summaryrefslogtreecommitdiff
path: root/drivers/gpu/drm/amd
diff options
context:
space:
mode:
Diffstat (limited to 'drivers/gpu/drm/amd')
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c1
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_device.c48
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c5
-rw-r--r--drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c12
-rw-r--r--drivers/gpu/drm/amd/amdgpu/gfx_v9_0_cleaner_shader.h44
-rw-r--r--drivers/gpu/drm/amd/amdgpu/gfx_v9_4_2_cleaner_shader.asm153
-rw-r--r--drivers/gpu/drm/amd/amdgpu/hdp_v4_0.c12
-rw-r--r--drivers/gpu/drm/amd/amdgpu/hdp_v5_0.c7
-rw-r--r--drivers/gpu/drm/amd/amdgpu/hdp_v5_2.c6
-rw-r--r--drivers/gpu/drm/amd/amdgpu/hdp_v6_0.c6
-rw-r--r--drivers/gpu/drm/amd/amdgpu/hdp_v7_0.c6
-rw-r--r--drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c30
-rw-r--r--drivers/gpu/drm/amd/amdgpu/vega20_ih.c27
-rw-r--r--drivers/gpu/drm/amd/amdkfd/kfd_crat.c6
-rw-r--r--drivers/gpu/drm/amd/amdkfd/kfd_device.c3
-rw-r--r--drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c25
-rw-r--r--drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c17
-rw-r--r--drivers/gpu/drm/amd/display/dc/core/dc.c18
-rw-r--r--drivers/gpu/drm/amd/display/dc/core/dc_debug.c40
-rw-r--r--drivers/gpu/drm/amd/display/dc/core/dc_resource.c57
-rw-r--r--drivers/gpu/drm/amd/display/dc/core/dc_stream.c6
-rw-r--r--drivers/gpu/drm/amd/display/dc/dc.h3
-rw-r--r--drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c6
-rw-r--r--drivers/gpu/drm/amd/display/dc/dio/dcn314/dcn314_dio_stream_encoder.c10
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/display_mode_core.c1
-rw-r--r--drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_translation_helper.c29
-rw-r--r--drivers/gpu/drm/amd/display/dc/hwss/dcn314/dcn314_hwseq.c20
-rw-r--r--drivers/gpu/drm/amd/display/dc/inc/core_status.h2
-rw-r--r--drivers/gpu/drm/amd/display/dc/inc/core_types.h1
-rw-r--r--drivers/gpu/drm/amd/display/dc/resource/dcn20/dcn20_resource.c23
-rw-r--r--drivers/gpu/drm/amd/display/dc/resource/dcn21/dcn21_resource.c2
-rw-r--r--drivers/gpu/drm/amd/display/dc/resource/dcn30/dcn30_resource.c1
-rw-r--r--drivers/gpu/drm/amd/display/dc/resource/dcn302/dcn302_resource.c1
-rw-r--r--drivers/gpu/drm/amd/display/dc/resource/dcn303/dcn303_resource.c1
-rw-r--r--drivers/gpu/drm/amd/display/dc/resource/dcn31/dcn31_resource.c7
-rw-r--r--drivers/gpu/drm/amd/display/dc/resource/dcn31/dcn31_resource.h3
-rw-r--r--drivers/gpu/drm/amd/display/dc/resource/dcn314/dcn314_resource.c1
-rw-r--r--drivers/gpu/drm/amd/display/dc/resource/dcn315/dcn315_resource.c1
-rw-r--r--drivers/gpu/drm/amd/display/dc/resource/dcn316/dcn316_resource.c1
-rw-r--r--drivers/gpu/drm/amd/display/dc/resource/dcn32/dcn32_resource.c1
-rw-r--r--drivers/gpu/drm/amd/display/dc/resource/dcn321/dcn321_resource.c1
-rw-r--r--drivers/gpu/drm/amd/display/dc/resource/dcn35/dcn35_resource.c2
-rw-r--r--drivers/gpu/drm/amd/display/dc/resource/dcn351/dcn351_resource.c2
-rw-r--r--drivers/gpu/drm/amd/display/dc/resource/dcn401/dcn401_resource.c1
-rw-r--r--drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h3
-rw-r--r--drivers/gpu/drm/amd/display/modules/freesync/freesync.c13
-rw-r--r--drivers/gpu/drm/amd/pm/amdgpu_pm.c6
-rw-r--r--drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c150
-rw-r--r--drivers/gpu/drm/amd/pm/swsmu/inc/amdgpu_smu.h15
-rw-r--r--drivers/gpu/drm/amd/pm/swsmu/smu11/arcturus_ppt.c166
-rw-r--r--drivers/gpu/drm/amd/pm/swsmu/smu11/navi10_ppt.c167
-rw-r--r--drivers/gpu/drm/amd/pm/swsmu/smu11/sienna_cichlid_ppt.c168
-rw-r--r--drivers/gpu/drm/amd/pm/swsmu/smu11/vangogh_ppt.c41
-rw-r--r--drivers/gpu/drm/amd/pm/swsmu/smu12/renoir_ppt.c43
-rw-r--r--drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c167
-rw-r--r--drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c138
-rw-r--r--drivers/gpu/drm/amd/pm/swsmu/smu14/smu_v14_0_2_ppt.c168
-rw-r--r--drivers/gpu/drm/amd/pm/swsmu/smu_cmn.c25
-rw-r--r--drivers/gpu/drm/amd/pm/swsmu/smu_cmn.h4
59 files changed, 1334 insertions, 589 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c
index 7dd55ed57c1d..b8d4e07d2043 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c
@@ -800,6 +800,7 @@ int amdgpu_acpi_power_shift_control(struct amdgpu_device *adev,
return -EIO;
}
+ kfree(info);
return 0;
}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index 1f08cb88d51b..51904906545e 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -3666,7 +3666,7 @@ static int amdgpu_device_ip_resume_phase1(struct amdgpu_device *adev)
*
* @adev: amdgpu_device pointer
*
- * First resume function for hardware IPs. The list of all the hardware
+ * Second resume function for hardware IPs. The list of all the hardware
* IPs that make up the asic is walked and the resume callbacks are run for
* all blocks except COMMON, GMC, and IH. resume puts the hardware into a
* functional state after a suspend and updates the software state as
@@ -3684,6 +3684,7 @@ static int amdgpu_device_ip_resume_phase2(struct amdgpu_device *adev)
if (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_COMMON ||
adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_GMC ||
adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_IH ||
+ adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_DCE ||
adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_PSP)
continue;
r = adev->ip_blocks[i].version->funcs->resume(adev);
@@ -3699,6 +3700,36 @@ static int amdgpu_device_ip_resume_phase2(struct amdgpu_device *adev)
}
/**
+ * amdgpu_device_ip_resume_phase3 - run resume for hardware IPs
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Third resume function for hardware IPs. The list of all the hardware
+ * IPs that make up the asic is walked and the resume callbacks are run for
+ * all DCE. resume puts the hardware into a functional state after a suspend
+ * and updates the software state as necessary. This function is also used
+ * for restoring the GPU after a GPU reset.
+ *
+ * Returns 0 on success, negative error code on failure.
+ */
+static int amdgpu_device_ip_resume_phase3(struct amdgpu_device *adev)
+{
+ int i, r;
+
+ for (i = 0; i < adev->num_ip_blocks; i++) {
+ if (!adev->ip_blocks[i].status.valid || adev->ip_blocks[i].status.hw)
+ continue;
+ if (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_DCE) {
+ r = adev->ip_blocks[i].version->funcs->resume(adev);
+ if (r)
+ return r;
+ }
+ }
+
+ return 0;
+}
+
+/**
* amdgpu_device_ip_resume - run resume for hardware IPs
*
* @adev: amdgpu_device pointer
@@ -3727,6 +3758,13 @@ static int amdgpu_device_ip_resume(struct amdgpu_device *adev)
if (adev->mman.buffer_funcs_ring->sched.ready)
amdgpu_ttm_set_buffer_funcs_status(adev, true);
+ if (r)
+ return r;
+
+ amdgpu_fence_driver_hw_init(adev);
+
+ r = amdgpu_device_ip_resume_phase3(adev);
+
return r;
}
@@ -4809,7 +4847,6 @@ int amdgpu_device_resume(struct drm_device *dev, bool fbcon)
dev_err(adev->dev, "amdgpu_device_ip_resume failed (%d).\n", r);
goto exit;
}
- amdgpu_fence_driver_hw_init(adev);
if (!adev->in_s0ix) {
r = amdgpu_amdkfd_resume(adev, adev->in_runpm);
@@ -5431,6 +5468,10 @@ int amdgpu_do_asic_reset(struct list_head *device_list_handle,
if (tmp_adev->mman.buffer_funcs_ring->sched.ready)
amdgpu_ttm_set_buffer_funcs_status(tmp_adev, true);
+ r = amdgpu_device_ip_resume_phase3(tmp_adev);
+ if (r)
+ goto out;
+
if (vram_lost)
amdgpu_device_fill_reset_magic(tmp_adev);
@@ -6344,6 +6385,9 @@ bool amdgpu_device_cache_pci_state(struct pci_dev *pdev)
struct amdgpu_device *adev = drm_to_adev(dev);
int r;
+ if (amdgpu_sriov_vf(adev))
+ return false;
+
r = pci_save_state(pdev);
if (!r) {
kfree(adev->pci_state);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index 74adb983ab03..9f922ec50ea2 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -812,7 +812,7 @@ static int amdgpu_ttm_tt_pin_userptr(struct ttm_device *bdev,
/* Map SG to device */
r = dma_map_sgtable(adev->dev, ttm->sg, direction, 0);
if (r)
- goto release_sg;
+ goto release_sg_table;
/* convert SG to linear array of pages and dma addresses */
drm_prime_sg_to_dma_addr_array(ttm->sg, gtt->ttm.dma_address,
@@ -820,6 +820,8 @@ static int amdgpu_ttm_tt_pin_userptr(struct ttm_device *bdev,
return 0;
+release_sg_table:
+ sg_free_table(ttm->sg);
release_sg:
kfree(ttm->sg);
ttm->sg = NULL;
@@ -1849,6 +1851,7 @@ int amdgpu_ttm_init(struct amdgpu_device *adev)
mutex_init(&adev->mman.gtt_window_lock);
+ dma_set_max_seg_size(adev->dev, UINT_MAX);
/* No others user of address space so set it to 0 */
r = ttm_device_init(&adev->mman.bdev, &amdgpu_bo_driver, adev->dev,
adev_to_drm(adev)->anon_inode->i_mapping,
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
index 785a343a95f0..e7cd51c95141 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
@@ -2223,6 +2223,18 @@ static int gfx_v9_0_sw_init(void *handle)
}
switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+ case IP_VERSION(9, 4, 2):
+ adev->gfx.cleaner_shader_ptr = gfx_9_4_2_cleaner_shader_hex;
+ adev->gfx.cleaner_shader_size = sizeof(gfx_9_4_2_cleaner_shader_hex);
+ if (adev->gfx.mec_fw_version >= 88) {
+ adev->gfx.enable_cleaner_shader = true;
+ r = amdgpu_gfx_cleaner_shader_sw_init(adev, adev->gfx.cleaner_shader_size);
+ if (r) {
+ adev->gfx.enable_cleaner_shader = false;
+ dev_err(adev->dev, "Failed to initialize cleaner shader\n");
+ }
+ }
+ break;
default:
adev->gfx.enable_cleaner_shader = false;
break;
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0_cleaner_shader.h b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0_cleaner_shader.h
index 36c0292b5110..0b6bd09b7529 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0_cleaner_shader.h
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0_cleaner_shader.h
@@ -1,6 +1,6 @@
/* SPDX-License-Identifier: MIT */
/*
- * Copyright 2018 Advanced Micro Devices, Inc.
+ * Copyright 2024 Advanced Micro Devices, Inc.
*
* Permission is hereby granted, free of charge, to any person obtaining a
* copy of this software and associated documentation files (the "Software"),
@@ -24,3 +24,45 @@
static const u32 __maybe_unused gfx_9_0_cleaner_shader_hex[] = {
/* Add the cleaner shader code here */
};
+
+/* Define the cleaner shader gfx_9_4_2 */
+static const u32 gfx_9_4_2_cleaner_shader_hex[] = {
+ 0xbf068100, 0xbf84003b,
+ 0xbf8a0000, 0xb07c0000,
+ 0xbe8200ff, 0x00000078,
+ 0xbf110802, 0x7e000280,
+ 0x7e020280, 0x7e040280,
+ 0x7e060280, 0x7e080280,
+ 0x7e0a0280, 0x7e0c0280,
+ 0x7e0e0280, 0x80828802,
+ 0xbe803202, 0xbf84fff5,
+ 0xbf9c0000, 0xbe8200ff,
+ 0x80000000, 0x86020102,
+ 0xbf840011, 0xbefe00c1,
+ 0xbeff00c1, 0xd28c0001,
+ 0x0001007f, 0xd28d0001,
+ 0x0002027e, 0x10020288,
+ 0xbe8200bf, 0xbefc00c1,
+ 0xd89c2000, 0x00020201,
+ 0xd89c6040, 0x00040401,
+ 0x320202ff, 0x00000400,
+ 0x80828102, 0xbf84fff8,
+ 0xbefc00ff, 0x0000005c,
+ 0xbf800000, 0xbe802c80,
+ 0xbe812c80, 0xbe822c80,
+ 0xbe832c80, 0x80fc847c,
+ 0xbf84fffa, 0xbee60080,
+ 0xbee70080, 0xbeea0180,
+ 0xbeec0180, 0xbeee0180,
+ 0xbef00180, 0xbef20180,
+ 0xbef40180, 0xbef60180,
+ 0xbef80180, 0xbefa0180,
+ 0xbf810000, 0xbf8d0001,
+ 0xbefc00ff, 0x0000005c,
+ 0xbf800000, 0xbe802c80,
+ 0xbe812c80, 0xbe822c80,
+ 0xbe832c80, 0x80fc847c,
+ 0xbf84fffa, 0xbee60080,
+ 0xbee70080, 0xbeea01ff,
+ 0x000000ee, 0xbf810000,
+};
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_2_cleaner_shader.asm b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_2_cleaner_shader.asm
new file mode 100644
index 000000000000..35b8cf9070bd
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_2_cleaner_shader.asm
@@ -0,0 +1,153 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2024 Advanced Micro Devices, 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.
+ */
+
+// This shader is to clean LDS, SGPRs and VGPRs. It is first 64 Dwords or 256 bytes of 192 Dwords cleaner shader.
+//To turn this shader program on for complitaion change this to main and lower shader main to main_1
+
+// MI200 : Clear SGPRs, VGPRs and LDS
+// Uses two kernels launched separately:
+// 1. Clean VGPRs, LDS, and lower SGPRs
+// Launches one workgroup per CU, each workgroup with 4x wave64 per SIMD in the CU
+// Waves are "wave64" and have 128 VGPRs each, which uses all 512 VGPRs per SIMD
+// Waves in the workgroup share the 64KB of LDS
+// Each wave clears SGPRs 0 - 95. Because there are 4 waves/SIMD, this is physical SGPRs 0-383
+// Each wave clears 128 VGPRs, so all 512 in the SIMD
+// The first wave of the workgroup clears its 64KB of LDS
+// The shader starts with "S_BARRIER" to ensure SPI has launched all waves of the workgroup
+// before any wave in the workgroup could end. Without this, it is possible not all SGPRs get cleared.
+// 2. Clean remaining SGPRs
+// Launches a workgroup with 24 waves per workgroup, yielding 6 waves per SIMD in each CU
+// Waves are allocating 96 SGPRs
+// CP sets up SPI_RESOURCE_RESERVE_* registers to prevent these waves from allocating SGPRs 0-223.
