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

Commit e34f44b3 authored by Matthew Wilcox's avatar Matthew Wilcox
Browse files

pool: Improve memory usage for devices which can't cross boundaries 



The previous implementation simply refused to allocate more than a
boundary's worth of data from an entire page.  Some users didn't know
this, so specified things like SMP_CACHE_BYTES, not realising the
horrible waste of memory that this was.  It's fairly easy to correct
this problem, just by ensuring we don't cross a boundary within a page.
This even helps drivers like EHCI (which can't cross a 4k boundary)
on machines with larger page sizes.

Signed-off-by: default avatarMatthew Wilcox <willy@linux.intel.com>
Acked-by: default avatarDavid S. Miller <davem@davemloft.net>
parent a35a3455

mm/dmapool.c

+20 −16
Original line number Original line Diff line number Diff line
@@ -43,6 +43,7 @@ struct dma_pool { /* the pool */ 
	size_t size;

	size_t size;

	struct device *dev;

	struct device *dev;

	size_t allocation;

	size_t allocation;

	size_t boundary;

	char name[32];

	char name[32];

	wait_queue_head_t waitq;

	wait_queue_head_t waitq;

	struct list_head pools;

	struct list_head pools;
@@ -107,7 +108,7 @@ static DEVICE_ATTR(pools, S_IRUGO, show_pools, NULL); 
 * @dev: device that will be doing the DMA

 * @dev: device that will be doing the DMA

 * @size: size of the blocks in this pool.

 * @size: size of the blocks in this pool.

 * @align: alignment requirement for blocks; must be a power of two

 * @align: alignment requirement for blocks; must be a power of two

 * @allocation: returned blocks won't cross this boundary (or zero)

 * @boundary: returned blocks won't cross this power of two boundary

 * Context: !in_interrupt()

 * Context: !in_interrupt()

 *

 *

 * Returns a dma allocation pool with the requested characteristics, or

 * Returns a dma allocation pool with the requested characteristics, or
@@ -117,15 +118,16 @@ static DEVICE_ATTR(pools, S_IRUGO, show_pools, NULL); 
 * cache flushing primitives.  The actual size of blocks allocated may be

 * cache flushing primitives.  The actual size of blocks allocated may be

 * larger than requested because of alignment.

 * larger than requested because of alignment.

 *

 *

 * If allocation is nonzero, objects returned from dma_pool_alloc() won't

 * If @boundary is nonzero, objects returned from dma_pool_alloc() won't

 * cross that size boundary.  This is useful for devices which have

 * cross that size boundary.  This is useful for devices which have

 * addressing restrictions on individual DMA transfers, such as not crossing

 * addressing restrictions on individual DMA transfers, such as not crossing

 * boundaries of 4KBytes.

 * boundaries of 4KBytes.

 */

 */

struct dma_pool *dma_pool_create(const char *name, struct device *dev,

struct dma_pool *dma_pool_create(const char *name, struct device *dev,

				 size_t size, size_t align, size_t allocation)

				 size_t size, size_t align, size_t boundary)

{

{

	struct dma_pool *retval;

	struct dma_pool *retval;

	size_t allocation;





	if (align == 0) {

	if (align == 0) {

		align = 1;

		align = 1;
@@ -142,27 +144,26 @@ struct dma_pool *dma_pool_create(const char *name, struct device *dev, 
	if ((size % align) != 0)

	if ((size % align) != 0)

		size = ALIGN(size, align);

		size = ALIGN(size, align);





	if (allocation == 0) {

	allocation = max_t(size_t, size, PAGE_SIZE);

		if (PAGE_SIZE < size)



			allocation = size;

	if (!boundary) {

		else

		boundary = allocation;

			allocation = PAGE_SIZE;

	} else if ((boundary < size) || (boundary & (boundary - 1))) {

		/* FIXME: round up for less fragmentation */

	} else if (allocation < size)

		return NULL;

		return NULL;

	}





	if (!

	retval = kmalloc_node(sizeof(*retval), GFP_KERNEL, dev_to_node(dev));

	    (retval =

	if (!retval)

	     kmalloc_node(sizeof *retval, GFP_KERNEL, dev_to_node(dev))))

		return retval;

		return retval;





	strlcpy(retval->name, name, sizeof retval->name);

	strlcpy(retval->name, name, sizeof(retval->name));





	retval->dev = dev;

	retval->dev = dev;





	INIT_LIST_HEAD(&retval->page_list);

	INIT_LIST_HEAD(&retval->page_list);

	spin_lock_init(&retval->lock);

	spin_lock_init(&retval->lock);

	retval->size = size;

	retval->size = size;

	retval->boundary = boundary;

	retval->allocation = allocation;

	retval->allocation = allocation;

	init_waitqueue_head(&retval->waitq);

	init_waitqueue_head(&retval->waitq);
@@ -192,11 +193,14 @@ EXPORT_SYMBOL(dma_pool_create); 
static void pool_initialise_page(struct dma_pool *pool, struct dma_page *page)

static void pool_initialise_page(struct dma_pool *pool, struct dma_page *page)

{

{

	unsigned int offset = 0;

	unsigned int offset = 0;

	unsigned int next_boundary = pool->boundary;





	do {

	do {

		unsigned int next = offset + pool->size;

		unsigned int next = offset + pool->size;

		if (unlikely((next + pool->size) >= pool->allocation))

		if (unlikely((next + pool->size) >= next_boundary)) {

			next = pool->allocation;

			next = next_boundary;

			next_boundary += pool->boundary;

		}

		*(int *)(page->vaddr + offset) = next;

		*(int *)(page->vaddr + offset) = next;

		offset = next;

		offset = next;

	} while (offset < pool->allocation);

	} while (offset < pool->allocation);