Closed telans closed 4 years ago
I've seen it, but I also see it in alot of other software too. It's just a side effect of using GentooLTO. Other than that I haven't had any major crashes with wine outside of virtual desktop which was fixed via *FLAGS-="${IPAPTA}"
.
You may want to try *FLAGS+=-fno-avx
to see if that clears up your application's crash though.
You may want to try
*FLAGS+=-fno-avx
to see if that clears up your application's crash though.
Wow. I think it did, works with wine-staging git master now too, thanks!
Is AVX known to cause problems? I haven't seen any warnings about it
I know DXVK (at least the winelib version) would crash if -march=native
is used. Turns out it has to do with AVX.
If I can find it I'll put a link to the bug report on it.
Found some bug reports related to march=native and AVX
https://bugs.winehq.org/show_bug.cgi?id=45289 https://bugs.winehq.org/show_bug.cgi?id=43516
Actually I don't believe that worked (nor -mno-avx
which I think you meant). It just seems like blind luck whether a build works with the program or not. I had a working version today, rebuilt with everything the same and it just crashes no matter what. I don't suppose you couldn't give the program a try? Would be a big help: https://www.ti.com/tool/TINA-TI Nevermind looks like you need a login to download from there :(
I've tried updating and rolling back the staging/wine git repo to a commit where I know it worked yesterday as well as trying pure 32/64bit wine with no luck. Looks like I might just have to trial and error until I get a working version, duplicate the ebuild and leave it alone in eselect wine
forever. Any ideas?
Cheers
Can you post the terminal output of the program crashing?
001d:err:ntoskrnl:ZwLoadDriver failed to create driver L"\\Registry\\Machine\\System\\CurrentControlSet\\Services\\wineusb": c0000142
000f:fixme:service:scmdatabase_autostart_services Auto-start service L"wineusb" failed to start: 1114
0025:fixme:mountmgr:harddisk_ioctl Unsupported ioctl 4d004 (device=4 access=3 func=401 method=0)
wine: Read access denied for device L"\\??\\Z:\\", FS volume label and serial are not available.
0009:err:ole:TLB_ReadTypeLib Loading of typelib L"" failed with error 5
0009:err:ole:CoRevokeClassObject COM was not initialized
wine: Unhandled exception 0x0eedfade in thread 9 at address 7B00E188 (thread 0009), starting debugger...
I notice the wineusb
error while the program is working, so I don't believe that is the cause. Same with read access is denied
.
It appears 0x0eedfade
is related to "Delphi", but I haven't seen any related wine bugs specifically related to fixing the error issue. I might file a wine bug to see if anyone can figure it out. Cheers
I'm guessing the program is written in pascal based on the Delphi reference.
Have you looked at WineHQ?
It needs to be updated. I'll see if I can get it working.
Also check this out.
I'm able to run the installer fine without any issues "Tina90-TIen.exe", and Crossover lists it as working too. I do have a working version of wine for TINA on my Arch partition I'm switching over from (wine-tkg-git). Perhaps I might look at installing that build from makepkg on here.
That was made with the same flags on Arch, too. The only change I can account for now is patching cross/binutils amd enabling posix threads for mingw. I'll try with a clean toolchain and see if it helps.
Does it crash as soon as it tries to load?
I just got it installed, and it started up just fine. I'm currently using my wine-staging-5.6 ebuild which is technically lutris-fshack-5.6
The first time I launched it the schematic symbols wouldn't render. They did show up the second time I launched TINA.EXE though.
Yeah no gui or anything, just immediately crashes. That makes me laugh, working on the first try haha. I'll give that ebuild a go once the toolchain is finished building. Cheers
telans@genbox ~ $ wine .wine/drive_c/Program\ Files\ \(x86\)/DesignSoft/Tina\ 9\ -\ TI/TINA.EXE
000b:fixme:winediag:__wine_start_process Wine TkG 5.6 is a testing version containing experimental patches.
