06:07colinmarc: <cwegener> "I already tried at a "low..." <- I can only get dmabufs to work at around 800x600. (Watch out for scale factor.)
06:07colinmarc: Thanks for testing, would be great to know it's not just me :)
06:10cwegener: Is there a way to ignore the available resolutions from EDID and force an unsupport resolution? I mainly do testing with HDMI.
06:11cwegener: Ah. Actually, I think 800x600 is offered by a test monitor I have here ... let me check
06:13cwegener: hmmm ... having trouble to switch to that mode with jay.
06:14colinmarc: What if you run in nested mode?
06:15colinmarc: Actually if it's the same issue, it should be enough to run vkcube at 800x600
06:15colinmarc: But if you have a prime system, I guess it's passing a dmabuf to the second gpu?
06:17cwegener: I've set the NVIDIA to be the primary in the BIOS. So all outputs are wired to the NVIDIA card.
06:21cwegener: hmmm .. ok ... I launched vkcube-wayland (starts at 960x540 i.e. half of the monitor) and then made it smaller and smaller. it kept working until the vkcube window reached a size somewhere in the 200 pixel range.
06:21cwegener: the HDMI output stopped updating at that point.
06:21colinmarc: That's with the OpenGL renderer, right? since you said you can't see anything on vulkan
06:21cwegener: Yes
06:22colinmarc: That's a separate issue, pretty sure. The only thing I can think of would be to run a nested instance of Jay with the Vulkan renderer, and see if that's all black.
06:22colinmarc: So jay (OpenGL) -> Jay (Vulkan) -> maybe vkcube?!
06:24cwegener: let me give that a try
06:25colinmarc: I found https://github.com/mahkoh/jay/issues/49 on the jay issue tracker, might be worth adding a comment there.
06:27cwegener: I was working via this issue for the initial problems regarding the issue with the driver refusing the commit when the input fence fd was present in OpenGL ... https://github.com/mahkoh/jay/issues/159
06:31cwegener: Getting the nesting test going will be a bit tricky (need to figure out sending keys to the nested one ...). I'll test it a bit later.
06:32colinmarc: No worries, thank you!
07:10wlb: wayland-protocols Issue #193 opened by Tobias Bengfort (xi) foreign toplevel: support for desktops https://gitlab.freedesktop.org/wayland/wayland-protocols/-/issues/193
07:52wlb: weston Issue #901 opened by Mateusz Fonfara (maatay) Weston and input devices hotplug https://gitlab.freedesktop.org/wayland/weston/-/issues/901
07:55pq: mclasen, Company, sounds good. Btw. did you already consider that an xdg_toplevel in fullscreen mode already implies compositor-added black bars if the surface is smaller than full screen?
08:00pq: I think those would be most sure way of getting CRTC background used if at all possible.
08:00pq: at least it cannot have any disagreements about which black
14:35wlb: weston Merge request !1506 opened by Michael Tretter (m.tretter) backend-drm: move mode parsing to frontend and add fallback mode https://gitlab.freedesktop.org/wayland/weston/-/merge_requests/1506
14:36wlb: weston Merge request !1254 closed (backend-drm: add fallback-mode for outputs without modes)
15:05wlb: weston/main: Pekka Paalanen * gl-renderer: consolidate pre and post curve LUT GLSL https://gitlab.freedesktop.org/wayland/weston/commit/8fc5e0a333c1 libweston/renderer-gl/fragment.glsl
15:05wlb: weston/main: Pekka Paalanen * gl-renderer: consolidate linpow sampling GLSL https://gitlab.freedesktop.org/wayland/weston/commit/044379f66ad3 libweston/renderer-gl/fragment.glsl
15:05wlb: weston/main: Pekka Paalanen * gl-renderer: consolidate powlin sampling GLSL https://gitlab.freedesktop.org/wayland/weston/commit/7d38e044dd08 libweston/renderer-gl/fragment.glsl
15:05wlb: weston Merge request !1486 merged \o/ (gl-renderer: consolidate GLSL functions https://gitlab.freedesktop.org/wayland/weston/-/merge_requests/1486)
20:02wlb: wayland Issue #457 opened by Sebastian Wick (swick) A new model for content update dependencies and constraints https://gitlab.freedesktop.org/wayland/wayland/-/issues/457