media: imx: silence a couple debug messages

Convert to dev_dbg the "subdev bound" and IPU-internal media-link
creation messages.

Signed-off-by: Steve Longerbeam <slongerbeam@gmail.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: imx: Create missing links from CSI-2 receiver

The entities external to the i.MX6 IPU and i.MX7 now create the links
to their fwnode-endpoint connected entities in their notifier bound
callbacks. Which means imx_media_create_of_links() and
imx_media_create_csi_of_links() are no longer needed and are removed.

However there is still one case in which imx-media needs to create
fwnode-endpoint based links at probe completion. The v4l2-async framework
does not allow multiple subdevice notifiers to contain a duplicate
subdevice in their asd_list. Only the first subdev notifier that discovers
and adds that one subdevice to its asd_list will receive a bound callback
for it. Other subdevices that also have firmware endpoint connections to
this duplicate subdevice will not have it in their asd_list, and thus will
never receive a bound callback for it. In the case of imx-media, the one
duplicate subdevice in question is the i.MX6 MIPI CSI-2 receiver.

Until there is a solution to that problem, rewrite imx_media_create_links()
to add the missing links from the CSI-2 receiver to the CSIs and CSI muxes.
The function is renamed imx_media_create_csi2_links().

Signed-off-by: Steve Longerbeam <slongerbeam@gmail.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: imx: csi: Lookup upstream endpoint with imx_media_get_pad_fwnode

Fix the 1:1 port-id:pad-index assumption for the upstream subdevice, by
searching the upstream subdevice's endpoints for one that maps to the
pad's index. This is carried out by a new reverse mapping function
imx_media_get_pad_fwnode().

Signed-off-by: Steve Longerbeam <slongerbeam@gmail.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: imx5/6/7: csi: Mark a bound video mux as a CSI mux

For i.MX5/6, if the bound subdev is a video mux, it must be one of the
CSI muxes, and for i.MX7, the bound subdev must always be a CSI mux.

So if the bound subdev is a video mux, mark it as a CSI mux with a new
group id IMX_MEDIA_GRP_ID_CSI_MUX.

In the process use the new group id in csi_get_upstream_endpoint(), and
do some cleanup in that function for better readability.

Suggested-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Steve Longerbeam <slongerbeam@gmail.com>
Acked-by: Rui Miguel Silva <rmfrfs@gmail.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: imx7: csi: Remove imx7_csi_get_upstream_endpoint()

The function imx7_csi_get_upstream_endpoint() is not necessary for
imx7. First, the imx7 CSI only receives from the CSI mux, so much of
the code in there is pointless. Second, it is only used to determine
whether the CSI mux has selected the CSI-2 input or the parallel input.
This can be accomplished much more simply by getting the function type
of selected input entity to the CSI mux.

Signed-off-by: Steve Longerbeam <slongerbeam@gmail.com>
Acked-by: Rui Miguel Silva <rmfrfs@gmail.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: imx: csi: Create media links in bound notifier

Implement a notifier bound op to register media links from the remote
sub-device's source pad(s) to the CSI sink pad.

Signed-off-by: Steve Longerbeam <slongerbeam@gmail.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: imx7: csi: Create media links in bound notifier

Implement a notifier bound op to register media links from the remote
sub-device's source pad(s) to the CSI sink pad.

Signed-off-by: Steve Longerbeam <slongerbeam@gmail.com>
Reviewed-by: Rui Miguel Silva <rmfrfs@gmail.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: imx7: mipi csis: Create media links in bound notifier

Implement a notifier bound op to register media links from the remote
sub-device's source pad(s) to the mipi csi-2 receiver sink pad.

Signed-off-by: Steve Longerbeam <slongerbeam@gmail.com>
Acked-by: Rui Miguel Silva <rmfrfs@gmail.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: imx: mipi csi-2: Create media links in bound notifier

Implement a notifier bound op to register media links from the remote
sub-device's source pad(s) to the mipi csi-2 receiver sink pad.

Signed-off-by: Steve Longerbeam <slongerbeam@gmail.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: video-mux: Create media links in bound notifier

Implement a notifier bound op to register media links from the remote
sub-device's source pad(s) to the video-mux sink pad(s).

Signed-off-by: Steve Longerbeam <slongerbeam@gmail.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: imx: imx7-media-csi: Implement get_fwnode_pad op

Use v4l2_subdev_get_fwnode_pad_1_to_1() as the get_fwnode_pad operation.
The i.MX7 CSI maps port numbers and pad indexes 1:1.

Acked-by: Rui Miguel Silva <rmfrfs@gmail.com>
Signed-off-by: Steve Longerbeam <slongerbeam@gmail.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: imx: imx7-mipi-csis: Implement get_fwnode_pad op

Use v4l2_subdev_get_fwnode_pad_1_to_1() as the get_fwnode_pad operation.
The imx7-mipi-csis maps port numbers and pad indexes 1:1.

Signed-off-by: Steve Longerbeam <slongerbeam@gmail.com>
Acked-by: Rui Miguel Silva <rmfrfs@gmail.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: imx: mipi csi-2: Implement get_fwnode_pad op

Use v4l2_subdev_get_fwnode_pad_1_to_1() as the get_fwnode_pad operation.
The MIPI CSI-2 receiver maps port numbers and pad indexes 1:1.

Signed-off-by: Steve Longerbeam <slongerbeam@gmail.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: video-mux: Implement get_fwnode_pad op

Use v4l2_subdev_get_fwnode_pad_1_to_1() as the get_fwnode_pad operation.
The video mux maps fwnode port numbers and pad indexes 1:1.

Signed-off-by: Steve Longerbeam <slongerbeam@gmail.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: imx: csi: Implement get_fwnode_pad op

The CSI does not have a 1:1 relationship between fwnode port numbers and
pad indexes. In fact the CSI fwnode device is itself a port which is the
sink, containing only a single fwnode endpoint. Implement media_entity
operation get_fwnode_pad to first verify the given endpoint is the CSI's
sink endpoint, and if so return the CSI sink pad index.

Signed-off-by: Steve Longerbeam <slongerbeam@gmail.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: Revert "media: v4l2-fwnode: Add a convenience function for registering subdevs with notifiers"

The users of v4l2_async_register_fwnode_subdev() have switched to
parsing their endpoints and setting up async sub-device lists in their
notifiers locally, without using the endpoint parsing callbacks. There
are no more users of v4l2_async_register_fwnode_subdev() so this
convenience function can be removed.

This reverts commit 1634f0eded87d1f150e823fa56cd782ea0775eb2.

