My 'allmodconfig' build is _almost_ free of warnings, and most of the
remaining ones are for legacy drivers that just do bad things that I
can't find it in my black heart to care too much about.  But this one
was just annoying me:

   drivers/media/v4l2-core/videobuf2-core.c:3256:26: warning: unused variable ‘fileio’ [-Wunused-variable]

because commit 0e661006 ("[media] vb2: fix 'UNBALANCED' warnings
when calling vb2_thread_stop()") removed all users of 'fileio' and
instead calls "__vb2_cleanup_fileio(q)" to clean up q->fileio.  But the
now unused 'fileio' variable was left around.
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

The first time b is used, it is assigned to the first element of the rdy_queue
list. There is no need to set it to NULL before.
Signed-off-by: NPhilipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: NKamil Debski <k.debski@samsung.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@osg.samsung.com>

The vb2: fix bytesused == 0 handling (8a75ffb8) patch changed the behavior
of __fill_vb2_buffer function, so that if bytesused is 0 it is set to the
size of the buffer. However, bytesused set to 0 is used by older codec
drivers as as indication used to mark the end of stream.

To keep backward compatibility, this patch adds a flag passed to the
vb2_queue_init function - allow_zero_bytesused. If the flag is set upon
initialization of the queue, the videobuf2 keeps the value of bytesused
intact in the OUTPUT queue and passes it to the driver.
Reported-by: NNicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: NKamil Debski <k.debski@samsung.com>
Acked-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@osg.samsung.com>

This patch splits the io_flags member of vb2_queue into a bit field.
Instead of an enum with flags separate bit fields were introduced.
Signed-off-by: NKamil Debski <k.debski@samsung.com>
Acked-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@osg.samsung.com>

Please revert this patch: vm_area_cachep is not exported, so you cannot use
this in a module.

This reverts commit 5ed1c328.
Reported-by: NHans Verkuil <hverkuil@cisco.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@osg.samsung.com>

The slab for vm_area_struct which is vm_area_cachep is already prepared
for the general use. Instead of kmalloc() for the vma copy for userptr,
allocation from vm_area_cachep is more beneficial.

CC: Hans Verkuil <hans.verkuil@cisco.com>
CC: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: NCho KyongHo <pullip.cho@samsung.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@osg.samsung.com>

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

This issue was detected by using the Coccinelle software.
Signed-off-by: NMarkus Elfring <elfring@users.sourceforge.net>
Signed-off-by: NMauro Carvalho Chehab <mchehab@osg.samsung.com>

Add support for the new flag to v4l2_print_dv_timings().
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@osg.samsung.com>

Any control with V4L2_CTRL_FLAG_EXECUTE_ON_WRITE set should return
changed == true in cluster_changed.

This forces the value to be passed to the driver even if it has not
changed.
Signed-off-by: NRicardo Ribalda Delgado <ricardo.ribalda@gmail.com>
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@osg.samsung.com>

Any control that sets FLAG_WRITE_ONLY should OR it with
FLAG_EXECUTE_ON_WRITE.

So we can keep the current meaning of WRITE_ONLY.
Signed-off-by: NRicardo Ribalda Delgado <ricardo.ribalda@gmail.com>
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@osg.samsung.com>

Volatile controls should not generate CH_VALUE events.

Set has_changed to false to prevent this happening.
Signed-off-by: NRicardo Ribalda Delgado <ricardo.ribalda@gmail.com>
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@osg.samsung.com>

The function fills a link data structure with the device node and port
number at both the local and remote ends of a link defined by one of its
endpoint nodes.
Signed-off-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: NHans Verkuil <hans.verkuil@cisco.com>
Acked-by: NSakari Ailus <sakari.ailus@linux.intel.com>
Acked-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@osg.samsung.com>

V4L2 clocks, e.g. used by camera sensors for their master clock, do not
have to be supplied by a different V4L2 driver, they can also be
supplied by an independent source. In this case the standart kernel
clock API should be used to handle such clocks. This patch adds support
for such cases.
Signed-off-by: NGuennadi Liakhovetski <g.liakhovetski@gmx.de>
Acked-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Tested-by: NJosh Wu <josh.wu@atmel.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@osg.samsung.com>

