Laurent Pinchart [Mon, 6 Feb 2023 22:47:02 +0000 (23:47 +0100)]
media: i2c: imx296: Use v4l2_subdev_get_fmt()
The imx296 driver uses the subdev active state, there's no need to
implement the .get_fmt() operation manually. Use the
v4l2_subdev_get_fmt() helper instead.
The ov5670 driver tries to get a reference to the xvclk provider by using
the common cock framework and deflects to parsing the "clock-frequency"
property in case the clock provider is not specified in the firmware
interface, detected by checking if ov5670->xvclk == PTR_ERR(-ENOENT).
However, as reported by the Smatch static checker, if CONFIG_HAVE_CLK is
not enabled, devm_clk_get() returns 0 which when passed to PTR_ERR()
means success causing the driver to fail without propagating any error
code up.
Explicitly handle the case where ov5670->xvclk it set to NULL, forcing
the code to parse the "clock-frequency" property in case CONFIG_HAVE_CLK
is not enabled, as suggested by Dan Carpenter.
Reported-by: Dan Carpenter <error27@gmail.com> Suggested-by: Dan Carpenter <error27@gmail.com> Signed-off-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Luca Weiss [Fri, 10 Feb 2023 20:33:18 +0000 (21:33 +0100)]
media: i2c: ov5670: Support single-lane operation
Currently the driver always configures the sensor for dual-lane MIPI
output, but it also supports single-lane output. Add support for that by
checking the data-lanes fwnode property how many lanes are used and
configure the necessary registers based on that.
To achieve this we move setting register 0x3018 out of the general reg
sequence so we set it to the correct value. The pixel_rate value also
needs to be adjusted.
[Sakari Ailus: Use div_s64 to divide a 64-bit number]
Krzysztof Kozlowski [Mon, 13 Mar 2023 14:49:15 +0000 (15:49 +0100)]
media: dt-bindings: samsung,fimc: convert to dtschema
Convert the Samsung S5P/Exynos Camera Subsystem (FIMC) bindings to DT
schema. Changes during conversion - adjust to existing DTS and Linux
driver: add iommus and power-domains.
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> Reviewed-by: Rob Herring <robh@kernel.org> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Krzysztof Kozlowski [Mon, 13 Mar 2023 14:49:14 +0000 (15:49 +0100)]
media: dt-bindings: samsung,exynos4212-is: convert to dtschema
Convert the Samsung Exynos4212/4412 SoC Imaging Subsystem (FIMC-IS)
bindings to DT schema. Changes during conversion - adjust to existing
DTS and Linux driver: add iommus and power-domains.
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> Reviewed-by: Rob Herring <robh@kernel.org> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Krzysztof Kozlowski [Mon, 13 Mar 2023 14:49:13 +0000 (15:49 +0100)]
media: dt-bindings: samsung,exynos4212-fimc-lite: convert to dtschema
Convert the Samsung Exynos SoC series camera host interface (FIMC-LITE)
bindings to DT schema. Changes during conversion - adjust to existing
DTS and Linux driver: add iommus and power-domains.
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> Reviewed-by: Rob Herring <robh@kernel.org> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Krzysztof Kozlowski [Mon, 13 Mar 2023 14:49:12 +0000 (15:49 +0100)]
media: dt-bindings: samsung,exynos4210-csis: convert to dtschema
Convert the Samsung S5P/Exynos SoC series MIPI CSI-2 receiver (MIPI
CSIS) bindings to DT schema. Changes during conversion - adjust to
existing DTS and Linux driver:
1. Add phys and power-domains.
2. Move samsung,csis-wclk property to the endpoint node.
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> Reviewed-by: Rob Herring <robh@kernel.org> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Krzysztof Kozlowski [Mon, 13 Mar 2023 14:49:09 +0000 (15:49 +0100)]
media: samsung: exynos4-is: drop simple-bus from compatibles
The FIMC camera node wrapper is not a bus, so using simple-bus fallback
compatible just to instantiate its children nodes was never correct.
Driver should explicitly populate all its children devices.
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Krzysztof Kozlowski [Mon, 13 Mar 2023 14:49:08 +0000 (15:49 +0100)]
media: samsung: exynos4-is: do not require pinctrl
Driver does not handle pin configuration switching and several DTS
provide empty pinctrl property, just to satisfy the driver's requirement
for it. Drop requirement for pinctrl property as it is really optional.
