Experimental Physics and
Industrial Control System

Andrew Johnson <[email protected]> · Tue, 08 Jul 1997 10:43:15 +0100

Ralph,

The BSP in the mv167 has to set a tick-timer divider, the correct
setting of which depends on the clock speed of the board.  Once the
divider is set, the vxWorks system clock  tick rate can be set properly,
but with the wrong divider setting sysClockRateSet doesn't get the right
tick rate.  Whether the same divider setting is needed for the mv162 I
don't know, but the technique I explained is how the 167 BSP retains
knowledge of the clock rate.  It sounds like the mv162 BSP uses a
different method to store the CPU speed.  It's also possible that the
mv162 can determine its own speed by looking at hardware registers.  I
don't think earlier versions of the mv167 could do that (and I'm not
sure whether later versions can or not) which is why they put in the
test they did.

The '0xee' is a magic number which ensures a random number is unlikely
to be thought of as a clock speed, '0x21' is 33 decimal, being the clock
rate in MHz.

The simple answer is - different BSP, different method.

-- Andrew
   ___
 .'   `.	Andrew Johnson, Head of Electronics
/ Royal )	Royal Greenwich Observatory
\ Greenwich	Madingley Road, Cambridge, CB3 0EZ
| Observatory	Tel: +44 1223 374823   Fax: 374700
+----------	WWW: http://www.ast.cam.ac.uk/~anj

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System