#
✕ nick:integralpilot
#asahi-dev · 2026-01-31 08:06
integralpilotdefinetly understand that ghidra or anything looking at apple's compiled code is not allowed and won't do that.
#asahi-dev · 2026-01-31 08:06
integralpilotapplegpu has a built in disassemble.py tool that you can use on a metal program you write, seems useful but won't touch it if it's not allowed
#asahi-dev · 2026-01-31 08:02
integralpilotjust confirming, it's okay to write simple metal shaders myself, compile it, then look at the GPU instructions it produces? the copyright policy webpage seems to says it's okay, but i just wanted to 100% confirm before i do anything that's not right or might taint me
#asahi-dev · 2026-01-31 06:15
integralpilotthanks!
#asahi-dev · 2026-01-31 05:50
integralpilotoh also, why does the nvme/ANS bindings say "Do not add additional SoC to this list". what should be used instead of apple,t8122-nvme-ans2?
#asahi-dev · 2026-01-31 05:39
integralpilottalking from fedora on m3 btw, so can confirm PCIe and WiFi works for me too
#asahi-dev · 2026-01-31 05:39
integralpilotokay, will do that then, thanks
#asahi-dev · 2026-01-31 04:02
integralpilotso like compadible = "apple,t8122-pcie", "apple,t6020-pcie" should work i think
#asahi-dev · 2026-01-31 04:02
integralpilotbut we can omit apple,pcie as t6020-pcie is the main base compadible
#asahi-dev · 2026-01-31 04:02
integralpilottherefore i don't believe we can omit apple,t6020-pcie like you suggested on GH
#asahi-dev · 2026-01-31 04:02
integralpilotnoopwafel: using apple,pcie as the base compadible causes it to fail with -28 (no space i believe) but using apple,t6020-pcie causes it to succeed and PCIe to work
#asahi-dev · 2026-01-30 12:24
integralpilothave put WIP work at https://github.com/AsahiLinux/linux/pull/445
#asahi-dev · 2026-01-30 12:09
integralpilotthank you, that makes sense
#asahi-dev · 2026-01-30 12:04
integralpilotoh makes sense, thanks
#asahi-dev · 2026-01-30 12:03
integralpilotno?
#asahi-dev · 2026-01-30 12:03
integralpilotprehaps AIC3 support should be added to it, otherwise the branch is kinda useless on its own for testing
#asahi-dev · 2026-01-30 12:02
integralpiloti just realised the bits/001-devicetree-m3 doesn't have AIC3 support and that's why it was crashing in early boot, i spent far too long trying to debug it thinking it was an issue with the DT
#asahi-dev · 2026-01-30 11:22
integralpilotoh thanks!
#asahi-dev · 2026-01-30 11:19
integralpilotmight not be looking in the right place but it doesn't seem to be getting filled for me
#asahi-dev · 2026-01-30 11:19
integralpilotjannau: in the t8122-jxxx.dtsi you added the framebuffer but made it 0 and said "filled by loader" but i don't see any code in m1n1 that actually fills that?
#asahi-dev · 2026-01-30 09:01
integralpilotso one of those might be a better choice
#asahi-dev · 2026-01-30 09:01
integralpiloti dont think it would be a good idea either as later M5 products are going to ship with 26.2 or 26.3
#asahi-dev · 2026-01-30 09:00
integralpilot?
#asahi-dev · 2026-01-30 08:59
integralpilotsven: i think 26.0
#asahi-dev · 2026-01-30 08:56
integralpilotbut ofc that's harder to trace due to SPTM
#asahi-dev · 2026-01-30 08:56
integralpilotideally, after we fix the NVMe thing, it would be ideal to target 26.smth
#asahi-dev · 2026-01-30 08:55
integralpiloti think 14.x for the version we decide would be a mistake as then it doesn't apply to M4 & M5
#asahi-dev · 2026-01-30 08:32
integralpiloti did most of my stuff with 26.2 before switching for NVMe, so it works on that too besides NVMe i believe
#asahi-dev · 2026-01-30 08:30
integralpilot(i believe 14.8.3)
#asahi-dev · 2026-01-30 08:30
integralpilot14.latest
#asahi-dev · 2026-01-30 08:26
integralpiloti'll try and PR what i can based on my local DT tonight
#asahi-dev · 2026-01-30 07:18
integralpilotbeen a long first day back at school after summer, sorry!
#asahi-dev · 2026-01-30 07:14
integralpilotmeant the nodes in DTs
#asahi-dev · 2026-01-30 07:14
integralpilotbindings was the wrong word, got my terms mixed up sorry
#asahi-dev · 2026-01-30 07:06
integralpilotand using old MTP format for now
#asahi-dev · 2026-01-30 07:06
integralpilotoh sorry just re-read thoughs and think i wasn't clear: i'll work off the base on AsahiLinux/linux for GPU stuff, and will maybe send MBA board and SMC and MTP stuff for M3 DTs as a PR to that (not just bindings)
#asahi-dev · 2026-01-30 06:57
integralpilotmight also send some things like MBA board and SMC and MTP (though keyboard only, as i think my touchpad things are a bit hacky) bindings to AsahiLinux/linux as i think they are pretty stable and ready
#asahi-dev · 2026-01-30 06:56
integralpilotsounds good, i'll just work off that as a base then
#asahi-dev · 2026-01-30 06:14
integralpilotwell maybe we could integrate them into a branch on the main repo (i believe there's already one but it's very outdated with only one board variant and few peripherals), would be better then the current "everyone has it locally and posts a gist when we need to sync" model we've got now
#asahi-dev · 2026-01-30 05:57
integralpilotare there t8122 dts?
#asahi-dev · 2026-01-30 05:39
integralpilotprehaps we should make a github org or something for those working on m3 bringup so we can all commit to the one DT repo and have it be consistent
#asahi-dev · 2026-01-30 05:36
integralpilotwe might also have implemented MTP in different ways so would be good to get that consistent (if so, i'd just use your way i think as mine was a bit hacky)
#asahi-dev · 2026-01-30 05:35
integralpilotnoopwafel: is it possible for you to upload your current t8122 DT somewhere? i'm working on GPU stuff now based on my local DT but don't want our DTs to diverge too much so might be better to work off a consistent base
#asahi-dev · 2026-01-30 05:34
integralpilotjsk: you might want to look into the AOP ("always-on processor", though not actually always on). see if you can get some simple sensors like ambient light or hinge angle working
#asahi-dev · 2026-01-28 12:12
integralpiloti've decoded mov & basic arithmatic now, it's not changed too much and i think i've got the format fully now. but vertex shaders have changed a lot, looking into it now
#asahi-dev · 2026-01-28 11:49
integralpilotokay i've figured out for OpenGL 1.0, we only need to decode about 14 of the instructions (if i'm understanding what i'm reading correctly). i'm working on those 14 now.
#asahi-dev · 2026-01-28 11:23
integralpilotokay so there's a new mystery byte in the mov instruction, will experiment and try to find what it means
#asahi-dev · 2026-01-28 11:07
integralpilot:D
#asahi-dev · 2026-01-28 11:06
integralpilotsven: oh i was working off the main applegpu upstream, i'll take a look at that, thanks!
#asahi-dev · 2026-01-28 11:04
integralpilot(about some things the same, not about xz not working)

plus ›