| Message ID | 20220812131202.1331238-1-daniel.lezcano@linaro.org |
|---|---|
| State | New |
| Headers | show |
| Series | thermal/core: Fix lockdep_assert() warning | expand |
On 12/08/2022 15:34, Marek Szyprowski wrote: > On 12.08.2022 15:12, Daniel Lezcano wrote: >> The function thermal_zone_device_is_enabled() must be called with the >> thermal zone lock held. In the resume path, it is called without. >> >> As the thermal_zone_device_is_enabled() is also checked in >> thermal_zone_device_update(), do the check in resume() function is >> pointless, except for saving an extra initialization which does not >> hurt if it is done in all the cases. >> >> Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org> > > This fixes the warning I've reported. Feel free to add: > > Reported-by: Marek Szyprowski <m.szyprowski@samsung.com> > > Tested-by: Marek Szyprowski <m.szyprowski@samsung.com> Great, thanks for testing
On Fri, Aug 12, 2022 at 3:54 PM Daniel Lezcano <daniel.lezcano@linaro.org> wrote: > > On 12/08/2022 15:34, Marek Szyprowski wrote: > > On 12.08.2022 15:12, Daniel Lezcano wrote: > >> The function thermal_zone_device_is_enabled() must be called with the > >> thermal zone lock held. In the resume path, it is called without. > >> > >> As the thermal_zone_device_is_enabled() is also checked in > >> thermal_zone_device_update(), do the check in resume() function is > >> pointless, except for saving an extra initialization which does not > >> hurt if it is done in all the cases. > >> > >> Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org> > > > > This fixes the warning I've reported. Feel free to add: > > > > Reported-by: Marek Szyprowski <m.szyprowski@samsung.com> > > > > Tested-by: Marek Szyprowski <m.szyprowski@samsung.com> > > Great, thanks for testing Do you want me to apply this for -rc3?
diff --git a/drivers/thermal/thermal_core.c b/drivers/thermal/thermal_core.c index 50814009339d..dc8ff6a84df1 100644 --- a/drivers/thermal/thermal_core.c +++ b/drivers/thermal/thermal_core.c @@ -1438,9 +1438,6 @@ static int thermal_pm_notify(struct notifier_block *nb, case PM_POST_SUSPEND: atomic_set(&in_suspend, 0); list_for_each_entry(tz, &thermal_tz_list, node) { - if (!thermal_zone_device_is_enabled(tz)) - continue; - thermal_zone_device_init(tz); thermal_zone_device_update(tz, THERMAL_EVENT_UNSPECIFIED);
The function thermal_zone_device_is_enabled() must be called with the thermal zone lock held. In the resume path, it is called without. As the thermal_zone_device_is_enabled() is also checked in thermal_zone_device_update(), do the check in resume() function is pointless, except for saving an extra initialization which does not hurt if it is done in all the cases. Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org> --- drivers/thermal/thermal_core.c | 3 --- 1 file changed, 3 deletions(-)