Ezequiel Garcia [Thu, 26 Nov 2020 09:36:17 +0000 (10:36 +0100)]
media: uapi: move H264 stateless controls out of staging
The H.264 stateless 'uAPI' was staging and marked explicitly in the
V4L2 specification that it will change and is unstable.
Note that these control IDs were never exported as a public API,
they were only defined in kernel-local headers (h264-ctrls.h).
Now, the H264 stateless controls is ready to be part
of the stable uAPI.
While not too late, let's rename them and re-number their
control IDs, moving them to the newly created stateless
control class, and updating all the drivers accordingly.
Ezequiel Garcia [Thu, 26 Nov 2020 09:36:16 +0000 (10:36 +0100)]
media: controls: Log H264 stateless controls in .std_log
For now, only print the type of the control. In the future,
we might want to be able to print the values of all
the stateless codec controls currently set in the control handler,
which should be useful for debugging reasons.
Ezequiel Garcia [Thu, 26 Nov 2020 09:36:09 +0000 (10:36 +0100)]
media: Rename stateful codec control macros
For historical reasons, stateful codec controls are named
as {}_MPEG_{}. While we can't at this point sanely
change all control IDs (such as V4L2_CID_MPEG_VIDEO_VP8_FRAME_HEADER),
we can least change the more meaningful macros such as classes
macros.
Jonas Karlman [Thu, 26 Nov 2020 09:36:07 +0000 (10:36 +0100)]
media: rkvdec: h264: Support profile and level controls
The Rockchip Video Decoder used in RK3399 supports H.264 profiles from
Baseline to High 4:2:2 up to Level 5.1, except for the Extended profile.
Expose the V4L2_CID_MPEG_VIDEO_H264_PROFILE and the
V4L2_CID_MPEG_VIDEO_H264_LEVEL control, so that userspace can query the
driver for the list of supported profiles and level.
For now, we don't expose 4:2:2 since the driver doesn't
implement the required support.
[Ezequiel: Don't expose 4:2:2 profile for now]
Signed-off-by: Jonas Karlman <jonas@kwiboo.se> Signed-off-by: Ezequiel Garcia <ezequiel@collabora.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Add a debug message for control validation (validate_new)
failures. This is useful to debug issues with ioctls such
as VIDIOC_TRY_EXT_CTRLS and VIDIOC_S_EXT_CTRLS.
Sakari Ailus [Fri, 18 Sep 2020 09:36:38 +0000 (11:36 +0200)]
media: ccs: Wrap long lines, unwrap short ones
Over the years (and renaming) some lines that may well be wrapped ended up
being over 80 characters, likewise there are shorter lines that can be
merged. Do that.
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Sakari Ailus [Fri, 18 Sep 2020 09:23:39 +0000 (11:23 +0200)]
media: ccs: Clean up runtime PM usage
If pm_runtime_get_sync() fails, there's no need to set the device active
again. Also, in the same case to return the usage_count to zero,
pm_runtime_put_noidle() is enough.
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Sakari Ailus [Fri, 13 Nov 2020 15:07:28 +0000 (16:07 +0100)]
media: ccs: Allow range in between I²C retries
Make the delay between I²C access retries a range between 1 and 2 ms. Also
make the number of retries 10 instead of 5, in order not to reduce the
total amount of time.
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Sakari Ailus [Wed, 27 May 2020 21:59:40 +0000 (23:59 +0200)]
media: smiapp-pll: Rename as ccs-pll
MIPI CCS replaces SMIA and SMIA++ as the current standard. CCS brings new
features while existing functionality will be supported. Rename the
smiapp-pll as ccs-pll accordingly.
Also add Intel copyright to the files.
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Sakari Ailus [Wed, 27 May 2020 08:35:01 +0000 (10:35 +0200)]
media: ccs: Stop reading arrays after the first zero
The register arrays have a certain size but not all the entries will be
relevant. In practice reading can be stopped after encountering a zero
value in the array. Do that to avoid extra reads.
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Sakari Ailus [Fri, 14 Feb 2020 09:00:35 +0000 (10:00 +0100)]
media: ccs: Add CCS static data parser library
Add a parser library for parsing the CCS static data format.
The library may be also compiled in user space as the format has uses also
in the user space. Therefore it is dual licensed under the 3-clause BSD
license as well.
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Sakari Ailus [Wed, 30 Sep 2020 16:20:42 +0000 (18:20 +0200)]
media: ccs: Fix obtaining bus information from firmware
Let v4l2_fwnode_endpoint_alloc_parse to figure out the type of the data
bus. As the old bindings did not require the "bus-type" property, we need
to rely on guessing between CSI-2 D-PHY and CCP2. Setting the type to
CSI-2 D-PHY will parse just that and succeed even if no data-lanes are
set.