Signed-off-by: Steve Longerbeam <slongerbeam@gmail.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: imx: Parse information from firmware without using callbacks

Instead of using the convenience functions
v4l2_async_notifier_parse_fwnode_endpoints*() or
v4l2_async_register_fwnode_subdev(), parse the input endpoints
and set up the async sub-devices without using callbacks. The drivers
know which ports it must parse and how to handle unconnected remotes,
so it makes the code simpler to transfer control of endpoint parsing
to the driver.

Signed-off-by: Steve Longerbeam <slongerbeam@gmail.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: video-mux: Parse information from firmware without using callbacks

Instead of using the convenience function
v4l2_async_register_fwnode_subdev(), parse the video-mux input endpoints
and set up the async sub-devices without using callbacks. The video-mux
knows which ports it must parse (the input ports) and how to handle
unconnected remotes, so it makes the code simpler to transfer control
of endpoint parsing to the driver.

Signed-off-by: Steve Longerbeam <slongerbeam@gmail.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: v4l2-mc: add v4l2_create_fwnode_links helpers

Add functions to create media links between source and sink subdevices,
based on the fwnode endpoint connections between them:

v4l2_create_fwnode_links_to_pad() - create links from a source subdev to
                                    a single sink pad based on fwnode
                                    endpoint connections.

v4l2_create_fwnode_links() - create all links from a source to sink subdev
                             based on fwnode endpoint connections.

These functions can be used in a sink's v4l2-async notifier subdev
bound callback to make the links from the bound subdev.

Signed-off-by: Steve Longerbeam <slongerbeam@gmail.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: v4l2-subdev: add v4l2_subdev_get_fwnode_pad_1_to_1

Add a convenience function that can be used as the .get_fwnode_pad
operation for subdevices that map port numbers and pad indexes 1:1.
The function verifies the endpoint is owned by the subdevice, and if
so returns the endpoint port number.

Signed-off-by: Steve Longerbeam <slongerbeam@gmail.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: entity: Pass entity to get_fwnode_pad operation

Add a missing pointer to the entity in the media_entity operation
get_fwnode_pad. There are no implementers of this op yet, but a future
entity that does so will almost certainly need a reference to itself
to carry out the work.

operation")

Fixes: ae45cd5efc120 ("[media] media: entity: Add get_fwnode_pad entity
Signed-off-by: Steve Longerbeam <slongerbeam@gmail.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: i2c: imx219: Fix a bug in imx219_enum_frame_size

When enumerating the frame sizes, the value sent to
imx219_get_format_code should be fse->code
(the code from the ioctl) and not imx219->fmt.code
which is the code set currently in the driver.

Fixes: 22da1d56e982 ("media: i2c: imx219: Add support for RAW8 bit bayer format")
Signed-off-by: Dafna Hirschfeld <dafna.hirschfeld@collabora.com>
Reviewed-by: Helen Koike <helen.koike@collabora.com>
Reviewed-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
Reviewed-by: Lad Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: i2c: imx219: Implement get_selection

Implement the get_selection pad operation for the IMX219 sensor driver.
The supported targets report the sensor's native size, the crop default
rectangle and the crop rectangle.

Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: media: sh_veu: Remove driver

Since its inclusion in v3.9, no users of the SuperH VEU mem2mem video
processing driver have appeared upstream.  All VEU devices in SuperH
board code still bind to the "uio_pdrv_genirq" driver instead.
The original author marked the driver orphaned in v3.15.

Remove the driver; it can always be resurrected from git history when
needed.

Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Acked-by: Rob Landley <rob@landley.net>
Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: dvb-frontends: remove redundant initialization of variable status

The variable status is being initialized with a value that is never read
and it is being updated later with a new value.  The initialization is
redundant and can be removed.

Addresses-Coverity: ("Unused value")

Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: mantis_dvb: remove redundant initialization to variable result

The variable result is being initialized with a value that is never read
and it is being updated later with a new value.  The initialization is
redundant and can be removed.

Addresses-Coverity: ("Unused value")

Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: usb: ttusb-dec: reduce the number of memory reads in ttusb_dec_handle_irq()

In ttusb_dec_handle_irq(), buffer[4] is continuously read from memory
three times, without being modified.
To reduce the number of memory reads, buffer[4] is first assigned to a
local variable index, and then index is used to replace buffer[4].

Signed-off-by: Jia-Ju Bai <baijiaju1990@gmail.com>
Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: stv0900_core: remove redundant assignment to variable val

The variable val is being initializeed with a value that is never read
and it is being updated later with a new value. The initialization
is redundant and can be removed.

Addresses-Coverity: ("Unused value")

Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: dvb: remove redundant assignment to variable bw

The variable bw is being initialized with a value that is
never read and it is being updated later with a new value.  The
initialization is redundant and can be removed.

Addresses-Coverity: ("Unused value")

Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: gpio-ir-tx: allow transmission without carrier

Some IR protocols do not use a carrier.

Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: gpio-ir-tx: improve precision of transmitted signal due to scheduling

usleep_range() may take longer than the max argument due to scheduling,
especially under load. This is causing random errors in the transmitted
IR. Remove the usleep_range() in favour of busy-looping with udelay().

Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: v4l: document VIDIOC_SUBDEV_QUERYCAP

Add documentation for the new VIDIOC_SUBDEV_QUERYCAP ioctl.

Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: v4l2-subdev: add VIDIOC_SUBDEV_QUERYCAP ioctl

While normal video/radio/vbi/swradio nodes have a proper QUERYCAP ioctl
that apps can call to determine that it is indeed a V4L2 device, there
is currently no equivalent for v4l-subdev nodes. Adding this ioctl will
solve that, and it will allow utilities like v4l2-compliance to be used
with these devices as well.

SUBDEV_QUERYCAP currently returns the version and capabilities of the
subdevice. Define a capability flag to report if the subdevice is
registered in read-only mode.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: v4l2-subdev: Guard whole fops and ioctl hdlr

A sub-device device node can be registered in user space only if the
CONFIG_VIDEO_V4L2_SUBDEV_API Kconfig option is selected. Currently the
open/close file operations and the ioctl handler have some parts of
their implementations guarded by #if
defined(CONFIG_VIDEO_V4L2_SUBDEV_API), while they are actually not
accessible without a video device node registered to user space.

Guard the whole open, close and ioctl handler and provide stubs if the
VIDEO_V4L2_SUBDEV_API Kconfig option is not selected.

This slightly reduces the kernel size when the option is not selected
and simplifies the file ops and ioctl implementations.

Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: v4l2-dev: Add v4l2_device_register_ro_subdev_node()

Add to the V4L2 core a function to register device nodes for video
subdevices in read-only mode.

