[PATCH 0/7] drm/ttm: Pool allocator

[PATCH 0/7] drm/ttm: Pool allocator

[Pauli Nieminen]

When allocating wc/uc pages cache state transition requires cache flush which
is expensive operation. To avoid cache flushes allocation of wc/uc pages should
be done in large groups when only single cache flush is required for whole group
of pages.

In some cases drivers need t oallocate and deallocate many pages in a short time
frame. In this case we can avoid cache flushes if we keep pages in the pool before
actually freeing them later.

arch/x86 was missing set_pages_array_wc and set_memory_array_wc. Patch 6 and 7 add
missing functions and hooks set_pages_array_wc to the pool allocator.


------------------------------------------------------------------------------
Download Intel® Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
--

[PATCH 1/7] drm/ttm: add pool wc/uc page allocator

[Pauli Nieminen]

On AGP system we might allocate/free routinely uncached or wc memory,
changing page from cached (wb) to uc or wc is very expensive and involves
a lot of flushing. To improve performance this allocator use a pool
of uc,wc pages.

Pools are protected with spinlocks to allow multiple threads to allocate pages
simultanously. Expensive operations are done outside of spinlock to maximize
concurrency.

Pools are linked lists of pages that were recently freed. Shrink callback
is used for removing pages from pool. Half of pages not used between two
subsequent shrink calls will be freed.

Pool fill is using separate spinlock to protect that parallel allocations
won't result exceess allocations to the pool. To prevent dead lock chances
fill_lock is locked with spin_trylock if it is not locked.

Based on Jerome Glisse's and Dave Airlie's pool allocator.

Signed-off-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Pauli Nieminen <suokkos@gmail.com>
---
 drivers/gpu/drm/ttm/Makefile         |    2 +-
 drivers/gpu/drm/ttm/ttm_memory.c     |    7 +-
 drivers/gpu/drm/ttm/ttm_page_alloc.c |  775 ++++++++++++++++++++++++++++++++++
 drivers/gpu/drm/ttm/ttm_tt.c         |   44 +-
 include/drm/ttm/ttm_page_alloc.h     |   64 +++
 5 files changed, 867 insertions(+), 25 deletions(-)
 create mode 100644 drivers/gpu/drm/ttm/ttm_page_alloc.c
 create mode 100644 include/drm/ttm/ttm_page_alloc.h

diff --git a/drivers/gpu/drm/ttm/Makefile b/drivers/gpu/drm/ttm/Makefile
index 1e138f5..4256e20 100644
--- a/drivers/gpu/drm/ttm/Makefile
+++ b/drivers/gpu/drm/ttm/Makefile
@@ -4,6 +4,6 @@
 ccflags-y := -Iinclude/drm
 ttm-y := ttm_agp_backend.o ttm_memory.o ttm_tt.o ttm_bo.o \
 	ttm_bo_util.o ttm_bo_vm.o ttm_module.o ttm_global.o \
-	ttm_object.o ttm_lock.o ttm_execbuf_util.o
+	ttm_object.o ttm_lock.o ttm_execbuf_util.o ttm_page_alloc.o
 
 obj-$(CONFIG_DRM_TTM) += ttm.o
diff --git a/drivers/gpu/drm/ttm/ttm_memory.c b/drivers/gpu/drm/ttm/ttm_memory.c
index eb143e0..e4c7cea 100644
--- a/drivers/gpu/drm/ttm/ttm_memory.c
+++ b/drivers/gpu/drm/ttm/ttm_memory.c
@@ -27,6 +27,7 @@
 
 #include "ttm/ttm_memory.h"
 #include "ttm/ttm_module.h"
+#include "ttm/ttm_page_alloc.h"
 #include <linux/spinlock.h>
 #include <linux/sched.h>
 #include <linux/wait.h>
@@ -394,6 +395,7 @@ int ttm_mem_global_init(struct ttm_mem_global *glob)
 		       "Zone %7s: Available graphics memory: %llu kiB.\n",
 		       zone->name, (unsigned long long) zone->max_mem >> 10);
 	}
+	ttm_page_alloc_init(glob);
 	return 0;
 out_no_zone:
 	ttm_mem_global_release(glob);
@@ -406,6 +408,9 @@ void ttm_mem_global_release(struct ttm_mem_global *glob)
 	unsigned int i;
 	struct ttm_mem_zone *zone;
 
