[4646] in testers
solaris 9.0.3: su
daemon@ATHENA.MIT.EDU (Ken Raeburn)
Tue May 22 12:49:36 2001
To: testers@mit.edu
From: Ken Raeburn <raeburn@MIT.EDU>
Date: 22 May 2001 12:49:21 -0400
Message-ID: <tx1itit1g8e.fsf@mit.edu>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
dcl% su
Password:
PAM-KRB5: Kerberos V5 authentication failed Client not found in Kerberos database
Segmentation fault
dcl%
This is after fixing a problem where some old files Ted left around in
/usr/lib/gss/do were loaded in preference to those in /usr/lib/gss/gl,
causing a different crash. (Miki says a program to check for unwanted
files left around like that should be coming soon, but isn't ready
yet.)
This doesn't happen to me on joust, the other 9.0 Sun I've tried.
It didn't happen for Larry or Miki on their machines either.
Since it's setuid, there's no core file. But my guess is it's getting
an error trying to write something to the console, perhaps an error
returned by the krb5 module?
22676: stat64("/usr/lib/security/pam_krb5.so.1", 0xFFBEE780) = 0
22676: d=0x02180005 i=47405 m=0100755 l=1 u=0 g=3 sz=58628
22676: at = May 22 12:31:25 EDT 2001 [ 990549085 ]
22676: mt = Apr 3 01:01:36 EDT 2001 [ 986274096 ]
22676: ct = May 18 13:16:23 EDT 2001 [ 990206183 ]
22676: bsz=8192 blks=116 fs=ufs
22676: setuid(0) = 0
22676: munmap(0x7F6B2000, 658) = 0
22676: munmap(0x7F6A0000, 3434) = 0
22676: munmap(0x7F682000, 440) = 0
22676: munmap(0x7F670000, 2481) = 0
22676: munmap(0x7F876000, 16960) = 0
22676: munmap(0x7F850000, 88290) = 0
22676: munmap(0x7F968000, 3424) = 0
22676: munmap(0x7F950000, 29980) = 0
22676: munmap(0x7F7FA000, 22692) = 0
22676: munmap(0x7F780000, 432354) = 0
22676: munmap(0x7F83C000, 7915) = 0
22676: munmap(0x7F820000, 52386) = 0
22676: munmap(0x7FAD0000, 1238) = 0
22676: munmap(0x7F776000, 8913) = 0
22676: munmap(0x7F750000, 92620) = 0
22676: munmap(0x7F742000, 13300) = 0
22676: munmap(0x7F710000, 135388) = 0
22676: munmap(0x7F6FE000, 4536) = 0
22676: munmap(0x7F6E0000, 49367) = 0
22676: sigaction(SIGALRM, 0xFFBEE818, 0xFFBEE898) = 0
22676: new: hand = 0x7F91A6B4 mask = 0 0 0 0 flags = 0x0012
22676: old: hand = 0x00000000 mask = 0 0 0 0 flags = 0x0000
22676: alarm(30) = 0
22676: open("/dev/console", O_WRONLY|O_APPEND|O_CREAT, 0666) = 4
22676: llseek(4, 0, SEEK_END) = 0
22676: time() = 990549086
22676: Incurred fault #6, FLTBOUNDS %pc = 0x7F8B9EB8
22676: siginfo: SIGSEGV SEGV_MAPERR addr=0x7F968A98
22676: Received signal #11, SIGSEGV [default]
22676: siginfo: SIGSEGV SEGV_MAPERR addr=0x7F968A98
22676: *** process killed ***
However, the user call trace, if it's to be believed, makes it appear
that the error is *within* the alarm() routine; that just seems
wrong...
-> libpam:pam_end(0x25fb0, 0x0, 0x2321c, 0x0)
-> libc:free(0x27118, 0x27118, 0x0, 0x7f855704)
[...]
-> libc:mutex_unlock(0x7f93c558, 0x40, 0x25fb0, 0x7f9c512c)
<- libc:free() = 0
-> libc:strcmp(0x7f93e9a8, 0x254d0, 0x7f9d6000, 0x25fb0)
-> libc:signal(0xe, 0x1260c, 0x7f9d6000, 0x0)
sigaction(SIGALRM, 0xFFBEE818, 0xFFBEE898) = 0
-> libc:alarm(0x1e, 0x1260c, 0x7f9d6000, 0x0)
alarm(30) = 0
-> libc:fopen(0x254d0, 0x12cb0, 0x0, 0x0)
-> libc_psr:memset(0x7f93c0a8, 0x0, 0x18, 0x7f938000)
<- libc_psr:memset() = 0x7f93c0a8
open("/dev/console", O_WRONLY|O_APPEND|O_CREAT, 0666) = 4
llseek(4, 0, SEEK_END) = 0
<- libc:fopen() = 0x24568
-> libc:time(0x0, 0x12cb0, 0x254d0, 0x0)
time() = 990549554
<- libc:time() = 0x3b0a9632
-> libc:localtime(0xffbee8d4, 0x12cb0, 0x0, 0x0)
Incurred fault #6, FLTBOUNDS %pc = 0x7F8B9EB8
siginfo: SIGSEGV SEGV_MAPERR addr=0x7F968A98
Received signal #11, SIGSEGV [default]
siginfo: SIGSEGV SEGV_MAPERR addr=0x7F968A98
*** process killed ***
Ken
