10:30 danvet: j4ni, yeah
10:30 danvet: if there is anything
10:30 danvet: or did I miss it?
10:33 j4ni: I'll put it together, there's only a few fixes
13:48 danvet: agd5f, + * OK to resume normal operation. Use completion to allow
13:48 danvet: + * halted scsi ops to resume.
13:48 danvet: amdgpu now also a scsi driver? :-)
14:10 danvet: airlied, got the qxl patch, pimped commit message to taste, and applied
14:40 sumits: danvet: sent a merge-req for the maintainers-tool doc update in gitlab
14:40 danvet: sumits, it's patches
14:41 sumits: danvet - ack. sending
14:41 danvet: well I just hit merge
14:41 danvet: hitting that button is pretty quick :-)
14:42 danvet: since I can just look for the green circle and all that
15:04 valged3: you know guys, hardware gurus have done everything right, it's the developers who misunderstand stuff, telling how harddrive bottlenecks compute pipeline, while it is entirely sure that magnetic harddrive manufacturers know that this is not the case, otherwise they would not have had correctly modeled latencies and block counts and sizes of the elements etc. 0.2ms worst case head seek is very plauisible number when you look at at how magnet works, it is
15:04 valged3: 1/5th of a second, this is absolute bullet speed when you order things right which you by far have not done.
15:07 valged3: when i hear stories from my friends, the car manufacturing industry uses similar technologies but very strong robitic magnet movable arms. For instance mercedes benz and such, magneto based technology will never leave the scenes while we still can exist.
15:08 valged3: it more like starts to pick up the real momentum in mixed mode designs as i told, just a CVD machine needed to author those.
15:29 ajax: _mesa_is_winsys_fbo will return true for incomplete framebuffers
15:29 ajax: this seems suboptimal
16:04 zmike: mareko: what tests were you using to verify your inline uniforms feature?
16:21 ajax: why is fbconfig setup so horribly broken and stupid
16:26 Vanfanel: Hello. I am seeing these three lines when eglMakeCurrent() is called by the second time. Any idea of what they mean? There are no errors or fails in dmesg when the 2nd eglMakeCurrent() call occurs:
16:26 Vanfanel: mesa: for the --simplifycfg-sink-common option: may only occur zero or one times!
16:26 Vanfanel: mesa: for the --global-isel-abort option: may only occur zero or one times!
16:26 Vanfanel: mesa: for the --amdgpu-atomic-optimizations option: may only occur zero or one times!
16:26 ajax: Vanfanel: internal mesa issue. that's the internal llvm engine initializing itself twice, which shouldn't be happening
16:26 Vanfanel: ...It must be something in between, context is destroyed and rebuilt, etc. But no errors so no way for me to know.
16:27 Vanfanel: ajax: so it's nothing I am doing wrong?
16:27 Vanfanel: ajax: I am also dynamically loading and unloading the libs, but that should be OK, right?
16:28 ajax: correct, not your fault
16:29 Vanfanel: ajax: ok, then I can continue :) Thanks a lot!
16:29 amonakov: I've seen that a while ago, but a Mesa update resolved it for me
16:30 Vanfanel: amonakov: what MESA version are you using now? Did you update via your distro's package manager or manually?
16:31 amonakov: I'm currently on 20.2.2, distro updates
16:33 Vanfanel: amonakov: it seems that 20.0.8 is the latest version available on Lubuntu, so that's what I have on my desktop machine. Thanks! :)
17:55 dcbaker: itoral: I've got 9f2afe4170 and 2832cbea7a nominated for 20.3, but neither of them apply and I'm not sure how to resolve them. Do you want to de-nominate them, or could you provide me a backport please?
17:56 dcbaker: cmarcelo: I've got f65750d221 nominated for 20.3, but I can't figure out how to resolve the conflicts. Could you provide a backport?
18:00 dcbaker: kusma, zmike : I had to massage the last zink commit on the staging/20.3 branch a tiny bit. Could one of you look at it and let me know if it looks good?
18:00 zmike: dcbaker: sure, uh, can you give me a link? not at a real computer atm
18:05 dcbaker: zmike: https://gitlab.freedesktop.org/mesa/mesa/-/commit/6731e34f52b149e80537666732d567540d2a2f09
18:07 zmike: dcbaker: 👍 from me
18:18 dcbaker: zmike: Thanks! it was a very minor conflict, but I still like to ask
18:18 zmike: dcbaker: yup, appreciated!
18:32 ajax: #intel-gfx
18:32 ajax: ngh
19:19 cmarcelo: dcbaker: I'll make a backport. the MR should target staging/20.3 right?
19:19 dcbaker: cmarcelo: yes please
19:36 cmarcelo: dcbaker: f65750d2215 ("spirv: Implement OpArrayLength for OpenGL")
19:36 cmarcelo: dcbaker: f65750d2215 ("spirv: Implement OpArrayLength for OpenGL")
19:36 cmarcelo: ops
19:36 cmarcelo: wrong paste
19:37 cmarcelo: dcbaker: dcbaker: MR 8167... should I drop the original "Part of" tag?
