linux-origin
[Top] [All Lists]

BUG 792097 - 'top', 'vmstat', 'ps' cputime all say zero

To: hawkes@xxxxxxxxxxxxxxxxxxxx
Subject: BUG 792097 - 'top', 'vmstat', 'ps' cputime all say zero
From: pv@xxxxxxxxxxxxxxxxxxxxxx (hawkes@xxxxxxxxxxxx)
Date: Fri, 26 May 2000 15:50:28 -0700 (PDT)
Cc: linux-origin@xxxxxxxxxxx
Reply-to: sgi.bugs.linux-origin@xxxxxxxxxxxxxxxxx
Sender: owner-linux-origin@xxxxxxxxxxx
Webexec: webpvsubmit,PvProjectIncident
Webpv: www-proxy.engr.sgi.com
View Incident: 
http://co-op.engr.sgi.com/BugWorks/code/bwxquery.cgi?search=Search&wlong=1&view_type=Bug&wi=792097

Submitter : hawkes                    Submitter Domain : engr               
Assigned Engineer : hawkes            Assigned Domain : engr                
Assigned Group : linux-origin         Category : software                   
Customer Reported : F                 Priority : 2                          
Project : linux-origin                Status : open                         
Description :
Various commands that report cputime -- 'top', 'vmstat', 'ps' (e.g., ps -S) -- 
report zero elapsed time.

/proc/<pid>/cpu is correct.
/proc/<pid>/stat shows bogus values for user/system times.

The problem seems to be a mistyping of __kernel_clock_t in 
include/asm-mips64/posix_types.h.  The mips64 type is "int", though in all the 
other architectures it is "long".  Changing it to "long" seems to fix the 
problem (because proc_pid_stat() in fs/proc/array.c pushes these clock_t values 
through an sprintf() using a format of "%lu").

<Prev in Thread] Current Thread [Next in Thread>
  • BUG 792097 - 'top', 'vmstat', 'ps' cputime all say zero, hawkes@xxxxxxxxxxxx <=