[media] media/videobuf2-dma-contig: Save output from dma_map_sg

dma_map_sg returns the number of areas mapped by the hardware,
which could be different than the areas given as an input.
The output must be saved to nent.

Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@gmail.com>
Reviewed-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] media/videobuf2-dma-sg: Fix handling of sg_table structure

When sg_alloc_table_from_pages() does not fail it returns a sg_table
structure with nents and nents_orig initialized to the same value.

dma_map_sg returns the number of areas mapped by the hardware,
which could be different than the areas given as an input.
The output must be saved to nent.

The output of dma_map, should be used to transverse the scatter list.

dma_unmap_sg needs the value passed to dma_map_sg (nents_orig).

sg_free_tables uses also orig_nent.

This patch fix the file to follow this paradigm.

Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@gmail.com>
Reviewed-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] adv7604/cobalt: missing GPIOLIB dependency

The adv7604 driver depends on GPIOLIB, and therefore cobalt depends
on it as well.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] saa7164: fix sparse warning

drivers/media/pci/saa7164/saa7164-i2c.c:45:33: warning: Using plain integer as NULL pointer

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Reviewed-by: Steven Toth <stoth@kernellabs.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] cx24120: fix sparse warning

drivers/media/dvb-frontends/cx24120.c:837:6: warning: symbol 'cx24120_calculate_ber_window' was not declared. Should it be static?

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] cobalt: fix sparse warnings

drivers/media/pci/cobalt/cobalt-v4l2.c:189:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:191:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:192:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:193:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:194:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:195:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:196:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:197:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:198:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:199:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:201:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:202:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:203:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:234:17: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:240:17: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:246:17: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:266:17: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:267:17: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:271:28: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:275:17: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:276:17: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:312:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:313:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:314:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:315:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:317:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:320:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:321:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:321:36: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:324:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:327:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:328:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:328:41: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:329:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:331:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:332:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:334:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:335:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:336:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:362:17: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:367:17: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:368:17: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:420:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:421:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:422:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:423:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:516:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:516:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:518:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:518:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:518:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:525:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:525:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:531:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:531:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:531:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:531:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:531:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:531:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:531:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:546:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:546:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:548:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:549:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:550:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:551:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:552:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:553:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:554:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:555:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:556:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:556:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:556:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:563:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:564:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:564:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:569:9: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:595:16: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-v4l2.c:602:9: warning: dereference of noderef expression

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] cobalt: fix sparse warnings

drivers/media/pci/cobalt/cobalt-i2c.c:130:17: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-i2c.c:147:17: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-i2c.c:151:26: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-i2c.c:156:34: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-i2c.c:206:17: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-i2c.c:210:26: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-i2c.c:215:34: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-i2c.c:225:27: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-i2c.c:335:17: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-i2c.c:336:17: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-i2c.c:337:17: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-i2c.c:348:34: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-i2c.c:352:17: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-i2c.c:353:17: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-i2c.c:356:17: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-i2c.c:357:17: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-i2c.c:359:17: warning: dereference of noderef expression

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] cobalt: fix sparse warnings

drivers/media/pci/cobalt/cobalt-irq.c:62:33: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-irq.c:64:17: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-irq.c:65:23: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-irq.c:72:21: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-irq.c:73:25: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-irq.c:74:25: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-irq.c:82:33: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-irq.c:83:33: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-irq.c:91:25: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-irq.c:94:23: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-irq.c:103:25: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-irq.c:107:17: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-irq.c:109:17: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-irq.c:116:13: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-irq.c:119:17: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-irq.c:120:17: warning: dereference of noderef expression
drivers/media/pci/cobalt/cobalt-irq.c:122:17: warning: dereference of noderef expression

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] cobalt: fix sparse warnings

drivers/media/pci/cobalt/cobalt-flash.c:101:5: warning: symbol 'cobalt_flash_probe' was not declared. Should it be static?
drivers/media/pci/cobalt/cobalt-flash.c:126:6: warning: symbol 'cobalt_flash_remove' was not declared. Should it be static?
drivers/media/pci/cobalt/cobalt-cpld.c:101:6: warning: symbol 'cobalt_cpld_status' was not declared. Should it be static?
drivers/media/pci/cobalt/cobalt-cpld.c:240:6: warning: symbol 'cobalt_cpld_set_freq' was not declared. Should it be static?

