From mboxrd@z Thu Jan 1 00:00:00 1970 From: Andrew Morton Subject: Re: [Bugme-new] [Bug 18492] New: kernel softirq warning on boot Date: Tue, 14 Sep 2010 13:57:30 -0700 Message-ID: <20100914135730.8af6bc76.akpm@linux-foundation.org> References: <20100914120738.6a658f69.akpm@linux-foundation.org> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Cc: netdev@vger.kernel.org, bugzilla-daemon@bugzilla.kernel.org, bugme-daemon@bugzilla.kernel.org, Herbert Xu To: Michal Suchanek Return-path: Received: from smtp1.linux-foundation.org ([140.211.169.13]:37943 "EHLO smtp1.linux-foundation.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752294Ab0INU6e (ORCPT ); Tue, 14 Sep 2010 16:58:34 -0400 In-Reply-To: Sender: netdev-owner@vger.kernel.org List-ID: On Tue, 14 Sep 2010 22:45:19 +0200 Michal Suchanek wrote: > > rcu_read_unlock_bh() in netpoll_rx(), added by "netpoll: Fix RCU > > usage". > > Do you mean de85d99eb7b595f6751550184b94c1e2f74a828b? yup. > I am not sure reverting just that would work, nor that the kernel > around that point is in shape for trying on a system I would want to > still use. It's way back in history for a kernel which is just to be > released. No, we wouldn't want to revert it - it fixes stuff. It could be that we just don't need the _bh locking, if local irqs are reliably disabled. But I didn't really look.