+	/* let the page allocator first stop the shrink work. */
+	ttm_page_alloc_fini();
+
 	flush_workqueue(glob->swap_queue);
 	destroy_workqueue(glob->swap_queue);
 	glob->swap_queue = NULL;
@@ -413,7 +418,7 @@ void ttm_mem_global_release(struct ttm_mem_global *glob)
 		zone = glob->zones[i];
 		kobject_del(&zone->kobj);
 		kobject_put(&zone->kobj);
-	}
+			}
 	kobject_del(&glob->kobj);
 	kobject_put(&glob->kobj);
 }
diff --git a/drivers/gpu/drm/ttm/ttm_page_alloc.c b/drivers/gpu/drm/ttm/ttm_page_alloc.c
new file mode 100644
index 0000000..768d479
--- /dev/null
+++ b/drivers/gpu/drm/ttm/ttm_page_alloc.c
@@ -0,0 +1,775 @@
+/*
+ * Copyright (c) Red Hat Inc.
+
+ * 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, sub license,
+ * 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 NON-INFRINGEMENT. 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.
+ *
+ * Authors: Dave Airlie <airlied@redhat.com>
+ *          Jerome Glisse <jglisse@redhat.com>
+ *          Pauli Nieminen <suokkos@gmail.com>
+ */
+
+/* simple list based uncached page pool
+ * - Pool collects resently freed pages for reuse
+ * - Use page->lru to keep a free list
+ * - doesn't track currently in use pages
+ */
+#include <linux/list.h>
+#include <linux/spinlock.h>
+#include <linux/highmem.h>
+#include <linux/mm_types.h>
+#include <linux/jiffies.h>
+#include <linux/timer.h>
+#include <linux/workqueue.h>
+
+#include <asm/atomic.h>
+#include <asm/agp.h>
+
+#include "ttm/ttm_bo_driver.h"
+#include "ttm/ttm_page_alloc.h"
+
+
+#define NUM_PAGES_TO_ALLOC		256
+#define SMALL_ALLOCATION		64
+#define FREE_ALL_PAGES			1
+/* times are in msecs */
+#define PAGE_FREE_INTERVAL		1000
+
+/**
+ * struct ttm_page_pool - Pool to reuse recently allocated uc/wc pages.
+ *
+ * @lock: Protects the shared pool from concurrnet access. Must be used with
+ * irqsave/irqrestore variants because pool allocator maybe called from
+ * delayed work.
+ * @fill_lock: Prevent concurrent calls to fill.
+ * @list: Pool of free uc/wc pages for fast reuse
+ * @gfp_flags: Flags to pass for alloc_page.
+ * @npages: Number of pages in pool
+ * @nlowpages: Minimum nubmer of pages in pool since previous shrink
+ * @alloc_size: Allocation sizes of this pool.
+ * operation.
+ */
+struct ttm_page_pool {
+	spinlock_t		lock;
+	bool			fill_lock;
+	struct list_head	list;
+	int			gfp_flags;
+	unsigned		npages;
+	unsigned		nlowpages;
+	unsigned		alloc_size;
+};
+
+#define NUM_POOLS 4
+
+/**
+ * struct ttm_pool_manager - Holds memory pools for fst allocation
+ *
+ * Manager is read only object for pool code so it doesn't need locking.
+ *
+ * @free_interval: minimum number of jiffies between freeing pages from pool.
+ * @glob: Global memory object for shrinker registeration.
+ * @page_alloc_inited: reference counting for pool allocation.
+ * @work: Work that is used to shrink the pool. Work is only run when there is
+ * some pages to free.
+ * @small_allocation: Limit in number of pages what is small allocation.
+ *
+ * @pools: All pool objects in use.
+ **/
+struct ttm_pool_manager {
+	unsigned long		free_interval;
+	struct ttm_mem_global	*glob;
+	atomic_t		page_alloc_inited;
+	struct delayed_work	work;
+	unsigned		small_allocation;
+
+	union {
+		struct ttm_page_pool	pools[NUM_POOLS];
+		struct {
+			struct ttm_page_pool	wc_pool;
+			struct ttm_page_pool	uc_pool;
+			struct ttm_page_pool	wc_pool_dma32;
+			struct ttm_page_pool	uc_pool_dma32;
+		} ;
+	};
+};
+
+static struct ttm_pool_manager _manager = {
+	.page_alloc_inited	= ATOMIC_INIT(0)
+};
+
+#ifdef CONFIG_X86
+/* TODO: add this to x86 like _uc, this version here is inefficient */
+static int set_pages_array_wc(struct page **pages, int addrinarray)
+{
+	int i;
+
+	for (i = 0; i < addrinarray; i++)
+		set_memory_wc((unsigned long)page_address(pages[i]), 1);
+	return 0;
+}
+#else
+static int set_pages_array_wb(struct page **pages, int addrinarray)
+{
+#ifdef TTM_HAS_AGP
+	int i;
+
+	for (i = 0; i < addrinarray; i++)
+		unmap_page_from_agp(pages[i]);
+#endif
+	return 0;
+}
+
+static int set_pages_array_wc(struct page **pages, int addrinarray)
+{
+#ifdef TTM_HAS_AGP
+	int i;
+
+	for (i = 0; i < addrinarray; i++)
+		map_page_into_agp(pages[i]);
+#endif
+	return 0;
+}
+
+static int set_pages_array_uc(struct page **pages, int addrinarray)
+{
+#ifdef TTM_HAS_AGP
+	int i;
+
+	for (i = 0; i < addrinarray; i++)
+		map_page_into_agp(pages[i]);
+#endif
+	return 0;
+}
+#endif
+
+/**
+ * Select the right pool or requested caching state and ttm flags. */
+static struct ttm_page_pool *ttm_get_pool(int flags,
+		enum ttm_caching_state cstate)
+{
+	int pool_index;
+
+	if (cstate == tt_cached)
+		return NULL;
+
+	if (cstate == tt_wc)
+		pool_index = 0x0;
+	else
+		pool_index = 0x1;
+
+	if (flags & TTM_PAGE_FLAG_DMA32)
+		pool_index |= 0x2;
+
+	return &_manager.pools[pool_index];
+}
+
+/* set memory back to wb and free the pages. */
+static void ttm_pages_put(struct page *pages[], unsigned npages)
+{
+	unsigned i;
+	if (set_pages_array_wb(pages, npages))
+		printk(KERN_ERR "[ttm] Failed to set %d pages to wb!\n",
+				npages);
+	for (i = 0; i < npages; ++i)
+		__free_page(pages[i]);
+}
+
+/**
+ * reset nlowpages after all pools have been cleaned in this run.
+ **/
+static bool ttm_reset_pools(struct ttm_pool_manager *manager)
+{
+	unsigned long irq_flags;
+	bool pages_in_pool = false;
+	unsigned i;
+	for (i = 0; i < NUM_POOLS; ++i) {
+		spin_lock_irqsave(&manager->pools[i].lock, irq_flags);
+		manager->pools[i].nlowpages = manager->pools[i].npages;
+		pages_in_pool = pages_in_pool
+			|| manager->pools[i].npages > manager->pools[i].alloc_size;
+		spin_unlock_irqrestore(&manager->pools[i].lock, irq_flags);
+	}
+	return pages_in_pool;
+}
+
+/**
+ * Calculate amount of pages to free from pool in this run.
+ *
+ * Must be called with pool lock held.
+ **/
+static unsigned ttm_page_pool_get_npages_to_free_locked(struct ttm_page_pool *pool)
+{
+	unsigned r;
+	/* If less than alloc sizes was the lowest number of pages we don't
+	 * free any */
+	if (pool->nlowpages < pool->alloc_size)
+		return 0;
+	/* leave half of unused pages to pool */
+	r = (pool->nlowpages - pool->alloc_size)/2;
+	if (r)
+		return r;
+	/* make sure we remove all pages even when there is rounding down */
+	if (pool->nlowpages)
+		return 1;
+	return 0;
+}
+
+/**
+ * Update pool counters match pool state after freeing pages.
+ *
+ * Must be called with pool lock held.
+ */
+static bool ttm_page_pool_free_pages_locked(struct ttm_page_pool *pool,
+		unsigned freed_pages)
+{
+	unsigned tmp;
+	pool->npages -= freed_pages;
+	/* Calculate number of pages taken from nlowpages
+	 * npages_to_free = 1/2*nlowpages =>
+	 * nlowpages_delta = 2*freed_pages
+	 */
+	tmp = 2*freed_pages;
+	/* protect against rounding errors */
+	if (tmp < pool->nlowpages) {
+		pool->nlowpages -= tmp;
+		return true;
+	}
+
+	pool->nlowpages = 0;
+	return false;
+}
+
+/**
+ * Free pages from pool.
+ *
+ * To prevent hogging the ttm_swap process we only free NUM_PAGES_TO_ALLOC
+ * number of pages in one go.
+ *
+ * @pool: to free the pages from
+ * @free_all: If set to true will free all pages in pool
+ **/
+static bool ttm_page_pool_free(struct ttm_page_pool *pool, const int free_all)
+{
+	unsigned long irq_flags;
+	struct page *p;
+	struct page **pages_to_free;
+	unsigned freed_pages, npages_to_free;
+	bool more_work = false;
+
+	pages_to_free = kmalloc(NUM_PAGES_TO_ALLOC * sizeof(struct page *),
+			GFP_KERNEL);
+	if (!pages_to_free) {
+		printk(KERN_ERR "Failed to allocate memory for pool free operation.\n");
+		return true;
+	}
+
+restart:
+	spin_lock_irqsave(&pool->lock, irq_flags);
+
+	npages_to_free = ttm_page_pool_get_npages_to_free_locked(pool);
+
+	freed_pages = 0;
+	if (unlikely(free_all))
+		npages_to_free = pool->npages;
+
+	list_for_each_entry_reverse(p, &pool->list, lru) {
+		if (freed_pages >= npages_to_free)
+			break;
+
+		pages_to_free[freed_pages++] = p;
+		/* We can only remove NUM_PAGES_TO_ALLOC at a time. */
+		if (freed_pages >= NUM_PAGES_TO_ALLOC) {
+			/* remove range of page sfrom the pool */
+			__list_del(p->lru.prev, &pool->list);
+			/* update pool to counters match what is in pool.
+			 * return value tells us if we are finnished with this
+			 * free operation.
+			 **/
+			more_work = ttm_page_pool_free_pages_locked(pool, freed_pages);
+			/**
+			 * Because changing page caching is costly
+			 * we unlock the pool to prevent stalling.
+			 */
+			spin_unlock_irqrestore(&pool->lock, irq_flags);
+
+			ttm_pages_put(pages_to_free, freed_pages);
+
+			/* free all so restart the processing */
+			if (unlikely(free_all))
+				goto restart;
+			/* Now out of here to let others jobs run in ttm_swap */
+			goto out;
+
+		}
+	}
+
+	pool->npages -= freed_pages;
+	/* set nlowpages to zero to prevent extra freeing in thsi patch.
+	 * nlowpages is reseted later after all work has been finnished.
+	 **/
+	pool->nlowpages = 0;
+
+	/* remove range of pages from the pool */
+	if (freed_pages)
+		__list_del(&p->lru, &pool->list);
+
+	spin_unlock_irqrestore(&pool->lock, irq_flags);
+
+	if (freed_pages)
+		ttm_pages_put(pages_to_free, freed_pages);
+out:
+	kfree(pages_to_free);
+	return more_work;
+}
+
+/**
+ * Callback for workqueue.
+ *
+ * We limit the work that is done in single go to let others task run too in
+ * shared workqueue. If ttm_page_pool_free signals there is more work to do
+ * we immediately queue new work.
+ *
+ * @w: work structure that can't be used to find the related data.
+ **/
+static void ttm_pool_shrink(struct work_struct *w)
+{
+	struct delayed_work *work =
+		container_of(w, struct delayed_work, work);
+	struct ttm_pool_manager *manager =
+	    container_of(work, struct ttm_pool_manager, work);
+	unsigned i;
+	bool more_work = false;
+
+
+	for (i = 0; i < NUM_POOLS; ++i)
+		more_work = more_work || ttm_page_pool_free(&manager->pools[i], 0);
+
+	/* Queue more work to be done. This forces cleanup to use
+	 * cancel_delayed_work_sync() to break the loop. */
+	if (!more_work) {
+		/* reset the pool counters and queue more work if there is
+		 * pages to free */
+		if (ttm_reset_pools(manager))
+			(void)queue_delayed_work(manager->glob->swap_queue,
+					&manager->work,
+					round_jiffies(manager->free_interval));
+	} else /* Restart work with short delay because there is more work */
+		(void)queue_delayed_work(manager->glob->swap_queue,
+				&manager->work, HZ);
+
+}
+
+
+static int ttm_set_pages_caching(struct page **pages,
+		enum ttm_caching_state cstate, unsigned cpages)
+{
+	int r = 0;
+	/* Set page caching */
+	switch (cstate) {
+	case tt_uncached:
+		r = set_pages_array_uc(pages, cpages);
+		if (r)
+			printk(KERN_ERR "[ttm] Failed to set %d pages to uc!\n",
+					cpages);
+		break;
+	case tt_wc:
+		r = set_pages_array_wc(pages, cpages);
+		if (r)
+			printk(KERN_ERR "[ttm] Failed to set %d pages to wc!\n",
+					cpages);
+		break;
+	default:
+		break;
+	}
+	return r;
+}
+
+/**
+ * Free pages the pages that failed to change the caching state. If there is
+ * any pages that have changed their caching state already put them to the
+ * pool.
+ */
+static void ttm_handle_caching_state_failure(struct list_head *pages,
+		int ttm_flags, enum ttm_caching_state cstate,
+		struct page **failed_pages, unsigned cpages)
+{
+	unsigned i;
+	/* Failed pages has to be reed */
+	for (i = 0; i < cpages; ++i) {
+		list_del(&failed_pages[i]->lru);
+		__free_page(failed_pages[i]);
+	}
+}
+
+/**
+ * Allocate new pages with correct caching.
+ *
+ * This function is reentrant if caller updates count depending on number of
+ * pages returned in pages array.
+ */
+static int ttm_alloc_new_pages(struct list_head *pages, int gfp_flags,
+		int ttm_flags, enum ttm_caching_state cstate, unsigned count)
+{
+	struct page **caching_array;
+	struct page *p;
+	int r = 0;
+	unsigned i, cpages;
+	unsigned max_cpages = min(count,
+			(unsigned)(PAGE_SIZE/sizeof(struct page *)));
+
+	/* allocate array for page caching change */
+	caching_array = kmalloc(max_cpages*sizeof(struct page *), GFP_KERNEL);
+
+	if (!caching_array) {
+		printk(KERN_ERR "[ttm] unable to allocate table for new pages.");
+		return -ENOMEM;
+	}
+
+	for (i = 0, cpages = 0; i < count; ++i) {
+		p = alloc_page(gfp_flags);
+
+		if (!p) {
+			printk(KERN_ERR "[ttm] unable to get page %u\n", i);
+
+			/* store already allocated pages in the pool after
+			 * setting the caching state */
+			if (cpages) {
+				r = ttm_set_pages_caching(caching_array, cstate, cpages);
+				if (r)
+					ttm_handle_caching_state_failure(pages,
+						ttm_flags, cstate,
+						caching_array, cpages);
+			}
+			r = -ENOMEM;
+			goto out;
+		}
+
+#ifdef CONFIG_HIGHMEM
+		/* gfp flags of highmem page should never be dma32 so we
+		 * we should be fine in such case
+		 */
+		if (!PageHighMem(p))
+#endif
+		{
+			caching_array[cpages++] = p;
+			if (cpages == max_cpages) {
+
+				r = ttm_set_pages_caching(caching_array,
+						cstate, cpages);
+				if (r) {
+					ttm_handle_caching_state_failure(pages,
+						ttm_flags, cstate,
+						caching_array, cpages);
+					goto out;
+				}
+				cpages = 0;
+			}
+		}
+
+		list_add(&p->lru, pages);
+	}
+
+	if (cpages) {
+		r = ttm_set_pages_caching(caching_array, cstate, cpages);
+		if (r)
+			ttm_handle_caching_state_failure(pages,
+					ttm_flags, cstate,
+					caching_array, cpages);
+	}
+out:
+	kfree(caching_array);
+
+	return r;
+}
+
+/**
+ * Fill the given pool if there isn't enough pages and requested number of
+ * pages is small.
+ */
+static void ttm_page_pool_fill_locked(struct ttm_page_pool *pool,
+		int ttm_flags, enum ttm_caching_state cstate, unsigned count,
+		unsigned long irq_flags)
+{
+	struct list_head new_pages;
+	struct page *p, *tmp;
+	int r;
+	unsigned cpages = 0;
+	/**
+	 * Only allow one pool fill operation at a time.
+	 * If pool doesn't have enough pages for the allocation new pages are
+	 * allocated from outside of pool.
+	 */
+	if (pool->fill_lock)
+		return;
+
+	pool->fill_lock = true;
+
+	if (count < _manager.small_allocation
+		&& count > pool->npages) {
+		/* If allocation request is small and there is not enough
+		 * pages in pool we fill the pool first */
+		INIT_LIST_HEAD(&new_pages);
+
+		/**
+		 * Can't change page caching if in irqsave context. We have to
+		 * drop the pool->lock.
+		 */
+		spin_unlock_irqrestore(&pool->lock, irq_flags);
+		r = ttm_alloc_new_pages(&new_pages, pool->gfp_flags, ttm_flags,
+				cstate,	pool->alloc_size);
+		spin_lock_irqsave(&pool->lock, irq_flags);
+
+		if (!r) {
+			list_splice(&new_pages, &pool->list);
+			pool->npages += pool->alloc_size;
+			/* Have to remmber to update the low number of pages
+			 * too */
+			pool->nlowpages += pool->alloc_size;
+		} else {
+			printk(KERN_ERR "[ttm] Failed to fill pool (%p).", pool);
+			/* If we have any pages left put them to the pool. */
+			list_for_each_entry_safe(p, tmp, &pool->list, lru) {
+				++cpages;
+			}
+			list_splice(&new_pages, &pool->list);
+			pool->npages += cpages;
+			pool->nlowpages += cpages;
+		}
+
+	}
+	pool->fill_lock = false;
+}
+
+/**
+ * Cut count nubmer of pages from the pool and put them to return list
+ *
+ * @return count of pages still to allocate to fill the request.
+ */
+static unsigned ttm_page_pool_get_pages(struct ttm_page_pool *pool,
+		struct list_head *pages, int ttm_flags,
+		enum ttm_caching_state cstate, unsigned count)
+{
+	unsigned long irq_flags;
+	struct list_head *p;
+	unsigned i;
+
+	spin_lock_irqsave(&pool->lock, irq_flags);
+	ttm_page_pool_fill_locked(pool, ttm_flags, cstate, count, irq_flags);
+
+	if (count >= pool->npages) {
+		/* take all pages from the pool */
+		list_splice_init(&pool->list, pages);
+		count -= pool->npages;
+		pool->npages = 0;
+		pool->nlowpages = 0;
+		goto out;
+	}
+	/* find the last pages to include for requested number of pages */
+	if (count <= pool->npages/2) {
+		i = 0;
+		list_for_each(p, &pool->list) {
+			if (++i == count)
+				break;
+		}
+	} else {
+		i = pool->npages + 1;
+		list_for_each_prev(p, &pool->list) {
+			if (--i == count)
+				break;
+		}
+	}
+	/* Cut count number of pages from pool */
+	list_cut_position(pages, &pool->list, p);
+	pool->npages -= count;
+	count = 0;
+	if (pool->npages < pool->nlowpages)
+		pool->nlowpages = pool->npages;
+out:
+	spin_unlock_irqrestore(&pool->lock, irq_flags);
+	return count;
+}
+
+/*
+ * On success pages list will hold count number of correctly
+ * cached pages.
+ */
+int ttm_get_pages(struct list_head *pages, int flags,
+		enum ttm_caching_state cstate, unsigned count)
+{
+	struct ttm_page_pool *pool = ttm_get_pool(flags, cstate);
+	struct page *p = NULL;
+	int gfp_flags = 0;
+	int r;
+
+	/* set zero flag for page allocation if required */
+	if (flags & TTM_PAGE_FLAG_ZERO_ALLOC)
+		gfp_flags |= __GFP_ZERO;
+
+	/* No pool for cached pages */
+	if (pool == NULL) {
+		if (flags & TTM_PAGE_FLAG_DMA32)
+			gfp_flags |= GFP_DMA32;
+		else
+			gfp_flags |= __GFP_HIGHMEM;
+
+		for (r = 0; r < count; ++r) {
+			p = alloc_page(gfp_flags);
+			if (!p) {
+
+				printk(KERN_ERR "[ttm] unable to allocate page.");
+				return -ENOMEM;
+			}
+
+			list_add(&p->lru, pages);
+		}
+		return 0;
+	}
+
+
+	/* combine zero flag to pool flags */
+	gfp_flags |= pool->gfp_flags;
+
+	/* First we take pages from the pool */
+	count = ttm_page_pool_get_pages(pool, pages, flags, cstate, count);
+
+	/* clear the pages coming from the pool if requested */
+	if (flags & TTM_PAGE_FLAG_ZERO_ALLOC) {
+		list_for_each_entry(p, pages, lru) {
+			clear_page(page_address(p));
+		}
+	}
+
+	/* If pool didn't have enough pages allocate new one. */
+	if (count > 0) {
+		/* ttm_alloc_new_pages doesn't reference pool so we can run
+		 * multiple requests in parallel.
+		 **/
+		r = ttm_alloc_new_pages(pages, gfp_flags, flags, cstate, count);
+		if (r) {
+			/* If there is any pages in the list put them back to
+			 * the pool. */
+			printk(KERN_ERR "[ttm] Failed to allocate extra pages "
+					"for large request.");
+			ttm_put_pages(pages, flags, cstate);
+			return r;
+		}
+	}
+
+
+	return 0;
+}
+
+/* Put all pages in pages list to correct pool to wait for reuse */
+void ttm_put_pages(struct list_head *pages, int flags,
+		enum ttm_caching_state cstate)
+{
+	unsigned long irq_flags;
+	struct ttm_page_pool *pool = ttm_get_pool(flags, cstate);
+	struct page *p, *tmp;
+	unsigned page_count = 0;
+	bool work_for_shrink;
+
+	if (pool == NULL) {
+
+		/* No pool for this memory type so free the pages */
+
+		list_for_each_entry_safe(p, tmp, pages, lru) {
+			__free_page(p);
+		}
+		/* Make the pages list empty */
+		INIT_LIST_HEAD(pages);
+		return;
+	}
+
+	list_for_each_entry_safe(p, tmp, pages, lru) {
+
+#ifdef CONFIG_HIGHMEM
+		/* we don't have pool for highmem -> free them */
+		if (PageHighMem(p)) {
+			list_del(&p->lru);
+			__free_page(p);
+		} else
+#endif
+		{
+			++page_count;
+		}
+
+	}
+
+	spin_lock_irqsave(&pool->lock, irq_flags);
+	list_splice_init(pages, &pool->list);
+	pool->npages += page_count;
+	work_for_shrink = pool->npages > pool->alloc_size;
+	spin_unlock_irqrestore(&pool->lock, irq_flags);
+
+	if (work_for_shrink)
+		(void)queue_delayed_work(_manager.glob->swap_queue,
+				&_manager.work,
+				round_jiffies(_manager.free_interval));
+}
+
+static void ttm_page_pool_init_locked(struct ttm_page_pool *pool, int flags)
+{
+	spin_lock_init(&pool->lock);
+	pool->fill_lock = false;
+	INIT_LIST_HEAD(&pool->list);
+	pool->npages = pool->nlowpages = 0;
+	pool->alloc_size = NUM_PAGES_TO_ALLOC;
+	pool->gfp_flags = flags;
+}
+
+int ttm_page_alloc_init(struct ttm_mem_global *glob)
+{
+	if (atomic_add_return(1, &_manager.page_alloc_inited) > 1)
+		return 0;
+
+	printk(KERN_INFO "[ttm] Initializing pool allocator.\n");
+
+	ttm_page_pool_init_locked(&_manager.wc_pool, GFP_HIGHUSER);
+
+	ttm_page_pool_init_locked(&_manager.uc_pool, GFP_HIGHUSER);
+
+	ttm_page_pool_init_locked(&_manager.wc_pool_dma32, GFP_USER | GFP_DMA32);
+
+	ttm_page_pool_init_locked(&_manager.uc_pool_dma32, GFP_USER | GFP_DMA32);
+
+	_manager.free_interval = msecs_to_jiffies(PAGE_FREE_INTERVAL);
+	_manager.small_allocation = SMALL_ALLOCATION;
+	_manager.glob = glob;
+
+	INIT_DELAYED_WORK(&_manager.work, ttm_pool_shrink);
+
+	return 0;
+}
+
+void ttm_page_alloc_fini()
+{
+	int i;
+
+	if (atomic_sub_return(1, &_manager.page_alloc_inited) > 0)
+		return;
+
+	printk(KERN_INFO "[ttm] Finilizing pool allocator.\n");
+
+	/* stop the shrinker from running */
+	cancel_delayed_work_sync(&_manager.work);
+
+	for (i = 0; i < NUM_POOLS; ++i)
+		ttm_page_pool_free(&_manager.pools[i], FREE_ALL_PAGES);
+}
diff --git a/drivers/gpu/drm/ttm/ttm_tt.c b/drivers/gpu/drm/ttm/ttm_tt.c
index a759170..8a6fc01 100644
--- a/drivers/gpu/drm/ttm/ttm_tt.c
+++ b/drivers/gpu/drm/ttm/ttm_tt.c
@@ -38,6 +38,7 @@
 #include "ttm/ttm_module.h"
 #include "ttm/ttm_bo_driver.h"
 #include "ttm/ttm_placement.h"
+#include "ttm/ttm_page_alloc.h"
 
