03:54 mareko: zmike: why am I getting this with zink? heaven_x64: ../../src/xcb_in.c:746: xcb_request_check: Assertion `!reply' failed.
04:56 mareko: zmike: https://www.phoronix.com/forums/forum/linux-graphics-x-org-drivers/open-source-amd-linux/1391113-radv-zink-vs-radeonsi-opengl-performance-on-mesa-23-2-devel?p=1391944#post1391944
05:32 mupuf: DavidHeidelberg[m]: thanks a lot! You did the right thing :)
05:45 mupuf: Not sure yet what really happened there, seemed to be firewalls related on the switch side, but a reboot of the gateway fixed it?
05:46 mupuf:really wishes the hosting service would use a dumb switch
05:46 mupuf: Not the first time we had issues
05:47 mupuf: anyway, all back online
05:47 mupuf: I'll send an MR to bring the farm back
05:55 mupuf:will need to add a check on all the runners' last contact with gitlab in his monitoring
06:57 airlied: mareko: why is venus different to virgl?
12:07 zmike: mareko: old xlib version?
12:09 zmike: his results were odd for bioshock also; on radeonsi I only get half the fps he listed
12:12 DavidHeidelberg[m]: mupuf: thanks for bringing the farm back :)
12:13 mupuf: Hehe, you're welcome!
12:13 mupuf: I found two series I had to re-assign to Marge
12:13 mupuf: So, the incident should be over
14:13 FireBurn: Biscected to 58e67bb3c131da5ee14e4842b08e53f4888dce0a drm/amd/display: fix dpms_off issue when disabling bios mode
19:32 sassefa: I'm reading documentation that references drm and client capabilities. I'm not exactly sure what the difference b/w the two are? Are client capabilities just what the DRM master is capable of supporting?
19:32 sassefa: Also what is the difference b/w a DRM driver (which I presume is just a Mesa driver) and the DRM core?
19:48 gio: Hi, does softpipe support Vulkan? Reading the docs it wouldn't seem so, but I'm not terribly sure I'm understanding properly.
19:48 airlied: no\
19:49 gio: Ok, thanks.
21:24 mareko: airlied: venus isolates guest processes, while virgl runs all guest processes in the same host process using the same GL driver AFAIK
21:29 daniels: mareko: isolated ctx is new for venus tbf
21:31 mareko: the native context isolates guest processes too, but differently - it opens a separate drm_file for each guest process in the same host process
21:31 mareko: and each drm_file gets its own GPU address space
23:41 sassefa: Found the answer to the difference between the DRM core and driver in the wiki page :p
23:41 sassefa: https://en.wikipedia.org/wiki/Direct_Rendering_Manager#Software_architecture