From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jarek Poplawski Subject: Re: [PATCH 1/1] Use RCU for the UDP hash lock Date: Fri, 26 Sep 2008 21:50:24 +0200 Message-ID: <20080926195024.GA2548@ami.dom.local> References: <20080926031832.GA27289@minyard.local> <20080926040917.GB7209@linux.vnet.ibm.com> <48DCE874.8070704@acm.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: paulmck@linux.vnet.ibm.com, Linux Kernel , netdev@vger.kernel.org, shemminger@vyatta.com To: Corey Minyard Return-path: Received: from nf-out-0910.google.com ([64.233.182.186]:19946 "EHLO nf-out-0910.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751621AbYIZTt6 (ORCPT ); Fri, 26 Sep 2008 15:49:58 -0400 Received: by nf-out-0910.google.com with SMTP id d3so423512nfc.21 for ; Fri, 26 Sep 2008 12:49:56 -0700 (PDT) Content-Disposition: inline In-Reply-To: <48DCE874.8070704@acm.org> Sender: netdev-owner@vger.kernel.org List-ID: On Fri, Sep 26, 2008 at 08:49:40AM -0500, Corey Minyard wrote: ... > I don't believe udp_disconnect() releases the socket. > sk_common_release() seems to be the place where the refcount is > decremented. But wherever it is done, it would have to be after the > unhash. ...Which, BTW, could be repeated then, so hlist_del_init() matters yet. (Another sign of such possibility should be this "if ()" in udp_lib_unhash().) Jarek P.