diff options
Diffstat (limited to 'drivers/gpu/drm/xe')
26 files changed, 260 insertions, 136 deletions
diff --git a/drivers/gpu/drm/xe/regs/xe_gt_regs.h b/drivers/gpu/drm/xe/regs/xe_gt_regs.h index 06cb6b02ec64..51f2a03847f9 100644 --- a/drivers/gpu/drm/xe/regs/xe_gt_regs.h +++ b/drivers/gpu/drm/xe/regs/xe_gt_regs.h @@ -342,6 +342,7 @@ #define POWERGATE_ENABLE XE_REG(0xa210) #define RENDER_POWERGATE_ENABLE REG_BIT(0) #define MEDIA_POWERGATE_ENABLE REG_BIT(1) +#define MEDIA_SAMPLERS_POWERGATE_ENABLE REG_BIT(2) #define VDN_HCP_POWERGATE_ENABLE(n) REG_BIT(3 + 2 * (n)) #define VDN_MFXVDENC_POWERGATE_ENABLE(n) REG_BIT(4 + 2 * (n)) diff --git a/drivers/gpu/drm/xe/tests/xe_pci.c b/drivers/gpu/drm/xe/tests/xe_pci.c index 69e2840c7ef0..663a79ec960d 100644 --- a/drivers/gpu/drm/xe/tests/xe_pci.c +++ b/drivers/gpu/drm/xe/tests/xe_pci.c @@ -66,6 +66,7 @@ KUNIT_ARRAY_PARAM(platform, cases, xe_pci_fake_data_desc); /** * xe_pci_fake_data_gen_params - Generate struct xe_pci_fake_data parameters + * @test: test context object * @prev: the pointer to the previous parameter to iterate from or NULL * @desc: output buffer with minimum size of KUNIT_PARAM_DESC_SIZE * @@ -242,6 +243,7 @@ KUNIT_ARRAY_PARAM(pci_id, pciidlist, xe_pci_id_kunit_desc); /** * xe_pci_graphics_ip_gen_param - Generate graphics struct xe_ip parameters + * @test: test context object * @prev: the pointer to the previous parameter to iterate from or NULL * @desc: output buffer with minimum size of KUNIT_PARAM_DESC_SIZE * @@ -266,6 +268,7 @@ EXPORT_SYMBOL_IF_KUNIT(xe_pci_graphics_ip_gen_param); /** * xe_pci_media_ip_gen_param - Generate media struct xe_ip parameters + * @test: test context object * @prev: the pointer to the previous parameter to iterate from or NULL * @desc: output buffer with minimum size of KUNIT_PARAM_DESC_SIZE * @@ -290,6 +293,7 @@ EXPORT_SYMBOL_IF_KUNIT(xe_pci_media_ip_gen_param); /** * xe_pci_id_gen_param - Generate struct pci_device_id parameters + * @test: test context object * @prev: the pointer to the previous parameter to iterate from or NULL * @desc: output buffer with minimum size of KUNIT_PARAM_DESC_SIZE * @@ -376,6 +380,7 @@ EXPORT_SYMBOL_IF_KUNIT(xe_pci_fake_device_init); /** * xe_pci_live_device_gen_param - Helper to iterate Xe devices as KUnit parameters + * @test: test context object * @prev: the previously returned value, or NULL for the first iteration * @desc: the buffer for a parameter name * diff --git a/drivers/gpu/drm/xe/xe_bo_evict.c b/drivers/gpu/drm/xe/xe_bo_evict.c index d5dbc51e8612..bc5b4c5fab81 100644 --- a/drivers/gpu/drm/xe/xe_bo_evict.c +++ b/drivers/gpu/drm/xe/xe_bo_evict.c @@ -182,7 +182,6 @@ int xe_bo_evict_all(struct xe_device *xe) static int xe_bo_restore_and_map_ggtt(struct xe_bo *bo) { - struct xe_device *xe = xe_bo_device(bo); int ret; ret = xe_bo_restore_pinned(bo); @@ -201,13 +200,6 @@ static int xe_bo_restore_and_map_ggtt(struct xe_bo *bo) } } - /* - * We expect validate to trigger a move VRAM and our move code - * should setup the iosys map. - */ - xe_assert(xe, !(bo->flags & XE_BO_FLAG_PINNED_LATE_RESTORE) || - !iosys_map_is_null(&bo->vmap)); - return 0; } diff --git a/drivers/gpu/drm/xe/xe_device.c b/drivers/gpu/drm/xe/xe_device.c index 2883b39c9b37..456899238377 100644 --- a/drivers/gpu/drm/xe/xe_device.c +++ b/drivers/gpu/drm/xe/xe_device.c @@ -988,16 +988,16 @@ void xe_device_shutdown(struct xe_device *xe) drm_dbg(&xe->drm, "Shutting down device\n"); - if (xe_driver_flr_disabled(xe)) { - xe_display_pm_shutdown(xe); + xe_display_pm_shutdown(xe); - xe_irq_suspend(xe); + xe_irq_suspend(xe); - for_each_gt(gt, xe, id) - xe_gt_shutdown(gt); + for_each_gt(gt, xe, id) + xe_gt_shutdown(gt); - xe_display_pm_shutdown_late(xe); - } else { + xe_display_pm_shutdown_late(xe); + + if (!xe_driver_flr_disabled(xe)) { /* BOOM! */ __xe_driver_flr(xe); } @@ -1070,7 +1070,7 @@ void xe_device_l2_flush(struct xe_device *xe) spin_lock(>->global_invl_lock); xe_mmio_write32(>->mmio, XE2_GLOBAL_INVAL, 0x1); - if (xe_mmio_wait32(>->mmio, XE2_GLOBAL_INVAL, 0x1, 0x0, 500, NULL, true)) + if (xe_mmio_wait32(>->mmio, XE2_GLOBAL_INVAL, 0x1, 0x0, 1000, NULL, true)) xe_gt_err_once(gt, "Global invalidation timeout\n"); spin_unlock(>->global_invl_lock); diff --git a/drivers/gpu/drm/xe/xe_exec.c b/drivers/gpu/drm/xe/xe_exec.c index 7715e74bb945..a8ab363a8046 100644 --- a/drivers/gpu/drm/xe/xe_exec.c +++ b/drivers/gpu/drm/xe/xe_exec.c @@ -165,7 +165,8 @@ int xe_exec_ioctl(struct drm_device *dev, void *data, struct drm_file *file) for (num_syncs = 0; num_syncs < args->num_syncs; num_syncs++) { err = xe_sync_entry_parse(xe, xef, &syncs[num_syncs], - &syncs_user[num_syncs], SYNC_PARSE_FLAG_EXEC | + &syncs_user[num_syncs], NULL, 0, + SYNC_PARSE_FLAG_EXEC | (xe_vm_in_lr_mode(vm) ? SYNC_PARSE_FLAG_LR_MODE : 0)); if (err) diff --git a/drivers/gpu/drm/xe/xe_exec_queue.c b/drivers/gpu/drm/xe/xe_exec_queue.c index 37b2b93b73d6..cb5f204c08ed 100644 --- a/drivers/gpu/drm/xe/xe_exec_queue.c +++ b/drivers/gpu/drm/xe/xe_exec_queue.c @@ -10,6 +10,7 @@ #include <drm/drm_device.h> #include <drm/drm_drv.h> #include <drm/drm_file.h> +#include <drm/drm_syncobj.h> #include <uapi/drm/xe_drm.h> #include "xe_dep_scheduler.h" @@ -324,6 +325,16 @@ struct xe_exec_queue *xe_exec_queue_create_bind(struct xe_device *xe, } xe_vm_put(migrate_vm); + if (!IS_ERR(q)) { + int err = drm_syncobj_create(&q->ufence_syncobj, + DRM_SYNCOBJ_CREATE_SIGNALED, + NULL); + if (err) { + xe_exec_queue_put(q); + return ERR_PTR(err); + } + } + return q; } ALLOW_ERROR_INJECTION(xe_exec_queue_create_bind, ERRNO); @@ -333,6 +344,9 @@ void xe_exec_queue_destroy(struct kref *ref) struct xe_exec_queue *q = container_of(ref, struct xe_exec_queue, refcount); struct xe_exec_queue *eq, *next; + if (q->ufence_syncobj) + drm_syncobj_put(q->ufence_syncobj); + if (xe_exec_queue_uses_pxp(q)) xe_pxp_exec_queue_remove(gt_to_xe(q->gt)->pxp, q); diff --git a/drivers/gpu/drm/xe/xe_exec_queue_types.h b/drivers/gpu/drm/xe/xe_exec_queue_types.h index 27b76cf9da89..df1c69dc81f1 100644 --- a/drivers/gpu/drm/xe/xe_exec_queue_types.h +++ b/drivers/gpu/drm/xe/xe_exec_queue_types.h @@ -15,6 +15,7 @@ #include "xe_hw_fence_types.h" #include "xe_lrc_types.h" +struct drm_syncobj; struct xe_execlist_exec_queue; struct xe_gt; struct xe_guc_exec_queue; @@ -155,6 +156,12 @@ struct xe_exec_queue { struct list_head link; } pxp; + /** @ufence_syncobj: User fence syncobj */ + struct drm_syncobj *ufence_syncobj; + + /** @ufence_timeline_value: User fence timeline value */ + u64 ufence_timeline_value; + /** @ops: submission backend exec queue operations */ const struct xe_exec_queue_ops *ops; diff --git a/drivers/gpu/drm/xe/xe_ggtt.c b/drivers/gpu/drm/xe/xe_ggtt.c index 7fdd0a97a628..5edc0cad47e2 100644 --- a/drivers/gpu/drm/xe/xe_ggtt.c +++ b/drivers/gpu/drm/xe/xe_ggtt.c @@ -292,6 +292,9 @@ int xe_ggtt_init_early(struct xe_ggtt *ggtt) ggtt->pt_ops = &xelp_pt_ops; ggtt->wq = alloc_workqueue("xe-ggtt-wq", 0, WQ_MEM_RECLAIM); + if (!ggtt->wq) + return -ENOMEM; + __xe_ggtt_init_early(ggtt, xe_wopcm_size(xe)); err = drmm_add_action_or_reset(&xe->drm, ggtt_fini_early, ggtt); diff --git a/drivers/gpu/drm/xe/xe_gt.c b/drivers/gpu/drm/xe/xe_gt.c index 3e0ad7e5b5df..6d3db5e55d98 100644 --- a/drivers/gpu/drm/xe/xe_gt.c +++ b/drivers/gpu/drm/xe/xe_gt.c @@ -813,12 +813,16 @@ static int gt_reset(struct xe_gt *gt) unsigned int fw_ref; int err; - if (xe_device_wedged(gt_to_xe(gt))) - return -ECANCELED; + if (xe_device_wedged(gt_to_xe(gt))) { + err = -ECANCELED; + goto err_pm_put; + } /* We only support GT resets with GuC submission */ - if (!xe_device_uc_enabled(gt_to_xe(gt))) - return -ENODEV; + if (!xe_device_uc_enabled(gt_to_xe(gt))) { + err = -ENODEV; + goto err_pm_put; + } xe_gt_info(gt, "reset started\n"); @@ -826,8 +830,6 @@ static int gt_reset(struct xe_gt *gt) if (!err) xe_gt_warn(gt, "reset block failed to get lifted"); - xe_pm_runtime_get(gt_to_xe(gt)); - if (xe_fault_inject_gt_reset()) { err = -ECANCELED; goto err_fail; @@ -874,6 +876,7 @@ err_fail: xe_gt_err(gt, "reset failed (%pe)\n", ERR_PTR(err)); xe_device_declare_wedged(gt_to_xe(gt)); +err_pm_put: xe_pm_runtime_put(gt_to_xe(gt)); return err; @@ -895,7 +898,9 @@ void xe_gt_reset_async(struct xe_gt *gt) return; xe_gt_info(gt, "reset queued\n"); - queue_work(gt->ordered_wq, >->reset.worker); + xe_pm_runtime_get_noresume(gt_to_xe(gt)); + if (!queue_work(gt->ordered_wq, >->reset.worker)) + xe_pm_runtime_put(gt_to_xe(gt)); } void xe_gt_suspend_prepare(struct xe_gt *gt) diff --git a/drivers/gpu/drm/xe/xe_gt_idle.c b/drivers/gpu/drm/xe/xe_gt_idle.c index f8950a52d0a4..bdc9d9877ec4 100644 --- a/drivers/gpu/drm/xe/xe_gt_idle.c +++ b/drivers/gpu/drm/xe/xe_gt_idle.c @@ -124,6 +124,9 @@ void xe_gt_idle_enable_pg(struct xe_gt *gt) if (xe_gt_is_main_type(gt)) gtidle->powergate_enable |= RENDER_POWERGATE_ENABLE; + if (MEDIA_VERx100(xe) >= 1100 && MEDIA_VERx100(xe) < 1255) + gtidle->powergate_enable |= MEDIA_SAMPLERS_POWERGATE_ENABLE; + if (xe->info.platform != XE_DG1) { for (i = XE_HW_ENGINE_VCS0, j = 0; i <= XE_HW_ENGINE_VCS7; ++i, ++j) { if ((gt->info.engine_mask & BIT(i))) @@ -246,6 +249,11 @@ int xe_gt_idle_pg_print(struct xe_gt *gt, struct drm_printer *p) drm_printf(p, "Media Slice%d Power Gate Status: %s\n", n, str_up_down(pg_status & media_slices[n].status_bit)); } + + if (MEDIA_VERx100(xe) >= 1100 && MEDIA_VERx100(xe) < 1255) + drm_printf(p, "Media Samplers Power Gating Enabled: %s\n", + str_yes_no(pg_enabled & MEDIA_SAMPLERS_POWERGATE_ENABLE)); + return 0; } diff --git a/drivers/gpu/drm/xe/xe_guc_ct.c b/drivers/gpu/drm/xe/xe_guc_ct.c index 18f6327bf552..283d846c3512 