Also add a comment on the matter to the driver to avoid breaking this in
the future.
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Sakari Ailus [Wed, 2 Dec 2020 14:16:57 +0000 (15:16 +0100)]
media: Documentation: ccs: Rename ccs-regs.txt as ccs-regs.asc
As documentation used to be in .txt files before converting to ReST,
rename ccs-regs.txt to avoid it being taken as documentation that
pre-dates ReST conversion and so target for further conversion to ReST.
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
The debugfs_create_dir() function never returns NULLs so this code will
never be executed. It's not intended that callers will check for
debugfs errors in the normal case and it's not necessary in this driver,
so we can just delete this code.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Acked-by: Helen Koike <helen.koike@collabora.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
The wrappers in include/linux/pci-dma-compat.h should go away.
The patch has been generated with the coccinelle script below and has been
hand modified to replace GFP_ with a correct flag.
It has been compile tested.
When memory is allocated in 'saa7146_pgtable_alloc()' GFP_KERNEL can be
used because the callers are either .buf_prepare functions or function that
already use GFP_KERNEL (hidden in a 'vmalloc_32()' call).
When memory is allocated in 'saa7146_init_one()' GFP_KERNEL can be used
because it is probe function and no lock is taken in the between.
When memory is allocated in 'saa7146_vv_init()' GFP_KERNEL can be used
because this function already uses GFP_KERNEL and no lock is taken in the
between.
When memory is allocated in 'vbi_workaround()' GFP_KERNEL can be used
because it is only called from a .open function.
The wrappers in include/linux/pci-dma-compat.h should go away.
The patch has been generated with the coccinelle script below and has been
hand modified to replace GFP_ with a correct flag.
It has been compile tested.
When memory is allocated in 'av7110_attach()' GFP_KERNEL can be used
because this flag is already used above in the same function and no lock is
taken in the between.
The wrappers in include/linux/pci-dma-compat.h should go away.
The patch has been generated with the coccinelle script below and has been
hand modified to replace GFP_ with a correct flag.
It has been compile tested.
When memory is allocated in 'snd_solo_pcm_open()' (solo6x10-g723.c)
GFP_KERNEL can be used because this flag is already used jew a few lines
above.
When memory is allocated in 'solo_enc_alloc()' (solo6x10-v4l2-enc.c)
GFP_KERNEL can be used because this flag is already used jew a few lines
above.
When memory is allocated in 'solo_enc_v4l2_init()' (solo6x10-v4l2-enc.c)
GFP_KERNEL can be used because calls 'solo_enc_alloc()' which already uses
this flag.
Andrey Konovalov [Thu, 26 Nov 2020 15:01:57 +0000 (16:01 +0100)]
media: camss: Make use of V4L2_CAP_IO_MC
Implement mbus_code filtering for format enumeration.
Without this patch libcamera errors out with:
"ERROR V4L2 v4l2_videodevice.cpp:982 /dev/video0[cap]: Media bus code
filtering not supported by the device"
This is the second version of this change which handles the case of
several pixel formats corresponding to one media bus format correctly.
Signed-off-by: Andrey Konovalov <andrey.konovalov@linaro.org> Reviewed-by: Robert Foss <robert.foss@linaro.org> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
The assumption of "Each entry in formats[] table has unique mbus_code"
is valid for the RDI entities, but not for the PIX ones.
Reverting this patch and creating a new one which handles the PIX entities
correctly results in smaller and more straightforward patch than doing the
changes on top of the current version.
Signed-off-by: Andrey Konovalov <andrey.konovalov@linaro.org> Reviewed-by: Robert Foss <robert.foss@linaro.org> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Fabio Estevam [Wed, 25 Nov 2020 12:04:47 +0000 (13:04 +0100)]
media: fsl-viu: Use the ioread/write32be() accessors
Since commit 29d750686331 ("media: fsl-viu: allow building it with
COMPILE_TEST") there were several attempts to fix build failures on
PPC, MICROBLAZE and M68K.
However, such errors are still present as reported by kbuild robot.
Fix the problem by using the ioread/write32be() accessors, which build
fine on all platforms.
Reported-by: kernel test robot <lkp@intel.com> Suggested-by: Michael Ellerman <mpe@ellerman.id.au> Signed-off-by: Fabio Estevam <festevam@gmail.com> Reviewed-by: Michael Ellerman <mpe@ellerman.id.au> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Marek Szyprowski [Tue, 24 Nov 2020 08:41:28 +0000 (09:41 +0100)]
media: platform: exynos4-is: remove all references to physical addresses
This driver always operates on the DMA/IOVA addresses, so calling them
physical addresses is misleading, although when no IOMMU is used they
equal each other. Fix this by renaming all such entries to 'addr' and
adjusting comments.
