platform/kernel/linux-starfive.git
6 years agomedia: v4l: fwnode: Print bus type
Sakari Ailus [Tue, 2 Jan 2018 15:44:34 +0000 (10:44 -0500)]
media: v4l: fwnode: Print bus type

Print bus type either as set by the driver or as parsed from the bus-type
property, as well as the guessed V4L2 media bus type.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Tested-by: Steve Longerbeam <steve_longerbeam@mentor.com>
Tested-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: v4l: fwnode: Use default parallel flags
Sakari Ailus [Tue, 3 Jul 2018 11:09:25 +0000 (07:09 -0400)]
media: v4l: fwnode: Use default parallel flags

The caller may provide default flags for the endpoint. Change the
configuration based on what is available through the fwnode property API.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Tested-by: Steve Longerbeam <steve_longerbeam@mentor.com>
Reviewed-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
Tested-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: v4l: fwnode: Use media bus type for bus parser selection
Sakari Ailus [Tue, 17 Jul 2018 16:17:02 +0000 (12:17 -0400)]
media: v4l: fwnode: Use media bus type for bus parser selection

Use the media bus types instead of the fwnode bus types internally. This
is the interface to the drivers as well, making the use of the fwnode bus
types more localised to the V4L2 fwnode framework.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Tested-by: Steve Longerbeam <steve_longerbeam@mentor.com>
Tested-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: v4l: fwnode: Only zero the struct if bus type is set to V4L2_MBUS_UNKNOWN
Sakari Ailus [Tue, 31 Jul 2018 09:21:53 +0000 (05:21 -0400)]
media: v4l: fwnode: Only zero the struct if bus type is set to V4L2_MBUS_UNKNOWN

In order to prepare for allowing drivers to set the defaults for a given
bus, make zeroing the struct conditional based on detecting the bus.
All callers now set the bus type to zero which allows only zeroing the
remaining bus union.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Tested-by: Steve Longerbeam <steve_longerbeam@mentor.com>
Tested-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: v4l: fwnode: Initialise the V4L2 fwnode endpoints to zero
Sakari Ailus [Tue, 31 Jul 2018 09:15:50 +0000 (05:15 -0400)]
media: v4l: fwnode: Initialise the V4L2 fwnode endpoints to zero

Initialise the V4L2 fwnode endpoints to zero in all drivers using
v4l2_fwnode_endpoint_parse(). This prepares for setting default endpoint
flags as well as the bus type. Setting bus type to zero will continue to
guess the bus among the guessable set (parallel, Bt.656 and CSI-2 D-PHY).

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Tested-by: Steve Longerbeam <steve_longerbeam@mentor.com>
Tested-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: v4l: fwnode: Parse the graph endpoint as last
Sakari Ailus [Wed, 18 Jul 2018 13:20:34 +0000 (09:20 -0400)]
media: v4l: fwnode: Parse the graph endpoint as last

Parsing the graph endpoint is always successful; therefore parse it as
last.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Tested-by: Steve Longerbeam <steve_longerbeam@mentor.com>
Tested-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: v4l: fwnode: Support default CSI-2 lane mapping for drivers
Sakari Ailus [Fri, 20 Jul 2018 13:57:48 +0000 (09:57 -0400)]
media: v4l: fwnode: Support default CSI-2 lane mapping for drivers

Most hardware doesn't support re-mapping of the CSI-2 lanes. Especially
sensor drivers have a default number of lanes. Instead of requiring the
caller (the driver) to provide such a unit mapping, provide one if no
mapping is configured.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Tested-by: Steve Longerbeam <steve_longerbeam@mentor.com>
Tested-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: v4l: fwnode: Support driver-defined lane mapping defaults
Sakari Ailus [Fri, 20 Jul 2018 10:14:14 +0000 (06:14 -0400)]
media: v4l: fwnode: Support driver-defined lane mapping defaults

Make use of the default CSI-2 lane mapping from caller-passed
configuration.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Tested-by: Steve Longerbeam <steve_longerbeam@mentor.com>
Tested-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: v4l: fwnode: Only assign configuration if there is no error
Sakari Ailus [Tue, 3 Jul 2018 11:06:20 +0000 (07:06 -0400)]
media: v4l: fwnode: Only assign configuration if there is no error

Only assign endpoint configuration if the endpoint is parsed successfully.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Tested-by: Steve Longerbeam <steve_longerbeam@mentor.com>
Tested-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: v4l: fwnode: Read lane inversion information despite lane numbering
Sakari Ailus [Wed, 18 Jul 2018 13:33:45 +0000 (09:33 -0400)]
media: v4l: fwnode: Read lane inversion information despite lane numbering

Read the lane inversion independently of whether the "data-lanes" property
exists. This makes sense since the caller may pass the number of lanes as
the default configuration while the lane inversion configuration may still
be available in firmware.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Tested-by: Steve Longerbeam <steve_longerbeam@mentor.com>
Tested-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: v4l: fwnode: Make use of newly specified bus types
Sakari Ailus [Tue, 2 Jan 2018 13:14:30 +0000 (08:14 -0500)]
media: v4l: fwnode: Make use of newly specified bus types

Add support for parsing CSI-2 D-PHY, parallel or Bt.656 bus explicitly.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Tested-by: Steve Longerbeam <steve_longerbeam@mentor.com>
Reviewed-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
Tested-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: v4l: fwnode: Detect bus type correctly
Sakari Ailus [Tue, 31 Jul 2018 10:43:14 +0000 (06:43 -0400)]
media: v4l: fwnode: Detect bus type correctly

In case the device supports multiple video bus types on an endpoint, the
V4L2 fwnode framework attempts to detect the type based on the available
information. This wasn't working really well, and sometimes could lead to
the V4L2 fwnode endpoint struct as being mishandled between the bus types.

Default to Bt.656 if no properties suggesting a bus type are found.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Tested-by: Steve Longerbeam <steve_longerbeam@mentor.com>
Tested-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: v4l: fwnode: Let the caller provide V4L2 fwnode endpoint
Sakari Ailus [Sat, 2 Jun 2018 16:19:35 +0000 (12:19 -0400)]
media: v4l: fwnode: Let the caller provide V4L2 fwnode endpoint

Instead of allocating the V4L2 fwnode endpoint in
v4l2_fwnode_endpoint_alloc_parse, let the caller to do this. This allows
setting default parameters for the endpoint which is a very common need
for drivers.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Tested-by: Steve Longerbeam <steve_longerbeam@mentor.com>
Tested-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: v4l: mediabus: Recognise CSI-2 D-PHY and C-PHY
Sakari Ailus [Tue, 3 Jul 2018 21:19:27 +0000 (17:19 -0400)]
media: v4l: mediabus: Recognise CSI-2 D-PHY and C-PHY

The CSI-2 bus may use either D-PHY or C-PHY. Make this visible in media
bus enum.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Tested-by: Steve Longerbeam <steve_longerbeam@mentor.com>
Tested-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: v4l: fwnode: Add definitions for CSI-2 D-PHY, parallel and Bt.656 busses
Sakari Ailus [Tue, 3 Jul 2018 21:40:37 +0000 (17:40 -0400)]
media: v4l: fwnode: Add definitions for CSI-2 D-PHY, parallel and Bt.656 busses

Add definitions corresponding to DT bindings to the CSI-2 D-PHY, parallel
and Bt.656 busses.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Tested-by: Steve Longerbeam <steve_longerbeam@mentor.com>
Tested-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: dt-bindings: media: Specify bus type for MIPI D-PHY, others, explicitly
Sakari Ailus [Tue, 2 Jan 2018 13:12:35 +0000 (08:12 -0500)]
media: dt-bindings: media: Specify bus type for MIPI D-PHY, others, explicitly

Allow specifying the bus type explicitly for MIPI D-PHY, parallel and
Bt.656 busses. This is useful for devices that can make use of different
bus types. There are CSI-2 transmitters and receivers but the PHY
selection needs to be made between C-PHY and D-PHY; many devices also
support parallel and Bt.656 interfaces but the means to pass that
information to software wasn't there.

