This is true, but I reviewed the surrounding code and it appears
harmless. Anyway, let's verify that "rc_proto" is valid as a kernel
hardening measure.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Sean Young <sean@mess.org> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
media: venus: core: add shutdown callback for venus
After the SMMU translation is disabled in the
arm-smmu shutdown callback during reboot, if
any subsystem are still alive then IOVAs they
are using will become PAs on bus, which may
lead to crash.
So implemented shutdown callback, which detach iommu maps.
media: venus: handle use after free for iommu_map/iommu_unmap
In concurrency usecase and reboot scenario we are seeing muliple
crashes related to iommu_map/iommu_unamp of core->fw.iommu_domain.
In one case we are seeing "Unable to handle kernel NULL pointer
dereference at virtual address 0000000000000008" crash, this is
because of core->fw.iommu_domain in venus_firmware_deinit() and
trying to map in venus_boot() during venus_sys_error_handler()
In second case we are seeing "Unable to handle kernel paging request
at virtual address 006b6b6b6b6b6b9b" crash, this is because of
unmapping iommu domain which is already unmapped.
media: venus: core: handle race condititon for core ops
For core ops we are having only write protect but there
is no read protect, because of this in multithreading
and concurrency, one CPU core is reading without wait
which is causing the NULL pointer dereferece crash.
one such scenario is as show below, where in one CPU
core, core->ops becoming NULL and in another CPU core
calling core->ops->session_init().
media: venus: fix calculating mbps in calculate_inst_freq()
Currently in calculate_inst_freq(), video driver is calculating
macro blocks per frame instead of macro blocks per second(mpbs).
Which results frequency is always setting to lower frequency (150MHz)
as per frequency table for sc7180. Hence the playback is not smooth.
Corrected this by correcting the mbps calculation in calculate_inst_freq().
media: venus: put dummy vote on video-mem path after last session release
As per current implementation, video driver is unvoting "videom-mem" path
for last video session during vdec_session_release().
While video playback when we try to suspend device, we see video clock
warnings since votes are already removed during vdec_session_release().
corrected this by putting dummy vote on "video-mem" after last video
session release and unvoting it during suspend.
suspend")
Fixes: 07f8f22a33a9e ("media: venus: core: remove CNOC voting while device Signed-off-by: Mansur Alisha Shaik <mansur@codeaurora.org> Reviewed-by: Stephen Boyd <swboyd@chromium.org> Signed-off-by: Stanimir Varbanov <stanimir.varbanov@linaro.org> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Currently video driver is voting for venus0-ebi path during buffer
processing with an average bandwidth of all the instances and
unvoting during session release.
While video streaming when we try to do XO-SD using the command
"echo mem > /sys/power/state command" , device is not entering
to suspend state and from interconnect summary seeing votes for venus0-ebi
Corrected this by voting for venus0-ebi path in venus_runtime_resume()
and unvote during venus_runtime_suspend().
suspend")
Fixes: 07f8f22a33a9e ("media: venus: core: remove CNOC voting while device Signed-off-by: Mansur Alisha Shaik <mansur@codeaurora.org> Reviewed-by: Stephen Boyd <swboyd@chromium.org> Signed-off-by: Stanimir Varbanov <stanimir.varbanov@linaro.org> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
media: venus: core: change clk enable and disable order in resume and suspend
Currently video driver is voting after clk enable and un voting
before clk disable. This is incorrect, video driver should vote
before clk enable and unvote after clk disable.
Corrected this by changing the order of clk enable and clk disable.
suspend")
Fixes: 07f8f22a33a9e ("media: venus: core: remove CNOC voting while device Signed-off-by: Mansur Alisha Shaik <mansur@codeaurora.org> Reviewed-by: Stephen Boyd <swboyd@chromium.org> Signed-off-by: Stanimir Varbanov <stanimir.varbanov@linaro.org> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
media: venus: vdec: return parsed crop information from stream
Per the stateful codec specification, VIDIOC_G_SELECTION with a target
of V4L2_SEL_TGT_COMPOSE is supposed to return the crop area of capture
buffers containing the decoded frame. Until now the driver did not get
that information from the firmware and just returned the dimensions of
CAPTURE buffers.
The firmware unfortunately does not always provide the crop information
from the stream ; also make sure to detect when that happens and
fallback to providing the coded size in these cases.
