Hans Verkuil [Fri, 29 Aug 2014 08:45:59 +0000 (05:45 -0300)]
[media] cx88: increase API command timeout
The timeout is way too small. Especially complicated command like
CX2341X_ENC_STOP_CAPTURE takes much more time than 10 ms. Increase the
timeout to 1 second, just as ivtv does (the cx88-blackbird has the
same IP core for MPEG compression as ivtv).
This solves a nasty issue where STOP_CAPTURE would timeout and the
mailbox is left in a busy state, making it impossible to start streaming
a second time without reloading the driver first.
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Hans Verkuil [Sat, 20 Sep 2014 12:22:18 +0000 (09:22 -0300)]
[media] cx88: convert to vb2
As usual, this patch is very large due to the fact that half a vb2 conversion
isn't possible. And since this affects blackbird, alsa, core, dvb, vbi and
video the changes are all over.
What made this more difficult was the peculiar way the risc program was setup.
The driver allowed for running out of buffers in which case the DMA would stop
and restart when the next buffer was queued. There was also a complicated
timeout system for when buffers weren't filled. This was replaced by a much
simpler scheme where there is always one buffer around and the DMA will just
cycle that buffer until a new buffer is queued. In that case the previous
buffer will be chained to the new buffer. An interrupt is generated at the
start of the new buffer telling the driver that the previous buffer can be
passed on to userspace.
Much simpler and more robust.
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Hans Verkuil [Fri, 29 Aug 2014 06:46:05 +0000 (03:46 -0300)]
[media] cx88: remove fmt from the buffer struct
This is a duplicate of dev->fmt and can be removed. As a consequence a
lot of tests that check if the format has changed midstream can be
removed as well: the format cannot change midstream, so this is a bogus
check.
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
[media] cx231xx: change usage of I2C_1 to the real i2c port
change almost all instances of I2C_1 to I2C_1_MUX_3
Only these cases are changed to I2C_1_MUX_1:
* All that have dont_use_port_3 set.
* CX231XX_BOARD_HAUPPAUGE_EXETER, old code did explicitly not switch to port3.
* eeprom access for 930C
[media] cx231xx: give each master i2c bus a seperate name
Instead of using the same name for all 3 i2c physical buses
inside cx231xx, name them differently, adding a number
to it. This helps to better deal with the logs.
[mchehab@osg.samsung.com: removed an unused bus_name var from
the original patch] Signed-off-by: Matthias Schwarzott <zzam@gentoo.org> Reviewed-by: Antti Palosaari <crope@iki.fi> Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Tomas Melin [Tue, 30 Sep 2014 13:32:08 +0000 (10:32 -0300)]
[media] rc-main: fix lockdep splash for rc-main
lockdep reports a potential circular dependecy deadlock when registering input device.
Unlock mutex rc_dev->lock prior to calling ir_raw_event_register to avoid the circular
dependency since that function also calls input_register_device and rc_open.
======================================================
[ INFO: possible circular locking dependency detected ]
3.17.0-rc7+ #24 Not tainted
-------------------------------------------------------
modprobe/647 is trying to acquire lock:
(input_mutex){+.+.+.}, at: [<ffffffff812ed81c>] input_register_device+0x2ba/0x381
but task is already holding lock:
(ir_raw_handler_lock){+.+.+.}, at: [<ffffffff813186ed>] ir_raw_event_register+0x102/0x190
Frank Schaefer [Mon, 29 Sep 2014 18:17:35 +0000 (15:17 -0300)]
[media] af9005: fix kernel panic on init if compiled without IR
This patches fixes an ancient bug in the dvb_usb_af9005 driver, which
has been reported at least in the following threads:
https://lkml.org/lkml/2009/2/4/350
https://lkml.org/lkml/2014/9/18/558
If the driver is compiled in without any IR support (neither
DVB_USB_AF9005_REMOTE nor custom symbols), the symbol_request calls in
af9005_usb_module_init() return pointers != NULL although the IR
symbols are not available.
The prefered fix would be to convert the whole IR code to use the kernel IR
infrastructure (which wasn't available at the time this driver had been created).
Until anyone who still has this old hardware steps up an does the conversion,
fix it by not calling the symbol_request calls if the driver is compiled in
without the default IR symbols (CONFIG_DVB_USB_AF9005_REMOTE).
Due to the IR related pointers beeing NULL by default, IR support will then be disabled.
The downside of this solution is, that it will no longer be possible to
compile custom IR symbols (not using CONFIG_DVB_USB_AF9005_REMOTE) in.
Please note that this patch has NOT been tested with all possible cases.
I don't have the hardware and could only verify that it fixes the reported
bug.
Olli Salonen [Mon, 29 Sep 2014 07:44:16 +0000 (04:44 -0300)]
[media] cx23855: add support for DVBSky T980C (no CI support)
This patch adds basic support for DVBSky T980C card. CI interface is not supported.
DVBSky T980C is a PCIe card with the following components:
- CX23885 PCIe bridge
- Si2168-A20 demodulator
- Si2158-A20 tuner
- CIMaX SP2 CI chip
The demodulator and tuner need firmware. They're the same as used with TT CT2-4650 CI:
https://www.mail-archive.com/linux-media@vger.kernel.org/msg78033.html
Signed-off-by: Olli Salonen <olli.salonen@iki.fi> Reviewed-by: Antti Palosaari <crope@iki.fi> Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
[media] ti-fpe: LLVMLinux: Remove nested function from ti-vpe
Replace the use of nested functions where a normal function will suffice.
Nested functions are not liked by upstream kernel developers in general. Their
use breaks the use of clang as a compiler, and doesn't make the code any
better.
Olli Salonen [Thu, 30 Oct 2014 10:43:16 +0000 (07:43 -0300)]
[media] si2157: add support for SYS_DVBC_ANNEX_B
Set the property for delivery system also in case of SYS_DVBC_ANNEX_B.
This behaviour is observed in the sniffs taken with Hauppauge HVR-955Q
Windows driver.
Signed-off-by: Olli Salonen <olli.salonen@iki.fi> Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
This patch partially reverts 0fb220f2a5cb. What happened is
that the conversion of debug messages to use pr_debug() was a
bad idea, because one needing to debug would need to both
enable debug level via a modprobe parameter, and then to
enable the dynamic printk's.
So, for now, let's use printk() directly at dprintk().
Hans Verkuil [Mon, 22 Sep 2014 14:08:55 +0000 (11:08 -0300)]
[media] v4l2-ctrls: fix sparse warning
The warning is simple:
drivers/media/v4l2-core/v4l2-ctrls.c:1685:15: warning: incorrect type in assignment (different address spaces)
but the fix isn't.
The core problem was that the conversion from user to kernelspace was
done at too low a level and that needed to be moved up. That made it possible
to drop pointers to v4l2_ext_control from set_ctrl and validate_new and
clean up this sparse warning because those functions now always operate
on kernelspace pointers.
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
this patch improve vpbe_buffer_prepare() callback, as buf_prepare()
callback is never called with invalid state and check for
vb2_plane_vaddr(vb, 0) is dropped as payload check should
be done unconditionally.
ayaka [Wed, 22 Oct 2014 18:03:08 +0000 (15:03 -0300)]
[media] s5p-mfc: correct the formats info for encoder
The NV12M is supported by all the version of MFC, so it is better
to use it as default OUTPUT format.
MFC v5 doesn't support NV21, I have tested it, for the SEC doc
it is not supported either.
Philipp Zabel [Tue, 21 Oct 2014 16:25:52 +0000 (13:25 -0300)]
[media] coda: re-queue buffers if start_streaming fails
Patch b906352c2338 ([media] coda: dequeue buffers if start_streaming fails)
incorrectly marked buffers as DEQUEUED in case of start_streaming failure, when
in fact they should be set to QUEUED. The videobuf2 core warns about this.
Pawel Osciak [Tue, 21 Oct 2014 11:07:07 +0000 (08:07 -0300)]
[media] s5p-mfc: fix a race in interrupt flags handling
Interrupt result flags have to be cleared before a hardware job is run.
Otherwise, if they are cleared asynchronously, we may end up clearing them
after the interrupt for which we wanted to wait has already arrived, thus
overwriting the job results that we intended to wait for.
To prevent this, clear the flags only under hw_lock and before running
a hardware job.
Pawel Osciak [Tue, 21 Oct 2014 11:07:06 +0000 (08:07 -0300)]
[media] s5p-mfc: fix V4L2_CID_MIN_BUFFERS_FOR_CAPTURE on resolution change
G_CTRL on V4L2_CID_MIN_BUFFERS_FOR_CAPTURE will fail if userspace happens to
query it after getting a resolution change event and before the codec has
a chance to parse the header and switch to an initialized state.
Kiran AVND [Tue, 21 Oct 2014 11:07:04 +0000 (08:07 -0300)]
[media] s5p-mfc: flush dpbs when resolution changes
While resolution change is detected by MFC, we flush out
older dpbs, send the resolution change event to application,
and then accept further queuing of new src buffers.
