firefly-linux-kernel-4.4.55.git
13 years ago[media] OMAP_VOUT: CLEANUP: Move generic functions and macros to common files
archit taneja [Tue, 14 Jun 2011 06:54:45 +0000 (03:54 -0300)]
[media] OMAP_VOUT: CLEANUP: Move generic functions and macros to common files

Move the inline functions rotate_90_or_270(), rotation_enabled(), and
calc_rotation() from omap_vout.c to omap_voutdef.h.

Move the independent functions omap_vout_alloc_buffer() and
omap_vout_free_buffer() to omap_voutlib.c.

Remove extern identifier from function definitions in omap_voutlib.h

Add static identifier to functions that are used locally in omap_vout.c

Signed-off-by: Archit Taneja <archit@ti.com>
Signed-off-by: Vaibhav Hiremath <hvaibhav@ti.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] OMAP2: V4L2: Remove GFP_DMA allocation as ZONE_DMA is not configured on OMAP
Amber Jain [Tue, 31 May 2011 14:45:36 +0000 (11:45 -0300)]
[media] OMAP2: V4L2: Remove GFP_DMA allocation as ZONE_DMA is not configured on OMAP

Remove GFP_DMA from the __get_free_pages() call from omap24xxcam as ZONE_DMA
is not configured on OMAP. Earlier the page allocator used to return a page
from ZONE_NORMAL even when GFP_DMA is passed and CONFIG_ZONE_DMA is disabled.
As a result of commit a197b59ae6e8bee56fcef37ea2482dc08414e2ac, page allocator
returns null in such a scenario with a warning emitted to kernel log.

Signed-off-by: Amber Jain <amber@ti.com>
Acked-by: Sakari Ailus <sakari.ailus@iki.fi>
Signed-off-by: Vaibhav Hiremath <hvaibhav@ti.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] V4L2: omap_vout: Remove GFP_DMA allocation as ZONE_DMA is not configured...
Amber Jain [Tue, 31 May 2011 14:45:35 +0000 (11:45 -0300)]
[media] V4L2: omap_vout: Remove GFP_DMA allocation as ZONE_DMA is not configured on OMAP

Remove GFP_DMA from the __get_free_pages() call from omap_vout as ZONE_DMA
is not configured on OMAP. Earlier the page allocator used to return a page
from ZONE_NORMAL even when GFP_DMA is passed and CONFIG_ZONE_DMA is disabled.
As a result of commit a197b59ae6e8bee56fcef37ea2482dc08414e2ac, page allocator
returns null in such a scenario with a warning emitted to kernel log.

Signed-off-by: Amber Jain <amber@ti.com>
Signed-off-by: Vaibhav Hiremath <hvaibhav@ti.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] uvcvideo: Add FIX_BANDWIDTH quirk to HP Webcam on HP Mini 5103 netbook
Kirill Smelkov [Fri, 22 Jul 2011 14:47:22 +0000 (11:47 -0300)]
[media] uvcvideo: Add FIX_BANDWIDTH quirk to HP Webcam on HP Mini 5103 netbook

The camera there identifies itself as being manufactured by Cheng Uei
Precision Industry Co., Ltd (Foxlink), and product is titled as "HP
Webcam [2 MP Fixed]".

I was trying to get 2 USB video capture devices to work simultaneously,
and noticed that the above mentioned webcam always requires packet size
= 3072 bytes per micro frame (~= 23.4 MB/s isoc bandwidth), which is far
more than enough to get standard NTSC 640x480x2x30 = ~17.6 MB/s isoc
bandwidth.

As there are alt interfaces with smaller MxPS

    T:  Bus=01 Lev=01 Prnt=01 Port=03 Cnt=01 Dev#=  2 Spd=480  MxCh= 0
    D:  Ver= 2.00 Cls=ef(misc ) Sub=02 Prot=01 MxPS=64 #Cfgs=  1
    P:  Vendor=05c8 ProdID=0403 Rev= 1.06
    S:  Manufacturer=Foxlink
    S:  Product=HP Webcam [2 MP Fixed]
    S:  SerialNumber=200909240102
    C:* #Ifs= 2 Cfg#= 1 Atr=80 MxPwr=500mA
    A:  FirstIf#= 0 IfCount= 2 Cls=0e(video) Sub=03 Prot=00
    I:* If#= 0 Alt= 0 #EPs= 1 Cls=0e(video) Sub=01 Prot=00 Driver=uvcvideo
    E:  Ad=83(I) Atr=03(Int.) MxPS=  16 Ivl=4ms
    I:* If#= 1 Alt= 0 #EPs= 0 Cls=0e(video) Sub=02 Prot=00 Driver=uvcvideo
    I:  If#= 1 Alt= 1 #EPs= 1 Cls=0e(video) Sub=02 Prot=00 Driver=uvcvideo
    E:  Ad=81(I) Atr=05(Isoc) MxPS= 128 Ivl=125us
    I:  If#= 1 Alt= 2 #EPs= 1 Cls=0e(video) Sub=02 Prot=00 Driver=uvcvideo
    E:  Ad=81(I) Atr=05(Isoc) MxPS= 512 Ivl=125us
    I:  If#= 1 Alt= 3 #EPs= 1 Cls=0e(video) Sub=02 Prot=00 Driver=uvcvideo
    E:  Ad=81(I) Atr=05(Isoc) MxPS=1024 Ivl=125us
    I:  If#= 1 Alt= 4 #EPs= 1 Cls=0e(video) Sub=02 Prot=00 Driver=uvcvideo
    E:  Ad=81(I) Atr=05(Isoc) MxPS=1536 Ivl=125us
    I:  If#= 1 Alt= 5 #EPs= 1 Cls=0e(video) Sub=02 Prot=00 Driver=uvcvideo
    E:  Ad=81(I) Atr=05(Isoc) MxPS=2048 Ivl=125us
    I:  If#= 1 Alt= 6 #EPs= 1 Cls=0e(video) Sub=02 Prot=00 Driver=uvcvideo
    E:  Ad=81(I) Atr=05(Isoc) MxPS=2688 Ivl=125us
    I:  If#= 1 Alt= 7 #EPs= 1 Cls=0e(video) Sub=02 Prot=00 Driver=uvcvideo
    E:  Ad=81(I) Atr=05(Isoc) MxPS=3072 Ivl=125us

UVC_QUIRK_FIX_BANDWIDTH helps here and NTSC video can be served with
MxPS=2688 i.e. 20.5 MB/s isoc bandwidth.

In terms of microframe time allocation, before the quirk NTSC video
required 60 usecs / microframe and 53 usecs / microframe after.

Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Kirill Smelkov <kirr@mns.spb.ru>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] au8522: set signal field to 100% when signal present
Devin Heitmueller [Sun, 24 Jul 2011 20:23:09 +0000 (17:23 -0300)]
[media] au8522: set signal field to 100% when signal present

The signal state field in G_TUNER is typically scaled from 0-100%.  Since we
don't know the signal level, we really would prefer the field to contain 100%
than 1/256, which in many utilities (such as v4l2-ctl) rounds to 0% even when
a signal is actually present.

This patch makes the behavior consistent with other drivers.

