From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7.6) Gecko/20050323 Firefox/1.0.2 Fedora/1.0.2-1.3.1 Description of problem: After updating to kernel-2.6.11-1.14_FC3 from updates, ntpd fails to restart, as a service, or will die of SIGSEGV if launched directly. This is confirmed on two different hosts, one is an old Athlon 700 on a AMD-751 MB, the other one is a more recent P4 (Dell PowerEdge 750), running kernel-smp-2.6.11-1.14_FC3, BTW. Version-Release number of selected component (if applicable): kernel-2.6.11-1.14_FC3 How reproducible: Sometimes Steps to Reproduce: 1. Create a small configuration file for ntpd, like this one: cat /tmp/ntp-tmp.conf server 0.pool.ntp.org server 1.pool.ntp.org server 2.pool.ntp.org 2. Depending on your selinux configuration, you may need to set the right type: chcon -t net_conf_t /tmp/ntp-tmp.conf 3. Run: ntpd -q -g -c /tmp/ntp-tmp.conf Actual Results: [root@vis01 ~]# ntpd -q -g -c /tmp/ntp-tmp.conf Segmentation fault [root@srv03 ~]# ntpd -q -g -c /tmp/ntp-tmp.conf Segmentation fault Expected Results: ntpd should set the date after a while (anyway, should not SIGSEGV) Additional info: This is a regression. I've got many hosts running either ntpd as a client (as a replacement for ntpdate) or as a server, all tested with all kernels from FC3 before 2.6.11-1.14_FC3. A quick test shows the problems exists also if ntpd is used in normal daemon mode. On the P4, reverting to 2.6.10-1.770_FC3 solves the problem. The Athlon server used to run 2.6.10-1.770_FC3 just fine. Weird enough, using strace it seems to fail at different, random, places: run1: read(10, "#\n# /etc/nsswitch.conf\n#\n# An ex"..., 4096) = 1686 read(10, "", 4096) = 0 close(10) = 0 munmap(0xb7ffb000, 4096) = 0 --- SIGSEGV (Segmentation fault) @ 0 (0) --- +++ killed by SIGSEGV +++ run2: gettimeofday({1114001558, 283781}, NULL) = 0 gettimeofday({1114001558, 284029}, NULL) = 0 --- SIGSEGV (Segmentation fault) @ 0 (0) --- +++ killed by SIGSEGV +++ run3: rt_sigprocmask(SIG_BLOCK, NULL, [], 8) = 0 adjtimex({modes=61, offset=0, freq=0, maxerror=16, esterror=16, status=64, constant=0, precision=1, tolerance=33554432, time={1114001571, 330177}}) = 5rt_sigaction(SIGSYS, {SIG_DFL}, NULL, 8) = 0 --- SIGSEGV (Segmentation fault) @ 0 (0) --- +++ killed by SIGSEGV +++ Sometimes it even works fine. I've never managed to have it not fail w/o strace. With strace, on the P4 it succeeds more often than on the Athlon (I'm not implying this is anyhow useful info - just reporting everything I noticed).
*** This bug has been marked as a duplicate of 145258 ***