| Message ID | 1372066617-23518-1-git-send-email-sachin.kamat@linaro.org |
|---|---|
| State | Accepted |
| Headers | show |
On Mon, Jun 24, 2013 at 03:06:57PM +0530, Sachin Kamat wrote: > s5m8767 regulator is used on Exynos platforms which use pin controller > to configure GPIOs. Update the example accordingly. This smells bad, why does a driver using GPIOs through the GPIO API see a change in the binding?
On 24 June 2013 20:55, Mark Brown <broonie@kernel.org> wrote: > On Mon, Jun 24, 2013 at 03:06:57PM +0530, Sachin Kamat wrote: >> s5m8767 regulator is used on Exynos platforms which use pin controller >> to configure GPIOs. Update the example accordingly. > > This smells bad, why does a driver using GPIOs through the GPIO API see > a change in the binding? There is no change in the bindings, but just a correction in the documentation to reflect the implementation. Earlier when Samsung platforms did not have pinctrl driver, legacy GPIO driver was used which took those 5 parameters. Now since we are using pinctrl, we need only 3 parameters. The document was somehow not updated to reflect this change.
On Tue, Jun 25, 2013 at 11:56:12AM +0530, Sachin Kamat wrote: > There is no change in the bindings, but just a correction in the > documentation to reflect the > implementation. Earlier when Samsung platforms did not have pinctrl > driver, legacy GPIO driver > was used which took those 5 parameters. Now since we are using > pinctrl, we need only 3 parameters. > The document was somehow not updated to reflect this change. So there was a previous change to the code that mistakenly didn't update the binding document?
On 28 June 2013 17:05, Mark Brown <broonie@kernel.org> wrote: > On Tue, Jun 25, 2013 at 11:56:12AM +0530, Sachin Kamat wrote: > >> There is no change in the bindings, but just a correction in the >> documentation to reflect the >> implementation. Earlier when Samsung platforms did not have pinctrl >> driver, legacy GPIO driver >> was used which took those 5 parameters. Now since we are using >> pinctrl, we need only 3 parameters. >> The document was somehow not updated to reflect this change. > > So there was a previous change to the code that mistakenly didn't update > the binding document? Yes, that is correct. In fact the changes happened during the review process itself. Hence there is no commit history of this change. However the binding document was not updated during the review phase and the first version itself got committed as it is.
On Mon, Jun 24, 2013 at 03:06:57PM +0530, Sachin Kamat wrote: > s5m8767 regulator is used on Exynos platforms which use pin controller > to configure GPIOs. Update the example accordingly. Applied, thanks. Please use subject lines that match the subsystem and try to make your changelogs clearer.
diff --git a/Documentation/devicetree/bindings/regulator/s5m8767-regulator.txt b/Documentation/devicetree/bindings/regulator/s5m8767-regulator.txt index a35ff99..7364f71 100644 --- a/Documentation/devicetree/bindings/regulator/s5m8767-regulator.txt +++ b/Documentation/devicetree/bindings/regulator/s5m8767-regulator.txt @@ -103,13 +103,13 @@ Example: s5m8767,pmic-buck-default-dvs-idx = <0>; - s5m8767,pmic-buck-dvs-gpios = <&gpx0 0 1 0 0>, /* DVS1 */ - <&gpx0 1 1 0 0>, /* DVS2 */ - <&gpx0 2 1 0 0>; /* DVS3 */ + s5m8767,pmic-buck-dvs-gpios = <&gpx0 0 0>, /* DVS1 */ + <&gpx0 1 0>, /* DVS2 */ + <&gpx0 2 0>; /* DVS3 */ - s5m8767,pmic-buck-ds-gpios = <&gpx2 3 1 0 0>, /* SET1 */ - <&gpx2 4 1 0 0>, /* SET2 */ - <&gpx2 5 1 0 0>; /* SET3 */ + s5m8767,pmic-buck-ds-gpios = <&gpx2 3 0>, /* SET1 */ + <&gpx2 4 0>, /* SET2 */ + <&gpx2 5 0>; /* SET3 */ s5m8767,pmic-buck2-dvs-voltage = <1350000>, <1300000>, <1250000>, <1200000>,
s5m8767 regulator is used on Exynos platforms which use pin controller to configure GPIOs. Update the example accordingly. Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org> --- .../bindings/regulator/s5m8767-regulator.txt | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-)