07:17luali: hi
07:27luali: i am new here, can Freedreno and Turnip be discussed in this channel?
07:40HdkR: luali: Yes
07:40HdkR: Although the #freedreno channel may be more appropriate
07:42luali: Thanks~
07:45luali: Does the community currently have plans to support the Qualcomm SM8750? At the moment, Qualcomm likely hasn't released any software documentation, or have community already obtained the relevant interfaces?
07:53HdkR: luali: afaik a8xx shouldn't be a huge departure from a7xx. Bigger issue is the lack of an HDK for sm8750 which means hacking on Android phones which isn't very fun. But it's also a relatively new SoC, it'll come in time.
07:59luali: I think Qualcomm hasn't released the HDK (Hardware Development Kit) for the SM8750 now.
08:00luali: I guess it will be in a few months. the community have any channel with Qualcomm?
08:01HdkR: There's at least some coordination on the kernel side.
08:01luali: OK~
08:02HdkR: Obviously everyone would love more public documentation :)
08:04luali: I understand. If there is anything I can do to help community, I will try to contact here.
08:33jfalempe: tzimmermann: did you look at https://lore.kernel.org/dri-devel/CAMwc25rKPKooaSp85zDq2eh-9q4UPZD=RqSDBRp1fAagDnmRmA@mail.gmail.com/T/#u ?
08:34tzimmermann: jfalempe, i'm about to send a patch in a minute
08:34jfalempe: tzimmermann: ok thanks :)
08:35tzimmermann: looks like an off-by-one bug to me. IIRC the matrox manuals are fuzzy on the exact meanings of soem of the registers. so i might have changed something
08:36tzimmermann: but i cannot test the patch. it appears to be happening only with some bmcs
09:17sima: tzimmermann, r-b: me on v3 of your drm_gem_is_imported() fix
09:17tzimmermann: sima, thanks
09:17sima: somehow I don't yet have it here
11:51dj-death: how come I have a test failing on the gitlab CI on dEQP-VK.spirv_assembly.instruction.compute.physical_storage_buffer.vertex_in_out_in
11:51dj-death: but if I take the CTS commit that the CI is using that test doesn't appear to exist??
11:52dj-death: am I correct it's using 76c1572eaba42d7ddd9bb8eb5788e52dd932068e ?
11:55dj-death: ah no, it's using vulkan-cts-1.4.1.1
11:55dj-death: derp
12:17eric_engestrom: dj-death: yeah, so far only radv uses the commit from main, all the other drivers use the release
12:28dj-death: eric_engestrom: thanks
12:56karolherbst: alyssa: fyi, I finally fixed the int conversion test fails 🙃 https://github.com/KhronosGroup/SPIRV-LLVM-Translator/pull/3120
13:16eric_engestrom: zmike, karolherbst: 🪓
13:16zmike: karolherbst: GOGOGOGOGOGGOGOGOGOGOGOGOGOGOGO
13:35alyssa: karolherbst: yay
13:49zmike: let's fucking gooooooooooooooooooooooooooooo
13:50zmike: actually wait
13:50daniels: this is going to be so anticlimactic when you trip over an unused-variable warning on the no-libdrm build
13:50zmike: yeah I just realized I forgot to delete some ci stuff
13:51daniels: as the first line of the MR description says: 'I'm not expecting this to be merged today.'
13:51glehmann: > removed clover state tracker
13:51glehmann: aren't they called frontends now?
13:54alyssa: glehmann: that's how long we've wanted to delete clover for :3
13:58daniels: zmike: I was going to type windows-msvc instead of no-libdrm, figured I wouldn't pick on poor jenatali, but yeah turns out that is exactly what should blow up
13:58zmike: wut
13:58daniels: .gitlab-ci/windows/mesa_build.ps1:-Dgallium-opencl=icd
13:58zmike: I deleted that?
13:59zmike: are you sure you're reviewing the latest diff
13:59daniels: looks like you raced marge
14:00daniels: and won/lost
14:02zmike: in a surprise turn of events, the build that I actually broke was rusticl
14:03daniels: delete it
14:03zmike: fucken done.
14:04daniels: also even if you don't break the build:
14:04daniels: Merge blocked: 2 checks failed
14:04daniels: Cannot merge until 17 Apr 2025, 00:00
14:04alyssa: ...what?
14:04daniels: so might want to delete that too
14:04daniels: alyssa: it's something you can explicitly set on MRs
14:04zmike: where?
14:05daniels: edit -> Merge can start -> After scheduled date -> [...]
14:05alyssa: TIL
14:05alyssa: neat
14:05zmike: wild
14:07daniels: I'd not previously considered its comedy value
15:09llyyr: it's clOVER
15:10austriancoder: \o/
15:11zmike: definitely it is CLover
15:12zmike: karolherbst: the time is now
15:12zmike: delete all the code
15:16glehmann: why are there still open clover issues: https://gitlab.freedesktop.org/mesa/mesa/-/issues/?label_name%5B%5D=clover
15:18karolherbst: it's clover 🙃
15:19glehmann: also someone should probably go over the remaining clover MRs and close the ones that are no longer useful (likely all of them?)
20:08alyssa: who maintains mesa/shader-db?
20:09alyssa:eyes merge button
20:09alyssa: apparently i do. ok
20:11olivial: thanks :)
20:12alyssa: olivial: sure heh
20:12alyssa: I don't think we've met, are you a new collaboran?
20:13olivial: as of like several months ago. I think we talked about the noperspective lowering stuff at one point?
20:13olivial: I changed names recently though
20:14alyssa: ah right
20:19zmike:gets a wild MR ping from 2024
20:22alyssa: zmike: i have "finish a shaderdb patch i started" on my todo list and have instead gotten sidetracked merging half the open MR's
20:29alyssa: in fairness said patch is gonna churn stuff bad so i'm trying to just land everyone's changes first :p