From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754992AbdBNRVq (ORCPT ); Tue, 14 Feb 2017 12:21:46 -0500 Received: from mga05.intel.com ([192.55.52.43]:33227 "EHLO mga05.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754929AbdBNRVi (ORCPT ); Tue, 14 Feb 2017 12:21:38 -0500 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.35,162,1484035200"; d="scan'208";a="933816668" From: Alexander Shishkin To: Peter Zijlstra Cc: Ingo Molnar , linux-kernel@vger.kernel.org, vince@deater.net, eranian@google.com, Arnaldo Carvalho de Melo , Borislav Petkov , Thomas Gleixner Subject: Re: [PATCH 1/2] perf/x86/intel/pt: Fail event scheduling on conflict with VMX In-Reply-To: <20170214162702.GR6500@twins.programming.kicks-ass.net> References: <20170214132416.26400-1-alexander.shishkin@linux.intel.com> <20170214132416.26400-2-alexander.shishkin@linux.intel.com> <20170214135620.GN6500@twins.programming.kicks-ass.net> <87d1eklpb9.fsf@ashishki-desk.ger.corp.intel.com> <20170214162702.GR6500@twins.programming.kicks-ass.net> User-Agent: Notmuch/0.23.5 (http://notmuchmail.org) Emacs/24.5.1 (x86_64-pc-linux-gnu) Date: Tue, 14 Feb 2017 19:21:34 +0200 Message-ID: <874lzwlmch.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 Tue, Feb 14, 2017 at 06:17:30PM +0200, Alexander Shishkin wrote: >> This is called by pmu::add(), which checks hw.state afterwards and if it >> finds HES_STOPPED, it returns an error, which event_sched_in() captures >> and keeps the event in INACTIVE state. Should I add a comment about it? > > Egads... so what if ->add() succeeds but we then hit this on > ->stop()/->start() due to throttle or period adjust? It will hang there with hw.state==PERF_HES_STOPPED till the next sched_out. But that will be the case anyway if VMXON kicks in while PT is running. > Now I suppose PT will never normally hit either of those, but you can do > IOC_PERIOD on it, just for giggles. True. Should we worry? > Yes, this very much needs a comment... Also, should not this then live > in ->add() in the first place? The output_stop() stuff in the mmap_close path relies on perf_aux_output_end() being in ->stop(), let me think if there's more. Regards, -- Alex