+// As such, these 6 waves per SIMD are allocated physical SGPRs 224-799
+// Barriers do not work for >16 waves per workgroup, so we cannot start with S_BARRIER
+// Instead, the shader starts with an S_SETHALT 1. Once all waves are launched CP will send unhalt command
+// The shader then clears all SGPRs allocated to it, cleaning out physical SGPRs 224-799
+
+shader main
+ asic(MI200)
+ type(CS)
+ wave_size(64)
+// Note: original source code from SQ team
+
+// (theorhetical fastest = ~512clks vgpr + 1536 lds + ~128 sgpr = 2176 clks)
+
+ s_cmp_eq_u32 s0, 1 // Bit0 is set, sgpr0 is set then clear VGPRS and LDS as FW set COMPUTE_USER_DATA_3
+ s_cbranch_scc0 label_0023 // Clean VGPRs and LDS if sgpr0 of wave is set, scc = (s3 == 1)
+ S_BARRIER
+
+ s_movk_i32 m0, 0x0000
+ s_mov_b32 s2, 0x00000078 // Loop 128/8=16 times (loop unrolled for performance)
+ //
+ // CLEAR VGPRs
+ //
+ s_set_gpr_idx_on s2, 0x8 // enable Dest VGPR indexing
+label_0005:
+ v_mov_b32 v0, 0
+ v_mov_b32 v1, 0
+ v_mov_b32 v2, 0
+ v_mov_b32 v3, 0
+ v_mov_b32 v4, 0
+ v_mov_b32 v5, 0
+ v_mov_b32 v6, 0
+ v_mov_b32 v7, 0
+ s_sub_u32 s2, s2, 8
+ s_set_gpr_idx_idx s2
+ s_cbranch_scc0 label_0005
+ s_set_gpr_idx_off
+
+ //
+ //
+
+ s_mov_b32 s2, 0x80000000 // Bit31 is first_wave
+ s_and_b32 s2, s2, s1 // sgpr0 has tg_size (first_wave) term as in ucode only COMPUTE_PGM_RSRC2.tg_size_en is set
+ s_cbranch_scc0 label_clean_sgpr_1 // Clean LDS if its first wave of ThreadGroup/WorkGroup
+ // CLEAR LDS
+ //
+ s_mov_b32 exec_lo, 0xffffffff
+ s_mov_b32 exec_hi, 0xffffffff
+ v_mbcnt_lo_u32_b32 v1, exec_hi, 0 // Set V1 to thread-ID (0..63)
+ v_mbcnt_hi_u32_b32 v1, exec_lo, v1 // Set V1 to thread-ID (0..63)
+ v_mul_u32_u24 v1, 0x00000008, v1 // * 8, so each thread is a double-dword address (8byte)
+ s_mov_b32 s2, 0x00000003f // 64 loop iterations
+ s_mov_b32 m0, 0xffffffff
+ // Clear all of LDS space
+ // Each FirstWave of WorkGroup clears 64kbyte block
+
+label_001F:
+ ds_write2_b64 v1, v[2:3], v[2:3] offset1:32
+ ds_write2_b64 v1, v[4:5], v[4:5] offset0:64 offset1:96
+ v_add_co_u32 v1, vcc, 0x00000400, v1
+ s_sub_u32 s2, s2, 1
+ s_cbranch_scc0 label_001F
+ //
+ // CLEAR SGPRs
+ //
+label_clean_sgpr_1:
+ s_mov_b32 m0, 0x0000005c // Loop 96/4=24 times (loop unrolled for performance)
+ s_nop 0
+label_sgpr_loop:
+ s_movreld_b32 s0, 0
+ s_movreld_b32 s1, 0
+ s_movreld_b32 s2, 0
+ s_movreld_b32 s3, 0
+ s_sub_u32 m0, m0, 4
+ s_cbranch_scc0 label_sgpr_loop
+
+ //clear vcc, flat scratch
+ s_mov_b32 flat_scratch_lo, 0 //clear flat scratch lo SGPR
+ s_mov_b32 flat_scratch_hi, 0 //clear flat scratch hi SGPR
+ s_mov_b64 vcc, 0 //clear vcc
+ s_mov_b64 ttmp0, 0 //Clear ttmp0 and ttmp1
+ s_mov_b64 ttmp2, 0 //Clear ttmp2 and ttmp3
+ s_mov_b64 ttmp4, 0 //Clear ttmp4 and ttmp5
+ s_mov_b64 ttmp6, 0 //Clear ttmp6 and ttmp7
+ s_mov_b64 ttmp8, 0 //Clear ttmp8 and ttmp9
+ s_mov_b64 ttmp10, 0 //Clear ttmp10 and ttmp11
+ s_mov_b64 ttmp12, 0 //Clear ttmp12 and ttmp13
+ s_mov_b64 ttmp14, 0 //Clear ttmp14 and ttmp15
+s_endpgm
+
+label_0023:
+
+ s_sethalt 1
+
+ s_mov_b32 m0, 0x0000005c // Loop 96/4=24 times (loop unrolled for performance)
+ s_nop 0
+label_sgpr_loop1:
+
+ s_movreld_b32 s0, 0
+ s_movreld_b32 s1, 0
+ s_movreld_b32 s2, 0
+ s_movreld_b32 s3, 0
+ s_sub_u32 m0, m0, 4
+ s_cbranch_scc0 label_sgpr_loop1
+
+ //clear vcc, flat scratch
+ s_mov_b32 flat_scratch_lo, 0 //clear flat scratch lo SGPR
+ s_mov_b32 flat_scratch_hi, 0 //clear flat scratch hi SGPR
+ s_mov_b64 vcc, 0xee //clear vcc
+
+s_endpgm
+end
+
diff --git a/drivers/gpu/drm/amd/amdgpu/hdp_v4_0.c b/drivers/gpu/drm/amd/amdgpu/hdp_v4_0.c
index e019249883fb..194026e9be33 100644
--- a/drivers/gpu/drm/amd/amdgpu/hdp_v4_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/hdp_v4_0.c
@@ -40,10 +40,12 @@
static void hdp_v4_0_flush_hdp(struct amdgpu_device *adev,
struct amdgpu_ring *ring)
{
- if (!ring || !ring->funcs->emit_wreg)
+ if (!ring || !ring->funcs->emit_wreg) {
WREG32((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, 0);
- else
+ RREG32((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2);
+ } else {
amdgpu_ring_emit_wreg(ring, (adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, 0);
+ }
}
static void hdp_v4_0_invalidate_hdp(struct amdgpu_device *adev,
@@ -54,11 +56,13 @@ static void hdp_v4_0_invalidate_hdp(struct amdgpu_device *adev,
amdgpu_ip_version(adev, HDP_HWIP, 0) == IP_VERSION(4, 4, 5))
return;
- if (!ring || !ring->funcs->emit_wreg)
+ if (!ring || !ring->funcs->emit_wreg) {
WREG32_SOC15_NO_KIQ(HDP, 0, mmHDP_READ_CACHE_INVALIDATE, 1);
- else
+ RREG32_SOC15_NO_KIQ(HDP, 0, mmHDP_READ_CACHE_INVALIDATE);
+ } else {
amdgpu_ring_emit_wreg(ring, SOC15_REG_OFFSET(
HDP, 0, mmHDP_READ_CACHE_INVALIDATE), 1);
+ }
}
static void hdp_v4_0_query_ras_error_count(struct amdgpu_device *adev,
diff --git a/drivers/gpu/drm/amd/amdgpu/hdp_v5_0.c b/drivers/gpu/drm/amd/amdgpu/hdp_v5_0.c
index ed7facacf2fe..d3962d469088 100644
--- a/drivers/gpu/drm/amd/amdgpu/hdp_v5_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/hdp_v5_0.c
@@ -31,10 +31,12 @@
static void hdp_v5_0_flush_hdp(struct amdgpu_device *adev,
struct amdgpu_ring *ring)
{
- if (!ring || !ring->funcs->emit_wreg)
+ if (!ring || !ring->funcs->emit_wreg) {
WREG32((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, 0);
- else
+ RREG32((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2);
+ } else {
amdgpu_ring_emit_wreg(ring, (adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, 0);
+ }
}
static void hdp_v5_0_invalidate_hdp(struct amdgpu_device *adev,
@@ -42,6 +44,7 @@ static void hdp_v5_0_invalidate_hdp(struct amdgpu_device *adev,
{
if (!ring || !ring->funcs->emit_wreg) {
WREG32_SOC15_NO_KIQ(HDP, 0, mmHDP_READ_CACHE_INVALIDATE, 1);
+ RREG32_SOC15_NO_KIQ(HDP, 0, mmHDP_READ_CACHE_INVALIDATE);
} else {
amdgpu_ring_emit_wreg(ring, SOC15_REG_OFFSET(
HDP, 0, mmHDP_READ_CACHE_INVALIDATE), 1);
diff --git a/drivers/gpu/drm/amd/amdgpu/hdp_v5_2.c b/drivers/gpu/drm/amd/amdgpu/hdp_v5_2.c
index 29c3484ae1f1..f52552c5fa27 100644
--- a/drivers/gpu/drm/amd/amdgpu/hdp_v5_2.c
+++ b/drivers/gpu/drm/amd/amdgpu/hdp_v5_2.c
@@ -31,13 +31,15 @@
static void hdp_v5_2_flush_hdp(struct amdgpu_device *adev,
struct amdgpu_ring *ring)
{
- if (!ring || !ring->funcs->emit_wreg)
+ if (!ring || !ring->funcs->emit_wreg) {
WREG32_NO_KIQ((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2,
0);
- else
+ RREG32_NO_KIQ((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2);
+ } else {
amdgpu_ring_emit_wreg(ring,
(adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2,
0);
+ }
}
static void hdp_v5_2_update_mem_power_gating(struct amdgpu_device *adev,
diff --git a/drivers/gpu/drm/amd/amdgpu/hdp_v6_0.c b/drivers/gpu/drm/amd/amdgpu/hdp_v6_0.c
index 33736d361dd0..6948fe9956ce 100644
--- a/drivers/gpu/drm/amd/amdgpu/hdp_v6_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/hdp_v6_0.c
@@ -34,10 +34,12 @@
static void hdp_v6_0_flush_hdp(struct amdgpu_device *adev,
struct amdgpu_ring *ring)
{
- if (!ring || !ring->funcs->emit_wreg)
+ if (!ring || !ring->funcs->emit_wreg) {
WREG32((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, 0);
- else
+ RREG32((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2);
+ } else {
amdgpu_ring_emit_wreg(ring, (adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, 0);
+ }
}
static void hdp_v6_0_update_clock_gating(struct amdgpu_device *adev,
diff --git a/drivers/gpu/drm/amd/amdgpu/hdp_v7_0.c b/drivers/gpu/drm/amd/amdgpu/hdp_v7_0.c
index 1c99bb09e2a1..63820329f67e 100644
--- a/drivers/gpu/drm/amd/amdgpu/hdp_v7_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/hdp_v7_0.c
@@ -31,10 +31,12 @@
static void hdp_v7_0_flush_hdp(struct amdgpu_device *adev,
struct amdgpu_ring *ring)
{
- if (!ring || !ring->funcs->emit_wreg)
+ if (!ring || !ring->funcs->emit_wreg) {
WREG32((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, 0);
- else
+ RREG32((adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2);
+ } else {
amdgpu_ring_emit_wreg(ring, (adev->rmmio_remap.reg_offset + KFD_MMIO_REMAP_HDP_MEM_FLUSH_CNTL) >> 2, 0);
+ }
}
static void hdp_v7_0_update_clock_gating(struct amdgpu_device *adev,
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c
index 0fda70336300..6fca2915ea8f 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c
@@ -116,6 +116,20 @@ static int vcn_v4_0_3_early_init(void *handle)
return amdgpu_vcn_early_init(adev);
}
+static int vcn_v4_0_3_fw_shared_init(struct amdgpu_device *adev, int inst_idx)
+{
+ struct amdgpu_vcn4_fw_shared *fw_shared;
+
+ fw_shared = adev->vcn.inst[inst_idx].fw_shared.cpu_addr;
+ fw_shared->present_flag_0 = cpu_to_le32(AMDGPU_FW_SHARED_FLAG_0_UNIFIED_QUEUE);
+ fw_shared->sq.is_enabled = 1;
+
+ if (amdgpu_vcnfw_log)
+ amdgpu_vcn_fwlog_init(&adev->vcn.inst[inst_idx]);
+
+ return 0;
+}
+
/**
* vcn_v4_0_3_sw_init - sw init for VCN block
*
@@ -148,8 +162,6 @@ static int vcn_v4_0_3_sw_init(void *handle)
return r;
for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
- volatile struct amdgpu_vcn4_fw_shared *fw_shared;
-
vcn_inst = GET_INST(VCN, i);
ring = &adev->vcn.inst[i].ring_enc[0];
@@ -172,12 +184,7 @@ static int vcn_v4_0_3_sw_init(void *handle)
if (r)
return r;
- fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr;
- fw_shared->present_flag_0 = cpu_to_le32(AMDGPU_FW_SHARED_FLAG_0_UNIFIED_QUEUE);
- fw_shared->sq.is_enabled = true;
-
- if (amdgpu_vcnfw_log)
- amdgpu_vcn_fwlog_init(&adev->vcn.inst[i]);
+ vcn_v4_0_3_fw_shared_init(adev, i);
}
if (amdgpu_sriov_vf(adev)) {
@@ -273,6 +280,8 @@ static int vcn_v4_0_3_hw_init(void *handle)
}
} else {
for (i = 0; i < adev->vcn.num_vcn_inst; ++i) {
+ struct amdgpu_vcn4_fw_shared *fw_shared;
+
vcn_inst = GET_INST(VCN, i);
ring = &adev->vcn.inst[i].ring_enc[0];
@@ -296,6 +305,11 @@ static int vcn_v4_0_3_hw_init(void *handle)
regVCN_RB1_DB_CTRL);
}
+ /* Re-init fw_shared when RAS fatal error occurred */
+ fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr;
+ if (!fw_shared->sq.is_enabled)
+ vcn_v4_0_3_fw_shared_init(adev, i);
+
r = amdgpu_ring_test_helper(ring);
if (r)
return r;
diff --git a/drivers/gpu/drm/amd/amdgpu/vega20_ih.c b/drivers/gpu/drm/amd/amdgpu/vega20_ih.c
index ac439f0565e3..16f5561fb86e 100644
--- a/drivers/gpu/drm/amd/amdgpu/vega20_ih.c
+++ b/drivers/gpu/drm/amd/amdgpu/vega20_ih.c
@@ -114,6 +114,33 @@ static int vega20_ih_toggle_ring_interrupts(struct amdgpu_device *adev,
tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, RB_ENABLE, (enable ? 1 : 0));
tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, RB_GPU_TS_ENABLE, 1);
+ if (enable) {
+ /* Unset the CLEAR_OVERFLOW bit to make sure the next step
+ * is switching the bit from 0 to 1
+ */
+ tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 0);
+ if (amdgpu_sriov_vf(adev) && amdgpu_sriov_reg_indirect_ih(adev)) {
+ if (psp_reg_program(&adev->psp, ih_regs->psp_reg_id, tmp))
+ return -ETIMEDOUT;
+ } else {
+ WREG32_NO_KIQ(ih_regs->ih_rb_cntl, tmp);
+ }
+
+ /* Clear RB_OVERFLOW bit */
+ tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 1);
+ if (amdgpu_sriov_vf(adev) && amdgpu_sriov_reg_indirect_ih(adev)) {
+ if (psp_reg_program(&adev->psp, ih_regs->psp_reg_id, tmp))
+ return -ETIMEDOUT;
+ } else {
+ WREG32_NO_KIQ(ih_regs->ih_rb_cntl, tmp);
+ }
+
+ /* Unset the CLEAR_OVERFLOW bit immediately so new overflows
+ * can be detected.
+ */
+ tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, WPTR_OVERFLOW_CLEAR, 0);
+ }
+
/* enable_intr field is only valid in ring0 */
if (ih == &adev->irq.ih)
tmp = REG_SET_FIELD(tmp, IH_RB_CNTL, ENABLE_INTR, (enable ? 1 : 0));
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_crat.c b/drivers/gpu/drm/amd/amdkfd/kfd_crat.c
index 48caecf7e72e..8de61cc524c9 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_crat.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_crat.c
@@ -1509,6 +1509,8 @@ static int kfd_fill_gpu_cache_info_from_gfx_config_v2(struct kfd_dev *kdev,
if (adev->gfx.config.gc_tcp_size_per_cu) {
pcache_info[i].cache_size = adev->gfx.config.gc_tcp_size_per_cu;
pcache_info[i].cache_level = 1;
+ /* Cacheline size not available in IP discovery for gc943,gc944 */
+ pcache_info[i].cache_line_size = 128;
pcache_info[i].flags = (CRAT_CACHE_FLAGS_ENABLED |
CRAT_CACHE_FLAGS_DATA_CACHE |
CRAT_CACHE_FLAGS_SIMD_CACHE);
@@ -1520,6 +1522,7 @@ static int kfd_fill_gpu_cache_info_from_gfx_config_v2(struct kfd_dev *kdev,
pcache_info[i].cache_size =
adev->gfx.config.gc_l1_instruction_cache_size_per_sqc;
pcache_info[i].cache_level = 1;
+ pcache_info[i].cache_line_size = 64;
pcache_info[i].flags = (CRAT_CACHE_FLAGS_ENABLED |
CRAT_CACHE_FLAGS_INST_CACHE |
CRAT_CACHE_FLAGS_SIMD_CACHE);
@@ -1530,6 +1533,7 @@ static int kfd_fill_gpu_cache_info_from_gfx_config_v2(struct kfd_dev *kdev,
if (adev->gfx.config.gc_l1_data_cache_size_per_sqc) {
pcache_info[i].cache_size = adev->gfx.config.gc_l1_data_cache_size_per_sqc;
pcache_info[i].cache_level = 1;
+ pcache_info[i].cache_line_size = 64;
pcache_info[i].flags = (CRAT_CACHE_FLAGS_ENABLED |
CRAT_CACHE_FLAGS_DATA_CACHE |
CRAT_CACHE_FLAGS_SIMD_CACHE);
@@ -1540,6 +1544,7 @@ static int kfd_fill_gpu_cache_info_from_gfx_config_v2(struct kfd_dev *kdev,
if (adev->gfx.config.gc_tcc_size) {
pcache_info[i].cache_size = adev->gfx.config.gc_tcc_size;
pcache_info[i].cache_level = 2;
+ pcache_info[i].cache_line_size = 128;
pcache_info[i].flags = (CRAT_CACHE_FLAGS_ENABLED |
CRAT_CACHE_FLAGS_DATA_CACHE |
CRAT_CACHE_FLAGS_SIMD_CACHE);
@@ -1550,6 +1555,7 @@ static int kfd_fill_gpu_cache_info_from_gfx_config_v2(struct kfd_dev *kdev,
if (adev->gmc.mall_size) {
pcache_info[i].cache_size = adev->gmc.mall_size / 1024;
pcache_info[i].cache_level = 3;
+ pcache_info[i].cache_line_size = 64;
pcache_info[i].flags = (CRAT_CACHE_FLAGS_ENABLED |
CRAT_CACHE_FLAGS_DATA_CACHE |
CRAT_CACHE_FLAGS_SIMD_CACHE);
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device.c b/drivers/gpu/drm/amd/amdkfd/kfd_device.c
index fad1c8f2bc83..b05be24531e1 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device.c
@@ -235,6 +235,9 @@ static void kfd_device_info_init(struct kfd_dev *kfd,
*/
kfd->device_info.needs_pci_atomics = true;
kfd->device_info.no_atomic_fw_version = kfd->adev->gfx.rs64_enable ? 509 : 0;
+ } else if (gc_version < IP_VERSION(13, 0, 0)) {
+ kfd->device_info.needs_pci_atomics = true;
+ kfd->device_info.no_atomic_fw_version = 2090;
} else {
kfd->device_info.needs_pci_atomics = true;
}
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
index 24fbde7dd1c4..ad3a3aa72b51 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
@@ -1910,7 +1910,11 @@ static int amdgpu_dm_init(struct amdgpu_device *adev)
else
init_data.flags.gpu_vm_support = (amdgpu_sg_display != 0);
} else {
- init_data.flags.gpu_vm_support = (amdgpu_sg_display != 0) && (adev->flags & AMD_IS_APU);
+ if (amdgpu_ip_version(adev, DCE_HWIP, 0) == IP_VERSION(2, 0, 3))
+ init_data.flags.gpu_vm_support = (amdgpu_sg_display == 1);
+ else
+ init_data.flags.gpu_vm_support =
+ (amdgpu_sg_display != 0) && (adev->flags & AMD_IS_APU);
}
adev->mode_info.gpu_vm_support = init_data.flags.gpu_vm_support;
@@ -7337,10 +7341,15 @@ create_validate_stream_for_sink(struct amdgpu_dm_connector *aconnector,
const struct drm_connector_state *drm_state = dm_state ? &dm_state->base : NULL;
int requested_bpc = drm_state ? drm_state->max_requested_bpc : 8;
enum dc_status dc_result = DC_OK;
+ uint8_t bpc_limit = 6;
if (!dm_state)
return NULL;
+ if (aconnector->dc_link->connector_signal == SIGNAL_TYPE_HDMI_TYPE_A ||
+ aconnector->dc_link->dpcd_caps.dongle_type == DISPLAY_DONGLE_DP_HDMI_CONVERTER)
+ bpc_limit = 8;
+
do {
stream = create_stream_for_sink(connector, drm_mode,
dm_state, old_stream,
@@ -7361,11 +7370,12 @@ create_validate_stream_for_sink(struct amdgpu_dm_connector *aconnector,
dc_result = dm_validate_stream_and_context(adev->dm.dc, stream);
if (dc_result != DC_OK) {
- DRM_DEBUG_KMS("Mode %dx%d (clk %d) failed DC validation with error %d (%s)\n",
+ DRM_DEBUG_KMS("Mode %dx%d (clk %d) pixel_encoding:%s color_depth:%s failed validation -- %s\n",
drm_mode->hdisplay,
drm_mode->vdisplay,
drm_mode->clock,
- dc_result,
+ dc_pixel_encoding_to_str(stream->timing.pixel_encoding),
+ dc_color_depth_to_str(stream->timing.display_color_depth),
dc_status_to_str(dc_result));
dc_stream_release(stream);
@@ -7373,10 +7383,13 @@ create_validate_stream_for_sink(struct amdgpu_dm_connector *aconnector,
requested_bpc -= 2; /* lower bpc to retry validation */
}
- } while (stream == NULL && requested_bpc >= 6);
+ } while (stream == NULL && requested_bpc >= bpc_limit);
- if (dc_result == DC_FAIL_ENC_VALIDATE && !aconnector->force_yuv420_output) {
- DRM_DEBUG_KMS("Retry forcing YCbCr420 encoding\n");
+ if ((dc_result == DC_FAIL_ENC_VALIDATE ||
+ dc_result == DC_EXCEED_DONGLE_CAP) &&
+ !aconnector->force_yuv420_output) {
+ DRM_DEBUG_KMS("%s:%d Retry forcing yuv420 encoding\n",
+ __func__, __LINE__);
aconnector->force_yuv420_output = true;
stream = create_validate_stream_for_sink(aconnector, drm_mode,
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c
index b46a3afe48ca..3bd0d46c1701 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c
@@ -132,6 +132,8 @@ static void dcn35_disable_otg_wa(struct clk_mgr *clk_mgr_base, struct dc_state *
