Patchwork [ext4:dev,47/50] fs/ext4/extents.c:4949 ext4_ext_punch_hole() warn: inconsistent returns mutex:&inode->i_mutex: locked (4813,4815) unlocked (4806,4949)

login
register
mail settings
Submitter Dmitri Monakho
Date Oct. 1, 2012, 3:16 a.m.
Message ID <87mx06vq3w.fsf@openvz.org>
Download mbox | patch
Permalink /patch/188209/
State Superseded
Headers show

Comments

Dmitri Monakho - Oct. 1, 2012, 3:16 a.m.
On Mon, 1 Oct 2012 10:50:47 +0800, Fengguang Wu <fengguang.wu@intel.com> wrote:
> Hi Dmitry,
> 
> FYI, there are new smatch warnings show up in
> 
> tree:   git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4.git dev
> head:   07043757c1a4018a04aded14deb1f28269a4b2ad
> commit: c6a59106b4106d13dc824a388c5572ddd65a0779 [47/50] ext4: punch_hole should wait for DIO writers
> 
>   arch/x86/include/asm/jump_label.h:25 arch_static_branch() info: ignoring unreachable code.
> + fs/ext4/extents.c:4949 ext4_ext_punch_hole() warn: inconsistent returns mutex:&inode->i_mutex: locked (4813,4815) unlocked (4806,4949)
Yes, that is correct, my bad. Sorry, patch attached
BTW  What tool(compile option) you use to catch this issue?

> 
> vim +4949 fs/ext4/extents.c
> 
> c6a59106 (Dmitry Monakhov   2012-09-29  4808) 
> c6a59106 (Dmitry Monakhov   2012-09-29  4809) 	mutex_lock(&inode->i_mutex);
> c6a59106 (Dmitry Monakhov   2012-09-29  4810) 	/* Need recheck file flags under mutex */
> c6a59106 (Dmitry Monakhov   2012-09-29  4811) 	/* It's not possible punch hole on append only file */
> c6a59106 (Dmitry Monakhov   2012-09-29  4812) 	if (IS_APPEND(inode) || IS_IMMUTABLE(inode))
> c6a59106 (Dmitry Monakhov   2012-09-29  4813) 		return -EPERM;
                                                               ^^^^^^^^^^^^^^^^
> c6a59106 (Dmitry Monakhov   2012-09-29  4814) 	if (IS_SWAPFILE(inode))
> c6a59106 (Dmitry Monakhov   2012-09-29  4815) 		return -ETXTBSY;
                                                               ^^^^^^^^^^^^^^^^^^
> c6a59106 (Dmitry Monakhov   2012-09-29  4816) 
> 2be4751b (Allison Henderson 2011-09-03  4817) 	/* No need to punch hole beyond i_size */
> 2be4751b (Allison Henderson 2011-09-03  4818) 	if (offset >= inode->i_size)
> c6a59106 (Dmitry Monakhov   2012-09-29  4819) 		goto out_mutex;
> 2be4751b (Allison Henderson 2011-09-03  4820)
Theodore Ts'o - Oct. 1, 2012, 3:20 a.m.
On Mon, Oct 01, 2012 at 07:16:35AM +0400, Dmitry Monakhov wrote:
> On Mon, 1 Oct 2012 10:50:47 +0800, Fengguang Wu <fengguang.wu@intel.com> wrote:
> > Hi Dmitry,
> > 
> > FYI, there are new smatch warnings show up in
> > 
> > tree:   git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4.git dev
> > head:   07043757c1a4018a04aded14deb1f28269a4b2ad
> > commit: c6a59106b4106d13dc824a388c5572ddd65a0779 [47/50] ext4: punch_hole should wait for DIO writers
> > 
> >   arch/x86/include/asm/jump_label.h:25 arch_static_branch() info: ignoring unreachable code.
> > + fs/ext4/extents.c:4949 ext4_ext_punch_hole() warn: inconsistent returns mutex:&inode->i_mutex: locked (4813,4815) unlocked (4806,4949)
> Yes, that is correct, my bad. Sorry, patch attached
> BTW  What tool(compile option) you use to catch this issue?

Fengguang is using Dan Carpenter's "smatch" tool.  For more
information, see:

	http://www.linuxplumbersconf.org/2011/ocw/sessions/165

						- Ted
--
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
Wu Fengguang - Oct. 1, 2012, 3:22 a.m.
On Mon, Oct 01, 2012 at 07:16:35AM +0400, Dmitry Monakhov wrote:
> On Mon, 1 Oct 2012 10:50:47 +0800, Fengguang Wu <fengguang.wu@intel.com> wrote:
> > Hi Dmitry,
> > 
> > FYI, there are new smatch warnings show up in
> > 
> > tree:   git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4.git dev
> > head:   07043757c1a4018a04aded14deb1f28269a4b2ad
> > commit: c6a59106b4106d13dc824a388c5572ddd65a0779 [47/50] ext4: punch_hole should wait for DIO writers
> > 
> >   arch/x86/include/asm/jump_label.h:25 arch_static_branch() info: ignoring unreachable code.
> > + fs/ext4/extents.c:4949 ext4_ext_punch_hole() warn: inconsistent returns mutex:&inode->i_mutex: locked (4813,4815) unlocked (4806,4949)
> Yes, that is correct, my bad. Sorry, patch attached
> BTW  What tool(compile option) you use to catch this issue?

