Re: [PATCH v3 3/5] OPP: Improve require-opps linking

From: Saravana Kannan <saravanak@google.com>
To: Viresh Kumar <viresh.kumar@linaro.org>
Cc: MyungJoo Ham <myungjoo.ham@samsung.com>,
	Kyungmin Park <kyungmin.park@samsung.com>,
	Chanwoo Choi <cw00.choi@samsung.com>,
	Viresh Kumar <vireshk@kernel.org>, Nishanth Menon <nm@ti.com>,
	Stephen Boyd <sboyd@kernel.org>,
	"Rafael J. Wysocki" <rjw@rjwysocki.net>,
	Sibi Sankar <sibis@codeaurora.org>,
	Android Kernel Team <kernel-team@android.com>,
	Linux PM <linux-pm@vger.kernel.org>,
	LKML <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH v3 3/5] OPP: Improve require-opps linking
Date: Tue, 23 Jul 2019 13:36:03 -0700	[thread overview]
Message-ID: <CAGETcx_4N8uBbA0yhMdcTHpw_DkMVqQwbZeSDTLOzR7t-uZnVg@mail.gmail.com> (raw)
In-Reply-To: <20190723102842.t2s45zzylsjuccm4@vireshk-i7>

Resending again due to accidental HTML (minor rewording/typo fixes too).

On Tue, Jul 23, 2019 at 3:28 AM Viresh Kumar <viresh.kumar@linaro.org> wrote:
>
> $subject doesn't have correct property name.
>
> On 17-07-19, 15:23, Saravana Kannan wrote:
> > Currently, the linking of required-opps fails silently if the
> > destination OPP table hasn't been added before the source OPP table is
> > added. This puts an unnecessary requirement that the destination table
> > be added before the source table is added.
> >
> > In reality, the destination table is needed only when we try to
> > translate from source OPP to destination OPP. So, instead of
> > completely failing, retry linking the tables when the translation is
> > attempted.
> >
> > Signed-off-by: Saravana Kannan <saravanak@google.com>
> > ---
> >  drivers/opp/core.c | 32 +++++++++++-----
> >  drivers/opp/of.c   | 91 ++++++++++++++++++++++------------------------
> >  drivers/opp/opp.h  |  5 +++
> >  3 files changed, 71 insertions(+), 57 deletions(-)
>
> Here is the general feedback and requirements I have:
>
> - We shouldn't do it from _set_required_opps() but way earlier, it
>   shouldn't affect the fast path (where we change the frequency).

I don't think there's any other intermediate OPP API call where we can
try to lazily link the tables. Also if the tables are already fully
linked, I don't think this is really going to slow down the fast path
in anyway (because it's just a few NULL checks).

If you have other ideas, I'm willing to look at it. But as it is right
now seems the best to me.

> - Programming required-opps for half of the properties isn't correct,
>   i.e. in case only few of the required-opps are parsed until now. So
>   setting of rate shouldn't even start unless the OPP table is fully
>   initialized with all required-opps in it.

Ok, I can fix this.

-Saravana