compute_tiles_uniform() and compute_tiles_non_uniform() loop over
num_tile_columns_minus1 + 1 / num_tile_rows_minus1 + 1 entries, and
assemble_hw_pps() writes one COLUMN_WIDTH / ROW_HEIGHT register per tile
and indexes priv_tbl->param_set[] by pic_parameter_set_id, all taken from
the untrusted PPS. Use the bounded v4l2_hevc_pps_num_tile_columns() /
v4l2_hevc_pps_num_tile_rows() helpers for the tile loops, and bail out of
assemble_hw_pps() before indexing priv_tbl->param_set[] with an
out-of-range pic_parameter_set_id, so the writes stay within the hardware
tables.
Fixes: 3595375c23 ("media: rkvdec: Add HEVC backend")
Fixes: c9a59dc2ac ("media: rkvdec: Add HEVC support for the VDPU381 variant")
Assisted-by: Claude:claude-opus-4-8
Cc: stable@vger.kernel.org
Signed-off-by: Michael Bommarito <michael.bommarito@gmail.com>
Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org>
Use DIV_ROUND_UP() when computing HEVC coding tree block counts
instead of open-coding the same rounding expression. This keeps the
rounding intent explicit without changing behavior.
Found using a Coccinelle rule generated from the DIV_ROUND_UP() macro
definition.
Signed-off-by: Hiroki Nakajima <3na7nanana@gmail.com>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org>
platform_get_irq() returns a positive IRQ number on success and a
negative error code on failure. It no longer returns zero. The driver
currently returns -ENXIO for all failures, which loses useful errors
such as -EPROBE_DEFER.
Return the error from platform_get_irq() directly.
Fixes: cd33c83044 ("media: rkvdec: Add the rkvdec driver")
Cc: stable@vger.kernel.org
Signed-off-by: Narasimharao Vadlamudi <ahmisaranrao@gmail.com>
Reviewed-by: Detlev Casanova <detlev.casanova@collabora.com>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org>
st_ref_pic_set_prediction() computes the reference RPS index as
st_rps_idx - (delta_idx_minus1 + 1) per HEVC spec equation 7-59.
Both operands are u8, so when delta_idx_minus1 + 1 exceeds the
current index the subtraction wraps and the subsequent array access
at calculated_rps_st_sets[ref_rps_idx] reads far out of bounds.
A userspace V4L2 client that can open the RKVDEC m2m decoder can
submit an EXT_SPS_ST_RPS control with INTER_REF_PIC_SET_PRED set
and delta_idx_minus1 crafted to trigger the underflow.
Reject the entry early when the reference index would underflow.
Fixes: c9a59dc2ac ("media: rkvdec: Add HEVC support for the VDPU381 variant")
Cc: stable@vger.kernel.org
Suggested-by: Detlev Casanova <detlev.casanova@collabora.com>
Assisted-by: Claude:claude-opus-4-7
Signed-off-by: Michael Bommarito <michael.bommarito@gmail.com>
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org>
The VDPU381 HEVC driver registers V4L2_CID_STATELESS_HEVC_EXT_SPS_ST_RPS
and V4L2_CID_STATELESS_HEVC_EXT_SPS_LT_RPS with .cfg.dims = { 65 }, but
the HEVC spec caps num_short_term_ref_pic_sets at 64 (ITU-T H.265 7.4.8)
and num_long_term_ref_pics_sps at 32 (7.4.3.2.1). The hardware descriptor
table (struct rkvdec_rps) sizes match those spec limits: refs[32] and
short_term_ref_sets[64].
Reduce the dims to { 64 } and { 32 } respectively so the V4L2 control
framework rejects oversized payloads before any driver code runs.
