media: vim2m: allow setting the default transaction time via parameter
While there's a control to allow setting it at runtime, as the
control handler is per file handler, only the application setting
the m2m device can change it. As this is a custom control, it is
unlikely that existing apps would be able to set it.
Due to that, and due to the fact that v4l2-mem2mem serializes all
accesses to a m2m device, trying to setup two GStreamer
v4l2videoconvert instance at the same time will cause frame drops.
So, add an alternate way of setting its default via a modprobe parameter.
It doesn't make sense to have a per-device work queue, as the
scheduler should be called per file handler. Having a single
one causes failures if multiple streams are filtered by vim2m.
media: vim2m: fix driver for it to handle different fourcc formats
Despite vim2m is reporting that it supports RGB565BE and YUYV,
that's not true.
Right now, it just says that it supports both format, but it
doesn't actually support them.
Also, horizontal flip is not properly implemented. It sounds
that it was designed to do a pseudo-horizontal flip using 8
tiles. Yet, as it doesn't do format conversion, the result
is a mess.
I suspect that it was done this way in order to save CPU time,
at the time of OMAP2 days.
That's messy and doesn't really help if someone wants to
use vim2m to test a pipeline.
Worse than that, the unique RGB format it says it supports is
RGB565BE, with is not supported by Gstreamer. That prevents
practical usage of it, even for tests.
So, instead, properly implement fourcc format conversions,
adding a few more RGB formats:
- RGB and BGR with 24 bits
- RGB565LE (known as RGB16 at gstreamer)
Also allows using any of the 5 supported formats as either
capture or output.
Note: The YUYV conversion routines are based on the conversion code
written by Hans de Goede inside libv4lconvert (part of v4l-utils),
released under LGPGL 2.1 (GPL 2.0 compatible).
Tested all possible format combinations except for RGB565BE,
as Gstreamer currently doesn't support it.
Hans Verkuil [Fri, 25 Jan 2019 15:40:04 +0000 (13:40 -0200)]
media: vivid: fix vid_out_buf_prepare()
The wrong size check was performed for output formats like NV24 which
set vfmt->buffers to 1, but vfmt->planes is 2. It was incorrectly
checking the payload size for plane 1, which doesn't exist.
Note: vfmt->buffers refers to the number of per-plane-buffers that
should be allocated. vfmt->planes refers to the number of planes
that make up an image. vfmt->planes may be > vfmt->buffers.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Hans Verkuil [Wed, 16 Jan 2019 12:01:16 +0000 (10:01 -0200)]
media: cedrus: add buf_out_validate callback
Validate the field for an output buffer. This ensures that the
field is validated when the buffer is queued to a request, and
not when the request itself is queued, which is too late.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Hans Verkuil [Wed, 16 Jan 2019 12:01:15 +0000 (10:01 -0200)]
media: vivid: add buf_out_validate callback
Validate the field for an output buffer. This ensures that the
field is validated when the buffer is queued to a request, and
not when the request itself is queued, which is too late.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Hans Verkuil [Wed, 16 Jan 2019 12:01:14 +0000 (10:01 -0200)]
media: vim2m: add buf_out_validate callback
Validate the field for an output buffer. This ensures that the
field is validated when the buffer is queued to a request, and
not when the request itself is queued, which is too late.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Hans Verkuil [Wed, 16 Jan 2019 12:01:13 +0000 (10:01 -0200)]
media: vb2: add buf_out_validate callback
When queueing a buffer to a request the 'field' value is not validated.
That field is only validated when the _buf_prepare() is called,
which happens when the request is queued.
However, this validation should happen at QBUF time, since you want
to know about this as soon as possible. Also, the spec requires that
the 'field' value is validated at QBUF time.
This patch adds a new buf_out_validate callback to validate the
output buffer at buf_prepare time or when QBUF queues an unprepared
buffer to a request. This callback is mandatory for output queues
that support requests.
This issue was found by v4l2-compliance since it failed to replace
V4L2_FIELD_ANY by a proper field value when testing the vivid video
output in combination with requests.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Pawel Osciak [Thu, 24 Jan 2019 09:51:55 +0000 (07:51 -0200)]
media: vb2: Keep dma-buf buffers mapped until they are freed
When using vb2 for video decoding, dequeued capture buffers may still
be accessed by the hardware: this is the case when they are used as
reference frames for decoding subsequent frames.
When the buffer is imported with dma-buf, it needs to be mapped before
access. Until now, it was mapped when queuing and unmapped when
dequeuing, which doesn't work for access as a reference frames.