All uses of the v4l2_clk API so far only register one clock with a fixed
name. This allows us to get rid of it, which also will make CCF and DT
integration easier.
Signed-off-by: NGuennadi Liakhovetski <g.liakhovetski@gmx.de>
Acked-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Tested-by: NJosh Wu <josh.wu@atmel.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@osg.samsung.com>

The check_ext_ctrls() function in v4l2-ioctl.c checks if all controls in the
control array are from the same control class as c->ctrl_class. However,
that check should only be done if c->ctrl_class != 0. A 0 value means
that this restriction does not apply.

So return 1 (OK) if c->ctrl_class == 0.

Found by running v4l2-compliance on the uvc driver.
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Acked-by: NSakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@osg.samsung.com>

The selection/cropping ioctls are only valid for video nodes, not for vbi.

Found by v4l2-compliance when run on a VBI device node.
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Acked-by: NSakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@osg.samsung.com>

Make sure the format fields are reported consistently.
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Acked-by: NSakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@osg.samsung.com>

Return -EINVAL if read() or write() is not supported by the queue. This
makes it possible to provide both vb2_fop_read and vb2_fop_write in a
struct v4l2_file_operations since the vb2_fop_* function will check if
the file operation is allowed.

A similar check exists in __vb2_init_fileio() which is called from
__vb2_perform_fileio(), but that check is only done if no file I/O is
active. So the sequence of read() followed by write() would be allowed,
which is obviously a bug.

In addition, vb2_fop_write/read should always return -EINVAL if the
operation is not allowed, and by putting the check in the lower levels
of the code it is possible that other error codes are returned (EBUSY
or ERESTARTSYS).

All these issues are avoided by just doing a quick explicit check.
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Acked-by: NSakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@osg.samsung.com>

Add lane_polarities field to struct v4l2_of_bus_mipi_csi2 and write the
contents of the lane-polarities property to it. The field tells the polarity
of the physical lanes starting from the first one. Any unused lanes are
ignored, i.e. only the polarity of the used lanes is specified.

Also rework reading the "data-lanes" property a little.
Signed-off-by: NSakari Ailus <sakari.ailus@iki.fi>
Acked-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@osg.samsung.com>

The last argument of vb2_dc_get_user_pages() is of type enum
dma_data_direction, but the caller, vb2_dc_get_userptr() passes a value
which is the result of comparison dma_dir == DMA_FROM_DEVICE. This results
in the write parameter to get_user_pages() being zero in all cases, i.e.
that the caller has no intent to write there.

This was broken by patch "vb2: replace 'write' by 'dma_dir'".
Signed-off-by: NSakari Ailus <sakari.ailus@linux.intel.com>
Cc: stable@vger.kernel.org  # for v3.19
Acked-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@osg.samsung.com>

Return an error if an invalid 'which' valid is passed in.
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Acked-by: NLad, Prabhakar <prabhakar.csengg@gmail.com>
Tested-by: NLad, Prabhakar <prabhakar.csengg@gmail.com>
Acked-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@osg.samsung.com>

If a subdevice pad op is called from a bridge driver, then there is
no v4l2_subdev_fh struct that can be passed to the subdevice. This
made it hard to use such subdevs from a bridge driver.

This patch replaces the v4l2_subdev_fh pointer by a v4l2_subdev_pad_config
pointer in the pad ops. This allows bridge drivers to use the various
try_ pad ops by creating a v4l2_subdev_pad_config struct and passing it
along to the pad op.

The v4l2_subdev_get_try_* macros had to be changed because of this, so
I also took the opportunity to use the full name of the v4l2_subdev_get_try_*
functions in the __V4L2_SUBDEV_MK_GET_TRY macro arguments: if you now do
'git grep v4l2_subdev_get_try_format' you will actually find the header
where it is defined.

