platform/kernel/linux-rpi.git
18 months agodt-bindings: media: gpio-ir-receiver: Document wakeup-souce property
Florian Fainelli [Fri, 24 Mar 2023 20:38:32 +0000 (13:38 -0700)]
dt-bindings: media: gpio-ir-receiver: Document wakeup-souce property

The GPIO IR receiver can be used as a wake-up source for the system,
document that.

Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Acked-by: Krzysztof Kozlowski <krzk@kernel.org>
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: i2c: imx290: Add missing \n on dev_err_probe() message
Alexander Stein [Thu, 6 Apr 2023 09:55:51 +0000 (11:55 +0200)]
media: i2c: imx290: Add missing \n on dev_err_probe() message

Also dev_err_probe message require a trailing \n.

Signed-off-by: Alexander Stein <alexander.stein@ew.tq-group.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: ccs: Document CCS static data file names
Sakari Ailus [Tue, 28 Mar 2023 11:18:59 +0000 (14:18 +0300)]
media: ccs: Document CCS static data file names

Document the firmware file names for CCS static data for CCS, SMIA++ and
SMIA devices.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: ccs: Differentiate SMIA and MIPI vendors in static data
Sakari Ailus [Fri, 13 May 2022 13:12:54 +0000 (16:12 +0300)]
media: ccs: Differentiate SMIA and MIPI vendors in static data

MIPI CCS uses a 16-bit MIPI manufacturer identifier for firmware filenames
compared to older 8-bit identifier used by SMIA. The requested firmware
files used the MIPI manufacturer ID even for SMIA compliant devices,
effectively making the manufacturer ID 0.

While CCS static data is a feature of CCS 1.1, it has no hardware
dependencies. Making this feature available for SMIA devices helps adding
support for them and avoids requesting ill-generated CCS static data file
names. The files are named as:

ccs/smiapp-module-vv-mmmm-rrrr.fw,
ccs/smiapp-sensor-vv-mmmm-rr.fw and
ccs/smia-sensor-vv-mmmm-rr.fw

where vv is the manufacturer ID, mmmm is the model ID and rr or rrrr is
the revision number.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: ccs: Apply module manufacturer hack on non-CCS devices only
Sakari Ailus [Thu, 6 Apr 2023 16:41:03 +0000 (19:41 +0300)]
media: ccs: Apply module manufacturer hack on non-CCS devices only

Some modules don't have any model identification information in the
register address space. The driver as a SMIA++ driver attempted to use
sensor information in this case. This workaround is definitely not for CCS
devices.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: ccs: Support 16-bit sensor revision number register
Sakari Ailus [Tue, 28 Mar 2023 11:11:00 +0000 (14:11 +0300)]
media: ccs: Support 16-bit sensor revision number register

Read 16-bit sensor revision number if the 8-bit register has value 0.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: ccs: Add V4L2 controls from properties
Sakari Ailus [Tue, 28 Mar 2023 11:39:26 +0000 (14:39 +0300)]
media: ccs: Add V4L2 controls from properties

Add V4L2 controls (currently CAMERA_SENSOR_ROTATION and
CAMERA_SENSOR_ORIENTATION) from properties.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: ccs: Align flipping behaviour with other drivers
Sakari Ailus [Tue, 28 Mar 2023 11:34:42 +0000 (14:34 +0300)]
media: ccs: Align flipping behaviour with other drivers

No longer mirror flipping controls if the sensor is mounted upside down.

This way the behaviour of the flipping controls and rotation of the sensor
are aligned with the rest of the drivers.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: imx258: Remove mandatory 180 degrees rotation
Jacopo Mondi [Tue, 17 Jan 2023 10:06:03 +0000 (11:06 +0100)]
media: imx258: Remove mandatory 180 degrees rotation

The "rotation" fwnode device property is intended to allow specify the
sensor's physical mounting rotation, so that it can be exposed through
the read-only V4L2_CID_CAMERA_SENSOR_ROTATION control and applications
can decide how to compensate for that.

The imx258 driver has read-only VFLIP and HFLIP enabled, resulting in
a 180 degrees image rotation being produced by the sensor. But this
doesn't imply that the physical mounting rotation should match the
driver's implementation.

Now that the driver registers V4L2_CID_CAMERA_SENSOR_ROTATION
and V4L2_CID_HFLIP/VFLIP correctly, userspace has all the required
information to handle the rotation correctly, hence it is not necessary
to require the 'rotation' property to be fixed to 180 degrees in DTS.

Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: imx258: Register H/V flip controls
Jacopo Mondi [Tue, 17 Jan 2023 10:06:02 +0000 (11:06 +0100)]
media: imx258: Register H/V flip controls

Register controls for V4L2_CID_HFLIP and V4L2_CID_VFLIP.

The controls are registered as read-only and enabled by default, as the
driver embeds a 180 degrees rotation in its programming sequences and
only supports that mode of operations.

Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: dt-bindings: ov8856: Assign maintainership to myself
Sakari Ailus [Fri, 31 Mar 2023 09:17:45 +0000 (12:17 +0300)]
media: dt-bindings: ov8856: Assign maintainership to myself

Dongchun's e-mail mailbox is no longer available, assign the
maintainership to myself.

The following message to <dongchun.zhu@mediatek.com> was
undeliverable. The reason for the problem: 5.1.0 - Unknown address
error 550-'Relaying mail to dongchun.zhu@mediatek.com is not
allowed'

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Acked-by: Rob Herring <robh@kernel.org>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agoMAINTAINERS: Assume ov8856 driver maintainership
Sakari Ailus [Fri, 31 Mar 2023 08:57:53 +0000 (11:57 +0300)]
MAINTAINERS: Assume ov8856 driver maintainership

Dongchun's e-mail address is no longer valid, assign the ov8856 driver to
myself.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: ipu3-cio2: support more camera sensors in cio2-bridge
Bingbu Cao [Fri, 3 Mar 2023 18:15:03 +0000 (02:15 +0800)]
media: ipu3-cio2: support more camera sensors in cio2-bridge

Add more camera sensors into the supported camera sensors list
to make cio2-bridge to support more camera sensors.

Signed-off-by: Bingbu Cao <bingbu.cao@intel.com>
Reviewed-by: Daniel Scally <dan.scally@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: hi556: add 2592x1444 resolution
Jim Lai [Wed, 26 Oct 2022 01:35:45 +0000 (09:35 +0800)]
media: hi556: add 2592x1444 resolution

Adding additional cropped 2592x1444 resolution
for QHD output from Hi556 sensor

Also implement the get_selection pad operation for the Hi556 sensor driver.
The supported targets report the sensor's native size, the crop default
rectangle and the crop rectangle.

