<konradybcio>
No, but I know a guy that knows this and that about PEP
<konradybcio>
Test test my matrix client seems to have randomly died for the past 20 minutes
<bamse>
no worries, we waited for you to come back ;)
pevik_ has quit [Ping timeout: 480 seconds]
<Tooniis[m]>
bamse: my msm8996 locks up something like 4/5 times with both cpuidle and cpufreq enabled, so I disabled cpuidle to make it work. maybe you have something similar?
Daanct12 is now known as Danct12
marvin24 has joined #linux-msm
marvin24_ has quit [Ping timeout: 480 seconds]
<Danct12>
konradybcio, do i need to make those pon keys disabled?
<Danct12>
i plan to submit a v3 tonight or tomorrow
hfink_ has quit []
hfink has joined #linux-msm
svarbanov has quit [Ping timeout: 480 seconds]
link2xt has quit [Ping timeout: 480 seconds]
<konradybcio>
The idea is that you disable it in the common pmic tree because not every device uses them
<konradybcio>
Some just really on gpio-keys
<konradybcio>
the configuration is the same across all (if used) but what key is electrically wired, that's up tho the vendor
<konradybcio>
from the PMIC perspective, it doesn't care what it is, might as well be 2 wires that you touch with a graphite stick
<konradybcio>
(ugh, with some resistors, probably :P)
pevik_ has joined #linux-msm
<Danct12>
konradybcio, so basically just set pon keys as disabled by default
<Danct12>
and then enable it in specific tree?
<konradybcio>
yes, you then only need to add the keycode and statusokay in device-specific dt
<Danct12>
alright, thanks. i'll push a v3 this afternoon when i'm done and tested it
<minecrell>
konradybcio: Being able to start cores via PSCI is already quite nice, since it's the really hacky part that requires changes to arm64 core code.
<minecrell>
The cpuidle part is abstracted through the cpuidle subsystem so allowing other drivers there has absolutely minimal maintenance overhead
<konradybcio>
Breaking news: "apparently" sony software is *not* nice (not only) in this regard
<konradybcio>
But hey, 8992/4 lumias can even do EL2 with some setup
<minecrell>
konradybcio: then without the hacky EL2 PSCI?
<konradybcio>
Yes, PSCI is good enough to do nonhacky-EL1-normal-startup
<konradybcio>
I wonder if/how cpufreq (once/if that gets to exist one day..) will be impacted by only having one core on half the boards..