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Krzysztof Kozlowski [Mon, 13 Mar 2023 14:49:07 +0000 (15:49 +0100)]
media: dt-bindings: samsung-fimc: drop simple-bus
The FIMC camera node wrapper is not a bus, so using simple-bus fallback
compatible just to instantiate its children nodes was never correct.
Drop the simple-bus compatible and expect driver to explicitly populate
children devices.
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> Acked-by: Rob Herring <robh@kernel.org> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Dorota Czaplejewicz [Sun, 19 Feb 2023 11:01:21 +0000 (12:01 +0100)]
media: media api: Try to make enum usage clearer
This clarifies which side of the calls is responsible for doing what to
which parts of the struct. It also expands the terse description of the
access algorithm into more prose-like, active voice description, which
trades conciseness for ease of comprehension.
Fixed: typo "format" -> "frame size" in enum-frame-size
Added: no holes in the enumeration
Added: enumerations per what?
Added: who fills in what in calls
Changed: "given" -> "specified"
[Sakari Ailus: Rewrap text]
Signed-off-by: Dorota Czaplejewicz <dorota.czaplejewicz@puri.sm> Acked-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Dorota Czaplejewicz [Sun, 19 Feb 2023 11:00:34 +0000 (12:00 +0100)]
media: doc/media api: Try to make enum usage clearer
This clarifies which side of the calls is responsible for doing what to
which parts of the struct. This also explicitly states that repeating
values are disallowed. It also expands the terse description of the access
algorithm into more prose-like, active voice description, which trades
conciseness for ease of comprehension.
Added: mbus codes must not repeat
Added: no holes in the enumeration
Added: enumerations per what?
Added: who fills in what in calls
Changed: "zero" -> "0"
Changed: "given" -> "specified"
Still unclear how it works so didn't describe: "which". What is a "try
format" vs "active format"?
[Sakari Ailus: Rewrap lines, fix build issue]
Signed-off-by: Dorota Czaplejewicz <dorota.czaplejewicz@puri.sm> Acked-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Alexander Stein [Fri, 17 Feb 2023 09:52:21 +0000 (10:52 +0100)]
media: i2c: imx290: Add support for imx327 variant
Both sensors are quite similar. Their specs only differ regarding LVDS
and parallel output but are identical regarding MIPI-CSI-2 interface.
But they use a different init setting of hard-coded values, taken from
the datasheet.
Signed-off-by: Alexander Stein <alexander.stein@ew.tq-group.com> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Alexander Stein [Fri, 17 Feb 2023 09:52:20 +0000 (10:52 +0100)]
media: dt-bindings: media: i2c: Add imx327 version to IMX327 bindings
The imx290 driver can be used for both imx290 and imx327 as they have a
similar register set and configuration. imx327 lacks 8 lanes LVDS and
120 FPS support.
Signed-off-by: Alexander Stein <alexander.stein@ew.tq-group.com> Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Dave Stevenson [Wed, 15 Feb 2023 22:30:02 +0000 (23:30 +0100)]
media: i2c: imx290: Add support for H & V Flips
The sensor supports H & V flips, so add the relevant hooks for
V4L2_CID_HFLIP and V4L2_CID_VFLIP to configure them.
Note that the Bayer order is maintained as the readout area
shifts by 1 pixel in the appropriate direction (note the
comment about the top margin being 8 pixels whilst the bottom
margin is 9). The V4L2_SEL_TGT_CROP region is therefore
adjusted appropriately.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com> Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Dave Stevenson [Wed, 15 Feb 2023 22:29:59 +0000 (23:29 +0100)]
media: i2c: imx290: VMAX is mode dependent
The default VMAX for 60fps in 720p mode is 750 according to the
datasheet, however the driver always left it at 1125 thereby stopping
60fps being achieved.
Make VMAX (and therefore V4L2_CID_VBLANK) mode dependent so that 720p60
can be achieved.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com> Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Dave Stevenson [Wed, 15 Feb 2023 22:29:58 +0000 (23:29 +0100)]
media: i2c: imx290: Convert V4L2_CID_VBLANK to read/write
The driver exposed V4L2_CID_VBLANK as a read only control to allow
for exposure calculations and determination of the frame rate.
