This is the mail archive of the
libc-hacker@sourceware.cygnus.com
mailing list for the glibc project.
linuxconf complexity
- To: libc-alpha@cygnus.com
- Subject: linuxconf complexity
- From: "Jack Howarth" <howarth@bromo.med.uc.edu>
- Date: Sat, 31 Oct 1998 16:38:58 -0500
I thought everyone interested in this might like to read the explanation
Andrew Morgan who maintains pam gave me as to the complexity we are dealing
with in linuxconf with password authentication through pam. He said...
-----------------------------------------------------------------------
Jack,
I see now: linuxconf loads redhat.so which loads libpam.so which tries
to load the pam_pwdb.so module. Does redhat.so happen to load libpam.so
with the RTLD_GLOBAL flag?
dlopen("libpam.so", RTLD_GLOBAL);
If this is done, based on the pre-glibc behavior, pam_pwdb.so should be
able to resolve the symbols in libpam.
[I don't run glibc (yet) so I'm not really able to debug the problem you
are seeing.]
Cheers
Andrew
-------------------------------------------------------------------------
Andreas Jaeger has kindly provided me with test code. Unfortunately
I am seeing no failures with it when the test module is loaded either
with RTLD_LAZY or RTLD_GLOBAL. However I think we don't quite have the
required level of complexity. From what I can see I need a test program
which loads a shared module which in turn calls its own shared module.
I am wondering if we really are passing the symbols all the way back up
to the orginal calling program.
Jack
--
------------------------------------------------------------------------------
Jack W. Howarth, Ph.D. 231 Bethesda Avenue
NMR Facility Director Cincinnati, Ohio 45267-0524
Dept. of Molecular Genetics phone: (513) 558-4420
Univ. of Cincinnati College of Medicine fax: (513) 558-8474