Skip to contentSkip to conversation

perf, x86, uncore: Fix minor race in box set up

PATCH New Andi Kleen 2 participants 2 messages Sep 22, 2014 – Oct 3, 2014

Subsystems (2)
lore
Share & tools
lore

Andi Kleen

author
#0 lore raw

PATCH 1/1 perf, x86, uncore: Fix minor race in box set up

From: Andi Kleen <[email protected]>

I was looking for the trinity oops cause in the uncore driver. (so far didn't found it)

However I found this tiny race: when a box is set up two threads on the same CPU, they may be setting up the box in parallel (e.g. with kernel preemption). This could lead to the reference count being increasing too much. Always recheck there is no existing cpu reference inside the lock.

---
 arch/x86/kernel/cpu/perf_event_intel_uncore.c | 4 ++++
 1 file changed, 4 insertions(+)
arch/x86/kernel/cpu/perf_event_intel_uncore.c +4 -0
index 812ec5d..76fa7e3 100644
@@ -42,6 +42,9 @@ struct intel_uncore_box *uncore_pmu_to_box(struct intel_uncore_pmu *pmu, int cpu
 		return box;
 
 	raw_spin_lock(&uncore_box_lock);
+	/* Recheck in lock to handle races. */
+	if (*per_cpu_ptr(pmu->box, cpu))
+		goto out;
 	list_for_each_entry(box, &pmu->box_list, list) {
 		if (box->phys_id == topology_physical_package_id(cpu)) {
 			atomic_inc(&box->refcnt);
@@ -49,6 +52,7 @@ struct intel_uncore_box *uncore_pmu_to_box(struct intel_uncore_pmu *pmu, int cpu
 			break;
 		}
 	}
+out:
 	raw_spin_unlock(&uncore_box_lock);
 
 	return *per_cpu_ptr(pmu->box, cpu);
Signature · 1 line
1.9.3
#1 lore raw

[tip:perf/core] perf/x86/intel/uncore: Fix minor race in box set up

perf/x86/intel/uncore: Fix minor race in box set up
Commit-ID:  4f971248bc6ad2bb2a89a25a072ebfec5757d298
Gitweb:     http://git.kernel.org/tip/4f971248bc6ad2bb2a89a25a072ebfec5757d298
Author:     Andi Kleen <[email protected]>
AuthorDate: Mon, 22 Sep 2014 15:27:06 -0700
Committer:  Ingo Molnar <[email protected]>
CommitDate: Fri, 3 Oct 2014 06:02:49 +0200

perf/x86/intel/uncore: Fix minor race in box set up

I was looking for the trinity oops cause in the uncore driver. (so far didn't found it)

However I found this tiny race: when a box is set up two threads on the same CPU, they may be setting up the box in parallel (e.g. with kernel preemption). This could lead to the reference count being increasing too much. Always recheck there is no existing cpu reference inside the lock.

---
 arch/x86/kernel/cpu/perf_event_intel_uncore.c | 4 ++++
 1 file changed, 4 insertions(+)
arch/x86/kernel/cpu/perf_event_intel_uncore.c +4 -0
index 42d00e5..9762dbd 100644
@@ -42,6 +42,9 @@ struct intel_uncore_box *uncore_pmu_to_box(struct intel_uncore_pmu *pmu, int cpu
 		return box;
 
 	raw_spin_lock(&uncore_box_lock);
+	/* Recheck in lock to handle races. */
+	if (*per_cpu_ptr(pmu->box, cpu))
+		goto out;
 	list_for_each_entry(box, &pmu->box_list, list) {
 		if (box->phys_id == topology_physical_package_id(cpu)) {
 			atomic_inc(&box->refcnt);
@@ -49,6 +52,7 @@ struct intel_uncore_box *uncore_pmu_to_box(struct intel_uncore_pmu *pmu, int cpu
 			break;
 		}
 	}
+out:
 	raw_spin_unlock(&uncore_box_lock);
 
 	return *per_cpu_ptr(pmu->box, cpu);