for (i = 0; i < dc->res_pool->pipe_count; ++i) {
struct pipe_ctx *old_pipe = &dc->current_state->res_ctx.pipe_ctx[i];
struct pipe_ctx *new_pipe = &context->res_ctx.pipe_ctx[i];
+ struct clk_mgr_internal *clk_mgr_internal = TO_CLK_MGR_INTERNAL(clk_mgr_base);
+ struct dccg *dccg = clk_mgr_internal->dccg;
struct pipe_ctx *pipe = safe_to_lower
? &context->res_ctx.pipe_ctx[i]
: &dc->current_state->res_ctx.pipe_ctx[i];
@@ -148,8 +150,13 @@ static void dcn35_disable_otg_wa(struct clk_mgr *clk_mgr_base, struct dc_state *
new_pipe->stream_res.stream_enc &&
new_pipe->stream_res.stream_enc->funcs->is_fifo_enabled &&
new_pipe->stream_res.stream_enc->funcs->is_fifo_enabled(new_pipe->stream_res.stream_enc);
- if (pipe->stream && (pipe->stream->dpms_off || dc_is_virtual_signal(pipe->stream->signal) ||
- !pipe->stream->link_enc) && !stream_changed_otg_dig_on) {
+ bool has_active_hpo = dccg->ctx->dc->link_srv->dp_is_128b_132b_signal(old_pipe) && dccg->ctx->dc->link_srv->dp_is_128b_132b_signal(new_pipe);
+
+ if (!has_active_hpo && !dccg->ctx->dc->link_srv->dp_is_128b_132b_signal(pipe) &&
+ (pipe->stream && (pipe->stream->dpms_off || dc_is_virtual_signal(pipe->stream->signal) ||
+ !pipe->stream->link_enc) && !stream_changed_otg_dig_on)) {
+
+
/* This w/a should not trigger when we have a dig active */
if (disable) {
if (pipe->stream_res.tg && pipe->stream_res.tg->funcs->immediate_disable_crtc)
@@ -257,11 +264,11 @@ static void dcn35_notify_host_router_bw(struct clk_mgr *clk_mgr_base, struct dc_
struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
uint32_t host_router_bw_kbps[MAX_HOST_ROUTERS_NUM] = { 0 };
int i;
-
for (i = 0; i < context->stream_count; ++i) {
const struct dc_stream_state *stream = context->streams[i];
const struct dc_link *link = stream->link;
- uint8_t lowest_dpia_index = 0, hr_index = 0;
+ uint8_t lowest_dpia_index = 0;
+ unsigned int hr_index = 0;
if (!link)
continue;
@@ -271,6 +278,8 @@ static void dcn35_notify_host_router_bw(struct clk_mgr *clk_mgr_base, struct dc_
continue;
hr_index = (link->link_index - lowest_dpia_index) / 2;
+ if (hr_index >= MAX_HOST_ROUTERS_NUM)
+ continue;
host_router_bw_kbps[hr_index] += dc_bandwidth_in_kbps_from_timing(
&stream->timing, dc_link_get_highest_encoding_format(link));
}
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc.c b/drivers/gpu/drm/amd/display/dc/core/dc.c
index a6911bb2cf0c..9f570d447c20 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc.c
@@ -6006,3 +6006,21 @@ struct dc_power_profile dc_get_power_profile_for_dc_state(const struct dc_state
return profile;
}
+
+/*
+ **********************************************************************************
+ * dc_get_det_buffer_size_from_state() - extracts detile buffer size from dc state
+ *
+ * Called when DM wants to log detile buffer size from dc_state
+ *
+ **********************************************************************************
+ */
+unsigned int dc_get_det_buffer_size_from_state(const struct dc_state *context)
+{
+ struct dc *dc = context->clk_mgr->ctx->dc;
+
+ if (dc->res_pool->funcs->get_det_buffer_size)
+ return dc->res_pool->funcs->get_det_buffer_size(context);
+ else
+ return 0;
+}
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_debug.c b/drivers/gpu/drm/amd/display/dc/core/dc_debug.c
index 801cdbc8117d..e255c204b7e8 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_debug.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_debug.c
@@ -434,3 +434,43 @@ char *dc_status_to_str(enum dc_status status)
return "Unexpected status error";
}
+
+char *dc_pixel_encoding_to_str(enum dc_pixel_encoding pixel_encoding)
+{
+ switch (pixel_encoding) {
+ case PIXEL_ENCODING_RGB:
+ return "RGB";
+ case PIXEL_ENCODING_YCBCR422:
+ return "YUV422";
+ case PIXEL_ENCODING_YCBCR444:
+ return "YUV444";
+ case PIXEL_ENCODING_YCBCR420:
+ return "YUV420";
+ default:
+ return "Unknown";
+ }
+}
+
+char *dc_color_depth_to_str(enum dc_color_depth color_depth)
+{
+ switch (color_depth) {
+ case COLOR_DEPTH_666:
+ return "6-bpc";
+ case COLOR_DEPTH_888:
+ return "8-bpc";
+ case COLOR_DEPTH_101010:
+ return "10-bpc";
+ case COLOR_DEPTH_121212:
+ return "12-bpc";
+ case COLOR_DEPTH_141414:
+ return "14-bpc";
+ case COLOR_DEPTH_161616:
+ return "16-bpc";
+ case COLOR_DEPTH_999:
+ return "9-bpc";
+ case COLOR_DEPTH_111111:
+ return "11-bpc";
+ default:
+ return "Unknown";
+ }
+}
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_resource.c b/drivers/gpu/drm/amd/display/dc/core/dc_resource.c
index c7599c40d4be..d915020a4295 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_resource.c
@@ -765,25 +765,6 @@ static inline void get_vp_scan_direction(
*flip_horz_scan_dir = !*flip_horz_scan_dir;
}
-/*
- * This is a preliminary vp size calculation to allow us to check taps support.
- * The result is completely overridden afterwards.
- */
-static void calculate_viewport_size(struct pipe_ctx *pipe_ctx)
-{
- struct scaler_data *data = &pipe_ctx->plane_res.scl_data;
-
- data->viewport.width = dc_fixpt_ceil(dc_fixpt_mul_int(data->ratios.horz, data->recout.width));
- data->viewport.height = dc_fixpt_ceil(dc_fixpt_mul_int(data->ratios.vert, data->recout.height));
- data->viewport_c.width = dc_fixpt_ceil(dc_fixpt_mul_int(data->ratios.horz_c, data->recout.width));
- data->viewport_c.height = dc_fixpt_ceil(dc_fixpt_mul_int(data->ratios.vert_c, data->recout.height));
- if (pipe_ctx->plane_state->rotation == ROTATION_ANGLE_90 ||
- pipe_ctx->plane_state->rotation == ROTATION_ANGLE_270) {
- swap(data->viewport.width, data->viewport.height);
- swap(data->viewport_c.width, data->viewport_c.height);
- }
-}
-
static struct rect intersect_rec(const struct rect *r0, const struct rect *r1)
{
struct rect rec;
@@ -1468,6 +1449,7 @@ bool resource_build_scaling_params(struct pipe_ctx *pipe_ctx)
const struct dc_plane_state *plane_state = pipe_ctx->plane_state;
struct dc_crtc_timing *timing = &pipe_ctx->stream->timing;
const struct rect odm_slice_src = resource_get_odm_slice_src_rect(pipe_ctx);
+ struct scaling_taps temp = {0};
bool res = false;
DC_LOGGER_INIT(pipe_ctx->stream->ctx->logger);
@@ -1519,14 +1501,16 @@ bool resource_build_scaling_params(struct pipe_ctx *pipe_ctx)
res = spl_calculate_scaler_params(spl_in, spl_out);
// Convert respective out params from SPL to scaler data
translate_SPL_out_params_to_pipe_ctx(pipe_ctx, spl_out);
+
+ /* Ignore scaler failure if pipe context plane is phantom plane */
+ if (!res && plane_state->is_phantom)
+ res = true;
} else {
#endif
/* depends on h_active */
calculate_recout(pipe_ctx);
/* depends on pixel format */
calculate_scaling_ratios(pipe_ctx);
- /* depends on scaling ratios and recout, does not calculate offset yet */
- calculate_viewport_size(pipe_ctx);
/*
* LB calculations depend on vp size, h/v_active and scaling ratios
@@ -1547,6 +1531,24 @@ bool resource_build_scaling_params(struct pipe_ctx *pipe_ctx)
pipe_ctx->plane_res.scl_data.lb_params.alpha_en = plane_state->per_pixel_alpha;
+ // get TAP value with 100x100 dummy data for max scaling qualify, override
+ // if a new scaling quality required
+ pipe_ctx->plane_res.scl_data.viewport.width = 100;
+ pipe_ctx->plane_res.scl_data.viewport.height = 100;
+ pipe_ctx->plane_res.scl_data.viewport_c.width = 100;
+ pipe_ctx->plane_res.scl_data.viewport_c.height = 100;
+ if (pipe_ctx->plane_res.xfm != NULL)
+ res = pipe_ctx->plane_res.xfm->funcs->transform_get_optimal_number_of_taps(
+ pipe_ctx->plane_res.xfm, &pipe_ctx->plane_res.scl_data, &plane_state->scaling_quality);
+
+ if (pipe_ctx->plane_res.dpp != NULL)
+ res = pipe_ctx->plane_res.dpp->funcs->dpp_get_optimal_number_of_taps(
+ pipe_ctx->plane_res.dpp, &pipe_ctx->plane_res.scl_data, &plane_state->scaling_quality);
+
+ temp = pipe_ctx->plane_res.scl_data.taps;
+
+ calculate_inits_and_viewports(pipe_ctx);
+
if (pipe_ctx->plane_res.xfm != NULL)
res = pipe_ctx->plane_res.xfm->funcs->transform_get_optimal_number_of_taps(
pipe_ctx->plane_res.xfm, &pipe_ctx->plane_res.scl_data, &plane_state->scaling_quality);
@@ -1573,11 +1575,14 @@ bool resource_build_scaling_params(struct pipe_ctx *pipe_ctx)
&plane_state->scaling_quality);
}
- /*
- * Depends on recout, scaling ratios, h_active and taps
- * May need to re-check lb size after this in some obscure scenario
- */
- if (res)
+ /* Ignore scaler failure if pipe context plane is phantom plane */
+ if (!res && plane_state->is_phantom)
+ res = true;
+
+ if (res && (pipe_ctx->plane_res.scl_data.taps.v_taps != temp.v_taps ||
+ pipe_ctx->plane_res.scl_data.taps.h_taps != temp.h_taps ||
+ pipe_ctx->plane_res.scl_data.taps.v_taps_c != temp.v_taps_c ||
+ pipe_ctx->plane_res.scl_data.taps.h_taps_c != temp.h_taps_c))
calculate_inits_and_viewports(pipe_ctx);
/*
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_stream.c b/drivers/gpu/drm/amd/display/dc/core/dc_stream.c
index 9a406d74c0dd..3d93efdc1026 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_stream.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_stream.c
@@ -819,12 +819,12 @@ void dc_stream_log(const struct dc *dc, const struct dc_stream_state *stream)
stream->dst.height,
stream->output_color_space);
DC_LOG_DC(
- "\tpix_clk_khz: %d, h_total: %d, v_total: %d, pixelencoder:%d, displaycolorDepth:%d\n",
+ "\tpix_clk_khz: %d, h_total: %d, v_total: %d, pixel_encoding:%s, color_depth:%s\n",
stream->timing.pix_clk_100hz / 10,
stream->timing.h_total,
stream->timing.v_total,
- stream->timing.pixel_encoding,
- stream->timing.display_color_depth);
+ dc_pixel_encoding_to_str(stream->timing.pixel_encoding),
+ dc_color_depth_to_str(stream->timing.display_color_depth));
DC_LOG_DC(
"\tlink: %d\n",
stream->link->link_index);
diff --git a/drivers/gpu/drm/amd/display/dc/dc.h b/drivers/gpu/drm/amd/display/dc/dc.h
index 3992ad73165b..7c163aa7e8bd 100644
--- a/drivers/gpu/drm/amd/display/dc/dc.h
+++ b/drivers/gpu/drm/amd/display/dc/dc.h
@@ -285,6 +285,7 @@ struct dc_caps {
uint16_t subvp_vertical_int_margin_us;
bool seamless_odm;
uint32_t max_v_total;
+ bool vtotal_limited_by_fp2;
uint32_t max_disp_clock_khz_at_vmin;
uint8_t subvp_drr_vblank_start_margin_us;
bool cursor_not_scaled;
@@ -2543,6 +2544,8 @@ struct dc_power_profile {
struct dc_power_profile dc_get_power_profile_for_dc_state(const struct dc_state *context);
+unsigned int dc_get_det_buffer_size_from_state(const struct dc_state *context);
+
/* DSC Interfaces */
#include "dc_dsc.h"
diff --git a/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c b/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c
index 1e7de0f03290..ec5009f411eb 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c
+++ b/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c
@@ -1294,6 +1294,8 @@ static void dc_dmub_srv_notify_idle(const struct dc *dc, bool allow_idle)
memset(&new_signals, 0, sizeof(new_signals));
+ new_signals.bits.allow_idle = 1; /* always set */
+
if (dc->config.disable_ips == DMUB_IPS_ENABLE ||
dc->config.disable_ips == DMUB_IPS_DISABLE_DYNAMIC) {
new_signals.bits.allow_pg = 1;
@@ -1389,7 +1391,7 @@ static void dc_dmub_srv_exit_low_power_state(const struct dc *dc)
*/
dc_dmub_srv->needs_idle_wake = false;
- if (prev_driver_signals.bits.allow_ips2 &&
+ if ((prev_driver_signals.bits.allow_ips2 || prev_driver_signals.all == 0) &&
(!dc->debug.optimize_ips_handshake ||
ips_fw->signals.bits.ips2_commit || !ips_fw->signals.bits.in_idle)) {
DC_LOG_IPS(
@@ -1450,7 +1452,7 @@ static void dc_dmub_srv_exit_low_power_state(const struct dc *dc)
}
dc_dmub_srv_notify_idle(dc, false);
- if (prev_driver_signals.bits.allow_ips1) {
+ if (prev_driver_signals.bits.allow_ips1 || prev_driver_signals.all == 0) {
DC_LOG_IPS(
"wait for IPS1 commit clear (ips1_commit=%u ips2_commit=%u)",
ips_fw->signals.bits.ips1_commit,
diff --git a/drivers/gpu/drm/amd/display/dc/dio/dcn314/dcn314_dio_stream_encoder.c b/drivers/gpu/drm/amd/display/dc/dio/dcn314/dcn314_dio_stream_encoder.c
index 5b343f745cf3..ae81451a3a72 100644
--- a/drivers/gpu/drm/amd/display/dc/dio/dcn314/dcn314_dio_stream_encoder.c
+++ b/drivers/gpu/drm/amd/display/dc/dio/dcn314/dcn314_dio_stream_encoder.c
@@ -83,6 +83,15 @@ void enc314_disable_fifo(struct stream_encoder *enc)
REG_UPDATE(DIG_FIFO_CTRL0, DIG_FIFO_ENABLE, 0);
}
+static bool enc314_is_fifo_enabled(struct stream_encoder *enc)
+{
+ struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc);
+ uint32_t reset_val;
+
+ REG_GET(DIG_FIFO_CTRL0, DIG_FIFO_ENABLE, &reset_val);
+ return (reset_val != 0);
+}
+
void enc314_dp_set_odm_combine(
struct stream_encoder *enc,
bool odm_combine)
@@ -468,6 +477,7 @@ static const struct stream_encoder_funcs dcn314_str_enc_funcs = {
.enable_fifo = enc314_enable_fifo,
.disable_fifo = enc314_disable_fifo,
+ .is_fifo_enabled = enc314_is_fifo_enabled,
.set_input_mode = enc314_set_dig_input_mode,
};
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/display_mode_core.c b/drivers/gpu/drm/amd/display/dc/dml2/display_mode_core.c
index d851c081e376..8dabb1ac0b68 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/display_mode_core.c
+++ b/drivers/gpu/drm/amd/display/dc/dml2/display_mode_core.c
@@ -1222,6 +1222,7 @@ static dml_bool_t CalculatePrefetchSchedule(struct display_mode_lib_scratch_st *
s->dst_y_prefetch_oto = s->Tvm_oto_lines + 2 * s->Tr0_oto_lines + s->Lsw_oto;
s->dst_y_prefetch_equ = p->VStartup - (*p->TSetup + dml_max(p->TWait + p->TCalc, *p->Tdmdl)) / s->LineTime - (*p->DSTYAfterScaler + (dml_float_t) *p->DSTXAfterScaler / (dml_float_t)p->myPipe->HTotal);
+ s->dst_y_prefetch_equ = dml_min(s->dst_y_prefetch_equ, 63.75); // limit to the reg limit of U6.2 for DST_Y_PREFETCH
#ifdef __DML_VBA_DEBUG__
dml_print("DML::%s: HTotal = %u\n", __func__, p->myPipe->HTotal);
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_translation_helper.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_translation_helper.c
index 8697eac1e1f7..8dee0d397e03 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_translation_helper.c
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_translation_helper.c
@@ -339,11 +339,22 @@ void dml21_apply_soc_bb_overrides(struct dml2_initialize_instance_in_out *dml_in
// }
}
+static unsigned int calc_max_hardware_v_total(const struct dc_stream_state *stream)
+{
+ unsigned int max_hw_v_total = stream->ctx->dc->caps.max_v_total;
+
+ if (stream->ctx->dc->caps.vtotal_limited_by_fp2) {
+ max_hw_v_total -= stream->timing.v_front_porch + 1;
+ }
+
+ return max_hw_v_total;
+}
+
static void populate_dml21_timing_config_from_stream_state(struct dml2_timing_cfg *timing,
struct dc_stream_state *stream,
struct dml2_context *dml_ctx)
{
- unsigned int hblank_start, vblank_start;
+ unsigned int hblank_start, vblank_start, min_hardware_refresh_in_uhz;
timing->h_active = stream->timing.h_addressable + stream->timing.h_border_left + stream->timing.h_border_right;
timing->v_active = stream->timing.v_addressable + stream->timing.v_border_bottom + stream->timing.v_border_top;
@@ -371,11 +382,23 @@ static void populate_dml21_timing_config_from_stream_state(struct dml2_timing_cf
- stream->timing.v_border_top - stream->timing.v_border_bottom;
timing->drr_config.enabled = stream->ignore_msa_timing_param;
- timing->drr_config.min_refresh_uhz = stream->timing.min_refresh_in_uhz;
timing->drr_config.drr_active_variable = stream->vrr_active_variable;
timing->drr_config.drr_active_fixed = stream->vrr_active_fixed;
timing->drr_config.disallowed = !stream->allow_freesync;
+ /* limit min refresh rate to DC cap */
+ min_hardware_refresh_in_uhz = stream->timing.min_refresh_in_uhz;
+ if (stream->ctx->dc->caps.max_v_total != 0) {
+ min_hardware_refresh_in_uhz = div64_u64((stream->timing.pix_clk_100hz * 100000000ULL),
+ (stream->timing.h_total * (long long)calc_max_hardware_v_total(stream)));
+ }
+
+ if (stream->timing.min_refresh_in_uhz > min_hardware_refresh_in_uhz) {
+ timing->drr_config.min_refresh_uhz = stream->timing.min_refresh_in_uhz;
+ } else {
+ timing->drr_config.min_refresh_uhz = min_hardware_refresh_in_uhz;
+ }
+
if (dml_ctx->config.callbacks.get_max_flickerless_instant_vtotal_increase &&
stream->ctx->dc->config.enable_fpo_flicker_detection == 1)
timing->drr_config.max_instant_vtotal_delta = dml_ctx->config.callbacks.get_max_flickerless_instant_vtotal_increase(stream, false);
@@ -859,7 +882,7 @@ static void populate_dml21_plane_config_from_plane_state(struct dml2_context *dm
plane->immediate_flip = plane_state->flip_immediate;
plane->composition.rect_out_height_spans_vactive =
- plane_state->dst_rect.height >= stream->timing.v_addressable &&
+ plane_state->dst_rect.height >= stream->src.height &&
stream->dst.height >= stream->timing.v_addressable;
}
diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn314/dcn314_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn314/dcn314_hwseq.c
index 4e93eeedfc1b..efcc1a6b364c 100644
--- a/drivers/gpu/drm/amd/display/dc/hwss/dcn314/dcn314_hwseq.c
+++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn314/dcn314_hwseq.c
@@ -355,6 +355,20 @@ void dcn314_calculate_pix_rate_divider(
}
}
+static bool dcn314_is_pipe_dig_fifo_on(struct pipe_ctx *pipe)
+{
+ return pipe && pipe->stream
+ // Check dig's otg instance.
