[Bug 19931] New: winedbg crashes entering an infinite self-launch loops

http://bugs.winehq.org/show_bug.cgi?id=19931

Summary: winedbg crashes entering an infinite self-launch loops
           Product: Wine
           Version: 1.1.29
          Platform: PC
        OS/Version: Linux
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: winedbg
        AssignedTo: [e-mail]
        ReportedBy: [e-mail]

When some app crashes winedbg starts, but it crashes itself causing to start
another winedbg and so on.

[Bug 19931] winedbg crashes entering an infinite self-launch loops

http://bugs.winehq.org/show_bug.cgi?id=19931

--- Comment #1 from Dmitry Timoshkov   2009-09-04 01:44:41 ---
Any particular app? Steps to reproduce? Console output?

[Bug 19931] winedbg crashes entering an infinite self-launch loops

http://bugs.winehq.org/show_bug.cgi?id=19931

--- Comment #2 from The Source   2009-09-04 08:58:41 ---
Here's output on winedbg with 64bit version of wine launching winecfg:

[thesource@MAXBOOK ~]$ wine winecfg
Trying to load PE image for unsupported architecture (I386)
Trying to load PE image for unsupported architecture (I386)
Could not load wine-gecko. HTML rendering will be disabled.
wine: configuration in '/home/thesource/.wine' has been updated.
wine: Unhandled page fault on write access to 0xf213d5f0 at address
0x30e38826e3 (thread 0009), starting debugger...
wine: Unhandled page fault on read access to 0x00000028 at address
0x7fc8e97a0fc7 (thread 0023), starting debugger...
wine: Unhandled page fault on read access to 0x00000028 at address
0x7f2ab1d20fc7 (thread 0025), starting debugger...
wine: Unhandled page fault on read access to 0x00000028 at address
0x7f1e68750fc7 (thread 0027), starting debugger...

and so on...

[Bug 19931] winedbg crashes entering an infinite self-launch loops

http://bugs.winehq.org/show_bug.cgi?id=19931

Dmitry Timoshkov  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|                            |INVALID

--- Comment #3 from Dmitry Timoshkov   2009-09-04 11:48:07 ---
64-bit doesn't count, Wine doesn't officially support 64-bit yet.

[Bug 19931] winedbg crashes entering an infinite self-launch loops

http://bugs.winehq.org/show_bug.cgi?id=19931

Dmitry Timoshkov  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |CLOSED

--- Comment #4 from Dmitry Timoshkov   2009-09-04 11:48:43 ---
Closing.

[Bug 19931] winedbg crashes entering an infinite self-launch loops

http://bugs.winehq.org/show_bug.cgi?id=19931

Vitaliy Margolen  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |source, win64
           Priority|P2                          |P5
             Status|CLOSED                      |RESOLVED

--- Comment #5 from Vitaliy Margolen   2009-09-04 18:35:58 ---
It's a valid bug - since we do have wine-64 all bugs for it count as well. Just
not something with a high priority.

[Bug 19931] winedbg crashes entering an infinite self-launch loops

http://bugs.winehq.org/show_bug.cgi?id=19931

Vitaliy Margolen  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |UNCONFIRMED
         Resolution|INVALID                     |

--- Comment #6 from Vitaliy Margolen   2009-09-04 19:30:45 ---
Meant to reopen.

[Bug 19931] winedbg crashes entering an infinite self-launch loops

http://bugs.winehq.org/show_bug.cgi?id=19931

--- Comment #7 from Austin English   2009-09-04 22:53:11 ---
'wine winecfg' works fine for me in 64bit.

[Bug 19931] winedbg crashes entering an infinite self-launch loops

http://bugs.winehq.org/show_bug.cgi?id=19931

--- Comment #8 from The Source   2009-09-05 01:37:56 ---
Try to switch to audio tab (I don't know if some sound plugins are required to
reproduce, I have pulse, alsa, oss, jack and esound).
If I use wine32 winecfg then it works fine (well, mostly. sometimes it does not
start saying something about 'impossible to allocate memory'), but with wine64
winecfg switching to audio tab results in crash and the following winedbg
launch loop.