One way to solve this would be to map the buffer again when it is
needed as a reference, but the mapping/unmapping operations can
seriously impact performance. As a result, map the buffer once (when it
is first needed when queued) and keep it mapped until it is freed.
Reviewed-on: https://chromium-review.googlesource.com/334103
[Paul: Updated for mainline and changed commit message]
Signed-off-by: Pawel Osciak <posciak@chromium.org> Signed-off-by: Paul Kocialkowski <paul.kocialkowski@bootlin.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
The vb2_find_timestamp helper was modified to allow finding buffers
regardless of their current state in the queue. This means that we
no longer have to take particular care of references to the current
capture buffer.
Signed-off-by: Paul Kocialkowski <paul.kocialkowski@bootlin.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Yunfei Dong [Thu, 17 Jan 2019 05:39:20 +0000 (03:39 -0200)]
media: mtk-vcodec: Using common interface to manage vdec/venc clock
Vdec: Using standard CCF interface to set parent clock and
clock rate in dtsi and using common interface to open/close
video decoder clock.
Venc: Using standard CCF interface to set parent clock in dtsi
and using common interface to open/close video encoder clock.
Signed-off-by: Yunfei Dong <yunfei.dong@mediatek.com> Signed-off-by: Qianqian Yan <qianqian.yan@mediatek.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Steve Longerbeam [Mon, 21 Jan 2019 23:35:52 +0000 (21:35 -0200)]
media: imx: prpencvf: Stop upstream before disabling IDMA channel
Upstream must be stopped immediately after receiving the last EOF and
before disabling the IDMA channel. This can be accomplished by moving
upstream stream off to just after receiving the last EOF completion in
prp_stop(). For symmetry also move upstream stream on to end of
prp_start().
This fixes a complete system hard lockup on the SabreAuto when streaming
from the ADV7180, by repeatedly sending a stream off immediately followed
by stream on:
while true; do v4l2-ctl -d1 --stream-mmap --stream-count=3; done
Eventually this either causes the system lockup or EOF timeouts at all
subsequent stream on, until a system reset.
The lockup occurs when disabling the IDMA channel at stream off. Stopping
the video data stream entering the IDMA channel before disabling the
channel itself appears to be a reliable fix for the hard lockup.
Fixes: f0d9c8924e2c3 ("[media] media: imx: Add IC subdev drivers") Reported-by: Gaël PORTAY <gael.portay@collabora.com> Tested-by: Gaël PORTAY <gael.portay@collabora.com> Signed-off-by: Steve Longerbeam <slongerbeam@gmail.com> Cc: stable@vger.kernel.org # for 4.13 and up Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Steve Longerbeam [Mon, 21 Jan 2019 23:35:51 +0000 (21:35 -0200)]
media: imx: csi: Stop upstream before disabling IDMA channel
Move upstream stream off to just after receiving the last EOF completion
and disabling the CSI (and thus before disabling the IDMA channel) in
csi_stop(). For symmetry also move upstream stream on to beginning of
csi_start().
Doing this makes csi_s_stream() more symmetric with prp_s_stream() which
will require the same change to fix a hard lockup.
Signed-off-by: Steve Longerbeam <slongerbeam@gmail.com> Cc: stable@vger.kernel.org # for 4.13 and up Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Steve Longerbeam [Mon, 21 Jan 2019 23:35:50 +0000 (21:35 -0200)]
media: imx: csi: Disable CSI immediately after last EOF
Disable the CSI immediately after receiving the last EOF before stream
off (and thus before disabling the IDMA channel). Do this by moving the
wait for EOF completion into a new function csi_idmac_wait_last_eof().
This fixes a complete system hard lockup on the SabreAuto when streaming
from the ADV7180, by repeatedly sending a stream off immediately followed
by stream on:
while true; do v4l2-ctl -d4 --stream-mmap --stream-count=3; done
Eventually this either causes the system lockup or EOF timeouts at all
subsequent stream on, until a system reset.
The lockup occurs when disabling the IDMA channel at stream off. Disabling
the CSI before disabling the IDMA channel appears to be a reliable fix for
the hard lockup.
