OpenSSH crashes - able to repeat the cause

OpenSSH crashes - able to repeat the cause

[Jack Radigan]

I've been having the same crashes as others have been reporting here with
the latest cygwin and OpenSSH code.  In short, the NT resource kits seem
to be the common cause here.

I was able to isolate it to the reskit (tried supplament 3 & 4).  Better
still, by backing out the reskit and removing the following:

	%NTRESKIT%
	the reskit path from %PATH%
	HKLM\Software\Microsoft\Reskit

After rebooting OpenSSH no longer gets a Dr. Watson on id.exe.  I've been
able to reproduce this on both a PDC and a member server, both are runing
NT4 SP6a.

I not got a clue why this is happening but hope someone else can get the
same results.

-jack-

Jack Radigan -  Dun & Bradstreet, Inc.



--
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple

Re: OpenSSH crashes - able to repeat the cause

[Erdely, Michael]

I'm having a hard time seeing this. I have installed this on Windows 2000
machines both with and without the Win2000 Reskit.  All of my machines have
had the "crash".

-ME

----- Original Message -----
From: "Jack Radigan" <jprad@dnb.com>
To: <cygwin@cygwin.com>
Sent: Sunday, January 21, 2001 10:05 PM
Subject: OpenSSH crashes - able to repeat the cause


> I've been having the same crashes as others have been reporting here with
> the latest cygwin and OpenSSH code.  In short, the NT resource kits seem
> to be the common cause here.
>
> I was able to isolate it to the reskit (tried supplament 3 & 4).  Better
> still, by backing out the reskit and removing the following:
>
> %NTRESKIT%
> the reskit path from %PATH%
> HKLM\Software\Microsoft\Reskit
>
> After rebooting OpenSSH no longer gets a Dr. Watson on id.exe.  I've been
> able to reproduce this on both a PDC and a member server, both are runing
> NT4 SP6a.
>
> I not got a clue why this is happening but hope someone else can get the
> same results.
>
> -jack-
>
> Jack Radigan -  Dun & Bradstreet, Inc.
>
>
>
> --
> Want to unsubscribe from this list?
> Check out: http://cygwin.com/ml/#unsubscribe-simple
>
>


--
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple

Re: OpenSSH crashes - able to repeat the cause

[Jack Radigan]

I've only been doing this on NT4, not Win2k.

Sorry.

-jack-


On Sun, 21 Jan 2001, Erdely, Michael wrote:

> I'm having a hard time seeing this. I have installed this on Windows 2000
> machines both with and without the Win2000 Reskit.  All of my machines have
> had the "crash".
>
> -ME
>
> ----- Original Message -----
> From: "Jack Radigan" <jprad@dnb.com>
> To: <cygwin@cygwin.com>
> Sent: Sunday, January 21, 2001 10:05 PM
> Subject: OpenSSH crashes - able to repeat the cause
>
>
> > I've been having the same crashes as others have been reporting here with
> > the latest cygwin and OpenSSH code.  In short, the NT resource kits seem
> > to be the common cause here.
> >
> > I was able to isolate it to the reskit (tried supplament 3 & 4).  Better
> > still, by backing out the reskit and removing the following:
> >
> > %NTRESKIT%
> > the reskit path from %PATH%
> > HKLM\Software\Microsoft\Reskit
> >
> > After rebooting OpenSSH no longer gets a Dr. Watson on id.exe.  I've been
> > able to reproduce this on both a PDC and a member server, both are runing
> > NT4 SP6a.
> >
> > I not got a clue why this is happening but hope someone else can get the
> > same results.
> >
> > -jack-
> >
> > Jack Radigan -  Dun & Bradstreet, Inc.
> >
> >
> >
> > --
> > Want to unsubscribe from this list?
> > Check out: http://cygwin.com/ml/#unsubscribe-simple
> >
> >
>
>
> --
> Want to unsubscribe from this list?
> Check out: http://cygwin.com/ml/#unsubscribe-simple
>
>



--
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple

Re: OpenSSH crashes - able to repeat the cause

[Ross Smith]

Jack Radigan wrote:
> 
> I've only been doing this on NT4, not Win2k.
> 
> Sorry.
> 
> -jack-

I get the crashes on NT4, and I haven't got the reskit installed.

