openwrt/staging/blogic.git
9 years ago[media] mxl111sf: missing return values validation
Insu Yun [Thu, 15 Oct 2015 20:22:45 +0000 (17:22 -0300)]
[media] mxl111sf: missing return values validation

Return values of mxl111sf_enable_usb_output and mxl1x1sf_top_master_ctrl
are not validated.

Signed-off-by: Insu Yun <wuninsu@gmail.com>
Reviewed-by: Michael Ira Krufky <mkrufky@linuxtv.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] rcar_vin: Remove obsolete platform data support
Geert Uytterhoeven [Wed, 7 Oct 2015 10:39:36 +0000 (07:39 -0300)]
[media] rcar_vin: Remove obsolete platform data support

Since commit 3d7608e4c169af03 ("ARM: shmobile: bockw: remove legacy
board file and config"), Renesas R-Car SoCs are only supported in
generic DT-only ARM multi-platform builds.  The driver doesn't need to
use platform data anymore, hence remove platform data configuration.

Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] move media platform data to linux/platform_data/media
Mauro Carvalho Chehab [Tue, 17 Nov 2015 09:11:13 +0000 (07:11 -0200)]
[media] move media platform data to linux/platform_data/media

Now that media has its own subdirectory inside platform_data,
let's move the headers that are already there to such subdir.

After moving those files, the references were adjusted using this
script:

    MAIN_DIR="linux/platform_data/"
    PREV_DIR="linux/platform_data/"
    DIRS="media/"

    echo "Checking affected files" >&2
    for i in $DIRS; do
for j in $(find include/$MAIN_DIR/$i -type f -name '*.h'); do
 n=`basename $j`
git grep -l $n
done
    done|sort|uniq >files && (
echo "Handling files..." >&2;
echo "for i in \$(cat files|grep -v Documentation); do cat \$i | \\";
(
cd include/$MAIN_DIR;
for j in $DIRS; do
for i in $(ls $j); do
echo "perl -ne 's,(include [\\\"\\<])$PREV_DIR($i)([\\\"\\>]),\1$MAIN_DIR$j\2\3,; print \$_' |\\";
done;
done;
echo "cat > a && mv a \$i; done";
);
echo "Handling documentation..." >&2;
echo "for i in MAINTAINERS \$(cat files); do cat \$i | \\";
(
cd include/$MAIN_DIR;
for j in $DIRS; do
for i in $(ls $j); do
echo "  perl -ne 's,include/$PREV_DIR($i)\b,include/$MAIN_DIR$j\1,; print \$_' |\\";
done;
done;
echo "cat > a && mv a \$i; done"
);
    ) >script && . ./script

Suggested-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Mark Brown <broonie@kernel.org>
Acked-by: Robert Jarzmik <robert.jarzmik@free.fr>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] V4L2: platform: rcar_jpu: switch off clock on release later
Mikhail Ulyanov [Thu, 1 Oct 2015 12:03:32 +0000 (09:03 -0300)]
[media] V4L2: platform: rcar_jpu: switch off clock on release later

Give JPU peripheral chance to finish current job.
Don't switch off clock until context release.

Signed-off-by: Mikhail Ulyanov <mikhail.ulyanov@cogentembedded.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] V4L2: platform: rcar_jpu: remove redundant code
Mikhail Ulyanov [Thu, 1 Oct 2015 12:03:31 +0000 (09:03 -0300)]
[media] V4L2: platform: rcar_jpu: remove redundant code

Remove redundant code. Following code line do what we want.

[mchehab@osg.samsung.com: fix conflict with VB2 split patches that renamed some structs]
Signed-off-by: Mikhail Ulyanov <mikhail.ulyanov@cogentembedded.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years agoDocBook: only copy stuff to media_api if media xml is generated
Mauro Carvalho Chehab [Thu, 19 Nov 2015 09:41:40 +0000 (07:41 -0200)]
DocBook: only copy stuff to media_api if media xml is generated

It is possible to use:
make DOCBOOKS=device-drivers.xml htmldocs

To produce just a few docbooks. In such case, the media docs
won't be built, causing the makefile target to return an error.

While this is ok for human eyes, if the above is used on an script,
it would cause troubles.

Fix it by only creating/filling the media_api directory if the
media_api.xml is found at DOCBOOKS.

Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] rc: gpio-ir-recv: add timeout on idle
Eric Nelson [Wed, 23 Sep 2015 14:07:08 +0000 (11:07 -0300)]
[media] rc: gpio-ir-recv: add timeout on idle

Many decoders require a trailing space (period without IR illumination)
to be delivered before completing a decode.

Since the gpio-ir-recv driver only delivers events on gpio transitions,
a single IR symbol (caused by a quick touch on an IR remote) will not
be properly decoded without the use of a timer to flush the tail end
state of the IR receiver.

This patch initializes and uses a timer and the timeout field of rcdev
to complete the stream and allow decode.

The timeout can be overridden through the use of the LIRC_SET_REC_TIMEOUT
ioctl.

Signed-off-by: Eric Nelson <eric@nelint.com>
Acked-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] rc-core: define a default timeout for drivers
Eric Nelson [Sat, 3 Oct 2015 15:18:50 +0000 (08:18 -0700)]
[media] rc-core: define a default timeout for drivers

A default timeout value of 125 ms should work for all decoders.

Declare a constant to help standardize its' use.

Signed-off-by: Eric Nelson <eric@nelint.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] media: st-rc: remove misuse of IRQF_NO_SUSPEND flag
Sudeep Holla [Mon, 21 Sep 2015 15:47:10 +0000 (12:47 -0300)]
[media] media: st-rc: remove misuse of IRQF_NO_SUSPEND flag

The device is set as wakeup capable using proper wakeup API but the
driver misuses IRQF_NO_SUSPEND to set the interrupt as wakeup source
which is incorrect.

This patch removes the use of IRQF_NO_SUSPEND flags replacing it with
enable_irq_wake instead.

Cc: Srinivas Kandagatla <srinivas.kandagatla@gmail.com>
Cc: Maxime Coquelin <maxime.coquelin@st.com>
Cc: Patrice Chotard <patrice.chotard@st.com>
Cc: linux-arm-kernel@lists.infradead.org
Cc: kernel@stlinux.com
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] s5c73m3-spi: fix compilation breakage when compiled as Module
Mauro Carvalho Chehab [Wed, 18 Nov 2015 18:46:33 +0000 (16:46 -0200)]
[media] s5c73m3-spi: fix compilation breakage when compiled as Module

commit f934a94bb566a629b7e0be52d087a686145d1f14 ('s5c73m3: Export OF
module alias information') broke compilation when compiled as module:

In file included from drivers/media/i2c/s5c73m3/s5c73m3-spi.c:22:0:
drivers/media/i2c/s5c73m3/s5c73m3-spi.c:40:40: error: expected ')' before ';' token
 MODULE_DEVICE_TABLE(of, s5c73m3_spi_ids;);
                                        ^

Fix it by removing the extra ';'

Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] c8sectpfe: Update DT binding doc with some minor fixes
Peter Griffin [Thu, 3 Sep 2015 17:59:53 +0000 (14:59 -0300)]
[media] c8sectpfe: Update DT binding doc with some minor fixes

Signed-off-by: Peter Griffin <peter.griffin@linaro.org>
Acked-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] c8sectpfe: Update binding to reset-gpios
Peter Griffin [Thu, 3 Sep 2015 17:59:52 +0000 (14:59 -0300)]
[media] c8sectpfe: Update binding to reset-gpios

reset-gpios is more clear than rst-gpio.

This change has been done as one atomic commit but it
does breaks compatability with older dtbs.

Signed-off-by: Peter Griffin <peter.griffin@linaro.org>
Acked-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] ARM: DT: STi: STiH407: Add c8sectpfe LinuxDVB DT node
Peter Griffin [Thu, 3 Sep 2015 17:59:50 +0000 (14:59 -0300)]
[media] ARM: DT: STi: STiH407: Add c8sectpfe LinuxDVB DT node

This patch adds in the required DT node for the c8sectpfe
Linux DVB demux driver which allows the tsin channels
to be used on an upstream kernel.

Signed-off-by: Peter Griffin <peter.griffin@linaro.org>
Acked-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] ARM: DT: STi: stihxxx-b2120: Add pulse-width properties to ssc2 & ssc3
Peter Griffin [Thu, 3 Sep 2015 17:59:49 +0000 (14:59 -0300)]
[media] ARM: DT: STi: stihxxx-b2120: Add pulse-width properties to ssc2 & ssc3

Adding these properties makes the I2C bus to the demodulators much
more reliable, and we no longer suffer from I2C errors when tuning.

Signed-off-by: Peter Griffin <peter.griffin@linaro.org>
Acked-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] media: dvb-core: Don't force CAN_INVERSION_AUTO in oneshot mode
Malcolm Priestley [Mon, 31 Aug 2015 09:13:45 +0000 (06:13 -0300)]
[media] media: dvb-core: Don't force CAN_INVERSION_AUTO in oneshot mode

When in FE_TUNE_MODE_ONESHOT the frontend must report
the actual capabilities so user can take appropriate
action.

With frontends that can't do auto inversion this is done
by dvb-core automatically so CAN_INVERSION_AUTO is valid.

However, when in FE_TUNE_MODE_ONESHOT this is not true.

So only set FE_CAN_INVERSION_AUTO in modes other than
FE_TUNE_MODE_ONESHOT

Signed-off-by: Malcolm Priestley <tvboxspy@gmail.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] s5c73m3: Export OF module alias information
Javier Martinez Canillas [Thu, 20 Aug 2015 07:07:21 +0000 (04:07 -0300)]
[media] s5c73m3: Export OF module alias information

The SPI core always reports the MODALIAS uevent as "spi:<modalias>"
regardless of the mechanism that was used to register the device
(i.e: OF or board code) and the table that is used later to match
the driver with the device (i.e: SPI id table or OF match table).

So drivers needs to export the SPI id table and this be built into
the module or udev won't have the necessary information to autoload
the needed driver module when the device is added.

But this means that OF-only drivers needs to have both OF and SPI id
tables that have to be kept in sync and also the dev node compatible
manufacturer prefix is stripped when reporting the MODALIAS. Which can
lead to issues if two vendors use the same SPI device name for example.

To avoid the above, the SPI core behavior may be changed in the future
to not require an SPI device table for OF-only drivers and report the
OF module alias. So, it's better to also export the OF table even when
is unused now to prevent breaking module loading when the core changes.

Signed-off-by: Javier Martinez Canillas <javier@osg.samsung.com>
Reviewed-by: Andrzej Hajda <a.hajda@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] ddbridge: fix wait_event_timeout return handling
Nicholas Mc Guire [Wed, 10 Jun 2015 07:40:02 +0000 (04:40 -0300)]
[media] ddbridge: fix wait_event_timeout return handling

API conformance testing for completions with coccinelle spatches are being
used to locate API usage inconsistencies:
./drivers/media/pci/ddbridge/ddbridge-core.c:89
        incorrect check for negative return

Return type of wait_event_timeout is signed long not int and the
return type is >=0 always thus the negative check is unnecessary..
As stat is used here exclusively its type is simply changed and the
negative return check dropped.

Signed-off-by: Nicholas Mc Guire <hofrat@osadl.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] i2c: constify v4l2_ctrl_ops structures
Julia Lawall [Fri, 13 Nov 2015 22:05:17 +0000 (20:05 -0200)]
[media] i2c: constify v4l2_ctrl_ops structures

These v4l2_ctrl_ops structures are never modified, like all the other
v4l2_ctrl_ops structures, so declare them as const.

Done with the help of Coccinelle.

Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] atmel-isi: support RGB565 output when sensor output YUV formats
Josh Wu [Tue, 3 Nov 2015 05:45:12 +0000 (03:45 -0200)]
[media] atmel-isi: support RGB565 output when sensor output YUV formats

This patch enable Atmel ISI preview path to convert the YUV to RGB format.

Signed-off-by: Josh Wu <josh.wu@atmel.com>
[g.liakhovetski@gmx.de: removed superfluous parentheses]
Signed-off-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] atmel-isi: setup YCC_SWAP correctly when using preview path
Josh Wu [Tue, 3 Nov 2015 05:45:11 +0000 (03:45 -0200)]
[media] atmel-isi: setup YCC_SWAP correctly when using preview path

The preview path only can convert UYVY format to RGB data.

To make preview path work correctly, we need to set up YCC_SWAP
according to sensor output and convert them to UYVY.

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] atmel-isi: add code to setup correct resolution for preview path
Josh Wu [Tue, 3 Nov 2015 05:45:10 +0000 (03:45 -0200)]
[media] atmel-isi: add code to setup correct resolution for preview path

Not like codec path, preview path can do downsampling, so we should setup
a extra preview width, height for it.

This patch add preview resolution setup without down sampling. So currently
preview path will output same size as sensor output size.

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] atmel-isi: prepare for the support of preview path
Josh Wu [Tue, 3 Nov 2015 05:45:09 +0000 (03:45 -0200)]
[media] atmel-isi: prepare for the support of preview path

Atmel ISI support a preview path which can output RGB data.

So this patch introduces a bool variable to choose which path is
enabled currently. And also we need setup corresponding path registers.

By default the preview path is disabled. We only use Codec path.

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] atmel-isi: correct yuv swap according to different sensor outputs
Josh Wu [Tue, 3 Nov 2015 05:45:08 +0000 (03:45 -0200)]
[media] atmel-isi: correct yuv swap according to different sensor outputs

we need to configure the YCC_SWAP bits in ISI_CFG2 according to current
sensor output and Atmel ISI output format.

Current there are two cases Atmel ISI supported:
  1. Atmel ISI outputs YUYV format.
     This case we need to setup YCC_SWAP according to sensor output
     format.
  2. Atmel ISI output a pass-through formats, which means no swap.
     Just setup YCC_SWAP as default with no swap.

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] v4l2-clk: v4l2_clk_get() also need to find the of_fullname clock
Josh Wu [Wed, 28 Oct 2015 09:48:55 +0000 (07:48 -0200)]
[media] v4l2-clk: v4l2_clk_get() also need to find the of_fullname clock

The soc-camera host will be probed and register a v4l2_clk, but if at
that moment, the i2c device is not available, then the registered
v4l2_clk name is an OF string not a I2C string.

So when i2c sensor probes and calls v4l2_clk_get(), it only searches a
clock with I2C string, like "1-0030".

This patch will search the clock with OF string name if no clock with
I2C string name could be found.

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] v4l2-clk: add new definition: V4L2_CLK_NAME_SIZE
Josh Wu [Wed, 28 Oct 2015 09:48:54 +0000 (07:48 -0200)]
[media] v4l2-clk: add new definition: V4L2_CLK_NAME_SIZE

Make all v4l2-clk's clock name use V4L2_CLK_NAME_SIZE definition.

In future, if the string is increased we just need to change the
V4L2_CLK_NAME_SIZE once.

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] v4l2-clk: add new macro for v4l2_clk_name_of()
Josh Wu [Wed, 28 Oct 2015 09:48:53 +0000 (07:48 -0200)]
[media] v4l2-clk: add new macro for v4l2_clk_name_of()

This macro is used to generate an OF string for a v4l2 clock.

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: get the clock name by using macro: v4l2_clk_name_i2c()
Josh Wu [Wed, 28 Oct 2015 09:48:52 +0000 (07:48 -0200)]
[media] soc_camera: get the clock name by using macro: v4l2_clk_name_i2c()

Since v4l2_clk_name_i2c() is defined, reuse it.

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] vivid: add support for reduced frame rate in video capture
Prashant Laddha [Tue, 22 Sep 2015 14:27:31 +0000 (11:27 -0300)]
[media] vivid: add support for reduced frame rate in video capture

With this patch, vivid capture thread can now generate a video with
frame rate reduced by a factor of 1000 / 1001. This option can be
selected using a control Reduced Framerate from gui.

Signed-off-by: Prashant Laddha <prladdha@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] vivid-capture: add control for reduced frame rate
Prashant Laddha [Tue, 22 Sep 2015 14:27:30 +0000 (11:27 -0300)]
[media] vivid-capture: add control for reduced frame rate

A boolean control Reduced Framerate is added to vivid controls for
controlling the reduced fps option for vivid capture from gui.

Signed-off-by: Prashant Laddha <prladdha@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] vivid: add support for reduced fps in video out
Prashant Laddha [Tue, 22 Sep 2015 14:27:29 +0000 (11:27 -0300)]
[media] vivid: add support for reduced fps in video out

If reduced fps flag is set then check if other necessary conditions
are true for the given bt timing. If yes, then reduce the frame rate.
For vivid transmitter, timeperframe_vid_out controls the frame rate.
Adjusting the timeperframe_vid_out by scaling down pixel clock by
factor of 1000 / 1001.

