#asahi-dev · 2025-12-16 11:54
chadmedVRR will help with that on the desktop but that will need more work#asahi-dev · 2025-12-16 11:54
j`eymight want to be configurable then?#asahi-dev · 2025-12-16 11:54
chaos_princessit will#asahi-dev · 2025-12-16 11:54
j`eydoes 120mhz consume more power? #asahi-dev · 2025-12-16 11:53
chaos_princessoliver_: 24mhz is the "standard apple clock rate". and yes, macos reads timesync before every frame submit#asahi-dev · 2025-12-16 11:53
oliver_j`ey: never heard that before ;p#asahi-dev · 2025-12-16 11:53
j`ey(sorry, the joke had to be made :P)#asahi-dev · 2025-12-16 11:52
j`eyoliver_: nice patch, you're the best, man!#asahi-dev · 2025-12-16 11:52
oliver_chaos_princess: i am new to the kernel stuff, but the 24mhz fits neatly with the conversion factor of 41.67 that i need to use. would the timesync thing be useful ensure that the dcp clock does not drift from the kernels ktime_get_ns clock?#asahi-dev · 2025-12-16 11:47
chaos_princessthat would be constant#asahi-dev · 2025-12-16 11:47
lm_oliver_: is it a constant 120hz or variable?#asahi-dev · 2025-12-16 11:46
chaos_princessand, i would be interested in how the constants were derived#asahi-dev · 2025-12-16 11:45
chaos_princessoliver_: the only other issue i see is that you might want to look into something called "timesync", there is a separate mmio region on some coprocessors that contains the 24mhz clock from the coprocessor's perspective#asahi-dev · 2025-12-16 11:43
lm_Early christmas present#asahi-dev · 2025-12-16 11:43
lm_Oh my god #asahi-dev · 2025-12-16 11:43
chaos_princesshell yes#asahi-dev · 2025-12-16 11:43
oliver_you're now probably telling me that this is nice but there are other issues that I dont know about why this was disabled. bring it on!#asahi-dev · 2025-12-16 11:43
oliver_i got 120hz to work on my m1 pro: https://github.com/AsahiLinux/linux/pull/428#asahi-dev · 2025-12-15 09:14
nickchanwell that makes things simpler#asahi-dev · 2025-12-15 09:14
nickchanah so all the hvcs are just pv apple pac#asahi-dev · 2025-12-15 09:13
nickchanhm. that file is on opensource.apple.com's archive tarballs but not on github so i guess that was another slipup#asahi-dev · 2025-12-15 09:08
svennickchan: i don't expect gl1 to be available, my suggestion is to just virtualize everything in el1 with shadow page tables, etc.#asahi-dev · 2025-12-15 09:07
nickchanhttps://github.com/matteyeux/darwin-xnu/blob/master/doc/vmapple_pac.md as for pv pac this is actually documented#asahi-dev · 2025-12-15 09:04
nickchanignore the gic stuffs cause here the aic kext is used#asahi-dev · 2025-12-15 09:04
nickchanhttps://lore.kernel.org/qemu-devel/774802D2-E1B7-40B7-A189-82B4CE18B152@unpredictable.fr/ this is one of the HVC#asahi-dev · 2025-12-15 09:03
nickchanright not helpful#asahi-dev · 2025-12-15 09:03
chaos_princessbut 12 is not helpful for m4?#asahi-dev · 2025-12-15 09:02
nickchanmacos 12 vmapples would be the easiest to handle#asahi-dev · 2025-12-15 09:00
nickchanit tries to hypercall quite early i think#asahi-dev · 2025-12-15 08:59
chaos_princessinteresting. i sorta want to try it on one of my machines now and see how it goes#asahi-dev · 2025-12-15 08:59
nickchanthis is actually the same kext that has to be removed when using OSS XNU but here it's the vmapple kernel in KDK#asahi-dev · 2025-12-15 08:57
nickchanchaos_princess: omit com.apple.driver.SEPHibernation and the KC can be created#asahi-dev · 2025-12-15 08:50
chaos_princesswouldn't kmutil complain when trying to build that?#asahi-dev · 2025-12-15 08:50
chaos_princessits more that, well, how is this supposed to work#asahi-dev · 2025-12-15 08:50
nickchaneven on m4 normal kernelcache (!)#asahi-dev · 2025-12-15 08:50
nickchanAs for kexts SPTM is supposed to manage DART on hw but at this moment DART kexts are still available#asahi-dev · 2025-12-15 08:50
nickchanthe vmapple CPU is about as generic as what XNU kernel expect CPUs to be#asahi-dev · 2025-12-15 08:50
nickchancore xnu pretty only manages the CPU and practically all MMIO accesses are in kexts#asahi-dev · 2025-12-15 08:50
nickchanchaos_princess: the idea is to use vmapple's core xnu and combining with kexts that would normally be loaded on hardware.#asahi-dev · 2025-12-15 08:38
chaos_princessnickchan: so, mind elaborating on that? since i've heard you propose that before and i am a bit confused as to how that is supposed to work#asahi-dev · 2025-12-15 08:36
nickchannamely - mangling vmapple kernel and hardware kexts together#asahi-dev · 2025-12-15 08:31
nickchanthere may be methods that does not need sptm but i don't think m4 is any good to new reverse enginners#asahi-dev · 2025-12-15 08:30
nickchanif gxf isn't available to raw boot objects in el2 i don't think it would be available in el1 either#asahi-dev · 2025-12-15 04:14
lm_sakompella: M4 needs m1n1 to emulate SPTM properly before it can be reverse engineered so that would be the first step. see https://social.treehouse.systems/@sven/114278224116678776 and https://asahilinux.org/docs/sw/tethered-boot/#asahi-dev · 2025-12-15 04:09
b074with a draft PR I could also try to reproduce and help out with some progress. I'm also running on a t6030#asahi-dev · 2025-12-15 04:07
b074integralpilot - ive been reading the logs of the last few days, and I'm on the edge of my seat!#asahi-dev · 2025-12-14 21:25
sakompellaHello, I'm interested in contributing to the asahi project -- I have some systems programming experience (just finished a course on it, moderate experience with rust), but I have no experience with reverse-engineering. How can I get started? I have a M4 air machine, which I understand has minimal support.#asahi-dev · 2025-12-14 17:04
svenlet’s hope atcphy is picked up this cycle then 🤞#asahi-dev · 2025-12-14 16:44
jannauyes, downstream only. I don't think I'll submit rust rtkit this cycle