platform/kernel/linux-rpi.git
7 years ago[media] ov2640: fix vflip control
Frank Schaefer [Sun, 16 Apr 2017 17:35:45 +0000 (14:35 -0300)]
[media] ov2640: fix vflip control

Enabling vflip currently causes wrong colors.
It seems that (at least with the current sensor setup) REG04_VFLIP_IMG only
changes the vertical readout direction.
Because pixels are arranged RGRG... in odd lines and GBGB... in even lines,
either a one line shift or even/odd line swap is required, too, but
apparently this doesn't happen.

I finally figured out that this can be done manually by setting
REG04_VREF_EN.
Looking at hflip, it turns out that bit REG04_HREF_EN is set there
permanetly, but according to my tests has no effect on the pixel readout
order.
So my conclusion is that the current documentation of sensor register 0x04
is wrong (has changed after preliminary datasheet version 2.2).

I'm pretty sure that automatic vertical line shift/switch can be enabled,
too, but until anyone finds ot how this works, we have to stick with manual
switching.

Signed-off-by: Frank Schäfer <fschaefer.oss@googlemail.com>
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] ov2640: fix duplicate width+height returning from ov2640_select_win()
Frank Schaefer [Sun, 16 Apr 2017 17:35:44 +0000 (14:35 -0300)]
[media] ov2640: fix duplicate width+height returning from ov2640_select_win()

ov2640_select_win() returns height and width values as part of struct
ov2640_win_size, so there is no point in modifying the passed height and
width parameters, too.

Signed-off-by: Frank Schäfer <fschaefer.oss@googlemail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] ov2640: add missing write to size change preamble
Frank Schaefer [Sun, 16 Apr 2017 17:35:43 +0000 (14:35 -0300)]
[media] ov2640: add missing write to size change preamble

HSIZE and VSIZE bits 0 to 2 and HSIZE bit 11 are encoded in DSP register
SIZEL.

Signed-off-by: Frank Schäfer <fschaefer.oss@googlemail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] ov2640: add information about DSP register 0xc7
Frank Schaefer [Sun, 16 Apr 2017 17:35:42 +0000 (14:35 -0300)]
[media] ov2640: add information about DSP register 0xc7

According to ov2640 software application notes, there are two Automatic
White Balance (AWB) modes, which are selected by DSP register 0xc7:

1) Simple AWB: assumes the average color is gray
   + independent from lens
   - doesn't work well if captured area contains unbalanced colors
     (e.g. large blue background)

2) Advanced AWB: uses color temperature information
   + more accurate, works with all image contents
   - lens specific, requires calibration

Signed-off-by: Frank Schäfer <fschaefer.oss@googlemail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] ov2640: improve banding filter register definitions/documentation
Frank Schaefer [Sun, 16 Apr 2017 17:35:41 +0000 (14:35 -0300)]
[media] ov2640: improve banding filter register definitions/documentation

- add missing initialisation of sensor register COM25 (2 MSBs of banding
  filter AEC values)
- add macros for setting the banding filter AEC values
- add definitions for sensor register 0x5a, which is documented in
  Omnivisions software application notes

Signed-off-by: Frank Schäfer <fschaefer.oss@googlemail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] ov2640: fix init sequence alignment
Frank Schaefer [Sun, 16 Apr 2017 17:35:40 +0000 (14:35 -0300)]
[media] ov2640: fix init sequence alignment

While we are at it, remove a misleading comment (copy/paste mistake)

Signed-off-by: Frank Schäfer <fschaefer.oss@googlemail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] ov2640: make GPIOLIB an optional dependency
Mauro Carvalho Chehab [Wed, 19 Apr 2017 11:43:49 +0000 (08:43 -0300)]
[media] ov2640: make GPIOLIB an optional dependency

As warned by kbuild test robot:
warning: (VIDEO_EM28XX_V4L2) selects VIDEO_OV2640 which has unmet direct dependencies (MEDIA_SUPPORT && VIDEO_V4L2 && I2C && GPIOLIB && MEDIA_CAMERA_SUPPORT)

The em28xx driver can use ov2640, but it doesn't depend
(or use) the GPIOLIB in order to power off/on the sensor.

So, as we want to allow both usages with and without
GPIOLIB, make its dependency optional.

Reported-by: kbuild test robot <fengguang.wu@intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] xc5000: fix spelling mistake: "calibration"
Colin Ian King [Wed, 12 Apr 2017 14:04:13 +0000 (11:04 -0300)]
[media] xc5000: fix spelling mistake: "calibration"

Trivial fix to spelling mistake on calibration, make Self lowercase
and re-join multiple lined printk since checkpatch allows this
coding style.

Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] vidioc-queryctrl.rst: fix menu/int menu references
Mauro Carvalho Chehab [Sun, 9 Apr 2017 10:56:08 +0000 (07:56 -0300)]
[media] vidioc-queryctrl.rst: fix menu/int menu references

The documentation incorrectly mentions MENU and INTEGER_MENU
at struct v4l2_querymenu table as if they were flags. They're
not: they're types.

Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Reviewed-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media-entity: only call dev_dbg_obj if mdev is not NULL
Helen Fornazier [Thu, 6 Apr 2017 19:25:15 +0000 (16:25 -0300)]
[media] media-entity: only call dev_dbg_obj if mdev is not NULL

Fix kernel Oops NULL pointer deference
Call dev_dbg_obj only after checking if gobj->mdev is not NULL

Signed-off-by: Helen Koike <helen.koike@collabora.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] pixfmt-meta-vsp1-hgo.rst: remove spurious '-'
Mauro Carvalho Chehab [Wed, 19 Apr 2017 11:01:18 +0000 (08:01 -0300)]
[media] pixfmt-meta-vsp1-hgo.rst: remove spurious '-'

Remove spurious '-' in the VSP1 hgo table.

This resulted in a weird dot character that also caused
the row to be double-height.

We used to have it on other tables, but we got rid of them
on changeset 8ed29e302dd1 ("[media] subdev-formats.rst: remove
spurious '-'").

