Red Hat Bugzilla – Bug 18187
SegFault of __strcpy() in glibc.so.6
Last modified: 2014-03-16 22:16:31 EDT
/sbin/checkpc and syslogd both segfault for the same reason (ie, a call to __strcpy() in glibc.so.6, or atleast that's the point gdb returns a
Platform is an upgraded RH6.2 -> 7.0, K6-3 box.
No, that probably means just a segfault somewhere in the main program which
has stripped symbols (try to compile a simple program calling strcpy and then
*(int *)0 = 0 in main, strip it and see what gdb shows you. The __strcpy symbol
comes from the .plt and as there are no more symbols, you get __strcpy+something.
So, could you please fill this as a sysklogd bug (but with description on how
you get it to segfault, it does not segfault for me - BTW: do you have the
security errata for sysklogd?) and I don't know where checkpc comes from?
checkpc is part of LPRng. What does rpm -q glibc say?
Okay... this isn't a sysklogd problem alone. checkpc, syslogd and ncftp all segfault. Nothing special needs to be done, just run them.
syslogd will run but rang (annoying as start up). Using the -d flag will make it segfault and core dump.
ncftp and checkpc just need to be run in order to create a segfault and core dump (yes, both).
rpm -q glibc returns glibc-2.1.92-14
Can you rebuild a debugging version of ncftp and get a backtrace
of where it crashes?
Do you have your locale set to anything in particular?
Created attachment 3698 [details]
strace of ncftp (rpm standard)
locale is set to en_US
Can you post your /etc/hosts file, and your /etc/nsswitch.conf?
Well... /etc/hosts was revealing. Removed the the last entry. Things don't segfault now. Closed(?).
126.96.36.199 Minion.gmu.edu Minion
127.0.0.1 localhost.localdomain localhost
188.8.131.52 Minion Minion