-- 
Ross Smith <ross.s@ihug.co.nz> The Internet Group, Auckland, New Zealand
========================================================================
          "Normally he was insane, but he had lucid moments
          when he was merely stupid."     -- Heinrich Heine

--
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple

Re: OpenSSH crashes - able to repeat the cause

[Henry S. Thompson]

"Jack Radigan" <jprad@dnb.com> writes:

> I've only been doing this on NT4, not Win2k.
> 
> Sorry.
> 
> -jack-
> 
> 
> On Sun, 21 Jan 2001, Erdely, Michael wrote:
> 
> > I'm having a hard time seeing this. I have installed this on Windows 2000
> > machines both with and without the Win2000 Reskit.  All of my machines have
> > had the "crash".

Another Win2k user with this crash.

But I note that occasionally (not always) after a reboot things work
OK for a short while, then the crash re-appears.  So you may have seen
some joy because you were rebooting after removing the ResKit.

ht
-- 
  Henry S. Thompson, HCRC Language Technology Group, University of Edinburgh
          W3C Fellow 1999--2001, part-time member of W3C Team
     2 Buccleuch Place, Edinburgh EH8 9LW, SCOTLAND -- (44) 131 650-4440
	    Fax: (44) 131 650-4587, e-mail: ht@cogsci.ed.ac.uk
		     URL: http://www.ltg.ed.ac.uk/~ht/

--
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple

Re: OpenSSH crashes - able to repeat the cause

[Jack Radigan]

Still haven't played with this on Win2K.  As for things here,
it's been solid ever since.  I'll throw an expect script
together to see if it dies overnight.

-jack-

Jack Radigan - Dun & Bradstreet, Inc.

On 22 Jan 2001 ht@cogsci.ed.ac.uk wrote:

> "Jack Radigan" <jprad@dnb.com> writes:
>
> > I've only been doing this on NT4, not Win2k.
> >
> > Sorry.
> >
> > -jack-
> >
> >
> > On Sun, 21 Jan 2001, Erdely, Michael wrote:
> >
> > > I'm having a hard time seeing this. I have installed this on Windows 2000
> > > machines both with and without the Win2000 Reskit.  All of my machines have
> > > had the "crash".
>
> Another Win2k user with this crash.
>
> But I note that occasionally (not always) after a reboot things work
> OK for a short while, then the crash re-appears.  So you may have seen
> some joy because you were rebooting after removing the ResKit.
>
> ht
> --
>   Henry S. Thompson, HCRC Language Technology Group, University of Edinburgh
>           W3C Fellow 1999--2001, part-time member of W3C Team
>      2 Buccleuch Place, Edinburgh EH8 9LW, SCOTLAND -- (44) 131 650-4440
> 	    Fax: (44) 131 650-4587, e-mail: ht@cogsci.ed.ac.uk
> 		     URL: http://www.ltg.ed.ac.uk/~ht/
>



--
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple

Re: OpenSSH crashes - able to repeat the cause

[Corinna Vinschen]

On Mon, Jan 22, 2001 at 09:05:34PM +0000, Henry S. Thompson wrote:
> "Jack Radigan" <jprad@dnb.com> writes:
> 
> > I've only been doing this on NT4, not Win2k.
> > 
> > Sorry.
> > 
> > -jack-
> > 
> > 
> > On Sun, 21 Jan 2001, Erdely, Michael wrote:
> > 
> > > I'm having a hard time seeing this. I have installed this on Windows 2000
> > > machines both with and without the Win2000 Reskit.  All of my machines have
> > > had the "crash".
> 
> Another Win2k user with this crash.

I like these "me too"s like headaches.

Guess what? I'm using W2K and I'm not seeing this. Where is the guy
having that problem trying actually to DEBUG the problem?

Corinna

-- 
Corinna Vinschen                  Please, send mails regarding Cygwin to
Cygwin Developer                                mailto:cygwin@cygwin.com
Red Hat, Inc.

--
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple

Re: OpenSSH crashes - able to repeat the cause

[Henry S. Thompson]

Corinna Vinschen <cygwin@cygwin.com> writes:

> On Mon, Jan 22, 2001 at 09:05:34PM +0000, Henry S. Thompson wrote:
> > "Jack Radigan" <jprad@dnb.com> writes:
> > 
> > > I've only been doing this on NT4, not Win2k.
> > > 
> > > Sorry.
> > > 
> > > -jack-
> > > 
> > > 
> > > On Sun, 21 Jan 2001, Erdely, Michael wrote:
> > > 
> > > > I'm having a hard time seeing this. I have installed this on Windows 2000
> > > > machines both with and without the Win2000 Reskit.  All of my machines have
> > > > had the "crash".
> > 
> > Another Win2k user with this crash.
> 
> I like these "me too"s like headaches.
> 
> Guess what? I'm using W2K and I'm not seeing this. Where is the guy
> having that problem trying actually to DEBUG the problem?