One remark regarding the drivers/staging/media/davinci_vpfe patches: the
*_init_formats() functions assumed that fh could be NULL. However, that's
not true for this driver, it's always set. This is almost certainly a copy
and paste from the omap3isp driver. I've updated the code to reflect the
fact that fh is never NULL.
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Acked-by: NLad, Prabhakar <prabhakar.csengg@gmail.com>
Tested-by: NLad, Prabhakar <prabhakar.csengg@gmail.com>
Acked-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: NSakari Ailus <sakari.ailus@linux.intel.com>
Acked-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@osg.samsung.com>

The handling of VIDIOC_G/S_PRIORITY is now entirely done by the V4L2
core, so we can drop the g/s_priority ioctl ops.

We do have to make sure though that when S_PRIORITY is called we check
that the driver used struct v4l2_fh. This check can be removed once all
drivers are converted to that structure.
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@osg.samsung.com>

To keep V4L2 drivers that did not yet convert to unlocked_ioctl happy,
the v4l2 core had a .ioctl file operation that took a V4L2 lock.

The last drivers are now converted to unlocked_ioctl, so all this
old code can now be removed.
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@osg.samsung.com>

Stopping the vb2 thread (as used by several DVB devices) can result
in an 'UNBALANCED' warning such as this:

vb2: counters for queue ffff880407ee9828: UNBALANCED!
vb2:     setup: 1 start_streaming: 1 stop_streaming: 1
vb2:     wait_prepare: 249333 wait_finish: 249334

This is due to a race condition between stopping the thread and
calling vb2_internal_streamoff(). While I have not been able to deduce
the exact mechanism how this race condition can produce this warning,
I can see that the way the stream is stopped is likely to lead to a
race somewhere.

This patch simplifies how this is done by first ensuring that the
thread is completely stopped before cleaning up the vb2 queue. It
does that by setting threadio->stop to true, followed by a call to
vb2_queue_error() which will wake up the thread. The thread sees that
'stop' is true and it will exit.

The call to kthread_stop() waits until the thread has exited, and only
then is the queue cleaned up by calling __vb2_cleanup_fileio().

This is a much cleaner sequence and the warning has now disappeared.
Reported-by: NJurgen Kramer <gtmkramer@xs4all.nl>
Tested-by: NJurgen Kramer <gtmkramer@xs4all.nl>
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Cc: <stable@vger.kernel.org>      # for v3.18 and up
Signed-off-by: NMauro Carvalho Chehab <mchehab@osg.samsung.com>

drivers/media/v4l2-core/tuner-core.c:440:7: error: 'struct v4l2_subdev' has no member named 'entity'
     t->sd.entity.name = t->name;
Reported-by: Nkbuild test robot <fengguang.wu@intel.com>
Acked-by: NLad, Prabhakar <prabhakar.csengg@gmail.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@osg.samsung.com>

Properly initialize tuner core subdev at the media controller.

That requires a new subtype at the media controller API.
Signed-off-by: NMauro Carvalho Chehab <mchehab@osg.samsung.com>

The previous provision for DVB media controller support were to
define an ID (likely meaning the adapter number) for the DVB
devnodes.

This is just plain wrong. Just like V4L, DVB devices (and any other
device node)) are uniquely identified via a (major, minor) tuple.

This is enough to uniquely identify a devnode, no matter what
API it implements.

So, before we go too far, let's mark the old v4l, fb, dvb and alsa
"devnode" info as deprecated, and just call it as "dev".

We can latter add fields specific to each API if needed.

As we don't want to break compilation on already existing apps,
let's just keep the old definitions as-is, adding a note that
those are deprecated at media-entity.h.
Signed-off-by: NMauro Carvalho Chehab <mchehab@osg.samsung.com>

The videobuf_dma_init* and videobuf_dma_map() functions are no longer
used except in videobuf-dma-sg.c itself. Make them static.

These functions were abused in various drivers. All those drivers
have now been fixed, so by no longer exporting these functions
future abuse is now prevented.
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@osg.samsung.com>

The locking scheme inside the vb2 thread is unsafe when stopping the
thread. In particular kthread_stop was called *after* internal data
structures were cleaned up instead of doing that before. In addition,
internal vb2 functions were called after threadio->stop was set to
true and vb2_internal_streamoff was called. This is also not allowed.

