[media] lirc_zilog: Move constants from ir_probe() into the lirc_driver template

ir_probe() makes a number of constant assignments into the lirc_driver
object after copying in a template.  Make better use of the template.

Signed-off-by: Andy Walls <awalls@md.metrocast.net>
Signed-off-by: Jarod Wilson <jarod@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] lirc_zilog: Always allocate a Rx lirc_buffer object

Always allocate a lirc_buffer object, instead of just upon setup of
the Rx i2c_client.  If we do not allocate a lirc_buffer object, because
we are not handling the Rx i2c_client, lirc_dev will allocate its own
lirc_buffer anyway and not tell us about its location.

Signed-off-by: Andy Walls <awalls@md.metrocast.net>
Signed-off-by: Jarod Wilson <jarod@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] lirc_zilog: Remove unneeded rx->buf_lock

Remove the rx->buf_lock that protected the rx->buf lirc_buffer.  The
underlying operations on the objects within the lirc_buffer are already
protected by spinlocks, or the objects are constant (e.g. chunk_size).

Signed-off-by: Andy Walls <awalls@md.metrocast.net>
Signed-off-by: Jarod Wilson <jarod@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] lirc_zilog: Don't acquire the rx->buf_lock in the poll() function

There is no need to take the rx->buf_lock in the the poll() function
as all the underling calls made on objects in the rx->buf lirc_buffer object
are protected by spinlocks.

Corrected a bad error return value in poll(): return POLLERR instead
of -ENODEV.

Added some comments to poll() for when, in the future, I forget what
poll() and poll_wait() are supposed to do.

[Jarod: minor debug spew fix]

Signed-off-by: Andy Walls <awalls@md.metrocast.net>
Signed-off-by: Jarod Wilson <jarod@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] lirc_zilog: Use kernel standard methods for marking device non-seekable

lirc_zilog had its own llseek stub that returned -ESPIPE.  Get rid of
it and use the kernel's no_llseek() and nonseekable_open() functions
instead.

Signed-off-by: Andy Walls <awalls@md.metrocast.net>
Signed-off-by: Jarod Wilson <jarod@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] lirc_zilog: Convert the instance open count to an atomic_t

The open count is simply used for deciding if the Rx polling thread
needs to poll the IR chip for userspace.  Simplify the manipulation
of the open count by using an atomic_t and not requiring a lock
The polling thread errantly didn't try to take the lock anyway.

Signed-off-by: Andy Walls <awalls@md.metrocast.net>
Signed-off-by: Jarod Wilson <jarod@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] lirc_zilog: Convert ir_device instance array to a linked list

Signed-off-by: Andy Walls <awalls@md.metrocast.net>
Signed-off-by: Jarod Wilson <jarod@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] lirc_zilog: Remove broken, ineffective reference counting

The set_use_inc() and set_use_dec() functions tried to lock
the underlying bridge driver device instance in memory by
changing the use count on the device's i2c_clients.  This
worked for PCI devices (ivtv, cx18, bttv).  It doesn't
work for hot-pluggable usb devices (pvrusb2 and hdpvr).
With usb device instances, the driver may get locked into
memory, but the unplugged hardware is gone.

The set_use_inc() set_use_dec() functions also tried to have
lirc_zilog change its own module refernce count, which is
racy and not guaranteed to work.  The lirc_dev module does
actually perform proper module ref count manipulation on the
lirc_zilog module, so there is need for lirc_zilog to
attempt a buggy module get on itself anyway.

lirc_zilog also errantly called these functions on itself
in open() and close(), but lirc_dev did that already too.

So let's just gut the bodies of the set_use_*() functions,
and remove the extra calls to them from within lirc_zilog.

Proper reference counting of the struct IR, IR_rx, and IR_tx
objects -- to handle the case when the underlying
bttv, ivtv, cx18, hdpvr, or pvrusb2 bridge driver module or
device instance goes away -- will be added in subsequent
patches.

Signed-off-by: Andy Walls <awalls@md.metrocast.net>
Signed-off-by: Jarod Wilson <jarod@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] lirc_zilog: Restore checks for existence of the IR_tx object

This reverts commit 8090232a237ab62e22307fc060097da1a283dd66 and
adds an additional check for ir->tx == NULL.