100644 --- a/drivers/gpu/drm/xe/xe_guc_ct.c +++ b/drivers/gpu/drm/xe/xe_guc_ct.c @@ -200,6 +200,9 @@ static void guc_ct_fini(struct drm_device *drm, void *arg) { struct xe_guc_ct *ct = arg; +#if IS_ENABLED(CONFIG_DRM_XE_DEBUG) + cancel_work_sync(&ct->dead.worker); +#endif ct_exit_safe_mode(ct); destroy_workqueue(ct->g2h_wq); xa_destroy(&ct->fence_lookup); diff --git a/drivers/gpu/drm/xe/xe_guc_submit.c b/drivers/gpu/drm/xe/xe_guc_submit.c index 53024eb5670b..94ed8159496f 100644 --- a/drivers/gpu/drm/xe/xe_guc_submit.c +++ b/drivers/gpu/drm/xe/xe_guc_submit.c @@ -44,6 +44,7 @@ #include "xe_ring_ops_types.h" #include "xe_sched_job.h" #include "xe_trace.h" +#include "xe_uc_fw.h" #include "xe_vm.h" static struct xe_guc * @@ -1489,7 +1490,17 @@ static void __guc_exec_queue_process_msg_cleanup(struct xe_sched_msg *msg) xe_gt_assert(guc_to_gt(guc), !(q->flags & EXEC_QUEUE_FLAG_PERMANENT)); trace_xe_exec_queue_cleanup_entity(q); - if (exec_queue_registered(q)) + /* + * Expected state transitions for cleanup: + * - If the exec queue is registered and GuC firmware is running, we must first + * disable scheduling and deregister the queue to ensure proper teardown and + * resource release in the GuC, then destroy the exec queue on driver side. + * - If the GuC is already stopped (e.g., during driver unload or GPU reset), + * we cannot expect a response for the deregister request. In this case, + * it is safe to directly destroy the exec queue on driver side, as the GuC + * will not process further requests and all resources must be cleaned up locally. + */ + if (exec_queue_registered(q) && xe_uc_fw_is_running(&guc->fw)) disable_scheduling_deregister(guc, q); else __guc_exec_queue_destroy(guc, q); diff --git a/drivers/gpu/drm/xe/xe_migrate.c b/drivers/gpu/drm/xe/xe_migrate.c index 1d667fa36cf3..a36ce7dce8cc 100644 --- a/drivers/gpu/drm/xe/xe_migrate.c +++ b/drivers/gpu/drm/xe/xe_migrate.c @@ -434,7 +434,7 @@ int xe_migrate_init(struct xe_migrate *m) err = xe_migrate_lock_prepare_vm(tile, m, vm); if (err) - return err; + goto err_out; if (xe->info.has_usm) { struct xe_hw_engine *hwe = xe_gt_hw_engine(primary_gt, @@ -2113,7 +2113,9 @@ int xe_migrate_access_memory(struct xe_migrate *m, struct xe_bo *bo, if (current_bytes & ~PAGE_MASK) { int pitch = 4; - current_bytes = min_t(int, current_bytes, S16_MAX * pitch); + current_bytes = min_t(int, current_bytes, + round_down(S16_MAX * pitch, + XE_CACHELINE_BYTES)); } __fence = xe_migrate_vram(m, current_bytes, diff --git a/drivers/gpu/drm/xe/xe_oa.c b/drivers/gpu/drm/xe/xe_oa.c index a4894eb0d7f3..125698a9ecf1 100644 --- a/drivers/gpu/drm/xe/xe_oa.c +++ b/drivers/gpu/drm/xe/xe_oa.c @@ -10,6 +10,7 @@ #include <drm/drm_drv.h> #include <drm/drm_managed.h> +#include <drm/drm_syncobj.h> #include <uapi/drm/xe_drm.h> #include <generated/xe_wa_oob.h> @@ -1389,7 +1390,9 @@ static int xe_oa_user_extensions(struct xe_oa *oa, enum xe_oa_user_extn_from fro return 0; } -static int xe_oa_parse_syncs(struct xe_oa *oa, struct xe_oa_open_param *param) +static int xe_oa_parse_syncs(struct xe_oa *oa, + struct xe_oa_stream *stream, + struct xe_oa_open_param *param) { int ret, num_syncs, num_ufence = 0; @@ -1409,7 +1412,9 @@ static int xe_oa_parse_syncs(struct xe_oa *oa, struct xe_oa_open_param *param) for (num_syncs = 0; num_syncs < param->num_syncs; num_syncs++) { ret = xe_sync_entry_parse(oa->xe, param->xef, ¶m->syncs[num_syncs], - ¶m->syncs_user[num_syncs], 0); + ¶m->syncs_user[num_syncs], + stream->ufence_syncobj, + ++stream->ufence_timeline_value, 0); if (ret) goto err_syncs; @@ -1539,7 +1544,7 @@ static long xe_oa_config_locked(struct xe_oa_stream *stream, u64 arg) return -ENODEV; param.xef = stream->xef; - err = xe_oa_parse_syncs(stream->oa, ¶m); + err = xe_oa_parse_syncs(stream->oa, stream, ¶m); if (err) goto err_config_put; @@ -1635,6 +1640,7 @@ static void xe_oa_destroy_locked(struct xe_oa_stream *stream) if (stream->exec_q) xe_exec_queue_put(stream->exec_q); + drm_syncobj_put(stream->ufence_syncobj); kfree(stream); } @@ -1826,6 +1832,7 @@ static int xe_oa_stream_open_ioctl_locked(struct xe_oa *oa, struct xe_oa_open_param *param) { struct xe_oa_stream *stream; + struct drm_syncobj *ufence_syncobj; int stream_fd; int ret; @@ -1836,17 +1843,31 @@ static int xe_oa_stream_open_ioctl_locked(struct xe_oa *oa, goto exit; } + ret = drm_syncobj_create(&ufence_syncobj, DRM_SYNCOBJ_CREATE_SIGNALED, + NULL); + if (ret) + goto exit; + stream = kzalloc(sizeof(*stream), GFP_KERNEL); if (!stream) { ret = -ENOMEM; - goto exit; + goto err_syncobj; } - + stream->ufence_syncobj = ufence_syncobj; stream->oa = oa; - ret = xe_oa_stream_init(stream, param); + + ret = xe_oa_parse_syncs(oa, stream, param); if (ret) goto err_free; + ret = xe_oa_stream_init(stream, param); + if (ret) { + while (param->num_syncs--) + xe_sync_entry_cleanup(¶m->syncs[param->num_syncs]); + kfree(param->syncs); + goto err_free; + } + if (!param->disabled) { ret = xe_oa_enable_locked(stream); if (ret) @@ -1870,6 +1891,8 @@ err_destroy: xe_oa_stream_destroy(stream); err_free: kfree(stream); +err_syncobj: + drm_syncobj_put(ufence_syncobj); exit: return ret; } @@ -2083,22 +2106,14 @@ int xe_oa_stream_open_ioctl(struct drm_device *dev, u64 data, struct drm_file *f goto err_exec_q; } - ret = xe_oa_parse_syncs(oa, ¶m); - if (ret) - goto err_exec_q; - mutex_lock(¶m.hwe->gt->oa.gt_lock); ret = xe_oa_stream_open_ioctl_locked(oa, ¶m); mutex_unlock(¶m.hwe->gt->oa.gt_lock); if (ret < 0) - goto err_sync_cleanup; + goto err_exec_q; return ret; -err_sync_cleanup: - while (param.num_syncs--) - xe_sync_entry_cleanup(¶m.syncs[param.num_syncs]); - kfree(param.syncs); err_exec_q: if (param.exec_q) xe_exec_queue_put(param.exec_q); diff --git a/drivers/gpu/drm/xe/xe_oa_types.h b/drivers/gpu/drm/xe/xe_oa_types.h index 2628f78c4e8d..daf701b5d48b 100644 --- a/drivers/gpu/drm/xe/xe_oa_types.h +++ b/drivers/gpu/drm/xe/xe_oa_types.h @@ -15,6 +15,8 @@ #include "regs/xe_reg_defs.h" #include "xe_hw_engine_types.h" +struct drm_syncobj; + #define DEFAULT_XE_OA_BUFFER_SIZE SZ_16M enum xe_oa_report_header { @@ -248,6 +250,12 @@ struct xe_oa_stream { /** @xef: xe_file with which the stream was opened */ struct xe_file *xef; + /** @ufence_syncobj: User fence syncobj */ + struct drm_syncobj *ufence_syncobj; + + /** @ufence_timeline_value: User fence timeline value */ + u64 ufence_timeline_value; + /** @last_fence: fence to use in stream destroy when needed */ struct dma_fence *last_fence; diff --git a/drivers/gpu/drm/xe/xe_pci.c b/drivers/gpu/drm/xe/xe_pci.c index be91343829dd..9a6df79fc5b6 100644 --- a/drivers/gpu/drm/xe/xe_pci.c +++ b/drivers/gpu/drm/xe/xe_pci.c @@ -867,6 +867,8 @@ static int xe_pci_probe(struct pci_dev *pdev, const struct pci_device_id *ent) if (err) return err; + xe_vram_resize_bar(xe); + err = xe_device_probe_early(xe); /* * In Boot Survivability mode, no drm card is exposed and driver diff --git a/drivers/gpu/drm/xe/xe_pt.c b/drivers/gpu/drm/xe/xe_pt.c index a1c88f9a6c76..07f96bda638a 100644 --- a/drivers/gpu/drm/xe/xe_pt.c +++ b/drivers/gpu/drm/xe/xe_pt.c @@ -2022,7 +2022,7 @@ static int op_prepare(struct xe_vm *vm, case DRM_GPUVA_OP_MAP: if ((!op->map.immediate && xe_vm_in_fault_mode(vm) && !op->map.invalidate_on_bind) || - op->map.is_cpu_addr_mirror) + (op->map.vma_flags & XE_VMA_SYSTEM_ALLOCATOR)) break; err = bind_op_prepare(vm, tile, pt_update_ops, op->map.vma, @@ -2252,7 +2252,7 @@ static void op_commit(struct xe_vm *vm, switch (op->base.op) { case DRM_GPUVA_OP_MAP: if ((!op->map.immediate && xe_vm_in_fault_mode(vm)) || - op->map.is_cpu_addr_mirror) + (op->map.vma_flags & XE_VMA_SYSTEM_ALLOCATOR)) break; bind_op_commit(vm, tile, pt_update_ops, op->map.vma, fence, diff --git a/drivers/gpu/drm/xe/xe_svm.c b/drivers/gpu/drm/xe/xe_svm.c index 7e2db71ff34e..129e7818565c 100644 --- a/drivers/gpu/drm/xe/xe_svm.c +++ b/drivers/gpu/drm/xe/xe_svm.c @@ -302,6 +302,11 @@ static int xe_svm_range_set_default_attr(struct xe_vm *vm, u64 range_start, u64 if (!vma) return -EINVAL; + if (!(vma->gpuva.flags & XE_VMA_MADV_AUTORESET)) { + drm_dbg(&vm->xe->drm, "Skipping madvise reset for vma.\n"); + return 0; + } + if (xe_vma_has_default_mem_attrs(vma)) return 0; @@ -1034,6 +1039,9 @@ retry: if (err) return err; + dpagemap = xe_vma_resolve_pagemap(vma, tile); + if (!dpagemap && !ctx.devmem_only) + ctx.device_private_page_owner = NULL; range = xe_svm_range_find_or_insert(vm, fault_addr, vma, &ctx); if (IS_ERR(range)) @@ -1054,7 +1062,6 @@ retry: range_debug(range, "PAGE FAULT"); - dpagemap = xe_vma_resolve_pagemap(vma, tile); if (--migrate_try_count >= 0 && xe_svm_range_needs_migrate_to_vram(range, vma, !!dpagemap || ctx.devmem_only)) { ktime_t migrate_start = xe_svm_stats_ktime_get(); @@ -1073,7 +1080,17 @@ retry: drm_dbg(&vm->xe->drm, "VRAM allocation failed, falling back to retrying fault, asid=%u, errno=%pe\n", vm->usm.asid, ERR_PTR(err)); - goto retry; + + /* + * In the devmem-only case, mixed mappings may + * be found. The get_pages function will fix + * these up to a single location, allowing the + * page fault handler to make forward progress. + */ + if (ctx.devmem_only) + goto get_pages; + else + goto retry; } else { drm_err(&vm->xe->drm, "VRAM allocation failed, retry count exceeded, asid=%u, errno=%pe\n", @@ -1083,6 +1100,7 @@ retry: } } +get_pages: get_pages_start = xe_svm_stats_ktime_get(); range_debug(range, "GET PAGES"); diff --git a/drivers/gpu/drm/xe/xe_sync.c b/drivers/gpu/drm/xe/xe_sync.c index 82872a51f098..d48ab7b32ca5 100644 --- a/drivers/gpu/drm/xe/xe_sync.c +++ b/drivers/gpu/drm/xe/xe_sync.c @@ -113,6 +113,8 @@ static void