[media] move soc_camera to its own directory

That helps to better organize the soc_camera items.

While here, cleanup Makefiles, removing uneeded include dirs.

Signed-off-by: Mauro Carvalho Chehab <[email protected]>

sh_mobile_csi2: move it to the right place

make[4]: *** No rule to make target `drivers/media/platform/sh_mobile_csi2.c',
needed by `drivers/media/platform/sh_mobile_csi2.o'.  Stop.

Reported-by: Hans Verkuil <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] b2c2: export b2c2_flexcop_debug symbol

ERROR: "b2c2_flexcop_debug" [drivers/media/pci/b2c2/b2c2-flexcop-pci.ko] undefined!

Reported-by: Fengguang Wu <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] Fix some Makefile rules

On a few places, := were using instead of +=, causing drivers to
not compile.

While here, standardize the usage of += on all cases where multiple
lines are needed, and for obj-y/obj-m targets, and := when just one
line is needed, on <module>-obj rules.

Reported-by: Hans Verkuil <[email protected]>
Identified-by: Antti Polosaari <[email protected]>
Tested-by: Antti Palosaari <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] cx23885-cards: fix netup card default revision

Netup cards revision 1 are not manufactured anymore.
So, rev. 4 should be default.

Signed-off-by: Anton Nurkin <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] media: soc_camera: don't clear pix->sizeimage in JPEG mode

In JPEG mode, the size of image is variable due to different JPEG compression
rate. We only can get the pix->sizeimage from the user.

If we clear pix->sizeimage in soc_camera_try_fmt() then we will get it from:
ret = soc_mbus_image_size(xlate->host_fmt, pix->bytesperline,
pix->height);
if (ret < 0)
return ret;

pix->sizeimage = max_t(u32, pix->sizeimage, ret);

In general, this sizeimage will be larger than the actul JPEG image size.

But vb2 will check the buffer and size of image in __qbuf_userptr():
/* Check if the provided plane buffer is large enough */
if (planes[plane].length < q->plane_sizes[plane])

So we shouldn't clear the pix->sizeimage and also shouldn't re-calculate
the pix->sizeimage in soc_mbus_image_size() in JPEG mode

We also shouldn't re-calculate pix->bytesperline:
ret = soc_mbus_bytes_per_line(pix->width, xlate->host_fmt);
if (ret < 0)
return ret;

pix->bytesperline = max_t(u32, pix->bytesperline, ret);

pix->bytesperline also should be set by the user or by the driver's
try_fmt() implementation.

Change-Id: I700690a2287346127a624b5260922eaa5427a596

Signed-off-by: Albert Wang <[email protected]>
Signed-off-by: Guennadi Liakhovetski <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] media: mx3_camera: Improve data bus width check code for probe

This patch contains code change only to use the present macro-
MX3_CAMERA_DATAWIDTH_MASK to check valid camera platform data
bus width instead of enumerating every possible data bus width.

Signed-off-by: Liu Ying <[email protected]>
Signed-off-by: Guennadi Liakhovetski <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] mt9v022: Add support for mt9v024

Driver for mt9v022 camera sensor is fully compatible for mt9v024 camera sensor
with the exception of several registers which have been changed addresses.
mt9v024 also has improved and additional features, but they are currently not in use.

Signed-off-by: Alex Gershgorin <[email protected]>
Signed-off-by: Guennadi Liakhovetski <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] media: mx2_camera: Fix clock handling for i.MX27

On i.MX27 two clocks are required: emma-ipg and emma-ahb. The ahb clock
has to be requested using both a device and a connection ID.

Signed-off-by: Javier Martin <[email protected]>
[[email protected]: rebase to the current media tree]
Signed-off-by: Guennadi Liakhovetski <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] media: mx3_camera: buf_init() add buffer state check

This patch checks the state of the buffer when calling .buf_init() method.
This is needed for the USERPTR buffer type, because in that case
.buf_init() is called every time a buffer is queued, and not only once
during the preparation stage, like in the MMAP case. Without this check
buffers get initialised repeatedly, which also leads to the allocation
of new DMA descriptors, of which there is only a final relatively small
number available. Both MMAP and USERPTR methods were successfully tested.

