel *channel; struct device_node *ddc_node; struct device_node *ep; + int bus_format; ret = of_property_read_u32(child, "reg", &i); if (ret || i < 0 || i > 1) @@ -632,21 +680,22 @@ static int imx_ldb_bind(struct device *dev, struct device *master, void *data) } } - channel->bus_format = of_get_bus_format(dev, child); - if (channel->bus_format == -EINVAL) { + bus_format = of_get_bus_format(dev, child); + if (bus_format == -EINVAL) { /* * If no bus format was specified in the device tree, * we can still get it from the connected panel later. */ if (channel->panel && channel->panel->funcs && channel->panel->funcs->get_modes) - channel->bus_format = 0; + bus_format = 0; } - if (channel->bus_format < 0) { + if (bus_format < 0) { dev_err(dev, "could not determine data mapping: %d\n", - channel->bus_format); - return channel->bus_format; + bus_format); + return bus_format; } + channel->bus_format = bus_format; ret = imx_ldb_register(drm, channel); if (ret) diff --git a/drivers/gpu/drm/imx/imx-tve.c b/drivers/gpu/drm/imx/imx-tve.c index baf788121287..5e875944ffa2 100644 --- a/drivers/gpu/drm/imx/imx-tve.c +++ b/drivers/gpu/drm/imx/imx-tve.c @@ -23,6 +23,7 @@ #include <linux/spinlock.h> #include <linux/videodev2.h> #include <drm/drmP.h> +#include <drm/drm_atomic_helper.h> #include <drm/drm_fb_helper.h> #include <drm/drm_crtc_helper.h> #include <video/imx-ipu-v3.h> @@ -97,9 +98,6 @@ /* TVE_TST_MODE_REG */ #define TVE_TVDAC_TEST_MODE_MASK (0x7 << 0) -#define con_to_tve(x) container_of(x, struct imx_tve, connector) -#define enc_to_tve(x) container_of(x, struct imx_tve, encoder) - enum { TVE_MODE_TVOUT, TVE_MODE_VGA, @@ -112,6 +110,8 @@ struct imx_tve { spinlock_t lock; /* register lock */ bool enabled; int mode; + int di_hsync_pin; + int di_vsync_pin; struct regmap *regmap; struct regulator *dac_reg; @@ -120,10 +120,18 @@ struct imx_tve { struct clk *di_sel_clk; struct clk_hw clk_hw_di; struct clk *di_clk; - int vsync_pin; - int hsync_pin; }; +static inline struct imx_tve *con_to_tve(struct drm_connector *c) +{ + return container_of(c, struct imx_tve, connector); +} + +static inline struct imx_tve *enc_to_tve(struct drm_encoder *e) +{ + return container_of(e, struct imx_tve, encoder); +} + static void tve_lock(void *__tve) __acquires(&tve->lock) { @@ -148,8 +156,7 @@ static void tve_enable(struct imx_tve *tve) tve->enabled = true; clk_prepare_enable(tve->clk); ret = regmap_update_bits(tve->regmap, TVE_COM_CONF_REG, - TVE_IPU_CLK_EN | TVE_EN, - TVE_IPU_CLK_EN | TVE_EN); + TVE_EN, TVE_EN); } /* clear interrupt status register */ @@ -172,7 +179,7 @@ static void tve_disable(struct imx_tve *tve) if (tve->enabled) { tve->enabled = false; ret = regmap_update_bits(tve->regmap, TVE_COM_CONF_REG, - TVE_IPU_CLK_EN | TVE_EN, 0); + TVE_EN, 0); clk_disable_unprepare(tve->clk); } } @@ -275,36 +282,6 @@ static struct drm_encoder *imx_tve_connector_best_encoder( return &tve->encoder; } -static void imx_tve_encoder_dpms(struct drm_encoder *encoder, int mode) -{ - struct imx_tve *tve = enc_to_tve(encoder); - int ret; - - ret = regmap_update_bits(tve->regmap, TVE_COM_CONF_REG, - TVE_TV_OUT_MODE_MASK, TVE_TV_OUT_DISABLE); - if (ret < 0) - dev_err(tve->dev, "failed to disable TVOUT: %d\n", ret); -} - -static void imx_tve_encoder_prepare(struct drm_encoder *encoder) -{ - struct imx_tve *tve = enc_to_tve(encoder); - - tve_disable(tve); - - switch (tve->mode) { - case TVE_MODE_VGA: - imx_drm_set_bus_config(encoder, MEDIA_BUS_FMT_GBR888_1X24, - tve->hsync_pin, tve->vsync_pin, - DRM_BUS_FLAG_DE_HIGH | - DRM_BUS_FLAG_PIXDATA_NEGEDGE); - break; - case TVE_MODE_TVOUT: - imx_drm_set_bus_format(encoder, MEDIA_BUS_FMT_YUV8_1X24); - break; - } -} - static void imx_tve_encoder_mode_set(struct drm_encoder *encoder, struct drm_display_mode *orig_mode, struct drm_display_mode *mode) @@ -333,6 +310,9 @@ static void imx_tve_encoder_mode_set(struct drm_encoder *encoder, ret); } + regmap_update_bits(tve->regmap, TVE_COM_CONF_REG, + TVE_IPU_CLK_EN, TVE_IPU_CLK_EN); + if (tve->mode == TVE_MODE_VGA) ret = tve_setup_vga(tve); else @@ -341,7 +321,7 @@ static void imx_tve_encoder_mode_set(struct drm_encoder *encoder, dev_err(tve->dev, "failed to set configuration: %d\n", ret); } -static void imx_tve_encoder_commit(struct drm_encoder *encoder) +static void imx_tve_encoder_enable(struct drm_encoder *encoder) { struct imx_tve *tve = enc_to_tve(encoder); @@ -355,11 +335,28 @@ static void imx_tve_encoder_disable(struct drm_encoder *encoder) tve_disable(tve); } +static int imx_tve_atomic_check(struct drm_encoder *encoder, + struct drm_crtc_state *crtc_state, + struct drm_connector_state *conn_state) +{ + struct imx_crtc_state *imx_crtc_state = to_imx_crtc_state(crtc_state); + struct imx_tve *tve = enc_to_tve(encoder); + + imx_crtc_state->bus_format = MEDIA_BUS_FMT_GBR888_1X24; + imx_crtc_state->di_hsync_pin = tve->di_hsync_pin; + imx_crtc_state->di_vsync_pin = tve->di_vsync_pin; + + return 0; +} + static const struct drm_connector_funcs imx_tve_connector_funcs = { - .dpms = drm_helper_connector_dpms, + .dpms = drm_atomic_helper_connector_dpms, .fill_modes = drm_helper_probe_single_connector_modes, .detect = imx_tve_connector_detect, .destroy = imx_drm_connector_destroy, + .reset = drm_atomic_helper_connector_reset, + .atomic_duplicate_state = drm_atomic_helper_connector_duplicate_state, + .atomic_destroy_state = drm_atomic_helper_connector_destroy_state, }; static const struct drm_connector_helper_funcs imx_tve_connector_helper_funcs = { @@ -373,11 +370,10 @@ static const struct drm_encoder_funcs imx_tve_encoder_funcs = { }; static const struct drm_encoder_helper_funcs imx_tve_encoder_helper_funcs = { - .dpms = imx_tve_encoder_dpms, - .prepare = imx_tve_encoder_prepare, .mode_set = imx_tve_encoder_mode_set, - .commit = imx_tve_encoder_commit, + .enable = imx_tve_encoder_enable, .disable = imx_tve_encoder_disable, + .atomic_check = imx_tve_atomic_check, }; static irqreturn_t imx_tve_irq_handler(int irq, void *data) @@ -495,8 +491,7 @@ static int imx_tve_register(struct drm_device *drm, struct imx_tve *tve) encoder_type = tve->mode == TVE_MODE_VGA ? DRM_MODE_ENCODER_DAC : DRM_MODE_ENCODER_TVDAC; - ret = imx_drm_encoder_parse_of(drm, &tve->encoder, - tve->dev->of_node); + ret = imx_drm_encoder_parse_of(drm, &tve->encoder, tve->dev->of_node); if (ret) return ret; @@ -587,15 +582,15 @@ static int imx_tve_bind(struct device *dev, struct device *master, void *data) if (tve->mode == TVE_MODE_VGA) { ret = of_property_read_u32(np, "fsl,hsync-pin", - &tve->hsync_pin); + &tve->di_hsync_pin); if (ret < 0) { - dev_err(dev, "failed to get vsync pin\n"); + dev_err(dev, "failed to get hsync pin\n"); return ret; } - ret |= of_property_read_u32(np, "fsl,vsync-pin", - &tve->vsync_pin); + ret = of_property_read_u32(np, "fsl,vsync-pin", + &tve->di_vsync_pin); if (ret < 0) { dev_err(dev, "failed to get vsync pin\n"); @@ -633,7 +628,9 @@ static int imx_tve_bind(struct device *dev, struct device *master, void *data) tve->dac_reg = devm_regulator_get(dev, "dac"); if (!IS_ERR(tve->dac_reg)) { - regulator_set_voltage(tve->dac_reg, 2750000, 2750000); + ret = regulator_set_voltage(tve->dac_reg, 2750000, 2750000); + if (ret) + return ret; ret = regulator_enable(tve->dac_reg); if (ret) return ret; diff --git a/drivers/gpu/drm/imx/ipuv3-crtc.c b/drivers/gpu/drm/imx/ipuv3-crtc.c index fc040417e1e8..08e188bc10fc 100644 --- a/drivers/gpu/drm/imx/ipuv3-crtc.c +++ b/drivers/gpu/drm/imx/ipuv3-crtc.c @@ -18,12 +18,12 @@ #include <linux/device.h> #include <linux/platform_device.h> #include <drm/drmP.h> +#include <drm/drm_atomic.h> +#include <drm/drm_atomic_helper.h> #include <drm/drm_crtc_helper.h> #include <linux/fb.h> #include <linux/clk.h> #include <linux/errno.h> -#include <linux/reservation.h> -#include <linux/dma-buf.h> #include <drm/drm_gem_cma_helper.h> #include <drm/drm_fb_cma_helper.h> @@ -33,23 +33,6 @@ #define DRIVER_DESC "i.MX IPUv3 Graphics" -enum ipu_flip_status { - IPU_FLIP_NONE, - IPU_FLIP_PENDING, - IPU_FLIP_SUBMITTED, -}; - -struct ipu_flip_work { - struct work_struct unref_work; - struct drm_gem_object *bo; - struct drm_pending_vblank_event *page_flip_event; - struct work_struct fence_work; - struct ipu_crtc *crtc; - struct fence *excl; - unsigned shared_count; - struct fence **shared; -}; - struct ipu_crtc { struct device *dev; struct drm_crtc base; @@ -60,201 +43,166 @@ struct ipu_crtc { struct ipu_dc *dc; struct ipu_di *di; - int enabled; - enum ipu_flip_status flip_state; - struct workqueue_struct *flip_queue; - struct ipu_flip_work *flip_work; int irq; - u32 bus_format; - u32 bus_flags; - int di_hsync_pin; - int di_vsync_pin; }; -#define to_ipu_crtc(x) container_of(x, struct ipu_crtc, base) +static inline struct ipu_crtc *to_ipu_crtc(struct drm_crtc *crtc) +{ + return container_of(crtc, struct ipu_crtc, base); +} -static void ipu_fb_enable(struct ipu_crtc *ipu_crtc) +static void ipu_crtc_enable(struct drm_crtc *crtc) { + struct ipu_crtc *ipu_crtc = to_ipu_crtc(crtc); struct ipu_soc *ipu = dev_get_drvdata(ipu_crtc->dev->parent); - if (ipu_crtc->enabled) - return; - ipu_dc_enable(ipu); - ipu_plane_enable(ipu_crtc->plane[0]); - /* Start DC channel and DI after IDMAC */ ipu_dc_enable_channel(ipu_crtc->dc); ipu_di_enable(ipu_crtc->di); - drm_crtc_vblank_on(&ipu_crtc->base); - - ipu_crtc->enabled = 1; } -static void ipu_fb_disable(struct ipu_crtc *ipu_crtc) +static void ipu_crtc_disable(struct drm_crtc *crtc) { + struct ipu_crtc *ipu_crtc = to_ipu_crtc(crtc); struct ipu_soc *ipu = dev_get_drvdata(ipu_crtc->dev->parent); - if (!ipu_crtc->enabled) - return; - - /* Stop DC channel and DI before IDMAC */ ipu_dc_disable_channel(ipu_crtc->dc); ipu_di_disable(ipu_crtc->di); - ipu_plane_disable(ipu_crtc->plane[0]); ipu_dc_disable(ipu); - drm_crtc_vblank_off(&ipu_crtc->base); - ipu_crtc->enabled = 0; + spin_lock_irq(&crtc->dev->event_lock); + if (crtc->state->event) { + drm_crtc_send_vblank_event(crtc, crtc->state->event); + crtc->state->event = NULL; + } + spin_unlock_irq(&crtc->dev->event_lock); } -static void ipu_crtc_dpms(struct drm_crtc *crtc, int mode) +static void imx_drm_crtc_reset(struct drm_crtc *crtc) { - struct ipu_crtc *ipu_crtc = to_ipu_crtc(crtc); + struct imx_crtc_state *state; - dev_dbg(ipu_crtc->dev, "%s mode: %d\n", __func__, mode); - - switch (mode) { - case DRM_MODE_DPMS_ON: - ipu_fb_enable(ipu_crtc); - break; - case DRM_MODE_DPMS_STANDBY: - case DRM_MODE_DPMS_SUSPEND: - case DRM_MODE_DPMS_OFF: - ipu_fb_disable(ipu_crtc); - break; + if (crtc->state) { + if (crtc->state->mode_blob) + drm_property_unreference_blob(crtc->state->mode_blob); + + state = to_imx_crtc_state(crtc->state); + memset(state, 0, sizeof(*state)); + } else { + state = kzalloc(sizeof(*state), GFP_KERNEL); + if (!state) + return; + crtc->state = &state->base; } + + state->base.crtc = crtc; } -static void ipu_flip_unref_work_func(struct work_struct *__work) +static struct drm_crtc_state *imx_drm_crtc_duplicate_state(struct drm_crtc *crtc) { - struct ipu_flip_work *work = - container_of(__work, struct ipu_flip_work, unref_work); + struct imx_crtc_state *state; + + state = kzalloc(sizeof(*state), GFP_KERNEL); + if (!state) + return NULL; - drm_gem_object_unreference_unlocked(work->bo); - kfree(work); + __drm_atomic_helper_crtc_duplicate_state(crtc, &state->base); + + WARN_ON(state->base.crtc != crtc); + state->base.crtc = crtc; + + return &state->base; } -static void ipu_flip_fence_work_func(struct work_struct *__work) +static void imx_drm_crtc_destroy_state(struct drm_crtc *crtc, + struct drm_crtc_state *state) { - struct ipu_flip_work *work = - container_of(__work, struct ipu_flip_work, fence_work); - int i; - - /* wait for all fences attached to the FB obj to signal */ - if (work->excl) { - fence_wait(work->excl, false); - fence_put(work->excl); - } - for (i = 0; i < work->shared_count; i++) { - fence_wait(work->shared[i], false); - fence_put(work->shared[i]); - } + __drm_atomic_helper_crtc_destroy_state(state); + kfree(to_imx_crtc_state(state)); +} + +static const struct drm_crtc_funcs ipu_crtc_funcs = { + .set_config = drm_atomic_helper_set_config, + .destroy = drm_crtc_cleanup, + .page_flip = drm_atomic_helper_page_flip, + .reset = imx_drm_crtc_reset, + .atomic_duplicate_state = imx_drm_crtc_duplicate_state, + .atomic_destroy_state = imx_drm_crtc_destroy_state, +}; - work->crtc->flip_state = IPU_FLIP_SUBMITTED; +static irqreturn_t ipu_irq_handler(int irq, void *dev_id) +{ + struct ipu_crtc *ipu_crtc = dev_id; + + imx_drm_handle_vblank(ipu_crtc->imx_crtc); + + return IRQ_HANDLED; } -static int ipu_page_flip(struct drm_crtc *crtc, - struct drm_framebuffer *fb, - struct drm_pending_vblank_event *event, - uint32_t page_flip_flags) +static bool ipu_crtc_mode_fixup(struct drm_crtc *crtc, + const struct drm_display_mode *mode, + struct drm_display_mode *adjusted_mode) { - struct drm_gem_cma_object *cma_obj = drm_fb_cma_get_gem_obj(fb, 0); struct ipu_crtc *ipu_crtc = to_ipu_crtc(crtc); - struct ipu_flip_work *flip_work; + struct videomode vm; int ret; - if (ipu_crtc->flip_state != IPU_FLIP_NONE) - return -EBUSY; - - ret = imx_drm_crtc_vblank_get(ipu_crtc->imx_crtc); - if (ret) { - dev_dbg(ipu_crtc->dev, "failed to acquire vblank counter\n"); - list_del(&event->base.link); - - return ret; - } + drm_display_mode_to_videomode(adjusted_mode, &vm); - flip_work = kzalloc(sizeof *flip_work, GFP_KERNEL); - if (!flip_work) { - ret = -ENOMEM; - goto put_vblank; - } - INIT_WORK(&flip_work->unref_work, ipu_flip_unref_work_func); - flip_work->page_flip_event = event; + ret = ipu_di_adjust_videomode(ipu_crtc->di, &vm); + if (ret) + return false; - /* get BO backing the old framebuffer and take a reference */ - flip_work->bo = &drm_fb_cma_get_gem_obj(crtc->primary->fb, 0)->base; - drm_gem_object_reference(flip_work->bo); + if ((vm.vsync_len == 0) || (vm.hsync_len == 0)) + return false; - ipu_crtc->flip_work = flip_work; - /* - * If the object has a DMABUF attached, we need to wait on its fences - * if there are any. - */ - if (cma_obj->base.dma_buf) { - INIT_WORK(&flip_work->fence_work, ipu_flip_fence_work_func); - flip_work->crtc = ipu_crtc; + drm_display_mode_from_videomode(&vm, adjusted_mode); - ret = reservation_object_get_fences_rcu( - cma_obj->base.dma_buf->resv, &flip_work->excl, - &flip_work->shared_count, &flip_work->shared); + return true; +} - if (unlikely(ret)) { - DRM_ERROR("failed to get fences for buffer\n"); - goto free_flip_work; - } +static int ipu_crtc_atomic_check(struct drm_crtc *crtc, + struct drm_crtc_state *state) +{ + u32 primary_plane_mask = 1 << drm_plane_index(crtc->primary); - /* No need to queue the worker if the are no fences */ - if (!flip_work->excl && !flip_work->shared_count) { - ipu_crtc->flip_state = IPU_FLIP_SUBMITTED; - } else { - ipu_crtc->flip_state = IPU_FLIP_PENDING; - queue_work(ipu_crtc->flip_queue, - &flip_work->fence_work); - } - } else { - ipu_crtc->flip_state = IPU_FLIP_SUBMITTED; - } + if (state->active && (primary_plane_mask & state->plane_mask) == 0) + return -EINVAL; return 0; - -free_flip_work: - drm_gem_object_unreference_unlocked(flip_work->bo); - kfree(flip_work); - ipu_crtc->flip_work = NULL; -put_vblank: - imx_drm_crtc_vblank_put(ipu_crtc->imx_crtc); - - return ret; } -static const struct drm_crtc_funcs ipu_crtc_funcs = { - .set_config = drm_crtc_helper_set_config, - .destroy = drm_crtc_cleanup, - .page_flip = ipu_page_flip, -}; +static void ipu_crtc_atomic_begin(struct drm_crtc *crtc, + struct drm_crtc_state *old_crtc_state) +{ + spin_lock_irq(&crtc->dev->event_lock); + if (crtc->state->event) { + WARN_ON(drm_crtc_vblank_get(crtc)); + drm_crtc_arm_vblank_event(crtc, crtc->state->event); + crtc->state->event = NULL; + } + spin_unlock_irq(&crtc->dev->event_lock); +} -static int ipu_crtc_mode_set(struct drm_crtc *crtc, - struct drm_display_mode *orig_mode, - struct drm_display_mode *mode, - int x, int y, - struct drm_framebuffer *old_fb) +static void ipu_crtc_mode_set_nofb(struct drm_crtc *crtc) { struct drm_device *dev = crtc->dev; struct drm_encoder *encoder; struct ipu_crtc *ipu_crtc = to_ipu_crtc(crtc); + struct drm_display_mode *mode = &crtc->state->adjusted_mode; + struct imx_crtc_state *imx_crtc_state = to_imx_crtc_state(crtc->state); struct ipu_di_signal_cfg sig_cfg = {}; unsigned long encoder_types = 0; - int ret; dev_dbg(ipu_crtc->dev, "%s: mode->hdisplay: %d\n", __func__, mode->hdisplay); dev_dbg(ipu_crtc->dev, "%s: mode->vdisplay: %d\n", __func__, mode->vdisplay); - list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) + list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) { if (encoder->crtc == crtc) encoder_types |= BIT(encoder->encoder_type); + } dev_dbg(ipu_crtc->dev, "%s: attached to encoder types 0x%lx\n", __func__, encoder_types); @@ -272,114 +220,30 @@ static int ipu_crtc_mode_set(struct drm_crtc *crtc, else sig_cfg.clkflags = 0; - sig_cfg.enable_pol = !(ipu_crtc->bus_flags & DRM_BUS_FLAG_DE_LOW); + sig_cfg.enable_pol = !(imx_crtc_state->bus_flags & DRM_BUS_FLAG_DE_LOW); /* Default to driving pixel data on negative clock edges */ - sig_cfg.clk_pol = !!(ipu_crtc->bus_flags & + sig_cfg.clk_pol = !!(imx_crtc_state->bus_flags & DRM_BUS_FLAG_PIXDATA_POSEDGE); - sig_cfg.bus_format = ipu_crtc->bus_format; + sig_cfg.bus_format = imx_crtc_state->bus_format; sig_cfg.v_to_h_sync = 0; - sig_cfg.hsync_pin = ipu_crtc->di_hsync_pin; - sig_cfg.vsync_pin = ipu_crtc->di_vsync_pin; + sig_cfg.hsync_pin = imx_crtc_state->di_hsync_pin; + sig_cfg.vsync_pin = imx_crtc_state->di_vsync_pin; drm_display_mode_to_videomode(mode, &sig_cfg.mode); - ret = ipu_dc_init_sync(ipu_crtc->dc, ipu_crtc->di, - mode->flags & DRM_MODE_FLAG_INTERLACE, - ipu_crtc->bus_format, mode->hdisplay); - if (ret) { - dev_err(ipu_crtc->dev, - "initializing display controller failed with %d\n", - ret); - return ret; - } - - ret = ipu_di_init_sync_panel(ipu_crtc->di, &sig_cfg); - if (ret) { - dev_err(ipu_crtc->dev, - "initializing panel failed with %d\n", ret); - return ret; - } - - return ipu_plane_mode_set(ipu_crtc->plane[0], crtc, mode, - crtc->primary->fb, - 0, 0, mode->hdisplay, mode->vdisplay, - x, y, mode->hdisplay, mode->vdisplay, - mode->flags & DRM_MODE_FLAG_INTERLACE); -} - -static void ipu_crtc_handle_pageflip(struct ipu_crtc *ipu_crtc) -{ - unsigned long flags; - struct drm_device *drm = ipu_crtc->base.dev; - struct ipu_flip_work *work = ipu_crtc->flip_work; - - spin_lock_irqsave(&drm->event_lock, flags); - if (work->page_flip_event) - drm_crtc_send_vblank_event(&ipu_crtc->base, - work->page_flip_event); - imx_drm_crtc_vblank_put(ipu_crtc->imx_crtc); - spin_unlock_irqrestore(&drm->event_lock, flags); -} - -static irqreturn_t ipu_irq_handler(int irq, void *dev_id) -{ - struct ipu_crtc *ipu_crtc = dev_id; - - imx_drm_handle_vblank(ipu_crtc->imx_crtc); - - if (ipu_crtc->flip_state == IPU_FLIP_SUBMITTED) { - struct ipu_plane *plane = ipu_crtc->plane[0]; - - ipu_plane_set_base(plane, ipu_crtc->base.primary->fb, - plane->x, plane->y); - ipu_crtc_handle_pageflip(ipu_crtc); - queue_work(ipu_crtc->flip_queue, - &ipu_crtc->flip_work->unref_work); - ipu_crtc->flip_state = IPU_FLIP_NONE; - } - - return IRQ_HANDLED; -} - -static bool ipu_crtc_mode_fixup(struct drm_crtc *crtc, - const struct drm_display_mode *mode, - struct drm_display_mode *adjusted_mode) -{ - struct ipu_crtc *ipu_crtc = to_ipu_crtc(crtc); - struct videomode vm; - int ret; - - drm_display_mode_to_videomode(adjusted_mode, &vm); - - ret = ipu_di_adjust_videomode(ipu_crtc->di, &vm); - if (ret) - return false; - - drm_display_mode_from_videomode(&vm, adjusted_mode); - - return true; -} - -static void ipu_crtc_prepare(struct drm_crtc *crtc) -{ - struct ipu_crtc *ipu_crtc = to_ipu_crtc(crtc); - - ipu_fb_disable(ipu_crtc); -} - -static void ipu_crtc_commit(struct drm_crtc *crtc) -{ - struct ipu_crtc *ipu_crtc = to_ipu_crtc(crtc); - - ipu_fb_enable(ipu_crtc); + ipu_dc_init_sync(ipu_crtc->dc, ipu_crtc->di, + mode->flags & DRM_MODE_FLAG_INTERLACE, + imx_crtc_state->bus_format, mode->hdisplay); + ipu_di_init_sync_panel(ipu_crtc->di, &sig_cfg); } static const struct drm_crtc_helper_funcs ipu_helper_funcs = { - .dpms = ipu_crtc_dpms, .mode_fixup = ipu_crtc_mode_fixup, - .mode_set = ipu_crtc_mode_set, - .prepare = ipu_crtc_prepare, - .commit = ipu_crtc_commit, + .mode_set_nofb = ipu_crtc_mode_set_nofb, + .atomic_check = ipu_crtc_atomic_check, + .atomic_begin = ipu_crtc_atomic_begin, + .disable = ipu_crtc_disable, + .enable = ipu_crtc_enable, }; static int ipu_enable_vblank(struct drm_crtc *crtc) @@ -398,23 +262,9 @@ static void ipu_disable_vblank(struct drm_crtc *crtc) disable_irq_nosync(ipu_crtc->irq); } -static int ipu_set_interface_pix_fmt(struct drm_crtc *crtc, - u32 bus_format, int hsync_pin, int vsync_pin, u32 bus_flags) -{ - struct ipu_crtc *ipu_crtc = to_ipu_crtc(crtc); - - ipu_crtc->bus_format = bus_format; - ipu_crtc->bus_flags = bus_flags; - ipu_crtc->di_hsync_pin = hsync_pin; - ipu_crtc->di_vsync_pin = vsync_pin; - - return 0; -} - static const struct imx_drm_crtc_helper_funcs ipu_crtc_helper_funcs = { .enable_vblank = ipu_enable_vblank, .disable_vblank = ipu_disable_vblank, - .set_interface_pix_fmt = ipu_set_interface_pix_fmt, .crtc_funcs = &ipu_crtc_funcs, .crtc_helper_funcs = &ipu_helper_funcs, }; @@ -496,8 +346,16 @@ static int ipu_crtc_init(struct ipu_crtc *ipu_crtc, IPU_DP_FLOW_SYNC_FG, drm_crtc_mask(&ipu_crtc->base), DRM_PLANE_TYPE_OVERLAY); - if (IS_ERR(ipu_crtc->plane[1])) + if (IS_ERR(ipu_crtc->plane[1])) { ipu_crtc->plane[1] = NULL; + } else { + ret = ipu_plane_get_resources(ipu_crtc->plane[1]); + if (ret) { + dev_err(ipu_crtc->dev, "getting plane 1 " + "resources failed with %d.\n", ret); + goto err_put_plane0_res; + } + } } ipu_crtc->irq = ipu_plane_irq(ipu_crtc->plane[0]); @@ -505,16 +363,17 @@ static int ipu_crtc_init(struct ipu_crtc *ipu_crtc, "imx_drm", ipu_crtc); if (ret < 0) { dev_err(ipu_crtc->dev, "irq request failed with %d.\n", ret); - goto err_put_plane_res; + goto err_put_plane1_res; } /* Only enable IRQ when we actually need it to trigger work. */ disable_irq(ipu_crtc->irq); - ipu_crtc->flip_queue = create_singlethread_workqueue("ipu-crtc-flip"); - return 0; -err_put_plane_res: +err_put_plane1_res: + if (ipu_crtc->plane[1]) + ipu_plane_put_resources(ipu_crtc->plane[1]); +err_put_plane0_res: ipu_plane_put_resources(ipu_crtc->plane[0]); err_remove_crtc: imx_drm_remove_crtc(ipu_crtc->imx_crtc); @@ -553,9 +412,10 @@ static void ipu_drm_unbind(struct device *dev, struct device *master, imx_drm_remove_crtc(ipu_crtc->imx_crtc); - destroy_workqueue(ipu_crtc->flip_queue); - ipu_plane_put_resources(ipu_crtc->plane[0]); ipu_put_resources(ipu_crtc); + if (ipu_crtc->plane[1]) + ipu_plane_put_resources(ipu_crtc->plane[1]); + ipu_plane_put_resources(ipu_crtc->plane[0]); } static const struct component_ops ipu_crtc_ops = { diff --git a/drivers/gpu/drm/imx/ipuv3-plane.c b/drivers/gpu/drm/imx/ipuv3-plane.c index a4bb44118d33..4ad67d015ec7 100644 --- a/drivers/gpu/drm/imx/ipuv3-plane.c +++ b/drivers/gpu/drm/imx/ipuv3-plane.c @@ -14,13 +14,19 @@ */ #include <drm/drmP.h> +#include <drm/drm_atomic.h> +#include <drm/drm_atomic_helper.h> #include <drm/drm_fb_cma_helper.h> #include <drm/drm_gem_cma_helper.h> +#include <drm/drm_plane_helper.h> #include "video/imx-ipu-v3.h" #include "ipuv3-plane.h" -#define to_ipu_plane(x) container_of(x, struct ipu_plane, base) +static inline struct ipu_plane *to_ipu_plane(struct drm_plane *p) +{ + return container_of(p, struct ipu_plane, base); +} static const uint32_t ipu_plane_formats[] = { DRM_FORMAT_ARGB1555, @@ -53,62 +59,67 @@ int ipu_plane_irq(struct ipu_plane *ipu_plane) IPU_IRQ_EOF); } -static int calc_vref(struct drm_display_mode *mode) +static inline unsigned long +drm_plane_state_to_eba(struct drm_plane_state *state) { - unsigned long htotal, vtotal; + struct drm_framebuffer *fb = state->fb; + struct drm_gem_cma_object *cma_obj; - htotal = mode->htotal; - vtotal = mode->vtotal; + cma_obj = drm_fb_cma_get_gem_obj(fb, 0); + BUG_ON(!cma_obj); - if (!htotal || !vtotal) - return 60; - - return DIV_ROUND_UP(mode->clock * 1000, vtotal * htotal); + return cma_obj->paddr + fb->offsets[0] + + fb->pitches[0] * (state->src_y >> 16) + + (fb->bits_per_pixel >> 3) * (state->src_x >> 16); } -static inline int calc_bandwidth(int width, int height, unsigned int vref) +static inline unsigned long +drm_plane_state_to_ubo(struct drm_plane_state *state) { - return width * height * vref; -} + struct drm_framebuffer *fb = state->fb; + struct drm_gem_cma_object *cma_obj; + unsigned long eba = drm_plane_state_to_eba(state); -int ipu_plane_set_base(struct ipu_plane *ipu_plane, struct drm_framebuffer *fb, - int x, int y) -{ - struct drm_gem_cma_object *cma_obj[3]; - unsigned long eba, ubo, vbo; - int active, i; + cma_obj = drm_fb_cma_get_gem_obj(fb, 1); + BUG_ON(!cma_obj); - for (i = 0; i < drm_format_num_planes(fb->pixel_format); i++) { - cma_obj[i] = drm_fb_cma_get_gem_obj(fb, i); - if (!cma_obj[i]) { - DRM_DEBUG_KMS("plane %d entry is null.\n", i); - return -EFAULT; - } - } + return cma_obj->paddr + fb->offsets[1] + + fb->pitches[1] * (state->src_y >> 16) / 2 + + (state->src_x >> 16) / 2 - eba; +} - eba = cma_obj[0]->paddr + fb->offsets[0] + - fb->pitches[0] * y + (fb->bits_per_pixel >> 3) * x; +static inline unsigned long +drm_plane_state_to_vbo(struct drm_plane_state *state) +{ + struct drm_framebuffer *fb = state->fb; + struct drm_gem_cma_object *cma_obj; + unsigned long eba = drm_plane_state_to_eba(state); - if (eba & 0x7) { - DRM_DEBUG_KMS("base address must be a multiple of 8.\n"); - return -EINVAL; - } + cma_obj = drm_fb_cma_get_gem_obj(fb, 2); + BUG_ON(!cma_obj); - if (fb->pitches[0] < 1 || fb->pitches[0] > 16384) { - DRM_DEBUG_KMS("pitches out of range.\n"); - return -EINVAL; - } + return cma_obj->paddr + fb->offsets[2] + + fb->pitches[2] * (state->src_y >> 16) / 2 + + (state->src_x >> 16) / 2 - eba; +} - if (ipu_plane->enabled && fb->pitches[0] != ipu_plane->stride[0]) { - DRM_DEBUG_KMS("pitches must not change while plane is enabled.\n"); - return -EINVAL; - } +static void ipu_plane_atomic_set_base(struct ipu_plane *ipu_plane, + struct drm_plane_state *old_state) +{ + struct drm_plane *plane = &ipu_plane->base; + struct drm_plane_state *state = plane->state; + struct drm_framebuffer *fb = state->fb; + unsigned long eba, ubo, vbo; + int active; - ipu_plane->stride[0] = fb->pitches[0]; + eba = drm_plane_state_to_eba(state); switch (fb->pixel_format) { case DRM_FORMAT_YUV420: case DRM_FORMAT_YVU420: + if (old_state->fb) + break; + /* * Multiplanar formats have to meet the following restrictions: * - The (up to) three plane addresses are EBA, EBA+UBO, EBA+VBO @@ -117,59 +128,28 @@ int ipu_plane_set_base(struct ipu_plane *ipu_plane, struct drm_framebuffer *fb, * - Only EBA may be changed while scanout is active * - The strides of U and V planes must be identical. */ - ubo = cma_obj[1]->paddr + fb->offsets[1] + - fb->pitches[1] * y / 2 + x / 2 - eba; - vbo = cma_obj[2]->paddr + fb->offsets[2] + - fb->pitches[2] * y / 2 + x / 2 - eba; + ubo = drm_plane_state_to_ubo(state); + vbo = drm_plane_state_to_vbo(state); - if ((ubo & 0x7) || (vbo & 0x7)) { - DRM_DEBUG_KMS("U/V buffer offsets must be a multiple of 8.\n"); - return -EINVAL; - } - - if ((ubo > 0xfffff8) || (vbo > 0xfffff8)) { - DRM_DEBUG_KMS("U/V buffer offsets must be positive and not larger than 0xfffff8.\n"); - return -EINVAL; - } - - if (ipu_plane->enabled && ((ipu_plane->u_offset != ubo) || - (ipu_plane->v_offset != vbo))) { - DRM_DEBUG_KMS("U/V buffer offsets must not change while plane is enabled.\n"); - return -EINVAL; - } - - if (fb->pitches[1] != fb->pitches[2]) { - DRM_DEBUG_KMS("U/V pitches must be identical.\n"); - return -EINVAL; - } - - if (fb->pitches[1] < 1 || fb->pitches[1] > 16384) { - DRM_DEBUG_KMS("U/V pitches out of range.\n"); - return -EINVAL; - } - - if (ipu_plane->enabled && - (ipu_plane->stride[1] != fb->pitches[1])) { - DRM_DEBUG_KMS("U/V pitches must not change while plane is enabled.\n"); - return -EINVAL; - } - - ipu_plane->u_offset = ubo; - ipu_plane->v_offset = vbo; - ipu_plane->stride[1] = fb->pitches[1]; + if (fb->pixel_format == DRM_FORMAT_YUV420) + ipu_cpmem_set_yuv_planar_full(ipu_plane->ipu_ch, + fb->pitches[1], ubo, vbo); + else + ipu_cpmem_set_yuv_planar_full(ipu_plane->ipu_ch, + fb->pitches[1], vbo, ubo); dev_dbg(ipu_plane->base.dev->dev, - "phys = %pad %pad %pad, x = %d, y = %d", - &cma_obj[0]->paddr, &cma_obj[1]->paddr, - &cma_obj[2]->paddr, x, y); + "phy = %lu %lu %lu, x = %d, y = %d", eba, ubo, vbo, + state->src_x >> 16, state->src_y >> 16); break; default: - dev_dbg(ipu_plane->base.dev->dev, "phys = %pad, x = %d, y = %d", - &cma_obj[0]->paddr, x, y); + dev_dbg(ipu_plane->base.dev->dev, "phys = %lu, x = %d, y = %d", + eba, state->src_x >> 16, state->src_y >> 16); + break; } - if (ipu_plane->enabled) { + if (old_state->fb) { active = ipu_idmac_get_current_buffer(ipu_plane->ipu_ch); ipu_cpmem_set_buffer(ipu_plane->ipu_ch, !active, eba); ipu_idmac_select_buffer(ipu_plane->ipu_ch, !active); @@ -177,155 +157,6 @@ int ipu_plane_set_base(struct ipu_plane *ipu_plane, struct drm_framebuffer *fb, ipu_cpmem_set_buffer(ipu_plane->ipu_ch, 0, eba); ipu_cpmem_set_buffer(ipu_plane->ipu_ch, 1, eba); } - - /* cache offsets for subsequent pageflips */ - ipu_plane->x = x; - ipu_plane->y = y; - - return 0; -} - -int ipu_plane_mode_set(struct ipu_plane *ipu_plane, struct drm_crtc *crtc, - struct drm_display_mode *mode, - struct drm_framebuffer *fb, int crtc_x, int crtc_y, - unsigned int crtc_w, unsigned int crtc_h, - uint32_t src_x, uint32_t src_y, - uint32_t src_w, uint32_t src_h, bool interlaced) -{ - struct device *dev = ipu_plane->base.dev->dev; - int ret; - - /* no scaling */ - if (src_w != crtc_w || src_h != crtc_h) - return -EINVAL; - - /* clip to crtc bounds */ - if (crtc_x < 0) { - if (-crtc_x > crtc_w) - return -EINVAL; - src_x += -crtc_x; - src_w -= -crtc_x; - crtc_w -= -crtc_x; - crtc_x = 0; - } - if (crtc_y < 0) { - if (-crtc_y > crtc_h) - return -EINVAL; - src_y += -crtc_y; - src_h -= -crtc_y; - crtc_h -= -crtc_y; - crtc_y = 0; - } - if (crtc_x + crtc_w > mode->hdisplay) { - if (crtc_x > mode->hdisplay) - return -EINVAL; - crtc_w = mode->hdisplay - crtc_x; - src_w = crtc_w; - } - if (crtc_y + crtc_h > mode->vdisplay) { - if (crtc_y > mode->vdisplay) - return -EINVAL; - crtc_h = mode->vdisplay - crtc_y; - src_h = crtc_h; - } - /* full plane minimum width is 13 pixels */ - if (crtc_w < 13 && (ipu_plane->dp_flow != IPU_DP_FLOW_SYNC_FG)) - return -EINVAL; - if (crtc_h < 2) - return -EINVAL; - - /* - * since we cannot touch active IDMAC channels, we do not support - * resizing the enabled plane or changing its format - */ - if (ipu_plane->enabled) { - if (src_w != ipu_plane->w || src_h != ipu_plane->h || - fb->pixel_format != ipu_plane->base.fb->pixel_format) - return -EINVAL; - - return ipu_plane_set_base(ipu_plane, fb, src_x, src_y); - } - - switch (ipu_plane->dp_flow) { - case IPU_DP_FLOW_SYNC_BG: - ret = ipu_dp_setup_channel(ipu_plane->dp, - IPUV3_COLORSPACE_RGB, - IPUV3_COLORSPACE_RGB); - if (ret) { - dev_err(dev, - "initializing display processor failed with %d\n", - ret); - return ret; - } - ipu_dp_set_global_alpha(ipu_plane->dp, true, 0, true); - break; - case IPU_DP_FLOW_SYNC_FG: - ipu_dp_setup_channel(ipu_plane->dp, - ipu_drm_fourcc_to_colorspace(fb->pixel_format), - IPUV3_COLORSPACE_UNKNOWN); - ipu_dp_set_window_pos(ipu_plane->dp, crtc_x, crtc_y); - /* Enable local alpha on partial plane */ - switch (fb->pixel_format) { - case DRM_FORMAT_ARGB1555: - case DRM_FORMAT_ABGR1555: - case DRM_FORMAT_RGBA5551: - case DRM_FORMAT_BGRA5551: - case DRM_FORMAT_ARGB4444: - case DRM_FORMAT_ARGB8888: - case DRM_FORMAT_ABGR8888: - case DRM_FORMAT_RGBA8888: - case DRM_FORMAT_BGRA8888: - ipu_dp_set_global_alpha(ipu_plane->dp, false, 0, false); - break; - default: - break; - } - } - - ret = ipu_dmfc_alloc_bandwidth(ipu_plane->dmfc, - calc_bandwidth(crtc_w, crtc_h, - calc_vref(mode)), 64); - if (ret) { - dev_err(dev, "allocating dmfc bandwidth failed with %d\n", ret); - return ret; - } - - ipu_dmfc_config_wait4eot(ipu_plane->dmfc, crtc_w); - - ipu_cpmem_zero(ipu_plane->ipu_ch); - ipu_cpmem_set_resolution(ipu_plane->ipu_ch, src_w, src_h); - ret = ipu_cpmem_set_fmt(ipu_plane->ipu_ch, fb->pixel_format); - if (ret < 0) { - dev_err(dev, "unsupported pixel format 0x%08x\n", - fb->pixel_format); - return ret; - } - ipu_cpmem_set_high_priority(ipu_plane->ipu_ch); - ipu_idmac_set_double_buffer(ipu_plane->ipu_ch, 1); - ipu_cpmem_set_stride(ipu_plane->ipu_ch, fb->pitches[0]); - - ret = ipu_plane_set_base(ipu_plane, fb, src_x, src_y); - if (ret < 0) - return ret; - if (interlaced) - ipu_cpmem_interlaced_scan(ipu_plane->ipu_ch, fb->pitches[0]); - - if (fb->pixel_format == DRM_FORMAT_YUV420) { - ipu_cpmem_set_yuv_planar_full(ipu_plane->ipu_ch, - ipu_plane->stride[1], - ipu_plane->u_offset, - ipu_plane->v_offset); - } else if (fb->pixel_format == DRM_FORMAT_YVU420) { - ipu_cpmem_set_yuv_planar_full(ipu_plane->ipu_ch, - ipu_plane->stride[1], - ipu_plane->v_offset, - ipu_plane->u_offset); - } - - ipu_plane->w = src_w; - ipu_plane->h = src_h; - - return 0; } void ipu_plane_put_resources(struct ipu_plane *ipu_plane) @@ -372,7 +203,7 @@ err_out: return ret; } -void ipu_plane_enable(struct ipu_plane *ipu_plane) +static void ipu_plane_enable(struct ipu_plane *ipu_plane) { if (ipu_plane->dp) ipu_dp_enable(ipu_plane->ipu); @@ -380,14 +211,10 @@ void ipu_plane_enable(struct ipu_plane *ipu_plane) ipu_idmac_enable_channel(ipu_plane->ipu_ch); if (ipu_plane->dp) ipu_dp_enable_channel(ipu_plane->dp); - - ipu_plane->enabled = true; } -void ipu_plane_disable(struct ipu_plane *ipu_plane) +static void ipu_plane_disable(struct ipu_plane *ipu_plane) { - ipu_plane->enabled = false; - ipu_idmac_wait_busy(ipu_plane->ipu_ch, 50); if (ipu_plane->dp) @@ -398,74 +225,225 @@ void ipu_plane_disable(struct ipu_plane *ipu_plane) ipu_dp_disable(ipu_plane->ipu); } -/* - * drm_plane API - */ - -static int ipu_update_plane(struct drm_plane *plane, struct drm_crtc *crtc, - struct drm_framebuffer *fb, int crtc_x, int crtc_y, - unsigned int crtc_w, unsigned int crtc_h, - uint32_t src_x, uint32_t src_y, - uint32_t src_w, uint32_t src_h) +static int ipu_disable_plane(struct drm_plane *plane) { struct ipu_plane *ipu_plane = to_ipu_plane(plane); - int ret = 0; - - DRM_DEBUG_KMS("plane - %p\n", plane); - - if (!ipu_plane->enabled) - ret = ipu_plane_get_resources(ipu_plane); - if (ret < 0) - return ret; - - ret = ipu_plane_mode_set(ipu_plane, crtc, &crtc->hwmode, fb, - crtc_x, crtc_y, crtc_w, crtc_h, - src_x >> 16, src_y >> 16, src_w >> 16, src_h >> 16, - false); - if (ret < 0) { - ipu_plane_put_resources(ipu_plane); - return ret; - } - if (crtc != plane->crtc) - dev_dbg(plane->dev->dev, "crtc change: %p -> %p\n", - plane->crtc, crtc); + DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__); - if (!ipu_plane->enabled) - ipu_plane_enable(ipu_plane); + ipu_plane_disable(ipu_plane); return 0; } -static int ipu_disable_plane(struct drm_plane *plane) +static void ipu_plane_destroy(struct drm_plane *plane) { struct ipu_plane *ipu_plane = to_ipu_plane(plane); DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__); - if (ipu_plane->enabled) - ipu_plane_disable(ipu_plane); + ipu_disable_plane(plane); + drm_plane_cleanup(plane); + kfree(ipu_plane); +} - ipu_plane_put_resources(ipu_plane); +static const struct drm_plane_funcs ipu_plane_funcs = { + .update_plane = drm_atomic_helper_update_plane, + .disable_plane = drm_atomic_helper_disable_plane, + .destroy = ipu_plane_destroy, + .reset = drm_atomic_helper_plane_reset, + .atomic_duplicate_state = drm_atomic_helper_plane_duplicate_state, + .atomic_destroy_state = drm_atomic_helper_plane_destroy_state, +}; + +static int ipu_plane_atomic_check(struct drm_plane *plane, + struct drm_plane_state *state) +{ + struct drm_plane_state *old_state = plane->state; + struct drm_crtc_state *crtc_state; + struct device *dev = plane->dev->dev; + struct drm_framebuffer *fb = state->fb; + struct drm_framebuffer *old_fb = old_state->fb; + unsigned long eba, ubo, vbo, old_ubo, old_vbo; + + /* Ok to disable */ + if (!fb) + return 0; + + if (!state->crtc) + return -EINVAL; + + crtc_state = + drm_atomic_get_existing_crtc_state(state->state, state->crtc); + if (WARN_ON(!crtc_state)) + return -EINVAL; + + /* CRTC should be enabled */ + if (!crtc_state->enable) + return -EINVAL; + + /* no scaling */ + if (state->src_w >> 16 != state->crtc_w || + state->src_h >> 16 != state->crtc_h) + return -EINVAL; + + switch (plane->type) { + case DRM_PLANE_TYPE_PRIMARY: + /* full plane doesn't support partial off screen */ + if (state->crtc_x || state->crtc_y || + state->crtc_w != crtc_state->adjusted_mode.hdisplay || + state->crtc_h != crtc_state->adjusted_mode.vdisplay) + return -EINVAL; + + /* full plane minimum width is 13 pixels */ + if (state->crtc_w < 13) + return -EINVAL; + break; + case DRM_PLANE_TYPE_OVERLAY: + if (state->crtc_x < 0 || state->crtc_y < 0) + return -EINVAL; + + if (state->crtc_x + state->crtc_w > + crtc_state->adjusted_mode.hdisplay) + return -EINVAL; + if (state->crtc_y + state->crtc_h > + crtc_state->adjusted_mode.vdisplay) + return -EINVAL; + break; + default: + dev_warn(dev, "Unsupported plane type\n"); + return -EINVAL; + } + + if (state->crtc_h < 2) + return -EINVAL; + + /* + * since we cannot touch active IDMAC channels, we do not support + * resizing the enabled plane or changing its format + */ + if (old_fb && (state->src_w != old_state->src_w || + state->src_h != old_state->src_h || + fb->pixel_format != old_fb->pixel_format)) + return -EINVAL; + + eba = drm_plane_state_to_eba(state); + + if (eba & 0x7) + return -EINVAL; + + if (fb->pitches[0] < 1 || fb->pitches[0] > 16384) + return -EINVAL; + + if (old_fb && fb->pitches[0] != old_fb->pitches[0]) + return -EINVAL; + + switch (fb->pixel_format) { + case DRM_FORMAT_YUV420: + case DRM_FORMAT_YVU420: + /* + * Multiplanar formats have to meet the following restrictions: + * - The (up to) three plane addresses are EBA, EBA+UBO, EBA+VBO + * - EBA, UBO and VBO are a multiple of 8 + * - UBO and VBO are unsigned and not larger than 0xfffff8 + * - Only EBA may be changed while scanout is active + * - The strides of U and V planes must be identical. + */ + ubo = drm_plane_state_to_ubo(state); + vbo = drm_plane_state_to_vbo(state); + + if ((ubo & 0x7) || (vbo & 0x7)) + return -EINVAL; + + if ((ubo > 0xfffff8) || (vbo > 0xfffff8)) + return -EINVAL; + + if (old_fb) { + old_ubo = drm_plane_state_to_ubo(old_state); + old_vbo = drm_plane_state_to_vbo(old_state); + if (ubo != old_ubo || vbo != old_vbo) + return -EINVAL; + } + + if (fb->pitches[1] != fb->pitches[2]) + return -EINVAL; + + if (fb->pitches[1] < 1 || fb->pitches[1] > 16384) + return -EINVAL; + + if (old_fb && old_fb->pitches[1] != fb->pitches[1]) + return -EINVAL; + } return 0; } -static void ipu_plane_destroy(struct drm_plane *plane) +static void ipu_plane_atomic_disable(struct drm_plane *plane, + struct drm_plane_state *old_state) +{ + ipu_disable_plane(plane); +} + +static void ipu_plane_atomic_update(struct drm_plane *plane, + struct drm_plane_state *old_state) { struct ipu_plane *ipu_plane = to_ipu_plane(plane); + struct drm_plane_state *state = plane->state; + enum ipu_color_space ics; - DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__); + if (old_state->fb) { + ipu_plane_atomic_set_base(ipu_plane, old_state); + return; + } - ipu_disable_plane(plane); - drm_plane_cleanup(plane); - kfree(ipu_plane); + switch (ipu_plane->dp_flow) { + case IPU_DP_FLOW_SYNC_BG: + ipu_dp_setup_channel(ipu_plane->dp, + IPUV3_COLORSPACE_RGB, + IPUV3_COLORSPACE_RGB); + ipu_dp_set_global_alpha(ipu_plane->dp, true, 0, true); + break; + case IPU_DP_FLOW_SYNC_FG: + ics = ipu_drm_fourcc_to_colorspace(state->fb->pixel_format); + ipu_dp_setup_channel(ipu_plane->dp, ics, + IPUV3_COLORSPACE_UNKNOWN); + ipu_dp_set_window_pos(ipu_plane->dp, state->crtc_x, + state->crtc_y); + /* Enable local alpha on partial plane */ + switch (state->fb->pixel_format) { + case DRM_FORMAT_ARGB1555: + case DRM_FORMAT_ABGR1555: + case DRM_FORMAT_RGBA5551: + case DRM_FORMAT_BGRA5551: + case DRM_FORMAT_ARGB4444: + case DRM_FORMAT_ARGB8888: + case DRM_FORMAT_ABGR8888: + case DRM_FORMAT_RGBA8888: + case DRM_FORMAT_BGRA8888: + ipu_dp_set_global_alpha(ipu_plane->dp, false, 0, false); + break; + default: + break; + } + } + + ipu_dmfc_config_wait4eot(ipu_plane->dmfc, state->crtc_w); + + ipu_cpmem_zero(ipu_plane->ipu_ch); + ipu_cpmem_set_resolution(ipu_plane->ipu_ch, state->src_w >> 16, + state->src_h >> 16); + ipu_cpmem_set_fmt(ipu_plane->ipu_ch, state->fb->pixel_format); + ipu_cpmem_set_high_priority(ipu_plane->ipu_ch); + ipu_idmac_set_double_buffer(ipu_plane->ipu_ch, 1); + ipu_cpmem_set_stride(ipu_plane->ipu_ch, state->fb->pitches[0]); + ipu_plane_atomic_set_base(ipu_plane, old_state); + ipu_plane_enable(ipu_plane); } -static const struct drm_plane_funcs ipu_plane_funcs = { - .update_plane = ipu_update_plane, - .disable_plane = ipu_disable_plane, - .destroy = ipu_plane_destroy, +static const struct drm_plane_helper_funcs ipu_plane_helper_funcs = { + .atomic_check = ipu_plane_atomic_check, + .atomic_disable = ipu_plane_atomic_disable, + .atomic_update = ipu_plane_atomic_update, }; struct ipu_plane *ipu_plane_init(struct drm_device *dev, struct ipu_soc *ipu, @@ -498,5 +476,7 @@ struct ipu_plane *ipu_plane_init(struct drm_device *dev, struct ipu_soc *ipu, return ERR_PTR(ret); } + drm_plane_helper_add(&ipu_plane->base, &ipu_plane_helper_funcs); + return ipu_plane; } diff --git a/drivers/gpu/drm/imx/ipuv3-plane.h b/drivers/gpu/drm/imx/ipuv3-plane.h index 4448fd4ad4eb..338b88a74eb6 100644 --- a/drivers/gpu/drm/imx/ipuv3-plane.h +++ b/drivers/gpu/drm/imx/ipuv3-plane.h @@ -23,17 +23,6 @@ struct ipu_plane { int dma; int dp_flow; - - int x; - int y; - int w; - int h; - - unsigned int u_offset; - unsigned int v_offset; - unsigned int stride[2]; - - bool enabled; }; struct ipu_plane *ipu_plane_init(struct drm_device *dev, struct ipu_soc *ipu, @@ -48,11 +37,6 @@ int ipu_plane_mode_set(struct ipu_plane *plane, struct drm_crtc *crtc, uint32_t src_x, uint32_t src_y, uint32_t src_w, uint32_t src_h, bool interlaced); -void ipu_plane_enable(struct ipu_plane *plane); -void ipu_plane_disable(struct ipu_plane *plane); -int ipu_plane_set_base(struct ipu_plane *plane, struct drm_framebuffer *fb, - int x, int y); - int ipu_plane_get_resources(struct ipu_plane *plane); void ipu_plane_put_resources(struct ipu_plane *plane); diff --git a/drivers/gpu/drm/imx/parallel-display.c b/drivers/gpu/drm/imx/parallel-display.c index 2d1fd02cd3d6..1dad297b01fd 100644 --- a/drivers/gpu/drm/imx/parallel-display.c +++ b/drivers/gpu/drm/imx/parallel-display.c @@ -16,6 +16,7 @@ #include <linux/component.h> #include <linux/module.h> #include <drm/drmP.h> +#include <drm/drm_atomic_helper.h> #include <drm/drm_fb_helper.h> #include <drm/drm_crtc_helper.h> #include <drm/drm_panel.h> @@ -25,9 +26,6 @@ #include "imx-drm.h" -#define con_to_imxpd(x) container_of(x, struct imx_parallel_display, connector) -#define enc_to_imxpd(x) container_of(x, struct imx_parallel_display, encoder) - struct imx_parallel_display { struct drm_connector connector; struct drm_encoder encoder; @@ -37,8 +35,19 @@ struct imx_parallel_display { u32 bus_format; struct drm_display_mode mode; struct drm_panel *panel; + struct drm_bridge *bridge; }; +static inline struct imx_parallel_display *con_to_imxpd(struct drm_connector *c) +{ + return container_of(c, struct imx_parallel_display, connector); +} + +static inline struct imx_parallel_display *enc_to_imxpd(struct drm_encoder *e) +{ + return container_of(e, struct imx_parallel_display, encoder); +} + static enum drm_connector_status imx_pd_connector_detect( struct drm_connector *connector, bool force) { @@ -53,11 +62,7 @@ static int imx_pd_connector_get_modes(struct drm_connector *connector) if (imxpd->panel && imxpd->panel->funcs && imxpd->panel->funcs->get_modes) { - struct drm_display_info *di = &connector->display_info; - num_modes = imxpd->panel->funcs->get_modes(imxpd->panel); - if (!imxpd->bus_format && di->num_bus_formats) - imxpd->bus_format = di->bus_formats[0]; if (num_modes > 0) return num_modes; } @@ -69,10 +74,16 @@ static int imx_pd_connector_get_modes(struct drm_connector *connector) if (np) { struct drm_display_mode *mode = drm_mode_create(connector->dev); + int ret; if (!mode) return -EINVAL; - of_get_drm_display_mode(np, &imxpd->mode, OF_USE_NATIVE_MODE); + + ret = of_get_drm_display_mode(np, &imxpd->mode, + OF_USE_NATIVE_MODE); + if (ret) + return ret; + drm_mode_copy(mode, &imxpd->mode); mode->type |= DRM_MODE_TYPE_DRIVER | DRM_MODE_TYPE_PREFERRED, drm_mode_probed_add(connector, mode); @@ -90,24 +101,7 @@ static struct drm_encoder *imx_pd_connector_best_encoder( return &imxpd->encoder; } -static void imx_pd_encoder_dpms(struct drm_encoder *encoder, int mode) -{ - struct imx_parallel_display *imxpd = enc_to_imxpd(encoder); - - if (mode != DRM_MODE_DPMS_ON) - drm_panel_disable(imxpd->panel); - else - drm_panel_enable(imxpd->panel); -} - -static void imx_pd_encoder_prepare(struct drm_encoder *encoder) -{ - struct imx_parallel_display *imxpd = enc_to_imxpd(encoder); - imx_drm_set_bus_config(encoder, imxpd->bus_format, 2, 3, - imxpd->connector.display_info.bus_flags); -} - -static void imx_pd_encoder_commit(struct drm_encoder *encoder) +static void imx_pd_encoder_enable(struct drm_encoder *encoder) { struct imx_parallel_display *imxpd = enc_to_imxpd(encoder); @@ -115,12 +109,6 @@ static void imx_pd_encoder_commit(struct drm_encoder *encoder) drm_panel_enable(imxpd->panel); } -static void imx_pd_encoder_mode_set(struct drm_encoder *encoder, - struct drm_display_mode *orig_mode, - struct drm_display_mode *mode) -{ -} - static void imx_pd_encoder_disable(struct drm_encoder *encoder) { struct imx_parallel_display *imxpd = enc_to_imxpd(encoder); @@ -129,11 +117,33 @@ static void imx_pd_encoder_disable(struct drm_encoder *encoder) drm_panel_unprepare(imxpd->panel); } +static int imx_pd_encoder_atomic_check(struct drm_encoder *encoder, + struct drm_crtc_state *crtc_state, + struct drm_connector_state *conn_state) +{ + struct imx_crtc_state *imx_crtc_state = to_imx_crtc_state(crtc_state); + struct drm_display_info *di = &conn_state->connector->display_info; + struct imx_parallel_display *imxpd = enc_to_imxpd(encoder); + + imx_crtc_state->bus_flags = di->bus_flags; + if (!imxpd->bus_format && di->num_bus_formats) + imx_crtc_state->bus_format = di->bus_formats[0]; + else + imx_crtc_state->bus_format = imxpd->bus_format; + imx_crtc_state->di_hsync_pin = 2; + imx_crtc_state->di_vsync_pin = 3; + + return 0; +} + static const struct drm_connector_funcs imx_pd_connector_funcs = { - .dpms = drm_helper_connector_dpms, + .dpms = drm_atomic_helper_connector_dpms, .fill_modes = drm_helper_probe_single_connector_modes, .detect = imx_pd_connector_detect, .destroy = imx_drm_connector_destroy, + .reset = drm_atomic_helper_connector_reset, + .atomic_duplicate_state = drm_atomic_helper_connector_duplicate_state, + .atomic_destroy_state = drm_atomic_helper_connector_destroy_state, }; static const struct drm_connector_helper_funcs imx_pd_connector_helper_funcs = { @@ -146,20 +156,18 @@ static const struct drm_encoder_funcs imx_pd_encoder_funcs = { }; static const struct drm_encoder_helper_funcs imx_pd_encoder_helper_funcs = { - .dpms = imx_pd_encoder_dpms, - .prepare = imx_pd_encoder_prepare, - .commit = imx_pd_encoder_commit, - .mode_set = imx_pd_encoder_mode_set, + .enable = imx_pd_encoder_enable, .disable = imx_pd_encoder_disable, + .atomic_check = imx_pd_encoder_atomic_check, }; static int imx_pd_register(struct drm_device *drm, struct imx_parallel_display *imxpd) { + struct drm_encoder *encoder = &imxpd->encoder; int ret; - ret = imx_drm_encoder_parse_of(drm, &imxpd->encoder, - imxpd->dev->of_node); + ret = imx_drm_encoder_parse_of(drm, encoder, imxpd->dev->of_node); if (ret) return ret; @@ -170,19 +178,33 @@ static int imx_pd_register(struct drm_device *drm, */ imxpd->connector.dpms = DRM_MODE_DPMS_OFF; - drm_encoder_helper_add(&imxpd->encoder, &imx_pd_encoder_helper_funcs); - drm_encoder_init(drm, &imxpd->encoder, &imx_pd_encoder_funcs, + drm_encoder_helper_add(encoder, &imx_pd_encoder_helper_funcs); + drm_encoder_init(drm, encoder, &imx_pd_encoder_funcs, DRM_MODE_ENCODER_NONE, NULL); - drm_connector_helper_add(&imxpd->connector, - &imx_pd_connector_helper_funcs); - drm_connector_init(drm, &imxpd->connector, &imx_pd_connector_funcs, - DRM_MODE_CONNECTOR_VGA); + if (!imxpd->bridge) { + drm_connector_helper_add(&imxpd->connector, + &imx_pd_connector_helper_funcs); + drm_connector_init(drm, &imxpd->connector, + &imx_pd_connector_funcs, + DRM_MODE_CONNECTOR_VGA); + } if (imxpd->panel) drm_panel_attach(imxpd->panel, &imxpd->connector); - drm_mode_connector_attach_encoder(&imxpd->connector, &imxpd->encoder); + if (imxpd->bridge) { + imxpd->bridge->encoder = encoder; + encoder->bridge = imxpd->bridge; + ret = drm_bridge_attach(drm, imxpd->bridge); + if (ret < 0) { + dev_err(imxpd->dev, "failed to attach bridge: %d\n", + ret); + return ret; + } + } else { + drm_mode_connector_attach_encoder(&imxpd->connector, encoder); + } return 0; } @@ -195,6 +217,7 @@ static int imx_pd_bind(struct device *dev, struct device *master, void *data) const u8 *edidp; struct imx_parallel_display *imxpd; int ret; + u32 bus_format = 0; const char *fmt; imxpd = devm_kzalloc(dev, sizeof(*imxpd), GFP_KERNEL); @@ -208,14 +231,15 @@ static int imx_pd_bind(struct device *dev, struct device *master, void *data) ret = of_property_read_string(np, "interface-pix-fmt", &fmt); if (!ret) { if (!strcmp(fmt, "rgb24")) - imxpd->bus_format = MEDIA_BUS_FMT_RGB888_1X24; + bus_format = MEDIA_BUS_FMT_RGB888_1X24; else if (!strcmp(fmt, "rgb565")) - imxpd->bus_format = MEDIA_BUS_FMT_RGB565_1X16; + bus_format = MEDIA_BUS_FMT_RGB565_1X16; else if (!strcmp(fmt, "bgr666")) - imxpd->bus_format = MEDIA_BUS_FMT_RGB666_1X18; + bus_format = MEDIA_BUS_FMT_RGB666_1X18; else if (!strcmp(fmt, "lvds666")) - imxpd->bus_format = MEDIA_BUS_FMT_RGB666_1X24_CPADHI; + bus_format = MEDIA_BUS_FMT_RGB666_1X24_CPADHI; } + imxpd->bus_format = bus_format; /* port@1 is the output port */ ep = of_graph_get_endpoint_by_regs(np, 1, -1); @@ -223,13 +247,30 @@ static int imx_pd_bind(struct device *dev, struct device *master, void *data) struct device_node *remote; remote = of_graph_get_remote_port_parent(ep); + if (!remote) { + dev_warn(dev, "endpoint %s not connected\n", + ep->full_name); + of_node_put(ep); + return -ENODEV; + } of_node_put(ep); - if (remote) { - imxpd->panel = of_drm_find_panel(remote); - of_node_put(remote); + + imxpd->panel = of_drm_find_panel(remote); + if (imxpd->panel) { + dev_dbg(dev, "found panel %s\n", remote->full_name); + } else { + imxpd->bridge = of_drm_find_bridge(remote); + if (imxpd->bridge) + dev_dbg(dev, "found bridge %s\n", + remote->full_name); } - if (!imxpd->panel) + if (!imxpd->panel && !imxpd->bridge) { + dev_dbg(dev, "waiting for panel or bridge %s\n", + remote->full_name); + of_node_put(remote); return -EPROBE_DEFER; + } + of_node_put(remote); } imxpd->dev = dev; diff --git a/drivers/gpu/drm/mediatek/mtk_hdmi.c b/drivers/gpu/drm/mediatek/mtk_hdmi.c index ba812ef2c9d1..334562d06731 100644 --- a/drivers/gpu/drm/mediatek/mtk_hdmi.c +++ b/drivers/gpu/drm/mediatek/mtk_hdmi.c @@ -1535,7 +1535,7 @@ static int mtk_hdmi_dt_parse_pdata(struct mtk_hdmi *hdmi, * HDMI audio codec callbacks */ -static int mtk_hdmi_audio_hw_params(struct device *dev, +static int mtk_hdmi_audio_hw_params(struct device *dev, void *data, struct hdmi_codec_daifmt *daifmt, struct hdmi_codec_params *params) { @@ -1604,7 +1604,7 @@ static int mtk_hdmi_audio_hw_params(struct device *dev, return 0; } -static int mtk_hdmi_audio_startup(struct device *dev) +static int mtk_hdmi_audio_startup(struct device *dev, void *data) { struct mtk_hdmi *hdmi = dev_get_drvdata(dev); @@ -1615,7 +1615,7 @@ static int mtk_hdmi_audio_startup(struct device *dev) return 0; } -static void mtk_hdmi_audio_shutdown(struct device *dev) +static void mtk_hdmi_audio_shutdown(struct device *dev, void *data) { struct mtk_hdmi *hdmi = dev_get_drvdata(dev); @@ -1624,7 +1624,7 @@ static void mtk_hdmi_audio_shutdown(struct device *dev) mtk_hdmi_audio_disable(hdmi); } -int mtk_hdmi_audio_digital_mute(struct device *dev, bool enable) +int mtk_hdmi_audio_digital_mute(struct device *dev, void *data, bool enable) { struct mtk_hdmi *hdmi = dev_get_drvdata(dev); @@ -1638,7 +1638,7 @@ int mtk_hdmi_audio_digital_mute(struct device *dev, bool enable) return 0; } -static int mtk_hdmi_audio_get_eld(struct device *dev, uint8_t *buf, size_t len) +static int mtk_hdmi_audio_get_eld(struct device *dev, void *data, uint8_t *buf, size_t len) { struct mtk_hdmi *hdmi = dev_get_drvdata(dev); diff --git a/drivers/gpu/drm/mediatek/mtk_mipi_tx.c b/drivers/gpu/drm/mediatek/mtk_mipi_tx.c index cf8f38d39e10..1c366f8cb2d0 100644 --- a/drivers/gpu/drm/mediatek/mtk_mipi_tx.c +++ b/drivers/gpu/drm/mediatek/mtk_mipi_tx.c @@ -431,7 +431,7 @@ static int mtk_mipi_tx_probe(struct platform_device *pdev) phy_set_drvdata(phy, mipi_tx); phy_provider = devm_of_phy_provider_register(dev, of_phy_simple_xlate); - if (IS_ERR(phy)) { + if (IS_ERR(phy_provider)) { ret = PTR_ERR(phy_provider); return ret; } diff --git a/drivers/gpu/drm/mgag200/Kconfig b/drivers/gpu/drm/mgag200/Kconfig index 3a1c5fbae54a..520e5e668d6c 100644 --- a/drivers/gpu/drm/mgag200/Kconfig +++ b/drivers/gpu/drm/mgag200/Kconfig @@ -1,11 +1,7 @@ config DRM_MGAG200 tristate "Kernel modesetting driver for MGA G200 server engines" depends on DRM && PCI - select FB_SYS_FILLRECT - select FB_SYS_COPYAREA - select FB_SYS_IMAGEBLIT select DRM_KMS_HELPER - select DRM_KMS_FB_HELPER select DRM_TTM help This is a KMS driver for the MGA G200 server chips, it diff --git a/drivers/gpu/drm/mgag200/mgag200_main.c b/drivers/gpu/drm/mgag200/mgag200_main.c index 615cbb08ba29..13798b3e6beb 100644 --- a/drivers/gpu/drm/mgag200/mgag200_main.c +++ b/drivers/gpu/drm/mgag200/mgag200_main.c @@ -17,8 +17,8 @@ static void mga_user_framebuffer_destroy(struct drm_framebuffer *fb) { struct mga_framebuffer *mga_fb = to_mga_framebuffer(fb); - if (mga_fb->obj) - drm_gem_object_unreference_unlocked(mga_fb->obj); + + drm_gem_object_unreference_unlocked(mga_fb->obj); drm_framebuffer_cleanup(fb); kfree(fb); } diff --git a/drivers/gpu/drm/mgag200/mgag200_ttm.c b/drivers/gpu/drm/mgag200/mgag200_ttm.c index 9d5083d0f1ee..68268e55d595 100644 --- a/drivers/gpu/drm/mgag200/mgag200_ttm.c +++ b/drivers/gpu/drm/mgag200/mgag200_ttm.c @@ -186,17 +186,6 @@ static void mgag200_ttm_io_mem_free(struct ttm_bo_device *bdev, struct ttm_mem_r { } -static int mgag200_bo_move(struct ttm_buffer_object *bo, - bool evict, bool interruptible, - bool no_wait_gpu, - struct ttm_mem_reg *new_mem) -{ - int r; - r = ttm_bo_move_memcpy(bo, evict, no_wait_gpu, new_mem); - return r; -} - - static void mgag200_ttm_backend_destroy(struct ttm_tt *tt) { ttm_tt_fini(tt); @@ -241,7 +230,7 @@ struct ttm_bo_driver mgag200_bo_driver = { .ttm_tt_unpopulate = mgag200_ttm_tt_unpopulate, .init_mem_type = mgag200_bo_init_mem_type, .evict_flags = mgag200_bo_evict_flags, - .move = mgag200_bo_move, + .move = NULL, .verify_access = mgag200_bo_verify_access, .io_mem_reserve = &mgag200_ttm_io_mem_reserve, .io_mem_free = &mgag200_ttm_io_mem_free, diff --git a/drivers/gpu/drm/msm/Kconfig b/drivers/gpu/drm/msm/Kconfig index 167a4971f47c..7c7a0314a756 100644 --- a/drivers/gpu/drm/msm/Kconfig +++ b/drivers/gpu/drm/msm/Kconfig @@ -10,6 +10,7 @@ config DRM_MSM select SHMEM select TMPFS select QCOM_SCM + select SND_SOC_HDMI_CODEC if SND_SOC default y help DRM/KMS driver for MSM/snapdragon. diff --git a/drivers/gpu/drm/msm/Makefile b/drivers/gpu/drm/msm/Makefile index 60cb02624dc0..4e2806cf778c 100644 --- a/drivers/gpu/drm/msm/Makefile +++ b/drivers/gpu/drm/msm/Makefile @@ -35,6 +35,7 @@ msm-y := \ mdp/mdp5/mdp5_crtc.o \ mdp/mdp5/mdp5_encoder.o \ mdp/mdp5/mdp5_irq.o \ + mdp/mdp5/mdp5_mdss.o \ mdp/mdp5/mdp5_kms.o \ mdp/mdp5/mdp5_plane.o \ mdp/mdp5/mdp5_smp.o \ @@ -45,6 +46,7 @@ msm-y := \ msm_fence.o \ msm_gem.o \ msm_gem_prime.o \ + msm_gem_shrinker.o \ msm_gem_submit.o \ msm_gpu.o \ msm_iommu.o \ diff --git a/drivers/gpu/drm/msm/adreno/adreno_gpu.c b/drivers/gpu/drm/msm/adreno/adreno_gpu.c index fbe304ee6c80..f386f463278d 100644 --- a/drivers/gpu/drm/msm/adreno/adreno_gpu.c +++ b/drivers/gpu/drm/msm/adreno/adreno_gpu.c @@ -139,7 +139,7 @@ void adreno_submit(struct msm_gpu *gpu, struct msm_gem_submit *submit, struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu); struct msm_drm_private *priv = gpu->dev->dev_private; struct msm_ringbuffer *ring = gpu->rb; - unsigned i, ibs = 0; + unsigned i; for (i = 0; i < submit->nr_cmds; i++) { switch (submit->cmd[i].type) { @@ -155,18 +155,11 @@ void adreno_submit(struct msm_gpu *gpu, struct msm_gem_submit *submit, CP_INDIRECT_BUFFER_PFE : CP_INDIRECT_BUFFER_PFD, 2); OUT_RING(ring, submit->cmd[i].iova); OUT_RING(ring, submit->cmd[i].size); - ibs++; + OUT_PKT2(ring); break; } } - /* on a320, at least, we seem to need to pad things out to an - * even number of qwords to avoid issue w/ CP hanging on wrap- - * around: - */ - if (ibs % 2) - OUT_PKT2(ring); - OUT_PKT0(ring, REG_AXXX_CP_SCRATCH_REG2, 1); OUT_RING(ring, submit->fence->seqno); @@ -407,8 +400,8 @@ int adreno_gpu_init(struct drm_device *drm, struct platform_device *pdev, return ret; } - adreno_gpu->memptrs = msm_gem_vaddr(adreno_gpu->memptrs_bo); - if (!adreno_gpu->memptrs) { + adreno_gpu->memptrs = msm_gem_get_vaddr(adreno_gpu->memptrs_bo); + if (IS_ERR(adreno_gpu->memptrs)) { dev_err(drm->dev, "could not vmap memptrs\n"); return -ENOMEM; } @@ -426,8 +419,12 @@ int adreno_gpu_init(struct drm_device *drm, struct platform_device *pdev, void adreno_gpu_cleanup(struct adreno_gpu *gpu) { if (gpu->memptrs_bo) { + if (gpu->memptrs) + msm_gem_put_vaddr(gpu->memptrs_bo); + if (gpu->memptrs_iova) msm_gem_put_iova(gpu->memptrs_bo, gpu->base.id); + drm_gem_object_unreference_unlocked(gpu->memptrs_bo); } release_firmware(gpu->pm4); diff --git a/drivers/gpu/drm/msm/dsi/dsi.c b/drivers/gpu/drm/msm/dsi/dsi.c index 6edcd6f57e70..ec572f8389ed 100644 --- a/drivers/gpu/drm/msm/dsi/dsi.c +++ b/drivers/gpu/drm/msm/dsi/dsi.c @@ -29,7 +29,7 @@ static int dsi_get_phy(struct msm_dsi *msm_dsi) struct platform_device *phy_pdev; struct device_node *phy_node; - phy_node = of_parse_phandle(pdev->dev.of_node, "qcom,dsi-phy", 0); + phy_node = of_parse_phandle(pdev->dev.of_node, "phys", 0); if (!phy_node) { dev_err(&pdev->dev, "cannot find phy device\n"); return -ENXIO; diff --git a/drivers/gpu/drm/msm/dsi/dsi_cfg.c b/drivers/gpu/drm/msm/dsi/dsi_cfg.c index 93c1ee094eac..63436d8ee470 100644 --- a/drivers/gpu/drm/msm/dsi/dsi_cfg.c +++ b/drivers/gpu/drm/msm/dsi/dsi_cfg.c @@ -29,6 +29,8 @@ static const struct msm_dsi_config apq8064_dsi_cfg = { }, .bus_clk_names = dsi_v2_bus_clk_names, .num_bus_clks = ARRAY_SIZE(dsi_v2_bus_clk_names), + .io_start = { 0x4700000, 0x5800000 }, + .num_dsi = 2, }; static const char * const dsi_6g_bus_clk_names[] = { @@ -48,6 +50,8 @@ static const struct msm_dsi_config msm8974_apq8084_dsi_cfg = { }, .bus_clk_names = dsi_6g_bus_clk_names, .num_bus_clks = ARRAY_SIZE(dsi_6g_bus_clk_names), + .io_start = { 0xfd922800, 0xfd922b00 }, + .num_dsi = 2, }; static const char * const dsi_8916_bus_clk_names[] = { @@ -66,6 +70,8 @@ static const struct msm_dsi_config msm8916_dsi_cfg = { }, .bus_clk_names = dsi_8916_bus_clk_names, .num_bus_clks = ARRAY_SIZE(dsi_8916_bus_clk_names), + .io_start = { 0x1a98000 }, + .num_dsi = 1, }; static const struct msm_dsi_config msm8994_dsi_cfg = { @@ -84,6 +90,8 @@ static const struct msm_dsi_config msm8994_dsi_cfg = { }, .bus_clk_names = dsi_6g_bus_clk_names, .num_bus_clks = ARRAY_SIZE(dsi_6g_bus_clk_names), + .io_start = { 0xfd998000, 0xfd9a0000 }, + .num_dsi = 2, }; static const struct msm_dsi_cfg_handler dsi_cfg_handlers[] = { diff --git a/drivers/gpu/drm/msm/dsi/dsi_cfg.h b/drivers/gpu/drm/msm/dsi/dsi_cfg.h index a68c836744a3..eeacc3232494 100644 --- a/drivers/gpu/drm/msm/dsi/dsi_cfg.h +++ b/drivers/gpu/drm/msm/dsi/dsi_cfg.h @@ -34,6 +34,8 @@ struct msm_dsi_config { struct dsi_reg_config reg_cfg; const char * const *bus_clk_names; const int num_bus_clks; + const resource_size_t io_start[DSI_MAX]; + const int num_dsi; }; struct msm_dsi_cfg_handler { diff --git a/drivers/gpu/drm/msm/dsi/dsi_host.c b/drivers/gpu/drm/msm/dsi/dsi_host.c index a3e47ad83eb3..f05ed0e1f3d6 100644 --- a/drivers/gpu/drm/msm/dsi/dsi_host.c +++ b/drivers/gpu/drm/msm/dsi/dsi_host.c @@ -1066,7 +1066,7 @@ static int dsi_cmd_dma_add(struct msm_dsi_host *msm_host, } if (cfg_hnd->major == MSM_DSI_VER_MAJOR_6G) { - data = msm_gem_vaddr(msm_host->tx_gem_obj); + data = msm_gem_get_vaddr(msm_host->tx_gem_obj); if (IS_ERR(data)) { ret = PTR_ERR(data); pr_err("%s: get vaddr failed, %d\n", __func__, ret); @@ -1094,6 +1094,9 @@ static int dsi_cmd_dma_add(struct msm_dsi_host *msm_host, if (packet.size < len) memset(data + packet.size, 0xff, len - packet.size); + if (cfg_hnd->major == MSM_DSI_VER_MAJOR_6G) + msm_gem_put_vaddr(msm_host->tx_gem_obj); + return len; } @@ -1543,7 +1546,7 @@ static int dsi_host_parse_lane_data(struct msm_dsi_host *msm_host, u32 lane_map[4]; int ret, i, len, num_lanes; - prop = of_find_property(ep, "qcom,data-lane-map", &len); + prop = of_find_property(ep, "data-lanes", &len); if (!prop) { dev_dbg(dev, "failed to find data lane mapping\n"); return -EINVAL; @@ -1558,7 +1561,7 @@ static int dsi_host_parse_lane_data(struct msm_dsi_host *msm_host, msm_host->num_data_lanes = num_lanes; - ret = of_property_read_u32_array(ep, "qcom,data-lane-map", lane_map, + ret = of_property_read_u32_array(ep, "data-lanes", lane_map, num_lanes); if (ret) { dev_err(dev, "failed to read lane data\n"); @@ -1573,8 +1576,19 @@ static int dsi_host_parse_lane_data(struct msm_dsi_host *msm_host, const int *swap = supported_data_lane_swaps[i]; int j; + /* + * the data-lanes array we get from DT has a logical->physical + * mapping. The "data lane swap" register field represents + * supported configurations in a physical->logical mapping. + * Translate the DT mapping to what we understand and find a + * configuration that works. + */ for (j = 0; j < num_lanes; j++) { - if (swap[j] != lane_map[j]) + if (lane_map[j] < 0 || lane_map[j] > 3) + dev_err(dev, "bad physical lane entry %u\n", + lane_map[j]); + + if (swap[lane_map[j]] != j) break; } @@ -1594,20 +1608,13 @@ static int dsi_host_parse_dt(struct msm_dsi_host *msm_host) struct device_node *endpoint, *device_node; int ret; - ret = of_property_read_u32(np, "qcom,dsi-host-index", &msm_host->id); - if (ret) { - dev_err(dev, "%s: host index not specified, ret=%d\n", - __func__, ret); - return ret; - } - /* - * Get the first endpoint node. In our case, dsi has one output port - * to which the panel is connected. Don't return an error if a port - * isn't defined. It's possible that there is nothing connected to - * the dsi output. + * Get the endpoint of the output port of the DSI host. In our case, + * this is mapped to port number with reg = 1. Don't return an error if + * the remote endpoint isn't defined. It's possible that there is + * nothing connected to the dsi output. */ - endpoint = of_graph_get_next_endpoint(np, NULL); + endpoint = of_graph_get_endpoint_by_regs(np, 1, -1); if (!endpoint) { dev_dbg(dev, "%s: no endpoint\n", __func__); return 0; @@ -1648,6 +1655,25 @@ err: return ret; } +static int dsi_host_get_id(struct msm_dsi_host *msm_host) +{ + struct platform_device *pdev = msm_host->pdev; + const struct msm_dsi_config *cfg = msm_host->cfg_hnd->cfg; + struct resource *res; + int i; + + res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "dsi_ctrl"); + if (!res) + return -EINVAL; + + for (i = 0; i < cfg->num_dsi; i++) { + if (cfg->io_start[i] == res->start) + return i; + } + + return -EINVAL; +} + int msm_dsi_host_init(struct msm_dsi *msm_dsi) { struct msm_dsi_host *msm_host = NULL; @@ -1684,6 +1710,13 @@ int msm_dsi_host_init(struct msm_dsi *msm_dsi) goto fail; } + msm_host->id = dsi_host_get_id(msm_host); + if (msm_host->id < 0) { + ret = msm_host->id; + pr_err("%s: unable to identify DSI host index\n", __func__); + goto fail; + } + /* fixup base address by io offset */ msm_host->ctrl_base += msm_host->cfg_hnd->cfg->io_offset; @@ -2245,9 +2278,9 @@ int msm_dsi_host_set_display_mode(struct mipi_dsi_host *host, } msm_host->mode = drm_mode_duplicate(msm_host->dev, mode); - if (IS_ERR(msm_host->mode)) { + if (!msm_host->mode) { pr_err("%s: cannot duplicate mode\n", __func__); - return PTR_ERR(msm_host->mode); + return -ENOMEM; } return 0; diff --git a/drivers/gpu/drm/msm/dsi/phy/dsi_phy.c b/drivers/gpu/drm/msm/dsi/phy/dsi_phy.c index e2f42d8ea294..f39386ed75e4 100644 --- a/drivers/gpu/drm/msm/dsi/phy/dsi_phy.c +++ b/drivers/gpu/drm/msm/dsi/phy/dsi_phy.c @@ -271,6 +271,30 @@ static const struct of_device_id dsi_phy_dt_match[] = { {} }; +/* + * Currently, we only support one SoC for each PHY type. When we have multiple + * SoCs for the same PHY, we can try to make the index searching a bit more + * clever. + */ +static int dsi_phy_get_id(struct msm_dsi_phy *phy) +{ + struct platform_device *pdev = phy->pdev; + const struct msm_dsi_phy_cfg *cfg = phy->cfg; + struct resource *res; + int i; + + res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "dsi_phy"); + if (!res) + return -EINVAL; + + for (i = 0; i < cfg->num_dsi_phy; i++) { + if (cfg->io_start[i] == res->start) + return i; + } + + return -EINVAL; +} + static int dsi_phy_driver_probe(struct platform_device *pdev) { struct msm_dsi_phy *phy; @@ -289,10 +313,10 @@ static int dsi_phy_driver_probe(struct platform_device *pdev) phy->cfg = match->data; phy->pdev = pdev; - ret = of_property_read_u32(dev->of_node, - "qcom,dsi-phy-index", &phy->id); - if (ret) { - dev_err(dev, "%s: PHY index not specified, %d\n", + phy->id = dsi_phy_get_id(phy); + if (phy->id < 0) { + ret = phy->id; + dev_err(dev, "%s: couldn't identify PHY index, %d\n", __func__, ret); goto fail; } diff --git a/drivers/gpu/drm/msm/dsi/phy/dsi_phy.h b/drivers/gpu/drm/msm/dsi/phy/dsi_phy.h index 0d54ed00386d..f24a85439b94 100644 --- a/drivers/gpu/drm/msm/dsi/phy/dsi_phy.h +++ b/drivers/gpu/drm/msm/dsi/phy/dsi_phy.h @@ -38,6 +38,8 @@ struct msm_dsi_phy_cfg { * Fill default H/W values in illegal cells, eg. cell {0, 1}. */ bool src_pll_truthtable[DSI_MAX][DSI_MAX]; + const resource_size_t io_start[DSI_MAX]; + const int num_dsi_phy; }; extern const struct msm_dsi_phy_cfg dsi_phy_28nm_hpm_cfgs; diff --git a/drivers/gpu/drm/msm/dsi/phy/dsi_phy_20nm.c b/drivers/gpu/drm/msm/dsi/phy/dsi_phy_20nm.c index f4bc11af849a..c757e2070cac 100644 --- a/drivers/gpu/drm/msm/dsi/phy/dsi_phy_20nm.c +++ b/drivers/gpu/drm/msm/dsi/phy/dsi_phy_20nm.c @@ -145,6 +145,8 @@ const struct msm_dsi_phy_cfg dsi_phy_20nm_cfgs = { .ops = { .enable = dsi_20nm_phy_enable, .disable = dsi_20nm_phy_disable, - } + }, + .io_start = { 0xfd998300, 0xfd9a0300 }, + .num_dsi_phy = 2, }; diff --git a/drivers/gpu/drm/msm/dsi/phy/dsi_phy_28nm.c b/drivers/gpu/drm/msm/dsi/phy/dsi_phy_28nm.c index 96d1852af418..63d7fba31380 100644 --- a/drivers/gpu/drm/msm/dsi/phy/dsi_phy_28nm.c +++ b/drivers/gpu/drm/msm/dsi/phy/dsi_phy_28nm.c @@ -145,6 +145,8 @@ const struct msm_dsi_phy_cfg dsi_phy_28nm_hpm_cfgs = { .enable = dsi_28nm_phy_enable, .disable = dsi_28nm_phy_disable, }, + .io_start = { 0xfd922b00, 0xfd923100 }, + .num_dsi_phy = 2, }; const struct msm_dsi_phy_cfg dsi_phy_28nm_lp_cfgs = { @@ -160,5 +162,7 @@ const struct msm_dsi_phy_cfg dsi_phy_28nm_lp_cfgs = { .enable = dsi_28nm_phy_enable, .disable = dsi_28nm_phy_disable, }, + .io_start = { 0x1a98500 }, + .num_dsi_phy = 1, }; diff --git a/drivers/gpu/drm/msm/dsi/phy/dsi_phy_28nm_8960.c b/drivers/gpu/drm/msm/dsi/phy/dsi_phy_28nm_8960.c index 213355a3e767..7bdb9de54968 100644 --- a/drivers/gpu/drm/msm/dsi/phy/dsi_phy_28nm_8960.c +++ b/drivers/gpu/drm/msm/dsi/phy/dsi_phy_28nm_8960.c @@ -192,4 +192,6 @@ const struct msm_dsi_phy_cfg dsi_phy_28nm_8960_cfgs = { .enable = dsi_28nm_phy_enable, .disable = dsi_28nm_phy_disable, }, + .io_start = { 0x4700300, 0x5800300 }, + .num_dsi_phy = 2, }; diff --git a/drivers/gpu/drm/msm/hdmi/hdmi.c b/drivers/gpu/drm/msm/hdmi/hdmi.c index 51b9ea552f97..973720792236 100644 --- a/drivers/gpu/drm/msm/hdmi/hdmi.c +++ b/drivers/gpu/drm/msm/hdmi/hdmi.c @@ -19,6 +19,7 @@ #include <linux/of_irq.h> #include <linux/of_gpio.h> +#include <sound/hdmi-codec.h> #include "hdmi.h" void msm_hdmi_set_mode(struct hdmi *hdmi, bool power_on) @@ -434,6 +435,111 @@ static int msm_hdmi_get_gpio(struct device_node *of_node, const char *name) return gpio; } +/* + * HDMI audio codec callbacks + */ +static int msm_hdmi_audio_hw_params(struct device *dev, void *data, + struct hdmi_codec_daifmt *daifmt, + struct hdmi_codec_params *params) +{ + struct hdmi *hdmi = dev_get_drvdata(dev); + unsigned int chan; + unsigned int channel_allocation = 0; + unsigned int rate; + unsigned int level_shift = 0; /* 0dB */ + bool down_mix = false; + + dev_dbg(dev, "%u Hz, %d bit, %d channels\n", params->sample_rate, + params->sample_width, params->cea.channels); + + switch (params->cea.channels) { + case 2: + /* FR and FL speakers */ + channel_allocation = 0; + chan = MSM_HDMI_AUDIO_CHANNEL_2; + break; + case 4: + /* FC, LFE, FR and FL speakers */ + channel_allocation = 0x3; + chan = MSM_HDMI_AUDIO_CHANNEL_4; + break; + case 6: + /* RR, RL, FC, LFE, FR and FL speakers */ + channel_allocation = 0x0B; + chan = MSM_HDMI_AUDIO_CHANNEL_6; + break; + case 8: + /* FRC, FLC, RR, RL, FC, LFE, FR and FL speakers */ + channel_allocation = 0x1F; + chan = MSM_HDMI_AUDIO_CHANNEL_8; + break; + default: + return -EINVAL; + } + + switch (params->sample_rate) { + case 32000: + rate = HDMI_SAMPLE_RATE_32KHZ; + break; + case 44100: + rate = HDMI_SAMPLE_RATE_44_1KHZ; + break; + case 48000: + rate = HDMI_SAMPLE_RATE_48KHZ; + break; + case 88200: + rate = HDMI_SAMPLE_RATE_88_2KHZ; + break; + case 96000: + rate = HDMI_SAMPLE_RATE_96KHZ; + break; + case 176400: + rate = HDMI_SAMPLE_RATE_176_4KHZ; + break; + case 192000: + rate = HDMI_SAMPLE_RATE_192KHZ; + break; + default: + dev_err(dev, "rate[%d] not supported!\n", + params->sample_rate); + return -EINVAL; + } + + msm_hdmi_audio_set_sample_rate(hdmi, rate); + msm_hdmi_audio_info_setup(hdmi, 1, chan, channel_allocation, + level_shift, down_mix); + + return 0; +} + +static void msm_hdmi_audio_shutdown(struct device *dev, void *data) +{ + struct hdmi *hdmi = dev_get_drvdata(dev); + + msm_hdmi_audio_info_setup(hdmi, 0, 0, 0, 0, 0); +} + +static const struct hdmi_codec_ops msm_hdmi_audio_codec_ops = { + .hw_params = msm_hdmi_audio_hw_params, + .audio_shutdown = msm_hdmi_audio_shutdown, +}; + +static struct hdmi_codec_pdata codec_data = { + .ops = &msm_hdmi_audio_codec_ops, + .max_i2s_channels = 8, + .i2s = 1, +}; + +static int msm_hdmi_register_audio_driver(struct hdmi *hdmi, struct device *dev) +{ + hdmi->audio_pdev = platform_device_register_data(dev, + HDMI_CODEC_DRV_NAME, + PLATFORM_DEVID_AUTO, + &codec_data, + sizeof(codec_data)); + return PTR_ERR_OR_ZERO(hdmi->audio_pdev); +} + static int msm_hdmi_bind(struct device *dev, struct device *master, void *data) { struct drm_device *drm = dev_get_drvdata(master); @@ -441,7 +547,7 @@ static int msm_hdmi_bind(struct device *dev, struct device *master, void *data) static struct hdmi_platform_config *hdmi_cfg; struct hdmi *hdmi; struct device_node *of_node = dev->of_node; - int i; + int i, err; hdmi_cfg = (struct hdmi_platform_config *) of_device_get_match_data(dev); @@ -468,6 +574,12 @@ static int msm_hdmi_bind(struct device *dev, struct device *master, void *data) return PTR_ERR(hdmi); priv->hdmi = hdmi; + err = msm_hdmi_register_audio_driver(hdmi, dev); + if (err) { + DRM_ERROR("Failed to attach an audio codec %d\n", err); + hdmi->audio_pdev = NULL; + } + return 0; } @@ -477,6 +589,9 @@ static void msm_hdmi_unbind(struct device *dev, struct device *master, struct drm_device *drm = dev_get_drvdata(master); struct msm_drm_private *priv = drm->dev_private; if (priv->hdmi) { + if (priv->hdmi->audio_pdev) + platform_device_unregister(priv->hdmi->audio_pdev); + msm_hdmi_destroy(priv->hdmi); priv->hdmi = NULL; } diff --git a/drivers/gpu/drm/msm/hdmi/hdmi.h b/drivers/gpu/drm/msm/hdmi/hdmi.h index bc7ba0bdee07..accc9a61611d 100644 --- a/drivers/gpu/drm/msm/hdmi/hdmi.h +++ b/drivers/gpu/drm/msm/hdmi/hdmi.h @@ -50,6 +50,7 @@ struct hdmi_hdcp_ctrl; struct hdmi { struct drm_device *dev; struct platform_device *pdev; + struct platform_device *audio_pdev; const struct hdmi_platform_config *config; @@ -210,6 +211,19 @@ static inline int msm_hdmi_pll_8996_init(struct platform_device *pdev) /* * audio: */ +/* Supported HDMI Audio channels and rates */ +#define MSM_HDMI_AUDIO_CHANNEL_2 0 +#define MSM_HDMI_AUDIO_CHANNEL_4 1 +#define MSM_HDMI_AUDIO_CHANNEL_6 2 +#define MSM_HDMI_AUDIO_CHANNEL_8 3 + +#define HDMI_SAMPLE_RATE_32KHZ 0 +#define HDMI_SAMPLE_RATE_44_1KHZ 1 +#define HDMI_SAMPLE_RATE_48KHZ 2 +#define HDMI_SAMPLE_RATE_88_2KHZ 3 +#define HDMI_SAMPLE_RATE_96KHZ 4 +#define HDMI_SAMPLE_RATE_176_4KHZ 5 +#define HDMI_SAMPLE_RATE_192KHZ 6 int msm_hdmi_audio_update(struct hdmi *hdmi); int msm_hdmi_audio_info_setup(struct hdmi *hdmi, bool enabled, diff --git a/drivers/gpu/drm/msm/hdmi/hdmi_hdcp.c b/drivers/gpu/drm/msm/hdmi/hdmi_hdcp.c index 0baaaaabd002..6e767979aab3 100644 --- a/drivers/gpu/drm/msm/hdmi/hdmi_hdcp.c +++ b/drivers/gpu/drm/msm/hdmi/hdmi_hdcp.c @@ -1430,7 +1430,7 @@ struct hdmi_hdcp_ctrl *msm_hdmi_hdcp_init(struct hdmi *hdmi) void msm_hdmi_hdcp_destroy(struct hdmi *hdmi) { - if (hdmi && hdmi->hdcp_ctrl) { + if (hdmi) { kfree(hdmi->hdcp_ctrl); hdmi->hdcp_ctrl = NULL; } diff --git a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_dtv_encoder.c b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_dtv_encoder.c index 35ad78a1dc1c..24258e3025e3 100644 --- a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_dtv_encoder.c +++ b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_dtv_encoder.c @@ -23,7 +23,6 @@ struct mdp4_dtv_encoder { struct drm_encoder base; - struct clk *src_clk; struct clk *hdmi_clk; struct clk *mdp_clk; unsigned long int pixclock; @@ -179,7 +178,6 @@ static void mdp4_dtv_encoder_disable(struct drm_encoder *encoder) */ mdp_irq_wait(&mdp4_kms->base, MDP4_IRQ_EXTERNAL_VSYNC); - clk_disable_unprepare(mdp4_dtv_encoder->src_clk); clk_disable_unprepare(mdp4_dtv_encoder->hdmi_clk); clk_disable_unprepare(mdp4_dtv_encoder->mdp_clk); @@ -208,19 +206,21 @@ static void mdp4_dtv_encoder_enable(struct drm_encoder *encoder) bs_set(mdp4_dtv_encoder, 1); - DBG("setting src_clk=%lu", pc); + DBG("setting mdp_clk=%lu", pc); - ret = clk_set_rate(mdp4_dtv_encoder->src_clk, pc); + ret = clk_set_rate(mdp4_dtv_encoder->mdp_clk, pc); if (ret) - dev_err(dev->dev, "failed to set src_clk to %lu: %d\n", pc, ret); - clk_prepare_enable(mdp4_dtv_encoder->src_clk); - ret = clk_prepare_enable(mdp4_dtv_encoder->hdmi_clk); - if (ret) - dev_err(dev->dev, "failed to enable hdmi_clk: %d\n", ret); + dev_err(dev->dev, "failed to set mdp_clk to %lu: %d\n", + pc, ret); + ret = clk_prepare_enable(mdp4_dtv_encoder->mdp_clk); if (ret) dev_err(dev->dev, "failed to enabled mdp_clk: %d\n", ret); + ret = clk_prepare_enable(mdp4_dtv_encoder->hdmi_clk); + if (ret) + dev_err(dev->dev, "failed to enable hdmi_clk: %d\n", ret); + mdp4_write(mdp4_kms, REG_MDP4_DTV_ENABLE, 1); mdp4_dtv_encoder->enabled = true; @@ -235,7 +235,7 @@ static const struct drm_encoder_helper_funcs mdp4_dtv_encoder_helper_funcs = { long mdp4_dtv_round_pixclk(struct drm_encoder *encoder, unsigned long rate) { struct mdp4_dtv_encoder *mdp4_dtv_encoder = to_mdp4_dtv_encoder(encoder); - return clk_round_rate(mdp4_dtv_encoder->src_clk, rate); + return clk_round_rate(mdp4_dtv_encoder->mdp_clk, rate); } /* initialize encoder */ @@ -257,13 +257,6 @@ struct drm_encoder *mdp4_dtv_encoder_init(struct drm_device *dev) DRM_MODE_ENCODER_TMDS, NULL); drm_encoder_helper_add(encoder, &mdp4_dtv_encoder_helper_funcs); - mdp4_dtv_encoder->src_clk = devm_clk_get(dev->dev, "src_clk"); - if (IS_ERR(mdp4_dtv_encoder->src_clk)) { - dev_err(dev->dev, "failed to get src_clk\n"); - ret = PTR_ERR(mdp4_dtv_encoder->src_clk); - goto fail; - } - mdp4_dtv_encoder->hdmi_clk = devm_clk_get(dev->dev, "hdmi_clk"); if (IS_ERR(mdp4_dtv_encoder->hdmi_clk)) { dev_err(dev->dev, "failed to get hdmi_clk\n"); @@ -271,9 +264,9 @@ struct drm_encoder *mdp4_dtv_encoder_init(struct drm_device *dev) goto fail; } - mdp4_dtv_encoder->mdp_clk = devm_clk_get(dev->dev, "mdp_clk"); + mdp4_dtv_encoder->mdp_clk = devm_clk_get(dev->dev, "tv_clk"); if (IS_ERR(mdp4_dtv_encoder->mdp_clk)) { - dev_err(dev->dev, "failed to get mdp_clk\n"); + dev_err(dev->dev, "failed to get tv_clk\n"); ret = PTR_ERR(mdp4_dtv_encoder->mdp_clk); goto fail; } diff --git a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_kms.c b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_kms.c index f145d256e332..7b39e89fbc2b 100644 --- a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_kms.c +++ b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_kms.c @@ -158,6 +158,7 @@ static const char * const iommu_ports[] = { static void mdp4_destroy(struct msm_kms *kms) { struct mdp4_kms *mdp4_kms = to_mdp4_kms(to_mdp_kms(kms)); + struct device *dev = mdp4_kms->dev->dev; struct msm_mmu *mmu = mdp4_kms->mmu; if (mmu) { @@ -167,8 +168,11 @@ static void mdp4_destroy(struct msm_kms *kms) if (mdp4_kms->blank_cursor_iova) msm_gem_put_iova(mdp4_kms->blank_cursor_bo, mdp4_kms->id); - if (mdp4_kms->blank_cursor_bo) - drm_gem_object_unreference_unlocked(mdp4_kms->blank_cursor_bo); + drm_gem_object_unreference_unlocked(mdp4_kms->blank_cursor_bo); + + if (mdp4_kms->rpm_enabled) + pm_runtime_disable(dev); + kfree(mdp4_kms); } @@ -436,7 +440,7 @@ struct msm_kms *mdp4_kms_init(struct drm_device *dev) struct mdp4_kms *mdp4_kms; struct msm_kms *kms = NULL; struct msm_mmu *mmu; - int ret; + int irq, ret; mdp4_kms = kzalloc(sizeof(*mdp4_kms), GFP_KERNEL); if (!mdp4_kms) { @@ -457,6 +461,15 @@ struct msm_kms *mdp4_kms_init(struct drm_device *dev) goto fail; } + irq = platform_get_irq(pdev, 0); + if (irq < 0) { + ret = irq; + dev_err(dev->dev, "failed to get irq: %d\n", ret); + goto fail; + } + + kms->irq = irq; + /* NOTE: driver for this regulator still missing upstream.. use * _get_exclusive() and ignore the error if it does not exist * (and hope that the bootloader left it on for us) @@ -492,7 +505,7 @@ struct msm_kms *mdp4_kms_init(struct drm_device *dev) goto fail; } - mdp4_kms->axi_clk = devm_clk_get(&pdev->dev, "mdp_axi_clk"); + mdp4_kms->axi_clk = devm_clk_get(&pdev->dev, "bus_clk"); if (IS_ERR(mdp4_kms->axi_clk)) { dev_err(dev->dev, "failed to get axi_clk\n"); ret = PTR_ERR(mdp4_kms->axi_clk); @@ -502,6 +515,9 @@ struct msm_kms *mdp4_kms_init(struct drm_device *dev) clk_set_rate(mdp4_kms->clk, config->max_clk); clk_set_rate(mdp4_kms->lut_clk, config->max_clk); + pm_runtime_enable(dev->dev); + mdp4_kms->rpm_enabled = true; + /* make sure things are off before attaching iommu (bootloader could * have left things on, in which case we'll start getting faults if * we don't disable): diff --git a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_kms.h b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_kms.h index c5d045d5680d..25fb83997119 100644 --- a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_kms.h +++ b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_kms.h @@ -47,6 +47,8 @@ struct mdp4_kms { struct mdp_irq error_handler; + bool rpm_enabled; + /* empty/blank cursor bo to use when cursor is "disabled" */ struct drm_gem_object *blank_cursor_bo; uint32_t blank_cursor_iova; diff --git a/drivers/gpu/drm/msm/mdp/mdp5/mdp5.xml.h b/drivers/gpu/drm/msm/mdp/mdp5/mdp5.xml.h index b275ce11b24b..ca6ca30650a0 100644 --- a/drivers/gpu/drm/msm/mdp/mdp5/mdp5.xml.h +++ b/drivers/gpu/drm/msm/mdp/mdp5/mdp5.xml.h @@ -8,19 +8,11 @@ http://github.com/freedreno/envytools/ git clone https://github.com/freedreno/envytools.git The rules-ng-ng source files this header was generated from are: -- /home/robclark/src/freedreno/envytools/rnndb/msm.xml ( 676 bytes, from 2015-05-20 20:03:14) -- /home/robclark/src/freedreno/envytools/rnndb/freedreno_copyright.xml ( 1572 bytes, from 2016-02-10 17:07:21) -- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp4.xml ( 20915 bytes, from 2015-05-20 20:03:14) -- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp_common.xml ( 2849 bytes, from 2015-09-18 12:07:28) -- /home/robclark/src/freedreno/envytools/rnndb/mdp/mdp5.xml ( 37194 bytes, from 2015-09-18 12:07:28) -- /home/robclark/src/freedreno/envytools/rnndb/dsi/dsi.xml ( 27887 bytes, from 2015-10-22 16:34:52) -- /home/robclark/src/freedreno/envytools/rnndb/dsi/sfpb.xml ( 602 bytes, from 2015-10-22 16:35:02) -- /home/robclark/src/freedreno/envytools/rnndb/dsi/mmss_cc.xml ( 1686 bytes, from 2015-05-20 20:03:14) -- /home/robclark/src/freedreno/envytools/rnndb/hdmi/qfprom.xml ( 600 bytes, from 2015-05-20 20:03:07) -- /home/robclark/src/freedreno/envytools/rnndb/hdmi/hdmi.xml ( 41472 bytes, from 2016-01-22 18:18:18) -- /home/robclark/src/freedreno/envytools/rnndb/edp/edp.xml ( 10416 bytes, from 2015-05-20 20:03:14) - -Copyright (C) 2013-2015 by the following authors: +- /local/mnt/workspace/source_trees/envytools/rnndb/../rnndb/mdp/mdp5.xml ( 36965 bytes, from 2016-05-10 05:06:30) +- /local/mnt/workspace/source_trees/envytools/rnndb/freedreno_copyright.xml ( 1572 bytes, from 2016-05-09 06:32:54) +- /local/mnt/workspace/source_trees/envytools/rnndb/mdp/mdp_common.xml ( 2849 bytes, from 2016-01-07 08:45:55) + +Copyright (C) 2013-2016 by the following authors: - Rob Clark <robdclark@gmail.com> (robclark) - Ilia Mirkin <imirkin@alum.mit.edu> (imirkin) @@ -198,118 +190,109 @@ static inline uint32_t MDSS_HW_VERSION_MAJOR(uint32_t val) #define MDSS_HW_INTR_STATUS_INTR_HDMI 0x00000100 #define MDSS_HW_INTR_STATUS_INTR_EDP 0x00001000 -static inline uint32_t __offset_MDP(uint32_t idx) -{ - switch (idx) { - case 0: return (mdp5_cfg->mdp.base[0]); - default: return INVALID_IDX(idx); - } -} -static inline uint32_t REG_MDP5_MDP(uint32_t i0) { return 0x00000000 + __offset_MDP(i0); } - -static inline uint32_t REG_MDP5_MDP_HW_VERSION(uint32_t i0) { return 0x00000000 + __offset_MDP(i0); } -#define MDP5_MDP_HW_VERSION_STEP__MASK 0x0000ffff -#define MDP5_MDP_HW_VERSION_STEP__SHIFT 0 -static inline uint32_t MDP5_MDP_HW_VERSION_STEP(uint32_t val) +#define REG_MDP5_HW_VERSION 0x00000000 +#define MDP5_HW_VERSION_STEP__MASK 0x0000ffff +#define MDP5_HW_VERSION_STEP__SHIFT 0 +static inline uint32_t MDP5_HW_VERSION_STEP(uint32_t val) { - return ((val) << MDP5_MDP_HW_VERSION_STEP__SHIFT) & MDP5_MDP_HW_VERSION_STEP__MASK; + return ((val) << MDP5_HW_VERSION_STEP__SHIFT) & MDP5_HW_VERSION_STEP__MASK; } -#define MDP5_MDP_HW_VERSION_MINOR__MASK 0x0fff0000 -#define MDP5_MDP_HW_VERSION_MINOR__SHIFT 16 -static inline uint32_t MDP5_MDP_HW_VERSION_MINOR(uint32_t val) +#define MDP5_HW_VERSION_MINOR__MASK 0x0fff0000 +#define MDP5_HW_VERSION_MINOR__SHIFT 16 +static inline uint32_t MDP5_HW_VERSION_MINOR(uint32_t val) { - return ((val) << MDP5_MDP_HW_VERSION_MINOR__SHIFT) & MDP5_MDP_HW_VERSION_MINOR__MASK; + return ((val) << MDP5_HW_VERSION_MINOR__SHIFT) & MDP5_HW_VERSION_MINOR__MASK; } -#define MDP5_MDP_HW_VERSION_MAJOR__MASK 0xf0000000 -#define MDP5_MDP_HW_VERSION_MAJOR__SHIFT 28 -static inline uint32_t MDP5_MDP_HW_VERSION_MAJOR(uint32_t val) +#define MDP5_HW_VERSION_MAJOR__MASK 0xf0000000 +#define MDP5_HW_VERSION_MAJOR__SHIFT 28 +static inline uint32_t MDP5_HW_VERSION_MAJOR(uint32_t val) { - return ((val) << MDP5_MDP_HW_VERSION_MAJOR__SHIFT) & MDP5_MDP_HW_VERSION_MAJOR__MASK; + return ((val) << MDP5_HW_VERSION_MAJOR__SHIFT) & MDP5_HW_VERSION_MAJOR__MASK; } -static inline uint32_t REG_MDP5_MDP_DISP_INTF_SEL(uint32_t i0) { return 0x00000004 + __offset_MDP(i0); } -#define MDP5_MDP_DISP_INTF_SEL_INTF0__MASK 0x000000ff -#define MDP5_MDP_DISP_INTF_SEL_INTF0__SHIFT 0 -static inline uint32_t MDP5_MDP_DISP_INTF_SEL_INTF0(enum mdp5_intf_type val) +#define REG_MDP5_DISP_INTF_SEL 0x00000004 +#define MDP5_DISP_INTF_SEL_INTF0__MASK 0x000000ff +#define MDP5_DISP_INTF_SEL_INTF0__SHIFT 0 +static inline uint32_t MDP5_DISP_INTF_SEL_INTF0(enum mdp5_intf_type val) { - return ((val) << MDP5_MDP_DISP_INTF_SEL_INTF0__SHIFT) & MDP5_MDP_DISP_INTF_SEL_INTF0__MASK; + return ((val) << MDP5_DISP_INTF_SEL_INTF0__SHIFT) & MDP5_DISP_INTF_SEL_INTF0__MASK; } -#define MDP5_MDP_DISP_INTF_SEL_INTF1__MASK 0x0000ff00 -#define MDP5_MDP_DISP_INTF_SEL_INTF1__SHIFT 8 -static inline uint32_t MDP5_MDP_DISP_INTF_SEL_INTF1(enum mdp5_intf_type val) +#define MDP5_DISP_INTF_SEL_INTF1__MASK 0x0000ff00 +#define MDP5_DISP_INTF_SEL_INTF1__SHIFT 8 +static inline uint32_t MDP5_DISP_INTF_SEL_INTF1(enum mdp5_intf_type val) { - return ((val) << MDP5_MDP_DISP_INTF_SEL_INTF1__SHIFT) & MDP5_MDP_DISP_INTF_SEL_INTF1__MASK; + return ((val) << MDP5_DISP_INTF_SEL_INTF1__SHIFT) & MDP5_DISP_INTF_SEL_INTF1__MASK; } -#define MDP5_MDP_DISP_INTF_SEL_INTF2__MASK 0x00ff0000 -#define MDP5_MDP_DISP_INTF_SEL_INTF2__SHIFT 16 -static inline uint32_t MDP5_MDP_DISP_INTF_SEL_INTF2(enum mdp5_intf_type val) +#define MDP5_DISP_INTF_SEL_INTF2__MASK 0x00ff0000 +#define MDP5_DISP_INTF_SEL_INTF2__SHIFT 16 +static inline uint32_t MDP5_DISP_INTF_SEL_INTF2(enum mdp5_intf_type val) { - return ((val) << MDP5_MDP_DISP_INTF_SEL_INTF2__SHIFT) & MDP5_MDP_DISP_INTF_SEL_INTF2__MASK; + return ((val) << MDP5_DISP_INTF_SEL_INTF2__SHIFT) & MDP5_DISP_INTF_SEL_INTF2__MASK; } -#define MDP5_MDP_DISP_INTF_SEL_INTF3__MASK 0xff000000 -#define MDP5_MDP_DISP_INTF_SEL_INTF3__SHIFT 24 -static inline uint32_t MDP5_MDP_DISP_INTF_SEL_INTF3(enum mdp5_intf_type val) +#define MDP5_DISP_INTF_SEL_INTF3__MASK 0xff000000 +#define MDP5_DISP_INTF_SEL_INTF3__SHIFT 24 +static inline uint32_t MDP5_DISP_INTF_SEL_INTF3(enum mdp5_intf_type val) { - return ((val) << MDP5_MDP_DISP_INTF_SEL_INTF3__SHIFT) & MDP5_MDP_DISP_INTF_SEL_INTF3__MASK; + return ((val) << MDP5_DISP_INTF_SEL_INTF3__SHIFT) & MDP5_DISP_INTF_SEL_INTF3__MASK; } -static inline uint32_t REG_MDP5_MDP_INTR_EN(uint32_t i0) { return 0x00000010 + __offset_MDP(i0); } +#define REG_MDP5_INTR_EN 0x00000010 -static inline uint32_t REG_MDP5_MDP_INTR_STATUS(uint32_t i0) { return 0x00000014 + __offset_MDP(i0); } +#define REG_MDP5_INTR_STATUS 0x00000014 -static inline uint32_t REG_MDP5_MDP_INTR_CLEAR(uint32_t i0) { return 0x00000018 + __offset_MDP(i0); } +#define REG_MDP5_INTR_CLEAR 0x00000018 -static inline uint32_t REG_MDP5_MDP_HIST_INTR_EN(uint32_t i0) { return 0x0000001c + __offset_MDP(i0); } +#define REG_MDP5_HIST_INTR_EN 0x0000001c -static inline uint32_t REG_MDP5_MDP_HIST_INTR_STATUS(uint32_t i0) { return 0x00000020 + __offset_MDP(i0); } +#define REG_MDP5_HIST_INTR_STATUS 0x00000020 -static inline uint32_t REG_MDP5_MDP_HIST_INTR_CLEAR(uint32_t i0) { return 0x00000024 + __offset_MDP(i0); } +#define REG_MDP5_HIST_INTR_CLEAR 0x00000024 -static inline uint32_t REG_MDP5_MDP_SPARE_0(uint32_t i0) { return 0x00000028 + __offset_MDP(i0); } -#define MDP5_MDP_SPARE_0_SPLIT_DPL_SINGLE_FLUSH_EN 0x00000001 +#define REG_MDP5_SPARE_0 0x00000028 +#define MDP5_SPARE_0_SPLIT_DPL_SINGLE_FLUSH_EN 0x00000001 -static inline uint32_t REG_MDP5_MDP_SMP_ALLOC_W(uint32_t i0, uint32_t i1) { return 0x00000080 + __offset_MDP(i0) + 0x4*i1; } +static inline uint32_t REG_MDP5_SMP_ALLOC_W(uint32_t i0) { return 0x00000080 + 0x4*i0; } -static inline uint32_t REG_MDP5_MDP_SMP_ALLOC_W_REG(uint32_t i0, uint32_t i1) { return 0x00000080 + __offset_MDP(i0) + 0x4*i1; } -#define MDP5_MDP_SMP_ALLOC_W_REG_CLIENT0__MASK 0x000000ff -#define MDP5_MDP_SMP_ALLOC_W_REG_CLIENT0__SHIFT 0 -static inline uint32_t MDP5_MDP_SMP_ALLOC_W_REG_CLIENT0(uint32_t val) +static inline uint32_t REG_MDP5_SMP_ALLOC_W_REG(uint32_t i0) { return 0x00000080 + 0x4*i0; } +#define MDP5_SMP_ALLOC_W_REG_CLIENT0__MASK 0x000000ff +#define MDP5_SMP_ALLOC_W_REG_CLIENT0__SHIFT 0 +static inline uint32_t MDP5_SMP_ALLOC_W_REG_CLIENT0(uint32_t val) { - return ((val) << MDP5_MDP_SMP_ALLOC_W_REG_CLIENT0__SHIFT) & MDP5_MDP_SMP_ALLOC_W_REG_CLIENT0__MASK; + return ((val) << MDP5_SMP_ALLOC_W_REG_CLIENT0__SHIFT) & MDP5_SMP_ALLOC_W_REG_CLIENT0__MASK; } -#define MDP5_MDP_SMP_ALLOC_W_REG_CLIENT1__MASK 0x0000ff00 -#define MDP5_MDP_SMP_ALLOC_W_REG_CLIENT1__SHIFT 8 -static inline uint32_t MDP5_MDP_SMP_ALLOC_W_REG_CLIENT1(uint32_t val) +#define MDP5_SMP_ALLOC_W_REG_CLIENT1__MASK 0x0000ff00 +#define MDP5_SMP_ALLOC_W_REG_CLIENT1__SHIFT 8 +static inline uint32_t MDP5_SMP_ALLOC_W_REG_CLIENT1(uint32_t val) { - return ((val) << MDP5_MDP_SMP_ALLOC_W_REG_CLIENT1__SHIFT) & MDP5_MDP_SMP_ALLOC_W_REG_CLIENT1__MASK; + return ((val) << MDP5_SMP_ALLOC_W_REG_CLIENT1__SHIFT) & MDP5_SMP_ALLOC_W_REG_CLIENT1__MASK; } -#define MDP5_MDP_SMP_ALLOC_W_REG_CLIENT2__MASK 0x00ff0000 -#define MDP5_MDP_SMP_ALLOC_W_REG_CLIENT2__SHIFT 16 -static inline uint32_t MDP5_MDP_SMP_ALLOC_W_REG_CLIENT2(uint32_t val) +#define MDP5_SMP_ALLOC_W_REG_CLIENT2__MASK 0x00ff0000 +#define MDP5_SMP_ALLOC_W_REG_CLIENT2__SHIFT 16 +static inline uint32_t MDP5_SMP_ALLOC_W_REG_CLIENT2(uint32_t val) { - return ((val) << MDP5_MDP_SMP_ALLOC_W_REG_CLIENT2__SHIFT) & MDP5_MDP_SMP_ALLOC_W_REG_CLIENT2__MASK; + return ((val) << MDP5_SMP_ALLOC_W_REG_CLIENT2__SHIFT) & MDP5_SMP_ALLOC_W_REG_CLIENT2__MASK; } -static inline uint32_t REG_MDP5_MDP_SMP_ALLOC_R(uint32_t i0, uint32_t i1) { return 0x00000130 + __offset_MDP(i0) + 0x4*i1; } +static inline uint32_t REG_MDP5_SMP_ALLOC_R(uint32_t i0) { return 0x00000130 + 0x4*i0; } -static inline uint32_t REG_MDP5_MDP_SMP_ALLOC_R_REG(uint32_t i0, uint32_t i1) { return 0x00000130 + __offset_MDP(i0) + 0x4*i1; } -#define MDP5_MDP_SMP_ALLOC_R_REG_CLIENT0__MASK 0x000000ff -#define MDP5_MDP_SMP_ALLOC_R_REG_CLIENT0__SHIFT 0 -static inline uint32_t MDP5_MDP_SMP_ALLOC_R_REG_CLIENT0(uint32_t val) +static inline uint32_t REG_MDP5_SMP_ALLOC_R_REG(uint32_t i0) { return 0x00000130 + 0x4*i0; } +#define MDP5_SMP_ALLOC_R_REG_CLIENT0__MASK 0x000000ff +#define MDP5_SMP_ALLOC_R_REG_CLIENT0__SHIFT 0 +static inline uint32_t MDP5_SMP_ALLOC_R_REG_CLIENT0(uint32_t val) { - return ((val) << MDP5_MDP_SMP_ALLOC_R_REG_CLIENT0__SHIFT) & MDP5_MDP_SMP_ALLOC_R_REG_CLIENT0__MASK; + return ((val) << MDP5_SMP_ALLOC_R_REG_CLIENT0__SHIFT) & MDP5_SMP_ALLOC_R_REG_CLIENT0__MASK; } -#define MDP5_MDP_SMP_ALLOC_R_REG_CLIENT1__MASK 0x0000ff00 -#define MDP5_MDP_SMP_ALLOC_R_REG_CLIENT1__SHIFT 8 -static inline uint32_t