Re: On I/O and meory address space

[jdahlin@precisionimages.com]

On Mon, 10 Feb 1997, Ali Najafi <alinajafi@simei.aztech.com.sg> wrote:
<text deleted>
>
>1- How difficult really is it to debug the device driver of a PCI device
>mapped anywhere on a 32 bit memory address space?
>
>A. Najafi
>

Our system can be mapped anywhere in the 32-bit memory space, but there are 
times that I want to have it mapped below the 1Meg boundary for testing and 
debugging (sometimes I work in DOS mode with batch files).  

What we have done is to set the base address of our board to a location below 
the 1Meg boundary after the system has booted (our software people provided a 
couple of programs, CPEEK and CPOKE, that peek and poke to the PCI config 
space).  You run the risk of picking a "bad" address that won't work because 
someone else is using it, like in the days before plug-n-play.  We use address 
0xD0000, which I believe is the address for a monochrome display.   Since we 
don't have a monochrome display in our test system we know it is probably an 
unused address.  

Maybe you can use a similar scheme during the debug phase of you device driver 
developement, where your driver sets the value of the base address to something 
below the 1Meg boundary so the software tools you have can be used.

Of course, this is done for TESTING ONLY!  It cannot be done in a product.

Jeff Dahlin
jdahlin@precisionimages.com

======  ___ \==  __ \==   /== Precision Digital Images
=====  /    /=  /   /=   /=== http://www.precisionimages.com
====  _____/=  /   /=   /==== 8520 154th Avenue NE
===  /======  /   /=   /===== Redmond, WA  98053  U S A
==__/======______/=___/====== Tel (206)882-0218   Fax 867-9177
E X C E L L E N C E    I N    A D V A N C E D    I M A G I N G

		Please visit our Home Page
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
v��