Signed-off-by: Devin Heitmueller <dheitmueller@kernellabs.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] cx231xx: Provide signal lock status in G_INPUT
Devin Heitmueller [Sun, 24 Jul 2011 20:07:07 +0000 (17:07 -0300)]
[media] cx231xx: Provide signal lock status in G_INPUT

Make use of the signal state registers to properly populate the signal lock
registers in the cx231xx driver.

This allows applications to know whether there is a signal present even in
devices which lack a tuner (since such apps typically won't call G_TUNER if
no tuner is present).

[mchehab@redhat.com: Fix CodingStyle: don't use {} for one-line if's]
Signed-off-by: Devin Heitmueller <dheitmueller@kernellabs.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] cx231xx: Fix power ramping issue
Devin Heitmueller [Sun, 24 Jul 2011 18:44:43 +0000 (15:44 -0300)]
[media] cx231xx: Fix power ramping issue

On platforms that have CONFIG_HZ set to 100, the power ramp time effectively
ends up being 10ms.  However, on those that have a higher CONFIG_HZ, the time
ends up *actually* being 5ms, which doesn't allow enough time for the hardware
to be fully powered up before attempting to address it via i2c.

Change the constant to 10ms, which is long enough for the hardware to power
up, and won't really be anymore time than it was previously on platforms
with CONFIG_HZ being 100.

Credit goes to Mauro Carvalho Chehab and Gerd Hoffmann who previously
investigated this issue.

Tested with the Hauppauge USBLive 2, with which the problem was readily
reproducible after setting CONFIG_HZ to 1000.

Signed-off-by: Devin Heitmueller <dheitmueller@kernellabs.com>
Cc: Mauro Carvalho Chehab <mchehab@redhat.com>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] Fix regression introduced which broke the Hauppauge USBLive 2
Devin Heitmueller [Sun, 24 Jul 2011 01:17:17 +0000 (22:17 -0300)]
[media] Fix regression introduced which broke the Hauppauge USBLive 2

The following patch addresses the regression introduced in the cx231xx
driver which stopped the Hauppauge USBLive2 from working.

Confirmed working by both myself and the user who reported the issue
on the KernelLabs blog (Robert DeLuca).

At some point during refactoring of the cx231xx driver, the USBLive 2 device
became broken.  This patch results in the device working again.

Thanks to Robert DeLuca for sponsoring this work.

Signed-off-by: Devin Heitmueller <dheitmueller@kernellabs.com>
Cc: Robert DeLuca <robertdeluca@me.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] uvcvideo: Fix control mapping for devices with multiple chains
Stephan Lachowsky [Tue, 31 May 2011 22:24:21 +0000 (19:24 -0300)]
[media] uvcvideo: Fix control mapping for devices with multiple chains

The search for matching extension units fails to take account of the
current chain.  In the case where you have two distinct video chains,
both containing an XU with the same GUID but different unit ids, you
will be unable to perform a mapping on the second chain because entity
on the first chain will always be found first

Fix this by only searching the current chain when performing a control
mapping.  This is analogous to the search used by uvc_find_control(),
and is the correct behaviour.

Signed-off-by: Stephan Lachowsky <stephan.lachowsky@maxim-ic.com>
Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] bt8xx: use pci_dev->subsystem_{vendor|device}
Sergei Shtylyov [Fri, 22 Jul 2011 18:16:04 +0000 (15:16 -0300)]
[media] bt8xx: use pci_dev->subsystem_{vendor|device}

The driver reads PCI subsystem IDs from the PCI configuration registers while
they are already stored by the PCI subsystem in the 'subsystem_{vendor|device}'
fields of 'struct pci_dev'...

Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] drxk: Fix the logic that selects between DVB-C annex A and C
Mauro Carvalho Chehab [Sun, 24 Jul 2011 12:11:36 +0000 (09:11 -0300)]
[media] drxk: Fix the logic that selects between DVB-C annex A and C

Fix the DRX-K logic that selects between DVB-C annex A and C

Fix a typo where DVB-C annex type is set via setEnvParameters, but
the driver, uses, instead, setParamParameters[2].

While here, cleans up the code, fixing a bad identation at the fallback
code for other types of firmware, and put the multiple-line comments
into the Linux CodingStyle.

Acked-by: Oliver Endriss <o.endriss@gmx.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] drxk: Fix read debug message
Mauro Carvalho Chehab [Fri, 22 Jul 2011 01:30:27 +0000 (22:30 -0300)]
[media] drxk: Fix read debug message

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] drxk: Fix error return code during drxk init
Mauro Carvalho Chehab [Thu, 21 Jul 2011 22:35:04 +0000 (19:35 -0300)]
[media] drxk: Fix error return code during drxk init

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] drxk: Remove goto/break after return
Mauro Carvalho Chehab [Fri, 15 Jul 2011 12:35:58 +0000 (09:35 -0300)]
[media] drxk: Remove goto/break after return

After return, we don't need any other statement to change the
function flux ;)

Reported-by: Oliver Endriss <o.endriss@gmx.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] drxk: Fix a bug at some switches that broke DVB-T
Mauro Carvalho Chehab [Fri, 15 Jul 2011 04:14:17 +0000 (01:14 -0300)]
[media] drxk: Fix a bug at some switches that broke DVB-T

The error propagation changeset c23bf4402 broke the DVB-T
code.

The legacy way for propagate errors was:

do {
status = foo_func()
if (status < 0)
break;
} while (0);
return status;

However, on a few places, it was doing:
do {
switch(foo) {
case bar:
status = foo_func()
if (status < 0)
break;
break;
}
switch(foo2) {
case bar:
status = foo_func()
if (status < 0)
break;
break;
}
...
} while (0);
return (status)

The inner error break were not working, as it were breaking only
the switch, instead of the do. The solution used were to do a
s/break/goto error/ at the inner breaks, but preserving the last
break. Onfortunately, on a few switches, the replacement were
applied also to the final break for the case statements.

Fix the broken logic, by reverting them to break, where pertinent,
in order to fix DVB-T support.

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] Remove the double symbol increment hack from drxk_hard
Mauro Carvalho Chehab [Fri, 15 Jul 2011 01:23:18 +0000 (22:23 -0300)]
[media] Remove the double symbol increment hack from drxk_hard

Both ngene and ddbrige calls dvb_attach once for drxk_attach.
The logic used there, and by tda18271c2dd driver is different
from similar logic on other frontends.

The right fix is to change them to use the same logic, but,
while we don't do that, we need to patch em28xx-dvb in order
to do cope with ngene/ddbridge magic.

While here, document why drxk_t_release should do nothing.

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] em28xx: Add other Terratec H5 USB ID's
Mauro Carvalho Chehab [Fri, 15 Jul 2011 00:48:00 +0000 (21:48 -0300)]
[media] em28xx: Add other Terratec H5 USB ID's

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years agoRevert "[media] DVB: dvb_frontend: off by one in dtv_property_dump()"
Mauro Carvalho Chehab [Fri, 15 Jul 2011 11:49:06 +0000 (08:49 -0300)]
Revert "[media] DVB: dvb_frontend: off by one in dtv_property_dump()"

This reverts commit a3e4adf274f86b2363fedaa964297cb38526cef0.

As pointed by Andread Oberritter <obi@linuxtv.org>:

That's wrong, because the array size is DTV_MAX_COMMAND + 1.
Using the ARRAY_SIZE macro instead might reduce the confusion.

Also, changeset 3995223038 already fixed this issue.