Fixes: c9a59dc2ac ("media: rkvdec: Add HEVC support for the VDPU381 variant")
Cc: stable@vger.kernel.org
Suggested-by: Detlev Casanova <detlev.casanova@collabora.com>
Assisted-by: Claude:claude-opus-4-7
Signed-off-by: Michael Bommarito <michael.bommarito@gmail.com>
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org>
Don't change the quantization of an RGB format when converting between
RGB and YUV with the RGA3. As the RGA3 only supports doing conversions
to full range YUV with BT601, it wants to announce it through
try_fmt/s_fmt.
As it is only relevant, when converting between RGB and YUV, it's
guarded by a given condition. But the condition also causes the
RGB format quantization to be adjusted to limited range, which
is incorrect.
Therefore simplify the condition to only apply when the current format
is a YUV format. Also document the reason for checking if the other
format is an RGB format.
Fixes: 24a63d4c9d ("media: rockchip: rga: add rga3 support")
Cc: stable@vger.kernel.org
Signed-off-by: Sven Püschel <s.pueschel@pengutronix.de>
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org>
The RK3588 Video Capture (VICAP) unit features a Digital Video Port (DVP)
and six MIPI CSI-2 capture interfaces. Add initial support for this variant
to the rkcif driver and enable the MIPI CSI-2 capture interfaces.
Reviewed-by: Mehdi Djait <mehdi.djait@linux.intel.com>
Signed-off-by: Michael Riesch <michael.riesch@collabora.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Add support for the RGA3 unit contained in the RK3588.
Only a basic feature set consisting of scaling and color conversion is
implemented. Currently unimplemented features include:
- Advanced formats like 10bit YUV, FBCE mode and Tile8x8 mode
- Background color (V4L2_CID_BG_COLOR)
- Configurable alpha value (V4L2_CID_ALPHA_COMPONENT)
- Image flipping (V4L2_CID_HFLIP and V4L2_CID_VFLIP)
- Image rotation (V4L2_CID_ROTATE)
- Image cropping/composing (VIDIOC_S_SELECTION)
- Only very basic output cropping for 1088 -> 1080 cases is implemented
The register address defines were copied from the
vendor Rockchip kernel sources and slightly adjusted to not start at 0
again for the cmd registers.
During testing it has been noted that the scaling of the hardware is
slightly incorrect. A test conversion of 128x128 RGBA to 256x256 RGBA
causes a slightly larger scaling. The scaling is suddle, as it seems
that the image is scaled to a 2px larger version and then cropped to
it's final size. Trying to use the RGA2 scaling factor calculation
didn't work. As the calculation matches the vendor kernel driver, no
further research has been utilized to check if there may be some kind of
better scaling factor calculation.
Furthermore comparing the RGA3 conversion with the GStreamer
videoconvertscale element, the chroma-site is different. A quick testing
didn't reveal a chroma-site that creates the same image with the
GStreamer Element. Also when converting from YUV to RGB the RGB values
differ by 1 or 2. This doesn't seem to be a colorspace conversion issue
but rather a slightly different precision on the calculation.
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Sven Püschel <s.pueschel@pengutronix.de>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org>
Disable multi-core support in preparation of the RGA3 addition. The
RK3588 SoC features two equal RGA3 cores. This allows scheduling of the
work between both cores, which is not yet implemented. Until it is
implemented avoid exposing both cores as independent video devices to
prevent an ABI breakage when multi-core support is added.
This patch is copied from the Hantro driver patch to disable multi core
support by Sebastian Reichel. See
commit ccdeb8d57f ("media: hantro: Disable multicore support")
Link: https://lore.kernel.org/all/20240618183816.77597-4-sebastian.reichel@collabora.com/
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Sven Püschel <s.pueschel@pengutronix.de>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org>
In preparation to the RGA3 addition add feature flags, which can limit
the exposed feature set of the video device, like rotating or selection
support. This is necessary as the RGA3 doesn't initially implement the
full feature set currently exposed by the driver.
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Sven Püschel <s.pueschel@pengutronix.de>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org>
Move rga_fmt to rga-hw in preparation of the RGA3 addition, as the struct
contains many RGA2 specific values. They are used to write the correct
register values quickly based on the chosen format. Therefore the
pointer to the rga_fmt struct is kept but changed to an opaque void
pointer outside of the rga-hw.h.
