00:18 airlied: are you adventurous enough to try a 6.16-rc1 kernel (either build yourself or see if a distro has it), I might try and track down the problem if I get a chance
00:23 anarsoul: 6.16-rc1 exact or master branch from Linus' tree would also do?
00:31 airlied: Linus master
00:32 anarsoul: OK, compiling
05:36 anarsoul: airlied: https://gist.github.com/anarsoul/bf80683558b3c8359309995205b970e6
05:40 anarsoul: fwiw according to power consumption, the gpu isn't in low-power state
05:52 airlied: anarsoul: okay that isn't good
05:52 airlied: I didn't expect things to get worse
05:53 airlied[d]: skeggsb9778: ^
06:00 anarsoul: airlied: I'd suggest at least submitting a patch that changes runpm default until it's fixed
06:40 airlied: anarsoul: I'll just see if I can reproduce on my tu117 first
07:19 airlied: anarsoul: can you see what gsp: RM version is reported for the linus kernel boot?
07:20 airlied: oh i see it above
08:22 pavlo_kozlenko[d]: Has anyone used the rustctl driver with nouveau? Does anyone have it working, and what is the maximum version of OpenCL available?
08:36 karolherbst[d]: it's not really working, because the gallium driver uses codegen and it would be a massive amount of work to fix all the corner caases
08:38 sonicadvance1[d]: Rusticl on Zink on NVK here we go
08:40 mangodev[d]: wait
08:40 mangodev[d]: ctl or icl?
08:41 mangodev[d]: nvm
08:41 mangodev[d]: probably icl
08:52 karolherbst[d]: yeah, just use zink
09:25 karolherbst[d]: not that I'm testing rusticl on zink on nvk, but I expect it to work well enough
18:18 anarsoul: airlied: I really wonder why the changes like this do not go through CI before getting merged
18:19 karolherbst: hw CI is hard
18:20 anarsoul: but mesa has it nevertheless
18:23 karolherbst: sure, but somebody needs to wire it up for nouveau as well and maintain the hw farms
18:28 anarsoul: True, someone would have to invest time and resources into it. Although it is important for the quality of the driver
18:29 anarsoul: anyway, who am I to complain... It is opensource software, so fix it if it doesn't work, right? :)
18:44 snowycoder[d]: karolherbst: Weren't there some 4060 cards that were donated to mesa by Valve for CI integration?
18:44 snowycoder[d]: I heard it some weeks ago around the Gitlab maintenance week
19:02 mohamexiety[d]: This is a kernel change isn’t it
19:02 mohamexiety[d]: That CI is for mesa
19:07 airlied[d]: Also CI for laptop power mgmt is like more insane than even basic CI for an NVIDIA GPU, which is also insane because it costs a lot of time and money to maintain
19:09 mohamexiety[d]: Yeah
19:54 anarsoul: mohamexiety[d]: if your device under test uses up-to-date kernel, it can also catch kernel issues
19:59 anarsoul: airlied[d]: while I understand all the complexeties of CI for laptops I find it crazy to test potentially breaking changes on users. But again, I understand the the project is mostly driven by volunteers, so there is always not enough resources.
20:07 airlied[d]: Linux has always been reliant on users for testing
22:00 airlied[d]: won't say I've conquered depth/stencil, but it's not handing and passing a lot more tests, needs some dma-copy hackery
23:04 pavlo_kozlenko[d]: Will I be able to use my Kepler gt 630 video card to record video to an OBS via va-api? karolherbst[d]