Skip to contentSkip to conversation

dt-bindings: gpio: pca95xx: add onnn,pca9655 compatible

PATCH In review Rosen Penev 4 participants 5 messages Oct 4 – Oct 9

Subsystems (2)
Device tree bindings GPIO
lore
Share & tools
lore

Rosen Penev

author
#0 lore raw

PATCH v2 1/1 dt-bindings: gpio: pca95xx: add onnn,pca9655 compatible

The pca9655 is the On Semi variant of the pca9555 16-bit I2C GPIO expander. "onnn,pca9655" is already used standalone by in-tree DTS (armada-388-clearfog, armada-388-helios4) and has a dedicated entry in the pca953x driver's OF match table, but was never documented, causing dtbs_check failures.

Document it alongside the other standalone entries such as onnn,pca9654, without an nxp,pca9555 fallback, so that the existing DTS remain valid.

Reported-by kernel test robot <[email protected]>
Closes https://lore.kernel.org/oe-kbuild-all/[email protected]/ open here
Assisted-by LLM
Signed-off-by Rosen Penev <[email protected]>

---
v2: rewrite description. Use tags from kernel test bot.

 Documentation/devicetree/bindings/gpio/gpio-pca95xx.yaml | 1 +
 1 file changed, 1 insertion(+)
Documentation/devicetree/bindings/gpio/gpio-pca95xx.yaml +1 -0
index 4631388a7d91..563f61e6e543 100644
@@ -67,6 +67,7 @@ properties:
               - nxp,pcal9555a
               - onnn,cat9554
               - onnn,pca9654
+              - onnn,pca9655
               - ti,pca6107
               - ti,pca9536
               - ti,tca6408
Signature · 1 line
2.56.0
AI review
#1 lore raw
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider: - [Medium] dt-bindings: commit message falsely claims in-tree DTS usage --

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] dt-bindings: commit message falsely claims in-tree DTS usage
--

commit 7b4c1f570780f70afdcc949de9998c051c2323e5
Author: Rosen Penev <[email protected]>

dt-bindings: gpio: pca95xx: add onnn,pca9655 compatible

This commit documents the onnn,pca9655 compatible string for the On Semi variant of the pca9555 16-bit I2C GPIO expander. It adds the string alongside other standalone entries without a fallback.

Closes https://lore.kernel.org/oe-kbuild-all/[email protected]/ open here
Quoted 5 lines

The pca9655 is the On Semi variant of the pca9555 16-bit I2C GPIO expander. "onnn,pca9655" is already used standalone by in-tree DTS (armada-388-clearfog, armada-388-helios4) and has a dedicated entry in the pca953x driver's OF match table, but was never documented, causing dtbs_check failures.

[Severity: Medium]
Is this claim about in-tree DTS usage accurate?

Looking at the armada-388-clearfog.dtsi file, the compatible string appears to be only used within a comment:

/*
* This is how it should be:
* compatible = "onnn,pca9655", "nxp,pca9555";
* but you can't do this because of the way I2C works. */
compatible = "nxp,pca9555";

Because the string is only present in a comment and the active property is nxp,pca9555, it would not trigger dtbs_check failures. Could the commit message be updated to reflect why the binding actually needs to be added?

Signature · 1 line
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=1
author
#2 lore raw
Quoted 35 lines · On Sun, Oct 4, 2026 at 12:17 PM <[email protected]> wrote:

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] dt-bindings: commit message falsely claims in-tree DTS usage
--

commit 7b4c1f570780f70afdcc949de9998c051c2323e5
Author: Rosen Penev <[email protected]>

dt-bindings: gpio: pca95xx: add onnn,pca9655 compatible

This commit documents the onnn,pca9655 compatible string for the On Semi variant of the pca9555 16-bit I2C GPIO expander. It adds the string alongside other standalone entries without a fallback.

Closes https://lore.kernel.org/oe-kbuild-all/[email protected]/ open here
Quoted 5 lines

The pca9655 is the On Semi variant of the pca9555 16-bit I2C GPIO expander. "onnn,pca9655" is already used standalone by in-tree DTS (armada-388-clearfog, armada-388-helios4) and has a dedicated entry in the pca953x driver's OF match table, but was never documented, causing dtbs_check failures.

[Severity: Medium]
Is this claim about in-tree DTS usage accurate?

Looking at the armada-388-clearfog.dtsi file, the compatible string appears to be only used within a comment:

/*
* This is how it should be:
* compatible = "onnn,pca9655", "nxp,pca9555";
* but you can't do this because of the way I2C works. */
compatible = "nxp,pca9555";

Quoted 3 lines

Because the string is only present in a comment and the active property is nxp,pca9555, it would not trigger dtbs_check failures. Could the commit message be updated to reflect why the binding actually needs to be added?

This is outdated. 6dfda79476538a9aad1d2dac45453e62c6939522 in linux-next fixes this.

Quoted 3 lines

--
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=1

#3 lore raw
In reply to Rosen Penev
Quoted 2 lines · Rosen Penev

The pca9655 is the On Semi variant of the pca9555 16-bit I2C GPIO expander. "onnn,pca9655" is already used standalone by in-tree DTS

"Already" is a stretch - YOU added it here:
https://lore.kernel.org/all/[email protected]/

and that patch explained nothing about missing compatible documentation. That patch should clearly document missing binding, so the SoC maintainer would not pick it up.

Quoted 7 lines

(armada-388-clearfog, armada-388-helios4) and has a dedicated entry in the pca953x driver's OF match table, but was never documented, causing dtbs_check failures.

Document it alongside the other standalone entries such as onnn,pca9654, without an nxp,pca9555 fallback, so that the existing DTS remain valid.

There is no existing DTs, git grep. Next is not existing, because I can reject that DTS pull.

And what about clear comment in mainline:

* This is how it should be:
* compatible = "onnn,pca9655", "nxp,pca9555";
* but you can't do this because of the way I2C works.

Best regards,
Krzysztof