dm thin metadata: try to avoid ever aborting transactions

Committing a transaction can consume some metadata of it's own, we now
reserve a small amount of metadata to cover this.  Free metadata
reported by the kernel will not include this reserve.

If any of the reserve has been used after a commit we enter a new
internal state PM_OUT_OF_METADATA_SPACE.  This is reported as
PM_READ_ONLY, so no userland changes are needed.  If the metadata
device is resized the pool will move back to PM_WRITE.

These changes mean we never need to abort and rollback a transaction due
to running out of metadata space.  This is particularly important
because there have been a handful of reports of data corruption against
DM thin-provisioning that can all be attributed to the thin-pool having
ran out of metadata space.

Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>

1 files changed, 35 insertions, 1 deletions

diff --git a/drivers/md/dm-thin-metadata.c b/drivers/md/dm-thin-metadata.c
index 72142021b5c9..74f6770c70b1 100644
--- a/drivers/md/dm-thin-metadata.c
+++ b/drivers/md/dm-thin-metadata.c
@@ -189,6 +189,12 @@ struct dm_pool_metadata {
 	sector_t data_block_size;
 
 	/*
+	 * We reserve a section of the metadata for commit overhead.
+	 * All reported space does *not* include this.
+	 */
+	dm_block_t metadata_reserve;
+
+	/*
 	 * Set if a transaction has to be aborted but the attempt to roll back
 	 * to the previous (good) transaction failed.  The only pool metadata
 	 * operation possible in this state is the closing of the device.
@@ -816,6 +822,22 @@ static int __commit_transaction(struct dm_pool_metadata *pmd)
 	return dm_tm_commit(pmd->tm, sblock);
 }
 
+static void __set_metadata_reserve(struct dm_pool_metadata *pmd)
+{
+	int r;
+	dm_block_t total;
+	dm_block_t max_blocks = 4096; /* 16M */
+
+	r = dm_sm_get_nr_blocks(pmd->metadata_sm, &total);
+	if (r) {
+		DMERR("could not get size of metadata device");
+		pmd->metadata_reserve = max_blocks;
+	} else {
+		sector_div(total, 10);
+		pmd->metadata_reserve = min(max_blocks, total);
+	}
+}
+
 struct dm_pool_metadata *dm_pool_metadata_open(struct block_device *bdev,
 					       sector_t data_block_size,
 					       bool format_device)
@@ -849,6 +871,8 @@ struct dm_pool_metadata *dm_pool_metadata_open(struct block_device *bdev,
 		return ERR_PTR(r);
 	}
 
+	__set_metadata_reserve(pmd);
+
 	return pmd;
 }
 
@@ -1820,6 +1844,13 @@ int dm_pool_get_free_metadata_block_count(struct dm_pool_metadata *pmd,
 	down_read(&pmd->root_lock);
 	if (!pmd->fail_io)
 		r = dm_sm_get_nr_free(pmd->metadata_sm, result);
+
+	if (!r) {
+		if (*result < pmd->metadata_reserve)
+			*result = 0;
+		else
+			*result -= pmd->metadata_reserve;
+	}
 	up_read(&pmd->root_lock);
 
 	return r;
@@ -1932,8 +1963,11 @@ int dm_pool_resize_metadata_dev(struct dm_pool_metadata *pmd, dm_block_t new_cou
 	int r = -EINVAL;
 
 	down_write(&pmd->root_lock);
-	if (!pmd->fail_io)
+	if (!pmd->fail_io) {
 		r = __resize_space_map(pmd->metadata_sm, new_count);
+		if (!r)
+			__set_metadata_reserve(pmd);
+	}
 	up_write(&pmd->root_lock);
 
 	return r;