On Mon, Oct 16, 2023 at 01:47:28PM +0100, Mark Brown wrote:
> On Sat, Oct 14, 2023 at 07:26:22PM +0200, Otto Pflüger wrote:
>
> > + for (i = 0; i < g_core->svc_version->num_services; i++) {
> > + struct avcs_svc_info *info;
> > +
> > + info = &g_core->svc_version->svc_api_info[i];
> > + if (info->service_id != APR_SVC_ADSP_CORE)
> > + continue;
> > +
> > + switch (info->version) {
> > + case AVCS_CMDRSP_Q6_ID_2_6:
> > + core->adsp_version = Q6_ADSP_VERSION_2_6;
> > + break;
> > + case AVCS_CMDRSP_Q6_ID_2_7:
> > + core->adsp_version = Q6_ADSP_VERSION_2_7;
> > + break;
> > + case AVCS_CMDRSP_Q6_ID_2_8:
> > + core->adsp_version = Q6_ADSP_VERSION_2_8;
> > + break;
> > + }
>
> This doesn't handle unknown versions at all.
The adsp_version is initialized to Q6_ADSP_VERSION_UNKNOWN in
q6core_probe, so it should stay unknown if it doesn't match any of
these values. I don't see any big problems here, but I agree that some
additional handling such as a warning message could be useful.
