netdev - Re: [PATCH net-next] alx: use fine-grained locking instead of RTNL

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: <8e2a8567d79141f8f5cb180be656efb378dcfee8.camel@sipsolutions.net>
Date:   Fri, 14 May 2021 13:22:56 +0200
From:   Johannes Berg <johannes@...solutions.net>
To:     Nikolai Zhubr <zhubr.2@...il.com>
Cc:     netdev@...r.kernel.org
Subject: Re: [PATCH net-next] alx: use fine-grained locking instead of RTNL

On Fri, 2021-05-14 at 13:59 +0300, Nikolai Zhubr wrote:
> Hi Johannes,
> 
> 12.05.2021 13:19, Johannes Berg:
> > In the alx driver, all locking depended on the RTNL, but
> [...]
> > @@ -232,7 +240,7 @@ static int alx_set_pauseparam(struct net_device *netdev,
> >   	if (pause->autoneg)
> >   		fc |= ALX_FC_ANEG;
> > 
> > -	ASSERT_RTNL();
> > +	mutex_lock(&alx->mtx);
> > 
> >   	/* restart auto-neg for auto-mode */
> >   	if (hw->adv_cfg&  ADVERTISED_Autoneg) {
> > @@ -254,6 +262,7 @@ static int alx_set_pauseparam(struct net_device *netdev,
> >   		alx_cfg_mac_flowcontrol(hw, fc);
> > 
> >   	hw->flowctrl = fc;
> > +	mutex_unlock(&alx->mtx);
> > 
> >   	return 0;
> >   }
> 
> Isn't this fragment missing a mutex_unlock(&alx->mtx) for the "return 
> err" codepath in the middle? I'm not sure, its like very suspicious, 
> please have a look.

I clearly missed it, somebody already picked up on it:

https://lore.kernel.org/r/20210514082405.91011-1-pulehui@huawei.com

johannes