netdev - Re: [PATCH] netconsole: fix deadlock when removing net driver that netconsole is using (v2)

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: <20110422213145.GA2332@neilslaptop.think-freely.org>
Date:	Fri, 22 Apr 2011 17:31:45 -0400
From:	Neil Horman <nhorman@...driver.com>
To:	David Miller <davem@...emloft.net>
Cc:	netdev@...r.kernel.org
Subject: Re: [PATCH] netconsole: fix deadlock when removing net driver that
 netconsole is using (v2)

On Fri, Apr 22, 2011 at 11:40:33AM -0700, David Miller wrote:
> From: Neil Horman <nhorman@...driver.com>
> Date: Fri, 22 Apr 2011 14:10:59 -0400
> 
> > @@ -683,9 +684,16 @@ static int netconsole_netdev_event(struct notifier_block *this,
> >  				 * rtnl_lock already held
> >  				 */
> >  				if (nt->np.dev) {
> > +					spin_unlock_irqrestore(
> > +							      &target_list_lock,
> > +							      flags);
> >  					__netpoll_cleanup(&nt->np);
> > +					spin_lock_irqsave(&target_list_lock,
> > +							  flags);
> >  					dev_put(nt->np.dev);
> >  					nt->np.dev = NULL;
> > +					netconsole_target_put(nt);
> > +					goto restart;
> 
> If you drop the lock here, another cpu can put the device and set it
> to NULL.
> 
> In which case you'll double release when you regrab the lock here.
> 
> Too bad you can't NULL out the device before dropping the lock,
> because that way would be able to ensure you'd be the only releaser.
> 
> But that won't work because __netpoll_cleanup() wants the device
> pointer to be set.
> 
> Hmmm...
> 
I understand what you're saying here, but I think we're ok in this particular
case.  I say that because all other callers of __netpoll_cleanup, call it via
netpoll_cleanup, which does the dev_put under protection of the rtnl_lock.  This
call is also under the rtnl_lock protection, its just taken when the event
notification is made (thats why we call __netpoll_cleanup instead of
netpoll_cleanup).  The target_list_lock just protects the integrity of the
netconsole_target list.  If someone disables a netconsole via configfs, they'll
block on the rtnl_lock.  Since no path through configfs takes the
target_list_lock and rtnl (via netpoll_cleanup) in any nested fashion, we're
safe from deadlock.

Best
Neil

--
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