[Bug 819] New: uptime wrong on x86-64 (fwd)

[Bug 819] New: uptime wrong on x86-64 (fwd)

[Martin J. Bligh]

           Summary: uptime wrong on x86-64
    Kernel Version: 2.5.72
            Status: NEW
          Severity: normal
             Owner: johnstul@us.ibm.com
         Submitter: johnstul@us.ibm.com


Distribution: SLES8  
Hardware Environment: AMD64 Melody box  
Software Environment: 2.5.72  
Problem Description:   
  
uptime reports wrong value after booting  
  
Steps to reproduce:  
jstultz@elm3b31:~/linux-2.5> uptime  
  
Actual result:  
  1:19pm  up 12220 days, 20:19,  1 user,  load average: 0.00, 0.00, 0.00  
  
Expected result:  
  1:19pm  up 15 min,  1 user,  load average: 0.00, 0.01, 0.02  
  
Additional info:  
jstultz@elm3b31:~/linux-2.5> date +%s  
1055881186  
jstultz@elm3b31:~/linux-2.5> cat /proc/stat | grep btime  
btime 1055880294  
  
I've been unable to reproduce the problem on i386, however this has also been  
reported against sparc64 by Daniel Whitener.    
  
I suspect it might have something to do w/ do_posix_clock_monotonic_gettime() on  
these arches.

Re: [Bug 819] New: uptime wrong on x86-64 (fwd)

[george anzinger]

Martin J. Bligh wrote:
>            Summary: uptime wrong on x86-64
>     Kernel Version: 2.5.72
>             Status: NEW
>           Severity: normal
>              Owner: johnstul@us.ibm.com
>          Submitter: johnstul@us.ibm.com
> 
> 
> Distribution: SLES8  
> Hardware Environment: AMD64 Melody box  
> Software Environment: 2.5.72  
> Problem Description:   
>   
> uptime reports wrong value after booting  
>   
> Steps to reproduce:  
> jstultz@elm3b31:~/linux-2.5> uptime  
>   
> Actual result:  
>   1:19pm  up 12220 days, 20:19,  1 user,  load average: 0.00, 0.00, 0.00  
>   
> Expected result:  
>   1:19pm  up 15 min,  1 user,  load average: 0.00, 0.01, 0.02  
>   
> Additional info:  
> jstultz@elm3b31:~/linux-2.5> date +%s  
> 1055881186  
> jstultz@elm3b31:~/linux-2.5> cat /proc/stat | grep btime  
> btime 1055880294  
>   
> I've been unable to reproduce the problem on i386, however this has also been  
> reported against sparc64 by Daniel Whitener.    
>   
> I suspect it might have something to do w/ do_posix_clock_monotonic_gettime() on  
> these arches.

Yes, most likely not setting wall_to_monotonic to -xtime at boot (see 
i386/kernel/time.c last few lines).
-- 
George Anzinger   george@mvista.com
High-res-timers:  http://sourceforge.net/projects/high-res-timers/
Preemption patch: http://www.kernel.org/pub/linux/kernel/people/rml