Reported-by: Andread Oberritter <obi@linuxtv.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] cx88: implemented luma notch filter control
istvan_v@mailbox.hu [Mon, 11 Jul 2011 14:07:43 +0000 (11:07 -0300)]
[media] cx88: implemented luma notch filter control

The following patch adds a new control that makes it possible to set the
luma notch filter type to finetune picture quality.

Signed-off-by: Istvan Varga <istvan_v@mailbox.hu>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] cx88: implemented sharpness control
istvan_v@mailbox.hu [Mon, 11 Jul 2011 14:02:19 +0000 (11:02 -0300)]
[media] cx88: implemented sharpness control

This patch implements support for a sharpness control, using the luma
peaking filter feature of cx2388x.

[mchehab@redhat.com: use cx_andor instead of cx_write]
Signed-off-by: Istvan Varga <istvan_v@mailbox.hu>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] v4l: s5p-tv: add TV Mixer driver for Samsung S5P platform
Tomasz Stanislawski [Wed, 2 Feb 2011 08:40:08 +0000 (05:40 -0300)]
[media] v4l: s5p-tv: add TV Mixer driver for Samsung S5P platform

Add driver for TV Mixer on Samsung platforms from S5P family.
Mixer is responsible for merging images from three layers and
passing result to the output.

Drivers are using:
- v4l2 framework
- videobuf2
- runtime PM

Signed-off-by: Tomasz Stanislawski <t.stanislaws@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Reviewed-by: Marek Szyprowski <m.szyprowski@samsung.com>
Reviewed-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Reviewed-by: Hans Verkuil <hverkuil@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] v4l: s5p-tv: add SDO driver for Samsung S5P platform
Tomasz Stanislawski [Wed, 2 Feb 2011 08:40:08 +0000 (05:40 -0300)]
[media] v4l: s5p-tv: add SDO driver for Samsung S5P platform

Add drivers for Standard Definition output (SDO) on Samsung platforms
from S5P family. The driver provides control over streaming analog TV
via Composite connector.

Driver is using:
- v4l2 framework
- runtime PM

Signed-off-by: Tomasz Stanislawski <t.stanislaws@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Reviewed-by: Marek Szyprowski <m.szyprowski@samsung.com>
Reviewed-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Reviewed-by: Hans Verkuil <hverkuil@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] v4l: s5p-tv: add drivers for HDMI on Samsung S5P platform
Tomasz Stanislawski [Wed, 2 Mar 2011 16:16:37 +0000 (13:16 -0300)]
[media] v4l: s5p-tv: add drivers for HDMI on Samsung S5P platform

Add drivers for HDMI outputs on Samsung platforms from S5P family.

Drivers are using:
- v4l2 framework
- runtime PM

Signed-off-by: Tomasz Stanislawski <t.stanislaws@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Reviewed-by: Marek Szyprowski <m.szyprowski@samsung.com>
Reviewed-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Reviewed-by: Hans Verkuil <hverkuil@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] v4l: add g_std_output callback to V4L2 subdev
Tomasz Stanislawski [Wed, 29 Jun 2011 10:16:15 +0000 (07:16 -0300)]
[media] v4l: add g_std_output callback to V4L2 subdev

Callback is used to acquire current analog TV standard from a subdev.
It is used to avoid keeping current standard in top-level driver.

Signed-off-by: Tomasz Stanislawski <t.stanislaws@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] v4l: add g_dv_preset callback to V4L2 subdev
Tomasz Stanislawski [Wed, 29 Jun 2011 10:01:28 +0000 (07:01 -0300)]
[media] v4l: add g_dv_preset callback to V4L2 subdev

Callback is used to acquire current digital video preset from a subdev.
It is used to avoid keeping dv preset in top-level driver.

Signed-off-by: Tomasz Stanislawski <t.stanislaws@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] v4l: add g_tvnorms_output callback to V4L2 subdev
Tomasz Stanislawski [Wed, 2 Mar 2011 16:16:37 +0000 (13:16 -0300)]
[media] v4l: add g_tvnorms_output callback to V4L2 subdev

Callback is used to acquire TV norms supported by a subdev.
It is used to avoid having standards in top-level driver.

Signed-off-by: Tomasz Stanislawski <t.stanislaws@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Reviewed-by: Hans Verkuil <hverkuil@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] cx88: properly maintain decoder config when using MPEG encoder
Devin Heitmueller [Thu, 14 Jul 2011 15:44:46 +0000 (12:44 -0300)]
[media] cx88: properly maintain decoder config when using MPEG encoder

The cx88 driver would force core->input to always be zero when doing the
the request_acquire().  While it wasn't actually changing the input register
in the hardware, the driver makes decision based on the current input.  In
particular, it decides whether to do things like enabling the comb filter
when on a composite input but disabling it on s-video.  So for example, on
the HVR-1300, using the s-video input with the MPEG encoder would end up with
the video decoder core configured as though the input type were composite.

In short, the driver state did not match the hardware state.

This patch does two things:

1.  It forces the input to zero only if actually switching to DVB mode.  This
prevents the input from changing when the blackbird driver opens the device.

2.  Keep track of what the input was set to when switching to DVB, and reset
it back when done.  This eliminates a condition where for example the user
had the analog side of the board set to capture on the s-video input, then
he used DVB for a bit, then the analog input would unexpectedly be set to
the tuner input.

This work was sponsored by Anevia S.A.

Signed-off-by: Devin Heitmueller <dheitmueller@kernellabs.com>
Cc: Florent Audebert <florent.audebert@anevia.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] MFC: Add MFC 5.1 V4L2 driver
Kamil Debski [Tue, 21 Jun 2011 13:51:26 +0000 (10:51 -0300)]
[media] MFC: Add MFC 5.1 V4L2 driver

Multi Format Codec 5.1 is a hardware video coding acceleration
module found in the S5PV210 and Exynos4 Samsung SoCs. It is
capable of handling a range of video codecs and this driver
provides a V4L2 interface for video decoding and encoding.

Signed-off-by: Kamil Debski <k.debski@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Cc: Jeongtae Park <jtp.park@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] redrat3: improve compat with lirc userspace decode
Jarod Wilson [Wed, 13 Jul 2011 21:26:07 +0000 (18:26 -0300)]
[media] redrat3: improve compat with lirc userspace decode

This is admittedly a bit of a hack, but if we change our timeout value
to something longer and fudge our synthesized trailing space sample
based on the initial pulse sample, rc-core decode continues to work just
fine with both rc-6 and rc-5, and now lirc userspace decode shows proper
repeats for both of those protocols as well. Also tested NEC
successfully with both decode options.

We do still need a reset timer callback using the hardware's timeout
value to make sure we actually process samples correctly, regardless of
our somewhat hacky timeout and synthesized trailer above.

This also adds a missing del_timer_sync call to the module unload path.

CC: Chris Dodge <chris@redrat.co.uk>
CC: Andrew Vincer <andrew.vincer@redrat.co.uk>
CC: Stephen Cox <scox_nz@yahoo.com>
Signed-off-by: Jarod Wilson <jarod@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] redrat3: cap duration in the right place
Jarod Wilson [Wed, 13 Jul 2011 21:26:06 +0000 (18:26 -0300)]
[media] redrat3: cap duration in the right place

Trying to cap duration before multiplying it was obviously wrong.

