https://bugs.winehq.org/show_bug.cgi?id=43530
Bug ID: 43530 Summary: A lot of games (mostly old) crash with Nvidia drivers 384.59 Product: Wine-staging Version: 2.14 Hardware: x86 OS: Linux Status: UNCONFIRMED Severity: major Priority: P2 Component: -unknown Assignee: wine-bugs@winehq.org Reporter: qptainnemo@gmail.com CC: erich.e.hoover@wine-staging.com, michael@fds-team.de, sebastian@fds-team.de Distribution: ---
Created attachment 58926 --> https://bugs.winehq.org/attachment.cgi?id=58926 crash report from wine
Pretty much all old games stopped working, as in anything pre-directx 9. Examples include Anachronox, Arcanum, Kult: Heretic Kingdoms. Some non-old games also do, such as Test Drive Unlimited 2 and Driftmoon. I can't discern the pattern between the games that work and the games that don't but it seems to be affecting mostly pre-dx9 games.
I'm running Arch Linux and using the currently latest proprietary Nvidia drivers 384.59 (arch package 384.59-4).
All of the mentioned games launch in wine staging 2.13 and in vanilla Wine 2.14.
An example of a crash report is in the attachment.
Here's my steam game list if you want me to test something else http://steamcommunity.com/id/qptain_nemo/games/?tab=all
https://bugs.winehq.org/show_bug.cgi?id=43530
qptain Nemo qptainnemo@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Distribution|--- |ArchLinux
https://bugs.winehq.org/show_bug.cgi?id=43530
--- Comment #1 from Sebastian Lackner sebastian@fds-team.de --- (In reply to qptain Nemo from comment #0)
All of the mentioned games launch in wine staging 2.13 and in vanilla Wine 2.14.
Could you please attach the terminal output? There were no Staging specific wined3d changes between 2.13 and 2.14 and we also test pre-dx9 games during our release testing, so it would really surprise me if we somehow missed such a major regression. I assume you already did try to reboot since upgrading your nVidia graphic drivers?
https://bugs.winehq.org/show_bug.cgi?id=43530
--- Comment #2 from qptain Nemo qptainnemo@gmail.com --- The terminal output seems to offer the identical output to the crash report but I just discovered that this is entirely Arch-specific. I downloaded the build of 2.14 provided by Lutris and there was no issue.
https://bugs.winehq.org/show_bug.cgi?id=43530
--- Comment #3 from qptain Nemo qptainnemo@gmail.com --- Created attachment 58927 --> https://bugs.winehq.org/attachment.cgi?id=58927 terminal output
https://bugs.winehq.org/show_bug.cgi?id=43530
Fabian Maurer dark.shadow4@web.de changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |dark.shadow4@web.de
--- Comment #4 from Fabian Maurer dark.shadow4@web.de --- Looks like a nvidia driver bug to me. It really shouldn't crash in libglx_nvidia.so.0. Does it work with mesa?
https://bugs.winehq.org/show_bug.cgi?id=43530
Erik Hanson erik@slackbuilds.org changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |erik@slackbuilds.org
https://bugs.winehq.org/show_bug.cgi?id=43530
solusoperandi solusoperandi@protonmail.ch changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |solusoperandi@protonmail.ch
https://bugs.winehq.org/show_bug.cgi?id=43530
felixonmars@archlinux.org changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |felixonmars@archlinux.org
--- Comment #5 from felixonmars@archlinux.org --- FYI: Nekopara Vol.3 crashes with nvidia drivers but works with mesa here. Very similar backtrace.
https://bugs.winehq.org/show_bug.cgi?id=43530
Sebastian Lackner sebastian@fds-team.de changed:
What |Removed |Added ---------------------------------------------------------------------------- See Also| |https://bugs.winehq.org/sho | |w_bug.cgi?id=43594
https://bugs.winehq.org/show_bug.cgi?id=43530
--- Comment #6 from Sebastian Lackner sebastian@fds-team.de --- (In reply to qptain Nemo from comment #2)
The terminal output seems to offer the identical output to the crash report but I just discovered that this is entirely Arch-specific. I downloaded the build of 2.14 provided by Lutris and there was no issue.
Have you been able to find out more details about this issue in the meantime? Is it maybe related to a compiler change or optimization? Does it work when building Wine with "-O0" for example?
https://bugs.winehq.org/show_bug.cgi?id=43530
andy andy86@fastwebnet.it changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |andy86@fastwebnet.it
--- Comment #7 from andy andy86@fastwebnet.it --- Same crash for me with most non steam games, also in directx9.