 static int ttm_tt_swapin(struct ttm_tt *ttm);
 
@@ -72,21 +73,6 @@ static void ttm_tt_free_page_directory(struct ttm_tt *ttm)
 	ttm->pages = NULL;
 }
 
-static struct page *ttm_tt_alloc_page(unsigned page_flags)
-{
-	gfp_t gfp_flags = GFP_USER;
-
-	if (page_flags & TTM_PAGE_FLAG_ZERO_ALLOC)
-		gfp_flags |= __GFP_ZERO;
-
-	if (page_flags & TTM_PAGE_FLAG_DMA32)
-		gfp_flags |= __GFP_DMA32;
-	else
-		gfp_flags |= __GFP_HIGHMEM;
-
-	return alloc_page(gfp_flags);
-}
-
 static void ttm_tt_free_user_pages(struct ttm_tt *ttm)
 {
 	int write;
@@ -127,15 +113,21 @@ static void ttm_tt_free_user_pages(struct ttm_tt *ttm)
 static struct page *__ttm_tt_get_page(struct ttm_tt *ttm, int index)
 {
 	struct page *p;
+	struct list_head h;
 	struct ttm_mem_global *mem_glob = ttm->glob->mem_glob;
 	int ret;
 
 	while (NULL == (p = ttm->pages[index])) {
-		p = ttm_tt_alloc_page(ttm->page_flags);
 
-		if (!p)
+		INIT_LIST_HEAD(&h);
+
+		ret = ttm_get_pages(&h, ttm->page_flags, ttm->caching_state, 1);
+
+		if (ret != 0)
 			return NULL;
 
+		p = list_first_entry(&h, struct page, lru);
+
 		ret = ttm_mem_global_alloc_page(mem_glob, p, false, false);
 		if (unlikely(ret != 0))
 			goto out_err;
@@ -244,10 +236,10 @@ static int ttm_tt_set_caching(struct ttm_tt *ttm,
 	if (ttm->caching_state == c_state)
 		return 0;
 
-	if (c_state != tt_cached) {
-		ret = ttm_tt_populate(ttm);
-		if (unlikely(ret != 0))
-			return ret;
+	if (ttm->state == tt_unpopulated) {
+		/* Change caching but don't populate */
+		ttm->caching_state = c_state;
+		return 0;
 	}
 
 	if (ttm->caching_state == tt_cached)
@@ -298,13 +290,17 @@ EXPORT_SYMBOL(ttm_tt_set_placement_caching);
 static void ttm_tt_free_alloced_pages(struct ttm_tt *ttm)
 {
 	int i;
+	unsigned count = 0;
+	struct list_head h;
 	struct page *cur_page;
 	struct ttm_backend *be = ttm->be;
 
