Commit 528cbd17 authored by Chris Wilson's avatar Chris Wilson

drm/i915: Move vma lookup to its own lock

Remove the struct_mutex requirement for looking up the vma for an
object.

v2: Highlight how the race for duplicate vma creation is resolved on
reacquiring the lock with a short comment.
Signed-off-by: default avatarChris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: default avatarTvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190128102356.15037-3-chris@chris-wilson.co.uk
parent 09d7e46b
...@@ -159,14 +159,14 @@ describe_obj(struct seq_file *m, struct drm_i915_gem_object *obj) ...@@ -159,14 +159,14 @@ describe_obj(struct seq_file *m, struct drm_i915_gem_object *obj)
obj->mm.madv == I915_MADV_DONTNEED ? " purgeable" : ""); obj->mm.madv == I915_MADV_DONTNEED ? " purgeable" : "");
if (obj->base.name) if (obj->base.name)
seq_printf(m, " (name: %d)", obj->base.name); seq_printf(m, " (name: %d)", obj->base.name);
list_for_each_entry(vma, &obj->vma_list, obj_link) { list_for_each_entry(vma, &obj->vma.list, obj_link) {
if (i915_vma_is_pinned(vma)) if (i915_vma_is_pinned(vma))
pin_count++; pin_count++;
} }
seq_printf(m, " (pinned x %d)", pin_count); seq_printf(m, " (pinned x %d)", pin_count);
if (obj->pin_global) if (obj->pin_global)
seq_printf(m, " (global)"); seq_printf(m, " (global)");
list_for_each_entry(vma, &obj->vma_list, obj_link) { list_for_each_entry(vma, &obj->vma.list, obj_link) {
if (!drm_mm_node_allocated(&vma->node)) if (!drm_mm_node_allocated(&vma->node))
continue; continue;
...@@ -322,7 +322,7 @@ static int per_file_stats(int id, void *ptr, void *data) ...@@ -322,7 +322,7 @@ static int per_file_stats(int id, void *ptr, void *data)
if (obj->base.name || obj->base.dma_buf) if (obj->base.name || obj->base.dma_buf)
stats->shared += obj->base.size; stats->shared += obj->base.size;
list_for_each_entry(vma, &obj->vma_list, obj_link) { list_for_each_entry(vma, &obj->vma.list, obj_link) {
if (!drm_mm_node_allocated(&vma->node)) if (!drm_mm_node_allocated(&vma->node))
continue; continue;
......
...@@ -437,15 +437,19 @@ int i915_gem_object_unbind(struct drm_i915_gem_object *obj) ...@@ -437,15 +437,19 @@ int i915_gem_object_unbind(struct drm_i915_gem_object *obj)
if (ret) if (ret)
return ret; return ret;
while ((vma = list_first_entry_or_null(&obj->vma_list, spin_lock(&obj->vma.lock);
struct i915_vma, while (!ret && (vma = list_first_entry_or_null(&obj->vma.list,
obj_link))) { struct i915_vma,
obj_link))) {
list_move_tail(&vma->obj_link, &still_in_list); list_move_tail(&vma->obj_link, &still_in_list);
spin_unlock(&obj->vma.lock);
ret = i915_vma_unbind(vma); ret = i915_vma_unbind(vma);
if (ret)
break; spin_lock(&obj->vma.lock);
} }
list_splice(&still_in_list, &obj->vma_list); list_splice(&still_in_list, &obj->vma.list);
spin_unlock(&obj->vma.lock);
return ret; return ret;
} }
...@@ -3489,7 +3493,7 @@ int i915_gem_object_set_cache_level(struct drm_i915_gem_object *obj, ...@@ -3489,7 +3493,7 @@ int i915_gem_object_set_cache_level(struct drm_i915_gem_object *obj,
* reading an invalid PTE on older architectures. * reading an invalid PTE on older architectures.