Signed-off-by: Alex Gershgorin <[email protected]>
[[email protected]: remove mx3_camera_buffer::state completely]
Signed-off-by: Guennadi Liakhovetski <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] dvb_usb_v2: ce6230, rtl28xxu use .reset_resume

All these seems to survive .reset_resume.

Signed-off-by: Antti Palosaari <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] dvb_usb_v2: af9015, af9035, anysee use .reset_resume

All these seems to survive .reset_resume.

Signed-off-by: Antti Palosaari <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] dvb_usb_v2: .reset_resume() support

Add .reset_resume() support.
Also some other small changes for suspend / resume.

Signed-off-by: Antti Palosaari <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] dvb_frontend: implement suspend / resume

Move initial suspend / resume support from dvb_usb_v2 to dvb_frontend
as it is dvb general feature that could be used all dvb devices.

Signed-off-by: Antti Palosaari <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] dvb_usb_v2: implement power-management for suspend

Put device full sleep on suspend, wake-up it on resume and acquire
retune in order to return same television channel.

Signed-off-by: Antti Palosaari <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] dvb_core: export function to perform retune

We need to retune when resume from suspend.

Signed-off-by: Antti Palosaari <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] media: coda: add const qualifiers

The commit 98d7bbb9 changed *of_device_id.data to const
which introduced warnings in various places that have mostly
been fixed. This patch fixes one such warning by introducing
two const qualifiers.

GCC warning:
drivers/media/platform/coda.c:1785:16: warning:
        assignment discards ‘const’ qualifier
        from pointer target type [enabled by default]

Signed-off-by: Emil Goode <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] anysee: convert Kernel dev_* logging

Signed-off-by: Antti Palosaari <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] anysee: fix compiler warning

debug_dump macro was defined twice when CONFIG_DVB_USB_DEBUG was
not set. Move debug_dump macro to correct place.

Reported-by: Randy Dunlap <[email protected]>
Signed-off-by: Antti Palosaari <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] drivers/media/usb/{s2255drv.c, tm6000/tm6000-alsa.c, tm6000/tm6000-input.c}: Remove potential NULL dereferences

If the NULL test is necessary, the initialization involving a dereference of
the tested value should be moved after the NULL test.

The sematic patch that fixes this problem is as follows:
(http://coccinelle.lip6.fr/)

// <smpl>
@@
type T;
expression E;
identifier i,fld;
statement S;
@@

- T i = E->fld;
+ T i;
  ... when != E
      when != i
  if (E == NULL) S
+ i = E->fld;
// </smpl>

Signed-off-by: Julia Lawall <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] s5p-tv: Use devm_* functions in sii9234_drv.c file

devm_* functions are device managed functions and make error handling
and cleanup cleaner and simpler.

Signed-off-by: Sachin Kamat <[email protected]>
Signed-off-by: Sylwester Nawrocki <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] s5p-tv: Replace printk with pr_* functions

Replace printk with pr_* functions to silence checkpatch warnings.

Signed-off-by: Sachin Kamat <[email protected]>
Signed-off-by: Sylwester Nawrocki <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] s5p-tv: Use devm_regulator_get() in sdo_drv.c file

devm_regulator_get() is a device managed function and makes the exit code
a bit simpler and cleaner.

Signed-off-by: Sachin Kamat <[email protected]>
Signed-off-by: Sylwester Nawrocki <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] drivers/media/platform/mx2_emmaprp.c: use devm_kzalloc and devm_clk_get

Using devm_kzalloc simplifies the code and ensures that the use of
devm_request_irq is safe.  When kzalloc and kfree were used, the interrupt
could be triggered after the handler's data argument had been freed.

This also introduces some missing initializations of the return variable
ret, and uses devm_request_and_ioremap instead of the combination of
devm_request_mem_region and devm_ioremap.

The problem of a free after a devm_request_irq was found using the
following semantic match (http://coccinelle.lip6.fr/)

// <smpl>
@r exists@
expression e1,e2,x,a,b,c,d;
identifier free;
position p1,p2;
@@

  devm_request_irq@p1(e1,e2,...,x)
  ... when any
      when != e2 = a
      when != x = b
  if (...) {
    ... when != e2 = c
        when != x = d
    free@p2(...,x,...);
    ...
    return ...;
  }
// </smpl>

Signed-off-by: Julia Lawall <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] DocBook validation fixes

More validation fixes as reported by xmllint.

