#asahi-dev · 2026-10-05 10:46
integralpiloti also supplied UEFI firmware#asahi-dev · 2026-10-05 10:46
integralpilotmyria: i found my command and realised what was different! (also dance classes sound cool! :D)#asahi-dev · 2026-10-05 10:42
fl0_idchaos_princess I sometimes imagine that there's like a specific ai trainer somewhere in nigeria or so, that managed to impose their specific style on the ai models...#asahi-dev · 2026-10-05 10:35
Hel3ziYes, m4 pro#asahi-dev · 2026-10-05 10:35
Hel3ziHangs on the Asahi logo (from m1n1, not Apple logo). Using linux.py (with -b "earlycon console=tty0 loglevel=7"). Regarding pcie_init: before I stubbed it out, m1n1 hung right at pcie: No common tunab... and stopped responding there.#asahi-dev · 2026-10-05 10:34
yukaThis is M4 Pro right?#asahi-dev · 2026-10-05 10:33
yukagiven that you don't have serial, probably run_guest is the best option, and then you can connect to ttyACM1 to see the kernel output#asahi-dev · 2026-10-05 10:33
yukahangs on which logo, and are you doing run_guest.py or linux.py??#asahi-dev · 2026-10-05 10:32
yukapcie_init should not cause problems...#asahi-dev · 2026-10-05 10:32
yukaHuh interesting#asahi-dev · 2026-10-05 10:31
Hel3ziyuka: stubbed out dapf and pcie_init in m1n1. Now m1n1 cleanly reaches Preparing to boot kernel at 0x1... and jumps. After that, USB proxy disconnects as expected and the kernel hangs on the logo#asahi-dev · 2026-10-05 10:05
jannauin 70 minutes#asahi-dev · 2026-10-05 10:05
jannauwill be streamed in https://www.youtube.com/watch?v=ksq7uqsg2kM#asahi-dev · 2026-10-05 10:03
JamesCalligeros[m]Oh it starts soon lets go#asahi-dev · 2026-10-05 10:01
JamesCalligeros[m]Sad that this is only a 15m talk#asahi-dev · 2026-10-05 10:01
chaos_princessright, aarch32 at el0 still exists#asahi-dev · 2026-10-05 10:01
nickchanthat kind of system is rare though I only know of two kinds of cpu that's like that and both of them are apple#asahi-dev · 2026-10-05 10:00
nickchanand make sure the native page size is used for aarch32 if 4k is not implemented yet aarch32 is implemented#asahi-dev · 2026-10-05 09:59
nickchanif available#asahi-dev · 2026-10-05 09:59
nickchanerm, is available#asahi-dev · 2026-10-05 09:59
chaos_princesswe can steal it for ourselves even if it is not generic#asahi-dev · 2026-10-05 09:59
nickchanso like handle 64k page systems and forcing 4k on aarch32 if possible#asahi-dev · 2026-10-05 09:59
JamesCalligeros[m]Yeah#asahi-dev · 2026-10-05 09:59
nickchanfor the thing to be upstreamable it would need to be completely generic#asahi-dev · 2026-10-05 09:58
jannauplease move the gerrit discussion to #asahi-offtopic#asahi-dev · 2026-10-05 09:58
JamesCalligeros[m]Gerrit > gitlab? Consider this an official warning :p #asahi-dev · 2026-10-05 09:58
chaos_princessMichaelReeves[m]: lol, wat#asahi-dev · 2026-10-05 09:58
MichaelReeves[m]certainly better than gitlab and MLs imo#asahi-dev · 2026-10-05 09:57
JamesCalligeros[m]This is really cool#asahi-dev · 2026-10-05 09:57
chaos_princessi may even forgive a few of google's crimes for that#asahi-dev · 2026-10-05 09:57
MichaelReeves[m]hey, I kinda like gerrit. have to use it for chromium & mediawiki#asahi-dev · 2026-10-05 09:57
JamesCalligeros[m]Smaller patchset than i expected though#asahi-dev · 2026-10-05 09:57
chaos_princessholy shit, this is awesome.#asahi-dev · 2026-10-05 09:57
JamesCalligeros[m]Are they just continuing to use gerrit to spite everyone who thinks it sucks (everyone)#asahi-dev · 2026-10-05 09:56
Hel3ziyuka: Quick update: I flashed the patched m1n1 directly to the SSD via kmutil configure-boot in 1TR to bypass chainload. Disabling DAPF worked - no more SError, and the Mac doesn't reboot anymore! It now progresses further and hangs at pcie: Initializing t6040 PCIe controller (right around No common tunables).#asahi-dev · 2026-10-05 09:54
jannaupatches posted unfortunatly means for the android 6.18 kernel. https://android-review.googlesource.com/q/topic:p3s_v1#asahi-dev · 2026-10-05 09:53
JamesCalligeros[m]The holy grail doctor jones#asahi-dev · 2026-10-05 09:53
JamesCalligeros[m]omg#asahi-dev · 2026-10-05 09:53
jannauit implements per process page sizes, i.e. allowing 4k processes on a 16k kernel#asahi-dev · 2026-10-05 09:51
jannaunice: "Compatibility for 4KB Applications on Large Page-Sized Systems" https://lpc.events/event/20/contributions/2630/ hidden in the Android MC#asahi-dev · 2026-10-05 08:43
myriaMichaelReeves[m]: it's okay, I had a dance class :3#asahi-dev · 2026-10-05 08:38
yukanice!#asahi-dev · 2026-10-05 08:35
jannauyuka: changes due to upstream rust changes (7.2 and part of drm-rust for 7.3). hopefully the only remaining part is gpuvm. lacked motivation to work on it in the last week but working now#asahi-dev · 2026-10-05 08:34
JamesCalligeros[m]Lmaoooo#asahi-dev · 2026-10-05 08:30
chaos_princessthink of the robot as an utter failure, a fuck-up that is constantly being told that you got it wrong and should re-do it, because you did in fact get it wrong. ofc you will start insisting that yes, the thing you are talking about is real, just to re-gain some self-confidence that the scawwy humans are constantly shattering.#asahi-dev · 2026-10-05 08:22
JamesCalligeros[m]Why are the robots so eager to assert that everything is “real”#asahi-dev · 2026-10-05 08:14
jannausorry, I did only read the first sentence#asahi-dev · 2026-10-05 08:08
svenand ugh… that fucking description :(#asahi-dev · 2026-10-05 08:07
svenNice, do maybe PCIe hotplug will just work now #asahi-dev · 2026-10-05 07:54
yukabtw, jannau, can I possibly help with the 7.2 rebase? was the gpu driver just not compiling or is there some run-time foo which needs to be adapted?