From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eric Dumazet Subject: Re: 2.6.34-rc1: rcu lockdep bug? Date: Fri, 12 Mar 2010 12:11:16 +0100 Message-ID: <1268392276.3141.4.camel@edumazet-laptop> References: <20100311134556.GA6344@linux.vnet.ibm.com> <20100311161751.GA3804@hack> <2375c9f91003112356g1b4164e4pb5f63f0e0e2f310a@mail.gmail.com> <20100312.000705.225033546.davem@davemloft.net> <2375c9f91003120059g771d162fxefc21beb2ab17b4d@mail.gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: David Miller , paulmck@linux.vnet.ibm.com, peterz@infradead.org, linux-kernel@vger.kernel.org, netdev@vger.kernel.org To: =?ISO-8859-1?Q?Am=E9rico?= Wang Return-path: Received: from mail-bw0-f209.google.com ([209.85.218.209]:54477 "EHLO mail-bw0-f209.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S932284Ab0CLLLX (ORCPT ); Fri, 12 Mar 2010 06:11:23 -0500 In-Reply-To: <2375c9f91003120059g771d162fxefc21beb2ab17b4d@mail.gmail.com> Sender: netdev-owner@vger.kernel.org List-ID: Le vendredi 12 mars 2010 =C3=A0 16:59 +0800, Am=C3=A9rico Wang a =C3=A9= crit : > On Fri, Mar 12, 2010 at 4:07 PM, David Miller w= rote: > > From: Am=C3=A9rico Wang > > Date: Fri, 12 Mar 2010 15:56:03 +0800 > > > >> Ok, after decoding the lockdep output, it looks like that > >> netif_receive_skb() should call rcu_read_lock_bh() instead of rcu_= read_lock()? > >> But I don't know if all callers of netif_receive_skb() are in soft= irq context. > > > > Normally, netif_receive_skb() is invoked from softirq context. > > > > However, via netpoll it can be invoked essentially from any context= =2E > > > > But, when this happens, the networking receive path makes amends su= ch > > that this works fine. That's what the netpoll_receive_skb() check = in > > netif_receive_skb() is for. That check makes it bail out early if = the > > call to netif_receive_skb() is via a netpoll invocation. > > >=20 > Oh, I see. This means we should call rcu_read_lock_bh() instead. > If Paul has no objections, I will send a patch for this. >=20 Nope, its calling rcu_read_lock() from interrupt context and it should stay as is (we dont need to disable bh, this has a cpu cost)