On Thu, 2023-11-30 at 14:20 +0200, Ilpo Järvinen wrote:
> On Tue, 28 Nov 2023, Srinivas Pandruvada wrote:
>
> > When a feature is read blocked, don't continue to read SST
> > information
> > and register with SST core.
> >
> > When the feature is write blocked, continue to offer read interface
> > for
> > SST parameters, but don't allow any operation to change state. A
> > state
> > change results from SST level change, feature change or class of
> > service
> > change.
> >
> > Signed-off-by: Srinivas Pandruvada
> > <srinivas.pandruvada@xxxxxxxxxxxxxxx>
> > ---
> > .../intel/speed_select_if/isst_tpmi_core.c | 25
> > +++++++++++++++++++
> > 1 file changed, 25 insertions(+)
> >
> > diff --git
> > a/drivers/platform/x86/intel/speed_select_if/isst_tpmi_core.c
> > b/drivers/platform/x86/intel/speed_select_if/isst_tpmi_core.c
> > index 0b6d2c864437..ed3a04d6c99c 100644
> > --- a/drivers/platform/x86/intel/speed_select_if/isst_tpmi_core.c
> > +++ b/drivers/platform/x86/intel/speed_select_if/isst_tpmi_core.c
> > @@ -514,6 +516,9 @@ static long isst_if_clos_param(void __user
> > *argp)
> > if (!power_domain_info)
> > return -EINVAL;
> >
> > + if (power_domain_info->write_blocked)
> > + return -EPERM;
> > +
>
> I don't understand this, doesn't this now -EPERM both
> _write_cp_info() AND
> _read_cp_info()??? Does _read_cp_info() also change state??
You have a point here. Unlike other SST features, CP access is useful
for OS as it know workloads and priorities.
But I will change for consistency.
Thanks,
Srinivas
>
> > if (clos_param.get_set) {
> > _write_cp_info("clos.min_freq",
> > clos_param.min_freq_mhz,
> > (SST_CLOS_CONFIG_0_OFFSET +
> > clos_param.clos * SST_REG_SIZE),
> > @@ -602,6 +607,9 @@ static long isst_if_clos_assoc(void __user
> > *argp)
> >
> > power_domain_info = &sst_inst-
> > >power_domain_info[punit_id];
> >
> > + if (power_domain_info->write_blocked)
> > + return -EPERM;
>
> Same here, this blocks also the get path?
>
>