Convert to a read/write control so that the frame rate can be
controlled.
V4L2_CID_VBLANK also sets the limits for the exposure control,
therefore exposure ranges have to be updated when vblank changes
(either via s_ctrl, or via changing mode).
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com> Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Acked-by: Alexander Stein <alexander.stein@ew.tq-group.com> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Dave Stevenson [Wed, 15 Feb 2023 22:29:56 +0000 (23:29 +0100)]
media: i2c: imx290: Use CSI timings as per datasheet
Commit "98e0500eadb7 media: i2c: imx290: Add configurable link frequency
and pixel rate" added support for the increased link frequencies
on 2 data lanes, but didn't update the CSI timing registers in
accordance with the datasheet.
Use the specified settings.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com> Reviewed-by: Alexander Stein <alexander.stein@ew.tq-group.com> Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Dave Stevenson [Wed, 15 Feb 2023 22:29:55 +0000 (23:29 +0100)]
media: i2c: imx290: Support 60fps in 2 lane operation
Commit "97589ad61c73 media: i2c: imx290: Add support for 2 data lanes"
added support for running in two lane mode (instead of 4), but
without changing the link frequency that resulted in a max of 30fps.
Commit "98e0500eadb7 media: i2c: imx290: Add configurable link frequency
and pixel rate" then doubled the link frequency when in 2 lane mode,
but didn't undo the correction for running at only 30fps, just extending
horizontal blanking instead.
Remove the 30fps limit on 2 lane by correcting the register config
in accordance with the datasheet for 60fps operation over 2 lanes.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com> Reviewed-by: Alexander Stein <alexander.stein@ew.tq-group.com> Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Dave Stevenson [Wed, 15 Feb 2023 22:29:54 +0000 (23:29 +0100)]
media: i2c: imx290: Fix the pixel rate at 148.5Mpix/s
The datasheet lists the link frequency changes between
1080p and 720p modes. This is correct that the link frequency
changes as measured on an oscilloscope.
Link frequency is not necessarily the same as pixel rate.
The datasheet gives standard configurations for 1080p and 720p
modes at a number of frame rates.
Looking at the 1080p mode it gives:
HMAX = 0x898 = 2200
VMAX = 0x465 = 1125
2200 * 1125 * 60fps = 148.5MPix/s
Looking at the 720p mode it gives:
HMAX = 0xce4 = 3300
VMAX = 0x2ee = 750
3300 * 750 * 60fps = 148.5Mpix/s
This driver currently scales the pixel rate proportionally to the
link frequency, however the above shows that this is not the
correct thing to do, and currently all frame rate and exposure
calculations give incorrect results.
Correctly report the pixel rate as being 148.5MPix/s under any
mode.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com> Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Dave Stevenson [Wed, 15 Feb 2023 22:29:53 +0000 (23:29 +0100)]
media: i2c: imx290: Add V4L2_SUBDEV_FL_HAS_EVENTS and subscribe hooks
Any V4L2 subdevice that implements controls and declares
V4L2_SUBDEV_FL_HAS_DEVNODE should also declare V4L2_SUBDEV_FL_HAS_EVENTS
and implement subscribe_event and unsubscribe_event hooks.
This driver didn't and would therefore fail v4l2-compliance
testing.
Add the relevant hooks.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com> Reviewed-by: Alexander Stein <alexander.stein@ew.tq-group.com> Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Dave Stevenson [Wed, 15 Feb 2023 22:29:49 +0000 (23:29 +0100)]
media: dt-bindings: media: i2c: Add mono version to IMX290 bindings
The IMX290 module is available as either monochrome or colour and
the variant is not detectable at runtime.
Add a new compatible string for the monochrome version, based on the
full device name IMX290LLR. For consistency, add a new compatible string
for the colour version based on the IMX290LQR full device name, and
deprecate the current ambiguous compatible string.
Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com> Acked-by: Rob Herring <robh@kernel.org> Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Tested-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Jacopo Mondi [Tue, 21 Feb 2023 17:10:48 +0000 (18:10 +0100)]
media: i2c: ov5647: Use bus-locked i2c_transfer()
The ov5647_read() functions calls i2c_master_send() and
i2c_master_read() in sequence. However this leaves space for other
clients to contend the bus and insert an unrelated transaction in between
the two calls.