To enumerate and set the correct formats, two helper functions need to
be exposed in the rga_hw struct:
enum_format just get's the vidioc_enum_fmt format and it's return value
is also returned from vidioc_enum_fmt. This is a simple pass-through,
as the implementation is very simple.
adjust_and_map_format is a simple abstraction around the previous
rga_find_format. But unlike rga_find_format, it always returns a valid
format. Therefore the passed format value is also a pointer to update
it in case the values are not supported by the hardware.
Due to the RGA3 supporting different formats on the capture and output
side, an additional parameter is_capture has been added to support
this use-case. The additional ctx parameter is also added to allow the
RGA3 to limit the colorimetry only if an RGB<->YCrCb transformation
happens.
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Sven Püschel <s.pueschel@pengutronix.de>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org>
Remove the stride variable from rga_frame. Despite the comment it
didn't involve any calculation and is just a copy of the
plane_fmt[0].bytesperline value. Therefore avoid this struct member
and use the bytesperline value directly in the places where it is
required.
Also drop the dependency on the depth format member, which was only
used to calculate the stride of the default format. This is already done
by the v4l2_fill_pixfmt_mp_aligned helper and used as stride in try_fmt.
Therefore using it's value also for the default format stride is just
more consistent.
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Sven Püschel <s.pueschel@pengutronix.de>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org>
The size member is only used for the mmu page table mapping.
Therefore avoid storing the value and instead only calculate it
in place. This also avoids the calculation entirely when an external
iommu is used.
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Sven Püschel <s.pueschel@pengutronix.de>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org>
The RGA3 and the corresponding iommu share the interrupt. So in that
case, request a shared interrupt so that the iommu driver can request
it as well.
Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Sven Püschel <s.pueschel@pengutronix.de>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org>
In preparation for the RGA3 add support for external iommus. This is a
transition step to just disable the RGA2 specific mmu table setup code.
Currently a simple rga_hw struct field is used to set the internal iommu.
But to handle the case of more sophisticated detection mechanisms
(e.g. check for an iommu property in the device tree), it is abstracted
by an inline function.
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Sven Püschel <s.pueschel@pengutronix.de>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org>
Change the offset to dma_addresses, as the current naming is misleading.
The offset naming comes from the fact that it references the offset in
the mapped iommu address space. But from the hardware point of view this
is an address, as also pointed out by the register naming
(e.g. RGA_DST_Y_RGB_BASE_ADDR). Therefore also change the type to
dma_addr_t, as with an external iommu driver this would also be the
correct type.
This change is a preparation for the RGA3 support, which uses an external
iommu and therefore just gets an dma_addr_t for each buffer. The field
renaming allows to reuse the existing fields of rga_vb_buffer to store
these values.
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Sven Püschel <s.pueschel@pengutronix.de>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org>
In preparation of the RGA3 support add a filed to the rga_hw struct
to specify the desired card type value. This allows the user to
differentiate the RGA2 and RGA3 video device nodes.
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Sven Püschel <s.pueschel@pengutronix.de>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org>
Check the scaling factor to avoid potential problems. This is relevant
for the upcoming RGA3 support, as it can hang when the scaling factor
is exceeded.
The check is done at streamon when the other side is already streaming
to avoid incorrectly failing if the application configures the other
side after calling streamon. As try_fmt shouldn't be state aware,
it cannot be used to limit the format based on the scaling factor.
Therefore the check is done just before the actual streaming would be
started.
As the driver allows changing the rotation and selection while
streaming, add additional checks to ensure these changes
don't exceed the scaling factor.
Signed-off-by: Sven Püschel <s.pueschel@pengutronix.de>
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org>
Reuse the command buffer contents instead of completely writing it
for every frame. Therefore we only need to replace the source and
destination addresses for each frame. This reduces the amount of CPU
and memory operations done in each frame. A new cmdbuf_dirty flag notes
if the cmdbuf has to be rewritten on the next frame.
