LU-6285 libcfs: Do not unnecessarily copy cpumask

Copying a mask just to calculate a static value of siblings
is overkill, esp. if CPUMASK_OFFSTACK is set and the mask is huge,
additionally that'll workaround the incorrectness of kernel code
dealing with such a code until that is actually fixed upstream.

Change-Id: I860cc7d5b54adcde2e7622bfbc64b4e27046b083
Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
Reviewed-on: http://review.whamcloud.com/13904
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: Liang Zhen <liang.zhen@intel.com>
Reviewed-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>

diff --git a/libcfs/libcfs/linux/linux-cpu.c b/libcfs/libcfs/linux/linux-cpu.c
index e55678e..6e879d4 100644 (file)
--- a/libcfs/libcfs/linux/linux-cpu.c
+++ b/libcfs/libcfs/linux/linux-cpu.c
@@ -118,12 +118,7 @@ cfs_cpu_ht_nsiblings(int cpu)
 {
        int     num;
 
-       mutex_lock(&cpt_data.cpt_mutex);
-
-       cfs_cpu_ht_siblings(cpu, cpt_data.cpt_cpumask);
-       num = cpus_weight(*cpt_data.cpt_cpumask);
-
-       mutex_unlock(&cpt_data.cpt_mutex);
+       num = cpus_weight(*topology_thread_cpumask(cpu));
 
        return num;
 }