On Fri, Jul 24, 2015 at 07:46:55AM +1000, Benjamin Herrenschmidt wrote: > On Thu, 2015-07-23 at 22:00 +0100, Peter Maydell wrote: > > (Also, none of our PCI device models actually try to do > > the "BAR at zero means I won't respond" behaviour, which > > presumably they might do in real life.) > > Devices aren't supposed to do that ... In fact, I don't see HW designers > adding a comparator if they don't have to :-) > > Cheers, > Ben. >
You never can tell with HW designers. For example, maybe they used this as a trick to avoid implementing memory/io enable properly. -- MST
