<apritzel>
that really mostly for power off (no regulator setup for those two new PMICs), but it now detects the right PMIC at runtime, and also supports I2C
<apritzel>
would be glad if someone could give this a test
<apritzel>
acmeplus, tokyovigilante, macromorgan: ^^^^ also for your AXP717 devices
<acmeplus>
apritzel: been busy these days, but I went with the latest mainline and added your AXP717 patch as well as the cpufreq one. Still need to look at the AXP717 spec to see if I win the prize :)
<acmeplus>
I'll update the TF-A with those changes too
<acmeplus>
For the kernel I'm applying these two patches: v2_20240318_andre_przywara_cpufreq_sun50i_add_allwinner_h616_support, and 20240310_andre_przywara_regulator_add_x_powers_axp717_pmic_support
<macromorgan>
sweet, I was going to do that then got sidetracked writing another PMIC driver... I guess when I'm done with this one I can look at the missing bits for the AXP717
BroderTuck has joined #linux-sunxi
<BroderTuck>
s/one at least/on at least/ on the "refactor" patch in the tf-a link above
BroderTuck has quit []
warpme has quit []
warpme has joined #linux-sunxi
JohnDoe_71Rus has joined #linux-sunxi
<apritzel>
thanks BroderTuck, fixed locally
gsz has joined #linux-sunxi
Schimsalabim has quit [Ping timeout: 480 seconds]
Schimsalabim has joined #linux-sunxi
digetx has quit [Read error: Connection reset by peer]
Schimsalabim has quit [Read error: Connection reset by peer]
Schimsalabim has joined #linux-sunxi
<Jookia>
after staring at the alsa soc subsystem i think the sunxi i2s driver might be doing something .... unexpected
<Jookia>
hmm. i'm not sure. we'll see
<Jookia>
it's super easy to make linux do something by editing the source code! but super hard to add functionality and preserve existing behaviour for mainline :(
<Jookia>
there's a lot of behaviour that is nice but probably really should be explicit and moved to the sound card layer
<Jookia>
(moved as in for my use case)
<apritzel>
Jookia: "super easy to make linux do something..."> Oh yeah, that's a common misconception: getting something to run is one thing, getting it in an upstreamable shape can be a quite different story
Schimsalabim has quit [Read error: Connection reset by peer]
<Jookia>
basically: different sound DAIs handle 'automatic' TDM mode differently
<Jookia>
on sunxi i2s if you ask for channels > 2 then you get automatic TDM
<Jookia>
this is a cool feature as it means at runtime you can use a slower clock for less channels
<Jookia>
unfortunately this is also kind of unpredictable and error prone, i think IDEALLY the sound card should be calling set_dai_tdm_slot
Schimsalabim has joined #linux-sunxi
<Jookia>
so i think i've kind of untangled that out of my blob of weird codec stuff: adding variable TDM support to audio-graph-card2 or something