Donate to e Foundation | Murena handsets with /e/OS | Own a part of Murena! Learn more

Commit be6a0376 authored by Daniel Vetter's avatar Daniel Vetter
Browse files

drm/i915: Extract i915_gem_shrinker.c 



Two code changes:
- Extract i915_gem_shrinker_init.
- Inline i915_gem_object_is_purgeable since we open-code it everywhere
  else too.

This already has the benefit of pulling all the shrinker code
together, next patch adds a bit of kerneldoc.

Signed-off-by: default avatarDaniel Vetter <daniel.vetter@intel.com>
Reviewed-by: default avatarChris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: default avatarDaniel Vetter <daniel.vetter@ffwll.ch>
parent 6f4b12f8

drivers/gpu/drm/i915/Makefile

+1 −0
Original line number Diff line number Diff line
@@ -28,6 +28,7 @@ i915-y += i915_cmd_parser.o \ 
	  i915_gem_execbuffer.o \

	  i915_gem_gtt.o \

	  i915_gem.o \

	  i915_gem_shrinker.o \

	  i915_gem_stolen.o \

	  i915_gem_tiling.o \

	  i915_gem_userptr.o \

drivers/gpu/drm/i915/i915_drv.h

+11 −6
Original line number Diff line number Diff line
@@ -2600,12 +2600,6 @@ int i915_gem_get_aperture_ioctl(struct drm_device *dev, void *data, 
int i915_gem_wait_ioctl(struct drm_device *dev, void *data,

			struct drm_file *file_priv);

void i915_gem_load(struct drm_device *dev);

unsigned long i915_gem_shrink(struct drm_i915_private *dev_priv,

			      long target,

			      unsigned flags);

#define I915_SHRINK_PURGEABLE 0x1

#define I915_SHRINK_UNBOUND 0x2

#define I915_SHRINK_BOUND 0x4

void *i915_gem_object_alloc(struct drm_device *dev);

void i915_gem_object_free(struct drm_i915_gem_object *obj);

void i915_gem_object_init(struct drm_i915_gem_object *obj,
@@ -2953,6 +2947,17 @@ i915_gem_object_create_stolen_for_preallocated(struct drm_device *dev, 
					       u32 gtt_offset,

					       u32 size);



/* i915_gem_shrinker.c */

unsigned long i915_gem_shrink(struct drm_i915_private *dev_priv,

			      long target,

			      unsigned flags);

#define I915_SHRINK_PURGEABLE 0x1

#define I915_SHRINK_UNBOUND 0x2

#define I915_SHRINK_BOUND 0x4

unsigned long i915_gem_shrink_all(struct drm_i915_private *dev_priv);

void i915_gem_shrinker_init(struct drm_i915_private *dev_priv);





/* i915_gem_tiling.c */

static inline bool i915_gem_object_needs_bit17_swizzle(struct drm_i915_gem_object *obj)

{

drivers/gpu/drm/i915/i915_gem.c

+3 −271
Original line number Diff line number Diff line 
/*

 * Copyright © 2008 Intel Corporation

 * Copyright © 2008-2015 Intel Corporation

 *

 * Permission is hereby granted, free of charge, to any person obtaining a

 * copy of this software and associated documentation files (the "Software"),
@@ -32,7 +32,6 @@ 
#include "i915_vgpu.h"

#include "i915_trace.h"

#include "intel_drv.h"

#include <linux/oom.h>

#include <linux/shmem_fs.h>

#include <linux/slab.h>

#include <linux/swap.h>
@@ -53,15 +52,6 @@ static void i915_gem_object_update_fence(struct drm_i915_gem_object *obj, 
					 struct drm_i915_fence_reg *fence,

					 bool enable);



static unsigned long i915_gem_shrinker_count(struct shrinker *shrinker,

					     struct shrink_control *sc);

static unsigned long i915_gem_shrinker_scan(struct shrinker *shrinker,

					    struct shrink_control *sc);

static int i915_gem_shrinker_oom(struct notifier_block *nb,

				 unsigned long event,

				 void *ptr);

static unsigned long i915_gem_shrink_all(struct drm_i915_private *dev_priv);



static bool cpu_cache_is_coherent(struct drm_device *dev,

				  enum i915_cache_level level)

