As of_node_full_name() now returns only the basename, the endpoint
information printed became useless:
adv748x 4-0070: Endpoint endpoint on port 7
adv748x 4-0070: Endpoint endpoint on port 8
adv748x 4-0070: Endpoint endpoint on port 10
adv748x 4-0070: Endpoint endpoint on port 11
Restore the old behavior by using "%pOF" instead:
adv748x 4-0070: Endpoint /soc/i2c@e66d8000/video-receiver@70/port@7/endpoint on port 7
adv748x 4-0070: Endpoint /soc/i2c@e66d8000/video-receiver@70/port@8/endpoint on port 8
adv748x 4-0070: Endpoint /soc/i2c@e66d8000/video-receiver@70/port@10/endpoint on port 10
adv748x 4-0070: Endpoint /soc/i2c@e66d8000/video-receiver@70/port@11/endpoint on port 11
Fixes: a7e4cfb0a7 ("of/fdt: only store the device node basename in full_name")
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Acked-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
timespec is generally deprecated because of the y2038 overflow.
In vivid, the usage is fine, since we are dealing with monotonic
timestamps, but we can also simplify the code by going to ktime_t.
Using ktime_divns() should be roughly as efficient as the old code,
since the constant 64-bit division gets turned into a multiplication
on modern platforms, and we save multiple 32-bit divisions that can be
expensive e.g. on ARMv7.
Tested-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
The imx media driver passes around monotonic timestamps in the deprecated
'timespec' format. This is not a problem for the driver, as they won't
overflow, but moving to either timespec64 or ktime_t is preferred.
I'm picking ktime_t for simplicity here. frame_interval_monitor() is
the main function that changes, as it tries to compare a time interval
in microseconds. The algorithm slightly changes here, to avoid 64-bit
division. The code previously assumed that the error was at most 32-bit
worth of microseconds here, so I'm making the same assumption but add
an explicit test for it.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
solo6x10 correctly deals with time stamps and will never
suffer from overflows, but it uses the deprecated 'struct timespec'
type and 'ktime_get_ts()' interface to read the monotonic clock.
This changes it to use ktime_get_ts64() instead, so we can
eventually remove ktime_get_ts().
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Ismael Luceno <ismael@iodev.co.uk>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Deblocking filter allows to reduce blocking artifacts and improve
visual quality. This is configurable via the V4L2 API but eventually
not applied to the encoder.
Note that alpha and beta deblocking values are 32-bit signed (-6;+6).
Signed-off-by: Loic Poulain <loic.poulain@linaro.org>
Reviewed-by: Stanimir Varbanov <stanimir.varbanov@linaro.org>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
H264 entropy mode can be selected via V4L2 API but is eventually not
applied. Configure encoder with selected mode, CALVC (def) or CABAC.
Note that hw/firmware also expects a CABAC model configuration which
currently doesn't have existing V4L2 API control. For now, use model_0
which seems always supported and so the default one.
Signed-off-by: Loic Poulain <loic.poulain@linaro.org>
Reviewed-by: Stanimir Varbanov <stanimir.varbanov@linaro.org>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Check return value from call to devm_kzalloc() in order to prevent
a NULL pointer dereference.
This issue was detected with the help of Coccinelle.
Fixes: 4a5f8ae50b ("[media] davinci: vpif_capture: get subdevs from DT when available")
Signed-off-by: Gustavo A. R. Silva <garsilva@embeddedor.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
This change resolves a new compile-time warning
when built as a loadable module:
WARNING: modpost: missing MODULE_LICENSE() in drivers/media/platform/tegra-cec/tegra_cec.o
see include/linux/module.h for more information
This adds the license as "GPL v2", which matches the header of the file.
MODULE_DESCRIPTION and MODULE_AUTHOR are also added.
Signed-off-by: Jesse Chan <jc@linux.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
This change resolves a new compile-time warning
when built as a loadable module:
WARNING: modpost: missing MODULE_LICENSE() in drivers/media/platform/soc_camera/soc_scale_crop.o
see include/linux/module.h for more information
This adds the license as "GPL", which matches the header of the file.
MODULE_DESCRIPTION and MODULE_AUTHOR are also added.
