dm thin: ensure user takes action to validate data and metadata consi…

…stency

If a thin metadata operation fails the current transaction will abort,
whereby causing potential for IO layers up the stack (e.g. filesystems)
to have data loss.  As such, set THIN_METADATA_NEEDS_CHECK_FLAG in the
thin metadata's superblock which:
1) requires the user verify the thin metadata is consistent (e.g. use
   thin_check, etc)
2) suggests the user verify the thin data is consistent (e.g. use fsck)

The only way to clear the superblock's THIN_METADATA_NEEDS_CHECK_FLAG is
to run thin_repair.

On metadata operation failure: abort current metadata transaction, set
pool in read-only mode, and now set the needs_check flag.

As part of this change, constraints are introduced or relaxed:
* don't allow a pool to transition to write mode if needs_check is set
* don't allow data or metadata space to be resized if needs_check is set
* if a thin pool's metadata space is exhausted: the kernel will now
  force the user to take the pool offline for repair before the kernel
  will allow the metadata space to be extended.

Also, update Documentation to include information about when the thin
provisioning target commits metadata, how it handles metadata failures
and running out of space.

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

Documentation/device-mapper/cache.txt

@@ -124,12 +124,11 @@ the default being 204800 sectors (or 100MB).
 Updating on-disk metadata
 -------------------------
 
-On-disk metadata is committed every time a REQ_SYNC or REQ_FUA bio is
-written.  If no such requests are made then commits will occur every
-second.  This means the cache behaves like a physical disk that has a
-write cache (the same is true of the thin-provisioning target).  If
-power is lost you may lose some recent writes.  The metadata should
-always be consistent in spite of any crash.
+On-disk metadata is committed every time a FLUSH or FUA bio is written.
+If no such requests are made then commits will occur every second.  This
+means the cache behaves like a physical disk that has a volatile write
+cache.  If power is lost you may lose some recent writes.  The metadata
+should always be consistent in spite of any crash.
 
 The 'dirty' state for a cache block changes far too frequently for us
 to keep updating it on the fly.  So we treat it as a hint.  In normal

Documentation/device-mapper/thin-provisioning.txt

@@ -116,6 +116,35 @@ Resuming a device with a new table itself triggers an event so the
 userspace daemon can use this to detect a situation where a new table
 already exceeds the threshold.
 
+A low water mark for the metadata device is maintained in the kernel and
+will trigger a dm event if free space on the metadata device drops below
+it.
+
+Updating on-disk metadata
+-------------------------
+
+On-disk metadata is committed every time a FLUSH or FUA bio is written.
+If no such requests are made then commits will occur every second.  This
+means the thin-provisioning target behaves like a physical disk that has
+a volatile write cache.  If power is lost you may lose some recent
+writes.  The metadata should always be consistent in spite of any crash.
+
+If data space is exhausted the pool will either error or queue IO
+according to the configuration (see: error_if_no_space).  If metadata
+space is exhausted or a metadata operation fails: the pool will error IO
+until the pool is taken offline and repair is performed to 1) fix any
+potential inconsistencies and 2) clear the flag that imposes repair.
+Once the pool's metadata device is repaired it may be resized, which
+will allow the pool to return to normal operation.  Note that if a pool
+is flagged as needing repair, the pool's data and metadata devices
+cannot be resized until repair is performed.  It should also be noted
+that when the pool's metadata space is exhausted the current metadata
+transaction is aborted.  Given that the pool will cache IO whose
+completion may have already been acknowledged to upper IO layers
+(e.g. filesystem) it is strongly suggested that consistency checks
+(e.g. fsck) be performed on those layers when repair of the pool is
+required.
+
 Thin provisioning
 -----------------
 

drivers/md/dm-thin-metadata.c

@@ -76,7 +76,7 @@
 
 #define THIN_SUPERBLOCK_MAGIC 27022010
 #define THIN_SUPERBLOCK_LOCATION 0
-#define THIN_VERSION 1
+#define THIN_VERSION 2
 #define THIN_METADATA_CACHE_SIZE 64
 #define SECTOR_TO_BLOCK_SHIFT 3
 