+ && pipe->stream_res.stream_enc
+ && pipe->stream_res.stream_enc->funcs->dig_source_otg
+ && pipe->stream_res.tg->inst == pipe->stream_res.stream_enc->funcs->dig_source_otg(pipe->stream_res.stream_enc)
+ && pipe->stream->link && pipe->stream->link->link_enc
+ && pipe->stream->link->link_enc->funcs->is_dig_enabled
+ && pipe->stream->link->link_enc->funcs->is_dig_enabled(pipe->stream->link->link_enc)
+ && pipe->stream_res.stream_enc->funcs->is_fifo_enabled
+ && pipe->stream_res.stream_enc->funcs->is_fifo_enabled(pipe->stream_res.stream_enc);
+}
+
void dcn314_resync_fifo_dccg_dio(struct dce_hwseq *hws, struct dc *dc, struct dc_state *context, unsigned int current_pipe_idx)
{
unsigned int i;
@@ -371,7 +385,11 @@ void dcn314_resync_fifo_dccg_dio(struct dce_hwseq *hws, struct dc *dc, struct dc
if (pipe->top_pipe || pipe->prev_odm_pipe)
continue;
- if (pipe->stream && (pipe->stream->dpms_off || dc_is_virtual_signal(pipe->stream->signal))) {
+ if (pipe->stream && (pipe->stream->dpms_off || dc_is_virtual_signal(pipe->stream->signal)) &&
+ !pipe->stream->apply_seamless_boot_optimization &&
+ !pipe->stream->apply_edp_fast_boot_optimization) {
+ if (dcn314_is_pipe_dig_fifo_on(pipe))
+ continue;
pipe->stream_res.tg->funcs->disable_crtc(pipe->stream_res.tg);
reset_sync_context_for_pipe(dc, context, i);
otg_disabled[i] = true;
diff --git a/drivers/gpu/drm/amd/display/dc/inc/core_status.h b/drivers/gpu/drm/amd/display/dc/inc/core_status.h
index fa5edd03d004..b5afd8c3103d 100644
--- a/drivers/gpu/drm/amd/display/dc/inc/core_status.h
+++ b/drivers/gpu/drm/amd/display/dc/inc/core_status.h
@@ -60,5 +60,7 @@ enum dc_status {
};
char *dc_status_to_str(enum dc_status status);
+char *dc_pixel_encoding_to_str(enum dc_pixel_encoding pixel_encoding);
+char *dc_color_depth_to_str(enum dc_color_depth color_depth);
#endif /* _CORE_STATUS_H_ */
diff --git a/drivers/gpu/drm/amd/display/dc/inc/core_types.h b/drivers/gpu/drm/amd/display/dc/inc/core_types.h
index bfb8b8502d20..e1e3142cdc00 100644
--- a/drivers/gpu/drm/amd/display/dc/inc/core_types.h
+++ b/drivers/gpu/drm/amd/display/dc/inc/core_types.h
@@ -215,6 +215,7 @@ struct resource_funcs {
void (*get_panel_config_defaults)(struct dc_panel_config *panel_config);
void (*build_pipe_pix_clk_params)(struct pipe_ctx *pipe_ctx);
+ unsigned int (*get_det_buffer_size)(const struct dc_state *context);
};
struct audio_support{
diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn20/dcn20_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn20/dcn20_resource.c
index eea2b3b307cd..45e4de8d5cff 100644
--- a/drivers/gpu/drm/amd/display/dc/resource/dcn20/dcn20_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/resource/dcn20/dcn20_resource.c
@@ -1511,6 +1511,7 @@ bool dcn20_split_stream_for_odm(
if (prev_odm_pipe->plane_state) {
struct scaler_data *sd = &prev_odm_pipe->plane_res.scl_data;
+ struct output_pixel_processor *opp = next_odm_pipe->stream_res.opp;
int new_width;
/* HACTIVE halved for odm combine */
@@ -1544,7 +1545,28 @@ bool dcn20_split_stream_for_odm(
sd->viewport_c.x += dc_fixpt_floor(dc_fixpt_mul_int(
sd->ratios.horz_c, sd->h_active - sd->recout.x));
sd->recout.x = 0;
+
+ /*
+ * When odm is used in YcbCr422 or 420 colour space, a split screen
+ * will be seen with the previous calculations since the extra left
+ * edge pixel is accounted for in fmt but not in viewport.
+ *
+ * Below are calculations which fix the split by fixing the calculations
+ * if there is an extra left edge pixel.
+ */
+ if (opp && opp->funcs->opp_get_left_edge_extra_pixel_count
+ && opp->funcs->opp_get_left_edge_extra_pixel_count(
+ opp, next_odm_pipe->stream->timing.pixel_encoding,
+ resource_is_pipe_type(next_odm_pipe, OTG_MASTER)) == 1) {
+ sd->h_active += 1;
+ sd->recout.width += 1;
+ sd->viewport.x -= dc_fixpt_ceil(dc_fixpt_mul_int(sd->ratios.horz, 1));
+ sd->viewport_c.x -= dc_fixpt_ceil(dc_fixpt_mul_int(sd->ratios.horz, 1));
+ sd->viewport_c.width += dc_fixpt_ceil(dc_fixpt_mul_int(sd->ratios.horz, 1));
+ sd->viewport.width += dc_fixpt_ceil(dc_fixpt_mul_int(sd->ratios.horz, 1));
+ }
}
+
if (!next_odm_pipe->top_pipe)
next_odm_pipe->stream_res.opp = pool->opps[next_odm_pipe->pipe_idx];
else
@@ -2133,6 +2155,7 @@ bool dcn20_fast_validate_bw(
ASSERT(0);
}
}
+
/* Actual dsc count per stream dsc validation*/
if (!dcn20_validate_dsc(dc, context)) {
context->bw_ctx.dml.vba.ValidationStatus[context->bw_ctx.dml.vba.soc.num_states] =
diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn21/dcn21_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn21/dcn21_resource.c
index 347e6aaea582..14b28841657d 100644
--- a/drivers/gpu/drm/amd/display/dc/resource/dcn21/dcn21_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/resource/dcn21/dcn21_resource.c
@@ -1298,7 +1298,7 @@ static struct link_encoder *dcn21_link_encoder_create(
kzalloc(sizeof(struct dcn21_link_encoder), GFP_KERNEL);
int link_regs_id;
- if (!enc21)
+ if (!enc21 || enc_init_data->hpd_source >= ARRAY_SIZE(link_enc_hpd_regs))
return NULL;
link_regs_id =
diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn30/dcn30_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn30/dcn30_resource.c
index 5040a4c6ed18..75cc84473a57 100644
--- a/drivers/gpu/drm/amd/display/dc/resource/dcn30/dcn30_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/resource/dcn30/dcn30_resource.c
@@ -2354,6 +2354,7 @@ static bool dcn30_resource_construct(
dc->caps.dp_hdmi21_pcon_support = true;
dc->caps.max_v_total = (1 << 15) - 1;
+ dc->caps.vtotal_limited_by_fp2 = true;
/* read VBIOS LTTPR caps */
{
diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn302/dcn302_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn302/dcn302_resource.c
index 5791b5cc2875..320b040d591d 100644
--- a/drivers/gpu/drm/amd/display/dc/resource/dcn302/dcn302_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/resource/dcn302/dcn302_resource.c
@@ -1234,6 +1234,7 @@ static bool dcn302_resource_construct(
dc->caps.extended_aux_timeout_support = true;
dc->caps.dmcub_support = true;
dc->caps.max_v_total = (1 << 15) - 1;
+ dc->caps.vtotal_limited_by_fp2 = true;
/* Color pipeline capabilities */
dc->caps.color.dpp.dcn_arch = 1;
diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn303/dcn303_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn303/dcn303_resource.c
index 63f0f882c861..297cf4b5600d 100644
--- a/drivers/gpu/drm/amd/display/dc/resource/dcn303/dcn303_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/resource/dcn303/dcn303_resource.c
@@ -1179,6 +1179,7 @@ static bool dcn303_resource_construct(
dc->caps.extended_aux_timeout_support = true;
dc->caps.dmcub_support = true;
dc->caps.max_v_total = (1 << 15) - 1;
+ dc->caps.vtotal_limited_by_fp2 = true;
/* Color pipeline capabilities */
dc->caps.color.dpp.dcn_arch = 1;
diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn31/dcn31_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn31/dcn31_resource.c
index ac8cb20e2e3b..80386f698ae4 100644
--- a/drivers/gpu/drm/amd/display/dc/resource/dcn31/dcn31_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/resource/dcn31/dcn31_resource.c
@@ -1721,6 +1721,12 @@ int dcn31_populate_dml_pipes_from_context(
return pipe_cnt;
}
+unsigned int dcn31_get_det_buffer_size(
+ const struct dc_state *context)
+{
+ return context->bw_ctx.dml.ip.det_buffer_size_kbytes;
+}
+
void dcn31_calculate_wm_and_dlg(
struct dc *dc, struct dc_state *context,
display_e2e_pipe_params_st *pipes,
@@ -1843,6 +1849,7 @@ static struct resource_funcs dcn31_res_pool_funcs = {
.update_bw_bounding_box = dcn31_update_bw_bounding_box,
.patch_unknown_plane_state = dcn20_patch_unknown_plane_state,
.get_panel_config_defaults = dcn31_get_panel_config_defaults,
+ .get_det_buffer_size = dcn31_get_det_buffer_size,
};
static struct clock_source *dcn30_clock_source_create(
diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn31/dcn31_resource.h b/drivers/gpu/drm/amd/display/dc/resource/dcn31/dcn31_resource.h
index 901436591ed4..551ad912f7be 100644
--- a/drivers/gpu/drm/amd/display/dc/resource/dcn31/dcn31_resource.h
+++ b/drivers/gpu/drm/amd/display/dc/resource/dcn31/dcn31_resource.h
@@ -63,6 +63,9 @@ struct resource_pool *dcn31_create_resource_pool(
const struct dc_init_data *init_data,
struct dc *dc);
+unsigned int dcn31_get_det_buffer_size(
+ const struct dc_state *context);
+
/*temp: B0 specific before switch to dcn313 headers*/
#ifndef regPHYPLLF_PIXCLK_RESYNC_CNTL
#define regPHYPLLF_PIXCLK_RESYNC_CNTL 0x007e
diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn314/dcn314_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn314/dcn314_resource.c
index 169924d0a839..01d95108ce66 100644
--- a/drivers/gpu/drm/amd/display/dc/resource/dcn314/dcn314_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/resource/dcn314/dcn314_resource.c
@@ -1778,6 +1778,7 @@ static struct resource_funcs dcn314_res_pool_funcs = {
.patch_unknown_plane_state = dcn20_patch_unknown_plane_state,
.get_panel_config_defaults = dcn314_get_panel_config_defaults,
.get_preferred_eng_id_dpia = dcn314_get_preferred_eng_id_dpia,
+ .get_det_buffer_size = dcn31_get_det_buffer_size,
};
static struct clock_source *dcn30_clock_source_create(
diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn315/dcn315_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn315/dcn315_resource.c
index 3f4b9dba4112..f2ce687c0e03 100644
--- a/drivers/gpu/drm/amd/display/dc/resource/dcn315/dcn315_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/resource/dcn315/dcn315_resource.c
@@ -1840,6 +1840,7 @@ static struct resource_funcs dcn315_res_pool_funcs = {
.update_bw_bounding_box = dcn315_update_bw_bounding_box,
.patch_unknown_plane_state = dcn20_patch_unknown_plane_state,
.get_panel_config_defaults = dcn315_get_panel_config_defaults,
+ .get_det_buffer_size = dcn31_get_det_buffer_size,
};
static bool dcn315_resource_construct(
diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn316/dcn316_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn316/dcn316_resource.c
index 5fd52c5fcee4..af82e13029c9 100644
--- a/drivers/gpu/drm/amd/display/dc/resource/dcn316/dcn316_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/resource/dcn316/dcn316_resource.c
@@ -1720,6 +1720,7 @@ static struct resource_funcs dcn316_res_pool_funcs = {
.update_bw_bounding_box = dcn316_update_bw_bounding_box,
.patch_unknown_plane_state = dcn20_patch_unknown_plane_state,
.get_panel_config_defaults = dcn316_get_panel_config_defaults,
+ .get_det_buffer_size = dcn31_get_det_buffer_size,
};
static bool dcn316_resource_construct(
diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn32/dcn32_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn32/dcn32_resource.c
index a124ad9bd108..6b889c8be0ca 100644
--- a/drivers/gpu/drm/amd/display/dc/resource/dcn32/dcn32_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/resource/dcn32/dcn32_resource.c
@@ -2186,6 +2186,7 @@ static bool dcn32_resource_construct(
dc->caps.dmcub_support = true;
dc->caps.seamless_odm = true;
dc->caps.max_v_total = (1 << 15) - 1;
+ dc->caps.vtotal_limited_by_fp2 = true;
/* Color pipeline capabilities */
dc->caps.color.dpp.dcn_arch = 1;
diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn321/dcn321_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn321/dcn321_resource.c
index 827a94f84f10..74113c578bac 100644
--- a/drivers/gpu/drm/amd/display/dc/resource/dcn321/dcn321_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/resource/dcn321/dcn321_resource.c
@@ -1743,6 +1743,7 @@ static bool dcn321_resource_construct(
dc->caps.extended_aux_timeout_support = true;
dc->caps.dmcub_support = true;
dc->caps.max_v_total = (1 << 15) - 1;
+ dc->caps.vtotal_limited_by_fp2 = true;
/* Color pipeline capabilities */
dc->caps.color.dpp.dcn_arch = 1;
diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn35/dcn35_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn35/dcn35_resource.c
index 893a9d9ee870..d0c4693c1224 100644
--- a/drivers/gpu/drm/amd/display/dc/resource/dcn35/dcn35_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/resource/dcn35/dcn35_resource.c
@@ -1779,6 +1779,7 @@ static struct resource_funcs dcn35_res_pool_funcs = {
.patch_unknown_plane_state = dcn20_patch_unknown_plane_state,
.get_panel_config_defaults = dcn35_get_panel_config_defaults,
.get_preferred_eng_id_dpia = dcn35_get_preferred_eng_id_dpia,
+ .get_det_buffer_size = dcn31_get_det_buffer_size,
};
static bool dcn35_resource_construct(
@@ -1850,6 +1851,7 @@ static bool dcn35_resource_construct(
dc->caps.zstate_support = true;
dc->caps.ips_support = true;
dc->caps.max_v_total = (1 << 15) - 1;
+ dc->caps.vtotal_limited_by_fp2 = true;
/* Color pipeline capabilities */
dc->caps.color.dpp.dcn_arch = 1;
diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn351/dcn351_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn351/dcn351_resource.c
index 70abd32ce2ad..575c0aa12229 100644
--- a/drivers/gpu/drm/amd/display/dc/resource/dcn351/dcn351_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/resource/dcn351/dcn351_resource.c
@@ -1758,6 +1758,7 @@ static struct resource_funcs dcn351_res_pool_funcs = {
.patch_unknown_plane_state = dcn20_patch_unknown_plane_state,
.get_panel_config_defaults = dcn35_get_panel_config_defaults,
.get_preferred_eng_id_dpia = dcn351_get_preferred_eng_id_dpia,
+ .get_det_buffer_size = dcn31_get_det_buffer_size,
};
static bool dcn351_resource_construct(
@@ -1829,6 +1830,7 @@ static bool dcn351_resource_construct(
dc->caps.zstate_support = true;
dc->caps.ips_support = true;
dc->caps.max_v_total = (1 << 15) - 1;
+ dc->caps.vtotal_limited_by_fp2 = true;
/* Color pipeline capabilities */
dc->caps.color.dpp.dcn_arch = 1;
diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn401/dcn401_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn401/dcn401_resource.c
index 9d56fbdcd06a..4aa975418fb1 100644
--- a/drivers/gpu/drm/amd/display/dc/resource/dcn401/dcn401_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/resource/dcn401/dcn401_resource.c
@@ -1826,6 +1826,7 @@ static bool dcn401_resource_construct(
dc->caps.extended_aux_timeout_support = true;
dc->caps.dmcub_support = true;
dc->caps.max_v_total = (1 << 15) - 1;
+ dc->caps.vtotal_limited_by_fp2 = true;
if (ASICREV_IS_GC_12_0_1_A0(dc->ctx->asic_id.hw_internal_rev))
dc->caps.dcc_plane_width_limit = 7680;
diff --git a/drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h b/drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h
index ebcf68bfae2b..7835100b37c4 100644
--- a/drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h
+++ b/drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h
@@ -747,7 +747,8 @@ union dmub_shared_state_ips_driver_signals {
uint32_t allow_ips1 : 1; /**< 1 is IPS1 is allowed */
uint32_t allow_ips2 : 1; /**< 1 is IPS1 is allowed */
uint32_t allow_z10 : 1; /**< 1 if Z10 is allowed */
- uint32_t reserved_bits : 28; /**< Reversed bits */
+ uint32_t allow_idle : 1; /**< 1 if driver is allowing idle */
+ uint32_t reserved_bits : 27; /**< Reversed bits */
} bits;
uint32_t all;
};
diff --git a/drivers/gpu/drm/amd/display/modules/freesync/freesync.c b/drivers/gpu/drm/amd/display/modules/freesync/freesync.c
index bbd259cea4f4..ab62a76d48cf 100644
--- a/drivers/gpu/drm/amd/display/modules/freesync/freesync.c
+++ b/drivers/gpu/drm/amd/display/modules/freesync/freesync.c
@@ -121,6 +121,17 @@ static unsigned int calc_duration_in_us_from_v_total(
return duration_in_us;
}
+static unsigned int calc_max_hardware_v_total(const struct dc_stream_state *stream)
+{
+ unsigned int max_hw_v_total = stream->ctx->dc->caps.max_v_total;
+
+ if (stream->ctx->dc->caps.vtotal_limited_by_fp2) {
+ max_hw_v_total -= stream->timing.v_front_porch + 1;
+ }
+
+ return max_hw_v_total;
+}
+
unsigned int mod_freesync_calc_v_total_from_refresh(
const struct dc_stream_state *stream,
unsigned int refresh_in_uhz)
@@ -1002,7 +1013,7 @@ void mod_freesync_build_vrr_params(struct mod_freesync *mod_freesync,
if (stream->ctx->dc->caps.max_v_total != 0 && stream->timing.h_total != 0) {
min_hardware_refresh_in_uhz = div64_u64((stream->timing.pix_clk_100hz * 100000000ULL),
- (stream->timing.h_total * (long long)stream->ctx->dc->caps.max_v_total));
+ (stream->timing.h_total * (long long)calc_max_hardware_v_total(stream)));
}
/* Limit minimum refresh rate to what can be supported by hardware */
min_refresh_in_uhz = min_hardware_refresh_in_uhz > in_config->min_refresh_in_uhz ?
diff --git a/drivers/gpu/drm/amd/pm/amdgpu_pm.c b/drivers/gpu/drm/amd/pm/amdgpu_pm.c
index d5d6ab484e5a..0fa6fbee1978 100644
--- a/drivers/gpu/drm/amd/pm/amdgpu_pm.c
+++ b/drivers/gpu/drm/amd/pm/amdgpu_pm.c
@@ -1409,7 +1409,11 @@ static ssize_t amdgpu_set_pp_mclk_od(struct device *dev,
* create a custom set of heuristics, write a string of numbers to the file
* starting with the number of the custom profile along with a setting
* for each heuristic parameter. Due to differences across asic families
- * the heuristic parameters vary from family to family.
