Ticket History #383: com_err_hook declared static in source, extern in header

From hartmans@MIT.EDU Thu Feb 27 21:58:21 1997
Received: from MIT.EDU (SOUTH-STATION-ANNEX.MIT.EDU [18.72.1.2]) by rt-11.MIT.EDU (8.7.5/8.7.3) with SMTP id VAA14216 for <bugs@RT-11.MIT.EDU>; Thu, 27 Feb 1997 21:58:20 -0500
Received: from LUMINOUS.MIT.EDU by MIT.EDU with SMTP
id AA29057; Thu, 27 Feb 97 21:58:19 EST
Received: (from hartmans@localhost) by luminous.MIT.EDU (8.8.5/8.6.12) id VAA29387; Thu, 27 Feb 1997 21:58:18 -0500 (EST)
Message-Id: <199702280258.VAA29387@luminous.MIT.EDU>
Date: Thu, 27 Feb 1997 21:58:18 -0500 (EST)
From: Sam Hartman <hartmans@MIT.EDU>
Reply-To: hartmans@MIT.EDU
To: krb5-bugs@MIT.EDU
Cc: probe@MIT.EDU
Subject: klist dumps core if ccache not found
X-Send-Pr-Version: 3.99

MIT

System: NetBSD luminous 1.2B NetBSD 1.2B (LUMINOUS) #17: Thu Dec 19 02:15:35 EST 1996 hartmans@luminous:/u1/usr/src/sys/arch/i386/compile/LUMINOUS i386



I built the current sources and klist dumps core if the ccache cannot be found. I know I am using the right shared libraries.


It turns out that com_err_hook was declared extern in
com_err.h and static in com_err.c. Both gas and gld gave warnings
that something strange was going on; they were right.


hartmans@luminous:tmp(1)> cd /tmp
hartmans@luminous:/tmp(2)> KRB5CCNAME=/tmp/foobazuu klist
Segmentation fault (core dumped)
hartmans@luminous:/tmp(3)> which klist
/usr/local/bin/klist
hartmans@luminous:/tmp(4)> gdb /usr/local/build/krb5/clients/klist/klist klist.core
GDB is free software and you are welcome to distribute copies of it
under certain conditions; type "show copying" to see the conditions.
There is absolutely no warranty for GDB; type "show warranty" for details.
GDB 4.11 (i386-netbsd), Copyright 1993 Free Software Foundation, Inc...
Core was generated by `klist'.
Program terminated with signal 11, Segmentation fault.
Reading symbols from /usr/libexec/ld.so...done.
Reading symbols from /usr/local/lib/libkrb5.so.1.1...done.
Reading symbols from /usr/local/lib/libcrypto.so.1.1...done.
Reading symbols from /usr/local/lib/libcom_err.so.3.0...done.
Reading symbols from /usr/lib/libc.so.12.10...done.
#0 0xb018680 in end ()
(gdb) bt
#0 0xb018680 in end ()
#1 0x1006942e in com_err_va (whoami=0xf7bfdb74 "klist", code=-1765328189,
fmt=0x21ed "while setting cache flags (ticket cache %s)",
ap=0xf7bfd648 "@`") at ../../../../src/krb5/src/util/et/com_err.c:104
#2 0x1006945a in com_err (whoami=0xf7bfdb74 "klist", code=-1765328189,
fmt=0x21ed "while setting cache flags (ticket cache %s)")
at ../../../../src/krb5/src/util/et/com_err.c:127
#3 0x2462 in do_ccache (name=0x0)
at ../../../../src/krb5/src/clients/klist/klist.c:283
#4 0x1cb2 in main (argc=1, argv=0xf7bfd6d8)
at ../../../../src/krb5/src/clients/klist/klist.c:156
(gdb) up
#1 0x1006942e in com_err_va (whoami=0xf7bfdb74 "klist", code=-1765328189,
fmt=0x21ed "while setting cache flags (ticket cache %s)",
ap=0xf7bfd648 "@`") at ../../../../src/krb5/src/util/et/com_err.c:104
104 (*com_err_hook)(whoami, code, fmt, ap);
(gdb) p com_err_hook
$1 = (void (*)()) 0x63657865 <end+1398322573>
(gdb) down
#0 0xb018680 in end ()
x/i $eip
0xb018680 <end+184632152>: Cannot access memory at address 0xb018680.
(gdb) p $eip
$2 = (void *) 0xb018680
(gdb) quit
hartmans@luminous:/tmp(5)> ldd /usr/local/bin/klist
/usr/local/bin/klist:
-lkrb5.1 => /usr/local/lib/libkrb5.so.1.1 (0x1001a000)
-lcrypto.1 => /usr/local/lib/libcrypto.so.1.1 (0x1005b000)
-lcom_err.3 => /usr/local/lib/libcom_err.so.3.0 (0x10069000)
-lc.12 => /usr/lib/libc.so.12.10 (0x1006b000)
hartmans@luminous:/tmp(6)> rm klist.core
hartmans@luminous:/tmp(7)>

Do not declare the same identifier with both internal and external linkage.


From: "Richard Basch" <basch@lehman.com>
To: hartmans@MIT.EDU
Cc: krb5-bugs@MIT.EDU
Subject: Re: krb5-clients/383: klist dumps core if ccache not found
Date: Thu, 27 Feb 1997 22:23:28 -0500

I can't reproduce the coredump, so if you can narrow down the problem,
that would be appreciated... I don't have access to a NetBSD box, and
it doesn't fail on any other platform on which I have access.

