Bug#186393: time error (random time jump) in libc

[GOTO Masanori <gotom@debian.or.jp>]

reassign 186393 kernel-source-2.4.20
thanks

At Wed, 26 Mar 2003 14:21:20 -0600,
Jeff Hahn wrote:
> KERNEL: Linux bambusa 2.4.20-686 #1 Mon Jan 13 22:22:30 EST 2003 i686
> unknown unknown GNU/Linux
> 
> Kernel: kernel-image-2.4-686_2.4.20-5_i386.deb
> Libc: libc6-2.3.1-14
> Gcc: gcc-2.95.4-17
> Gcc: gcc-3.0-3.0.4-7
> Ntp: ntp-1:4.1.0-8
> 
> I noticed out of sequences in various log files, etc.  Was not restricted
> just to syslog, or Exim, or any other particular program.  NTP is locked
> solid to a stratum 1 server and is not reporting any changes of server,
> synchronization lost, etc.
> 
> The error is always the same (within a couple of seconds), about 72 minutes
> ahead, then reverts to correct time. 
> 
> The source code of error test is below (c1.c).  Error seems to occur
> randomly, is frequent enough that you'll get several over a 10-15 minute
> period.

> Source code (c1.c)
> #include <stdio.h>
> #include <time.h>
> #include <unistd.h>
> 
> int main (int argc, char **argv) {
> 
>         time_t current, elapsed, past=0;
> 
>         while ( -1 ) {
>                 current = time(NULL);

time() gets the time current value from system call time(2).

> Current: 1048713075   Last: 1048708779  Elapsed: 4296
> Current: 1048713076   Last: 1048713075  Elapsed: 1
> Current: 1048708783   Last: 1048713076  Elapsed: -4293

Interesting, but I think your RTC is broken.  So I don't think it's
kernel bug, but I reassign it to the kernel package to make sure.

Regards,
-- gotom