Autodetection (value 0) is removed as an option as the property could be
simply omitted in that case.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Acked-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: v4l: fwnode: The CSI-2 clock is continuous if it's not non-continuous
Sakari Ailus [Fri, 20 Jul 2018 09:04:29 +0000 (05:04 -0400)]
media: v4l: fwnode: The CSI-2 clock is continuous if it's not non-continuous

The continuous clock flag was only set if there was a clock or data lanes.
This isn't needed as such a configuration is invalid to begin with. Always
set the continuous clock flag if the non-continuous property is not found.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Tested-by: Steve Longerbeam <steve_longerbeam@mentor.com>
Tested-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: v4l: fwnode: Use fwnode_graph_for_each_endpoint
Sakari Ailus [Thu, 21 Dec 2017 12:11:19 +0000 (07:11 -0500)]
media: v4l: fwnode: Use fwnode_graph_for_each_endpoint

Use fwnode_graph_for_each_endpoint iterator for better readability.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Tested-by: Steve Longerbeam <steve_longerbeam@mentor.com>
Tested-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: v4l: fwnode: Add debug prints for V4L2 endpoint property parsing
Sakari Ailus [Mon, 4 Dec 2017 21:25:06 +0000 (16:25 -0500)]
media: v4l: fwnode: Add debug prints for V4L2 endpoint property parsing

Print debug info as standard V4L2 endpoint are parsed.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Tested-by: Steve Longerbeam <steve_longerbeam@mentor.com>
Tested-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: v4l2-subdev.rst: Update doc regarding subdev descriptors
Steve Longerbeam [Sat, 29 Sep 2018 19:54:20 +0000 (15:54 -0400)]
media: v4l2-subdev.rst: Update doc regarding subdev descriptors

Update the doc to describe the new method of adding subdevice
descriptors to async notifiers.

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+samsung@kernel.org>
6 years agomedia: v4l2: async: Remove notifier subdevs array
Steve Longerbeam [Sat, 29 Sep 2018 19:54:19 +0000 (15:54 -0400)]
media: v4l2: async: Remove notifier subdevs array

All platform drivers have been converted to use
v4l2_async_notifier_add_subdev(), in place of adding
asd's to the notifier subdevs array. So the subdevs
array can now be removed from struct v4l2_async_notifier,
and remove the backward compatibility support for that
array in v4l2-async.c.

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+samsung@kernel.org>
6 years agomedia: platform: Switch to v4l2_async_notifier_add_subdev
Steve Longerbeam [Sat, 29 Sep 2018 19:54:18 +0000 (15:54 -0400)]
media: platform: Switch to v4l2_async_notifier_add_subdev

Switch all media platform drivers to call v4l2_async_notifier_add_subdev()
to add asd's to a notifier, in place of referencing the notifier->subdevs[]
array. These drivers also must now call v4l2_async_notifier_init() before
adding asd's to their notifiers.

There may still be cases where a platform driver maintains a list of
asd's that is a duplicate of the notifier asd_list, in which case its
possible the platform driver list can be removed, and can reference the
notifier asd_list instead. One example of where a duplicate list has
been removed in this patch is xilinx-vipp.c. If there are such cases
remaining, those drivers should be optimized to remove the duplicate
platform driver asd lists.

None of the changes to the platform drivers in this patch have been
tested. Verify that the async subdevices needed by the platform are
bound at load time, and that the driver unloads and reloads correctly
with no memory leaking of asd objects.

Suggested-by: Sakari Ailus <sakari.ailus@linux.intel.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+samsung@kernel.org>
6 years agomedia: staging/imx: TODO: Remove one assumption about OF graph parsing
Steve Longerbeam [Sat, 29 Sep 2018 19:54:17 +0000 (15:54 -0400)]
media: staging/imx: TODO: Remove one assumption about OF graph parsing

The move to subdev notifiers fixes one assumption of OF graph parsing.
If a subdevice has non-video related ports, the subdev driver knows not
to follow those ports when adding remote devices to its subdev notifier.

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+samsung@kernel.org>
6 years agomedia: staging/imx: Switch to v4l2_async_notifier_add_*_subdev
Steve Longerbeam [Sat, 29 Sep 2018 19:54:16 +0000 (15:54 -0400)]
media: staging/imx: Switch to v4l2_async_notifier_add_*_subdev

Switch to v4l2_async_notifier_add_*_subdev() when adding async subdevs
to the imx-media root notifier. This removes the need to check for
an already added asd, since v4l2_async_notifier_add_*_subdev() does this
check. Also no need to allocate a subdevs array when registering the
root notifier, or keeping an internal master asd_list, since this is
moved to the notifier's asd_list.

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+samsung@kernel.org>
6 years agomedia: staging/imx: Rename root notifier
Steve Longerbeam [Sat, 29 Sep 2018 19:54:15 +0000 (15:54 -0400)]
media: staging/imx: Rename root notifier

Rename the imx-media root async notifier from "subdev_notifier" to
simply "notifier", so as not to confuse it with true subdev notifiers.
No functional changes.

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+samsung@kernel.org>
6 years agomedia: staging/imx: Loop through all registered subdevs for media links
Steve Longerbeam [Sat, 29 Sep 2018 19:54:14 +0000 (15:54 -0400)]
media: staging/imx: Loop through all registered subdevs for media links

The root imx-media notifier no longer sees all bound subdevices because
some of them will be bound to subdev notifiers. So imx_media_create_links()
now needs to loop through all subdevices registered with the v4l2-device,
not just the ones in the root notifier's done list. This should be safe
because imx_media_create_of_links() checks if a fwnode link already
exists before creating.

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+samsung@kernel.org>
6 years agomedia: staging/imx: of: Remove recursive graph walk
Steve Longerbeam [Sat, 29 Sep 2018 19:54:13 +0000 (15:54 -0400)]
media: staging/imx: of: Remove recursive graph walk

After moving to subdev notifiers, it's no longer necessary to recursively
walk the OF graph, because the subdev notifiers will discover and add
devices from the graph for us.

So the recursive of_parse_subdev() function is gone, replaced with
of_add_csi() which adds only the CSI port fwnodes to the imx-media
root notifier.

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+samsung@kernel.org>
6 years agomedia: imx: mipi csi-2: Register a subdev notifier
Steve Longerbeam [Sat, 29 Sep 2018 19:54:12 +0000 (15:54 -0400)]
media: imx: mipi csi-2: Register a subdev notifier

Parse neighbor remote devices on the MIPI CSI-2 input port, add
them to a subdev notifier, and register the subdev notifier for the
MIPI CSI-2 receiver, by calling v4l2_async_register_fwnode_subdev().

csi2_parse_endpoints() is modified to be the parse_endpoint callback.

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+samsung@kernel.org>
6 years agomedia: imx: csi: Register a subdev notifier
Steve Longerbeam [Sat, 29 Sep 2018 19:54:11 +0000 (15:54 -0400)]
media: imx: csi: Register a subdev notifier

Parse neighbor remote devices on the CSI port, and add them to a subdev
notifier, by calling v4l2_async_notifier_parse_fwnode_endpoints_by_port()
using the CSI's port id. And register the subdev notifier for the CSI.

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+samsung@kernel.org>
6 years agomedia: platform: video-mux: Register a subdev notifier
Steve Longerbeam [Sat, 29 Sep 2018 19:54:10 +0000 (15:54 -0400)]
media: platform: video-mux: Register a subdev notifier

Parse neighbor remote devices on the video muxes input ports, add them to a
subdev notifier, and register the subdev notifier for the video mux, by
calling v4l2_async_register_fwnode_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+samsung@kernel.org>
6 years agomedia: v4l2-fwnode: Add a convenience function for registering subdevs with notifiers
Steve Longerbeam [Sat, 29 Sep 2018 19:54:09 +0000 (15:54 -0400)]
media: v4l2-fwnode: Add a convenience function for registering subdevs with notifiers

Adds v4l2_async_register_fwnode_subdev(), which is a convenience function
for parsing a sub-device's fwnode port endpoints for connected remote
sub-devices, registering a sub-device notifier, and then registering
the sub-device itself.

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+samsung@kernel.org>
6 years agomedia: v4l2-fwnode: Switch to v4l2_async_notifier_add_subdev
Steve Longerbeam [Sat, 29 Sep 2018 19:54:08 +0000 (15:54 -0400)]
media: v4l2-fwnode: Switch to v4l2_async_notifier_add_subdev

The fwnode endpoint and reference parsing functions in v4l2-fwnode.c
are modified to make use of v4l2_async_notifier_add_subdev().
As a result the notifier->subdevs array is no longer allocated or
re-allocated, and by extension the max_subdevs value is also no
longer needed.

Callers of the fwnode endpoint and reference parsing functions must now
first initialize the notifier with a call to v4l2_async_notifier_init().
This includes the function v4l2_async_register_subdev_sensor_common(),
and the intel-ipu3, omap3isp, and rcar-vin drivers.

Since the notifier->subdevs array is no longer allocated in the
fwnode endpoint and reference parsing functions, the callers of
those functions must never reference that array, since it is now
NULL. Of the drivers that make use of the fwnode/ref parsing,
only the intel-ipu3 driver references the ->subdevs[] array,
(in the notifier completion callback), so that driver has been
modified to iterate through the notifier->asd_list instead.

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+samsung@kernel.org>
6 years agomedia: v4l2: async: Add convenience functions to allocate and add asd's
Steve Longerbeam [Sat, 29 Sep 2018 19:54:07 +0000 (15:54 -0400)]
media: v4l2: async: Add convenience functions to allocate and add asd's

Add these convenience functions, which allocate an asd of match type
fwnode, i2c, or device-name, of size asd_struct_size, and then adds
them to the notifier asd_list.

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+samsung@kernel.org>
6 years agomedia: v4l2: async: Add v4l2_async_notifier_add_subdev
Steve Longerbeam [Sat, 29 Sep 2018 19:54:06 +0000 (15:54 -0400)]
media: v4l2: async: Add v4l2_async_notifier_add_subdev

v4l2_async_notifier_add_subdev() adds an asd to the notifier. It checks
that no other equivalent asd's have already been added to this notifier's
asd list, or to other registered notifier's waiting or done lists, and
increments num_subdevs.

v4l2_async_notifier_add_subdev() does not make use of the notifier subdevs
array, otherwise it would have to re-allocate the array every time the
function was called. In place of the subdevs array, the function adds
the newly allocated asd to a new master asd_list. The function will
return error with a WARN() if it is ever called with the subdevs array
allocated.

Drivers are now required to call a v4l2_async_notifier_init(), before the
first call to v4l2_async_notifier_add_subdev(), in order to initialize
the asd_list.

In v4l2_async_notifier_has_async_subdev(), __v4l2_async_notifier_register(),
and v4l2_async_notifier_cleanup(), maintain backward compatibility with
the subdevs array, by alternatively operate on the subdevs array or a
non-empty notifier->asd_list.

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+samsung@kernel.org>
6 years agomedia: v4l2: async: Allow searching for asd of any type
Steve Longerbeam [Sat, 29 Sep 2018 19:54:05 +0000 (15:54 -0400)]
media: v4l2: async: Allow searching for asd of any type

Generalize v4l2_async_notifier_fwnode_has_async_subdev() to allow
searching for any type of async subdev, not just fwnodes. Rename to
v4l2_async_notifier_has_async_subdev() and pass it an asd pointer.

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+samsung@kernel.org>
6 years agomedia: v4l2-fwnode: ignore endpoints that have no remote port parent
Steve Longerbeam [Sat, 29 Sep 2018 19:54:04 +0000 (15:54 -0400)]
media: v4l2-fwnode: ignore endpoints that have no remote port parent

Documentation/devicetree/bindings/media/video-interfaces.txt states that
the 'remote-endpoint' property is optional.

So v4l2_async_notifier_fwnode_parse_endpoint() should not return error
if the endpoint has no remote port parent. Just ignore the endpoint,
skip adding an asd to the notifier and return 0.
__v4l2_async_notifier_parse_fwnode_endpoints() will then continue
parsing the remaining port endpoints of the device.

Signed-off-by: Steve Longerbeam <slongerbeam@gmail.com>
Acked-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: i2c: adv748x: fix typo in comment for TXB CSI-2 transmitter power down
Niklas Söderlund [Tue, 18 Sep 2018 01:45:09 +0000 (21:45 -0400)]
media: i2c: adv748x: fix typo in comment for TXB CSI-2 transmitter power down

Fix copy-and-paste error in comment for TXB CSI-2 transmitter power down
sequence.

Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: i2c: adv748x: Register only enabled inputs
Jacopo Mondi [Mon, 17 Sep 2018 11:30:57 +0000 (07:30 -0400)]
media: i2c: adv748x: Register only enabled inputs

The adv748x assumes input endpoints are always enabled, and registers
a subdevice for each of them when the corresponding output subdevice
is registered.

Fix this by conditionally registering the input subdevice only if it is
actually described in device tree.

Tested-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
Signed-off-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: i2c: adv748x: Conditionally enable only CSI-2 outputs
Jacopo Mondi [Mon, 17 Sep 2018 11:30:56 +0000 (07:30 -0400)]
media: i2c: adv748x: Conditionally enable only CSI-2 outputs

The ADV748x has two CSI-2 output port and one TTL input/output port for
digital video reception/transmission. The TTL digital pad is unconditionally
enabled during the device reset even if not used. Same goes for the TXA
and TXB CSI-2 outputs, which are enabled by the initial settings blob
programmed into the chip.

In order to improve power saving, do not enable unused output interfaces:
keep TTL output disabled, as it is not used, and drop CSI-2 output enabling
from the initial settings list, as they get conditionally enabled later.

Tested-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
Signed-off-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: i2c: adv748x: Handle TX[A|B] power management
Jacopo Mondi [Mon, 17 Sep 2018 11:30:55 +0000 (07:30 -0400)]
media: i2c: adv748x: Handle TX[A|B] power management

As the driver is now allowed to probe with a single output endpoint,
power management routines shall now take into account the case a CSI-2 TX
is not enabled.

Unify the adv748x_tx_power() routine to handle transparently TXA and TXB,
and enable the CSI-2 outputs conditionally.

Tested-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
Signed-off-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: i2c: adv748x: Support probing a single output
Jacopo Mondi [Mon, 17 Sep 2018 11:30:54 +0000 (07:30 -0400)]
media: i2c: adv748x: Support probing a single output

Currently the adv748x driver will fail to probe unless both of its
output endpoints (TXA and TXB) are connected.

Make the driver support probing provided that there is at least one
input, and one output connected and protect the clean-up function from
accessing un-initialized fields.

Following patches will fix other uses of un-initialized TXs in the driver,
such as power management functions.

Tested-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
Signed-off-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: MAINTAINERS: fix reference to STI CEC driver
Benjamin Gaignard [Mon, 1 Oct 2018 07:46:40 +0000 (03:46 -0400)]
media: MAINTAINERS: fix reference to STI CEC driver

STI CEC driver has move from staging directory to media/platform/sti/cec/

Signed-off-by: Benjamin Gaignard <benjamin.gaignard@st.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: zoran: fix spelling mistake "queing" -> "queuing"
Colin Ian King [Sat, 29 Sep 2018 11:25:39 +0000 (07:25 -0400)]
media: zoran: fix spelling mistake "queing" -> "queuing"

Trivial fix to spelling mistake in kernel warning message

Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: rcar-vin: fix redeclaration of symbol
Niklas Söderlund [Wed, 26 Sep 2018 21:40:06 +0000 (17:40 -0400)]
media: rcar-vin: fix redeclaration of symbol

When adding support for parallel subdev for Gen3 it was missed that the
symbol 'i' in rvin_group_link_notify() was already declared, remove the
dupe as it's only used as a loop variable this have no functional
change. This fixes warning:

    rcar-core.c:117:52: originally declared here
    rcar-core.c:173:30: warning: symbol 'i' shadows an earlier one

Fixes: 1284605dc821cebd ("media: rcar-vin: Handle parallel subdev in link_notify")

Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Acked-by: Jacopo Mondi <jacopo+renesas@jmondi.org>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: imx-pxp: include linux/interrupt.h
Arnd Bergmann [Wed, 26 Sep 2018 13:01:26 +0000 (09:01 -0400)]
media: imx-pxp: include linux/interrupt.h

The newly added driver fails to build in some configurations due to a
missing header inclusion:

drivers/media/platform/imx-pxp.c:988:8: error: unknown type name 'irqreturn_t'
 static irqreturn_t pxp_irq_handler(int irq, void *dev_id)
        ^~~~~~~~~~~
drivers/media/platform/imx-pxp.c: In function 'pxp_irq_handler':
drivers/media/platform/imx-pxp.c:1012:9: error: 'IRQ_HANDLED' undeclared (first use in this function); did you mean 'IRQ_MODE'?
  return IRQ_HANDLED;
         ^~~~~~~~~~~
         IRQ_MODE
drivers/media/platform/imx-pxp.c:1012:9: note: each undeclared identifier is reported only once for each function it appears in
drivers/media/platform/imx-pxp.c: In function 'pxp_probe':
drivers/media/platform/imx-pxp.c:1660:8: error: implicit declaration of function 'devm_request_threaded_irq'; did you mean 'devm_request_region'? [-Werror=implicit-function-declaration]
  ret = devm_request_threaded_irq(&pdev->dev, irq, NULL, pxp_irq_handler,
        ^~~~~~~~~~~~~~~~~~~~~~~~~
        devm_request_region
drivers/media/platform/imx-pxp.c:1661:4: error: 'IRQF_ONESHOT' undeclared (first use in this function); did you mean 'SA_ONESHOT'?
    IRQF_ONESHOT, dev_name(&pdev->dev), dev);

Fixes: 51abcf7fdb70 ("media: imx-pxp: add i.MX Pixel Pipeline driver")

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: stm32-dcmi: only enable IT frame on JPEG capture
Hugues Fruchet [Mon, 24 Sep 2018 12:54:59 +0000 (08:54 -0400)]
media: stm32-dcmi: only enable IT frame on JPEG capture

Only enable IT frame on JPEG capture, this saves some CPU
interruptions and processing on all the other cases.

Signed-off-by: Hugues Fruchet <hugues.fruchet@st.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: pxa_camera: Fix check for pdev->dev.of_node
Nathan Chancellor [Fri, 21 Sep 2018 10:00:45 +0000 (06:00 -0400)]
media: pxa_camera: Fix check for pdev->dev.of_node

Clang warns that the address of a pointer will always evaluated as true
in a boolean context.

drivers/media/platform/pxa_camera.c:2400:17: warning: address of
'pdev->dev.of_node' will always evaluate to 'true'
[-Wpointer-bool-conversion]
        if (&pdev->dev.of_node && !pcdev->pdata) {
             ~~~~~~~~~~^~~~~~~ ~~
1 warning generated.

Judging from the rest of the kernel, it seems like this was an error and
just the value of of_node should be checked rather than the address.

Reported-by: Nick Desaulniers <ndesaulniers@google.com>
Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: qcom: remove duplicated include file
zhong jiang [Fri, 21 Sep 2018 03:30:20 +0000 (23:30 -0400)]
media: qcom: remove duplicated include file

We include device.h twice in camss.h. It's unnecessary.
hence just remove it.

Signed-off-by: zhong jiang <zhongjiang@huawei.com>
Acked-by: Todor Tomov <todor.tomov@linaro.org>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: imx: use well defined 32-bit RGB pixel format
Philipp Zabel [Tue, 18 Sep 2018 09:42:31 +0000 (05:42 -0400)]
media: imx: use well defined 32-bit RGB pixel format

The documentation in Documentation/media/uapi/v4l/pixfmt-packed-rgb.rst
tells us that the V4L2_PIX_FMT_RGB32 format is deprecated and must not
be used by new drivers. Replace it with V4L2_PIX_FMT_XRGB32.

Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Acked-by: Steve Longerbeam <steve_longerbeam@mentor.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: VPU: mediatek: don't pass an unused parameter
Dan Carpenter [Mon, 17 Sep 2018 13:56:22 +0000 (09:56 -0400)]
media: VPU: mediatek: don't pass an unused parameter

The load_requested_vpu() function returns a freed vpu_fw pointer.  It's
not used so it doesn't cause any problems, but Smatch complains about
it:

    drivers/media/platform/mtk-vpu/mtk_vpu.c:578 vpu_load_firmware()
    warn: passing freed memory 'vpu_fw'

We can clean up the code a bit and silence the static checker warning
by not passing the parameter at all.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: vidioc-dqevent.rst: clarify V4L2_EVENT_SRC_CH_RESOLUTION
Hans Verkuil [Mon, 17 Sep 2018 11:37:18 +0000 (07:37 -0400)]
media: vidioc-dqevent.rst: clarify V4L2_EVENT_SRC_CH_RESOLUTION

Clarify that when you receive V4L2_EVENT_SOURCE_CHANGE with flag
V4L2_EVENT_SRC_CH_RESOLUTION set, and the new resolution appears
identical to the old resolution, then you still must restart the
streaming I/O.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: rc: ir-rc6-decoder: enable toggle bit for Kathrein RCU-676 remote
Matthias Reichl [Tue, 28 Aug 2018 13:49:42 +0000 (09:49 -0400)]
media: rc: ir-rc6-decoder: enable toggle bit for Kathrein RCU-676 remote

The Kathrein RCU-676 remote uses the 32-bit rc6 protocol and toggles
bit 15 (0x8000) on repeated button presses, like MCE remotes.

Add it's customer code 0x80460000 to the 32-bit rc6 toggle
handling code to get proper scancodes and toggle reports.

Signed-off-by: Matthias Reichl <hias@horus.com>
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: rc: Remove init_ir_raw_event and DEFINE_IR_RAW_EVENT macros
Sean Young [Tue, 21 Aug 2018 19:57:52 +0000 (15:57 -0400)]
media: rc: Remove init_ir_raw_event and DEFINE_IR_RAW_EVENT macros

This can be done with c99 initializers, which makes the code cleaner
and more transparent. It does require gcc 4.6, because of this bug
in earlier versions:

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=10676

Since commit cafa0010cd51 ("Raise the minimum required gcc version to
4.6"), this is the case.

Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: rc: nec keymaps should specify the nec variant they use
Sean Young [Tue, 17 Jul 2018 21:33:05 +0000 (17:33 -0400)]
media: rc: nec keymaps should specify the nec variant they use

The rc_proto field should list the exact variant used by the remote. This
does not change the decoder used, but helps with using keymaps for
transmit purposes.

Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: smiapp: Remove unused loop
Ricardo Ribalda Delgado [Wed, 26 Sep 2018 06:12:42 +0000 (02:12 -0400)]
media: smiapp: Remove unused loop

The loop seemed to be made to calculate max, but max is not used in that
function.

Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@gmail.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: ov5640: use JPEG mode 3 for 720p
Hugues Fruchet [Mon, 24 Sep 2018 13:33:27 +0000 (09:33 -0400)]
media: ov5640: use JPEG mode 3 for 720p

Change 720p JPEG mode to mode 3 as per other resolutions.

