From: Jakub Kicinski <kuba@kernel.org>
To: netdev@vger.kernel.org, davem@davemloft.net, edumazet@google.com,
pabeni@redhat.com
Cc: sdf@google.com, jacob.e.keller@intel.com, vadfed@fb.com,
johannes@sipsolutions.net, jiri@resnulli.us, dsahern@kernel.org,
stephen@networkplumber.org, fw@strlen.de,
linux-doc@vger.kernel.org, Jakub Kicinski <kuba@kernel.org>
Subject: [RFC net-next 1/4] ynl: add intro docs for the concept
Date: Wed, 10 Aug 2022 19:23:01 -0700 [thread overview]
Message-ID: <20220811022304.583300-2-kuba@kernel.org> (raw)
In-Reply-To: <20220811022304.583300-1-kuba@kernel.org>
Short overview of the sections. I presume most people will start
by copy'n'pasting existing schemas rather than studying the docs,
but FWIW...
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
---
Documentation/index.rst | 1 +
Documentation/netlink/index.rst | 13 +++
Documentation/netlink/netlink-bindings.rst | 104 +++++++++++++++++++++
3 files changed, 118 insertions(+)
create mode 100644 Documentation/netlink/index.rst
create mode 100644 Documentation/netlink/netlink-bindings.rst
diff --git a/Documentation/index.rst b/Documentation/index.rst
index 67036a05b771..130e39c18fe0 100644
--- a/Documentation/index.rst
+++ b/Documentation/index.rst
@@ -112,6 +112,7 @@ needed).
infiniband/index
leds/index
netlabel/index
+ netlink/index
networking/index
pcmcia/index
power/index
diff --git a/Documentation/netlink/index.rst b/Documentation/netlink/index.rst
new file mode 100644
index 000000000000..a7f063f31ff3
--- /dev/null
+++ b/Documentation/netlink/index.rst
@@ -0,0 +1,13 @@
+.. SPDX-License-Identifier: BSD-3-Clause
+
+====================
+Netlink API Handbook
+====================
+
+Netlink documentation.
+
+.. toctree::
+ :maxdepth: 2
+
+ netlink-bindings
+
diff --git a/Documentation/netlink/netlink-bindings.rst b/Documentation/netlink/netlink-bindings.rst
new file mode 100644
index 000000000000..af0c069001f3
--- /dev/null
+++ b/Documentation/netlink/netlink-bindings.rst
@@ -0,0 +1,104 @@
+.. SPDX-License-Identifier: BSD-3-Clause
+
+Netlink protocol specifications
+===============================
+
+Netlink protocol specifications are complete, machine readable descriptions of
+genetlink protocols written in YAML. The schema (in ``jsonschema``) can be found
+in the same directory as this documentation file.
+
+Schema structure
+----------------
+
+YAML schema has the following conceptual sections. Most properties in the schema
+accept (or in fact require) a ``description`` sub-property documenting the defined
+object.
+
+Globals
+~~~~~~~
+
+There is a handful of global attributes such as the family name, version of
+the protocol, and additional C headers (used only for uAPI and C-compatible
+codegen).
+
+Global level also contains a handful of customization properties, like
+``attr-cnt-suffix`` which allow accommodating quirks of existing families.
+Those properties should not be used in new families.
+
+Attribute Spaces
+~~~~~~~~~~~~~~~~
+
+First of the main two sections is ``attribute-spaces``. This property contains
+information about netlink attributes of the family. All families have at least
+one attribute space, most have multiple. ``attribute-spaces`` is an array/list,
+with each entry describing a single space. The ``name`` of the space is not used
+in uAPI/C codegen, it's internal to the spec itself, used by operations and nested
+attributes to refer to a space.
+
+Each attribute space has properties used to render uAPI header enums. ``name-prefix``
+is prepended to the name of each attribute, allowing the attribute names to be shorter
+compared to the enum names in the uAPI.
+Optionally attribute space may contain ``enum-name`` if the uAPI header's enum should
+have a name. Most netlink uAPIs do not name attribute enums, the attribute names are
+heavily prefixed, which is sufficient.
+
+Most importantly each attribute space contains a list of attributes under the ``attributes``
+property. The properties of an attribute should look fairly familiar to anyone who ever
+wrote netlink code (``name``, ``type``, optional validation constraints like ``len`` and
+reference to the internal space for nests).
+
+Note that attribute spaces do not themselves nest, nested attributes refer to their internal
+space via a ``nested-attributes`` property, so the YAML spec does not resemble the format
+of the netlink messages directly.
+
+YAML spec may also contain fractional spaces - spaces which contain a ``subspace-of``
+property. Such spaces describe a section of a full space, allowing narrowing down which
+attributes are allowed in a nest or refining the validation criteria. Fractional spaces
+can only be used in nests. They are not rendered to the uAPI in any fashion.
+
+Operations and notifications
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+This section describes messages passed between the kernel and the user space.
+There are three types of entries in this section - operations, notifications
+and events.
+
+Notifications and events both refer to the asynchronous messages sent by the kernel
+to members of a multicast group. The difference between the two is that a notification
+shares its contents with a GET operation (the name of the GET operation is specified
+in the ``notify`` property). This arrangement is commonly used for notifications about
+objects where the notification carries the full object definition.
