netdev - Re: [PATCH net-next] tcp: avoid retransmits of TCP packets hanging in host queues

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

Message-Id: <20140422.212825.238523997379918414.davem@davemloft.net>
Date:	Tue, 22 Apr 2014 21:28:25 -0400 (EDT)
From:	David Miller <davem@...emloft.net>
To:	eric.dumazet@...il.com
Cc:	ycheng@...gle.com, ncardwell@...gle.com, netdev@...r.kernel.org
Subject: Re: [PATCH net-next] tcp: avoid retransmits of TCP packets hanging
 in host queues

From: Eric Dumazet <eric.dumazet@...il.com>
Date: Sun, 20 Apr 2014 17:58:17 -0700

> From: Eric Dumazet <edumazet@...gle.com>
> 
> In commit 0e280af026a5 ("tcp: introduce TCPSpuriousRtxHostQueues SNMP
> counter") we added a logic to detect when a packet was retransmitted
> while the prior clone was still in a qdisc or driver queue.
> 
> We are now confident we can do better, and catch the problem before
> we fragment a TSO packet before retransmit, or in TLP path.
> 
> This patch fully exploits the logic by simply canceling the spurious
> retransmit.
> Original packet is in a queue and will eventually leave the host.
> 
> This helps to avoid network collapses when some events make the RTO
> estimations very wrong, particularly when dealing with huge number of
> sockets with synchronized blast.
> 
> Signed-off-by: Eric Dumazet <edumazet@...gle.com>
> Signed-off-by: Neal Cardwell <ncardwell@...gle.com>
> Signed-off-by: Yuchung Cheng <ycheng@...gle.com>

Great idea, looks good, applied.
--
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