Hello Kate
This is a great idea, maybe we can ask Anas to join one of our upcoming calls?
-- Artem
On Wed, 2020-07-08 at 18:23 -0500, Kate Stewart wrote:
On Wed, 8 Jul 2020, Kate Stewart wrote:
> On Wed, Jul 8, 2020 at 12:47 PM Nathalie Chan King Choy <nathalie@xxxxxxxxxx> wrote:
...
> o Stefano S: Email Kate to see if she knows how Linux kernel has solved reverse linking problem for kernel certification.
>
> Not sure which reverse linking is being discussed... V model to requirements, or linking of evidence to the standards clauses? Can
> you please clarify?
Hi Kate,
We were discussing how to link the documents to relevant code (e.g.
header files under xen.git) and the code to the documents under the
new separate repository.
At the moment we are working on documents for backware-compatible
interfaces that rarely change so it is a relatively easy problem to
solve because we don't expect many changes on the xen.git side.
However, if we were working on documents for interfaces that change more
often things would be different.
Eventually we came up with the idea of introducing a special header in
the document under the new repository with a "link" to the original
code, but we have no reverse link (no link from the code to the doc, but
again in these cases the code is not expected to change.)
See for instance:
.. #########################################################
.. version: $Id$
.. src: xen/include/public/vcpu.h@RELEASE-4.13.0
.. #########################################################
https://gitlab.com/xen-project/fusa/-/blob/master/hypercalls/runstate.rst
Linux kernel isn't pursuing this direction (at least so far), but the Zephyr project is.
Anas has been working on this and is scheduled to present at an upcoming
TSC, so as soon as his details are ready to present, it may be worthwhile
arranging a call to discuss, and see if the projects can leverage each other's
work.
Thoughts?
Kate
|