From mboxrd@z Thu Jan 1 00:00:00 1970 From: Luiz Capitulino Subject: Re: [PATCH RESEND -rt] mm: perform lru_add_drain_all() remotely Date: Thu, 12 May 2016 15:52:02 -0400 Message-ID: <20160512155202.46666531@redhat.com> References: <20160509105037.1555d0e0@redhat.com> <20160512084230.GA19035@linutronix.de> <20160512095149.2a6f7357@redhat.com> <20160512144943.GD19035@linutronix.de> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Cc: linux-rt-users@vger.kernel.org, riel@redhat.com, tglx@linutronix.de, srostedt@redhat.com, williams@redhat.com To: Sebastian Andrzej Siewior Return-path: Received: from mx1.redhat.com ([209.132.183.28]:47034 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753156AbcELTwE (ORCPT ); Thu, 12 May 2016 15:52:04 -0400 In-Reply-To: <20160512144943.GD19035@linutronix.de> Sender: linux-rt-users-owner@vger.kernel.org List-ID: On Thu, 12 May 2016 16:49:43 +0200 Sebastian Andrzej Siewior wrote: > * Luiz Capitulino | 2016-05-12 09:51:49 [-0400]: > > >That's correct. This is an -RT patch for two reasons: > > > > 1. This solution builds on top of the per-cpu locks API implemented > > by -RT and already in place in the pagevecs drain code > > > > 2. As SCHED_FIFO is the norm in -RT, the problem is more common there > > more common so not RT-only. Would you people prefer this for mainline or > would you rather something else in case this needs to be fixed > mainline, too? To have this for mainline we'd have to forward-port local_locks. It's doable, but I think it's a bit too much only to solve this problem. The alternative, which is what I had in mind, was to forward-port this change to mainline when local_locks gets merged upstream. But I guess this will only happen when sleeping spinlocks are merged?