All of lore.kernel.org
 help / color / mirror / Atom feed
From: Jacopo Mondi <jacopo@jmondi.org>
To: linux-media@vger.kernel.org, libcamera-devel@lists.libcamera.org
Cc: Jacopo Mondi <jacopo@jmondi.org>,
	mchehab@kernel.org, hverkuil-cisco@xs4all.nl,
	sakari.ailus@linux.intel.com, andrey.konovalov@linaro.org,
	laurent.pinchart@ideasonboard.com
Subject: [v3 0/3] media: Register read-only sub-dev devnode
Date: Fri, 27 Mar 2020 23:35:19 +0100	[thread overview]
Message-ID: <20200327223522.506832-1-jacopo@jmondi.org> (raw)

Add new function v4l2_device_register_ro_subdev_nodes() to pair with
v4l2_device_register_subdev_nodes() that allows a bridge driver to register the
device node for its subdevices in read-only mode.

Devnode-centric (aka non-MC) bridge drivers control their subdevices through
direct calls to v4l2 subdev operations and do not want userspace to be able
to control the subdevice configuration by calling ioctls on the sub-device
devnode. For this reason, they usually refrain from registering any devnode at
all for their subdevices.

However it is sometimes required for userspace to access the sub-dev device
nodes to collect information on the actual configuration, without changing
the one currently applied to the device.

This requirement became pressing while working on libcamera on devnode-centric
platforms that do not expose any sub-device for their camera sensor to prevent
userspace from changing their configuration. To allow them to register device
node and being guaranteed to retain control of the subdevice configuration this
series proposes a way to register device nodes in read-only to restrict
access to all ioctls that could potentially affect the sub-dev configuration.

Thanks
   j

v1->v2:
- Documentation:
  - Add a new patch using Laurent's suggestion to update the sub-device
    userspace API introduction
  - Take in some of Laurent's suggestions in v4l2-subdev.rst and add a new
    section in dev-subdev.rst
- Implementation:
  - As noted by Andrey, V4L2_FL_* are meant to be used as bitmasks. Use
    test_bit()/set_bit() as the rest of the v4l2 core does. It's a bit an
    overkill compared to use plain BIT() as noted by Sakari but I preferred
    consistency with the rest of the core
  - Make v4l2_device_register_subdev_nodes() and
    v4l2_device_register_ro_subdev_nodes() to v4l2-device.h and make them
    inline functions. Documentation style has been copied from other functions
    with similar implementations, such as __video_register_device() in
    v4l2-dev.h

Jacopo Mondi (3):
  Documentation: media: Update sub-device API intro
  Documentation: media: Document read-only subdevice
  media: v4l2-dev: Add v4l2_device_register_ro_subdev_node()

 Documentation/media/kapi/v4l2-subdev.rst      | 53 ++++++++++++++++++-
 Documentation/media/uapi/v4l/dev-subdev.rst   |  5 ++
 .../media/uapi/v4l/vidioc-g-dv-timings.rst    |  6 +++
 Documentation/media/uapi/v4l/vidioc-g-std.rst |  6 +++
 .../media/uapi/v4l/vidioc-subdev-g-crop.rst   |  9 ++++
 .../media/uapi/v4l/vidioc-subdev-g-fmt.rst    |  8 +++
 .../v4l/vidioc-subdev-g-frame-interval.rst    |  8 +++
 .../uapi/v4l/vidioc-subdev-g-selection.rst    |  8 +++
 drivers/media/v4l2-core/v4l2-device.c         |  7 ++-
 drivers/media/v4l2-core/v4l2-subdev.c         | 19 +++++++
 include/media/v4l2-dev.h                      |  7 +++
 include/media/v4l2-device.h                   | 42 +++++++++++++--
 12 files changed, 170 insertions(+), 8 deletions(-)

--
2.25.1


             reply	other threads:[~2020-03-27 22:32 UTC|newest]

Thread overview: 17+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-03-27 22:35 Jacopo Mondi [this message]
2020-03-27 22:35 ` [v2 1/3] Documentation: media: Update sub-device API intro Jacopo Mondi
2020-03-27 22:35 ` [v2 2/3] Documentation: media: Document read-only subdevice Jacopo Mondi
2020-03-27 23:33   ` Sakari Ailus
2020-04-01 22:10     ` Jacopo Mondi
2020-04-06  7:00       ` Sakari Ailus
2020-04-01 11:19   ` Hans Verkuil
2020-04-01 11:46     ` Laurent Pinchart
2020-04-01 11:54       ` Hans Verkuil
2020-04-01 12:39         ` Jacopo Mondi
2020-04-04  1:35         ` Laurent Pinchart
2020-04-06 11:07           ` Jacopo Mondi
2020-04-06 12:39             ` Laurent Pinchart
2020-04-06 12:47               ` Jacopo Mondi
2020-03-27 22:35 ` [v2 3/3] media: v4l2-dev: Add v4l2_device_register_ro_subdev_node() Jacopo Mondi
2020-04-01 11:28   ` Hans Verkuil
2020-04-01 12:26     ` Jacopo Mondi

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=20200327223522.506832-1-jacopo@jmondi.org \
    --to=jacopo@jmondi.org \
    --cc=andrey.konovalov@linaro.org \
    --cc=hverkuil-cisco@xs4all.nl \
    --cc=laurent.pinchart@ideasonboard.com \
    --cc=libcamera-devel@lists.libcamera.org \
    --cc=linux-media@vger.kernel.org \
    --cc=mchehab@kernel.org \
    --cc=sakari.ailus@linux.intel.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.