[ntp:hackers] BusError from ntpd on FreeBSD 9.1

Hal Murray hmurray at megapathdsl.net
Sun Aug 4 00:05:17 UTC 2013


Does anybody recognize this one?  Or have suggestions on how to get more info?

It starts normally at boot time.  It crashes if I restart it.


Aug  3 07:54:10 new-ted ntpd[26953]: ntpd 4.2.7p380-HGM at 1.2948 Sat Aug  3 
14:50:26 UTC 2013 (1): Starting
Aug  3 07:54:10 new-ted ntpd[26954]: ntpd exiting on signal 10 (Bus error: 10)
Aug  3 07:54:10 new-ted kernel: pid 26954 (ntpd), uid 0: exited on signal 10 
(core dumped)

gdb says:
...
#7  <signal handler called>
#8  init_auth () at ../../libntp/authkeys.c:115
#9  0x0805a00c in ntpdmain (argc=0, argv=0xbfbfebe4) at ../../ntpd/ntpd.c:775
#10 0x0805a602 in main (argc=Cannot access memory at address 0x53
) at ../../ntpd/ntpd.c:274

>From libntp/authkeys.c:
        size_t newalloc;
        newalloc = authhashbuckets * sizeof(key_hash[0]);
        key_hash = erealloc(key_hash, newalloc);    <== line 115
        INIT_DLIST(key_listhead, llink);

newalloc is 256, key_hash is 0

It's configured with   --without-crypto
But I didn't see any interesting ifdefs in that area.


This is a new(ish) system, but it works well enough to run the compiler and 
such and ntpd seems to work OK when it doesn't do this.  (I haven't got the 
serial ports working yet, but I don't use FreeBSD enough to be familar with 
them and I haven't tried very hard.)

9.1-RELEASE FreeBSD

I normally grab the new bits whenever Harlan pushes out a new point release.  
I saw this before but didn't track it this far and/or it went away when I 
rebooted and I didn't try restart back then.

I think it doesn't do this most of the time, but I'm not sure.

The motherboard is an Intel D2500CCE
  CPU: Intel(R) Atom(TM) CPU D2500   @ 1.86GHz (1866.77-MHz 686-class CPU)
  real memory  = 4294967296 (4096 MB)
  avail memory = 3387453440 (3230 MB)


-- 
These are my opinions.  I hate spam.





More information about the hackers mailing list