Skip to contentSkip to conversation

brd: support discard

PATCH In review Nick Piggin 3 participants 4 messages May 26, 2010 – May 27, 2010

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 brd: support discard

Hi Jens,

What do you think of this patch? Can you possibly merge it via your block tree if it looks OK?

Thanks,
Nick
--

brd: support discard

Support discard requests in brd by zeroing or deleting the underlying backing pages. This is simply to help with testing and documentation nature of brd code.

---

drivers/block/brd.c +52 -1
Index: linux-2.6/drivers/block/brd.c
===================================================================
@@ -133,6 +133,28 @@ static struct page *brd_insert_page(stru
 	return page;
 }
 
+static void brd_free_page(struct brd_device *brd, sector_t sector)
+{
+	struct page *page;
+	pgoff_t idx;
+
+	spin_lock(&brd->brd_lock);
+	idx = sector >> PAGE_SECTORS_SHIFT;
+	page = radix_tree_delete(&brd->brd_pages, idx);
+	spin_unlock(&brd->brd_lock);
+	if (page)
+		__free_page(page);
+}
+
+static void brd_zero_page(struct brd_device *brd, sector_t sector)
+{
+	struct page *page;
+
+	page = brd_lookup_page(brd, sector);
+	if (page)
+		clear_highpage(page);
+}
+
 /*
  * Free all backing store pages and radix tree. This must only be called when
  * there are no other users of the device.
@@ -189,6 +211,24 @@ static int copy_to_brd_setup(struct brd_
 	return 0;
 }
 
+static void discard_from_brd(struct brd_device *brd,
+			sector_t sector, size_t n)
+{
+	while (n >= PAGE_SIZE) {
+		/*
+		 * Don't want to actually discard pages here because
+		 * re-allocating the pages can result in writeback
+		 * deadlocks under heavy load.
+		 */
+		if (0)
+			brd_free_page(brd, sector);
+		else
+			brd_zero_page(brd, sector);
+		sector += PAGE_SIZE >> SECTOR_SHIFT;
+		n -= PAGE_SIZE;
+	}
+}
+
 /*
  * Copy n bytes from src to the brd starting at sector. Does not sleep.
  */
@@ -300,6 +340,12 @@ static int brd_make_request(struct reque
 						get_capacity(bdev->bd_disk))
 		goto out;
 
+	if (unlikely(bio_rw_flagged(bio, BIO_RW_DISCARD))) {
+		err = 0;
+		discard_from_brd(brd, sector, bio->bi_size);
+		goto out;
+	}
+
 	rw = bio_rw(bio);
 	if (rw == READA)
 		rw = READ;
@@ -320,7 +366,7 @@ out:
 }
 
 #ifdef CONFIG_BLK_DEV_XIP
-static int brd_direct_access (struct block_device *bdev, sector_t sector,
+static int brd_direct_access(struct block_device *bdev, sector_t sector,
 			void **kaddr, unsigned long *pfn)
 {
 	struct brd_device *brd = bdev->bd_disk->private_data;
@@ -437,6 +483,11 @@ static struct brd_device *brd_alloc(int
 	blk_queue_max_hw_sectors(brd->brd_queue, 1024);
 	blk_queue_bounce_limit(brd->brd_queue, BLK_BOUNCE_ANY);
 
+	brd->brd_queue->limits.discard_granularity = PAGE_SIZE;
+	brd->brd_queue->limits.max_discard_sectors = UINT_MAX;
+	brd->brd_queue->limits.discard_zeroes_data = 1;
+	queue_flag_set_unlocked(QUEUE_FLAG_DISCARD, brd->brd_queue);
+
 	disk = brd->brd_disk = alloc_disk(1 << part_shift);
 	if (!disk)
 		goto out_free_queue;
#1 lore raw
Quoted 4 lines · Nick Piggin

Hi Jens,

What do you think of this patch? Can you possibly merge it via your block tree if it looks OK?

Sure, looks trivial enough (and correct, from a quick look through).

Signature · 1 line
Jens Axboe
#2 lore raw
Quoted 7 lines · Jens Axboe
Quoted 4 lines · Nick Piggin

Hi Jens,

What do you think of this patch? Can you possibly merge it via your block tree if it looks OK?

Sure, looks trivial enough (and correct, from a quick look through).

Thanks. It passes some basic testing (via blockdev discard ioctl and btrfs -odiscard mount) and it looks to be working properly.

#3 lore raw
In reply to Nick Piggin

drivers/block/brd.c static int copy_to_brd_setup(struct brd_ · line 226

Quoted 13 lines · Nick Piggin
drivers/block/brd.c +13 -0
+static void discard_from_brd(struct brd_device *brd,
+			sector_t sector, size_t n)
+{
+	while (n >= PAGE_SIZE) {
+		/*
+		 * Don't want to actually discard pages here because
+		 * re-allocating the pages can result in writeback
+		 * deadlocks under heavy load.
+		 */
+		if (0)
+			brd_free_page(brd, sector);
+		else
+			brd_zero_page(brd, sector);

That if (0) looks a little suspicious... ;-)

regards, Kyle