Skip to contentSkip to conversation

block/file-posix: Truncate in xfs_write_zeroes()

PATCH Applied Max Reitz 2 participants 2 messages May 10, 2019 – May 13, 2019 Block layer

lore
Share & tools
lore

Max Reitz

author
#0 lore raw

PATCH 1/1 block/file-posix: Truncate in xfs_write_zeroes()

XFS_IOC_ZERO_RANGE does not increase the file length:

$ touch foo
$ xfs_io -c 'zero 0 65536' foo
$ stat -c "size=%s, blocks=%b" foo
size=0, blocks=128

We do want writes beyond the EOF to automatically increase the file length, however. This is evidenced by the fact that iotest 061 is broken on XFS since qcow2's check implementation checks for blocks beyond the EOF.

---
 block/file-posix.c | 13 +++++++++++++
 1 file changed, 13 insertions(+)
block/file-posix.c +13 -0
index 1cf4ee49eb..e09e15bbf8 100644
@@ -1444,9 +1444,22 @@ out:
 #ifdef CONFIG_XFS
 static int xfs_write_zeroes(BDRVRawState *s, int64_t offset, uint64_t bytes)
 {
+    int64_t len;
     struct xfs_flock64 fl;
     int err;
 
+    len = lseek(s->fd, 0, SEEK_END);
+    if (len < 0) {
+        return -errno;
+    }
+
+    if (offset + bytes > len) {
+        /* XFS_IOC_ZERO_RANGE does not increase the file length */
+        if (ftruncate(s->fd, offset + bytes) < 0) {
+            return -errno;
+        }
+    }
+
     memset(&fl, 0, sizeof(fl));
     fl.l_whence = SEEK_SET;
     fl.l_start = offset;
Signature · 1 line
2.21.0
Applied
#1 lore raw

Am 10.05.2019 um 23:12 hat Max Reitz geschrieben:

Quoted 6 lines

XFS_IOC_ZERO_RANGE does not increase the file length:

$ touch foo
$ xfs_io -c 'zero 0 65536' foo
$ stat -c "size=%s, blocks=%b" foo
size=0, blocks=128
Quoted 7 lines

We do want writes beyond the EOF to automatically increase the file length, however. This is evidenced by the fact that iotest 061 is broken on XFS since qcow2's check implementation checks for blocks beyond the EOF.

Just for the record, the commit that made the problem visible in 061 is commit a5fff8d4.

Thanks, applied to the block branch.

Though I wonder if we should prefer FALLOC_FL_ZERO_RANGE now if available, which is a single syscall and consistent for all filesystems.

Kevin