[RFC PATCH 3/7] firmware: arm_scmi: Add Telemetry protocol support

[Cristian Marussi]

On Fri, Jun 20, 2025 at 11:46:15PM +0300, Dan Carpenter wrote:
> On Fri, Jun 20, 2025 at 08:28:09PM +0100, Cristian Marussi wrote:
> > +static int
> > +scmi_telemetry_protocol_attributes_get(const struct scmi_protocol_handle *ph,
> > +				       struct telemetry_info *ti)
> > +{

Hi Dan,

thanks for having a look.

> > +	int ret;
> > +	struct scmi_xfer *t;
> > +	struct scmi_msg_resp_telemetry_protocol_attributes *resp;
> > +
> > +	ret = ph->xops->xfer_get_init(ph, PROTOCOL_ATTRIBUTES,
> > +				      0, sizeof(*resp), &t);
> > +	if (ret)
> > +		return ret;
> > +
> > +	resp = t->rx.buf;
> > +	ret = ph->xops->do_xfer(ph, t);
> > +	if (!ret) {
> > +		__le32 attr = resp->attributes;
> > +
> > +		ti->info.num_de = le32_to_cpu(resp->de_num);
> > +		ti->info.num_groups = le32_to_cpu(resp->groups_num);
> > +		for (int i = 0; i < SCMI_TLM_MAX_DWORD; i++)
> > +			ti->info.de_impl_version[i] =
> > +				le32_to_cpu(resp->de_implementation_rev_dword[i]);
> > +		ti->info.single_read_support = SUPPORTS_SINGLE_READ(attr);
> > +		ti->info.continuos_update_support = SUPPORTS_CONTINUOS_UPDATE(attr);
> > +		ti->info.per_group_config_support = SUPPORTS_PER_GROUP_CONFIG(attr);
> > +		ti->info.reset_support = SUPPORTS_RESET(attr);
> > +		ti->info.fc_support = SUPPORTS_FC(attr);
> > +		ti->num_shmti = le32_get_bits(attr, GENMASK(15, 0));
> > +		/* Allocate DEs descriptors */
> > +		ti->info.des = devm_kcalloc(ph->dev, ti->info.num_de,
> > +					    sizeof(*ti->info.des), GFP_KERNEL);
> > +		if (!ti->info.des)
> > +			ret = -ENOMEM;
> > +
> > +		/* Allocate DE GROUPS descriptors */
> > +		ti->info.des_groups = devm_kcalloc(ph->dev, ti->info.num_groups,
> > +						   sizeof(*ti->info.des_groups),
> > +						   GFP_KERNEL);
> > +		if (!ti->info.des_groups)
> > +			ret = -ENOMEM;
> 
> It the allocation fails we need to jump to the ->xfer_put
> 
> > +
> > +		for (int i = 0; i < ti->info.num_groups; i++)
> > +			ti->info.des_groups[i].id = i;
> 
> otherwise it leads to a NULL dereference.
>

Indeed...I will fix.

There are a few more of this bugs around probably being an RFC...bots also are
complaining a lot at this series indeed :D

Thanks,
Cristian