MDP5_MDP_SMP_ALLOC_R_REG_CLIENT1(uint32_t val) +#define MDP5_SMP_ALLOC_R_REG_CLIENT1__MASK 0x0000ff00 +#define MDP5_SMP_ALLOC_R_REG_CLIENT1__SHIFT 8 +static inline uint32_t MDP5_SMP_ALLOC_R_REG_CLIENT1(uint32_t val) { - return ((val) << MDP5_MDP_SMP_ALLOC_R_REG_CLIENT1__SHIFT) & MDP5_MDP_SMP_ALLOC_R_REG_CLIENT1__MASK; + return ((val) << MDP5_SMP_ALLOC_R_REG_CLIENT1__SHIFT) & MDP5_SMP_ALLOC_R_REG_CLIENT1__MASK; } -#define MDP5_MDP_SMP_ALLOC_R_REG_CLIENT2__MASK 0x00ff0000 -#define MDP5_MDP_SMP_ALLOC_R_REG_CLIENT2__SHIFT 16 -static inline uint32_t MDP5_MDP_SMP_ALLOC_R_REG_CLIENT2(uint32_t val) +#define MDP5_SMP_ALLOC_R_REG_CLIENT2__MASK 0x00ff0000 +#define MDP5_SMP_ALLOC_R_REG_CLIENT2__SHIFT 16 +static inline uint32_t MDP5_SMP_ALLOC_R_REG_CLIENT2(uint32_t val) { - return ((val) << MDP5_MDP_SMP_ALLOC_R_REG_CLIENT2__SHIFT) & MDP5_MDP_SMP_ALLOC_R_REG_CLIENT2__MASK; + return ((val) << MDP5_SMP_ALLOC_R_REG_CLIENT2__SHIFT) & MDP5_SMP_ALLOC_R_REG_CLIENT2__MASK; } static inline uint32_t __offset_IGC(enum mdp5_igc_type idx) @@ -322,35 +305,35 @@ static inline uint32_t __offset_IGC(enum mdp5_igc_type idx) default: return INVALID_IDX(idx); } } -static inline uint32_t REG_MDP5_MDP_IGC(uint32_t i0, enum mdp5_igc_type i1) { return 0x00000000 + __offset_MDP(i0) + __offset_IGC(i1); } +static inline uint32_t REG_MDP5_IGC(enum mdp5_igc_type i0) { return 0x00000000 + __offset_IGC(i0); } -static inline uint32_t REG_MDP5_MDP_IGC_LUT(uint32_t i0, enum mdp5_igc_type i1, uint32_t i2) { return 0x00000000 + __offset_MDP(i0) + __offset_IGC(i1) + 0x4*i2; } +static inline uint32_t REG_MDP5_IGC_LUT(enum mdp5_igc_type i0, uint32_t i1) { return 0x00000000 + __offset_IGC(i0) + 0x4*i1; } -static inline uint32_t REG_MDP5_MDP_IGC_LUT_REG(uint32_t i0, enum mdp5_igc_type i1, uint32_t i2) { return 0x00000000 + __offset_MDP(i0) + __offset_IGC(i1) + 0x4*i2; } -#define MDP5_MDP_IGC_LUT_REG_VAL__MASK 0x00000fff -#define MDP5_MDP_IGC_LUT_REG_VAL__SHIFT 0 -static inline uint32_t MDP5_MDP_IGC_LUT_REG_VAL(uint32_t val) +static inline uint32_t REG_MDP5_IGC_LUT_REG(enum mdp5_igc_type i0, uint32_t i1) { return 0x00000000 + __offset_IGC(i0) + 0x4*i1; } +#define MDP5_IGC_LUT_REG_VAL__MASK 0x00000fff +#define MDP5_IGC_LUT_REG_VAL__SHIFT 0 +static inline uint32_t MDP5_IGC_LUT_REG_VAL(uint32_t val) { - return ((val) << MDP5_MDP_IGC_LUT_REG_VAL__SHIFT) & MDP5_MDP_IGC_LUT_REG_VAL__MASK; + return ((val) << MDP5_IGC_LUT_REG_VAL__SHIFT) & MDP5_IGC_LUT_REG_VAL__MASK; } -#define MDP5_MDP_IGC_LUT_REG_INDEX_UPDATE 0x02000000 -#define MDP5_MDP_IGC_LUT_REG_DISABLE_PIPE_0 0x10000000 -#define MDP5_MDP_IGC_LUT_REG_DISABLE_PIPE_1 0x20000000 -#define MDP5_MDP_IGC_LUT_REG_DISABLE_PIPE_2 0x40000000 +#define MDP5_IGC_LUT_REG_INDEX_UPDATE 0x02000000 +#define MDP5_IGC_LUT_REG_DISABLE_PIPE_0 0x10000000 +#define MDP5_IGC_LUT_REG_DISABLE_PIPE_1 0x20000000 +#define MDP5_IGC_LUT_REG_DISABLE_PIPE_2 0x40000000 -static inline uint32_t REG_MDP5_MDP_SPLIT_DPL_EN(uint32_t i0) { return 0x000002f4 + __offset_MDP(i0); } +#define REG_MDP5_SPLIT_DPL_EN 0x000002f4 -static inline uint32_t REG_MDP5_MDP_SPLIT_DPL_UPPER(uint32_t i0) { return 0x000002f8 + __offset_MDP(i0); } -#define MDP5_MDP_SPLIT_DPL_UPPER_SMART_PANEL 0x00000002 -#define MDP5_MDP_SPLIT_DPL_UPPER_SMART_PANEL_FREE_RUN 0x00000004 -#define MDP5_MDP_SPLIT_DPL_UPPER_INTF1_SW_TRG_MUX 0x00000010 -#define MDP5_MDP_SPLIT_DPL_UPPER_INTF2_SW_TRG_MUX 0x00000100 +#define REG_MDP5_SPLIT_DPL_UPPER 0x000002f8 +#define MDP5_SPLIT_DPL_UPPER_SMART_PANEL 0x00000002 +#define MDP5_SPLIT_DPL_UPPER_SMART_PANEL_FREE_RUN 0x00000004 +#define MDP5_SPLIT_DPL_UPPER_INTF1_SW_TRG_MUX 0x00000010 +#define MDP5_SPLIT_DPL_UPPER_INTF2_SW_TRG_MUX 0x00000100 -static inline uint32_t REG_MDP5_MDP_SPLIT_DPL_LOWER(uint32_t i0) { return 0x000003f0 + __offset_MDP(i0); } -#define MDP5_MDP_SPLIT_DPL_LOWER_SMART_PANEL 0x00000002 -#define MDP5_MDP_SPLIT_DPL_LOWER_SMART_PANEL_FREE_RUN 0x00000004 -#define MDP5_MDP_SPLIT_DPL_LOWER_INTF1_TG_SYNC 0x00000010 -#define MDP5_MDP_SPLIT_DPL_LOWER_INTF2_TG_SYNC 0x00000100 +#define REG_MDP5_SPLIT_DPL_LOWER 0x000003f0 +#define MDP5_SPLIT_DPL_LOWER_SMART_PANEL 0x00000002 +#define MDP5_SPLIT_DPL_LOWER_SMART_PANEL_FREE_RUN 0x00000004 +#define MDP5_SPLIT_DPL_LOWER_INTF1_TG_SYNC 0x00000010 +#define MDP5_SPLIT_DPL_LOWER_INTF2_TG_SYNC 0x00000100 static inline uint32_t __offset_CTL(uint32_t idx) { diff --git a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_cfg.c b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_cfg.c index 57f73f0c120d..ac9e4cde1380 100644 --- a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_cfg.c +++ b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_cfg.c @@ -26,7 +26,6 @@ const struct mdp5_cfg_hw msm8x74v1_config = { .name = "msm8x74v1", .mdp = { .count = 1, - .base = { 0x00100 }, .caps = MDP_CAP_SMP | 0, }, @@ -41,12 +40,12 @@ const struct mdp5_cfg_hw msm8x74v1_config = { }, .ctl = { .count = 5, - .base = { 0x00600, 0x00700, 0x00800, 0x00900, 0x00a00 }, + .base = { 0x00500, 0x00600, 0x00700, 0x00800, 0x00900 }, .flush_hw_mask = 0x0003ffff, }, .pipe_vig = { .count = 3, - .base = { 0x01200, 0x01600, 0x01a00 }, + .base = { 0x01100, 0x01500, 0x01900 }, .caps = MDP_PIPE_CAP_HFLIP | MDP_PIPE_CAP_VFLIP | MDP_PIPE_CAP_SCALE | @@ -55,7 +54,7 @@ const struct mdp5_cfg_hw msm8x74v1_config = { }, .pipe_rgb = { .count = 3, - .base = { 0x01e00, 0x02200, 0x02600 }, + .base = { 0x01d00, 0x02100, 0x02500 }, .caps = MDP_PIPE_CAP_HFLIP | MDP_PIPE_CAP_VFLIP | MDP_PIPE_CAP_SCALE | @@ -63,26 +62,26 @@ const struct mdp5_cfg_hw msm8x74v1_config = { }, .pipe_dma = { .count = 2, - .base = { 0x02a00, 0x02e00 }, + .base = { 0x02900, 0x02d00 }, .caps = MDP_PIPE_CAP_HFLIP | MDP_PIPE_CAP_VFLIP | 0, }, .lm = { .count = 5, - .base = { 0x03200, 0x03600, 0x03a00, 0x03e00, 0x04200 }, + .base = { 0x03100, 0x03500, 0x03900, 0x03d00, 0x04100 }, .nb_stages = 5, }, .dspp = { .count = 3, - .base = { 0x04600, 0x04a00, 0x04e00 }, + .base = { 0x04500, 0x04900, 0x04d00 }, }, .pp = { .count = 3, - .base = { 0x21b00, 0x21c00, 0x21d00 }, + .base = { 0x21a00, 0x21b00, 0x21c00 }, }, .intf = { - .base = { 0x21100, 0x21300, 0x21500, 0x21700 }, + .base = { 0x21000, 0x21200, 0x21400, 0x21600 }, .connect = { [0] = INTF_eDP, [1] = INTF_DSI, @@ -97,7 +96,6 @@ const struct mdp5_cfg_hw msm8x74v2_config = { .name = "msm8x74", .mdp = { .count = 1, - .base = { 0x00100 }, .caps = MDP_CAP_SMP | 0, }, @@ -112,48 +110,48 @@ const struct mdp5_cfg_hw msm8x74v2_config = { }, .ctl = { .count = 5, - .base = { 0x00600, 0x00700, 0x00800, 0x00900, 0x00a00 }, + .base = { 0x00500, 0x00600, 0x00700, 0x00800, 0x00900 }, .flush_hw_mask = 0x0003ffff, }, .pipe_vig = { .count = 3, - .base = { 0x01200, 0x01600, 0x01a00 }, + .base = { 0x01100, 0x01500, 0x01900 }, .caps = MDP_PIPE_CAP_HFLIP | MDP_PIPE_CAP_VFLIP | MDP_PIPE_CAP_SCALE | MDP_PIPE_CAP_CSC | MDP_PIPE_CAP_DECIMATION, }, .pipe_rgb = { .count = 3, - .base = { 0x01e00, 0x02200, 0x02600 }, + .base = { 0x01d00, 0x02100, 0x02500 }, .caps = MDP_PIPE_CAP_HFLIP | MDP_PIPE_CAP_VFLIP | MDP_PIPE_CAP_SCALE | MDP_PIPE_CAP_DECIMATION, }, .pipe_dma = { .count = 2, - .base = { 0x02a00, 0x02e00 }, + .base = { 0x02900, 0x02d00 }, .caps = MDP_PIPE_CAP_HFLIP | MDP_PIPE_CAP_VFLIP, }, .lm = { .count = 5, - .base = { 0x03200, 0x03600, 0x03a00, 0x03e00, 0x04200 }, + .base = { 0x03100, 0x03500, 0x03900, 0x03d00, 0x04100 }, .nb_stages = 5, .max_width = 2048, .max_height = 0xFFFF, }, .dspp = { .count = 3, - .base = { 0x04600, 0x04a00, 0x04e00 }, + .base = { 0x04500, 0x04900, 0x04d00 }, }, .ad = { .count = 2, - .base = { 0x13100, 0x13300 }, + .base = { 0x13000, 0x13200 }, }, .pp = { .count = 3, - .base = { 0x12d00, 0x12e00, 0x12f00 }, + .base = { 0x12c00, 0x12d00, 0x12e00 }, }, .intf = { - .base = { 0x12500, 0x12700, 0x12900, 0x12b00 }, + .base = { 0x12400, 0x12600, 0x12800, 0x12a00 }, .connect = { [0] = INTF_eDP, [1] = INTF_DSI, @@ -168,7 +166,6 @@ const struct mdp5_cfg_hw apq8084_config = { .name = "apq8084", .mdp = { .count = 1, - .base = { 0x00100 }, .caps = MDP_CAP_SMP | 0, }, @@ -190,49 +187,49 @@ const struct mdp5_cfg_hw apq8084_config = { }, .ctl = { .count = 5, - .base = { 0x00600, 0x00700, 0x00800, 0x00900, 0x00a00 }, + .base = { 0x00500, 0x00600, 0x00700, 0x00800, 0x00900 }, .flush_hw_mask = 0x003fffff, }, .pipe_vig = { .count = 4, - .base = { 0x01200, 0x01600, 0x01a00, 0x01e00 }, + .base = { 0x01100, 0x01500, 0x01900, 0x01d00 }, .caps = MDP_PIPE_CAP_HFLIP | MDP_PIPE_CAP_VFLIP | MDP_PIPE_CAP_SCALE | MDP_PIPE_CAP_CSC | MDP_PIPE_CAP_DECIMATION, }, .pipe_rgb = { .count = 4, - .base = { 0x02200, 0x02600, 0x02a00, 0x02e00 }, + .base = { 0x02100, 0x02500, 0x02900, 0x02d00 }, .caps = MDP_PIPE_CAP_HFLIP | MDP_PIPE_CAP_VFLIP | MDP_PIPE_CAP_SCALE | MDP_PIPE_CAP_DECIMATION, }, .pipe_dma = { .count = 2, - .base = { 0x03200, 0x03600 }, + .base = { 0x03100, 0x03500 }, .caps = MDP_PIPE_CAP_HFLIP | MDP_PIPE_CAP_VFLIP, }, .lm = { .count = 6, - .base = { 0x03a00, 0x03e00, 0x04200, 0x04600, 0x04a00, 0x04e00 }, + .base = { 0x03900, 0x03d00, 0x04100, 0x04500, 0x04900, 0x04d00 }, .nb_stages = 5, .max_width = 2048, .max_height = 0xFFFF, }, .dspp = { .count = 4, - .base = { 0x05200, 0x05600, 0x05a00, 0x05e00 }, + .base = { 0x05100, 0x05500, 0x05900, 0x05d00 }, }, .ad = { .count = 3, - .base = { 0x13500, 0x13700, 0x13900 }, + .base = { 0x13400, 0x13600, 0x13800 }, }, .pp = { .count = 4, - .base = { 0x12f00, 0x13000, 0x13100, 0x13200 }, + .base = { 0x12e00, 0x12f00, 0x13000, 0x13100 }, }, .intf = { - .base = { 0x12500, 0x12700, 0x12900, 0x12b00, 0x12d00 }, + .base = { 0x12400, 0x12600, 0x12800, 0x12a00, 0x12c00 }, .connect = { [0] = INTF_eDP, [1] = INTF_DSI, @@ -247,7 +244,7 @@ const struct mdp5_cfg_hw msm8x16_config = { .name = "msm8x16", .mdp = { .count = 1, - .base = { 0x01000 }, + .base = { 0x0 }, .caps = MDP_CAP_SMP | 0, }, @@ -261,41 +258,41 @@ const struct mdp5_cfg_hw msm8x16_config = { }, .ctl = { .count = 5, - .base = { 0x02000, 0x02200, 0x02400, 0x02600, 0x02800 }, + .base = { 0x01000, 0x01200, 0x01400, 0x01600, 0x01800 }, .flush_hw_mask = 0x4003ffff, }, .pipe_vig = { .count = 1, - .base = { 0x05000 }, + .base = { 0x04000 }, .caps = MDP_PIPE_CAP_HFLIP | MDP_PIPE_CAP_VFLIP | MDP_PIPE_CAP_SCALE | MDP_PIPE_CAP_CSC | MDP_PIPE_CAP_DECIMATION, }, .pipe_rgb = { .count = 2, - .base = { 0x15000, 0x17000 }, + .base = { 0x14000, 0x16000 }, .caps = MDP_PIPE_CAP_HFLIP | MDP_PIPE_CAP_VFLIP | MDP_PIPE_CAP_SCALE | MDP_PIPE_CAP_DECIMATION, }, .pipe_dma = { .count = 1, - .base = { 0x25000 }, + .base = { 0x24000 }, .caps = MDP_PIPE_CAP_HFLIP | MDP_PIPE_CAP_VFLIP, }, .lm = { .count = 2, /* LM0 and LM3 */ - .base = { 0x45000, 0x48000 }, + .base = { 0x44000, 0x47000 }, .nb_stages = 5, .max_width = 2048, .max_height = 0xFFFF, }, .dspp = { .count = 1, - .base = { 0x55000 }, + .base = { 0x54000 }, }, .intf = { - .base = { 0x00000, 0x6b800 }, + .base = { 0x00000, 0x6a800 }, .connect = { [0] = INTF_DISABLED, [1] = INTF_DSI, @@ -308,7 +305,6 @@ const struct mdp5_cfg_hw msm8x94_config = { .name = "msm8x94", .mdp = { .count = 1, - .base = { 0x01000 }, .caps = MDP_CAP_SMP | 0, }, @@ -330,49 +326,49 @@ const struct mdp5_cfg_hw msm8x94_config = { }, .ctl = { .count = 5, - .base = { 0x02000, 0x02200, 0x02400, 0x02600, 0x02800 }, + .base = { 0x01000, 0x01200, 0x01400, 0x01600, 0x01800 }, .flush_hw_mask = 0xf0ffffff, }, .pipe_vig = { .count = 4, - .base = { 0x05000, 0x07000, 0x09000, 0x0b000 }, + .base = { 0x04000, 0x06000, 0x08000, 0x0a000 }, .caps = MDP_PIPE_CAP_HFLIP | MDP_PIPE_CAP_VFLIP | MDP_PIPE_CAP_SCALE | MDP_PIPE_CAP_CSC | MDP_PIPE_CAP_DECIMATION, }, .pipe_rgb = { .count = 4, - .base = { 0x15000, 0x17000, 0x19000, 0x1b000 }, + .base = { 0x14000, 0x16000, 0x18000, 0x1a000 }, .caps = MDP_PIPE_CAP_HFLIP | MDP_PIPE_CAP_VFLIP | MDP_PIPE_CAP_SCALE | MDP_PIPE_CAP_DECIMATION, }, .pipe_dma = { .count = 2, - .base = { 0x25000, 0x27000 }, + .base = { 0x24000, 0x26000 }, .caps = MDP_PIPE_CAP_HFLIP | MDP_PIPE_CAP_VFLIP, }, .lm = { .count = 6, - .base = { 0x45000, 0x46000, 0x47000, 0x48000, 0x49000, 0x4a000 }, + .base = { 0x44000, 0x45000, 0x46000, 0x47000, 0x48000, 0x49000 }, .nb_stages = 8, .max_width = 2048, .max_height = 0xFFFF, }, .dspp = { .count = 4, - .base = { 0x55000, 0x57000, 0x59000, 0x5b000 }, + .base = { 0x54000, 0x56000, 0x58000, 0x5a000 }, }, .ad = { .count = 3, - .base = { 0x79000, 0x79800, 0x7a000 }, + .base = { 0x78000, 0x78800, 0x79000 }, }, .pp = { .count = 4, - .base = { 0x71000, 0x71800, 0x72000, 0x72800 }, + .base = { 0x70000, 0x70800, 0x71000, 0x71800 }, }, .intf = { - .base = { 0x6b000, 0x6b800, 0x6c000, 0x6c800, 0x6d000 }, + .base = { 0x6a000, 0x6a800, 0x6b000, 0x6b800, 0x6c000 }, .connect = { [0] = INTF_DISABLED, [1] = INTF_DSI, @@ -387,19 +383,18 @@ const struct mdp5_cfg_hw msm8x96_config = { .name = "msm8x96", .mdp = { .count = 1, - .base = { 0x01000 }, .caps = MDP_CAP_DSC | MDP_CAP_CDM | 0, }, .ctl = { .count = 5, - .base = { 0x02000, 0x02200, 0x02400, 0x02600, 0x02800 }, + .base = { 0x01000, 0x01200, 0x01400, 0x01600, 0x01800 }, .flush_hw_mask = 0xf4ffffff, }, .pipe_vig = { .count = 4, - .base = { 0x05000, 0x07000, 0x09000, 0x0b000 }, + .base = { 0x04000, 0x06000, 0x08000, 0x0a000 }, .caps = MDP_PIPE_CAP_HFLIP | MDP_PIPE_CAP_VFLIP | MDP_PIPE_CAP_SCALE | @@ -410,7 +405,7 @@ const struct mdp5_cfg_hw msm8x96_config = { }, .pipe_rgb = { .count = 4, - .base = { 0x15000, 0x17000, 0x19000, 0x1b000 }, + .base = { 0x14000, 0x16000, 0x18000, 0x1a000 }, .caps = MDP_PIPE_CAP_HFLIP | MDP_PIPE_CAP_VFLIP | MDP_PIPE_CAP_SCALE | @@ -420,7 +415,7 @@ const struct mdp5_cfg_hw msm8x96_config = { }, .pipe_dma = { .count = 2, - .base = { 0x25000, 0x27000 }, + .base = { 0x24000, 0x26000 }, .caps = MDP_PIPE_CAP_HFLIP | MDP_PIPE_CAP_VFLIP | MDP_PIPE_CAP_SW_PIX_EXT | @@ -428,33 +423,33 @@ const struct mdp5_cfg_hw msm8x96_config = { }, .lm = { .count = 6, - .base = { 0x45000, 0x46000, 0x47000, 0x48000, 0x49000, 0x4a000 }, + .base = { 0x44000, 0x45000, 0x46000, 0x47000, 0x48000, 0x49000 }, .nb_stages = 8, .max_width = 2560, .max_height = 0xFFFF, }, .dspp = { .count = 2, - .base = { 0x55000, 0x57000 }, + .base = { 0x54000, 0x56000 }, }, .ad = { .count = 3, - .base = { 0x79000, 0x79800, 0x7a000 }, + .base = { 0x78000, 0x78800, 0x79000 }, }, .pp = { .count = 4, - .base = { 0x71000, 0x71800, 0x72000, 0x72800 }, + .base = { 0x70000, 0x70800, 0x71000, 0x71800 }, }, .cdm = { .count = 1, - .base = { 0x7a200 }, + .base = { 0x79200 }, }, .dsc = { .count = 2, - .base = { 0x81000, 0x81400 }, + .base = { 0x80000, 0x80400 }, }, .intf = { - .base = { 0x6b000, 0x6b800, 0x6c000, 0x6c800, 0x6d000 }, + .base = { 0x6a000, 0x6a800, 0x6b000, 0x6b800, 0x6c000 }, .connect = { [0] = INTF_DISABLED, [1] = INTF_DSI, diff --git a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_cmd_encoder.c b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_cmd_encoder.c index 69094cb28103..c627ab6d0061 100644 --- a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_cmd_encoder.c +++ b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_cmd_encoder.c @@ -272,22 +272,22 @@ int mdp5_cmd_encoder_set_split_display(struct drm_encoder *encoder, * start signal for the slave encoder */ if (intf_num == 1) - data |= MDP5_MDP_SPLIT_DPL_UPPER_INTF2_SW_TRG_MUX; + data |= MDP5_SPLIT_DPL_UPPER_INTF2_SW_TRG_MUX; else if (intf_num == 2) - data |= MDP5_MDP_SPLIT_DPL_UPPER_INTF1_SW_TRG_MUX; + data |= MDP5_SPLIT_DPL_UPPER_INTF1_SW_TRG_MUX; else return -EINVAL; /* Smart Panel, Sync mode */ - data |= MDP5_MDP_SPLIT_DPL_UPPER_SMART_PANEL; + data |= MDP5_SPLIT_DPL_UPPER_SMART_PANEL; /* Make sure clocks are on when connectors calling this function. */ mdp5_enable(mdp5_kms); - mdp5_write(mdp5_kms, REG_MDP5_MDP_SPLIT_DPL_UPPER(0), data); + mdp5_write(mdp5_kms, REG_MDP5_SPLIT_DPL_UPPER, data); - mdp5_write(mdp5_kms, REG_MDP5_MDP_SPLIT_DPL_LOWER(0), - MDP5_MDP_SPLIT_DPL_LOWER_SMART_PANEL); - mdp5_write(mdp5_kms, REG_MDP5_MDP_SPLIT_DPL_EN(0), 1); + mdp5_write(mdp5_kms, REG_MDP5_SPLIT_DPL_LOWER, + MDP5_SPLIT_DPL_LOWER_SMART_PANEL); + mdp5_write(mdp5_kms, REG_MDP5_SPLIT_DPL_EN, 1); mdp5_disable(mdp5_kms); return 0; diff --git a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_crtc.c b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_crtc.c index 4e8ed739f558..fa2be7ce9468 100644 --- a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_crtc.c +++ b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_crtc.c @@ -490,8 +490,7 @@ static int mdp5_crtc_cursor_set(struct drm_crtc *crtc, struct mdp5_kms *mdp5_kms = get_kms(crtc); struct drm_gem_object *cursor_bo, *old_bo = NULL; uint32_t blendcfg, cursor_addr, stride; - int ret, bpp, lm; - unsigned int depth; + int ret, lm; enum mdp5_cursor_alpha cur_alpha = CURSOR_ALPHA_PER_PIXEL; uint32_t flush_mask = mdp_ctl_flush_mask_cursor(0); uint32_t roi_w, roi_h; @@ -521,8 +520,7 @@ static int mdp5_crtc_cursor_set(struct drm_crtc *crtc, return -EINVAL; lm = mdp5_crtc->lm; - drm_fb_get_bpp_depth(DRM_FORMAT_ARGB8888, &depth, &bpp); - stride = width * (bpp >> 3); + stride = width * drm_format_plane_cpp(DRM_FORMAT_ARGB8888, 0); spin_lock_irqsave(&mdp5_crtc->cursor.lock, flags); old_bo = mdp5_crtc->cursor.scanout_bo; diff --git a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_ctl.c b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_ctl.c index 4e81ca4f964a..d021edc3b307 100644 --- a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_ctl.c +++ b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_ctl.c @@ -118,31 +118,31 @@ static void set_display_intf(struct mdp5_kms *mdp5_kms, u32 intf_sel; spin_lock_irqsave(&mdp5_kms->resource_lock, flags); - intf_sel = mdp5_read(mdp5_kms, REG_MDP5_MDP_DISP_INTF_SEL(0)); + intf_sel = mdp5_read(mdp5_kms, REG_MDP5_DISP_INTF_SEL); switch (intf->num) { case 0: - intf_sel &= ~MDP5_MDP_DISP_INTF_SEL_INTF0__MASK; - intf_sel |= MDP5_MDP_DISP_INTF_SEL_INTF0(intf->type); + intf_sel &= ~MDP5_DISP_INTF_SEL_INTF0__MASK; + intf_sel |= MDP5_DISP_INTF_SEL_INTF0(intf->type); break; case 1: - intf_sel &= ~MDP5_MDP_DISP_INTF_SEL_INTF1__MASK; - intf_sel |= MDP5_MDP_DISP_INTF_SEL_INTF1(intf->type); + intf_sel &= ~MDP5_DISP_INTF_SEL_INTF1__MASK; + intf_sel |= MDP5_DISP_INTF_SEL_INTF1(intf->type); break; case 2: - intf_sel &= ~MDP5_MDP_DISP_INTF_SEL_INTF2__MASK; - intf_sel |= MDP5_MDP_DISP_INTF_SEL_INTF2(intf->type); + intf_sel &= ~MDP5_DISP_INTF_SEL_INTF2__MASK; + intf_sel |= MDP5_DISP_INTF_SEL_INTF2(intf->type); break; case 3: - intf_sel &= ~MDP5_MDP_DISP_INTF_SEL_INTF3__MASK; - intf_sel |= MDP5_MDP_DISP_INTF_SEL_INTF3(intf->type); + intf_sel &= ~MDP5_DISP_INTF_SEL_INTF3__MASK; + intf_sel |= MDP5_DISP_INTF_SEL_INTF3(intf->type); break; default: BUG(); break; } - mdp5_write(mdp5_kms, REG_MDP5_MDP_DISP_INTF_SEL(0), intf_sel); + mdp5_write(mdp5_kms, REG_MDP5_DISP_INTF_SEL, intf_sel); spin_unlock_irqrestore(&mdp5_kms->resource_lock, flags); } @@ -557,7 +557,7 @@ int mdp5_ctl_pair(struct mdp5_ctl *ctlx, struct mdp5_ctl *ctly, bool enable) if (!enable) { ctlx->pair = NULL; ctly->pair = NULL; - mdp5_write(mdp5_kms, REG_MDP5_MDP_SPARE_0(0), 0); + mdp5_write(mdp5_kms, REG_MDP5_SPARE_0, 0); return 0; } else if ((ctlx->pair != NULL) || (ctly->pair != NULL)) { dev_err(ctl_mgr->dev->dev, "CTLs already paired\n"); @@ -570,8 +570,8 @@ int mdp5_ctl_pair(struct mdp5_ctl *ctlx, struct mdp5_ctl *ctly, bool enable) ctlx->pair = ctly; ctly->pair = ctlx; - mdp5_write(mdp5_kms, REG_MDP5_MDP_SPARE_0(0), - MDP5_MDP_SPARE_0_SPLIT_DPL_SINGLE_FLUSH_EN); + mdp5_write(mdp5_kms, REG_MDP5_SPARE_0, + MDP5_SPARE_0_SPLIT_DPL_SINGLE_FLUSH_EN); return 0; } diff --git a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_encoder.c b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_encoder.c index 1d95f9fd9dc7..fe0c22230883 100644 --- a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_encoder.c +++ b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_encoder.c @@ -322,18 +322,18 @@ int mdp5_encoder_set_split_display(struct drm_encoder *encoder, * to use the master's enable signal for the slave encoder. */ if (intf_num == 1) - data |= MDP5_MDP_SPLIT_DPL_LOWER_INTF2_TG_SYNC; + data |= MDP5_SPLIT_DPL_LOWER_INTF2_TG_SYNC; else if (intf_num == 2) - data |= MDP5_MDP_SPLIT_DPL_LOWER_INTF1_TG_SYNC; + data |= MDP5_SPLIT_DPL_LOWER_INTF1_TG_SYNC; else return -EINVAL; /* Make sure clocks are on when connectors calling this function. */ mdp5_enable(mdp5_kms); /* Dumb Panel, Sync mode */ - mdp5_write(mdp5_kms, REG_MDP5_MDP_SPLIT_DPL_UPPER(0), 0); - mdp5_write(mdp5_kms, REG_MDP5_MDP_SPLIT_DPL_LOWER(0), data); - mdp5_write(mdp5_kms, REG_MDP5_MDP_SPLIT_DPL_EN(0), 1); + mdp5_write(mdp5_kms, REG_MDP5_SPLIT_DPL_UPPER, 0); + mdp5_write(mdp5_kms, REG_MDP5_SPLIT_DPL_LOWER, data); + mdp5_write(mdp5_kms, REG_MDP5_SPLIT_DPL_EN, 1); mdp5_ctl_pair(mdp5_encoder->ctl, mdp5_slave_enc->ctl, true); diff --git a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_irq.c b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_irq.c index 73bc3e312fd4..d53e5510fd7c 100644 --- a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_irq.c +++ b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_irq.c @@ -15,7 +15,6 @@ * this program. If not, see <http://www.gnu.org/licenses/>. */ -#include <linux/irqdomain.h> #include <linux/irq.h> #include "msm_drv.h" @@ -24,9 +23,9 @@ void mdp5_set_irqmask(struct mdp_kms *mdp_kms, uint32_t irqmask, uint32_t old_irqmask) { - mdp5_write(to_mdp5_kms(mdp_kms), REG_MDP5_MDP_INTR_CLEAR(0), - irqmask ^ (irqmask & old_irqmask)); - mdp5_write(to_mdp5_kms(mdp_kms), REG_MDP5_MDP_INTR_EN(0), irqmask); + mdp5_write(to_mdp5_kms(mdp_kms), REG_MDP5_INTR_CLEAR, + irqmask ^ (irqmask & old_irqmask)); + mdp5_write(to_mdp5_kms(mdp_kms), REG_MDP5_INTR_EN, irqmask); } static void mdp5_irq_error_handler(struct mdp_irq *irq, uint32_t irqstatus) @@ -38,8 +37,8 @@ void mdp5_irq_preinstall(struct msm_kms *kms) { struct mdp5_kms *mdp5_kms = to_mdp5_kms(to_mdp_kms(kms)); mdp5_enable(mdp5_kms); - mdp5_write(mdp5_kms, REG_MDP5_MDP_INTR_CLEAR(0), 0xffffffff); - mdp5_write(mdp5_kms, REG_MDP5_MDP_INTR_EN(0), 0x00000000); + mdp5_write(mdp5_kms, REG_MDP5_INTR_CLEAR, 0xffffffff); + mdp5_write(mdp5_kms, REG_MDP5_INTR_EN, 0x00000000); mdp5_disable(mdp5_kms); } @@ -55,7 +54,9 @@ int mdp5_irq_postinstall(struct msm_kms *kms) MDP5_IRQ_INTF2_UNDER_RUN | MDP5_IRQ_INTF3_UNDER_RUN; + mdp5_enable(mdp5_kms); mdp_irq_register(mdp_kms, error_handler); + mdp5_disable(mdp5_kms); return 0; } @@ -64,21 +65,22 @@ void mdp5_irq_uninstall(struct msm_kms *kms) { struct mdp5_kms *mdp5_kms = to_mdp5_kms(to_mdp_kms(kms)); mdp5_enable(mdp5_kms); - mdp5_write(mdp5_kms, REG_MDP5_MDP_INTR_EN(0), 0x00000000); + mdp5_write(mdp5_kms, REG_MDP5_INTR_EN, 0x00000000); mdp5_disable(mdp5_kms); } -static void mdp5_irq_mdp(struct mdp_kms *mdp_kms) +irqreturn_t mdp5_irq(struct msm_kms *kms) { + struct mdp_kms *mdp_kms = to_mdp_kms(kms); struct mdp5_kms *mdp5_kms = to_mdp5_kms(mdp_kms); struct drm_device *dev = mdp5_kms->dev; struct msm_drm_private *priv = dev->dev_private; unsigned int id; uint32_t status, enable; - enable = mdp5_read(mdp5_kms, REG_MDP5_MDP_INTR_EN(0)); - status = mdp5_read(mdp5_kms, REG_MDP5_MDP_INTR_STATUS(0)) & enable; - mdp5_write(mdp5_kms, REG_MDP5_MDP_INTR_CLEAR(0), status); + enable = mdp5_read(mdp5_kms, REG_MDP5_INTR_EN); + status = mdp5_read(mdp5_kms, REG_MDP5_INTR_STATUS) & enable; + mdp5_write(mdp5_kms, REG_MDP5_INTR_CLEAR, status); VERB("status=%08x", status); @@ -87,29 +89,6 @@ static void mdp5_irq_mdp(struct mdp_kms *mdp_kms) for (id = 0; id < priv->num_crtcs; id++) if (status & mdp5_crtc_vblank(priv->crtcs[id])) drm_handle_vblank(dev, id); -} - -irqreturn_t mdp5_irq(struct msm_kms *kms) -{ - struct mdp_kms *mdp_kms = to_mdp_kms(kms); - struct mdp5_kms *mdp5_kms = to_mdp5_kms(mdp_kms); - uint32_t intr; - - intr = mdp5_read(mdp5_kms, REG_MDSS_HW_INTR_STATUS); - - VERB("intr=%08x", intr); - - if (intr & MDSS_HW_INTR_STATUS_INTR_MDP) { - mdp5_irq_mdp(mdp_kms); - intr &= ~MDSS_HW_INTR_STATUS_INTR_MDP; - } - - while (intr) { - irq_hw_number_t hwirq = fls(intr) - 1; - generic_handle_irq(irq_find_mapping( - mdp5_kms->irqcontroller.domain, hwirq)); - intr &= ~(1 << hwirq); - } return IRQ_HANDLED; } @@ -135,81 +114,3 @@ void mdp5_disable_vblank(struct msm_kms *kms, struct drm_crtc *crtc) mdp5_crtc_vblank(crtc), false); mdp5_disable(mdp5_kms); } - -/* - * interrupt-controller implementation, so sub-blocks (hdmi/eDP/dsi/etc) - * can register to get their irq's delivered - */ - -#define VALID_IRQS (MDSS_HW_INTR_STATUS_INTR_DSI0 | \ - MDSS_HW_INTR_STATUS_INTR_DSI1 | \ - MDSS_HW_INTR_STATUS_INTR_HDMI | \ - MDSS_HW_INTR_STATUS_INTR_EDP) - -static void mdp5_hw_mask_irq(struct irq_data *irqd) -{ - struct mdp5_kms *mdp5_kms = irq_data_get_irq_chip_data(irqd); - smp_mb__before_atomic(); - clear_bit(irqd->hwirq, &mdp5_kms->irqcontroller.enabled_mask); - smp_mb__after_atomic(); -} - -static void mdp5_hw_unmask_irq(struct irq_data *irqd) -{ - struct mdp5_kms *mdp5_kms = irq_data_get_irq_chip_data(irqd); - smp_mb__before_atomic(); - set_bit(irqd->hwirq, &mdp5_kms->irqcontroller.enabled_mask); - smp_mb__after_atomic(); -} - -static struct irq_chip mdp5_hw_irq_chip = { - .name = "mdp5", - .irq_mask = mdp5_hw_mask_irq, - .irq_unmask = mdp5_hw_unmask_irq, -}; - -static int mdp5_hw_irqdomain_map(struct irq_domain *d, - unsigned int irq, irq_hw_number_t hwirq) -{ - struct mdp5_kms *mdp5_kms = d->host_data; - - if (!(VALID_IRQS & (1 << hwirq))) - return -EPERM; - - irq_set_chip_and_handler(irq, &mdp5_hw_irq_chip, handle_level_irq); - irq_set_chip_data(irq, mdp5_kms); - - return 0; -} - -static struct irq_domain_ops mdp5_hw_irqdomain_ops = { - .map = mdp5_hw_irqdomain_map, - .xlate = irq_domain_xlate_onecell, -}; - - -int mdp5_irq_domain_init(struct mdp5_kms *mdp5_kms) -{ - struct device *dev = mdp5_kms->dev->dev; - struct irq_domain *d; - - d = irq_domain_add_linear(dev->of_node, 32, - &mdp5_hw_irqdomain_ops, mdp5_kms); - if (!d) { - dev_err(dev, "mdp5 irq domain add failed\n"); - return -ENXIO; - } - - mdp5_kms->irqcontroller.enabled_mask = 0; - mdp5_kms->irqcontroller.domain = d; - - return 0; -} - -void mdp5_irq_domain_fini(struct mdp5_kms *mdp5_kms) -{ - if (mdp5_kms->irqcontroller.domain) { - irq_domain_remove(mdp5_kms->irqcontroller.domain); - mdp5_kms->irqcontroller.domain = NULL; - } -} diff --git a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.c b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.c index f0c285b1c027..ed7143d35b25 100644 --- a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.c +++ b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.c @@ -16,6 +16,7 @@ * this program. If not, see <http://www.gnu.org/licenses/>. */ +#include <linux/of_irq.h> #include "msm_drv.h" #include "msm_mmu.h" @@ -28,10 +29,11 @@ static const char *iommu_ports[] = { static int mdp5_hw_init(struct msm_kms *kms) { struct mdp5_kms *mdp5_kms = to_mdp5_kms(to_mdp_kms(kms)); - struct drm_device *dev = mdp5_kms->dev; + struct platform_device *pdev = mdp5_kms->pdev; unsigned long flags; - pm_runtime_get_sync(dev->dev); + pm_runtime_get_sync(&pdev->dev); + mdp5_enable(mdp5_kms); /* Magic unknown register writes: * @@ -58,12 +60,13 @@ static int mdp5_hw_init(struct msm_kms *kms) */ spin_lock_irqsave(&mdp5_kms->resource_lock, flags); - mdp5_write(mdp5_kms, REG_MDP5_MDP_DISP_INTF_SEL(0), 0); + mdp5_write(mdp5_kms, REG_MDP5_DISP_INTF_SEL, 0); spin_unlock_irqrestore(&mdp5_kms->resource_lock, flags); mdp5_ctlm_hw_reset(mdp5_kms->ctlm); - pm_runtime_put_sync(dev->dev); + mdp5_disable(mdp5_kms); + pm_runtime_put_sync(&pdev->dev); return 0; } @@ -111,26 +114,15 @@ static int mdp5_set_split_display(struct msm_kms *kms, return mdp5_encoder_set_split_display(encoder, slave_encoder); } -static void mdp5_destroy(struct msm_kms *kms) +static void mdp5_kms_destroy(struct msm_kms *kms) { struct mdp5_kms *mdp5_kms = to_mdp5_kms(to_mdp_kms(kms)); struct msm_mmu *mmu = mdp5_kms->mmu; - mdp5_irq_domain_fini(mdp5_kms); - if (mmu) { mmu->funcs->detach(mmu, iommu_ports, ARRAY_SIZE(iommu_ports)); mmu->funcs->destroy(mmu); } - - if (mdp5_kms->ctlm) - mdp5_ctlm_destroy(mdp5_kms->ctlm); - if (mdp5_kms->smp) - mdp5_smp_destroy(mdp5_kms->smp); - if (mdp5_kms->cfg) - mdp5_cfg_destroy(mdp5_kms->cfg); - - kfree(mdp5_kms); } static const struct mdp_kms_funcs kms_funcs = { @@ -148,7 +140,7 @@ static const struct mdp_kms_funcs kms_funcs = { .get_format = mdp_get_format, .round_pixclk = mdp5_round_pixclk, .set_split_display = mdp5_set_split_display, - .destroy = mdp5_destroy, + .destroy = mdp5_kms_destroy, }, .set_irqmask = mdp5_set_irqmask, }; @@ -345,13 +337,6 @@ static int modeset_init(struct mdp5_kms *mdp5_kms) hw_cfg = mdp5_cfg_get_hw_config(mdp5_kms->cfg); - /* register our interrupt-controller for hdmi/eDP/dsi/etc - * to use for irqs routed through mdp: - */ - ret = mdp5_irq_domain_init(mdp5_kms); - if (ret) - goto fail; - /* construct CRTCs and their private planes: */ for (i = 0; i < hw_cfg->pipe_rgb.count; i++) { struct drm_plane *plane; @@ -419,17 +404,17 @@ fail: return ret; } -static void read_hw_revision(struct mdp5_kms *mdp5_kms, - uint32_t *major, uint32_t *minor) +static void read_mdp_hw_revision(struct mdp5_kms *mdp5_kms, + u32 *major, u32 *minor) { - uint32_t version; + u32 version; mdp5_enable(mdp5_kms); - version = mdp5_read(mdp5_kms, REG_MDSS_HW_VERSION); + version = mdp5_read(mdp5_kms, REG_MDP5_HW_VERSION); mdp5_disable(mdp5_kms); - *major = FIELD(version, MDSS_HW_VERSION_MAJOR); - *minor = FIELD(version, MDSS_HW_VERSION_MINOR); + *major = FIELD(version, MDP5_HW_VERSION_MAJOR); + *minor = FIELD(version, MDP5_HW_VERSION_MINOR); DBG("MDP5 version v%d.%d", *major, *minor); } @@ -574,51 +559,146 @@ static u32 mdp5_get_vblank_counter(struct drm_device *dev, unsigned int pipe) struct msm_kms *mdp5_kms_init(struct drm_device *dev) { - struct platform_device *pdev = dev->platformdev; - struct mdp5_cfg *config; + struct msm_drm_private *priv = dev->dev_private; + struct platform_device *pdev; struct mdp5_kms *mdp5_kms; - struct msm_kms *kms = NULL; + struct mdp5_cfg *config; + struct msm_kms *kms; struct msm_mmu *mmu; - uint32_t major, minor; - int i, ret; + int irq, i, ret; - mdp5_kms = kzalloc(sizeof(*mdp5_kms), GFP_KERNEL); - if (!mdp5_kms) { - dev_err(dev->dev, "failed to allocate kms\n"); - ret = -ENOMEM; + /* priv->kms would have been populated by the MDP5 driver */ + kms = priv->kms; + if (!kms) + return NULL; + + mdp5_kms = to_mdp5_kms(to_mdp_kms(kms)); + + mdp_kms_init(&mdp5_kms->base, &kms_funcs); + + pdev = mdp5_kms->pdev; + + irq = irq_of_parse_and_map(pdev->dev.of_node, 0); + if (irq < 0) { + ret = irq; + dev_err(&pdev->dev, "failed to get irq: %d\n", ret); goto fail; } - spin_lock_init(&mdp5_kms->resource_lock); + kms->irq = irq; - mdp_kms_init(&mdp5_kms->base, &kms_funcs); + config = mdp5_cfg_get_config(mdp5_kms->cfg); - kms = &mdp5_kms->base.base; + /* make sure things are off before attaching iommu (bootloader could + * have left things on, in which case we'll start getting faults if + * we don't disable): + */ + mdp5_enable(mdp5_kms); + for (i = 0; i < MDP5_INTF_NUM_MAX; i++) { + if (mdp5_cfg_intf_is_virtual(config->hw->intf.connect[i]) || + !config->hw->intf.base[i]) + continue; + mdp5_write(mdp5_kms, REG_MDP5_INTF_TIMING_ENGINE_EN(i), 0); - mdp5_kms->dev = dev; + mdp5_write(mdp5_kms, REG_MDP5_INTF_FRAME_LINE_COUNT_EN(i), 0x3); + } + mdp5_disable(mdp5_kms); + mdelay(16); - /* mdp5_kms->mmio actually represents the MDSS base address */ - mdp5_kms->mmio = msm_ioremap(pdev, "mdp_phys", "MDP5"); - if (IS_ERR(mdp5_kms->mmio)) { - ret = PTR_ERR(mdp5_kms->mmio); + if (config->platform.iommu) { + mmu = msm_iommu_new(&pdev->dev, config->platform.iommu); + if (IS_ERR(mmu)) { + ret = PTR_ERR(mmu); + dev_err(&pdev->dev, "failed to init iommu: %d\n", ret); + iommu_domain_free(config->platform.iommu); + goto fail; + } + + ret = mmu->funcs->attach(mmu, iommu_ports, + ARRAY_SIZE(iommu_ports)); + if (ret) { + dev_err(&pdev->dev, "failed to attach iommu: %d\n", + ret); + mmu->funcs->destroy(mmu); + goto fail; + } + } else { + dev_info(&pdev->dev, + "no iommu, fallback to phys contig buffers for scanout\n"); + mmu = NULL; + } + mdp5_kms->mmu = mmu; + + mdp5_kms->id = msm_register_mmu(dev, mmu); + if (mdp5_kms->id < 0) { + ret = mdp5_kms->id; + dev_err(&pdev->dev, "failed to register mdp5 iommu: %d\n", ret); goto fail; } - mdp5_kms->vbif = msm_ioremap(pdev, "vbif_phys", "VBIF"); - if (IS_ERR(mdp5_kms->vbif)) { - ret = PTR_ERR(mdp5_kms->vbif); + ret = modeset_init(mdp5_kms); + if (ret) { + dev_err(&pdev->dev, "modeset_init failed: %d\n", ret); goto fail; } - mdp5_kms->vdd = devm_regulator_get(&pdev->dev, "vdd"); - if (IS_ERR(mdp5_kms->vdd)) { - ret = PTR_ERR(mdp5_kms->vdd); + dev->mode_config.min_width = 0; + dev->mode_config.min_height = 0; + dev->mode_config.max_width = config->hw->lm.max_width; + dev->mode_config.max_height = config->hw->lm.max_height; + + dev->driver->get_vblank_timestamp = mdp5_get_vblank_timestamp; + dev->driver->get_scanout_position = mdp5_get_scanoutpos; + dev->driver->get_vblank_counter = mdp5_get_vblank_counter; + dev->max_vblank_count = 0xffffffff; + dev->vblank_disable_immediate = true; + + return kms; +fail: + if (kms) + mdp5_kms_destroy(kms); + return ERR_PTR(ret); +} + +static void mdp5_destroy(struct platform_device *pdev) +{ + struct mdp5_kms *mdp5_kms = platform_get_drvdata(pdev); + + if (mdp5_kms->ctlm) + mdp5_ctlm_destroy(mdp5_kms->ctlm); + if (mdp5_kms->smp) + mdp5_smp_destroy(mdp5_kms->smp); + if (mdp5_kms->cfg) + mdp5_cfg_destroy(mdp5_kms->cfg); + + if (mdp5_kms->rpm_enabled) + pm_runtime_disable(&pdev->dev); +} + +static int mdp5_init(struct platform_device *pdev, struct drm_device *dev) +{ + struct msm_drm_private *priv = dev->dev_private; + struct mdp5_kms *mdp5_kms; + struct mdp5_cfg *config; + u32 major, minor; + int ret; + + mdp5_kms = devm_kzalloc(&pdev->dev, sizeof(*mdp5_kms), GFP_KERNEL); + if (!mdp5_kms) { + ret = -ENOMEM; goto fail; } - ret = regulator_enable(mdp5_kms->vdd); - if (ret) { - dev_err(dev->dev, "failed to enable regulator vdd: %d\n", ret); + platform_set_drvdata(pdev, mdp5_kms); + + spin_lock_init(&mdp5_kms->resource_lock); + + mdp5_kms->dev = dev; + mdp5_kms->pdev = pdev; + + mdp5_kms->mmio = msm_ioremap(pdev, "mdp_phys", "MDP5"); + if (IS_ERR(mdp5_kms->mmio)) { + ret = PTR_ERR(mdp5_kms->mmio); goto fail; } @@ -629,9 +709,6 @@ struct msm_kms *mdp5_kms_init(struct drm_device *dev) ret = get_clk(pdev, &mdp5_kms->ahb_clk, "iface_clk", true); if (ret) goto fail; - ret = get_clk(pdev, &mdp5_kms->src_clk, "core_clk_src", true); - if (ret) - goto fail; ret = get_clk(pdev, &mdp5_kms->core_clk, "core_clk", true); if (ret) goto fail; @@ -646,9 +723,12 @@ struct msm_kms *mdp5_kms_init(struct drm_device *dev) * rate first, then figure out hw revision, and then set a * more optimal rate: */ - clk_set_rate(mdp5_kms->src_clk, 200000000); + clk_set_rate(mdp5_kms->core_clk, 200000000); + + pm_runtime_enable(&pdev->dev); + mdp5_kms->rpm_enabled = true; - read_hw_revision(mdp5_kms, &major, &minor); + read_mdp_hw_revision(mdp5_kms, &major, &minor); mdp5_kms->cfg = mdp5_cfg_init(mdp5_kms, major, minor); if (IS_ERR(mdp5_kms->cfg)) { @@ -661,7 +741,7 @@ struct msm_kms *mdp5_kms_init(struct drm_device *dev) mdp5_kms->caps = config->hw->mdp.caps; /* TODO: compute core clock rate at runtime */ - clk_set_rate(mdp5_kms->src_clk, config->hw->max_clk); + clk_set_rate(mdp5_kms->core_clk, config->hw->max_clk); /* * Some chipsets have a Shared Memory Pool (SMP), while others @@ -684,73 +764,76 @@ struct msm_kms *mdp5_kms_init(struct drm_device *dev) goto fail; } - /* make sure things are off before attaching iommu (bootloader could - * have left things on, in which case we'll start getting faults if - * we don't disable): - */ - mdp5_enable(mdp5_kms); - for (i = 0; i < MDP5_INTF_NUM_MAX; i++) { - if (mdp5_cfg_intf_is_virtual(config->hw->intf.connect[i]) || - !config->hw->intf.base[i]) - continue; - mdp5_write(mdp5_kms, REG_MDP5_INTF_TIMING_ENGINE_EN(i), 0); + /* set uninit-ed kms */ + priv->kms = &mdp5_kms->base.base; - mdp5_write(mdp5_kms, REG_MDP5_INTF_FRAME_LINE_COUNT_EN(i), 0x3); - } - mdp5_disable(mdp5_kms); - mdelay(16); + return 0; +fail: + mdp5_destroy(pdev); + return ret; +} - if (config->platform.iommu) { - mmu = msm_iommu_new(&pdev->dev, config->platform.iommu); - if (IS_ERR(mmu)) { - ret = PTR_ERR(mmu); - dev_err(dev->dev, "failed to init iommu: %d\n", ret); - iommu_domain_free(config->platform.iommu); - goto fail; - } +static int mdp5_bind(struct device *dev, struct device *master, void *data) +{ + struct drm_device *ddev = dev_get_drvdata(master); + struct platform_device *pdev = to_platform_device(dev); - ret = mmu->funcs->attach(mmu, iommu_ports, - ARRAY_SIZE(iommu_ports)); - if (ret) { - dev_err(dev->dev, "failed to attach iommu: %d\n", ret); - mmu->funcs->destroy(mmu); - goto fail; - } - } else { - dev_info(dev->dev, "no iommu, fallback to phys " - "contig buffers for scanout\n"); - mmu = NULL; - } - mdp5_kms->mmu = mmu; + DBG(""); - mdp5_kms->id = msm_register_mmu(dev, mmu); - if (mdp5_kms->id < 0) { - ret = mdp5_kms->id; - dev_err(dev->dev, "failed to register mdp5 iommu: %d\n", ret); - goto fail; - } + return mdp5_init(pdev, ddev); +} - ret = modeset_init(mdp5_kms); - if (ret) { - dev_err(dev->dev, "modeset_init failed: %d\n", ret); - goto fail; - } +static void mdp5_unbind(struct device *dev, struct device *master, + void *data) +{ + struct platform_device *pdev = to_platform_device(dev); - dev->mode_config.min_width = 0; - dev->mode_config.min_height = 0; - dev->mode_config.max_width = config->hw->lm.max_width; - dev->mode_config.max_height = config->hw->lm.max_height; + mdp5_destroy(pdev); +} - dev->driver->get_vblank_timestamp = mdp5_get_vblank_timestamp; - dev->driver->get_scanout_position = mdp5_get_scanoutpos; - dev->driver->get_vblank_counter = mdp5_get_vblank_counter; - dev->max_vblank_count = 0xffffffff; - dev->vblank_disable_immediate = true; +static const struct component_ops mdp5_ops = { + .bind = mdp5_bind, + .unbind = mdp5_unbind, +}; - return kms; +static int mdp5_dev_probe(struct platform_device *pdev) +{ + DBG(""); + return component_add(&pdev->dev, &mdp5_ops); +} -fail: - if (kms) - mdp5_destroy(kms); - return ERR_PTR(ret); +static int mdp5_dev_remove(struct platform_device *pdev) +{ + DBG(""); + component_del(&pdev->dev, &mdp5_ops); + return 0; +} + +static const struct of_device_id mdp5_dt_match[] = { + { .compatible = "qcom,mdp5", }, + /* to support downstream DT files */ + { .compatible = "qcom,mdss_mdp", }, + {} +}; +MODULE_DEVICE_TABLE(of, mdp5_dt_match); + +static struct platform_driver mdp5_driver = { + .probe = mdp5_dev_probe, + .remove = mdp5_dev_remove, + .driver = { + .name = "msm_mdp", + .of_match_table = mdp5_dt_match, + }, +}; + +void __init msm_mdp_register(void) +{ + DBG(""); + platform_driver_register(&mdp5_driver); +} + +void __exit msm_mdp_unregister(void) +{ + DBG(""); + platform_driver_unregister(&mdp5_driver); } diff --git a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.h b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.h index 9a25898239d3..03738927be10 100644 --- a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.h +++ b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.h @@ -31,6 +31,8 @@ struct mdp5_kms { struct drm_device *dev; + struct platform_device *pdev; + struct mdp5_cfg_handler *cfg; uint32_t caps; /* MDP capabilities (MDP_CAP_XXX bits) */ @@ -43,29 +45,23 @@ struct mdp5_kms { struct mdp5_ctl_manager *ctlm; /* io/register spaces: */ - void __iomem *mmio, *vbif; - - struct regulator *vdd; + void __iomem *mmio; struct clk *axi_clk; struct clk *ahb_clk; - struct clk *src_clk; struct clk *core_clk; struct clk *lut_clk; struct clk *vsync_clk; /* * lock to protect access to global resources: ie., following register: - * - REG_MDP5_MDP_DISP_INTF_SEL + * - REG_MDP5_DISP_INTF_SEL */ spinlock_t resource_lock; - struct mdp_irq error_handler; + bool rpm_enabled; - struct { - volatile unsigned long enabled_mask; - struct irq_domain *domain; - } irqcontroller; + struct mdp_irq error_handler; }; #define to_mdp5_kms(x) container_of(x, struct mdp5_kms, base) diff --git a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_mdss.c b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_mdss.c new file mode 100644 index 000000000000..d444a6901fff --- /dev/null +++ b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_mdss.c @@ -0,0 +1,235 @@ +/* + * Copyright (c) 2016, The Linux Foundation. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License version 2 as published by + * the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program. If not, see <http://www.gnu.org/licenses/>. + */ + +#include <linux/irqdomain.h> +#include <linux/irq.h> + +#include "msm_drv.h" +#include "mdp5_kms.h" + +/* + * If needed, this can become more specific: something like struct mdp5_mdss, + * which contains a 'struct msm_mdss base' member. + */ +struct msm_mdss { + struct drm_device *dev; + + void __iomem *mmio, *vbif; + + struct regulator *vdd; + + struct { + volatile unsigned long enabled_mask; + struct irq_domain *domain; + } irqcontroller; +}; + +static inline void mdss_write(struct msm_mdss *mdss, u32 reg, u32 data) +{ + msm_writel(data, mdss->mmio + reg); +} + +static inline u32 mdss_read(struct msm_mdss *mdss, u32 reg) +{ + return msm_readl(mdss->mmio + reg); +} + +static irqreturn_t mdss_irq(int irq, void *arg) +{ + struct msm_mdss *mdss = arg; + u32 intr; + + intr = mdss_read(mdss, REG_MDSS_HW_INTR_STATUS); + + VERB("intr=%08x", intr); + + while (intr) { + irq_hw_number_t hwirq = fls(intr) - 1; + + generic_handle_irq(irq_find_mapping( + mdss->irqcontroller.domain, hwirq)); + intr &= ~(1 << hwirq); + } + + return IRQ_HANDLED; +} + +/* + * interrupt-controller implementation, so sub-blocks (MDP/HDMI/eDP/DSI/etc) + * can register to get their irq's delivered + */ + +#define VALID_IRQS (MDSS_HW_INTR_STATUS_INTR_MDP | \ + MDSS_HW_INTR_STATUS_INTR_DSI0 | \ + MDSS_HW_INTR_STATUS_INTR_DSI1 | \ + MDSS_HW_INTR_STATUS_INTR_HDMI | \ + MDSS_HW_INTR_STATUS_INTR_EDP) + +static void mdss_hw_mask_irq(struct irq_data *irqd) +{ + struct msm_mdss *mdss = irq_data_get_irq_chip_data(irqd); + + smp_mb__before_atomic(); + clear_bit(irqd->hwirq, &mdss->irqcontroller.enabled_mask); + smp_mb__after_atomic(); +} + +static void mdss_hw_unmask_irq(struct irq_data *irqd) +{ + struct msm_mdss *mdss = irq_data_get_irq_chip_data(irqd); + + smp_mb__before_atomic(); + set_bit(irqd->hwirq, &mdss->irqcontroller.enabled_mask); + smp_mb__after_atomic(); +} + +static struct irq_chip mdss_hw_irq_chip = { + .name = "mdss", + .irq_mask = mdss_hw_mask_irq, + .irq_unmask = mdss_hw_unmask_irq, +}; + +static int mdss_hw_irqdomain_map(struct irq_domain *d, unsigned int irq, + irq_hw_number_t hwirq) +{ + struct msm_mdss *mdss = d->host_data; + + if (!(VALID_IRQS & (1 << hwirq))) + return -EPERM; + + irq_set_chip_and_handler(irq, &mdss_hw_irq_chip, handle_level_irq); + irq_set_chip_data(irq, mdss); + + return 0; +} + +static struct irq_domain_ops mdss_hw_irqdomain_ops = { + .map = mdss_hw_irqdomain_map, + .xlate = irq_domain_xlate_onecell, +}; + + +static int mdss_irq_domain_init(struct msm_mdss *mdss) +{ + struct device *dev = mdss->dev->dev; + struct irq_domain *d; + + d = irq_domain_add_linear(dev->of_node, 32, &mdss_hw_irqdomain_ops, + mdss); + if (!d) { + dev_err(dev, "mdss irq domain add failed\n"); + return -ENXIO; + } + + mdss->irqcontroller.enabled_mask = 0; + mdss->irqcontroller.domain = d; + + return 0; +} + +void msm_mdss_destroy(struct drm_device *dev) +{ + struct msm_drm_private *priv = dev->dev_private; + struct msm_mdss *mdss = priv->mdss; + + if (!mdss) + return; + + irq_domain_remove(mdss->irqcontroller.domain); + mdss->irqcontroller.domain = NULL; + + regulator_disable(mdss->vdd); + + pm_runtime_put_sync(dev->dev); + + pm_runtime_disable(dev->dev); +} + +int msm_mdss_init(struct drm_device *dev) +{ + struct platform_device *pdev = dev->platformdev; + struct msm_drm_private *priv = dev->dev_private; + struct msm_mdss *mdss; + int ret; + + DBG(""); + + if (!of_device_is_compatible(dev->dev->of_node, "qcom,mdss")) + return 0; + + mdss = devm_kzalloc(dev->dev, sizeof(*mdss), GFP_KERNEL); + if (!mdss) { + ret = -ENOMEM; + goto fail; + } + + mdss->dev = dev; + + mdss->mmio = msm_ioremap(pdev, "mdss_phys", "MDSS"); + if (IS_ERR(mdss->mmio)) { + ret = PTR_ERR(mdss->mmio); + goto fail; + } + + mdss->vbif = msm_ioremap(pdev, "vbif_phys", "VBIF"); + if (IS_ERR(mdss->vbif)) { + ret = PTR_ERR(mdss->vbif); + goto fail; + } + + /* Regulator to enable GDSCs in downstream kernels */ + mdss->vdd = devm_regulator_get(dev->dev, "vdd"); + if (IS_ERR(mdss->vdd)) { + ret = PTR_ERR(mdss->vdd); + goto fail; + } + + ret = regulator_enable(mdss->vdd); + if (ret) { + dev_err(dev->dev, "failed to enable regulator vdd: %d\n", + ret); + goto fail; + } + + ret = devm_request_irq(dev->dev, platform_get_irq(pdev, 0), + mdss_irq, 0, "mdss_isr", mdss); + if (ret) { + dev_err(dev->dev, "failed to init irq: %d\n", ret); + goto fail_irq; + } + + ret = mdss_irq_domain_init(mdss); + if (ret) { + dev_err(dev->dev, "failed to init sub-block irqs: %d\n", ret); + goto fail_irq; + } + + priv->mdss = mdss; + + pm_runtime_enable(dev->dev); + + /* + * TODO: This is needed as the MDSS GDSC is only tied to MDSS's power + * domain. Remove this once runtime PM is adapted for all the devices. + */ + pm_runtime_get_sync(dev->dev); + + return 0; +fail_irq: + regulator_disable(mdss->vdd); +fail: + return ret; +} diff --git a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_smp.c b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_smp.c index 6f425c25d9fe..27d7b55b52c9 100644 --- a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_smp.c +++ b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_smp.c @@ -42,7 +42,7 @@ * * configured: * The block is allocated to some client, and assigned to that - * client in MDP5_MDP_SMP_ALLOC registers. + * client in MDP5_SMP_ALLOC registers. * * inuse: * The block is being actively used by a client. @@ -59,7 +59,7 @@ * mdp5_smp_commit. * * 2) mdp5_smp_configure(): - * As hw is programmed, before FLUSH, MDP5_MDP_SMP_ALLOC registers + * As hw is programmed, before FLUSH, MDP5_SMP_ALLOC registers * are configured for the union(pending, inuse) * Current pending is copied to configured. * It is assumed that mdp5_smp_request and mdp5_smp_configure not run @@ -311,25 +311,25 @@ static void update_smp_state(struct mdp5_smp *smp, int idx = blk / 3; int fld = blk % 3; - val = mdp5_read(mdp5_kms, REG_MDP5_MDP_SMP_ALLOC_W_REG(0, idx)); + val = mdp5_read(mdp5_kms, REG_MDP5_SMP_ALLOC_W_REG(idx)); switch (fld) { case 0: - val &= ~MDP5_MDP_SMP_ALLOC_W_REG_CLIENT0__MASK; - val |= MDP5_MDP_SMP_ALLOC_W_REG_CLIENT0(cid); + val &= ~MDP5_SMP_ALLOC_W_REG_CLIENT0__MASK; + val |= MDP5_SMP_ALLOC_W_REG_CLIENT0(cid); break; case 1: - val &= ~MDP5_MDP_SMP_ALLOC_W_REG_CLIENT1__MASK; - val |= MDP5_MDP_SMP_ALLOC_W_REG_CLIENT1(cid); + val &= ~MDP5_SMP_ALLOC_W_REG_CLIENT1__MASK; + val |= MDP5_SMP_ALLOC_W_REG_CLIENT1(cid); break; case 2: - val &= ~MDP5_MDP_SMP_ALLOC_W_REG_CLIENT2__MASK; - val |= MDP5_MDP_SMP_ALLOC_W_REG_CLIENT2(cid); + val &= ~MDP5_SMP_ALLOC_W_REG_CLIENT2__MASK; + val |= MDP5_SMP_ALLOC_W_REG_CLIENT2(cid); break; } - mdp5_write(mdp5_kms, REG_MDP5_MDP_SMP_ALLOC_W_REG(0, idx), val); - mdp5_write(mdp5_kms, REG_MDP5_MDP_SMP_ALLOC_R_REG(0, idx), val); + mdp5_write(mdp5_kms, REG_MDP5_SMP_ALLOC_W_REG(idx), val); + mdp5_write(mdp5_kms, REG_MDP5_SMP_ALLOC_R_REG(idx), val); } } diff --git a/drivers/gpu/drm/msm/msm_drv.c b/drivers/gpu/drm/msm/msm_drv.c index a02dc2b27739..26f859ec24b3 100644 --- a/drivers/gpu/drm/msm/msm_drv.c +++ b/drivers/gpu/drm/msm/msm_drv.c @@ -21,6 +21,16 @@ #include "msm_gpu.h" #include "msm_kms.h" + +/* + * MSM driver version: + * - 1.0.0 - initial interface + * - 1.1.0 - adds madvise, and support for submits with > 4 cmd buffers + */ +#define MSM_VERSION_MAJOR 1 +#define MSM_VERSION_MINOR 1 +#define MSM_VERSION_PATCHLEVEL 0 + static void msm_fb_output_poll_changed(struct drm_device *dev) { struct msm_drm_private *priv = dev->dev_private; @@ -195,6 +205,8 @@ static int msm_drm_uninit(struct device *dev) kfree(vbl_ev); } + msm_gem_shrinker_cleanup(ddev); + drm_kms_helper_poll_fini(ddev); drm_dev_unregister(ddev); @@ -215,10 +227,8 @@ static int msm_drm_uninit(struct device *dev) flush_workqueue(priv->atomic_wq); destroy_workqueue(priv->atomic_wq); - if (kms) { - pm_runtime_disable(dev); + if (kms) kms->funcs->destroy(kms); - } if (gpu) { mutex_lock(&ddev->struct_mutex); @@ -237,6 +247,8 @@ static int msm_drm_uninit(struct device *dev) component_unbind_all(dev, ddev); + msm_mdss_destroy(ddev); + ddev->dev_private = NULL; drm_dev_unref(ddev); @@ -282,6 +294,7 @@ static int msm_init_vram(struct drm_device *dev) if (node) { struct resource r; ret = of_address_to_resource(node, 0, &r); + of_node_put(node); if (ret) return ret; size = r.end - r.start; @@ -350,6 +363,14 @@ static int msm_drm_init(struct device *dev, struct drm_driver *drv) } ddev->dev_private = priv; + priv->dev = ddev; + + ret = msm_mdss_init(ddev); + if (ret) { + kfree(priv); + drm_dev_unref(ddev); + return ret; + } priv->wq = alloc_ordered_workqueue("msm", 0); priv->atomic_wq = alloc_ordered_workqueue("msm:atomic", 0); @@ -365,6 +386,7 @@ static int msm_drm_init(struct device *dev, struct drm_driver *drv) /* Bind all our sub-components: */ ret = component_bind_all(dev, ddev); if (ret) { + msm_mdss_destroy(ddev); kfree(priv); drm_dev_unref(ddev); return ret; @@ -374,9 +396,12 @@ static int msm_drm_init(struct device *dev, struct drm_driver *drv) if (ret) goto fail; + msm_gem_shrinker_init(ddev); + switch (get_mdp_ver(pdev)) { case 4: kms = mdp4_kms_init(ddev); + priv->kms = kms; break; case 5: kms = mdp5_kms_init(ddev); @@ -398,10 +423,7 @@ static int msm_drm_init(struct device *dev, struct drm_driver *drv) goto fail; } - priv->kms = kms; - if (kms) { - pm_runtime_enable(dev); ret = kms->funcs->hw_init(kms); if (ret) { dev_err(dev, "kms hw init failed: %d\n", ret); @@ -417,12 +439,14 @@ static int msm_drm_init(struct device *dev, struct drm_driver *drv) goto fail; } - pm_runtime_get_sync(dev); - ret = drm_irq_install(ddev, platform_get_irq(pdev, 0)); - pm_runtime_put_sync(dev); - if (ret < 0) { - dev_err(dev, "failed to install IRQ handler\n"); - goto fail; + if (kms) { + pm_runtime_get_sync(dev); + ret = drm_irq_install(ddev, kms->irq); + pm_runtime_put_sync(dev); + if (ret < 0) { + dev_err(dev, "failed to install IRQ handler\n"); + goto fail; + } } ret = drm_dev_register(ddev, 0); @@ -682,6 +706,44 @@ static int msm_ioctl_wait_fence(struct drm_device *dev, void *data, return msm_wait_fence(priv->gpu->fctx, args->fence, &timeout, true); } +static int msm_ioctl_gem_madvise(struct drm_device *dev, void *data, + struct drm_file *file) +{ + struct drm_msm_gem_madvise *args = data; + struct drm_gem_object *obj; + int ret; + + switch (args->madv) { + case MSM_MADV_DONTNEED: + case MSM_MADV_WILLNEED: + break; + default: + return -EINVAL; + } + + ret = mutex_lock_interruptible(&dev->struct_mutex); + if (ret) + return ret; + + obj = drm_gem_object_lookup(file, args->handle); + if (!obj) { + ret = -ENOENT; + goto unlock; + } + + ret = msm_gem_madvise(obj, args->madv); + if (ret >= 0) { + args->retained = ret; + ret = 0; + } + + drm_gem_object_unreference(obj); + +unlock: + mutex_unlock(&dev->struct_mutex); + return ret; +} + static const struct drm_ioctl_desc msm_ioctls[] = { DRM_IOCTL_DEF_DRV(MSM_GET_PARAM, msm_ioctl_get_param, DRM_AUTH|DRM_RENDER_ALLOW), DRM_IOCTL_DEF_DRV(MSM_GEM_NEW, msm_ioctl_gem_new, DRM_AUTH|DRM_RENDER_ALLOW), @@ -690,6 +752,7 @@ static const struct drm_ioctl_desc msm_ioctls[] = { DRM_IOCTL_DEF_DRV(MSM_GEM_CPU_FINI, msm_ioctl_gem_cpu_fini, DRM_AUTH|DRM_RENDER_ALLOW), DRM_IOCTL_DEF_DRV(MSM_GEM_SUBMIT, msm_ioctl_gem_submit, DRM_AUTH|DRM_RENDER_ALLOW), DRM_IOCTL_DEF_DRV(MSM_WAIT_FENCE, msm_ioctl_wait_fence, DRM_AUTH|DRM_RENDER_ALLOW), + DRM_IOCTL_DEF_DRV(MSM_GEM_MADVISE, msm_ioctl_gem_madvise, DRM_AUTH|DRM_RENDER_ALLOW), }; static const struct vm_operations_struct vm_ops = { @@ -755,8 +818,9 @@ static struct drm_driver msm_driver = { .name = "msm", .desc = "MSM Snapdragon DRM", .date = "20130625", - .major = 1, - .minor = 0, + .major = MSM_VERSION_MAJOR, + .minor = MSM_VERSION_MINOR, + .patchlevel = MSM_VERSION_PATCHLEVEL, }; #ifdef CONFIG_PM_SLEEP @@ -796,22 +860,146 @@ static int compare_of(struct device *dev, void *data) return dev->of_node == data; } -static int add_components(struct device *dev, struct component_match **matchptr, - const char *name) +/* + * Identify what components need to be added by parsing what remote-endpoints + * our MDP output ports are connected to. In the case of LVDS on MDP4, there + * is no external component that we need to add since LVDS is within MDP4 + * itself. + */ +static int add_components_mdp(struct device *mdp_dev, + struct component_match **matchptr) +{ + struct device_node *np = mdp_dev->of_node; + struct device_node *ep_node; + struct device *master_dev; + + /* + * on MDP4 based platforms, the MDP platform device is the component + * master that adds other display interface components to itself. + * + * on MDP5 based platforms, the MDSS platform device is the component + * master that adds MDP5 and other display interface components to + * itself. + */ + if (of_device_is_compatible(np, "qcom,mdp4")) + master_dev = mdp_dev; + else + master_dev = mdp_dev->parent; + + for_each_endpoint_of_node(np, ep_node) { + struct device_node *intf; + struct of_endpoint ep; + int ret; + + ret = of_graph_parse_endpoint(ep_node, &ep); + if (ret) { + dev_err(mdp_dev, "unable to parse port endpoint\n"); + of_node_put(ep_node); + return ret; + } + + /* + * The LCDC/LVDS port on MDP4 is a speacial case where the + * remote-endpoint isn't a component that we need to add + */ + if (of_device_is_compatible(np, "qcom,mdp4") && + ep.port == 0) { + of_node_put(ep_node); + continue; + } + + /* + * It's okay if some of the ports don't have a remote endpoint + * specified. It just means that the port isn't connected to + * any external interface. + */ + intf = of_graph_get_remote_port_parent(ep_node); + if (!intf) { + of_node_put(ep_node); + continue; + } + + component_match_add(master_dev, matchptr, compare_of, intf); + + of_node_put(intf); + of_node_put(ep_node); + } + + return 0; +} + +static int compare_name_mdp(struct device *dev, void *data) { - struct device_node *np = dev->of_node; - unsigned i; + return (strstr(dev_name(dev), "mdp") != NULL); +} + +static int add_display_components(struct device *dev, + struct component_match **matchptr) +{ + struct device *mdp_dev; + int ret; + + /* + * MDP5 based devices don't have a flat hierarchy. There is a top level + * parent: MDSS, and children: MDP5, DSI, HDMI, eDP etc. Populate the + * children devices, find the MDP5 node, and then add the interfaces + * to our components list. + */ + if (of_device_is_compatible(dev->of_node, "qcom,mdss")) { + ret = of_platform_populate(dev->of_node, NULL, NULL, dev); + if (ret) { + dev_err(dev, "failed to populate children devices\n"); + return ret; + } - for (i = 0; ; i++) { - struct device_node *node; + mdp_dev = device_find_child(dev, NULL, compare_name_mdp); + if (!mdp_dev) { + dev_err(dev, "failed to find MDSS MDP node\n"); + of_platform_depopulate(dev); + return -ENODEV; + } - node = of_parse_phandle(np, name, i); - if (!node) - break; + put_device(mdp_dev); - component_match_add(dev, matchptr, compare_of, node); + /* add the MDP component itself */ + component_match_add(dev, matchptr, compare_of, + mdp_dev->of_node); + } else { + /* MDP4 */ + mdp_dev = dev; } + ret = add_components_mdp(mdp_dev, matchptr); + if (ret) + of_platform_depopulate(dev); + + return ret; +} + +/* + * We don't know what's the best binding to link the gpu with the drm device. + * Fow now, we just hunt for all the possible gpus that we support, and add them + * as components. + */ +static const struct of_device_id msm_gpu_match[] = { + { .compatible = "qcom,adreno-3xx" }, + { .compatible = "qcom,kgsl-3d0" }, + { }, +}; + +static int add_gpu_components(struct device *dev, + struct component_match **matchptr) +{ + struct device_node *np; + + np = of_find_matching_node(NULL, msm_gpu_match); + if (!np) + return 0; + + component_match_add(dev, matchptr, compare_of, np); + + of_node_put(np); + return 0; } @@ -837,9 +1025,15 @@ static const struct component_master_ops msm_drm_ops = { static int msm_pdev_probe(struct platform_device *pdev) { struct component_match *match = NULL; + int ret; - add_components(&pdev->dev, &match, "connectors"); - add_components(&pdev->dev, &match, "gpus"); + ret = add_display_components(&pdev->dev, &match); + if (ret) + return ret; + + ret = add_gpu_components(&pdev->dev, &match); + if (ret) + return ret; pdev->dev.coherent_dma_mask = DMA_BIT_MASK(32); return component_master_add_with_match(&pdev->dev, &msm_drm_ops, match); @@ -848,20 +1042,14 @@ static int msm_pdev_probe(struct platform_device *pdev) static int msm_pdev_remove(struct platform_device *pdev) { component_master_del(&pdev->dev, &msm_drm_ops); + of_platform_depopulate(&pdev->dev); return 0; } -static const struct platform_device_id msm_id[] = { - { "mdp", 0 }, - { } -}; - static const struct of_device_id dt_match[] = { - { .compatible = "qcom,mdp4", .data = (void *) 4 }, /* mdp4 */ - { .compatible = "qcom,mdp5", .data = (void *) 5 }, /* mdp5 */ - /* to support downstream DT files */ - { .compatible = "qcom,mdss_mdp", .data = (void *) 5 }, /* mdp5 */ + { .compatible = "qcom,mdp4", .data = (void *)4 }, /* MDP4 */ + { .compatible = "qcom,mdss", .data = (void *)5 }, /* MDP5 MDSS */ {} }; MODULE_DEVICE_TABLE(of, dt_match); @@ -874,12 +1062,12 @@ static struct platform_driver msm_platform_driver = { .of_match_table = dt_match, .pm = &msm_pm_ops, }, - .id_table = msm_id, }; static int __init msm_drm_register(void) { DBG("init"); + msm_mdp_register(); msm_dsi_register(); msm_edp_register(); msm_hdmi_register(); @@ -895,6 +1083,7 @@ static void __exit msm_drm_unregister(void) adreno_unregister(); msm_edp_unregister(); msm_dsi_unregister(); + msm_mdp_unregister(); } module_init(msm_drm_register); diff --git a/drivers/gpu/drm/msm/msm_drv.h b/drivers/gpu/drm/msm/msm_drv.h index 5b2963f32291..b4bc7f1ef717 100644 --- a/drivers/gpu/drm/msm/msm_drv.h +++ b/drivers/gpu/drm/msm/msm_drv.h @@ -46,6 +46,7 @@ struct msm_kms; struct msm_gpu; struct msm_mmu; +struct msm_mdss; struct msm_rd_state; struct msm_perf_state; struct msm_gem_submit; @@ -77,11 +78,16 @@ struct msm_vblank_ctrl { struct msm_drm_private { + struct drm_device *dev; + struct msm_kms *kms; /* subordinate devices, if present: */ struct platform_device *gpu_pdev; + /* top level MDSS wrapper device (for MDP5 only) */ + struct msm_mdss *mdss; + /* possibly this should be in the kms component, but it is * shared by both mdp4 and mdp5.. */ @@ -147,6 +153,9 @@ struct msm_drm_private { struct drm_mm mm; } vram; + struct notifier_block vmap_notifier; + struct shrinker shrinker; + struct msm_vblank_ctrl vblank_ctrl; }; @@ -165,6 +174,9 @@ void msm_gem_submit_free(struct msm_gem_submit *submit); int msm_ioctl_gem_submit(struct drm_device *dev, void *data, struct drm_file *file); +void msm_gem_shrinker_init(struct drm_device *dev); +void msm_gem_shrinker_cleanup(struct drm_device *dev); + int msm_gem_mmap_obj(struct drm_gem_object *obj, struct vm_area_struct *vma); int msm_gem_mmap(struct file *filp, struct vm_area_struct *vma); @@ -189,8 +201,13 @@ struct drm_gem_object *msm_gem_prime_import_sg_table(struct drm_device *dev, struct dma_buf_attachment *attach, struct sg_table *sg); int msm_gem_prime_pin(struct drm_gem_object *obj); void msm_gem_prime_unpin(struct drm_gem_object *obj); -void *msm_gem_vaddr_locked(struct drm_gem_object *obj); -void *msm_gem_vaddr(struct drm_gem_object *obj); +void *msm_gem_get_vaddr_locked(struct drm_gem_object *obj); +void *msm_gem_get_vaddr(struct drm_gem_object *obj); +void msm_gem_put_vaddr_locked(struct drm_gem_object *obj); +void msm_gem_put_vaddr(struct drm_gem_object *obj); +int msm_gem_madvise(struct drm_gem_object *obj, unsigned madv); +void msm_gem_purge(struct drm_gem_object *obj); +void msm_gem_vunmap(struct drm_gem_object *obj); int msm_gem_sync_object(struct drm_gem_object *obj, struct msm_fence_context *fctx, bool exclusive); void msm_gem_move_to_active(struct drm_gem_object *obj, @@ -257,6 +274,9 @@ static inline int msm_dsi_modeset_init(struct msm_dsi *msm_dsi, } #endif +void __init msm_mdp_register(void); +void __exit msm_mdp_unregister(void); + #ifdef CONFIG_DEBUG_FS void msm_gem_describe(struct drm_gem_object *obj, struct seq_file *m); void msm_gem_describe_objects(struct list_head *list, struct seq_file *m); diff --git a/drivers/gpu/drm/msm/msm_fb.c b/drivers/gpu/drm/msm/msm_fb.c index 7919c24c6ddd..95cf8fe72ee5 100644 --- a/drivers/gpu/drm/msm/msm_fb.c +++ b/drivers/gpu/drm/msm/msm_fb.c @@ -49,8 +49,8 @@ static void msm_framebuffer_destroy(struct drm_framebuffer *fb) for (i = 0; i < n; i++) { struct drm_gem_object *bo = msm_fb->planes[i]; - if (bo) - drm_gem_object_unreference_unlocked(bo); + + drm_gem_object_unreference_unlocked(bo); } kfree(msm_fb); diff --git a/drivers/gpu/drm/msm/msm_fbdev.c b/drivers/gpu/drm/msm/msm_fbdev.c index 1a061e3e8b9e..ffd4a338ca12 100644 --- a/drivers/gpu/drm/msm/msm_fbdev.c +++ b/drivers/gpu/drm/msm/msm_fbdev.c @@ -158,7 +158,11 @@ static int msm_fbdev_create(struct drm_fb_helper *helper, dev->mode_config.fb_base = paddr; - fbi->screen_base = msm_gem_vaddr_locked(fbdev->bo); + fbi->screen_base = msm_gem_get_vaddr_locked(fbdev->bo); + if (IS_ERR(fbi->screen_base)) { + ret = PTR_ERR(fbi->screen_base); + goto fail_unlock; + } fbi->screen_size = fbdev->bo->size; fbi->fix.smem_start = paddr; fbi->fix.smem_len = fbdev->bo->size; @@ -247,6 +251,7 @@ void msm_fbdev_free(struct drm_device *dev) /* this will free the backing object */ if (fbdev->fb) { + msm_gem_put_vaddr(fbdev->bo); drm_framebuffer_unregister_private(fbdev->fb); drm_framebuffer_remove(fbdev->fb); } diff --git a/drivers/gpu/drm/msm/msm_gem.c b/drivers/gpu/drm/msm/msm_gem.c index 7daf4054dd2b..6cd4af443139 100644 --- a/drivers/gpu/drm/msm/msm_gem.c +++ b/drivers/gpu/drm/msm/msm_gem.c @@ -276,6 +276,26 @@ uint64_t msm_gem_mmap_offset(struct drm_gem_object *obj) return offset; } +static void +put_iova(struct drm_gem_object *obj) +{ + struct drm_device *dev = obj->dev; + struct msm_drm_private *priv = obj->dev->dev_private; + struct msm_gem_object *msm_obj = to_msm_bo(obj); + int id; + + WARN_ON(!mutex_is_locked(&dev->struct_mutex)); + + for (id = 0; id < ARRAY_SIZE(msm_obj->domain); id++) { + struct msm_mmu *mmu = priv->mmus[id]; + if (mmu && msm_obj->domain[id].iova) { + uint32_t offset = msm_obj->domain[id].iova; + mmu->funcs->unmap(mmu, offset, msm_obj->sgt, obj->size); + msm_obj->domain[id].iova = 0; + } + } +} + /* should be called under struct_mutex.. although it can be called * from atomic context without struct_mutex to acquire an extra * iova ref if you know one is already held. @@ -388,7 +408,7 @@ fail: return ret; } -void *msm_gem_vaddr_locked(struct drm_gem_object *obj) +void *msm_gem_get_vaddr_locked(struct drm_gem_object *obj) { struct msm_gem_object *msm_obj = to_msm_bo(obj); WARN_ON(!mutex_is_locked(&obj->dev->struct_mutex)); @@ -398,19 +418,94 @@ void *msm_gem_vaddr_locked(struct drm_gem_object *obj) return ERR_CAST(pages); msm_obj->vaddr = vmap(pages, obj->size >> PAGE_SHIFT, VM_MAP, pgprot_writecombine(PAGE_KERNEL)); + if (msm_obj->vaddr == NULL) + return ERR_PTR(-ENOMEM); } + msm_obj->vmap_count++; return msm_obj->vaddr; } -void *msm_gem_vaddr(struct drm_gem_object *obj) +void *msm_gem_get_vaddr(struct drm_gem_object *obj) { void *ret; mutex_lock(&obj->dev->struct_mutex); - ret = msm_gem_vaddr_locked(obj); + ret = msm_gem_get_vaddr_locked(obj); mutex_unlock(&obj->dev->struct_mutex); return ret; } +void msm_gem_put_vaddr_locked(struct drm_gem_object *obj) +{ + struct msm_gem_object *msm_obj = to_msm_bo(obj); + WARN_ON(!mutex_is_locked(&obj->dev->struct_mutex)); + WARN_ON(msm_obj->vmap_count < 1); + msm_obj->vmap_count--; +} + +void msm_gem_put_vaddr(struct drm_gem_object *obj) +{ + mutex_lock(&obj->dev->struct_mutex); + msm_gem_put_vaddr_locked(obj); + mutex_unlock(&obj->dev->struct_mutex); +} + +/* Update madvise status, returns true if not purged, else + * false or -errno. + */ +int msm_gem_madvise(struct drm_gem_object *obj, unsigned madv) +{ + struct msm_gem_object *msm_obj = to_msm_bo(obj); + + WARN_ON(!mutex_is_locked(&obj->dev->struct_mutex)); + + if (msm_obj->madv != __MSM_MADV_PURGED) + msm_obj->madv = madv; + + return (msm_obj->madv != __MSM_MADV_PURGED); +} + +void msm_gem_purge(struct drm_gem_object *obj) +{ + struct drm_device *dev = obj->dev; + struct msm_gem_object *msm_obj = to_msm_bo(obj); + + WARN_ON(!mutex_is_locked(&dev->struct_mutex)); + WARN_ON(!is_purgeable(msm_obj)); + WARN_ON(obj->import_attach); + + put_iova(obj); + + msm_gem_vunmap(obj); + + put_pages(obj); + + msm_obj->madv = __MSM_MADV_PURGED; + + drm_vma_node_unmap(&obj->vma_node, dev->anon_inode->i_mapping); + drm_gem_free_mmap_offset(obj); + + /* Our goal here is to return as much of the memory as + * is possible back to the system as we are called from OOM. + * To do this we must instruct the shmfs to drop all of its + * backing pages, *now*. + */ + shmem_truncate_range(file_inode(obj->filp), 0, (loff_t)-1); + + invalidate_mapping_pages(file_inode(obj->filp)->i_mapping, + 0, (loff_t)-1); +} + +void msm_gem_vunmap(struct drm_gem_object *obj) +{ + struct msm_gem_object *msm_obj = to_msm_bo(obj); + + if (!msm_obj->vaddr || WARN_ON(!is_vunmapable(msm_obj))) + return; + + vunmap(msm_obj->vaddr); + msm_obj->vaddr = NULL; +} + /* must be called before _move_to_active().. */ int msm_gem_sync_object(struct drm_gem_object *obj, struct msm_fence_context *fctx, bool exclusive) @@ -462,6 +557,7 @@ void msm_gem_move_to_active(struct drm_gem_object *obj, struct msm_gpu *gpu, bool exclusive, struct fence *fence) { struct msm_gem_object *msm_obj = to_msm_bo(obj); + WARN_ON(msm_obj->madv != MSM_MADV_WILLNEED); msm_obj->gpu = gpu; if (exclusive) reservation_object_add_excl_fence(msm_obj->resv, fence); @@ -530,13 +626,27 @@ void msm_gem_describe(struct drm_gem_object *obj, struct seq_file *m) struct reservation_object_list *fobj; struct fence *fence; uint64_t off = drm_vma_node_start(&obj->vma_node); + const char *madv; WARN_ON(!mutex_is_locked(&obj->dev->struct_mutex)); - seq_printf(m, "%08x: %c %2d (%2d) %08llx %p %zu\n", + switch (msm_obj->madv) { + case __MSM_MADV_PURGED: + madv = " purged"; + break; + case MSM_MADV_DONTNEED: + madv = " purgeable"; + break; + case MSM_MADV_WILLNEED: + default: + madv = ""; + break; + } + + seq_printf(m, "%08x: %c %2d (%2d) %08llx %p %zu%s\n", msm_obj->flags, is_active(msm_obj) ? 'A' : 'I', obj->name, obj->refcount.refcount.counter, - off, msm_obj->vaddr, obj->size); + off, msm_obj->vaddr, obj->size, madv); rcu_read_lock(); fobj = rcu_dereference(robj->fence); @@ -576,9 +686,7 @@ void msm_gem_describe_objects(struct list_head *list, struct seq_file *m) void msm_gem_free_object(struct drm_gem_object *obj) { struct drm_device *dev = obj->dev; - struct msm_drm_private *priv = obj->dev->dev_private; struct msm_gem_object *msm_obj = to_msm_bo(obj); - int id; WARN_ON(!mutex_is_locked(&dev->struct_mutex)); @@ -587,13 +695,7 @@ void msm_gem_free_object(struct drm_gem_object *obj) list_del(&msm_obj->mm_list); - for (id = 0; id < ARRAY_SIZE(msm_obj->domain); id++) { - struct msm_mmu *mmu = priv->mmus[id]; - if (mmu && msm_obj->domain[id].iova) { - uint32_t offset = msm_obj->domain[id].iova; - mmu->funcs->unmap(mmu, offset, msm_obj->sgt, obj->size); - } - } + put_iova(obj); if (obj->import_attach) { if (msm_obj->vaddr) @@ -607,7 +709,7 @@ void msm_gem_free_object(struct drm_gem_object *obj) drm_prime_gem_destroy(obj, msm_obj->sgt); } else { - vunmap(msm_obj->vaddr); + msm_gem_vunmap(obj); put_pages(obj); } @@ -686,6 +788,7 @@ static int msm_gem_new_impl(struct drm_device *dev, msm_obj->vram_node = (void *)&msm_obj[1]; msm_obj->flags = flags; + msm_obj->madv = MSM_MADV_WILLNEED; if (resv) { msm_obj->resv = resv; @@ -727,9 +830,7 @@ struct drm_gem_object *msm_gem_new(struct drm_device *dev, return obj; fail: - if (obj) - drm_gem_object_unreference(obj); - + drm_gem_object_unreference(obj); return ERR_PTR(ret); } @@ -772,8 +873,6 @@ struct drm_gem_object *msm_gem_import(struct drm_device *dev, return obj; fail: - if (obj) - drm_gem_object_unreference_unlocked(obj); - + drm_gem_object_unreference_unlocked(obj); return ERR_PTR(ret); } diff --git a/drivers/gpu/drm/msm/msm_gem.h b/drivers/gpu/drm/msm/msm_gem.h index 9facd4b6ffd9..b2f13cfe945e 100644 --- a/drivers/gpu/drm/msm/msm_gem.h +++ b/drivers/gpu/drm/msm/msm_gem.h @@ -29,6 +29,16 @@ struct msm_gem_object { uint32_t flags; + /** + * Advice: are the backing pages purgeable? + */ + uint8_t madv; + + /** + * count of active vmap'ing + */ + uint8_t vmap_count; + /* And object is either: * inactive - on priv->inactive_list * active - on one one of the gpu's active_list.. well, at @@ -72,7 +82,16 @@ static inline bool is_active(struct msm_gem_object *msm_obj) return msm_obj->gpu != NULL; } -#define MAX_CMDS 4 +static inline bool is_purgeable(struct msm_gem_object *msm_obj) +{ + return (msm_obj->madv == MSM_MADV_DONTNEED) && msm_obj->sgt && + !msm_obj->base.dma_buf && !msm_obj->base.import_attach; +} + +static inline bool is_vunmapable(struct msm_gem_object *msm_obj) +{ + return (msm_obj->vmap_count == 0) && msm_obj->vaddr; +} /* Created per submit-ioctl, to track bo's and cmdstream bufs, etc, * associated with the cmdstream submission for synchronization (and @@ -95,7 +114,7 @@ struct msm_gem_submit { uint32_t size; /* in dwords */ uint32_t iova; uint32_t idx; /* cmdstream buffer idx in bos[] */ - } cmd[MAX_CMDS]; + } *cmd; /* array of size nr_cmds */ struct { uint32_t flags; struct msm_gem_object *obj; diff --git a/drivers/gpu/drm/msm/msm_gem_prime.c b/drivers/gpu/drm/msm/msm_gem_prime.c index 6b90890faffe..60bb290700ce 100644 --- a/drivers/gpu/drm/msm/msm_gem_prime.c +++ b/drivers/gpu/drm/msm/msm_gem_prime.c @@ -33,12 +33,12 @@ struct sg_table *msm_gem_prime_get_sg_table(struct drm_gem_object *obj) void *msm_gem_prime_vmap(struct drm_gem_object *obj) { - return msm_gem_vaddr(obj); + return msm_gem_get_vaddr(obj); } void msm_gem_prime_vunmap(struct drm_gem_object *obj, void *vaddr) { - /* TODO msm_gem_vunmap() */ + msm_gem_put_vaddr(obj); } int msm_gem_prime_mmap(struct drm_gem_object *obj, struct vm_area_struct *vma) diff --git a/drivers/gpu/drm/msm/msm_gem_shrinker.c b/drivers/gpu/drm/msm/msm_gem_shrinker.c new file mode 100644 index 000000000000..283d2841ba58 --- /dev/null +++ b/drivers/gpu/drm/msm/msm_gem_shrinker.c @@ -0,0 +1,168 @@ +/* + * Copyright (C) 2016 Red Hat + * Author: Rob Clark <robdclark@gmail.com> + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License version 2 as published by + * the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program. If not, see <http://www.gnu.org/licenses/>. + */ + +#include "msm_drv.h" +#include "msm_gem.h" + +static bool mutex_is_locked_by(struct mutex *mutex, struct task_struct *task) +{ + if (!mutex_is_locked(mutex)) + return false; + +#if defined(CONFIG_SMP) || defined(CONFIG_DEBUG_MUTEXES) + return mutex->owner == task; +#else + /* Since UP may be pre-empted, we cannot assume that we own the lock */ + return false; +#endif +} + +static bool msm_gem_shrinker_lock(struct drm_device *dev, bool *unlock) +{ + if (!mutex_trylock(&dev->struct_mutex)) { + if (!mutex_is_locked_by(&dev->struct_mutex, current)) + return false; + *unlock = false; + } else { + *unlock = true; + } + + return true; +} + + +static unsigned long +msm_gem_shrinker_count(struct shrinker *shrinker, struct shrink_control *sc) +{ + struct msm_drm_private *priv = + container_of(shrinker, struct msm_drm_private, shrinker); + struct drm_device *dev = priv->dev; + struct msm_gem_object *msm_obj; + unsigned long count = 0; + bool unlock; + + if (!msm_gem_shrinker_lock(dev, &unlock)) + return 0; + + list_for_each_entry(msm_obj, &priv->inactive_list, mm_list) { + if (is_purgeable(msm_obj)) + count += msm_obj->base.size >> PAGE_SHIFT; + } + + if (unlock) + mutex_unlock(&dev->struct_mutex); + + return count; +} + +static unsigned long +msm_gem_shrinker_scan(struct shrinker *shrinker, struct shrink_control *sc) +{ + struct msm_drm_private *priv = + container_of(shrinker, struct msm_drm_private, shrinker); + struct drm_device *dev = priv->dev; + struct msm_gem_object *msm_obj; + unsigned long freed = 0; + bool unlock; + + if (!msm_gem_shrinker_lock(dev, &unlock)) + return SHRINK_STOP; + + list_for_each_entry(msm_obj, &priv->inactive_list, mm_list) { + if (freed >= sc->nr_to_scan) + break; + if (is_purgeable(msm_obj)) { + msm_gem_purge(&msm_obj->base); + freed += msm_obj->base.size >> PAGE_SHIFT; + } + } + + if (unlock) + mutex_unlock(&dev->struct_mutex); + + if (freed > 0) + pr_info_ratelimited("Purging %lu bytes\n", freed << PAGE_SHIFT); + + return freed; +} + +static int +msm_gem_shrinker_vmap(struct notifier_block *nb, unsigned long event, void *ptr) +{ + struct msm_drm_private *priv = + container_of(nb, struct msm_drm_private, vmap_notifier); + struct drm_device *dev = priv->dev; + struct msm_gem_object *msm_obj; + unsigned unmapped = 0; + bool unlock; + + if (!msm_gem_shrinker_lock(dev, &unlock)) + return NOTIFY_DONE; + + list_for_each_entry(msm_obj, &priv->inactive_list, mm_list) { + if (is_vunmapable(msm_obj)) { + msm_gem_vunmap(&msm_obj->base); + /* since we don't know any better, lets bail after a few + * and if necessary the shrinker will be invoked again. + * Seems better than unmapping *everything* + */ + if (++unmapped >= 15) + break; + } + } + + if (unlock) + mutex_unlock(&dev->struct_mutex); + + *(unsigned long *)ptr += unmapped; + + if (unmapped > 0) + pr_info_ratelimited("Purging %u vmaps\n", unmapped); + + return NOTIFY_DONE; +} + +/** + * msm_gem_shrinker_init - Initialize msm shrinker + * @dev_priv: msm device + * + * This function registers and sets up the msm shrinker. + */ +void msm_gem_shrinker_init(struct drm_device *dev) +{ + struct msm_drm_private *priv = dev->dev_private; + priv->shrinker.count_objects = msm_gem_shrinker_count; + priv->shrinker.scan_objects = msm_gem_shrinker_scan; + priv->shrinker.seeks = DEFAULT_SEEKS; + WARN_ON(register_shrinker(&priv->shrinker)); + + priv->vmap_notifier.notifier_call = msm_gem_shrinker_vmap; + WARN_ON(register_vmap_purge_notifier(&priv->vmap_notifier)); +} + +/** + * msm_gem_shrinker_cleanup - Clean up msm shrinker + * @dev_priv: msm device + * + * This function unregisters the msm shrinker. + */ +void msm_gem_shrinker_cleanup(struct drm_device *dev) +{ + struct msm_drm_private *priv = dev->dev_private; + WARN_ON(unregister_vmap_purge_notifier(&priv->vmap_notifier)); + unregister_shrinker(&priv->shrinker); +} diff --git a/drivers/gpu/drm/msm/msm_gem_submit.c b/drivers/gpu/drm/msm/msm_gem_submit.c index b89ca5174863..9766f9ae4b7d 100644 --- a/drivers/gpu/drm/msm/msm_gem_submit.c +++ b/drivers/gpu/drm/msm/msm_gem_submit.c @@ -29,10 +29,11 @@ #define BO_PINNED 0x2000 static struct msm_gem_submit *submit_create(struct drm_device *dev, - struct msm_gpu *gpu, int nr) + struct msm_gpu *gpu, int nr_bos, int nr_cmds) { struct msm_gem_submit *submit; - int sz = sizeof(*submit) + (nr * sizeof(submit->bos[0])); + int sz = sizeof(*submit) + (nr_bos * sizeof(submit->bos[0])) + + (nr_cmds * sizeof(*submit->cmd)); submit = kmalloc(sz, GFP_TEMPORARY | __GFP_NOWARN | __GFP_NORETRY); if (!submit) @@ -40,12 +41,15 @@ static struct msm_gem_submit *submit_create(struct drm_device *dev, submit->dev = dev; submit->gpu = gpu; + submit->fence = NULL; submit->pid = get_pid(task_pid(current)); + submit->cmd = (void *)&submit->bos[nr_bos]; /* initially, until copy_from_user() and bo lookup succeeds: */ submit->nr_bos = 0; submit->nr_cmds = 0; + INIT_LIST_HEAD(&submit->node); INIT_LIST_HEAD(&submit->bo_list); ww_acquire_init(&submit->ticket, &reservation_ww_class); @@ -75,6 +79,11 @@ static int submit_lookup_objects(struct msm_gem_submit *submit, void __user *userptr = u64_to_user_ptr(args->bos + (i * sizeof(submit_bo))); + /* make sure we don't have garbage flags, in case we hit + * error path before flags is initialized: + */ + submit->bos[i].flags = 0; + ret = copy_from_user(&submit_bo, userptr, sizeof(submit_bo)); if (ret) { ret = -EFAULT; @@ -272,7 +281,7 @@ static int submit_reloc(struct msm_gem_submit *submit, struct msm_gem_object *ob /* For now, just map the entire thing. Eventually we probably * to do it page-by-page, w/ kmap() if not vmap()d.. */ - ptr = msm_gem_vaddr_locked(&obj->base); + ptr = msm_gem_get_vaddr_locked(&obj->base); if (IS_ERR(ptr)) { ret = PTR_ERR(ptr); @@ -325,6 +334,8 @@ static int submit_reloc(struct msm_gem_submit *submit, struct msm_gem_object *ob last_offset = off; } + msm_gem_put_vaddr_locked(&obj->base); + return 0; } @@ -362,14 +373,15 @@ int msm_ioctl_gem_submit(struct drm_device *dev, void *data, if (args->pipe != MSM_PIPE_3D0) return -EINVAL; - if (args->nr_cmds > MAX_CMDS) - return -EINVAL; - - submit = submit_create(dev, gpu, args->nr_bos); - if (!submit) - return -ENOMEM; + ret = mutex_lock_interruptible(&dev->struct_mutex); + if (ret) + return ret; - mutex_lock(&dev->struct_mutex); + submit = submit_create(dev, gpu, args->nr_bos, args->nr_cmds); + if (!submit) { + ret = -ENOMEM; + goto out_unlock; + } ret = submit_lookup_objects(submit, args, file); if (ret) @@ -455,6 +467,7 @@ out: submit_cleanup(submit); if (ret) msm_gem_submit_free(submit); +out_unlock: mutex_unlock(&dev->struct_mutex); return ret; } diff --git a/drivers/gpu/drm/msm/msm_iommu.c b/drivers/gpu/drm/msm/msm_iommu.c index a7a0b6d9b057..3a294d0da3a0 100644 --- a/drivers/gpu/drm/msm/msm_iommu.c +++ b/drivers/gpu/drm/msm/msm_iommu.c @@ -59,10 +59,10 @@ static int msm_iommu_map(struct msm_mmu *mmu, uint32_t iova, return -EINVAL; for_each_sg(sgt->sgl, sg, sgt->nents, i) { - u32 pa = sg_phys(sg) - sg->offset; + dma_addr_t pa = sg_phys(sg) - sg->offset; size_t bytes = sg->length + sg->offset; - VERB("map[%d]: %08x %08x(%zx)", i, iova, pa, bytes); + VERB("map[%d]: %08x %08lx(%zx)", i, da, (unsigned long)pa, bytes); ret = iommu_map(domain, da, pa, bytes, prot); if (ret) @@ -101,7 +101,7 @@ static int msm_iommu_unmap(struct msm_mmu *mmu, uint32_t iova, if (unmapped < bytes) return unmapped; - VERB("unmap[%d]: %08x(%zx)", i, iova, bytes); + VERB("unmap[%d]: %08x(%zx)", i, da, bytes); BUG_ON(!PAGE_ALIGNED(bytes)); diff --git a/drivers/gpu/drm/msm/msm_kms.h b/drivers/gpu/drm/msm/msm_kms.h index e32222c3d44f..40e41e5cdbc6 100644 --- a/drivers/gpu/drm/msm/msm_kms.h +++ b/drivers/gpu/drm/msm/msm_kms.h @@ -61,10 +61,8 @@ struct msm_kms_funcs { struct msm_kms { const struct msm_kms_funcs *funcs; - /* irq handling: */ - bool in_irq; - struct list_head irq_list; /* list of mdp4_irq */ - uint32_t vblank_mask; /* irq bits set for userspace vblank */ + /* irq number to be passed on to drm_irq_install */ + int irq; }; static inline void msm_kms_init(struct msm_kms *kms, @@ -75,5 +73,7 @@ static inline void msm_kms_init(struct msm_kms *kms, struct msm_kms *mdp4_kms_init(struct drm_device *dev); struct msm_kms *mdp5_kms_init(struct drm_device *dev); +int msm_mdss_init(struct drm_device *dev); +void msm_mdss_destroy(struct drm_device *dev); #endif /* __MSM_KMS_H__ */ diff --git a/drivers/gpu/drm/msm/msm_perf.c b/drivers/gpu/drm/msm/msm_perf.c index 830857c47c86..17fe4e53e0d1 100644 --- a/drivers/gpu/drm/msm/msm_perf.c +++ b/drivers/gpu/drm/msm/msm_perf.c @@ -132,7 +132,7 @@ static ssize_t perf_read(struct file *file, char __user *buf, size_t sz, loff_t *ppos) { struct msm_perf_state *perf = file->private_data; - int n = 0, ret; + int n = 0, ret = 0; mutex_lock(&perf->read_lock); @@ -143,9 +143,10 @@ static ssize_t perf_read(struct file *file, char __user *buf, } n = min((int)sz, perf->buftot - perf->bufpos); - ret = copy_to_user(buf, &perf->buf[perf->bufpos], n); - if (ret) + if (copy_to_user(buf, &perf->buf[perf->bufpos], n)) { + ret = -EFAULT; goto out; + } perf->bufpos += n; *ppos += n; diff --git a/drivers/gpu/drm/msm/msm_rd.c b/drivers/gpu/drm/msm/msm_rd.c index b48f73ac6389..3a5fdfcd67ae 100644 --- a/drivers/gpu/drm/msm/msm_rd.c +++ b/drivers/gpu/drm/msm/msm_rd.c @@ -27,6 +27,11 @@ * This bypasses drm_debugfs_create_files() mainly because we need to use * our own fops for a bit more control. In particular, we don't want to * do anything if userspace doesn't have the debugfs file open. + * + * The module-param "rd_full", which defaults to false, enables snapshotting + * all (non-written) buffers in the submit, rather than just cmdstream bo's. + * This is useful to capture the contents of (for example) vbo's or textures, + * or shader programs (if not emitted inline in cmdstream). */ #ifdef CONFIG_DEBUG_FS @@ -40,6 +45,10 @@ #include "msm_gpu.h" #include "msm_gem.h" +static bool rd_full = false; +MODULE_PARM_DESC(rd_full, "If true, $debugfs/.../rd will snapshot all buffer contents"); +module_param_named(rd_full, rd_full, bool, 0600); + enum rd_sect_type { RD_NONE, RD_TEST, /* ascii text */ @@ -140,9 +149,10 @@ static ssize_t rd_read(struct file *file, char __user *buf, goto out; n = min_t(int, sz, circ_count_to_end(&rd->fifo)); - ret = copy_to_user(buf, fptr, n); - if (ret) + if (copy_to_user(buf, fptr, n)) { + ret = -EFAULT; goto out; + } fifo->tail = (fifo->tail + n) & (BUF_SZ - 1); *ppos += n; @@ -277,6 +287,31 @@ void msm_rd_debugfs_cleanup(struct drm_minor *minor) kfree(rd); } +static void snapshot_buf(struct msm_rd_state *rd, + struct msm_gem_submit *submit, int idx, + uint32_t iova, uint32_t size) +{ + struct msm_gem_object *obj = submit->bos[idx].obj; + const char *buf; + + buf = msm_gem_get_vaddr_locked(&obj->base); + if (IS_ERR(buf)) + return; + + if (iova) { + buf += iova - submit->bos[idx].iova; + } else { + iova = submit->bos[idx].iova; + size = obj->base.size; + } + + rd_write_section(rd, RD_GPUADDR, + (uint32_t[2]){ iova, size }, 8); + rd_write_section(rd, RD_BUFFER_CONTENTS, buf, size); + + msm_gem_put_vaddr_locked(&obj->base); +} + /* called under struct_mutex */ void msm_rd_dump_submit(struct msm_gem_submit *submit) { @@ -300,24 +335,27 @@ void msm_rd_dump_submit(struct msm_gem_submit *submit) rd_write_section(rd, RD_CMD, msg, ALIGN(n, 4)); - /* could be nice to have an option (module-param?) to snapshot - * all the bo's associated with the submit. Handy to see vtx - * buffers, etc. For now just the cmdstream bo's is enough. - */ + if (rd_full) { + for (i = 0; i < submit->nr_bos; i++) { + /* buffers that are written to probably don't start out + * with anything interesting: + */ + if (submit->bos[i].flags & MSM_SUBMIT_BO_WRITE) + continue; + + snapshot_buf(rd, submit, i, 0, 0); + } + } for (i = 0; i < submit->nr_cmds; i++) { - uint32_t idx = submit->cmd[i].idx; uint32_t iova = submit->cmd[i].iova; uint32_t szd = submit->cmd[i].size; /* in dwords */ - struct msm_gem_object *obj = submit->bos[idx].obj; - const char *buf = msm_gem_vaddr_locked(&obj->base); - buf += iova - submit->bos[idx].iova; - - rd_write_section(rd, RD_GPUADDR, - (uint32_t[2]){ iova, szd * 4 }, 8); - rd_write_section(rd, RD_BUFFER_CONTENTS, - buf, szd * 4); + /* snapshot cmdstream bo's (if we haven't already): */ + if (!rd_full) { + snapshot_buf(rd, submit, submit->cmd[i].idx, + submit->cmd[i].iova, szd * 4); + } switch (submit->cmd[i].type) { case MSM_SUBMIT_CMD_IB_TARGET_BUF: diff --git a/drivers/gpu/drm/msm/msm_ringbuffer.c b/drivers/gpu/drm/msm/msm_ringbuffer.c index 1f14b908b221..f326cf6a32e6 100644 --- a/drivers/gpu/drm/msm/msm_ringbuffer.c +++ b/drivers/gpu/drm/msm/msm_ringbuffer.c @@ -39,7 +39,11 @@ struct msm_ringbuffer *msm_ringbuffer_new(struct msm_gpu *gpu, int size) goto fail; } - ring->start = msm_gem_vaddr_locked(ring->bo); + ring->start = msm_gem_get_vaddr_locked(ring->bo); + if (IS_ERR(ring->start)) { + ret = PTR_ERR(ring->start); + goto fail; + } ring->end = ring->start + (size / 4); ring->cur = ring->start; @@ -55,7 +59,9 @@ fail: void msm_ringbuffer_destroy(struct msm_ringbuffer *ring) { - if (ring->bo) + if (ring->bo) { + msm_gem_put_vaddr(ring->bo); drm_gem_object_unreference_unlocked(ring->bo); + } kfree(ring); } diff --git a/drivers/gpu/drm/nouveau/Kconfig b/drivers/gpu/drm/nouveau/Kconfig index 5ab13e7939db..2922a82cba8e 100644 --- a/drivers/gpu/drm/nouveau/Kconfig +++ b/drivers/gpu/drm/nouveau/Kconfig @@ -3,13 +3,7 @@ config DRM_NOUVEAU depends on DRM && PCI select FW_LOADER select DRM_KMS_HELPER - select DRM_KMS_FB_HELPER select DRM_TTM - select FB_CFB_FILLRECT - select FB_CFB_COPYAREA - select FB_CFB_IMAGEBLIT - select FB - select FRAMEBUFFER_CONSOLE if !EXPERT select FB_BACKLIGHT if DRM_NOUVEAU_BACKLIGHT select ACPI_VIDEO if ACPI && X86 && BACKLIGHT_CLASS_DEVICE && INPUT select X86_PLATFORM_DEVICES if ACPI && X86 diff --git a/drivers/gpu/drm/nouveau/dispnv04/disp.c b/drivers/gpu/drm/nouveau/dispnv04/disp.c index aea81a547e85..34c0f2f67548 100644 --- a/drivers/gpu/drm/nouveau/dispnv04/disp.c +++ b/drivers/gpu/drm/nouveau/dispnv04/disp.c @@ -125,18 +125,8 @@ nv04_display_destroy(struct drm_device *dev) struct nv04_display *disp = nv04_display(dev); struct nouveau_drm *drm = nouveau_drm(dev); struct nouveau_encoder *encoder; - struct drm_crtc *crtc; struct nouveau_crtc *nv_crtc; - /* Turn every CRTC off. */ - list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) { - struct drm_mode_set modeset = { - .crtc = crtc, - }; - - drm_mode_set_config_internal(&modeset); - } - /* Restore state */ list_for_each_entry(encoder, &dev->mode_config.encoder_list, base.base.head) encoder->enc_restore(&encoder->base.base); diff --git a/drivers/gpu/drm/nouveau/dispnv04/tvnv17.c b/drivers/gpu/drm/nouveau/dispnv04/tvnv17.c index a665b78b2af5..434d1e29f279 100644 --- a/drivers/gpu/drm/nouveau/dispnv04/tvnv17.c +++ b/drivers/gpu/drm/nouveau/dispnv04/tvnv17.c @@ -749,13 +749,8 @@ static int nv17_tv_set_property(struct drm_encoder *encoder, /* Disable the crtc to ensure a full modeset is * performed whenever it's turned on again. */ - if (crtc) { - struct drm_mode_set modeset = { - .crtc = crtc, - }; - - drm_mode_set_config_internal(&modeset); - } + if (crtc) + drm_crtc_force_disable(crtc); } return 0; diff --git a/drivers/gpu/drm/nouveau/include/nvif/cl0080.h b/drivers/gpu/drm/nouveau/include/nvif/cl0080.h index 331620a52afa..287a7d6fa480 100644 --- a/drivers/gpu/drm/nouveau/include/nvif/cl0080.h +++ b/drivers/gpu/drm/nouveau/include/nvif/cl0080.h @@ -29,6 +29,7 @@ struct nv_device_info_v0 { #define NV_DEVICE_INFO_V0_FERMI 0x07 #define NV_DEVICE_INFO_V0_KEPLER 0x08 #define NV_DEVICE_INFO_V0_MAXWELL 0x09 +#define NV_DEVICE_INFO_V0_PASCAL 0x0a __u8 family; __u8 pad06[2]; __u64 ram_size; diff --git a/drivers/gpu/drm/nouveau/include/nvif/class.h b/drivers/gpu/drm/nouveau/include/nvif/class.h index 982aad8fa645..e6e9537537cf 100644 --- a/drivers/gpu/drm/nouveau/include/nvif/class.h +++ b/drivers/gpu/drm/nouveau/include/nvif/class.h @@ -39,6 +39,7 @@ #define KEPLER_CHANNEL_GPFIFO_A /* cla06f.h */ 0x0000a06f #define KEPLER_CHANNEL_GPFIFO_B /* cla06f.h */ 0x0000a16f #define MAXWELL_CHANNEL_GPFIFO_A /* cla06f.h */ 0x0000b06f +#define PASCAL_CHANNEL_GPFIFO_A /* cla06f.h */ 0x0000c06f #define NV50_DISP /* cl5070.h */ 0x00005070 #define G82_DISP /* cl5070.h */ 0x00008270 @@ -50,6 +51,8 @@ #define GK110_DISP /* cl5070.h */ 0x00009270 #define GM107_DISP /* cl5070.h */ 0x00009470 #define GM200_DISP /* cl5070.h */ 0x00009570 +#define GP100_DISP /* cl5070.h */ 0x00009770 +#define GP104_DISP /* cl5070.h */ 0x00009870 #define NV31_MPEG 0x00003174 #define G82_MPEG 0x00008274 @@ -86,6 +89,8 @@ #define GK110_DISP_CORE_CHANNEL_DMA /* cl507d.h */ 0x0000927d #define GM107_DISP_CORE_CHANNEL_DMA /* cl507d.h */ 0x0000947d #define GM200_DISP_CORE_CHANNEL_DMA /* cl507d.h */ 0x0000957d +#define GP100_DISP_CORE_CHANNEL_DMA /* cl507d.h */ 0x0000977d +#define GP104_DISP_CORE_CHANNEL_DMA /* cl507d.h */ 0x0000987d #define NV50_DISP_OVERLAY_CHANNEL_DMA /* cl507e.h */ 0x0000507e #define G82_DISP_OVERLAY_CHANNEL_DMA /* cl507e.h */ 0x0000827e @@ -105,6 +110,8 @@ #define MAXWELL_A /* cl9097.h */ 0x0000b097 #define MAXWELL_B /* cl9097.h */ 0x0000b197 +#define PASCAL_A /* cl9097.h */ 0x0000c097 + #define NV74_BSP 0x000074b0 #define GT212_MSVLD 0x000085b1 @@ -128,6 +135,8 @@ #define FERMI_DMA 0x000090b5 #define KEPLER_DMA_COPY_A 0x0000a0b5 #define MAXWELL_DMA_COPY_A 0x0000b0b5 +#define PASCAL_DMA_COPY_A 0x0000c0b5 +#define PASCAL_DMA_COPY_B 0x0000c1b5 #define FERMI_DECOMPRESS 0x000090b8 @@ -137,6 +146,7 @@ #define KEPLER_COMPUTE_B 0x0000a1c0 #define MAXWELL_COMPUTE_A 0x0000b0c0 #define MAXWELL_COMPUTE_B 0x0000b1c0 +#define PASCAL_COMPUTE_A 0x0000c0c0 #define NV74_CIPHER 0x000074c1 #endif diff --git a/drivers/gpu/drm/nouveau/include/nvkm/core/device.h b/drivers/gpu/drm/nouveau/include/nvkm/core/device.h index c612dc1f1eb4..7ea8aa7ca408 100644 --- a/drivers/gpu/drm/nouveau/include/nvkm/core/device.h +++ b/drivers/gpu/drm/nouveau/include/nvkm/core/device.h @@ -16,9 +16,9 @@ enum nvkm_devidx { NVKM_SUBDEV_MC, NVKM_SUBDEV_BUS, NVKM_SUBDEV_TIMER, + NVKM_SUBDEV_INSTMEM, NVKM_SUBDEV_FB, NVKM_SUBDEV_LTC, - NVKM_SUBDEV_INSTMEM, NVKM_SUBDEV_MMU, NVKM_SUBDEV_BAR, NVKM_SUBDEV_PMU, @@ -33,7 +33,10 @@ enum nvkm_devidx { NVKM_ENGINE_CE0, NVKM_ENGINE_CE1, NVKM_ENGINE_CE2, - NVKM_ENGINE_CE_LAST = NVKM_ENGINE_CE2, + NVKM_ENGINE_CE3, + NVKM_ENGINE_CE4, + NVKM_ENGINE_CE5, + NVKM_ENGINE_CE_LAST = NVKM_ENGINE_CE5, NVKM_ENGINE_CIPHER, NVKM_ENGINE_DISP, @@ -50,7 +53,8 @@ enum nvkm_devidx { NVKM_ENGINE_NVENC0, NVKM_ENGINE_NVENC1, - NVKM_ENGINE_NVENC_LAST = NVKM_ENGINE_NVENC1, + NVKM_ENGINE_NVENC2, + NVKM_ENGINE_NVENC_LAST = NVKM_ENGINE_NVENC2, NVKM_ENGINE_NVDEC, NVKM_ENGINE_PM, @@ -102,6 +106,7 @@ struct nvkm_device { NV_C0 = 0xc0, NV_E0 = 0xe0, GM100 = 0x110, + GP100 = 0x130, } card_type; u32 chipset; u8 chiprev; @@ -136,7 +141,7 @@ struct nvkm_device { struct nvkm_volt *volt; struct nvkm_engine *bsp; - struct nvkm_engine *ce[3]; + struct nvkm_engine *ce[6]; struct nvkm_engine *cipher; struct nvkm_disp *disp; struct nvkm_dma *dma; @@ -149,7 +154,7 @@ struct nvkm_device { struct nvkm_engine *mspdec; struct nvkm_engine *msppp; struct nvkm_engine *msvld; - struct nvkm_engine *nvenc[2]; + struct nvkm_engine *nvenc[3]; struct nvkm_engine *nvdec; struct nvkm_pm *pm; struct nvkm_engine *sec; @@ -170,7 +175,6 @@ struct nvkm_device_func { void (*fini)(struct nvkm_device *, bool suspend); resource_size_t (*resource_addr)(struct nvkm_device *, unsigned bar); resource_size_t (*resource_size)(struct nvkm_device *, unsigned bar); - bool cpu_coherent; }; struct nvkm_device_quirk { @@ -206,7 +210,7 @@ struct nvkm_device_chip { int (*volt )(struct nvkm_device *, int idx, struct nvkm_volt **); int (*bsp )(struct nvkm_device *, int idx, struct nvkm_engine **); - int (*ce[3] )(struct nvkm_device *, int idx, struct nvkm_engine **); + int (*ce[6] )(struct nvkm_device *, int idx, struct nvkm_engine **); int (*cipher )(struct nvkm_device *, int idx, struct nvkm_engine **); int (*disp )(struct nvkm_device *, int idx, struct nvkm_disp **); int (*dma )(struct nvkm_device *, int idx, struct nvkm_dma **); @@ -219,7 +223,7 @@ struct nvkm_device_chip { int (*mspdec )(struct nvkm_device *, int idx, struct nvkm_engine **); int (*msppp )(struct nvkm_device *, int idx, struct nvkm_engine **); int (*msvld )(struct nvkm_device *, int idx, struct nvkm_engine **); - int (*nvenc[2])(struct nvkm_device *, int idx, struct nvkm_engine **); + int (*nvenc[3])(struct nvkm_device *, int idx, struct nvkm_engine **); int (*nvdec )(struct nvkm_device *, int idx, struct nvkm_engine **); int (*pm )(struct nvkm_device *, int idx, struct nvkm_pm **); int (*sec )(struct nvkm_device *, int idx, struct nvkm_engine **); diff --git a/drivers/gpu/drm/nouveau/include/nvkm/core/tegra.h b/drivers/gpu/drm/nouveau/include/nvkm/core/tegra.h index b5370cb56e3c..e5c9b6268dcc 100644 --- a/drivers/gpu/drm/nouveau/include/nvkm/core/tegra.h +++ b/drivers/gpu/drm/nouveau/include/nvkm/core/tegra.h @@ -28,6 +28,7 @@ struct nvkm_device_tegra { } iommu; int gpu_speedo; + int gpu_speedo_id; }; struct nvkm_device_tegra_func { diff --git a/drivers/gpu/drm/nouveau/include/nvkm/engine/ce.h b/drivers/gpu/drm/nouveau/include/nvkm/engine/ce.h index 594d719ba41e..d3d26a1e215d 100644 --- a/drivers/gpu/drm/nouveau/include/nvkm/engine/ce.h +++ b/drivers/gpu/drm/nouveau/include/nvkm/engine/ce.h @@ -7,4 +7,6 @@ int gf100_ce_new(struct nvkm_device *, int, struct nvkm_engine **); int gk104_ce_new(struct nvkm_device *, int, struct nvkm_engine **); int gm107_ce_new(struct nvkm_device *, int, struct nvkm_engine **); int gm200_ce_new(struct nvkm_device *, int, struct nvkm_engine **); +int gp100_ce_new(struct nvkm_device *, int, struct nvkm_engine **); +int gp104_ce_new(struct nvkm_device *, int, struct nvkm_engine **); #endif diff --git a/drivers/gpu/drm/nouveau/include/nvkm/engine/disp.h b/drivers/gpu/drm/nouveau/include/nvkm/engine/disp.h index d4fdce27b297..e82049667ce4 100644 --- a/drivers/gpu/drm/nouveau/include/nvkm/engine/disp.h +++ b/drivers/gpu/drm/nouveau/include/nvkm/engine/disp.h @@ -32,4 +32,6 @@ int gk104_disp_new(struct nvkm_device *, int, struct nvkm_disp **); int gk110_disp_new(struct nvkm_device *, int, struct nvkm_disp **); int gm107_disp_new(struct nvkm_device *, int, struct nvkm_disp **); int gm200_disp_new(struct nvkm_device *, int, struct nvkm_disp **); +int gp100_disp_new(struct nvkm_device *, int, struct nvkm_disp **); +int gp104_disp_new(struct nvkm_device *, int, struct nvkm_disp **); #endif diff --git a/drivers/gpu/drm/nouveau/include/nvkm/engine/fifo.h b/drivers/gpu/drm/nouveau/include/nvkm/engine/fifo.h index 15ddfcf5e8db..ed92fec5292c 100644 --- a/drivers/gpu/drm/nouveau/include/nvkm/engine/fifo.h +++ b/drivers/gpu/drm/nouveau/include/nvkm/engine/fifo.h @@ -66,4 +66,5 @@ int gk20a_fifo_new(struct nvkm_device *, int, struct nvkm_fifo **); int gm107_fifo_new(struct nvkm_device *, int, struct nvkm_fifo **); int gm200_fifo_new(struct nvkm_device *, int, struct nvkm_fifo **); int gm20b_fifo_new(struct nvkm_device *, int, struct nvkm_fifo **); +int gp100_fifo_new(struct nvkm_device *, int, struct nvkm_fifo **); #endif diff --git a/drivers/gpu/drm/nouveau/include/nvkm/engine/gr.h b/drivers/gpu/drm/nouveau/include/nvkm/engine/gr.h index 6515f5810a26..89cf99307828 100644 --- a/drivers/gpu/drm/nouveau/include/nvkm/engine/gr.h +++ b/drivers/gpu/drm/nouveau/include/nvkm/engine/gr.h @@ -42,4 +42,5 @@ int gk20a_gr_new(struct nvkm_device *, int, struct nvkm_gr **); int gm107_gr_new(struct nvkm_device *, int, struct nvkm_gr **); int gm200_gr_new(struct nvkm_device *, int, struct nvkm_gr **); int gm20b_gr_new(struct nvkm_device *, int, struct nvkm_gr **); +int gp100_gr_new(struct nvkm_device *, int, struct nvkm_gr **); #endif diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/bios.