tmpfs: fix regression hang in fallocate undo
The well-spotted fallocate undo fix is good in most cases, but not when
fallocate failed on the very first page. index 0 then passes lend -1
to shmem_undo_range(), and that has two bad effects: (a) that it will
undo every fallocation throughout the file, unrestricted by the current
range; but more importantly (b) it can cause the undo to hang, because
lend -1 is treated as truncation, which makes it keep on retrying until
every page has gone, but those already fully instantiated will never go
away. Big thank you to xfstests generic/269 which demonstrates this.
Fixes: b9b4bb26af
("tmpfs: don't undo fallocate past its last page")
Cc: stable@vger.kernel.org
Signed-off-by: Hugh Dickins <hughd@google.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This commit is contained in:
parent
617a8d6bc1
commit
7f55656703
1 changed files with 5 additions and 3 deletions
|
@ -2225,9 +2225,11 @@ static long shmem_fallocate(struct file *file, int mode, loff_t offset,
|
||||||
error = shmem_getpage(inode, index, &page, SGP_FALLOC);
|
error = shmem_getpage(inode, index, &page, SGP_FALLOC);
|
||||||
if (error) {
|
if (error) {
|
||||||
/* Remove the !PageUptodate pages we added */
|
/* Remove the !PageUptodate pages we added */
|
||||||
shmem_undo_range(inode,
|
if (index > start) {
|
||||||
(loff_t)start << PAGE_SHIFT,
|
shmem_undo_range(inode,
|
||||||
((loff_t)index << PAGE_SHIFT) - 1, true);
|
(loff_t)start << PAGE_SHIFT,
|
||||||
|
((loff_t)index << PAGE_SHIFT) - 1, true);
|
||||||
|
}
|
||||||
goto undone;
|
goto undone;
|
||||||
}
|
}
|
||||||
|
|
||||||
|
|
Loading…
Reference in a new issue