Commit Graph

9 Commits

Author SHA1 Message Date
Mauro Carvalho Chehab
d2a72e1f27 tools: docs: parse_data_structs.py: accept more reftypes
Current regex is limited to only some c-domain reftypes.
There are several others.

Change the code to pick the name specified there.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Message-ID: <5c146923d1e3183893f290216fb1378954e2e540.1759329363.git.mchehab+huawei@kernel.org>
2025-10-17 13:56:59 -06:00
Mauro Carvalho Chehab
7f809e6a6f tools: docs: parse_data_structs.py: add namespace support
C domain supports a ".. c:namespace::" tag that allows setting a
symbol namespace. This is used within the kernel to avoid warnings
about duplicated symbols. This is specially important for syscalls,
as each subsystem may have their own documentation for them.
This is specially true for ioctl.

When such tag is used, all C domain symbols have c++ style,
e.g. they'll become "{namespace}.{reference}".

Allow specifying C namespace at the exception files, avoiding
the need of override rules for every symbol.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Message-ID: <cc27ec60ceb3bdac4197fb7266d2df8155edacda.1759329363.git.mchehab+huawei@kernel.org>
2025-10-17 13:56:59 -06:00
Mauro Carvalho Chehab
2cdd27a708 tools: docs: parse_data_structs.py: get rid of process_exceptions()
Add an extra parameter to parse_file to make it handle exceptions
internally, cleaning up the API.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Message-ID: <8575bbc94ff706aa7e7cc3a188399ca17a3169e6.1759329363.git.mchehab+huawei@kernel.org>
2025-10-17 13:56:59 -06:00
Mauro Carvalho Chehab
9e4173432e tools: docs: parse_data_structs: make process_exceptions two stages
Split the logic which parses exceptions on two stages, preparing
the exceptions file to have rules that will affect xref generation.

For now, preserve the original API.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Message-ID: <da9ca5f2ce1ffcfb355e32e676ff013607c227e0.1759329363.git.mchehab+huawei@kernel.org>
2025-10-17 13:56:59 -06:00
Mauro Carvalho Chehab
3ed9521772 docs: kernel_include.py: fix line numbers for TOC
On TOC output, we need to embeed line numbers with ViewList.

Change the parse class to produce a line-number parsed result,
and adjust the output accordingly.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Message-ID: <74eed96e32f79eaaef7a99ffe7c3224fed369c27.1759329363.git.mchehab+huawei@kernel.org>
2025-10-17 13:56:59 -06:00
Mauro Carvalho Chehab
ba9fbb3d9a tools: docs: parse_data_structs.py: output a line number
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Message-ID: <dcffa6844dede00052f5fb851a857991468f22b5.1759329363.git.mchehab+huawei@kernel.org>
2025-10-17 13:56:59 -06:00
Mauro Carvalho Chehab
ba99100687 tools: docs: parse_data_structs.py: drop contents header
When used in practice, one may want to have multiple header
files on a single rst file, like:

	***********************
	Digital TV uAPI symbols
	***********************

	.. contents:: Table of Contents
	   :depth: 2
	   :local:

	Frontend
	========

	.. kernel-include:: include/uapi/linux/dvb/frontend.h
	    :generate-cross-refs:
	    :toc:

	Demux
	=====

	.. kernel-include:: include/uapi/linux/dvb/dmx.h
	    :generate-cross-refs:
	    :toc:

	...

So, don't add ..contents:: here.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Message-ID: <4bf353e5248133a3b0abd82519a38453402fe7c6.1759329363.git.mchehab+huawei@kernel.org>
2025-10-17 13:56:59 -06:00
Mauro Carvalho Chehab
242cfe3f77 tools: docs: parse_data_structs.py: add methods to return output
When running it from command line, we want to write an output
file, but when used as a class, one may just want the output
content returned as a string.

Split write_output() on two methods to allow both usecases.

Also add an extra method to produce a TOC.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Link: https://lore.kernel.org/r/c98bdec3380aad54178baf2751a2f1fcd128576b.1755872208.git.mchehab+huawei@kernel.org
2025-08-29 15:54:42 -06:00
Mauro Carvalho Chehab
cde494660f tools: docs: parse-headers.py: move it from sphinx dir
As suggested by Jon, we should start having a tools/docs
directory, instead of placing everything under scripts.

In the specific case of parse-headers.py, the previous
location is where we're placing Sphinx extensions, which is
not the right place for execs.

Move it to tools/docs/parse-headers.py.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Link: https://lore.kernel.org/r/0f5ac2d704cffe9834e589b39549d2393e1237ef.1755872208.git.mchehab+huawei@kernel.org
2025-08-29 15:54:42 -06:00