Re: Explicit pointer provenance APIs for CHERI

On 2026-06-24 12:45 -04:00 EDT, "Jessica Clarke" <[email protected]> wrote:
> On 24 Jun 2026, at 17:34, Poul-Henning Kamp <[email protected]> wrote:
>>=20
>> --------
>> Alexander Ziaee writes:
>>=20
>>> Can we massage it into memory(7)? I think that'd be the most obvious pl=
ace.
>>> The draft is very focused on just memory statistics, but I think it wou=
ldn't hurt
>>> it to tweak that a little bit.
>>>=20
>>> https://reviews.freebsd.org/D18768
>>=20
>> As long as there is a cheri(7) manualpage to get people started=E2=80=A6
>=20
> Why? It=E2=80=99s not FreeBSD=E2=80=99s job to tell you everything about =
CHERI, only to
> document FreeBSD=E2=80=99s APIs and reference external documentation as
> appropriate. It=E2=80=99s not like we have an mmu(7) to teach you about h=
ow
> virtual addressing works, for example.

This is kind of what I was thinking. We don't have pages for processor exte=
nsions
or processors, we have arch(7) and a draft for memory(7). Personally, these=
 seem
to be what we're looking for here, but I'll defer to senior colleagues to t=
ake this
information and make the best decision.

Best,
Alex=