#asahi-dev · 2026-06-16 19:09
yukaconfirmed SYS_IMP_APL_VM_TMR_FIQ_ENA_EL2 write does not crash on 26.6 beta 2 on A18 Pro#asahi-dev · 2026-06-16 18:58
svennot sure how we would pass that to u-boot, but that can be bikeshedded later#asahi-dev · 2026-06-16 18:57
svenu-boot just has to be taught to pass along yet another opaque config table and the right memory mappings to ensure m1n1 is mapped#asahi-dev · 2026-06-16 18:56
jannauhow would that work with u-boot in the mix?#asahi-dev · 2026-06-16 18:49
sven[ 0.000000] psci: PSCI-via-EFI conduit#asahi-dev · 2026-06-16 18:49
sven[ 0.000000] efi: RTPROP=0x81508c040 #asahi-dev · 2026-06-16 18:49
sven[ 0.000000] EFI: PSCI = (____ptrval____)#asahi-dev · 2026-06-16 18:49
sven[ 0.000000] efi: EFI v2.0 by m1n1#asahi-dev · 2026-06-16 18:49
sventried to prototype psci-via-efi and ended up with a very hacky thing. it doesn't seem to call my m1n1 code yet, but everything should be in place:#asahi-dev · 2026-06-16 18:47
jannaunickchan: I did diff HID* and other Apple sysregs. one flipped bit in each HID4 and HID18 but those make no difference. Did you have anything else in mind?#asahi-dev · 2026-06-16 16:35
nickchanjannau: have you checked sysreg state?#asahi-dev · 2026-06-16 15:49
jannauany idea why the first perf core (boot core) on M3 would be slower than the other 3? only 41% in coremark independent of p-state#asahi-dev · 2026-06-16 14:31
chaos_princessit is a guess#asahi-dev · 2026-06-16 14:30
nickchanand let me guess; no reply on the feedback#asahi-dev · 2026-06-16 14:18
fl0id_oftcgl then :)#asahi-dev · 2026-06-16 14:17
mazfl0id_oftc: she also has an M1 air, and just finished her degree anyway... :D#asahi-dev · 2026-06-16 14:16
fl0id_oftcwhich do they get instead?#asahi-dev · 2026-06-16 14:16
fl0id_oftcmaz haha lol#asahi-dev · 2026-06-16 14:15
sven:-)#asahi-dev · 2026-06-16 14:15
mazholy crap. now I need to pry the M4 from the SO's hands...#asahi-dev · 2026-06-16 13:40
jannauSYS_IMP_APL_VM_TMR_FIQ_ENA_EL2 access is fixed in 26.6 beta 2 (only tested M4/t8132)#asahi-dev · 2026-06-16 13:38
chaos_princessyes, i have the t6031#asahi-dev · 2026-06-16 13:36
yukat6030 specifically?#asahi-dev · 2026-06-16 13:33
chaos_princessthere are now 4 fields instead of 3#asahi-dev · 2026-06-16 13:33
chaos_princesswell, not unchanged, but rather not changed in a way that affects the driver#asahi-dev · 2026-06-16 13:32
chaos_princessalso, pmgr misc seems to be unchanged from t602x, will send that once someone sends me the t6030 xnu logs, as i am not sure how to figure out which specific pmgr mmio is responsible for it w/o that#asahi-dev · 2026-06-16 13:30
yukaI don't think I'll find time to do that today#asahi-dev · 2026-06-16 13:30
yukaAhh, I did not run the pmp specific tracer yet. Just have an older full trace lying around.#asahi-dev · 2026-06-16 13:29
chaos_princessjust run macos under the tracer, it will be obvious as to what is needed#asahi-dev · 2026-06-16 13:28
chaos_princessalso, change dev.flags.notify_pmp to dev.id1 !=0, there is some random stuff like the afnc bits that need to be always-on, but aren't marked as such#asahi-dev · 2026-06-16 13:27
chaos_princessi actually thought of doing it later today :P#asahi-dev · 2026-06-16 13:26
yukanice#asahi-dev · 2026-06-16 13:26
chaos_princesshttps://github.com/asahilinux/m1n1/blob/main/proxyclient/tools/pmp_assist.py - just use this#asahi-dev · 2026-06-16 13:26
yukaI think the t6031 pmp-report offsets might be unchanged from t6020, at least they appear a lot in my trace#asahi-dev · 2026-06-16 13:26
yukaI'm looking into it, yeah#asahi-dev · 2026-06-16 13:26
chaos_princessare you trying to add pmp support to t603x?#asahi-dev · 2026-06-16 13:25
yukachaos_princess: is the report-entry reg unk1_2 in the adt pmgr devices?#asahi-dev · 2026-06-16 13:06
chadmedyou can (and sometimes need to) disassemble your own shaders to poke at the ISA etc#asahi-dev · 2026-06-16 13:05
fl0id_oftcthis from alyssa might be helpful, though seems to more of a disassembler... https://github.com/AsahiLinux/gpu#asahi-dev · 2026-06-16 13:03
yukahttps://www.youtube.com/watch?v=SDJCzJ1ETsM#asahi-dev · 2026-06-16 13:02
yukahttps://www.youtube.com/watch?v=ObS6sdfus2w#asahi-dev · 2026-06-16 13:02
yukaTo get some idea of what you're getting yourself into, I can also recommend watching the XDC conference talks on the previous Apple GPU work#asahi-dev · 2026-06-16 13:01
chaos_princessstart here https://developer.apple.com/metal/#asahi-dev · 2026-06-16 13:01
JailbreakHat[m]Is there any sample shader program I can look at to get an idea?#asahi-dev · 2026-06-16 13:00
chaos_princesswell, the harness is objc(++), shader itself is msl#asahi-dev · 2026-06-16 13:00
chadmedor yeah objc#asahi-dev · 2026-06-16 13:00
chadmedyoull have to write them in MSL which much like other shader languages is just spicy c++#asahi-dev · 2026-06-16 13:00
chaos_princessobjc or objc++. Metal supports those#asahi-dev · 2026-06-16 12:59
JailbreakHat[m]fl0id_oftc: I am already doing that.