Patchwork vfs: add missing check for __O_TMPFILE in fcntl_init() (Re: [PATCH] ext4: fix...)

login
register
mail settings
Submitter Zheng Liu
Date July 25, 2013, 12:13 a.m.
Message ID <20130725001319.GA12994@gmail.com>
Download mbox | patch
Permalink /patch/261563/
State New
Headers show

Comments

Zheng Liu - July 25, 2013, 12:13 a.m.
On Wed, Jul 24, 2013 at 09:58:45PM +0000, jon ernst wrote:
> > My purpose is to trigger this bug and fix it.  So I manually define this
> > flag for my convenience.
> >
> >> I saw it has
> >> been defined in header file. (fcntl.h) Did I miss anything?
> >
> > I guess that you might 'include <fcntl.h>' header file, right?  But the
> > O_TMPFILE is defined in $LINUX/include/uapi/asm-generic/fcntl.h.  So
> > maybe compiler couldn't find this header file.
> >
> > Regards,
> >                                                 - Zheng
> --
> Got it. Thank you.  BTW, the weird thing is I didn't see any
> discussion on ext4 mailing list before Al Viro committed this change.
> The first time I heard about this is someone reporting a bug about it.
> Document about this FLAG might be necessary.

Oh, this flag has been discussed in linux-fsdevel mailing list [1].
Maybe you don't subscribe it.

1. https://lkml.org/lkml/2013/7/3/181

> Also, about this code:
> 
> /* a horrid kludge trying to make sure that this will fail on old kernels */
> #define O_TMPFILE (__O_TMPFILE | O_DIRECTORY)
> 
> why not doing something like this:
> 
> #ifndef O_TMPFILE
> /* a horrid kludge trying to make sure that this will fail on old kernels */
> #define O_TMPFILE (__O_TMPFILE | O_DIRECTORY)
> #endif

Sorry, I don't take a closer look at that thread.  I look at the code,
and it seems that there is no any rule about adding '#ifndef XXX'.
Meanwhile, I notice a comment:

/*
 * When introducing new O_* bits, please check its uniqueness in fcntl_init().
 */

But we are missing adding O_TMPFILE flag in fcntl_init().  So I attach a
patch below.

Al, could you please answer this question?  Thanks.

Regards,
                                                - Zheng

Subject: [PATCH] vfs: add missing check for __O_TMPFILE in fcntl_init()

From: Zheng Liu <wenqing.lz@taobao.com>

As comment in include/uapi/asm-generic/fcntl.h described, when
introducing new O_* bits, we need to check its uniqueness in
fcntl_init().  But __O_TMPFILE bit is missing.  So fix it.

Signed-off-by: Zheng Liu <wenqing.lz@taobao.com>
---
*This patch is against ext4/dev tree.*

 fs/fcntl.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
jon ernst - July 25, 2013, 1:17 a.m.
On Thu, Jul 25, 2013 at 12:13 AM, Zheng Liu <gnehzuil.liu@gmail.com> wrote:
> On Wed, Jul 24, 2013 at 09:58:45PM +0000, jon ernst wrote:
>> > My purpose is to trigger this bug and fix it.  So I manually define this
>> > flag for my convenience.
>> >
>> >> I saw it has
>> >> been defined in header file. (fcntl.h) Did I miss anything?
>> >
>> > I guess that you might 'include <fcntl.h>' header file, right?  But the
>> > O_TMPFILE is defined in $LINUX/include/uapi/asm-generic/fcntl.h.  So
>> > maybe compiler couldn't find this header file.
>> >
>> > Regards,
>> >                                                 - Zheng
>> --
>> Got it. Thank you.  BTW, the weird thing is I didn't see any
>> discussion on ext4 mailing list before Al Viro committed this change.
>> The first time I heard about this is someone reporting a bug about it.
>> Document about this FLAG might be necessary.
>
> Oh, this flag has been discussed in linux-fsdevel mailing list [1].
> Maybe you don't subscribe it.
>
> 1. https://lkml.org/lkml/2013/7/3/181
>
>> Also, about this code:
>>
>> /* a horrid kludge trying to make sure that this will fail on old kernels */
>> #define O_TMPFILE (__O_TMPFILE | O_DIRECTORY)
>>
>> why not doing something like this:
>>
>> #ifndef O_TMPFILE
>> /* a horrid kludge trying to make sure that this will fail on old kernels */
>> #define O_TMPFILE (__O_TMPFILE | O_DIRECTORY)
>> #endif
>
> Sorry, I don't take a closer look at that thread.  I look at the code,
> and it seems that there is no any rule about adding '#ifndef XXX'.
> Meanwhile, I notice a comment:
>
> /*
>  * When introducing new O_* bits, please check its uniqueness in fcntl_init().
>  */
>
> But we are missing adding O_TMPFILE flag in fcntl_init().  So I attach a
> patch below.
>