The initial idea of initializing the cmdbuf on streamon broke the
ability to update controls while streaming (e.g. mirroring).
Signed-off-by: Sven Püschel <s.pueschel@pengutronix.de>
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org>
Add an alignment setting to rga_hw to set the desired stride alignment.
As the RGA2 register for the stride counts in word units, the code
already divides the bytesperline value by 4 when writing it into the
register. Therefore fix the alignment to a multiple of 4 to avoid
potential off by one errors due from the division.
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Sven Püschel <s.pueschel@pengutronix.de>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org>
Move the command buffer to the rga_ctx struct in preparation to reuse
an already prepared command buffer. This allows to split the command
buffer setup in a further commit to setup a template for the command
buffer at streamon and only update the buffer addresses in device_run
and trigger the command stream. No sync point is added, as one command
buffer should only be used for one conversion at a time.
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Sven Püschel <s.pueschel@pengutronix.de>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org>
Calculate the x_div and y_div variables with the information from
v4l2_format_info instead of storing these in the rga_fmt struct.
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Sven Püschel <s.pueschel@pengutronix.de>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org>
Avoid odd frame sizes for YUV formats, as they may cause undefined
behavior. This is done in preparation for the RGA3, which hangs when the
output format is set to 129x129 pixel YUV420 SP (NV12).
This requirement is documented explicitly for the RGA3 in section 5.6.3
of the RK3588 TRM Part 2. For the RGA2 the RK3588 TRM Part 2
(section 6.1.2) and RK3568 TRM Part 2 (section 14.2) only mentions the
x/y offsets and stride aligning requirements. But the vendor driver for
the RGA2 also contains checks for the width and height to be aligned to
2 bytes.
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Sven Püschel <s.pueschel@pengutronix.de>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org>
In preparation for the RGA3 unit, move RGA2 specific parts from rga.c
to rga-hw.c and create a struct to reference the RGA2 specific functions
and formats. This also allows to remove the rga-hw.h reference from the
include list of the rga driver.
Also document the command finish interrupt with a dedicated define.
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Sven Püschel <s.pueschel@pengutronix.de>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org>
Announce the capability to adjust the quantization and ycbcr_enc on the
capture side and check if the SET_CSC flag is set when the colorimetry
is changed. Furthermore copy the colorimetry from the output to the
capture side to fix the currently failing v4l2-compliance tests, which
expect exactly this behavior.
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Sven Püschel <s.pueschel@pengutronix.de>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org>
Remove the redundant rga_frame variables width, height and color space.
The value of these variables is already contained in the pix member
of rga_frame. The code also keeps these values in sync. Therefore drop
them in favor of the existing pix member.
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Sven Püschel <s.pueschel@pengutronix.de>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org>
Use the stride instead of the width for the offset calculation. This
ensures that the bytesperline value doesn't need to match the width
value of the image.
Furthermore this patch removes the dependency on the uv_factor property
and instead reuses the v4l2_format_info to determine the correct
division factor.
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Sven Püschel <s.pueschel@pengutronix.de>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org>
Use the clk_bulk API to avoid code duplication for each of the three
clocks.
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Sven Püschel <s.pueschel@pengutronix.de>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org>
Fix the command buffer size being only a quarter of the actual size.
The RGA_CMDBUF_SIZE macro was potentially intended to specify the length
of the cmdbuf u32 array pointer. But as it's used to specify the size of
the allocation, which is counted in bytes. Therefore adjust the macro
size to bytes as it better matches the variable name and adjust it's
users accordingly.
As the command buffer is relatively small, it probably didn't caused
an issue due to being smaller than a single page.
