The previous attempt to fix for metadata read-ahead during truncate was
incorrect: for files with a height > 2 (1006989312 bytes with a block
size of 4096 bytes), read-ahead requests were not being issued for some
of the indirect blocks discovered while walking the metadata tree,
leading to significant slow-downs when deleting large files. Fix that.
In addition, only issue read-ahead requests in the first pass through
the meta-data tree, while deallocating data blocks.
Fixes: c3ce5aa9b0 ("gfs2: Fix metadata read-ahead during truncate")
Cc: stable@xxxxxxxxxxxxxxx # v4.16+
Signed-off-by: Andreas Gruenbacher <agruenba@xxxxxxxxxx>
---
fs/gfs2/bmap.c | 14 ++++++++++----
1 file changed, 10 insertions(+), 4 deletions(-)
diff --git a/fs/gfs2/bmap.c b/fs/gfs2/bmap.c
index 5f3ea07ef5e2..38d88fcb6988 100644
--- a/fs/gfs2/bmap.c
+++ b/fs/gfs2/bmap.c
@@ -1908,10 +1908,16 @@ static int punch_hole(struct gfs2_inode *ip, u64 offset, u64 length)
if (ret < 0)
goto out;
- /* issue read-ahead on metadata */
- if (mp.mp_aheight > 1) {
- for (; ret > 1; ret--) {
- metapointer_range(&mp, mp.mp_aheight - ret,
+ /* On the first pass, issue read-ahead on metadata. */
+ if (mp.mp_aheight > 1 && strip_h == ip->i_height - 1) {
+ unsigned int height = mp.mp_aheight - 1;
+
+ /* No read-ahead for data blocks. */
+ if (mp.mp_aheight - 1 == strip_h)
+ height--;
+
+ for (; height >= mp.mp_aheight - ret; height--) {
+ metapointer_range(&mp, height,
start_list, start_aligned,
end_list, end_aligned,
&start, &end);
--
2.19.1.546.g028f9c799.dirty
