#asahi-dev · 2026-09-17 06:59
nomad_of_tatooine[m]Hi everyone,... (full message at <https://matrix.org/oftc/media/v1/media/download/AXDtYjBvsdu0Uhq4QAyMpu9cyj7iM0hqC79JPBEFO6LD2SSJ6gnoIDS9J0zSB7lSUNgDVVzqvIOrIJOyWV4-sytCeg9kJ4OAAG1hdHJpeC5vcmcva0lNRmJwcHRxVnJSYXRNYnpFeXhac25D>)#asahi-dev · 2026-09-16 19:47
mmediouni[m]M1 is effectively the A14X, so yes#asahi-dev · 2026-09-16 19:46
chaos_princessisn't "run on A14 linux" the ~same as "run on m1 linux"?#asahi-dev · 2026-09-16 19:45
myriaah, so close. And yeah, if that's the case, then it's really a question if we want things to run on A14 linux before an A14 can even run linux#asahi-dev · 2026-09-16 18:11
mmediouni[m]the disclosed securerom bug (in dubious circumstances, there's a lawsuit by a private exploit firm over it) works all the way to A13#asahi-dev · 2026-09-16 18:07
mmediouni[m]myria: A14 and M1 are the same core. A15 and M2 are also the same cores#asahi-dev · 2026-09-16 18:04
myria*because everything I'm looking at tells me that A14/15 will share the M1/2 PMC, and thus should be fine to use the current driver for, and in writing bindings for libpfm4, wasn't sure if I should include their MIDR values#asahi-dev · 2026-09-16 18:03
myriaAsking in the context of pmu/pmc, becayse #asahi-dev · 2026-09-16 18:01
myrianickchan: that reminds me, has there been any work on running linux on A14/15, I think I heard about A11/12 with the usb vuln, but don't know about later.#asahi-dev · 2026-09-16 16:23
mmediouni[m]and don't even have the same decoder width apparently...#asahi-dev · 2026-09-16 16:23
mmediouni[m]M6 and A20 Pro super cores are not the same#asahi-dev · 2026-09-16 16:11
nickchanA20 Pro does have "Super Cores" in marketing now#asahi-dev · 2026-09-16 16:11
nickchananother thing that i have missed#asahi-dev · 2026-09-16 14:46
JPeisachactually i'll figure it out.. dittoing to #asahi#asahi-dev · 2026-09-16 14:39
JPeisachunder hypervisor? :/#asahi-dev · 2026-09-16 14:33
chaos_princessproxyclient/tools/smccli.py#asahi-dev · 2026-09-16 14:33
JPeisachreal dev question: is there a tool to poke at (read) SMC key values at a given time, in hopes of figuring out what it means#asahi-dev · 2026-09-16 12:31
JPeisachif i can throw in my two cents without going into offtopic realm: sure, you can be efficient. but it's about caring. for example, when burying someone (just go with me for a second), some cultures prefer to have a human do it. it's more preferred than having a machine do it for efficiency. Same idea can apply to projects. You can use a LLM, but do you really care about its output?#asahi-dev · 2026-09-16 12:30
mmediouni[m]because LLM usage could have resulted in it anyway...#asahi-dev · 2026-09-16 12:30
mmediouni[m]well, there were things that'd have been interesting projects before, now people just go meh if you do it#asahi-dev · 2026-09-16 12:22
chadmedi wouldnt trust any proclamation of mobile soc support that didnt come from you at this point lmao #asahi-dev · 2026-09-16 12:20
NickChan[m]If you come across some linux supporting A5/A6 it's probably slop 🫠 #asahi-dev · 2026-09-16 12:20
NickChan[m]so since the topic is slop#asahi-dev · 2026-09-16 11:51
chadmedas much as we absolutely 1000% cannot and do not want to get sued by apple, im sure the linux foundation would also like to avoid it #asahi-dev · 2026-09-16 11:51
chadmedand as much as the agent puppeteer has been claiming to the contrary, he is rather untrustworthy when it comes to all of this and has already even disregarded his own rather lax LLM policy#asahi-dev · 2026-09-16 11:50
neobrain[m]ah, that's fair lol#asahi-dev · 2026-09-16 11:49
chadmedyou cant "make no mistakes" your way around that#asahi-dev · 2026-09-16 11:49
chadmedwell the code still needs to be good#asahi-dev · 2026-09-16 11:48
neobrain[m]I wonder what would happen if they tried upstreaming their kernel bits though... assuming they would bother upstreaming, anyway#asahi-dev · 2026-09-16 11:44
chadmedit's a pile of slop and can be ignored#asahi-dev · 2026-09-16 11:43
chadmedoh yeah lol#asahi-dev · 2026-09-16 11:42
neobrain[m]chadmed: vibe-coded m4 gpu driver 🙈#asahi-dev · 2026-09-16 10:18
chadmedneobrain[m]: did i miss something specific today#asahi-dev · 2026-09-16 09:22
nickchannot enough to explode nvme cause adminq io buffer is only used for things like the smart data but still#asahi-dev · 2026-09-16 09:08
nickchansent a reply hopefully stable won't be broken#asahi-dev · 2026-09-16 08:47
nickchans/prp check/drop prp check/#asahi-dev · 2026-09-16 08:47
nickchanso i guess i have to care about the stable backports a bit#asahi-dev · 2026-09-16 08:47
nickchanwithout the page size quirk#asahi-dev · 2026-09-16 08:46
nickchanah right, the the adminq would stop working properly if the prp check is applied in isolation in stable#asahi-dev · 2026-09-16 08:28
neobrain[m]Ugh, sorry for today's nonsense that people are having to deal with :(#asahi-dev · 2026-09-16 08:06
chadmedthis Works For Me(tm)? #asahi-dev · 2026-09-16 08:06
chadmedonly thing i can think of is making the shmem region bigger but like#asahi-dev · 2026-09-16 08:06
yukaI'm switching back to 7.1 for now since I need to get some work done, but if you have something I should try I can test it / trace it later / tomorrow#asahi-dev · 2026-09-16 08:04
chadmedthe last message we get is setshmem #asahi-dev · 2026-09-16 08:04
chadmedand the init sequence does not start because we do not get any epic messages in the trace#asahi-dev · 2026-09-16 08:03
chadmedidk how/why that would fail but its basically the last thing that can fail before the init sequence starts#asahi-dev · 2026-09-16 08:03
chadmedare we setting the shmem wrong for m3 maybe #asahi-dev · 2026-09-16 08:00
yukaMaybe another parameter I forgot?#asahi-dev · 2026-09-16 07:58
chadmedhow fun #asahi-dev · 2026-09-16 07:58
chadmedi have like a gigabyte of traces from macos i got with that :P