#asahi-dev · 2026-10-09 09:54
NickChan[m]use boot args and bputil boot args filtering bit#asahi-dev · 2026-10-09 09:50
NickChan[m]as a bonus, this doesn't break the security model in terms of one OS interfering another OS#asahi-dev · 2026-10-09 09:49
NickChan[m]jannau: csrutil debugging restriction bit could be used#asahi-dev · 2026-10-08 17:58
NickChan[m]it would not be possible to know for sure it is the guest you just exited for sure even though it probably is#asahi-dev · 2026-10-08 17:58
NickChan[m]can't really put an SError back into the guest i think because it's asynchronous #asahi-dev · 2026-10-07 07:11
NickChan[m]<MichaelReeves[m]> "they'd probably just say QEMU/..." <- this was true but in more recent versions of the architecture there are seperate feature reg for stage1 and stage2 translation granule supoort#asahi-dev · 2026-09-16 12:20
NickChan[m]If you come across some linux supporting A5/A6 it's probably slop 🫠 #asahi-dev · 2026-09-16 12:20
NickChan[m]so since the topic is slop#asahi-dev · 2026-09-09 23:22
NickChan[m]I hope I am wrong though#asahi-dev · 2026-09-09 23:21
NickChan[m](mostly the cpu cluster stuff is changed and that's not good news since raw boot object is very untested)#asahi-dev · 2026-09-09 23:21
NickChan[m]with the device tree changes shown wouldn't be surprised if raw boot objects don't work at all#asahi-dev · 2026-08-28 16:18
NickChan[m]about exclaves disable#asahi-dev · 2026-08-28 16:17
NickChan[m]...what does that even mean? There's no GL in raw boot mode#asahi-dev · 2026-08-27 07:05
NickChan[m]so the ifdef is not triggered and all the new hardware code and stuff is shipped to public#asahi-dev · 2026-08-27 07:04
NickChan[m]the theory is that they have an ifdef to hide m6 stuff from beta ("seed") builds, but then they backported stuff to 26.7 and forgot backports/security updates aren't seeds#asahi-dev · 2026-08-27 04:42
NickChan[m]TXM and ExclaveKit runs in GL0, GL0 always existed I think#asahi-dev · 2026-08-22 05:28
NickChan[m]chaos_princess: it also applies to a12, so maybe 5 users?#asahi-dev · 2026-08-17 09:55
NickChan[m]m7 motion coprocessor !!! /j#asahi-dev · 2026-08-17 09:54
NickChan[m]though I suppose it's cursed in reality given how the ways it has been broken in the past#asahi-dev · 2026-08-17 09:53
NickChan[m]It ought to be mostly stable otherwise the iBoot1->iBoot2 handoff for the boot logo would quickly become very cursed#asahi-dev · 2026-08-08 12:12
NickChan[m]pac is an example#asahi-dev · 2026-08-08 12:12
NickChan[m]In the form of "this standard defined feature registers says this feature is implemented as IMPDEF but it does not work"#asahi-dev · 2026-08-08 12:11
NickChan[m](approximately, some disable actually breaks the standard)#asahi-dev · 2026-08-08 12:07
NickChan[m]another solution would be to buy ARM but regulators wojld probably stop that#asahi-dev · 2026-08-08 11:57
NickChan[m]there's also a "APB_RESET_N" in m1n1's atc python #asahi-dev · 2026-08-08 11:51
NickChan[m]"APB" do exist on apple hw i think but probably not stand for the same thing#asahi-dev · 2026-08-02 12:07
NickChan[m]can't exactly commit to 14.8.3 anyways because m3 ultra mac studio needs at least 15.1#asahi-dev · 2026-07-28 08:54
NickChan[m]And if you make it think it has full SPRR then it would try to use the EL0 switching registers which can't easily be trapped#asahi-dev · 2026-07-28 08:54
NickChan[m]so the problem is you can't make it think the machine is a VMAPPLE because then it would try to make rwx mappings that won't work if you are virtualizing SPTM and SPRR#asahi-dev · 2026-07-28 08:48
NickChan[m]arm specification says it is implementation defined for each encoding#asahi-dev · 2026-07-28 08:48
NickChan[m]I do not know if TIDCP traps the EL0 registers on apple silicon#asahi-dev · 2026-07-28 08:47
NickChan[m](using the same physical memory but mapped rx in one location rw in another could also work, but safari does not do that)#asahi-dev · 2026-07-28 08:46
NickChan[m]and changing pagetable requires sptm #asahi-dev · 2026-07-28 08:46
NickChan[m]sven: if there's no jit fast switching then i think it does jit by changing pagetable since rwx don't exist under sprr #asahi-dev · 2026-07-26 17:34
NickChan[m]to do stuff like that on macOS I think people uses a kernel read write kext instead of patching that#asahi-dev · 2026-07-26 17:34
NickChan[m]chaos_princess: that's one check but you would also need to patch the check for kernel_pmap and probably the check when converting the task port#asahi-dev · 2026-07-25 13:39
NickChan[m]sven: why do you need to replace registers? TIDCP should trap them#asahi-dev · 2026-07-23 10:02
NickChan[m]looking at S3C2410_UCON_CLKMASK#asahi-dev · 2026-07-23 10:02
NickChan[m]yuka: i believe linux clears it because the uart clock in device tree is clk_uart_baud0 and as the only baud clock #asahi-dev · 2026-07-20 16:57
NickChan[m]m3?#asahi-dev · 2026-07-14 09:16
NickChan[m]assuming the concern is about the spmi generation i think it's fair to consider both spmi,gen0 and spmi,t8015 as generation 0#asahi-dev · 2026-07-09 07:14
NickChan[m]make sure bit 0 in CYC_OVRD is cleared#asahi-dev · 2026-07-09 06:55
NickChan[m]could happen if irq/fiq is not disabled in CYC_OVRD#asahi-dev · 2026-07-09 06:54
NickChan[m]I have seen cores ending up at rvbar in m1n1 multiple times before it looks exactly like that#asahi-dev · 2026-07-08 13:59
NickChan[m]maybe top_of_memory_alloc() can be used?#asahi-dev · 2026-07-07 17:50
NickChan[m]ideally it would have been jumping into asm that doesn't use stack but that if everything use psci, that's another way to get rid of the reserved stack#asahi-dev · 2026-07-07 17:49
NickChan[m]reserving the stack for spin table has always felt been a bit of a hack to me#asahi-dev · 2026-07-03 14:13
NickChan[m]out of 69 of those dts files i say maybe 30 of them is tested#asahi-dev · 2026-07-03 14:12
NickChan[m]that isn't Asahi Linux is it#asahi-dev · 2026-07-02 03:47
NickChan[m]in which case... (full message at <https://matrix.org/oftc/media/v1/media/download/AZGUH4kLBBXK0oIZaUyA8vBs7vJqQ-Vl6bLN9tEN12KJGIjtMfPLaf8ghlfhSD1skp65PlIw_pqU3TIKvtNhHExCefaQnBdwAG1hdHJpeC5vcmcvWGNwbVhESG9FcUxCanJrSFhYTHZJZWlT>)