#
#asahi-dev · 2026-01-31 22:16
jannauiirc I added it to boot multiple concatenated cpios
#asahi-dev · 2026-01-31 22:14
jannaunoopwafel: you could add a initramfs_magic header like run_guest_kernel.sh
#asahi-dev · 2026-01-31 22:08
noopwafel^- me :o
#asahi-dev · 2026-01-31 22:08
noopwafel"We could handle this, but who uses uncompressed initramfs?
#asahi-dev · 2026-01-31 21:51
jannaukettenis: how well tested / used is nvme write in u-boot via efi? https://codeberg.org/Limine/Limine/commit/0491f550c1b3 is reported to avoid https://github.com/AsahiLinux/linux/issues/421
#asahi-dev · 2026-01-31 20:45
bartalso saw other report on reddit that was one port as well
#asahi-dev · 2026-01-31 20:45
bartsame hardware, same port
#asahi-dev · 2026-01-31 20:45
bartyeah
#asahi-dev · 2026-01-31 20:45
noopwafelhuh. still only port 2?
#asahi-dev · 2026-01-31 20:35
bartrecent kernel upgrade also made it worse as the ports seem to be initiated at boot, causing the device to be unbootable
#asahi-dev · 2026-01-31 20:34
bartanyone know what could be causing this? https://github.com/AsahiLinux/linux/issues/156
#asahi-dev · 2026-01-31 19:54
bigmacthank you leio
#asahi-dev · 2026-01-31 18:08
leioprobably would just take installing libxkbcommon distro dev package and reconfiguring whatever you're actually building.
#asahi-dev · 2026-01-31 18:08
leioPersonally I would investigate why it's building libxkbcommon as a subproject in the first place; if it found the system libxkbcommon development pkg-config file and headers, it probably wouldn't, used your actual system-wide libxkbcommon install for this and not exhibit any of these problems by sidestepping all of that.
#asahi-dev · 2026-01-31 18:07
leioso this is strictly a problem with your configuration and local build
#asahi-dev · 2026-01-31 18:07
leioit configures it as such in its meson.build and all; this is why you see these #define's
#asahi-dev · 2026-01-31 18:07
leiolibxkbcommon will look for its files from where they are supposed to be in its knowledge. If you don't pass a -Dprefix (probably inherited from whatever you are actually building that grabs libxkbcommon as a subrpoject), then that will default to /usr/local, so of course it's going to be trying to load its files at runtime from the place it assumes you are going to be installing it - under /usr/local
#asahi-dev · 2026-01-31 18:05
bigmac(that's why I started it here)
#asahi-dev · 2026-01-31 18:05
bigmacI thought that the problem was on our side where we did not set a needed env var pointing to the place where we have the xkb configs
#asahi-dev · 2026-01-31 18:04
bigmacYes it totally is atp. I thought it was relevant to this room.
#asahi-dev · 2026-01-31 18:04
bigmachttps://paste.centos.org/view/bfdc9aee never had this happen before, but maybe it is common and I just haven't came across such issues before
#asahi-dev · 2026-01-31 18:04
leioI don't know what you are building that has the problem (something using wlroots?), also I doubt this is the correct venue to discuss this
#asahi-dev · 2026-01-31 18:01
bigmacleio: I do not indent to install it system wide and setting the mentioned env var solves it for me, but I do not recall having such difficulties building on other systems, so I am curious where the hurdle comes from. They either have it in the default location /usr/local or have an env set by default that I did not
#asahi-dev · 2026-01-31 17:55
leioThey do not, it's probably because you built it without specifying a prefix for the install, then it defaults to /usr/local instead of /usr, and the default lookup paths are coded to be in the same prefix as the install of it, but then you also run it locally perhaps. If you don't install it anyways to system (do you?), maybe just configuring with /usr as install prefix is sufficient.
#asahi-dev · 2026-01-31 17:27
bigmacwelp, this kinda sucks as it took me quite some time to figure out. Can we do anything about it? Do other systems have it in usr/local?