@@ -1755,3 +1755,38 @@ int dm_pool_register_metadata_threshold(struct dm_pool_metadata *pmd,
 
 	return r;
 }
+
+int dm_pool_metadata_set_needs_check(struct dm_pool_metadata *pmd)
+{
+	int r;
+	struct dm_block *sblock;
+	struct thin_disk_superblock *disk_super;
+
+	down_write(&pmd->root_lock);
+	pmd->flags |= THIN_METADATA_NEEDS_CHECK_FLAG;
+
+	r = superblock_lock(pmd, &sblock);
+	if (r) {
+		DMERR("couldn't read superblock");
+		goto out;
+	}
+
+	disk_super = dm_block_data(sblock);
+	disk_super->flags = cpu_to_le32(pmd->flags);
+
+	dm_bm_unlock(sblock);
+out:
+	up_write(&pmd->root_lock);
+	return r;
+}
+
+bool dm_pool_metadata_needs_check(struct dm_pool_metadata *pmd)
+{
+	bool needs_check;
+
+	down_read(&pmd->root_lock);
+	needs_check = pmd->flags & THIN_METADATA_NEEDS_CHECK_FLAG;
+	up_read(&pmd->root_lock);
+
+	return needs_check;
+}

drivers/md/dm-thin-metadata.h

@@ -25,6 +25,11 @@
 
 /*----------------------------------------------------------------*/
 
+/*
+ * Thin metadata superblock flags.
+ */
+#define THIN_METADATA_NEEDS_CHECK_FLAG (1 << 0)
+
 struct dm_pool_metadata;
 struct dm_thin_device;
 
@@ -202,6 +207,12 @@ int dm_pool_register_metadata_threshold(struct dm_pool_metadata *pmd,
 					dm_sm_threshold_fn fn,
 					void *context);
 
+/*
+ * Updates the superblock immediately.
+ */
+int dm_pool_metadata_set_needs_check(struct dm_pool_metadata *pmd);
+bool dm_pool_metadata_needs_check(struct dm_pool_metadata *pmd);
+
 /*----------------------------------------------------------------*/
 
 #endif

drivers/md/dm-thin.c

@@ -1403,7 +1403,28 @@ static void set_pool_mode(struct pool *pool, enum pool_mode new_mode)
 {
 	int r;
 	struct pool_c *pt = pool->ti->private;
-	enum pool_mode old_mode = pool->pf.mode;
+	bool needs_check = dm_pool_metadata_needs_check(pool->pmd);
+	enum pool_mode old_mode = get_pool_mode(pool);
+
+	/*
+	 * Never allow the pool to transition to PM_WRITE mode if user
+	 * intervention is required to verify metadata and data consistency.
+	 */
+	if (new_mode == PM_WRITE && needs_check) {
+		DMERR("%s: unable to switch pool to write mode until repaired.",
+		      dm_device_name(pool->pool_md));
+		if (old_mode != new_mode)
+			new_mode = old_mode;
+		else
+			new_mode = PM_READ_ONLY;
+	}
+	/*
+	 * If we were in PM_FAIL mode, rollback of metadata failed.  We're
+	 * not going to recover without a thin_repair.	So we never let the
+	 * pool move out of the old mode.
+	 */
+	if (old_mode == PM_FAIL)
+		new_mode = old_mode;
 
 	switch (new_mode) {
 	case PM_FAIL:
@@ -1467,19 +1488,28 @@ static void out_of_data_space(struct pool *pool)
 	set_pool_mode(pool, PM_READ_ONLY);
 }
 
-static void metadata_operation_failed(struct pool *pool, const char *op, int r)
+static void abort_transaction(struct pool *pool)
 {
-	dm_block_t free_blocks;
+	const char *dev_name = dm_device_name(pool->pool_md);
+
+	DMERR_LIMIT("%s: aborting current metadata transaction", dev_name);
+	if (dm_pool_abort_metadata(pool->pmd)) {
+		DMERR("%s: failed to abort metadata transaction", dev_name);
+		set_pool_mode(pool, PM_FAIL);
+	}
+
+	if (dm_pool_metadata_set_needs_check(pool->pmd)) {
+		DMERR("%s: failed to set 'needs_check' flag in metadata", dev_name);
+		set_pool_mode(pool, PM_FAIL);
+	}
+}
 