There are still three xmllint errors after this remaining regarding SVG file support.

Signed-off-by: Hans Verkuil <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] DocBook: update RDS references to the latest RDS standards

Signed-off-by: Hans Verkuil <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] stk1160: remove unneeded check

"card" is a valid pointer here because we checked snd_card_create() for
error returns.  Checking after a dereference makes the static checkers
complain.

Signed-off-by: Dan Carpenter <[email protected]>
Acked-by: Ezequiel Garcia <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] stk1160: unlock on error path stk1160_set_alternate()

There are some unlocks missing on error.

Signed-off-by: Dan Carpenter <[email protected]>
Acked-by: Ezequiel Garcia <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] mem2mem_testdev: unlock and return error code properly

We recently added locking to this function, but there was an error path
which accidentally returned holding a lock.  Also we returned zero on
failure on some paths instead of the error code.

Signed-off-by: Dan Carpenter <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] em28xx: use after free in em28xx_v4l2_close()

We need to move the unlock before the kfree(dev);

Signed-off-by: Dan Carpenter <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] it913x-fe: use ARRAY_SIZE() as a cleanup

This code looks suspicious, but it turns out that "nv" is an array of u8
so sizeof() is the same as ARRAY_SIZE().  Using ARRAY_SIZE() is more
readable though.

Signed-off-by: Dan Carpenter <[email protected]>
Cc: Malcolm Priestley <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] V4L: soc-camera: add selection API host operations

Add .get_selection() and .set_selection() soc-camera host driver
operations. Additionally check, that the user is not trying to change the
output sizes during a running capture.

Signed-off-by: Guennadi Liakhovetski <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] media: mx2_camera: Add YUYV output format

Add explicit conversions from UYVY and YUYV to YUYV so that
csicr1 configuration can be set properly for each format.

Signed-off-by: Javier Martin <[email protected]>
Signed-off-by: Guennadi Liakhovetski <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] video: mx2_camera: Use clk_prepare_enable/clk_disable_unprepare

Prepare the clock before enabling it.

Cc: Guennadi Liakhovetski <[email protected]>
Cc: <[email protected]>
Signed-off-by: Fabio Estevam <[email protected]>
Signed-off-by: Guennadi Liakhovetski <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] video: mx1_camera: Use clk_prepare_enable/clk_disable_unprepare

Prepare the clock before enabling it.

Cc: Guennadi Liakhovetski <[email protected]>
Cc: <[email protected]>
Signed-off-by: Fabio Estevam <[email protected]>
Signed-off-by: Guennadi Liakhovetski <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] ov772x: Stop sensor readout right after reset

The sensor starts streaming video as soon as it gets powered or is
reset. Disable the output in the reset function.

Signed-off-by: Laurent Pinchart <[email protected]>
Signed-off-by: Guennadi Liakhovetski <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] ov772x: Compute window size registers at runtime

Instead of hardcoding register arrays, compute the values at runtime.

Signed-off-by: Laurent Pinchart <[email protected]>
[[email protected]: keep (Q)VGA_* macros for now]
Signed-off-by: Guennadi Liakhovetski <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] ov772x: Add support for SBGGR10 format

Signed-off-by: Laurent Pinchart <[email protected]>
Signed-off-by: Guennadi Liakhovetski <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] ov772x: Add ov772x_read() and ov772x_write() functions

And use them instead of calling SMBus access functions directly.

Signed-off-by: Laurent Pinchart <[email protected]>
Signed-off-by: Guennadi Liakhovetski <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] ov772x: Make to_ov772x convert from v4l2_subdev to ov772x_priv

Conversion from i2c_client to ov772x_priv is only needed in a single
location, while conversion from v4l2_subdev to ov772x_priv is needed in
several locations. Perform the former manually, and use to_ov772x for
the later.

Signed-off-by: Laurent Pinchart <[email protected]>
Signed-off-by: Guennadi Liakhovetski <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] ov772x: try_fmt must not default to the current format

If the requested format isn't supported, return a fixed default format
instead of the current format.