*/ */
restart: restart:
list_for_each_entry(vma, &obj->vma_list, obj_link) { list_for_each_entry(vma, &obj->vma.list, obj_link) {
if (!drm_mm_node_allocated(&vma->node)) if (!drm_mm_node_allocated(&vma->node))
continue; continue;
...@@ -3567,7 +3571,7 @@ int i915_gem_object_set_cache_level(struct drm_i915_gem_object *obj, ...@@ -3567,7 +3571,7 @@ int i915_gem_object_set_cache_level(struct drm_i915_gem_object *obj,
*/ */
} }
list_for_each_entry(vma, &obj->vma_list, obj_link) { list_for_each_entry(vma, &obj->vma.list, obj_link) {
if (!drm_mm_node_allocated(&vma->node)) if (!drm_mm_node_allocated(&vma->node))
continue; continue;
...@@ -3577,7 +3581,7 @@ int i915_gem_object_set_cache_level(struct drm_i915_gem_object *obj, ...@@ -3577,7 +3581,7 @@ int i915_gem_object_set_cache_level(struct drm_i915_gem_object *obj,
} }
} }
list_for_each_entry(vma, &obj->vma_list, obj_link) list_for_each_entry(vma, &obj->vma.list, obj_link)
vma->node.color = cache_level; vma->node.color = cache_level;
i915_gem_object_set_cache_coherency(obj, cache_level); i915_gem_object_set_cache_coherency(obj, cache_level);
obj->cache_dirty = true; /* Always invalidate stale cachelines */ obj->cache_dirty = true; /* Always invalidate stale cachelines */
...@@ -4153,7 +4157,9 @@ void i915_gem_object_init(struct drm_i915_gem_object *obj, ...@@ -4153,7 +4157,9 @@ void i915_gem_object_init(struct drm_i915_gem_object *obj,
{ {
mutex_init(&obj->mm.lock); mutex_init(&obj->mm.lock);
INIT_LIST_HEAD(&obj->vma_list); spin_lock_init(&obj->vma.lock);
INIT_LIST_HEAD(&obj->vma.list);
INIT_LIST_HEAD(&obj->lut_list); INIT_LIST_HEAD(&obj->lut_list);
INIT_LIST_HEAD(&obj->batch_pool_link); INIT_LIST_HEAD(&obj->batch_pool_link);
...@@ -4319,14 +4325,13 @@ static void __i915_gem_free_objects(struct drm_i915_private *i915, ...@@ -4319,14 +4325,13 @@ static void __i915_gem_free_objects(struct drm_i915_private *i915,
mutex_lock(&i915->drm.struct_mutex); mutex_lock(&i915->drm.struct_mutex);
GEM_BUG_ON(i915_gem_object_is_active(obj)); GEM_BUG_ON(i915_gem_object_is_active(obj));
list_for_each_entry_safe(vma, vn, list_for_each_entry_safe(vma, vn, &obj->vma.list, obj_link) {
&obj->vma_list, obj_link) {
GEM_BUG_ON(i915_vma_is_active(vma)); GEM_BUG_ON(i915_vma_is_active(vma));
vma->flags &= ~I915_VMA_PIN_MASK; vma->flags &= ~I915_VMA_PIN_MASK;
i915_vma_destroy(vma); i915_vma_destroy(vma);
} }
GEM_BUG_ON(!list_empty(&obj->vma_list)); GEM_BUG_ON(!list_empty(&obj->vma.list));
GEM_BUG_ON(!RB_EMPTY_ROOT(&obj->vma_tree)); GEM_BUG_ON(!RB_EMPTY_ROOT(&obj->vma.tree));
/* This serializes freeing with the shrinker. Since the free /* This serializes freeing with the shrinker. Since the free
* is delayed, first by RCU then by the workqueue, we want the * is delayed, first by RCU then by the workqueue, we want the
......
...@@ -87,24 +87,33 @@ struct drm_i915_gem_object { ...@@ -87,24 +87,33 @@ struct drm_i915_gem_object {
const struct drm_i915_gem_object_ops *ops; const struct drm_i915_gem_object_ops *ops;
/** struct {
* @vma_list: List of VMAs backed by this object /**
* * @vma.lock: protect the list/tree of vmas
* The VMA on this list are ordered by type, all GGTT vma are placed */
* at the head and all ppGTT vma are placed at the tail. The different spinlock_t lock;
* types of GGTT vma are unordered between themselves, use the
* @vma_tree (which has a defined order between all VMA) to find an /**
* exact match. * @vma.list: List of VMAs backed by this object
*/ *
struct list_head vma_list; * The VMA on this list are ordered by type, all GGTT vma are
/** * placed at the head and all ppGTT vma are placed at the tail.