Registering a device node in read-only mode is useful to expose to
userspace the current sub-device configuration, without allowing
application to change it by using the V4L2 subdevice ioctls.

Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: Documentation: media: Document read-only subdevice

Document a new kAPI function to register subdev device nodes in read only
mode and for each affected ioctl report how access is restricted.

Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: Documentation: media: Update sub-device API intro

Update the V4L2 sub-device userspace API introduction to provide more
details on why complex devices might want to register devnodes for the
connected subdevices.

Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Suggested-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: ipu3.rst: fix a build warning

Sphinx 2.4.4 produces this warning:

Documentation/admin-guide/media/ipu3.rst:235: WARNING: Title underline too short.

Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: vidioc-enum-fmt.rst: make the ENUM_FMT text clearer

Rework the documentation to make it easier for the reader to understand
the differences in behavior of this ioctl between MC and non-MC drivers.

Note the addition of the 'video-node-centric' and 'MC-centric' terms to
help understand what the IO_MC capability really means.

Also mention in the beginning that mbus_code is one of the fields that
application should initialize, and add META_OUTPUT as one of the types that
this ioctl supports (that was never added here when the META_OUTPUT buffer
type was added).

Finally document that EINVAL will be returned if mbus_code is unsupported.

Fixes: e5b6b07a1b45 ("media: v4l2: Extend VIDIOC_ENUM_FMT to support MC-centric devices")
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: vimc: Make use of V4L2_CAP_IO_MC

Set the V4L2_CAP_IO_MC capability flag to report this vimc
inputs/outputs are controlled by the media graph.

Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: staging/intel-ipu3: Make use of V4L2_CAP_IO_MC

Set the V4L2_CAP_IO_MC capability flag and remove the driver specific
vidioc_enum_{input,output}, vidioc_g_{input,output} and
vidioc_s_{input,output} callbacks.

Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: rcar-vin: Make use of V4L2_CAP_IO_MC

Set the V4L2_CAP_IO_MC capability flag and remove the driver specific
vidioc_enum_input, vidioc_g_input and vidioc_s_input callbacks for the
media controller enabled part of the driver. Also add support mbus_code
filtering for format enumeration.

Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: v4l2: Extend VIDIOC_ENUM_FMT to support MC-centric devices

The VIDIOC_ENUM_FMT ioctl enumerates all formats supported by a video
node. For MC-centric devices, its behaviour has always been ill-defined,
with drivers implementing one of the following behaviours:

- No support for VIDIOC_ENUM_FMT at all
- Enumerating all formats supported by the video node, regardless of the
  configuration of the pipeline
- Enumerating formats supported by the video node for the active
  configuration of the connected subdevice

The first behaviour is obviously useless for applications. The second
behaviour provides the most information, but doesn't offer a way to find
what formats are compatible with a given pipeline configuration. The
third behaviour fixes that, but with the drawback that applications
can't enumerate all supported formats anymore, and have to modify the
active configuration of the pipeline to enumerate formats.

The situation is messy as none of the implemented behaviours are ideal,
and userspace can't predict what will happen as the behaviour is
driver-specific.

To fix this, let's extend the VIDIOC_ENUM_FMT with a missing capability:
enumerating pixel formats for a given media bus code. The media bus code
is passed through the v4l2_fmtdesc structure in a new mbus_code field
(repurposed from the reserved fields). With this capability in place,
applications can enumerate pixel formats for a given media bus code
without modifying the active configuration of the device.

The current behaviour of the ioctl is preserved when the new mbus_code
field is set to 0, ensuring compatibility with existing userspace. The
API extension is documented as mandatory for MC-centric devices (as
advertised through the V4L2_CAP_IO_MC capability), allowing applications
and compliance tools to easily determine the availability of the
VIDIOC_ENUM_FMT extension.

Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: pci: Fill v4l2_fmtdesc with designated initializers

Replace initialization of static const v4l2_fmtdesc instances that
specify every struct member with designated initializers. This allows
not zeroing the reserved fields explicitly, and will avoid a need to
patch these drivers every time a reserved field is repurposed.

Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: v4l2-dev/ioctl: Add V4L2_CAP_IO_MC

Add a video device capability flag to indicate that its inputs and/or
outputs are controlled by the Media Controller instead of the V4L2 API.
When this flag is set, ioctl for enum inputs and outputs are
automatically enabled and programmed to call a helper function.

Suggested-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Helen Koike <helen.koike@collabora.com>
Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: admin-guide: split cardlist.rst file

After adding all cardlists, this file became too big. Split
it on smaller files, in order to make easier to maintain.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: admin-guide: make clear about PCI subsystem IDs

On several parts of the document, it mentions "PCI ID", when
it is actually referring to the subsystem's part of the PCI
ID.

Change the language to let it be clear.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: cec: silence shift wrapping warning in __cec_s_log_addrs()

The log_addrs->log_addr_type[i] value is a u8 which is controlled by
the user and comes from the ioctl.  If it's over 31 then that results in
undefined behavior (shift wrapping) and that leads to a Smatch static
checker warning.  We already cap the value later so we can silence the
warning just by re-ordering the existing checks.

I think the UBSan checker will also catch this bug at runtime and
generate a warning.  But otherwise the bug is harmless.

Fixes: 9881fe0ca187 ("[media] cec: add HDMI CEC framework (adapter)")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: sun8i: Fix an error handling path in 'deinterlace_runtime_resume()'

It is spurious to call 'clk_disable_unprepare()' when
'clk_prepare_enable()' has not been called yet.
Re-order the error handling path to avoid it.

Fixes: a4260ea49547 ("media: sun4i: Add H3 deinterlace driver")
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Acked-by: Chen-Yu Tsai <wens@csie.org>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
[hverkuil-cisco@xs4all.nl: err_exlusive_rate -> err_exclusive_rate]
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: vimc: deb: Add support for {RGB,BGR,GBR}888 bus formats on source pad

Add support for RGB888_*, BGR888_* and GBR888_* media bus formats on
the source pad of debayer subdevices.

Acked-by: Helen Koike <helen.koike@collabora.com>
Co-developed-by: Vitor Massaru Iha <vitor@massaru.org>
Signed-off-by: Vitor Massaru Iha <vitor@massaru.org>
Signed-off-by: Nícolas F. R. A. Prado <nfraprado@protonmail.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: vimc: Add missing {RGB,BGR,GBR}888 media bus codes

Add missing RGB888_*, BGR888_* and GBR888_* media bus codes in the
vimc_pix_map_list. Since there is no GBR24 pixelformat, use the RGB24
pixelformat for MEDIA_BUS_FMT_GBR888_1X24.

