Closed WheezyE closed 2 years ago
WheezyE
commented
3 years ago Just wanted to update this ticket and say thanks for the d0ac1e6 and 664dab7 commits you made to get us farther on this. We're getting close! It's awesome getting to work with you on the discord.
WheezyE
commented
3 years ago Found a bug in RMS Express after commit 549ff980 (that was done to help get VARA running). I just wanted to report it here to help debug. I don't think this was here before:
WheezyE
commented
3 years ago New opcode request for VARA after compiling Box86 2375d707
0308:fixme:olepicture:OleLoadPictureEx (10CB0E04,7366,0,{7bf80980-bf32-101a-8bbb-00aa00300cab},x=0,y=0,f=0,1020FB10), partially implemented.
9173|SIGSEGV @0xb5d88a04 (???) (x86pc=0x7bc5f880/???:"???"), for accessing (nil) (code=1), db=0x4967a28(0x7bc5f880/???)
9173|0x6640accc: Unimplemented Opcode 0F 01 4C 24 4A 0F B7 54Not sure when this cropped up because I haven’t tried to run the Winlink installer in a while, but I just tried to run Winlink installer with Box86 eaf7d0a2 (box86 version w signal handling for dotnet30) and got these new errors:
This all be a bit moot anyway though until signal handling works for dotnet35sp1 since dotnet35sp1 is a pre-requisite for Winlink, but just wanted to post this error here for future debugging
WheezyE
commented
3 years ago Nice :) with 269bc9e28901e67a3a8a8a6d72b0903bccf1e153 (I believe), the Winlink installer runs again.
WheezyE
commented
3 years ago With dotnet35sp1 now installing in box86 (#227), I'm looking at Winlink again.
I'm getting some sigsegv's when trying to run it with just dotnet35sp1 installed. I've logged them with log1 here: log1.log
I also tried updating box86 with today's commits, then installing some of the other components that VARA needs (winetricks -q corefonts dotnet35sp1 vb6run vcrun2015 win7) and logged again here: morecomponents_log1.log
Log2 looks like this: morecomponents_log2.log.zip
WheezyE
commented
3 years ago Thanks for the updated signal handling for dotnet!
WheezyE
commented
3 years ago Getting even farther with eca39e6a ! With a fresh wineprefix, new dotnet35sp1 install, I run RMS Express and get Winlink Express dialogue box telling me an error happened in kernelbase32.dll. Here's the log1: log1file.log
WheezyE
commented
3 years ago Log2 here: log2file (copy 1).log.zip :)
WheezyE
commented
3 years ago We're getting so close now that .net is functioning.
Winlink appears to error every OTHER time it opens (how odd), but not every time it opens.
I'm still suspicious of ngen. A winlink error dialog box pops up whenever ngen appears in the terminal (".NET Runtime version 2.0.50727.3053").
Trying to install (and override in winecfg) any version of a native iphlpapi.dll from online that I can find gives me a bunch of new errors.
I know ngen tries to do some kind of compiling thing on files so that they run faster the next time they're loaded. Maybe it's caching things (that are broken), crashing when it tries to load its cached files, then just rebuilding from scratch on the next run of the program.
I'm still working on figuring out the best installation conditions, but so far I have:
BOX86_NOBANNER=1 winetricks corefonts dotnet35sp1 vb6run win7 riched30 richtx32
I'm also currently experimenting with other winetricks packages (any related to dll warnings I've seen in the log1 output) to see if any help. So far, I'm trying winetricks comctl32 comctl32ocx comdlg32ocx usp10 gdiplus wininet winhttp secur32 vb6run vcrun2005 (as well as ole32 crypt32 msvcrt40 after renaming the downloaded windowsxp-kb936929...exe to the filename that winetricks is actually looking for - which is also pictured above in the lower terminal window), but I've been getting less stability this way so far.
WheezyE
commented
3 years ago wtf. It appears to not be erroring anymore. Sorry, this whole process is confusing.
After installing all those winetricks packages, I was switching them off/on (by commenting in/out the dll overrides) in ~/.wine/user.reg
I'll try to figure out which packages those are exactly
WheezyE
commented
3 years ago (Sorry to spam - just kinda excited) I think I narrowed this ngen crash down to wine just missing winetricks crypt32. Maybe we just need to wrap crypt32.so?
EDIT: I found out that if I decide to install vb6run (so that winlink's internet works), then I needed to install these two things too winetricks comctl32ocx comdlg32ocx (maybe I just need one of them?)
I do still have all the other dll files in place in my wine system32 folder, so I'll try to see if it's truly just this dll or if Winlink/ngen need others too, but this is pretty exciting to think that ngen woes might just be from missing this crypt32.dll
(ps I'm running Box86 dd733c41, wine_5.21-devel, TwisterOS)
Another edit: I just tried some other combinations of things and wanted to post them here too.
WheezyE
commented
3 years ago With box86 9c5a9ebb when I try to run RMS Express, the program seems to freeze now when trying to open. Log 1 shows this: Warning: Cannot dlopen("/home/pi/wine/lib/wine/shell32.so"/0x7db9d178, 2) at the freeze.
I’ll try to bisect and find the commit. Maybe this is related to the new crashes in ngen/dotnet?
WheezyE
commented
3 years ago That error appears to be a separate issue. I went back to the commit that has libcrypt wrapping and it appears that the wrapping isn’t playing nice with ngen. I’ll bisect that and log1 soon
WheezyE
commented
3 years ago I’m having some trouble today pinning down which commit causes RMS Express to not open anymore. Part of my problem is that I broke test.sh somehow (I think by experimenting with RAM stuff). The other half of my problem is that RMS Express still has that initial crash that pops up and I’m having trouble sorting out maybe two different crashes. I’ll probably focus on dotnet35sp1 installer first and come back to this issue after that’s resolved
WheezyE
commented
3 years ago Just wanted to say I’m excited to have put up a GitHub page to host an install script and hopefully make install easy for the hams I know: https://github.com/WheezyE/Winelink
I’m so excited that VARA runs and am really looking forward to ironing out even more bugs and streamlining the dotnet install process. I’ve been trying to do this for two years and it wouldn’t have been possible to do it open-source without box86 and your help. Lots of work still to be done with Winlink on box86, but this is a pretty amazing proof of concept.
