On Sun, 3 Dec 1995, Frank Wille wrote:

> Some days ago, I reported a Bus Error during configuration of
> grfcv0.
> 
> 
> Now, after two days of debugging, I made it to localize the reason
> for this crash:
> 
> The bus error is caused in cvattach(), while determining the size
> of the RAM on the CV64. I have a 2MB CV64, so that cv_has_4mb()
> writes some test patterns to non existent memory - Bus Error!

On at least one system with 2MB CV64 it was reported to work, that's why 
I integrated this routine. Perhaps I should follow Michael Hitch's 
suggestion for safely testing nonexistant memory ...

> After commenting out the cv_has_4mb() call and hardcoding a 2 MB
> RAM size, the configuration of grfcv0 works!

With the driver in it's current state you can't use more than 2MB anyway:) 

-- 
Michael Teske (teske@dice2.desy.de)