+	INIT_LIST_HEAD(&h);
+
 	if (be)
 		be->func->clear(be);
-	(void)ttm_tt_set_caching(ttm, tt_cached);
 	for (i = 0; i < ttm->num_pages; ++i) {
+
 		cur_page = ttm->pages[i];
 		ttm->pages[i] = NULL;
 		if (cur_page) {
@@ -314,9 +310,11 @@ static void ttm_tt_free_alloced_pages(struct ttm_tt *ttm)
 				       "Leaking pages.\n");
 			ttm_mem_global_free_page(ttm->glob->mem_glob,
 						 cur_page);
-			__free_page(cur_page);
+			list_add(&cur_page->lru, &h);
+			count++;
 		}
 	}
+	ttm_put_pages(&h, ttm->page_flags, ttm->caching_state);
 	ttm->state = tt_unpopulated;
 	ttm->first_himem_page = ttm->num_pages;
 	ttm->last_lomem_page = -1;
diff --git a/include/drm/ttm/ttm_page_alloc.h b/include/drm/ttm/ttm_page_alloc.h
new file mode 100644
index 0000000..485514a
--- /dev/null
+++ b/include/drm/ttm/ttm_page_alloc.h
@@ -0,0 +1,64 @@
+/*
+ * Copyright (c) Red Hat Inc.
+
+ * 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, sub license,
+ * 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 NON-INFRINGEMENT. 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.
+ *
+ * Authors: Dave Airlie <airlied@redhat.com>
+ *          Jerome Glisse <jglisse@redhat.com>
+ */
+#ifndef TTM_PAGE_ALLOC
+#define TTM_PAGE_ALLOC
+
+#include "ttm_bo_driver.h"
+#include "ttm_memory.h"
+
+/**
+ * Get count number of pages from pool to pages list.
+ *
+ * @pages: heado of empty linked list where pages are filled.
+ * @flags: ttm flags for page allocation.
+ * @cstate: ttm caching state for the page.
+ * @count: number of pages to allocate.
+ */
+int ttm_get_pages(struct list_head *pages, int flags,
+		enum ttm_caching_state cstate, unsigned count);
+/**
+ * Put linked list of pages to pool.
+ *
+ * @pages: list of pages to free.
+ * @flags: ttm flags for page allocation.
+ * @cstate: ttm caching state.
+ */
+void ttm_put_pages(struct list_head *pages, int flags,
+		enum ttm_caching_state cstate);
+/**
+ * Initialize pool allocator.
+ *
+ * Pool allocator is internaly reference counted so it can be initialized
+ * multiple times but ttm_page_alloc_fini has to be called same number of
+ * times.
+ */
+int ttm_page_alloc_init(struct ttm_mem_global *glob);
+/**
+ * Free pool allocator.
+ */
+void ttm_page_alloc_fini(void);
+
+#endif
-- 
1.6.3.3


------------------------------------------------------------------------------
Download Intel&#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
--

[PATCH 2/7] drm/ttm: Add debugfs output entry to pool allocator.

[Pauli Nieminen]

ttm_page_alloc_debugfs can be registered to output the state
of pools.

Debugfs file will output number of pages freed from the pool,
number of pages in pool now and the lowes number of pages in
pool since previous shrink.

Signed-off-by: Pauli Nieminen <suokkos@gmail.com>
---
 drivers/gpu/drm/ttm/ttm_page_alloc.c |   46 +++++++++++++++++++++++++++++----
 include/drm/ttm/ttm_page_alloc.h     |    4 +++
 2 files changed, 44 insertions(+), 6 deletions(-)

diff --git a/drivers/gpu/drm/ttm/ttm_page_alloc.c b/drivers/gpu/drm/ttm/ttm_page_alloc.c
index 768d479..206bee9 100644
--- a/drivers/gpu/drm/ttm/ttm_page_alloc.c
+++ b/drivers/gpu/drm/ttm/ttm_page_alloc.c
@@ -34,6 +34,7 @@
 #include <linux/spinlock.h>
 #include <linux/highmem.h>
 #include <linux/mm_types.h>
+#include <linux/module.h>
 #include <linux/jiffies.h>
 #include <linux/timer.h>
 #include <linux/workqueue.h>
@@ -73,6 +74,9 @@ struct ttm_page_pool {
 	unsigned		npages;
 	unsigned		nlowpages;
 	unsigned		alloc_size;
+	char			*name;
+	unsigned long		nfrees;
+	unsigned long		nrefills;
 };
 
 #define NUM_POOLS 4
@@ -240,6 +244,7 @@ static bool ttm_page_pool_free_pages_locked(struct ttm_page_pool *pool,
 {
 	unsigned tmp;
 	pool->npages -= freed_pages;
+	pool->nfrees += freed_pages;
 	/* Calculate number of pages taken from nlowpages
 	 * npages_to_free = 1/2*nlowpages =>
 	 * nlowpages_delta = 2*freed_pages
@@ -320,6 +325,7 @@ restart:
 	}
 
 	pool->npages -= freed_pages;
+	pool->nfrees += freed_pages;
 	/* set nlowpages to zero to prevent extra freeing in thsi patch.
 	 * nlowpages is reseted later after all work has been finnished.
 	 **/
@@ -537,6 +543,7 @@ static void ttm_page_pool_fill_locked(struct ttm_page_pool *pool,
 
 		if (!r) {
 			list_splice(&new_pages, &pool->list);
+			++pool->nrefills;
 			pool->npages += pool->alloc_size;
 			/* Have to remmber to update the low number of pages
 			 * too */
@@ -724,14 +731,16 @@ void ttm_put_pages(struct list_head *pages, int flags,
 				round_jiffies(_manager.free_interval));
 }
 
-static void ttm_page_pool_init_locked(struct ttm_page_pool *pool, int flags)
+static void ttm_page_pool_init_locked(struct ttm_page_pool *pool, int flags,
+		char *name)
 {
 	spin_lock_init(&pool->lock);
 	pool->fill_lock = false;
 	INIT_LIST_HEAD(&pool->list);
-	pool->npages = pool->nlowpages = 0;
+	pool->npages = pool->nlowpages = pool->nfrees = 0;
 	pool->alloc_size = NUM_PAGES_TO_ALLOC;
 	pool->gfp_flags = flags;
+	pool->name = name;
 }
 
 int ttm_page_alloc_init(struct ttm_mem_global *glob)
@@ -741,13 +750,15 @@ int ttm_page_alloc_init(struct ttm_mem_global *glob)
 
 	printk(KERN_INFO "[ttm] Initializing pool allocator.\n");
 
-	ttm_page_pool_init_locked(&_manager.wc_pool, GFP_HIGHUSER);
+	ttm_page_pool_init_locked(&_manager.wc_pool, GFP_HIGHUSER, "wc");
 
-	ttm_page_pool_init_locked(&_manager.uc_pool, GFP_HIGHUSER);
+	ttm_page_pool_init_locked(&_manager.uc_pool, GFP_HIGHUSER, "uc");
 
-	ttm_page_pool_init_locked(&_manager.wc_pool_dma32, GFP_USER | GFP_DMA32);
+	ttm_page_pool_init_locked(&_manager.wc_pool_dma32, GFP_USER | GFP_DMA32,
+			"wc dma");
 
-	ttm_page_pool_init_locked(&_manager.uc_pool_dma32, GFP_USER | GFP_DMA32);
+	ttm_page_pool_init_locked(&_manager.uc_pool_dma32, GFP_USER | GFP_DMA32,
+			"uc dma");
 
 	_manager.free_interval = msecs_to_jiffies(PAGE_FREE_INTERVAL);
 	_manager.small_allocation = SMALL_ALLOCATION;
@@ -773,3 +784,26 @@ void ttm_page_alloc_fini()
 	for (i = 0; i < NUM_POOLS; ++i)
 		ttm_page_pool_free(&_manager.pools[i], FREE_ALL_PAGES);
 }
+
+int ttm_page_alloc_debugfs(struct seq_file *m, void *data)
+{
+	struct ttm_page_pool *p;
+	unsigned i;
+	char *h[] = {"pool", "refills", "pages freed", "size", "min size"};
+	if (atomic_read(&_manager.page_alloc_inited) == 0) {
+		seq_printf(m, "No pool allocator running.\n");
+		return 0;
+	}
+	seq_printf(m, "%6s %12s %13s %8s %8s\n",
+			h[0], h[1], h[2], h[3], h[4]);
+	for (i = 0; i < NUM_POOLS; ++i) {
+		p = &_manager.pools[i];
+
+		seq_printf(m, "%6s %12ld %13ld %8d %8d\n",
+				p->name, p->nrefills,
+				p->nfrees, p->npages,
+				p->nlowpages);
+	}
+	return 0;
+}
+EXPORT_SYMBOL(ttm_page_alloc_debugfs);
diff --git a/include/drm/ttm/ttm_page_alloc.h b/include/drm/ttm/ttm_page_alloc.h
index 485514a..2df0caa 100644
--- a/include/drm/ttm/ttm_page_alloc.h
+++ b/include/drm/ttm/ttm_page_alloc.h
@@ -61,4 +61,8 @@ int ttm_page_alloc_init(struct ttm_mem_global *glob);
  */
 void ttm_page_alloc_fini(void);
 
+/**
+ * Output the state of pools to debugfs file
+ */
+extern int ttm_page_alloc_debugfs(struct seq_file *m, void *data);
 #endif
-- 
1.6.3.3


------------------------------------------------------------------------------
Download Intel&#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
--

[PATCH 3/7] drm/radeon/kms: Add ttm page pool debugfs file.

[Pauli Nieminen]

ttm_page_pool file is hooked ttm_page_alloc_debugfs for pool
allocator state.

Signed-off-by: Pauli Nieminen <suokkos@gmail.com>
---
 drivers/gpu/drm/radeon/radeon_ttm.c |   13 ++++++++++---
 1 files changed, 10 insertions(+), 3 deletions(-)

diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c
index 43c5ab3..fc787e8 100644
--- a/drivers/gpu/drm/radeon/radeon_ttm.c
+++ b/drivers/gpu/drm/radeon/radeon_ttm.c
@@ -33,6 +33,7 @@
 #include <ttm/ttm_bo_driver.h>
 #include <ttm/ttm_placement.h>
 #include <ttm/ttm_module.h>
+#include <ttm/ttm_page_alloc.h>
 #include <drm/drmP.h>
 #include <drm/radeon_drm.h>
 #include <linux/seq_file.h>
@@ -744,8 +745,8 @@ static int radeon_mm_dump_table(struct seq_file *m, void *data)
 static int radeon_ttm_debugfs_init(struct radeon_device *rdev)
 {
 #if defined(CONFIG_DEBUG_FS)
-	static struct drm_info_list radeon_mem_types_list[RADEON_DEBUGFS_MEM_TYPES];
-	static char radeon_mem_types_names[RADEON_DEBUGFS_MEM_TYPES][32];
+	static struct drm_info_list radeon_mem_types_list[RADEON_DEBUGFS_MEM_TYPES+1];
+	static char radeon_mem_types_names[RADEON_DEBUGFS_MEM_TYPES+1][32];
 	unsigned i;
 
 	for (i = 0; i < RADEON_DEBUGFS_MEM_TYPES; i++) {
@@ -762,7 +763,13 @@ static int radeon_ttm_debugfs_init(struct radeon_device *rdev)
 			radeon_mem_types_list[i].data = &rdev->mman.bdev.man[TTM_PL_TT].manager;
 
 	}
-	return radeon_debugfs_add_files(rdev, radeon_mem_types_list, RADEON_DEBUGFS_MEM_TYPES);
+	/* Add ttm page pool to debugfs */
+	sprintf(radeon_mem_types_names[i], "ttm_page_pool");
+	radeon_mem_types_list[i].name = radeon_mem_types_names[i];
+	radeon_mem_types_list[i].show = &ttm_page_alloc_debugfs;
+	radeon_mem_types_list[i].driver_features = 0;
+	radeon_mem_types_list[i].data = NULL;
+	return radeon_debugfs_add_files(rdev, radeon_mem_types_list, RADEON_DEBUGFS_MEM_TYPES+1);
 
 #endif
 	return 0;
-- 
1.6.3.3


------------------------------------------------------------------------------
Download Intel&#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
--

[PATCH 4/7] drm/nouveau: Add ttm page pool debugfs file.

[Pauli Nieminen]

ttm_page_pool file is hooked ttm_page_alloc_debugfs for pool
allocator state.