I’ll keep looking for bugs 🙂
WheezyE
commented
3 years ago These crashes took a while to track down since they happen inconsistently (and I got kind of busy this month + side-tracked working on my Winelink script)
RMS Express crashes happen whenever ngen crashes appear in the terminal (when running RMS Express): 0274:err:eventlog:ReportEventW L".NET Runtime version 2.0.50727.3053 - Fatal Execution Engine Error (7A097706) (80131506)"
The .NET ngen.exe that RMS Express relies on is located here: ~/.wine/drive_c/windows/Microsoft.NET/Framework/v2.0.50727/ngen.exe
After installing dotnet35sp1 (commit db5efa89085a085d733c859662799ebcf4e5c3c2 works to install dotnet35sp1), I believe we can test ngen alone by telling it to create a native image of any exe:
cd ~/.wine/drive_c/windows/Microsoft.NET/Framework/v2.0.50727 # must be in this directory to run ngen manually w wine?
wine ngen.exe vbc.exengen stores the native images that it creates in ~/.wine/drive_c/windows/assembly/NativeImages_v2.0.50727_32
ptitSeb
commented
3 years ago Ok! So commit "d3d431d5" introduced some mutex mecanism, so I assume there is a deadlock happening here. I tried to reproduce on my Pandora (can't try on my Pi for now) but I don't get any lock, so it maybe sensitive to timming.
If you can reproduce on your side (with latest box86 sources): once it's stuck, open another terminal, and with ps -A find the pid of "RMS Express.exe" (I assume the deadlock is in there), then attach a gdb to it using gdb -pid PID (with PID the pid number of the process).
In gdb use thread apply all bt to get the back trace in all thread: I need that output.
(you can then kill the app with k and quit gdb with q)
(in the mean time, I'll try to think of what could go wrong, but the backtrace would certainly help)
WheezyE
commented
3 years ago Awesome, alright here's the backtrace of RMS Express with box86 be82ccb. ngen wasn't in the PID list.
The terminal running RMS Express (frozen) reads:
...
0024:fixme:ntdll:EtwRegisterTraceGuidsW register trace class {861f5339-19d6-4873-b350-7b03228bda7c}
0024:fixme:ntdll:EtwRegisterTraceGuidsW (7A030638, 001255C8, {cc2bcbba-16b6-4cf3-8990-d74c2e8af500}, 1, 7A3BEE38, (null), (null), 7A3BDC78): stub
0024:fixme:ntdll:EtwRegisterTraceGuidsW register trace class {ea40c74d-4f65-4561-bb26-656231c8967f}
# stuck hereThanks for the update! Hm the freeze seems to still be happening at the same spot.
0024:fixme:ntdll:EtwRegisterTraceGuidsW register trace class {861f5339-19d6-4873-b350-7b03228bda7c}
0024:fixme:ntdll:EtwRegisterTraceGuidsW (7A030638, 00125610, {cc2bcbba-16b6-4cf3-8990-d74c2e8af500}, 1, 7A3BEE38, (null), (null), 7A3BDC78): stub
0024:fixme:ntdll:EtwRegisterTraceGuidsW register trace class {ea40c74d-4f65-4561-bb26-656231c8967f}I never thought to try this till now, but I turned off dynarec and that got past the freeze in the latest commit (c254dc31) and in the Feb 9 (a062a8ef) commit (though in both cases after turning off Dynarec, after Winlink loads it then has internet errors and always crashes w ngen problems - though these are probably separate issues).
WheezyE
commented
3 years ago Thank you again for your work on this error. The freeze is unfortunately still persisting.
ptitSeb
commented
3 years ago well, it seem the deadlock doesn't happens on main "RMS Express.exe" program, as it is waiting for an answer from another program. Maybe wineserver, maybe another, I cann't tell, be if it was a deadlock, you would see some thread of the backtrace waiting on nanosleep or pthread_mutex_lock.
WheezyE
commented
3 years ago I will say, with dynarec off and with these patches, the other RMS Express crashes seem to be gone now and the internet works again
ptitSeb
commented
3 years ago At least something progress in the right direction! :)
WheezyE
commented
3 years ago I thought maybe I should post a log1 and backtrace of ngen too since .NET's ngen freezes in the same spot as RMS Express in the terminal (I believe RMS Express invokes ngen while opening).
WheezyE
commented
3 years ago I gave gdb another shot today with the C freeze with box86 74d3bb76 and found a reference to "nanosleep" in winedevice.exe
I ran ngen again until freeze and then attached gdb to various processes and searched thread apply all bt. Then I copy-pasted the output into a text file and searched through there for "nanosleep", "pthread_mutex_lock", and "mutex".
WheezyE
commented
3 years ago With 00ac8ec1aa12d8679a619f38c54dc880b504a597 , the freeze persists but nanosleep is gone from the backtrace of winedevice.exe. Nice work! I’ll look for more nanosleeps or mutex errors today in a bit
WheezyE
commented
3 years ago Oh, my mistake: two winedevice.exe processes were run by ngen.
ptitSeb
commented
3 years ago With commit 017d82d16e919acc9e57c35fe850a289ffe722d7 I have changed again the Memory Protection and Dynablock handling.
I think it should have improve the behaviour (both in term of speed and stability) for C# / Mono / Unity3D software, so i think it's worth giving it a new try.
WheezyE
commented
3 years ago Thank you again for the update. I’m unfortunately still seeing the freeze (and the backtrace in winedevice.exe still says nanosleep) with 017d82d1
ptitSeb
commented
3 years ago I don't understand why it would freeze. It's probably not what I thought it was initialy.
Can you test something? In scr/tools/bridge.c, comment all the pthread_mutex_lock(...) and pthread_mutex_unlock(..) you see try again.
WheezyE
commented
3 years ago Thanks! I’ll try this out this afternoon (about 4:30pm MST) after I get back from work
On Sat, Mar 27, 2021 at 7:46 AM ptitSeb @.***> wrote:
I don't understand why it would freeze. It's probably not what I thought it was initialy.
Can you test something? In scr/tools/bridge.c, comment all the pthread_mutex_lock(...) and pthread_mutex_unlock(..) you see try again.
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/ptitSeb/box86/issues/217#issuecomment-808735477, or unsubscribe https://github.com/notifications/unsubscribe-auth/AHCGVPWZ3HN6ENCWQYPXGXLTFXOUHANCNFSM4SECK3RQ .