Acked-by: Helen Koike <helen.koike@collabora.com>
Co-developed-by: Vitor Massaru Iha <vitor@massaru.org>
Signed-off-by: Vitor Massaru Iha <vitor@massaru.org>
Signed-off-by: Nícolas F. R. A. Prado <nfraprado@protonmail.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: vimc: Support multiple media bus codes for each pixelformat

Change vimc_pix_map_list to allow multiple media bus codes to map to the
same pixelformat, making it possible to add media bus codes for which
there are no pixelformat.

Acked-by: Helen Koike <helen.koike@collabora.com>
Signed-off-by: Nícolas F. R. A. Prado <nfraprado@protonmail.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
[hverkuil-cisco@xs4all.nl: fix sparse warning: const u32 -> u32]
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: ti-vpe: avoid gcc-9 warning

gcc warns about empty array declarations, which we get in this driver
when compile-testing without CONFIG_OF:

drivers/media/platform/ti-vpe/cal.c:2194:34: warning: array 'cal_of_match' assumed to have one element
 2194 | static const struct of_device_id cal_of_match[];

Since all users of this driver do need CONFIG_OF anyway, there is no
point in making the array definition conditional to save space, so
just remove the #ifdef and move the array up a little.

Fixes: 343e89a792a5 ("[media] media: ti-vpe: Add CAL v4l2 camera capture driver")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Benoit Parrot <bparrot@ti.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: staging: media: usbvision: usbvision-core Correct spelling

Correct spelling in comment

Signed-off-by: John Oldman <john.oldman@polehill.co.uk>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: staging: media: imx: no need to check return value of debugfs_create functions

When calling debugfs functions, there is no need to ever check the
return value.  The function can work or not, but the code logic should
never do something different based on this.

Cc: Steve Longerbeam <slongerbeam@gmail.com>
Cc: Philipp Zabel <p.zabel@pengutronix.de>
Cc: Shawn Guo <shawnguo@kernel.org>
Cc: Sascha Hauer <s.hauer@pengutronix.de>
Cc: Pengutronix Kernel Team <kernel@pengutronix.de>
Cc: Fabio Estevam <festevam@gmail.com>
Cc: NXP Linux Team <linux-imx@nxp.com>
Cc: devel@driverdev.osuosl.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Reviewed-by: Rui Miguel Silva <rmfrfs@gmail.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: MAINTAINERS: update the Amlogic VDEC driver maintainer entry

Add myself as co-maintainer of the Amlogic VDEC driver, and add the
missing vdec DT yaml bindings.

Signed-off-by: Neil Armstrong <narmstrong@baylibre.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: meson: vdec: enable mcrcc for VP9

The motion compensation reference cache controller allows caching
parts of reference frames for faster decoding.

Fixes: 00c43088aa68 ("media: meson: vdec: add VP9 decoder support")
Signed-off-by: Maxime Jourdan <mjourdan@baylibre.com>
Signed-off-by: Neil Armstrong <narmstrong@baylibre.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: cx231xx: Remove unneeded semicolon

Fixes coccicheck warnings:

drivers/media/usb/cx231xx/cx231xx-avcore.c:610:3-4: Unneeded semicolon

Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Zou Wei <zou_wei@huawei.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: rcar-csi2: Fix comment of VCDT/VCDT2 register

According to latest H/W manual v1.50, the description of channel
number in the VCDT/VCDT2 register is decremented by one.
Therefore, this patch fixes it about comment.

Signed-off-by: Koji Matsuoka <koji.matsuoka.xm@renesas.com>
Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: imx: utils: Default colorspace to SRGB

The function imx_media_init_mbus_fmt() initializes the imx subdevice
mbus colorimetry to some sane defaults when the subdevice is registered.
Currently it guesses at a colorspace based on the passed mbus pixel
format. If the format is RGB, it chooses colorspace V4L2_COLORSPACE_SRGB,
and if the format is YUV, it chooses V4L2_COLORSPACE_SMPTE170M.

While that might be a good guess, it's not necessarily true that a RGB
pixel format encoding uses a SRGB colorspace, or that a YUV encoding
uses a SMPTE170M colorspace. Instead of making this dubious guess,
just default the colorspace to SRGB.

Reported-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Steve Longerbeam <slongerbeam@gmail.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: dt-bindings: media: rockchip-vpu: fix interrupt-names

A test with the command below gives for example this error:

arch/arm64/boot/dts/rockchip/rk3328-evb.dt.yaml: video-codec@ff350000:
interrupts: [[0, 9, 4]] is too short
arch/arm64/boot/dts/rockchip/rk3328-evb.dt.yaml: video-codec@ff350000:
interrupt-names: ['vdpu'] is too short
arch/arm64/boot/dts/rockchip/rk3328-evb.dt.yaml: video-codec@ff350000:
interrupt-names:0: 'vepu' was expected

With the conversion of rockchip-vpu.txt to yaml the correct
'interrupt-names' for rk3328 was not included, so add them now.
Also add 'minItems' to 'interrupts'for the completeness.

make ARCH=arm64 dtbs_check
DT_SCHEMA_FILES=Documentation/devicetree/bindings/media/rockchip-vpu.yaml

Signed-off-by: Johan Jonker <jbx6244@gmail.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: staging: rkisp1: cap: enable RGB capture format with YUV media bus

In selfpath, RGB capture formats are received in the sink pad as YUV
and are converted to RGB only when writing to memory. So the validation
function should accept YUV bus formats with RGB capture encoding.

Signed-off-by: Dafna Hirschfeld <dafna.hirschfeld@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: staging: rkisp1: rsz: change (hv)div only if capture format is YUV

RGB formats in selfpath should receive input format as YUV422.
The resizer input format is always YUV422 and therefore
if the capture format is RGB, the resizer should not change
the YUV rations.

Signed-off-by: Dafna Hirschfeld <dafna.hirschfeld@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: staging: rkisp1: rsz: remove redundant if statement and add inline doc

The statement "if (rsz->fmt_type == V4L2_PIXEL_ENC_YUV)"
can be removed since the value of rsz->fmt_type is either
V4L2_PIXEL_ENC_YUV or V4L2_PIXEL_ENC_BAYER and the function
returns if it is bayer. In addition some doc with clarification
is added.

Signed-off-by: Dafna Hirschfeld <dafna.hirschfeld@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: staging: rkisp1: rsz: get the capture format info from the capture struct

Currently the format info of the capture is retrieved by calling
the function v4l2_format_info. This is not needed since it is
already saved in the capture object.

