2022-01-06 03:31:57

by Qinghua Jin

[permalink] [raw]
Subject: [PATCH] minix: fix bug when opening a file with O_DIRECT will

Testcase:
1. create a minix file system and mount it
2. open a file on the file system with O_RDWR|O_CREAT|O_TRUNC|O_DIRECT
3. open fails with -EINVAL but leaves an empty file behind. All other
open() failures don't leave the failed open files behind.

It is hard to check the direct_IO op before creating the inode. Just as
ext4 and btrfs do, this patch will resolve the issue by allowing to
create the file with O_DIRECT but returning error when writing the file.

Signed-off-by: Qinghua Jin <[email protected]>
---
fs/minix/inode.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/fs/minix/inode.c b/fs/minix/inode.c
index a71f1cf894b9..d4bd94234ef7 100644
--- a/fs/minix/inode.c
+++ b/fs/minix/inode.c
@@ -447,7 +447,8 @@ static const struct address_space_operations minix_aops = {
.writepage = minix_writepage,
.write_begin = minix_write_begin,
.write_end = generic_write_end,
- .bmap = minix_bmap
+ .bmap = minix_bmap,
+ .direct_IO = noop_direct_IO
};

static const struct inode_operations minix_symlink_inode_operations = {
--
2.30.2



2022-01-06 15:43:05

by Jan Kara

[permalink] [raw]
Subject: Re: [PATCH] minix: fix bug when opening a file with O_DIRECT will

On Thu 06-01-22 11:30:41, Qinghua Jin wrote:
> Testcase:
> 1. create a minix file system and mount it
> 2. open a file on the file system with O_RDWR|O_CREAT|O_TRUNC|O_DIRECT
> 3. open fails with -EINVAL but leaves an empty file behind. All other
> open() failures don't leave the failed open files behind.
>
> It is hard to check the direct_IO op before creating the inode. Just as
> ext4 and btrfs do, this patch will resolve the issue by allowing to
> create the file with O_DIRECT but returning error when writing the file.
>
> Signed-off-by: Qinghua Jin <[email protected]>

The patch looks good. Thanks. Feel free to add:

Reviewed-by: Jan Kara <[email protected]>

Just out of curiosity: Do you happen to really use minix filesystem or was
this just a fallout from some fuzz testing or something like that?

Honza
> ---
> fs/minix/inode.c | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/fs/minix/inode.c b/fs/minix/inode.c
> index a71f1cf894b9..d4bd94234ef7 100644
> --- a/fs/minix/inode.c
> +++ b/fs/minix/inode.c
> @@ -447,7 +447,8 @@ static const struct address_space_operations minix_aops = {
> .writepage = minix_writepage,
> .write_begin = minix_write_begin,
> .write_end = generic_write_end,
> - .bmap = minix_bmap
> + .bmap = minix_bmap,
> + .direct_IO = noop_direct_IO
> };
>
> static const struct inode_operations minix_symlink_inode_operations = {
> --
> 2.30.2
>
--
Jan Kara <[email protected]>
SUSE Labs, CR

2022-01-07 11:33:09

by Christian Brauner

[permalink] [raw]
Subject: Re: [PATCH] minix: fix bug when opening a file with O_DIRECT will

On Fri, Jan 07, 2022 at 07:12:39AM +0800, Qinghua Jin wrote:
> On Thu, Jan 6, 2022 at 11:43 PM Jan Kara <[email protected]> wrote:
>
> > On Thu 06-01-22 11:30:41, Qinghua Jin wrote:
> > > Testcase:
> > > 1. create a minix file system and mount it
> > > 2. open a file on the file system with O_RDWR|O_CREAT|O_TRUNC|O_DIRECT
> > > 3. open fails with -EINVAL but leaves an empty file behind. All other
> > > open() failures don't leave the failed open files behind.
> > >
> > > It is hard to check the direct_IO op before creating the inode. Just as
> > > ext4 and btrfs do, this patch will resolve the issue by allowing to
> > > create the file with O_DIRECT but returning error when writing the file.
> > >
> > > Signed-off-by: Qinghua Jin <[email protected]>
> >
> > The patch looks good. Thanks. Feel free to add:
> >
> > Reviewed-by: Jan Kara <[email protected]>
> >
> > Just out of curiosity: Do you happen to really use minix filesystem or was
> > this just a fallout from some fuzz testing or something like that?
> >
> > Thanks, It's a bug reported by [email protected]
> https://bugzilla.kernel.org/show_bug.cgi?id=213041
> I don't include the email because it's not reachable.

Fwiw, Colin's now at Intel:
Colin Ian King <[email protected]>

(While I'm here:
Acked-by: Christian Brauner <[email protected]>)

2022-01-07 13:37:20

by Qinghua Jin

[permalink] [raw]
Subject: [PATCH v2] minix: fix bug when opening a file with O_DIRECT

Testcase:
1. create a minix file system and mount it
2. open a file on the file system with O_RDWR|O_CREAT|O_TRUNC|O_DIRECT
3. open fails with -EINVAL but leaves an empty file behind. All other
open() failures don't leave the failed open files behind.

It is hard to check the direct_IO op before creating the inode. Just as
ext4 and btrfs do, this patch will resolve the issue by allowing to
create the file with O_DIRECT but returning error when writing the file.

Reported-by: Colin Ian King <[email protected]>
Reviewed-by: Jan Kara <[email protected]>
Acked-by: Christian Brauner <[email protected]>
Signed-off-by: Qinghua Jin <[email protected]>
---
fs/minix/inode.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/fs/minix/inode.c b/fs/minix/inode.c
index a71f1cf894b9..d4bd94234ef7 100644
--- a/fs/minix/inode.c
+++ b/fs/minix/inode.c
@@ -447,7 +447,8 @@ static const struct address_space_operations minix_aops = {
.writepage = minix_writepage,
.write_begin = minix_write_begin,
.write_end = generic_write_end,
- .bmap = minix_bmap
+ .bmap = minix_bmap,
+ .direct_IO = noop_direct_IO
};

static const struct inode_operations minix_symlink_inode_operations = {
--
2.30.2