From: Dave Chinner <dchinner@xxxxxxxxxx>
xfs_alloc_file_space ends up in an endless loop when
xfs_bmapi_write() returns nimaps == 0 at ENOSPC. The process is
unkillable, and so just runs around in a tight circle burning CPU
until the system is rebooted.
This is a regression introduced by commit 35dc55b9e80c ("xfs: handle
nimaps=0 from xfs_bmapi_write in xfs_alloc_file_space") which
specifically removed ENOSPC detection from xfs_alloc_file_space()
and replaces it with an endless loop. This attempts to fix an issue
converting a delalloc extent when not enough contiguous free space
is available to convert the entire delalloc extent.
Right now just revert the change as it only manifested on code under
development and isn't currently a real-world problem.
Fixes: 35dc55b9e80c ("xfs: handle nimaps=0 from xfs_bmapi_write in xfs_alloc_file_space")
Signed-off-by: Dave Chinner <dchinner@xxxxxxxxxx>
---
fs/xfs/xfs_bmap_util.c | 18 ++++++------------
1 file changed, 6 insertions(+), 12 deletions(-)
diff --git a/fs/xfs/xfs_bmap_util.c b/fs/xfs/xfs_bmap_util.c
index 19e11d1da660..262557735d4d 100644
--- a/fs/xfs/xfs_bmap_util.c
+++ b/fs/xfs/xfs_bmap_util.c
@@ -735,19 +735,13 @@ xfs_alloc_file_space(
if (error)
break;
- /*
- * If the allocator cannot find a single free extent large
- * enough to cover the start block of the requested range,
- * xfs_bmapi_write will return 0 but leave *nimaps set to 0.
- *
- * In that case we simply need to keep looping with the same
- * startoffset_fsb so that one of the following allocations
- * will eventually reach the requested range.
- */
- if (nimaps) {
- startoffset_fsb += imapp->br_blockcount;
- allocatesize_fsb -= imapp->br_blockcount;
+ if (nimaps == 0) {
+ error = ENOSPC;
+ break;
}
+
+ startoffset_fsb += imapp->br_blockcount;
+ allocatesize_fsb -= imapp->br_blockcount;
}
return error;
--
2.43.0
