[Bug 49631] New: Wine does not fill in Vulkan device LUID property when running in Virtual Desktop
https://bugs.winehq.org/show_bug.cgi?id=49631 Bug ID: 49631 Summary: Wine does not fill in Vulkan device LUID property when running in Virtual Desktop Product: Wine Version: 5.13 Hardware: x86-64 OS: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: winevulkan Assignee: wine-bugs(a)winehq.org Reporter: sashok.olen(a)gmail.com Distribution: ArchLinux Hi. Previously, this game has an issue with Vulkan backend not being selectable in the video settings, even in normal fullscreen mode, but in recent versions this issue was resolved by Wine exposing Vulkan device LUID property. However, it doesn't seem to work when the game is running in Virtual Desktop mode, it just reverts to its DX11 backend, which suggests there is a bug with LUID property when Wine is in virtual desktop mode. I've additionally tested this with standalone vkd3d-proton and it also reports that it can't find LUID property when in virtual desktop. I could attach a log for WWZ but I honestly see no need as there seems to be absolutely nothing Vulkan related in the output and you can reproduce the issue with standalone (Windows DLL) Vkd3d-Proton as well. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=49631 Olivier F. R. Dierick <o.dierick(a)piezo-forte.be> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |o.dierick(a)piezo-forte.be --- Comment #1 from Olivier F. R. Dierick <o.dierick(a)piezo-forte.be> --- Hello A WINEDEBUG=+vulkan debug log will surely have vulkan related contents. Could you provide a link to the standalone vkd3d-proton application, so that we don't have to search for it? Regards. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=49631 --- Comment #2 from Alexandr Oleynikov <sashok.olen(a)gmail.com> --- I think the easiest way to reproduce this would be to just run the vulkan test from dlls/vulkan-1/tests/ with virtual desktop enabled. If I run it in normal mode, there are no faliures and LUID is set properly: vulkan.c:212: Device 'GeForce GTX 1070', device UUID: {b2da86a7-e117-c09f-d36d-f9052e8030fa}, driver UUID: {76fcdb08-9c22-4d0e-c1c5-24b51a62a1f5}, device LUID: 00000000:000003f2. But in virtual desktop, 2 tests fails and LUID is zeroed out: vulkan.c:212: Device 'GeForce GTX 1070', device UUID: {b2da86a7-e117-c09f-d36d-f9052e8030fa}, driver UUID: {76fcdb08-9c22-4d0e-c1c5-24b51a62a1f5}, device LUID: 00000000:00000000. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=49631 Liam Middlebrook <loothelion(a)nvidia.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |loothelion(a)nvidia.com --- Comment #3 from Liam Middlebrook <loothelion(a)nvidia.com> --- I was discussing this with Alexandr in a chat earlier today and noted that the issue likely stems from some difference in the RROutput values being passed into vkGetRandROutputDisplayEXT() in order to match up VkPhysicalDevice objects to RROutput objects. This is how winex11 gets the vulkan_uuid information, so that winevulkan can query it and be sure that it's grabbing the proper LUID for a given VkPhysicalDevice. I think the immediate questions here are: 1. Is winex11 hitting the path where get_gpu_properties_from_vulkan() is being called for any x11drv_gpu objects? 2. If so, what are the differences in RROutputs being passed to vkGetRandrOutputDisplayEXT()? Under a virtual desktop is there ever a case where the call succeeds? -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=49631 Zhiyi Zhang <zzhang(a)codeweavers.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |zzhang(a)codeweavers.com Ever confirmed|0 |1 Status|UNCONFIRMED |NEW --- Comment #4 from Zhiyi Zhang <zzhang(a)codeweavers.com> --- get_gpu_properties_from_vulkan() is not getting called in Virtual Desktop mode. It requires the XRandR display device handler. I think we can make it work in Virtual Desktop mode as well, e.g., by passing the primary randr output to vkGetRandROutputDisplayEXT if RandR is available even if running in Virtual Desktop. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=49631 --- Comment #5 from Zhiyi Zhang <zzhang(a)codeweavers.com> --- We can probably just return the host primary GPU for Virtual Desktop. The Virtual Desktop mode is used mostly for virtualizing the monitor, not the GPUs. So it's probably alright. I will look into this. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=49631 --- Comment #6 from Zhiyi Zhang <zzhang(a)codeweavers.com> --- Created attachment 67842 --> https://bugs.winehq.org/attachment.cgi?id=67842 patch -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=49631 --- Comment #7 from Alexandr Oleynikov <sashok.olen(a)gmail.com> --- The patch works, thanks! -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=49631 Zhiyi Zhang <zzhang(a)codeweavers.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Resolution|--- |FIXED Status|NEW |RESOLVED Fixed by SHA1| |aadae4d1ea173dfa30e09d6cc81 | |6bb174e35f90d --- Comment #8 from Zhiyi Zhang <zzhang(a)codeweavers.com> --- Upstreammed in aadae4d1ea173dfa30e09d6cc816bb174e35f90d -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
https://bugs.winehq.org/show_bug.cgi?id=49631 Alexandre Julliard <julliard(a)winehq.org> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED --- Comment #9 from Alexandre Julliard <julliard(a)winehq.org> --- Closing bugs fixed in 5.16. -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.
participants (1)
-
WineHQ Bugzilla