#
✕ nick:chadmed
#asahi-dev · 2026-06-30 09:40
chadmedgave you maintainer access to the fw repo which should let you do ~everything except org level admin
#asahi-dev · 2026-06-30 09:38
chadmedall done :)
#asahi-dev · 2026-06-30 09:37
chadmedyou should be able to transfer the repo now. once thats done ill add you as a collaborator to that repo
#asahi-dev · 2026-06-30 09:36
chadmedsofus: check your emails
#asahi-dev · 2026-06-30 09:35
chadmedok one second
#asahi-dev · 2026-06-30 09:33
chadmedaa sorry theyve changed how this works again
#asahi-dev · 2026-06-30 09:30
chadmedone sec
#asahi-dev · 2026-06-30 09:18
chadmedlooking at isci which is dual gpl/bsd they included it
#asahi-dev · 2026-06-30 09:17
chadmedlooks like it
#asahi-dev · 2026-06-30 09:16
chadmedso we submit built blobs _and_ the source?
#asahi-dev · 2026-06-30 09:10
chadmedif youre in doubt just do gpl only. i dont think thered be anything useable that is also copyrightable in there for the bsds anyway
#asahi-dev · 2026-06-30 09:08
chadmedthe repo is just a bunch of blobs so perhaps not
#asahi-dev · 2026-06-30 09:05
chadmedalso consider moving the fw to the asahilinux github org for discoverability etc. we can give you read/write perms over the repo so you dont really lose anything.
#asahi-dev · 2026-06-30 09:04
chadmeddual licence allows the bsds to look at the driver and copy parts of it for their own use
#asahi-dev · 2026-06-30 09:02
chadmedbut it cannot _not_ be gpl
#asahi-dev · 2026-06-30 09:02
chadmeddriver also must be gpl and optionally -or-MIT
#asahi-dev · 2026-06-30 09:01
chadmedif so yes, you will need to add Copyright (c) xxxx Rockchip or whatever
#asahi-dev · 2026-06-30 09:01
chadmedwhen you say "based on" do you mean "copied"
#asahi-dev · 2026-06-30 08:58
chadmedso pick whichever licence you actually want to use, and take my S-o-b for my commits
#asahi-dev · 2026-06-30 08:58
chadmedi only said that for the benefit of the bsds but then i realised they dont need to actually incorporate the code at all and can just take the blobs because its firmware
#asahi-dev · 2026-06-30 08:57
chadmedsofus: yea pretty much. probably also add a LICENCE file to the fw repo
#asahi-dev · 2026-06-30 08:50
chadmedthat would be a nice goal eventually but i dont think its realistic until it is some semblance of complete with the release cadence of linux-firmware
#asahi-dev · 2026-06-30 08:48
chadmedalso please actually assign copyright and such to the firmware, we cant really ship it until you do
#asahi-dev · 2026-06-30 08:47
chadmedfor the driver itself one commit is typically acceptable for a new driver but making it atomic at the "add hevc support" level or support for various soc quirks might also make sense
#asahi-dev · 2026-06-30 08:47
chadmedimportant thing to make everyones lives easier: make sure your dt binding is its own commit, your dts changes are their own commits for each soc (or soc series)
#asahi-dev · 2026-06-29 12:04
chadmedi mean a good chunk of the pages are still just trace dumps and "i think bit x of 0x69420 is the interrupt enable bit but idk"
#asahi-dev · 2026-06-29 12:03
chadmedenverbalalic: the sidebar is really supposed to be navigation to like landing pages where you go and find the interesting stuff but we never got the docs into a well-hyperlinked state
#asahi-dev · 2026-06-29 11:58
chadmedthat should probably include a "how to read the output of m1n1.adt and avoid common papercuts like forgetting to offset the arm-io addresses" paragraph too :P
#asahi-dev · 2026-06-29 11:56
chadmedprobably a recent thing to prevent slop bots from training on our chats
#asahi-dev · 2026-06-29 11:55
chadmedi used to try and keep the feature support pages at least up to date but i stopped being an unemployed bum uni student :p
#asahi-dev · 2026-06-29 11:54
chadmedanyway re docs iteration kinda slowed down when we moved from a wiki to a proper repo but for every good edit someone made we got like 5 or 6 griefers so it was a necessary change
#asahi-dev · 2026-06-29 11:53
chadmedthe "eureka" (lol) moment is so addicting, why would anyone want to gatekeep it
#asahi-dev · 2026-06-29 11:52
chadmedyeah we dont want that
#asahi-dev · 2026-06-29 11:51
chadmedsemi-related but the recent news about vibeos is especially funny to me. i thought i was clever making an "operating system" in ms powerpoint in my lunch breaks in grade 3. these people are doing the same thing but it cost billions of dollars and theyre ostensibly fully functioning adults
#asahi-dev · 2026-06-29 11:48
chadmed"the code is the documentation :3
#asahi-dev · 2026-06-29 11:47
chadmeddevelopers suck at writing docs and dont want to do it, more news at 11
#asahi-dev · 2026-06-29 11:45
chadmedoh yeah lmfao
#asahi-dev · 2026-06-29 11:44
chadmedi would be curious to see how far a base m4 machine could get by dumping the adt, writing an fdt, then yoloing a cpus=1 kernel on to the thing
#asahi-dev · 2026-06-29 11:42
chadmedthis will at least give you a safe playground with a fixed scope to play around in and learn how to stare at hexdumps and stuff like that
#asahi-dev · 2026-06-29 11:42
chadmedi think your best bet would be to wait until we can at least reliably run m1n1 and macos in a hypervisor, pick a piece of hardware that _worked on m1/m2_ but _does not on m4_ and use the prior art in m1n1 to explore what has changed
#asahi-dev · 2026-06-29 11:40
chadmed(do as i say not as i do in this regard btw)
#asahi-dev · 2026-06-29 11:39
chadmeddedication and a certain set of skills. this is by no means a simple platform and doing early bringup work on new socs is not a suitable "id like to learn hardware reverse engineering" project
#asahi-dev · 2026-06-29 11:36
chadmedbut to answer your question, it should be totally safe to trial and error test them. it will either work or it will not, there shouldnt be any flakiness
#asahi-dev · 2026-06-29 11:35
chadmedjannau: i cant think of any reason the idle masks would be inverted other than the amps being yet another 2764 variant where the bus keeper's config registers are the opposite... the active tx slots are the same....
#asahi-dev · 2026-06-29 11:11
chadmedmight be obvious but is rvbar at m1n1?
#asahi-dev · 2026-06-29 09:25
chadmedpr added, should be sorted
#asahi-dev · 2026-06-29 09:12
chadmedok its just j413 that needs fixing
#asahi-dev · 2026-06-29 08:11
chadmedand 313
#asahi-dev · 2026-06-29 08:10
chadmedi set the indexes to 1 accidentally for j413
#asahi-dev · 2026-06-29 08:10
chadmedah fuck this one's my fault

plus ›