Message ID | 20250519121512.5657-1-wsa+renesas@sang-engineering.com |
---|---|
Headers | show |
Series | archs: use proper node names for GPIO based I2C busses | expand |
Hi Wolfram, On Mon, 19 May 2025 at 14:15, Wolfram Sang <wsa+renesas@sang-engineering.com> wrote: > A lot of boards across various archs have a superfluous '-' in their > node name for GPIO based I2C busses. 'dtbs_check' complains, so fix > them. > > Based on linux-next as of 20250516. Build bots are happy. ARM patches > depend on another cleanup series[1], the rest has no dependencies. I > suggest that patches go via their subsystem trees. > > Tested on Calao USB boards using AT91 chipsets and a Renesas Lager board > using R-Car H2. Thanks for your series! > arm64: dts: exynos: use proper node names for GPIO based I2C busses > arm64: dts: mediatek: use proper node names for GPIO based I2C busses > ARM: dts: microchip: use proper node names for GPIO based I2C busses > ARM: dts: samsung: use proper node names for GPIO based I2C busses > ARM: dts: stm32: use proper node names for GPIO based I2C busses > LoongArch: dts: use proper node names for GPIO based I2C busses > riscv: dts: allwinner: use proper node names for GPIO based I2C busses Hmmm... which of these changes affect the R-Car board? ;-) Am I missing something? Gr{oetje,eeting}s, Geert
On 19/05/2025 14:15, Wolfram Sang wrote: > There shall not be a '-' before the number. > > Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com> > --- > arch/arm64/boot/dts/exynos/exynos5433-tm2-common.dtsi | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/arch/arm64/boot/dts/exynos/exynos5433-tm2-common.dtsi b/arch/arm64/boot/dts/exynos/exynos5433-tm2-common.dtsi > index 8f02de8480b6..197de398dd9b 100644 > --- a/arch/arm64/boot/dts/exynos/exynos5433-tm2-common.dtsi > +++ b/arch/arm64/boot/dts/exynos/exynos5433-tm2-common.dtsi > @@ -85,7 +85,7 @@ homepage-key { > }; > }; > > - i2c_max98504: i2c-gpio-0 { > + i2c_max98504: i2c-gpio0 { You did not paste the warning in commit msg, so I don't know the rationale, but at first glance this is not correct. "-0" is the preferred suffix. "0" is not. Best regards, Krzysztof
> You did not paste the warning in commit msg, so I don't know the > rationale, but at first glance this is not correct. "-0" is the > preferred suffix. "0" is not. Here is an example: .../Kernel/linux/arch/arm/boot/dts/microchip/usb_a9g20.dtb: i2c-gpio-0 (i2c-gpio): $nodename:0: 'i2c-gpio-0' does not match '^i2c(@.+|-[a-z0-9]+)?$' from schema $id: http://devicetree.org/schemas/i2c/i2c-gpio.yaml#
On 19/05/2025 16:45, Wolfram Sang wrote: > >> You did not paste the warning in commit msg, so I don't know the >> rationale, but at first glance this is not correct. "-0" is the >> preferred suffix. "0" is not. > > Here is an example: > > .../Kernel/linux/arch/arm/boot/dts/microchip/usb_a9g20.dtb: i2c-gpio-0 (i2c-gpio): $nodename:0: 'i2c-gpio-0' does not match '^i2c(@.+|-[a-z0-9]+)?$' > from schema $id: http://devicetree.org/schemas/i2c/i2c-gpio.yaml# > ... so clearly the suffix is "-foo" or "-0", as we usually prefer. This should be replaced into i2c-X, by dropping "gpio", instead of using less favored suffixing (one without -). Best regards, Krzysztof
On 19/05/2025 17:14, Wolfram Sang wrote: > >> ... so clearly the suffix is "-foo" or "-0", as we usually prefer. This >> should be replaced into i2c-X, by dropping "gpio", instead of using less >> favored suffixing (one without -). > > Hmm, I can't automate this because it will then need to be aligned with > the bus numbering of other existing non-GPIO-busses. Which is highly > individual per board. That means we need to drop this series? I think either we use i2c-X or commit 57138f5b8c92 ("schemas: i2c: Avoid extra characters in i2c nodename pattern") from Herve was not correct and needs to be fixed. Best regards, Krzysztof
> I think either we use i2c-X or commit 57138f5b8c92 ("schemas: i2c: Avoid > extra characters in i2c nodename pattern") from Herve was not correct > and needs to be fixed. I will look if I can fix dt-schema instead. Thanks for the pointer!
On Mon, 19 May 2025 18:11:29 +0200 Krzysztof Kozlowski <krzk@kernel.org> wrote: > On 19/05/2025 17:14, Wolfram Sang wrote: > > > >> ... so clearly the suffix is "-foo" or "-0", as we usually prefer. This > >> should be replaced into i2c-X, by dropping "gpio", instead of using less > >> favored suffixing (one without -). > > > > Hmm, I can't automate this because it will then need to be aligned with > > the bus numbering of other existing non-GPIO-busses. Which is highly > > individual per board. That means we need to drop this series? > > I think either we use i2c-X or commit 57138f5b8c92 ("schemas: i2c: Avoid > extra characters in i2c nodename pattern") from Herve was not correct > and needs to be fixed. > I don't know if relevant for this case but Rob did the fix 647181a ("schemas: i2c: Allow for 'i2c-.*' node names") on top of my commit. Best regards, Hervé
On Mon, May 19, 2025 at 06:25:39PM +0200, Wolfram Sang wrote: > > > I think either we use i2c-X or commit 57138f5b8c92 ("schemas: i2c: Avoid > > extra characters in i2c nodename pattern") from Herve was not correct > > and needs to be fixed. > > I will look if I can fix dt-schema instead. Thanks for the pointer! Found it! Patch sent.
On Mon, May 19, 2025 at 02:15:00PM +0200, Wolfram Sang wrote: > A lot of boards across various archs have a superfluous '-' in their > node name for GPIO based I2C busses. 'dtbs_check' complains, so fix > them. Please drop this series. We found out that upstream dt-schema is broken instead. I sent a fix.