Re: [PATCH v1 2/2] x86, perf,cqm: handle CPU hotplug

From: Vikas Shivappa
Date: Tue Jun 23 2015 - 16:35:32 EST




On Thu, 18 Jun 2015, Thomas Gleixner wrote:

On Thu, 18 Jun 2015, Kanaka Juvva wrote:

Added lock in event reader function. The cqm_pick_event_reader() function
accesses cqm_cpumask and it is critical section between this and
cqm_stable().

This situation is true when a CPU is hotplugged. Mutex is used to protect
the critical section.

Signed-off-by: Kanaka Juvva <kanaka.d.juvva@xxxxxxxxxxxxxxx>
---
arch/x86/kernel/cpu/perf_event_intel_cqm.c | 5 ++++-
1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/arch/x86/kernel/cpu/perf_event_intel_cqm.c b/arch/x86/kernel/cpu/perf_event_intel_cqm.c
index 1880761..e17e37f 100644
--- a/arch/x86/kernel/cpu/perf_event_intel_cqm.c
+++ b/arch/x86/kernel/cpu/perf_event_intel_cqm.c
@@ -1239,12 +1239,15 @@ static inline void cqm_pick_event_reader(int cpu)
int phys_id = topology_physical_package_id(cpu);
int i;

+ mutex_lock(&cache_mutex);

I already explained it to Vikas. You CANNOT take a mutex in that code
path as it runs with interrupts disabled on a CPU which cannot
schedule.

This patch also needs to be merged with the new package mask changes that was added. Will merge and send the fix.

Thanks,
Vikas


Sigh.

tglx

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/