user_fence_cb(struct dma_fence *fence, struct dma_fence_cb *cb) int xe_sync_entry_parse(struct xe_device *xe, struct xe_file *xef, struct xe_sync_entry *sync, struct drm_xe_sync __user *sync_user, + struct drm_syncobj *ufence_syncobj, + u64 ufence_timeline_value, unsigned int flags) { struct drm_xe_sync sync_in; @@ -192,10 +194,15 @@ int xe_sync_entry_parse(struct xe_device *xe, struct xe_file *xef, if (exec) { sync->addr = sync_in.addr; } else { + sync->ufence_timeline_value = ufence_timeline_value; sync->ufence = user_fence_create(xe, sync_in.addr, sync_in.timeline_value); if (XE_IOCTL_DBG(xe, IS_ERR(sync->ufence))) return PTR_ERR(sync->ufence); + sync->ufence_chain_fence = dma_fence_chain_alloc(); + if (!sync->ufence_chain_fence) + return -ENOMEM; + sync->ufence_syncobj = ufence_syncobj; } break; @@ -239,7 +246,12 @@ void xe_sync_entry_signal(struct xe_sync_entry *sync, struct dma_fence *fence) } else if (sync->ufence) { int err; - dma_fence_get(fence); + drm_syncobj_add_point(sync->ufence_syncobj, + sync->ufence_chain_fence, + fence, sync->ufence_timeline_value); + sync->ufence_chain_fence = NULL; + + fence = drm_syncobj_fence_get(sync->ufence_syncobj); user_fence_get(sync->ufence); err = dma_fence_add_callback(fence, &sync->ufence->cb, user_fence_cb); @@ -259,7 +271,8 @@ void xe_sync_entry_cleanup(struct xe_sync_entry *sync) drm_syncobj_put(sync->syncobj); dma_fence_put(sync->fence); dma_fence_chain_free(sync->chain_fence); - if (sync->ufence) + dma_fence_chain_free(sync->ufence_chain_fence); + if (!IS_ERR_OR_NULL(sync->ufence)) user_fence_put(sync->ufence); } diff --git a/drivers/gpu/drm/xe/xe_sync.h b/drivers/gpu/drm/xe/xe_sync.h index 256ffc1e54dc..51f2d803e977 100644 --- a/drivers/gpu/drm/xe/xe_sync.h +++ b/drivers/gpu/drm/xe/xe_sync.h @@ -8,6 +8,7 @@ #include "xe_sync_types.h" +struct drm_syncobj; struct xe_device; struct xe_exec_queue; struct xe_file; @@ -21,6 +22,8 @@ struct xe_vm; int xe_sync_entry_parse(struct xe_device *xe, struct xe_file *xef, struct xe_sync_entry *sync, struct drm_xe_sync __user *sync_user, + struct drm_syncobj *ufence_syncobj, + u64 ufence_timeline_value, unsigned int flags); int xe_sync_entry_add_deps(struct xe_sync_entry *sync, struct xe_sched_job *job); diff --git a/drivers/gpu/drm/xe/xe_sync_types.h b/drivers/gpu/drm/xe/xe_sync_types.h index 30ac3f51993b..b88f1833e28c 100644 --- a/drivers/gpu/drm/xe/xe_sync_types.h +++ b/drivers/gpu/drm/xe/xe_sync_types.h @@ -18,9 +18,12 @@ struct xe_sync_entry { struct drm_syncobj *syncobj; struct dma_fence *fence; struct dma_fence_chain *chain_fence; + struct dma_fence_chain *ufence_chain_fence; + struct drm_syncobj *ufence_syncobj; struct xe_user_fence *ufence; u64 addr; u64 timeline_value; + u64 ufence_timeline_value; u32 type; u32 flags; }; diff --git a/drivers/gpu/drm/xe/xe_validation.h b/drivers/gpu/drm/xe/xe_validation.h index fec331d791e7..b2d09c596714 100644 --- a/drivers/gpu/drm/xe/xe_validation.h +++ b/drivers/gpu/drm/xe/xe_validation.h @@ -166,10 +166,10 @@ xe_validation_device_init(struct xe_validation_device *val) */ DEFINE_CLASS(xe_validation, struct xe_validation_ctx *, if (_T) xe_validation_ctx_fini(_T);, - ({_ret = xe_validation_ctx_init(_ctx, _val, _exec, _flags); - _ret ? NULL : _ctx; }), + ({*_ret = xe_validation_ctx_init(_ctx, _val, _exec, _flags); + *_ret ? NULL : _ctx; }), struct xe_validation_ctx *_ctx, struct xe_validation_device *_val, - struct drm_exec *_exec, const struct xe_val_flags _flags, int _ret); + struct drm_exec *_exec, const struct xe_val_flags _flags, int *_ret); static inline void *class_xe_validation_lock_ptr(class_xe_validation_t *_T) {return *_T; } #define class_xe_validation_is_conditional true @@ -186,7 +186,7 @@ static inline void *class_xe_validation_lock_ptr(class_xe_validation_t *_T) * exhaustive eviction. */ #define xe_validation_guard(_ctx, _val, _exec, _flags, _ret) \ - scoped_guard(xe_validation, _ctx, _val, _exec, _flags, _ret) \ + scoped_guard(xe_validation, _ctx, _val, _exec, _flags, &_ret) \ drm_exec_until_all_locked(_exec) #endif diff --git a/drivers/gpu/drm/xe/xe_vm.c b/drivers/gpu/drm/xe/xe_vm.c index 027e6ce648c5..ccb09ef4ec9e 100644 --- a/drivers/gpu/drm/xe/xe_vm.c +++ b/drivers/gpu/drm/xe/xe_vm.c @@ -616,6 +616,13 @@ static void xe_vma_ops_incr_pt_update_ops(struct xe_vma_ops *vops, u8 tile_mask, vops->pt_update_ops[i].num_ops += inc_val; } +#define XE_VMA_CREATE_MASK ( \ + XE_VMA_READ_ONLY | \ + XE_VMA_DUMPABLE | \ + XE_VMA_SYSTEM_ALLOCATOR | \ + DRM_GPUVA_SPARSE | \ + XE_VMA_MADV_AUTORESET) + static void xe_vm_populate_rebind(struct xe_vma_op *op, struct xe_vma *vma, u8 tile_mask) { @@ -628,8 +635,7 @@ static void xe_vm_populate_rebind(struct xe_vma_op *op, struct xe_vma *vma, op->base.map.gem.offset = vma->gpuva.gem.offset; op->map.vma = vma; op->map.immediate = true; - op->map.dumpable = vma->gpuva.flags & XE_VMA_DUMPABLE; - op->map.is_null = xe_vma_is_null(vma); + op->map.vma_flags = vma->gpuva.flags & XE_VMA_CREATE_MASK; } static int xe_vm_ops_add_rebind(struct xe_vma_ops *vops, struct xe_vma *vma, @@ -932,11 +938,6 @@ static void xe_vma_free(struct xe_vma *vma) kfree(vma); } -#define VMA_CREATE_FLAG_READ_ONLY BIT(0) -#define VMA_CREATE_FLAG_IS_NULL BIT(1) -#define VMA_CREATE_FLAG_DUMPABLE BIT(2) -#define VMA_CREATE_FLAG_IS_SYSTEM_ALLOCATOR BIT(3) - static struct xe_vma *xe_vma_create(struct xe_vm *vm, struct xe_bo *bo, u64 bo_offset_or_userptr, @@ -947,11 +948,8 @@ static struct xe_vma *xe_vma_create(struct xe_vm *vm, struct xe_vma *vma; struct xe_tile *tile; u8 id; - bool read_only = (flags & VMA_CREATE_FLAG_READ_ONLY); - bool is_null = (flags & VMA_CREATE_FLAG_IS_NULL); - bool dumpable = (flags & VMA_CREATE_FLAG_DUMPABLE); - bool is_cpu_addr_mirror = - (flags & VMA_CREATE_FLAG_IS_SYSTEM_ALLOCATOR); + bool is_null = (flags & DRM_GPUVA_SPARSE); + bool is_cpu_addr_mirror = (flags & XE_VMA_SYSTEM_ALLOCATOR); xe_assert(vm->xe, start < end); xe_assert(vm->xe, end < vm->size); @@ -972,10 +970,6 @@ static struct xe_vma *xe_vma_create(struct xe_vm *vm, if (!vma) return ERR_PTR(-ENOMEM); - if (is_cpu_addr_mirror) - vma->gpuva.flags |= XE_VMA_SYSTEM_ALLOCATOR; - if (is_null) - vma->gpuva.flags |= DRM_GPUVA_SPARSE; if (bo) vma->gpuva.gem.obj = &bo->ttm.base; } @@ -986,10 +980,7 @@ static struct xe_vma *xe_vma_create(struct xe_vm *vm, vma->gpuva.vm = &vm->gpuvm; vma->gpuva.va.addr = start; vma->gpuva.va.range = end - start + 1; - if (read_only) - vma->gpuva.flags |= XE_VMA_READ_ONLY; - if (dumpable) - vma->gpuva.flags |= XE_VMA_DUMPABLE; + vma->gpuva.flags = flags; for_each_tile(tile, vm->xe, id) vma->tile_mask |= 0x1 << id; @@ -2272,12 +2263,16 @@ vm_bind_ioctl_ops_create(struct xe_vm *vm, struct xe_vma_ops *vops, if (__op->op == DRM_GPUVA_OP_MAP) { op->map.immediate = flags & DRM_XE_VM_BIND_FLAG_IMMEDIATE; - op->map.read_only = - flags & DRM_XE_VM_BIND_FLAG_READONLY; - op->map.is_null = flags & DRM_XE_VM_BIND_FLAG_NULL; - op->map.is_cpu_addr_mirror = flags & - DRM_XE_VM_BIND_FLAG_CPU_ADDR_MIRROR; - op->map.dumpable = flags & DRM_XE_VM_BIND_FLAG_DUMPABLE; + if (flags & DRM_XE_VM_BIND_FLAG_READONLY) + op->map.vma_flags |= XE_VMA_READ_ONLY; + if (flags & DRM_XE_VM_BIND_FLAG_NULL) + op->map.vma_flags |= DRM_GPUVA_SPARSE; + if (flags & DRM_XE_VM_BIND_FLAG_CPU_ADDR_MIRROR) + op->map.vma_flags |= XE_VMA_SYSTEM_ALLOCATOR; + if (flags & DRM_XE_VM_BIND_FLAG_DUMPABLE) + op->map.vma_flags |= XE_VMA_DUMPABLE; + if (flags & DRM_XE_VM_BIND_FLAG_MADVISE_AUTORESET) + op->map.vma_flags |= XE_VMA_MADV_AUTORESET; op->map.pat_index = pat_index; op->map.invalidate_on_bind = __xe_vm_needs_clear_scratch_pages(vm, flags); @@ -2590,14 +2585,7 @@ static int vm_bind_ioctl_ops_parse(struct xe_vm *vm, struct drm_gpuva_ops *ops, .pat_index = op->map.pat_index, }; - flags |= op->map.read_only ? - VMA_CREATE_FLAG_READ_ONLY : 0; - flags |= op->map.is_null ? - VMA_CREATE_FLAG_IS_NULL : 0; - flags |= op->map.dumpable ? - VMA_CREATE_FLAG_DUMPABLE : 0; - flags |= op->map.is_cpu_addr_mirror ? - VMA_CREATE_FLAG_IS_SYSTEM_ALLOCATOR : 0; + flags |= op->map.vma_flags & XE_VMA_CREATE_MASK; vma = new_vma(vm, &op->base.map, &default_attr, flags); @@ -2606,7 +2594,7 @@ static int vm_bind_ioctl_ops_parse(struct xe_vm *vm, struct drm_gpuva_ops *ops, op->map.vma = vma; if (((op->map.immediate || !xe_vm_in_fault_mode(vm)) && - !op->map.is_cpu_addr_mirror) || + !(op->map.vma_flags & XE_VMA_SYSTEM_ALLOCATOR)) || op->map.invalidate_on_bind) xe_vma_ops_incr_pt_update_ops(vops, op->tile_mask, 1); @@ -2637,18 +2625,7 @@ static int vm_bind_ioctl_ops_parse(struct xe_vm *vm, struct drm_gpuva_ops *ops, op->remap.start = xe_vma_start(old); op->remap.range = xe_vma_size(old); - flags |= op->base.remap.unmap->va->flags & - XE_VMA_READ_ONLY ? - VMA_CREATE_FLAG_READ_ONLY : 0; - flags |= op->base.remap.unmap->va->flags & - DRM_GPUVA_SPARSE ? - VMA_CREATE_FLAG_IS_NULL : 0; - flags |= op->base.remap.unmap->va->flags & - XE_VMA_DUMPABLE ? - VMA_CREATE_FLAG_DUMPABLE : 0; - flags |= xe_vma_is_cpu_addr_mirror(old) ? - VMA_CREATE_FLAG_IS_SYSTEM_ALLOCATOR : 0; - + flags |= op->base.remap.unmap->va->flags & XE_VMA_CREATE_MASK; if (op->base.remap.prev) { vma = new_vma(vm, op->base.remap.prev, &old->attr, flags); @@ -2832,7 +2809,7 @@ static void vm_bind_ioctl_ops_unwind(struct xe_vm *vm, } static int vma_lock_and_validate(struct drm_exec *exec, struct xe_vma *vma, - bool validate) + bool res_evict, bool validate) { struct xe_bo *bo = xe_vma_bo(vma); struct xe_vm *vm = xe_vma_vm(vma); @@ -2843,7 +2820,8 @@ static int vma_lock_and_validate(struct drm_exec *exec, struct xe_vma *vma, err = drm_exec_lock_obj(exec, &bo->ttm.base); if (!err && validate) err = xe_bo_validate(bo, vm, - !xe_vm_in_preempt_fence_mode(vm), exec); + !xe_vm_in_preempt_fence_mode(vm) && + res_evict, exec); } return err; @@ -2913,14 +2891,23 @@ static int prefetch_ranges(struct xe_vm *vm, struct xe_vma_op *op) } static int op_lock_and_prep(struct drm_exec *exec, struct xe_vm *vm, - struct xe_vma_op *op) + struct xe_vma_ops *vops, struct xe_vma_op *op) { int err = 0; + bool res_evict; + + /* + * We only allow evicting a BO within the VM if it is not part of an + * array of binds, as an array of binds can evict another BO within the + * bind. + */ + res_evict = !