Trivial USB ID addition for Avermedia H335.
Signed-off-by: NMalcolm Priestley <tvboxspy@gmail.com>
Cc: stable@vger.kernel.org # v3.11+
Signed-off-by: NMauro Carvalho Chehab <m.chehab@samsung.com>

Use the %zu and %pad printk specifiers to print size_t and dma_addr_t
variables. This fixes warnings on platforms where dma_addr_t has a
different size than int.

Cc: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Signed-off-by: NLaurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Signed-off-by: NMauro Carvalho Chehab <m.chehab@samsung.com>

PCTV 461e requires that small delay.
Signed-off-by: NAntti Palosaari <crope@iki.fi>
Signed-off-by: NMauro Carvalho Chehab <m.chehab@samsung.com>

New chip version, which is very similar than EM28174.
Signed-off-by: NAntti Palosaari <crope@iki.fi>
Signed-off-by: NMauro Carvalho Chehab <m.chehab@samsung.com>

In the ISI driver it reads the config register to get original value,
then set the correct FRATE_DIV and YCC_SWAP_MODE directly. This will
cause some bits overlap.

So we need to clear these bits first, then set correct value. This patch
fix it.
Signed-off-by: NJosh Wu <josh.wu@atmel.com>
Acked-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: NMauro Carvalho Chehab <m.chehab@samsung.com>

The ISI_CFG2.YCC_SWAP field controls color component ordering. The
datasheet lists the following orderings for the memory formats.

YCC_SWAP	Byte 0	Byte 1	Byte 2	Byte 3
00: Default	Cb(i)	Y(i)	Cr(i)	Y(i+1)
01: Mode1	Cr(i)	Y(i)	Cb(i)	Y(i+1)
10: Mode2	Y(i)	Cb(i)	Y(i+1)	Cr(i)
11: Mode3	Y(i)	Cr(i)	Y(i+1)	Cb(i)

This is based on a sensor format set to CbYCrY (UYVY). The driver
hardcodes the output memory format to YUYV, configure the ordering
accordingly.
Signed-off-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: NJosh Wu <josh.wu@atmel.com>
Signed-off-by: NMauro Carvalho Chehab <m.chehab@samsung.com>

ISI_MCK is the sensor master clock. It should be handled by the sensor
driver directly, as the ISI has no use for that clock. Make the clock
optional here while platforms transition to the correct model.
Signed-off-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: NJosh Wu <josh.wu@atmel.com>
Signed-off-by: NMauro Carvalho Chehab <m.chehab@samsung.com>

The queue setup operation isn't the right place to reset the ISI. Move
the reset call to the start streaming operation.
Signed-off-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: NJosh Wu <josh.wu@atmel.com>
Signed-off-by: NMauro Carvalho Chehab <m.chehab@samsung.com>

The PCLK and MCK clocks are prepared and unprepared at probe and remove
time. Clock (un)preparation isn't needed before enabling/disabling the
clocks, and the enable/disable operation happen in non-atomic context.
We can thus defer (un)preparation to enable/disable time.
Signed-off-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: NJosh Wu <josh.wu@atmel.com>
Signed-off-by: NMauro Carvalho Chehab <m.chehab@samsung.com>

This simplifies error and cleanup code paths.
Signed-off-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: NJosh Wu <josh.wu@atmel.com>
Signed-off-by: NMauro Carvalho Chehab <m.chehab@samsung.com>

when a userspace applications calls the VIDIOC_STREAMON ioctl. The
V4L2 core calls the soc_camera_streamon function, which is responsible
for starting the video stream. It does so by first starting the atmel-isi
host by a call to the vb2_streamon function, and then starting the sensor
by a call to the video.s_stream sensor subdev operation.
That means we wait for a SOF in start_streaming() before call sensor's
s_stream(). It is possible no VSYNC interrupt arrive as the sensor
hasn't been started yet.
To avoid such case, this patch remove the code to wait for the VSYNC
interrupt. And such code is not necessary.
Reported-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: NJosh Wu <josh.wu@atmel.com>
Signed-off-by: NMauro Carvalho Chehab <m.chehab@samsung.com>

