PERFORCE change 51597 for review

[Marcel Moolenaar]

On Thu, Apr 22, 2004 at 11:33:50PM -0600, M. Warner Losh wrote:
> : ==== //depot/projects/gdb/sys/dev/pci/pci.c#7 (text+ko) ====
> : 
> : @@ -825,6 +825,12 @@
> :  	if (base == 0)
> :  		return 1;
> :  
> : +	/* Sun EBus bridges contain the ranges for the devices beyond them */
> : +	if ((pci_get_class(dev) == PCIC_BRIDGE) &&
> : +	    (pci_get_vendor(dev) == 0x108e) &&
> : +	    (pci_get_device(dev) == 0x1000 || pci_get_device(dev) == 0x1100))
> : +		return 1;
> : +
> :  	start = base;
> :  	end = base + (1 << ln2size) - 1;
> :  	count = 1 << ln2size;
> 
> Shouldn't the ebus bridge device be allocating the resources and
> giving them to their children?  This isn't the right fix and I'd like
> to work with you to the fix at the right level...

Yes. tmm@ is working on that. I want to see the stopgap fix committed
in the mean time to unbreak sparc64. I'm waiting to hear back from
tmm at . Don't worry. The above change is not a solution. It's a quick
fix to give us some breathing room.

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