[Date Prev][Date Next] [Chronological] [Thread] [Top]

SASL/GSSAPI with 2.1.4 causes segfaults


I'm wondering if anyone else has seen the following problem with 2.1.4.

I'm running:
Berkeley DB 4.0.14

I upgraded my installation from 2.1.3 leaving the rest the same, and now every time I try to do a K5 bind (which worked flawlessly under 2.1.3), it seg faults. Note that the database was deleted, so as to start fresh. I then import the initial part of our database. Using ldapsearch -x works without problem. ldapsearch with GSSAPI results in:

dnPrettyNormal: <>
<<< dnPrettyNormal: <>, <>
do_sasl_bind: dn () mech GSSAPI
conn=0 op=3 BIND dn="" method=163
==> sasl_bind: dn="" mech=<continuing> datalen=53
SASL Canonicalize [conn=0]: authcid="quanah"
slap_sasl_getdn: id=quanah
getdn: u:id converted to uid=quanah,cn=stanford.edu,cn=GSSAPI,cn=auth
dnNormalize: <uid=quanah,cn=stanford.edu,cn=GSSAPI,cn=auth>
=> ldap_bv2dn(uid=quanah,cn=stanford.edu,cn=GSSAPI,cn=auth,0)
<= ldap_bv2dn(uid=quanah,cn=stanford.edu,cn=GSSAPI,cn=auth,0)=0
=> ldap_dn2bv(272)
<= ldap_dn2bv(uid=quanah,cn=stanford.edu,cn=gssapi,cn=auth,272)=0
<<< dnNormalize: <uid=quanah,cn=stanford.edu,cn=gssapi,cn=auth>
==>slap_sasl2dn: converting SASL name uid=quanah,cn=stanford.edu,cn=gssapi,cn=auth to a DN
slap_sasl_regexp: converting SASL name uid=quanah,cn=stanford.edu,cn=gssapi,cn=auth
Segmentation Fault

Running truss shows nothing interesting except:

Incurred fault #6, FLTBOUNDS %pc = 0xFEDB337C
siginfo: SIGSEGV SEGV_MAPERR addr=0x00000000
Received signal #11, SIGSEGV [default]
siginfo: SIGSEGV SEGV_MAPERR addr=0x00000000
*** process killed ***

Quanah Gibson-Mount
Senior Systems Administrator
ITSS/TSS/Computing Systems
Stanford University
GnuPG Public Key: http://www.stanford.edu/~quanah/pgp.html

Attachment: pgpyL5L8zw8V5.pgp
Description: PGP signature