Fix the non-whitespace checkpatch errors/warnings.
Replace msleep with usleep_range and the jiffies comparison with
time_is_after_jiffies().
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Acked-by: NEduardo Valentin <edubezval@gmail.com>
Acked-by: NDinesh Ram <dinesh.ram@cern.ch>
Signed-off-by: NMauro Carvalho Chehab <m.chehab@samsung.com>

Fix most whitespace-related checkpatch errors/warnings.
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Acked-by: NEduardo Valentin <edubezval@gmail.com>
Acked-by: NDinesh Ram <dinesh.ram@cern.ch>
Signed-off-by: NMauro Carvalho Chehab <m.chehab@samsung.com>

Print the PN value, useful to check what chip the dev board has.
Signed-off-by: NEduardo Valentin <edubezval@gmail.com>
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NMauro Carvalho Chehab <m.chehab@samsung.com>

si4713_set_rds_radio_text will overwrite the terminating zero at the
end of the rds radio text string in order to send out a carriage return
as per the RDS spec.
Use a separate char buffer for the CR instead of corrupting the control
string.
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Tested-by: NEduardo Valentin <edubezval@gmail.com>
Acked-by: NEduardo Valentin <edubezval@gmail.com>
Signed-off-by: NMauro Carvalho Chehab <m.chehab@samsung.com>

The supply list is needed by the platform driver, but not by the usb driver.
So this information belongs to the platform data and should not be hardcoded
in the subdevice driver.
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Tested-by: NEduardo Valentin <edubezval@gmail.com>
Acked-by: NEduardo Valentin <edubezval@gmail.com>
Signed-off-by: NMauro Carvalho Chehab <m.chehab@samsung.com>

Hans Verkuil <hverkuil@xs4all.nl> will maintain the USB driver for si4713
Signed-off-by: NDinesh Ram <dinesh.ram@cern.ch>
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Tested-by: NEduardo Valentin <edubezval@gmail.com>
Acked-by: NEduardo Valentin <edubezval@gmail.com>
Signed-off-by: NMauro Carvalho Chehab <m.chehab@samsung.com>

This is the USB driver for the Silicon Labs development board.
It contains the Si4713 FM transmitter chip.
Signed-off-by: NDinesh Ram <dinesh.ram@cern.ch>
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Tested-by: NEduardo Valentin <edubezval@gmail.com>
Acked-by: NEduardo Valentin <edubezval@gmail.com>
Signed-off-by: NMauro Carvalho Chehab <m.chehab@samsung.com>

The Si4713 development board contains a Si4713 FM transmitter chip
and is handled by the radio-usb-si4713 driver.
The board reports itself as (10c4:8244) Cygnal Integrated Products, Inc.
and misidentifies itself as a HID device in its USB interface descriptor.
This patch ignores this device as an HID device and hence loads the custom driver.
Signed-off-by: NDinesh Ram <dinesh.ram@cern.ch>
Signed-off-by: NJiri Kosina <jkosina@suse.cz>
Acked-by: NJiri Kosina <jkosina@suse.cz>
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Tested-by: NEduardo Valentin <edubezval@gmail.com>
Acked-by: NEduardo Valentin <edubezval@gmail.com>
Signed-off-by: NMauro Carvalho Chehab <m.chehab@samsung.com>

In the si4713_tx_tune_power() method, the args array element 'power'
can take values between SI4713_MIN_POWER and SI4713_MAX_POWER. power = 0
is also valid. All the values (0 > power < SI4713_MIN_POWER) are illegal
and hence are all mapped to SI4713_MIN_POWER.
Signed-off-by: NDinesh Ram <dinesh.ram@cern.ch>
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Tested-by: NEduardo Valentin <edubezval@gmail.com>
Acked-by: NEduardo Valentin <edubezval@gmail.com>
Signed-off-by: NMauro Carvalho Chehab <m.chehab@samsung.com>