Strange but steam games work well without bug.
https://bugs.winehq.org/show_bug.cgi?id=43530
mrdeathjr28@yahoo.es changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |mrdeathjr28@yahoo.es
--- Comment #8 from mrdeathjr28@yahoo.es --- In my case works anachronox, test drive unlimited 2 + mods with nvidia drivers 384.59 and 384.69
However i use xubuntu 16.04 x64
https://bugs.winehq.org/show_bug.cgi?id=43530
--- Comment #9 from andy andy86@fastwebnet.it ---
but I just discovered that this is entirely Arch-specific.
Hello. I note that file "libglx_nvidia.so.0" mentioned in the bug is not present in my filesystem, instead is present "libglx.so.0", both in lib and lib32.
Maybe that Arch package for nvidia drivers rename libs unlike wine expect?
I try to make symbolic link between two mentioned file, but unlucky.
https://bugs.winehq.org/show_bug.cgi?id=43530
--- Comment #10 from Felix Yan felixonmars@archlinux.org --- I believe it has something to do with case sensitivity. The real file should be:
/usr/lib/libGLX_nvidia.so.0 /usr/lib32/libGLX_nvidia.so.0
https://bugs.winehq.org/show_bug.cgi?id=43530
--- Comment #11 from Fabian Maurer dark.shadow4@web.de ---
I believe it has something to do with case sensitivity.
What makes you think so? Does it work when you make a symlink for those two files?
https://bugs.winehq.org/show_bug.cgi?id=43530
--- Comment #12 from Felix Yan felixonmars@archlinux.org --- Created attachment 59052 --> https://bugs.winehq.org/attachment.cgi?id=59052 Build environment difference between wine-staging 2.13 and 2.14
I was replying to andy, and I think the case sensitivity doesn't have anything to do with the bug.
I have attached the build environment difference between wine-staging 2.13 and 2.14. Since wine-staging 2.13 on the current system still works, the build environment is somewhat suspicious.
https://bugs.winehq.org/show_bug.cgi?id=43530
--- Comment #13 from Sebastian Lackner sebastian@fds-team.de --- (In reply to Felix Yan from comment #12)
I have attached the build environment difference between wine-staging 2.13 and 2.14. Since wine-staging 2.13 on the current system still works, the build environment is somewhat suspicious.
Did you try to rebuild an old Wine version with the new build tools? That would probably be a good test to make sure its really a build environment issue.
https://bugs.winehq.org/show_bug.cgi?id=43530
--- Comment #14 from Felix Yan felixonmars@archlinux.org --- (In reply to Sebastian Lackner from comment #13)
(In reply to Felix Yan from comment #12)
I have attached the build environment difference between wine-staging 2.13 and 2.14. Since wine-staging 2.13 on the current system still works, the build environment is somewhat suspicious.
Did you try to rebuild an old Wine version with the new build tools? That would probably be a good test to make sure its really a build environment issue.
Just tried to build wine-staging 2.13 with the latest build tools, and the same error appears. I have looked through the diff but didn't have an idea yet. Which ones would you suggest me to try downgrading first?
https://bugs.winehq.org/show_bug.cgi?id=43530
--- Comment #15 from andy andy86@fastwebnet.it --- @Felix
Sorry, you're in right, i've forget '-i' in 'locate' command. :Z
I try to downgrade lib32-nvida-utils, but nothing change. Then problem seems not to be in that file (but in wine?).
https://bugs.winehq.org/show_bug.cgi?id=43530
Sebastian Lackner sebastian@fds-team.de changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |aarda.uunlu@gmail.com
--- Comment #16 from Sebastian Lackner sebastian@fds-team.de --- *** Bug 43623 has been marked as a duplicate of this bug. ***
https://bugs.winehq.org/show_bug.cgi?id=43530
Sebastian Lackner sebastian@fds-team.de changed:
What |Removed |Added ---------------------------------------------------------------------------- Product|Wine-staging |Wine Component|-unknown |-unknown
https://bugs.winehq.org/show_bug.cgi?id=43530
Sebastian Lackner sebastian@fds-team.de changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |michael.kogan@gmx.net
--- Comment #17 from Sebastian Lackner sebastian@fds-team.de --- *** Bug 43622 has been marked as a duplicate of this bug. ***
https://bugs.winehq.org/show_bug.cgi?id=43530
Sebastian Lackner sebastian@fds-team.de changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |steveamigauk@yahoo.co.uk
--- Comment #18 from Sebastian Lackner sebastian@fds-team.de --- *** Bug 43594 has been marked as a duplicate of this bug. ***
https://bugs.winehq.org/show_bug.cgi?id=43530
--- Comment #19 from Sebastian Lackner sebastian@fds-team.de --- (In reply to Sebastian Lackner from comment #13)
(In reply to Felix Yan from comment #12)
I have attached the build environment difference between wine-staging 2.13 and 2.14. Since wine-staging 2.13 on the current system still works, the build environment is somewhat suspicious.
