Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1339371

[patch V2 17/28] x86/perf/cqm: Get rid of the silly for_each_cpu lookups

Show all headers | View raw

CQM is a strict per package facility. Use the proper cpumasks to lookup the
readers.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
---
 arch/x86/kernel/cpu/perf_event_intel_cqm.c |   34 ++++++++++-------------------
 1 file changed, 12 insertions(+), 22 deletions(-)

Index: b/arch/x86/kernel/cpu/perf_event_intel_cqm.c
===================================================================
--- a/arch/x86/kernel/cpu/perf_event_intel_cqm.c
+++ b/arch/x86/kernel/cpu/perf_event_intel_cqm.c
@@ -1244,15 +1244,12 @@ static struct pmu intel_cqm_pmu = {
 
 static inline void cqm_pick_event_reader(int cpu)
 {
-	int phys_id = topology_physical_package_id(cpu);
-	int i;
+	int reader;
 
-	for_each_cpu(i, &cqm_cpumask) {
-		if (phys_id == topology_physical_package_id(i))
-			return;	/* already got reader for this socket */
-	}
-
-	cpumask_set_cpu(cpu, &cqm_cpumask);
+	/* First online cpu in package becomes the reader */
+	reader = cpumask_any_and(&cqm_cpumask, topology_core_cpumask(cpu));
+	if (reader >= nr_cpu_ids)
+		cpumask_set_cpu(cpu, &cqm_cpumask);
 }
 
 static void intel_cqm_cpu_starting(unsigned int cpu)
@@ -1270,24 +1267,17 @@ static void intel_cqm_cpu_starting(unsig
 
 static void intel_cqm_cpu_exit(unsigned int cpu)
 {
-	int phys_id = topology_physical_package_id(cpu);
-	int i;
+	int target;
 
-	/*
-	 * Is @cpu a designated cqm reader?
-	 */
+	/* Is @cpu the current cqm reader for this package ? */
 	if (!cpumask_test_and_clear_cpu(cpu, &cqm_cpumask))
 		return;
 
-	for_each_online_cpu(i) {
-		if (i == cpu)
-			continue;
-
-		if (phys_id == topology_physical_package_id(i)) {
-			cpumask_set_cpu(i, &cqm_cpumask);
-			break;
-		}
-	}
+	/* Find another online reader in this package */
+	target = cpumask_any_but(topology_core_cpumask(cpu), cpu);
+
+	if (target < nr_cpu_ids)
+		cpumask_set_cpu(target, &cqm_cpumask);
 }
 
 static int intel_cqm_cpu_notifier(struct notifier_block *nb,

Back to linux.kernel | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread

Thread

[patch V2 00/28] x86/perf/intel/uncore|rapl: Fix error handling and  sanitize pmu management Thomas Gleixner <tglx@linutronix.de> - 2016-02-22 12:20 +0100
  [patch V2 18/28] x86/perf/intel_rapl: Make Knights Landings support  functional Thomas Gleixner <tglx@linutronix.de> - 2016-02-22 12:20 +0100
    Re: [patch V2 18/28] x86/perf/intel_rapl: Make Knights Landings  support functional Borislav Petkov <bp@alien8.de> - 2016-02-22 21:40 +0100
  [patch V2 13/28] x86/perf/intel_uncore: Clear all hardware state on  exit Thomas Gleixner <tglx@linutronix.de> - 2016-02-22 12:20 +0100
    Re: [patch V2 13/28] x86/perf/intel_uncore: Clear all hardware state  on exit Peter Zijlstra <peterz@infradead.org> - 2016-02-22 13:00 +0100
      Re: [patch V2 13/28] x86/perf/intel_uncore: Clear all hardware state  on exit Thomas Gleixner <tglx@linutronix.de> - 2016-02-22 13:10 +0100
  [patch V2 17/28] x86/perf/cqm: Get rid of the silly for_each_cpu  lookups Thomas Gleixner <tglx@linutronix.de> - 2016-02-22 12:20 +0100
  [patch V2 16/28] x86/perf/intel_uncore: Make it modular Thomas Gleixner <tglx@linutronix.de> - 2016-02-22 12:20 +0100
  [patch V2 12/28] x86/perf/uncore: Track packages not per cpu data Thomas Gleixner <tglx@linutronix.de> - 2016-02-22 12:20 +0100
  [patch V2 24/28] x86/perf/intel/rapl: Make pmu lock raw Thomas Gleixner <tglx@linutronix.de> - 2016-02-22 12:20 +0100
  [patch V2 02/28] x86/perf/intel/uncore: Simplify error rollback Thomas Gleixner <tglx@linutronix.de> - 2016-02-22 12:20 +0100
  [patch V2 03/28] x86/perf/intel/uncore: Fix error handling Thomas Gleixner <tglx@linutronix.de> - 2016-02-22 12:20 +0100
  [patch V2 22/28] x86/perf/intel/rapl: Cleanup the printk output Thomas Gleixner <tglx@linutronix.de> - 2016-02-22 12:20 +0100
    Re: [patch V2 22/28] x86/perf/intel/rapl: Cleanup the printk output Andi Kleen <andi@firstfloor.org> - 2016-02-22 19:50 +0100
      Re: [patch V2 22/28] x86/perf/intel/rapl: Cleanup the printk  output Thomas Gleixner <tglx@linutronix.de> - 2016-02-22 20:00 +0100
  [patch V2 08/28] x86/perf/uncore: Make uncore_pcibus_to_physid static Thomas Gleixner <tglx@linutronix.de> - 2016-02-22 12:20 +0100
  [patch V2 11/28] x86/topology: Create logical package id Thomas Gleixner <tglx@linutronix.de> - 2016-02-22 12:20 +0100
    Re: [patch V2 11/28] x86/topology: Create logical package id Andi Kleen <andi@firstfloor.org> - 2016-02-22 20:00 +0100
      Re: [patch V2 11/28] x86/topology: Create logical package id Thomas Gleixner <tglx@linutronix.de> - 2016-02-22 20:10 +0100
  [patch V2 14/28] x86/perf/intel_uncore: Make PCI and MSR uncore  independent Thomas Gleixner <tglx@linutronix.de> - 2016-02-22 12:20 +0100
  [patch V2 23/28] x86/perf/intel/rapl: Refactor code some more Thomas Gleixner <tglx@linutronix.de> - 2016-02-22 12:20 +0100
  [patch V2 07/28] x86/perf/intel_uncore: Make code readable Thomas Gleixner <tglx@linutronix.de> - 2016-02-22 12:20 +0100
  [patch V2 20/28] x86/perf/intel/rapl: Sanitize the quirk handling Thomas Gleixner <tglx@linutronix.de> - 2016-02-22 12:20 +0100
  [patch V2 25/28] x86/perf/intel/rapl: Utilize event->pmu_private Thomas Gleixner <tglx@linutronix.de> - 2016-02-22 12:20 +0100
  [patch V2 05/28] x86/perf/intel_uncore: Cleanup hardware on exit Thomas Gleixner <tglx@linutronix.de> - 2016-02-22 12:20 +0100
  Re: [patch V2 00/28] x86/perf/intel/uncore|rapl: Fix error handling and sanitize pmu management Andi Kleen <andi@firstfloor.org> - 2016-02-22 20:00 +0100

csiph-web