Fixes: 14d665387165 ("[media] v4l: Define a pixel format for the R-Car VSP1 1-D histogram engine")
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] mtk-vcodec: avoid warnings because of empty macros
Mauro Carvalho Chehab [Wed, 19 Apr 2017 10:48:56 +0000 (07:48 -0300)]
[media] mtk-vcodec: avoid warnings because of empty macros

Remove those gcc warnings:

drivers/media/platform/mtk-vcodec/mtk_vcodec_dec_pm.c: In function 'mtk_vcodec_dec_pw_on':
drivers/media/platform/mtk-vcodec/mtk_vcodec_dec_pm.c:114:51: warning: suggest braces around empty body in an 'if' statement [-Wempty-body]
   mtk_v4l2_err("pm_runtime_get_sync fail %d", ret);
                                                   ^

By adding braces.

Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] coda: bump maximum number of internal framebuffers to 17
Lucas Stach [Wed, 5 Apr 2017 13:11:15 +0000 (10:11 -0300)]
[media] coda: bump maximum number of internal framebuffers to 17

The h.264 standard allows up to 16 reference frame for the high profile
and we need one additional internal framebuffer when the VDOA is in use.

Lift the current maximum of 8 internal framebuffers to allow playback
of those video streams.

Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Acked-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: mtk-vcodec: remove informative log
Minghsiu Tsai [Wed, 5 Apr 2017 10:54:29 +0000 (07:54 -0300)]
[media] media: mtk-vcodec: remove informative log

Driver is stable. Remove DEBUG definition from driver.

There are debug message in /var/log/messages if DEBUG is defined,
such as:
[MTK_V4L2] level=0 fops_vcodec_open(),170: decoder capability 0
[MTK_V4L2] level=0 fops_vcodec_open(),177: 16000000.vcodec decoder [0]
[MTK_V4L2] level=0 fops_vcodec_release(),200: [0] decoder

Signed-off-by: Minghsiu Tsai <minghsiu.tsai@mediatek.com>
Acked-by: Tiffany Lin <Tiffany.lin@mediatek.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] subdev-formats.rst: remove spurious '-'
Hans Verkuil [Mon, 3 Apr 2017 10:05:45 +0000 (07:05 -0300)]
[media] subdev-formats.rst: remove spurious '-'

Remove spurious duplicate '-' in the Bayer Formats description. This resulted in a
weird dot character that also caused the row to be double-height.

The - character was probably used originally as indicator of an unused bit, but as the
number of columns was increased it was never used for the new columns.

Other tables do not use '-' either, so just remove it.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] dw2102: limit messages to buffer size
Alyssa Milburn [Sat, 1 Apr 2017 17:34:49 +0000 (14:34 -0300)]
[media] dw2102: limit messages to buffer size

Otherwise the i2c transfer functions can read or write beyond the end of
stack or heap buffers.

Signed-off-by: Alyssa Milburn <amilburn@zall.org>
Cc: stable@vger.kernel.org
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] ttusb2: limit messages to buffer size
Alyssa Milburn [Sat, 1 Apr 2017 17:34:32 +0000 (14:34 -0300)]
[media] ttusb2: limit messages to buffer size

Otherwise ttusb2_i2c_xfer can read or write beyond the end of static and
heap buffers.

Signed-off-by: Alyssa Milburn <amilburn@zall.org>
Cc: stable@vger.kernel.org
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] cec: add MEDIA_CEC_RC config option
Hans Verkuil [Mon, 17 Apr 2017 11:05:10 +0000 (08:05 -0300)]
[media] cec: add MEDIA_CEC_RC config option

Add an explicit config option to select whether the CEC remote control
messages are to be passed on to the RC subsystem or not.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] cec.h: merge cec-edid.h into cec.h
Hans Verkuil [Mon, 17 Apr 2017 10:54:37 +0000 (07:54 -0300)]
[media] cec.h: merge cec-edid.h into cec.h

Drop the separate cec-edid.h header and merge it into cec.h.

There was really no need to have a separate header for this.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] cec: Kconfig cleanup
Hans Verkuil [Mon, 17 Apr 2017 10:44:35 +0000 (07:44 -0300)]
[media] cec: Kconfig cleanup

The Kconfig options for the CEC subsystem were a bit messy. In
addition there were two cec sources (cec-edid.c and cec-notifier.c)
that were outside of the media/cec directory, which was weird.

Move those sources to media/cec as well.

The cec-edid and cec-notifier functionality is now part of the cec
module and these are no longer separate modules.

Also remove the MEDIA_CEC_EDID config option and include it with the
main CEC config option (which defined CEC_EDID anyway).

Added static inlines to cec-edid.h for dummy functions when CEC_CORE
isn't defined.

CEC drivers should now depend on CEC_CORE.

CEC drivers that need the cec-notifier functionality must explicitly
select CEC_NOTIFIER.

The s5p-cec and stih-cec drivers depended on VIDEO_DEV instead of
CEC_CORE, fix that as well.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Acked-by: Benjamin Gaignard <benjamin.gaignard@linaro.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] zr364xx: enforce minimum size when reading header
Alyssa Milburn [Sat, 1 Apr 2017 17:34:08 +0000 (14:34 -0300)]
[media] zr364xx: enforce minimum size when reading header

This code copies actual_length-128 bytes from the header, which will
underflow if the received buffer is too small.

Signed-off-by: Alyssa Milburn <amilburn@zall.org>
Cc: stable@vger.kernel.org
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] digitv: limit messages to buffer size
Alyssa Milburn [Sat, 1 Apr 2017 17:33:42 +0000 (14:33 -0300)]
[media] digitv: limit messages to buffer size

Return an error rather than memcpy()ing beyond the end of the buffer.
Internal callers use appropriate sizes, but digitv_i2c_xfer may not.

Signed-off-by: Alyssa Milburn <amilburn@zall.org>
Cc: stable@vger.kernel.org
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: stk1160: Add Kconfig help on snd-usb-audio requirement
Ezequiel Garcia [Sat, 25 Mar 2017 19:14:15 +0000 (16:14 -0300)]
[media] media: stk1160: Add Kconfig help on snd-usb-audio requirement

The Kconfig currently makes no reference to the snd-usb-audio
driver, which supports audio capture for this type of devices.
Just in case, let's make sure the requirement is mentioned
in the description.

Signed-off-by: Ezequiel Garcia <ezequiel@vanguardiasur.com.ar>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] staging: media: omap4iss: Replace a bit shift by a use of BIT
Arushi Singhal [Wed, 22 Mar 2017 04:26:09 +0000 (01:26 -0300)]
[media] staging: media: omap4iss: Replace a bit shift by a use of BIT

This patch replaces bit shifting on 1 with the BIT(x) macro.
This was done with coccinelle:
@@
constant c;
@@

-1 << c
+BIT(c)

Signed-off-by: Arushi Singhal <arushisinghal19971997@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] dvb-frontends/cxd2841er: define symbol_rate_min/max in T/C fe-ops
Daniel Scheller [Sun, 19 Mar 2017 15:26:39 +0000 (12:26 -0300)]
[media] dvb-frontends/cxd2841er: define symbol_rate_min/max in T/C fe-ops

Fixes "w_scan -f c" complaining with

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

Signed-off-by: Daniel Scheller <d.scheller@gmx.net>
Acked-by: Abylay Ospan <aospan@netup.ru>
Cc: stable@vger.kernel.org
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] em28xx: add support for V4L2_PIX_FMT_SRGGB8
Frank Schaefer [Sat, 15 Apr 2017 10:05:04 +0000 (07:05 -0300)]
[media] em28xx: add support for V4L2_PIX_FMT_SRGGB8

Adding support for SRGGB8 is as simple as adding a new entry at
struct em28xx_fmt.

Signed-off-by: Frank Schäfer <fschaefer.oss@googlemail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] em28xx: shed some light on video input formats
Frank Schaefer [Sat, 15 Apr 2017 10:05:03 +0000 (07:05 -0300)]
[media] em28xx: shed some light on video input formats

CbYCrY has been identified by looking into the tvp5150 driver and the
saa7115 datasheet.
YUV formats have been verified with em2765 + ov2640 (VAD Laplace webcam).
RGB8 formats have been verified with em2710/em2820 + mt9v011 (Silvercrest
webcam 1.3mpix).
I also did some cross-checking with these two camera devices and 0x08-0x0b
are at least 16 bits per pixel formats on em2710/em2820, too, and
0x0c-0x0f are at least 8 bits per pixel formats on em2765, too.

Signed-off-by: Frank Schäfer <fschaefer.oss@googlemail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] em28xx: don't treat device as webcam if an unknown sensor is detected
Frank Schaefer [Sat, 15 Apr 2017 10:05:02 +0000 (07:05 -0300)]
[media] em28xx: don't treat device as webcam if an unknown sensor is detected

With an unknown sensor, norm_maxw() and norm_maxh() return 0 as max.
height and width values, which causes a devide by zero in size_to_scale().
Of course we could use speculative default values for unknown sensors,
but the chance that the device works at this resolution without any
driver/setup is very low and therefore not worth the efforts.
Instead, just don't treat the device as camera.
A message will then be printed to the log that the device isn't supported.

Signed-off-by: Frank Schäfer <fschaefer.oss@googlemail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] em28xx: add missing auto-selections for build
Frank Schaefer [Sat, 15 Apr 2017 10:05:01 +0000 (07:05 -0300)]
[media] em28xx: add missing auto-selections for build

With MEDIA_SUBDRV_AUTOSELECT enabled in the kernel config, the em28xx
driver currently does't select some used subdrivers.
Fix this by adding the missing auto-selections to the Kconfig file.

Signed-off-by: Frank Schäfer <fschaefer.oss@googlemail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] em28xx: get rid of the dummy clock source
Frank Schaefer [Sat, 15 Apr 2017 10:05:00 +0000 (07:05 -0300)]
[media] em28xx: get rid of the dummy clock source

The v4l2 dummy clock has been added with commit fc5d0f8a8878
("V4L2: em28xx: register a V4L2 clock source") to be able to use the ov2640
soc_camera driver.
Since commit 46796cfcd346 ("ov2640: use standard clk and enable it") it is
no longer required.

Signed-off-by: Frank Schäfer <fschaefer.oss@googlemail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] em28xx: simplify ID-reading from Micron sensors
Frank Schaefer [Sun, 16 Apr 2017 17:41:52 +0000 (14:41 -0300)]
[media] em28xx: simplify ID-reading from Micron sensors

Use i2c_smbus_read_word_data() instead of i2c_master_send() and
i2c_master_recv() for reading the ID of Micorn sensors.

i2c_smbus_read_word_data() assumes that byes are in little-endian,
so, it uses:
data->word = msgbuf1[0] | (msgbuf1[1] << 8);

However, Micron datasheet describes the ID as if they were read
in big-endian. So, we need to change the byte order in order to
match the ID number as described on their datasheets.

Signed-off-by: Frank Schäfer <fschaefer.oss@googlemail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media/dvb-core: Race condition when writing to CAM
Jasmin J [Sat, 18 Mar 2017 02:04:20 +0000 (23:04 -0300)]
[media] media/dvb-core: Race condition when writing to CAM

It started with a sporadic message in syslog: "CAM tried to send a
buffer larger than the ecount size" This message is not the fault
itself, but a consecutive fault, after a read error from the CAM. This
happens only on several CAMs, several hardware, and of course sporadic.

It is a consecutive fault, if the last read from the CAM did fail. I
guess this will not happen on all CAMs, but at least it did on mine.
There was a write error to the CAM and during the re-initialization
procedure, the CAM finished the last read, although it got a RS.

The write error to the CAM happened because a race condition between HC
write, checking DA and FR.

This patch added an additional check for DA(RE), just after checking FR.
It is important to read the CAMs status register again, to give the CAM
the necessary time for a proper reaction to HC. Please note the
description within the source code (patch below).

[mchehab@s-opensource.com: make checkpatch happy]

Signed-off-by: Jasmin jessich <jasmin@anw.at>
Tested-by: Ralph Metzler <rjkm@metzlerbros.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] dvb-frontends/drxk: don't log errors on unsupported operation mode
Daniel Scheller [Tue, 14 Mar 2017 22:22:37 +0000 (19:22 -0300)]
[media] dvb-frontends/drxk: don't log errors on unsupported operation mode

When fe_ops.read_status is called and no channel is tuned (yet), the
subsequent calls to get_lock_status() causes the kernel log to be filled
with

    drxk: Error -22 on get_lock_status

which either means a NULL pointer was passed for the p_lock_status var,
or neither QAM nor OFDM/DVBT operation mode are active. Instead of
filling the kernel log in the latter case, print out a message to the debug
level and return 0 (this isn't used in the calling drxk_get_stats() anyway).