Signed-off-by: Prashant Laddha <prladdha@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] v4l2-dv-timings: add condition checks for reduced fps
Prashant Laddha [Tue, 22 Sep 2015 14:27:28 +0000 (11:27 -0300)]
[media] v4l2-dv-timings: add condition checks for reduced fps

Added a helper function to check necessary conditions required for
reduced fps. The reduced fps is supported for CVT and CEA861 timings.
CVT supports reduced fps only if reduced blanking v2 (indicated by
vsync = 8) is true. Whereas CEA861 supports reduced fps if
V4L2_DV_FL_CAN_REDUCE_FPS flag is true.

Signed-off-by: Prashant Laddha <prladdha@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] v4l2-device: Don't unregister ACPI/Device Tree based devices
Tommi Franttila [Thu, 12 Nov 2015 09:01:07 +0000 (07:01 -0200)]
[media] v4l2-device: Don't unregister ACPI/Device Tree based devices

When a V4L2 sub-device backed by a DT or ACPI based device was removed,
the device was unregistered as well which certainly was not intentional,
as the client device would not be re-created by simply reinstating the
V4L2 sub-device (indeed the device would have to be there first!).

Skip unregistering the device in case it has non-NULL of_node or fwnode.

Signed-off-by: Tommi Franttila <tommi.franttila@intel.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.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: adv7180: increase delay after reset to 5ms
Ulrich Hecht [Tue, 10 Nov 2015 13:39:00 +0000 (11:39 -0200)]
[media] media: adv7180: increase delay after reset to 5ms

Initialization of the ADV7180 chip fails on the Renesas R8A7790-based
Lager board about 50% of the time.  This patch resolves the issue by
increasing the minimum delay after reset from 2 ms to 5 ms, following the
recommendation in the ADV7180 datasheet:

"Executing a software reset takes approximately 2 ms. However, it is
recommended to wait 5 ms before any further I2C writes are performed."

Signed-off-by: Ulrich Hecht <ulrich.hecht+renesas@gmail.com>
Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] usbvision fix overflow of interfaces array
Oliver Neukum [Tue, 27 Oct 2015 11:51:34 +0000 (09:51 -0200)]
[media] usbvision fix overflow of interfaces array

This fixes the crash reported in:
http://seclists.org/bugtraq/2015/Oct/35
The interface number needs a sanity check.

Signed-off-by: Oliver Neukum <oneukum@suse.com>
Cc: Vladis Dronov <vdronov@redhat.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] sh-vou: clarify videobuf2 dependency
Arnd Bergmann [Fri, 16 Oct 2015 00:35:40 +0000 (21:35 -0300)]
[media] sh-vou: clarify videobuf2 dependency

The sh-vou driver has been converted from videobuf to videobuf2, but
the Kconfig file still lists VIDEOBUF_DMA_CONTIG as a dependency.
Consequently we can build the driver without VIDEOBUF2_DMA_CONTIG
and get a link error:

drivers/built-in.o: In function `sh_vou_probe':
vf610-ocotp.c:(.text+0x2dbf5c): undefined reference to `vb2_dma_contig_init_ctx'
vf610-ocotp.c:(.text+0x2dc0b4): undefined reference to `vb2_dma_contig_cleanup_ctx'
vf610-ocotp.c:(.text+0x2dc144): undefined reference to `vb2_dma_contig_memops'
drivers/built-in.o: In function `sh_vou_remove':
vf610-ocotp.c:(.text+0x2dc190): undefined reference to `vb2_dma_contig_cleanup_ctx'

This changes the dependency to VIDEOBUF2_DMA_CONTIG instead.

Fixes: 57af3ad59d95 ("[media] sh-vou: convert to vb2")
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: fix waitqueue_active without memory barrier in cpia2 driver
Kosuke Tatsukawa [Fri, 9 Oct 2015 00:35:40 +0000 (21:35 -0300)]
[media] media: fix waitqueue_active without memory barrier in cpia2 driver

cpia2_usb_disconnect() seems to be missing a memory barrier which might
cause the waker to not notice the waiter and miss sending a wake_up as
in the following figure.

cpia2_usb_disconnect sync
------------------------------------------------------------------------
mutex_unlock(&cam->v4l2_lock);
if (waitqueue_active(&cam->wq_stream))
/* The CPU might reorder the test for
   the waitqueue up here, before
   prior writes complete */
/* wait_event_interruptible */
 /* __wait_event_interruptible */
  /* ___wait_event */
  long __int = prepare_to_wait_event(
    &wq, &__wait, state);
  if (!cam->streaming ||
    frame->status == FRAME_READY)
cam->curbuff->status = FRAME_READY;
cam->curbuff->length = 0;
  schedule()
------------------------------------------------------------------------

The attached patch removes the call to waitqueue_active() leaving just
wake_up() behind.  This fixes the problem because the call to
spin_lock_irqsave() in wake_up() will be an ACQUIRE operation.

I found this issue when I was looking through the linux source code
for places calling waitqueue_active() before wake_up*(), but without
preceding memory barriers, after sending a patch to fix a similar
issue in drivers/tty/n_tty.c  (Details about the original issue can be
found here: https://lkml.org/lkml/2015/9/28/849).

Signed-off-by: Kosuke Tatsukawa <tatsu@ab.jp.nec.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] gspca: ov534/topro: prevent a division by 0
Antonio Ospite [Fri, 2 Oct 2015 20:33:13 +0000 (17:33 -0300)]
[media] gspca: ov534/topro: prevent a division by 0

v4l2-compliance sends a zeroed struct v4l2_streamparm in
v4l2-test-formats.cpp::testParmType(), and this results in a division by
0 in some gspca subdrivers:

  divide error: 0000 [#1] SMP
  Modules linked in: gspca_ov534 gspca_main ...
  CPU: 0 PID: 17201 Comm: v4l2-compliance Not tainted 4.3.0-rc2-ao2 #1
  Hardware name: System manufacturer System Product Name/M2N-E SLI, BIOS
    ASUS M2N-E SLI ACPI BIOS Revision 1301 09/16/2010
  task: ffff8800818306c0 ti: ffff880095c4c000 task.ti: ffff880095c4c000
  RIP: 0010:[<ffffffffa079bd62>]  [<ffffffffa079bd62>] sd_set_streamparm+0x12/0x60 [gspca_ov534]
  RSP: 0018:ffff880095c4fce8  EFLAGS: 00010296
  RAX: 0000000000000000 RBX: ffff8800c9522000 RCX: ffffffffa077a140
  RDX: 0000000000000000 RSI: ffff880095e0c100 RDI: ffff8800c9522000
  RBP: ffff880095e0c100 R08: ffffffffa077a100 R09: 00000000000000cc
  R10: ffff880067ec7740 R11: 0000000000000016 R12: ffffffffa07bb400
  R13: 0000000000000000 R14: ffff880081b6a800 R15: 0000000000000000
  FS:  00007fda0de78740(0000) GS:ffff88012fc00000(0000) knlGS:0000000000000000
  CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
  CR2: 00000000014630f8 CR3: 00000000cf349000 CR4: 00000000000006f0
  Stack:
   ffffffffa07a6431 ffff8800c9522000 ffffffffa077656e 00000000c0cc5616
   ffff8800c9522000 ffffffffa07a5e20 ffff880095e0c100 0000000000000000
   ffff880067ec7740 ffffffffa077a140 ffff880067ec7740 0000000000000016
  Call Trace:
   [<ffffffffa07a6431>] ? v4l_s_parm+0x21/0x50 [videodev]
   [<ffffffffa077656e>] ? vidioc_s_parm+0x4e/0x60 [gspca_main]
   [<ffffffffa07a5e20>] ? __video_do_ioctl+0x280/0x2f0 [videodev]
   [<ffffffffa07a5ba0>] ? video_ioctl2+0x20/0x20 [videodev]
   [<ffffffffa07a59b9>] ? video_usercopy+0x319/0x4e0 [videodev]
   [<ffffffff81182dc1>] ? page_add_new_anon_rmap+0x71/0xa0
   [<ffffffff811afb92>] ? mem_cgroup_commit_charge+0x52/0x90
   [<ffffffff81179b18>] ? handle_mm_fault+0xc18/0x1680
   [<ffffffffa07a15cc>] ? v4l2_ioctl+0xac/0xd0 [videodev]
   [<ffffffff811c846f>] ? do_vfs_ioctl+0x28f/0x480
   [<ffffffff811c86d4>] ? SyS_ioctl+0x74/0x80
   [<ffffffff8154a8b6>] ? entry_SYSCALL_64_fastpath+0x16/0x75
  Code: c7 93 d9 79 a0 5b 5d e9 f1 f3 9a e0 0f 1f 00 66 2e 0f 1f 84 00
    00 00 00 00 66 66 66 66 90 53 31 d2 48 89 fb 48 83 ec 08 8b 46 10 <f7>
    76 0c 80 bf ac 0c 00 00 00 88 87 4e 0e 00 00 74 09 80 bf 4f
  RIP  [<ffffffffa079bd62>] sd_set_streamparm+0x12/0x60 [gspca_ov534]
   RSP <ffff880095c4fce8>
  ---[ end trace 279710c2c6c72080 ]---

Following what the doc says about a zeroed timeperframe (see
http://www.linuxtv.org/downloads/v4l-dvb-apis/vidioc-g-parm.html):

  ...
  To reset manually applications can just set this field to zero.

fix the issue by resetting the frame rate to a default value in case of
an unusable timeperframe.

