media: videobuf2: move cache_hints handling to allocators
This moves cache hints handling from the videobuf2 core down
to the allocator's level, because allocators do the sync/flush
caches eventually and may take better decisions. Besides,
allocators already decide whether cache sync/flush should
be done or can be skipped. This patch moves the scattered
buffer cache sync logic to one common place.
V4L2 is not the perfect place to manage vb2 buffer cache hints.
It works for V4L2 users, but there are backends that use vb2 core
and don't use V4L2. Factor out buffer cache hints initialisation
and call it when we allocate vb2 buffer.
It would be less error prone if the default cache hints value
(we kzalloc() structs, so it's zeroed out by default) would be
to "always sync/flush" caches. Inverse and rename cache hints
flags.
With the new DMA API we need an extension of the videobuf2 API.
Previously, videobuf2 core would set the non-coherent DMA bit
in the vb2_queue dma_attr field (if user-space would pass a
corresponding memory hint); the vb2 core then would pass the
vb2_queue dma_attrs to the vb2 allocators. The vb2 allocator
would use the queue's dma_attr and the DMA API would allocate
either coherent or non-coherent memory.
But we cannot do this anymore, since there is no corresponding DMA
attr flag and, hence, there is no way for the allocator to become
aware of what type of allocation user-space has requested. So we
need to pass more context from videobuf2 core to the allocators.
Fix this by changing the call_ptr_memop() macro to pass the
vb2 pointer to the corresponding op callbacks.
Chad Fraleigh [Sun, 29 Aug 2021 23:51:27 +0000 (01:51 +0200)]
media: gspca: Limit frame size to sizeimage.
Limit frame size to what userland code expects. This can happen when
cameras, such as Kensington VideoCAM, use fixed sized transfer packets
which includes trailing junk in the final packet.
Signed-off-by: Chad Fraleigh <chadf@triularity.org> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Deborah Brouwer [Thu, 26 Aug 2021 03:44:11 +0000 (05:44 +0200)]
media: vivid: add signal-free time for cec message xfer
Currently, the vivid emulation of cec message transmission does not force
adapters to wait until the cec bus has been signal-free for a certain
number of bit periods before transmitting or re-transmitting a message.
Without enforcing the signal-free time requirements, adapters do not share
the bus very well and some messages are sent too quickly while other
messages are lost. By emulating a signal-free time and forcing adapters
to wait their turn to transmit, the vivid emulation of cec transmission
is much more reliable.
Ezequiel Garcia [Wed, 25 Aug 2021 00:23:37 +0000 (02:23 +0200)]
media: Request API is no longer experimental
The Request API is currently used and specified as part of the
Memory-to-memory Stateless Video Decoder Interface [1].
This can now be considered as non-experimental and stable, given
the decoder API has been used by products since a couple years,
supported by several drivers and userspace frameworks,
media: imx-jpeg: Fix the error handling path of 'mxc_jpeg_probe()'
A successful 'mxc_jpeg_attach_pm_domains()' call should be balanced by a
corresponding 'mxc_jpeg_detach_pm_domains()' call in the error handling
path of the probe, as already done in the remove function.
media: mtk-vpu: Fix a resource leak in the error handling path of 'mtk_vpu_probe()'
A successful 'clk_prepare()' call should be balanced by a corresponding
'clk_unprepare()' call in the error handling path of the probe, as already
done in the remove function.
Tom Rix [Wed, 18 Aug 2021 19:41:05 +0000 (21:41 +0200)]
media: camss: vfe: simplify vfe_get_wm_sizes()
Static analysis reports this representative problem
camss-vfe-4-1.c:333: The result of the left shift is undefined because
the left operand is negative
reg |= (height - 1) << 4;
~~~~~~~~~ ^
The is a false positive. height is set in vfe_get_wm_sizes() which
has a switch statement without a default.
Reviewing the switch, the cases contain redundant assignments.
So simplify to assignments.
Signed-off-by: Tom Rix <trix@redhat.com> 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>
Tom Rix [Thu, 12 Aug 2021 17:00:43 +0000 (19:00 +0200)]
media: TDA1997x: handle short reads of hdmi info frame.
Static analysis reports this representative problem
tda1997x.c:1939: warning: 7th function call argument is an uninitialized
value
The 7th argument is buffer[0], which is set in the earlier call to
io_readn(). When io_readn() call to io_read() fails with the first
read, buffer[0] is not set and 0 is returned and stored in len.
The later call to hdmi_infoframe_unpack()'s size parameter is the
static size of buffer, always 40, so a short read is not caught
in hdmi_infoframe_unpacks()'s checking. The variable len should be
used instead.
Zero initialize buffer to 0 so it is in a known start state.
