GitHub/moto-9609/android_kernel_motorola_exynos9610.git
9 years ago[media] rtl2830: fix logging
Antti Palosaari [Tue, 9 Dec 2014 02:32:19 +0000 (23:32 -0300)]
[media] rtl2830: fix logging

Pass correct device for dev_foo() logging in order to print logs
correctly.

Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] rtl2830: carry pointer to I2C client for every function
Antti Palosaari [Tue, 9 Dec 2014 01:47:21 +0000 (22:47 -0300)]
[media] rtl2830: carry pointer to I2C client for every function

As a I2C driver struct i2c_client is top level structure representing
the driver. Use it as parameter to carry all needed information for
each function in order to simplify things.

Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] rtl2830: rename 'priv' to 'dev'
Antti Palosaari [Tue, 9 Dec 2014 01:31:28 +0000 (22:31 -0300)]
[media] rtl2830: rename 'priv' to 'dev'

Use name 'dev' for device state instance as it is more common and
also one letter shorter.

Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] rtl2830: get rid of legacy DVB driver binding
Antti Palosaari [Tue, 9 Dec 2014 01:20:55 +0000 (22:20 -0300)]
[media] rtl2830: get rid of legacy DVB driver binding

Remove legacy DVB binding as all users are using I2C binding.

Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] rtl28xxu: use I2C binding for RTL2830 demod driver
Antti Palosaari [Mon, 8 Dec 2014 01:02:23 +0000 (22:02 -0300)]
[media] rtl28xxu: use I2C binding for RTL2830 demod driver

rtl2830 driver supports now I2C model too. Start using it.

Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] rtl2830: convert driver to kernel I2C model
Antti Palosaari [Sun, 7 Dec 2014 07:07:29 +0000 (04:07 -0300)]
[media] rtl2830: convert driver to kernel I2C model

Convert driver to kernel I2C model. Old DVB proprietary model is
still left there also.

Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] dvb-usb-v2: add pointer to 'struct usb_interface' for driver usage
Antti Palosaari [Wed, 17 Dec 2014 11:10:49 +0000 (08:10 -0300)]
[media] dvb-usb-v2: add pointer to 'struct usb_interface' for driver usage

Top level pointer on USB probe is struct usb_interface *. Add that
pointer to struct dvb_usb_device that drivers could use it, for
dev_* logging and more.

Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] smipcie: return more proper value in interrupt handler
Nibble Max [Thu, 25 Dec 2014 08:18:31 +0000 (05:18 -0300)]
[media] smipcie: return more proper value in interrupt handler

Although IRQ_HANDLED is "safe" value to return,
it is better to let the kernel know whether the driver handle the interrupt or not.

Signed-off-by: Nibble Max <nibble.max@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] media: platform: fix platform_no_drv_owner.cocci warnings
Fengguang Wu [Fri, 30 Jan 2015 02:26:10 +0000 (23:26 -0300)]
[media] media: platform: fix platform_no_drv_owner.cocci warnings

drivers/media/platform/am437x/am437x-vpfe.c:2767:3-8: No need to set .owner here. The core will do it.

 Remove .owner field if calls are used which set it automatically

Generated by: scripts/coccinelle/api/platform_no_drv_owner.cocci

CC: Benoit Parrot <bparrot@ti.com>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Acked-by: Lad, Prabhakar <prabhakar.csengg@gmail.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] marvell-ccic: MMP_CAMERA no longer builds
Arnd Bergmann [Wed, 28 Jan 2015 22:11:20 +0000 (19:11 -0300)]
[media] marvell-ccic: MMP_CAMERA no longer builds

The mmp ccic driver expects a platform_data structure that does not exist
in the mainline kernel and presumably was changed in a kernel fork, which
leads to build errors now:

media/platform/marvell-ccic/mmp-driver.c: In function 'mmpcam_calc_dphy':
media/platform/marvell-ccic/mmp-driver.c:252:15: error: 'struct mmp_camera_platform_data' has no member named 'dphy3_algo'
  switch (pdata->dphy3_algo) {
               ^
media/platform/marvell-ccic/mmp-driver.c:253:7: error: 'DPHY3_ALGO_PXA910' undeclared (first use in this function)
  case DPHY3_ALGO_PXA910:
       ^
media/platform/marvell-ccic/mmp-driver.c:253:7: note: each undeclared identifier is reported only once for each function it appears in
media/platform/marvell-ccic/mmp-driver.c:257:8: error: 'struct mmp_camera_platform_data' has no member named 'dphy'

This marks the driver as 'BROKEN' but keeps the code around.
Alternatively it could be removed entirely.

Fixes: 05fed81625bf75 ("[media] marvell-ccic: add MIPI support for marvell-ccic driver")

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Jonathan Corbet <corbet@lwn.net>
Cc: Libin Yang <lbyang@marvell.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] staging: bcm2048: Delete an unnecessary check before the function call "video...
Markus Elfring [Mon, 2 Feb 2015 12:20:23 +0000 (09:20 -0300)]
[media] staging: bcm2048: Delete an unnecessary check before the function call "video_unregister_device"

The video_unregister_device() function tests whether its argument is NULL
and then returns immediately. Thus the test around the call is not needed.

This issue was detected by using the Coccinelle software.

Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] media: ti-vpe: Use mem-to-mem ioctl helpers
Prabhakar Lad [Fri, 19 Dec 2014 13:52:04 +0000 (10:52 -0300)]
[media] media: ti-vpe: Use mem-to-mem ioctl helpers

1: Simplify the vpe mem-to-mem driver by using the m2m ioctl
   and vb2 helpers.
2: Align and arranged the v4l2_ioctl_ops.
3: Fixes a typo.
4: Use of_match_ptr() instead of explicitly defining the macro
   to NULL in case CONFIG_OF is not defined.

