http://bugs.winehq.org/show_bug.cgi?id=58812
Bug ID: 58812 Summary: PlayOnline Viewer: Application runs and works otherwise, but only shows a black screen. Product: Wine Version: 10.17 Hardware: x86-64 OS: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: winex11.drv Assignee: wine-bugs@winehq.org Reporter: chiitoo@gentoo.org Regression SHA1: 7abf4432a7dfdc044f0458f2457829038dc497f0. Distribution: Gentoo
Created attachment 79467 --> http://bugs.winehq.org/attachment.cgi?id=79467 terminal output from a bad run
After 7abf4432a7d [1], the PlayOnline Viewer runs and seems functional, but only shows a black screen.
There doesn't seem to be anything out of the ordinary in the terminal output, aside from a flood of:
libEGL warning: FIXME: egl/x11 doesn't support front buffer rendering.
Last tested at 'wine-10.17'.
Thank you!
1. https://gitlab.winehq.org/wine/wine/-/commit/7abf4432a7dfdc044f0458f24578290...
http://bugs.winehq.org/show_bug.cgi?id=58812
Chiitoo chiitoo@gentoo.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Keywords| |download, regression URL| |https://web.archive.org/web | |/20210810150839/http://www. | |playonline.com/ff11eu/downl | |oad/media/install_win.html
http://bugs.winehq.org/show_bug.cgi?id=58812
Alexandre Julliard julliard@winehq.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Regression SHA1|7abf4432a7dfdc044f0458f2457 |7abf4432a7dfdc044f0458f2457 |829038dc497f0. |829038dc497f0
http://bugs.winehq.org/show_bug.cgi?id=58812
Rémi Bernon rbernon@codeweavers.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |rbernon@codeweavers.com
--- Comment #1 from Rémi Bernon rbernon@codeweavers.com --- Could you please attach a WINEDEBUG=+wgl log?
http://bugs.winehq.org/show_bug.cgi?id=58812
Krzysztof Sobiecki sobkas@sobkas.io changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |sobkas@sobkas.io
--- Comment #2 from Krzysztof Sobiecki sobkas@sobkas.io --- This is a MESA warning: libEGL warning: FIXME: egl/x11 doesn't support front buffer rendering. In dri3_flush_front_buffer at src/egl/drivers/dri2/platform_x11_dri3.c - /* There does not seem to be any kind of consensus on whether we should - * support front-buffer rendering or not: - * http://lists.freedesktop.org/archives/mesa-dev/2013-June/040129.html - */
dri3_flush_front_buffer doesn't do anything so it might be a reason it shows black screen...
http://bugs.winehq.org/show_bug.cgi?id=58812
--- Comment #3 from Chiitoo chiitoo@gentoo.org --- Created attachment 79490 --> http://bugs.winehq.org/attachment.cgi?id=79490 terminal output with +wgl
(In reply to Rémi Bernon from comment #1)
Could you please attach a WINEDEBUG=+wgl log?
Certainly!
Snipped ~100k lines of hopefully only duplication to make it fit into 10 MiB.
(In reply to Krzysztof Sobiecki from comment #2)
This is a MESA warning:
Yeah, figured that part too.
Using the open amdgpu driver with a Radeon RX 5700 XT with mesa-25.2.5 at this time in case it helps.
http://bugs.winehq.org/show_bug.cgi?id=58812
Rimiz cramea@proton.me changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |cramea@proton.me
--- Comment #4 from Rimiz cramea@proton.me --- WineX11 now has the same issues as WineWayland. On WineWayland, I've seen games where videos don't play; you just hear sound because the game is running via WineD3D on WineWayland now with Wine 10.17 Winex11 has this issue as well. The difference between winewayland and winex11, though, is that in the logs it spams FIXME: egl/x11 doesn't support front buffer rendering, but when running in winewalyand, the log is silent, and you don't know why the application isn't functioning properly.
http://bugs.winehq.org/show_bug.cgi?id=58812
--- Comment #5 from Rimiz cramea@proton.me --- Btw, if you can somehow force the application to use Vulkan, all the issues disappear, this is a major OpenGL issue; it's pretty bad.
http://bugs.winehq.org/show_bug.cgi?id=58812
--- Comment #6 from Rémi Bernon rbernon@codeweavers.com --- Created attachment 79491 --> http://bugs.winehq.org/attachment.cgi?id=79491 Possible fix
It's arguably a missing feature from Wayland and EGL, especially when the EGL and GL spec explicitly allow it, but as modern platforms have apparently decided to not bother implementing front buffer rendering and not even emulate it, we need to do it ourselves. It's just the occasion to fix them both winewayland and winex11 at once now.
