From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752821AbbETH1r (ORCPT ); Wed, 20 May 2015 03:27:47 -0400 Received: from mx1.redhat.com ([209.132.183.28]:56530 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750966AbbETH1n (ORCPT ); Wed, 20 May 2015 03:27:43 -0400 Date: Wed, 20 May 2015 09:27:15 +0200 From: Milos Vyletel To: Ingo Molnar Cc: Jiri Olsa , Peter Zijlstra , Paul Mackerras , Ingo Molnar , Arnaldo Carvalho de Melo , Masami Hiramatsu , Namhyung Kim , Jiri Olsa , He Kuang , Adrian Hunter , "open list:PERFORMANCE EVENT..." Subject: Re: [PATCH 2/2] perf/tools: put new buildid locks to use Message-ID: <20150520072715.GA2194@lenivo.usersys.redhat.com> References: <1431588493-1474-1-git-send-email-milos@redhat.com> <1431588493-1474-3-git-send-email-milos@redhat.com> <20150514104059.GA27771@gmail.com> <20150514113821.GB1313@krava.redhat.com> <20150514173807.GA19338@gmail.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20150514173807.GA19338@gmail.com> Organization: Red Hat Czech s.r.o. User-Agent: Mutt/1.5.23.1-rc1 (2014-03-12) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Thu, May 14, 2015 at 07:38:08PM +0200, Ingo Molnar wrote: > > * Milos Vyletel wrote: > > > On Thu, May 14, 2015 at 01:38:21PM +0200, Jiri Olsa wrote: > > > > > On Thu, May 14, 2015 at 12:40:59PM +0200, Ingo Molnar wrote: > > > > > > > > * Milos Vyletel wrote: > > > > > > > > > Use new read/write locks when accesing buildid directory on places where > > > > > we may race if multiple instances are run simultaneously. > > > > > > > > Dunno, this will create locking interaction between multiple instances > > > > of perf - hanging each other, etc. > > > > > > > > And it seems unnecessary: the buildid hierarchy is already spread out. > > > > What kind of races might there be? > > > > > > there was just recently one fixed by commit: > > > 0635b0f71424 perf tools: Fix race in build_id_cache__add_s() > > > > > > havent checked the final patch yet, but the idea is to > > > protect us from similar bugs > > > > right. on top of race with EEXIST couple more are possible (EMLINK, > > ENOSPC, EDQUOT, ENOMEM... the only way to prevent them all is to > > lock this kind of operations and make sure we run one at a time. > > Yeah, so the race pointed out in 0635b0f71424 can be (and should be) > fixed without locking: > > - first create the file under a process-private name under > ~/.debug/tmp/ if the target does not exist yet > > - then fully fill it in with content > > - then link(2) it to the public target name, which VFS operation is > atomic and may fail safely: at which point it got already created > by someone else. > > - finally unlink() the private instance name and the target will now > be the only instance left: either created by us, or by some other > perf instance in the rare racy case. > > Since all of ~/.debug is on the same filesystem this should work fine. > > Beyond avoiding locking this approach has another advantage: it's > transaction safe, so a crashed/interrupted perf instance won't corrupt > the debug database, it will only put fully constructed files into the > public build-id namespace. It at most leaves a stale private file > around in ~/.debug/tmp/. > > Really, we should be following the example of Git, which is using a > similar append-mostly flow to handle data, and generally avoids file > locking as much as possible - which is a whole new can of worms. > Thanks Ingo. I'll take a look at this later this week. Milos