mbox series

[v2,0/2] DDR/L3 Scaling support on SC7280 SoCs

Message ID 1620111510-31455-1-git-send-email-sibis@codeaurora.org
Headers show
Series DDR/L3 Scaling support on SC7280 SoCs | expand

Message

Sibi Sankar May 4, 2021, 6:58 a.m. UTC 
The patch series adds support for DDR/L3 Scaling on SC7280 SoCs.

V2:
 * Add a new opp table for cpu 7 to account for the additional frequencies
   supported by it.

Depends on the following patch series:
L3 Provider Support: https://lore.kernel.org/lkml/1618556290-28303-1-git-send-email-okukatla@codeaurora.org/
CPUfreq Support: https://lore.kernel.org/lkml/1618020280-5470-2-git-send-email-tdas@codeaurora.org/
RPMH Provider Support: https://lore.kernel.org/lkml/1619517059-12109-1-git-send-email-okukatla@codeaurora.org/

It also depends on L3 and cpufreq dt nodes from the ^^ series to not have
overlapping memory regions.

Sibi Sankar (2):
  cpufreq: blacklist SC7280 in cpufreq-dt-platdev
  arm64: dts: qcom: sc7280: Add cpu OPP tables

 arch/arm64/boot/dts/qcom/sc7280.dtsi | 215 +++++++++++++++++++++++++++++++++++
 drivers/cpufreq/cpufreq-dt-platdev.c |   1 +
 2 files changed, 216 insertions(+)

Comments

Sibi Sankar May 5, 2021, 10:11 a.m. UTC | #1 
Hey Doug,

Thanks for the review!

On 2021-05-05 01:32, Doug Anderson wrote:
> Hi,
> 
> On Mon, May 3, 2021 at 11:59 PM Sibi Sankar <sibis@codeaurora.org> 
> wrote:
>> 
>> +       cpu0_opp_table: cpu0_opp_table {
>> +               compatible = "operating-points-v2";
>> +               opp-shared;
>> +
>> +               cpu0_opp1: opp-300000000 {
> 
> It seems like it might be nicer to give the node labels a less
> arbitrary name. How about?
> 
> cpu0_opp_300mhz: opp-300000000
> 
> That has advantes:
> 
> * If, for some reason, you have to mess with some operating point in
> another dts it'll be less fragile.
> 
> * It'll make diffing easier between SoCs.
> 
> * If you end up putting a new operating point in the middle you don't
> need to rename everything below.

sure makes sense, will fix it in v3.

> 
> Other than that, I can't say that I'm a huge expert on the
> interconnect stuff and whether those make sense, but I'm still OK
> with:
> 
> Reviewed-by: Douglas Anderson <dianders@chromium.org>