Signed-off-by: Daniel Scheller <d.scheller@gmx.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years agoMAINTAINERS: Add file patterns for media device tree bindings
Geert Uytterhoeven [Sun, 12 Mar 2017 13:16:56 +0000 (10:16 -0300)]
MAINTAINERS: Add file patterns for media device tree bindings

Submitters of device tree binding documentation may forget to CC
the subsystem maintainer if this is missing.

Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
Reviewed-by: Javier Martinez Canillas <javier@osg.samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] staging: lirc_zilog: Clean up tests if NULL returned on failure
simran singhal [Fri, 10 Mar 2017 05:13:12 +0000 (02:13 -0300)]
[media] staging: lirc_zilog: Clean up tests if NULL returned on failure

Some functions like kmalloc/kzalloc return NULL on failure.
When NULL represents failure, !x is commonly used.

This was done using Coccinelle:
@@
expression *e;
identifier l1;
@@

e = \(kmalloc\|kzalloc\|kcalloc\|devm_kzalloc\)(...);
...
- e == NULL
+ !e

Signed-off-by: simran singhal <singhalsimran0@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] docs-rst: media: better document refcount in struct dvb_frontend
Sakari Ailus [Thu, 9 Mar 2017 15:19:32 +0000 (12:19 -0300)]
[media] docs-rst: media: better document refcount in struct dvb_frontend

The refcount field was added to the struct but it was not
properly documented.

Document it.

[mchehab@s-opensource.com: fix a merge conflict]

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] docs-rst: Make the CSI-2 bus initialisation documentation match reality
Sakari Ailus [Thu, 9 Mar 2017 10:22:11 +0000 (07:22 -0300)]
[media] docs-rst: Make the CSI-2 bus initialisation documentation match reality

Update the CSI-2 bus documentation to tell that the LP-11 mode is not
mandatory as there are transmitters that cannot be explicitly set to LP-11
mode. Instead, say that this what the transmitter drivers shall do if
possible.

Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] em28xx: Add new USB ID eb1a:5051
Alexandre-Xavier Labonté-Lamoureux [Sat, 25 Feb 2017 08:03:38 +0000 (05:03 -0300)]
[media] em28xx: Add new USB ID eb1a:5051

Add new usbid eb1a:5051 for the Ion Video 2 PC MKII, Startech svid2usb23
and Raygo R12-41373.

Signed-off-by: Alexandre-Xavier Labonté-Lamoureux <alexandrexavier@live.ca>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] tm6000: Fix resource freeing in 'tm6000_prepare_isoc()'
Christophe JAILLET [Wed, 22 Feb 2017 22:32:19 +0000 (19:32 -0300)]
[media] tm6000: Fix resource freeing in 'tm6000_prepare_isoc()'

'usb_free_urb(urb)' is a no-op, because urb is known to be NULL.

It is likelly that releasing resources allocated by
'tm6000_alloc_urb_buffers()' just a few lines above is expected here.

This has been spotted by the following coccinelle script:
@@
expression ret, x, e;
identifier f;
@@

*   if (x == NULL)
    {
     ... when != x = e;
(
*    f(<+...x...+>);
|
*    ret = f(<+...x...+>);
)
     ...
    }

Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] tc358743: Add OF device ID table
Javier Martinez Canillas [Wed, 22 Feb 2017 16:11:28 +0000 (13:11 -0300)]
[media] tc358743: Add OF device ID table

The driver doesn't have a struct of_device_id table but supported devices
are registered via Device Trees. This is working on the assumption that a
I2C device registered via OF will always match a legacy I2C device ID and
that the MODALIAS reported will always be of the form i2c:<device>.

But this could change in the future so the correct approach is to have an
OF device ID table if the devices are registered via OF.

Signed-off-by: Javier Martinez Canillas <javier@osg.samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: pci: saa7164: remove dead code
Gustavo A. R. Silva [Tue, 21 Feb 2017 03:49:59 +0000 (00:49 -0300)]
[media] media: pci: saa7164: remove dead code

Remove dead code. The following line of code is never reached:
return SAA_OK;

Addresses-Coverity-ID: 114283

Signed-off-by: Gustavo A. R. Silva <garsilva@embeddedor.com>
Reviewed-by: Peter Senna Tschudin <peter.senna@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: pci: saa7164: remove unnecessary code
Gustavo A. R. Silva [Tue, 21 Feb 2017 03:46:58 +0000 (00:46 -0300)]
[media] media: pci: saa7164: remove unnecessary code

Remove unnecessary variable 'loop'.

Signed-off-by: Gustavo A. R. Silva <garsilva@embeddedor.com>
Reviewed-by: Peter Senna Tschudin <peter.senna@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] et8ek8: Export OF device ID as module aliases
Javier Martinez Canillas [Mon, 20 Feb 2017 20:16:16 +0000 (17:16 -0300)]
[media] et8ek8: Export OF device ID as module aliases

The I2C core always reports a MODALIAS of the form i2c:<foo> even if the
device was registered via OF, this means that exporting the OF device ID
table device aliases in the module is not needed. But in order to change
how the core reports modaliases to user-space, it's better to export it.

Before this patch:

$ modinfo drivers/media/i2c/et8ek8/et8ek8.ko | grep alias
alias:          i2c:et8ek8

After this patch:

$ modinfo drivers/media/i2c/et8ek8/et8ek8.ko | grep alias
alias:          i2c:et8ek8
alias:          of:N*T*Ctoshiba,et8ek8C*
alias:          of:N*T*Ctoshiba,et8ek8

Signed-off-by: Javier Martinez Canillas <javier@osg.samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: pci: constify stv0299_config structures
Bhumika Goyal [Sun, 19 Feb 2017 18:04:41 +0000 (15:04 -0300)]
[media] media: pci: constify stv0299_config structures

Declare stv0299_config structures as const as they are only passed as
an argument to the function dvb_attach. dvb_attach
calls its first argument on the rest of its arguments. The first
argument of dvb_attach in the changed cases is stv0299_attach and
the parameter of this function to which the object references are passed
is of type const. So, stv0299_config structures having this property
can be made const.