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/bios.h index e39a1fea930b..a72f3290528a 100644 --- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/bios.h +++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/bios.h @@ -7,6 +7,9 @@ struct nvkm_bios { u32 size; u8 *data; + u32 image0_size; + u32 imaged_addr; + u32 bmp_offset; u32 bit_offset; @@ -22,10 +25,9 @@ struct nvkm_bios { u8 nvbios_checksum(const u8 *data, int size); u16 nvbios_findstr(const u8 *data, int size, const char *str, int len); int nvbios_memcmp(struct nvkm_bios *, u32 addr, const char *, u32 len); - -#define nvbios_rd08(b,o) (b)->data[(o)] -#define nvbios_rd16(b,o) get_unaligned_le16(&(b)->data[(o)]) -#define nvbios_rd32(b,o) get_unaligned_le32(&(b)->data[(o)]) +u8 nvbios_rd08(struct nvkm_bios *, u32 addr); +u16 nvbios_rd16(struct nvkm_bios *, u32 addr); +u32 nvbios_rd32(struct nvkm_bios *, u32 addr); int nvkm_bios_new(struct nvkm_device *, int, struct nvkm_bios **); #endif diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/bios/disp.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/bios/disp.h index db10c11f0595..c5a6ebd5a478 100644 --- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/bios/disp.h +++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/bios/disp.h @@ -25,7 +25,8 @@ u16 nvbios_outp_match(struct nvkm_bios *, u16 type, u16 mask, u8 *ver, u8 *hdr, u8 *cnt, u8 *len, struct nvbios_outp *); struct nvbios_ocfg { - u16 match; + u8 proto; + u8 flags; u16 clkcmp[2]; }; @@ -33,7 +34,7 @@ u16 nvbios_ocfg_entry(struct nvkm_bios *, u16 outp, u8 idx, u8 *ver, u8 *hdr, u8 *cnt, u8 *len); u16 nvbios_ocfg_parse(struct nvkm_bios *, u16 outp, u8 idx, u8 *ver, u8 *hdr, u8 *cnt, u8 *len, struct nvbios_ocfg *); -u16 nvbios_ocfg_match(struct nvkm_bios *, u16 outp, u16 type, +u16 nvbios_ocfg_match(struct nvkm_bios *, u16 outp, u8 proto, u8 flags, u8 *ver, u8 *hdr, u8 *cnt, u8 *len, struct nvbios_ocfg *); u16 nvbios_oclk_match(struct nvkm_bios *, u16 cmp, u32 khz); #endif diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/fb.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/fb.h index 0a734fd06acf..3a410275fa71 100644 --- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/fb.h +++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/fb.h @@ -56,6 +56,8 @@ struct nvkm_fb { int regions; } tile; + u8 page; + struct nvkm_memory *mmu_rd; struct nvkm_memory *mmu_wr; }; @@ -91,6 +93,8 @@ int gk104_fb_new(struct nvkm_device *, int, struct nvkm_fb **); int gk20a_fb_new(struct nvkm_device *, int, struct nvkm_fb **); int gm107_fb_new(struct nvkm_device *, int, struct nvkm_fb **); int gm200_fb_new(struct nvkm_device *, int, struct nvkm_fb **); +int gp100_fb_new(struct nvkm_device *, int, struct nvkm_fb **); +int gp104_fb_new(struct nvkm_device *, int, struct nvkm_fb **); #include <subdev/bios.h> #include <subdev/bios/ramcfg.h> diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/ltc.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/ltc.h index c6b90b6543b3..cd755baf9cab 100644 --- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/ltc.h +++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/ltc.h @@ -38,4 +38,5 @@ int gk104_ltc_new(struct nvkm_device *, int, struct nvkm_ltc **); int gk20a_ltc_new(struct nvkm_device *, int, struct nvkm_ltc **); int gm107_ltc_new(struct nvkm_device *, int, struct nvkm_ltc **); int gm200_ltc_new(struct nvkm_device *, int, struct nvkm_ltc **); +int gp100_ltc_new(struct nvkm_device *, int, struct nvkm_ltc **); #endif diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/mc.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/mc.h index 2e80682b2da1..27d25b18d85c 100644 --- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/mc.h +++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/mc.h @@ -7,11 +7,14 @@ struct nvkm_mc { struct nvkm_subdev subdev; }; -void nvkm_mc_intr(struct nvkm_mc *, bool *handled); -void nvkm_mc_intr_unarm(struct nvkm_mc *); -void nvkm_mc_intr_rearm(struct nvkm_mc *); -void nvkm_mc_reset(struct nvkm_mc *, enum nvkm_devidx); -void nvkm_mc_unk260(struct nvkm_mc *, u32 data); +void nvkm_mc_enable(struct nvkm_device *, enum nvkm_devidx); +void nvkm_mc_disable(struct nvkm_device *, enum nvkm_devidx); +void nvkm_mc_reset(struct nvkm_device *, enum nvkm_devidx); +void nvkm_mc_intr(struct nvkm_device *, bool *handled); +void nvkm_mc_intr_unarm(struct nvkm_device *); +void nvkm_mc_intr_rearm(struct nvkm_device *); +void nvkm_mc_intr_mask(struct nvkm_device *, enum nvkm_devidx, bool enable); +void nvkm_mc_unk260(struct nvkm_device *, u32 data); int nv04_mc_new(struct nvkm_device *, int, struct nvkm_mc **); int nv11_mc_new(struct nvkm_device *, int, struct nvkm_mc **); @@ -24,4 +27,5 @@ int gt215_mc_new(struct nvkm_device *, int, struct nvkm_mc **); int gf100_mc_new(struct nvkm_device *, int, struct nvkm_mc **); int gk104_mc_new(struct nvkm_device *, int, struct nvkm_mc **); int gk20a_mc_new(struct nvkm_device *, int, struct nvkm_mc **); +int gp100_mc_new(struct nvkm_device *, int, struct nvkm_mc **); #endif diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/pci.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/pci.h index ddb913889d7e..e6523e2cea9f 100644 --- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/pci.h +++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/pci.h @@ -47,6 +47,7 @@ int g94_pci_new(struct nvkm_device *, int, struct nvkm_pci **); int gf100_pci_new(struct nvkm_device *, int, struct nvkm_pci **); int gf106_pci_new(struct nvkm_device *, int, struct nvkm_pci **); int gk104_pci_new(struct nvkm_device *, int, struct nvkm_pci **); +int gp100_pci_new(struct nvkm_device *, int, struct nvkm_pci **); /* pcie functions */ int nvkm_pcie_set_link(struct nvkm_pci *, enum nvkm_pcie_speed, u8 width); diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/secboot.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/secboot.h index c6edd95a5b69..b04c38c07761 100644 --- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/secboot.h +++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/secboot.h @@ -43,9 +43,8 @@ struct nvkm_secboot { const struct nvkm_secboot_func *func; struct nvkm_subdev subdev; + enum nvkm_devidx devidx; u32 base; - u32 irq_mask; - u32 enable_mask; }; #define nvkm_secboot(p) container_of((p), struct nvkm_secboot, subdev) diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/top.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/top.h index 8fb575a92c48..71ebbfd4484f 100644 --- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/top.h +++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/top.h @@ -8,10 +8,11 @@ struct nvkm_top { struct list_head device; }; -u32 nvkm_top_reset(struct nvkm_top *, enum nvkm_devidx); -u32 nvkm_top_intr(struct nvkm_top *, u32 intr, u64 *subdevs); -enum nvkm_devidx nvkm_top_fault(struct nvkm_top *, int fault); -enum nvkm_devidx nvkm_top_engine(struct nvkm_top *, int, int *runl, int *engn); +u32 nvkm_top_reset(struct nvkm_device *, enum nvkm_devidx); +u32 nvkm_top_intr(struct nvkm_device *, u32 intr, u64 *subdevs); +u32 nvkm_top_intr_mask(struct nvkm_device *, enum nvkm_devidx); +enum nvkm_devidx nvkm_top_fault(struct nvkm_device *, int fault); +enum nvkm_devidx nvkm_top_engine(struct nvkm_device *, int, int *runl, int *engn); int gk104_top_new(struct nvkm_device *, int, struct nvkm_top **); #endif diff --git a/drivers/gpu/drm/nouveau/include/nvkm/subdev/volt.h b/drivers/gpu/drm/nouveau/include/nvkm/subdev/volt.h index feff55cff05b..b765f4ffcde6 100644 --- a/drivers/gpu/drm/nouveau/include/nvkm/subdev/volt.h +++ b/drivers/gpu/drm/nouveau/include/nvkm/subdev/volt.h @@ -12,6 +12,9 @@ struct nvkm_volt { u32 uv; u8 vid; } vid[256]; + + u32 max_uv; + u32 min_uv; }; int nvkm_volt_get(struct nvkm_volt *); diff --git a/drivers/gpu/drm/nouveau/nouveau_abi16.c b/drivers/gpu/drm/nouveau/nouveau_abi16.c index eb7de487a2b3..7bd4683216d0 100644 --- a/drivers/gpu/drm/nouveau/nouveau_abi16.c +++ b/drivers/gpu/drm/nouveau/nouveau_abi16.c @@ -100,6 +100,7 @@ nouveau_abi16_swclass(struct nouveau_drm *drm) case NV_DEVICE_INFO_V0_FERMI: case NV_DEVICE_INFO_V0_KEPLER: case NV_DEVICE_INFO_V0_MAXWELL: + case NV_DEVICE_INFO_V0_PASCAL: return NVIF_CLASS_SW_GF100; } diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c index 5e3f3e826476..528bdeffb339 100644 --- a/drivers/gpu/drm/nouveau/nouveau_bo.c +++ b/drivers/gpu/drm/nouveau/nouveau_bo.c @@ -209,8 +209,7 @@ nouveau_bo_new(struct drm_device *dev, int size, int align, nvbo->tile_flags = tile_flags; nvbo->bo.bdev = &drm->ttm.bdev; - if (!nvxx_device(&drm->device)->func->cpu_coherent) - nvbo->force_coherent = flags & TTM_PL_FLAG_UNCACHED; + nvbo->force_coherent = flags & TTM_PL_FLAG_UNCACHED; nvbo->page_shift = 12; if (drm->client.vm) { @@ -424,13 +423,7 @@ nouveau_bo_map(struct nouveau_bo *nvbo) if (ret) return ret; - /* - * TTM buffers allocated using the DMA API already have a mapping, let's - * use it instead. - */ - if (!nvbo->force_coherent) - ret = ttm_bo_kmap(&nvbo->bo, 0, nvbo->bo.mem.num_pages, - &nvbo->kmap); + ret = ttm_bo_kmap(&nvbo->bo, 0, nvbo->bo.mem.num_pages, &nvbo->kmap); ttm_bo_unreserve(&nvbo->bo); return ret; @@ -442,12 +435,7 @@ nouveau_bo_unmap(struct nouveau_bo *nvbo) if (!nvbo) return; - /* - * TTM buffers allocated using the DMA API already had a coherent - * mapping which we used, no need to unmap. - */ - if (!nvbo->force_coherent) - ttm_bo_kunmap(&nvbo->kmap); + ttm_bo_kunmap(&nvbo->kmap); } void @@ -506,35 +494,13 @@ nouveau_bo_validate(struct nouveau_bo *nvbo, bool interruptible, return 0; } -static inline void * -_nouveau_bo_mem_index(struct nouveau_bo *nvbo, unsigned index, void *mem, u8 sz) -{ - struct ttm_dma_tt *dma_tt; - u8 *m = mem; - - index *= sz; - - if (m) { - /* kmap'd address, return the corresponding offset */ - m += index; - } else { - /* DMA-API mapping, lookup the right address */ - dma_tt = (struct ttm_dma_tt *)nvbo->bo.ttm; - m = dma_tt->cpu_address[index / PAGE_SIZE]; - m += index % PAGE_SIZE; - } - - return m; -} -#define nouveau_bo_mem_index(o, i, m) _nouveau_bo_mem_index(o, i, m, sizeof(*m)) - void nouveau_bo_wr16(struct nouveau_bo *nvbo, unsigned index, u16 val) { bool is_iomem; u16 *mem = ttm_kmap_obj_virtual(&nvbo->kmap, &is_iomem); - mem = nouveau_bo_mem_index(nvbo, index, mem); + mem += index; if (is_iomem) iowrite16_native(val, (void __force __iomem *)mem); @@ -548,7 +514,7 @@ nouveau_bo_rd32(struct nouveau_bo *nvbo, unsigned index) bool is_iomem; u32 *mem = ttm_kmap_obj_virtual(&nvbo->kmap, &is_iomem); - mem = nouveau_bo_mem_index(nvbo, index, mem); + mem += index; if (is_iomem) return ioread32_native((void __force __iomem *)mem); @@ -562,7 +528,7 @@ nouveau_bo_wr32(struct nouveau_bo *nvbo, unsigned index, u32 val) bool is_iomem; u32 *mem = ttm_kmap_obj_virtual(&nvbo->kmap, &is_iomem); - mem = nouveau_bo_mem_index(nvbo, index, mem); + mem += index; if (is_iomem) iowrite32_native(val, (void __force __iomem *)mem); @@ -1082,7 +1048,6 @@ nouveau_bo_move_m2mf(struct ttm_buffer_object *bo, int evict, bool intr, ret = ttm_bo_move_accel_cleanup(bo, &fence->base, evict, - no_wait_gpu, new_mem); nouveau_fence_unref(&fence); } @@ -1104,6 +1069,10 @@ nouveau_bo_move_init(struct nouveau_drm *drm) struct ttm_mem_reg *, struct ttm_mem_reg *); int (*init)(struct nouveau_channel *, u32 handle); } _methods[] = { + { "COPY", 4, 0xc1b5, nve0_bo_move_copy, nve0_bo_move_init }, + { "GRCE", 0, 0xc1b5, nve0_bo_move_copy, nvc0_bo_move_init }, + { "COPY", 4, 0xc0b5, nve0_bo_move_copy, nve0_bo_move_init }, + { "GRCE", 0, 0xc0b5, nve0_bo_move_copy, nvc0_bo_move_init }, { "COPY", 4, 0xb0b5, nve0_bo_move_copy, nve0_bo_move_init }, { "GRCE", 0, 0xb0b5, nve0_bo_move_copy, nvc0_bo_move_init }, { "COPY", 4, 0xa0b5, nve0_bo_move_copy, nve0_bo_move_init }, @@ -1289,6 +1258,10 @@ nouveau_bo_move(struct ttm_buffer_object *bo, bool evict, bool intr, struct nouveau_drm_tile *new_tile = NULL; int ret = 0; + ret = ttm_bo_wait(bo, intr, no_wait_gpu); + if (ret) + return ret; + if (nvbo->pin_refcnt) NV_WARN(drm, "Moving pinned object %p!\n", nvbo); @@ -1324,7 +1297,7 @@ nouveau_bo_move(struct ttm_buffer_object *bo, bool evict, bool intr, /* Fallback to software copy. */ ret = ttm_bo_wait(bo, intr, no_wait_gpu); if (ret == 0) - ret = ttm_bo_move_memcpy(bo, evict, no_wait_gpu, new_mem); + ret = ttm_bo_move_memcpy(bo, evict, intr, no_wait_gpu, new_mem); out: if (drm->device.info.family < NV_DEVICE_INFO_V0_TESLA) { @@ -1488,14 +1461,6 @@ nouveau_ttm_tt_populate(struct ttm_tt *ttm) dev = drm->dev; pdev = device->dev; - /* - * Objects matching this condition have been marked as force_coherent, - * so use the DMA API for them. - */ - if (!nvxx_device(&drm->device)->func->cpu_coherent && - ttm->caching_state == tt_uncached) - return ttm_dma_populate(ttm_dma, dev->dev); - #if IS_ENABLED(CONFIG_AGP) if (drm->agp.bridge) { return ttm_agp_tt_populate(ttm); @@ -1553,16 +1518,6 @@ nouveau_ttm_tt_unpopulate(struct ttm_tt *ttm) dev = drm->dev; pdev = device->dev; - /* - * Objects matching this condition have been marked as force_coherent, - * so use the DMA API for them. - */ - if (!nvxx_device(&drm->device)->func->cpu_coherent && - ttm->caching_state == tt_uncached) { - ttm_dma_unpopulate(ttm_dma, dev->dev); - return; - } - #if IS_ENABLED(CONFIG_AGP) if (drm->agp.bridge) { ttm_agp_tt_unpopulate(ttm); diff --git a/drivers/gpu/drm/nouveau/nouveau_chan.c b/drivers/gpu/drm/nouveau/nouveau_chan.c index b1d2527c5625..f9b3c811187e 100644 --- a/drivers/gpu/drm/nouveau/nouveau_chan.c +++ b/drivers/gpu/drm/nouveau/nouveau_chan.c @@ -191,7 +191,8 @@ static int nouveau_channel_ind(struct nouveau_drm *drm, struct nvif_device *device, u32 engine, struct nouveau_channel **pchan) { - static const u16 oclasses[] = { MAXWELL_CHANNEL_GPFIFO_A, + static const u16 oclasses[] = { PASCAL_CHANNEL_GPFIFO_A, + MAXWELL_CHANNEL_GPFIFO_A, KEPLER_CHANNEL_GPFIFO_B, KEPLER_CHANNEL_GPFIFO_A, FERMI_CHANNEL_GPFIFO, diff --git a/drivers/gpu/drm/nouveau/nouveau_display.c b/drivers/gpu/drm/nouveau/nouveau_display.c index 6072fe292db8..afbf557b23d4 100644 --- a/drivers/gpu/drm/nouveau/nouveau_display.c +++ b/drivers/gpu/drm/nouveau/nouveau_display.c @@ -47,7 +47,7 @@ nouveau_display_vblank_handler(struct nvif_notify *notify) { struct nouveau_crtc *nv_crtc = container_of(notify, typeof(*nv_crtc), vblank); - drm_handle_vblank(nv_crtc->base.dev, nv_crtc->index); + drm_crtc_handle_vblank(&nv_crtc->base); return NVIF_NOTIFY_KEEP; } @@ -495,6 +495,8 @@ nouveau_display_create(struct drm_device *dev) if (nouveau_modeset != 2 && drm->vbios.dcb.entries) { static const u16 oclass[] = { + GP104_DISP, + GP100_DISP, GM200_DISP, GM107_DISP, GK110_DISP, @@ -554,6 +556,7 @@ nouveau_display_destroy(struct drm_device *dev) nouveau_display_vblank_fini(dev); drm_kms_helper_poll_fini(dev); + drm_crtc_force_disable_all(dev); drm_mode_config_cleanup(dev); if (disp->dtor) diff --git a/drivers/gpu/drm/nouveau/nouveau_drm.c b/drivers/gpu/drm/nouveau/nouveau_drm.c index 295e7621cc68..66c1280c0f1f 100644 --- a/drivers/gpu/drm/nouveau/nouveau_drm.c +++ b/drivers/gpu/drm/nouveau/nouveau_drm.c @@ -198,6 +198,7 @@ nouveau_accel_init(struct nouveau_drm *drm) case KEPLER_CHANNEL_GPFIFO_A: case KEPLER_CHANNEL_GPFIFO_B: case MAXWELL_CHANNEL_GPFIFO_A: + case PASCAL_CHANNEL_GPFIFO_A: ret = nvc0_fence_create(drm); break; default: @@ -316,7 +317,16 @@ static int nouveau_drm_probe(struct pci_dev *pdev, if (vga_switcheroo_client_probe_defer(pdev)) return -EPROBE_DEFER; - /* remove conflicting drivers (vesafb, efifb etc) */ + /* We need to check that the chipset is supported before booting + * fbdev off the hardware, as there's no way to put it back. + */ + ret = nvkm_device_pci_new(pdev, NULL, "error", true, false, 0, &device); + if (ret) + return ret; + + nvkm_device_del(&device); + + /* Remove conflicting drivers (vesafb, efifb etc). */ aper = alloc_apertures(3); if (!aper) return -ENOMEM; @@ -430,6 +440,11 @@ nouveau_drm_load(struct drm_device *dev, unsigned long flags) nouveau_vga_init(drm); if (drm->device.info.family >= NV_DEVICE_INFO_V0_TESLA) { + if (!nvxx_device(&drm->device)->mmu) { + ret = -ENOSYS; + goto fail_device; + } + ret = nvkm_vm_new(nvxx_device(&drm->device), 0, (1ULL << 40), 0x1000, NULL, &drm->client.vm); if (ret) @@ -490,7 +505,11 @@ nouveau_drm_unload(struct drm_device *dev) { struct nouveau_drm *drm = nouveau_drm(dev); - pm_runtime_get_sync(dev->dev); + if (nouveau_runtime_pm != 0) { + pm_runtime_get_sync(dev->dev); + pm_runtime_forbid(dev->dev); + } + nouveau_fbcon_fini(dev); nouveau_accel_fini(drm); nouveau_hwmon_fini(dev); diff --git a/drivers/gpu/drm/nouveau/nouveau_fbcon.c b/drivers/gpu/drm/nouveau/nouveau_fbcon.c index 57aaf98a26f9..d1f248fd3506 100644 --- a/drivers/gpu/drm/nouveau/nouveau_fbcon.c +++ b/drivers/gpu/drm/nouveau/nouveau_fbcon.c @@ -552,6 +552,8 @@ nouveau_fbcon_init(struct drm_device *dev) if (ret) goto fini; + if (fbcon->helper.fbdev) + fbcon->helper.fbdev->pixmap.buf_align = 4; return 0; fini: diff --git a/drivers/gpu/drm/nouveau/nouveau_hwmon.c b/drivers/gpu/drm/nouveau/nouveau_hwmon.c index 1ff4166af26e..71f764bf4cc6 100644 --- a/drivers/gpu/drm/nouveau/nouveau_hwmon.c +++ b/drivers/gpu/drm/nouveau/nouveau_hwmon.c @@ -535,6 +535,40 @@ static SENSOR_DEVICE_ATTR(in0_input, S_IRUGO, nouveau_hwmon_get_in0_input, NULL, 0); static ssize_t +nouveau_hwmon_get_in0_min(struct device *d, + struct device_attribute *a, char *buf) +{ + struct drm_device *dev = dev_get_drvdata(d); + struct nouveau_drm *drm = nouveau_drm(dev); + struct nvkm_volt *volt = nvxx_volt(&drm->device); + + if (!volt || !volt->min_uv) + return -ENODEV; + + return sprintf(buf, "%i\n", volt->min_uv / 1000); +} + +static SENSOR_DEVICE_ATTR(in0_min, S_IRUGO, + nouveau_hwmon_get_in0_min, NULL, 0); + +static ssize_t +nouveau_hwmon_get_in0_max(struct device *d, + struct device_attribute *a, char *buf) +{ + struct drm_device *dev = dev_get_drvdata(d); + struct nouveau_drm *drm = nouveau_drm(dev); + struct nvkm_volt *volt = nvxx_volt(&drm->device); + + if (!volt || !volt->max_uv) + return -ENODEV; + + return sprintf(buf, "%i\n", volt->max_uv / 1000); +} + +static SENSOR_DEVICE_ATTR(in0_max, S_IRUGO, + nouveau_hwmon_get_in0_max, NULL, 0); + +static ssize_t nouveau_hwmon_get_in0_label(struct device *d, struct device_attribute *a, char *buf) { @@ -594,6 +628,8 @@ static struct attribute *hwmon_pwm_fan_attributes[] = { static struct attribute *hwmon_in0_attributes[] = { &sensor_dev_attr_in0_input.dev_attr.attr, + &sensor_dev_attr_in0_min.dev_attr.attr, + &sensor_dev_attr_in0_max.dev_attr.attr, &sensor_dev_attr_in0_label.dev_attr.attr, NULL }; diff --git a/drivers/gpu/drm/nouveau/nouveau_ttm.c b/drivers/gpu/drm/nouveau/nouveau_ttm.c index bcee91497eb9..1825dbc33192 100644 --- a/drivers/gpu/drm/nouveau/nouveau_ttm.c +++ b/drivers/gpu/drm/nouveau/nouveau_ttm.c @@ -164,6 +164,7 @@ nouveau_gart_manager_new(struct ttm_mem_type_manager *man, case NV_DEVICE_INFO_V0_FERMI: case NV_DEVICE_INFO_V0_KEPLER: case NV_DEVICE_INFO_V0_MAXWELL: + case NV_DEVICE_INFO_V0_PASCAL: node->memtype = (nvbo->tile_flags & 0xff00) >> 8; break; default: diff --git a/drivers/gpu/drm/nouveau/nv04_fbcon.c b/drivers/gpu/drm/nouveau/nv04_fbcon.c index 0f3e4bb411cc..7d9248b8c664 100644 --- a/drivers/gpu/drm/nouveau/nv04_fbcon.c +++ b/drivers/gpu/drm/nouveau/nv04_fbcon.c @@ -82,7 +82,6 @@ nv04_fbcon_imageblit(struct fb_info *info, const struct fb_image *image) uint32_t fg; uint32_t bg; uint32_t dsize; - uint32_t width; uint32_t *data = (uint32_t *)image->data; int ret; @@ -93,9 +92,6 @@ nv04_fbcon_imageblit(struct fb_info *info, const struct fb_image *image) if (ret) return ret; - width = ALIGN(image->width, 8); - dsize = ALIGN(width * image->height, 32) >> 5; - if (info->fix.visual == FB_VISUAL_TRUECOLOR || info->fix.visual == FB_VISUAL_DIRECTCOLOR) { fg = ((uint32_t *) info->pseudo_palette)[image->fg_color]; @@ -111,10 +107,11 @@ nv04_fbcon_imageblit(struct fb_info *info, const struct fb_image *image) ((image->dx + image->width) & 0xffff)); OUT_RING(chan, bg); OUT_RING(chan, fg); - OUT_RING(chan, (image->height << 16) | width); + OUT_RING(chan, (image->height << 16) | image->width); OUT_RING(chan, (image->height << 16) | image->width); OUT_RING(chan, (image->dy << 16) | (image->dx & 0xffff)); + dsize = ALIGN(image->width * image->height, 32) >> 5; while (dsize) { int iter_len = dsize > 128 ? 128 : dsize; diff --git a/drivers/gpu/drm/nouveau/nv50_display.c b/drivers/gpu/drm/nouveau/nv50_display.c index 7a7788212df7..7d0edcbcfca7 100644 --- a/drivers/gpu/drm/nouveau/nv50_display.c +++ b/drivers/gpu/drm/nouveau/nv50_display.c @@ -297,6 +297,8 @@ nv50_core_create(struct nvif_device *device, struct nvif_object *disp, .pushbuf = 0xb0007d00, }; static const s32 oclass[] = { + GP104_DISP_CORE_CHANNEL_DMA, + GP100_DISP_CORE_CHANNEL_DMA, GM200_DISP_CORE_CHANNEL_DMA, GM107_DISP_CORE_CHANNEL_DMA, GK110_DISP_CORE_CHANNEL_DMA, diff --git a/drivers/gpu/drm/nouveau/nv50_fbcon.c b/drivers/gpu/drm/nouveau/nv50_fbcon.c index 33d9ee0fac40..1aeb698e9707 100644 --- a/drivers/gpu/drm/nouveau/nv50_fbcon.c +++ b/drivers/gpu/drm/nouveau/nv50_fbcon.c @@ -95,7 +95,7 @@ nv50_fbcon_imageblit(struct fb_info *info, const struct fb_image *image) struct nouveau_fbdev *nfbdev = info->par; struct nouveau_drm *drm = nouveau_drm(nfbdev->dev); struct nouveau_channel *chan = drm->channel; - uint32_t width, dwords, *data = (uint32_t *)image->data; + uint32_t dwords, *data = (uint32_t *)image->data; uint32_t mask = ~(~0 >> (32 - info->var.bits_per_pixel)); uint32_t *palette = info->pseudo_palette; int ret; @@ -107,9 +107,6 @@ nv50_fbcon_imageblit(struct fb_info *info, const struct fb_image *image) if (ret) return ret; - width = ALIGN(image->width, 32); - dwords = (width * image->height) >> 5; - BEGIN_NV04(chan, NvSub2D, 0x0814, 2); if (info->fix.visual == FB_VISUAL_TRUECOLOR || info->fix.visual == FB_VISUAL_DIRECTCOLOR) { @@ -128,6 +125,7 @@ nv50_fbcon_imageblit(struct fb_info *info, const struct fb_image *image) OUT_RING(chan, 0); OUT_RING(chan, image->dy); + dwords = ALIGN(image->width * image->height, 32) >> 5; while (dwords) { int push = dwords > 2047 ? 2047 : dwords; diff --git a/drivers/gpu/drm/nouveau/nvc0_fbcon.c b/drivers/gpu/drm/nouveau/nvc0_fbcon.c index a0913359ac05..839f4c8c1805 100644 --- a/drivers/gpu/drm/nouveau/nvc0_fbcon.c +++ b/drivers/gpu/drm/nouveau/nvc0_fbcon.c @@ -95,7 +95,7 @@ nvc0_fbcon_imageblit(struct fb_info *info, const struct fb_image *image) struct nouveau_fbdev *nfbdev = info->par; struct nouveau_drm *drm = nouveau_drm(nfbdev->dev); struct nouveau_channel *chan = drm->channel; - uint32_t width, dwords, *data = (uint32_t *)image->data; + uint32_t dwords, *data = (uint32_t *)image->data; uint32_t mask = ~(~0 >> (32 - info->var.bits_per_pixel)); uint32_t *palette = info->pseudo_palette; int ret; @@ -107,9 +107,6 @@ nvc0_fbcon_imageblit(struct fb_info *info, const struct fb_image *image) if (ret) return ret; - width = ALIGN(image->width, 32); - dwords = (width * image->height) >> 5; - BEGIN_NVC0(chan, NvSub2D, 0x0814, 2); if (info->fix.visual == FB_VISUAL_TRUECOLOR || info->fix.visual == FB_VISUAL_DIRECTCOLOR) { @@ -128,6 +125,7 @@ nvc0_fbcon_imageblit(struct fb_info *info, const struct fb_image *image) OUT_RING (chan, 0); OUT_RING (chan, image->dy); + dwords = ALIGN(image->width * image->height, 32) >> 5; while (dwords) { int push = dwords > 2047 ? 2047 : dwords; diff --git a/drivers/gpu/drm/nouveau/nvkm/core/subdev.c b/drivers/gpu/drm/nouveau/nvkm/core/subdev.c index b18557858f19..19044aba265e 100644 --- a/drivers/gpu/drm/nouveau/nvkm/core/subdev.c +++ b/drivers/gpu/drm/nouveau/nvkm/core/subdev.c @@ -57,6 +57,9 @@ nvkm_subdev_name[NVKM_SUBDEV_NR] = { [NVKM_ENGINE_CE0 ] = "ce0", [NVKM_ENGINE_CE1 ] = "ce1", [NVKM_ENGINE_CE2 ] = "ce2", + [NVKM_ENGINE_CE3 ] = "ce3", + [NVKM_ENGINE_CE4 ] = "ce4", + [NVKM_ENGINE_CE5 ] = "ce5", [NVKM_ENGINE_CIPHER ] = "cipher", [NVKM_ENGINE_DISP ] = "disp", [NVKM_ENGINE_DMAOBJ ] = "dma", @@ -71,6 +74,7 @@ nvkm_subdev_name[NVKM_SUBDEV_NR] = { [NVKM_ENGINE_MSVLD ] = "msvld", [NVKM_ENGINE_NVENC0 ] = "nvenc0", [NVKM_ENGINE_NVENC1 ] = "nvenc1", + [NVKM_ENGINE_NVENC2 ] = "nvenc2", [NVKM_ENGINE_NVDEC ] = "nvdec", [NVKM_ENGINE_PM ] = "pm", [NVKM_ENGINE_SEC ] = "sec", @@ -105,7 +109,7 @@ nvkm_subdev_fini(struct nvkm_subdev *subdev, bool suspend) } } - nvkm_mc_reset(device->mc, subdev->index); + nvkm_mc_reset(device, subdev->index); time = ktime_to_us(ktime_get()) - time; nvkm_trace(subdev, "%s completed in %lldus\n", action, time); diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/ce/Kbuild b/drivers/gpu/drm/nouveau/nvkm/engine/ce/Kbuild index 9c19d59b47df..a4458a8eb30a 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/ce/Kbuild +++ b/drivers/gpu/drm/nouveau/nvkm/engine/ce/Kbuild @@ -3,3 +3,5 @@ nvkm-y += nvkm/engine/ce/gf100.o nvkm-y += nvkm/engine/ce/gk104.o nvkm-y += nvkm/engine/ce/gm107.o nvkm-y += nvkm/engine/ce/gm200.o +nvkm-y += nvkm/engine/ce/gp100.o +nvkm-y += nvkm/engine/ce/gp104.o diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/ce/gp100.c b/drivers/gpu/drm/nouveau/nvkm/engine/ce/gp100.c new file mode 100644 index 000000000000..c7710456bc30 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/engine/ce/gp100.c @@ -0,0 +1,102 @@ +/* + * Copyright 2015 Red Hat Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Ben Skeggs + */ +#include "priv.h" +#include <core/enum.h> + +#include <nvif/class.h> + +static const struct nvkm_enum +gp100_ce_launcherr_report[] = { + { 0x0, "NO_ERR" }, + { 0x1, "2D_LAYER_EXCEEDS_DEPTH" }, + { 0x2, "INVALID_ALIGNMENT" }, + { 0x3, "MEM2MEM_RECT_OUT_OF_BOUNDS" }, + { 0x4, "SRC_LINE_EXCEEDS_PITCH" }, + { 0x5, "SRC_LINE_EXCEEDS_NEG_PITCH" }, + { 0x6, "DST_LINE_EXCEEDS_PITCH" }, + { 0x7, "DST_LINE_EXCEEDS_NEG_PITCH" }, + { 0x8, "BAD_SRC_PIXEL_COMP_REF" }, + { 0x9, "INVALID_VALUE" }, + { 0xa, "UNUSED_FIELD" }, + { 0xb, "INVALID_OPERATION" }, + { 0xc, "NO_RESOURCES" }, + { 0xd, "INVALID_CONFIG" }, + {} +}; + +static void +gp100_ce_intr_launcherr(struct nvkm_engine *ce, const u32 base) +{ + struct nvkm_subdev *subdev = &ce->subdev; + struct nvkm_device *device = subdev->device; + u32 stat = nvkm_rd32(device, 0x104418 + base); + const struct nvkm_enum *en = + nvkm_enum_find(gp100_ce_launcherr_report, stat & 0x0000000f); + nvkm_warn(subdev, "LAUNCHERR %08x [%s]\n", stat, en ? en->name : ""); +} + +void +gp100_ce_intr(struct nvkm_engine *ce) +{ + const u32 base = (ce->subdev.index - NVKM_ENGINE_CE0) * 0x80; + struct nvkm_subdev *subdev = &ce->subdev; + struct nvkm_device *device = subdev->device; + u32 mask = nvkm_rd32(device, 0x10440c + base); + u32 intr = nvkm_rd32(device, 0x104410 + base) & mask; + if (intr & 0x00000001) { //XXX: guess + nvkm_warn(subdev, "BLOCKPIPE\n"); + nvkm_wr32(device, 0x104410 + base, 0x00000001); + intr &= ~0x00000001; + } + if (intr & 0x00000002) { //XXX: guess + nvkm_warn(subdev, "NONBLOCKPIPE\n"); + nvkm_wr32(device, 0x104410 + base, 0x00000002); + intr &= ~0x00000002; + } + if (intr & 0x00000004) { + gp100_ce_intr_launcherr(ce, base); + nvkm_wr32(device, 0x104410 + base, 0x00000004); + intr &= ~0x00000004; + } + if (intr) { + nvkm_warn(subdev, "intr %08x\n", intr); + nvkm_wr32(device, 0x104410 + base, intr); + } +} + +static const struct nvkm_engine_func +gp100_ce = { + .intr = gp100_ce_intr, + .sclass = { + { -1, -1, PASCAL_DMA_COPY_A }, + {} + } +}; + +int +gp100_ce_new(struct nvkm_device *device, int index, + struct nvkm_engine **pengine) +{ + return nvkm_engine_new_(&gp100_ce, device, index, true, pengine); +} diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/ce/gp104.c b/drivers/gpu/drm/nouveau/nvkm/engine/ce/gp104.c new file mode 100644 index 000000000000..20e019788a53 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/engine/ce/gp104.c @@ -0,0 +1,44 @@ +/* + * Copyright 2015 Red Hat Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Ben Skeggs + */ +#include "priv.h" +#include <core/enum.h> + +#include <nvif/class.h> + +static const struct nvkm_engine_func +gp104_ce = { + .intr = gp100_ce_intr, + .sclass = { + { -1, -1, PASCAL_DMA_COPY_B }, + { -1, -1, PASCAL_DMA_COPY_A }, + {} + } +}; + +int +gp104_ce_new(struct nvkm_device *device, int index, + struct nvkm_engine **pengine) +{ + return nvkm_engine_new_(&gp104_ce, device, index, true, pengine); +} diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/ce/priv.h b/drivers/gpu/drm/nouveau/nvkm/engine/ce/priv.h index e2fa8b161943..2dce405976ad 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/ce/priv.h +++ b/drivers/gpu/drm/nouveau/nvkm/engine/ce/priv.h @@ -4,4 +4,5 @@ void gt215_ce_intr(struct nvkm_falcon *, struct nvkm_fifo_chan *); void gk104_ce_intr(struct nvkm_engine *); +void gp100_ce_intr(struct nvkm_engine *); #endif diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/device/base.c b/drivers/gpu/drm/nouveau/nvkm/engine/device/base.c index 4572debcb0c9..7218a067a6c5 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/device/base.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/device/base.c @@ -2148,6 +2148,67 @@ nv12b_chipset = { .sw = gf100_sw_new, }; +static const struct nvkm_device_chip +nv130_chipset = { + .name = "GP100", + .bar = gf100_bar_new, + .bios = nvkm_bios_new, + .bus = gf100_bus_new, + .devinit = gm200_devinit_new, + .fb = gp100_fb_new, + .fuse = gm107_fuse_new, + .gpio = gk104_gpio_new, + .i2c = gm200_i2c_new, + .ibus = gm200_ibus_new, + .imem = nv50_instmem_new, + .ltc = gp100_ltc_new, + .mc = gp100_mc_new, + .mmu = gf100_mmu_new, + .secboot = gm200_secboot_new, + .pci = gp100_pci_new, + .timer = gk20a_timer_new, + .top = gk104_top_new, + .ce[0] = gp100_ce_new, + .ce[1] = gp100_ce_new, + .ce[2] = gp100_ce_new, + .ce[3] = gp100_ce_new, + .ce[4] = gp100_ce_new, + .ce[5] = gp100_ce_new, + .dma = gf119_dma_new, + .disp = gp100_disp_new, + .fifo = gp100_fifo_new, + .gr = gp100_gr_new, + .sw = gf100_sw_new, +}; + +static const struct nvkm_device_chip +nv134_chipset = { + .name = "GP104", + .bar = gf100_bar_new, + .bios = nvkm_bios_new, + .bus = gf100_bus_new, + .devinit = gm200_devinit_new, + .fb = gp104_fb_new, + .fuse = gm107_fuse_new, + .gpio = gk104_gpio_new, + .i2c = gm200_i2c_new, + .ibus = gm200_ibus_new, + .imem = nv50_instmem_new, + .ltc = gp100_ltc_new, + .mc = gp100_mc_new, + .mmu = gf100_mmu_new, + .pci = gp100_pci_new, + .timer = gk20a_timer_new, + .top = gk104_top_new, + .ce[0] = gp104_ce_new, + .ce[1] = gp104_ce_new, + .ce[2] = gp104_ce_new, + .ce[3] = gp104_ce_new, + .disp = gp104_disp_new, + .dma = gf119_dma_new, + .fifo = gp100_fifo_new, +}; + static int nvkm_device_event_ctor(struct nvkm_object *object, void *data, u32 size, struct nvkm_notify *notify) @@ -2221,6 +2282,9 @@ nvkm_device_engine(struct nvkm_device *device, int index) _(CE0 , device->ce[0] , device->ce[0]); _(CE1 , device->ce[1] , device->ce[1]); _(CE2 , device->ce[2] , device->ce[2]); + _(CE3 , device->ce[3] , device->ce[3]); + _(CE4 , device->ce[4] , device->ce[4]); + _(CE5 , device->ce[5] , device->ce[5]); _(CIPHER , device->cipher , device->cipher); _(DISP , device->disp , &device->disp->engine); _(DMAOBJ , device->dma , &device->dma->engine); @@ -2235,6 +2299,7 @@ nvkm_device_engine(struct nvkm_device *device, int index) _(MSVLD , device->msvld , device->msvld); _(NVENC0 , device->nvenc[0], device->nvenc[0]); _(NVENC1 , device->nvenc[1], device->nvenc[1]); + _(NVENC2 , device->nvenc[2], device->nvenc[2]); _(NVDEC , device->nvdec , device->nvdec); _(PM , device->pm , &device->pm->engine); _(SEC , device->sec , device->sec); @@ -2492,6 +2557,7 @@ nvkm_device_ctor(const struct nvkm_device_func *func, case 0x100: device->card_type = NV_E0; break; case 0x110: case 0x120: device->card_type = GM100; break; + case 0x130: device->card_type = GP100; break; default: break; } @@ -2576,6 +2642,8 @@ nvkm_device_ctor(const struct nvkm_device_func *func, case 0x124: device->chip = &nv124_chipset; break; case 0x126: device->chip = &nv126_chipset; break; case 0x12b: device->chip = &nv12b_chipset; break; + case 0x130: device->chip = &nv130_chipset; break; + case 0x134: device->chip = &nv134_chipset; break; default: nvdev_error(device, "unknown chipset (%08x)\n", boot0); goto done; @@ -2659,6 +2727,9 @@ nvkm_device_ctor(const struct nvkm_device_func *func, _(NVKM_ENGINE_CE0 , ce[0]); _(NVKM_ENGINE_CE1 , ce[1]); _(NVKM_ENGINE_CE2 , ce[2]); + _(NVKM_ENGINE_CE3 , ce[3]); + _(NVKM_ENGINE_CE4 , ce[4]); + _(NVKM_ENGINE_CE5 , ce[5]); _(NVKM_ENGINE_CIPHER , cipher); _(NVKM_ENGINE_DISP , disp); _(NVKM_ENGINE_DMAOBJ , dma); @@ -2673,6 +2744,7 @@ nvkm_device_ctor(const struct nvkm_device_func *func, _(NVKM_ENGINE_MSVLD , msvld); _(NVKM_ENGINE_NVENC0 , nvenc[0]); _(NVKM_ENGINE_NVENC1 , nvenc[1]); + _(NVKM_ENGINE_NVENC2 , nvenc[2]); _(NVKM_ENGINE_NVDEC , nvdec); _(NVKM_ENGINE_PM , pm); _(NVKM_ENGINE_SEC , sec); diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/device/pci.c b/drivers/gpu/drm/nouveau/nvkm/engine/device/pci.c index 18fab3973ce5..b1b693219db3 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/device/pci.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/device/pci.c @@ -1614,7 +1614,6 @@ nvkm_device_pci_func = { .fini = nvkm_device_pci_fini, .resource_addr = nvkm_device_pci_resource_addr, .resource_size = nvkm_device_pci_resource_size, - .cpu_coherent = !IS_ENABLED(CONFIG_ARM) && !IS_ENABLED(CONFIG_ARM64), }; int diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/device/tegra.c b/drivers/gpu/drm/nouveau/nvkm/engine/device/tegra.c index ec12efb4689a..939682f18788 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/device/tegra.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/device/tegra.c @@ -191,13 +191,11 @@ static irqreturn_t nvkm_device_tegra_intr(int irq, void *arg) { struct nvkm_device_tegra *tdev = arg; - struct nvkm_mc *mc = tdev->device.mc; + struct nvkm_device *device = &tdev->device; bool handled = false; - if (likely(mc)) { - nvkm_mc_intr_unarm(mc); - nvkm_mc_intr(mc, &handled); - nvkm_mc_intr_rearm(mc); - } + nvkm_mc_intr_unarm(device); + nvkm_mc_intr(device, &handled); + nvkm_mc_intr_rearm(device); return handled ? IRQ_HANDLED : IRQ_NONE; } @@ -247,7 +245,6 @@ nvkm_device_tegra_func = { .fini = nvkm_device_tegra_fini, .resource_addr = nvkm_device_tegra_resource_addr, .resource_size = nvkm_device_tegra_resource_size, - .cpu_coherent = false, }; int @@ -313,6 +310,7 @@ nvkm_device_tegra_new(const struct nvkm_device_tegra_func *func, goto remove; tdev->gpu_speedo = tegra_sku_info.gpu_speedo_value; + tdev->gpu_speedo_id = tegra_sku_info.gpu_speedo_id; ret = nvkm_device_ctor(&nvkm_device_tegra_func, NULL, &pdev->dev, NVKM_DEVICE_TEGRA, pdev->id, NULL, cfg, dbg, detect, mmio, subdev_mask, diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/device/user.c b/drivers/gpu/drm/nouveau/nvkm/engine/device/user.c index 137066426ed7..79a8f71cf788 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/device/user.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/device/user.c @@ -102,6 +102,7 @@ nvkm_udevice_info(struct nvkm_udevice *udev, void *data, u32 size) case NV_C0: args->v0.family = NV_DEVICE_INFO_V0_FERMI; break; case NV_E0: args->v0.family = NV_DEVICE_INFO_V0_KEPLER; break; case GM100: args->v0.family = NV_DEVICE_INFO_V0_MAXWELL; break; + case GP100: args->v0.family = NV_DEVICE_INFO_V0_PASCAL; break; default: args->v0.family = 0; break; diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/Kbuild b/drivers/gpu/drm/nouveau/nvkm/engine/disp/Kbuild index a74c5dd27dc0..77a52b54a31e 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/Kbuild +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/Kbuild @@ -10,6 +10,8 @@ nvkm-y += nvkm/engine/disp/gk104.o nvkm-y += nvkm/engine/disp/gk110.o nvkm-y += nvkm/engine/disp/gm107.o nvkm-y += nvkm/engine/disp/gm200.o +nvkm-y += nvkm/engine/disp/gp100.o +nvkm-y += nvkm/engine/disp/gp104.o nvkm-y += nvkm/engine/disp/outp.o nvkm-y += nvkm/engine/disp/outpdp.o @@ -18,6 +20,7 @@ nvkm-y += nvkm/engine/disp/piornv50.o nvkm-y += nvkm/engine/disp/sornv50.o nvkm-y += nvkm/engine/disp/sorg94.o nvkm-y += nvkm/engine/disp/sorgf119.o +nvkm-y += nvkm/engine/disp/sorgm107.o nvkm-y += nvkm/engine/disp/sorgm200.o nvkm-y += nvkm/engine/disp/dport.o @@ -44,12 +47,15 @@ nvkm-y += nvkm/engine/disp/rootgk104.o nvkm-y += nvkm/engine/disp/rootgk110.o nvkm-y += nvkm/engine/disp/rootgm107.o nvkm-y += nvkm/engine/disp/rootgm200.o +nvkm-y += nvkm/engine/disp/rootgp100.o +nvkm-y += nvkm/engine/disp/rootgp104.o nvkm-y += nvkm/engine/disp/channv50.o nvkm-y += nvkm/engine/disp/changf119.o nvkm-y += nvkm/engine/disp/dmacnv50.o nvkm-y += nvkm/engine/disp/dmacgf119.o +nvkm-y += nvkm/engine/disp/dmacgp104.o nvkm-y += nvkm/engine/disp/basenv50.o nvkm-y += nvkm/engine/disp/baseg84.o @@ -58,6 +64,7 @@ nvkm-y += nvkm/engine/disp/basegt215.o nvkm-y += nvkm/engine/disp/basegf119.o nvkm-y += nvkm/engine/disp/basegk104.o nvkm-y += nvkm/engine/disp/basegk110.o +nvkm-y += nvkm/engine/disp/basegp104.o nvkm-y += nvkm/engine/disp/corenv50.o nvkm-y += nvkm/engine/disp/coreg84.o @@ -69,6 +76,8 @@ nvkm-y += nvkm/engine/disp/coregk104.o nvkm-y += nvkm/engine/disp/coregk110.o nvkm-y += nvkm/engine/disp/coregm107.o nvkm-y += nvkm/engine/disp/coregm200.o +nvkm-y += nvkm/engine/disp/coregp100.o +nvkm-y += nvkm/engine/disp/coregp104.o nvkm-y += nvkm/engine/disp/ovlynv50.o nvkm-y += nvkm/engine/disp/ovlyg84.o @@ -76,6 +85,7 @@ nvkm-y += nvkm/engine/disp/ovlygt200.o nvkm-y += nvkm/engine/disp/ovlygt215.o nvkm-y += nvkm/engine/disp/ovlygf119.o nvkm-y += nvkm/engine/disp/ovlygk104.o +nvkm-y += nvkm/engine/disp/ovlygp104.o nvkm-y += nvkm/engine/disp/piocnv50.o nvkm-y += nvkm/engine/disp/piocgf119.o diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/basegp104.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/basegp104.c new file mode 100644 index 000000000000..51688e37c54e --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/basegp104.c @@ -0,0 +1,38 @@ +/* + * Copyright 2016 Red Hat Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Ben Skeggs <bskeggs@redhat.com> + */ +#include "dmacnv50.h" +#include "rootnv50.h" + +#include <nvif/class.h> + +const struct nv50_disp_dmac_oclass +gp104_disp_base_oclass = { + .base.oclass = GK110_DISP_BASE_CHANNEL_DMA, + .base.minver = 0, + .base.maxver = 0, + .ctor = nv50_disp_base_new, + .func = &gp104_disp_dmac_func, + .mthd = &gf119_disp_base_chan_mthd, + .chid = 1, +}; diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/channv50.h b/drivers/gpu/drm/nouveau/nvkm/engine/disp/channv50.h index aee374884c96..f5f683d9fd20 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/channv50.h +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/channv50.h @@ -85,6 +85,7 @@ extern const struct nv50_disp_mthd_list gf119_disp_core_mthd_pior; extern const struct nv50_disp_chan_mthd gf119_disp_base_chan_mthd; extern const struct nv50_disp_chan_mthd gk104_disp_core_chan_mthd; +extern const struct nv50_disp_chan_mthd gk104_disp_ovly_chan_mthd; struct nv50_disp_pioc_oclass { int (*ctor)(const struct nv50_disp_chan_func *, diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/coregf119.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/coregf119.c index 6b1dc703dac7..21fbf89b6319 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/coregf119.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/coregf119.c @@ -171,7 +171,7 @@ gf119_disp_core_chan_mthd = { } }; -static void +void gf119_disp_core_fini(struct nv50_disp_dmac *chan) { struct nv50_disp *disp = chan->base.root->disp; diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/coregp100.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/coregp100.c new file mode 100644 index 000000000000..d5dff6619d4d --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/coregp100.c @@ -0,0 +1,38 @@ +/* + * Copyright 2015 Red Hat Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Ben Skeggs <bskeggs@redhat.com> + */ +#include "dmacnv50.h" +#include "rootnv50.h" + +#include <nvif/class.h> + +const struct nv50_disp_dmac_oclass +gp100_disp_core_oclass = { + .base.oclass = GP100_DISP_CORE_CHANNEL_DMA, + .base.minver = 0, + .base.maxver = 0, + .ctor = nv50_disp_core_new, + .func = &gf119_disp_core_func, + .mthd = &gk104_disp_core_chan_mthd, + .chid = 0, +}; diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/coregp104.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/coregp104.c new file mode 100644 index 000000000000..6922f4007b61 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/coregp104.c @@ -0,0 +1,78 @@ +/* + * Copyright 2016 Red Hat Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Ben Skeggs <bskeggs@redhat.com> + */ +#include "dmacnv50.h" +#include "rootnv50.h" + +#include <subdev/timer.h> + +#include <nvif/class.h> + +static int +gp104_disp_core_init(struct nv50_disp_dmac *chan) +{ + struct nv50_disp *disp = chan->base.root->disp; + struct nvkm_subdev *subdev = &disp->base.engine.subdev; + struct nvkm_device *device = subdev->device; + + /* enable error reporting */ + nvkm_mask(device, 0x6100a0, 0x00000001, 0x00000001); + + /* initialise channel for dma command submission */ + nvkm_wr32(device, 0x611494, chan->push); + nvkm_wr32(device, 0x611498, 0x00010000); + nvkm_wr32(device, 0x61149c, 0x00000001); + nvkm_mask(device, 0x610490, 0x00000010, 0x00000010); + nvkm_wr32(device, 0x640000, 0x00000000); + nvkm_wr32(device, 0x610490, 0x01000013); + + /* wait for it to go inactive */ + if (nvkm_msec(device, 2000, + if (!(nvkm_rd32(device, 0x610490) & 0x80000000)) + break; + ) < 0) { + nvkm_error(subdev, "core init: %08x\n", + nvkm_rd32(device, 0x610490)); + return -EBUSY; + } + + return 0; +} + +const struct nv50_disp_dmac_func +gp104_disp_core_func = { + .init = gp104_disp_core_init, + .fini = gf119_disp_core_fini, + .bind = gf119_disp_dmac_bind, +}; + +const struct nv50_disp_dmac_oclass +gp104_disp_core_oclass = { + .base.oclass = GP104_DISP_CORE_CHANNEL_DMA, + .base.minver = 0, + .base.maxver = 0, + .ctor = nv50_disp_core_new, + .func = &gp104_disp_core_func, + .mthd = &gk104_disp_core_chan_mthd, + .chid = 0, +}; diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/dmacgf119.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/dmacgf119.c index 876b14549a58..a57f7cef307a 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/dmacgf119.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/dmacgf119.c @@ -36,7 +36,7 @@ gf119_disp_dmac_bind(struct nv50_disp_dmac *chan, chan->base.chid << 27 | 0x00000001); } -static void +void gf119_disp_dmac_fini(struct nv50_disp_dmac *chan) { struct nv50_disp *disp = chan->base.root->disp; diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/dmacgp104.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/dmacgp104.c new file mode 100644 index 000000000000..ad24c2c57696 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/dmacgp104.c @@ -0,0 +1,66 @@ +/* + * Copyright 2016 Red Hat Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Ben Skeggs <bskeggs@redhat.com> + */ +#include "dmacnv50.h" +#include "rootnv50.h" + +#include <subdev/timer.h> + +static int +gp104_disp_dmac_init(struct nv50_disp_dmac *chan) +{ + struct nv50_disp *disp = chan->base.root->disp; + struct nvkm_subdev *subdev = &disp->base.engine.subdev; + struct nvkm_device *device = subdev->device; + int chid = chan->base.chid; + + /* enable error reporting */ + nvkm_mask(device, 0x6100a0, 0x00000001 << chid, 0x00000001 << chid); + + /* initialise channel for dma command submission */ + nvkm_wr32(device, 0x611494 + (chid * 0x0010), chan->push); + nvkm_wr32(device, 0x611498 + (chid * 0x0010), 0x00010000); + nvkm_wr32(device, 0x61149c + (chid * 0x0010), 0x00000001); + nvkm_mask(device, 0x610490 + (chid * 0x0010), 0x00000010, 0x00000010); + nvkm_wr32(device, 0x640000 + (chid * 0x1000), 0x00000000); + nvkm_wr32(device, 0x610490 + (chid * 0x0010), 0x00000013); + + /* wait for it to go inactive */ + if (nvkm_msec(device, 2000, + if (!(nvkm_rd32(device, 0x610490 + (chid * 0x10)) & 0x80000000)) + break; + ) < 0) { + nvkm_error(subdev, "ch %d init: %08x\n", chid, + nvkm_rd32(device, 0x610490 + (chid * 0x10))); + return -EBUSY; + } + + return 0; +} + +const struct nv50_disp_dmac_func +gp104_disp_dmac_func = { + .init = gp104_disp_dmac_init, + .fini = gf119_disp_dmac_fini, + .bind = gf119_disp_dmac_bind, +}; diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/dmacnv50.h b/drivers/gpu/drm/nouveau/nvkm/engine/disp/dmacnv50.h index fc84eb8b5c45..43ac05857853 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/dmacnv50.h +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/dmacnv50.h @@ -25,8 +25,12 @@ int nv50_disp_dmac_bind(struct nv50_disp_dmac *, struct nvkm_object *, u32); extern const struct nv50_disp_dmac_func nv50_disp_core_func; extern const struct nv50_disp_dmac_func gf119_disp_dmac_func; +void gf119_disp_dmac_fini(struct nv50_disp_dmac *); int gf119_disp_dmac_bind(struct nv50_disp_dmac *, struct nvkm_object *, u32); extern const struct nv50_disp_dmac_func gf119_disp_core_func; +void gf119_disp_core_fini(struct nv50_disp_dmac *); + +extern const struct nv50_disp_dmac_func gp104_disp_dmac_func; struct nv50_disp_dmac_oclass { int (*ctor)(const struct nv50_disp_dmac_func *, @@ -88,4 +92,10 @@ extern const struct nv50_disp_dmac_oclass gk110_disp_base_oclass; extern const struct nv50_disp_dmac_oclass gm107_disp_core_oclass; extern const struct nv50_disp_dmac_oclass gm200_disp_core_oclass; + +extern const struct nv50_disp_dmac_oclass gp100_disp_core_oclass; + +extern const struct nv50_disp_dmac_oclass gp104_disp_core_oclass; +extern const struct nv50_disp_dmac_oclass gp104_disp_base_oclass; +extern const struct nv50_disp_dmac_oclass gp104_disp_ovly_oclass; #endif diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gf119.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gf119.c index f0314664349c..29e84b241cca 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gf119.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gf119.c @@ -79,8 +79,7 @@ exec_lookup(struct nv50_disp *disp, int head, int or, u32 ctrl, list_for_each_entry(outp, &disp->base.outp, head) { if ((outp->info.hasht & 0xff) == type && (outp->info.hashm & mask) == mask) { - *data = nvbios_outp_match(bios, outp->info.hasht, - outp->info.hashm, + *data = nvbios_outp_match(bios, outp->info.hasht, mask, ver, hdr, cnt, len, info); if (!*data) return NULL; @@ -155,25 +154,21 @@ exec_clkcmp(struct nv50_disp *disp, int head, int id, u32 pclk, u32 *conf) if (!outp) return NULL; + *conf = (ctrl & 0x00000f00) >> 8; switch (outp->info.type) { case DCB_OUTPUT_TMDS: - *conf = (ctrl & 0x00000f00) >> 8; if (*conf == 5) *conf |= 0x0100; break; case DCB_OUTPUT_LVDS: - *conf = disp->sor.lvdsconf; + *conf |= disp->sor.lvdsconf; break; - case DCB_OUTPUT_DP: - *conf = (ctrl & 0x00000f00) >> 8; - break; - case DCB_OUTPUT_ANALOG: default: - *conf = 0x00ff; break; } - data = nvbios_ocfg_match(bios, data, *conf, &ver, &hdr, &cnt, &len, &info2); + data = nvbios_ocfg_match(bios, data, *conf & 0xff, *conf >> 8, + &ver, &hdr, &cnt, &len, &info2); if (data && id < 0xff) { data = nvbios_oclk_match(bios, info2.clkcmp[id], pclk); if (data) { @@ -418,7 +413,7 @@ gf119_disp_intr_supervisor(struct work_struct *work) nvkm_wr32(device, 0x6101d0, 0x80000000); } -static void +void gf119_disp_intr_error(struct nv50_disp *disp, int chid) { struct nvkm_subdev *subdev = &disp->base.engine.subdev; @@ -466,7 +461,7 @@ gf119_disp_intr(struct nv50_disp *disp) u32 stat = nvkm_rd32(device, 0x61009c); int chid = ffs(stat) - 1; if (chid >= 0) - gf119_disp_intr_error(disp, chid); + disp->func->intr_error(disp, chid); intr &= ~0x00000002; } @@ -510,6 +505,7 @@ gf119_disp_new_(const struct nv50_disp_func *func, struct nvkm_device *device, static const struct nv50_disp_func gf119_disp = { .intr = gf119_disp_intr, + .intr_error = gf119_disp_intr_error, .uevent = &gf119_disp_chan_uevent, .super = gf119_disp_intr_supervisor, .root = &gf119_disp_root_oclass, diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gk104.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gk104.c index a86384b8e388..37f145cf30d7 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gk104.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gk104.c @@ -27,6 +27,7 @@ static const struct nv50_disp_func gk104_disp = { .intr = gf119_disp_intr, + .intr_error = gf119_disp_intr_error, .uevent = &gf119_disp_chan_uevent, .super = gf119_disp_intr_supervisor, .root = &gk104_disp_root_oclass, diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gk110.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gk110.c index 0d574c7e594a..e14ac946608c 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gk110.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gk110.c @@ -27,6 +27,7 @@ static const struct nv50_disp_func gk110_disp = { .intr = gf119_disp_intr, + .intr_error = gf119_disp_intr_error, .uevent = &gf119_disp_chan_uevent, .super = gf119_disp_intr_supervisor, .root = &gk110_disp_root_oclass, diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gm107.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gm107.c index b6944142d616..2f2437cc5891 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gm107.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gm107.c @@ -27,6 +27,7 @@ static const struct nv50_disp_func gm107_disp = { .intr = gf119_disp_intr, + .intr_error = gf119_disp_intr_error, .uevent = &gf119_disp_chan_uevent, .super = gf119_disp_intr_supervisor, .root = &gm107_disp_root_oclass, @@ -36,7 +37,7 @@ gm107_disp = { .outp.internal.crt = nv50_dac_output_new, .outp.internal.tmds = nv50_sor_output_new, .outp.internal.lvds = nv50_sor_output_new, - .outp.internal.dp = gf119_sor_dp_new, + .outp.internal.dp = gm107_sor_dp_new, .dac.nr = 3, .dac.power = nv50_dac_power, .dac.sense = nv50_dac_sense, diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gm200.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gm200.c index 67eec8620719..9f368d4ee61e 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gm200.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gm200.c @@ -27,6 +27,7 @@ static const struct nv50_disp_func gm200_disp = { .intr = gf119_disp_intr, + .intr_error = gf119_disp_intr_error, .uevent = &gf119_disp_chan_uevent, .super = gf119_disp_intr_supervisor, .root = &gm200_disp_root_oclass, diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gp100.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gp100.c new file mode 100644 index 000000000000..4f81bf31435e --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gp100.c @@ -0,0 +1,55 @@ +/* + * Copyright 2015 Red Hat Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Ben Skeggs <bskeggs@redhat.com> + */ +#include "nv50.h" +#include "rootnv50.h" + +static const struct nv50_disp_func +gp100_disp = { + .intr = gf119_disp_intr, + .intr_error = gf119_disp_intr_error, + .uevent = &gf119_disp_chan_uevent, + .super = gf119_disp_intr_supervisor, + .root = &gp100_disp_root_oclass, + .head.vblank_init = gf119_disp_vblank_init, + .head.vblank_fini = gf119_disp_vblank_fini, + .head.scanoutpos = gf119_disp_root_scanoutpos, + .outp.internal.crt = nv50_dac_output_new, + .outp.internal.tmds = nv50_sor_output_new, + .outp.internal.lvds = nv50_sor_output_new, + .outp.internal.dp = gm200_sor_dp_new, + .dac.nr = 3, + .dac.power = nv50_dac_power, + .dac.sense = nv50_dac_sense, + .sor.nr = 4, + .sor.power = nv50_sor_power, + .sor.hda_eld = gf119_hda_eld, + .sor.hdmi = gk104_hdmi_ctrl, + .sor.magic = gm200_sor_magic, +}; + +int +gp100_disp_new(struct nvkm_device *device, int index, struct nvkm_disp **pdisp) +{ + return gf119_disp_new_(&gp100_disp, device, index, pdisp); +} diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/gp104.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gp104.c new file mode 100644 index 000000000000..3bf3380336e4 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/gp104.c @@ -0,0 +1,81 @@ +/* + * Copyright 2016 Red Hat Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Ben Skeggs <bskeggs@redhat.com> + */ +#include "nv50.h" +#include "rootnv50.h" + +static void +gp104_disp_intr_error(struct nv50_disp *disp, int chid) +{ + struct nvkm_subdev *subdev = &disp->base.engine.subdev; + struct nvkm_device *device = subdev->device; + u32 mthd = nvkm_rd32(device, 0x6111f0 + (chid * 12)); + u32 data = nvkm_rd32(device, 0x6111f4 + (chid * 12)); + u32 unkn = nvkm_rd32(device, 0x6111f8 + (chid * 12)); + + nvkm_error(subdev, "chid %d mthd %04x data %08x %08x %08x\n", + chid, (mthd & 0x0000ffc), data, mthd, unkn); + + if (chid < ARRAY_SIZE(disp->chan)) { + switch (mthd & 0xffc) { + case 0x0080: + nv50_disp_chan_mthd(disp->chan[chid], NV_DBG_ERROR); + break; + default: + break; + } + } + + nvkm_wr32(device, 0x61009c, (1 << chid)); + nvkm_wr32(device, 0x6111f0 + (chid * 12), 0x90000000); +} + +static const struct nv50_disp_func +gp104_disp = { + .intr = gf119_disp_intr, + .intr_error = gp104_disp_intr_error, + .uevent = &gf119_disp_chan_uevent, + .super = gf119_disp_intr_supervisor, + .root = &gp104_disp_root_oclass, + .head.vblank_init = gf119_disp_vblank_init, + .head.vblank_fini = gf119_disp_vblank_fini, + .head.scanoutpos = gf119_disp_root_scanoutpos, + .outp.internal.crt = nv50_dac_output_new, + .outp.internal.tmds = nv50_sor_output_new, + .outp.internal.lvds = nv50_sor_output_new, + .outp.internal.dp = gm200_sor_dp_new, + .dac.nr = 3, + .dac.power = nv50_dac_power, + .dac.sense = nv50_dac_sense, + .sor.nr = 4, + .sor.power = nv50_sor_power, + .sor.hda_eld = gf119_hda_eld, + .sor.hdmi = gk104_hdmi_ctrl, + .sor.magic = gm200_sor_magic, +}; + +int +gp104_disp_new(struct nvkm_device *device, int index, struct nvkm_disp **pdisp) +{ + return gf119_disp_new_(&gp104_disp, device, index, pdisp); +} diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/nv50.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/nv50.c index 4226d2153b9c..fbb8c7dc18fd 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/nv50.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/nv50.c @@ -32,6 +32,7 @@ #include <subdev/bios/init.h> #include <subdev/bios/pll.h> #include <subdev/devinit.h> +#include <subdev/timer.h> static const struct nvkm_disp_oclass * nv50_disp_root_(struct nvkm_disp *base) @@ -269,8 +270,7 @@ exec_lookup(struct nv50_disp *disp, int head, int or, u32 ctrl, list_for_each_entry(outp, &disp->base.outp, head) { if ((outp->info.hasht & 0xff) == type && (outp->info.hashm & mask) == mask) { - *data = nvbios_outp_match(bios, outp->info.hasht, - outp->info.hashm, + *data = nvbios_outp_match(bios, outp->info.hasht, mask, ver, hdr, cnt, len, info); if (!*data) return NULL; @@ -387,22 +387,17 @@ exec_clkcmp(struct nv50_disp *disp, int head, int id, u32 pclk, u32 *conf) if (!outp) return NULL; + *conf = (ctrl & 0x00000f00) >> 8; if (outp->info.location == 0) { switch (outp->info.type) { case DCB_OUTPUT_TMDS: - *conf = (ctrl & 0x00000f00) >> 8; if (*conf == 5) *conf |= 0x0100; break; case DCB_OUTPUT_LVDS: - *conf = disp->sor.lvdsconf; + *conf |= disp->sor.lvdsconf; break; - case DCB_OUTPUT_DP: - *conf = (ctrl & 0x00000f00) >> 8; - break; - case DCB_OUTPUT_ANALOG: default: - *conf = 0x00ff; break; } } else { @@ -410,7 +405,8 @@ exec_clkcmp(struct nv50_disp *disp, int head, int id, u32 pclk, u32 *conf) pclk = pclk / 2; } - data = nvbios_ocfg_match(bios, data, *conf, &ver, &hdr, &cnt, &len, &info2); + data = nvbios_ocfg_match(bios, data, *conf & 0xff, *conf >> 8, + &ver, &hdr, &cnt, &len, &info2); if (data && id < 0xff) { data = nvbios_oclk_match(bios, info2.clkcmp[id], pclk); if (data) { @@ -430,6 +426,134 @@ exec_clkcmp(struct nv50_disp *disp, int head, int id, u32 pclk, u32 *conf) return outp; } +static bool +nv50_disp_dptmds_war(struct nvkm_device *device) +{ + switch (device->chipset) { + case 0x94: + case 0x96: + case 0x98: + case 0xaa: + case 0xac: + return true; + default: + break; + } + return false; +} + +static bool +nv50_disp_dptmds_war_needed(struct nv50_disp *disp, struct dcb_output *outp) +{ + struct nvkm_device *device = disp->base.engine.subdev.device; + const u32 soff = __ffs(outp->or) * 0x800; + if (nv50_disp_dptmds_war(device) && outp->type == DCB_OUTPUT_TMDS) { + switch (nvkm_rd32(device, 0x614300 + soff) & 0x00030000) { + case 0x00000000: + case 0x00030000: + return true; + default: + break; + } + } + return false; + +} + +static void +nv50_disp_dptmds_war_2(struct nv50_disp *disp, struct dcb_output *outp) +{ + struct nvkm_device *device = disp->base.engine.subdev.device; + const u32 soff = __ffs(outp->or) * 0x800; + + if (!nv50_disp_dptmds_war_needed(disp, outp)) + return; + + nvkm_mask(device, 0x00e840, 0x80000000, 0x80000000); + nvkm_mask(device, 0x614300 + soff, 0x03000000, 0x03000000); + nvkm_mask(device, 0x61c10c + soff, 0x00000001, 0x00000001); + + nvkm_mask(device, 0x61c00c + soff, 0x0f000000, 0x00000000); + nvkm_mask(device, 0x61c008 + soff, 0xff000000, 0x14000000); + nvkm_usec(device, 400, NVKM_DELAY); + nvkm_mask(device, 0x61c008 + soff, 0xff000000, 0x00000000); + nvkm_mask(device, 0x61c00c + soff, 0x0f000000, 0x01000000); + + if (nvkm_rd32(device, 0x61c004 + soff) & 0x00000001) { + u32 seqctl = nvkm_rd32(device, 0x61c030 + soff); + u32 pu_pc = seqctl & 0x0000000f; + nvkm_wr32(device, 0x61c040 + soff + pu_pc * 4, 0x1f008000); + } +} + +static void +nv50_disp_dptmds_war_3(struct nv50_disp *disp, struct dcb_output *outp) +{ + struct nvkm_device *device = disp->base.engine.subdev.device; + const u32 soff = __ffs(outp->or) * 0x800; + u32 sorpwr; + + if (!nv50_disp_dptmds_war_needed(disp, outp)) + return; + + sorpwr = nvkm_rd32(device, 0x61c004 + soff); + if (sorpwr & 0x00000001) { + u32 seqctl = nvkm_rd32(device, 0x61c030 + soff); + u32 pd_pc = (seqctl & 0x00000f00) >> 8; + u32 pu_pc = seqctl & 0x0000000f; + + nvkm_wr32(device, 0x61c040 + soff + pd_pc * 4, 0x1f008000); + + nvkm_msec(device, 2000, + if (!(nvkm_rd32(device, 0x61c030 + soff) & 0x10000000)) + break; + ); + nvkm_mask(device, 0x61c004 + soff, 0x80000001, 0x80000000); + nvkm_msec(device, 2000, + if (!(nvkm_rd32(device, 0x61c030 + soff) & 0x10000000)) + break; + ); + + nvkm_wr32(device, 0x61c040 + soff + pd_pc * 4, 0x00002000); + nvkm_wr32(device, 0x61c040 + soff + pu_pc * 4, 0x1f000000); + } + + nvkm_mask(device, 0x61c10c + soff, 0x00000001, 0x00000000); + nvkm_mask(device, 0x614300 + soff, 0x03000000, 0x00000000); + + if (sorpwr & 0x00000001) { + nvkm_mask(device, 0x61c004 + soff, 0x80000001, 0x80000001); + } +} + +static void +nv50_disp_update_sppll1(struct nv50_disp *disp) +{ + struct nvkm_device *device = disp->base.engine.subdev.device; + bool used = false; + int sor; + + if (!nv50_disp_dptmds_war(device)) + return; + + for (sor = 0; sor < disp->func->sor.nr; sor++) { + u32 clksor = nvkm_rd32(device, 0x614300 + (sor * 0x800)); + switch (clksor & 0x03000000) { + case 0x02000000: + case 0x03000000: + used = true; + break; + default: + break; + } + } + + if (used) + return; + + nvkm_mask(device, 0x00e840, 0x80000000, 0x00000000); +} + static void nv50_disp_intr_unk10_0(struct nv50_disp *disp, int head) { @@ -683,6 +807,8 @@ nv50_disp_intr_unk20_2(struct nv50_disp *disp, int head) nvkm_mask(device, hreg, 0x0000000f, hval); nvkm_mask(device, oreg, mask, oval); + + nv50_disp_dptmds_war_2(disp, &outp->info); } /* If programming a TMDS output on a SOR that can also be configured for @@ -724,6 +850,7 @@ nv50_disp_intr_unk40_0(struct nv50_disp *disp, int head) if (outp->info.location == 0 && outp->info.type == DCB_OUTPUT_TMDS) nv50_disp_intr_unk40_0_tmds(disp, &outp->info); + nv50_disp_dptmds_war_3(disp, &outp->info); } void @@ -771,6 +898,7 @@ nv50_disp_intr_supervisor(struct work_struct *work) continue; nv50_disp_intr_unk40_0(disp, head); } + nv50_disp_update_sppll1(disp); } nvkm_wr32(device, 0x610030, 0x80000000); diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/nv50.h b/drivers/gpu/drm/nouveau/nvkm/engine/disp/nv50.h index aecebd8717e5..1e1de6bfe85a 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/nv50.h +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/nv50.h @@ -68,6 +68,7 @@ struct nv50_disp_func_outp { struct nv50_disp_func { void (*intr)(struct nv50_disp *); + void (*intr_error)(struct nv50_disp *, int chid); const struct nvkm_event_func *uevent; void (*super)(struct work_struct *); @@ -114,4 +115,5 @@ void gf119_disp_vblank_init(struct nv50_disp *, int); void gf119_disp_vblank_fini(struct nv50_disp *, int); void gf119_disp_intr(struct nv50_disp *); void gf119_disp_intr_supervisor(struct work_struct *); +void gf119_disp_intr_error(struct nv50_disp *, int); #endif diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/outpdp.h b/drivers/gpu/drm/nouveau/nvkm/engine/disp/outpdp.h index e9067ba4e179..4e983f6d7032 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/outpdp.h +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/outpdp.h @@ -62,7 +62,12 @@ int g94_sor_dp_lnk_pwr(struct nvkm_output_dp *, int); int gf119_sor_dp_new(struct nvkm_disp *, int, struct dcb_output *, struct nvkm_output **); int gf119_sor_dp_lnk_ctl(struct nvkm_output_dp *, int, int, bool); +int gf119_sor_dp_drv_ctl(struct nvkm_output_dp *, int, int, int, int); -int gm200_sor_dp_new(struct nvkm_disp *, int, struct dcb_output *, - struct nvkm_output **); +int gm107_sor_dp_new(struct nvkm_disp *, int, struct dcb_output *, + struct nvkm_output **); +int gm107_sor_dp_pattern(struct nvkm_output_dp *, int); + +int gm200_sor_dp_new(struct nvkm_disp *, int, struct dcb_output *, + struct nvkm_output **); #endif diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/ovlygk104.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/ovlygk104.c index 2e2dc0641ef2..2f0220b39f34 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/ovlygk104.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/ovlygk104.c @@ -80,7 +80,7 @@ gk104_disp_ovly_mthd_base = { } }; -static const struct nv50_disp_chan_mthd +const struct nv50_disp_chan_mthd gk104_disp_ovly_chan_mthd = { .name = "Overlay", .addr = 0x001000, diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/ovlygp104.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/ovlygp104.c new file mode 100644 index 000000000000..97e2dd2d908e --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/ovlygp104.c @@ -0,0 +1,38 @@ +/* + * Copyright 2012 Red Hat Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Ben Skeggs + */ +#include "dmacnv50.h" +#include "rootnv50.h" + +#include <nvif/class.h> + +const struct nv50_disp_dmac_oclass +gp104_disp_ovly_oclass = { + .base.oclass = GK104_DISP_OVERLAY_CONTROL_DMA, + .base.minver = 0, + .base.maxver = 0, + .ctor = nv50_disp_ovly_new, + .func = &gp104_disp_dmac_func, + .mthd = &gk104_disp_ovly_chan_mthd, + .chid = 5, +}; diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/rootgp100.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/rootgp100.c new file mode 100644 index 000000000000..ac8fdd728ec6 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/rootgp100.c @@ -0,0 +1,58 @@ +/* + * Copyright 2015 Red Hat Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Ben Skeggs <bskeggs@redhat.com> + */ +#include "rootnv50.h" +#include "dmacnv50.h" + +#include <nvif/class.h> + +static const struct nv50_disp_root_func +gp100_disp_root = { + .init = gf119_disp_root_init, + .fini = gf119_disp_root_fini, + .dmac = { + &gp100_disp_core_oclass, + &gk110_disp_base_oclass, + &gk104_disp_ovly_oclass, + }, + .pioc = { + &gk104_disp_oimm_oclass, + &gk104_disp_curs_oclass, + }, +}; + +static int +gp100_disp_root_new(struct nvkm_disp *disp, const struct nvkm_oclass *oclass, + void *data, u32 size, struct nvkm_object **pobject) +{ + return nv50_disp_root_new_(&gp100_disp_root, disp, oclass, + data, size, pobject); +} + +const struct nvkm_disp_oclass +gp100_disp_root_oclass = { + .base.oclass = GP100_DISP, + .base.minver = -1, + .base.maxver = -1, + .ctor = gp100_disp_root_new, +}; diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/rootgp104.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/rootgp104.c new file mode 100644 index 000000000000..8443e04dc626 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/rootgp104.c @@ -0,0 +1,58 @@ +/* + * Copyright 2016 Red Hat Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Ben Skeggs <bskeggs@redhat.com> + */ +#include "rootnv50.h" +#include "dmacnv50.h" + +#include <nvif/class.h> + +static const struct nv50_disp_root_func +gp104_disp_root = { + .init = gf119_disp_root_init, + .fini = gf119_disp_root_fini, + .dmac = { + &gp104_disp_core_oclass, + &gp104_disp_base_oclass, + &gp104_disp_ovly_oclass, + }, + .pioc = { + &gk104_disp_oimm_oclass, + &gk104_disp_curs_oclass, + }, +}; + +static int +gp104_disp_root_new(struct nvkm_disp *disp, const struct nvkm_oclass *oclass, + void *data, u32 size, struct nvkm_object **pobject) +{ + return nv50_disp_root_new_(&gp104_disp_root, disp, oclass, + data, size, pobject); +} + +const struct nvkm_disp_oclass +gp104_disp_root_oclass = { + .base.oclass = GP104_DISP, + .base.minver = -1, + .base.maxver = -1, + .ctor = gp104_disp_root_new, +}; diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/rootnv50.h b/drivers/gpu/drm/nouveau/nvkm/engine/disp/rootnv50.h index cb449ed8d92c..ad00f1724b72 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/rootnv50.h +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/rootnv50.h @@ -40,4 +40,6 @@ extern const struct nvkm_disp_oclass gk104_disp_root_oclass; extern const struct nvkm_disp_oclass gk110_disp_root_oclass; extern const struct nvkm_disp_oclass gm107_disp_root_oclass; extern const struct nvkm_disp_oclass gm200_disp_root_oclass; +extern const struct nvkm_disp_oclass gp100_disp_root_oclass; +extern const struct nvkm_disp_oclass gp104_disp_root_oclass; #endif diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgf119.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgf119.c index b4b41b135643..49bd5da194e1 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgf119.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgf119.c @@ -40,8 +40,8 @@ static int gf119_sor_dp_pattern(struct nvkm_output_dp *outp, int pattern) { struct nvkm_device *device = outp->base.disp->engine.subdev.device; - const u32 loff = gf119_sor_loff(outp); - nvkm_mask(device, 0x61c110 + loff, 0x0f0f0f0f, 0x01010101 * pattern); + const u32 soff = gf119_sor_soff(outp); + nvkm_mask(device, 0x61c110 + soff, 0x0f0f0f0f, 0x01010101 * pattern); return 0; } @@ -64,7 +64,7 @@ gf119_sor_dp_lnk_ctl(struct nvkm_output_dp *outp, int nr, int bw, bool ef) return 0; } -static int +int gf119_sor_dp_drv_ctl(struct nvkm_output_dp *outp, int ln, int vs, int pe, int pc) { diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgm107.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgm107.c new file mode 100644 index 000000000000..37790b2617c5 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgm107.c @@ -0,0 +1,53 @@ +/* + * Copyright 2016 Red Hat Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Ben Skeggs <bskeggs@redhat.com> + */ +#include "nv50.h" +#include "outpdp.h" + +int +gm107_sor_dp_pattern(struct nvkm_output_dp *outp, int pattern) +{ + struct nvkm_device *device = outp->base.disp->engine.subdev.device; + const u32 soff = outp->base.or * 0x800; + const u32 data = 0x01010101 * pattern; + if (outp->base.info.sorconf.link & 1) + nvkm_mask(device, 0x61c110 + soff, 0x0f0f0f0f, data); + else + nvkm_mask(device, 0x61c12c + soff, 0x0f0f0f0f, data); + return 0; +} + +static const struct nvkm_output_dp_func +gm107_sor_dp_func = { + .pattern = gm107_sor_dp_pattern, + .lnk_pwr = g94_sor_dp_lnk_pwr, + .lnk_ctl = gf119_sor_dp_lnk_ctl, + .drv_ctl = gf119_sor_dp_drv_ctl, +}; + +int +gm107_sor_dp_new(struct nvkm_disp *disp, int index, + struct dcb_output *dcbE, struct nvkm_output **poutp) +{ + return nvkm_output_dp_new_(&gm107_sor_dp_func, disp, index, dcbE, poutp); +} diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgm200.c b/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgm200.c index 2cfbef9c344f..c44fa7ea672a 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgm200.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/disp/sorgm200.c @@ -57,19 +57,6 @@ gm200_sor_dp_lane_map(struct nvkm_device *device, u8 lane) } static int -gm200_sor_dp_pattern(struct nvkm_output_dp *outp, int pattern) -{ - struct nvkm_device *device = outp->base.disp->engine.subdev.device; - const u32 soff = gm200_sor_soff(outp); - const u32 data = 0x01010101 * pattern; - if (outp->base.info.sorconf.link & 1) - nvkm_mask(device, 0x61c110 + soff, 0x0f0f0f0f, data); - else - nvkm_mask(device, 0x61c12c + soff, 0x0f0f0f0f, data); - return 0; -} - -static int gm200_sor_dp_lnk_pwr(struct nvkm_output_dp *outp, int nr) { struct nvkm_device *device = outp->base.disp->engine.subdev.device; @@ -129,7 +116,7 @@ gm200_sor_dp_drv_ctl(struct nvkm_output_dp *outp, static const struct nvkm_output_dp_func gm200_sor_dp_func = { - .pattern = gm200_sor_dp_pattern, + .pattern = gm107_sor_dp_pattern, .lnk_pwr = gm200_sor_dp_lnk_pwr, .lnk_ctl = gf119_sor_dp_lnk_ctl, .drv_ctl = gm200_sor_dp_drv_ctl, diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/Kbuild b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/Kbuild index 65e5d291ecda..98651a43bc12 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/Kbuild +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/Kbuild @@ -13,6 +13,7 @@ nvkm-y += nvkm/engine/fifo/gk20a.o nvkm-y += nvkm/engine/fifo/gm107.o nvkm-y += nvkm/engine/fifo/gm200.o nvkm-y += nvkm/engine/fifo/gm20b.o +nvkm-y += nvkm/engine/fifo/gp100.o nvkm-y += nvkm/engine/fifo/chan.o nvkm-y += nvkm/engine/fifo/channv50.o @@ -31,3 +32,4 @@ nvkm-y += nvkm/engine/fifo/gpfifogf100.o nvkm-y += nvkm/engine/fifo/gpfifogk104.o nvkm-y += nvkm/engine/fifo/gpfifogk110.o nvkm-y += nvkm/engine/fifo/gpfifogm200.o +nvkm-y += nvkm/engine/fifo/gpfifogp100.o diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/changk104.h b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/changk104.h index e06f4d46f802..230f64e5f731 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/changk104.h +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/changk104.h @@ -27,4 +27,5 @@ int gk104_fifo_gpfifo_new(struct nvkm_fifo *, const struct nvkm_oclass *, extern const struct nvkm_fifo_chan_oclass gk104_fifo_gpfifo_oclass; extern const struct nvkm_fifo_chan_oclass gk110_fifo_gpfifo_oclass; extern const struct nvkm_fifo_chan_oclass gm200_fifo_gpfifo_oclass; +extern const struct nvkm_fifo_chan_oclass gp100_fifo_gpfifo_oclass; #endif diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk104.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk104.c index 743f3a189f28..103c0afaaa6d 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk104.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gk104.c @@ -329,7 +329,7 @@ gk104_fifo_intr_fault(struct gk104_fifo *fifo, int unit) } if (eu == NULL) { - enum nvkm_devidx engidx = nvkm_top_fault(device->top, unit); + enum nvkm_devidx engidx = nvkm_top_fault(device, unit); if (engidx < NVKM_SUBDEV_NR) { const char *src = nvkm_subdev_name[engidx]; char *dst = en; @@ -589,7 +589,6 @@ gk104_fifo_oneinit(struct nvkm_fifo *base) struct gk104_fifo *fifo = gk104_fifo(base); struct nvkm_subdev *subdev = &fifo->base.engine.subdev; struct nvkm_device *device = subdev->device; - struct nvkm_top *top = device->top; int engn, runl, pbid, ret, i, j; enum nvkm_devidx engidx; u32 *map; @@ -608,7 +607,7 @@ gk104_fifo_oneinit(struct nvkm_fifo *base) /* Determine runlist configuration from topology device info. */ i = 0; - while ((int)(engidx = nvkm_top_engine(top, i++, &runl, &engn)) >= 0) { + while ((int)(engidx = nvkm_top_engine(device, i++, &runl, &engn)) >= 0) { /* Determine which PBDMA handles requests for this engine. */ for (j = 0, pbid = -1; j < fifo->pbdma_nr; j++) { if (map[j] & (1 << runl)) { @@ -617,8 +616,8 @@ gk104_fifo_oneinit(struct nvkm_fifo *base) } } - nvkm_debug(subdev, "engine %2d: runlist %2d pbdma %2d\n", - engn, runl, pbid); + nvkm_debug(subdev, "engine %2d: runlist %2d pbdma %2d (%s)\n", + engn, runl, pbid, nvkm_subdev_name[engidx]); fifo->engine[engn].engine = nvkm_device_engine(device, engidx); fifo->engine[engn].runl = runl; diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gp100.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gp100.c new file mode 100644 index 000000000000..eff83f7fb705 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gp100.c @@ -0,0 +1,67 @@ +/* + * Copyright 2016 Red Hat Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Ben Skeggs + */ +#include "gk104.h" +#include "changk104.h" + +static const struct nvkm_enum +gp100_fifo_fault_engine[] = { + { 0x01, "DISPLAY" }, + { 0x03, "IFB", NULL, NVKM_ENGINE_IFB }, + { 0x04, "BAR1", NULL, NVKM_SUBDEV_BAR }, + { 0x05, "BAR2", NULL, NVKM_SUBDEV_INSTMEM }, + { 0x06, "HOST0" }, + { 0x07, "HOST1" }, + { 0x08, "HOST2" }, + { 0x09, "HOST3" }, + { 0x0a, "HOST4" }, + { 0x0b, "HOST5" }, + { 0x0c, "HOST6" }, + { 0x0d, "HOST7" }, + { 0x0e, "HOST8" }, + { 0x0f, "HOST9" }, + { 0x10, "HOST10" }, + { 0x13, "PERF" }, + { 0x17, "PMU" }, + { 0x18, "PTP" }, + { 0x1f, "PHYSICAL" }, + {} +}; + +static const struct gk104_fifo_func +gp100_fifo = { + .fault.engine = gp100_fifo_fault_engine, + .fault.reason = gk104_fifo_fault_reason, + .fault.hubclient = gk104_fifo_fault_hubclient, + .fault.gpcclient = gk104_fifo_fault_gpcclient, + .chan = { + &gp100_fifo_gpfifo_oclass, + NULL + }, +}; + +int +gp100_fifo_new(struct nvkm_device *device, int index, struct nvkm_fifo **pfifo) +{ + return gk104_fifo_new_(&gp100_fifo, device, index, 4096, pfifo); +} diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gpfifogp100.c b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gpfifogp100.c new file mode 100644 index 000000000000..1530a9217aea --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/engine/fifo/gpfifogp100.c @@ -0,0 +1,34 @@ +/* + * Copyright 2016 Red Hat Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Ben Skeggs + */ +#include "changk104.h" + +#include <nvif/class.h> + +const struct nvkm_fifo_chan_oclass +gp100_fifo_gpfifo_oclass = { + .base.oclass = PASCAL_CHANNEL_GPFIFO_A, + .base.minver = 0, + .base.maxver = 0, + .ctor = gk104_fifo_gpfifo_new, +}; diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/Kbuild b/drivers/gpu/drm/nouveau/nvkm/engine/gr/Kbuild index 290ed0db8047..f1c494182248 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/Kbuild +++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/Kbuild @@ -31,6 +31,7 @@ nvkm-y += nvkm/engine/gr/gk20a.o nvkm-y += nvkm/engine/gr/gm107.o nvkm-y += nvkm/engine/gr/gm200.o nvkm-y += nvkm/engine/gr/gm20b.o +nvkm-y += nvkm/engine/gr/gp100.o nvkm-y += nvkm/engine/gr/ctxnv40.o nvkm-y += nvkm/engine/gr/ctxnv50.o @@ -48,3 +49,4 @@ nvkm-y += nvkm/engine/gr/ctxgk20a.o nvkm-y += nvkm/engine/gr/ctxgm107.o nvkm-y += nvkm/engine/gr/ctxgm200.o nvkm-y += nvkm/engine/gr/ctxgm20b.o +nvkm-y += nvkm/engine/gr/ctxgp100.o diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/ctxgf100.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/ctxgf100.c index b02d8f50ea6a..bc77eea351a5 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/ctxgf100.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/ctxgf100.c @@ -1240,7 +1240,7 @@ gf100_grctx_generate_main(struct gf100_gr *gr, struct gf100_grctx *info) const struct gf100_grctx_func *grctx = gr->func->grctx; u32 idle_timeout; - nvkm_mc_unk260(device->mc, 0); + nvkm_mc_unk260(device, 0); gf100_gr_mmio(gr, grctx->hub); gf100_gr_mmio(gr, grctx->gpc); @@ -1264,7 +1264,7 @@ gf100_grctx_generate_main(struct gf100_gr *gr, struct gf100_grctx *info) gf100_gr_icmd(gr, grctx->icmd); nvkm_wr32(device, 0x404154, idle_timeout); gf100_gr_mthd(gr, grctx->mthd); - nvkm_mc_unk260(device->mc, 1); + nvkm_mc_unk260(device, 1); } int diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/ctxgf100.h b/drivers/gpu/drm/nouveau/nvkm/engine/gr/ctxgf100.h index ac895edce164..52048b5a5274 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/ctxgf100.h +++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/ctxgf100.h @@ -101,6 +101,8 @@ void gm200_grctx_generate_405b60(struct gf100_gr *); extern const struct gf100_grctx_func gm20b_grctx; +extern const struct gf100_grctx_func gp100_grctx; + /* context init value lists */ extern const struct gf100_gr_pack gf100_grctx_pack_icmd[]; diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/ctxgf117.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/ctxgf117.c index f521de11a299..c925ade5880e 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/ctxgf117.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/ctxgf117.c @@ -226,7 +226,7 @@ gf117_grctx_generate_main(struct gf100_gr *gr, struct gf100_grctx *info) u32 idle_timeout; int i; - nvkm_mc_unk260(device->mc, 0); + nvkm_mc_unk260(device, 0); gf100_gr_mmio(gr, grctx->hub); gf100_gr_mmio(gr, grctx->gpc); @@ -253,7 +253,7 @@ gf117_grctx_generate_main(struct gf100_gr *gr, struct gf100_grctx *info) gf100_gr_icmd(gr, grctx->icmd); nvkm_wr32(device, 0x404154, idle_timeout); gf100_gr_mthd(gr, grctx->mthd); - nvkm_mc_unk260(device->mc, 1); + nvkm_mc_unk260(device, 1); } const struct gf100_grctx_func diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/ctxgk104.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/ctxgk104.c index 9ba337778ef5..c46b3fdf7203 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/ctxgk104.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/ctxgk104.c @@ -950,7 +950,7 @@ gk104_grctx_generate_main(struct gf100_gr *gr, struct gf100_grctx *info) u32 idle_timeout; int i; - nvkm_mc_unk260(device->mc, 0); + nvkm_mc_unk260(device, 0); gf100_gr_mmio(gr, grctx->hub); gf100_gr_mmio(gr, grctx->gpc); @@ -979,7 +979,7 @@ gk104_grctx_generate_main(struct gf100_gr *gr, struct gf100_grctx *info) gf100_gr_icmd(gr, grctx->icmd); nvkm_wr32(device, 0x404154, idle_timeout); gf100_gr_mthd(gr, grctx->mthd); - nvkm_mc_unk260(device->mc, 1); + nvkm_mc_unk260(device, 1); nvkm_mask(device, 0x418800, 0x00200000, 0x00200000); nvkm_mask(device, 0x41be10, 0x00800000, 0x00800000); diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/ctxgp100.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/ctxgp100.c new file mode 100644 index 000000000000..3d1ae7ddf7dd --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/ctxgp100.c @@ -0,0 +1,179 @@ +/* + * Copyright 2016 Red Hat Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Ben Skeggs <bskeggs@redhat.com> + */ +#include "ctxgf100.h" + +#include <subdev/fb.h> + +/******************************************************************************* + * PGRAPH context implementation + ******************************************************************************/ + +static void +gp100_grctx_generate_pagepool(struct gf100_grctx *info) +{ + const struct gf100_grctx_func *grctx = info->gr->func->grctx; + const u32 access = NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS; + const int s = 8; + const int b = mmio_vram(info, grctx->pagepool_size, (1 << s), access); + mmio_refn(info, 0x40800c, 0x00000000, s, b); + mmio_wr32(info, 0x408010, 0x80000000); + mmio_refn(info, 0x419004, 0x00000000, s, b); + mmio_wr32(info, 0x419008, 0x00000000); +} + +static void +gp100_grctx_generate_attrib(struct gf100_grctx *info) +{ + struct gf100_gr *gr = info->gr; + const struct gf100_grctx_func *grctx = gr->func->grctx; + const u32 alpha = grctx->alpha_nr; + const u32 attrib = grctx->attrib_nr; + const u32 pertpc = 0x20 * (grctx->attrib_nr_max + grctx->alpha_nr_max); + const u32 size = roundup(gr->tpc_total * pertpc, 0x80); + const u32 access = NV_MEM_ACCESS_RW; + const int s = 12; + const int b = mmio_vram(info, size, (1 << s), access); + const int max_batches = 0xffff; + u32 ao = 0; + u32 bo = ao + grctx->alpha_nr_max * gr->tpc_total; + int gpc, ppc, n = 0; + + mmio_refn(info, 0x418810, 0x80000000, s, b); + mmio_refn(info, 0x419848, 0x10000000, s, b); + mmio_refn(info, 0x419c2c, 0x10000000, s, b); + mmio_refn(info, 0x419b00, 0x00000000, s, b); + mmio_wr32(info, 0x419b04, 0x80000000 | size >> 7); + mmio_wr32(info, 0x405830, attrib); + mmio_wr32(info, 0x40585c, alpha); + mmio_wr32(info, 0x4064c4, ((alpha / 4) << 16) | max_batches); + + for (gpc = 0; gpc < gr->gpc_nr; gpc++) { + for (ppc = 0; ppc < gr->ppc_nr[gpc]; ppc++, n++) { + const u32 as = alpha * gr->ppc_tpc_nr[gpc][ppc]; + const u32 bs = attrib * gr->ppc_tpc_nr[gpc][ppc]; + const u32 u = 0x418ea0 + (n * 0x04); + const u32 o = PPC_UNIT(gpc, ppc, 0); + if (!(gr->ppc_mask[gpc] & (1 << ppc))) + continue; + mmio_wr32(info, o + 0xc0, bs); + mmio_wr32(info, o + 0xf4, bo); + mmio_wr32(info, o + 0xf0, bs); + bo += grctx->attrib_nr_max * gr->ppc_tpc_nr[gpc][ppc]; + mmio_wr32(info, o + 0xe4, as); + mmio_wr32(info, o + 0xf8, ao); + ao += grctx->alpha_nr_max * gr->ppc_tpc_nr[gpc][ppc]; + mmio_wr32(info, u, bs); + } + } + + mmio_wr32(info, 0x418eec, 0x00000000); + mmio_wr32(info, 0x41befc, 0x00000000); +} + +static void +gp100_grctx_generate_405b60(struct gf100_gr *gr) +{ + struct nvkm_device *device = gr->base.engine.subdev.device; + const u32 dist_nr = DIV_ROUND_UP(gr->tpc_total, 4); + u32 dist[TPC_MAX / 4] = {}; + u32 gpcs[GPC_MAX * 2] = {}; + u8 tpcnr[GPC_MAX]; + int tpc, gpc, i; + + memcpy(tpcnr, gr->tpc_nr, sizeof(gr->tpc_nr)); + + /* won't result in the same distribution as the binary driver where + * some of the gpcs have more tpcs than others, but this shall do + * for the moment. the code for earlier gpus has this issue too. + */ + for (gpc = -1, i = 0; i < gr->tpc_total; i++) { + do { + gpc = (gpc + 1) % gr->gpc_nr; + } while(!tpcnr[gpc]); + tpc = gr->tpc_nr[gpc] - tpcnr[gpc]--; + + dist[i / 4] |= ((gpc << 4) | tpc) << ((i % 4) * 8); + gpcs[gpc + (gr->gpc_nr * (tpc / 4))] |= i << (tpc * 8); + } + + for (i = 0; i < dist_nr; i++) + nvkm_wr32(device, 0x405b60 + (i * 4), dist[i]); + for (i = 0; i < gr->gpc_nr * 2; i++) + nvkm_wr32(device, 0x405ba0 + (i * 4), gpcs[i]); +} + +static void +gp100_grctx_generate_main(struct gf100_gr *gr, struct gf100_grctx *info) +{ + struct nvkm_device *device = gr->base.engine.subdev.device; + const struct gf100_grctx_func *grctx = gr->func->grctx; + u32 idle_timeout, tmp; + int i; + + gf100_gr_mmio(gr, gr->fuc_sw_ctx); + + idle_timeout = nvkm_mask(device, 0x404154, 0xffffffff, 0x00000000); + + grctx->pagepool(info); + grctx->bundle(info); + grctx->attrib(info); + grctx->unkn(gr); + + gm200_grctx_generate_tpcid(gr); + gf100_grctx_generate_r406028(gr); + gk104_grctx_generate_r418bb8(gr); + + for (i = 0; i < 8; i++) + nvkm_wr32(device, 0x4064d0 + (i * 0x04), 0x00000000); + nvkm_wr32(device, 0x406500, 0x00000000); + + nvkm_wr32(device, 0x405b00, (gr->tpc_total << 8) | gr->gpc_nr); + + for (tmp = 0, i = 0; i < gr->gpc_nr; i++) + tmp |= ((1 << gr->tpc_nr[i]) - 1) << (i * 5); + nvkm_wr32(device, 0x4041c4, tmp); + + gp100_grctx_generate_405b60(gr); + + gf100_gr_icmd(gr, gr->fuc_bundle); + nvkm_wr32(device, 0x404154, idle_timeout); + gf100_gr_mthd(gr, gr->fuc_method); +} + +const struct gf100_grctx_func +gp100_grctx = { + .main = gp100_grctx_generate_main, + .unkn = gk104_grctx_generate_unkn, + .bundle = gm107_grctx_generate_bundle, + .bundle_size = 0x3000, + .bundle_min_gpm_fifo_depth = 0x180, + .bundle_token_limit = 0x1080, + .pagepool = gp100_grctx_generate_pagepool, + .pagepool_size = 0x20000, + .attrib = gp100_grctx_generate_attrib, + .attrib_nr_max = 0x660, + .attrib_nr = 0x440, + .alpha_nr_max = 0xc00, + .alpha_nr = 0x800, +}; diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf100.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf100.c index 9513badb8220..157919c788e6 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf100.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf100.c @@ -949,22 +949,41 @@ gf100_gr_trap_gpc_rop(struct gf100_gr *gr, int gpc) } static const struct nvkm_enum gf100_mp_warp_error[] = { - { 0x00, "NO_ERROR" }, - { 0x01, "STACK_MISMATCH" }, + { 0x01, "STACK_ERROR" }, + { 0x02, "API_STACK_ERROR" }, + { 0x03, "RET_EMPTY_STACK_ERROR" }, + { 0x04, "PC_WRAP" }, { 0x05, "MISALIGNED_PC" }, - { 0x08, "MISALIGNED_GPR" }, - { 0x09, "INVALID_OPCODE" }, - { 0x0d, "GPR_OUT_OF_BOUNDS" }, - { 0x0e, "MEM_OUT_OF_BOUNDS" }, - { 0x0f, "UNALIGNED_MEM_ACCESS" }, + { 0x06, "PC_OVERFLOW" }, + { 0x07, "MISALIGNED_IMMC_ADDR" }, + { 0x08, "MISALIGNED_REG" }, + { 0x09, "ILLEGAL_INSTR_ENCODING" }, + { 0x0a, "ILLEGAL_SPH_INSTR_COMBO" }, + { 0x0b, "ILLEGAL_INSTR_PARAM" }, + { 0x0c, "INVALID_CONST_ADDR" }, + { 0x0d, "OOR_REG" }, + { 0x0e, "OOR_ADDR" }, + { 0x0f, "MISALIGNED_ADDR" }, { 0x10, "INVALID_ADDR_SPACE" }, - { 0x11, "INVALID_PARAM" }, + { 0x11, "ILLEGAL_INSTR_PARAM2" }, + { 0x12, "INVALID_CONST_ADDR_LDC" }, + { 0x13, "GEOMETRY_SM_ERROR" }, + { 0x14, "DIVERGENT" }, + { 0x15, "WARP_EXIT" }, {} }; static const struct nvkm_bitfield gf100_mp_global_error[] = { + { 0x00000001, "SM_TO_SM_FAULT" }, + { 0x00000002, "L1_ERROR" }, { 0x00000004, "MULTIPLE_WARP_ERRORS" }, - { 0x00000008, "OUT_OF_STACK_SPACE" }, + { 0x00000008, "PHYSICAL_STACK_OVERFLOW" }, + { 0x00000010, "BPT_INT" }, + { 0x00000020, "BPT_PAUSE" }, + { 0x00000040, "SINGLE_STEP_COMPLETE" }, + { 0x20000000, "ECC_SEC_ERROR" }, + { 0x40000000, "ECC_DED_ERROR" }, + { 0x80000000, "TIMEOUT" }, {} }; @@ -1438,24 +1457,30 @@ gf100_gr_init_ctxctl(struct gf100_gr *gr) struct nvkm_device *device = subdev->device; struct nvkm_secboot *sb = device->secboot; int i; + int ret = 0; if (gr->firmware) { /* load fuc microcode */ - nvkm_mc_unk260(device->mc, 0); + nvkm_mc_unk260(device, 0); /* securely-managed falcons must be reset using secure boot */ if (nvkm_secboot_is_managed(sb, NVKM_SECBOOT_FALCON_FECS)) - nvkm_secboot_reset(sb, NVKM_SECBOOT_FALCON_FECS); + ret = nvkm_secboot_reset(sb, NVKM_SECBOOT_FALCON_FECS); else gf100_gr_init_fw(gr, 0x409000, &gr->fuc409c, &gr->fuc409d); + if (ret) + return ret; + if (nvkm_secboot_is_managed(sb, NVKM_SECBOOT_FALCON_GPCCS)) - nvkm_secboot_reset(sb, NVKM_SECBOOT_FALCON_GPCCS); + ret = nvkm_secboot_reset(sb, NVKM_SECBOOT_FALCON_GPCCS); else gf100_gr_init_fw(gr, 0x41a000, &gr->fuc41ac, &gr->fuc41ad); + if (ret) + return ret; - nvkm_mc_unk260(device->mc, 1); + nvkm_mc_unk260(device, 1); /* start both of them running */ nvkm_wr32(device, 0x409840, 0xffffffff); @@ -1557,7 +1582,7 @@ gf100_gr_init_ctxctl(struct gf100_gr *gr) } /* load HUB microcode */ - nvkm_mc_unk260(device->mc, 0); + nvkm_mc_unk260(device, 0); nvkm_wr32(device, 0x4091c0, 0x01000000); for (i = 0; i < gr->func->fecs.ucode->data.size / 4; i++) nvkm_wr32(device, 0x4091c4, gr->func->fecs.ucode->data.data[i]); @@ -1580,7 +1605,7 @@ gf100_gr_init_ctxctl(struct gf100_gr *gr) nvkm_wr32(device, 0x41a188, i >> 6); nvkm_wr32(device, 0x41a184, gr->func->gpccs.ucode->code.data[i]); } - nvkm_mc_unk260(device->mc, 1); + nvkm_mc_unk260(device, 1); /* load register lists */ gf100_gr_init_csdata(gr, grctx->hub, 0x409000, 0x000, 0x000000); diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf100.h b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf100.h index 2b98abdb9270..268b8d60ff73 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf100.h +++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gf100.h @@ -292,4 +292,6 @@ extern const struct gf100_gr_init gm107_gr_init_l1c_0[]; extern const struct gf100_gr_init gm107_gr_init_wwdx_0[]; extern const struct gf100_gr_init gm107_gr_init_cbm_0[]; void gm107_gr_init_bios(struct gf100_gr *); + +void gm200_gr_init_gpc_mmu(struct gf100_gr *); #endif diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gk20a.c index 4ca8ed15191c..de8b806b88fd 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gk20a.