Re: Guidance on emulating "sparse" address spaces

[Richard Henderson]

On 6/23/21 5:53 PM, Jason Thorpe wrote:
> Thanks for the pointer!  Yes, it seems like that might be similar to what I need... 
> create new address spaces (rather than using the "system" address space) for 
> the PCI memory and I/O regions (I need to do this for PCI configuration space, too, 
> FWIW...), then in my sparse ops handlers, decode the bus address and size, and then call 
> address_space_{read,write}().  Have I got the gist of it?

Yep.  That bitband device is a good example.


r~