Yu Kuai [Wed, 4 Nov 2020 12:41:36 +0000 (13:41 +0100)]
media: platform: add missing put_device() call in mtk_jpeg_probe() and mtk_jpeg_remove()
if mtk_jpeg_clk_init() succeed, mtk_jpeg_probe() and mtk_jpeg_remove()
doesn't have a corresponding put_device(). Thus add a new helper
mtk_jpeg_clk_release() to fix it.
Ezequiel Garcia [Wed, 4 Nov 2020 17:43:11 +0000 (18:43 +0100)]
media: coda: Add a V4L2 user for control error macroblocks count
To avoid potentially overflowing the kernel logs in the case
of corrupted streams, this commit replaces an error message with
a per-stream counter to be read through a driver-specific
control.
Applications can read the per-stream accumulated
error macroblocks count.
The old error message is replaced by a rate-limited debug message.
Ezequiel Garcia [Tue, 3 Nov 2020 21:32:37 +0000 (22:32 +0100)]
media: coda: coda_buffer_meta housekeeping fix
It's possible that the VPU was initialized using just one buffer,
containing only codec headers.
In this case, right after the initialization and after updating
the FIFO read pointer, we need to iterate through all the coda_buffer_meta
and release any metas that have been already used by the VPU.
This issue is affecting indirectly the bitstream buffer fill
threshold, which depends on the meta end position of the first
queued meta, which is passed to coda_bitstream_can_fetch_past().
Without this fix, it's possible that for certain videos, the
bitstream buffer level is not filled properly, resulting in a PIC_RUN
timeout.
Reported-by: Benjamin Bara <benjamin.bara@skidata.com> Signed-off-by: Ezequiel Garcia <ezequiel@collabora.com> Reviewed-by: Philipp Zabel <p.zabel@pengutronix.de> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Dafna Hirschfeld [Fri, 30 Oct 2020 13:46:09 +0000 (14:46 +0100)]
media: staging: rkisp1: isp: set metadata pads to MEDIA_BUS_FMT_METADATA_FIXED
Set the code of the metadata pads of the isp entity to
MEDIA_BUS_FMT_METADATA_FIXED and set the width and
height of their formats to 0. This solves the TODO
item:
"Fix pad format size for statistics and parameters entities."
Signed-off-by: Dafna Hirschfeld <dafna.hirschfeld@collabora.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>
Dafna Hirschfeld [Fri, 30 Oct 2020 13:46:08 +0000 (14:46 +0100)]
media: uapi: add MEDIA_BUS_FMT_METADATA_FIXED media bus format.
MEDIA_BUS_FMT_METADATA_FIXED should be used when
the same driver handles both sides of the link and
the bus format is a fixed metadata format that is
not configurable from userspace.
The width and height will be set to 0 for this format.
Signed-off-by: Dafna Hirschfeld <dafna.hirschfeld@collabora.com> Acked-by: Helen Koike <helen.koike@collabora.com> Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Zhang Qilong [Mon, 2 Nov 2020 14:26:22 +0000 (15:26 +0100)]
media: cedrus: fix reference leak in cedrus_start_streaming
pm_runtime_get_sync will increment pm usage counter even it
failed. Forgetting to pm_runtime_put_noidle will result in
reference leak in cedrus_start_streaming. We should fix it.
Zhang Qilong [Mon, 2 Nov 2020 14:28:52 +0000 (15:28 +0100)]
media: staging: rkisp1: cap: fix runtime PM imbalance on error
pm_runtime_get_sync will increment pm usage counter even it
failed. Forgetting to pm_runtime_put_noidle will result in
reference imbalance in rkisp1_vb2_start_streaming, so we
should fix it.
Hsin-Yi Wang [Thu, 3 Sep 2020 05:48:33 +0000 (07:48 +0200)]
media: mtk-vcodec: remove allocated dma_parms
Commit 9495b7e92f71 ("driver core: platform: Initialize dma_parms for
platform devices") included dma_parms in platform_device. There's no need
to allocate again.
Fixes: 13483fc2f20f ("media: mtk-vcodec: set dma max segment size") Suggested-by: Robin Murphy <robin.murphy@arm.com> Signed-off-by: Hsin-Yi Wang <hsinyi@chromium.org> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
isif_probe() invokes iounmap() on error handling paths, but it does not
reset the global state. So, later it can invoke iounmap() even when
ioremap() fails. This is the case also for isif_remove(). The patch
resets the global state after invoking iounmap() to avoid this.