I'm trying as hard as I can with the resources I have available to
debug, by giving as much detail on the problem which has not otherwise 
been provided, in the hopes that they will spark an 'aha!' reaction
from someone.

Would compiling id (the program which fails) from source with -g be a
useful step?  Presumably not, because the crash goes in to the Win2k
debugger directly -- what would be required to get symbols available
for debugging at that point?  I have a bit of experience debugging C
under Win2K, and would be happy to try to follow suggestions in this
area to help sort this out once and for all.

Any program at all with symbols etc. would do, actually, as anything
you run from the command line after blowing off the initial error dies.

ht
-- 
  Henry S. Thompson, HCRC Language Technology Group, University of Edinburgh
          W3C Fellow 1999--2001, part-time member of W3C Team
     2 Buccleuch Place, Edinburgh EH8 9LW, SCOTLAND -- (44) 131 650-4440
	    Fax: (44) 131 650-4587, e-mail: ht@cogsci.ed.ac.uk
		     URL: http://www.ltg.ed.ac.uk/~ht/

--
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple

Re: OpenSSH crashes - able to repeat the cause

[Corinna Vinschen]

On Tue, Jan 23, 2001 at 11:01:12AM +0000, Henry S. Thompson wrote:
> I'm trying as hard as I can with the resources I have available to
> debug, by giving as much detail on the problem which has not otherwise 
> been provided, in the hopes that they will spark an 'aha!' reaction
> from someone.
> 
> Would compiling id (the program which fails) from source with -g be a
> useful step?  Presumably not, because the crash goes in to the Win2k
> debugger directly -- what would be required to get symbols available
> for debugging at that point?  I have a bit of experience debugging C

You need a non stripped version of the Cygwin DLL and gdb. The
non stripped version is easily produced by downloading the sources
from cvs as descibed on http://www.cygwin.com/cvs.html and following
the directions given on the same side.

> under Win2K, and would be happy to try to follow suggestions in this
> area to help sort this out once and for all.

You (and others having that problem) could do me a tiny favour,
please. Please download the latest developers snapshot from
http://www.cygwin.com/snapshots and try that DLL instead of the
official 1.1.7. If it suddenly works - good for me. If not, please
send a stacktrace  to the mailing list if available.

Corinna

-- 
Corinna Vinschen                  Please, send mails regarding Cygwin to
Cygwin Developer                                mailto:cygwin@cygwin.com
Red Hat, Inc.

--
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple

Re: OpenSSH crashes - able to repeat the cause

[Henry S. Thompson]

Corinna Vinschen <cygwin@cygwin.com> writes:

> On Tue, Jan 23, 2001 at 11:01:12AM +0000, Henry S. Thompson wrote:
> > I'm trying as hard as I can with the resources I have available to
> > debug, by giving as much detail on the problem which has not otherwise 
> > been provided, in the hopes that they will spark an 'aha!' reaction
> > from someone.
> > 
> > Would compiling id (the program which fails) from source with -g be a
> > useful step?  Presumably not, because the crash goes in to the Win2k
> > debugger directly -- what would be required to get symbols available
> > for debugging at that point?  I have a bit of experience debugging C
> 
> You need a non stripped version of the Cygwin DLL and gdb. The
> non stripped version is easily produced by downloading the sources
> from cvs as descibed on http://www.cygwin.com/cvs.html and following

I'll try that tomorrow.

> > under Win2K, and would be happy to try to follow suggestions in this
> > area to help sort this out once and for all.
> 
> You (and others having that problem) could do me a tiny favour,
> please. Please download the latest developers snapshot from
> http://www.cygwin.com/snapshots and try that DLL instead of the
> official 1.1.7. If it suddenly works - good for me. If not, please
> send a stacktrace  to the mailing list if available.

The snapshot at .../cygwin1-20010119.dll.bz2 does not fix the
problem.  Do you want the raw hex trace?

ht
-- 
  Henry S. Thompson, HCRC Language Technology Group, University of Edinburgh
          W3C Fellow 1999--2001, part-time member of W3C Team
     2 Buccleuch Place, Edinburgh EH8 9LW, SCOTLAND -- (44) 131 650-4440
	    Fax: (44) 131 650-4587, e-mail: ht@cogsci.ed.ac.uk
		     URL: http://www.ltg.ed.ac.uk/~ht/

--
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple

Re: OpenSSH crashes - able to repeat the cause

[Corinna Vinschen]

On Tue, Jan 23, 2001 at 05:35:12PM +0000, Henry S. Thompson wrote:
> Corinna Vinschen <cygwin@cygwin.com> writes:
> > You (and others having that problem) could do me a tiny favour,
> > please. Please download the latest developers snapshot from
> > http://www.cygwin.com/snapshots and try that DLL instead of the
> > official 1.1.7. If it suddenly works - good for me. If not, please
> > send a stacktrace  to the mailing list if available.
> 
> The snapshot at .../cygwin1-20010119.dll.bz2 does not fix the
> problem.  Do you want the raw hex trace?

Yes, please.

Corinna

-- 
Corinna Vinschen                  Please, send mails regarding Cygwin to
Cygwin Developer                                mailto:cygwin@cygwin.com
Red Hat, Inc.

--
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple

Re: OpenSSH crashes - able to repeat the cause

[Henry S. Thompson]

Stacktrace of id.exe crash with cygwin1-20010119.dll:

You are successfully logged in to this server!!!
    920 [main]  a 0 handle_exceptions: Exception:
STATUS_ACCESS_VIOLATION
Exception: STATUS_ACCESS_VIOLATION at eip=61001BD0
eax=00000001 ebx=00000001 ecx=00000000 edx=12014988 esi=000206F0
edi=00000000
ebp=0240F9E0 esp=0240FE7C program=H/    a
cs=001B ds=0023 es=0023 fs=0038 gs=0000 ss=0023
Stack trace:
Frame     Function  Args
0240F9E0  61001BD0  (00000000, 00000000, 00000000, 00000000)
0240FEA4  61001BD0  (00000110, 02400001, 0240FEE4, 610033D6)
0240FEE4  61003510  (7FFDF000, 004072CC, 6108513C, 000A7000)
0240FF74  61003D15  (00000008, 0247D7F1, 00000000, 00000000)
0240FF94  61003D64  (00407230, 004011D0, 0240FFC0, 00401038)
0240FFA4  00405206  (004011D0, 80430F7B, 77E992A3, FFFFFFFE)
0240FFC0  00401038  (0247D7F1, 00000008, 7FFDF000, 77FCA3F7)
0240FFF0  77E992A6  (00401000, 00000000, 000000C8, 00000100)
End of stack traceht> exit
-- 
  Henry S. Thompson, HCRC Language Technology Group, University of Edinburgh
          W3C Fellow 1999--2001, part-time member of W3C Team
     2 Buccleuch Place, Edinburgh EH8 9LW, SCOTLAND -- (44) 131 650-4440
	    Fax: (44) 131 650-4587, e-mail: ht@cogsci.ed.ac.uk
		     URL: http://www.ltg.ed.ac.uk/~ht/

--
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple

Re: OpenSSH crashes - able to repeat the cause

[Henry S. Thompson]

Corinna Vinschen <cygwin@cygwin.com> writes:

> On Tue, Jan 23, 2001 at 11:01:12AM +0000, Henry S. Thompson wrote:
> > I'm trying as hard as I can with the resources I have available to
> > debug, by giving as much detail on the problem which has not otherwise 
> > been provided, in the hopes that they will spark an 'aha!' reaction
> > from someone.
> > 
> > Would compiling id (the program which fails) from source with -g be a
> > useful step?  Presumably not, because the crash goes in to the Win2k
> > debugger directly -- what would be required to get symbols available
> > for debugging at that point?  I have a bit of experience debugging C
> 
> You need a non stripped version of the Cygwin DLL and gdb. The
> non stripped version is easily produced by downloading the sources
> from cvs as descibed on http://www.cygwin.com/cvs.html and following
> the directions given on the same side.

