netdev - Re: hung task in mac80211

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: <20170906152709.673f230d@elisabeth>
Date:   Wed, 6 Sep 2017 15:27:09 +0200
From:   Stefano Brivio <sbrivio@...hat.com>
To:     Johannes Berg <johannes@...solutions.net>
Cc:     Matteo Croce <mcroce@...hat.com>, linux-wireless@...r.kernel.org,
        netdev@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: hung task in mac80211

On Wed, 06 Sep 2017 15:21:00 +0200
Johannes Berg <johannes@...solutions.net> wrote:

> On Wed, 2017-09-06 at 15:19 +0200, Stefano Brivio wrote:
> > On Wed, 06 Sep 2017 14:48:35 +0200
> > Johannes Berg <johannes@...solutions.net> wrote:
> >   
> > > I'll look in a bit - but
> > >   
> > > > +			mutex_unlock(&sta->ampdu_mlme.mtx);
> > > >  			___ieee80211_stop_rx_ba_session(
> > > >  				sta, tid, WLAN_BACK_RECIPIENT,
> > > >  				WLAN_REASON_QSTA_TIMEOUT,
> > > > true);    
> > > 
> > > This already has three underscores so shouldn't drop.  
> > 
> > Right, of course.
> >   
> > > [...]  
> > > > +			mutex_unlock(&sta->ampdu_mlme.mtx);
> > > >  			__ieee80211_start_rx_ba_session(sta, 0,
> > > > 0,
> > > > 0, 1, tid,    
> > > 
> > > maybe this one needs a ___ version then?  
> > 
> > Either that, or as it's a single call, perhaps just the following?
> > Matter of taste I guess...  
> 
> I don't think it's a matter of taste - for me, in principle, dropping
> locks for small sections of code where the larger section holds it is a
> bug waiting to happen. It may (may, I don't even know) be OK here, but
> in general it's something to avoid.

Yes, that was based on the assumption that the initial part of
__ieee80211_start_rx_ba_session() can't really affect the AMPDU
state-machine in any way.

But sure, one small change there in the future and the assumption
doesn't hold anymore.


--
Stefano