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gk20a.c @@ -361,6 +361,5 @@ gk20a_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr) if (ret) return ret; - return 0; } diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gm200.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gm200.c index 4dfa4513bb6c..6435f1257572 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gm200.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gm200.c @@ -38,7 +38,7 @@ gm200_gr_rops(struct gf100_gr *gr) return nvkm_rd32(gr->base.engine.subdev.device, 0x12006c); } -static void +void gm200_gr_init_gpc_mmu(struct gf100_gr *gr) { struct nvkm_device *device = gr->base.engine.subdev.device; diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/gr/gp100.c b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gp100.c new file mode 100644 index 000000000000..26ad79def0ff --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/engine/gr/gp100.c @@ -0,0 +1,171 @@ +/* + * Copyright 2016 Red Hat Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Ben Skeggs <bskeggs@redhat.com> + */ +#include "gf100.h" +#include "ctxgf100.h" + +#include <nvif/class.h> + +/******************************************************************************* + * PGRAPH engine/subdev functions + ******************************************************************************/ + +static void +gp100_gr_init_rop_active_fbps(struct gf100_gr *gr) +{ + struct nvkm_device *device = gr->base.engine.subdev.device; + /*XXX: otherwise identical to gm200 aside from mask.. do everywhere? */ + const u32 fbp_count = nvkm_rd32(device, 0x12006c) & 0x0000000f; + nvkm_mask(device, 0x408850, 0x0000000f, fbp_count); /* zrop */ + nvkm_mask(device, 0x408958, 0x0000000f, fbp_count); /* crop */ +} + +static int +gp100_gr_init(struct gf100_gr *gr) +{ + struct nvkm_device *device = gr->base.engine.subdev.device; + const u32 magicgpc918 = DIV_ROUND_UP(0x00800000, gr->tpc_total); + u32 data[TPC_MAX / 8] = {}; + u8 tpcnr[GPC_MAX]; + int gpc, tpc, rop; + int i; + + gr->func->init_gpc_mmu(gr); + + gf100_gr_mmio(gr, gr->fuc_sw_nonctx); + + nvkm_wr32(device, GPC_UNIT(0, 0x3018), 0x00000001); + + memset(data, 0x00, sizeof(data)); + memcpy(tpcnr, gr->tpc_nr, sizeof(gr->tpc_nr)); + for (i = 0, gpc = -1; i < gr->tpc_total; i++) { + do { + gpc = (gpc + 1) % gr->gpc_nr; + } while (!tpcnr[gpc]); + tpc = gr->tpc_nr[gpc] - tpcnr[gpc]--; + + data[i / 8] |= tpc << ((i % 8) * 4); + } + + nvkm_wr32(device, GPC_BCAST(0x0980), data[0]); + nvkm_wr32(device, GPC_BCAST(0x0984), data[1]); + nvkm_wr32(device, GPC_BCAST(0x0988), data[2]); + nvkm_wr32(device, GPC_BCAST(0x098c), data[3]); + + for (gpc = 0; gpc < gr->gpc_nr; gpc++) { + nvkm_wr32(device, GPC_UNIT(gpc, 0x0914), + gr->screen_tile_row_offset << 8 | gr->tpc_nr[gpc]); + nvkm_wr32(device, GPC_UNIT(gpc, 0x0910), 0x00040000 | + gr->tpc_total); + nvkm_wr32(device, GPC_UNIT(gpc, 0x0918), magicgpc918); + } + + nvkm_wr32(device, GPC_BCAST(0x3fd4), magicgpc918); + nvkm_wr32(device, GPC_BCAST(0x08ac), nvkm_rd32(device, 0x100800)); + nvkm_wr32(device, GPC_BCAST(0x033c), nvkm_rd32(device, 0x100804)); + + gr->func->init_rop_active_fbps(gr); + + nvkm_wr32(device, 0x400500, 0x00010001); + nvkm_wr32(device, 0x400100, 0xffffffff); + nvkm_wr32(device, 0x40013c, 0xffffffff); + nvkm_wr32(device, 0x400124, 0x00000002); + nvkm_wr32(device, 0x409c24, 0x000f0002); + nvkm_wr32(device, 0x405848, 0xc0000000); + nvkm_mask(device, 0x40584c, 0x00000000, 0x00000001); + nvkm_wr32(device, 0x404000, 0xc0000000); + nvkm_wr32(device, 0x404600, 0xc0000000); + nvkm_wr32(device, 0x408030, 0xc0000000); + nvkm_wr32(device, 0x404490, 0xc0000000); + nvkm_wr32(device, 0x406018, 0xc0000000); + nvkm_wr32(device, 0x407020, 0x40000000); + nvkm_wr32(device, 0x405840, 0xc0000000); + nvkm_wr32(device, 0x405844, 0x00ffffff); + nvkm_mask(device, 0x419cc0, 0x00000008, 0x00000008); + + nvkm_mask(device, 0x419c9c, 0x00010000, 0x00010000); + nvkm_mask(device, 0x419c9c, 0x00020000, 0x00020000); + + gr->func->init_ppc_exceptions(gr); + + for (gpc = 0; gpc < gr->gpc_nr; gpc++) { + nvkm_wr32(device, GPC_UNIT(gpc, 0x0420), 0xc0000000); + nvkm_wr32(device, GPC_UNIT(gpc, 0x0900), 0xc0000000); + nvkm_wr32(device, GPC_UNIT(gpc, 0x1028), 0xc0000000); + nvkm_wr32(device, GPC_UNIT(gpc, 0x0824), 0xc0000000); + for (tpc = 0; tpc < gr->tpc_nr[gpc]; tpc++) { + nvkm_wr32(device, TPC_UNIT(gpc, tpc, 0x508), 0xffffffff); + nvkm_wr32(device, TPC_UNIT(gpc, tpc, 0x50c), 0xffffffff); + nvkm_wr32(device, TPC_UNIT(gpc, tpc, 0x224), 0xc0000000); + nvkm_wr32(device, TPC_UNIT(gpc, tpc, 0x48c), 0xc0000000); + nvkm_wr32(device, TPC_UNIT(gpc, tpc, 0x084), 0xc0000000); + nvkm_wr32(device, TPC_UNIT(gpc, tpc, 0x430), 0xc0000000); + nvkm_wr32(device, TPC_UNIT(gpc, tpc, 0x644), 0x00dffffe); + nvkm_wr32(device, TPC_UNIT(gpc, tpc, 0x64c), 0x00000105); + } + nvkm_wr32(device, GPC_UNIT(gpc, 0x2c90), 0xffffffff); + nvkm_wr32(device, GPC_UNIT(gpc, 0x2c94), 0xffffffff); + } + + for (rop = 0; rop < gr->rop_nr; rop++) { + nvkm_wr32(device, ROP_UNIT(rop, 0x144), 0x40000000); + nvkm_wr32(device, ROP_UNIT(rop, 0x070), 0x40000000); + nvkm_wr32(device, ROP_UNIT(rop, 0x204), 0xffffffff); + nvkm_wr32(device, ROP_UNIT(rop, 0x208), 0xffffffff); + } + + nvkm_wr32(device, 0x400108, 0xffffffff); + nvkm_wr32(device, 0x400138, 0xffffffff); + nvkm_wr32(device, 0x400118, 0xffffffff); + nvkm_wr32(device, 0x400130, 0xffffffff); + nvkm_wr32(device, 0x40011c, 0xffffffff); + nvkm_wr32(device, 0x400134, 0xffffffff); + + gf100_gr_zbc_init(gr); + + return gf100_gr_init_ctxctl(gr); +} + +static const struct gf100_gr_func +gp100_gr = { + .init = gp100_gr_init, + .init_gpc_mmu = gm200_gr_init_gpc_mmu, + .init_rop_active_fbps = gp100_gr_init_rop_active_fbps, + .init_ppc_exceptions = gk104_gr_init_ppc_exceptions, + .rops = gm200_gr_rops, + .ppc_nr = 2, + .grctx = &gp100_grctx, + .sclass = { + { -1, -1, FERMI_TWOD_A }, + { -1, -1, KEPLER_INLINE_TO_MEMORY_B }, + { -1, -1, PASCAL_A, &gf100_fermi }, + { -1, -1, PASCAL_COMPUTE_A }, + {} + } +}; + +int +gp100_gr_new(struct nvkm_device *device, int index, struct nvkm_gr **pgr) +{ + return gm200_gr_new_(&gp100_gr, device, index, pgr); +} diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bios/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/bios/base.c index e15b9627b07e..f3c30b2a788e 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/bios/base.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bios/base.c @@ -26,6 +26,49 @@ #include <subdev/bios.h> #include <subdev/bios/bmp.h> #include <subdev/bios/bit.h> +#include <subdev/bios/image.h> + +static bool +nvbios_addr(struct nvkm_bios *bios, u32 *addr, u8 size) +{ + u32 p = *addr; + + if (*addr > bios->image0_size && bios->imaged_addr) { + *addr -= bios->image0_size; + *addr += bios->imaged_addr; + } + + if (unlikely(*addr + size >= bios->size)) { + nvkm_error(&bios->subdev, "OOB %d %08x %08x\n", size, p, *addr); + return false; + } + + return true; +} + +u8 +nvbios_rd08(struct nvkm_bios *bios, u32 addr) +{ + if (likely(nvbios_addr(bios, &addr, 1))) + return bios->data[addr]; + return 0x00; +} + +u16 +nvbios_rd16(struct nvkm_bios *bios, u32 addr) +{ + if (likely(nvbios_addr(bios, &addr, 2))) + return get_unaligned_le16(&bios->data[addr]); + return 0x0000; +} + +u32 +nvbios_rd32(struct nvkm_bios *bios, u32 addr) +{ + if (likely(nvbios_addr(bios, &addr, 4))) + return get_unaligned_le32(&bios->data[addr]); + return 0x00000000; +} u8 nvbios_checksum(const u8 *data, int size) @@ -100,8 +143,9 @@ int nvkm_bios_new(struct nvkm_device *device, int index, struct nvkm_bios **pbios) { struct nvkm_bios *bios; + struct nvbios_image image; struct bit_entry bit_i; - int ret; + int ret, idx = 0; if (!(bios = *pbios = kzalloc(sizeof(*bios), GFP_KERNEL))) return -ENOMEM; @@ -111,6 +155,19 @@ nvkm_bios_new(struct nvkm_device *device, int index, struct nvkm_bios **pbios) if (ret) return ret; + /* Some tables have weird pointers that need adjustment before + * they're dereferenced. I'm not entirely sure why... + */ + if (nvbios_image(bios, idx++, &image)) { + bios->image0_size = image.size; + while (nvbios_image(bios, idx++, &image)) { + if (image.type == 0xe0) { + bios->imaged_addr = image.base; + break; + } + } + } + /* detect type of vbios we're dealing with */ bios->bmp_offset = nvbios_findstr(bios->data, bios->size, "\xff\x7f""NV\0", 5); diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bios/disp.c b/drivers/gpu/drm/nouveau/nvkm/subdev/bios/disp.c index a5e92135cd77..9efb1b48cd54 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/bios/disp.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bios/disp.c @@ -141,7 +141,8 @@ nvbios_ocfg_parse(struct nvkm_bios *bios, u16 outp, u8 idx, { u16 data = nvbios_ocfg_entry(bios, outp, idx, ver, hdr, cnt, len); if (data) { - info->match = nvbios_rd16(bios, data + 0x00); + info->proto = nvbios_rd08(bios, data + 0x00); + info->flags = nvbios_rd16(bios, data + 0x01); info->clkcmp[0] = nvbios_rd16(bios, data + 0x02); info->clkcmp[1] = nvbios_rd16(bios, data + 0x04); } @@ -149,12 +150,13 @@ nvbios_ocfg_parse(struct nvkm_bios *bios, u16 outp, u8 idx, } u16 -nvbios_ocfg_match(struct nvkm_bios *bios, u16 outp, u16 type, +nvbios_ocfg_match(struct nvkm_bios *bios, u16 outp, u8 proto, u8 flags, u8 *ver, u8 *hdr, u8 *cnt, u8 *len, struct nvbios_ocfg *info) { u16 data, idx = 0; while ((data = nvbios_ocfg_parse(bios, outp, idx++, ver, hdr, cnt, len, info))) { - if (info->match == type) + if ((info->proto == proto || info->proto == 0xff) && + (info->flags == flags)) break; } return data; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bios/dp.c b/drivers/gpu/drm/nouveau/nvkm/subdev/bios/dp.c index 05332476354a..d89e78c4e689 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/bios/dp.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bios/dp.c @@ -40,6 +40,7 @@ nvbios_dp_table(struct nvkm_bios *bios, u8 *ver, u8 *hdr, u8 *cnt, u8 *len) case 0x30: case 0x40: case 0x41: + case 0x42: *hdr = nvbios_rd08(bios, data + 0x01); *len = nvbios_rd08(bios, data + 0x02); *cnt = nvbios_rd08(bios, data + 0x03); @@ -70,6 +71,7 @@ nvbios_dpout_entry(struct nvkm_bios *bios, u8 idx, break; case 0x40: case 0x41: + case 0x42: *hdr = nvbios_rd08(bios, data + 0x04); *cnt = 0; *len = 0; @@ -109,6 +111,7 @@ nvbios_dpout_parse(struct nvkm_bios *bios, u8 idx, break; case 0x40: case 0x41: + case 0x42: info->flags = nvbios_rd08(bios, data + 0x04); info->script[0] = nvbios_rd16(bios, data + 0x05); info->script[1] = nvbios_rd16(bios, data + 0x07); @@ -180,6 +183,11 @@ nvbios_dpcfg_parse(struct nvkm_bios *bios, u16 outp, u8 idx, info->pe = nvbios_rd08(bios, data + 0x02); info->tx_pu = nvbios_rd08(bios, data + 0x03); break; + case 0x42: + info->dc = nvbios_rd08(bios, data + 0x00); + info->pe = nvbios_rd08(bios, data + 0x01); + info->tx_pu = nvbios_rd08(bios, data + 0x02); + break; default: data = 0x0000; break; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bios/image.c b/drivers/gpu/drm/nouveau/nvkm/subdev/bios/image.c index 74b14cf09308..1dbff7aeafec 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/bios/image.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bios/image.c @@ -68,11 +68,16 @@ nvbios_imagen(struct nvkm_bios *bios, struct nvbios_image *image) bool nvbios_image(struct nvkm_bios *bios, int idx, struct nvbios_image *image) { + u32 imaged_addr = bios->imaged_addr; memset(image, 0x00, sizeof(*image)); + bios->imaged_addr = 0; do { image->base += image->size; - if (image->last || !nvbios_imagen(bios, image)) + if (image->last || !nvbios_imagen(bios, image)) { + bios->imaged_addr = imaged_addr; return false; + } } while(idx--); + bios->imaged_addr = imaged_addr; return true; } diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bios/pll.c b/drivers/gpu/drm/nouveau/nvkm/subdev/bios/pll.c index 91a7dc56e406..2ca23a9157ab 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/bios/pll.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bios/pll.c @@ -77,15 +77,17 @@ g84_pll_mapping[] = { {} }; -static u16 +static u32 pll_limits_table(struct nvkm_bios *bios, u8 *ver, u8 *hdr, u8 *cnt, u8 *len) { struct bit_entry bit_C; - u16 data = 0x0000; + u32 data = 0x0000; if (!bit_entry(bios, 'C', &bit_C)) { if (bit_C.version == 1 && bit_C.length >= 10) data = nvbios_rd16(bios, bit_C.offset + 8); + if (bit_C.version == 2 && bit_C.length >= 4) + data = nvbios_rd32(bios, bit_C.offset + 0); if (data) { *ver = nvbios_rd08(bios, data + 0); *hdr = nvbios_rd08(bios, data + 1); @@ -137,12 +139,12 @@ pll_map(struct nvkm_bios *bios) } } -static u16 +static u32 pll_map_reg(struct nvkm_bios *bios, u32 reg, u32 *type, u8 *ver, u8 *len) { struct pll_mapping *map; u8 hdr, cnt; - u16 data; + u32 data; data = pll_limits_table(bios, ver, &hdr, &cnt, len); if (data && *ver >= 0x30) { @@ -160,7 +162,7 @@ pll_map_reg(struct nvkm_bios *bios, u32 reg, u32 *type, u8 *ver, u8 *len) map = pll_map(bios); while (map && map->reg) { if (map->reg == reg && *ver >= 0x20) { - u16 addr = (data += hdr); + u32 addr = (data += hdr); *type = map->type; while (cnt--) { if (nvbios_rd32(bios, data) == map->reg) @@ -179,12 +181,12 @@ pll_map_reg(struct nvkm_bios *bios, u32 reg, u32 *type, u8 *ver, u8 *len) return 0x0000; } -static u16 +static u32 pll_map_type(struct nvkm_bios *bios, u8 type, u32 *reg, u8 *ver, u8 *len) { struct pll_mapping *map; u8 hdr, cnt; - u16 data; + u32 data; data = pll_limits_table(bios, ver, &hdr, &cnt, len); if (data && *ver >= 0x30) { @@ -202,7 +204,7 @@ pll_map_type(struct nvkm_bios *bios, u8 type, u32 *reg, u8 *ver, u8 *len) map = pll_map(bios); while (map && map->reg) { if (map->type == type && *ver >= 0x20) { - u16 addr = (data += hdr); + u32 addr = (data += hdr); *reg = map->reg; while (cnt--) { if (nvbios_rd32(bios, data) == map->reg) @@ -228,7 +230,7 @@ nvbios_pll_parse(struct nvkm_bios *bios, u32 type, struct nvbios_pll *info) struct nvkm_device *device = subdev->device; u8 ver, len; u32 reg = type; - u16 data; + u32 data; if (type > PLL_MAX) { reg = type; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bios/pmu.c b/drivers/gpu/drm/nouveau/nvkm/subdev/bios/pmu.c index c268e5afe852..b4a308f3cf7b 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/bios/pmu.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bios/pmu.c @@ -26,21 +26,6 @@ #include <subdev/bios/image.h> #include <subdev/bios/pmu.h> -static u32 -weirdo_pointer(struct nvkm_bios *bios, u32 data) -{ - struct nvbios_image image; - int idx = 0; - if (nvbios_image(bios, idx++, &image)) { - data -= image.size; - while (nvbios_image(bios, idx++, &image)) { - if (image.type == 0xe0) - return image.base + data; - } - } - return 0; -} - u32 nvbios_pmuTe(struct nvkm_bios *bios, u8 *ver, u8 *hdr, u8 *cnt, u8 *len) { @@ -50,7 +35,7 @@ nvbios_pmuTe(struct nvkm_bios *bios, u8 *ver, u8 *hdr, u8 *cnt, u8 *len) if (!bit_entry(bios, 'p', &bit_p)) { if (bit_p.version == 2 && bit_p.length >= 4) data = nvbios_rd32(bios, bit_p.offset + 0x00); - if ((data = weirdo_pointer(bios, data))) { + if (data) { *ver = nvbios_rd08(bios, data + 0x00); /* maybe? */ *hdr = nvbios_rd08(bios, data + 0x01); *len = nvbios_rd08(bios, data + 0x02); @@ -97,8 +82,7 @@ nvbios_pmuRm(struct nvkm_bios *bios, u8 type, struct nvbios_pmuR *info) u32 data; memset(info, 0x00, sizeof(*info)); while ((data = nvbios_pmuEp(bios, idx++, &ver, &hdr, &pmuE))) { - if ( pmuE.type == type && - (data = weirdo_pointer(bios, pmuE.data))) { + if (pmuE.type == type && (data = pmuE.data)) { info->init_addr_pmu = nvbios_rd32(bios, data + 0x08); info->args_addr_pmu = nvbios_rd32(bios, data + 0x0c); info->boot_addr = data + 0x30; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/bios/rammap.c b/drivers/gpu/drm/nouveau/nvkm/subdev/bios/rammap.c index d0ae7454764e..b57c370c725d 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/bios/rammap.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/bios/rammap.c @@ -30,11 +30,11 @@ nvbios_rammapTe(struct nvkm_bios *bios, u8 *ver, u8 *hdr, u8 *cnt, u8 *len, u8 *snr, u8 *ssz) { struct bit_entry bit_P; - u16 rammap = 0x0000; + u32 rammap = 0x0000; if (!bit_entry(bios, 'P', &bit_P)) { if (bit_P.version == 2) - rammap = nvbios_rd16(bios, bit_P.offset + 4); + rammap = nvbios_rd32(bios, bit_P.offset + 4); if (rammap) { *ver = nvbios_rd08(bios, rammap + 0); @@ -61,7 +61,7 @@ nvbios_rammapEe(struct nvkm_bios *bios, int idx, u8 *ver, u8 *hdr, u8 *cnt, u8 *len) { u8 snr, ssz; - u16 rammap = nvbios_rammapTe(bios, ver, hdr, cnt, len, &snr, &ssz); + u32 rammap = nvbios_rammapTe(bios, ver, hdr, cnt, len, &snr, &ssz); if (rammap && idx < *cnt) { rammap = rammap + *hdr + (idx * (*len + (snr * ssz))); *hdr = *len; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gf100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gf100.c index 78c449b417b7..89d5543118cf 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gf100.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gf100.c @@ -99,7 +99,7 @@ read_div(struct gf100_clk *clk, int doff, u32 dsrc, u32 dctl) { struct nvkm_device *device = clk->base.subdev.device; u32 ssrc = nvkm_rd32(device, dsrc + (doff * 4)); - u32 sctl = nvkm_rd32(device, dctl + (doff * 4)); + u32 sclk, sctl, sdiv = 2; switch (ssrc & 0x00000003) { case 0: @@ -109,13 +109,21 @@ read_div(struct gf100_clk *clk, int doff, u32 dsrc, u32 dctl) case 2: return 100000; case 3: - if (sctl & 0x80000000) { - u32 sclk = read_vco(clk, dsrc + (doff * 4)); - u32 sdiv = (sctl & 0x0000003f) + 2; - return (sclk * 2) / sdiv; + sclk = read_vco(clk, dsrc + (doff * 4)); + + /* Memclk has doff of 0 despite its alt. location */ + if (doff <= 2) { + sctl = nvkm_rd32(device, dctl + (doff * 4)); + + if (sctl & 0x80000000) { + if (ssrc & 0x100) + sctl >>= 8; + + sdiv = (sctl & 0x3f) + 2; + } } - return read_vco(clk, dsrc + (doff * 4)); + return (sclk * 2) / sdiv; default: return 0; } @@ -366,11 +374,17 @@ gf100_clk_prog_2(struct gf100_clk *clk, int idx) if (info->coef) { nvkm_wr32(device, addr + 0x04, info->coef); nvkm_mask(device, addr + 0x00, 0x00000001, 0x00000001); + + /* Test PLL lock */ + nvkm_mask(device, addr + 0x00, 0x00000010, 0x00000000); nvkm_msec(device, 2000, if (nvkm_rd32(device, addr + 0x00) & 0x00020000) break; ); - nvkm_mask(device, addr + 0x00, 0x00020004, 0x00000004); + nvkm_mask(device, addr + 0x00, 0x00000010, 0x00000010); + + /* Enable sync mode */ + nvkm_mask(device, addr + 0x00, 0x00000004, 0x00000004); } } } diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk104.c b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk104.c index 975c401bccab..06bc0d2d6ae1 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk104.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk104.c @@ -393,11 +393,17 @@ gk104_clk_prog_2(struct gk104_clk *clk, int idx) if (info->coef) { nvkm_wr32(device, addr + 0x04, info->coef); nvkm_mask(device, addr + 0x00, 0x00000001, 0x00000001); + + /* Test PLL lock */ + nvkm_mask(device, addr + 0x00, 0x00000010, 0x00000000); nvkm_msec(device, 2000, if (nvkm_rd32(device, addr + 0x00) & 0x00020000) break; ); - nvkm_mask(device, addr + 0x00, 0x00020004, 0x00000004); + nvkm_mask(device, addr + 0x00, 0x00000010, 0x00000010); + + /* Enable sync mode */ + nvkm_mask(device, addr + 0x00, 0x00000004, 0x00000004); } } diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk20a.c index 5f0ee24e31b8..218893e3e5f9 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk20a.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk20a.c @@ -28,69 +28,6 @@ #include <core/tegra.h> #include <subdev/timer.h> -#define KHZ (1000) -#define MHZ (KHZ * 1000) - -#define MASK(w) ((1 << w) - 1) - -#define GPCPLL_CFG (SYS_GPCPLL_CFG_BASE + 0) -#define GPCPLL_CFG_ENABLE BIT(0) -#define GPCPLL_CFG_IDDQ BIT(1) -#define GPCPLL_CFG_LOCK_DET_OFF BIT(4) -#define GPCPLL_CFG_LOCK BIT(17) - -#define GPCPLL_COEFF (SYS_GPCPLL_CFG_BASE + 4) -#define GPCPLL_COEFF_M_SHIFT 0 -#define GPCPLL_COEFF_M_WIDTH 8 -#define GPCPLL_COEFF_N_SHIFT 8 -#define GPCPLL_COEFF_N_WIDTH 8 -#define GPCPLL_COEFF_P_SHIFT 16 -#define GPCPLL_COEFF_P_WIDTH 6 - -#define GPCPLL_CFG2 (SYS_GPCPLL_CFG_BASE + 0xc) -#define GPCPLL_CFG2_SETUP2_SHIFT 16 -#define GPCPLL_CFG2_PLL_STEPA_SHIFT 24 - -#define GPCPLL_CFG3 (SYS_GPCPLL_CFG_BASE + 0x18) -#define GPCPLL_CFG3_PLL_STEPB_SHIFT 16 - -#define GPC_BCASE_GPCPLL_CFG_BASE 0x00132800 -#define GPCPLL_NDIV_SLOWDOWN (SYS_GPCPLL_CFG_BASE + 0x1c) -#define GPCPLL_NDIV_SLOWDOWN_NDIV_LO_SHIFT 0 -#define GPCPLL_NDIV_SLOWDOWN_NDIV_MID_SHIFT 8 -#define GPCPLL_NDIV_SLOWDOWN_STEP_SIZE_LO2MID_SHIFT 16 -#define GPCPLL_NDIV_SLOWDOWN_SLOWDOWN_USING_PLL_SHIFT 22 -#define GPCPLL_NDIV_SLOWDOWN_EN_DYNRAMP_SHIFT 31 - -#define SEL_VCO (SYS_GPCPLL_CFG_BASE + 0x100) -#define SEL_VCO_GPC2CLK_OUT_SHIFT 0 - -#define GPC2CLK_OUT (SYS_GPCPLL_CFG_BASE + 0x250) -#define GPC2CLK_OUT_SDIV14_INDIV4_WIDTH 1 -#define GPC2CLK_OUT_SDIV14_INDIV4_SHIFT 31 -#define GPC2CLK_OUT_SDIV14_INDIV4_MODE 1 -#define GPC2CLK_OUT_VCODIV_WIDTH 6 -#define GPC2CLK_OUT_VCODIV_SHIFT 8 -#define GPC2CLK_OUT_VCODIV1 0 -#define GPC2CLK_OUT_VCODIV_MASK (MASK(GPC2CLK_OUT_VCODIV_WIDTH) << \ - GPC2CLK_OUT_VCODIV_SHIFT) -#define GPC2CLK_OUT_BYPDIV_WIDTH 6 -#define GPC2CLK_OUT_BYPDIV_SHIFT 0 -#define GPC2CLK_OUT_BYPDIV31 0x3c -#define GPC2CLK_OUT_INIT_MASK ((MASK(GPC2CLK_OUT_SDIV14_INDIV4_WIDTH) << \ - GPC2CLK_OUT_SDIV14_INDIV4_SHIFT)\ - | (MASK(GPC2CLK_OUT_VCODIV_WIDTH) << GPC2CLK_OUT_VCODIV_SHIFT)\ - | (MASK(GPC2CLK_OUT_BYPDIV_WIDTH) << GPC2CLK_OUT_BYPDIV_SHIFT)) -#define GPC2CLK_OUT_INIT_VAL ((GPC2CLK_OUT_SDIV14_INDIV4_MODE << \ - GPC2CLK_OUT_SDIV14_INDIV4_SHIFT) \ - | (GPC2CLK_OUT_VCODIV1 << GPC2CLK_OUT_VCODIV_SHIFT) \ - | (GPC2CLK_OUT_BYPDIV31 << GPC2CLK_OUT_BYPDIV_SHIFT)) - -#define GPC_BCAST_NDIV_SLOWDOWN_DEBUG (GPC_BCASE_GPCPLL_CFG_BASE + 0xa0) -#define GPC_BCAST_NDIV_SLOWDOWN_DEBUG_PLL_DYNRAMP_DONE_SYNCED_SHIFT 24 -#define GPC_BCAST_NDIV_SLOWDOWN_DEBUG_PLL_DYNRAMP_DONE_SYNCED_MASK \ - (0x1 << GPC_BCAST_NDIV_SLOWDOWN_DEBUG_PLL_DYNRAMP_DONE_SYNCED_SHIFT) - static const u8 _pl_to_div[] = { /* PL: 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14 */ /* p: */ 1, 2, 3, 4, 5, 6, 8, 10, 12, 16, 12, 16, 20, 24, 32, @@ -124,7 +61,7 @@ static const struct gk20a_clk_pllg_params gk20a_pllg_params = { .min_pl = 1, .max_pl = 32, }; -static void +void gk20a_pllg_read_mnp(struct gk20a_clk *clk, struct gk20a_pll *pll) { struct nvkm_device *device = clk->base.subdev.device; @@ -136,20 +73,33 @@ gk20a_pllg_read_mnp(struct gk20a_clk *clk, struct gk20a_pll *pll) pll->pl = (val >> GPCPLL_COEFF_P_SHIFT) & MASK(GPCPLL_COEFF_P_WIDTH); } -static u32 -gk20a_pllg_calc_rate(struct gk20a_clk *clk) +void +gk20a_pllg_write_mnp(struct gk20a_clk *clk, const struct gk20a_pll *pll) +{ + struct nvkm_device *device = clk->base.subdev.device; + u32 val; + + val = (pll->m & MASK(GPCPLL_COEFF_M_WIDTH)) << GPCPLL_COEFF_M_SHIFT; + val |= (pll->n & MASK(GPCPLL_COEFF_N_WIDTH)) << GPCPLL_COEFF_N_SHIFT; + val |= (pll->pl & MASK(GPCPLL_COEFF_P_WIDTH)) << GPCPLL_COEFF_P_SHIFT; + nvkm_wr32(device, GPCPLL_COEFF, val); +} + +u32 +gk20a_pllg_calc_rate(struct gk20a_clk *clk, struct gk20a_pll *pll) { u32 rate; u32 divider; - rate = clk->parent_rate * clk->pll.n; - divider = clk->pll.m * clk->pl_to_div(clk->pll.pl); + rate = clk->parent_rate * pll->n; + divider = pll->m * clk->pl_to_div(pll->pl); return rate / divider / 2; } -static int -gk20a_pllg_calc_mnp(struct gk20a_clk *clk, unsigned long rate) +int +gk20a_pllg_calc_mnp(struct gk20a_clk *clk, unsigned long rate, + struct gk20a_pll *pll) { struct nvkm_subdev *subdev = &clk->base.subdev; u32 target_clk_f, ref_clk_f, target_freq; @@ -163,16 +113,13 @@ gk20a_pllg_calc_mnp(struct gk20a_clk *clk, unsigned long rate) target_clk_f = rate * 2 / KHZ; ref_clk_f = clk->parent_rate / KHZ; - max_vco_f = clk->params->max_vco; + target_vco_f = target_clk_f + target_clk_f / 50; + max_vco_f = max(clk->params->max_vco, target_vco_f); min_vco_f = clk->params->min_vco; best_m = clk->params->max_m; best_n = clk->params->min_n; best_pl = clk->params->min_pl; - target_vco_f = target_clk_f + target_clk_f / 50; - if (max_vco_f < target_vco_f) - max_vco_f = target_vco_f; - /* min_pl <= high_pl <= max_pl */ high_pl = (max_vco_f + target_vco_f - 1) / target_vco_f; high_pl = min(high_pl, clk->params->max_pl); @@ -195,9 +142,7 @@ gk20a_pllg_calc_mnp(struct gk20a_clk *clk, unsigned long rate) target_vco_f = target_clk_f * clk->pl_to_div(pl); for (m = clk->params->min_m; m <= clk->params->max_m; m++) { - u32 u_f, vco_f; - - u_f = ref_clk_f / m; + u32 u_f = ref_clk_f / m; if (u_f < clk->params->min_u) break; @@ -211,6 +156,8 @@ gk20a_pllg_calc_mnp(struct gk20a_clk *clk, unsigned long rate) break; for (; n <= n2; n++) { + u32 vco_f; + if (n < clk->params->min_n) continue; if (n > clk->params->max_n) @@ -247,16 +194,16 @@ found_match: "no best match for target @ %dMHz on gpc_pll", target_clk_f / KHZ); - clk->pll.m = best_m; - clk->pll.n = best_n; - clk->pll.pl = best_pl; + pll->m = best_m; + pll->n = best_n; + pll->pl = best_pl; - target_freq = gk20a_pllg_calc_rate(clk); + target_freq = gk20a_pllg_calc_rate(clk, pll); nvkm_debug(subdev, - "actual target freq %d MHz, M %d, N %d, PL %d(div%d)\n", - target_freq / MHZ, clk->pll.m, clk->pll.n, clk->pll.pl, - clk->pl_to_div(clk->pll.pl)); + "actual target freq %d KHz, M %d, N %d, PL %d(div%d)\n", + target_freq / KHZ, pll->m, pll->n, pll->pl, + clk->pl_to_div(pll->pl)); return 0; } @@ -265,45 +212,36 @@ gk20a_pllg_slide(struct gk20a_clk *clk, u32 n) { struct nvkm_subdev *subdev = &clk->base.subdev; struct nvkm_device *device = subdev->device; - u32 val; - int ramp_timeout; + struct gk20a_pll pll; + int ret = 0; /* get old coefficients */ - val = nvkm_rd32(device, GPCPLL_COEFF); + gk20a_pllg_read_mnp(clk, &pll); /* do nothing if NDIV is the same */ - if (n == ((val >> GPCPLL_COEFF_N_SHIFT) & MASK(GPCPLL_COEFF_N_WIDTH))) + if (n == pll.n) return 0; - /* setup */ - nvkm_mask(device, GPCPLL_CFG2, 0xff << GPCPLL_CFG2_PLL_STEPA_SHIFT, - 0x2b << GPCPLL_CFG2_PLL_STEPA_SHIFT); - nvkm_mask(device, GPCPLL_CFG3, 0xff << GPCPLL_CFG3_PLL_STEPB_SHIFT, - 0xb << GPCPLL_CFG3_PLL_STEPB_SHIFT); - /* pll slowdown mode */ nvkm_mask(device, GPCPLL_NDIV_SLOWDOWN, BIT(GPCPLL_NDIV_SLOWDOWN_SLOWDOWN_USING_PLL_SHIFT), BIT(GPCPLL_NDIV_SLOWDOWN_SLOWDOWN_USING_PLL_SHIFT)); /* new ndiv ready for ramp */ - val = nvkm_rd32(device, GPCPLL_COEFF); - val &= ~(MASK(GPCPLL_COEFF_N_WIDTH) << GPCPLL_COEFF_N_SHIFT); - val |= (n & MASK(GPCPLL_COEFF_N_WIDTH)) << GPCPLL_COEFF_N_SHIFT; + pll.n = n; udelay(1); - nvkm_wr32(device, GPCPLL_COEFF, val); + gk20a_pllg_write_mnp(clk, &pll); /* dynamic ramp to new ndiv */ - val = nvkm_rd32(device, GPCPLL_NDIV_SLOWDOWN); - val |= 0x1 << GPCPLL_NDIV_SLOWDOWN_EN_DYNRAMP_SHIFT; udelay(1); - nvkm_wr32(device, GPCPLL_NDIV_SLOWDOWN, val); + nvkm_mask(device, GPCPLL_NDIV_SLOWDOWN, + BIT(GPCPLL_NDIV_SLOWDOWN_EN_DYNRAMP_SHIFT), + BIT(GPCPLL_NDIV_SLOWDOWN_EN_DYNRAMP_SHIFT)); - for (ramp_timeout = 500; ramp_timeout > 0; ramp_timeout--) { - udelay(1); - val = nvkm_rd32(device, GPC_BCAST_NDIV_SLOWDOWN_DEBUG); - if (val & GPC_BCAST_NDIV_SLOWDOWN_DEBUG_PLL_DYNRAMP_DONE_SYNCED_MASK) - break; - } + /* wait for ramping to complete */ + if (nvkm_wait_usec(device, 500, GPC_BCAST_NDIV_SLOWDOWN_DEBUG, + GPC_BCAST_NDIV_SLOWDOWN_DEBUG_PLL_DYNRAMP_DONE_SYNCED_MASK, + GPC_BCAST_NDIV_SLOWDOWN_DEBUG_PLL_DYNRAMP_DONE_SYNCED_MASK) < 0) + ret = -ETIMEDOUT; /* exit slowdown mode */ nvkm_mask(device, GPCPLL_NDIV_SLOWDOWN, @@ -311,21 +249,35 @@ gk20a_pllg_slide(struct gk20a_clk *clk, u32 n) BIT(GPCPLL_NDIV_SLOWDOWN_EN_DYNRAMP_SHIFT), 0); nvkm_rd32(device, GPCPLL_NDIV_SLOWDOWN); - if (ramp_timeout <= 0) { - nvkm_error(subdev, "gpcpll dynamic ramp timeout\n"); - return -ETIMEDOUT; - } - - return 0; + return ret; } -static void +static int gk20a_pllg_enable(struct gk20a_clk *clk) { struct nvkm_device *device = clk->base.subdev.device; + u32 val; nvkm_mask(device, GPCPLL_CFG, GPCPLL_CFG_ENABLE, GPCPLL_CFG_ENABLE); nvkm_rd32(device, GPCPLL_CFG); + + /* enable lock detection */ + val = nvkm_rd32(device, GPCPLL_CFG); + if (val & GPCPLL_CFG_LOCK_DET_OFF) { + val &= ~GPCPLL_CFG_LOCK_DET_OFF; + nvkm_wr32(device, GPCPLL_CFG, val); + } + + /* wait for lock */ + if (nvkm_wait_usec(device, 300, GPCPLL_CFG, GPCPLL_CFG_LOCK, + GPCPLL_CFG_LOCK) < 0) + return -ETIMEDOUT; + + /* switch to VCO mode */ + nvkm_mask(device, SEL_VCO, BIT(SEL_VCO_GPC2CLK_OUT_SHIFT), + BIT(SEL_VCO_GPC2CLK_OUT_SHIFT)); + + return 0; } static void @@ -333,117 +285,81 @@ gk20a_pllg_disable(struct gk20a_clk *clk) { struct nvkm_device *device = clk->base.subdev.device; + /* put PLL in bypass before disabling it */ + nvkm_mask(device, SEL_VCO, BIT(SEL_VCO_GPC2CLK_OUT_SHIFT), 0); + nvkm_mask(device, GPCPLL_CFG, GPCPLL_CFG_ENABLE, 0); nvkm_rd32(device, GPCPLL_CFG); } static int -_gk20a_pllg_program_mnp(struct gk20a_clk *clk, bool allow_slide) +gk20a_pllg_program_mnp(struct gk20a_clk *clk, const struct gk20a_pll *pll) { struct nvkm_subdev *subdev = &clk->base.subdev; struct nvkm_device *device = subdev->device; - u32 val, cfg; - struct gk20a_pll old_pll; - u32 n_lo; - - /* get old coefficients */ - gk20a_pllg_read_mnp(clk, &old_pll); - - /* do NDIV slide if there is no change in M and PL */ - cfg = nvkm_rd32(device, GPCPLL_CFG); - if (allow_slide && clk->pll.m == old_pll.m && - clk->pll.pl == old_pll.pl && (cfg & GPCPLL_CFG_ENABLE)) { - return gk20a_pllg_slide(clk, clk->pll.n); - } - - /* slide down to NDIV_LO */ - if (allow_slide && (cfg & GPCPLL_CFG_ENABLE)) { - int ret; - - n_lo = DIV_ROUND_UP(old_pll.m * clk->params->min_vco, - clk->parent_rate / KHZ); - ret = gk20a_pllg_slide(clk, n_lo); + struct gk20a_pll cur_pll; + int ret; - if (ret) - return ret; - } + gk20a_pllg_read_mnp(clk, &cur_pll); - /* split FO-to-bypass jump in halfs by setting out divider 1:2 */ + /* split VCO-to-bypass jump in half by setting out divider 1:2 */ nvkm_mask(device, GPC2CLK_OUT, GPC2CLK_OUT_VCODIV_MASK, - 0x2 << GPC2CLK_OUT_VCODIV_SHIFT); - - /* put PLL in bypass before programming it */ - val = nvkm_rd32(device, SEL_VCO); - val &= ~(BIT(SEL_VCO_GPC2CLK_OUT_SHIFT)); + GPC2CLK_OUT_VCODIV2 << GPC2CLK_OUT_VCODIV_SHIFT); + /* Intentional 2nd write to assure linear divider operation */ + nvkm_mask(device, GPC2CLK_OUT, GPC2CLK_OUT_VCODIV_MASK, + GPC2CLK_OUT_VCODIV2 << GPC2CLK_OUT_VCODIV_SHIFT); + nvkm_rd32(device, GPC2CLK_OUT); udelay(2); - nvkm_wr32(device, SEL_VCO, val); - - /* get out from IDDQ */ - val = nvkm_rd32(device, GPCPLL_CFG); - if (val & GPCPLL_CFG_IDDQ) { - val &= ~GPCPLL_CFG_IDDQ; - nvkm_wr32(device, GPCPLL_CFG, val); - nvkm_rd32(device, GPCPLL_CFG); - udelay(2); - } gk20a_pllg_disable(clk); - nvkm_debug(subdev, "%s: m=%d n=%d pl=%d\n", __func__, - clk->pll.m, clk->pll.n, clk->pll.pl); - - n_lo = DIV_ROUND_UP(clk->pll.m * clk->params->min_vco, - clk->parent_rate / KHZ); - val = clk->pll.m << GPCPLL_COEFF_M_SHIFT; - val |= (allow_slide ? n_lo : clk->pll.n) << GPCPLL_COEFF_N_SHIFT; - val |= clk->pll.pl << GPCPLL_COEFF_P_SHIFT; - nvkm_wr32(device, GPCPLL_COEFF, val); + gk20a_pllg_write_mnp(clk, pll); - gk20a_pllg_enable(clk); - - val = nvkm_rd32(device, GPCPLL_CFG); - if (val & GPCPLL_CFG_LOCK_DET_OFF) { - val &= ~GPCPLL_CFG_LOCK_DET_OFF; - nvkm_wr32(device, GPCPLL_CFG, val); - } - - if (nvkm_usec(device, 300, - if (nvkm_rd32(device, GPCPLL_CFG) & GPCPLL_CFG_LOCK) - break; - ) < 0) - return -ETIMEDOUT; - - /* switch to VCO mode */ - nvkm_mask(device, SEL_VCO, BIT(SEL_VCO_GPC2CLK_OUT_SHIFT), - BIT(SEL_VCO_GPC2CLK_OUT_SHIFT)); + ret = gk20a_pllg_enable(clk); + if (ret) + return ret; /* restore out divider 1:1 */ - val = nvkm_rd32(device, GPC2CLK_OUT); - if ((val & GPC2CLK_OUT_VCODIV_MASK) != - (GPC2CLK_OUT_VCODIV1 << GPC2CLK_OUT_VCODIV_SHIFT)) { - val &= ~GPC2CLK_OUT_VCODIV_MASK; - val |= GPC2CLK_OUT_VCODIV1 << GPC2CLK_OUT_VCODIV_SHIFT; - udelay(2); - nvkm_wr32(device, GPC2CLK_OUT, val); - /* Intentional 2nd write to assure linear divider operation */ - nvkm_wr32(device, GPC2CLK_OUT, val); - nvkm_rd32(device, GPC2CLK_OUT); - } + udelay(2); + nvkm_mask(device, GPC2CLK_OUT, GPC2CLK_OUT_VCODIV_MASK, + GPC2CLK_OUT_VCODIV1 << GPC2CLK_OUT_VCODIV_SHIFT); + /* Intentional 2nd write to assure linear divider operation */ + nvkm_mask(device, GPC2CLK_OUT, GPC2CLK_OUT_VCODIV_MASK, + GPC2CLK_OUT_VCODIV1 << GPC2CLK_OUT_VCODIV_SHIFT); + nvkm_rd32(device, GPC2CLK_OUT); - /* slide up to new NDIV */ - return allow_slide ? gk20a_pllg_slide(clk, clk->pll.n) : 0; + return 0; } static int -gk20a_pllg_program_mnp(struct gk20a_clk *clk) +gk20a_pllg_program_mnp_slide(struct gk20a_clk *clk, const struct gk20a_pll *pll) { - int err; + struct gk20a_pll cur_pll; + int ret; - err = _gk20a_pllg_program_mnp(clk, true); - if (err) - err = _gk20a_pllg_program_mnp(clk, false); + if (gk20a_pllg_is_enabled(clk)) { + gk20a_pllg_read_mnp(clk, &cur_pll); + + /* just do NDIV slide if there is no change to M and PL */ + if (pll->m == cur_pll.m && pll->pl == cur_pll.pl) + return gk20a_pllg_slide(clk, pll->n); + + /* slide down to current NDIV_LO */ + cur_pll.n = gk20a_pllg_n_lo(clk, &cur_pll); + ret = gk20a_pllg_slide(clk, cur_pll.n); + if (ret) + return ret; + } + + /* program MNP with the new clock parameters and new NDIV_LO */ + cur_pll = *pll; + cur_pll.n = gk20a_pllg_n_lo(clk, &cur_pll); + ret = gk20a_pllg_program_mnp(clk, &cur_pll); + if (ret) + return ret; - return err; + /* slide up to new NDIV */ + return gk20a_pllg_slide(clk, pll->n); } static struct nvkm_pstate @@ -546,13 +462,14 @@ gk20a_clk_read(struct nvkm_clk *base, enum nv_clk_src src) struct gk20a_clk *clk = gk20a_clk(base); struct nvkm_subdev *subdev = &clk->base.subdev; struct nvkm_device *device = subdev->device; + struct gk20a_pll pll; switch (src) { case nv_clk_src_crystal: return device->crystal; case nv_clk_src_gpc: - gk20a_pllg_read_mnp(clk, &clk->pll); - return gk20a_pllg_calc_rate(clk) / GK20A_CLK_GPC_MDIV; + gk20a_pllg_read_mnp(clk, &pll); + return gk20a_pllg_calc_rate(clk, &pll) / GK20A_CLK_GPC_MDIV; default: nvkm_error(subdev, "invalid clock source %d\n", src); return -EINVAL; @@ -565,15 +482,20 @@ gk20a_clk_calc(struct nvkm_clk *base, struct nvkm_cstate *cstate) struct gk20a_clk *clk = gk20a_clk(base); return gk20a_pllg_calc_mnp(clk, cstate->domain[nv_clk_src_gpc] * - GK20A_CLK_GPC_MDIV); + GK20A_CLK_GPC_MDIV, &clk->pll); } int gk20a_clk_prog(struct nvkm_clk *base) { struct gk20a_clk *clk = gk20a_clk(base); + int ret; + + ret = gk20a_pllg_program_mnp_slide(clk, &clk->pll); + if (ret) + ret = gk20a_pllg_program_mnp(clk, &clk->pll); - return gk20a_pllg_program_mnp(clk); + return ret; } void @@ -581,29 +503,62 @@ gk20a_clk_tidy(struct nvkm_clk *base) { } +int +gk20a_clk_setup_slide(struct gk20a_clk *clk) +{ + struct nvkm_subdev *subdev = &clk->base.subdev; + struct nvkm_device *device = subdev->device; + u32 step_a, step_b; + + switch (clk->parent_rate) { + case 12000000: + case 12800000: + case 13000000: + step_a = 0x2b; + step_b = 0x0b; + break; + case 19200000: + step_a = 0x12; + step_b = 0x08; + break; + case 38400000: + step_a = 0x04; + step_b = 0x05; + break; + default: + nvkm_error(subdev, "invalid parent clock rate %u KHz", + clk->parent_rate / KHZ); + return -EINVAL; + } + + nvkm_mask(device, GPCPLL_CFG2, 0xff << GPCPLL_CFG2_PLL_STEPA_SHIFT, + step_a << GPCPLL_CFG2_PLL_STEPA_SHIFT); + nvkm_mask(device, GPCPLL_CFG3, 0xff << GPCPLL_CFG3_PLL_STEPB_SHIFT, + step_b << GPCPLL_CFG3_PLL_STEPB_SHIFT); + + return 0; +} + void gk20a_clk_fini(struct nvkm_clk *base) { struct nvkm_device *device = base->subdev.device; struct gk20a_clk *clk = gk20a_clk(base); - u32 val; /* slide to VCO min */ - val = nvkm_rd32(device, GPCPLL_CFG); - if (val & GPCPLL_CFG_ENABLE) { + if (gk20a_pllg_is_enabled(clk)) { struct gk20a_pll pll; u32 n_lo; gk20a_pllg_read_mnp(clk, &pll); - n_lo = DIV_ROUND_UP(pll.m * clk->params->min_vco, - clk->parent_rate / KHZ); + n_lo = gk20a_pllg_n_lo(clk, &pll); gk20a_pllg_slide(clk, n_lo); } - /* put PLL in bypass before disabling it */ - nvkm_mask(device, SEL_VCO, BIT(SEL_VCO_GPC2CLK_OUT_SHIFT), 0); - gk20a_pllg_disable(clk); + + /* set IDDQ */ + nvkm_mask(device, GPCPLL_CFG, GPCPLL_CFG_IDDQ, 1); } static int @@ -614,9 +569,18 @@ gk20a_clk_init(struct nvkm_clk *base) struct nvkm_device *device = subdev->device; int ret; + /* get out from IDDQ */ + nvkm_mask(device, GPCPLL_CFG, GPCPLL_CFG_IDDQ, 0); + nvkm_rd32(device, GPCPLL_CFG); + udelay(5); + nvkm_mask(device, GPC2CLK_OUT, GPC2CLK_OUT_INIT_MASK, GPC2CLK_OUT_INIT_VAL); + ret = gk20a_clk_setup_slide(clk); + if (ret) + return ret; + /* Start with lowest frequency */ base->func->calc(base, &base->func->pstates[0].base); ret = base->func->prog(&clk->base); @@ -646,7 +610,7 @@ gk20a_clk = { }; int -_gk20a_clk_ctor(struct nvkm_device *device, int index, +gk20a_clk_ctor(struct nvkm_device *device, int index, const struct nvkm_clk_func *func, const struct gk20a_clk_pllg_params *params, struct gk20a_clk *clk) @@ -685,7 +649,7 @@ gk20a_clk_new(struct nvkm_device *device, int index, struct nvkm_clk **pclk) return -ENOMEM; *pclk = &clk->base; - ret = _gk20a_clk_ctor(device, index, &gk20a_clk, &gk20a_pllg_params, + ret = gk20a_clk_ctor(device, index, &gk20a_clk, &gk20a_pllg_params, clk); clk->pl_to_div = pl_to_div; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk20a.h b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk20a.h index 13c46740197d..0d1450972162 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk20a.h +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk20a.h @@ -24,9 +24,79 @@ #ifndef __NVKM_CLK_GK20A_H__ #define __NVKM_CLK_GK20A_H__ +#define KHZ (1000) +#define MHZ (KHZ * 1000) + +#define MASK(w) ((1 << (w)) - 1) + #define GK20A_CLK_GPC_MDIV 1000 #define SYS_GPCPLL_CFG_BASE 0x00137000 +#define GPCPLL_CFG (SYS_GPCPLL_CFG_BASE + 0) +#define GPCPLL_CFG_ENABLE BIT(0) +#define GPCPLL_CFG_IDDQ BIT(1) +#define GPCPLL_CFG_LOCK_DET_OFF BIT(4) +#define GPCPLL_CFG_LOCK BIT(17) + +#define GPCPLL_CFG2 (SYS_GPCPLL_CFG_BASE + 0xc) +#define GPCPLL_CFG2_SETUP2_SHIFT 16 +#define GPCPLL_CFG2_PLL_STEPA_SHIFT 24 + +#define GPCPLL_CFG3 (SYS_GPCPLL_CFG_BASE + 0x18) +#define GPCPLL_CFG3_VCO_CTRL_SHIFT 0 +#define GPCPLL_CFG3_VCO_CTRL_WIDTH 9 +#define GPCPLL_CFG3_VCO_CTRL_MASK \ + (MASK(GPCPLL_CFG3_VCO_CTRL_WIDTH) << GPCPLL_CFG3_VCO_CTRL_SHIFT) +#define GPCPLL_CFG3_PLL_STEPB_SHIFT 16 +#define GPCPLL_CFG3_PLL_STEPB_WIDTH 8 + +#define GPCPLL_COEFF (SYS_GPCPLL_CFG_BASE + 4) +#define GPCPLL_COEFF_M_SHIFT 0 +#define GPCPLL_COEFF_M_WIDTH 8 +#define GPCPLL_COEFF_N_SHIFT 8 +#define GPCPLL_COEFF_N_WIDTH 8 +#define GPCPLL_COEFF_N_MASK \ + (MASK(GPCPLL_COEFF_N_WIDTH) << GPCPLL_COEFF_N_SHIFT) +#define GPCPLL_COEFF_P_SHIFT 16 +#define GPCPLL_COEFF_P_WIDTH 6 + +#define GPCPLL_NDIV_SLOWDOWN (SYS_GPCPLL_CFG_BASE + 0x1c) +#define GPCPLL_NDIV_SLOWDOWN_NDIV_LO_SHIFT 0 +#define GPCPLL_NDIV_SLOWDOWN_NDIV_MID_SHIFT 8 +#define GPCPLL_NDIV_SLOWDOWN_STEP_SIZE_LO2MID_SHIFT 16 +#define GPCPLL_NDIV_SLOWDOWN_SLOWDOWN_USING_PLL_SHIFT 22 +#define GPCPLL_NDIV_SLOWDOWN_EN_DYNRAMP_SHIFT 31 + +#define GPC_BCAST_GPCPLL_CFG_BASE 0x00132800 +#define GPC_BCAST_NDIV_SLOWDOWN_DEBUG (GPC_BCAST_GPCPLL_CFG_BASE + 0xa0) +#define GPC_BCAST_NDIV_SLOWDOWN_DEBUG_PLL_DYNRAMP_DONE_SYNCED_SHIFT 24 +#define GPC_BCAST_NDIV_SLOWDOWN_DEBUG_PLL_DYNRAMP_DONE_SYNCED_MASK \ + (0x1 << GPC_BCAST_NDIV_SLOWDOWN_DEBUG_PLL_DYNRAMP_DONE_SYNCED_SHIFT) + +#define SEL_VCO (SYS_GPCPLL_CFG_BASE + 0x100) +#define SEL_VCO_GPC2CLK_OUT_SHIFT 0 + +#define GPC2CLK_OUT (SYS_GPCPLL_CFG_BASE + 0x250) +#define GPC2CLK_OUT_SDIV14_INDIV4_WIDTH 1 +#define GPC2CLK_OUT_SDIV14_INDIV4_SHIFT 31 +#define GPC2CLK_OUT_SDIV14_INDIV4_MODE 1 +#define GPC2CLK_OUT_VCODIV_WIDTH 6 +#define GPC2CLK_OUT_VCODIV_SHIFT 8 +#define GPC2CLK_OUT_VCODIV1 0 +#define GPC2CLK_OUT_VCODIV2 2 +#define GPC2CLK_OUT_VCODIV_MASK (MASK(GPC2CLK_OUT_VCODIV_WIDTH) << \ + GPC2CLK_OUT_VCODIV_SHIFT) +#define GPC2CLK_OUT_BYPDIV_WIDTH 6 +#define GPC2CLK_OUT_BYPDIV_SHIFT 0 +#define GPC2CLK_OUT_BYPDIV31 0x3c +#define GPC2CLK_OUT_INIT_MASK ((MASK(GPC2CLK_OUT_SDIV14_INDIV4_WIDTH) << \ + GPC2CLK_OUT_SDIV14_INDIV4_SHIFT)\ + | (MASK(GPC2CLK_OUT_VCODIV_WIDTH) << GPC2CLK_OUT_VCODIV_SHIFT)\ + | (MASK(GPC2CLK_OUT_BYPDIV_WIDTH) << GPC2CLK_OUT_BYPDIV_SHIFT)) +#define GPC2CLK_OUT_INIT_VAL ((GPC2CLK_OUT_SDIV14_INDIV4_MODE << \ + GPC2CLK_OUT_SDIV14_INDIV4_SHIFT) \ + | (GPC2CLK_OUT_VCODIV1 << GPC2CLK_OUT_VCODIV_SHIFT) \ + | (GPC2CLK_OUT_BYPDIV31 << GPC2CLK_OUT_BYPDIV_SHIFT)) /* All frequencies in Khz */ struct gk20a_clk_pllg_params { @@ -54,7 +124,29 @@ struct gk20a_clk { }; #define gk20a_clk(p) container_of((p), struct gk20a_clk, base) -int _gk20a_clk_ctor(struct nvkm_device *, int, const struct nvkm_clk_func *, +u32 gk20a_pllg_calc_rate(struct gk20a_clk *, struct gk20a_pll *); +int gk20a_pllg_calc_mnp(struct gk20a_clk *, unsigned long, struct gk20a_pll *); +void gk20a_pllg_read_mnp(struct gk20a_clk *, struct gk20a_pll *); +void gk20a_pllg_write_mnp(struct gk20a_clk *, const struct gk20a_pll *); + +static inline bool +gk20a_pllg_is_enabled(struct gk20a_clk *clk) +{ + struct nvkm_device *device = clk->base.subdev.device; + u32 val; + + val = nvkm_rd32(device, GPCPLL_CFG); + return val & GPCPLL_CFG_ENABLE; +} + +static inline u32 +gk20a_pllg_n_lo(struct gk20a_clk *clk, struct gk20a_pll *pll) +{ + return DIV_ROUND_UP(pll->m * clk->params->min_vco, + clk->parent_rate / KHZ); +} + +int gk20a_clk_ctor(struct nvkm_device *, int, const struct nvkm_clk_func *, const struct gk20a_clk_pllg_params *, struct gk20a_clk *); void gk20a_clk_fini(struct nvkm_clk *); int gk20a_clk_read(struct nvkm_clk *, enum nv_clk_src); @@ -62,4 +154,6 @@ int gk20a_clk_calc(struct nvkm_clk *, struct nvkm_cstate *); int gk20a_clk_prog(struct nvkm_clk *); void gk20a_clk_tidy(struct nvkm_clk *); +int gk20a_clk_setup_slide(struct gk20a_clk *); + #endif diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gm20b.c b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gm20b.c index 71b2bbb61973..b284e949f732 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gm20b.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gm20b.c @@ -21,20 +21,123 @@ */ #include <subdev/clk.h> +#include <subdev/volt.h> +#include <subdev/timer.h> #include <core/device.h> +#include <core/tegra.h> #include "priv.h" #include "gk20a.h" -#define KHZ (1000) -#define MHZ (KHZ * 1000) - -#define MASK(w) ((1 << w) - 1) +#define GPCPLL_CFG_SYNC_MODE BIT(2) #define BYPASSCTRL_SYS (SYS_GPCPLL_CFG_BASE + 0x340) #define BYPASSCTRL_SYS_GPCPLL_SHIFT 0 #define BYPASSCTRL_SYS_GPCPLL_WIDTH 1 +#define GPCPLL_CFG2_SDM_DIN_SHIFT 0 +#define GPCPLL_CFG2_SDM_DIN_WIDTH 8 +#define GPCPLL_CFG2_SDM_DIN_MASK \ + (MASK(GPCPLL_CFG2_SDM_DIN_WIDTH) << GPCPLL_CFG2_SDM_DIN_SHIFT) +#define GPCPLL_CFG2_SDM_DIN_NEW_SHIFT 8 +#define GPCPLL_CFG2_SDM_DIN_NEW_WIDTH 15 +#define GPCPLL_CFG2_SDM_DIN_NEW_MASK \ + (MASK(GPCPLL_CFG2_SDM_DIN_NEW_WIDTH) << GPCPLL_CFG2_SDM_DIN_NEW_SHIFT) +#define GPCPLL_CFG2_SETUP2_SHIFT 16 +#define GPCPLL_CFG2_PLL_STEPA_SHIFT 24 + +#define GPCPLL_DVFS0 (SYS_GPCPLL_CFG_BASE + 0x10) +#define GPCPLL_DVFS0_DFS_COEFF_SHIFT 0 +#define GPCPLL_DVFS0_DFS_COEFF_WIDTH 7 +#define GPCPLL_DVFS0_DFS_COEFF_MASK \ + (MASK(GPCPLL_DVFS0_DFS_COEFF_WIDTH) << GPCPLL_DVFS0_DFS_COEFF_SHIFT) +#define GPCPLL_DVFS0_DFS_DET_MAX_SHIFT 8 +#define GPCPLL_DVFS0_DFS_DET_MAX_WIDTH 7 +#define GPCPLL_DVFS0_DFS_DET_MAX_MASK \ + (MASK(GPCPLL_DVFS0_DFS_DET_MAX_WIDTH) << GPCPLL_DVFS0_DFS_DET_MAX_SHIFT) + +#define GPCPLL_DVFS1 (SYS_GPCPLL_CFG_BASE + 0x14) +#define GPCPLL_DVFS1_DFS_EXT_DET_SHIFT 0 +#define GPCPLL_DVFS1_DFS_EXT_DET_WIDTH 7 +#define GPCPLL_DVFS1_DFS_EXT_STRB_SHIFT 7 +#define GPCPLL_DVFS1_DFS_EXT_STRB_WIDTH 1 +#define GPCPLL_DVFS1_DFS_EXT_CAL_SHIFT 8 +#define GPCPLL_DVFS1_DFS_EXT_CAL_WIDTH 7 +#define GPCPLL_DVFS1_DFS_EXT_SEL_SHIFT 15 +#define GPCPLL_DVFS1_DFS_EXT_SEL_WIDTH 1 +#define GPCPLL_DVFS1_DFS_CTRL_SHIFT 16 +#define GPCPLL_DVFS1_DFS_CTRL_WIDTH 12 +#define GPCPLL_DVFS1_EN_SDM_SHIFT 28 +#define GPCPLL_DVFS1_EN_SDM_WIDTH 1 +#define GPCPLL_DVFS1_EN_SDM_BIT BIT(28) +#define GPCPLL_DVFS1_EN_DFS_SHIFT 29 +#define GPCPLL_DVFS1_EN_DFS_WIDTH 1 +#define GPCPLL_DVFS1_EN_DFS_BIT BIT(29) +#define GPCPLL_DVFS1_EN_DFS_CAL_SHIFT 30 +#define GPCPLL_DVFS1_EN_DFS_CAL_WIDTH 1 +#define GPCPLL_DVFS1_EN_DFS_CAL_BIT BIT(30) +#define GPCPLL_DVFS1_DFS_CAL_DONE_SHIFT 31 +#define GPCPLL_DVFS1_DFS_CAL_DONE_WIDTH 1 +#define GPCPLL_DVFS1_DFS_CAL_DONE_BIT BIT(31) + +#define GPC_BCAST_GPCPLL_DVFS2 (GPC_BCAST_GPCPLL_CFG_BASE + 0x20) +#define GPC_BCAST_GPCPLL_DVFS2_DFS_EXT_STROBE_BIT BIT(16) + +#define GPCPLL_CFG3_PLL_DFS_TESTOUT_SHIFT 24 +#define GPCPLL_CFG3_PLL_DFS_TESTOUT_WIDTH 7 + +#define DFS_DET_RANGE 6 /* -2^6 ... 2^6-1 */ +#define SDM_DIN_RANGE 12 /* -2^12 ... 2^12-1 */ + +struct gm20b_clk_dvfs_params { + s32 coeff_slope; + s32 coeff_offs; + u32 vco_ctrl; +}; + +static const struct gm20b_clk_dvfs_params gm20b_dvfs_params = { + .coeff_slope = -165230, + .coeff_offs = 214007, + .vco_ctrl = 0x7 << 3, +}; + +/* + * base.n is now the *integer* part of the N factor. + * sdm_din contains n's decimal part. + */ +struct gm20b_pll { + struct gk20a_pll base; + u32 sdm_din; +}; + +struct gm20b_clk_dvfs { + u32 dfs_coeff; + s32 dfs_det_max; + s32 dfs_ext_cal; +}; + +struct gm20b_clk { + /* currently applied parameters */ + struct gk20a_clk base; + struct gm20b_clk_dvfs dvfs; + u32 uv; + + /* new parameters to apply */ + struct gk20a_pll new_pll; + struct gm20b_clk_dvfs new_dvfs; + u32 new_uv; + + const struct gm20b_clk_dvfs_params *dvfs_params; + + /* fused parameters */ + s32 uvdet_slope; + s32 uvdet_offs; + + /* safe frequency we can use at minimum voltage */ + u32 safe_fmax_vmin; +}; +#define gm20b_clk(p) container_of((gk20a_clk(p)), struct gm20b_clk, base) + static u32 pl_to_div(u32 pl) { return pl; @@ -53,6 +156,484 @@ static const struct gk20a_clk_pllg_params gm20b_pllg_params = { .min_pl = 1, .max_pl = 31, }; +static void +gm20b_pllg_read_mnp(struct gm20b_clk *clk, struct gm20b_pll *pll) +{ + struct nvkm_subdev *subdev = &clk->base.base.subdev; + struct nvkm_device *device = subdev->device; + u32 val; + + gk20a_pllg_read_mnp(&clk->base, &pll->base); + val = nvkm_rd32(device, GPCPLL_CFG2); + pll->sdm_din = (val >> GPCPLL_CFG2_SDM_DIN_SHIFT) & + MASK(GPCPLL_CFG2_SDM_DIN_WIDTH); +} + +static void +gm20b_pllg_write_mnp(struct gm20b_clk *clk, const struct gm20b_pll *pll) +{ + struct nvkm_device *device = clk->base.base.subdev.device; + + nvkm_mask(device, GPCPLL_CFG2, GPCPLL_CFG2_SDM_DIN_MASK, + pll->sdm_din << GPCPLL_CFG2_SDM_DIN_SHIFT); + gk20a_pllg_write_mnp(&clk->base, &pll->base); +} + +/* + * Determine DFS_COEFF for the requested voltage. Always select external + * calibration override equal to the voltage, and set maximum detection + * limit "0" (to make sure that PLL output remains under F/V curve when + * voltage increases). + */ +static void +gm20b_dvfs_calc_det_coeff(struct gm20b_clk *clk, s32 uv, + struct gm20b_clk_dvfs *dvfs) +{ + struct nvkm_subdev *subdev = &clk->base.base.subdev; + const struct gm20b_clk_dvfs_params *p = clk->dvfs_params; + u32 coeff; + /* Work with mv as uv would likely trigger an overflow */ + s32 mv = DIV_ROUND_CLOSEST(uv, 1000); + + /* coeff = slope * voltage + offset */ + coeff = DIV_ROUND_CLOSEST(mv * p->coeff_slope, 1000) + p->coeff_offs; + coeff = DIV_ROUND_CLOSEST(coeff, 1000); + dvfs->dfs_coeff = min_t(u32, coeff, MASK(GPCPLL_DVFS0_DFS_COEFF_WIDTH)); + + dvfs->dfs_ext_cal = DIV_ROUND_CLOSEST(uv - clk->uvdet_offs, + clk->uvdet_slope); + /* should never happen */ + if (abs(dvfs->dfs_ext_cal) >= BIT(DFS_DET_RANGE)) + nvkm_error(subdev, "dfs_ext_cal overflow!\n"); + + dvfs->dfs_det_max = 0; + + nvkm_debug(subdev, "%s uv: %d coeff: %x, ext_cal: %d, det_max: %d\n", + __func__, uv, dvfs->dfs_coeff, dvfs->dfs_ext_cal, + dvfs->dfs_det_max); +} + +/* + * Solve equation for integer and fractional part of the effective NDIV: + * + * n_eff = n_int + 1/2 + (SDM_DIN / 2^(SDM_DIN_RANGE + 1)) + + * (DVFS_COEFF * DVFS_DET_DELTA) / 2^DFS_DET_RANGE + * + * The SDM_DIN LSB is finally shifted out, since it is not accessible by sw. + */ +static void +gm20b_dvfs_calc_ndiv(struct gm20b_clk *clk, u32 n_eff, u32 *n_int, u32 *sdm_din) +{ + struct nvkm_subdev *subdev = &clk->base.base.subdev; + const struct gk20a_clk_pllg_params *p = clk->base.params; + u32 n; + s32 det_delta; + u32 rem, rem_range; + + /* calculate current ext_cal and subtract previous one */ + det_delta = DIV_ROUND_CLOSEST(((s32)clk->uv) - clk->uvdet_offs, + clk->uvdet_slope); + det_delta -= clk->dvfs.dfs_ext_cal; + det_delta = min(det_delta, clk->dvfs.dfs_det_max); + det_delta *= clk->dvfs.dfs_coeff; + + /* integer part of n */ + n = (n_eff << DFS_DET_RANGE) - det_delta; + /* should never happen! */ + if (n <= 0) { + nvkm_error(subdev, "ndiv <= 0 - setting to 1...\n"); + n = 1 << DFS_DET_RANGE; + } + if (n >> DFS_DET_RANGE > p->max_n) { + nvkm_error(subdev, "ndiv > max_n - setting to max_n...\n"); + n = p->max_n << DFS_DET_RANGE; + } + *n_int = n >> DFS_DET_RANGE; + + /* fractional part of n */ + rem = ((u32)n) & MASK(DFS_DET_RANGE); + rem_range = SDM_DIN_RANGE + 1 - DFS_DET_RANGE; + /* subtract 2^SDM_DIN_RANGE to account for the 1/2 of the equation */ + rem = (rem << rem_range) - BIT(SDM_DIN_RANGE); + /* lose 8 LSB and clip - sdm_din only keeps the most significant byte */ + *sdm_din = (rem >> BITS_PER_BYTE) & MASK(GPCPLL_CFG2_SDM_DIN_WIDTH); + + nvkm_debug(subdev, "%s n_eff: %d, n_int: %d, sdm_din: %d\n", __func__, + n_eff, *n_int, *sdm_din); +} + +static int +gm20b_pllg_slide(struct gm20b_clk *clk, u32 n) +{ + struct nvkm_subdev *subdev = &clk->base.base.subdev; + struct nvkm_device *device = subdev->device; + struct gm20b_pll pll; + u32 n_int, sdm_din; + int ret = 0; + + /* calculate the new n_int/sdm_din for this n/uv */ + gm20b_dvfs_calc_ndiv(clk, n, &n_int, &sdm_din); + + /* get old coefficients */ + gm20b_pllg_read_mnp(clk, &pll); + /* do nothing if NDIV is the same */ + if (n_int == pll.base.n && sdm_din == pll.sdm_din) + return 0; + + /* pll slowdown mode */ + nvkm_mask(device, GPCPLL_NDIV_SLOWDOWN, + BIT(GPCPLL_NDIV_SLOWDOWN_SLOWDOWN_USING_PLL_SHIFT), + BIT(GPCPLL_NDIV_SLOWDOWN_SLOWDOWN_USING_PLL_SHIFT)); + + /* new ndiv ready for ramp */ + /* in DVFS mode SDM is updated via "new" field */ + nvkm_mask(device, GPCPLL_CFG2, GPCPLL_CFG2_SDM_DIN_NEW_MASK, + sdm_din << GPCPLL_CFG2_SDM_DIN_NEW_SHIFT); + pll.base.n = n_int; + udelay(1); + gk20a_pllg_write_mnp(&clk->base, &pll.base); + + /* dynamic ramp to new ndiv */ + udelay(1); + nvkm_mask(device, GPCPLL_NDIV_SLOWDOWN, + BIT(GPCPLL_NDIV_SLOWDOWN_EN_DYNRAMP_SHIFT), + BIT(GPCPLL_NDIV_SLOWDOWN_EN_DYNRAMP_SHIFT)); + + /* wait for ramping to complete */ + if (nvkm_wait_usec(device, 500, GPC_BCAST_NDIV_SLOWDOWN_DEBUG, + GPC_BCAST_NDIV_SLOWDOWN_DEBUG_PLL_DYNRAMP_DONE_SYNCED_MASK, + GPC_BCAST_NDIV_SLOWDOWN_DEBUG_PLL_DYNRAMP_DONE_SYNCED_MASK) < 0) + ret = -ETIMEDOUT; + + /* in DVFS mode complete SDM update */ + nvkm_mask(device, GPCPLL_CFG2, GPCPLL_CFG2_SDM_DIN_MASK, + sdm_din << GPCPLL_CFG2_SDM_DIN_SHIFT); + + /* exit slowdown mode */ + nvkm_mask(device, GPCPLL_NDIV_SLOWDOWN, + BIT(GPCPLL_NDIV_SLOWDOWN_SLOWDOWN_USING_PLL_SHIFT) | + BIT(GPCPLL_NDIV_SLOWDOWN_EN_DYNRAMP_SHIFT), 0); + nvkm_rd32(device, GPCPLL_NDIV_SLOWDOWN); + + return ret; +} + +static int +gm20b_pllg_enable(struct gm20b_clk *clk) +{ + struct nvkm_device *device = clk->base.base.subdev.device; + + nvkm_mask(device, GPCPLL_CFG, GPCPLL_CFG_ENABLE, GPCPLL_CFG_ENABLE); + nvkm_rd32(device, GPCPLL_CFG); + + /* In DVFS mode lock cannot be used - so just delay */ + udelay(40); + + /* set SYNC_MODE for glitchless switch out of bypass */ + nvkm_mask(device, GPCPLL_CFG, GPCPLL_CFG_SYNC_MODE, + GPCPLL_CFG_SYNC_MODE); + nvkm_rd32(device, GPCPLL_CFG); + + /* switch to VCO mode */ + nvkm_mask(device, SEL_VCO, BIT(SEL_VCO_GPC2CLK_OUT_SHIFT), + BIT(SEL_VCO_GPC2CLK_OUT_SHIFT)); + + return 0; +} + +static void +gm20b_pllg_disable(struct gm20b_clk *clk) +{ + struct nvkm_device *device = clk->base.base.subdev.device; + + /* put PLL in bypass before disabling it */ + nvkm_mask(device, SEL_VCO, BIT(SEL_VCO_GPC2CLK_OUT_SHIFT), 0); + + /* clear SYNC_MODE before disabling PLL */ + nvkm_mask(device, GPCPLL_CFG, GPCPLL_CFG_SYNC_MODE, 0); + + nvkm_mask(device, GPCPLL_CFG, GPCPLL_CFG_ENABLE, 0); + nvkm_rd32(device, GPCPLL_CFG); +} + +static int +gm20b_pllg_program_mnp(struct gm20b_clk *clk, const struct gk20a_pll *pll) +{ + struct nvkm_subdev *subdev = &clk->base.base.subdev; + struct nvkm_device *device = subdev->device; + struct gm20b_pll cur_pll; + u32 n_int, sdm_din; + /* if we only change pdiv, we can do a glitchless transition */ + bool pdiv_only; + int ret; + + gm20b_dvfs_calc_ndiv(clk, pll->n, &n_int, &sdm_din); + gm20b_pllg_read_mnp(clk, &cur_pll); + pdiv_only = cur_pll.base.n == n_int && cur_pll.sdm_din == sdm_din && + cur_pll.base.m == pll->m; + + /* need full sequence if clock not enabled yet */ + if (!gk20a_pllg_is_enabled(&clk->base)) + pdiv_only = false; + + /* split VCO-to-bypass jump in half by setting out divider 1:2 */ + nvkm_mask(device, GPC2CLK_OUT, GPC2CLK_OUT_VCODIV_MASK, + GPC2CLK_OUT_VCODIV2 << GPC2CLK_OUT_VCODIV_SHIFT); + /* Intentional 2nd write to assure linear divider operation */ + nvkm_mask(device, GPC2CLK_OUT, GPC2CLK_OUT_VCODIV_MASK, + GPC2CLK_OUT_VCODIV2 << GPC2CLK_OUT_VCODIV_SHIFT); + nvkm_rd32(device, GPC2CLK_OUT); + udelay(2); + + if (pdiv_only) { + u32 old = cur_pll.base.pl; + u32 new = pll->pl; + + /* + * we can do a glitchless transition only if the old and new PL + * parameters share at least one bit set to 1. If this is not + * the case, calculate and program an interim PL that will allow + * us to respect that rule. + */ + if ((old & new) == 0) { + cur_pll.base.pl = min(old | BIT(ffs(new) - 1), + new | BIT(ffs(old) - 1)); + gk20a_pllg_write_mnp(&clk->base, &cur_pll.base); + } + + cur_pll.base.pl = new; + gk20a_pllg_write_mnp(&clk->base, &cur_pll.base); + } else { + /* disable before programming if more than pdiv changes */ + gm20b_pllg_disable(clk); + + cur_pll.base = *pll; + cur_pll.base.n = n_int; + cur_pll.sdm_din = sdm_din; + gm20b_pllg_write_mnp(clk, &cur_pll); + + ret = gm20b_pllg_enable(clk); + if (ret) + return ret; + } + + /* restore out divider 1:1 */ + udelay(2); + nvkm_mask(device, GPC2CLK_OUT, GPC2CLK_OUT_VCODIV_MASK, + GPC2CLK_OUT_VCODIV1 << GPC2CLK_OUT_VCODIV_SHIFT); + /* Intentional 2nd write to assure linear divider operation */ + nvkm_mask(device, GPC2CLK_OUT, GPC2CLK_OUT_VCODIV_MASK, + GPC2CLK_OUT_VCODIV1 << GPC2CLK_OUT_VCODIV_SHIFT); + nvkm_rd32(device, GPC2CLK_OUT); + + return 0; +} + +static int +gm20b_pllg_program_mnp_slide(struct gm20b_clk *clk, const struct gk20a_pll *pll) +{ + struct gk20a_pll cur_pll; + int ret; + + if (gk20a_pllg_is_enabled(&clk->base)) { + gk20a_pllg_read_mnp(&clk->base, &cur_pll); + + /* just do NDIV slide if there is no change to M and PL */ + if (pll->m == cur_pll.m && pll->pl == cur_pll.pl) + return gm20b_pllg_slide(clk, pll->n); + + /* slide down to current NDIV_LO */ + cur_pll.n = gk20a_pllg_n_lo(&clk->base, &cur_pll); + ret = gm20b_pllg_slide(clk, cur_pll.n); + if (ret) + return ret; + } + + /* program MNP with the new clock parameters and new NDIV_LO */ + cur_pll = *pll; + cur_pll.n = gk20a_pllg_n_lo(&clk->base, &cur_pll); + ret = gm20b_pllg_program_mnp(clk, &cur_pll); + if (ret) + return ret; + + /* slide up to new NDIV */ + return gm20b_pllg_slide(clk, pll->n); +} + +static int +gm20b_clk_calc(struct nvkm_clk *base, struct nvkm_cstate *cstate) +{ + struct gm20b_clk *clk = gm20b_clk(base); + struct nvkm_subdev *subdev = &base->subdev; + struct nvkm_volt *volt = base->subdev.device->volt; + int ret; + + ret = gk20a_pllg_calc_mnp(&clk->base, cstate->domain[nv_clk_src_gpc] * + GK20A_CLK_GPC_MDIV, &clk->new_pll); + if (ret) + return ret; + + clk->new_uv = volt->vid[cstate->voltage].uv; + gm20b_dvfs_calc_det_coeff(clk, clk->new_uv, &clk->new_dvfs); + + nvkm_debug(subdev, "%s uv: %d uv\n", __func__, clk->new_uv); + + return 0; +} + +/* + * Compute PLL parameters that are always safe for the current voltage + */ +static void +gm20b_dvfs_calc_safe_pll(struct gm20b_clk *clk, struct gk20a_pll *pll) +{ + u32 rate = gk20a_pllg_calc_rate(&clk->base, pll) / KHZ; + u32 parent_rate = clk->base.parent_rate / KHZ; + u32 nmin, nsafe; + + /* remove a safe margin of 10% */ + if (rate > clk->safe_fmax_vmin) + rate = rate * (100 - 10) / 100; + + /* gpc2clk */ + rate *= 2; + + nmin = DIV_ROUND_UP(pll->m * clk->base.params->min_vco, parent_rate); + nsafe = pll->m * rate / (clk->base.parent_rate); + + if (nsafe < nmin) { + pll->pl = DIV_ROUND_UP(nmin * parent_rate, pll->m * rate); + nsafe = nmin; + } + + pll->n = nsafe; +} + +static void +gm20b_dvfs_program_coeff(struct gm20b_clk *clk, u32 coeff) +{ + struct nvkm_device *device = clk->base.base.subdev.device; + + /* strobe to read external DFS coefficient */ + nvkm_mask(device, GPC_BCAST_GPCPLL_DVFS2, + GPC_BCAST_GPCPLL_DVFS2_DFS_EXT_STROBE_BIT, + GPC_BCAST_GPCPLL_DVFS2_DFS_EXT_STROBE_BIT); + + nvkm_mask(device, GPCPLL_DVFS0, GPCPLL_DVFS0_DFS_COEFF_MASK, + coeff << GPCPLL_DVFS0_DFS_COEFF_SHIFT); + + udelay(1); + nvkm_mask(device, GPC_BCAST_GPCPLL_DVFS2, + GPC_BCAST_GPCPLL_DVFS2_DFS_EXT_STROBE_BIT, 0); +} + +static void +gm20b_dvfs_program_ext_cal(struct gm20b_clk *clk, u32 dfs_det_cal) +{ + struct nvkm_device *device = clk->base.base.subdev.device; + u32 val; + + nvkm_mask(device, GPC_BCAST_GPCPLL_DVFS2, MASK(DFS_DET_RANGE + 1), + dfs_det_cal); + udelay(1); + + val = nvkm_rd32(device, GPCPLL_DVFS1); + if (!(val & BIT(25))) { + /* Use external value to overwrite calibration value */ + val |= BIT(25) | BIT(16); + nvkm_wr32(device, GPCPLL_DVFS1, val); + } +} + +static void +gm20b_dvfs_program_dfs_detection(struct gm20b_clk *clk, + struct gm20b_clk_dvfs *dvfs) +{ + struct nvkm_device *device = clk->base.base.subdev.device; + + /* strobe to read external DFS coefficient */ + nvkm_mask(device, GPC_BCAST_GPCPLL_DVFS2, + GPC_BCAST_GPCPLL_DVFS2_DFS_EXT_STROBE_BIT, + GPC_BCAST_GPCPLL_DVFS2_DFS_EXT_STROBE_BIT); + + nvkm_mask(device, GPCPLL_DVFS0, + GPCPLL_DVFS0_DFS_COEFF_MASK | GPCPLL_DVFS0_DFS_DET_MAX_MASK, + dvfs->dfs_coeff << GPCPLL_DVFS0_DFS_COEFF_SHIFT | + dvfs->dfs_det_max << GPCPLL_DVFS0_DFS_DET_MAX_SHIFT); + + udelay(1); + nvkm_mask(device, GPC_BCAST_GPCPLL_DVFS2, + GPC_BCAST_GPCPLL_DVFS2_DFS_EXT_STROBE_BIT, 0); + + gm20b_dvfs_program_ext_cal(clk, dvfs->dfs_ext_cal); +} + +static int +gm20b_clk_prog(struct nvkm_clk *base) +{ + struct gm20b_clk *clk = gm20b_clk(base); + u32 cur_freq; + int ret; + + /* No change in DVFS settings? */ + if (clk->uv == clk->new_uv) + goto prog; + + /* + * Interim step for changing DVFS detection settings: low enough + * frequency to be safe at at DVFS coeff = 0. + * + * 1. If voltage is increasing: + * - safe frequency target matches the lowest - old - frequency + * - DVFS settings are still old + * - Voltage already increased to new level by volt, but maximum + * detection limit assures PLL output remains under F/V curve + * + * 2. If voltage is decreasing: + * - safe frequency target matches the lowest - new - frequency + * - DVFS settings are still old + * - Voltage is also old, it will be lowered by volt afterwards + * + * Interim step can be skipped if old frequency is below safe minimum, + * i.e., it is low enough to be safe at any voltage in operating range + * with zero DVFS coefficient. + */ + cur_freq = nvkm_clk_read(&clk->base.base, nv_clk_src_gpc); + if (cur_freq > clk->safe_fmax_vmin) { + struct gk20a_pll pll_safe; + + if (clk->uv < clk->new_uv) + /* voltage will raise: safe frequency is current one */ + pll_safe = clk->base.pll; + else + /* voltage will drop: safe frequency is new one */ + pll_safe = clk->new_pll; + + gm20b_dvfs_calc_safe_pll(clk, &pll_safe); + ret = gm20b_pllg_program_mnp_slide(clk, &pll_safe); + if (ret) + return ret; + } + + /* + * DVFS detection settings transition: + * - Set DVFS coefficient zero + * - Set calibration level to new voltage + * - Set DVFS coefficient to match new voltage + */ + gm20b_dvfs_program_coeff(clk, 0); + gm20b_dvfs_program_ext_cal(clk, clk->new_dvfs.dfs_ext_cal); + gm20b_dvfs_program_coeff(clk, clk->new_dvfs.dfs_coeff); + gm20b_dvfs_program_dfs_detection(clk, &clk->new_dvfs); + +prog: + clk->uv = clk->new_uv; + clk->dvfs = clk->new_dvfs; + clk->base.pll = clk->new_pll; + + return gm20b_pllg_program_mnp_slide(clk, &clk->base.pll); +} + static struct nvkm_pstate gm20b_pstates[] = { { @@ -133,9 +714,99 @@ gm20b_pstates[] = { .voltage = 12, }, }, - }; +static void +gm20b_clk_fini(struct nvkm_clk *base) +{ + struct nvkm_device *device = base->subdev.device; + struct gm20b_clk *clk = gm20b_clk(base); + + /* slide to VCO min */ + if (gk20a_pllg_is_enabled(&clk->base)) { + struct gk20a_pll pll; + u32 n_lo; + + gk20a_pllg_read_mnp(&clk->base, &pll); + n_lo = gk20a_pllg_n_lo(&clk->base, &pll); + gm20b_pllg_slide(clk, n_lo); + } + + gm20b_pllg_disable(clk); + + /* set IDDQ */ + nvkm_mask(device, GPCPLL_CFG, GPCPLL_CFG_IDDQ, 1); +} + +static int +gm20b_clk_init_dvfs(struct gm20b_clk *clk) +{ + struct nvkm_subdev *subdev = &clk->base.base.subdev; + struct nvkm_device *device = subdev->device; + bool fused = clk->uvdet_offs && clk->uvdet_slope; + static const s32 ADC_SLOPE_UV = 10000; /* default ADC detection slope */ + u32 data; + int ret; + + /* Enable NA DVFS */ + nvkm_mask(device, GPCPLL_DVFS1, GPCPLL_DVFS1_EN_DFS_BIT, + GPCPLL_DVFS1_EN_DFS_BIT); + + /* Set VCO_CTRL */ + if (clk->dvfs_params->vco_ctrl) + nvkm_mask(device, GPCPLL_CFG3, GPCPLL_CFG3_VCO_CTRL_MASK, + clk->dvfs_params->vco_ctrl << GPCPLL_CFG3_VCO_CTRL_SHIFT); + + if (fused) { + /* Start internal calibration, but ignore results */ + nvkm_mask(device, GPCPLL_DVFS1, GPCPLL_DVFS1_EN_DFS_CAL_BIT, + GPCPLL_DVFS1_EN_DFS_CAL_BIT); + + /* got uvdev parameters from fuse, skip calibration */ + goto calibrated; + } + + /* + * If calibration parameters are not fused, start internal calibration, + * wait for completion, and use results along with default slope to + * calculate ADC offset during boot. + */ + nvkm_mask(device, GPCPLL_DVFS1, GPCPLL_DVFS1_EN_DFS_CAL_BIT, + GPCPLL_DVFS1_EN_DFS_CAL_BIT); + + /* Wait for internal calibration done (spec < 2us). */ + ret = nvkm_wait_usec(device, 10, GPCPLL_DVFS1, + GPCPLL_DVFS1_DFS_CAL_DONE_BIT, + GPCPLL_DVFS1_DFS_CAL_DONE_BIT); + if (ret < 0) { + nvkm_error(subdev, "GPCPLL calibration timeout\n"); + return -ETIMEDOUT; + } + + data = nvkm_rd32(device, GPCPLL_CFG3) >> + GPCPLL_CFG3_PLL_DFS_TESTOUT_SHIFT; + data &= MASK(GPCPLL_CFG3_PLL_DFS_TESTOUT_WIDTH); + + clk->uvdet_slope = ADC_SLOPE_UV; + clk->uvdet_offs = ((s32)clk->uv) - data * ADC_SLOPE_UV; + + nvkm_debug(subdev, "calibrated DVFS parameters: offs %d, slope %d\n", + clk->uvdet_offs, clk->uvdet_slope); + +calibrated: + /* Compute and apply initial DVFS parameters */ + gm20b_dvfs_calc_det_coeff(clk, clk->uv, &clk->dvfs); + gm20b_dvfs_program_coeff(clk, 0); + gm20b_dvfs_program_ext_cal(clk, clk->dvfs.dfs_ext_cal); + gm20b_dvfs_program_coeff(clk, clk->dvfs.dfs_coeff); + gm20b_dvfs_program_dfs_detection(clk, &clk->new_dvfs); + + return 0; +} + +/* Forward declaration to detect speedo >=1 in gm20b_clk_init() */ +static const struct nvkm_clk_func gm20b_clk; + static int gm20b_clk_init(struct nvkm_clk *base) { @@ -143,15 +814,56 @@ gm20b_clk_init(struct nvkm_clk *base) struct nvkm_subdev *subdev = &clk->base.subdev; struct nvkm_device *device = subdev->device; int ret; + u32 data; + + /* get out from IDDQ */ + nvkm_mask(device, GPCPLL_CFG, GPCPLL_CFG_IDDQ, 0); + nvkm_rd32(device, GPCPLL_CFG); + udelay(5); + + nvkm_mask(device, GPC2CLK_OUT, GPC2CLK_OUT_INIT_MASK, + GPC2CLK_OUT_INIT_VAL); /* Set the global bypass control to VCO */ nvkm_mask(device, BYPASSCTRL_SYS, MASK(BYPASSCTRL_SYS_GPCPLL_WIDTH) << BYPASSCTRL_SYS_GPCPLL_SHIFT, 0); + ret = gk20a_clk_setup_slide(clk); + if (ret) + return ret; + + /* If not fused, set RAM SVOP PDP data 0x2, and enable fuse override */ + data = nvkm_rd32(device, 0x021944); + if (!(data & 0x3)) { + data |= 0x2; + nvkm_wr32(device, 0x021944, data); + + data = nvkm_rd32(device, 0x021948); + data |= 0x1; + nvkm_wr32(device, 0x021948, data); + } + + /* Disable idle slow down */ + nvkm_mask(device, 0x20160, 0x003f0000, 0x0); + + /* speedo >= 1? */ + if (clk->base.func == &gm20b_clk) { + struct gm20b_clk *_clk = gm20b_clk(base); + struct nvkm_volt *volt = device->volt; + + /* Get current voltage */ + _clk->uv = nvkm_volt_get(volt); + + /* Initialize DVFS */ + ret = gm20b_clk_init_dvfs(_clk); + if (ret) + return ret; + } + /* Start with lowest frequency */ base->func->calc(base, &base->func->pstates[0].base); - ret = base->func->prog(&clk->base); + ret = base->func->prog(base); if (ret) { nvkm_error(subdev, "cannot initialize clock\n"); return ret; @@ -169,6 +881,7 @@ gm20b_clk_speedo0 = { .prog = gk20a_clk_prog, .tidy = gk20a_clk_tidy, .pstates = gm20b_pstates, + /* Speedo 0 only supports 12 voltages */ .nr_pstates = ARRAY_SIZE(gm20b_pstates) - 1, .domains = { { nv_clk_src_crystal, 0xff }, @@ -177,8 +890,26 @@ gm20b_clk_speedo0 = { }, }; -int -gm20b_clk_new(struct nvkm_device *device, int index, struct nvkm_clk **pclk) +static const struct nvkm_clk_func +gm20b_clk = { + .init = gm20b_clk_init, + .fini = gm20b_clk_fini, + .read = gk20a_clk_read, + .calc = gm20b_clk_calc, + .prog = gm20b_clk_prog, + .tidy = gk20a_clk_tidy, + .pstates = gm20b_pstates, + .nr_pstates = ARRAY_SIZE(gm20b_pstates), + .domains = { + { nv_clk_src_crystal, 0xff }, + { nv_clk_src_gpc, 0xff, 0, "core", GK20A_CLK_GPC_MDIV }, + { nv_clk_src_max }, + }, +}; + +static int +gm20b_clk_new_speedo0(struct nvkm_device *device, int index, + struct nvkm_clk **pclk) { struct gk20a_clk *clk; int ret; @@ -188,11 +919,156 @@ gm20b_clk_new(struct nvkm_device *device, int index, struct nvkm_clk **pclk) return -ENOMEM; *pclk = &clk->base; - ret = _gk20a_clk_ctor(device, index, &gm20b_clk_speedo0, - &gm20b_pllg_params, clk); + ret = gk20a_clk_ctor(device, index, &gm20b_clk_speedo0, + &gm20b_pllg_params, clk); clk->pl_to_div = pl_to_div; clk->div_to_pl = div_to_pl; return ret; } + +/* FUSE register */ +#define FUSE_RESERVED_CALIB0 0x204 +#define FUSE_RESERVED_CALIB0_INTERCEPT_FRAC_SHIFT 0 +#define FUSE_RESERVED_CALIB0_INTERCEPT_FRAC_WIDTH 4 +#define FUSE_RESERVED_CALIB0_INTERCEPT_INT_SHIFT 4 +#define FUSE_RESERVED_CALIB0_INTERCEPT_INT_WIDTH 10 +#define FUSE_RESERVED_CALIB0_SLOPE_FRAC_SHIFT 14 +#define FUSE_RESERVED_CALIB0_SLOPE_FRAC_WIDTH 10 +#define FUSE_RESERVED_CALIB0_SLOPE_INT_SHIFT 24 +#define FUSE_RESERVED_CALIB0_SLOPE_INT_WIDTH 6 +#define FUSE_RESERVED_CALIB0_FUSE_REV_SHIFT 30 +#define FUSE_RESERVED_CALIB0_FUSE_REV_WIDTH 2 + +static int +gm20b_clk_init_fused_params(struct gm20b_clk *clk) +{ + struct nvkm_subdev *subdev = &clk->base.base.subdev; + u32 val = 0; + u32 rev = 0; + +#if IS_ENABLED(CONFIG_ARCH_TEGRA) + tegra_fuse_readl(FUSE_RESERVED_CALIB0, &val); + rev = (val >> FUSE_RESERVED_CALIB0_FUSE_REV_SHIFT) & + MASK(FUSE_RESERVED_CALIB0_FUSE_REV_WIDTH); +#endif + + /* No fused parameters, we will calibrate later */ + if (rev == 0) + return -EINVAL; + + /* Integer part in mV + fractional part in uV */ + clk->uvdet_slope = ((val >> FUSE_RESERVED_CALIB0_SLOPE_INT_SHIFT) & + MASK(FUSE_RESERVED_CALIB0_SLOPE_INT_WIDTH)) * 1000 + + ((val >> FUSE_RESERVED_CALIB0_SLOPE_FRAC_SHIFT) & + MASK(FUSE_RESERVED_CALIB0_SLOPE_FRAC_WIDTH)); + + /* Integer part in mV + fractional part in 100uV */ + clk->uvdet_offs = ((val >> FUSE_RESERVED_CALIB0_INTERCEPT_INT_SHIFT) & + MASK(FUSE_RESERVED_CALIB0_INTERCEPT_INT_WIDTH)) * 1000 + + ((val >> FUSE_RESERVED_CALIB0_INTERCEPT_FRAC_SHIFT) & + MASK(FUSE_RESERVED_CALIB0_INTERCEPT_FRAC_WIDTH)) * 100; + + nvkm_debug(subdev, "fused calibration data: slope %d, offs %d\n", + clk->uvdet_slope, clk->uvdet_offs); + return 0; +} + +static int +gm20b_clk_init_safe_fmax(struct gm20b_clk *clk) +{ + struct nvkm_subdev *subdev = &clk->base.base.subdev; + struct nvkm_volt *volt = subdev->device->volt; + struct nvkm_pstate *pstates = clk->base.base.func->pstates; + int nr_pstates = clk->base.base.func->nr_pstates; + int vmin, id = 0; + u32 fmax = 0; + int i; + + /* find lowest voltage we can use */ + vmin = volt->vid[0].uv; + for (i = 1; i < volt->vid_nr; i++) { + if (volt->vid[i].uv <= vmin) { + vmin = volt->vid[i].uv; + id = volt->vid[i].vid; + } + } + + /* find max frequency at this voltage */ + for (i = 0; i < nr_pstates; i++) + if (pstates[i].base.voltage == id) + fmax = max(fmax, + pstates[i].base.domain[nv_clk_src_gpc]); + + if (!fmax) { + nvkm_error(subdev, "failed to evaluate safe fmax\n"); + return -EINVAL; + } + + /* we are safe at 90% of the max frequency */ + clk->safe_fmax_vmin = fmax * (100 - 10) / 100; + nvkm_debug(subdev, "safe fmax @ vmin = %u Khz\n", clk->safe_fmax_vmin); + + return 0; +} + +int +gm20b_clk_new(struct nvkm_device *device, int index, struct nvkm_clk **pclk) +{ + struct nvkm_device_tegra *tdev = device->func->tegra(device); + struct gm20b_clk *clk; + struct nvkm_subdev *subdev; + struct gk20a_clk_pllg_params *clk_params; + int ret; + + /* Speedo 0 GPUs cannot use noise-aware PLL */ + if (tdev->gpu_speedo_id == 0) + return gm20b_clk_new_speedo0(device, index, pclk); + + /* Speedo >= 1, use NAPLL */ + clk = kzalloc(sizeof(*clk) + sizeof(*clk_params), GFP_KERNEL); + if (!clk) + return -ENOMEM; + *pclk = &clk->base.base; + subdev = &clk->base.base.subdev; + + /* duplicate the clock parameters since we will patch them below */ + clk_params = (void *) (clk + 1); + *clk_params = gm20b_pllg_params; + ret = gk20a_clk_ctor(device, index, &gm20b_clk, clk_params, + &clk->base); + if (ret) + return ret; + + /* + * NAPLL can only work with max_u, clamp the m range so + * gk20a_pllg_calc_mnp always uses it + */ + clk_params->max_m = clk_params->min_m = DIV_ROUND_UP(clk_params->max_u, + (clk->base.parent_rate / KHZ)); + if (clk_params->max_m == 0) { + nvkm_warn(subdev, "cannot use NAPLL, using legacy clock...\n"); + kfree(clk); + return gm20b_clk_new_speedo0(device, index, pclk); + } + + clk->base.pl_to_div = pl_to_div; + clk->base.div_to_pl = div_to_pl; + + clk->dvfs_params = &gm20b_dvfs_params; + + ret = gm20b_clk_init_fused_params(clk); + /* + * we will calibrate during init - should never happen on + * prod parts + */ + if (ret) + nvkm_warn(subdev, "no fused calibration parameters\n"); + + ret = gm20b_clk_init_safe_fmax(clk); + if (ret) + return ret; + + return 0; +} diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/Kbuild b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/Kbuild index 842d5de96d73..edcc157e6ac8 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/Kbuild +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/Kbuild @@ -24,6 +24,8 @@ nvkm-y += nvkm/subdev/fb/gk104.o nvkm-y += nvkm/subdev/fb/gk20a.o nvkm-y += nvkm/subdev/fb/gm107.o nvkm-y += nvkm/subdev/fb/gm200.o +nvkm-y += nvkm/subdev/fb/gp100.o +nvkm-y += nvkm/subdev/fb/gp104.o nvkm-y += nvkm/subdev/fb/ram.o nvkm-y += nvkm/subdev/fb/ramnv04.o @@ -41,6 +43,7 @@ nvkm-y += nvkm/subdev/fb/rammcp77.o nvkm-y += nvkm/subdev/fb/ramgf100.o nvkm-y += nvkm/subdev/fb/ramgk104.o nvkm-y += nvkm/subdev/fb/ramgm107.o +nvkm-y += nvkm/subdev/fb/ramgp100.o nvkm-y += nvkm/subdev/fb/sddr2.o nvkm-y += nvkm/subdev/fb/sddr3.o nvkm-y += nvkm/subdev/fb/gddr3.o diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/base.c index ce90242b8cce..a7049c041594 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/base.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/base.c @@ -25,6 +25,7 @@ #include "ram.h" #include <core/memory.h> +#include <core/option.h> #include <subdev/bios.h> #include <subdev/bios/M0203.h> #include <engine/gr.h> @@ -134,6 +135,10 @@ nvkm_fb_init(struct nvkm_subdev *subdev) if (fb->func->init) fb->func->init(fb); + if (fb->func->init_page) + fb->func->init_page(fb); + if (fb->func->init_unkn) + fb->func->init_unkn(fb); return 0; } @@ -171,6 +176,7 @@ nvkm_fb_ctor(const struct nvkm_fb_func *func, struct nvkm_device *device, nvkm_subdev_ctor(&nvkm_fb, device, index, &fb->subdev); fb->func = func; fb->tile.regions = fb->func->tile.regions; + fb->page = nvkm_longopt(device->cfgopt, "NvFbBigPage", 0); } int diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gf100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gf100.c index e649ead5ccfc..76433cc66fff 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gf100.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gf100.c @@ -72,6 +72,22 @@ gf100_fb_oneinit(struct nvkm_fb *fb) } void +gf100_fb_init_page(struct nvkm_fb *fb) +{ + struct nvkm_device *device = fb->subdev.device; + switch (fb->page) { + case 16: + nvkm_mask(device, 0x100c80, 0x00000001, 0x00000001); + break; + case 17: + default: + nvkm_mask(device, 0x100c80, 0x00000001, 0x00000000); + fb->page = 17; + break; + } +} + +void gf100_fb_init(struct nvkm_fb *base) { struct gf100_fb *fb = gf100_fb(base); @@ -79,8 +95,6 @@ gf100_fb_init(struct nvkm_fb *base) if (fb->r100c10_page) nvkm_wr32(device, 0x100c10, fb->r100c10 >> 8); - - nvkm_mask(device, 0x100c80, 0x00000001, 0x00000000); /* 128KiB lpg */ } void * @@ -125,6 +139,7 @@ gf100_fb = { .dtor = gf100_fb_dtor, .oneinit = gf100_fb_oneinit, .init = gf100_fb_init, + .init_page = gf100_fb_init_page, .intr = gf100_fb_intr, .ram_new = gf100_ram_new, .memtype_valid = gf100_fb_memtype_valid, diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gf100.h b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gf100.h index 2160e5a39c9a..449f431644b3 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gf100.h +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gf100.h @@ -14,4 +14,6 @@ int gf100_fb_new_(const struct nvkm_fb_func *, struct nvkm_device *, void *gf100_fb_dtor(struct nvkm_fb *); void gf100_fb_init(struct nvkm_fb *); void gf100_fb_intr(struct nvkm_fb *); + +void gp100_fb_init(struct nvkm_fb *); #endif diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gk104.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gk104.c index b41f0f70038c..4245e2e6e604 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gk104.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gk104.c @@ -29,6 +29,7 @@ gk104_fb = { .dtor = gf100_fb_dtor, .oneinit = gf100_fb_oneinit, .init = gf100_fb_init, + .init_page = gf100_fb_init_page, .intr = gf100_fb_intr, .ram_new = gk104_ram_new, .memtype_valid = gf100_fb_memtype_valid, diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gk20a.c index 7306f7dfc3b9..f815fe2bbf08 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gk20a.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gk20a.c @@ -27,7 +27,6 @@ static void gk20a_fb_init(struct nvkm_fb *fb) { struct nvkm_device *device = fb->subdev.device; - nvkm_mask(device, 0x100c80, 0x00000001, 0x00000000); /* 128KiB lpg */ nvkm_wr32(device, 0x100cc8, nvkm_memory_addr(fb->mmu_wr) >> 8); nvkm_wr32(device, 0x100ccc, nvkm_memory_addr(fb->mmu_rd) >> 8); } @@ -36,6 +35,7 @@ static const struct nvkm_fb_func gk20a_fb = { .oneinit = gf100_fb_oneinit, .init = gk20a_fb_init, + .init_page = gf100_fb_init_page, .memtype_valid = gf100_fb_memtype_valid, }; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gm107.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gm107.c index 4869fdb753c9..db699025f546 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gm107.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gm107.c @@ -29,6 +29,7 @@ gm107_fb = { .dtor = gf100_fb_dtor, .oneinit = gf100_fb_oneinit, .init = gf100_fb_init, + .init_page = gf100_fb_init_page, .intr = gf100_fb_intr, .ram_new = gm107_ram_new, .memtype_valid = gf100_fb_memtype_valid, diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gm200.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gm200.c index 44f5716f64d8..62f653240be3 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gm200.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gm200.c @@ -26,6 +26,24 @@ #include <core/memory.h> +void +gm200_fb_init_page(struct nvkm_fb *fb) +{ + struct nvkm_device *device = fb->subdev.device; + switch (fb->page) { + case 16: + nvkm_mask(device, 0x100c80, 0x00000801, 0x00000001); + break; + case 17: + nvkm_mask(device, 0x100c80, 0x00000801, 0x00000000); + break; + default: + nvkm_mask(device, 0x100c80, 0x00000800, 0x00000800); + fb->page = 0; + break; + } +} + static void gm200_fb_init(struct nvkm_fb *base) { @@ -48,6 +66,7 @@ gm200_fb = { .dtor = gf100_fb_dtor, .oneinit = gf100_fb_oneinit, .init = gm200_fb_init, + .init_page = gm200_fb_init_page, .intr = gf100_fb_intr, .ram_new = gm107_ram_new, .memtype_valid = gf100_fb_memtype_valid, diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gp100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gp100.c new file mode 100644 index 000000000000..98474aec1921 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gp100.c @@ -0,0 +1,69 @@ +/* + * Copyright 2016 Red Hat Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Ben Skeggs <bskeggs@redhat.com> + */ +#include "gf100.h" +#include "ram.h" + +#include <core/memory.h> + +static void +gp100_fb_init_unkn(struct nvkm_fb *base) +{ + struct nvkm_device *device = gf100_fb(base)->base.subdev.device; + nvkm_wr32(device, 0x1fac80, nvkm_rd32(device, 0x100c80)); + nvkm_wr32(device, 0x1facc4, nvkm_rd32(device, 0x100cc4)); + nvkm_wr32(device, 0x1facc8, nvkm_rd32(device, 0x100cc8)); + nvkm_wr32(device, 0x1faccc, nvkm_rd32(device, 0x100ccc)); +} + +void +gp100_fb_init(struct nvkm_fb *base) +{ + struct gf100_fb *fb = gf100_fb(base); + struct nvkm_device *device = fb->base.subdev.device; + + if (fb->r100c10_page) + nvkm_wr32(device, 0x100c10, fb->r100c10 >> 8); + + nvkm_wr32(device, 0x100cc8, nvkm_memory_addr(fb->base.mmu_wr) >> 8); + nvkm_wr32(device, 0x100ccc, nvkm_memory_addr(fb->base.mmu_rd) >> 8); + nvkm_mask(device, 0x100cc4, 0x00060000, + max(nvkm_memory_size(fb->base.mmu_rd) >> 16, (u64)2) << 17); +} + +static const struct nvkm_fb_func +gp100_fb = { + .dtor = gf100_fb_dtor, + .oneinit = gf100_fb_oneinit, + .init = gp100_fb_init, + .init_page = gm200_fb_init_page, + .init_unkn = gp100_fb_init_unkn, + .ram_new = gp100_ram_new, + .memtype_valid = gf100_fb_memtype_valid, +}; + +int +gp100_fb_new(struct nvkm_device *device, int index, struct nvkm_fb **pfb) +{ + return gf100_fb_new_(&gp100_fb, device, index, pfb); +} diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gp104.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gp104.c new file mode 100644 index 000000000000..92cb71861bec --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/gp104.c @@ -0,0 +1,43 @@ +/* + * Copyright 2016 Red Hat Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Ben Skeggs <bskeggs@redhat.com> + */ +#include "gf100.h" +#include "ram.h" + +#include <core/memory.h> + +static const struct nvkm_fb_func +gp104_fb = { + .dtor = gf100_fb_dtor, + .oneinit = gf100_fb_oneinit, + .init = gp100_fb_init, + .init_page = gm200_fb_init_page, + .ram_new = gp100_ram_new, + .memtype_valid = gf100_fb_memtype_valid, +}; + +int +gp104_fb_new(struct nvkm_device *device, int index, struct nvkm_fb **pfb) +{ + return gf100_fb_new_(&gp104_fb, device, index, pfb); +} diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/priv.h b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/priv.h index d97d640e60a0..e905d44fa1d5 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/priv.h +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/priv.h @@ -8,6 +8,8 @@ struct nvkm_fb_func { void *(*dtor)(struct nvkm_fb *); int (*oneinit)(struct nvkm_fb *); void (*init)(struct nvkm_fb *); + void (*init_page)(struct nvkm_fb *); + void (*init_unkn)(struct nvkm_fb *); void (*intr)(struct nvkm_fb *); struct { @@ -60,5 +62,8 @@ void nv46_fb_tile_init(struct nvkm_fb *, int i, u32 addr, u32 size, u32 pitch, u32 flags, struct nvkm_fb_tile *); int gf100_fb_oneinit(struct nvkm_fb *); +void gf100_fb_init_page(struct nvkm_fb *); bool gf100_fb_memtype_valid(struct nvkm_fb *, u32); + +void gm200_fb_init_page(struct nvkm_fb *); #endif diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ram.h b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ram.h index f816cbf2ced3..b9ec0ae6723a 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ram.h +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ram.h @@ -47,4 +47,5 @@ int mcp77_ram_new(struct nvkm_fb *, struct nvkm_ram **); int gf100_ram_new(struct nvkm_fb *, struct nvkm_ram **); int gk104_ram_new(struct nvkm_fb *, struct nvkm_ram **); int gm107_ram_new(struct nvkm_fb *, struct nvkm_ram **); +int gp100_ram_new(struct nvkm_fb *, struct nvkm_ram **); #endif diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ramgp100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ramgp100.c new file mode 100644 index 000000000000..f3be408b5e5e --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/fb/ramgp100.c @@ -0,0 +1,146 @@ +/* + * Copyright 2013 Red Hat Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Ben Skeggs + */ +#include "ram.h" + +#include <subdev/bios.h> +#include <subdev/bios/init.h> +#include <subdev/bios/rammap.h> + +static int +gp100_ram_init(struct nvkm_ram *ram) +{ + struct nvkm_subdev *subdev = &ram->fb->subdev; + struct nvkm_device *device = subdev->device; + struct nvkm_bios *bios = device->bios; + u8 ver, hdr, cnt, len, snr, ssz; + u32 data; + int i; + + /* run a bunch of tables from rammap table. there's actually + * individual pointers for each rammap entry too, but, nvidia + * seem to just run the last two entries' scripts early on in + * their init, and never again.. we'll just run 'em all once + * for now. + * + * i strongly suspect that each script is for a separate mode + * (likely selected by 0x9a065c's lower bits?), and the + * binary driver skips the one that's already been setup by + * the init tables. + */ + data = nvbios_rammapTe(bios, &ver, &hdr, &cnt, &len, &snr, &ssz); + if (!data || hdr < 0x15) + return -EINVAL; + + cnt = nvbios_rd08(bios, data + 0x14); /* guess at count */ + data = nvbios_rd32(bios, data + 0x10); /* guess u32... */ + if (cnt) { + u32 save = nvkm_rd32(device, 0x9a065c) & 0x000000f0; + for (i = 0; i < cnt; i++, data += 4) { + if (i != save >> 4) { + nvkm_mask(device, 0x9a065c, 0x000000f0, i << 4); + nvbios_exec(&(struct nvbios_init) { + .subdev = subdev, + .bios = bios, + .offset = nvbios_rd32(bios, data), + .execute = 1, + }); + } + } + nvkm_mask(device, 0x9a065c, 0x000000f0, save); + } + + nvkm_mask(device, 0x9a0584, 0x11000000, 0x00000000); + nvkm_wr32(device, 0x10ecc0, 0xffffffff); + nvkm_mask(device, 0x9a0160, 0x00000010, 0x00000010); + return 0; +} + +static const struct nvkm_ram_func +gp100_ram_func = { + .init = gp100_ram_init, + .get = gf100_ram_get, + .put = gf100_ram_put, +}; + +int +gp100_ram_new(struct nvkm_fb *fb, struct nvkm_ram **pram) +{ + struct nvkm_ram *ram; + struct nvkm_subdev *subdev = &fb->subdev; + struct nvkm_device *device = subdev->device; + enum nvkm_ram_type type = nvkm_fb_bios_memtype(device->bios); + const u32 rsvd_head = ( 256 * 1024); /* vga memory */ + const u32 rsvd_tail = (1024 * 1024); /* vbios etc */ + u32 fbpa_num = nvkm_rd32(device, 0x022438), fbpa; + u32 fbio_opt = nvkm_rd32(device, 0x021c14); + u64 part, size = 0, comm = ~0ULL; + bool mixed = false; + int ret; + + nvkm_debug(subdev, "022438: %08x\n", fbpa_num); + nvkm_debug(subdev, "021c14: %08x\n", fbio_opt); + for (fbpa = 0; fbpa < fbpa_num; fbpa++) { + if (!(fbio_opt & (1 << fbpa))) { + part = nvkm_rd32(device, 0x90020c + (fbpa * 0x4000)); + nvkm_debug(subdev, "fbpa %02x: %lld MiB\n", fbpa, part); + part = part << 20; + if (part != comm) { + if (comm != ~0ULL) + mixed = true; + comm = min(comm, part); + } + size = size + part; + } + } + + ret = nvkm_ram_new_(&gp100_ram_func, fb, type, size, 0, &ram); + *pram = ram; + if (ret) + return ret; + + nvkm_mm_fini(&ram->vram); + + if (mixed) { + ret = nvkm_mm_init(&ram->vram, rsvd_head >> NVKM_RAM_MM_SHIFT, + ((comm * fbpa_num) - rsvd_head) >> + NVKM_RAM_MM_SHIFT, 1); + if (ret) + return ret; + + ret = nvkm_mm_init(&ram->vram, (0x1000000000ULL + comm) >> + NVKM_RAM_MM_SHIFT, + (size - (comm * fbpa_num) - rsvd_tail) >> + NVKM_RAM_MM_SHIFT, 1); + if (ret) + return ret; + } else { + ret = nvkm_mm_init(&ram->vram, rsvd_head >> NVKM_RAM_MM_SHIFT, + (size - rsvd_head - rsvd_tail) >> + NVKM_RAM_MM_SHIFT, 1); + if (ret) + return ret; + } + + return 0; +} diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/iccsense/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/iccsense/base.c index 323c79abe468..41bd5d0f7692 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/iccsense/base.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/iccsense/base.c @@ -276,6 +276,8 @@ nvkm_iccsense_oneinit(struct nvkm_subdev *subdev) struct pwr_rail_t *r = &stbl.rail[i]; struct nvkm_iccsense_rail *rail; struct nvkm_iccsense_sensor *sensor; + int (*read)(struct nvkm_iccsense *, + struct nvkm_iccsense_rail *); if (!r->mode || r->resistor_mohm == 0) continue; @@ -284,31 +286,31 @@ nvkm_iccsense_oneinit(struct nvkm_subdev *subdev) if (!sensor) continue; - rail = kmalloc(sizeof(*rail), GFP_KERNEL); - if (!rail) - return -ENOMEM; - switch (sensor->type) { case NVBIOS_EXTDEV_INA209: if (r->rail != 0) continue; - rail->read = nvkm_iccsense_ina209_read; + read = nvkm_iccsense_ina209_read; break; case NVBIOS_EXTDEV_INA219: if (r->rail != 0) continue; - rail->read = nvkm_iccsense_ina219_read; + read = nvkm_iccsense_ina219_read; break; case NVBIOS_EXTDEV_INA3221: if (r->rail >= 3) continue; - rail->read = nvkm_iccsense_ina3221_read; + read = nvkm_iccsense_ina3221_read; break; default: continue; } + rail = kmalloc(sizeof(*rail), GFP_KERNEL); + if (!rail) + return -ENOMEM; sensor->rail_mask |= 1 << r->rail; + rail->read = read; rail->sensor = sensor; rail->idx = r->rail; rail->mohm = r->resistor_mohm; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/Kbuild b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/Kbuild index 932b366598aa..12d6f4f102cb 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/Kbuild +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/Kbuild @@ -3,3 +3,4 @@ nvkm-y += nvkm/subdev/ltc/gf100.o nvkm-y += nvkm/subdev/ltc/gk104.o nvkm-y += nvkm/subdev/ltc/gm107.o nvkm-y += nvkm/subdev/ltc/gm200.o +nvkm-y += nvkm/subdev/ltc/gp100.o diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gf100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gf100.c index c9eb677967a8..4a0fa0a9b802 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gf100.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gf100.c @@ -23,7 +23,6 @@ */ #include "priv.h" -#include <core/enum.h> #include <subdev/fb.h> #include <subdev/timer.h> @@ -71,7 +70,7 @@ gf100_ltc_zbc_clear_depth(struct nvkm_ltc *ltc, int i, const u32 depth) nvkm_wr32(device, 0x17ea58, depth); } -static const struct nvkm_bitfield +const struct nvkm_bitfield gf100_ltc_lts_intr_name[] = { { 0x00000001, "IDLE_ERROR_IQ" }, { 0x00000002, "IDLE_ERROR_CBC" }, diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gm107.c b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gm107.c index e292f5679418..ec0a3844b2d1 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gm107.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gm107.c @@ -68,18 +68,22 @@ gm107_ltc_zbc_clear_depth(struct nvkm_ltc *ltc, int i, const u32 depth) nvkm_wr32(device, 0x17e34c, depth); } -static void -gm107_ltc_lts_isr(struct nvkm_ltc *ltc, int c, int s) +void +gm107_ltc_intr_lts(struct nvkm_ltc *ltc, int c, int s) { struct nvkm_subdev *subdev = <c->subdev; struct nvkm_device *device = subdev->device; - u32 base = 0x140000 + (c * 0x2000) + (s * 0x200); - u32 stat = nvkm_rd32(device, base + 0x00c); + u32 base = 0x140400 + (c * 0x2000) + (s * 0x200); + u32 intr = nvkm_rd32(device, base + 0x00c); + u16 stat = intr & 0x0000ffff; + char msg[128]; if (stat) { - nvkm_error(subdev, "LTC%d_LTS%d: %08x\n", c, s, stat); - nvkm_wr32(device, base + 0x00c, stat); + nvkm_snprintbf(msg, sizeof(msg), gf100_ltc_lts_intr_name, stat); + nvkm_error(subdev, "LTC%d_LTS%d: %08x [%s]\n", c, s, intr, msg); } + + nvkm_wr32(device, base + 0x00c, intr); } void @@ -92,7 +96,7 @@ gm107_ltc_intr(struct nvkm_ltc *ltc) while (mask) { u32 s, c = __ffs(mask); for (s = 0; s < ltc->lts_nr; s++) - gm107_ltc_lts_isr(ltc, c, s); + gm107_ltc_intr_lts(ltc, c, s); mask &= ~(1 << c); } } diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gm200.c b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gm200.c index 2a29bfd5125a..e18e0dc19ec8 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gm200.