* @vma_tree: Ordered tree of VMAs backed by this object * The different types of GGTT vma are unordered between
* * themselves, use the @vma.tree (which has a defined order
* All VMA created for this object are placed in the @vma_tree for * between all VMA) to quickly find an exact match.
* fast retrieval via a binary search in i915_vma_instance(). */
* They are also added to @vma_list for easy iteration. struct list_head list;
*/
struct rb_root vma_tree; /**
* @vma.tree: Ordered tree of VMAs backed by this object
*
* All VMA created for this object are placed in the @vma.tree
* for fast retrieval via a binary search in
* i915_vma_instance(). They are also added to @vma.list for
* easy iteration.
*/
struct rb_root tree;
} vma;
/** /**
* @lut_list: List of vma lookup entries in use for this object. * @lut_list: List of vma lookup entries in use for this object.
......
...@@ -187,32 +187,52 @@ vma_create(struct drm_i915_gem_object *obj, ...@@ -187,32 +187,52 @@ vma_create(struct drm_i915_gem_object *obj,
i915_gem_object_get_stride(obj)); i915_gem_object_get_stride(obj));
GEM_BUG_ON(!is_power_of_2(vma->fence_alignment)); GEM_BUG_ON(!is_power_of_2(vma->fence_alignment));
/*
* We put the GGTT vma at the start of the vma-list, followed
* by the ppGGTT vma. This allows us to break early when
* iterating over only the GGTT vma for an object, see
* for_each_ggtt_vma()
*/
vma->flags |= I915_VMA_GGTT; vma->flags |= I915_VMA_GGTT;
list_add(&vma->obj_link, &obj->vma_list);
} else {
list_add_tail(&vma->obj_link, &obj->vma_list);
} }
spin_lock(&obj->vma.lock);
rb = NULL; rb = NULL;
p = &obj->vma_tree.rb_node; p = &obj->vma.tree.rb_node;
while (*p) { while (*p) {
struct i915_vma *pos; struct i915_vma *pos;
long cmp;
rb = *p; rb = *p;
pos = rb_entry(rb, struct i915_vma, obj_node); pos = rb_entry(rb, struct i915_vma, obj_node);
if (i915_vma_compare(pos, vm, view) < 0)
/*
* If the view already exists in the tree, another thread
* already created a matching vma, so return the older instance
* and dispose of ours.
*/
cmp = i915_vma_compare(pos, vm, view);
if (cmp == 0) {
spin_unlock(&obj->vma.lock);
kmem_cache_free(vm->i915->vmas, vma);
return pos;
}
if (cmp < 0)
p = &rb->rb_right; p = &rb->rb_right;
else else
p = &rb->rb_left; p = &rb->rb_left;
} }
rb_link_node(&vma->obj_node, rb, p); rb_link_node(&vma->obj_node, rb, p);
rb_insert_color(&vma->obj_node, &obj->vma_tree); rb_insert_color(&vma->obj_node, &obj->vma.tree);
if (i915_vma_is_ggtt(vma))
/*
* We put the GGTT vma at the start of the vma-list, followed
* by the ppGGTT vma. This allows us to break early when
* iterating over only the GGTT vma for an object, see
* for_each_ggtt_vma()
*/
list_add(&vma->obj_link, &obj->vma.list);
else
list_add_tail(&vma->obj_link, &obj->vma.list);
spin_unlock(&obj->vma.lock);
mutex_lock(&vm->mutex); mutex_lock(&vm->mutex);
list_add(&vma->vm_link, &vm->unbound_list); list_add(&vma->vm_link, &vm->unbound_list);
...@@ -232,7 +252,7 @@ vma_lookup(struct drm_i915_gem_object *obj, ...@@ -232,7 +252,7 @@ vma_lookup(struct drm_i915_gem_object *obj,
{ {
struct rb_node *rb; struct rb_node *rb;
rb = obj->vma_tree.rb_node; rb = obj->vma.tree.rb_node;
while (rb) { while (rb) {
struct i915_vma *vma = rb_entry(rb, struct i915_vma, obj_node); struct i915_vma *vma = rb_entry(rb, struct i915_vma, obj_node);
long cmp; long cmp;
