[josef-btrfs:master 9/11] fs/btrfs/delayed-inode.c:1013:1-11: second lock on line 1013

[josef-btrfs:master 9/11] fs/btrfs/delayed-inode.c:1013:1-11: second lock on line 1013

[Fengguang Wu]

[-- Attachment #1: Type: text/plain, Size: 978 bytes --]

Hi Josef,

There are (old) coccinelle warnings in

tree:   git://git.kernel.org/pub/scm/linux/kernel/git/josef/btrfs-next.git master
head:   233564d3f5bde271885f5b784b63cfe627bba0a5
commit: ec93ee7ac77b2e95934253823918c3c684c1a36c [9/11] Btrfs: barrier before waitqueue_active

All coccinelle warnings:

+ fs/btrfs/delayed-inode.c:1013:1-11: second lock on line 1013
--
+ fs/btrfs/volumes.c:2888:1-3: WARNING: PTR_RET can be used (patch attached)

vim +1013 fs/btrfs/delayed-inode.c
  1010		int ret = 0;
  1011	
  1012	do_again:
> 1013		mutex_lock(&node->mutex);
  1014		curr = __btrfs_first_delayed_deletion_item(node);
  1015		if (!curr)
  1016			goto delete_fail;
...
  1031                 if (curr)
> 1032                         goto do_again;
  1033                 else
  1034                         goto delete_fail;

---
0-DAY kernel build testing backend         Open Source Technology Centre
Fengguang Wu <wfg@linux.intel.com>                     Intel Corporation

[-- Attachment #2: ptr_ret:volumes.patch --]
[-- Type: text/x-diff, Size: 413 bytes --]

Generated by: api/ptr_ret.cocci


 Use PTR_RET rather than if(IS_ERR(...)) + PTR_ERR

--- a/fs/btrfs/volumes.c
+++ b/fs/btrfs/volumes.c
@@ -2885,10 +2885,7 @@ int btrfs_resume_balance_async(struct bt
 	}
 
 	tsk = kthread_run(balance_kthread, fs_info, "btrfs-balance");
-	if (IS_ERR(tsk))
-		return PTR_ERR(tsk);
-
-	return 0;
+	return PTR_RET(tsk);
 }
 
 int btrfs_recover_balance(struct btrfs_fs_info *fs_info)

Re: [josef-btrfs:master 9/11] fs/btrfs/delayed-inode.c:1013:1-11: second lock on line 1013

[Josef Bacik]

On Fri, Aug 03, 2012 at 07:16:59AM -0600, Fengguang Wu wrote:
> Hi Josef,
> 
> There are (old) coccinelle warnings in
> 
> tree:   git://git.kernel.org/pub/scm/linux/kernel/git/josef/btrfs-next.git master
> head:   233564d3f5bde271885f5b784b63cfe627bba0a5
> commit: ec93ee7ac77b2e95934253823918c3c684c1a36c [9/11] Btrfs: barrier before waitqueue_active
> 
> All coccinelle warnings:
> 
> + fs/btrfs/delayed-inode.c:1013:1-11: second lock on line 1013
> --
> + fs/btrfs/volumes.c:2888:1-3: WARNING: PTR_RET can be used (patch attached)
> 
> vim +1013 fs/btrfs/delayed-inode.c
>   1010		int ret = 0;
>   1011	
>   1012	do_again:
> > 1013		mutex_lock(&node->mutex);
>   1014		curr = __btrfs_first_delayed_deletion_item(node);
>   1015		if (!curr)
>   1016			goto delete_fail;
> ...
>   1031                 if (curr)
> > 1032                         goto do_again;
>   1033                 else
>   1034                         goto delete_fail;
> 

Ok that wasn't related to my patch so I'm not going to fold it in, do you want
to send me a real patch with a Signed-off-by that I can apply?  Thanks,

Josef

Re: [josef-btrfs:master 9/11] fs/btrfs/delayed-inode.c:1013:1-11: second lock on line 1013

[Fengguang Wu]

On Fri, Aug 03, 2012 at 05:11:32PM -0400, Josef Bacik wrote:
> On Fri, Aug 03, 2012 at 07:16:59AM -0600, Fengguang Wu wrote:
> > Hi Josef,
> > 
> > There are (old) coccinelle warnings in
> > 
> > tree:   git://git.kernel.org/pub/scm/linux/kernel/git/josef/btrfs-next.git master
> > head:   233564d3f5bde271885f5b784b63cfe627bba0a5
> > commit: ec93ee7ac77b2e95934253823918c3c684c1a36c [9/11] Btrfs: barrier before waitqueue_active
> > 
> > All coccinelle warnings:
> > 
> > + fs/btrfs/delayed-inode.c:1013:1-11: second lock on line 1013
> > --
> > + fs/btrfs/volumes.c:2888:1-3: WARNING: PTR_RET can be used (patch attached)
> > 
> > vim +1013 fs/btrfs/delayed-inode.c
> >   1010		int ret = 0;
> >   1011	
> >   1012	do_again:
> > > 1013		mutex_lock(&node->mutex);
> >   1014		curr = __btrfs_first_delayed_deletion_item(node);
> >   1015		if (!curr)
> >   1016			goto delete_fail;
> > ...
> >   1031                 if (curr)
> > > 1032                         goto do_again;
> >   1033                 else
> >   1034                         goto delete_fail;
> > 
> 
> Ok that wasn't related to my patch so I'm not going to fold it in, do you want
> to send me a real patch with a Signed-off-by that I can apply?  Thanks,

Sure. I'll send one bug fix and another code cleanup for each of the
warnings.

Thanks,
Fengguang