#
#asahi-gpu · 2023-03-03 15:42
alyssaand seeing that all render targets were correct except the very last
#asahi-gpu · 2023-03-03 15:42
alyssathe way I got there was by looking at the results.xml for the failing test
#asahi-gpu · 2023-03-03 15:41
alyssa:\
#asahi-gpu · 2023-03-03 15:41
linaI'm not sure how long it's been but it's been a while of digging through render target code to get to offsets > 64 (which I got to just now) ^^;;
#asahi-gpu · 2023-03-03 15:40
alyssai'm sorry to hear that
#asahi-gpu · 2023-03-03 15:40
alyssaoh, um
#asahi-gpu · 2023-03-03 15:40
linaI mean I spent that long getting to where you were ;;
#asahi-gpu · 2023-03-03 15:40
alyssaif I had figured it out, I would've fixed it ;)
#asahi-gpu · 2023-03-03 15:38
linaI do feel like sometimes I spend hours figuring out stuff you've already figured out ;;
#asahi-gpu · 2023-03-03 15:37
alyssaAwesome :)
#asahi-gpu · 2023-03-03 15:37
linaI think I spent the past 2-3 hours getting there...
#asahi-gpu · 2023-03-03 15:37
linaWell either way at least for GLES3 I'm pretty sure I don't have any sync-related regressions... it's that, some depth/stencil stuff, and some pack/unpack stuff. All of which sound completely unrelated to sync.
#asahi-gpu · 2023-03-03 15:37
alyssastill grateful for help debugging them because I got as far as "yeah tib offsets >= 64 broken ???? why ????"
#asahi-gpu · 2023-03-03 15:36
linaOh...........
#asahi-gpu · 2023-03-03 15:36
alyssaand I haven't uprevved deqp in a long time
#asahi-gpu · 2023-03-03 15:36
alyssaThose tests used to be in dEQP-GLES31 but then Khronos realized "Wait there's literally nothing in here that depends on GLES31" so they moved it to GLES3
#asahi-gpu · 2023-03-03 15:36
alyssaoh. yeah. lmao
#asahi-gpu · 2023-03-03 15:36
linaIt wasn't on your existing GLES3 fail list... but I see it is in the GLES31 one...
#asahi-gpu · 2023-03-03 15:35
alyssai mean you're welcome to debug it but it's not a blocker
#asahi-gpu · 2023-03-03 15:35
alyssalina: That's an existing fail, not a regression
#asahi-gpu · 2023-03-03 15:34
linaalyssa: I'm stuck on max draw buffers right now, TIB offsets >= 64 do not work. I'm going to macOS to see if I'm missing something here, even kernel side...
#asahi-gpu · 2023-03-03 15:31
alyssaI use --deqp-surface-type=pbuffer, same as CI, fwiw
#asahi-gpu · 2023-03-03 15:30
alyssamaybe the colour masking opts are still not quite right
#asahi-gpu · 2023-03-03 15:30
alyssalina: dEQP-GLES2.functional.fragment_ops.interaction.basic_shader.77 could easily be an upstream regression, yeah
#asahi-gpu · 2023-03-03 12:53
linaSooo --deqp-surface-type=fbo is very, very broken... ;;
#asahi-gpu · 2023-03-03 12:30
linaAnd now I'm running into others... and I'm getting the feeling none of this has anything to do with explicit sync and it's just all very broken and I don't get it...
#asahi-gpu · 2023-03-03 12:30
linaOK so I fixed that one with a state tracker patch... but I still don't know why this isn't breaking other drivers...
#asahi-gpu · 2023-03-03 12:15
linaalyssa: Ummm so I'm running into ASAN issues with constant buffers and blitting and I can't figure out why this isn't broken for everyone...
#asahi-gpu · 2023-03-03 10:00
linaIt turns out you can render to a BO and *then* export it and so we need to post-facto insert a DMA-BUF fence into it at export time...
#asahi-gpu · 2023-03-03 09:59
linaalyssa: I can't repro the pink square thing with stk but I can with nautilus. Fixed ^^
#asahi-gpu · 2023-03-03 09:08
linaI don't recall ever seeing this back when I tested dEQP-GLES2 with the explicit sync stuff... could it be an upstream mesa regression recently? I tried just UAPI on top of mesa/mesa:main and it still happens...
#asahi-gpu · 2023-03-03 09:07
linaalyssa: I have a flake regression in dEQP-GLES2 and I don't know where it comes from... it's easy to reproduce running dEQP-GLES2.functional.fragment_ops.interaction.basic_shader.77 in a loop. The images have random bad pixels... any ideas? It still happens in sync mode, and also with the flush_resource stuff disabled...
#asahi-gpu · 2023-03-03 08:05
linaOkay, the flush_resource thing was completely broken... Maybe that's what was breaking Sway... ^^;;
#asahi-gpu · 2023-03-03 07:14
linaAnd at least some are due to the flush_resource stuff
#asahi-gpu · 2023-03-03 07:14
linadEQP-EGL was all kinds of broken now...
#asahi-gpu · 2023-03-03 07:14
linaalyssa: So I think a lot of the regressions are actually from the git work/rebase for upstreaming the device stuff ^^;; (fd leaks...)
#asahi-gpu · 2023-03-03 05:58
lina^^
#asahi-gpu · 2023-03-03 05:58
alyssaand o look it's your stream o'clock which means it's my sleep o'clock
#asahi-gpu · 2023-03-03 05:58
alyssa:+1:
#asahi-gpu · 2023-03-03 05:57
linaI don't think we want to do anything fancy with IRQs delivered to userspace... the equivalent of that with the existing UAPI is to just have a thread poll on syncobjs and do the cleanup there, but that sounds like a "might be a good idea some day, probably not now" thing...
#asahi-gpu · 2023-03-03 05:56
linaAnd I'll do something to make batches not roll around all the time. I think I can just do a poll on all submitted syncobjs twice after each submission (you can do several in one syscall, doing it twice means we potentially clean up 2 batches for each new batch, which should keep memory usage under control)
#asahi-gpu · 2023-03-03 05:55
linathe flush_resource thing
#asahi-gpu · 2023-03-03 05:55
linaSo I'm going to move the SHAREABLE fix thing to before the batch tracking because that was actually regressed (well, it was always technically wrong, it just started mattering) by the UAPI change, not the actual explicit sync
#asahi-gpu · 2023-03-03 05:54
alyssaand we can't release if the driver is like this. what we're shipping now was rock solid. if we can't even run basic test suites anymore without crashing, it's almost guaranteed that something will regress for a user if we push it out like this.
#asahi-gpu · 2023-03-03 05:52
alyssabut this was working before at any rate
#asahi-gpu · 2023-03-03 05:52
alyssaunsure if it's related to sync. it's kinda hard to bisect over hard uapi changes.
#asahi-gpu · 2023-03-03 05:51
alyssanow, as in, this is a regression
#asahi-gpu · 2023-03-03 05:51
linaIt's not a hard leak but I don't know how much memory it's *supposed* to use...
#asahi-gpu · 2023-03-03 05:51
linaOh yeah, deqp-EGL hits the OOM killer on 8GB with >1 thread (or maybe 2 fit, barely), I don't know why...
#asahi-gpu · 2023-03-03 05:50
alyssaYeah. deqp-gles31 is going down in under a minute (8GB mini)

plus ›