Re: sysfs: duplicate filename '74' can not be created

[Alexey Dobriyan <adobriyan@sw.ru>]

On Wed, Mar 05, 2008 at 10:00:38PM -0800, Greg KH wrote:
> On Wed, Mar 05, 2008 at 02:29:22PM +0530, Dhaval Giani wrote:
> > On Tue, Mar 04, 2008 at 11:24:36AM -0800, Greg KH wrote:
> > > On Tue, Mar 04, 2008 at 10:03:39PM +0530, Dhaval Giani wrote:
> > > > On Tue, Mar 04, 2008 at 07:09:40PM +0300, Alexey Dobriyan wrote:
> > > > > [Cc'ing /sys/kernel/uids guys]
> > > > > 
> > > > 
> > > > [Also adding gregkh to the cc :)]
> > > > 
> > > > > > Mar  3 13:50:01 helena kernel: sysfs: duplicate filename '74' can not be created
> > > > > > Mar  3 13:50:01 helena kernel: WARNING: at fs/sysfs/dir.c:424 sysfs_add_one()
> > > > > > Mar  3 13:50:01 helena kernel: Pid: 21911, comm: sshd Not tainted 2.6.24.3 #1
> > > 
> > > What kernel version is this?  2.6.24 had lots of "issues" with the uids
> > > code, I'm amazed it worked at all :)
> > > 
> > 
> > :). Seems to be a stable series kernel (2.6.24.3). Can you please tell
> > me some of the issues, so that I can try to go fix it up.
> 
> Ick, there were a lot of them.  Take a look at the changes to
> kernel/user.c (I think that was the file).  A lot of the work was
> reworking it due to the kobject changes, but at the same time lots of
> code was fixed.
> 
> It is a harmless warning, everything should still work properly, right?

The whole existence of uids_mutex should guarantee that duplicate name
can't happen, and that logic with async kobject removal is still in
place. And there was an oops afterwards:
http://marc.info/?l=linux-kernel&m=120455796407423&w=2