kernel-doc nano-HOWTO
=====================

Many places in the source tree have extractable documentation in the
form of block comments above functions.  The components of this system
are:

- scripts/kernel-doc

  This is a perl script that hunts for the block comments and can mark
  them up directly into DocBook, man, text, and HTML. (No, not
  texinfo.)

- Documentation/DocBook/*.tmpl

  These are SGML template files, which are normal SGML files with
  special place-holders for where the extracted documentation should
  go.

- scripts/basic/docproc.c

  This is a program for converting SGML template files into SGML
  files. When a file is referenced it is searched for symbols
  exported (EXPORT_SYMBOL), to be able to distinguish between internal
  and external functions.
  It invokes kernel-doc, giving it the list of functions that
  are to be documented.
  Additionally it is used to scan the SGML template files to locate
  all the files referenced herein. This is used to generate dependency
  information as used by make.

- Makefile

  The targets 'sgmldocs', 'psdocs', 'pdfdocs', and 'htmldocs' are used
  to build DocBook files, PostScript files, PDF files, and html files
  in Documentation/DocBook.

- Documentation/DocBook/Makefile

  This is where C files are associated with SGML templates.


How to extract the documentation
--------------------------------

If you just want to read the ready-made books on the various
subsystems (see Documentation/DocBook/*.tmpl), just type 'make
psdocs', or 'make pdfdocs', or 'make htmldocs', depending on your
preference.  If you would rather read a different format, you can type
'make sgmldocs' and then use DocBook tools to convert
Documentation/DocBook/*.sgml to a format of your choice (for example,
'db2html ...' if 'make htmldocs' was not defined).

If you want to see man pages instead, you can do this:

$ cd linux
$ scripts/kernel-doc -man $(find -name '*.c') | split-man.pl /tmp/man
$ scripts/kernel-doc -man $(find -name '*.h') | split-man.pl /tmp/man

Here is split-man.pl:

-->
#!/usr/bin/perl

if ($#ARGV < 0) {
   die "where do I put the results?\n";
}

mkdir $ARGV[0],0777;
$state = 0;
while (<STDIN>) {
    if (/^\.TH \"[^\"]*\" 4 \"([^\"]*)\"/) {
	if ($state == 1) { close OUT }
	$state = 1;
	$fn = "$ARGV[0]/$1.4";
	print STDERR "Creating $fn\n";
	open OUT, ">$fn" or die "can't open $fn: $!\n";
	print OUT $_;
    } elsif ($state != 0) {
	print OUT $_;
    }
}

close OUT;
<--

If you just want to view the documentation for one function in one
file, you can do this:

$ scripts/kernel-doc -man -function fn file | nroff -man | less

or this:

$ scripts/kernel-doc -text -function fn file


How to add extractable documentation to your source files
---------------------------------------------------------

The format of the block comment is like this:

/**
 * function_name(:)? (- short description)?
(* @parameterx(space)*: (description of parameter x)?)*
(* a blank line)?
 * (Description:)? (Description of function)?
 * (section header: (section description)? )*
(*)?*/

The short function description ***cannot be multiline***, but the other
descriptions can be (and they can contain blank lines).  If you continue
that initial short description onto a second line, that second line will
appear further down at the beginning of the description section, which is
almost certainly not what you had in mind.

Avoid putting a spurious blank line after the function name, or else the
description will be repeated!

All descriptive text is further processed, scanning for the following special
patterns, which are highlighted appropriately.

'funcname()' - function
'$ENVVAR' - environment variable
'&struct_name' - name of a structure (up to two words including 'struct')
'@parameter' - name of a parameter
'%CONST' - name of a constant.

NOTE 1:  The multi-line descriptive text you provide does *not* recognize
line breaks, so if you try to format some text nicely, as in:

  Return codes
    0 - cool
    1 - invalid arg
    2 - out of memory

this will all run together and produce:

  Return codes 0 - cool 1 - invalid arg 2 - out of memory

NOTE 2:  If the descriptive text you provide has lines that begin with
some phrase followed by a colon, each of those phrases will be taken as
a new section heading, which means you should similarly try to avoid text
like:

  Return codes:
    0: cool
    1: invalid arg
    2: out of memory

every line of which would start a new section.  Again, probably not
what you were after.

Take a look around the source tree for examples.


kernel-doc for structs, unions, enums, and typedefs
---------------------------------------------------

Beside functions you can also write documentation for structs, unions,
enums and typedefs. Instead of the function name you must write the name
of the declaration;  the struct/union/enum/typedef must always precede
the name. Nesting of declarations is not supported.
Use the argument mechanism to document members or constants.

Inside a struct description, you can use the "private:" and "public:"
comment tags.  Structure fields that are inside a "private:" area
are not listed in the generated output documentation.

Example:

/**
 * struct my_struct - short description
 * @a: first member
 * @b: second member
 *
 * Longer description
 */
struct my_struct {
    int a;
    int b;
/* private: */
    int c;
};


How to make new SGML template files
-----------------------------------

SGML template files (*.tmpl) are like normal SGML files, except that
they can contain escape sequences where extracted documentation should
be inserted.

!E<filename> is replaced by the documentation, in <filename>, for
functions that are exported using EXPORT_SYMBOL: the function list is
collected from files listed in Documentation/DocBook/Makefile.

!I<filename> is replaced by the documentation for functions that are
_not_ exported using EXPORT_SYMBOL.

!D<filename> is used to name additional files to search for functions
exported using EXPORT_SYMBOL.

!F<filename> <function [functions...]> is replaced by the
documentation, in <filename>, for the functions listed.


Tim.
*/ <twaugh@redhat.com>