Em Tue, 10 Jun 2025 14:07:24 -0700
Jakub Kicinski [off-list ref] escreveu:

On Tue, 10 Jun 2025 22:59:11 +0200 Mauro Carvalho Chehab wrote:

quoted

quoted

The question is, are we OK with the templates that need to be created
for netlink specs?!     

If there's no other way, one might have a tool for maintainers to use
to update templates, but yeah, having one template per each yaml
is not ideal. I think we need to investigate it better and seek for
some alternatives to avoid it.  

FWIW we have tools/net/ynl/ynl-regen.sh, it regenerates the C code 
we have committed in the tree (uAPI headers mostly).
We could add it there. Which is not to distract from your main
point that not having the templates would be ideal.

With the new Sphinx extension for netlink specs I posted:

	https://lore.kernel.org/linux-doc/cover.1749723671.git.mchehab+huawei@kernel.org/T/#t (local)
	https://lore.kernel.org/linux-doc/20250612142438.MED5SEN3C-3RDQI5I1ELC-u8QJEjH8W4vUQRBdyK1tI@z/T/#t

There's no need for a template for each file, although it does require
updating Documentation/netlink/specs/index.rst. There are a couple
of reasons:

	1. on my tests, I got some errors auto-generating it while
	   using:
		 make SPHINXDIRS="networking netlink" htmldocs

	2. a dynamically-generated file will cause a extra
	   warnings at the userspace files that contain the name of the
           netlink spec index.html. Basically, kernel build runs a script
	   which validates that all files under Documentation/ actually
	   exist

	3. adding/renaming files typically require changing
	   MAINTAINERS and/or Makefiles. Updating index.rst
	   accordingly is already expected for documentation.

In any case, as I didn't drop the existing script, you could add a
call inside tools/net/ynl/ynl-regen.sh to:

	tools/net/ynl/pyynl/ynl_gen_rst.py -x  -v -o Documentation/netlink/specs/index.rst

To ensure that nobody would forget updating it.

Regards,
Mauro