Signed-off-by: Laurent Pinchart <[email protected]>
Signed-off-by: Guennadi Liakhovetski <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] ov772x: Don't fail in s_fmt if the requested format isn't supported

Select a default format instead.

Signed-off-by: Laurent Pinchart <[email protected]>
Signed-off-by: Guennadi Liakhovetski <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] ov772x: Select the default format at probe time

The format and window size are only initialized during the first g_fmt
call. This leaves the device in an inconsistent state after
initialization, which will cause problems when implementing pad
operations. Move the format and window size initialization to probe
time.

Signed-off-by: Laurent Pinchart <[email protected]>
Signed-off-by: Guennadi Liakhovetski <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] ov772x: Fix memory leak in probe error path

The control handler isn't freed if its initialization fails. Fix it.

Signed-off-by: Laurent Pinchart <[email protected]>
Signed-off-by: Guennadi Liakhovetski <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] soc-camera: Push probe-time power management to drivers

Several client drivers access the hardware at probe time, for instance
to read the probe chip ID. Such chips need to be powered up when being
probed.

soc-camera handles this by powering chips up in the soc-camera probe
implementation. However, this will break with non soc-camera hosts that
don't perform the same operations.

Fix the problem by pushing the power up/down from the soc-camera core
down to individual drivers on a needs basis.

Signed-off-by: Laurent Pinchart <[email protected]>
Signed-off-by: Guennadi Liakhovetski <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] soc-camera: Add and use soc_camera_power_[on|off]() helper functions

Instead of forcing all soc-camera drivers to go through the mid-layer to
handle power management, create soc_camera_power_[on|off]() functions
that can be called from the subdev .s_power() operation to manage
regulators and platform-specific power handling. This allows non
soc-camera hosts to use soc-camera-aware clients.

Signed-off-by: Laurent Pinchart <[email protected]>
[[email protected]: fix compile breakage]
Signed-off-by: Guennadi Liakhovetski <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] soc-camera: Continue the power off sequence if one of the steps fails

Powering off a device is a "best effort" task: failure to execute one of
the steps should not prevent the next steps to be executed. For
instance, an I2C communication error when putting the chip in stand-by
mode should not prevent the more agressive next step of turning the
chip's power supply off.

Signed-off-by: Laurent Pinchart <[email protected]>
Signed-off-by: Guennadi Liakhovetski <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] soc_camera: Don't call .s_power() during probe

The .s_power() call only covers the .g_mbus_fmt() operation call.
Several clients required to be powered on to retrieve the current mbus
format but have now been fixed. The .s_power() call is thus not needed
anymore and can be removed.

Signed-off-by: Laurent Pinchart <[email protected]>
Signed-off-by: Guennadi Liakhovetski <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] rename drivers/media/video as .../platform

The remaining drivers are mostly platform drivers. Name the
dir to reflect it.

It makes sense to latter break it into a few other dirs.

Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] move soc_camera i2c drivers into its own dir

Move all soc_camera i2c drivers into drivers/media/i2c/soc_camera/.

Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] move i2c files into drivers/media/i2c

Move ancillary I2C drivers into drivers/media/i2c, in order to
better organize them.

Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] reorganize the API core items

Reorganize the API core changes for them to appear closer to
the items that enable them, and not at the drivers part of
the menu.

Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] mmc/Kconfig: Improve driver name for siano mmc/sdio driver

Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] move parallel port/isa video drivers to drivers/media/parport/

We should keep just the I2C drivers under drivers/media/video, and
then rename it to drivers/media/i2c.

Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] move the remaining PCI devices to drivers/media/pci

Move meye and sta2x11_vip into the drivers/media/pci subdirs.

Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] move analog PCI saa7146 drivers to its own dir

Instead of having them under drivers/media/video, move them
to their own directory.

Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] rename most media/video pci drivers to media/pci

Rename all PCI drivers with their own directory under
drivers/media/video into drivers/media/pci and update the
building system.

Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] bt8xx: move analog TV part to be together with DTV one

Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] move the remaining USB drivers to drivers/media/usb

Move the 3 remaining usb drivers to their proper space.

Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] rename most media/video usb drivers to media/usb

Rename all USB drivers with their own directory under
drivers/media/video into drivers/media/usb and update the
building system.

Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] b2c2: fix driver's build due to the lack of pci DMA code

Fix a one-character typo at the makefile.

drivers/built-in.o: In function `flexcop_pci_remove':
flexcop-pci.c:(.text+0x20d1a28): undefined reference to `flexcop_dma_free'
flexcop-pci.c:(.text+0x20d1a3e): undefined reference to `flexcop_dma_free'
drivers/built-in.o: In function `flexcop_pci_stream_control':
flexcop-pci.c:(.text+0x20d2396): undefined reference to `flexcop_dma_config'
flexcop-pci.c:(.text+0x20d23b3): undefined reference to `flexcop_dma_config'
flexcop-pci.c:(.text+0x20d23cf): undefined reference to `flexcop_dma_config_timer'
flexcop-pci.c:(.text+0x20d23f5): undefined reference to `flexcop_dma_xfer_control'
flexcop-pci.c:(.text+0x20d244c): undefined reference to `flexcop_dma_control_timer_irq'
flexcop-pci.c:(.text+0x20d24b0): undefined reference to `flexcop_dma_control_timer_irq'
flexcop-pci.c:(.text+0x20d2507): undefined reference to `flexcop_dma_xfer_control'
drivers/built-in.o: In function `flexcop_pci_probe':
flexcop-pci.c:(.text+0x20d28d2): undefined reference to `flexcop_dma_allocate'
flexcop-pci.c:(.text+0x20d2907): undefined reference to `flexcop_dma_allocate'
flexcop-pci.c:(.text+0x20d2923): undefined reference to `flexcop_dma_free'

Reported-by: Fengguang Wu <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] media: davinci: fix section mismatch warnings

This patch fixes section mismatch warnings for
davinci video drivers.

[[email protected]: applied it as a fixup over the previous (and wrong) changeset c6afbf2]
Signed-off-by: Lad, Prabhakar <[email protected]>
Signed-off-by: Manjunath Hadli <[email protected]>
Acked-by: Laurent Pinchart <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

ioctl-number.txt: Remove legacy private ioctl's from media drivers

None of those drivers use private ioctl's, as they all got converted
to the standard V4L2 ones.

Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] b2c2: frontends/tuners are not needed at the bridge binding

The frontends/tuners are used inside the common part of the driver.

Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] siano: break it into common, mmc and usb

siano is, in fact, 2 drivers: one for MMC and one for USB, plus
a common bus-independent code. Break it accordingly.

Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] saa7146: Move it to its own directory

In order to better organize the directory tree, move the
saa7146 common driver to its own directory.

Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] common: move media/common/tuners to media/tuners

Move the tuners one level up, as the "common" directory will be used
by drivers that are shared between more than one driver.

Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] b2c2: break it into common/pci/usb directories

b2c2 is, in fact, 2 drivers: one for PCI and one for USB, plus
a common bus-independent code. Break it accordingly.

Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] Rename media/dvb as media/pci

The remaining dvb drivers are pci, so rename them to match the
bus.

Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] dvb-usb: move it to drivers/media/usb/dvb-usb

As media/dvb will be removed, move it to a proper place.

Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] firewire: move it one level up

Move firewire to one level up, as the dvb subdirectory will be
removed.

Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] move the dvb/frontends to drivers/media/dvb-frontends

Raise the DVB frontends one level up, as the intention is to remove
the drivers/media/dvb directory.

Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] dvb: move the dvb core one level up

just like the V4L2 core, move the DVB core to drivers/media, as the
intention is to get rid of both "video" and "dvb" directories.

Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] v4l: move v4l2 core into a separate directory

Currently, the v4l2 core is mixed together with other non-core drivers.
Move them into a separate directory.

Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] rc/Kconfig: Fix a warning

drivers/media/rc/Kconfig:291:warning: multi-line strings not supported

Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] it913x ver 1.32 driver moved to dvb-usb-v2

Functional changes

PID filter is default to off and controlled from dvb-usb-v2

Driver now supports suspend and resume changes in dvb-usb-v2

USB bus repeater functions have been removed.

