[PATCH 2/2] arm64: dts: rockchip: introduce rk3399-op1 operating points
heiko at sntech.de
Tue May 23 09:38:23 PDT 2017
Am Dienstag, 23. Mai 2017, 09:35:05 CEST schrieb Doug Anderson:
> On Tue, May 23, 2017 at 9:10 AM, Heiko Stuebner <heiko at sntech.de> wrote:
> > Am Dienstag, 23. Mai 2017, 08:57:01 CEST schrieb Doug Anderson:
> >> Hi,
> >> On Tue, May 23, 2017 at 2:51 AM, Heiko Stuebner <heiko at sntech.de> wrote:
> >> > The OP1 is a rk3399 variant used in ChromeOS devices with a slightly
> >> > higher frequency rating. Therefore introduce separate operting points,
> >> > from the ChromeOS tree, for it and use it in Gru devices.
> >> >
> >> > Signed-off-by: Heiko Stuebner <heiko at sntech.de>
> >> > ---
> >> > arch/arm64/boot/dts/rockchip/rk3399-gru.dtsi | 2 +-
> >> > arch/arm64/boot/dts/rockchip/rk3399-op1-opp.dtsi | 145 +++++++++++++++++++++++
> >> > 2 files changed, 146 insertions(+), 1 deletion(-)
> >> > create mode 100644 arch/arm64/boot/dts/rockchip/rk3399-op1-opp.dtsi
> >> The numbers here match what's in the Chrome OS tree and those were all
> >> approved by Rockchip.
> >> One slight nit is that the descriptions these two patches together
> >> make it look like this will be a no-op change for OP1 devices. This
> >> is not true. After things were posted upstream there have been
> >> several adjustments to the opp points in the Chrome OS tree. Since
> >> what you have currently matches the ToT Chrome OS tree, that should
> >> mean that there's a net difference. It would be nice to mention that
> >> in the commit message.
> > In the mainline kernel it is a no-op change, as we're still running with
> > the cpu-regulators disabled on gru ;-)
> Ah, that's true.
> > But to address your very valid point, I guess you mean something like
> > ----
> > The OP1 is a rk3399 variant used in ChromeOS devices with a slightly
> > higher frequency rating compared to the regular rk3399, but right now
> > the only available operating points are somewhere in between both
> > variants with both needing adjustments to actually fit their specs.
> That would be fine and would make it obvious that it's a no-op.
> Ironically, the OPP points that were in mainline were actually not
> totally in between. There were some numbers that were raised for
> both. :)
wording fun ... then we'll make that part a
"[...] but right now
the only available operating points don't match either variant
with both needing adjustments to actually fit their specs."
> > Therefore introduce separate operting points, from the ChromeOS kernel,
> > for the OP1 and use it on Gru devices.
> > ----
> > And I guess we should also swap patch ordering, first move Gru away from
> > the general opps and then adust them for the regular rk3399.
> Yeah, I think swapping the order would make it slightly better.
More information about the Linux-rockchip