Re: [PATCH V2 1/7] sched/topology: Adding function partition_sched_domains_locked()

[Mathieu Poirier]

On 2 February 2018 at 03:19, Juri Lelli <juri.lelli@xxxxxxxxxx> wrote:
> Hi Mathieu,
>
> On 01/02/18 09:51, Mathieu Poirier wrote:
>> Introducing function partition_sched_domains_locked() by taking
>> the mutex locking code out of the original function.  That way
>> the work done by partition_sched_domains_locked() can be reused
>> without dropping the mutex lock.
>>
>> This patch doesn't change the functionality provided by the
>> original code.
>>
>> Signed-off-by: Mathieu Poirier <mathieu.poirier@xxxxxxxxxx>
>> ---
>
> [...]
>
>> +/*
>> + * Call with hotplug lock held
>
> Is this the one that we can actually check if it's locked with
>
> lockdep_assert_cpus_held()
>
> ?

Hi Juri,

You are correct - we could call lockdep_assert_cpus_held() but in my
opinion it would be in a separate patch and probably outside the scope
of this work.  The sole purpose of this patch is to get the
locking/unlocking operations of mutex sched_domains_mutex out of
function partition_sched_domains_locked().

Mathieu

>
>> + */
>> +void partition_sched_domains(int ndoms_new, cpumask_var_t doms_new[],
>> +                          struct sched_domain_attr *dattr_new)
>> +{
>> +     mutex_lock(&sched_domains_mutex);
>> +     partition_sched_domains_locked(ndoms_new, doms_new, dattr_new);
>>       mutex_unlock(&sched_domains_mutex);
>>  }
>
> Best,
>
> - Juri