{
@@ -1936,12 +1926,6 @@ i915_gem_mmap_gtt_ioctl(struct drm_device *dev, void *data, 
	return i915_gem_mmap_gtt(file, dev, args->handle, &args->offset);

}



static inline int

i915_gem_object_is_purgeable(struct drm_i915_gem_object *obj)

{

	return obj->madv == I915_MADV_DONTNEED;

}



/* Immediately discard the backing storage */

static void

i915_gem_object_truncate(struct drm_i915_gem_object *obj)
@@ -2047,85 +2031,6 @@ i915_gem_object_put_pages(struct drm_i915_gem_object *obj) 
	return 0;

}



unsigned long

i915_gem_shrink(struct drm_i915_private *dev_priv,

		long target, unsigned flags)

{

	const struct {

		struct list_head *list;

		unsigned int bit;

	} phases[] = {

		{ &dev_priv->mm.unbound_list, I915_SHRINK_UNBOUND },

		{ &dev_priv->mm.bound_list, I915_SHRINK_BOUND },

		{ NULL, 0 },

	}, *phase;

	unsigned long count = 0;



	/*

	 * As we may completely rewrite the (un)bound list whilst unbinding

	 * (due to retiring requests) we have to strictly process only

	 * one element of the list at the time, and recheck the list

	 * on every iteration.

	 *

	 * In particular, we must hold a reference whilst removing the

	 * object as we may end up waiting for and/or retiring the objects.

	 * This might release the final reference (held by the active list)

	 * and result in the object being freed from under us. This is

	 * similar to the precautions the eviction code must take whilst

	 * removing objects.

	 *

	 * Also note that although these lists do not hold a reference to

	 * the object we can safely grab one here: The final object

	 * unreferencing and the bound_list are both protected by the

	 * dev->struct_mutex and so we won't ever be able to observe an

	 * object on the bound_list with a reference count equals 0.

	 */

	for (phase = phases; phase->list; phase++) {

		struct list_head still_in_list;



		if ((flags & phase->bit) == 0)

			continue;



		INIT_LIST_HEAD(&still_in_list);

		while (count < target && !list_empty(phase->list)) {

			struct drm_i915_gem_object *obj;

			struct i915_vma *vma, *v;



			obj = list_first_entry(phase->list,

					       typeof(*obj), global_list);

			list_move_tail(&obj->global_list, &still_in_list);



			if (flags & I915_SHRINK_PURGEABLE &&

			    !i915_gem_object_is_purgeable(obj))

				continue;



			drm_gem_object_reference(&obj->base);



			/* For the unbound phase, this should be a no-op! */

			list_for_each_entry_safe(vma, v,

						 &obj->vma_list, vma_link)

				if (i915_vma_unbind(vma))

					break;



			if (i915_gem_object_put_pages(obj) == 0)

				count += obj->base.size >> PAGE_SHIFT;



			drm_gem_object_unreference(&obj->base);

		}

		list_splice(&still_in_list, phase->list);

	}



	return count;

}



static unsigned long

i915_gem_shrink_all(struct drm_i915_private *dev_priv)

{

	i915_gem_evict_everything(dev_priv->dev);

	return i915_gem_shrink(dev_priv, LONG_MAX,

			       I915_SHRINK_BOUND | I915_SHRINK_UNBOUND);

}



static int

i915_gem_object_get_pages_gtt(struct drm_i915_gem_object *obj)

{
@@ -4425,7 +4330,7 @@ i915_gem_madvise_ioctl(struct drm_device *dev, void *data, 
		obj->madv = args->madv;



