Messages in this thread Patch in this message | | | Date | Wed, 9 May 2018 14:58:23 +0530 | From | Viresh Kumar <> | Subject | Re: [RFC PATCH] sched/cpufreq/schedutil: handling urgent frequency requests |
| |
On 09-05-18, 02:02, Joel Fernandes wrote: > On Wed, May 09, 2018 at 02:10:01PM +0530, Viresh Kumar wrote: > > Right, none of the above changes are required now. > > I didn't follow what you mean the changes are not required? I was developing > against Linus mainline. Also I replied to Rafael's comment in the other > thread.
At least for the shared policy case the entire sequence of sugov_should_update_freq() followed by sugov_update_commit() is executed from within spinlock protected region and you are using the same lock below. And so either the above two routines or the kthread routine below will execute at a given point of time.
So in case kthread has started doing the update and acquired the lock, the util update handler will wait until the time work_in_progress is set to false, that's not a problem we are trying to solve here.
And if kthread hasn't acquired the lock yet and util handler has started executing sugov_should_update_freq() ....
And ^^^ this is where I understood that your earlier change is actually required, so that we accumulate the latest updated next_freq value.
And with all that we wouldn't require a while loop in the kthread code. > > > > @@ -381,13 +381,23 @@ sugov_update_shared(struct update_util_data *hook, u64 time, unsigned int flags) > > > > static void sugov_work(struct kthread_work *work) > > > > { > > > > struct sugov_policy *sg_policy = container_of(work, struct sugov_policy, work); > > > > + unsigned int freq; > > > > + > > > > + /* > > > > + * Hold sg_policy->update_lock just enough to handle the case where: > > > > + * if sg_policy->next_freq is updated before work_in_progress is set to > > > > + * false, we may miss queueing the new update request since > > > > + * work_in_progress would appear to be true. > > > > + */ > > > > + raw_spin_lock(&sg_policy->update_lock); > > > > + freq = sg_policy->next_freq; > > > > + sg_policy->work_in_progress = false; > > > > + raw_spin_unlock(&sg_policy->update_lock); > > > > One problem we still have is that sg_policy->update_lock is only used > > in the shared policy case and not in the single CPU per policy case, > > so the race isn't solved there yet. > > True.. I can make the single CPU case acquire the update_lock very briefly > around sugov_update_commit call in sugov_update_single.
Rafael was very clear from the beginning that he wouldn't allow a spin lock in the un-shared policy case :)
-- viresh
| |