Hi,

I have noticed that if I do a ssh connection from my Solaris s10_72 x86
machine to any other and close the connection afterwards (i.e. exit the
remote shell), I end up with a core file. E.g.

lokrum% uname -a
SunOS lokrum 5.10 s10_72 i86pc i386 i86pc
lokrum% which ssh
/bin/ssh
lokrum% ssh titan 
Last login: Tue Dec  7 12:09:21 2004 from lokrum.tor.soli
Sun Microsystems Inc.   SunOS 5.8       Generic Patch   October 2001
titan$ exit
exit
Segmentation fault (core dumped)
lokrum% ls -al core.ssh.*
-rw-------   1 dragan   sax      8320475 Dec  7 12:25 core.ssh.644

If I do pstack on the core file, I always get the same backtrace:

lokrum% pstack core.ssh.644 
core 'core.ssh.644' of 644:     ssh titan
 d2512ac1 strlen   (807e3ef, 8047644, 8047610, 0) + 31
 d256aae0 snprintf (804766c, 64, 807e3dc, d) + 74
 080610da client_loop (1, 7e, 0) + 509
 0805a5c2 ???????? (80479d4, 80478bc, d27fb824, 1, 0, 2)
 0805941b main     (0, 8047908, 804790c) + c96
 080586ba ???????? (2, 8047a3c, 8047a40, 0, 8047a46, 8047a6c)

(dbx gives the following one
     (dbx) where                                                                  
=>[1] strlen(0x807e3ef), at 0xd2512ac1 
  [2] _ndoprnt(0x807e3ef, 0x8047644, 0x8047610, 0x0), at 0xd2567c46 
  [3] snprintf(0x804766c, 0x64, 0x807e3dc, 0xd), at 0xd256aae0 
  [4] client_loop(0x1, 0x7e, 0x0), at 0x80610da 
  [5] 0x805a5c2(), at 0x805a5c1 
  [6] main(0x0, 0x8047908, 0x804790c), at 0x805941b 
)

Truss output (of another ssh connection) ends with 

657:	write(4, "B0 +05 .97\f7F98CE\r1BB4".., 32)	= 32
657:	sigaction(SIGWINCH, 0x08047590, 0x08047610)	= 0
657:	ioctl(0, TCGETA, 0x08047628)			= 0
657:	ioctl(0, TCGETA, 0x08047628)			= 0
657:	ioctl(0, TCGETA, 0x08047628)			= 0
657:	    Incurred fault #6, FLTBOUNDS  %pc = 0xD2512AC1
657:	      siginfo: SIGSEGV SEGV_MAPERR addr=0x000000B6
657:	    Received signal #11, SIGSEGV [default]
657:	      siginfo: SIGSEGV SEGV_MAPERR addr=0x000000B6


That has started since I updated to s10_72. What is going on here? Is this
a (known) bug or is this just me?

Thanks and bye, Dragan

-- 
Dragan Cvetkovic, 

To be or not to be is true. G. Boole      No it isn't.  L. E. J. Brouwer

!!! Sender/From address is bogus. Use reply-to one !!!

Dragan Cvetkovic <me@privacy.net> writes:

>I have noticed that if I do a ssh connection from my Solaris s10_72 x86
>machine to any other and close the connection afterwards (i.e. exit the
>remote shell), I end up with a core file. E.g.



Do you have the SUNWcry/SUNWcryr (extra crypto) packages installed?

If so, you need to remove the old ones and upgrade to the
s10_72 specific ones.  You need to keep the two in sync.

Casper
-- 
Expressed in this posting are my opinions.  They are in no way related
to opinions held by my employer, Sun Microsystems.
Statements on Sun products included here are not gospel and may
be fiction rather than truth.

Casper H.S. Dik <Casper.Dik@Sun.COM> writes:

> Dragan Cvetkovic <me@privacy.net> writes:
>
>>I have noticed that if I do a ssh connection from my Solaris s10_72 x86
>>machine to any other and close the connection afterwards (i.e. exit the
>>remote shell), I end up with a core file. E.g.

> Do you have the SUNWcry/SUNWcryr (extra crypto) packages installed?

Yes,  I do.

> If so, you need to remove the old ones and upgrade to the
> s10_72 specific ones.  You need to keep the two in sync.

That's was it, thanks Casper. I couldn't find the newer version of extra
crypto packages, so I simply removed them. Hope nothing will break.

Thanks and bye, Dragan

-- 
Dragan Cvetkovic, 

To be or not to be is true. G. Boole      No it isn't.  L. E. J. Brouwer

!!! Sender/From address is bogus. Use reply-to one !!!

Dragan Cvetkovic <me@privacy.net> writes:

>> Do you have the SUNWcry/SUNWcryr (extra crypto) packages installed?

>Yes,  I do.

This was a fun bug (not) to chase; we went through that
internally a bit earlier.  I've send a note to the platbeta people
to get it mentioned in the release notes.

>That's was it, thanks Casper. I couldn't find the newer version of extra
>crypto packages, so I simply removed them. Hope nothing will break.

Long keys; the crypto stuff should be out there shortly.

Casper
-- 
Expressed in this posting are my opinions.  They are in no way related
to opinions held by my employer, Sun Microsystems.
Statements on Sun products included here are not gospel and may
be fiction rather than truth.