[Bug 19931] winedbg crashes entering an infinite self-launch loops

http://bugs.winehq.org/show_bug.cgi?id=19931

--- Comment #9 from Austin English   2009-09-05 03:05:30 ---
(In reply to comment #8)
> Try to switch to audio tab (I don't know if some sound plugins are required to
> reproduce, I have pulse, alsa, oss, jack and esound).
> If I use wine32 winecfg then it works fine (well, mostly. sometimes it does not
> start saying something about 'impossible to allocate memory'), but with wine64
> winecfg switching to audio tab results in crash and the following winedbg
> launch loop.

Crashes, but doesn't loop for me.

[Bug 19931] winedbg crashes entering an infinite self-launch loops

http://bugs.winehq.org/show_bug.cgi?id=19931

--- Comment #10 from The Source   2009-09-05 03:34:49 ---
Just tested on 64 bit WinRAR:

[thesource@MAXBOOK WinRAR]$ ./WinRAR.exe 
fixme:heap:HeapSetInformation 0x7f6291f90000 0 0x7f629550fd10 4
wine: Unhandled page fault on read access to 0x9a9c7a20 at address
0x7f6295170e89 (thread 0009), starting debugger...
wine: Unhandled page fault on read access to 0x00000028 at address
0x7ff473590fc7 (thread 001b), starting debugger...
wine: Unhandled page fault on read access to 0x00000028 at address
0x7f5a08100fc7 (thread 001d), starting debugger...
wine: Unhandled page fault on read access to 0x00000028 at address
0x7fcffc530fc7 (thread 001f), starting debugger...
wine: Unhandled page fault on read access to 0x00000028 at address
0x7ffbb7eb0fc7 (thread 0021), starting debugger...

The first error window says WinRAR.exe has encountered an error and must be
closed. When I press 'close' button, the same window appears but for
winedbg.exe, I close it, it appears again, and so on.

[Bug 19931] winedbg crashes entering an infinite self-launch loops

http://bugs.winehq.org/show_bug.cgi?id=19931

--- Comment #11 from Austin English   2009-09-05 12:40:02 ---
Created an attachment (id=23455)
 --> (http://bugs.winehq.org/attachment.cgi?id=23455)
crash dump

Still only crashes once here. What wine version?

[Bug 19931] winedbg crashes entering an infinite self-launch loops

http://bugs.winehq.org/show_bug.cgi?id=19931

--- Comment #12 from The Source   2009-09-05 12:42:12 ---
wine-1.1.29-1.fc11.x86_64
wine-1.1.29-1.fc11.i586

[Bug 19931] winedbg crashes entering an infinite self-launch loops

http://bugs.winehq.org/show_bug.cgi?id=19931

André H.  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
                 CC|                            |[e-mail]
     Ever Confirmed|0                           |1

--- Comment #13 from André H.   2009-09-09 08:01:30 ---
can confrim it in wine64 1.1.29 with "wine winecfg" and the click on audiotap
austin: its wine64 not wineon64, maybe thats the point

[Bug 19931] winedbg crashes entering an infinite self-launch loops

http://bugs.winehq.org/show_bug.cgi?id=19931

--- Comment #14 from Austin English   2009-09-09 12:54:52 ---
(In reply to comment #13)
> austin: its wine64 not wineon64, maybe thats the point

Yes, I'm aware. I still can't confirm it here though. I keep two trees:
$HOME/wine-git # regular tree, compiled with ./configure --without-libmpg123
$HOME/64wine-git # regular tree, compiled with ./configure --enable-win64

I used the 64wine-git tree:
$ ./wine winecfg

I don't use the wineon64, --with-wine64, etc. mess. The 32-bit tree is
installed to /usr/local/bin, the 64-bit is only used within tree, not installed
anywhere.

