Here's the driver for the Renesas R-Car JPEG processing unit.
The driver is implemented within the V4L2 framework as a memory-to-memory
device. It presents two video nodes to userspace, one for the encoding part,
and one for the decoding part.
It was found that the only working mode for encoding is no markers output, so we
generate markers with software. In the current version of driver we also use
software JPEG header parsing because with hardware parsing performance is lower
than desired.
>From a userspace point of view the process is typical (S_FMT, REQBUF,
optionally QUERYBUF, QBUF, STREAMON, DQBUF) for both the source and destination
queues. STREAMON can return -EINVAL in case of mismatch of output and capture
queues format. Also during decoding driver can return buffers if queued
buffer with JPEG image contains image with inappropriate subsampling (e.g.
4:2:0 in JPEG and 4:2:2 in capture). If JPEG image and queue format dimensions
differ driver will return buffer on QBUF with VB2_BUF_STATE_ERROR flag.
During encoding the available formats are: V4L2_PIX_FMT_NV12M,
V4L2_PIX_FMT_NV12, V4L2_PIX_FMT_NV16, V4L2_PIX_FMT_NV16M for source and
V4L2_PIX_FMT_JPEG for destination.
During decoding the available formats are: V4L2_PIX_FMT_JPEG for source and
V4L2_PIX_FMT_NV12M, V4L2_PIX_FMT_NV16M, V4L2_PIX_FMT_NV12, V4L2_PIX_FMT_NV16
for destination.
Performance of current version:
1280x800 NV12 image encoding/decoding
decoding ~122 FPS
encoding ~191 FPS
Signed-off-by: Mikhail Ulyanov <mikhail.ulyanov@cogentembedded.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Currently, the warning for missing device_caps gives a backtrace like so:
[<ffffffff8175c199>] dump_stack+0x45/0x57
[<ffffffff8109ad5a>] warn_slowpath_common+0x8a/0xc0
[<ffffffff8109ae8a>] warn_slowpath_null+0x1a/0x20
[<ffffffffa0237453>] v4l_querycap+0x43/0x80 [videodev]
[<ffffffffa0237734>] __video_do_ioctl+0x2a4/0x320 [videodev]
[<ffffffff812207e5>] ? do_last+0x195/0x1210
[<ffffffffa023a11e>] video_usercopy+0x22e/0x5b0 [videodev]
[<ffffffffa0237490>] ? v4l_querycap+0x80/0x80 [videodev]
[<ffffffffa023a4b5>] video_ioctl2+0x15/0x20 [videodev]
[<ffffffffa0233733>] v4l2_ioctl+0x113/0x150 [videodev]
[<ffffffff81225798>] do_vfs_ioctl+0x2f8/0x4f0
[<ffffffff8113b2d4>] ? __audit_syscall_entry+0xb4/0x110
[<ffffffff81022d7c>] ? do_audit_syscall_entry+0x6c/0x70
[<ffffffff81225a11>] SyS_ioctl+0x81/0xa0
[<ffffffff8113b526>] ? __audit_syscall_exit+0x1f6/0x2a0
[<ffffffff81763549>] system_call_fastpath+0x12/0x17
This indicates that device_caps are missing but doesn't give
much of a clue which driver is actually at fault. Improve
the warning output by showing the capabilities and the
responsible driver.
Signed-off-by: Laura Abbott <labbott@fedoraproject.org>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
The bt pointer was never used, remove it.
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
video_unregister_device() calls device_unregister(), which calls
put_device(), which calls kobject_put(), and if this is the last reference
then kobject_release() is called, which calls kobject_cleanup(), which
calls ktype's release method which happens to be device_release() in this
case, which calls dev->release(), which happens to be
v4l2_device_release() in this case, which calls vdev->release(), which
happens to be video_device_release(). But video_device_release() is
called explicitly both in error recovery path of s5p_jpeg_probe() and
in s5p_jpeg_remove(). The pointers in question are not nullified between
the two calls, so this is harmful.
This patch fixes the driver so that video_device_release() is not called
twice for the same object.
Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
MFC hardware is known to trash random memory if one tries to use a
buffer which has lower DMA addresses than the configured DMA base
address. This patch adds a check for this case and proper error
handling.
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Acked-by: Kamil Debski <kamil@wypas.org>
[s.nawrocki@samsung.com: fixed typo (addres -> address]
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
MFCINST_GOT_INST state is set to encoder context with set_format
only for capture buffer. In queue_setup of encoder called during
reqbufs, it is checked MFCINST_GOT_INST state for both capture
and output buffer. So this patch fixes encoder to check
MFCINST_GOT_INST state only for capture buffer from queue_setup.
