Currently, the Imagination E5010 JPEG Encoder is only present on Texas
Instruments K3 SoCs. Hence add a dependency on ARCH_K3, to prevent
asking the user about this driver when configuring a kernel without TI
K3 SoC support. The dependency can be relaxed if/when the encoder
appears on other SoC families.
Fixes: a1e2940458 ("media: imagination: Add E5010 JPEG Encoder driver")
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Sebastian Fricke <sebastian.fricke@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
RK3588 handling is exactly the same as RK3568. This is not
handled using fallback compatibles to avoid exposing multiple
video devices on kernels not having the multicore disable
patch.
Signed-off-by: Sebastian Reichel <sebastian.reichel@collabora.com>
Signed-off-by: Sebastian Fricke <sebastian.fricke@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Avoid exposing equal Hantro video codecs to userspace. Equal video
codecs allow scheduling work between the cores. For that kernel support
is required, which does not yet exist. Until that is implemented avoid
exposing each core separately to userspace so that multicore can be
added in the future without breaking userspace ABI.
This was written with Rockchip RK3588 in mind (which has 4 Hantro H1
cores), but applies to all SoCs.
Signed-off-by: Sebastian Reichel <sebastian.reichel@collabora.com>
Signed-off-by: Sebastian Fricke <sebastian.fricke@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
RK3588 has four Hantro H1 VEPUs (encoder-only) modules and one combined
Hantro H1/G1 VPU (decoder and encoder). These are not described as
separate IP, since they are sharing an internal cache. This adds the
RK3588 specific compatible string for the combined VPU, which seems to
be identical to the version found in the RK3568.
Signed-off-by: Jianfeng Liu <liujianfeng1994@gmail.com>
Acked-by: Conor Dooley <conor.dooley@microchip.com>
Signed-off-by: Sebastian Reichel <sebastian.reichel@collabora.com>
Signed-off-by: Sebastian Fricke <sebastian.fricke@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
This encoder-only device is present four times on this SoC, and should
support everything the rk3568 vepu supports (so JPEG, H.264 and VP8
encoding). No fallback compatible has been added, since the operating
systems might already support RK3568 VEPU and want to avoid registering
four of them separately considering they can be used as a cluster.
Signed-off-by: Emmanuel Gil Peyrot <linkmauve@linkmauve.fr>
Acked-by: Conor Dooley <conor.dooley@microchip.com>
Signed-off-by: Sebastian Reichel <sebastian.reichel@collabora.com>
Signed-off-by: Sebastian Fricke <sebastian.fricke@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
- v4l2_av1_segmentation table formatting
- incorrect member name uv_sec_strength
- incorrect struct name v4l2_av1_global_motion
- incorrect link label for v4l2_av1_loop_filter
- add max_frame_height_minus_1
- struct ordering
Signed-off-by: Fritz Koenig <frkoenig@chromium.org>
Reviewed-by: Daniel Almeida <daniel.almeida@collabora.com>
Signed-off-by: Sebastian Fricke <sebastian.fricke@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Reference frame compression is a feature added in the G2 decoder to
compress frame buffers so that the bandwidth of storing/loading
reference frames can be reduced, especially with high resolution decoded
streams.
The impact of compressed frames is confirmed when using perf to monitor
the number of memory accesses with or without the compression feature.
The following command:
perf stat -a -e \
imx8_ddr0/cycles/,imx8_ddr0/read-cycles/,imx8_ddr0/write-cycles/ \
gst-launch-1.0 filesrc \
location=Jockey_3840x2160_120fps_420_8bit_HEVC_RAW.hevc ! queue ! \
h265parse ! v4l2slh265dec ! video/x-raw,format=NV12 ! fakesink
Gives us these results without the compression feature:
Performance counter stats for 'system wide':
1711300345 imx8_ddr0/cycles/
892207924 imx8_ddr0/read-cycles/
1291785864 imx8_ddr0/write-cycles/
13.760048353 seconds time elapsed
With the compression feature:
Performance counter stats for 'system wide':
274526799 imx8_ddr0/cycles/
453120194 imx8_ddr0/read-cycles/
833391434 imx8_ddr0/write-cycles/
18.257831534 seconds time elapsed
As expected the number of read/write cycles are really lower when
compression is used.
Since storing the compression data requires more memory a module
parameter named 'hevc_use_compression' is used to enable/disable
this feature and, by default, compression isn't used.
Enabling the compression feature means that the output-frames of the
decoder
are stored with a specific compression pixel-format. Since this
pixel format is unknown, this patch restrains the compression feature
usage to the cases where post-processor pixel-formats (NV12 or NV15)
are selected by the applications.
The Fluster compliance HEVC test suite score is still 141/147 with this
patch.
Signed-off-by: Benjamin Gaignard <benjamin.gaignard@collabora.com>
Tested-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Sebastian Fricke <sebastian.fricke@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
"struct vb2_ops" are not modified in this driver.
Constifying this structure moves some data to a read-only section, so
increase overall security.
