Re: [PATCH v3 6/7] thermal: netlink: Add a new event to notify CPU capabilities change

[Rafael J. Wysocki]

On Wed, Jan 12, 2022 at 10:32 PM Srinivas Pandruvada
<srinivas.pandruvada@xxxxxxxxxxxxxxx> wrote:
>
> On Wed, 2022-01-12 at 20:25 +0100, Rafael J. Wysocki wrote:
> > On Thu, Jan 6, 2022 at 3:49 AM Ricardo Neri
> > <ricardo.neri-calderon@xxxxxxxxxxxxxxx> wrote:
> > > From: Srinivas Pandruvada <srinivas.pandruvada@xxxxxxxxxxxxxxx>
> > >
> > > Add a new netlink event to notify change in CPU capabilities in
> > > terms of
> > > performance and efficiency.
> > >
> > > Firmware may change CPU capabilities as a result of thermal events
> > > in the
> > > system or to account for changes in the TDP (thermal design power)
> > > level.
> > >
> > > This notification type will allow user space to avoid running
> > > workloads
> > > on certain CPUs or proactively adjust power limits to avoid future
> > > events.
> > >
> > > The netlink message consists of a nested attribute
> > > (THERMAL_GENL_ATTR_CPU_CAPABILITY) with three attributes:
> > >
> > >  * THERMAL_GENL_ATTR_CPU_CAPABILITY_ID (type u32):
> > >    -- logical CPU number
> > >  * THERMAL_GENL_ATTR_CPU_CAPABILITY_PERFORMANCE (type u32):
> > >    -- Scaled performance from 0-1023
> > >  * THERMAL_GENL_ATTR_CPU_CAPABILITY_EFFICIENCY (type u32):
> > >    -- Scaled efficiency from 0-1023
> > >
> > > Cc: Andi Kleen <ak@xxxxxxxxxxxxxxx>
> > > Cc: Aubrey Li <aubrey.li@xxxxxxxxxxxxxxx>
> > > Cc: Lukasz Luba <lukasz.luba@xxxxxxx>
> > > Cc: Tim Chen <tim.c.chen@xxxxxxxxxxxxxxx>
> > > Cc: "Ravi V. Shankar" <ravi.v.shankar@xxxxxxxxx>
> > > Reviewed-by: Len Brown <len.brown@xxxxxxxxx>
> > > Signed-off-by: Srinivas Pandruvada <
> > > srinivas.pandruvada@xxxxxxxxxxxxxxx>
> >
> > Of course, I need to know if Daniel and Lukasz agree with this patch.
> >
> I pinged Daniel offline. I accommodated comments from Lukasz.
>
> > > ---
> > >
>
> [...]
>
> > > +static int thermal_genl_event_cpu_capability_change(struct param
> > > *p)
> > > +{
> > > +       struct cpu_capability *cpu_cap = p->cpu_capabilities;
> > > +       struct sk_buff *msg = p->msg;
> > > +       struct nlattr *start_cap;
> > > +       int i, ret;
> > > +
> > > +       start_cap = nla_nest_start(msg,
> > > THERMAL_GENL_ATTR_CPU_CAPABILITY);
> > > +       if (!start_cap)
> > > +               return -EMSGSIZE;
> > > +
> > > +       for (i = 0; i < p->cpu_capabilities_count; ++i) {
> > > +               if (nla_put_u32(msg,
> > > THERMAL_GENL_ATTR_CPU_CAPABILITY_ID,
> > > +                               cpu_cap->cpu)) {
> > > +                       ret = -EMSGSIZE;
> > > +                       goto out_cancel_nest;
> > > +               }
> > > +               if (nla_put_u32(msg,
> > > THERMAL_GENL_ATTR_CPU_CAPABILITY_PERFORMANCE,
> > > +                               cpu_cap->performance)) {
> > > +                       ret = -EMSGSIZE;
> > > +                       goto out_cancel_nest;
> > > +               }
> > > +               if (nla_put_u32(msg,
> > > THERMAL_GENL_ATTR_CPU_CAPABILITY_EFFICIENCY,
> > > +                               cpu_cap->efficiency)) {
> > > +                       ret = -EMSGSIZE;
> > > +                       goto out_cancel_nest;
> > > +               }
> > > +               ++cpu_cap;
> > > +       }
> > > +
> > > +       nla_nest_end(msg, start_cap);
> > > +
> > > +       return 0;
> > > +out_cancel_nest:
> > > +       nla_nest_cancel(msg, start_cap);
> > > +
> > > +       return ret;
> >
> > It looks like ret is never different from -EMSGSIZE here, so I'd just
> > return that error and drop the ret variable.
> >
> ret is initialized for every case when it will be returned.

Right, but it is redundant.

> But agree
> that we can just return -EMSGSIZE as there is no other return value
> here.
>
> > > +}
> > > +
> > >
>
> [...]
>
> > > +struct cpu_capability {
> >
> > I'm wondering if the struct name is not too generic as the purpose it
> > is used for is rather narrow and specific.
> >
> This was named something else before. What about cpu_energy_perf_cap?

Because it is only used in the thermal_genl_cpu_capability_event()
interface, it would be good to make the name reflect that IMO.
Something like thermal_genl_cpu_caps would work in this regard.