(vops->flags & XE_VMA_OPS_ARRAY_OF_BINDS); switch (op->base.op) { case DRM_GPUVA_OP_MAP: if (!op->map.invalidate_on_bind) err = vma_lock_and_validate(exec, op->map.vma, + res_evict, !xe_vm_in_fault_mode(vm) || op->map.immediate); break; @@ -2931,11 +2918,13 @@ static int op_lock_and_prep(struct drm_exec *exec, struct xe_vm *vm, err = vma_lock_and_validate(exec, gpuva_to_vma(op->base.remap.unmap->va), - false); + res_evict, false); if (!err && op->remap.prev) - err = vma_lock_and_validate(exec, op->remap.prev, true); + err = vma_lock_and_validate(exec, op->remap.prev, + res_evict, true); if (!err && op->remap.next) - err = vma_lock_and_validate(exec, op->remap.next, true); + err = vma_lock_and_validate(exec, op->remap.next, + res_evict, true); break; case DRM_GPUVA_OP_UNMAP: err = check_ufence(gpuva_to_vma(op->base.unmap.va)); @@ -2944,7 +2933,7 @@ static int op_lock_and_prep(struct drm_exec *exec, struct xe_vm *vm, err = vma_lock_and_validate(exec, gpuva_to_vma(op->base.unmap.va), - false); + res_evict, false); break; case DRM_GPUVA_OP_PREFETCH: { @@ -2959,7 +2948,7 @@ static int op_lock_and_prep(struct drm_exec *exec, struct xe_vm *vm, err = vma_lock_and_validate(exec, gpuva_to_vma(op->base.prefetch.va), - false); + res_evict, false); if (!err && !xe_vma_has_no_bo(vma)) err = xe_bo_migrate(xe_vma_bo(vma), region_to_mem_type[region], @@ -3005,7 +2994,7 @@ static int vm_bind_ioctl_ops_lock_and_prep(struct drm_exec *exec, return err; list_for_each_entry(op, &vops->list, link) { - err = op_lock_and_prep(exec, vm, op); + err = op_lock_and_prep(exec, vm, vops, op); if (err) return err; } @@ -3267,7 +3256,8 @@ ALLOW_ERROR_INJECTION(vm_bind_ioctl_ops_execute, ERRNO); DRM_XE_VM_BIND_FLAG_NULL | \ DRM_XE_VM_BIND_FLAG_DUMPABLE | \ DRM_XE_VM_BIND_FLAG_CHECK_PXP | \ - DRM_XE_VM_BIND_FLAG_CPU_ADDR_MIRROR) + DRM_XE_VM_BIND_FLAG_CPU_ADDR_MIRROR | \ + DRM_XE_VM_BIND_FLAG_MADVISE_AUTORESET) #ifdef TEST_VM_OPS_ERROR #define SUPPORTED_FLAGS (SUPPORTED_FLAGS_STUB | FORCE_OP_ERROR) @@ -3382,7 +3372,9 @@ static int vm_bind_ioctl_check_args(struct xe_device *xe, struct xe_vm *vm, XE_IOCTL_DBG(xe, (prefetch_region != DRM_XE_CONSULT_MEM_ADVISE_PREF_LOC && !(BIT(prefetch_region) & xe->info.mem_region_mask))) || XE_IOCTL_DBG(xe, obj && - op == DRM_XE_VM_BIND_OP_UNMAP)) { + op == DRM_XE_VM_BIND_OP_UNMAP) || + XE_IOCTL_DBG(xe, (flags & DRM_XE_VM_BIND_FLAG_MADVISE_AUTORESET) && + (!is_cpu_addr_mirror || op != DRM_XE_VM_BIND_OP_MAP))) { err = -EINVAL; goto free_bind_ops; } @@ -3614,8 +3606,12 @@ int xe_vm_bind_ioctl(struct drm_device *dev, void *data, struct drm_file *file) syncs_user = u64_to_user_ptr(args->syncs); for (num_syncs = 0; num_syncs < args->num_syncs; num_syncs++) { + struct xe_exec_queue *__q = q ?: vm->q[0]; + err = xe_sync_entry_parse(xe, xef, &syncs[num_syncs], &syncs_user[num_syncs], + __q->ufence_syncobj, + ++__q->ufence_timeline_value, (xe_vm_in_lr_mode(vm) ? SYNC_PARSE_FLAG_LR_MODE : 0) | (!args->num_binds ? @@ -3638,6 +3634,8 @@ int xe_vm_bind_ioctl(struct drm_device *dev, void *data, struct drm_file *file) } xe_vma_ops_init(&vops, vm, q, syncs, num_syncs); + if (args->num_binds > 1) + vops.flags |= XE_VMA_OPS_ARRAY_OF_BINDS; for (i = 0; i < args->num_binds; ++i) { u64 range = bind_ops[i].range; u64 addr = bind_ops[i].addr; @@ -4198,7 +4196,7 @@ static int xe_vm_alloc_vma(struct xe_vm *vm, struct xe_vma_ops vops; struct drm_gpuva_ops *ops = NULL; struct drm_gpuva_op *__op; - bool is_cpu_addr_mirror = false; + unsigned int vma_flags = 0; bool remap_op = false; struct xe_vma_mem_attr tmp_attr; u16 default_pat; @@ -4228,15 +4226,17 @@ static int xe_vm_alloc_vma(struct xe_vm *vm, vma = gpuva_to_vma(op->base.unmap.va); XE_WARN_ON(!xe_vma_has_default_mem_attrs(vma)); default_pat = vma->attr.default_pat_index; + vma_flags = vma->gpuva.flags; } if (__op->op == DRM_GPUVA_OP_REMAP) { vma = gpuva_to_vma(op->base.remap.unmap->va); default_pat = vma->attr.default_pat_index; + vma_flags = vma->gpuva.flags; } if (__op->op == DRM_GPUVA_OP_MAP) { - op->map.is_cpu_addr_mirror = true; + op->map.vma_flags |= vma_flags & XE_VMA_CREATE_MASK; op->map.pat_index = default_pat; } } else { @@ -4245,11 +4245,7 @@ static int xe_vm_alloc_vma(struct