Great. I think this is the reason O_TMPFILE is unrecognizable in user
space. As the problem I reported 2 days ago in the thread that you
fixed another bug.

> Al, could you please answer this question?  Thanks.
>
> Regards,
>                                                 - Zheng
>
> Subject: [PATCH] vfs: add missing check for __O_TMPFILE in fcntl_init()
>
> From: Zheng Liu <wenqing.lz@taobao.com>
>
> As comment in include/uapi/asm-generic/fcntl.h described, when
> introducing new O_* bits, we need to check its uniqueness in
> fcntl_init().  But __O_TMPFILE bit is missing.  So fix it.
>
> Signed-off-by: Zheng Liu <wenqing.lz@taobao.com>
> ---
> *This patch is against ext4/dev tree.*
>
>  fs/fcntl.c |    4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/fs/fcntl.c b/fs/fcntl.c
> index 6599222..65343c3 100644
> --- a/fs/fcntl.c
> +++ b/fs/fcntl.c
> @@ -730,14 +730,14 @@ static int __init fcntl_init(void)
>          * Exceptions: O_NONBLOCK is a two bit define on parisc; O_NDELAY
>          * is defined as O_NONBLOCK on some platforms and not on others.
>          */
> -       BUILD_BUG_ON(19 - 1 /* for O_RDONLY being 0 */ != HWEIGHT32(
> +       BUILD_BUG_ON(20 - 1 /* for O_RDONLY being 0 */ != HWEIGHT32(
>                 O_RDONLY        | O_WRONLY      | O_RDWR        |
>                 O_CREAT         | O_EXCL        | O_NOCTTY      |
>                 O_TRUNC         | O_APPEND      | /* O_NONBLOCK | */
>                 __O_SYNC        | O_DSYNC       | FASYNC        |
>                 O_DIRECT        | O_LARGEFILE   | O_DIRECTORY   |
>                 O_NOFOLLOW      | O_NOATIME     | O_CLOEXEC     |
> -               __FMODE_EXEC    | O_PATH
> +               __FMODE_EXEC    | O_PATH        | __O_TMPFILE
>                 ));
>
>         fasync_cache = kmem_cache_create("fasync_cache",
> --
> 1.7.9.7
>
--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Zheng Liu - July 29, 2013, 6:17 a.m.
[To linux-fsdevel mailing list]

