Supporting Intel/AMD memory encryption

Posted Apr 26, 2018 16:05 UTC (Thu) by sjfriedl (✭ supporter ✭, #10111)
In reply to: Supporting Intel/AMD memory encryption by farnz
Parent article: Supporting Intel/AMD memory encryption

I completely agree with the assessments that we have a long way to go before this becomes a thing, and it probably won't ever be a thing, but those of us who are older remember enough times when $AMOUNT of memory was "enough for the forseeable future" that ended up being a bummer, that we have a knee-jerk chuckle about this one :-)

Supporting Intel/AMD memory encryption

Posted Apr 26, 2018 16:08 UTC (Thu) by farnz (subscriber, #17727) [Link] (2 responses)

There is a reason I said "move to 128-bit virtual addresses first", not "256 PiB will be enough for the foreseeable future" :)

Supporting Intel/AMD memory encryption

Posted Apr 26, 2018 16:57 UTC (Thu) by sjfriedl (✭ supporter ✭, #10111) [Link] (1 responses)

How useful will 128-bit address space be with 6 bits taken out of the middle? :-)

Supporting Intel/AMD memory encryption

Posted Apr 26, 2018 16:59 UTC (Thu) by farnz (subscriber, #17727) [Link]

These are physical addresses, not virtual, so absolutely fine - no more of an issue than the weirdness you find in the PC's address space at 640k to 1MiB.