CC: Chris Dodge <chris@redrat.co.uk>
CC: Andrew Vincer <andrew.vincer@redrat.co.uk>
CC: Stephen Cox <scox_nz@yahoo.com>
Signed-off-by: Jarod Wilson <jarod@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] redrat3: sending extra trailing space was useless
Jarod Wilson [Wed, 13 Jul 2011 21:26:05 +0000 (18:26 -0300)]
[media] redrat3: sending extra trailing space was useless

We already add a trailing space, this wasn't doing anything useful, and
actually confused lirc userspace a bit. Rip it out.

CC: Chris Dodge <chris@redrat.co.uk>
CC: Andrew Vincer <andrew.vincer@redrat.co.uk>
CC: Stephen Cox <scox_nz@yahoo.com>
Signed-off-by: Jarod Wilson <jarod@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] rc-core support for Microsoft IR keyboard/mouse
Jarod Wilson [Wed, 13 Jul 2011 21:09:48 +0000 (18:09 -0300)]
[media] rc-core support for Microsoft IR keyboard/mouse

This is a custom IR protocol decoder, for the RC-6-ish protocol used by
the Microsoft Remote Keyboard, apparently developed internally at
Microsoft, and officially dubbed MCIR-2, per their March 2011 remote and
transceiver requirements and specifications document, which also touches
on this IR keyboard/mouse device.

Its a standard keyboard with embedded thumb stick mouse pointer and
mouse buttons, along with a number of media keys. The media keys are
standard RC-6, identical to the signals from the stock MCE remotes, and
will be handled as such. The keyboard and mouse signals will be decoded
and delivered to the system by an input device registered specifically
by this driver.

Successfully tested with multiple mceusb-driven transceivers, as well as
with fintek-cir and redrat3 hardware. Essentially, any raw IR hardware
with enough sampling resolution should be able to use this decoder,
nothing about it is at all receiver-hardware-specific.

This work is inspired by lirc_mod_mce:

The documentation there and code aided in understanding and decoding the
protocol, but the bulk of the code is actually borrowed more from the
existing in-kernel decoders than anything. I did recycle the keyboard
keycode table, a few defines, and some of the keyboard and mouse data
parsing bits from lirc_mod_mce though.

Special thanks to James Meyer for providing the hardware, and being
patient with me as I took forever to get around to writing this.

callback routine to ensure we don't get any stuck keys, and used
symbolic names for the keytable. Also cc'ing Florian this time, who I
believe is the original mod-mce author...

CC: Florian Demski <fdemski@users.sourceforge.net>
Signed-off-by: Jarod Wilson <jarod@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] af9015: add support for Sveon STV22 [1b80:e401]
Emilio David Diaus Lopez [Tue, 12 Jul 2011 22:53:39 +0000 (19:53 -0300)]
[media] af9015: add support for Sveon STV22 [1b80:e401]

Signed-off-by: Emilio David Diaus Lopez <reality_es@yahoo.es>
Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] af9015: remove 2nd I2C-adapter
Antti Palosaari [Sat, 18 Jun 2011 13:24:53 +0000 (10:24 -0300)]
[media] af9015: remove 2nd I2C-adapter

It is useless. There is only one physical I2C-adapter.

2nd adapter was added originally due to some plans for allowing only one
demod to access bus at time. But I never implemented proper locking...

Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] af9015: remove old FW based IR polling code
Antti Palosaari [Sat, 18 Jun 2011 00:21:32 +0000 (21:21 -0300)]
[media] af9015: remove old FW based IR polling code

Remove old code which is not used anymore since IR code is read
directly from memory nowadays.

Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] af9015: add more I2C msg checks
Antti Palosaari [Sat, 18 Jun 2011 00:16:38 +0000 (21:16 -0300)]
[media] af9015: add more I2C msg checks

Return EOPNOTSUPP for too long messages.

Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] af9015: small optimization
Antti Palosaari [Thu, 16 Jun 2011 23:02:41 +0000 (20:02 -0300)]
[media] af9015: small optimization

Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] af9015: map remote for MSI DIGIVOX Duo
Antti Palosaari [Wed, 15 Jun 2011 14:29:47 +0000 (11:29 -0300)]
[media] af9015: map remote for MSI DIGIVOX Duo

Reported-by: Jacek M. Holeczek <jacek.m.holeczek@gmail.com>
Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] af9015: setup rc keytable for LC-Power LC-USB-DVBT
Juergen Lock [Sun, 12 Jun 2011 20:25:12 +0000 (17:25 -0300)]
[media] af9015: setup rc keytable for LC-Power LC-USB-DVBT

That's this tuner:

The credit card sized remote more or less works if I set remote=4,
so I added the hash to get it autodetected.  (`more or less' there
meaning sometimes buttons are `stuck on repeat', i.e. ir-keytable -t
keeps repeating the same scancode until i press another button.)

Signed-off-by: Juergen Lock <nox@jelal.kn-bremen.de>
Signed-off-by: Antti Palosaari <crope@iki.fi>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] media, Micronas dvb-t: Fix mem leaks, don't needlessly zero mem, fix spelling
Jesper Juhl [Wed, 13 Jul 2011 20:58:13 +0000 (17:58 -0300)]
[media] media, Micronas dvb-t: Fix mem leaks, don't needlessly zero mem, fix spelling

In drivers/media/dvb/frontends/drxd_hard.c::load_firmware() I see 3
small issues:

 1) When the 'fw' variable goes out of scope we'll leak the memory
 allocated to it by request_firmware() by neglecting to call
 release_firmware().

 2) After a successful request_firmware() we allocate fw->size bytes
 of memory using kzalloc() only to immediately overwrite all that
 memory with memcpy(), so asking for zeroed memory seems like wasted
 effort - just use kmalloc().

 3) In one of the error messages "no memory" lacks a space and is
 written as "nomemory".

This patch fixes all 3 issues.

Signed-off-by: Jesper Juhl <jj@chaosbits.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] rc-rc6-mce: minor keymap updates
Jarod Wilson [Wed, 13 Jul 2011 20:57:42 +0000 (17:57 -0300)]
[media] rc-rc6-mce: minor keymap updates

Microsoft's Windows Media Center specification and requirements doc from
2011.03.18 now refers to the former Power Toggle button as the Sleep
Toggle, and recommends using a new moon sleep icon for it. Its the same
key, but its apparently always been meant to put the system to sleep,
not power it off. Adjust accordingly. While we're here, lets also remove
the duplicate KEY_PLAYPAUSE entry.

Signed-off-by: Jarod Wilson <jarod@redhat.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] drivers/media/video/cx231xx/cx231xx-cards.c: add missing kfree
Julia Lawall [Mon, 4 Jul 2011 14:11:41 +0000 (11:11 -0300)]
[media] drivers/media/video/cx231xx/cx231xx-cards.c: add missing kfree

Clear the cx231xx_devused variable and free dev in the error handling code,
as done in the error handling code nearby.

The semantic match that finds this problem is as follows:

// <smpl>
@r@
identifier x;
@@

kfree(x)

@@
identifier r.x;
expression E1!=0,E2,E3,E4;
statement S;
@@

(
if (<+...x...+>) S
|
if (...) { ... when != kfree(x)
               when != if (...) { ... kfree(x); ... }
               when != x = E3
* return E1;
}
... when != x = E2
if (...) { ... when != x = E4
 kfree(x); ... return ...; }
)
// </smpl>

Signed-off-by: Julia Lawall <julia@diku.dk>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] drivers/media/video: add missing kfree
Julia Lawall [Mon, 4 Jul 2011 14:11:42 +0000 (11:11 -0300)]
[media] drivers/media/video: add missing kfree

Free the recently allocated qcam in each case.