Signed-off-by: Lad, Prabhakar <prabhakar.csengg@gmail.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] dib8000: upd_demod_gain_period should be u32
Mauro Carvalho Chehab [Tue, 23 Dec 2014 14:21:12 +0000 (11:21 -0300)]
[media] dib8000: upd_demod_gain_period should be u32

X-Patchwork-Delegate: m.chehab@samsung.com
As shown at the code, upd_demod_gain_period is used to write
to two 16-bit registers:
    dib8000_write_word(state, 1946, upd_demod_gain_period & 0xFFFF);
    dib8000_write_word(state, 1947, reg | (1<<14) | ((upd_demod_gain_period >> 16) & 0xFF));

So, it should be declared as u32.

This fixes the following smatch warning:
drivers/media/dvb-frontends/dib8000.c:1282 dib8000_agc_startup() warn: right shifting more than type allows

Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] cx23885: move CI/MAC registration to a separate function
Mauro Carvalho Chehab [Tue, 23 Dec 2014 15:48:07 +0000 (13:48 -0200)]
[media] cx23885: move CI/MAC registration to a separate function

As reported by smatch:
drivers/media/pci/cx23885/cx23885-dvb.c:2080 dvb_register() Function too hairy.  Giving up.

This is indeed a too complex function, with lots of stuff inside.
Breaking this into two functions makes it a little bit less hairy.

Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] media: au0828 - convert to use videobuf2
Shuah Khan [Thu, 29 Jan 2015 16:41:32 +0000 (13:41 -0300)]
[media] media: au0828 - convert to use videobuf2

Convert au0828 to use videobuf2. Tested with NTSC.
Tested video and vbi devices with xawtv, tvtime,
and vlc. Ran v4l2-compliance to ensure there are
no failures.

Video compliance test results summary:
Total: 75, Succeeded: 75, Failed: 0, Warnings: 18

Vbi compliance test results summary:
Total: 75, Succeeded: 75, Failed: 0, Warnings: 0

Signed-off-by: Shuah Khan <shuahkh@osg.samsung.com>
Reviewed-by: Lad, Prabhakar <prabhakar.csengg@gmail.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] media: fix au0828 compile error from au0828_boards initialization
Shuah Khan [Tue, 13 Jan 2015 02:56:55 +0000 (23:56 -0300)]
[media] media: fix au0828 compile error from au0828_boards initialization

au0828 picked up UNSET from videobuf-core.h and fails to compile
if videobuf-core.h isn't included. Change it to use -1U instead
to fix the problem.

    drivers/media/usb/au0828/au0828-cards.c:47:17: error: â€˜UNSET’ undeclared here (not in a function)
       .tuner_type = UNSET,

Signed-off-by: Shuah Khan <shuahkh@osg.samsung.com>
Acked-by: Lad, Prabhakar <prabhakar.csengg@gmail.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] media: fix au0828_analog_register() to not free au0828_dev
Shuah Khan [Sun, 21 Dec 2014 03:24:48 +0000 (00:24 -0300)]
[media] media: fix au0828_analog_register() to not free au0828_dev

au0828_analog_register() frees au0828_dev when it fails to
locate isoc endpoint. au0828_usb_probe() continues with dvb
and rc probe and registration assuming dev is still valid.
When au0828_analog_register() fails to locate isoc endpoint,
it should return without free'ing au0828_dev. Otherwise, the
probe will fail as dev is null when au0828_dvb_register() is
called.

Signed-off-by: Shuah Khan <shuahkh@osg.samsung.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] Add MAINTAINERS entry for the adv7180
Lars-Peter Clausen [Fri, 23 Jan 2015 15:52:34 +0000 (12:52 -0300)]
[media] Add MAINTAINERS entry for the adv7180

Add myself as the maintainer for the adv7180 video subdev driver.

Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Acked-by: Federico Vaga <federico.vaga@gmail.com>
Acked-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] adv7180: Add fast switch support
Lars-Peter Clausen [Fri, 23 Jan 2015 15:52:33 +0000 (12:52 -0300)]
[media] adv7180: Add fast switch support

In fast switch mode the adv7180 (and similar) can lock onto a new signal
faster when switching between different inputs. As a downside though it is
no longer able to auto-detect the incoming format.

The fast switch mode is exposed as a boolean v4l control that allows
userspace applications to either enable or disable fast switch mode.

Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Acked-by: Federico Vaga <federico.vaga@gmail.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] adv7180: Add I2P support
Lars-Peter Clausen [Fri, 23 Jan 2015 15:52:32 +0000 (12:52 -0300)]
[media] adv7180: Add I2P support

Some of the devices supported by this driver have a interlaced-to-
progressive converter which can optionally be enabled. This patch adds
support for enabling and disabling the I2P converter on such devices.

I2P mode can be enabled by selecting V4L2_FIELD_NONE instead of
V4L2_FIELD_INTERLACED for the format.

Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Acked-by: Federico Vaga <federico.vaga@gmail.com>
Acked-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] adv7180: Add support for the adv7280-m/adv7281-m/adv7281-ma/adv7282-m
Lars-Peter Clausen [Fri, 23 Jan 2015 15:52:31 +0000 (12:52 -0300)]
[media] adv7180: Add support for the adv7280-m/adv7281-m/adv7281-ma/adv7282-m

This patch adds support for the adv7280-m/adv2781-m/adv7281-ma/adv7282-m
devices to the adv7180 driver. They are very similar to the
adv7280/adv7281/adv7282 but instead of parallel video out they feature a
MIPI CSI2 transmitter.

The CSI2 transmitter is configured via a separate I2C address, so we need to
register a dummy device for it.

Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Acked-by: Federico Vaga <federico.vaga@gmail.com>
Acked-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] adv7180: Add support for the adv7280/adv7281/adv7282
Lars-Peter Clausen [Fri, 23 Jan 2015 15:52:30 +0000 (12:52 -0300)]
[media] adv7180: Add support for the adv7280/adv7281/adv7282

This patch adds support for the adv7280/adv7281/adv7282 devices to the
adv7180 driver. They are very similar to the adv7182, the main difference
from the drivers point of view are some different tuning constants for
improved video performance.

Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Acked-by: Federico Vaga <federico.vaga@gmail.com>
Acked-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] adv7180: Add support for the adv7182
Lars-Peter Clausen [Fri, 23 Jan 2015 15:52:29 +0000 (12:52 -0300)]
[media] adv7180: Add support for the adv7182

This patch adds support for the adv7182 to the adv7180 driver. The adv7182
is similar to the adv7180, the main difference from the driver's point of
view is how the video input and how the input format are selected.

Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Acked-by: Federico Vaga <federico.vaga@gmail.com>
Acked-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] adv7180: Prepare for multi-chip support
Lars-Peter Clausen [Fri, 23 Jan 2015 15:52:28 +0000 (12:52 -0300)]
[media] adv7180: Prepare for multi-chip support

The adv7180 is part of a larger family of device, which have all a very
similar register map layout. This patch prepares the adv7180 driver for
support for multiple different devices. For now the only difference we care
about is the number of input channel configurations. Also the way the input
format is configured slightly differs between some devices.

Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Acked-by: Federico Vaga <federico.vaga@gmail.com>
Acked-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] adv7180: Consolidate video mode setting
Lars-Peter Clausen [Fri, 23 Jan 2015 15:52:27 +0000 (12:52 -0300)]
[media] adv7180: Consolidate video mode setting

We have basically the same code to set the video standard in init_device()
and adv7180_s_std(). Factor this out into a common helper function.

Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Acked-by: Federico Vaga <federico.vaga@gmail.com>
Acked-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] adv7180: Add media controller support
Lars-Peter Clausen [Fri, 23 Jan 2015 15:52:26 +0000 (12:52 -0300)]
[media] adv7180: Add media controller support

Add media controller support to the adv7180 driver by registering a media
entity instance for it as well as implementing pad ops for configuring the
format.

As there currently don't seem to be any users of the video ops format
operations those are removed as well in this patch.

Also set the V4L2_SUBDEV_FL_HAS_DEVNODE flag for the subdevice so it is
possible to create a subdevice device node.

Since the driver now depends on VIDEO_V4L2_SUBDEV_API all drivers which
select the driver need to depend on that symbol as well.

Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Acked-by: Federico Vaga <federico.vaga@gmail.com>
Acked-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] adv7180: Reset the device before initialization
Lars-Peter Clausen [Fri, 23 Jan 2015 15:52:25 +0000 (12:52 -0300)]
[media] adv7180: Reset the device before initialization

Reset the device when initializing it so it is in a good known state and the
assumed register settings matches the actual register settings.

Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Acked-by: Federico Vaga <federico.vaga@gmail.com>
Acked-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] adv7180: Do implicit register paging
Lars-Peter Clausen [Fri, 23 Jan 2015 15:52:24 +0000 (12:52 -0300)]
[media] adv7180: Do implicit register paging

The ad7180 has multiple register pages which can be switched between by
writing to a register. Currently the driver manually switches between pages
whenever a register outside of the default register map is accessed and
switches back after it has been accessed. This is a bit tedious and also
potential source for bugs.

This patch adds two helper functions that take care of switching between
pages and reading/writing the register. The register numbers for registers
are updated to encode both the page (in the upper 8-bits) and the register
(in the lower 8-bits) numbers.

Having multiple pages means that a register access is not a single atomic
i2c_smbus_write_byte_data() or i2c_smbus_read_byte_data() call and we need
to make sure that concurrent register access does not race against each
other.

Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Acked-by: Federico Vaga <federico.vaga@gmail.com>
Acked-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] adv7180: Cleanup register define naming
Lars-Peter Clausen [Fri, 23 Jan 2015 15:52:23 +0000 (12:52 -0300)]
[media] adv7180: Cleanup register define naming

Consistently prefix register defines with ADV7180_REG. Also remove the "ADI"
from register names, the ADV7180 prefix should provide enough of a namespace
separation.

Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Acked-by: Federico Vaga <federico.vaga@gmail.com>
Acked-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] adv7180: Pass correct flags to request_threaded_irq()
Lars-Peter Clausen [Fri, 23 Jan 2015 15:52:21 +0000 (12:52 -0300)]
[media] adv7180: Pass correct flags to request_threaded_irq()

Most IRQ controllers support different types of interrupts. The adv7180
generates falling edge interrupts, so make sure to pass IRQF_TRIGGER_FALLING
to request_threaded_irq() so the IRQ controller is configured for the
correct mode.

Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Acked-by: Federico Vaga <federico.vaga@gmail.com>
Acked-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] adv7180: Do not request the IRQ again during resume
Lars-Peter Clausen [Fri, 23 Jan 2015 15:52:20 +0000 (12:52 -0300)]
[media] adv7180: Do not request the IRQ again during resume

Currently the IRQ is requested from within the init_device() function. This
function is not only called during device probe, but also during resume
causing the driver to try to request the IRQ again. Move requesting the IRQ
from init_device() to the probe function to make sure that it is only
requested once.

Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Acked-by: Federico Vaga <federico.vaga@gmail.com>
Acked-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] davinci: add V4L2 dependencies
Arnd Bergmann [Thu, 29 Jan 2015 16:12:05 +0000 (13:12 -0300)]
[media] davinci: add V4L2 dependencies

The davinci media drivers use videobuf2, which they enable through
a 'select' statement. If one of these drivers is built-in, but
the v4l2 core is a loadable modules, we end up with a link
error:

drivers/built-in.o: In function `vb2_fop_mmap':
:(.text+0x113e84): undefined reference to `video_devdata'
drivers/built-in.o: In function `vb2_ioctl_create_bufs':
:(.text+0x114710): undefined reference to `video_devdata'
drivers/built-in.o: In function `vb2_ioctl_reqbufs':
:(.text+0x114ed8): undefined reference to `video_devdata'
drivers/built-in.o: In function `vb2_ioctl_querybuf':
:(.text+0x115530): undefined reference to `video_devdata'

To solve this, we need to add a dependency on VIDEO_V4L2,
which enforces that the davinci drivers themselves can only
be loadable modules if V4L2 is not built-in, and they do
not cause the videobuf2 code to be built-in.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Lad, Prabhakar <prabhakar.csengg@gmail.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] vivid: use consistent colorspace/Y'CbCr Encoding strings
Hans Verkuil [Thu, 29 Jan 2015 11:22:37 +0000 (08:22 -0300)]
[media] vivid: use consistent colorspace/Y'CbCr Encoding strings

Keep the colorspace and encoding names consistent with what is
used elsewhere (primarily the utilities in v4l-utils.git).

