From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752564AbaJFJIi (ORCPT ); Mon, 6 Oct 2014 05:08:38 -0400 Received: from mga11.intel.com ([192.55.52.93]:15564 "EHLO mga11.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752406AbaJFJIg (ORCPT ); Mon, 6 Oct 2014 05:08:36 -0400 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.04,662,1406617200"; d="scan'208";a="601126757" From: Alexander Shishkin To: Peter Zijlstra Cc: Ingo Molnar , linux-kernel@vger.kernel.org, Robert Richter , Frederic Weisbecker , Mike Galbraith , Paul Mackerras , Stephane Eranian , Andi Kleen , kan.liang@intel.com Subject: Re: [PATCH v4 02/22] perf: Add AUX area to ring buffer for raw data streams In-Reply-To: <20140908131230.GE6758@twins.programming.kicks-ass.net> References: <1408538179-792-1-git-send-email-alexander.shishkin@linux.intel.com> <1408538179-792-3-git-send-email-alexander.shishkin@linux.intel.com> <20140908070242.GS19379@twins.programming.kicks-ass.net> <87lhpuuztj.fsf@ashishki-desk.ger.corp.intel.com> <20140908113422.GD6758@twins.programming.kicks-ass.net> <87a96auv9s.fsf@ashishki-desk.ger.corp.intel.com> <20140908131230.GE6758@twins.programming.kicks-ass.net> User-Agent: Notmuch/0.17+49~gaa57e9d (http://notmuchmail.org) Emacs/24.3.1 (x86_64-pc-linux-gnu) Date: Mon, 06 Oct 2014 12:08:19 +0300 Message-ID: <87k34do9a4.fsf@ashishki-desk.ger.corp.intel.com> MIME-Version: 1.0 Content-Type: text/plain Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Peter Zijlstra writes: > On Mon, Sep 08, 2014 at 03:55:11PM +0300, Alexander Shishkin wrote: > >> Fair enough. Then I'd like to disable the ACTIVE ones before freeing AUX >> stuff and then re-enabling them since perf_event_{en,dis}able() already >> provide the convenient cross-cpu calls, which would also avoid >> concurrency between pmu::{add,del} callbacks and this unmap path. Makes >> sense? > > But why? The buffer stuff is RCU freed, so if the hardware observes > pages and does get_page_unless_zero() on them its good. The memory will > not be freed from underneath the hardware writer because of the > get_page(). > > Then when the buffer is full and we 'swap', we'll find there is no next > buffer. At that point we can not provide a new buffer, effectively > stopping the hardware writes and release the old buffer, freeing the > memory. There are several problems with this. Firstly, aux buffers can be quite large, which means that we have to do get_page() on thousands of pages on every pmu::add, which is a hot path and free_page() again in pmu::del. Secondly, all the sg bookkeeping that the driver keeps in aux_priv needs to be refcounted. Right now, in the mmap_close path we just free everything. But if we want to free the aux_pages in pmu::del, we need to keep a list of these pages still around after mmap_close() and same goes for the actual sg tables. I can see a way of doing that on the ring buffer side (as opposed to the driver side), but are you quite sure we should go down this road? Regards, -- Alex