linux-kernel - Re: 3.3.0-rc5: OOps in dql_completed (Broadcom tg3 driver)

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]

Date:	Fri, 2 Mar 2012 09:23:20 -0800
From:	Tom Herbert <therbert@...gle.com>
To:	Eric Dumazet <eric.dumazet@...il.com>
Cc:	Christoph Lameter <cl@...two.org>, mcarlson@...adcom.com,
	netdev@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: 3.3.0-rc5: OOps in dql_completed (Broadcom tg3 driver)

Hi Christoph, Eric,

Looks like we're hitting BUG_ON(count > dql->num_queued -
dql->num_completed).  This is indicative of mis-accounting occurring
between netdev_sent_queue and netdev_completed_queue.  I don't
immediately see how this could happen here in the tg3_tso_bug path,
this is still using the same calls to transmit and complete as other
paths.  I suppose it's possible that skb_gso_segment is somehow
munging skb->len in segments.

I'll try to reproduce this.

> In case we hit BQL limit in one of the tg3_start_xmit() calls, we should
> 'abort' the following ones, dont you think ?

I'm not sure what you mean Eric.  It should be okay to exceed the BQL
limit up to the point that num_queued rolls over (probably should be a
BUG_ON for that).  This is a much greater number than could ever be
queued to tg3.

Thanks,
Tom

>
> Or maybe thats irrelevant, and only dql_queued() comment is wrong.
>
> Thanks
>
>
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/