From mboxrd@z Thu Jan 1 00:00:00 1970 From: Andi Kleen Subject: Re: TCP always advertises zero window. Date: Mon, 25 Oct 2010 20:18:14 +0200 Message-ID: <87tyka3zmh.fsf@basil.nowhere.org> References: <4CBFE924.2060103@gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: "netdev\@vger.kernel.org" To: Li Yu Return-path: Received: from one.firstfloor.org ([213.235.205.2]:44098 "EHLO one.firstfloor.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751542Ab0JYSSY (ORCPT ); Mon, 25 Oct 2010 14:18:24 -0400 In-Reply-To: <4CBFE924.2060103@gmail.com> (Li Yu's message of "Thu, 21 Oct 2010 15:17:56 +0800") Sender: netdev-owner@vger.kernel.org List-ID: Li Yu writes: > > According to source code, only MTU probing success and receive some non-zero length of L7 payload could grow tcp_sock->rcv_ssthresh. Because of we turn off MTU probing and TCP only received some zero-window probe from another end, so it seem that we have not any chance to update tcp_sock->rcv_ssthresh at all, so a dead loop come here. The question is why are you turning off MTU probing? -Andi