#
✕ nick:nickchan
#asahi-dev · 2026-07-18 05:15
nickchancpu6 - cpu9 is cluster-type = P (marketing name: super core)
#asahi-dev · 2026-07-18 05:14
nickchanand /cpus/cpu0 to /cpus/cpu5 in adt seems to be efficiency core
#asahi-dev · 2026-07-18 05:14
nickchanthere are 10 cpu cores total in base m5
#asahi-dev · 2026-07-18 05:14
nickchanrhlvista[m]: that doesn't seem to be a problem
#asahi-dev · 2026-07-18 05:10
nickchanexpect more multi die shenanigans in the mac studio soon
#asahi-dev · 2026-07-18 05:09
nickchanand their device trees both have double the amount of cpu in m5 max
#asahi-dev · 2026-07-18 05:09
nickchangiven that m5 pro and m5 max is both t6050
#asahi-dev · 2026-07-18 05:02
nickchanfollowed by even more unifdef ofc
#asahi-dev · 2026-07-18 05:02
nickchanthen it's a19 pro and a19 though the comments seems to be copy pasted wrong
#asahi-dev · 2026-07-18 05:01
nickchanm5 pro and m5 max is sorta_s and sorta_c and there's something unifdef's immediately after based on the spacing
#asahi-dev · 2026-07-18 04:59
nickchanbase m5 is hidra
#asahi-dev · 2026-07-18 04:59
nickchanrefer to /System/Library/Frameworks/Kernel.framework/Versions/A/Headers/arm/cpuid.h in macos 27 sdk
#asahi-dev · 2026-07-18 04:59
nickchanyuka: tupai is base A18
#asahi-dev · 2026-07-15 07:46
nickchanDTK goes like this: Units returned to Apple => Shipped to China for recycling and destruction => Each unit is damage in some way but all in *different* parts => these scrap metal with different damages ended up in people hands => started cannibalizing parts => some working units are made => (about a year) => apple broke the activation
#asahi-dev · 2026-07-15 07:39
nickchanregardless of activation status which apple broke like a year ago
#asahi-dev · 2026-07-15 07:38
nickchan actually, definitely possible without no-no software because of the bootrom exploit, but the universal app quick start program ended years ago
#asahi-dev · 2026-07-15 07:37
nickchanbut then again, no one is supposed to have it now
#asahi-dev · 2026-07-15 07:36
nickchanthe main difference with dtk is it is actually possible to run custom code on them without any no-no software, not sure about it
#asahi-dev · 2026-07-15 07:33
nickchanmany of them come from recyclers i think, but even ignoring the device issue, the software needed to run custom code on them is definitely not fine
#asahi-dev · 2026-07-14 19:32
nickchanit can be put another way: any slop output is XNU disassembly under the policy
#asahi-dev · 2026-07-14 09:09
nickchani know for sure that t8012/t8015/t8020 is compatible with how linux does things, not sure about t8027/t8030 but probably still compatible given that it's designed for t8103
#asahi-dev · 2026-07-14 09:08
nickchanyuka: t8012/t8015/t8030=>spmi,gen0; t8020/t8027=>spmi,t8015 (yes t8015 does not have t8015 and not t8015 has t8015)
#asahi-dev · 2026-07-13 17:43
nickchanthe boot arg controls the number of microseconds before a ticket spinlock times out
#asahi-dev · 2026-07-13 17:23
nickchantlto_us=900000 boot-arg may help
#asahi-dev · 2026-07-12 11:52
nickchanat leasr for me i would rather not have those tools potentially spitting out leaked materials
#asahi-dev · 2026-07-12 11:36
nickchanit's been unraveling at least for github copilot and claude code
#asahi-dev · 2026-07-12 11:25
nickchanAssisted-by: insert-ad-here-2.0
#asahi-dev · 2026-07-12 11:06
nickchanthe prove of work has been turned on really high
#asahi-dev · 2026-07-12 07:17
nickchanby the way i retract my previous assumption about a11 smc not having rtc it actually does it's just called sRTC instead of CLKM
#asahi-dev · 2026-07-12 07:14
nickchanthe macsmc-input also went nowhere and last i checked it doesn't build and it didn't break the kernel build only because of the kconfig option is wrong too (so there's no attempt to compile)
#asahi-dev · 2026-07-11 04:29
nickchanm1n1 gets chip-id from /chosen chip-id
#asahi-dev · 2026-07-11 04:28
nickchanall the big m5 are supposedly t6050
#asahi-dev · 2026-07-11 04:27
nickchanyuka: I *believe* T6051 is not a thing
#asahi-dev · 2026-07-10 13:48
nickchannow ofc el0 usage of private isa is disabled in that hypervisor implementation but like, you don't need custom kernel to document things accessible from el0, and people have been doing it since m1
#asahi-dev · 2026-07-10 13:46
nickchanso it's slightly funny
#asahi-dev · 2026-07-10 13:46
nickchani have a feeling that by essentially forcing people to soft model private isa there would be more documentation on it, not less
#asahi-dev · 2026-07-10 10:12
nickchanplus, could jsut brute force the offset
#asahi-dev · 2026-07-10 10:11
nickchanthe api is the same in every 64 bit soc but the cpu start offset differ
#asahi-dev · 2026-07-10 04:09
nickchanno
#asahi-dev · 2026-07-08 17:27
nickchani am typing this from a macbook pro 2021 14 inch with m1 pro
#asahi-dev · 2026-07-08 17:25
nickchanno operating system image support these machines
#asahi-dev · 2026-07-08 17:22
nickchan16.8.3
#asahi-dev · 2026-07-08 17:20
nickchanand for CPU_SUSPEND either state is kept like with retention deep wfi so kernel stack and all, or lose state then it's like CPU_ON
#asahi-dev · 2026-07-08 17:20
nickchanfor CPU_OFF the secondary stack would not even be used
#asahi-dev · 2026-07-08 17:19
nickchanfor CPU_ON the secondary is outside the kernel with mmu off so from that core's perspective the stack is perfectly accessible
#asahi-dev · 2026-07-08 17:19
nickchanput it as reserved memory no map; then
#asahi-dev · 2026-07-08 17:18
nickchani don't understand why the secondary stack must be accessible in the EFI PSCI context
#asahi-dev · 2026-07-08 16:20
nickchanon a11 the pcore and ecore is in same cluster (?) and there's a ACC_EBLK_OVRD that is should be but never set in m1n1
#asahi-dev · 2026-07-08 16:19
nickchana11 has been working by luck this whole time :)
#asahi-dev · 2026-07-08 16:19
nickchanhonestly though cpu sleep is magic

plus ›