First line shows the file size before patching and second one shows size
after patching.

   text    data     bss     dec     hex filename
   9572     926      40   10538    292a media/pci/dm1105/dm1105.o
   9636     862      40   10538    292a media/pci/dm1105/dm1105.o

  15133    5408       0   20541    503d media/pci/ttpci/budget-av.o
  15389    5152       0   20541    503d media/pci/ttpci/budget-av.o

  15703    2326      36   18065    4691 media/pci/ttpci/budget-ci.o
  15767    2262      36   18065    4691 media/pci/ttpci/budget-ci.o

  10555    1918       4   12477    30bd drivers/media/pci/ttpci/budget.o
  10683    1822       4   12509    30dd drivers/media/pci/ttpci/budget.o

Signed-off-by: Bhumika Goyal <bhumirks@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] pci: mantis: constify mb86a16_config structure
Bhumika Goyal [Sun, 19 Feb 2017 17:39:06 +0000 (14:39 -0300)]
[media] pci: mantis: constify mb86a16_config structure

Declare mb86a16_config structure as const as it is either passed as
an argument to the function dvb_attach or is dereferenced. dvb_attach
calls its first argument on the rest of its arguments. The first argument
of dvb_attach in the changed case is mb86a16_attach and the parameter of
this function to which the object reference is passed is of type
const. So, mb86a16_config structures having this property can be made
const.

Signed-off-by: Bhumika Goyal <bhumirks@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] cx88: constify mb86a16_config structure
Bhumika Goyal [Sun, 19 Feb 2017 17:29:07 +0000 (14:29 -0300)]
[media] cx88: constify mb86a16_config structure

Declare mb86a16_config structure as const as it is only passed as
an argument to the function dvb_attach. dvb_attach calls its first
argument on the rest of its arguments. The first argument of
dvb_attach in the changed case is mb86a16_attach and the parameter of
this function to which the object reference is passed is of type
const. So, mb86a16_config structures having this property can be made
const.

Signed-off-by: Bhumika Goyal <bhumirks@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] saa7134: constify nxt200x_config structures
Bhumika Goyal [Sun, 19 Feb 2017 13:06:38 +0000 (10:06 -0300)]
[media] saa7134: constify nxt200x_config structures

Declare nxt200x_config structures as const as they are only passed as
an argument to the function dvb_attach. dvb_attach calls its first
argument on the rest of its arguments. The first argument of
dvb_attach in the changed cases is nxt200x_attach and the parameter of
this function to which the object references are passed is of type
const. So, nxt200x_config structures having this property can be made
const.

File size before:
   text    data     bss     dec     hex filename
  21320    3776      16   25112    6218 saa7134/saa7134-dvb.o

File size after:
   text    data     bss     dec     hex filename
  21384    3744      16   25144    6238 saa7134/saa7134-dvb.o

Signed-off-by: Bhumika Goyal <bhumirks@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] b2c2: constify nxt200x_config structure
Bhumika Goyal [Sun, 19 Feb 2017 12:59:52 +0000 (09:59 -0300)]
[media] b2c2: constify nxt200x_config structure

Declare nxt200x_config structure as const as it is only passed as
an argument to the function dvb_attach. dvb_attach calls its first
argument on the rest of its arguments. The first argument of
dvb_attach in the changed case is nxt200x_attach and the parameter of
this function to which the object reference is passed is of type
const. So, nxt200x_config structures having this property can be made
const.

File size before:
   text    data     bss     dec     hex filename
   7566     568       0    8134    1fc6 common/b2c2/flexcop-fe-tuner.o

File size after:
   text    data     bss     dec     hex filename
   7582     536       0    8118    1fb6 common/b2c2/flexcop-fe-tuner.o

Signed-off-by: Bhumika Goyal <bhumirks@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] dvb-usb-dibusb-mc-common: Add MODULE_LICENSE
Ben Hutchings [Sat, 18 Feb 2017 00:30:51 +0000 (22:30 -0200)]
[media] dvb-usb-dibusb-mc-common: Add MODULE_LICENSE

dvb-usb-dibusb-mc-common is licensed under GPLv2, and if we don't say
so then it won't even load since it needs a GPL-only symbol.

Fixes: e91455a1495a ("[media] dvb-usb: split out common parts of dibusb")

Reported-by: Dominique Dumont <dod@debian.org>
Cc: stable@vger.kernel.org # 4.9+
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] Staging: media/lirc: don't call put_ir_rx on rx twice
Colin Ian King [Fri, 17 Feb 2017 16:17:30 +0000 (14:17 -0200)]
[media] Staging: media/lirc: don't call put_ir_rx on rx twice

There is an exit path where rx is kfree'd on put_ir_rx and then
a jump to label out_put_xx will again kfree it with another
call to put_ir_rx.  Fix this by adding a new label that avoids
this 2nd call to put_ir_rx for this specific case.

Detected with CoverityScan, CID#145119 ("Use after free")

Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] drivers/media: Convert remaining use of pr_warning to pr_warn
Joe Perches [Fri, 17 Feb 2017 07:11:35 +0000 (05:11 -0200)]
[media] drivers/media: Convert remaining use of pr_warning to pr_warn

To enable eventual removal of pr_warning

This makes pr_warn use consistent for drivers/media

Prior to this patch, there was 1 use of pr_warning and
310 uses of pr_warn in drivers/media

Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] vimc: Virtual Media Controller core, capture and sensor
Helen Koike [Fri, 7 Apr 2017 17:55:19 +0000 (14:55 -0300)]
[media] vimc: Virtual Media Controller core, capture and sensor

First version of the Virtual Media Controller.
Add a simple version of the core of the driver, the capture and
sensor nodes in the topology, generating a grey image in a hardcoded
format.

Signed-off-by: Helen Koike <helen.koike@collabora.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
[hans.verkuil@cisco.com: fix small typo in Kconfig]

Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: i2c: Add support for OV5647 sensor
Ramiro Oliveira [Wed, 22 Mar 2017 12:30:27 +0000 (09:30 -0300)]
[media] media: i2c: Add support for OV5647 sensor

The OV5647 sensor from Omnivision supports up to 2592x1944 @ 15 fps, RAW 8
and RAW 10 output formats, and MIPI CSI-2 interface.

The driver adds support for 640x480 RAW 8.

Signed-off-by: Ramiro Oliveira <roliveir@synopsys.com>
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: Vladimir Zapolskiy <vladimir_zapolskiy@mentor.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] Documentation: DT: Add OV5647 bindings
Ramiro Oliveira [Wed, 22 Mar 2017 12:30:26 +0000 (09:30 -0300)]
[media] Documentation: DT: Add OV5647 bindings

Create device tree bindings documentation.

