cpuidle: teo: Ignore disabled idle states that are too deep
authorRafael J. Wysocki <rafael.j.wysocki@intel.com>
Thu, 10 Oct 2019 21:32:17 +0000 (23:32 +0200)
committerRafael J. Wysocki <rafael.j.wysocki@intel.com>
Mon, 14 Oct 2019 08:40:32 +0000 (10:40 +0200)
Prevent disabled CPU idle state with target residencies beyond the
anticipated idle duration from being taken into account by the TEO
governor.

Fixes: b26bf6ab716f ("cpuidle: New timer events oriented governor for tickless systems")
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Cc: 5.1+ <stable@vger.kernel.org> # 5.1+
drivers/cpuidle/governors/teo.c

index b5a0e498f7989791192f79e72e166eb5ed862a85..8806db95a91388c5962595fb6794a8d9f5345cd2 100644 (file)
@@ -257,6 +257,13 @@ static int teo_select(struct cpuidle_driver *drv, struct cpuidle_device *dev,
                struct cpuidle_state_usage *su = &dev->states_usage[i];
 
                if (s->disabled || su->disable) {
+                       /*
+                        * Ignore disabled states with target residencies beyond
+                        * the anticipated idle duration.
+                        */
+                       if (s->target_residency > duration_us)
+                               continue;
+
                        /*
                         * If the "early hits" metric of a disabled state is
                         * greater than the current maximum, it should be taken