[Bug 19931] winedbg crashes entering an infinite self-launch loops

http://bugs.winehq.org/show_bug.cgi?id=19931

Aled Hughes  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

--- Comment #15 from Aled Hughes   2009-10-14 16:56:44 ---
I've just encountered this infinite winedbg.exe crash loop too. Trying to
install mdac25 via latest winetricks. First dasetup.exe crashes, then
winedbg.exe crashes. I have to do a "killall winedbg" to get out of the loop.

Fedora 11 x86 (2.6.30.8-64.fc11.i586)
wine-1.1.29

Here's last fragment from output:

fixme:setupapi:extract_cabinet_file awful hack: extracting cabinet
"C:\windows\temp\IXP001.TMP\MDACTWO.CAB"
wine: Call from 0x683ea6f2 to unimplemented function
oledb32.dll.DllRegisterServer, aborting
wine: Unimplemented function oledb32.dll.DllRegisterServer called at address
0x683ea6f2 (thread 002e), starting debugger...
wine: Unhandled page fault on read access to 0x0000002a at address 0x68462f23
(thread 002a), starting debugger...
wine: Unhandled page fault on read access to 0x0000002a at address 0x7cbddf23
(thread 0032), starting debugger...
wine: Unhandled page fault on read access to 0x0000002a at address 0x6848bf23
(thread 003b), starting debugger...
wine: Unhandled page fault on read access to 0x0000002a at address 0x685d7f23
(thread 0025), starting debugger...

[Bug 19931] winedbg crashes entering an infinite self-launch loops

http://bugs.winehq.org/show_bug.cgi?id=19931

Paul Bolle  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

--- Comment #16 from Paul Bolle   2009-10-25 19:05:06 ---
0) Trivially reproducible here:

winedbg notepad.exe
WineDbg starting on pid 003b
wine: Unhandled page fault on read access to 0x0000002a at address 0x6846f0c4
(thread 003a), starting debugger...
wine: Unhandled page fault on read access to 0x0000002a at address 0x684cf0c4
(thread 003e), starting debugger...
[...]

1) wine-1.1.29-3.fc12.i686 (current Fedora Rawhide).

2) Fedora seems to be the theme here. Packaging bug? SELinux issue?

[Bug 19931] winedbg crashes entering an infinite self-launch loops

http://bugs.winehq.org/show_bug.cgi?id=19931

--- Comment #17 from Vitaliy Margolen   2009-10-25 19:51:10 ---
Could you try with something newer, like wine-1.1.32? There were few patches to
address this issue.

[Bug 19931] winedbg crashes entering an infinite self-launch loops

http://bugs.winehq.org/show_bug.cgi?id=19931

--- Comment #18 from Paul Bolle   2009-10-26 03:40:21 ---
(In reply to comment #17)
> Could you try with something newer, like wine-1.1.32?

0) First step to do that (rebuild current Fedora Rawhide wine) ran into a
package conflict: see https://bugzilla.redhat.com/show_bug.cgi?id=530944 . Not
good.

1) I hope to investigate that build problem shortly so I can try to build my
own 1.1.32 package for Fedora Rawhide. Suggestions are appreciated in the mean
time.

[Bug 19931] winedbg crashes entering an infinite self-launch loops

http://bugs.winehq.org/show_bug.cgi?id=19931

--- Comment #19 from Paul Bolle   2009-10-26 09:54:22 ---
(In reply to comment #18)
> - I hope to investigate that build problem shortly so I can try to build my
>
> own 1.1.32 package for Fedora Rawhide.

0) Finally anaged to build and package wine-1.1.32.

1) Identical error/crash.

2) What can I do to further pinpoint this crash of winedbg (env variables,
commandline options, buildflags)? A way to make winedbg print some sort of 
backtrace?

[Bug 19931] winedbg crashes entering an infinite self-launch loops

http://bugs.winehq.org/show_bug.cgi?id=19931

--- Comment #20 from Paul Bolle   2009-10-28 07:35:35 ---
Created an attachment (id=24409)
 --> (http://bugs.winehq.org/attachment.cgi?id=24409)
Trivial fix, see comment.