	/* if the object is no longer attached, discard its backing storage */

	if (i915_gem_object_is_purgeable(obj) && obj->pages == NULL)

	if (obj->madv == I915_MADV_DONTNEED && obj->pages == NULL)

		i915_gem_object_truncate(obj);



	args->retained = obj->madv != __I915_MADV_PURGED;
@@ -5064,13 +4969,7 @@ i915_gem_load(struct drm_device *dev) 


	dev_priv->mm.interruptible = true;



	dev_priv->mm.shrinker.scan_objects = i915_gem_shrinker_scan;

	dev_priv->mm.shrinker.count_objects = i915_gem_shrinker_count;

	dev_priv->mm.shrinker.seeks = DEFAULT_SEEKS;

	register_shrinker(&dev_priv->mm.shrinker);



	dev_priv->mm.oom_notifier.notifier_call = i915_gem_shrinker_oom;

	register_oom_notifier(&dev_priv->mm.oom_notifier);

	i915_gem_shrinker_init(dev_priv);



	i915_gem_batch_pool_init(dev, &dev_priv->mm.batch_pool);
@@ -5162,77 +5061,6 @@ void i915_gem_track_fb(struct drm_i915_gem_object *old, 
	}

}



static bool mutex_is_locked_by(struct mutex *mutex, struct task_struct *task)

{

	if (!mutex_is_locked(mutex))

		return false;



#if defined(CONFIG_SMP) || defined(CONFIG_DEBUG_MUTEXES)

	return mutex->owner == task;

#else

	/* Since UP may be pre-empted, we cannot assume that we own the lock */

	return false;

#endif

}



static bool i915_gem_shrinker_lock(struct drm_device *dev, bool *unlock)

{

	if (!mutex_trylock(&dev->struct_mutex)) {

		if (!mutex_is_locked_by(&dev->struct_mutex, current))

			return false;



		if (to_i915(dev)->mm.shrinker_no_lock_stealing)

			return false;



		*unlock = false;

	} else

		*unlock = true;



	return true;

}



static int num_vma_bound(struct drm_i915_gem_object *obj)

{

	struct i915_vma *vma;

	int count = 0;



	list_for_each_entry(vma, &obj->vma_list, vma_link)

		if (drm_mm_node_allocated(&vma->node))

			count++;



	return count;

}



static unsigned long

i915_gem_shrinker_count(struct shrinker *shrinker, struct shrink_control *sc)

{

	struct drm_i915_private *dev_priv =

		container_of(shrinker, struct drm_i915_private, mm.shrinker);

	struct drm_device *dev = dev_priv->dev;

	struct drm_i915_gem_object *obj;

	unsigned long count;

	bool unlock;



	if (!i915_gem_shrinker_lock(dev, &unlock))

		return 0;



	count = 0;

	list_for_each_entry(obj, &dev_priv->mm.unbound_list, global_list)

		if (obj->pages_pin_count == 0)

			count += obj->base.size >> PAGE_SHIFT;



	list_for_each_entry(obj, &dev_priv->mm.bound_list, global_list) {

		if (!i915_gem_obj_is_pinned(obj) &&

		    obj->pages_pin_count == num_vma_bound(obj))

			count += obj->base.size >> PAGE_SHIFT;

	}



	if (unlock)

		mutex_unlock(&dev->struct_mutex);



	return count;

}



/* All the new VM stuff */

unsigned long

i915_gem_obj_offset(struct drm_i915_gem_object *o,
@@ -5333,102 +5161,6 @@ unsigned long i915_gem_obj_size(struct drm_i915_gem_object *o, 
	return 0;

}



static unsigned long

i915_gem_shrinker_scan(struct shrinker *shrinker, struct shrink_control *sc)

{

	struct drm_i915_private *dev_priv =

		container_of(shrinker, struct drm_i915_private, mm.shrinker);

	struct drm_device *dev = dev_priv->dev;

	unsigned long freed;

	bool unlock;



	if (!i915_gem_shrinker_lock(dev, &unlock))

		return SHRINK_STOP;



	freed = i915_gem_shrink(dev_priv,

				sc->nr_to_scan,

				I915_SHRINK_BOUND |

				I915_SHRINK_UNBOUND |

				I915_SHRINK_PURGEABLE);

	if (freed < sc->nr_to_scan)

		freed += i915_gem_shrink(dev_priv,

					 sc->nr_to_scan - freed,

					 I915_SHRINK_BOUND |

					 I915_SHRINK_UNBOUND);

	if (unlock)

		mutex_unlock(&dev->struct_mutex);



	return freed;