Signed-off-by: Seung-Woo Kim <sw0312.kim@samsung.com>
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
event API conformance testing with coccinelle spatches are being
used to locate API usage inconsistencies this triggert with:
./drivers/media/platform/s5p-tv/mixer_reg.c:364
incorrect check for negative return
Return type of wait_event_timeout is signed long not int and the
return type is >=0 always thus the negative check is unnecessary.
An appropriately named variable of type long is inserted and the
call fixed up aswell as the negative return check dropped.
Signed-off-by: Nicholas Mc Guire <hofrat@osadl.org>
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
i2c_driver does not need to set an owner because i2c_register_driver()
will set it.
Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Use eth_zero_addr to assign the zero address to the given address
array instead of memset when second argument is address of zero.
The Coccinelle semantic patch that makes this change is as follows:
// <smpl>
@eth_zero_addr@
expression e;
@@
-memset(e,0x00,ETH_ALEN);
+eth_zero_addr(e);
// </smpl>
Signed-off-by: Vaishali Thakkar <vthakkar1994@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Fix code indentation error by replacing tab in place of spaces.
Signed-off-by: Sunil Shahu <shshahu@gmail.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
COLORSPACE_JPEG should only be used for JPEGs. Use SMPTE170M instead,
which is how YCbCr images are usually encoded.
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Acked-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Format mac addresses with the normal kernel extension.
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
This is useful to subscribe to HDMI hotplug events via the
V4L2_CID_DV_RX_POWER_PRESENT control.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
When probed from device tree, the i2c client driver can handle the interrupt
on its own.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Add support for probing the TC358743 subdevice from device tree.
The reference clock must be supplied using the common clock bindings.
MIPI CSI-2 specific properties are parsed from the OF graph endpoint
node and support for a non-continuous MIPI CSI-2 clock is added.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Add V4L2_SUBDEV_FL_HAS_DEVNODE to subdev flags, in order to
enable a subdev device node.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Add support for registering the sensor subdevice using the v4l2-async API.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Currently the cobalt driver refuses to load if fewer than 8 PCIe lanes
are assigned. This patch changes this and just issues a warning. The
only time it will refuse to load is if the number of assigned lanes is less
than what the PCIe host is capable of since this suggests that the card
isn't seated correctly in the slot.
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
When vb2_is_busy() it should still be possible to call S_DV_TIMINGS
provided the new timings are the same as the current timings.
For input 1 (test generator) the size is always 1080p, so just return
that.
Fixes a v4l2-compliance issue.
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
There was a race condition where during cleanup/release operation
on-going streaming would cause a kernel panic because the hardware
module was disabled prematurely with IRQ still pending.
Fixes: 417d2e507e ("[media] media: platform: add VPFE capture driver support for AM437X")
Cc: <stable@vger.kernel.org> # v4.0+
Signed-off-by: Benoit Parrot <bparrot@ti.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Add videobuf2 specific vb2_qbuf and vb2_dqbuf trace events that mirror the
v4l2_qbuf and v4l2_dqbuf trace events, only they include additional
information about queue fill state and are emitted right before the buffer
is enqueued in the driver or userspace is woken up. This allows to make
sense of the timeline of trace events in combination with others that might
be triggered by __enqueue_in_driver.
Also two new trace events vb2_buf_queue and vb2_buf_done are added,
allowing to trace the handover between videobuf2 framework and driver.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
The queue owner will be used by videobuf2 trace events to determine and
record the device minor number. It is set in v4l2_m2m_reqbufs instead of
v4l2_m2m_ioctl_reqbufs because several drivers implement their own
vidioc_reqbufs handlers that still call v4l2_m2m_reqbufs directly.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Trace events with exactly the same parameters and trace output, such as
v4l2_qbuf and v4l2_dqbuf, are supposed to use the DECLARE_EVENT_CLASS and
DEFINE_EVENT macros instead of duplicated TRACE_EVENT macro calls.
Suggested-by: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
The chroma interleaved NV12 format has higher memory bandwidth efficiency
because the chroma planes can be read/written with longer burst lengths.
Use NV12 as default format if available and consistently sort it first.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Storing internal frames in macroblock tiled order improves memory
access patterns by allowing increased burst sizes when transferring
the uncompressed macroblocks to or from main memory.
The translation logic only supports a single chroma base address,
so this is only supported for the chroma interleaved NV12 format.
Since the rotator used to copy the decoder output into the v4l2
capture buffers does not seem to support the tiled format correctly,
only enable it in the encoder for now.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
The frame cache should be set up correctly to encode NV12 source frames.