Signed-off-by: Hugues Fruchet <hugues.fruchet@st.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: v4l: Remove support for crop default target in subdev drivers
Sakari Ailus [Mon, 24 Sep 2018 14:38:11 +0000 (10:38 -0400)]
media: v4l: Remove support for crop default target in subdev drivers

The V4L2 sub-device API does not support the crop default target. A number
of drivers apparently still did support this, likely as it was needed by
the SoC camera framework. Drop support for the default crop rectaingle in
sub-device drivers, and use the bounds rectangle in SoC camera instead.

Reported-by: Helmut Grohne <h.grohne@intenta.de>
Suggested-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Acked-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: v4l: i2c: Add a comment not to use static sub-device names in the future
Sakari Ailus [Mon, 24 Sep 2018 14:24:43 +0000 (10:24 -0400)]
media: v4l: i2c: Add a comment not to use static sub-device names in the future

A number of sub-device drivers used a static name for the sub-device, and
thus the media entity. As the entity name must be unique within a media
device, this makes it impossible to have more than one instance of each
device in a media device. This is a rather severe limitation.

Instead of fixing these drivers, add a comment to the drivers noting that
such static names may not be used in the future.

The alternative of fixing the drivers is troublesome as the entity (as
well as sub-device) name is part of the uAPI. Changing that is almost
certain to break something. As these devices are old but no-one has
encountered a problem with the static names, leave it as-is.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: davinci: Fix implicit enum conversion warning
Nathan Chancellor [Sat, 15 Sep 2018 06:16:15 +0000 (02:16 -0400)]
media: davinci: Fix implicit enum conversion warning

Clang warns when one enumerated type is implicitly converted to another.

drivers/media/platform/davinci/vpbe_display.c:524:24: warning: implicit
conversion from enumeration type 'enum osd_v_exp_ratio' to different
enumeration type 'enum osd_h_exp_ratio' [-Wenum-conversion]
                        layer_info->h_exp = V_EXP_6_OVER_5;
                                          ~ ^~~~~~~~~~~~~~
1 warning generated.

This appears to be a copy and paste error judging from the couple of
lines directly above this statement and the way that height is handled
in the if block above this one.

Reported-by: Nick Desaulniers <ndesaulniers@google.com>
Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: bt8xx: Remove unnecessary self-assignment
Nathan Chancellor [Sat, 15 Sep 2018 06:06:13 +0000 (02:06 -0400)]
media: bt8xx: Remove unnecessary self-assignment

Clang warns when a variable is assigned to itself.

drivers/media/pci/bt8xx/bttv-driver.c:2043:13: warning: explicitly
assigning value of variable of type '__s32' (aka 'int') to itself
[-Wself-assign]
        min_height = min_height;
        ~~~~~~~~~~ ^ ~~~~~~~~~~
1 warning generated.

There doesn't appear to be any good reason for this and this statement
was added in commit e5bd0260e7d3 ("V4L/DVB (5077): Bttv cropping
support") back in 2007. Just remove it.

Reported-by: Nick Desaulniers <ndesaulniers@google.com>
Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: v4l2-common: fix typo in documentation for v4l_bound_align_image()
Niklas Söderlund [Thu, 13 Sep 2018 00:07:38 +0000 (20:07 -0400)]
media: v4l2-common: fix typo in documentation for v4l_bound_align_image()

Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: Simon Horman <horms+renesas@verge.net.au>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: pci: cx23885: handle adding to list failure
Nicholas Mc Guire [Sun, 9 Sep 2018 16:02:32 +0000 (12:02 -0400)]
media: pci: cx23885: handle adding to list failure

altera_hw_filt_init() which calls append_internal() assumes
that the node was successfully linked in while in fact it can
silently fail. So the call-site needs to set return to -ENOMEM
on append_internal() returning NULL and exit through the err path.

Fixes: 349bcf02e361 ("[media] Altera FPGA based CI driver module")

Signed-off-by: Nicholas Mc Guire <hofrat@osadl.org>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: au0828: Fix incorrect error messages
Brad Love [Thu, 6 Sep 2018 21:07:49 +0000 (17:07 -0400)]
media: au0828: Fix incorrect error messages

Correcting red herring error messages.

Where appropriate, replaces au0282_dev_register with:
- au0828_analog_register
- au0828_dvb_register

Signed-off-by: Brad Love <brad@nextdimension.cc>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: au0828: cannot kfree dev before usb disconnect
Brad Love [Thu, 6 Sep 2018 21:07:48 +0000 (17:07 -0400)]
media: au0828: cannot kfree dev before usb disconnect

If au0828_analog_register fails, the dev is kfree'd and then flow
jumps to done, which can call au0828_usb_disconnect. Since all USB
error codes are negative, au0828_usb_disconnect will be called. The
problem is au0828_usb_disconnect uses dev, if dev is NULL then there
is immediate oops encountered.

[    7.454307] au0828: au0828_usb_probe() au0282_dev_register failed to register on V4L2
[    7.454323] BUG: unable to handle kernel NULL pointer dereference at 0000000000000050
[    7.454421] PGD 0 P4D 0
[    7.454457] Oops: 0002 [#1] SMP PTI
[    7.454500] CPU: 1 PID: 262 Comm: systemd-udevd Tainted: P           O      4.18.3 #1
[    7.454584] Hardware name: Google Panther/Panther, BIOS MattDevo 04/27/2015
[    7.454670] RIP: 0010:_raw_spin_lock_irqsave+0x2c/0x50
[    7.454725] Code: 44 00 00 55 48 89 e5 41 54 53 48 89 fb 9c 58 0f 1f 44 00 00 49 89 c4 fa 66 0f 1f 44 00 00 e8 db 23 1b ff 31 c0 ba 01 00 00 00 <f0> 0f b1 13 85 c0 75 08 4c 89 e0 5b 41 5c 5d c3 89 c6 48 89 df e8
[    7.455004] RSP: 0018:ffff9130f53ef988 EFLAGS: 00010046
[    7.455063] RAX: 0000000000000000 RBX: 0000000000000050 RCX: 0000000000000000
[    7.455139] RDX: 0000000000000001 RSI: 0000000000000003 RDI: 0000000000000050
[    7.455216] RBP: ffff9130f53ef998 R08: 0000000000000018 R09: 0000000000000090
[    7.455292] R10: ffffed4cc53cb000 R11: ffffed4cc53cb108 R12: 0000000000000082
[    7.455369] R13: ffff9130cf2c6188 R14: 0000000000000000 R15: 0000000000000018
[    7.455447] FS:  00007f2ff8514cc0(0000) GS:ffff9130fcb00000(0000) knlGS:0000000000000000
[    7.455535] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[    7.455597] CR2: 0000000000000050 CR3: 00000001753f0002 CR4: 00000000000606a0
[    7.455675] Call Trace:
[    7.455713]  __wake_up_common_lock+0x65/0xc0
[    7.455764]  __wake_up+0x13/0x20
[    7.455808]  ir_lirc_unregister+0x57/0xe0 [rc_core]
[    7.455865]  rc_unregister_device+0xa0/0xc0 [rc_core]
[    7.455935]  au0828_rc_unregister+0x25/0x40 [au0828]
[    7.455999]  au0828_usb_disconnect+0x33/0x80 [au0828]
[    7.456064]  au0828_usb_probe.cold.16+0x8d/0x2aa [au0828]
[    7.456130]  usb_probe_interface+0xf1/0x300
[    7.456184]  driver_probe_device+0x2e3/0x460
[    7.456235]  __driver_attach+0xe4/0x110
[    7.456282]  ? driver_probe_device+0x460/0x460
[    7.456335]  bus_for_each_dev+0x74/0xb0
[    7.456385]  ? kmem_cache_alloc_trace+0x15d/0x1d0
[    7.456441]  driver_attach+0x1e/0x20
[    7.456485]  bus_add_driver+0x159/0x230
[    7.456532]  driver_register+0x70/0xc0
[    7.456578]  usb_register_driver+0x7f/0x140
[    7.456626]  ? 0xffffffffc0474000
[    7.456674]  au0828_init+0xbc/0x1000 [au0828]
[    7.456725]  do_one_initcall+0x4a/0x1c9
[    7.456771]  ? _cond_resched+0x19/0x30
[    7.456817]  ? kmem_cache_alloc_trace+0x15d/0x1d0
[    7.456873]  do_init_module+0x60/0x210
[    7.456918]  load_module+0x221b/0x2710
[    7.456966]  ? vfs_read+0xf5/0x120
[    7.457010]  __do_sys_finit_module+0xbd/0x120
[    7.457061]  ? __do_sys_finit_module+0xbd/0x120
[    7.457115]  __x64_sys_finit_module+0x1a/0x20
[    7.457166]  do_syscall_64+0x5b/0x110
[    7.457210]  entry_SYSCALL_64_after_hwframe+0x49/0xbe

Signed-off-by: Brad Love <brad@nextdimension.cc>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: venus: helpers: use true and false for boolean values
Gustavo A. R. Silva [Sun, 5 Aug 2018 00:12:57 +0000 (20:12 -0400)]
media: venus: helpers: use true and false for boolean values

Return statements in functions returning bool should use true or false
instead of an integer value.