Signed-off-by: Pauli Nieminen <suokkos@gmail.com>
---
 drivers/gpu/drm/nouveau/nouveau_debugfs.c |    3 +++
 1 files changed, 3 insertions(+), 0 deletions(-)

diff --git a/drivers/gpu/drm/nouveau/nouveau_debugfs.c b/drivers/gpu/drm/nouveau/nouveau_debugfs.c
index 8ff9ef5..59416c8 100644
--- a/drivers/gpu/drm/nouveau/nouveau_debugfs.c
+++ b/drivers/gpu/drm/nouveau/nouveau_debugfs.c
@@ -33,6 +33,8 @@
 #include "drmP.h"
 #include "nouveau_drv.h"
 
+#include <ttm/ttm_page_alloc.h>
+
 static int
 nouveau_debugfs_channel_info(struct seq_file *m, void *data)
 {
@@ -160,6 +162,7 @@ static struct drm_info_list nouveau_debugfs_list[] = {
 	{ "chipset", nouveau_debugfs_chipset_info, 0, NULL },
 	{ "memory", nouveau_debugfs_memory_info, 0, NULL },
 	{ "vbios.rom", nouveau_debugfs_vbios_image, 0, NULL },
+	{ "ttm_page_pool", ttm_page_alloc_debugfs, 0, NULL },
 };
 #define NOUVEAU_DEBUGFS_ENTRIES ARRAY_SIZE(nouveau_debugfs_list)
 
-- 
1.6.3.3


------------------------------------------------------------------------------
Download Intel&#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
--

[PATCH 5/7] drm/ttm: Dynamically scale the allocation sizes of pools.

[Pauli Nieminen]

If we need large number of pages allocations in short time frame allocating
large number of pages in single operation performs better than many small
operations. If there is less allocation operations allocating large number of
pages in a fill operation would waste memory and performance.

The allocation sizes of pools is doubled everytime refill is done more than
once between shrink calls. While the size is halfed happens if there is no
refills for manager->shrink_alloc_interval seconds.

Signed-off-by: Pauli Nieminen <suokkos@gmail.com>
---
 drivers/gpu/drm/ttm/ttm_page_alloc.c |  105 ++++++++++++++++++++++++++--------
 1 files changed, 81 insertions(+), 24 deletions(-)

diff --git a/drivers/gpu/drm/ttm/ttm_page_alloc.c b/drivers/gpu/drm/ttm/ttm_page_alloc.c
index 206bee9..af91994 100644
--- a/drivers/gpu/drm/ttm/ttm_page_alloc.c
+++ b/drivers/gpu/drm/ttm/ttm_page_alloc.c
@@ -47,8 +47,9 @@
 
 
 #define NUM_PAGES_TO_ALLOC		256
-#define SMALL_ALLOCATION		64
+#define SMALL_ALLOCATION		16
 #define FREE_ALL_PAGES			1
+#define SHRINK_ALLOC_INTERVAL		8
 /* times are in msecs */
 #define PAGE_FREE_INTERVAL		1000
 
@@ -63,7 +64,10 @@
  * @gfp_flags: Flags to pass for alloc_page.
  * @npages: Number of pages in pool
  * @nlowpages: Minimum nubmer of pages in pool since previous shrink
+ * @freed_pages: Number of pages freed in this shrinker run.
  * @alloc_size: Allocation sizes of this pool.
+ * @last_refill: Number of shrink calls (with more than alloc_size pages free)
+ * since last refill call.
  * operation.
  */
 struct ttm_page_pool {
@@ -73,7 +77,9 @@ struct ttm_page_pool {
 	int			gfp_flags;
 	unsigned		npages;
 	unsigned		nlowpages;
+	unsigned		freed_pages;
 	unsigned		alloc_size;
+	unsigned		last_refill;
 	char			*name;
 	unsigned long		nfrees;
 	unsigned long		nrefills;
@@ -101,6 +107,7 @@ struct ttm_pool_manager {
 	atomic_t		page_alloc_inited;
 	struct delayed_work	work;
 	unsigned		small_allocation;
+	unsigned		shrink_alloc_interval;
 
 	union {
 		struct ttm_page_pool	pools[NUM_POOLS];
@@ -199,17 +206,33 @@ static void ttm_pages_put(struct page *pages[], unsigned npages)
  **/
 static bool ttm_reset_pools(struct ttm_pool_manager *manager)
 {
+	struct ttm_page_pool *pool;
 	unsigned long irq_flags;
-	bool pages_in_pool = false;
+	bool more_delayed_work = false;
 	unsigned i;
 	for (i = 0; i < NUM_POOLS; ++i) {
-		spin_lock_irqsave(&manager->pools[i].lock, irq_flags);
-		manager->pools[i].nlowpages = manager->pools[i].npages;
-		pages_in_pool = pages_in_pool
-			|| manager->pools[i].npages > manager->pools[i].alloc_size;
-		spin_unlock_irqrestore(&manager->pools[i].lock, irq_flags);
+		pool = &manager->pools[i];
+		spin_lock_irqsave(&pool->lock, irq_flags);
+		/**
+		 * Check if pool has shrink work in a second.
+		 **/
+		if (pool->alloc_size > manager->small_allocation) {
+			/* Reduce pool sizes if there is no refills for
+			 * manager->shrink_alloc_interval seconds. */
+			if (++pool->last_refill % manager->shrink_alloc_interval == 0)
+				pool->alloc_size /= 2;
+
+			more_delayed_work = true;
+		} else
+			more_delayed_work = more_delayed_work
+				|| pool->npages > pool->alloc_size;
+
+		pool->nlowpages = pool->npages;
+		pool->freed_pages = 0;
+
+		spin_unlock_irqrestore(&pool->lock, irq_flags);
 	}
-	return pages_in_pool;
+	return more_delayed_work;
 }
 
 /**
@@ -220,16 +243,19 @@ static bool ttm_reset_pools(struct ttm_pool_manager *manager)
 static unsigned ttm_page_pool_get_npages_to_free_locked(struct ttm_page_pool *pool)
 {
 	unsigned r;
+	unsigned low = pool->nlowpages - pool->freed_pages;
+	if (pool->nlowpages < pool->freed_pages)
+		low = 0;
 	/* If less than alloc sizes was the lowest number of pages we don't
 	 * free any */
-	if (pool->nlowpages < pool->alloc_size)
+	if (low < pool->alloc_size)
 		return 0;
 	/* leave half of unused pages to pool */
-	r = (pool->nlowpages - pool->alloc_size)/2;
+	r = (low - pool->alloc_size)/2;
 	if (r)
 		return r;
 	/* make sure we remove all pages even when there is rounding down */
-	if (pool->nlowpages)
+	if (low)
 		return 1;
 	return 0;
 }
@@ -251,12 +277,12 @@ static bool ttm_page_pool_free_pages_locked(struct ttm_page_pool *pool,
 	 */
 	tmp = 2*freed_pages;
 	/* protect against rounding errors */
-	if (tmp < pool->nlowpages) {
-		pool->nlowpages -= tmp;
+	if (tmp < pool->nlowpages - pool->freed_pages) {
+		pool->freed_pages += tmp;
 		return true;
 	}
 
-	pool->nlowpages = 0;
+	pool->freed_pages = pool->nlowpages;
 	return false;
 }
 
@@ -329,7 +355,7 @@ restart:
 	/* set nlowpages to zero to prevent extra freeing in thsi patch.
 	 * nlowpages is reseted later after all work has been finnished.
 	 **/
-	pool->nlowpages = 0;
+	pool->freed_pages = pool->nlowpages;
 
 	/* remove range of pages from the pool */
 	if (freed_pages)
@@ -516,6 +542,8 @@ static void ttm_page_pool_fill_locked(struct ttm_page_pool *pool,
 	struct page *p, *tmp;
 	int r;
 	unsigned cpages = 0;
+	unsigned alloc_size;
+	bool queue_shrink = false;
 	/**
 	 * Only allow one pool fill operation at a time.
 	 * If pool doesn't have enough pages for the allocation new pages are
@@ -528,6 +556,24 @@ static void ttm_page_pool_fill_locked(struct ttm_page_pool *pool,
 
 	if (count < _manager.small_allocation
 		&& count > pool->npages) {
+		/* Increase allocation sizes if we are refilling pool multiple
+		 * times between shrink calls. Clamp alloc size of the pool
+		 * to NUM_PAGES_TO_ALLOC * 2. */
+		if (pool->last_refill == 0
+			&& pool->alloc_size < NUM_PAGES_TO_ALLOC * 2) {
+			pool->alloc_size *= 2;
+			queue_shrink = true;
+		}
+
+		pool->last_refill = 0;
+		alloc_size = pool->alloc_size;
+
+		/* clamp size to NUM_PAGES_TO_ALLOC because it is maximum
+		 * number of pages in a single caching change.
+		 */
+		if (alloc_size > NUM_PAGES_TO_ALLOC)
+			alloc_size = NUM_PAGES_TO_ALLOC;
+
 		/* If allocation request is small and there is not enough
 		 * pages in pool we fill the pool first */
 		INIT_LIST_HEAD(&new_pages);
@@ -538,16 +584,20 @@ static void ttm_page_pool_fill_locked(struct ttm_page_pool *pool,
 		 */
 		spin_unlock_irqrestore(&pool->lock, irq_flags);
 		r = ttm_alloc_new_pages(&new_pages, pool->gfp_flags, ttm_flags,
-				cstate,	pool->alloc_size);
+				cstate,	alloc_size);
+		if (queue_shrink)
+			(void)queue_delayed_work(_manager.glob->swap_queue,
+					&_manager.work,
+					round_jiffies(_manager.free_interval));
 		spin_lock_irqsave(&pool->lock, irq_flags);
 
 		if (!r) {
 			list_splice(&new_pages, &pool->list);
 			++pool->nrefills;
-			pool->npages += pool->alloc_size;
+			pool->npages += alloc_size;
 			/* Have to remmber to update the low number of pages
 			 * too */
-			pool->nlowpages += pool->alloc_size;
+			pool->nlowpages += alloc_size;
 		} else {
 			printk(KERN_ERR "[ttm] Failed to fill pool (%p).", pool);
 			/* If we have any pages left put them to the pool. */
@@ -738,7 +788,9 @@ static void ttm_page_pool_init_locked(struct ttm_page_pool *pool, int flags,
 	pool->fill_lock = false;
 	INIT_LIST_HEAD(&pool->list);
 	pool->npages = pool->nlowpages = pool->nfrees = 0;
-	pool->alloc_size = NUM_PAGES_TO_ALLOC;
+	pool->last_refill = 0;
+	pool->freed_pages = 0;
+	pool->alloc_size = SMALL_ALLOCATION;
 	pool->gfp_flags = flags;
 	pool->name = name;
 }
@@ -762,10 +814,14 @@ int ttm_page_alloc_init(struct ttm_mem_global *glob)
 
 	_manager.free_interval = msecs_to_jiffies(PAGE_FREE_INTERVAL);
 	_manager.small_allocation = SMALL_ALLOCATION;
+	_manager.shrink_alloc_interval = SHRINK_ALLOC_INTERVAL;
 	_manager.glob = glob;
 
 	INIT_DELAYED_WORK(&_manager.work, ttm_pool_shrink);
 
+	(void)queue_delayed_work(_manager.glob->swap_queue,
+			&_manager.work,
+			round_jiffies(_manager.free_interval));
 	return 0;
 }
 
@@ -789,20 +845,21 @@ int ttm_page_alloc_debugfs(struct seq_file *m, void *data)
 {
 	struct ttm_page_pool *p;
 	unsigned i;
-	char *h[] = {"pool", "refills", "pages freed", "size", "min size"};
+	char *h[] = {"pool", "refills", "pages freed", "size", "min size",
+		"alloc size"};
 	if (atomic_read(&_manager.page_alloc_inited) == 0) {
 		seq_printf(m, "No pool allocator running.\n");
 		return 0;
 	}
-	seq_printf(m, "%6s %12s %13s %8s %8s\n",
-			h[0], h[1], h[2], h[3], h[4]);
+	seq_printf(m, "%6s %12s %13s %8s %8s %8s\n",
+			h[0], h[1], h[2], h[3], h[4], h[5]);
 	for (i = 0; i < NUM_POOLS; ++i) {
 		p = &_manager.pools[i];
 
-		seq_printf(m, "%6s %12ld %13ld %8d %8d\n",
+		seq_printf(m, "%6s %12ld %13ld %8d %8d %8d\n",
 				p->name, p->nrefills,
 				p->nfrees, p->npages,
-				p->nlowpages);
+				p->nlowpages, p->alloc_size);
 	}
 	return 0;
 }
-- 
1.6.3.3


------------------------------------------------------------------------------
Download Intel&#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
--

[PATCH 6/7] arch/x86: Add array variants for setting memory to wc caching.