Signed-off-by: Jim Lai <jim.lai@intel.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: venus: drop unused opp_table field documentation
Krzysztof Kozlowski [Tue, 27 Dec 2022 14:35:09 +0000 (15:35 +0100)]
media: venus: drop unused opp_table field documentation

The struct venus_format does not have a "opp_table" field.

Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: renesas: fdp1: remove R-Car H3 ES1.* handling
Wolfram Sang [Tue, 7 Mar 2023 16:30:33 +0000 (17:30 +0100)]
media: renesas: fdp1: remove R-Car H3 ES1.* handling

R-Car H3 ES1.* was only available to an internal development group and
needed a lot of quirks and workarounds. These become a maintenance
burden now, so our development group decided to remove upstream support
and disable booting for this SoC. Public users only have ES2 onwards.

Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Reviewed-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: rcar-vin: csi2: remove R-Car H3 ES1.* handling
Wolfram Sang [Tue, 7 Mar 2023 16:30:32 +0000 (17:30 +0100)]
media: rcar-vin: csi2: remove R-Car H3 ES1.* handling

R-Car H3 ES1.* was only available to an internal development group and
needed a lot of quirks and workarounds. These become a maintenance
burden now, so our development group decided to remove upstream support
and disable booting for this SoC. Public users only have ES2 onwards.

Reviewed-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: rcar-vin: remove R-Car H3 ES1.* handling
Wolfram Sang [Tue, 7 Mar 2023 16:30:31 +0000 (17:30 +0100)]
media: rcar-vin: remove R-Car H3 ES1.* handling

R-Car H3 ES1.* was only available to an internal development group and
needed a lot of quirks and workarounds. These become a maintenance
burden now, so our development group decided to remove upstream support
and disable booting for this SoC. Public users only have ES2 onwards.

Reviewed-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: Use of_property_present() for testing DT property presence
Rob Herring [Fri, 10 Mar 2023 14:47:11 +0000 (08:47 -0600)]
media: Use of_property_present() for testing DT property presence

It is preferred to use typed property access functions (i.e.
of_property_read_<type> functions) rather than low-level
of_get_property/of_find_property functions for reading properties. As
part of this, convert of_get_property/of_find_property calls to the
recently added of_property_present() helper when we just want to test
for presence of a property and nothing more.

Signed-off-by: Rob Herring <robh@kernel.org>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: ipu3-cio2: support multiple sensors and VCMs with same HID
Bingbu Cao [Fri, 10 Mar 2023 15:19:10 +0000 (23:19 +0800)]
media: ipu3-cio2: support multiple sensors and VCMs with same HID

In current cio2-bridge, it is using the hid name to register software
node and software node will create kobject and sysfs entry according to
the node name, if there are multiple sensors and VCMs which are sharing
same HID name, it will cause the software nodes registration failure:

sysfs: cannot create duplicate filename '/kernel/software_nodes/dw9714'
...
Call Trace:
software_node_register_nodes
cio2_bridge_init
...
kobject_add_internal failed for dw9714 with -EEXIST,
don't try to register things with the same name in the same directory.

One solution is appending the sensor link(Mipi Port) in SSDB as suffix
of the node name to fix this problem.

Signed-off-by: Bingbu Cao <bingbu.cao@intel.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: Accept non-subdev sinks in v4l2_create_fwnode_links_to_pad()
Laurent Pinchart [Fri, 24 Mar 2023 10:35:29 +0000 (12:35 +0200)]
media: Accept non-subdev sinks in v4l2_create_fwnode_links_to_pad()

The v4l2_create_fwnode_links_to_pad() helper requires the sink pad
passed to it to belong to a subdev. This requirement can be lifted
easily. Make the function usable for non-subdev sinks, which allows
using it with video_device sinks.

Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: i2c: adv7604: Fix range of hue control
Laurent Pinchart [Sat, 25 Mar 2023 20:56:10 +0000 (22:56 +0200)]
media: i2c: adv7604: Fix range of hue control

The ADV7604, ADV7611 and ADV7612 software manuals document the CP_HUE
value differently:

- For ADV7604 and ADV7611, the hue is specified as an unsigned 8-bit
  value, and calculated as

  (CP_HUE[7:0] * 180) / 256 - 90

  with the range set to [-90°, 90°]. Additionally, the values 0x00, 0x0f
  and 0xff are documented as corresponding to -90°, 0° and 90°
  respectively.

- For ADV7612, the hue is specified as a signed 8-bit value in the range
  [0°, 360°[ without any formula. Additionally, the value 0x00 is
  documented as corresponding to 0°.

The ADV7604 and ADV7611 documentation is clearly wrong as the example
values don't correspond to the formula. Furthermore, the [-90°, 90°]
range seems incorrect as it would cover only half of the hue space.

The ADV7612 documentation is better, although the range should likely be
[-180°, 180°[ if the value is signed. Given that the values wrap around,
this makes no difference in practice.

The hue values have been verified on ADV7612 to follow the
documentation. There is a high chance they do as well on ADV7604 and
ADV7611.

In any case, all software manuals document the register as 8-bit, so the
current range of the V4L2 hue control [0, 128] is not correct. Expand it
to cover the full 8-bit space, using unsigned values to avoid breaking
any application that may rely on 128 being a valid value.

Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Tested-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: i2c: adv7604: Enable video adjustment
Laurent Pinchart [Sat, 25 Mar 2023 20:56:09 +0000 (22:56 +0200)]
media: i2c: adv7604: Enable video adjustment

The video adjustments (contrast, brightness, saturation and hue) are
ignored by default by the device when the VID_ADJ_EN bit is clear. The
corresponding V4L2 controls exposed by the drivers have thus no effect.
Fix this by setting the VID_ADJ_EN bit.

Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Tested-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: i2c: ov2685: Make reset gpio optional
Luca Weiss [Thu, 23 Mar 2023 17:57:51 +0000 (18:57 +0100)]
media: i2c: ov2685: Make reset gpio optional

In some setups XSHUTDOWN is connected to DOVDD when it's unused,
therefore treat the reset gpio as optional.

Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Signed-off-by: Luca Weiss <luca@z3ntu.xyz>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: dt-bindings: ov2685: convert to dtschema
Luca Weiss [Thu, 23 Mar 2023 17:57:50 +0000 (18:57 +0100)]
media: dt-bindings: ov2685: convert to dtschema

Convert the text-based dt-bindings to yaml.

Changes from original txt:
* Take wording for various properties from other yaml bindings, this
  removes e.g. volt amount from schema since it isn't really relevant
  and the datasheet is a better source.
* Don't make reset-gpios a required property since it can be tied to
  DOVDD instead.

Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Luca Weiss <luca@z3ntu.xyz>
[Sakari Ailus: bump the maxItems for data-lanes to two]
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: dt-bindings: media: rkisp1: complete ov2685 example
Luca Weiss [Thu, 23 Mar 2023 17:57:49 +0000 (18:57 +0100)]
media: dt-bindings: media: rkisp1: complete ov2685 example

With the upcoming conversion of ov2685 to dt-schema let's complete the
example so validation succeeds. At the same time fix the indentation of
the port node in this example.

Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Luca Weiss <luca@z3ntu.xyz>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: mc-device: remove unnecessary __must_check
Jason Kim [Tue, 28 Mar 2023 03:14:59 +0000 (12:14 +0900)]
media: mc-device: remove unnecessary __must_check

In the file mc-device.c, the function
media_device_register_entity_notify() does not need to have the
__must_check attribute since it returns only a value of 0.
Therefore, we can remove this attribute and change the function's
return type.

Signed-off-by: Jason Kim <sukbeom.kim@gmail.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: hi846: Fix memleak in hi846_init_controls()
Wei Chen [Mon, 27 Mar 2023 11:58:09 +0000 (11:58 +0000)]
media: hi846: Fix memleak in hi846_init_controls()

hi846_init_controls doesn't clean the allocated ctrl_hdlr
in case there is a failure, which causes memleak. Add
v4l2_ctrl_handler_free to free the resource properly.

Fixes: e8c0882685f9 ("media: i2c: add driver for the SK Hynix Hi-846 8M pixel camera")
Signed-off-by: Wei Chen <harperchen1110@gmail.com>
Reviewed-by: Martin Kepplinger <martin.kepplinger@puri.sm>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: ov8856: Do not check for for module version
Ricardo Ribalda [Thu, 23 Mar 2023 22:44:20 +0000 (23:44 +0100)]
media: ov8856: Do not check for for module version

It the device is probed in non-zero ACPI D state, the module
identification is delayed until the first streamon.

The module identification has two parts: deviceID and version. To rea
the version we have to enable OTP read. This cannot be done during
streamon, becase it modifies REG_MODE_SELECT.

Since the driver has the same behaviour for all the module versions, do
not read the module version from the sensor's OTP.

Cc: stable@vger.kernel.org
Fixes: 0e014f1a8d54 ("media: ov8856: support device probe in non-zero ACPI D state")
Signed-off-by: Ricardo Ribalda <ribalda@chromium.org>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: dt-bindings: Drop unneeded quotes
Rob Herring [Mon, 20 Mar 2023 23:39:42 +0000 (18:39 -0500)]
media: dt-bindings: Drop unneeded quotes

Cleanup bindings dropping unneeded quotes. Once all these are fixed,
checking for this can be enabled in yamllint.

Signed-off-by: Rob Herring <robh@kernel.org>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Reviewed-by: Joe Tessler <jrt@google.com>
Reviewed-by: Heiko Stuebner <heiko@sntech.de> #rockchip
Acked-by: Jernej Skrabec <jernej.skrabec@gmail.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: Philipp Zabel <p.zabel@pengutronix.de>
Reviewed-by: Neil Armstrong <neil.armstrong@linaro.org>
[Sakari Ailus: Drop changes for files in which they had been done already]
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: i2c: ov7670: Use the devm_clk_get_optional() helper
Christophe JAILLET [Sun, 26 Mar 2023 14:21:16 +0000 (16:21 +0200)]
media: i2c: ov7670: Use the devm_clk_get_optional() helper

Use devm_clk_get_optional() instead of hand writing it.
This saves some loC and improves the semantic.

Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: v4l: async: Return async sub-devices to subnotifier list
Sakari Ailus [Tue, 7 Mar 2023 13:49:44 +0000 (15:49 +0200)]
media: v4l: async: Return async sub-devices to subnotifier list

When an async notifier is unregistered, the async sub-devices in the
notifier's done list will disappear with the notifier. However this is
currently also done to the sub-notifiers that remain registered. Their
sub-devices only need to be unbound while the async sub-devices themselves
need to be returned to the sub-notifier's waiting list. Do this now.

Fixes: 2cab00bb076b ("media: v4l: async: Allow binding notifiers to sub-devices")
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agostaging: media: imx: Make imx_media_of_add_csi() static
Laurent Pinchart [Thu, 7 Jul 2022 00:26:40 +0000 (03:26 +0300)]
staging: media: imx: Make imx_media_of_add_csi() static

The imx_media_of_add_csi() function is only called in its compilation
unit. Make it static and don't export it.

Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agostaging: media: imx: Drop unused helper functions
Laurent Pinchart [Tue, 10 May 2022 11:27:40 +0000 (14:27 +0300)]
staging: media: imx: Drop unused helper functions

The imx_media_find_subdev_by_fwnode(),
imx_media_find_subdev_by_devname(), imx_media_pipeline_csi2_channel()
and imx_media_pipeline_video_device() helper functions are not used.
Drop them.

Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agostaging: media: imx: Drop IMX_MEDIA_GRP_ID_CSI
Laurent Pinchart [Sun, 6 Feb 2022 06:07:33 +0000 (08:07 +0200)]
staging: media: imx: Drop IMX_MEDIA_GRP_ID_CSI

The IMX_MEDIA_GRP_ID_CSI group ID isn't used anymore. Drop it.

Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agostaging: media: imx: Drop imx_media_subdev_bound()
Laurent Pinchart [Sun, 6 Feb 2022 05:42:20 +0000 (07:42 +0200)]
staging: media: imx: Drop imx_media_subdev_bound()

The imx_media_subdev_bound() function is a no-op, and the async notifier
.bound() operation is optional. Drop it.

Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: imx: imx8mq-mipi-csi2: Remove unneeded state variable and function
Martin Kepplinger [Tue, 7 Mar 2023 15:00:47 +0000 (16:00 +0100)]
media: imx: imx8mq-mipi-csi2: Remove unneeded state variable and function

Clean up the driver a bit by inlining the imx8mq_mipi_csi_system_enable()
function to the callsites and removing the hs_settle variable from the
driver state.

Signed-off-by: Martin Kepplinger <martin.kepplinger@puri.sm>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: imx: imx8mq-mipi-csi2: Use V4L2 subdev active state
Martin Kepplinger [Tue, 7 Mar 2023 15:00:46 +0000 (16:00 +0100)]
media: imx: imx8mq-mipi-csi2: Use V4L2 subdev active state

Simplify the driver by using the V4L2 subdev active state API to store
the active format.