Dmitry, I'm running smatch, a very powerful tool, written and maintained by Dan Carpenter.

http://smatch.sourceforge.net/

Thanks,
Fengguang

> > 
> > vim +4949 fs/ext4/extents.c
> > 
> > c6a59106 (Dmitry Monakhov   2012-09-29  4808) 
> > c6a59106 (Dmitry Monakhov   2012-09-29  4809) 	mutex_lock(&inode->i_mutex);
> > c6a59106 (Dmitry Monakhov   2012-09-29  4810) 	/* Need recheck file flags under mutex */
> > c6a59106 (Dmitry Monakhov   2012-09-29  4811) 	/* It's not possible punch hole on append only file */
> > c6a59106 (Dmitry Monakhov   2012-09-29  4812) 	if (IS_APPEND(inode) || IS_IMMUTABLE(inode))
> > c6a59106 (Dmitry Monakhov   2012-09-29  4813) 		return -EPERM;
>                                                                ^^^^^^^^^^^^^^^^
> > c6a59106 (Dmitry Monakhov   2012-09-29  4814) 	if (IS_SWAPFILE(inode))
> > c6a59106 (Dmitry Monakhov   2012-09-29  4815) 		return -ETXTBSY;
>                                                                ^^^^^^^^^^^^^^^^^^
> > c6a59106 (Dmitry Monakhov   2012-09-29  4816) 
> > 2be4751b (Allison Henderson 2011-09-03  4817) 	/* No need to punch hole beyond i_size */
> > 2be4751b (Allison Henderson 2011-09-03  4818) 	if (offset >= inode->i_size)
> > c6a59106 (Dmitry Monakhov   2012-09-29  4819) 		goto out_mutex;
> > 2be4751b (Allison Henderson 2011-09-03  4820) 

> >From 6cf5cfc5db6c792c76c68bdbc765b733b5ad5488 Mon Sep 17 00:00:00 2001
> From: Dmitry Monakhov <dmonakhov@openvz.org>
> Date: Mon, 1 Oct 2012 07:09:49 +0400
> Subject: [PATCH] ext4: ext4_ext_punch_hole put i_mutex on error
> 
> Fix mistypo introduced in 'ext4: punch_hole should wait for DIO writers' patch.
> 
> Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
> ---
>  fs/ext4/extents.c |   12 ++++++++----
>  1 files changed, 8 insertions(+), 4 deletions(-)
> 
> diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c
> index f02ffce..0a720d4 100644
> --- a/fs/ext4/extents.c
> +++ b/fs/ext4/extents.c
> @@ -4811,10 +4811,14 @@ int ext4_ext_punch_hole(struct file *file, loff_t offset, loff_t length)
>  	mutex_lock(&inode->i_mutex);
>  	/* Need recheck file flags under mutex */
>  	/* It's not possible punch hole on append only file */
> -	if (IS_APPEND(inode) || IS_IMMUTABLE(inode))
> -		return -EPERM;
> -	if (IS_SWAPFILE(inode))
> -		return -ETXTBSY;
> +	if (IS_APPEND(inode) || IS_IMMUTABLE(inode)) {
> +		err = -EPERM;
> +		goto out_mutex;
> +	}
> +	if (IS_SWAPFILE(inode)) {
> +		err = -ETXTBSY;
> +		goto out_mutex;
> +	}
>  
>  	/* No need to punch hole beyond i_size */
>  	if (offset >= inode->i_size)
> -- 
> 1.7.7.6
> 

--
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

From 6cf5cfc5db6c792c76c68bdbc765b733b5ad5488 Mon Sep 17 00:00:00 2001
From: Dmitry Monakhov <dmonakhov@openvz.org>
Date: Mon, 1 Oct 2012 07:09:49 +0400
Subject: [PATCH] ext4: ext4_ext_punch_hole put i_mutex on error

Fix mistypo introduced in 'ext4: punch_hole should wait for DIO writers' patch.

Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org>
---
 fs/ext4/extents.c |   12 ++++++++----
 1 files changed, 8 insertions(+), 4 deletions(-)

diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c
index f02ffce..0a720d4 100644
--- a/fs/ext4/extents.c
+++ b/fs/ext4/extents.c
@@ -4811,10 +4811,14 @@  int ext4_ext_punch_hole(struct file *file, loff_t offset, loff_t length)
 	mutex_lock(&inode->i_mutex);
 	/* Need recheck file flags under mutex */
 	/* It's not possible punch hole on append only file */
-	if (IS_APPEND(inode) || IS_IMMUTABLE(inode))
-		return -EPERM;
-	if (IS_SWAPFILE(inode))
-		return -ETXTBSY;
+	if (IS_APPEND(inode) || IS_IMMUTABLE(inode)) {
+		err = -EPERM;
+		goto out_mutex;
+	}
+	if (IS_SWAPFILE(inode)) {
+		err = -ETXTBSY;
+		goto out_mutex;
+	}
 
 	/* No need to punch hole beyond i_size */
 	if (offset >= inode->i_size)
-- 
1.7.7.6