11:43 <mal> aka_[m]: I tried to look in that apcs, I can't directly use msm8916 code because it has a53 while msm8226 is a7, there is a7pll in mainline but it appears to be for some different platform currently and might need some changes based on a quick look

11:43 <minecrell> mal: the a53pll code works for the cortex-a7 in msm8909 so the name doesn't really mean something

11:44 <mal> ah, ok

13:13 <aka_[m]> These names makes no sense

13:13 <aka_[m]> A7pll refers to lucid pll which is alpha type

13:14 <aka_[m]> It should probably sit inside apss-ipq-pll.c driver

14:47 <mal> aka_[m]: no luck yet getting that apcs to work, device hangs at boot for a moment and then reboots after display turns blue

15:26 <aka_[m]> Could try dumping register values

15:26 <aka_[m]> Do you have changes anywhere?

15:27 <mal> just a moment, I can push that hacky branch

15:27 <mal> I also might have found the commits that break my device on 6.4 rcs

15:28 <mal> when I add couple of relatively simple changes from 6.4 to 6.3 it breaks the same way

15:28 <mal> https://lore.kernel.org/linux-arm-msm/20230121192540.9177-1-rayyan@ansari.sh/

15:28 <mal> that is the patchset that causes problems for me

15:29 <mal> just not sure why

15:32 <mal> aka_[m]: https://github.com/mlehtima/linux/commits/qcom-msm8226-6.3.y-apcs

15:42 <aka_[m]> minecrell[m]: woukd that work without wiring it to cpus and opp tables?

15:43 <mal> a very good question

15:44 <minecrell> I think it's extremely unlikely that you would not have something vote for XO permanently

15:45 <minecrell> still, you can try https://lore.kernel.org/linux-arm-msm/20230501-topic-rpmcc_xo_a-v1-0-93f18e47b607@linaro.org/ to be sure

15:45 <minecrell> then at least the cpu vote will be always there independent of drivers

15:48 <mal> minecrell: that fixed booting with those patches I linked, I will test 6.4 also to be sure it works

15:51 <aka_[m]> mal: I would try reading to which source apcs mux is set

15:51 <aka_[m]> Not sure whicb it was

15:51 <aka_[m]> 0x50 or 0x54

15:52 <aka_[m]> From apcs base

15:52 <aka_[m]> I'm sick in bed so cannot see a flat

15:56 <mal> aka_[m]: I started to wonder about this https://github.com/LineageOS/android_kernel_samsung_msm8226/blob/lineage-18.1/arch/arm/boot/dts/msm8226/msm8226.dtsi#L1125

15:56 <minecrell> mal: well you'd need to scale the cpu voltage for higher cpu freqs which is a bit "involved"

16:01 <mal> minecrell: also 6.4 kernel now works on my device (with the usb related reverts in addition to those patches you gave)

16:02 <mal> so those patches are quite critical

16:04 <mal> I can reply with tested-by to that patchset later today

17:10 <aka_[m]> minecrell: on 8976 i was able to reach max just by setting min-max for spmi apc regulator

17:16 <Rayyan> did someone ping me?

17:16 <aka_[m]> .

17:16 <Rayyan> oh it was from a link

18:01 <minecrell> aka_[m]: well if you're lucky you can run at low voltages with high freqs and maybe even save power

18:01 <minecrell> but that's like "undervolting", might become unstable in some situations

18:41 <DavidHeidelberg[m]> @aka I would add it also depends on the piece of silicon you have = depends on where it was on wafer :D