Did you try to rebuild an old Wine version with the new build tools? That would probably be a good test to make sure its really a build environment issue.
My first few guesses would be:
lib32-mesa lib32-glibc mesa
Please note that you will have to a full rebuild of Wine after downgrading those packages. My theory is that the underlying issue is either in a header file or in a static lib.
To confirm that theory: Is it possible to get a working build when building with Nvidia drivers installed?
https://bugs.winehq.org/show_bug.cgi?id=43530
--- Comment #20 from Photon michael.kogan@gmx.net --- Is it sure that https://bugs.winehq.org/show_bug.cgi?id=43622 is a duplicate of this? In #43622 the problem appears in 2.15 only but 2.14 works with 384.59 drivers perfectly well!
https://bugs.winehq.org/show_bug.cgi?id=43530
Sebastian Lackner sebastian@fds-team.de changed:
What |Removed |Added ---------------------------------------------------------------------------- Summary|A lot of games (mostly old) |A lot of games (mostly old) |crash with Nvidia drivers |crash in Nvidia driver when |384.59 |Wine is compiled on current | |ArchLinux build environment
--- Comment #21 from Sebastian Lackner sebastian@fds-team.de --- (In reply to Photon from comment #20)
Is it sure that https://bugs.winehq.org/show_bug.cgi?id=43622 is a duplicate of this? In #43622 the problem appears in 2.15 only but 2.14 works with 384.59 drivers perfectly well!
Yes, its really the same bug, the backtrace is exactly identical. The title of this bug is a bit misleading though, I have updated it based on the current information. The issue seems to occur only when Wine is compiled with a current ArchLinux build environment. So far it is not known yet which packages are exactly the issue, but using a build from a different source works around the issue (see comment 2). Also, I'm not yet aware of any reports for other distributions, so its maybe even an ArchLinux specific change (or other distros just haven't updated their packages yet).
https://bugs.winehq.org/show_bug.cgi?id=43530
--- Comment #22 from Photon michael.kogan@gmx.net --- I see, thanks for the explanation!
https://bugs.winehq.org/show_bug.cgi?id=43530
--- Comment #23 from Felix Yan felixonmars@archlinux.org --- (In reply to Sebastian Lackner from comment #19)
My first few guesses would be:
lib32-mesa lib32-glibc mesa
Thanks, I'll try with them first.
To confirm that theory: Is it possible to get a working build when building with Nvidia drivers installed?
No, building with nvidia drivers installed didn't fix the issue here. However, building in my current "dirty" environment (all packages are up to date, but many more installed packages than the clean build environment) does get a working build.
https://bugs.winehq.org/show_bug.cgi?id=43530
--- Comment #24 from Felix Yan felixonmars@archlinux.org --- By comparing the config.log between the build environments, the problem turns out to be "-fno-plt" in CFLAGS and "-z,now" in LDFLAGS, which Arch added recently. Wine built with these two flags enabled will have this issue.
I have pushed wine-staging 2.15-2 to the repos so you can try yourself. Other variants will follow soon.
https://bugs.winehq.org/show_bug.cgi?id=43530
--- Comment #25 from andy andy86@fastwebnet.it --- Hello. 2.15-2 working for me. Thank-you Felix Yan.
https://bugs.winehq.org/show_bug.cgi?id=43530
--- Comment #26 from steveamigauk@yahoo.co.uk --- Confirmed - 2.15-2 seems to be working fine.
https://bugs.winehq.org/show_bug.cgi?id=43530
--- Comment #27 from Photon michael.kogan@gmx.net --- -2 works as well, thanks for the quick fix!
https://bugs.winehq.org/show_bug.cgi?id=43530
--- Comment #28 from andy andy86@fastwebnet.it --- Bug is still present with 2.16-staging. Need downgrade.
https://bugs.winehq.org/show_bug.cgi?id=43530
--- Comment #29 from steveamigauk@yahoo.co.uk --- 2.16-2 fixes it again
https://bugs.winehq.org/show_bug.cgi?id=43530
--- Comment #30 from andy andy86@fastwebnet.it --- Hello. Anyone have try 2.18-1 from repo? Is problem still present?
https://bugs.winehq.org/show_bug.cgi?id=43530
--- Comment #31 from steveamigauk@yahoo.co.uk --- (In reply to andy from comment #30)
Hello. Anyone have try 2.18-1 from repo? Is problem still present?
