3.4-stable review patch.  If anyone has any objections, please let me know.

------------------

From: Joe Thornber <redacted>

commit 12c91a5c2d2a8e8cc40a9552313e1e7b0a2d9ee3 upstream.

When extending a low level space map we should update nr_blocks at
the start so the new space is used for the index entries.

Otherwise extend can fail, e.g.: sm_metadata_extend call sequence
that fails:
 -> sm_ll_extend
    -> dm_tm_new_block -> dm_sm_new_block -> sm_bootstrap_new_block
    => returns -ENOSPC because smm->begin == smm->ll.nr_blocks

Signed-off-by: Joe Thornber <redacted>
Signed-off-by: Mike Snitzer <redacted>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

---
 drivers/md/persistent-data/dm-space-map-common.c |    6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

--- a/drivers/md/persistent-data/dm-space-map-common.c
+++ b/drivers/md/persistent-data/dm-space-map-common.c

@@ -244,6 +244,10 @@ int sm_ll_extend(struct ll_disk *ll, dm_
 		return -EINVAL;
 	}
 
+	/*
+	 * We need to set this before the dm_tm_new_block() call below.
+	 */
+	ll->nr_blocks = nr_blocks;
 	for (i = old_blocks; i < blocks; i++) {
 		struct dm_block *b;
 		struct disk_index_entry idx;

@@ -251,6 +255,7 @@ int sm_ll_extend(struct ll_disk *ll, dm_
 		r = dm_tm_new_block(ll->tm, &dm_sm_bitmap_validator, &b);
 		if (r < 0)
 			return r;
+
 		idx.blocknr = cpu_to_le64(dm_block_location(b));
 
 		r = dm_tm_unlock(ll->tm, b);

@@ -265,7 +270,6 @@ int sm_ll_extend(struct ll_disk *ll, dm_
 			return r;
 	}
 
-	ll->nr_blocks = nr_blocks;
 	return 0;
 }