Moved the header <linux/regulator/consumer.h> from si4713.c to si4713.h
Signed-off-by: NDinesh Ram <dinesh.ram@cern.ch>
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Tested-by: NEduardo Valentin <edubezval@gmail.com>
Acked-by: NEduardo Valentin <edubezval@gmail.com>
Signed-off-by: NMauro Carvalho Chehab <m.chehab@samsung.com>

Checks have been introduced at several places in the code to test if an
interrupt is set or not. For devices which do not use the interrupt, to
get a valid response, within a specified timeout, the device is polled
instead.
Signed-off-by: NDinesh Ram <dinesh.ram@cern.ch>
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Tested-by: NEduardo Valentin <edubezval@gmail.com>
Acked-by: NEduardo Valentin <edubezval@gmail.com>
Signed-off-by: NMauro Carvalho Chehab <m.chehab@samsung.com>

Added a new si4713 directory which will contain all si4713 related files.
Also updated Makefile and Kconfig
Signed-off-by: NDinesh Ram <dinesh.ram@cern.ch>
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Tested-by: NEduardo Valentin <edubezval@gmail.com>
Acked-by: NEduardo Valentin <edubezval@gmail.com>
Signed-off-by: NMauro Carvalho Chehab <m.chehab@samsung.com>

Rewrite the return statement in vpfe_video.c. This will prevent
the checkpatch.pl script from generating a warning saying
to remove () from this particular return statement.
Signed-off-by: NLisa Nguyen <lisa@xenapiadmin.com>
Acked-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: NLad, Prabhakar <prabhakar.csengg@gmail.com>
Signed-off-by: NMauro Carvalho Chehab <m.chehab@samsung.com>

Remove unnecessary spaces before semicolons to meet kernel
coding style.
Signed-off-by: NLisa Nguyen <lisa@xenapiadmin.com>
Acked-by: NLad, Prabhakar <prabhakar.csengg@gmail.com>
Signed-off-by: NMauro Carvalho Chehab <m.chehab@samsung.com>

The v4l2_of_get_remote_port() function acquires a reference to an
endpoint node through a phandle and then returns the node's parent,
without dropping the reference to the endpoint node. Fix it.
Signed-off-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMauro Carvalho Chehab <m.chehab@samsung.com>

The field isn't set when parsing the endpoint. Remove it.
Signed-off-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMauro Carvalho Chehab <m.chehab@samsung.com>

When CONFIG_OF is not defined the v4l2_of_parse_endpoint() function is
defined as a stub that returns -ENOSYS. Make the real function return an
integer as well to be able to differentiate between the two cases.
Signed-off-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMauro Carvalho Chehab <m.chehab@samsung.com>

A missing break resulted in all done buffers being flagged with
V4L2_BUF_FLAG_QUEUED. Fix it.
Signed-off-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: NSakari Ailus <sakari.ailus@iki.fi>
Signed-off-by: NMauro Carvalho Chehab <m.chehab@samsung.com>

Replace devm_request_mem_region() and devm_ioremap_nocache() with
devm_ioremap_resource(). The behaviour remains the same and the code is
simplified.
Signed-off-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: NSakari Ailus <sakari.ailus@iki.fi>
Signed-off-by: NMauro Carvalho Chehab <m.chehab@samsung.com>

The clock core code is going to be modified so clk_get() takes
reference on the clock provider module. Until the potential circular
reference issue is properly addressed, we pass NULL as the first
argument to clk_register(), in order to disallow sub-devices taking
a reference on the ISP module back trough clk_get(). This should
prevent locking the modules in memory.
Signed-off-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NKyungmin Park <kyungmin.park@samsung.com>
Acked-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: NMauro Carvalho Chehab <m.chehab@samsung.com>

This patch adds proper initialization of the in-driver
cached state of JPEG controls with V4L2 core.
Signed-off-by: NJacek Anaszewski <j.anaszewski@samsung.com>
Signed-off-by: NKyungmin Park <kyungmin.park@samsung.com>
Acked-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMauro Carvalho Chehab <m.chehab@samsung.com>

s5p_jpeg_device_run and s5p_jpeg_runtime_resume callbacks should
have spin_lock protection as they alter device registers.
Signed-off-by: NJacek Anaszewski <j.anaszewski@samsung.com>
Signed-off-by: NKyungmin Park <kyungmin.park@samsung.com>
Acked-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMauro Carvalho Chehab <m.chehab@samsung.com>

Standard suspend/resume path is called after runtime resume
of the given device, so suspend/resume callbacks must do all
clock management done also by runtime pm to allow for proper
power domain shutdown. Moreover, JPEG clock is enabled from
probe function but is is not necessary. This patch also moves
control of jpeg clock to runtime_pm callbacks.
Signed-off-by: NMarek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: NSeung-Woo Kim <sw0312.kim@samsung.com>
Signed-off-by: NJacek Anaszewski <j.anaszewski@samsung.com>
Signed-off-by: NKyungmin Park <kyungmin.park@samsung.com>
Acked-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMauro Carvalho Chehab <m.chehab@samsung.com>

Signed-off-by: NJacek Anaszewski <j.anaszewski@samsung.com>
Signed-off-by: NKyungmin Park <kyungmin.park@samsung.com>
Acked-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMauro Carvalho Chehab <m.chehab@samsung.com>

Aligning capture queue image dimensions while enqueuing output
queue doesn't make a sense as the S_FMT ioctl might have not
been called for the capture queue until that moment, whereas
it is required to know capture format as the type of alignment
heavily depends on it.
Signed-off-by: NJacek Anaszewski <j.anaszewski@samsung.com>
Signed-off-by: NKyungmin Park <kyungmin.park@samsung.com>
Acked-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMauro Carvalho Chehab <m.chehab@samsung.com>

The aim of the condition is ensuring that the bytesperline
value set by the user space application is proper for the
given format and adjusting it if isn't. As the depth value
of the format description entry is expressed in bits then
the bytesperline value needs to be divided, not multiplied,
by that value to get the number of bytes required to store
single line of image samples.
Signed-off-by: NJacek Anaszewski <j.anaszewski@samsung.com>
Signed-off-by: NKyungmin Park <kyungmin.park@samsung.com>
Acked-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMauro Carvalho Chehab <m.chehab@samsung.com>

Output samples during decoding phase for the YUV 4:2:0 format
are arranged in the manner compatible with 2-planar NV12,
not 3-planar YUV420 fourcc.
Signed-off-by: NJacek Anaszewski <j.anaszewski@samsung.com>
Signed-off-by: NKyungmin Park <kyungmin.park@samsung.com>
Acked-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMauro Carvalho Chehab <m.chehab@samsung.com>

Reorder quantization tables so that their elements
are arranged in the manner reflecting compression
quality level that is in accordance with V4L2 documentation,
i.e. the larger value of the V4L2_CID_JPEG_COMPRESSION_QUALITY
control the better image quality, and thus lower compression
quality. The modification allows also to get rid of
reverse logic in the s_ctrl callback while assigning
user space value to the ctx->compr_quality variable.
Signed-off-by: NJacek Anaszewski <j.anaszewski@samsung.com>
Signed-off-by: NKyungmin Park <kyungmin.park@samsung.com>
Acked-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMauro Carvalho Chehab <m.chehab@samsung.com>

This patch enables the JPEG codec on S5PV210 and Exynos4210 SoCs. There are
some differences in newer versions of the JPEG codec IP on SoCs like Exynos4x12
and Exynos5 series and support for them will be added in subsequent patches.
Signed-off-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Acked-by: NStephen Warren <swarren@nvidia.com>
Signed-off-by: NMauro Carvalho Chehab <m.chehab@samsung.com>