| Message ID | 20210721140424.725744-34-maxime@cerno.tech (mailing list archive) |
|---|---|
| State | Not Applicable, archived |
| Headers | show |
| Series | None | expand |
On Wed, 21 Jul 2021 16:04:03 +0200, Maxime Ripard wrote: > Even though the previous binding made it a required child node, the > implementation in Linux never made it mandatory and just ignored thermal > zones without trip points. > > This was even effectively encouraged, since the thermal core wouldn't > allow a thermal sensor to probe without a thermal zone. > > In the case where you had a thermal device that had multiple sensors but > with enough knowledge to provide trip points for only a few of them, > this meant that the only way to make that driver probe was to provide a > thermal zone without the trips node required by the binding. > > This obviously led to a fair number of device trees doing exactly that, > making the initial binding requirement ineffective. > > Let's make it clear by dropping that requirement. > > Cc: Amit Kucheria <amitk@kernel.org> > Cc: Daniel Lezcano <daniel.lezcano@linaro.org> > Cc: linux-pm@vger.kernel.org > Cc: Zhang Rui <rui.zhang@intel.com> > Signed-off-by: Maxime Ripard <maxime@cerno.tech> > --- > Documentation/devicetree/bindings/thermal/thermal-zones.yaml | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > Reviewed-by: Rob Herring <robh@kernel.org>
On 21/07/2021 16:04, Maxime Ripard wrote: > Even though the previous binding made it a required child node, the > implementation in Linux never made it mandatory and just ignored thermal > zones without trip points. > > This was even effectively encouraged, since the thermal core wouldn't > allow a thermal sensor to probe without a thermal zone. > > In the case where you had a thermal device that had multiple sensors but > with enough knowledge to provide trip points for only a few of them, > this meant that the only way to make that driver probe was to provide a > thermal zone without the trips node required by the binding. > > This obviously led to a fair number of device trees doing exactly that, > making the initial binding requirement ineffective. > > Let's make it clear by dropping that requirement. > > Cc: Amit Kucheria <amitk@kernel.org> > Cc: Daniel Lezcano <daniel.lezcano@linaro.org> > Cc: linux-pm@vger.kernel.org > Cc: Zhang Rui <rui.zhang@intel.com> > Signed-off-by: Maxime Ripard <maxime@cerno.tech> Applied, thanks! -- D.
diff --git a/Documentation/devicetree/bindings/thermal/thermal-zones.yaml b/Documentation/devicetree/bindings/thermal/thermal-zones.yaml index 164f71598c59..a07de5ed0ca6 100644 --- a/Documentation/devicetree/bindings/thermal/thermal-zones.yaml +++ b/Documentation/devicetree/bindings/thermal/thermal-zones.yaml @@ -215,7 +215,7 @@ patternProperties: - polling-delay - polling-delay-passive - thermal-sensors - - trips + additionalProperties: false additionalProperties: false
Even though the previous binding made it a required child node, the implementation in Linux never made it mandatory and just ignored thermal zones without trip points. This was even effectively encouraged, since the thermal core wouldn't allow a thermal sensor to probe without a thermal zone. In the case where you had a thermal device that had multiple sensors but with enough knowledge to provide trip points for only a few of them, this meant that the only way to make that driver probe was to provide a thermal zone without the trips node required by the binding. This obviously led to a fair number of device trees doing exactly that, making the initial binding requirement ineffective. Let's make it clear by dropping that requirement. Cc: Amit Kucheria <amitk@kernel.org> Cc: Daniel Lezcano <daniel.lezcano@linaro.org> Cc: linux-pm@vger.kernel.org Cc: Zhang Rui <rui.zhang@intel.com> Signed-off-by: Maxime Ripard <maxime@cerno.tech> --- Documentation/devicetree/bindings/thermal/thermal-zones.yaml | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)