As reported by sparse:

	drivers/media/test-drivers/vidtv/vidtv_ts.h:47:47: warning: array of flexible structures
	drivers/media/test-drivers/vidtv/vidtv_channel.c:458:54: warning: incorrect type in argument 3 (different base types)
	drivers/media/test-drivers/vidtv/vidtv_channel.c:458:54:    expected unsigned short [usertype] service_id
	drivers/media/test-drivers/vidtv/vidtv_channel.c:458:54:    got restricted __be16 [usertype] service_id
	drivers/media/test-drivers/vidtv/vidtv_s302m.c:471 vidtv_s302m_encoder_init() warn: possible memory leak of 'e'

Address such warnings.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Some functions has a different name between their prototypes
and the corresponding kernel-doc markups.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Currently the call to vidtv_psi_pat_table_destroy frees the object
m->si.pat however m->si.pat->num_pmt is being accessed after the
free.  Fix this by destroying m->si.pat after the m->si.pmt_secs[]
objects have been freed.

Addresses-Coverity: ("Read from pointer after free")

Reported-by: Cengiz Can <cengiz@kernel.wtf> # sent a similar fix about the same time

Fixes: 039b7cae ("media: vidtv: add a PID entry for the NIT table")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

We need to set ->need_cache_sync_on_prepare and
->need_cache_sync_on_finish when we initialize vb2 buffer.

Currently these flags are set/adjusted only in V4L2's
vb2_queue_or_prepare_buf(), which means that for the code
paths that don't use V4L2 vb2 will always tell videobuf2
core to skip ->prepare() and ->finish() cache syncs/flushes.
Fix this by setting cache sync hints for new buffers; except
VB2_MEMORY_DMABUF buffers, for which DMA exporter syncs
caches.

Fixes: f5f5fa73 ("media: videobuf2: handle V4L2 buffer cache flags")
Reported-by: Tomasz Figa <tfiga@chromium.org>
Signed-off-by: Sergey Senozhatsky <senozhatsky@chromium.org>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Starting with firmware version 10 the GET/SET_HDMI_VERSION message
was removed and GET/SET_AUTO_POWER_ON was added.

The removal of GET/SET_HDMI_VERSION caused the probe of the
Pulse-Eight to fail. Add a version check to handle this gracefully.

Also show (but do not set) the Auto Power On value.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Reported-by: Maxime Ripard <mripard@kernel.org>
Tested-by: Maxime Ripard <mripard@kernel.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Commit 601282d6 ("media: pulse8-cec: use adap_free callback") used
the adap_free callback to clean up on disconnect. What I forgot was that
in the probe it will call cec_delete_adapter() followed by kfree(pulse8)
if an error occurs. But by using the adap_free callback,
cec_delete_adapter() is already freeing the pulse8 struct.

This wasn't noticed since normally the probe works fine, but Pulse-Eight
published a new firmware version that caused a probe error, so now it
hits this bug. This affects firmware version 12, but probably any
version >= 10.

Commit aa9eda76 ("media: pulse8-cec: close serio in disconnect, not
adap_free") made this worse by adding the line 'pulse8->serio = NULL'
right after the call to cec_unregister_adapter in the disconnect()
function. Unfortunately, cec_unregister_adapter will typically call
cec_delete_adapter (unless a filehandle to the cec device is still
open), which frees the pulse8 struct. So now it will also crash on a
simple unplug of the Pulse-Eight device.

With this fix both the unplug issue and a probe() error situation are
handled correctly again.

It will still fail to probe() with a v12 firmware, that's something
to look at separately.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Reported-by: Maxime Ripard <mripard@kernel.org>
Tested-by: Maxime Ripard <mripard@kernel.org>
Fixes: aa9eda76 ("media: pulse8-cec: close serio in disconnect, not adap_free")
Fixes: 601282d6 ("media: pulse8-cec: use adap_free callback")
Cc: <stable@vger.kernel.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Since commit 528222d8 ("media: rc: harmonize infrared durations to
microseconds"), the calculation of the chk period is wrong. As a result,
all reported IR will have incorrect timings.

Now that the calculations are done in microseconds rather than nanoseconds,
we can fold the calculations in a simpler form with less rounding error.
Tested-by: Frank Wunderlich <frank-w@public-files.de>
Fixes: 528222d8 ("media: rc: harmonize infrared durations to microseconds")
Signed-off-by: Sean Young <sean@mess.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Fix existing issues at the kernel-doc markups and add them to
the vidtv.rst file.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Update the vidtv documentation with the relevant changes
after the last patches.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

- pass struct vidtv_psi_eit_write_args as a pointer;
- avoid initializing struct fields multiple times.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

- pass struct vidtv_psi_nit_write_args as a pointer;
- avoid initializing struct fields multiple times.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

- pass struct vidtv_psi_sdt_write_args as a pointer;
- avoid initializing struct fields multiple times.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

- Pass struct vidtv_psi_pmt_write_args as a pointer;
- Avoid initializing structs multiple times.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Avoid initializing the structs multiple times and pass the
PAT struct as a pointer, instead of a var.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

- Pass struct header_write_args as a pointer, instead of
  passing as a var;

- Initialize the psi_args struct only once.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

This function initializes the psi_args twice, and receives
a struct, instead of a pointer to a struct.

