From mboxrd@z Thu Jan 1 00:00:00 1970 From: Peter Zijlstra Subject: Re: [PATCH 6/9] locking/qrwlock: allow architectures to hook in to contended paths Date: Wed, 8 Jul 2015 12:06:57 +0200 Message-ID: <20150708100657.GD3644@twins.programming.kicks-ass.net> References: <1436289865-2331-1-git-send-email-will.deacon@arm.com> <1436289865-2331-7-git-send-email-will.deacon@arm.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Received: from bombadil.infradead.org ([198.137.202.9]:58268 "EHLO bombadil.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S932693AbbGHKHF (ORCPT ); Wed, 8 Jul 2015 06:07:05 -0400 Content-Disposition: inline In-Reply-To: <1436289865-2331-7-git-send-email-will.deacon@arm.com> Sender: linux-arch-owner@vger.kernel.org List-ID: To: Will Deacon Cc: linux-arch@vger.kernel.org, linux-arm-kernel@lists.infradead.org, Waiman.Long@hp.com, mingo@redhat.com On Tue, Jul 07, 2015 at 06:24:22PM +0100, Will Deacon wrote: > When contended, architectures may be able to reduce the polling overhead > in ways which aren't expressible using a simple relax() primitive. > > This patch allows architectures to override the use of > cpu_relax_lowlatency() in the qrwlock code and also implement their own > unlock macros in case explicit signalling is required to wake up a > `relaxed' CPU spinning on an unlock event. No real objection, but could you do this _after_ you've converted AARGH64 to use the normal qrwlock, such that you can show the benefit with numbers? From mboxrd@z Thu Jan 1 00:00:00 1970 From: peterz@infradead.org (Peter Zijlstra) Date: Wed, 8 Jul 2015 12:06:57 +0200 Subject: [PATCH 6/9] locking/qrwlock: allow architectures to hook in to contended paths In-Reply-To: <1436289865-2331-7-git-send-email-will.deacon@arm.com> References: <1436289865-2331-1-git-send-email-will.deacon@arm.com> <1436289865-2331-7-git-send-email-will.deacon@arm.com> Message-ID: <20150708100657.GD3644@twins.programming.kicks-ass.net> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Tue, Jul 07, 2015 at 06:24:22PM +0100, Will Deacon wrote: > When contended, architectures may be able to reduce the polling overhead > in ways which aren't expressible using a simple relax() primitive. > > This patch allows architectures to override the use of > cpu_relax_lowlatency() in the qrwlock code and also implement their own > unlock macros in case explicit signalling is required to wake up a > `relaxed' CPU spinning on an unlock event. No real objection, but could you do this _after_ you've converted AARGH64 to use the normal qrwlock, such that you can show the benefit with numbers?