Re: comments on PMP enhancements
toggle quoted message Show quoted text
It seems like the current trade-off in supported use cases is due to the combination of two constraints: backward compatibility and avoiding use of a Reserved PMP bit. A year ago the exact degree of this trade-off may not have been obvious, but at this point adopting this new spec would then just lead to development of a PMPv3 spec that addresses the unaddressed use cases and uses a Reserved bit - thus still ending up architecturally at the place that the current spec tries to avoid. (And tbd how easily that spec could maintain full backward compatibility with this spec.)
The key question is whether it is better to evolve/expand the current spec into what it ultimately wants or needs to be, or to follow a two-step approach (that architecturally, at least to me, seems messy)? If we're only putting off using a Reserved bit for half a year or a year, let's just cleanly move to that architectural solution directly.
On Wed, Feb 12, 2020, 4:42 AM Joe Xie <joxie@...> wrote: