From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1755276Ab3I3Xo2 (ORCPT ); Mon, 30 Sep 2013 19:44:28 -0400 Received: from webmail.solarflare.com ([12.187.104.25]:39322 "EHLO webmail.solarflare.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754512Ab3I3Xo1 (ORCPT ); Mon, 30 Sep 2013 19:44:27 -0400 Message-ID: <1380584663.2562.27.camel@bwh-desktop.uk.level5networks.com> Subject: Re: IRQ affinity notifiers vs RT From: Ben Hutchings To: Sebastian Andrzej Siewior CC: Steven Rostedt , "Alexandra N. Kossovsky" , linux-kernel , Thomas Gleixner Date: Tue, 1 Oct 2013 00:44:23 +0100 In-Reply-To: <52403B0B.1060305@linutronix.de> References: <1374708683.1732.38.camel@bwh-desktop.uk.level5networks.com> <1377898155.4629.11.camel@bwh-desktop.uk.level5networks.com> <52403B0B.1060305@linutronix.de> Organization: Solarflare Content-Type: text/plain; charset="UTF-8" X-Mailer: Evolution 3.6.4 (3.6.4-3.fc18) MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Originating-IP: [10.17.20.137] X-TM-AS-Product-Ver: SMEX-10.0.0.1412-7.000.1014-20184.005 X-TM-AS-Result: No--40.180000-0.000000-31 X-TM-AS-User-Approved-Sender: Yes X-TM-AS-User-Blocked-Sender: No Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Mon, 2013-09-23 at 14:58 +0200, Sebastian Andrzej Siewior wrote: > On 08/30/2013 11:29 PM, Ben Hutchings wrote: > > Sebastian, I saw you came up with a fix for this but apparently without > > seeing my earlier message: > > Yes Ben, I haven't seen it. If I was on Cc I very sorry for overlooking > it. You weren't, as I didn't realise you were maintaining the RT patch set then. > > On Thu, 2013-07-25 at 00:31 +0100, Ben Hutchings wrote: > > >> Workqueue code uses spin_lock_irq() on the workqueue lock, which with > >> PREEMPT_RT enabled doesn't actually block IRQs. > >> > >> In 3.6, the irq_cpu_rmap functions relies on a workqueue flush to finish > >> any outstanding notifications before freeing the cpu_rmap that they use. > >> This won't be reliable if the notification is scheduled after releasing > >> the irq_desc lock. > >> > >> However, following commit 896f97ea95c1 ('lib: cpu_rmap: avoid flushing > >> all workqueues') in 3.8, I think that it is sufficient to do only > >> kref_get(&desc->affinity_notify->kref) in __irq_set_affinity_locked() > >> and then call schedule_work() in irq_set_affinity() after releasing the > >> lock. Something like this (untested): > > > > Does the following make sense to you? > > This was suggested by the original submitter on rt-users@v.k.o (Joe > Korty) where I've been made aware of this for the first time. This okay > except for the part where the workqueue is not scheduled if calling by > the __ function (i.e. the mips case). If I read the code correctly, the > CPU goes offline and affinity change should be updated / users notified > and this is not the case with this patch. > > It is a valid question why only one mips SoC needs the function. If you > look at commit 0c3263870f ("MIPS: Octeon: Rewrite interrupt handling > code.") you can see that tglx himself made this adjustment so it might > be valid :) Therefore I assume that we may get more callers of this > function and the workqueue should be executed and I made something > simple that works on RT. Right, but it looks quite strange to have this thread just for a (probably) quite rare event. Maybe some day someone will work out how to make Octeon's IRQ management less unusual and then you can use the simpler approach for RT... Ben. -- Ben Hutchings, Staff Engineer, Solarflare Not speaking for my employer; that's the marketing department's job. They asked us to note that Solarflare product names are trademarked.