Philipp Zabel [Mon, 8 Apr 2019 12:32:49 +0000 (08:32 -0400)]
media: coda: clear error return value before picture run
The error return value is not written by some firmware codecs, such as
MPEG-2 decode on CodaHx4. Clear the error return value before starting
the picture run to avoid misinterpreting unrelated values returned by
sequence initialization as error return value.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Philipp Zabel [Mon, 8 Apr 2019 12:32:47 +0000 (08:32 -0400)]
media: coda: set codec earlier
The chosen codec depends on the coded format, which is known as soon as
the S_FMT call on the coded queue. This allows to use the codec in
callbacks that may be called before start_streaming, such as buf_queue.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Hans Verkuil [Mon, 19 Nov 2018 11:09:00 +0000 (06:09 -0500)]
media: vb2: add waiting_in_dqbuf flag
Calling VIDIOC_DQBUF can release the core serialization lock pointed to
by vb2_queue->lock if it has to wait for a new buffer to arrive.
However, if userspace dup()ped the video device filehandle, then it is
possible to read or call DQBUF from two filehandles at the same time.
It is also possible to call REQBUFS from one filehandle while the other
is waiting for a buffer. This will remove all the buffers and reallocate
new ones. Removing all the buffers isn't the problem here (that's already
handled correctly by DQBUF), but the reallocating part is: DQBUF isn't
aware that the buffers have changed.
This is fixed by setting a flag whenever the lock is released while waiting
for a buffer to arrive. And checking the flag where needed so we can return
-EBUSY.
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl> Reported-by: Syzbot <syzbot+4180ff9ca6810b06c1e9@syzkaller.appspotmail.com> Reviewed-by: Tomasz Figa <tfiga@chromium.org> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Dan Carpenter [Mon, 8 Apr 2019 09:52:38 +0000 (05:52 -0400)]
media: pvrusb2: Prevent a buffer overflow
The ctrl_check_input() function is called from pvr2_ctrl_range_check().
It's supposed to validate user supplied input and return true or false
depending on whether the input is valid or not. The problem is that
negative shifts or shifts greater than 31 are undefined in C. In
practice with GCC they result in shift wrapping so this function returns
true for some inputs which are not valid and this could result in a
buffer overflow:
drivers/media/usb/pvrusb2/pvrusb2-ctrl.c:205 pvr2_ctrl_get_valname()
warn: uncapped user index 'names[val]'
The cptr->hdw->input_allowed_mask mask is configured in pvr2_hdw_create()
and the highest valid bit is BIT(4).
Fixes: 7fb20fa38caa ("V4L/DVB (7299): pvrusb2: Improve logic which handles input choice availability") 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+samsung@kernel.org>
Currently, MEDIA_IOC_G_TOPOLOGY ioctl on cedrus fails due to incorrect
initialization order. Fix that by moving video_register_device() before
v4l2_m2m_register_media_controller() and while at it, fix error path.
Reported-by: Jonas Karlman <jonas@kwiboo.se> Signed-off-by: Jernej Skrabec <jernej.skrabec@siol.net> Acked-by: Paul Kocialkowski <paul.kocialkowski@bootlin.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Alexander Potapenko [Thu, 4 Apr 2019 14:56:46 +0000 (10:56 -0400)]
media: vivid: use vfree() instead of kfree() for dev->bitmap_cap
syzkaller reported crashes on kfree() called from
vivid_vid_cap_s_selection(). This looks like a simple typo, as
dev->bitmap_cap is allocated with vzalloc() throughout the file.
Fixes: ef834f7836ec0 ("[media] vivid: add the video capture and output
parts")
Signed-off-by: Alexander Potapenko <glider@google.com> Reported-by: Syzbot <syzbot+6c0effb5877f6b0344e2@syzkaller.appspotmail.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Hans Verkuil [Mon, 1 Apr 2019 12:41:58 +0000 (08:41 -0400)]
media: videobuf2-v4l2.c: move up STATE_DEQUEUED check
If a buffer is queued to a request, followed by an attempt to queue
the same buffer again, then the second qbuf returns an error since
the buffer is not in the DEQUEUED state anymore.
