[dm-devel] [PATCH 2/3] dm-thin: fix discard support
Mikulas Patocka
mpatocka at redhat.com
Mon Jul 16 18:35:18 UTC 2012
dm-thin: fix discard support
There is a bug in dm_thin regarding processing discards.
When dm-thin receives a discard request with size equal to block size
that is not aligned on block size boundary, io_overlaps_block returns
true, process_discard treats this discard as a full block discard,
deletes the full block - the result is that some data that shouldn't be
discarded are discarded.
This patch sets the variable "ti->split_discard_requests", so that
device mapper core splits discard requests on a block boundary.
Consequently, a discard request that spans multiple blocks is never sent
to dm-thin. The patch also removes some code in process_discard that
deals with discards that span multiple blocks.
Signed-off-by: Mikulas Patocka <mpatocka at redhat.com>
---
drivers/md/dm-thin.c | 18 +++++++-----------
1 file changed, 7 insertions(+), 11 deletions(-)
Index: linux-3.5-rc6-fast/drivers/md/dm-thin.c
===================================================================
--- linux-3.5-rc6-fast.orig/drivers/md/dm-thin.c 2012-07-16 18:46:18.000000000 +0200
+++ linux-3.5-rc6-fast/drivers/md/dm-thin.c 2012-07-16 20:07:19.000000000 +0200
@@ -1246,17 +1246,10 @@ static void process_discard(struct thin_
}
} else {
/*
- * This path is hit if people are ignoring
- * limits->discard_granularity. It ignores any
- * part of the discard that is in a subsequent
- * block.
+ * The dm makes sure that the discard doesn't span
+ * a block boundary. So we submit the discard
+ * to the appropriate block.
*/
- sector_t offset = pool->sectors_per_block_shift >= 0 ?
- bio->bi_sector & (pool->sectors_per_block - 1) :
- bio->bi_sector - block * pool->sectors_per_block;
- unsigned remaining = (pool->sectors_per_block - offset) << SECTOR_SHIFT;
- bio->bi_size = min(bio->bi_size, remaining);
-
cell_release_singleton(cell, bio);
cell_release_singleton(cell2, bio);
remap_and_issue(tc, bio, lookup_result.block);
@@ -2506,7 +2499,8 @@ static void set_discard_limits(struct po
/*
* This is just a hint, and not enforced. We have to cope with
- * bios that overlap 2 blocks.
+ * bios cover a block partially. A discard that spans a block boundary
+ * is not sent to this target.
*/
limits->discard_granularity = pool->sectors_per_block << SECTOR_SHIFT;
limits->discard_zeroes_data = pool->pf.zero_new_blocks;
@@ -2648,6 +2642,8 @@ static int thin_ctr(struct dm_target *ti
if (tc->pool->pf.discard_enabled) {
ti->discards_supported = 1;
ti->num_discard_requests = 1;
+ /* Discard requests must be split on a chunk boundary */
+ ti->split_discard_requests = 1;
}
dm_put(pool_md);
More information about the dm-devel
mailing list