The semantic match that finds this problem is as follows:

// <smpl>
@r@
identifier x;
@@

kfree(x)

@@
identifier r.x;
expression E1!=0,E2,E3,E4;
statement S;
@@

(
if (<+...x...+>) S
|
if (...) { ... when != kfree(x)
               when != if (...) { ... kfree(x); ... }
               when != x = E3
* return E1;
}
... when != x = E2
if (...) { ... when != x = E4
 kfree(x); ... return ...; }
)
// </smpl>

Signed-off-by: Julia Lawall <julia@diku.dk>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] TM6000: alsa: Clean up kernel coding style errors
Adam M. Dutko [Thu, 23 Jun 2011 14:57:45 +0000 (11:57 -0300)]
[media] TM6000: alsa: Clean up kernel coding style errors

There were several coding style errors as reported by checkpatch.pl. This
patch should fix those errors with the single exception of the open square
bracket issue on line 45.

[mchehab@redhat.com: Fix a merge conflict]
Signed-off-by: Adam M. Dutko <dutko.adam@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] tm6000: remove a check for NO_PCM_LOCK
Mauro Carvalho Chehab [Wed, 13 Jul 2011 22:15:57 +0000 (19:15 -0300)]
[media] tm6000: remove a check for NO_PCM_LOCK

This was likely used by some test. There's no sense on keeping it
upstream.

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] DVB: dvb_frontend: off by one in dtv_property_dump()
Dan Carpenter [Thu, 26 May 2011 08:44:52 +0000 (05:44 -0300)]
[media] DVB: dvb_frontend: off by one in dtv_property_dump()

If the tvp->cmd == DTV_MAX_COMMAND then we read past the end of the
array.

Signed-off-by: Dan Carpenter <error27@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] cx23885: added support for card 107d:6f39
istvan_v@mailbox.hu [Mon, 11 Jul 2011 13:58:35 +0000 (10:58 -0300)]
[media] cx23885: added support for card 107d:6f39

This patch, based on code by Mirek Slugen, implements support for the
Leadtek WinFast PxDVR3200 H card with XC4000 tuner (107d:6f39).

Signed-off-by: Istvan Varga <istvan_v@mailbox.hu>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] drivers/media: static should be at beginning of declaration
Jesper Juhl [Sat, 9 Jul 2011 21:22:17 +0000 (18:22 -0300)]
[media] drivers/media: static should be at beginning of declaration

Make sure that the 'static' keywork is at the beginning of declaration
for drivers/media/video/omap/omap_vout.c

This gets rid of warnings like
when building with -Wold-style-declaration (and/or -Wextra which also
enables it).

Signed-off-by: Jesper Juhl <jj@chaosbits.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] marvell-cam: clean up a couple of unused cam structure fields
Jonathan Corbet [Fri, 8 Jul 2011 20:50:50 +0000 (17:50 -0300)]
[media] marvell-cam: clean up a couple of unused cam structure fields

Delete a couple of leftover fields whose time has passed.

Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] marvell-cam: Allow selection of supported buffer modes
Jonathan Corbet [Fri, 8 Jul 2011 20:50:49 +0000 (17:50 -0300)]
[media] marvell-cam: Allow selection of supported buffer modes

The Marvell camera core can support all three videobuf2 buffer modes, which
is slick, but it also requires that all three modes be built and present,
even though only one is likely to be used.  This patch allows the supported
modes to be selected at configuration time, reducing the footprint of the
driver.  Prior to this patch, the MMP camera driver looked like this:

mmp_camera             19092  0
videobuf2_core         15542  1 mmp_camera
videobuf2_dma_sg        3173  1 mmp_camera
videobuf2_dma_contig     2188  1 mmp_camera
videobuf2_vmalloc       1718  1 mmp_camera
videobuf2_memops        2100  3 videobuf2_dma_sg,videobuf2_dma_contig,videobuf2_vmalloc

Afterward, instead, with scatter/gather only configured:

mmp_camera             16021  0
videobuf2_core         15542  1 mmp_camera
videobuf2_dma_sg        3173  1 mmp_camera
videobuf2_memops        2100  1 videobuf2_dma_sg

The total goes from 43,813 bytes to 36,836.

The emphasis has been on simplicity and minimal #ifdef use rather than on
squeezing out every possible byte of code.  For configuration, the driver
simply looks at which videobuf2 modes have been configured in and supports
them all; it's simplistic but should be good enough.

The cafe driver is set to support vmalloc and dma-contig; mmp supports only
dma-sg, since that's the only mode that really makes sense to use.

Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] marvell-cam: power down mmp camera on registration failure
Jonathan Corbet [Fri, 8 Jul 2011 20:50:48 +0000 (17:50 -0300)]
[media] marvell-cam: power down mmp camera on registration failure

If registration does not work, we don't want to leave the sensor powered on.

Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] marvell-cam: remove {min,max}_buffers parameters
Jonathan Corbet [Fri, 8 Jul 2011 20:50:47 +0000 (17:50 -0300)]
[media] marvell-cam: remove {min,max}_buffers parameters

Somewhere along the way the code stopped actually paying any attention to
them, and I doubt anybody has ever made use of them.

Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] marvell-cam: core code reorganization
Jonathan Corbet [Fri, 8 Jul 2011 20:50:46 +0000 (17:50 -0300)]
[media] marvell-cam: core code reorganization

This code shows signs of having been mucked with over the last five years
or so; things were kind of mixed up.  This patch reorders functions into a
more rational organization which, with luck, will facilitate making the
buffer modes selectable at configuration time.  Code movement only: no
functional changes here.

Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] marvell-cam: delete struct mcam_sio_buffer
Jonathan Corbet [Fri, 8 Jul 2011 20:50:45 +0000 (17:50 -0300)]
[media] marvell-cam: delete struct mcam_sio_buffer

This structure got passed over in the videobuf2 conversion; it has no
reason to exist now.

Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] tea5764: Fix module parameter permissions
Jean Delvare [Fri, 8 Jul 2011 09:00:37 +0000 (06:00 -0300)]
[media] tea5764: Fix module parameter permissions

The third parameter of module_param is supposed to represent sysfs
file permissions. A value of "1" leads to the following:

$ ls -l /sys/module/radio_tea5764/parameters/
total 0

I am changing it to "0" to align with the other module parameters in
this driver.

Signed-off-by: Jean Delvare <jdelvare@suse.de>
Cc: Mauro Carvalho Chehab <mchehab@infradead.org>
Cc: Fabio Belavenuto <belavenuto@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] Make Compro VideoMate Vista T750F actually work
Carlos Corbacho [Sat, 25 Jun 2011 13:24:28 +0000 (10:24 -0300)]
[media] Make Compro VideoMate Vista T750F actually work

Based on the work of John Newbigin, Davor Emard and others who contributed
on the mailing lists.

The previous 'support' for this card was a partial merge of John's changes
that, as far as I can tell, never actually got the thing working (no DVB-T,
analog tuner not initialised).

Initialise the analog tuner properly and hook up the DVB tuner and demodulator.