+static void metadata_operation_failed(struct pool *pool, const char *op, int r)
+{
 	DMERR_LIMIT("%s: metadata operation '%s' failed: error = %d",
 		    dm_device_name(pool->pool_md), op, r);
 
-	if (r == -ENOSPC &&
-	    !dm_pool_get_free_metadata_block_count(pool->pmd, &free_blocks) &&
-	    !free_blocks)
-		DMERR_LIMIT("%s: no free metadata space available.",
-			    dm_device_name(pool->pool_md));
-
+	abort_transaction(pool);
 	set_pool_mode(pool, PM_READ_ONLY);
 }
 
@@ -1693,7 +1723,7 @@ static int bind_control_target(struct pool *pool, struct dm_target *ti)
 	/*
 	 * We want to make sure that a pool in PM_FAIL mode is never upgraded.
 	 */
-	enum pool_mode old_mode = pool->pf.mode;
+	enum pool_mode old_mode = get_pool_mode(pool);
 	enum pool_mode new_mode = pt->adjusted_pf.mode;
 
 	/*
@@ -1707,16 +1737,6 @@ static int bind_control_target(struct pool *pool, struct dm_target *ti)
 	pool->pf = pt->adjusted_pf;
 	pool->low_water_blocks = pt->low_water_blocks;
 
-	/*
-	 * If we were in PM_FAIL mode, rollback of metadata failed.  We're
-	 * not going to recover without a thin_repair.  So we never let the
-	 * pool move out of the old mode.  On the other hand a PM_READ_ONLY
-	 * may have been due to a lack of metadata or data space, and may
-	 * now work (ie. if the underlying devices have been resized).
-	 */
-	if (old_mode == PM_FAIL)
-		new_mode = old_mode;
-
 	set_pool_mode(pool, new_mode);
 
 	return 0;
@@ -2259,6 +2279,12 @@ static int maybe_resize_data_dev(struct dm_target *ti, bool *need_commit)
 		return -EINVAL;
 
 	} else if (data_size > sb_data_size) {
+		if (dm_pool_metadata_needs_check(pool->pmd)) {
+			DMERR("%s: unable to grow the data device until repaired.",
+			      dm_device_name(pool->pool_md));
+			return 0;
+		}
+
 		if (sb_data_size)
 			DMINFO("%s: growing the data device from %llu to %llu blocks",
 			       dm_device_name(pool->pool_md),
@@ -2300,6 +2326,12 @@ static int maybe_resize_metadata_dev(struct dm_target *ti, bool *need_commit)
 		return -EINVAL;
 
 	} else if (metadata_dev_size > sb_metadata_dev_size) {
+		if (dm_pool_metadata_needs_check(pool->pmd)) {
+			DMERR("%s: unable to grow the metadata device until repaired.",
+			      dm_device_name(pool->pool_md));
+			return 0;
+		}
+
 		warn_if_metadata_device_too_big(pool->md_dev);
 		DMINFO("%s: growing the metadata device from %llu to %llu blocks",
 		       dm_device_name(pool->pool_md),
@@ -2801,7 +2833,7 @@ static struct target_type pool_target = {
 	.name = "thin-pool",
 	.features = DM_TARGET_SINGLETON | DM_TARGET_ALWAYS_WRITEABLE |
 		    DM_TARGET_IMMUTABLE,
-	.version = {1, 10, 0},
+	.version = {1, 11, 0},
 	.module = THIS_MODULE,
 	.ctr = pool_ctr,
 	.dtr = pool_dtr,
@@ -3091,7 +3123,7 @@ static int thin_iterate_devices(struct dm_target *ti,
 
 static struct target_type thin_target = {
 	.name = "thin",
-	.version = {1, 10, 0},
+	.version = {1, 11, 0},
 	.module	= THIS_MODULE,
 	.ctr = thin_ctr,
 	.dtr = thin_dtr,