+ * the heuristic parameters vary from family to family. Additionally,
+ * you can apply the custom heuristics to different clock domains. Each
+ * clock domain is considered a distinct operation so if you modify the
+ * gfxclk heuristics and then the memclk heuristics, the all of the
+ * custom heuristics will be retained until you switch to another profile.
*
*/
diff --git a/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c b/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c
index 32bdeac2676b..0c0b9aa44dfa 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c
@@ -72,6 +72,10 @@ static int smu_set_power_limit(void *handle, uint32_t limit);
static int smu_set_fan_speed_rpm(void *handle, uint32_t speed);
static int smu_set_gfx_cgpg(struct smu_context *smu, bool enabled);
static int smu_set_mp1_state(void *handle, enum pp_mp1_state mp1_state);
+static void smu_power_profile_mode_get(struct smu_context *smu,
+ enum PP_SMC_POWER_PROFILE profile_mode);
+static void smu_power_profile_mode_put(struct smu_context *smu,
+ enum PP_SMC_POWER_PROFILE profile_mode);
static int smu_sys_get_pp_feature_mask(void *handle,
char *buf)
@@ -1257,35 +1261,19 @@ static int smu_sw_init(void *handle)
INIT_WORK(&smu->interrupt_work, smu_interrupt_work_fn);
atomic64_set(&smu->throttle_int_counter, 0);
smu->watermarks_bitmap = 0;
- smu->power_profile_mode = PP_SMC_POWER_PROFILE_BOOTUP_DEFAULT;
- smu->default_power_profile_mode = PP_SMC_POWER_PROFILE_BOOTUP_DEFAULT;
atomic_set(&smu->smu_power.power_gate.vcn_gated, 1);
atomic_set(&smu->smu_power.power_gate.jpeg_gated, 1);
atomic_set(&smu->smu_power.power_gate.vpe_gated, 1);
atomic_set(&smu->smu_power.power_gate.umsch_mm_gated, 1);
- smu->workload_prority[PP_SMC_POWER_PROFILE_BOOTUP_DEFAULT] = 0;
- smu->workload_prority[PP_SMC_POWER_PROFILE_FULLSCREEN3D] = 1;
- smu->workload_prority[PP_SMC_POWER_PROFILE_POWERSAVING] = 2;
- smu->workload_prority[PP_SMC_POWER_PROFILE_VIDEO] = 3;
- smu->workload_prority[PP_SMC_POWER_PROFILE_VR] = 4;
- smu->workload_prority[PP_SMC_POWER_PROFILE_COMPUTE] = 5;
- smu->workload_prority[PP_SMC_POWER_PROFILE_CUSTOM] = 6;
-
if (smu->is_apu ||
!smu_is_workload_profile_available(smu, PP_SMC_POWER_PROFILE_FULLSCREEN3D))
- smu->workload_mask = 1 << smu->workload_prority[PP_SMC_POWER_PROFILE_BOOTUP_DEFAULT];
+ smu->power_profile_mode = PP_SMC_POWER_PROFILE_BOOTUP_DEFAULT;
else
- smu->workload_mask = 1 << smu->workload_prority[PP_SMC_POWER_PROFILE_FULLSCREEN3D];
-
- smu->workload_setting[0] = PP_SMC_POWER_PROFILE_BOOTUP_DEFAULT;
- smu->workload_setting[1] = PP_SMC_POWER_PROFILE_FULLSCREEN3D;
- smu->workload_setting[2] = PP_SMC_POWER_PROFILE_POWERSAVING;
- smu->workload_setting[3] = PP_SMC_POWER_PROFILE_VIDEO;
- smu->workload_setting[4] = PP_SMC_POWER_PROFILE_VR;
- smu->workload_setting[5] = PP_SMC_POWER_PROFILE_COMPUTE;
- smu->workload_setting[6] = PP_SMC_POWER_PROFILE_CUSTOM;
+ smu->power_profile_mode = PP_SMC_POWER_PROFILE_FULLSCREEN3D;
+ smu_power_profile_mode_get(smu, smu->power_profile_mode);
+
smu->display_config = &adev->pm.pm_display_cfg;
smu->smu_dpm.dpm_level = AMD_DPM_FORCED_LEVEL_AUTO;
@@ -1338,6 +1326,11 @@ static int smu_sw_fini(void *handle)
return ret;
}
+ if (smu->custom_profile_params) {
+ kfree(smu->custom_profile_params);
+ smu->custom_profile_params = NULL;
+ }
+
smu_fini_microcode(smu);
return 0;
@@ -2117,6 +2110,9 @@ static int smu_suspend(void *handle)
if (!ret)
adev->gfx.gfx_off_entrycount = count;
+ /* clear this on suspend so it will get reprogrammed on resume */
+ smu->workload_mask = 0;
+
return 0;
}
@@ -2229,25 +2225,49 @@ static int smu_enable_umd_pstate(void *handle,
}
static int smu_bump_power_profile_mode(struct smu_context *smu,
- long *param,
- uint32_t param_size)
+ long *custom_params,
+ u32 custom_params_max_idx)
{
- int ret = 0;
+ u32 workload_mask = 0;
+ int i, ret = 0;
+
+ for (i = 0; i < PP_SMC_POWER_PROFILE_COUNT; i++) {
+ if (smu->workload_refcount[i])
+ workload_mask |= 1 << i;
+ }
+
+ if (smu->workload_mask == workload_mask)
+ return 0;
if (smu->ppt_funcs->set_power_profile_mode)
- ret = smu->ppt_funcs->set_power_profile_mode(smu, param, param_size);
+ ret = smu->ppt_funcs->set_power_profile_mode(smu, workload_mask,
+ custom_params,
+ custom_params_max_idx);
+
+ if (!ret)
+ smu->workload_mask = workload_mask;
return ret;
}
+static void smu_power_profile_mode_get(struct smu_context *smu,
+ enum PP_SMC_POWER_PROFILE profile_mode)
+{
+ smu->workload_refcount[profile_mode]++;
+}
+
+static void smu_power_profile_mode_put(struct smu_context *smu,
+ enum PP_SMC_POWER_PROFILE profile_mode)
+{
+ if (smu->workload_refcount[profile_mode])
+ smu->workload_refcount[profile_mode]--;
+}
+
static int smu_adjust_power_state_dynamic(struct smu_context *smu,
enum amd_dpm_forced_level level,
- bool skip_display_settings,
- bool init)
+ bool skip_display_settings)
{
int ret = 0;
- int index = 0;
- long workload[1];
struct smu_dpm_context *smu_dpm_ctx = &(smu->smu_dpm);
if (!skip_display_settings) {
@@ -2284,14 +2304,8 @@ static int smu_adjust_power_state_dynamic(struct smu_context *smu,
}
if (smu_dpm_ctx->dpm_level != AMD_DPM_FORCED_LEVEL_MANUAL &&
- smu_dpm_ctx->dpm_level != AMD_DPM_FORCED_LEVEL_PERF_DETERMINISM) {
- index = fls(smu->workload_mask);
- index = index > 0 && index <= WORKLOAD_POLICY_MAX ? index - 1 : 0;
- workload[0] = smu->workload_setting[index];
-
- if (init || smu->power_profile_mode != workload[0])
- smu_bump_power_profile_mode(smu, workload, 0);
- }
+ smu_dpm_ctx->dpm_level != AMD_DPM_FORCED_LEVEL_PERF_DETERMINISM)
+ smu_bump_power_profile_mode(smu, NULL, 0);
return ret;
}
@@ -2310,13 +2324,13 @@ static int smu_handle_task(struct smu_context *smu,
ret = smu_pre_display_config_changed(smu);
if (ret)
return ret;
- ret = smu_adjust_power_state_dynamic(smu, level, false, false);
+ ret = smu_adjust_power_state_dynamic(smu, level, false);
break;
case AMD_PP_TASK_COMPLETE_INIT:
- ret = smu_adjust_power_state_dynamic(smu, level, true, true);
+ ret = smu_adjust_power_state_dynamic(smu, level, true);
break;
case AMD_PP_TASK_READJUST_POWER_STATE:
- ret = smu_adjust_power_state_dynamic(smu, level, true, false);
+ ret = smu_adjust_power_state_dynamic(smu, level, true);
break;
default:
break;
@@ -2338,12 +2352,11 @@ static int smu_handle_dpm_task(void *handle,
static int smu_switch_power_profile(void *handle,
enum PP_SMC_POWER_PROFILE type,
- bool en)
+ bool enable)
{
struct smu_context *smu = handle;
struct smu_dpm_context *smu_dpm_ctx = &(smu->smu_dpm);
- long workload[1];
- uint32_t index;
+ int ret;
if (!smu->pm_enabled || !smu->adev->pm.dpm_enabled)
return -EOPNOTSUPP;
@@ -2351,21 +2364,21 @@ static int smu_switch_power_profile(void *handle,
if (!(type < PP_SMC_POWER_PROFILE_CUSTOM))
return -EINVAL;
- if (!en) {
- smu->workload_mask &= ~(1 << smu->workload_prority[type]);
- index = fls(smu->workload_mask);
- index = index > 0 && index <= WORKLOAD_POLICY_MAX ? index - 1 : 0;
- workload[0] = smu->workload_setting[index];
- } else {
- smu->workload_mask |= (1 << smu->workload_prority[type]);
- index = fls(smu->workload_mask);
- index = index <= WORKLOAD_POLICY_MAX ? index - 1 : 0;
- workload[0] = smu->workload_setting[index];
- }
-
if (smu_dpm_ctx->dpm_level != AMD_DPM_FORCED_LEVEL_MANUAL &&
- smu_dpm_ctx->dpm_level != AMD_DPM_FORCED_LEVEL_PERF_DETERMINISM)
- smu_bump_power_profile_mode(smu, workload, 0);
+ smu_dpm_ctx->dpm_level != AMD_DPM_FORCED_LEVEL_PERF_DETERMINISM) {
+ if (enable)
+ smu_power_profile_mode_get(smu, type);
+ else
+ smu_power_profile_mode_put(smu, type);
+ ret = smu_bump_power_profile_mode(smu, NULL, 0);
+ if (ret) {
+ if (enable)
+ smu_power_profile_mode_put(smu, type);
+ else
+ smu_power_profile_mode_get(smu, type);
+ return ret;
+ }
+ }
return 0;
}
@@ -3053,12 +3066,35 @@ static int smu_set_power_profile_mode(void *handle,
uint32_t param_size)
{
struct smu_context *smu = handle;
+ bool custom = false;
+ int ret = 0;
if (!smu->pm_enabled || !smu->adev->pm.dpm_enabled ||
!smu->ppt_funcs->set_power_profile_mode)
return -EOPNOTSUPP;
- return smu_bump_power_profile_mode(smu, param, param_size);
+ if (param[param_size] == PP_SMC_POWER_PROFILE_CUSTOM) {
+ custom = true;
+ /* clear frontend mask so custom changes propogate */
+ smu->workload_mask = 0;
+ }
+
+ if ((param[param_size] != smu->power_profile_mode) || custom) {
+ /* clear the old user preference */
+ smu_power_profile_mode_put(smu, smu->power_profile_mode);
+ /* set the new user preference */
+ smu_power_profile_mode_get(smu, param[param_size]);
+ ret = smu_bump_power_profile_mode(smu,
+ custom ? param : NULL,
+ custom ? param_size : 0);
+ if (ret)
+ smu_power_profile_mode_put(smu, param[param_size]);
+ else
+ /* store the user's preference */
+ smu->power_profile_mode = param[param_size];
+ }
+
+ return ret;
}
static int smu_get_fan_control_mode(void *handle, u32 *fan_mode)
diff --git a/drivers/gpu/drm/amd/pm/swsmu/inc/amdgpu_smu.h b/drivers/gpu/drm/amd/pm/swsmu/inc/amdgpu_smu.h
index b44a185d07e8..2b8a18ce25d9 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/inc/amdgpu_smu.h
+++ b/drivers/gpu/drm/amd/pm/swsmu/inc/amdgpu_smu.h
@@ -556,11 +556,13 @@ struct smu_context {
uint32_t hard_min_uclk_req_from_dal;
bool disable_uclk_switch;
+ /* asic agnostic workload mask */
uint32_t workload_mask;
- uint32_t workload_prority[WORKLOAD_POLICY_MAX];
- uint32_t workload_setting[WORKLOAD_POLICY_MAX];
+ /* default/user workload preference */
uint32_t power_profile_mode;
- uint32_t default_power_profile_mode;
+ uint32_t workload_refcount[PP_SMC_POWER_PROFILE_COUNT];
+ /* backend specific custom workload settings */
+ long *custom_profile_params;
bool pm_enabled;
bool is_apu;
@@ -731,9 +733,12 @@ struct pptable_funcs {
* @set_power_profile_mode: Set a power profile mode. Also used to
* create/set custom power profile modes.
* &input: Power profile mode parameters.
- * &size: Size of &input.
+ * &workload_mask: mask of workloads to enable
+ * &custom_params: custom profile parameters
+ * &custom_params_max_idx: max valid idx into custom_params
*/
- int (*set_power_profile_mode)(struct smu_context *smu, long *input, uint32_t size);
+ int (*set_power_profile_mode)(struct smu_context *smu, u32 workload_mask,
+ long *custom_params, u32 custom_params_max_idx);
/**
* @dpm_set_vcn_enable: Enable/disable VCN engine dynamic power
diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu11/arcturus_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu11/arcturus_ppt.c
index d52512f5f1bd..fc1297fecc62 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/smu11/arcturus_ppt.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu11/arcturus_ppt.c
@@ -1445,98 +1445,120 @@ static int arcturus_get_power_profile_mode(struct smu_context *smu,
return size;
}
-static int arcturus_set_power_profile_mode(struct smu_context *smu,
- long *input,
- uint32_t size)
+#define ARCTURUS_CUSTOM_PARAMS_COUNT 10
+#define ARCTURUS_CUSTOM_PARAMS_CLOCK_COUNT 2
+#define ARCTURUS_CUSTOM_PARAMS_SIZE (ARCTURUS_CUSTOM_PARAMS_CLOCK_COUNT * ARCTURUS_CUSTOM_PARAMS_COUNT * sizeof(long))
+
+static int arcturus_set_power_profile_mode_coeff(struct smu_context *smu,
+ long *input)
{
DpmActivityMonitorCoeffInt_t activity_monitor;
- int workload_type = 0;
- uint32_t profile_mode = input[size];
- int ret = 0;
+ int ret, idx;
- if (profile_mode > PP_SMC_POWER_PROFILE_CUSTOM) {
- dev_err(smu->adev->dev, "Invalid power profile mode %d\n", profile_mode);
- return -EINVAL;
+ ret = smu_cmn_update_table(smu,
+ SMU_TABLE_ACTIVITY_MONITOR_COEFF,
+ WORKLOAD_PPLIB_CUSTOM_BIT,
+ (void *)(&activity_monitor),
+ false);
+ if (ret) {
+ dev_err(smu->adev->dev, "[%s] Failed to get activity monitor!", __func__);
+ return ret;
}
+ idx = 0 * ARCTURUS_CUSTOM_PARAMS_COUNT;
+ if (input[idx]) {
+ /* Gfxclk */
+ activity_monitor.Gfx_FPS = input[idx + 1];
+ activity_monitor.Gfx_UseRlcBusy = input[idx + 2];
+ activity_monitor.Gfx_MinActiveFreqType = input[idx + 3];
+ activity_monitor.Gfx_MinActiveFreq = input[idx + 4];
+ activity_monitor.Gfx_BoosterFreqType = input[idx + 5];
+ activity_monitor.Gfx_BoosterFreq = input[idx + 6];
+ activity_monitor.Gfx_PD_Data_limit_c = input[idx + 7];
+ activity_monitor.Gfx_PD_Data_error_coeff = input[idx + 8];
+ activity_monitor.Gfx_PD_Data_error_rate_coeff = input[idx + 9];
+ }
+ idx = 1 * ARCTURUS_CUSTOM_PARAMS_COUNT;
+ if (input[idx]) {
+ /* Uclk */
+ activity_monitor.Mem_FPS = input[idx + 1];
+ activity_monitor.Mem_UseRlcBusy = input[idx + 2];
+ activity_monitor.Mem_MinActiveFreqType = input[idx + 3];
+ activity_monitor.Mem_MinActiveFreq = input[idx + 4];
+ activity_monitor.Mem_BoosterFreqType = input[idx + 5];
+ activity_monitor.Mem_BoosterFreq = input[idx + 6];
+ activity_monitor.Mem_PD_Data_limit_c = input[idx + 7];
+ activity_monitor.Mem_PD_Data_error_coeff = input[idx + 8];
+ activity_monitor.Mem_PD_Data_error_rate_coeff = input[idx + 9];
+ }
- if ((profile_mode == PP_SMC_POWER_PROFILE_CUSTOM) &&
- (smu->smc_fw_version >= 0x360d00)) {
- if (size != 10)
- return -EINVAL;
+ ret = smu_cmn_update_table(smu,
+ SMU_TABLE_ACTIVITY_MONITOR_COEFF,
+ WORKLOAD_PPLIB_CUSTOM_BIT,
+ (void *)(&activity_monitor),
+ true);
+ if (ret) {
+ dev_err(smu->adev->dev, "[%s] Failed to set activity monitor!", __func__);
+ return ret;
+ }
- ret = smu_cmn_update_table(smu,
- SMU_TABLE_ACTIVITY_MONITOR_COEFF,
- WORKLOAD_PPLIB_CUSTOM_BIT,
- (void *)(&activity_monitor),
- false);
- if (ret) {
- dev_err(smu->adev->dev, "[%s] Failed to get activity monitor!", __func__);
- return ret;
- }
+ return ret;
+}
- switch (input[0]) {
- case 0: /* Gfxclk */
- activity_monitor.Gfx_FPS = input[1];
- activity_monitor.Gfx_UseRlcBusy = input[2];
- activity_monitor.Gfx_MinActiveFreqType = input[3];
- activity_monitor.Gfx_MinActiveFreq = input[4];
- activity_monitor.Gfx_BoosterFreqType = input[5];
- activity_monitor.Gfx_BoosterFreq = input[6];
- activity_monitor.Gfx_PD_Data_limit_c = input[7];
- activity_monitor.Gfx_PD_Data_error_coeff = input[8];
- activity_monitor.Gfx_PD_Data_error_rate_coeff = input[9];
- break;
- case 1: /* Uclk */
- activity_monitor.Mem_FPS = input[1];
- activity_monitor.Mem_UseRlcBusy = input[2];
- activity_monitor.Mem_MinActiveFreqType = input[3];
- activity_monitor.Mem_MinActiveFreq = input[4];
- activity_monitor.Mem_BoosterFreqType = input[5];
- activity_monitor.Mem_BoosterFreq = input[6];
- activity_monitor.Mem_PD_Data_limit_c = input[7];
- activity_monitor.Mem_PD_Data_error_coeff = input[8];
- activity_monitor.Mem_PD_Data_error_rate_coeff = input[9];
- break;
- default:
+static int arcturus_set_power_profile_mode(struct smu_context *smu,
+ u32 workload_mask,
+ long *custom_params,
+ u32 custom_params_max_idx)
+{
+ u32 backend_workload_mask = 0;
+ int ret, idx = -1, i;
+
+ smu_cmn_get_backend_workload_mask(smu, workload_mask,
+ &backend_workload_mask);
+
+ if (workload_mask & (1 << PP_SMC_POWER_PROFILE_CUSTOM)) {
+ if (smu->smc_fw_version < 0x360d00)
return -EINVAL;
+ if (!smu->custom_profile_params) {
+ smu->custom_profile_params =
+ kzalloc(ARCTURUS_CUSTOM_PARAMS_SIZE, GFP_KERNEL);
+ if (!smu->custom_profile_params)
+ return -ENOMEM;
}
-
- ret = smu_cmn_update_table(smu,
- SMU_TABLE_ACTIVITY_MONITOR_COEFF,
- WORKLOAD_PPLIB_CUSTOM_BIT,
- (void *)(&activity_monitor),
- true);
+ if (custom_params && custom_params_max_idx) {
+ if (custom_params_max_idx != ARCTURUS_CUSTOM_PARAMS_COUNT)
+ return -EINVAL;
+ if (custom_params[0] >= ARCTURUS_CUSTOM_PARAMS_CLOCK_COUNT)
+ return -EINVAL;
+ idx = custom_params[0] * ARCTURUS_CUSTOM_PARAMS_COUNT;
+ smu->custom_profile_params[idx] = 1;
+ for (i = 1; i < custom_params_max_idx; i++)
+ smu->custom_profile_params[idx + i] = custom_params[i];
+ }
+ ret = arcturus_set_power_profile_mode_coeff(smu,
+ smu->custom_profile_params);
if (ret) {
- dev_err(smu->adev->dev, "[%s] Failed to set activity monitor!", __func__);
+ if (idx != -1)
+ smu->custom_profile_params[idx] = 0;
return ret;
}
- }
-
- /*
- * Conv PP_SMC_POWER_PROFILE* to WORKLOAD_PPLIB_*_BIT
- * Not all profile modes are supported on arcturus.
