From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1422751AbcHEPyw (ORCPT ); Fri, 5 Aug 2016 11:54:52 -0400 Received: from merlin.infradead.org ([205.233.59.134]:52152 "EHLO merlin.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1422676AbcHEPyt (ORCPT ); Fri, 5 Aug 2016 11:54:49 -0400 Date: Fri, 5 Aug 2016 17:54:42 +0200 From: Peter Zijlstra To: Mathieu Poirier Cc: Arnaldo Carvalho de Melo , jolsa@kernel.org, Ingo Molnar , "linux-kernel@vger.kernel.org" , "linux-arm-kernel@lists.infradead.org" , Alexander Shishkin Subject: Re: [PATCH V3 1/6] perf/core: Adding PMU driver specific configuration Message-ID: <20160805155442.GW6879@twins.programming.kicks-ass.net> References: <1469742143-22245-1-git-send-email-mathieu.poirier@linaro.org> <1469742143-22245-2-git-send-email-mathieu.poirier@linaro.org> <20160804165951.GU6862@twins.programming.kicks-ass.net> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.5.23.1 (2014-03-12) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Fri, Aug 05, 2016 at 09:41:15AM -0600, Mathieu Poirier wrote: > On 4 August 2016 at 10:59, Peter Zijlstra wrote: > > On Thu, Jul 28, 2016 at 03:42:18PM -0600, Mathieu Poirier wrote: > >> This patch somewhat mimics the work done on address filters to > >> add the infrastructure needed to pass PMU specific HW > >> configuration to the driver before a session starts. > > > > I'm thinking we want to specify a syntax and validate the string matches > > the syntax in the generic code. > > The syntax is checked in the lexer making sure that nothing other than > @cfg or @cfg=config is sent to the kernel. From there validation is > done in the PMU driver that implements the set_drv_configs() > interface. > > I am not sure to get you point here - can I ask you to provide more details? What keeps random userspace from sending malformed strings in? There's more userspace than just the perf tool. The kernel needs to validate structure.