From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1756012AbZHYUdO (ORCPT ); Tue, 25 Aug 2009 16:33:14 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1755523AbZHYUdO (ORCPT ); Tue, 25 Aug 2009 16:33:14 -0400 Received: from e23smtp01.au.ibm.com ([202.81.31.143]:57815 "EHLO e23smtp01.au.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755281AbZHYUdN (ORCPT ); Tue, 25 Aug 2009 16:33:13 -0400 Date: Wed, 26 Aug 2009 02:03:07 +0530 From: "K.Prasad" To: Frederic Weisbecker , Ingo Molnar , Peter Zijlstra Cc: LKML , Lai Jiangshan , Steven Rostedt , Mathieu Desnoyers , Alan Stern Subject: Re: [Patch 0/1] HW-BKPT: Allow per-cpu kernel-space Hardware Breakpoint requests Message-ID: <20090825203307.GA5834@in.ibm.com> Reply-To: prasad@linux.vnet.ibm.com References: <20090817124641.GA13354@in.ibm.com> <20090819161119.GA3633@in.ibm.com> <20090819173259.GF4972@nowhere> <20090820172719.GA16499@in.ibm.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20090820172719.GA16499@in.ibm.com> User-Agent: Mutt/1.5.19 (2009-01-05) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Thu, Aug 20, 2009 at 10:57:19PM +0530, K.Prasad wrote: > On Wed, Aug 19, 2009 at 07:33:00PM +0200, Frederic Weisbecker wrote: > > On Wed, Aug 19, 2009 at 09:41:19PM +0530, K.Prasad wrote: > > > On Mon, Aug 17, 2009 at 06:16:41PM +0530, K.Prasad wrote: > > > > Hi All, > > > > Please find a patch that enables kernel-space breakpoints to be > > > > requested for a subset of the available CPUs in the system. This allows > > > > per-CPU breakpoints and comes with the associated benefit of reduced > > > > overhead during (un)registration. > > > > > > > > This enhancement allows exploitation of hardware breakpoint registers by > > > > 'perf' which produces a CPU-wise information. > > > > > [edited] > > > > > > Hi Frederic, > > > Do you find these patches, that provide the ability to restrict > > > kernel-space breakpoints to any given subset of CPUs, to bring the > > > requisite features for exploitation of hw-bkpt by 'perf tools'? > > > > > > Also of interest would be the reduced overhead associated with > > > (un)register_kernel_hw_breakpoint() operations (no IPI in case of > > > single-CPU breakpoint request). > > > [edited] > > - having a callback that quickly swap two breakpoints in order to support > > the hardware register multiplexing. I guess the pmu object would just need > > to call it when the multiplexing is decided. > > > > > > Are you suggesting something like a modify_kernel_hw_breakpoint() that > can quickly change a breakpoint address/characteristics? > > That's quite doable...it requires a quick validation through > arch_validate_hwbkpt_settings() and the requisite IPIs (depending on > what the new cpumask is). > > I will send a patch to that effect soon. > Hi Frederic, I just sent a patchset that adds the ability to specify per-cpu kernel-space breakpoints + a (relatively) lightweight function to modify the characteristics of a kernel-space breakpoint that can be used to swap between two breakpoint requests. Please pull them into -tip tree if you find them mature and ready. With these new feature additions, I see the HW-Breakpoint infrastructure code ready to meet the needs for exploitation by perf-tools and I presume you would restart your effort on the same? Thanks, K.Prasad