cpufreq: Avoid false-positive WARN_ON()s in cpufreq_update_policy()
authorRafael J. Wysocki <rafael.j.wysocki@intel.com>
Tue, 28 Jun 2016 01:29:29 +0000 (03:29 +0200)
committerRafael J. Wysocki <rafael.j.wysocki@intel.com>
Tue, 28 Jun 2016 01:29:29 +0000 (03:29 +0200)
CPU notifications from the firmware coming in when cpufreq is
suspended cause cpufreq_update_current_freq() to return 0 which
triggers the WARN_ON() in cpufreq_update_policy() for no reason.

Avoid that by checking cpufreq_suspended before calling
cpufreq_update_current_freq().

Fixes: c9d9c929e674 (cpufreq: Abort cpufreq_update_current_freq() for cpufreq_suspended set)
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Cc: 4.6+ <stable@vger.kernel.org> # 4.6+
drivers/cpufreq/cpufreq.c

index 9009295f5134247374cacf2a55717e07ff61a057..5617c7087d775ff5f4f93cc084b6bc8769be77ec 100644 (file)
@@ -2261,6 +2261,10 @@ int cpufreq_update_policy(unsigned int cpu)
         * -> ask driver for current freq and notify governors about a change
         */
        if (cpufreq_driver->get && !cpufreq_driver->setpolicy) {
+               if (cpufreq_suspended) {
+                       ret = -EAGAIN;
+                       goto unlock;
+               }
                new_policy.cur = cpufreq_update_current_freq(policy);
                if (WARN_ON(!new_policy.cur)) {
                        ret = -EIO;