Does this patch help?
http://bugs.winehq.org/show_bug.cgi?id=58812
imaxm memax@gmx.fr changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |memax@gmx.fr
--- Comment #7 from imaxm memax@gmx.fr --- Created attachment 79493 --> http://bugs.winehq.org/attachment.cgi?id=79493 Terminal output - Voyage (Kheops Studio game)
For your information, the same problem (blank screen) affects Kheops Studio games when software rendering is used: * Egypt III: The Fate of Ramses, aka The Egyptian Prophecy * Return to Mysterious Island * Echo: Secrets of the Lost Cavern * Voyage: Inspired by Jules Verne, aka Journey to the Moon
http://bugs.winehq.org/show_bug.cgi?id=58812
--- Comment #8 from imaxm memax@gmx.fr --- (In reply to Rémi Bernon from comment #6)
Created attachment 79491 [details] Possible fix
It's arguably a missing feature from Wayland and EGL, especially when the EGL and GL spec explicitly allow it, but as modern platforms have apparently decided to not bother implementing front buffer rendering and not even emulate it, we need to do it ourselves. It's just the occasion to fix them both winewayland and winex11 at once now.
Does this patch help?
Hello,
Your patch resolves the issue with Kheops Studio games (I tested Voyage: Inspired by Jules Verne, also known as Journey to the Moon). Thank you.
For your information, there is a new “fixme” in the terminal output: 0180:fixme:d3d:surface_cpu_blt Filter WINED3D_TEXF_LINEAR not supported in software blit.
http://bugs.winehq.org/show_bug.cgi?id=58812
--- Comment #9 from imaxm memax@gmx.fr --- Created attachment 79494 --> http://bugs.winehq.org/attachment.cgi?id=79494 Terminal output - With patch - Voyage
http://bugs.winehq.org/show_bug.cgi?id=58812
--- Comment #10 from Chiitoo chiitoo@gentoo.org --- (In reply to Rémi Bernon from comment #6)
Does this patch help?
Indeed it does!
Looks good on my side so far.
Thank you!
http://bugs.winehq.org/show_bug.cgi?id=58812
--- Comment #11 from Chiitoo chiitoo@gentoo.org --- Created attachment 79499 --> http://bugs.winehq.org/attachment.cgi?id=79499 page fault backtrace with patch 79491 applied
After further testing, there appears to either be a memory leak of sorts, and quite a bit more memory used in general. Well, virtual memory at least.
Without the patch, and the commit reverted, the virtual memory usage goes like this on 'top', in stages.
PlayOnline Viewer before log-in:
3343040
Final Fantasy XI started, in the first screen before title:
3401136
Title screen:
3421604
Back to PlayOnline Viewer from Final Fantasy XI:
3390496
Now the same stages with the patch:
3828208
3987620
4015256
4027220
At this point, after a while of doing nothing, the PlayOnline Viewer crashes. The same thing happens if Final Fantasy XI is started again before the crash.
These are with a debug build though, so the numbers might in general be a bit higher. Might see what happens with a non-debug build if I have a moment for it.
Backtrace attached with this comment.
http://bugs.winehq.org/show_bug.cgi?id=58812
--- Comment #12 from Rimiz cramea@proton.me --- This patch 79491 fixes all the games that were either starting up with a black screen or showing black screens during video playback for me in both WineWayland and WineX11. Idk about increased memory, though; I haven't tested that much.
http://bugs.winehq.org/show_bug.cgi?id=58812
Chiitoo chiitoo@gentoo.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|UNCONFIRMED |RESOLVED Fixed by SHA1| |63e6cfa106a0b12351e8d01eb5b | |fefd925c3a74e Resolution|--- |FIXED
--- Comment #13 from Chiitoo chiitoo@gentoo.org --- Since the changes to fix this are now in git, I suppose I'll open a new report for the memory issue.
I did give a non-debug build a go too out of interest, and the issue shows up there, too, but so far the virtual size has stayed just below the crashing point (still over 4 G).
Thank you!
http://bugs.winehq.org/show_bug.cgi?id=58812
Erwin Iosef erwiniosef@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |erwiniosef@gmail.com
--- Comment #14 from Erwin Iosef erwiniosef@gmail.com --- *** Bug 58843 has been marked as a duplicate of this bug. ***
http://bugs.winehq.org/show_bug.cgi?id=58812
Alexandre Julliard julliard@winehq.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED
--- Comment #15 from Alexandre Julliard julliard@winehq.org --- Closing bugs fixed in 10.18.