The user may need us to handle an RX only unit.  Apparently
there are TV capture units in existence with Rx only wiring
and/or RX only firmware for the on-board Zilog Z8 IR unit.

Signed-off-by: Andy Walls <awalls@md.metrocast.net>
Signed-off-by: Jarod Wilson <jarod@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] hdpvr: i2c master enhancements

Make the hdpvr's i2c master implementation more closely mirror that of
the pvrusb2 driver. Currently makes no significant difference in IR
reception behavior with ir-kbd-i2c (i.e., it still sucks).

Signed-off-by: Jarod Wilson <jarod@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] imon: add more panel scancode mappings

Signed-off-by: Jarod Wilson <jarod@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] docs: fix typo in lirc_device_interface.xml

Reported-by: Daniel Burr <dburr@topcon.com>
Signed-off-by: Jarod Wilson <jarod@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] tm6000: fix s-video input

Add compatibility for composite and s-video inputs. Some TV cards hasn't
it.

Fix S-Video input, the s-video cable has only video signals no
audio. Call the function of audio configure kill chroma in signal. only
b/w video.

Known bugs:
  - after s-video the audio for radio didn't work, TV crashed hardly
  - after composite TV crashed hardly too.

P.S. After this patch I'll want to rework the procedure of configure
     video. Now it has a lot of junk and dubles.

With my best regards, Dmitry.

Signed-off-by: Beholder Intl. Ltd. Dmitry Belimov <d.belimov@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] lirc: remove staging lirc_it87 and lirc_ite8709 drivers

Remove older drivers lirc_it87 and lirc_ite8709 from the LIRC staging area,
since they're now superceded by ite-cir.

Signed-off-by: Juan J. Garcia de Soria <skandalfo@gmail.com>
Tested-by: Stephan Raue <stephan@openelec.tv>
Signed-off-by: Jarod Wilson <jarod@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] ite-cir: Fix some CodingStyle issues

Cc: Juan J. Garcia de Soria <skandalfo@gmail.com>
Cc: Stephan Raue <stephan@openelec.tv>
Cc: Jarod Wilson <jarod@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] rc: New rc-based ite-cir driver for several ITE CIRs

This is a second version of an rc-core based driver for the ITE Tech IT8712F
CIR and now for a pair of other variants of the IT8512 CIR too.

This driver should replace the lirc_it87 and lirc_ite8709 currently living in
the LIRC staging directory.

The driver should support the ITE8704, ITE8713, ITE8708 and ITE8709 (this last
one yet untested) PNP ID's.

The code doesn'te reuse code from the pre-existing LIRC drivers, but has been
written from scratch using the nuvoton.cir driver as a skeleton.

This new driver shouldn't exhibit timing problems when running under load (or
with interrupts disabled for relatively long times). It works OOTB with the
RC6 MCE remote bundled with the ASUS EEEBox. TX support is implemented, but
I'm unable to test it since my hardware lacks TX capability.

Signed-off-by: Juan J. Garcia de Soria <skandalfo@gmail.com>
Tested-by: Stephan Raue <stephan@openelec.tv>
Signed-off-by: Jarod Wilson <jarod@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] drivers/media/rc/Kconfig: use tabs, instead of spaces

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] saa7134: Fix strange kconfig dependency on RC_CORE

As the code in saa7134-input is not a module, but the config for it is
set as a boolean instead of a tristate, this causes a strange dependency
on RC_CORE.

VIDEO_SAA7134_RC (which determines if saa7134-input.o is built) depends
on RC_CORE and VIDEO_SAA7134. If VIDEO_SAA7134 is compiled as 'y' but
RC_CORE is compiled as 'm' VIDEO_SAA7134_RC can still be set to 'y'
which causes undefined symbols that it needs from RC_CORE.

The simplest solution is to not allow VIDEO_SAA7134_RC be enabled if
RC_CORE compiled as a module (m) and VIDEO_SA7134 is compiled into the
kernel (y).

Suggested-by: Mauro Carvalho Chehab <mchehab@redhat.com>
Cc: Michal Marek <mmarek@suse.cz>
Cc: linux-kbuild <linux-kbuild@vger.kernel.org>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] STV0288 added full frontend status

status now returns
 FE_HAS_CARRIER
 FE_HAS_SIGNAL
 FE_HAS_VITERBI

Signed-off-by: Malcolm Priestley <tvboxspy@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] gspca - zc3xx: Add exposure control for sensor hv7131r

Signed-off-by: Jean-François Moine <moinejf@free.fr>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] gspca - main: Add endpoint direction test in alt_xfer

This patch fixes a bug in gspca, more precisely in alt_xfer().

This function looks for an input transfer endpoint in an alternate setting.
By default it returns the first endpoint corresponding to the transfer type
indicated in parameter.
But with some USB devices, the first endpoint corresponding to the transfer
type is not always an INPUT endpoint but an OUTPUT one.

This patch adds the endpoint direction test to be sure to return an INPUT endpoint

Signed-off-by: Patrice CHOTARD <patricechotard@free.fr>
Signed-off-by: Jean-François Moine <moinejf@free.fr>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] gspca - nw80x: Fix exposure for some webcams

Signed-off-by: Jean-François Moine <moinejf@free.fr>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] gspca - nw80x: Get the sensor ID when bridge et31x110

Signed-off-by: Jean-François Moine <moinejf@free.fr>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] gspca - nw80x: Fix some image resolutions

No 352x288 for Ds3303 and no 320x240 for Generic802.

Signed-off-by: Jean-François Moine <moinejf@free.fr>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] gspca - nw80x: Check the bridge from the webcam type

Signed-off-by: Jean-François Moine <moinejf@free.fr>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] gspca - nw80x: Fix the gain, exposure and autogain

The autogain now uses common functions.

Signed-off-by: Jean-François Moine <moinejf@free.fr>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] gspca - nw80x: Do some initialization at probe time

Signed-off-by: Jean-François Moine <moinejf@free.fr>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] gspca - nw80x: The webcam dsb-c110 is the same as the twinkle

Signed-off-by: Jean-François Moine <moinejf@free.fr>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] gspca - nw80x: Cleanup source

Signed-off-by: Jean-François Moine <moinejf@free.fr>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] v4l2: use new flag to enable core priority handling

Rather than guess which driver supports core priority handling, require drivers
that do to explicitly set the V4L2_FL_USE_FH_PRIO flag in video_device.

Updated the core prio handling accordingly and set the flag in the three
drivers that do.

Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] ivtv: replace ugly casts with a proper container_of

ivtv-ioctl cast the 'void *fh' directly to 'ivtv_open_id *'. This should be
done properly with a contained_of since the 'void *fh' is really a
'struct v4l2_fh *'.

It worked because the v4l2_fh field is also the first field in the ivtv_open_id
struct, but it is not clean code.

Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] ivtv: add missing v4l2_fh_exit

Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] vivi: convert to core priority handling

Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] dsbr100: ensure correct disconnect sequence

Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] dsbr100: convert to unlocked_ioctl

Use core-assisted locking so .ioctl can be replaced by .unlocked_ioctl.

Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] v4l2-framework.txt: document new v4l2_device release() callback

Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] v4l2-device: add kref and a release function

The video_device struct has proper ref counting and its release function
will be called when the last user releases it. But no such support was
available for struct v4l2_device. This made it hard to determine when a
USB driver can release the device if it has multiple device nodes.

With one device node it is easy of course, since when the device node is
released, the whole device can be released.

This patch adds refcounting to v4l2_device. When registering device nodes
the v4l2_device refcount will be increased, when releasing device nodes
it will be decreased. The (optional) release function will be called when
the last device node was released.

Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] cx18: use core priority handling

VIDIOC_S/G_PRIORITY handling is now done by the v4l2 core framework.

Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] cx18: use v4l2_fh as preparation for adding core priority support

Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] v4l2-framework.txt: improve v4l2_fh/priority documentation

Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] ivtv: convert to core priority handling

Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] v4l2-ioctl: add priority handling support

Drivers that use v4l2_fh can now use the core framework support of g/s_priority.

Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] v4l2-fh: add v4l2_fh_is_singular

Several drivers need to do something when the first filehandle is opened
or the last filehandle is closed. Most implement some use count mechanism,
but if they use v4l2_fh, then you can also just check if this is the only
filehandle for the device node. A simple helper function can do this.

Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] v4l2-fh: add v4l2_fh_open and v4l2_fh_release helper functions

Add two new functions: v4l2_fh_open allocates and initializes a struct v4l2_fh
based on a struct file pointer and v4l2_fh_release releases and frees a struct
v4l2_fh.

Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] v4l2-fh: implement v4l2_priority support

Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] v4l2: add v4l2_prio_state to v4l2_device and video_device

Integrate the v4l2_prio_state into the core, ready for use.

One struct v4l2_prio_state is added to v4l2_device and a pointer
to a prio state is added to video_device.

Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] v4l2_prio: move from v4l2-common to v4l2-dev

We are going to move priority handling into the v4l2 core. As a consequence
the v4l2_prio helper functions need to be moved into the core videodev
module as well to prevent circular dependencies.

Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] saa7134-input: key up events not sent after suspend/resume

On my AverMedia AverTV Studio 507, key up events are no longer sent after
a suspend-to-disk/resume cycle, resulting in "stuck" keys.

Apparently, for key up events to be generated, a certain GPIO pin must be set.
Currently it's set in saa7134_input_init1(), but that function is not called
on device resume. I suggest that code be moved to __saa7134_ir_start(), which
is called both on init and resume.

Signed-off-by: Vadim Solomin <vadic052@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] stv0367: typo in function parameter

The CellsCoeffs arrays are [3][6][5] not [2][6][5].

Signed-off-by: Dan Carpenter <error27@gmail.com>
Cc: Igor M. Liplianin <liplianin@netup.ru>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] stv0367: signedness bug in stv0367_get_tuner_freq()

We use err to store negative error codes so it should be signed.  And
if we return an error from stv0367_get_tuner_freq() that needs to be
handled properly as well.  (param->frequency is a u32).

Signed-off-by: Dan Carpenter <error27@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] via-camera: Fix OLPC serial check

The code that checks the OLPC serial port is never built at the moment,
because CONFIG_OLPC_XO_1_5 doesn't exist and probably won't be added.

Fix it so that it gets compiled in, only executes on OLPC laptops, and
move the check into the probe routine.

The compiler is smart enough to eliminate this code when CONFIG_OLPC=n
(due to machine_is_olpc() always returning false).

Signed-off-by: Daniel Drake <dsd@laptop.org>
Acked-by: Jonathan Corbet <corbet@lwn.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] Change to 32 bit and add other remote controls for lme2510

These bubble button remote controls appear to be generic from China.

These are the three variants known to be supplied with DM04/QQBOX DVB-S

They could well be supplied with other devices from the region.

Signed-off-by: Malcolm Priestley <tvboxspy@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] DM04/QQBOX Update V1.76 - use 32 bit remote decoding

Use 32 bit decoding to add support for more than one variant of remote
control.

Signed-off-by: Malcolm Priestley <tvboxspy@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] DM04 LME2510(C) Sharp BS2F7HZ0194 Firmware Information

DM04 LME2510(C) Sharp BS2F7HZ0194 Firmware Information

Signed-off-by: Malcolm Priestley <tvboxspy@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] v180 - DM04/QQBOX added support for BS2F7HZ0194 versions

Old versions of these boxes have the BS2F7HZ0194 tuner module on
both the LME2510 and LME2510C.

Firmware dvb-usb-lme2510-s0194.fw  and/or dvb-usb-lme2510c-s0194.fw
files are required.

See Documentation/dvb/lmedm04.txt

Signed-off-by: Malcolm Priestley <tvboxspy@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] drivers/media/video/tlg2300/pd-video.c: Remove second mutex_unlock in pd_vidioc_s_fmt

Error path in file drivers/media/video/tlg2300/pd-video.c:
1. First mutex_unlock on &pd->lock in line 767 (in function that
   called from line 805)
2. Second in line  806

 805        pd_vidioc_s_fmt(pd, &f->fmt.pix);
 806        mutex_unlock(&pd->lock);

Found by Linux Device Drivers Verification Project

Signed-off-by: Alexander Strakh <strakh@ispras.ru>
Acked-by: Huang Shijie <shijie8@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] radio-wl1273: remove unused wl1273_device->work

wl1273_device->work is unused.  Remove it along with the spurious
flush_scheduled_work() call in wl1273_fm_module_exit().

Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Matti Aaltonen <matti.j.aaltonen@nokia.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] Zarlink zl10036 DVB-S: Fix mem leak in zl10036_attach

On Thu, 17 Feb 2011, Matthias Schwarzott wrote:

> On Sunday 06 February 2011, Jesper Juhl wrote:
> > If the memory allocation to 'state' succeeds but we jump to the 'error'
> > label before 'state' is assigned to fe->tuner_priv, then the call to
> > 'zl10036_release(fe)' at the 'error:' label will not free 'state', but
> > only what was previously assigned to 'tuner_priv', thus leaking the memory
> > allocated to 'state'.
> > There are may ways to fix this, including assigning the allocated memory
> > directly to 'fe->tuner_priv', but I did not go for that since the
> > additional pointer derefs are more expensive than the local variable, so I
> > just added a 'kfree(state)' call. I guess the call to 'zl10036_release'
> > might not even be needed in this case, but I wasn't sure, so I left it in.
> >
> Yeah, that call to zl10036_release can be completely eleminated.
> Another thing is: jumping to the error label only makes sense when memory was
> already allocated. So the jump in line 471 can be replaced by "return NULL",
> as the other error handling before allocation:
>         if (NULL == config) {
>                 printk(KERN_ERR "%s: no config specified", __func__);
>                 goto error;
>         }
>
> I suggest to improve the patch to clean the code up when changing that.
>
> But I am fine with commiting this patch also if you do not want to change it.
>

Thank you for your feedback. It makes a lot of sense.
Changing it is not a problem :)
How about the updated patch below?

If the memory allocation to 'state' succeeds but we jump to the 'error'
label before 'state' is assigned to fe->tuner_priv, then the call to
'zl10036_release(fe)' at the 'error:' label will not free 'state', but
only what was previously assigned to 'tuner_priv', thus leaking the memory
allocated to 'state'.
This patch fixes the leak and also does not jump to 'error:' before mem
has been allocated but instead just returns. Also some small style
cleanups.

Signed-off-by: Jesper Juhl <jj@chaosbits.net>
Signed-off-by: Matthias Schwarzott <zzam@gentoo.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] V4L: videobuf, don't use dma addr as physical

mem->dma_handle is a dma address obtained by dma_alloc_coherent which
needn't be a physical address in presence of IOMMU, as
a hardware IOMMU can (and most likely) will return a bus address where
physical != bus address.

So ensure we are remapping (remap_pfn_range) the right page in
__videobuf_mmap_mapper by using virt_to_phys(mem->vaddr) and not
mem->dma_handle.

While at it, use PFN_DOWN instead of explicit shift.

[mchehab@redhat.com: Fix compilation breakage due to the lack of a comma]
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] uvcvideo: Fix descriptor parsing for video output devices

Commit 4057ac6ca9a77c4275b34b5925ab5c99557913b1

    V4L/DVB (13505): uvcvideo: Refactor chain scan

broke output terminals parsing. Fix it.

Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Cc: stable@kernel.org
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] uvcvideo: Fix uvc_fixup_video_ctrl() format search

The scheme used to index format in uvc_fixup_video_ctrl() is not robust:
format index is based on descriptor ordering, which does not necessarily
match bFormatIndex ordering.  Searching for first matching format will
prevent uvc_fixup_video_ctrl() from using the wrong format/frame to make
adjustments.

Signed-off-by: Stephan Lachowsky <stephan.lachowsky@maxim-ic.com>
Cc: stable@kernel.org
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] vb2: Handle return value from start_streaming callback

Fix vb2 not handling return value from start_streaming() callback.

Signed-off-by: Pawel Osciak <pawel@osciak.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] vb2: vb2_poll() fix return values for file I/O mode

poll() should be returning poll-specific error values, not E* errors.

Signed-off-by: Pawel Osciak <pawel@osciak.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] Update Pawel Osciak's e-mail address

Signed-off-by: Pawel Osciak <pawel@osciak.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] Make 2.6.39 not 2.6.38 the version when Multi-planar API was added

Multi-planar API was added to 2.6.39 version of Video for Linux 2 API.