Fixes: 9ac0038db9a7 ("media: i2c: Add TDA1997x HDMI receiver driver") Signed-off-by: Tom Rix <trix@redhat.com> Reviewed-by: Tim Harvey <tharvey@gateworks.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Tuo Li [Thu, 5 Aug 2021 07:55:35 +0000 (09:55 +0200)]
media: s5p-mfc: fix possible null-pointer dereference in s5p_mfc_probe()
The variable pdev is assigned to dev->plat_dev, and dev->plat_dev is
checked in:
if (!dev->plat_dev)
This indicates both dev->plat_dev and pdev can be NULL. If so, the
function dev_err() is called to print error information.
dev_err(&pdev->dev, "No platform data specified\n");
However, &pdev->dev is an illegal address, and it is dereferenced in
dev_err().
To fix this possible null-pointer dereference, replace dev_err() with
mfc_err().
Reported-by: TOTE Robot <oslab@tsinghua.edu.cn> Signed-off-by: Tuo Li <islituo@gmail.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Zev Weiss [Thu, 17 Jun 2021 22:02:29 +0000 (00:02 +0200)]
media: aspeed-video: ignore interrupts that aren't enabled
As partially addressed in commit 65d270acb2d6 ("media: aspeed: clear
garbage interrupts"), the ASpeed video engine sometimes asserts
interrupts that the driver hasn't enabled. In addition to the
CAPTURE_COMPLETE and FRAME_COMPLETE interrupts dealt with in that
patch, COMP_READY has also been observed. Instead of playing
whack-a-mole with each one individually, we can instead just blanket
ignore everything we haven't explicitly enabled.
Signed-off-by: Zev Weiss <zev@bewilderbeest.net> Tested-by: Lei YU <yulei.sh@bytedance.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Dan Carpenter [Mon, 7 Jun 2021 15:23:48 +0000 (17:23 +0200)]
media: firewire: firedtv-avc: fix a buffer overflow in avc_ca_pmt()
The bounds checking in avc_ca_pmt() is not strict enough. It should
be checking "read_pos + 4" because it's reading 5 bytes. If the
"es_info_length" is non-zero then it reads a 6th byte so there needs to
be an additional check for that.
I also added checks for the "write_pos". I don't think these are
required because "read_pos" and "write_pos" are tied together so
checking one ought to be enough. But they make the code easier to
understand for me. The check on write_pos is:
if (write_pos + 4 >= sizeof(c->operand) - 4) {
The first "+ 4" is because we're writing 5 bytes and the last " - 4"
is to leave space for the CRC.
The other problem is that "length" can be invalid. It comes from
"data_length" in fdtv_ca_pmt().
Cc: stable@vger.kernel.org Reported-by: Luo Likang <luolikang@nsfocus.com> 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>
lijian [Fri, 28 May 2021 08:47:40 +0000 (10:47 +0200)]
media: via-camera: deleted these redundant semicolons
Macros should not use a trailing semicolon，and it was used
the semicolon after macro cam_dbg called in viacam_vb2_prepare,
so deleted these redundant semicolons.
Liu Shixin [Thu, 27 May 2021 13:17:01 +0000 (15:17 +0200)]
media: pvrusb2: Replaced simple_strtol() with kstrtoint()
It looks like that memcpy() is a superfluous operation in
parse_token()/parse_mtoken(). Simple these two functions and
use kstrtoint() instead of simple_strtol() to avoid data
overflow.
Signed-off-by: Liu Shixin <liushixin2@huawei.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Ricardo Ribalda [Fri, 18 Jun 2021 12:29:16 +0000 (14:29 +0200)]
media: uvcvideo: Check controls flags before accessing them
We can figure out if reading/writing a set of controls can fail without
accessing them by checking their flags.
This way we can honor the API closer:
If an error is found when validating the list of controls passed with
VIDIOC_G_EXT_CTRLS, then error_idx shall be set to ctrls->count to
indicate to userspace that no actual hardware was touched.
Fixes v4l2-compliance:
Control ioctls (Input 0):
warn: v4l2-test-controls.cpp(765): g_ext_ctrls(0) invalid error_idx 0
fail: v4l2-test-controls.cpp(645): invalid error index write only control
test VIDIOC_G/S/TRY_EXT_CTRLS: FAIL
Ricardo Ribalda [Fri, 18 Jun 2021 12:29:14 +0000 (14:29 +0200)]
media: uvcvideo: Increase the size of UVC_METADATA_BUF_SIZE
Hans has discovered that in his test device, for the H264 format
bytesused goes up to about 570, for YUYV it will actually go up
to a bit over 5000 bytes, and for MJPG up to about 2706 bytes.
We should also, according to V4L2_META_FMT_UVC docs, drop headers when
the buffer is full.