Signed-off-by: Dafna Hirschfeld <dafna.hirschfeld@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: staging: ipu3-imgu: Add a sanity check for the parameter struct size

There have been cases where seemingly innocuous patches have broken the
uAPI by changing the memory layout of the parameter struct. Generally such
changes also introduce a change in the size of the entire struct. This
patch adds a sanity check to avoid such cases happening in the future.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Tested-by: Tested-by: Bingbu Cao <bingbu.cao@intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: staging: ipu3-imgu: Move alignment attribute to field

Move the alignment attribute of struct ipu3_uapi_awb_fr_config_s to the
field in struct ipu3_uapi_4a_config, the other location where the struct
is used.

Fixes: commit c9d52c114a9f ("media: staging: imgu: Address a compiler warning on alignment")
Reported-by: Tomasz Figa <tfiga@chromium.org>
Tested-by: Bingbu Cao <bingbu.cao@intel.com>
Cc: stable@vger.kernel.org # for v5.3 and up
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: Revert "staging: imgu: Address a compiler warning on alignment"

This reverts commit c9d52c114a9fcc61c30512c7f810247a9f2812af.

The patch being reverted changed the memory layout of struct
ipu3_uapi_acc_param. Revert it, and address the compiler warning issues in
further patches.

Fixes: commit c9d52c114a9f ("media: staging: imgu: Address a compiler warning on alignment")
Reported-by: Tomasz Figa <tfiga@chromium.org>
Tested-by: Bingbu Cao <bingbu.cao@intel.com>
Cc: stable@vger.kernel.org # for v5.3 and up
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: MAINTAINERS: ipu3-imgu: Add Tian Shu and Bingbu as reviewers

Add Bingbu Cao and Tian Shu Qiu as reviewers for the IPU3 ImgU driver.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Cc: Bingbu Cao <bingbu.cao@intel.com>
Cc: Tian Shu Qiu <tian.shu.qiu@intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: ipu3.rst: add explanation for running mode

Add some explanation of the ImgU running mode and add more information
about firmware selection and running mode usage.

Signed-off-by: Bingbu Cao <bingbu.cao@intel.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: staging: ipu3: Fix stale list entries on parameter queue failure

When queuing parameters fails, current code bails out without deleting
the corresponding vb2 buffer from the driver buffer list, but the buffer
is returned to vb2. This leads to stale list entries and a crash when
the driver stops streaming:

[  224.935561] ipu3-imgu 0000:00:05.0: set parameters failed.
[  224.998932] ipu3-imgu 0000:00:05.0: set parameters failed.
[  225.064430] ipu3-imgu 0000:00:05.0: set parameters failed.
[  225.128534] ipu3-imgu 0000:00:05.0: set parameters failed.
[  225.194945] ipu3-imgu 0000:00:05.0: set parameters failed.
[  225.360363] ------------[ cut here ]------------
[  225.360372] WARNING: CPU: 0 PID: 6704 at
drivers/media/common/videobuf2/videobuf2-core.c:927
vb2_buffer_done+0x20f/0x21a [videobuf2_common]
[  225.360374] Modules linked in: snd_seq_dummy snd_seq snd_seq_device
veth bridge stp llc tun nf_nat_tftp nf_conntrack_tftp nf_nat_ftp
nf_conntrack_ftp esp6 ah6 ip6t_REJECT ip6t_ipv6header cmac rfcomm uinput
ipu3_imgu(C) ipu3_cio2 iova videobuf2_v4l2 videobuf2_common
videobuf2_dma_sg videobuf2_memops ov13858 ov5670 v4l2_fwnode dw9714
acpi_als xt_MASQUERADE fuse iio_trig_sysfs cros_ec_sensors_ring
cros_ec_light_prox cros_ec_sensors cros_ec_sensors_core
industrialio_triggered_buffer kfifo_buf industrialio
cros_ec_sensorsupport cdc_ether btusb btrtl btintel btbcm usbnet
bluetooth ecdh_generic ecc hid_google_hammer iwlmvm iwl7000_mac80211
r8152 mii lzo_rle lzo_compress iwlwifi zram cfg80211 joydev
[  225.360400] CPU: 0 PID: 6704 Comm: CameraDeviceOps Tainted: G
C        5.4.30 #5
[  225.360402] Hardware name: HP Soraka/Soraka, BIOS
Google_Soraka.10431.106.0 12/03/2019
[  225.360405] RIP: 0010:vb2_buffer_done+0x20f/0x21a [videobuf2_common]
[  225.360408] Code: 5e 41 5f 5d e9 e0 16 5a d4 41 8b 55 08 48 c7 c7 8f
8b 5c c0 48 c7 c6 36 9a 5c c0 44 89 f9 31 c0 e8 a5 1c 5b d4 e9 53 fe ff
ff <0f> 0b eb a3 e8 12 d7 43 d4 eb 97 0f 1f 44 00 00 55 48 89 e5 41 56
[  225.360410] RSP: 0018:ffff9468ab32fba8 EFLAGS: 00010297
[  225.360412] RAX: ffff8aa7a51577a8 RBX: dead000000000122 RCX:
ffff8aa7a51577a8
[  225.360414] RDX: 0000000000000000 RSI: 0000000000000006 RDI:
ffff8aa7a5157400
[  225.360416] RBP: ffff9468ab32fbd8 R08: ffff8aa64e47e600 R09:
0000000000000000
[  225.360418] R10: 0000000000000000 R11: ffffffffc06036e6 R12:
dead000000000100
[  225.360420] R13: ffff8aa7820f1940 R14: ffff8aa7a51577a8 R15:
0000000000000006
[  225.360422] FS:  00007c1146ffd700(0000) GS:ffff8aa7baa00000(0000)
knlGS:0000000000000000
[  225.360424] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[  225.360426] CR2: 00007aea3473a000 CR3: 00000000537d6004 CR4:
00000000003606f0
[  225.360427] Call Trace:
[  225.360434]  imgu_return_all_buffers+0x6f/0x8e [ipu3_imgu]
[  225.360438]  imgu_vb2_stop_streaming+0xd6/0xf0 [ipu3_imgu]
[  225.360441]  __vb2_queue_cancel+0x33/0x22d [videobuf2_common]
[  225.360443]  vb2_core_streamoff+0x16/0x78 [videobuf2_common]
[  225.360448]  __video_do_ioctl+0x33d/0x42a
[  225.360452]  video_usercopy+0x34a/0x615
[  225.360455]  ? video_ioctl2+0x16/0x16
[  225.360458]  v4l2_ioctl+0x46/0x53
[  225.360462]  do_vfs_ioctl+0x50a/0x787
[  225.360465]  ksys_ioctl+0x58/0x83
[  225.360468]  __x64_sys_ioctl+0x1a/0x1e
[  225.360470]  do_syscall_64+0x54/0x68
[  225.360474]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
[  225.360476] RIP: 0033:0x7c118030f497
[  225.360479] Code: 8a 66 90 48 8b 05 d1 d9 2b 00 64 c7 00 26 00 00 00
48 c7 c0 ff ff ff ff c3 66 2e 0f 1f 84 00 00 00 00 00 b8 10 00 00 00 0f
05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d a1 d9 2b 00 f7 d8 64 89 01 48
[  225.360480] RSP: 002b:00007c1146ffa5a8 EFLAGS: 00000246 ORIG_RAX:
0000000000000010
[  225.360483] RAX: ffffffffffffffda RBX: 00007c1140010018 RCX:
00007c118030f497
[  225.360484] RDX: 00007c114001019c RSI: 0000000040045613 RDI:
000000000000004c
[  225.360486] RBP: 00007c1146ffa700 R08: 00007c1140010048 R09:
0000000000000000
[  225.360488] R10: 0000000000000000 R11: 0000000000000246 R12:
00007c11400101b0
[  225.360489] R13: 00007c1140010200 R14: 00007c1140010048 R15:
0000000000000001
[  225.360492] ---[ end trace 73625ecfbd1c930e ]---
[  225.360498] general protection fault: 0000 [#1] PREEMPT SMP PTI
[  225.360501] CPU: 0 PID: 6704 Comm: CameraDeviceOps Tainted: G
WC        5.4.30 #5
[  225.360502] Hardware name: HP Soraka/Soraka, BIOS
Google_Soraka.10431.106.0 12/03/2019
[  225.360505] RIP: 0010:imgu_return_all_buffers+0x52/0x8e [ipu3_imgu]
[  225.360507] Code: d4 49 8b 85 70 0a 00 00 49 81 c5 70 0a 00 00 49 39
c5 74 3b 49 bc 00 01 00 00 00 00 ad de 49 8d 5c 24 22 4c 8b 30 48 8b 48
08 <49> 89 4e 08 4c 89 31 4c 89 20 48 89 58 08 48 8d b8 58 fc ff ff 44
[  225.360509] RSP: 0018:ffff9468ab32fbe8 EFLAGS: 00010293
[  225.360511] RAX: ffff8aa7a51577a8 RBX: dead000000000122 RCX:
dead000000000122
[  225.360512] RDX: 0000000000000000 RSI: 0000000000000006 RDI:
ffff8aa7a5157400
[  225.360514] RBP: ffff9468ab32fc18 R08: ffff8aa64e47e600 R09:
0000000000000000
[  225.360515] R10: 0000000000000000 R11: ffffffffc06036e6 R12:
dead000000000100
[  225.360517] R13: ffff8aa7820f1940 R14: dead000000000100 R15:
0000000000000006
[  225.360519] FS:  00007c1146ffd700(0000) GS:ffff8aa7baa00000(0000)
knlGS:0000000000000000
[  225.360521] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[  225.360523] CR2: 00007aea3473a000 CR3: 00000000537d6004 CR4:
00000000003606f0
[  225.360525] Call Trace:
[  225.360528]  imgu_vb2_stop_streaming+0xd6/0xf0 [ipu3_imgu]
[  225.360531]  __vb2_queue_cancel+0x33/0x22d [videobuf2_common]
[  225.360534]  vb2_core_streamoff+0x16/0x78 [videobuf2_common]
[  225.360537]  __video_do_ioctl+0x33d/0x42a
[  225.360540]  video_usercopy+0x34a/0x615
[  225.360542]  ? video_ioctl2+0x16/0x16
[  225.360546]  v4l2_ioctl+0x46/0x53
[  225.360548]  do_vfs_ioctl+0x50a/0x787
[  225.360551]  ksys_ioctl+0x58/0x83
[  225.360554]  __x64_sys_ioctl+0x1a/0x1e
[  225.360556]  do_syscall_64+0x54/0x68
[  225.360559]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
[  225.360561] RIP: 0033:0x7c118030f497
[  225.360563] Code: 8a 66 90 48 8b 05 d1 d9 2b 00 64 c7 00 26 00 00 00
48 c7 c0 ff ff ff ff c3 66 2e 0f 1f 84 00 00 00 00 00 b8 10 00 00 00 0f
05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d a1 d9 2b 00 f7 d8 64 89 01 48
[  225.360565] RSP: 002b:00007c1146ffa5a8 EFLAGS: 00000246 ORIG_RAX:
0000000000000010
[  225.360567] RAX: ffffffffffffffda RBX: 00007c1140010018 RCX:
00007c118030f497
[  225.360569] RDX: 00007c114001019c RSI: 0000000040045613 RDI:
000000000000004c
[  225.360570] RBP: 00007c1146ffa700 R08: 00007c1140010048 R09:
0000000000000000
[  225.360572] R10: 0000000000000000 R11: 0000000000000246 R12:
00007c11400101b0
[  225.360574] R13: 00007c1140010200 R14: 00007c1140010048 R15:
0000000000000001
[  225.360576] Modules linked in: snd_seq_dummy snd_seq snd_seq_device
veth bridge stp llc tun nf_nat_tftp nf_conntrack_tftp nf_nat_ftp
nf_conntrack_ftp esp6 ah6 ip6t_REJECT ip6t_ipv6header cmac rfcomm uinput
ipu3_imgu(C) ipu3_cio2 iova videobuf2_v4l2 videobuf2_common
videobuf2_dma_sg videobuf2_memops ov13858 ov567

Fix this by moving the list_del() call just below the list_first_entry()
call when the buffer no longer needs to be in the list.

Fixes: 8ecc7c9da013 ("media: staging/intel-ipu3: parameter buffer refactoring")
Signed-off-by: Tomasz Figa <tfiga@chromium.org>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Bingbu Cao <bingbu.cao@intel.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: venus: venc,vdec: Return EBUSY on S_FMT while streaming

According to the v4l spec s_fmt must return EBUSY while the
particular queue is streaming. Add such check in encoder and
decoder s_fmt methods.

Signed-off-by: Stanimir Varbanov <stanimir.varbanov@linaro.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: venus: Mark last capture buffer

According to stateful Codec API the decoder will process all
remaining buffers from before the source change event in
dynamic-resolution-change state and mark the last buffer with
V4L2_BUF_FLAG_LAST.