However, before it gets to that check it executes the code under the
'if (!vb->prepared)' condition. This clears previously set data needed
for request handling, and now querybuf will no longer report that this
buffer is part of a request.
Move the state check to before the 'if' and make sure to only do the
state check when called from QBUF and if V4L2_BUF_FLAG_REQUEST_FD is
set.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
media: mtk-vcodec: fix vp9 content playback error with show exist frame
Update referenced frame buffer's reference count when playing vp9
content which has show_existing_frame flag, and copy enough buffer
data to current shown frame.
Current ImgU driver processes and releases the parameter buffer
immediately after queued from user. This does not align with other
image buffers which are grouped in sets and used for the same frame.
If user queues multiple parameter buffers continuously, only the last
one will take effect.
To make consistent buffers usage, this patch changes the parameter
buffer handling and group parameter buffer with other image buffers
for each frame.
Each time driver will queue one more group of buffers when previous
frame processed and buffers consumed by css.
Signed-off-by: Tianshu Qiu <tian.shu.qiu@intel.com> Signed-off-by: Bingbu Cao <bingbu.cao@intel.com> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Akinobu Mita [Sat, 30 Mar 2019 14:01:31 +0000 (10:01 -0400)]
media: ov2659: make S_FMT succeed even if requested format doesn't match
This driver returns an error if unsupported media bus pixel code is
requested by VIDIOC_SUBDEV_S_FMT.
But according to Documentation/media/uapi/v4l/vidioc-subdev-g-fmt.rst,
Drivers must not return an error solely because the requested format
doesn't match the device capabilities. They must instead modify the
format to match what the hardware can provide.
So select default format code and return success in that case.
Janusz Krzysztofik [Sat, 30 Mar 2019 01:06:09 +0000 (21:06 -0400)]
media: ov6650: Move v4l2_clk_get() to ov6650_video_probe() helper
In preparation for adding asynchronous subdevice support to the driver,
don't acquire v4l2_clk from the driver .probe() callback as that may
fail if the clock is provided by a bridge driver which may be not yet
initialized. Move the v4l2_clk_get() to ov6650_video_probe() helper
which is going to be converted to v4l2_subdev_internal_ops.registered()
callback, executed only when the bridge driver is ready.
Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Janusz Krzysztofik [Mon, 25 Mar 2019 00:21:12 +0000 (20:21 -0400)]
media: ov6650: Fix sensor possibly not detected on probe
After removal of clock_start() from before soc_camera_init_i2c() in
soc_camera_probe() by commit 9aea470b399d ("[media] soc-camera: switch
I2C subdevice drivers to use v4l2-clk") introduced in v3.11, the ov6650
driver could no longer probe the sensor successfully because its clock
was no longer turned on in advance. The issue was initially worked
around by adding that missing clock_start() equivalent to OMAP1 camera
interface driver - the only user of this sensor - but a propoer fix
should be rather implemented in the sensor driver code itself.
Fix the issue by inserting a delay between the clock is turned on and
the sensor I2C registers are read for the first time.
Tested on Amstrad Delta with now out of tree but still locally
maintained omap1_camera host driver.
Fixes: 9aea470b399d ("[media] soc-camera: switch I2C subdevice drivers to use v4l2-clk") Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Cc: stable@vger.kernel.org Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Fix media_devnode_unregister() to clear devnode->media_dev while holding
media_devnode_lock. media_devnode_register()'s cdev_add_error handling
does this correctly.
media: au0828: Fix NULL pointer dereference in au0828_analog_stream_enable()
Fix au0828_analog_stream_enable() to check if device is in the right
state first. When unbind happens while bind is in progress, usbdev
pointer could be invalid in au0828_analog_stream_enable() and a call
to usb_ifnum_to_if() will result in the null pointer dereference.
This problem is found with the new media_dev_allocator.sh test.
Media Device Allocator API to allows multiple drivers share a media device.
This API solves a very common use-case for media devices where one physical
device (an USB stick) provides both audio and video. When such media device
exposes a standard USB Audio class, a proprietary Video class, two or more
independent drivers will share a single physical USB bridge. In such cases,
it is necessary to coordinate access to the shared resource.
Using this API, drivers can allocate a media device with the shared struct
device as the key. Once the media device is allocated by a driver, other
drivers can get a reference to it. The media device is released when all
the references are released.
This test does a series of unbind/bind tests to make sure media device
is released correctly when it is no longer is use and when the last
driver releases the reference.
media: au0828: fix enable and disable source audio and video inconsistencies
Enable and disable source interfaces aren't consistent in enforcing
how video and audio share the tuner resource.
Fix these issues to enforce the following rules and allow
sharing between audio and video applications.
- When DVB is streaming, audio/video/vbi/s-video/composite
should find the resource busy. DVB holds the tuner in
exclusive mode.
- When video/VBI is streaming, audio can share the tuner and vice versa.
- v4l2 allows multiple applications to open video device.
- Video applications call enable source multiple times during their
run-time. Resource should stay locked until the last application
releases it.
- A shared resource should stay in shared state and locked when it is
in use by audio and video. More than one video application is allowed
to use the tuner as long as video streaming protocol allows such usage.
Resource is released when the last video/audio application releases it.
- S-Video and Composite hold the resource in exclusive mode.
media: sound/usb: Use Media Controller API to share media resources
Media Device Allocator API to allows multiple drivers share a media device.
This API solves a very common use-case for media devices where one physical
device (an USB stick) provides both audio and video. When such media device
exposes a standard USB Audio class, a proprietary Video class, two or more
independent drivers will share a single physical USB bridge. In such cases,
it is necessary to coordinate access to the shared resource.
Using this API, drivers can allocate a media device with the shared struct
device as the key. Once the media device is allocated by a driver, other
drivers can get a reference to it. The media device is released when all
the references are released.
Change the ALSA driver to use the Media Controller API to share media
resources with DVB, and V4L2 drivers on a AU0828 media device.
The Media Controller specific initialization is done after sound card is
registered. ALSA creates Media interface and entity function graph nodes
for Control, Mixer, PCM Playback, and PCM Capture devices.
snd_usb_hw_params() will call Media Controller enable source handler
interface to request the media resource. If resource request is granted,
it will release it from snd_usb_hw_free(). If resource is busy, -EBUSY is
returned.
Media specific cleanup is done in usb_audio_disconnect().
Move PCM_CAPTURE, PCM_PLAYBACK, and CONTROL ALSA MEDIA_INTF_T* interface
types back into __KERNEL__ scope to get ready for adding ALSA support for
these to the media controller.
media: change au0828 to use Media Device Allocator API
Media Device Allocator API to allows multiple drivers share a media device.
This API solves a very common use-case for media devices where one physical
device (an USB stick) provides both audio and video. When such media device
exposes a standard USB Audio class, a proprietary Video class, two or more
independent drivers will share a single physical USB bridge. In such cases,
it is necessary to coordinate access to the shared resource.
Using this API, drivers can allocate a media device with the shared struct
device as the key. Once the media device is allocated by a driver, other
drivers can get a reference to it. The media device is released when all
the references are released.
Change au0828 to use Media Device Allocator API to allocate media device
with the parent usb struct device as the key, so it can be shared with the
snd_usb_audio driver.
Media Device Allocator API to allows multiple drivers share a media device.
This API solves a very common use-case for media devices where one physical
device (an USB stick) provides both audio and video. When such media device
exposes a standard USB Audio class, a proprietary Video class, two or more
independent drivers will share a single physical USB bridge. In such cases,
it is necessary to coordinate access to the shared resource.
Using this API, drivers can allocate a media device with the shared struct
device as the key. Once the media device is allocated by a driver, other
drivers can get a reference to it. The media device is released when all
the references are released.
Colin Ian King [Sat, 30 Mar 2019 15:25:03 +0000 (11:25 -0400)]
media: dvb: clean up redundant break statements
There are several places where a break statement occurs before
a following break statement; these are unnecessary and can be
removed to clean up the code a little.
Signed-off-by: Colin Ian King <colin.king@canonical.com> Reviewed-by: Mukesh Ojha <mojha@codeaurora.org> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
The image renderer (IMR), or the distortion correction engine, is a
drawing processor with a simple instruction system capable of referencing
video capture data or data in an external memory as the 2D texture data
and performing texture mapping and drawing with respect to any shape that
is split into triangular objects.
Document the device tree bindings for the image renderer light extended 4
(IMR-LX4) found in the R-Car gen3 SoCs...
Signed-off-by: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com> Acked-by: Rob Herring <robh@kernel.org> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Helen Fornazier [Wed, 13 Mar 2019 18:29:37 +0000 (14:29 -0400)]
media: vimc: propagate pixel format in the stream
Media bus codes were being mapped to pixelformats, which causes a
limitation on vimc because not all pixelformats can be mapped to media
bus codes.
Also, media bus codes are an internal configuration from the device.
Userspace only assures media bus codes matches between pads and expects
the image in a given pixelformat. So we can allow almost any media bus
format to be configured between pads, except for debayer that expects a
media bus code of type bayer in the sink pad.
[hverkuil-cisco@xs4all.nl: drop use of v4l2_get_fourcc_name: not yet available]
[hverkuil-cisco@xs4all.nl: made vimc_mbus_list static]
Signed-off-by: Helen Koike <helen.koike@collabora.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Ezequiel Garcia [Thu, 28 Mar 2019 18:07:03 +0000 (14:07 -0400)]
media: rockchip/vpu: Rename pixel format helpers
The rockchip VPU driver uses generic names for its pixel format
helpers. We want to use the same names for generic versions
of these helpers, so rename the rockchip ones.
The driver will be switched to the generic helpers later.
Yannick Fertre [Fri, 29 Mar 2019 12:39:42 +0000 (08:39 -0400)]
media: platform: stm32: wait end of transmission
It is mandatory to write CEC_CFGR only when CECEN=0. To protect
transmission, a check have been added to delayed logical address
modification. This patch is necessary tp pass all tests of compliance.
Eugen Hristev [Fri, 29 Mar 2019 07:38:31 +0000 (03:38 -0400)]
media: atmel: atmel-isc: removed ARGB32 added ABGR32 and XBGR32
ISC will output the "ARGB32" configuration in byte order: B, G, R, Alpha.
This is in fact the format BGRA, aka ABGR32.
If alpha is missing, the same format is equivalent to XBGR32.
Added both formats and removed ARGB32 which is wrong.
Signed-off-by: Eugen Hristev <eugen.hristev@microchip.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Eugen Hristev [Fri, 29 Mar 2019 07:38:28 +0000 (03:38 -0400)]
media: atmel: atmel-isc: reworked driver and formats
This change is a redesign in the formats and the way the ISC is
configured w.r.t. sensor format and the output format from the ISC.
I have changed the splitting between sensor output (which is also ISC input)
and ISC output.
The sensor format represents the way the sensor is configured, and what ISC
is receiving.
The format configuration represents the way ISC is interpreting the data and
formatting the output to the subsystem.
Now it's much easier to figure out what is the ISC configuration for input, and
what is the configuration for output.
The non-raw format can be obtained directly from sensor or it can be done
inside the ISC. The controller format list will include a configuration for
each format.
The old supported formats are still in place, if we want to dump the sensor
format directly to the output, the try format routine will detect and
configure the pipeline accordingly.
This also fixes the previous issues when the raw format was NULL which
resulted in many crashes for sensors which did not have the expected/tested
formats.
Signed-off-by: Eugen Hristev <eugen.hristev@microchip.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Fish Lin [Fri, 29 Mar 2019 03:20:46 +0000 (23:20 -0400)]
media: v4l: add I / P frame min max QP definitions
Add following V4L2 QP parameters for H.264:
* V4L2_CID_MPEG_VIDEO_H264_I_FRAME_MIN_QP
* V4L2_CID_MPEG_VIDEO_H264_I_FRAME_MAX_QP
* V4L2_CID_MPEG_VIDEO_H264_P_FRAME_MIN_QP
* V4L2_CID_MPEG_VIDEO_H264_P_FRAME_MAX_QP
These controls will limit QP range for intra and inter frame,
provide more manual control to improve video encode quality.
Signed-off-by: Fish Lin <linfish@google.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Akinobu Mita [Mon, 18 Mar 2019 15:25:43 +0000 (11:25 -0400)]
media: ov7740: enable to get exposure control in autoexposure mode
The exposure control is clustered with the autoexposure control and
flagged as volatile, but the g_volatile_ctrl() doesn't handle
V4L2_CID_EXPOSURE_AUTO. So, the value of the exposure control can't be
read in autoexposure mode.
This enables to get the exposure control in autoexposure mode by making
ov7740_get_volatile_ctrl() deal with V4L2_CID_EXPOSURE_AUTO.
This also sets the exposure control as volatile by specifying the
argument to v4l2_ctrl_auto_cluster() instead of manually flagging it.
Cc: Wenyou Yang <wenyou.yang@microchip.com> Cc: Eugen Hristev <eugen.hristev@microchip.com> Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com> Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Dan Carpenter [Tue, 26 Mar 2019 05:12:07 +0000 (01:12 -0400)]
media: wl128x: prevent two potential buffer overflows
Smatch marks skb->data as untrusted so it warns that "evt_hdr->dlen"
can copy up to 255 bytes and we only have room for two bytes. Even
if this comes from the firmware and we trust it, the new policy
generally is just to fix it as kernel hardenning.
I can't test this code so I tried to be very conservative. I considered
not allowing "evt_hdr->dlen == 1" because it doesn't initialize the
whole variable but in the end I decided to allow it and manually
initialized "asic_id" and "asic_ver" to zero.
Fixes: e8454ff7b9a4 ("[media] drivers:media:radio: wl128x: FM Driver Common sources") 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+samsung@kernel.org>
Alexandre Courbot [Tue, 26 Mar 2019 07:44:23 +0000 (03:44 -0400)]
media: mtk-vcodec: fix access to incorrect planes member
Commit 0650a91499e0 ("media: mtk-vcodec: Correct return type for mem2mem
buffer helpers") fixed the return types for mem2mem buffer helper
functions by changing a few local variables from vb2_buffer to
vb2_v4l2_buffer. However, it left a few accesses to vb2_buffer::planes
as-is, accidentally turning them into accesses to
vb2_v4l2_buffer::planes and resulting in values being read from/written
to the wrong place.
Fix this by inserting vb2_buf into these accesses so they mimic their
original behavior.
Fixes: 0650a91499e0 ("media: mtk-vcodec: Correct return type for mem2mem buffer helpers") Signed-off-by: Alexandre Courbot <acourbot@chromium.org> Reviewed-by: Ezequiel Garcia <ezequiel@collabora.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Arnd Bergmann [Fri, 22 Mar 2019 14:34:22 +0000 (10:34 -0400)]
media: davinci-isif: avoid uninitialized variable use
clang warns about a possible variable use that gcc never
complained about:
drivers/media/platform/davinci/isif.c:982:32: error: variable 'frame_size' is uninitialized when used here
[-Werror,-Wuninitialized]
dm365_vpss_set_pg_frame_size(frame_size);
^~~~~~~~~~
drivers/media/platform/davinci/isif.c:887:2: note: variable 'frame_size' is declared here
struct vpss_pg_frame_size frame_size;
^
1 error generated.