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] cobalt: fix sparse warnings

drivers/media/pci/cobalt/cobalt-flash.c:39:36: warning: incorrect type in initializer (different address spaces)
drivers/media/pci/cobalt/cobalt-flash.c:54:36: warning: incorrect type in initializer (different address spaces)
drivers/media/pci/cobalt/cobalt-flash.c:63:36: warning: incorrect type in initializer (different address spaces)
drivers/media/pci/cobalt/cobalt-flash.c:82:36: warning: incorrect type in initializer (different address spaces)
drivers/media/pci/cobalt/cobalt-flash.c:107:19: warning: incorrect type in assignment (different address spaces)

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] e4000: fix compiler warning

drivers/media/tuners/e4000.c:287:3: warning: this decimal constant is unsigned only in ISO C90
   .rangehigh  =  2208000000L,
   ^

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Cc: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] cobalt: fix compiler warnings on 32 bit OSes

Fixes these warnings:

drivers/media/pci/cobalt/cobalt-omnitek.c: In function 'omni_sg_dma_start':
drivers/media/pci/cobalt/cobalt-omnitek.c:112:28: warning: right shift count >= width of type [-Wshift-count-overflow]
  iowrite32((u32)(desc->bus >> 32), DESCRIPTOR(s->dma_channel) + 4);
                            ^
drivers/media/pci/cobalt/cobalt-omnitek.c: In function 'descriptor_list_create':
drivers/media/pci/cobalt/cobalt-omnitek.c:222:28: warning: right shift count >= width of type [-Wshift-count-overflow]
     d->next_h = (u32)(next >> 32);
                            ^
drivers/media/pci/cobalt/cobalt-omnitek.c:268:32: warning: right shift count >= width of type [-Wshift-count-overflow]
    d->next_h = (u32)(desc->bus >> 32);
                                ^
drivers/media/pci/cobalt/cobalt-omnitek.c:275:27: warning: right shift count >= width of type [-Wshift-count-overflow]
    d->next_h = (u32)(next >> 32);
                           ^
drivers/media/pci/cobalt/cobalt-omnitek.c: In function 'descriptor_list_chain':
drivers/media/pci/cobalt/cobalt-omnitek.c:293:31: warning: right shift count >= width of type [-Wshift-count-overflow]
   d->next_h = (u32)(next->bus >> 32);
                               ^
drivers/media/pci/cobalt/cobalt-omnitek.c: In function 'descriptor_list_loopback':
drivers/media/pci/cobalt/cobalt-omnitek.c:332:30: warning: right shift count >= width of type [-Wshift-count-overflow]
  d->next_h = (u32)(desc->bus >> 32);
                              ^

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] cobalt: fix 64-bit division link error

[linuxtv-media:master 1023/1029] ERROR: "__aeabi_uldivmod" [drivers/media/pci/cobalt/cobalt.ko] undefined!

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Reported-by: kbuild test robot <fengguang.wu@intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] cobalt: fix irqs used for the adv7511 transmitter

The interrupt bit assignments use for the adv7511 were off by one.
This means that the current scheme (bit << (4 * stream_index)) can
no longer be used.

Fix this by precalculating and storing the correct masks in the
cobalt_stream struct.

This wasn't noticed before because the adv7511 interrupts are very
rare. But for CEC support these interrupts are essential, so this made
me realize that it wasn't working correctly.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] DocBook: Update DocBook version and fix a few legacy things

The DVB part of the media API documentation has several
legacy things on it:
- Examples that don't work;
- APIs unused and deprecated;
- places mentioning the wrong API version.

Fix them and bump the documentation version, in order to
reflect the cleanup efforts to make it more consistent with
the current status of the API.

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

[media] DocBook: Fix arguments on some ioctl documentation

Due to a cut-and-paste error, the argument is missing or wrong
on 3 ioctl documentation. Fix them.

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

[media] DocBook: cleaup the notes about DTV properties

The notes there are somewhat confusing and assumes that the
reader would have read the DVBv3 way. This is not true anymore,
as the DVBv3 is now on a separate section that is marked as
deprecated.

