netdev - Re: [PATCH 1/1] tg3: Avoid Send BD corruption

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: <1221175168.12785.40.camel@HP1>
Date:	Thu, 11 Sep 2008 16:19:28 -0700
From:	"Michael Chan" <mchan@...adcom.com>
To:	"David Miller" <davem@...emloft.net>
cc:	"Matthew Carlson" <mcarlson@...adcom.com>,
	"netdev@...r.kernel.org" <netdev@...r.kernel.org>,
	"andy@...yhouse.net" <andy@...yhouse.net>
Subject: Re: [PATCH 1/1] tg3: Avoid Send BD corruption


On Thu, 2008-09-11 at 16:01 -0700, David Miller wrote:
> From: "Michael Chan" <mchan@...adcom.com>
> Date: Thu, 11 Sep 2008 15:55:58 -0700
> 
> > Send BD corruption is quite serious as it can cause tg3_tx() to crash.
> > We've seen a number of similar crashes over the years caused by
> > re-ordered IOs and the nr_frags getting modified by HTB.
> >
> > Since regression/security/oops fixes are allowed, shouldn't this qualify
> > since it prevents a crash in tg3_tx()?
> 
> It's at best borderline, since tg3_tx() has reordering detection logic
> which will reset and recover the card.

The re-ordering detection logic will allow corruption to happen once
only.  It will then enable flushing of IOs and if it detects corruption
again, it will hit BUG().  That's why it wasn't able to prevent the
crash caused by HTB.

> 
> Until I see a real user trigger this I'm still leaning towards no.
> 
> Otherwise any driver maintainer can give me this spiel to get their
> changes installed outside of the merge window, and that subverts the
> entire purpose of the rules.
> 

OK, we'll resubmit this for net-next then.  But I still think this one
should qualify.  We do internal testing to find as many problems as
possible so that real users don't see too many of them.


--
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