From mboxrd@z Thu Jan 1 00:00:00 1970 From: Petr Holasek Subject: Re: numactl problems with -C option specifying cpus outside of cpus_allowed Date: Thu, 3 Oct 2013 17:56:06 +0200 Message-ID: <20131003155606.GA30005@thinkpad-work.redhat.com> References: <00000141799702db-dbec8157-ae1f-43b6-8a30-ec1653be5d31-000000@email.amazonses.com> <20131002190149.GA6882@two.firstfloor.org> <000001417e93dd7e-d18881f0-6ad9-4f67-bcef-89842502d68c-000000@email.amazonses.com> Mime-Version: 1.0 Return-path: Content-Disposition: inline In-Reply-To: <000001417e93dd7e-d18881f0-6ad9-4f67-bcef-89842502d68c-000000@email.amazonses.com> Sender: linux-numa-owner@vger.kernel.org List-ID: Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: Christoph Lameter Cc: Andi Kleen , linux-numa@vger.kernel.org On Thu, 03 Oct 2013, Christoph Lameter wrote: > On Wed, 2 Oct 2013, Andi Kleen wrote: > > > On Wed, Oct 02, 2013 at 02:34:01PM +0000, Christoph Lameter wrote: > > > It seems that the check for processors specified with C is always done > > > regardless of CAP_SYS_NICE or the possibility that root may be calling > > > numactl. > > > > > > taskset does not have that problem. Nor did releases of numactl before > > > 2.0.8. > > > > The recently added --all option may help? > > Which version added --all? Would the correct approach not be to check if > its root runninng numactl or if CAP_SYS_NICE is set? --all option was added in version 2.0.9-rc5. When CAP_SYS_NICE is not set, numactl will fail in sched_setaffinity() call even so.