- */
- workload_type = smu_cmn_to_asic_specific_index(smu,
- CMN2ASIC_MAPPING_WORKLOAD,
- profile_mode);
- if (workload_type < 0) {
- dev_dbg(smu->adev->dev, "Unsupported power profile mode %d on arcturus\n", profile_mode);
- return -EINVAL;
+ } else if (smu->custom_profile_params) {
+ memset(smu->custom_profile_params, 0, ARCTURUS_CUSTOM_PARAMS_SIZE);
}
ret = smu_cmn_send_smc_msg_with_param(smu,
- SMU_MSG_SetWorkloadMask,
- 1 << workload_type,
- NULL);
+ SMU_MSG_SetWorkloadMask,
+ backend_workload_mask,
+ NULL);
if (ret) {
- dev_err(smu->adev->dev, "Fail to set workload type %d\n", workload_type);
+ dev_err(smu->adev->dev, "Failed to set workload mask 0x%08x\n",
+ workload_mask);
+ if (idx != -1)
+ smu->custom_profile_params[idx] = 0;
return ret;
}
- smu->power_profile_mode = profile_mode;
-
- return 0;
+ return ret;
}
static int arcturus_set_performance_level(struct smu_context *smu,
diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu11/navi10_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu11/navi10_ppt.c
index 16af1a329621..27c1892b2c74 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/smu11/navi10_ppt.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu11/navi10_ppt.c
@@ -2004,87 +2004,122 @@ static int navi10_get_power_profile_mode(struct smu_context *smu, char *buf)
return size;
}
-static int navi10_set_power_profile_mode(struct smu_context *smu, long *input, uint32_t size)
+#define NAVI10_CUSTOM_PARAMS_COUNT 10
+#define NAVI10_CUSTOM_PARAMS_CLOCKS_COUNT 3
+#define NAVI10_CUSTOM_PARAMS_SIZE (NAVI10_CUSTOM_PARAMS_CLOCKS_COUNT * NAVI10_CUSTOM_PARAMS_COUNT * sizeof(long))
+
+static int navi10_set_power_profile_mode_coeff(struct smu_context *smu,
+ long *input)
{
DpmActivityMonitorCoeffInt_t activity_monitor;
- int workload_type, ret = 0;
+ int ret, idx;
- smu->power_profile_mode = input[size];
+ ret = smu_cmn_update_table(smu,
+ SMU_TABLE_ACTIVITY_MONITOR_COEFF, WORKLOAD_PPLIB_CUSTOM_BIT,
+ (void *)(&activity_monitor), false);
+ if (ret) {
+ dev_err(smu->adev->dev, "[%s] Failed to get activity monitor!", __func__);
+ return ret;
+ }
- if (smu->power_profile_mode > PP_SMC_POWER_PROFILE_CUSTOM) {
- dev_err(smu->adev->dev, "Invalid power profile mode %d\n", smu->power_profile_mode);
- return -EINVAL;
+ idx = 0 * NAVI10_CUSTOM_PARAMS_COUNT;
+ if (input[idx]) {
+ /* Gfxclk */
+ activity_monitor.Gfx_FPS = input[idx + 1];
+ activity_monitor.Gfx_MinFreqStep = input[idx + 2];
+ activity_monitor.Gfx_MinActiveFreqType = input[idx + 3];
+ activity_monitor.Gfx_MinActiveFreq = input[idx + 4];
+ activity_monitor.Gfx_BoosterFreqType = input[idx + 5];
+ activity_monitor.Gfx_BoosterFreq = input[idx + 6];
+ activity_monitor.Gfx_PD_Data_limit_c = input[idx + 7];
+ activity_monitor.Gfx_PD_Data_error_coeff = input[idx + 8];
+ activity_monitor.Gfx_PD_Data_error_rate_coeff = input[idx + 9];
+ }
+ idx = 1 * NAVI10_CUSTOM_PARAMS_COUNT;
+ if (input[idx]) {
+ /* Socclk */
+ activity_monitor.Soc_FPS = input[idx + 1];
+ activity_monitor.Soc_MinFreqStep = input[idx + 2];
+ activity_monitor.Soc_MinActiveFreqType = input[idx + 3];
+ activity_monitor.Soc_MinActiveFreq = input[idx + 4];
+ activity_monitor.Soc_BoosterFreqType = input[idx + 5];
+ activity_monitor.Soc_BoosterFreq = input[idx + 6];
+ activity_monitor.Soc_PD_Data_limit_c = input[idx + 7];
+ activity_monitor.Soc_PD_Data_error_coeff = input[idx + 8];
+ activity_monitor.Soc_PD_Data_error_rate_coeff = input[idx + 9];
+ }
+ idx = 2 * NAVI10_CUSTOM_PARAMS_COUNT;
+ if (input[idx]) {
+ /* Memclk */
+ activity_monitor.Mem_FPS = input[idx + 1];
+ activity_monitor.Mem_MinFreqStep = input[idx + 2];
+ activity_monitor.Mem_MinActiveFreqType = input[idx + 3];
+ activity_monitor.Mem_MinActiveFreq = input[idx + 4];
+ activity_monitor.Mem_BoosterFreqType = input[idx + 5];
+ activity_monitor.Mem_BoosterFreq = input[idx + 6];
+ activity_monitor.Mem_PD_Data_limit_c = input[idx + 7];
+ activity_monitor.Mem_PD_Data_error_coeff = input[idx + 8];
+ activity_monitor.Mem_PD_Data_error_rate_coeff = input[idx + 9];
+ }
+
+ ret = smu_cmn_update_table(smu,
+ SMU_TABLE_ACTIVITY_MONITOR_COEFF, WORKLOAD_PPLIB_CUSTOM_BIT,
+ (void *)(&activity_monitor), true);
+ if (ret) {
+ dev_err(smu->adev->dev, "[%s] Failed to set activity monitor!", __func__);
+ return ret;
}
- if (smu->power_profile_mode == PP_SMC_POWER_PROFILE_CUSTOM) {
- if (size != 10)
- return -EINVAL;
+ return ret;
+}
- ret = smu_cmn_update_table(smu,
- SMU_TABLE_ACTIVITY_MONITOR_COEFF, WORKLOAD_PPLIB_CUSTOM_BIT,
- (void *)(&activity_monitor), false);
- if (ret) {
- dev_err(smu->adev->dev, "[%s] Failed to get activity monitor!", __func__);
- return ret;
- }
+static int navi10_set_power_profile_mode(struct smu_context *smu,
+ u32 workload_mask,
+ long *custom_params,
+ u32 custom_params_max_idx)
+{
+ u32 backend_workload_mask = 0;
+ int ret, idx = -1, i;
- switch (input[0]) {
- case 0: /* Gfxclk */
- activity_monitor.Gfx_FPS = input[1];
- activity_monitor.Gfx_MinFreqStep = input[2];
- activity_monitor.Gfx_MinActiveFreqType = input[3];
- activity_monitor.Gfx_MinActiveFreq = input[4];
- activity_monitor.Gfx_BoosterFreqType = input[5];
- activity_monitor.Gfx_BoosterFreq = input[6];
- activity_monitor.Gfx_PD_Data_limit_c = input[7];
- activity_monitor.Gfx_PD_Data_error_coeff = input[8];
- activity_monitor.Gfx_PD_Data_error_rate_coeff = input[9];
- break;
- case 1: /* Socclk */
- activity_monitor.Soc_FPS = input[1];
- activity_monitor.Soc_MinFreqStep = input[2];
- activity_monitor.Soc_MinActiveFreqType = input[3];
- activity_monitor.Soc_MinActiveFreq = input[4];
- activity_monitor.Soc_BoosterFreqType = input[5];
- activity_monitor.Soc_BoosterFreq = input[6];
- activity_monitor.Soc_PD_Data_limit_c = input[7];
- activity_monitor.Soc_PD_Data_error_coeff = input[8];
- activity_monitor.Soc_PD_Data_error_rate_coeff = input[9];
- break;
- case 2: /* Memclk */
- activity_monitor.Mem_FPS = input[1];
- activity_monitor.Mem_MinFreqStep = input[2];
- activity_monitor.Mem_MinActiveFreqType = input[3];
- activity_monitor.Mem_MinActiveFreq = input[4];
- activity_monitor.Mem_BoosterFreqType = input[5];
- activity_monitor.Mem_BoosterFreq = input[6];
- activity_monitor.Mem_PD_Data_limit_c = input[7];
- activity_monitor.Mem_PD_Data_error_coeff = input[8];
- activity_monitor.Mem_PD_Data_error_rate_coeff = input[9];
- break;
- default:
- return -EINVAL;
- }
+ smu_cmn_get_backend_workload_mask(smu, workload_mask,
+ &backend_workload_mask);
- ret = smu_cmn_update_table(smu,
- SMU_TABLE_ACTIVITY_MONITOR_COEFF, WORKLOAD_PPLIB_CUSTOM_BIT,
- (void *)(&activity_monitor), true);
+ if (workload_mask & (1 << PP_SMC_POWER_PROFILE_CUSTOM)) {
+ if (!smu->custom_profile_params) {
+ smu->custom_profile_params = kzalloc(NAVI10_CUSTOM_PARAMS_SIZE, GFP_KERNEL);
+ if (!smu->custom_profile_params)
+ return -ENOMEM;
+ }
+ if (custom_params && custom_params_max_idx) {
+ if (custom_params_max_idx != NAVI10_CUSTOM_PARAMS_COUNT)
+ return -EINVAL;
+ if (custom_params[0] >= NAVI10_CUSTOM_PARAMS_CLOCKS_COUNT)
+ return -EINVAL;
+ idx = custom_params[0] * NAVI10_CUSTOM_PARAMS_COUNT;
+ smu->custom_profile_params[idx] = 1;
+ for (i = 1; i < custom_params_max_idx; i++)
+ smu->custom_profile_params[idx + i] = custom_params[i];
+ }
+ ret = navi10_set_power_profile_mode_coeff(smu,
+ smu->custom_profile_params);
if (ret) {
- dev_err(smu->adev->dev, "[%s] Failed to set activity monitor!", __func__);
+ if (idx != -1)
+ smu->custom_profile_params[idx] = 0;
return ret;
}
+ } else if (smu->custom_profile_params) {
+ memset(smu->custom_profile_params, 0, NAVI10_CUSTOM_PARAMS_SIZE);
}
- /* conv PP_SMC_POWER_PROFILE* to WORKLOAD_PPLIB_*_BIT */
- workload_type = smu_cmn_to_asic_specific_index(smu,
- CMN2ASIC_MAPPING_WORKLOAD,
- smu->power_profile_mode);
- if (workload_type < 0)
- return -EINVAL;
ret = smu_cmn_send_smc_msg_with_param(smu, SMU_MSG_SetWorkloadMask,
- 1 << workload_type, NULL);
- if (ret)
- dev_err(smu->adev->dev, "[%s] Failed to set work load mask!", __func__);
+ backend_workload_mask, NULL);
+ if (ret) {
+ dev_err(smu->adev->dev, "Failed to set workload mask 0x%08x\n",
+ workload_mask);
+ if (idx != -1)
+ smu->custom_profile_params[idx] = 0;
+ return ret;
+ }
return ret;
}
diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu11/sienna_cichlid_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu11/sienna_cichlid_ppt.c
index 9c3c48297cba..1af90990d05c 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/smu11/sienna_cichlid_ppt.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu11/sienna_cichlid_ppt.c
@@ -1706,90 +1706,126 @@ static int sienna_cichlid_get_power_profile_mode(struct smu_context *smu, char *
return size;
}
-static int sienna_cichlid_set_power_profile_mode(struct smu_context *smu, long *input, uint32_t size)
+#define SIENNA_CICHLID_CUSTOM_PARAMS_COUNT 10
+#define SIENNA_CICHLID_CUSTOM_PARAMS_CLOCK_COUNT 3
+#define SIENNA_CICHLID_CUSTOM_PARAMS_SIZE (SIENNA_CICHLID_CUSTOM_PARAMS_CLOCK_COUNT * SIENNA_CICHLID_CUSTOM_PARAMS_COUNT * sizeof(long))
+
+static int sienna_cichlid_set_power_profile_mode_coeff(struct smu_context *smu,
+ long *input)
{
DpmActivityMonitorCoeffIntExternal_t activity_monitor_external;
DpmActivityMonitorCoeffInt_t *activity_monitor =
&(activity_monitor_external.DpmActivityMonitorCoeffInt);
- int workload_type, ret = 0;
+ int ret, idx;
- smu->power_profile_mode = input[size];
+ ret = smu_cmn_update_table(smu,
+ SMU_TABLE_ACTIVITY_MONITOR_COEFF, WORKLOAD_PPLIB_CUSTOM_BIT,
+ (void *)(&activity_monitor_external), false);
+ if (ret) {
+ dev_err(smu->adev->dev, "[%s] Failed to get activity monitor!", __func__);
+ return ret;
+ }
- if (smu->power_profile_mode > PP_SMC_POWER_PROFILE_CUSTOM) {
- dev_err(smu->adev->dev, "Invalid power profile mode %d\n", smu->power_profile_mode);
- return -EINVAL;
+ idx = 0 * SIENNA_CICHLID_CUSTOM_PARAMS_COUNT;
+ if (input[idx]) {
+ /* Gfxclk */
+ activity_monitor->Gfx_FPS = input[idx + 1];
+ activity_monitor->Gfx_MinFreqStep = input[idx + 2];
+ activity_monitor->Gfx_MinActiveFreqType = input[idx + 3];
+ activity_monitor->Gfx_MinActiveFreq = input[idx + 4];
+ activity_monitor->Gfx_BoosterFreqType = input[idx + 5];
+ activity_monitor->Gfx_BoosterFreq = input[idx + 6];
+ activity_monitor->Gfx_PD_Data_limit_c = input[idx + 7];
+ activity_monitor->Gfx_PD_Data_error_coeff = input[idx + 8];
+ activity_monitor->Gfx_PD_Data_error_rate_coeff = input[idx + 9];
+ }
+ idx = 1 * SIENNA_CICHLID_CUSTOM_PARAMS_COUNT;
+ if (input[idx]) {
+ /* Socclk */
+ activity_monitor->Fclk_FPS = input[idx + 1];
+ activity_monitor->Fclk_MinFreqStep = input[idx + 2];
+ activity_monitor->Fclk_MinActiveFreqType = input[idx + 3];
+ activity_monitor->Fclk_MinActiveFreq = input[idx + 4];
+ activity_monitor->Fclk_BoosterFreqType = input[idx + 5];
+ activity_monitor->Fclk_BoosterFreq = input[idx + 6];
+ activity_monitor->Fclk_PD_Data_limit_c = input[idx + 7];
+ activity_monitor->Fclk_PD_Data_error_coeff = input[idx + 8];
+ activity_monitor->Fclk_PD_Data_error_rate_coeff = input[idx + 9];
+ }
+ idx = 2 * SIENNA_CICHLID_CUSTOM_PARAMS_COUNT;
+ if (input[idx]) {
+ /* Memclk */
+ activity_monitor->Mem_FPS = input[idx + 1];
+ activity_monitor->Mem_MinFreqStep = input[idx + 2];
+ activity_monitor->Mem_MinActiveFreqType = input[idx + 3];
+ activity_monitor->Mem_MinActiveFreq = input[idx + 4];
+ activity_monitor->Mem_BoosterFreqType = input[idx + 5];
+ activity_monitor->Mem_BoosterFreq = input[idx + 6];
+ activity_monitor->Mem_PD_Data_limit_c = input[idx + 7];
+ activity_monitor->Mem_PD_Data_error_coeff = input[idx + 8];
+ activity_monitor->Mem_PD_Data_error_rate_coeff = input[idx + 9];
}
- if (smu->power_profile_mode == PP_SMC_POWER_PROFILE_CUSTOM) {
- if (size != 10)
- return -EINVAL;
+ ret = smu_cmn_update_table(smu,
+ SMU_TABLE_ACTIVITY_MONITOR_COEFF, WORKLOAD_PPLIB_CUSTOM_BIT,
+ (void *)(&activity_monitor_external), true);
+ if (ret) {
+ dev_err(smu->adev->dev, "[%s] Failed to set activity monitor!", __func__);
+ return ret;
+ }
- ret = smu_cmn_update_table(smu,
- SMU_TABLE_ACTIVITY_MONITOR_COEFF, WORKLOAD_PPLIB_CUSTOM_BIT,
- (void *)(&activity_monitor_external), false);
- if (ret) {
- dev_err(smu->adev->dev, "[%s] Failed to get activity monitor!", __func__);
- return ret;
- }
+ return ret;
+}
- switch (input[0]) {
- case 0: /* Gfxclk */
- activity_monitor->Gfx_FPS = input[1];
- activity_monitor->Gfx_MinFreqStep = input[2];
- activity_monitor->Gfx_MinActiveFreqType = input[3];
- activity_monitor->Gfx_MinActiveFreq = input[4];
- activity_monitor->Gfx_BoosterFreqType = input[5];
- activity_monitor->Gfx_BoosterFreq = input[6];
- activity_monitor->Gfx_PD_Data_limit_c = input[7];
- activity_monitor->Gfx_PD_Data_error_coeff = input[8];
- activity_monitor->Gfx_PD_Data_error_rate_coeff = input[9];
- break;
- case 1: /* Socclk */
- activity_monitor->Fclk_FPS = input[1];
- activity_monitor->Fclk_MinFreqStep = input[2];
- activity_monitor->Fclk_MinActiveFreqType = input[3];
- activity_monitor->Fclk_MinActiveFreq = input[4];
- activity_monitor->Fclk_BoosterFreqType = input[5];
- activity_monitor->Fclk_BoosterFreq = input[6];
- activity_monitor->Fclk_PD_Data_limit_c = input[7];
- activity_monitor->Fclk_PD_Data_error_coeff = input[8];
- activity_monitor->Fclk_PD_Data_error_rate_coeff = input[9];
- break;
- case 2: /* Memclk */
- activity_monitor->Mem_FPS = input[1];
- activity_monitor->Mem_MinFreqStep = input[2];
- activity_monitor->Mem_MinActiveFreqType = input[3];
- activity_monitor->Mem_MinActiveFreq = input[4];
- activity_monitor->Mem_BoosterFreqType = input[5];
- activity_monitor->Mem_BoosterFreq = input[6];
