openssh bugs - Mar 2007 - [Bug 1290] sshd dies if passed host key with relative path on command line

http://bugzilla.mindrot.org/show_bug.cgi?id=1290

           Summary: sshd dies if passed host key with relative path on
                    command line
           Product: Portable OpenSSH
           Version: 4.5p1
          Platform: All
        OS/Version: FreeBSD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: sshd
        AssignedTo: bitbucket at mindrot.org
        ReportedBy: bugzilla.midrot.org at beej.org


the following works fine on linux:
root at ubuntu1:/test# /usr/sbin/sshd -D -h ssh_host_dsa_key -p 6789

on FreeBSD and OSX, once a client connects, the server exits.
changing the path for the host key to an absolute one allows the server
to function normally:
beejmacbook:/test beej$ /opt/local/sbin/sshd -h /test/ssh_host_dsa_key
-D -ddd -p 6789




------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

http://bugzilla.mindrot.org/show_bug.cgi?id=1290


dtucker at zip.com.au changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         OS/Version|FreeBSD                     |All
             Status|NEW                         |ASSIGNED




------- Comment #1 from dtucker at zip.com.au  2007-03-05 11:51 -------
What version is it that you're testing on Linux, and do you see the
same results with a version built form vanilla source?

I suspect it's not dependant on the the OS but rather on whether or not
re-exec is enabled.




------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

http://bugzilla.mindrot.org/show_bug.cgi?id=1290





------- Comment #2 from bugzilla.midrot.org at beej.org  2007-03-05 16:15
-------
the testing i did for this bug report was with ubuntu's sshd 
OpenSSH_4.3p2 from their 4.3p2-5ubuntu1 deb.

anyway, using vanilla 4.5p1 source with all defaults on the same linux
test box does indeed break things like the other OS's when using a
relative path for host key.  and adding the -r flag fixes things again
so you're right about this being related to re-exec :)




------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.