All this led to a variety of race conditions and kernel warnings and/or
oopses.

Fixed by moving the kthread_stop call up before the cleanup takes
place, and by checking threadio->stop before calling internal vb2
queuing operations.
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Cc: <stable@vger.kernel.org>      # for v3.16 and up
Signed-off-by: NMauro Carvalho Chehab <mchehab@osg.samsung.com>

If NO_DMA=y:

drivers/built-in.o: In function `vb2_vmalloc_dmabuf_ops_detach':
videobuf2-vmalloc.c:(.text+0x6f11b0): undefined reference to `dma_unmap_sg'
drivers/built-in.o: In function `vb2_vmalloc_dmabuf_ops_map':
videobuf2-vmalloc.c:(.text+0x6f1266): undefined reference to `dma_unmap_sg'
videobuf2-vmalloc.c:(.text+0x6f1282): undefined reference to `dma_map_sg'

As we don't want to make the core VIDEOBUF2_VMALLOC depend on HAS_DMA
(it's v4l2 core code, and selected by a lot of drivers), stub out the
DMA support if HAS_DMA is not set.
Signed-off-by: NGeert Uytterhoeven <geert@linux-m68k.org>
Acked-by: NMarek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@osg.samsung.com>

Fix this warning:

drivers/media/v4l2-core/videobuf2-vmalloc.c:98:28: warning: incorrect type in assignment (different address spaces)
drivers/media/v4l2-core/videobuf2-vmalloc.c:158:28: warning: incorrect type in argument 1 (different address spaces)

The warning is correct, but we have no other choice here to forcibly cast.
At least it is now explicit that such a cast is needed.
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@osg.samsung.com>

The old debug field is renamed to dev_debug to ensure that existing drivers
(including out-of-tree drivers) that try to use the old name will no longer
compile. A comment has also been added that makes it explicit that drivers
shouldn't use this field.

Additional bits have been added to the debug flag to be more fine-grained
when debugging, especially when dealing with streaming ioctls and read,
write and poll. You want to enable those explicitly to prevent flooding
the log when streaming unless you actually want to do that.
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Acked-by: NSakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@osg.samsung.com>

Drop the duplicate get/set_crop pad ops and only use get/set_selection.
It makes no sense to have two duplicate ops in the internal subdev API.
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Acked-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@osg.samsung.com>

1) This is not allowed by the kernel coding style
2) Just configure your editor correctly
3) It's really ugly
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Acked-by: NSakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@osg.samsung.com>

This is easy to forget to do in drivers. While v4l2-compliance will check for it,
not everyone remembers to run it. So warn about it.
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@osg.samsung.com>

The function releases the queue if the file being released is the queue
owner. The check reads the queue->owner field without taking the queue
lock, creating a race condition with functions that set the queue owner,
such as vb2_ioctl_reqbufs() for instance.

Fix this by moving the queue->owner check within the mutex protected
section.
Signed-off-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: NHans Verkuil <hans.verkuil@cisco.com>
Acked-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@osg.samsung.com>

The vb2_fop_poll() implementation tries to be clever on whether it needs
to lock the queue mutex by checking whether polling might start fileio.
The test requires reading the q->num_buffer field, which is racy if we
don't hold the queue mutex in the first place.

Remove the extra cleverness and just lock the mutex.
Signed-off-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@osg.samsung.com>

The vb2_put_vma() function tests whether its argument is NULL and then
returns immediately. Thus the test around the call is not needed.

This issue was detected by using the Coccinelle software.
Signed-off-by: NMarkus Elfring <elfring@users.sourceforge.net>
Acked-by: NMarek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@osg.samsung.com>

Log the new ycbcr_enc and quantization fields. Note that it now
also logs the flags field for the multiplanar buffer type. This was
forgotten when the flags field was added.
Signed-off-by: NHans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: NMauro Carvalho Chehab <mchehab@osg.samsung.com>