Ricardo Ribalda [Fri, 18 Jun 2021 12:29:11 +0000 (14:29 +0200)]
media: uvcvideo: Add support for V4L2_CTRL_TYPE_CTRL_CLASS
Create all the class controls for the device defined controls.
Fixes v4l2-compliance:
Control ioctls (Input 0):
fail: v4l2-test-controls.cpp(216): missing control class for class 00980000
fail: v4l2-test-controls.cpp(216): missing control tclass for class 009a0000
test VIDIOC_QUERY_EXT_CTRL/QUERYMENU: FAIL
Ricardo Ribalda [Fri, 18 Jun 2021 12:29:09 +0000 (14:29 +0200)]
media: uvcvideo: Return -EIO for control errors
The device is doing something unexpected with the control. Either because
the protocol is not properly implemented or there has been a HW error.
Fixes v4l2-compliance:
Control ioctls (Input 0):
fail: v4l2-test-controls.cpp(448): s_ctrl returned an error (22)
test VIDIOC_G/S_CTRL: FAIL
fail: v4l2-test-controls.cpp(698): s_ext_ctrls returned an error (22)
test VIDIOC_G/S/TRY_EXT_CTRLS: FAIL
Ricardo Ribalda [Fri, 18 Jun 2021 12:29:06 +0000 (14:29 +0200)]
media: v4l2-ioctl: S_CTRL output the right value
If the driver does not implement s_ctrl, but it does implement
s_ext_ctrls, we convert the call.
When that happens we have also to convert back the response from
s_ext_ctrls.
Fixes v4l2_compliance:
Control ioctls (Input 0):
fail: v4l2-test-controls.cpp(411): returned control value out of range
fail: v4l2-test-controls.cpp(507): invalid control 00980900
test VIDIOC_G/S_CTRL: FAIL
Fixes: 35ea11ff8471 ("V4L/DVB (8430): videodev: move some functions from v4l2-dev.h to v4l2-common.h or v4l2-ioctl.h") Reviewed-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Ricardo Ribalda <ribalda@chromium.org> Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Ricardo Ribalda [Fri, 18 Jun 2021 12:29:03 +0000 (14:29 +0200)]
media: v4l2-ioctl: Fix check_ext_ctrls
Drivers that do not use the ctrl-framework use this function instead.
Fix the following issues:
- Do not check for multiple classes when getting the DEF_VAL.
- Return -EINVAL for request_api calls
- Default value cannot be changed, return EINVAL as soon as possible.
- Return the right error_idx
[If an error is found when validating the list of controls passed with
VIDIOC_G_EXT_CTRLS, then error_idx shall be set to ctrls->count to
indicate to userspace that no actual hardware was touched.
It would have been much nicer of course if error_idx could point to the
control index that failed the validation, but sadly that's not how the
API was designed.]
Fixes v4l2-compliance:
Control ioctls (Input 0):
warn: v4l2-test-controls.cpp(834): error_idx should be equal to count
warn: v4l2-test-controls.cpp(855): error_idx should be equal to count
fail: v4l2-test-controls.cpp(813): doioctl(node, VIDIOC_G_EXT_CTRLS, &ctrls)
test VIDIOC_G/S/TRY_EXT_CTRLS: FAIL
Buffer ioctls (Input 0):
fail: v4l2-test-buffers.cpp(1994): ret != EINVAL && ret != EBADR && ret != ENOTTY
test Requests: FAIL
David Plowman [Mon, 16 Aug 2021 11:39:09 +0000 (13:39 +0200)]
media: v4l2-ctrls: Document V4L2_CID_NOTIFY_GAINS control
Add documentation for the V4L2_CID_NOTIFY_GAINS control.
This control is required by sensors that need to know what colour
gains will be applied to pixels by downstream processing (such as by
an ISP), though the sensor does not apply these gains itself.
Signed-off-by: David Plowman <david.plowman@raspberrypi.com> Reviewed-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>
David Plowman [Mon, 16 Aug 2021 11:39:08 +0000 (13:39 +0200)]
media: v4l2-ctrls: Add V4L2_CID_NOTIFY_GAINS control
We add a new control V4L2_CID_NOTIFY_GAINS which allows the sensor to
be notified what gains will be applied to the different colour
channels by subsequent processing (such as by an ISP), even though the
sensor will not apply any of these gains itself.
For Bayer sensors this will be an array control taking 4 values which
are the 4 gains arranged in the fixed order B, Gb, Gr and R,
irrespective of the exact Bayer order of the sensor itself. The use of
an array makes it straightforward to extend this control to non-Bayer
sensors (for example, sensors with an RGBW pattern) in future.