Signed-off-by: Pawel Osciak <pawel@osciak.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] af9013: reimplement firmware download

Split FW download packages smarter way and bug free. Implementation is
based of Andrea Merello's example he provided for tda18218 driver.
Count remaining FW bytes down in loop instead of division and modulo
combination used earlier.

Thanks to: Andrea Merello <andrea.merello@gmail.com>

Signed-off-by: Antti Palosaari <crope@iki.fi>
Cc: Andrea Merello <andrea.merello@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] af9013: download FW earlier in attach()

Quick test shows that FW must be running before other checks
so make it happen earlier.

Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] af9015: reimplement firmware download

Split FW download packages smarter way and bug free. Implementation is
based of Andrea Merello's example he provided for tda18218 driver.
Count remaining FW bytes down in loop instead of division and modulo
combination used earlier.

Thanks-to: Andrea Merello <andrea.merello@gmail.com>
Signed-off-by: Antti Palosaari <crope@iki.fi>
Cc: Andrea Merello <andrea.merello@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] af9015: map remote for TerraTec Cinergy T Stick RC

Map rc-terratec-slim-2 for TerraTec Cinergy T Stick RC device.

Signed-off-by: Antti Palosaari <crope@iki.fi>
Cc: Martin Groszhauser <mgroszhauser@gmail.com>
Cc: TerraTux <TerraTux@terratec.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] add TerraTec remote

Remote used for TerraTec Cinergy T Stick RC.
Keytable from Martin Groszhauser <mgroszhauser@gmail.com>

Signed-off-by: Antti Palosaari <crope@iki.fi>
Cc: Martin Groszhauser <mgroszhauser@gmail.com>
Cc: TerraTux <TerraTux@terratec.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] af9015: small RC change

Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] af9015: enhance RC

Patch from Ian Armstrong.

I've encountered a couple of problems with the current af9015 driver as
supplied with the 2.6.37 kernel, that the attached patch appears to fix. (I've
generated this patch against the current v4l-dvb git).

Some key-presses are lost. A key-press is only generated upon 'valid' data
(buf[14] == (u8) ~buf[15]), but the buffer is wiped before this check.
Sometimes the 15th byte has not been set at the time of read, so the data
isn't valid & ignored. On the next poll the 15th byte has been set, but the
rest of the data was wiped previously, so the data is still invalid & the key
is lost.

Weird repeat error, where an old key press is sometimes repeated in error. ie.
button sequence '1 (pause) 2 (pause) 3 (pause) 4' generates output like '1
(pause) 2 (pause) 23 (pause) 4'. The current driver zeroes the data for the
key pushed, but sometimes this data is already zero but with other bytes set
suggesting a repeat code. This results in the last key being incorrectly
repeated.

This patch attempts to reduce the risk of a missed key-press & also stop the
random repeat of an old key-press when a new key is pressed.

Signed-off-by: Ian Armstrong <ian@iarmst.demon.co.uk>
Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] gspca - nw80x: New subdriver for Divio based webcams

[mchehab@redhat.com: Fix a few CodingStyle issues]
Tested-by: Kjell Claesson <kjell.claesson@epost.tidanet.se>
Tested-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Jean-François Moine <moinejf@free.fr>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] gspca - main: Cleanup source

- change copyright and erroneous comment
- small code optimization
- have constant the device template

Signed-off-by: Jean-François Moine <moinejf@free.fr>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] gspca - ov519: Add exposure and autogain controls for ov2610/2610ae

Signed-off-by: Jean-François Moine <moinejf@free.fr>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] gspca - jeilinj / stv06xx: Fix some warnings

Signed-off-by: Jean-François Moine <moinejf@free.fr>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] gspca - sonixb: Clenup source

- update copyright and module author
- set the sensor table as constant

Signed-off-by: Jean-François Moine <moinejf@free.fr>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] gspca - sonixb: Update inactive flags to reflect autogain setting

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Jean-François Moine <moinejf@free.fr>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] gspca - sonixb: Use the new control mechanism

Acked-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Jean-François Moine <moinejf@free.fr>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] gspca: New file autogain_functions.h

This file contains functions for the auto gain.
It must be used with the new control mechanism.

Acked-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Jean-François Moine <moinejf@free.fr>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] gspca - zc3xx: Cleanup source

- change some comments and copyright
- change an erroneous register name
- change hdcs2020b to hdcs2020

Signed-off-by: Jean-François Moine <moinejf@free.fr>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] gspca - zc3xx: Remove double definition

Signed-off-by: Jean-François Moine <moinejf@free.fr>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] gspca_cpia1: Add support for button

Not only the qx3 microscope has a button, but some cameras too. Tested
with the Trust sp@cecam 100 (and with a creative and ezcam without button).

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] pvrusb2: check kmalloc return value

 allows the driver to proceed and initialize the below two values
 even if the kmalloc() fails.
        hdw->std_info_enum.def.type_enum.value_names
        hdw->std_info_enum.def.type_enum.count

Signed-off-by: Xiaochen Wang <wangxiaochen0@gmail.com>
Signed-off-by: Mike Isely <isely@pobox.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] pvrusb2: Remove dead code

This was caught via a compiler warning.  Amazingly enough this bit of
benign dreck dates all the way back to 2008.

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] pvrusb2: Implement support for Terratec Grabster AV400

Signed-off-by: Mike Isely <isely@pobox.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] Fix 'ID nv12mt already defined' error

DocBook validation fix.

Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] V4L DocBook: update V4L2 version

Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] V4L doc fixes

The xmlto validation produced a number of errors that are now fixed.
Sadly, the DocBook/Makefile still adds --skip-validation to xmlto, so
these errors are missed during a normal compile.

Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] v4l2: vb2-dma-sg: fix potential security hole

Memory allocated by alloc_page() function might contain some potentially
important data from other system processes. The patch adds a flag to
zero the allocated page before giving it to videobuf2 (and then to
userspace).

Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] v4l2: vb2: simplify __vb2_queue_free function

__vb2_queue_free function doesn't really return anything useful. This patch
removes support for the return value to simplify the code.

Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] v4l2: vb2: one more fix for REQBUFS()

Return immediately if the target number of buffers is the same as
the current one and memory access type doesn't change.

Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] v4l2: vb2-dma-sg: fix memory leak

This patch fixes two minor memory leaks in videobuf2-dma-sg module. They
might happen only in case some other operations (like memory allocation)
failed.

Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] v4l2: vb2: fix queue reallocation and REQBUFS(0) case

This patch fixes 2 minor bugs in videobuf2 core:
1. Queue should be reallocated if one change the memory access
method without changing the number of buffers.
2. In case of REQBUFS(0), the request should not be passed to the
driver.

Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] v4l: Documentation for the NV12MT format

Added documentation for V4L2_PIX_FMT_NV12MT format. This is a YUV 4:2:0
format with macro block size of 64x32 and specific order of macro blocks
in the memory.

Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] s5p-fimc: Fix G_FMT ioctl handler

Use pix_mp member of struct v4l2_format to return a format
description rather than pix. Also fill in the plane_fmt array.
This is a missing bit of conversion to the multiplanar API.

Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] s5p-fimc: Use dynamic debug

Use pr_debug instead of printk so it is possible to control
debug traces at runtime.
E.g. to enable debug trace in file fimc-core.c use command:
echo -n 'file fimc-core.c +p' > /sys/kernel/debug/dynamic_debug/control
or
echo -n 'file fimc-core.c -p' > /sys/kernel/debug/dynamic_debug/control
to disable.

Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] s5p-fimc: Add a platform data entry for MIPI-CSI data alignment

Allow the MIPI-CSI data alignment to be defined in the board setup
as it may be different across various camera sensors.

Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] s5p-fimc: Allow defining number of sensors at runtime

Add num_clients field to struct s5p_fimc_isp_info to define exactly
size of clients array which simplifies a bit the sensors management.

Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

[media] s5p-fimc: Prevent hanging on device close and fix the locking

Rework the locking in m2m driver to assure proper operation on SMP systems.

When job_abort or stop_streaming was called to immediately shutdown
a memory-to-memory transaction video buffers scheduled for processing
were never returned to vb2 and v4l2_m2m_job_finish was not called
which led to hanging.

Correct this and also return the unprocessed buffers to vb2 marking
them as erroneous, in case the end of frame interrupt do not occur.

Reported-by: Sewoon Park <seuni.park@samsung.com>
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>