...@@ -272,16 +292,18 @@ i915_vma_instance(struct drm_i915_gem_object *obj, ...@@ -272,16 +292,18 @@ i915_vma_instance(struct drm_i915_gem_object *obj,
{ {
struct i915_vma *vma; struct i915_vma *vma;
lockdep_assert_held(&obj->base.dev->struct_mutex);
GEM_BUG_ON(view && !i915_is_ggtt(vm)); GEM_BUG_ON(view && !i915_is_ggtt(vm));
GEM_BUG_ON(vm->closed); GEM_BUG_ON(vm->closed);
spin_lock(&obj->vma.lock);
vma = vma_lookup(obj, vm, view); vma = vma_lookup(obj, vm, view);
if (!vma) spin_unlock(&obj->vma.lock);
/* vma_create() will resolve the race if another creates the vma */
if (unlikely(!vma))
vma = vma_create(obj, vm, view); vma = vma_create(obj, vm, view);
GEM_BUG_ON(!IS_ERR(vma) && i915_vma_compare(vma, vm, view)); GEM_BUG_ON(!IS_ERR(vma) && i915_vma_compare(vma, vm, view));
GEM_BUG_ON(!IS_ERR(vma) && vma_lookup(obj, vm, view) != vma);
return vma; return vma;
} }
...@@ -808,14 +830,18 @@ static void __i915_vma_destroy(struct i915_vma *vma) ...@@ -808,14 +830,18 @@ static void __i915_vma_destroy(struct i915_vma *vma)
GEM_BUG_ON(i915_gem_active_isset(&vma->last_fence)); GEM_BUG_ON(i915_gem_active_isset(&vma->last_fence));
list_del(&vma->obj_link);
mutex_lock(&vma->vm->mutex); mutex_lock(&vma->vm->mutex);
list_del(&vma->vm_link); list_del(&vma->vm_link);
mutex_unlock(&vma->vm->mutex); mutex_unlock(&vma->vm->mutex);
if (vma->obj) if (vma->obj) {
rb_erase(&vma->obj_node, &vma->obj->vma_tree); struct drm_i915_gem_object *obj = vma->obj;
spin_lock(&obj->vma.lock);
list_del(&vma->obj_link);
rb_erase(&vma->obj_node, &vma->obj->vma.tree);
spin_unlock(&obj->vma.lock);
}
rbtree_postorder_for_each_entry_safe(iter, n, &vma->active, node) { rbtree_postorder_for_each_entry_safe(iter, n, &vma->active, node) {
GEM_BUG_ON(i915_gem_active_isset(&iter->base)); GEM_BUG_ON(i915_gem_active_isset(&iter->base));
......
...@@ -425,7 +425,7 @@ void i915_vma_parked(struct drm_i915_private *i915); ...@@ -425,7 +425,7 @@ void i915_vma_parked(struct drm_i915_private *i915);
* or the list is empty ofc. * or the list is empty ofc.
*/ */
#define for_each_ggtt_vma(V, OBJ) \ #define for_each_ggtt_vma(V, OBJ) \
list_for_each_entry(V, &(OBJ)->vma_list, obj_link) \ list_for_each_entry(V, &(OBJ)->vma.list, obj_link) \
for_each_until(!i915_vma_is_ggtt(V)) for_each_until(!i915_vma_is_ggtt(V))
#endif #endif
...@@ -672,7 +672,7 @@ static int igt_vma_partial(void *arg) ...@@ -672,7 +672,7 @@ static int igt_vma_partial(void *arg)
} }
count = 0; count = 0;
list_for_each_entry(vma, &obj->vma_list, obj_link) list_for_each_entry(vma, &obj->vma.list, obj_link)
count++; count++;
if (count != nvma) { if (count != nvma) {
pr_err("(%s) All partial vma were not recorded on the obj->vma_list: found %u, expected %u\n", pr_err("(%s) All partial vma were not recorded on the obj->vma_list: found %u, expected %u\n",
...@@ -701,7 +701,7 @@ static int igt_vma_partial(void *arg) ...@@ -701,7 +701,7 @@ static int igt_vma_partial(void *arg)
i915_vma_unpin(vma); i915_vma_unpin(vma);
count = 0; count = 0;
list_for_each_entry(vma, &obj->vma_list, obj_link) list_for_each_entry(vma, &obj->vma.list, obj_link)
count++; count++;
if (count != nvma) { if (count != nvma) {
pr_err("(%s) allocated an extra full vma!\n", p->name); pr_err("(%s) allocated an extra full vma!\n", p->name);
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment