You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The current specification is ambiguous concerning the interaction of the SPMP with the Hypervisor extension.
The sentence describes that "there are extensions" that are being developed to support SPMP in a dual-stage model.
What are those "extensions"? The way I see it, it will be a single extension that will include part (or all) of the content that was removed recently in this commit: 185d2ee
Apart from polishing the jargon, I also suggest having a note with an updated roadmap status toward the formalization of such an extension (to provide clarity).
Btw, I have yet to see an open discussion on this model/profile where we may have virtualization without virtual memory (MMU) but rather based on a dual-stage sPMP. This is needed because many players don't understand the need for such a profile - this was pretty much corroborated by the recent comments provided by Qualcomm to the sPMP spec.
I will suggest Anup to add this topic to the agenda of the upcoming Hypervisor SIG meeting (25 Oct). I will also volunteer to prepare a few slides to drive the discussion and to share what is happening behind the scenes by some players and chip designers/manufacturers (to the extent I may disclose).
The text was updated successfully, but these errors were encountered:
The current specification is ambiguous concerning the interaction of the SPMP with the Hypervisor extension.
The sentence describes that "there are extensions" that are being developed to support SPMP in a dual-stage model.
What are those "extensions"? The way I see it, it will be a single extension that will include part (or all) of the content that was removed recently in this commit:
185d2ee
Apart from polishing the jargon, I also suggest having a note with an updated roadmap status toward the formalization of such an extension (to provide clarity).
Btw, I have yet to see an open discussion on this model/profile where we may have virtualization without virtual memory (MMU) but rather based on a dual-stage sPMP. This is needed because many players don't understand the need for such a profile - this was pretty much corroborated by the recent comments provided by Qualcomm to the sPMP spec.
I will suggest Anup to add this topic to the agenda of the upcoming Hypervisor SIG meeting (25 Oct). I will also volunteer to prepare a few slides to drive the discussion and to share what is happening behind the scenes by some players and chip designers/manufacturers (to the extent I may disclose).
The text was updated successfully, but these errors were encountered: