cyrozap has quit [Quit: ZNC 1.8.2+deb3.1+deb12u1 - https://znc.in]
cyrozap has joined #linux-rockchip
ldevulder has joined #linux-rockchip
raster has joined #linux-rockchip
<hrw>
[ +0.003515] cpufreq: cpufreq_online: CPU0: Running at unlisted initial frequency: 816000 KHz, changing to: 1008000 KHz
<hrw>
Linux does not have 816MHz speed?
<qschulz>
hrw: upstream kernel?
<hrw>
qschulz: yes. 6.11-rc6
<qschulz>
hrw: could be that the CPU really is running at 816KHz by default when booting
<hrw>
qschulz: and u-boot 2024.10-git
<qschulz>
nothing changes it in TPL/U-Boot
<qschulz>
so it stays there and then the kernel reads the clock frequency and sees it doesn't match the expectation and switches it for something it knows?
<hrw>
looks like
<qschulz>
(also, you probably want to send a patch to fix the unit in the log message :) )
<hrw>
ah. KHz -> kHz?
<qschulz>
hrw: no, KHz -> Hz :)
<qschulz>
otherwise RK3588 really is FAAAAAAAAAAAAAAAAAAAAAAAAST
<qschulz>
ah no
<qschulz>
After KHz there is MHz Quentin
<hrw>
816 000 Hz == 816 kHz == 0.8 MHz
<qschulz>
should have slept more last night
<hrw>
;)
<qschulz>
sorry for the noise
<hrw>
qschulz: still KHz -> kHz it should be
<qschulz>
hrw: meh
<hrw>
not meh. next step is mHz instead of MHz and people wondering about minihertz
<qschulz>
diederik: yes, this was a deliberate choice
<diederik>
ok
<qschulz>
the point raised by the contributor was that frequencies for the same voltage wouldn't bring much benefit
<qschulz>
so instead they decided to keep only the highest frequency "per voltage" point
<hrw>
if power use is the same why bother with lower freqs?
<qschulz>
hrw: I didn't say the power use was the same, I said the voltage was the same. I don't have data to say it doesn't impact power consumption to run at a lower freq :)
<hrw>
k
Spirit532 has quit [Read error: Connection reset by peer]
Spirit532 has joined #linux-rockchip
Stat_headcrabed has joined #linux-rockchip
mps has quit [Changing host]
mps has joined #linux-rockchip
<dsimic>
hrw: qschulz: diederik: my plan is to do detailed testing, to prove that the "redundant" OPPs are actually useful
<dsimic>
i.e. the "same-voltage, lower-frequency" OPPs
<dsimic>
we've got them already defined for other Rockchip SoCs, and it's just the IPA governor that logs warnings about them being inefficient
<dsimic>
and not everybody uses IPA, in the end
<hrw>
dsimic: thanks
<diederik>
thx
<dsimic>
anytime :)
Stat_headcrabed has quit [Quit: Stat_headcrabed]
Stat_headcrabed has joined #linux-rockchip
raster has quit [Quit: Gettin' stinky!]
raster has joined #linux-rockchip
Stat_headcrabed has quit [Quit: Stat_headcrabed]
testaccount has joined #linux-rockchip
testaccount has quit [Client Quit]
testaccount has joined #linux-rockchip
testaccount has quit [Quit: Going offline for now]
testaccount has joined #linux-rockchip
testaccount has quit [Client Quit]
<Kwiboo>
diederik: I hope you noticed that I pushed rebased hdmi series yesterday, including 2 new fixes/patches, would appreciate if you could re-test the two hdmi series :-)
<Kwiboo>
one of the fixes should address obbardc's issue with 4K -> 1080p hotplugging
<diederik>
I saw them arrive in my Inbox :) I don't know how I can check if the first series (CEC/EDID) changes/fixes something though
<Kwiboo>
hehe, yeah that series is fixing very rare issues, so hard to validate, as long as you continue to get hdmi signal I am happy :-)
<diederik>
My HDMI and/or power cables aren't very long so I don't know how I could even try it
<diederik>
I could try one and then the other, but not the hotplug thing
<diederik>
Which would exactly NOT test the thing which is supposed to be fixed. So I guess you'd have to ask obbardc to check again?
<Kwiboo>
e.g. one issue it should fix is enabling display before EDID is read during bootup, remember that I have seen this issue long time ago, sometimes boot console / splash on libreelec would show at 1027x768 res or similar while others times it was native resolution
<diederik>
I did notice (and describe) that ... and that I actually liked the (apparently?) 'wrong' behavior
<diederik>
Reading console msgs on a 4k screen is only possible with a <50 cm distance from the screen ...
<dsimic>
perhaps a different text resolution could be configured?
<diederik>
I didn't know it had sth to do with EDID though. But if the tiny console msgs are back, I guess that means the patch is working ;-P
<qschulz>
diederik: gotta put a magnifier at hand on the couch ;)
<Kwiboo>
hehe, yeah, it was 'wrong' in that the screen/hdmi signal wrongly would/could be enabled before the drm core had finalized initialization, hotplug irq was triggered during probing not after probing had finalized
<Kwiboo>
also hoping to get new revision of rkvdec high10 and hevc series out later tonight, need to re-run fluster tests with the code that is actually sent out / referenced