On a x86_64, with allmodconfig:
Before:
======
text data bss dec hex filename
18059 3096 16 21171 52b3 drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_stateful.o
After:
=====
text data bss dec hex filename
18171 2968 16 21155 52a3 drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_stateful.o
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: Sebastian Fricke <sebastian.fricke@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
When introducing the ability for drivers to indicate the minimum number
of buffers they require an application to allocate, commit 6662edcd32
("media: videobuf2: Add min_reqbufs_allocation field to vb2_queue
structure") also introduced a global minimum of 2 buffers. It turns out
this breaks the Renesas R-Car VSP test suite, where a test that
allocates a single buffer fails when two buffers are used.
One may consider debatable whether test suite failures without failures
in production use cases should be considered as a regression, but
operation with a single buffer is a valid use case. While full frame
rate can't be maintained, memory-to-memory devices can still be used
with a decent efficiency, and requiring applications to allocate
multiple buffers for single-shot use cases with capture devices would
just waste memory.
For those reasons, fix the regression by dropping the global minimum of
buffers. Individual drivers can still set their own minimum.
Fixes: 6662edcd32 ("media: videobuf2: Add min_reqbufs_allocation field to vb2_queue structure")
Cc: stable@vger.kernel.org
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Acked-by: Tomasz Figa <tfiga@chromium.org>
Link: https://lore.kernel.org/r/20240825232449.25905-1-laurent.pinchart+renesas@ideasonboard.com
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
The v4l2_subdev_link_validate() helper prints a warning if the
.link_validate() operation is not implemented for video devices
connected to the subdevs. Implement the operation to silence the
warning.
Ideally validation of the link between the video device and the subdev
should be implemented in that operation. That would however break
userspace that does not configure formats on all video devices before
starting streaming. While this mode of operation may not be considered
valid by the V4L2 API specification (interpretation differ), it is
nonetheless supported by the vsp1 driver at the moment and used by at
least the vsp1 unit test suite, and possibly other userspace
applications. Removing it would be a regression.
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: Tomi Valkeinen <tomi.valkeinen+renesas@ideasonboard.com>
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
The v4l2_subdev_link_validate() helper function is meant to be used as a
drop-in implementation of a V4L2 subdev entity .link_validate() handler.
It supports subdev-to-subdev links only, and complains if one end of the
link is not a subdev. This forces drivers that have video output devices
connected to subdevs to implement a custom .link_validate() handler,
calling v4l2_subdev_link_validate() for the subdev-to-subdev links, and
performing manual link validation for the video-to-subdev links.
Video devices embed a media entity, and therefore also have a
.link_validate() operation. For video capture devices, the operation
should be manually implemented by drivers for validate the
subdev-to-video links. For video output devices, on the other hand, that
operation is never called, as link validation is performed in the
context of the sink entity.
As a result, we end up forcing drivers to implement a custom
.link_validate() handler for subdevs connected to video output devices,
when the video devices provide an operation that could be used for that
purpose.
To improve that situation, make v4l2_subdev_link_validate() delegate
link validation to the source's .link_validate() operation when the link
source is a video device and the link sink is a subdev. This allows
broader usage of v4l2_subdev_link_validate(), and simplifies drivers by
making video device link validation easy to implement in the video
device .link_validate(), regardless of whether the video device is an
output device or a capture device.
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: Tomi Valkeinen <tomi.valkeinen+renesas@ideasonboard.com>
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
The v4l2_subdev_link_validate() function prints a one-time warning if it
gets called on a link whose source or sink is not a subdev. As links get
validated in the context of their sink, a call to the helper when the
link's sink is not a subdev indicates that the driver has set its
.link_validate() handler to v4l2_subdev_link_validate() on a non-subdev
entity, which is a clear driver bug. On the other hand, the link's
source not being a subdev indicates that the helper is used for a subdev
connected to a video output device, which is a lesser issue, if only
because this is currently common practice.
There are no drivers left in the kernel that use
v4l2_subdev_link_validate() in a context where it may get called on a
non-subdev sink. Replace the pr_warn_once() with a WARN_ON_ONCE() in
this case to make sure that new offenders won't be introduced.
A subsequent change will improve the v4l2_subdev_link_validate() helper
to properly support validating video device to subdev links.
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: Tomi Valkeinen <tomi.valkeinen+renesas@ideasonboard.com>
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
The v4l2_subdev_link_validate() function is a helper designed to
validate links whose sink is a subdev. When called on a link whose sink
is a video device, it only prints a warning and returns. Its usage in
the sun4i_csi driver is wrong, leaving the link from the sub4i_csi
subdev to the capture video device unvalidated.
Planned improvements to the v4l2_subdev_link_validate() function will
turn the warning into an error, breaking the sun4i_csi driver. As an
interim measure, move the warning to the sun4i_csi driver in a custom
validation handler, and drop the call to the helper.
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Acked-by: Chen-Yu Tsai <wens@csie.org>
Reviewed-by: Tomi Valkeinen <tomi.valkeinen+renesas@ideasonboard.com>
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
The v4l2_subdev_link_validate() function is a helper designed to
validate links whose sink is a subdev. When called on a link whose sink
is a video device, it only prints a warning and returns. As the
microchip-isc driver implements manual validate of the subdev to video
device link, we can just drop the v4l2_subdev_link_validate() to avoid
the warning.