Fixes: f7e7b48e6d ("[media] rockchip/rga: v4l2 m2m support")
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Sven Püschel <s.pueschel@pengutronix.de>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org>
The CAC block implements chromatic aberration correction. Expose it to
userspace using the extensible parameters format. This was tested on the
i.MX8MP platform, but based on available documentation it is also present
in the RK3399 variant (V10). Thus presumably also in later versions,
so no feature flag is introduced.
Signed-off-by: Barnabás Pőcze <barnabas.pocze@ideasonboard.com>
Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Link: https://patch.msgid.link/20260511150957.581049-1-barnabas.pocze@ideasonboard.com
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org>
The VDPU383 support for hevc and h264 use structs with bitfields to
represent the SPS and PPS.
Because the fields are mostly unaligned and numerous, it brings compiler
issues, especially with clang.
To prevent that, switch to using the global bitwriter previously
introduced instead.
Signed-off-by: Detlev Casanova <detlev.casanova@collabora.com>
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org>
Currently, the common code files for hevc and h264 use structs with
bitfields to represent the HW RPS buffer.
Because the bitfields are mostly unaligned and numerous, it brings compiler
issues, especially with clang.
To prevent that, switch to using the global bitwriter previously
introduced instead.
Signed-off-by: Detlev Casanova <detlev.casanova@collabora.com>
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org>
Both rkvdec-h264.c and rkvdec-hevc.c use their own bitwriter
function and macros.
Move to using the global one introduced before.
Signed-off-by: Detlev Casanova <detlev.casanova@collabora.com>
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org>
The use of structures with bitfields is good when the values are
somewhat aligned.
More mis-alignement means that compilers need to do more gymnastics
to edit the fields values.
Some cases have been reported with CLang on specific architectures
like armhf and hexagon, where the compiler would allocate a bigger
local stack than needed or even completely freeze during compilation.
Some fixes have been provided to ease the issues, but the real fix
here is to use a bitwriter instead of heavily unaligned bitfields.
This is a preparation commit to provide a global bitwriter interface
for the whole driver.
Signed-off-by: Detlev Casanova <detlev.casanova@collabora.com>
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org>
-----BEGIN PGP SIGNATURE-----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=3RtB
-----END PGP SIGNATURE-----
Merge tag 'media/v7.1-1' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-media
Pull media updates from Mauro Carvalho Chehab:
- new CSI tegra support, covering Tegra20 and Tegra30
- new camera sensor drivers: T4ka3 and ov2732
- m88ds3103: add 3103c chip support
- uvcvideo: add support for Intel RealSense D436/D555 and P010 pixel format
- synopsys csi2rx: add i.MX93 support
- imx8-isi: add i.MX95 support
- imx8mq-mipi-csi2: add i.MX8ULP support
- dw100: add V4L2 requests support
- support for DTV devices from Hauppauge got some improvements
- media staging: dropped starfive-camss driver
- media docs: document multi-committers model and improve maint profile
- media core:
- add v4l2_subdev_get_frame_desc_passthrough() helper
- improve error handling in fwnode parsing
- lots of driver fixes, cleanups and improvements
* tag 'media/v7.1-1' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-media: (251 commits)
Revert "media: cx231xx: add USB ID 2040:8360 for Hauppauge WinTV-HVR-935"
media: synopsys: csi2rx: add i.MX93 support
media: dt-bindings: add NXP i.MX93 compatible string
media: synopsys: csi2rx: Use enum and u32 array for register offsets
media: synopsys: csi2rx: implement .get_frame_desc() callback
media: synopsys: csi2rx: only check errors from devm_clk_bulk_get_all()
media: synopsys: csi2rx: use devm_reset_control_get_optional_exclusive()
media: i2c: imx283: add support for non-continuous MIPI clock mode
media: i2c: ov08d10: add support for 24 MHz input clock
media: i2c: ov08d10: add support for reset and power management
media: i2c: ov08d10: add support for binding via device tree
dt-bindings: media: i2c: document Omnivision OV08D10 CMOS image sensor
media: i2c: ov08d10: add missing newline to prints
media: i2c: ov08d10: fix some typos in comments
media: i2c: ov08d10: remove duplicate register write
media: i2c: ov08d10: fix image vertical start setting
media: i2c: ov08d10: fix runtime PM handling in probe
staging: media: ipu7: Update TODO
media: Add t4ka3 camera sensor driver
media: i2c: Add ov2732 image sensor driver
...