+
+Events are more focused and carry only a subset of information rather than full
+object state (a made up example would be a link state change event with just
+the interface name and the new link state).
+Events are considered less idiomatic for netlink and notifications
+should be preferred. After all, if the information in an event is sufficiently
+complete to be useful, it should also be useful enough to have a corresponding
+GET command.
+
+Operations describe the most common request - response communication. User
+sends a request and kernel replies. Each operation may contain any combination
+of the two modes familiar to netlink users - ``do`` and ``dump``.
+``do`` and ``dump`` in turn contain a combination of ``request`` and ``response``
+properties. If no explicit message with attributes is passed in a given
+direction (e.g. a ``dump`` which doesn't not accept filter, or a ``do``
+of a SET operation to which the kernel responds with just the netlink error code)
+``request`` or ``response`` section can be skipped. ``request`` and ``response``
+sections list the attributes allowed in a message. The list contains only
+the names of attributes from a space referred to by the ``attribute-space``
+property.
+
+An astute reader will notice that there are two ways of defining sub-spaces.
+A full fractional space with a ``subspace-of`` property and a de facto subspace
+created by list attributes for an operation. This is only for convenience.
+The abilities to refine the selection of attributes and change their definition
+afforded by the fractional space result in much more verbose YAML, and the full
+definition of a space (i.e. containing all attributes) is always required to render
+the uAPI header, anyway. So the per-operation attribute selection is a form of
+a shorthand.
+
+Multicast groups
+~~~~~~~~~~~~~~~~
+
+This section lists the multicast groups of the family, not much to be said.
--
2.37.1
next prev parent reply other threads:[~2022-08-11 2:23 UTC|newest]
Thread overview: 44+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-08-11 2:23 [RFC net-next 0/4] ynl: YAML netlink protocol descriptions Jakub Kicinski
2022-08-11 2:23 ` Jakub Kicinski [this message]
2022-08-11 20:17 ` [RFC net-next 1/4] ynl: add intro docs for the concept Edward Cree
2022-08-12 22:23 ` Jakub Kicinski
2022-08-15 20:09 ` Johannes Berg
2022-08-16 0:32 ` Jakub Kicinski
2022-08-16 7:07 ` Johannes Berg
2022-08-11 2:23 ` [RFC net-next 2/4] ynl: add the schema for the schemas Jakub Kicinski
2022-08-15 20:03 ` Johannes Berg
2022-08-15 20:09 ` Johannes Berg
2022-08-16 0:47 ` Jakub Kicinski
2022-08-16 7:21 ` Johannes Berg
2022-08-16 15:53 ` Jakub Kicinski
2022-08-16 19:30 ` Johannes Berg
2022-09-26 16:10 ` Rob Herring
2022-09-27 21:56 ` Jakub Kicinski
2022-09-28 12:32 ` Rob Herring
2022-08-11 2:23 ` [RFC net-next 3/4] ynl: add a sample python library Jakub Kicinski
2022-08-11 5:48 ` Benjamin Poirier
2022-08-11 15:50 ` Jakub Kicinski
2022-08-11 20:09 ` Stephen Hemminger
2022-08-12 22:53 ` Jakub Kicinski
2022-08-15 20:00 ` Johannes Berg
2022-08-12 1:04 ` Stephen Hemminger
2022-08-12 15:42 ` Edward Cree
2022-08-12 23:07 ` Jakub Kicinski
2022-08-18 21:26 ` Keller, Jacob E
2022-08-11 2:23 ` [RFC net-next 4/4] ynl: add a sample user for ethtool Jakub Kicinski
2022-08-11 16:18 ` sdf
2022-08-11 19:35 ` Jakub Kicinski
2022-08-11 22:55 ` Stanislav Fomichev
2022-08-11 23:31 ` Jakub Kicinski
2022-08-12 16:26 ` Stanislav Fomichev
2022-08-12 22:48 ` Jakub Kicinski
2022-08-14 12:27 ` Ido Schimmel
2022-08-11 4:15 ` [RFC net-next 0/4] ynl: YAML netlink protocol descriptions Stephen Hemminger
2022-08-11 4:47 ` Jakub Kicinski
2022-08-11 15:01 ` Stephen Hemminger
2022-08-11 15:34 ` Jakub Kicinski
2022-08-11 16:28 ` sdf
2022-08-11 19:42 ` Jakub Kicinski
2022-08-12 17:00 ` Florian Fainelli
2022-08-12 22:26 ` Jakub Kicinski
2022-08-19 19:56 ` Jakub Kicinski
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20220811022304.583300-2-kuba@kernel.org \
--to=kuba@kernel.org \
--cc=davem@davemloft.net \
--cc=dsahern@kernel.org \
--cc=edumazet@google.com \
--cc=fw@strlen.de \
--cc=jacob.e.keller@intel.com \
--cc=jiri@resnulli.us \
--cc=johannes@sipsolutions.net \
--cc=linux-doc@vger.kernel.org \
--cc=netdev@vger.kernel.org \
--cc=pabeni@redhat.com \
--cc=sdf@google.com \
--cc=stephen@networkplumber.org \
--cc=vadfed@fb.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).