contification paper
   
    Henry Cejtin
     
    henry@sourcelight.com
       
    Thu, 8 Mar 2001 17:22:45 -0600
    
    
  
Would  that  (the  combination  of  getruage()  and  times()) explain all the
negatives?  If so I could believe it: one CPU is updating the times  in  some
order, so there is a race condition.
...
No,  I looked at the code in the kernel, and the getruasage() system call and
the times() system call read the same variables.
Can some one send me the code which fails?  Probably both binary  and  source
in case it is some version-itis thing.