diff options
Diffstat (limited to 'drivers/gpu/drm/xe')
29 files changed, 269 insertions, 255 deletions
diff --git a/drivers/gpu/drm/xe/display/xe_display.c b/drivers/gpu/drm/xe/display/xe_display.c index 75736faf2a80..c6e0c8d77a70 100644 --- a/drivers/gpu/drm/xe/display/xe_display.c +++ b/drivers/gpu/drm/xe/display/xe_display.c @@ -309,18 +309,7 @@ static void xe_display_flush_cleanup_work(struct xe_device *xe) } /* TODO: System and runtime suspend/resume sequences will be sanitized as a follow-up. */ -void xe_display_pm_runtime_suspend(struct xe_device *xe) -{ - if (!xe->info.probe_display) - return; - - if (xe->d3cold.allowed) - xe_display_pm_suspend(xe, true); - - intel_hpd_poll_enable(xe); -} - -void xe_display_pm_suspend(struct xe_device *xe, bool runtime) +static void __xe_display_pm_suspend(struct xe_device *xe, bool runtime) { struct intel_display *display = &xe->display; bool s2idle = suspend_to_idle(); @@ -353,28 +342,38 @@ void xe_display_pm_suspend(struct xe_device *xe, bool runtime) intel_opregion_suspend(display, s2idle ? PCI_D1 : PCI_D3cold); intel_dmc_suspend(xe); + + if (runtime && has_display(xe)) + intel_hpd_poll_enable(xe); } -void xe_display_pm_suspend_late(struct xe_device *xe) +void xe_display_pm_suspend(struct xe_device *xe) +{ + __xe_display_pm_suspend(xe, false); +} + +void xe_display_pm_runtime_suspend(struct xe_device *xe) { - bool s2idle = suspend_to_idle(); if (!xe->info.probe_display) return; - intel_power_domains_suspend(xe, s2idle); + if (xe->d3cold.allowed) { + __xe_display_pm_suspend(xe, true); + return; + } - intel_display_power_suspend_late(xe); + intel_hpd_poll_enable(xe); } -void xe_display_pm_runtime_resume(struct xe_device *xe) +void xe_display_pm_suspend_late(struct xe_device *xe) { + bool s2idle = suspend_to_idle(); if (!xe->info.probe_display) return; - intel_hpd_poll_disable(xe); + intel_power_domains_suspend(xe, s2idle); - if (xe->d3cold.allowed) - xe_display_pm_resume(xe, true); + intel_display_power_suspend_late(xe); } void xe_display_pm_resume_early(struct xe_device *xe) @@ -387,7 +386,7 @@ void xe_display_pm_resume_early(struct xe_device *xe) intel_power_domains_resume(xe); } -void xe_display_pm_resume(struct xe_device *xe, bool runtime) +static void __xe_display_pm_resume(struct xe_device *xe, bool runtime) { struct intel_display *display = &xe->display; @@ -411,9 +410,11 @@ void xe_display_pm_resume(struct xe_device *xe, bool runtime) intel_display_driver_resume(xe); drm_kms_helper_poll_enable(&xe->drm); intel_display_driver_enable_user_access(xe); - intel_hpd_poll_disable(xe); } + if (has_display(xe)) + intel_hpd_poll_disable(xe); + intel_opregion_resume(display); intel_fbdev_set_suspend(&xe->drm, FBINFO_STATE_RUNNING, false); @@ -421,6 +422,26 @@ void xe_display_pm_resume(struct xe_device *xe, bool runtime) intel_power_domains_enable(xe); } +void xe_display_pm_resume(struct xe_device *xe) +{ + __xe_display_pm_resume(xe, false); +} + +void xe_display_pm_runtime_resume(struct xe_device *xe) +{ + if (!xe->info.probe_display) + return; + + if (xe->d3cold.allowed) { + __xe_display_pm_resume(xe, true); + return; + } + + intel_hpd_init(xe); + intel_hpd_poll_disable(xe); +} + + static void display_device_remove(struct drm_device *dev, void *arg) { struct xe_device *xe = arg; diff --git a/drivers/gpu/drm/xe/display/xe_display.h b/drivers/gpu/drm/xe/display/xe_display.h index 53d727fd792b..bed55fd26f30 100644 --- a/drivers/gpu/drm/xe/display/xe_display.h +++ b/drivers/gpu/drm/xe/display/xe_display.h @@ -34,10 +34,10 @@ void xe_display_irq_enable(struct xe_device *xe, u32 gu_misc_iir); void xe_display_irq_reset(struct xe_device *xe); void xe_display_irq_postinstall(struct xe_device *xe, struct xe_gt *gt); -void xe_display_pm_suspend(struct xe_device *xe, bool runtime); +void xe_display_pm_suspend(struct xe_device *xe); void xe_display_pm_suspend_late(struct xe_device *xe); void xe_display_pm_resume_early(struct xe_device *xe); -void xe_display_pm_resume(struct xe_device *xe, bool runtime); +void xe_display_pm_resume(struct xe_device *xe); void xe_display_pm_runtime_suspend(struct xe_device *xe); void xe_display_pm_runtime_resume(struct xe_device *xe); @@ -65,10 +65,10 @@ static inline void xe_display_irq_enable(struct xe_device *xe, u32 gu_misc_iir) static inline void xe_display_irq_reset(struct xe_device *xe) {} static inline void xe_display_irq_postinstall(struct xe_device *xe, struct xe_gt *gt) {} -static inline void xe_display_pm_suspend(struct xe_device *xe, bool runtime) {} +static inline void xe_display_pm_suspend(struct xe_device *xe) {} static inline void xe_display_pm_suspend_late(struct xe_device *xe) {} static inline void xe_display_pm_resume_early(struct xe_device *xe) {} -static inline void xe_display_pm_resume(struct xe_device *xe, bool runtime) {} +static inline void xe_display_pm_resume(struct xe_device *xe) {} static inline void xe_display_pm_runtime_suspend(struct xe_device *xe) {} static inline void xe_display_pm_runtime_resume(struct xe_device *xe) {} diff --git