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gm200.c @@ -46,7 +46,7 @@ static const struct nvkm_ltc_func gm200_ltc = { .oneinit = gm200_ltc_oneinit, .init = gm200_ltc_init, - .intr = gm107_ltc_intr, /*XXX: not validated */ + .intr = gm107_ltc_intr, .cbc_clear = gm107_ltc_cbc_clear, .cbc_wait = gm107_ltc_cbc_wait, .zbc = 16, diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gp100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gp100.c new file mode 100644 index 000000000000..0bdfb2f40266 --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/gp100.c @@ -0,0 +1,75 @@ +/* + * Copyright 2016 Red Hat Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Ben Skeggs + */ +#include "priv.h" + +static void +gp100_ltc_intr(struct nvkm_ltc *ltc) +{ + struct nvkm_device *device = ltc->subdev.device; + u32 mask; + + mask = nvkm_rd32(device, 0x0001c0); + while (mask) { + u32 s, c = __ffs(mask); + for (s = 0; s < ltc->lts_nr; s++) + gm107_ltc_intr_lts(ltc, c, s); + mask &= ~(1 << c); + } +} + +static int +gp100_ltc_oneinit(struct nvkm_ltc *ltc) +{ + struct nvkm_device *device = ltc->subdev.device; + ltc->ltc_nr = nvkm_rd32(device, 0x12006c); + ltc->lts_nr = nvkm_rd32(device, 0x17e280) >> 28; + /*XXX: tagram allocation - TBD */ + return nvkm_mm_init(<c->tags, 0, 0, 1); +} + +static void +gp100_ltc_init(struct nvkm_ltc *ltc) +{ + /*XXX: PMU LS call to setup tagram address */ +} + +static const struct nvkm_ltc_func +gp100_ltc = { + .oneinit = gp100_ltc_oneinit, + .init = gp100_ltc_init, + .intr = gp100_ltc_intr, + .cbc_clear = gm107_ltc_cbc_clear, + .cbc_wait = gm107_ltc_cbc_wait, + .zbc = 16, + .zbc_clear_color = gm107_ltc_zbc_clear_color, + .zbc_clear_depth = gm107_ltc_zbc_clear_depth, + .invalidate = gf100_ltc_invalidate, + .flush = gf100_ltc_flush, +}; + +int +gp100_ltc_new(struct nvkm_device *device, int index, struct nvkm_ltc **pltc) +{ + return nvkm_ltc_new_(&gp100_ltc, device, index, pltc); +} diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/priv.h b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/priv.h index 6d81c695ed0d..8b95f96e3ffa 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/priv.h +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/ltc/priv.h @@ -2,6 +2,7 @@ #define __NVKM_LTC_PRIV_H__ #define nvkm_ltc(p) container_of((p), struct nvkm_ltc, subdev) #include <subdev/ltc.h> +#include <core/enum.h> int nvkm_ltc_new_(const struct nvkm_ltc_func *, struct nvkm_device *, int index, struct nvkm_ltc **); @@ -31,8 +32,10 @@ void gf100_ltc_zbc_clear_color(struct nvkm_ltc *, int, const u32[4]); void gf100_ltc_zbc_clear_depth(struct nvkm_ltc *, int, const u32); void gf100_ltc_invalidate(struct nvkm_ltc *); void gf100_ltc_flush(struct nvkm_ltc *); +extern const struct nvkm_bitfield gf100_ltc_lts_intr_name[]; void gm107_ltc_intr(struct nvkm_ltc *); +void gm107_ltc_intr_lts(struct nvkm_ltc *, int ltc, int lts); void gm107_ltc_cbc_clear(struct nvkm_ltc *, u32, u32); void gm107_ltc_cbc_wait(struct nvkm_ltc *); void gm107_ltc_zbc_clear_color(struct nvkm_ltc *, int, const u32[4]); diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/Kbuild b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/Kbuild index 49695ac7be2e..12943f92c206 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/Kbuild +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/Kbuild @@ -10,3 +10,4 @@ nvkm-y += nvkm/subdev/mc/gt215.o nvkm-y += nvkm/subdev/mc/gf100.o nvkm-y += nvkm/subdev/mc/gk104.o nvkm-y += nvkm/subdev/mc/gk20a.o +nvkm-y += nvkm/subdev/mc/gp100.o diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/base.c index 350a8caa84c8..6b25e25f9eba 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/base.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/base.c @@ -27,43 +27,67 @@ #include <subdev/top.h> void -nvkm_mc_unk260(struct nvkm_mc *mc, u32 data) +nvkm_mc_unk260(struct nvkm_device *device, u32 data) { - if (mc->func->unk260) + struct nvkm_mc *mc = device->mc; + if (likely(mc) && mc->func->unk260) mc->func->unk260(mc, data); } void -nvkm_mc_intr_unarm(struct nvkm_mc *mc) +nvkm_mc_intr_mask(struct nvkm_device *device, enum nvkm_devidx devidx, bool en) { - return mc->func->intr_unarm(mc); + struct nvkm_mc *mc = device->mc; + const struct nvkm_mc_map *map; + if (likely(mc) && mc->func->intr_mask) { + u32 mask = nvkm_top_intr_mask(device, devidx); + for (map = mc->func->intr; !mask && map->stat; map++) { + if (map->unit == devidx) + mask = map->stat; + } + mc->func->intr_mask(mc, mask, en ? mask : 0); + } +} + +void +nvkm_mc_intr_unarm(struct nvkm_device *device) +{ + struct nvkm_mc *mc = device->mc; + if (likely(mc)) + mc->func->intr_unarm(mc); } void -nvkm_mc_intr_rearm(struct nvkm_mc *mc) +nvkm_mc_intr_rearm(struct nvkm_device *device) { - return mc->func->intr_rearm(mc); + struct nvkm_mc *mc = device->mc; + if (likely(mc)) + mc->func->intr_rearm(mc); } static u32 -nvkm_mc_intr_mask(struct nvkm_mc *mc) +nvkm_mc_intr_stat(struct nvkm_mc *mc) { - u32 intr = mc->func->intr_mask(mc); + u32 intr = mc->func->intr_stat(mc); if (WARN_ON_ONCE(intr == 0xffffffff)) intr = 0; /* likely fallen off the bus */ return intr; } void -nvkm_mc_intr(struct nvkm_mc *mc, bool *handled) +nvkm_mc_intr(struct nvkm_device *device, bool *handled) { - struct nvkm_device *device = mc->subdev.device; + struct nvkm_mc *mc = device->mc; struct nvkm_subdev *subdev; - const struct nvkm_mc_map *map = mc->func->intr; - u32 stat, intr = nvkm_mc_intr_mask(mc); + const struct nvkm_mc_map *map; + u32 stat, intr; u64 subdevs; - stat = nvkm_top_intr(device->top, intr, &subdevs); + if (unlikely(!mc)) + return; + + intr = nvkm_mc_intr_stat(mc); + stat = nvkm_top_intr(device, intr, &subdevs); while (subdevs) { enum nvkm_devidx subidx = __ffs64(subdevs); subdev = nvkm_device_subdev(device, subidx); @@ -72,14 +96,13 @@ nvkm_mc_intr(struct nvkm_mc *mc, bool *handled) subdevs &= ~BIT_ULL(subidx); } - while (map->stat) { + for (map = mc->func->intr; map->stat; map++) { if (intr & map->stat) { subdev = nvkm_device_subdev(device, map->unit); if (subdev) nvkm_subdev_intr(subdev); stat &= ~map->stat; } - map++; } if (stat) @@ -87,22 +110,32 @@ nvkm_mc_intr(struct nvkm_mc *mc, bool *handled) *handled = intr != 0; } -static void -nvkm_mc_reset_(struct nvkm_mc *mc, enum nvkm_devidx devidx) +static u32 +nvkm_mc_reset_mask(struct nvkm_device *device, bool isauto, + enum nvkm_devidx devidx) { - struct nvkm_device *device = mc->subdev.device; + struct nvkm_mc *mc = device->mc; const struct nvkm_mc_map *map; - u64 pmc_enable; - - if (!(pmc_enable = nvkm_top_reset(device->top, devidx))) { - for (map = mc->func->reset; map && map->stat; map++) { - if (map->unit == devidx) { - pmc_enable = map->stat; - break; + u64 pmc_enable = 0; + if (likely(mc)) { + if (!(pmc_enable = nvkm_top_reset(device, devidx))) { + for (map = mc->func->reset; map && map->stat; map++) { + if (!isauto || !map->noauto) { + if (map->unit == devidx) { + pmc_enable = map->stat; + break; + } + } } } } + return pmc_enable; +} +void +nvkm_mc_reset(struct nvkm_device *device, enum nvkm_devidx devidx) +{ + u64 pmc_enable = nvkm_mc_reset_mask(device, true, devidx); if (pmc_enable) { nvkm_mask(device, 0x000200, pmc_enable, 0x00000000); nvkm_mask(device, 0x000200, pmc_enable, pmc_enable); @@ -111,17 +144,27 @@ nvkm_mc_reset_(struct nvkm_mc *mc, enum nvkm_devidx devidx) } void -nvkm_mc_reset(struct nvkm_mc *mc, enum nvkm_devidx devidx) +nvkm_mc_disable(struct nvkm_device *device, enum nvkm_devidx devidx) { - if (likely(mc)) - nvkm_mc_reset_(mc, devidx); + u64 pmc_enable = nvkm_mc_reset_mask(device, false, devidx); + if (pmc_enable) + nvkm_mask(device, 0x000200, pmc_enable, 0x00000000); +} + +void +nvkm_mc_enable(struct nvkm_device *device, enum nvkm_devidx devidx) +{ + u64 pmc_enable = nvkm_mc_reset_mask(device, false, devidx); + if (pmc_enable) { + nvkm_mask(device, 0x000200, pmc_enable, pmc_enable); + nvkm_rd32(device, 0x000200); + } } static int nvkm_mc_fini(struct nvkm_subdev *subdev, bool suspend) { - struct nvkm_mc *mc = nvkm_mc(subdev); - nvkm_mc_intr_unarm(mc); + nvkm_mc_intr_unarm(subdev->device); return 0; } @@ -131,7 +174,7 @@ nvkm_mc_init(struct nvkm_subdev *subdev) struct nvkm_mc *mc = nvkm_mc(subdev); if (mc->func->init) mc->func->init(mc); - nvkm_mc_intr_rearm(mc); + nvkm_mc_intr_rearm(subdev->device); return 0; } @@ -148,16 +191,21 @@ nvkm_mc = { .fini = nvkm_mc_fini, }; +void +nvkm_mc_ctor(const struct nvkm_mc_func *func, struct nvkm_device *device, + int index, struct nvkm_mc *mc) +{ + nvkm_subdev_ctor(&nvkm_mc, device, index, &mc->subdev); + mc->func = func; +} + int nvkm_mc_new_(const struct nvkm_mc_func *func, struct nvkm_device *device, int index, struct nvkm_mc **pmc) { struct nvkm_mc *mc; - if (!(mc = *pmc = kzalloc(sizeof(*mc), GFP_KERNEL))) return -ENOMEM; - - nvkm_subdev_ctor(&nvkm_mc, device, index, &mc->subdev); - mc->func = func; + nvkm_mc_ctor(func, device, index, *pmc); return 0; } diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/g84.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/g84.c index 5c85b47f071d..c3d66ef5dc12 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/g84.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/g84.c @@ -57,7 +57,7 @@ g84_mc = { .intr = g84_mc_intr, .intr_unarm = nv04_mc_intr_unarm, .intr_rearm = nv04_mc_intr_rearm, - .intr_mask = nv04_mc_intr_mask, + .intr_stat = nv04_mc_intr_stat, .reset = g84_mc_reset, }; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/g98.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/g98.c index 0280b43cc10c..93ad4982ce5f 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/g98.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/g98.c @@ -57,7 +57,7 @@ g98_mc = { .intr = g98_mc_intr, .intr_unarm = nv04_mc_intr_unarm, .intr_rearm = nv04_mc_intr_rearm, - .intr_mask = nv04_mc_intr_mask, + .intr_stat = nv04_mc_intr_stat, .reset = g98_mc_reset, }; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gf100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gf100.c index 8397e223bd43..d2c4d6033abb 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gf100.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gf100.c @@ -76,7 +76,7 @@ gf100_mc_intr_rearm(struct nvkm_mc *mc) } u32 -gf100_mc_intr_mask(struct nvkm_mc *mc) +gf100_mc_intr_stat(struct nvkm_mc *mc) { struct nvkm_device *device = mc->subdev.device; u32 intr0 = nvkm_rd32(device, 0x000100); @@ -85,6 +85,14 @@ gf100_mc_intr_mask(struct nvkm_mc *mc) } void +gf100_mc_intr_mask(struct nvkm_mc *mc, u32 mask, u32 stat) +{ + struct nvkm_device *device = mc->subdev.device; + nvkm_mask(device, 0x000640, mask, stat); + nvkm_mask(device, 0x000644, mask, stat); +} + +void gf100_mc_unk260(struct nvkm_mc *mc, u32 data) { nvkm_wr32(mc->subdev.device, 0x000260, data); @@ -97,6 +105,7 @@ gf100_mc = { .intr_unarm = gf100_mc_intr_unarm, .intr_rearm = gf100_mc_intr_rearm, .intr_mask = gf100_mc_intr_mask, + .intr_stat = gf100_mc_intr_stat, .reset = gf100_mc_reset, .unk260 = gf100_mc_unk260, }; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gk104.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gk104.c index 317464212c7d..7b8c6ecad1a5 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gk104.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gk104.c @@ -26,6 +26,7 @@ const struct nvkm_mc_map gk104_mc_reset[] = { { 0x00000100, NVKM_ENGINE_FIFO }, + { 0x00002000, NVKM_SUBDEV_PMU, true }, {} }; @@ -53,6 +54,7 @@ gk104_mc = { .intr_unarm = gf100_mc_intr_unarm, .intr_rearm = gf100_mc_intr_rearm, .intr_mask = gf100_mc_intr_mask, + .intr_stat = gf100_mc_intr_stat, .reset = gk104_mc_reset, .unk260 = gf100_mc_unk260, }; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gk20a.c index 60b044f517ed..ca1bf3279dbe 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gk20a.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gk20a.c @@ -30,6 +30,7 @@ gk20a_mc = { .intr_unarm = gf100_mc_intr_unarm, .intr_rearm = gf100_mc_intr_rearm, .intr_mask = gf100_mc_intr_mask, + .intr_stat = gf100_mc_intr_stat, .reset = gk104_mc_reset, }; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gp100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gp100.c new file mode 100644 index 000000000000..4d22f4abd6de --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gp100.c @@ -0,0 +1,103 @@ +/* + * Copyright 2012 Red Hat Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Ben Skeggs + */ +#define gp100_mc(p) container_of((p), struct gp100_mc, base) +#include "priv.h" + +struct gp100_mc { + struct nvkm_mc base; + spinlock_t lock; + bool intr; + u32 mask; +}; + +static void +gp100_mc_intr_update(struct gp100_mc *mc) +{ + struct nvkm_device *device = mc->base.subdev.device; + u32 mask = mc->intr ? mc->mask : 0, i; + for (i = 0; i < 2; i++) { + nvkm_wr32(device, 0x000180 + (i * 0x04), ~mask); + nvkm_wr32(device, 0x000160 + (i * 0x04), mask); + } +} + +static void +gp100_mc_intr_unarm(struct nvkm_mc *base) +{ + struct gp100_mc *mc = gp100_mc(base); + unsigned long flags; + spin_lock_irqsave(&mc->lock, flags); + mc->intr = false; + gp100_mc_intr_update(mc); + spin_unlock_irqrestore(&mc->lock, flags); +} + +static void +gp100_mc_intr_rearm(struct nvkm_mc *base) +{ + struct gp100_mc *mc = gp100_mc(base); + unsigned long flags; + spin_lock_irqsave(&mc->lock, flags); + mc->intr = true; + gp100_mc_intr_update(mc); + spin_unlock_irqrestore(&mc->lock, flags); +} + +static void +gp100_mc_intr_mask(struct nvkm_mc *base, u32 mask, u32 intr) +{ + struct gp100_mc *mc = gp100_mc(base); + unsigned long flags; + spin_lock_irqsave(&mc->lock, flags); + mc->mask = (mc->mask & ~mask) | intr; + gp100_mc_intr_update(mc); + spin_unlock_irqrestore(&mc->lock, flags); +} + +static const struct nvkm_mc_func +gp100_mc = { + .init = nv50_mc_init, + .intr = gk104_mc_intr, + .intr_unarm = gp100_mc_intr_unarm, + .intr_rearm = gp100_mc_intr_rearm, + .intr_mask = gp100_mc_intr_mask, + .intr_stat = gf100_mc_intr_stat, + .reset = gk104_mc_reset, +}; + +int +gp100_mc_new(struct nvkm_device *device, int index, struct nvkm_mc **pmc) +{ + struct gp100_mc *mc; + + if (!(mc = kzalloc(sizeof(*mc), GFP_KERNEL))) + return -ENOMEM; + nvkm_mc_ctor(&gp100_mc, device, index, &mc->base); + *pmc = &mc->base; + + spin_lock_init(&mc->lock); + mc->intr = false; + mc->mask = 0x7fffffff; + return 0; +} diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gt215.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gt215.c index aad0ba95bf18..99d50a3d956f 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gt215.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/gt215.c @@ -53,13 +53,20 @@ gt215_mc_intr[] = { {}, }; +static void +gt215_mc_intr_mask(struct nvkm_mc *mc, u32 mask, u32 stat) +{ + nvkm_mask(mc->subdev.device, 0x000640, mask, stat); +} + static const struct nvkm_mc_func gt215_mc = { .init = nv50_mc_init, .intr = gt215_mc_intr, .intr_unarm = nv04_mc_intr_unarm, .intr_rearm = nv04_mc_intr_rearm, - .intr_mask = nv04_mc_intr_mask, + .intr_mask = gt215_mc_intr_mask, + .intr_stat = nv04_mc_intr_stat, .reset = gt215_mc_reset, }; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv04.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv04.c index a062624e906b..6509defd1460 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv04.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv04.c @@ -56,7 +56,7 @@ nv04_mc_intr_rearm(struct nvkm_mc *mc) } u32 -nv04_mc_intr_mask(struct nvkm_mc *mc) +nv04_mc_intr_stat(struct nvkm_mc *mc) { return nvkm_rd32(mc->subdev.device, 0x000100); } @@ -75,7 +75,7 @@ nv04_mc = { .intr = nv04_mc_intr, .intr_unarm = nv04_mc_intr_unarm, .intr_rearm = nv04_mc_intr_rearm, - .intr_mask = nv04_mc_intr_mask, + .intr_stat = nv04_mc_intr_stat, .reset = nv04_mc_reset, }; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv11.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv11.c index 55f0b9166b52..9213107901e6 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv11.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv11.c @@ -39,7 +39,7 @@ nv11_mc = { .intr = nv11_mc_intr, .intr_unarm = nv04_mc_intr_unarm, .intr_rearm = nv04_mc_intr_rearm, - .intr_mask = nv04_mc_intr_mask, + .intr_stat = nv04_mc_intr_stat, .reset = nv04_mc_reset, }; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv17.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv17.c index c40fa67f79a5..64bf5bbf8146 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv17.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv17.c @@ -48,7 +48,7 @@ nv17_mc = { .intr = nv17_mc_intr, .intr_unarm = nv04_mc_intr_unarm, .intr_rearm = nv04_mc_intr_rearm, - .intr_mask = nv04_mc_intr_mask, + .intr_stat = nv04_mc_intr_stat, .reset = nv17_mc_reset, }; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv44.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv44.c index cc56271db564..65fa44a64b98 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv44.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv44.c @@ -43,7 +43,7 @@ nv44_mc = { .intr = nv17_mc_intr, .intr_unarm = nv04_mc_intr_unarm, .intr_rearm = nv04_mc_intr_rearm, - .intr_mask = nv04_mc_intr_mask, + .intr_stat = nv04_mc_intr_stat, .reset = nv17_mc_reset, }; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv50.c b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv50.c index 343b6078580d..fe93b4fd7100 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv50.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/nv50.c @@ -50,7 +50,7 @@ nv50_mc = { .intr = nv50_mc_intr, .intr_unarm = nv04_mc_intr_unarm, .intr_rearm = nv04_mc_intr_rearm, - .intr_mask = nv04_mc_intr_mask, + .intr_stat = nv04_mc_intr_stat, .reset = nv17_mc_reset, }; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/priv.h b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/priv.h index a12038118512..4f0576a06d24 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/mc/priv.h +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/mc/priv.h @@ -3,12 +3,15 @@ #define nvkm_mc(p) container_of((p), struct nvkm_mc, subdev) #include <subdev/mc.h> +void nvkm_mc_ctor(const struct nvkm_mc_func *, struct nvkm_device *, + int index, struct nvkm_mc *); int nvkm_mc_new_(const struct nvkm_mc_func *, struct nvkm_device *, int index, struct nvkm_mc **); struct nvkm_mc_map { u32 stat; u32 unit; + bool noauto; }; struct nvkm_mc_func { @@ -18,8 +21,10 @@ struct nvkm_mc_func { void (*intr_unarm)(struct nvkm_mc *); /* enable reporting of interrupts to host */ void (*intr_rearm)(struct nvkm_mc *); + /* (un)mask delivery of specific interrupts */ + void (*intr_mask)(struct nvkm_mc *, u32 mask, u32 stat); /* retrieve pending interrupt mask (NV_PMC_INTR) */ - u32 (*intr_mask)(struct nvkm_mc *); + u32 (*intr_stat)(struct nvkm_mc *); const struct nvkm_mc_map *reset; void (*unk260)(struct nvkm_mc *, u32); }; @@ -27,7 +32,7 @@ struct nvkm_mc_func { void nv04_mc_init(struct nvkm_mc *); void nv04_mc_intr_unarm(struct nvkm_mc *); void nv04_mc_intr_rearm(struct nvkm_mc *); -u32 nv04_mc_intr_mask(struct nvkm_mc *); +u32 nv04_mc_intr_stat(struct nvkm_mc *); extern const struct nvkm_mc_map nv04_mc_reset[]; extern const struct nvkm_mc_map nv17_mc_intr[]; @@ -39,7 +44,8 @@ void nv50_mc_init(struct nvkm_mc *); void gf100_mc_intr_unarm(struct nvkm_mc *); void gf100_mc_intr_rearm(struct nvkm_mc *); -u32 gf100_mc_intr_mask(struct nvkm_mc *); +void gf100_mc_intr_mask(struct nvkm_mc *, u32, u32); +u32 gf100_mc_intr_stat(struct nvkm_mc *); void gf100_mc_unk260(struct nvkm_mc *, u32); extern const struct nvkm_mc_map gk104_mc_intr[]; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/Kbuild b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/Kbuild index 3c2519fdeb81..2a31b7d66a6d 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/Kbuild +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/Kbuild @@ -10,3 +10,4 @@ nvkm-y += nvkm/subdev/pci/g94.o nvkm-y += nvkm/subdev/pci/gf100.o nvkm-y += nvkm/subdev/pci/gf106.o nvkm-y += nvkm/subdev/pci/gk104.o +nvkm-y += nvkm/subdev/pci/gp100.o diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/base.c index 6b0328bd7eed..eb9b278198b2 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/base.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/base.c @@ -69,15 +69,13 @@ static irqreturn_t nvkm_pci_intr(int irq, void *arg) { struct nvkm_pci *pci = arg; - struct nvkm_mc *mc = pci->subdev.device->mc; + struct nvkm_device *device = pci->subdev.device; bool handled = false; - if (likely(mc)) { - nvkm_mc_intr_unarm(mc); - if (pci->msi) - pci->func->msi_rearm(pci); - nvkm_mc_intr(mc, &handled); - nvkm_mc_intr_rearm(mc); - } + nvkm_mc_intr_unarm(device); + if (pci->msi) + pci->func->msi_rearm(pci); + nvkm_mc_intr(device, &handled); + nvkm_mc_intr_rearm(device); return handled ? IRQ_HANDLED : IRQ_NONE; } diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/gp100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/gp100.c new file mode 100644 index 000000000000..82c5234a06ff --- /dev/null +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/gp100.c @@ -0,0 +1,44 @@ +/* + * Copyright 2015 Red Hat Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: Ben Skeggs <bskeggs@redhat.com> + */ +#include "priv.h" + +static void +gp100_pci_msi_rearm(struct nvkm_pci *pci) +{ + nvkm_pci_wr32(pci, 0x0704, 0x00000000); +} + +static const struct nvkm_pci_func +gp100_pci_func = { + .rd32 = nv40_pci_rd32, + .wr08 = nv40_pci_wr08, + .wr32 = nv40_pci_wr32, + .msi_rearm = gp100_pci_msi_rearm, +}; + +int +gp100_pci_new(struct nvkm_device *device, int index, struct nvkm_pci **ppci) +{ + return nvkm_pci_new_(&gp100_pci_func, device, index, ppci); +} diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/secboot/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/secboot/base.c index 213fdba6cfa0..314be2192b7d 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/secboot/base.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/secboot/base.c @@ -19,8 +19,9 @@ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER * DEALINGS IN THE SOFTWARE. */ - #include "priv.h" + +#include <subdev/mc.h> #include <subdev/timer.h> static const char * @@ -70,12 +71,11 @@ nvkm_secboot_falcon_enable(struct nvkm_secboot *sb) int ret; /* enable engine */ - nvkm_mask(device, 0x200, sb->enable_mask, sb->enable_mask); - nvkm_rd32(device, 0x200); + nvkm_mc_enable(device, sb->devidx); ret = nvkm_wait_msec(device, 10, sb->base + 0x10c, 0x6, 0x0); if (ret < 0) { - nvkm_mask(device, 0x200, sb->enable_mask, 0x0); nvkm_error(&sb->subdev, "Falcon mem scrubbing timeout\n"); + nvkm_mc_disable(device, sb->devidx); return ret; } @@ -85,8 +85,7 @@ nvkm_secboot_falcon_enable(struct nvkm_secboot *sb) /* enable IRQs */ nvkm_wr32(device, sb->base + 0x010, 0xff); - nvkm_mask(device, 0x640, sb->irq_mask, sb->irq_mask); - nvkm_mask(device, 0x644, sb->irq_mask, sb->irq_mask); + nvkm_mc_intr_mask(device, sb->devidx, true); return 0; } @@ -97,14 +96,13 @@ nvkm_secboot_falcon_disable(struct nvkm_secboot *sb) struct nvkm_device *device = sb->subdev.device; /* disable IRQs and wait for any previous code to complete */ - nvkm_mask(device, 0x644, sb->irq_mask, 0x0); - nvkm_mask(device, 0x640, sb->irq_mask, 0x0); + nvkm_mc_intr_mask(device, sb->devidx, false); nvkm_wr32(device, sb->base + 0x014, 0xff); falcon_wait_idle(device, sb->base); /* disable engine */ - nvkm_mask(device, 0x200, sb->enable_mask, 0x0); + nvkm_mc_disable(device, sb->devidx); return 0; } @@ -216,14 +214,7 @@ nvkm_secboot_oneinit(struct nvkm_subdev *subdev) return ret; } - /* - * Build all blobs - the same blobs can be used to perform secure boot - * multiple times - */ - if (sb->func->prepare_blobs) - ret = sb->func->prepare_blobs(sb); - - return ret; + return 0; } static int @@ -270,9 +261,8 @@ nvkm_secboot_ctor(const struct nvkm_secboot_func *func, /* setup the performing falcon's base address and masks */ switch (func->boot_falcon) { case NVKM_SECBOOT_FALCON_PMU: + sb->devidx = NVKM_SUBDEV_PMU; sb->base = 0x10a000; - sb->irq_mask = 0x1000000; - sb->enable_mask = 0x2000; break; default: nvkm_error(&sb->subdev, "invalid secure boot falcon\n"); diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/secboot/gm200.c b/drivers/gpu/drm/nouveau/nvkm/subdev/secboot/gm200.c index cc100dc940ea..f1e2dc914366 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/secboot/gm200.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/secboot/gm200.c @@ -860,6 +860,8 @@ gm200_secboot_prepare_ls_blob(struct gm200_secboot *gsb) /* Write LS blob */ ret = ls_ucode_mgr_write_wpr(gsb, &mgr, gsb->ls_blob); + if (ret) + nvkm_gpuobj_del(&gsb->ls_blob); cleanup: ls_ucode_mgr_cleanup(&mgr); @@ -1023,29 +1025,34 @@ gm20x_secboot_prepare_blobs(struct gm200_secboot *gsb) int ret; /* Load and prepare the managed falcon's firmwares */ - ret = gm200_secboot_prepare_ls_blob(gsb); - if (ret) - return ret; + if (!gsb->ls_blob) { + ret = gm200_secboot_prepare_ls_blob(gsb); + if (ret) + return ret; + } /* Load the HS firmware that will load the LS firmwares */ - ret = gm200_secboot_prepare_hs_blob(gsb, "acr/ucode_load", - &gsb->acr_load_blob, - &gsb->acr_load_bl_desc, true); - if (ret) - return ret; + if (!gsb->acr_load_blob) { + ret = gm200_secboot_prepare_hs_blob(gsb, "acr/ucode_load", + &gsb->acr_load_blob, + &gsb->acr_load_bl_desc, true); + if (ret) + return ret; + } /* Load the HS firmware bootloader */ - ret = gm200_secboot_prepare_hsbl_blob(gsb); - if (ret) - return ret; + if (!gsb->hsbl_blob) { + ret = gm200_secboot_prepare_hsbl_blob(gsb); + if (ret) + return ret; + } return 0; } static int -gm200_secboot_prepare_blobs(struct nvkm_secboot *sb) +gm200_secboot_prepare_blobs(struct gm200_secboot *gsb) { - struct gm200_secboot *gsb = gm200_secboot(sb); int ret; ret = gm20x_secboot_prepare_blobs(gsb); @@ -1053,15 +1060,37 @@ gm200_secboot_prepare_blobs(struct nvkm_secboot *sb) return ret; /* dGPU only: load the HS firmware that unprotects the WPR region */ - ret = gm200_secboot_prepare_hs_blob(gsb, "acr/ucode_unload", - &gsb->acr_unload_blob, - &gsb->acr_unload_bl_desc, false); - if (ret) - return ret; + if (!gsb->acr_unload_blob) { + ret = gm200_secboot_prepare_hs_blob(gsb, "acr/ucode_unload", + &gsb->acr_unload_blob, + &gsb->acr_unload_bl_desc, false); + if (ret) + return ret; + } return 0; } +static int +gm200_secboot_blobs_ready(struct gm200_secboot *gsb) +{ + struct nvkm_subdev *subdev = &gsb->base.subdev; + int ret; + + /* firmware already loaded, nothing to do... */ + if (gsb->firmware_ok) + return 0; + + ret = gsb->func->prepare_blobs(gsb); + if (ret) { + nvkm_error(subdev, "failed to load secure firmware\n"); + return ret; + } + + gsb->firmware_ok = true; + + return 0; +} /* @@ -1234,6 +1263,11 @@ gm200_secboot_reset(struct nvkm_secboot *sb, enum nvkm_secboot_falcon falcon) struct gm200_secboot *gsb = gm200_secboot(sb); int ret; + /* Make sure all blobs are ready */ + ret = gm200_secboot_blobs_ready(gsb); + if (ret) + return ret; + /* * Dummy GM200 implementation: perform secure boot each time we are * called on FECS. Since only FECS and GPCCS are managed and started @@ -1373,7 +1407,6 @@ gm200_secboot = { .dtor = gm200_secboot_dtor, .init = gm200_secboot_init, .fini = gm200_secboot_fini, - .prepare_blobs = gm200_secboot_prepare_blobs, .reset = gm200_secboot_reset, .start = gm200_secboot_start, .managed_falcons = BIT(NVKM_SECBOOT_FALCON_FECS) | @@ -1415,6 +1448,7 @@ gm200_secboot_func = { .bl_desc_size = sizeof(struct gm200_flcn_bl_desc), .fixup_bl_desc = gm200_secboot_fixup_bl_desc, .fixup_hs_desc = gm200_secboot_fixup_hs_desc, + .prepare_blobs = gm200_secboot_prepare_blobs, }; int @@ -1487,3 +1521,19 @@ MODULE_FIRMWARE("nvidia/gm206/gr/sw_ctx.bin"); MODULE_FIRMWARE("nvidia/gm206/gr/sw_nonctx.bin"); MODULE_FIRMWARE("nvidia/gm206/gr/sw_bundle_init.bin"); MODULE_FIRMWARE("nvidia/gm206/gr/sw_method_init.bin"); + +MODULE_FIRMWARE("nvidia/gp100/acr/bl.bin"); +MODULE_FIRMWARE("nvidia/gp100/acr/ucode_load.bin"); +MODULE_FIRMWARE("nvidia/gp100/acr/ucode_unload.bin"); +MODULE_FIRMWARE("nvidia/gp100/gr/fecs_bl.bin"); +MODULE_FIRMWARE("nvidia/gp100/gr/fecs_inst.bin"); +MODULE_FIRMWARE("nvidia/gp100/gr/fecs_data.bin"); +MODULE_FIRMWARE("nvidia/gp100/gr/fecs_sig.bin"); +MODULE_FIRMWARE("nvidia/gp100/gr/gpccs_bl.bin"); +MODULE_FIRMWARE("nvidia/gp100/gr/gpccs_inst.bin"); +MODULE_FIRMWARE("nvidia/gp100/gr/gpccs_data.bin"); +MODULE_FIRMWARE("nvidia/gp100/gr/gpccs_sig.bin"); +MODULE_FIRMWARE("nvidia/gp100/gr/sw_ctx.bin"); +MODULE_FIRMWARE("nvidia/gp100/gr/sw_nonctx.bin"); +MODULE_FIRMWARE("nvidia/gp100/gr/sw_bundle_init.bin"); +MODULE_FIRMWARE("nvidia/gp100/gr/sw_method_init.bin"); diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/secboot/gm20b.c b/drivers/gpu/drm/nouveau/nvkm/subdev/secboot/gm20b.c index 684320484b70..d5395ebfe8d3 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/secboot/gm20b.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/secboot/gm20b.c @@ -42,6 +42,32 @@ struct gm20b_flcn_bl_desc { u32 data_size; }; +static int +gm20b_secboot_prepare_blobs(struct gm200_secboot *gsb) +{ + struct nvkm_subdev *subdev = &gsb->base.subdev; + int acr_size; + int ret; + + ret = gm20x_secboot_prepare_blobs(gsb); + if (ret) + return ret; + + acr_size = gsb->acr_load_blob->size; + /* + * On Tegra the WPR region is set by the bootloader. It is illegal for + * the HS blob to be larger than this region. + */ + if (acr_size > gsb->wpr_size) { + nvkm_error(subdev, "WPR region too small for FW blob!\n"); + nvkm_error(subdev, "required: %dB\n", acr_size); + nvkm_error(subdev, "WPR size: %dB\n", gsb->wpr_size); + return -ENOSPC; + } + + return 0; +} + /** * gm20b_secboot_fixup_bl_desc - adapt BL descriptor to format used by GM20B FW * @@ -88,6 +114,7 @@ gm20b_secboot_func = { .bl_desc_size = sizeof(struct gm20b_flcn_bl_desc), .fixup_bl_desc = gm20b_secboot_fixup_bl_desc, .fixup_hs_desc = gm20b_secboot_fixup_hs_desc, + .prepare_blobs = gm20b_secboot_prepare_blobs, }; @@ -147,32 +174,6 @@ gm20b_tegra_read_wpr(struct gm200_secboot *gsb) #endif static int -gm20b_secboot_prepare_blobs(struct nvkm_secboot *sb) -{ - struct gm200_secboot *gsb = gm200_secboot(sb); - int acr_size; - int ret; - - ret = gm20x_secboot_prepare_blobs(gsb); - if (ret) - return ret; - - acr_size = gsb->acr_load_blob->size; - /* - * On Tegra the WPR region is set by the bootloader. It is illegal for - * the HS blob to be larger than this region. - */ - if (acr_size > gsb->wpr_size) { - nvkm_error(&sb->subdev, "WPR region too small for FW blob!\n"); - nvkm_error(&sb->subdev, "required: %dB\n", acr_size); - nvkm_error(&sb->subdev, "WPR size: %dB\n", gsb->wpr_size); - return -ENOSPC; - } - - return 0; -} - -static int gm20b_secboot_init(struct nvkm_secboot *sb) { struct gm200_secboot *gsb = gm200_secboot(sb); @@ -189,7 +190,6 @@ static const struct nvkm_secboot_func gm20b_secboot = { .dtor = gm200_secboot_dtor, .init = gm20b_secboot_init, - .prepare_blobs = gm20b_secboot_prepare_blobs, .reset = gm200_secboot_reset, .start = gm200_secboot_start, .managed_falcons = BIT(NVKM_SECBOOT_FALCON_FECS), diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/secboot/priv.h b/drivers/gpu/drm/nouveau/nvkm/subdev/secboot/priv.h index f2b09dee7c5d..a9a8a0e1017e 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/secboot/priv.h +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/secboot/priv.h @@ -30,7 +30,6 @@ struct nvkm_secboot_func { int (*init)(struct nvkm_secboot *); int (*fini)(struct nvkm_secboot *, bool suspend); void *(*dtor)(struct nvkm_secboot *); - int (*prepare_blobs)(struct nvkm_secboot *); int (*reset)(struct nvkm_secboot *, enum nvkm_secboot_falcon); int (*start)(struct nvkm_secboot *, enum nvkm_secboot_falcon); @@ -147,10 +146,8 @@ struct hsflcn_acr_desc { * @inst: instance block for HS falcon * @pgd: page directory for the HS falcon * @vm: address space used by the HS falcon - * @bl_desc_size: size of the BL descriptor used by this chip. - * @fixup_bl_desc: hook that generates the proper BL descriptor format from - * the generic GM200 format into a data array of size - * bl_desc_size + * @falcon_state: current state of the managed falcons + * @firmware_ok: whether the firmware blobs have been created */ struct gm200_secboot { struct nvkm_secboot base; @@ -196,9 +193,19 @@ struct gm200_secboot { RUNNING, } falcon_state[NVKM_SECBOOT_FALCON_END]; + bool firmware_ok; }; #define gm200_secboot(sb) container_of(sb, struct gm200_secboot, base) +/** + * Contains functions we wish to abstract between GM200-like implementations + * @bl_desc_size: size of the BL descriptor used by this chip. + * @fixup_bl_desc: hook that generates the proper BL descriptor format from + * the generic GM200 format into a data array of size + * bl_desc_size + * @fixup_hs_desc: hook that twiddles the HS descriptor before it is used + * @prepare_blobs: prepares the various blobs needed for secure booting + */ struct gm200_secboot_func { /* * Size of the bootloader descriptor for this chip. A block of this @@ -214,6 +221,7 @@ struct gm200_secboot_func { * we want the HS FW to set up. */ void (*fixup_hs_desc)(struct gm200_secboot *, struct hsflcn_acr_desc *); + int (*prepare_blobs)(struct gm200_secboot *); }; int gm200_secboot_init(struct nvkm_secboot *); diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/top/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/top/base.c index a1b264664aad..fe063d5728e2 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/top/base.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/top/base.c @@ -41,8 +41,9 @@ nvkm_top_device_new(struct nvkm_top *top) } u32 -nvkm_top_reset(struct nvkm_top *top, enum nvkm_devidx index) +nvkm_top_reset(struct nvkm_device *device, enum nvkm_devidx index) { + struct nvkm_top *top = device->top; struct nvkm_top_device *info; if (top) { @@ -56,8 +57,25 @@ nvkm_top_reset(struct nvkm_top *top, enum nvkm_devidx index) } u32 -nvkm_top_intr(struct nvkm_top *top, u32 intr, u64 *psubdevs) +nvkm_top_intr_mask(struct nvkm_device *device, enum nvkm_devidx devidx) { + struct nvkm_top *top = device->top; + struct nvkm_top_device *info; + + if (top) { + list_for_each_entry(info, &top->device, head) { + if (info->index == devidx && info->intr >= 0) + return BIT(info->intr); + } + } + + return 0; +} + +u32 +nvkm_top_intr(struct nvkm_device *device, u32 intr, u64 *psubdevs) +{ + struct nvkm_top *top = device->top; struct nvkm_top_device *info; u64 subdevs = 0; u32 handled = 0; @@ -78,8 +96,9 @@ nvkm_top_intr(struct nvkm_top *top, u32 intr, u64 *psubdevs) } enum nvkm_devidx -nvkm_top_fault(struct nvkm_top *top, int fault) +nvkm_top_fault(struct nvkm_device *device, int fault) { + struct nvkm_top *top = device->top; struct nvkm_top_device *info; list_for_each_entry(info, &top->device, head) { @@ -91,8 +110,9 @@ nvkm_top_fault(struct nvkm_top *top, int fault) } enum nvkm_devidx -nvkm_top_engine(struct nvkm_top *top, int index, int *runl, int *engn) +nvkm_top_engine(struct nvkm_device *device, int index, int *runl, int *engn) { + struct nvkm_top *top = device->top; struct nvkm_top_device *info; int n = 0; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/top/gk104.c b/drivers/gpu/drm/nouveau/nvkm/subdev/top/gk104.c index e06acc340e99..efac3402f9dd 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/top/gk104.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/top/gk104.c @@ -29,7 +29,7 @@ gk104_top_oneinit(struct nvkm_top *top) struct nvkm_subdev *subdev = &top->subdev; struct nvkm_device *device = subdev->device; struct nvkm_top_device *info = NULL; - u32 data, type; + u32 data, type, inst; int i; for (i = 0; i < 64; i++) { @@ -37,6 +37,7 @@ gk104_top_oneinit(struct nvkm_top *top) if (!(info = nvkm_top_device_new(top))) return -ENOMEM; type = ~0; + inst = 0; } data = nvkm_rd32(device, 0x022700 + (i * 0x04)); @@ -45,6 +46,7 @@ gk104_top_oneinit(struct nvkm_top *top) case 0x00000000: /* NOT_VALID */ continue; case 0x00000001: /* DATA */ + inst = (data & 0x3c000000) >> 26; info->addr = (data & 0x00fff000); info->fault = (data & 0x000000f8) >> 3; break; @@ -67,27 +69,32 @@ gk104_top_oneinit(struct nvkm_top *top) continue; /* Translate engine type to NVKM engine identifier. */ +#define A_(A) if (inst == 0) info->index = NVKM_ENGINE_##A +#define B_(A) if (inst + NVKM_ENGINE_##A##0 < NVKM_ENGINE_##A##_LAST + 1) \ + info->index = NVKM_ENGINE_##A##0 + inst switch (type) { - case 0x00000000: info->index = NVKM_ENGINE_GR; break; - case 0x00000001: info->index = NVKM_ENGINE_CE0; break; - case 0x00000002: info->index = NVKM_ENGINE_CE1; break; - case 0x00000003: info->index = NVKM_ENGINE_CE2; break; - case 0x00000008: info->index = NVKM_ENGINE_MSPDEC; break; - case 0x00000009: info->index = NVKM_ENGINE_MSPPP; break; - case 0x0000000a: info->index = NVKM_ENGINE_MSVLD; break; - case 0x0000000b: info->index = NVKM_ENGINE_MSENC; break; - case 0x0000000c: info->index = NVKM_ENGINE_VIC; break; - case 0x0000000d: info->index = NVKM_ENGINE_SEC; break; - case 0x0000000e: info->index = NVKM_ENGINE_NVENC0; break; - case 0x0000000f: info->index = NVKM_ENGINE_NVENC1; break; - case 0x00000010: info->index = NVKM_ENGINE_NVDEC; break; + case 0x00000000: A_(GR ); break; + case 0x00000001: A_(CE0 ); break; + case 0x00000002: A_(CE1 ); break; + case 0x00000003: A_(CE2 ); break; + case 0x00000008: A_(MSPDEC); break; + case 0x00000009: A_(MSPPP ); break; + case 0x0000000a: A_(MSVLD ); break; + case 0x0000000b: A_(MSENC ); break; + case 0x0000000c: A_(VIC ); break; + case 0x0000000d: A_(SEC ); break; + case 0x0000000e: B_(NVENC ); break; + case 0x0000000f: A_(NVENC1); break; + case 0x00000010: A_(NVDEC ); break; + case 0x00000013: B_(CE ); break; break; default: break; } - nvkm_debug(subdev, "%02x (%8s): addr %06x fault %2d engine %2d " - "runlist %2d intr %2d reset %2d\n", type, + nvkm_debug(subdev, "%02x.%d (%8s): addr %06x fault %2d " + "engine %2d runlist %2d intr %2d " + "reset %2d\n", type, inst, info->index == NVKM_SUBDEV_NR ? NULL : nvkm_subdev_name[info->index], info->addr, info->fault, info->engine, info->runlist, diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/volt/base.c b/drivers/gpu/drm/nouveau/nvkm/subdev/volt/base.c index 6b2d7531a7ff..1c3d23b0e84a 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/volt/base.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/volt/base.c @@ -120,6 +120,8 @@ nvkm_volt_parse_bios(struct nvkm_bios *bios, struct nvkm_volt *volt) data = nvbios_volt_parse(bios, &ver, &hdr, &cnt, &len, &info); if (data && info.vidmask && info.base && info.step) { + volt->min_uv = info.min; + volt->max_uv = info.max; for (i = 0; i < info.vidmask + 1; i++) { if (info.base >= info.min && info.base <= info.max) { @@ -131,6 +133,8 @@ nvkm_volt_parse_bios(struct nvkm_bios *bios, struct nvkm_volt *volt) } volt->vid_mask = info.vidmask; } else if (data && info.vidmask) { + volt->min_uv = 0xffffffff; + volt->max_uv = 0; for (i = 0; i < cnt; i++) { data = nvbios_volt_entry_parse(bios, i, &ver, &hdr, &ivid); @@ -138,9 +142,14 @@ nvkm_volt_parse_bios(struct nvkm_bios *bios, struct nvkm_volt *volt) volt->vid[volt->vid_nr].uv = ivid.voltage; volt->vid[volt->vid_nr].vid = ivid.vid; volt->vid_nr++; + volt->min_uv = min(volt->min_uv, ivid.voltage); + volt->max_uv = max(volt->max_uv, ivid.voltage); } } volt->vid_mask = info.vidmask; + } else if (data && info.type == NVBIOS_VOLT_PWM) { + volt->min_uv = info.base; + volt->max_uv = info.base + info.pwm_range; } } @@ -181,8 +190,11 @@ nvkm_volt_ctor(const struct nvkm_volt_func *func, struct nvkm_device *device, volt->func = func; /* Assuming the non-bios device should build the voltage table later */ - if (bios) + if (bios) { nvkm_volt_parse_bios(bios, volt); + nvkm_debug(&volt->subdev, "min: %iuv max: %iuv\n", + volt->min_uv, volt->max_uv); + } if (volt->vid_nr) { for (i = 0; i < volt->vid_nr; i++) { diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gk20a.c index d554455326da..ce5d83cdc7cf 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gk20a.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gk20a.c @@ -77,18 +77,19 @@ gk20a_volt_get_cvb_t_voltage(int speedo, int temp, int s_scale, int t_scale, return mv; } -int +static int gk20a_volt_calc_voltage(const struct cvb_coef *coef, int speedo) { + static const int v_scale = 1000; int mv; mv = gk20a_volt_get_cvb_t_voltage(speedo, -10, 100, 10, coef); - mv = DIV_ROUND_UP(mv, 1000); + mv = DIV_ROUND_UP(mv, v_scale); return mv * 1000; } -int +static int gk20a_volt_vid_get(struct nvkm_volt *base) { struct gk20a_volt *volt = gk20a_volt(base); @@ -103,7 +104,7 @@ gk20a_volt_vid_get(struct nvkm_volt *base) return -EINVAL; } -int +static int gk20a_volt_vid_set(struct nvkm_volt *base, u8 vid) { struct gk20a_volt *volt = gk20a_volt(base); @@ -113,7 +114,7 @@ gk20a_volt_vid_set(struct nvkm_volt *base, u8 vid) return regulator_set_voltage(volt->vdd, volt->base.vid[vid].uv, 1200000); } -int +static int gk20a_volt_set_id(struct nvkm_volt *base, u8 id, int condition) { struct gk20a_volt *volt = gk20a_volt(base); @@ -143,9 +144,9 @@ gk20a_volt = { }; int -_gk20a_volt_ctor(struct nvkm_device *device, int index, - const struct cvb_coef *coefs, int nb_coefs, - struct gk20a_volt *volt) +gk20a_volt_ctor(struct nvkm_device *device, int index, + const struct cvb_coef *coefs, int nb_coefs, + int vmin, struct gk20a_volt *volt) { struct nvkm_device_tegra *tdev = device->func->tegra(device); int i, uv; @@ -160,9 +161,9 @@ _gk20a_volt_ctor(struct nvkm_device *device, int index, volt->base.vid_nr = nb_coefs; for (i = 0; i < volt->base.vid_nr; i++) { volt->base.vid[i].vid = i; - volt->base.vid[i].uv = - gk20a_volt_calc_voltage(&coefs[i], - tdev->gpu_speedo); + volt->base.vid[i].uv = max( + gk20a_volt_calc_voltage(&coefs[i], tdev->gpu_speedo), + vmin); nvkm_debug(&volt->base.subdev, "%2d: vid=%d, uv=%d\n", i, volt->base.vid[i].vid, volt->base.vid[i].uv); } @@ -180,6 +181,6 @@ gk20a_volt_new(struct nvkm_device *device, int index, struct nvkm_volt **pvolt) return -ENOMEM; *pvolt = &volt->base; - return _gk20a_volt_ctor(device, index, gk20a_cvb_coef, - ARRAY_SIZE(gk20a_cvb_coef), volt); + return gk20a_volt_ctor(device, index, gk20a_cvb_coef, + ARRAY_SIZE(gk20a_cvb_coef), 0, volt); } diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gk20a.h b/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gk20a.h index 0fa3b502bcf8..6a6c97f9684e 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gk20a.h +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gk20a.h @@ -37,13 +37,8 @@ struct gk20a_volt { struct regulator *vdd; }; -int _gk20a_volt_ctor(struct nvkm_device *device, int index, - const struct cvb_coef *coefs, int nb_coefs, - struct gk20a_volt *volt); - -int gk20a_volt_calc_voltage(const struct cvb_coef *coef, int speedo); -int gk20a_volt_vid_get(struct nvkm_volt *volt); -int gk20a_volt_vid_set(struct nvkm_volt *volt, u8 vid); -int gk20a_volt_set_id(struct nvkm_volt *volt, u8 id, int condition); +int gk20a_volt_ctor(struct nvkm_device *device, int index, + const struct cvb_coef *coefs, int nb_coefs, + int vmin, struct gk20a_volt *volt); #endif diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gm20b.c b/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gm20b.c index 49b5ecb701e4..74db4d28930f 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gm20b.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/volt/gm20b.c @@ -41,16 +41,52 @@ const struct cvb_coef gm20b_cvb_coef[] = { /* 921600 */ { 2647676, -106455, 1632 }, }; +static const struct cvb_coef gm20b_na_cvb_coef[] = { + /* KHz, c0, c1, c2, c3, c4, c5 */ + /* 76800 */ { 814294, 8144, -940, 808, -21583, 226 }, + /* 153600 */ { 856185, 8144, -940, 808, -21583, 226 }, + /* 230400 */ { 898077, 8144, -940, 808, -21583, 226 }, + /* 307200 */ { 939968, 8144, -940, 808, -21583, 226 }, + /* 384000 */ { 981860, 8144, -940, 808, -21583, 226 }, + /* 460800 */ { 1023751, 8144, -940, 808, -21583, 226 }, + /* 537600 */ { 1065642, 8144, -940, 808, -21583, 226 }, + /* 614400 */ { 1107534, 8144, -940, 808, -21583, 226 }, + /* 691200 */ { 1149425, 8144, -940, 808, -21583, 226 }, + /* 768000 */ { 1191317, 8144, -940, 808, -21583, 226 }, + /* 844800 */ { 1233208, 8144, -940, 808, -21583, 226 }, + /* 921600 */ { 1275100, 8144, -940, 808, -21583, 226 }, + /* 998400 */ { 1316991, 8144, -940, 808, -21583, 226 }, +}; + +const u32 speedo_to_vmin[] = { + /* 0, 1, 2, 3, 4, */ + 950000, 840000, 818750, 840000, 810000, +}; + int gm20b_volt_new(struct nvkm_device *device, int index, struct nvkm_volt **pvolt) { + struct nvkm_device_tegra *tdev = device->func->tegra(device); struct gk20a_volt *volt; + u32 vmin; + + if (tdev->gpu_speedo_id >= ARRAY_SIZE(speedo_to_vmin)) { + nvdev_error(device, "unsupported speedo %d\n", + tdev->gpu_speedo_id); + return -EINVAL; + } volt = kzalloc(sizeof(*volt), GFP_KERNEL); if (!volt) return -ENOMEM; *pvolt = &volt->base; - return _gk20a_volt_ctor(device, index, gm20b_cvb_coef, - ARRAY_SIZE(gm20b_cvb_coef), volt); + vmin = speedo_to_vmin[tdev->gpu_speedo_id]; + + if (tdev->gpu_speedo_id >= 1) + return gk20a_volt_ctor(device, index, gm20b_na_cvb_coef, + ARRAY_SIZE(gm20b_na_cvb_coef), vmin, volt); + else + return gk20a_volt_ctor(device, index, gm20b_cvb_coef, + ARRAY_SIZE(gm20b_cvb_coef), vmin, volt); } diff --git a/drivers/gpu/drm/omapdrm/Kconfig b/drivers/gpu/drm/omapdrm/Kconfig index 336ad4de9981..556f81f6b2c7 100644 --- a/drivers/gpu/drm/omapdrm/Kconfig +++ b/drivers/gpu/drm/omapdrm/Kconfig @@ -4,11 +4,6 @@ config DRM_OMAP depends on ARCH_OMAP2PLUS || ARCH_MULTIPLATFORM select OMAP2_DSS select DRM_KMS_HELPER - select DRM_KMS_FB_HELPER - select FB_SYS_FILLRECT - select FB_SYS_COPYAREA - select FB_SYS_IMAGEBLIT - select FB_SYS_FOPS default n help DRM display driver for OMAP2/3/4 based boards. diff --git a/drivers/gpu/drm/omapdrm/displays/connector-dvi.c b/drivers/gpu/drm/omapdrm/displays/connector-dvi.c index 75f7827525cf..684b7aeda411 100644 --- a/drivers/gpu/drm/omapdrm/displays/connector-dvi.c +++ b/drivers/gpu/drm/omapdrm/displays/connector-dvi.c @@ -255,6 +255,7 @@ static int dvic_probe_of(struct platform_device *pdev) adapter_node = of_parse_phandle(node, "ddc-i2c-bus", 0); if (adapter_node) { adapter = of_get_i2c_adapter_by_node(adapter_node); + of_node_put(adapter_node); if (adapter == NULL) { dev_err(&pdev->dev, "failed to parse ddc-i2c-bus\n"); omap_dss_put_device(ddata->in); diff --git a/drivers/gpu/drm/omapdrm/displays/panel-dsi-cm.c b/drivers/gpu/drm/omapdrm/displays/panel-dsi-cm.c index 1b0cf2d8224b..0eae8afaed90 100644 --- a/drivers/gpu/drm/omapdrm/displays/panel-dsi-cm.c +++ b/drivers/gpu/drm/omapdrm/displays/panel-dsi-cm.c @@ -1284,8 +1284,7 @@ static int dsicm_probe(struct platform_device *pdev) return 0; err_sysfs_create: - if (bldev != NULL) - backlight_device_unregister(bldev); + backlight_device_unregister(bldev); err_bl: destroy_workqueue(ddata->workqueue); err_reg: diff --git a/drivers/gpu/drm/omapdrm/dss/dsi.c b/drivers/gpu/drm/omapdrm/dss/dsi.c index 6f45e9d00b41..e1be5e795cd8 100644 --- a/drivers/gpu/drm/omapdrm/dss/dsi.c +++ b/drivers/gpu/drm/omapdrm/dss/dsi.c @@ -1167,7 +1167,6 @@ static int dsi_regulator_init(struct platform_device *dsidev) { struct dsi_data *dsi = dsi_get_dsidrv_data(dsidev); struct regulator *vdds_dsi; - int r; if (dsi->vdds_dsi_reg != NULL) return 0; diff --git a/drivers/gpu/drm/omapdrm/dss/dss-of.c b/drivers/gpu/drm/omapdrm/dss/dss-of.c index dfd4e9621e3b..e256d879b25c 100644 --- a/drivers/gpu/drm/omapdrm/dss/dss-of.c +++ b/drivers/gpu/drm/omapdrm/dss/dss-of.c @@ -125,15 +125,16 @@ u32 dss_of_port_get_port_number(struct device_node *port) static struct device_node *omapdss_of_get_remote_port(const struct device_node *node) { - struct device_node *np; + struct device_node *np, *np_parent; np = of_parse_phandle(node, "remote-endpoint", 0); if (!np) return NULL; - np = of_get_next_parent(np); + np_parent = of_get_next_parent(np); + of_node_put(np); - return np; + return np_parent; } struct device_node * diff --git a/drivers/gpu/drm/omapdrm/dss/hdmi5.c b/drivers/gpu/drm/omapdrm/dss/hdmi5.c index 061f9bab4c9b..0c0a5139a301 100644 --- a/drivers/gpu/drm/omapdrm/dss/hdmi5.c +++ b/drivers/gpu/drm/omapdrm/dss/hdmi5.c @@ -120,7 +120,6 @@ static irqreturn_t hdmi_irq_handler(int irq, void *data) static int hdmi_init_regulator(void) { - int r; struct regulator *reg; if (hdmi.vdda_reg != NULL) diff --git a/drivers/gpu/drm/omapdrm/omap_fb.c b/drivers/gpu/drm/omapdrm/omap_fb.c index 983c8cf2441c..31f5178c22c7 100644 --- a/drivers/gpu/drm/omapdrm/omap_fb.c +++ b/drivers/gpu/drm/omapdrm/omap_fb.c @@ -115,8 +115,8 @@ static void omap_framebuffer_destroy(struct drm_framebuffer *fb) for (i = 0; i < n; i++) { struct plane *plane = &omap_fb->planes[i]; - if (plane->bo) - drm_gem_object_unreference_unlocked(plane->bo); + + drm_gem_object_unreference_unlocked(plane->bo); } kfree(omap_fb); diff --git a/drivers/gpu/drm/panel/panel-simple.c b/drivers/gpu/drm/panel/panel-simple.c index 3a7bdf1c842b..85143d1b9b31 100644 --- a/drivers/gpu/drm/panel/panel-simple.c +++ b/drivers/gpu/drm/panel/panel-simple.c @@ -168,6 +168,7 @@ static int panel_simple_disable(struct drm_panel *panel) if (p->backlight) { p->backlight->props.power = FB_BLANK_POWERDOWN; + p->backlight->props.state |= BL_CORE_FBBLANK; backlight_update_status(p->backlight); } @@ -235,6 +236,7 @@ static int panel_simple_enable(struct drm_panel *panel) msleep(p->desc->delay.enable); if (p->backlight) { + p->backlight->props.state &= ~BL_CORE_FBBLANK; p->backlight->props.power = FB_BLANK_UNBLANK; backlight_update_status(p->backlight); } @@ -964,8 +966,8 @@ static const struct panel_desc innolux_zj070na_01p = { .num_modes = 1, .bpc = 6, .size = { - .width = 1024, - .height = 600, + .width = 154, + .height = 90, }, }; @@ -1017,6 +1019,51 @@ static const struct panel_desc lg_lb070wv8 = { .bus_format = MEDIA_BUS_FMT_RGB888_1X7X4_SPWG, }; +static const struct drm_display_mode lg_lp079qx1_sp0v_mode = { + .clock = 200000, + .hdisplay = 1536, + .hsync_start = 1536 + 12, + .hsync_end = 1536 + 12 + 16, + .htotal = 1536 + 12 + 16 + 48, + .vdisplay = 2048, + .vsync_start = 2048 + 8, + .vsync_end = 2048 + 8 + 4, + .vtotal = 2048 + 8 + 4 + 8, + .vrefresh = 60, + .flags = DRM_MODE_FLAG_NVSYNC | DRM_MODE_FLAG_NHSYNC, +}; + +static const struct panel_desc lg_lp079qx1_sp0v = { + .modes = &lg_lp079qx1_sp0v_mode, + .num_modes = 1, + .size = { + .width = 129, + .height = 171, + }, +}; + +static const struct drm_display_mode lg_lp097qx1_spa1_mode = { + .clock = 205210, + .hdisplay = 2048, + .hsync_start = 2048 + 150, + .hsync_end = 2048 + 150 + 5, + .htotal = 2048 + 150 + 5 + 5, + .vdisplay = 1536, + .vsync_start = 1536 + 3, + .vsync_end = 1536 + 3 + 1, + .vtotal = 1536 + 3 + 1 + 9, + .vrefresh = 60, +}; + +static const struct panel_desc lg_lp097qx1_spa1 = { + .modes = &lg_lp097qx1_spa1_mode, + .num_modes = 1, + .size = { + .width = 208, + .height = 147, + }, +}; + static const struct drm_display_mode lg_lp120up1_mode = { .clock = 162300, .hdisplay = 1920, @@ -1224,6 +1271,28 @@ static const struct panel_desc qd43003c0_40 = { .bus_format = MEDIA_BUS_FMT_RGB888_1X24, }; +static const struct drm_display_mode samsung_lsn122dl01_c01_mode = { + .clock = 271560, + .hdisplay = 2560, + .hsync_start = 2560 + 48, + .hsync_end = 2560 + 48 + 32, + .htotal = 2560 + 48 + 32 + 80, + .vdisplay = 1600, + .vsync_start = 1600 + 2, + .vsync_end = 1600 + 2 + 5, + .vtotal = 1600 + 2 + 5 + 57, + .vrefresh = 60, +}; + +static const struct panel_desc samsung_lsn122dl01_c01 = { + .modes = &samsung_lsn122dl01_c01_mode, + .num_modes = 1, + .size = { + .width = 263, + .height = 164, + }, +}; + static const struct drm_display_mode samsung_ltn101nt05_mode = { .clock = 54030, .hdisplay = 1024, @@ -1242,8 +1311,8 @@ static const struct panel_desc samsung_ltn101nt05 = { .num_modes = 1, .bpc = 6, .size = { - .width = 1024, - .height = 600, + .width = 223, + .height = 125, }, }; @@ -1270,6 +1339,53 @@ static const struct panel_desc samsung_ltn140at29_301 = { }, }; +static const struct display_timing sharp_lq101k1ly04_timing = { + .pixelclock = { 60000000, 65000000, 80000000 }, + .hactive = { 1280, 1280, 1280 }, + .hfront_porch = { 20, 20, 20 }, + .hback_porch = { 20, 20, 20 }, + .hsync_len = { 10, 10, 10 }, + .vactive = { 800, 800, 800 }, + .vfront_porch = { 4, 4, 4 }, + .vback_porch = { 4, 4, 4 }, + .vsync_len = { 4, 4, 4 }, + .flags = DISPLAY_FLAGS_PIXDATA_POSEDGE, +}; + +static const struct panel_desc sharp_lq101k1ly04 = { + .timings = &sharp_lq101k1ly04_timing, + .num_timings = 1, + .bpc = 8, + .size = { + .width = 217, + .height = 136, + }, + .bus_format = MEDIA_BUS_FMT_RGB888_1X7X4_JEIDA, +}; + +static const struct drm_display_mode sharp_lq123p1jx31_mode = { + .clock = 252750, + .hdisplay = 2400, + .hsync_start = 2400 + 48, + .hsync_end = 2400 + 48 + 32, + .htotal = 2400 + 48 + 32 + 80, + .vdisplay = 1600, + .vsync_start = 1600 + 3, + .vsync_end = 1600 + 3 + 10, + .vtotal = 1600 + 3 + 10 + 33, + .vrefresh = 60, + .flags = DRM_MODE_FLAG_NVSYNC | DRM_MODE_FLAG_NHSYNC, +}; + +static const struct panel_desc sharp_lq123p1jx31 = { + .modes = &sharp_lq123p1jx31_mode, + .num_modes = 1, + .size = { + .width = 259, + .height = 173, + }, +}; + static const struct drm_display_mode shelly_sca07010_bfn_lnn_mode = { .clock = 33300, .hdisplay = 800, @@ -1293,6 +1409,29 @@ static const struct panel_desc shelly_sca07010_bfn_lnn = { .bus_format = MEDIA_BUS_FMT_RGB666_1X18, }; +static const struct drm_display_mode starry_kr122ea0sra_mode = { + .clock = 147000, + .hdisplay = 1920, + .hsync_start = 1920 + 16, + .hsync_end = 1920 + 16 + 16, + .htotal = 1920 + 16 + 16 + 32, + .vdisplay = 1200, + .vsync_start = 1200 + 15, + .vsync_end = 1200 + 15 + 2, + .vtotal = 1200 + 15 + 2 + 18, + .vrefresh = 60, + .flags = DRM_MODE_FLAG_NVSYNC | DRM_MODE_FLAG_NHSYNC, +}; + +static const struct panel_desc starry_kr122ea0sra = { + .modes = &starry_kr122ea0sra_mode, + .num_modes = 1, + .size = { + .width = 263, + .height = 164, + }, +}; + static const struct drm_display_mode tpk_f07a_0102_mode = { .clock = 33260, .hdisplay = 800, @@ -1457,6 +1596,12 @@ static const struct of_device_id platform_of_match[] = { .compatible = "lg,lb070wv8", .data = &lg_lb070wv8, }, { + .compatible = "lg,lp079qx1-sp0v", + .data = &lg_lp079qx1_sp0v, + }, { + .compatible = "lg,lp097qx1-spa1", + .data = &lg_lp097qx1_spa1, + }, { .compatible = "lg,lp120up1", .data = &lg_lp120up1, }, { @@ -1481,15 +1626,27 @@ static const struct of_device_id platform_of_match[] = { .compatible = "qiaodian,qd43003c0-40", .data = &qd43003c0_40, }, { + .compatible = "samsung,lsn122dl01-c01", + .data = &samsung_lsn122dl01_c01, + }, { .compatible = "samsung,ltn101nt05", .data = &samsung_ltn101nt05, }, { .compatible = "samsung,ltn140at29-301", .data = &samsung_ltn140at29_301, }, { + .compatible = "sharp,lq101k1ly04", + .data = &sharp_lq101k1ly04, + }, { + .compatible = "sharp,lq123p1jx31", + .data = &sharp_lq123p1jx31, + }, { .compatible = "shelly,sca07010-bfn-lnn", .data = &shelly_sca07010_bfn_lnn, }, { + .compatible = "starry,kr122ea0sra", + .data = &starry_kr122ea0sra, + }, { .compatible = "tpk,f07a-0102", .data = &tpk_f07a_0102, }, { @@ -1701,7 +1858,6 @@ static const struct panel_desc_dsi panasonic_vvx10f004b00 = { .lanes = 4, }; - static const struct of_device_id dsi_of_match[] = { { .compatible = "auo,b080uan01", diff --git a/drivers/gpu/drm/qxl/Kconfig b/drivers/gpu/drm/qxl/Kconfig index 38c2bb72e456..da45b11b66b8 100644 --- a/drivers/gpu/drm/qxl/Kconfig +++ b/drivers/gpu/drm/qxl/Kconfig @@ -1,12 +1,7 @@ config DRM_QXL tristate "QXL virtual GPU" depends on DRM && PCI - select FB_SYS_FILLRECT - select FB_SYS_COPYAREA - select FB_SYS_IMAGEBLIT - select FB_DEFERRED_IO select DRM_KMS_HELPER - select DRM_KMS_FB_HELPER select DRM_TTM select CRC32 help diff --git a/drivers/gpu/drm/qxl/qxl_cmd.c b/drivers/gpu/drm/qxl/qxl_cmd.c index b5d4b41361bd..04270f5d110c 100644 --- a/drivers/gpu/drm/qxl/qxl_cmd.c +++ b/drivers/gpu/drm/qxl/qxl_cmd.c @@ -203,7 +203,7 @@ qxl_push_cursor_ring_release(struct qxl_device *qdev, struct qxl_release *releas bool qxl_queue_garbage_collect(struct qxl_device *qdev, bool flush) { if (!qxl_check_idle(qdev->release_ring)) { - queue_work(qdev->gc_queue, &qdev->gc_work); + schedule_work(&qdev->gc_work); if (flush) flush_work(&qdev->gc_work); return true; diff --git a/drivers/gpu/drm/qxl/qxl_display.c b/drivers/gpu/drm/qxl/qxl_display.c index ad429683fef7..3aef12742a53 100644 --- a/drivers/gpu/drm/qxl/qxl_display.c +++ b/drivers/gpu/drm/qxl/qxl_display.c @@ -468,8 +468,7 @@ void qxl_user_framebuffer_destroy(struct drm_framebuffer *fb) { struct qxl_framebuffer *qxl_fb = to_qxl_framebuffer(fb); - if (qxl_fb->obj) - drm_gem_object_unreference_unlocked(qxl_fb->obj); + drm_gem_object_unreference_unlocked(qxl_fb->obj); drm_framebuffer_cleanup(fb); kfree(qxl_fb); } diff --git a/drivers/gpu/drm/qxl/qxl_draw.c b/drivers/gpu/drm/qxl/qxl_draw.c index 56e1d633875e..ffe885395145 100644 --- a/drivers/gpu/drm/qxl/qxl_draw.c +++ b/drivers/gpu/drm/qxl/qxl_draw.c @@ -37,7 +37,6 @@ static int alloc_clips(struct qxl_device *qdev, * the qxl_clip_rects. This is *not* the same as the memory allocated * on the device, it is offset to qxl_clip_rects.chunk.data */ static struct qxl_rect *drawable_set_clipping(struct qxl_device *qdev, - struct qxl_drawable *drawable, unsigned num_clips, struct qxl_bo *clips_bo) { @@ -136,6 +135,8 @@ static int qxl_palette_create_1bit(struct qxl_bo *palette_bo, * correctly globaly, since that would require * tracking all of our palettes. */ ret = qxl_bo_kmap(palette_bo, (void **)&pal); + if (ret) + return ret; pal->num_ents = 2; pal->unique = unique++; if (visual == FB_VISUAL_TRUECOLOR || visual == FB_VISUAL_DIRECTCOLOR) { @@ -349,7 +350,7 @@ void qxl_draw_dirty_fb(struct qxl_device *qdev, if (ret) goto out_release_backoff; - rects = drawable_set_clipping(qdev, drawable, num_clips, clips_bo); + rects = drawable_set_clipping(qdev, num_clips, clips_bo); if (!rects) goto out_release_backoff; diff --git a/drivers/gpu/drm/qxl/qxl_drv.h b/drivers/gpu/drm/qxl/qxl_drv.h index 3ad6604b34ce..8e633caa4078 100644 --- a/drivers/gpu/drm/qxl/qxl_drv.h +++ b/drivers/gpu/drm/qxl/qxl_drv.h @@ -321,7 +321,6 @@ struct qxl_device { struct qxl_bo *current_release_bo[3]; int current_release_bo_offset[3]; - struct workqueue_struct *gc_queue; struct work_struct gc_work; struct drm_property *hotplug_mode_update_property; diff --git a/drivers/gpu/drm/qxl/qxl_kms.c b/drivers/gpu/drm/qxl/qxl_kms.c index 2319800b7add..e642242728c0 100644 --- a/drivers/gpu/drm/qxl/qxl_kms.c +++ b/drivers/gpu/drm/qxl/qxl_kms.c @@ -258,7 +258,6 @@ static int qxl_device_init(struct qxl_device *qdev, (unsigned long)qdev->surfaceram_size); - qdev->gc_queue = create_singlethread_workqueue("qxl_gc"); INIT_WORK(&qdev->gc_work, qxl_gc_work); return 0; @@ -270,10 +269,7 @@ static void qxl_device_fini(struct qxl_device *qdev) qxl_bo_unref(&qdev->current_release_bo[0]); if (qdev->current_release_bo[1]) qxl_bo_unref(&qdev->current_release_bo[1]); - flush_workqueue(qdev->gc_queue); - destroy_workqueue(qdev->gc_queue); - qdev->gc_queue = NULL; - + flush_work(&qdev->gc_work); qxl_ring_free(qdev->command_ring); qxl_ring_free(qdev->cursor_ring); qxl_ring_free(qdev->release_ring); @@ -310,10 +306,6 @@ int qxl_driver_load(struct drm_device *dev, unsigned long flags) struct qxl_device *qdev; int r; - /* require kms */ - if (!drm_core_check_feature(dev, DRIVER_MODESET)) - return -ENODEV; - qdev = kzalloc(sizeof(struct qxl_device), GFP_KERNEL); if (qdev == NULL) return -ENOMEM; diff --git a/drivers/gpu/drm/qxl/qxl_ttm.c b/drivers/gpu/drm/qxl/qxl_ttm.c index 0738d74c8d04..d50c9679e631 100644 --- a/drivers/gpu/drm/qxl/qxl_ttm.c +++ b/drivers/gpu/drm/qxl/qxl_ttm.c @@ -350,11 +350,19 @@ static int qxl_bo_move(struct ttm_buffer_object *bo, struct ttm_mem_reg *new_mem) { struct ttm_mem_reg *old_mem = &bo->mem; + int ret; + + ret = ttm_bo_wait(bo, interruptible, no_wait_gpu); + if (ret) + return ret; + + if (old_mem->mem_type == TTM_PL_SYSTEM && bo->ttm == NULL) { qxl_move_null(bo, new_mem); return 0; } - return ttm_bo_move_memcpy(bo, evict, no_wait_gpu, new_mem); + return ttm_bo_move_memcpy(bo, evict, interruptible, + no_wait_gpu, new_mem); } static void qxl_bo_move_notify(struct ttm_buffer_object *bo, diff --git a/drivers/gpu/drm/radeon/atombios_crtc.c b/drivers/gpu/drm/radeon/atombios_crtc.c index e91763d5d800..a97abc8af657 100644 --- a/drivers/gpu/drm/radeon/atombios_crtc.c +++ b/drivers/gpu/drm/radeon/atombios_crtc.c @@ -589,7 +589,8 @@ static u32 atombios_adjust_pll(struct drm_crtc *crtc, if (ASIC_IS_DCE41(rdev) || ASIC_IS_DCE61(rdev) || ASIC_IS_DCE8(rdev)) radeon_crtc->pll_flags |= RADEON_PLL_USE_FRAC_FB_DIV; /* use frac fb div on RS780/RS880 */ - if ((rdev->family == CHIP_RS780) || (rdev->family == CHIP_RS880)) + if (((rdev->family == CHIP_RS780) || (rdev->family == CHIP_RS880)) + && !radeon_crtc->ss_enabled) radeon_crtc->pll_flags |= RADEON_PLL_USE_FRAC_FB_DIV; if (ASIC_IS_DCE32(rdev) && mode->clock > 165000) radeon_crtc->pll_flags |= RADEON_PLL_USE_FRAC_FB_DIV; @@ -626,7 +627,7 @@ static u32 atombios_adjust_pll(struct drm_crtc *crtc, if (radeon_crtc->ss.refdiv) { radeon_crtc->pll_flags |= RADEON_PLL_USE_REF_DIV; radeon_crtc->pll_reference_div = radeon_crtc->ss.refdiv; - if (ASIC_IS_AVIVO(rdev)) + if (rdev->family >= CHIP_RV770) radeon_crtc->pll_flags |= RADEON_PLL_USE_FRAC_FB_DIV; } } diff --git a/drivers/gpu/drm/radeon/atombios_encoders.c b/drivers/gpu/drm/radeon/atombios_encoders.c index 587cae4e73c9..56bb758f4e33 100644 --- a/drivers/gpu/drm/radeon/atombios_encoders.c +++ b/drivers/gpu/drm/radeon/atombios_encoders.c @@ -120,6 +120,7 @@ atombios_set_backlight_level(struct radeon_encoder *radeon_encoder, u8 level) case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA: case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1: case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2: + case ENCODER_OBJECT_ID_INTERNAL_UNIPHY3: if (dig->backlight_level == 0) atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_LCD_BLOFF, 0, 0); else { diff --git a/drivers/gpu/drm/radeon/ci_dpm.c b/drivers/gpu/drm/radeon/ci_dpm.c index 35e0fc3ae8a7..7ba450832e6b 100644 --- a/drivers/gpu/drm/radeon/ci_dpm.c +++ b/drivers/gpu/drm/radeon/ci_dpm.c @@ -3843,7 +3843,10 @@ static void ci_find_dpm_states_clocks_in_dpm_table(struct radeon_device *rdev, if (i >= sclk_table->count) { pi->need_update_smu7_dpm_table |= DPMTABLE_OD_UPDATE_SCLK; } else { - /* XXX check display min clock requirements */ + /* XXX The current code always reprogrammed the sclk levels, + * but we don't currently handle disp sclk requirements + * so just skip it. + */ if (CISLAND_MINIMUM_ENGINE_CLOCK != CISLAND_MINIMUM_ENGINE_CLOCK) pi->need_update_smu7_dpm_table |= DPMTABLE_UPDATE_SCLK; } diff --git a/drivers/gpu/drm/radeon/cik.c b/drivers/gpu/drm/radeon/cik.c index ba192a35c607..0c1b9ff433af 100644 --- a/drivers/gpu/drm/radeon/cik.c +++ b/drivers/gpu/drm/radeon/cik.c @@ -53,6 +53,7 @@ MODULE_FIRMWARE("radeon/bonaire_mc.bin"); MODULE_FIRMWARE("radeon/bonaire_rlc.bin"); MODULE_FIRMWARE("radeon/bonaire_sdma.bin"); MODULE_FIRMWARE("radeon/bonaire_smc.bin"); +MODULE_FIRMWARE("radeon/bonaire_k_smc.bin"); MODULE_FIRMWARE("radeon/HAWAII_pfp.bin"); MODULE_FIRMWARE("radeon/HAWAII_me.bin"); @@ -72,6 +73,7 @@ MODULE_FIRMWARE("radeon/hawaii_mc.bin"); MODULE_FIRMWARE("radeon/hawaii_rlc.bin"); MODULE_FIRMWARE("radeon/hawaii_sdma.bin"); MODULE_FIRMWARE("radeon/hawaii_smc.bin"); +MODULE_FIRMWARE("radeon/hawaii_k_smc.bin"); MODULE_FIRMWARE("radeon/KAVERI_pfp.bin"); MODULE_FIRMWARE("radeon/KAVERI_me.bin"); @@ -1990,12 +1992,17 @@ static int cik_init_microcode(struct radeon_device *rdev) int new_fw = 0; int err; int num_fw; + bool new_smc = false; DRM_DEBUG("\n"); switch (rdev->family) { case CHIP_BONAIRE: chip_name = "BONAIRE"; + if ((rdev->pdev->revision == 0x80) || + (rdev->pdev->revision == 0x81) || + (rdev->pdev->device == 0x665f)) + new_smc = true; new_chip_name = "bonaire"; pfp_req_size = CIK_PFP_UCODE_SIZE * 4; me_req_size = CIK_ME_UCODE_SIZE * 4; @@ -2010,6 +2017,8 @@ static int cik_init_microcode(struct radeon_device *rdev) break; case CHIP_HAWAII: chip_name = "HAWAII"; + if (rdev->pdev->revision == 0x80) + new_smc = true; new_chip_name = "hawaii"; pfp_req_size = CIK_PFP_UCODE_SIZE * 4; me_req_size = CIK_ME_UCODE_SIZE * 4; @@ -2259,7 +2268,10 @@ static int cik_init_microcode(struct radeon_device *rdev) } } - snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", new_chip_name); + if (new_smc) + snprintf(fw_name, sizeof(fw_name), "radeon/%s_k_smc.bin", new_chip_name); + else + snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", new_chip_name); err = request_firmware(&rdev->smc_fw, fw_name, rdev->dev); if (err) { snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", chip_name); @@ -8354,7 +8366,8 @@ static int cik_startup(struct radeon_device *rdev) } } rdev->rlc.cs_data = ci_cs_data; - rdev->rlc.cp_table_size = CP_ME_TABLE_SIZE * 5 * 4; + rdev->rlc.cp_table_size = ALIGN(CP_ME_TABLE_SIZE * 5 * 4, 2048); /* CP JT */ + rdev->rlc.cp_table_size += 64 * 1024; /* GDS */ r = sumo_rlc_init(rdev); if (r) { DRM_ERROR("Failed to init rlc BOs!\n"); diff --git a/drivers/gpu/drm/radeon/evergreen_cs.c b/drivers/gpu/drm/radeon/evergreen_cs.c index 0d3f744de35a..d960d3915408 100644 --- a/drivers/gpu/drm/radeon/evergreen_cs.c +++ b/drivers/gpu/drm/radeon/evergreen_cs.c @@ -2209,6 +2209,12 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p, } break; } + case PACKET3_PFP_SYNC_ME: + if (pkt->count) { + DRM_ERROR("bad PFP_SYNC_ME\n"); + return -EINVAL; + } + break; case PACKET3_SURFACE_SYNC: if (pkt->count != 3) { DRM_ERROR("bad SURFACE_SYNC\n"); @@ -3381,6 +3387,7 @@ static int evergreen_vm_packet3_check(struct radeon_device *rdev, case PACKET3_MPEG_INDEX: case PACKET3_WAIT_REG_MEM: case PACKET3_MEM_WRITE: + case PACKET3_PFP_SYNC_ME: case PACKET3_SURFACE_SYNC: case PACKET3_EVENT_WRITE: case PACKET3_EVENT_WRITE_EOP: diff --git a/drivers/gpu/drm/radeon/evergreend.h b/drivers/gpu/drm/radeon/evergreend.h index 0b174e14e9a6..c8e3d394cde7 100644 --- a/drivers/gpu/drm/radeon/evergreend.h +++ b/drivers/gpu/drm/radeon/evergreend.h @@ -1624,6 +1624,7 @@ */ # define PACKET3_CP_DMA_CMD_SAIC (1 << 28) # define PACKET3_CP_DMA_CMD_DAIC (1 << 29) +#define PACKET3_PFP_SYNC_ME 0x42 #define PACKET3_SURFACE_SYNC 0x43 # define PACKET3_CB0_DEST_BASE_ENA (1 << 6) # define PACKET3_CB1_DEST_BASE_ENA (1 << 7) diff --git a/drivers/gpu/drm/radeon/radeon_acpi.c b/drivers/gpu/drm/radeon/radeon_acpi.c index 59acd0e5c2c6..31c9a92d6a1b 100644 --- a/drivers/gpu/drm/radeon/radeon_acpi.c +++ b/drivers/gpu/drm/radeon/radeon_acpi.c @@ -741,13 +741,6 @@ int radeon_acpi_init(struct radeon_device *rdev) } atif->encoder_for_bl = target; - if (!target) { - /* Brightness change notification is enabled, but we - * didn't find a backlight controller, this should - * never happen. - */ - DRM_ERROR("Cannot find a backlight controller\n"); - } } if (atif->functions.sbios_requests && !atif->functions.system_params) { diff --git a/drivers/gpu/drm/radeon/radeon_atombios.c b/drivers/gpu/drm/radeon/radeon_atombios.c index f8097a0e7a79..5df3ec73021b 100644 --- a/drivers/gpu/drm/radeon/radeon_atombios.c +++ b/drivers/gpu/drm/radeon/radeon_atombios.c @@ -1155,7 +1155,7 @@ bool radeon_atom_get_clock_info(struct drm_device *dev) le16_to_cpu(firmware_info->info.usReferenceClock); p1pll->reference_div = 0; - if (crev < 2) + if ((frev < 2) && (crev < 2)) p1pll->pll_out_min = le16_to_cpu(firmware_info->info.usMinPixelClockPLL_Output); else @@ -1164,7 +1164,7 @@ bool radeon_atom_get_clock_info(struct drm_device *dev) p1pll->pll_out_max = le32_to_cpu(firmware_info->info.ulMaxPixelClockPLL_Output); - if (crev >= 4) { + if (((frev < 2) && (crev >= 4)) || (frev >= 2)) { p1pll->lcd_pll_out_min = le16_to_cpu(firmware_info->info_14.usLcdMinPixelClockPLL_Output) * 100; if (p1pll->lcd_pll_out_min == 0) diff --git a/drivers/gpu/drm/radeon/radeon_atpx_handler.c b/drivers/gpu/drm/radeon/radeon_atpx_handler.c index 95f4fea89302..6de342861202 100644 --- a/drivers/gpu/drm/radeon/radeon_atpx_handler.c +++ b/drivers/gpu/drm/radeon/radeon_atpx_handler.c @@ -10,6 +10,7 @@ #include <linux/slab.h> #include <linux/acpi.h> #include <linux/pci.h> +#include <linux/delay.h> #include "radeon_acpi.h" @@ -27,6 +28,7 @@ struct radeon_atpx_functions { struct radeon_atpx { acpi_handle handle; struct radeon_atpx_functions functions; + bool is_hybrid; }; static struct radeon_atpx_priv { @@ -62,6 +64,14 @@ bool radeon_has_atpx(void) { return radeon_atpx_priv.atpx_detected; } +bool radeon_has_atpx_dgpu_power_cntl(void) { + return radeon_atpx_priv.atpx.functions.power_cntl; +} + +bool radeon_is_atpx_hybrid(void) { + return radeon_atpx_priv.atpx.is_hybrid; +} + /** * radeon_atpx_call - call an ATPX method * @@ -141,18 +151,12 @@ static void radeon_atpx_parse_functions(struct radeon_atpx_functions *f, u32 mas */ static int radeon_atpx_validate(struct radeon_atpx *atpx) { - /* make sure required functions are enabled */ - /* dGPU power control is required */ - if (atpx->functions.power_cntl == false) { - printk("ATPX dGPU power cntl not present, forcing\n"); - atpx->functions.power_cntl = true; - } + u32 valid_bits = 0; if (atpx->functions.px_params) { union acpi_object *info; struct atpx_px_params output; size_t size; - u32 valid_bits; info = radeon_atpx_call(atpx->handle, ATPX_FUNCTION_GET_PX_PARAMETERS, NULL); if (!info) @@ -171,19 +175,42 @@ static int radeon_atpx_validate(struct radeon_atpx *atpx) memcpy(&output, info->buffer.pointer, size); valid_bits = output.flags & output.valid_flags; - /* if separate mux flag is set, mux controls are required */ - if (valid_bits & ATPX_SEPARATE_MUX_FOR_I2C) { - atpx->functions.i2c_mux_cntl = true; - atpx->functions.disp_mux_cntl = true; - } - /* if any outputs are muxed, mux controls are required */ - if (valid_bits & (ATPX_CRT1_RGB_SIGNAL_MUXED | - ATPX_TV_SIGNAL_MUXED | - ATPX_DFP_SIGNAL_MUXED)) - atpx->functions.disp_mux_cntl = true; kfree(info); } + + /* if separate mux flag is set, mux controls are required */ + if (valid_bits & ATPX_SEPARATE_MUX_FOR_I2C) { + atpx->functions.i2c_mux_cntl = true; + atpx->functions.disp_mux_cntl = true; + } + /* if any outputs are muxed, mux controls are required */ + if (valid_bits & (ATPX_CRT1_RGB_SIGNAL_MUXED | + ATPX_TV_SIGNAL_MUXED | + ATPX_DFP_SIGNAL_MUXED)) + atpx->functions.disp_mux_cntl = true; + + /* some bioses set these bits rather than flagging power_cntl as supported */ + if (valid_bits & (ATPX_DYNAMIC_PX_SUPPORTED | + ATPX_DYNAMIC_DGPU_POWER_OFF_SUPPORTED)) + atpx->functions.power_cntl = true; + + atpx->is_hybrid = false; + if (valid_bits & ATPX_MS_HYBRID_GFX_SUPPORTED) { + printk("ATPX Hybrid Graphics\n"); +#if 1 + /* This is a temporary hack until the D3 cold support + * makes it upstream. The ATPX power_control method seems + * to still work on even if the system should be using + * the new standardized hybrid D3 cold ACPI interface. + */ + atpx->functions.power_cntl = true; +#else + atpx->functions.power_cntl = false; +#endif + atpx->is_hybrid = true; + } + return 0; } @@ -258,6 +285,10 @@ static int radeon_atpx_set_discrete_state(struct radeon_atpx *atpx, u8 state) if (!info) return -EIO; kfree(info); + + /* 200ms delay is required after off */ + if (state == 0) + msleep(200); } return 0; } @@ -505,7 +536,6 @@ static int radeon_atpx_get_client_id(struct pci_dev *pdev) static const struct vga_switcheroo_handler radeon_atpx_handler = { .switchto = radeon_atpx_switchto, .power_state = radeon_atpx_power_state, - .init = radeon_atpx_init, .get_client_id = radeon_atpx_get_client_id, }; @@ -541,6 +571,7 @@ static bool radeon_atpx_detect(void) printk(KERN_INFO "vga_switcheroo: detected switching method %s handle\n", acpi_method_name); radeon_atpx_priv.atpx_detected = true; + radeon_atpx_init(); return true; } return false; diff --git a/drivers/gpu/drm/radeon/radeon_connectors.c b/drivers/gpu/drm/radeon/radeon_connectors.c index 81a63d7f5cd9..b79f3b002471 100644 --- a/drivers/gpu/drm/radeon/radeon_connectors.c +++ b/drivers/gpu/drm/radeon/radeon_connectors.c @@ -2064,7 +2064,6 @@ radeon_add_atom_connector(struct drm_device *dev, RADEON_OUTPUT_CSC_BYPASS); /* no HPD on analog connectors */ radeon_connector->hpd.hpd = RADEON_HPD_NONE; - connector->polled = DRM_CONNECTOR_POLL_CONNECT; connector->interlace_allowed = true; connector->doublescan_allowed = true; break; @@ -2314,8 +2313,10 @@ radeon_add_atom_connector(struct drm_device *dev, } if (radeon_connector->hpd.hpd == RADEON_HPD_NONE) { - if (i2c_bus->valid) - connector->polled = DRM_CONNECTOR_POLL_CONNECT; + if (i2c_bus->valid) { + connector->polled = DRM_CONNECTOR_POLL_CONNECT | + DRM_CONNECTOR_POLL_DISCONNECT; + } } else connector->polled = DRM_CONNECTOR_POLL_HPD; @@ -2391,7 +2392,6 @@ radeon_add_legacy_connector(struct drm_device *dev, 1); /* no HPD on analog connectors */ radeon_connector->hpd.hpd = RADEON_HPD_NONE; - connector->polled = DRM_CONNECTOR_POLL_CONNECT; connector->interlace_allowed = true; connector->doublescan_allowed = true; break; @@ -2476,10 +2476,13 @@ radeon_add_legacy_connector(struct drm_device *dev, } if (radeon_connector->hpd.hpd == RADEON_HPD_NONE) { - if (i2c_bus->valid) - connector->polled = DRM_CONNECTOR_POLL_CONNECT; + if (i2c_bus->valid) { + connector->polled = DRM_CONNECTOR_POLL_CONNECT | + DRM_CONNECTOR_POLL_DISCONNECT; + } } else connector->polled = DRM_CONNECTOR_POLL_HPD; + connector->display_info.subpixel_order = subpixel_order; drm_connector_register(connector); } diff --git a/drivers/gpu/drm/radeon/radeon_device.c b/drivers/gpu/drm/radeon/radeon_device.c index e721e6b2766e..a00dd2f74527 100644 --- a/drivers/gpu/drm/radeon/radeon_device.c +++ b/drivers/gpu/drm/radeon/radeon_device.c @@ -30,6 +30,7 @@ #include <drm/drmP.h> #include <drm/drm_crtc_helper.h> #include <drm/radeon_drm.h> +#include <linux/pm_runtime.h> #include <linux/vgaarb.h> #include <linux/vga_switcheroo.h> #include <linux/efi.h> @@ -630,6 +631,23 @@ void radeon_gtt_location(struct radeon_device *rdev, struct radeon_mc *mc) /* * GPU helpers function. */ + +/** + * radeon_device_is_virtual - check if we are running is a virtual environment + * + * Check if the asic has been passed through to a VM (all asics). + * Used at driver startup. + * Returns true if virtual or false if not. + */ +static bool radeon_device_is_virtual(void) +{ +#ifdef CONFIG_X86 + return boot_cpu_has(X86_FEATURE_HYPERVISOR); +#else + return false; +#endif +} + /** * radeon_card_posted - check if the hw has already been initialized * @@ -643,6 +661,10 @@ bool radeon_card_posted(struct radeon_device *rdev) { uint32_t reg; + /* for pass through, always force asic_init */ + if (radeon_device_is_virtual()) + return false; + /* required for EFI mode on macbook2,1 which uses an r5xx asic */ if (efi_enabled(EFI_BOOT) && (rdev->pdev->subsystem_vendor == PCI_VENDOR_ID_APPLE) && @@ -1505,6 +1527,9 @@ int radeon_device_init(struct radeon_device *rdev, return 0; failed: + /* balance pm_runtime_get_sync() in radeon_driver_unload_kms() */ + if (radeon_is_px(ddev)) + pm_runtime_put_noidle(ddev->dev); if (runtime) vga_switcheroo_fini_domain_pm_ops(rdev->dev); return r; @@ -1631,7 +1656,7 @@ int radeon_suspend_kms(struct drm_device *dev, bool suspend, radeon_agp_suspend(rdev); pci_save_state(dev->pdev); - if (freeze && rdev->family >= CHIP_R600) { + if (freeze && rdev->family >= CHIP_CEDAR) { rdev->asic->asic_reset(rdev, true); pci_restore_state(dev->pdev); } else if (suspend) { diff --git a/drivers/gpu/drm/radeon/radeon_display.c b/drivers/gpu/drm/radeon/radeon_display.c index 3965d1916b9c..c3206fb8f4cf 100644 --- a/drivers/gpu/drm/radeon/radeon_display.c +++ b/drivers/gpu/drm/radeon/radeon_display.c @@ -714,7 +714,7 @@ static void radeon_crtc_init(struct drm_device *dev, int index) drm_mode_crtc_set_gamma_size(&radeon_crtc->base, 256); radeon_crtc->crtc_id = index; - radeon_crtc->flip_queue = create_singlethread_workqueue("radeon-crtc"); + radeon_crtc->flip_queue = alloc_workqueue("radeon-crtc", WQ_HIGHPRI, 0); rdev->mode_info.crtcs[index] = radeon_crtc; if (rdev->family >= CHIP_BONAIRE) { @@ -1324,9 +1324,7 @@ static void radeon_user_framebuffer_destroy(struct drm_framebuffer *fb) { struct radeon_framebuffer *radeon_fb = to_radeon_framebuffer(fb); - if (radeon_fb->obj) { - drm_gem_object_unreference_unlocked(radeon_fb->obj); - } + drm_gem_object_unreference_unlocked(radeon_fb->obj); drm_framebuffer_cleanup(fb); kfree(radeon_fb); } @@ -1711,6 +1709,7 @@ void radeon_modeset_fini(struct radeon_device *rdev) radeon_afmt_fini(rdev); drm_kms_helper_poll_fini(rdev->ddev); radeon_hpd_fini(rdev); + drm_crtc_force_disable_all(rdev->ddev); drm_mode_config_cleanup(rdev->ddev); rdev->mode_info.mode_config_initialized = false; } diff --git a/drivers/gpu/drm/radeon/radeon_drv.c b/drivers/gpu/drm/radeon/radeon_drv.c index a455dc7d4aa1..c01a7c6abb49 100644 --- a/drivers/gpu/drm/radeon/radeon_drv.c +++ b/drivers/gpu/drm/radeon/radeon_drv.c @@ -93,9 +93,10 @@ * 2.43.0 - RADEON_INFO_GPU_RESET_COUNTER * 2.44.0 - SET_APPEND_CNT packet3 support * 2.45.0 - Allow setting shader registers using DMA/COPY packet3 on SI + * 2.46.0 - Add PFP_SYNC_ME support on evergreen */ #define KMS_DRIVER_MAJOR 2 -#define KMS_DRIVER_MINOR 45 +#define KMS_DRIVER_MINOR 46 #define KMS_DRIVER_PATCHLEVEL 0 int radeon_driver_load_kms(struct drm_device *dev, unsigned long flags); int radeon_driver_unload_kms(struct drm_device *dev); @@ -162,9 +163,13 @@ void radeon_debugfs_cleanup(struct drm_minor *minor); #if defined(CONFIG_VGA_SWITCHEROO) void radeon_register_atpx_handler(void); void radeon_unregister_atpx_handler(void); +bool radeon_has_atpx_dgpu_power_cntl(void); +bool radeon_is_atpx_hybrid(void); #else static inline void radeon_register_atpx_handler(void) {} static inline void radeon_unregister_atpx_handler(void) {} +static inline bool radeon_has_atpx_dgpu_power_cntl(void) { return false; } +static inline bool radeon_is_atpx_hybrid(void) { return false; } #endif int radeon_no_wb; @@ -404,7 +409,10 @@ static int radeon_pmops_runtime_suspend(struct device *dev) pci_save_state(pdev); pci_disable_device(pdev); pci_ignore_hotplug(pdev); - pci_set_power_state(pdev, PCI_D3cold); + if (radeon_is_atpx_hybrid()) + pci_set_power_state(pdev, PCI_D3cold); + else if (!radeon_has_atpx_dgpu_power_cntl()) + pci_set_power_state(pdev, PCI_D3hot); drm_dev->switch_power_state = DRM_SWITCH_POWER_DYNAMIC_OFF; return 0; @@ -421,7 +429,9 @@ static int radeon_pmops_runtime_resume(struct device *dev) drm_dev->switch_power_state = DRM_SWITCH_POWER_CHANGING; - pci_set_power_state(pdev, PCI_D0); + if (radeon_is_atpx_hybrid() || + !radeon_has_atpx_dgpu_power_cntl()) + pci_set_power_state(pdev, PCI_D0); pci_restore_state(pdev); ret = pci_enable_device(pdev); if (ret) diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c index 414953c46a38..835563c1f0ed 100644 --- a/drivers/gpu/drm/radeon/radeon_kms.c +++ b/drivers/gpu/drm/radeon/radeon_kms.c @@ -63,7 +63,10 @@ int radeon_driver_unload_kms(struct drm_device *dev) if (rdev->rmmio == NULL) goto done_free; - pm_runtime_get_sync(dev->dev); + if (radeon_is_px(dev)) { + pm_runtime_get_sync(dev->dev); + pm_runtime_forbid(dev->dev); + } radeon_kfd_device_fini(rdev); diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c index 590b0377fbe2..ffdad81ef964 100644 --- a/drivers/gpu/drm/radeon/radeon_ttm.c +++ b/drivers/gpu/drm/radeon/radeon_ttm.c @@ -300,8 +300,7 @@ static int radeon_move_blit(struct ttm_buffer_object *bo, if (IS_ERR(fence)) return PTR_ERR(fence); - r = ttm_bo_move_accel_cleanup(bo, &fence->base, - evict, no_wait_gpu, new_mem); + r = ttm_bo_move_accel_cleanup(bo, &fence->base, evict, new_mem); radeon_fence_unref(&fence); return r; } @@ -403,6 +402,10 @@ static int radeon_bo_move(struct ttm_buffer_object *bo, struct ttm_mem_reg *old_mem = &bo->mem; int r; + r = ttm_bo_wait(bo, interruptible, no_wait_gpu); + if (r) + return r; + /* Can't move a pinned BO */ rbo = container_of(bo, struct radeon_bo, tbo); if (WARN_ON_ONCE(rbo->pin_count > 0)) @@ -441,7 +444,8 @@ static int radeon_bo_move(struct ttm_buffer_object *bo, if (r) { memcpy: - r = ttm_bo_move_memcpy(bo, evict, no_wait_gpu, new_mem); + r = ttm_bo_move_memcpy(bo, evict, interruptible, + no_wait_gpu, new_mem); if (r) { return r; } diff --git a/drivers/gpu/drm/radeon/si.c b/drivers/gpu/drm/radeon/si.c index b30e719dd56d..2523ca96c6c7 100644 --- a/drivers/gpu/drm/radeon/si.c +++ b/drivers/gpu/drm/radeon/si.c @@ -50,6 +50,7 @@ MODULE_FIRMWARE("radeon/tahiti_ce.bin"); MODULE_FIRMWARE("radeon/tahiti_mc.bin"); MODULE_FIRMWARE("radeon/tahiti_rlc.bin"); MODULE_FIRMWARE("radeon/tahiti_smc.bin"); +MODULE_FIRMWARE("radeon/tahiti_k_smc.bin"); MODULE_FIRMWARE("radeon/PITCAIRN_pfp.bin"); MODULE_FIRMWARE("radeon/PITCAIRN_me.bin"); @@ -65,6 +66,7 @@ MODULE_FIRMWARE("radeon/pitcairn_ce.bin"); MODULE_FIRMWARE("radeon/pitcairn_mc.bin"); MODULE_FIRMWARE("radeon/pitcairn_rlc.bin"); MODULE_FIRMWARE("radeon/pitcairn_smc.bin"); +MODULE_FIRMWARE("radeon/pitcairn_k_smc.bin"); MODULE_FIRMWARE("radeon/VERDE_pfp.bin"); MODULE_FIRMWARE("radeon/VERDE_me.bin"); @@ -80,6 +82,7 @@ MODULE_FIRMWARE("radeon/verde_ce.bin"); MODULE_FIRMWARE("radeon/verde_mc.bin"); MODULE_FIRMWARE("radeon/verde_rlc.bin"); MODULE_FIRMWARE("radeon/verde_smc.bin"); +MODULE_FIRMWARE("radeon/verde_k_smc.bin"); MODULE_FIRMWARE("radeon/OLAND_pfp.bin"); MODULE_FIRMWARE("radeon/OLAND_me.bin"); @@ -95,6 +98,7 @@ MODULE_FIRMWARE("radeon/oland_ce.bin"); MODULE_FIRMWARE("radeon/oland_mc.bin"); MODULE_FIRMWARE("radeon/oland_rlc.bin"); MODULE_FIRMWARE("radeon/oland_smc.bin"); +MODULE_FIRMWARE("radeon/oland_k_smc.bin"); MODULE_FIRMWARE("radeon/HAINAN_pfp.bin"); MODULE_FIRMWARE("radeon/HAINAN_me.bin"); @@ -110,6 +114,7 @@ MODULE_FIRMWARE("radeon/hainan_ce.bin"); MODULE_FIRMWARE("radeon/hainan_mc.bin"); MODULE_FIRMWARE("radeon/hainan_rlc.bin"); MODULE_FIRMWARE("radeon/hainan_smc.bin"); +MODULE_FIRMWARE("radeon/hainan_k_smc.bin"); static u32 si_get_cu_active_bitmap(struct radeon_device *rdev, u32 se, u32 sh); static void si_pcie_gen3_enable(struct radeon_device *rdev); @@ -1653,12 +1658,16 @@ static int si_init_microcode(struct radeon_device *rdev) char fw_name[30]; int err; int new_fw = 0; + bool new_smc = false; DRM_DEBUG("\n"); switch (rdev->family) { case CHIP_TAHITI: chip_name = "TAHITI"; + /* XXX: figure out which Tahitis need the new ucode */ + if (0) + new_smc = true; new_chip_name = "tahiti"; pfp_req_size = SI_PFP_UCODE_SIZE * 4; me_req_size = SI_PM4_UCODE_SIZE * 4; @@ -1670,6 +1679,13 @@ static int si_init_microcode(struct radeon_device *rdev) break; case CHIP_PITCAIRN: chip_name = "PITCAIRN"; + if ((rdev->pdev->revision == 0x81) || + (rdev->pdev->device == 0x6810) || + (rdev->pdev->device == 0x6811) || + (rdev->pdev->device == 0x6816) || + (rdev->pdev->device == 0x6817) || + (rdev->pdev->device == 0x6806)) + new_smc = true; new_chip_name = "pitcairn"; pfp_req_size = SI_PFP_UCODE_SIZE * 4; me_req_size = SI_PM4_UCODE_SIZE * 4; @@ -1681,6 +1697,16 @@ static int si_init_microcode(struct radeon_device *rdev) break; case CHIP_VERDE: chip_name = "VERDE"; + if ((rdev->pdev->revision == 0x81) || + (rdev->pdev->revision == 0x83) || + (rdev->pdev->revision == 0x87) || + (rdev->pdev->device == 0x6820) || + (rdev->pdev->device == 0x6821) || + (rdev->pdev->device == 0x6822) || + (rdev->pdev->device == 0x6823) || + (rdev->pdev->device == 0x682A) || + (rdev->pdev->device == 0x682B)) + new_smc = true; new_chip_name = "verde"; pfp_req_size = SI_PFP_UCODE_SIZE * 4; me_req_size = SI_PM4_UCODE_SIZE * 4; @@ -1692,6 +1718,13 @@ static int si_init_microcode(struct radeon_device *rdev) break; case CHIP_OLAND: chip_name = "OLAND"; + if ((rdev->pdev->revision == 0xC7) || + (rdev->pdev->revision == 0x80) || + (rdev->pdev->revision == 0x81) || + (rdev->pdev->revision == 0x83) || + (rdev->pdev->device == 0x6604) || + (rdev->pdev->device == 0x6605)) + new_smc = true; new_chip_name = "oland"; pfp_req_size = SI_PFP_UCODE_SIZE * 4; me_req_size = SI_PM4_UCODE_SIZE * 4; @@ -1702,6 +1735,13 @@ static int si_init_microcode(struct radeon_device *rdev) break; case CHIP_HAINAN: chip_name = "HAINAN"; + if ((rdev->pdev->revision == 0x81) || + (rdev->pdev->revision == 0x83) || + (rdev->pdev->revision == 0xC3) || + (rdev->pdev->device == 0x6664) || + (rdev->pdev->device == 0x6665) || + (rdev->pdev->device == 0x6667)) + new_smc = true; new_chip_name = "hainan"; pfp_req_size = SI_PFP_UCODE_SIZE * 4; me_req_size = SI_PM4_UCODE_SIZE * 4; @@ -1847,7 +1887,10 @@ static int si_init_microcode(struct radeon_device *rdev) } } - snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", new_chip_name); + if (new_smc) + snprintf(fw_name, sizeof(fw_name), "radeon/%s_k_smc.bin", new_chip_name); + else + snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", new_chip_name); err = request_firmware(&rdev->smc_fw, fw_name, rdev->dev); if (err) { snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", chip_name); diff --git a/drivers/gpu/drm/rcar-du/Kconfig b/drivers/gpu/drm/rcar-du/Kconfig index 7fc3ca5ce6c7..4c2fd056dd6d 100644 --- a/drivers/gpu/drm/rcar-du/Kconfig +++ b/drivers/gpu/drm/rcar-du/Kconfig @@ -6,7 +6,6 @@ config DRM_RCAR_DU select DRM_KMS_HELPER select DRM_KMS_CMA_HELPER select DRM_GEM_CMA_HELPER - select DRM_KMS_FB_HELPER select VIDEOMODE_HELPERS help Choose this option if you have an R-Car chipset. diff --git a/drivers/gpu/drm/rcar-du/Makefile b/drivers/gpu/drm/rcar-du/Makefile index 827711e28226..d3b44651061a 100644 --- a/drivers/gpu/drm/rcar-du/Makefile +++ b/drivers/gpu/drm/rcar-du/Makefile @@ -7,8 +7,8 @@ rcar-du-drm-y := rcar_du_crtc.o \ rcar_du_plane.o \ rcar_du_vgacon.o -rcar-du-drm-$(CONFIG_DRM_RCAR_HDMI) += rcar_du_hdmicon.o \ - rcar_du_hdmienc.o +rcar-du-drm-$(CONFIG_DRM_RCAR_HDMI) += rcar_du_hdmienc.o + rcar-du-drm-$(CONFIG_DRM_RCAR_LVDS) += rcar_du_lvdsenc.o rcar-du-drm-$(CONFIG_DRM_RCAR_VSP) += rcar_du_vsp.o diff --git a/drivers/gpu/drm/rcar-du/rcar_du_crtc.c b/drivers/gpu/drm/rcar-du/rcar_du_crtc.c index 0d8bdda736f9..e39fcef2e033 100644 --- a/drivers/gpu/drm/rcar-du/rcar_du_crtc.c +++ b/drivers/gpu/drm/rcar-du/rcar_du_crtc.c @@ -552,7 +552,7 @@ static irqreturn_t rcar_du_crtc_irq(int irq, void *arg) rcar_du_crtc_write(rcrtc, DSRCR, status & DSRCR_MASK); if (status & DSSR_FRM) { - drm_handle_vblank(rcrtc->crtc.dev, rcrtc->index); + drm_crtc_handle_vblank(&rcrtc->crtc); rcar_du_crtc_finish_page_flip(rcrtc); ret = IRQ_HANDLED; } diff --git a/drivers/gpu/drm/rcar-du/rcar_du_encoder.c b/drivers/gpu/drm/rcar-du/rcar_du_encoder.c index 55149e9ce28e..ab8645c57e2d 100644 --- a/drivers/gpu/drm/rcar-du/rcar_du_encoder.c +++ b/drivers/gpu/drm/rcar-du/rcar_du_encoder.c @@ -19,7 +19,6 @@ #include "rcar_du_drv.h" #include "rcar_du_encoder.h" -#include "rcar_du_hdmicon.h" #include "rcar_du_hdmienc.h" #include "rcar_du_kms.h" #include "rcar_du_lvdscon.h" @@ -174,7 +173,7 @@ int rcar_du_encoder_init(struct rcar_du_device *rcdu, break; case DRM_MODE_ENCODER_TMDS: - ret = rcar_du_hdmi_connector_init(rcdu, renc); + /* connector managed by the bridge driver */ break; default: diff --git a/drivers/gpu/drm/rcar-du/rcar_du_encoder.h b/drivers/gpu/drm/rcar-du/rcar_du_encoder.h index a8669c3e0dd5..7fc10a9c34c3 100644 --- a/drivers/gpu/drm/rcar-du/rcar_du_encoder.h +++ b/drivers/gpu/drm/rcar-du/rcar_du_encoder.h @@ -15,7 +15,6 @@ #define __RCAR_DU_ENCODER_H__ #include <drm/drm_crtc.h> -#include <drm/drm_encoder_slave.h> struct rcar_du_device; struct rcar_du_hdmienc; @@ -30,16 +29,16 @@ enum rcar_du_encoder_type { }; struct rcar_du_encoder { - struct drm_encoder_slave slave; + struct drm_encoder base; enum rcar_du_output output; struct rcar_du_hdmienc *hdmi; struct rcar_du_lvdsenc *lvds; }; #define to_rcar_encoder(e) \ - container_of(e, struct rcar_du_encoder, slave.base) + container_of(e, struct rcar_du_encoder, base) -#define rcar_encoder_to_drm_encoder(e) (&(e)->slave.base) +#define rcar_encoder_to_drm_encoder(e) (&(e)->base) struct rcar_du_connector { struct drm_connector connector; diff --git a/drivers/gpu/drm/rcar-du/rcar_du_hdmicon.c b/drivers/gpu/drm/rcar-du/rcar_du_hdmicon.c deleted file mode 100644 index 612b4d5ae098..000000000000 --- a/drivers/gpu/drm/rcar-du/rcar_du_hdmicon.c +++ /dev/null @@ -1,116 +0,0 @@ -/* - * R-Car Display Unit HDMI Connector - * - * Copyright (C) 2014 Renesas Electronics Corporation - * - * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com) - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; either version 2 of the License, or - * (at your option) any later version. - */ - -#include <drm/drmP.h> -#include <drm/drm_atomic_helper.h> -#include <drm/drm_crtc.h> -#include <drm/drm_crtc_helper.h> -#include <drm/drm_encoder_slave.h> - -#include "rcar_du_drv.h" -#include "rcar_du_encoder.h" -#include "rcar_du_hdmicon.h" -#include "rcar_du_kms.h" - -#define to_slave_funcs(e) (to_rcar_encoder(e)->slave.slave_funcs) - -static int rcar_du_hdmi_connector_get_modes(struct drm_connector *connector) -{ - struct rcar_du_connector *con = to_rcar_connector(connector); - struct drm_encoder *encoder = rcar_encoder_to_drm_encoder(con->encoder); - const struct drm_encoder_slave_funcs *sfuncs = to_slave_funcs(encoder); - - if (sfuncs->get_modes == NULL) - return 0; - - return sfuncs->get_modes(encoder, connector); -} - -static int rcar_du_hdmi_connector_mode_valid(struct drm_connector *connector, - struct drm_display_mode *mode) -{ - struct rcar_du_connector *con = to_rcar_connector(connector); - struct drm_encoder *encoder = rcar_encoder_to_drm_encoder(con->encoder); - const struct drm_encoder_slave_funcs *sfuncs = to_slave_funcs(encoder); - - if (sfuncs->mode_valid == NULL) - return MODE_OK; - - return sfuncs->mode_valid(encoder, mode); -} - -static const struct drm_connector_helper_funcs connector_helper_funcs = { - .get_modes = rcar_du_hdmi_connector_get_modes, - .mode_valid = rcar_du_hdmi_connector_mode_valid, -}; - -static enum drm_connector_status -rcar_du_hdmi_connector_detect(struct drm_connector *connector, bool force) -{ - struct rcar_du_connector *con = to_rcar_connector(connector); - struct drm_encoder *encoder = rcar_encoder_to_drm_encoder(con->encoder); - const struct drm_encoder_slave_funcs *sfuncs = to_slave_funcs(encoder); - - if (sfuncs->detect == NULL) - return connector_status_unknown; - - return sfuncs->detect(encoder, connector); -} - -static const struct drm_connector_funcs connector_funcs = { - .dpms = drm_atomic_helper_connector_dpms, - .reset = drm_atomic_helper_connector_reset, - .detect = rcar_du_hdmi_connector_detect, - .fill_modes = drm_helper_probe_single_connector_modes, - .destroy = drm_connector_cleanup, - .atomic_duplicate_state = drm_atomic_helper_connector_duplicate_state, - .atomic_destroy_state = drm_atomic_helper_connector_destroy_state, -}; - -int rcar_du_hdmi_connector_init(struct rcar_du_device *rcdu, - struct rcar_du_encoder *renc) -{ - struct drm_encoder *encoder = rcar_encoder_to_drm_encoder(renc); - struct rcar_du_connector *rcon; - struct drm_connector *connector; - int ret; - - rcon = devm_kzalloc(rcdu->dev, sizeof(*rcon), GFP_KERNEL); - if (rcon == NULL) - return -ENOMEM; - - connector = &rcon->connector; - connector->display_info.width_mm = 0; - connector->display_info.height_mm = 0; - connector->interlace_allowed = true; - connector->polled = DRM_CONNECTOR_POLL_HPD; - - ret = drm_connector_init(rcdu->ddev, connector, &connector_funcs, - DRM_MODE_CONNECTOR_HDMIA); - if (ret < 0) - return ret; - - drm_connector_helper_add(connector, &connector_helper_funcs); - - connector->dpms = DRM_MODE_DPMS_OFF; - drm_object_property_set_value(&connector->base, - rcdu->ddev->mode_config.dpms_property, DRM_MODE_DPMS_OFF); - - ret = drm_mode_connector_attach_encoder(connector, encoder); - if (ret < 0) - return ret; - - rcon->encoder = renc; - - return 0; -} diff --git a/drivers/gpu/drm/rcar-du/rcar_du_hdmicon.h b/drivers/gpu/drm/rcar-du/rcar_du_hdmicon.h deleted file mode 100644 index 87daa949227f..000000000000 --- a/drivers/gpu/drm/rcar-du/rcar_du_hdmicon.h +++ /dev/null @@ -1,31 +0,0 @@ -/* - * R-Car Display Unit HDMI Connector - * - * Copyright (C) 2014 Renesas Electronics Corporation - * - * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com) - * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; either version 2 of the License, or - * (at your option) any later version. - */ - -#ifndef __RCAR_DU_HDMICON_H__ -#define __RCAR_DU_HDMICON_H__ - -struct rcar_du_device; -struct rcar_du_encoder; - -#if IS_ENABLED(CONFIG_DRM_RCAR_HDMI) -int rcar_du_hdmi_connector_init(struct rcar_du_device *rcdu, - struct rcar_du_encoder *renc); -#else -static inline int rcar_du_hdmi_connector_init(struct rcar_du_device *rcdu, - struct rcar_du_encoder *renc) -{ - return -ENOSYS; -} -#endif - -#endif /* __RCAR_DU_HDMICON_H__ */ diff --git a/drivers/gpu/drm/rcar-du/rcar_du_hdmienc.c b/drivers/gpu/drm/rcar-du/rcar_du_hdmienc.c index 461662d231e2..4de3ff0dbebd 100644 --- a/drivers/gpu/drm/rcar-du/rcar_du_hdmienc.c +++ b/drivers/gpu/drm/rcar-du/rcar_du_hdmienc.c @@ -16,7 +16,6 @@ #include <drm/drmP.h> #include <drm/drm_crtc.h> #include <drm/drm_crtc_helper.h> -#include <drm/drm_encoder_slave.h> #include "rcar_du_drv.h" #include "rcar_du_encoder.h" @@ -25,20 +24,14 @@ struct rcar_du_hdmienc { struct rcar_du_encoder *renc; - struct device *dev; bool enabled; }; #define to_rcar_hdmienc(e) (to_rcar_encoder(e)->hdmi) -#define to_slave_funcs(e) (to_rcar_encoder(e)->slave.slave_funcs) static void rcar_du_hdmienc_disable(struct drm_encoder *encoder) { struct rcar_du_hdmienc *hdmienc = to_rcar_hdmienc(encoder); - const struct drm_encoder_slave_funcs *sfuncs = to_slave_funcs(encoder); - - if (sfuncs->dpms) - sfuncs->dpms(encoder, DRM_MODE_DPMS_OFF); if (hdmienc->renc->lvds) rcar_du_lvdsenc_enable(hdmienc->renc->lvds, encoder->crtc, @@ -50,15 +43,11 @@ static void rcar_du_hdmienc_disable(struct drm_encoder *encoder) static void rcar_du_hdmienc_enable(struct drm_encoder *encoder) { struct rcar_du_hdmienc *hdmienc = to_rcar_hdmienc(encoder); - const struct drm_encoder_slave_funcs *sfuncs = to_slave_funcs(encoder); if (hdmienc->renc->lvds) rcar_du_lvdsenc_enable(hdmienc->renc->lvds, encoder->crtc, true); - if (sfuncs->dpms) - sfuncs->dpms(encoder, DRM_MODE_DPMS_ON); - hdmienc->enabled = true; } @@ -67,29 +56,21 @@ static int rcar_du_hdmienc_atomic_check(struct drm_encoder *encoder, struct drm_connector_state *conn_state) { struct rcar_du_hdmienc *hdmienc = to_rcar_hdmienc(encoder); - const struct drm_encoder_slave_funcs *sfuncs = to_slave_funcs(encoder); struct drm_display_mode *adjusted_mode = &crtc_state->adjusted_mode; - const struct drm_display_mode *mode = &crtc_state->mode; if (hdmienc->renc->lvds) rcar_du_lvdsenc_atomic_check(hdmienc->renc->lvds, adjusted_mode); - if (sfuncs->mode_fixup == NULL) - return 0; - - return sfuncs->mode_fixup(encoder, mode, adjusted_mode) ? 