On Thu, Jul 25, 2013 at 08:13:19AM +0800, Zheng Liu wrote:
[...]
> > Also, about this code:
> > 
> > /* a horrid kludge trying to make sure that this will fail on old kernels */
> > #define O_TMPFILE (__O_TMPFILE | O_DIRECTORY)
> > 
> > why not doing something like this:
> > 
> > #ifndef O_TMPFILE
> > /* a horrid kludge trying to make sure that this will fail on old kernels */
> > #define O_TMPFILE (__O_TMPFILE | O_DIRECTORY)
> > #endif
> 
> Sorry, I don't take a closer look at that thread.  I look at the code,
> and it seems that there is no any rule about adding '#ifndef XXX'.
> Meanwhile, I notice a comment:
> 
> /*
>  * When introducing new O_* bits, please check its uniqueness in fcntl_init().
>  */
> 
> But we are missing adding O_TMPFILE flag in fcntl_init().  So I attach a
> patch below.
> 
> Al, could you please answer this question?  Thanks.
> 
> Regards,
>                                                 - Zheng
> 
> Subject: [PATCH] vfs: add missing check for __O_TMPFILE in fcntl_init()
> 
> From: Zheng Liu <wenqing.lz@taobao.com>
> 
> As comment in include/uapi/asm-generic/fcntl.h described, when
> introducing new O_* bits, we need to check its uniqueness in
> fcntl_init().  But __O_TMPFILE bit is missing.  So fix it.
> 
> Signed-off-by: Zheng Liu <wenqing.lz@taobao.com>

Any comment?

Thanks,
                                                - Zheng

> ---
> *This patch is against ext4/dev tree.*
> 
>  fs/fcntl.c |    4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/fs/fcntl.c b/fs/fcntl.c
> index 6599222..65343c3 100644
> --- a/fs/fcntl.c
> +++ b/fs/fcntl.c
> @@ -730,14 +730,14 @@ static int __init fcntl_init(void)
>  	 * Exceptions: O_NONBLOCK is a two bit define on parisc; O_NDELAY
>  	 * is defined as O_NONBLOCK on some platforms and not on others.
>  	 */
> -	BUILD_BUG_ON(19 - 1 /* for O_RDONLY being 0 */ != HWEIGHT32(
> +	BUILD_BUG_ON(20 - 1 /* for O_RDONLY being 0 */ != HWEIGHT32(
>  		O_RDONLY	| O_WRONLY	| O_RDWR	|
>  		O_CREAT		| O_EXCL	| O_NOCTTY	|
>  		O_TRUNC		| O_APPEND	| /* O_NONBLOCK	| */
>  		__O_SYNC	| O_DSYNC	| FASYNC	|
>  		O_DIRECT	| O_LARGEFILE	| O_DIRECTORY	|
>  		O_NOFOLLOW	| O_NOATIME	| O_CLOEXEC	|
> -		__FMODE_EXEC	| O_PATH
> +		__FMODE_EXEC	| O_PATH	| __O_TMPFILE
>  		));
>  
>  	fasync_cache = kmem_cache_create("fasync_cache",
> -- 
> 1.7.9.7
> 
--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Patch

diff --git a/fs/fcntl.c b/fs/fcntl.c
index 6599222..65343c3 100644
--- a/fs/fcntl.c
+++ b/fs/fcntl.c
@@ -730,14 +730,14 @@  static int __init fcntl_init(void)
 	 * Exceptions: O_NONBLOCK is a two bit define on parisc; O_NDELAY
 	 * is defined as O_NONBLOCK on some platforms and not on others.
 	 */
-	BUILD_BUG_ON(19 - 1 /* for O_RDONLY being 0 */ != HWEIGHT32(
+	BUILD_BUG_ON(20 - 1 /* for O_RDONLY being 0 */ != HWEIGHT32(
 		O_RDONLY	| O_WRONLY	| O_RDWR	|
 		O_CREAT		| O_EXCL	| O_NOCTTY	|
 		O_TRUNC		| O_APPEND	| /* O_NONBLOCK	| */
 		__O_SYNC	| O_DSYNC	| FASYNC	|
 		O_DIRECT	| O_LARGEFILE	| O_DIRECTORY	|
 		O_NOFOLLOW	| O_NOATIME	| O_CLOEXEC	|
-		__FMODE_EXEC	| O_PATH
+		__FMODE_EXEC	| O_PATH	| __O_TMPFILE
 		));
 
 	fasync_cache = kmem_cache_create("fasync_cache",