On Wed, Jul 20, 2022 at 10:29 AM Robert Chyla <Robert.Chyla@...> wrote:

Only now (after your explanations ...) I can understand what Sm1-12 and Ss1-12 listed here:

    https://wiki.riscv.org/display/HOME/Recently+Ratified+Extensions

I believe the use of hyphens in the names was an early and incorrect (i.e. disallowed) usage.  The proper and official names are as I mentioned in my last email.  Please ask the tpm's to correct this wiki page.

Greg

 

as follows:

mean. Navigation of 'extension landscape' is hard (and I am not new to RISC-V ...).

>These are a subset of the contents of the Priv 1.11 and 1.12 documents.

IMO it may be really hard for anyone to 'extract' what 'subset of the contents' Sm1-12 is exactly. Especially that other extensions in same PDF ('Svnapot' for example ...) have dedicated chapters.

Maybe it is time to mandate that each ratified extension has separated PDF (or at least separated chapter). Otherwise it may be really, really hard (for implementer and verifier ...) to know what particular extension is exactly. IMO saying 'Sm1-12' is 'subset of the contents' is not precise enough ...

/Robert

On 7/20/2022 10:13 AM, Greg Favor wrote:

On Tue, Jul 19, 2022 at 7:32 AM Allen Baum <allen.baum@...> wrote:

Note that some of the version numbers (specifically priv spec 1.11 and 1.12) have names, and therefore can be represented by binary choices

I just want to note ( to make sure there isn't any confusion) that "Priv 1.11" and "Priv 1.12" are NOT extensions.  They effectively refer to documents that contain a collection of extensions (a superset of the following extensions):

The actual extensions that people are trying to refer to when they say "Priv 1.11" and "Priv 1.12", and their names are:  Sm1p11, Sm1p12, Ss1p11, and Ss1p12.  These are a subset of the contents of the Priv 1.11 and 1.12 documents.

For any ratified extension one shouldn't need to refer to a dated release of the document that contains the extension - since ratified extensions are frozen and any new "version" of an extension must be a separate new extension.  Going forward, version numbers on an extension should be meaningless.

But, to be more precise, the text of a ratified spec is allowed to change only to the extent of incorporating "typo" corrections and clarifications.  To that extent it is useful to look at the latest release of the document containing a ratified extension so as to see the latest typo fixes and clarifications.

Greg