So, cleanup the notes.

While here, add a note about using libdvbv5, instead of using
the DVBv5 API directly.

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

[media] DocBook: Add an example for using FE_SET_PROPERTY

In order to make it clearer about how to use the DVBv5 calls,
add an example of its usage. That should make it clearer about
what's actually required for the DVBv5 calls to work.

While here, also mentions the libdvbv5 library.

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

[media] DocBook: improve documentation of the properties structs

Rename the tytle of the struct documentation to reflect
the name of the structures, and use links to do cross-ref.

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

[media] DocBook: improve documentation for hierarchy

Format it as a table and links it with the legacy API xml.

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

[media] DocBook: improve documentation for guard interval

Format it as a table and add more details, in special for
DTMB guard intervals.

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

[media] DocBook: improve documentation for FEC fields

Format it as a table and add more details. Also, remove the
duplicated occurrences.

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

[media] DocBook: move fe_bandwidth to the frontend legacy section

fe_bandwidth/fe_bandwidth_t is used only on DVBv3 API. So, move
it to the frontend legacy xml, and convert it into a table.

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

[media] DocBook: improve documentation for OFDM transmission mode

Format it as a table and add more details, in special, for
the DTMB modes.

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

[media] DocBook: improve documentation for DVB spectral inversion

Format it as a table and provide more details.

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

[media] DocBook: cross-reference enum fe_modulation where needed

At frontend legacy API description, there are three places
where fe_modulation_t is defined. Cross-reference it to point
to the right place at the documentation.

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

[media] DocBook/Makefile: improve typedef parser

The typedef parser is wrong and doesn't get some of the
types defined at the DVB API. Improve it, as we want to add
cross-references to those types.

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

[media] DocBook: some fixes at FE_GET_INFO

There are a few issues at FE_GET_INFO documentation:
- name is a string, not a pointer to a string;
- the return text should be after the paragraph.

While here, better to bold that two fields of the structure used
by FE_GET_INFO are actually deprecated.

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

[media] DocBook: Better document enum fe_modulation

Instead of using programlisting, use a table, as this provides
a better view of the structure.

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

[media] DocBook: add a proper description for dvb_frontend_info.fe_type

The fe_type is deprecated at the DVB API. However, it may still
be used by legacy DVBv3 applications. While this works with old
devices, modern devices may support more than one delivery
system.

Add an explanation about that and a point to what should be
used, instead, in order for legacy apps to support newer hardware.

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

[media] DocBook: Provide a high-level description for DVB frontend

Instead of just showing ioctls, let's add an introdutory text
briefly explaining the DVB frontend API.

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

[media] DocBook: fix FE_READ_STATUS argument description

What's written there about the arguments for this ioctl
is bogus: it doesn't return an enum (or a typedef)
for enum fe_status. Instead, it returns a bitmask with the
values defined by enum fe_status.

Also, the size of the integer returned is not 16 bits, but,
instead, sizeof(fe_status_t), e. g. sizeof(enum), with is
arch-dependent.

This should of course be fixed, but this should be done on
a separate patch.

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

[media] DocBook: better organize the function descriptions for frontend

Move the function ioctl definitions to the end of the chapter,
at their importance. That makes the document better organized,
as the DVB frontend system call index will look like:

open()
close()
ioctl FE_GET_INFO — Query DVB frontend capabilities and returns information about the front-end. This call only requires read-only access to the device
ioctl FE_READ_STATUS — Returns status information about the front-end. This call only requires read-only access to the device
ioctl FE_SET_PROPERTY, FE_GET_PROPERTY — FE_SET_PROPERTY sets one or more frontend properties. FE_GET_PROPERTY returns one or more frontend properties.
ioctl FE_DISEQC_RESET_OVERLOAD — Restores the power to the antenna subsystem, if it was powered off due to power overload.
ioctl FE_DISEQC_SEND_MASTER_CMD — Sends a DiSEqC command
ioctl FE_DISEQC_RECV_SLAVE_REPLY — Receives reply from a DiSEqC 2.0 command
ioctl FE_DISEQC_SEND_BURST — Sends a 22KHz tone burst for 2x1 mini DiSEqC satellite selection.
ioctl FE_SET_TONE — Sets/resets the generation of the continuous 22kHz tone.
ioctl FE_SET_VOLTAGE — Allow setting the DC level sent to the antenna subsystem.
ioctl FE_ENABLE_HIGH_LNB_VOLTAGE — Select output DC level between normal LNBf voltages or higher LNBf voltages.
ioctl FE_SET_FRONTEND_TUNE_MODE — Allow setting tuner mode flags to the frontend.