[Pauli Nieminen]

Setting single memory pages at a time to wc takes a lot time in cache flush. To
reduce number of cache flush set_pages_array_wc and set_memory_array_wc can be
used to set multiple pages to WC with single cache flush.

This improves allocation performance for wc cached pages in drm/ttm.

Signed-off-by: Pauli Nieminen <suokkos@gmail.com>
---
 arch/x86/include/asm/cacheflush.h |    2 +
 arch/x86/mm/pageattr.c            |   53 +++++++++++++++++++++++++++++++-----
 2 files changed, 47 insertions(+), 8 deletions(-)

diff --git a/arch/x86/include/asm/cacheflush.h b/arch/x86/include/asm/cacheflush.h
index 634c40a..d92d63a 100644
--- a/arch/x86/include/asm/cacheflush.h
+++ b/arch/x86/include/asm/cacheflush.h
@@ -139,9 +139,11 @@ int set_memory_np(unsigned long addr, int numpages);
 int set_memory_4k(unsigned long addr, int numpages);
 
 int set_memory_array_uc(unsigned long *addr, int addrinarray);
+int set_memory_array_wc(unsigned long *addr, int addrinarray);
 int set_memory_array_wb(unsigned long *addr, int addrinarray);
 
 int set_pages_array_uc(struct page **pages, int addrinarray);
+int set_pages_array_wc(struct page **pages, int addrinarray);
 int set_pages_array_wb(struct page **pages, int addrinarray);
 
 /*
diff --git a/arch/x86/mm/pageattr.c b/arch/x86/mm/pageattr.c
index cf07c26..0c98a75 100644
--- a/arch/x86/mm/pageattr.c
+++ b/arch/x86/mm/pageattr.c
@@ -997,7 +997,8 @@ out_err:
 }
 EXPORT_SYMBOL(set_memory_uc);
 
-int set_memory_array_uc(unsigned long *addr, int addrinarray)
+int _set_memory_array(unsigned long *addr, int addrinarray,
+		unsigned long new_type)
 {
 	int i, j;
 	int ret;
@@ -1007,13 +1008,19 @@ int set_memory_array_uc(unsigned long *addr, int addrinarray)
 	 */
 	for (i = 0; i < addrinarray; i++) {
 		ret = reserve_memtype(__pa(addr[i]), __pa(addr[i]) + PAGE_SIZE,
-					_PAGE_CACHE_UC_MINUS, NULL);
+					new_type, NULL);
 		if (ret)
 			goto out_free;
 	}
 
 	ret = change_page_attr_set(addr, addrinarray,
 				    __pgprot(_PAGE_CACHE_UC_MINUS), 1);
+
+	if (!ret && new_type == _PAGE_CACHE_WC)
+		ret = change_page_attr_set_clr(addr, addrinarray,
+					       __pgprot(_PAGE_CACHE_WC),
+					       __pgprot(_PAGE_CACHE_MASK),
+					       0, CPA_ARRAY, NULL);
 	if (ret)
 		goto out_free;
 
@@ -1025,8 +1032,19 @@ out_free:
 
 	return ret;
 }
+
+int set_memory_array_uc(unsigned long *addr, int addrinarray)
+{
+	return _set_memory_array(addr, addrinarray, _PAGE_CACHE_UC_MINUS);
+}
 EXPORT_SYMBOL(set_memory_array_uc);
 
+int set_memory_array_wc(unsigned long *addr, int addrinarray)
+{
+	return _set_memory_array(addr, addrinarray, _PAGE_CACHE_WC);
+}
+EXPORT_SYMBOL(set_memory_array_wc);
+
 int _set_memory_wc(unsigned long addr, int numpages)
 {
 	int ret;
@@ -1153,26 +1171,34 @@ int set_pages_uc(struct page *page, int numpages)
 }
 EXPORT_SYMBOL(set_pages_uc);
 
-int set_pages_array_uc(struct page **pages, int addrinarray)
+static int _set_pages_array(struct page **pages, int addrinarray,
+		unsigned long new_type)
 {
 	unsigned long start;
 	unsigned long end;
 	int i;
 	int free_idx;
+	int ret;
 
 	for (i = 0; i < addrinarray; i++) {
 		if (PageHighMem(pages[i]))
 			continue;
 		start = page_to_pfn(pages[i]) << PAGE_SHIFT;
 		end = start + PAGE_SIZE;
-		if (reserve_memtype(start, end, _PAGE_CACHE_UC_MINUS, NULL))
+		if (reserve_memtype(start, end, new_type, NULL))
 			goto err_out;
 	}
 
-	if (cpa_set_pages_array(pages, addrinarray,
-			__pgprot(_PAGE_CACHE_UC_MINUS)) == 0) {
-		return 0; /* Success */
-	}
+	ret = cpa_set_pages_array(pages, addrinarray,
+			__pgprot(_PAGE_CACHE_UC_MINUS));
+	if (!ret && new_type == _PAGE_CACHE_WC)
+		ret = change_page_attr_set_clr(NULL, addrinarray,
+					       __pgprot(_PAGE_CACHE_WC),
+					       __pgprot(_PAGE_CACHE_MASK),
+					       0, CPA_PAGES_ARRAY, pages);
+	if (ret)
+		goto err_out;
+	return 0; /* Success */
 err_out:
 	free_idx = i;
 	for (i = 0; i < free_idx; i++) {
@@ -1184,8 +1210,19 @@ err_out:
 	}
 	return -EINVAL;
 }
+
+int set_pages_array_uc(struct page **pages, int addrinarray)
+{
+	return _set_pages_array(pages, addrinarray, _PAGE_CACHE_UC_MINUS);
+}
 EXPORT_SYMBOL(set_pages_array_uc);
 
+int set_pages_array_wc(struct page **pages, int addrinarray)
+{
+	return _set_pages_array(pages, addrinarray, _PAGE_CACHE_WC);
+}
+EXPORT_SYMBOL(set_pages_array_wc);
+
 int set_pages_wb(struct page *page, int numpages)
 {
 	unsigned long addr = (unsigned long)page_address(page);
-- 
1.6.3.3


------------------------------------------------------------------------------
Download Intel&#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
--

[PATCH 7/7] drm/ttm: Use set_pages_array_wc instead of set_memory_wc.

[Pauli Nieminen]

Using single call to set multiple pages to wc reduces number of expensive cache
flushes.

Signed-off-by: Pauli Nieminen <suokkos@gmail.com>
---
 drivers/gpu/drm/ttm/ttm_page_alloc.c |   12 +-----------
 1 files changed, 1 insertions(+), 11 deletions(-)

diff --git a/drivers/gpu/drm/ttm/ttm_page_alloc.c b/drivers/gpu/drm/ttm/ttm_page_alloc.c
index af91994..111c70b 100644
--- a/drivers/gpu/drm/ttm/ttm_page_alloc.c
+++ b/drivers/gpu/drm/ttm/ttm_page_alloc.c
@@ -124,17 +124,7 @@ static struct ttm_pool_manager _manager = {
 	.page_alloc_inited	= ATOMIC_INIT(0)
 };
 
-#ifdef CONFIG_X86
-/* TODO: add this to x86 like _uc, this version here is inefficient */
-static int set_pages_array_wc(struct page **pages, int addrinarray)
-{
-	int i;
-
-	for (i = 0; i < addrinarray; i++)
-		set_memory_wc((unsigned long)page_address(pages[i]), 1);
-	return 0;
-}
-#else
+#ifndef CONFIG_X86
 static int set_pages_array_wb(struct page **pages, int addrinarray)
 {
 #ifdef TTM_HAS_AGP
-- 
1.6.3.3


------------------------------------------------------------------------------
Download Intel&#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
--

Re: [PATCH 6/7] arch/x86: Add array variants for setting memory to wc caching.

[Dave Airlie]

On Thu, Mar 18, 2010 at 6:50 AM, Pauli Nieminen <suokkos@gmail.com> wrote:
> Setting single memory pages at a time to wc takes a lot time in cache flush. To
> reduce number of cache flush set_pages_array_wc and set_memory_array_wc can be
> used to set multiple pages to WC with single cache flush.

I don't think this is correct, I've cc'ed Suresh and Venki who looked
at this before,
but I think there is an array in the x86 code that stores the state
and it only has a
single bit in it, it needs to be expanded in order to at WC support.

Dave.