Signed-off-by: Jesse Chan <jc@linux.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
This change resolves a new compile-time warning
when built as a loadable module:
WARNING: modpost: missing MODULE_LICENSE() in drivers/media/platform/mtk-vcodec/mtk-vcodec-common.o
see include/linux/module.h for more information
This adds the license as "GPL v2", which matches the header of the file.
MODULE_DESCRIPTION is also added.
Signed-off-by: Jesse Chan <jc@linux.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
There is no need to #define the license of the driver, just put it in
the MODULE_LICENSE() line directly as a text string.
This allows tools that check that the module license matches the source
code license to work properly, as there is no need to unwind the
unneeded dereference.
Cc: Hans Verkuil <hverkuil@xs4all.nl>
Cc: Johan Hovold <johan@kernel.org>
Cc: Davidlohr Bueso <dave@stgolabs.net>
Cc: Sakari Ailus <sakari.ailus@linux.intel.com>
Reported-by: Philippe Ombredanne <pombredanne@nexb.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Reviewed-by: Philippe Ombredanne <pombredanne@nexb.com>
Acked-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
time_to_tm() takes a time_t value that overflows in 2038 on 32-bit
systems. time64_to_tm() doesn't have this problem, so let's use that in
combination with ktime_get_real_seconds() to read a 64-bit time
value.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
The cam->buffers[] array has cam->num_frames elements so the > needs to
be changed to >= to avoid going beyond the end of the array. The
->buffers[] array is allocated in cpia2_allocate_buffers() if you want
to confirm.
Fixes: ab33d5071d ("V4L/DVB (3376): Add cpia2 camera support")
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@s-opensource.com>
According to i.MX VPU API Reference Manuals the MJPG video codec is
referenced to by number 7, not 3.
Also Philipp pointed out that this value is only meant to fill in
CMD_ENC_SEQ_COD_STD for encoding, only on i.MX53. It was never written
to any register, and even if defined correctly, wouldn't be needed
for i.MX6.
So avoid confusion and remove this definition.
Signed-off-by: Martin Kepplinger <martink@posteo.de>
Acked-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Fix inconsistent IS_ERR and PTR_ERR in csi_link_validate.
The proper pointer to be passed as argument is sensor.
This issue was detected with the help of Coccinelle.
Reported-by: Julia Lawall <julia.lawall@lip6.fr>
Signed-off-by: Gustavo A. R. Silva <garsilva@embeddedor.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
When the device is being unregistered disable the hardware, don't wait
until cec_delete_adapter is called as the hardware may have disappeared by
then. This would be the case for hotplugable devices.
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Reported-by: Bård Eirik Winther <bwinther@cisco.com>
Tested-by: Bård Eirik Winther <bwinther@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Some devices can monitor the CEC pin using an interrupt, but you
only want to enable the interrupt if you actually switch to pin
monitoring mode.
So add a new op that is called when pin monitoring needs to be
switched on or off.
Also fix a small bug where the initial CEC pin event was sent
again when calling S_MODE twice with the same CEC_MODE_MONITOR_PIN
mode.
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
The firmware timestamp is an unsigned 32-bit value, but we copy it into
a signed 32-bit variable, so we can theoretically get an overflow in
the calculation when the timestamp is between 2038 and 2106.
This changes the temporary variable to time64_t and changes the deprecated
time_to_tm() over to time64_to_tm() accordingly.
There is still an overflow in y2106, but that is a limitation of the
firmware interface, not a kernel problem.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Add support for the test_pattern control and the h/vflip controls.
This makes it possible to switch to more interesting test patterns and to
test control handling in v4l-subdevs.
There are more tpg-related controls that can be added, but this is a good
start.
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
[hans.verkuil@cisco.com: fix small whitespace checkpatch warning]
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
fixes intermittent soft reboot issue with firmware load
increases wait time of reset, as required by HW
Signed-off-by: Dean Anderson <dean@sensoray.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
kasprintf() can fail here and we must check its return value.
Signed-off-by: Arvind Yadav <arvind.yadav.cs@gmail.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
When used as part of a display pipeline, the VSP is stopped and
restarted explicitly by the DU from its suspend and resume handlers.