WheezyE
commented
3 years ago Alright, I made a fork of the box86 repo, commented out two lines, and did the box86 build instructions but with the fork under my github name. I'm still getting the freeze and same nanosleep backtrace in winedevice.exe.
I re-tested my commit bisect too just to make sure it's reproducible (6409f5d4 doesn't have the freeze, d3d431d5 still has the freeze). I also tested the latest 9bf5c20a9a2bb576f6436aee678c3c87768b992e and still had the freeze.
ptitSeb
commented
3 years ago What is strange is that commit d3d431d5b1cbe0fd051a9a2560ae75748a854ced basicaly added a mutex around Memory Protection access, and with commit 017d82d16e919acc9e57c35fe850a289ffe722d7 I completly removed it. So the freeze must be something else now.
WheezyE
commented
3 years ago I'll try to make another box86 fork and try going through different commits while patching out d3d431d on each commit along the way to see if that helps. I got kind of busy yesterday, but I'll see if I can work on this today and/or tomorrow
WheezyE
commented
3 years ago I guess I'm kinda stuck lol. I'll try to circle back around to this
WheezyE
commented
3 years ago Thanks again for trying to get this running. I really appreciate it! I built bd48cae19f32dc18b0c94b4af7a0bbd684bdef07 and ran it, restarted me pi, ran again, but still getting the same freeze in RMS Express and ngen.
I’ve been trying to fork and then branch box86 from the commit before d3d431d5b1cbe0fd051a9a2560ae75748a854ced to try to see if adding your code changes (not including d3d431d) would allow us to find which other commit might be causing the freeze in ngen, but I’ve been running into problems with just not knowing how to use GitHub correctly
bbbruni
commented
3 years ago Hello!
I had a similar problem with a hidden buggy commit in wine a few years ago. There was a game that had worked correctly before a wine developer introduced some regression. I created a bug report on wine bugzilla. On the bug report page, that developer said he had committed a fix before I created the report. Then, we found out that the fix had not solved the regression because there was another buggy commit between those two. There was another commit he had accidentally overlooked.
See https://bugs.winehq.org/show_bug.cgi?id=28491#c6 and below.
We determined the right buggy commit there eventually and the developer fixed the regression.
ptitSeb says that the problematic d3d431d5b1cbe0fd051a9a2560ae75748a854ced was cancelled by 017d82d16e919acc9e57c35fe850a289ffe722d7 here.
So you may try to do what we did on wine bugzilla.
Download box86 into a new location, then make sure you use fresh wine prefix and run:
git clone https://github.com/ptitSeb/box86
git checkout d3d431d5b1cbe0fd051a9a2560ae75748a854ced git cherry-pick 017d82d16e919acc9e57c35fe850a289ffe722d7
cd box86 mkdir build; cd build; cmake .. -DRPI4=1 -DCMAKE_BUILD_TYPE=RelWithDebInfo make -j4 sudo make install sudo systemctl restart systemd-binfmt
Say if you observe the freeze after that. Regards.
WheezyE
commented
3 years ago I super appreciate the help and advice. I'd heard of cherry-picking but wasn't really sure of the use-case for it. When I cherry pick in this case though, I'm running into some conflicts with the upstream "Removed lock on Memory Protection," commit (I'll see if I can resolve the conflicts, but I'm kind of a C noob). This gives me some ideas though for maybe trying other functions like revert?
EDIT: I tried deleting all my box86 build folders again (like you suggested to start from scratch) and now github is letting me build from my bugfinder branch of my box86 fork. Not sure why this is the case, but I've been able to patch out d3d431d and continue adding Seb's other commits sequentially. (My builds seem to not be freezing so far without d3d431d, but I'll keep adding commits to see if there's a second freeze point).
Thank you both again!
bbbruni
commented
3 years ago Do you have prior experience in bisecting git repos?
WheezyE
commented
3 years ago I've heard of the general process, though I haven't actually used git to do it. I usually "bisect" manually using sort of like a "method of limits" (for lack of a better term) by gradually decreasing my search range before and after a bug I've found.
etc generally narrowing my search down. I think this is what git does too though just with a script.
WheezyE
commented
3 years ago I've been adding old commits by hand one at a time to my fork/branch of box86 (my branch starts at 6409f5d40a26f115d2c0034f25d7b5cbb7131048 and doesn't have d3d431d, but has later commits), but I hit a roadblock with adding the 4e5473150a9e76d427c3946a115f8ede554579b5 commit (seems like I broke box86 somehow). I'll try to reconcile the differences, but I might be stuck again.
So far, I can say that Box86 loads RMS express if d3d431d is patched out of commits between d3d431d and 4e5473150a9e76d427c3946a115f8ede554579b5.
bbbruni
commented
3 years ago Okay. You had dynarec enabled all the way, right?
Then clean up wine and box86 and git directories (paths may vary):
cd box86 rm -rf .git rm -rf .wine cd .. && rm -rf box86
Then run:
git clone https://github.com/ptitSeb/box86
git checkout 4e54731 git cherry-pick c254dc3121dea773677d4fcbbe204a497a009aa5
cd box86 mkdir build; cd build; cmake .. -DRPI4=1 -DCMAKE_BUILD_TYPE=RelWithDebInfo make -j4 sudo make install sudo systemctl restart systemd-binfmt
Say if you observe a broken compilation or broken merge or freeze after that.
EDIT: MY BAD.
git clone https://github.com/ptitSeb/box86
git checkout 4e54731 git cherry-pick c254dc3121dea773677d4fcbbe204a497a009aa5
cd box86 mkdir build; cd build; cmake .. -DRPI4=1 -DCMAKE_BUILD_TYPE=RelWithDebInfo make -j4 sudo make install sudo systemctl restart systemd-binfmt
Say if you observe a broken compilation or broken merge or freeze after that.
Then
Then run:
git clone https://github.com/ptitSeb/box86
git checkout 4e54731 git cherry-pick c254dc3121dea773677d4fcbbe204a497a009aa5 git cherry-pick 017d82d
cd box86 mkdir build; cd build; cmake .. -DRPI4=1 -DCMAKE_BUILD_TYPE=RelWithDebInfo make -j4 sudo make install sudo systemctl restart systemd-binfmt
Say if you observe a broken compilation or broken merge or freeze after that.
If that doesn't help, we will try doing bisect.
You have to do two consequent test: 1st with one cherry-pick c254dc3121dea773677d4fcbbe204a497a009aa5
Then clean up all
2nd with two consequent cherry-picks cherry-pick c254dc3121dea773677d4fcbbe204a497a009aa5 && cherry-pick 017d82d
Please write if you understood it. Regards.
