diff options
Diffstat (limited to 'drivers/gpu/drm/i915/i915_gem.c')
| -rw-r--r-- | drivers/gpu/drm/i915/i915_gem.c | 132 |
1 files changed, 47 insertions, 85 deletions
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index 19404c96eeb1..73eeb6b1f1cd 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -179,7 +179,7 @@ i915_gem_object_get_pages_phys(struct drm_i915_gem_object *obj) * the alignment of the buddy allocation will naturally match. */ phys = drm_pci_alloc(obj->base.dev, - obj->base.size, + roundup_pow_of_two(obj->base.size), roundup_pow_of_two(obj->base.size)); if (!phys) return ERR_PTR(-ENOMEM); @@ -694,10 +694,10 @@ flush_write_domain(struct drm_i915_gem_object *obj, unsigned int flush_domains) switch (obj->base.write_domain) { case I915_GEM_DOMAIN_GTT: - if (INTEL_GEN(dev_priv) >= 6 && !HAS_LLC(dev_priv)) { + if (!HAS_LLC(dev_priv)) { intel_runtime_pm_get(dev_priv); spin_lock_irq(&dev_priv->uncore.lock); - POSTING_READ_FW(RING_ACTHD(dev_priv->engine[RCS]->mmio_base)); + POSTING_READ_FW(RING_HEAD(dev_priv->engine[RCS]->mmio_base)); spin_unlock_irq(&dev_priv->uncore.lock); intel_runtime_pm_put(dev_priv); } @@ -1013,17 +1013,20 @@ gtt_user_read(struct io_mapping *mapping, loff_t base, int offset, char __user *user_data, int length) { - void *vaddr; + void __iomem *vaddr; unsigned long unwritten; /* We can use the cpu mem copy function because this is X86. */ - vaddr = (void __force *)io_mapping_map_atomic_wc(mapping, base); - unwritten = __copy_to_user_inatomic(user_data, vaddr + offset, length); + vaddr = io_mapping_map_atomic_wc(mapping, base); + unwritten = __copy_to_user_inatomic(user_data, + (void __force *)vaddr + offset, + length); io_mapping_unmap_atomic(vaddr); if (unwritten) { - vaddr = (void __force *) - io_mapping_map_wc(mapping, base, PAGE_SIZE); - unwritten = copy_to_user(user_data, vaddr + offset, length); + vaddr = io_mapping_map_wc(mapping, base, PAGE_SIZE); + unwritten = copy_to_user(user_data, + (void __force *)vaddr + offset, + length); io_mapping_unmap(vaddr); } return unwritten; @@ -1189,18 +1192,18 @@ ggtt_write(struct io_mapping *mapping, loff_t base, int offset, char __user *user_data, int length) { - void *vaddr; + void __iomem *vaddr; unsigned long unwritten; /* We can use the cpu mem copy function because this is X86. */ - vaddr = (void __force *)io_mapping_map_atomic_wc(mapping, base); - unwritten = __copy_from_user_inatomic_nocache(vaddr + offset, + vaddr = io_mapping_map_atomic_wc(mapping, base); + unwritten = __copy_from_user_inatomic_nocache((void __force *)vaddr + offset, user_data, length); io_mapping_unmap_atomic(vaddr); if (unwritten) { - vaddr = (void __force *) - io_mapping_map_wc(mapping, base, PAGE_SIZE); - unwritten = copy_from_user(vaddr + offset, user_data, length); + vaddr = io_mapping_map_wc(mapping, base, PAGE_SIZE); + unwritten = copy_from_user((void __force *)vaddr + offset, + user_data, length); io_mapping_unmap(vaddr); } @@ -2300,7 +2303,7 @@ i915_gem_object_get_pages_gtt(struct drm_i915_gem_object *obj) struct sgt_iter sgt_iter; struct page *page; unsigned long last_pfn = 0; /* suppress gcc warning */ - unsigned int max_segment; + unsigned int max_segment = i915_sg_segment_size(); gfp_t noreclaim; int ret; @@ -2311,10 +2314,6 @@ i915_gem_object_get_pages_gtt(struct drm_i915_gem_object *obj) GEM_BUG_ON(obj->base.read_domains & I915_GEM_GPU_DOMAINS); GEM_BUG_ON(obj->base.write_domain & I915_GEM_GPU_DOMAINS); - max_segment = swiotlb_max_segment(); - if (!max_segment) - max_segment = rounddown(UINT_MAX, PAGE_SIZE); - st = kmalloc(sizeof(*st), GFP_KERNEL); if (st == NULL) return ERR_PTR(-ENOMEM); @@ -2476,8 +2475,6 @@ static int ____i915_gem_object_get_pages(struct drm_i915_gem_object *obj) { struct sg_table *pages; - GEM_BUG_ON(i915_gem_object_has_pinned_pages(obj)); - if (unlikely(obj->mm.madv != I915_MADV_WILLNEED)) { DRM_DEBUG("Attempting to obtain a purgeable object\n"); return -EFAULT; @@ -2507,6 +2504,8 @@ int __i915_gem_object_get_pages(struct drm_i915_gem_object *obj) return err; if (unlikely(IS_ERR_OR_NULL(obj->mm.pages))) { + GEM_BUG_ON(i915_gem_object_has_pinned_pages(obj)); + err = ____i915_gem_object_get_pages(obj); if (err) goto unlock; @@ -2590,6 +2589,8 @@ void *i915_gem_object_pin_map(struct drm_i915_gem_object *obj, if (!atomic_inc_not_zero(&obj->mm.pages_pin_count)) { if (unlikely(IS_ERR_OR_NULL(obj->mm.pages))) { + GEM_BUG_ON(i915_gem_object_has_pinned_pages(obj)); + ret = ____i915_gem_object_get_pages(obj); if (ret) goto err_unlock; @@ -2814,8 +2815,8 @@ i915_gem_reset_prepare_engine(struct intel_engine_cs *engine) * Turning off the engine->irq_tasklet until the reset is over * prevents the race. */ - tasklet_kill(&engine->irq_tasklet); - tasklet_disable(&engine->irq_tasklet); + tasklet_kill(&engine->execlists.irq_tasklet); + tasklet_disable(&engine->execlists.irq_tasklet); if (engine->irq_seqno_barrier) engine->irq_seqno_barrier(engine); @@ -2994,7 +2995,7 @@ void i915_gem_reset(struct drm_i915_private *dev_priv) void i915_gem_reset_finish_engine(struct intel_engine_cs *engine) { - tasklet_enable(&engine->irq_tasklet); + tasklet_enable(&engine->execlists.irq_tasklet); kthread_unpark(engine->breadcrumbs.signaler); } @@ -3021,9 +3022,6 @@ static void nop_submit_request(struct drm_i915_gem_request *request) static void engine_set_wedged(struct intel_engine_cs *engine) { - struct drm_i915_gem_request *request; - unsigned long flags; - /* We need to be sure that no thread is running the old callback as * we install the nop handler (otherwise we would submit a request * to hardware that will never complete). In order to prevent this @@ -3033,40 +3031,7 @@ static void engine_set_wedged(struct intel_engine_cs *engine) engine->submit_request = nop_submit_request; /* Mark all executing requests as skipped */ - spin_lock_irqsave(&engine->timeline->lock, flags); - list_for_each_entry(request, &engine->timeline->requests, link) - if (!i915_gem_request_completed(request)) - dma_fence_set_error(&request->fence, -EIO); - spin_unlock_irqrestore(&engine->timeline->lock, flags); - - /* - * Clear the execlists queue up before freeing the requests, as those - * are the ones that keep the context and ringbuffer backing objects - * pinned in place. - */ - - if (i915.enable_execlists) { - struct execlist_port *port = engine->execlist_port; - unsigned long flags; - unsigned int n; - - spin_lock_irqsave(&engine->timeline->lock, flags); - - for (n = 0; n < ARRAY_SIZE(engine->execlist_port); n++) - i915_gem_request_put(port_request(&port[n])); - memset(engine->execlist_port, 0, sizeof(engine->execlist_port)); - engine->execlist_queue = RB_ROOT; - engine->execlist_first = NULL; - - spin_unlock_irqrestore(&engine->timeline->lock, flags); - - /* The port is checked prior to scheduling a tasklet, but - * just in case we have suspended the tasklet to do the - * wedging make sure that when it wakes, it decides there - * is no work to do by clearing the irq_posted bit. - */ - clear_bit(ENGINE_IRQ_EXECLIST, &engine->irq_posted); - } + engine->cancel_requests(engine); /* Mark all pending requests as complete so that any concurrent * (lockless) lookup doesn't try and wait upon the request as we @@ -3257,11 +3222,11 @@ void i915_gem_close_object(struct drm_gem_object *gem, struct drm_file *file) struct i915_gem_context *ctx = lut->ctx; struct i915_vma *vma; + GEM_BUG_ON(ctx->file_priv == ERR_PTR(-EBADF)); if (ctx->file_priv != fpriv) continue; vma = radix_tree_delete(&ctx->handles_vma, lut->handle); - GEM_BUG_ON(vma->obj != obj); /* We allow the process to have multiple handles to the same @@ -3375,24 +3340,12 @@ static int wait_for_timeline(struct i915_gem_timeline *tl, unsigned int flags) return 0; } -static int wait_for_engine(struct intel_engine_cs *engine, int timeout_ms) -{ - return wait_for(intel_engine_is_idle(engine), timeout_ms); -} - static int wait_for_engines(struct drm_i915_private *i915) { - struct intel_engine_cs *engine; - enum intel_engine_id id; - - for_each_engine(engine, i915, id) { - if (GEM_WARN_ON(wait_for_engine(engine, 50))) { - i915_gem_set_wedged(i915); - return -EIO; - } - - GEM_BUG_ON(intel_engine_get_seqno(engine) != - intel_engine_last_submit(engine)); + if (wait_for(intel_engines_are_idle(i915), 50)) { + DRM_ERROR("Failed to idle engines, declaring wedged!\n"); + i915_gem_set_wedged(i915); + return -EIO; } return 0; @@ -4426,6 +4379,7 @@ static void __i915_gem_free_objects(struct drm_i915_private *i915, llist_for_each_entry_safe(obj, on, freed, freed) { GEM_BUG_ON(obj->bind_count); GEM_BUG_ON(atomic_read(&obj->frontbuffer_bits)); + GEM_BUG_ON(!list_empty(&obj->lut_list)); if (obj->ops->release) obj->ops->release(obj); @@ -4533,6 +4487,12 @@ static void assert_kernel_context_is_current(struct drm_i915_private *dev_priv) void i915_gem_sanitize(struct drm_i915_private *i915) { + if (i915_terminally_wedged(&i915->gpu_error)) { + mutex_lock(&i915->drm.struct_mutex); + i915_gem_unset_wedged(i915); + mutex_unlock(&i915->drm.struct_mutex); + } + /* * If we inherit context state from the BIOS or earlier occupants * of the GPU, the GPU may be in an inconsistent state when we @@ -4572,7 +4532,7 @@ int i915_gem_suspend(struct drm_i915_private *dev_priv) ret = i915_gem_wait_for_idle(dev_priv, I915_WAIT_INTERRUPTIBLE | I915_WAIT_LOCKED); - if (ret) + if (ret && ret != -EIO) goto err_unlock; assert_kernel_context_is_current(dev_priv); @@ -4594,7 +4554,8 @@ int i915_gem_suspend(struct drm_i915_private *dev_priv) * reset the GPU back to its idle, low power state. */ WARN_ON(dev_priv->gt.awake); - WARN_ON(!intel_engines_are_idle(dev_priv)); + if (WARN_ON(!intel_engines_are_idle(dev_priv))) + i915_gem_set_wedged(dev_priv); /* no hope, discard everything */ /* * Neither the BIOS, ourselves or any other kernel @@ -4616,11 +4577,12 @@ int i915_gem_suspend(struct drm_i915_private *dev_priv) * machine in an unusable condition. */ i915_gem_sanitize(dev_priv); - goto out_rpm_put; + + intel_runtime_pm_put(dev_priv); + return 0; err_unlock: mutex_unlock(&dev->struct_mutex); -out_rpm_put: intel_runtime_pm_put(dev_priv); return ret; } @@ -4776,7 +4738,7 @@ bool intel_sanitize_semaphores(struct drm_i915_private *dev_priv, int value) return false; /* TODO: make semaphores and Execlists play nicely together */ - if (i915.enable_execlists) + if (i915_modparams.enable_execlists) return false; if (value >= 0) @@ -4797,7 +4759,7 @@ int i915_gem_init(struct drm_i915_private *dev_priv) dev_priv->mm.unordered_timeline = dma_fence_context_alloc(1); - if (!i915.enable_execlists) { + if (!i915_modparams.enable_execlists) { dev_priv->gt.resume = intel_legacy_submission_resume; dev_priv->gt.cleanup_engine = intel_engine_cleanup; } else { |
