Messages in this thread | | | Subject | Re: [PATCH 3/3] PM: domains: Add a ->dev_get_performance_state() callback to genpd | From | Dmitry Osipenko <> | Date | Tue, 7 Sep 2021 11:10:12 +0300 |
| |
07.09.2021 06:40, Viresh Kumar пишет: > On 06-09-21, 17:35, Dmitry Osipenko wrote: >> Viresh, are you okay with going back to the variant with the >> dev_pm_opp_sync() helper? > > I have missed a lot of stuff in between and wasn't following this > carefully as I thought my half was resolved :) > > Can you describe what to propose to do again ? From what I remember, > doing this one time from probe() is okay, doing it from > suspend/resume, not so much.
Hmm.. actually, it's not a problem to set up the performance state from probe() now with that recent change that was made to the PD core. [1]
[1] https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/?id=3c5a272202c28c1f9309566f206ba40787246149
And then we indeed don't need neither the dev_get_performance_state() callback, nor the dev_pm_opp_sync() helper.
The devm_tegra_core_dev_init_opp_table() already supports performance state syncing, so I will just need to call it after the RPM setup made by consumer driver, allowing PD core to set the rpm_pstate. I already gave it a quick test and it works perfectly.
Ulf, are you okay with abandoning the dev_get_performance_state() callback? We don't need it anymore.
| |