Sometimes, we error out during dpb flush because of lack of src
buffers. Since we have not started decoding new resolution yet,
it is simpler to push zero-size buffer until we flush out all dpbs.
This is already been done while handling EOS command, and this patch
extends the same logic to resolution change as well.
Pawel Osciak [Tue, 21 Oct 2014 11:07:01 +0000 (08:07 -0300)]
[media] s5p-mfc: Don't crash the kernel if the watchdog kicks in
If the software watchdog kicks in, the watchdog worker is not synchronized
with hardware interrupts and does not block other instances. It's possible
for it to clear the hw_lock, making other instances trigger a BUG() on
hw_lock checks. Since it's not fatal to clear the hw_lock to zero twice,
just WARN in those cases for now. We should not explode, as firmware will
return errors as needed for other instances after it's reloaded, or they
will time out.
A clean fix should involve killing other instances when watchdog kicks in,
but requires a major redesign of locking in the driver.
[media] s5p-mfc: declare s5p_mfc_bus_reset as static
drivers/media/platform/s5p-mfc/s5p_mfc_ctrl.c:132:5: warning: no previous prototype for 's5p_mfc_bus_reset' [-Wmissing-prototypes]
int s5p_mfc_bus_reset(struct s5p_mfc_dev *dev)
^
Kiran AVND [Tue, 21 Oct 2014 11:07:00 +0000 (08:07 -0300)]
[media] s5p-mfc: check mfc bus ctrl before reset
during reset sequence, it is advisable to follow the below
sequence, in order to avoid unexpected behavior from MFC
. set SFR 0x7110 MFC_BUS_RESET_CTRL 0x1
// wait for REQ_STATUS to be 1
. get SFR 0x7110 MFC_BUS_RESET_CTRL 0x3
// reset now
Ilja Friedel [Tue, 21 Oct 2014 11:06:58 +0000 (08:06 -0300)]
[media] s5p-mfc: Only set timestamp/timecode for new frames
Timestamp i of a previously decoded buffer was overwritten for some
H.264 streams with timestamp i+1 of the next buffer. This happened when
encountering frame_type S5P_FIMV_DECODE_FRAME_SKIPPED, indicating no
new frame.
In most cases this wrong indexing might not have been noticed except
for a one frame delay in frame presentation. For H.264 streams though
that require reordering of frames for presentation, it caused a slightly
erratic presentation time lookup and consequently dropped frames in the
Pepper Flash plugin.
Prathyush K [Tue, 21 Oct 2014 11:06:57 +0000 (08:06 -0300)]
[media] s5p-mfc: clear 'enter_suspend' flag if suspend fails
The enter_suspend flag is set after we enter mfc suspend function but if
suspend fails after that due to any reason (like hardware timeout etc),
this flag must be cleared before returning an error. Otherwise, this
flag never gets cleared and the MFC suspend will always return an error
on subsequent tries. If clock off fails, disable hw_lock also.
Kiran AVND [Tue, 21 Oct 2014 11:06:55 +0000 (08:06 -0300)]
[media] s5p-mfc: support MIN_BUFFERS query for encoder
Add V4L2_CID_MIN_BUFFERS_FOR_OUTPUT query for encoder.
Once mfc encoder state is HEAD_PARSED, which is sequence
header produced, dpb_count is avaialable. Let user space
query this value.
Philipp Zabel [Wed, 8 Oct 2014 16:09:27 +0000 (13:09 -0300)]
[media] coda: drop JPEG buffers not framed by SOI and EOI markers
This patch adds a quick check for valid JPEG frames before feeding them into
the bitstream buffer: Frames that do not begin with the JPEG start of image
marker and end with the end of image marker are dropped.
Philipp Zabel [Thu, 2 Oct 2014 17:08:35 +0000 (14:08 -0300)]
[media] coda: allow userspace to set compressed buffer size in a certain range
For small frame sizes, allocating 1 MiB per compressed buffer is a waste of
space. On the other hand, incompressible 1080p data can produce JPEGs larger
than 1 MiB at higher quality settings. Allow userspace to set the compressed
buffer size and clamp the value to a sensible range.
Also set the initial sizeimage to a value inside the range allowed by try_fmt.
While at it, reduce the default image size to a maximum of 1920*1088 (otherwise
JPEG will default to 8k*8k and 96 MiB buffers).
Philipp Zabel [Thu, 2 Oct 2014 17:08:34 +0000 (14:08 -0300)]
[media] coda: try to only queue a single JPEG into the bitstream
With bitstream padding, it is possible to decode a single JPEG in the bitstream
immediately. This allows us to only ever queue a single JPEG into the bitstream
buffer, except to increase payload over 512 bytes or to back out of hold state.