Replace the two calls with a single i2c_transfer() one, that locks the
bus in between the transactions.
A common case with subdev routing is that on the subdevice just before
the DMA engines (video nodes), no multiplexing is allowed on the source
pads, as the DMA engine can only handle a single stream.
In some other situations one might also want to do the same check on the
sink side.
Add new routing validation flags to check these:
V4L2_SUBDEV_ROUTING_NO_SINK_MULTIPLEXING and
V4L2_SUBDEV_ROUTING_NO_SOURCE_MULTIPLEXING.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ideasonboard.com> Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
V4L2_SUBDEV_ROUTING_NO_STREAM_MIX routing validation flag means that all
routes from a sink pad must go to the same source pad and all routes
going to the same source pad must originate from the same sink pad.
This does not cover all use cases. For example, if a device routes
all streams from a single sink pad to any of the source pads, but
streams from multiple sink pads can go to the same source pad, the
current flag is too restrictive.
Split the flag into two parts, V4L2_SUBDEV_ROUTING_NO_SINK_STREAM_MIX
and V4L2_SUBDEV_ROUTING_NO_SOURCE_STREAM_MIX, which add the restriction
only on one side of the device. Together they mean the same as
V4L2_SUBDEV_ROUTING_NO_STREAM_MIX.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ideasonboard.com> Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Reviewed-by: Jacopo Mondi <jacopo.mondi@ideasonboard.com> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Benjamin Mugnier [Wed, 1 Feb 2023 14:04:15 +0000 (15:04 +0100)]
media: i2c: st-vgxy61: Move 'detect' call to 'power_on'
Previously the device detection was performed after patching.
Move it right after the reset to make sure we have the correct sensor
before trying to patch it.
Signed-off-by: Benjamin Mugnier <benjamin.mugnier@foss.st.com> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Hans Verkuil [Thu, 2 Mar 2023 12:57:27 +0000 (13:57 +0100)]
media: vivid: drop bitmap and clipping output overlay support
This test driver is the only remaining driver still using
the clipping and bitmap method. Drop support for this so
we can remove this in the V4L2 API as well.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Hans Verkuil [Thu, 2 Mar 2023 12:57:26 +0000 (13:57 +0100)]
media: vivid: drop overlay support
Destructive overlay support (i.e. where the video frame is DMA-ed
straight into a framebuffer) is effectively dead. It was a
necessary evil in the early days when computers were not fast enough
to copy SDTV video frames around, but today that's no longer a problem.
It requires access to the framebuffer memory, which is a bad idea and
very hard to do safely. In addition, in drm it is today almost
impossible to get hold of the framebuffer address.
So drop support for this.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Hans Verkuil [Thu, 2 Mar 2023 12:57:25 +0000 (13:57 +0100)]
media: bttv: drop overlay support
Destructive overlay support (i.e. where the video frame is DMA-ed
straight into a framebuffer) is effectively dead. It was a
necessary evil in the early days when computers were not fast enough
to copy SDTV video frames around, but today that's no longer a problem.
It requires access to the framebuffer memory, which is a bad idea and
very hard to do safely. In addition, in drm it is today almost
impossible to get hold of the framebuffer address.
So drop support for this.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Hans Verkuil [Thu, 2 Mar 2023 12:57:24 +0000 (13:57 +0100)]
media: saa7134: drop overlay support
Destructive overlay support (i.e. where the video frame is DMA-ed
straight into a framebuffer) is effectively dead. It was a
necessary evil in the early days when computers were not fast enough
to copy SDTV video frames around, but today that's no longer a problem.
It requires access to the framebuffer memory, which is a bad idea and
very hard to do safely. In addition, in drm it is today almost
impossible to get hold of the framebuffer address.
So drop support for this.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Hans Verkuil [Thu, 2 Mar 2023 12:57:23 +0000 (13:57 +0100)]
media: saa7146: drop overlay support
Destructive overlay support (i.e. where the video frame is DMA-ed
straight into a framebuffer) is effectively dead. It was a
necessary evil in the early days when computers were not fast enough
to copy SDTV video frames around, but today that's no longer a problem.
