On 7/11/20 12:30 PM, Richard Henderson wrote: >> The old implementation returns true for >> HOST_LONG_BITS <= TARGET_VIRT_ADDR_SPACE_BITS >> (because there's a different definition of guest_addr_valid() there) >> but this one does a range check even in that case. > > It's part and parcel with patch 1, wherein we are in fact attempting to limit > the guest address space to GUEST_ADDR_MAX. > > That's why I put patch 1 first, so the behaviour change happens there.
Ho hum. I've just realized the messages are sorted oddly in the mbox here, and that the behaviour change is actually coming later in patch 7. So, to summarize, I am intending a change here, it's just a matter of sorting things so that one thing happens at a time. r~
