Fix dio_bio_cleanup() to advance the head index into the list of pages past
the pages it has released, as __blockdev_direct_IO() will call it twice if
do_direct_IO() fails.
The issue was causing:
WARNING: CPU: 6 PID: 2220 at mm/gup.c:76 try_get_folio
This can be triggered by setting up a clean pair of UDF filesystems on
loopback devices and running the generic/451 xfstest with them as the
scratch and test partitions. Something like the following:
fallocate /mnt2/udf_scratch -l 1G
fallocate /mnt2/udf_test -l 1G
mknod /dev/lo0 b 7 0
mknod /dev/lo1 b 7 1
losetup lo0 /mnt2/udf_scratch
losetup lo1 /mnt2/udf_test
mkfs -t udf /dev/lo0
mkfs -t udf /dev/lo1
cd xfstests
./check generic/451
with xfstests configured by putting the following into local.config:
export FSTYP=udf
export DISABLE_UDF_TEST=1
export TEST_DEV=/dev/lo1
export TEST_DIR=/xfstest.test
export SCRATCH_DEV=/dev/lo0
export SCRATCH_MNT=/xfstest.scratch
Fixes: 1ccf164ec866 ("block: Use iov_iter_extract_pages() and page pinning in direct-io.c")
Reported-by: kernel test robot <[email protected]>
Closes: https://lore.kernel.org/oe-lkp/[email protected]
Signed-off-by: David Howells <[email protected]>
cc: Christoph Hellwig <[email protected]>
cc: David Hildenbrand <[email protected]>
cc: Andrew Morton <[email protected]>
cc: Jens Axboe <[email protected]>
cc: Al Viro <[email protected]>
cc: Matthew Wilcox <[email protected]>
cc: Jan Kara <[email protected]>
cc: Jeff Layton <[email protected]>
cc: Jason Gunthorpe <[email protected]>
cc: Logan Gunthorpe <[email protected]>
cc: Hillf Danton <[email protected]>
cc: Christian Brauner <[email protected]>
cc: Linus Torvalds <[email protected]>
cc: [email protected]
cc: [email protected]
cc: [email protected]
cc: [email protected]
---
fs/direct-io.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/fs/direct-io.c b/fs/direct-io.c
index 0643f1bb4b59..2ceb378b93c0 100644
--- a/fs/direct-io.c
+++ b/fs/direct-io.c
@@ -459,6 +459,7 @@ static inline void dio_cleanup(struct dio *dio, struct dio_submit *sdio)
if (dio->is_pinned)
unpin_user_pages(dio->pages + sdio->head,
sdio->tail - sdio->head);
+ sdio->head = sdio->tail;
}
/*
On Tue, Jun 13, 2023 at 10:54:39PM +0100, David Howells wrote:
> --- a/fs/direct-io.c
> +++ b/fs/direct-io.c
> @@ -459,6 +459,7 @@ static inline void dio_cleanup(struct dio *dio, struct dio_submit *sdio)
> if (dio->is_pinned)
> unpin_user_pages(dio->pages + sdio->head,
> sdio->tail - sdio->head);
> + sdio->head = sdio->tail;
So looking at the original patch, it does:
- while (sdio->head < sdio->tail)
- put_page(dio->pages[sdio->head++]);
+ if (dio->is_pinned)
+ unpin_user_pages(dio->pages + sdio->head,
+ sdio->tail - sdio->head);
so yes, we're this looks correct:
Reviewed-by: Christoph Hellwig <[email protected]>
On 13.06.23 23:54, David Howells wrote:
>
> Fix dio_bio_cleanup() to advance the head index into the list of pages past
> the pages it has released, as __blockdev_direct_IO() will call it twice if
> do_direct_IO() fails.
>
> The issue was causing:
>
> WARNING: CPU: 6 PID: 2220 at mm/gup.c:76 try_get_folio
>
> This can be triggered by setting up a clean pair of UDF filesystems on
> loopback devices and running the generic/451 xfstest with them as the
> scratch and test partitions. Something like the following:
>
> fallocate /mnt2/udf_scratch -l 1G
> fallocate /mnt2/udf_test -l 1G
> mknod /dev/lo0 b 7 0
> mknod /dev/lo1 b 7 1
> losetup lo0 /mnt2/udf_scratch
> losetup lo1 /mnt2/udf_test
> mkfs -t udf /dev/lo0
> mkfs -t udf /dev/lo1
> cd xfstests
> ./check generic/451
>
> with xfstests configured by putting the following into local.config:
>
> export FSTYP=udf
> export DISABLE_UDF_TEST=1
> export TEST_DEV=/dev/lo1
> export TEST_DIR=/xfstest.test
> export SCRATCH_DEV=/dev/lo0
> export SCRATCH_MNT=/xfstest.scratch
>
> Fixes: 1ccf164ec866 ("block: Use iov_iter_extract_pages() and page pinning in direct-io.c")
> Reported-by: kernel test robot <[email protected]>
> Closes: https://lore.kernel.org/oe-lkp/[email protected]
> Signed-off-by: David Howells <[email protected]>
> cc: Christoph Hellwig <[email protected]>
> cc: David Hildenbrand <[email protected]>
> cc: Andrew Morton <[email protected]>
> cc: Jens Axboe <[email protected]>
> cc: Al Viro <[email protected]>
> cc: Matthew Wilcox <[email protected]>
> cc: Jan Kara <[email protected]>
> cc: Jeff Layton <[email protected]>
> cc: Jason Gunthorpe <[email protected]>
> cc: Logan Gunthorpe <[email protected]>
> cc: Hillf Danton <[email protected]>
> cc: Christian Brauner <[email protected]>
> cc: Linus Torvalds <[email protected]>
> cc: [email protected]
> cc: [email protected]
> cc: [email protected]
> cc: [email protected]
> ---
> fs/direct-io.c | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/fs/direct-io.c b/fs/direct-io.c
> index 0643f1bb4b59..2ceb378b93c0 100644
> --- a/fs/direct-io.c
> +++ b/fs/direct-io.c
> @@ -459,6 +459,7 @@ static inline void dio_cleanup(struct dio *dio, struct dio_submit *sdio)
> if (dio->is_pinned)
> unpin_user_pages(dio->pages + sdio->head,
> sdio->tail - sdio->head);
> + sdio->head = sdio->tail;
> }
>
> /*
>
Reviewed-by: David Hildenbrand <[email protected]>
--
Cheers,
David / dhildenb
On Tue, 13 Jun 2023 22:54:39 +0100, David Howells wrote:
>
> Fix dio_bio_cleanup() to advance the head index into the list of pages past
> the pages it has released, as __blockdev_direct_IO() will call it twice if
> do_direct_IO() fails.
>
> The issue was causing:
>
> [...]
Applied, thanks!
[1/1] block: Fix dio_cleanup() to advance the head index
commit: d44c404207831dfe3b301ff479e964b77914488b
Best regards,
--
Jens Axboe