https://bugs.winehq.org/show_bug.cgi?id=42502
Bug ID: 42502 Summary: SlingPlayer 1.5.1 crashes on load (Assertion failure) Product: Wine Version: 2.2 Hardware: x86 URL: http://download.slingmedia.com/player/pc/SlingPlayer-S etup-EU-1.5.1.343.exe OS: Linux Status: NEW Keywords: download, regression Severity: normal Priority: P2 Component: directx-d3d Assignee: wine-bugs@winehq.org Reporter: imwellcushtymelike@gmail.com CC: stefan@codeweavers.com Blocks: 17972, 18010, 31651, 38938 Regression SHA1: 34d8b987c493544f640bf5ca0a5a6cbb3d3f4e7b Distribution: Ubuntu
Created attachment 57376 --> https://bugs.winehq.org/attachment.cgi?id=57376 Wine 2.1 console output
SlingPlayer 1.5.1 crashes around 98% of the time with an assertion failure, 1% of the time with a page fault, and 1% of the time it loads, but crashes later in a similar manner. The attached log shows an assertion failure, though the debugger doesn't run.
wine: Assertion failed at address 0xb77df428 (thread 0009), starting debugger... *** Error in `/home/test/.wine/drive_c/Program Files/Sling Media/SlingPlayer/SlingPlayer.exe': corrupted double-linked list: 0x7c3ba500 ***
This is a regression.
34d8b987c493544f640bf5ca0a5a6cbb3d3f4e7b is the first bad commit commit 34d8b987c493544f640bf5ca0a5a6cbb3d3f4e7b Author: Stefan Dösinger stefan@codeweavers.com Date: Thu Jul 30 18:50:15 2015 +0200
wined3d: Try to detect the polygon offset scale value.
FEAR draws the same geometry twice, the second time using zfunc=equal. In both cases it sets a huge depth bias of -0.5, presumably to get better precision for the fragile Z comparison. The GL polygon offset we set ends up being so large that it pulls the geometry into the negative Z range. It isn't clipped (or no longer, older NV drivers probably had a separate bug there), but the Z value gets clamped to 0.0 in the first draw and doesn't match the incoming Z in the second draw.
This commit cannot be automatically reverted in latest git. No known workaround.
The crash is avoided by running it under Valgrind.
Optional native gdiplus to work around Bug 30899.
https://bugs.winehq.org/show_bug.cgi?id=42502
Stefan Dösinger stefandoesinger@gmx.at changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |stefandoesinger@gmx.at
--- Comment #1 from Stefan Dösinger stefandoesinger@gmx.at --- The error message seems unrelated to the change the patch made. I could imagine that wined3d_adapter_find_polyoffset_scale causes problems, but then I'd expect the crash to occur early during startup. You can try to disable the call to this function and see what happens. Just set the scale value to 0, or try (1 << 23).
What graphics driver are you using?
https://bugs.winehq.org/show_bug.cgi?id=42502
--- Comment #2 from Ken Sharp imwellcushtymelike@gmail.com --- (In reply to Stefan Dösinger from comment #1)
Just set the scale value to 0, or try (1 << 23).
Setting the scale to 0 stops the crashes.
What graphics driver are you using?
NVIDIA legacy binary driver - version 173.14.39 which is apparently the latest version that supports my card.
https://bugs.winehq.org/show_bug.cgi?id=42502
--- Comment #3 from Stefan Dösinger stefandoesinger@gmx.at --- That confirms that the bisect result makes sense. What happens if you set the scale to (1 << 23) (which gives 2^24)? That way we can tell if the crash is caused by the testing function or using the polygon offset function when the app renders something.
I guess your GPU is a Geforce FX (sometimes called geforce 5 I believe) card, right? Geforce 4 and earlier would need driver 96.xx, Geforce 6 and 7 up to 304.xx.
https://bugs.winehq.org/show_bug.cgi?id=42502
--- Comment #4 from Ken Sharp imwellcushtymelike@gmail.com --- Created attachment 57394 --> https://bugs.winehq.org/attachment.cgi?id=57394 wine-2.2-33-g60c98ca +d3d
NVIDIA Corporation NV34 [GeForce FX 5200] (rev a1) I should have included that, sorry.
( 1 << 23 ) also works without a crash. I've attached a +d3d trace though I don't know if it's any help.
https://bugs.winehq.org/show_bug.cgi?id=42502
--- Comment #5 from Matteo Bruni matteo.mystral@gmail.com ---
NVIDIA Corporation NV34 [GeForce FX 5200] (rev a1)
NVIDIA legacy binary driver - version 173.14.39 which is apparently the latest version that supports my card.
Generally, you might be better off with Nouveau. Nvidia binary drivers for your GPU haven't seen fixes for many years by now, aside from minor maintenance updates.
https://bugs.winehq.org/show_bug.cgi?id=42502
--- Comment #6 from Stefan Dösinger stefandoesinger@gmx.at --- Can you try to find out which line in wined3d_adapter_find_polyoffset_scale causes the crash? Unfortunately the 173 driver needs an outdated X.org version, so I cannot easily downgrade the driver to test.
https://bugs.winehq.org/show_bug.cgi?id=42502
--- Comment #7 from Ken Sharp imwellcushtymelike@gmail.com --- Took me all morning to try and find a common theme, because I don't really know what I'm doing. Anyway...
