From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1764620AbYDVOxr (ORCPT ); Tue, 22 Apr 2008 10:53:47 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1756386AbYDVOxh (ORCPT ); Tue, 22 Apr 2008 10:53:37 -0400 Received: from brick.kernel.dk ([87.55.233.238]:27889 "EHLO kernel.dk" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754522AbYDVOxg (ORCPT ); Tue, 22 Apr 2008 10:53:36 -0400 Date: Tue, 22 Apr 2008 16:53:29 +0200 From: Jens Axboe To: Avi Kivity Cc: Peter Zijlstra , linux-arch@vger.kernel.org, linux-kernel@vger.kernel.org, npiggin@suse.de, torvalds@linux-foundation.org Subject: Re: [PATCH 1/11] Add generic helpers for arch IPI function calls Message-ID: <20080422145329.GK12774@kernel.dk> References: <1208851058-8500-1-git-send-email-jens.axboe@oracle.com> <1208851058-8500-2-git-send-email-jens.axboe@oracle.com> <480DACDD.7040108@qumranet.com> <20080422092230.GW12774@kernel.dk> <20080422111403.GX12774@kernel.dk> <1208869254.7115.258.camel@twins> <20080422142543.GG12774@kernel.dk> <480DF861.6000705@qumranet.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <480DF861.6000705@qumranet.com> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Tue, Apr 22 2008, Avi Kivity wrote: > Jens Axboe wrote: > >>Introducing this fallback will make any usage from irq disabled context > >>deadlock prone. > >> > >>I rather like the current interface. > >> > > > >Hmm good point, I'll back that bit out again. > > > > But then you need to fix all callers to handle errors. They should already, there were various error conditions before depending on the arch. I'm aware that some do not, so I'd consider that a general cleanup (that must also be done, agree). > Perhaps we need a new flag for failable async IPIs (or better, API). > I'm pretty sure no one uses s_c_f_s() from irq disabled context on x86, > since it will WARN_ON(irqs_disabled()). That'd work. I tried to keep things really simple here, I have other ideas for cleanups and improvements once this batch goes in. -- Jens Axboe