[Bug 59394] New: Tex Murphy: Overseer fails to start (segfault) using NVK/Mesa
http://bugs.winehq.org/show_bug.cgi?id=59394 Bug ID: 59394 Summary: Tex Murphy: Overseer fails to start (segfault) using NVK/Mesa Product: Wine Version: 11.1 Hardware: x86-64 URL: https://store.steampowered.com/app/302370/Tex_Murphy_O verseer/ OS: Linux Status: NEW Keywords: regression Severity: normal Priority: P2 Component: opengl Assignee: wine-bugs@list.winehq.org Reporter: gyebro69@gmail.com CC: rbernon@codeweavers.com Regression SHA1: 6802b32b3859e3b1c3e59bdb462d373fcf0608cf Distribution: ArchLinux Created attachment 80353 --> http://bugs.winehq.org/attachment.cgi?id=80353 glxinfo -B, eglinfo -B The game crashes immediately on start in Wine-11.1 and 11.2 with a segmentation fault when using the open source NVK driver and Mesa 25.3.4. There is no such problem with the Nvidia binary driver 580.126.09. This happens in Wayland (KDE Plasma 6.5.5) and under X11 (XFCE) too. Switching to the GLX backend resolves the crash under X11 and XWayland, but not when using the native Wayland driver. Reverting the following commit on top of Wine-11.2 fixes the problem for me: commit 6802b32b3859e3b1c3e59bdb462d373fcf0608cf opengl32: Fix front buffer emulation glBlitFramebuffer parameters. There is nothing obvious in the terminal. I'm mostly using the binary Nvidia driver, I just happened to bump into this bug. Let me know if you need any further log. I'm attaching here the output of my 'glxinfo -B' and 'eglinfo -B'. I *think* the vga card is properly recognized and used by the OS with the open source driver. Wine-11.2 Mesa 25.3.4 NVIDIA GeForce GTX 1660 SUPER (NVK TU116) -- 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.
http://bugs.winehq.org/show_bug.cgi?id=59394 --- Comment #1 from Rémi Bernon <rbernon@codeweavers.com> --- The native Wayland driver always uses EGL, this is is probably something EGL specific, and probably because with GLX front buffer rendering is possible and used instead. The change from that commit only touches the front buffer rendering emulation code path, which explains that at least. It seems to me that the change is correct nonetheless, and perhaps the NVK driver is crashing because we are actually trying to blit front buffer to back whereas we were using empty/invalid coordinates before. Possibly this is a NVK / zink bug, but I don't have such setup at hand so I cannot be sure for now. -- 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.
http://bugs.winehq.org/show_bug.cgi?id=59394 --- Comment #2 from Rémi Bernon <rbernon@codeweavers.com> --- Fwiw that specific change also fixed bug 59232, so it becomes even more likely that it is correct, and that this issue is instead a zink / NVK bug. It's possible that they don't like blitting front buffer to back. -- 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.
http://bugs.winehq.org/show_bug.cgi?id=59394 Béla Gyebrószki <gyebro69@gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution|--- |NOTOURBUG --- Comment #3 from Béla Gyebrószki <gyebro69@gmail.com> --- (In reply to Rémi Bernon from comment #2)
Fwiw that specific change also fixed bug 59232, so it becomes even more likely that it is correct, and that this issue is instead a zink / NVK bug. It's possible that they don't like blitting front buffer to back.
OK, thanks for the feedback. I'm closing this as not a Wine bug (upstream issue). -- 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.
http://bugs.winehq.org/show_bug.cgi?id=59394 Béla Gyebrószki <gyebro69@gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED --- Comment #4 from Béla Gyebrószki <gyebro69@gmail.com> --- Closing. -- 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.
http://bugs.winehq.org/show_bug.cgi?id=59394 Shmerl <shtetldik@gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |shtetldik@gmail.com --- Comment #5 from Shmerl <shtetldik@gmail.com> --- (In reply to Rémi Bernon from comment #2)
Fwiw that specific change also fixed bug 59232, so it becomes even more likely that it is correct, and that this issue is instead a zink / NVK bug. It's possible that they don't like blitting front buffer to back.
The problem could be related to incorrect multithreading access to pipe_context. See: * https://bugs.winehq.org/show_bug.cgi?id=59677 * https://gitlab.freedesktop.org/mesa/mesa/-/work_items/15335 -- 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