Signed-off-by: Malcolm Priestley <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] Schedule removal of i.MX25 support in mx2_camera.c

Support for i.MX25 in mx2_camera.c has been broken
for a year. Furthermore, i.MX25 video capture HW
doesn't have much in common with i.MX27. A separate
driver would be desirable.

Signed-off-by: Javier Martin <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] media: mx2_camera: Mark i.MX25 support as BROKEN

i.MX25 support is known to have been broken for almost a year.

Signed-off-by: Javier Martin <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] s5k6aa: Use devm_kzalloc function

devm_kzalloc() eliminates the need to free explicitly thereby
making the code a bit simpler.

Signed-off-by: Sachin Kamat <[email protected]>
Acked-by: Sylwester Nawrocki <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] saa7134-dvb: Fix kworld sbtvd I2C gate control

The correct place to put i2c_gate_ctrl is before calling tda18271_attach,
because the driver tda18271 will use it to enable or disable the i2c-bus
from the demodulator to the tuner.

And thus eliminate the error message: "Unknown device (255) detected
@ 1-00c0, device not supported" in the driver tda18271.

In the device kworld_sbtvd (hybrid analog and digital TV) the control
of the i2c-bus to tuner is done in the analog demodulator and not in
the digital demodulator.

Signed-off-by: Manoel Pinheiro <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] rc-core: move timeout and checks to lirc

The lirc TX functionality expects the process which writes (TX) data to
the lirc dev to sleep until the actual data has been transmitted by the
hardware.

Since the same timeout calculation is duplicated in more than one driver
(and would have to be duplicated in even more drivers as they gain TX
support), it makes sense to move this timeout calculation to the lirc
layer instead.

At the same time, centralize some of the sanity checks.

Signed-off-by: David Härdeman <[email protected]>
Cc: Jarod Wilson <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] lmedm04: fix data usage past the end of the buffer

On Mon, 2012-08-13 at 19:58 +0300, Dan Carpenter wrote:
> Hello Mauro Carvalho Chehab,
>
> The patch db6651a9ebb3: "[media] lmedm04: fix build" from Aug 12,
> 2012, leads to the following warning:
> drivers/media/dvb/dvb-usb-v2/lmedm04.c:769 lme2510_download_firmware()
>   error: usb_control_msg() 'data' too small (128 vs 265)
>
>    737          data = kzalloc(128, GFP_KERNEL);
>                                ^^^
> data is 128 bytes.

Control isn't used, so remove it.

Reported-by: Dan Carpenter <[email protected]>
Signed-off-by: Malcolm Priestley <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] DVB: improve handling of TS packets containing a raised TEI bit

When the TEI bit is raised, we should not trust any of the contents of
the packet in question, including but not limited to its PID number.

Considering that we don't trust the PID number of this packet, we should
not proceed to check the packet counter (if dvb_demux_tscheck is set).

We should expect to see at least one discontinuity after a bad packet is
received, so any time a TEI is detected, a following TS packet counter
mismatch is to be expected.

There is no real reason to ever allow bad packets to pass through the
kernel demux, other than for purposes of attempting error correction via
software or statistical information.

However, since we have always passed these bad packets though the demux,
we should not change this default behavior.

Without altering module options, this patch merely prevents the
TS packet counter check on packets containing a raised TEI.

If module option dvb_demux_feed_err_pkts is set to 0, the kernel demux
will drop these error packets entirely, preventing any possibility of
corruption caused by userspace programs that are expecting valid data.

Signed-off-by: Michael Krufky <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] cx88: Remove duplicate const

This patch fixes the following sparse warnings
by removing use of duplicate const.

drivers/media/video/cx88/cx88.h:152:40:
warning: duplicate const
drivers/media/video/cx88/cx88-core.c:256:33:
warning: duplicate const
drivers/media/video/cx88/cx88-alsa.c:769:41:
warning: duplicate const

As commented by Jonathan Nieder:

These double "const" were introduced in v2.6.37-rc1~64^2~464 (V4L/DVB:
drivers/media: Make static data tables and strings const, 2010-08-25).