Found by Linux Driver Verification project (linuxtesting.org).
Hugues Fruchet [Tue, 13 Oct 2020 09:02:23 +0000 (11:02 +0200)]
media: ov5640: fix support of BT656 bus mode
Fix PCLK polarity not being taken into account.
Add comments about BT656 register control.
Remove useless ov5640_set_stream_bt656() function.
Refine comments about MIPI IO register control.
Sakari Ailus [Tue, 13 Oct 2020 14:02:27 +0000 (16:02 +0200)]
media: Documentation: v4l: Document required controls on camera sensors
Raw camera sensors connected to parallel or CSI-2 busses need to implement
V4L2_CID_LINK_FREQ control so receiver drivers can configure the hardware
accordingly.
Document this. Also say V4L2_CID_PIXEL_RATE is recommended for such
hardware.
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Sakari Ailus [Tue, 13 Oct 2020 13:11:37 +0000 (15:11 +0200)]
media: Documentation: v4l: Fix language
Remove extra "is" from sentence explaining V4L2_CID_PIXEL_RATE may be used
to tell the pixel rate used by a transmitter driver. Also call
V4L2_CID_PIXEL_RATE a control.
Sakari Ailus [Tue, 13 Oct 2020 12:12:02 +0000 (14:12 +0200)]
media: ipu3-cio2: Check receved the size against payload size, not buffer size
Compare the received size of the payload size, not the allocated size of
the buffer that is page aligned. This way also images that aren't aligned
to page size are not warned about.
Replace dmaengine_terminate_all() by dmaengine_terminate_sync()
to ensure that all pending dma operations are really finished.
This is not ensured by dmaengine_terminate_all() and this API
is deprecated, so better to use the _sync() variant.
Signed-off-by: Hugues Fruchet <hugues.fruchet@st.com> Reviewed-by: Philippe CORNU <philippe.cornu@st.com> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Sakari Ailus [Tue, 5 Jul 2016 09:41:15 +0000 (11:41 +0200)]
media: v4l: subdev: Improve link format validation debug messages
The existing link format validation failure debug message in media-entity.c
helped to pinpoint the point of failure but provided no additional
information what's wrong. Tell the user exactly why the validation failed.
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Lad Prabhakar [Fri, 2 Oct 2020 22:23:22 +0000 (00:23 +0200)]
media: i2c: ov772x: Add support for BT.656 mode
Add support to read the bus-type for V4L2_MBUS_BT656 and enable BT.656
mode in the sensor if needed.
For backward compatibility with older DTS where the bus-type property was
not mandatory, assume V4L2_MBUS_PARALLEL as it was the only supported bus
at the time. v4l2_fwnode_endpoint_alloc_parse() will not fail if
'bus-type' is not specified.
media: rcar-csi2: Switch to using fwnode instead of OF
Use the fwnode_graph_get_endpoint_by_id() interface instead of
of_graph_get_endpoint_by_regs() to fetch the fwnode. This saves
translating between a device_node and fwnode_handle.
Sakari Ailus [Wed, 30 Sep 2020 14:34:43 +0000 (16:34 +0200)]
media: v4l2-fwnode: Say it's fine to use v4l2_fwnode_endpoint_parse
Earlier it was expected that there would be more variable size endpoint
properties and that most if not all drivers would need them. For that
reason it was expected also that v4l2_fwnode_endpoint_parse would no
longer be needed.
What actually happened that not all drivers require "link-frequencies",
the only variable size media endpoint property without a small upper
limit. Therefore drivers that do not need that information are fine using
v4l2_fwnode_endpoint_parse. So don't tell drivers to use
v4l2_fwnode_endpoint_alloc_parse in all cases.
Sakari Ailus [Wed, 30 Sep 2020 12:28:54 +0000 (14:28 +0200)]
media: v4l2-fwnode: Don't zero parts of struct v4l2_fwnode_endpoint anymore
Don't zero parts of the vep argument to v4l2_fwnode_endpoint_parse()
anymore as this can no longer be done while still supporting defaults on
multiple bus types.
Sakari Ailus [Wed, 30 Sep 2020 12:08:03 +0000 (14:08 +0200)]
media: v4l2-fwnode: v4l2_fwnode_endpoint_parse caller must init vep argument
Document that the caller of v4l2_fwnode_endpoint_parse() must init the
fields of struct v4l2_fwnode_endpoint (vep argument) fields.