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com> Reviewed-by: Sylwester Nawrocki <s.nawrocki@samsung.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Niklas Söderlund [Sun, 22 Nov 2020 16:36:37 +0000 (17:36 +0100)]
media: adv748x: Configure device when resuming from sleep
If the device is powered off (for example during system suspend to ram)
the devices loses its configuration, especially the slave i2c mappings
and other configuration set at probe time. This renders the device
unusable and the only way to recover is to unbind and rebind the device
to the driver to run the probe setup again.
Add an early resume callback that reinitializes the device and setup the
slave i2c address mappings and other probe time configuration.
Niklas Söderlund [Sun, 22 Nov 2020 16:36:36 +0000 (17:36 +0100)]
media: adv748x: csi2: Set virtual channel when device is reset
It's not enough to set the CSI-2 virtual channel for TXA and TXB during
probe: it also needs to be set when the device is reset. Move the virtual
channel selection to adv748x_reset() that is called during probe and
when the device needs to be reset.
Niklas Söderlund [Sun, 22 Nov 2020 16:36:35 +0000 (17:36 +0100)]
media: adv748x: afe: Select input port when device is reset
It's not enough to select the AFE input port during probe: it also needs
to be set when the device is reset. Move the port selection to
adv748x_reset() that is called during probe and when the device needs to
be reset.
Niklas Söderlund [Sun, 22 Nov 2020 16:30:48 +0000 (17:30 +0100)]
media: adv748x: Only set i2c addresses once during probe
During probe the i2c slave addresses are set two times in a row, once in
adv748x_initialise_clients() and once directly after in adv748x_reset().
Remove the call to adv748x_set_slave_addresses() in
adv748x_initialise_clients() as it's only called during probe while
adv748x_reset() is called during probe and normal operation.
Dan Carpenter [Tue, 17 Nov 2020 07:23:40 +0000 (08:23 +0100)]
media: saa7146: fix array overflow in vidioc_s_audio()
The "a->index" value comes from the user via the ioctl. The problem is
that the shift can wrap resulting in setting "mxb->cur_audinput" to an
invalid value, which later results in an array overflow.
Fixes: 6680427791c9 ("[media] mxb: fix audio handling") Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Jernej Skrabec [Tue, 10 Nov 2020 22:35:40 +0000 (23:35 +0100)]
media: cedrus: Add support for VP8 decoding
VP8 in Cedrus shares same engine as H264.
Note that it seems necessary to call bitstream parsing functions,
to parse frame header, otherwise decoded image is garbage. This is
contrary to what is driver supposed to do. However, values are not
really used, so this might be acceptable. It's possible that bitstream
parsing functions set some internal VPU state, which is later necessary
for proper decoding. Biggest suspect is "VP8 probs update" trigger.
Daniel Lee Kruse [Wed, 30 Sep 2020 03:36:35 +0000 (05:36 +0200)]
media: cx23885: add more quirks for reset DMA on some AMD IOMMU
On AMD Family 15h (Models 30h-3fh), I/O Memory Management Unit
RiSC engine sometimes stalls, requiring a reset.
As result, MythTV and w-scan won't scan channels on the AMD Kaveri
APU with the Hauppauge QuadHD TV tuner card.
For the solution I added the Input/Output Memory Management Unit's PCI
Identity of 0x1423 to the broken_dev_id[] array, which is used by
a quirks logic meant to fix similar problems with other AMD
chipsets.
Signed-off-by: Daniel Lee Kruse <daniel.lee.kruse@protonmail.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Hans Verkuil [Wed, 5 Aug 2020 10:57:21 +0000 (12:57 +0200)]
media: i2c: imx219: Selection compliance fixes
To comply with the intended usage of the V4L2 selection target when
used to retrieve a sensor image properties, adjust the rectangles
returned by the imx219 driver.
The top/left crop coordinates of the TGT_CROP rectangle were set to
(0, 0) instead of (8, 8) which is the offset from the larger physical
pixel array rectangle. This was also a mismatch with the default values
crop rectangle value, so this is corrected. Found with v4l2-compliance.
While at it, add V4L2_SEL_TGT_CROP_BOUNDS support: CROP_DEFAULT and
CROP_BOUNDS have the same size as the non-active pixels are not readable
using the selection API. Found with v4l2-compliance.
[reword commit message, use macros for pixel offsets]
Rob Herring [Tue, 17 Nov 2020 01:39:47 +0000 (02:39 +0100)]
media: dt-bindings: media: Use OF graph schema
Now that we have a graph schema, rework the media related schemas to
use it. Mostly this is adding a reference to graph.yaml and dropping
duplicate parts from schemas.