Fixes: 4a34ec8e470cb ("[media] media: imx: Add CSI subdev driver") Reported-by: Gaël PORTAY <gael.portay@collabora.com> Signed-off-by: Steve Longerbeam <slongerbeam@gmail.com> Cc: stable@vger.kernel.org # for 4.13 and up Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
drivers/media/platform/vicodec/vicodec-core.c:drivers/media/platform/vicodec/vicodec-core.c:210:23: warning: symbol 'get_next_header' was not declared. Should it be static?
drivers/media/platform/vicodec/vicodec-core.c:210:23: warning: no previous prototype for 'get_next_header' [-Wmissing-prototypes]
enum vb2_buffer_state get_next_header(struct vicodec_ctx *ctx, u8 **pp, u32 sz)
^~~~~~~~~~~~~~~
Dafna Hirschfeld [Mon, 21 Jan 2019 11:46:18 +0000 (09:46 -0200)]
media: vicodec: Add support for resolution change event.
If the the queues are not streaming then the first resolution
change is handled in the buf_queue callback.
The following resolution change events are handled in job_ready.
Dafna Hirschfeld [Mon, 21 Jan 2019 11:46:14 +0000 (09:46 -0200)]
media: vicodec: add support for CROP and COMPOSE selection
Add support for the selection api for the crop and compose targets.
The driver rounds up the coded width and height such that
all planes dimensions are multiple of 8.
The "Codec Interface" chapter is poorly named since codecs are just one
use-case of the Memory-to-Memory Interface. Rename it and clean up the
text a bit.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Reviewed-by: Tomasz Figa <tfiga@chromium.org> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
media: venus: firmware: check fw size against DT memory region size
By historical reasons we defined firmware memory size to be 6MB even
that the firmware size for all supported Venus versions is 5MBs. Correct
that by compare the required firmware size returned from mdt loader and
the one provided by DT reserved memory region. We proceed further if the
required firmware size is smaller than provided by DT memory region.
Akinobu Mita [Tue, 22 Jan 2019 13:42:59 +0000 (11:42 -0200)]
media: ov2640: fix initial try format
Set initial try format with default configuration instead of current one.
Fixes: 8d3b307a150a ("media: ov2640: make VIDIOC_SUBDEV_G_FMT ioctl work with V4L2_SUBDEV_FORMAT_TRY") Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Akinobu Mita [Tue, 8 Jan 2019 14:51:44 +0000 (12:51 -0200)]
media: mt9m001: switch s_power callback to runtime PM
Switch s_power() callback to runtime PM framework. This also removes
soc_camera specific power management code and introduces reset and standby
gpios instead.
Cc: Guennadi Liakhovetski <g.liakhovetski@gmx.de> Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Ben Kao [Tue, 15 Jan 2019 09:36:07 +0000 (07:36 -0200)]
media: ov8856: Modify ov8856 register reading function to be simplified
We use put_unaligned_be16() to be simplified for setting register address
in ov8856_read_reg() and use sizeof() to be better suited for bytes
copying.
Signed-off-by: Ben Kao <ben.kao@intel.com> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Lubomir Rintel [Tue, 15 Jan 2019 08:54:46 +0000 (06:54 -0200)]
media: ov7670: control clock along with power
This provides more power saving when the sensor is off.
While at that, do the delay on power/clock enable even if the sensor driver
itself doesn't control the GPIOs. This is required for the OLPC XO-1
platform, that lacks the proper power/reset properties in its DT, but
needs the delay after the sensor is clocked up.
Lubomir Rintel [Tue, 15 Jan 2019 08:54:45 +0000 (06:54 -0200)]
media: ov7670: hook s_power onto v4l2 core
The commit 71862f63f351 ("media: ov7670: Add the ov7670_s_power function")
added a power control routing. However, it was not good enough to use as
a s_power() callback: it merely flipped on the power GPIOs without
restoring the register settings.
Fix this now and register an actual power callback.
Sakari Ailus [Tue, 15 Jan 2019 08:54:44 +0000 (06:54 -0200)]
media: ov7670: Remove useless use of a ret variable
Instead of assigning the return value to ret and then checking and
returning it, just return the value to the caller directly. The success
value is always 0.
Ben Kao [Tue, 15 Jan 2019 08:30:29 +0000 (06:30 -0200)]
media: ov8856: Add support for OV8856 sensor
This patch adds driver for Omnivision's ov8856 sensor,
the driver supports following features:
- manual exposure/gain(analog and digital) control support
- two link frequencies
- VBLANK/HBLANK support
- test pattern support
- media controller support
- runtime PM support
- enable Vsync signal output
- supported resolutions
+ 3280x2464 at 30FPS
+ 1640x1232 at 30FPS
Signed-off-by: Ben Kao <ben.kao@intel.com> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Steve Longerbeam [Thu, 17 Jan 2019 20:58:37 +0000 (18:58 -0200)]
media: imx-csi: Input connections to CSI should be optional
Some imx platforms do not have fwnode connections to all CSI input
ports, and should not be treated as an error. This includes the
imx6q SabreAuto, which has no connections to ipu1_csi1 and ipu2_csi0.