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] siano: fix Kconfig dependencies
Arnd Bergmann [Wed, 28 Jan 2015 21:17:44 +0000 (18:17 -0300)]
[media] siano: fix Kconfig dependencies

The USB and MMC front-ends to the siano driver both only make
sense when combined with the SMS_SIANO_MDTV driver. That driver
already requires RC_CORE to not be a module, so we also need
to add that dependency here.

drivers/built-in.o: In function `smssdio_remove':
:(.text+0x155bd8): undefined reference to `smscore_putbuffer'
:(.text+0x155bdc): undefined reference to `smscore_unregister_device'
drivers/built-in.o: In function `smssdio_interrupt':
:(.text+0x155e4c): undefined reference to `smsendian_handle_rx_message'
:(.text+0x155e50): undefined reference to `smscore_onresponse'
:(.text+0x155e54): undefined reference to `smscore_getbuffer'
:(.text+0x155e58): undefined reference to `smscore_putbuffer'
drivers/built-in.o: In function `smssdio_sendrequest':
:(.text+0x155f20): undefined reference to `smsendian_handle_tx_message'
drivers/built-in.o: In function `smssdio_probe':
:(.text+0x15610c): undefined reference to `sms_get_board'
:(.text+0x156114): undefined reference to `smscore_register_device'
:(.text+0x156118): undefined reference to `smscore_set_board_id'
:(.text+0x156128): undefined reference to `smscore_unregister_device'
:(.text+0x156140): undefined reference to `smscore_start_device'

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] radio/aimslab: use mdelay instead of udelay
Arnd Bergmann [Wed, 28 Jan 2015 21:17:42 +0000 (18:17 -0300)]
[media] radio/aimslab: use mdelay instead of udelay

Large udelay values are not allowed on the ARM architecture
and result in a build error like

ERROR: "__bad_udelay" [drivers/media/radio/radio-aimslab.ko] undefined!

This changes the aimslab radio driver to use an equivalent mdelay
statement.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Cc: Hans Verkuil <hverkuil@xs4all.nl>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] timberdale: do not select TIMB_DMA
Arnd Bergmann [Wed, 28 Jan 2015 21:17:41 +0000 (18:17 -0300)]
[media] timberdale: do not select TIMB_DMA

The timberdale media driver requires the use of the respective
dma engine driver, but that may not be enabled, causing a
Kconfig warning:

warning: (VIDEO_TIMBERDALE) selects TIMB_DMA which has unmet direct dependencies (DMADEVICES && MFD_TIMBERDALE)

This fixes the dependency by removing the inappropriate 'select'
statement and replacing it with a direct dependency on the
drivers that provide the services this needs.

Fixes: 7155043c2d027 ("[media] enable COMPILE_TEST for media drivers")

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] media: am437x: fix sparse warnings
Prabhakar Lad [Mon, 26 Jan 2015 14:50:15 +0000 (11:50 -0300)]
[media] media: am437x: fix sparse warnings

This patch fixes following spare warnings:

drivers/media/platform/am437x/am437x-vpfe.c:66:28: warning: symbol 'vpfe_standards' was not declared. Should it be static?
drivers/media/platform/am437x/am437x-vpfe.c:2202:57: warning: incorrect type in argument 2 (different address spaces)
drivers/media/platform/am437x/am437x-vpfe.c:2202:57:    expected void [noderef] <asn:1>*params
drivers/media/platform/am437x/am437x-vpfe.c:2202:57:    got void *param
include/linux/spinlock.h:364:9: warning: context imbalance in 'vpfe_start_streaming' - unexpected unlock

Reported-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Lad, Prabhakar <prabhakar.csengg@gmail.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] pvrusb2: use msecs_to_jiffies for conversion
Nicholas Mc Guire [Thu, 22 Jan 2015 10:39:11 +0000 (07:39 -0300)]
[media] pvrusb2: use msecs_to_jiffies for conversion

This is only an API consolidation and should make things more readable

Signed-off-by: Nicholas Mc Guire <der.herr@hofr.at>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] MAINTAINERS: Update solo6x10 entry
Ismael Luceno [Tue, 20 Jan 2015 14:43:50 +0000 (11:43 -0300)]
[media] MAINTAINERS: Update solo6x10 entry

Re-add Ismael Luceno as co-maintainer (with personal email address).

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@osg.samsung.com>
9 years ago[media] cx231xx: don't use dev it not allocated
Mauro Carvalho Chehab [Mon, 2 Feb 2015 12:52:26 +0000 (10:52 -0200)]
[media] cx231xx: don't use dev it not allocated

changeset 5eeb3014827f added a fixup at the error check
code. However, it introduced a new error:

drivers/media/usb/cx231xx/cx231xx-cards.c:1586 cx231xx_usb_probe() error: we previously assumed 'dev' could be null (see line 1430)

This happens when dev = kmalloc() fails. So, instead of relying
on it to succeed, just change the parameter of clear_bit() from
'dev->devno' to 'nr'.

Cc: Alexey Khoroshilov <khoroshilov@ispras.ru>
Cc: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] cx231xx: fix usbdev leak on failure paths in cx231xx_usb_probe()
Alexey Khoroshilov [Fri, 16 Jan 2015 22:55:26 +0000 (19:55 -0300)]
[media] cx231xx: fix usbdev leak on failure paths in cx231xx_usb_probe()

Commit b7085c086475 ("cx231xx: convert from pr_foo to dev_foo")
moves usb_get_dev(interface_to_usbdev(interface)) to the beginning
of cx231xx_usb_probe() to use udev->dev in dev_err(),
but it does not make sure usbdev is put on all failure paths.

Later dev_err(udev->dev) was replaced by dev_err(d).
So the patch moves usb_get_dev() below (before the first use)
and fixes another failure path.

Found by Linux Driver Verification project (linuxtesting.org).