This was not done before, so move the cache setup out of start_decoding
into its own function and call it from both start_encoding and
start_decoding.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
The encoder needs to know the nominal framerate for the constant bitrate
control mechanism to work. Currently the only way to set the framerate is
by using VIDIOC_S_PARM on the output queue.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
The encoder allows to specify the VBV model reference decoder's initial
delay and buffer size. Export the corresponding V4L2 controls.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
On i.MX6, sometimes after decoding a stream, encoding will produce macroblock
errors caused by missing 8-byte sequences in the output stream. Until the cause
for this is found, reset the hardware after sequence end, which seems to help.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Keep count of number of buffer meta structures in the list and use
a separate spinlock for operations on this counted list instead
of reusing the bitstream mutex in some places and none at all in
others.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
The v4l2_m2m_num_src_bufs_ready() function is called in multiple places
in coda_cob_ready, and there already is a variable src_bufs that is
assigned to its result. Move it to the beginning and use it everywhere.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Trace events with exactly the same parameters and trace output, such as
coda_enc_pic_run and coda_enc_pic_done, are supposed to use the
DECLARE_EVENT_CLASS and DEFINE_EVENT macros instead of duplicated
TRACE_EVENT macro calls.
This patch changes the order of parameters to coda_dec_rot_done and adds
a timestamp so it can share an event class with coda_bit_queue.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Since commit ba3002045f ("[media] v4l2-ioctl: fill in the description
for VIDIOC_ENUM_FMT"), all pixel formats are assigned their description
in a central place. We can now drop the custom list.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Allow coda_seq_end_work to be called multiple times, move the setting
of ctx->initialized from coda_start/stop_streaming() into
coda_start_encoding/decoding and coda_seq_end_work, respectively,
and skip the SEQ_END command in coda_seq_end_work if the context is
already deinitialized before.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
In stream end mode the hardware will read the bitstream to its end,
overshooting the write pointer. Do not write additional data into
the bitstream in this mode.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
All decoder instances using the BIT processor should preload buffers
into the bitstream ring buffer, including MPEG4 decoding. Fix this
by explicitly stating the above condition instead of listing all
relevant input formats.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
The mvcol buffer is allocated at the end of the first internal buffer.
This patch fixes an out of bounds array access.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
This is already done for one side of the comparison with the expectation
that the HW counter rolls over at the 16 bit boundary. This is true when
decoding a h.264 stream, but doesn't hold for at least MJPEG. As we don't
know the exact wrap-around point for this format just clamp the HW counter
to the same 16 bits. This should be enough to detect most of the errors
and saves us from doing different comparisons based on the decoded format.
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
The vsp1_entity_is_streaming() function is called in atomic context when
queuing buffers, and sleeps due to a mutex. As the mutex just protects
access to one structure field, fix this by replace the mutex with a
spinlock.
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
The checks need to be performed on up to two planes, as the third plane,
if present, must have the same stride and size as the second plane.
The code incorrectly performs the checks on at least two planes instead
of at most two planes, fix it.
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
We should set "ret" on this error path instead of returning success.
Fixes: df3305156f ('[media] v4l: xilinx: Add Xilinx Video IP core')
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Acked-by: Hyun Kwon <hyun.kwon@xilinx.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
The driver is tested on our hardware and all the implemented features
works as expected.
Missing features:
- CEC support
- HDCP repeater support
- IR support
Signed-off-by: Mats Randgaard <matrandg@cisco.com>
[hans.verkuil@cisco.com: updated copyright year to 2015]
[hans.verkuil@cisco.com: update confusing confctl_mutex comment]
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
bdisp_dev->dbg.copy_node shall be a copy of (and not point to)
bdisp_ctx->node, since this resource is freed upon driver release.
Signed-off-by: Fabien Dessenne <fabien.dessenne@st.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Support the composing (at VIDEO_CAPTURE) with the _selection API.
v4l2-compliance successfully run ("test Composing: OK")
Signed-off-by: Fabien Dessenne <fabien.dessenne@st.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
This commit implements frame decimation for stk1160, which allows
to support format changes instead of a static frame size.
The stk1160 supports independent row and column decimation, in two
different modes:
* set a number of rows/columns units to skip for each unit sent.
* set a number of rows/columns units to send for each unit skipped.
This effectively allows to achieve different frame scaling ratios.
The unit number can be set to either two row/columns sent/skipped,
or four row/columns sent/skipped. Since the video format (UYVY)
has 4-bytes, using a unit number of two row/columns results
in frame color 'shifting', so set to four row/columns sent/skipped.
Signed-off-by: Michael Stegemann <michael@stegemann.it>
Signed-off-by: Dale Hamel <dale.hamel@srvthe.net>
Signed-off-by: Ezequiel Garcia <ezequiel@vanguardiasur.com.ar>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
