#
✕ nick:lina
#asahi-gpu · 2023-03-29 04:08
linaFrom what alyssa told me, AGX/PowerVR does not support *fragment* shaders, it only supports pixel shaders, that's why it's called PDM ^^
#asahi-gpu · 2023-03-28 12:27
linaI did not! ^^
#asahi-gpu · 2023-03-28 10:43
linaI did see framebuffer-sized things...
#asahi-gpu · 2023-03-28 10:23
linajannau: we map to DCP anything that "could be" a framebuffer (to make kmsro work), so it can have stuff that has nothing to do with framebuffers... though we might be too eager right now. I'll take a look tomorrow ^^
#asahi-gpu · 2023-03-27 14:23
linaAlso I tried to reproduce the issue and I ended up crashing the GPU firmware instead... that's not good...
#asahi-gpu · 2023-03-27 14:23
linaAh yeah, I doubt kwin uses that
#asahi-gpu · 2023-03-27 14:03
linaI haven't looked at the KMS side API but if it supports both an explicit fence as well as implicit buffer fences, I would expect only implicit ones since that's how the usual WSI stuff works so far as far as I know
#asahi-gpu · 2023-03-27 14:02
linaI meant like no fence attached at all
#asahi-gpu · 2023-03-26 15:16
linaGEM_BIND should never fail unless an invariant was violated (debug build should catch that) or something went very weirdly wrong with VA tracking or memory exhaustion or something...
#asahi-gpu · 2023-03-26 15:15
linaBut I suspect there is a bad codepath somewhere around that
#asahi-gpu · 2023-03-26 15:15
linajannau: It is expected that the FB has no explicit fence if and only if the render was issued before export, actually completed, and then the buffer was exported when it was known to be idle.
#asahi-gpu · 2023-03-25 06:16
linaTF2 does not work because there is a general problem with Steam crashing when games connect to it, and therefore games with a hard Steam requirement don't work
#asahi-gpu · 2023-03-24 13:41
linaalyssa: Portal/Portal2/HL2 are smooth now, and even Crysis runs too though with some glitches. 34 tested working games now ^^ Going to get dinner and actually send the MRs after!
#asahi-gpu · 2023-03-22 18:31
linaIt probably does have that dependency though, but I've never seen it fail due to that
#asahi-gpu · 2023-03-22 18:31
linaThat's unrelated, the problem I actually ran into was pmp getting shut down early and then the DCP power domains failing to probe...
#asahi-gpu · 2023-03-22 18:28
linaIt goes in the build dir
#asahi-gpu · 2023-03-22 18:25
linajannau: It's not that, it's the power domain probe order thing. It was breaking DCP...
#asahi-gpu · 2023-03-22 18:15
linaalyssa: Total 24 games tested working today! ^^
#asahi-gpu · 2023-03-22 18:14
linaLet me get dinner first ^^;;
#asahi-gpu · 2023-03-22 18:12
linaMesa in asahilina/mesa:uapi-32bit-buildfix ^^
#asahi-gpu · 2023-03-22 18:11
linaalyssa: Kernel in gpu/rust-wip as usual, you might need to disable CONFIG_HUGETLBFS and CONFIG_TRANSPARENT_HUGEPAGE if enabled ^^
#asahi-gpu · 2023-03-22 08:15
linaalyssa: Hitting assert(bary->intrinsic == nir_intrinsic_load_barycentric_pixel); on Portal
#asahi-gpu · 2023-03-21 18:39
lina^^
#asahi-gpu · 2023-03-21 18:39
linaalyssa: Signed-off-by: Asahi Lina <lina@asahilina.net>
#asahi-gpu · 2023-03-21 18:12
linaalyssa: Ah yes, sounds good to me! Do you want a signoff here or on the patch?
#asahi-gpu · 2023-03-21 16:48
linaalyssa: I already did that ^^
#asahi-gpu · 2023-03-21 15:07
linaalyssa: What do you want me to do about changes to the UAPI integration commit? Open a PR or just fix it and force push directly? (I have some 32bit build fixes and I think they're all in that code ^^)
#asahi-gpu · 2023-03-20 18:34
linaSo yes, nouveau, panfrost, freedreno, etc. should all say "Mesa" now
#asahi-gpu · 2023-03-20 18:34
linaIt used to say "Asahi" I think
#asahi-gpu · 2023-03-20 18:34
linaI think there was a patch recently to unify all community-supported Mesa drivers and make them all report "Mesa" as the vendor.
#asahi-gpu · 2023-03-15 14:20
linaThe latest kernel branch is gpu/rust-wip (back to the usual one), and asahilina:agx/release-rebased-uapi has our release branch rebased on the updated uapi MR branch plus fixes ^^
#asahi-gpu · 2023-03-15 14:19
linaalyssa: UAPI MR comments should be done, the kernel oopses were another drm_scheduler bug (^^;;), and the firmware crashes seem to be related to having faults while running compute jobs, but it doesn't always happen and I haven't been able to identify a reliable trigger or find anything wrong on my side that could cause it. It could be a firmware bug/race...
#asahi-gpu · 2023-03-15 07:50
linaI think the kernel oopses were a drm_scheduler core bug I just found... ^^;;
#asahi-gpu · 2023-03-14 15:13
linaI know several people have hit it but not how...
#asahi-gpu · 2023-03-14 15:12
linastill not sure if I'll be able to repro the crashes... I'm just hoping kasan catches something obvious
#asahi-gpu · 2023-03-14 06:31
linaStill, I bet there is something wrong somewhere around this area...
#asahi-gpu · 2023-03-14 06:30
linaAh but import should do that...
#asahi-gpu · 2023-03-14 06:30
linaI think we don't set the shareable flag at all in the kmsro path... because it exits early...
#asahi-gpu · 2023-03-14 06:30
linaOh no, I just found it...
#asahi-gpu · 2023-03-14 06:28
linaMaybe that codepath ends up not hitting the post-facto export stuff I added, or kmsro somehow confuses everything...
#asahi-gpu · 2023-03-14 06:28
linaI need to look at what it does to export the first framebuffer for the display controller
#asahi-gpu · 2023-03-14 06:28
linaI wonder if it's a thing with kwin startup... I might have an idea
#asahi-gpu · 2023-03-14 05:33
linaI'm going to look at that tomorrow
#asahi-gpu · 2023-03-14 05:33
linaIf anyone has a reliable reproducer for the pink squares stuff or kernel oopses please tell me ^^
#asahi-gpu · 2023-03-14 05:32
lina^^;;
#asahi-gpu · 2023-03-13 13:48
linaPhew! I thought I must've broken the PM code again ^^;;
#asahi-gpu · 2023-03-13 12:12
linachadmed: It's 1, 0 is effectively off
#asahi-gpu · 2023-03-12 15:31
linaalyssa: Sure! ^^
#asahi-gpu · 2023-03-12 02:35
linaThanks ^^
#asahi-gpu · 2023-03-12 02:34
linaalyssa: Which tests?

plus ›