In Venus case the firmware doesn't mark that last buffer and
some mechanism have to be created in v4l decoder driver.
Fortunately the firmware interface (HFI) claims that the
decoder output buffers will be returned to v4l decoder
driver before it send the insufficient event.

In order to do that we save last queued in the driver capture
buffer in the event_notify and issue flush on output firmware
buffers queue. Once the saved buffer is returned (as a result of
flush command) we mark it as LAST. For all that possible we
extend HFI flush command with one more argument and one more
flush_done HFI driver callback.

Signed-off-by: Stanimir Varbanov <stanimir.varbanov@linaro.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: venus: vdec: Init registered list unconditionally

Presently the list initialization is done only in
dynamic-resolution-change state, which leads to list corruptions
and use-after-free. Init list_head unconditionally in
vdec_stop_capture called by vb2 stop_streaming without takeing
into account current codec state.

Signed-off-by: Stanimir Varbanov <stanimir.varbanov@linaro.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: venus: vdec: Mark flushed buffers with error state

Once the hfi_session_flush is issued by the vdec all queued
buffers to firmware should be returned to the v4l driver. Some
of those buffers are not processed at the time of flush command,
those buffers has filled len zero (no data). Catch that in
buffer_done callback and mark not filled capture buffers with
error state so that client can discard them.

Signed-off-by: Stanimir Varbanov <stanimir.varbanov@linaro.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: venus: helpers: Done buffers per queue type

Currently calling venus_helper_buffers_done() will return buffers to
user for both capture and output queues in the same call. This is
wrong because both queues are really separate and calling
stop_streaming on one queue shouldn't return buffers for the other.
Solve this by add a new queue type argument and fix the clients of
the helper function.

Signed-off-by: Stanimir Varbanov <stanimir.varbanov@linaro.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: venus: core: Constify codec frequency data array

The array is not changed in the code, so make it const.

Signed-off-by: Stanimir Varbanov <stanimir.varbanov@linaro.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: venus: core: Fix mutex destroy in remove

The hfi_destroy function is called too early in remove method. It
destroys a mutex which is used later in the .remove from pmruntime.
Solve the issue by moving hfi_destroy after last usage of the mutex.

Signed-off-by: Stanimir Varbanov <stanimir.varbanov@linaro.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: venus: core: Add missing mutex destroy

This adds missing mutex_destroy in remove method of venus core driver.

Signed-off-by: Stanimir Varbanov <stanimir.varbanov@linaro.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: v4l2-ctrls.h: clarify the p_def argument of v4l2_ctrl_new_std_compound

It was not sufficiently clear how to create and use p_def.
Improve the documentation.

Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: cec-gpio: handle gpiod_get_value errors correctly

gpiod_get_value() can return negative values if an error occurs.
In several places this error code was ignored.

Ensure that errors codes are handled correctly throughout the CEC
pin framework and CEC pin drivers.

The return code of the cec_pin_ops read() callback had to be changed
from 'bool' to 'int', which mean the prototype of that callback in the
sun4i drm driver also had to be changed.

Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: vidioc-reqbufs/create-bufs.rst: fix typo

any others capabilities -> any other capabilities

Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: MAINTAINERS: adjust entries to moving CEC USB drivers

Commit a81068181aad ("media: move CEC USB drivers to a separate directory")
moved drivers/media/usb/{pulse8,rainshadow}-cec to
drivers/media/cec/usb/{rainshadow,pulse8}, but did not adjust the entries
in MAINTAINERS.

Since then, ./scripts/get_maintainer.pl --self-test=patterns complains:

  warning: no file matches F: drivers/media/usb/pulse8-cec/*
  warning: no file matches F: drivers/media/usb/rainshadow-cec/*

Update the MAINTAINERS entries to the new file locations.

Signed-off-by: Lukas Bulwahn <lukas.bulwahn@gmail.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: vimc: cap: Report a colorspace

The colorspace reported by a video nodes should not be
V4L2_COLORSPACE_DEFAULT. Instead a default colorspace should be picked
by the driver if V4L2_COLORSPACE_DEFAULT is given by userspace to
{G,S,TRY}_FMT.

The colorspace V4L2_COLORSPACE_SRGB is arbitrary chosen as the vimc
default format to report as it's used for most webcams.

Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Acked-by: Helen Koike <helen.koike@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: si2157: ensure wait_status is initialized

smatch reports wait_status is uninitialized, because smatch cannot assume
the loop body is ever executed.

Clarify the code so that wait_status is retrieved at least once.

Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: rc: no need for decoder state if decoder not enabled

One struct ir_raw_event_ctrl is allocated per raw IR device; reduce the
amount allocated if not all decoders are enabled.

Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: m88ds3103: error in set_frontend is swallowed and not reported

Bail out if registers can not be updated.

Addresses-Coverity-ID: 1461655 ("Code maintainability issues")

Reported-by: coverity-bot <keescook+coverity-bot@chromium.org>
Fixes: e6089feca460 ("media: m88ds3103: Add support for ds3103b demod")
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: Kconfig: fix some dvb-usb-v2 dependencies

There are some tuners used by a few pure DVB boards that use
internally V4L2 function calls. Due to that, such drivers now
depends on v4l2 core support, and can't be auto-selected
if !VIDEO_V4L2:

WARNING: unmet direct dependencies detected for DVB_RTL2832_SDR
  Depends on [n]: MEDIA_SUPPORT [=y] && MEDIA_DIGITAL_TV_SUPPORT [=y] && DVB_CORE [=y] && I2C [=y] && I2C_MUX [=y] && VIDEO_V4L2 [=n] && MEDIA_SDR_SUPPORT [=y] && USB [=y]
  Selected by [y]:
  - DVB_USB_RTL28XXU [=y] && USB [=y] && MEDIA_SUPPORT [=y] && MEDIA_USB_SUPPORT [=y] && I2C [=y] && MEDIA_DIGITAL_TV_SUPPORT [=y] && DVB_USB_V2 [=y] && I2C_MUX [=y] && MEDIA_SUBDRV_AUTOSELECT [=y] && MEDIA_SDR_SUPPORT [=y]

WARNING: unmet direct dependencies detected for MEDIA_TUNER_E4000
  Depends on [n]: (MEDIA_ANALOG_TV_SUPPORT [=y] || MEDIA_DIGITAL_TV_SUPPORT [=y] || MEDIA_RADIO_SUPPORT [=y] || MEDIA_SDR_SUPPORT [=y]) && MEDIA_SUPPORT [=y] && I2C [=y] && VIDEO_V4L2 [=n]
  Selected by [y]:
  - DVB_USB_RTL28XXU [=y] && USB [=y] && MEDIA_SUPPORT [=y] && MEDIA_USB_SUPPORT [=y] && I2C [=y] && MEDIA_DIGITAL_TV_SUPPORT [=y] && DVB_USB_V2 [=y] && I2C_MUX [=y] && MEDIA_SUBDRV_AUTOSELECT [=y]

