[Bug 34588] New: rpcrt4_test.exe fails to delay load secur32.dll on ARM

Bug #: 34588 Summary: rpcrt4_test.exe fails to delay load secur32.dll on ARM Product: Wine Version: 1.7.2 Platform: arm OS/Version: Linux Status: UNCONFIRMED Severity: normal Priority: P2 Component: secur32 AssignedTo: wine-bugs@winehq.org ReportedBy: Stefan.Leichter@camLine.com Classification: Unclassified

Created attachment 46063 --> http://bugs.winehq.org/attachment.cgi?id=46063 "wine rpcrt4_test.exe rpc" console log with +seh enabled

While trying to find the reason for the timeouts in my winetests runs (Tag arm-sle85276) i found that secur32.dll can not be delay loaded because of an exception in function DllMain().

The exception happens when calling the fork() function in function fork_helper() http://source.winehq.org/git/wine.git/blob/81213a88e83b19c73e244e00bffac2e40...

The console log of "wine rpcrt4_test.exe rpc" with +seh is attached.

-- Configure bugmail: http://bugs.winehq.org/userprefs.cgi?tab=email Do not reply to this email, post in Bugzilla using the above URL to reply. ------- You are receiving this mail because: ------- You are watching all bug changes.

Show replies by date

wine-bugs＠winehq.org

22 Oct 22 Oct

5:39 p.m.

New subject: [Bug 34588] rpcrt4_test.exe fails to delay load secur32.dll on ARM

André H. nerv@dawncrow.de changed:

What |Removed |Added ---------------------------------------------------------------------------- CC| |nerv@dawncrow.de

wine-bugs＠winehq.org

24 Oct 24 Oct

9:51 p.m.

New subject: [Bug 34588] rpcrt4_test.exe fails to delay load secur32.dll on ARM

--- Comment #1 from André H. nerv@dawncrow.de 2013-10-24 16:51:27 CDT --- thx for analyzing it!

i wasn't able to reproduce it in 1.7.2 so i updated my source tree i'm sadly still not able to reproduce it in wine-1.7.4-300-ga562af5

have you used some special compiler/linker/configure flags? does it help to rebuild secur32, rpcrt4 and the test?

wine-bugs＠winehq.org

29 Oct 29 Oct

9:52 p.m.

New subject: [Bug 34588] rpcrt4_test.exe fails to delay load secur32.dll on ARM

--- Comment #2 from Stefan Leichter Stefan.Leichter@camLine.com 2013-10-29 16:52:30 CDT --- (In reply to comment #1)

have you used some special compiler/linker/configure flags?

options for configure were CFLAGS="-g -mcpu=cortex-a8 -mfpu=neon -mfloat-abi=softfp" --prefix=/mnt/utmp/wine

does it help to rebuild secur32, rpcrt4 and the test?

The build tree gets removed before starting a new build, what can i do more?

wine-bugs＠winehq.org

11 Nov 11 Nov

9:35 p.m.

New subject: [Bug 34588] rpcrt4_test.exe fails to delay load secur32.dll on ARM

--- Comment #3 from André H. nerv@dawncrow.de 2013-11-11 15:35:42 CST --- (In reply to comment #2)

(In reply to comment #1)

...
have you used some special compiler/linker/configure flags?

options for configure were CFLAGS="-g -mcpu=cortex-a8 -mfpu=neon -mfloat-abi=softfp" --prefix=/mnt/utmp/wine

...
does it help to rebuild secur32, rpcrt4 and the test?

The build tree gets removed before starting a new build, what can i do more?

I was only able to compile Wine with configure options CFLAGS="-g -mcpu=cortex-a8 -mfpu=neon" and still can't reproduce this bug :(

A wild guess... Could you please add some ERRs or TRACEs in SECUR32_initializeProviders to see which fails? (in case it is there) and then iterate that way through the subfunction that might crashed until you find the point where that happens. Of course you could also run Wine in GDB...

wine-bugs＠winehq.org

20 Nov 20 Nov

10:07 p.m.

New subject: [Bug 34588] rpcrt4_test.exe fails to delay load secur32.dll on ARM

--- Comment #4 from Stefan Leichter Stefan.Leichter@camLine.com 2013-11-20 16:07:38 CST --- Created attachment 46592 --> http://bugs.winehq.org/attachment.cgi?id=46592 diff containing the additional ERR's

Here is poor man's backtrace:

fixme:rpc:RpcNetworkIsProtseqValidW Unknown protseq L"foo" rpc.c:228: Test failed: wrong RpcMgmtStopServerListening error (0) rpc.c:248: Test failed: RpcServerListen failed (1713) rpc.c:253: Test failed: wrong RpcServerListen error (0) err:secur32:SECUR32_initializeProviders before Schannel err:secur32:SECUR32_initializeProviders Schannel -> NTLM err:ntlm:fork_helper 0x40e0f638, "ntlm_auth", 0x40e0f62c err:ntlm:fork_helper err:ntlm:fork_helper 0x4095d3b8 0x400ab448 err:module:DelayLoadFailureHook failed to delay load secur32.dll.EnumerateSecurityPackagesA

wine-bugs＠winehq.org

10:23 p.m.