000b:fixme:winediag:__wine_start_process Please don't report bugs about it on winehq.org and use https://github.com/Tk-Glitch/PKGBUILDS/issues instead.
0025:fixme:mountmgr:harddisk_ioctl Unsupported ioctl 4d004 (device=4 access=3 func=401 method=0)
wine: Read access denied for device L"\\??\\Z:\\", FS volume label and serial are not available.
0009:err:ole:TLB_ReadTypeLib Loading of typelib L"" failed with error 5
0009:err:ole:CoRevokeClassObject COM was not initialized
wine: Unhandled exception 0x0eedfade in thread 9 at address 7B00E1A8 (thread 0009), starting debugger...
This is with a fresh toolchain with the patches and thread changes, using wine-staging-5.6-r1
along with a fresh wine prefix.
The only thing I imagine that's different between my build and yours would be -march=haswell
being passed to mingw. I'll see how it goes without it. Thank you for testing this, I really do appreciate it
What USE flags do you have enabled? I'll mirror those too
This is what I have in my package.use
app-emulation/wine-staging realtime sdl2 netapi gstreamer vkd3d custom-cflags faudio vaapi mingw
This is what's in my make.conf
USE="libglvnd pulseaudio xinerama realtime vulkan vdpau acl opencl wavpack x264 x265 fontforge zsh-completion unicode"
My cpu is an ivybridge 3570K, and I use -march=native
.
Just to be sure have you tried making a prefix with WINEARCH=win32?
I was looking at the error you posted and noticed this liine.
wine: Read access denied for device L"\\??\\Z:\\", FS volume label and serial are not available.
Have you tried going into winecfg's Drives tab, and removing drive Z:? (Also try removing all drives except for C:)
No dice with WINEARCH=win32
, removing the Z: drive results in loading errors as that's my root drive (same with removing it before installing TINA):
002b:err:module:import_dll Library bcbsmp60.bpl (which is needed by L"unix\\home\\telans\\.wine\\drive_c\\Program Files\\DesignSoft\\Tina 9 - TI\\vhdl_dll2.dll") not found
002b:err:module:import_dll Library bcbsmp60.bpl (which is needed by L"unix\\home\\telans\\.wine\\drive_c\\Program Files\\DesignSoft\\Tina 9 - TI\\vhdl_dll2.dll") not found
002b:err:module:import_dll Library vhdl_dll2.dll (which is needed by L"unix\\home\\telans\\.wine\\drive_c\\Program Files\\DesignSoft\\Tina 9 - TI\\TINA.EXE") not found
002b:err:module:LdrInitializeThunk Importing dlls for L"\\\\?\\unix\\home\\telans\\.wine\\drive_c\\Program Files\\DesignSoft\\Tina 9 - TI\\TINA.EXE" failed, status c0000135
I'll get back to you whether change my use flags fixes it. Cheers
EDIT: Shouldn't wine be avoiding "Drive Z" if it can, and instead use "C:"?
5.6 build with the same USE flags apart from net-api
, opencl
, & vkd3d
. I assume you use the default GentooLTO cflags?
The installer seems to be crashing at the end now, but TINA itself is reporting a different error which looks to be even less helpful:
telans@genbox ~ $ wine .wine/drive_c/Program\ Files\ \(x86\)/DesignSoft/Tina\ 9\ -\ TI/TINA.EXE
000b:fixme:winediag:__wine_start_process Wine TkG 5.6 is a testing version containing experimental patches.
000b:fixme:winediag:__wine_start_process Please don't report bugs about it on winehq.org and use https://github.com/Tk-Glitch/PKGBUILDS/issues instead.
0025:fixme:mountmgr:harddisk_ioctl Unsupported ioctl 4d004 (device=4 access=3 func=401 method=0)
wine: Unhandled page fault on read access to 043FC938 at address 00404E82 (thread 0009), starting debugger...
