Thermal: pid_thermal_gov: Apply compenstaion even if sum_err is 0 or max

Applied compensation even if sum_err is 0 or max to prevent determining
target state to 0 in short time.

Bug 1200111

Change-Id: I52a8bbeb4ae8553f14401cd0eaa4508337111955
Signed-off-by: Jinyoung Park <jinyoungp@nvidia.com>
Reviewed-on: http://git-master/r/211960
(cherry picked from commit 221bc8d9307876a099c917100ef1066cf8cef308)
Signed-off-by: Jinyoung Park <jinyoungp@nvidia.com>
Reviewed-on: http://git-master/r/215535
Reviewed-by: Diwakar Tundlam <dtundlam@nvidia.com>
(cherry picked from commit b510de2a4f4eeba4fa69d2633407d1696235c4f0)
Signed-off-by: Srikar Srimath Tirumala <srikars@nvidia.com>

1 files changed, 1 insertions, 8 deletions

diff --git a/drivers/thermal/pid_thermal_gov.c b/drivers/thermal/pid_thermal_gov.c
index c29033fb6..f326d8e15 100644
--- a/drivers/thermal/pid_thermal_gov.c
+++ b/drivers/thermal/pid_thermal_gov.c
@@ -382,15 +382,8 @@ pid_thermal_gov_get_target(struct thermal_zone_device *tz,
                         derivative = min_t(s64, derivative, max_dout);
         }
 
-        sum_err = proportional + derivative;
-        sum_err = max_t(s64, sum_err, 0);
-        if (sum_err == 0)
-                return 0;
-
+        sum_err = max_t(s64, proportional + derivative, 0);
         sum_err = min_t(s64, sum_err, max_err);
-        if (sum_err == max_err)
-                return max_state;
-
         sum_err = sum_err * max_state + max_err - 1;
         target = (unsigned long)div64_s64(sum_err, max_err);