Re: [Patch] cpufreq: tegra194: remove opp table in exit hook

[Viresh Kumar]

On 14-08-23, 23:43, Sumit Gupta wrote:
> On 10/08/23 11:01, Viresh Kumar wrote:
> > External email: Use caution opening links or attachments
> > 
> > 
> > On 09-08-23, 21:04, Sumit Gupta wrote:
> > > Add exit hook and remove OPP table when all the CPU's in a policy
> > > are offlined. It will fix the below error messages when onlining
> > > first CPU from a policy whose all CPU's were previously offlined.
> > > 
> > >   debugfs: File 'cpu5' in directory 'opp' already present!
> > >   debugfs: File 'cpu6' in directory 'opp' already present!
> > >   debugfs: File 'cpu7' in directory 'opp' already present!
> > > 
> > > Fixes: f41e1442ac5b ("cpufreq: tegra194: add OPP support and set bandwidth")
> > > Signed-off-by: Sumit Gupta <sumitg@xxxxxxxxxx>
> > > ---
> > >   drivers/cpufreq/tegra194-cpufreq.c | 13 +++++++++++++
> > >   1 file changed, 13 insertions(+)
> > > 
> > > diff --git a/drivers/cpufreq/tegra194-cpufreq.c b/drivers/cpufreq/tegra194-cpufreq.c
> > > index c90b30469165..66a9c23544db 100644
> > > --- a/drivers/cpufreq/tegra194-cpufreq.c
> > > +++ b/drivers/cpufreq/tegra194-cpufreq.c
> > > @@ -454,6 +454,8 @@ static int tegra_cpufreq_init_cpufreq_table(struct cpufreq_policy *policy,
> > >                if (ret < 0)
> > >                        return ret;
> > > 
> > > +             dev_pm_opp_put(opp);
> > > +
> > >                freq_table[j].driver_data = pos->driver_data;
> > >                freq_table[j].frequency = pos->frequency;
> > >                j++;
> > > @@ -508,6 +510,16 @@ static int tegra194_cpufreq_init(struct cpufreq_policy *policy)
> > >        return 0;
> > >   }
> > > 
> > > +static int tegra194_cpufreq_exit(struct cpufreq_policy *policy)
> > > +{
> > > +     struct device *cpu_dev = get_cpu_device(policy->cpu);
> > > +
> > > +     dev_pm_opp_remove_all_dynamic(cpu_dev);
> > > +     dev_pm_opp_of_cpumask_remove_table(policy->related_cpus);
> > > +
> > > +     return 0;
> > > +}
> > > +
> > >   static int tegra194_cpufreq_set_target(struct cpufreq_policy *policy,
> > >                                       unsigned int index)
> > >   {
> > > @@ -535,6 +547,7 @@ static struct cpufreq_driver tegra194_cpufreq_driver = {
> > >        .target_index = tegra194_cpufreq_set_target,
> > >        .get = tegra194_get_speed,
> > >        .init = tegra194_cpufreq_init,
> > > +     .exit = tegra194_cpufreq_exit,
> > >        .attr = cpufreq_generic_attr,
> > >   };
> > 
> > If it is only about hotplugging of the CPUs, then you can also do this I guess.
> > 
> > commit 263abfe74b5f ("cpufreq: dt: Implement online/offline() callbacks")

You should do this as well, this makes hotplugging paths much faster. i.e. on
top of this patch.

> > But since your driver is capable of being built as a module, I suggest you try
> > to build it as one and insert remove it multiple times. It must cause you some
> > trouble as you don't implement an .exit() before this patch.
> > 
> > Eventually, I think you need to do both, what this patch and 263abfe74b5f do.
> > Just that the reasons need to be correct for both the changes.
> > 
> > --
> > viresh
> 
> Hi Viresh,
> I got the same message on inserting and removing the module multiple times
> as you suggested. After applying this change, the message is not coming. So,
> the current change is resolving both scenarios as __cpufreq_offline() calls
> either exit() or offline().
> I can update the commit message to mention both scenarios and keep change as
> it is?
> 
>   cpufreq_remove_dev
>   |-__cpufreq_offline
>   |--tegra194_cpufreq_exit
> 
>   cpuhp_cpufreq_offline
>   |-__cpufreq_offline
>   |--tegra194_cpufreq_exit

-- 
viresh