DVB-T and analog now work (though I can't tune every DVB channel, but I think
there's an issue with the aerial and signal boosters here that is causing
me problems).

Signed-off-by: Carlos Corbacho <carlos@strangeworlds.co.uk>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years agotda18271c2dd: Fix compilation when module is not selected
Mauro Carvalho Chehab [Mon, 11 Jul 2011 18:47:40 +0000 (15:47 -0300)]
tda18271c2dd: Fix compilation when module is not selected

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] tda18271c2dd.h: Don't add the same symbol twice
Mauro Carvalho Chehab [Mon, 11 Jul 2011 18:28:12 +0000 (15:28 -0300)]
[media] tda18271c2dd.h: Don't add the same symbol twice

Should be folded with the next patch when sending upstream

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] dvb: don't cause missing symbols for drxk/tda18271c2dd
Mauro Carvalho Chehab [Mon, 11 Jul 2011 17:56:30 +0000 (14:56 -0300)]
[media] dvb: don't cause missing symbols for drxk/tda18271c2dd

If those demods are unselected, but a bridge driver requires them,
produce a runtime message, instead of missing symbols.

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] drxk: remove a now unused variable
Mauro Carvalho Chehab [Mon, 11 Jul 2011 17:56:03 +0000 (14:56 -0300)]
[media] drxk: remove a now unused variable

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] em28xx: Change firmware name for Terratec H5 DRX-K
Mauro Carvalho Chehab [Mon, 11 Jul 2011 17:33:51 +0000 (14:33 -0300)]
[media] em28xx: Change firmware name for Terratec H5 DRX-K

Use a name convention for the firmware file that matches on the
current firmware namespacing. Also, add it to the firmware
download script.

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] drxk: Add a fallback method for QAM parameter setting
Mauro Carvalho Chehab [Sun, 10 Jul 2011 17:33:29 +0000 (14:33 -0300)]
[media] drxk: Add a fallback method for QAM parameter setting

The QAM standard is set using this scu_command:
SCU_RAM_COMMAND_STANDARD_QAM |
SCU_RAM_COMMAND_CMD_DEMOD_SET_PARAM

The driver implements a version that has 4 parameters, however,
Terratec H5 needs to break this into two separate commands, otherwise,
DVB-C doesn't work.

With this fix, scan is now properly working and getting the
channel list:
>>> tune to: 609000000:INVERSION_AUTO:5217000:FEC_3_4:QAM_256
>>> tuning status == 0x00
>>> tuning status == 0x07
>>> tuning status == 0x1f

0x0093 0x0026: pmt_pid 0x0758 (null) -- SporTV2 (running, scrambled)
0x0093 0x0027: pmt_pid 0x0748 (null) -- SporTV (running, scrambled)
0x0093 0x0036: pmt_pid 0x0768 (null) -- FX (running, scrambled)
0x0093 0x0052: pmt_pid 0x0788 (null) -- The History Channel (running, scrambled)

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] drxk: Improve the scu_command error message
Mauro Carvalho Chehab [Sun, 10 Jul 2011 16:25:48 +0000 (13:25 -0300)]
[media] drxk: Improve the scu_command error message

Now, it outputs:

[10927.639641] drxk: SCU_RESULT_INVPAR while sending cmd 0x0203 with params:
[10927.646283] drxk: 02 00 00 00 10 00 07 00 03 02                    ..........

Better than ERROR -3. This happens with Terratec H5 firmware.

It adds 2 new error conditions, and something useful to track
what the heck is that.

I suspect that the scu_command is dependent on the firmware
revision.

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] drxk: Simplify the DVB-C set mode logic
Mauro Carvalho Chehab [Sun, 10 Jul 2011 16:08:44 +0000 (13:08 -0300)]
[media] drxk: Simplify the DVB-C set mode logic

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] drxk: Fix driver removal
Mauro Carvalho Chehab [Sun, 10 Jul 2011 13:26:06 +0000 (10:26 -0300)]
[media] drxk: Fix driver removal

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] drxk: Improves the UIO handling
Mauro Carvalho Chehab [Sun, 10 Jul 2011 12:36:30 +0000 (09:36 -0300)]
[media] drxk: Improves the UIO handling

The driver is too limited: it assumes that UIO is used only for
controlling the antenna, and that only UIO-1 is in usage. However,
from Terratec H7 driver [1], 3 UIO's can be used. In fact, it seems
that H7 needs to use all 3. So, make the code generic enough to handle
the most complex scenario. For now, only antena GPIO can be specified,
but is is easier now to add the other GPIO/UIO needs.

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] drxk: Print detected configuration
Mauro Carvalho Chehab [Sun, 10 Jul 2011 11:38:18 +0000 (08:38 -0300)]
[media] drxk: Print detected configuration

DRX-K configuration is interesting when writing/testing
new devices. Add an info line showing the discovered info.

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] drxk: Fix the antenna switch logic
Mauro Carvalho Chehab [Sun, 10 Jul 2011 11:24:26 +0000 (08:24 -0300)]
[media] drxk: Fix the antenna switch logic

Terratec H5 doesn't require to switch mode, but generates
an error due to this logic. Also, GPIO's are board-dependent.

So, add it at the board config struct.

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] drxk: change mode before calling the set mode routines
Mauro Carvalho Chehab [Sun, 10 Jul 2011 16:07:36 +0000 (13:07 -0300)]
[media] drxk: change mode before calling the set mode routines

The set mode routines assume that state were changed to the
new mode, otherwise, they'll fail.

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] drxk: Proper handle/propagate the error codes
Mauro Carvalho Chehab [Sun, 10 Jul 2011 04:49:53 +0000 (01:49 -0300)]
[media] drxk: Proper handle/propagate the error codes

This driver is very big and complex. An error happening in the middle
of any initialization may cause the frontend to not work. So, it
needs to properly propagate error codes internally and to userspace.

Also, printing the error codes at the places it happened helps to
discover were's a bug at the code.

Before this change, a do { } while (0) loop and lots of breaks inside
were used to propagate errors. While this works, if there are
loops inside other loops, it could be easy to forget to add another
break, causing the error to not abort the function.

Also, as not all functions were reporting errors, it is hard to
discover why something failed.

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] drxk: Allow to disable I2C Bridge control switch
Mauro Carvalho Chehab [Sun, 10 Jul 2011 00:59:33 +0000 (21:59 -0300)]
[media] drxk: Allow to disable I2C Bridge control switch

On em28xx, tda18271C2 is accessible when the i2c port
is not touched. Touching on it breaks the driver.

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] em28xx-i2c: Add a read after I2C write
Mauro Carvalho Chehab [Sat, 9 Jul 2011 22:36:11 +0000 (19:36 -0300)]
[media] em28xx-i2c: Add a read after I2C write

All I2C logs we got for em28xx does that. With Terratec H5, at
400MHz speed, it seems that this is required, to avoid having
troubles at the I2C bus.

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] drxk: Add a parameter for the microcode name
Mauro Carvalho Chehab [Sat, 9 Jul 2011 20:35:26 +0000 (17:35 -0300)]
[media] drxk: Add a parameter for the microcode name

The microcode firmware provided on Terratec H5 seems to be
different. Add a parameter to allow specifying a different
firmware per-device.

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] Add initial support for Terratec H5
Mauro Carvalho Chehab [Mon, 4 Jul 2011 00:05:06 +0000 (21:05 -0300)]
[media] Add initial support for Terratec H5

Not working yet. There are some fixes at the DRX-K that are needed
for it to work.

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] drxk: Print an error if firmware is not loaded
Mauro Carvalho Chehab [Sat, 9 Jul 2011 22:23:44 +0000 (19:23 -0300)]
[media] drxk: Print an error if firmware is not loaded

If something bad happens during firmware load, an error
should be printed at dmesg.

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] drxk: Avoid OOPSes if firmware is corrupted
Mauro Carvalho Chehab [Sat, 9 Jul 2011 21:57:54 +0000 (18:57 -0300)]
[media] drxk: Avoid OOPSes if firmware is corrupted

Don't read paste the buffer, if the firmware is corrupted.
Instead, print an error message.

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] drxk: Convert an #ifdef logic as a new config parameter
Mauro Carvalho Chehab [Sat, 9 Jul 2011 16:06:12 +0000 (13:06 -0300)]
[media] drxk: Convert an #ifdef logic as a new config parameter

Instead of using #ifdef I2C_LONG_ADR for some devices, convert
it into a parameter. Terratec H5 logs from the original driver
seems to need this mode.

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] drxk: Move I2C address into a config structure
Mauro Carvalho Chehab [Sat, 9 Jul 2011 15:36:58 +0000 (12:36 -0300)]
[media] drxk: Move I2C address into a config structure

Currently, the only parameter to be configured is the I2C
address. However, Terratec H5 logs shows that it needs a different
setting for some things, and it has its own firmware.

So, move the addr into a config structure, in order to allow adding
the required configuration bits.

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] drxk: remove _0 from read/write routines
Mauro Carvalho Chehab [Sat, 9 Jul 2011 12:50:21 +0000 (09:50 -0300)]
[media] drxk: remove _0 from read/write routines

The normal 16-bits read routine is called as "Read16_0". This is
due to a flags that could optionally be passed. Yet, on no places
at the code, a flag is passed there.

The same happens with 16-bits write and 32-read/write routines,
and with WriteBlock.

Also, using flags, is an exception: there's no place currently using
flags, except for an #ifdef at WriteBlock.

Rename the function as just "read16", and the one that requires flags,
as "read16_flags".

This helps to see where the flags are used, and also avoid using
CamelCase on Kernel.

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] drxk: Add debug printk's
Mauro Carvalho Chehab [Mon, 4 Jul 2011 20:39:21 +0000 (17:39 -0300)]
[media] drxk: Add debug printk's

This is a complex driver. Adding support for other devices with drxk
requires to be able to debug it and see where it is failing. So, add
optional printk messages to allow debugging it.

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] tda18271c2dd: add tda18271c2dd prefix to the errors
Mauro Carvalho Chehab [Mon, 4 Jul 2011 17:00:25 +0000 (14:00 -0300)]
[media] tda18271c2dd: add tda18271c2dd prefix to the errors

It is hard to identify the origin for those errors without a
prefix to indicate which driver produced them:

[ 1390.220984] i2c_write error
[ 1390.224133] I2C Write error
[ 1391.284202] i2c_read error
[ 1392.288685] i2c_read error

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] drxk: add drxk prefix to the errors
Mauro Carvalho Chehab [Mon, 4 Jul 2011 11:27:47 +0000 (08:27 -0300)]
[media] drxk: add drxk prefix to the errors

It is hard to identify the origin for those errors without a
prefix to indicate which driver produced them:

[ 1390.220984] i2c_write error
[ 1390.224133] I2C Write error
[ 1391.284202] i2c_read error
[ 1392.288685] i2c_read error

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] ddbridge: use linux/io.h, instead of asm/io.h
Mauro Carvalho Chehab [Sun, 3 Jul 2011 22:00:32 +0000 (19:00 -0300)]
[media] ddbridge: use linux/io.h, instead of asm/io.h

WARNING: Use #include <linux/io.h> instead of <asm/io.h>
+#include <asm/io.h>

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] ddbridge: Avoid duplicated symbol definition
Mauro Carvalho Chehab [Sun, 3 Jul 2011 21:59:08 +0000 (18:59 -0300)]
[media] ddbridge: Avoid duplicated symbol definition

drivers/media/dvb/ddbridge/built-in.o: In function `my_dvb_dmx_ts_card_init':
/home/v4l/v4l/patchwork/drivers/media/dvb/ddbridge/ddbridge-core.c:718: multiple definition of `my_dvb_dmx_ts_card_init'
drivers/media/dvb/ngene/built-in.o:/home/v4l/v4l/patchwork/drivers/media/dvb/ngene/ngene-dvb.c:227: first defined here
drivers/media/dvb/ddbridge/built-in.o: In function `my_dvb_dmxdev_ts_card_init':
/home/v4l/v4l/patchwork/drivers/media/dvb/ddbridge/ddbridge-core.c:737: multiple definition of `my_dvb_dmxdev_ts_card_init'
drivers/media/dvb/ngene/built-in.o:/home/v4l/v4l/patchwork/drivers/media/dvb/ngene/ngene-dvb.c:246: first defined here

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] cxd2099: Update Kconfig description (ddbridge support)
Oliver Endriss [Sun, 3 Jul 2011 21:27:51 +0000 (18:27 -0300)]
[media] cxd2099: Update Kconfig description (ddbridge support)

Update Kconfig description (ddbridge with cxd2099)

Signed-off-by: Oliver Endriss <o.endriss@gmx.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] cxd2099: Fix compilation of ngene/ddbridge for DVB_CXD2099=n
Oliver Endriss [Sun, 3 Jul 2011 21:26:22 +0000 (18:26 -0300)]
[media] cxd2099: Fix compilation of ngene/ddbridge for DVB_CXD2099=n

Fix compilation of ngene/ddbridge for DVB_CXD2099=n.

Note: Bug was introduced by commit 'cxd2099: Update to latest version'.

Signed-off-by: Oliver Endriss <o.endriss@gmx.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] ddbridge: Allow compiling of the driver
Oliver Endriss [Sun, 3 Jul 2011 21:25:29 +0000 (18:25 -0300)]
[media] ddbridge: Allow compiling of the driver

Driver added to Makefile and Kconfig.

Signed-off-by: Oliver Endriss <o.endriss@gmx.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] ddbridge: Codingstyle fixes
Oliver Endriss [Sun, 3 Jul 2011 21:24:07 +0000 (18:24 -0300)]
[media] ddbridge: Codingstyle fixes

Codingstyle fixes

Signed-off-by: Oliver Endriss <o.endriss@gmx.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] ddbridge: Initial check-in
Ralph Metzler [Sun, 3 Jul 2011 21:23:11 +0000 (18:23 -0300)]
[media] ddbridge: Initial check-in

Driver support for Digital Devices ddbridge-based cards:
Octopus, Octopus mini, Octopus LE, cineS2(v6)
with DuoFlex S2 and/or DuoFlex CT tuners.

Driver was taken from ddbridge-0.6.1.tar.bz2.

