- 29 May, 2019 16 commits
-
-
Hans Verkuil authored
In vb2_vmalloc_get_userptr() the framevector is created with the 'write' argument set to false when vb2_create_framevec() is called for OUTPUT buffers. So the pages are marked as read-only. However, userspace will write to these buffers since it will fill in the data to output. Since get_userptr is only called if the userptr of the queued buffer has changed since the last time that same buffer was queued, this will fail when the buffer contents is updated and the buffer is queued again. E.g., userspace fills buffer 1 with the output video and queues it. The first time get_userptr is called and the pages are grabbed and pinned in memory and marked read-only. The second time buffer 1 is filled with different video data and queued again. Since the userptr hasn't changed the get_userptr() callback isn't called again. Since the pages were marked as read-only the new contents isn't updated. Just always call vb2_create_framevec() with FOLL_WRITE to always allow writing to the buffers. Using USERPTR streaming with OUTPUT devices is almost never done. And when it is done it is via v4l2-compliance and a driver like vim2m. But since v4l2-compliance doesn't actually inspect the capture buffer and compare it to the original output buffer, this issue was never noticed. But the vicodec driver actually needs to parse the bitstream in the OUTPUT buffers and any errors there will be immediately noticed. So this time v4l2-compliance failed the USERPTR streaming test. Signed-off-by:
Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by:
Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
-
Michael Tretter authored
The allegro hardware encoder does not write SPS/PPS nal units into the encoded video stream. Therefore, we need to write the units in software. The implementation follows Rec. ITU-T H.264 (04/2017) to allow to convert between a C struct and the RBSP representation of the SPS and PPS nal units. The allegro driver writes the nal units into the v4l2 capture buffer in front of the actual video data which is written at an offset by the IP core. The remaining gap is filled with a filler nal unit. [hverkuil-cisco@xs4all.nl: added missing @dev description for nal_h264_read_sps()] [mchehab+samsung@kernel.org: removed uneeded "-ccflags-y $(src)" from Makefile] Signed-off-by:
Michael Tretter <m.tretter@pengutronix.de> Signed-off-by:
-
Michael Tretter authored
Add a V4L2 mem-to-mem driver for Allegro DVT video IP cores as found in the EV family of the Xilinx ZynqMP SoC. The Zynq UltraScale+ Device Technical Reference Manual uses the term VCU (Video Codec Unit) for the encoder, decoder and system integration block. This driver takes care of interacting with the MicroBlaze MCU that controls the actual IP cores. The IP cores and MCU are integrated in the FPGA. The xlnx_vcu driver is responsible for configuring the clocks and providing information about the codec configuration. The driver currently only supports the H.264 video encoder. Signed-off-by:
-
Michael Tretter authored
Add vendor prefix for Allegro DVT, a provider of H.264/AVC, H.265/HEVC, AVS2, VP9 and AV1 compliance test suites and H.264/AVC, H.265/HEVC, and VP9 encoder, codec and decoder hardware (RTL) IPs. Signed-off-by:
Rob Herring <robh@kernel.org> Signed-off-by:
-
Michael Tretter authored
Add device-tree bindings for the Allegro DVT video IP core found on the Xilinx ZynqMP EV family. Signed-off-by:
-
Hans Verkuil authored
last_buffer_dequeued was set to true in __fill_v4l2_buffer, but this is called for qbuf as well. Move it to vb2_dqbuf. Signed-off-by:
-
Maxime Ripard authored
Introduce some basic H264 decoding support in cedrus. So far, only the baseline profile videos have been tested, and some more advanced features used in higher profiles are not even implemented. Reviewed-by:
Jernej Skrabec <jernej.skrabec@siol.net> Reviewed-by:
Paul Kocialkowski <paul.kocialkowski@bootlin.com> Signed-off-by:
Maxime Ripard <maxime.ripard@bootlin.com> Signed-off-by:
-
Maxime Ripard authored
The H264_SLICE_RAW format introduced before is meant for stateless decoders that will need the H264 parsed slice data without the start code. Let's document it. Signed-off-by:
-
Maxime Ripard authored
The H264_SLICE_RAW format is meant to hold the parsed slice data without the start code. This will be needed by stateless decoders. Signed-off-by:
-
Pawel Osciak authored
Stateless video codecs will require both the H264 metadata and slices in order to be able to decode frames. This introduces the definitions for the structures used to pass the metadata from the userspace to the kernel. [hverkuil-cisco@xs4all.nl: add space after . in ".For"] [hverkuil-cisco@xs4all.nl: sync v4l2_ctrl_h264_decode_params struct layout with header] Co-developed-by:
Tomasz Figa <tfiga@chromium.org> Signed-off-by:
Pawel Osciak <posciak@chromium.org> Signed-off-by:
Guenter Roeck <groeck@chromium.org> Signed-off-by:
-
Jernej Skrabec authored