So I've built and installed a non-stripped version, and I've read the
FAQ, but I don't see how to get GDB involved!  I run ssh, I get the
crash. . . then what?  I can't get GDB running to catch the error???

ht
-- 
  Henry S. Thompson, HCRC Language Technology Group, University of Edinburgh
          W3C Fellow 1999--2001, part-time member of W3C Team
     2 Buccleuch Place, Edinburgh EH8 9LW, SCOTLAND -- (44) 131 650-4440
	    Fax: (44) 131 650-4587, e-mail: ht@cogsci.ed.ac.uk
		     URL: http://www.ltg.ed.ac.uk/~ht/

--
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple

'descriptor is a file, not a socket' returns (was Re: OpenSSH crashes - able to repeat the cause)

[Henry S. Thompson]

I was slowed down in building an unstripped cygwin1.dll by the return
of the cvs/socket problem???  Having successfully logged in to the
:pserver:anoncvs . . ., cvs checkout winsup failed with the above
message.  I'm running cvs 1.11.0-1.  Have we somehow lost Corinna's
patch described in
http://www.delorie.com/archives/browse.cgi?p=cygwin/2000/10/26/06:43:29 ?

I'm running Win2k.

NB this happened both before (with Cygwin1.1.7) _and_ after I switched 
to the locally built unstripped cygwin1.dll.

ht
-- 
  Henry S. Thompson, HCRC Language Technology Group, University of Edinburgh
          W3C Fellow 1999--2001, part-time member of W3C Team
     2 Buccleuch Place, Edinburgh EH8 9LW, SCOTLAND -- (44) 131 650-4440
	    Fax: (44) 131 650-4587, e-mail: ht@cogsci.ed.ac.uk
		     URL: http://www.ltg.ed.ac.uk/~ht/

--
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple

Re: OpenSSH crashes - able to repeat the cause

[Corinna Vinschen]

Sorry but... could you please send your /etc/passwd and /etc/group
files as well as the output of id when it not crashes?

Thanks,
Corinna

On Wed, Jan 24, 2001 at 08:31:36AM +0000, Henry S. Thompson wrote:
> Stacktrace of id.exe crash with cygwin1-20010119.dll:
> 
> You are successfully logged in to this server!!!
>     920 [main]  a 0 handle_exceptions: Exception:
> STATUS_ACCESS_VIOLATION
> Exception: STATUS_ACCESS_VIOLATION at eip=61001BD0
> eax=00000001 ebx=00000001 ecx=00000000 edx=12014988 esi=000206F0
> edi=00000000
> ebp=0240F9E0 esp=0240FE7C program=H/    a
> cs=001B ds=0023 es=0023 fs=0038 gs=0000 ss=0023
> Stack trace:
> Frame     Function  Args
> 0240F9E0  61001BD0  (00000000, 00000000, 00000000, 00000000)
> 0240FEA4  61001BD0  (00000110, 02400001, 0240FEE4, 610033D6)
> 0240FEE4  61003510  (7FFDF000, 004072CC, 6108513C, 000A7000)
> 0240FF74  61003D15  (00000008, 0247D7F1, 00000000, 00000000)
> 0240FF94  61003D64  (00407230, 004011D0, 0240FFC0, 00401038)
> 0240FFA4  00405206  (004011D0, 80430F7B, 77E992A3, FFFFFFFE)
> 0240FFC0  00401038  (0247D7F1, 00000008, 7FFDF000, 77FCA3F7)
> 0240FFF0  77E992A6  (00401000, 00000000, 000000C8, 00000100)
> End of stack traceht> exit
> -- 
>   Henry S. Thompson, HCRC Language Technology Group, University of Edinburgh
>           W3C Fellow 1999--2001, part-time member of W3C Team
>      2 Buccleuch Place, Edinburgh EH8 9LW, SCOTLAND -- (44) 131 650-4440
> 	    Fax: (44) 131 650-4587, e-mail: ht@cogsci.ed.ac.uk
> 		     URL: http://www.ltg.ed.ac.uk/~ht/
> 
> --
> Want to unsubscribe from this list?
> Check out: http://cygwin.com/ml/#unsubscribe-simple

-- 
Corinna Vinschen                  Please, send mails regarding Cygwin to
Cygwin Developer                                mailto:cygwin@cygwin.com
Red Hat, Inc.

--
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple

Re: OpenSSH crashes - able to repeat the cause

[Henry S. Thompson]

Corinna Vinschen <cygwin@cygwin.com> writes:

> Sorry but... could you please send your /etc/passwd and /etc/group
> files as well as the output of id when it not crashes?