This code was detected with the help of Coccinelle.

Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: coda: don't overwrite h.264 profile_idc on decoder instance
Lucas Stach [Wed, 1 Aug 2018 14:18:04 +0000 (10:18 -0400)]
media: coda: don't overwrite h.264 profile_idc on decoder instance

On a decoder instance, after the profile has been parsed from the stream
__v4l2_ctrl_s_ctrl() is called to notify userspace about changes in the
read-only profile control. This ends up calling back into the CODA driver
where a missing check on the s_ctrl caused the profile information that has
just been parsed from the stream to be overwritten with the default
baseline profile.

Later on the driver fails to enable frame reordering, based on the wrong
profile information.

Fixes: 347de126d1da (media: coda: add read-only h.264 decoder
                     profile/level controls)

Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Reviewed-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: vb2: check for sane values from queue_setup
Johan Fjeldtvedt [Mon, 17 Sep 2018 08:36:47 +0000 (04:36 -0400)]
media: vb2: check for sane values from queue_setup

Warn and return error from the reqbufs ioctl when driver sets 0 number
of planes or 0 as plane sizes, as these values don't make any sense.
Checking this here stops obviously wrong values from propagating
further and causing various problems that are hard to trace back to
either of these values being 0.

Signed-off-by: Johan Fjeldtvedt <johfjeld@cisco.com>
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: platform: remove redundant null pointer check before of_node_put
zhong jiang [Sun, 16 Sep 2018 13:16:51 +0000 (09:16 -0400)]
media: platform: remove redundant null pointer check before of_node_put

of_node_put has taken the null pinter check into account. So it is
safe to remove the duplicated check before of_node_put.

Signed-off-by: zhong jiang <zhongjiang@huawei.com>
Reviewed-by: Benoit Parrot <bparrot@ti.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: coda: remove redundant null pointer check before of_node_put
zhong jiang [Sun, 16 Sep 2018 13:04:49 +0000 (09:04 -0400)]
media: coda: remove redundant null pointer check before of_node_put

of_node_put has taken the null pointer check into account. So it is
safe to remove the duplicated check before of_node_put.

Signed-off-by: zhong jiang <zhongjiang@huawei.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: media colorspaces*.rst: rename AdobeRGB to opRGB
Hans Verkuil [Thu, 13 Sep 2018 11:47:28 +0000 (07:47 -0400)]
media: media colorspaces*.rst: rename AdobeRGB to opRGB

Drop all Adobe references and use the official opRGB standard
instead.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Cc: stable@vger.kernel.org
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: hdmi.h: rename ADOBE_RGB to OPRGB and ADOBE_YCC to OPYCC
Hans Verkuil [Thu, 13 Sep 2018 11:47:29 +0000 (07:47 -0400)]
media: hdmi.h: rename ADOBE_RGB to OPRGB and ADOBE_YCC to OPYCC

These names have been renamed in the CTA-861 standard due to trademark
issues. Replace them here as well so they are in sync with the standard.

Signed-off-by: Hans Verkuil <hansverk@cisco.com>
Cc: stable@vger.kernel.org
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: replace ADOBERGB by OPRGB
Hans Verkuil [Fri, 14 Sep 2018 08:58:03 +0000 (04:58 -0400)]
media: replace ADOBERGB by OPRGB

The CTA-861 standards have been updated to refer to opRGB instead
of AdobeRGB. The official standard is in fact named opRGB, so
switch to that.

The two old defines referring to ADOBERGB in the public API are
put under #ifndef __KERNEL__ and a comment mentions that they are
deprecated.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Cc: stable@vger.kernel.org
Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: imx-pxp: fix compilation on i386 or x86_64
Philipp Zabel [Fri, 14 Sep 2018 07:10:56 +0000 (03:10 -0400)]
media: imx-pxp: fix compilation on i386 or x86_64

Include the missing interrupt.h header to fix compilation on i386 or
x86_64:

 ../drivers/media/platform/imx-pxp.c:988:1: error: unknown type name 'irqreturn_t'
  static irqreturn_t pxp_irq_handler(int irq, void *dev_id)
  ^
 ../drivers/media/platform/imx-pxp.c: In function 'pxp_irq_handler':
 ../drivers/media/platform/imx-pxp.c:1012:9: error: 'IRQ_HANDLED' undeclared (first use in this function)
   return IRQ_HANDLED;
          ^
 ../drivers/media/platform/imx-pxp.c:1012:9: note: each undeclared identifier is reported only once for each function it appears in
 ../drivers/media/platform/imx-pxp.c: In function 'pxp_probe':
 ../drivers/media/platform/imx-pxp.c:1660:2: error: implicit declaration of function 'devm_request_threaded_irq' [-Werror=implicit-function-declaration]
   ret = devm_request_threaded_irq(&pdev->dev, irq, NULL, pxp_irq_handler,
   ^
 ../drivers/media/platform/imx-pxp.c:1661:4: error: 'IRQF_ONESHOT' undeclared (first use in this function)
     IRQF_ONESHOT, dev_name(&pdev->dev), dev);

Fixes: 51abcf7fdb70 ("media: imx-pxp: add i.MX Pixel Pipeline driver")

Reported-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Acked-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: vicodec: Drop unused job_abort()
Ezequiel Garcia [Mon, 10 Sep 2018 15:21:54 +0000 (11:21 -0400)]
media: vicodec: Drop unused job_abort()

The vicodec does not use the aborting field. In fact, this driver
can't really cancel any work, since it performs all the work
in device_run().

Signed-off-by: Ezequiel Garcia <ezequiel@collabora.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: vicodec: Drop unneeded symbol dependency
Ezequiel Garcia [Mon, 10 Sep 2018 15:21:53 +0000 (11:21 -0400)]
media: vicodec: Drop unneeded symbol dependency

The vicodec doesn't use the Subdev API, so drop the dependency.

Signed-off-by: Ezequiel Garcia <ezequiel@collabora.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: vicodec: set state->info before calling the encode/decode funcs
Hans Verkuil [Mon, 10 Sep 2018 15:00:40 +0000 (11:00 -0400)]
media: vicodec: set state->info before calling the encode/decode funcs

state->info was NULL since I completely forgot to set state->info.
Oops.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Reported-by: Ezequiel Garcia <ezequiel@collabora.com>
Tested-by: Ezequiel Garcia <ezequiel@collabora.com>
[hans.verkuil@cisco.com: re-add q_out which was removed by commit 703fe34bac]
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: vicodec: check for valid format in v4l2_fwht_en/decode
Hans Verkuil [Mon, 10 Sep 2018 15:00:39 +0000 (11:00 -0400)]
media: vicodec: check for valid format in v4l2_fwht_en/decode

These functions did not return an error if state->info was NULL
or an unsupported pixelformat was selected (should not happen,
but just to be on the safe side).