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

[media] DocBook: better document FE_DISEQC_RESET_OVERLOAD

Use the new format for the ioctl documentation and put the
struct dvb_diseqc_slave_reply together with the ioctl.

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

[media] DocBook: better document FE_DISEQC_SEND_MASTER_CMD

Use the new format for the ioctl documentation and put the
struct dvb_diseqc_slave_reply together with the ioctl.

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

[media] DocBook: better document FE_DISEQC_RECV_SLAVE_REPLY

Use the new format for the ioctl documentation and put the
struct dvb_diseqc_slave_reply together with the ioctl.

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

[media] DocBook: better document FE_DISEQC_SEND_BURST ioctl

Use the proper format for FE_DISEQC_SEND_BURST documentation
and  improve the documentation.

Keep the enum fe_sec_mini_cmd description together with
the ioctl, as both are used together.

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

[media] DocBook: better document FE_SET_TONE ioctl

Use the proper format for FE_SET_TONE documentation and
improve the documentation.

Keep the enum fe_sec_tone_mode description together with
the ioctl, as both are used together.

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

[media] DocBook: better document FE_SET_VOLTAGE ioctl

Use the proper format for FE_SET_VOLTAGE documentation and fix
the documentation. The description for the enum is not 100%,
and it is missing the voltage off value.

Also, it is better to keep the enum description together with
the ioctl, as both are used together.

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

[media] DocBook: reformat FE_ENABLE_HIGH_LNB_VOLTAGE ioctl

Use the proper format for FE_ENABLE_HIGH_LNB_VOLTAGE documentation.

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

[media] DocBook: reformat FE_SET_FRONTEND_TUNE_MODE ioctl

Use the proper format for FE_SET_FRONTEND_TUNE_MODE documentation.

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

[media] DocBook: move FE_GET_PROPERTY to its own xml file

That helps the xref logic at the Makefile to point to the
right place. Also, it becomes more organized and easier to
maintain if each ioctl have its own xml file.

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

[media] DocBook: Rename ioctl xml files

for the xml files describing ioctls, use the same nomenclature
as on V4L2: the ioctl name, in lower case, using - instead of _.

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

[media] DocBook: Fix false positive undefined ioctl references

The new code that detects undocumented ioctls hits some false
positives:

This one is not documented, nor it should, as this is
there just to reserve namespace:

Warning: can't find reference for VIDIOC_RESERVED ioctl

But those are already documented together with other ioctls:

Warning: can't find reference for VIDIOC_UNSUBSCRIBE_EVENT ioctl
Warning: can't find reference for FE_GET_PROPERTY ioctl
Warning: can't find reference for VIDIOC_SUBDEV_G_EDID ioctl
Warning: can't find reference for VIDIOC_SUBDEV_S_EDID ioctl
Warning: can't find reference for VIDIOC_SUBDEV_S_DV_TIMINGS ioctl
Warning: can't find reference for VIDIOC_SUBDEV_G_DV_TIMINGS ioctl
Warning: can't find reference for VIDIOC_SUBDEV_QUERY_DV_TIMINGS ioctl

So, we need to just be sure to point to the right documentation.

Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Acked-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] DocBook: remove duplicated ioctl from v4l2-subdev

Those ioctls are already parsed. No need to explicitly add
them.

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

[media] DocBook: Improve xref check for undocumented ioctls

There are several badly documented undocumented ioctls.
Currently, it just generates an empty link. Instead of doing that,
only add references to the ones that exists, and add a warning
for all references that weren't found.

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

[media] DocBook: Add xref links for DTV propeties

Create xref links for all DTV properties and link the frontend.h
to each. Also use them at the DVB frontent API example.

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

[media] DocBook: Improve the description of the properties API

Make the text clearer about what the properties API does.