> id
uid=1000(ht) gid=513(None) groups=513(None)

> cat /etc/group
Everyone:S-1-1-0:0:
SYSTEM:S-1-5-18:18:
None:S-1-5-21-1216639349-1515132614-219625367-513:513:
Administrators:S-1-5-32-544:544:
Backup Operators:S-1-5-32-551:551:
Guests:S-1-5-32-546:546:
Power Users:S-1-5-32-547:547:
Replicator:S-1-5-32-552:552:
Users:S-1-5-32-545:545:

> cat /etc/passwd
Everyone:*:0:0:,S-1-1-0::
SYSTEM:*:18:18:,S-1-5-18::
Administrator::500:513:,S-1-5-21-1216639349-1515132614-219625367-500::/bin/sh
Guest::501:513:,S-1-5-21-1216639349-1515132614-219625367-501::/bin/sh
hisarm::1001:513:Ruli Manurung,S-1-5-21-1216639349-1515132614-219625367-1001:/home/hisarm:/usr/bin/bash
ht::1000:513:Henry S. Thompson,S-1-5-21-1216639349-1515132614-219625367-1000:/home/ht:/usr/bin/bash
-- 
  Henry S. Thompson, HCRC Language Technology Group, University of Edinburgh
          W3C Fellow 1999--2001, part-time member of W3C Team
     2 Buccleuch Place, Edinburgh EH8 9LW, SCOTLAND -- (44) 131 650-4440
	    Fax: (44) 131 650-4587, e-mail: ht@cogsci.ed.ac.uk
		     URL: http://www.ltg.ed.ac.uk/~ht/

--
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple

Re: OpenSSH crashes - able to repeat the cause

[Henry S. Thompson]

Corinna Vinschen <cygwin@cygwin.com> writes:

> On Tue, Jan 23, 2001 at 11:01:12AM +0000, Henry S. Thompson wrote:
> > I'm trying as hard as I can with the resources I have available to
> > debug, by giving as much detail on the problem which has not otherwise 
> > been provided, in the hopes that they will spark an 'aha!' reaction
> > from someone.
> > 
> > Would compiling id (the program which fails) from source with -g be a
> > useful step?  Presumably not, because the crash goes in to the Win2k
> > debugger directly -- what would be required to get symbols available
> > for debugging at that point?  I have a bit of experience debugging C
> 
> You need a non stripped version of the Cygwin DLL and gdb. The
> non stripped version is easily produced by downloading the sources
> from cvs as descibed on http://www.cygwin.com/cvs.html and following
> the directions given on the same side.

I've built a non-stripped version, but am still waiting for advice on
how to get gdb involved when the problem occurs -- I've tried various
hacks with inetd.conf and /usr/bin/id to get it involved, so far
without success -- suggestions welcomed.

ht
-- 
  Henry S. Thompson, HCRC Language Technology Group, University of Edinburgh
          W3C Fellow 1999--2001, part-time member of W3C Team
     2 Buccleuch Place, Edinburgh EH8 9LW, SCOTLAND -- (44) 131 650-4440
	    Fax: (44) 131 650-4587, e-mail: ht@cogsci.ed.ac.uk
		     URL: http://www.ltg.ed.ac.uk/~ht/

--
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple

Re: OpenSSH crashes - able to repeat the cause

[Jack Radigan]

I'm doing the same at the moment.  When id.exe crashes go over to the
system and do a 'ps' to find out the pid of id.exe.  Then, start gdb with:

	gdb program <pid>

I'm able to see the registers but the stack is lost in space somewhere
and doesn't show anything relevant.

Still playing, maybe you'll have better luck.

-jack-

Jack Radigan - Dun & Bradstreet, Inc.

On 25 Jan 2001 ht@cogsci.ed.ac.uk wrote:

> Corinna Vinschen <cygwin@cygwin.com> writes:
>
> > On Tue, Jan 23, 2001 at 11:01:12AM +0000, Henry S. Thompson wrote:
> > > I'm trying as hard as I can with the resources I have available to
> > > debug, by giving as much detail on the problem which has not otherwise
> > > been provided, in the hopes that they will spark an 'aha!' reaction
> > > from someone.
> > >
> > > Would compiling id (the program which fails) from source with -g be a
> > > useful step?  Presumably not, because the crash goes in to the Win2k
> > > debugger directly -- what would be required to get symbols available
> > > for debugging at that point?  I have a bit of experience debugging C
> >
> > You need a non stripped version of the Cygwin DLL and gdb. The
> > non stripped version is easily produced by downloading the sources
> > from cvs as descibed on http://www.cygwin.com/cvs.html and following
> > the directions given on the same side.
>
> I've built a non-stripped version, but am still waiting for advice on
> how to get gdb involved when the problem occurs -- I've tried various
> hacks with inetd.conf and /usr/bin/id to get it involved, so far
> without success -- suggestions welcomed.
>
> ht
> --
>   Henry S. Thompson, HCRC Language Technology Group, University of Edinburgh
>           W3C Fellow 1999--2001, part-time member of W3C Team
>      2 Buccleuch Place, Edinburgh EH8 9LW, SCOTLAND -- (44) 131 650-4440
> 	    Fax: (44) 131 650-4587, e-mail: ht@cogsci.ed.ac.uk
> 		     URL: http://www.ltg.ed.ac.uk/~ht/
>
> --
> Want to unsubscribe from this list?
> Check out: http://cygwin.com/ml/#unsubscribe-simple
>
>