Some VPU variants may run at higher clock speeds. They actually need extra speed to be capable of decoding more complex codecs like HEVC or bigger image sizes (4K). Expand variant structure with mod_rate information. Signed-off-by:
-
Philipp Zabel authored
Use v4l2_m2m2_buf_copy_metadata to let BIT encoder contexts copy buffer field, timestamp, timestamp flags, and optionally timecode. Signed-off-by:
Philipp Zabel <p.zabel@pengutronix.de> Signed-off-by:
-
Philipp Zabel authored
coda_command_sync, coda_hw_reset, and __coda_start_decoding all expect to be called under the coda_mutex device lock. Signed-off-by:
-
Philipp Zabel authored
The MPEG-2 decoder firmware reports profile and level indication that can be used to set V4L2 MPEG-2 profile and level controls Signed-off-by:
-
Philipp Zabel authored
Add MPEG-2 CID definitions for profiles and levels defined in ITU-T Rec. H.262. Signed-off-by:
-
Philipp Zabel authored
The MPEG-4 decoder firmware reports profile and level indication values that can be used to update V4L2 MPEG-4 profile and level controls. Signed-off-by:
-
- 28 May, 2019 24 commits
-
-
Maciej S. Szmigiero authored
This adds raw (unprocessed) BT.656 stream capturing support for the analog part of Medion 95700. It can be enabled by setting CXUSB_EXTENDEDMODE_CAPTURE_RAW flag in parm.capture.extendedmode passed to VIDIOC_S_PARM. Signed-off-by:
Maciej S. Szmigiero <mail@maciej.szmigiero.name> Signed-off-by:
-
Maciej S. Szmigiero authored
This patch adds support for analog part of Medion 95700 in the cxusb driver. What works: * Video capture at various sizes with sequential fields, * Input switching (TV Tuner, Composite, S-Video), * TV and radio tuning, * Video standard switching and auto detection, * Radio mode switching (stereo / mono), * Unplugging while capturing, * DVB / analog coexistence. What does not work yet: * Audio, * VBI, * Picture controls. Signed-off-by:
-
Maciej S. Szmigiero authored
This patch prepares cxusb driver for supporting the analog part of Medion 95700 (previously only the digital - DVB - mode was supported). Specifically, it adds support for: * switching the device between analog and digital modes of operation, * enforcing that only one mode is active at the same time due to hardware limitations. Actual implementation of the analog mode will be provided by the next commit. Signed-off-by:
-
Maciej S. Szmigiero authored
According to V4L2 API set_fmt subdev operation should not return an error on out-of-range picture sizes, the values should be clamped instead to the supported range. The cx25840 datasheet says that the chip is capable of scaling down the picture width and height, respectively, 16 and 8 times. These values agree with what the old implementation enforced. Signed-off-by:
-
Maciej S. Szmigiero authored
This commit adds pin to pad mapping and output format configuration support in CX2584x-series chips to cx25840 driver. This functionality is then used to allow disabling ivtv-specific hacks and configuration values (called a "generic mode"), so cx25840 driver can be used for other devices not needing them without risking compatibility problems. Signed-off-by:
-
Maciej S. Szmigiero authored
This commit implements g_std operation in cx25840 driver by returning the last set video standard. Signed-off-by:
-
Maciej S. Szmigiero authored
cx25840 driver g_std operation queries the currently detected video signal, however this is what querystd operation should do, so let's rename the handler. None of the existing cx25840 driver users ever called the g_std operation, one of them calls querystd on each of its subdevs but then the result is only used to implement VIDIOC_QUERYSTD (as it should). Signed-off-by:
-
Maciej S. Szmigiero authored
cx25840_load_fw() does the same thing as cx25840_reset(), only keeps "is_initialized" flag so any further invocation of this function besides the first one is a NOP. Let's just call cx25840_reset() directly from cx25840_load_fw() instead of open coding it there. While we are at it, let's also improve comments about cx25840_load_fw() so they are current and in the proper style (one of them even referred to a non-existing cx25840 init operation). Signed-off-by:
-
Fabio Estevam authored
imx7_csi_enable() always return 0 and its return value is never checked, so convert it to void. Signed-off-by:
Fabio Estevam <festevam@gmail.com> Tested-by:
Rui Miguel Silva <rmfrfs@gmail.com> Reviewed-by:
-
Fabio Estevam authored
clk_prepare_enable() may fail, so we should better check its return value and propagate it in the case of error. Signed-off-by:
-
Fabio Estevam authored
Remove unneeded 'break' right after the 'return' statement as pointed out by checkpatch. Signed-off-by:
-
Fabio Estevam authored
In the case of devm_request_irq() failure, let's propagate the real error code instead of a fake one. Signed-off-by:
-
Fabio Estevam authored
In the case of platform_get_irq() failure, let's propagate the real error code instead of a fake one. Signed-off-by:
-
Fabio Estevam authored
