From mboxrd@z Thu Jan 1 00:00:00 1970 From: Evgeniy Polyakov Subject: Re: 2.6.23 WARNING: at kernel/softirq.c:139 local_bh_enable() Date: Fri, 23 Nov 2007 22:15:24 +0300 Message-ID: <20071123191524.GA6273@2ka.mipt.ru> References: <4745DCD7.8070805@simon.arlott.org.uk> <20071123002157.cb27f4a1.akpm@linux-foundation.org> <20071123105518.GA22062@2ka.mipt.ru> <20071123170756.GV19691@waste.org> <20071123175757.GA23991@2ka.mipt.ru> <20071123185943.GZ19691@waste.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Andrew Morton , Simon Arlott , Linux Kernel Mailing List , netdev@vger.kernel.org, Ingo Molnar To: Matt Mackall Return-path: Received: from relay.2ka.mipt.ru ([194.85.82.65]:34460 "EHLO 2ka.mipt.ru" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754595AbXKWTQK (ORCPT ); Fri, 23 Nov 2007 14:16:10 -0500 Content-Disposition: inline In-Reply-To: <20071123185943.GZ19691@waste.org> Sender: netdev-owner@vger.kernel.org List-Id: netdev.vger.kernel.org On Fri, Nov 23, 2007 at 12:59:43PM -0600, Matt Mackall (mpm@selenic.com) wrote: > So I'd be surprised if that was a problem. But I can imagine having > problems for skbs without destructors which run into one of these in > __kfree_skb: > > dst_release > secpath_put > nf_conntrack_put > nf_conntrack_put_reasm > nf_bridge_put > > ..some or all of which assume a softirq context. bridging is ok, others require softirq context. I've sent a patch (the last one should be ok) to guard against xfrm and connection tracking. > > No matter if we are under memory pressure or whatever - it is not > > allowed - a lot of skbs are supposed to be freed in softirq context, > > that is why dev_kfree_skb_any() exists. > > Some skbs we definitely -can- free in irq context. The only ones we > care about are the ones generated by netpoll. If there's a reason you > think netpoll's own skbs can't be freed, please describe it. Only some and to distinguish them we can not use destructor - if it is set (even empty function) it will fire an alarm. > > I think we can drop skbs _without_ destructor from the queue though in > > that conditions given that we actually need only one. > > Huh? Don't mind - friday... I posted a patch (third one should be ok) to fix this issue. -- Evgeniy Polyakov