--
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple

Re: OpenSSH crashes - able to repeat the cause

[Christopher Faylor]

On Thu, Jan 25, 2001 at 10:46:19AM -0500, Jack Radigan wrote:
>I'm doing the same at the moment.  When id.exe crashes go over to the
>system and do a 'ps' to find out the pid of id.exe.  Then, start gdb with:
>
>	gdb program <pid>
>
>I'm able to see the registers but the stack is lost in space somewhere
>and doesn't show anything relevant.

Switch to the main thread (probably thread 1).

cgf

--
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple

Re: OpenSSH crashes - able to repeat the cause

[Christopher Faylor]

On Thu, Jan 25, 2001 at 11:14:30AM -0500, Christopher Faylor wrote:
>On Thu, Jan 25, 2001 at 10:46:19AM -0500, Jack Radigan wrote:
>>I'm doing the same at the moment.  When id.exe crashes go over to the
>>system and do a 'ps' to find out the pid of id.exe.  Then, start gdb with:
>>
>>	gdb program <pid>
>>
>>I'm able to see the registers but the stack is lost in space somewhere
>>and doesn't show anything relevant.
>
>Switch to the main thread (probably thread 1).

Also configure cygwin with "--enable-debugging" and you can set the CYGWIN_SLEEP
environment variable to some number of milliseconds (like 10000).  This will
cause every cygwin application to wait for that number of milliseconds, printing
its pid.  You can then attach gdb to the pid.

You'll have to set CYGWIN_SLEEP as a global environment variable.  Setting it in
.profile or .bashrc probably won't work.

cgf

--
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple

Re: OpenSSH crashes - able to repeat the cause

[Henry S. Thompson]

"Jack Radigan" <jprad@dnb.com> writes:

> I'm doing the same at the moment.  When id.exe crashes go over to the
> system and do a 'ps' to find out the pid of id.exe.  Then, start gdb with:
> 
> 	gdb program <pid>
> 
> I'm able to see the registers but the stack is lost in space somewhere
> and doesn't show anything relevant.
> 
> Still playing, maybe you'll have better luck.

So I don't understand, I tried this yesterday and found that id was
running as root and I couldn't attach, but it isn't now -- maybe it
was sshd that was in that situation. . .

Anyway, I do now have a bit of stack information, not much use I
guess:

(gdb) info threads
  3 thread 26324.0x6770  0x77f9eeaa in _libkernel32_a_iname ()
* 2 thread 26324.0x66d8  0x77f820f0 in _libkernel32_a_iname ()
  1 thread 26324.0x66b0  0x77f820f0 in _libkernel32_a_iname ()


Attaching to program `/usr/bin/id.exe', process 26324
[Switching to thread 26324.0x6770]
(gdb) where
#0  0x77f9eeaa in _libkernel32_a_iname ()
#1  0x77ea8983 in _libkernel32_a_iname ()
#2  0x77e837cd in _libkernel32_a_iname ()

(gdb) info threads
* 3 thread 26324.0x6770  0x77f9eeaa in _libkernel32_a_iname ()
  2 thread 26324.0x66d8  0x77f820f0 in _libkernel32_a_iname ()
  1 thread 26324.0x66b0  0x77f820f0 in _libkernel32_a_iname ()

(gdb) thread 2
[Switching to thread 2 (thread 26324.0x66d8)]
#0  0x77f820f0 in _libkernel32_a_iname ()
(gdb) where
#0  0x77f820f0 in _libkernel32_a_iname ()
#1  0x77e8366e in _libkernel32_a_iname ()
#2  0x77e860c6 in _libkernel32_a_iname ()
#3  0x610041a2 in thread_stub (arg=0x61084554)
    at ../../../../winsup/cygwin/debug.cc:97
#4  0x77e837cd in _libkernel32_a_iname ()