It requires access to the framebuffer memory, which is a bad idea and
very hard to do safely. In addition, in drm it is today almost
impossible to get hold of the framebuffer address.
So drop support for this.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
The documentation on how to create your own Raspberry Pi CEC debugger was a
bit out of date. Update it to the Raspberry Pi 4B, drop the mention of the RTC
and a link to a picture that no longer works.
Also reorganize the text to make it easier to follow and change the pins to
match the pins I use.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Dan Carpenter [Tue, 7 Mar 2023 10:00:23 +0000 (11:00 +0100)]
media: av7110: prevent underflow in write_ts_to_decoder()
The buf[4] value comes from the user via ts_play(). It is a value in
the u8 range. The final length we pass to av7110_ipack_instant_repack()
is "len - (buf[4] + 1) - 4" so add a check to ensure that the length is
not negative. It's not clear that passing a negative len value does
anything bad necessarily, but it's not best practice.
With the new bounds checking the "if (!len)" condition is no longer
possible or required so remove that.
Set allow_cache_hints to 1 for the vb2_queue capture queue in the
STM32MP15xx DCMI V4L2 driver. This allows us to allocate buffers
with the V4L2_MEMORY_FLAG_NON_COHERENT set. On STM32MP15xx SoCs,
this enables caching for this memory, which improves performance
when being read from CPU.
This change should be safe from race conditions since videobuf2
already invalidates or flushes the appropriate cache lines in
its prepare() and finish() methods.
Tested on a STM32MP157F SoC. Resulted in 4x buffer access speedup.
Signed-off-by: Marek Vasut <marex@denx.de> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
harperchen [Fri, 3 Mar 2023 15:30:11 +0000 (16:30 +0100)]
media: pci: tw68: Fix null-ptr-deref bug in buf prepare and finish
When the driver calls tw68_risc_buffer() to prepare the buffer, the
function call dma_alloc_coherent may fail, resulting in a empty buffer
buf->cpu. Later when we free the buffer or access the buffer, null ptr
deref is triggered.
We believe the bug can be also dynamically triggered from user side.
Similarly, we fix this by checking the return value of tw68_risc_buffer()
and the value of buf->cpu before buffer free.
harperchen [Thu, 2 Mar 2023 12:39:05 +0000 (13:39 +0100)]
media: cx23885: Fix a null-ptr-deref bug in buffer_prepare() and buffer_finish()
When the driver calls cx23885_risc_buffer() to prepare the buffer, the
function call dma_alloc_coherent may fail, resulting in a empty buffer
risc->cpu. Later when we free the buffer or access the buffer, null ptr
deref is triggered.
We believe the bug can be also dynamically triggered from user side.
Similarly, we fix this by checking the return value of cx23885_risc_buffer()
and the value of risc->cpu before buffer free.
amphion vpu support a low latency mode,
when V4L2_CID_MPEG_VIDEO_DEC_DISPLAY_DELAY_ENABLE is enabled,
decoder can display frame immediately after it's decoded.
Only h264 is support yet.
Jiasheng Jiang [Thu, 9 Feb 2023 13:52:45 +0000 (14:52 +0100)]
media: platform: mtk-mdp3: Add missing check and free for ida_alloc
Add the check for the return value of the ida_alloc in order to avoid
NULL pointer dereference.
Moreover, free allocated "ctx->id" if mdp_m2m_open fails later in order
to avoid memory leak.
Fixes: 61890ccaefaf ("media: platform: mtk-mdp3: add MediaTek MDP3 driver") Signed-off-by: Jiasheng Jiang <jiasheng@iscas.ac.cn> Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Ye Xingchen [Wed, 8 Feb 2023 09:01:06 +0000 (10:01 +0100)]
media: platform: stm32: use devm_platform_get_and_ioremap_resource()
Convert platform_get_resource(), devm_ioremap_resource() to a single
call to devm_platform_get_and_ioremap_resource(), as this is exactly
what this function does.
Signed-off-by: Ye Xingchen <ye.xingchen@zte.com.cn> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Ye Xingchen [Wed, 8 Feb 2023 08:59:40 +0000 (09:59 +0100)]
media: platform: renesas: use devm_platform_get_and_ioremap_resource()
Convert platform_get_resource(), devm_ioremap_resource() to a single
call to devm_platform_get_and_ioremap_resource(), as this is exactly
what this function does.