Return -ENOTCONN in imx_csi_parse_endpoint() so that v4l2-fwnode
endpoint parsing will not treat an unconnected CSI input port as
an error.
Fixes: c893500a16baf ("media: imx: csi: Register a subdev notifier") Signed-off-by: Steve Longerbeam <slongerbeam@gmail.com> Reviewed-by: Philipp Zabel <p.zabel@pengutronix.de> Acked-by: Tim Harvey <tharvey@gateworks.com> Cc: stable@vger.kernel.org Tested-by: Fabio Estevam <festevam@gmail.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Implement the vidioc_(un)subscribe_event operations. Imx will allow
subscribing to the imx-specific frame interval error events, events
from subdevices (V4L2_EVENT_SOURCE_CHANGE), and control events.
Signed-off-by: Steve Longerbeam <slongerbeam@gmail.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Steve Longerbeam [Tue, 15 Jan 2019 19:15:46 +0000 (17:15 -0200)]
media: imx: queue subdev events to reachable video devices
In order to receive events generated by subdevices on the video capture
nodes, those events need to be forwarded to the subdevice's list of
reachable video capture devices.
Note this will queue the event to a video device even if there is
no actual _enabled_ media path from the sub-device to the video device.
So a future improvement is to skip the video device if there is no enabled
path to it from the sub-device. The entity->pipe pointer can't be
used for this check because in imx-media a sub-device can be a
member to more than one streaming pipeline at a time.
Signed-off-by: Steve Longerbeam <slongerbeam@gmail.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Philipp Zabel [Thu, 17 Jan 2019 15:51:54 +0000 (13:51 -0200)]
media: imx: lift CSI and PRP ENC/VF width alignment restriction
The CSI, PRP ENC, and PRP VF subdevices shouldn't have to care about
IDMAC line start address alignment. With compose rectangle support in
the capture driver, they don't have to anymore.
If the direct CSI -> IC path is enabled, the CSI output width must
still be aligned to 8 pixels (IC burst length).
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de> Reviewed-by: Steve Longerbeam <slongerbeam@gmail.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Philipp Zabel [Tue, 8 Jan 2019 17:20:43 +0000 (15:20 -0200)]
media: coda: fix decoder capture buffer payload
It is not correct to calculate decoder capture payload dynamically from
the decoded frame width and height reported by the firmware. These tell
us what the decoder wrote into the internal framebuffers. The rotator or
VDOA always write the full sizeimage when copying the previously decoded
frame from the internal framebuffers into the capture queue.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Philipp Zabel [Tue, 8 Jan 2019 17:20:16 +0000 (15:20 -0200)]
media: coda: use macroblock tiling on CODA960 only
Coda7541 and earlier do not support macroblock tiling. They do support
the NV12 format, though. Enable macroblock tiling for NV12 only on
CODA960. This fixes crashes when trying to use NV12 support on CodaHx4.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
media: cedrus: Allow using the current dst buffer as reference
It was reported that some cases of interleaved video decoding require
using the current destination buffer as a reference. However, this is
no longer possible after the move to vb2_find_timestamp because only
dequeued and done buffers are considered.
Add a helper in our driver that also considers the current destination
buffer before resorting to vb2_find_timestamp and use it in MPEG-2.
Signed-off-by: Paul Kocialkowski <paul.kocialkowski@bootlin.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Niklas Söderlund [Fri, 14 Dec 2018 06:18:21 +0000 (04:18 -0200)]
media: rcar-vin: fix wrong return value in rvin_set_channel_routing()
If the operation in rvin_set_channel_routing() is successful the 'ret'
variable contains the runtime PM use count for the VIN master device.
The intention is not to return the use count to the caller but to return
0 on success else none zero.
Fix this by always returning 0 if the operation is successful.
Yangtao Li [Wed, 12 Dec 2018 16:27:03 +0000 (14:27 -0200)]
media: platform: sti: remove bdisp_dbg_declare() and hva_dbg_declare()
We already have the DEFINE_SHOW_ATTRIBUTE. There is no need to define
bdisp_dbg_declare and hva_dbg_declare, so remove them. Also use
DEFINE_SHOW_ATTRIBUTE to simplify some code.