Use devm_platform_ioremap_resource() to simplify the code a bit. While at it, propagate the real error value in case of devm_platform_ioremap_resource() failure. Signed-off-by:
-
Fabio Estevam authored
As per the i.MX7D Reference Manual only the MCLK is used for the CSI block, so only handle this single clock. Signed-off-by:
-
Steve Longerbeam authored
The media device is already available via multiple methods, there is no need to set driver data for v4l2_dev to the media device. In imx_media_link_notify(), get media device from link->graph_obj.mdev. In imx_media_capture_device_register(), get media device from v4l2_dev->mdev. Signed-off-by:
Steve Longerbeam <slongerbeam@gmail.com> Signed-off-by:
-
Steve Longerbeam authored
Export find_pipeline_pad(), renaming to imx_media_pipeline_pad(), and extend its functionality to allow searching for video devices in the enabled pipeline in addition to sub-devices. As part of this: - Rename imx_media_find_mipi_csi2_channel() to imx_media_pipeline_csi2_channel(). - Remove imx_media_find_upstream_pad(), it is redundant now. - Rename imx_media_find_upstream_subdev() to imx_media_pipeline_subdev() with an additional boolean argument for searching upstream or downstream. - Add imx_media_pipeline_video_device() which is analogous to imx_media_pipeline_subdev() but searches for video devices. - Remove imxmd pointer arg from all of the functions above, it was never used in those functions. With that change the i.MX5/6 CSI, VDIC, and IC sub-devices no longer require the media_device. Signed-off-by:
-
Steve Longerbeam authored
Re-organize modules, and which objects are linked into those modules, so that: - imx6-media (renamed from imx-media) is the media driver module for imx5/6 only, and has no symbol exports. - imx6-media-csi (renamed from imx-media-csi) is the subdev driver module for imx5/6 CSI. It is now linked direcly with imx-media-fim, since only the imx5/6 CSI makes use of the frame interval monitor. - imx-media-common now only contains common code between imx5/6 and imx7 media drivers. It contains imx-media-utils, imx-media-of, imx-media-dev-common, and imx-media-capture. In order to acheive that, some functions common to imx5/6 and imx7 have been moved out of imx-media-dev.c and into imx-media-dev-common.c. Signed-off-by:
-
Steve Longerbeam authored
Don't propagate the source pad format to the connected capture device. It's now the responsibility of userspace to call VIDIOC_S_FMT on the capture device to ensure the capture format and compose rectangle are compatible with the connected source. To check this, validate the capture format with the source before streaming starts. Signed-off-by:
-
Steve Longerbeam authored
Revert this commit, as imx_media_capture_device_set_format() will be removed. The arguments to mx_media_mbus_fmt_to_pix_fmt() and imx_media_capture_device_set_format() in imx7_csi_set_fmt() are also reverted. This reverts commit 5964cbd8. Signed-off-by:
-
Steve Longerbeam authored
Move imx_media_add_video_device() into imx_media_capture_device_register(). Also the former has no error conditions to convert to void. Signed-off-by:
-
Steve Longerbeam authored
Allocate and free a DMA coherent buffer in imx_media_alloc/free_dma_buf() from the given device. This allows DMA alloc and free using a device that is backed by real hardware, which for the imx5/6/7 CSI is the CSI unit, and for the internal IPU sub-devices, is the parent IPU. Signed-off-by:
-
Steve Longerbeam authored
Because the IPU sub-devices VDIC and IC are not present in the device-tree, platform devices were created for them instead. This allowed these sub-devices to be added to the media device's async notifier and registered asynchronously along with the other sub-devices that do have a device-tree presence (CSI and devices external to the IPU and SoC). But that approach isn't really necessary. The IPU sub-devices don't actually require a backing device (sd->dev is allowed to be NULL). And that approach can't get around the fact that the IPU sub-devices are not part of a device hierarchy, which makes it awkward to retrieve the parent IPU of these devices. By registering them synchronously, they can be registered from the CSI async bound notifier, so the init function for them can be given the CSI subdev, who's dev->parent is the IPU. That is a somewhat cleaner way to retrieve the parent IPU. So convert to synchronous registration for the VDIC and IC task sub-devices, at the time a CSI sub-device is bound. There is no longer a backing device for them (sd->dev is NULL), but that's ok. Also set the VDIC/IC sub-device owner as the IPU, so that a reference can be taken on the IPU module. Since the VDIC and IC task drivers are no longer platform drivers, they are now statically linked to imx-media module. Signed-off-by:
-
Steve Longerbeam authored
The imx6-specific subdevs that register a capture device will no longer hold a reference to the media device, so this commit must be reverted. This reverts commit 16204b8a. Signed-off-by:
-