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

[media] DocBook: Merge FE_SET_PROPERTY/FE_GET_PROPERTY ioctl description

Instead of having two refentries, merge them into just one,
like what's done with other similar ioctls at V4L2 side.

That makes the entry cleaner and will allow to add the associated
structures together with the refentry.

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

[media] DocBook: fix xref to the FE open() function

Instead of going to the V4L2 open(), use the xref to the
proper place at the frontend ioctls that were already
reformatted.

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

[media] DocBook: rewrite FE_GET_PROPERTY/FE_SET_PROPERTY to use the std way

Use the standard way of documenting ioctls for FE_GET_PROPERTY
and FE_SET_PROPERTY.

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

[media] DocBook: move DVB properties to happen earlier at the document

The DVBv5 API uses DVB properties as the main way to set the frontend
and collect statistics. Move the definition to happen earlier, in
order to reflect its importance.

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

[media] DocBook: improve documentation for FE_READ_STATUS

Move the ioctl and enum fe_status to a separate xml file and
put it into a better format.

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

[media] DocBook: move FE_GET_INFO to a separate xml file

Keeping everything altogether makes harder to reorganize the
DocBook. So, move the FE_GET_INFO ioctl and the associated structures
into a separate file.

No functional changes.

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

[media] DocBook: reformat FE_GET_INFO ioctl documentation

The DVB part of the docbook has a completely different format
than the V4L2 part, as it was written as a separate document.

As the V4L2 documentation is on better shape, and its format
allows adding more information, let's use it for FE_GET_INFO
and gradually update the non-legacy DVB ioctls using the new
format.

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

[media] dvb: split enum from typedefs at frontend.h

Using typedefs is already bad enough, but doing it together
with enum declaration is even worse.

Also, it breaks the scripts at DocBook that would be generating
reference pointers for the enums.

Well, we can't get rid of typedef right now, but let's at least
declare it on a separate line, and let the scripts to generate
the cross-reference, as this is needed for the next DocBook
patches.

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

[media] DocBook: move DVBv3 frontend bits to a separate section

Although the recommended usage is the DVBv5 API calls, the
documentation doesn't make it clear about what's the recommended
calls and what's legacy.

So, move the legacy API bits to a separate xml, putting them into
a new section.

Please notice that more changes are needed, since some of the
bits there are cross-referenced elsewhere.

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

[media] DocBook: Improve DVB frontend description

The DVB frontend API got bitrotten. Update it.

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

[media] DocBook: fix emphasis at the DVB documentation

Currently, it is using 'role="tt"', but this is not defined at
the DocBook 4.5 spec. The net result is that no emphasis happens.

So, replace them to bold emphasis.

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

[media] DocBook: add drawing with a typical media device

An illustration of what's considered a typical media device
may help people to better understand the contents of the media
infrastructure API docbook. So, add it.

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

[media] DocBook: add a note about the ALSA API

The Docbook mistakenly makes to believe that all needed APIs for
media devices are there. Add a note there pointing that some
sub-devices are actually be controlled via ALSA API.

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

[media] DocBook: Update DVB supported standards at introduction

The list of standards at the media docbook is incomplete, and it
is mentioning that the DVB-S2 & friends is "currently being updated".
That's wrong, as such update occurred back in 2008.

So, provide a more complete list of supported standards and add
a reference to the actual list.

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

[media] cx24120: Assume ucb registers is a counter

The ucblocks register is probably a counter and not a rate; assume
it is so and change the calculations as required.

Signed-off-by: Jemma Denson <jdenson@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] dvb-core: fix 32-bit overflow during bandwidth calculation

Frontend bandwidth calculation overflows on very high DVB-S/S2
symbol rates. Use mult_frac() macro in order to keep calculation
correct.

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

[media] e4000: Fix rangehigh value

As reported by smatch:

drivers/media/tuners/e4000.c:287:32: warning: constant 2208000000 is so big it is long long
drivers/media/tuners/e4000.c:287:32: warning: decimal constant 2208000000 is between LONG_MAX and ULONG_MAX. For C99 that means long long, C90 compilers are very likely to produce unsigned long (and a warning) here
drivers/media/tuners/e4000.c:287:3: warning: this decimal constant is unsigned only in ISO C90
   .rangehigh  =  2208000000,
   ^

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

[media] e4000: implement V4L2 subdevice tuner and core ops

Implement V4L2 subdevice tuner and core ops. After that this driver
is hybrid driver implementing both V4L2 and DVB ops.

