PERFORCE change 64481 for review

[Marcel Moolenaar]

On Nov 10, 2004, at 1:18 AM, M. Warner Losh wrote:

> : 	Add resource allocation and release related code. With this,
> : 	uart(4) probes as a child of scc(4). Next, IRQ handling so
> : 	that we can do something other than wait for the interrupt
> : 	storm to pass over...
> :
> : +		rle->res = malloc(sizeof(struct resource), M_SCC,
> : +		    M_WAITOK | M_ZERO);
>
> We gotta find a way to put this in fewer drivers, not more...
> Encoding sizeof struct resource is why you need
> __RMAN_RESOURCE_VISIBLE, which generally shouldn't be defined.

Agreed. I don't like it either. My first thought was to add a wrapper
for the malloc and my second was that I don't want to get sidetracked
right now. I'll get sidetracked for the newbus issues (the bus-specific 
interrupt handlers and adding a mechanism for time-multiplexing) that
I still need to work out. I'll probably deal with the resource stuff
then. If I can defer the newbus issues, I'll fix the resource stuff
before I commit.

-- 
  Marcel Moolenaar         USPA: A-39004          marcel at xcllnt.net