#
#asahi-gpu · 2022-11-19 19:23
alyssaneed to rebase asahi/main
#asahi-gpu · 2022-11-19 19:23
alyssaoh right yes
#asahi-gpu · 2022-11-19 19:20
jannauI think the patch works fine the Xorg problems I saw were unrelated
#asahi-gpu · 2022-11-19 19:19
jannaualyssa: that has the broken stride check in agx_resource_from_handle()
#asahi-gpu · 2022-11-19 19:12
alyssa95.2%
#asahi-gpu · 2022-11-19 19:12
alyssa^^ deqp-gles3 is getting there
#asahi-gpu · 2022-11-19 19:12
alyssaPass: 41976, Fail: 1533, Crash: 359, Warn: 87, Skip: 209, Missing: 225, Duration: 10:48, Remaining: 0
#asahi-gpu · 2022-11-19 19:11
alyssaEverything else can and should be considered known-broken until proven otherwise ;-)
#asahi-gpu · 2022-11-19 19:11
alyssayuka: The main branch of https://gitlab.freedesktop.org/asahi/mesa is the one to use :-)
#asahi-gpu · 2022-11-19 18:27
linayuka: It should be good right now, but it's not a long term branch or anything
#asahi-gpu · 2022-11-19 18:25
yukaor should I stick with asahi/wip?
#asahi-gpu · 2022-11-19 18:24
yukais asahi/batch-tracking-fixes a branch that should be usable?
#asahi-gpu · 2022-11-19 04:32
alyssajust.. what
#asahi-gpu · 2022-11-19 04:32
alyssa Passed: 939/940 (99.9%)
#asahi-gpu · 2022-11-19 04:32
alyssastill being taunted by dEQP-GLES3.functional.fragment_out.random.51
#asahi-gpu · 2022-11-19 01:29
alyssaback to linux 🐧
#asahi-gpu · 2022-11-19 01:29
alyssafor now anyway
#asahi-gpu · 2022-11-19 01:29
alyssaeither way I think I've seen enough
#asahi-gpu · 2022-11-19 01:26
alyssaunclear if that's an ISA constraint, or maybe it's an optimization?
#asahi-gpu · 2022-11-19 01:26
alyssait lowers sample_mask to zs_emit if zs_emit is used, I guess
#asahi-gpu · 2022-11-19 01:26
alyssanot sure i've ever seen both sample_mask and zs_emit used in the same shader
#asahi-gpu · 2022-11-19 01:25
alyssaif sample mask comes from the shader
#asahi-gpu · 2022-11-19 01:25
alyssabut at least one needs to be executed statically
#asahi-gpu · 2022-11-19 01:24
alyssaalso looks like multiple sample_mask writes are allowed, and they have the effect of masking out each other
#asahi-gpu · 2022-11-19 01:23
alyssawith MSAA + sample mask write + a stencil write, it uses zs_emit only
#asahi-gpu · 2022-11-19 01:21
alyssa> 2a: c17f0003 sample_mask 255, r0l
#asahi-gpu · 2022-11-19 01:21
alyssathe latter being as simple as
#asahi-gpu · 2022-11-19 01:21
alyssaoh so Metal does use that lowering for discard with MSAA and a sample mask write
#asahi-gpu · 2022-11-19 01:20
alyssa> 18: c1008000 sample_mask r0l, 0
#asahi-gpu · 2022-11-19 01:20
alyssa> a: 0280c002023f0110c002 fcmpsel eq, r0l, r0.discard, -0.5, 255, 0
#asahi-gpu · 2022-11-19 01:13
alyssaditto with MSAA
#asahi-gpu · 2022-11-19 01:09
alyssawith no MSAA
#asahi-gpu · 2022-11-19 01:09
alyssa> being ld tile + discard together
#asahi-gpu · 2022-11-19 01:08
alyssayou were right to revert the pass type commit then
#asahi-gpu · 2022-11-19 01:08
alyssalina: >.>
#asahi-gpu · 2022-11-19 01:08
alyssa> Pass type: unknown 3 (XXX)
#asahi-gpu · 2022-11-19 01:08
alyssa---
#asahi-gpu · 2022-11-19 01:08
alyssa< Pass type: Translucent
#asahi-gpu · 2022-11-17 23:33
alyssanini
#asahi-gpu · 2022-11-17 23:33
jannaugood night
#asahi-gpu · 2022-11-17 23:29
jannauno crash but terminiating plasmashell after "[destroyed object]: error 7: importing the supplied dmabufs failed"
#asahi-gpu · 2022-11-17 23:28
alyssacrash?
#asahi-gpu · 2022-11-17 23:28
alyssa?
#asahi-gpu · 2022-11-17 23:28
alyssa23:27 < jannau> without the stride patch it was impossible to get there
#asahi-gpu · 2022-11-17 23:27
jannauwithout the stride patch it was impossible to get there
#asahi-gpu · 2022-11-17 23:26
jannauit worked with the rebased branch
#asahi-gpu · 2022-11-17 23:26
jannauinstead the screen gets bright
#asahi-gpu · 2022-11-17 23:26
alyssadid it before?
#asahi-gpu · 2022-11-17 23:25
jannauthe log out/restart/shutdown dialog doesn't render
#asahi-gpu · 2022-11-17 23:23
alyssablink

plus ›