}



static int

i915_gem_shrinker_oom(struct notifier_block *nb, unsigned long event, void *ptr)

{

	struct drm_i915_private *dev_priv =

		container_of(nb, struct drm_i915_private, mm.oom_notifier);

	struct drm_device *dev = dev_priv->dev;

	struct drm_i915_gem_object *obj;

	unsigned long timeout = msecs_to_jiffies(5000) + 1;

	unsigned long pinned, bound, unbound, freed_pages;

	bool was_interruptible;

	bool unlock;



	while (!i915_gem_shrinker_lock(dev, &unlock) && --timeout) {

		schedule_timeout_killable(1);

		if (fatal_signal_pending(current))

			return NOTIFY_DONE;

	}

	if (timeout == 0) {

		pr_err("Unable to purge GPU memory due lock contention.\n");

		return NOTIFY_DONE;

	}



	was_interruptible = dev_priv->mm.interruptible;

	dev_priv->mm.interruptible = false;



	freed_pages = i915_gem_shrink_all(dev_priv);



	dev_priv->mm.interruptible = was_interruptible;



	/* Because we may be allocating inside our own driver, we cannot

	 * assert that there are no objects with pinned pages that are not

	 * being pointed to by hardware.

	 */

	unbound = bound = pinned = 0;

	list_for_each_entry(obj, &dev_priv->mm.unbound_list, global_list) {

		if (!obj->base.filp) /* not backed by a freeable object */

			continue;



		if (obj->pages_pin_count)

			pinned += obj->base.size;

		else

			unbound += obj->base.size;

	}

	list_for_each_entry(obj, &dev_priv->mm.bound_list, global_list) {

		if (!obj->base.filp)

			continue;



		if (obj->pages_pin_count)

			pinned += obj->base.size;

		else

			bound += obj->base.size;

	}



	if (unlock)

		mutex_unlock(&dev->struct_mutex);



	if (freed_pages || unbound || bound)

		pr_info("Purging GPU memory, %lu bytes freed, %lu bytes still pinned.\n",

			freed_pages << PAGE_SHIFT, pinned);

	if (unbound || bound)

		pr_err("%lu and %lu bytes still available in the "

		       "bound and unbound GPU page lists.\n",

		       bound, unbound);



	*(unsigned long *)ptr += freed_pages;

	return NOTIFY_DONE;

}



bool i915_gem_obj_is_pinned(struct drm_i915_gem_object *obj)

{

	struct i915_vma *vma;

drivers/gpu/drm/i915/i915_gem_shrinker.c

0 → 100644
+291 −0
Original line number Diff line number Diff line 
/*

 * Copyright © 2008-2015 Intel Corporation

 *

 * Permission is hereby granted, free of charge, to any person obtaining a

 * copy of this software and associated documentation files (the "Software"),

 * to deal in the Software without restriction, including without limitation

 * the rights to use, copy, modify, merge, publish, distribute, sublicense,

 * and/or sell copies of the Software, and to permit persons to whom the

 * Software is furnished to do so, subject to the following conditions:

 *

 * The above copyright notice and this permission notice (including the next

 * paragraph) shall be included in all copies or substantial portions of the

 * Software.

 *

 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR

 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,

 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL

 * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER

 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING

 * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS

 * IN THE SOFTWARE.

 *

 */



#include <linux/oom.h>

#include <linux/shmem_fs.h>

#include <linux/slab.h>

#include <linux/swap.h>

#include <linux/pci.h>

#include <linux/dma-buf.h>

#include <drm/drmP.h>

#include <drm/i915_drm.h>



#include "i915_drv.h"

#include "i915_trace.h"



static bool mutex_is_locked_by(struct mutex *mutex, struct task_struct *task)

{

	if (!mutex_is_locked(mutex))

		return false;



#if defined(CONFIG_SMP) || defined(CONFIG_DEBUG_MUTEXES)

	return mutex->owner == task;

#else

	/* Since UP may be pre-empted, we cannot assume that we own the lock */

	return false;

