From mboxrd@z Thu Jan 1 00:00:00 1970 From: Hannes Frederic Sowa Subject: Re: [PATCH] ipv6: Don't depend on per socket memory for neighbour discovery messages Date: Wed, 28 Aug 2013 15:31:00 +0200 Message-ID: <20130828133100.GA12353@order.stressinduktion.org> References: <361d4275efc79e3048191306df25ad003ed497d7.1377644830.git.tgraf@suug.ch> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Cc: davem@davemloft.net, netdev@vger.kernel.org, Eric Dumazet To: Thomas Graf Return-path: Received: from order.stressinduktion.org ([87.106.68.36]:51535 "EHLO order.stressinduktion.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751896Ab3H1NbD (ORCPT ); Wed, 28 Aug 2013 09:31:03 -0400 Content-Disposition: inline In-Reply-To: <361d4275efc79e3048191306df25ad003ed497d7.1377644830.git.tgraf@suug.ch> Sender: netdev-owner@vger.kernel.org List-ID: On Wed, Aug 28, 2013 at 01:07:25AM +0200, Thomas Graf wrote: > Allocating skbs when sending out neighbour discovery messages > currently uses sock_alloc_send_skb() based on a per net namespace > socket and thus share a socket wmem buffer space. > > If a netdevice is temporarily unable to transmit due to carrier > loss or for other reasons, the queued up ndisc messages will cosnume > all of the wmem space and will thus prevent from any more skbs to > be allocated even for netdevices that are able to transmit packets. > > The number of neighbour discovery messages sent is very limited, > simply use alloc_skb() and don't depend on any socket wmem space any > longer. > > This patch has orginally been posted by Eric Dumazet in a modified > form. > > Signed-off-by: Thomas Graf > Cc: Eric Dumazet Acked-by: Hannes Frederic Sowa This fixes a real bug (I also had this issue) and thus would propose it for stable, too. Thanks!