Hi Julien,
On 01/10/2024 12:59, Julien Grall wrote:
> Hi Andrei,
>
> On 30/09/2024 12:47, Andrei Cherechesu (OSS) wrote:
>> +/*
>> + * Generic handler for SCMI-SMC requests, currently only forwarding the
>> + * request to FW running at EL3 if it came from Dom0. Is called from the
>> vSMC
>> + * layer for SiP SMCs, since SCMI calls are usually provided this way.
>> + * Can also be called from `platform_smc()` plat-specific callback.
>> + *
>> + * Returns true if SMC was handled (regardless of response), false
>> otherwise.
>> + */
>> +bool scmi_handle_smc(struct cpu_user_regs *regs)
>> +{
>> + struct arm_smccc_res res;
>> +
>> + /* Only the hardware domain should use SCMI calls */
>> + if ( !is_hardware_domain(current->domain) )
>> + {
>> + gprintk(XENLOG_ERR, "SCMI: Unprivileged d%d cannot use SCMI.\n",
>> + current->domain->domain_id);
>> + return false;
>> + }
>> +
>> + /* For the moment, forward the SCMI Request to FW running at EL3 */
>> + arm_smccc_1_1_smc(scmi_smc_id,
>
> Actually, shouldn't this be get_user_reg(regs, 0) so we don't rely on
> scmi_handle_smc() to be called only when this is equal to scmi_smc_id?
>
> The other option is to move the check for scmi_smc_id in this function.
I'll move the check for scmi_smc_id in this function and make it static,
to avoid exposing unnecessary complexity to the users of this layer.
Thanks for the suggestion & review.
>
> Cheers,
>
Regards,
Andrei Cherechesu