19:37 dcbaker: cmarcelo: No, please leave it. A normal backported patch would have it
19:37 dcbaker: thank you
20:45 cmarcelo: dcbaker: not clear to me that CI failure is related to the patch (the backport)
20:45 dcbaker: It's not your fault, I saw that on the staging/20.3 branch as well
20:46 dcbaker: Can you force push over your branch to get the broken commit from Eric out of it?
20:47 dcbaker: cmarcelo: ^
20:51 cmarcelo: dcbaker: ok
20:56 dcbaker: thanks, sorry about that, I should have checked that staging/20.3 was green before assigning to marge
21:32 cmarcelo: dcbaker: np
22:59 dcbaker: anholt: https://gitlab.freedesktop.org/mesa/mesa/-/commit/1e32cb0675b9a2e7de200c2ed254abe8d193c86f/pipelines?ref=staging%2F20.3 is coming up with some unexpected passes, and they look a little odd to me. Should I just mark those tests as expected pass, or does that surprise you too?
23:02 anholt: dcbaker: looks like we need to remove dEQP-GLES31.functional.texture.multisample.samples_1.use_texture_depth_2d.* from bypass-fails in the backport.
23:03 anholt: the deqp 1.2.5.0 bump had dropped them from xfails since they were in the flakes list, but you don't have that on the stable branch
23:04 dcbaker: Here's what it says unexpected passed, it's a bit of an odd mixture:
23:04 dcbaker: 0-12-18 22:18:01 R SERIAL-CPU> dEQP-GLES31.functional.stencil_texturing.render.depth24_stencil8_draw,UnexpectedPass
23:04 dcbaker: 20-12-18 22:18:01 R SERIAL-CPU> dEQP-GLES31.functional.texture.multisample.samples_3.use_texture_depth_2d,UnexpectedPass
23:04 dcbaker: 20-12-18 22:18:01 R SERIAL-CPU> dEQP-GLES31.functional.texture.multisample.samples_4.use_texture_depth_2d_array,UnexpectedPass
23:05 anholt: yeah. need to remove them from the xfails list.
23:05 anholt: they had been listed in flakes because they didn't fail 100% of the time (it's a cache flush bug)
23:06 anholt: so they were in both xfail (they almost always fail! nothing to see here if they do) and flake (don't stop the pipeline if they unexpectedpass). now the bug is fixed and it should never fail, so drop it from both.
23:09 masterino: you probably have noticed those fuckers here in koh rong island, who press raping charges against me, and talk about their gangsters and stalk me to tell their disgusting bi and lesbo show khmer porno partners about my private life , some wanker and gangsters stories again, like biritish clowns did, they were sent to take a dump mostly. Don't bother those are high end clowns. MY pals are ready to kill them if needed.
23:10 anholt:gets XML conversion and QPA file cleanup into deqp-runner, now to see about junit.
23:10 dcbaker: cool, I'll just remove those form the list
23:11 anholt: thanks!
23:11 masterino: estonian and finish stalkers , welcome to my world of issues, one of the nastiest kinds.
23:11 masterino: *finnish
23:15 masterino: in two thousound and eleven i had 50of such from british international cockblocking stalkers, next to experiences of deep kind with estonian ones , where all the country declered death to me, because of some intrigues around abortionleftovers like this, so professional athletes what i also almost used to be and was about to come until taken down came to help me, and sent those clowns away.
23:16 masterino: we had this problem in our country, when the estonian military slaughtered people on the grounds of racism and cockblocking.
23:16 masterino: and i appeared on that list for some whatever reason as well.
23:21 masterino: they continuosly threattened and freaked out all my fans during those periods of time, and terrorised me , and today are very desprate that they could not make any progress against me, cause i was way too smart for idiots like that.
23:22 masterino: cause my grannies uncle died 42 years old due to estonian picketers and fuckers and agents who talked stories in europe and world , i knew round about what to expect from such.
23:24 masterino: so you do not have to bother about those gangster wannabees, top athletes olympic medalists are behind trying to help me out there.
23:24 masterino: and basically those will rip out the gangsters hands with pretty much raw powers used, they lose their hand alltogether if needed
23:24 masterino: ligaments and everything torned
23:26 masterino: couple of stronger policemen may come as well, and gangstas go straight back to jail
23:27 masterino: those are the most highest end ridiculous idiots
23:31 masterino: my knee has been soar for their knee surgery deliberate 5times protocol violation , only on single occasion for 19years in a row, i may get it resolved this year finally, my dad has been coworking with authorities who made it happen.
23:31 masterino: suprisingly complex sitution very like retarded formation, but yeah what can i say
23:32 masterino: so i have two big injuries quite painful, plus a follow up diaogosis from the desperation and anxiety from one of those, not being able to finish me back times.
23:34 masterino: one of my friends was professional wrestler, and he knows what happens there, he was about to get the same faith, butcher was allready waiting.
23:34 masterino: but he did not go to the surgery, and got three gold medals in wrestling insead.
23:34 masterino: instead.