Signed-off-by: Alexey Khoroshilov <khoroshilov@ispras.ru>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] rcar_vin: move buffer management to .stop_streaming handler
William Towle [Mon, 26 Jan 2015 17:08:40 +0000 (14:08 -0300)]
[media] rcar_vin: move buffer management to .stop_streaming handler

This commit moves the "buffer in use" logic from the .buf_cleanup
handler into .stop_streaming, based on advice that this is its
proper logical home.

By ensuring the list of pointers in priv->queue_buf[] is managed
as soon as possible, we avoid warnings concerning buffers in ACTIVE
state when the system cleans up after streaming stops. This fixes a
problem with modification of buffers after their content has been
cleared for passing to userspace.

After the refactoring, the buf_init and buf_cleanup functions were
found to contain only initialisation/release steps as are carried out
elsewhere if omitted; these functions and references were removed.

Signed-off-by: William Towle <william.towle@codethink.co.uk>
Signed-off-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] rcar_vin: helper function for streaming stop
Ian Molton [Mon, 26 Jan 2015 17:08:39 +0000 (14:08 -0300)]
[media] rcar_vin: helper function for streaming stop

The code that tests that capture from a stream has stopped is
presently insufficient and the potential for a race condition
exists where frame capture may generate an interrupt between
requesting the capture process halt and freeing buffers.

This patch refactors code out of rcar_vin_videobuf_release() and
into rcar_vin_wait_stop_streaming(), and ensures there are calls
in places where we need to know that capturing has finished.

Signed-off-by: Ian Molton <ian.molton@codethink.co.uk>
Signed-off-by: William Towle <william.towle@codethink.co.uk>
Signed-off-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] ov2640: use the v4l2 size definitions
Josh Wu [Tue, 25 Nov 2014 08:54:28 +0000 (05:54 -0300)]
[media] ov2640: use the v4l2 size definitions

Reuse the v4l2 size definitions from v4l2-image-sizes.h.
So we can remove the rudundent definitions from ov2640.c.

Signed-off-by: Josh Wu <josh.wu@atmel.com>
Signed-off-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] soc-camera: remove redundant code
Guennadi Liakhovetski [Sun, 1 Feb 2015 15:06:17 +0000 (12:06 -0300)]
[media] soc-camera: remove redundant code

A hunk, removing this code has been lost between versions 6 and 7 of
patch "soc-camera: add V4L2-async support." The code is harmless, but
redundant.

Signed-off-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] soc_camera: use vb2_ops_wait_prepare/finish helper
Lad, Prabhakar [Wed, 26 Nov 2014 22:42:27 +0000 (19:42 -0300)]
[media] soc_camera: use vb2_ops_wait_prepare/finish helper

This patch drops driver specific wait_prepare() and
wait_finish() callbacks from vb2_ops and instead uses
the the helpers vb2_ops_wait_prepare/finish() provided
by the vb2 core, the lock member of the queue needs
to be initalized to a mutex so that vb2 helpers
vb2_ops_wait_prepare/finish() can make use of it.

Signed-off-by: Lad, Prabhakar <prabhakar.csengg@gmail.com>
Cc: Josh Wu <josh.wu@atmel.com>
Cc: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Signed-off-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] coda: simplify check in coda_buf_queue
Philipp Zabel [Fri, 23 Jan 2015 16:51:35 +0000 (13:51 -0300)]
[media] coda: simplify check in coda_buf_queue

Now that the bitstream buffer is only allocated for the BIT decoder
case, we can use bitstream.size to check for bitstream ringbuffer
operation.

Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] coda: allocate bitstream ringbuffer only for BIT decoder
Philipp Zabel [Fri, 23 Jan 2015 16:51:34 +0000 (13:51 -0300)]
[media] coda: allocate bitstream ringbuffer only for BIT decoder

The BIT encoder does not use a per-context bitstream ringbuffer as it encodes
directly into the videobuf2 capture queue's buffers. Avoid allocation of the
bitstream ringbuffer for encoder contexts.

Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] coda: add support for contexts that do not use the BIT processor
Philipp Zabel [Fri, 23 Jan 2015 16:51:33 +0000 (13:51 -0300)]
[media] coda: add support for contexts that do not use the BIT processor

In preparation for CODA9 JPEG support, allow contexts that
control hardware units directly, without the BIT processor.

Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] coda: free context buffers under buffer mutex
Philipp Zabel [Fri, 23 Jan 2015 16:51:32 +0000 (13:51 -0300)]
[media] coda: free context buffers under buffer mutex

Make sure the buffer_mutex lock is taken in coda_bit_release
while coda_free_framebuffers and coda_free_context_buffers
are called.

Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] coda: make seq_end_work optional
Philipp Zabel [Fri, 23 Jan 2015 16:51:31 +0000 (13:51 -0300)]
[media] coda: make seq_end_work optional

In preparation for CODA9 JPEG support, which doesn't have to call
SEQ_END on the BIT processor.

Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] coda: switch BIT decoder source queue to vmalloc
Philipp Zabel [Fri, 23 Jan 2015 16:51:30 +0000 (13:51 -0300)]
[media] coda: switch BIT decoder source queue to vmalloc

Since we have to copy from input buffers into the bitstream ringbuffer
with the CPU, there is no need for contiguous DMA buffers on the decoder
input side.

Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] coda: add coda_estimate_sizeimage and use it in set_defaults
Philipp Zabel [Fri, 23 Jan 2015 16:51:29 +0000 (13:51 -0300)]
[media] coda: add coda_estimate_sizeimage and use it in set_defaults

Call coda_estimate_sizeimage from both try_fmt and set_defaults to
avoid this v4l2-compliance warning on the h.264 decoder video device:

    G_FMT:     1920x1088, 32315559, 1, 1920, 1048576, 3, 0, 0, feedcafe
    TRY/S_FMT: 1920x1088, 32315559, 1, 1920, 3133440, 3, 0, 0, feedcafe
    fail: v4l2-test-formats.cpp(948): Video Capture: S_FMT(G_FMT) != G_FMT

Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] coda: don't ever use subsampling ping-pong buffers as reconstructed reference...
Philipp Zabel [Fri, 23 Jan 2015 16:51:28 +0000 (13:51 -0300)]
[media] coda: don't ever use subsampling ping-pong buffers as reconstructed reference buffers

On i.MX6, two subsampling ping-pong buffers are used for motion estimation and
deblocking They should not be counted as framebuffers, or they will be also used
to store reconstructed frames, causing visible artifacts in P-frames.

Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] coda: issue seq_end_work during stop_streaming
Philipp Zabel [Fri, 23 Jan 2015 16:51:27 +0000 (13:51 -0300)]
[media] coda: issue seq_end_work during stop_streaming

This patch queues seq_end_work and flushes the queue during stop_streaming
and clears the ctx->initialized flag. This allows to start streaming again
after stopping streaming without releasing the context.

Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] coda: remove unused isequence, reset qsequence in stop_streaming
Philipp Zabel [Fri, 23 Jan 2015 16:51:26 +0000 (13:51 -0300)]
[media] coda: remove unused isequence, reset qsequence in stop_streaming

The isequence counter is never used, qsequence counts the buffers queued into
the bit decoder bitstream ringbuffer. It needs to be reset in stop_streaming.

Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] coda: fix width validity check when starting to decode
Markus Pargmann [Fri, 23 Jan 2015 16:51:25 +0000 (13:51 -0300)]
[media] coda: fix width validity check when starting to decode

Compare rounded up width to fit into bytesperline.

Signed-off-by: Markus Pargmann <mpa@pengutronix.de>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] coda: clear RET_DEC_PIC_SUCCESS flag in prepare_decode
Philipp Zabel [Fri, 23 Jan 2015 16:51:24 +0000 (13:51 -0300)]
[media] coda: clear RET_DEC_PIC_SUCCESS flag in prepare_decode

To make sure a set RET_DEC_PIC_SUCCESS flag is not a leftover from
a previous successful run, clear it in prepare_decode.

Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] coda: initialize SRAM on probe
Philipp Zabel [Fri, 23 Jan 2015 16:51:23 +0000 (13:51 -0300)]
[media] coda: initialize SRAM on probe

Zeroing the SRAM on probe helps with debugging SRAM contents.

Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] coda: properly clear f_cap in coda_s_fmt_vid_out
Philipp Zabel [Fri, 23 Jan 2015 16:51:22 +0000 (13:51 -0300)]
[media] coda: properly clear f_cap in coda_s_fmt_vid_out

Properly zero the structure on the stack before using it.

Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] coda: fix try_fmt_vid_out colorspace setting
Philipp Zabel [Fri, 23 Jan 2015 16:51:21 +0000 (13:51 -0300)]
[media] coda: fix try_fmt_vid_out colorspace setting

v4l2-compliance complains about invalid colorspace settings being accepted
on the output side. This patch only allows REC709 and JPEG.

Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] coda: fix job_ready debug reporting for bitstream decoding
Philipp Zabel [Fri, 23 Jan 2015 16:51:20 +0000 (13:51 -0300)]
[media] coda: fix job_ready debug reporting for bitstream decoding

Clarify whether job_ready returns false because the context is on hold, waiting
for new input buffers, whether there are not enough input buffers to fill two
into the bitstream, or whether there is not enough data in the bitstream buffer
for the bitstream reader hardware to read a whole frame.

Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] coda: adjust sequence offset after unexpected decoded frame
Lucas Stach [Fri, 23 Jan 2015 16:51:19 +0000 (13:51 -0300)]
[media] coda: adjust sequence offset after unexpected decoded frame

If userspace doesn't properly separate the bitstream input into
individual frames (which may happen for example on slightly
corrupted streams) the CODA hardware may decode more frames
than we expect. We already log an error in this case, but it's
also necessary to adjust the sequence offset. Otherwise we
spam the log with a sequence number mismatch on every frame
frame after the unexpected one.

Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] coda: move meta out of padding
Philipp Zabel [Fri, 23 Jan 2015 16:51:18 +0000 (13:51 -0300)]
[media] coda: move meta out of padding

Handle an empty buffer metadata list without crashing. This can happen
if the decoder is fed a broken stream, or multiple compressed frames in
a single queued buffer.

Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] coda: remove context debugfs entry last
Philipp Zabel [Fri, 23 Jan 2015 16:51:17 +0000 (13:51 -0300)]
[media] coda: remove context debugfs entry last

Do not remove the per-context debugfs directory before the
per-buffer debugfs entries contained therein.

Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] coda: fix encoder rate control parameter masks
Philipp Zabel [Fri, 23 Jan 2015 16:51:15 +0000 (13:51 -0300)]
[media] coda: fix encoder rate control parameter masks

This patch fixes the ENC_SEQ_RC_PARA initial delay and bitrate masks.
These bit fields are 15 bit wide, not 7 bit.

Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] s5p-mfc: remove unnecessary version.h inclusion
Fabian Frederick [Mon, 29 Dec 2014 14:29:41 +0000 (11:29 -0300)]
[media] s5p-mfc: remove unnecessary version.h inclusion

Based on versioncheck.

Signed-off-by: Fabian Frederick <fabf@skynet.be>
Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] s5p-g2d: remove unnecessary version.h inclusion
Fabian Frederick [Mon, 29 Dec 2014 14:29:40 +0000 (11:29 -0300)]
[media] s5p-g2d: remove unnecessary version.h inclusion

Based on versioncheck.

Signed-off-by: Fabian Frederick <fabf@skynet.be>
Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] media-doc: Fix MFC display delay control doc
Nicolas Dufresne [Mon, 15 Dec 2014 21:10:59 +0000 (18:10 -0300)]
[media] media-doc: Fix MFC display delay control doc

The V4L2_CID_MPEG_MFC51_VIDEO_DECODER_H264_DISPLAY_DELAY_ENABLE control
is a boolean but was documented as a integer. The documentation was
also slightly miss-leading.

Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Acked-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] s5p-mfc-dec: Don't use encoder stop command
Nicolas Dufresne [Mon, 15 Dec 2014 21:10:58 +0000 (18:10 -0300)]
[media] s5p-mfc-dec: Don't use encoder stop command

The decoder should handle V4L2_DEC_CMD_STOP to trigger drain,
but it currently expecting V4L2_ENC_CMD_STOP.

Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Acked-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] s5p-mfc-v6+: Use display_delay_enable CID
Nicolas Dufresne [Mon, 15 Dec 2014 21:10:57 +0000 (18:10 -0300)]
[media] s5p-mfc-v6+: Use display_delay_enable CID

The MFC driver has two controls, DISPLAY_DELAY and DISPLAY_DELAY_ENABLE
that allow forcing the decoder to return a decoded frame sooner
regardless of the order. The added support for firmware version 6 and
higher was not taking into account the DISPLAY_DELAY_ENABLE boolean.
Instead it had a comment stating that DISPLAY_DELAY should be set to a
negative value to disable it. This is not possible since the control
range is from 0 to 65535. This feature was also supposed to be disabled
by default in order to produce frames in display order.

Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Acked-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] Documentation/video4linux: remove obsolete text files
Hans Verkuil [Tue, 23 Dec 2014 12:20:53 +0000 (09:20 -0300)]
[media] Documentation/video4linux: remove obsolete text files

Remove obsolete text files for drivers that have been removed.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] bw/c-qcam, w9966, pms: remove deprecated staging drivers
Hans Verkuil [Tue, 23 Dec 2014 12:17:09 +0000 (09:17 -0300)]
[media] bw/c-qcam, w9966, pms: remove deprecated staging drivers

These drivers haven't been tested in a long, long time. The hardware is
ancient and hopelessly obsolete. These drivers also need to be converted
to newer media frameworks but due to the lack of hardware that's going
to be impossible. In addition, cheaper and vastly better hardware is
available today.

These drivers are already deprecated, so now remove them altogether.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] vino/saa7191: remove deprecated drivers
Hans Verkuil [Tue, 23 Dec 2014 12:14:37 +0000 (09:14 -0300)]
[media] vino/saa7191: remove deprecated drivers

These drivers haven't been tested in a long, long time. The hardware is
ancient and hopelessly obsolete. These drivers also need to be converted
to newer media frameworks but due to the lack of hardware that's going
to be impossible.

So these drivers are a prime candidate for removal. If someone is
interested in working on these drivers to prevent their removal, then
please contact the linux-media mailinglist.

These drivers are already deprecated, so now remove them altogether.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] tlg2300: remove deprecated staging driver
Hans Verkuil [Tue, 23 Dec 2014 12:11:48 +0000 (09:11 -0300)]
[media] tlg2300: remove deprecated staging driver

This driver hasn't been tested in a long, long time. The company that made
this chip has gone bust many years ago and hardware using this chip is next
to impossible to find.

This driver needs to be converted to newer media frameworks but due to the
lack of hardware that's going to be impossible. Since cheap alternatives are
easily available, there is little point in keeping this driver alive.

This driver is already deprecated, so now remove it altogether.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Cc: Huang Shijie <shijie8@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] si2157: change firmware variable name and type
Antti Palosaari [Sat, 6 Dec 2014 21:15:18 +0000 (18:15 -0300)]
[media] si2157: change firmware variable name and type

Rename firmware variable from fw_file to fw_name and change its
type from u8 to const char as request_firmware() input is.

Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] si2157: print chip version
Antti Palosaari [Sat, 6 Dec 2014 21:12:39 +0000 (18:12 -0300)]
[media] si2157: print chip version

Print chip version once using log level into when init() is called.
Remove cold/warm state printing as those are not very useful.

Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] si2157: print firmware version
Antti Palosaari [Sat, 6 Dec 2014 20:25:24 +0000 (17:25 -0300)]
[media] si2157: print firmware version

Firmware version could be printed similarly than si2168 driver does.

Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] si2157: add own goto label for kfree() on probe error
Antti Palosaari [Sat, 6 Dec 2014 18:13:31 +0000 (15:13 -0300)]
[media] si2157: add own goto label for kfree() on probe error

Use own goto label for error case mem free is needed, even kfree could
be called with NULL. I think it is better to have it, even not required.

Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] si2157: trivial ID table changes
Antti Palosaari [Sat, 6 Dec 2014 18:07:45 +0000 (15:07 -0300)]
[media] si2157: trivial ID table changes

- Rename ID table.
- Remove magic numbers from ID table driver data field.

Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] si2157: change firmware download error handling
Antti Palosaari [Sat, 6 Dec 2014 17:51:17 +0000 (14:51 -0300)]
[media] si2157: change firmware download error handling

Rename firmare download error path goto label. Remove firmware NULL
set as NULL value is not needed anymore, due to recent change which
started using goto labels for firmware error handling.

Cc: Olli Salonen <olli.salonen@iki.fi>
Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] si2157: carry pointer to client instead of state in tuner_priv
Antti Palosaari [Sat, 6 Dec 2014 17:40:06 +0000 (14:40 -0300)]
[media] si2157: carry pointer to client instead of state in tuner_priv

Carry struct i2c_client pointer in tuner_priv. This driver is I2C driver,
which is represented as a struct i2c_client, so better to carry this top
level structure for each routine in order to unify things.

Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] si2157: simplify si2157_cmd_execute() error path
Antti Palosaari [Sat, 6 Dec 2014 17:04:05 +0000 (14:04 -0300)]
[media] si2157: simplify si2157_cmd_execute() error path

Remove if () from firmware command error path as there should not be
any error prone conditional logic there. Use goto labels instead.

Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] si2157: rename device state variable from 's' to 'dev'
Antti Palosaari [Fri, 5 Dec 2014 20:46:30 +0000 (17:46 -0300)]
[media] si2157: rename device state variable from 's' to 'dev'

'dev' is likely most common name in kernel for structure containing
device state instance, so rename it in order to keep things
consistent.

Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] si2168: change firmware variable name and type
Antti Palosaari [Sat, 6 Dec 2014 20:53:16 +0000 (17:53 -0300)]
[media] si2168: change firmware variable name and type

Rename firmware variable from fw_file to fw_name and change its type
from u8 to const char as request_firmware() input defines.

Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] si2168: print chip version
Antti Palosaari [Sat, 6 Dec 2014 20:43:27 +0000 (17:43 -0300)]
[media] si2168: print chip version

Print chip version once using log level into when init() is called.
Remove cold/warm state printing as those are not very useful.

old printing:
si2168 6-0064: found a 'Silicon Labs Si2168' in cold state
si2168 6-0064: downloading firmware from file 'dvb-demod-si2168-b40-01.fw'
si2168 6-0064: firmware version: 4.0.11
si2168 6-0064: found a 'Silicon Labs Si2168' in warm state

new printing:
si2168 6-0064: found a 'Silicon Labs Si2168-B40'
si2168 6-0064: downloading firmware from file 'dvb-demod-si2168-b40-01.fw'
si2168 6-0064: firmware version: 4.0.11

Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] si2168: remove unneeded fw variable initialization
Antti Palosaari [Fri, 5 Dec 2014 20:22:42 +0000 (17:22 -0300)]
[media] si2168: remove unneeded fw variable initialization

commit 034e1ec0ce299b9e90056793dcb3187e7add6b62
si2168: One function call less in si2168_init() after error detection

That commit added goto label for error path to release firmware,
but forgets to remove variable NULL set. Remove those now.

Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] si2168: enhance firmware download routine
Antti Palosaari [Fri, 5 Dec 2014 20:08:22 +0000 (17:08 -0300)]
[media] si2168: enhance firmware download routine

All known old firmware firmware formats are downloaded using 8 byte
chunks. Reject firmware if it could not be divided to 8 byte chunks
and because of that we could simplify some calculations. Now both
supported firmware download routines are rather similar.

Cc: Olli Salonen <olli.salonen@iki.fi>
Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] si2168: add own goto label for kzalloc failure
Antti Palosaari [Fri, 5 Dec 2014 19:02:42 +0000 (16:02 -0300)]
[media] si2168: add own goto label for kzalloc failure

Use own label for kzalloc failure in which does not call kfree().
kfree() could be called with NULL, but it is still better to have
own label which skips unnecessary kfree().

Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] si2168: change stream id debug log formatter
Antti Palosaari [Fri, 5 Dec 2014 18:57:03 +0000 (15:57 -0300)]
[media] si2168: change stream id debug log formatter

Change formatter from signed to unsigned as stream_id is 32bit
unsigned variable.

Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] si2168: change firmware version print from debug to info
Antti Palosaari [Fri, 5 Dec 2014 18:04:12 +0000 (15:04 -0300)]
[media] si2168: change firmware version print from debug to info

Even firmware version is not needed to know, it is still interesting
and useful to know some cases. Due to that increase its printing to
info log level.

Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] si2168: rename few things
Antti Palosaari [Fri, 5 Dec 2014 17:54:14 +0000 (14:54 -0300)]
[media] si2168: rename few things

Rename some goto labels and more. No functionality changes.

Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] si2168: simplify si2168_cmd_execute() error path
Antti Palosaari [Fri, 5 Dec 2014 17:30:44 +0000 (14:30 -0300)]
[media] si2168: simplify si2168_cmd_execute() error path

Remove if () from firmware command error path as there should not be
any error prone conditional logic there. Use goto labels instead.

Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] si2168: get rid of own struct i2c_client pointer
Antti Palosaari [Tue, 25 Nov 2014 20:53:21 +0000 (17:53 -0300)]
[media] si2168: get rid of own struct i2c_client pointer

We don't need that anymore as same pointer is passed to each
routine via struct dvb_frontend private field.

Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] si2168: carry pointer to client instead of state
Antti Palosaari [Tue, 25 Nov 2014 20:31:43 +0000 (17:31 -0300)]
[media] si2168: carry pointer to client instead of state

Carry struct i2c_client pointer inside struct dvb_frontend private
pointer. This driver is I2C driver, which is represented as a
struct i2c_client, so better to carry this top level structure for
each routine in order to unify things. This allows further
simplification.

Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] si2168: rename device state variable from 's' to 'dev'
Antti Palosaari [Tue, 25 Nov 2014 19:46:16 +0000 (16:46 -0300)]
[media] si2168: rename device state variable from 's' to 'dev'

'dev' is most common name in kernel for structure containing device
state instance, so rename it.

Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] si2168: define symbol rate limits
Antti Palosaari [Tue, 25 Nov 2014 19:26:49 +0000 (16:26 -0300)]
[media] si2168: define symbol rate limits

w_scan complains about missing symbol rate limits:
This dvb driver is *buggy*: the symbol rate limits are undefined - please report to linuxtv.org

Chip supports 1 to 7.2 MSymbol/s on DVB-C.

Cc: stable@vger.kernel.org
Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] lgdt3305: add support for fixed tp clock mode
Michael Ira Krufky [Sun, 21 Dec 2014 21:54:50 +0000 (18:54 -0300)]
[media] lgdt3305: add support for fixed tp clock mode

Add support for controlling TP clock mode for VSB and QAM annex-B/C mode.
Gated clock mode is the default value, and does not support QAM annex-C.
The patch enables setting this control to fixed clock mode.

Signed-off-by: Michael Ira Krufky <mkrufky@linuxtv.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] lgdt3305: we only need to pass state into lgdt3305_mpeg_mode_polarity()
Michael Ira Krufky [Sun, 21 Dec 2014 21:48:19 +0000 (18:48 -0300)]
[media] lgdt3305: we only need to pass state into lgdt3305_mpeg_mode_polarity()

Simplify the code a little bit by removing some uneeded arguments
to lgdt3305_mpeg_mode_polarity().

Signed-off-by: Michael Ira Krufky <mkrufky@linuxtv.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] gspca_vc032x: Fix wrong bytesperline
Luca Bonissi [Mon, 26 Jan 2015 10:38:17 +0000 (07:38 -0300)]
[media] gspca_vc032x: Fix wrong bytesperline

I found a problem on vc032x gspca usb webcam subdriver: "bytesperline"
property is wrong for YUYV and YVYU formats.
With recent v4l-utils library (>=0.9.1), that uses "bytesperline" for
pixel format conversion, the result is a wrong jerky image.

Patch tested on my laptop (USB webcam Logitech Orbicam 046d:0892).

BugLink: https://bugzilla.kernel.org/show_bug.cgi?id=91181
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>