11:24paulk: hi folks, working with an imx8mp using mesa (24 something) on the GC7000 here
11:25paulk: and it seems that scaling 720p to 1080p + converting NV12 to RGB in weston doesn't keep up to 60 fps
11:25paulk: is that the "expected" level of performance for etnaviv?
11:25paulk: also the GPU is clocked at 1 GHz
11:40austriancoder: paulk: you could give https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3418 a try - maybe it helps
11:41paulk: oh that's interesting, thanks :)
11:59wooosaiiii: I have an issue with panel-simple driver and suspend/resume on 6.6 and i.MX6 board. Panel is working until suspend/resume. Afterward the panel comes up for 1 second before it turns off. I tracked it down to a panel_simple_suspend being called after resume and in turn setting my enable-gpios to off... Any ideas?
12:08paulk: wooosaiiii: could be your kms driver misbehaving, you could trace what it's doing with the panel
12:09wooosaiiii: paulk: Any idea how to do that? Anyway, I can get it back working by doing: systemctl restart weston.service
12:42paulk: you could use printk, backtrace, ftrace, anything works to find out what's happening
13:13alyssa: the latest in my witchcraft https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32339
13:45rgallaispou: Hi, fyi there was a trivial conflict merge for drm/xe/xe_bo.c and include/drm/intel/xe_pciids.h when rebuilding drm-tip. I have solved it and pushed to drm-tip after being sure that it compiled.
13:45rgallaispou: If there is any trouble do not hesitate to ping me
14:56davestevenson: Quick query for drm maintainers. I'm about to merge the vc4 series https://lore.kernel.org/linux-arm-kernel/20241025-drm-vc4-2712-support-v2-0-35efa83c8fc0@raspberrypi.com/. Do the dtbinding changes go through drm or dt tree?
15:02luc: are there any circumstances where drm_prime_lookup_buf_handle() will return -ENOENT in drm_gem_prime_fd_to_handle()? self-importing as docs says? in most cases, i think it would get the exported dmabuf and return 0, wouldn't it?
17:01eric_engestrom: can I request a review of a trivial meson change to support megadrivers on windows & macos? https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32333
17:01eric_engestrom: /me has 2.5 work days left this year
17:22mattst88: eric_engestrom: whole MR looks good to me
17:26eric_engestrom: mattst88: thanks ❤️
17:26eric_engestrom: between that and the original reporter confirming it fixes their issue, I'll merge it now so that I have time to fix or revert if something arises
17:27mattst88: awesome
17:29eric_engestrom: ao2_collabora: I just saw your commit https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30895/diffs?commit_id=890e929783fd30069634669018ed759f496084ba which can actually be taken a bit further and greatly simplifies things for me
17:29eric_engestrom: I've included an extended version of your commit in my MR (except adding deqp-vk to the android build) as https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28168/diffs?commit_id=312e94cfee7a8d717dce9f466d7b01b2cf339156; please take a look to confirm I'm not making a mistake as I can only build-test android stuff :)
17:33eric_engestrom: actually no, it's not even build-tested, the android test image build job is disabled since 7101ed7f885853c2976b ("ci: temporarily disable Android test builds")
17:58dcbaker: alyssa: "aae0c1d5a8 asahi,hk: reenable rgb32 buffer textures" is getting tagged for 24.3 because it's a revert, but it doesn't apply, and its not clear to me that you even want it backported
18:07dcbaker: olv: "4d35002949 panvk: expand top-of-pipe and bottom-of-pipe" is tagged as fixes for 24.3, but the majority of the code it's fixing isn't in the 24.3 branch, do you want to send a backport, or do you want me to drop it?
18:09alyssa: dcbaker: no need to backport as discused
18:09dcbaker: alyssa: sounds good, thanks!
19:16olv: Yeah, it is fine to drop. Thanks for checking!
19:26dj-death: am I the only one to reproduce deadlocks in the cache db multipart on main?
19:26dj-death: dEQP-VK.pipeline.pipeline_library.dynamic_offset.graphics.single_set.uniform_buffer.numcmdbuffers_1.sameorder.numdescriptorsetbindings_1.numdynamicbindings_2.numnondynamicbindings_1 just deadlocks on my laptop
19:29dj-death: ah.. I had another process locked on the same file in gdb in another process...