WheezyE
commented
3 years ago I never did try the cherry picks since my previous experience with them ran into ‘collisions’ with code and since I’d been kind of already doing my own version of cherry-picking with my own branch of box86, but I did appreciate the help and ideas with those. I kind of got frustrated and decided to take a break with Winlink for a bit. Thank you though for the suggestions
WheezyE
commented
3 years ago So, after seeing on the discord that people were installing .net 4+ successfully, I decided to reflash my Pi4 and start over in case my testing environment was borked somehow. I updated box86 and was successfully able to install dotnet35sp1 with no ngen errors!! I did have to turn off dynarec and install was 90+ minutes. I’m wondering if this had been fixed earlier and it was just errors I was encountering in my old wineprefix (I had installed dotnet with an older box86 commit and then upgraded to test to see if ngen would register files or if Winlink would run with newer box86 commits).
I can’t believe ngen isn’t hanging anymore. I am encountering some fresh errors though: A freeze happens with dynarec on.
ptitSeb
commented
3 years ago The "Library libXXXX needs YYYYY, but it was not found" is ok, that's part of the new "quit" protocol...
The SIGSEGV with code=2 and prot!=0 can be normal, especially with .NET stuff. A Double SIGSEGV should not happens, I'll check that. Can you do addr2line -e /usr/local/bin/box86 -a 0x62872294 with the same box86 build that was used with the trace?
WheezyE
commented
3 years ago Thank again you for looking into it. Here are those addresses
pi@raspberrypi:~/.wine/drive_c/RMS Express $ addr2line -e /usr/local/bin/box86 -a 0x62872294
0x62872294
/home/pi/box86/box86/src/emu/x86run.c:673 (discriminator 22)
pi@raspberrypi:~/.wine/drive_c/RMS Express $ addr2line -e /usr/local/bin/box86 -a 0x6286dd70
0x6286dd70
/home/pi/box86/box86/src/emu/x86run.c:192
pi@raspberrypi:~/.wine/drive_c/RMS Express $ addr2line -e /usr/local/bin/box86 -a 0x62873bc4
0x62873bc4
/home/pi/box86/box86/src/emu/x86run.c:899
pi@raspberrypi:~/.wine/drive_c/RMS Express $ addr2line -e /usr/local/bin/box86 -a 0x62875330
0x62875330
/home/pi/box86/box86/src/emu/x86run.c:1570In case it helps, I ran this again and got this this time:
Library libdl.so.2 (0x6396aa28) needs 0x6396b5a0, but it was not found. Ignoring.
Stop waiting for remaining thread 19896
Stop waiting for remaining thread 19897
19897|SIGSEGV @0x62875330 (???(/usr/local/bin/box86/0x62875330)) (x86pc=0x7bc8022f/�M�cp��c:"???", esp=0x323e840, stack=0x3140000:0x3244000 own=(nil) fp=0x323e858), for accessing 0x405b66a8 (code=1/prot=7), db=(nil)((nil):(nil)/(nil):(nil)/???:clean, hash:0/0)
19896|SIGSEGV @0x62873bc4 (???(/usr/local/bin/box86/0x62873bc4)) (x86pc=0x40580394/???:"???", esp=0x108dfd0, stack=0xf90000:0x1094000 own=(nil) fp=0x108dff8), for accessing 0x40580394 (code=1/prot=7), db=(nil)((nil):(nil)/(nil):(nil)/???:clean, hash:0/0)
19896|SIGSEGV @0x6286dd70 (???(/usr/local/bin/box86/0x6286dd70)) (x86pc=0x40583580/???:"???", esp=0x3fff7ca0, stack=0xf90000:0x1094000 own=(nil) fp=0x108dff8), for accessing 0x40583580 (code=1/prot=7), db=(nil)((nil):(nil)/(nil):(nil)/???:clean, hash:0/0)
19896|Double SIGSEGV (code=1, pc=0x6286dd70, addr=0x40583580)!Alright, new commits, more progress, another hurdle: Working from b588709ad; Methods and logs below.
dotnet35sp1 (fresh wineprefixes)
ntdll:NtQuerySystemInformation info_class SYSTEM_PERFORMANCE_INFORMATION). 60 min install. Log1: dotnet35sp1_dyn.txtWinlink (using dotnet35sp1 successfully installed prefix above - no other winetricks components were installed) still freezing - I believe from ngen failing to register native images
v2 ngen.exe (using dotnet35sp1 successfully installed prefix above) - freezes
pi@raspberrypi:~/.wine/drive_c/windows/Microsoft.NET/Framework/v2.0.50727 $ BOX86_LOG=1 BOX86_DYNAREC=0 wine ngen.exe vbc.exe 2>&1 | tee ~/ngen_nodyn.txtWheezyE
commented
3 years ago Ah the ngen.exe & winlink crashes both begin with 0x62877234 sigsegv's
pi@raspberrypi:~ $ addr2line -e /usr/local/bin/box86 -a 0x62877234
0x62877234
/home/pi/box86/box86/src/emu/x86run.c:673 (discriminator 22)
pi@raspberrypi:~ $
Seb, thank you again for all your hard work with this finicky program. The commits from today (commits near c7a4dbe I think?) are getting RMS Express to open (and connect to Telnet!) with mono - which is farther than ExaGear got using mono.
I did some analysis today using mono vs using dotnet35sp1+vcrun2015 (from my pre-loaded dotnet35sp1+vcrun2015 wineprefix from my Debian 10 x86 VMWare virtual computer on my Windows laptop copied to my Pi --- vcrun2015 and dotnet35sp1 still crash when trying to install in box86 on the Pi).
Total packages needed for RMS Express and VARA to work (as learned from a functioning/clunky wine+ExaGear install) are:
winetricks -q corefonts dotnet35sp1 vb6run vcrun2015 sound=alsa win7 # VARA also needs an old WinNT4.0 pdh.dllEDIT 8/1/2021: More testing has revealed that these are the only packages needed for each program ...
