netdev - Re: [PATCH] Make sure GRED backlog does not get screwed up.

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>] [day] [month] [year] [list]

Date:	Mon, 08 Apr 2013 06:26:44 -0400
From:	Jamal Hadi Salim <jhs@...atatu.com>
To:	Magnus Gille <mgille@...il.com>
CC:	netdev@...r.kernel.org
Subject: Re: [PATCH] Make sure GRED backlog does not get screwed up.

Hi,

please use my proper address (in the From:) for faster response.

On 13-04-02 02:32 PM, Magnus Gille wrote:
> First patch, comments, feedback, suggestions are welcome. Before this patch
> we could end up with a 4GB backlog on certain DPs without any packets. The
> patch is created against net-next.
>
> If a packet is enqueued on a GRED qdisc before the DPs are setup the packet
> will not be accounted for in the DP backlog queue.

In such a case, the packet will end up in the default queue

>   When the packet later
> gets
> dequeued from the DP queue it will subtract the length from the DP backlog
> but,
> since the packet was never accounted for we now wrap the backlog around.
> Fixed
> this by making sure we never subtract more than what's present in the
> backlog.

When no DPs set: if the packet was enqueued in a default queue,
at dequeue it will be pulled from the default queue.
What am i missing?

cheers,
jamal

--
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