openssh bugs - Aug 2006 - [Bug 926] pam_session_close called as user or not at all

http://bugzilla.mindrot.org/show_bug.cgi?id=926





------- Comment #21 from t8m at centrum.cz  2006-08-05 01:18 -------
The patch causes a regression with pam_krb5 module.

See https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=201341

As I said above I think that the only correct solution which would
solve all cases (privsep yes/no, root/regular user) would be to add
another fork before the setuid calls and shell process exec.

login does this:
1. call pam_open_session
2. fork
3. parent waits for child, child impersonates user, execs shell
4. when child exits, parent calls pam_close_session




------- 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=926


dtucker at zip.com.au changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
OtherBugsDependingO|1155                        |
              nThis|                            |




------- Comment #22 from dtucker at zip.com.au  2006-08-20 15:58 -------
(In reply to comment #21)
> The patch causes a regression with pam_krb5 module.
> See https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=201341
Thanks for giving it a spin in Fedora.  Does this particular problem
also occur with PrivSep=no?
> As I said above I think that the only correct solution which would
> solve all cases (privsep yes/no, root/regular user) would be to add
> another fork before the setuid calls and shell process exec.
Would there be any downside to setting KRB5CCNAME in the parent too?

(since it causes a regression, I'm taking this bug out of the list for
4.4 pending further work.)




------- 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=926





------- Comment #23 from t8m at centrum.cz  2006-08-23 22:03 -------
(In reply to comment #22)
> (In reply to comment #21)
> > The patch causes a regression with pam_krb5 module.
> > See https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=201341
> 
> Thanks for giving it a spin in Fedora.  Does this particular problem
> also occur with PrivSep=no?
I don't think that this occurs with privsep disabled.
> > As I said above I think that the only correct solution which would
> > solve all cases (privsep yes/no, root/regular user) would be to add
> > another fork before the setuid calls and shell process exec.
> 
> Would there be any downside to setting KRB5CCNAME in the parent too?
I don't know of any however note that with privsep disabled or when
called as root the pam_session_close still won't be called correctly.




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