Signed-off-by: Martin Kepplinger <martin.kepplinger@puri.sm>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: imx: imx7-media-csi: Fail on invalid type in VIDIOC_G_SELECTION
Alexander Stein [Tue, 21 Mar 2023 07:27:07 +0000 (08:27 +0100)]
media: imx: imx7-media-csi: Fail on invalid type in VIDIOC_G_SELECTION

This device only supports video capture, so bail out if invalid
selection type is passed.

Signed-off-by: Alexander Stein <alexander.stein@ew.tq-group.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Rui Miguel Silva <rmfrfs@gmail.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: imx: imx7-media-csi: Fix mbus framefmt field init
Alexander Stein [Tue, 21 Mar 2023 07:27:06 +0000 (08:27 +0100)]
media: imx: imx7-media-csi: Fix mbus framefmt field init

'field' is zero-initialized to V4L2_FIELD_ANY, which is an invalid value
to return to userspace. Instead default to non-interleaving.

Signed-off-by: Alexander Stein <alexander.stein@ew.tq-group.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Rui Miguel Silva <rmfrfs@gmail.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: imx: imx7-media-csi: Fix error handling in imx7_csi_async_register()
Frieder Schrempf [Wed, 15 Feb 2023 17:16:38 +0000 (18:16 +0100)]
media: imx: imx7-media-csi: Fix error handling in imx7_csi_async_register()

The CSI requires a connected source subdev to operate. If
fwnode_graph_get_endpoint_by_id() fails and returns NULL, there is no
point in going on. Print an error message and abort instead.

Also we don't need to check for an existing asd. Any failure of
v4l2_async_nf_add_fwnode_remote() should abort the probe.

Suggested-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Frieder Schrempf <frieder.schrempf@kontron.de>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: rkisp1: Implement ENUM_FRAMESIZES
Paul Elder [Wed, 20 Jul 2022 17:22:31 +0000 (02:22 +0900)]
media: rkisp1: Implement ENUM_FRAMESIZES

Implement VIDIOC_ENUM_FRAMESIZES for the rkisp1 capture devices.

Signed-off-by: Paul Elder <paul.elder@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Dafna Hirschfeld <dafna@fastmail.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: rkisp1: Add NV16M and NV61M to output formats
Paul Elder [Wed, 13 Jul 2022 07:24:58 +0000 (16:24 +0900)]
media: rkisp1: Add NV16M and NV61M to output formats

Add support for NV16M and NV61M as output formats. As NV16, NV61, NV12M
and NV21M are already supported, the infrastructure is already in place
to support NV16M and NV61M, so it is sufficient to simply add relevant
entries to the list of output formats.

Signed-off-by: Paul Elder <paul.elder@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Dafna Hirschfeld <dafna@fastmail.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: rcar_fdp1: Fix refcount leak in probe and remove function
Miaoqian Lin [Fri, 6 Jan 2023 07:58:09 +0000 (11:58 +0400)]
media: rcar_fdp1: Fix refcount leak in probe and remove function

rcar_fcp_get() take reference, which should be balanced with
rcar_fcp_put(). Add missing rcar_fcp_put() in fdp1_remove and
the error paths of fdp1_probe() to fix this.

Fixes: 4710b752e029 ("[media] v4l: Add Renesas R-Car FDP1 Driver")
Signed-off-by: Miaoqian Lin <linmq006@gmail.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
[hverkuil: resolve merge conflict, remove() is now void]
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: vsp1: Add underrun debug print
Tomi Valkeinen [Thu, 16 Feb 2023 09:41:15 +0000 (11:41 +0200)]
media: vsp1: Add underrun debug print

Print underrun interrupts with ratelimited print.

Note that we don't enable the underrun interrupt. If we have underruns,
we don't want to get flooded with interrupts about them. It's enough to
see that an underrun happened at the end of a frame.

Signed-off-by: Tomi Valkeinen <tomi.valkeinen+renesas@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: vsp1: Replace vb2_is_streaming() with vb2_start_streaming_called()
Laurent Pinchart [Fri, 20 Jan 2023 20:20:31 +0000 (22:20 +0200)]
media: vsp1: Replace vb2_is_streaming() with vb2_start_streaming_called()

The vsp1 driver uses the vb2_is_streaming() function in its .buf_queue()
handler to check if the .start_streaming() operation has been called,
and decide whether to just add the buffer to an internal queue, or also
trigger a hardware run. vb2_is_streaming() relies on the vb2_queue
structure's streaming field, which used to be set only after calling the
.start_streaming() operation.

Commit a10b21532574 ("media: vb2: add (un)prepare_streaming queue ops")
changed this, setting the .streaming field in vb2_core_streamon() before
enqueuing buffers to the driver and calling .start_streaming(). This
broke the vsp1 driver which now believes that .start_streaming() has
been called when it hasn't, leading to a crash:

[  881.058705] Unable to handle kernel NULL pointer dereference at virtual address 0000000000000020
[  881.067495] Mem abort info:
[  881.070290]   ESR = 0x0000000096000006
[  881.074042]   EC = 0x25: DABT (current EL), IL = 32 bits
[  881.079358]   SET = 0, FnV = 0
[  881.082414]   EA = 0, S1PTW = 0
[  881.085558]   FSC = 0x06: level 2 translation fault
[  881.090439] Data abort info:
[  881.093320]   ISV = 0, ISS = 0x00000006
[  881.097157]   CM = 0, WnR = 0
[  881.100126] user pgtable: 4k pages, 48-bit VAs, pgdp=000000004fa51000
[  881.106573] [0000000000000020] pgd=080000004f36e003, p4d=080000004f36e003, pud=080000004f7ec003, pmd=0000000000000000
[  881.117217] Internal error: Oops: 0000000096000006 [#1] PREEMPT SMP
[  881.123494] Modules linked in: rcar_fdp1 v4l2_mem2mem
[  881.128572] CPU: 0 PID: 1271 Comm: yavta Tainted: G    B              6.2.0-rc1-00023-g6c94e2e99343 #556
[  881.138061] Hardware name: Renesas Salvator-X 2nd version board based on r8a77965 (DT)
[  881.145981] pstate: 400000c5 (nZcv daIF -PAN -UAO -TCO -DIT -SSBS BTYPE=--)
[  881.152951] pc : vsp1_dl_list_add_body+0xa8/0xe0
[  881.157580] lr : vsp1_dl_list_add_body+0x34/0xe0
[  881.162206] sp : ffff80000c267710
[  881.165522] x29: ffff80000c267710 x28: ffff000010938ae8 x27: ffff000013a8dd98
[  881.172683] x26: ffff000010938098 x25: ffff000013a8dc00 x24: ffff000010ed6ba8
[  881.179841] x23: ffff00000faa4000 x22: 0000000000000000 x21: 0000000000000020
[  881.186998] x20: ffff00000faa4000 x19: 0000000000000000 x18: 0000000000000000
[  881.194154] x17: 0000000000000000 x16: 0000000000000000 x15: 0000000000000000
[  881.201309] x14: 0000000000000000 x13: 746e696174206c65 x12: ffff70000157043d
[  881.208465] x11: 1ffff0000157043c x10: ffff70000157043c x9 : dfff800000000000
[  881.215622] x8 : ffff80000ab821e7 x7 : 00008ffffea8fbc4 x6 : 0000000000000001
[  881.222779] x5 : ffff80000ab821e0 x4 : ffff70000157043d x3 : 0000000000000020
[  881.229936] x2 : 0000000000000020 x1 : ffff00000e4f6400 x0 : 0000000000000000
[  881.237092] Call trace:
[  881.239542]  vsp1_dl_list_add_body+0xa8/0xe0
[  881.243822]  vsp1_video_pipeline_run+0x270/0x2a0
[  881.248449]  vsp1_video_buffer_queue+0x1c0/0x1d0
[  881.253076]  __enqueue_in_driver+0xbc/0x260
[  881.257269]  vb2_start_streaming+0x48/0x200
[  881.261461]  vb2_core_streamon+0x13c/0x280
[  881.265565]  vb2_streamon+0x3c/0x90
[  881.269064]  vsp1_video_streamon+0x2fc/0x3e0
[  881.273344]  v4l_streamon+0x50/0x70
[  881.276844]  __video_do_ioctl+0x2bc/0x5d0
[  881.280861]  video_usercopy+0x2a8/0xc80
[  881.284704]  video_ioctl2+0x20/0x40
[  881.288201]  v4l2_ioctl+0xa4/0xc0
[  881.291525]  __arm64_sys_ioctl+0xe8/0x110
[  881.295543]  invoke_syscall+0x68/0x190
[  881.299303]  el0_svc_common.constprop.0+0x88/0x170
[  881.304105]  do_el0_svc+0x4c/0xf0
[  881.307430]  el0_svc+0x4c/0xa0
[  881.310494]  el0t_64_sync_handler+0xbc/0x140
[  881.314773]  el0t_64_sync+0x190/0x194
[  881.318450] Code: d50323bf d65f03c0 91008263 f9800071 (885f7c60)
[  881.324551] ---[ end trace 0000000000000000 ]---
[  881.329173] note: yavta[1271] exited with preempt_count 1

A different regression report sent to the linux-media mailing list ([1])
was answered with a claim that the vb2_is_streaming() function has never
been meant for this purpose. The document of the function, as well as of
the struct vb2_queue streaming field, is sparse, so this claim may be
hard to verify.

The information needed by the vsp1 driver to decide how to process
queued buffers is also available from the vb2_start_streaming_called()
function. Use it instead of vb2_is_streaming() to fix the problem.

[1] https://lore.kernel.org/linux-media/545610e7-3446-2b82-60dc-7385fea3774f@redhat.com/

Fixes: a10b21532574 ("media: vb2: add (un)prepare_streaming queue ops")
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Tested-by: Duy Nguyen <duy.nguyen.rh@renesas.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: vsp1: Remove unused vsp1_subdev_internal_ops declaration
Gaosheng Cui [Sun, 11 Sep 2022 09:15:13 +0000 (17:15 +0800)]
media: vsp1: Remove unused vsp1_subdev_internal_ops declaration

vsp1_subdev_internal_ops has been removed since commit 0efdf0f5eaaf
("[media] v4l: vsp1: Implement and use the subdev pad::init_cfg
configuration"), so remove it.

Signed-off-by: Gaosheng Cui <cuigaosheng1@huawei.com>
Reviewed-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: platform: mtk-mdp3: reconfigure shared memory
Moudy Ho [Mon, 27 Mar 2023 03:13:35 +0000 (11:13 +0800)]
media: platform: mtk-mdp3: reconfigure shared memory

For performance and multi-chip support, use dynamic layout instead of
statically configured pools.
Divide the shared memory into the 3 64-bit aligned layouts listed below:

    vpu->param_addr  -> +-----------------------------------------+
                        |                                         |
                        | To SCP : Input frame parameters         |
                        |          (struct img_ipi_frameparam)    |
                        |                                         |
                        +-----------------------------------------+

    vpu->work_addr   -> +-----------------------------------------+
                        |                                         |
                        | In SCP : Reserve for SCP calculation    |
                        |                                         |
                        +-----------------------------------------+

    vpu->config_addr -> +-----------------------------------------+
                        |                                         |
                        | From SCP : Output component config      |
                        |            (struct img_config)          |
                        |                                         |
                        +-----------------------------------------+

Signed-off-by: Moudy Ho <moudy.ho@mediatek.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: platform: mtk-mdp3: decompose hardware-related information in shared memory
Moudy Ho [Mon, 27 Mar 2023 03:13:34 +0000 (11:13 +0800)]
media: platform: mtk-mdp3: decompose hardware-related information in shared memory

The communication between the MDP3 kernel driver and SCP is to
pass a shared memory through the cooperation of "mtk-mdp3-vpu.c" and
remoteproc driver.
The data structure of this shared memory is defined in "mtk-img-ipi.h",
as shown below:

vpu->work_addr -> +-----------------------------------------+
                  |                                         |
                  | To SCP : Input frame parameters         |
                  |          (struct img_ipi_frameparam)    |
                  |                                         |
     vpu->pool -> +-----------------------------------------+
                  |                                         |
                  | From SCP : Output component config pool |
                  |            (struct img_config)          |
                  |                                         |
                  |           *struct img_config 1          |
                  |                    |                    |
                  |                    |                    |
                  |                    v                    |
                  |           *struct img_config N          |
                  |            (N = MDP_CONFIG_POOL_SIZE)   |
                  +-----------------------------------------+

One output component configuration contains the components
currently used by the pipeline, and has the register settings
that each component needs to set.

Since the quantity, type and function of components on each chip
will vary, the effect is that the size of the "struct img_config"
and its substructures will be different on each chip.
In addition, all chips will have to update their SCP firmware for
every change if the output component config structure is defined
and shared by a common header.

Therefore, all functions that operate on "struct img_config" and
its substructures must be separated by chips and so are the
relevant definations.

Signed-off-by: Moudy Ho <moudy.ho@mediatek.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: platform: mtk-mdp3: Split general definitions used in MDP3
Moudy Ho [Mon, 27 Mar 2023 03:13:33 +0000 (11:13 +0800)]
media: platform: mtk-mdp3: Split general definitions used in MDP3

Separate the generic definitions used in MDP3 to avoid recursive includes
when splitting chip-related definitions in further.