a/drivers/gpu/drm/xe/regs/xe_gt_regs.h b/drivers/gpu/drm/xe/regs/xe_gt_regs.h index ac9c437e103d..bd604b9f08e4 100644 --- a/drivers/gpu/drm/xe/regs/xe_gt_regs.h +++ b/drivers/gpu/drm/xe/regs/xe_gt_regs.h @@ -393,9 +393,6 @@ #define XE2_GLOBAL_INVAL XE_REG(0xb404) -#define SCRATCH1LPFC XE_REG(0xb474) -#define EN_L3_RW_CCS_CACHE_FLUSH REG_BIT(0) - #define XE2LPM_L3SQCREG2 XE_REG_MCR(0xb604) #define XE2LPM_L3SQCREG3 XE_REG_MCR(0xb608) @@ -520,7 +517,7 @@ * [4-6] RSVD * [7] Disabled */ -#define CCS_MODE XE_REG(0x14804) +#define CCS_MODE XE_REG(0x14804, XE_REG_OPTION_MASKED) #define CCS_MODE_CSLICE_0_3_MASK REG_GENMASK(11, 0) /* 3 bits per cslice */ #define CCS_MODE_CSLICE_MASK 0x7 /* CCS0-3 + rsvd */ #define CCS_MODE_CSLICE_WIDTH ilog2(CCS_MODE_CSLICE_MASK + 1) diff --git a/drivers/gpu/drm/xe/xe_bo.c b/drivers/gpu/drm/xe/xe_bo.c index e5f51fd23c65..2a093540354e 100644 --- a/drivers/gpu/drm/xe/xe_bo.c +++ b/drivers/gpu/drm/xe/xe_bo.c @@ -886,8 +886,8 @@ int xe_bo_evict_pinned(struct xe_bo *bo) if (WARN_ON(!xe_bo_is_pinned(bo))) return -EINVAL; - if (WARN_ON(!xe_bo_is_vram(bo))) - return -EINVAL; + if (!xe_bo_is_vram(bo)) + return 0; ret = ttm_bo_mem_space(&bo->ttm, &placement, &new_mem, &ctx); if (ret) @@ -937,6 +937,7 @@ int xe_bo_restore_pinned(struct xe_bo *bo) .interruptible = false, }; struct ttm_resource *new_mem; + struct ttm_place *place = &bo->placements[0]; int ret; xe_bo_assert_held(bo); @@ -947,9 +948,15 @@ int xe_bo_restore_pinned(struct xe_bo *bo) if (WARN_ON(!xe_bo_is_pinned(bo))) return -EINVAL; - if (WARN_ON(xe_bo_is_vram(bo) || !bo->ttm.ttm)) + if (WARN_ON(xe_bo_is_vram(bo))) + return -EINVAL; + + if (WARN_ON(!bo->ttm.ttm && !xe_bo_is_stolen(bo))) return -EINVAL; + if (!mem_type_is_vram(place->mem_type)) + return 0; + ret = ttm_bo_mem_space(&bo->ttm, &bo->placement, &new_mem, &ctx); if (ret) return ret; @@ -1719,6 +1726,7 @@ int xe_bo_pin_external(struct xe_bo *bo) int xe_bo_pin(struct xe_bo *bo) { + struct ttm_place *place = &bo->placements[0]; struct xe_device *xe = xe_bo_device(bo); int err; @@ -1749,21 +1757,21 @@ int xe_bo_pin(struct xe_bo *bo) */ if (IS_DGFX(xe) && !(IS_ENABLED(CONFIG_DRM_XE_DEBUG) && bo->flags & XE_BO_FLAG_INTERNAL_TEST)) { - struct ttm_place *place = &(bo->placements[0]); - if (mem_type_is_vram(place->mem_type)) { xe_assert(xe, place->flags & TTM_PL_FLAG_CONTIGUOUS); place->fpfn = (xe_bo_addr(bo, 0, PAGE_SIZE) - vram_region_gpu_offset(bo->ttm.resource)) >> PAGE_SHIFT; place->lpfn = place->fpfn + (bo->size >> PAGE_SHIFT); - - spin_lock(&xe->pinned.lock); - list_add_tail(&bo->pinned_link, &xe->pinned.kernel_bo_present); - spin_unlock(&xe->pinned.lock); } } + if (mem_type_is_vram(place->mem_type) || bo->flags & XE_BO_FLAG_GGTT) { + spin_lock(&xe->pinned.lock); + list_add_tail(&bo->pinned_link, &xe->pinned.kernel_bo_present); + spin_unlock(&xe->pinned.lock); + } + ttm_bo_pin(&bo->ttm); /* @@ -1809,23 +1817,18 @@ void xe_bo_unpin_external(struct xe_bo *bo) void xe_bo_unpin(struct xe_bo *bo) { + struct ttm_place *place = &bo->placements[0]; struct xe_device *xe = xe_bo_device(bo); xe_assert(xe, !bo->ttm.base.import_attach); xe_assert(xe, xe_bo_is_pinned(bo)); - if (IS_DGFX(xe) && !(IS_ENABLED(CONFIG_DRM_XE_DEBUG) && - bo->flags & XE_BO_FLAG_INTERNAL_TEST)) { - struct ttm_place *place = &(bo->placements[0]); - - if (mem_type_is_vram(place->mem_type)) { - spin_lock(&xe->pinned.lock); - xe_assert(xe, !list_empty(&bo->pinned_link)); - list_del_init(&bo->pinned_link); - spin_unlock(&xe->pinned.lock); - } + if (mem_type_is_vram(place->mem_type) || bo->flags & XE_BO_FLAG_GGTT) { + spin_lock(&xe->pinned.lock); + xe_assert(xe, !list_empty(&bo->pinned_link)); + list_del_init(&bo->pinned_link); + spin_unlock(&xe->pinned.lock); } - ttm_bo_unpin(&bo->ttm); } diff --git a/drivers/gpu/drm/xe/xe_bo_evict.c b/drivers/gpu/drm/xe/xe_bo_evict.c index 541b49007d73..8fb2be061003 100644 --- a/drivers/gpu/drm/xe/xe_bo_evict.c +++ b/drivers/gpu/drm/xe/xe_bo_evict.c @@ -34,14 +34,22 @@ int xe_bo_evict_all(struct xe_device *xe) u8 id; int ret; - if (!IS_DGFX(xe)) - return 0; - /* User memory */ - for (mem_type = XE_PL_VRAM0; mem_type <= XE_PL_VRAM1; ++mem_type) { + for (mem_type = XE_PL_TT; mem_type <= XE_PL_VRAM1; ++mem_type) { struct ttm_resource_manager *man = ttm_manager_type(bdev, mem_type); + /* + * On igpu platforms with flat CCS we need to ensure we save and restore any CCS + * state since this state lives inside graphics stolen memory which doesn't survive + * hibernation. + * + * This can be further improved by only evicting objects that we know have actually + * used a compression enabled PAT index. + */ + if (mem_type == XE_PL_TT && (IS_DGFX(xe) || !xe_device_has_flat_ccs(xe))) + continue; + if (man) { ret = ttm_resource_manager_evict_all(bdev, man); if (ret) @@ -125,9 +133,6 @@ int xe_bo_restore_kernel(struct xe_device *xe) struct xe_bo *bo; int ret; - if (!IS_DGFX(xe)) - return 0; - spin_lock(&xe->pinned.lock); for (;;) { bo = list_first_entry_or_null(&xe->pinned.evicted, @@ -159,7 +164,6 @@ int xe_bo_restore_kernel(struct