The fix is done in the subdrivers instead of gspca.c because only the
subdrivers have notion of a default frame rate to reset the camera to.

Signed-off-by: Antonio Ospite <ao2@ao2.it>
Cc: stable@vger.kernel.org
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] cx231xx: fix bulk transfer mode
Terry Heo [Tue, 22 Sep 2015 08:18:05 +0000 (05:18 -0300)]
[media] cx231xx: fix bulk transfer mode

The current cx231xx driver doesn't work with bulk transfer mode.
This patch makes it possible to use bulk transfer mode.

Signed-off-by: Terry Heo <terryheo@google.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] radio-shark2: constify radio_tea5777_ops structures
Julia Lawall [Fri, 13 Nov 2015 11:38:32 +0000 (09:38 -0200)]
[media] radio-shark2: constify radio_tea5777_ops structures

The radio_tea5777_ops structure is never modified, so declare it as const.

Done with the help of Coccinelle.

Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] media: videobuf2: fix compare_const_fl.cocci warnings
Julia Lawall [Fri, 13 Nov 2015 10:46:00 +0000 (08:46 -0200)]
[media] media: videobuf2: fix compare_const_fl.cocci warnings

Move constants to the right of binary operators.

Generated by: scripts/coccinelle/misc/compare_const_fl.cocci

CC: Junghak Sung <jh1009.sung@samsung.com>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Julia Lawall <julia.lawall@lip6.fr>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] Docbook: media: Document changes on struct v4l2_ext_controls
Ricardo Ribalda [Thu, 29 Oct 2015 10:10:32 +0000 (08:10 -0200)]
[media] Docbook: media: Document changes on struct v4l2_ext_controls

Vidioc-g-ext-ctrls can now be used to get the default value of the
controls.

Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@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/usb/pvrusb2: Support for V4L2_CTRL_WHICH_DEF_VAL
Ricardo Ribalda [Thu, 29 Oct 2015 10:10:31 +0000 (08:10 -0200)]
[media] media/usb/pvrusb2: Support for V4L2_CTRL_WHICH_DEF_VAL

This driver does not use the control infrastructure.
Add support for the new field which on structure
 v4l2_ext_controls

Acked-by: Mike Isely <isely@pobox.com>
Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@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] usb/uvc: Support for V4L2_CTRL_WHICH_DEF_VAL
Ricardo Ribalda [Thu, 29 Oct 2015 10:10:30 +0000 (08:10 -0200)]
[media] usb/uvc: Support for V4L2_CTRL_WHICH_DEF_VAL

This driver does not use the control infrastructure.
Add support for the new field which on structure
 v4l2_ext_controls

Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@gmail.com>
Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.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/v4l2-core: struct struct v4l2_ext_controls param which
Ricardo Ribalda [Thu, 29 Oct 2015 10:10:29 +0000 (08:10 -0200)]
[media] media/v4l2-core: struct struct v4l2_ext_controls param which

Support for new field which on v4l2_ext_controls, used to get the
default value of one or more controls.

Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@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/core: Replace ctrl_class with which
Ricardo Ribalda [Thu, 29 Oct 2015 10:10:28 +0000 (08:10 -0200)]
[media] media/core: Replace ctrl_class with which

Replace the obsolete field ctrl_class with "which".

Make sure it not used in future modules by commenting out the field with
ifndef __KERNEL_ .

The field cannot be simply removed because that would be change on the
kenel API to the userspace (and we don't like that).

Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@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] videodev2.h: Extend struct v4l2_ext_controls
Ricardo Ribalda [Thu, 29 Oct 2015 10:10:27 +0000 (08:10 -0200)]
[media] videodev2.h: Extend struct v4l2_ext_controls

So it can be used to get the default value of a control.

Without this change it is not possible to get the
default value of array controls.

Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@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] adv7511: fix incorrect bit offset
Hans Verkuil [Thu, 12 Nov 2015 15:30:39 +0000 (13:30 -0200)]
[media] adv7511: fix incorrect bit offset

The quantization bits are in bits 7-6, not 7-4, so shift by 6 instead of 4.

This bug is caused by a typo in the adv7511 datasheet.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] v4l2-dv-timings: Compare horizontal blanking
Mats Randgaard [Thu, 12 Nov 2015 12:34:55 +0000 (10:34 -0200)]
[media] v4l2-dv-timings: Compare horizontal blanking

hsync and hbackporch must also be compared

Signed-off-by: Mats Randgaard <matrandg@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] v4l2-core/v4l2-ctrls: Filter NOOP CH_RANGE events
Ricardo Ribalda [Wed, 11 Nov 2015 11:58:34 +0000 (09:58 -0200)]
[media] v4l2-core/v4l2-ctrls: Filter NOOP CH_RANGE events

If modify_range is called but no range is changed, do not send the
CH_RANGE event.

Reported-by: Dimitrios Katsaros <patcherwork@gmail.com>
Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@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] av7110: potential divide by zero
Dan Carpenter [Tue, 10 Nov 2015 22:23:53 +0000 (20:23 -0200)]
[media] av7110: potential divide by zero

"len" comes from dvb_video_ioctl() and there is a possibility that it is
zero.  We do a divide by len later in the function so that's not ok.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] av7110: don't allow negative volumes
Dan Carpenter [Tue, 10 Nov 2015 22:22:49 +0000 (20:22 -0200)]
[media] av7110: don't allow negative volumes

The issue here is that we there is a static checker warning because we
have a user controlled volume setting and we cap the upper bound but we
allow negative numbers.  Negative volumes don't make sense, so let's
make these variables unsigned.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] vb2: fix a regression in poll() behavior for output,streams
Hans Verkuil [Thu, 29 Oct 2015 05:02:06 +0000 (03:02 -0200)]
[media] vb2: fix a regression in poll() behavior for output,streams

In the 3.17 kernel the poll() behavior changed for output streams:
as long as not all buffers were queued up poll() would return that
userspace can write. This is fine for the write() call, but when
using stream I/O this changed the behavior since the expectation
was that it would wait for buffers to become available for dequeuing.

This patch only enables the check whether you can queue buffers
for file I/O only, and skips it for stream I/O.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Cc: <stable@vger.kernel.org> # for v3.17 and up
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] vivid: fix compliance error
Hans Verkuil [Thu, 29 Oct 2015 04:59:37 +0000 (02:59 -0200)]
[media] vivid: fix compliance error

If vivid is loaded with the no_error_inj=1 option, then v4l2-compliance will
fail for the video and vbi output nodes because the vivid control class has no
controls.

Don't add the control class for video and vbi output if no_error_inj is true.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] go7007: fix broken test
Hans Verkuil [Thu, 22 Oct 2015 08:56:21 +0000 (06:56 -0200)]
[media] go7007: fix broken test

The wrong flags field was tested for the GO7007_BOARD_HAS_AUDIO flag: that
flag is in board->main_info.flags, not in board->flags.

Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] DocBook media: s/input stream/capture stream/
Hans Verkuil [Thu, 22 Oct 2015 08:46:53 +0000 (06:46 -0200)]
[media] DocBook media: s/input stream/capture stream/

The term "input stream" is confusing since it is not clear whether this is an
input stream from the point of view of the application or from the point of
view of the hardware. So replace it with the more standard term "capture stream".

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] staging: media: davinci_vpfe: fix ipipe_mode type
Andrzej Hajda [Mon, 21 Sep 2015 13:33:57 +0000 (10:33 -0300)]
[media] staging: media: davinci_vpfe: fix ipipe_mode type

The variable can take negative values.