Cc: Maxime Ripard <mripard@kernel.org> Cc: Jacopo Mondi <jacopo@jmondi.org> Cc: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com> Signed-off-by: Rob Herring <robh@kernel.org> Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Jacopo Mondi [Fri, 20 Nov 2020 16:15:28 +0000 (17:15 +0100)]
media: max9271: Fix GPIO enable/disable
Fix GPIO enable/disable operations which wrongly read the 0x0f register
to obtain the current mask of the enabled lines instead of using
the correct 0x0e register.
Also fix access to bit 0 of the register which is marked as reserved.
There are no users left of this helper and as it implements an
undesirable and too simple behaviour that should instead be implemented
directly by drivers remove it to prevent future uses of it.
Suggested-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Niklas Söderlund [Wed, 25 Nov 2020 16:44:49 +0000 (17:44 +0100)]
media: rcar-vin: Rework CSI-2 firmware parsing
Rework the CSI-2 firmware parsing code to not use the soon to be
removed v4l2_async_notifier_parse_fwnode_endpoints_by_port() helper. The
change only aims to prepare for the removing of the old helper and there
are no functional change.
Niklas Söderlund [Wed, 25 Nov 2020 16:44:48 +0000 (17:44 +0100)]
media: rcar-vin: Use v4l2_async_subdev instead of fwnode_handle to match subdevices
In preparation of removing the usage of the old helper
v4l2_async_notifier_parse_fwnode_endpoints_by_port() use the
v4l2_async_subdev instead of fwnode_handle to match subdevices.
Niklas Söderlund [Wed, 25 Nov 2020 16:44:47 +0000 (17:44 +0100)]
media: rcar-vin: Rework parallel firmware parsing
Rework the parallel firmware parsing code to not use the soon to be
removed v4l2_async_notifier_parse_fwnode_endpoints_by_port() helper. The
change only aims to prepare for the removing of the old helper and there
are no functional change.
Niklas Söderlund [Wed, 25 Nov 2020 16:44:46 +0000 (17:44 +0100)]
media: rcar-vin: Only dynamically allocate v4l2_async_subdev
In preparation of removing the usage of the old helper
v4l2_async_notifier_parse_fwnode_endpoints_by_port() do not dynamically
allocate the whole structure containing the parameters for the parallel
interface, instead only allocate the v4l2_async_subdev structure. There
is no functional change.
Colin Ian King [Thu, 26 Nov 2020 11:49:32 +0000 (12:49 +0100)]
media: ov2740: fix dereference before null check on pointer nvm
Currently pointer nvm is being dereferenced before it is being null
checked. Fix this by moving the assignments of pointers client and
ov2740 so that are after the null check hence avoiding any potential
null pointer dereferences on pointer nvm.
Fixes: 5e6fd339b68d ("media: ov2740: allow OTP data access during streaming") Signed-off-by: Colin Ian King <colin.king@canonical.com> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Bingbu Cao [Fri, 27 Nov 2020 03:48:12 +0000 (04:48 +0100)]
media: ov9734: hold lock to check streaming state
The lock in ov9734 is used to protect the streaming state and
serialize the stream on and off callbacks, it should be hold before
checking the streaming state in ov9734_set_stream().
Reported-by: Tomasz Figa <tfiga@chromium.org> Reported-by: Sergey Senozhatsky <senozhatsky@chromium.org> Signed-off-by: Bingbu Cao <bingbu.cao@intel.com> Reviewed-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Sakari Ailus [Thu, 15 Aug 2019 07:36:49 +0000 (09:36 +0200)]
media: v4l: subdev: Set sd->devnode before registering the subdev
The subdev's video device node was only assigned after registering the
device node in the system. While it is unlikely that a driver needed to
use this field in handling system calls to its file handle, there remains
a slim chance the devnode field remains NULL while the driver expects to
find a video node there.
Assign the devnode field before registering the device, and assign it back
to NULL if the registration failed.
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Reviewed-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Sakari Ailus [Thu, 18 Jun 2020 11:12:51 +0000 (13:12 +0200)]
media: ccs: Request for "reset" GPIO
The DT bindings documented "reset-gpios" property but the driver never
made use of it. Instead it used a GPIO called "xshutdown", with apprently
wrong polarity.
Fix this by requesting "reset" GPIO with the right polarity first, and if
that fails, then request "xshutdown" GPIO with the old polarity. This way
it works for new users as expected while if someone, somewhere, depended
on "xshutdown" GPIO, that continues to work as well.
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Sakari Ailus [Tue, 11 Feb 2020 21:26:25 +0000 (22:26 +0100)]
media: dt-bindings: nokia,smia: Amend SMIA bindings with MIPI CCS support
Amend the existing SMIA bindings by adding MIPI CCS support, with separate
compatible strings for CCS 1.0 and CCS 1.1. Rename the old bindings
accordingly as CCS is the current standard.
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Reviewed-by: Rob Herring <robh@kernel.org> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
projects / thirdparty / kernel / linux.git / log
