All of lore.kernel.org
 help / color / mirror / Atom feed
From: Greg KH <greg@kroah.com>
To: Maneesh Soni <maneesh@in.ibm.com>
Cc: Keith Owens <kaos@sgi.com>, Andrew Morton <akpm@osdl.org>,
	linux-kernel@vger.kernel.org
Subject: Re: 2.6.13-rc4 use after free in class_device_attr_show
Date: Wed, 10 Aug 2005 15:35:53 -0700	[thread overview]
Message-ID: <20050810223552.GB6045@kroah.com> (raw)
In-Reply-To: <20050810100636.GB5334@in.ibm.com>

On Wed, Aug 10, 2005 at 03:36:36PM +0530, Maneesh Soni wrote:
> On Wed, Aug 10, 2005 at 04:26:51PM +1000, Keith Owens wrote:
> > FYI, the intermittent free after use in sysfs is still there in
> > 2.6.13-rc6.
> > 
> 
> The race condition is known here. It is some thing in the upper layer. 
> In this case "driver/base/class.c" which frees the kobject's attributes 
> even if there are live references to kobject.
> 
> 
> open sysfs file				unregister class device
> sysfs_open_file()			class_device_del()
>   -> takes a ref on kobject		  -> kfree attribute struct
>      -> accesses attributes		  -> kobject_del()
> 					      -> kref_put()	
> close sysfs file				  
> sysfs_release()				    
>   -> acesses attributes using s_element
>   -> drops ref to kobject
> 
> Solution could be either we have reference counting for attributes also
> or keep attributes alive till the last reference to the kobject. Both these
> needs changes in the driver core.
> 
> Greg, will the following patch make sense? This postpones the kfree() of
> devt_attr till class_dev_release() is called. 

Yes, that patch looks good, if you fix up the space vs. tabs issue :)

But will that really fix this race?  I was under the impression the oops
didn't come from trying to access the devt_attr, but the sysfs s_element
pointer?

> Please check this patch out, if this helps or not.

I'd be interested in seeing if this fixes it.

thanks,

greg k-h

  reply	other threads:[~2005-08-10 22:36 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2005-07-30  5:47 2.6.13-rc4 use after free in class_device_attr_show Keith Owens
2005-07-30  9:29 ` Andrew Morton
2005-08-01 12:14   ` Keith Owens
2005-08-01 14:03     ` Keith Owens
2005-08-01 19:03     ` Andrew Morton
2005-08-02  3:05       ` Keith Owens
2005-08-02  3:32         ` Keith Owens
2005-08-02  8:04         ` Maneesh Soni
2005-08-02 17:33           ` Greg KH
2005-08-10  6:26           ` Keith Owens
2005-08-10 10:06             ` Maneesh Soni
2005-08-10 22:35               ` Greg KH [this message]
2005-08-11  5:34                 ` Maneesh Soni

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20050810223552.GB6045@kroah.com \
    --to=greg@kroah.com \
    --cc=akpm@osdl.org \
    --cc=kaos@sgi.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=maneesh@in.ibm.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.