#asahi-gpu · 2026-08-24 16:21
waltmckI'm not using muvm but I don't think that should make a difference (virgl should inherit the error from the host driver)#asahi-gpu · 2026-08-24 16:20
waltmckOne thing I'm kind of confused about is that there is a report on reddit (https://old.reddit.com/r/AsahiLinux/comments/1ll30ey/controller_cant_be_used_in_steam_games/) of getting NMS running. So either that is incorrect or they hardcoded 16-bit query addresses in some update in the last year#asahi-gpu · 2026-08-24 16:18
chaos_princessactually, it may be the instruction displacement limit? If it is, you will get errors when compiling shaders i think#asahi-gpu · 2026-08-24 16:17
chaos_princessi haven't seen an obvious reason (very much does not mean that there isn't one), so yea, why not try it and see what happens#asahi-gpu · 2026-08-24 16:16
waltmckYeah I'm tryng that right now (rebuilding mesa). I suspect that there must be some reason for this limit (maybe the hardware only supports 15 bit identifiers or something) in which case there might be issues when the game tries to use more than 32768 queries, but I'll test it empirically#asahi-gpu · 2026-08-24 16:14
chaos_princessstupid question - what if you just rebuild the driver with the limit raised?#asahi-gpu · 2026-08-24 16:11
waltmckHave you gotten No Man's Sky to run? For me (under WINE/FEX) it calls `vkCreateQueryPool(queryType = VK_QUERY_TYPE_OCCLUSION, queryCount = 65536)` which fails with `VK_ERROR_OUT_OF_DEVICE_MEMORY`. This makes sense because it looks like on honeykrisp `AGX_MAX_OCCLUSION_QUERIES = 32768`.#asahi-gpu · 2026-08-23 19:55
oliverYea, iirc it did not really make a distinction there, it was "only subtitles" in raster order. So only 16x16px tiles in raster order.#asahi-gpu · 2026-08-23 06:22
chadmedoh i think i got the description of interchange wrong in the dcp branch. are the tiles _and_ subtiles/pixel data are both raster order? the difference from AGX tiling confused me#asahi-gpu · 2026-08-23 04:51
oliverchadmed: awesome! #asahi-gpu · 2026-08-22 12:32
chadmedbut yes there is no enablement work for new hardware here, this is for direct scanout support on machines we already support#asahi-gpu · 2026-08-22 12:32
chadmedi presume this will continue working on m3 and up#asahi-gpu · 2026-08-22 12:28
Sobek[m]That’s only M1/M2, right ?#asahi-gpu · 2026-08-22 11:57
chadmed(fyi for oliver too)#asahi-gpu · 2026-08-22 11:57
chadmedthis branch doesnt have the hacks in it and will only use interchange where we expect it to be used#asahi-gpu · 2026-08-22 11:56
chadmedim not 100% if the vulkan stuff is done correctly or if there's a better way to make that sane#asahi-gpu · 2026-08-22 11:56
chadmedhttps://gitlab.freedesktop.org/chadmed/mesa/-/tree/asahi/interchange #asahi-gpu · 2026-08-22 11:56
chadmedwith that, there are no regressions in vk or gles/gl cts vs main#asahi-gpu · 2026-08-22 11:55
chadmedokay with interchange hacked in as the driver's default/preferred format when twiddling and compression are allowed, the only thing that seems to be 100% broken are formats with blocksize > 4 and z/s buffers but i've banned array and compressed images too for safety#asahi-gpu · 2026-08-20 19:41
ArmandoLara[m]Update: filed the Mesa issue for the vkGetQueryPoolResults fault, with sanitised logs from both 26.0.3 and 26.1.7: https://gitlab.freedesktop.org/mesa/mesa/-/work_items/16145 — happy to test patches or debug flags there.#asahi-gpu · 2026-08-20 18:53
ArmandoLara[m]Hi — Honeykrisp crash report from an M1 Pro (Fedora Asahi Remix 44), reproduced on two driver versions, plus a FEX rootfs question.... (full message at <https://matrix.org/oftc/media/v1/media/download/AWEnIRSgRfk0Lw6p9tVx9NVAY0pl2M8T4UnwkgjJWFuTD3Yx-0_9FDcZqzz-xBCtk5i836HvCMjrmdRNzO6wUDxCegaJ5pugAG1hdHJpeC5vcmcvSnlUSkpURVpmZUFla3dKeWZSYlJBdWtX>)#asahi-gpu · 2026-08-20 10:08