[Bug 19931] winedbg crashes entering an infinite self-launch loops

http://bugs.winehq.org/show_bug.cgi?id=19931

--- Comment #21 from Paul Bolle   2009-10-28 07:37:00 ---
0) I think I have found the problem(s). There seem to be two issues here.

1) On Fedora /usr/bin/wine is a shell script that (basically) either calls
/usr/bin/wine32 or /usr/bin/wine64. Somehow this confuses winedbg, at least it
feels like winedbg behaves better if [...]/wine32 is copied to [...]/wine.
Packaging issue, I'd guess.

2) Another issue is an apparent bug in elf_load_file(). I pinpointed the
winedbg repeating crashes (with a lot of printfs) to:
   esm->fmap->sect[esm->sidx].mapped = ELF_NO_MAP;
in elf_unmap_section().

The last steps in the call chain always looked like:
elf_load_file()
    elf_unmap_file()
        elf_unmap_section()

It turns out elf_load_file() calls elf_unmap_file() if elf_map_file() failed.
That is incorrect (and really should be incorrect, given the names of those two
functions).

3) Trivial patch is attached.

[Bug 19931] winedbg crashes entering an infinite self-launch loops

http://bugs.winehq.org/show_bug.cgi?id=19931

--- Comment #22 from Paul Bolle   2009-10-28 11:33:10 ---
Fixed included in master as commit 4c97910387cc72bc1e7e600fc71aa8b285dae31e
(see
http://source.winehq.org/git/wine.git/?a=commit;h=4c97910387cc72bc1e7e600fc7...
). Can this be closed?

[Bug 19931] winedbg crashes entering an infinite self-launch loops

http://bugs.winehq.org/show_bug.cgi?id=19931

--- Comment #23 from André H.   2009-10-28 12:00:04 ---
if that patch really fixed it(please really try it), then this bug can be
resolved fixed, on next release it will be closed

[Bug 19931] winedbg crashes entering an infinite self-launch loops

http://bugs.winehq.org/show_bug.cgi?id=19931

Austin English  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #24 from Austin English   2009-10-28 13:19:35 ---
Should be fixed.

[Bug 19931] winedbg crashes entering an infinite self-launch loops

http://bugs.winehq.org/show_bug.cgi?id=19931

The Source  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
         Resolution|FIXED                       |

--- Comment #25 from The Source   2009-11-06 08:24:01 ---
Still an issue in 1.1.32

wine64 ./WinRAR.exe 
fixme:heap:HeapSetInformation 0x7fe9a4a40000 0 0x7fe9a81efd10 4
wine: Unhandled page fault on read access to 0xad6d43e0 at address
0x7fe9a7e50639 (thread 0009), starting debugger...
wine: Unhandled page fault on read access to 0x00000028 at address
0x7f0e98861057 (thread 001c), starting debugger...
wine: Unhandled page fault on read access to 0x00000028 at address
0x7f7b20011057 (thread 001e), starting debugger...
wine: Unhandled page fault on read access to 0x00000028 at address
0x7fa19ea21057 (thread 0020), starting debugger...

[Bug 19931] winedbg crashes entering an infinite self-launch loops

http://bugs.winehq.org/show_bug.cgi?id=19931

Vitaliy Margolen  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|REOPENED                    |RESOLVED
         Resolution|                            |FIXED

--- Comment #26 from Vitaliy Margolen   2009-11-06 09:05:15 ---
The fix is in the git, not wine-1.1.32

[Bug 19931] winedbg crashes entering an infinite self-launch loops

http://bugs.winehq.org/show_bug.cgi?id=19931

Alexandre Julliard  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |CLOSED

--- Comment #27 from Alexandre Julliard   2009-11-13 12:44:18 ---
Closing bugs fixed in 1.1.33.

[Bug 19931] winedbg crashes entering an infinite self-launch loops

http://bugs.winehq.org/show_bug.cgi?id=19931

Anastasius Focht  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
      Fixed by SHA1|                            |4c97910387cc72bc1e7e600fc71
                   |                            |aa8b285dae31e
                 CC|                            |[e-mail]