Timothy Miller wrote:
> Regarding real-time download, the primary purpose of this is to
> provide a human with a way to examine PCI activity.  A human cannot
> examine PCI activity in real-time  At least, I don't know anyone who
> can think that fast.
> 
> What we need to focus on first are the human factors.  Assuming we
> have managed to get a trace into the display computer (all of it or
> the right pieces on demand), what does a human need to see, and how
> can we best help him/her to see it and understand it?

1) we need to decode the dump. it would be good if the decoder could be
more intelligent and do also some error recovery and indicate the
mis-timed parts (possible from oversampled input). Simply: low level
decoding & validation

2a) the decoded information is now human-friendly, reading &
understanding is possible (contains addresses+data/wr/rd/bulk
transfers/int assertions..)

2b) write another "validator" to do machine checks of higher level stuff
(maybe in conjunction with an OGP simulator/sw model) .. for debugging
bios/drivers/etc..

In short for the start: replace the correct waveforms with a symbolic
notation (rd/wr) and indicate incorrect ones.
_______________________________________________
Open-graphics mailing list
[email protected]
http://lists.duskglow.com/mailman/listinfo/open-graphics
List service provided by Duskglow Consulting, LLC (www.duskglow.com)

Reply via email to