- activity_monitor->Mem_PD_Data_limit_c = input[7];
- activity_monitor->Mem_PD_Data_error_coeff = input[8];
- activity_monitor->Mem_PD_Data_error_rate_coeff = input[9];
- break;
- default:
- return -EINVAL;
+static int sienna_cichlid_set_power_profile_mode(struct smu_context *smu,
+ u32 workload_mask,
+ long *custom_params,
+ u32 custom_params_max_idx)
+{
+ u32 backend_workload_mask = 0;
+ int ret, idx = -1, i;
+
+ smu_cmn_get_backend_workload_mask(smu, workload_mask,
+ &backend_workload_mask);
+
+ if (workload_mask & (1 << PP_SMC_POWER_PROFILE_CUSTOM)) {
+ if (!smu->custom_profile_params) {
+ smu->custom_profile_params =
+ kzalloc(SIENNA_CICHLID_CUSTOM_PARAMS_SIZE, GFP_KERNEL);
+ if (!smu->custom_profile_params)
+ return -ENOMEM;
}
-
- ret = smu_cmn_update_table(smu,
- SMU_TABLE_ACTIVITY_MONITOR_COEFF, WORKLOAD_PPLIB_CUSTOM_BIT,
- (void *)(&activity_monitor_external), true);
+ if (custom_params && custom_params_max_idx) {
+ if (custom_params_max_idx != SIENNA_CICHLID_CUSTOM_PARAMS_COUNT)
+ return -EINVAL;
+ if (custom_params[0] >= SIENNA_CICHLID_CUSTOM_PARAMS_CLOCK_COUNT)
+ return -EINVAL;
+ idx = custom_params[0] * SIENNA_CICHLID_CUSTOM_PARAMS_COUNT;
+ smu->custom_profile_params[idx] = 1;
+ for (i = 1; i < custom_params_max_idx; i++)
+ smu->custom_profile_params[idx + i] = custom_params[i];
+ }
+ ret = sienna_cichlid_set_power_profile_mode_coeff(smu,
+ smu->custom_profile_params);
if (ret) {
- dev_err(smu->adev->dev, "[%s] Failed to set activity monitor!", __func__);
+ if (idx != -1)
+ smu->custom_profile_params[idx] = 0;
return ret;
}
+ } else if (smu->custom_profile_params) {
+ memset(smu->custom_profile_params, 0, SIENNA_CICHLID_CUSTOM_PARAMS_SIZE);
}
- /* conv PP_SMC_POWER_PROFILE* to WORKLOAD_PPLIB_*_BIT */
- workload_type = smu_cmn_to_asic_specific_index(smu,
- CMN2ASIC_MAPPING_WORKLOAD,
- smu->power_profile_mode);
- if (workload_type < 0)
- return -EINVAL;
ret = smu_cmn_send_smc_msg_with_param(smu, SMU_MSG_SetWorkloadMask,
- 1 << workload_type, NULL);
- if (ret)
- dev_err(smu->adev->dev, "[%s] Failed to set work load mask!", __func__);
+ backend_workload_mask, NULL);
+ if (ret) {
+ dev_err(smu->adev->dev, "Failed to set workload mask 0x%08x\n",
+ workload_mask);
+ if (idx != -1)
+ smu->custom_profile_params[idx] = 0;
+ return ret;
+ }
return ret;
}
diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu11/vangogh_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu11/vangogh_ppt.c
index 1fe020f1f4db..9bca748ac2e9 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/smu11/vangogh_ppt.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu11/vangogh_ppt.c
@@ -1054,42 +1054,27 @@ static int vangogh_get_power_profile_mode(struct smu_context *smu,
return size;
}
-static int vangogh_set_power_profile_mode(struct smu_context *smu, long *input, uint32_t size)
+static int vangogh_set_power_profile_mode(struct smu_context *smu,
+ u32 workload_mask,
+ long *custom_params,
+ u32 custom_params_max_idx)
{
- int workload_type, ret;
- uint32_t profile_mode = input[size];
+ u32 backend_workload_mask = 0;
+ int ret;
- if (profile_mode >= PP_SMC_POWER_PROFILE_COUNT) {
- dev_err(smu->adev->dev, "Invalid power profile mode %d\n", profile_mode);
- return -EINVAL;
- }
-
- if (profile_mode == PP_SMC_POWER_PROFILE_BOOTUP_DEFAULT ||
- profile_mode == PP_SMC_POWER_PROFILE_POWERSAVING)
- return 0;
-
- /* conv PP_SMC_POWER_PROFILE* to WORKLOAD_PPLIB_*_BIT */
- workload_type = smu_cmn_to_asic_specific_index(smu,
- CMN2ASIC_MAPPING_WORKLOAD,
- profile_mode);
- if (workload_type < 0) {
- dev_dbg(smu->adev->dev, "Unsupported power profile mode %d on VANGOGH\n",
- profile_mode);
- return -EINVAL;
- }
+ smu_cmn_get_backend_workload_mask(smu, workload_mask,
+ &backend_workload_mask);
ret = smu_cmn_send_smc_msg_with_param(smu, SMU_MSG_ActiveProcessNotify,
- 1 << workload_type,
- NULL);
+ backend_workload_mask,
+ NULL);
if (ret) {
- dev_err_once(smu->adev->dev, "Fail to set workload type %d\n",
- workload_type);
+ dev_err_once(smu->adev->dev, "Fail to set workload mask 0x%08x\n",
+ workload_mask);
return ret;
}
- smu->power_profile_mode = profile_mode;
-
- return 0;
+ return ret;
}
static int vangogh_set_soft_freq_limited_range(struct smu_context *smu,
diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu12/renoir_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu12/renoir_ppt.c
index cc0504b063fa..1a8a42b176e5 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/smu12/renoir_ppt.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu12/renoir_ppt.c
@@ -862,44 +862,27 @@ static int renoir_force_clk_levels(struct smu_context *smu,
return ret;
}
-static int renoir_set_power_profile_mode(struct smu_context *smu, long *input, uint32_t size)
+static int renoir_set_power_profile_mode(struct smu_context *smu,
+ u32 workload_mask,
+ long *custom_params,
+ u32 custom_params_max_idx)
{
- int workload_type, ret;
- uint32_t profile_mode = input[size];
+ int ret;
+ u32 backend_workload_mask = 0;
- if (profile_mode > PP_SMC_POWER_PROFILE_CUSTOM) {
- dev_err(smu->adev->dev, "Invalid power profile mode %d\n", profile_mode);
- return -EINVAL;
- }
-
- if (profile_mode == PP_SMC_POWER_PROFILE_BOOTUP_DEFAULT ||
- profile_mode == PP_SMC_POWER_PROFILE_POWERSAVING)
- return 0;
-
- /* conv PP_SMC_POWER_PROFILE* to WORKLOAD_PPLIB_*_BIT */
- workload_type = smu_cmn_to_asic_specific_index(smu,
- CMN2ASIC_MAPPING_WORKLOAD,
- profile_mode);
- if (workload_type < 0) {
- /*
- * TODO: If some case need switch to powersave/default power mode
- * then can consider enter WORKLOAD_COMPUTE/WORKLOAD_CUSTOM for power saving.
- */
- dev_dbg(smu->adev->dev, "Unsupported power profile mode %d on RENOIR\n", profile_mode);
- return -EINVAL;
- }
+ smu_cmn_get_backend_workload_mask(smu, workload_mask,
+ &backend_workload_mask);
ret = smu_cmn_send_smc_msg_with_param(smu, SMU_MSG_ActiveProcessNotify,
- 1 << workload_type,
- NULL);
+ backend_workload_mask,
+ NULL);
if (ret) {
- dev_err_once(smu->adev->dev, "Fail to set workload type %d\n", workload_type);
+ dev_err_once(smu->adev->dev, "Failed to set workload mask 0x08%x\n",
+ workload_mask);
return ret;
}
- smu->power_profile_mode = profile_mode;
-
- return 0;
+ return ret;
}
static int renoir_set_peak_clock_by_device(struct smu_context *smu)
diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c
index d53e162dcd8d..a93739688071 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c
@@ -2477,82 +2477,76 @@ static int smu_v13_0_0_get_power_profile_mode(struct smu_context *smu,
return size;
}
-static int smu_v13_0_0_set_power_profile_mode(struct smu_context *smu,
- long *input,
- uint32_t size)
+#define SMU_13_0_0_CUSTOM_PARAMS_COUNT 9
+#define SMU_13_0_0_CUSTOM_PARAMS_CLOCK_COUNT 2
+#define SMU_13_0_0_CUSTOM_PARAMS_SIZE (SMU_13_0_0_CUSTOM_PARAMS_CLOCK_COUNT * SMU_13_0_0_CUSTOM_PARAMS_COUNT * sizeof(long))
+
+static int smu_v13_0_0_set_power_profile_mode_coeff(struct smu_context *smu,
+ long *input)
{
DpmActivityMonitorCoeffIntExternal_t activity_monitor_external;
DpmActivityMonitorCoeffInt_t *activity_monitor =
&(activity_monitor_external.DpmActivityMonitorCoeffInt);
- int workload_type, ret = 0;
- u32 workload_mask, selected_workload_mask;
-
- smu->power_profile_mode = input[size];
+ int ret, idx;
- if (smu->power_profile_mode >= PP_SMC_POWER_PROFILE_COUNT) {
- dev_err(smu->adev->dev, "Invalid power profile mode %d\n", smu->power_profile_mode);
- return -EINVAL;
+ ret = smu_cmn_update_table(smu,
+ SMU_TABLE_ACTIVITY_MONITOR_COEFF,
+ WORKLOAD_PPLIB_CUSTOM_BIT,
+ (void *)(&activity_monitor_external),
+ false);
+ if (ret) {
+ dev_err(smu->adev->dev, "[%s] Failed to get activity monitor!", __func__);
+ return ret;
}
- if (smu->power_profile_mode == PP_SMC_POWER_PROFILE_CUSTOM) {
- if (size != 9)
- return -EINVAL;
-
- ret = smu_cmn_update_table(smu,
- SMU_TABLE_ACTIVITY_MONITOR_COEFF,
- WORKLOAD_PPLIB_CUSTOM_BIT,
- (void *)(&activity_monitor_external),
- false);
- if (ret) {
- dev_err(smu->adev->dev, "[%s] Failed to get activity monitor!", __func__);
- return ret;
- }
-
- switch (input[0]) {
- case 0: /* Gfxclk */
- activity_monitor->Gfx_FPS = input[1];
- activity_monitor->Gfx_MinActiveFreqType = input[2];
- activity_monitor->Gfx_MinActiveFreq = input[3];
- activity_monitor->Gfx_BoosterFreqType = input[4];
- activity_monitor->Gfx_BoosterFreq = input[5];
- activity_monitor->Gfx_PD_Data_limit_c = input[6];
- activity_monitor->Gfx_PD_Data_error_coeff = input[7];
- activity_monitor->Gfx_PD_Data_error_rate_coeff = input[8];
- break;
- case 1: /* Fclk */
- activity_monitor->Fclk_FPS = input[1];
- activity_monitor->Fclk_MinActiveFreqType = input[2];
- activity_monitor->Fclk_MinActiveFreq = input[3];
- activity_monitor->Fclk_BoosterFreqType = input[4];
- activity_monitor->Fclk_BoosterFreq = input[5];
- activity_monitor->Fclk_PD_Data_limit_c = input[6];
- activity_monitor->Fclk_PD_Data_error_coeff = input[7];
- activity_monitor->Fclk_PD_Data_error_rate_coeff = input[8];
- break;
- default:
- return -EINVAL;
- }
+ idx = 0 * SMU_13_0_0_CUSTOM_PARAMS_COUNT;
+ if (input[idx]) {
+ /* Gfxclk */
+ activity_monitor->Gfx_FPS = input[idx + 1];
+ activity_monitor->Gfx_MinActiveFreqType = input[idx + 2];
+ activity_monitor->Gfx_MinActiveFreq = input[idx + 3];
+ activity_monitor->Gfx_BoosterFreqType = input[idx + 4];
+ activity_monitor->Gfx_BoosterFreq = input[idx + 5];
+ activity_monitor->Gfx_PD_Data_limit_c = input[idx + 6];
+ activity_monitor->Gfx_PD_Data_error_coeff = input[idx + 7];
+ activity_monitor->Gfx_PD_Data_error_rate_coeff = input[idx + 8];
+ }
+ idx = 1 * SMU_13_0_0_CUSTOM_PARAMS_COUNT;
+ if (input[idx]) {
+ /* Fclk */
+ activity_monitor->Fclk_FPS = input[idx + 1];
+ activity_monitor->Fclk_MinActiveFreqType = input[idx + 2];
+ activity_monitor->Fclk_MinActiveFreq = input[idx + 3];
+ activity_monitor->Fclk_BoosterFreqType = input[idx + 4];
+ activity_monitor->Fclk_BoosterFreq = input[idx + 5];
+ activity_monitor->Fclk_PD_Data_limit_c = input[idx + 6];
+ activity_monitor->Fclk_PD_Data_error_coeff = input[idx + 7];
+ activity_monitor->Fclk_PD_Data_error_rate_coeff = input[idx + 8];
+ }
- ret = smu_cmn_update_table(smu,
- SMU_TABLE_ACTIVITY_MONITOR_COEFF,
- WORKLOAD_PPLIB_CUSTOM_BIT,
- (void *)(&activity_monitor_external),
- true);
- if (ret) {
- dev_err(smu->adev->dev, "[%s] Failed to set activity monitor!", __func__);
- return ret;
- }
+ ret = smu_cmn_update_table(smu,
+ SMU_TABLE_ACTIVITY_MONITOR_COEFF,
+ WORKLOAD_PPLIB_CUSTOM_BIT,
+ (void *)(&activity_monitor_external),
+ true);
+ if (ret) {
+ dev_err(smu->adev->dev, "[%s] Failed to set activity monitor!", __func__);
+ return ret;
}
- /* conv PP_SMC_POWER_PROFILE* to WORKLOAD_PPLIB_*_BIT */
- workload_type = smu_cmn_to_asic_specific_index(smu,
- CMN2ASIC_MAPPING_WORKLOAD,
- smu->power_profile_mode);
+ return ret;
+}
- if (workload_type < 0)
- return -EINVAL;
+static int smu_v13_0_0_set_power_profile_mode(struct smu_context *smu,
+ u32 workload_mask,
+ long *custom_params,
+ u32 custom_params_max_idx)
+{
+ u32 backend_workload_mask = 0;
+ int workload_type, ret, idx = -1, i;
- selected_workload_mask = workload_mask = 1 << workload_type;
+ smu_cmn_get_backend_workload_mask(smu, workload_mask,
+ &backend_workload_mask);
/* Add optimizations for SMU13.0.0/10. Reuse the power saving profile */
if ((amdgpu_ip_version(smu->adev, MP1_HWIP, 0) == IP_VERSION(13, 0, 0) &&
@@ -2564,15 +2558,48 @@ static int smu_v13_0_0_set_power_profile_mode(struct smu_context *smu,
CMN2ASIC_MAPPING_WORKLOAD,
PP_SMC_POWER_PROFILE_POWERSAVING);
if (workload_type >= 0)
- workload_mask |= 1 << workload_type;
+ backend_workload_mask |= 1 << workload_type;
+ }
+
+ if (workload_mask & (1 << PP_SMC_POWER_PROFILE_CUSTOM)) {
+ if (!smu->custom_profile_params) {
+ smu->custom_profile_params =
+ kzalloc(SMU_13_0_0_CUSTOM_PARAMS_SIZE, GFP_KERNEL);
+ if (!smu->custom_profile_params)
+ return -ENOMEM;
+ }
+ if (custom_params && custom_params_max_idx) {
+ if (custom_params_max_idx != SMU_13_0_0_CUSTOM_PARAMS_COUNT)
+ return -EINVAL;
+ if (custom_params[0] >= SMU_13_0_0_CUSTOM_PARAMS_CLOCK_COUNT)
+ return -EINVAL;
+ idx = custom_params[0] * SMU_13_0_0_CUSTOM_PARAMS_COUNT;
+ smu->custom_profile_params[idx] = 1;
+ for (i = 1; i < custom_params_max_idx; i++)
+ smu->custom_profile_params[idx + i] = custom_params[i];
+ }
+ ret = smu_v13_0_0_set_power_profile_mode_coeff(smu,
+ smu->custom_profile_params);
+ if (ret) {
+ if (idx != -1)
+ smu->custom_profile_params[idx] = 0;
+ return ret;
+ }
+ } else if (smu->custom_profile_params) {
+ memset(smu->custom_profile_params, 0, SMU_13_0_0_CUSTOM_PARAMS_SIZE);
}
ret = smu_cmn_send_smc_msg_with_param(smu,
- SMU_MSG_SetWorkloadMask,
- workload_mask,
- NULL);
- if (!ret)
- smu->workload_mask = selected_workload_mask;
+ SMU_MSG_SetWorkloadMask,
+ backend_workload_mask,
+ NULL);
+ if (ret) {
+ dev_err(smu->adev->dev, "Failed to set workload mask 0x%08x\n",
+ workload_mask);
+ if (idx != -1)
+ smu->custom_profile_params[idx] = 0;
+ return ret;
+ }
return ret;
}
diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c
index ceaf4572db25..d0e6d051e9cf 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c
@@ -2436,78 +2436,110 @@ out:
return result;
}
-static int smu_v13_0_7_set_power_profile_mode(struct smu_context *smu, long *input, uint32_t size)
+#define SMU_13_0_7_CUSTOM_PARAMS_COUNT 8
+#define SMU_13_0_7_CUSTOM_PARAMS_CLOCK_COUNT 2
+#define SMU_13_0_7_CUSTOM_PARAMS_SIZE (SMU_13_0_7_CUSTOM_PARAMS_CLOCK_COUNT * SMU_13_0_7_CUSTOM_PARAMS_COUNT * sizeof(long))
+
+static int smu_v13_0_7_set_power_profile_mode_coeff(struct smu_context *smu,
+ long *input)
{
