#
#asahi-dev · 2026-10-14 19:26
chaos_princessyou need to power up the pmgr before letting pmp touch stuff
#asahi-dev · 2026-10-14 19:26
jannauchaos_princess: have you tried reversing the power-domain dependencies? i.e. ps_disp_cpu uses pmp_report_disp power-domain? that would make the dtbs backwards compatibility easier
#asahi-dev · 2026-10-14 19:24
yukabut we should do it
#asahi-dev · 2026-10-14 19:23
yukaworked fine for me
#asahi-dev · 2026-10-14 19:23
yukadevices without dcp as in they don't have the drivers enabled yet?
#asahi-dev · 2026-10-14 19:23
yukaoeehhhh
#asahi-dev · 2026-10-14 19:23
jannaudon't we need to change framebuffer0's power-domains as well? certainly on devices without dcp
#asahi-dev · 2026-10-14 19:22
jannauyuka: 1 minor conflict in t6031-die0.dtsi, testing now
#asahi-dev · 2026-10-14 18:26
yukaOr I guess close it, since at that point it would go directly against upstream?
#asahi-dev · 2026-10-14 18:24
yukait's multiple commits, I believe some of them have conflicts. If there are no plans to add PMP to the M3 device trees for now, I'll not rebase the PR but make it a draft.
#asahi-dev · 2026-10-14 18:07
jannaushould be mostly fine with pmp-report but power-domain changes might be very disruptive w.r.t. dtb vs. kernel mismatches
#asahi-dev · 2026-10-14 18:06
jannaumain reason why pmp is not enabled are the unreviewed dt-bindings
#asahi-dev · 2026-10-14 18:05
jannau not sure if it needs to be rebased. `git fetch asahi pull/$NUM/head` and `git cherry-pick FETCH_HEAD` works with rebased base branches
#asahi-dev · 2026-10-14 18:00
yukaI'm running it on my t6034 since months at this point, and it's still disabled by default
#asahi-dev · 2026-10-14 18:00
yukajannau: could you have a look at M3 PMP if I rebased it?
#asahi-dev · 2026-10-14 17:26
jannauhandled a few of the simple PRs
#asahi-dev · 2026-10-14 16:52
tryhard[m]jannau: awesome
#asahi-dev · 2026-10-14 16:38
jannauI'm looking forward to v7.3 next week. the rebasing from v7.2.2 to v7.2.10 was a bit annoying
#asahi-dev · 2026-10-14 16:36
jannauthe idle=nop and WFxT override changes are included so M4 and later should work in linux
#asahi-dev · 2026-10-14 16:30
svennice work!!
#asahi-dev · 2026-10-14 16:28
yukathank you!
#asahi-dev · 2026-10-14 16:28
jannauasahi-wip is finally based on v7.2.10, only lightly tested here. I haven't yet looked at the PRs arrived in the last weeks
#asahi-dev · 2026-10-14 14:51
Alex0[m]Also idk why pwm-leds is used inside the dts and leds-pwm (ik about the prefix based convention) xD
#asahi-dev · 2026-10-14 14:51
Alex0[m]Thxx
#asahi-dev · 2026-10-14 14:22
chaos_princessthen just use pwm-leds (an existing generic driver) to expose the pwm controller as a led
#asahi-dev · 2026-10-14 14:22
chaos_princessyou do not need to look at the led system at all, what you are interested is the pwm one
#asahi-dev · 2026-10-14 14:19
Alex0[m]chadmed: Should I look at other led implantations or just the led drivers or just look at <linux/leds.h>
#asahi-dev · 2026-10-14 12:19
MichaelReeves[m]I'll try and rewrite the frontend around WS now (probably going to bed soon, will continue tomorrow). looking at it I designed it pretty de-coupled so I don't think it should be too hard
#asahi-dev · 2026-10-14 12:18
MichaelReeves[m]ah ok then. sounds good.
#asahi-dev · 2026-10-14 12:16
chadmedwhich youre going to need regardless
#asahi-dev · 2026-10-14 12:16
chadmedit seems like its only major direct dependencies are fastapi, pydantic and websockets
#asahi-dev · 2026-10-14 12:15
chadmedwe're gonna have to pull a bunch of nonsense in anyway
#asahi-dev · 2026-10-14 12:13
MichaelReeves[m]s/it/in
#asahi-dev · 2026-10-14 12:12
MichaelReeves[m]ehhh im not sure about that it looks like its pulling it a bunch of native deps, would that complicate packaging?
#asahi-dev · 2026-10-14 11:41
JamesCalligeros[m]https://pypi.org/project/fastapi-ws-rpc/
#asahi-dev · 2026-10-14 11:40
MichaelReeves[m]oh that sounds good, thank you! I think web sockets sound good then
#asahi-dev · 2026-10-14 11:39
chadmedyou just throw json messages in one end and they come out the other and you can do whatever you want with them
#asahi-dev · 2026-10-14 11:38
chadmedfastapi comes with one
#asahi-dev · 2026-10-14 11:38
MichaelReeves[m]how big a dep would a RPC library be? I was hesitant because I thought it would massive but maybe it's manageable
#asahi-dev · 2026-10-14 11:38
chaos_princesswebsocket provides message framing for you, so the message can be a json blob or sth if you do not want to use some pre-existing thing
#asahi-dev · 2026-10-14 11:37
MichaelReeves[m]ah, true
#asahi-dev · 2026-10-14 11:37
chaos_princessand you are designing a message protocol when using rest, even if you do not see it
#asahi-dev · 2026-10-14 11:36
Faelar[m]that's where the rpc part is used no?
#asahi-dev · 2026-10-14 11:36
MichaelReeves[m]REST is just the existing really simple ones in the plan GET /api/info, POST /api/install, GET /api/telemetry, POST /api/shutdown
#asahi-dev · 2026-10-14 11:36
chaos_princessas i have mentioned, insert-rpc-solution-here
#asahi-dev · 2026-10-14 11:36
MichaelReeves[m]yeah but then you still need to decide the message protocol and everything I think
#asahi-dev · 2026-10-14 11:35
chaos_princessit is just a pipe you can send messages into and they come on the other side
#asahi-dev · 2026-10-14 11:35
chaos_princesswebsockets are less than doing full rest and sse?
#asahi-dev · 2026-10-14 11:34
jannau /kick Faelar[m]
#asahi-dev · 2026-10-14 11:34
Faelar[m]learn from the best, ask claude to fix your finnish penguin something OS 😎

plus ›