There is thus no need to stop or restart pipelines in the VSP suspend
and resume handlers, and doing so would cause the hardware to be
left in a misconfigured state.
Ensure that the VSP suspend and resume handlers do not affect DRM-based
pipelines.
Signed-off-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
It's not clear if the transmit status that is printed when debugging
is enabled is hex or decimal. Add 0x prefix to make this explicit.
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
A 'return' was missing when detecting Arbitration Lost and
calling transmit_done. With the return transmit_done could be
called a second time, confusing the CEC framework. Luckily
the Arbitration Lost condition is very rare.
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
It is possible to call STREAMON without the minimum number of
buffers queued. In this case the vb2_queue state will be set to
streaming but the start_streaming vb2_op will not be called.
Later when enough buffers are queued, start_streaming will
be called but vb2_is_streaming will already return true.
Also removed the queue state check in stop_streaming since it's
not valid there either.
Signed-off-by: Ian Jamison <ian.dev@arkver.com>
Reviewed-by: Steve Longerbeam <steve_longerbeam@mentor.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
This change fixes the calculation of B-frames and GOP size by
adopt v4l2 controls with the firmware interface expectations.
Signed-off-by: Stanimir Varbanov <stanimir.varbanov@linaro.org>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
[hans.verkuil@cisco.com: fixed two small checkpatch comments]
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Note that the inputs for the Composite and S-Video are inverted in
regard to most of the other boards using a cx231xx chip.
Signed-off-by: Romain Reignier <r.reignier@robopec.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Since i2c_unregister_device() became NULL-aware we may remove duplicate
NULL check.
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Acked-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Since i2c_unregister_device() became NULL-aware we may remove duplicate checks.
Cc: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
The pointer h is already initialized to codeclist_top so the second
identical assignment is redundant and can be removed. Cleans up clang
warning:
drivers/media/pci/zoran/videocodec.c:322:21: warning: Value stored to 'h'
during its initialization is never read
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@s-opensource.com>
Variable mask is being set to 0x80 and then set to this value again
in the following for-loop. Remove the extraneous first setting of mask.
Cleans up clang warning:
drivers/media/pci/pt3/pt3_i2c.c:88:2: warning: Value stored to 'mask'
is never read
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@s-opensource.com>
Using the ARRAY_SIZE macro improves the readability of the code. Also,
it is not always useful to use a variable to store this constant
calculated at compile time.
Found with Coccinelle with the following semantic patch:
@r depends on (org || report)@
type T;
T[] E;
position p;
@@
(
(sizeof(E)@p /sizeof(*E))
|
(sizeof(E)@p /sizeof(E[...]))
|
(sizeof(E)@p /sizeof(T))
)
Signed-off-by: Jérémy Lefaure <jeremy.lefaure@lse.epita.fr>
Reviewed-by: Michael Ira Krufky <mkrufky@linuxtv.org>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Use a more typical logging style.
The current macro hides the gspca_dev argument so add it to the
macro uses instead.
Miscellanea:
o Add missing '\n' terminations to formats
o Realign arguments to open parenthesis
o Remove commented out uses of PDEBUG
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Use a more typical kernel logging style.
The current macro hides the gspca_dev argument so add it to the
macro uses instead.
Miscellanea:
o Add missing '\n' terminations to formats
o Realign arguments to open parenthesis
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
We free the stk_camera device too early. It's allocate first in probe
and it should be freed last in stk_camera_disconnect().
Reported-by: Andrey Konovalov <andreyknvl@google.com>
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.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@s-opensource.com>
Here, hdpvr_register_videodev() is responsible for setup and
register a video device. Also defining and initializing a worker.
hdpvr_register_videodev() is calling by hdpvr_probe at last.
So no need to flush any work here.
Unregister v4l2, free buffers and memory. If hdpvr_probe() will fail.