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

[media] e4000: various small changes

* Rename device state from 's' to 'dev'.
* Move single include to driver private header.
* Change error handling type of each function to one I tend use
nowadays.
* Remove dummy register write from init. Even Windows driver does this
multiple times remove it as I have never seen any I2C errors.
* Define I2C client pointer for each function and use it.
* Do not clean tuner ops during driver remove - not needed.
* Disable sysfs device bind / unbind. We are not allowed manually
bind / unbind device from the driver currently.
* Rename some other variables.

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

[media] e4000: revise synthesizer calculation

Update synthesizer calculation to model I prefer nowadays. It is
mostly just renaming some variables to ones I think are most standard.
Also add 'schematic' of synthesizer following my current understanding.

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

[media] rtl2832_sdr: cleanup some set_bit() calls

This code works fine but static checkers complain.  The test_bit()
function takes the bit number and not a mask.  Then the other issue is
that we were using USB_STATE_URB_BUF which is BIT(0) instead of URB_BUF.
Also we were open coding that instead of using the test/clear/set_bit()
functions.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] cobalt: add new driver

The cobalt device is a PCIe card with 4 HDMI inputs (adv7604) and a
connector that can be used to hook up an adv7511 transmitter or an
adv7842 receiver daughterboard.

This device is used within Cisco but is sadly not available outside
of Cisco. Nevertheless it is a very interesting driver that can serve
as an example of how to support HDMI hardware and how to use the popular
adv devices.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] adv7604/adv7842: replace FMT_CHANGED by V4L2_DEVICE_NOTIFY_EVENT

This makes it easier for the bridge driver to just passthrough such
events to the corresponding device node.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] v4l2-subdev: allow subdev to send an event to the v4l2_device notify function

All drivers use custom notifications, in particular when source changes.
The bridge only has to map the subdev that sends it to whatever video node it is connected to.

Signed-off-by: Jean-Michel Hautbois <jean-michel.hautbois@vodalys.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] vb2: allow requeuing buffers while streaming

vb2_buffer_done() already allows STATE_QUEUED, but currently only when not
streaming. It is useful to allow it while streaming as well, as this makes
it possible for drivers to requeue buffers while waiting for a stable
video signal.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] adv7842: Make output format configurable through pad format operations

Replace the dummy video format operations by pad format operations that
configure the output format.

Copied from the adv7604 driver.

Note: while arch/blackfin/mach-bf609/boards/ezkit.c uses adv7842_platform_data
this source has not been updated because it is broken since the very
beginning. It depends on a struct adv7842_output_format that does not
exist.

And besides that gcc has no support for bf609 so nobody can compile it
except by installing a toolchain from ADI.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Cc: Scott Jiang <scott.jiang.linux@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] cx24120: fix minor checkpatch-error

Remove multiple blank likes.

Signed-off-by: Patrick Boettcher <patrick.boettcher@posteo.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] cx24120: Update comment & fix typo

These state vars are used for more than just ber calculation, also
fix typo.

Signed-off-by: Jemma Denson <jdenson@gmail.com>
Signed-off-by: Patrick Boettcher <patrick.boettcher@posteo.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] cx24120: Check for lock before updating BER & UCB

BER & UCB aren't available unless we're locked; don't update dvbv5
stats when not locked and mark these counters as unavailable.

Signed-off-by: Jemma Denson <jdenson@gmail.com>
Signed-off-by: Patrick Boettcher <patrick.boettcher@posteo.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] cx24120: Convert ucblocks to dvbv5 stats

DVBv3 is a legacy API. Drivers should use DVBv5, in order to support
modern applications. So, implement UCB using dvbv5.

Signed-off-by: Jemma Denson <jdenson@gmail.com>
Signed-off-by: Patrick Boettcher <patrick.boettcher@posteo.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] cx24120: Convert read_ber to retrieve from cache

Instead of reading BER again for DVBv3 call, use the value from the cache.