(gdb) thread 1
[Switching to thread 1 (thread 26324.0x66b0)]
#0  0x77f820f0 in _libkernel32_a_iname ()
(gdb) where
#0  0x77f820f0 in _libkernel32_a_iname ()
#1  0x77e8366e in _libkernel32_a_iname ()
#2  0x77e860c6 in _libkernel32_a_iname ()
#3  0x610461b6 in _spawnve (hToken=0x0, mode=3, path=0xa014c58
"/usr/bin/id",
    argv=0xa014cd8, envp=0xa0147f8) at
../../../../winsup/cygwin/spawn.cc:882
#4  0x6100e169 in _execve (path=0xa014c58 "/usr/bin/id",
argv=0xa014cd8,
    envp=0xa0147f8) at ../../../../winsup/cygwin/exec.cc:34
#5  0x6106978a in execve (path=0xa014c58 "/usr/bin/id",
argv=0xa014cd8,
    envp=0xa0147f8) at ../../../../../newlib/libc/posix/execve.c:13
#6  0x411e54 in ?? ()
#7  0x411d80 in ?? ()
#8  0x4112f9 in ?? ()
#9  0x40e5ef in ?? ()
#10 0x444ccf in ?? ()
#11 0x420a12 in ?? ()
#12 0x423cbb in ?? ()
#13 0x41ee7c in ?? ()
#14 0x41eef4 in ?? ()
#15 0x41ef25 in ?? ()
#16 0x41ec91 in ?? ()
#17 0x41e76b in ?? ()
#18 0x41e93c in ?? ()
#19 0x424e84 in ?? ()
#20 0x424918 in ?? ()
#21 0x410cbc in ?? ()
#22 0x40e5ef in ?? ()
#23 0x444ccf in ?? ()
#24 0x44528e in ?? ()
#25 0x445330 in ?? ()
#26 0x401ff7 in _size_of_heap_reserve__ ()
#27 0x401796 in _size_of_heap_reserve__ ()
#28 0x61003825 in dll_crt0_1 () at
../../../../winsup/cygwin/dcrt0.cc:865
#29 0x61003a0d in _dll_crt0 () at
../../../../winsup/cygwin/dcrt0.cc:931
#30 0x61003a4c in dll_crt0 (uptr=0x0) at
../../../../winsup/cygwin/dcrt0.cc:943
#31 0x468d2d in ?? ()
#32 0x40103d in _size_of_heap_reserve__ ()
#33 0x77e992a6 in _libkernel32_a_iname ()

(gdb) up
#1  0x77e8366e in _libkernel32_a_iname ()
(gdb) up
#2  0x77e860c6 in _libkernel32_a_iname ()
(gdb) up
#3  0x610461b6 in _spawnve (hToken=0x0, mode=3, path=0xa014c58
"/usr/bin/id",
    argv=0xa014cd8, envp=0xa0147f8) at
../../../../winsup/cygwin/spawn.cc:882
882           spawn_guts (hToken, path, argv, envp, mode);
Current language:  auto; currently c++
(gdb) p argv
$1 = (char **) 0xa014cd8
(gdb) p *argv
$2 = 0xa014c18 "id"
(gdb) p argv[0]
$3 = 0xa014c18 "id"
(gdb) p argv[1]
$4 = 0xa014708 "-un"
(gdb) p argv[2]
$5 = 0x0


What more information should I now look for?

ht
-- 
  Henry S. Thompson, HCRC Language Technology Group, University of Edinburgh
          W3C Fellow 1999--2001, part-time member of W3C Team
     2 Buccleuch Place, Edinburgh EH8 9LW, SCOTLAND -- (44) 131 650-4440
	    Fax: (44) 131 650-4587, e-mail: ht@cogsci.ed.ac.uk
		     URL: http://www.ltg.ed.ac.uk/~ht/

--
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple

Re: OpenSSH crashes - able to repeat the cause

[Corinna Vinschen]

Hi,

I have uploaded a test version of bash to

	ftp://ftp.hirmke.de/corinna/bash.exe.gz

Please try if that version solves the problem. Don't expect
too much. It's just a shot in the dark.

I would be pleased, if everybody which had that problem would try
this bash. Nobody else, please. The server is connected via 64K
line only.

Corinna

-- 
Corinna Vinschen                  Please, send mails regarding Cygwin to
Cygwin Developer                                mailto:cygwin@cygwin.com
Red Hat, Inc.

--
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple