#
✕ nick:lina
#asahi-gpu · 2022-10-27 02:19
linaAnd that's why it looks like it works
#asahi-gpu · 2022-10-27 02:19
linaBut you're still faulting on intermediate flushes, before the TVB grows in Mesa, without the metadata buffer
#asahi-gpu · 2022-10-27 02:18
linaSo you probably found the flag to make Z uncompressed on *final store*, and I am setting that right
#asahi-gpu · 2022-10-27 02:18
linaSo only the intermediate flushes use the acceleration buffer!
#asahi-gpu · 2022-10-27 02:17
linaalyssa: The depth buffer *is* written unconditionally. However, it's written uncompressed!
#asahi-gpu · 2022-10-27 02:17
linaOhhh....
#asahi-gpu · 2022-10-27 02:17
linaOr... hold on.
#asahi-gpu · 2022-10-27 02:14
linaMind you, the mesa I'm using for python testing is very outdated right now, so if you fixed this recently, that might explain it. Or I got the depth flags wrong.
#asahi-gpu · 2022-10-27 02:13
linaIt sure is with my bunny framedumps...
#asahi-gpu · 2022-10-27 02:11
linaWith my bunny renders, the depth buffer is unwritten if I make the TVB large enough that it can render in a single pass, I only get data there if it ends up splitting due to a small TVB
#asahi-gpu · 2022-10-27 02:11
linaIf the app isn't actually reading out the depth buffer there is no need to flush it out to RAM
#asahi-gpu · 2022-10-27 02:10
linaAnd those go away after a few frames
#asahi-gpu · 2022-10-27 02:10
linaKeep in mind that the depth buffer, unless the app really needs it, is only written to on TVB overflows
#asahi-gpu · 2022-10-27 02:09
linaphire: It "works" but it also faults sometimes. I did get frames out of it without the patch, but the hypervisor showed GPU faults.
#asahi-gpu · 2022-10-26 15:49
linaStill need to look at the deflake buffer sizes... 2 of them are obvious due to adjacency, but do you remember how you figured out the third bound?
#asahi-gpu · 2022-10-26 15:48
linaIt seems every 8x4 block of Z pixels maps to one accel buffer byte, 0x03 means clear.
#asahi-gpu · 2022-10-26 15:48
linaLooks like it's 1/32 compression and it uses POT addressing, so align zbuffer size to POT and divide by 32 for the accel buffer size
#asahi-gpu · 2022-10-26 15:29
linaThe stencil one is just a guess though, haven't actually seen it yet
#asahi-gpu · 2022-10-26 15:25
linaLast 3 commits here: https://gitlab.freedesktop.org/asahilina/mesa/-/commits/asahi/wip
#asahi-gpu · 2022-10-26 15:22
linaThe size is just random though, haven't worked out how to calculate any of it yet.
#asahi-gpu · 2022-10-26 15:21
linaLet me do that!
#asahi-gpu · 2022-10-26 15:14
linaalyssa: I've never seen it enabled on the M1 Mini, ever. I have zero hits for those pointers in all my historical hypervisor logs. But on this one, it showed up, and mesa was faulting without it... so I'm not sure you disabled it ^^
#asahi-gpu · 2022-10-26 15:07
linaParallelizing fragment processing is trivial, but fragment is not, since it interacts with the tiler/sorting stuff
#asahi-gpu · 2022-10-26 15:06
linaI get the feeling they're actually trying to balance work within single jobs between dies, which would explain why they need some new buffers to transfer things around.
#asahi-gpu · 2022-10-26 15:05
linaThere's a Z acceleration/hierarchical Z thing that showed up, and then 5 new buffers adjacent to the tiler?
#asahi-gpu · 2022-10-26 15:05
linaSo the M1 Ultra just needed a bunch of new buffers and some larger bigger ones, initdata changes, and a couple constants changed, and then it worked.
#asahi-gpu · 2022-10-26 05:55
linaNice!! <3
#asahi-gpu · 2022-10-24 09:39
lina(Default is 0x10 so 2MiB)
#asahi-gpu · 2022-10-24 09:38
linaThe units are 128K buffer blocks
#asahi-gpu · 2022-10-24 09:37
linaI do get some improvement on terrain at small framebuffer sizes, but nothing to write home about, so I think most of the slowness there is just shader stuff, not TVB overflows
#asahi-gpu · 2022-10-24 09:37
linaJust added an initial_tvb_size module param (can be changed at runtime, only affects new GPU clients)
#asahi-gpu · 2022-10-24 09:26
linajannau: I think the crash should be fixed (turns out that tvb_something buffer does need to be dynamically sized based on resolution...)
#asahi-gpu · 2022-10-24 09:23
linaalyssa: I'm also pretty sure terrain is overflowing my fixed size TVB, possibly a lot, and I don't have any grow logic in yet...
#asahi-gpu · 2022-10-23 16:15
linaSo now it should be really obvious if you're doing bad texture/vertex/etc accesses (before it just silently ignored it)
#asahi-gpu · 2022-10-23 16:15
linaAlso just added forcing hard faults (on by default) and fault address/info decoding ^^
#asahi-gpu · 2022-10-23 09:04
linajannau: Now pushed!
#asahi-gpu · 2022-10-23 04:52
linaIt wasn't too bad, I just had to do cursed filter-branch sed games again because Rust-for-Linux renamed Ref<> to Arc<>
#asahi-gpu · 2022-10-23 04:51
linajannau: Just rebased on top of asahi-wip with your dcp branch merged! I'm doing it remote right now, so I'll push when I get home ^^
#asahi-gpu · 2022-10-22 23:10
linachain: Thank you!
#asahi-gpu · 2022-10-22 23:10
linaella-0: By the way, the new UAPI has the OpenGL depth clipping bit in there (there's a comment in mesa) ^^
#asahi-gpu · 2022-10-22 08:30
linajannau: It depends on the rust-for-linux merge. I haven't checked the upstream repo yet... if they already rebased on 6.1 then I want to do it, otherwise I'll wait for that
#asahi-gpu · 2022-10-20 14:21
linaphire: You mean t8103?
#asahi-gpu · 2022-10-18 02:17
linamarcan: Thank you!!
#asahi-gpu · 2022-10-13 17:01
lina(I have whole-TLB invals spammed everywhere in the MMU code right now)
#asahi-gpu · 2022-10-13 17:01
linaI think I'm invalidating in so many codepaths I'm effectively already invalidating twice... ^^
#asahi-gpu · 2022-10-13 16:04
linaYeah, I'm thinking either TLBs or virtually tagged caches clashing on remapped VAs or physically tagged caches clobbering unmapped memory with dirty data or both...
#asahi-gpu · 2022-10-13 15:59
linakarolherbst: The reason I think it's TLBs (or closely related caches) is all the problems go away if I never free/unmap anything, as far as I can tell
#asahi-gpu · 2022-10-13 15:15
linaSo not a firmware issue
#asahi-gpu · 2022-10-13 15:15
linakarolherbst: The problems lately have been GPU timeouts, which happen when the GPU side buffers are bad (they aren't faults so it's not crazy pointers, but just "something that wedges the GPU")
#asahi-gpu · 2022-10-13 15:13
linaAnd now I crashed kmscube, okay, so it's just not as bad as before.

plus ›