From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eric Dumazet Subject: Re: Fw: [Bug 54601] New: PPP regression on 3.2.39 stop work need restart server ont version 3.2.21 not found this problem. Date: Mon, 04 Mar 2013 16:05:15 -0800 Message-ID: <1362441915.15793.107.camel@edumazet-glaptop> References: <20130301084705.4169c429@nehalam.linuxnetplumber.net> <20130304224837.GX9079@decadent.org.uk> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 7bit Cc: Stephen Hemminger , netdev@vger.kernel.org, micron10@gmail.com, Eric Dumazet To: Ben Hutchings Return-path: Received: from mail-pb0-f41.google.com ([209.85.160.41]:52634 "EHLO mail-pb0-f41.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1758329Ab3CEAFT (ORCPT ); Mon, 4 Mar 2013 19:05:19 -0500 Received: by mail-pb0-f41.google.com with SMTP id um15so3618222pbc.0 for ; Mon, 04 Mar 2013 16:05:18 -0800 (PST) In-Reply-To: <20130304224837.GX9079@decadent.org.uk> Sender: netdev-owner@vger.kernel.org List-ID: On Mon, 2013-03-04 at 22:48 +0000, Ben Hutchings wrote: > On Fri, Mar 01, 2013 at 08:47:05AM -0800, Stephen Hemminger wrote: > > > > > > Begin forwarded message: > > > > Date: Fri, 1 Mar 2013 01:24:14 -0800 > > From: "bugzilla-daemon@bugzilla.kernel.org" > > To: "stephen@networkplumber.org" > > Subject: [Bug 54601] New: PPP regression on 3.2.39 stop work need restart server ont version 3.2.21 not found this problem. > > > > > > https://bugzilla.kernel.org/show_bug.cgi?id=54601 > > > > Summary: PPP regression on 3.2.39 stop work need restart server > > ont version 3.2.21 not found this problem. > > Product: Networking > > Version: 2.5 > > Kernel Version: 3.2.39 > > Platform: All > > OS/Version: Linux > > Tree: Mainline > > Status: NEW > > Severity: high > > Priority: P1 > > Component: Other > > AssignedTo: shemminger@linux-foundation.org > > ReportedBy: micron10@gmail.com > > Regression: Yes > > > > > > Hi > > watch a problem with PPP after upgrade to version 3.2.39 ppp service stop work > > 2 hours after start with this crash : > > > > [79619.180026] unregister_netdevice: waiting for ppp618 to become free. Usage > > count = 562 > [...] > > [79862.540040] unregister_netdevice: waiting for ppp618 to become free. Usage > > count = 30 > > [79872.680037] unregister_netdevice: waiting for ppp618 to become free. Usage > > count = 9 > [...] > > So there are many references left to the device, which slowly go away > (very slowly, about 2 per second). Maybe there are skbs in a queue somewhere > that still refer to the PPP device. > > Since this a regression between 3.2.21 and 3.2.29, it might have been > introduced by: > > commit 73a3346556281fd56f39f0a9475249e5039d8807 > Author: Eric Dumazet > Date: Thu Jun 14 06:42:44 2012 +0000 > > net: remove skb_orphan_try() > > commit 62b1a8ab9b3660bb820d8dfe23148ed6cda38574 upstream. > > But this is a bit of a guess. And I don't think that this change > would do more than expose/exacerbate a bug that already existed. > > Ben. > I wonder which component could possibly take one dev refcnt for every packet. maybe xdst_queue_output in net/xfrm/xfrm_policy.c ? Then maybe the cleanup is not properly done at device dismantle