Signed-off-by: Jemma Denson <jdenson@gmail.com>
Signed-off-by: Patrick Boettcher <patrick.boettcher@posteo.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] cx24120: Add in dvbv5 stats for bit error rate

DVBv3 is a legacy API. Drivers should use DVBv5, in order to support
modern applications. So, implement BER using dvbv5.

Signed-off-by: Jemma Denson <jdenson@gmail.com>
Signed-off-by: Patrick Boettcher <patrick.boettcher@posteo.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] cx24120: constify static data

Use const on the static data, as gcc may optimize better the
code. Also, would prevent that some code would override the
data there.

Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Signed-off-by: Patrick Boettcher <patrick.boettcher@posteo.de>

[media] cx24120: declare cx24120_init() as static

drivers/media/dvb-frontends/cx24120.c:1182:5: warning: no previous prototype for 'cx24120_init' [-Wmissing-prototypes]
 int cx24120_init(struct dvb_frontend *fe)
     ^

Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Signed-off-by: Patrick Boettcher <patrick.boettcher@posteo.de>

[media] cx24120: don't initialize a var that won't be used

As reported by smatch:
drivers/media/dvb-frontends/cx24120.c: In function 'cx24120_message_send':
drivers/media/dvb-frontends/cx24120.c:368:6: warning: variable 'ret' set but not used [-Wunused-but-set-variable]
  int ret, ficus;
      ^

The values written by cx24120 are never checked. So, remove the
check here too. That's said, the best would be to do the reverse,
but globally: to properly handle the error codes.

Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Signed-off-by: Patrick Boettcher <patrick.boettcher@posteo.de>

[media] cx24120: Complete modfec_table

Add in all the likely values for setting fec on tuning. Some values
yet to be tested but very likely to be correct.
In the very unlikely event that these are wrong tuning with
FEC_AUTO will still work.

Signed-off-by: Jemma Denson <jdenson@gmail.com>
Signed-off-by: Patrick Boettcher <patrick.boettcher@posteo.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] b2c2: Reset no_base_addr on skystarS2 attach failure

If set, all the other drivers reset no_base_addr on failure to
attach - git commit for this being added to some of them seems to
be eccd15aad72f774b2059f708bc422dbb8493bb30

This driver has been floating around outside the mainline for so
long it hasn't had this fix, so add it in.

Whilst here tidy surrounding code to fix style issues.

Signed-off-by: Jemma Denson <jdenson@gmail.com>
Signed-off-by: Patrick Boettcher <patrick.boettcher@posteo.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] cx24120: More tidying in cx24120_init

Remove reg1 by refactoring for loops
Change ret into an int as it's only used for return values now,
and remove reset_result by using ret instead.

Signed-off-by: Jemma Denson <jdenson@gmail.com>
Signed-off-by: Patrick Boettcher <patrick.boettcher@posteo.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] cx24120: Tidy cx24120_init

Use reg instead of ret for all calls to readreg, remove ret_EA
as it doesn't need to be a separate var anymore.

Signed-off-by: Jemma Denson <jdenson@gmail.com>
Signed-off-by: Patrick Boettcher <patrick.boettcher@posteo.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] cx24120: Remove unneccesary assignments in cx24120_init

Probably a remnant of this driver being reverse engineered,
cx24120_init assigned ret on each call to writereg - they're not
used for anything so remove them to clear up the codebase.

Signed-off-by: Jemma Denson <jdenson@gmail.com>
Signed-off-by: Patrick Boettcher <patrick.boettcher@posteo.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] cx24120: Tidy up calls to dev_dbg

Remove __func__ from calls to dev_dbg as dynamic debug can add in
the function name anyway.
Remove debug call in dvbv3 read_signal_strength as userspace has
this value anyway.
Reword some strings to make them simpler / more obvious.

Signed-off-by: Jemma Denson <jdenson@gmail.com>
Signed-off-by: Patrick Boettcher <patrick.boettcher@posteo.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] cx24120: Move CNR to DVBv5 stats

Better to report CNR via DVBv5 stats, as scale can be used.