It used to be that the fields were zeroed by v4l2_fwnode_endpoint_parse
when bus type was set to V4L2_MBUS_UNKNOWN but with recent changes (Fixes:
line below) that no longer makes sense.
Fixes: bb4bba9232fc ("media: v4l2-fwnode: Make bus configuration a struct") Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Reviewed-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se> Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Sakari Ailus [Wed, 30 Sep 2020 11:49:15 +0000 (13:49 +0200)]
media: adv748x: Zero entire struct v4l2_fwnode_endpoint
The v4l2_fwnode_parse_endpoint() function can make use of defaults in
multiple bus types. To use this feature, all callers must zero the rest of
the fields of this struct, too. All other drivers appear to do that
already apart from this one.
Philipp Zabel [Wed, 23 Sep 2020 08:21:12 +0000 (10:21 +0200)]
media: dt-bindings: convert CODA VPU bindings to yaml
Convert to YAML and add generic IP core compatibles "cnm,codadx6",
"cnm,codahx4", "cnm,coda7541", and "cnm,coda960" in addition to the SoC
specific compatibles. The new generic compatibles are already used in
the SoC device trees and replace the free form comments. For example:
- compatible : should be "fsl,<chip>-src" for i.MX SoCs:
(a) "fsl,imx27-vpu" for CodaDx6 present in i.MX27
This allows to properly specify the secondary JPEG unit interrupt that
is only present on cnm,coda960.
Also add the missing "fsl,imx6dl-vpu", "cnm,coda960" compatible.
The AXI bus connection to the internal SRAM is different between i.MX6Q
and i.MX6DL, which requires the driver to load a different firmware
depending on the SoC.
Further, specify the power-domain property for i.MX6 and change the
clock order from "ahb", "per" to "per", "ahb". This order is currently
used in all SoC device trees.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de> Reviewed-by: Rob Herring <robh@kernel.org> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Document endpoint properties for the parallel bus type and
add them to the example.
Specify a few constraints:
- If the bus type is BT.656 no hsync or vsync polarities can be
specified.
- If the bus width is 10 bits, not data-shift can be applied.
media: dt-bindings: media: ov772x: Make bus-type mandatory
In order to establish required properties based on the selected
bus type, make the 'bus-type' property mandatory. As this change
documents an endpoint property, also document the 'remote-endpoint'
one now that the 'endpoint' schema has been expanded.
Binary compatibility with existing DTB is kept as the driver does not
enforce the property to be present, and shall fall-back to default
parallel bus configuration, which was the only supported bus type, if
the property is not specified.
Eugen Hristev [Tue, 15 Sep 2020 09:04:42 +0000 (11:04 +0200)]
media: imx274: add support for sensor mode6, 1280x540
Add support for the mode 6 for the sensor, this mode uses
3/8 subsampling and 3 horizontal binning.
Aspect ratio is changed.
Split the bin_ratio variable into two parts, one for
width and one for height, as the ratio is no longer preserved
when doing subsampling in this mode.
Signed-off-by: Eugen Hristev <eugen.hristev@microchip.com> Reviewed-by: Luca Ceresoli <luca@lucaceresoli.net> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
As per Sony IMX274 Y_OUT_SIZE should be the height of effective
image output from the sensor which are the actual total lines
sent over MIPI CSI to receiver.
So, Y_OUT_SIZE should be same as crop height and this patch fixes it.
media: v4l2-fwnode: Return -EINVAL for invalid bus-type
Return -EINVAL if invalid bus-type is detected while parsing endpoints.
Fixes: 26c1126c9b56 ("media: v4l: fwnode: Use media bus type for bus parser selection") Signed-off-by: Lad Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Colin Ian King [Tue, 20 Oct 2020 14:46:55 +0000 (16:46 +0200)]
media: staging: rkisp1: rsz: make const array static, makes object smaller
Don't populate the const array dev_names on the stack but instead it
static. Makes the object code smaller by 15 bytes.
Before:
text data bss dec hex filename
17091 2648 64 19803 4d5b media/rkisp1/rkisp1-resizer.o
After:
text data bss dec hex filename
17012 2712 64 19788 4d4c media/rkisp1/rkisp1-resizer.o
(gcc version 10.2.0)
Signed-off-by: Colin Ian King <colin.king@canonical.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>