0 : -EINVAL; + return 0; } + static void rcar_du_hdmienc_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode, struct drm_display_mode *adjusted_mode) { struct rcar_du_hdmienc *hdmienc = to_rcar_hdmienc(encoder); - const struct drm_encoder_slave_funcs *sfuncs = to_slave_funcs(encoder); - - if (sfuncs->mode_set) - sfuncs->mode_set(encoder, mode, adjusted_mode); rcar_du_crtc_route_output(encoder->crtc, hdmienc->renc->output); } @@ -109,7 +90,6 @@ static void rcar_du_hdmienc_cleanup(struct drm_encoder *encoder) rcar_du_hdmienc_disable(encoder); drm_encoder_cleanup(encoder); - put_device(hdmienc->dev); } static const struct drm_encoder_funcs encoder_funcs = { @@ -120,8 +100,7 @@ int rcar_du_hdmienc_init(struct rcar_du_device *rcdu, struct rcar_du_encoder *renc, struct device_node *np) { struct drm_encoder *encoder = rcar_encoder_to_drm_encoder(renc); - struct drm_i2c_encoder_driver *driver; - struct i2c_client *i2c_slave; + struct drm_bridge *bridge; struct rcar_du_hdmienc *hdmienc; int ret; @@ -129,44 +108,29 @@ int rcar_du_hdmienc_init(struct rcar_du_device *rcdu, if (hdmienc == NULL) return -ENOMEM; - /* Locate the slave I2C device and driver. */ - i2c_slave = of_find_i2c_device_by_node(np); - if (!i2c_slave || !i2c_get_clientdata(i2c_slave)) { - dev_dbg(rcdu->dev, - "can't get I2C slave for %s, deferring probe\n", - of_node_full_name(np)); + /* Locate drm bridge from the hdmi encoder DT node */ + bridge = of_drm_find_bridge(np); + if (!bridge) return -EPROBE_DEFER; - } - - hdmienc->dev = &i2c_slave->dev; - - if (hdmienc->dev->driver == NULL) { - dev_dbg(rcdu->dev, - "I2C slave %s not probed yet, deferring probe\n", - dev_name(hdmienc->dev)); - ret = -EPROBE_DEFER; - goto error; - } - - /* Initialize the slave encoder. */ - driver = to_drm_i2c_encoder_driver(to_i2c_driver(hdmienc->dev->driver)); - ret = driver->encoder_init(i2c_slave, rcdu->ddev, &renc->slave); - if (ret < 0) - goto error; ret = drm_encoder_init(rcdu->ddev, encoder, &encoder_funcs, DRM_MODE_ENCODER_TMDS, NULL); if (ret < 0) - goto error; + return ret; drm_encoder_helper_add(encoder, &encoder_helper_funcs); renc->hdmi = hdmienc; hdmienc->renc = renc; - return 0; + /* Link drm_bridge to encoder */ + bridge->encoder = encoder; + + ret = drm_bridge_attach(rcdu->ddev, bridge); + if (ret) { + drm_encoder_cleanup(encoder); + return ret; + } -error: - put_device(hdmienc->dev); - return ret; + return 0; } diff --git a/drivers/gpu/drm/rockchip/Kconfig b/drivers/gpu/drm/rockchip/Kconfig index e48611e83c03..3c58669a06ce 100644 --- a/drivers/gpu/drm/rockchip/Kconfig +++ b/drivers/gpu/drm/rockchip/Kconfig @@ -4,11 +4,7 @@ config DRM_ROCKCHIP depends on RESET_CONTROLLER select DRM_GEM_CMA_HELPER select DRM_KMS_HELPER - select DRM_KMS_FB_HELPER select DRM_PANEL - select FB_CFB_FILLRECT - select FB_CFB_COPYAREA - select FB_CFB_IMAGEBLIT select VIDEOMODE_HELPERS help Choose this option if you have a Rockchip soc chipset. diff --git a/drivers/gpu/drm/rockchip/analogix_dp-rockchip.c b/drivers/gpu/drm/rockchip/analogix_dp-rockchip.c index c120172add5c..89aadbf465f8 100644 --- a/drivers/gpu/drm/rockchip/analogix_dp-rockchip.c +++ b/drivers/gpu/drm/rockchip/analogix_dp-rockchip.c @@ -14,6 +14,7 @@ #include <linux/component.h> #include <linux/mfd/syscon.h> +#include <linux/of_device.h> #include <linux/of_graph.h> #include <linux/regmap.h> #include <linux/reset.h> @@ -33,13 +34,28 @@ #include "rockchip_drm_drv.h" #include "rockchip_drm_vop.h" +#define RK3288_GRF_SOC_CON6 0x25c +#define RK3288_EDP_LCDC_SEL BIT(5) +#define RK3399_GRF_SOC_CON20 0x6250 +#define RK3399_EDP_LCDC_SEL BIT(5) + +#define HIWORD_UPDATE(val, mask) (val | (mask) << 16) + #define to_dp(nm) container_of(nm, struct rockchip_dp_device, nm) -/* dp grf register offset */ -#define GRF_SOC_CON6 0x025c -#define GRF_EDP_LCD_SEL_MASK BIT(5) -#define GRF_EDP_SEL_VOP_LIT BIT(5) -#define GRF_EDP_SEL_VOP_BIG 0 +/** + * struct rockchip_dp_chip_data - splite the grf setting of kind of chips + * @lcdsel_grf_reg: grf register offset of lcdc select + * @lcdsel_big: reg value of selecting vop big for eDP + * @lcdsel_lit: reg value of selecting vop little for eDP + * @chip_type: specific chip type + */ +struct rockchip_dp_chip_data { + u32 lcdsel_grf_reg; + u32 lcdsel_big; + u32 lcdsel_lit; + u32 chip_type; +}; struct rockchip_dp_device { struct drm_device *drm_dev; @@ -48,9 +64,12 @@ struct rockchip_dp_device { struct drm_display_mode mode; struct clk *pclk; + struct clk *grfclk; struct regmap *grf; struct reset_control *rst; + const struct rockchip_dp_chip_data *data; + struct analogix_dp_plat_data plat_data; }; @@ -77,6 +96,7 @@ static int rockchip_dp_poweron(struct analogix_dp_plat_data *plat_data) ret = rockchip_dp_pre_init(dp); if (ret < 0) { dev_err(dp->dev, "failed to dp pre init %d\n", ret); + clk_disable_unprepare(dp->pclk); return ret; } @@ -92,6 +112,23 @@ static int rockchip_dp_powerdown(struct analogix_dp_plat_data *plat_data) return 0; } +static int rockchip_dp_get_modes(struct analogix_dp_plat_data *plat_data, + struct drm_connector *connector) +{ + struct drm_display_info *di = &connector->display_info; + /* VOP couldn't output YUV video format for eDP rightly */ + u32 mask = DRM_COLOR_FORMAT_YCRCB444 | DRM_COLOR_FORMAT_YCRCB422; + + if ((di->color_formats & mask)) { + DRM_DEBUG_KMS("Swapping display color format from YUV to RGB\n"); + di->color_formats &= ~mask; + di->color_formats |= DRM_COLOR_FORMAT_RGB444; + di->bpc = 8; + } + + return 0; +} + static bool rockchip_dp_drm_encoder_mode_fixup(struct drm_encoder *encoder, const struct drm_display_mode *mode, @@ -119,17 +156,23 @@ static void rockchip_dp_drm_encoder_enable(struct drm_encoder *encoder) return; if (ret) - val = GRF_EDP_SEL_VOP_LIT | (GRF_EDP_LCD_SEL_MASK << 16); + val = dp->data->lcdsel_lit; else - val = GRF_EDP_SEL_VOP_BIG | (GRF_EDP_LCD_SEL_MASK << 16); + val = dp->data->lcdsel_big; dev_dbg(dp->dev, "vop %s output to dp\n", (ret) ? "LIT" : "BIG"); - ret = regmap_write(dp->grf, GRF_SOC_CON6, val); - if (ret != 0) { - dev_err(dp->dev, "Could not write to GRF: %d\n", ret); + ret = clk_prepare_enable(dp->grfclk); + if (ret < 0) { + dev_err(dp->dev, "failed to enable grfclk %d\n", ret); return; } + + ret = regmap_write(dp->grf, dp->data->lcdsel_grf_reg, val); + if (ret != 0) + dev_err(dp->dev, "Could not write to GRF: %d\n", ret); + + clk_disable_unprepare(dp->grfclk); } static void rockchip_dp_drm_encoder_nop(struct drm_encoder *encoder) @@ -143,22 +186,29 @@ rockchip_dp_drm_encoder_atomic_check(struct drm_encoder *encoder, struct drm_connector_state *conn_state) { struct rockchip_crtc_state *s = to_rockchip_crtc_state(crtc_state); + struct rockchip_dp_device *dp = to_dp(encoder); + int ret; /* - * FIXME(Yakir): driver should configure the CRTC output video - * mode with the display information which indicated the monitor - * support colorimetry. - * - * But don't know why the CRTC driver seems could only output the - * RGBaaa rightly. For example, if connect the "innolux,n116bge" - * eDP screen, EDID would indicated that screen only accepted the - * 6bpc mode. But if I configure CRTC to RGB666 output, then eDP - * screen would show a blue picture (RGB888 show a green picture). - * But if I configure CTRC to RGBaaa, and eDP driver still keep - * RGB666 input video mode, then screen would works prefect. + * The hardware IC designed that VOP must output the RGB10 video + * format to eDP controller, and if eDP panel only support RGB8, + * then eDP controller should cut down the video data, not via VOP + * controller, that's why we need to hardcode the VOP output mode + * to RGA10 here. */ + s->output_mode = ROCKCHIP_OUT_MODE_AAAA; s->output_type = DRM_MODE_CONNECTOR_eDP; + if (dp->data->chip_type == RK3399_EDP) { + /* + * For RK3399, VOP Lit must code the out mode to RGB888, + * VOP Big must code the out mode to RGB10. + */ + ret = drm_of_encoder_active_endpoint_id(dp->dev->of_node, + encoder); + if (ret > 0) + s->output_mode = ROCKCHIP_OUT_MODE_P888; + } return 0; } @@ -192,6 +242,16 @@ static int rockchip_dp_init(struct rockchip_dp_device *dp) return PTR_ERR(dp->grf); } + dp->grfclk = devm_clk_get(dev, "grf"); + if (PTR_ERR(dp->grfclk) == -ENOENT) { + dp->grfclk = NULL; + } else if (PTR_ERR(dp->grfclk) == -EPROBE_DEFER) { + return -EPROBE_DEFER; + } else if (IS_ERR(dp->grfclk)) { + dev_err(dev, "failed to get grf clock\n"); + return PTR_ERR(dp->grfclk); + } + dp->pclk = devm_clk_get(dev, "pclk"); if (IS_ERR(dp->pclk)) { dev_err(dev, "failed to get pclk property\n"); @@ -213,6 +273,7 @@ static int rockchip_dp_init(struct rockchip_dp_device *dp) ret = rockchip_dp_pre_init(dp); if (ret < 0) { dev_err(dp->dev, "failed to pre init %d\n", ret); + clk_disable_unprepare(dp->pclk); return ret; } @@ -246,6 +307,7 @@ static int rockchip_dp_bind(struct device *dev, struct device *master, void *data) { struct rockchip_dp_device *dp = dev_get_drvdata(dev); + const struct rockchip_dp_chip_data *dp_data; struct drm_device *drm_dev = data; int ret; @@ -256,10 +318,15 @@ static int rockchip_dp_bind(struct device *dev, struct device *master, */ dev_set_drvdata(dev, NULL); + dp_data = of_device_get_match_data(dev); + if (!dp_data) + return -ENODEV; + ret = rockchip_dp_init(dp); if (ret < 0) return ret; + dp->data = dp_data; dp->drm_dev = drm_dev; ret = rockchip_dp_drm_create_encoder(dp); @@ -270,9 +337,10 @@ static int rockchip_dp_bind(struct device *dev, struct device *master, dp->plat_data.encoder = &dp->encoder; - dp->plat_data.dev_type = RK3288_DP; + dp->plat_data.dev_type = dp->data->chip_type; dp->plat_data.power_on = rockchip_dp_poweron; dp->plat_data.power_off = rockchip_dp_powerdown; + dp->plat_data.get_modes = rockchip_dp_get_modes; return analogix_dp_bind(dev, dp->drm_dev, &dp->plat_data); } @@ -292,38 +360,33 @@ static int rockchip_dp_probe(struct platform_device *pdev) { struct device *dev = &pdev->dev; struct device_node *panel_node, *port, *endpoint; + struct drm_panel *panel = NULL; struct rockchip_dp_device *dp; - struct drm_panel *panel; port = of_graph_get_port_by_id(dev->of_node, 1); - if (!port) { - dev_err(dev, "can't find output port\n"); - return -EINVAL; - } - - endpoint = of_get_child_by_name(port, "endpoint"); - of_node_put(port); - if (!endpoint) { - dev_err(dev, "no output endpoint found\n"); - return -EINVAL; - } - - panel_node = of_graph_get_remote_port_parent(endpoint); - of_node_put(endpoint); - if (!panel_node) { - dev_err(dev, "no output node found\n"); - return -EINVAL; - } - - panel = of_drm_find_panel(panel_node); - if (!panel) { - DRM_ERROR("failed to find panel\n"); + if (port) { + endpoint = of_get_child_by_name(port, "endpoint"); + of_node_put(port); + if (!endpoint) { + dev_err(dev, "no output endpoint found\n"); + return -EINVAL; + } + + panel_node = of_graph_get_remote_port_parent(endpoint); + of_node_put(endpoint); + if (!panel_node) { + dev_err(dev, "no output node found\n"); + return -EINVAL; + } + + panel = of_drm_find_panel(panel_node); of_node_put(panel_node); - return -EPROBE_DEFER; + if (!panel) { + DRM_ERROR("failed to find panel\n"); + return -EPROBE_DEFER; + } } - of_node_put(panel_node); - dp = devm_kzalloc(dev, sizeof(*dp), GFP_KERNEL); if (!dp) return -ENOMEM; @@ -356,8 +419,23 @@ static const struct dev_pm_ops rockchip_dp_pm_ops = { #endif }; +static const struct rockchip_dp_chip_data rk3399_edp = { + .lcdsel_grf_reg = RK3399_GRF_SOC_CON20, + .lcdsel_big = HIWORD_UPDATE(0, RK3399_EDP_LCDC_SEL), + .lcdsel_lit = HIWORD_UPDATE(RK3399_EDP_LCDC_SEL, RK3399_EDP_LCDC_SEL), + .chip_type = RK3399_EDP, +}; + +static const struct rockchip_dp_chip_data rk3288_dp = { + .lcdsel_grf_reg = RK3288_GRF_SOC_CON6, + .lcdsel_big = HIWORD_UPDATE(0, RK3288_EDP_LCDC_SEL), + .lcdsel_lit = HIWORD_UPDATE(RK3288_EDP_LCDC_SEL, RK3288_EDP_LCDC_SEL), + .chip_type = RK3288_DP, +}; + static const struct of_device_id rockchip_dp_dt_ids[] = { - {.compatible = "rockchip,rk3288-dp",}, + {.compatible = "rockchip,rk3288-dp", .data = &rk3288_dp }, + {.compatible = "rockchip,rk3399-edp", .data = &rk3399_edp }, {} }; MODULE_DEVICE_TABLE(of, rockchip_dp_dt_ids); diff --git a/drivers/gpu/drm/rockchip/dw_hdmi-rockchip.c b/drivers/gpu/drm/rockchip/dw_hdmi-rockchip.c index 801110f65a63..0665fb915579 100644 --- a/drivers/gpu/drm/rockchip/dw_hdmi-rockchip.c +++ b/drivers/gpu/drm/rockchip/dw_hdmi-rockchip.c @@ -15,7 +15,6 @@ #include <drm/drmP.h> #include <drm/drm_crtc_helper.h> #include <drm/drm_edid.h> -#include <drm/drm_encoder_slave.h> #include <drm/bridge/dw_hdmi.h> #include "rockchip_drm_drv.h" diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_drv.c b/drivers/gpu/drm/rockchip/rockchip_drm_drv.c index d665fb04d264..a822d49a255a 100644 --- a/drivers/gpu/drm/rockchip/rockchip_drm_drv.c +++ b/drivers/gpu/drm/rockchip/rockchip_drm_drv.c @@ -79,7 +79,7 @@ int rockchip_register_crtc_funcs(struct drm_crtc *crtc, int pipe = drm_crtc_index(crtc); struct rockchip_drm_private *priv = crtc->dev->dev_private; - if (pipe > ROCKCHIP_MAX_CRTC) + if (pipe >= ROCKCHIP_MAX_CRTC) return -EINVAL; priv->crtc_funcs[pipe] = crtc_funcs; @@ -92,7 +92,7 @@ void rockchip_unregister_crtc_funcs(struct drm_crtc *crtc) int pipe = drm_crtc_index(crtc); struct rockchip_drm_private *priv = crtc->dev->dev_private; - if (pipe > ROCKCHIP_MAX_CRTC) + if (pipe >= ROCKCHIP_MAX_CRTC) return; priv->crtc_funcs[pipe] = NULL; @@ -257,7 +257,7 @@ static void rockchip_drm_unbind(struct device *dev) dev_set_drvdata(dev, NULL); } -void rockchip_drm_lastclose(struct drm_device *dev) +static void rockchip_drm_lastclose(struct drm_device *dev) { struct rockchip_drm_private *priv = dev->dev_private; @@ -433,6 +433,7 @@ static int rockchip_drm_platform_probe(struct platform_device *pdev) is_support_iommu = false; } + of_node_put(iommu); component_match_add(dev, &match, compare_of, port->parent); of_node_put(port); } diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_fb.c b/drivers/gpu/drm/rockchip/rockchip_drm_fb.c index 20f12bc5a386..55c52734c52d 100644 --- a/drivers/gpu/drm/rockchip/rockchip_drm_fb.c +++ b/drivers/gpu/drm/rockchip/rockchip_drm_fb.c @@ -20,6 +20,7 @@ #include <drm/drm_crtc_helper.h> #include "rockchip_drm_drv.h" +#include "rockchip_drm_fb.h" #include "rockchip_drm_gem.h" #define to_rockchip_fb(x) container_of(x, struct rockchip_drm_fb, fb) @@ -43,14 +44,10 @@ struct drm_gem_object *rockchip_fb_get_gem_obj(struct drm_framebuffer *fb, static void rockchip_drm_fb_destroy(struct drm_framebuffer *fb) { struct rockchip_drm_fb *rockchip_fb = to_rockchip_fb(fb); - struct drm_gem_object *obj; int i; - for (i = 0; i < ROCKCHIP_MAX_FB_BUFFER; i++) { - obj = rockchip_fb->obj[i]; - if (obj) - drm_gem_object_unreference_unlocked(obj); - } + for (i = 0; i < ROCKCHIP_MAX_FB_BUFFER; i++) + drm_gem_object_unreference_unlocked(rockchip_fb->obj[i]); drm_framebuffer_cleanup(fb); kfree(rockchip_fb); @@ -245,7 +242,7 @@ rockchip_atomic_commit_tail(struct drm_atomic_state *state) drm_atomic_helper_cleanup_planes(dev, state); } -struct drm_mode_config_helper_funcs rockchip_mode_config_helpers = { +static struct drm_mode_config_helper_funcs rockchip_mode_config_helpers = { .atomic_commit_tail = rockchip_atomic_commit_tail, }; diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_vop.c b/drivers/gpu/drm/rockchip/rockchip_drm_vop.c index 6255e5bcd954..91305eb7d312 100644 --- a/drivers/gpu/drm/rockchip/rockchip_drm_vop.c +++ b/drivers/gpu/drm/rockchip/rockchip_drm_vop.c @@ -328,9 +328,9 @@ static void scl_vop_cal_scl_fac(struct vop *vop, const struct vop_win_data *win, scl_cal_scale2(src_h, dst_h)); if (is_yuv) { VOP_SCL_SET(vop, win, scale_cbcr_x, - scl_cal_scale2(src_w, dst_w)); + scl_cal_scale2(cbcr_src_w, dst_w)); VOP_SCL_SET(vop, win, scale_cbcr_y, - scl_cal_scale2(src_h, dst_h)); + scl_cal_scale2(cbcr_src_h, dst_h)); } return; } @@ -798,7 +798,7 @@ static const struct drm_plane_helper_funcs plane_helper_funcs = { .atomic_disable = vop_plane_atomic_disable, }; -void vop_atomic_plane_reset(struct drm_plane *plane) +static void vop_atomic_plane_reset(struct drm_plane *plane) { struct vop_plane_state *vop_plane_state = to_vop_plane_state(plane->state); @@ -815,7 +815,7 @@ void vop_atomic_plane_reset(struct drm_plane *plane) plane->state->plane = plane; } -struct drm_plane_state * +static struct drm_plane_state * vop_atomic_plane_duplicate_state(struct drm_plane *plane) { struct vop_plane_state *old_vop_plane_state; @@ -1052,6 +1052,17 @@ static void vop_crtc_destroy(struct drm_crtc *crtc) drm_crtc_cleanup(crtc); } +static void vop_crtc_reset(struct drm_crtc *crtc) +{ + if (crtc->state) + __drm_atomic_helper_crtc_destroy_state(crtc->state); + kfree(crtc->state); + + crtc->state = kzalloc(sizeof(struct rockchip_crtc_state), GFP_KERNEL); + if (crtc->state) + crtc->state->crtc = crtc; +} + static struct drm_crtc_state *vop_crtc_duplicate_state(struct drm_crtc *crtc) { struct rockchip_crtc_state *rockchip_state; @@ -1077,7 +1088,7 @@ static const struct drm_crtc_funcs vop_crtc_funcs = { .set_config = drm_atomic_helper_set_config, .page_flip = drm_atomic_helper_page_flip, .destroy = vop_crtc_destroy, - .reset = drm_atomic_helper_crtc_reset, + .reset = vop_crtc_reset, .atomic_duplicate_state = vop_crtc_duplicate_state, .atomic_destroy_state = vop_crtc_destroy_state, }; diff --git a/drivers/gpu/drm/rockchip/rockchip_vop_reg.c b/drivers/gpu/drm/rockchip/rockchip_vop_reg.c index 3166b46a5893..919992cdc97e 100644 --- a/drivers/gpu/drm/rockchip/rockchip_vop_reg.c +++ b/drivers/gpu/drm/rockchip/rockchip_vop_reg.c @@ -190,7 +190,7 @@ static const struct vop_data rk3288_vop = { .win_size = ARRAY_SIZE(rk3288_vop_win_data), }; -static const struct vop_scl_regs rk3066_win_scl = { +static const struct vop_scl_regs rk3036_win_scl = { .scale_yrgb_x = VOP_REG(RK3036_WIN0_SCL_FACTOR_YRGB, 0xffff, 0x0), .scale_yrgb_y = VOP_REG(RK3036_WIN0_SCL_FACTOR_YRGB, 0xffff, 16), .scale_cbcr_x = VOP_REG(RK3036_WIN0_SCL_FACTOR_CBR, 0xffff, 0x0), @@ -198,7 +198,7 @@ static const struct vop_scl_regs rk3066_win_scl = { }; static const struct vop_win_phy rk3036_win0_data = { - .scl = &rk3066_win_scl, + .scl = &rk3036_win_scl, .data_formats = formats_win_full, .nformats = ARRAY_SIZE(formats_win_full), .enable = VOP_REG(RK3036_SYS_CTRL, 0x1, 0), @@ -210,6 +210,7 @@ static const struct vop_win_phy rk3036_win0_data = { .yrgb_mst = VOP_REG(RK3036_WIN0_YRGB_MST, 0xffffffff, 0), .uv_mst = VOP_REG(RK3036_WIN0_CBR_MST, 0xffffffff, 0), .yrgb_vir = VOP_REG(RK3036_WIN0_VIR, 0xffff, 0), + .uv_vir = VOP_REG(RK3036_WIN0_VIR, 0x1fff, 16), }; static const struct vop_win_phy rk3036_win1_data = { @@ -299,7 +300,7 @@ static int vop_remove(struct platform_device *pdev) return 0; } -struct platform_driver vop_platform_driver = { +static struct platform_driver vop_platform_driver = { .probe = vop_probe, .remove = vop_remove, .driver = { diff --git a/drivers/gpu/drm/shmobile/Kconfig b/drivers/gpu/drm/shmobile/Kconfig index 8d17d00ddb4b..c987c826daa3 100644 --- a/drivers/gpu/drm/shmobile/Kconfig +++ b/drivers/gpu/drm/shmobile/Kconfig @@ -6,7 +6,6 @@ config DRM_SHMOBILE select BACKLIGHT_CLASS_DEVICE select BACKLIGHT_LCD_SUPPORT select DRM_KMS_HELPER - select DRM_KMS_FB_HELPER select DRM_KMS_CMA_HELPER select DRM_GEM_CMA_HELPER help diff --git a/drivers/gpu/drm/sti/Kconfig b/drivers/gpu/drm/sti/Kconfig index 5ad43a1bb260..494ab257f77c 100644 --- a/drivers/gpu/drm/sti/Kconfig +++ b/drivers/gpu/drm/sti/Kconfig @@ -7,5 +7,6 @@ config DRM_STI select DRM_KMS_CMA_HELPER select DRM_PANEL select FW_LOADER + select SND_SOC_HDMI_CODEC if SND_SOC help Choose this option to enable DRM on STM stiH41x chipset diff --git a/drivers/gpu/drm/sti/sti_awg_utils.c b/drivers/gpu/drm/sti/sti_awg_utils.c index a516eb869f6f..2da7d6866d5d 100644 --- a/drivers/gpu/drm/sti/sti_awg_utils.c +++ b/drivers/gpu/drm/sti/sti_awg_utils.c @@ -6,6 +6,8 @@ #include "sti_awg_utils.h" +#define AWG_DELAY (-5) + #define AWG_OPCODE_OFFSET 10 #define AWG_MAX_ARG 0x3ff @@ -125,7 +127,7 @@ static int awg_generate_line_signal( val = timing->blanking_level; ret |= awg_generate_instr(RPLSET, val, 0, 0, fwparams); - val = timing->trailing_pixels - 1; + val = timing->trailing_pixels - 1 + AWG_DELAY; ret |= awg_generate_instr(SKIP, val, 0, 0, fwparams); } diff --git a/drivers/gpu/drm/sti/sti_compositor.c b/drivers/gpu/drm/sti/sti_compositor.c index 794148ff0e57..134201ecc6fd 100644 --- a/drivers/gpu/drm/sti/sti_compositor.c +++ b/drivers/gpu/drm/sti/sti_compositor.c @@ -254,12 +254,12 @@ static int sti_compositor_probe(struct platform_device *pdev) } /* Get reset resources */ - compo->rst_main = devm_reset_control_get(dev, "compo-main"); + compo->rst_main = devm_reset_control_get_shared(dev, "compo-main"); /* Take compo main out of reset */ if (!IS_ERR(compo->rst_main)) reset_control_deassert(compo->rst_main); - compo->rst_aux = devm_reset_control_get(dev, "compo-aux"); + compo->rst_aux = devm_reset_control_get_shared(dev, "compo-aux"); /* Take compo aux out of reset */ if (!IS_ERR(compo->rst_aux)) reset_control_deassert(compo->rst_aux); @@ -267,10 +267,12 @@ static int sti_compositor_probe(struct platform_device *pdev) vtg_np = of_parse_phandle(pdev->dev.of_node, "st,vtg", 0); if (vtg_np) compo->vtg_main = of_vtg_find(vtg_np); + of_node_put(vtg_np); vtg_np = of_parse_phandle(pdev->dev.of_node, "st,vtg", 1); if (vtg_np) compo->vtg_aux = of_vtg_find(vtg_np); + of_node_put(vtg_np); platform_set_drvdata(pdev, compo); diff --git a/drivers/gpu/drm/sti/sti_crtc.c b/drivers/gpu/drm/sti/sti_crtc.c index 7fab3af7473b..c7d734dc3cf4 100644 --- a/drivers/gpu/drm/sti/sti_crtc.c +++ b/drivers/gpu/drm/sti/sti_crtc.c @@ -23,22 +23,11 @@ static void sti_crtc_enable(struct drm_crtc *crtc) { struct sti_mixer *mixer = to_sti_mixer(crtc); - struct device *dev = mixer->dev; - struct sti_compositor *compo = dev_get_drvdata(dev); DRM_DEBUG_DRIVER("\n"); mixer->status = STI_MIXER_READY; - /* Prepare and enable the compo IP clock */ - if (mixer->id == STI_MIXER_MAIN) { - if (clk_prepare_enable(compo->clk_compo_main)) - DRM_INFO("Failed to prepare/enable compo_main clk\n"); - } else { - if (clk_prepare_enable(compo->clk_compo_aux)) - DRM_INFO("Failed to prepare/enable compo_aux clk\n"); - } - drm_crtc_vblank_on(crtc); } @@ -57,9 +46,8 @@ sti_crtc_mode_set(struct drm_crtc *crtc, struct drm_display_mode *mode) struct sti_mixer *mixer = to_sti_mixer(crtc); struct device *dev = mixer->dev; struct sti_compositor *compo = dev_get_drvdata(dev); - struct clk *clk; + struct clk *compo_clk, *pix_clk; int rate = mode->clock * 1000; - int res; DRM_DEBUG_KMS("CRTC:%d (%s) mode:%d (%s)\n", crtc->base.id, sti_mixer_to_str(mixer), @@ -74,32 +62,46 @@ sti_crtc_mode_set(struct drm_crtc *crtc, struct drm_display_mode *mode) mode->vsync_start, mode->vsync_end, mode->vtotal, mode->type, mode->flags); - /* Set rate and prepare/enable pixel clock */ - if (mixer->id == STI_MIXER_MAIN) - clk = compo->clk_pix_main; - else - clk = compo->clk_pix_aux; + if (mixer->id == STI_MIXER_MAIN) { + compo_clk = compo->clk_compo_main; + pix_clk = compo->clk_pix_main; + } else { + compo_clk = compo->clk_compo_aux; + pix_clk = compo->clk_pix_aux; + } + + /* Prepare and enable the compo IP clock */ + if (clk_prepare_enable(compo_clk)) { + DRM_INFO("Failed to prepare/enable compositor clk\n"); + goto compo_error; + } - res = clk_set_rate(clk, rate); - if (res < 0) { + /* Set rate and prepare/enable pixel clock */ + if (clk_set_rate(pix_clk, rate) < 0) { DRM_ERROR("Cannot set rate (%dHz) for pix clk\n", rate); - return -EINVAL; + goto pix_error; } - if (clk_prepare_enable(clk)) { + if (clk_prepare_enable(pix_clk)) { DRM_ERROR("Failed to prepare/enable pix clk\n"); - return -EINVAL; + goto pix_error; } sti_vtg_set_config(mixer->id == STI_MIXER_MAIN ? compo->vtg_main : compo->vtg_aux, &crtc->mode); - res = sti_mixer_active_video_area(mixer, &crtc->mode); - if (res) { + if (sti_mixer_active_video_area(mixer, &crtc->mode)) { DRM_ERROR("Can't set active video area\n"); - return -EINVAL; + goto mixer_error; } - return res; + return 0; + +mixer_error: + clk_disable_unprepare(pix_clk); +pix_error: + clk_disable_unprepare(compo_clk); +compo_error: + return -EINVAL; } static void sti_crtc_disable(struct drm_crtc *crtc) @@ -130,7 +132,6 @@ static void sti_crtc_disable(struct drm_crtc *crtc) static void sti_crtc_mode_set_nofb(struct drm_crtc *crtc) { - sti_crtc_enable(crtc); sti_crtc_mode_set(crtc, &crtc->state->adjusted_mode); } @@ -221,9 +222,7 @@ static void sti_crtc_atomic_flush(struct drm_crtc *crtc, static const struct drm_crtc_helper_funcs sti_crtc_helper_funcs = { .enable = sti_crtc_enable, .disable = sti_crtc_disabling, - .mode_set = drm_helper_crtc_mode_set, .mode_set_nofb = sti_crtc_mode_set_nofb, - .mode_set_base = drm_helper_crtc_mode_set_base, .atomic_begin = sti_crtc_atomic_begin, .atomic_flush = sti_crtc_atomic_flush, }; diff --git a/drivers/gpu/drm/sti/sti_dvo.c b/drivers/gpu/drm/sti/sti_dvo.c index ec3108074350..00881eb4536e 100644 --- a/drivers/gpu/drm/sti/sti_dvo.c +++ b/drivers/gpu/drm/sti/sti_dvo.c @@ -580,6 +580,7 @@ static int sti_dvo_probe(struct platform_device *pdev) dvo->panel_node = of_parse_phandle(np, "sti,panel", 0); if (!dvo->panel_node) DRM_ERROR("No panel associated to the dvo output\n"); + of_node_put(dvo->panel_node); platform_set_drvdata(pdev, dvo); diff --git a/drivers/gpu/drm/sti/sti_hdmi.c b/drivers/gpu/drm/sti/sti_hdmi.c index 8d1402b245bf..fedc17f98d9b 100644 --- a/drivers/gpu/drm/sti/sti_hdmi.c +++ b/drivers/gpu/drm/sti/sti_hdmi.c @@ -18,6 +18,8 @@ #include <drm/drm_crtc_helper.h> #include <drm/drm_edid.h> +#include <sound/hdmi-codec.h> + #include "sti_hdmi.h" #include "sti_hdmi_tx3g4c28phy.h" #include "sti_hdmi_tx3g0c55phy.h" @@ -35,6 +37,8 @@ #define HDMI_DFLT_CHL0_DAT 0x0110 #define HDMI_DFLT_CHL1_DAT 0x0114 #define HDMI_DFLT_CHL2_DAT 0x0118 +#define HDMI_AUDIO_CFG 0x0200 +#define HDMI_SPDIF_FIFO_STATUS 0x0204 #define HDMI_SW_DI_1_HEAD_WORD 0x0210 #define HDMI_SW_DI_1_PKT_WORD0 0x0214 #define HDMI_SW_DI_1_PKT_WORD1 0x0218 @@ -44,6 +48,9 @@ #define HDMI_SW_DI_1_PKT_WORD5 0x0228 #define HDMI_SW_DI_1_PKT_WORD6 0x022C #define HDMI_SW_DI_CFG 0x0230 +#define HDMI_SAMPLE_FLAT_MASK 0x0244 +#define HDMI_AUDN 0x0400 +#define HDMI_AUD_CTS 0x0404 #define HDMI_SW_DI_2_HEAD_WORD 0x0600 #define HDMI_SW_DI_2_PKT_WORD0 0x0604 #define HDMI_SW_DI_2_PKT_WORD1 0x0608 @@ -103,6 +110,7 @@ #define HDMI_INT_DLL_LCK BIT(5) #define HDMI_INT_NEW_FRAME BIT(6) #define HDMI_INT_GENCTRL_PKT BIT(7) +#define HDMI_INT_AUDIO_FIFO_XRUN BIT(8) #define HDMI_INT_SINK_TERM_PRESENT BIT(11) #define HDMI_DEFAULT_INT (HDMI_INT_SINK_TERM_PRESENT \ @@ -111,6 +119,7 @@ | HDMI_INT_GLOBAL) #define HDMI_WORKING_INT (HDMI_INT_SINK_TERM_PRESENT \ + | HDMI_INT_AUDIO_FIFO_XRUN \ | HDMI_INT_GENCTRL_PKT \ | HDMI_INT_NEW_FRAME \ | HDMI_INT_DLL_LCK \ @@ -121,6 +130,27 @@ #define HDMI_STA_SW_RST BIT(1) +#define HDMI_AUD_CFG_8CH BIT(0) +#define HDMI_AUD_CFG_SPDIF_DIV_2 BIT(1) +#define HDMI_AUD_CFG_SPDIF_DIV_3 BIT(2) +#define HDMI_AUD_CFG_SPDIF_CLK_DIV_4 (BIT(1) | BIT(2)) +#define HDMI_AUD_CFG_CTS_CLK_256FS BIT(12) +#define HDMI_AUD_CFG_DTS_INVALID BIT(16) +#define HDMI_AUD_CFG_ONE_BIT_INVALID (BIT(18) | BIT(19) | BIT(20) | BIT(21)) +#define HDMI_AUD_CFG_CH12_VALID BIT(28) +#define HDMI_AUD_CFG_CH34_VALID BIT(29) +#define HDMI_AUD_CFG_CH56_VALID BIT(30) +#define HDMI_AUD_CFG_CH78_VALID BIT(31) + +/* sample flat mask */ +#define HDMI_SAMPLE_FLAT_NO 0 +#define HDMI_SAMPLE_FLAT_SP0 BIT(0) +#define HDMI_SAMPLE_FLAT_SP1 BIT(1) +#define HDMI_SAMPLE_FLAT_SP2 BIT(2) +#define HDMI_SAMPLE_FLAT_SP3 BIT(3) +#define HDMI_SAMPLE_FLAT_ALL (HDMI_SAMPLE_FLAT_SP0 | HDMI_SAMPLE_FLAT_SP1 |\ + HDMI_SAMPLE_FLAT_SP2 | HDMI_SAMPLE_FLAT_SP3) + #define HDMI_INFOFRAME_HEADER_TYPE(x) (((x) & 0xff) << 0) #define HDMI_INFOFRAME_HEADER_VERSION(x) (((x) & 0xff) << 8) #define HDMI_INFOFRAME_HEADER_LEN(x) (((x) & 0x0f) << 16) @@ -171,6 +201,10 @@ static irqreturn_t hdmi_irq_thread(int irq, void *arg) wake_up_interruptible(&hdmi->wait_event); } + /* Audio FIFO underrun IRQ */ + if (hdmi->irq_status & HDMI_INT_AUDIO_FIFO_XRUN) + DRM_INFO("Warning: audio FIFO underrun occurs!"); + return IRQ_HANDLED; } @@ -441,26 +475,29 @@ static int hdmi_avi_infoframe_config(struct sti_hdmi *hdmi) */ static int hdmi_audio_infoframe_config(struct sti_hdmi *hdmi) { - struct hdmi_audio_infoframe infofame; + struct hdmi_audio_params *audio = &hdmi->audio; u8 buffer[HDMI_INFOFRAME_SIZE(AUDIO)]; - int ret; - - ret = hdmi_audio_infoframe_init(&infofame); - if (ret < 0) { - DRM_ERROR("failed to setup audio infoframe: %d\n", ret); - return ret; - } - - infofame.channels = 2; - - ret = hdmi_audio_infoframe_pack(&infofame, buffer, sizeof(buffer)); - if (ret < 0) { - DRM_ERROR("failed to pack audio infoframe: %d\n", ret); - return ret; + int ret, val; + + DRM_DEBUG_DRIVER("enter %s, AIF %s\n", __func__, + audio->enabled ? "enable" : "disable"); + if (audio->enabled) { + /* set audio parameters stored*/ + ret = hdmi_audio_infoframe_pack(&audio->cea, buffer, + sizeof(buffer)); + if (ret < 0) { + DRM_ERROR("failed to pack audio infoframe: %d\n", ret); + return ret; + } + hdmi_infoframe_write_infopack(hdmi, buffer, ret); + } else { + /*disable audio info frame transmission */ + val = hdmi_read(hdmi, HDMI_SW_DI_CFG); + val &= ~HDMI_IFRAME_CFG_DI_N(HDMI_IFRAME_MASK, + HDMI_IFRAME_SLOT_AUDIO); + hdmi_write(hdmi, val, HDMI_SW_DI_CFG); } - hdmi_infoframe_write_infopack(hdmi, buffer, ret); - return 0; } @@ -650,6 +687,10 @@ static int hdmi_dbg_show(struct seq_file *s, void *data) DBGFS_DUMP("", HDMI_SW_DI_CFG); hdmi_dbg_sw_di_cfg(s, hdmi_read(hdmi, HDMI_SW_DI_CFG)); + DBGFS_DUMP("\n", HDMI_AUDIO_CFG); + DBGFS_DUMP("\n", HDMI_SPDIF_FIFO_STATUS); + DBGFS_DUMP("\n", HDMI_AUDN); + seq_printf(s, "\n AVI Infoframe (Data Island slot N=%d):", HDMI_IFRAME_SLOT_AVI); DBGFS_DUMP_DI(HDMI_SW_DI_N_HEAD_WORD, HDMI_IFRAME_SLOT_AVI); @@ -854,6 +895,7 @@ static int sti_hdmi_connector_get_modes(struct drm_connector *connector) count = drm_add_edid_modes(connector, edid); drm_mode_connector_update_edid_property(connector, edid); + drm_edid_to_eld(connector, edid); kfree(edid); return count; @@ -1036,6 +1078,207 @@ static struct drm_encoder *sti_hdmi_find_encoder(struct drm_device *dev) return NULL; } +/** + * sti_hdmi_audio_get_non_coherent_n() - get N parameter for non-coherent + * clocks. None-coherent clocks means that audio and TMDS clocks have not the + * same source (drifts between clocks). In this case assumption is that CTS is + * automatically calculated by hardware. + * + * @audio_fs: audio frame clock frequency in Hz + * + * Values computed are based on table described in HDMI specification 1.4b + * + * Returns n value. + */ +static int sti_hdmi_audio_get_non_coherent_n(unsigned int audio_fs) +{ + unsigned int n; + + switch (audio_fs) { + case 32000: + n = 4096; + break; + case 44100: + n = 6272; + break; + case 48000: + n = 6144; + break; + case 88200: + n = 6272 * 2; + break; + case 96000: + n = 6144 * 2; + break; + case 176400: + n = 6272 * 4; + break; + case 192000: + n = 6144 * 4; + break; + default: + /* Not pre-defined, recommended value: 128 * fs / 1000 */ + n = (audio_fs * 128) / 1000; + } + + return n; +} + +static int hdmi_audio_configure(struct sti_hdmi *hdmi, + struct hdmi_audio_params *params) +{ + int audio_cfg, n; + struct hdmi_audio_infoframe *info = ¶ms->cea; + + DRM_DEBUG_DRIVER("\n"); + + if (!hdmi->enabled) + return 0; + + /* update N parameter */ + n = sti_hdmi_audio_get_non_coherent_n(params->sample_rate); + + DRM_DEBUG_DRIVER("Audio rate = %d Hz, TMDS clock = %d Hz, n = %d\n", + params->sample_rate, hdmi->mode.clock * 1000, n); + hdmi_write(hdmi, n, HDMI_AUDN); + + /* update HDMI registers according to configuration */ + audio_cfg = HDMI_AUD_CFG_SPDIF_DIV_2 | HDMI_AUD_CFG_DTS_INVALID | + HDMI_AUD_CFG_ONE_BIT_INVALID; + + switch (info->channels) { + case 8: + audio_cfg |= HDMI_AUD_CFG_CH78_VALID; + case 6: + audio_cfg |= HDMI_AUD_CFG_CH56_VALID; + case 4: + audio_cfg |= HDMI_AUD_CFG_CH34_VALID | HDMI_AUD_CFG_8CH; + case 2: + audio_cfg |= HDMI_AUD_CFG_CH12_VALID; + break; + default: + DRM_ERROR("ERROR: Unsupported number of channels (%d)!\n", + info->channels); + return -EINVAL; + } + + hdmi_write(hdmi, audio_cfg, HDMI_AUDIO_CFG); + + hdmi->audio = *params; + + return hdmi_audio_infoframe_config(hdmi); +} + +static void hdmi_audio_shutdown(struct device *dev, void *data) +{ + struct sti_hdmi *hdmi = dev_get_drvdata(dev); + int audio_cfg; + + DRM_DEBUG_DRIVER("\n"); + + /* disable audio */ + audio_cfg = HDMI_AUD_CFG_SPDIF_DIV_2 | HDMI_AUD_CFG_DTS_INVALID | + HDMI_AUD_CFG_ONE_BIT_INVALID; + hdmi_write(hdmi, audio_cfg, HDMI_AUDIO_CFG); + + hdmi->audio.enabled = 0; + hdmi_audio_infoframe_config(hdmi); +} + +static int hdmi_audio_hw_params(struct device *dev, + void *data, + struct hdmi_codec_daifmt *daifmt, + struct hdmi_codec_params *params) +{ + struct sti_hdmi *hdmi = dev_get_drvdata(dev); + int ret; + struct hdmi_audio_params audio = { + .sample_width = params->sample_width, + .sample_rate = params->sample_rate, + .cea = params->cea, + }; + + DRM_DEBUG_DRIVER("\n"); + + if (!hdmi->enabled) + return 0; + + if ((daifmt->fmt != HDMI_I2S) || daifmt->bit_clk_inv || + daifmt->frame_clk_inv || daifmt->bit_clk_master || + daifmt->frame_clk_master) { + dev_err(dev, "%s: Bad flags %d %d %d %d\n", __func__, + daifmt->bit_clk_inv, daifmt->frame_clk_inv, + daifmt->bit_clk_master, + daifmt->frame_clk_master); + return -EINVAL; + } + + audio.enabled = 1; + + ret = hdmi_audio_configure(hdmi, &audio); + if (ret < 0) + return ret; + + return 0; +} + +static int hdmi_audio_digital_mute(struct device *dev, void *data, bool enable) +{ + struct sti_hdmi *hdmi = dev_get_drvdata(dev); + + DRM_DEBUG_DRIVER("%s\n", enable ? "enable" : "disable"); + + if (enable) + hdmi_write(hdmi, HDMI_SAMPLE_FLAT_ALL, HDMI_SAMPLE_FLAT_MASK); + else + hdmi_write(hdmi, HDMI_SAMPLE_FLAT_NO, HDMI_SAMPLE_FLAT_MASK); + + return 0; +} + +static int hdmi_audio_get_eld(struct device *dev, void *data, uint8_t *buf, size_t len) +{ + struct sti_hdmi *hdmi = dev_get_drvdata(dev); + struct drm_connector *connector = hdmi->drm_connector; + + DRM_DEBUG_DRIVER("\n"); + memcpy(buf, connector->eld, min(sizeof(connector->eld), len)); + + return 0; +} + +static const struct hdmi_codec_ops audio_codec_ops = { + .hw_params = hdmi_audio_hw_params, + .audio_shutdown = hdmi_audio_shutdown, + .digital_mute = hdmi_audio_digital_mute, + .get_eld = hdmi_audio_get_eld, +}; + +static int sti_hdmi_register_audio_driver(struct device *dev, + struct sti_hdmi *hdmi) +{ + struct hdmi_codec_pdata codec_data = { + .ops = &audio_codec_ops, + .max_i2s_channels = 8, + .i2s = 1, + }; + + DRM_DEBUG_DRIVER("\n"); + + hdmi->audio.enabled = 0; + + hdmi->audio_pdev = platform_device_register_data( + dev, HDMI_CODEC_DRV_NAME, PLATFORM_DEVID_AUTO, + &codec_data, sizeof(codec_data)); + + if (IS_ERR(hdmi->audio_pdev)) + return PTR_ERR(hdmi->audio_pdev); + + DRM_INFO("%s Driver bound %s\n", HDMI_CODEC_DRV_NAME, dev_name(dev)); + + return 0; +} + static int sti_hdmi_bind(struct device *dev, struct device *master, void *data) { struct sti_hdmi *hdmi = dev_get_drvdata(dev); @@ -1082,12 +1325,27 @@ static int sti_hdmi_bind(struct device *dev, struct device *master, void *data) /* initialise property */ sti_hdmi_connector_init_property(drm_dev, drm_connector); + hdmi->drm_connector = drm_connector; + err = drm_mode_connector_attach_encoder(drm_connector, encoder); if (err) { DRM_ERROR("Failed to attach a connector to a encoder\n"); goto err_sysfs; } + err = sti_hdmi_register_audio_driver(dev, hdmi); + if (err) { + DRM_ERROR("Failed to attach an audio codec\n"); + goto err_sysfs; + } + + /* Initialize audio infoframe */ + err = hdmi_audio_infoframe_init(&hdmi->audio.cea); + if (err) { + DRM_ERROR("Failed to init audio infoframe\n"); + goto err_sysfs; + } + /* Enable default interrupts */ hdmi_write(hdmi, HDMI_DEFAULT_INT, HDMI_INT_EN); @@ -1095,6 +1353,7 @@ static int sti_hdmi_bind(struct device *dev, struct device *master, void *data) err_sysfs: drm_bridge_remove(bridge); + hdmi->drm_connector = NULL; return -EINVAL; } @@ -1244,6 +1503,8 @@ static int sti_hdmi_remove(struct platform_device *pdev) struct sti_hdmi *hdmi = dev_get_drvdata(&pdev->dev); i2c_put_adapter(hdmi->ddc_adapt); + if (hdmi->audio_pdev) + platform_device_unregister(hdmi->audio_pdev); component_del(&pdev->dev, &sti_hdmi_ops); return 0; diff --git a/drivers/gpu/drm/sti/sti_hdmi.h b/drivers/gpu/drm/sti/sti_hdmi.h index ef3a94583bbd..119bc3582ac7 100644 --- a/drivers/gpu/drm/sti/sti_hdmi.h +++ b/drivers/gpu/drm/sti/sti_hdmi.h @@ -23,6 +23,13 @@ struct hdmi_phy_ops { void (*stop)(struct sti_hdmi *hdmi); }; +struct hdmi_audio_params { + bool enabled; + unsigned int sample_width; + unsigned int sample_rate; + struct hdmi_audio_infoframe cea; +}; + /* values for the framing mode property */ enum sti_hdmi_modes { HDMI_MODE_HDMI, @@ -67,6 +74,9 @@ static const struct drm_prop_enum_list colorspace_mode_names[] = { * @ddc_adapt: i2c ddc adapter * @colorspace: current colorspace selected * @hdmi_mode: select framing for HDMI or DVI + * @audio_pdev: ASoC hdmi-codec platform device + * @audio: hdmi audio parameters. + * @drm_connector: hdmi connector */ struct sti_hdmi { struct device dev; @@ -89,6 +99,9 @@ struct sti_hdmi { struct i2c_adapter *ddc_adapt; enum hdmi_colorspace colorspace; enum sti_hdmi_modes hdmi_mode; + struct platform_device *audio_pdev; + struct hdmi_audio_params audio; + struct drm_connector *drm_connector; }; u32 hdmi_read(struct sti_hdmi *hdmi, int offset); diff --git a/drivers/gpu/drm/sti/sti_hqvdp.c b/drivers/gpu/drm/sti/sti_hqvdp.c index 33d2f42550cc..b03232247966 100644 --- a/drivers/gpu/drm/sti/sti_hqvdp.c +++ b/drivers/gpu/drm/sti/sti_hqvdp.c @@ -1363,6 +1363,7 @@ static int sti_hqvdp_probe(struct platform_device *pdev) vtg_np = of_parse_phandle(pdev->dev.of_node, "st,vtg", 0); if (vtg_np) hqvdp->vtg = of_vtg_find(vtg_np); + of_node_put(vtg_np); platform_set_drvdata(pdev, hqvdp); diff --git a/drivers/gpu/drm/sti/sti_plane.c b/drivers/gpu/drm/sti/sti_plane.c index 85cee9098439..0cf3335ef37c 100644 --- a/drivers/gpu/drm/sti/sti_plane.c +++ b/drivers/gpu/drm/sti/sti_plane.c @@ -45,25 +45,15 @@ const char *sti_plane_to_str(struct sti_plane *plane) #define STI_FPS_INTERVAL_MS 3000 -static int sti_plane_timespec_ms_diff(struct timespec lhs, struct timespec rhs) -{ - struct timespec tmp_ts = timespec_sub(lhs, rhs); - u64 tmp_ns = (u64)timespec_to_ns(&tmp_ts); - - do_div(tmp_ns, NSEC_PER_MSEC); - - return (u32)tmp_ns; -} - void sti_plane_update_fps(struct sti_plane *plane, bool new_frame, bool new_field) { - struct timespec now; + ktime_t now; struct sti_fps_info *fps; int fpks, fipks, ms_since_last, num_frames, num_fields; - getrawmonotonic(&now); + now = ktime_get(); /* Compute number of frame updates */ fps = &plane->fps_info; @@ -76,7 +66,7 @@ void sti_plane_update_fps(struct sti_plane *plane, return; fps->curr_frame_counter++; - ms_since_last = sti_plane_timespec_ms_diff(now, fps->last_timestamp); + ms_since_last = ktime_to_ms(ktime_sub(now, fps->last_timestamp)); num_frames = fps->curr_frame_counter - fps->last_frame_counter; if (num_frames <= 0 || ms_since_last < STI_FPS_INTERVAL_MS) diff --git a/drivers/gpu/drm/sti/sti_plane.h b/drivers/gpu/drm/sti/sti_plane.h index 39d39f5b7dd9..e0ea1dd3bb88 100644 --- a/drivers/gpu/drm/sti/sti_plane.h +++ b/drivers/gpu/drm/sti/sti_plane.h @@ -55,7 +55,7 @@ struct sti_fps_info { unsigned int last_frame_counter; unsigned int curr_field_counter; unsigned int last_field_counter; - struct timespec last_timestamp; + ktime_t last_timestamp; char fps_str[FPS_LENGTH]; char fips_str[FPS_LENGTH]; }; diff --git a/drivers/gpu/drm/sti/sti_vtg.c b/drivers/gpu/drm/sti/sti_vtg.c index 6bf4ce466d20..0bdc385eec17 100644 --- a/drivers/gpu/drm/sti/sti_vtg.c +++ b/drivers/gpu/drm/sti/sti_vtg.c @@ -65,7 +65,7 @@ #define HDMI_DELAY (5) /* Delay introduced by the DVO in nb of pixel */ -#define DVO_DELAY (2) +#define DVO_DELAY (7) /* delay introduced by the Arbitrary Waveform Generator in nb of pixels */ #define AWG_DELAY_HD (-9) @@ -432,6 +432,7 @@ static int vtg_probe(struct platform_device *pdev) np = of_parse_phandle(pdev->dev.of_node, "st,slave", 0); if (np) { vtg->slave = of_vtg_find(np); + of_node_put(np); if (!vtg->slave) return -EPROBE_DEFER; diff --git a/drivers/gpu/drm/sun4i/Kconfig b/drivers/gpu/drm/sun4i/Kconfig index 99510e64e91a..a4b357db8856 100644 --- a/drivers/gpu/drm/sun4i/Kconfig +++ b/drivers/gpu/drm/sun4i/Kconfig @@ -1,6 +1,6 @@ config DRM_SUN4I tristate "DRM Support for Allwinner A10 Display Engine" - depends on DRM && ARM + depends on DRM && ARM && COMMON_CLK depends on ARCH_SUNXI || COMPILE_TEST select DRM_GEM_CMA_HELPER select DRM_KMS_HELPER diff --git a/drivers/gpu/drm/sun4i/sun4i_backend.c b/drivers/gpu/drm/sun4i/sun4i_backend.c index f7a15c1a93bf..3ab560450a82 100644 --- a/drivers/gpu/drm/sun4i/sun4i_backend.c +++ b/drivers/gpu/drm/sun4i/sun4i_backend.c @@ -190,7 +190,7 @@ int sun4i_backend_update_layer_buffer(struct sun4i_backend *backend, /* Get the physical address of the buffer in memory */ gem = drm_fb_cma_get_gem_obj(fb, 0); - DRM_DEBUG_DRIVER("Using GEM @ 0x%x\n", gem->paddr); + DRM_DEBUG_DRIVER("Using GEM @ %pad\n", &gem->paddr); /* Compute the start of the displayed memory */ bpp = drm_format_plane_cpp(fb->pixel_format, 0); @@ -198,7 +198,7 @@ int sun4i_backend_update_layer_buffer(struct sun4i_backend *backend, paddr += (state->src_x >> 16) * bpp; paddr += (state->src_y >> 16) * fb->pitches[0]; - DRM_DEBUG_DRIVER("Setting buffer address to 0x%x\n", paddr); + DRM_DEBUG_DRIVER("Setting buffer address to %pad\n", &paddr); /* Write the 32 lower bits of the address (in bits) */ lo_paddr = paddr << 3; diff --git a/drivers/gpu/drm/sun4i/sun4i_crtc.c b/drivers/gpu/drm/sun4i/sun4i_crtc.c index f628b6d8f23f..4a192210574f 100644 --- a/drivers/gpu/drm/sun4i/sun4i_crtc.c +++ b/drivers/gpu/drm/sun4i/sun4i_crtc.c @@ -77,6 +77,14 @@ static void sun4i_crtc_disable(struct drm_crtc *crtc) DRM_DEBUG_DRIVER("Disabling the CRTC\n"); sun4i_tcon_disable(drv->tcon); + + if (crtc->state->event && !crtc->state->active) { + spin_lock_irq(&crtc->dev->event_lock); + drm_crtc_send_vblank_event(crtc, crtc->state->event); + spin_unlock_irq(&crtc->dev->event_lock); + + crtc->state->event = NULL; + } } static void sun4i_crtc_enable(struct drm_crtc *crtc) diff --git a/drivers/gpu/drm/sun4i/sun4i_dotclock.c b/drivers/gpu/drm/sun4i/sun4i_dotclock.c index 3ff668cb463c..5b3463197c48 100644 --- a/drivers/gpu/drm/sun4i/sun4i_dotclock.c +++ b/drivers/gpu/drm/sun4i/sun4i_dotclock.c @@ -72,14 +72,40 @@ static unsigned long sun4i_dclk_recalc_rate(struct clk_hw *hw, static long sun4i_dclk_round_rate(struct clk_hw *hw, unsigned long rate, unsigned long *parent_rate) { - return *parent_rate / DIV_ROUND_CLOSEST(*parent_rate, rate); + unsigned long best_parent = 0; + u8 best_div = 1; + int i; + + for (i = 6; i < 127; i++) { + unsigned long ideal = rate * i; + unsigned long rounded; + + rounded = clk_hw_round_rate(clk_hw_get_parent(hw), + ideal); + + if (rounded == ideal) { + best_parent = rounded; + best_div = i; + goto out; + } + + if ((rounded < ideal) && (rounded > best_parent)) { + best_parent = rounded; + best_div = i; + } + } + +out: + *parent_rate = best_parent; + + return best_parent / best_div; } static int sun4i_dclk_set_rate(struct clk_hw *hw, unsigned long rate, unsigned long parent_rate) { struct sun4i_dclk *dclk = hw_to_dclk(hw); - int div = DIV_ROUND_CLOSEST(parent_rate, rate); + u8 div = parent_rate / rate; return regmap_update_bits(dclk->regmap, SUN4I_TCON0_DCLK_REG, GENMASK(6, 0), div); @@ -127,10 +153,14 @@ int sun4i_dclk_create(struct device *dev, struct sun4i_tcon *tcon) const char *clk_name, *parent_name; struct clk_init_data init; struct sun4i_dclk *dclk; + int ret; parent_name = __clk_get_name(tcon->sclk0); - of_property_read_string_index(dev->of_node, "clock-output-names", 0, - &clk_name); + ret = of_property_read_string_index(dev->of_node, + "clock-output-names", 0, + &clk_name); + if (ret) + return ret; dclk = devm_kzalloc(dev, sizeof(*dclk), GFP_KERNEL); if (!dclk) @@ -140,6 +170,7 @@ int sun4i_dclk_create(struct device *dev, struct sun4i_tcon *tcon) init.ops = &sun4i_dclk_ops; init.parent_names = &parent_name; init.num_parents = 1; + init.flags = CLK_SET_RATE_PARENT; dclk->regmap = tcon->regs; dclk->hw.init = &init; diff --git a/drivers/gpu/drm/sun4i/sun4i_drv.c b/drivers/gpu/drm/sun4i/sun4i_drv.c index 9a67f927a53e..7092daaf6c43 100644 --- a/drivers/gpu/drm/sun4i/sun4i_drv.c +++ b/drivers/gpu/drm/sun4i/sun4i_drv.c @@ -92,11 +92,27 @@ static struct drm_driver sun4i_drv_driver = { /* Frame Buffer Operations */ /* VBlank Operations */ - .get_vblank_counter = drm_vblank_count, + .get_vblank_counter = drm_vblank_no_hw_counter, .enable_vblank = sun4i_drv_enable_vblank, .disable_vblank = sun4i_drv_disable_vblank, }; +static void sun4i_remove_framebuffers(void) +{ + struct apertures_struct *ap; + + ap = alloc_apertures(1); + if (!ap) + return; + + /* The framebuffer can be located anywhere in RAM */ + ap->ranges[0].base = 0; + ap->ranges[0].size = ~0; + + remove_conflicting_framebuffers(ap, "sun4i-drm-fb", false); + kfree(ap); +} + static int sun4i_drv_bind(struct device *dev) { struct drm_device *drm; @@ -140,6 +156,9 @@ static int sun4i_drv_bind(struct device *dev) } drm->irq_enabled = true; + /* Remove early framebuffers (ie. simplefb) */ + sun4i_remove_framebuffers(); + /* Create our framebuffer */ drv->fbdev = sun4i_framebuffer_init(drm); if (IS_ERR(drv->fbdev)) { @@ -280,6 +299,7 @@ static int sun4i_drv_probe(struct platform_device *pdev) count += sun4i_drv_add_endpoints(&pdev->dev, &match, pipeline); + of_node_put(pipeline); DRM_DEBUG_DRIVER("Queued %d outputs on pipeline %d\n", count, i); diff --git a/drivers/gpu/drm/sun4i/sun4i_framebuffer.c b/drivers/gpu/drm/sun4i/sun4i_framebuffer.c index a0b30c216a5b..70688febd7ac 100644 --- a/drivers/gpu/drm/sun4i/sun4i_framebuffer.c +++ b/drivers/gpu/drm/sun4i/sun4i_framebuffer.c @@ -20,8 +20,7 @@ static void sun4i_de_output_poll_changed(struct drm_device *drm) { struct sun4i_drv *drv = drm->dev_private; - if (drv->fbdev) - drm_fbdev_cma_hotplug_event(drv->fbdev); + drm_fbdev_cma_hotplug_event(drv->fbdev); } static const struct drm_mode_config_funcs sun4i_de_mode_config_funcs = { diff --git a/drivers/gpu/drm/sun4i/sun4i_rgb.c b/drivers/gpu/drm/sun4i/sun4i_rgb.c index 442cfe271688..f5bbac6efb4c 100644 --- a/drivers/gpu/drm/sun4i/sun4i_rgb.c +++ b/drivers/gpu/drm/sun4i/sun4i_rgb.c @@ -54,8 +54,13 @@ static int sun4i_rgb_get_modes(struct drm_connector *connector) static int sun4i_rgb_mode_valid(struct drm_connector *connector, struct drm_display_mode *mode) { + struct sun4i_rgb *rgb = drm_connector_to_sun4i_rgb(connector); + struct sun4i_drv *drv = rgb->drv; + struct sun4i_tcon *tcon = drv->tcon; u32 hsync = mode->hsync_end - mode->hsync_start; u32 vsync = mode->vsync_end - mode->vsync_start; + unsigned long rate = mode->clock * 1000; + long rounded_rate; DRM_DEBUG_DRIVER("Validating modes...\n"); @@ -87,6 +92,15 @@ static int sun4i_rgb_mode_valid(struct drm_connector *connector, DRM_DEBUG_DRIVER("Vertical parameters OK\n"); + rounded_rate = clk_round_rate(tcon->dclk, rate); + if (rounded_rate < rate) + return MODE_CLOCK_LOW; + + if (rounded_rate > rate) + return MODE_CLOCK_HIGH; + + DRM_DEBUG_DRIVER("Clock rate OK\n"); + return MODE_OK; } @@ -193,7 +207,7 @@ int sun4i_rgb_init(struct drm_device *drm) int ret; /* If we don't have a panel, there's no point in going on */ - if (!tcon->panel) + if (IS_ERR(tcon->panel)) return -ENODEV; rgb = devm_kzalloc(drm->dev, sizeof(*rgb), GFP_KERNEL); diff --git a/drivers/gpu/drm/sun4i/sun4i_tcon.c b/drivers/gpu/drm/sun4i/sun4i_tcon.c index 9f19b0e08560..652385f09735 100644 --- a/drivers/gpu/drm/sun4i/sun4i_tcon.c +++ b/drivers/gpu/drm/sun4i/sun4i_tcon.c @@ -425,11 +425,11 @@ static struct drm_panel *sun4i_tcon_find_panel(struct device_node *node) remote = of_graph_get_remote_port_parent(end_node); if (!remote) { - DRM_DEBUG_DRIVER("Enable to parse remote node\n"); + DRM_DEBUG_DRIVER("Unable to parse remote node\n"); return ERR_PTR(-EINVAL); } - return of_drm_find_panel(remote); + return of_drm_find_panel(remote) ?: ERR_PTR(-EPROBE_DEFER); } static int sun4i_tcon_bind(struct device *dev, struct device *master, @@ -490,7 +490,11 @@ static int sun4i_tcon_bind(struct device *dev, struct device *master, return 0; } - return sun4i_rgb_init(drm); + ret = sun4i_rgb_init(drm); + if (ret < 0) + goto err_free_clocks; + + return 0; err_free_clocks: sun4i_tcon_free_clocks(tcon); @@ -522,12 +526,13 @@ static int sun4i_tcon_probe(struct platform_device *pdev) * Defer the probe. */ panel = sun4i_tcon_find_panel(node); - if (IS_ERR(panel)) { - /* - * If we don't have a panel endpoint, just go on - */ - if (PTR_ERR(panel) != -ENODEV) - return -EPROBE_DEFER; + + /* + * If we don't have a panel endpoint, just go on + */ + if (PTR_ERR(panel) == -EPROBE_DEFER) { + DRM_DEBUG_DRIVER("Still waiting for our panel. Deferring...\n"); + return -EPROBE_DEFER; } return component_add(&pdev->dev, &sun4i_tcon_ops); diff --git a/drivers/gpu/drm/tegra/dc.c b/drivers/gpu/drm/tegra/dc.c index 39940f5b7c91..8495bd01b544 100644 --- a/drivers/gpu/drm/tegra/dc.c +++ b/drivers/gpu/drm/tegra/dc.c @@ -10,6 +10,7 @@ #include <linux/clk.h> #include <linux/debugfs.h> #include <linux/iommu.h> +#include <linux/pm_runtime.h> #include <linux/reset.h> #include <soc/tegra/pmc.h> @@ -1216,6 +1217,8 @@ static void tegra_crtc_disable(struct drm_crtc *crtc) tegra_dc_stats_reset(&dc->stats); drm_crtc_vblank_off(crtc); + + pm_runtime_put_sync(dc->dev); } static void tegra_crtc_enable(struct drm_crtc *crtc) @@ -1225,6 +1228,48 @@ static void tegra_crtc_enable(struct drm_crtc *crtc) struct tegra_dc *dc = to_tegra_dc(crtc); u32 value; + pm_runtime_get_sync(dc->dev); + + /* initialize display controller */ + if (dc->syncpt) { + u32 syncpt = host1x_syncpt_id(dc->syncpt); + + value = SYNCPT_CNTRL_NO_STALL; + tegra_dc_writel(dc, value, DC_CMD_GENERAL_INCR_SYNCPT_CNTRL); + + value = SYNCPT_VSYNC_ENABLE | syncpt; + tegra_dc_writel(dc, value, DC_CMD_CONT_SYNCPT_VSYNC); + } + + value = WIN_A_UF_INT | WIN_B_UF_INT | WIN_C_UF_INT | + WIN_A_OF_INT | WIN_B_OF_INT | WIN_C_OF_INT; + tegra_dc_writel(dc, value, DC_CMD_INT_TYPE); + + value = WIN_A_UF_INT | WIN_B_UF_INT | WIN_C_UF_INT | + WIN_A_OF_INT | WIN_B_OF_INT | WIN_C_OF_INT; + tegra_dc_writel(dc, value, DC_CMD_INT_POLARITY); + + /* initialize timer */ + value = CURSOR_THRESHOLD(0) | WINDOW_A_THRESHOLD(0x20) | + WINDOW_B_THRESHOLD(0x20) | WINDOW_C_THRESHOLD(0x20); + tegra_dc_writel(dc, value, DC_DISP_DISP_MEM_HIGH_PRIORITY); + + value = CURSOR_THRESHOLD(0) | WINDOW_A_THRESHOLD(1) | + WINDOW_B_THRESHOLD(1) | WINDOW_C_THRESHOLD(1); + tegra_dc_writel(dc, value, DC_DISP_DISP_MEM_HIGH_PRIORITY_TIMER); + + value = VBLANK_INT | WIN_A_UF_INT | WIN_B_UF_INT | WIN_C_UF_INT | + WIN_A_OF_INT | WIN_B_OF_INT | WIN_C_OF_INT; + tegra_dc_writel(dc, value, DC_CMD_INT_ENABLE); + + value = WIN_A_UF_INT | WIN_B_UF_INT | WIN_C_UF_INT | + WIN_A_OF_INT | WIN_B_OF_INT | WIN_C_OF_INT; + tegra_dc_writel(dc, value, DC_CMD_INT_MASK); + + if (dc->soc->supports_border_color) + tegra_dc_writel(dc, 0, DC_DISP_BORDER_COLOR); + + /* apply PLL and pixel clock changes */ tegra_dc_commit_state(dc, state); /* program display mode */ @@ -1685,7 +1730,6 @@ static int tegra_dc_init(struct host1x_client *client) struct tegra_drm *tegra = drm->dev_private; struct drm_plane *primary = NULL; struct drm_plane *cursor = NULL; - u32 value; int err; dc->syncpt = host1x_syncpt_request(dc->dev, flags); @@ -1755,47 +1799,6 @@ static int tegra_dc_init(struct host1x_client *client) goto cleanup; } - /* initialize display controller */ - if (dc->syncpt) { - u32 syncpt = host1x_syncpt_id(dc->syncpt); - - value = SYNCPT_CNTRL_NO_STALL; - tegra_dc_writel(dc, value, DC_CMD_GENERAL_INCR_SYNCPT_CNTRL); - - value = SYNCPT_VSYNC_ENABLE | syncpt; - tegra_dc_writel(dc, value, DC_CMD_CONT_SYNCPT_VSYNC); - } - - value = WIN_A_UF_INT | WIN_B_UF_INT | WIN_C_UF_INT | - WIN_A_OF_INT | WIN_B_OF_INT | WIN_C_OF_INT; - tegra_dc_writel(dc, value, DC_CMD_INT_TYPE); - - value = WIN_A_UF_INT | WIN_B_UF_INT | WIN_C_UF_INT | - WIN_A_OF_INT | WIN_B_OF_INT | WIN_C_OF_INT; - tegra_dc_writel(dc, value, DC_CMD_INT_POLARITY); - - /* initialize timer */ - value = CURSOR_THRESHOLD(0) | WINDOW_A_THRESHOLD(0x20) | - WINDOW_B_THRESHOLD(0x20) | WINDOW_C_THRESHOLD(0x20); - tegra_dc_writel(dc, value, DC_DISP_DISP_MEM_HIGH_PRIORITY); - - value = CURSOR_THRESHOLD(0) | WINDOW_A_THRESHOLD(1) | - WINDOW_B_THRESHOLD(1) | WINDOW_C_THRESHOLD(1); - tegra_dc_writel(dc, value, DC_DISP_DISP_MEM_HIGH_PRIORITY_TIMER); - - value = VBLANK_INT | WIN_A_UF_INT | WIN_B_UF_INT | WIN_C_UF_INT | - WIN_A_OF_INT | WIN_B_OF_INT | WIN_C_OF_INT; - tegra_dc_writel(dc, value, DC_CMD_INT_ENABLE); - - value = WIN_A_UF_INT | WIN_B_UF_INT | WIN_C_UF_INT | - WIN_A_OF_INT | WIN_B_OF_INT | WIN_C_OF_INT; - tegra_dc_writel(dc, value, DC_CMD_INT_MASK); - - if (dc->soc->supports_border_color) - tegra_dc_writel(dc, 0, DC_DISP_BORDER_COLOR); - - tegra_dc_stats_reset(&dc->stats); - return 0; cleanup: @@ -1987,33 +1990,15 @@ static int tegra_dc_probe(struct platform_device *pdev) return PTR_ERR(dc->rst); } + reset_control_assert(dc->rst); + if (dc->soc->has_powergate) { if (dc->pipe == 0) dc->powergate = TEGRA_POWERGATE_DIS; else dc->powergate = TEGRA_POWERGATE_DISB; - err = tegra_powergate_sequence_power_up(dc->powergate, dc->clk, - dc->rst); - if (err < 0) { - dev_err(&pdev->dev, "failed to power partition: %d\n", - err); - return err; - } - } else { - err = clk_prepare_enable(dc->clk); - if (err < 0) { - dev_err(&pdev->dev, "failed to enable clock: %d\n", - err); - return err; - } - - err = reset_control_deassert(dc->rst); - if (err < 0) { - dev_err(&pdev->dev, "failed to deassert reset: %d\n", - err); - return err; - } + tegra_powergate_power_off(dc->powergate); } regs = platform_get_resource(pdev, IORESOURCE_MEM, 0); @@ -2027,16 +2012,19 @@ static int tegra_dc_probe(struct platform_device *pdev) return -ENXIO; } - INIT_LIST_HEAD(&dc->client.list); - dc->client.ops = &dc_client_ops; - dc->client.dev = &pdev->dev; - err = tegra_dc_rgb_probe(dc); if (err < 0 && err != -ENODEV) { dev_err(&pdev->dev, "failed to probe RGB output: %d\n", err); return err; } + platform_set_drvdata(pdev, dc); + pm_runtime_enable(&pdev->dev); + + INIT_LIST_HEAD(&dc->client.list); + dc->client.ops = &dc_client_ops; + dc->client.dev = &pdev->dev; + err = host1x_client_register(&dc->client); if (err < 0) { dev_err(&pdev->dev, "failed to register host1x client: %d\n", @@ -2044,8 +2032,6 @@ static int tegra_dc_probe(struct platform_device *pdev) return err; } - platform_set_drvdata(pdev, dc); - return 0; } @@ -2067,7 +2053,22 @@ static int tegra_dc_remove(struct platform_device *pdev) return err; } - reset_control_assert(dc->rst); + pm_runtime_disable(&pdev->dev); + + return 0; +} + +#ifdef CONFIG_PM +static int tegra_dc_suspend(struct device *dev) +{ + struct tegra_dc *dc = dev_get_drvdata(dev); + int err; + + err = reset_control_assert(dc->rst); + if (err < 0) { + dev_err(dev, "failed to assert reset: %d\n", err); + return err; + } if (dc->soc->has_powergate) tegra_powergate_power_off(dc->powergate); @@ -2077,10 +2078,45 @@ static int tegra_dc_remove(struct platform_device *pdev) return 0; } +static int tegra_dc_resume(struct device *dev) +{ + struct tegra_dc *dc = dev_get_drvdata(dev); + int err; + + if (dc->soc->has_powergate) { + err = tegra_powergate_sequence_power_up(dc->powergate, dc->clk, + dc->rst); + if (err < 0) { + dev_err(dev, "failed to power partition: %d\n", err); + return err; + } + } else { + err = clk_prepare_enable(dc->clk); + if (err < 0) { + dev_err(dev, "failed to enable clock: %d\n", err); + return err; + } + + err = reset_control_deassert(dc->rst); + if (err < 0) { + dev_err(dev, "failed to deassert reset: %d\n", err); + return err; + } + } + + return 0; +} +#endif + +static const struct dev_pm_ops tegra_dc_pm_ops = { + SET_RUNTIME_PM_OPS(tegra_dc_suspend, tegra_dc_resume, NULL) +}; + struct platform_driver tegra_dc_driver = { .driver = { .name = "tegra-dc", .of_match_table = tegra_dc_of_match, + .pm = &tegra_dc_pm_ops, }, .probe = tegra_dc_probe, .remove = tegra_dc_remove, diff --git a/drivers/gpu/drm/tegra/dpaux.c b/drivers/gpu/drm/tegra/dpaux.c index b24a0f14821a..059f409556d5 100644 --- a/drivers/gpu/drm/tegra/dpaux.c +++ b/drivers/gpu/drm/tegra/dpaux.c @@ -12,6 +12,9 @@ #include <linux/interrupt.h> #include <linux/io.h> #include <linux/of_gpio.h> +#include <linux/pinctrl/pinconf-generic.h> +#include <linux/pinctrl/pinctrl.h> +#include <linux/pinctrl/pinmux.h> #include <linux/platform_device.h> #include <linux/reset.h> #include <linux/regulator/consumer.h> @@ -44,6 +47,11 @@ struct tegra_dpaux { struct completion complete; struct work_struct work; struct list_head list; + +#ifdef CONFIG_GENERIC_PINCONF + struct pinctrl_dev *pinctrl; + struct pinctrl_desc desc; +#endif }; static inline struct tegra_dpaux *to_dpaux(struct drm_dp_aux *aux) @@ -267,6 +275,148 @@ static irqreturn_t tegra_dpaux_irq(int irq, void *data) return ret; } +enum tegra_dpaux_functions { + DPAUX_PADCTL_FUNC_AUX, + DPAUX_PADCTL_FUNC_I2C, + DPAUX_PADCTL_FUNC_OFF, +}; + +static void tegra_dpaux_pad_power_down(struct tegra_dpaux *dpaux) +{ + u32 value = tegra_dpaux_readl(dpaux, DPAUX_HYBRID_SPARE); + + value |= DPAUX_HYBRID_SPARE_PAD_POWER_DOWN; + + tegra_dpaux_writel(dpaux, value, DPAUX_HYBRID_SPARE); +} + +static void tegra_dpaux_pad_power_up(struct tegra_dpaux *dpaux) +{ + u32 value = tegra_dpaux_readl(dpaux, DPAUX_HYBRID_SPARE); + + value &= ~DPAUX_HYBRID_SPARE_PAD_POWER_DOWN; + + tegra_dpaux_writel(dpaux, value, DPAUX_HYBRID_SPARE); +} + +static int tegra_dpaux_pad_config(struct tegra_dpaux *dpaux, unsigned function) +{ + u32 value; + + switch (function) { + case DPAUX_PADCTL_FUNC_AUX: + value = DPAUX_HYBRID_PADCTL_AUX_CMH(2) | + DPAUX_HYBRID_PADCTL_AUX_DRVZ(4) | + DPAUX_HYBRID_PADCTL_AUX_DRVI(0x18) | + DPAUX_HYBRID_PADCTL_AUX_INPUT_RCV | + DPAUX_HYBRID_PADCTL_MODE_AUX; + break; + + case DPAUX_PADCTL_FUNC_I2C: + value = DPAUX_HYBRID_PADCTL_I2C_SDA_INPUT_RCV | + DPAUX_HYBRID_PADCTL_I2C_SCL_INPUT_RCV | + DPAUX_HYBRID_PADCTL_MODE_I2C; + break; + + case DPAUX_PADCTL_FUNC_OFF: + tegra_dpaux_pad_power_down(dpaux); + return 0; + + default: + return -ENOTSUPP; + } + + tegra_dpaux_writel(dpaux, value, DPAUX_HYBRID_PADCTL); + tegra_dpaux_pad_power_up(dpaux); + + return 0; +} + +#ifdef CONFIG_GENERIC_PINCONF +static const struct pinctrl_pin_desc tegra_dpaux_pins[] = { + PINCTRL_PIN(0, "DP_AUX_CHx_P"), + PINCTRL_PIN(1, "DP_AUX_CHx_N"), +}; + +static const unsigned tegra_dpaux_pin_numbers[] = { 0, 1 }; + +static const char * const tegra_dpaux_groups[] = { + "dpaux-io", +}; + +static const char * const tegra_dpaux_functions[] = { + "aux", + "i2c", + "off", +}; + +static int tegra_dpaux_get_groups_count(struct pinctrl_dev *pinctrl) +{ + return ARRAY_SIZE(tegra_dpaux_groups); +} + +static const char *tegra_dpaux_get_group_name(struct pinctrl_dev *pinctrl, + unsigned int group) +{ + return tegra_dpaux_groups[group]; +} + +static int tegra_dpaux_get_group_pins(struct pinctrl_dev *pinctrl, + unsigned group, const unsigned **pins, + unsigned *num_pins) +{ + *pins = tegra_dpaux_pin_numbers; + *num_pins = ARRAY_SIZE(tegra_dpaux_pin_numbers); + + return 0; +} + +static const struct pinctrl_ops tegra_dpaux_pinctrl_ops = { + .get_groups_count = tegra_dpaux_get_groups_count, + .get_group_name = tegra_dpaux_get_group_name, + .get_group_pins = tegra_dpaux_get_group_pins, + .dt_node_to_map = pinconf_generic_dt_node_to_map_group, + .dt_free_map = pinconf_generic_dt_free_map, +}; + +static int tegra_dpaux_get_functions_count(struct pinctrl_dev *pinctrl) +{ + return ARRAY_SIZE(tegra_dpaux_functions); +} + +static const char *tegra_dpaux_get_function_name(struct pinctrl_dev *pinctrl, + unsigned int function) +{ + return tegra_dpaux_functions[function]; +} + +static int tegra_dpaux_get_function_groups(struct pinctrl_dev *pinctrl, + unsigned int function, + const char * const **groups, + unsigned * const num_groups) +{ + *num_groups = ARRAY_SIZE(tegra_dpaux_groups); + *groups = tegra_dpaux_groups; + + return 0; +} + +static int tegra_dpaux_set_mux(struct pinctrl_dev *pinctrl, + unsigned int function, unsigned int group) +{ + struct tegra_dpaux *dpaux = pinctrl_dev_get_drvdata(pinctrl); + + return tegra_dpaux_pad_config(dpaux, function); +} + +static const struct pinmux_ops tegra_dpaux_pinmux_ops = { + .get_functions_count = tegra_dpaux_get_functions_count, + .get_function_name = tegra_dpaux_get_function_name, + .get_function_groups = tegra_dpaux_get_function_groups, + .set_mux = tegra_dpaux_set_mux, +}; +#endif + static int tegra_dpaux_probe(struct platform_device *pdev) { struct tegra_dpaux *dpaux; @@ -294,11 +444,14 @@ static int tegra_dpaux_probe(struct platform_device *pdev) return -ENXIO; } - dpaux->rst = devm_reset_control_get(&pdev->dev, "dpaux"); - if (IS_ERR(dpaux->rst)) { - dev_err(&pdev->dev, "failed to get reset control: %ld\n", - PTR_ERR(dpaux->rst)); - return PTR_ERR(dpaux->rst); + if (!pdev->dev.pm_domain) { + dpaux->rst = devm_reset_control_get(&pdev->dev, "dpaux"); + if (IS_ERR(dpaux->rst)) { + dev_err(&pdev->dev, + "failed to get reset control: %ld\n", + PTR_ERR(dpaux->rst)); + return PTR_ERR(dpaux->rst); + } } dpaux->clk = devm_clk_get(&pdev->dev, NULL); @@ -315,34 +468,37 @@ static int tegra_dpaux_probe(struct platform_device *pdev) return err; } - reset_control_deassert(dpaux->rst); + if (dpaux->rst) + reset_control_deassert(dpaux->rst); dpaux->clk_parent = devm_clk_get(&pdev->dev, "parent"); if (IS_ERR(dpaux->clk_parent)) { dev_err(&pdev->dev, "failed to get parent clock: %ld\n", PTR_ERR(dpaux->clk_parent)); - return PTR_ERR(dpaux->clk_parent); + err = PTR_ERR(dpaux->clk_parent); + goto assert_reset; } err = clk_prepare_enable(dpaux->clk_parent); if (err < 0) { dev_err(&pdev->dev, "failed to enable parent clock: %d\n", err); - return err; + goto assert_reset; } err = clk_set_rate(dpaux->clk_parent, 270000000); if (err < 0) { dev_err(&pdev->dev, "failed to set clock to 270 MHz: %d\n", err); - return err; + goto disable_parent_clk; } dpaux->vdd = devm_regulator_get(&pdev->dev, "vdd"); if (IS_ERR(dpaux->vdd)) { dev_err(&pdev->dev, "failed to get VDD supply: %ld\n", PTR_ERR(dpaux->vdd)); - return PTR_ERR(dpaux->vdd); + err = PTR_ERR(dpaux->vdd); + goto disable_parent_clk; } err = devm_request_irq(dpaux->dev, dpaux->irq, tegra_dpaux_irq, 0, @@ -350,7 +506,7 @@ static int tegra_dpaux_probe(struct platform_device *pdev) if (err < 0) { dev_err(dpaux->dev, "failed to request IRQ#%u: %d\n", dpaux->irq, err); - return err; + goto disable_parent_clk; } disable_irq(dpaux->irq); @@ -360,7 +516,7 @@ static int tegra_dpaux_probe(struct platform_device *pdev) err = drm_dp_aux_register(&dpaux->aux); if (err < 0) - return err; + goto disable_parent_clk; /* * Assume that by default the DPAUX/I2C pads will be used for HDMI, @@ -370,16 +526,24 @@ static int tegra_dpaux_probe(struct platform_device *pdev) * is no possibility to perform the I2C mode configuration in the * HDMI path. */ - value = tegra_dpaux_readl(dpaux, DPAUX_HYBRID_SPARE); - value &= ~DPAUX_HYBRID_SPARE_PAD_POWER_DOWN; - tegra_dpaux_writel(dpaux, value, DPAUX_HYBRID_SPARE); - - value = tegra_dpaux_readl(dpaux, DPAUX_HYBRID_PADCTL); - value = DPAUX_HYBRID_PADCTL_I2C_SDA_INPUT_RCV | - DPAUX_HYBRID_PADCTL_I2C_SCL_INPUT_RCV | - DPAUX_HYBRID_PADCTL_MODE_I2C; - tegra_dpaux_writel(dpaux, value, DPAUX_HYBRID_PADCTL); + err = tegra_dpaux_pad_config(dpaux, DPAUX_HYBRID_PADCTL_MODE_I2C); + if (err < 0) + return err; +#ifdef CONFIG_GENERIC_PINCONF + dpaux->desc.name = dev_name(&pdev->dev); + dpaux->desc.pins = tegra_dpaux_pins; + dpaux->desc.npins = ARRAY_SIZE(tegra_dpaux_pins); + dpaux->desc.pctlops = &tegra_dpaux_pinctrl_ops; + dpaux->desc.pmxops = &tegra_dpaux_pinmux_ops; + dpaux->desc.owner = THIS_MODULE; + + dpaux->pinctrl = devm_pinctrl_register(&pdev->dev, &dpaux->desc, dpaux); + if (!dpaux->pinctrl) { + dev_err(&pdev->dev, "failed to register pincontrol\n"); + return -ENODEV; + } +#endif /* enable and clear all interrupts */ value = DPAUX_INTR_AUX_DONE | DPAUX_INTR_IRQ_EVENT | DPAUX_INTR_UNPLUG_EVENT | DPAUX_INTR_PLUG_EVENT; @@ -393,17 +557,24 @@ static int tegra_dpaux_probe(struct platform_device *pdev) platform_set_drvdata(pdev, dpaux); return 0; + +disable_parent_clk: + clk_disable_unprepare(dpaux->clk_parent); +assert_reset: + if (dpaux->rst) + reset_control_assert(dpaux->rst); + + clk_disable_unprepare(dpaux->clk); + + return err; } static int tegra_dpaux_remove(struct platform_device *pdev) { struct tegra_dpaux *dpaux = platform_get_drvdata(pdev); - u32 value; /* make sure pads are powered down when not in use */ - value = tegra_dpaux_readl(dpaux, DPAUX_HYBRID_SPARE); - value |= DPAUX_HYBRID_SPARE_PAD_POWER_DOWN; - tegra_dpaux_writel(dpaux, value, DPAUX_HYBRID_SPARE); + tegra_dpaux_pad_power_down(dpaux); drm_dp_aux_unregister(&dpaux->aux); @@ -414,7 +585,10 @@ static int tegra_dpaux_remove(struct platform_device *pdev) cancel_work_sync(&dpaux->work); clk_disable_unprepare(dpaux->clk_parent); - reset_control_assert(dpaux->rst); + + if (dpaux->rst) + reset_control_assert(dpaux->rst); + clk_disable_unprepare(dpaux->clk); return 0; @@ -528,30 +702,15 @@ enum drm_connector_status drm_dp_aux_detect(struct drm_dp_aux *aux) int drm_dp_aux_enable(struct drm_dp_aux *aux) { struct tegra_dpaux *dpaux = to_dpaux(aux); - u32 value; - - value = DPAUX_HYBRID_PADCTL_AUX_CMH(2) | - DPAUX_HYBRID_PADCTL_AUX_DRVZ(4) | - DPAUX_HYBRID_PADCTL_AUX_DRVI(0x18) | - DPAUX_HYBRID_PADCTL_AUX_INPUT_RCV | - DPAUX_HYBRID_PADCTL_MODE_AUX; - tegra_dpaux_writel(dpaux, value, DPAUX_HYBRID_PADCTL); - - value = tegra_dpaux_readl(dpaux, DPAUX_HYBRID_SPARE); - value &= ~DPAUX_HYBRID_SPARE_PAD_POWER_DOWN; - tegra_dpaux_writel(dpaux, value, DPAUX_HYBRID_SPARE); - return 0; + return tegra_dpaux_pad_config(dpaux, DPAUX_PADCTL_FUNC_AUX); } int drm_dp_aux_disable(struct drm_dp_aux *aux) { struct tegra_dpaux *dpaux = to_dpaux(aux); - u32 value; - value = tegra_dpaux_readl(dpaux, DPAUX_HYBRID_SPARE); - value |= DPAUX_HYBRID_SPARE_PAD_POWER_DOWN; - tegra_dpaux_writel(dpaux, value, DPAUX_HYBRID_SPARE); + tegra_dpaux_pad_power_down(dpaux); return 0; } diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c index a177a42a9849..755264d9db22 100644 --- a/drivers/gpu/drm/tegra/drm.c +++ b/drivers/gpu/drm/tegra/drm.c @@ -56,8 +56,8 @@ static void tegra_atomic_complete(struct tegra_drm *tegra, */ drm_atomic_helper_commit_modeset_disables(drm, state); - drm_atomic_helper_commit_planes(drm, state, false); drm_atomic_helper_commit_modeset_enables(drm, state); + drm_atomic_helper_commit_planes(drm, state, true); drm_atomic_helper_wait_for_vblanks(drm, state); diff --git a/drivers/gpu/drm/tegra/dsi.c b/drivers/gpu/drm/tegra/dsi.c index 099cccb2fbcb..3d228ad90e0f 100644 --- a/drivers/gpu/drm/tegra/dsi.c +++ b/drivers/gpu/drm/tegra/dsi.c @@ -13,6 +13,7 @@ #include <linux/of.h> #include <linux/of_platform.h> #include <linux/platform_device.h> +#include <linux/pm_runtime.h> #include <linux/reset.h> #include <linux/regulator/consumer.h> @@ -677,6 +678,45 @@ static void tegra_dsi_ganged_disable(struct tegra_dsi *dsi) tegra_dsi_writel(dsi, 0, DSI_GANGED_MODE_CONTROL); } +static int tegra_dsi_pad_enable(struct tegra_dsi *dsi) +{ + u32 value; + + value = DSI_PAD_CONTROL_VS1_PULLDN(0) | DSI_PAD_CONTROL_VS1_PDIO(0); + tegra_dsi_writel(dsi, value, DSI_PAD_CONTROL_0); + + return 0; +} + +static int tegra_dsi_pad_calibrate(struct tegra_dsi *dsi) +{ + u32 value; + + /* + * XXX Is this still needed? The module reset is deasserted right + * before this function is called. + */ + tegra_dsi_writel(dsi, 0, DSI_PAD_CONTROL_0); + tegra_dsi_writel(dsi, 0, DSI_PAD_CONTROL_1); + tegra_dsi_writel(dsi, 0, DSI_PAD_CONTROL_2); + tegra_dsi_writel(dsi, 0, DSI_PAD_CONTROL_3); + tegra_dsi_writel(dsi, 0, DSI_PAD_CONTROL_4); + + /* start calibration */ + tegra_dsi_pad_enable(dsi); + + value = DSI_PAD_SLEW_UP(0x7) | DSI_PAD_SLEW_DN(0x7) | + DSI_PAD_LP_UP(0x1) | DSI_PAD_LP_DN(0x1) | + DSI_PAD_OUT_CLK(0x0); + tegra_dsi_writel(dsi, value, DSI_PAD_CONTROL_2); + + value = DSI_PAD_PREEMP_PD_CLK(0x3) | DSI_PAD_PREEMP_PU_CLK(0x3) | + DSI_PAD_PREEMP_PD(0x03) | DSI_PAD_PREEMP_PU(0x3); + tegra_dsi_writel(dsi, value, DSI_PAD_CONTROL_3); + + return tegra_mipi_calibrate(dsi->mipi); +} + static void tegra_dsi_set_timeout(struct tegra_dsi *dsi, unsigned long bclk, unsigned int vrefresh) { @@ -836,7 +876,7 @@ static void tegra_dsi_encoder_disable(struct drm_encoder *encoder) tegra_dsi_disable(dsi); - return; + pm_runtime_put(dsi->dev); } static void tegra_dsi_encoder_enable(struct drm_encoder *encoder) @@ -847,6 +887,13 @@ static void tegra_dsi_encoder_enable(struct drm_encoder *encoder) struct tegra_dsi *dsi = to_dsi(output); struct tegra_dsi_state *state; u32 value; + int err; + + pm_runtime_get_sync(dsi->dev); + + err = tegra_dsi_pad_calibrate(dsi); + if (err < 0) + dev_err(dsi->dev, "MIPI calibration failed: %d\n", err); state = tegra_dsi_get_state(dsi); @@ -875,8 +922,6 @@ static void tegra_dsi_encoder_enable(struct drm_encoder *encoder) if (output->panel) drm_panel_enable(output->panel); - - return; } static int @@ -966,55 +1011,12 @@ static const struct drm_encoder_helper_funcs tegra_dsi_encoder_helper_funcs = { .atomic_check = tegra_dsi_encoder_atomic_check, }; -static int tegra_dsi_pad_enable(struct tegra_dsi *dsi) -{ - u32 value; - - value = DSI_PAD_CONTROL_VS1_PULLDN(0) | DSI_PAD_CONTROL_VS1_PDIO(0); - tegra_dsi_writel(dsi, value, DSI_PAD_CONTROL_0); - - return 0; -} - -static int tegra_dsi_pad_calibrate(struct tegra_dsi *dsi) -{ - u32 value; - - tegra_dsi_writel(dsi, 0, DSI_PAD_CONTROL_0); - tegra_dsi_writel(dsi, 0, DSI_PAD_CONTROL_1); - tegra_dsi_writel(dsi, 0, DSI_PAD_CONTROL_2); - tegra_dsi_writel(dsi, 0, DSI_PAD_CONTROL_3); - tegra_dsi_writel(dsi, 0, DSI_PAD_CONTROL_4); - - /* start calibration */ - tegra_dsi_pad_enable(dsi); - - value = DSI_PAD_SLEW_UP(0x7) | DSI_PAD_SLEW_DN(0x7) | - DSI_PAD_LP_UP(0x1) | DSI_PAD_LP_DN(0x1) | - DSI_PAD_OUT_CLK(0x0); - tegra_dsi_writel(dsi, value, DSI_PAD_CONTROL_2); - - value = DSI_PAD_PREEMP_PD_CLK(0x3) | DSI_PAD_PREEMP_PU_CLK(0x3) | - DSI_PAD_PREEMP_PD(0x03) | DSI_PAD_PREEMP_PU(0x3); - tegra_dsi_writel(dsi, value, DSI_PAD_CONTROL_3); - - return tegra_mipi_calibrate(dsi->mipi); -} - static int tegra_dsi_init(struct host1x_client *client) { struct drm_device *drm = dev_get_drvdata(client->parent); struct tegra_dsi *dsi = host1x_client_to_dsi(client); int err; - reset_control_deassert(dsi->rst); - - err = tegra_dsi_pad_calibrate(dsi); - if (err < 0) { - dev_err(dsi->dev, "MIPI calibration failed: %d\n", err); - goto reset; - } - /* Gangsters must not register their own outputs. */ if (!dsi->master) { dsi->output.dev = client->dev; @@ -1037,12 +1039,9 @@ static int tegra_dsi_init(struct host1x_client *client) drm_connector_register(&dsi->output.connector); err = tegra_output_init(drm, &dsi->output); - if (err < 0) { - dev_err(client->dev, - "failed to initialize output: %d\n", + if (err < 0) + dev_err(dsi->dev, "failed to initialize output: %d\n", err); - goto reset; - } dsi->output.encoder.possible_crtcs = 0x3; } @@ -1054,10 +1053,6 @@ static int tegra_dsi_init(struct host1x_client *client) } return 0; - -reset: - reset_control_assert(dsi->rst); - return err; } static int tegra_dsi_exit(struct host1x_client *client) @@ -1069,7 +1064,7 @@ static int tegra_dsi_exit(struct host1x_client *client) if (IS_ENABLED(CONFIG_DEBUG_FS)) tegra_dsi_debugfs_exit(dsi); - reset_control_assert(dsi->rst); + regulator_disable(dsi->vdd); return 0; } @@ -1493,74 +1488,50 @@ static int tegra_dsi_probe(struct platform_device *pdev) dsi->format = MIPI_DSI_FMT_RGB888; dsi->lanes = 4; - dsi->rst = devm_reset_control_get(&pdev->dev, "dsi"); - if (IS_ERR(dsi->rst)) - return PTR_ERR(dsi->rst); + if (!pdev->dev.pm_domain) { + dsi->rst = devm_reset_control_get(&pdev->dev, "dsi"); + if (IS_ERR(dsi->rst)) + return PTR_ERR(dsi->rst); + } dsi->clk = devm_clk_get(&pdev->dev, NULL); if (IS_ERR(dsi->clk)) { dev_err(&pdev->dev, "cannot get DSI clock\n"); - err = PTR_ERR(dsi->clk); - goto reset; - } - - err = clk_prepare_enable(dsi->clk); - if (err < 0) { - dev_err(&pdev->dev, "cannot enable DSI clock\n"); - goto reset; + return PTR_ERR(dsi->clk); } dsi->clk_lp = devm_clk_get(&pdev->dev, "lp"); if (IS_ERR(dsi->clk_lp)) { dev_err(&pdev->dev, "cannot get low-power clock\n"); - err = PTR_ERR(dsi->clk_lp); - goto disable_clk; - } - - err = clk_prepare_enable(dsi->clk_lp); - if (err < 0) { - dev_err(&pdev->dev, "cannot enable low-power clock\n"); - goto disable_clk; + return PTR_ERR(dsi->clk_lp); } dsi->clk_parent = devm_clk_get(&pdev->dev, "parent"); if (IS_ERR(dsi->clk_parent)) { dev_err(&pdev->dev, "cannot get parent clock\n"); - err = PTR_ERR(dsi->clk_parent); - goto disable_clk_lp; + return PTR_ERR(dsi->clk_parent); } dsi->vdd = devm_regulator_get(&pdev->dev, "avdd-dsi-csi"); if (IS_ERR(dsi->vdd)) { dev_err(&pdev->dev, "cannot get VDD supply\n"); - err = PTR_ERR(dsi->vdd); - goto disable_clk_lp; - } - - err = regulator_enable(dsi->vdd); - if (err < 0) { - dev_err(&pdev->dev, "cannot enable VDD supply\n"); - goto disable_clk_lp; + return PTR_ERR(dsi->vdd); } err = tegra_dsi_setup_clocks(dsi); if (err < 0) { dev_err(&pdev->dev, "cannot setup clocks\n"); - goto disable_vdd; + return err; } regs = platform_get_resource(pdev, IORESOURCE_MEM, 0); dsi->regs = devm_ioremap_resource(&pdev->dev, regs); - if (IS_ERR(dsi->regs)) { - err = PTR_ERR(dsi->regs); - goto disable_vdd; - } + if (IS_ERR(dsi->regs)) + return PTR_ERR(dsi->regs); dsi->mipi = tegra_mipi_request(&pdev->dev); - if (IS_ERR(dsi->mipi)) { - err = PTR_ERR(dsi->mipi); - goto disable_vdd; - } + if (IS_ERR(dsi->mipi)) + return PTR_ERR(dsi->mipi); dsi->host.ops = &tegra_dsi_host_ops; dsi->host.dev = &pdev->dev; @@ -1571,6 +1542,9 @@ static int tegra_dsi_probe(struct platform_device *pdev) goto mipi_free; } + platform_set_drvdata(pdev, dsi); + pm_runtime_enable(&pdev->dev); + INIT_LIST_HEAD(&dsi->client.list); dsi->client.ops = &dsi_client_ops; dsi->client.dev = &pdev->dev; @@ -1582,22 +1556,12 @@ static int tegra_dsi_probe(struct platform_device *pdev) goto unregister; } - platform_set_drvdata(pdev, dsi); - return 0; unregister: mipi_dsi_host_unregister(&dsi->host); mipi_free: tegra_mipi_free(dsi->mipi); -disable_vdd: - regulator_disable(dsi->vdd); -disable_clk_lp: - clk_disable_unprepare(dsi->clk_lp); -disable_clk: - clk_disable_unprepare(dsi->clk); -reset: - reset_control_assert(dsi->rst); return err; } @@ -1606,6 +1570,8 @@ static int tegra_dsi_remove(struct platform_device *pdev) struct tegra_dsi *dsi = platform_get_drvdata(pdev); int err; + pm_runtime_disable(&pdev->dev); + err = host1x_client_unregister(&dsi->client); if (err < 0) { dev_err(&pdev->dev, "failed to unregister host1x client: %d\n", @@ -1618,14 +1584,82 @@ static int tegra_dsi_remove(struct platform_device *pdev) mipi_dsi_host_unregister(&dsi->host); tegra_mipi_free(dsi->mipi); - regulator_disable(dsi->vdd); + return 0; +} + +#ifdef CONFIG_PM +static int tegra_dsi_suspend(struct device *dev) +{ + struct tegra_dsi *dsi = dev_get_drvdata(dev); + int err; + + if (dsi->rst) { + err = reset_control_assert(dsi->rst); + if (err < 0) { + dev_err(dev, "failed to assert reset: %d\n", err); + return err; + } + } + + usleep_range(1000, 2000); + clk_disable_unprepare(dsi->clk_lp); clk_disable_unprepare(dsi->clk); - reset_control_assert(dsi->rst); + + regulator_disable(dsi->vdd); return 0; } +static int tegra_dsi_resume(struct device *dev) +{ + struct tegra_dsi *dsi = dev_get_drvdata(dev); + int err; + + err = regulator_enable(dsi->vdd); + if (err < 0) { + dev_err(dsi->dev, "failed to enable VDD supply: %d\n", err); + return err; + } + + err = clk_prepare_enable(dsi->clk); + if (err < 0) { + dev_err(dev, "cannot enable DSI clock: %d\n", err); + goto disable_vdd; + } + + err = clk_prepare_enable(dsi->clk_lp); + if (err < 0) { + dev_err(dev, "cannot enable low-power clock: %d\n", err); + goto disable_clk; + } + + usleep_range(1000, 2000); + + if (dsi->rst) { + err = reset_control_deassert(dsi->rst); + if (err < 0) { + dev_err(dev, "cannot assert reset: %d\n", err); + goto disable_clk_lp; + } + } + + return 0; + +disable_clk_lp: + clk_disable_unprepare(dsi->clk_lp); +disable_clk: + clk_disable_unprepare(dsi->clk); +disable_vdd: + regulator_disable(dsi->vdd); + return err; +} +#endif + +static const struct dev_pm_ops tegra_dsi_pm_ops = { + SET_RUNTIME_PM_OPS(tegra_dsi_suspend, tegra_dsi_resume, NULL) +}; + static const struct of_device_id tegra_dsi_of_match[] = { { .compatible = "nvidia,tegra210-dsi", }, { .compatible = "nvidia,tegra132-dsi", }, @@ -1639,6 +1673,7 @@ struct platform_driver tegra_dsi_driver = { .driver = { .name = "tegra-dsi", .of_match_table = tegra_dsi_of_match, + .pm = &tegra_dsi_pm_ops, }, .probe = tegra_dsi_probe, .remove = tegra_dsi_remove, diff --git a/drivers/gpu/drm/tegra/fb.c b/drivers/gpu/drm/tegra/fb.c index 1b12aa7a715e..e6d71fa4028e 100644 --- a/drivers/gpu/drm/tegra/fb.c +++ b/drivers/gpu/drm/tegra/fb.c @@ -68,7 +68,7 @@ static void tegra_fb_destroy(struct drm_framebuffer *framebuffer) struct tegra_bo *bo = fb->planes[i]; if (bo) { - if (bo->pages && bo->vaddr) + if (bo->pages) vunmap(bo->vaddr); drm_gem_object_unreference_unlocked(&bo->gem); diff --git a/drivers/gpu/drm/tegra/hdmi.c b/drivers/gpu/drm/tegra/hdmi.c index 2fdb8796443e..cda0491ed6bf 100644 --- a/drivers/gpu/drm/tegra/hdmi.c +++ b/drivers/gpu/drm/tegra/hdmi.c @@ -11,6 +11,7 @@ #include <linux/debugfs.h> #include <linux/gpio.h> #include <linux/hdmi.h> +#include <linux/pm_runtime.h> #include <linux/regulator/consumer.h> #include <linux/reset.h> @@ -18,10 +19,14 @@ #include <drm/drm_crtc.h> #include <drm/drm_crtc_helper.h> +#include <sound/hda_verbs.h> + #include "hdmi.h" #include "drm.h" #include "dc.h" +#define HDMI_ELD_BUFFER_SIZE 96 + struct tmds_config { unsigned int pclk; u32 pll0; @@ -39,6 +44,8 @@ struct tegra_hdmi_config { u32 fuse_override_value; bool has_sor_io_peak_current; + bool has_hda; + bool has_hbr; }; struct tegra_hdmi { @@ -60,7 +67,10 @@ struct tegra_hdmi { const struct tegra_hdmi_config *config; unsigned int audio_source; - unsigned int audio_freq; + unsigned int audio_sample_rate; + unsigned int audio_channels; + + unsigned int pixel_clock; bool stereo; bool dvi; @@ -402,11 +412,11 @@ static const struct tmds_config tegra124_tmds_config[] = { }; static const struct tegra_hdmi_audio_config * -tegra_hdmi_get_audio_config(unsigned int audio_freq, unsigned int pclk) +tegra_hdmi_get_audio_config(unsigned int sample_rate, unsigned int pclk) { const struct tegra_hdmi_audio_config *table; - switch (audio_freq) { + switch (sample_rate) { case 32000: table = tegra_hdmi_audio_32k; break; @@ -476,44 +486,114 @@ static void tegra_hdmi_setup_audio_fs_tables(struct tegra_hdmi *hdmi) } } -static int tegra_hdmi_setup_audio(struct tegra_hdmi *hdmi, unsigned int pclk) +static void tegra_hdmi_write_aval(struct tegra_hdmi *hdmi, u32 value) +{ + static const struct { + unsigned int sample_rate; + unsigned int offset; + } regs[] = { + { 32000, HDMI_NV_PDISP_SOR_AUDIO_AVAL_0320 }, + { 44100, HDMI_NV_PDISP_SOR_AUDIO_AVAL_0441 }, + { 48000, HDMI_NV_PDISP_SOR_AUDIO_AVAL_0480 }, + { 88200, HDMI_NV_PDISP_SOR_AUDIO_AVAL_0882 }, + { 96000, HDMI_NV_PDISP_SOR_AUDIO_AVAL_0960 }, + { 176400, HDMI_NV_PDISP_SOR_AUDIO_AVAL_1764 }, + { 192000, HDMI_NV_PDISP_SOR_AUDIO_AVAL_1920 }, + }; + unsigned int i; + + for (i = 0; i < ARRAY_SIZE(regs); i++) { + if (regs[i].sample_rate == hdmi->audio_sample_rate) { + tegra_hdmi_writel(hdmi, value, regs[i].offset); + break; + } + } +} + +static int tegra_hdmi_setup_audio(struct tegra_hdmi *hdmi) { - struct device_node *node = hdmi->dev->of_node; const struct tegra_hdmi_audio_config *config; - unsigned int offset = 0; - u32 value; + u32 source, value; switch (hdmi->audio_source) { case HDA: - value = AUDIO_CNTRL0_SOURCE_SELECT_HDAL; + if (hdmi->config->has_hda) + source = SOR_AUDIO_CNTRL0_SOURCE_SELECT_HDAL; + else + return -EINVAL; + break; case SPDIF: - value = AUDIO_CNTRL0_SOURCE_SELECT_SPDIF; + if (hdmi->config->has_hda) + source = SOR_AUDIO_CNTRL0_SOURCE_SELECT_SPDIF; + else + source = AUDIO_CNTRL0_SOURCE_SELECT_SPDIF; break; default: - value = AUDIO_CNTRL0_SOURCE_SELECT_AUTO; + if (hdmi->config->has_hda) + source = SOR_AUDIO_CNTRL0_SOURCE_SELECT_AUTO; + else + source = AUDIO_CNTRL0_SOURCE_SELECT_AUTO; break; } - if (of_device_is_compatible(node, "nvidia,tegra30-hdmi")) { - value |= AUDIO_CNTRL0_ERROR_TOLERANCE(6) | - AUDIO_CNTRL0_FRAMES_PER_BLOCK(0xc0); - tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_AUDIO_CNTRL0); - } else { - value |= AUDIO_CNTRL0_INJECT_NULLSMPL; + /* + * Tegra30 and later use a slightly modified version of the register + * layout to accomodate for changes related to supporting HDA as the + * audio input source for HDMI. The source select field has moved to + * the SOR_AUDIO_CNTRL0 register, but the error tolerance and frames + * per block fields remain in the AUDIO_CNTRL0 register. + */ + if (hdmi->config->has_hda) { + /* + * Inject null samples into the audio FIFO for every frame in + * which the codec did not receive any samples. This applies + * to stereo LPCM only. + * + * XXX: This seems to be a remnant of MCP days when this was + * used to work around issues with monitors not being able to + * play back system startup sounds early. It is possibly not + * needed on Linux at all. + */ + if (hdmi->audio_channels == 2) + value = SOR_AUDIO_CNTRL0_INJECT_NULLSMPL; + else + value = 0; + + value |= source; + tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_SOR_AUDIO_CNTRL0); + } - value = AUDIO_CNTRL0_ERROR_TOLERANCE(6) | - AUDIO_CNTRL0_FRAMES_PER_BLOCK(0xc0); - tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_AUDIO_CNTRL0); + /* + * On Tegra20, HDA is not a supported audio source and the source + * select field is part of the AUDIO_CNTRL0 register. + */ + value = AUDIO_CNTRL0_FRAMES_PER_BLOCK(0xc0) | + AUDIO_CNTRL0_ERROR_TOLERANCE(6); + + if (!hdmi->config->has_hda) + value |= source; + + tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_AUDIO_CNTRL0); + + /* + * Advertise support for High Bit-Rate on Tegra114 and later. + */ + if (hdmi->config->has_hbr) { + value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_SOR_AUDIO_SPARE0); + value |= SOR_AUDIO_SPARE0_HBR_ENABLE; + tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_SOR_AUDIO_SPARE0); } - config = tegra_hdmi_get_audio_config(hdmi->audio_freq, pclk); + config = tegra_hdmi_get_audio_config(hdmi->audio_sample_rate, + hdmi->pixel_clock); if (!config) { - dev_err(hdmi->dev, "cannot set audio to %u at %u pclk\n", - hdmi->audio_freq, pclk); + dev_err(hdmi->dev, + "cannot set audio to %u Hz at %u Hz pixel clock\n", + hdmi->audio_sample_rate, hdmi->pixel_clock); return -EINVAL; } @@ -526,8 +606,8 @@ static int tegra_hdmi_setup_audio(struct tegra_hdmi *hdmi, unsigned int pclk) tegra_hdmi_writel(hdmi, ACR_SUBPACK_N(config->n) | ACR_ENABLE, HDMI_NV_PDISP_HDMI_ACR_0441_SUBPACK_HIGH); - value = ACR_SUBPACK_CTS(config->cts); - tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_HDMI_ACR_0441_SUBPACK_LOW); + tegra_hdmi_writel(hdmi, ACR_SUBPACK_CTS(config->cts), + HDMI_NV_PDISP_HDMI_ACR_0441_SUBPACK_LOW); value = SPARE_HW_CTS | SPARE_FORCE_SW_CTS | SPARE_CTS_RESET_VAL(1); tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_HDMI_SPARE); @@ -536,43 +616,53 @@ static int tegra_hdmi_setup_audio(struct tegra_hdmi *hdmi, unsigned int pclk) value &= ~AUDIO_N_RESETF; tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_AUDIO_N); - if (of_device_is_compatible(node, "nvidia,tegra30-hdmi")) { - switch (hdmi->audio_freq) { - case 32000: - offset = HDMI_NV_PDISP_SOR_AUDIO_AVAL_0320; - break; + if (hdmi->config->has_hda) + tegra_hdmi_write_aval(hdmi, config->aval); - case 44100: - offset = HDMI_NV_PDISP_SOR_AUDIO_AVAL_0441; - break; + tegra_hdmi_setup_audio_fs_tables(hdmi); - case 48000: - offset = HDMI_NV_PDISP_SOR_AUDIO_AVAL_0480; - break; + return 0; +} - case 88200: - offset = HDMI_NV_PDISP_SOR_AUDIO_AVAL_0882; - break; +static void tegra_hdmi_disable_audio(struct tegra_hdmi *hdmi) +{ + u32 value; - case 96000: - offset = HDMI_NV_PDISP_SOR_AUDIO_AVAL_0960; - break; + value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_HDMI_GENERIC_CTRL); + value &= ~GENERIC_CTRL_AUDIO; + tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_HDMI_GENERIC_CTRL); +} - case 176400: - offset = HDMI_NV_PDISP_SOR_AUDIO_AVAL_1764; - break; +static void tegra_hdmi_enable_audio(struct tegra_hdmi *hdmi) +{ + u32 value; - case 192000: - offset = HDMI_NV_PDISP_SOR_AUDIO_AVAL_1920; - break; - } + value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_HDMI_GENERIC_CTRL); + value |= GENERIC_CTRL_AUDIO; + tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_HDMI_GENERIC_CTRL); +} - tegra_hdmi_writel(hdmi, config->aval, offset); - } +static void tegra_hdmi_write_eld(struct tegra_hdmi *hdmi) +{ + size_t length = drm_eld_size(hdmi->output.connector.eld), i; + u32 value; - tegra_hdmi_setup_audio_fs_tables(hdmi); + for (i = 0; i < length; i++) + tegra_hdmi_writel(hdmi, i << 8 | hdmi->output.connector.eld[i], + HDMI_NV_PDISP_SOR_AUDIO_HDA_ELD_BUFWR); - return 0; + /* + * The HDA codec will always report an ELD buffer size of 96 bytes and + * the HDA codec driver will check that each byte read from the buffer + * is valid. Therefore every byte must be written, even if no 96 bytes + * were parsed from EDID. + */ + for (i = length; i < HDMI_ELD_BUFFER_SIZE; i++) + tegra_hdmi_writel(hdmi, i << 8 | 0, + HDMI_NV_PDISP_SOR_AUDIO_HDA_ELD_BUFWR); + + value = SOR_AUDIO_HDA_PRESENSE_VALID | SOR_AUDIO_HDA_PRESENSE_PRESENT; + tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_SOR_AUDIO_HDA_PRESENSE); } static inline u32 tegra_hdmi_subpack(const u8 *ptr, size_t size) @@ -644,12 +734,6 @@ static void tegra_hdmi_setup_avi_infoframe(struct tegra_hdmi *hdmi, u8 buffer[17]; ssize_t err; - if (hdmi->dvi) { - tegra_hdmi_writel(hdmi, 0, - HDMI_NV_PDISP_HDMI_AVI_INFOFRAME_CTRL); - return; - } - err = drm_hdmi_avi_infoframe_from_display_mode(&frame, mode); if (err < 0) { dev_err(hdmi->dev, "failed to setup AVI infoframe: %zd\n", err); @@ -663,9 +747,24 @@ static void tegra_hdmi_setup_avi_infoframe(struct tegra_hdmi *hdmi, } tegra_hdmi_write_infopack(hdmi, buffer, err); +} + +static void tegra_hdmi_disable_avi_infoframe(struct tegra_hdmi *hdmi) +{ + u32 value; - tegra_hdmi_writel(hdmi, INFOFRAME_CTRL_ENABLE, - HDMI_NV_PDISP_HDMI_AVI_INFOFRAME_CTRL); + value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_HDMI_AVI_INFOFRAME_CTRL); + value &= ~INFOFRAME_CTRL_ENABLE; + tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_HDMI_AVI_INFOFRAME_CTRL); +} + +static void tegra_hdmi_enable_avi_infoframe(struct tegra_hdmi *hdmi) +{ + u32 value; + + value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_HDMI_AVI_INFOFRAME_CTRL); + value |= INFOFRAME_CTRL_ENABLE; + tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_HDMI_AVI_INFOFRAME_CTRL); } static void tegra_hdmi_setup_audio_infoframe(struct tegra_hdmi *hdmi) @@ -674,12 +773,6 @@ static void tegra_hdmi_setup_audio_infoframe(struct tegra_hdmi *hdmi) u8 buffer[14]; ssize_t err; - if (hdmi->dvi) { - tegra_hdmi_writel(hdmi, 0, - HDMI_NV_PDISP_HDMI_AUDIO_INFOFRAME_CTRL); - return; - } - err = hdmi_audio_infoframe_init(&frame); if (err < 0) { dev_err(hdmi->dev, "failed to setup audio infoframe: %zd\n", @@ -687,7 +780,7 @@ static void tegra_hdmi_setup_audio_infoframe(struct tegra_hdmi *hdmi) return; } - frame.channels = 2; + frame.channels = hdmi->audio_channels; err = hdmi_audio_infoframe_pack(&frame, buffer, sizeof(buffer)); if (err < 0) { @@ -703,9 +796,24 @@ static void tegra_hdmi_setup_audio_infoframe(struct tegra_hdmi *hdmi) * bytes can be programmed. */ tegra_hdmi_write_infopack(hdmi, buffer, min_t(size_t, 10, err)); +} - tegra_hdmi_writel(hdmi, INFOFRAME_CTRL_ENABLE, - HDMI_NV_PDISP_HDMI_AUDIO_INFOFRAME_CTRL); +static void tegra_hdmi_disable_audio_infoframe(struct tegra_hdmi *hdmi) +{ + u32 value; + + value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_HDMI_AUDIO_INFOFRAME_CTRL); + value &= ~INFOFRAME_CTRL_ENABLE; + tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_HDMI_AUDIO_INFOFRAME_CTRL); +} + +static void tegra_hdmi_enable_audio_infoframe(struct tegra_hdmi *hdmi) +{ + u32 value; + + value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_HDMI_AUDIO_INFOFRAME_CTRL); + value |= INFOFRAME_CTRL_ENABLE; + tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_HDMI_AUDIO_INFOFRAME_CTRL); } static void tegra_hdmi_setup_stereo_infoframe(struct tegra_hdmi *hdmi) @@ -713,14 +821,6 @@ static void tegra_hdmi_setup_stereo_infoframe(struct tegra_hdmi *hdmi) struct hdmi_vendor_infoframe frame; u8 buffer[10]; ssize_t err; - u32 value; - - if (!hdmi->stereo) { - value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_HDMI_GENERIC_CTRL); - value &= ~GENERIC_CTRL_ENABLE; - tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_HDMI_GENERIC_CTRL); - return; - } hdmi_vendor_infoframe_init(&frame); frame.s3d_struct = HDMI_3D_STRUCTURE_FRAME_PACKING; @@ -733,6 +833,20 @@ static void tegra_hdmi_setup_stereo_infoframe(struct tegra_hdmi *hdmi) } tegra_hdmi_write_infopack(hdmi, buffer, err); +} + +static void tegra_hdmi_disable_stereo_infoframe(struct tegra_hdmi *hdmi) +{ + u32 value; + + value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_HDMI_GENERIC_CTRL); + value &= ~GENERIC_CTRL_ENABLE; + tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_HDMI_GENERIC_CTRL); +} + +static void tegra_hdmi_enable_stereo_infoframe(struct tegra_hdmi *hdmi) +{ + u32 value; value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_HDMI_GENERIC_CTRL); value |= GENERIC_CTRL_ENABLE; @@ -772,10 +886,25 @@ static bool tegra_output_is_hdmi(struct tegra_output *output) return drm_detect_hdmi_monitor(edid); } +static enum drm_connector_status +tegra_hdmi_connector_detect(struct drm_connector *connector, bool force) +{ + struct tegra_output *output = connector_to_output(connector); + struct tegra_hdmi *hdmi = to_hdmi(output); + enum drm_connector_status status; + + status = tegra_output_connector_detect(connector, force); + if (status == connector_status_connected) + return status; + + tegra_hdmi_writel(hdmi, 0, HDMI_NV_PDISP_SOR_AUDIO_HDA_PRESENSE); + return status; +} + static const struct drm_connector_funcs tegra_hdmi_connector_funcs = { .dpms = drm_atomic_helper_connector_dpms, .reset = drm_atomic_helper_connector_reset, - .detect = tegra_output_connector_detect, + .detect = tegra_hdmi_connector_detect, .fill_modes = drm_helper_probe_single_connector_modes, .destroy = tegra_output_connector_destroy, .atomic_duplicate_state = drm_atomic_helper_connector_duplicate_state, @@ -814,7 +943,9 @@ static const struct drm_encoder_funcs tegra_hdmi_encoder_funcs = { static void tegra_hdmi_encoder_disable(struct drm_encoder *encoder) { + struct tegra_output *output = encoder_to_output(encoder); struct tegra_dc *dc = to_tegra_dc(encoder->crtc); + struct tegra_hdmi *hdmi = to_hdmi(output); u32 value; /* @@ -828,6 +959,20 @@ static void tegra_hdmi_encoder_disable(struct drm_encoder *encoder) tegra_dc_commit(dc); } + + if (!hdmi->dvi) { + if (hdmi->stereo) + tegra_hdmi_disable_stereo_infoframe(hdmi); + + tegra_hdmi_disable_audio_infoframe(hdmi); + tegra_hdmi_disable_avi_infoframe(hdmi); + tegra_hdmi_disable_audio(hdmi); + } + + tegra_hdmi_writel(hdmi, 0, HDMI_NV_PDISP_INT_ENABLE); + tegra_hdmi_writel(hdmi, 0, HDMI_NV_PDISP_INT_MASK); + + pm_runtime_put(hdmi->dev); } static void tegra_hdmi_encoder_enable(struct drm_encoder *encoder) @@ -836,21 +981,28 @@ static void tegra_hdmi_encoder_enable(struct drm_encoder *encoder) unsigned int h_sync_width, h_front_porch, h_back_porch, i, rekey; struct tegra_output *output = encoder_to_output(encoder); struct tegra_dc *dc = to_tegra_dc(encoder->crtc); - struct device_node *node = output->dev->of_node; struct tegra_hdmi *hdmi = to_hdmi(output); - unsigned int pulse_start, div82, pclk; + unsigned int pulse_start, div82; int retries = 1000; u32 value; int err; - hdmi->dvi = !tegra_output_is_hdmi(output); + pm_runtime_get_sync(hdmi->dev); - pclk = mode->clock * 1000; + /* + * Enable and unmask the HDA codec SCRATCH0 register interrupt. This + * is used for interoperability between the HDA codec driver and the + * HDMI driver. + */ + tegra_hdmi_writel(hdmi, INT_CODEC_SCRATCH0, HDMI_NV_PDISP_INT_ENABLE); + tegra_hdmi_writel(hdmi, INT_CODEC_SCRATCH0, HDMI_NV_PDISP_INT_MASK); + + hdmi->pixel_clock = mode->clock * 1000; h_sync_width = mode->hsync_end - mode->hsync_start; h_back_porch = mode->htotal - mode->hsync_end; h_front_porch = mode->hsync_start - mode->hdisplay; - err = clk_set_rate(hdmi->clk, pclk); + err = clk_set_rate(hdmi->clk, hdmi->pixel_clock); if (err < 0) { dev_err(hdmi->dev, "failed to set HDMI clock frequency: %d\n", err); @@ -909,17 +1061,15 @@ static void tegra_hdmi_encoder_enable(struct drm_encoder *encoder) value = SOR_REFCLK_DIV_INT(div82 >> 2) | SOR_REFCLK_DIV_FRAC(div82); tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_SOR_REFCLK); + hdmi->dvi = !tegra_output_is_hdmi(output); if (!hdmi->dvi) { - err = tegra_hdmi_setup_audio(hdmi, pclk); + err = tegra_hdmi_setup_audio(hdmi); if (err < 0) hdmi->dvi = true; } - if (of_device_is_compatible(node, "nvidia,tegra20-hdmi")) { - /* - * TODO: add ELD support - */ - } + if (hdmi->config->has_hda) + tegra_hdmi_write_eld(hdmi); rekey = HDMI_REKEY_DEFAULT; value = HDMI_CTRL_REKEY(rekey); @@ -931,20 +1081,17 @@ static void tegra_hdmi_encoder_enable(struct drm_encoder *encoder) tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_HDMI_CTRL); - if (hdmi->dvi) - tegra_hdmi_writel(hdmi, 0x0, - HDMI_NV_PDISP_HDMI_GENERIC_CTRL); - else - tegra_hdmi_writel(hdmi, GENERIC_CTRL_AUDIO, - HDMI_NV_PDISP_HDMI_GENERIC_CTRL); + if (!hdmi->dvi) { + tegra_hdmi_setup_avi_infoframe(hdmi, mode); + tegra_hdmi_setup_audio_infoframe(hdmi); - tegra_hdmi_setup_avi_infoframe(hdmi, mode); - tegra_hdmi_setup_audio_infoframe(hdmi); - tegra_hdmi_setup_stereo_infoframe(hdmi); + if (hdmi->stereo) + tegra_hdmi_setup_stereo_infoframe(hdmi); + } /* TMDS CONFIG */ for (i = 0; i < hdmi->config->num_tmds; i++) { - if (pclk <= hdmi->config->tmds[i].pclk) { + if (hdmi->pixel_clock <= hdmi->config->tmds[i].pclk) { tegra_hdmi_setup_tmds(hdmi, &hdmi->config->tmds[i]); break; } @@ -1031,6 +1178,15 @@ static void tegra_hdmi_encoder_enable(struct drm_encoder *encoder) tegra_dc_commit(dc); + if (!hdmi->dvi) { + tegra_hdmi_enable_avi_infoframe(hdmi); + tegra_hdmi_enable_audio_infoframe(hdmi); + tegra_hdmi_enable_audio(hdmi); + + if (hdmi->stereo) + tegra_hdmi_enable_stereo_infoframe(hdmi); + } + /* TODO: add HDCP support */ } @@ -1235,8 +1391,14 @@ static int tegra_hdmi_show_regs(struct seq_file *s, void *data) DUMP_REG(HDMI_NV_PDISP_KEY_HDCP_KEY_TRIG); DUMP_REG(HDMI_NV_PDISP_KEY_SKEY_INDEX); DUMP_REG(HDMI_NV_PDISP_SOR_AUDIO_CNTRL0); + DUMP_REG(HDMI_NV_PDISP_SOR_AUDIO_SPARE0); + DUMP_REG(HDMI_NV_PDISP_SOR_AUDIO_HDA_CODEC_SCRATCH0); + DUMP_REG(HDMI_NV_PDISP_SOR_AUDIO_HDA_CODEC_SCRATCH1); DUMP_REG(HDMI_NV_PDISP_SOR_AUDIO_HDA_ELD_BUFWR); DUMP_REG(HDMI_NV_PDISP_SOR_AUDIO_HDA_PRESENSE); + DUMP_REG(HDMI_NV_PDISP_INT_STATUS); + DUMP_REG(HDMI_NV_PDISP_INT_MASK); + DUMP_REG(HDMI_NV_PDISP_INT_ENABLE); DUMP_REG(HDMI_NV_PDISP_SOR_IO_PEAK_CURRENT); #undef DUMP_REG @@ -1360,14 +1522,6 @@ static int tegra_hdmi_init(struct host1x_client *client) return err; } - err = clk_prepare_enable(hdmi->clk); - if (err < 0) { - dev_err(hdmi->dev, "failed to enable clock: %d\n", err); - return err; - } - - reset_control_deassert(hdmi->rst); - return 0; } @@ -1377,9 +1531,6 @@ static int tegra_hdmi_exit(struct host1x_client *client) tegra_output_exit(&hdmi->output); - reset_control_assert(hdmi->rst); - clk_disable_unprepare(hdmi->clk); - regulator_disable(hdmi->vdd); regulator_disable(hdmi->pll); regulator_disable(hdmi->hdmi); @@ -1401,6 +1552,8 @@ static const struct tegra_hdmi_config tegra20_hdmi_config = { .fuse_override_offset = HDMI_NV_PDISP_SOR_LANE_DRIVE_CURRENT, .fuse_override_value = 1 << 31, .has_sor_io_peak_current = false, + .has_hda = false, + .has_hbr = false, }; static const struct tegra_hdmi_config tegra30_hdmi_config = { @@ -1409,6 +1562,8 @@ static const struct tegra_hdmi_config tegra30_hdmi_config = { .fuse_override_offset = HDMI_NV_PDISP_SOR_LANE_DRIVE_CURRENT, .fuse_override_value = 1 << 31, .has_sor_io_peak_current = false, + .has_hda = true, + .has_hbr = false, }; static const struct tegra_hdmi_config tegra114_hdmi_config = { @@ -1417,6 +1572,8 @@ static const struct tegra_hdmi_config tegra114_hdmi_config = { .fuse_override_offset = HDMI_NV_PDISP_SOR_PAD_CTLS0, .fuse_override_value = 1 << 31, .has_sor_io_peak_current = true, + .has_hda = true, + .has_hbr = true, }; static const struct tegra_hdmi_config tegra124_hdmi_config = { @@ -1425,6 +1582,8 @@ static const struct tegra_hdmi_config tegra124_hdmi_config = { .fuse_override_offset = HDMI_NV_PDISP_SOR_PAD_CTLS0, .fuse_override_value = 1 << 31, .has_sor_io_peak_current = true, + .has_hda = true, + .has_hbr = true, }; static const struct of_device_id tegra_hdmi_of_match[] = { @@ -1436,6 +1595,67 @@ static const struct of_device_id tegra_hdmi_of_match[] = { }; MODULE_DEVICE_TABLE(of, tegra_hdmi_of_match); +static void hda_format_parse(unsigned int format, unsigned int *rate, + unsigned int *channels) +{ + unsigned int mul, div; + + if (format & AC_FMT_BASE_44K) + *rate = 44100; + else + *rate = 48000; + + mul = (format & AC_FMT_MULT_MASK) >> AC_FMT_MULT_SHIFT; + div = (format & AC_FMT_DIV_MASK) >> AC_FMT_DIV_SHIFT; + + *rate = *rate * (mul + 1) / (div + 1); + + *channels = (format & AC_FMT_CHAN_MASK) >> AC_FMT_CHAN_SHIFT; +} + +static irqreturn_t tegra_hdmi_irq(int irq, void *data) +{ + struct tegra_hdmi *hdmi = data; + u32 value; + int err; + + value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_INT_STATUS); + tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_INT_STATUS); + + if (value & INT_CODEC_SCRATCH0) { + unsigned int format; + u32 value; + + value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_SOR_AUDIO_HDA_CODEC_SCRATCH0); + + if (value & SOR_AUDIO_HDA_CODEC_SCRATCH0_VALID) { + unsigned int sample_rate, channels; + + format = value & SOR_AUDIO_HDA_CODEC_SCRATCH0_FMT_MASK; + + hda_format_parse(format, &sample_rate, &channels); + + hdmi->audio_sample_rate = sample_rate; + hdmi->audio_channels = channels; + + err = tegra_hdmi_setup_audio(hdmi); + if (err < 0) { + tegra_hdmi_disable_audio_infoframe(hdmi); + tegra_hdmi_disable_audio(hdmi); + } else { + tegra_hdmi_setup_audio_infoframe(hdmi); + tegra_hdmi_enable_audio_infoframe(hdmi); + tegra_hdmi_enable_audio(hdmi); + } + } else { + tegra_hdmi_disable_audio_infoframe(hdmi); + tegra_hdmi_disable_audio(hdmi); + } + } + + return IRQ_HANDLED; +} + static int tegra_hdmi_probe(struct platform_device *pdev) { const struct of_device_id *match; @@ -1453,8 +1673,10 @@ static int tegra_hdmi_probe(struct platform_device *pdev) hdmi->config = match->data; hdmi->dev = &pdev->dev; + hdmi->audio_source = AUTO; - hdmi->audio_freq = 44100; + hdmi->audio_sample_rate = 48000; + hdmi->audio_channels = 2; hdmi->stereo = false; hdmi->dvi = false; @@ -1515,6 +1737,17 @@ static int tegra_hdmi_probe(struct platform_device *pdev) hdmi->irq = err; + err = devm_request_irq(hdmi->dev, hdmi->irq, tegra_hdmi_irq, 0, + dev_name(hdmi->dev), hdmi); + if (err < 0) { + dev_err(&pdev->dev, "failed to request IRQ#%u: %d\n", + hdmi->irq, err); + return err; + } + + platform_set_drvdata(pdev, hdmi); + pm_runtime_enable(&pdev->dev); + INIT_LIST_HEAD(&hdmi->client.list); hdmi->client.ops = &hdmi_client_ops; hdmi->client.dev = &pdev->dev; @@ -1526,8 +1759,6 @@ static int tegra_hdmi_probe(struct platform_device *pdev) return err; } - platform_set_drvdata(pdev, hdmi); - return 0; } @@ -1536,6 +1767,8 @@ static int tegra_hdmi_remove(struct platform_device *pdev) struct tegra_hdmi *hdmi = platform_get_drvdata(pdev); int err; + pm_runtime_disable(&pdev->dev); + err = host1x_client_unregister(&hdmi->client); if (err < 0) { dev_err(&pdev->dev, "failed to unregister host1x client: %d\n", @@ -1545,17 +1778,61 @@ static int tegra_hdmi_remove(struct platform_device *pdev) tegra_output_remove(&hdmi->output); - clk_disable_unprepare(hdmi->clk_parent); + return 0; +} + +#ifdef CONFIG_PM +static int tegra_hdmi_suspend(struct device *dev) +{ + struct tegra_hdmi *hdmi = dev_get_drvdata(dev); + int err; + + err = reset_control_assert(hdmi->rst); + if (err < 0) { + dev_err(dev, "failed to assert reset: %d\n", err); + return err; + } + + usleep_range(1000, 2000); + clk_disable_unprepare(hdmi->clk); return 0; } +static int tegra_hdmi_resume(struct device *dev) +{ + struct tegra_hdmi *hdmi = dev_get_drvdata(dev); + int err; + + err = clk_prepare_enable(hdmi->clk); + if (err < 0) { + dev_err(dev, "failed to enable clock: %d\n", err); + return err; + } + + usleep_range(1000, 2000); + + err = reset_control_deassert(hdmi->rst); + if (err < 0) { + dev_err(dev, "failed to deassert reset: %d\n", err); + clk_disable_unprepare(hdmi->clk); + return err; + } + + return 0; +} +#endif + +static const struct dev_pm_ops tegra_hdmi_pm_ops = { + SET_RUNTIME_PM_OPS(tegra_hdmi_suspend, tegra_hdmi_resume, NULL) +}; + struct platform_driver tegra_hdmi_driver = { .driver = { .name = "tegra-hdmi", - .owner = THIS_MODULE, .of_match_table = tegra_hdmi_of_match, + .pm = &tegra_hdmi_pm_ops, }, .probe = tegra_hdmi_probe, .remove = tegra_hdmi_remove, diff --git a/drivers/gpu/drm/tegra/hdmi.h b/drivers/gpu/drm/tegra/hdmi.h index a882514389cd..2339f134a09a 100644 --- a/drivers/gpu/drm/tegra/hdmi.h +++ b/drivers/gpu/drm/tegra/hdmi.h @@ -468,9 +468,20 @@ #define HDMI_NV_PDISP_KEY_SKEY_INDEX 0xa3 #define HDMI_NV_PDISP_SOR_AUDIO_CNTRL0 0xac -#define AUDIO_CNTRL0_INJECT_NULLSMPL (1 << 29) +#define SOR_AUDIO_CNTRL0_SOURCE_SELECT_AUTO (0 << 20) +#define SOR_AUDIO_CNTRL0_SOURCE_SELECT_SPDIF (1 << 20) +#define SOR_AUDIO_CNTRL0_SOURCE_SELECT_HDAL (2 << 20) +#define SOR_AUDIO_CNTRL0_INJECT_NULLSMPL (1 << 29) +#define HDMI_NV_PDISP_SOR_AUDIO_SPARE0 0xae +#define SOR_AUDIO_SPARE0_HBR_ENABLE (1 << 27) +#define HDMI_NV_PDISP_SOR_AUDIO_HDA_CODEC_SCRATCH0 0xba +#define SOR_AUDIO_HDA_CODEC_SCRATCH0_VALID (1 << 30) +#define SOR_AUDIO_HDA_CODEC_SCRATCH0_FMT_MASK 0xffff +#define HDMI_NV_PDISP_SOR_AUDIO_HDA_CODEC_SCRATCH1 0xbb #define HDMI_NV_PDISP_SOR_AUDIO_HDA_ELD_BUFWR 0xbc #define HDMI_NV_PDISP_SOR_AUDIO_HDA_PRESENSE 0xbd +#define SOR_AUDIO_HDA_PRESENSE_VALID (1 << 1) +#define SOR_AUDIO_HDA_PRESENSE_PRESENT (1 << 0) #define HDMI_NV_PDISP_SOR_AUDIO_AVAL_0320 0xbf #define HDMI_NV_PDISP_SOR_AUDIO_AVAL_0441 0xc0 @@ -481,6 +492,14 @@ #define HDMI_NV_PDISP_SOR_AUDIO_AVAL_1920 0xc5 #define HDMI_NV_PDISP_SOR_AUDIO_AVAL_DEFAULT 0xc5 +#define HDMI_NV_PDISP_INT_STATUS 0xcc +#define INT_SCRATCH (1 << 3) +#define INT_CP_REQUEST (1 << 2) +#define INT_CODEC_SCRATCH1 (1 << 1) +#define INT_CODEC_SCRATCH0 (1 << 0) +#define HDMI_NV_PDISP_INT_MASK 0xcd +#define HDMI_NV_PDISP_INT_ENABLE 0xce + #define HDMI_NV_PDISP_SOR_IO_PEAK_CURRENT 0xd1 #define PEAK_CURRENT_LANE0(x) (((x) & 0x7f) << 0) #define PEAK_CURRENT_LANE1(x) (((x) & 0x7f) << 8) diff --git a/drivers/gpu/drm/tegra/output.c b/drivers/gpu/drm/tegra/output.c index 1480f6aaffe4..595d1ec3e02e 100644 --- a/drivers/gpu/drm/tegra/output.c +++ b/drivers/gpu/drm/tegra/output.c @@ -36,6 +36,7 @@ int tegra_output_connector_get_modes(struct drm_connector *connector) if (edid) { err = drm_add_edid_modes(connector, edid); + drm_edid_to_eld(connector, edid); kfree(edid); } diff --git a/drivers/gpu/drm/tegra/sor.c b/drivers/gpu/drm/tegra/sor.c index 34958d71284b..74d0540b8d4c 100644 --- a/drivers/gpu/drm/tegra/sor.c +++ b/drivers/gpu/drm/tegra/sor.c @@ -7,11 +7,13 @@ */ #include <linux/clk.h> +#include <linux/clk-provider.h> #include <linux/debugfs.h> #include <linux/gpio.h> #include <linux/io.h> #include <linux/of_device.h> #include <linux/platform_device.h> +#include <linux/pm_runtime.h> #include <linux/regulator/consumer.h> #include <linux/reset.h> @@ -149,6 +151,8 @@ struct tegra_sor_soc { const struct tegra_sor_hdmi_settings *settings; unsigned int num_settings; + + const u8 *xbar_cfg; }; struct tegra_sor; @@ -169,7 +173,9 @@ struct tegra_sor { struct reset_control *rst; struct clk *clk_parent; + struct clk *clk_brick; struct clk *clk_safe; + struct clk *clk_src; struct clk *clk_dp; struct clk *clk; @@ -190,6 +196,18 @@ struct tegra_sor { struct regulator *hdmi_supply; }; +struct tegra_sor_state { + struct drm_connector_state base; + + unsigned int bpc; +}; + +static inline struct tegra_sor_state * +to_sor_state(struct drm_connector_state *state) +{ + return container_of(state, struct tegra_sor_state, base); +} + struct tegra_sor_config { u32 bits_per_pixel; @@ -225,6 +243,118 @@ static inline void tegra_sor_writel(struct tegra_sor *sor, u32 value, writel(value, sor->regs + (offset << 2)); } +static int tegra_sor_set_parent_clock(struct tegra_sor *sor, struct clk *parent) +{ + int err; + + clk_disable_unprepare(sor->clk); + + err = clk_set_parent(sor->clk, parent); + if (err < 0) + return err; + + err = clk_prepare_enable(sor->clk); + if (err < 0) + return err; + + return 0; +} + +struct tegra_clk_sor_brick { + struct clk_hw hw; + struct tegra_sor *sor; +}; + +static inline struct tegra_clk_sor_brick *to_brick(struct clk_hw *hw) +{ + return container_of(hw, struct tegra_clk_sor_brick, hw); +} + +static const char * const tegra_clk_sor_brick_parents[] = { + "pll_d2_out0", "pll_dp" +}; + +static int tegra_clk_sor_brick_set_parent(struct clk_hw *hw, u8 index) +{ + struct tegra_clk_sor_brick *brick = to_brick(hw); + struct tegra_sor *sor = brick->sor; + u32 value; + + value = tegra_sor_readl(sor, SOR_CLK_CNTRL); + value &= ~SOR_CLK_CNTRL_DP_CLK_SEL_MASK; + + switch (index) { + case 0: + value |= SOR_CLK_CNTRL_DP_CLK_SEL_SINGLE_PCLK; + break; + + case 1: + value |= SOR_CLK_CNTRL_DP_CLK_SEL_SINGLE_DPCLK; + break; + } + + tegra_sor_writel(sor, value, SOR_CLK_CNTRL); + + return 0; +} + +static u8 tegra_clk_sor_brick_get_parent(struct clk_hw *hw) +{ + struct tegra_clk_sor_brick *brick = to_brick(hw); + struct tegra_sor *sor = brick->sor; + u8 parent = U8_MAX; + u32 value; + + value = tegra_sor_readl(sor, SOR_CLK_CNTRL); + + switch (value & SOR_CLK_CNTRL_DP_CLK_SEL_MASK) { + case SOR_CLK_CNTRL_DP_CLK_SEL_SINGLE_PCLK: + case SOR_CLK_CNTRL_DP_CLK_SEL_DIFF_PCLK: + parent = 0; + break; + + case SOR_CLK_CNTRL_DP_CLK_SEL_SINGLE_DPCLK: + case SOR_CLK_CNTRL_DP_CLK_SEL_DIFF_DPCLK: + parent = 1; + break; + } + + return parent; +} + +static const struct clk_ops tegra_clk_sor_brick_ops = { + .set_parent = tegra_clk_sor_brick_set_parent, + .get_parent = tegra_clk_sor_brick_get_parent, +}; + +static struct clk *tegra_clk_sor_brick_register(struct tegra_sor *sor, + const char *name) +{ + struct tegra_clk_sor_brick *brick; + struct clk_init_data init; + struct clk *clk; + + brick = devm_kzalloc(sor->dev, sizeof(*brick), GFP_KERNEL); + if (!brick) + return ERR_PTR(-ENOMEM); + + brick->sor = sor; + + init.name = name; + init.flags = 0; + init.parent_names = tegra_clk_sor_brick_parents; + init.num_parents = ARRAY_SIZE(tegra_clk_sor_brick_parents); + init.ops = &tegra_clk_sor_brick_ops; + + brick->hw.init = &init; + + clk = devm_clk_register(sor->dev, &brick->hw); + if (IS_ERR(clk)) + kfree(brick); + + return clk; +} + static int tegra_sor_dp_train_fast(struct tegra_sor *sor, struct drm_dp_link *link) { @@ -569,10 +699,10 @@ static int tegra_sor_compute_params(struct tegra_sor *sor, return false; } -static int tegra_sor_calc_config(struct tegra_sor *sor, - const struct drm_display_mode *mode, - struct tegra_sor_config *config, - struct drm_dp_link *link) +static int tegra_sor_compute_config(struct tegra_sor *sor, + const struct drm_display_mode *mode, + struct tegra_sor_config *config, + struct drm_dp_link *link) { const u64 f = 100000, link_rate = link->rate * 1000; const u64 pclk = mode->clock * 1000; @@ -661,6 +791,135 @@ static int tegra_sor_calc_config(struct tegra_sor *sor, return 0; } +static void tegra_sor_apply_config(struct tegra_sor *sor, + const struct tegra_sor_config *config) +{ + u32 value; + + value = tegra_sor_readl(sor, SOR_DP_LINKCTL0); + value &= ~SOR_DP_LINKCTL_TU_SIZE_MASK; + value |= SOR_DP_LINKCTL_TU_SIZE(config->tu_size); + tegra_sor_writel(sor, value, SOR_DP_LINKCTL0); + + value = tegra_sor_readl(sor, SOR_DP_CONFIG0); + value &= ~SOR_DP_CONFIG_WATERMARK_MASK; + value |= SOR_DP_CONFIG_WATERMARK(config->watermark); + + value &= ~SOR_DP_CONFIG_ACTIVE_SYM_COUNT_MASK; + value |= SOR_DP_CONFIG_ACTIVE_SYM_COUNT(config->active_count); + + value &= ~SOR_DP_CONFIG_ACTIVE_SYM_FRAC_MASK; + value |= SOR_DP_CONFIG_ACTIVE_SYM_FRAC(config->active_frac); + + if (config->active_polarity) + value |= SOR_DP_CONFIG_ACTIVE_SYM_POLARITY; + else + value &= ~SOR_DP_CONFIG_ACTIVE_SYM_POLARITY; + + value |= SOR_DP_CONFIG_ACTIVE_SYM_ENABLE; + value |= SOR_DP_CONFIG_DISPARITY_NEGATIVE; + tegra_sor_writel(sor, value, SOR_DP_CONFIG0); + + value = tegra_sor_readl(sor, SOR_DP_AUDIO_HBLANK_SYMBOLS); + value &= ~SOR_DP_AUDIO_HBLANK_SYMBOLS_MASK; + value |= config->hblank_symbols & 0xffff; + tegra_sor_writel(sor, value, SOR_DP_AUDIO_HBLANK_SYMBOLS); + + value = tegra_sor_readl(sor, SOR_DP_AUDIO_VBLANK_SYMBOLS); + value &= ~SOR_DP_AUDIO_VBLANK_SYMBOLS_MASK; + value |= config->vblank_symbols & 0xffff; + tegra_sor_writel(sor, value, SOR_DP_AUDIO_VBLANK_SYMBOLS); +} + +static void tegra_sor_mode_set(struct tegra_sor *sor, + const struct drm_display_mode *mode, + struct tegra_sor_state *state) +{ + struct tegra_dc *dc = to_tegra_dc(sor->output.encoder.crtc); + unsigned int vbe, vse, hbe, hse, vbs, hbs; + u32 value; + + value = tegra_sor_readl(sor, SOR_STATE1); + value &= ~SOR_STATE_ASY_PIXELDEPTH_MASK; + value &= ~SOR_STATE_ASY_CRC_MODE_MASK; + value &= ~SOR_STATE_ASY_OWNER_MASK; + + value |= SOR_STATE_ASY_CRC_MODE_COMPLETE | + SOR_STATE_ASY_OWNER(dc->pipe + 1); + + if (mode->flags & DRM_MODE_FLAG_PHSYNC) + value &= ~SOR_STATE_ASY_HSYNCPOL; + + if (mode->flags & DRM_MODE_FLAG_NHSYNC) + value |= SOR_STATE_ASY_HSYNCPOL; + + if (mode->flags & DRM_MODE_FLAG_PVSYNC) + value &= ~SOR_STATE_ASY_VSYNCPOL; + + if (mode->flags & DRM_MODE_FLAG_NVSYNC) + value |= SOR_STATE_ASY_VSYNCPOL; + + switch (state->bpc) { + case 16: + value |= SOR_STATE_ASY_PIXELDEPTH_BPP_48_444; + break; + + case 12: + value |= SOR_STATE_ASY_PIXELDEPTH_BPP_36_444; + break; + + case 10: + value |= SOR_STATE_ASY_PIXELDEPTH_BPP_30_444; + break; + + case 8: + value |= SOR_STATE_ASY_PIXELDEPTH_BPP_24_444; + break; + + case 6: + value |= SOR_STATE_ASY_PIXELDEPTH_BPP_18_444; + break; + + default: + value |= SOR_STATE_ASY_PIXELDEPTH_BPP_24_444; + break; + } + + tegra_sor_writel(sor, value, SOR_STATE1); + + /* + * TODO: The video timing programming below doesn't seem to match the + * register definitions. + */ + + value = ((mode->vtotal & 0x7fff) << 16) | (mode->htotal & 0x7fff); + tegra_sor_writel(sor, value, SOR_HEAD_STATE1(dc->pipe)); + + /* sync end = sync width - 1 */ + vse = mode->vsync_end - mode->vsync_start - 1; + hse = mode->hsync_end - mode->hsync_start - 1; + + value = ((vse & 0x7fff) << 16) | (hse & 0x7fff); + tegra_sor_writel(sor, value, SOR_HEAD_STATE2(dc->pipe)); + + /* blank end = sync end + back porch */ + vbe = vse + (mode->vtotal - mode->vsync_end); + hbe = hse + (mode->htotal - mode->hsync_end); + + value = ((vbe & 0x7fff) << 16) | (hbe & 0x7fff); + tegra_sor_writel(sor, value, SOR_HEAD_STATE3(dc->pipe)); + + /* blank start = blank end + active */ + vbs = vbe + mode->vdisplay; + hbs = hbe + mode->hdisplay; + + value = ((vbs & 0x7fff) << 16) | (hbs & 0x7fff); + tegra_sor_writel(sor, value, SOR_HEAD_STATE4(dc->pipe)); + + /* XXX interlacing support */ + tegra_sor_writel(sor, 0x001, SOR_HEAD_STATE5(dc->pipe)); +} + static int tegra_sor_detach(struct tegra_sor *sor) { unsigned long value, timeout; @@ -733,7 +992,8 @@ static int tegra_sor_power_down(struct tegra_sor *sor) if ((value & SOR_PWR_TRIGGER) != 0) return -ETIMEDOUT; - err = clk_set_parent(sor->clk, sor->clk_safe); + /* switch to safe parent clock */ + err = tegra_sor_set_parent_clock(sor, sor->clk_safe); if (err < 0) dev_err(sor->dev, "failed to set safe parent clock: %d\n", err); @@ -1038,6 +1298,22 @@ static void tegra_sor_debugfs_exit(struct tegra_sor *sor) sor->debugfs = NULL; } +static void tegra_sor_connector_reset(struct drm_connector *connector) +{ + struct tegra_sor_state *state; + + state = kzalloc(sizeof(*state), GFP_KERNEL); + if (!state) + return; + + if (connector->state) { + __drm_atomic_helper_connector_destroy_state(connector->state); + kfree(connector->state); + } + + __drm_atomic_helper_connector_reset(connector, &state->base); +} + static enum drm_connector_status tegra_sor_connector_detect(struct drm_connector *connector, bool force) { @@ -1050,13 +1326,28 @@ tegra_sor_connector_detect(struct drm_connector *connector, bool force) return tegra_output_connector_detect(connector, force); } +static struct drm_connector_state * +tegra_sor_connector_duplicate_state(struct drm_connector *connector) +{ + struct tegra_sor_state *state = to_sor_state(connector->state); + struct tegra_sor_state *copy; + + copy = kmemdup(state, sizeof(*state), GFP_KERNEL); + if (!copy) + return NULL; + + __drm_atomic_helper_connector_duplicate_state(connector, ©->base); + + return ©->base; +} + static const struct drm_connector_funcs tegra_sor_connector_funcs = { .dpms = drm_atomic_helper_connector_dpms, - .reset = drm_atomic_helper_connector_reset, + .reset = tegra_sor_connector_reset, .detect = tegra_sor_connector_detect, .fill_modes = drm_helper_probe_single_connector_modes, .destroy = tegra_output_connector_destroy, - .atomic_duplicate_state = drm_atomic_helper_connector_duplicate_state, + .atomic_duplicate_state = tegra_sor_connector_duplicate_state, .atomic_destroy_state = drm_atomic_helper_connector_destroy_state, }; @@ -1081,6 +1372,10 @@ static enum drm_mode_status tegra_sor_connector_mode_valid(struct drm_connector *connector, struct drm_display_mode *mode) { + /* HDMI 2.0 modes are not yet supported */ + if (mode->clock > 340000) + return MODE_NOCLOCK; + return MODE_OK; } @@ -1140,8 +1435,7 @@ static void tegra_sor_edp_disable(struct drm_encoder *encoder) if (output->panel) drm_panel_unprepare(output->panel); - reset_control_assert(sor->rst); - clk_disable_unprepare(sor->clk); + pm_runtime_put(sor->dev); } #if 0 @@ -1191,19 +1485,18 @@ static void tegra_sor_edp_enable(struct drm_encoder *encoder) struct drm_display_mode *mode = &encoder->crtc->state->adjusted_mode; struct tegra_output *output = encoder_to_output(encoder); struct tegra_dc *dc = to_tegra_dc(encoder->crtc); - unsigned int vbe, vse, hbe, hse, vbs, hbs, i; struct tegra_sor *sor = to_sor(output); struct tegra_sor_config config; + struct tegra_sor_state *state; struct drm_dp_link link; u8 rate, lanes; + unsigned int i; int err = 0; u32 value; - err = clk_prepare_enable(sor->clk); - if (err < 0) - dev_err(sor->dev, "failed to enable clock: %d\n", err); + state = to_sor_state(output->connector.state); - reset_control_deassert(sor->rst); + pm_runtime_get_sync(sor->dev); if (output->panel) drm_panel_prepare(output->panel); @@ -1218,17 +1511,17 @@ static void tegra_sor_edp_enable(struct drm_encoder *encoder) return; } - err = clk_set_parent(sor->clk, sor->clk_safe); + /* switch to safe parent clock */ + err = tegra_sor_set_parent_clock(sor, sor->clk_safe); if (err < 0) dev_err(sor->dev, "failed to set safe parent clock: %d\n", err); memset(&config, 0, sizeof(config)); - config.bits_per_pixel = output->connector.display_info.bpc * 3; + config.bits_per_pixel = state->bpc * 3; - err = tegra_sor_calc_config(sor, mode, &config, &link); + err = tegra_sor_compute_config(sor, mode, &config, &link); if (err < 0) - dev_err(sor->dev, "failed to compute link configuration: %d\n", - err); + dev_err(sor->dev, "failed to compute configuration: %d\n", err); value = tegra_sor_readl(sor, SOR_CLK_CNTRL); value &= ~SOR_CLK_CNTRL_DP_CLK_SEL_MASK; @@ -1325,10 +1618,18 @@ static void tegra_sor_edp_enable(struct drm_encoder *encoder) value &= ~SOR_PLL2_PORT_POWERDOWN; tegra_sor_writel(sor, value, SOR_PLL2); - /* switch to DP clock */ - err = clk_set_parent(sor->clk, sor->clk_dp); + /* XXX not in TRM */ + for (value = 0, i = 0; i < 5; i++) + value |= SOR_XBAR_CTRL_LINK0_XSEL(i, sor->soc->xbar_cfg[i]) | + SOR_XBAR_CTRL_LINK1_XSEL(i, i); + + tegra_sor_writel(sor, 0x00000000, SOR_XBAR_POL); + tegra_sor_writel(sor, value, SOR_XBAR_CTRL); + + /* switch to DP parent clock */ + err = tegra_sor_set_parent_clock(sor, sor->clk_dp); if (err < 0) - dev_err(sor->dev, "failed to set DP parent clock: %d\n", err); + dev_err(sor->dev, "failed to set parent clock: %d\n", err); /* power DP lanes */ value = tegra_sor_readl(sor, SOR_DP_PADCTL0); @@ -1374,13 +1675,11 @@ static void tegra_sor_edp_enable(struct drm_encoder *encoder) value |= drm_dp_link_rate_to_bw_code(link.rate) << 2; tegra_sor_writel(sor, value, SOR_CLK_CNTRL); - /* set linkctl */ + tegra_sor_apply_config(sor, &config); + + /* enable link */ value = tegra_sor_readl(sor, SOR_DP_LINKCTL0); value |= SOR_DP_LINKCTL_ENABLE; - - value &= ~SOR_DP_LINKCTL_TU_SIZE_MASK; - value |= SOR_DP_LINKCTL_TU_SIZE(config.tu_size); - value |= SOR_DP_LINKCTL_ENHANCED_FRAME; tegra_sor_writel(sor, value, SOR_DP_LINKCTL0); @@ -1393,35 +1692,6 @@ static void tegra_sor_edp_enable(struct drm_encoder *encoder) tegra_sor_writel(sor, value, SOR_DP_TPG); - value = tegra_sor_readl(sor, SOR_DP_CONFIG0); - value &= ~SOR_DP_CONFIG_WATERMARK_MASK; - value |= SOR_DP_CONFIG_WATERMARK(config.watermark); - - value &= ~SOR_DP_CONFIG_ACTIVE_SYM_COUNT_MASK; - value |= SOR_DP_CONFIG_ACTIVE_SYM_COUNT(config.active_count); - - value &= ~SOR_DP_CONFIG_ACTIVE_SYM_FRAC_MASK; - value |= SOR_DP_CONFIG_ACTIVE_SYM_FRAC(config.active_frac); - - if (config.active_polarity) - value |= SOR_DP_CONFIG_ACTIVE_SYM_POLARITY; - else - value &= ~SOR_DP_CONFIG_ACTIVE_SYM_POLARITY; - - value |= SOR_DP_CONFIG_ACTIVE_SYM_ENABLE; - value |= SOR_DP_CONFIG_DISPARITY_NEGATIVE; - tegra_sor_writel(sor, value, SOR_DP_CONFIG0); - - value = tegra_sor_readl(sor, SOR_DP_AUDIO_HBLANK_SYMBOLS); - value &= ~SOR_DP_AUDIO_HBLANK_SYMBOLS_MASK; - value |= config.hblank_symbols & 0xffff; - tegra_sor_writel(sor, value, SOR_DP_AUDIO_HBLANK_SYMBOLS); - - value = tegra_sor_readl(sor, SOR_DP_AUDIO_VBLANK_SYMBOLS); - value &= ~SOR_DP_AUDIO_VBLANK_SYMBOLS_MASK; - value |= config.vblank_symbols & 0xffff; - tegra_sor_writel(sor, value, SOR_DP_AUDIO_VBLANK_SYMBOLS); - /* enable pad calibration logic */ value = tegra_sor_readl(sor, SOR_DP_PADCTL0); value |= SOR_DP_PADCTL_PAD_CAL_PD; @@ -1477,75 +1747,19 @@ static void tegra_sor_edp_enable(struct drm_encoder *encoder) if (err < 0) dev_err(sor->dev, "failed to power up SOR: %d\n", err); - /* - * configure panel (24bpp, vsync-, hsync-, DP-A protocol, complete - * raster, associate with display controller) - */ - value = SOR_STATE_ASY_PROTOCOL_DP_A | - SOR_STATE_ASY_CRC_MODE_COMPLETE | - SOR_STATE_ASY_OWNER(dc->pipe + 1); - - if (mode->flags & DRM_MODE_FLAG_PHSYNC) - value &= ~SOR_STATE_ASY_HSYNCPOL; - - if (mode->flags & DRM_MODE_FLAG_NHSYNC) - value |= SOR_STATE_ASY_HSYNCPOL; - - if (mode->flags & DRM_MODE_FLAG_PVSYNC) - value &= ~SOR_STATE_ASY_VSYNCPOL; - - if (mode->flags & DRM_MODE_FLAG_NVSYNC) - value |= SOR_STATE_ASY_VSYNCPOL; - - switch (config.bits_per_pixel) { - case 24: - value |= SOR_STATE_ASY_PIXELDEPTH_BPP_24_444; - break; - - case 18: - value |= SOR_STATE_ASY_PIXELDEPTH_BPP_18_444; - break; - - default: - BUG(); - break; - } - - tegra_sor_writel(sor, value, SOR_STATE1); - - /* - * TODO: The video timing programming below doesn't seem to match the - * register definitions. - */ - - value = ((mode->vtotal & 0x7fff) << 16) | (mode->htotal & 0x7fff); - tegra_sor_writel(sor, value, SOR_HEAD_STATE1(dc->pipe)); - - vse = mode->vsync_end - mode->vsync_start - 1; - hse = mode->hsync_end - mode->hsync_start - 1; - - value = ((vse & 0x7fff) << 16) | (hse & 0x7fff); - tegra_sor_writel(sor, value, SOR_HEAD_STATE2(dc->pipe)); - - vbe = vse + (mode->vsync_start - mode->vdisplay); - hbe = hse + (mode->hsync_start - mode->hdisplay); - - value = ((vbe & 0x7fff) << 16) | (hbe & 0x7fff); - tegra_sor_writel(sor, value, SOR_HEAD_STATE3(dc->pipe)); - - vbs = vbe + mode->vdisplay; - hbs = hbe + mode->hdisplay; - - value = ((vbs & 0x7fff) << 16) | (hbs & 0x7fff); - tegra_sor_writel(sor, value, SOR_HEAD_STATE4(dc->pipe)); - - tegra_sor_writel(sor, 0x1, SOR_HEAD_STATE5(dc->pipe)); - /* CSTM (LVDS, link A/B, upper) */ value = SOR_CSTM_LVDS | SOR_CSTM_LINK_ACT_A | SOR_CSTM_LINK_ACT_B | SOR_CSTM_UPPER; tegra_sor_writel(sor, value, SOR_CSTM); + /* use DP-A protocol */ + value = tegra_sor_readl(sor, SOR_STATE1); + value &= ~SOR_STATE_ASY_PROTOCOL_MASK; + value |= SOR_STATE_ASY_PROTOCOL_DP_A; + tegra_sor_writel(sor, value, SOR_STATE1); + + tegra_sor_mode_set(sor, mode, state); + /* PWM setup */ err = tegra_sor_setup_pwm(sor, 250); if (err < 0) @@ -1577,11 +1791,15 @@ tegra_sor_encoder_atomic_check(struct drm_encoder *encoder, struct drm_connector_state *conn_state) { struct tegra_output *output = encoder_to_output(encoder); + struct tegra_sor_state *state = to_sor_state(conn_state); struct tegra_dc *dc = to_tegra_dc(conn_state->crtc); unsigned long pclk = crtc_state->mode.clock * 1000; struct tegra_sor *sor = to_sor(output); + struct drm_display_info *info; int err; + info = &output->connector.display_info; + err = tegra_dc_state_setup_clock(dc, crtc_state, sor->clk_parent, pclk, 0); if (err < 0) { @@ -1589,6 +1807,18 @@ tegra_sor_encoder_atomic_check(struct drm_encoder *encoder, return err; } + switch (info->bpc) { + case 8: + case 6: + state->bpc = info->bpc; + break; + + default: + DRM_DEBUG_KMS("%u bits-per-color not supported\n", info->bpc); + state->bpc = 8; + break; + } + return 0; } @@ -1751,9 +1981,7 @@ static void tegra_sor_hdmi_disable(struct drm_encoder *encoder) if (err < 0) dev_err(sor->dev, "failed to power off HDMI rail: %d\n", err); - reset_control_assert(sor->rst); - usleep_range(1000, 2000); - clk_disable_unprepare(sor->clk); + pm_runtime_put(sor->dev); } static void tegra_sor_hdmi_enable(struct drm_encoder *encoder) @@ -1761,26 +1989,21 @@ static void tegra_sor_hdmi_enable(struct drm_encoder *encoder) struct tegra_output *output = encoder_to_output(encoder); unsigned int h_ref_to_sync = 1, pulse_start, max_ac; struct tegra_dc *dc = to_tegra_dc(encoder->crtc); - unsigned int vbe, vse, hbe, hse, vbs, hbs, div; struct tegra_sor_hdmi_settings *settings; struct tegra_sor *sor = to_sor(output); + struct tegra_sor_state *state; struct drm_display_mode *mode; - struct drm_display_info *info; + unsigned int div, i; u32 value; int err; + state = to_sor_state(output->connector.state); mode = &encoder->crtc->state->adjusted_mode; - info = &output->connector.display_info; - err = clk_prepare_enable(sor->clk); - if (err < 0) - dev_err(sor->dev, "failed to enable clock: %d\n", err); + pm_runtime_get_sync(sor->dev); - usleep_range(1000, 2000); - - reset_control_deassert(sor->rst); - - err = clk_set_parent(sor->clk, sor->clk_safe); + /* switch to safe parent clock */ + err = tegra_sor_set_parent_clock(sor, sor->clk_safe); if (err < 0) dev_err(sor->dev, "failed to set safe parent clock: %d\n", err); @@ -1876,22 +2099,20 @@ static void tegra_sor_hdmi_enable(struct drm_encoder *encoder) value = SOR_REFCLK_DIV_INT(div) | SOR_REFCLK_DIV_FRAC(div); tegra_sor_writel(sor, value, SOR_REFCLK); - /* XXX don't hardcode */ - value = SOR_XBAR_CTRL_LINK1_XSEL(4, 4) | - SOR_XBAR_CTRL_LINK1_XSEL(3, 3) | - SOR_XBAR_CTRL_LINK1_XSEL(2, 2) | - SOR_XBAR_CTRL_LINK1_XSEL(1, 1) | - SOR_XBAR_CTRL_LINK1_XSEL(0, 0) | - SOR_XBAR_CTRL_LINK0_XSEL(4, 4) | - SOR_XBAR_CTRL_LINK0_XSEL(3, 3) | - SOR_XBAR_CTRL_LINK0_XSEL(2, 0) | - SOR_XBAR_CTRL_LINK0_XSEL(1, 1) | - SOR_XBAR_CTRL_LINK0_XSEL(0, 2); - tegra_sor_writel(sor, value, SOR_XBAR_CTRL); + /* XXX not in TRM */ + for (value = 0, i = 0; i < 5; i++) + value |= SOR_XBAR_CTRL_LINK0_XSEL(i, sor->soc->xbar_cfg[i]) | + SOR_XBAR_CTRL_LINK1_XSEL(i, i); tegra_sor_writel(sor, 0x00000000, SOR_XBAR_POL); + tegra_sor_writel(sor, value, SOR_XBAR_CTRL); - err = clk_set_parent(sor->clk, sor->clk_parent); + /* switch to parent clock */ + err = clk_set_parent(sor->clk_src, sor->clk_parent); + if (err < 0) + dev_err(sor->dev, "failed to set source clock: %d\n", err); + + err = tegra_sor_set_parent_clock(sor, sor->clk_src); if (err < 0) dev_err(sor->dev, "failed to set parent clock: %d\n", err); @@ -2001,7 +2222,7 @@ static void tegra_sor_hdmi_enable(struct drm_encoder *encoder) value &= ~DITHER_CONTROL_MASK; value &= ~BASE_COLOR_SIZE_MASK; - switch (info->bpc) { + switch (state->bpc) { case 6: value |= BASE_COLOR_SIZE_666; break; @@ -2011,7 +2232,8 @@ static void tegra_sor_hdmi_enable(struct drm_encoder *encoder) break; default: - WARN(1, "%u bits-per-color not supported\n", info->bpc); + WARN(1, "%u bits-per-color not supported\n", state->bpc); + value |= BASE_COLOR_SIZE_888; break; } @@ -2021,83 +2243,19 @@ static void tegra_sor_hdmi_enable(struct drm_encoder *encoder) if (err < 0) dev_err(sor->dev, "failed to power up SOR: %d\n", err); - /* configure mode */ - value = tegra_sor_readl(sor, SOR_STATE1); - value &= ~SOR_STATE_ASY_PIXELDEPTH_MASK; - value &= ~SOR_STATE_ASY_CRC_MODE_MASK; - value &= ~SOR_STATE_ASY_OWNER_MASK; - - value |= SOR_STATE_ASY_CRC_MODE_COMPLETE | - SOR_STATE_ASY_OWNER(dc->pipe + 1); - - if (mode->flags & DRM_MODE_FLAG_PHSYNC) - value &= ~SOR_STATE_ASY_HSYNCPOL; - - if (mode->flags & DRM_MODE_FLAG_NHSYNC) - value |= SOR_STATE_ASY_HSYNCPOL; - - if (mode->flags & DRM_MODE_FLAG_PVSYNC) - value &= ~SOR_STATE_ASY_VSYNCPOL; - - if (mode->flags & DRM_MODE_FLAG_NVSYNC) - value |= SOR_STATE_ASY_VSYNCPOL; - - switch (info->bpc) { - case 8: - value |= SOR_STATE_ASY_PIXELDEPTH_BPP_24_444; - break; - - case 6: - value |= SOR_STATE_ASY_PIXELDEPTH_BPP_18_444; - break; - - default: - BUG(); - break; - } - - tegra_sor_writel(sor, value, SOR_STATE1); - + /* configure dynamic range of output */ value = tegra_sor_readl(sor, SOR_HEAD_STATE0(dc->pipe)); value &= ~SOR_HEAD_STATE_RANGECOMPRESS_MASK; value &= ~SOR_HEAD_STATE_DYNRANGE_MASK; tegra_sor_writel(sor, value, SOR_HEAD_STATE0(dc->pipe)); + /* configure colorspace */ value = tegra_sor_readl(sor, SOR_HEAD_STATE0(dc->pipe)); value &= ~SOR_HEAD_STATE_COLORSPACE_MASK; value |= SOR_HEAD_STATE_COLORSPACE_RGB; tegra_sor_writel(sor, value, SOR_HEAD_STATE0(dc->pipe)); - /* - * TODO: The video timing programming below doesn't seem to match the - * register definitions. - */ - - value = ((mode->vtotal & 0x7fff) << 16) | (mode->htotal & 0x7fff); - tegra_sor_writel(sor, value, SOR_HEAD_STATE1(dc->pipe)); - - /* sync end = sync width - 1 */ - vse = mode->vsync_end - mode->vsync_start - 1; - hse = mode->hsync_end - mode->hsync_start - 1; - - value = ((vse & 0x7fff) << 16) | (hse & 0x7fff); - tegra_sor_writel(sor, value, SOR_HEAD_STATE2(dc->pipe)); - - /* blank end = sync end + back porch */ - vbe = vse + (mode->vtotal - mode->vsync_end); - hbe = hse + (mode->htotal - mode->hsync_end); - - value = ((vbe & 0x7fff) << 16) | (hbe & 0x7fff); - tegra_sor_writel(sor, value, SOR_HEAD_STATE3(dc->pipe)); - - /* blank start = blank end + active */ - vbs = vbe + mode->vdisplay; - hbs = hbe + mode->hdisplay; - - value = ((vbs & 0x7fff) << 16) | (hbs & 0x7fff); - tegra_sor_writel(sor, value, SOR_HEAD_STATE4(dc->pipe)); - - tegra_sor_writel(sor, 0x1, SOR_HEAD_STATE5(dc->pipe)); + tegra_sor_mode_set(sor, mode, state); tegra_sor_update(sor); @@ -2195,10 +2353,13 @@ static int tegra_sor_init(struct host1x_client *client) * XXX: Remove this reset once proper hand-over from firmware to * kernel is possible. */ - err = reset_control_assert(sor->rst); - if (err < 0) { - dev_err(sor->dev, "failed to assert SOR reset: %d\n", err); - return err; + if (sor->rst) { + err = reset_control_assert(sor->rst); + if (err < 0) { + dev_err(sor->dev, "failed to assert SOR reset: %d\n", + err); + return err; + } } err = clk_prepare_enable(sor->clk); @@ -2209,10 +2370,13 @@ static int tegra_sor_init(struct host1x_client *client) usleep_range(1000, 3000); - err = reset_control_deassert(sor->rst); - if (err < 0) { - dev_err(sor->dev, "failed to deassert SOR reset: %d\n", err); - return err; + if (sor->rst) { + err = reset_control_deassert(sor->rst); + if (err < 0) { + dev_err(sor->dev, "failed to deassert SOR reset: %d\n", + err); + return err; + } } err = clk_prepare_enable(sor->clk_safe); @@ -2323,11 +2487,16 @@ static const struct tegra_sor_ops tegra_sor_hdmi_ops = { .remove = tegra_sor_hdmi_remove, }; +static const u8 tegra124_sor_xbar_cfg[5] = { + 0, 1, 2, 3, 4 +}; + static const struct tegra_sor_soc tegra124_sor = { .supports_edp = true, .supports_lvds = true, .supports_hdmi = false, .supports_dp = false, + .xbar_cfg = tegra124_sor_xbar_cfg, }; static const struct tegra_sor_soc tegra210_sor = { @@ -2335,6 +2504,11 @@ static const struct tegra_sor_soc tegra210_sor = { .supports_lvds = false, .supports_hdmi = false, .supports_dp = false, + .xbar_cfg = tegra124_sor_xbar_cfg, +}; + +static const u8 tegra210_sor_xbar_cfg[5] = { + 2, 1, 0, 3, 4 }; static const struct tegra_sor_soc tegra210_sor1 = { @@ -2345,6 +2519,8 @@ static const struct tegra_sor_soc tegra210_sor1 = { .num_settings = ARRAY_SIZE(tegra210_sor_hdmi_defaults), .settings = tegra210_sor_hdmi_defaults, + + .xbar_cfg = tegra210_sor_xbar_cfg, }; static const struct of_device_id tegra_sor_of_match[] = { @@ -2434,11 +2610,14 @@ static int tegra_sor_probe(struct platform_device *pdev) goto remove; } - sor->rst = devm_reset_control_get(&pdev->dev, "sor"); - if (IS_ERR(sor->rst)) { - err = PTR_ERR(sor->rst); - dev_err(&pdev->dev, "failed to get reset control: %d\n", err); - goto remove; + if (!pdev->dev.pm_domain) { + sor->rst = devm_reset_control_get(&pdev->dev, "sor"); + if (IS_ERR(sor->rst)) { + err = PTR_ERR(sor->rst); + dev_err(&pdev->dev, "failed to get reset control: %d\n", + err); + goto remove; + } } sor->clk = devm_clk_get(&pdev->dev, NULL); @@ -2448,6 +2627,16 @@ static int tegra_sor_probe(struct platform_device *pdev) goto remove; } + if (sor->soc->supports_hdmi || sor->soc->supports_dp) { + sor->clk_src = devm_clk_get(&pdev->dev, "source"); + if (IS_ERR(sor->clk_src)) { + err = PTR_ERR(sor->clk_src); + dev_err(sor->dev, "failed to get source clock: %d\n", + err); + goto remove; + } + } + sor->clk_parent = devm_clk_get(&pdev->dev, "parent"); if (IS_ERR(sor->clk_parent)) { err = PTR_ERR(sor->clk_parent); @@ -2469,6 +2658,19 @@ static int tegra_sor_probe(struct platform_device *pdev) goto remove; } + platform_set_drvdata(pdev, sor); + pm_runtime_enable(&pdev->dev); + + pm_runtime_get_sync(&pdev->dev); + sor->clk_brick = tegra_clk_sor_brick_register(sor, "sor1_brick"); + pm_runtime_put(&pdev->dev); + + if (IS_ERR(sor->clk_brick)) { + err = PTR_ERR(sor->clk_brick); + dev_err(&pdev->dev, "failed to register SOR clock: %d\n", err); + goto remove; + } + INIT_LIST_HEAD(&sor->client.list); sor->client.ops = &sor_client_ops; sor->client.dev = &pdev->dev; @@ -2480,8 +2682,6 @@ static int tegra_sor_probe(struct platform_device *pdev) goto remove; } - platform_set_drvdata(pdev, sor); - return 0; remove: @@ -2497,6 +2697,8 @@ static int tegra_sor_remove(struct platform_device *pdev) struct tegra_sor *sor = platform_get_drvdata(pdev); int err; + pm_runtime_disable(&pdev->dev); + err = host1x_client_unregister(&sor->client); if (err < 0) { dev_err(&pdev->dev, "failed to unregister host1x client: %d\n", @@ -2515,10 +2717,62 @@ static int tegra_sor_remove(struct platform_device *pdev) return 0; } +#ifdef CONFIG_PM +static int tegra_sor_suspend(struct device *dev) +{ + struct tegra_sor *sor = dev_get_drvdata(dev); + int err; + + if (sor->rst) { + err = reset_control_assert(sor->rst); + if (err < 0) { + dev_err(dev, "failed to assert reset: %d\n", err); + return err; + } + } + + usleep_range(1000, 2000); + + clk_disable_unprepare(sor->clk); + + return 0; +} + +static int tegra_sor_resume(struct device *dev) +{ + struct tegra_sor *sor = dev_get_drvdata(dev); + int err; + + err = clk_prepare_enable(sor->clk); + if (err < 0) { + dev_err(dev, "failed to enable clock: %d\n", err); + return err; + } + + usleep_range(1000, 2000); + + if (sor->rst) { + err = reset_control_deassert(sor->rst); + if (err < 0) { + dev_err(dev, "failed to deassert reset: %d\n", err); + clk_disable_unprepare(sor->clk); + return err; + } + } + + return 0; +} +#endif + +static const struct dev_pm_ops tegra_sor_pm_ops = { + SET_RUNTIME_PM_OPS(tegra_sor_suspend, tegra_sor_resume, NULL) +}; + struct platform_driver tegra_sor_driver = { .driver = { .name = "tegra-sor", .of_match_table = tegra_sor_of_match, + .pm = &tegra_sor_pm_ops, }, .probe = tegra_sor_probe, .remove = tegra_sor_remove, diff --git a/drivers/gpu/drm/tegra/sor.h b/drivers/gpu/drm/tegra/sor.h index 2d31d027e3f6..865c73b48968 100644 --- a/drivers/gpu/drm/tegra/sor.h +++ b/drivers/gpu/drm/tegra/sor.h @@ -27,6 +27,9 @@ #define SOR_STATE_ASY_PIXELDEPTH_MASK (0xf << 17) #define SOR_STATE_ASY_PIXELDEPTH_BPP_18_444 (0x2 << 17) #define SOR_STATE_ASY_PIXELDEPTH_BPP_24_444 (0x5 << 17) +#define SOR_STATE_ASY_PIXELDEPTH_BPP_30_444 (0x6 << 17) +#define SOR_STATE_ASY_PIXELDEPTH_BPP_36_444 (0x8 << 17) +#define SOR_STATE_ASY_PIXELDEPTH_BPP_48_444 (0x9 << 17) #define SOR_STATE_ASY_VSYNCPOL (1 << 13) #define SOR_STATE_ASY_HSYNCPOL (1 << 12) #define SOR_STATE_ASY_PROTOCOL_MASK (0xf << 8) diff --git a/drivers/gpu/drm/tilcdc/Kconfig b/drivers/gpu/drm/tilcdc/Kconfig index f60a1ec84fa4..28fed7e206d0 100644 --- a/drivers/gpu/drm/tilcdc/Kconfig +++ b/drivers/gpu/drm/tilcdc/Kconfig @@ -2,7 +2,6 @@ config DRM_TILCDC tristate "DRM Support for TI LCDC Display Controller" depends on DRM && OF && ARM select DRM_KMS_HELPER - select DRM_KMS_FB_HELPER select DRM_KMS_CMA_HELPER select DRM_GEM_CMA_HELPER select VIDEOMODE_HELPERS diff --git a/drivers/gpu/drm/tilcdc/tilcdc_crtc.c b/drivers/gpu/drm/tilcdc/tilcdc_crtc.c index 79027b1c64d3..107c8bd04f6d 100644 --- a/drivers/gpu/drm/tilcdc/tilcdc_crtc.c +++ b/drivers/gpu/drm/tilcdc/tilcdc_crtc.c @@ -697,7 +697,7 @@ irqreturn_t tilcdc_crtc_irq(struct drm_crtc *crtc) spin_unlock_irqrestore(&tilcdc_crtc->irq_lock, flags); - drm_handle_vblank(dev, 0); + drm_crtc_handle_vblank(crtc); if (!skip_event) { struct drm_pending_vblank_event *event; diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c index 39386f50af87..4054d804fe06 100644 --- a/drivers/gpu/drm/ttm/ttm_bo.c +++ b/drivers/gpu/drm/ttm/ttm_bo.c @@ -146,10 +146,9 @@ static void ttm_bo_release_list(struct kref *list_kref) BUG_ON(bo->mem.mm_node != NULL); BUG_ON(!list_empty(&bo->lru)); BUG_ON(!list_empty(&bo->ddestroy)); - - if (bo->ttm) - ttm_tt_destroy(bo->ttm); + ttm_tt_destroy(bo->ttm); atomic_dec(&bo->glob->bo_count); + fence_put(bo->moving); if (bo->resv == &bo->ttm_resv) reservation_object_fini(&bo->ttm_resv); mutex_destroy(&bo->wu_mutex); @@ -360,7 +359,8 @@ static int ttm_bo_handle_move_mem(struct ttm_buffer_object *bo, ret = bdev->driver->move(bo, evict, interruptible, no_wait_gpu, mem); else - ret = ttm_bo_move_memcpy(bo, evict, no_wait_gpu, mem); + ret = ttm_bo_move_memcpy(bo, evict, interruptible, + no_wait_gpu, mem); if (ret) { if (bdev->driver->move_notify) { @@ -396,8 +396,7 @@ moved: out_err: new_man = &bdev->man[bo->mem.mem_type]; - if ((new_man->flags & TTM_MEMTYPE_FLAG_FIXED) && bo->ttm) { - ttm_tt_unbind(bo->ttm); + if (new_man->flags & TTM_MEMTYPE_FLAG_FIXED) { ttm_tt_destroy(bo->ttm); bo->ttm = NULL; } @@ -418,11 +417,8 @@ static void ttm_bo_cleanup_memtype_use(struct ttm_buffer_object *bo) if (bo->bdev->driver->move_notify) bo->bdev->driver->move_notify(bo, NULL); - if (bo->ttm) { - ttm_tt_unbind(bo->ttm); - ttm_tt_destroy(bo->ttm); - bo->ttm = NULL; - } + ttm_tt_destroy(bo->ttm); + bo->ttm = NULL; ttm_bo_mem_put(bo, &bo->mem); ww_mutex_unlock (&bo->resv->lock); @@ -688,15 +684,6 @@ static int ttm_bo_evict(struct ttm_buffer_object *bo, bool interruptible, struct ttm_placement placement; int ret = 0; - ret = ttm_bo_wait(bo, interruptible, no_wait_gpu); - - if (unlikely(ret != 0)) { - if (ret != -ERESTARTSYS) { - pr_err("Failed to expire sync object before buffer eviction\n"); - } - goto out; - } - lockdep_assert_held(&bo->resv->lock.base); evict_mem = bo->mem; @@ -720,7 +707,7 @@ static int ttm_bo_evict(struct ttm_buffer_object *bo, bool interruptible, ret = ttm_bo_handle_move_mem(bo, &evict_mem, true, interruptible, no_wait_gpu); - if (ret) { + if (unlikely(ret)) { if (ret != -ERESTARTSYS) pr_err("Buffer eviction failed\n"); ttm_bo_mem_put(bo, &evict_mem); @@ -800,6 +787,34 @@ void ttm_bo_mem_put(struct ttm_buffer_object *bo, struct ttm_mem_reg *mem) EXPORT_SYMBOL(ttm_bo_mem_put); /** + * Add the last move fence to the BO and reserve a new shared slot. + */ +static int ttm_bo_add_move_fence(struct ttm_buffer_object *bo, + struct ttm_mem_type_manager *man, + struct ttm_mem_reg *mem) +{ + struct fence *fence; + int ret; + + spin_lock(&man->move_lock); + fence = fence_get(man->move); + spin_unlock(&man->move_lock); + + if (fence) { + reservation_object_add_shared_fence(bo->resv, fence); + + ret = reservation_object_reserve_shared(bo->resv); + if (unlikely(ret)) + return ret; + + fence_put(bo->moving); + bo->moving = fence; + } + + return 0; +} + +/** * Repeatedly evict memory from the LRU for @mem_type until we create enough * space, or we've evicted everything and there isn't enough space. */ @@ -825,10 +840,8 @@ static int ttm_bo_mem_force_space(struct ttm_buffer_object *bo, if (unlikely(ret != 0)) return ret; } while (1); - if (mem->mm_node == NULL) - return -ENOMEM; mem->mem_type = mem_type; - return 0; + return ttm_bo_add_move_fence(bo, man, mem); } static uint32_t ttm_bo_select_caching(struct ttm_mem_type_manager *man, @@ -898,6 +911,10 @@ int ttm_bo_mem_space(struct ttm_buffer_object *bo, bool has_erestartsys = false; int i, ret; + ret = reservation_object_reserve_shared(bo->resv); + if (unlikely(ret)) + return ret; + mem->mm_node = NULL; for (i = 0; i < placement->num_placement; ++i) { const struct ttm_place *place = &placement->placement[i]; @@ -931,9 +948,15 @@ int ttm_bo_mem_space(struct ttm_buffer_object *bo, ret = (*man->func->get_node)(man, bo, place, mem); if (unlikely(ret)) return ret; - - if (mem->mm_node) + + if (mem->mm_node) { + ret = ttm_bo_add_move_fence(bo, man, mem); + if (unlikely(ret)) { + (*man->func->put_node)(man, mem); + return ret; + } break; + } } if ((type_ok && (mem_type == TTM_PL_SYSTEM)) || mem->mm_node) { @@ -1000,20 +1023,6 @@ static int ttm_bo_move_buffer(struct ttm_buffer_object *bo, lockdep_assert_held(&bo->resv->lock.base); - /* - * Don't wait for the BO on initial allocation. This is important when - * the BO has an imported reservation object. - */ - if (bo->mem.mem_type != TTM_PL_SYSTEM || bo->ttm != NULL) { - /* - * FIXME: It's possible to pipeline buffer moves. - * Have the driver move function wait for idle when necessary, - * instead of doing it here. - */ - ret = ttm_bo_wait(bo, interruptible, no_wait_gpu); - if (ret) - return ret; - } mem.num_pages = bo->num_pages; mem.size = mem.num_pages << PAGE_SHIFT; mem.page_alignment = bo->mem.page_alignment; @@ -1034,9 +1043,9 @@ out_unlock: return ret; } -static bool ttm_bo_mem_compat(struct ttm_placement *placement, - struct ttm_mem_reg *mem, - uint32_t *new_flags) +bool ttm_bo_mem_compat(struct ttm_placement *placement, + struct ttm_mem_reg *mem, + uint32_t *new_flags) { int i; @@ -1068,6 +1077,7 @@ static bool ttm_bo_mem_compat(struct ttm_placement *placement, return false; } +EXPORT_SYMBOL(ttm_bo_mem_compat); int ttm_bo_validate(struct ttm_buffer_object *bo, struct ttm_placement *placement, @@ -1165,7 +1175,7 @@ int ttm_bo_init(struct ttm_bo_device *bdev, bo->mem.page_alignment = page_alignment; bo->mem.bus.io_reserved_vm = false; bo->mem.bus.io_reserved_count = 0; - bo->priv_flags = 0; + bo->moving = NULL; bo->mem.placement = (TTM_PL_FLAG_SYSTEM | TTM_PL_FLAG_CACHED); bo->persistent_swap_storage = persistent_swap_storage; bo->acc_size = acc_size; @@ -1277,6 +1287,7 @@ static int ttm_bo_force_list_clean(struct ttm_bo_device *bdev, { struct ttm_mem_type_manager *man = &bdev->man[mem_type]; struct ttm_bo_global *glob = bdev->glob; + struct fence *fence; int ret; /* @@ -1297,6 +1308,23 @@ static int ttm_bo_force_list_clean(struct ttm_bo_device *bdev, spin_lock(&glob->lru_lock); } spin_unlock(&glob->lru_lock); + + spin_lock(&man->move_lock); + fence = fence_get(man->move); + spin_unlock(&man->move_lock); + + if (fence) { + ret = fence_wait(fence, false); + fence_put(fence); + if (ret) { + if (allow_errors) { + return ret; + } else { + pr_err("Cleanup eviction failed\n"); + } + } + } + return 0; } @@ -1316,6 +1344,7 @@ int ttm_bo_clean_mm(struct ttm_bo_device *bdev, unsigned mem_type) mem_type); return ret; } + fence_put(man->move); man->use_type = false; man->has_type = false; @@ -1361,6 +1390,7 @@ int ttm_bo_init_mm(struct ttm_bo_device *bdev, unsigned type, man->io_reserve_fastpath = true; man->use_io_reserve_lru = false; mutex_init(&man->io_reserve_mutex); + spin_lock_init(&man->move_lock); INIT_LIST_HEAD(&man->io_reserve_lru); ret = bdev->driver->init_mem_type(bdev, type, man); @@ -1379,6 +1409,7 @@ int ttm_bo_init_mm(struct ttm_bo_device *bdev, unsigned type, man->size = p_size; INIT_LIST_HEAD(&man->lru); + man->move = NULL; return 0; } @@ -1572,47 +1603,17 @@ EXPORT_SYMBOL(ttm_bo_unmap_virtual); int ttm_bo_wait(struct ttm_buffer_object *bo, bool interruptible, bool no_wait) { - struct reservation_object_list *fobj; - struct reservation_object *resv; - struct fence *excl; - long timeout = 15 * HZ; - int i; - - resv = bo->resv; - fobj = reservation_object_get_list(resv); - excl = reservation_object_get_excl(resv); - if (excl) { - if (!fence_is_signaled(excl)) { - if (no_wait) - return -EBUSY; - - timeout = fence_wait_timeout(excl, - interruptible, timeout); - } - } - - for (i = 0; fobj && timeout > 0 && i < fobj->shared_count; ++i) { - struct fence *fence; - fence = rcu_dereference_protected(fobj->shared[i], - reservation_object_held(resv)); - - if (!fence_is_signaled(fence)) { - if (no_wait) - return -EBUSY; - - timeout = fence_wait_timeout(fence, - interruptible, timeout); - } - } + long timeout = no_wait ? 0 : 15 * HZ; + timeout = reservation_object_wait_timeout_rcu(bo->resv, true, + interruptible, timeout); if (timeout < 0) return timeout; if (timeout == 0) return -EBUSY; - reservation_object_add_excl_fence(resv, NULL); - clear_bit(TTM_BO_PRIV_FLAG_MOVING, &bo->priv_flags); + reservation_object_add_excl_fence(bo->resv, NULL); return 0; } EXPORT_SYMBOL(ttm_bo_wait); @@ -1682,14 +1683,9 @@ static int ttm_bo_swapout(struct ttm_mem_shrink *shrink) ttm_bo_list_ref_sub(bo, put_count, true); /** - * Wait for GPU, then move to system cached. + * Move to system cached */ - ret = ttm_bo_wait(bo, false, false); - - if (unlikely(ret != 0)) - goto out; - if ((bo->mem.placement & swap_placement) != swap_placement) { struct ttm_mem_reg evict_mem; @@ -1704,6 +1700,14 @@ static int ttm_bo_swapout(struct ttm_mem_shrink *shrink) goto out; } + /** + * Make sure BO is idle. + */ + + ret = ttm_bo_wait(bo, false, false); + if (unlikely(ret != 0)) + goto out; + ttm_bo_unmap_virtual(bo); /** diff --git a/drivers/gpu/drm/ttm/ttm_bo_util.c b/drivers/gpu/drm/ttm/ttm_bo_util.c index d9831559706e..2df602a35f92 100644 --- a/drivers/gpu/drm/ttm/ttm_bo_util.c +++ b/drivers/gpu/drm/ttm/ttm_bo_util.c @@ -321,7 +321,8 @@ static int ttm_copy_ttm_io_page(struct ttm_tt *ttm, void *dst, } int ttm_bo_move_memcpy(struct ttm_buffer_object *bo, - bool evict, bool no_wait_gpu, + bool evict, bool interruptible, + bool no_wait_gpu, struct ttm_mem_reg *new_mem) { struct ttm_bo_device *bdev = bo->bdev; @@ -337,6 +338,10 @@ int ttm_bo_move_memcpy(struct ttm_buffer_object *bo, unsigned long add = 0; int dir; + ret = ttm_bo_wait(bo, interruptible, no_wait_gpu); + if (ret) + return ret; + ret = ttm_mem_reg_ioremap(bdev, old_mem, &old_iomap); if (ret) return ret; @@ -401,8 +406,7 @@ out2: *old_mem = *new_mem; new_mem->mm_node = NULL; - if ((man->flags & TTM_MEMTYPE_FLAG_FIXED) && (ttm != NULL)) { - ttm_tt_unbind(ttm); + if (man->flags & TTM_MEMTYPE_FLAG_FIXED) { ttm_tt_destroy(ttm); bo->ttm = NULL; } @@ -462,6 +466,7 @@ static int ttm_buffer_object_transfer(struct ttm_buffer_object *bo, INIT_LIST_HEAD(&fbo->lru); INIT_LIST_HEAD(&fbo->swap); INIT_LIST_HEAD(&fbo->io_reserve_lru); + fbo->moving = NULL; drm_vma_node_reset(&fbo->vma_node); atomic_set(&fbo->cpu_writers, 0); @@ -634,7 +639,6 @@ EXPORT_SYMBOL(ttm_bo_kunmap); int ttm_bo_move_accel_cleanup(struct ttm_buffer_object *bo, struct fence *fence, bool evict, - bool no_wait_gpu, struct ttm_mem_reg *new_mem) { struct ttm_bo_device *bdev = bo->bdev; @@ -649,9 +653,7 @@ int ttm_bo_move_accel_cleanup(struct ttm_buffer_object *bo, if (ret) return ret; - if ((man->flags & TTM_MEMTYPE_FLAG_FIXED) && - (bo->ttm != NULL)) { - ttm_tt_unbind(bo->ttm); + if (man->flags & TTM_MEMTYPE_FLAG_FIXED) { ttm_tt_destroy(bo->ttm); bo->ttm = NULL; } @@ -665,7 +667,8 @@ int ttm_bo_move_accel_cleanup(struct ttm_buffer_object *bo, * operation has completed. */ - set_bit(TTM_BO_PRIV_FLAG_MOVING, &bo->priv_flags); + fence_put(bo->moving); + bo->moving = fence_get(fence); ret = ttm_buffer_object_transfer(bo, &ghost_obj); if (ret) @@ -694,3 +697,95 @@ int ttm_bo_move_accel_cleanup(struct ttm_buffer_object *bo, return 0; } EXPORT_SYMBOL(ttm_bo_move_accel_cleanup); + +int ttm_bo_pipeline_move(struct ttm_buffer_object *bo, + struct fence *fence, bool evict, + struct ttm_mem_reg *new_mem) +{ + struct ttm_bo_device *bdev = bo->bdev; + struct ttm_mem_reg *old_mem = &bo->mem; + + struct ttm_mem_type_manager *from = &bdev->man[old_mem->mem_type]; + struct ttm_mem_type_manager *to = &bdev->man[new_mem->mem_type]; + + int ret; + + reservation_object_add_excl_fence(bo->resv, fence); + + if (!evict) { + struct ttm_buffer_object *ghost_obj; + + /** + * This should help pipeline ordinary buffer moves. + * + * Hang old buffer memory on a new buffer object, + * and leave it to be released when the GPU + * operation has completed. + */ + + fence_put(bo->moving); + bo->moving = fence_get(fence); + + ret = ttm_buffer_object_transfer(bo, &ghost_obj); + if (ret) + return ret; + + reservation_object_add_excl_fence(ghost_obj->resv, fence); + + /** + * If we're not moving to fixed memory, the TTM object + * needs to stay alive. Otherwhise hang it on the ghost + * bo to be unbound and destroyed. + */ + + if (!(to->flags & TTM_MEMTYPE_FLAG_FIXED)) + ghost_obj->ttm = NULL; + else + bo->ttm = NULL; + + ttm_bo_unreserve(ghost_obj); + ttm_bo_unref(&ghost_obj); + + } else if (from->flags & TTM_MEMTYPE_FLAG_FIXED) { + + /** + * BO doesn't have a TTM we need to bind/unbind. Just remember + * this eviction and free up the allocation + */ + + spin_lock(&from->move_lock); + if (!from->move || fence_is_later(fence, from->move)) { + fence_put(from->move); + from->move = fence_get(fence); + } + spin_unlock(&from->move_lock); + + ttm_bo_free_old_node(bo); + + fence_put(bo->moving); + bo->moving = fence_get(fence); + + } else { + /** + * Last resort, wait for the move to be completed. + * + * Should never happen in pratice. + */ + + ret = ttm_bo_wait(bo, false, false); + if (ret) + return ret; + + if (to->flags & TTM_MEMTYPE_FLAG_FIXED) { + ttm_tt_destroy(bo->ttm); + bo->ttm = NULL; + } + ttm_bo_free_old_node(bo); + } + + *old_mem = *new_mem; + new_mem->mm_node = NULL; + + return 0; +} +EXPORT_SYMBOL(ttm_bo_pipeline_move); diff --git a/drivers/gpu/drm/ttm/ttm_bo_vm.c b/drivers/gpu/drm/ttm/ttm_bo_vm.c index 3216878bced3..a6ed9d5e5167 100644 --- a/drivers/gpu/drm/ttm/ttm_bo_vm.c +++ b/drivers/gpu/drm/ttm/ttm_bo_vm.c @@ -48,15 +48,14 @@ static int ttm_bo_vm_fault_idle(struct ttm_buffer_object *bo, { int ret = 0; - if (likely(!test_bit(TTM_BO_PRIV_FLAG_MOVING, &bo->priv_flags))) + if (likely(!bo->moving)) goto out_unlock; /* * Quick non-stalling check for idle. */ - ret = ttm_bo_wait(bo, false, true); - if (likely(ret == 0)) - goto out_unlock; + if (fence_is_signaled(bo->moving)) + goto out_clear; /* * If possible, avoid waiting for GPU with mmap_sem @@ -68,17 +67,23 @@ static int ttm_bo_vm_fault_idle(struct ttm_buffer_object *bo, goto out_unlock; up_read(&vma->vm_mm->mmap_sem); - (void) ttm_bo_wait(bo, true, false); + (void) fence_wait(bo->moving, true); goto out_unlock; } /* * Ordinary wait. */ - ret = ttm_bo_wait(bo, true, false); - if (unlikely(ret != 0)) + ret = fence_wait(bo->moving, true); + if (unlikely(ret != 0)) { ret = (ret != -ERESTARTSYS) ? VM_FAULT_SIGBUS : VM_FAULT_NOPAGE; + goto out_unlock; + } + +out_clear: + fence_put(bo->moving); + bo->moving = NULL; out_unlock: return ret; diff --git a/drivers/gpu/drm/ttm/ttm_tt.c b/drivers/gpu/drm/ttm/ttm_tt.c index 077ae9b2865d..526e5a78eea5 100644 --- a/drivers/gpu/drm/ttm/ttm_tt.c +++ b/drivers/gpu/drm/ttm/ttm_tt.c @@ -166,12 +166,10 @@ EXPORT_SYMBOL(ttm_tt_set_placement_caching); void ttm_tt_destroy(struct ttm_tt *ttm) { - if (unlikely(ttm == NULL)) + if (ttm == NULL) return; - if (ttm->state == tt_bound) { - ttm_tt_unbind(ttm); - } + ttm_tt_unbind(ttm); if (ttm->state == tt_unbound) ttm_tt_unpopulate(ttm); diff --git a/drivers/gpu/drm/udl/Kconfig b/drivers/gpu/drm/udl/Kconfig index 613ab0622d6e..1616ec4f4d84 100644 --- a/drivers/gpu/drm/udl/Kconfig +++ b/drivers/gpu/drm/udl/Kconfig @@ -4,12 +4,7 @@ config DRM_UDL depends on USB_SUPPORT depends on USB_ARCH_HAS_HCD select USB - select FB_SYS_FILLRECT - select FB_SYS_COPYAREA - select FB_SYS_IMAGEBLIT - select FB_DEFERRED_IO select DRM_KMS_HELPER - select DRM_KMS_FB_HELPER help This is a KMS driver for the USB displaylink video adapters. Say M/Y to add support for these devices via drm/kms interfaces. diff --git a/drivers/gpu/drm/udl/udl_drv.c b/drivers/gpu/drm/udl/udl_drv.c index c20408940cd0..17d34e0edbdd 100644 --- a/drivers/gpu/drm/udl/udl_drv.c +++ b/drivers/gpu/drm/udl/udl_drv.c @@ -94,7 +94,6 @@ static void udl_usb_disconnect(struct usb_interface *interface) struct drm_device *dev = usb_get_intfdata(interface); drm_kms_helper_poll_disable(dev); - drm_connector_unregister_all(dev); udl_fbdev_unplug(dev); udl_drop_usb(dev); drm_unplug_dev(dev); diff --git a/drivers/gpu/drm/vc4/vc4_crtc.c b/drivers/gpu/drm/vc4/vc4_crtc.c index 4c0f26a644a3..8fc2b731b59a 100644 --- a/drivers/gpu/drm/vc4/vc4_crtc.c +++ b/drivers/gpu/drm/vc4/vc4_crtc.c @@ -46,12 +46,17 @@ struct vc4_crtc { const struct vc4_crtc_data *data; void __iomem *regs; + /* Timestamp at start of vblank irq - unaffected by lock delays. */ + ktime_t t_vblank; + /* Which HVS channel we're using for our CRTC. */ int channel; u8 lut_r[256]; u8 lut_g[256]; u8 lut_b[256]; + /* Size in pixels of the COB memory allocated to this CRTC. */ + u32 cob_size; struct drm_pending_vblank_event *event; }; @@ -146,6 +151,144 @@ int vc4_crtc_debugfs_regs(struct seq_file *m, void *unused) } #endif +int vc4_crtc_get_scanoutpos(struct drm_device *dev, unsigned int crtc_id, + unsigned int flags, int *vpos, int *hpos, + ktime_t *stime, ktime_t *etime, + const struct drm_display_mode *mode) +{ + struct vc4_dev *vc4 = to_vc4_dev(dev); + struct vc4_crtc *vc4_crtc = vc4->crtc[crtc_id]; + u32 val; + int fifo_lines; + int vblank_lines; + int ret = 0; + + /* + * XXX Doesn't work well in interlaced mode yet, partially due + * to problems in vc4 kms or drm core interlaced mode handling, + * so disable for now in interlaced mode. + */ + if (mode->flags & DRM_MODE_FLAG_INTERLACE) + return ret; + + /* preempt_disable_rt() should go right here in PREEMPT_RT patchset. */ + + /* Get optional system timestamp before query. */ + if (stime) + *stime = ktime_get(); + + /* + * Read vertical scanline which is currently composed for our + * pixelvalve by the HVS, and also the scaler status. + */ + val = HVS_READ(SCALER_DISPSTATX(vc4_crtc->channel)); + + /* Get optional system timestamp after query. */ + if (etime) + *etime = ktime_get(); + + /* preempt_enable_rt() should go right here in PREEMPT_RT patchset. */ + + /* Vertical position of hvs composed scanline. */ + *vpos = VC4_GET_FIELD(val, SCALER_DISPSTATX_LINE); + + /* No hpos info available. */ + if (hpos) + *hpos = 0; + + /* This is the offset we need for translating hvs -> pv scanout pos. */ + fifo_lines = vc4_crtc->cob_size / mode->crtc_hdisplay; + + if (fifo_lines > 0) + ret |= DRM_SCANOUTPOS_VALID; + + /* HVS more than fifo_lines into frame for compositing? */ + if (*vpos > fifo_lines) { + /* + * We are in active scanout and can get some meaningful results + * from HVS. The actual PV scanout can not trail behind more + * than fifo_lines as that is the fifo's capacity. Assume that + * in active scanout the HVS and PV work in lockstep wrt. HVS + * refilling the fifo and PV consuming from the fifo, ie. + * whenever the PV consumes and frees up a scanline in the + * fifo, the HVS will immediately refill it, therefore + * incrementing vpos. Therefore we choose HVS read position - + * fifo size in scanlines as a estimate of the real scanout + * position of the PV. + */ + *vpos -= fifo_lines + 1; + if (mode->flags & DRM_MODE_FLAG_INTERLACE) + *vpos /= 2; + + ret |= DRM_SCANOUTPOS_ACCURATE; + return ret; + } + + /* + * Less: This happens when we are in vblank and the HVS, after getting + * the VSTART restart signal from the PV, just started refilling its + * fifo with new lines from the top-most lines of the new framebuffers. + * The PV does not scan out in vblank, so does not remove lines from + * the fifo, so the fifo will be full quickly and the HVS has to pause. + * We can't get meaningful readings wrt. scanline position of the PV + * and need to make things up in a approximative but consistent way. + */ + ret |= DRM_SCANOUTPOS_IN_VBLANK; + vblank_lines = mode->crtc_vtotal - mode->crtc_vdisplay; + + if (flags & DRM_CALLED_FROM_VBLIRQ) { + /* + * Assume the irq handler got called close to first + * line of vblank, so PV has about a full vblank + * scanlines to go, and as a base timestamp use the + * one taken at entry into vblank irq handler, so it + * is not affected by random delays due to lock + * contention on event_lock or vblank_time lock in + * the core. + */ + *vpos = -vblank_lines; + + if (stime) + *stime = vc4_crtc->t_vblank; + if (etime) + *etime = vc4_crtc->t_vblank; + + /* + * If the HVS fifo is not yet full then we know for certain + * we are at the very beginning of vblank, as the hvs just + * started refilling, and the stime and etime timestamps + * truly correspond to start of vblank. + */ + if ((val & SCALER_DISPSTATX_FULL) != SCALER_DISPSTATX_FULL) + ret |= DRM_SCANOUTPOS_ACCURATE; + } else { + /* + * No clue where we are inside vblank. Return a vpos of zero, + * which will cause calling code to just return the etime + * timestamp uncorrected. At least this is no worse than the + * standard fallback. + */ + *vpos = 0; + } + + return ret; +} + +int vc4_crtc_get_vblank_timestamp(struct drm_device *dev, unsigned int crtc_id, + int *max_error, struct timeval *vblank_time, + unsigned flags) +{ + struct vc4_dev *vc4 = to_vc4_dev(dev); + struct vc4_crtc *vc4_crtc = vc4->crtc[crtc_id]; + struct drm_crtc *crtc = &vc4_crtc->base; + struct drm_crtc_state *state = crtc->state; + + /* Helper routine in DRM core does all the work: */ + return drm_calc_vbltimestamp_from_scanoutpos(dev, crtc_id, max_error, + vblank_time, flags, + &state->adjusted_mode); +} + static void vc4_crtc_destroy(struct drm_crtc *crtc) { drm_crtc_cleanup(crtc); @@ -449,14 +592,6 @@ static void vc4_crtc_atomic_flush(struct drm_crtc *crtc, WARN_ON_ONCE(dlist_next - dlist_start != vc4_state->mm.size); - HVS_WRITE(SCALER_DISPLISTX(vc4_crtc->channel), - vc4_state->mm.start); - - if (debug_dump_regs) { - DRM_INFO("CRTC %d HVS after:\n", drm_crtc_index(crtc)); - vc4_hvs_dump_state(dev); - } - if (crtc->state->event) { unsigned long flags; @@ -466,8 +601,20 @@ static void vc4_crtc_atomic_flush(struct drm_crtc *crtc, spin_lock_irqsave(&dev->event_lock, flags); vc4_crtc->event = crtc->state->event; - spin_unlock_irqrestore(&dev->event_lock, flags); crtc->state->event = NULL; + + HVS_WRITE(SCALER_DISPLISTX(vc4_crtc->channel), + vc4_state->mm.start); + + spin_unlock_irqrestore(&dev->event_lock, flags); + } else { + HVS_WRITE(SCALER_DISPLISTX(vc4_crtc->channel), + vc4_state->mm.start); + } + + if (debug_dump_regs) { + DRM_INFO("CRTC %d HVS after:\n", drm_crtc_index(crtc)); + vc4_hvs_dump_state(dev); } } @@ -493,12 +640,17 @@ static void vc4_crtc_handle_page_flip(struct vc4_crtc *vc4_crtc) { struct drm_crtc *crtc = &vc4_crtc->base; struct drm_device *dev = crtc->dev; + struct vc4_dev *vc4 = to_vc4_dev(dev); + struct vc4_crtc_state *vc4_state = to_vc4_crtc_state(crtc->state); + u32 chan = vc4_crtc->channel; unsigned long flags; spin_lock_irqsave(&dev->event_lock, flags); - if (vc4_crtc->event) { + if (vc4_crtc->event && + (vc4_state->mm.start == HVS_READ(SCALER_DISPLACTX(chan)))) { drm_crtc_send_vblank_event(crtc, vc4_crtc->event); vc4_crtc->event = NULL; + drm_crtc_vblank_put(crtc); } spin_unlock_irqrestore(&dev->event_lock, flags); } @@ -510,6 +662,7 @@ static irqreturn_t vc4_crtc_irq_handler(int irq, void *data) irqreturn_t ret = IRQ_NONE; if (stat & PV_INT_VFP_START) { + vc4_crtc->t_vblank = ktime_get(); CRTC_WRITE(PV_INTSTAT, PV_INT_VFP_START); drm_crtc_handle_vblank(&vc4_crtc->base); vc4_crtc_handle_page_flip(vc4_crtc); @@ -549,6 +702,7 @@ vc4_async_page_flip_complete(struct vc4_seqno_cb *cb) spin_unlock_irqrestore(&dev->event_lock, flags); } + drm_crtc_vblank_put(crtc); drm_framebuffer_unreference(flip_state->fb); kfree(flip_state); @@ -591,6 +745,8 @@ static int vc4_async_page_flip(struct drm_crtc *crtc, return ret; } + WARN_ON(drm_crtc_vblank_get(crtc) != 0); + /* Immediately update the plane's legacy fb pointer, so that later * modeset prep sees the state that will be present when the semaphore * is released. @@ -711,6 +867,22 @@ static void vc4_set_crtc_possible_masks(struct drm_device *drm, } } +static void +vc4_crtc_get_cob_allocation(struct vc4_crtc *vc4_crtc) +{ + struct drm_device *drm = vc4_crtc->base.dev; + struct vc4_dev *vc4 = to_vc4_dev(drm); + u32 dispbase = HVS_READ(SCALER_DISPBASEX(vc4_crtc->channel)); + /* Top/base are supposed to be 4-pixel aligned, but the + * Raspberry Pi firmware fills the low bits (which are + * presumably ignored). + */ + u32 top = VC4_GET_FIELD(dispbase, SCALER_DISPBASEX_TOP) & ~3; + u32 base = VC4_GET_FIELD(dispbase, SCALER_DISPBASEX_BASE) & ~3; + + vc4_crtc->cob_size = top - base + 4; +} + static int vc4_crtc_bind(struct device *dev, struct device *master, void *data) { struct platform_device *pdev = to_platform_device(dev); @@ -787,6 +959,8 @@ static int vc4_crtc_bind(struct device *dev, struct device *master, void *data) crtc->cursor = cursor_plane; } + vc4_crtc_get_cob_allocation(vc4_crtc); + CRTC_WRITE(PV_INTEN, 0); CRTC_WRITE(PV_INTSTAT, PV_INT_VFP_START); ret = devm_request_irq(dev, platform_get_irq(pdev, 0), diff --git a/drivers/gpu/drm/vc4/vc4_dpi.c b/drivers/gpu/drm/vc4/vc4_dpi.c index dba1114297e4..275fedbdbd9e 100644 --- a/drivers/gpu/drm/vc4/vc4_dpi.c +++ b/drivers/gpu/drm/vc4/vc4_dpi.c @@ -227,14 +227,12 @@ static struct drm_connector *vc4_dpi_connector_init(struct drm_device *dev, { struct drm_connector *connector = NULL; struct vc4_dpi_connector *dpi_connector; - int ret = 0; dpi_connector = devm_kzalloc(dev->dev, sizeof(*dpi_connector), GFP_KERNEL); - if (!dpi_connector) { - ret = -ENOMEM; - goto fail; - } + if (!dpi_connector) + return ERR_PTR(-ENOMEM); + connector = &dpi_connector->base; dpi_connector->encoder = dpi->encoder; @@ -251,12 +249,6 @@ static struct drm_connector *vc4_dpi_connector_init(struct drm_device *dev, drm_mode_connector_attach_encoder(connector, dpi->encoder); return connector; - - fail: - if (connector) - vc4_dpi_connector_destroy(connector); - - return ERR_PTR(ret); } static const struct drm_encoder_funcs vc4_dpi_encoder_funcs = { diff --git a/drivers/gpu/drm/vc4/vc4_drv.c b/drivers/gpu/drm/vc4/vc4_drv.c index 9e88231b8906..8b42d31a7f0e 100644 --- a/drivers/gpu/drm/vc4/vc4_drv.c +++ b/drivers/gpu/drm/vc4/vc4_drv.c @@ -14,6 +14,7 @@ #include <linux/module.h> #include <linux/of_platform.h> #include <linux/platform_device.h> +#include <linux/pm_runtime.h> #include "drm_fb_cma_helper.h" #include "uapi/drm/vc4_drm.h" @@ -43,12 +44,54 @@ void __iomem *vc4_ioremap_regs(struct platform_device *dev, int index) return map; } +static int vc4_get_param_ioctl(struct drm_device *dev, void *data, + struct drm_file *file_priv) +{ + struct vc4_dev *vc4 = to_vc4_dev(dev); + struct drm_vc4_get_param *args = data; + int ret; + + if (args->pad != 0) + return -EINVAL; + + switch (args->param) { + case DRM_VC4_PARAM_V3D_IDENT0: + ret = pm_runtime_get_sync(&vc4->v3d->pdev->dev); + if (ret) + return ret; + args->value = V3D_READ(V3D_IDENT0); + pm_runtime_put(&vc4->v3d->pdev->dev); + break; + case DRM_VC4_PARAM_V3D_IDENT1: + ret = pm_runtime_get_sync(&vc4->v3d->pdev->dev); + if (ret) + return ret; + args->value = V3D_READ(V3D_IDENT1); + pm_runtime_put(&vc4->v3d->pdev->dev); + break; + case DRM_VC4_PARAM_V3D_IDENT2: + ret = pm_runtime_get_sync(&vc4->v3d->pdev->dev); + if (ret) + return ret; + args->value = V3D_READ(V3D_IDENT2); + pm_runtime_put(&vc4->v3d->pdev->dev); + break; + case DRM_VC4_PARAM_SUPPORTS_BRANCHES: + args->value = true; + break; + default: + DRM_DEBUG("Unknown parameter %d\n", args->param); + return -EINVAL; + } + + return 0; +} + static void vc4_lastclose(struct drm_device *dev) { struct vc4_dev *vc4 = to_vc4_dev(dev); - if (vc4->fbdev) - drm_fbdev_cma_restore_mode(vc4->fbdev); + drm_fbdev_cma_restore_mode(vc4->fbdev); } static const struct file_operations vc4_drm_fops = { @@ -66,14 +109,15 @@ static const struct file_operations vc4_drm_fops = { }; static const struct drm_ioctl_desc vc4_drm_ioctls[] = { - DRM_IOCTL_DEF_DRV(VC4_SUBMIT_CL, vc4_submit_cl_ioctl, 0), - DRM_IOCTL_DEF_DRV(VC4_WAIT_SEQNO, vc4_wait_seqno_ioctl, 0), - DRM_IOCTL_DEF_DRV(VC4_WAIT_BO, vc4_wait_bo_ioctl, 0), - DRM_IOCTL_DEF_DRV(VC4_CREATE_BO, vc4_create_bo_ioctl, 0), - DRM_IOCTL_DEF_DRV(VC4_MMAP_BO, vc4_mmap_bo_ioctl, 0), - DRM_IOCTL_DEF_DRV(VC4_CREATE_SHADER_BO, vc4_create_shader_bo_ioctl, 0), + DRM_IOCTL_DEF_DRV(VC4_SUBMIT_CL, vc4_submit_cl_ioctl, DRM_RENDER_ALLOW), + DRM_IOCTL_DEF_DRV(VC4_WAIT_SEQNO, vc4_wait_seqno_ioctl, DRM_RENDER_ALLOW), + DRM_IOCTL_DEF_DRV(VC4_WAIT_BO, vc4_wait_bo_ioctl, DRM_RENDER_ALLOW), + DRM_IOCTL_DEF_DRV(VC4_CREATE_BO, vc4_create_bo_ioctl, DRM_RENDER_ALLOW), + DRM_IOCTL_DEF_DRV(VC4_MMAP_BO, vc4_mmap_bo_ioctl, DRM_RENDER_ALLOW), + DRM_IOCTL_DEF_DRV(VC4_CREATE_SHADER_BO, vc4_create_shader_bo_ioctl, DRM_RENDER_ALLOW), DRM_IOCTL_DEF_DRV(VC4_GET_HANG_STATE, vc4_get_hang_state_ioctl, DRM_ROOT_ONLY), + DRM_IOCTL_DEF_DRV(VC4_GET_PARAM, vc4_get_param_ioctl, DRM_RENDER_ALLOW), }; static struct drm_driver vc4_drm_driver = { @@ -91,7 +135,9 @@ static struct drm_driver vc4_drm_driver = { .enable_vblank = vc4_enable_vblank, .disable_vblank = vc4_disable_vblank, - .get_vblank_counter = drm_vblank_count, + .get_vblank_counter = drm_vblank_no_hw_counter, + .get_scanout_position = vc4_crtc_get_scanoutpos, + .get_vblank_timestamp = vc4_crtc_get_vblank_timestamp, #if defined(CONFIG_DEBUG_FS) .debugfs_init = vc4_debugfs_init, @@ -195,8 +241,6 @@ static int vc4_drm_bind(struct device *dev) vc4_bo_cache_init(drm); drm_mode_config_init(drm); - if (ret) - goto unref; vc4_gem_init(drm); @@ -218,7 +262,6 @@ unbind_all: component_unbind_all(dev, drm); gem_destroy: vc4_gem_destroy(drm); -unref: drm_dev_unref(drm); vc4_bo_cache_destroy(drm); return ret; @@ -246,8 +289,8 @@ static const struct component_master_ops vc4_drm_ops = { static struct platform_driver *const component_drivers[] = { &vc4_hdmi_driver, &vc4_dpi_driver, - &vc4_crtc_driver, &vc4_hvs_driver, + &vc4_crtc_driver, &vc4_v3d_driver, }; diff --git a/drivers/gpu/drm/vc4/vc4_drv.h b/drivers/gpu/drm/vc4/vc4_drv.h index c799baabc008..489e3de0c050 100644 --- a/drivers/gpu/drm/vc4/vc4_drv.h +++ b/drivers/gpu/drm/vc4/vc4_drv.h @@ -355,6 +355,9 @@ struct vc4_validated_shader_info { uint32_t uniforms_src_size; uint32_t num_texture_samples; struct vc4_texture_sample_info *texture_samples; + + uint32_t num_uniform_addr_offsets; + uint32_t *uniform_addr_offsets; }; /** @@ -415,6 +418,13 @@ extern struct platform_driver vc4_crtc_driver; int vc4_enable_vblank(struct drm_device *dev, unsigned int crtc_id); void vc4_disable_vblank(struct drm_device *dev, unsigned int crtc_id); int vc4_crtc_debugfs_regs(struct seq_file *m, void *arg); +int vc4_crtc_get_scanoutpos(struct drm_device *dev, unsigned int crtc_id, + unsigned int flags, int *vpos, int *hpos, + ktime_t *stime, ktime_t *etime, + const struct drm_display_mode *mode); +int vc4_crtc_get_vblank_timestamp(struct drm_device *dev, unsigned int crtc_id, + int *max_error, struct timeval *vblank_time, + unsigned flags); /* vc4_debugfs.c */ int vc4_debugfs_init(struct drm_minor *minor); diff --git a/drivers/gpu/drm/vc4/vc4_hdmi.c b/drivers/gpu/drm/vc4/vc4_hdmi.c index 68df91c3f860..4452f3631cac 100644 --- a/drivers/gpu/drm/vc4/vc4_hdmi.c +++ b/drivers/gpu/drm/vc4/vc4_hdmi.c @@ -456,12 +456,6 @@ static int vc4_hdmi_bind(struct device *dev, struct device *master, void *data) if (IS_ERR(hdmi->hd_regs)) return PTR_ERR(hdmi->hd_regs); - ddc_node = of_parse_phandle(dev->of_node, "ddc", 0); - if (!ddc_node) { - DRM_ERROR("Failed to find ddc node in device tree\n"); - return -ENODEV; - } - hdmi->pixel_clock = devm_clk_get(dev, "pixel"); if (IS_ERR(hdmi->pixel_clock)) { DRM_ERROR("Failed to get pixel clock\n"); @@ -473,7 +467,14 @@ static int vc4_hdmi_bind(struct device *dev, struct device *master, void *data) return PTR_ERR(hdmi->hsm_clock); } + ddc_node = of_parse_phandle(dev->of_node, "ddc", 0); + if (!ddc_node) { + DRM_ERROR("Failed to find ddc node in device tree\n"); + return -ENODEV; + } + hdmi->ddc = of_find_i2c_adapter_by_node(ddc_node); + of_node_put(ddc_node); if (!hdmi->ddc) { DRM_DEBUG("Failed to get ddc i2c adapter by node\n"); return -EPROBE_DEFER; diff --git a/drivers/gpu/drm/vc4/vc4_kms.c b/drivers/gpu/drm/vc4/vc4_kms.c index 8f4d5ffc32be..4ac894d993cd 100644 --- a/drivers/gpu/drm/vc4/vc4_kms.c +++ b/drivers/gpu/drm/vc4/vc4_kms.c @@ -26,8 +26,7 @@ static void vc4_output_poll_changed(struct drm_device *dev) { struct vc4_dev *vc4 = to_vc4_dev(dev); - if (vc4->fbdev) - drm_fbdev_cma_hotplug_event(vc4->fbdev); + drm_fbdev_cma_hotplug_event(vc4->fbdev); } struct vc4_commit { @@ -119,10 +118,18 @@ static int vc4_atomic_commit(struct drm_device *dev, return -ENOMEM; /* Make sure that any outstanding modesets have finished. */ - ret = down_interruptible(&vc4->async_modeset); - if (ret) { - kfree(c); - return ret; + if (nonblock) { + ret = down_trylock(&vc4->async_modeset); + if (ret) { + kfree(c); + return -EBUSY; + } + } else { + ret = down_interruptible(&vc4->async_modeset); + if (ret) { + kfree(c); + return ret; + } } ret = drm_atomic_helper_prepare_planes(dev, state); diff --git a/drivers/gpu/drm/vc4/vc4_plane.c b/drivers/gpu/drm/vc4/vc4_plane.c index 5d2c3d9fd17a..29e4b400e25e 100644 --- a/drivers/gpu/drm/vc4/vc4_plane.c +++ b/drivers/gpu/drm/vc4/vc4_plane.c @@ -94,6 +94,14 @@ static const struct hvs_format { .pixel_order = HVS_PIXEL_ORDER_ABGR, .has_alpha = true, }, { + .drm = DRM_FORMAT_ABGR8888, .hvs = HVS_PIXEL_FORMAT_RGBA8888, + .pixel_order = HVS_PIXEL_ORDER_ARGB, .has_alpha = true, + }, + { + .drm = DRM_FORMAT_XBGR8888, .hvs = HVS_PIXEL_FORMAT_RGBA8888, + .pixel_order = HVS_PIXEL_ORDER_ARGB, .has_alpha = false, + }, + { .drm = DRM_FORMAT_RGB565, .hvs = HVS_PIXEL_FORMAT_RGB565, .pixel_order = HVS_PIXEL_ORDER_XRGB, .has_alpha = false, }, diff --git a/drivers/gpu/drm/vc4/vc4_qpu_defines.h b/drivers/gpu/drm/vc4/vc4_qpu_defines.h index d5c2f3c85ebb..f4e795a0d3f6 100644 --- a/drivers/gpu/drm/vc4/vc4_qpu_defines.h +++ b/drivers/gpu/drm/vc4/vc4_qpu_defines.h @@ -70,7 +70,7 @@ enum qpu_raddr { QPU_R_ELEM_QPU = 38, QPU_R_NOP, QPU_R_XY_PIXEL_COORD = 41, - QPU_R_MS_REV_FLAGS = 41, + QPU_R_MS_REV_FLAGS = 42, QPU_R_VPM = 48, QPU_R_VPM_LD_BUSY, QPU_R_VPM_LD_WAIT, @@ -230,6 +230,15 @@ enum qpu_unpack_r4 { #define QPU_COND_MUL_SHIFT 46 #define QPU_COND_MUL_MASK QPU_MASK(48, 46) +#define QPU_BRANCH_COND_SHIFT 52 +#define QPU_BRANCH_COND_MASK QPU_MASK(55, 52) + +#define QPU_BRANCH_REL ((uint64_t)1 << 51) +#define QPU_BRANCH_REG ((uint64_t)1 << 50) + +#define QPU_BRANCH_RADDR_A_SHIFT 45 +#define QPU_BRANCH_RADDR_A_MASK QPU_MASK(49, 45) + #define QPU_SF ((uint64_t)1 << 45) #define QPU_WADDR_ADD_SHIFT 38 @@ -261,4 +270,10 @@ enum qpu_unpack_r4 { #define QPU_OP_ADD_SHIFT 24 #define QPU_OP_ADD_MASK QPU_MASK(28, 24) +#define QPU_LOAD_IMM_SHIFT 0 +#define QPU_LOAD_IMM_MASK QPU_MASK(31, 0) + +#define QPU_BRANCH_TARGET_SHIFT 0 +#define QPU_BRANCH_TARGET_MASK QPU_MASK(31, 0) + #endif /* VC4_QPU_DEFINES_H */ diff --git a/drivers/gpu/drm/vc4/vc4_regs.h b/drivers/gpu/drm/vc4/vc4_regs.h index 6163b95c5411..160942a9180e 100644 --- a/drivers/gpu/drm/vc4/vc4_regs.h +++ b/drivers/gpu/drm/vc4/vc4_regs.h @@ -341,6 +341,10 @@ #define SCALER_DISPLACT0 0x00000030 #define SCALER_DISPLACT1 0x00000034 #define SCALER_DISPLACT2 0x00000038 +#define SCALER_DISPLACTX(x) (SCALER_DISPLACT0 + \ + (x) * (SCALER_DISPLACT1 - \ + SCALER_DISPLACT0)) + #define SCALER_DISPCTRL0 0x00000040 # define SCALER_DISPCTRLX_ENABLE BIT(31) # define SCALER_DISPCTRLX_RESET BIT(30) @@ -362,7 +366,6 @@ # define SCALER_DISPBKGND_FILL BIT(24) #define SCALER_DISPSTAT0 0x00000048 -#define SCALER_DISPBASE0 0x0000004c # define SCALER_DISPSTATX_MODE_MASK VC4_MASK(31, 30) # define SCALER_DISPSTATX_MODE_SHIFT 30 # define SCALER_DISPSTATX_MODE_DISABLED 0 @@ -371,6 +374,24 @@ # define SCALER_DISPSTATX_MODE_EOF 3 # define SCALER_DISPSTATX_FULL BIT(29) # define SCALER_DISPSTATX_EMPTY BIT(28) +# define SCALER_DISPSTATX_FRAME_COUNT_MASK VC4_MASK(17, 12) +# define SCALER_DISPSTATX_FRAME_COUNT_SHIFT 12 +# define SCALER_DISPSTATX_LINE_MASK VC4_MASK(11, 0) +# define SCALER_DISPSTATX_LINE_SHIFT 0 + +#define SCALER_DISPBASE0 0x0000004c +/* Last pixel in the COB (display FIFO memory) allocated to this HVS + * channel. Must be 4-pixel aligned (and thus 4 pixels less than the + * next COB base). + */ +# define SCALER_DISPBASEX_TOP_MASK VC4_MASK(31, 16) +# define SCALER_DISPBASEX_TOP_SHIFT 16 +/* First pixel in the COB (display FIFO memory) allocated to this HVS + * channel. Must be 4-pixel aligned. + */ +# define SCALER_DISPBASEX_BASE_MASK VC4_MASK(15, 0) +# define SCALER_DISPBASEX_BASE_SHIFT 0 + #define SCALER_DISPCTRL1 0x00000050 #define SCALER_DISPBKGND1 0x00000054 #define SCALER_DISPBKGNDX(x) (SCALER_DISPBKGND0 + \ @@ -381,6 +402,9 @@ (x) * (SCALER_DISPSTAT1 - \ SCALER_DISPSTAT0)) #define SCALER_DISPBASE1 0x0000005c +#define SCALER_DISPBASEX(x) (SCALER_DISPBASE0 + \ + (x) * (SCALER_DISPBASE1 - \ + SCALER_DISPBASE0)) #define SCALER_DISPCTRL2 0x00000060 #define SCALER_DISPCTRLX(x) (SCALER_DISPCTRL0 + \ (x) * (SCALER_DISPCTRL1 - \ diff --git a/drivers/gpu/drm/vc4/vc4_validate.c b/drivers/gpu/drm/vc4/vc4_validate.c index 24c2c746e8f3..9ce1d0adf882 100644 --- a/drivers/gpu/drm/vc4/vc4_validate.c +++ b/drivers/gpu/drm/vc4/vc4_validate.c @@ -802,7 +802,7 @@ validate_gl_shader_rec(struct drm_device *dev, uint32_t src_offset = *(uint32_t *)(pkt_u + o); uint32_t *texture_handles_u; void *uniform_data_u; - uint32_t tex; + uint32_t tex, uni; *(uint32_t *)(pkt_v + o) = bo[i]->paddr + src_offset; @@ -840,6 +840,17 @@ validate_gl_shader_rec(struct drm_device *dev, } } + /* Fill in the uniform slots that need this shader's + * start-of-uniforms address (used for resetting the uniform + * stream in the presence of control flow). + */ + for (uni = 0; + uni < validated_shader->num_uniform_addr_offsets; + uni++) { + uint32_t o = validated_shader->uniform_addr_offsets[uni]; + ((uint32_t *)exec->uniforms_v)[o] = exec->uniforms_p; + } + *(uint32_t *)(pkt_v + o + 4) = exec->uniforms_p; exec->uniforms_u += validated_shader->uniforms_src_size; diff --git a/drivers/gpu/drm/vc4/vc4_validate_shaders.c b/drivers/gpu/drm/vc4/vc4_validate_shaders.c index f67124b4c534..46527e989ce3 100644 --- a/drivers/gpu/drm/vc4/vc4_validate_shaders.c +++ b/drivers/gpu/drm/vc4/vc4_validate_shaders.c @@ -39,7 +39,17 @@ #include "vc4_drv.h" #include "vc4_qpu_defines.h" +#define LIVE_REG_COUNT (32 + 32 + 4) + struct vc4_shader_validation_state { + /* Current IP being validated. */ + uint32_t ip; + + /* IP at the end of the BO, do not read shader[max_ip] */ + uint32_t max_ip; + + uint64_t *shader; + struct vc4_texture_sample_info tmu_setup[2]; int tmu_write_count[2]; @@ -49,8 +59,30 @@ struct vc4_shader_validation_state { * * This is used for the validation of direct address memory reads. */ - uint32_t live_min_clamp_offsets[32 + 32 + 4]; - bool live_max_clamp_regs[32 + 32 + 4]; + uint32_t live_min_clamp_offsets[LIVE_REG_COUNT]; + bool live_max_clamp_regs[LIVE_REG_COUNT]; + uint32_t live_immediates[LIVE_REG_COUNT]; + + /* Bitfield of which IPs are used as branch targets. + * + * Used for validation that the uniform stream is updated at the right + * points and clearing the texturing/clamping state. + */ + unsigned long *branch_targets; + + /* Set when entering a basic block, and cleared when the uniform + * address update is found. This is used to make sure that we don't + * read uniforms when the address is undefined. + */ + bool needs_uniform_address_update; + + /* Set when we find a backwards branch. If the branch is backwards, + * the taraget is probably doing an address reset to read uniforms, + * and so we need to be sure that a uniforms address is present in the + * stream, even if the shader didn't need to read uniforms in later + * basic blocks. + */ + bool needs_uniform_address_for_loop; }; static uint32_t @@ -129,11 +161,11 @@ record_texture_sample(struct vc4_validated_shader_info *validated_shader, } static bool -check_tmu_write(uint64_t inst, - struct vc4_validated_shader_info *validated_shader, +check_tmu_write(struct vc4_validated_shader_info *validated_shader, struct vc4_shader_validation_state *validation_state, bool is_mul) { + uint64_t inst = validation_state->shader[validation_state->ip]; uint32_t waddr = (is_mul ? QPU_GET_FIELD(inst, QPU_WADDR_MUL) : QPU_GET_FIELD(inst, QPU_WADDR_ADD)); @@ -162,7 +194,7 @@ check_tmu_write(uint64_t inst, return false; } - /* We assert that the the clamped address is the first + /* We assert that the clamped address is the first * argument, and the UBO base address is the second argument. * This is arbitrary, but simpler than supporting flipping the * two either way. @@ -212,8 +244,14 @@ check_tmu_write(uint64_t inst, /* Since direct uses a RADDR uniform reference, it will get counted in * check_instruction_reads() */ - if (!is_direct) + if (!is_direct) { + if (validation_state->needs_uniform_address_update) { + DRM_ERROR("Texturing with undefined uniform address\n"); + return false; + } + validated_shader->uniforms_size += 4; + } if (submit) { if (!record_texture_sample(validated_shader, @@ -227,23 +265,138 @@ check_tmu_write(uint64_t inst, return true; } +static bool require_uniform_address_uniform(struct vc4_validated_shader_info *validated_shader) +{ + uint32_t o = validated_shader->num_uniform_addr_offsets; + uint32_t num_uniforms = validated_shader->uniforms_size / 4; + + validated_shader->uniform_addr_offsets = + krealloc(validated_shader->uniform_addr_offsets, + (o + 1) * + sizeof(*validated_shader->uniform_addr_offsets), + GFP_KERNEL); + if (!validated_shader->uniform_addr_offsets) + return false; + + validated_shader->uniform_addr_offsets[o] = num_uniforms; + validated_shader->num_uniform_addr_offsets++; + + return true; +} + static bool -check_reg_write(uint64_t inst, - struct vc4_validated_shader_info *validated_shader, +validate_uniform_address_write(struct vc4_validated_shader_info *validated_shader, + struct vc4_shader_validation_state *validation_state, + bool is_mul) +{ + uint64_t inst = validation_state->shader[validation_state->ip]; + u32 add_b = QPU_GET_FIELD(inst, QPU_ADD_B); + u32 raddr_a = QPU_GET_FIELD(inst, QPU_RADDR_A); + u32 raddr_b = QPU_GET_FIELD(inst, QPU_RADDR_B); + u32 add_lri = raddr_add_a_to_live_reg_index(inst); + /* We want our reset to be pointing at whatever uniform follows the + * uniforms base address. + */ + u32 expected_offset = validated_shader->uniforms_size + 4; + + /* We only support absolute uniform address changes, and we + * require that they be in the current basic block before any + * of its uniform reads. + * + * One could potentially emit more efficient QPU code, by + * noticing that (say) an if statement does uniform control + * flow for all threads and that the if reads the same number + * of uniforms on each side. However, this scheme is easy to + * validate so it's all we allow for now. + */ + + if (QPU_GET_FIELD(inst, QPU_SIG) != QPU_SIG_NONE) { + DRM_ERROR("uniforms address change must be " + "normal math\n"); + return false; + } + + if (is_mul || QPU_GET_FIELD(inst, QPU_OP_ADD) != QPU_A_ADD) { + DRM_ERROR("Uniform address reset must be an ADD.\n"); + return false; + } + + if (QPU_GET_FIELD(inst, QPU_COND_ADD) != QPU_COND_ALWAYS) { + DRM_ERROR("Uniform address reset must be unconditional.\n"); + return false; + } + + if (QPU_GET_FIELD(inst, QPU_PACK) != QPU_PACK_A_NOP && + !(inst & QPU_PM)) { + DRM_ERROR("No packing allowed on uniforms reset\n"); + return false; + } + + if (add_lri == -1) { + DRM_ERROR("First argument of uniform address write must be " + "an immediate value.