#
#asahi-dev · 2024-07-11 00:16
aliceyeah the repo layout is.. interesting
#asahi-dev · 2024-07-10 21:11
ella-0Turns out this is going to be a pain to push upstream. While all the makefiles are identical that I want to modify, they are split across about 15 different repositories.
#asahi-dev · 2024-07-10 11:08
chadmedlegend
#asahi-dev · 2024-07-10 11:08
ella-0I'll try pushing the X stuff upstream after work today
#asahi-dev · 2024-07-10 11:08
chadmedneat!
#asahi-dev · 2024-07-10 11:07
ella-0yup
#asahi-dev · 2024-07-10 11:07
chadmedis this for iglunix bringup?
#asahi-dev · 2024-07-10 11:07
ella-0Ah. Yeah that'd be it. I don't have udev. I use libudev-zero
#asahi-dev · 2024-07-10 11:07
chadmedthe udev rules are in speakersafetyd and i cant remember if the makefile installs them with the install target or not
#asahi-dev · 2024-07-10 11:07
chadmedhrmmm do you have the udev rules installed? thats a classic symptom of ACP_IGNORE not being set on the visense pcm
#asahi-dev · 2024-07-10 11:06
chadmedella-0: i wonder if that would be welcome upstream. id like to ditch more X deps
#asahi-dev · 2024-07-10 11:06
ella-0I am having some other fun issues though. Like speakers aren't detected unless I start speakersafetyd after launching pipewire/wireplumber not before. And I have to clear old wireplumber state every time I restart it otherwis I get really fun routing issues.
#asahi-dev · 2024-07-10 11:05
ella-0chadmed: sorry typo
#asahi-dev · 2024-07-10 11:04
ella-0chamed: I had to patch all the makefiles to get it to build without X. But it does still work fine.
#asahi-dev · 2024-07-09 08:22
sam_i've bumped it a few times in the past and it wasn't fun
#asahi-dev · 2024-07-09 08:22
sam_lv2-plugins is grim tho
#asahi-dev · 2024-07-09 08:22
sam_lots of stuff still needs to be USE=gui but we're getting there..
#asahi-dev · 2024-07-09 07:59
leio(and others, less needed for asahi)
#asahi-dev · 2024-07-09 07:58
leioyeah, still X deps in USE=lv2
#asahi-dev · 2024-07-08 21:54
chadmed_ugh that just disables xdg stuff...
#asahi-dev · 2024-07-08 21:53
chadmed_ella-0: i build it with USE="-X" on gentoo and it _seems_ to do what it's supposed to
#asahi-dev · 2024-07-08 19:18
ella-0Has anyone perchance managed to build lsp-plugins without X11? What I've found so far is it seems it used to be possible a while ago but then the build system got refactored and I don't even know where to start with the mess of makefiles
#asahi-dev · 2024-07-08 10:38
chadmedand this machine has no fans!
#asahi-dev · 2024-07-08 10:38
chadmedbut they dont do anything even when you light up the whole cluster
#asahi-dev · 2024-07-08 10:37
chadmedmy thought when i first noticed was maybe the cores really are just that efficient...
#asahi-dev · 2024-07-08 10:36
jannauI think the "ftP*" and "ftG*" keys might be a better target than "Tp0*" and "Tg0*" but still no sensors which respond clearly to e-core load
#asahi-dev · 2024-07-08 10:32
jannauI just add all the sensors to the devicetree like https://paste.centos.org/view/60b9d88b and then run something like `watch -n 1 "sensors macsmc_hwmon-isa-0000 | grep -e '^T'" -e '^ft'"`
#asahi-dev · 2024-07-08 10:27
chadmedwe still need to get all the T*** ones and see what they do when we burn the cpu
#asahi-dev · 2024-07-08 10:26
chadmedoh yeah we have #KEY
#asahi-dev · 2024-07-08 10:24
jannauchadmed: we can enumerate the smc keys: experiments/smc.py
#asahi-dev · 2024-07-08 09:41
chadmedbut all the cpu-adjacent ones seem to have stupid names so we cant even really intuit which they might be
#asahi-dev · 2024-07-08 09:41
chadmedsome of the smc keys are sensiple like TB0T (battery 0 temp?) and PSTR (system total real?)
#asahi-dev · 2024-07-08 09:40
chadmedthe issue is more that we dont even know what sensors exist beyond ones we found aaaages ago :p
#asahi-dev · 2024-07-08 09:39
chadmedyeah ive been burning the cores with stress-ng doing matrix workers
#asahi-dev · 2024-07-08 09:39
chaos_princessyou could probably figure out per-core ones by getting the core to multiply matrices and looking at which one spikes
#asahi-dev · 2024-07-08 09:37
chadmedat least then we can have a list of any temp sensors that actually exist, then test for what they are (if there arent a stupid amount of them)
#asahi-dev · 2024-07-08 09:37
chadmedto avoid having to look at macos again my current strat is to just write a brute forcer for m1n1 that goes through every T[A-Za-z0-9][A-Za-z0-9][A-Za-z0-9] and tries to read it
#asahi-dev · 2024-07-08 09:32
chadmedmaybe something we might be able to fix in #asahi-re but nothing easily accessed/open
#asahi-dev · 2024-07-08 09:31
chaos_princessand i take it there is no convenient plist in macos that fixes this mess?
#asahi-dev · 2024-07-08 09:31
chadmedand the sensors that some foss macos tools _think_ are the ecore temp sensors dont actually seem to be the ecore temp sensors
#asahi-dev · 2024-07-08 09:30
chadmedthere doesnt seem to be any rhyme or reason between machines even with the same soc
#asahi-dev · 2024-07-08 09:30
chaos_princesswhats up with hwmon sensors?
#asahi-dev · 2024-07-08 09:30
chadmed(just for temperature)
#asahi-dev · 2024-07-08 09:30
chadmedfantastic
#asahi-dev · 2024-07-08 09:30
chadmedso whats that like 250k "possible" sensors
#asahi-dev · 2024-07-08 09:29
chadmedwait no ffs its not 36 its 62 because they can be A-Z a-z 0-9
#asahi-dev · 2024-07-08 09:26
chadmedP(36,3) is like 40,000 possible sensors
#asahi-dev · 2024-07-08 09:22
chadmedjannau: at this stage i cant see a solution to the hwmon sensor discrepancies other than checking every single machine semi-manually
#asahi-dev · 2024-07-07 10:03
jannaushould work with 12.3 (12.4) and 13.5 depending on $AGX_FW. experiments/dcp.py might be even able to determine that by itself, otherwise you can set it as env var
#asahi-dev · 2024-07-07 09:43
chaos_princess*not tracer, the dcp.py thing in experiments

plus ›