BOX86_NOBANNER=1 winetricks -q dotnet35sp1 win7 sound=alsa# for RMS Express (corefonts & vcrun2015 appear to be optional)BOX86_NOBANNER=1 winetricks -q vb6run pdh_nt4 win7 sound=alsa# for VARAOverview
With mono (not dotnet35sp1) installed, RMS Express opens(!) and the ARDOP and WINMOR modems open. But ARDOP and WINMOR don't function correctly (I learned from ExaGear that they need dotnet35sp1 to function correctly, VARA needs vcrun2015). With dotnet35sp1 installed (instead of mono), I get SIGSEGV's running RMS Express, ARDOP, WINMOR, and VARA. VARA installs, but doesn't run with mono nor with dotnet35sp1.
RMS Express
RMS Express w/ mono: Runs!!! No problems! Telnet works! modems don't work though
RMS Express w/ dotnet35sp1:
``` Using native(wrapped) libtinfo.so.6 0024:fixme:ntdll:NtQuerySystemInformation info_class SYSTEM_PERFORMANCE_INFORMATION Using emulated /home/pi/wine/lib/wine/advapi32.dll.so Using emulated /home/pi/wine/lib/wine/sechost.dll.so Using emulated /home/pi/wine/lib/wine/ucrtbase.dll.so Using emulated /home/pi/wine/lib/wine/api-ms-win-core-synch-l1-2-0.dll.so Using emulated /home/pi/wine/lib/wine/api-ms-win-core-fibers-l1-1-1.dll.so Using emulated /home/pi/wine/lib/wine/api-ms-win-core-localization-l1-2-1.dll.so 0024:fixme:ntdll:EtwEventRegister ({319dc449-ada5-50f7-428e-957db6791668}, 109BA1A0, 10A12460, 10A12478) stub. 0024:fixme:ntdll:EtwEventSetInformation (deadbeef, 2, 10993241, 28) stub Using emulated /home/pi/wine/lib/wine/shlwapi.dll.so Using emulated /home/pi/wine/lib/wine/shcore.dll.so Using emulated /home/pi/wine/lib/wine/user32.dll.so Using emulated /home/pi/wine/lib/wine/setupapi.dll.so Using emulated /home/pi/wine/lib/wine/version.dll.so Using emulated /home/pi/wine/lib/wine/rpcrt4.dll.so Using emulated /home/pi/wine/lib/wine/gdi32.dll.so Using emulated /home/pi/wine/lib/wine/ole32.dll.so Using native(wrapped) libfreetype.so.6 Using native(wrapped) libfontconfig.so.1 Using emulated /home/pi/wine/lib/wine/imm32.dll.so Using emulated /home/pi/wine/lib/wine/api-ms-win-appmodel-runtime-l1-1-2.dll.so 0024:fixme:ntdll:EtwEventRegister ({319dc449-ada5-50f7-428e-957db6791668}, 10C1B8F0, 11161058, 11161070) stub. 0024:fixme:ntdll:EtwEventSetInformation (deadbeef, 2, 10A309AD, 28) stub Using emulated /home/pi/wine/lib/wine/api-ms-win-core-quirks-l1-1-0.dll.so 0024:fixme:kernelbase:QuirkIsEnabled3 (1020F4E0, FFFFFFFF) stub! 0024:fixme:heap:GetNumaHighestNodeNumber semi-stub: 1020FC14 0024:fixme:ntdll:EtwEventRegister ({e13c0d23-ccbc-4e12-931b-d9cc2eee27e4}, 10D54330, 11161100, 1116C108) stub. 0024:fixme:ntdll:EtwEventRegister ({763fd754-7086-4dfe-95eb-c01a46faf4ca}, 10D54330, 11162258, 111622C8) stub. 0024:fixme:ntdll:EtwEventRegister ({a669021c-c450-4609-a035-5af59af4df18}, 10D54330, 1116AC30, 1116C118) stub. 0024:fixme:ntdll:EtwEventRegister ({cc2bcbba-16b6-4cf3-8990-d74c2e8af500}, 10D54330, 11161220, 11161288) stub. Using emulated /home/pi/wine/lib/wine/combase.dll.so 0024:fixme:seh:WerRegisterRuntimeExceptionModule (L"C:\\windows\\Microsoft.NET\\Framework\\v4.0.30319\\mscordacwks.dll", 10A20000) stub 23702|SIGSEGV @0x7cc5c264 (???) (x86pc=0x10bd0e63/???:"???"), for accessing 0x13640004 (code=2), db=0x59bf3f0(0x10bd0e63/???) Warning, context has been changed in Sigactionhanlder (EIP changed) 23702|Double SIGSEGV! Segmentation fault pi@raspberrypi:~/.wine/drive_c/RMS Express $ 00a0:err:rpc:I_RpcReceive we got fault packet with status 0x1c010003 ```RMS Express w/ dotnet35sp1+vcrun2015:
``` 00b8:fixme:ntdll:NtQuerySystemInformation info_class SYSTEM_PERFORMANCE_INFORMATION 00c4:fixme:process:SetProcessShutdownParameters (00000380, 00000000): partial stub. Using emulated /home/pi/wine/lib/wine/kernel32.dll.so Using emulated /home/pi/wine/lib/wine/kernelbase.dll.so Using emulated /home/pi/wine/lib/wine/mscoree.dll.so Using emulated /home/pi/wine/lib/wine/dbghelp.dll.so Using emulated /home/pi/wine/lib/wine/ucrtbase.dll.so Using emulated /home/pi/wine/lib/wine/shell32.dll.so Using emulated /home/pi/wine/lib/wine/shlwapi.dll.so Using emulated /home/pi/wine/lib/wine/shcore.dll.so Using emulated /home/pi/wine/lib/wine/user32.dll.so Using emulated /home/pi/wine/lib/wine/setupapi.dll.so Using emulated /home/pi/wine/lib/wine/version.dll.so Using emulated /home/pi/wine/lib/wine/advapi32.dll.so Using emulated /home/pi/wine/lib/wine/sechost.dll.so Using emulated /home/pi/wine/lib/wine/rpcrt4.dll.so Using emulated /home/pi/wine/lib/wine/gdi32.dll.so Using emulated /home/pi/wine/lib/wine/ole32.dll.so