New subject: [Bug 34588] rpcrt4_test.exe fails to delay load secur32.dll on ARM

--- Comment #5 from André H. nerv@dawncrow.de 2013-11-20 16:23:32 CST --- (In reply to comment #4)

Created attachment 46592 [details] diff containing the additional ERR's

Here is poor man's backtrace:

fixme:rpc:RpcNetworkIsProtseqValidW Unknown protseq L"foo" rpc.c:228: Test failed: wrong RpcMgmtStopServerListening error (0) rpc.c:248: Test failed: RpcServerListen failed (1713) rpc.c:253: Test failed: wrong RpcServerListen error (0) err:secur32:SECUR32_initializeProviders before Schannel err:secur32:SECUR32_initializeProviders Schannel -> NTLM err:ntlm:fork_helper 0x40e0f638, "ntlm_auth", 0x40e0f62c err:ntlm:fork_helper err:ntlm:fork_helper 0x4095d3b8 0x400ab448 err:module:DelayLoadFailureHook failed to delay load secur32.dll.EnumerateSecurityPackagesA

good job just to be sure, could you please add some ERRs in the _three_ different cases for helper->helper_pid and see if it makes a difference? anyway, it's really strange what happens here...

wine-bugs＠winehq.org

21 Nov 21 Nov

10:35 p.m.

New subject: [Bug 34588] rpcrt4_test.exe fails to delay load secur32.dll on ARM

Stefan Leichter Stefan.Leichter@camLine.com changed:

--- Comment #6 from Stefan Leichter Stefan.Leichter@camLine.com 2013-11-21 16:35:03 CST --- Created attachment 46599 --> http://bugs.winehq.org/attachment.cgi?id=46599 diff containing more additional ERR's

I don't get the ERR from any of the three cases

fixme:rpc:RpcNetworkIsProtseqValidW Unknown protseq L"foo" rpc.c:228: Test failed: wrong RpcMgmtStopServerListening error (0) rpc.c:248: Test failed: RpcServerListen failed (1713) rpc.c:253: Test failed: wrong RpcServerListen error (0) err:secur32:SECUR32_initializeProviders before Schannel err:secur32:SECUR32_initializeProviders Schannel -> NTLM err:ntlm:fork_helper 0x40def638, "ntlm_auth", 0x40def62c err:ntlm:fork_helper err:ntlm:fork_helper 0x409283b8 0x4026a448

wine-bugs＠winehq.org

22 Nov 22 Nov

12:56 p.m.

New subject: [Bug 34588] rpcrt4_test.exe fails to delay load secur32.dll on ARM

André H. nerv@dawncrow.de changed:

wine-bugs＠winehq.org

1:08 p.m.

New subject: [Bug 34588] rpcrt4_test.exe fails to delay load secur32.dll on ARM

--- Comment #7 from André H. nerv@dawncrow.de 2013-11-22 07:08:37 CST --- (In reply to comment #6)

Created attachment 46599 [details] diff containing more additional ERR's

I don't get the ERR from any of the three cases

Thx, so something is really wrong with fork(), could you try getting a strace from this fork() call please?

wine-bugs＠winehq.org

7:33 p.m.

New subject: [Bug 34588] rpcrt4_test.exe fails to delay load secur32.dll on ARM

--- Comment #8 from Stefan Leichter Stefan.Leichter@camLine.com 2013-11-22 13:33:21 CST --- Created attachment 46608 --> http://bugs.winehq.org/attachment.cgi?id=46608 strace log (bzip2 compressed)

wine-bugs＠winehq.org

23 Feb 23 Feb

7:22 a.m.

New subject: [Bug 34588] rpcrt4_test.exe fails to delay load secur32.dll on ARM

--- Comment #9 from Austin English austinenglish@gmail.com --- This is your friendly reminder that there has been no bug activity for over a year. Is this still an issue in current (1.7.37 or newer) wine? If so, please attach the terminal output in 1.7.37 (see http://wiki.winehq.org/FAQ#get_log).

-- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.

wine-bugs＠winehq.org

13 Dec 13 Dec

7:10 p.m.

New subject: [Bug 34588] rpcrt4_test.exe fails to delay load secur32.dll on ARM

--- Comment #10 from André H. nerv@dawncrow.de --- coming back here quite late...

clone(child_stack=0, flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|SIGCHLD, child_tidptr=0x40022048) = 2103 --- SIGSEGV (Segmentation fault) @ 0 (0) ---

I guess it was just a buggy kernel or libc, as i never was able to reproduce it, this is my best guess

-- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.

wine-bugs＠winehq.org

6 Oct 6 Oct

11:03 a.m.

New subject: [Bug 34588] rpcrt4_test.exe fails to delay load secur32.dll on ARM

Anastasius Focht focht@gmx.net changed:

--- Comment #11 from Anastasius Focht focht@gmx.net --- Hello folks,

resolving here based on last comment #10

--- quote --- I guess it was just a buggy kernel or libc, as i never was able to reproduce it, this is my best guess --- quote ---

Regards

-- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.

wine-bugs＠winehq.org

4 Nov 4 Nov

10:22 a.m.

New subject: [Bug 34588] rpcrt4_test.exe fails to delay load secur32.dll on ARM

Austin English austinenglish@gmail.com changed:

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

--- Comment #12 from Austin English austinenglish@gmail.com --- Closing.

-- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.

wine-bugs＠winehq.org

5 Nov 5 Nov

10:19 p.m.

New subject: [Bug 34588] rpcrt4_test.exe fails to delay load secur32.dll on ARM