The problem has been detected using proposed semantic patch
scripts/coccinelle/tests/unsigned_lesser_than_zero.cocci [1].

[1]: http://permalink.gmane.org/gmane.linux.kernel/2038576

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] staging: media: davinci_vpfe: Fix over 80 characters coding style issue
Junsu Shin [Mon, 10 Aug 2015 21:40:59 +0000 (18:40 -0300)]
[media] staging: media: davinci_vpfe: Fix over 80 characters coding style issue

This is a patch to the dm365_ipipe.c that fixes over 80 characters
warning detected.

Signed-off-by: Junsu Shin <jjunes0@gmail.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] staging/davinci/vpfe/dm365: add missing dependencies
Arnd Bergmann [Wed, 28 Jan 2015 21:17:43 +0000 (19:17 -0200)]
[media] staging/davinci/vpfe/dm365: add missing dependencies

This driver can only be built when VIDEO_V4L2_SUBDEV_API
and VIDEO_DAVINCI_VPBE_DISPLAY are also provided by the
kernel.

drivers/staging/media/davinci_vpfe/dm365_isif.c: In function '__isif_get_format':
drivers/staging/media/davinci_vpfe/dm365_isif.c:1410:3: error: implicit declaration of function 'v4l2_subdev_get_try_format' [-Werror=implicit-function-declaration]
   return v4l2_subdev_get_try_format(fh, pad);
   ^

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Lad, Prabhakar <prabhakar.csengg@gmail.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] staging: media: davinci_vpfe: drop condition with no effect
Nicholas Mc Guire [Mon, 26 Jan 2015 07:27:05 +0000 (05:27 -0200)]
[media] staging: media: davinci_vpfe: drop condition with no effect

As the if and else branch body are identical the condition has no effect and
can be dropped.

Signed-off-by: Nicholas Mc Guire <der.herr@hofr.at>
Acked-by: Lad, Prabhakar <prabhakar.csengg@gmail.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] drivers/staging/media/davinci_vpfe/vpfe_mc_capture.c: use correct structure...
Julia Lawall [Tue, 29 Jul 2014 15:16:43 +0000 (12:16 -0300)]
[media] drivers/staging/media/davinci_vpfe/vpfe_mc_capture.c: use correct structure type name in sizeof

Correct typo in the name of the type given to sizeof.  Because it is the
size of a pointer that is wanted, the typo has no impact on compilation or
execution.

This problem was found using Coccinelle (http://coccinelle.lip6.fr/).  The
semantic patch used can be found in message 0 of this patch series.

[Replace sizeof(type) by sizeof(variable)]]

Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] include/media: move platform_data to linux/platform_data/media
Mauro Carvalho Chehab [Mon, 16 Nov 2015 10:35:53 +0000 (08:35 -0200)]
[media] include/media: move platform_data to linux/platform_data/media

Let's not mix platform_data headers with the core headers. Instead, let's
create a subdir at linux/platform_data and move the headers to that
common place, adding it to MAINTAINERS.

The headers were moved with:
mkdir include/linux/platform_data/media/; git mv include/media/gpio-ir-recv.h include/media/ir-rx51.h include/media/mmp-camera.h include/media/omap1_camera.h include/media/omap4iss.h include/media/s5p_hdmi.h include/media/si4713.h include/media/sii9234.h include/media/smiapp.h include/media/soc_camera.h include/media/soc_camera_platform.h include/media/timb_radio.h include/media/timb_video.h include/linux/platform_data/media/

And the references fixed with this script:
    MAIN_DIR="linux/platform_data/"
    PREV_DIR="media/"
    DIRS="media/"

    echo "Checking affected files" >&2
    for i in $DIRS; do
for j in $(find include/$MAIN_DIR/$i -type f -name '*.h'); do
 n=`basename $j`
git grep -l $n
done
    done|sort|uniq >files && (
echo "Handling files..." >&2;
echo "for i in \$(cat files|grep -v Documentation); do cat \$i | \\";
(
cd include/$MAIN_DIR;
for j in $DIRS; do
for i in $(ls $j); do
echo "perl -ne 's,(include [\\\"\\<])$PREV_DIR($i)([\\\"\\>]),\1$MAIN_DIR$j\2\3,; print \$_' |\\";
done;
done;
echo "cat > a && mv a \$i; done";
);
echo "Handling documentation..." >&2;
echo "for i in MAINTAINERS \$(cat files); do cat \$i | \\";
(
cd include/$MAIN_DIR;
for j in $DIRS; do
for i in $(ls $j); do
echo "  perl -ne 's,include/$PREV_DIR($i)\b,include/$MAIN_DIR$j\1,; print \$_' |\\";
done;
done;
echo "cat > a && mv a \$i; done"
);
    ) >script && . ./script

Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Lee Jones <lee.jones@linaro.org>
Acked-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
9 years ago[media] include/media: move driver interface headers to a separate dir
Mauro Carvalho Chehab [Fri, 13 Nov 2015 21:40:07 +0000 (19:40 -0200)]
[media] include/media: move driver interface headers to a separate dir

Let's not mix headers used by the core with those headers that
are needed by some driver-specific interface header.

The headers used on drivers were manually moved using:
    mkdir include/media/drv-intf/
    git mv include/media/cx2341x.h include/media/cx25840.h \
include/media/exynos-fimc.h include/media/msp3400.h \
include/media/s3c_camif.h include/media/saa7146.h \
include/media/saa7146_vv.h  include/media/sh_mobile_ceu.h \
include/media/sh_mobile_csi2.h include/media/sh_vou.h \
include/media/si476x.h include/media/soc_mediabus.h \
include/media/tea575x.h include/media/drv-intf/

And the references for those headers were corrected using:

    MAIN_DIR="media/"
    PREV_DIR="media/"
    DIRS="drv-intf/"

    echo "Checking affected files" >&2
    for i in $DIRS; do
for j in $(find include/$MAIN_DIR/$i -type f -name '*.h'); do
 n=`basename $j`
git grep -l $n
done
    done|sort|uniq >files && (
echo "Handling files..." >&2;
echo "for i in \$(cat files|grep -v Documentation); do cat \$i | \\";
(
cd include/$MAIN_DIR;
for j in $DIRS; do
for i in $(ls $j); do
echo "perl -ne 's,(include [\\\"\\<])$PREV_DIR($i)([\\\"\\>]),\1$MAIN_DIR$j\2\3,; print \$_' |\\";
done;
done;
echo "cat > a && mv a \$i; done";
);
echo "Handling documentation..." >&2;
echo "for i in MAINTAINERS \$(cat files); do cat \$i | \\";
(
cd include/$MAIN_DIR;
for j in $DIRS; do
for i in $(ls $j); do
echo "  perl -ne 's,include/$PREV_DIR($i)\b,include/$MAIN_DIR$j\1,; print \$_' |\\";
done;
done;
echo "cat > a && mv a \$i; done"
);
    ) >script && . ./script

Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
9 years ago[media] include/media: split I2C headers from V4L2 core
Mauro Carvalho Chehab [Tue, 10 Nov 2015 14:01:44 +0000 (12:01 -0200)]
[media] include/media: split I2C headers from V4L2 core

Currently, include/media is messy, as it contains both the V4L2 core
headers and some driver-specific headers on the same place. That makes
harder to identify what core headers should be documented and what
headers belong to I2C drivers that are included only by bridge/main
drivers that would require the functions provided by them.

Let's move those i2c specific files to its own subdirectory.

