Re: [PATCH] coresight: cti: Check if the CPU activated for the CPU CTI

[Suzuki K Poulose]

On 27/03/2023 10:49, Tao Zhang wrote:
> Check whether the CPU corresponding to the CPU CTI is activated.
> If it is not activated, the CPU CTI node should not exist, and
> an error will be returned in the initialization function.
>
> Signed-off-by: Tao Zhang <quic_taozha@xxxxxxxxxxx>
> ---
>   drivers/hwtracing/coresight/coresight-cti-core.c | 6 ++++--
>   1 file changed, 4 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/hwtracing/coresight/coresight-cti-core.c b/drivers/hwtracing/coresight/coresight-cti-core.c
> index 277c890..aaa83ae 100644
> --- a/drivers/hwtracing/coresight/coresight-cti-core.c
> +++ b/drivers/hwtracing/coresight/coresight-cti-core.c
> @@ -899,10 +899,12 @@ static int cti_probe(struct amba_device *adev, const struct amba_id *id)
>   	drvdata->config.hw_powered = true;
>   
>   	/* set up device name - will depend if cpu bound or otherwise */
> -	if (drvdata->ctidev.cpu >= 0)
> +	if (drvdata->ctidev.cpu >= 0) {
> +		if (!cpu_active(drvdata->ctidev.cpu))
> +			return -ENXIO;
>   		cti_desc.name = devm_kasprintf(dev, GFP_KERNEL, "cti_cpu%d",
>   					       drvdata->ctidev.cpu);

But why ? As long as we do not enable or touch any CPU specific bits in 
the probe, why do we need to fail this ? What are you trying to fix ?

Please could you share the log if you are hitting something ? This looks
like masking a problem.

Suzuki


> -	else
> +	} else
>   		cti_desc.name = coresight_alloc_device_name(&cti_sys_devs, dev);
>   	if (!cti_desc.name)
>   		return -ENOMEM;