Skip to content

Commit

Permalink
drm/i915: Make the inactive object shrinker per-device
Browse files Browse the repository at this point in the history 
Eliminate the racy device unload by embedding a shrinker into each
device. Smaller, simpler code.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
  • Loading branch information
Chris Wilson committed Oct 28, 2010
1 parent 176f28e commit 17250b7
Show file tree
Hide file tree
Showing 4 changed files with 56 additions and 141 deletions.
3 changes: 3 additions & 0 deletions drivers/gpu/drm/i915/i915_dma.c
View file
Original file line number Diff line number Diff line change
Expand Up @@ -2096,6 +2096,9 @@ int i915_driver_unload(struct drm_device *dev)
i915_mch_dev = NULL;
spin_unlock(&mchdev_lock);

if (dev_priv->mm.inactive_shrinker.shrink)
unregister_shrinker(&dev_priv->mm.inactive_shrinker);

mutex_lock(&dev->struct_mutex);
ret = i915_gpu_idle(dev);
if (ret)
Expand Down
3 changes: 0 additions & 3 deletions drivers/gpu/drm/i915/i915_drv.c
View file
Original file line number Diff line number Diff line change
Expand Up @@ -660,8 +660,6 @@ static int __init i915_init(void)

driver.num_ioctls = i915_max_ioctl;

i915_gem_shrinker_init();

/*
* If CONFIG_DRM_I915_KMS is set, default to KMS unless
* explicitly disabled with the module pararmeter.
Expand Down Expand Up @@ -693,7 +691,6 @@ static int __init i915_init(void)

static void __exit i915_exit(void)
{
i915_gem_shrinker_exit();
drm_exit(&driver);
}

Expand Down
12 changes: 1 addition & 11 deletions drivers/gpu/drm/i915/i915_drv.h
View file
Original file line number Diff line number Diff line change
Expand Up @@ -542,14 +542,7 @@ typedef struct drm_i915_private {
struct io_mapping *gtt_mapping;
int gtt_mtrr;

/**
* Membership on list of all loaded devices, used to evict
* inactive buffers under memory pressure.
*
* Modifications should only be done whilst holding the
* shrink_list_lock spinlock.
*/
struct list_head shrink_list;
struct shrinker inactive_shrinker;

/**
* List of objects currently involved in rendering.
Expand Down Expand Up @@ -1079,9 +1072,6 @@ void i915_gem_detach_phys_object(struct drm_device *dev,
void i915_gem_free_all_phys_object(struct drm_device *dev);
void i915_gem_release(struct drm_device * dev, struct drm_file *file_priv);

void i915_gem_shrinker_init(void);
void i915_gem_shrinker_exit(void);

/* i915_gem_evict.c */
int i915_gem_evict_something(struct drm_device *dev, int min_size,
unsigned alignment, bool mappable);
Expand Down
179 changes: 52 additions & 127 deletions drivers/gpu/drm/i915/i915_gem.c
View file
Original file line number Diff line number Diff line change
Expand Up @@ -65,8 +65,10 @@ i915_gem_object_get_pages(struct drm_gem_object *obj,
static void
i915_gem_object_put_pages(struct drm_gem_object *obj);

static LIST_HEAD(shrink_list);
static DEFINE_SPINLOCK(shrink_list_lock);
static int i915_gem_inactive_shrink(struct shrinker *shrinker,
int nr_to_scan,
gfp_t gfp_mask);


/* some bookkeeping */
static void i915_gem_info_add_obj(struct drm_i915_private *dev_priv,
Expand Down Expand Up @@ -4765,9 +4767,6 @@ i915_gem_load(struct drm_device *dev)
INIT_DELAYED_WORK(&dev_priv->mm.retire_work,
i915_gem_retire_work_handler);
init_completion(&dev_priv->error_completion);
spin_lock(&shrink_list_lock);
list_add(&dev_priv->mm.shrink_list, &shrink_list);
spin_unlock(&shrink_list_lock);

/* On GEN3 we really need to make sure the ARB C3 LP bit is set */
if (IS_GEN3(dev)) {
Expand Down Expand Up @@ -4810,6 +4809,10 @@ i915_gem_load(struct drm_device *dev)
}
i915_gem_detect_bit_6_swizzle(dev);
init_waitqueue_head(&dev_priv->pending_flip_queue);

dev_priv->mm.inactive_shrinker.shrink = i915_gem_inactive_shrink;
dev_priv->mm.inactive_shrinker.seeks = DEFAULT_SEEKS;
register_shrinker(&dev_priv->mm.inactive_shrinker);
}