xe_vm *vm, xe_assert(vm->xe, !remap_op); xe_assert(vm->xe, xe_vma_has_no_bo(vma)); remap_op = true; - - if (xe_vma_is_cpu_addr_mirror(vma)) - is_cpu_addr_mirror = true; - else - is_cpu_addr_mirror = false; + vma_flags = vma->gpuva.flags; } if (__op->op == DRM_GPUVA_OP_MAP) { @@ -4258,10 +4254,10 @@ static int xe_vm_alloc_vma(struct xe_vm *vm, /* * In case of madvise ops DRM_GPUVA_OP_MAP is * always after DRM_GPUVA_OP_REMAP, so ensure - * we assign op->map.is_cpu_addr_mirror true - * if REMAP is for xe_vma_is_cpu_addr_mirror vma + * to propagate the flags from the vma we're + * unmapping. */ - op->map.is_cpu_addr_mirror = is_cpu_addr_mirror; + op->map.vma_flags |= vma_flags & XE_VMA_CREATE_MASK; } } print_op(vm->xe, __op); diff --git a/drivers/gpu/drm/xe/xe_vm_types.h b/drivers/gpu/drm/xe/xe_vm_types.h index da39940501d8..d6e2a0fdd4b3 100644 --- a/drivers/gpu/drm/xe/xe_vm_types.h +++ b/drivers/gpu/drm/xe/xe_vm_types.h @@ -46,6 +46,7 @@ struct xe_vm_pgtable_update_op; #define XE_VMA_PTE_COMPACT (DRM_GPUVA_USERBITS << 7) #define XE_VMA_DUMPABLE (DRM_GPUVA_USERBITS << 8) #define XE_VMA_SYSTEM_ALLOCATOR (DRM_GPUVA_USERBITS << 9) +#define XE_VMA_MADV_AUTORESET (DRM_GPUVA_USERBITS << 10) /** * struct xe_vma_mem_attr - memory attributes associated with vma @@ -345,17 +346,10 @@ struct xe_vm { struct xe_vma_op_map { /** @vma: VMA to map */ struct xe_vma *vma; + unsigned int vma_flags; /** @immediate: Immediate bind */ bool immediate; /** @read_only: Read only */ - bool read_only; - /** @is_null: is NULL binding */ - bool is_null; - /** @is_cpu_addr_mirror: is CPU address mirror binding */ - bool is_cpu_addr_mirror; - /** @dumpable: whether BO is dumped on GPU hang */ - bool dumpable; - /** @invalidate: invalidate the VMA before bind */ bool invalidate_on_bind; /** @pat_index: The pat index to use for this operation. */ u16 pat_index; @@ -476,6 +470,7 @@ struct xe_vma_ops { /** @flag: signify the properties within xe_vma_ops*/ #define XE_VMA_OPS_FLAG_HAS_SVM_PREFETCH BIT(0) #define XE_VMA_OPS_FLAG_MADVISE BIT(1) +#define XE_VMA_OPS_ARRAY_OF_BINDS BIT(2) u32 flags; #ifdef TEST_VM_OPS_ERROR /** @inject_error: inject error to test error handling */ diff --git a/drivers/gpu/drm/xe/xe_vram.c b/drivers/gpu/drm/xe/xe_vram.c index b44ebf50fedb..652df7a5f4f6 100644 --- a/drivers/gpu/drm/xe/xe_vram.c +++ b/drivers/gpu/drm/xe/xe_vram.c @@ -26,15 +26,35 @@ #define BAR_SIZE_SHIFT 20 -static void -_resize_bar(struct xe_device *xe, int resno, resource_size_t size) +/* + * Release all the BARs that could influence/block LMEMBAR resizing, i.e. + * assigned IORESOURCE_MEM_64 BARs + */ +static void release_bars(struct pci_dev *pdev) +{ + struct resource *res; + int i; + + pci_dev_for_each_resource(pdev, res, i) { + /* Resource already un-assigned, do not reset it */ + if (!res->parent) + continue; + + /* No need to release unrelated BARs */ + if (!(res->flags & IORESOURCE_MEM_64)) + continue; + + pci_release_resource(pdev, i); + } +} + +static void resize_bar(struct xe_device *xe, int resno, resource_size_t size) { struct pci_dev *pdev = to_pci_dev(xe->drm.dev); int bar_size = pci_rebar_bytes_to_size(size); int ret; - if (pci_resource_len(pdev, resno)) - pci_release_resource(pdev, resno); + release_bars(pdev); ret = pci_resize_resource(pdev, resno, bar_size); if (ret) { @@ -50,7 +70,7 @@ _resize_bar(struct xe_device *xe, int resno, resource_size_t size) * if force_vram_bar_size is set, attempt to set to the requested size * else set to maximum possible size */ -static void resize_vram_bar(struct xe_device *xe) +void xe_vram_resize_bar(struct xe_device *xe) { int force_vram_bar_size = xe_modparam.force_vram_bar_size; struct pci_dev *pdev = to_pci_dev(xe->drm.dev); @@ -119,7 +139,7 @@ static void resize_vram_bar(struct xe_device *xe) pci_read_config_dword(pdev, PCI_COMMAND, &pci_cmd); pci_write_config_dword(pdev, PCI_COMMAND, pci_cmd & ~PCI_COMMAND_MEMORY); - _resize_bar(xe, LMEM_BAR, rebar_size); + resize_bar(xe, LMEM_BAR, rebar_size); pci_assign_unassigned_bus_resources(pdev->bus); pci_write_config_dword(pdev, PCI_COMMAND, pci_cmd); @@ -148,8 +168,6 @@ static int determine_lmem_bar_size(struct xe_device *xe, struct xe_vram_region * return -ENXIO; } - resize_vram_bar(xe); - lmem_bar->io_start = pci_resource_start(pdev, LMEM_BAR); lmem_bar->io_size = pci_resource_len(pdev, LMEM_BAR); if (!lmem_bar->io_size) diff --git a/drivers/gpu/drm/xe/xe_vram.h b/drivers/gpu/drm/xe/xe_vram.h index 72860f714fc6..13505cfb184d 100644 --- a/drivers/gpu/drm/xe/xe_vram.h +++ b/drivers/gpu/drm/xe/xe_vram.h @@ -11,6 +11,7 @@ struct xe_device; struct xe_vram_region; +void xe_vram_resize_bar(struct xe_device *xe); int xe_vram_probe(struct xe_device *xe); struct xe_vram_region *xe_vram_region_alloc(struct xe_device *xe, u8 id, u32 placement); |