Signed-off-by: Moudy Ho <moudy.ho@mediatek.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: platform: mtk-mdp3: dynamically allocate component clocks
Moudy Ho [Mon, 27 Mar 2023 03:13:32 +0000 (11:13 +0800)]
media: platform: mtk-mdp3: dynamically allocate component clocks

For extensibility of adding more MDP3 components in the further, the
magic number is removed by dynamically allocating component clocks.

Signed-off-by: Moudy Ho <moudy.ho@mediatek.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: platform: mtk-mdp3: extend mdp_color format for compressed mode
Moudy Ho [Mon, 27 Mar 2023 03:13:31 +0000 (11:13 +0800)]
media: platform: mtk-mdp3: extend mdp_color format for compressed mode

Due to future support for hardware compression formats, the MDP3
internal color format expression needs to be extended.

Signed-off-by: Moudy Ho <moudy.ho@mediatek.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: platform: mtk-mdp3: chip config split about pipe info
Moudy Ho [Mon, 27 Mar 2023 03:13:30 +0000 (11:13 +0800)]
media: platform: mtk-mdp3: chip config split about pipe info

In MDP3, a pipe is used to represent a data path which
consisting of different components and MUTEX, as shown in
the following diagram:

                     +----------------------------+
                     |           MUTEX [*1]       |
                     +----------------------------+
                        S ^ S ^  S ^  S ^  S ^
                        | | | |  | |  | |  | |
                        | | | |  | |  | |  | |
         +--------------+ | | |  | |  | |  | +-------------+
         | +--------------+ | |  | |  | |  +-------------+ |
         | |         +------+ |  | |  | +------+         | |
         | |         | +------+  | |  +------+ |         | |
         | |         | |         | |         | |         | |
         | |         | |         | |         | |         | |
         v E         v E         v E         v E         v E
      +-------+    +-----+    +-------+    +-----+    +-------+
B>--->| RDMA0 |>-->| PQ  |>-->| RSZ0  |>-->| PQ  |>-->| WROT0 |>-->B
      +-------+    +-----+    +-------+    +-----+    +-------+

[*1] MUTEX is SOF/EOF signal controller
[*2] S = Start of frame (SOF)
[*3] E = End of frame (EOF)
[*4] B = frame buffer

More pipes will be introduced in future chips for applications such as
higher frame rate frequency, which should integrate and reorder related
information into specific chip config file.

Signed-off-by: Moudy Ho <moudy.ho@mediatek.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: platform: mtk-mdp3: chip config split about resolution limitations
Moudy Ho [Mon, 27 Mar 2023 03:13:29 +0000 (11:13 +0800)]
media: platform: mtk-mdp3: chip config split about resolution limitations

Due to differences in hardware design, the supported max and min
resolutions and scaling capabilities will vary, and should be
integrated into specific config file.

Signed-off-by: Moudy Ho <moudy.ho@mediatek.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: platform: mtk-mdp3: chip config split about color format
Moudy Ho [Mon, 27 Mar 2023 03:13:28 +0000 (11:13 +0800)]
media: platform: mtk-mdp3: chip config split about color format

Due to differences in hardware design, the supported color formats
will vary and should be integrated into specific config file.

Signed-off-by: Moudy Ho <moudy.ho@mediatek.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: platform: mtk-mdp3: chip config split about subcomponents
Moudy Ho [Mon, 27 Mar 2023 03:13:27 +0000 (11:13 +0800)]
media: platform: mtk-mdp3: chip config split about subcomponents

Integrate subcomponent related information into specific config file for
further multi-chip compatibility.

Signed-off-by: Moudy Ho <moudy.ho@mediatek.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: platform: mtk-mdp3: chip config split about component settings
Moudy Ho [Mon, 27 Mar 2023 03:13:26 +0000 (11:13 +0800)]
media: platform: mtk-mdp3: chip config split about component settings

For different chips, different types and numbers of components are
configured in the MDP3 to achieve the desired application, which will
cause ID compatibility problems in the same driver.

Subdivide the component ID into two kinds:
1. public_id (compatible with other chips)
2. inner_id (what the current chip actually owns)

In addition, the chip configuration related structures of MDP3 components
are integrated into specific file.

Signed-off-by: Moudy Ho <moudy.ho@mediatek.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: platform: mtk-mdp3: add files for chip configuration
Moudy Ho [Mon, 27 Mar 2023 03:13:25 +0000 (11:13 +0800)]
media: platform: mtk-mdp3: add files for chip configuration

In order to be compatible with more MDP3 chip settings in further,
integrate and separate chip-related configurations into specific files.

Signed-off-by: Moudy Ho <moudy.ho@mediatek.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: platform: mtk-mdp3: fix potential frame size overflow in mdp_try_fmt_mplane()
Moudy Ho [Mon, 27 Mar 2023 03:13:24 +0000 (11:13 +0800)]
media: platform: mtk-mdp3: fix potential frame size overflow in mdp_try_fmt_mplane()

Fix overflow risk when setting certain formats whose frame size exceeds
a RGB24 with 7723x7723 resolution.

For example, a 7723x7724 RGB24 frame:
    1. bpl (byte per line) = 7723 * 3.
    2. Overflow occurs when bpl * 7724 * depth.

Fixes: 61890ccaefaf ("media: platform: mtk-mdp3: add MediaTek MDP3 driver")
Signed-off-by: Moudy Ho <moudy.ho@mediatek.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: au0828: remove unnecessary (void*) conversions
Yu Zhe [Mon, 20 Mar 2023 06:57:14 +0000 (14:57 +0800)]
media: au0828: remove unnecessary (void*) conversions

Pointer variables of void * type do not require type cast.

Signed-off-by: Yu Zhe <yuzhe@nfschina.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: saa7134: fix use after free bug in saa7134_finidev due to race condition
Zheng Wang [Sat, 18 Mar 2023 08:50:23 +0000 (16:50 +0800)]
media: saa7134: fix use after free bug in saa7134_finidev due to race condition

In saa7134_initdev, it will call saa7134_hwinit1. There are three
function invoking here: saa7134_video_init1, saa7134_ts_init1
and saa7134_vbi_init1.

All of them will init a timer with same function. Take
saa7134_video_init1 as an example. It'll bound &dev->video_q.timeout
with saa7134_buffer_timeout.

In buffer_activate, the timer funtcion is started.

If we remove the module or device which will call saa7134_finidev
to make cleanup, there may be a unfinished work. The
possible sequence is as follows, which will cause a
typical UAF bug.

Fix it by canceling the timer works accordingly before cleanup in
saa7134_finidev.

