[4/4] ARM: dts: qcom: msm8974: Add device tree for Samsung Galaxy S5 China

This device has little difference compared to Samsung Galaxy S5 (klte),
so the device tree is based on qcom-msm8974pro-samsung-klte.dts. The
only difference is the gpio pins of i2c_led_gpio. With pins corrected,
the LEDs and WiFi are able to work properly.

Signed-off-by: Rong Zhang <i@rong.moe>
---
 arch/arm/boot/dts/qcom/Makefile                  |  1 +
 .../dts/qcom/qcom-msm8974pro-samsung-kltechn.dts | 16 ++++++++++++++++
 2 files changed, 17 insertions(+)
 create mode 100644 arch/arm/boot/dts/qcom/qcom-msm8974pro-samsung-kltechn.dts

On 21/01/2024 16:39, Rong Zhang wrote:
> This device has little difference compared to Samsung Galaxy S5 (klte),
> so the device tree is based on qcom-msm8974pro-samsung-klte.dts. The
> only difference is the gpio pins of i2c_led_gpio. With pins corrected,
> the LEDs and WiFi are able to work properly.
> 
> Signed-off-by: Rong Zhang <i@rong.moe>


> diff --git a/arch/arm/boot/dts/qcom/qcom-msm8974pro-samsung-kltechn.dts b/arch/arm/boot/dts/qcom/qcom-msm8974pro-samsung-kltechn.dts
> new file mode 100644
> index 000000000000..5a8d59ea4439
> --- /dev/null
> +++ b/arch/arm/boot/dts/qcom/qcom-msm8974pro-samsung-kltechn.dts
> @@ -0,0 +1,16 @@
> +// SPDX-License-Identifier: GPL-2.0
> +#include "qcom-msm8974pro-samsung-klte.dts"
> +
> +/ {
> +	model = "Samsung Galaxy S5 China";
> +	compatible = "samsung,kltechn", "samsung,klte", "qcom,msm8974pro", "qcom,msm8974";

That's not what you said in the binding.

It does not look like you tested the DTS against bindings. Please run
`make dtbs_check W=1` (see
Documentation/devicetree/bindings/writing-schema.rst or
https://www.linaro.org/blog/tips-and-tricks-for-validating-devicetree-sources-with-the-devicetree-schema/
for instructions).

Best regards,
Krzysztof

在 2024-01-22星期一的 11:50 +0100，Konrad Dybcio写道：
> On 21.01.2024 16:39, Rong Zhang wrote:
> > This device has little difference compared to Samsung Galaxy S5
> > (klte),
> > so the device tree is based on qcom-msm8974pro-samsung-klte.dts.
> > The
> > only difference is the gpio pins of i2c_led_gpio. With pins
> > corrected,
> > the LEDs and WiFi are able to work properly.
> > 
> > Signed-off-by: Rong Zhang <i@rong.moe>
> > ---
> 
> Looks like you didn't change the brcm,board-type though?

This should be intentional to allow kltechn and klte to share Wi-Fi
NVRAM file.

> 
> [...]
> 
> > +++ b/arch/arm/boot/dts/qcom/qcom-msm8974pro-samsung-kltechn.dts
> > @@ -0,0 +1,16 @@
> > +// SPDX-License-Identifier: GPL-2.0
> > +#include "qcom-msm8974pro-samsung-klte.dts"
> 
> It's customary not to include .dts files, instead split the common
> parts
> into e.g. qcom-msm8974pro-samsung-klte-common.dtsi and include this
> in
> both the existing and the new one.
> 
> Konrad

On Mon, 2024-01-22 at 16:17 +0100, Krzysztof Kozlowski wrote:
> On 22/01/2024 15:51, Rong Zhang wrote:
> > I've glanced similar dts. To solve this, I think we could either:
> > 1. keep the dt-binding in [PATCH 3/4], and delete "samsung,klte" here
> > 2. rewrite dt-binding like crystalfontz,cfa100{36,37,49,55,56,57,58}:
[...]
> > My preference is (2.) since other variants of klte may be added in the
> > future. I would like to hear your preferences.
> 
> It depends whether the devices are compatible. IOW, entire klte DTB
> should work fine on kltechn, just without few new devices.

Yes, they are compatible. I'd used the klte DTB on kltechn for a long
time before I made this patchset. It worked totally fine expect for
LEDs and WiFi, as I've said in the cover letter.

Thanks,
Rong


> Best regards,
> Krzysztof
>

On Mon, 2024-01-22 at 11:50 +0100, Konrad Dybcio wrote:
> Looks like you didn't change the brcm,board-type though?

In "[PATCH 2/4] ARM: dts: qcom: msm8974-samsung-klte: Pin brcm,board-
type in wifi":

   /*
    * This aims to allow other klte* variants to load the same firmware,
    * as klte variants have little differences in the wifi part.
    */

So it is intentional, in order to let them share the same FW, in
particular, the NVRAM file.

Without [PATCH 2/4] and with this [PATCH 4/4]:
- klte DT probes "brcmfmac*.samsung,klte.txt"
- kltechn DT probes "brcmfmac*.samsung,kltechn.txt", but never probes
"brcmfmac*.samsung,klte.txt"

With both [PATCH 2/4] and this [PATCH 4/4]:
- klte DT probes "brcmfmac*.samsung,klte.txt"
- kltechn DT probes "brcmfmac*.samsung,klte.txt"

I pinned "brcm,board-type" in the klte DT instead of the kltechn one
because other klte* variants are known to have little difference in the
WiFi part. By pinning it in the klte DT, future ports could be easier.

If you'd prefer not doing this, I am OK to drop [PATCH 2/4] in the v2
patchset.

FYI:

LineageOS considers all klte variants use "common" WiFi FW:
https://github.com/search?q=repo%3ALineageOS%2Fandroid_kernel_samsung_klte+CONFIG_BCMDHD_NVRAM_PATH&type=code
https://github.com/LineageOS/android_device_samsung_klte-common/blob/8f71a63415397def5ba886f4030b0d91e2253262/common-proprietary-files.txt#L251

I've tested the klte port of PostmarketOS (with klte FW and this
patchset) on kltechn, and it worked fine. For the FW used by
PostmarketOS, see also:
https://gitlab.com/postmarketOS/pmaports/-/blob/439227770ffcd32eb7e26436598c804dc35637ad/device/testing/firmware-samsung-klte/APKBUILD#L17


> > +++ b/arch/arm/boot/dts/qcom/qcom-msm8974pro-samsung-kltechn.dts
> > @@ -0,0 +1,16 @@
> > +// SPDX-License-Identifier: GPL-2.0
> > +#include "qcom-msm8974pro-samsung-klte.dts"
> 
> It's customary not to include .dts files, instead split the common parts
> into e.g. qcom-msm8974pro-samsung-klte-common.dtsi and include this in
> both the existing and the new one.

At the very beginning, I thought including .dts could make the patchset
tiny (considering that the difference among klte* is also tiny).

I agree that splitting the common parts into a .dtsi will make things
more elegant and make klte* DTs consistent with the style of qcom DTs.

Will do in v2. Thanks for your advice.

Thanks,
Rong