The units are in all cases linear with the default value indicating a
gain of exactly 1.0. For example, if the default value were reported as
128 then the value 192 would represent a gain of exactly 1.5.
Signed-off-by: David Plowman <david.plowman@raspberrypi.com> Reviewed-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>
Hsin-Yi Wang [Mon, 5 Jul 2021 08:17:24 +0000 (10:17 +0200)]
media: ov8856: Set default mbus format but allow caller to alter
Setting the value of V_WIN_OFF (0x3818) from 0x02 to 0x01 to use GRBG
format still results in wrong color output if data is tuned in BGGR mode
before.
Set default mbus format for the supported modes, but allow the caller of
set(get)_fmt to change the bayer format between BGGR and GRBG.
Set the default mbus format for 3264x2448 (and 1632x1224) to BGGR as the
data sheet states the value of this reg should be 0x02 by default.
If new modes are added in the future, they can add the
mipi_data_mbus_{format} settings into bayer_offset_configs to adjust their
offset regs.
Fixes: 2984b0ddd557 ("media: ov8856: Configure sensor for GRBG Bayer for all modes") Signed-off-by: Hsin-Yi Wang <hsinyi@chromium.org> Reviewed-by: Robert Foss <robert.foss@linaro.org> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Arec Kao [Tue, 17 Aug 2021 04:46:07 +0000 (06:46 +0200)]
media: Add sensor driver support for the ov13b10 camera.
This driver supports following features:
- phase detection auto focus (PDAF)
- manual exposure and analog/digital gain control
- vblank/hblank control
- test pattern
- image vertical flip and horizontal mirror control
- 4208x3120 at 30FPS
- 2080x1170 at 60FPS
Signed-off-by: Arec Kao <arec.kao@intel.com> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Sakari Ailus [Mon, 16 Aug 2021 13:08:59 +0000 (15:08 +0200)]
media: imx258: Fix getting clock frequency
Obtain the clock frequency by reading the clock-frequency property if
there's no clock.
Fixes: 9fda25332c4b ("media: i2c: imx258: get clock from device properties and enable it via runtime PM") Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
There is a wrong comparison of the total size of the loaded firmware
css->fw->size with the size of a pointer to struct imgu_fw_header.
Turn binary_header into a flexible-array member[1][2], use the
struct_size() helper and fix the wrong size comparison. Notice
that the loaded firmware needs to contain at least one 'struct
imgu_fw_info' item in the binary_header[] array.
with binary_header declared as a flexible-array member is equivalent
to
"css->fw->size < sizeof(struct imgu_fw_header)"
with binary_header declared as a one-element array (as in the original
code).
The replacement of the one-element array with a flexible-array member
also helps with the ongoing efforts to globally enable -Warray-bounds
and get us closer to being able to tighten the FORTIFY_SOURCE routines
on memcpy().
media: ipu3.rst: Improve header formatting on tables
Use the header-rows option of the flat-table directive in order to have
the first row displayed as a header. Also capitalize these headers.
These changes make the tables easier to read.
Signed-off-by: Nícolas F. R. A. Prado <nfraprado@collabora.com> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
there is still uninitialized field sd_format of struct stm32_dcmi *dcmi.
If an interrupt occurs in the interval between the installation of the
interrupt handler and the initialization of this field, NULL pointer
dereference happens.
This field is dereferenced in the handler function without any check:
The patch moves interrupt handler installation
after initialization of the sd_format field that happens in
dcmi_graph_notify_complete() via dcmi_set_default_fmt().
Found by Linux Driver Verification project (linuxtesting.org).
media: staging: atomisp: fix the uninitialized use in gc2235_detect()
Inside function gc2235_detect(), variable "low" could be uninitialized
if ov5693_read_reg() returns error, however, it affects the value of
variable "id". The "id" is used in the later if statement, which is
potentially unsafe.
Evgeny Novikov [Tue, 10 Aug 2021 16:29:43 +0000 (18:29 +0200)]
media: atomisp: Fix error handling in probe
There were several issues with handling errors in lm3554_probe():
- Probe did not set the error code when v4l2_ctrl_handler_init() failed.
- It intermixed gotos for handling errors of v4l2_ctrl_handler_init()
and media_entity_pads_init().
- It did not set the error code for failures of v4l2_ctrl_new_custom().
- Probe did not free resources in case of failures of
atomisp_register_i2c_module().
The patch fixes all these issues.
Found by Linux Driver Verification project (linuxtesting.org).
Link: https://lore.kernel.org/linux-media/20210810162943.19852-1-novikov@ispras.ru Signed-off-by: Evgeny Novikov <novikov@ispras.ru> Reviewed-by: Dan Carpenter <dan.carpenter@oracle.com> Acked-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
projects / thirdparty / kernel / linux.git / log
