perf/x86/intel: Use ULL constant to prevent undefined shift behaviour
authorColin King <[email protected]>
Wed, 11 Jan 2017 11:43:10 +0000 (11:43 +0000)
committerThomas Gleixner <[email protected]>
Wed, 11 Jan 2017 15:43:30 +0000 (16:43 +0100)
When x86_pmu.num_counters is 32 the shift of the integer constant 1 is
exceeding 32bit and therefor undefined behaviour.

Fix this by shifting 1ULL instead of 1.

Reported-by: CoverityScan CID#1192105 ("Bad bit shift operation")
Signed-off-by: Colin Ian King <[email protected]>
Cc: Andi Kleen <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Kan Liang <[email protected]>
Cc: Stephane Eranian <[email protected]>
Cc: Alexander Shishkin <[email protected]>
Link: http://lkml.kernel.org/r/[email protected]
Signed-off-by: Thomas Gleixner <[email protected]>
arch/x86/events/intel/core.c

index 86138267b68a77753737ffd4c4be2ece6a24cc84..d611cab214a6050555eb4c5f6f5638ca46b4659e 100644 (file)
@@ -3987,7 +3987,7 @@ __init int intel_pmu_init(void)
                     x86_pmu.num_counters, INTEL_PMC_MAX_GENERIC);
                x86_pmu.num_counters = INTEL_PMC_MAX_GENERIC;
        }
-       x86_pmu.intel_ctrl = (1 << x86_pmu.num_counters) - 1;
+       x86_pmu.intel_ctrl = (1ULL << x86_pmu.num_counters) - 1;
 
        if (x86_pmu.num_counters_fixed > INTEL_PMC_MAX_FIXED) {
                WARN(1, KERN_ERR "hw perf events fixed %d > max(%d), clipping!",