WARNING: unmet direct dependencies detected for MEDIA_TUNER_FC2580
  Depends on [n]: (MEDIA_ANALOG_TV_SUPPORT [=y] || MEDIA_DIGITAL_TV_SUPPORT [=y] || MEDIA_RADIO_SUPPORT [=y] || MEDIA_SDR_SUPPORT [=y]) && MEDIA_SUPPORT [=y] && I2C [=y] && VIDEO_V4L2 [=n]
  Selected by [y]:
  - DVB_USB_RTL28XXU [=y] && USB [=y] && MEDIA_SUPPORT [=y] && MEDIA_USB_SUPPORT [=y] && I2C [=y] && MEDIA_DIGITAL_TV_SUPPORT [=y] && DVB_USB_V2 [=y] && I2C_MUX [=y] && MEDIA_SUBDRV_AUTOSELECT [=y]

Detected via randconfig builds.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: platform: fix VIA_CAMERA dependencies

As detected by a randconfig:

WARNING: unmet direct dependencies detected for VIDEO_OV7670
  Depends on [n]: MEDIA_SUPPORT [=y] && I2C [=y] && VIDEO_V4L2 [=n]
  Selected by [y]:
  - VIDEO_VIA_CAMERA [=y] && MEDIA_SUPPORT [=y] && MEDIA_PLATFORM_SUPPORT [=y] && V4L_PLATFORM_DRIVERS [=y] && FB_VIA [=y]

There's a missing dependency on VIDEO_V4L2 for VIA_CAMERA.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: tuners: Kconfig: add some missing VIDEO_V4L2 dependencies

There are three tuners that depend on some functions inside
V4L2 core:

$ git grep -l -E 'v4l2_[0-9a-zA-Z_]+\(' drivers/media/tuners/
drivers/media/tuners/e4000.c
drivers/media/tuners/fc2580.c
drivers/media/tuners/msi001.c

Make them dependent of VIDEO_V4L2, as otherwise, this would happen:

on x86_64:
CONFIG_VIDEO_V4L2=m
CONFIG_MEDIA_TUNER_E4000=y

ld: drivers/media/tuners/e4000.o: in function `e4000_remove':
e4000.c:(.text+0x34): undefined reference to `v4l2_ctrl_handler_free'
ld: drivers/media/tuners/e4000.o: in function `e4000_probe':
e4000.c:(.text+0x16c1): undefined reference to `v4l2_ctrl_handler_init_class'
ld: e4000.c:(.text+0x16eb): undefined reference to `v4l2_ctrl_new_std'
ld: e4000.c:(.text+0x1731): undefined reference to `v4l2_ctrl_new_std'
ld: e4000.c:(.text+0x1762): undefined reference to `v4l2_ctrl_auto_cluster'
ld: e4000.c:(.text+0x178c): undefined reference to `v4l2_ctrl_new_std'
ld: e4000.c:(.text+0x17d6): undefined reference to `v4l2_ctrl_new_std'
ld: e4000.c:(.text+0x1804): undefined reference to `v4l2_ctrl_auto_cluster'
ld: e4000.c:(.text+0x182e): undefined reference to `v4l2_ctrl_new_std'
ld: e4000.c:(.text+0x1878): undefined reference to `v4l2_ctrl_new_std'
ld: e4000.c:(.text+0x18a6): undefined reference to `v4l2_ctrl_auto_cluster'
ld: e4000.c:(.text+0x18d0): undefined reference to `v4l2_ctrl_new_std'
ld: e4000.c:(.text+0x191a): undefined reference to `v4l2_ctrl_new_std'
ld: e4000.c:(.text+0x1948): undefined reference to `v4l2_ctrl_auto_cluster'
ld: e4000.c:(.text+0x1972): undefined reference to `v4l2_ctrl_new_std'
ld: e4000.c:(.text+0x19a6): undefined reference to `v4l2_ctrl_handler_free'
ld: e4000.c:(.text+0x1a2a): undefined reference to `v4l2_i2c_subdev_init'

Reported-by: Randy Dunlap <rdunlap@infradead.org>
Acked-by: Randy Dunlap <rdunlap@infradead.org>
Tested-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

MAINTAINERS: reorder media attributes

Some media entry attributes got out of the order after the
bug was split into 3. Also, as reported by Johan, the
Rockchip VPU entry also had their attributes at the wrong
order.

As those entries weren't merged yet upstream, let's reorder
them with:

./scripts/parse-maintainers.pl --input=MAINTAINERS --output=MAINTAINERS --order

Reported-by: Johan Jonker <jbx6244@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: MAINTAINERS: Fix Hantro, Rga and Rkvdec entries

It seems recent merges introduced a couple issues
here, so let's fix them all. Also, reorder Rockchip
video decoder as per parse-maintainers.pl script
and add linux-rockchip mailing list.

Reported-by: Johan Jonker <jbx6244@gmail.com>
Signed-off-by: Ezequiel Garcia <ezequiel@collabora.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: rkisp1: fix missing mbus.type -> mbus_type conversion

There was one missing sensor->mbus.type to sensor->mbus_type
conversion which broke the build.

I suspect this was due to a merge conflict that was incorrectly resolved.

Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: admin-guide: split driver-specific indexes to new files

In order to better organize the main admin-guide index,
place the driver-specific indexes on separate files.

This ensures a more consistent numbering at the main index.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: admin-guide: reorganize the guide

Separate the generic documentation from the driver-specific
parts, and use a better title for the documentation.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: admin-guide: bt8xx.rst: fix a broken cross-reference

The old cards.rst file doesn't exist anymore. Also, it is not
the right reference there, as it should be pointing to the
bttv-specific cardlist.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: admin-guide: add test-drivers

While test drivers is probably something that normal users
won't bother, it doesn't hurt to add them to the card list.

This way, all drivers, except for the ones under staging,
would be listed there.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: admin-guide: add card lists for radio and firewire

Document the supported radio devices, plus document the
firewire driver.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: admin-guide: add SPI cards to platform drivers

The SPI cards are also under the concept of "platform" drivers
as defined at the section introduction.

So, add a SPI card list there.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

media: admin-guide: add a list of I2C drivers

For completeness, let's add a list of I2C drivers.

Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>