The deeply nested loop in rkvdec_init_v4l2_vp9_count_tbl() needs a lot
of registers, so when the clang register allocator runs out, it ends up
spilling countless temporaries to the stack:
drivers/media/platform/rockchip/rkvdec/rkvdec-vp9.c:966:12: error: stack frame size (1472) exceeds limit (1280) in 'rkvdec_vp9_start' [-Werror,-Wframe-larger-than]
Marking this function as noinline_for_stack keeps it out of
rkvdec_vp9_start(), giving the compiler more room for optimization.
The resulting code is good enough that both the total stack usage
and the loop get enough better to stay under the warning limit,
though it's still slow, and would need a larger rework if this
function ends up being called in a fast path.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
The rkvdec_pps had a large set of bitfields, all of which
as misaligned. This causes clang-21 and likely other versions to
produce absolutely awful object code and a warning about very
large stack usage, on targets without unaligned access:
drivers/media/platform/rockchip/rkvdec/rkvdec-vp9.c:966:12: error: stack frame size (1472) exceeds limit (1280) in 'rkvdec_vp9_start' [-Werror,-Wframe-larger-than]
Part of the problem here is how all the bitfield accesses are
inlined into a function that already has large structures on
the stack.
Mark set_field_order_cnt() as noinline_for_stack, and split out
the following accesses in assemble_hw_pps() into another noinline
function, both of which now using around 800 bytes of stack in the
same configuration.
There is clearly still something wrong with clang here, but
splitting it into multiple functions reduces the risk of stack
overflow.
Fixes: fde2490757 ("media: rkvdec: Add H264 support for the VDPU383 variant")
Link: https://godbolt.org/z/acP1eKeq9
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
The values of ext_sps_st_rps and ext_sps_lt_rps in struct rkvdec_hevc_run
are not initialized when the respective controls are not set by userspace.
When this is the case, set them to NULL so the rkvdec_hevc_run_preamble
function that parses controls does not access garbage data which leads to
a panic on unaccessible memory.
Fixes: c9a59dc2ac ("media: rkvdec: Add HEVC support for the VDPU381 variant")
Reported-by: Christian Hewitt <christianshewitt@gmail.com>
Suggested-by: Jonas Karlman <jonas@kwiboo.se>
Signed-off-by: Detlev Casanova <detlev.casanova@collabora.com>
Tested-by: Christian Hewitt <christianshewitt@gmail.com>
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Each stream requires CIF_REQ_BUFS_MIN=1 buffers to enable streaming.
However, it failed with only one buffer provided.
Comply with the minimum number of buffers requirement and accept
exactly one buffer.
Fixes: 501802e2ad ("media: rockchip: rkcif: add abstraction for dma blocks")
Cc: stable@kernel.org
Tested-by: Paul Elder <paul.elder@ideasonboard.com>
Tested-by: Chen-Yu Tsai <wens@kernel.org>
Signed-off-by: Michael Riesch <michael.riesch@collabora.com>
Reviewed-by: Isaac Scott <isaac.scott@ideasonboard.com>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Change these comparisons from > vs >= to avoid accessing one element
beyond the end of the arrays.
While at it, use ARRAY_SIZE instead of the _MAX enum values.
