#
✕ nick:nickchan
#asahi-dev · 2026-06-08 09:09
nickchanhttps://developer.arm.com/documentation/111107/2026-03/AArch64-Registers/ID-AA64DFR0-EL1--AArch64-Debug-Feature-Register-0
#asahi-dev · 2026-06-08 09:08
nickchan(this page is after the lkml arguments)
#asahi-dev · 2026-06-08 09:07
nickchanapple did mention https://developer.apple.com/documentation/virtualization/accelerating-the-performance-of-rosetta TSO here but they did not clearly define TSO either other than "it works with rosetta"
#asahi-dev · 2026-06-08 09:06
nickchanTSO is a switch in the spec says implementation defined register that switch between two spec compliant memory model
#asahi-dev · 2026-06-08 09:06
nickchanTSO is allowed by the arm spec, not in it
#asahi-dev · 2026-06-08 09:05
nickchannot like that matters for linux though so set that aside
#asahi-dev · 2026-06-08 09:05
nickchanwell the issue is custom algo register aren't accessible in non entitled mode yet the arm approved id registers say there should be impdef algo
#asahi-dev · 2026-06-08 09:03
nickchanotherwise it would be a violation of saying impdef exists but there's no impdef
#asahi-dev · 2026-06-08 09:02
nickchanunless they managed to somehow change ID_AA64ISAR1_EL1.API
#asahi-dev · 2026-06-08 09:02
nickchanBut by that logic, they can't disable Apple PAC either
#asahi-dev · 2026-06-08 09:01
nickchanso apple must allowed them in non entitled boot otherwise it's technically a violation
#asahi-dev · 2026-06-08 09:00
nickchanright, the standard register ID_AA64DFR0_EL1 says there's an IMP DEF performance counter
#asahi-dev · 2026-06-08 08:57
nickchani think the performance counters and uncore performance counters are also allowed in non entitled boot
#asahi-dev · 2026-06-08 08:56
nickchanlack of GIC is allowed
#asahi-dev · 2026-06-08 08:55
nickchanmeanwhile, it says this somewhere else on the box: https://developer.apple.com/documentation/virtualization/accelerating-the-performance-of-rosetta
#asahi-dev · 2026-06-08 08:52
nickchanpcc firmware's sepfw
#asahi-dev · 2026-06-08 08:51
nickchanit'd have to be like, before monterey is released since silicon take a long time
#asahi-dev · 2026-06-08 08:50
nickchani think even that's too late to be plausible
#asahi-dev · 2026-06-08 08:49
nickchanfor the lockout to be in m4 then someone must have been pissed off pretty soon after m1's release
#asahi-dev · 2026-06-08 08:47
nickchanhopefully rva23 can rectify some of the issues
#asahi-dev · 2026-06-08 08:46
nickchanm3 ultra would be last hw asahi linux fully functional on, defined as supporting features that: [(linux supported features on all previous HW OR'd together) AND (features that exist in hw of interest)]
#asahi-dev · 2026-06-08 08:41
nickchanin general, if a hw vendor makes private ISA, they have an incentive to spread it as widely as possible, so then people make stuff that only works on their on HW
#asahi-dev · 2026-06-08 08:40
nickchanit likely has something to do with Arm Limited
#asahi-dev · 2026-06-08 08:33
nickchanor in Apple wording, if a boot does not have private ISA, the boot is "non-Apple entitled"
#asahi-dev · 2026-06-08 08:32
nickchanSobek[m]: m4 doesn't have private ISA in raw boot mode
#asahi-dev · 2026-06-08 06:32
nickchanm1 pro and m1 don't have nested virt though
#asahi-dev · 2026-06-08 06:20
nickchanbut if you are asking about t6032, then that's only because it never shipped with macOS 14
#asahi-dev · 2026-06-08 06:17
nickchanas is current m1n1 without nested virtualization though, small patches are required even with private isa
#asahi-dev · 2026-06-08 06:16
nickchanm2/m3 has private isa and nested virtualization so virtualize sptm should not require any patching
#asahi-dev · 2026-06-08 05:24
nickchanfor m3 series though t6032 won't be able to use 14.something
#asahi-dev · 2026-06-08 05:16
nickchanthe gxf as seen on m1 really started on a14
#asahi-dev · 2026-06-08 05:15
nickchanbut the one on a13 behave differently and uses register redirection instead of seperate encoding for some of the GL registers
#asahi-dev · 2026-06-08 05:14
nickchangxf is one year before m1
#asahi-dev · 2026-06-07 14:22
nickchanokay so the status is still the same, optional linear sq on T8103, on T6000 it became required
#asahi-dev · 2026-06-07 14:22
nickchan"assert failed: [7442]:cmd fetch error for host Admin SQ, rresp code 2, status_reg: 0x1000000"
#asahi-dev · 2026-06-07 14:17
nickchanit wouldn't know which thread did what, so it panics
#asahi-dev · 2026-06-07 14:17
nickchanSError exception should be a kernel panic; it's async and all linux knows is that it came from "somewhere"
#asahi-dev · 2026-06-07 14:15
nickchanthe a11 driver is broke-ish until the patch that is still on the list so i do want to test again
#asahi-dev · 2026-06-07 14:14
nickchananyways nvmmu/linear sq is all there on M2+ still
#asahi-dev · 2026-06-07 14:14
nickchani need to do some additional testing on how the t8015-nvme-ans2 compatible behave if used on like t6000
#asahi-dev · 2026-06-07 14:12
nickchanso the ans2 driver does support ans2 :)
#asahi-dev · 2026-06-07 14:11
nickchan"ANS3" is the one found in A14/M1+ that had linear sq, nvmmu stuff
#asahi-dev · 2026-06-07 14:11
nickchan"ANS2" is on Apple A11-A13 and T2, it's roughly similar to the m1 device but it doesn't have linear SQ or nvmmu
#asahi-dev · 2026-06-07 14:10
nickchan"ANS" is the pcie nvme device in Apple A9 and Apple A10
#asahi-dev · 2026-06-06 16:48
nickchanmuch thanks!
#asahi-dev · 2026-06-06 16:33
nickchanwelp, reroll tomorrow I suppose
#asahi-dev · 2026-06-06 16:31
nickchani suppose fixing the submission function worked
#asahi-dev · 2026-06-06 16:31
nickchan[ 2.638555] EXT4-fs (nvme0n1p2): mounted filesystem affaf3a9-1dd4-4323-85f2-42e7fa66f7ea r/w with ordered data mode. Quota mode: none.
#asahi-dev · 2026-06-06 16:28
nickchan0
#asahi-dev · 2026-06-06 16:28
nickchanthe part where it sets q->sq_tail back to -

plus ›