CPU0                  CPU1

                    |saa7134_buffer_timeout
saa7134_finidev     |
  kfree(dev);       |
                    |
                    | saa7134_buffer_next
                    | //use dev

Fixes: 1e7126b4a86a ("media: saa7134: Convert timers to use timer_setup()")
Signed-off-by: Zheng Wang <zyytlz.wz@163.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: dm1105: Fix use after free bug in dm1105_remove due to race condition
Zheng Wang [Sat, 18 Mar 2023 08:15:06 +0000 (16:15 +0800)]
media: dm1105: Fix use after free bug in dm1105_remove due to race condition

In dm1105_probe, it called dm1105_ir_init and bound
&dm1105->ir.work with dm1105_emit_key.
When it handles IRQ request with dm1105_irq,
it may call schedule_work to start the work.

When we call dm1105_remove to remove the driver, there
may be a sequence as follows:

Fix it by finishing the work before cleanup in dm1105_remove

CPU0                  CPU1

                    |dm1105_emit_key
dm1105_remove      |
  dm1105_ir_exit       |
    rc_unregister_device |
    rc_free_device  |
    rc_dev_release  |
    kfree(dev);     |
                    |
                    | rc_keydown
                    |   //use

Fixes: 34d2f9bf189c ("V4L/DVB: dm1105: use dm1105_dev & dev instead of dm1105dvb")
Signed-off-by: Zheng Wang <zyytlz.wz@163.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agousbtv: usbtv_set_regs: the pipe is output
Oliver Neukum [Thu, 16 Mar 2023 13:15:07 +0000 (14:15 +0100)]
usbtv: usbtv_set_regs: the pipe is output

We are setting a value. That is output and
the pipe has to match that.

Signed-off-by: Oliver Neukum <oneukum@suse.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: atmel: atmel-isc: Use devm_platform_ioremap_resource()
Yang Li [Tue, 14 Mar 2023 05:58:26 +0000 (13:58 +0800)]
media: atmel: atmel-isc: Use devm_platform_ioremap_resource()

According to commit 7945f929f1a7 ("drivers: provide
devm_platform_ioremap_resource()"), convert platform_get_resource(),
devm_ioremap_resource() to a single call to use
devm_platform_ioremap_resource(), as this is exactly what this function
does.

Signed-off-by: Yang Li <yang.lee@linux.alibaba.com>
Reviewed-by: Claudiu Beznea <claudiu.beznea@microchip.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: i2c: adv748x: Report correct DV timings for pattern generator
Niklas Söderlund [Wed, 22 Feb 2023 22:12:49 +0000 (23:12 +0100)]
media: i2c: adv748x: Report correct DV timings for pattern generator

If the pattern generator is enabled the device shall not be queried for
timings. Instead the timings programmed shall be reported as they are
the ones being used to generate the pattern.

Before this change an external HDMI source needed to be connected for
the pattern generator to work. The driver would query this external
HDMI source for timings and program the pattern generator using those.

With this change the user can control the timings and have the pattern
generator work without the need of an external HDMI source being
connected.

Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: i2c: adv748x: Write initial DV timings to device
Niklas Söderlund [Wed, 22 Feb 2023 22:12:48 +0000 (23:12 +0100)]
media: i2c: adv748x: Write initial DV timings to device

When initializing the HDMI block during probe an initial set of timings
are selected. These timings are stored in the drivers private data, but
not written to the device.

This in itself is not bad, but in s_dv_timings() the timings stored in
the drivers private data are compared to the new timings, if they match
no action is taken. This creates the corner-case where the timing
selected at initialization is the first timings a user want to use as
the driver then never writes it to the device preventing proper
operation.

Fix this by writing the timings to the device at initialization in
addition to storing them in the drivers private data.

Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: i2c: adv748x: Fix lookup of DV timings
Niklas Söderlund [Wed, 22 Feb 2023 22:12:47 +0000 (23:12 +0100)]
media: i2c: adv748x: Fix lookup of DV timings

The loop to match the requested timings with the ones supported by the
driver is incorrect. It always iterates thru the whole array of
supported modes. The bounds check after the loop always triggers
resulting in adv748x_hdmi_set_video_timings() always returning -EINVAL.

Fix this by correcting the lookup to break the loop when a match is
found.

Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: camss: sm8250: Pipeline starting and stopping for multiple virtual channels
Milen Mitkov [Fri, 9 Dec 2022 09:40:37 +0000 (11:40 +0200)]
media: camss: sm8250: Pipeline starting and stopping for multiple virtual channels

Use the multistream series function video_device_pipeline_alloc_start
to allows multiple clients of the same pipeline.

If the VFE entity is used by another instance of the pipeline,
the pipeline won't be stopped. This allows for stopping and starting
streams at any point without disrupting the other running streams.

To prepare and start multiple virtual channels each CSID source pad
corresponding to a virtual channel must be linked to the corresponding
IFE entity. CSID pad 1 (1st source pad) corresponds to virtual
channel 0, CSID pad 2 corresponds to virtual channel 1 and so on.
Each of these must be linked to corresponding IFE RDI port.
E.g. to enable vc 0 on CSID0:

media-ctl -l '"msm_csid0":1->"msm_vfe0_rdi0":0[1]'

To enable vc1 on CSID0:

media-ctl -l '"msm_csid0":2->"msm_vfe0_rdi1":0[1]'

And so on. Note that on SM8250 each CSID is connected, at the
hardware level, to only one IFE. Thus, you must link CSID0
with IFE0, you can't link it with IFE1.

Example: the following media controller setup expects multiplexed
sensor data on CSIPHY2. Data will be passed on to CSID0, which will
demux it to 2 streams - for RDI0 and RD1 ports of IFE0:

media-ctl -v -d /dev/media0 -V '"imx577 '22-001a'":0[fmt:SRGGB10/3840x2160 field:none]'
media-ctl -V '"msm_csiphy2":0[fmt:SRGGB10/3840x2160]'
media-ctl -V '"msm_csid0":0[fmt:SRGGB10/3840x2160]'
media-ctl -V '"msm_csid0":1[fmt:SRGGB10/3840x2160]'
media-ctl -V '"msm_csid0":2[fmt:SRGGB10/3840x2160]'
media-ctl -V '"msm_vfe0_rdi0":0[fmt:SRGGB10/3840x2160]'
media-ctl -V '"msm_vfe0_rdi1":0[fmt:SRGGB10/3840x2160]'
media-ctl -l '"msm_csiphy2":1->"msm_csid0":0[1]'
media-ctl -l '"msm_csid0":1->"msm_vfe0_rdi0":0[1]'
media-ctl -l '"msm_csid0":2->"msm_vfe0_rdi1":0[1]'

Note: CSID's entity pad 0 is a sink pad, pads 1..4 are source pads
To start streaming a v4l2 client must open the corresponding
/dev/videoN node. For example, with yavta:

yavta -B capture-mplane -c -I -n 5 -f SRGGB10P -s 3840x2160 -F /dev/video0
yavta -B capture-mplane -c -I -n 5 -f SRGGB10P -s 3840x2160 -F /dev/video1

Note that IFEs (vfe0, vfe1) on SM8250 have 3 RDI ports and a single
PIX port and IFELites (vfe2, vfe3) have 4 RDI ports and no PIX port.

Signed-off-by: Milen Mitkov <quic_mmitkov@quicinc.com>
Reviewed-by: Robert Foss <robert.foss@linaro.org>
Tested-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Acked-by: Robert Foss <robert.foss@linaro.org>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: camss: vfe-480: Multiple outputs support for SM8250
Milen Mitkov [Fri, 9 Dec 2022 09:40:36 +0000 (11:40 +0200)]
media: camss: vfe-480: Multiple outputs support for SM8250

On SM8250 each VFE supports at least 3 RDI channels, or 4
in case of VFE-Lite, so add appropriate IRQ setup and handling.

Signed-off-by: Milen Mitkov <quic_mmitkov@quicinc.com>
Reviewed-by: Robert Foss <robert.foss@linaro.org>
Tested-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Acked-by: Robert Foss <robert.foss@linaro.org>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: camss: vfe: Reserve VFE lines on stream start and link to CSID
Milen Mitkov [Fri, 9 Dec 2022 09:40:35 +0000 (11:40 +0200)]
media: camss: vfe: Reserve VFE lines on stream start and link to CSID

For multiple virtual channels support, each VFE line can be in either
ON, RESERVED or OFF states. This allows the starting and stopping
of a VFE line independently of other active VFE lines (e.g. already-
running lines stay in ON state, and newly-added lines are RESERVED)

Also, link the CSID entity's source ports to corresponding VFE lines.

Signed-off-by: Milen Mitkov <quic_mmitkov@quicinc.com>
Reviewed-by: Robert Foss <robert.foss@linaro.org>
Tested-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Acked-by: Robert Foss <robert.foss@linaro.org>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: camss: sm8250: Virtual channels for CSID
Milen Mitkov [Fri, 9 Dec 2022 09:40:34 +0000 (11:40 +0200)]
media: camss: sm8250: Virtual channels for CSID

CSID hardware on SM8250 can demux up to 4 simultaneous streams
based on virtual channel (vc) or datatype (dt).
The CSID subdevice entity now has 4 source ports that can be
enabled/disabled and thus can control which virtual channels
are enabled. Datatype demuxing not tested.

In order to keep a valid internal state of the subdevice,
implicit format propagation from the sink to the source pads
has been preserved. However, the format on each source pad
can be different and in that case it must be configured explicitly.

CSID's s_stream is called when any stream is started or stopped.
It will call configure_streams() that will rewrite IRQ settings to HW.
When multiple streams are running simultaneously there is an issue
when writing IRQ settings for one stream while another is still
running, thus avoid re-writing settings if they were not changed
in link setup, or by fully powering off the CSID hardware.

Signed-off-by: Milen Mitkov <quic_mmitkov@quicinc.com>
Reviewed-by: Robert Foss <robert.foss@linaro.org>
Tested-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Acked-by: Robert Foss <robert.foss@linaro.org>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: it913x: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:32:24 +0000 (16:32 +0200)]
media: it913x: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: vivid-core: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:32:23 +0000 (16:32 +0200)]
media: vivid-core: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: visl-core: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:32:22 +0000 (16:32 +0200)]
media: visl-core: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: vimc-core: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:32:21 +0000 (16:32 +0200)]
media: vimc-core: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: vim2m: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:32:20 +0000 (16:32 +0200)]
media: vim2m: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: vidtv_bridge: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:32:19 +0000 (16:32 +0200)]
media: vidtv_bridge: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: vicodec-core: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:32:18 +0000 (16:32 +0200)]
media: vicodec-core: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: sunxi-cir: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:32:17 +0000 (16:32 +0200)]
media: sunxi-cir: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Acked-by: Jernej Skrabec <jernej.skrabec@gmail.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: st_rc: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:32:16 +0000 (16:32 +0200)]
media: st_rc: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: mtk-cir: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:32:15 +0000 (16:32 +0200)]
media: mtk-cir: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: meson-ir: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:32:14 +0000 (16:32 +0200)]
media: meson-ir: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: meson-ir-tx: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:32:13 +0000 (16:32 +0200)]
media: meson-ir-tx: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: ir-hix5hd2: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:32:12 +0000 (16:32 +0200)]
media: ir-hix5hd2: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: img-ir-core: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:32:11 +0000 (16:32 +0200)]
media: img-ir-core: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: gpio-ir-recv: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:32:10 +0000 (16:32 +0200)]
media: gpio-ir-recv: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: radio-platform-si4713: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:32:09 +0000 (16:32 +0200)]
media: radio-platform-si4713: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: radio-wl1273: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:32:08 +0000 (16:32 +0200)]
media: radio-wl1273: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: radio-timb: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:32:07 +0000 (16:32 +0200)]
media: radio-timb: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: radio-si476x: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:32:06 +0000 (16:32 +0200)]
media: radio-si476x: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: xilinx-vtc: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:32:05 +0000 (16:32 +0200)]
media: xilinx-vtc: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: xilinx-vipp: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:32:04 +0000 (16:32 +0200)]
media: xilinx-vipp: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: xilinx-tpg: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:32:03 +0000 (16:32 +0200)]
media: xilinx-tpg: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: xilinx-csi2rxss: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:32:02 +0000 (16:32 +0200)]
media: xilinx-csi2rxss: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: video-mux: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:32:01 +0000 (16:32 +0200)]
media: video-mux: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: via-camera: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:32:00 +0000 (16:32 +0200)]
media: via-camera: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: hantro_drv: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:31:59 +0000 (16:31 +0200)]
media: hantro_drv: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: vpe: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:31:58 +0000 (16:31 +0200)]
media: vpe: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: isp: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:31:57 +0000 (16:31 +0200)]
media: isp: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
18 months agomedia: omap_vout: Convert to platform remove callback returning void
Uwe Kleine-König [Sun, 26 Mar 2023 14:31:56 +0000 (16:31 +0200)]
media: omap_vout: Convert to platform remove callback returning void

The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is (mostly) ignored
and this typically results in resource leaks. To improve here there is a
quest to make the remove callback return void. In the first step of this
quest all drivers are converted to .remove_new() which already returns
void.

Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>