Signed-off-by: Jemma Denson <jdenson@gmail.com>
Signed-off-by: Patrick Boettcher <patrick.boettcher@posteo.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] cx24120: Fix disecq_send_burst command

Previous version of this driver had a memset before every call
to cmd. This meant the default value of cmd.arg[1] was zero unless
burst is set. Make sure it remains zero.
Also fe_sec_mini_cmd_t is an enum, so test against it.

Signed-off-by: Jemma Denson <jdenson@gmail.com>
Signed-off-by: Patrick Boettcher <patrick.boettcher@posteo.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] cx24120: More coding style fixes

Signed-off-by: Jemma Denson <jdenson@gmail.com>
Signed-off-by: Patrick Boettcher <patrick.boettcher@posteo.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] cx24120: Improve cooked signal strength value

Algorithm copied from cx24117.c - it seems to work. Showing 95% on
a signal I know to be somewhere around 74dBuV.

[mchehab@osg.samsung.net: merged two patches, to avoid breaking compilation]
Signed-off-by: Jemma Denson <jdenson@gmail.com>
Signed-off-by: Patrick Boettcher <patrick.boettcher@posteo.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] cx24120: Return DVBv3 signal strength from cache

This patch changes DVBv3 signal strength to read from the cache by
moving the signal strength reading routine into get_stat, and
reducing read_signal_strength to just returning the cached value.

Signed-off-by: Jemma Denson <jdenson@gmail.com>
Signed-off-by: Patrick Boettcher <patrick.boettcher@posteo.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] cx24120: Remove additional calls to read_status

get_frontend was making a call to read_status, which is being
called periodically anyway. Have read_status store fe_status in
the state struct instead and use that within get_frontend.

Signed-off-by: Jemma Denson <jdenson@gmail.com>
Signed-off-by: Patrick Boettcher <patrick.boettcher@posteo.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] cx24120: Enable DVBv5 signal strength stats

Previous patch added in collection, this patch enables them by
signalling they are available.

Signed-off-by: Jemma Denson <jdenson@gmail.com>
Signed-off-by: Patrick Boettcher <patrick.boettcher@posteo.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] cx24120: Add DVBv5 signal strength stats

Add new get_stats function, called from read_status, for
collecting DVBv5 stats into the frontend cache. Only signal
strength for now, can add others later.
Not currently marked as available, future patch will enable.

Signed-off-by: Jemma Denson <jdenson@gmail.com>
Signed-off-by: Patrick Boettcher <patrick.boettcher@posteo.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] cx24120: Rework vco function to remove xxyyzz variable

Change calculate_vco function to send_vco to tidy up cx24120_init
function.
xxyyzz variable is remnants of a manual div & remainder codepath
to perform a u64 rounded div; replace with kernel macro
DIV_ROUND_CLOSEST_ULL
Hex values provided to the message are mainly variables calculated
within this function, replace with these to remove hardcoding.

Signed-off-by: Jemma Denson <jdenson@gmail.com>
Signed-off-by: Patrick Boettcher <patrick.boettcher@posteo.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] cx24120: Fix hexdump length in writeregs

msg.len has been ++'d so msg.len is one too many.

Signed-off-by: Jemma Denson <jdenson@gmail.com>
Signed-off-by: Patrick Boettcher <patrick.boettcher@posteo.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

[media] cx24120: Add missing command to cx24120_check_cmd

This patch adds in a missing command to cx24120_check_cmd. This part
of the old reverse engineered driver was quite hard to follow - it's
presumed the compiler did some neat optimisation that wasn't easy to
decode. This command was checked for in that version but not copied
across to the new switch statement.

For reference, old reverse engineered code:
xxzz = cmd->id - 0x11; // look for specific message id
if ( xxzz <= 0x13 ) {
msg_cmd_mask = 1 << xxzz;
//0x0F8021 // if cmd_id 17 or 22 or 33-36, 42, 47, 57-61 etc. disable mpeg output
if ( msg_cmd_mask & 0x0F8021 ) { // 000011111000000000100001b
cx24120_msg_mpeg_output_global_config(state, 0);
msleep(100);
state->dvb_s2_mode &=  0xFD; // reset mpeg out enable flag
}
}

Signed-off-by: Jemma Denson <jdenson@gmail.com>
Signed-off-by: Patrick Boettcher <patrick.boettcher@posteo.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>