Fixes: 1f2353f5a1 ("media: rockchip: rkcif: add support for rk3568 vicap mipi capture")
Cc: stable@kernel.org
Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org>
Reviewed-by: Michael Riesch <michael.riesch@collabora.com>
Reviewed-by: Paul Elder <paul.elder@ideasonboard.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Tested-by: Chen-Yu Tsai <wens@kernel.org>
[fix cosmetic issues]
Signed-off-by: Michael Riesch <michael.riesch@collabora.com>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
This was done entirely with mindless brute force, using
git grep -l '\<k[vmz]*alloc_objs*(.*, GFP_KERNEL)' |
xargs sed -i 's/\(alloc_objs*(.*\), GFP_KERNEL)/\1)/'
to convert the new alloc_obj() users that had a simple GFP_KERNEL
argument to just drop that argument.
Note that due to the extreme simplicity of the scripting, any slightly
more complex cases spread over multiple lines would not be triggered:
they definitely exist, but this covers the vast bulk of the cases, and
the resulting diff is also then easier to check automatically.
For the same reason the 'flex' versions will be done as a separate
conversion.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This is the result of running the Coccinelle script from
scripts/coccinelle/api/kmalloc_objs.cocci. The script is designed to
avoid scalar types (which need careful case-by-case checking), and
instead replace kmalloc-family calls that allocate struct or union
object instances:
Single allocations: kmalloc(sizeof(TYPE), ...)
are replaced with: kmalloc_obj(TYPE, ...)
Array allocations: kmalloc_array(COUNT, sizeof(TYPE), ...)
are replaced with: kmalloc_objs(TYPE, COUNT, ...)
Flex array allocations: kmalloc(struct_size(PTR, FAM, COUNT), ...)
are replaced with: kmalloc_flex(*PTR, FAM, COUNT, ...)
(where TYPE may also be *VAR)
The resulting allocations no longer return "void *", instead returning
"TYPE *".
Signed-off-by: Kees Cook <kees@kernel.org>
The VDPU383 decoder is used on the RK3576 SoC and has support for HEVC.
This patch also moves some functions to a common rkvdec-hevc-common.c
file and adds a specific scaling matrix flatten function.
The fluster score for JCT-VC-HEVC_V1 is 146/147.
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Detlev Casanova <detlev.casanova@collabora.com>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org>
The VDPU381 supports HEVC decoding up to 7680x4320@30fps.
It could double that when using both decoder cores.
It support YUV420 (8 and 10 bits) as well as AFBC (not implemented
here)
The fluster score is 146/147 for JCT-VC-HEVC_V1, tested on ROCK 5B.
None of the other test suites works.
Tested-by: Diederik de Haas <didi.debian@cknow.org> # Rock 5B
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Detlev Casanova <detlev.casanova@collabora.com>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org>
This variant is used on the RK3576 SoC.
The moving vectors size requirements are slightly different so support
for a colmv_size function per variant is added.
Also, the link registers are used to start the decoder and read IRQ status.
The fluster score is 128/135 for JVT-AVC_V1, with MPS_MW_A failing in
addition to the usual ones.
The other test suites are not supported yet.
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Detlev Casanova <detlev.casanova@collabora.com>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org>
This decoder variant is found in Rockchip RK3588 SoC family.
Like for rkvdec on rk3399, it supports the NV12, NV15, NV16 and NV20
output formats and level up to 6.0.
The maximum width and height have been significantly increased
supporting up to 65520 pixels for both.
Support for named register sections is added for this variant and future
ones.
Fluster score for JVT-AVC_V1 is 129/135.
Tested-by: Diederik de Haas <didi.debian@cknow.org> # Rock 5B
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Detlev Casanova <detlev.casanova@collabora.com>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org>
Similarly to what is done in Hantro, avoid exposing equal 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 currently applies only to RK3588 which has 2 equal VDPU381 decoders,
but will be applied for all SoC supported by rkvdec that has multiple DTS
nodes with the same compatible.
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Detlev Casanova <detlev.casanova@collabora.com>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org>
For other variants, the clock names and number will differ.
There is no need to keep track of the clock names in the driver so drop
them to avoid having a list for each variant.
Tested-by: Diederik de Haas <didi.debian@cknow.org> # Rock 5B
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Detlev Casanova <detlev.casanova@collabora.com>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org>
