#asahi-dev · 2025-06-02 19:22
kettenishah, building with -Z build-std doesn't have rust/target ;)#asahi-dev · 2025-06-02 19:21
amarioguyah got it - it's not a major rush or anything, just makes it quite annoying to find what my actual changes are#asahi-dev · 2025-06-02 19:21
chaos_princessbecause nobody added it :P iirc i've added it there as a part of a not-yet-merged pr#asahi-dev · 2025-06-02 19:19
amarioguyprobably part of building the rust side, which makes sense, i guess my real question is why is rust/target not in gitignore :P#asahi-dev · 2025-06-02 19:18
amarioguybtw does anyone know why after an m1n1 build rust is leaving behind a lot of target stuff that isn't in the gitignore?#asahi-dev · 2025-06-02 19:00
svenADT had many weird hacks and cannot be trusted but at least address/size cells have been consistent everywhere I looked fwiw #asahi-dev · 2025-06-02 18:19
mixifl0_id: "#address-cells = 1", "#size-cells = 0" isn't unusual. its used to describe a bus where each device is identified by a single identifier (e.g. the id used to address a single device on the bus, numbered subdevices behind one parent device such as cpu cores, network ports) rather than a range of identifiers (e.g. memory addresses)#asahi-dev · 2025-06-02 16:22
amarioguyshouldn't matter#asahi-dev · 2025-06-02 16:22
fl0_iddoes the firmware version matter?#asahi-dev · 2025-06-02 16:21
fl0_idamarioguy well yeah that's good enough#asahi-dev · 2025-06-02 16:21
amarioguynot really smth that needs a script - i just need someone to run mrs(ICC_CTLR_EL1, call="el1") in the m1n1 proxyclient console#asahi-dev · 2025-06-02 16:20
amarioguytest for m3 max#asahi-dev · 2025-06-02 16:20
amarioguyfl0_id: i saw it#asahi-dev · 2025-06-02 16:01
fl0_idfl0_id thx. I guess the mistake was copying just the relevant files separately then#asahi-dev · 2025-06-02 15:57
chaos_princessyou are supposed to run `make dtbs_check` and it should do the thing#asahi-dev · 2025-06-02 15:56
fl0_idfor testing syntac compatibility of device trees, is there a good way to get better errors out of the compiler / should it be possible to compile outside of the kernel tree (with includes added into the relevant directory)? I get syntax errors quite at the beginning, but do not really see an error when compring it to others.#asahi-dev · 2025-06-02 15:27
ahdeb[mds]Unless I'm mistaken, this leaves a window where there's no boot.bin... would feel safer if first mv was instead cp... (the second is already mv -f so will overwrite): https://github.com/AsahiLinux/asahi-scripts/commit/3f13867d2147ff5c7cdc14dc404c27caabe899ec#asahi-dev · 2025-06-02 12:22
fl0_idchadmed in some other device trees it's the same, so it seems reasonable#asahi-dev · 2025-06-02 11:20
fl0_idsame haha#asahi-dev · 2025-06-02 11:18
chadmedthat seems sus to me but idk all the idiosyncracies of the ADT#asahi-dev · 2025-06-02 11:14
fl0_id newb question, in my m3 max device tree dump there is "#address-cells = 1 \n #size-cells = 0" - is that possible or does it mean that the dump is wrong?#asahi-dev · 2025-06-02 11:00
chadmedbut we have the gpu init stuff too so it will have to be soon-ish...#asahi-dev · 2025-06-02 10:59
chadmedyeah, any compulsion of rust will have to be a 2.0 release too#asahi-dev · 2025-06-02 10:58
kettenisand revisit if/when build-std becomes stable#asahi-dev · 2025-06-02 10:58
kettenischadmed: thanks, sounds like we should go with what I have now then#asahi-dev · 2025-06-02 07:13
fl0_idif amarioguy comes back, you can tell him I could test for m3 max, but ideally he would provide like a python script to run in proxy mode or similar...#asahi-dev · 2025-06-02 06:33
chadmedkettenis: id be in favour of making build-std the default once it's a stable feature#asahi-dev · 2025-06-02 01:40
amarioguynote that i think that register has to be read from EL1#asahi-dev · 2025-06-02 01:39
amarioguyneed to verify if the vGIC cpu interfaces gained support for extended SPI/PPI ranges (M1 and M2 along with Pro counterparts say no on that front, i want to verify if it's also true on the later chips, since for some of those core counts on the newer ones, seem to be pushing against the limit where those ranges would be needed)#asahi-dev · 2025-06-02 01:37
amarioguycan someone check the ICC_CTLR_EL1 on any M4 Pro models? (same for M3 and m3 pro if you have either)#asahi-dev · 2025-06-01 18:27
kettenisI marked it as draft as the other option is to just always build m1n1 with -Z build-std#asahi-dev · 2025-06-01 18:26
kettenisbtw, https://github.com/AsahiLinux/m1n1/pull/471#asahi-dev · 2025-06-01 17:23
svenif someone's looking for a yak to shave: we should get rid of all the "default ARCH_APPLE" that snuck into various Kconfig files and instead move them with m (and only y if required) to arm64 defconfig#asahi-dev · 2025-06-01 14:50
svenno. those names more or less come from xnus debug output so your guess is as good as mine #asahi-dev · 2025-06-01 14:49
nickchansven: do you know what USB2PHY_CTL_APB_RESET_N in atc.c is in more detail?#asahi-dev · 2025-06-01 12:53
nickchanugh I hit enter on accident#asahi-dev · 2025-06-01 12:52
nickchanafaik it's more #asahi-dev · 2025-06-01 12:32
svenhttps://git.kernel.org/pub/scm/linux/kernel/git/sven/linux.git/log/?h=wip/atcphy mostly happy with it now except for commit descriptions#asahi-dev · 2025-06-01 11:08
sven(either that or people have a bootrom exploit which is much less likely)#asahi-dev · 2025-06-01 11:08
svenwe actually know that these must've escaped containment because that's usually where those iboot keys come from#asahi-dev · 2025-06-01 11:07
svenprobably, but i really don't want touch those devices#asahi-dev · 2025-06-01 11:07
chadmedsurely one or two have fallen off the back of a truck in transit :P #asahi-dev · 2025-06-01 11:07
chadmeddamn it#asahi-dev · 2025-06-01 11:06
svenit only works on debug fused development SoCs#asahi-dev · 2025-06-01 11:05
svenoh, you get swd via usb ports. it's just completely useless#asahi-dev · 2025-06-01 11:05
chadmeddo we get jtag via the usb-c ports or do we have to start cracking machines open #asahi-dev · 2025-06-01 11:04
svennow I wish I actually had SWD. would be interesting to see what the other cores are up to just when the one core tries to jump to 0x20...#asahi-dev · 2025-06-01 11:03
chadmedneeds a bit of a cleanup though#asahi-dev · 2025-06-01 11:02
chadmedthere are a couple of gross things the C side does with memcpy and raw pointers that is going to be a pita to port but theyre the last two functions#asahi-dev · 2025-06-01 11:01
chadmedrust adt code now boots to linux and is almost done