On Fri, 2008-09-05 at 12:41 -0700, Henry B. Hotz wrote:
> On Sep 5, 2008, at 1:47 AM, Mark Phalan wrote:
> 
> > On Thu, 2008-09-04 at 11:41 -0700, Henry B. Hotz wrote:
> >> On Sep 4, 2008, at 4:37 AM, Mark Phalan wrote:
> >>>
> >>> Can you elaborate more on the seg-fault? I fixed this in S10u5 and
> >>> Nevada which sounds related:
> >>>
> >>> 6644742 kadmind cores when using an 'afs3' salt and password > 8  
> >>> chars
> >>>
> >>> you can see (a very little bit) more at bugs.opensolaris.org.
> >>>
> >>> -M
> >>
> >> OK, you got me.  I haven't checked that recently.  Maybe you've
> >> already done what I (minimally) want.  ;-)
> >
> > Perhaps :)
> >
> >> I've seen it in kinit and kpasswd.  It's been long enough since our
> >> migration that I don't have many test cases left, but the ones we
> >> still have are kind of awkward.  Presumably the real problem was  
> >> under
> >> krb5_get_init_creds_password().
> >>
> >> Did you mean kadmin (the user's client) vice kadmind (the server)?
> >
> > kadmind was crashing. I didn't see anything else crashing with afs3
> > passwords but that probably doesn't mean too much as we haven't been
> > testing much with an afs3 salt.
> > The problem was in string2key.c:
> >
> > http://src.opensolaris.org/source/diff/onnv/onnv-gate/usr/src/lib/gss_mechs/mech_krb5/crypto/des/string2key.c?r2=%2Fonnv%2Fonnv-gate%2Fusr%2Fsrc%2Flib%2Fgss_mechs%2Fmech_krb5%2Fcrypto%2Fdes%2Fstring2key.c%405826&r1=%2Fonnv%2Fonnv-gate%2Fusr%2Fsrc%2Flib%2Fgss_mechs%2Fmech_krb5%2Fcrypto%2Fdes%2Fstring2key.c%400
> >
> > The fix was just to resync some code back from MIT.
> >
> > I'll play around with afs3 salts some more and see if I can get  
> > anything
> > to crash/behave badly.
> >
> > -M
> 
> I'm very confused as to why kadmind (which should only see a service  
> ticket, not a password, right?) should care what salt type was used to  
> get that ticket.

kadmind was crashing when creating a new key with an afs salt.

-M

> 
> The client code doesn't always crash on AFS-salted passwords.  If it  
> doesn't crash outright, it usually seems to work.  I've had S9 kinit  
> work but not S9 kpasswd, and vice versa.  If you want help with a  
> server to test against contact me off-list, but I won't promise very  
> much.

Thanks for the offer.

-M


Reply via email to