Ye Xingchen [Wed, 8 Feb 2023 08:35:44 +0000 (09:35 +0100)]
media: dw100: use devm_platform_get_and_ioremap_resource()
Convert platform_get_resource(), devm_ioremap_resource() to a single
call to devm_platform_get_and_ioremap_resource(), as this is exactly
what this function does.
Signed-off-by: Ye Xingchen <ye.xingchen@zte.com.cn> Reviewed-by: Xavier Roumegue <xavier.roumegue@oss.nxp.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Kees Cook [Sat, 4 Feb 2023 18:38:05 +0000 (19:38 +0100)]
media: imx-jpeg: Bounds check sizeimage access
The call of mxc_jpeg_get_plane_size() from mxc_jpeg_dec_irq() sets
plane_no argument to 1. The compiler sees that it's possible to end up
with an access beyond the bounds of sizeimage, if mem_planes was too
large:
if (plane_no >= fmt->mem_planes) // mem_planes = 2+
return 0;
if (fmt->mem_planes == fmt->comp_planes) // comp_planes != mem_planes
return q_data->sizeimage[plane_no];
comp_planes == 0 or 1 is safe. comp_planes > 2 would be out of bounds.
(This isn't currently possible given the contents of mxc_formats, though.)
Silence the warning by bounds checking comp_planes for future
robustness. Seen with GCC 13:
In function 'mxc_jpeg_get_plane_size',
inlined from 'mxc_jpeg_dec_irq' at ../drivers/media/platform/nxp/imx-jpeg/mxc-jpeg.c:729:14:
../drivers/media/platform/nxp/imx-jpeg/mxc-jpeg.c:641:42: warning: array subscript 2 is above array bounds of 'u32[2]' {aka 'unsigned int[2]'} [-Warray-bounds=]
641 | size += q_data->sizeimage[i];
| ~~~~~~~~~~~~~~~~~^~~
In file included from ../drivers/media/platform/nxp/imx-jpeg/mxc-jpeg-hw.h:112,
from ../drivers/media/platform/nxp/imx-jpeg/mxc-jpeg.c:63:
../drivers/media/platform/nxp/imx-jpeg/mxc-jpeg.h: In function 'mxc_jpeg_dec_irq':
../drivers/media/platform/nxp/imx-jpeg/mxc-jpeg.h:84:41: note: while referencing 'sizeimage'
84 | u32 sizeimage[MXC_JPEG_MAX_PLANES];
| ^~~~~~~~~
Kees Cook [Sat, 4 Feb 2023 18:36:17 +0000 (19:36 +0100)]
media: mxl5005s: Bounds check size used for max array index
The use of state->CH_Ctrl[i].size in a shift operation implies that its
value can be as much as 32, but the state->CH_Ctrl[i].val array is only
25 in size. Bounds check the size before shifting and looping. Fixes
warnings seen with GCC 13:
../drivers/media/tuners/mxl5005s.c: In function 'MXL_ControlWrite_Group.isra':
../drivers/media/tuners/mxl5005s.c:3450:70: warning: array subscript 32 is above array bounds of 'u16[25]' {aka 'short unsigned int[25]'} [-Warray-bounds=]
3450 | state->CH_Ctrl[i].val[j] = (u8)((value >> j) & 0x01);
| ~~~~~~~~~~~~~~~~~~~~~^~~
../drivers/media/tuners/mxl5005s.c:238:13: note: while referencing 'val'
238 | u16 val[25]; /* Binary representation of Value */
| ^~~
Cc: Colin Ian King <colin.i.king@gmail.com> Signed-off-by: Kees Cook <keescook@chromium.org> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Jakub Kicinski [Fri, 3 Feb 2023 23:31:29 +0000 (00:31 +0100)]
media: drop unnecessary networking includes
dvb_net.h includes a bunch of core networking headers which increases
the number of objects rebuilt when we touch them. They are unnecessary
for the header itself and only one driver has an indirect dependency.
tveeprom.h includes if_packet to gain access to ETH_ALEN. This
is a bit of an overkill because if_packet.h pulls in skbuff.h.
The definition of ETH_ALEN is in the uAPI header, which is
very rarely touched, so switch to including that.
