From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753237Ab3AFWKy (ORCPT ); Sun, 6 Jan 2013 17:10:54 -0500 Received: from Mycroft.westnet.com ([216.187.52.7]:51031 "EHLO mycroft.westnet.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753085Ab3AFWKw (ORCPT ); Sun, 6 Jan 2013 17:10:52 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Message-ID: <20713.61807.385486.594545@quad.stoffel.home> Date: Sun, 6 Jan 2013 16:49:35 -0500 From: "John Stoffel" To: Willy Tarreau Cc: Eric Dumazet , netdev@vger.kernel.org, linux-kernel@vger.kernel.org Subject: Re: Major network performance regression in 3.7 In-Reply-To: <20130106193410.GL16031@1wt.eu> References: <1357484342.6919.61.camel@edumazet-glaptop> <20130106155123.GB16031@1wt.eu> <1357490393.6919.267.camel@edumazet-glaptop> <20130106164416.GF16031@1wt.eu> <1357492255.6919.336.camel@edumazet-glaptop> <20130106173543.GB22432@1wt.eu> <1357497541.6919.519.camel@edumazet-glaptop> <1357497809.6919.529.camel@edumazet-glaptop> <1357498276.6919.547.camel@edumazet-glaptop> <1357498815.6919.570.camel@edumazet-glaptop> <20130106193410.GL16031@1wt.eu> X-Mailer: VM 8.1.2 under 23.2.1 (x86_64-pc-linux-gnu) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org >>>>> "Willy" == Willy Tarreau writes: Willy> On Sun, Jan 06, 2013 at 11:00:15AM -0800, Eric Dumazet wrote: >> On Sun, 2013-01-06 at 10:51 -0800, Eric Dumazet wrote: >> > > >> > > (sd->len is usually 4096, which is expected, but sd->total_len value is >> > > huge in your case, so we always set the flag in fs/splice.c) >> > >> > I am testing : >> > >> > if (sd->len < sd->total_len && pipe->nrbufs > 1) >> > more |= MSG_SENDPAGE_NOTLAST; >> > >> >> Yes, this should fix the problem : >> >> If there is no following buffer in the pipe, we should not set NOTLAST. >> >> diff --git a/fs/splice.c b/fs/splice.c >> index 8890604..6909d89 100644 >> --- a/fs/splice.c >> +++ b/fs/splice.c >> @@ -696,8 +696,10 @@ static int pipe_to_sendpage(struct pipe_inode_info *pipe, >> return -EINVAL; >> >> more = (sd->flags & SPLICE_F_MORE) ? MSG_MORE : 0; >> - if (sd->len < sd->total_len) >> + >> + if (sd->len < sd->total_len && pipe->nrbufs > 1) >> more |= MSG_SENDPAGE_NOTLAST; >> + >> return file->f_op->sendpage(file, buf->page, buf->offset, sd-> len, &pos, more); >> } Willy> OK it works like a charm here now ! I can't break it anymore, so it Willy> looks like you finally got it ! It's still broken, there's no comments in the code to explain all this magic to mere mortals! *grin* John