[5146] in Athena Bugs

home help back first fref pref prev next nref lref last post

kinit/aklog

daemon@ATHENA.MIT.EDU (qjb@ATHENA.MIT.EDU)
Mon Jun 11 10:35:35 1990

From: qjb@ATHENA.MIT.EDU
Date: Mon, 11 Jun 90 10:35:13 -0400
To: bugs@ATHENA.MIT.EDU
In-Reply-To: [5145] in bugs


As a followup to the last message, after looking at the core
dump of aklog, I see that it dumped core during this line:

	    fprintf(stderr,":\n%s\n", krb_err_txt[status]);

in the midst of this piece of code:

	status = krb_get_cred(name, instance, realm_of_cell, &c);
	if (status != KSUCCESS) {
	    status = get_ad_tkt(name, instance, realm_of_cell, 255);
	    if (status == KSUCCESS)
		status = krb_get_cred(name, instance, realm_of_cell, &c);
	}
	
	if (status != KSUCCESS) {
	    fprintf(stderr, "%s: Couldn't get AFS tickets", progname);
	    fprintf(stderr," for cell %s", cell_to_use);
	    fprintf(stderr,":\n%s\n", krb_err_txt[status]);
	    return(AKLOG_KERBEROS);
	}


In other words, one of the three kerberos library routines above
returned a bogus error status that turned out to be outside the
bounds of krb_err_txt.


===== stack trace of core.aklog ======

$c
__doprnt() from _fprintf+35
_fprintf(10b54,c1df,242070) from 3c2
_auth_to_cell(7fffdcec,7fffe0ec) from fd9
_main(2,7fffe148,7fffe154) from start+3d

home help back first fref pref prev next nref lref last post