I'm going to rebuild with -mno-avx
, then remove custom-cflags
if that doesn't work. Even if it doesn't work after that, I think it may be time to give up on getting this program running with wine.
I have an last resort idea if you'd be willing, if none of that works. Would you mind using quickpkg
on a wine build you have working with TINA and send it over somehow? It appears that there are no differences between the two marches: diff <(gcc -march=haswell -Q --help=target) <(gcc -march=ivybridge -Q --help=target)
. Cheers
EDIT: -mno-avx
fixes the installation crash, not the program crash
I forgot to mention that I have
app-emulation/wine* *FLAGS+=-mno-avx *FLAGS-="${IPAPTA}"
in my package.cflags.
(to be honest I should just add -mno-avx
to my ebuild, but I don't think bobwya does that in their ebuilds.)
I'll rebuild my wine-staging without net-api
, opencl
, and vkd3d
. I'll also use -march=x86-64
, then make sure my build works before quickpkg it.
Thank you very much. Unfortunately adding *FLAGS-="${IPAPTA}"
still didn't change anything for me. There must be something going on but it's sure a mystery. What's your glibc version?
My glibc is 2.31-r3.
Here you go. wine-staging-5.6-r1.tbz2
Thank you! Installed perfectly fine, however I'm just running into:
wine: failed to initialize: libunwind.so.1: cannot open shared object file: No such file or directory
Which is strange, because neither version of sys-libs/libunwind provides libunwind.so.1
Anyways, this isn't your issue anymore and should be easily fixed so thank you for all your help.
I forgot I have sys-libs/llvm-libunwind
installed instead of sys-libs/libunwind
.
You should be able to temporarily replace sys-libs/libunwind
with sys-libs/llvm-libunwind
, to get it working.
wine: Unhandled exception 0x0eedfade in thread 9 at address 7B00ED57 (thread 0009), starting debugger...
Well, it's definitely a me problem. I am so stumped and clueless
Well hopefully you'll find a fix. I know wine can be pretty picky.
There was this time where I was trying to get Origin to work, but it kept crashing. It turns out the Windows fonts that I installed from my Win10 partition was causing Origin to crash.
What kernel are you on? I might give genkernel a go, but I don't even know if that's likely to cause a difference
I'm using liquorix-sources-5.6.15_p2 with a custom config.
I also use sakaki's buildkernel to update the kernel which uses genkernel for the initramfs.
@TheGreatMcPain, I have a question if you wouldn't mind. Do you build your glibc/other base libraries with custom-cflags
? (lto and such). I'm having trouble with Proton through Steam, where everything crashes with wine: unhandled exception ...
no matter what version or if I use a tkg/ge proton version. I'm thinking it would have to be something low level like that if multiple different builds crash the same, as it occurs regardless of using the steam runtime or not. Cheers
Well that wasn't it - I'm pretty stumped
I do use glibc with custom-cflags
, but it hasn't caused any issues so far.
Then again I build Proton locally which seems to get around some of the known issues with the newest version of Proton.
Well I built Proton locally through proton-tkg, yet still the same Unhandled exception: 0x20474343 in 64-bit code (0x000000007b011fbe).
in games which should be working flawlessly. I guess it's time to make an arch install for games, as ridiculous as that sounds
I probably should have updated vulkan-loader-9999
once in the last couple months, that was it. I've spent quite a few hours on this for it to be that simple holy shit
It's always user error
There's been a few times where I thaught about having an arch install as a backup incase I need to debug something in Gentoo.
For example there was a time when Origin kept crashing until I tested on Arch to find that it was because I installed Windows 10 fonts onto my system.
Round 2
These aren't all this issues that QA brings up, however I have noticed these exact ones for every wine build I have done (gcc/mingw/stable/staging/staging-9999) and I'm fairly sure it's the cause of a crash I am dealing with for a program.
Would you happen to have seen this? If so, I assume it's not actually all that important?
Cheers