cvs commit: src/sys/amd64/amd64 tsc.c src/sys/i386/i386 geode.c tsc.c src/sys/kern kern_clock.c kern_exit.c kern_proc.c kern_resource.c kern_synch.c kern_tc.c src/sys/sparc64/sparc64 tick.c src/sys/sys proc.h systm.h

[Julian Elischer]

Poul-Henning Kamp wrote:

>phk         2006-02-11 09:33:07 UTC
>
>  FreeBSD src repository
>
>  Modified files:
>    sys/amd64/amd64      tsc.c 
>    sys/i386/i386        geode.c tsc.c 
>    sys/kern             kern_clock.c kern_exit.c kern_proc.c 
>                         kern_resource.c kern_synch.c kern_tc.c 
>    sys/sparc64/sparc64  tick.c 
>    sys/sys              proc.h systm.h 
>  Log:
>  CPU time accounting speedup (step 2)
>  
>  Keep accounting time (in per-cpu) cputicks and the statistics counts
>  in the thread and summarize into struct proc when at context switch.
>  
>

exellent.  I have more to do in this line later. thread stats are still 
a bit of a mess.

>  
>  Don't reach across CPUs in calcru().
>  
>  Add code to calibrate the top speed of cpu_tickrate() for variable
>  cpu_tick hardware (like TSC on power managed machines).
>  
>  Don't enforce monotonicity (at least for now) in calcru.  While the
>  calibrated cpu_tickrate ramps up it may not be true.
>  
>  Use 27MHz counter on i386/Geode.
>  
>  Use TSC on amd64 & i386 if present.
>  
>  Use tick counter on sparc64
>  
>  Revision  Changes    Path
>  1.206     +1 -0      src/sys/amd64/amd64/tsc.c
>  1.7       +15 -0     src/sys/i386/i386/geode.c
>  1.205     +1 -0      src/sys/i386/i386/tsc.c
>  1.187     +3 -3      src/sys/kern/kern_clock.c
>  1.283     +3 -0      src/sys/kern/kern_exit.c
>  1.240     +1 -1      src/sys/kern/kern_proc.c
>  1.154     +45 -68    src/sys/kern/kern_resource.c
>  1.275     +6 -0      src/sys/kern/kern_synch.c
>  1.168     +133 -5    src/sys/kern/kern_tc.c
>  1.20      +1 -9      src/sys/sparc64/sparc64/tick.c
>  1.452     +3 -0      src/sys/sys/proc.h
>  1.240     +5 -2      src/sys/sys/systm.h
>  
>