From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S935567AbcKOODF (ORCPT ); Tue, 15 Nov 2016 09:03:05 -0500 Received: from Galois.linutronix.de ([146.0.238.70]:43227 "EHLO Galois.linutronix.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752225AbcKOODD (ORCPT ); Tue, 15 Nov 2016 09:03:03 -0500 Date: Tue, 15 Nov 2016 15:00:17 +0100 (CET) From: Thomas Gleixner To: Fenghua Yu cc: "H. Peter Anvin" , Ingo Molnar , Tony Luck , Ravi V Shankar , Sai Prakhya , Vikas Shivappa , linux-kernel , x86 Subject: Re: [PATCH 3/3] x86/intel_rdt: Update closid in PQR_ASSOC registers in synchronous mode when changing "cpus" In-Reply-To: <1478912558-55514-3-git-send-email-fenghua.yu@intel.com> Message-ID: References: <1478912558-55514-1-git-send-email-fenghua.yu@intel.com> <1478912558-55514-3-git-send-email-fenghua.yu@intel.com> User-Agent: Alpine 2.20 (DEB 67 2015-01-07) MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Fri, 11 Nov 2016, Fenghua Yu wrote: > +/* > + * MSR_IA32_PQR_ASSOC is scoped per logical CPU, so all updates > + * are always in thread context. And this comment tells us what? Nothing useful. It lacks the most important information why this is safe against a logical CPU switching the MSR right at this moment during context switch. It's safe because the pqr_switch_to function is called with interrupts disabled. > + */ > +static void rdt_update_pqr_assoc_closid(void *v) > +{ > + struct intel_pqr_state *state = this_cpu_ptr(&pqr_state); > + > + state->closid = *(int *)v; > + wrmsr(MSR_IA32_PQR_ASSOC, state->rmid, state->closid); > +} > + > +static void rdt_update_pqr_assoc(const struct cpumask *cpu_mask, int closid) > +{ > + int cpu = get_cpu(); > + > + /* Update PQR_ASSOC MSR on this cpu if it's in cpu_mask. */ > + if (cpumask_test_cpu(cpu, cpu_mask)) > + rdt_update_pqr_assoc_closid(&closid); > + /* Update PQR_ASSOC MSR on the rest of cpus. */ > + smp_call_function_many(cpu_mask, rdt_update_pqr_assoc_closid, > + &closid, 1); > + put_cpu(); > +} > + > static ssize_t rdtgroup_cpus_write(struct kernfs_open_file *of, > char *buf, size_t nbytes, loff_t off) > { > @@ -238,6 +263,9 @@ static ssize_t rdtgroup_cpus_write(struct kernfs_open_file *of, > &rdtgroup_default.cpu_mask, tmpmask); > for_each_cpu(cpu, tmpmask) > per_cpu(cpu_closid, cpu) = 0; > + > + /* Update PQR_ASSOC registers on the dropped cpus */ > + rdt_update_pqr_assoc(tmpmask, rdtgroup_default.closid); Grr. You store the new closid now in the for_each_cpu() loop and in the smp function call. > } > > /* > @@ -253,6 +281,9 @@ static ssize_t rdtgroup_cpus_write(struct kernfs_open_file *of, > } > for_each_cpu(cpu, tmpmask) > per_cpu(cpu_closid, cpu) = rdtgrp->closid; > + > + /* Update PQR_ASSOC registers on the added cpus */ > + rdt_update_pqr_assoc(tmpmask, rdtgrp->closid); Ditto. > } > > /* Done pushing/pulling - update this group with new mask */ > @@ -783,18 +814,6 @@ static int reset_all_cbms(struct rdt_resource *r) > } > > /* > - * MSR_IA32_PQR_ASSOC is scoped per logical CPU, so all updates > - * are always in thread context. > - */ > -static void rdt_reset_pqr_assoc_closid(void *v) > -{ > - struct intel_pqr_state *state = this_cpu_ptr(&pqr_state); > - > - state->closid = 0; > - wrmsr(MSR_IA32_PQR_ASSOC, state->rmid, 0); > -} > - > -/* > * Forcibly remove all of subdirectories under root. > */ > static void rmdir_all_sub(void) > @@ -809,13 +828,9 @@ static void rmdir_all_sub(void) > t->closid = 0; > read_unlock(&tasklist_lock); > > - get_cpu(); > - /* Reset PQR_ASSOC MSR on this cpu. */ > - rdt_reset_pqr_assoc_closid(NULL); > - /* Reset PQR_ASSOC MSR on the rest of cpus. */ > - smp_call_function_many(cpu_online_mask, rdt_reset_pqr_assoc_closid, > - NULL, 1); > - put_cpu(); > + /* Reset PQR_ASSOC MSR on all online cpus */ > + rdt_update_pqr_assoc(cpu_online_mask, 0); And here you have the extra closid loop later on. Really useful. Thanks, tglx