#asahi-re · 2026-07-30 09:03
svensobek[m]: siguza originally called genter "bless" and gexit "curse" :P#asahi-re · 2026-07-30 09:02
jannauarm documents it since 2025-09 as "Future Architecture Technologies" i.e. not yet assigned to a specific armv9.x extension#asahi-re · 2026-07-30 09:02
Sobek[m](Tenter really read to me as tempter, this is clearly evil code to push CPUs to sin 😈)#asahi-re · 2026-07-30 08:59
JamesCalligeros[m]We are in -re, safe to assume disasm. This is entirely for future hw#asahi-re · 2026-07-30 08:59
sventhis is obviously disassembly and i can't quite run this on M1 yet so i'm sure that check it does just above routes it to genter#asahi-re · 2026-07-30 08:59
chaos_princessthis is future code for m-whatever#asahi-re · 2026-07-30 08:59
chaos_princessno existing apple soc supports this feature#asahi-re · 2026-07-30 08:58
Sobek[m]sven: Uh, M1 supports that feature ? Or is this code that checks for the feature and does something else on M1 when unavailable ? (Also, public source release, or disassembly ?)#asahi-re · 2026-07-30 08:58
chaos_princessit is sprr/gxf that is a hard requirement for most kernel/fw bits and is particularly hard to emulate#asahi-re · 2026-07-30 08:58
svenlet me trace what sets that bit it checks, i bet its either some early init check for ID_AA64MMFR4 or just passed by iboot#asahi-re · 2026-07-30 08:58
chaos_princesssobek[m]: most other "ideas" are easy to turn off#asahi-re · 2026-07-30 08:57
jannaudo we have m5 / m5pro ID_AA64MMFR4?#asahi-re · 2026-07-30 08:57
Sobek[m](On the other hand, even if the move to FEAT_TEV, what tells us that Apple wont't keep innovating and shipping ideas before they get standardised by ARM ?)#asahi-re · 2026-07-30 08:57
svenyup#asahi-re · 2026-07-30 08:57
JamesCalligeros[m]Oh lol is this just GXF but architected#asahi-re · 2026-07-30 08:57
svenm1 xnu from the 27 beta#asahi-re · 2026-07-30 08:56
Sobek[m]Where did you see that code ?#asahi-re · 2026-07-30 08:55
svenso our HV woes are only temporary until M6 or M7 or Mwhatever just uses standard ARM stuff#asahi-re · 2026-07-30 08:53
svenhah... so it looks like they'll move to FEAT_TEV and tenter eventually#asahi-re · 2026-07-30 08:53
svenfffffe000c36a310: d65f03c0 ret#asahi-re · 2026-07-30 08:53
svenfffffe000c36a30c: d4e00000 tenter #0x0#asahi-re · 2026-07-30 08:53
svenfffffe000c36a308: d2e04010 mov x16, #0x200000000000000 // #144115188075855872#asahi-re · 2026-07-30 08:53
svenfffffe000c36a304: d65f03c0 ret#asahi-re · 2026-07-30 08:53
svenfffffe000c36a300: 00201420 .inst 0x00201420 ; undefined // genter#asahi-re · 2026-07-30 08:53
svenfffffe000c36a2f4: 3965e210 ldrb w16, [x16, #2424]#asahi-re · 2026-07-30 08:53
svenfffffe000c36a2fc: d2e04010 mov x16, #0x200000000000000 // #144115188075855872#asahi-re · 2026-07-30 08:53
svenfffffe000c36a2f0: 90fddb50 adrp x16, 0xfffffe0007ed2000#asahi-re · 2026-07-30 08:53
svenfffffe000c36a2f8: 35000090 cbnz w16, 0xfffffe000c36a308#asahi-re · 2026-07-30 08:53
svenfffffe000c36a2ec: d503245f bti c#asahi-re · 2026-07-24 08:06
sveni'll be very very happy if you (or anyone else) figures out what's wrong :D#asahi-re · 2026-07-24 08:05
svenyeah#asahi-re · 2026-07-24 07:59
jannauI still want to try the downstream atc.c from asahi-6.17 to see if that changes anything. easier to argue over code changes than mostly meaningless minor trace chnages#asahi-re · 2026-07-24 07:50
svenand then there’s also that weird tbt issue where the MMIO traces look similar enough and yet the tbt connection just won’t come up :(#asahi-re · 2026-07-24 07:44
JamesCalligeros[m]And if we have to do the xbar work for pro and base anyway#asahi-re · 2026-07-24 07:30
jannaumax and ultra have enough dcpexts for each atc backed type-c port for dp-altmode. keeping things simple and concerns of power use are reasons for not doing it#asahi-re · 2026-07-23 21:08
chaos_princesson pro there isn't enough#asahi-re · 2026-07-23 21:08
chaos_princessthere is an asmedia usb hub in front of them#asahi-re · 2026-07-23 21:08
JamesCalligeros[m]But pro only has two dcpexts on m1 and three on m2 right#asahi-re · 2026-07-23 21:08
JamesCalligeros[m]I havent tried#asahi-re · 2026-07-23 21:07
chaos_princessthe front ports on the max studio are not capable of dp output, right?#asahi-re · 2026-07-23 21:06
chaos_princessi think there is?#asahi-re · 2026-07-23 21:05
JamesCalligeros[m]On laptops sure but there arent enough dcpexts to do that for the desktops#asahi-re · 2026-07-23 21:04
chaos_princessmore complex on vanilla/pro, as there we have to care about crossbar configuration#asahi-re · 2026-07-23 21:03
chaos_princesswould not even be complicated on max SoCs, as we can just make every port the "blessed port"#asahi-re · 2026-07-23 21:03
JamesCalligeros[m]dp alt-mode works on a single blessed port in the fairydust kernel branch but requires work to get it going across all ports#asahi-re · 2026-07-23 21:01
chaos_princesswe have the atcphy parts, the missing bits are in other drivers#asahi-re · 2026-07-23 21:00
djwhat’s the status on it in general? do we have thunderbolt 4 / usb c displays close to working on m1/m2?#asahi-re · 2026-07-23 20:58
chaos_princessm5 may be useful one day#asahi-re · 2026-07-23 20:56
djm5 pro, m2 pro and planning to get my hands on a m1#asahi-re · 2026-07-23 19:51
chaos_princesswhich hw