xe_device *xe) * should setup the iosys map. */ xe_assert(xe, !iosys_map_is_null(&bo->vmap)); - xe_assert(xe, xe_bo_is_vram(bo)); xe_bo_put(bo); diff --git a/drivers/gpu/drm/xe/xe_debugfs.c b/drivers/gpu/drm/xe/xe_debugfs.c index 668615c6b172..fe4319eb13fd 100644 --- a/drivers/gpu/drm/xe/xe_debugfs.c +++ b/drivers/gpu/drm/xe/xe_debugfs.c @@ -187,7 +187,7 @@ void xe_debugfs_register(struct xe_device *xe) debugfs_create_file("forcewake_all", 0400, root, xe, &forcewake_all_fops); - debugfs_create_file("wedged_mode", 0400, root, xe, + debugfs_create_file("wedged_mode", 0600, root, xe, &wedged_mode_fops); for (mem_type = XE_PL_VRAM0; mem_type <= XE_PL_VRAM1; ++mem_type) { diff --git a/drivers/gpu/drm/xe/xe_device.c b/drivers/gpu/drm/xe/xe_device.c index 5a63d135ba96..a1987b554a8d 100644 --- a/drivers/gpu/drm/xe/xe_device.c +++ b/drivers/gpu/drm/xe/xe_device.c @@ -87,10 +87,6 @@ static int xe_file_open(struct drm_device *dev, struct drm_file *file) mutex_init(&xef->exec_queue.lock); xa_init_flags(&xef->exec_queue.xa, XA_FLAGS_ALLOC1); - spin_lock(&xe->clients.lock); - xe->clients.count++; - spin_unlock(&xe->clients.lock); - file->driver_priv = xef; kref_init(&xef->refcount); @@ -107,17 +103,12 @@ static int xe_file_open(struct drm_device *dev, struct drm_file *file) static void xe_file_destroy(struct kref *ref) { struct xe_file *xef = container_of(ref, struct xe_file, refcount); - struct xe_device *xe = xef->xe; xa_destroy(&xef->exec_queue.xa); mutex_destroy(&xef->exec_queue.lock); xa_destroy(&xef->vm.xa); mutex_destroy(&xef->vm.lock); - spin_lock(&xe->clients.lock); - xe->clients.count--; - spin_unlock(&xe->clients.lock); - xe_drm_client_put(xef->client); kfree(xef->process_name); kfree(xef); @@ -333,7 +324,6 @@ struct xe_device *xe_device_create(struct pci_dev *pdev, xe->info.force_execlist = xe_modparam.force_execlist; spin_lock_init(&xe->irq.lock); - spin_lock_init(&xe->clients.lock); init_waitqueue_head(&xe->ufence_wq); @@ -890,7 +880,7 @@ void xe_device_l2_flush(struct xe_device *xe) spin_lock(>->global_invl_lock); xe_mmio_write32(gt, XE2_GLOBAL_INVAL, 0x1); - if (xe_mmio_wait32(gt, XE2_GLOBAL_INVAL, 0x1, 0x0, 150, NULL, true)) + if (xe_mmio_wait32(gt, XE2_GLOBAL_INVAL, 0x1, 0x0, 500, NULL, true)) xe_gt_err_once(gt, "Global invalidation timeout\n"); spin_unlock(>->global_invl_lock); @@ -980,13 +970,13 @@ void xe_device_declare_wedged(struct xe_device *xe) return; } + xe_pm_runtime_get_noresume(xe); + if (drmm_add_action_or_reset(&xe->drm, xe_device_wedged_fini, xe)) { drm_err(&xe->drm, "Failed to register xe_device_wedged_fini clean-up. Although device is wedged.\n"); return; } - xe_pm_runtime_get_noresume(xe); - if (!atomic_xchg(&xe->wedged.flag, 1)) { xe->needs_flr_on_fini = true; drm_err(&xe->drm, diff --git a/drivers/gpu/drm/xe/xe_device.h b/drivers/gpu/drm/xe/xe_device.h index 894f04770454..34620ef855c0 100644 --- a/drivers/gpu/drm/xe/xe_device.h +++ b/drivers/gpu/drm/xe/xe_device.h @@ -178,4 +178,18 @@ void xe_device_declare_wedged(struct xe_device *xe); struct xe_file *xe_file_get(struct xe_file *xef); void xe_file_put(struct xe_file *xef); +/* + * Occasionally it is seen that the G2H worker starts running after a delay of more than + * a second even after being queued and activated by the Linux workqueue subsystem. This + * leads to G2H timeout error. The root cause of issue lies with scheduling latency of + * Lunarlake Hybrid CPU. Issue disappears if we disable Lunarlake atom cores from BIOS + * and this is beyond xe kmd. + * + * TODO: Drop this change once workqueue scheduling delay issue is fixed on LNL Hybrid CPU. + */ +#define LNL_FLUSH_WORKQUEUE(wq__) \ + flush_workqueue(wq__) +#define LNL_FLUSH_WORK(wrk__) \ + flush_work(wrk__) + #endif diff --git a/drivers/gpu/drm/xe/xe_device_types.h b/drivers/gpu/drm/xe/xe_device_types.h index 09d731a9125c..687f3a9039bb 100644 --- a/drivers/gpu/drm/xe/xe_device_types.h +++ b/drivers/gpu/drm/xe/xe_device_types.h @@ -353,15 +353,6 @@ struct xe_device { struct workqueue_struct *wq; } sriov; - /** @clients: drm clients info */ - struct { - /** @clients.lock: Protects drm clients info */ - spinlock_t lock; - - /** @clients.count: number of drm clients */ - u64 count; - } clients; - /** @usm: unified memory state */ struct { /** @usm.asid: convert a ASID to VM */ diff --git a/drivers/gpu/drm/xe/xe_exec.c b/drivers/gpu/drm/xe/xe_exec.c index 7b38485817dc..31cca938956f 100644 --- a/drivers/gpu/drm/xe/xe_exec.c +++ b/drivers/gpu/drm/xe/xe_exec.c @@ -41,11 +41,6 @@ * user knows an exec writes to a BO and reads from the BO in the next exec, it * is the user's responsibility to pass in / out fence between the two execs). * - * Implicit dependencies for external BOs are handled by using the dma-buf - * implicit dependency uAPI (TODO: add link). To make this works each exec must - * install the job's fence into the DMA_RESV_USAGE_WRITE slot of every external - * BO mapped in the VM. - * * We do not allow a user to trigger a bind at exec time rather we have a VM * bind IOCTL which uses the same in / out