>
> This improves allocation performance for wc cached pages in drm/ttm.
>
> Signed-off-by: Pauli Nieminen <suokkos@gmail.com>
> ---
>  arch/x86/include/asm/cacheflush.h |    2 +
>  arch/x86/mm/pageattr.c            |   53 +++++++++++++++++++++++++++++++-----
>  2 files changed, 47 insertions(+), 8 deletions(-)
>
> diff --git a/arch/x86/include/asm/cacheflush.h b/arch/x86/include/asm/cacheflush.h
> index 634c40a..d92d63a 100644
> --- a/arch/x86/include/asm/cacheflush.h
> +++ b/arch/x86/include/asm/cacheflush.h
> @@ -139,9 +139,11 @@ int set_memory_np(unsigned long addr, int numpages);
>  int set_memory_4k(unsigned long addr, int numpages);
>
>  int set_memory_array_uc(unsigned long *addr, int addrinarray);
> +int set_memory_array_wc(unsigned long *addr, int addrinarray);
>  int set_memory_array_wb(unsigned long *addr, int addrinarray);
>
>  int set_pages_array_uc(struct page **pages, int addrinarray);
> +int set_pages_array_wc(struct page **pages, int addrinarray);
>  int set_pages_array_wb(struct page **pages, int addrinarray);
>
>  /*
> diff --git a/arch/x86/mm/pageattr.c b/arch/x86/mm/pageattr.c
> index cf07c26..0c98a75 100644
> --- a/arch/x86/mm/pageattr.c
> +++ b/arch/x86/mm/pageattr.c
> @@ -997,7 +997,8 @@ out_err:
>  }
>  EXPORT_SYMBOL(set_memory_uc);
>
> -int set_memory_array_uc(unsigned long *addr, int addrinarray)
> +int _set_memory_array(unsigned long *addr, int addrinarray,
> +               unsigned long new_type)
>  {
>        int i, j;
>        int ret;
> @@ -1007,13 +1008,19 @@ int set_memory_array_uc(unsigned long *addr, int addrinarray)
>         */
>        for (i = 0; i < addrinarray; i++) {
>                ret = reserve_memtype(__pa(addr[i]), __pa(addr[i]) + PAGE_SIZE,
> -                                       _PAGE_CACHE_UC_MINUS, NULL);
> +                                       new_type, NULL);
>                if (ret)
>                        goto out_free;
>        }
>
>        ret = change_page_attr_set(addr, addrinarray,
>                                    __pgprot(_PAGE_CACHE_UC_MINUS), 1);
> +
> +       if (!ret && new_type == _PAGE_CACHE_WC)
> +               ret = change_page_attr_set_clr(addr, addrinarray,
> +                                              __pgprot(_PAGE_CACHE_WC),
> +                                              __pgprot(_PAGE_CACHE_MASK),
> +                                              0, CPA_ARRAY, NULL);
>        if (ret)
>                goto out_free;
>
> @@ -1025,8 +1032,19 @@ out_free:
>
>        return ret;
>  }
> +
> +int set_memory_array_uc(unsigned long *addr, int addrinarray)
> +{
> +       return _set_memory_array(addr, addrinarray, _PAGE_CACHE_UC_MINUS);
> +}
>  EXPORT_SYMBOL(set_memory_array_uc);
>
> +int set_memory_array_wc(unsigned long *addr, int addrinarray)
> +{
> +       return _set_memory_array(addr, addrinarray, _PAGE_CACHE_WC);
> +}
> +EXPORT_SYMBOL(set_memory_array_wc);
> +
>  int _set_memory_wc(unsigned long addr, int numpages)
>  {
>        int ret;
> @@ -1153,26 +1171,34 @@ int set_pages_uc(struct page *page, int numpages)
>  }
>  EXPORT_SYMBOL(set_pages_uc);
>
> -int set_pages_array_uc(struct page **pages, int addrinarray)
> +static int _set_pages_array(struct page **pages, int addrinarray,
> +               unsigned long new_type)
>  {
>        unsigned long start;
>        unsigned long end;
>        int i;
>        int free_idx;
> +       int ret;
>
>        for (i = 0; i < addrinarray; i++) {
>                if (PageHighMem(pages[i]))
>                        continue;
>                start = page_to_pfn(pages[i]) << PAGE_SHIFT;
>                end = start + PAGE_SIZE;
> -               if (reserve_memtype(start, end, _PAGE_CACHE_UC_MINUS, NULL))
> +               if (reserve_memtype(start, end, new_type, NULL))
>                        goto err_out;
>        }
>
> -       if (cpa_set_pages_array(pages, addrinarray,
> -                       __pgprot(_PAGE_CACHE_UC_MINUS)) == 0) {
> -               return 0; /* Success */
> -       }
> +       ret = cpa_set_pages_array(pages, addrinarray,
> +                       __pgprot(_PAGE_CACHE_UC_MINUS));
> +       if (!ret && new_type == _PAGE_CACHE_WC)
> +               ret = change_page_attr_set_clr(NULL, addrinarray,
> +                                              __pgprot(_PAGE_CACHE_WC),
> +                                              __pgprot(_PAGE_CACHE_MASK),
> +                                              0, CPA_PAGES_ARRAY, pages);
> +       if (ret)
> +               goto err_out;
> +       return 0; /* Success */
>  err_out:
>        free_idx = i;
>        for (i = 0; i < free_idx; i++) {
> @@ -1184,8 +1210,19 @@ err_out:
>        }
>        return -EINVAL;
>  }
> +
> +int set_pages_array_uc(struct page **pages, int addrinarray)
> +{
> +       return _set_pages_array(pages, addrinarray, _PAGE_CACHE_UC_MINUS);
> +}
>  EXPORT_SYMBOL(set_pages_array_uc);
>
> +int set_pages_array_wc(struct page **pages, int addrinarray)
> +{
> +       return _set_pages_array(pages, addrinarray, _PAGE_CACHE_WC);
> +}
> +EXPORT_SYMBOL(set_pages_array_wc);
> +
>  int set_pages_wb(struct page *page, int numpages)
>  {
>        unsigned long addr = (unsigned long)page_address(page);
> --
> 1.6.3.3
>
>

------------------------------------------------------------------------------
Download Intel&#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
--

Re: [PATCH 6/7] arch/x86: Add array variants for setting memory to wc caching.

[Pauli Nieminen]

On Thu, Mar 18, 2010 at 1:52 AM, Dave Airlie <airlied@gmail.com> wrote:
> On Thu, Mar 18, 2010 at 6:50 AM, Pauli Nieminen <suokkos@gmail.com> wrote:
>> Setting single memory pages at a time to wc takes a lot time in cache flush. To
>> reduce number of cache flush set_pages_array_wc and set_memory_array_wc can be
>> used to set multiple pages to WC with single cache flush.
>
> I don't think this is correct, I've cc'ed Suresh and Venki who looked
> at this before,
> but I think there is an array in the x86 code that stores the state
> and it only has a
> single bit in it, it needs to be expanded in order to at WC support.
>
> Dave.
>

This worked for me but there might be some hw specific cases when it
doesn't work. (all page act like wc cached ones when running memcpy
test on gtt bo)

This is only minor improvemnet for corner cases. Pool refills general
don't happen in normal use except for 3D. For others cases number of
pages was fairly static so pool code can avoid refills. Of course
there might be some corner case that I didn't know to test.
Even for them pool refills happened about once a second in average.
But allocation were happening in small bursts so this should avoid
frame rate hit in case of a lot of allocation in very short time.

>>
>> This improves allocation performance for wc cached pages in drm/ttm.
>>
>> Signed-off-by: Pauli Nieminen <suokkos@gmail.com>
>> ---
>>  arch/x86/include/asm/cacheflush.h |    2 +
>>  arch/x86/mm/pageattr.c            |   53 +++++++++++++++++++++++++++++++-----
>>  2 files changed, 47 insertions(+), 8 deletions(-)
>>
>> diff --git a/arch/x86/include/asm/cacheflush.h b/arch/x86/include/asm/cacheflush.h
>> index 634c40a..d92d63a 100644
>> --- a/arch/x86/include/asm/cacheflush.h
>> +++ b/arch/x86/include/asm/cacheflush.h
>> @@ -139,9 +139,11 @@ int set_memory_np(unsigned long addr, int numpages);
>>  int set_memory_4k(unsigned long addr, int numpages);
>>
>>  int set_memory_array_uc(unsigned long *addr, int addrinarray);
>> +int set_memory_array_wc(unsigned long *addr, int addrinarray);
>>  int set_memory_array_wb(unsigned long *addr, int addrinarray);
>>
>>  int set_pages_array_uc(struct page **pages, int addrinarray);
>> +int set_pages_array_wc(struct page **pages, int addrinarray);
>>  int set_pages_array_wb(struct page **pages, int addrinarray);
>>
>>  /*
>> diff --git a/arch/x86/mm/pageattr.c b/arch/x86/mm/pageattr.c
>> index cf07c26..0c98a75 100644
>> --- a/arch/x86/mm/pageattr.c
>> +++ b/arch/x86/mm/pageattr.c
>> @@ -997,7 +997,8 @@ out_err:
>>  }
>>  EXPORT_SYMBOL(set_memory_uc);
>>
>> -int set_memory_array_uc(unsigned long *addr, int addrinarray)
>> +int _set_memory_array(unsigned long *addr, int addrinarray,
>> +               unsigned long new_type)
>>  {
>>        int i, j;
>>        int ret;
>> @@ -1007,13 +1008,19 @@ int set_memory_array_uc(unsigned long *addr, int addrinarray)
>>         */
>>        for (i = 0; i < addrinarray; i++) {
>>                ret = reserve_memtype(__pa(addr[i]), __pa(addr[i]) + PAGE_SIZE,
>> -                                       _PAGE_CACHE_UC_MINUS, NULL);
>> +                                       new_type, NULL);
>>                if (ret)
>>                        goto out_free;
>>        }
>>
>>        ret = change_page_attr_set(addr, addrinarray,
>>                                    __pgprot(_PAGE_CACHE_UC_MINUS), 1);
>> +
>> +       if (!ret && new_type == _PAGE_CACHE_WC)
>> +               ret = change_page_attr_set_clr(addr, addrinarray,
>> +                                              __pgprot(_PAGE_CACHE_WC),
>> +                                              __pgprot(_PAGE_CACHE_MASK),
>> +                                              0, CPA_ARRAY, NULL);
>>        if (ret)
>>                goto out_free;
>>
>> @@ -1025,8 +1032,19 @@ out_free:
>>
>>        return ret;
>>  }
>> +
>> +int set_memory_array_uc(unsigned long *addr, int addrinarray)
>> +{
>> +       return _set_memory_array(addr, addrinarray, _PAGE_CACHE_UC_MINUS);
>> +}
>>  EXPORT_SYMBOL(set_memory_array_uc);
>>
>> +int set_memory_array_wc(unsigned long *addr, int addrinarray)
>> +{
>> +       return _set_memory_array(addr, addrinarray, _PAGE_CACHE_WC);
>> +}
>> +EXPORT_SYMBOL(set_memory_array_wc);
>> +
>>  int _set_memory_wc(unsigned long addr, int numpages)
>>  {
>>        int ret;
>> @@ -1153,26 +1171,34 @@ int set_pages_uc(struct page *page, int numpages)
>>  }
>>  EXPORT_SYMBOL(set_pages_uc);
>>
>> -int set_pages_array_uc(struct page **pages, int addrinarray)
>> +static int _set_pages_array(struct page **pages, int addrinarray,
>> +               unsigned long new_type)
>>  {
>>        unsigned long start;
>>        unsigned long end;
>>        int i;
>>        int free_idx;
>> +       int ret;
>>
>>        for (i = 0; i < addrinarray; i++) {
>>                if (PageHighMem(pages[i]))
>>                        continue;
>>                start = page_to_pfn(pages[i]) << PAGE_SHIFT;
>>                end = start + PAGE_SIZE;
>> -               if (reserve_memtype(start, end, _PAGE_CACHE_UC_MINUS, NULL))
>> +               if (reserve_memtype(start, end, new_type, NULL))
>>                        goto err_out;
>>        }
>>
>> -       if (cpa_set_pages_array(pages, addrinarray,
>> -                       __pgprot(_PAGE_CACHE_UC_MINUS)) == 0) {
>> -               return 0; /* Success */
>> -       }
>> +       ret = cpa_set_pages_array(pages, addrinarray,
>> +                       __pgprot(_PAGE_CACHE_UC_MINUS));
>> +       if (!ret && new_type == _PAGE_CACHE_WC)
>> +               ret = change_page_attr_set_clr(NULL, addrinarray,
>> +                                              __pgprot(_PAGE_CACHE_WC),
>> +                                              __pgprot(_PAGE_CACHE_MASK),
>> +                                              0, CPA_PAGES_ARRAY, pages);
>> +       if (ret)
>> +               goto err_out;
>> +       return 0; /* Success */
>>  err_out:
>>        free_idx = i;
>>        for (i = 0; i < free_idx; i++) {
>> @@ -1184,8 +1210,19 @@ err_out:
>>        }
>>        return -EINVAL;
>>  }
>> +
>> +int set_pages_array_uc(struct page **pages, int addrinarray)
>> +{
>> +       return _set_pages_array(pages, addrinarray, _PAGE_CACHE_UC_MINUS);
>> +}
>>  EXPORT_SYMBOL(set_pages_array_uc);
>>
>> +int set_pages_array_wc(struct page **pages, int addrinarray)
>> +{
>> +       return _set_pages_array(pages, addrinarray, _PAGE_CACHE_WC);
>> +}
>> +EXPORT_SYMBOL(set_pages_array_wc);
>> +
>>  int set_pages_wb(struct page *page, int numpages)
>>  {
>>        unsigned long addr = (unsigned long)page_address(page);
>> --
>> 1.6.3.3
>>
>>
>

------------------------------------------------------------------------------
Download Intel&#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
--

Re: [PATCH 6/7] arch/x86: Add array variants for setting memory to wc caching.