The address of an array is already immutable by definition, so for
what it's worth, with or without a clarified commit message,

[[email protected]: Tested with "gcc -s" (version 4.7.0):
 the produced asm is indeed equal before and after this patch]
Signed-off-by: Emil Goode <[email protected]>
Acked-by: Jonathan Nieder <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] frontend.h, Docbook: Improve status documentation

No functional changes. It just improves the description of the frontend
status, using Documentation/kernel-doc-nano-HOWTO.txt for the status
enumeration, and a table inside the DocBook.

Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] ARM: mach-omap2: board-rx51-peripherals: Add lirc-rx51 data

The IR diode on the RX51 is connected to the GPT9. This data is needed
for the IR driver to function.

Signed-off-by: Timo Kokkonen <[email protected]>
Cc: Tony Lindgren <[email protected]>
Cc: [email protected]
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] media: rc: Introduce RX51 IR transmitter driver

This is the driver for the IR transmitter diode found on the Nokia
N900 (also known as RX51) device. The driver is mostly the same as
found in the original 2.6.28 based kernel that comes with the device.

The following modifications have been made compared to the original
driver version:

- Adopt to the changes that has happen in the kernel during the past
  five years, such as the change in the include paths

- The OMAP DM-timers require much more care nowadays. The timers need
  to be enabled and disabled or otherwise many actions fail. Timers
  must not be freed without first stopping them or otherwise the timer
  cannot be requested again.

The code has been tested with sending IR codes with N900 device
running Debian userland. The device receiving the codes was Anysee
DVB-C USB receiver.

Signed-off-by: Timo Kokkonen <[email protected]>
Cc: Tony Lindgren <[email protected]>
Cc: [email protected]
Cc: Sakari Ailus <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] dvb_usb_v2: use %*ph to dump usb xfer debugs

Cc: Andy Shevchenko <[email protected]>
Signed-off-by: Antti Palosaari <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] dvb-usb: use %*ph to dump small buffers

[[email protected]: fix trivial merge conflict]
Signed-off-by: Andy Shevchenko <[email protected]>
Signed-off-by: Antti Palosaari <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] saa7127: use %*ph to print small buffers

Signed-off-by: Andy Shevchenko <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] ati_remote: use %*ph to dump small buffers

Signed-off-by: Andy Shevchenko <[email protected]>
Cc: Anssi Hannula <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] dvb: use %*ph to hexdump small buffers

Signed-off-by: Andy Shevchenko <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] gspca: use %*ph to print small buffers

Signed-off-by: Andy Shevchenko <[email protected]>
Cc: Hans de Goede <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] radio-shark2: use %*ph to print small buffers

Signed-off-by: Andy Shevchenko <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] dvb: frontends: use %*ph to dump small buffers

Signed-off-by: Andy Shevchenko <[email protected]>
Acked-by: Antti Palosaari <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] common: tunners: use %*ph to dump small buffers

Signed-off-by: Andy Shevchenko <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] dvb: get rid of fe_ioctl_override callback

This callback were meant to allow overriding a FE callback, before its
call, but it is not really needed, as the callback can be intercepted
after tuner attachment.

Worse than that, only DVBv3 calls are intercepted this way, so a DVBv5
application will produce different effects than a DVBv3 one.

So, get rid of it.

Signed-off-by: Mauro Carvalho Chehab <[email protected]>
Acked-by: Antti Palosaari <[email protected]>
Reviewed-by: Antti Palosaari <[email protected]>

[media] staging: lirc: use %*ph to print small buffers

Signed-off-by: Andy Shevchenko <[email protected]>
Cc: Mauro Carvalho Chehab <[email protected]>
Cc: [email protected]
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] drivers/media/radio/radio-si4713.c: use devm_ functions

The various devm_ functions allocate memory that is released when a driver
detaches.  This patch uses these functions for data that is allocated in
the probe function of a platform device and is only freed in the remove
function.

Signed-off-by: Julia Lawall <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>

[media] drivers/media/radio/radio-wl1273.c: use devm_ functions

The various devm_ functions allocate memory that is released when a driver
detaches.  This patch uses these functions for data that is allocated in
the probe function of a platform device and is only freed in the remove
function.

In two cases, the original memory allocation function was kmalloc, which
has been changed to a zeroing allocation to benefit from the devm function.

Signed-off-by: Julia Lawall <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>