This is a measure to decrease JPEG decoder latency.
Philipp Zabel [Thu, 2 Oct 2014 17:08:33 +0000 (14:08 -0300)]
[media] coda: pad input stream for JPEG decoder
Before starting a PIC_RUN, pad the bitstream with 0xff until 256 bytes
past the next multiple of 256 bytes, if the buffer to be decoded is the
last buffer in the bitstream.
Philipp Zabel [Thu, 2 Oct 2014 17:08:32 +0000 (14:08 -0300)]
[media] coda: store bitstream buffer position with buffer metadata
Storing the buffer position in the bitstream with the buffer metadata
allows to later use that information to drop metadata for skipped buffers
and to determine whether bitstream padding has to be applied.
This patch also renames struct coda_timestamp to struct coda_buffer_meta
to make clear that it contains more than only the buffer timestamp.
Philipp Zabel [Thu, 2 Oct 2014 17:08:31 +0000 (14:08 -0300)]
[media] coda: add CODA7541 JPEG support
This patch adds JPEG encoding and decoding support for CODA7541,
using the BIT processor. Separate JPEG encoder and decoder video
devices are created due to different streaming behaviour and
different supported pixel formats.
The hardware can not change subsampling on the fly, but encode
and decode 4:2:2 subsampled JPEG images from and into this format.
The CODA7541 JPEG decoder uses the bitstream buffer and thus can run
without new buffers queued if there is a buffer in the bitstream.
Since there is no standard way to store the colorspace used in
JPEGs, and to make v4l2-compliance happy, the JPEG format always
reports V4L2_COLORSPACE_JPEG.
Signed-off-by: Lucas Stach <l.stach@pengutronix.de> Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de> Signed-off-by: Kamil Debski <k.debski@samsung.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Philipp Zabel [Thu, 2 Oct 2014 17:08:28 +0000 (14:08 -0300)]
[media] coda: add coda_video_device descriptors
Each video device descriptor determines the name, callback ops, and input and
output formats on the corresponding video device. This simplifies coda_enum_fmt
and coda_try_fmt a bit and will simplify adding separate video devices for JPEG
codecs due to the slightly different behavior in the CodaDx6/CODA7542 case and
a separate hardware unit on CODA960.
Philipp Zabel [Mon, 29 Sep 2014 12:53:46 +0000 (09:53 -0300)]
[media] coda: simplify frame memory control register handling
Since the firmware newer writes to FRAME_MEM_CTRL, we can initialize it once
per context (incidentally, we already do write it in coda_hw_init) and never
have to read it back.
Philipp Zabel [Mon, 29 Sep 2014 12:53:44 +0000 (09:53 -0300)]
[media] coda: add coda_write_base helper
Add a helper function that writes a vb2_buffer's Y, Cb, and
Cr plane base addresses of into three consecutive registers.
This moves common code out of coda-bit.c.
Philipp Zabel [Mon, 29 Sep 2014 12:53:42 +0000 (09:53 -0300)]
[media] coda: clear aborting flag in stop_streaming
Clearing the aborting flag in stop_streaming is necessary if we want to start
streaming again without having to closing and reopening the device. Also,
do not explicitly set it in default_params; the context is zeroed by
kzalloc anyway.
Hans Verkuil [Mon, 22 Sep 2014 12:27:17 +0000 (09:27 -0300)]
[media] mem2mem_testdev: rename to vim2m
This is 1) *much* easier to type, and 2) is consistent with vivid
('vi' for virtual). More of such virtual drivers are planned, so keeping
the naming consistent makes sense.
Note that the old module name is retained as a module alias.
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Sakari Ailus [Fri, 26 Sep 2014 13:12:38 +0000 (10:12 -0300)]
[media] smiapp: Clean up smiapp_set_format()
smiapp_set_format() has accumulated a fair amount of changes without a
needed refactoring, do the cleanup now. There's also an unlocked version of
v4l2_ctrl_range_changed(), using that fixes a small serialisation issue with
the user space interface.
__v4l2_ctrl_modify_range() is used instead of v4l2_ctrl_modify_range() in
smiapp_set_format_source() since the mutex is now held during the function
call.
Sakari Ailus [Tue, 1 Apr 2014 22:18:09 +0000 (19:18 -0300)]
[media] smiapp-pll: Add pixel rate in pixel array as output parameters
The actual pixel array pixel rate may be something else than vt_pix_clk_freq
on some implementations. Add a new field which contains the corrected value.
projects / mirror_ubuntu-hirsute-kernel.git / log