chadmedso the condition is ail_can_compress && !util_is_depth_or_stencil#asahi-gpu · 2026-08-20 10:08
chadmedi should clarify that the above means in addition to the limitations already in place for compressible/twiddled images#asahi-gpu · 2026-08-20 10:02
chadmedquestion is do we still want to limit it to _just_ scanout/shared buffers (i think we do/should)#asahi-gpu · 2026-08-20 10:01
chadmedseems like the only thing interchange cant do is depth/stencil#asahi-gpu · 2026-08-16 10:12
dtwoodHi all, I'm trying to get the alpha build of the game Kitten Space Agency to run under Wine on an M1 MacBook Pro, and I'm in a bit over my head with debugging Mesa. What I'm seeing at the moment is one of two crashes (depending on... (full message at <https://matrix.org/oftc/media/v1/media/download/AYcEL7z95EPKo_u9TOnNmlHZJFU90d4CvCeDn524VxSwhZXiFTBGCLa6-sMymUyayqFBRnZp2g627vnS7HqtdWNCegUig_BgAG1hdHJpeC5vcmcvTW5QaFVHZEp1Vmp2anNjWGNldnV6cVdS>)#asahi-gpu · 2026-08-11 11:23
karolherbstthey apparently get unique IP addresses even if not registered or anything#asahi-gpu · 2026-08-11 11:11
jannauI ignored this one since they came via the matrix bridge#asahi-gpu · 2026-08-11 10:58
karolherbstplease also report spammers in #oftc (I already did) then those can be banned network wide (they also spammed e.g. #pipewire)#asahi-gpu · 2026-08-11 08:57
BrandiWhite[m]https://t.me/+pa-CiKYv9-5lNWM8#asahi-gpu · 2026-08-11 08:57
BrandiWhite[m]Get in touch with this platform for greatness you’ll definitely thank me later
ℹ️❤️
👇🏻👇🏻👇🏻#asahi-gpu · 2026-08-06 12:12
pac85Makes sense #asahi-gpu · 2026-08-06 10:59
noopwafelPorting all the initdata stuff back to rust (in m1n1) is the sensible next step on the driver side.. a lot of churn but mostly testable on M1/M2.#asahi-gpu · 2026-08-06 10:49
pac85noopwafel: ah ok. I'll try getting the compiler going then when I find time #asahi-gpu · 2026-08-06 10:44
noopwafelpac85: yes it's definitely time to make it work. adapting the compiler would be really great honestly. I don't think there's (fingers crossed) too much left to learn on the kernel side.\#asahi-gpu · 2026-08-02 16:20
aviv_beThanks a bunch#asahi-gpu · 2026-08-02 16:17
chaos_princesshttps://alyssarosenzweig.ca/blog/asahi-gpu-part-1.html - unfortunately not directly linked from the main website, so adjust the number in the url from 1 to 5#asahi-gpu · 2026-08-02 15:48
aviv_beWhat is the gpu reverse engineering process like? I always wondered how do you reverse engineer this kind of thing. #asahi-gpu · 2026-08-02 12:09
pac85noopwafel: does it sound good or did you have a preference for having a standalone triangle first so you can explore the kernel more. I guess we can either explore breadth first and finish the gfx re or go depth first and get compute fully working (like, moving towards the goal of having compute fully working on Mesa and kernel)#asahi-gpu · 2026-08-02 12:07
pac85And for now I'd get it up and running for pure compute stuff #asahi-gpu · 2026-08-02 12:07
pac85So, as we find more stuff out, we can just add support for emitting it #asahi-gpu · 2026-08-02 12:06
pac85I was wondering if it would make sense to start adapting the compiler #asahi-gpu · 2026-08-02 12:05
pac85sven: lol, executive dysfunction doesn't help with this #asahi-gpu · 2026-08-02 12:04
sven*that#asahi-gpu · 2026-08-02 12:04
pac85I had the whole fragment/vertex special instructions mostly figured out but I have very few memories #asahi-gpu · 2026-08-02 12:04
sventhat’s sounds familiar :D#asahi-gpu · 2026-08-02 12:03
pac85Trying to pick the M3 re project back up and I really regret not having taken more notes :(#asahi-gpu · 2026-07-31 16:18
jannauhttps://gitlab.freedesktop.org/noopwafel/mesa/-/commit/7c0dc5067f883649716404d7396cdf5f79a7c354 and I need to remove the fault_control parameter from the kernel driver#asahi-gpu · 2026-07-31 16:01
Sobek[m]De rien !#asahi-gpu · 2026-07-31 15:53
alyssaThanks haha