DpmActivityMonitorCoeffIntExternal_t activity_monitor_external;
DpmActivityMonitorCoeffInt_t *activity_monitor =
&(activity_monitor_external.DpmActivityMonitorCoeffInt);
- int workload_type, ret = 0;
+ int ret, idx;
- smu->power_profile_mode = input[size];
+ ret = smu_cmn_update_table(smu,
+ SMU_TABLE_ACTIVITY_MONITOR_COEFF, WORKLOAD_PPLIB_CUSTOM_BIT,
+ (void *)(&activity_monitor_external), false);
+ if (ret) {
+ dev_err(smu->adev->dev, "[%s] Failed to get activity monitor!", __func__);
+ return ret;
+ }
- if (smu->power_profile_mode > PP_SMC_POWER_PROFILE_WINDOW3D) {
- dev_err(smu->adev->dev, "Invalid power profile mode %d\n", smu->power_profile_mode);
- return -EINVAL;
+ idx = 0 * SMU_13_0_7_CUSTOM_PARAMS_COUNT;
+ if (input[idx]) {
+ /* Gfxclk */
+ activity_monitor->Gfx_ActiveHystLimit = input[idx + 1];
+ activity_monitor->Gfx_IdleHystLimit = input[idx + 2];
+ activity_monitor->Gfx_FPS = input[idx + 3];
+ activity_monitor->Gfx_MinActiveFreqType = input[idx + 4];
+ activity_monitor->Gfx_BoosterFreqType = input[idx + 5];
+ activity_monitor->Gfx_MinActiveFreq = input[idx + 6];
+ activity_monitor->Gfx_BoosterFreq = input[idx + 7];
+ }
+ idx = 1 * SMU_13_0_7_CUSTOM_PARAMS_COUNT;
+ if (input[idx]) {
+ /* Fclk */
+ activity_monitor->Fclk_ActiveHystLimit = input[idx + 1];
+ activity_monitor->Fclk_IdleHystLimit = input[idx + 2];
+ activity_monitor->Fclk_FPS = input[idx + 3];
+ activity_monitor->Fclk_MinActiveFreqType = input[idx + 4];
+ activity_monitor->Fclk_BoosterFreqType = input[idx + 5];
+ activity_monitor->Fclk_MinActiveFreq = input[idx + 6];
+ activity_monitor->Fclk_BoosterFreq = input[idx + 7];
}
- if (smu->power_profile_mode == PP_SMC_POWER_PROFILE_CUSTOM) {
- if (size != 8)
- return -EINVAL;
+ ret = smu_cmn_update_table(smu,
+ SMU_TABLE_ACTIVITY_MONITOR_COEFF, WORKLOAD_PPLIB_CUSTOM_BIT,
+ (void *)(&activity_monitor_external), true);
+ if (ret) {
+ dev_err(smu->adev->dev, "[%s] Failed to set activity monitor!", __func__);
+ return ret;
+ }
- ret = smu_cmn_update_table(smu,
- SMU_TABLE_ACTIVITY_MONITOR_COEFF, WORKLOAD_PPLIB_CUSTOM_BIT,
- (void *)(&activity_monitor_external), false);
- if (ret) {
- dev_err(smu->adev->dev, "[%s] Failed to get activity monitor!", __func__);
- return ret;
- }
+ return ret;
+}
- switch (input[0]) {
- case 0: /* Gfxclk */
- activity_monitor->Gfx_ActiveHystLimit = input[1];
- activity_monitor->Gfx_IdleHystLimit = input[2];
- activity_monitor->Gfx_FPS = input[3];
- activity_monitor->Gfx_MinActiveFreqType = input[4];
- activity_monitor->Gfx_BoosterFreqType = input[5];
- activity_monitor->Gfx_MinActiveFreq = input[6];
- activity_monitor->Gfx_BoosterFreq = input[7];
- break;
- case 1: /* Fclk */
- activity_monitor->Fclk_ActiveHystLimit = input[1];
- activity_monitor->Fclk_IdleHystLimit = input[2];
- activity_monitor->Fclk_FPS = input[3];
- activity_monitor->Fclk_MinActiveFreqType = input[4];
- activity_monitor->Fclk_BoosterFreqType = input[5];
- activity_monitor->Fclk_MinActiveFreq = input[6];
- activity_monitor->Fclk_BoosterFreq = input[7];
- break;
- default:
- return -EINVAL;
+static int smu_v13_0_7_set_power_profile_mode(struct smu_context *smu,
+ u32 workload_mask,
+ long *custom_params,
+ u32 custom_params_max_idx)
+{
+ u32 backend_workload_mask = 0;
+ int ret, idx = -1, i;
+
+ smu_cmn_get_backend_workload_mask(smu, workload_mask,
+ &backend_workload_mask);
+
+ if (workload_mask & (1 << PP_SMC_POWER_PROFILE_CUSTOM)) {
+ if (!smu->custom_profile_params) {
+ smu->custom_profile_params =
+ kzalloc(SMU_13_0_7_CUSTOM_PARAMS_SIZE, GFP_KERNEL);
+ if (!smu->custom_profile_params)
+ return -ENOMEM;
}
-
- ret = smu_cmn_update_table(smu,
- SMU_TABLE_ACTIVITY_MONITOR_COEFF, WORKLOAD_PPLIB_CUSTOM_BIT,
- (void *)(&activity_monitor_external), true);
+ if (custom_params && custom_params_max_idx) {
+ if (custom_params_max_idx != SMU_13_0_7_CUSTOM_PARAMS_COUNT)
+ return -EINVAL;
+ if (custom_params[0] >= SMU_13_0_7_CUSTOM_PARAMS_CLOCK_COUNT)
+ return -EINVAL;
+ idx = custom_params[0] * SMU_13_0_7_CUSTOM_PARAMS_COUNT;
+ smu->custom_profile_params[idx] = 1;
+ for (i = 1; i < custom_params_max_idx; i++)
+ smu->custom_profile_params[idx + i] = custom_params[i];
+ }
+ ret = smu_v13_0_7_set_power_profile_mode_coeff(smu,
+ smu->custom_profile_params);
if (ret) {
- dev_err(smu->adev->dev, "[%s] Failed to set activity monitor!", __func__);
+ if (idx != -1)
+ smu->custom_profile_params[idx] = 0;
return ret;
}
+ } else if (smu->custom_profile_params) {
+ memset(smu->custom_profile_params, 0, SMU_13_0_7_CUSTOM_PARAMS_SIZE);
}
- /* conv PP_SMC_POWER_PROFILE* to WORKLOAD_PPLIB_*_BIT */
- workload_type = smu_cmn_to_asic_specific_index(smu,
- CMN2ASIC_MAPPING_WORKLOAD,
- smu->power_profile_mode);
- if (workload_type < 0)
- return -EINVAL;
ret = smu_cmn_send_smc_msg_with_param(smu, SMU_MSG_SetWorkloadMask,
- 1 << workload_type, NULL);
+ backend_workload_mask, NULL);
- if (ret)
- dev_err(smu->adev->dev, "[%s] Failed to set work load mask!", __func__);
- else
- smu->workload_mask = (1 << workload_type);
+ if (ret) {
+ dev_err(smu->adev->dev, "Failed to set workload mask 0x%08x\n",
+ workload_mask);
+ if (idx != -1)
+ smu->custom_profile_params[idx] = 0;
+ return ret;
+ }
return ret;
}
diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu14/smu_v14_0_2_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu14/smu_v14_0_2_ppt.c
index 82aef8626afa..b22fb7eafcd3 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/smu14/smu_v14_0_2_ppt.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu14/smu_v14_0_2_ppt.c
@@ -1751,90 +1751,120 @@ static int smu_v14_0_2_get_power_profile_mode(struct smu_context *smu,
return size;
}
-static int smu_v14_0_2_set_power_profile_mode(struct smu_context *smu,
- long *input,
- uint32_t size)
+#define SMU_14_0_2_CUSTOM_PARAMS_COUNT 9
+#define SMU_14_0_2_CUSTOM_PARAMS_CLOCK_COUNT 2
+#define SMU_14_0_2_CUSTOM_PARAMS_SIZE (SMU_14_0_2_CUSTOM_PARAMS_CLOCK_COUNT * SMU_14_0_2_CUSTOM_PARAMS_COUNT * sizeof(long))
+
+static int smu_v14_0_2_set_power_profile_mode_coeff(struct smu_context *smu,
+ long *input)
{
DpmActivityMonitorCoeffIntExternal_t activity_monitor_external;
DpmActivityMonitorCoeffInt_t *activity_monitor =
&(activity_monitor_external.DpmActivityMonitorCoeffInt);
- int workload_type, ret = 0;
- uint32_t current_profile_mode = smu->power_profile_mode;
- smu->power_profile_mode = input[size];
+ int ret, idx;
- if (smu->power_profile_mode >= PP_SMC_POWER_PROFILE_COUNT) {
- dev_err(smu->adev->dev, "Invalid power profile mode %d\n", smu->power_profile_mode);
- return -EINVAL;
+ ret = smu_cmn_update_table(smu,
+ SMU_TABLE_ACTIVITY_MONITOR_COEFF,
+ WORKLOAD_PPLIB_CUSTOM_BIT,
+ (void *)(&activity_monitor_external),
+ false);
+ if (ret) {
+ dev_err(smu->adev->dev, "[%s] Failed to get activity monitor!", __func__);
+ return ret;
}
- if (smu->power_profile_mode == PP_SMC_POWER_PROFILE_CUSTOM) {
- if (size != 9)
- return -EINVAL;
+ idx = 0 * SMU_14_0_2_CUSTOM_PARAMS_COUNT;
+ if (input[idx]) {
+ /* Gfxclk */
+ activity_monitor->Gfx_FPS = input[idx + 1];
+ activity_monitor->Gfx_MinActiveFreqType = input[idx + 2];
+ activity_monitor->Gfx_MinActiveFreq = input[idx + 3];
+ activity_monitor->Gfx_BoosterFreqType = input[idx + 4];
+ activity_monitor->Gfx_BoosterFreq = input[idx + 5];
+ activity_monitor->Gfx_PD_Data_limit_c = input[idx + 6];
+ activity_monitor->Gfx_PD_Data_error_coeff = input[idx + 7];
+ activity_monitor->Gfx_PD_Data_error_rate_coeff = input[idx + 8];
+ }
+ idx = 1 * SMU_14_0_2_CUSTOM_PARAMS_COUNT;
+ if (input[idx]) {
+ /* Fclk */
+ activity_monitor->Fclk_FPS = input[idx + 1];
+ activity_monitor->Fclk_MinActiveFreqType = input[idx + 2];
+ activity_monitor->Fclk_MinActiveFreq = input[idx + 3];
+ activity_monitor->Fclk_BoosterFreqType = input[idx + 4];
+ activity_monitor->Fclk_BoosterFreq = input[idx + 5];
+ activity_monitor->Fclk_PD_Data_limit_c = input[idx + 6];
+ activity_monitor->Fclk_PD_Data_error_coeff = input[idx + 7];
+ activity_monitor->Fclk_PD_Data_error_rate_coeff = input[idx + 8];
+ }
- ret = smu_cmn_update_table(smu,
- SMU_TABLE_ACTIVITY_MONITOR_COEFF,
- WORKLOAD_PPLIB_CUSTOM_BIT,
- (void *)(&activity_monitor_external),
- false);
- if (ret) {
- dev_err(smu->adev->dev, "[%s] Failed to get activity monitor!", __func__);
- return ret;
- }
+ ret = smu_cmn_update_table(smu,
+ SMU_TABLE_ACTIVITY_MONITOR_COEFF,
+ WORKLOAD_PPLIB_CUSTOM_BIT,
+ (void *)(&activity_monitor_external),
+ true);
+ if (ret) {
+ dev_err(smu->adev->dev, "[%s] Failed to set activity monitor!", __func__);
+ return ret;
+ }
- switch (input[0]) {
- case 0: /* Gfxclk */
- activity_monitor->Gfx_FPS = input[1];
- activity_monitor->Gfx_MinActiveFreqType = input[2];
- activity_monitor->Gfx_MinActiveFreq = input[3];
- activity_monitor->Gfx_BoosterFreqType = input[4];
- activity_monitor->Gfx_BoosterFreq = input[5];
- activity_monitor->Gfx_PD_Data_limit_c = input[6];
- activity_monitor->Gfx_PD_Data_error_coeff = input[7];
- activity_monitor->Gfx_PD_Data_error_rate_coeff = input[8];
- break;
- case 1: /* Fclk */
- activity_monitor->Fclk_FPS = input[1];
- activity_monitor->Fclk_MinActiveFreqType = input[2];
- activity_monitor->Fclk_MinActiveFreq = input[3];
- activity_monitor->Fclk_BoosterFreqType = input[4];
- activity_monitor->Fclk_BoosterFreq = input[5];
- activity_monitor->Fclk_PD_Data_limit_c = input[6];
- activity_monitor->Fclk_PD_Data_error_coeff = input[7];
- activity_monitor->Fclk_PD_Data_error_rate_coeff = input[8];
- break;
- default:
- return -EINVAL;
- }
+ return ret;
+}
- ret = smu_cmn_update_table(smu,
- SMU_TABLE_ACTIVITY_MONITOR_COEFF,
- WORKLOAD_PPLIB_CUSTOM_BIT,
- (void *)(&activity_monitor_external),
- true);
- if (ret) {
- dev_err(smu->adev->dev, "[%s] Failed to set activity monitor!", __func__);
- return ret;
- }
- }
+static int smu_v14_0_2_set_power_profile_mode(struct smu_context *smu,
+ u32 workload_mask,
+ long *custom_params,
+ u32 custom_params_max_idx)
+{
+ u32 backend_workload_mask = 0;
+ int ret, idx = -1, i;
+
+ smu_cmn_get_backend_workload_mask(smu, workload_mask,
+ &backend_workload_mask);
- if (smu->power_profile_mode == PP_SMC_POWER_PROFILE_COMPUTE)
+ /* disable deep sleep if compute is enabled */
+ if (workload_mask & (1 << PP_SMC_POWER_PROFILE_COMPUTE))
smu_v14_0_deep_sleep_control(smu, false);
- else if (current_profile_mode == PP_SMC_POWER_PROFILE_COMPUTE)
+ else
smu_v14_0_deep_sleep_control(smu, true);
- /* conv PP_SMC_POWER_PROFILE* to WORKLOAD_PPLIB_*_BIT */
- workload_type = smu_cmn_to_asic_specific_index(smu,
- CMN2ASIC_MAPPING_WORKLOAD,
- smu->power_profile_mode);
- if (workload_type < 0)
- return -EINVAL;
+ if (workload_mask & (1 << PP_SMC_POWER_PROFILE_CUSTOM)) {
+ if (!smu->custom_profile_params) {
+ smu->custom_profile_params =
+ kzalloc(SMU_14_0_2_CUSTOM_PARAMS_SIZE, GFP_KERNEL);
+ if (!smu->custom_profile_params)
+ return -ENOMEM;
+ }
+ if (custom_params && custom_params_max_idx) {
+ if (custom_params_max_idx != SMU_14_0_2_CUSTOM_PARAMS_COUNT)
+ return -EINVAL;
+ if (custom_params[0] >= SMU_14_0_2_CUSTOM_PARAMS_CLOCK_COUNT)
+ return -EINVAL;
+ idx = custom_params[0] * SMU_14_0_2_CUSTOM_PARAMS_COUNT;
+ smu->custom_profile_params[idx] = 1;
+ for (i = 1; i < custom_params_max_idx; i++)
+ smu->custom_profile_params[idx + i] = custom_params[i];
+ }
+ ret = smu_v14_0_2_set_power_profile_mode_coeff(smu,
+ smu->custom_profile_params);
+ if (ret) {
+ if (idx != -1)
+ smu->custom_profile_params[idx] = 0;
+ return ret;
+ }
+ } else if (smu->custom_profile_params) {
+ memset(smu->custom_profile_params, 0, SMU_14_0_2_CUSTOM_PARAMS_SIZE);
+ }
- ret = smu_cmn_send_smc_msg_with_param(smu,
- SMU_MSG_SetWorkloadMask,
- 1 << workload_type,
- NULL);
- if (!ret)
- smu->workload_mask = 1 << workload_type;
+ ret = smu_cmn_send_smc_msg_with_param(smu, SMU_MSG_SetWorkloadMask,
+ backend_workload_mask, NULL);
+ if (ret) {
+ dev_err(smu->adev->dev, "Failed to set workload mask 0x%08x\n",
+ workload_mask);
+ if (idx != -1)
+ smu->custom_profile_params[idx] = 0;
+ return ret;
+ }
return ret;
}
diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu_cmn.c b/drivers/gpu/drm/amd/pm/swsmu/smu_cmn.c
index 91ad434bcdae..0d71db7be325 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/smu_cmn.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu_cmn.c
@@ -1215,3 +1215,28 @@ void smu_cmn_generic_plpd_policy_desc(struct smu_dpm_policy *policy)
{
policy->desc = &xgmi_plpd_policy_desc;
}
+
+void smu_cmn_get_backend_workload_mask(struct smu_context *smu,
+ u32 workload_mask,
+ u32 *backend_workload_mask)
+{
+ int workload_type;
+ u32 profile_mode;
+
+ *backend_workload_mask = 0;
+
+ for (profile_mode = 0; profile_mode < PP_SMC_POWER_PROFILE_COUNT; profile_mode++) {
+ if (!(workload_mask & (1 << profile_mode)))
+ continue;
+
+ /* conv PP_SMC_POWER_PROFILE* to WORKLOAD_PPLIB_*_BIT */
+ workload_type = smu_cmn_to_asic_specific_index(smu,
+ CMN2ASIC_MAPPING_WORKLOAD,
+ profile_mode);
+
+ if (workload_type < 0)
+ continue;
+
+ *backend_workload_mask |= 1 << workload_type;
+ }
+}
diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu_cmn.h b/drivers/gpu/drm/amd/pm/swsmu/smu_cmn.h
index 1de685defe85..a020277dec3e 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/smu_cmn.h
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu_cmn.h
@@ -147,5 +147,9 @@ bool smu_cmn_is_audio_func_enabled(struct amdgpu_device *adev);
void smu_cmn_generic_soc_policy_desc(struct smu_dpm_policy *policy);
void smu_cmn_generic_plpd_policy_desc(struct smu_dpm_policy *policy);
+void smu_cmn_get_backend_workload_mask(struct smu_context *smu,
+ u32 workload_mask,
+ u32 *backend_workload_mask);
+
#endif
#endif