#endif

}



unsigned long

i915_gem_shrink(struct drm_i915_private *dev_priv,

		long target, unsigned flags)

{

	const struct {

		struct list_head *list;

		unsigned int bit;

	} phases[] = {

		{ &dev_priv->mm.unbound_list, I915_SHRINK_UNBOUND },

		{ &dev_priv->mm.bound_list, I915_SHRINK_BOUND },

		{ NULL, 0 },

	}, *phase;

	unsigned long count = 0;



	/*

	 * As we may completely rewrite the (un)bound list whilst unbinding

	 * (due to retiring requests) we have to strictly process only

	 * one element of the list at the time, and recheck the list

	 * on every iteration.

	 *

	 * In particular, we must hold a reference whilst removing the

	 * object as we may end up waiting for and/or retiring the objects.

	 * This might release the final reference (held by the active list)

	 * and result in the object being freed from under us. This is

	 * similar to the precautions the eviction code must take whilst

	 * removing objects.

	 *

	 * Also note that although these lists do not hold a reference to

	 * the object we can safely grab one here: The final object

	 * unreferencing and the bound_list are both protected by the

	 * dev->struct_mutex and so we won't ever be able to observe an

	 * object on the bound_list with a reference count equals 0.

	 */

	for (phase = phases; phase->list; phase++) {

		struct list_head still_in_list;



		if ((flags & phase->bit) == 0)

			continue;



		INIT_LIST_HEAD(&still_in_list);

		while (count < target && !list_empty(phase->list)) {

			struct drm_i915_gem_object *obj;

			struct i915_vma *vma, *v;



			obj = list_first_entry(phase->list,

					       typeof(*obj), global_list);

			list_move_tail(&obj->global_list, &still_in_list);



			if (flags & I915_SHRINK_PURGEABLE &&

			    obj->madv != I915_MADV_DONTNEED)

				continue;



			drm_gem_object_reference(&obj->base);



			/* For the unbound phase, this should be a no-op! */

			list_for_each_entry_safe(vma, v,

						 &obj->vma_list, vma_link)

				if (i915_vma_unbind(vma))

					break;



			if (i915_gem_object_put_pages(obj) == 0)

				count += obj->base.size >> PAGE_SHIFT;



			drm_gem_object_unreference(&obj->base);

		}

		list_splice(&still_in_list, phase->list);

	}



	return count;

}



unsigned long i915_gem_shrink_all(struct drm_i915_private *dev_priv)

{

	i915_gem_evict_everything(dev_priv->dev);

	return i915_gem_shrink(dev_priv, LONG_MAX,

			       I915_SHRINK_BOUND | I915_SHRINK_UNBOUND);

}



static bool i915_gem_shrinker_lock(struct drm_device *dev, bool *unlock)

{

	if (!mutex_trylock(&dev->struct_mutex)) {

		if (!mutex_is_locked_by(&dev->struct_mutex, current))

			return false;



		if (to_i915(dev)->mm.shrinker_no_lock_stealing)

			return false;



		*unlock = false;

	} else

		*unlock = true;



	return true;

}



static int num_vma_bound(struct drm_i915_gem_object *obj)

{

	struct i915_vma *vma;

	int count = 0;



	list_for_each_entry(vma, &obj->vma_list, vma_link)

		if (drm_mm_node_allocated(&vma->node))

			count++;



	return count;

}



static unsigned long

i915_gem_shrinker_count(struct shrinker *shrinker, struct shrink_control *sc)

{

	struct drm_i915_private *dev_priv =

		container_of(shrinker, struct drm_i915_private, mm.shrinker);

	struct drm_device *dev = dev_priv->dev;

	struct drm_i915_gem_object *obj;

	unsigned long count;

	bool unlock;



	if (!i915_gem_shrinker_lock(dev, &unlock))

		return 0;



	count = 0;

	list_for_each_entry(obj, &dev_priv->mm.unbound_list, global_list)

		if (obj->pages_pin_count == 0)

			count += obj->base.size >> PAGE_SHIFT;



	list_for_each_entry(obj, &dev_priv->mm.bound_list, global_list) {

		if (!i915_gem_obj_is_pinned(obj) &&

		    obj->pages_pin_count == num_vma_bound(obj))

			count += obj->base.size >> PAGE_SHIFT;

