On 04/23/2017 08:39 PM, John Hubbard wrote: > Actually, MEMORY_DEVICE_PRIVATE / _PUBLIC seems like a good choice to > me, because the memory may not remain CPU-unaddressable in the future. > By that, I mean that I know of at least one company (ours) that is > working on products that will support hardware-based memory coherence > (and access counters to go along with that). If someone were to enable > HMM on such a system, then the device memory would be, in fact, directly > addressable by a CPU--thus exactly contradicting the "unaddressable" name.
I'm expecting similar with CCIX-like coherently attached accelerators running within FPGAs and as discrete devices as well. Everyone and their dog is working on hardware based coherence as a programming convenience and so the notion of ZONE_DEVICE as it stood is going to rapidly evolve over the next 18 months, maybe less. Short term anyway. Jon.