#
#asahi-dev · 2026-04-08 19:43
chaos_princessnot sure what you mean
#asahi-dev · 2026-04-08 19:42
tryhard[m]chaos_princess: it is possible to make linux use notify_pmp right?
#asahi-dev · 2026-04-08 19:41
tryhard[m]there is still these 0x204 messages, which don't appear to do anything other than appear after specific DEV devpwr messages. ill check the fabric power state, see if it does something there.
#asahi-dev · 2026-04-08 19:39
chaos_princesstime for a linux driver then
#asahi-dev · 2026-04-08 19:39
tryhard[m]chaos_princess: yeah
#asahi-dev · 2026-04-08 19:39
tryhard[m]chaos_princess: yeah i figured, but i thought there is something i forgot to setup because nothing actually changed from my view
#asahi-dev · 2026-04-08 19:39
chaos_princesscan you turn it on without macos assistance?
#asahi-dev · 2026-04-08 19:38
tryhard[m]ok. well other than that, it doesn't appear there is something else to do with pmp.py. sending messages are working as long as pmgr actually turned it on
#asahi-dev · 2026-04-08 19:38
chaos_princessyou kinda shouldn't care what exactly pmp is doing (i assume it is doing quite a few things), just make sure it is on and it gets the info it wants
#asahi-dev · 2026-04-08 19:36
chaos_princessyou don't. read the power usage with smc
#asahi-dev · 2026-04-08 19:36
tryhard[m]how can i check the fabric power state?
#asahi-dev · 2026-04-08 19:33
chaos_princesson v2 it has communication links to gpu and dcps and it can control fabric power state in response to gpu/dcp load
#asahi-dev · 2026-04-08 19:32
tryhard[m]if it doesn't actually cut off the power, then what does it actually do with that info?
#asahi-dev · 2026-04-08 19:29
chaos_princesspmgr controls the power, but pmp needs to be told that you are about to turn the power off, or it will crash. in case of pmp v2, pmp can actually shut down dcps and fabric links
#asahi-dev · 2026-04-08 19:26
tryhard[m]is it the pmgr that actually controls the power? because whenever something happening, like closing the lid and the disp0 turns off, there is a message to pmp but pmgr is the one i think that actually powers off. it happens with every soc.
#asahi-dev · 2026-04-07 15:58
tryhard[m]btw p.pmgr_adt_power_disable("/arm-io/disp0") doesnt work in macos.
#asahi-dev · 2026-04-07 15:58
tryhard[m]shit xD
#asahi-dev · 2026-04-07 15:56
chaos_princessidk
#asahi-dev · 2026-04-07 15:51
tryhard[m]so whats up with that? you have any idea why pmp devpwr doesn't work?
#asahi-dev · 2026-04-07 15:50
tryhard[m]oh yeah it works
#asahi-dev · 2026-04-07 15:49
chaos_princesshttps://github.com/AsahiLinux/m1n1/blob/main/proxyclient/m1n1/proxy.py#L1030
#asahi-dev · 2026-04-07 15:47
tryhard[m]doesn't seem like pmgr_adt_power_disable_index is exposed in python
#asahi-dev · 2026-04-07 15:40
chaos_princessabout right, yes. (can use p directly)
#asahi-dev · 2026-04-07 15:39
tryhard[m]this is correct right: pmp.p.pmgr_adt_power_disable("/arm-io/dcp")
#asahi-dev · 2026-04-07 15:38
chaos_princessmaybe try turning things off by index, not by path then?
#asahi-dev · 2026-04-07 15:38
tryhard[m]nope
#asahi-dev · 2026-04-07 15:37
chaos_princesstry also /arm-io/dcp
#asahi-dev · 2026-04-07 15:36
tryhard[m]/arm-io/disp0
#asahi-dev · 2026-04-07 15:35
chaos_princessok, then that bug is probably fixed. what dev path?
#asahi-dev · 2026-04-07 15:35
tryhard[m]yeah i guess
#asahi-dev · 2026-04-07 15:35
chaos_princessfrom 2025 or 2026
#asahi-dev · 2026-04-07 15:34
tryhard[m]wdym new
#asahi-dev · 2026-04-07 15:34
chaos_princessis you stage 1 m1n1 new?
#asahi-dev · 2026-04-07 15:32
tryhard[m]even disabling through pmgr it doesnt turn off, assuming i got the dev path correctly.
#asahi-dev · 2026-04-07 15:31
tryhard[m]macos does turn it off successfully with a devpwr msg
#asahi-dev · 2026-04-07 15:30
tryhard[m]yet, disp0 doesnt turn off when sending a message. turned on all the stuff it uses before turning off the display.
#asahi-dev · 2026-04-07 15:29
tryhard[m]thx
#asahi-dev · 2026-04-07 15:29
tryhard[m]noopwafel: yeah i turned it on with pmgr and it was fixed.
#asahi-dev · 2026-04-07 15:04
noopwafel(and that address does seem to be in ane[0], so my first thought would also be to make sure pmgr turned it on..)
#asahi-dev · 2026-04-07 14:53
noopwafelyou can ignore the "wrong PA; not mapped", sorry, that's just noting that it isn't mapped on the host side (to be expected)
#asahi-dev · 2026-04-07 13:21
tryhard[m]in addition, turning off the display still doesn't work. might be i need to turn on something first...
#asahi-dev · 2026-04-07 13:19
tryhard[m]if i dont turn it on, then no crash.
#asahi-dev · 2026-04-07 13:19
tryhard[m]ok it seems like its actually related to turning on ANE. specifcally self.send(PMP_DevPwr(DEV=0x68, STATE=1)).
#asahi-dev · 2026-04-07 13:15
tryhard[m]chaos_princess: don't think its related then.
#asahi-dev · 2026-04-07 12:57
jannaus/clocks/power/ in main
#asahi-dev · 2026-04-07 12:56
chaos_princesslook in other tracers for examples of how they use pmgr_adt_clocks_enable
#asahi-dev · 2026-04-07 12:55
tryhard[m]you happen to know the command for that?
#asahi-dev · 2026-04-07 12:53
chaos_princessturn on the devices with pmgr before sending pmp commands to touch them, pmp is most likely getting a serror when it touches a powered down device
#asahi-dev · 2026-04-07 12:52
tryhard[m]chaos_princess: https://pastebin.com/HDhTUM3n
#asahi-dev · 2026-04-07 12:49
chaos_princesstryhard[m]: post your code somewhere

plus ›