fence interface as exec. In that * sense, a VM bind is basically the same operation as an exec from the user @@ -59,8 +54,8 @@ * behind any pending kernel operations on any external BOs in VM or any BOs * private to the VM. This is accomplished by the rebinds waiting on BOs * DMA_RESV_USAGE_KERNEL slot (kernel ops) and kernel ops waiting on all BOs - * slots (inflight execs are in the DMA_RESV_USAGE_BOOKING for private BOs and - * in DMA_RESV_USAGE_WRITE for external BOs). + * slots (inflight execs are in the DMA_RESV_USAGE_BOOKKEEP for private BOs and + * for external BOs). * * Rebinds / dma-resv usage applies to non-compute mode VMs only as for compute * mode VMs we use preempt fences and a rebind worker (TODO: add link). @@ -137,12 +132,16 @@ int xe_exec_ioctl(struct drm_device *dev, void *data, struct drm_file *file) if (XE_IOCTL_DBG(xe, !q)) return -ENOENT; - if (XE_IOCTL_DBG(xe, q->flags & EXEC_QUEUE_FLAG_VM)) - return -EINVAL; + if (XE_IOCTL_DBG(xe, q->flags & EXEC_QUEUE_FLAG_VM)) { + err = -EINVAL; + goto err_exec_queue; + } if (XE_IOCTL_DBG(xe, args->num_batch_buffer && - q->width != args->num_batch_buffer)) - return -EINVAL; + q->width != args->num_batch_buffer)) { + err = -EINVAL; + goto err_exec_queue; + } if (XE_IOCTL_DBG(xe, q->ops->reset_status(q))) { err = -ECANCELED; @@ -204,14 +203,14 @@ retry: write_locked = false; } if (err) - goto err_syncs; + goto err_hw_exec_mode; if (write_locked) { err = xe_vm_userptr_pin(vm); downgrade_write(&vm->lock); write_locked = false; if (err) - goto err_hw_exec_mode; + goto err_unlock_list; } if (!args->num_batch_buffer) { @@ -225,6 +224,7 @@ retry: fence = xe_sync_in_fence_get(syncs, num_syncs, q, vm); if (IS_ERR(fence)) { err = PTR_ERR(fence); + xe_vm_unlock(vm); goto err_unlock_list; } for (i = 0; i < num_syncs; i++) @@ -304,7 +304,8 @@ retry: xe_sched_job_arm(job); if (!xe_vm_in_lr_mode(vm)) drm_gpuvm_resv_add_fence(&vm->gpuvm, exec, &job->drm.s_fence->finished, - DMA_RESV_USAGE_BOOKKEEP, DMA_RESV_USAGE_WRITE); + DMA_RESV_USAGE_BOOKKEEP, + DMA_RESV_USAGE_BOOKKEEP); for (i = 0; i < num_syncs; i++) { xe_sync_entry_signal(&syncs[i], &job->drm.s_fence->finished); diff --git a/drivers/gpu/drm/xe/xe_exec_queue.c b/drivers/gpu/drm/xe/xe_exec_queue.c index d098d2dd1b2d..fd0f3b3c9101 100644 --- a/drivers/gpu/drm/xe/xe_exec_queue.c +++ b/drivers/gpu/drm/xe/xe_exec_queue.c @@ -260,8 +260,14 @@ void xe_exec_queue_fini(struct xe_exec_queue *q) { int i; + /* + * Before releasing our ref to lrc and xef, accumulate our run ticks + */ + xe_exec_queue_update_run_ticks(q); + for (i = 0; i < q->width; ++i) xe_lrc_put(q->lrc[i]); + __xe_exec_queue_free(q); } diff --git a/drivers/gpu/drm/xe/xe_force_wake.c b/drivers/gpu/drm/xe/xe_force_wake.c index b263fff15273..7d9fc489dcb8 100644 --- a/drivers/gpu/drm/xe/xe_force_wake.c +++ b/drivers/gpu/drm/xe/xe_force_wake.c @@ -115,9 +115,15 @@ static int __domain_wait(struct xe_gt *gt, struct xe_force_wake_domain *domain, XE_FORCE_WAKE_ACK_TIMEOUT_MS * USEC_PER_MSEC, &value, true); if (ret) - xe_gt_notice(gt, "Force wake domain %d failed to ack %s (%pe) reg[%#x] = %#x\n", - domain->id, str_wake_sleep(wake), ERR_PTR(ret), - domain->reg_ack.addr, value); + xe_gt_err(gt, "Force wake domain %d failed to ack %s (%pe) reg[%#x] = %#x\n", + domain->id, str_wake_sleep(wake), ERR_PTR(ret), + domain->reg_ack.addr, value); + if (value == ~0) { + xe_gt_err(gt, + "Force wake domain %d: %s. MMIO unreliable (forcewake register returns 0xFFFFFFFF)!\n", + domain->id, str_wake_sleep(wake)); + ret = -EIO; + } return ret; } diff --git a/drivers/gpu/drm/xe/xe_ggtt.c b/drivers/gpu/drm/xe/xe_ggtt.c index 2895f154654c..ff19eca5d358 100644 --- a/drivers/gpu/drm/xe/xe_ggtt.c +++ b/drivers/gpu/drm/xe/xe_ggtt.c @@ -397,6 +397,16 @@ static void ggtt_invalidate_gt_tlb(struct xe_gt *gt) static void xe_ggtt_invalidate(struct xe_ggtt *ggtt) { + struct xe_device *xe = tile_to_xe(ggtt->tile); + + /* + * XXX: Barrier for GGTT pages. Unsure exactly why this required but + * without this LNL is having issues with the GuC reading scratch page + * vs. correct GGTT page. Not particularly a hot code path so blindly + * do a mmio read here which results in GuC reading correct GGTT page. + */ + xe_mmio_read32(xe_root_mmio_gt(xe), VF_CAP_REG); + /* Each GT in a tile has its own TLB to cache GGTT lookups */ ggtt_invalidate_gt_tlb(ggtt->tile->primary_gt); ggtt_invalidate_gt_tlb(ggtt->tile->media_gt); diff --git a/drivers/gpu/drm/xe/xe_gpu_scheduler.h b/drivers/gpu/drm/xe/xe_gpu_scheduler.h index 5ad5629a6c60..64b2ae6839db 100644 --- a/drivers/gpu/drm/xe/xe_gpu_scheduler.h +++ b/drivers/gpu/drm/xe/xe_gpu_scheduler.h @@ -63,7 +63,9 @@ xe_sched_invalidate_job(struct xe_sched_job *job, int threshold) static inline void xe_sched_add_pending_job(struct xe_gpu_scheduler *sched, struct xe_sched_job *job) { + spin_lock(&sched->base.job_list_lock); list_add(&job->drm.list, &sched->base.pending_list); + spin_unlock(&sched->base.job_list_lock); } static inline diff --git a/drivers/gpu/drm/xe/xe_gt.c b/drivers/gpu/drm/xe/xe_gt.c