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: saa7134: add P7131_4871 analog inputs
Alexandre GRIVEAUX [Sun, 2 Sep 2018 12:36:40 +0000 (08:36 -0400)]
media: saa7134: add P7131_4871 analog inputs

The saa7134 Tiger board has a front panel connector at the back (labeled
panel 2 on the PCB), with S-VIDEO, composite and audio.

This patch adds those inputs.

Signed-off-by: Alexandre GRIVEAUX <agriveaux@deutnet.info>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: pci: ivtv: Fix a sleep-in-atomic-context bug in ivtv_yuv_init()
Jia-Ju Bai [Sat, 1 Sep 2018 11:44:09 +0000 (07:44 -0400)]
media: pci: ivtv: Fix a sleep-in-atomic-context bug in ivtv_yuv_init()

The driver may sleep in a interrupt handler.

The function call paths (from bottom to top) in Linux-4.16 are:

[FUNC] kzalloc(GFP_KERNEL)
drivers/media/pci/ivtv/ivtv-yuv.c, 938:
kzalloc in ivtv_yuv_init
drivers/media/pci/ivtv/ivtv-yuv.c, 960:
ivtv_yuv_init in ivtv_yuv_next_free
drivers/media/pci/ivtv/ivtv-yuv.c, 1126:
ivtv_yuv_next_free in ivtv_yuv_setup_stream_frame
drivers/media/pci/ivtv/ivtv-irq.c, 827:
ivtv_yuv_setup_stream_frame in ivtv_irq_dec_data_req
drivers/media/pci/ivtv/ivtv-irq.c, 1013:
ivtv_irq_dec_data_req in ivtv_irq_handler

To fix this bug, GFP_KERNEL is replaced with GFP_ATOMIC.

This bug is found by my static analysis tool DSAC.

Signed-off-by: Jia-Ju Bai <baijiaju1990@gmail.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: imx: work around false-positive warning, again
Arnd Bergmann [Mon, 13 Aug 2018 21:32:17 +0000 (17:32 -0400)]
media: imx: work around false-positive warning, again

A warning that I thought to be solved by a previous patch of mine
has resurfaced with gcc-8:

media/imx/imx-media-csi.c: In function 'csi_link_validate':
media/imx/imx-media-csi.c:1025:20: error: 'upstream_ep' may be used uninitialized in this function [-Werror=maybe-uninitialized]
media/imx/imx-media-csi.c:1026:24: error: 'upstream_ep.bus_type' may be used uninitialized in this function [-Werror=maybe-uninitialized]
media/imx/imx-media-csi.c:127:19: error: 'upstream_ep.bus.parallel.bus_width' may be used uninitialized in this function [-Werror=maybe-uninitialized]
media/imx/imx-media-csi.c: In function 'csi_enum_mbus_code':
media/imx/imx-media-csi.c:132:9: error: '*((void *)&upstream_ep+12)' may be used uninitialized in this function [-Werror=maybe-uninitialized]
media/imx/imx-media-csi.c:132:48: error: 'upstream_ep.bus.parallel.bus_width' may be used uninitialized in this function [-Werror=maybe-uninitialized]

I spent some more time digging in this time, and think I have a better
fix, bailing out of the function that either initializes or errors
out here, which simplifies the code enough for gcc to figure out
what is going on. The earlier partial workaround can be removed now,
as the new workaround is better.

Fixes: 890f27693f2a ("media: imx: work around false-positive warning")

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: cec: remove cec-edid.c
Hans Verkuil [Thu, 13 Sep 2018 12:00:39 +0000 (08:00 -0400)]
media: cec: remove cec-edid.c

Move cec_get_edid_phys_addr() to cec-adap.c. It's not worth keeping
a separate source for this.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Cc: <stable@vger.kernel.org> # for v4.17 and up
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: cec/v4l2: move V4L2 specific CEC functions to V4L2
Hans Verkuil [Thu, 13 Sep 2018 07:40:56 +0000 (03:40 -0400)]
media: cec/v4l2: move V4L2 specific CEC functions to V4L2

Several CEC functions are actually specific for use with receivers,
i.e. they should be part of the V4L2 subsystem, not CEC.

These functions deal with validating and modifying EDIDs for (HDMI)
receivers, and they do not actually have anything to do with the CEC
subsystem and whether or not CEC is enabled. The problem was that if
the CEC_CORE config option was not set, then these functions would
become stubs, but that's not right: they should always be valid.

So replace the cec_ prefix by v4l2_ and move them to v4l2-dv-timings.c.
Update all drivers that call these accordingly.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Reported-by: Lars-Peter Clausen <lars@metafoo.de>
Cc: <stable@vger.kernel.org> # for v4.17 and up
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: cec: integrate cec_validate_phys_addr() in cec-api.c
Hans Verkuil [Thu, 13 Sep 2018 07:36:29 +0000 (03:36 -0400)]
media: cec: integrate cec_validate_phys_addr() in cec-api.c

The cec_phys_addr_validate() function will be moved to V4L2,
so use a simplified variant of that function in cec-api.c.
cec now no longer calls cec_phys_addr_validate() and it can
be safely moved to V4L2.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Cc: <stable@vger.kernel.org> # for v4.17 and up
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: cec: make cec_get_edid_spa_location() an inline function
Hans Verkuil [Thu, 13 Sep 2018 07:25:59 +0000 (03:25 -0400)]
media: cec: make cec_get_edid_spa_location() an inline function

This function is needed by both V4L2 and CEC, so move this to
cec.h as a static inline since there are no obvious shared
modules between the two subsystems.

This patch, together with the following ones, fixes a
dependency bug: if CEC_CORE is disabled, then building adv7604
(and other HDMI receivers) will fail because an essential
function is now stubbed out.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Cc: <stable@vger.kernel.org> # for v4.17 and up
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: ov5640: fix restore of last mode set
Hugues Fruchet [Tue, 11 Sep 2018 13:48:21 +0000 (09:48 -0400)]
media: ov5640: fix restore of last mode set

Mode setting depends on last mode set, in particular
because of exposure calculation when downscale mode
change between subsampling and scaling.
At stream on the last mode was wrongly set to current mode,
so no change was detected and exposure calculation
was not made, fix this.

Signed-off-by: Hugues Fruchet <hugues.fruchet@st.com>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
Tested-by: Jacopo Mondi <jacopo@jmondi.org>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: ov5640: fix auto controls values when switching to manual mode
Hugues Fruchet [Tue, 11 Sep 2018 13:48:20 +0000 (09:48 -0400)]
media: ov5640: fix auto controls values when switching to manual mode

When switching from auto to manual mode, V4L2 core is calling
g_volatile_ctrl() in manual mode in order to get the manual initial value.
Remove the manual mode check/return to not break this behaviour.

Signed-off-by: Hugues Fruchet <hugues.fruchet@st.com>
Tested-by: Jacopo Mondi <jacopo@jmondi.org>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: ov5640: fix wrong binning value in exposure calculation
Hugues Fruchet [Tue, 11 Sep 2018 13:48:19 +0000 (09:48 -0400)]
media: ov5640: fix wrong binning value in exposure calculation

ov5640_set_mode_exposure_calc() is checking binning value but
binning value read is buggy, fix this.
Rename ov5640_binning_on() to ov5640_get_binning() as per other
similar functions.

Signed-off-by: Hugues Fruchet <hugues.fruchet@st.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: ov5640: fix auto gain & exposure when changing mode
Hugues Fruchet [Tue, 11 Sep 2018 13:48:18 +0000 (09:48 -0400)]
media: ov5640: fix auto gain & exposure when changing mode

Ensure that auto gain and auto exposure are well restored
when changing mode.