The files to move were produced via the following script:
mkdir include/media/i2c
(cd include/media; for i in *.h; do n=`echo $i|sed s/.h$/.c/`; if [ -e ../../drivers/media/i2c/$n ]; then echo $i; git mv $i i2c/; fi; done)
(cd include/media; for i in *.h; do n=`echo $i|sed s/.h$/.c/`; if [ -e ../../drivers/media/*/i2c/$n ]; then echo $i; git mv $i i2c/; fi; done)
for i in include/media/*.h; do n=`basename $i`;  (for j in $(git grep -l $n); do dirname $j; done)|sort|uniq|grep -ve '^.$' > list; num=$(wc -l list|cut -d' ' -f1); if [ $num == 1 ]; then if [ "`grep i2c list`" != "" ]; then git mv $i include/media/i2c; fi; fi; done

And the references corrected via this script:
    MAIN_DIR="media/"
    PREV_DIR="media/"
    DIRS="i2c/"

    echo "Checking affected files" >&2
    for i in $DIRS; do
for j in $(find include/$MAIN_DIR/$i -type f -name '*.h'); do
 n=`basename $j`
git grep -l $n
done
    done|sort|uniq >files && (
echo "Handling files..." >&2;
echo "for i in \$(cat files|grep -v Documentation); do cat \$i | \\";
(
cd include/$MAIN_DIR;
for j in $DIRS; do
for i in $(ls $j); do
echo "perl -ne 's,(include [\\\"\\<])$PREV_DIR($i)([\\\"\\>]),\1$MAIN_DIR$j\2\3,; print \$_' |\\";
done;
done;
echo "cat > a && mv a \$i; done";
);
echo "Handling documentation..." >&2;
echo "for i in MAINTAINERS \$(cat files); do cat \$i | \\";
(
cd include/$MAIN_DIR;
for j in $DIRS; do
for i in $(ls $j); do
echo "  perl -ne 's,include/$PREV_DIR($i)\b,include/$MAIN_DIR$j\1,; print \$_' |\\";
done;
done;
echo "cat > a && mv a \$i; done"
);
    ) >script && . ./script

Merged Sakari Ailus patch that moves smiapp.h to include/media/i2c.

Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
9 years ago[media] bz#75751: Move internal header file lirc.h to uapi/
Alec Leamas [Thu, 12 Nov 2015 18:03:00 +0000 (16:03 -0200)]
[media] bz#75751: Move internal header file lirc.h to uapi/

The file include/media/lirc.h describes a public interface and
should thus be a public header. See kernel bug
https://bugzilla.kernel.org/show_bug.cgi?id=75751 which has
a manpage describing the interface + an acknowledgment that this
info belongs to uapi.

Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] DocBook/media/Makefile: Do not fail mkdir if dir already exists
Graham Whaley [Thu, 12 Nov 2015 20:51:51 +0000 (18:51 -0200)]
[media] DocBook/media/Makefile: Do not fail mkdir if dir already exists

Commit 5240f4e68d42 ("[media] DocBook/media/Makefile: Avoid make htmldocs
to fail") introduced a mkdir which is always called through
install_media_images from the Documentation/DocBook/Makefile htmldocs rule.
If you run 'make htmldocs' more than once you get:

 mkdir: cannot create directory â€˜./Documentation/DocBook//media_api’:
  File exists

Add -p to the mkdir to continue no matter if the dir already exists.

Signed-off-by: Graham Whaley <graham.whaley@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] dvb_frontend.h: improve documentation for struct dvb_tuner_ops
Mauro Carvalho Chehab [Wed, 11 Nov 2015 21:14:44 +0000 (19:14 -0200)]
[media] dvb_frontend.h: improve documentation for struct dvb_tuner_ops

Improve the comments at the header, removing kernel-doc
tag from where it doesn't belong, grouping the legacy tuner
functions, and improving the text.

No functional changes.

Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] dvb_frontend: get rid of set_state ops & related data
Mauro Carvalho Chehab [Wed, 11 Nov 2015 20:57:16 +0000 (18:57 -0200)]
[media] dvb_frontend: get rid of set_state ops & related data

The get_state()/set_state and the corresponding data types
(struct tuner_state and enum tuner_param) are old DVB interfaces
that came from the DVBv3 time.

Nowadays, set_params() provide a better way to set the tuner
and demod parameters. So, no need to keep those legacy stuff,
as all drivers that were using it got converted.

With this patch, all kABI elements at dvb_frontend.h are now
documented.

Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] stb6100: get rid of get_state()/set_state()
Mauro Carvalho Chehab [Wed, 11 Nov 2015 20:42:00 +0000 (18:42 -0200)]
[media] stb6100: get rid of get_state()/set_state()

It is tricky to get rid of those ops here, as the stv0299 driver
wants to set frequency in separate from setting the bandwidth.

So, we use a small trick: we temporarely fill the cache with
0 for either frequency or bandwidth and add some logic at
set_params to only change the property(ies) that aren't zero.

Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] tda6655: get rid of get_state()/set_state()
Mauro Carvalho Chehab [Wed, 11 Nov 2015 20:14:08 +0000 (18:14 -0200)]
[media] tda6655: get rid of get_state()/set_state()

Those ops aren't used by any driver, with is weird. I suspect
that mantis_vb3030 driver were not working properly...

Anyway, now that the driver uses the set_parms, the DVB
frontend core should do the right thing.

Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] tda8261: don't use set_state/get_state callbacks
Mauro Carvalho Chehab [Wed, 11 Nov 2015 20:07:44 +0000 (18:07 -0200)]
[media] tda8261: don't use set_state/get_state callbacks

Those callbacks are meant to be used only on some very specific
cases. There's absolutely no need to do that at tda8261, as
the only parameter that it allows to be set/get is the frequency.

So, use the standard get_params() and get_frequency() kABI
ops.

There's no need to touch at any bridge driver, as all interactions
are done via the macros at tda8261_cfg.h.

Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] tda666x: add support for set_parms() and get_frequency()
Mauro Carvalho Chehab [Wed, 11 Nov 2015 19:40:27 +0000 (17:40 -0200)]
[media] tda666x: add support for set_parms() and get_frequency()

Those two callbacks are the ones that should be used by normal
DVB frontend drivers.

Add support for them.

Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] tda665x: split set_frequency from set_state
Mauro Carvalho Chehab [Wed, 11 Nov 2015 19:33:59 +0000 (17:33 -0200)]
[media] tda665x: split set_frequency from set_state

On tda665x, set_state only sets frequency. As the kABI for set_state
is meant to be used only on special cases, split the function
into two, in order to allow it to be latter used by a DVBv5
cache params logic.

Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] stb6100: get rid of tuner_state at struct stb6100_state
Mauro Carvalho Chehab [Wed, 11 Nov 2015 18:52:46 +0000 (16:52 -0200)]
[media] stb6100: get rid of tuner_state at struct stb6100_state

The stb6100 driver has a struct tuner_state on its state
struct, that it is used only to store the bandwidth. Even so,
this struct is not really used, as every time the bandwidth
is get or set, it goes through the hardware.

So, get rid of that.

Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] dvb_frontend.h: get rid of unused tuner params/states
Mauro Carvalho Chehab [Wed, 11 Nov 2015 17:52:25 +0000 (15:52 -0200)]
[media] dvb_frontend.h: get rid of unused tuner params/states

There are several tuner_param values that aren't by any driver or core:
DVBFE_TUNER_TUNERSTEP
DVBFE_TUNER_IFFREQ
DVBFE_TUNER_REFCLOCK
DVBFE_TUNER_IQSENSE
DVBFE_TUNER_DUMMY

Several of those correspond to the values at the tuner_state
struct with is also only initialized by not used anyware:
u32 tunerstep;
u32 ifreq;
u32 refclock;

It doesn't make sense to keep anything at the kABI that it is
not used. So, get rid of them.

Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] dvb_frontend.h: Document suspend/resume functions
Mauro Carvalho Chehab [Tue, 10 Nov 2015 13:54:15 +0000 (11:54 -0200)]
[media] dvb_frontend.h: Document suspend/resume functions

Those functions should be implemented on all drivers. So, document
them.

Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] dvb_frontend: resume tone and voltage
Mauro Carvalho Chehab [Tue, 10 Nov 2015 13:50:30 +0000 (11:50 -0200)]
[media] dvb_frontend: resume tone and voltage

As SEC tone and voltage could have changed during
suspend(), restore them to their previous values at
resume().

Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] demux.h: Some documentation fixups for the header
Mauro Carvalho Chehab [Tue, 10 Nov 2015 13:14:52 +0000 (11:14 -0200)]
[media] demux.h: Some documentation fixups for the header

The DocBook description of this header has two issues:
- It calls the Kernel ABI as API, instead of kABI;
- It mentions that the DVB frontend kABI is not described
  within the document. As this will actually generate a
  single DocBook, this is actually not true, now that
  the documentation for the frontend was added.

So, fix both issues.

Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] dvb_frontend.h: Add a description for the header
Mauro Carvalho Chehab [Tue, 10 Nov 2015 13:12:30 +0000 (11:12 -0200)]
[media] dvb_frontend.h: Add a description for the header

This header file provides the kABI functions used by the
Digital TV Frontend core support. Add a description for
this kABI, to add at the device_drivers Kernel DocBook.

Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] dvb_frontend: document the most used functions
Mauro Carvalho Chehab [Tue, 10 Nov 2015 12:46:25 +0000 (10:46 -0200)]
[media] dvb_frontend: document the most used functions

Documents the most used functions at the Digital TV
kABI: dvb_frontend_register(), dvb_frontend_unregister()
and dvb_frontend_detach().

Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] Document the obscure dvb_frontend_reinitialise()
Mauro Carvalho Chehab [Tue, 10 Nov 2015 12:26:39 +0000 (10:26 -0200)]
[media] Document the obscure dvb_frontend_reinitialise()

The dvb_frontend_reinitialise() function is a special case
used by just one frontend. Document it, for completeness.

Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] dvb: document dvb_frontend_sleep_until()
Mauro Carvalho Chehab [Tue, 10 Nov 2015 01:24:10 +0000 (23:24 -0200)]
[media] dvb: document dvb_frontend_sleep_until()

This function is used mainly at the DVB core, in order to provide
emulation for a legacy ioctl. The only current exception is
the stv0299 driver, with takes more than 8ms to switch voltage,
breaking the emulation for FE_DISHNETWORK_SEND_LEGACY_CMD.

Document that.

Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] device-drivers.tmpl: better organize DVB function calls
Mauro Carvalho Chehab [Wed, 4 Nov 2015 14:07:02 +0000 (12:07 -0200)]
[media] device-drivers.tmpl: better organize DVB function calls

Classify the functions at the DVB core per API. That makes easier
to understand how they're related to the userspace API.

Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] demux.h: move documentation overview from device-drivers.tmpl
Mauro Carvalho Chehab [Wed, 4 Nov 2015 13:07:09 +0000 (11:07 -0200)]
[media] demux.h: move documentation overview from device-drivers.tmpl

It is better to keep the documentation overview at the header file,
as this makes easier for developers to remember to fix when needed.

Suggested-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] ivtv: avoid going past input/audio array
Mauro Carvalho Chehab [Wed, 11 Nov 2015 11:27:42 +0000 (09:27 -0200)]
[media] ivtv: avoid going past input/audio array

As reported by smatch:
drivers/media/pci/ivtv/ivtv-driver.c:832 ivtv_init_struct2() error: buffer overflow 'itv->card->video_inputs' 6 <= 6

Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years ago[media] Revert "[media] ivtv: avoid going past input/audio array"
Mauro Carvalho Chehab [Wed, 11 Nov 2015 11:22:36 +0000 (09:22 -0200)]
[media] Revert "[media] ivtv: avoid going past input/audio array"

This patch broke ivtv logic, as reported at
 https://bugzilla.redhat.com/show_bug.cgi?id=1278942

This reverts commit 09290cc885937cab3b2d60a6d48fe3d2d3e04061.

Cc: stable@vger.kernel.org # for v4.1 and upper
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
9 years agoMerge tag 'v4.4-rc1' into patchwork
Mauro Carvalho Chehab [Mon, 16 Nov 2015 09:27:01 +0000 (07:27 -0200)]
Merge tag 'v4.4-rc1' into patchwork

Linux 4.4-rc1

* tag 'v4.4-rc1': (12900 commits)
  Linux 4.4-rc1
  ARC: Fix silly typo in MAINTAINERS file
  ARC: cpu_relax() to be compiler barrier even for UP
  ARC: use ASL assembler mnemonic
  ARC: [arcompact] Handle bus error from userspace as Interrupt not exception
  ARC: remove extraneous header include
  f2fs: xattr simplifications
  squashfs: xattr simplifications
  9p: xattr simplifications
  xattr handlers: Pass handler to operations instead of flags
  jffs2: Add missing capability check for listing trusted xattrs
  hfsplus: Remove unused xattr handler list operations
  ubifs: Remove unused security xattr handler
  vfs: Fix the posix_acl_xattr_list return value
  vfs: Check attribute names in posix acl xattr handers
  mpt3sas: fix inline markers on non inline function declarations
  dax: fix __dax_pmd_fault crash
  Revert "drm/rockchip: Convert the probe function to the generic drm_of_component_probe()"
  drm: Don't oops in drm_calc_timestamping_constants() if drm_vblank_init() wasn't called
  ALSA: pci: depend on ZONE_DMA
  ...

9 years agoLinux 4.4-rc1
Linus Torvalds [Mon, 16 Nov 2015 01:00:27 +0000 (17:00 -0800)]
Linux 4.4-rc1

9 years agoMerge branch 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 15 Nov 2015 17:36:24 +0000 (09:36 -0800)]
Merge branch 'perf-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull perf updates from Thomas Gleixner:
 "Mostly updates to the perf tool plus two fixes to the kernel core code:

   - Handle tracepoint filters correctly for inherited events (Peter
     Zijlstra)

   - Prevent a deadlock in perf_lock_task_context (Paul McKenney)

   - Add missing newlines to some pr_err() calls (Arnaldo Carvalho de
     Melo)

   - Print full source file paths when using 'perf annotate --print-line
     --full-paths' (Michael Petlan)

   - Fix 'perf probe -d' when just one out of uprobes and kprobes is
     enabled (Wang Nan)

   - Add compiler.h to list.h to fix 'make perf-tar-src-pkg' generated
     tarballs, i.e. out of tree building (Arnaldo Carvalho de Melo)

   - Add the llvm-src-base.c and llvm-src-kbuild.c files, generated by
     the 'perf test' LLVM entries, when running it in-tree, to
     .gitignore (Yunlong Song)

   - libbpf error reporting improvements, using a strerror interface to
     more precisely tell the user about problems with the provided
     scriptlet, be it in C or as a ready made object file (Wang Nan)

   - Do not be case sensitive when searching for matching 'perf test'
     entries (Arnaldo Carvalho de Melo)

   - Inform the user about objdump failures in 'perf annotate' (Andi
     Kleen)

   - Improve the LLVM 'perf test' entry, introduce a new ones for BPF
     and kbuild tests to check the environment used by clang to compile
     .c scriptlets (Wang Nan)"

* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (32 commits)
  perf/x86/intel/rapl: Remove the unused RAPL_EVENT_DESC() macro
  tools include: Add compiler.h to list.h
  perf probe: Verify parameters in two functions
  perf session: Add missing newlines to some pr_err() calls
  perf annotate: Support full source file paths for srcline fix
  perf test: Add llvm-src-base.c and llvm-src-kbuild.c to .gitignore
  perf: Fix inherited events vs. tracepoint filters
  perf: Disable IRQs across RCU RS CS that acquires scheduler lock
  perf test: Do not be case sensitive when searching for matching tests
  perf test: Add 'perf test BPF'
  perf test: Enhance the LLVM tests: add kbuild test
  perf test: Enhance the LLVM test: update basic BPF test program
  perf bpf: Improve BPF related error messages
  perf tools: Make fetch_kernel_version() publicly available
  bpf tools: Add new API bpf_object__get_kversion()
  bpf tools: Improve libbpf error reporting
  perf probe: Cleanup find_perf_probe_point_from_map to reduce redundancy
  perf annotate: Inform the user about objdump failures in --stdio
  perf stat: Make stat options global
  perf sched latency: Fix thread pid reuse issue
  ...

9 years agoMerge branch 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 15 Nov 2015 17:35:33 +0000 (09:35 -0800)]
Merge branch 'sched-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull scheduler fix from Thomas Gleixner:
 "A single fix to prevent math underflow in the numa balancing code"

* 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  sched/numa: Fix math underflow in task_tick_numa()

9 years agoMerge branch 'locking-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 15 Nov 2015 17:34:32 +0000 (09:34 -0800)]
Merge branch 'locking-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull liblockdep fixes from Thomas Gleixner:
 "Three small patches to synchronize liblockdep with the latest core
  changes"

* 'locking-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  tools/liblockdep: explicitly declare lockdep API we call from liblockdep
  tools/liblockdep: add userspace versions of WRITE_ONCE and RCU_INIT_POINTER
  tools/liblockdep: remove task argument from debug_check_no_locks_held

9 years agoMerge branch 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 15 Nov 2015 17:32:59 +0000 (09:32 -0800)]
Merge branch 'x86-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull x86 fixes from Thomas Gleixner:
 "A couple of fixes and updates related to x86:

   - Fix the W+X check regression on XEN

   - The real fix for the low identity map trainwreck

   - Probe legacy PIC early instead of unconditionally allocating legacy
     irqs

   - Add cpu verification to long mode entry

   - Adjust the cache topology to AMD Fam17H systems

   - Let Merrifield use the TSC across S3"

* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/cpu: Call verify_cpu() after having entered long mode too
  x86/setup: Fix low identity map for >= 2GB kernel range
  x86/mm: Skip the hypervisor range when walking PGD
  x86/AMD: Fix last level cache topology for AMD Fam17h systems
  x86/irq: Probe for PIC presence before allocating descs for legacy IRQs
  x86/cpu/intel: Enable X86_FEATURE_NONSTOP_TSC_S3 for Merrifield

9 years agoMerge branches 'irq-urgent-for-linus' and 'timers-urgent-for-linus' of git://git...
Linus Torvalds [Sun, 15 Nov 2015 17:30:48 +0000 (09:30 -0800)]
Merge branches 'irq-urgent-for-linus' and 'timers-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull irq and timer fixes from Thomas Gleixner:

 - An irq regression fix to restore the wakeup behaviour of chained
   interrupts.

