Thread (21 messages) flat view 21 messages, 4 authors, 9d ago

Re: [PATCH 3/6] dt-bindings: rtc: rzn1: interrupts are not required

From: sashiko-bot@kernel.org
Date: 2026-09-14 21:29:37
Also in: linux-devicetree

Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:
- [Medium] Globally removing `interrupts` and `interrupt-names` from `required` weakens schema validation for existing SoCs.
- [Medium] Making `interrupts` and `interrupt-names` optional without adding a schema dependency allows incomplete interrupt specifications.
--

commit 963405de649fb7157801d57b7a47cf959108ca2c
Author: Wolfram Sang [off-list ref]

dt-bindings: rtc: rzn1: interrupts are not required

This commit drops the interrupts and interrupt-names properties from the
required list in the rzn1-rtc binding. This change is made to support an
upcoming SoC where interrupts are not wired to the RTC.
quoted hunk ↗ jump to hunk
diff --git a/Documentation/devicetree/bindings/rtc/renesas,rzn1-rtc.yaml b/Documentation/devicetree/bindings/rtc/renesas,rzn1-rtc.yaml
index 7ea77cd0bee15..b426a34958d6d 100644
--- a/Documentation/devicetree/bindings/rtc/renesas,rzn1-rtc.yaml
+++ b/Documentation/devicetree/bindings/rtc/renesas,rzn1-rtc.yaml
@@ -52,8 +52,6 @@ properties:
 required:
   - compatible
   - reg
-  - interrupts
-  - interrupt-names
[Severity: Medium]
Does globally removing these properties weaken schema validation for existing
SoCs? Device trees for existing SoCs where interrupts are physically wired
could now incorrectly omit the interrupts and silently pass the schema
checker, resulting in incomplete hardware descriptions.

To accommodate the new SoC without affecting existing ones, should we
conditionally require these properties for the older compatible strings
using an allOf -> if -> then block?

[Severity: Medium]
By making these properties optional independently, could this allow
incomplete interrupt specifications? If a device tree specifies interrupts
but omits interrupt-names, schema validation will pass. However, the driver
uses platform_get_irq_byname_optional(), which will fail with -EINVAL at
runtime if names are missing, silently disabling alarm functionality.

Would it be better to include a dependencies block (e.g.,
interrupts: [ interrupt-names ]) to ensure complete descriptions when
interrupts are provided?
   - clocks
   - clock-names
   - power-domains
-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260914211826.73727-8-wsa+renesas@sang-engineering.com?part=3
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help