Signed-off-by: Hugues Fruchet <hugues.fruchet@st.com>
Reviewed-by: Jacopo Mondi <jacopo@jmondi.org>
Tested-by: Jacopo Mondi <jacopo@jmondi.org>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: ov5640: fix exposure regression
Hugues Fruchet [Tue, 11 Sep 2018 13:48:17 +0000 (09:48 -0400)]
media: ov5640: fix exposure regression

Symptom was black image when capturing HD or 5Mp picture
due to manual exposure set to 1 while it was intended to
set autoexposure to "manual", fix this.

Fixes: bf4a4b518c20 ("media: ov5640: Don't force the auto exposure state at start time").

Signed-off-by: Hugues Fruchet <hugues.fruchet@st.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Tested-by: Jacopo Mondi <jacopo@jmondi.org>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: ov5640: Fix timings setup code
Jacopo Mondi [Wed, 18 Jul 2018 10:06:23 +0000 (06:06 -0400)]
media: ov5640: Fix timings setup code

As of: commit 476dec012f4c ("media: ov5640: Add horizontal and vertical
totals") the timings parameters gets programmed separately from the
static register values array.

When changing capture mode, the vertical and horizontal totals gets
inspected by the set_mode_exposure_calc() functions, and only later
programmed with the new values. This means exposure, light banding
filter and shutter gain are calculated using the previous timings, and
are thus not correct.

Fix this by programming timings right after the static register value
table has been sent to the sensor in the ov5640_load_regs() function.

Fixes: 476dec012f4c ("media: ov5640: Add horizontal and vertical totals")

Tested-by: Steve Longerbeam <slongerbeam@gmail.com> # i.MX6q SabreSD, CSI-2
Tested-by: Loic Poulain <loic.poulain@linaro.org> # Dragonboard-410c, CSI-2
Signed-off-by: Samuel Bobrowicz <sam@elite-embedded.com>
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.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+samsung@kernel.org>
6 years agomedia: ov5640: Re-work MIPI startup sequence
Jacopo Mondi [Fri, 6 Jul 2018 09:51:52 +0000 (05:51 -0400)]
media: ov5640: Re-work MIPI startup sequence

Rework the MIPI interface startup sequence with the following changes:

- Remove MIPI bus initialization from the initial settings blob
- At set_power(1) time power up MIPI Tx/Rx and set data and clock lanes in
  LP11 during 'sleep' and 'idle' with MIPI clock in non-continuous mode.
- At s_stream time enable/disable the MIPI interface output.
- Restore default settings at set_power(0) time.

Before this commit the sensor MIPI interface was initialized with settings
that require a start/stop sequence at power-up time in order to force lanes
into LP11 state, as they were initialized in LP00 when in 'sleep mode',
which is assumed to be the sensor manual definition for the D-PHY defined
stop mode.

The stream start/stop was performed by enabling disabling clock gating,
and had the side effect to change the lanes sleep mode configuration when
stream was stopped.

Clock gating/ungating:
-       ret = ov5640_mod_reg(sensor, OV5640_REG_MIPI_CTRL00, BIT(5),
-                            on ? 0 : BIT(5));
-       if (ret)

Set lanes in LP11 when in 'sleep mode':
-       ret = ov5640_write_reg(sensor, OV5640_REG_PAD_OUTPUT00,
-                              on ? 0x00 : 0x70);

This commit fixes an issue reported by Jagan Teki on i.MX6 platforms that
prevents the host interface from powering up correctly:
https://lkml.org/lkml/2018/6/1/38

It also improves MIPI capture operations stability on my testing platform
where MIPI capture often failed and returned all-purple frames.

Fixes: f22996db44e2 ("media: ov5640: add support of DVP parallel interface")

Tested-by: Steve Longerbeam <slongerbeam@gmail.com> (i.MX6q SabreSD, CSI-2)
Tested-by: Loic Poulain <loic.poulain@linaro.org> (Dragonboard-410c, CSI-2)
Reported-by: Jagan Teki <jagan@amarulasolutions.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+samsung@kernel.org>
6 years agomedia: ov2680: rename ov2680_v4l2_init() to ov2680_v4l2_register()
Javier Martinez Canillas [Sat, 1 Sep 2018 12:46:30 +0000 (08:46 -0400)]
media: ov2680: rename ov2680_v4l2_init() to ov2680_v4l2_register()

The function not only does initialization but also registers the subdevice
so change its name to make this more clear.

Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: ov2680: don't register the v4l2 subdevice before checking chip ID
Javier Martinez Canillas [Sat, 1 Sep 2018 12:46:29 +0000 (08:46 -0400)]
media: ov2680: don't register the v4l2 subdevice before checking chip ID

The driver registers the v4l2 subdevice before attempting to power on the
chip and checking its ID. This means that a media device driver that it's
waiting for this subdevice to be bound, will prematurely expose its media
device node to userspace because if something goes wrong the media entity
will be cleaned up again on the ov2680 probe function.

This also simplifies the probe function error path since no initialization
is made before attempting to enable the resources or checking the chip ID.

Fixes: 3ee47cad3e69 ("media: ov2680: Add Omnivision OV2680 sensor driver")

Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: sr030pc30: remove NULL in sr030pc30_base_config()
Dan Carpenter [Fri, 31 Aug 2018 13:20:18 +0000 (09:20 -0400)]
media: sr030pc30: remove NULL in sr030pc30_base_config()

This code doesn't check for NULL consistently and it generates a Smatch
warning:

    drivers/media/i2c/sr030pc30.c:575 sr030pc30_base_config()
    error: we previously assumed 'info->pdata' could be null (see line 572)

Fortunately, "info->pdata" can't be NULL to that check can be removed.
The other thing is that if "ret" is an error code here, then we don't
want to do the next call to cam_i2c_write(), so actually let's flip that
test around and return the error.  This is more of a theoretical issue
than something which is likely to affect real life.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: v4l: sr030pc30: Remove redundant setting of sub-device name
Sakari Ailus [Wed, 29 Aug 2018 10:01:48 +0000 (06:01 -0400)]
media: v4l: sr030pc30: Remove redundant setting of sub-device name

The sub-device name is set right after in v4l2_i2c_subdev_init(). Remove
the redundant strcpy() call.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Acked-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: smiapp: Use v4l2_i2c_subdev_set_name
Sakari Ailus [Wed, 29 Aug 2018 09:57:36 +0000 (05:57 -0400)]
media: smiapp: Use v4l2_i2c_subdev_set_name

Use v4l2_i2c_subdev_set_name() to set the name of the smiapp driver's
sub-devices. There is no functional change.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Acked-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: v4l: subdev: Add a function to set an I²C sub-device's name
Sakari Ailus [Wed, 29 Aug 2018 09:52:46 +0000 (05:52 -0400)]
media: v4l: subdev: Add a function to set an I²C sub-device's name

v4l2_i2c_subdev_set_name() can be used to assign a name to a sub-device.
This way uniform names can be formed easily without having to resort to
things such as snprintf in drivers.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Acked-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: imx274: switch to SPDX license identifier
Luca Ceresoli [Fri, 24 Aug 2018 16:35:25 +0000 (12:35 -0400)]
media: imx274: switch to SPDX license identifier

Signed-off-by: Luca Ceresoli <luca@lucaceresoli.net>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: imx274: add helper to read multibyte registers
Luca Ceresoli [Fri, 24 Aug 2018 16:35:24 +0000 (12:35 -0400)]
media: imx274: add helper to read multibyte registers

Currently 2-bytes and 3-bytes registers are read one byte at a time,
doing the needed shift & mask each time.

Replace all of this code by a unique helper function that calls
regmap_bulk_read(), which has two advantages:
 - reads all the bytes in a unique I2C transaction
 - simplifies code to read multibyte registers

Signed-off-by: Luca Ceresoli <luca@lucaceresoli.net>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
6 years agomedia: imx274: fix error in function docs
Luca Ceresoli [Fri, 24 Aug 2018 16:35:23 +0000 (12:35 -0400)]
media: imx274: fix error in function docs

This parameter holds the number of bytes, not bits.

Signed-off-by: Luca Ceresoli <luca@lucaceresoli.net>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>