Skip to contentSkip to conversation

xfs: drop extra transaction roll from inode extent truncate

PATCH In review Brian Foster 2 participants 2 messages Sep 10, 2020 – Sep 16, 2020 XFS filesystem

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 xfs: drop extra transaction roll from inode extent truncate

The inode extent truncate path unmaps extents from the inode block mapping, finishes deferred ops to free the associated extents and then explicitly rolls the transaction before processing the next extent. The latter extent roll is spurious as xfs_defer_finish() always returns a clean transaction and automatically relogs inodes attached to the transaction (with lock_flags == 0). This can unnecessarily increase the number of log ticket regrants that occur during a long running truncate operation. Remove the explicit transaction roll.

---

Just something I noticed when reading through the code based on Dave's recent EFI recovery reservation patches..

Brian

 fs/xfs/xfs_inode.c | 9 +--------
 1 file changed, 1 insertion(+), 8 deletions(-)
fs/xfs/xfs_inode.c +1 -8
index c06129cffba9..7af99c7a2821 100644
@@ -1532,17 +1532,10 @@ xfs_itruncate_extents_flags(
 		if (error)
 			goto out;
 
-		/*
-		 * Duplicate the transaction that has the permanent
-		 * reservation and commit the old transaction.
-		 */
+		/* free the just unmapped extents */
 		error = xfs_defer_finish(&tp);
 		if (error)
 			goto out;
-
-		error = xfs_trans_roll_inode(&tp, ip);
-		if (error)
-			goto out;
 	}
 
 	if (whichfork == XFS_DATA_FORK) {
Signature · 1 line
2.25.4
Reviewed-by on v1 1/1
#1 lore raw
Quoted 11 lines · Brian Foster

The inode extent truncate path unmaps extents from the inode block mapping, finishes deferred ops to free the associated extents and then explicitly rolls the transaction before processing the next extent. The latter extent roll is spurious as xfs_defer_finish() always returns a clean transaction and automatically relogs inodes attached to the transaction (with lock_flags == 0). This can unnecessarily increase the number of log ticket regrants that occur during a long running truncate operation. Remove the explicit transaction roll.

Looks ok to me,

--D

Quoted 36 lines

---

Just something I noticed when reading through the code based on Dave's recent EFI recovery reservation patches..

Brian

 fs/xfs/xfs_inode.c | 9 +--------
 1 file changed, 1 insertion(+), 8 deletions(-)
fs/xfs/xfs_inode.c +1 -8
index c06129cffba9..7af99c7a2821 100644
@@ -1532,17 +1532,10 @@ xfs_itruncate_extents_flags(
 		if (error)
 			goto out;
 
-		/*
-		 * Duplicate the transaction that has the permanent
-		 * reservation and commit the old transaction.
-		 */
+		/* free the just unmapped extents */
 		error = xfs_defer_finish(&tp);
 		if (error)
 			goto out;
-
-		error = xfs_trans_roll_inode(&tp, ip);
-		if (error)
-			goto out;
 	}
 
 	if (whichfork == XFS_DATA_FORK) {
Signature · 1 line
2.25.4