00:08 Lyude: clangcat[d]: which issue was this? if there's a patch we're missing upstream I can push it if it's already been reviewed (or get it reviewed if it was some wip from me)
00:10 phomes_[d]: a heads up for those testing games. Multiple games are crashing since aa1f00cf. I filed an issue (11505)
00:23 airlied[d]: The sg debug is a mess, I think once I resend the last workaround patch it might survive, just need to get a spot in my week
07:34 clangcat[d]: Lyude: The Crash on boot was your `SG_DEBUG fix` patch. But that has been reverted and the newer mainline kernels still don't work for me. As for the EGL/VK(happens with your patch applied) bad allocations that could technically be compositors/mesa or kernel driver unsure where exactly the problem is with that one.
07:39 clangcat[d]: airlied[d]: Yea it does seem like, and I get it's more of a headache as from what I've seen patching it so my card doesn't crash on boot, leads to other peoples(different cards) having issues so blegh.
07:39 clangcat[d]: I've also tried just disabling SG_DEBUG which also leads to bad allocations in EGL/VK code but unsure if that is mesa or kernel driver issues. (Or even compositor issues but it happens in sway, KDE and my own compositor so technically possible we are all misusing mesa EGL/NVK)
15:17 mhenning[d]: gfxstrand[d]: https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30166 is tiny and fixes a bunch of breakage on main, if you want to review
15:26 karolherbst[d]: gfxstrand[d]: not really. Maybe the encoding is wrong?
15:27 karolherbst[d]: mhhh
15:27 karolherbst[d]: codegen does use different encodings
15:28 karolherbst[d]: dunno.. maybe check if `nvdisasm` prints the same
15:31 gfxstrand[d]: Nvdisasm is happy with it. 😫
15:31 karolherbst[d]: mhh
15:32 karolherbst[d]: sure it's the shf it complains about?
15:33 karolherbst[d]: but anyway.. codegen emits .high only for .r
15:33 karolherbst[d]: soo.. maybe it's just `shf.l.high`
18:46 gfxstrand[d]: I think it's probably l.high
18:55 karolherbst[d]: could also be that a specific source has to be zero with l.high or something like that
19:10 dadschoorse[d]: gfxstrand[d]: do you have an opinion on this given that in the current state it also affects nak https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30069
19:16 ericonr:seems to be running Firefox on Zink+NVK on a 1050 Ti
19:17 ericonr: wanted to congratulate folks on the recent developments
23:16 maba_kalox[d]: Can somebody recommend tutorial for vulkan layers (I want to render overlay on top of vulkan app)? (Yes, I have tried google, but found only outdated 9years old articles)