Signed-off-by: Ralph Metzler <rmetzler@digitaldevices.de>
Signed-off-by: Oliver Endriss <o.endriss@gmx.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years agocxd2099: Remove the CHK_ERROR macro
Mauro Carvalho Chehab [Sun, 3 Jul 2011 21:45:37 +0000 (18:45 -0300)]
cxd2099: Remove the CHK_ERROR macro

The CHK_ERROR macro does a flow control, violating chapter 12
of the Documentation/CodingStyle. Doing flow controls inside
macros is a bad idea, as it hides what's happening. It also
hides the var "status" with is also a bad idea.

The changes were done by this small perl script:
my $blk=0;
while (<>) {
s/^\s+// if ($blk);
$f =~ s/\s+$// if ($blk && /^\(/);
$blk = 1 if (!m/\#/ && m/CHK_ERROR/);
$blk=0 if ($blk && m/\;/);
s/\n/ / if ($blk);
$f.=$_;
};
$f=~ s,\n(\t+)CHK_ERROR\((.*)\)\;([^\n]*),\n\1status = \2;\3\n\1if (status < 0)\n\1\tbreak;,g;

print $f;

And manually fixed.

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] drxk: fix warning: ‘status’ may be used uninitialized in this function
Mauro Carvalho Chehab [Sun, 3 Jul 2011 21:21:59 +0000 (18:21 -0300)]
[media] drxk: fix warning: ‘status’ may be used uninitialized in this function

One of the problems of the old CHECK_ERROR is that it was hiding
the status parameter. Maybe due to that, on a few places, the return
code might lead to return incorrect status:

drivers/media/dvb/frontends/drxk_hard.c: In function ‘load_microcode.clone.0’:
drivers/media/dvb/frontends/drxk_hard.c:1281: warning: ‘status’ may be used uninitialized in this function
drivers/media/dvb/frontends/drxk_hard.c:1281: note: ‘status’ was declared here
drivers/media/dvb/frontends/drxk_hard.c: In function ‘GetLockStatus’:
drivers/media/dvb/frontends/drxk_hard.c:1792: warning: ‘status’ may be used uninitialized in this function
drivers/media/dvb/frontends/drxk_hard.c: In function ‘Start.clone.7’:
drivers/media/dvb/frontends/drxk_hard.c:1734: warning: ‘status’ may be used uninitialized in this function

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] drxk: Return -EINVAL if an invalid bandwidth is used
Mauro Carvalho Chehab [Sun, 3 Jul 2011 21:18:14 +0000 (18:18 -0300)]
[media] drxk: Return -EINVAL if an invalid bandwidth is used

drivers/media/dvb/frontends/drxk_hard.c: In function ‘SetDVBT’:
drivers/media/dvb/frontends/drxk_hard.c:3766: warning: enumeration value ‘BANDWIDTH_5_MHZ’ not handled in switch
drivers/media/dvb/frontends/drxk_hard.c:3766: warning: enumeration value ‘BANDWIDTH_10_MHZ’ not handled in switch
drivers/media/dvb/frontends/drxk_hard.c:3766: warning: enumeration value ‘BANDWIDTH_1_712_MHZ’ not handled in switch

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] tda18271c2dd: Remove the CHK_ERROR macro
Mauro Carvalho Chehab [Sun, 3 Jul 2011 21:12:26 +0000 (18:12 -0300)]
[media] tda18271c2dd: Remove the CHK_ERROR macro

The CHK_ERROR macro does a flow control, violating chapter 12
of the Documentation/CodingStyle. Doing flow controls inside
macros is a bad idea, as it hides what's happening. It also
hides the var "status" with is also a bad idea.

The changes were done by this small perl script:
my $blk=0;
while (<>) {
s/^\s+// if ($blk);
$f =~ s/\s+$// if ($blk && /^\(/);
$blk = 1 if (!m/\#/ && m/CHK_ERROR/);
$blk=0 if ($blk && m/\;/);
s/\n/ / if ($blk);
$f.=$_;
};
$f=~ s,\n(\t+)CHK_ERROR\((.*)\)\;([^\n]*),\n\1status = \2;\3\n\1if (status < 0)\n\1\tbreak;,g;

print $f;

And manually fixed.
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] drxk: Remove the CHK_ERROR macro
Mauro Carvalho Chehab [Sun, 3 Jul 2011 21:06:07 +0000 (18:06 -0300)]
[media] drxk: Remove the CHK_ERROR macro

The CHK_ERROR macro does a flow control, violating chapter 12
of the Documentation/CodingStyle. Doing flow controls inside
macros is a bad idea, as it hides what's happening. It also
hides the var "status" with is also a bad idea.

The changes were done by this small perl script:
my $blk=0;
while (<>) {
s /^\s+// if ($blk);
$f =~ s/\s+$// if ($blk && /^\(/);
$blk = 1 if (!m/\#/ && m/CHK_ERROR/);
$blk=0 if ($blk && m/\;/);
s/\n/ / if ($blk);
$f.=$_;
};
$f=~ s,\n(\t+)CHK_ERROR\((.*)\)\;([^\n]*),\n\1status = \2;\3\n\1if (status < 0)\n\1\tbreak;,g;
print $f;

And manually fixed.

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] drxd/drxk: Don't export MulDiv32 symbol
Mauro Carvalho Chehab [Sun, 3 Jul 2011 20:18:57 +0000 (17:18 -0300)]
[media] drxd/drxk: Don't export MulDiv32 symbol

/home/v4l/v4l/patchwork/drivers/media/dvb/frontends/drxk_hard.c:181: multiple definition of `MulDiv32'
drivers/media/dvb/frontends/drxd.o:/home/v4l/v4l/patchwork/drivers/media/dvb/frontends/drxd_hard.c:236: first defined here

Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] ngene: Strip dummy packets inserted by the driver
Oliver Endriss [Sun, 3 Jul 2011 17:04:46 +0000 (14:04 -0300)]
[media] ngene: Strip dummy packets inserted by the driver

As the CI requires a continuous data stream, the driver inserts dummy
packets when necessary. Do not pass these packets to userspace anymore.

Signed-off-by: Oliver Endriss <o.endriss@gmx.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] ngene: Update for latest cxd2099
Oliver Endriss [Sun, 3 Jul 2011 17:03:49 +0000 (14:03 -0300)]
[media] ngene: Update for latest cxd2099

Modifications for latest cxd2099.

Signed-off-by: Oliver Endriss <o.endriss@gmx.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] cxd2099: Codingstyle fixes
Oliver Endriss [Sun, 3 Jul 2011 17:02:24 +0000 (14:02 -0300)]
[media] cxd2099: Codingstyle fixes

Codingstyle fixes.

Signed-off-by: Oliver Endriss <o.endriss@gmx.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] cxd2099: Update to latest version
Ralph Metzler [Sun, 3 Jul 2011 17:00:57 +0000 (14:00 -0300)]
[media] cxd2099: Update to latest version

Import latest driver from ddbridge-0.6.1.tar.bz2.

Signed-off-by: Ralph Metzler <rmetzler@digitaldevices.de>
Signed-off-by: Oliver Endriss <o.endriss@gmx.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
13 years ago[media] ngene: Support DuoFlex CT attached to CineS2 and SaTiX-S2
Oliver Endriss [Sun, 3 Jul 2011 16:59:30 +0000 (13:59 -0300)]
[media] ngene: Support DuoFlex CT attached to CineS2 and SaTiX-S2

Support DuoFlex CT with Digital Devices CineS2 and Mystique SaTiX-S2.

Signed-off-by: Oliver Endriss <o.endriss@gmx.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>