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: Tomi Valkeinen <tomi.valkeinen+renesas@ideasonboard.com>
Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com>
"struct dec_bufsize_ops and "struct enc_bufsize_ops" are not modified in
this driver.
Constifying these structures moves some data to a read-only section, so
increase overall security.
On a x86_64, with allmodconfig:
Before:
text data bss dec hex filename
12494 822 0 13316 3404 drivers/media/platform/qcom/venus/hfi_plat_bufs_v6.o
After:
text data bss dec hex filename
12766 566 0 13332 3414 drivers/media/platform/qcom/venus/hfi_plat_bufs_v6.o
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Reviewed-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Signed-off-by: Stanimir Varbanov <stanimir.k.varbanov@gmail.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
in venus_probe, core->work is bound with venus_sys_error_handler, which is
used to handle error. The code use core->sys_err_done to make sync work.
The core->work is started in venus_event_notify.
If we call venus_remove, there might be an unfished work. The possible
sequence is as follows:
CPU0 CPU1
|venus_sys_error_handler
venus_remove |
hfi_destroy |
venus_hfi_destroy |
kfree(hdev); |
|hfi_reinit
|venus_hfi_queues_reinit
|//use hdev
Fix it by canceling the work in venus_remove.
Cc: stable@vger.kernel.org
Fixes: af2c3834c8 ("[media] media: venus: adding core part and helper functions")
Signed-off-by: Zheng Wang <zyytlz.wz@163.com>
Signed-off-by: Dikshita Agarwal <quic_dikshita@quicinc.com>
Signed-off-by: Stanimir Varbanov <stanimir.k.varbanov@gmail.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
There is no need to compile the Rockchip specific AV1 drivers if
CONFIG_VIDEO_HANTRO_ROCKCHIP is not set.
All the driver functions are only referenced by rockchip_vpu_hw.c which
is already under this option.
Fixes: 727a400686 ("media: verisilicon: Add Rockchip AV1 decoder")
Fixes: d8ebe59e7b ("media: verisilicon: Add film grain feature to AV1 driver")
Fixes: c0d0e579db ("media: verisilicon: Add AV1 entropy helpers")
Signed-off-by: Alexander Stein <alexander.stein@ew.tq-group.com>
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Sebastian Fricke <sebastian.fricke@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
There is no need to compile the Rockchip specific drivers if
CONFIG_VIDEO_HANTRO_ROCKCHIP is not set.
All the driver functions are only referenced by rockchip_vpu_hw.c which
is already under this option.
Fixes: c9caebd57b ("media: hantro: merge Rockchip platform drivers")
Fixes: c07665f993 ("media: hantro: Add H.264 support for Rockchip VDPU2")
Signed-off-by: Alexander Stein <alexander.stein@ew.tq-group.com>
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Sebastian Fricke <sebastian.fricke@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
If the width and height are 0xffff (or close), the result of a
multiplication will overflow.
Cast to a larger type to avoid undefined behavior.
Such values are possible in CODA7, but unlikely.
Found by Linux Verification Center (linuxtesting.org) with SVACE.
Fixes: 918c66fd41 ("[media] coda: add CODA7541 decoding support")
Signed-off-by: Anastasia Belova <abelova@astralinux.ru>
Signed-off-by: Sebastian Fricke <sebastian.fricke@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Vicodec will ignore en/decoder commands if the CAPTURE queue is not
streaming. But this prevents CMD_STOP from being used during a dynamic
resolution change to mark the last source buffer. Since CMD_STOP is
ignored, but doesn’t fail, there is no warning that the command needs to
be resent, and CAPTURE will continue to attempt to dequeue buffers
waiting, futilely, for one with V4L2_BUF_FLAG_LAST.
Fix this problem by carrying out en/decoder commands even if CAPTURE is
not streaming.
Fixes: d17589afa9 vicodec: improve handling of ENC_CMD_STOP/START
Reported by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Deborah Brouwer <deborah.brouwer@collabora.com>
Signed-off-by: Sebastian Fricke <sebastian.fricke@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
The driver has an OF match table, still, it uses an ID lookup table for
retrieving match data. Currently, the driver is working on the
assumption that an I2C device registered via OF will always match a
legacy I2C device ID. The correct approach is to have an OF device ID
table using i2c_get_match_data() if the devices are registered via OF/ID.
Unify the OF/ID table by using MEDIA_BUS_FMT as match data for both these
tables and replace the ID lookup table for the match data by
i2c_get_match_data() and simplifly probe() and mt9p031_init_state().
While at it, remove the trailing comma in the terminator entry for the OF
table making code robust against (theoretical) misrebases or other similar
things where the new entry goes _after_ the termination without the
compiler noticing.
Signed-off-by: Biju Das <biju.das.jz@bp.renesas.com>
Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Link: https://lore.kernel.org/r/20230910160126.70122-2-biju.das.jz@bp.renesas.com
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
