2009-07-14 15:57:37

by Peter Oberparleiter

[permalink] [raw]
Subject: [PATCH] sysfs: fix hardlink count on device_move

From: Peter Oberparleiter <[email protected]>

Update directory hardlink count when moving kobjects to a new parent.
Fixes the following problem which occurs when several devices are
moved to the same parent and then unregistered:

> ls -laF /sys/devices/css0/defunct/
> total 0
> drwxr-xr-x 4294967295 root root 0 2009-07-14 17:02 ./
> drwxr-xr-x 114 root root 0 2009-07-14 17:02 ../
> drwxr-xr-x 2 root root 0 2009-07-14 17:01 power/
> -rw-r--r-- 1 root root 4096 2009-07-14 17:01 uevent

Signed-off-by: Peter Oberparleiter <[email protected]>
---
fs/sysfs/dir.c | 2 ++
1 file changed, 2 insertions(+)

Index: linux-2.6.31-rc3/fs/sysfs/dir.c
===================================================================
--- linux-2.6.31-rc3.orig/fs/sysfs/dir.c
+++ linux-2.6.31-rc3/fs/sysfs/dir.c
@@ -939,8 +939,10 @@ again:
/* Remove from old parent's list and insert into new parent's list. */
sysfs_unlink_sibling(sd);
sysfs_get(new_parent_sd);
+ drop_nlink(old_parent->d_inode);
sysfs_put(sd->s_parent);
sd->s_parent = new_parent_sd;
+ inc_nlink(new_parent->d_inode);
sysfs_link_sibling(sd);

out_unlock:


2009-07-14 16:05:25

by Greg KH

[permalink] [raw]
Subject: Re: [PATCH] sysfs: fix hardlink count on device_move

On Tue, Jul 14, 2009 at 05:56:15PM +0200, Peter Oberparleiter wrote:
> From: Peter Oberparleiter <[email protected]>
>
> Update directory hardlink count when moving kobjects to a new parent.
> Fixes the following problem which occurs when several devices are
> moved to the same parent and then unregistered:
>
> > ls -laF /sys/devices/css0/defunct/
> > total 0
> > drwxr-xr-x 4294967295 root root 0 2009-07-14 17:02 ./
> > drwxr-xr-x 114 root root 0 2009-07-14 17:02 ../
> > drwxr-xr-x 2 root root 0 2009-07-14 17:01 power/
> > -rw-r--r-- 1 root root 4096 2009-07-14 17:01 uevent
>
> Signed-off-by: Peter Oberparleiter <[email protected]>

Thanks for the patch.

Is this something new in 2.6.31-rc, or does it also show up in .30
and/or .27?

thanks,

greg k-h

2009-07-15 09:19:24

by Peter Oberparleiter

[permalink] [raw]
Subject: Re: [PATCH] sysfs: fix hardlink count on device_move

Greg KH wrote:
> On Tue, Jul 14, 2009 at 05:56:15PM +0200, Peter Oberparleiter wrote:
>> From: Peter Oberparleiter <[email protected]>
>>
>> Update directory hardlink count when moving kobjects to a new parent.
>> Fixes the following problem which occurs when several devices are
>> moved to the same parent and then unregistered:
>>
>>> ls -laF /sys/devices/css0/defunct/
>>> total 0
>>> drwxr-xr-x 4294967295 root root 0 2009-07-14 17:02 ./
>>> drwxr-xr-x 114 root root 0 2009-07-14 17:02 ../
>>> drwxr-xr-x 2 root root 0 2009-07-14 17:01 power/
>>> -rw-r--r-- 1 root root 4096 2009-07-14 17:01 uevent
>> Signed-off-by: Peter Oberparleiter <[email protected]>
>
> Thanks for the patch.
>
> Is this something new in 2.6.31-rc, or does it also show up in .30
> and/or .27?

This is an old problem that existed since the introduction of
sysfs_move_dir() in 2.6.20.

2009-07-15 13:22:30

by Greg KH

[permalink] [raw]
Subject: Re: [PATCH] sysfs: fix hardlink count on device_move

On Wed, Jul 15, 2009 at 11:17:43AM +0200, Peter Oberparleiter wrote:
> Greg KH wrote:
> > On Tue, Jul 14, 2009 at 05:56:15PM +0200, Peter Oberparleiter wrote:
> >> From: Peter Oberparleiter <[email protected]>
> >>
> >> Update directory hardlink count when moving kobjects to a new parent.
> >> Fixes the following problem which occurs when several devices are
> >> moved to the same parent and then unregistered:
> >>
> >>> ls -laF /sys/devices/css0/defunct/
> >>> total 0
> >>> drwxr-xr-x 4294967295 root root 0 2009-07-14 17:02 ./
> >>> drwxr-xr-x 114 root root 0 2009-07-14 17:02 ../
> >>> drwxr-xr-x 2 root root 0 2009-07-14 17:01 power/
> >>> -rw-r--r-- 1 root root 4096 2009-07-14 17:01 uevent
> >> Signed-off-by: Peter Oberparleiter <[email protected]>
> >
> > Thanks for the patch.
> >
> > Is this something new in 2.6.31-rc, or does it also show up in .30
> > and/or .27?
>
> This is an old problem that existed since the introduction of
> sysfs_move_dir() in 2.6.20.

Thanks for the info, I'll mark it as such.

greg k-h