There is no initialization for this variable at all, and there
has never been one in the mainline kernel, so we really should
not put that stack data into an mmio register.
On the other hand, I suspect that gcc checks the condition
more closely and notices that the global
isif_cfg.bayer.config_params.test_pat_gen flag is initialized
to zero and never written to from any code path, so anything
depending on it can be eliminated.
To shut up the clang warning, just remove the dead code manually,
it has probably never been used because any attempt to do so
would have resulted in undefined behavior.
When CONFIG_DEBUGFS is enabled, we get a warning about an
incorrect section annotation that can lead to undefined
behavior:
WARNING: vmlinux.o(.text+0xd3c7c4): Section mismatch in reference from the function mipi_csis_probe() to the function .init.text:mipi_csis_debugfs_init()
The function mipi_csis_probe() references
the function __init mipi_csis_debugfs_init().
This is often because mipi_csis_probe lacks a __init
annotation or the annotation of mipi_csis_debugfs_init is wrong.
The same function for an unknown reason has a different
version for !CONFIG_DEBUGFS, which does not have this problem,
but behaves the same way otherwise (it does nothing when debugfs
is disabled).
Consolidate the two versions, using the correct section from
one version, and the implementation from the other.
Signed-off-by: Arnd Bergmann <arnd@arndb.de> Reviewed-by: Rui Miguel Silva <rmfrfs@gmail.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
Arnd Bergmann [Wed, 13 Mar 2019 21:10:29 +0000 (17:10 -0400)]
media: staging: media: davinci_vpfe: fix large stack usage with clang
clang is unable to optimize the isif_ioctl() in the same way that
gcc does, as it fails to prove that the local copy of
the 'struct vpfe_isif_raw_config' argument is unnecessary:
drivers/staging/media/davinci_vpfe/dm365_isif.c:622:13: error: stack frame size of 1344 bytes in function 'isif_ioctl' [-Werror,-Wframe-larger-than=]
Marking it as 'const' while passing the data down clearly shows us that
the copy is never modified, and we can skip copying it entirely, which
reduces the stack usage to just eight bytes.
media: hfi_parser: don't trick gcc with a wrong expected size
Smatch warns about small size on two structs:
drivers/media/platform/qcom/venus/hfi_parser.c:103 parse_profile_level() error: memcpy() 'proflevel' too small (8 vs 128)
drivers/media/platform/qcom/venus/hfi_parser.c: drivers/media/platform/qcom/venus/hfi_parser.c:129 parse_caps() error: memcpy() 'cap' too small (16 vs 512)
The reason is that the hfi_parser actually expects:
- multiple data entries on hfi_capabilities
- multiple profile_level on hfi_profile_level_supported
However, the structs trick gcc, making it to believe that
there's just one value for each.
The only way for p_set to be NULL would be if vin_coef_set would be an
empty array.
On such case, the driver will OOPS, as it would try to de-reference a
NULL value. So, the check if p_set is not NULL doesn't make any sense.
Solves those two smatch warnings:
drivers/media/platform/rcar-vin/rcar-dma.c:489 rvin_set_coeff() warn: variable dereferenced before check 'p_set' (see line 484)
drivers/media/platform/rcar-vin/rcar-dma.c:494 rvin_set_coeff() error: we previously assumed 'p_set' could be null (see line 489)
The way the code works, compression will be a valid value (less or equal to 3)
on both set_video_mode_foo() calls at the beginning of the while() loop.
So, the value for pChoose can't be NULL.