index f0dc2bf24c7b..d5fd6a089b7c 100644 --- a/drivers/gpu/drm/xe/xe_gt.c +++ b/drivers/gpu/drm/xe/xe_gt.c @@ -108,7 +108,6 @@ static void xe_gt_enable_host_l2_vram(struct xe_gt *gt) return; if (!xe_gt_is_media_type(gt)) { - xe_mmio_write32(gt, SCRATCH1LPFC, EN_L3_RW_CCS_CACHE_FLUSH); reg = xe_gt_mcr_unicast_read_any(gt, XE2_GAMREQSTRM_CTRL); reg |= CG_DIS_CNTLBUS; xe_gt_mcr_multicast_write(gt, XE2_GAMREQSTRM_CTRL, reg); @@ -874,7 +873,9 @@ int xe_gt_sanitize_freq(struct xe_gt *gt) int ret = 0; if ((!xe_uc_fw_is_available(>->uc.gsc.fw) || - xe_uc_fw_is_loaded(>->uc.gsc.fw)) && XE_WA(gt, 22019338487)) + xe_uc_fw_is_loaded(>->uc.gsc.fw) || + xe_uc_fw_is_in_error_state(>->uc.gsc.fw)) && + XE_WA(gt, 22019338487)) ret = xe_guc_pc_restore_stashed_freq(>->uc.guc.pc); return ret; diff --git a/drivers/gpu/drm/xe/xe_gt_ccs_mode.c b/drivers/gpu/drm/xe/xe_gt_ccs_mode.c index d2e4dc3aaf61..ffcbd05671fc 100644 --- a/drivers/gpu/drm/xe/xe_gt_ccs_mode.c +++ b/drivers/gpu/drm/xe/xe_gt_ccs_mode.c @@ -68,6 +68,12 @@ static void __xe_gt_apply_ccs_mode(struct xe_gt *gt, u32 num_engines) } } + /* + * Mask bits need to be set for the register. Though only Xe2+ + * platforms require setting of mask bits, it won't harm for older + * platforms as these bits are unused there. + */ + mode |= CCS_MODE_CSLICE_0_3_MASK << 16; xe_mmio_write32(gt, CCS_MODE, mode); xe_gt_dbg(gt, "CCS_MODE=%x config:%08x, num_engines:%d, num_slices:%d\n", @@ -133,9 +139,10 @@ ccs_mode_store(struct device *kdev, struct device_attribute *attr, } /* CCS mode can only be updated when there are no drm clients */ - spin_lock(&xe->clients.lock); - if (xe->clients.count) { - spin_unlock(&xe->clients.lock); + mutex_lock(&xe->drm.filelist_mutex); + if (!list_empty(&xe->drm.filelist)) { + mutex_unlock(&xe->drm.filelist_mutex); + xe_gt_dbg(gt, "Rejecting compute mode change as there are active drm clients\n"); return -EBUSY; } @@ -146,7 +153,7 @@ ccs_mode_store(struct device *kdev, struct device_attribute *attr, xe_gt_reset_async(gt); } - spin_unlock(&xe->clients.lock); + mutex_unlock(&xe->drm.filelist_mutex); return count; } diff --git a/drivers/gpu/drm/xe/xe_gt_sriov_pf_config.c b/drivers/gpu/drm/xe/xe_gt_sriov_pf_config.c index 8250ef71e685..afdb477ecf83 100644 --- a/drivers/gpu/drm/xe/xe_gt_sriov_pf_config.c +++ b/drivers/gpu/drm/xe/xe_gt_sriov_pf_config.c @@ -387,6 +387,8 @@ static void pf_release_ggtt(struct xe_tile *tile, struct xe_ggtt_node *node) * the xe_ggtt_clear() called by below xe_ggtt_remove_node(). */ xe_ggtt_node_remove(node, false); + } else { + xe_ggtt_node_fini(node); } } @@ -442,7 +444,7 @@ static int pf_provision_vf_ggtt(struct xe_gt *gt, unsigned int vfid, u64 size) config->ggtt_region = node; return 0; err: - xe_ggtt_node_fini(node); + pf_release_ggtt(tile, node); return err; } diff --git a/drivers/gpu/drm/xe/xe_gt_tlb_invalidation.c b/drivers/gpu/drm/xe/xe_gt_tlb_invalidation.c index cca9cf536f76..9d82ea30f4df 100644 --- a/drivers/gpu/drm/xe/xe_gt_tlb_invalidation.c +++ b/drivers/gpu/drm/xe/xe_gt_tlb_invalidation.c @@ -37,6 +37,15 @@ static long tlb_timeout_jiffies(struct xe_gt *gt) return hw_tlb_timeout + 2 * delay; } +static void xe_gt_tlb_invalidation_fence_fini(struct xe_gt_tlb_invalidation_fence *fence) +{ + if (WARN_ON_ONCE(!fence->gt)) + return; + + xe_pm_runtime_put(gt_to_xe(fence->gt)); + fence->gt = NULL; /* fini() should be called once */ +} + static void __invalidation_fence_signal(struct xe_device *xe, struct xe_gt_tlb_invalidation_fence *fence) { @@ -63,6 +72,8 @@ static void xe_gt_tlb_fence_timeout(struct work_struct *work) struct xe_device *xe = gt_to_xe(gt); struct xe_gt_tlb_invalidation_fence *fence, *next; + LNL_FLUSH_WORK(>->uc.guc.ct.g2h_worker); + spin_lock_irq(>->tlb_invalidation.pending_lock); list_for_each_entry_safe(fence, next, >->tlb_invalidation.pending_fences, link) { @@ -204,7 +215,7 @@ static int send_tlb_invalidation(struct xe_guc *guc, tlb_timeout_jiffies(gt)); } spin_unlock_irq(>->tlb_invalidation.pending_lock); - } else if (ret < 0) { + } else { __invalidation_fence_signal(xe, fence); } if (!ret) { @@ -267,10 +278,8 @@ int xe_gt_tlb_invalidation_ggtt(struct xe_gt *gt) xe_gt_tlb_invalidation_fence_init(gt, &fence, true); ret = xe_gt_tlb_invalidation_guc(gt, &fence); - if (ret < 0) { - xe_gt_tlb_invalidation_fence_fini(&fence); + if (ret) return ret; - } xe_gt_tlb_invalidation_fence_wait(&fence); } else if (xe_device_uc_enabled(xe) && !xe_device_wedged(xe)) { @@ -496,7 +505,8 @@ static const struct dma_fence_ops invalidation_fence_ops = { * @stack: fence is stack variable * * Initialize TLB invalidation fence for use. xe_gt_tlb_invalidation_fence_fini - * must be called if fence is not signaled. + * will be automatically called when fence is signalled (all fences must signal), + * even on error. */ void xe_gt_tlb_invalidation_fence_init(struct xe_gt *gt, struct xe_gt_tlb_invalidation_fence *fence, @@ -516,14 +526,3 @@ void xe_gt_tlb_invalidation_fence_init(struct xe_gt *gt, dma_fence_get(&fence->base); fence->gt = gt; } - -/** - * xe_gt_tlb_invalidation_fence_fini - Finalize TLB invalidation fence - * @fence: TLB invalidation fence to finalize - * - * Drop PM ref which fence took durinig init. - */ -void xe_gt_tlb_invalidation_fence_fini(struct xe_gt_tlb_invalidation_fence *fence) -{ - xe_pm_runtime_put(gt_to_xe(fence->gt)); -} diff --git a/drivers/gpu/drm/xe/xe_gt_tlb_invalidation.h b/drivers/gpu/drm/xe/xe_gt_tlb_invalidation.h index a84065fa324c..f430d5797af7 100644 --- a/drivers/gpu/drm/xe/xe_gt_tlb_invalidation.h +++ b/drivers/gpu/drm/xe/xe_gt_tlb_invalidation.h @@ -28,7 +28,6 @@ int xe_guc_tlb_invalidation_done_handler(struct xe_guc *guc, u32 *msg, u32 len); void xe_gt_tlb_invalidation_fence_init(struct xe_gt *gt, struct xe_gt_tlb_invalidation_fence *fence, bool stack); -void xe_gt_tlb_invalidation_fence_fini(struct xe_gt_tlb_invalidation_fence *fence); static inline void xe_gt_tlb_invalidation_fence_wait(struct xe_gt_tlb_invalidation_fence *fence) diff --git a/drivers/gpu/drm/xe/xe_guc_ct.c b/drivers/gpu/drm/xe/xe_guc_ct.c index f24dd5223926..9c505d3517cd 100644 --- a/drivers/gpu/drm/xe/xe_guc_ct.c +++ b/drivers/gpu/drm/xe/xe_guc_ct.c @@ -667,16 +667,12 @@ static int __guc_ct_send_locked(struct xe_guc_ct *ct, const u32 *action, num_g2h = 1; if (g2h_fence_needs_alloc(g2h_fence)) { - void *ptr; - g2h_fence->seqno = next_ct_seqno(ct, true); - ptr = xa_store(&ct->fence_lookup, - g2h_fence->seqno, - g2h_fence, GFP_ATOMIC); - if (IS_ERR(ptr)) { - ret = PTR_ERR(ptr); + ret = xa_err(xa_store(&ct->fence_lookup, + g2h_fence->seqno, g2h_fence, + GFP_ATOMIC)); + if (ret) goto out; - } } seqno = g2h_fence->seqno; @@ -879,14 +875,11 @@ retry: retry_same_fence: ret = guc_ct_send(ct, action, len, 0, 0, &g2h_fence); if (unlikely(ret == -ENOMEM)) { - void *ptr; - /* Retry allocation /w GFP_KERNEL */ - ptr = xa_store(&ct->fence_lookup, - g2h_fence.seqno, - &g2h_fence, GFP_KERNEL); - if (IS_ERR(ptr)) - return PTR_ERR(ptr); + ret = xa_err(xa_store(&ct->fence_lookup, g2h_fence.seqno, + &g2h_fence, GFP_KERNEL)); + if (ret) + return ret; goto retry_same_fence; } else if (unlikely(ret)) { @@ -903,16 +896,35 @@ retry_same_fence: } ret = wait_event_timeout(ct->g2h_fence_wq, g2h_fence.done, HZ); + if (!ret) { - xe_gt_err(gt, "Timed out wait for G2H, fence %u, action %04x", - g2h_fence.seqno, action[0]); + LNL_FLUSH_WORK(&ct->g2h_worker); + if (g2h_fence.done) { + xe_gt_warn(gt, "G2H fence %u, action %04x, done\n", + g2h_fence.seqno, action[0]); + ret = 1; + } + } + + /* + * Ensure we serialize with completion side to prevent UAF with fence going out of scope on + * the stack, since we have no clue if it will fire after the timeout before we can erase + * from the xa. Also we have some dependent loads and stores below for which we need the + * correct ordering, and we lack the needed barriers. + */ + mutex_lock(&ct->lock); + if (!ret) { + xe_gt_err(gt, "Timed out wait for G2H, fence %u, action %04x, done %s", + g2h_fence.seqno, action[0], str_yes_no(g2h_fence.done)); xa_erase_irq(&ct->fence_lookup, g2h_fence.seqno); + mutex_unlock(&ct->lock); return -ETIME; } if (g2h_fence.retry) { xe_gt_dbg(gt, "H2G action %#x retrying: reason %#x\n", action[0], g2h_fence.reason); + mutex_unlock(&ct->lock); goto retry; } if (g2h_fence.fail) { @@ -921,7 +933,12 @@ retry_same_fence: ret = -EIO; } - return ret > 0 ? response_buffer ? g2h_fence.response_len : g2h_fence.response_data : ret; + if (ret > 0) + ret = response_buffer ? g2h_fence.response_len : g2h_fence.response_data; + + mutex_unlock(&ct->lock); + + return ret; } /** diff --git a/drivers/gpu/drm/xe/xe_guc_submit.c b/drivers/gpu/drm/xe/xe_guc_submit.c index 80062e1d3f66..4f5d00aea716 100644 --- a/drivers/gpu/drm/xe/xe_guc_submit.c +++ b/drivers/gpu/drm/xe/xe_guc_submit.c @@ -224,80 +224,11 @@ static bool exec_queue_killed_or_banned_or_wedged(struct xe_exec_queue *q) EXEC_QUEUE_STATE_BANNED)); } -#ifdef CONFIG_PROVE_LOCKING -static int alloc_submit_wq(struct xe_guc *guc) -{ - int i; - - for (i = 0; i < NUM_SUBMIT_WQ; ++i) { - guc->submission_state.submit_wq_pool[i] = - alloc_ordered_workqueue("submit_wq", 0); - if (!guc->submission_state.submit_wq_pool[i]) - goto err_free; - } - - return 0; - -err_free: - while (i) - destroy_workqueue(guc->submission_state.submit_wq_pool[--i]); - - return -ENOMEM; -} - -static void free_submit_wq(struct xe_guc *guc) -{ - int i; - - for (i = 0; i < NUM_SUBMIT_WQ; ++i) - destroy_workqueue(guc->submission_state.submit_wq_pool[i]); -} - -static struct workqueue_struct *get_submit_wq(struct xe_guc *guc) -{ - int idx = guc->submission_state.submit_wq_idx++ % NUM_SUBMIT_WQ; - - return guc->submission_state.submit_wq_pool[idx]; -} -#else -static int alloc_submit_wq(struct xe_guc *guc) -{ - return 0; -} - -static void free_submit_wq(struct xe_guc *guc) -{ - -} - -static struct workqueue_struct *get_submit_wq(struct xe_guc *guc) -{ - return NULL; -} -#endif - -static void xe_guc_submit_fini(struct xe_guc *guc) -{ - struct xe_device *xe = guc_to_xe(guc); - struct xe_gt *gt = guc_to_gt(guc); - int ret; - - ret = wait_event_timeout(guc->submission_state.fini_wq, - xa_empty(&guc->submission_state.exec_queue_lookup), - HZ * 5); - - drain_workqueue(xe->destroy_wq); - - xe_gt_assert(gt, ret); -} - static void guc_submit_fini(struct drm_device *drm, void *arg) { struct xe_guc *guc = arg; - xe_guc_submit_fini(guc); xa_destroy(&guc->submission_state.exec_queue_lookup); - free_submit_wq(guc); } static void guc_submit_wedged_fini(void *arg) @@ -359,10 +290,6 @@ int xe_guc_submit_init(struct xe_guc *guc, unsigned int num_ids) if (err) return err; - err = alloc_submit_wq(guc); - if (err) - return err; - gt->exec_queue_ops = &guc_exec_queue_ops; xa_init(&guc->submission_state.exec_queue_lookup); @@ -393,7 +320,6 @@ static void __release_guc_id(struct xe_guc *guc, struct xe_exec_queue *q, u32 xa static int alloc_guc_id(struct xe_guc *guc, struct xe_exec_queue *q) { int ret; - void *ptr; int i; /* @@ -413,12 +339,10 @@ static int alloc_guc_id(struct xe_guc *guc, struct xe_exec_queue *q) q->guc->id = ret; for (i = 0; i < q->width; ++i) { - ptr = xa_store(&guc->submission_state.exec_queue_lookup, - q->guc->id + i, q, GFP_NOWAIT); - if (IS_ERR(ptr)) { - ret = PTR_ERR(ptr); + ret = xa_err(xa_store(&guc->submission_state.exec_queue_lookup, + q->guc->id + i, q, GFP_NOWAIT)); + if (ret) goto err_release; - } } return 0; @@ -821,8 +745,6 @@ static void guc_exec_queue_free_job(struct drm_sched_job *drm_job) { struct xe_sched_job *job = to_xe_sched_job(drm_job); - xe_exec_queue_update_run_ticks(job->q); - trace_xe_sched_job_free(job); xe_sched_job_put(job); } @@ -992,12 +914,22 @@ static void xe_guc_exec_queue_lr_cleanup(struct work_struct *w) static bool check_timeout(struct xe_exec_queue *q, struct xe_sched_job *job) { struct xe_gt *gt = guc_to_gt(exec_queue_to_guc(q)); - u32 ctx_timestamp = xe_lrc_ctx_timestamp(q->lrc[0]); - u32 ctx_job_timestamp = xe_lrc_ctx_job_timestamp(q->lrc[0]); + u32 ctx_timestamp, ctx_job_timestamp; u32 timeout_ms = q->sched_props.job_timeout_ms; u32 diff; u64 running_time_ms; + if (!xe_sched_job_started(job)) { + xe_gt_warn(gt, "Check job timeout: seqno=%u, lrc_seqno=%u, guc_id=%d, not started", + xe_sched_job_seqno(job), xe_sched_job_lrc_seqno(job), + q->guc->id); + + return xe_sched_invalidate_job(job, 2); + } + + ctx_timestamp = xe_lrc_ctx_timestamp(q->lrc[0]); + ctx_job_timestamp = xe_lrc_ctx_job_timestamp(q->lrc[0]); + /* * Counter wraps at ~223s at the usual 19.2MHz, be paranoid catch * possible overflows with a high timeout. @@ -1106,10 +1038,13 @@ guc_exec_queue_timedout_job(struct drm_sched_job *drm_job) /* * TDR has fired before free job worker. Common if exec queue - * immediately closed after last fence signaled. + * immediately closed after last fence signaled. Add back to pending + * list so job can be freed and kick scheduler ensuring free job is not + * lost. */ if (test_bit(DMA_FENCE_FLAG_SIGNALED_BIT, &job->fence->flags)) { - guc_exec_queue_free_job(drm_job); + xe_sched_add_pending_job(sched, job); + xe_sched_submission_start(sched); return DRM_GPU_SCHED_STAT_NOMINAL; } @@ -1122,10 +1057,6 @@ guc_exec_queue_timedout_job(struct drm_sched_job *drm_job) exec_queue_killed_or_banned_or_wedged(q) || exec_queue_destroyed(q); - /* Job hasn't started, can't be timed out */ - if (!skip_timeout_check && !xe_sched_job_started(job)) - goto rearm; - /* * XXX: Sampling timeout doesn't work in wedged mode as we have to * modify scheduling state to read timestamp. We could read the @@ -1482,8 +1413,7 @@ static int guc_exec_queue_init(struct xe_exec_queue *q) timeout = (q->vm && xe_vm_in_lr_mode(q->vm)) ? MAX_SCHEDULE_TIMEOUT : msecs_to_jiffies(q->sched_props.job_timeout_ms); err = xe_sched_init(&ge->sched, &drm_sched_ops, &xe_sched_ops, - get_submit_wq(guc), - q->lrc[0]->ring.size / MAX_JOB_SIZE_BYTES, 64, + NULL, q->lrc[0]->ring.size / MAX_JOB_SIZE_BYTES, 64, timeout, guc_to_gt(guc)->ordered_wq, NULL, q->name, gt_to_xe(q->gt)->drm.dev); if (err) @@ -1800,8 +1730,13 @@ void xe_guc_submit_stop(struct xe_guc *guc) mutex_lock(&guc->submission_state.lock); - xa_for_each(&guc->submission_state.exec_queue_lookup, index, q) + xa_for_each(&guc->submission_state.exec_queue_lookup, index, q) { + /* Prevent redundant attempts to stop parallel queues */ + if (q->guc->id != index) + continue; + guc_exec_queue_stop(guc, q); + } mutex_unlock(&guc->submission_state.lock); @@ -1839,8 +1774,13 @@ int xe_guc_submit_start(struct xe_guc *guc) mutex_lock(&guc->submission_state.lock); atomic_dec(&guc->submission_state.stopped); - xa_for_each(&guc->submission_state.exec_queue_lookup, index, q) + xa_for_each(&guc->submission_state.exec_queue_lookup, index, q) { + /* Prevent redundant attempts to start parallel queues */ + if (q->guc->id != index) + continue; + guc_exec_queue_start(q); + } mutex_unlock(&guc->submission_state.lock); wake_up_all(&guc->ct.wq); diff --git a/drivers/gpu/drm/xe/xe_guc_types.h b/drivers/gpu/drm/xe/xe_guc_types.h index 69046f698271..ed150fc09ad0 100644 --- a/drivers/gpu/drm/xe/xe_guc_types.h +++ b/drivers/gpu/drm/xe/xe_guc_types.h @@ -72,13 +72,6 @@ struct xe_guc { atomic_t stopped; /** @submission_state.lock: protects submission state */ struct mutex lock; -#ifdef CONFIG_PROVE_LOCKING -#define NUM_SUBMIT_WQ 256 - /** @submission_state.submit_wq_pool: submission ordered workqueues pool */ - struct workqueue_struct *submit_wq_pool[NUM_SUBMIT_WQ]; - /** @submission_state.submit_wq_idx: submission ordered workqueue index */ - int submit_wq_idx; -#endif /** @submission_state.enabled: submission is enabled */ bool enabled; /** @submission_state.fini_wq: submit fini wait queue */ diff --git a/drivers/gpu/drm/xe/xe_oa.c b/drivers/gpu/drm/xe/xe_oa.c index 2804f14f8f29..78823f53d290 100644 --- a/drivers/gpu/drm/xe/xe_oa.c +++ b/drivers/gpu/drm/xe/xe_oa.c @@ -1206,9 +1206,11 @@ static int xe_oa_release(struct inode *inode, struct file *file) struct xe_oa_stream *stream = file->private_data; struct xe_gt *gt = stream->gt; + xe_pm_runtime_get(gt_to_xe(gt)); mutex_lock(>->oa.gt_lock); xe_oa_destroy_locked(stream); mutex_unlock(>->oa.gt_lock); + xe_pm_runtime_put(gt_to_xe(gt)); /* Release the reference the OA stream kept on the driver */ drm_dev_put(>_to_xe(gt)->drm); diff --git a/drivers/gpu/drm/xe/xe_pm.c b/drivers/gpu/drm/xe/xe_pm.c index 7cf2160fe040..33eb039053e4 100644 --- a/drivers/gpu/drm/xe/xe_pm.c +++ b/drivers/gpu/drm/xe/xe_pm.c @@ -123,7 +123,7 @@ int xe_pm_suspend(struct xe_device *xe) for_each_gt(gt, xe, id) xe_gt_suspend_prepare(gt); - xe_display_pm_suspend(xe, false); + xe_display_pm_suspend(xe); /* FIXME: Super racey... */ err = xe_bo_evict_all(xe); @@ -133,7 +133,7 @@ int xe_pm_suspend(struct xe_device *xe) for_each_gt(gt, xe, id) { err = xe_gt_suspend(gt); if (err) { - xe_display_pm_resume(xe, false); + xe_display_pm_resume(xe); goto err; } } @@ -187,7 +187,7 @@ int xe_pm_resume(struct xe_device *xe) for_each_gt(gt, xe, id) xe_gt_resume(gt); - xe_display_pm_resume(xe, false); + xe_display_pm_resume(xe); err = xe_bo_restore_user(xe); if (err) diff --git a/drivers/gpu/drm/xe/xe_query.c b/drivers/gpu/drm/xe/xe_query.c index 28d9bb3b825d..848da8e68c7a 100644 --- a/drivers/gpu/drm/xe/xe_query.c +++ b/drivers/gpu/drm/xe/xe_query.c @@ -161,7 +161,11 @@ query_engine_cycles(struct xe_device *xe, cpu_clock); xe_force_wake_put(gt_to_fw(gt), XE_FORCEWAKE_ALL); - resp.width = 36; + + if (GRAPHICS_VER(xe) >= 20) + resp.width = 64; + else + resp.width = 36; /* Only write to the output fields of user query */ if (put_user(resp.cpu_timestamp, &query_ptr->cpu_timestamp)) diff --git a/drivers/gpu/drm/xe/xe_sync.c b/drivers/gpu/drm/xe/xe_sync.c index bb3c2a830362..2e72c06fd40d 100644 --- a/drivers/gpu/drm/xe/xe_sync.c +++ b/drivers/gpu/drm/xe/xe_sync.c @@ -54,11 +54,12 @@ static struct xe_user_fence *user_fence_create(struct xe_device *xe, u64 addr, { struct xe_user_fence *ufence; u64 __user *ptr = u64_to_user_ptr(addr); + u64 __maybe_unused prefetch_val; - if (!access_ok(ptr, sizeof(*ptr))) + if (get_user(prefetch_val, ptr)) return ERR_PTR(-EFAULT); - ufence = kmalloc(sizeof(*ufence), GFP_KERNEL); + ufence = kzalloc(sizeof(*ufence), GFP_KERNEL); if (!ufence) return ERR_PTR(-ENOMEM); diff --git a/drivers/gpu/drm/xe/xe_vm.c b/drivers/gpu/drm/xe/xe_vm.c index ce9dca4d4e87..c99380271de6 100644 --- a/drivers/gpu/drm/xe/xe_vm.c +++ b/drivers/gpu/drm/xe/xe_vm.c @@ -3199,10 +3199,8 @@ int xe_vm_invalidate_vma(struct xe_vma *vma) ret = xe_gt_tlb_invalidation_vma(tile->primary_gt, &fence[fence_id], vma); - if (ret < 0) { - xe_gt_tlb_invalidation_fence_fini(&fence[fence_id]); + if (ret) goto wait; - } ++fence_id; if (!tile->media_gt) @@ -3214,10 +3212,8 @@ int xe_vm_invalidate_vma(struct xe_vma *vma) ret = xe_gt_tlb_invalidation_vma(tile->media_gt, &fence[fence_id], vma); - if (ret < 0) { - xe_gt_tlb_invalidation_fence_fini(&fence[fence_id]); + if (ret) goto wait; - } ++fence_id; } } diff --git a/drivers/gpu/drm/xe/xe_wa.c b/drivers/gpu/drm/xe/xe_wa.c index d424992514a4..353936a0f877 100644 --- a/drivers/gpu/drm/xe/xe_wa.c +++ b/drivers/gpu/drm/xe/xe_wa.c @@ -710,6 +710,10 @@ static const struct xe_rtp_entry_sr lrc_was[] = { DIS_PARTIAL_AUTOSTRIP | DIS_AUTOSTRIP)) }, + { XE_RTP_NAME("15016589081"), + XE_RTP_RULES(GRAPHICS_VERSION(2004), ENGINE_CLASS(RENDER)), + XE_RTP_ACTIONS(SET(CHICKEN_RASTER_1, DIS_CLIP_NEGATIVE_BOUNDING_BOX)) + }, /* Xe2_HPG */ { XE_RTP_NAME("15010599737"), diff --git a/drivers/gpu/drm/xe/xe_wait_user_fence.c b/drivers/gpu/drm/xe/xe_wait_user_fence.c index d46fa8374980..5b4264ea38bd 100644 --- a/drivers/gpu/drm/xe/xe_wait_user_fence.c +++ b/drivers/gpu/drm/xe/xe_wait_user_fence.c @@ -155,6 +155,13 @@ int xe_wait_user_fence_ioctl(struct drm_device *dev, void *data, } if (!timeout) { + LNL_FLUSH_WORKQUEUE(xe->ordered_wq); + err = do_compare(addr, args->value, args->mask, + args->op); + if (err <= 0) { + drm_dbg(&xe->drm, "LNL_FLUSH_WORKQUEUE resolved ufence timeout\n"); + break; + } err = -ETIME; break; } @@ -169,9 +176,6 @@ int xe_wait_user_fence_ioctl(struct drm_device *dev, void *data, args->timeout = 0; } - if (!timeout && !(err < 0)) - err = -ETIME; - if (q) xe_exec_queue_put(q); |