Signed-off-by: Ramiro Oliveira <roliveir@synopsys.com>
Acked-by: Rob Herring <robh@kernel.org>
Reviewed-by: Vladimir Zapolskiy <vladimir_zapolskiy@mentor.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: Add a driver for the ov5645 camera sensor
Todor Tomov [Tue, 11 Apr 2017 11:28:46 +0000 (08:28 -0300)]
[media] media: Add a driver for the ov5645 camera sensor

The ov5645 sensor from Omnivision supports up to 2592x1944
and CSI2 interface.

The driver adds support for the following modes:
- 1280x960
- 1920x1080
- 2592x1944

Output format is packed 8bit UYVY.

Signed-off-by: Todor Tomov <todor.tomov@linaro.org>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Hans Verkuil <hansverk@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: i2c/ov5645: add the device tree binding document
Todor Tomov [Tue, 11 Apr 2017 11:28:30 +0000 (08:28 -0300)]
[media] media: i2c/ov5645: add the device tree binding document

Add the document for ov5645 device tree binding.

Signed-off-by: Todor Tomov <todor.tomov@linaro.org>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Acked-by: Rob Herring <robh@kernel.org>
Signed-off-by: Hans Verkuil <hansverk@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] vsp1: set V4L2_CTRL_FLAG_MODIFY_LAYOUT for histogram controls
Hans Verkuil [Mon, 10 Apr 2017 19:18:06 +0000 (16:18 -0300)]
[media] vsp1: set V4L2_CTRL_FLAG_MODIFY_LAYOUT for histogram controls

The two histogram controls will modify the layout of the
metadata, so this flag should be set.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] buffer.rst: clarify how V4L2_CTRL_FLAG_MODIFY_LAYOUT/GRABBER are used
Hans Verkuil [Mon, 10 Apr 2017 19:17:06 +0000 (16:17 -0300)]
[media] buffer.rst: clarify how V4L2_CTRL_FLAG_MODIFY_LAYOUT/GRABBER are used

Explain when the V4L2_CTRL_FLAG_MODIFY_LAYOUT and
V4L2_CTRL_FLAG_MODIFY_GRABBER flags should be used.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] v4l2-ctrls.c: set V4L2_CTRL_FLAG_MODIFY_LAYOUT for ROTATE
Hans Verkuil [Mon, 10 Apr 2017 19:16:16 +0000 (16:16 -0300)]
[media] v4l2-ctrls.c: set V4L2_CTRL_FLAG_MODIFY_LAYOUT for ROTATE

The rotate control will modify the layout by definition. Always
set this flag.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] videodev.h: add V4L2_CTRL_FLAG_MODIFY_LAYOUT
Hans Verkuil [Mon, 10 Apr 2017 19:15:26 +0000 (16:15 -0300)]
[media] videodev.h: add V4L2_CTRL_FLAG_MODIFY_LAYOUT

Add new flag to indicate that changing this control will change the
buffer/mediabus layout as well.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] vidioc-queryctrl.rst: document V4L2_CTRL_FLAG_MODIFY_LAYOUT
Hans Verkuil [Mon, 10 Apr 2017 19:14:25 +0000 (16:14 -0300)]
[media] vidioc-queryctrl.rst: document V4L2_CTRL_FLAG_MODIFY_LAYOUT

Document this new control flag.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] v4l: vsp1: Add HGT support
Niklas Söderlund [Tue, 6 Sep 2016 14:38:56 +0000 (11:38 -0300)]
[media] v4l: vsp1: Add HGT support

The HGT is a Histogram Generator Two-Dimensions. It computes a weighted
frequency histograms for hue and saturation areas over a configurable
region of the image with optional subsampling.

Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Reviewed-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] v4l: Define a pixel format for the R-Car VSP1 2-D histogram engine
Niklas Söderlund [Tue, 6 Sep 2016 14:38:55 +0000 (11:38 -0300)]
[media] v4l: Define a pixel format for the R-Car VSP1 2-D histogram engine

The format is used on the R-Car VSP1 video queues that carry
2-D histogram statistics data.

Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Reviewed-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] v4l: vsp1: Add HGO support
Laurent Pinchart [Wed, 24 Feb 2016 23:40:22 +0000 (20:40 -0300)]
[media] v4l: vsp1: Add HGO support

The HGO is a Histogram Generator One-Dimension. It computes per-channel
histograms over a configurable region of the image with optional
subsampling.

Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] v4l: Define a pixel format for the R-Car VSP1 1-D histogram engine
Laurent Pinchart [Sun, 10 Apr 2016 07:37:48 +0000 (04:37 -0300)]
[media] v4l: Define a pixel format for the R-Car VSP1 1-D histogram engine

The format is used on the R-Car VSP1 video queues that carry
1-D histogram statistics data.

Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] v4l: vsp1: Fix HGO and HGT routing register addresses
Laurent Pinchart [Wed, 7 Sep 2016 12:36:31 +0000 (09:36 -0300)]
[media] v4l: vsp1: Fix HGO and HGT routing register addresses

The addresses are incorrect, fix them.

Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] v4l: vsp1: Support histogram generators in pipeline configuration
Laurent Pinchart [Wed, 7 Sep 2016 12:09:53 +0000 (09:09 -0300)]
[media] v4l: vsp1: Support histogram generators in pipeline configuration

Histogram generators are single-pad entities that branch as leaf nodes
at any point in the pipeline. Make sure that pipeline traversal and
routing configuration support them correctly.

Support for the actual HGO and HGT operation will come later.

Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] v4l: vsp1: Add histogram support
Laurent Pinchart [Wed, 7 Sep 2016 11:58:49 +0000 (08:58 -0300)]
[media] v4l: vsp1: Add histogram support

The histogram common code will be used to implement support for both the
HGO and HGT histogram computation engines.

Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] v4l: Add metadata buffer type and format
Laurent Pinchart [Tue, 12 Apr 2016 22:40:46 +0000 (19:40 -0300)]
[media] v4l: Add metadata buffer type and format

The metadata buffer type is used to transfer metadata between userspace
and kernelspace through a V4L2 buffers queue. It comes with a new
metadata capture capability and format description.

Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Tested-by: Guennadi Liakhovetski <guennadi.liakhovetski@intel.com>
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Acked-by: Hans Verkuil <hans.verkuil@cisco.com>
[hans.verkuil@cisco.com: removed left-over 'experimental' note]
[hans.verkuil@cisco.com: add newline after _v4l2-meta-format label]

Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] v4l: vsp1: wpf: Implement rotation support
Laurent Pinchart [Mon, 20 Jun 2016 09:07:08 +0000 (06:07 -0300)]
[media] v4l: vsp1: wpf: Implement rotation support

Some WPF instances, on Gen3 devices, can perform 90° rotation when
writing frames to memory. Implement support for this using the
V4L2_CID_ROTATE control.

Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] v4l: Clearly document interactions between formats, controls and buffers
Laurent Pinchart [Sun, 26 Feb 2017 12:44:18 +0000 (09:44 -0300)]
[media] v4l: Clearly document interactions between formats, controls and buffers

V4L2 exposes parameters that influence buffers sizes through the format
ioctls (VIDIOC_G_FMT, VIDIOC_TRY_FMT, VIDIOC_S_FMT, and possibly
VIDIOC_G_SELECTION and VIDIOC_S_SELECTION). Other parameters not part of
the format structure may also influence buffer sizes or buffer layout in
general. One existing such parameter is rotation, which is implemented
by the V4L2_CID_ROTATE control and thus exposed through the V4L2 control
ioctls.

The interaction between those parameters and buffers is currently only
partially specified by the V4L2 API. In particular interactions between
controls and buffers isn't specified at all. The behaviour of the
VIDIOC_S_FMT and VIDIOC_S_SELECTION ioctls when buffers are allocated is
also not fully specified.

This patch clearly defines and documents the interactions between
formats, selections, controls and buffers.

The preparatory discussions for the documentation change considered
completely disallowing controls that change the buffer size or layout,
in favour of extending the format API with a new ioctl that would bundle
those controls with format information. The idea has been rejected, as
this would essentially be a restricted version of the upcoming request
API that wouldn't bring any additional value.

Another option we have considered was to mandate the use of the request
API to modify controls that influence buffer size or layout. This has
also been rejected on the grounds that requiring the request API to
change rotation even when streaming is stopped would significantly
complicate implementation of drivers and usage of the V4L2 API for
applications.

Applications will however be required to use the upcoming request API to
change at runtime formats or controls that influence the buffer size or
layout, because of the need to synchronize buffers with the formats and
controls. Otherwise there would be no way to interpret the content of a
buffer correctly.

Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Acked-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] ov2640: avoid calling ov2640_select_win() twice
Hans Verkuil [Mon, 4 May 2015 10:25:51 +0000 (07:25 -0300)]
[media] ov2640: avoid calling ov2640_select_win() twice

Simplify ov2640_set_params and ov2640_set_fmt.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Acked-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] ov9740: avoid calling ov9740_res_roundup() twice
Hans Verkuil [Mon, 4 May 2015 10:25:55 +0000 (07:25 -0300)]
[media] ov9740: avoid calling ov9740_res_roundup() twice

Simplify ov9740_s_fmt.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Reported-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] ov9640: avoid calling ov9640_res_roundup() twice
Hans Verkuil [Mon, 4 May 2015 10:25:54 +0000 (07:25 -0300)]
[media] ov9640: avoid calling ov9640_res_roundup() twice

Simplify ov9640_s_fmt and ov9640_set_fmt

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Acked-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] ov772x: avoid calling ov772x_select_params() twice
Hans Verkuil [Mon, 4 May 2015 10:25:53 +0000 (07:25 -0300)]
[media] ov772x: avoid calling ov772x_select_params() twice

Merge ov772x_s_fmt into ov772x_set_fmt.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Acked-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] ov5642: avoid calling ov5642_find_datafmt() twice
Hans Verkuil [Mon, 4 May 2015 10:25:52 +0000 (07:25 -0300)]
[media] ov5642: avoid calling ov5642_find_datafmt() twice

Simplify ov5642_set_fmt().

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Reported-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] mt9v022: avoid calling mt9v022_find_datafmt() twice
Hans Verkuil [Mon, 4 May 2015 10:25:50 +0000 (07:25 -0300)]
[media] mt9v022: avoid calling mt9v022_find_datafmt() twice

Simplify mt9v022_s_fmt and mt9v022_set_fmt.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Acked-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] mt9m001: avoid calling mt9m001_find_datafmt() twice
Hans Verkuil [Mon, 4 May 2015 10:25:49 +0000 (07:25 -0300)]
[media] mt9m001: avoid calling mt9m001_find_datafmt() twice

Simplify mt9m001_s_fmt and mt9m001_set_fmt.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Acked-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] imx074: avoid calling imx074_find_datafmt() twice
Hans Verkuil [Mon, 4 May 2015 10:25:48 +0000 (07:25 -0300)]
[media] imx074: avoid calling imx074_find_datafmt() twice

Simplify imx074_set_fmt().

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Reported-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] si2157: Add support for Si2141-A10
Stefan Brüns [Fri, 17 Feb 2017 00:55:31 +0000 (22:55 -0200)]
[media] si2157: Add support for Si2141-A10

The Si2141 needs two distinct commands for powerup/reset, otherwise it
will not respond to chip revision requests. It also needs a firmware
to run properly.

Cc: Evgeny Plehov <EvgenyPlehov@ukr.net>
Signed-off-by: Stefan Brüns <stefan.bruens@rwth-aachen.de>
Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] si2157: revert si2157: Si2141/2151 tuner support
Antti Palosaari [Thu, 6 Apr 2017 12:17:42 +0000 (09:17 -0300)]
[media] si2157: revert si2157: Si2141/2151 tuner support

'Reset' loop does not look correct. I tested it very many times and
it never repeated those commands. If problem, it tries to solve,
really occurs on some situations better solution should be find out.

There is another patch which does not have such hackish looking loop.
Lets change to it.

Cc: Evgeny Plehov <EvgenyPlehov@ukr.net>
Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] v4l2-tpg: don't clamp XV601/709 to lim range
Hans Verkuil [Mon, 10 Apr 2017 07:54:30 +0000 (04:54 -0300)]
[media] v4l2-tpg: don't clamp XV601/709 to lim range

The XV601/709 encodings are special: they signal limited range, but use the full range
to encode a larger gamut with R', G' and B' values outside the [0-1] range.

So don't clamp to limited range for these two encodings.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] videodev2.h: fix outdated comment
Hans Verkuil [Mon, 10 Apr 2017 07:51:44 +0000 (04:51 -0300)]
[media] videodev2.h: fix outdated comment

The XV601/709 Y'CbCr encoding was changed to limited range, but the comment
still indicates full range.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] imon: use setup_timer
Geliang Tang [Sun, 9 Apr 2017 01:34:08 +0000 (22:34 -0300)]
[media] imon: use setup_timer

Use setup_timer() instead of init_timer() to simplify the code.

