From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S263557AbUEKTzV (ORCPT ); Tue, 11 May 2004 15:55:21 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S263574AbUEKTzV (ORCPT ); Tue, 11 May 2004 15:55:21 -0400 Received: from mailgate1b.savvis.net ([216.91.182.6]:39329 "EHLO mailgate1b.savvis.net") by vger.kernel.org with ESMTP id S263557AbUEKTzR (ORCPT ); Tue, 11 May 2004 15:55:17 -0400 From: "Dan A. Dickey" Reply-To: dan.dickey@savvis.net Organization: WAM!NET a Division of SAVVIS, Inc. To: "David S. Miller" Subject: Re: Sock leak in net/ipv4/af_inet.c - 2.4.26 Date: Tue, 11 May 2004 14:55:11 -0500 User-Agent: KMail/1.6.2 Cc: , , References: <20040511115934.0c591667.davem@redhat.com> In-Reply-To: <20040511115934.0c591667.davem@redhat.com> MIME-Version: 1.0 Content-Disposition: inline Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Message-Id: <200405111455.11935.dan.dickey@savvis.net> X-ECS-MailScanner: No virus is found Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Tuesday 11 May 2004 13:59, David S. Miller wrote: > The sk_free() should occur when the final sock_put() call brings the count > down to zero, then the socket destroy function is called and the eventual > sk_free() occurs there. Yes, I'm figuring this out. I think my earlier report of this problem was a bit premature. I'm refining my debug code and will let you know what I find. -Dan -- Dan A. Dickey dan.dickey@savvis.net