	}



	if (unlock)

		mutex_unlock(&dev->struct_mutex);



	return count;

}



static unsigned long

i915_gem_shrinker_scan(struct shrinker *shrinker, struct shrink_control *sc)

{

	struct drm_i915_private *dev_priv =

		container_of(shrinker, struct drm_i915_private, mm.shrinker);

	struct drm_device *dev = dev_priv->dev;

	unsigned long freed;

	bool unlock;



	if (!i915_gem_shrinker_lock(dev, &unlock))

		return SHRINK_STOP;



	freed = i915_gem_shrink(dev_priv,

				sc->nr_to_scan,

				I915_SHRINK_BOUND |

				I915_SHRINK_UNBOUND |

				I915_SHRINK_PURGEABLE);

	if (freed < sc->nr_to_scan)

		freed += i915_gem_shrink(dev_priv,

					 sc->nr_to_scan - freed,

					 I915_SHRINK_BOUND |

					 I915_SHRINK_UNBOUND);

	if (unlock)

		mutex_unlock(&dev->struct_mutex);



	return freed;

}



static int

i915_gem_shrinker_oom(struct notifier_block *nb, unsigned long event, void *ptr)

{

	struct drm_i915_private *dev_priv =

		container_of(nb, struct drm_i915_private, mm.oom_notifier);

	struct drm_device *dev = dev_priv->dev;

	struct drm_i915_gem_object *obj;

	unsigned long timeout = msecs_to_jiffies(5000) + 1;

	unsigned long pinned, bound, unbound, freed_pages;

	bool was_interruptible;

	bool unlock;



	while (!i915_gem_shrinker_lock(dev, &unlock) && --timeout) {

		schedule_timeout_killable(1);

		if (fatal_signal_pending(current))

			return NOTIFY_DONE;

	}

	if (timeout == 0) {

		pr_err("Unable to purge GPU memory due lock contention.\n");

		return NOTIFY_DONE;

	}



	was_interruptible = dev_priv->mm.interruptible;

	dev_priv->mm.interruptible = false;



	freed_pages = i915_gem_shrink_all(dev_priv);



	dev_priv->mm.interruptible = was_interruptible;



	/* Because we may be allocating inside our own driver, we cannot

	 * assert that there are no objects with pinned pages that are not

	 * being pointed to by hardware.

	 */

	unbound = bound = pinned = 0;

	list_for_each_entry(obj, &dev_priv->mm.unbound_list, global_list) {

		if (!obj->base.filp) /* not backed by a freeable object */

			continue;



		if (obj->pages_pin_count)

			pinned += obj->base.size;

		else

			unbound += obj->base.size;

	}

	list_for_each_entry(obj, &dev_priv->mm.bound_list, global_list) {

		if (!obj->base.filp)

			continue;



		if (obj->pages_pin_count)

			pinned += obj->base.size;

		else

			bound += obj->base.size;

	}



	if (unlock)

		mutex_unlock(&dev->struct_mutex);



	if (freed_pages || unbound || bound)

		pr_info("Purging GPU memory, %lu bytes freed, %lu bytes still pinned.\n",

			freed_pages << PAGE_SHIFT, pinned);

	if (unbound || bound)

		pr_err("%lu and %lu bytes still available in the "

		       "bound and unbound GPU page lists.\n",

		       bound, unbound);



	*(unsigned long *)ptr += freed_pages;

	return NOTIFY_DONE;

}



void i915_gem_shrinker_init(struct drm_i915_private *dev_priv)

{

	dev_priv->mm.shrinker.scan_objects = i915_gem_shrinker_scan;

	dev_priv->mm.shrinker.count_objects = i915_gem_shrinker_count;

	dev_priv->mm.shrinker.seeks = DEFAULT_SEEKS;

	register_shrinker(&dev_priv->mm.shrinker);



	dev_priv->mm.oom_notifier.notifier_call = i915_gem_shrinker_oom;

	register_oom_notifier(&dev_priv->mm.oom_notifier);

}