\n"); + return false; + } + + if (validation_state->live_immediates[add_lri] != expected_offset) { + DRM_ERROR("Resetting uniforms with offset %db instead of %db\n", + validation_state->live_immediates[add_lri], + expected_offset); + return false; + } + + if (!(add_b == QPU_MUX_A && raddr_a == QPU_R_UNIF) && + !(add_b == QPU_MUX_B && raddr_b == QPU_R_UNIF)) { + DRM_ERROR("Second argument of uniform address write must be " + "a uniform.\n"); + return false; + } + + validation_state->needs_uniform_address_update = false; + validation_state->needs_uniform_address_for_loop = false; + return require_uniform_address_uniform(validated_shader); +} + +static bool +check_reg_write(struct vc4_validated_shader_info *validated_shader, struct vc4_shader_validation_state *validation_state, bool is_mul) { + uint64_t inst = validation_state->shader[validation_state->ip]; uint32_t waddr = (is_mul ? QPU_GET_FIELD(inst, QPU_WADDR_MUL) : QPU_GET_FIELD(inst, QPU_WADDR_ADD)); + uint32_t sig = QPU_GET_FIELD(inst, QPU_SIG); + bool ws = inst & QPU_WS; + bool is_b = is_mul ^ ws; + u32 lri = waddr_to_live_reg_index(waddr, is_b); + + if (lri != -1) { + uint32_t cond_add = QPU_GET_FIELD(inst, QPU_COND_ADD); + uint32_t cond_mul = QPU_GET_FIELD(inst, QPU_COND_MUL); + + if (sig == QPU_SIG_LOAD_IMM && + QPU_GET_FIELD(inst, QPU_PACK) == QPU_PACK_A_NOP && + ((is_mul && cond_mul == QPU_COND_ALWAYS) || + (!is_mul && cond_add == QPU_COND_ALWAYS))) { + validation_state->live_immediates[lri] = + QPU_GET_FIELD(inst, QPU_LOAD_IMM); + } else { + validation_state->live_immediates[lri] = ~0; + } + } switch (waddr) { case QPU_W_UNIFORMS_ADDRESS: - /* XXX: We'll probably need to support this for reladdr, but - * it's definitely a security-related one. - */ - DRM_ERROR("uniforms address load unsupported\n"); - return false; + if (is_b) { + DRM_ERROR("relative uniforms address change " + "unsupported\n"); + return false; + } + + return validate_uniform_address_write(validated_shader, + validation_state, + is_mul); case QPU_W_TLB_COLOR_MS: case QPU_W_TLB_COLOR_ALL: @@ -261,7 +414,7 @@ check_reg_write(uint64_t inst, case QPU_W_TMU1_T: case QPU_W_TMU1_R: case QPU_W_TMU1_B: - return check_tmu_write(inst, validated_shader, validation_state, + return check_tmu_write(validated_shader, validation_state, is_mul); case QPU_W_HOST_INT: @@ -294,10 +447,10 @@ check_reg_write(uint64_t inst, } static void -track_live_clamps(uint64_t inst, - struct vc4_validated_shader_info *validated_shader, +track_live_clamps(struct vc4_validated_shader_info *validated_shader, struct vc4_shader_validation_state *validation_state) { + uint64_t inst = validation_state->shader[validation_state->ip]; uint32_t op_add = QPU_GET_FIELD(inst, QPU_OP_ADD); uint32_t waddr_add = QPU_GET_FIELD(inst, QPU_WADDR_ADD); uint32_t waddr_mul = QPU_GET_FIELD(inst, QPU_WADDR_MUL); @@ -369,10 +522,10 @@ track_live_clamps(uint64_t inst, } static bool -check_instruction_writes(uint64_t inst, - struct vc4_validated_shader_info *validated_shader, +check_instruction_writes(struct vc4_validated_shader_info *validated_shader, struct vc4_shader_validation_state *validation_state) { + uint64_t inst = validation_state->shader[validation_state->ip]; uint32_t waddr_add = QPU_GET_FIELD(inst, QPU_WADDR_ADD); uint32_t waddr_mul = QPU_GET_FIELD(inst, QPU_WADDR_MUL); bool ok; @@ -382,20 +535,44 @@ check_instruction_writes(uint64_t inst, return false; } - ok = (check_reg_write(inst, validated_shader, validation_state, - false) && - check_reg_write(inst, validated_shader, validation_state, - true)); + ok = (check_reg_write(validated_shader, validation_state, false) && + check_reg_write(validated_shader, validation_state, true)); - track_live_clamps(inst, validated_shader, validation_state); + track_live_clamps(validated_shader, validation_state); return ok; } static bool -check_instruction_reads(uint64_t inst, - struct vc4_validated_shader_info *validated_shader) +check_branch(uint64_t inst, + struct vc4_validated_shader_info *validated_shader, + struct vc4_shader_validation_state *validation_state, + int ip) +{ + int32_t branch_imm = QPU_GET_FIELD(inst, QPU_BRANCH_TARGET); + uint32_t waddr_add = QPU_GET_FIELD(inst, QPU_WADDR_ADD); + uint32_t waddr_mul = QPU_GET_FIELD(inst, QPU_WADDR_MUL); + + if ((int)branch_imm < 0) + validation_state->needs_uniform_address_for_loop = true; + + /* We don't want to have to worry about validation of this, and + * there's no need for it. + */ + if (waddr_add != QPU_W_NOP || waddr_mul != QPU_W_NOP) { + DRM_ERROR("branch instruction at %d wrote a register.\n", + validation_state->ip); + return false; + } + + return true; +} + +static bool +check_instruction_reads(struct vc4_validated_shader_info *validated_shader, + struct vc4_shader_validation_state *validation_state) { + uint64_t inst = validation_state->shader[validation_state->ip]; uint32_t raddr_a = QPU_GET_FIELD(inst, QPU_RADDR_A); uint32_t raddr_b = QPU_GET_FIELD(inst, QPU_RADDR_B); uint32_t sig = QPU_GET_FIELD(inst, QPU_SIG); @@ -407,40 +584,204 @@ check_instruction_reads(uint64_t inst, * already be OOM. */ validated_shader->uniforms_size += 4; + + if (validation_state->needs_uniform_address_update) { + DRM_ERROR("Uniform read with undefined uniform " + "address\n"); + return false; + } + } + + return true; +} + +/* Make sure that all branches are absolute and point within the shader, and + * note their targets for later. + */ +static bool +vc4_validate_branches(struct vc4_shader_validation_state *validation_state) +{ + uint32_t max_branch_target = 0; + bool found_shader_end = false; + int ip; + int shader_end_ip = 0; + int last_branch = -2; + + for (ip = 0; ip < validation_state->max_ip; ip++) { + uint64_t inst = validation_state->shader[ip]; + int32_t branch_imm = QPU_GET_FIELD(inst, QPU_BRANCH_TARGET); + uint32_t sig = QPU_GET_FIELD(inst, QPU_SIG); + uint32_t after_delay_ip = ip + 4; + uint32_t branch_target_ip; + + if (sig == QPU_SIG_PROG_END) { + shader_end_ip = ip; + found_shader_end = true; + continue; + } + + if (sig != QPU_SIG_BRANCH) + continue; + + if (ip - last_branch < 4) { + DRM_ERROR("Branch at %d during delay slots\n", ip); + return false; + } + last_branch = ip; + + if (inst & QPU_BRANCH_REG) { + DRM_ERROR("branching from register relative " + "not supported\n"); + return false; + } + + if (!(inst & QPU_BRANCH_REL)) { + DRM_ERROR("relative branching required\n"); + return false; + } + + /* The actual branch target is the instruction after the delay + * slots, plus whatever byte offset is in the low 32 bits of + * the instruction. Make sure we're not branching beyond the + * end of the shader object. + */ + if (branch_imm % sizeof(inst) != 0) { + DRM_ERROR("branch target not aligned\n"); + return false; + } + + branch_target_ip = after_delay_ip + (branch_imm >> 3); + if (branch_target_ip >= validation_state->max_ip) { + DRM_ERROR("Branch at %d outside of shader (ip %d/%d)\n", + ip, branch_target_ip, + validation_state->max_ip); + return false; + } + set_bit(branch_target_ip, validation_state->branch_targets); + + /* Make sure that the non-branching path is also not outside + * the shader. + */ + if (after_delay_ip >= validation_state->max_ip) { + DRM_ERROR("Branch at %d continues past shader end " + "(%d/%d)\n", + ip, after_delay_ip, validation_state->max_ip); + return false; + } + set_bit(after_delay_ip, validation_state->branch_targets); + max_branch_target = max(max_branch_target, after_delay_ip); + + /* There are two delay slots after program end is signaled + * that are still executed, then we're finished. + */ + if (found_shader_end && ip == shader_end_ip + 2) + break; + } + + if (max_branch_target > shader_end_ip) { + DRM_ERROR("Branch landed after QPU_SIG_PROG_END"); + return false; } return true; } +/* Resets any known state for the shader, used when we may be branched to from + * multiple locations in the program (or at shader start). + */ +static void +reset_validation_state(struct vc4_shader_validation_state *validation_state) +{ + int i; + + for (i = 0; i < 8; i++) + validation_state->tmu_setup[i / 4].p_offset[i % 4] = ~0; + + for (i = 0; i < LIVE_REG_COUNT; i++) { + validation_state->live_min_clamp_offsets[i] = ~0; + validation_state->live_max_clamp_regs[i] = false; + validation_state->live_immediates[i] = ~0; + } +} + +static bool +texturing_in_progress(struct vc4_shader_validation_state *validation_state) +{ + return (validation_state->tmu_write_count[0] != 0 || + validation_state->tmu_write_count[1] != 0); +} + +static bool +vc4_handle_branch_target(struct vc4_shader_validation_state *validation_state) +{ + uint32_t ip = validation_state->ip; + + if (!test_bit(ip, validation_state->branch_targets)) + return true; + + if (texturing_in_progress(validation_state)) { + DRM_ERROR("Branch target landed during TMU setup\n"); + return false; + } + + /* Reset our live values tracking, since this instruction may have + * multiple predecessors. + * + * One could potentially do analysis to determine that, for + * example, all predecessors have a live max clamp in the same + * register, but we don't bother with that. + */ + reset_validation_state(validation_state); + + /* Since we've entered a basic block from potentially multiple + * predecessors, we need the uniforms address to be updated before any + * unforms are read. We require that after any branch point, the next + * uniform to be loaded is a uniform address offset. That uniform's + * offset will be marked by the uniform address register write + * validation, or a one-off the end-of-program check. + */ + validation_state->needs_uniform_address_update = true; + + return true; +} + struct vc4_validated_shader_info * vc4_validate_shader(struct drm_gem_cma_object *shader_obj) { bool found_shader_end = false; int shader_end_ip = 0; - uint32_t ip, max_ip; - uint64_t *shader; - struct vc4_validated_shader_info *validated_shader; + uint32_t ip; + struct vc4_validated_shader_info *validated_shader = NULL; struct vc4_shader_validation_state validation_state; - int i; memset(&validation_state, 0, sizeof(validation_state)); + validation_state.shader = shader_obj->vaddr; + validation_state.max_ip = shader_obj->base.size / sizeof(uint64_t); - for (i = 0; i < 8; i++) - validation_state.tmu_setup[i / 4].p_offset[i % 4] = ~0; - for (i = 0; i < ARRAY_SIZE(validation_state.live_min_clamp_offsets); i++) - validation_state.live_min_clamp_offsets[i] = ~0; + reset_validation_state(&validation_state); - shader = shader_obj->vaddr; - max_ip = shader_obj->base.size / sizeof(uint64_t); + validation_state.branch_targets = + kcalloc(BITS_TO_LONGS(validation_state.max_ip), + sizeof(unsigned long), GFP_KERNEL); + if (!validation_state.branch_targets) + goto fail; validated_shader = kcalloc(1, sizeof(*validated_shader), GFP_KERNEL); if (!validated_shader) - return NULL; + goto fail; + + if (!vc4_validate_branches(&validation_state)) + goto fail; - for (ip = 0; ip < max_ip; ip++) { - uint64_t inst = shader[ip]; + for (ip = 0; ip < validation_state.max_ip; ip++) { + uint64_t inst = validation_state.shader[ip]; uint32_t sig = QPU_GET_FIELD(inst, QPU_SIG); + validation_state.ip = ip; + + if (!vc4_handle_branch_target(&validation_state)) + goto fail; + switch (sig) { case QPU_SIG_NONE: case QPU_SIG_WAIT_FOR_SCOREBOARD: @@ -450,13 +791,14 @@ vc4_validate_shader(struct drm_gem_cma_object *shader_obj) case QPU_SIG_LOAD_TMU1: case QPU_SIG_PROG_END: case QPU_SIG_SMALL_IMM: - if (!check_instruction_writes(inst, validated_shader, + if (!check_instruction_writes(validated_shader, &validation_state)) { DRM_ERROR("Bad write at ip %d\n", ip); goto fail; } - if (!check_instruction_reads(inst, validated_shader)) + if (!check_instruction_reads(validated_shader, + &validation_state)) goto fail; if (sig == QPU_SIG_PROG_END) { @@ -467,13 +809,18 @@ vc4_validate_shader(struct drm_gem_cma_object *shader_obj) break; case QPU_SIG_LOAD_IMM: - if (!check_instruction_writes(inst, validated_shader, + if (!check_instruction_writes(validated_shader, &validation_state)) { DRM_ERROR("Bad LOAD_IMM write at ip %d\n", ip); goto fail; } break; + case QPU_SIG_BRANCH: + if (!check_branch(inst, validated_shader, + &validation_state, ip)) + goto fail; + break; default: DRM_ERROR("Unsupported QPU signal %d at " "instruction %d\n", sig, ip); @@ -487,13 +834,28 @@ vc4_validate_shader(struct drm_gem_cma_object *shader_obj) break; } - if (ip == max_ip) { + if (ip == validation_state.max_ip) { DRM_ERROR("shader failed to terminate before " "shader BO end at %zd\n", shader_obj->base.size); goto fail; } + /* If we did a backwards branch and we haven't emitted a uniforms + * reset since then, we still need the uniforms stream to have the + * uniforms address available so that the backwards branch can do its + * uniforms reset. + * + * We could potentially prove that the backwards branch doesn't + * contain any uses of uniforms until program exit, but that doesn't + * seem to be worth the trouble. + */ + if (validation_state.needs_uniform_address_for_loop) { + if (!require_uniform_address_uniform(validated_shader)) + goto fail; + validated_shader->uniforms_size += 4; + } + /* Again, no chance of integer overflow here because the worst case * scenario is 8 bytes of uniforms plus handles per 8-byte * instruction. @@ -502,9 +864,12 @@ vc4_validate_shader(struct drm_gem_cma_object *shader_obj) (validated_shader->uniforms_size + 4 * validated_shader->num_texture_samples); + kfree(validation_state.branch_targets); + return validated_shader; fail: + kfree(validation_state.branch_targets); if (validated_shader) { kfree(validated_shader->texture_samples); kfree(validated_shader); diff --git a/drivers/gpu/drm/vgem/Makefile b/drivers/gpu/drm/vgem/Makefile index 3f4c7b842028..bfcdea1330e6 100644 --- a/drivers/gpu/drm/vgem/Makefile +++ b/drivers/gpu/drm/vgem/Makefile @@ -1,4 +1,4 @@ ccflags-y := -Iinclude/drm -vgem-y := vgem_drv.o +vgem-y := vgem_drv.o vgem_fence.o obj-$(CONFIG_DRM_VGEM) += vgem.o diff --git a/drivers/gpu/drm/vgem/vgem_drv.c b/drivers/gpu/drm/vgem/vgem_drv.c index 35ea5d02a827..c15bafb06665 100644 --- a/drivers/gpu/drm/vgem/vgem_drv.c +++ b/drivers/gpu/drm/vgem/vgem_drv.c @@ -42,81 +42,38 @@ #define DRIVER_MAJOR 1 #define DRIVER_MINOR 0 -void vgem_gem_put_pages(struct drm_vgem_gem_object *obj) -{ - drm_gem_put_pages(&obj->base, obj->pages, false, false); - obj->pages = NULL; -} - static void vgem_gem_free_object(struct drm_gem_object *obj) { struct drm_vgem_gem_object *vgem_obj = to_vgem_bo(obj); - drm_gem_free_mmap_offset(obj); - - if (vgem_obj->use_dma_buf && obj->dma_buf) { - dma_buf_put(obj->dma_buf); - obj->dma_buf = NULL; - } - drm_gem_object_release(obj); - - if (vgem_obj->pages) - vgem_gem_put_pages(vgem_obj); - - vgem_obj->pages = NULL; - kfree(vgem_obj); } -int vgem_gem_get_pages(struct drm_vgem_gem_object *obj) -{ - struct page **pages; - - if (obj->pages || obj->use_dma_buf) - return 0; - - pages = drm_gem_get_pages(&obj->base); - if (IS_ERR(pages)) { - return PTR_ERR(pages); - } - - obj->pages = pages; - - return 0; -} - static int vgem_gem_fault(struct vm_area_struct *vma, struct vm_fault *vmf) { struct drm_vgem_gem_object *obj = vma->vm_private_data; - loff_t num_pages; - pgoff_t page_offset; - int ret; - /* We don't use vmf->pgoff since that has the fake offset */ - page_offset = ((unsigned long)vmf->virtual_address - vma->vm_start) >> - PAGE_SHIFT; - - num_pages = DIV_ROUND_UP(obj->base.size, PAGE_SIZE); - - if (page_offset > num_pages) - return VM_FAULT_SIGBUS; - - ret = vm_insert_page(vma, (unsigned long)vmf->virtual_address, - obj->pages[page_offset]); - switch (ret) { - case 0: - return VM_FAULT_NOPAGE; - case -ENOMEM: - return VM_FAULT_OOM; - case -EBUSY: - return VM_FAULT_RETRY; - case -EFAULT: - case -EINVAL: - return VM_FAULT_SIGBUS; - default: - WARN_ON(1); - return VM_FAULT_SIGBUS; + unsigned long vaddr = (unsigned long)vmf->virtual_address; + struct page *page; + + page = shmem_read_mapping_page(file_inode(obj->base.filp)->i_mapping, + (vaddr - vma->vm_start) >> PAGE_SHIFT); + if (!IS_ERR(page)) { + vmf->page = page; + return 0; + } else switch (PTR_ERR(page)) { + case -ENOSPC: + case -ENOMEM: + return VM_FAULT_OOM; + case -EBUSY: + return VM_FAULT_RETRY; + case -EFAULT: + case -EINVAL: + return VM_FAULT_SIGBUS; + default: + WARN_ON_ONCE(PTR_ERR(page)); + return VM_FAULT_SIGBUS; } } @@ -126,6 +83,34 @@ static const struct vm_operations_struct vgem_gem_vm_ops = { .close = drm_gem_vm_close, }; +static int vgem_open(struct drm_device *dev, struct drm_file *file) +{ + struct vgem_file *vfile; + int ret; + + vfile = kzalloc(sizeof(*vfile), GFP_KERNEL); + if (!vfile) + return -ENOMEM; + + file->driver_priv = vfile; + + ret = vgem_fence_open(vfile); + if (ret) { + kfree(vfile); + return ret; + } + + return 0; +} + +static void vgem_preclose(struct drm_device *dev, struct drm_file *file) +{ + struct vgem_file *vfile = file->driver_priv; + + vgem_fence_close(vfile); + kfree(vfile); +} + /* ioctls */ static struct drm_gem_object *vgem_gem_create(struct drm_device *dev, @@ -134,57 +119,43 @@ static struct drm_gem_object *vgem_gem_create(struct drm_device *dev, unsigned long size) { struct drm_vgem_gem_object *obj; - struct drm_gem_object *gem_object; - int err; - - size = roundup(size, PAGE_SIZE); + int ret; obj = kzalloc(sizeof(*obj), GFP_KERNEL); if (!obj) return ERR_PTR(-ENOMEM); - gem_object = &obj->base; - - err = drm_gem_object_init(dev, gem_object, size); - if (err) - goto out; - - err = vgem_gem_get_pages(obj); - if (err) - goto out; - - err = drm_gem_handle_create(file, gem_object, handle); - if (err) - goto handle_out; + ret = drm_gem_object_init(dev, &obj->base, roundup(size, PAGE_SIZE)); + if (ret) + goto err_free; - drm_gem_object_unreference_unlocked(gem_object); + ret = drm_gem_handle_create(file, &obj->base, handle); + drm_gem_object_unreference_unlocked(&obj->base); + if (ret) + goto err; - return gem_object; + return &obj->base; -handle_out: - drm_gem_object_release(gem_object); -out: +err_free: kfree(obj); - return ERR_PTR(err); +err: + return ERR_PTR(ret); } static int vgem_gem_dumb_create(struct drm_file *file, struct drm_device *dev, struct drm_mode_create_dumb *args) { struct drm_gem_object *gem_object; - uint64_t size; - uint64_t pitch = args->width * DIV_ROUND_UP(args->bpp, 8); + u64 pitch, size; + pitch = args->width * DIV_ROUND_UP(args->bpp, 8); size = args->height * pitch; if (size == 0) return -EINVAL; gem_object = vgem_gem_create(dev, file, &args->handle, size); - - if (IS_ERR(gem_object)) { - DRM_DEBUG_DRIVER("object creation failed\n"); + if (IS_ERR(gem_object)) return PTR_ERR(gem_object); - } args->size = gem_object->size; args->pitch = pitch; @@ -194,26 +165,26 @@ static int vgem_gem_dumb_create(struct drm_file *file, struct drm_device *dev, return 0; } -int vgem_gem_dumb_map(struct drm_file *file, struct drm_device *dev, - uint32_t handle, uint64_t *offset) +static int vgem_gem_dumb_map(struct drm_file *file, struct drm_device *dev, + uint32_t handle, uint64_t *offset) { - int ret = 0; struct drm_gem_object *obj; + int ret; obj = drm_gem_object_lookup(file, handle); if (!obj) return -ENOENT; + if (!obj->filp) { + ret = -EINVAL; + goto unref; + } + ret = drm_gem_create_mmap_offset(obj); if (ret) goto unref; - BUG_ON(!obj->filp); - - obj->filp->private_data = obj; - *offset = drm_vma_node_offset_addr(&obj->vma_node); - unref: drm_gem_object_unreference_unlocked(obj); @@ -221,26 +192,134 @@ unref: } static struct drm_ioctl_desc vgem_ioctls[] = { + DRM_IOCTL_DEF_DRV(VGEM_FENCE_ATTACH, vgem_fence_attach_ioctl, DRM_AUTH|DRM_RENDER_ALLOW), + DRM_IOCTL_DEF_DRV(VGEM_FENCE_SIGNAL, vgem_fence_signal_ioctl, DRM_AUTH|DRM_RENDER_ALLOW), }; +static int vgem_mmap(struct file *filp, struct vm_area_struct *vma) +{ + unsigned long flags = vma->vm_flags; + int ret; + + ret = drm_gem_mmap(filp, vma); + if (ret) + return ret; + + /* Keep the WC mmaping set by drm_gem_mmap() but our pages + * are ordinary and not special. + */ + vma->vm_flags = flags | VM_DONTEXPAND | VM_DONTDUMP; + return 0; +} + static const struct file_operations vgem_driver_fops = { .owner = THIS_MODULE, .open = drm_open, - .mmap = drm_gem_mmap, + .mmap = vgem_mmap, .poll = drm_poll, .read = drm_read, .unlocked_ioctl = drm_ioctl, .release = drm_release, }; +static int vgem_prime_pin(struct drm_gem_object *obj) +{ + long n_pages = obj->size >> PAGE_SHIFT; + struct page **pages; + + /* Flush the object from the CPU cache so that importers can rely + * on coherent indirect access via the exported dma-address. + */ + pages = drm_gem_get_pages(obj); + if (IS_ERR(pages)) + return PTR_ERR(pages); + + drm_clflush_pages(pages, n_pages); + drm_gem_put_pages(obj, pages, true, false); + + return 0; +} + +static struct sg_table *vgem_prime_get_sg_table(struct drm_gem_object *obj) +{ + struct sg_table *st; + struct page **pages; + + pages = drm_gem_get_pages(obj); + if (IS_ERR(pages)) + return ERR_CAST(pages); + + st = drm_prime_pages_to_sg(pages, obj->size >> PAGE_SHIFT); + drm_gem_put_pages(obj, pages, false, false); + + return st; +} + +static void *vgem_prime_vmap(struct drm_gem_object *obj) +{ + long n_pages = obj->size >> PAGE_SHIFT; + struct page **pages; + void *addr; + + pages = drm_gem_get_pages(obj); + if (IS_ERR(pages)) + return NULL; + + addr = vmap(pages, n_pages, 0, pgprot_writecombine(PAGE_KERNEL)); + drm_gem_put_pages(obj, pages, false, false); + + return addr; +} + +static void vgem_prime_vunmap(struct drm_gem_object *obj, void *vaddr) +{ + vunmap(vaddr); +} + +static int vgem_prime_mmap(struct drm_gem_object *obj, + struct vm_area_struct *vma) +{ + int ret; + + if (obj->size < vma->vm_end - vma->vm_start) + return -EINVAL; + + if (!obj->filp) + return -ENODEV; + + ret = obj->filp->f_op->mmap(obj->filp, vma); + if (ret) + return ret; + + fput(vma->vm_file); + vma->vm_file = get_file(obj->filp); + vma->vm_flags |= VM_DONTEXPAND | VM_DONTDUMP; + vma->vm_page_prot = pgprot_writecombine(vm_get_page_prot(vma->vm_flags)); + + return 0; +} + static struct drm_driver vgem_driver = { - .driver_features = DRIVER_GEM, + .driver_features = DRIVER_GEM | DRIVER_PRIME, + .open = vgem_open, + .preclose = vgem_preclose, .gem_free_object_unlocked = vgem_gem_free_object, .gem_vm_ops = &vgem_gem_vm_ops, .ioctls = vgem_ioctls, + .num_ioctls = ARRAY_SIZE(vgem_ioctls), .fops = &vgem_driver_fops, + .dumb_create = vgem_gem_dumb_create, .dumb_map_offset = vgem_gem_dumb_map, + + .prime_handle_to_fd = drm_gem_prime_handle_to_fd, + .gem_prime_pin = vgem_prime_pin, + .gem_prime_export = drm_gem_prime_export, + .gem_prime_get_sg_table = vgem_prime_get_sg_table, + .gem_prime_vmap = vgem_prime_vmap, + .gem_prime_vunmap = vgem_prime_vunmap, + .gem_prime_mmap = vgem_prime_mmap, + .name = DRIVER_NAME, .desc = DRIVER_DESC, .date = DRIVER_DATE, @@ -248,7 +327,7 @@ static struct drm_driver vgem_driver = { .minor = DRIVER_MINOR, }; -struct drm_device *vgem_device; +static struct drm_device *vgem_device; static int __init vgem_init(void) { @@ -261,7 +340,6 @@ static int __init vgem_init(void) } ret = drm_dev_register(vgem_device, 0); - if (ret) goto out_unref; @@ -283,5 +361,6 @@ module_init(vgem_init); module_exit(vgem_exit); MODULE_AUTHOR("Red Hat, Inc."); +MODULE_AUTHOR("Intel Corporation"); MODULE_DESCRIPTION(DRIVER_DESC); MODULE_LICENSE("GPL and additional rights"); diff --git a/drivers/gpu/drm/vgem/vgem_drv.h b/drivers/gpu/drm/vgem/vgem_drv.h index e9f92f7ee275..1f8798ad329c 100644 --- a/drivers/gpu/drm/vgem/vgem_drv.h +++ b/drivers/gpu/drm/vgem/vgem_drv.h @@ -32,15 +32,25 @@ #include <drm/drmP.h> #include <drm/drm_gem.h> +#include <uapi/drm/vgem_drm.h> + +struct vgem_file { + struct idr fence_idr; + struct mutex fence_mutex; +}; + #define to_vgem_bo(x) container_of(x, struct drm_vgem_gem_object, base) struct drm_vgem_gem_object { struct drm_gem_object base; - struct page **pages; - bool use_dma_buf; }; -/* vgem_drv.c */ -extern void vgem_gem_put_pages(struct drm_vgem_gem_object *obj); -extern int vgem_gem_get_pages(struct drm_vgem_gem_object *obj); +int vgem_fence_open(struct vgem_file *file); +int vgem_fence_attach_ioctl(struct drm_device *dev, + void *data, + struct drm_file *file); +int vgem_fence_signal_ioctl(struct drm_device *dev, + void *data, + struct drm_file *file); +void vgem_fence_close(struct vgem_file *file); #endif diff --git a/drivers/gpu/drm/vgem/vgem_fence.c b/drivers/gpu/drm/vgem/vgem_fence.c new file mode 100644 index 000000000000..5c57c1ffa1f9 --- /dev/null +++ b/drivers/gpu/drm/vgem/vgem_fence.c @@ -0,0 +1,283 @@ +/* + * Copyright 2016 Intel Corporation + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software") + * to deal in the software without restriction, including without limitation + * on the rights to use, copy, modify, merge, publish, distribute, sub + * license, and/or sell copies of the Software, and to permit persons to whom + * them Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTIBILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS BE LIABLE FOR ANY CLAIM, DAMAGES, OR OTHER LIABILITY, WHETHER + * IN AN ACTION OF CONTRACT, TORT, OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +#include <linux/dma-buf.h> +#include <linux/reservation.h> + +#include "vgem_drv.h" + +#define VGEM_FENCE_TIMEOUT (10*HZ) + +struct vgem_fence { + struct fence base; + struct spinlock lock; + struct timer_list timer; +}; + +static const char *vgem_fence_get_driver_name(struct fence *fence) +{ + return "vgem"; +} + +static const char *vgem_fence_get_timeline_name(struct fence *fence) +{ + return "unbound"; +} + +static bool vgem_fence_signaled(struct fence *fence) +{ + return false; +} + +static bool vgem_fence_enable_signaling(struct fence *fence) +{ + return true; +} + +static void vgem_fence_release(struct fence *base) +{ + struct vgem_fence *fence = container_of(base, typeof(*fence), base); + + del_timer_sync(&fence->timer); + fence_free(&fence->base); +} + +static void vgem_fence_value_str(struct fence *fence, char *str, int size) +{ + snprintf(str, size, "%u", fence->seqno); +} + +static void vgem_fence_timeline_value_str(struct fence *fence, char *str, + int size) +{ + snprintf(str, size, "%u", fence_is_signaled(fence) ? fence->seqno : 0); +} + +static const struct fence_ops vgem_fence_ops = { + .get_driver_name = vgem_fence_get_driver_name, + .get_timeline_name = vgem_fence_get_timeline_name, + .enable_signaling = vgem_fence_enable_signaling, + .signaled = vgem_fence_signaled, + .wait = fence_default_wait, + .release = vgem_fence_release, + + .fence_value_str = vgem_fence_value_str, + .timeline_value_str = vgem_fence_timeline_value_str, +}; + +static void vgem_fence_timeout(unsigned long data) +{ + struct vgem_fence *fence = (struct vgem_fence *)data; + + fence_signal(&fence->base); +} + +static struct fence *vgem_fence_create(struct vgem_file *vfile, + unsigned int flags) +{ + struct vgem_fence *fence; + + fence = kzalloc(sizeof(*fence), GFP_KERNEL); + if (!fence) + return NULL; + + spin_lock_init(&fence->lock); + fence_init(&fence->base, &vgem_fence_ops, &fence->lock, + fence_context_alloc(1), 1); + + setup_timer(&fence->timer, vgem_fence_timeout, (unsigned long)fence); + + /* We force the fence to expire within 10s to prevent driver hangs */ + mod_timer(&fence->timer, jiffies + VGEM_FENCE_TIMEOUT); + + return &fence->base; +} + +static int attach_dmabuf(struct drm_device *dev, + struct drm_gem_object *obj) +{ + struct dma_buf *dmabuf; + + if (obj->dma_buf) + return 0; + + dmabuf = dev->driver->gem_prime_export(dev, obj, 0); + if (IS_ERR(dmabuf)) + return PTR_ERR(dmabuf); + + obj->dma_buf = dmabuf; + drm_gem_object_reference(obj); + return 0; +} + +/* + * vgem_fence_attach_ioctl (DRM_IOCTL_VGEM_FENCE_ATTACH): + * + * Create and attach a fence to the vGEM handle. This fence is then exposed + * via the dma-buf reservation object and visible to consumers of the exported + * dma-buf. If the flags contain VGEM_FENCE_WRITE, the fence indicates the + * vGEM buffer is being written to by the client and is exposed as an exclusive + * fence, otherwise the fence indicates the client is current reading from the + * buffer and all future writes should wait for the client to signal its + * completion. Note that if a conflicting fence is already on the dma-buf (i.e. + * an exclusive fence when adding a read, or any fence when adding a write), + * -EBUSY is reported. Serialisation between operations should be handled + * by waiting upon the dma-buf. + * + * This returns the handle for the new fence that must be signaled within 10 + * seconds (or otherwise it will automatically expire). See + * vgem_fence_signal_ioctl (DRM_IOCTL_VGEM_FENCE_SIGNAL). + * + * If the vGEM handle does not exist, vgem_fence_attach_ioctl returns -ENOENT. + */ +int vgem_fence_attach_ioctl(struct drm_device *dev, + void *data, + struct drm_file *file) +{ + struct drm_vgem_fence_attach *arg = data; + struct vgem_file *vfile = file->driver_priv; + struct reservation_object *resv; + struct drm_gem_object *obj; + struct fence *fence; + int ret; + + if (arg->flags & ~VGEM_FENCE_WRITE) + return -EINVAL; + + if (arg->pad) + return -EINVAL; + + obj = drm_gem_object_lookup(file, arg->handle); + if (!obj) + return -ENOENT; + + ret = attach_dmabuf(dev, obj); + if (ret) + goto err; + + fence = vgem_fence_create(vfile, arg->flags); + if (!fence) { + ret = -ENOMEM; + goto err; + } + + /* Check for a conflicting fence */ + resv = obj->dma_buf->resv; + if (!reservation_object_test_signaled_rcu(resv, + arg->flags & VGEM_FENCE_WRITE)) { + ret = -EBUSY; + goto err_fence; + } + + /* Expose the fence via the dma-buf */ + ret = 0; + mutex_lock(&resv->lock.base); + if (arg->flags & VGEM_FENCE_WRITE) + reservation_object_add_excl_fence(resv, fence); + else if ((ret = reservation_object_reserve_shared(resv)) == 0) + reservation_object_add_shared_fence(resv, fence); + mutex_unlock(&resv->lock.base); + + /* Record the fence in our idr for later signaling */ + if (ret == 0) { + mutex_lock(&vfile->fence_mutex); + ret = idr_alloc(&vfile->fence_idr, fence, 1, 0, GFP_KERNEL); + mutex_unlock(&vfile->fence_mutex); + if (ret > 0) { + arg->out_fence = ret; + ret = 0; + } + } +err_fence: + if (ret) { + fence_signal(fence); + fence_put(fence); + } +err: + drm_gem_object_unreference_unlocked(obj); + return ret; +} + +/* + * vgem_fence_signal_ioctl (DRM_IOCTL_VGEM_FENCE_SIGNAL): + * + * Signal and consume a fence ealier attached to a vGEM handle using + * vgem_fence_attach_ioctl (DRM_IOCTL_VGEM_FENCE_ATTACH). + * + * All fences must be signaled within 10s of attachment or otherwise they + * will automatically expire (and a vgem_fence_signal_ioctl returns -ETIMEDOUT). + * + * Signaling a fence indicates to all consumers of the dma-buf that the + * client has completed the operation associated with the fence, and that the + * buffer is then ready for consumption. + * + * If the fence does not exist (or has already been signaled by the client), + * vgem_fence_signal_ioctl returns -ENOENT. + */ +int vgem_fence_signal_ioctl(struct drm_device *dev, + void *data, + struct drm_file *file) +{ + struct vgem_file *vfile = file->driver_priv; + struct drm_vgem_fence_signal *arg = data; + struct fence *fence; + int ret = 0; + + if (arg->flags) + return -EINVAL; + + mutex_lock(&vfile->fence_mutex); + fence = idr_replace(&vfile->fence_idr, NULL, arg->fence); + mutex_unlock(&vfile->fence_mutex); + if (!fence) + return -ENOENT; + if (IS_ERR(fence)) + return PTR_ERR(fence); + + if (fence_is_signaled(fence)) + ret = -ETIMEDOUT; + + fence_signal(fence); + fence_put(fence); + return ret; +} + +int vgem_fence_open(struct vgem_file *vfile) +{ + mutex_init(&vfile->fence_mutex); + idr_init(&vfile->fence_idr); + + return 0; +} + +static int __vgem_fence_idr_fini(int id, void *p, void *data) +{ + fence_signal(p); + fence_put(p); + return 0; +} + +void vgem_fence_close(struct vgem_file *vfile) +{ + idr_for_each(&vfile->fence_idr, __vgem_fence_idr_fini, vfile); + idr_destroy(&vfile->fence_idr); +} diff --git a/drivers/gpu/drm/virtio/Kconfig b/drivers/gpu/drm/virtio/Kconfig index 9983eadb81b6..e1afc3d3f8d9 100644 --- a/drivers/gpu/drm/virtio/Kconfig +++ b/drivers/gpu/drm/virtio/Kconfig @@ -1,11 +1,7 @@ config DRM_VIRTIO_GPU tristate "Virtio GPU driver" depends on DRM && VIRTIO - select FB_SYS_FILLRECT - select FB_SYS_COPYAREA - select FB_SYS_IMAGEBLIT select DRM_KMS_HELPER - select DRM_KMS_FB_HELPER select DRM_TTM help This is the virtual GPU driver for virtio. It can be used with diff --git a/drivers/gpu/drm/virtio/virtgpu_display.c b/drivers/gpu/drm/virtio/virtgpu_display.c index ac758cdbc1bc..4e192aa2d021 100644 --- a/drivers/gpu/drm/virtio/virtgpu_display.c +++ b/drivers/gpu/drm/virtio/virtgpu_display.c @@ -53,8 +53,7 @@ static void virtio_gpu_user_framebuffer_destroy(struct drm_framebuffer *fb) struct virtio_gpu_framebuffer *virtio_gpu_fb = to_virtio_gpu_framebuffer(fb); - if (virtio_gpu_fb->obj) - drm_gem_object_unreference_unlocked(virtio_gpu_fb->obj); + drm_gem_object_unreference_unlocked(virtio_gpu_fb->obj); drm_framebuffer_cleanup(fb); kfree(virtio_gpu_fb); } @@ -326,8 +325,7 @@ virtio_gpu_user_framebuffer_create(struct drm_device *dev, ret = virtio_gpu_framebuffer_init(dev, virtio_gpu_fb, mode_cmd, obj); if (ret) { kfree(virtio_gpu_fb); - if (obj) - drm_gem_object_unreference_unlocked(obj); + drm_gem_object_unreference_unlocked(obj); return NULL; } @@ -348,7 +346,7 @@ static void vgdev_atomic_commit_tail(struct drm_atomic_state *state) drm_atomic_helper_cleanup_planes(dev, state); } -struct drm_mode_config_helper_funcs virtio_mode_config_helpers = { +static struct drm_mode_config_helper_funcs virtio_mode_config_helpers = { .atomic_commit_tail = vgdev_atomic_commit_tail, }; diff --git a/drivers/gpu/drm/virtio/virtgpu_ttm.c b/drivers/gpu/drm/virtio/virtgpu_ttm.c index a0580815629f..80482ac5f95d 100644 --- a/drivers/gpu/drm/virtio/virtgpu_ttm.c +++ b/drivers/gpu/drm/virtio/virtgpu_ttm.c @@ -375,6 +375,12 @@ static int virtio_gpu_bo_move(struct ttm_buffer_object *bo, bool no_wait_gpu, struct ttm_mem_reg *new_mem) { + int ret; + + ret = ttm_bo_wait(bo, interruptible, no_wait_gpu); + if (ret) + return ret; + virtio_gpu_move_null(bo, new_mem); return 0; } diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_dmabuf.c b/drivers/gpu/drm/vmwgfx/vmwgfx_dmabuf.c index 9b078a493996..0cd889015dc5 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_dmabuf.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_dmabuf.c @@ -49,6 +49,7 @@ int vmw_dmabuf_pin_in_placement(struct vmw_private *dev_priv, { struct ttm_buffer_object *bo = &buf->base; int ret; + uint32_t new_flags; ret = ttm_write_lock(&dev_priv->reservation_sem, interruptible); if (unlikely(ret != 0)) @@ -60,7 +61,12 @@ int vmw_dmabuf_pin_in_placement(struct vmw_private *dev_priv, if (unlikely(ret != 0)) goto err; - ret = ttm_bo_validate(bo, placement, interruptible, false); + if (buf->pin_count > 0) + ret = ttm_bo_mem_compat(placement, &bo->mem, + &new_flags) == true ? 0 : -EINVAL; + else + ret = ttm_bo_validate(bo, placement, interruptible, false); + if (!ret) vmw_bo_pin_reserved(buf, true); @@ -91,6 +97,7 @@ int vmw_dmabuf_pin_in_vram_or_gmr(struct vmw_private *dev_priv, { struct ttm_buffer_object *bo = &buf->base; int ret; + uint32_t new_flags; ret = ttm_write_lock(&dev_priv->reservation_sem, interruptible); if (unlikely(ret != 0)) @@ -102,6 +109,12 @@ int vmw_dmabuf_pin_in_vram_or_gmr(struct vmw_private *dev_priv, if (unlikely(ret != 0)) goto err; + if (buf->pin_count > 0) { + ret = ttm_bo_mem_compat(&vmw_vram_gmr_placement, &bo->mem, + &new_flags) == true ? 0 : -EINVAL; + goto out_unreserve; + } + ret = ttm_bo_validate(bo, &vmw_vram_gmr_placement, interruptible, false); if (likely(ret == 0) || ret == -ERESTARTSYS) @@ -161,6 +174,7 @@ int vmw_dmabuf_pin_in_start_of_vram(struct vmw_private *dev_priv, struct ttm_placement placement; struct ttm_place place; int ret = 0; + uint32_t new_flags; place = vmw_vram_placement.placement[0]; place.lpfn = bo->num_pages; @@ -185,10 +199,15 @@ int vmw_dmabuf_pin_in_start_of_vram(struct vmw_private *dev_priv, */ if (bo->mem.mem_type == TTM_PL_VRAM && bo->mem.start < bo->num_pages && - bo->mem.start > 0) + bo->mem.start > 0 && + buf->pin_count == 0) (void) ttm_bo_validate(bo, &vmw_sys_placement, false, false); - ret = ttm_bo_validate(bo, &placement, interruptible, false); + if (buf->pin_count > 0) + ret = ttm_bo_mem_compat(&placement, &bo->mem, + &new_flags) == true ? 0 : -EINVAL; + else + ret = ttm_bo_validate(bo, &placement, interruptible, false); /* For some reason we didn't end up at the start of vram */ WARN_ON(ret == 0 && bo->offset != 0); diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c index 60646644bef3..e8ae3dc476d1 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c @@ -233,6 +233,7 @@ static int vmw_force_iommu; static int vmw_restrict_iommu; static int vmw_force_coherent; static int vmw_restrict_dma_mask; +static int vmw_assume_16bpp; static int vmw_probe(struct pci_dev *, const struct pci_device_id *); static void vmw_master_init(struct vmw_master *); @@ -249,6 +250,8 @@ MODULE_PARM_DESC(force_coherent, "Force coherent TTM pages"); module_param_named(force_coherent, vmw_force_coherent, int, 0600); MODULE_PARM_DESC(restrict_dma_mask, "Restrict DMA mask to 44 bits with IOMMU"); module_param_named(restrict_dma_mask, vmw_restrict_dma_mask, int, 0600); +MODULE_PARM_DESC(assume_16bpp, "Assume 16-bpp when filtering modes"); +module_param_named(assume_16bpp, vmw_assume_16bpp, int, 0600); static void vmw_print_capabilities(uint32_t capabilities) @@ -660,6 +663,8 @@ static int vmw_driver_load(struct drm_device *dev, unsigned long chipset) dev_priv->vram_start = pci_resource_start(dev->pdev, 1); dev_priv->mmio_start = pci_resource_start(dev->pdev, 2); + dev_priv->assume_16bpp = !!vmw_assume_16bpp; + dev_priv->enable_fb = enable_fbdev; vmw_write(dev_priv, SVGA_REG_ID, SVGA_ID_2); @@ -706,6 +711,13 @@ static int vmw_driver_load(struct drm_device *dev, unsigned long chipset) vmw_read(dev_priv, SVGA_REG_SUGGESTED_GBOBJECT_MEM_SIZE_KB); + /* + * Workaround for low memory 2D VMs to compensate for the + * allocation taken by fbdev + */ + if (!(dev_priv->capabilities & SVGA_CAP_3D)) + mem_size *= 2; + dev_priv->max_mob_pages = mem_size * 1024 / PAGE_SIZE; dev_priv->prim_bb_mem = vmw_read(dev_priv, @@ -1041,8 +1053,7 @@ static struct vmw_master *vmw_master_check(struct drm_device *dev, struct vmw_fpriv *vmw_fp = vmw_fpriv(file_priv); struct vmw_master *vmaster; - if (file_priv->minor->type != DRM_MINOR_LEGACY || - !(flags & DRM_AUTH)) + if (!drm_is_primary_client(file_priv) || !(flags & DRM_AUTH)) return NULL; ret = mutex_lock_interruptible(&dev->master_mutex); diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h index 9a90f824814e..74304b03f9d4 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h @@ -387,6 +387,7 @@ struct vmw_private { spinlock_t hw_lock; spinlock_t cap_lock; bool has_dx; + bool assume_16bpp; /* * VGA registers. diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c index 1a1a87cbf109..dc5beff2b4aa 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c @@ -3625,9 +3625,7 @@ static int vmw_resize_cmd_bounce(struct vmw_sw_context *sw_context, (sw_context->cmd_bounce_size >> 1)); } - if (sw_context->cmd_bounce != NULL) - vfree(sw_context->cmd_bounce); - + vfree(sw_context->cmd_bounce); sw_context->cmd_bounce = vmalloc(sw_context->cmd_bounce_size); if (sw_context->cmd_bounce == NULL) { diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c b/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c index 679a4cb98ee3..d2d93959b119 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c @@ -517,28 +517,6 @@ static int vmw_fb_kms_framebuffer(struct fb_info *info) par->set_fb = &vfb->base; - if (!par->bo_ptr) { - /* - * Pin before mapping. Since we don't know in what placement - * to pin, call into KMS to do it for us. - */ - ret = vfb->pin(vfb); - if (ret) { - DRM_ERROR("Could not pin the fbdev framebuffer.\n"); - return ret; - } - - ret = ttm_bo_kmap(&par->vmw_bo->base, 0, - par->vmw_bo->base.num_pages, &par->map); - if (ret) { - vfb->unpin(vfb); - DRM_ERROR("Could not map the fbdev framebuffer.\n"); - return ret; - } - - par->bo_ptr = ttm_kmap_obj_virtual(&par->map, &par->bo_iowrite); - } - return 0; } @@ -601,6 +579,31 @@ static int vmw_fb_set_par(struct fb_info *info) if (ret) goto out_unlock; + if (!par->bo_ptr) { + struct vmw_framebuffer *vfb = vmw_framebuffer_to_vfb(set.fb); + + /* + * Pin before mapping. Since we don't know in what placement + * to pin, call into KMS to do it for us. + */ + ret = vfb->pin(vfb); + if (ret) { + DRM_ERROR("Could not pin the fbdev framebuffer.\n"); + goto out_unlock; + } + + ret = ttm_bo_kmap(&par->vmw_bo->base, 0, + par->vmw_bo->base.num_pages, &par->map); + if (ret) { + vfb->unpin(vfb); + DRM_ERROR("Could not map the fbdev framebuffer.\n"); + goto out_unlock; + } + + par->bo_ptr = ttm_kmap_obj_virtual(&par->map, &par->bo_iowrite); + } + + vmw_fb_dirty_mark(par, par->fb_x, par->fb_y, par->set_fb->width, par->set_fb->height); diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c index 8a69d4da40b5..bf28ccc150df 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c @@ -1555,14 +1555,10 @@ int vmw_du_connector_fill_modes(struct drm_connector *connector, DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) }; int i; - u32 assumed_bpp = 2; + u32 assumed_bpp = 4; - /* - * If using screen objects, then assume 32-bpp because that's what the - * SVGA device is assuming - */ - if (dev_priv->active_display_unit == vmw_du_screen_object) - assumed_bpp = 4; + if (dev_priv->assume_16bpp) + assumed_bpp = 2; if (dev_priv->active_display_unit == vmw_du_screen_target) { max_width = min(max_width, dev_priv->stdu_max_width); diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_msg.c b/drivers/gpu/drm/vmwgfx/vmwgfx_msg.c index 6de283c8fa3e..e57a0bad7a62 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_msg.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_msg.c @@ -28,6 +28,7 @@ #include <linux/slab.h> #include <linux/module.h> #include <linux/kernel.h> +#include <linux/frame.h> #include <asm/hypervisor.h> #include "drmP.h" #include "vmwgfx_msg.h" @@ -194,7 +195,7 @@ static int vmw_send_msg(struct rpc_channel *channel, const char *msg) return -EINVAL; } - +STACK_FRAME_NON_STANDARD(vmw_send_msg); /** @@ -299,11 +300,15 @@ static int vmw_recv_msg(struct rpc_channel *channel, void **msg, break; } + if (retries == RETRIES) + return -EINVAL; + *msg_len = reply_len; *msg = reply; return 0; } +STACK_FRAME_NON_STANDARD(vmw_recv_msg); /** diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_stdu.c b/drivers/gpu/drm/vmwgfx/vmwgfx_stdu.c index 9ca818fb034c..41932a7c4f79 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_stdu.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_stdu.c @@ -399,8 +399,10 @@ static int vmw_stdu_bind_fb(struct vmw_private *dev_priv, WARN_ON_ONCE(!stdu->defined); - if (!vfb->dmabuf && new_fb->width == mode->hdisplay && - new_fb->height == mode->vdisplay) + new_vfbs = (vfb->dmabuf) ? NULL : vmw_framebuffer_to_vfbs(new_fb); + + if (new_vfbs && new_vfbs->surface->base_size.width == mode->hdisplay && + new_vfbs->surface->base_size.height == mode->vdisplay) new_content_type = SAME_AS_DISPLAY; else if (vfb->dmabuf) new_content_type = SEPARATE_DMA; @@ -444,7 +446,6 @@ static int vmw_stdu_bind_fb(struct vmw_private *dev_priv, content_srf.mip_levels[0] = 1; content_srf.multisample_count = 0; } else { - new_vfbs = vmw_framebuffer_to_vfbs(new_fb); content_srf = *new_vfbs->surface; } @@ -464,7 +465,6 @@ static int vmw_stdu_bind_fb(struct vmw_private *dev_priv, return ret; } } else if (new_content_type == SAME_AS_DISPLAY) { - new_vfbs = vmw_framebuffer_to_vfbs(new_fb); new_display_srf = vmw_surface_reference(new_vfbs->surface); } diff --git a/drivers/gpu/host1x/cdma.c b/drivers/gpu/host1x/cdma.c index a18db4d5347c..c5d82a8a2ec9 100644 --- a/drivers/gpu/host1x/cdma.c +++ b/drivers/gpu/host1x/cdma.c @@ -96,12 +96,12 @@ fail: */ static void host1x_pushbuffer_push(struct push_buffer *pb, u32 op1, u32 op2) { - u32 pos = pb->pos; - u32 *p = (u32 *)((void *)pb->mapped + pos); - WARN_ON(pos == pb->fence); + u32 *p = (u32 *)((void *)pb->mapped + pb->pos); + + WARN_ON(pb->pos == pb->fence); *(p++) = op1; *(p++) = op2; - pb->pos = (pos + 8) & (pb->size_bytes - 1); + pb->pos = (pb->pos + 8) & (pb->size_bytes - 1); } /* @@ -134,14 +134,19 @@ unsigned int host1x_cdma_wait_locked(struct host1x_cdma *cdma, enum cdma_event event) { for (;;) { + struct push_buffer *pb = &cdma->push_buffer; unsigned int space; - if (event == CDMA_EVENT_SYNC_QUEUE_EMPTY) + switch (event) { + case CDMA_EVENT_SYNC_QUEUE_EMPTY: space = list_empty(&cdma->sync_queue) ? 1 : 0; - else if (event == CDMA_EVENT_PUSH_BUFFER_SPACE) { - struct push_buffer *pb = &cdma->push_buffer; + break; + + case CDMA_EVENT_PUSH_BUFFER_SPACE: space = host1x_pushbuffer_space(pb); - } else { + break; + + default: WARN_ON(1); return -EINVAL; } @@ -159,12 +164,14 @@ unsigned int host1x_cdma_wait_locked(struct host1x_cdma *cdma, mutex_lock(&cdma->lock); continue; } + cdma->event = event; mutex_unlock(&cdma->lock); down(&cdma->sem); mutex_lock(&cdma->lock); } + return 0; } @@ -234,6 +241,7 @@ static void update_cdma_locked(struct host1x_cdma *cdma) /* Start timer on next pending syncpt */ if (job->timeout) cdma_start_timer_locked(cdma, job); + break; } @@ -247,7 +255,9 @@ static void update_cdma_locked(struct host1x_cdma *cdma) /* Pop push buffer slots */ if (job->num_slots) { struct push_buffer *pb = &cdma->push_buffer; + host1x_pushbuffer_pop(pb, job->num_slots); + if (cdma->event == CDMA_EVENT_PUSH_BUFFER_SPACE) signal = true; } @@ -269,11 +279,9 @@ static void update_cdma_locked(struct host1x_cdma *cdma) void host1x_cdma_update_sync_queue(struct host1x_cdma *cdma, struct device *dev) { - u32 restart_addr; - u32 syncpt_incrs; - struct host1x_job *job = NULL; - u32 syncpt_val; struct host1x *host1x = cdma_to_host1x(cdma); + u32 restart_addr, syncpt_incrs, syncpt_val; + struct host1x_job *job = NULL; syncpt_val = host1x_syncpt_load(cdma->timeout.syncpt); @@ -342,9 +350,11 @@ void host1x_cdma_update_sync_queue(struct host1x_cdma *cdma, syncpt_val += syncpt_incrs; } - /* The following sumbits from the same client may be dependent on the + /* + * The following sumbits from the same client may be dependent on the * failed submit and therefore they may fail. Force a small timeout - * to make the queue cleanup faster */ + * to make the queue cleanup faster. + */ list_for_each_entry_from(job, &cdma->sync_queue, list) if (job->client == cdma->timeout.client) @@ -375,6 +385,7 @@ int host1x_cdma_init(struct host1x_cdma *cdma) err = host1x_pushbuffer_init(&cdma->push_buffer); if (err) return err; + return 0; } @@ -410,6 +421,7 @@ int host1x_cdma_begin(struct host1x_cdma *cdma, struct host1x_job *job) /* init state on first submit with timeout value */ if (!cdma->timeout.initialized) { int err; + err = host1x_hw_cdma_timeout_init(host1x, cdma, job->syncpt_id); if (err) { @@ -418,6 +430,7 @@ int host1x_cdma_begin(struct host1x_cdma *cdma, struct host1x_job *job) } } } + if (!cdma->running) host1x_hw_cdma_start(host1x, cdma); @@ -448,6 +461,7 @@ void host1x_cdma_push(struct host1x_cdma *cdma, u32 op1, u32 op2) slots_free = host1x_cdma_wait_locked(cdma, CDMA_EVENT_PUSH_BUFFER_SPACE); } + cdma->slots_free = slots_free - 1; cdma->slots_used++; host1x_pushbuffer_push(pb, op1, op2); diff --git a/drivers/gpu/host1x/channel.c b/drivers/gpu/host1x/channel.c index b4ae3affb987..8f437d924c10 100644 --- a/drivers/gpu/host1x/channel.c +++ b/drivers/gpu/host1x/channel.c @@ -83,9 +83,10 @@ EXPORT_SYMBOL(host1x_channel_put); struct host1x_channel *host1x_channel_request(struct device *dev) { struct host1x *host = dev_get_drvdata(dev->parent); - int max_channels = host->info->nb_channels; + unsigned int max_channels = host->info->nb_channels; struct host1x_channel *channel = NULL; - int index, err; + unsigned long index; + int err; mutex_lock(&host->chlist_mutex); diff --git a/drivers/gpu/host1x/debug.c b/drivers/gpu/host1x/debug.c index ee3d12b51c50..d9330fcc62ad 100644 --- a/drivers/gpu/host1x/debug.c +++ b/drivers/gpu/host1x/debug.c @@ -39,6 +39,7 @@ void host1x_debug_output(struct output *o, const char *fmt, ...) va_start(args, fmt); len = vsnprintf(o->buf, sizeof(o->buf), fmt, args); va_end(args); + o->fn(o->ctx, o->buf, len); } @@ -48,13 +49,17 @@ static int show_channels(struct host1x_channel *ch, void *data, bool show_fifo) struct output *o = data; mutex_lock(&ch->reflock); + if (ch->refcount) { mutex_lock(&ch->cdma.lock); + if (show_fifo) host1x_hw_show_channel_fifo(m, ch, o); + host1x_hw_show_channel_cdma(m, ch, o); mutex_unlock(&ch->cdma.lock); } + mutex_unlock(&ch->reflock); return 0; @@ -62,22 +67,27 @@ static int show_channels(struct host1x_channel *ch, void *data, bool show_fifo) static void show_syncpts(struct host1x *m, struct output *o) { - int i; + unsigned int i; + host1x_debug_output(o, "---- syncpts ----\n"); + for (i = 0; i < host1x_syncpt_nb_pts(m); i++) { u32 max = host1x_syncpt_read_max(m->syncpt + i); u32 min = host1x_syncpt_load(m->syncpt + i); + if (!min && !max) continue; - host1x_debug_output(o, "id %d (%s) min %d max %d\n", + + host1x_debug_output(o, "id %u (%s) min %d max %d\n", i, m->syncpt[i].name, min, max); } for (i = 0; i < host1x_syncpt_nb_bases(m); i++) { u32 base_val; + base_val = host1x_syncpt_load_wait_base(m->syncpt + i); if (base_val) - host1x_debug_output(o, "waitbase id %d val %d\n", i, + host1x_debug_output(o, "waitbase id %u val %d\n", i, base_val); } @@ -114,7 +124,9 @@ static int host1x_debug_show_all(struct seq_file *s, void *unused) .fn = write_to_seqfile, .ctx = s }; + show_all(s->private, &o); + return 0; } @@ -124,7 +136,9 @@ static int host1x_debug_show(struct seq_file *s, void *unused) .fn = write_to_seqfile, .ctx = s }; + show_all_no_fifo(s->private, &o); + return 0; } @@ -134,10 +148,10 @@ static int host1x_debug_open_all(struct inode *inode, struct file *file) } static const struct file_operations host1x_debug_all_fops = { - .open = host1x_debug_open_all, - .read = seq_read, - .llseek = seq_lseek, - .release = single_release, + .open = host1x_debug_open_all, + .read = seq_read, + .llseek = seq_lseek, + .release = single_release, }; static int host1x_debug_open(struct inode *inode, struct file *file) @@ -146,10 +160,10 @@ static int host1x_debug_open(struct inode *inode, struct file *file) } static const struct file_operations host1x_debug_fops = { - .open = host1x_debug_open, - .read = seq_read, - .llseek = seq_lseek, - .release = single_release, + .open = host1x_debug_open, + .read = seq_read, + .llseek = seq_lseek, + .release = single_release, }; static void host1x_debugfs_init(struct host1x *host1x) @@ -201,6 +215,7 @@ void host1x_debug_dump(struct host1x *host1x) struct output o = { .fn = write_to_printk }; + show_all(host1x, &o); } @@ -209,5 +224,6 @@ void host1x_debug_dump_syncpts(struct host1x *host1x) struct output o = { .fn = write_to_printk }; + show_syncpts(host1x, &o); } diff --git a/drivers/gpu/host1x/dev.c b/drivers/gpu/host1x/dev.c index ff348690df94..a62317af76ad 100644 --- a/drivers/gpu/host1x/dev.c +++ b/drivers/gpu/host1x/dev.c @@ -63,13 +63,13 @@ u32 host1x_ch_readl(struct host1x_channel *ch, u32 r) } static const struct host1x_info host1x01_info = { - .nb_channels = 8, - .nb_pts = 32, - .nb_mlocks = 16, - .nb_bases = 8, - .init = host1x01_init, - .sync_offset = 0x3000, - .dma_mask = DMA_BIT_MASK(32), + .nb_channels = 8, + .nb_pts = 32, + .nb_mlocks = 16, + .nb_bases = 8, + .init = host1x01_init, + .sync_offset = 0x3000, + .dma_mask = DMA_BIT_MASK(32), }; static const struct host1x_info host1x02_info = { @@ -102,7 +102,7 @@ static const struct host1x_info host1x05_info = { .dma_mask = DMA_BIT_MASK(34), }; -static struct of_device_id host1x_of_match[] = { +static const struct of_device_id host1x_of_match[] = { { .compatible = "nvidia,tegra210-host1x", .data = &host1x05_info, }, { .compatible = "nvidia,tegra124-host1x", .data = &host1x04_info, }, { .compatible = "nvidia,tegra114-host1x", .data = &host1x02_info, }, diff --git a/drivers/gpu/host1x/dev.h b/drivers/gpu/host1x/dev.h index dace124994bb..5220510f39da 100644 --- a/drivers/gpu/host1x/dev.h +++ b/drivers/gpu/host1x/dev.h @@ -45,7 +45,7 @@ struct host1x_cdma_ops { void (*start)(struct host1x_cdma *cdma); void (*stop)(struct host1x_cdma *cdma); void (*flush)(struct host1x_cdma *cdma); - int (*timeout_init)(struct host1x_cdma *cdma, u32 syncpt_id); + int (*timeout_init)(struct host1x_cdma *cdma, unsigned int syncpt); void (*timeout_destroy)(struct host1x_cdma *cdma); void (*freeze)(struct host1x_cdma *cdma); void (*resume)(struct host1x_cdma *cdma, u32 getptr); @@ -82,21 +82,21 @@ struct host1x_intr_ops { int (*init_host_sync)(struct host1x *host, u32 cpm, void (*syncpt_thresh_work)(struct work_struct *work)); void (*set_syncpt_threshold)( - struct host1x *host, u32 id, u32 thresh); - void (*enable_syncpt_intr)(struct host1x *host, u32 id); - void (*disable_syncpt_intr)(struct host1x *host, u32 id); + struct host1x *host, unsigned int id, u32 thresh); + void (*enable_syncpt_intr)(struct host1x *host, unsigned int id); + void (*disable_syncpt_intr)(struct host1x *host, unsigned int id); void (*disable_all_syncpt_intrs)(struct host1x *host); int (*free_syncpt_irq)(struct host1x *host); }; struct host1x_info { - int nb_channels; /* host1x: num channels supported */ - int nb_pts; /* host1x: num syncpoints supported */ - int nb_bases; /* host1x: num syncpoints supported */ - int nb_mlocks; /* host1x: number of mlocks */ - int (*init)(struct host1x *); /* initialize per SoC ops */ - int sync_offset; - u64 dma_mask; /* mask of addressable memory */ + unsigned int nb_channels; /* host1x: number of channels supported */ + unsigned int nb_pts; /* host1x: number of syncpoints supported */ + unsigned int nb_bases; /* host1x: number of syncpoint bases supported */ + unsigned int nb_mlocks; /* host1x: number of mlocks supported */ + int (*init)(struct host1x *host1x); /* initialize per SoC ops */ + unsigned int sync_offset; /* offset of syncpoint registers */ + u64 dma_mask; /* mask of addressable memory */ }; struct host1x { @@ -109,7 +109,6 @@ struct host1x { struct clk *clk; struct mutex intr_mutex; - struct workqueue_struct *intr_wq; int intr_syncpt_irq; const struct host1x_syncpt_ops *syncpt_op; @@ -183,19 +182,20 @@ static inline int host1x_hw_intr_init_host_sync(struct host1x *host, u32 cpm, } static inline void host1x_hw_intr_set_syncpt_threshold(struct host1x *host, - u32 id, u32 thresh) + unsigned int id, + u32 thresh) { host->intr_op->set_syncpt_threshold(host, id, thresh); } static inline void host1x_hw_intr_enable_syncpt_intr(struct host1x *host, - u32 id) + unsigned int id) { host->intr_op->enable_syncpt_intr(host, id); } static inline void host1x_hw_intr_disable_syncpt_intr(struct host1x *host, - u32 id) + unsigned int id) { host->intr_op->disable_syncpt_intr(host, id); } @@ -212,9 +212,9 @@ static inline int host1x_hw_intr_free_syncpt_irq(struct host1x *host) static inline int host1x_hw_channel_init(struct host1x *host, struct host1x_channel *channel, - int chid) + unsigned int id) { - return host->channel_op->init(channel, host, chid); + return host->channel_op->init(channel, host, id); } static inline int host1x_hw_channel_submit(struct host1x *host, @@ -243,9 +243,9 @@ static inline void host1x_hw_cdma_flush(struct host1x *host, static inline int host1x_hw_cdma_timeout_init(struct host1x *host, struct host1x_cdma *cdma, - u32 syncpt_id) + unsigned int syncpt) { - return host->cdma_op->timeout_init(cdma, syncpt_id); + return host->cdma_op->timeout_init(cdma, syncpt); } static inline void host1x_hw_cdma_timeout_destroy(struct host1x *host, diff --git a/drivers/gpu/host1x/hw/cdma_hw.c b/drivers/gpu/host1x/hw/cdma_hw.c index 305ea8f3382d..659c1bbfeeba 100644 --- a/drivers/gpu/host1x/hw/cdma_hw.c +++ b/drivers/gpu/host1x/hw/cdma_hw.c @@ -41,7 +41,7 @@ static void cdma_timeout_cpu_incr(struct host1x_cdma *cdma, u32 getptr, { struct host1x *host1x = cdma_to_host1x(cdma); struct push_buffer *pb = &cdma->push_buffer; - u32 i; + unsigned int i; for (i = 0; i < syncpt_incrs; i++) host1x_syncpt_incr(cdma->timeout.syncpt); @@ -58,6 +58,7 @@ static void cdma_timeout_cpu_incr(struct host1x_cdma *cdma, u32 getptr, &pb->phys, getptr); getptr = (getptr + 8) & (pb->size_bytes - 1); } + wmb(); } @@ -162,12 +163,14 @@ static void cdma_stop(struct host1x_cdma *cdma) struct host1x_channel *ch = cdma_to_channel(cdma); mutex_lock(&cdma->lock); + if (cdma->running) { host1x_cdma_wait_locked(cdma, CDMA_EVENT_SYNC_QUEUE_EMPTY); host1x_ch_writel(ch, HOST1X_CHANNEL_DMACTRL_DMASTOP, HOST1X_CHANNEL_DMACTRL); cdma->running = false; } + mutex_unlock(&cdma->lock); } @@ -213,11 +216,11 @@ static void cdma_resume(struct host1x_cdma *cdma, u32 getptr) u32 cmdproc_stop; dev_dbg(host1x->dev, - "resuming channel (id %d, DMAGET restart = 0x%x)\n", + "resuming channel (id %u, DMAGET restart = 0x%x)\n", ch->id, getptr); cmdproc_stop = host1x_sync_readl(host1x, HOST1X_SYNC_CMDPROC_STOP); - cmdproc_stop &= ~(BIT(ch->id)); + cmdproc_stop &= ~BIT(ch->id); host1x_sync_writel(host1x, cmdproc_stop, HOST1X_SYNC_CMDPROC_STOP); cdma->torndown = false; @@ -231,14 +234,11 @@ static void cdma_resume(struct host1x_cdma *cdma, u32 getptr) */ static void cdma_timeout_handler(struct work_struct *work) { + u32 prev_cmdproc, cmdproc_stop, syncpt_val; struct host1x_cdma *cdma; struct host1x *host1x; struct host1x_channel *ch; - u32 syncpt_val; - - u32 prev_cmdproc, cmdproc_stop; - cdma = container_of(to_delayed_work(work), struct host1x_cdma, timeout.wq); host1x = cdma_to_host1x(cdma); @@ -277,9 +277,9 @@ static void cdma_timeout_handler(struct work_struct *work) return; } - dev_warn(host1x->dev, "%s: timeout: %d (%s), HW thresh %d, done %d\n", - __func__, cdma->timeout.syncpt->id, cdma->timeout.syncpt->name, - syncpt_val, cdma->timeout.syncpt_val); + dev_warn(host1x->dev, "%s: timeout: %u (%s), HW thresh %d, done %d\n", + __func__, cdma->timeout.syncpt->id, cdma->timeout.syncpt->name, + syncpt_val, cdma->timeout.syncpt_val); /* stop HW, resetting channel/module */ host1x_hw_cdma_freeze(host1x, cdma); @@ -291,7 +291,7 @@ static void cdma_timeout_handler(struct work_struct *work) /* * Init timeout resources */ -static int cdma_timeout_init(struct host1x_cdma *cdma, u32 syncpt_id) +static int cdma_timeout_init(struct host1x_cdma *cdma, unsigned int syncpt) { INIT_DELAYED_WORK(&cdma->timeout.wq, cdma_timeout_handler); cdma->timeout.initialized = true; @@ -306,6 +306,7 @@ static void cdma_timeout_destroy(struct host1x_cdma *cdma) { if (cdma->timeout.initialized) cancel_delayed_work(&cdma->timeout.wq); + cdma->timeout.initialized = false; } diff --git a/drivers/gpu/host1x/hw/channel_hw.c b/drivers/gpu/host1x/hw/channel_hw.c index 946c332c3906..5e8df78b7acd 100644 --- a/drivers/gpu/host1x/hw/channel_hw.c +++ b/drivers/gpu/host1x/hw/channel_hw.c @@ -46,6 +46,7 @@ static void trace_write_gather(struct host1x_cdma *cdma, struct host1x_bo *bo, */ for (i = 0; i < words; i += TRACE_MAX_LENGTH) { u32 num_words = min(words - i, TRACE_MAX_LENGTH); + offset += i * sizeof(u32); trace_host1x_cdma_push_gather(dev_name(dev), bo, @@ -66,6 +67,7 @@ static void submit_gathers(struct host1x_job *job) struct host1x_job_gather *g = &job->gathers[i]; u32 op1 = host1x_opcode_gather(g->words); u32 op2 = g->base + g->offset; + trace_write_gather(cdma, g->bo, g->offset, op1 & 0xffff); host1x_cdma_push(cdma, op1, op2); } @@ -75,7 +77,8 @@ static inline void synchronize_syncpt_base(struct host1x_job *job) { struct host1x *host = dev_get_drvdata(job->channel->dev->parent); struct host1x_syncpt *sp = host->syncpt + job->syncpt_id; - u32 id, value; + unsigned int id; + u32 value; value = host1x_syncpt_read_max(sp); id = sp->base->id; diff --git a/drivers/gpu/host1x/hw/debug_hw.c b/drivers/gpu/host1x/hw/debug_hw.c index cc3f1825c735..7a4a3286e4a7 100644 --- a/drivers/gpu/host1x/hw/debug_hw.c +++ b/drivers/gpu/host1x/hw/debug_hw.c @@ -40,8 +40,7 @@ enum { static unsigned int show_channel_command(struct output *o, u32 val) { - unsigned mask; - unsigned subop; + unsigned int mask, subop; switch (val >> 28) { case HOST1X_OPCODE_SETCLASS: @@ -51,12 +50,11 @@ static unsigned int show_channel_command(struct output *o, u32 val) val >> 6 & 0x3ff, val >> 16 & 0xfff, mask); return hweight8(mask); - } else { - host1x_debug_output(o, "SETCL(class=%03x)\n", - val >> 6 & 0x3ff); - return 0; } + host1x_debug_output(o, "SETCL(class=%03x)\n", val >> 6 & 0x3ff); + return 0; + case HOST1X_OPCODE_INCR: host1x_debug_output(o, "INCR(offset=%03x, [", val >> 16 & 0xfff); @@ -143,7 +141,8 @@ static void show_channel_gathers(struct output *o, struct host1x_cdma *cdma) struct host1x_job *job; list_for_each_entry(job, &cdma->sync_queue, list) { - int i; + unsigned int i; + host1x_debug_output(o, "\n%p: JOB, syncpt_id=%d, syncpt_val=%d, first_get=%08x, timeout=%d num_slots=%d, num_handles=%d\n", job, job->syncpt_id, job->syncpt_end, job->first_get, job->timeout, @@ -190,7 +189,7 @@ static void host1x_debug_show_channel_cdma(struct host1x *host, cbread = host1x_sync_readl(host, HOST1X_SYNC_CBREAD(ch->id)); cbstat = host1x_sync_readl(host, HOST1X_SYNC_CBSTAT(ch->id)); - host1x_debug_output(o, "%d-%s: ", ch->id, dev_name(ch->dev)); + host1x_debug_output(o, "%u-%s: ", ch->id, dev_name(ch->dev)); if (HOST1X_CHANNEL_DMACTRL_DMASTOP_V(dmactrl) || !ch->cdma.push_buffer.mapped) { @@ -200,14 +199,13 @@ static void host1x_debug_show_channel_cdma(struct host1x *host, if (HOST1X_SYNC_CBSTAT_CBCLASS_V(cbstat) == HOST1X_CLASS_HOST1X && HOST1X_SYNC_CBSTAT_CBOFFSET_V(cbstat) == - HOST1X_UCLASS_WAIT_SYNCPT) + HOST1X_UCLASS_WAIT_SYNCPT) host1x_debug_output(o, "waiting on syncpt %d val %d\n", cbread >> 24, cbread & 0xffffff); else if (HOST1X_SYNC_CBSTAT_CBCLASS_V(cbstat) == - HOST1X_CLASS_HOST1X && - HOST1X_SYNC_CBSTAT_CBOFFSET_V(cbstat) == - HOST1X_UCLASS_WAIT_SYNCPT_BASE) { - + HOST1X_CLASS_HOST1X && + HOST1X_SYNC_CBSTAT_CBOFFSET_V(cbstat) == + HOST1X_UCLASS_WAIT_SYNCPT_BASE) { base = (cbread >> 16) & 0xff; baseval = host1x_sync_readl(host, HOST1X_SYNC_SYNCPT_BASE(base)); @@ -236,7 +234,7 @@ static void host1x_debug_show_channel_fifo(struct host1x *host, u32 val, rd_ptr, wr_ptr, start, end; unsigned int data_count = 0; - host1x_debug_output(o, "%d: fifo:\n", ch->id); + host1x_debug_output(o, "%u: fifo:\n", ch->id); val = host1x_ch_readl(ch, HOST1X_CHANNEL_FIFOSTAT); host1x_debug_output(o, "FIFOSTAT %08x\n", val); @@ -290,20 +288,22 @@ static void host1x_debug_show_channel_fifo(struct host1x *host, static void host1x_debug_show_mlocks(struct host1x *host, struct output *o) { - int i; + unsigned int i; host1x_debug_output(o, "---- mlocks ----\n"); + for (i = 0; i < host1x_syncpt_nb_mlocks(host); i++) { u32 owner = host1x_sync_readl(host, HOST1X_SYNC_MLOCK_OWNER(i)); if (HOST1X_SYNC_MLOCK_OWNER_CH_OWNS_V(owner)) - host1x_debug_output(o, "%d: locked by channel %d\n", + host1x_debug_output(o, "%u: locked by channel %u\n", i, HOST1X_SYNC_MLOCK_OWNER_CHID_V(owner)); else if (HOST1X_SYNC_MLOCK_OWNER_CPU_OWNS_V(owner)) - host1x_debug_output(o, "%d: locked by cpu\n", i); + host1x_debug_output(o, "%u: locked by cpu\n", i); else - host1x_debug_output(o, "%d: unlocked\n", i); + host1x_debug_output(o, "%u: unlocked\n", i); } + host1x_debug_output(o, "\n"); } diff --git a/drivers/gpu/host1x/hw/intr_hw.c b/drivers/gpu/host1x/hw/intr_hw.c index e1e31e9e67cd..dacb8009a605 100644 --- a/drivers/gpu/host1x/hw/intr_hw.c +++ b/drivers/gpu/host1x/hw/intr_hw.c @@ -38,14 +38,14 @@ static void host1x_intr_syncpt_handle(struct host1x_syncpt *syncpt) host1x_sync_writel(host, BIT_MASK(id), HOST1X_SYNC_SYNCPT_THRESH_CPU0_INT_STATUS(BIT_WORD(id))); - queue_work(host->intr_wq, &syncpt->intr.work); + schedule_work(&syncpt->intr.work); } static irqreturn_t syncpt_thresh_isr(int irq, void *dev_id) { struct host1x *host = dev_id; unsigned long reg; - int i, id; + unsigned int i, id; for (i = 0; i < DIV_ROUND_UP(host->info->nb_pts, 32); i++) { reg = host1x_sync_readl(host, @@ -62,7 +62,7 @@ static irqreturn_t syncpt_thresh_isr(int irq, void *dev_id) static void _host1x_intr_disable_all_syncpt_intrs(struct host1x *host) { - u32 i; + unsigned int i; for (i = 0; i < DIV_ROUND_UP(host->info->nb_pts, 32); ++i) { host1x_sync_writel(host, 0xffffffffu, @@ -72,10 +72,12 @@ static void _host1x_intr_disable_all_syncpt_intrs(struct host1x *host) } } -static int _host1x_intr_init_host_sync(struct host1x *host, u32 cpm, - void (*syncpt_thresh_work)(struct work_struct *)) +static int +_host1x_intr_init_host_sync(struct host1x *host, u32 cpm, + void (*syncpt_thresh_work)(struct work_struct *)) { - int i, err; + unsigned int i; + int err; host1x_hw_intr_disable_all_syncpt_intrs(host); @@ -106,18 +108,21 @@ static int _host1x_intr_init_host_sync(struct host1x *host, u32 cpm, } static void _host1x_intr_set_syncpt_threshold(struct host1x *host, - u32 id, u32 thresh) + unsigned int id, + u32 thresh) { host1x_sync_writel(host, thresh, HOST1X_SYNC_SYNCPT_INT_THRESH(id)); } -static void _host1x_intr_enable_syncpt_intr(struct host1x *host, u32 id) +static void _host1x_intr_enable_syncpt_intr(struct host1x *host, + unsigned int id) { host1x_sync_writel(host, BIT_MASK(id), HOST1X_SYNC_SYNCPT_THRESH_INT_ENABLE_CPU0(BIT_WORD(id))); } -static void _host1x_intr_disable_syncpt_intr(struct host1x *host, u32 id) +static void _host1x_intr_disable_syncpt_intr(struct host1x *host, + unsigned int id) { host1x_sync_writel(host, BIT_MASK(id), HOST1X_SYNC_SYNCPT_THRESH_INT_DISABLE(BIT_WORD(id))); @@ -127,8 +132,13 @@ static void _host1x_intr_disable_syncpt_intr(struct host1x *host, u32 id) static int _host1x_free_syncpt_irq(struct host1x *host) { + unsigned int i; + devm_free_irq(host->dev, host->intr_syncpt_irq, host); - flush_workqueue(host->intr_wq); + + for (i = 0; i < host->info->nb_pts; i++) + cancel_work_sync(&host->syncpt[i].intr.work); + return 0; } diff --git a/drivers/gpu/host1x/hw/syncpt_hw.c b/drivers/gpu/host1x/hw/syncpt_hw.c index 56e85395ac24..c93f74fcce72 100644 --- a/drivers/gpu/host1x/hw/syncpt_hw.c +++ b/drivers/gpu/host1x/hw/syncpt_hw.c @@ -26,8 +26,9 @@ */ static void syncpt_restore(struct host1x_syncpt *sp) { + u32 min = host1x_syncpt_read_min(sp); struct host1x *host = sp->host; - int min = host1x_syncpt_read_min(sp); + host1x_sync_writel(host, min, HOST1X_SYNC_SYNCPT(sp->id)); } @@ -37,6 +38,7 @@ static void syncpt_restore(struct host1x_syncpt *sp) static void syncpt_restore_wait_base(struct host1x_syncpt *sp) { struct host1x *host = sp->host; + host1x_sync_writel(host, sp->base_val, HOST1X_SYNC_SYNCPT_BASE(sp->id)); } @@ -47,6 +49,7 @@ static void syncpt_restore_wait_base(struct host1x_syncpt *sp) static void syncpt_read_wait_base(struct host1x_syncpt *sp) { struct host1x *host = sp->host; + sp->base_val = host1x_sync_readl(host, HOST1X_SYNC_SYNCPT_BASE(sp->id)); } @@ -85,6 +88,7 @@ static int syncpt_cpu_incr(struct host1x_syncpt *sp) if (!host1x_syncpt_client_managed(sp) && host1x_syncpt_idle(sp)) return -EINVAL; + host1x_sync_writel(host, BIT_MASK(sp->id), HOST1X_SYNC_SYNCPT_CPU_INCR(reg_offset)); wmb(); @@ -95,10 +99,10 @@ static int syncpt_cpu_incr(struct host1x_syncpt *sp) /* remove a wait pointed to by patch_addr */ static int syncpt_patch_wait(struct host1x_syncpt *sp, void *patch_addr) { - u32 override = host1x_class_host_wait_syncpt( - HOST1X_SYNCPT_RESERVED, 0); + u32 override = host1x_class_host_wait_syncpt(HOST1X_SYNCPT_RESERVED, 0); *((u32 *)patch_addr) = override; + return 0; } diff --git a/drivers/gpu/host1x/intr.c b/drivers/gpu/host1x/intr.c index 2491bf82e30c..8b4fad0ab35d 100644 --- a/drivers/gpu/host1x/intr.c +++ b/drivers/gpu/host1x/intr.c @@ -122,18 +122,20 @@ static void action_submit_complete(struct host1x_waitlist *waiter) static void action_wakeup(struct host1x_waitlist *waiter) { wait_queue_head_t *wq = waiter->data; + wake_up(wq); } static void action_wakeup_interruptible(struct host1x_waitlist *waiter) { wait_queue_head_t *wq = waiter->data; + wake_up_interruptible(wq); } typedef void (*action_handler)(struct host1x_waitlist *waiter); -static action_handler action_handlers[HOST1X_INTR_ACTION_COUNT] = { +static const action_handler action_handlers[HOST1X_INTR_ACTION_COUNT] = { action_submit_complete, action_wakeup, action_wakeup_interruptible, @@ -209,7 +211,7 @@ static void syncpt_thresh_work(struct work_struct *work) host1x_syncpt_load(host->syncpt + id)); } -int host1x_intr_add_action(struct host1x *host, u32 id, u32 thresh, +int host1x_intr_add_action(struct host1x *host, unsigned int id, u32 thresh, enum host1x_intr_action action, void *data, struct host1x_waitlist *waiter, void **ref) { @@ -254,7 +256,7 @@ int host1x_intr_add_action(struct host1x *host, u32 id, u32 thresh, return 0; } -void host1x_intr_put_ref(struct host1x *host, u32 id, void *ref) +void host1x_intr_put_ref(struct host1x *host, unsigned int id, void *ref) { struct host1x_waitlist *waiter = ref; struct host1x_syncpt *syncpt; @@ -277,9 +279,6 @@ int host1x_intr_init(struct host1x *host, unsigned int irq_sync) mutex_init(&host->intr_mutex); host->intr_syncpt_irq = irq_sync; - host->intr_wq = create_workqueue("host_syncpt"); - if (!host->intr_wq) - return -ENOMEM; for (id = 0; id < nb_pts; ++id) { struct host1x_syncpt *syncpt = host->syncpt + id; @@ -288,7 +287,7 @@ int host1x_intr_init(struct host1x *host, unsigned int irq_sync) INIT_LIST_HEAD(&syncpt->intr.wait_head); snprintf(syncpt->intr.thresh_irq_name, sizeof(syncpt->intr.thresh_irq_name), - "host1x_sp_%02d", id); + "host1x_sp_%02u", id); } host1x_intr_start(host); @@ -299,7 +298,6 @@ int host1x_intr_init(struct host1x *host, unsigned int irq_sync) void host1x_intr_deinit(struct host1x *host) { host1x_intr_stop(host); - destroy_workqueue(host->intr_wq); } void host1x_intr_start(struct host1x *host) @@ -342,7 +340,7 @@ void host1x_intr_stop(struct host1x *host) if (!list_empty(&syncpt[id].intr.wait_head)) { /* output diagnostics */ mutex_unlock(&host->intr_mutex); - pr_warn("%s cannot stop syncpt intr id=%d\n", + pr_warn("%s cannot stop syncpt intr id=%u\n", __func__, id); return; } diff --git a/drivers/gpu/host1x/intr.h b/drivers/gpu/host1x/intr.h index 2b8adf016a05..1370c2bb75b8 100644 --- a/drivers/gpu/host1x/intr.h +++ b/drivers/gpu/host1x/intr.h @@ -75,7 +75,7 @@ struct host1x_waitlist { * * This is a non-blocking api. */ -int host1x_intr_add_action(struct host1x *host, u32 id, u32 thresh, +int host1x_intr_add_action(struct host1x *host, unsigned int id, u32 thresh, enum host1x_intr_action action, void *data, struct host1x_waitlist *waiter, void **ref); @@ -84,7 +84,7 @@ int host1x_intr_add_action(struct host1x *host, u32 id, u32 thresh, * You must call this if you passed non-NULL as ref. * @ref the ref returned from host1x_intr_add_action() */ -void host1x_intr_put_ref(struct host1x *host, u32 id, void *ref); +void host1x_intr_put_ref(struct host1x *host, unsigned int id, void *ref); /* Initialize host1x sync point interrupt */ int host1x_intr_init(struct host1x *host, unsigned int irq_sync); diff --git a/drivers/gpu/host1x/job.c b/drivers/gpu/host1x/job.c index b4515d544039..a91b7c4a6110 100644 --- a/drivers/gpu/host1x/job.c +++ b/drivers/gpu/host1x/job.c @@ -161,7 +161,7 @@ static int do_waitchks(struct host1x_job *job, struct host1x *host, if (host1x_syncpt_is_expired(sp, wait->thresh)) { dev_dbg(host->dev, - "drop WAIT id %d (%s) thresh 0x%x, min 0x%x\n", + "drop WAIT id %u (%s) thresh 0x%x, min 0x%x\n", wait->syncpt_id, sp->name, wait->thresh, host1x_syncpt_read_min(sp)); @@ -464,6 +464,7 @@ static inline int copy_gathers(struct host1x_job *job, struct device *dev) for (i = 0; i < job->num_gathers; i++) { struct host1x_job_gather *g = &job->gathers[i]; + size += g->words * sizeof(u32); } @@ -514,6 +515,7 @@ int host1x_job_pin(struct host1x_job *job, struct device *dev) bitmap_zero(waitchk_mask, host1x_syncpt_nb_pts(host)); for (i = 0; i < job->num_waitchk; i++) { u32 syncpt_id = job->waitchk[i].syncpt_id; + if (syncpt_id < host1x_syncpt_nb_pts(host)) set_bit(syncpt_id, waitchk_mask); } @@ -571,14 +573,16 @@ void host1x_job_unpin(struct host1x_job *job) for (i = 0; i < job->num_unpins; i++) { struct host1x_job_unpin_data *unpin = &job->unpins[i]; + host1x_bo_unpin(unpin->bo, unpin->sgt); host1x_bo_put(unpin->bo); } + job->num_unpins = 0; if (job->gather_copy_size) dma_free_wc(job->channel->dev, job->gather_copy_size, - job->gather_copy_mapped, job->gather_copy); + job->gather_copy_mapped, job->gather_copy); } EXPORT_SYMBOL(host1x_job_unpin); diff --git a/drivers/gpu/host1x/syncpt.c b/drivers/gpu/host1x/syncpt.c index 6b7fdc1e2ed0..95589328ad52 100644 --- a/drivers/gpu/host1x/syncpt.c +++ b/drivers/gpu/host1x/syncpt.c @@ -73,7 +73,7 @@ static struct host1x_syncpt *host1x_syncpt_alloc(struct host1x *host, return NULL; } - name = kasprintf(GFP_KERNEL, "%02d-%s", sp->id, + name = kasprintf(GFP_KERNEL, "%02u-%s", sp->id, dev ? dev_name(dev) : NULL); if (!name) return NULL; @@ -110,12 +110,14 @@ EXPORT_SYMBOL(host1x_syncpt_incr_max); void host1x_syncpt_restore(struct host1x *host) { struct host1x_syncpt *sp_base = host->syncpt; - u32 i; + unsigned int i; for (i = 0; i < host1x_syncpt_nb_pts(host); i++) host1x_hw_syncpt_restore(host, sp_base + i); + for (i = 0; i < host1x_syncpt_nb_bases(host); i++) host1x_hw_syncpt_restore_wait_base(host, sp_base + i); + wmb(); } @@ -126,7 +128,7 @@ void host1x_syncpt_restore(struct host1x *host) void host1x_syncpt_save(struct host1x *host) { struct host1x_syncpt *sp_base = host->syncpt; - u32 i; + unsigned int i; for (i = 0; i < host1x_syncpt_nb_pts(host); i++) { if (host1x_syncpt_client_managed(sp_base + i)) @@ -146,6 +148,7 @@ void host1x_syncpt_save(struct host1x *host) u32 host1x_syncpt_load(struct host1x_syncpt *sp) { u32 val; + val = host1x_hw_syncpt_load(sp->host, sp); trace_host1x_syncpt_load_min(sp->id, val); @@ -157,10 +160,9 @@ u32 host1x_syncpt_load(struct host1x_syncpt *sp) */ u32 host1x_syncpt_load_wait_base(struct host1x_syncpt *sp) { - u32 val; host1x_hw_syncpt_load_wait_base(sp->host, sp); - val = sp->base_val; - return val; + + return sp->base_val; } /* @@ -179,6 +181,7 @@ EXPORT_SYMBOL(host1x_syncpt_incr); static bool syncpt_load_min_is_expired(struct host1x_syncpt *sp, u32 thresh) { host1x_hw_syncpt_load(sp->host, sp); + return host1x_syncpt_is_expired(sp, thresh); } @@ -186,7 +189,7 @@ static bool syncpt_load_min_is_expired(struct host1x_syncpt *sp, u32 thresh) * Main entrypoint for syncpoint value waits. */ int host1x_syncpt_wait(struct host1x_syncpt *sp, u32 thresh, long timeout, - u32 *value) + u32 *value) { DECLARE_WAIT_QUEUE_HEAD_ONSTACK(wq); void *ref; @@ -201,6 +204,7 @@ int host1x_syncpt_wait(struct host1x_syncpt *sp, u32 thresh, long timeout, if (host1x_syncpt_is_expired(sp, thresh)) { if (value) *value = host1x_syncpt_load(sp); + return 0; } @@ -209,6 +213,7 @@ int host1x_syncpt_wait(struct host1x_syncpt *sp, u32 thresh, long timeout, if (host1x_syncpt_is_expired(sp, thresh)) { if (value) *value = val; + goto done; } @@ -239,32 +244,42 @@ int host1x_syncpt_wait(struct host1x_syncpt *sp, u32 thresh, long timeout, /* wait for the syncpoint, or timeout, or signal */ while (timeout) { long check = min_t(long, SYNCPT_CHECK_PERIOD, timeout); - int remain = wait_event_interruptible_timeout(wq, + int remain; + + remain = wait_event_interruptible_timeout(wq, syncpt_load_min_is_expired(sp, thresh), check); if (remain > 0 || host1x_syncpt_is_expired(sp, thresh)) { if (value) *value = host1x_syncpt_load(sp); + err = 0; + break; } + if (remain < 0) { err = remain; break; } + timeout -= check; + if (timeout && check_count <= MAX_STUCK_CHECK_COUNT) { dev_warn(sp->host->dev, - "%s: syncpoint id %d (%s) stuck waiting %d, timeout=%ld\n", + "%s: syncpoint id %u (%s) stuck waiting %d, timeout=%ld\n", current->comm, sp->id, sp->name, thresh, timeout); host1x_debug_dump_syncpts(sp->host); + if (check_count == MAX_STUCK_CHECK_COUNT) host1x_debug_dump(sp->host); + check_count++; } } + host1x_intr_put_ref(sp->host, sp->id, ref); done: @@ -279,7 +294,9 @@ bool host1x_syncpt_is_expired(struct host1x_syncpt *sp, u32 thresh) { u32 current_val; u32 future_val; + smp_rmb(); + current_val = (u32)atomic_read(&sp->min_val); future_val = (u32)atomic_read(&sp->max_val); @@ -341,14 +358,14 @@ int host1x_syncpt_init(struct host1x *host) { struct host1x_syncpt_base *bases; struct host1x_syncpt *syncpt; - int i; + unsigned int i; - syncpt = devm_kzalloc(host->dev, sizeof(*syncpt) * host->info->nb_pts, + syncpt = devm_kcalloc(host->dev, host->info->nb_pts, sizeof(*syncpt), GFP_KERNEL); if (!syncpt) return -ENOMEM; - bases = devm_kzalloc(host->dev, sizeof(*bases) * host->info->nb_bases, + bases = devm_kcalloc(host->dev, host->info->nb_bases, sizeof(*bases), GFP_KERNEL); if (!bases) return -ENOMEM; @@ -378,6 +395,7 @@ struct host1x_syncpt *host1x_syncpt_request(struct device *dev, unsigned long flags) { struct host1x *host = dev_get_drvdata(dev->parent); + return host1x_syncpt_alloc(host, dev, flags); } EXPORT_SYMBOL(host1x_syncpt_request); @@ -398,8 +416,9 @@ EXPORT_SYMBOL(host1x_syncpt_free); void host1x_syncpt_deinit(struct host1x *host) { - int i; struct host1x_syncpt *sp = host->syncpt; + unsigned int i; + for (i = 0; i < host->info->nb_pts; i++, sp++) kfree(sp->name); } @@ -407,10 +426,11 @@ void host1x_syncpt_deinit(struct host1x *host) /* * Read max. It indicates how many operations there are in queue, either in * channel or in a software thread. - * */ + */ u32 host1x_syncpt_read_max(struct host1x_syncpt *sp) { smp_rmb(); + return (u32)atomic_read(&sp->max_val); } EXPORT_SYMBOL(host1x_syncpt_read_max); @@ -421,6 +441,7 @@ EXPORT_SYMBOL(host1x_syncpt_read_max); u32 host1x_syncpt_read_min(struct host1x_syncpt *sp) { smp_rmb(); + return (u32)atomic_read(&sp->min_val); } EXPORT_SYMBOL(host1x_syncpt_read_min); @@ -431,25 +452,26 @@ u32 host1x_syncpt_read(struct host1x_syncpt *sp) } EXPORT_SYMBOL(host1x_syncpt_read); -int host1x_syncpt_nb_pts(struct host1x *host) +unsigned int host1x_syncpt_nb_pts(struct host1x *host) { return host->info->nb_pts; } -int host1x_syncpt_nb_bases(struct host1x *host) +unsigned int host1x_syncpt_nb_bases(struct host1x *host) { return host->info->nb_bases; } -int host1x_syncpt_nb_mlocks(struct host1x *host) +unsigned int host1x_syncpt_nb_mlocks(struct host1x *host) { return host->info->nb_mlocks; } -struct host1x_syncpt *host1x_syncpt_get(struct host1x *host, u32 id) +struct host1x_syncpt *host1x_syncpt_get(struct host1x *host, unsigned int id) { if (host->info->nb_pts < id) return NULL; + return host->syncpt + id; } EXPORT_SYMBOL(host1x_syncpt_get); diff --git a/drivers/gpu/host1x/syncpt.h b/drivers/gpu/host1x/syncpt.h index 9056465ecd3f..f719205105ac 100644 --- a/drivers/gpu/host1x/syncpt.h +++ b/drivers/gpu/host1x/syncpt.h @@ -37,7 +37,7 @@ struct host1x_syncpt_base { }; struct host1x_syncpt { - int id; + unsigned int id; atomic_t min_val; atomic_t max_val; u32 base_val; @@ -58,13 +58,13 @@ int host1x_syncpt_init(struct host1x *host); void host1x_syncpt_deinit(struct host1x *host); /* Return number of sync point supported. */ -int host1x_syncpt_nb_pts(struct host1x *host); +unsigned int host1x_syncpt_nb_pts(struct host1x *host); /* Return number of wait bases supported. */ -int host1x_syncpt_nb_bases(struct host1x *host); +unsigned int host1x_syncpt_nb_bases(struct host1x *host); /* Return number of mlocks supported. */ -int host1x_syncpt_nb_mlocks(struct host1x *host); +unsigned int host1x_syncpt_nb_mlocks(struct host1x *host); /* * Check sync point sanity. If max is larger than min, there have too many diff --git a/drivers/gpu/ipu-v3/ipu-dc.c b/drivers/gpu/ipu-v3/ipu-dc.c index 2f29780e7c68..659475c1e44a 100644 --- a/drivers/gpu/ipu-v3/ipu-dc.c +++ b/drivers/gpu/ipu-v3/ipu-dc.c @@ -150,6 +150,9 @@ static void dc_write_tmpl(struct ipu_dc *dc, int word, u32 opcode, u32 operand, static int ipu_bus_format_to_map(u32 fmt) { switch (fmt) { + default: + WARN_ON(1); + /* fall-through */ case MEDIA_BUS_FMT_RGB888_1X24: return IPU_DC_MAP_RGB24; case MEDIA_BUS_FMT_RGB565_1X16: @@ -162,8 +165,6 @@ static int ipu_bus_format_to_map(u32 fmt) return IPU_DC_MAP_LVDS666; case MEDIA_BUS_FMT_BGR888_1X24: return IPU_DC_MAP_BGR24; - default: - return -EINVAL; } } @@ -178,10 +179,6 @@ int ipu_dc_init_sync(struct ipu_dc *dc, struct ipu_di *di, bool interlaced, dc->di = ipu_di_get_num(di); map = ipu_bus_format_to_map(bus_format); - if (map < 0) { - dev_dbg(priv->dev, "IPU_DISP: No MAP\n"); - return map; - } /* * In interlaced mode we need more counters to create the asymmetric diff --git a/drivers/gpu/ipu-v3/ipu-di.c b/drivers/gpu/ipu-v3/ipu-di.c index 359268e3a166..a8d87ddd8a17 100644 --- a/drivers/gpu/ipu-v3/ipu-di.c +++ b/drivers/gpu/ipu-v3/ipu-di.c @@ -572,9 +572,6 @@ int ipu_di_init_sync_panel(struct ipu_di *di, struct ipu_di_signal_cfg *sig) dev_dbg(di->ipu->dev, "disp %d: panel size = %d x %d\n", di->id, sig->mode.hactive, sig->mode.vactive); - if ((sig->mode.vsync_len == 0) || (sig->mode.hsync_len == 0)) - return -EINVAL; - dev_dbg(di->ipu->dev, "Clocks: IPU %luHz DI %luHz Needed %luHz\n", clk_get_rate(di->clk_ipu), clk_get_rate(di->clk_di), diff --git a/drivers/gpu/ipu-v3/ipu-dmfc.c b/drivers/gpu/ipu-v3/ipu-dmfc.c index 837b1ec22800..42705bb5aaa3 100644 --- a/drivers/gpu/ipu-v3/ipu-dmfc.c +++ b/drivers/gpu/ipu-v3/ipu-dmfc.c @@ -45,17 +45,6 @@ #define DMFC_DP_CHAN_6B_24 16 #define DMFC_DP_CHAN_6F_29 24 -#define DMFC_FIFO_SIZE_64 (3 << 3) -#define DMFC_FIFO_SIZE_128 (2 << 3) -#define DMFC_FIFO_SIZE_256 (1 << 3) -#define DMFC_FIFO_SIZE_512 (0 << 3) - -#define DMFC_SEGMENT(x) ((x & 0x7) << 0) -#define DMFC_BURSTSIZE_128 (0 << 6) -#define DMFC_BURSTSIZE_64 (1 << 6) -#define DMFC_BURSTSIZE_32 (2 << 6) -#define DMFC_BURSTSIZE_16 (3 << 6) - struct dmfc_channel_data { int ipu_channel; unsigned long channel_reg; @@ -104,9 +93,6 @@ struct ipu_dmfc_priv; struct dmfc_channel { unsigned slots; - unsigned slotmask; - unsigned segment; - int burstsize; struct ipu_soc *ipu; struct ipu_dmfc_priv *priv; const struct dmfc_channel_data *data; @@ -117,7 +103,6 @@ struct ipu_dmfc_priv { struct device *dev; struct dmfc_channel channels[DMFC_NUM_CHANNELS]; struct mutex mutex; - unsigned long bandwidth_per_slot; void __iomem *base; int use_count; }; @@ -172,184 +157,6 @@ void ipu_dmfc_disable_channel(struct dmfc_channel *dmfc) } EXPORT_SYMBOL_GPL(ipu_dmfc_disable_channel); -static int ipu_dmfc_setup_channel(struct dmfc_channel *dmfc, int slots, - int segment, int burstsize) -{ - struct ipu_dmfc_priv *priv = dmfc->priv; - u32 val, field; - - dev_dbg(priv->dev, - "dmfc: using %d slots starting from segment %d for IPU channel %d\n", - slots, segment, dmfc->data->ipu_channel); - - switch (slots) { - case 1: - field = DMFC_FIFO_SIZE_64; - break; - case 2: - field = DMFC_FIFO_SIZE_128; - break; - case 4: - field = DMFC_FIFO_SIZE_256; - break; - case 8: - field = DMFC_FIFO_SIZE_512; - break; - default: - return -EINVAL; - } - - switch (burstsize) { - case 16: - field |= DMFC_BURSTSIZE_16; - break; - case 32: - field |= DMFC_BURSTSIZE_32; - break; - case 64: - field |= DMFC_BURSTSIZE_64; - break; - case 128: - field |= DMFC_BURSTSIZE_128; - break; - } - - field |= DMFC_SEGMENT(segment); - - val = readl(priv->base + dmfc->data->channel_reg); - - val &= ~(0xff << dmfc->data->shift); - val |= field << dmfc->data->shift; - - writel(val, priv->base + dmfc->data->channel_reg); - - dmfc->slots = slots; - dmfc->segment = segment; - dmfc->burstsize = burstsize; - dmfc->slotmask = ((1 << slots) - 1) << segment; - - return 0; -} - -static int dmfc_bandwidth_to_slots(struct ipu_dmfc_priv *priv, - unsigned long bandwidth) -{ - int slots = 1; - - while (slots * priv->bandwidth_per_slot < bandwidth) - slots *= 2; - - return slots; -} - -static int dmfc_find_slots(struct ipu_dmfc_priv *priv, int slots) -{ - unsigned slotmask_need, slotmask_used = 0; - int i, segment = 0; - - slotmask_need = (1 << slots) - 1; - - for (i = 0; i < DMFC_NUM_CHANNELS; i++) - slotmask_used |= priv->channels[i].slotmask; - - while (slotmask_need <= 0xff) { - if (!(slotmask_used & slotmask_need)) - return segment; - - slotmask_need <<= 1; - segment++; - } - - return -EBUSY; -} - -void ipu_dmfc_free_bandwidth(struct dmfc_channel *dmfc) -{ - struct ipu_dmfc_priv *priv = dmfc->priv; - int i; - - dev_dbg(priv->dev, "dmfc: freeing %d slots starting from segment %d\n", - dmfc->slots, dmfc->segment); - - mutex_lock(&priv->mutex); - - if (!dmfc->slots) - goto out; - - dmfc->slotmask = 0; - dmfc->slots = 0; - dmfc->segment = 0; - - for (i = 0; i < DMFC_NUM_CHANNELS; i++) - priv->channels[i].slotmask = 0; - - for (i = 0; i < DMFC_NUM_CHANNELS; i++) { - if (priv->channels[i].slots > 0) { - priv->channels[i].segment = - dmfc_find_slots(priv, priv->channels[i].slots); - priv->channels[i].slotmask = - ((1 << priv->channels[i].slots) - 1) << - priv->channels[i].segment; - } - } - - for (i = 0; i < DMFC_NUM_CHANNELS; i++) { - if (priv->channels[i].slots > 0) - ipu_dmfc_setup_channel(&priv->channels[i], - priv->channels[i].slots, - priv->channels[i].segment, - priv->channels[i].burstsize); - } -out: - mutex_unlock(&priv->mutex); -} -EXPORT_SYMBOL_GPL(ipu_dmfc_free_bandwidth); - -int ipu_dmfc_alloc_bandwidth(struct dmfc_channel *dmfc, - unsigned long bandwidth_pixel_per_second, int burstsize) -{ - struct ipu_dmfc_priv *priv = dmfc->priv; - int slots = dmfc_bandwidth_to_slots(priv, bandwidth_pixel_per_second); - int segment = -1, ret = 0; - - dev_dbg(priv->dev, "dmfc: trying to allocate %ldMpixel/s for IPU channel %d\n", - bandwidth_pixel_per_second / 1000000, - dmfc->data->ipu_channel); - - ipu_dmfc_free_bandwidth(dmfc); - - mutex_lock(&priv->mutex); - - if (slots > 8) { - ret = -EBUSY; - goto out; - } - - /* For the MEM_BG channel, first try to allocate twice the slots */ - if (dmfc->data->ipu_channel == IPUV3_CHANNEL_MEM_BG_SYNC) - segment = dmfc_find_slots(priv, slots * 2); - else if (slots < 2) - /* Always allocate at least 128*4 bytes (2 slots) */ - slots = 2; - - if (segment >= 0) - slots *= 2; - else - segment = dmfc_find_slots(priv, slots); - if (segment < 0) { - ret = -EBUSY; - goto out; - } - - ipu_dmfc_setup_channel(dmfc, slots, segment, burstsize); - -out: - mutex_unlock(&priv->mutex); - - return ret; -} -EXPORT_SYMBOL_GPL(ipu_dmfc_alloc_bandwidth); - void ipu_dmfc_config_wait4eot(struct dmfc_channel *dmfc, int width) { struct ipu_dmfc_priv *priv = dmfc->priv; @@ -384,7 +191,6 @@ EXPORT_SYMBOL_GPL(ipu_dmfc_get); void ipu_dmfc_put(struct dmfc_channel *dmfc) { - ipu_dmfc_free_bandwidth(dmfc); } EXPORT_SYMBOL_GPL(ipu_dmfc_put); @@ -412,20 +218,15 @@ int ipu_dmfc_init(struct ipu_soc *ipu, struct device *dev, unsigned long base, priv->channels[i].priv = priv; priv->channels[i].ipu = ipu; priv->channels[i].data = &dmfcdata[i]; - } - - writel(0x0, priv->base + DMFC_WR_CHAN); - writel(0x0, priv->base + DMFC_DP_CHAN); - /* - * We have a total bandwidth of clkrate * 4pixel divided - * into 8 slots. - */ - priv->bandwidth_per_slot = clk_get_rate(ipu_clk) * 4 / 8; - - dev_dbg(dev, "dmfc: 8 slots with %ldMpixel/s bandwidth each\n", - priv->bandwidth_per_slot / 1000000); + if (dmfcdata[i].ipu_channel == IPUV3_CHANNEL_MEM_BG_SYNC || + dmfcdata[i].ipu_channel == IPUV3_CHANNEL_MEM_FG_SYNC || + dmfcdata[i].ipu_channel == IPUV3_CHANNEL_MEM_DC_SYNC) + priv->channels[i].slots = 2; + } + writel(0x00000050, priv->base + DMFC_WR_CHAN); + writel(0x00005654, priv->base + DMFC_DP_CHAN); writel(0x202020f6, priv->base + DMFC_WR_CHAN_DEF); writel(0x2020f6f6, priv->base + DMFC_DP_CHAN_DEF); writel(0x00000003, priv->base + DMFC_GENERAL1); diff --git a/drivers/gpu/vga/vga_switcheroo.c b/drivers/gpu/vga/vga_switcheroo.c index 2df216b39cc5..5f962bfcb43c 100644 --- a/drivers/gpu/vga/vga_switcheroo.c +++ b/drivers/gpu/vga/vga_switcheroo.c @@ -52,9 +52,9 @@ * * * muxed: Dual GPUs with a multiplexer chip to switch outputs between GPUs. * * muxless: Dual GPUs but only one of them is connected to outputs. - * The other one is merely used to offload rendering, its results - * are copied over PCIe into the framebuffer. On Linux this is - * supported with DRI PRIME. + * The other one is merely used to offload rendering, its results + * are copied over PCIe into the framebuffer. On Linux this is + * supported with DRI PRIME. * * Hybrid graphics started to appear in the late Naughties and were initially * all muxed. Newer laptops moved to a muxless architecture for cost reasons. @@ -560,21 +560,21 @@ EXPORT_SYMBOL(vga_switcheroo_unlock_ddc); * * OFF: Power off the device not in use. * * ON: Power on the device not in use. * * IGD: Switch to the integrated graphics device. - * Power on the integrated GPU if necessary, power off the discrete GPU. - * Prerequisite is that no user space processes (e.g. Xorg, alsactl) - * have opened device files of the GPUs or the audio client. If the - * switch fails, the user may invoke lsof(8) or fuser(1) on /dev/dri/ - * and /dev/snd/controlC1 to identify processes blocking the switch. + * Power on the integrated GPU if necessary, power off the discrete GPU. + * Prerequisite is that no user space processes (e.g. Xorg, alsactl) + * have opened device files of the GPUs or the audio client. If the + * switch fails, the user may invoke lsof(8) or fuser(1) on /dev/dri/ + * and /dev/snd/controlC1 to identify processes blocking the switch. * * DIS: Switch to the discrete graphics device. * * DIGD: Delayed switch to the integrated graphics device. - * This will perform the switch once the last user space process has - * closed the device files of the GPUs and the audio client. + * This will perform the switch once the last user space process has + * closed the device files of the GPUs and the audio client. * * DDIS: Delayed switch to the discrete graphics device. * * MIGD: Mux-only switch to the integrated graphics device. - * Does not remap console or change the power state of either gpu. - * If the integrated GPU is currently off, the screen will turn black. - * If it is on, the screen will show whatever happens to be in VRAM. - * Either way, the user has to blindly enter the command to switch back. + * Does not remap console or change the power state of either gpu. + * If the integrated GPU is currently off, the screen will turn black. + * If it is on, the screen will show whatever happens to be in VRAM. + * Either way, the user has to blindly enter the command to switch back. * * MDIS: Mux-only switch to the discrete graphics device. * * For GPUs whose power state is controlled by the driver's runtime pm, |