[Pauli Nieminen]

On Thu, Mar 18, 2010 at 9:29 PM, Suresh Siddha
<suresh.b.siddha@intel.com> wrote:
> On Thu, 2010-03-18 at 02:41 -0700, Pauli Nieminen wrote:
>> On Thu, Mar 18, 2010 at 1:52 AM, Dave Airlie <airlied@gmail.com> wrote:
>> > On Thu, Mar 18, 2010 at 6:50 AM, Pauli Nieminen <suokkos@gmail.com> wrote:
>> >> Setting single memory pages at a time to wc takes a lot time in cache flush. To
>> >> reduce number of cache flush set_pages_array_wc and set_memory_array_wc can be
>> >> used to set multiple pages to WC with single cache flush.
>> >
>> > I don't think this is correct, I've cc'ed Suresh and Venki who looked
>> > at this before,
>> > but I think there is an array in the x86 code that stores the state
>> > and it only has a
>> > single bit in it, it needs to be expanded in order to at WC support.
>> >
>
> hmm, I thought we had the array variants for the WC already. I don't
> think there is any explicit reason for not having it. Correcting Venki's
> address, to see if he has anything to add.
>
>> >> +       ret = cpa_set_pages_array(pages, addrinarray,
>> >> +                       __pgprot(_PAGE_CACHE_UC_MINUS));
>> >> +       if (!ret && new_type == _PAGE_CACHE_WC)
>> >> +               ret = change_page_attr_set_clr(NULL, addrinarray,
>> >> +                                              __pgprot(_PAGE_CACHE_WC),
>> >> +                                              __pgprot(_PAGE_CACHE_MASK),
>> >> +                                              0, CPA_PAGES_ARRAY, pages);
>
> Pauli, Is there any reason for not using cpa_set_pages_array() here
> aswell?
>

cpa_set_pages_array is setting mask parameter to zero while
_set_memory_wc is passing the mask parameter. I didn't look deep
enough to the code to know why _set_memory_wc works that way.

> thanks,
> suresh
>
>

------------------------------------------------------------------------------
Download Intel&#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
--

Re: [PATCH 6/7] arch/x86: Add array variants for setting memory to wc caching.

[Suresh Siddha]

On Thu, 2010-03-18 at 02:41 -0700, Pauli Nieminen wrote:
> On Thu, Mar 18, 2010 at 1:52 AM, Dave Airlie <airlied@gmail.com> wrote:
> > On Thu, Mar 18, 2010 at 6:50 AM, Pauli Nieminen <suokkos@gmail.com> wrote:
> >> Setting single memory pages at a time to wc takes a lot time in cache flush. To
> >> reduce number of cache flush set_pages_array_wc and set_memory_array_wc can be
> >> used to set multiple pages to WC with single cache flush.
> >
> > I don't think this is correct, I've cc'ed Suresh and Venki who looked
> > at this before,
> > but I think there is an array in the x86 code that stores the state
> > and it only has a
> > single bit in it, it needs to be expanded in order to at WC support.
> >

hmm, I thought we had the array variants for the WC already. I don't
think there is any explicit reason for not having it. Correcting Venki's
address, to see if he has anything to add.

> >> +       ret = cpa_set_pages_array(pages, addrinarray,
> >> +                       __pgprot(_PAGE_CACHE_UC_MINUS));
> >> +       if (!ret && new_type == _PAGE_CACHE_WC)
> >> +               ret = change_page_attr_set_clr(NULL, addrinarray,
> >> +                                              __pgprot(_PAGE_CACHE_WC),
> >> +                                              __pgprot(_PAGE_CACHE_MASK),
> >> +                                              0, CPA_PAGES_ARRAY, pages);

Pauli, Is there any reason for not using cpa_set_pages_array() here
aswell?

thanks,
suresh


------------------------------------------------------------------------------
Download Intel&#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
--

Re: [PATCH 0/7] drm/ttm: Pool allocator

[Pauli Nieminen]

2010/3/19 Thomas Hellström <thomas@shipmail.org>:
> Pauli, Dave and Jerome,
>
> Before reviewing this, Could you describe a bit how this interfaces with the
> TTM memory accounting. It's important for some systems to be able to set a
> limit beyond which TTM may not pin any pages.
>
> Am I right in assuming that TTM memory accounting is kicked in only when TTM
> allocs and frees pages from the pool?

yes.

TTM memory accounting is still handled in ttm_tt.c so pool is outside
of it. But I can move calls to memory accounting into pool if that is
preferred place.

With current implementation pool can hold about 512 pages more memory
than what is the TTM limit.

> Can the system reclaim *all* pages not used by TTM through a shrink
> mechanism?

Not with current version but I can modify patch so that system can
reclaim all pages. Current lowest limit is 16 pages in pool because
that avoids refills for 2D only desktop use.

The limit is already changed in runtime so making it scale to zero
sized pool is only minor change. But what should happen for pool
refill if system just before forced the pool size to zero?

>
> In the long run, I'd like to have a pool of non-kernel-mapped pages instead
> of a pool of uncached / write-combined pages, because then we'd have quite
> fast transition from write-combined to write-back, but I guess that will be
> something for the future.

I think this can be simulated with multiple pools if free logic is
changed from handling single pool at a time to combine multiple pools
to a single wb transition operation.
Trouble in making very large cache transition operations is that
allocating large continues arrays in kernel is problematic. Current
code is limiting the size of single cache transition operation to
avoid possible memory allocation problems.

>
> /Thomas
>
>
>
>
> Pauli Nieminen wrote:
>>
>> When allocating wc/uc pages cache state transition requires cache flush
>> which
>> is expensive operation. To avoid cache flushes allocation of wc/uc pages
>> should
>> be done in large groups when only single cache flush is required for whole
>> group
>> of pages.
>>
>> In some cases drivers need t oallocate and deallocate many pages in a
>> short time
>> frame. In this case we can avoid cache flushes if we keep pages in the
>> pool before
>> actually freeing them later.
>>
>> arch/x86 was missing set_pages_array_wc and set_memory_array_wc. Patch 6
>> and 7 add
>> missing functions and hooks set_pages_array_wc to the pool allocator.
>>
>>
>>
>> ------------------------------------------------------------------------------
>> Download Intel&#174; Parallel Studio Eval
>> Try the new software tools for yourself. Speed compiling, find bugs
>> proactively, and fine-tune applications for parallel performance.
>> See why Intel Parallel Studio got high marks during beta.
>> http://p.sf.net/sfu/intel-sw-dev
>> --
>> _______________________________________________
>> Dri-devel mailing list
>> Dri-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/dri-devel
>>
>
>
>
>

------------------------------------------------------------------------------
Download Intel&#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
--

Re: [PATCH 0/7] drm/ttm: Pool allocator

[Thomas Hellström]

Pauli Nieminen wrote:
> 2010/3/19 Thomas Hellström <thomas@shipmail.org>:
>   
>> Pauli, Dave and Jerome,
>>
>> Before reviewing this, Could you describe a bit how this interfaces with the
>> TTM memory accounting. It's important for some systems to be able to set a
>> limit beyond which TTM may not pin any pages.
>>
>> Am I right in assuming that TTM memory accounting is kicked in only when TTM
>> allocs and frees pages from the pool?
>>     
>
> yes.
>
> TTM memory accounting is still handled in ttm_tt.c so pool is outside
> of it. But I can move calls to memory accounting into pool if that is
> preferred place.
>
> With current implementation pool can hold about 512 pages more memory
> than what is the TTM limit.
>
>   

In that case I think we should keep the pool outside of TTM memory 
accounting.
Since the system can (or should be able to) reclaim the pool pages 
through the shrink mechanism, the
pool is sort of transparent from a memory accounting point of view 
although shrinking comes with a performance penalty.

>> Can the system reclaim *all* pages not used by TTM through a shrink
>> mechanism?
>>     
>
> Not with current version but I can modify patch so that system can
> reclaim all pages. Current lowest limit is 16 pages in pool because
> that avoids refills for 2D only desktop use.
>
> The limit is already changed in runtime so making it scale to zero
> sized pool is only minor change. But what should happen for pool
> refill if system just before forced the pool size to zero?
>   
16 pages sounds OK to me.

Perhaps we'd want to be able to change both this limit and the upper 
pool size limit from a sysfs interface, within
ttm/memory_accounting ?
>   
>> In the long run, I'd like to have a pool of non-kernel-mapped pages instead
>> of a pool of uncached / write-combined pages, because then we'd have quite
>> fast transition from write-combined to write-back, but I guess that will be
>> something for the future.
>>     
>
> I think this can be simulated with multiple pools if free logic is
> changed from handling single pool at a time to combine multiple pools
> to a single wb transition operation.
> Trouble in making very large cache transition operations is that
> allocating large continues arrays in kernel is problematic. Current
> code is limiting the size of single cache transition operation to
> avoid possible memory allocation problems.
>
>   
I think that's a good strategy.

However, the use-case I was thinking of was not fast freeing, but fast 
reading of GPU memory, like so:

*) Unbind from AGP.
*) Mark pages write-back() while the kernel map is still 'unmapped' 
(just like highmem pages).
*) Read
*) Cache flush pages and Mark pages write-combined(), while the kernel 
map is still 'unmapped'.

The operation will not involve any 'set_pages_x' , because the kernel 
map is not affected.
However, there is no exported api to unmap pages from the linear kernel 
map (yet).

/Thomas

>> /Thomas
>>
>>
>>
>>
>> Pauli Nieminen wrote:
>>     
>>> When allocating wc/uc pages cache state transition requires cache flush
>>> which
>>> is expensive operation. To avoid cache flushes allocation of wc/uc pages
>>> should
>>> be done in large groups when only single cache flush is required for whole
>>> group
>>> of pages.
>>>
>>> In some cases drivers need t oallocate and deallocate many pages in a
>>> short time
>>> frame. In this case we can avoid cache flushes if we keep pages in the
>>> pool before
>>> actually freeing them later.
>>>
>>> arch/x86 was missing set_pages_array_wc and set_memory_array_wc. Patch 6
>>> and 7 add
>>> missing functions and hooks set_pages_array_wc to the pool allocator.
>>>
>>>
>>>
>>> ------------------------------------------------------------------------------
>>> Download Intel&#174; Parallel Studio Eval
>>> Try the new software tools for yourself. Speed compiling, find bugs
>>> proactively, and fine-tune applications for parallel performance.
>>> See why Intel Parallel Studio got high marks during beta.
>>> http://p.sf.net/sfu/intel-sw-dev
>>> --
>>> _______________________________________________
>>> Dri-devel mailing list
>>> Dri-devel@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/dri-devel
>>>
>>>       
>>
>>
>>     




------------------------------------------------------------------------------
Download Intel&#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
--