The mm_list is used for the active/inactive LRUs. Since those LRUs are per address space, the link should be per VM area Because we'll only ever have 1 VMA before this point, it's not incorrect to defer this change until this point in the patch series, and doing it here makes the change much easier to understand. v2: only bump GGTT LRU in i915_gem_object_set_to_gtt_domain (Chris) Signed-off-by: Ben Widawsky <ben at bwidawsk.net> --- drivers/gpu/drm/i915/i915_debugfs.c | 49 ++++++++++++++++++++++------------ drivers/gpu/drm/i915/i915_drv.h | 5 ++-- drivers/gpu/drm/i915/i915_gem.c | 31 ++++++++++++++------- drivers/gpu/drm/i915/i915_gem_evict.c | 14 +++++----- drivers/gpu/drm/i915/i915_gem_stolen.c | 2 +- drivers/gpu/drm/i915/i915_irq.c | 13 +++++---- 6 files changed, 71 insertions(+), 43 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c index 7d01fb6..60d2a94 100644 --- a/drivers/gpu/drm/i915/i915_debugfs.c +++ b/drivers/gpu/drm/i915/i915_debugfs.c @@ -152,7 +152,7 @@ static int i915_gem_object_list_info(struct seq_file *m, void *data) struct list_head *head; struct drm_device *dev = node->minor->dev; drm_i915_private_t *dev_priv = dev->dev_private; - struct drm_i915_gem_object *obj; + struct i915_vma *vma; size_t total_obj_size, total_gtt_size; int count, ret; @@ -160,6 +160,7 @@ static int i915_gem_object_list_info(struct seq_file *m, void *data) if (ret) return ret; + /* FIXME: the user of this interface might want more than just GGTT */ switch (list) { case ACTIVE_LIST: seq_printf(m, "Active:\n"); @@ -175,13 +176,12 @@ static int i915_gem_object_list_info(struct seq_file *m, void *data) } total_obj_size = total_gtt_size = count = 0; - list_for_each_entry(obj, head, mm_list) { + list_for_each_entry(vma, head, mm_list) { seq_printf(m, " "); - describe_obj(m, obj); + describe_obj(m, vma->obj); seq_printf(m, "\n"); - total_obj_size += obj->base.size; - /* FIXME: Add size of all VMs */ - total_gtt_size += i915_gem_ggtt_size(obj); + total_obj_size += vma->obj->base.size; + total_gtt_size += i915_gem_ggtt_size(vma->obj); count++; } mutex_unlock(&dev->struct_mutex); @@ -229,6 +229,17 @@ static int per_file_stats(int id, void *ptr, void *data) return 0; } +#define count_vmas(list, member) do { \ + list_for_each_entry(vma, list, member) { \ + size += i915_gem_ggtt_size(vma->obj); \ + ++count; \ + if (vma->obj->map_and_fenceable) { \ + mappable_size += i915_gem_ggtt_size(vma->obj); \ + ++mappable_count; \ + } \ + } \ +} while (0) + static int i915_gem_object_info(struct seq_file *m, void* data) { struct drm_info_node *node = (struct drm_info_node *) m->private; @@ -238,6 +249,7 @@ static int i915_gem_object_info(struct seq_file *m, void* data) size_t size, mappable_size, purgeable_size; struct drm_i915_gem_object *obj; struct drm_file *file; + struct i915_vma *vma; int ret; ret = mutex_lock_interruptible(&dev->struct_mutex); @@ -254,12 +266,12 @@ static int i915_gem_object_info(struct seq_file *m, void* data) count, mappable_count, size, mappable_size); size = count = mappable_size = mappable_count = 0; - count_objects(ggtt_list(active_list), mm_list); + count_vmas(ggtt_list(active_list), mm_list); seq_printf(m, " %u [%u] active objects, %zu [%zu] bytes\n", count, mappable_count, size, mappable_size); size = count = mappable_size = mappable_count = 0; - count_objects(ggtt_list(inactive_list), mm_list); + count_vmas(ggtt_list(inactive_list), mm_list); seq_printf(m, " %u [%u] inactive objects, %zu [%zu] bytes\n", count, mappable_count, size, mappable_size); @@ -1966,6 +1978,8 @@ i915_drop_caches_set(void *data, u64 val) struct drm_device *dev = data; struct drm_i915_private *dev_priv = dev->dev_private; struct drm_i915_gem_object *obj, *next; + struct i915_address_space *vm; + struct i915_vma *vma, *x; int ret; DRM_DEBUG_DRIVER("Dropping caches: 0x%08llx\n", val); @@ -1986,15 +2000,16 @@ i915_drop_caches_set(void *data, u64 val) i915_gem_retire_requests(dev); if (val & DROP_BOUND) { - /* FIXME: Do this for all vms? */ - list_for_each_entry_safe(obj, next, ggtt_list(inactive_list), - mm_list) - if (obj->pin_count) - continue; - - ret = i915_gem_object_unbind(obj, &dev_priv->gtt.base); - if (ret) - goto unlock; + list_for_each_entry(vm, &dev_priv->vm_list, global_link) { + list_for_each_entry_safe(vma, x, &vm->inactive_list, + mm_list) + if (vma->obj->pin_count == 0) { + ret = i915_gem_object_unbind(vma->obj, + vm); + if (ret) + goto unlock; + } + } } if (val & DROP_UNBOUND) { diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index b0d1008..cc18349 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -545,6 +545,9 @@ struct i915_vma { /* Page aligned offset (helper for stolen) */ unsigned long deferred_offset; + /** This object's place on the active/inactive lists */ + struct list_head mm_list; + struct list_head vma_link; /* Link in the object's VMA list */ }; @@ -1237,9 +1240,7 @@ struct drm_i915_gem_object { struct drm_mm_node *stolen; struct list_head global_list; - /** This object's place on the active/inactive lists */ struct list_head ring_list; - struct list_head mm_list; /** This object's place in the batchbuffer or on the eviction list */ struct list_head exec_list; diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index 83e2eb3..950a14b 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -1878,6 +1878,7 @@ i915_gem_object_move_to_active(struct drm_i915_gem_object *obj, struct drm_device *dev = obj->base.dev; struct drm_i915_private *dev_priv = dev->dev_private; u32 seqno = intel_ring_get_seqno(ring); + struct i915_vma *vma; BUG_ON(ring == NULL); obj->ring = ring; @@ -1889,7 +1890,8 @@ i915_gem_object_move_to_active(struct drm_i915_gem_object *obj, } /* Move from whatever list we were on to the tail of execution. */ - list_move_tail(&obj->mm_list, &vm->active_list); + vma = i915_gem_obj_to_vma(obj, vm); + list_move_tail(&vma->mm_list, &vm->active_list); list_move_tail(&obj->ring_list, &ring->active_list); obj->last_read_seqno = seqno; @@ -1912,10 +1914,13 @@ static void i915_gem_object_move_to_inactive(struct drm_i915_gem_object *obj, struct i915_address_space *vm) { + struct i915_vma *vma; + BUG_ON(obj->base.write_domain & ~I915_GEM_GPU_DOMAINS); BUG_ON(!obj->active); - list_move_tail(&obj->mm_list, &vm->inactive_list); + vma = i915_gem_obj_to_vma(obj, vm); + list_move_tail(&vma->mm_list, &vm->inactive_list); list_del_init(&obj->ring_list); obj->ring = NULL; @@ -2290,9 +2295,9 @@ void i915_gem_restore_fences(struct drm_device *dev) bool i915_gem_reset(struct drm_device *dev) { struct drm_i915_private *dev_priv = dev->dev_private; - struct drm_i915_gem_object *obj; struct intel_ring_buffer *ring; struct i915_address_space *vm; + struct i915_vma *vma; int i; bool ctx_banned = false; @@ -2305,8 +2310,8 @@ bool i915_gem_reset(struct drm_device *dev) * necessary invalidation upon reuse. */ list_for_each_entry(vm, &dev_priv->vm_list, global_link) - list_for_each_entry(obj, &vm->inactive_list, mm_list) - obj->base.read_domains &= ~I915_GEM_GPU_DOMAINS; + list_for_each_entry(vma, &vm->inactive_list, mm_list) + vma->obj->base.read_domains &= ~I915_GEM_GPU_DOMAINS; /* The fence registers are invalidated so clear them out */ i915_gem_restore_fences(dev); @@ -2653,12 +2658,12 @@ i915_gem_object_unbind(struct drm_i915_gem_object *obj, i915_gem_gtt_finish_object(obj); i915_gem_object_unpin_pages(obj); - list_del(&obj->mm_list); /* Avoid an unnecessary call to unbind on rebind. */ if (is_i915_ggtt(vm)) obj->map_and_fenceable = true; vma = i915_gem_obj_to_vma(obj, vm); + list_del(&vma->mm_list); list_del(&vma->vma_link); drm_mm_remove_node(&vma->node); i915_gem_vma_destroy(vma); @@ -3208,7 +3213,7 @@ search_free: } list_move_tail(&obj->global_list, &dev_priv->mm.bound_list); - list_add_tail(&obj->mm_list, &vm->inactive_list); + list_add_tail(&vma->mm_list, &vm->inactive_list); /* Keep GGTT vmas first to make debug easier */ if (is_i915_ggtt(vm)) list_add(&vma->vma_link, &obj->vma_list); @@ -3364,8 +3369,14 @@ i915_gem_object_set_to_gtt_domain(struct drm_i915_gem_object *obj, bool write) old_write_domain); /* And bump the LRU for this access */ - if (i915_gem_object_is_inactive(obj)) - list_move_tail(&obj->mm_list, ggtt_list(inactive_list)); + if (i915_gem_object_is_inactive(obj)) { + struct i915_vma *vma = i915_gem_obj_to_vma(obj, + &dev_priv->gtt.base); + if (vma) + list_move_tail(&vma->mm_list, + &dev_priv->gtt.base.inactive_list); + + } return 0; } @@ -3943,7 +3954,6 @@ unlock: void i915_gem_object_init(struct drm_i915_gem_object *obj, const struct drm_i915_gem_object_ops *ops) { - INIT_LIST_HEAD(&obj->mm_list); INIT_LIST_HEAD(&obj->global_list); INIT_LIST_HEAD(&obj->ring_list); INIT_LIST_HEAD(&obj->exec_list); @@ -4086,6 +4096,7 @@ struct i915_vma *i915_gem_vma_create(struct drm_i915_gem_object *obj, return ERR_PTR(-ENOMEM); INIT_LIST_HEAD(&vma->vma_link); + INIT_LIST_HEAD(&vma->mm_list); vma->vm = vm; vma->obj = obj; diff --git a/drivers/gpu/drm/i915/i915_gem_evict.c b/drivers/gpu/drm/i915/i915_gem_evict.c index 7a210b8..028e8b1 100644 --- a/drivers/gpu/drm/i915/i915_gem_evict.c +++ b/drivers/gpu/drm/i915/i915_gem_evict.c @@ -86,8 +86,7 @@ i915_gem_evict_something(struct drm_device *dev, struct i915_address_space *vm, drm_mm_init_scan(&vm->mm, min_size, alignment, cache_level); /* First see if there is a large enough contiguous idle region... */ - list_for_each_entry(obj, &vm->inactive_list, mm_list) { - struct i915_vma *vma = i915_gem_obj_to_vma(obj, vm); + list_for_each_entry(vma, &vm->inactive_list, mm_list) { if (mark_free(vma, &unwind_list)) goto found; } @@ -96,8 +95,7 @@ i915_gem_evict_something(struct drm_device *dev, struct i915_address_space *vm, goto none; /* Now merge in the soon-to-be-expired objects... */ - list_for_each_entry(obj, &vm->active_list, mm_list) { - struct i915_vma *vma = i915_gem_obj_to_vma(obj, vm); + list_for_each_entry(vma, &vm->active_list, mm_list) { if (mark_free(vma, &unwind_list)) goto found; } @@ -158,8 +156,8 @@ int i915_gem_evict_everything(struct drm_device *dev) { drm_i915_private_t *dev_priv = dev->dev_private; - struct drm_i915_gem_object *obj, *next; struct i915_address_space *vm; + struct i915_vma *vma, *next; bool lists_empty = true; int ret; @@ -187,9 +185,9 @@ i915_gem_evict_everything(struct drm_device *dev) /* Having flushed everything, unbind() should never raise an error */ list_for_each_entry(vm, &dev_priv->vm_list, global_link) { - list_for_each_entry_safe(obj, next, &vm->inactive_list, mm_list) - if (obj->pin_count == 0) - WARN_ON(i915_gem_object_unbind(obj, vm)); + list_for_each_entry_safe(vma, next, &vm->inactive_list, mm_list) + if (vma->obj->pin_count == 0) + WARN_ON(i915_gem_object_unbind(vma->obj, vm)); } diff --git a/drivers/gpu/drm/i915/i915_gem_stolen.c b/drivers/gpu/drm/i915/i915_gem_stolen.c index 4863219..d393298 100644 --- a/drivers/gpu/drm/i915/i915_gem_stolen.c +++ b/drivers/gpu/drm/i915/i915_gem_stolen.c @@ -397,7 +397,7 @@ i915_gem_object_create_stolen_for_preallocated(struct drm_device *dev, obj->has_global_gtt_mapping = 1; list_add_tail(&obj->global_list, &dev_priv->mm.bound_list); - list_add_tail(&obj->mm_list, >t_vm->inactive_list); + list_add_tail(&vma->mm_list, >t_vm->inactive_list); return obj; } diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c index b786fcd..28880bf 100644 --- a/drivers/gpu/drm/i915/i915_irq.c +++ b/drivers/gpu/drm/i915/i915_irq.c @@ -1643,11 +1643,11 @@ static void capture_bo(struct drm_i915_error_buffer *err, static u32 capture_active_bo(struct drm_i915_error_buffer *err, int count, struct list_head *head) { - struct drm_i915_gem_object *obj; + struct i915_vma *vma; int i = 0; - list_for_each_entry(obj, head, mm_list) { - capture_bo(err++, obj); + list_for_each_entry(vma, head, mm_list) { + capture_bo(err++, vma->obj); if (++i == count) break; } @@ -1710,6 +1710,7 @@ i915_error_first_batchbuffer(struct drm_i915_private *dev_priv, struct intel_ring_buffer *ring) { struct i915_address_space *vm; + struct i915_vma *vma; struct drm_i915_gem_object *obj; u32 seqno; @@ -1733,7 +1734,8 @@ i915_error_first_batchbuffer(struct drm_i915_private *dev_priv, seqno = ring->get_seqno(ring, false); list_for_each_entry(vm, &dev_priv->vm_list, global_link) { - list_for_each_entry(obj, &vm->active_list, mm_list) { + list_for_each_entry(vma, &vm->active_list, mm_list) { + obj = vma->obj; if (obj->ring != ring) continue; @@ -1872,11 +1874,12 @@ static void i915_gem_capture_vm(struct drm_i915_private *dev_priv, const int ndx) { struct drm_i915_error_buffer *active_bo = NULL, *pinned_bo = NULL; + struct i915_vma *vma; struct drm_i915_gem_object *obj; int i; i = 0; - list_for_each_entry(obj, &vm->active_list, mm_list) + list_for_each_entry(vma, &vm->active_list, mm_list) i++; error->active_bo_count[ndx] = i; list_for_each_entry(obj, &dev_priv->mm.bound_list, global_list) -- 1.8.3.1