#asahi-re · 2026-07-21 07:13
svensure, go ahead#asahi-re · 2026-07-21 06:35
svenshutup, Claude :p#asahi-re · 2026-07-21 06:34
svensounds more likely they can just disable some regions or something #asahi-re · 2026-07-21 06:33
svenmy bet is that iboot is just broken and that’s why startup doesn’t work though :(#asahi-re · 2026-07-21 06:33
sventhere’s likely just some other init missing or something’s still powered off when you get SErrors #asahi-re · 2026-07-11 09:56
sven"If you have concerns about enforcement actions, please contact the board at board@asahilinux.org and refrain from public speculation. "#asahi-re · 2026-07-06 18:15
svenprobably better in #asahi-dev, but neal has an updated version and wanted to submit that patch#asahi-re · 2026-07-03 09:54
svenhave you tried just doing a normal tlb invalidation? if gxf is off there really shouldn’t be special maintenance requirements #asahi-re · 2026-07-02 16:23
svenbut i'd really expect that iboot sets all that up#asahi-re · 2026-07-02 16:23
svenand we also have to put the rtkit buffers in there#asahi-re · 2026-07-02 16:23
svenSART contains a few regions put there by iboot including the firmware itself iirc#asahi-re · 2026-07-02 16:23
sventhe NVMMU only works together with a NVMe command on either queue#asahi-re · 2026-07-02 16:22
svencould also be that you're missing some additional initialization that was added on t6041#asahi-re · 2026-07-02 16:17
svencould also be that SART is setup incorrectly #asahi-re · 2026-07-02 16:17
svenalso, uh, this happens before you send any command on the admin queue?#asahi-re · 2026-07-02 16:16
svensptm#asahi-re · 2026-07-01 11:18
sven+see#asahi-re · 2026-07-01 11:17
svenif you serrors there’s likely just some bringup missing #asahi-re · 2026-07-01 11:17
svenyou’ll be able to reach everything from el2#asahi-re · 2026-07-01 11:17
svenif you run Linux or m1n1 none of that applies #asahi-re · 2026-07-01 11:16
svenyup, it’s literally only pagetables setup in a way that only allows gl2 access to the nvmmu #asahi-re · 2026-06-18 06:39
svenyeah#asahi-re · 2026-06-18 06:38
svenor is MCORE the one that's in the middle now?#asahi-re · 2026-06-18 06:38
svenlet's just call them max cores then. or monster cores! :P#asahi-re · 2026-06-18 06:36
svenlol#asahi-re · 2026-06-17 14:52
svenyou should also see a difference to 26.2 beta 2 which enables write access to one more sysreg#asahi-re · 2026-05-13 14:57
svenyeah, xnu is especially painful :D#asahi-re · 2026-05-13 14:55
svennah, as mentioned early: there's just not that much disassembling required anyway#asahi-re · 2026-05-13 14:37
svencracking anything is certainly unwanted here#asahi-re · 2026-05-13 14:36
sventhat's what.. 6? 7? years ago now#asahi-re · 2026-05-13 14:35
sveni don't think i ever used ida since ghidra was released tbh#asahi-re · 2026-05-13 14:34
sveni've just switched over to ghidra completely since I couldn't justify the cost of a IDA license with decompilers. it's not unwanted#asahi-re · 2026-05-13 09:52
svensimple example, but blindly copying what macOS does which is tempting when disassembling would be weird there #asahi-re · 2026-05-13 09:52
svene.g. for nvme macOS creates two copies of the pages it needs to r/w but in Linux we can just point both hw regs to the same tables #asahi-re · 2026-05-13 09:51
svenbecause the hw is sometimes driven in macOS specific ways #asahi-re · 2026-05-13 09:50
svenand understanding what something does from tracing or testing sometimes even results in better drivers in the end #asahi-re · 2026-05-13 09:50
svenYeah, we’re just being as careful as possible here. You can do disassembly and still write code that’s effectively cleanroom but why take that risk when it’s just not necessary?#asahi-re · 2026-05-13 08:51
sventhat too #asahi-re · 2026-05-13 08:48
sventhat’s maybe overly cautious but as it turns out disassembling blobs usually isn’t required because tracing is to much faster #asahi-re · 2026-05-13 08:48
svenyour approach seems to be „looking for any reason to disassemble something“ while we prefer „looks for any reason NOT to disassemble anything and use that as a last resort“#asahi-re · 2026-05-13 08:47
svenbut why do we care?#asahi-re · 2026-05-13 04:29
svenagain, we’re not trying to cut corners here and since none of us are lawyers we’re being extra careful when possible #asahi-re · 2026-05-13 04:28
svenif it’s pagetable stuff why not just use the code we already have in m1n1?#asahi-re · 2026-05-11 08:19
svenhuh, I would’ve expected at least the wkdm compression stuff next to genter/exit #asahi-re · 2026-05-10 07:40
svenit should be possible to disable most extensions in userspace as well simply because vmapple userspace is the same and can’t use any extensions either #asahi-re · 2026-05-10 06:48
svenuserspace is also going to hit AMX but you may be able to disable that #asahi-re · 2026-05-10 06:48
svenyeah, which is fine because code segments are rx and not w anyway #asahi-re · 2026-05-10 06:46
svenalso lots of sysregs for stuff like apple pac but most of that can be ignored I think #asahi-re · 2026-05-10 06:45
svenI know about some compress/uncompress instructions #asahi-re · 2026-05-10 06:45
svenYeah, instructions