2.18 seems fine.
https://bugs.winehq.org/show_bug.cgi?id=43530
--- Comment #32 from andy andy86@fastwebnet.it --- Thank you. 2.18-1 works well for me, too.
https://bugs.winehq.org/show_bug.cgi?id=43530
--- Comment #33 from Sebastian Lackner sebastian@fds-team.de --- (In reply to andy from comment #30)
Hello. Anyone have try 2.18-1 from repo? Is problem still present?
Technically the bug is still present, but Archlinux packages contain a workaround for it and disable the problematic build flags.
(In reply to Felix Yan from comment #24)
By comparing the config.log between the build environments, the problem turns out to be "-fno-plt" in CFLAGS and "-z,now" in LDFLAGS, which Arch added recently. Wine built with these two flags enabled will have this issue.
Any news regarding this bug? Do you know if this is something which can be fixed in Wine, or is it a bug in the NVIDIA driver?
https://bugs.winehq.org/show_bug.cgi?id=43530
Robert Walker bob.mt.wya@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |bob.mt.wya@gmail.com
https://bugs.winehq.org/show_bug.cgi?id=43530
noabody@yahoo.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |noabody@yahoo.com
--- Comment #34 from noabody@yahoo.com --- Please check this bug report to see if the problem is related: https://bugs.winehq.org/show_bug.cgi?id=44395
https://bugs.winehq.org/show_bug.cgi?id=43530
--- Comment #35 from Felix Yan felixonmars@archlinux.org --- Someone from NVIDIA confirmed it's a bug in the driver. Hopefully this will be resolved in future releases.
https://bugs.winehq.org/show_bug.cgi?id=43530
Arthur Huillet arthur.huillet@free.fr changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |arthur.huillet@free.fr
https://bugs.winehq.org/show_bug.cgi?id=43530
tokktokk fdsfgs@krutt.org changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |fdsfgs@krutt.org
https://bugs.winehq.org/show_bug.cgi?id=43530
--- Comment #36 from coreyoconnor@gmail.com --- On NixOS the solution was to compile wine with the LD flag "-z now". Per documentation this results in: "When generating an executable or shared library, mark it to tell the dynamic linker to resolve all symbols when the program is started, or when the shared library is linked to using dlopen, instead of deferring function call resolution to the point when the function is first called."
* https://linux.die.net/man/1/ld * https://github.com/NixOS/nixpkgs/issues/31989
https://bugs.winehq.org/show_bug.cgi?id=43530
--- Comment #37 from Arthur Huillet arthur.huillet@free.fr --- Sorry I didn't update this bug earlier - I am the "someone from NVIDIA" fromhttps://bugs.winehq.org/show_bug.cgi?id=43530#c35
This was a NVIDIA driver bug which we fixed in our r415 branch - any driver 415.XX or more recent should carry the fix. Building Wine with full relro will now work properly.
https://bugs.winehq.org/show_bug.cgi?id=43530
Matteo Bruni matteo.mystral@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Resolution|--- |NOTOURBUG Status|UNCONFIRMED |RESOLVED
--- Comment #38 from Matteo Bruni matteo.mystral@gmail.com --- (In reply to Arthur Huillet from comment #37)
This was a NVIDIA driver bug which we fixed in our r415 branch - any driver 415.XX or more recent should carry the fix. Building Wine with full relro will now work properly.
Thanks for the update! Closing this bug as fixed upstream.
https://bugs.winehq.org/show_bug.cgi?id=43530
Matteo Bruni matteo.mystral@gmail.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED
--- Comment #39 from Matteo Bruni matteo.mystral@gmail.com --- Closing.
https://bugs.winehq.org/show_bug.cgi?id=43530
--- Comment #40 from Matteo Bruni matteo.mystral@gmail.com --- *** Bug 44395 has been marked as a duplicate of this bug. ***
-
wine-bugs＠winehq.org