Clean it up.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Cleanup the table_section_crc32_write_into() function
by initializing struct psi_write_args only once and by
passing the args as a pointer.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

The function vidtv_psi_ts_psi_write_into() initializes the
ts_header fields several times, and receives a struct
as argument, instead of using a pointer to struct.

Cleanup the function, in order to reduce its stack usage
and to avoid initializing the ts_header multiple times.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

The current event is using an undefined date. Instead, it
should be the timestamp when the EIT table was generated.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

The service_id there should be equal to the one used
on other tables, otherwise, EIT entries won't be valid.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

As the service currently broadcasts just audio, change the
service type to reflect that.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

On normal TS streams, the NIT table has its own entry at PAT,
but not at PMT.

While here, properly handle alloc problems when creating
PMT entries.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

The EIT header ID field should not contain the network ID, but,
instead, the service_id of the program described at EIT.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

As defined at ETSI TS 101 162, original network IDs up to 0xfebf
are reserved for registration at dvb.org.

Let's use, instead, an original network ID at the range
0xff00-0xffff, as this is for private temporary usage.

As the same value is also used for the network ID,
the range 0xff01-0xffff also fits better, as values
lower than that depend if the network is used for
satellite, terrestrial, cable of CI.

While here, move the TS ID to the bridge code, where it
is used, and change its value, as it was identical to
the value previously used by network ID. While we could
keep the same value, let's change it, just to make easier
to check for the new code while reading it with DVB tools
like dvbinspector.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Place some text at EIT data, and use ISO 8859-15 encoding for
the German letter "ü" (u mit umlat) letter.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Initialize the destination buffer/size and the initial
offset when creating the local var.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Instead of first zeroing all fields at the mux structs and
then filling, do some initialization for the const data
when they're created.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Right now, there's no need to access the length of some
tables. So, drop the unused functions.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Do some cleanups at the coding style of the driver:
- remove "inline" declarations;
- use reverse xmas-tree for local var declarations;
- Adjust some indent to avoid breaking 80-cols;
- Cleanup some comments.

No functional changes.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Minimize the number of data copies and initialization at
the code, passing them as pointers instead of duplicating
the data.

The only case where we're keeping the data copy is at
vidtv_pes_write_h(), as it needs a copy of the passed
arguments. On such case, we're being more explicit.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Initialize the fields of the arguments directly when
declaring it, and pass the args as a pointer, instead of
copying them.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

The sheet music used to generate the tones had a few
polyphonic notes. Due to that, its conversion to a
tones sequence had a few errors.

Also, due to a bug at the tone generator, it was missing
the pause at the initial compass.

Fix them.

While here, reduce the compass to 100bpm.

The music was converted from a Music XML file using
this small script:

<snip>
my $count = 0;
my $silent = 0;
my $note;
my $octave;
print "\t";
while (<>) {
	$note = $1 if (m,\<step\>(.*)\</step\>,);
	$octave = "_$1" if (m,\<octave\>(.*)\</octave\>,);
	if (m,\<alter\>1\</alter\>,) {
		$note .= "S";
		$sharp = 1;
	}
	if (m,\<rest/\>,) {
		$note = "SILENT";
		$silent = 1;
	}
	if (m,\<duration\>(.*)\</duration\>,) {
		printf "{ NOTE_${note}${octave}, %d},", $1 * 128 / 480;
		$count++;
		if ($silent || $count >= 3) {
			print "\n\t";
			$count = 0;
			$silent = 0;
		} else {
			print " ";
			print " " if (!$sharp);
		}
		$sharp = 0;
		$note = "";
		$octave = "";
	};
};
print "\n";
</snip>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

The tone generator logic were repeating the song after the
first silent. There's also a wrong logic at the note
offset calculus, which may create some noise.

Fix it.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

While the original plan was to use the first movement of
the 5th Symphony, it was opted to use the Für Elise song,
instead.

Fix it.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

The Linux stack is too short. So, using recursive functions
is a very bad idea. Convert those into non-recursive ones.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Currently, there are not checks if something gets bad during
memory allocation: it will simply use NULL pointers and
crash.

Add error path at the logic which allocates memory for the
MPEG-TS generator code, propagating the errors up to the
vidtv_bridge. Now, if something wents bad, start_streaming
will return an error that userspace can detect:

	ERROR    DMX_SET_PES_FILTER failed (PID = 0x2000): 12 Cannot allocate memory

and the driver doesn't crash.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

- Place the includes on alphabetical order;
- get rid of asm/byteorder.h;
- add bug.h at vidtv_s302m.c, as it is needed by
  inux/fixp-arith.h
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Some variables were only assigned once but were used in while
loops as if they had been updated at every iteration. Fix this.
Signed-off-by: Daniel W. S. Almeida <dwlsalmeida@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

A few fields used only by the tone generator in the s302m encoder
are stored in struct vidtv_encoder. Move them into
struct vidtv_s302m_ctx instead. While we are at it: fix a
checkpatch warning for long lines.
Signed-off-by: Daniel W. S. Almeida <dwlsalmeida@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

The code to append a descriptor to the end of a chain is repeated
throughout the psi generator code. Extract it into its own helper
function to avoid cluttering.
Signed-off-by: Daniel W. S. Almeida <dwlsalmeida@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>