Using native(wrapped) libncurses.so.6 Using native(wrapped) libtinfo.so.6 0024:fixme:ntdll:NtQuerySystemInformation info_class SYSTEM_PERFORMANCE_INFORMATION Using native(wrapped) libfreetype.so.6 Using native(wrapped) libfontconfig.so.1 Using emulated /home/pi/wine/lib/wine/imm32.dll.so Using emulated /home/pi/wine/lib/wine/mscorwks.dll.so Using emulated /home/pi/wine/lib/wine/msvcr80.dll.so 0024:fixme:ntdll:EtwRegisterTraceGuidsW (7A03261A, 1090D0C0, {e13c0d23-ccbc-4e12-931b-d9cc2eee27e4}, 14, 7A3BEE40, (null), (null), 7A3BDC70): stub 0024:fixme:ntdll:EtwRegisterTraceGuidsW register trace class {044973cd-251f-4dff-a3e9-9d6307286b05} 0024:fixme:ntdll:EtwRegisterTraceGuidsW register trace class {3044f61a-99b0-4c21-b203-d39423c73b00} 0024:fixme:ntdll:EtwRegisterTraceGuidsW register trace class {d00792da-07b7-40f5-97eb-5d974e054740} 0024:fixme:ntdll:EtwRegisterTraceGuidsW register trace class {cfc4ba53-fb42-4757-8b70-5f5d51fee2f4} 0024:fixme:ntdll:EtwRegisterTraceGuidsW register trace class {efdf1eac-1d5d-4e84-893a-19b80f692176} 0024:fixme:ntdll:EtwRegisterTraceGuidsW register trace class {c71408de-42cc-4f81-9c93-b8912abf2a0f} 0024:fixme:ntdll:EtwRegisterTraceGuidsW register trace class {840c8456-6457-4eb7-9cd0-d28f01c64f5e} 0024:fixme:ntdll:EtwRegisterTraceGuidsW register trace class {c424b3e3-2ae0-416e-a039-410c5d8e5f14} 0024:fixme:ntdll:EtwRegisterTraceGuidsW register trace class {300ce105-86d1-41f8-b9d2-83fcbff32d99} 0024:fixme:ntdll:EtwRegisterTraceGuidsW register trace class {561410f5-a138-4ab3-945e-516483cddfbc} 0024:fixme:ntdll:EtwRegisterTraceGuidsW register trace class {15447a14-b523-46ae-b75b-023f900b4393} 0024:fixme:ntdll:EtwRegisterTraceGuidsW register trace class {b17304d9-5afa-4da6-9f7b-5a4fa73129b6} 0024:fixme:ntdll:EtwRegisterTraceGuidsW register trace class {02d08a4f-fd01-4538-989b-03e437b950f4} 0024:fixme:ntdll:EtwRegisterTraceGuidsW register trace class {861f5339-19d6-4873-b350-7b03228bda7c} 0024:fixme:ntdll:EtwRegisterTraceGuidsW (7A030638, 1090D0C0, {cc2bcbba-16b6-4cf3-8990-d74c2e8af500}, 1, 7A3BEE38, (null), (null), 7A3BDC78): stub 0024:fixme:ntdll:EtwRegisterTraceGuidsW register trace class {ea40c74d-4f65-4561-bb26-656231c8967f} 23988|SIGSEGV @0x7c11bd58 (???) (x86pc=0x79f0a4f0/???:"???"), for accessing 0x12c30008 (code=2), db=0x4891230(0x79f0a4f0/???) Warning, context has been changed in Sigactionhanlder (EIP changed) 23988|Double SIGSEGV! Segmentation fault ```VARA
VARA install (still needs vcrun2015):
``` DYNAREC=0 wine VARA\ setup\ \(Run\ as\ Administrator\).exe BOX86_DYNAREC=0 BOX86_NOSIGSEGV=1 winetricks msxml3 BOX86_DYNAREC=0 winetricks -q vb6run BOX86_DYNAREC=0 wine ~/.wine/drive_c/VARA/VARA\ Install.exe #seems(?) to install mswinsck.ocx with a sigsegv #note: I think vcrun2015 is also a requirement for VARA but vcrun2015 doesn't install with box86 yet. BOX86_DYNAREC=0 wine ~/.wine/drive_c/VARA/VARA.exe # crashes with or without DYNAREC enabled ```VARA (with and without Dynarec) w/ mono (no vcrun2015):
``` Using emulated /home/pi/wine/lib/wine/sxs.dll.so 0024:fixme:olepicture:OleLoadPictureEx (10CB0DD4,7366,0,{7bf80980-bf32-101a-8bbb-00aa00300cab},x=0,y=0,f=0,1020FAB0), partially implemented. Using emulated /home/pi/wine/lib/wine/usp10.dll.so Using emulated /home/pi/wine/lib/wine/wsock32.dll.so Using emulated /home/pi/wine/lib/wine/ws2_32.dll.so Using emulated /home/pi/wine/lib/wine/iphlpapi.dll.so Using native(wrapped) libresolv.so.2 23790|SIGSEGV @0xb5b4370c (???) (x86pc=0x7b630df0//home/pi/wine/lib/wine/kernel32.dll.so:"/home/pi/wine/lib/wine/ntdll.dll.so/wcslen"), for accessing (nil) (code=1), db=0x3d611e0(0x7bc88ce0//home/pi/wine/lib/wine/ntdll.dll.so/wcslen) Warning, context has been changed in Sigactionhanlder (EIP changed) 23790|Double SIGSEGV! Segmentation fault ```VARA (with and without dynarec) w/ dotnet35sp1 (no vcrun2015):
``` Using emulated /home/pi/wine/lib/wine/sxs.dll.so 0024:fixme:olepicture:OleLoadPictureEx (10CB0DD4,7366,0,{7bf80980-bf32-101a-8bbb-00aa00300cab},x=0,y=0,f=0,1020FAB0), partially implemented. Using emulated /home/pi/wine/lib/wine/usp10.dll.so Using emulated /home/pi/wine/lib/wine/wsock32.dll.so Using emulated /home/pi/wine/lib/wine/ws2_32.dll.so Using emulated /home/pi/wine/lib/wine/iphlpapi.dll.so Using native(wrapped) libresolv.so.2 23644|SIGSEGV @0x284fb74 (???) (x86pc=0xffffffff/???:"???"), for accessing (nil) (code=1), db=(nil)((nil)/???) Warning, context has been changed in Sigactionhanlder (EIP changed) 23644|Double SIGSEGV! Segmentation fault ```VARA (with and without dynarec) w/ dotnet35sp1+vcrun2015:
``` Using emulated /home/pi/wine/lib/wine/sxs.dll.so 0024:fixme:olepicture:OleLoadPictureEx (10CB0DD4,7366,0,{7bf80980-bf32-101a-8bbb-00aa00300cab},x=0,y=0,f=0,1020FAB0), partially implemented. Using emulated /home/pi/wine/lib/wine/usp10.dll.so Using emulated /home/pi/wine/lib/wine/wsock32.dll.so Using emulated /home/pi/wine/lib/wine/ws2_32.dll.so Using emulated /home/pi/wine/lib/wine/iphlpapi.dll.so Using native(wrapped) libresolv.so.2 24059|SIGSEGV @0xb5bad70c (???) (x86pc=0x7b630df0//home/pi/wine/lib/wine/kernel32.dll.so:"/home/pi/wine/lib/wine/ntdll.dll.so/wcslen"), for accessing (nil) (code=1), db=0x4af4218(0x7bc88ce0//home/pi/wine/lib/wine/ntdll.dll.so/wcslen) Warning, context has been changed in Sigactionhanlder (EIP changed) 24059|Double SIGSEGV! Segmentation fault pi@raspberrypi:~/.wine/drive_c/VARA $ 009c:err:rpc:I_RpcReceive we got fault packet with status 0x1c010003 00c4:fixme:advapi:RegisterEventSourceW ((null),L".NET Runtime Optimization Service"): stub 00c4:fixme:advapi:DeregisterEventSource (0xcafe4242) stub 00b8:fixme:msvcrt:__clean_type_info_names_internal (0x64082a5c) stub ```WINMOR (EDIT 8/1/2021: WINMOR has been discontinued)
WINMOR w/ mono: Runs! Doesn't function correctly:
wine WINMOR\ TNC.exe #w/ dotnet35sp1 ``` 0024:fixme:ntdll:EtwEventRegister ({319dc449-ada5-50f7-428e-957db6791668}, 10C1B8F0, 11161058, 11161070) stub. 0024:fixme:ntdll:EtwEventSetInformation (deadbeef, 2, 10A309AD, 28) stub Using emulated /home/pi/wine/lib/wine/api-ms-win-core-quirks-l1-1-0.dll.so 0024:fixme:kernelbase:QuirkIsEnabled3 (1020F4E0, FFFFFFFF) stub! 0024:fixme:heap:GetNumaHighestNodeNumber semi-stub: 1020FC14 0024:fixme:ntdll:EtwEventRegister ({e13c0d23-ccbc-4e12-931b-d9cc2eee27e4}, 10D54330, 11161100, 1116C108) stub. 0024:fixme:ntdll:EtwEventRegister ({763fd754-7086-4dfe-95eb-c01a46faf4ca}, 10D54330, 11162258, 111622C8) stub. 0024:fixme:ntdll:EtwEventRegister ({a669021c-c450-4609-a035-5af59af4df18}, 10D54330, 1116AC30, 1116C118) stub. 0024:fixme:ntdll:EtwEventRegister ({cc2bcbba-16b6-4cf3-8990-d74c2e8af500}, 10D54330, 11161220, 11161288) stub. Using emulated /home/pi/wine/lib/wine/combase.dll.so 0024:fixme:seh:WerRegisterRuntimeExceptionModule (L"C:\\windows\\Microsoft.NET\\Framework\\v4.0.30319\\mscordacwks.dll", 10A20000) stub 20973|SIGSEGV @0x7cc5d064 (???) (x86pc=0x10bd0e63/???:"???"), for accessing 0x13640004 (code=2), db=0x4a23428(0x10bd0e63/???) Warning, context has been changed in Sigactionhanlder (EIP changed) 20973|Double SIGSEGV! Segmentation fault ```WINMOR w/ dotnet35sp1+vcrun2015 (WINMOR doesn't need vcrun2015 to function though):
``` 00b8:fixme:ntdll:NtQuerySystemInformation info_class SYSTEM_PERFORMANCE_INFORMATION 00c4:fixme:process:SetProcessShutdownParameters (00000380, 00000000): partial stub. Using emulated /home/pi/wine/lib/wine/kernel32.dll.so Using emulated /home/pi/wine/lib/wine/kernelbase.dll.so Using emulated /home/pi/wine/lib/wine/mscoree.dll.so Using emulated /home/pi/wine/lib/wine/dbghelp.dll.so Using emulated /home/pi/wine/lib/wine/ucrtbase.dll.so Using emulated /home/pi/wine/lib/wine/shell32.dll.so Using emulated /home/pi/wine/lib/wine/shlwapi.dll.so Using emulated /home/pi/wine/lib/wine/shcore.dll.so Using emulated /home/pi/wine/lib/wine/user32.dll.so Using emulated /home/pi/wine/lib/wine/setupapi.dll.so Using emulated /home/pi/wine/lib/wine/version.dll.so Using emulated /home/pi/wine/lib/wine/advapi32.dll.so Using emulated /home/pi/wine/lib/wine/sechost.dll.so Using emulated /home/pi/wine/lib/wine/rpcrt4.dll.so Using emulated /home/pi/wine/lib/wine/gdi32.dll.so Using emulated /home/pi/wine/lib/wine/ole32.dll.so Using native(wrapped) libncurses.so.6 Using native(wrapped) libtinfo.so.6 0024:fixme:ntdll:NtQuerySystemInformation info_class SYSTEM_PERFORMANCE_INFORMATION Using native(wrapped) libfreetype.so.6 Using native(wrapped) libfontconfig.so.1 Using emulated /home/pi/wine/lib/wine/imm32.dll.so Using emulated /home/pi/wine/lib/wine/mscorwks.dll.so Using emulated /home/pi/wine/lib/wine/msvcr80.dll.so 0024:fixme:ntdll:EtwRegisterTraceGuidsW (7A03261A, 1090D070, {e13c0d23-ccbc-4e12-931b-d9cc2eee27e4}, 14, 7A3BEE40, (null), (null), 7A3BDC70): stub 0024:fixme:ntdll:EtwRegisterTraceGuidsW register trace class {044973cd-251f-4dff-a3e9-9d6307286b05} 0024:fixme:ntdll:EtwRegisterTraceGuidsW register trace class {3044f61a-99b0-4c21-b203-d39423c73b00} 0024:fixme:ntdll:EtwRegisterTraceGuidsW register trace class {d00792da-07b7-40f5-97eb-5d974e054740} 0024:fixme:ntdll:EtwRegisterTraceGuidsW register trace class {cfc4ba53-fb42-4757-8b70-5f5d51fee2f4} 0024:fixme:ntdll:EtwRegisterTraceGuidsW register trace class {efdf1eac-1d5d-4e84-893a-19b80f692176} 0024:fixme:ntdll:EtwRegisterTraceGuidsW register trace class {c71408de-42cc-4f81-9c93-b8912abf2a0f} 0024:fixme:ntdll:EtwRegisterTraceGuidsW register trace class {840c8456-6457-4eb7-9cd0-d28f01c64f5e} 0024:fixme:ntdll:EtwRegisterTraceGuidsW register trace class {c424b3e3-2ae0-416e-a039-410c5d8e5f14} 0024:fixme:ntdll:EtwRegisterTraceGuidsW register trace class {300ce105-86d1-41f8-b9d2-83fcbff32d99} 0024:fixme:ntdll:EtwRegisterTraceGuidsW register trace class {561410f5-a138-4ab3-945e-516483cddfbc} 0024:fixme:ntdll:EtwRegisterTraceGuidsW register trace class {15447a14-b523-46ae-b75b-023f900b4393} 0024:fixme:ntdll:EtwRegisterTraceGuidsW register trace class {b17304d9-5afa-4da6-9f7b-5a4fa73129b6} 0024:fixme:ntdll:EtwRegisterTraceGuidsW register trace class {02d08a4f-fd01-4538-989b-03e437b950f4} 0024:fixme:ntdll:EtwRegisterTraceGuidsW register trace class {861f5339-19d6-4873-b350-7b03228bda7c} 0024:fixme:ntdll:EtwRegisterTraceGuidsW (7A030638, 1090D070, {cc2bcbba-16b6-4cf3-8990-d74c2e8af500}, 1, 7A3BEE38, (null), (null), 7A3BDC78): stub 0024:fixme:ntdll:EtwRegisterTraceGuidsW register trace class {ea40c74d-4f65-4561-bb26-656231c8967f} 24177|SIGSEGV @0x7c11e458 (???) (x86pc=0x79f0a4f0/???:"???"), for accessing 0x12c30008 (code=2), db=0x4f0d7b8(0x79f0a4f0/???) Warning, context has been changed in Sigactionhanlder (EIP changed) 24177|Double SIGSEGV! Segmentation fault pi@raspberrypi:~/.wine/drive_c/RMS Express $ 00b8:fixme:msvcrt:__clean_type_info_names_internal (0x64082a5c) stub ```ARDOP
ARDOP w/ mono: Runs! Doesn't function correctly:
wine ARDOP_Win.exe #w/ dotnet35sp1 ``` 0024:fixme:ntdll:NtQuerySystemInformation info_class SYSTEM_PERFORMANCE_INFORMATION Using emulated /home/pi/wine/lib/wine/advapi32.dll.so Using emulated /home/pi/wine/lib/wine/sechost.dll.so Using emulated /home/pi/wine/lib/wine/ucrtbase.dll.so Using emulated /home/pi/wine/lib/wine/api-ms-win-core-synch-l1-2-0.dll.so Using emulated /home/pi/wine/lib/wine/api-ms-win-core-fibers-l1-1-1.dll.so Using emulated /home/pi/wine/lib/wine/api-ms-win-core-localization-l1-2-1.dll.so 0024:fixme:ntdll:EtwEventRegister ({319dc449-ada5-50f7-428e-957db6791668}, 109BA1A0, 10A12460, 10A12478) stub. 0024:fixme:ntdll:EtwEventSetInformation (deadbeef, 2, 10993241, 28) stub Using emulated /home/pi/wine/lib/wine/shlwapi.dll.so Using emulated /home/pi/wine/lib/wine/shcore.dll.so Using emulated /home/pi/wine/lib/wine/user32.dll.so Using emulated /home/pi/wine/lib/wine/setupapi.dll.so Using emulated /home/pi/wine/lib/wine/version.dll.so Using emulated /home/pi/wine/lib/wine/rpcrt4.dll.so Using emulated /home/pi/wine/lib/wine/gdi32.dll.so Using emulated /home/pi/wine/lib/wine/ole32.dll.so Using native(wrapped) libfreetype.so.6 Using native(wrapped) libfontconfig.so.1 Using emulated /home/pi/wine/lib/wine/imm32.dll.so Using emulated /home/pi/wine/lib/wine/api-ms-win-appmodel-runtime-l1-1-2.dll.so 0024:fixme:ntdll:EtwEventRegister ({319dc449-ada5-50f7-428e-957db6791668}, 10C1B8F0, 11161058, 11161070) stub. 0024:fixme:ntdll:EtwEventSetInformation (deadbeef, 2, 10A309AD, 28) stub Using emulated /home/pi/wine/lib/wine/api-ms-win-core-quirks-l1-1-0.dll.so 0024:fixme:kernelbase:QuirkIsEnabled3 (1020F4E0, FFFFFFFF) stub! 0024:fixme:heap:GetNumaHighestNodeNumber semi-stub: 1020FC14 0024:fixme:ntdll:EtwEventRegister ({e13c0d23-ccbc-4e12-931b-d9cc2eee27e4}, 10D54330, 11161100, 1116C108) stub. 0024:fixme:ntdll:EtwEventRegister ({763fd754-7086-4dfe-95eb-c01a46faf4ca}, 10D54330, 11162258, 111622C8) stub. 0024:fixme:ntdll:EtwEventRegister ({a669021c-c450-4609-a035-5af59af4df18}, 10D54330, 1116AC30, 1116C118) stub. 0024:fixme:ntdll:EtwEventRegister ({cc2bcbba-16b6-4cf3-8990-d74c2e8af500}, 10D54330, 11161220, 11161288) stub. Using emulated /home/pi/wine/lib/wine/combase.dll.so 0024:fixme:seh:WerRegisterRuntimeExceptionModule (L"C:\\windows\\Microsoft.NET\\Framework\\v4.0.30319\\mscordacwks.dll", 10A20000) stub 21051|SIGSEGV @0x7cc59b64 (???) (x86pc=0x10bd0e63/???:"???"), for accessing 0x13640004 (code=2), db=0x51a7df0(0x10bd0e63/???) Warning, context has been changed in Sigactionhanlder (EIP changed) 21051|Double SIGSEGV! Segmentation fault pi@raspberrypi:~/.wine/drive_c/RMS Express $ 009c:err:rpc:I_RpcReceive we got fault packet with status 0x1c010003 ```