This results in roughly 250 fewer objects built when skbuff.h
is touched (6028 -> 5788).
Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
There are spelling mistakes in a literal string and a comment. Fix them.
Signed-off-by: Colin Ian King <colin.i.king@gmail.com> Reviewed-by: Alim Akhtar <alim.akhtar@samsung.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Christian Hewitt [Sat, 4 Feb 2023 07:45:01 +0000 (08:45 +0100)]
media: rc: add Beelink Mini MXIII keymap
Add a keymap for the simple IR (NEC) remote used with the Beelink
Mini MXIII Android STB device.
Signed-off-by: Christian Hewitt <christianshewitt@gmail.com> Signed-off-by: Sean Young <sean@mess.org> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Christian Hewitt [Sat, 4 Feb 2023 07:45:00 +0000 (08:45 +0100)]
media: dt-bindings: rc: add rc-beelink-mxiii
Add a binding for the rc-beelink-mxiii remote keymap
Signed-off-by: Christian Hewitt <christianshewitt@gmail.com> Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> Signed-off-by: Sean Young <sean@mess.org> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Emanuel Strobel [Fri, 3 Feb 2023 09:34:05 +0000 (10:34 +0100)]
media: rc: add common keymap for Dreambox RC10/RC0 and RC20/RC-BT remotes
Add a common keymap for the RC10/RC0 and RC20/RC-BT remotes used with
the Dreambox One and Dreambox Two DVB-S/T boxes. The maps are combined
since the IR codes do not conflict and both boxes have shipped with
both remote designs over time.
Both remote types can be programmed to control TVs, so include non-IR
keys that are used to switch-to or toggle the remote mode:
- DREAM in RC10/RC0 switches to (Dreambox) STB control mode
- TV in RC10/RC0 switches to TV control mode
- MODE in RC20/RC-BT toggles between STB/TV/BT control modes
In the RC20 keymap the Android MIC (voice search) key maps to KEY_HELP
and EXIT is mapped to KEY_ESC to replicate the go-backwards navigation
behaviour in the Android vendor OS that ships on Dreambox devices.
Signed-off-by: Emanuel Strobel <emanuel.strobel@yahoo.com> Signed-off-by: Christian Hewitt <christianshewitt@gmail.com> Signed-off-by: Sean Young <sean@mess.org> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Christian Hewitt [Fri, 3 Feb 2023 09:34:04 +0000 (10:34 +0100)]
media: dt-bindings: rc: add rc-dreambox
Add a binding for the rc-dreambox remote keymap
Signed-off-by: Christian Hewitt <christianshewitt@gmail.com> Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> Signed-off-by: Sean Young <sean@mess.org> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Hans Verkuil [Wed, 1 Feb 2023 12:56:20 +0000 (13:56 +0100)]
media: cx18: reorder fmt_vid_cap functions in cx18-ioctl.c
No actual code changed, this just reorders functions so that the
g/try/s_fmt_vid_cap functions are all together rather than mixed in
with the vbi format functions.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Hans Verkuil [Wed, 1 Feb 2023 12:54:20 +0000 (13:54 +0100)]
media: cx18: missing CAP_AUDIO for vbi stream
The v4l2-compliance test complained about this. You can change the
input connector for the vbi stream, but that means that you can have to
be able to do the same for the audio input.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Hans Verkuil [Wed, 1 Feb 2023 11:19:46 +0000 (12:19 +0100)]
media: cx18: fix incorrect input counting
Counting the number of video and audio inputs was wrong if the
number of inputs equalled CX18_CARD_MAX_VIDEO_INPUTS or
CX18_CARD_MAX_AUDIO_INPUTS. This was a copy-and-paste from the
ivtv driver. That driver has been fixed quite a long time ago,
but we missed that this driver had the same bug.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Hans Verkuil [Wed, 1 Feb 2023 13:06:47 +0000 (14:06 +0100)]
media: vb2: set owner before calling vb2_read
Before vb2_read is called, the owner must be set since vb2_read will
actually start streaming. If vb2_read returns an error and q->fileio is
NULL, then it failed to start streaming and the owner is set back to NULL.
When the vb2 start_streaming callback is called, it expects the owner
pointer to be set, and that wasn't the case in this particular situation.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
projects / users / jedix / linux-maple.git / log