Solves those smatch warnings:
drivers/media/usb/pwc/pwc-ctrl.c: drivers/media/usb/pwc/pwc-ctrl.c:252 set_video_mode_Timon() warn: variable dereferenced before check 'pChoose' (see line 248)
drivers/media/usb/pwc/pwc-ctrl.c: drivers/media/usb/pwc/pwc-ctrl.c:302 set_video_mode_Kiara() warn: variable dereferenced before check 'pChoose' (see line 298)
As warned by smatch:
drivers/media/i2c/imx214.c:591 imx214_set_format() warn: variable dereferenced before check 'format' (see line 589)
It turns that the code at imx214_set_format() has support for being
called with the format being NULL. I've no idea why, as it is only
called internally with the pointer set, and via subdev API (with
should also set it).
Also, the entire logic there depends on having format != NULL, so
just remove the bogus broken support for a null format.
Dafna Hirschfeld [Wed, 6 Mar 2019 21:13:41 +0000 (16:13 -0500)]
media: vicodec: Register another node for stateless decoder
Add stateless decoder instance field to the dev struct and
register another node for the statelsess decoder.
The stateless API for the node will be implemented in further patches.
Dafna Hirschfeld [Wed, 6 Mar 2019 21:13:35 +0000 (16:13 -0500)]
media: vicodec: Handle the case that the reference buffer is NULL
In the stateless decoder the reference buffer is null if the
frame is an I-frame (flagged with FWHT_FL_I_FRAME).
Make sure not to dereference it in that case.
Dafna Hirschfeld [Wed, 6 Mar 2019 21:13:34 +0000 (16:13 -0500)]
media: vicodec: rename v4l2_fwht_default_fmt to v4l2_fwht_find_nth_fmt
Rename 'v4l2_fwht_default_fmt' to 'v4l2_fwht_find_nth_fmt'
and add a function 'v4l2_fwht_validate_fmt' to check if
a format info matches the parameters.
This function will also be used to validate the stateless
params when adding support for stateless codecs.
Dafna Hirschfeld [Wed, 6 Mar 2019 21:13:32 +0000 (16:13 -0500)]
media: vicodec: keep the ref frame according to the format in decoder
In the decoder, save the inner reference frame in the same
format as the capture buffer.
The decoder writes directly to the capture buffer and then
the capture buffer is copied to the reference buffer.
This will simplify the stateless decoder.
Dafna Hirschfeld [Wed, 6 Mar 2019 21:13:31 +0000 (16:13 -0500)]
media: vicodec: add field 'buf' to fwht_raw_frame
Add the field 'buf' to fwht_raw_frame to indicate
the start of the raw frame buffer.
This field will be used to copy the capture buffer
to the reference buffer in the next patch.
Dafna Hirschfeld [Wed, 6 Mar 2019 21:13:28 +0000 (16:13 -0500)]
media: vicodec: bugfix - call v4l2_m2m_buf_copy_metadata also if decoding fails
The function 'v4l2_m2m_buf_copy_metadata' should
be called even if decoding/encoding ends with
status VB2_BUF_STATE_ERROR, so that the metadata
is copied from the source buffer to the dest buffer.
Dafna Hirschfeld [Wed, 6 Mar 2019 21:13:27 +0000 (16:13 -0500)]
media: vicodec: change variable name for the return value of v4l2_fwht_encode
v4l2_fwht_encode returns either an error code on
failure or the size of the compressed frame on
success. So change the var assigned to it from
'ret' to 'comp_sz_or_errcode' to clarify that.
Hans Verkuil [Wed, 20 Mar 2019 14:31:18 +0000 (10:31 -0400)]
media: media requests: return EBADR instead of EACCES
If requests are used when they shouldn't, or not used when they should,
then return EBADR (Invalid request descriptor) instead of EACCES.
The reason for this change is that EACCES has more to do with permissions
(not being the owner of the resource), but in this case the request file
descriptor is just wrong for the current mode of the device.
Update the documentation accordingly.
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
projects / users / willy / xarray.git / log