#asahi-dev · 2026-01-31 17:26
bigmac#define DFLT_XKB_LEGACY_ROOT "/usr/local/share/X11/xkb"
#asahi-dev · 2026-01-31 17:26
bigmac#define DFLT_XKB_CONFIG_ROOT "/usr/local/share/X11/xkb"
#asahi-dev · 2026-01-31 17:26
bigmacAnd I might have found the place where it's being set too cat ./subprojects/wlroots/build/subprojects/libxkbcommon/config.h | grep ROOT
#asahi-dev · 2026-01-31 17:25
bigmacHm, why is it looking into /usr/local then. Sure, when I set XKB_CONFIG_ROOT="/usr/share/X11/xkb" everything begins to work. It defaulted to the wrong location, I am trying to figure out why now..
#asahi-dev · 2026-01-31 17:00
leiobigmac: /usr/local is not something we'd have unless we make it ourselves. Stuff in the /usr/share variant comes from xkeyboard-config typically. Not sure what env vars you are referencing, but various XDG_* ones are supposed to be treated as a default value existing to a certain specified one by any consumers when the env var is not set
#asahi-dev · 2026-01-31 16:54
bigmachttps://xkbcommon.org/doc/current/group__include-path.html none of these env vars are set on my KDE install, so I am wondering whether I erased something by an accident
#asahi-dev · 2026-01-31 16:53
bigmac /usr/local/share/X11/xkb, do you have this directory on your asahi systems?
#asahi-dev · 2026-01-31 14:15
alyssaintegralpilot: yeah the intention of the r/e policy is specifically that you do that (disassembly.py on your own msl shaders) instead of popping their llvm backend into ghidra or whatever.
#asahi-dev · 2026-01-31 12:54
JoshuaPeisach[mds]Going to do a local rebuild for my J316, any branch specifically that I should use? Maybe a M3 WIP branch to test for regression?
#asahi-dev · 2026-01-31 11:48
integralpilotjust putting up WIP progress as it's in draft mode, thank you for all the detailed comments you left, will apply them!
#asahi-dev · 2026-01-31 11:48
integralpilotand there's still more alignment with dt-bindings to do
#asahi-dev · 2026-01-31 11:47
integralpilotyeah, haven't gotten around to extracting the new info yet, that's why i didn't add it to the PR yet
#asahi-dev · 2026-01-31 11:13
jannauthe latter is set by m1n1 but I think we have it anyway in the dts
#asahi-dev · 2026-01-31 11:10
jannauboard-type shouldn't be copied from t8112. board-type is module-instance in the adt and apple,antenna-sku is wifi-antenna-sku-info in the adt
#asahi-dev · 2026-01-31 11:05
jannauI'm missing wifi0/bluetooth0 nodes with apple,antenna-sku, brcm,board-type and local-mac-address/local-bd-address properties at least in the PR
#asahi-dev · 2026-01-31 11:02
noopwafelThe wifi didn't work for me, I assumed because I didn't add the wifi0 tree. I will take a look, but I can also just add a commit (as I mentioned I just copied from t8112 and it seemed fine).
#asahi-dev · 2026-01-31 10:04
noopwafelintegralpilot: thanks :)
#asahi-dev · 2026-01-31 09:02
integralpilotmade some decoding progress through lots of testing. documented what i know so far in -gpu
#asahi-dev · 2026-01-31 08:16
integralpilotthanks! :)
#asahi-dev · 2026-01-31 08:13
chadmed-re is for the ghidra type stuff that some folks may not want to be exposed to for legal reasons
#asahi-dev · 2026-01-31 08:12
chadmed-gpu is probably the place for this
#asahi-dev · 2026-01-31 08:12
integralpilotthank you both for confirming! i'm going to look into basic texture sampling now, will update if i find anything of value. does it need to go in #asahi-re or #asahi-gpu, or is here fine?
#asahi-dev · 2026-01-31 08:11
chadmedwell i guess this is still binary re in a way, but the output of apples shader compiler is your binary
#asahi-dev · 2026-01-31 08:10
chadmedall of those tools are cricket, only _binary_ re (as you say, decompiling apple software in ghidra) is not allowed
#asahi-dev · 2026-01-31 08:07
chaos_princessthat one is fine

plus ›