Re: [PATCH net 08/22] can: cc770: platform_get_irq(): propagate the error

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] can: cc770: missing IRQ returns -ENXIO instead of intended -ENODEV
--

commit d39a1b15e30aece42be9b57846cef8999da6e194
Author: Quchaosheng <quchaosheng000406@163.com>

can: cc770: platform_get_irq(): propagate the error

cc770_platform_probe() turns every failure of platform_get_irq() into
-ENODEV, including -EPROBE_DEFER. Return the error unchanged so that a
deferred probe stays deferred, and keep -ENODEV for the case where no
interrupt is described at all.

Link: https://patch.msgid.link/20260914055611.495105-1-quchaosheng000406@163.com

> diff --git a/drivers/net/can/cc770/cc770_platform.c b/drivers/net/can/cc770/cc770_platform.c
> index b6c4f02ffb97a..ac218acedb101 100644
> --- a/drivers/net/can/cc770/cc770_platform.c
> +++ b/drivers/net/can/cc770/cc770_platform.c
> @@ -156,8 +156,13 @@ static int cc770_platform_probe(struct platform_device *pdev)
>  	int err, irq;
>  
>  	mem = platform_get_resource(pdev, IORESOURCE_MEM, 0);
> +	if (!mem)
> +		return -ENODEV;
> +
>  	irq = platform_get_irq(pdev, 0);
> -	if (!mem || irq <= 0)
> +	if (irq < 0)
> +		return irq;
> +	if (!irq)
>  		return -ENODEV;

[Severity: Medium]
The commit message states the intent to "keep -ENODEV for the case where no
interrupt is described at all." Does this implementation achieve that?

When an interrupt is missing, platform_get_irq() returns -ENXIO. Since
-ENXIO is less than 0, the new check will return -ENXIO directly, bypassing
the intended -ENODEV return.

Additionally, is the check for !irq dead code? platform_get_irq() never
returns 0; if an underlying lookup returns 0, it triggers a WARN and
returns -EINVAL.

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260928193312.553632-1-mkl@pengutronix.de?part=8