Removing any ONE or more of the following lines stops the assertion failure:
gl_info->gl_ops.gl.p_glBindTexture(GL_TEXTURE_2D, color); gl_info->gl_ops.gl.p_glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MAX_LEVEL, 0); gl_info->gl_ops.gl.p_glTexImage2D(GL_TEXTURE_2D, 0, GL_RGBA8, 256, 1, 0, GL_BGRA, GL_UNSIGNED_INT_8_8_8_8_REV, 0);
gl_info->fbo_ops.glFramebufferTexture2D(GL_FRAMEBUFFER, GL_COLOR_ATTACHMENT0, GL_TEXTURE_2D, color, 0);
The "common theme" being the GL_TEXTURE_2D.
Of course the console fills with errors because it effectively breaks the function, but it does allow the application to load.
The current version of Nouveau available to me is rubbish, but when I get around to upgrading Ubuntu then I have little choice as the Nvidia drivers simply won't work with the newer kernels.
I tried this on a laptop with an Intel card and there is no crash, so it certainly seems to be related to the Nvidia drivers. (I have an AMD GPU to test as well when I get around to it.)
https://bugs.winehq.org/show_bug.cgi?id=42502
--- Comment #8 from Stefan Dösinger stefandoesinger@gmx.at --- It looks like it is crashing while setting up an FBO. That isn't too surprising - this part is the most complicated part of the test function.
Does the application run if you set OffscreenRenderingMode to backbuffer? (See https://wiki.winehq.org/Useful_Registry_Keys)
https://bugs.winehq.org/show_bug.cgi?id=42502
Ken Sharp imwellcushtymelike@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Summary|SlingPlayer 1.5.1 crashes |SlingPlayer 1.5.1 assertion |on load (Assertion failure) |failure on load on Nvidia Blocks|17972, 18010, 31651, 38938 | Severity|normal |minor
--- Comment #9 from Ken Sharp imwellcushtymelike@gmail.com --- Indeed it does.
Workaround: "winetricks orm=backbuffer"
https://bugs.winehq.org/show_bug.cgi?id=42502
--- Comment #10 from Stefan Dösinger stefandoesinger@gmx.at --- I'll probably leave it as this. It would be nice if this worked without such a workaround, but because this is an ancient driver that doesn't run on any modern Linux distro fixing this would mean a lot of setup effort on my side.
The bug should remain open though until Nouveau is able to run this. I'll probably be annoyed with having it on my regression list essentially unfixable in a few months though.
One question is if this driver is generally broken with Wine (since this is a function that runs for every app that uses DirectDraw or Direct3D) or if it is broken because the application does something weird (e.g. put the FPU into a non-standard state). If it's the former the bug is of some priority. Do you have any other Windows application you can try? What happens if you run make test in dlls/ddraw/tests? (I'd expect lots of tests to fail; the question is if they even start running)
https://bugs.winehq.org/show_bug.cgi?id=42502
--- Comment #11 from Henri Verbeet hverbeet@gmail.com --- (In reply to Stefan Dösinger from comment #10)
The bug should remain open though until Nouveau is able to run this. I'll probably be annoyed with having it on my regression list essentially unfixable in a few months though.
I'm all for trying to help people, but in principle an assert inside the driver is automatically NOTOURBUG. (Some exceptions like memory corruption aside.)
https://bugs.winehq.org/show_bug.cgi?id=42502
--- Comment #12 from Stefan Dösinger stefandoesinger@gmx.at --- Created attachment 57565 --> https://bugs.winehq.org/attachment.cgi?id=57565 Delete framebuffer before deleting attachments
Here's a shot in the dark: This patch reorders the delete calls to unbind and delete the framebuffer before deleting its attachments. Doing it in the other order should be fine, but I believe this way is less likely to hit bugs in the driver.
https://bugs.winehq.org/show_bug.cgi?id=42502
--- Comment #13 from Ken Sharp imwellcushtymelike@gmail.com --- VLC works fine. WMP won't install.
Quicktime crashes in the same way, same workaround. https://secure-appldnld.apple.com/QuickTime/031-43075-20160107-C0844134-B3CD...
Patch doesn't make any difference.
If this is unfixable and NOTOURBUG then so be it, but I'll let you guys decide what to do with this.
https://bugs.winehq.org/show_bug.cgi?id=42502
Ken Sharp imwellcushtymelike@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution|--- |ABANDONED
--- Comment #14 from Ken Sharp imwellcushtymelike@gmail.com --- I've finally upgraded to Ubuntu Xenial and the Nvidia drivers are no longer supported for my GPU model so I'm stuck with Nouveau.
There is therefore no crash and I don't much feel like messing around to continue testing the old drivers, so I'm marking this abandoned unless someone else feels like picking it up.
Happy for the resolution to be changed to something more appropriate.
https://bugs.winehq.org/show_bug.cgi?id=42502
André H. nerv@dawncrow.de changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED CC| |nerv@dawncrow.de
--- Comment #15 from André H. nerv@dawncrow.de --- closing abandoned
https://bugs.winehq.org/show_bug.cgi?id=42502
Anastasius Focht focht@gmx.net changed:
What |Removed |Added ---------------------------------------------------------------------------- URL|http://download.slingmedia. |https://web.archive.org/web |com/player/pc/SlingPlayer-S |/20210207134903/http://down |etup-EU-1.5.1.343.exe |load.slingmedia.com/player/ | |pc/SlingPlayer-Setup-EU-1.5 | |.1.343.exe CC| |focht@gmx.net
-
wine-bugs＠winehq.org -
WineHQ Bugzilla