From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eric Dumazet Subject: Re: Difficulties to get 1Gbps on be2net ethernet card Date: Wed, 30 May 2012 09:06:27 +0200 Message-ID: <1338361587.2760.94.camel@edumazet-glaptop> References: <1338360536.2760.85.camel@edumazet-glaptop> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 7bit Cc: netdev To: Jean-Michel Hautbois Return-path: Received: from mail-bk0-f46.google.com ([209.85.214.46]:38986 "EHLO mail-bk0-f46.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753743Ab2E3HGc (ORCPT ); Wed, 30 May 2012 03:06:32 -0400 Received: by bkcji2 with SMTP id ji2so3792934bkc.19 for ; Wed, 30 May 2012 00:06:31 -0700 (PDT) In-Reply-To: Sender: netdev-owner@vger.kernel.org List-ID: On Wed, 2012-05-30 at 08:51 +0200, Jean-Michel Hautbois wrote: > 2012/5/30 Eric Dumazet : > > On Wed, 2012-05-30 at 08:28 +0200, Jean-Michel Hautbois wrote: > > > >> If this can help, setting tx queue length to 5000 seems to make the > >> problem disappear. > > > > Then you should have drops at Qdisc layer (before your change to 5000) > > > > tc -s -d qdisc > > > >> I didn't specified it : MTU is 4096, UDP packets are 4000 bytes. > > > > Yes : > qdisc mq 0: dev eth1 root > Sent 5710049154383 bytes 1413544639 pkt (dropped 73078, overlimits 0 > requeues 281540) > backlog 0b 0p requeues 281540 > > Why ? With a 2.6.26 kernel it works well with a tx queue length of 1000. If you send big bursts of packets, then you need a large enough queue. Maybe your kernel is now faster than before and queue fills faster, or TX ring is smaller ? ethtool -g eth0 Note that everybody try to reduce dumb queue sizes because of latencies.