From siewerts@thinker.colorado.edu Wed Oct 17 17:26:41 2001
Received: (from siewerts@localhost)
	by thinker.colorado.edu (8.10.1/8.10.1/UnixOps+Hesiod) id f9HNOrQ17976;
	Wed, 17 Oct 2001 17:24:53 -0600 (MDT)
Message-Id: <200110172324.f9HNOrQ17976@thinker.colorado.edu>
Subject: Meet in Simlab Tonight!!
To: bf2man@earthlink.net (Hidekazu Miyoshi), chris_newman@maxtor.com,
   dan.vigil@seagate.com, darrin_laverenz@stortek.com, deanwschulze@home.com,
   dgardner@interfacecontrol.com, ehoskins@qwest.net, emiliod@linuxcrew.com,
   f.vang@gli-co.com, fsbh1@uaf.edu, gpayne@ball.com, gyounger@hwimfg.com,
   jeff_andre@stortek.com, jing_li@jdedwards.com, kansari@ball.com,
   ken.ordes@seagate.com, modhafarhussain@aol.com, noratheps@hotmail.com,
   sam.siewert@networkphotonics.com, siewerts@thinker.colorado.edu,
   shubharamani@yahoo.com, sombutsi@colorado.edu, sscranford@yahoo.com,
   sshekhar@qualcomm.com, sziesman@ball.com, tewahade@vlecom.com,
   vikram.shah@colorado.edu, wxie@qualcomm.com
Date: Wed, 17 Oct 2001 17:24:53 -0600 (MDT)
From: "Sam Siewert" <siewerts@thinker.colorado.edu>
X-Mailer: ELM [version 2.5 PL3]
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=us-ascii
Content-Length: 384
Status: RO

Meet in Simlab tonight!!

Sam

________________________________________________________________________
 Sam Siewert, Ph.D. -- Adjunct Faculty, Embedded Systems Cert. Program

 University of Colorado Boulder, Electrical and Computer Engineering

 siewerts@thinker.colorado.edu, www-sgc.colorado.edu/~siewerts/
________________________________________________________________________


From siewerts@thinker.colorado.edu Wed Oct 17 11:57:52 2001
Received: (from siewerts@localhost)
	by thinker.colorado.edu (8.10.1/8.10.1/UnixOps+Hesiod) id f9HHn0L00453;
	Wed, 17 Oct 2001 11:49:00 -0600 (MDT)
Message-Id: <200110171749.f9HHn0L00453@thinker.colorado.edu>
Subject: Example code helpful for HW
To: bf2man@earthlink.net (Hidekazu Miyoshi), chris_newman@maxtor.com,
   dan.vigil@seagate.com, darrin_laverenz@stortek.com, deanwschulze@home.com,
   dgardner@interfacecontrol.com, ehoskins@qwest.net, emiliod@linuxcrew.com,
   f.vang@gli-co.com, fsbh1@uaf.edu, gpayne@ball.com, gyounger@hwimfg.com,
   jeff_andre@stortek.com, jing_li@jdedwards.com, kansari@ball.com,
   ken.ordes@seagate.com, modhafarhussain@aol.com, noratheps@hotmail.com,
   sam.siewert@networkphotonics.com, siewerts@thinker.colorado.edu,
   shubharamani@yahoo.com, sombutsi@colorado.edu, sscranford@yahoo.com,
   sshekhar@qualcomm.com, sziesman@ball.com, tewahade@vlecom.com,
   vikram.shah@colorado.edu, wxie@qualcomm.com
Date: Wed, 17 Oct 2001 11:49:00 -0600 (MDT)
From: "Sam Siewert" <siewerts@thinker.colorado.edu>
X-Mailer: ELM [version 2.5 PL3]
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=us-ascii
Content-Length: 2266
Status: RO

Hi,

In case you haven't already looked at testtsh.c and time_stamp.c on
the class WWW page under the Solutions/Examples, please look at these to
get ideas on how to implement the taskSwitchHook call-back and how to
time-stamp events.

taskSwitchHookAdd, etc. are documented in the VxWorks manuals - please read
up on them!

You may time stamp events with ticks, but be aware that this is only
accurate to +/- 1 millisecond.

As an alternative, although not required, you may use code such as that found
in time_stamp.c to read the x86 timing oscillator base counter directly which
is typically good to +/- 1 microsecond given a period of about 0.4 microsecs.

Microsecond time-stamping is not a *requirement* for the HW, but does make
the HW much more interesting - so give it a try, but don't worry if you can't
get it to work as far as the grading of the HW.

The microsecond time-stamping calls are directly into 
/usr/local/wind/target/config/pcPentium BSP code and therefore are not 
documented in the generic VxWorks docs.

Be very careful with your taskSwitchHook call-back code - you can only 
have code in this function which is valid in kernel/interrupt context - i.e.
most C code is ok, but be careful about making API calls in this code - you
should note whether the VxWorks Ref. Man states that the API call is valid
in interrupt context - if it is valid in interrupt context (e.g. semGive),
then you can call such a function in the taskSwitchHook call-back - if not,
it will likely crash the kernel.

Likewise, if you write any "bad code" (e.g. bad pointer dereference) and it
is executed in the taskSwitchHook call-back, this will crash the kernel, so
you may want to test this code carefully before installing it.

Welcome to firmware/kernel code development!  It's not always easy - try and
write very well-formed code, be cognizant that you are executing as the
kernel, and you'll be ok.

Sam

________________________________________________________________________
 Sam Siewert, Ph.D. -- Adjunct Faculty, Embedded Systems Cert. Program

 University of Colorado Boulder, Electrical and Computer Engineering

 siewerts@thinker.colorado.edu, www-sgc.colorado.edu/~siewerts/
________________________________________________________________________