Signed-off-by: Yangtao Li <tiny.windzz@gmail.com> Reviewed-by: Fabien Dessenne <fabien.dessenne@st.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Matt Ranostay [Tue, 11 Dec 2018 15:17:01 +0000 (13:17 -0200)]
media: video-i2c: add Melexis MLX90640 thermal camera
Add initial support for MLX90640 thermal cameras which output an 32x24
greyscale pixel image along with 2 rows of coefficent data.
Because of this the data outputed is really 32x26 and needs the two rows
removed after using the coefficent information to generate processed
images in userspace.
Signed-off-by: Matt Ranostay <matt.ranostay@konsulko.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
media: ivtv: add parameter to enable ivtvfb on x86 PAT systems
ivtvfb was previously disabled for x86 PAT-enabled systems
by commit 1bf1735b4780 ("x86/mm/pat, drivers/media/ivtv:
Use arch_phys_wc_add() and require PAT disabled") as a
workaround to abstract MTRR code away from device drivers.
The driver is not easily upgradable to the PAT-aware
ioremap_wc() API since the firmware hides the address
ranges that should be marked write-combined from the driver.
However, since a write-combined cache on the framebuffer
is only a performance enhancement not a requirement for
the framebuffer to function, completely disabling the driver
in this configuration is not necessary.
Add force_pat module parameter and a corresponding kernel
configuration parameter to optionally force initialization
on PAT-enabled x86 systems with a warning about the lack of
write-combined caching, and document the reasons the driver
cannot be easily updated to support wc caching on all systems.
Signed-off-by: Nick French <naf@ou.edu> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
[hverkuil-cisco@xs4all.nl: fix typo, split long pr_ lines up] Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Hans Verkuil [Thu, 17 Jan 2019 13:47:55 +0000 (11:47 -0200)]
media: soc_camera_platform: remove obsolete soc_camera test driver
This is a test stub driver for soc_camera. Since soc_camera is
being deprecated (and in fact, nobody is using it anymore)
there's no sense in keeping this test driver.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
media: imx: Allow interweave with top/bottom lines swapped
Allow sequential->interlaced interweaving but with top/bottom
lines swapped to the output buffer.
This can be accomplished by adding one line length to IDMAC output
channel address, with a negative line length for the interlace offset.
This is to allow the seq-bt -> interlaced-bt transformation, where
bottom lines are still dominant (older in time) but with top lines
first in the interweaved output buffer.
With this support, the CSI can now allow seq-bt at its source pads,
e.g. the following transformations are allowed in CSI from sink to
source:
media: imx-csi: Move crop/compose reset after filling default mbus fields
If caller passes un-initialized field type V4L2_FIELD_ANY to CSI
sink pad, the reset CSI crop window would not be correct, because
the crop window depends on a valid input field type. To fix move
the reset of crop and compose windows to after the call to
imx_media_fill_default_mbus_fields().
Signed-off-by: Steve Longerbeam <slongerbeam@gmail.com> Reviewed-by: Philipp Zabel <p.zabel@pengutronix.de> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
media: imx: vdic: rely on VDIC for correct field order
prepare_vdi_in_buffers() was setting up the dma pointers as if the
VDIC is always programmed to receive the fields in bottom-top order,
i.e. as if ipu_vdi_set_field_order() only programs BT order in the VDIC.
But that's not true, ipu_vdi_set_field_order() is working correctly.
So fix prepare_vdi_in_buffers() to give the VDIC the fields in whatever
order they were received by the video source, and rely on the VDIC to
sort out which is top and which is bottom.
Signed-off-by: Steve Longerbeam <slongerbeam@gmail.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
media: imx: interweave and odd-chroma-row skip are incompatible
If IDMAC interweaving is enabled in a write channel, the channel must
write the odd chroma rows for 4:2:0 formats. Skipping writing the odd
chroma rows produces corrupted captured 4:2:0 images when interweave
is enabled.
Reported-by: Krzysztof Ha?asa <khalasa@piap.pl> Signed-off-by: Steve Longerbeam <slongerbeam@gmail.com> Reviewed-by: Philipp Zabel <p.zabel@pengutronix.de> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
media: imx-csi: Double crop height for alternate fields at sink
If the incoming sink field type is alternate, the reset crop height
and crop height bounds must be set to twice the incoming height,
because in alternate field mode, upstream will report only the
lines for a single field, and the CSI captures the whole frame.
Signed-off-by: Steve Longerbeam <slongerbeam@gmail.com> Reviewed-by: Philipp Zabel <p.zabel@pengutronix.de> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
projects / linux.git / log