netdev - Re: Kernel crash after using new Intel NIC (igb)

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <1306447292.2543.32.camel@edumazet-laptop>
Date:	Fri, 27 May 2011 00:01:32 +0200
From:	Eric Dumazet <eric.dumazet@...il.com>
To:	Arun Sharma <asharma@...com>
Cc:	Maximilian Engelhardt <maxi@...monizer.de>,
	linux-kernel@...r.kernel.org, netdev@...r.kernel.org,
	StuStaNet Vorstand <vorstand@...sta.mhn.de>
Subject: Re: Kernel crash after using new Intel NIC (igb)

Le jeudi 26 mai 2011 à 14:48 -0700, Arun Sharma a écrit :
> On 5/26/11 12:47 PM, Eric Dumazet wrote:
> 
> > You dont get the problem. Problem is : We can do the empty() test only
> > if protected by the lock.
> >
> > If not locked, result can be wrong. [ false positive or negative ]
> >
> 
> 
> Agreed. Failing to unlink from unused list when we should have sounds wrong.
> 
> >> The list modification under unused_peers.lock looks generally safe. But
> >> the control flow (based on refcnt) done outside the lock might have races.
> >>
> >
> > "might" is not a good word when dealing with this ;)
> 
> Potential race in the current code:
> 
> initial refcnt = 1
> 
>         T1:                                        T2
> 
> atomic_dec_and_lock(refcnt)
> // refcnt == 0
> 
>                                         atomic_add_unless(refcnt)
>                                         unlink_from_unused()
> 
> list_add_tail(unused)
>                                         // T2 using "unused" entry
> 
> 
> > Did you test my fix ?
> 
> I could try it on one or two machines - but it won't tell us anything 
> for weeks if not months. Unfortunately my next window to try a new 
> kernel on a large enough sample is several months away.
> 
> >
> > Its doing the right thing : Using refcnt as the only marker to say if
> > the item must be removed from unused list (and lock the central lock
> > protecting this list only when needed)
> >
> > Since we already must do an atomic operation on refcnt, using
> > atomic_inc_return [ or similar full barrier op ] is enough to tell us
> > the truth.
> 
> Yeah - using the refcnt seems better than list_empty(), but I'm not sure 
> that your patch addresses the race above.

It does.

It becomes

        T1:                                        T2
> 
> atomic_dec_and_lock(refcnt)
> // refcnt == 0
> 
>                          newref = atomic_add_unless_and_return(refcnt)
> 
> list_add_tail(unused)
  unlock();
>                                         
> 			   if (newref == 1) {
				   lock()
                                   unlink_from_unused()
				   unlock()
			   }



--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html