Signed-off-by: Arvind Yadav <arvind.yadav.cs@gmail.com>
Reported-by: Andrey Konovalov <andreyknvl@google.com>
Tested-by: Andrey Konovalov <andreyknvl@google.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Don't populate the read-only u8 array buf on the stack at run time but
instead make it static const; makes object code smaller saving over 480
bytes:
Before:
text data bss dec hex filename
33030 65936 192 99158 18356 drivers/media/usb/dvb-usb/cxusb.o
After:
text data bss dec hex filename
32446 66032 192 98670 1816e drivers/media/usb/dvb-usb/cxusb.o
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@s-opensource.com>
When building without CONFIG_GPIOLIB, we get a harmless
warning about an unused variable:
drivers/media/pci/solo6x10/solo6x10-gpio.c: In function 'solo_gpio_init':
drivers/media/pci/solo6x10/solo6x10-gpio.c:165:6: error: unused variable 'ret' [-Werror=unused-variable]
This adds another #ifdef around the declaration.
Fixes: d3202d1981 ("media: solo6x10: export hardware GPIO pins 8:31 to gpiolib interface")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Ismael Luceno <ismael@iodev.co.uk>
Acked-by: Anton Sviridenko <anton@corp.bluecherry.net>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Don't populate the arrays RegAddr on the stack, instead make them static
const. Makes the object code smaller by over 980 bytes:
Before:
text data bss dec hex filename
64923 304 0 65227 fecb drivers/media/tuners/mxl5005s.o
After:
text data bss dec hex filename
63779 464 0 64243 faf3 drivers/media/tuners/mxl5005s.o
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@s-opensource.com>
Use WARN_ON() instead of BUG_ON() to avoid crashing the kernel.
Signed-off-by: Srishti Sharma <srishtishar@gmail.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
The current implementation has to search the list of controls for the
pixel rate control, each time it is set. This can be optimised easily
by storing the ctrl pointer in the CSI/TX object, and referencing that
directly.
While at it, fix up a missing blank line also highlighted in review
comments.
Signed-off-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
Tested-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
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@s-opensource.com>
mt9v032_power_on() leaves clk enabled in case of errors,
but it is not expected by its callers.
There is a similar problem in mt9v032_registered().
Found by Linux Driver Verification project (linuxtesting.org).
Signed-off-by: Alexey Khoroshilov <khoroshilov@ispras.ru>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
The ov7670 has the test pattern generator features. This makes use of
it through V4L2_CID_TEST_PATTERN control.
Cc: Jonathan Corbet <corbet@lwn.net>
Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
The sub-device for ov7670 is registered by v4l2_async_register_subdev().
So it should be unregistered by v4l2_async_unregister_subdev() instead of
v4l2_device_unregister_subdev().
Cc: Jonathan Corbet <corbet@lwn.net>
Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Reviewed-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Connect voice coil lens to the camera sensor, so that it can be used
by camera applications.
Signed-off-by: Pavel Machek <pavel@ucw.cz>
Acked-by: Pali Rohár <pali.rohar@gmail.com>
Reviewed-by: Sebastian Reichel <sebastian.reichel@collabora.co.uk>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Document dts support of LED/focus.
Signed-off-by: Pavel Machek <pavel@ucw.cz>
Reviewed-by: Sebastian Reichel <sebastian.reichel@collabora.co.uk>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Add the missing MAINTAINERS entry for imx274, fix error handling in driver
probe and unregister the correct control handler in driver remove.
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
When matching and registering a sub-device from a sub-notifier use the
v4l2_device from the root parent notifier. Using the v4l2_dev stored in
the sub-notifier itself is incorrect as it might not be set.
This can be demonstrated by unbinding and rebinding the adv748x driver
and observing that it fails to probe due to the check !v4l2_dev in
v4l2_device_register_subdev().
# echo 4-0070 > /sys/bus/i2c/drivers/adv748x/unbind
# echo 4-0070 > /sys/bus/i2c/drivers/adv748x/bind
adv748x 4-0070: chip found @ 0xe0 revision 2143
adv748x 4-0070: Failed to probe TXA
adv748x: probe of 4-0070 failed with error -22
Looking at the commit which adds sub-notifiers to V4L2 it looks like
this is the intended behavior of the original commit. With this fix the
adv748x can be re-bound and still function properly.
Fixes: 2cab00bb07 ("media: v4l: async: Allow binding notifiers to sub-devices")
Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