 - A timer fix for a long standing race versus timers scheduled on a
   target cpu which got exposed by recent changes in the workqueue
   implementation.

* 'irq-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  genirq/PM: Restore system wake up from chained interrupts

* 'timers-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  timers: Use proper base migration in add_timer_on()

9 years agoMerge branch 'upstream' of git://git.linux-mips.org/pub/scm/ralf/upstream-linus
Linus Torvalds [Sun, 15 Nov 2015 17:10:53 +0000 (09:10 -0800)]
Merge branch 'upstream' of git://git.linux-mips.org/ralf/upstream-linus

Pull MIPS updates from Ralf Baechle:
 "These are the highlists of the main MIPS pull request for 4.4:

   - Add latencytop support
   - Support appended DTBs
   - VDSO support and initially use it for gettimeofday.
   - Drop the .MIPS.abiflags and ELF NOTE sections from vmlinux
   - Support for the 5KE, an internal test core.
   - Switch all MIPS platfroms to libata drivers.
   - Improved support, cleanups for ralink and Lantiq platforms.
   - Support for the new xilfpga platform.
   - A number of DTB improvments for BMIPS.
   - Improved support for CM and CPS.
   - Minor JZ4740 and BCM47xx enhancements"

* 'upstream' of git://git.linux-mips.org/pub/scm/ralf/upstream-linus: (120 commits)
  MIPS: idle: add case for CPU_5KE
  MIPS: Octeon: Support APPENDED_DTB
  MIPS: vmlinux: create a section for appended DTB
  MIPS: Clean up compat_siginfo_t
  MIPS: Fix PAGE_MASK definition
  MIPS: BMIPS: Enable GZIP ramdisk and timed printks
  MIPS: Add xilfpga defconfig
  MIPS: xilfpga: Add mipsfpga platform code
  MIPS: xilfpga: Add xilfpga device tree files.
  dt-bindings: MIPS: Document xilfpga bindings and boot style
  MIPS: Make MIPS_CMDLINE_DTB default
  MIPS: Make the kernel arguments from dtb available
  MIPS: Use USE_OF as the guard for appended dtb
  MIPS: BCM63XX: Use pr_* instead of printk
  MIPS: Loongson: Cleanup CONFIG_LOONGSON_SUSPEND.
  MIPS: lantiq: Disable xbar fpi burst mode
  MIPS: lantiq: Force the crossbar to big endian
  MIPS: lantiq: Initialize the USB core on boot
  MIPS: lantiq: Return correct value for fpi clock on ar9
  MIPS: ralink: Add missing clock on rt305x
  ...

9 years agoMerge tag 'sound-fix-4.4-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai...
Linus Torvalds [Sat, 14 Nov 2015 17:43:00 +0000 (09:43 -0800)]
Merge tag 'sound-fix-4.4-rc1' of git://git./linux/kernel/git/tiwai/sound

Pull sound fixes from Takashi Iwai:
 "Here are a collection of small fixes tha have been gathered for
  4.4-rc1.  The only significant changes are those in PCI drivers
  Kconfig, to use "depends on" instead of "select" for CONFIG_ZONE_DMA.
  A reverse select is often more user-friendly, but in this case, it
  makes hard to manage with the conflict with ZONE_DEVICE, so changed in
  such a way for now.

  Others are all small fixes and quirks: an error check in soundcore
  reigster_chrdev(), HD-audio HDMI/DP phantom jack fix, Intel Broxton DP
  quirk, USB-audio DSD device quirk, some constifications, etc"

* tag 'sound-fix-4.4-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound:
  ALSA: pci: depend on ZONE_DMA
  ALSA: hda - Simplify phantom jack handling for HDMI/DP
  ALSA: hda/hdmi - apply Skylake fix-ups to Broxton display codec
  ALSA: ctxfi: constify rsc ops structures
  ALSA: usb: Add native DSD support for Aune X1S
  ALSA: oxfw: add an comment to Kconfig for TASCAM FireOne
  sound: fix check for error condition of register_chrdev()

9 years agoMerge tag 'arc-4.4-rc1-part2' of git://git.kernel.org/pub/scm/linux/kernel/git/vgupta/arc
Linus Torvalds [Sat, 14 Nov 2015 17:09:37 +0000 (09:09 -0800)]
Merge tag 'arc-4.4-rc1-part2' of git://git./linux/kernel/git/vgupta/arc

Pull ARC fixes from Vineet Gupta:
 "Found a couple of brown paper bag bugs with the prev pull request
  (including a SMP build breakage report from Guenter).  Since these are
  urgent I also decided to send over a bunch of other pending fixes
  which could have otherwise waited an rc or two.

  Summary:

   - A bunch of brown paper bag bugs (MAINTAINERS list email, SMP build
     failure)
   - cpu_relax() now compiler barrier for UP as well
   - handling of userspace Bus Errors for ARCompact builds"

* tag 'arc-4.4-rc1-part2' of git://git.kernel.org/pub/scm/linux/kernel/git/vgupta/arc:
  ARC: Fix silly typo in MAINTAINERS file
  ARC: cpu_relax() to be compiler barrier even for UP
  ARC: use ASL assembler mnemonic
  ARC: [arcompact] Handle bus error from userspace as Interrupt not exception
  ARC: remove extraneous header include
  ARCv2: lib: memcpy: use local symbols

9 years agoARC: Fix silly typo in MAINTAINERS file
Vineet Gupta [Sat, 14 Nov 2015 07:28:53 +0000 (12:58 +0530)]
ARC: Fix silly typo in MAINTAINERS file

9 years agoARC: cpu_relax() to be compiler barrier even for UP
Vineet Gupta [Mon, 9 Nov 2015 12:18:34 +0000 (17:48 +0530)]
ARC: cpu_relax() to be compiler barrier even for UP

cpu_relax() on ARC has been barrier only for SMP (and no-op for UP). Per
recent discussions, it is safer to make it a compiler barrier
unconditionally.

Link: http://lkml.kernel.org/r/53A7D3AA.9020100@synopsys.com
Acked-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Vineet Gupta <vgupta@synopsys.com>
9 years agoARC: use ASL assembler mnemonic
Vineet Gupta [Thu, 5 Nov 2015 03:43:31 +0000 (09:13 +0530)]
ARC: use ASL assembler mnemonic

ARCompact and ARCv2 only have ASL, while binutils used to support LSL as
a alias mnemonic.

Newer binutils (upstream) don't want to do that so replace it.

Signed-off-by: Vineet Gupta <vgupta@synopsys.com>
9 years agoARC: [arcompact] Handle bus error from userspace as Interrupt not exception
Vineet Gupta [Fri, 30 Oct 2015 19:52:51 +0000 (01:22 +0530)]
ARC: [arcompact] Handle bus error from userspace as Interrupt not exception

Bus errors from userspace on ARCompact based cores are handled by core
as a high priority L2 interrupt but current code treated it as interrupt
Handling an interrupt like exception is certainly not going to go unnoticed.
(and it worked so far as we never saw a Bus error from userspace until
IPPK guys tested a DDR controller with ECC error detection etc hence
needed to explicitly trigger/handle such errors)

 - So move mem_service exception handler from common code into ARCv2 code.
 - In ARCompact code, define  mem_service as L2 interrupt handler which
   just drops down to pure kernel mode and goes of to enqueue SIGBUS

Reported-by: Nelson Pereira <npereira@synopsys.com>
Tested-by: Ana Martins <amartins@synopsys.com>
Signed-off-by: Vineet Gupta <vgupta@synopsys.com>
9 years agoARC: remove extraneous header include
Vineet Gupta [Fri, 30 Oct 2015 19:17:39 +0000 (00:47 +0530)]
ARC: remove extraneous header include

Signed-off-by: Vineet Gupta <vgupta@synopsys.com>