On 10/13/26 2:43 PM, Avi Weiss wrote:
> qcom_glink_smem_register() passes the result of of_irq_get() directly
> to devm_request_irq(). Negative errors are converted to unsigned IRQ
> numbers, causing the IRQ request to return -EINVAL instead of the
> original error.
>
> Check the IRQ before requesting it. Preserve negative errors and
> return -EINVAL if the IRQ mapping fails.
>
> Fixes: caf989c350e8e ("rpmsg: glink: Introduce glink smem based transport")
> Signed-off-by: Avi Weiss <[email protected]>
> ---
> A dev_err_probe() diagnostic can be added for IRQ acquisition failures
> if preferred.
>
> drivers/rpmsg/qcom_glink_smem.c | 5 +++++
> 1 file changed, 5 insertions(+)
>
> diff --git a/drivers/rpmsg/qcom_glink_smem.c b/drivers/rpmsg/qcom_glink_smem.c
> index 35bb03e67ae8..010a7da64906 100644
> --- a/drivers/rpmsg/qcom_glink_smem.c
> +++ b/drivers/rpmsg/qcom_glink_smem.c
> @@ -312,6 +312,11 @@ struct qcom_glink_smem *qcom_glink_smem_register(struct
> device *parent,
> }
>
> smem->irq = of_irq_get(smem->dev.of_node, 0);
> + if (smem->irq <= 0) {
> + ret = smem->irq ?: -EINVAL;
> + goto err_put_dev;
> + }
platform_get_irq() lets us drop the zero-dance
Konrad