/*
Expand Down Expand Up @@ -5022,152 +5025,74 @@ i915_gpu_is_active(struct drm_device *dev)
int lists_empty;

lists_empty = list_empty(&dev_priv->mm.flushing_list) &&
list_empty(&dev_priv->render_ring.active_list) &&
list_empty(&dev_priv->bsd_ring.active_list) &&
list_empty(&dev_priv->blt_ring.active_list);
list_empty(&dev_priv->mm.active_list);

return !lists_empty;
}

static int
i915_gem_shrink(struct shrinker *shrink, int nr_to_scan, gfp_t gfp_mask)
i915_gem_inactive_shrink(struct shrinker *shrinker,
int nr_to_scan,
gfp_t gfp_mask)
{
drm_i915_private_t *dev_priv, *next_dev;
struct drm_i915_gem_object *obj_priv, *next_obj;
int cnt = 0;
int would_deadlock = 1;
struct drm_i915_private *dev_priv =
container_of(shrinker,
struct drm_i915_private,
mm.inactive_shrinker);
struct drm_device *dev = dev_priv->dev;
struct drm_i915_gem_object *obj, *next;
int cnt;

if (!mutex_trylock(&dev->struct_mutex))
return nr_to_scan ? 0 : -1;

/* "fast-path" to count number of available objects */
if (nr_to_scan == 0) {
spin_lock(&shrink_list_lock);
list_for_each_entry(dev_priv, &shrink_list, mm.shrink_list) {
struct drm_device *dev = dev_priv->dev;

if (mutex_trylock(&dev->struct_mutex)) {
list_for_each_entry(obj_priv,
&dev_priv->mm.inactive_list,
mm_list)
cnt++;
mutex_unlock(&dev->struct_mutex);
}
}
spin_unlock(&shrink_list_lock);

return (cnt / 100) * sysctl_vfs_cache_pressure;
cnt = 0;
list_for_each_entry(obj,
&dev_priv->mm.inactive_list,
mm_list)
cnt++;
mutex_unlock(&dev->struct_mutex);
return cnt / 100 * sysctl_vfs_cache_pressure;
}

spin_lock(&shrink_list_lock);

rescan:
/* first scan for clean buffers */
list_for_each_entry_safe(dev_priv, next_dev,
&shrink_list, mm.shrink_list) {
struct drm_device *dev = dev_priv->dev;

if (! mutex_trylock(&dev->struct_mutex))
continue;

spin_unlock(&shrink_list_lock);
i915_gem_retire_requests(dev);
i915_gem_retire_requests(dev);

list_for_each_entry_safe(obj_priv, next_obj,
&dev_priv->mm.inactive_list,
mm_list) {
if (i915_gem_object_is_purgeable(obj_priv)) {
i915_gem_object_unbind(&obj_priv->base);
if (--nr_to_scan <= 0)
break;
}
list_for_each_entry_safe(obj, next,
&dev_priv->mm.inactive_list,
mm_list) {
if (i915_gem_object_is_purgeable(obj)) {
i915_gem_object_unbind(&obj->base);
if (--nr_to_scan == 0)
break;
}

spin_lock(&shrink_list_lock);
mutex_unlock(&dev->struct_mutex);

would_deadlock = 0;

if (nr_to_scan <= 0)
break;
}

/* second pass, evict/count anything still on the inactive list */
list_for_each_entry_safe(dev_priv, next_dev,
&shrink_list, mm.shrink_list) {
struct drm_device *dev = dev_priv->dev;

if (! mutex_trylock(&dev->struct_mutex))
continue;

spin_unlock(&shrink_list_lock);

list_for_each_entry_safe(obj_priv, next_obj,
&dev_priv->mm.inactive_list,
mm_list) {
if (nr_to_scan > 0) {
i915_gem_object_unbind(&obj_priv->base);
nr_to_scan--;
} else
cnt++;
}

spin_lock(&shrink_list_lock);
mutex_unlock(&dev->struct_mutex);

would_deadlock = 0;
}

if (nr_to_scan) {
int active = 0;

cnt = 0;
list_for_each_entry_safe(obj, next,
&dev_priv->mm.inactive_list,
mm_list) {
if (nr_to_scan) {
i915_gem_object_unbind(&obj->base);
nr_to_scan--;
} else
cnt++;
}

if (nr_to_scan && i915_gpu_is_active(dev)) {
/*
* We are desperate for pages, so as a last resort, wait
* for the GPU to finish and discard whatever we can.
* This has a dramatic impact to reduce the number of
* OOM-killer events whilst running the GPU aggressively.
*/
list_for_each_entry(dev_priv, &shrink_list, mm.shrink_list) {
struct drm_device *dev = dev_priv->dev;

if (!mutex_trylock(&dev->struct_mutex))
continue;

spin_unlock(&shrink_list_lock);

if (i915_gpu_is_active(dev)) {
i915_gpu_idle(dev);
active++;
}

spin_lock(&shrink_list_lock);
mutex_unlock(&dev->struct_mutex);
}

if (active)
if (i915_gpu_idle(dev) == 0)
goto rescan;
}

spin_unlock(&shrink_list_lock);

if (would_deadlock)
return -1;
else if (cnt > 0)
return (cnt / 100) * sysctl_vfs_cache_pressure;
else
return 0;
}

static struct shrinker shrinker = {
.shrink = i915_gem_shrink,
.seeks = DEFAULT_SEEKS,
};

__init void
i915_gem_shrinker_init(void)
{
register_shrinker(&shrinker);
}

__exit void
i915_gem_shrinker_exit(void)
{
unregister_shrinker(&shrinker);
mutex_unlock(&dev->struct_mutex);
return cnt / 100 * sysctl_vfs_cache_pressure;
}

0 comments on commit 17250b7

Please sign in to comment.