#asahi-dev · 2026-07-25 21:32
yukasplitting up the offsets of nvme and nvmmu into two regs is an M4+ thing#asahi-dev · 2026-07-25 21:31
yukaskipping APPLE_ANS_LINEAR_SQ, APPLE_ANS_UNKNOWN_CTRL & nvme_set_queue_count is FW 26.x thing (needed on M3 with 26.x too)#asahi-dev · 2026-07-25 21:29
yukathe hacks are here: https://cyberchaos.dev/yuka/linux/-/commit/4d9e35e5d9b01a4b711d654e252d4a9e902e7fcc#asahi-dev · 2026-07-25 21:19
yukaand nvme_set_queue_count is a command via the admin queue#asahi-dev · 2026-07-25 21:19
yukathe trace I have covers only the mmio address space and no the queues#asahi-dev · 2026-07-25 21:14
yukaOn A18 Pro too#asahi-dev · 2026-07-25 21:09
yukaseems to work though#asahi-dev · 2026-07-25 21:09
yukaI do not like that this involves commenting out "nvme_set_queue_count"#asahi-dev · 2026-07-25 21:08
yukaand M4 too#asahi-dev · 2026-07-25 21:08
yukaHmm, I was able to get M3 nvme working with 25.5 fw both in m1n1 and linux#asahi-dev · 2026-07-25 21:03
sofus(all os fw are reported as 10151.x)#asahi-dev · 2026-07-25 21:01
sofusyes its also new, 18000.x#asahi-dev · 2026-07-25 21:01
yukain the boot log#asahi-dev · 2026-07-25 21:00
yukam1n1 should tell you the system fw?#asahi-dev · 2026-07-25 21:00
sofusi did restore#asahi-dev · 2026-07-25 21:00
rhlvista[m]If you do revive#asahi-dev · 2026-07-25 20:59
sofusshould a dfu restore also change the system fw? the boot picker sometimes has the new symbols#asahi-dev · 2026-07-25 20:53
jannaucharging_end is broken in asahi-7.1.5-1 (I think the upstream driver dropped it). there will be asahi-7.1.5-2 in any case for the t600x/t8112 hevc fix. I'll look at it tomorrow#asahi-dev · 2026-07-25 20:51
sofusiirc i tested a sample where videotoolbox ~60fps and linux+gstreamer was ~30, but i also need to do a propper test#asahi-dev · 2026-07-25 20:48
jannaudecoding speed for a long high resolution sample looked okay but I did not compare it to videotoolbox#asahi-dev · 2026-07-25 20:46
sofushttps://aomedia.org/av1-video-decoder-verification-tool/#asahi-dev · 2026-07-25 20:46
sofusthe av1 tests are actually super cool, i have used the "standart" ones up until now#asahi-dev · 2026-07-25 20:43
sofusi plan to take a lot at stability/speed once av1 is good enough#asahi-dev · 2026-07-25 20:41
sofusthe adt describes some clocks, i gotta look into that at some point#asahi-dev · 2026-07-25 20:41
sofusvp9 is +15 sec for me, kinda weird its is less for you. #asahi-dev · 2026-07-25 20:36
jannausofus: t6001-vp9 takes 317s but feels faster because it's twice as many test cases#asahi-dev · 2026-07-25 20:33
svenand because it's just PPL and not the entire SPTM/TXM mess#asahi-dev · 2026-07-25 20:32
sveni emulated it on a m1 because it's easy to compare what should happen with the real HW vs. what my emulation does and because i don't have a m4 yet ;)#asahi-dev · 2026-07-25 20:31
svenanyway, now it's really enough for today :D#asahi-dev · 2026-07-25 20:31
svenwhich probably explains why booting takes like 10 minutes#asahi-dev · 2026-07-25 20:31
svenxnu's ttbr0 and ttbr1 get a guarded and a non-guarded shadow pagetable each which right now is fully rewritten whenever anything in what xnu believes the real pagetable to be is touched#asahi-dev · 2026-07-25 20:31
chaos_princesssec#asahi-dev · 2026-07-25 20:30
sveni don't think this old mac os release has txm yet#asahi-dev · 2026-07-25 20:30
chaos_princessmaybe its txm?#asahi-dev · 2026-07-25 20:30
svenchaos_princess: oh... nope#asahi-dev · 2026-07-25 20:30
svenessentially i patch every genter/gexit and a whole bunch of msr/mrs to hvc and emulate everything#asahi-dev · 2026-07-25 20:30
chaos_princesssven: just in case - are you sure that it is el0 and not gl0#asahi-dev · 2026-07-25 20:29
yukanice...#asahi-dev · 2026-07-25 20:29
svenyuka: yup#asahi-dev · 2026-07-25 20:29
chaos_princess1. awesome. 2. lol#asahi-dev · 2026-07-25 20:29
yukaThis is emulating SPRR on a device which would already have it?#asahi-dev · 2026-07-25 20:28
svenstill very slow and very unstable because EL0 still wants to update its own pagetables somehow, but sometimes i get far enough for this https://f.svpe.de/6a1342f428a1724f2055ed3dddb6896aaab4603551f8d8802966e4ab170253c7_cursed2.jpg#asahi-dev · 2026-07-25 20:28
jannaut6001-h265 took 324 seconds#asahi-dev · 2026-07-25 20:28
sofusnice! #asahi-dev · 2026-07-25 20:27
sofusyeah i noticed that to, vp9 feels like the slowest one, took ~60 seconds on my t6020 is t6001 much slower?#asahi-dev · 2026-07-25 20:27
jannaut6001-h265: 140/147 passed#asahi-dev · 2026-07-25 20:23
jannausofus: t6001 looks better but it's quite slow#asahi-dev · 2026-07-25 20:20
chaos_princesstarget is everything supported by m1/2 except gpu and dcp#asahi-dev · 2026-07-25 20:20
chaos_princessrelease will not have gpu#asahi-dev · 2026-07-25 20:20
sofuschaos_princess: what about gpu, is that not a lot of work?