765-1$ KRB5CCNAME=/ldj /opt/krb5/bin/klist
klist: No credentials cache file found while setting cache flags (ticket cache /ldj)
766-1$

I will try to run Purify over everything, in the hopes that maybe it
will turn up something... however, I do not have high hopes...
--
Richard Basch
Sr. Developer/Analyst, DSO URL: http://web.mit.edu/basch/www/home.html
Lehman Brothers, Inc. Email: basch@lehman.com, basch@mit.edu
101 Hudson St., 38th Floor Fax: +1-201-524-5828
Jersey City, NJ 07302-3988 Voice: +1-201-524-5049


From: "Richard Basch" <basch@lehman.com>
To: "Richard Basch" <basch@lehman.com>
Cc: hartmans@MIT.EDU, krb5-bugs@MIT.EDU
Subject: Re: krb5-clients/383: klist dumps core if ccache not found
Date: Fri, 28 Feb 1997 00:46:27 -0500

On Thu, 27-February-1997, "Richard Basch" wrote to "hartmans@MIT.EDU, krb5-bugs@MIT.EDU" saying:


Purify output:
Finished klist ( 0 errors, 0 leaked bytes)
Purify instrumented ./klist (pid 26490 at Fri Feb 28 00:39:26 1997)
Current file descriptors in use: 7
Memory leaked: 0 bytes (0%); potentially leaked: 0 bytes(0%)
Program exited with status code 1.

Doesn't NetBSD also have incompatible comerr libraries (version 2.x)?
Make sure your config.cache didn't cache the wrong value and that you
are linked against the correct version of the comerr library. Anyway,
without further information, I am not sure I can figure out what is
going on.

--
Richard Basch
Sr. Developer/Analyst, DSO URL: http://web.mit.edu/basch/www/home.html
Lehman Brothers, Inc. Email: basch@lehman.com, basch@mit.edu
101 Hudson St., 38th Floor Fax: +1-201-524-5828
Jersey City, NJ 07302-3988 Voice: +1-201-524-5049


From: Sam Hartman <hartmans@MIT.EDU>
To: "Richard Basch" <basch@lehman.com>
Cc: krb5-unassigned@RT-11.MIT.EDU, krb5-bugs@MIT.EDU
Subject: Re: krb5-clients/383: klist dumps core if ccache not found
Date: 28 Feb 1997 09:10:22 -0500



Yes, which is one good reason to seriously consider
maintaining compatability between new error table .c files and old
libcom_err libraries. (I think you will set the problem on BSDI and
many operating systems that have native KRB4 support).

However, I was linked against the right library. I will run through the code later today.

From: "Richard Basch" <basch@lehman.com>
To: Sam Hartman <hartmans@MIT.EDU>
Cc: "Richard Basch" <basch@lehman.com>, krb5-unassigned@RT-11.MIT.EDU,
krb5-bugs@MIT.EDU
Subject: Re: krb5-clients/383: klist dumps core if ccache not found
Date: Fri, 28 Feb 1997 10:05:24 -0500

On , 28-February-1997, "Sam Hartman" wrote to "Richard Basch, krb5-unassigned@RT-11.MIT.EDU, krb5-bugs@mit.edu" saying:


I believe I have maintained compatibility with the old comerr, as
written by MIT and as present in the Athena release. Since I do not
know what was done to the BSDI and NetBSD comerr, I do not know how to
maintain compatibility. If someone can give me a handle to that source
excerpt, I can see whether it should interoperate.

--
Richard Basch
Sr. Developer/Analyst, DSO URL: http://web.mit.edu/basch/www/home.html
Lehman Brothers, Inc. Email: basch@lehman.com, basch@mit.edu
101 Hudson St., 38th Floor Fax: +1-201-524-5828
Jersey City, NJ 07302-3988 Voice: +1-201-524-5049


From: Sam Hartman <hartmans@MIT.EDU>
To: "Richard Basch" <basch@lehman.com>
Cc: krb5-bugs@MIT.EDU
Subject: Re: krb5-clients/383: klist dumps core if ccache not found
Date: 28 Feb 1997 15:09:27 -0500

So, I understand why it is crashing, but not why the why is:

* com_err_hook is getting set to some large address in the middle of
no where by the time the first line of main is executed.

I'm looking into potential shared library build problems.

--Sam

Responsible-Changed-From-To: krb5-unassigned->hartmans
Responsible-Changed-By: hartmans
Responsible-Changed-When: Sat Mar 1 12:34:52 1997
Responsible-Changed-Why:
Dealing with problem.

State-Changed-From-To: open-closed
State-Changed-By: hartmans
State-Changed-When: Sat Mar 1 12:36:01 1997
State-Changed-Why:
ANSI 3.1.2.2 says you lose.



From: Sam Hartman <hartmans@MIT.EDU>
To: hartmans@MIT.EDU
Cc: krb5-bugs@MIT.EDU
Subject: Re: krb5-libs/383: com_err_hook declared static in source, extern in header
Date: Sat, 1 Mar 1997 12:35:43 -0500 (EST)

`Sam Hartman' made changes to this PR.

--- /tmp/gnatsa15756 Sat Mar 1 12:31:03 1997
+++ /tmp/gnatsb15756 Sat Mar 1 12:34:11 1997
@@ -13,8 +13,8 @@
X-Send-Pr-Version: 3.99

->Category: krb5-clients
->Synopsis: klist dumps core if ccache not found
+>Category: krb5-libs
+>Synopsis: com_err_hook declared static in source, extern in header
@@ -37,6 +37,10 @@

I built the current sources and klist dumps core if the ccache cannot be found. I know I am using the right shared libraries.

+
+ It turns out that com_err_hook was declared extern in
+com_err.h and static in com_err.c. Both gas and gld gave warnings
+that something strange was going on; they were right.


@@ -93,6 +97,8 @@
hartmans@luminous:/tmp(6)> rm klist.core
hartmans@luminous:/tmp(7)>
+
+Do not declare the same identifier with both internal and external linkage.