Signed-off-by: Geliang Tang <geliangtang@gmail.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] wl128x: use setup_timer
Geliang Tang [Sun, 9 Apr 2017 01:34:07 +0000 (22:34 -0300)]
[media] wl128x: use setup_timer

Use setup_timer() instead of init_timer() to simplify the code.

Signed-off-by: Geliang Tang <geliangtang@gmail.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] c8sectpfe: use setup_timer
Geliang Tang [Sun, 9 Apr 2017 01:34:06 +0000 (22:34 -0300)]
[media] c8sectpfe: use setup_timer

Use setup_timer() instead of init_timer() to simplify the code.

Signed-off-by: Geliang Tang <geliangtang@gmail.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] fsl-viu: use setup_timer
Geliang Tang [Sun, 9 Apr 2017 01:34:04 +0000 (22:34 -0300)]
[media] fsl-viu: use setup_timer

Use setup_timer() instead of init_timer() to simplify the code.

Signed-off-by: Geliang Tang <geliangtang@gmail.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] av7110: use setup_timer
Geliang Tang [Sun, 9 Apr 2017 01:34:03 +0000 (22:34 -0300)]
[media] av7110: use setup_timer

Use setup_timer() instead of init_timer() to simplify the code.

Signed-off-by: Geliang Tang <geliangtang@gmail.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] netup_unidvb: use setup_timer
Geliang Tang [Sun, 9 Apr 2017 01:34:02 +0000 (22:34 -0300)]
[media] netup_unidvb: use setup_timer

Use setup_timer() instead of init_timer() to simplify the code.

Signed-off-by: Geliang Tang <geliangtang@gmail.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] ivtv: use setup_timer
Geliang Tang [Sun, 9 Apr 2017 01:34:01 +0000 (22:34 -0300)]
[media] ivtv: use setup_timer

Use setup_timer() instead of init_timer() to simplify the code.

Signed-off-by: Geliang Tang <geliangtang@gmail.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] cx18: use setup_timer
Geliang Tang [Sun, 9 Apr 2017 01:34:00 +0000 (22:34 -0300)]
[media] cx18: use setup_timer

Use setup_timer() instead of init_timer() to simplify the code.

Signed-off-by: Geliang Tang <geliangtang@gmail.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] bt8xx: use setup_timer
Geliang Tang [Sun, 9 Apr 2017 01:33:59 +0000 (22:33 -0300)]
[media] bt8xx: use setup_timer

Use setup_timer() instead of init_timer() to simplify the code.

Signed-off-by: Geliang Tang <geliangtang@gmail.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] saa7146: use setup_timer
Geliang Tang [Sun, 9 Apr 2017 01:33:58 +0000 (22:33 -0300)]
[media] saa7146: use setup_timer

Use setup_timer() instead of init_timer() to simplify the code.

Signed-off-by: Geliang Tang <geliangtang@gmail.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] saa7134: use setup_timer
Geliang Tang [Sun, 9 Apr 2017 01:33:57 +0000 (22:33 -0300)]
[media] saa7134: use setup_timer

Use setup_timer() instead of init_timer() to simplify the code.

Signed-off-by: Geliang Tang <geliangtang@gmail.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] media: bcm2048: fix several macros
Nikola Jelic [Sat, 8 Apr 2017 15:44:41 +0000 (12:44 -0300)]
[media] media: bcm2048: fix several macros

Some of the macros didn't use the parenthesis around the parameters when
used in the body of the macro.

Signed-off-by: Nikola Jelic <nikola.jelic83@gmail.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] m2m-deinterlace: don't return zero on failure paths in deinterlace_probe()
Alexey Khoroshilov [Fri, 7 Apr 2017 23:09:17 +0000 (20:09 -0300)]
[media] m2m-deinterlace: don't return zero on failure paths in deinterlace_probe()

If DMA does not support INTERLEAVE, deinterlace_probe() breaks off
initialization, releases dma channel, but returns zero.

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@s-opensource.com>
7 years ago[media] coda: do not enumerate YUYV if VDOA is not available
Philipp Zabel [Thu, 6 Apr 2017 14:03:40 +0000 (11:03 -0300)]
[media] coda: do not enumerate YUYV if VDOA is not available

TRY_FMT already disables the YUYV format if the VDOA is not available.
ENUM_FMT must do the same.

Fixes: d40e98c13b3e ("[media] coda: support YUYV output if VDOA is used")

Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Reviewed-by: Lucas Stach <l.stach@pengutronix.de>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
7 years ago[media] tvp5150: fix pad format frame height
Philipp Zabel [Fri, 31 Mar 2017 13:33:26 +0000 (10:33 -0300)]
[media] tvp5150: fix pad format frame height

Even if field order is set to V4L2_FIELD_ALTERNATE, the width and height
values in struct v4l2_mbus_framefmt still refer to frame size, not field
size.

Fixes: 4f57d27be2a5 ("[media] tvp5150: fix tvp5150_fill_fmt()")

Signed-off-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>
7 years ago[media] tvp5150: allow get/set_fmt on the video source pad
Philipp Zabel [Fri, 31 Mar 2017 13:33:25 +0000 (10:33 -0300)]
[media] tvp5150: allow get/set_fmt on the video source pad

To let userspace propagate formats downstream in a media controller
scenario, the video source pad (now pad 1, DEMOD_PAD_VID_OUT) must allow
setting and getting the format. Incidentally, tvp5150_fill_fmt was
implemented for this pad, not for the new analog input pad (now pad 0,
DEMOD_PAD_IF_INPUT).

Fixes: 55606310e77f ("[media] tvp5150: create the expected number of pads")

Signed-off-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>
7 years ago[media] coda: remove redundant call to v4l2_m2m_get_vq
Colin Ian King [Thu, 23 Mar 2017 11:57:46 +0000 (08:57 -0300)]
[media] coda: remove redundant call to v4l2_m2m_get_vq

The call to v4ls_m2m_get_vq is only used to get the return value
which is not being used, so it appears to be redundant and can
be removed.

Detected with CoverityScan, CID#1420674 ("Useless call")

Signed-off-by: Colin Ian King <colin.king@canonical.com>
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>