[media] v4l: vsp1: Repair race between frame end and qbuf handler
The frame-end function releases and completes the buffers on the input
and output entities of the pipe before marking the pipe->state as
'STOPPED'. This introduces a race whereby with the pipe->state still
'RUNNING', a QBUF handler can commence processing a frame before the
frame_end function has completed.
In the event that this happens, a frame queued by QBUF hangs due to the
incorrect pipe->state setting which prevents vsp1_pipeline_run from
issuing a CMD_STRCMD.
By locking the entire function we prevent this from occurring, but we
also change the locking state of the buffer release code. This has been
analysed visually as acceptable, but it must be considered that this now
causes the video->irqlock to be taken under the pipe->irqlock context.
[media] v4l: vsp1: Prevent pipelines from running when not streaming
Pipelines can only be run if all their video nodes are streaming. Commit b4dfb9b35a19 ("[media] v4l: vsp1: Stop the pipeline upon the first
STREAMOFF") fixed the pipeline stop sequence, but introduced a race
condition that makes it possible to run a pipeline after stopping the
stream on a video node by queuing a buffer on the other side of the
pipeline.
Fix this by clearing the buffers ready flag when stopping the stream,
which will prevent the QBUF handler from running the pipeline.
Fixes: b4dfb9b35a19 ("[media] v4l: vsp1: Stop the pipeline upon the first STREAMOFF") Reported-by: Kieran Bingham <kieran@bingham.xyz> Tested-by: Kieran Bingham <kieran@bingham.xyz> Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Changeset ab6343956f9c ("[media] V4L2: Add documentation for SDI timings
and related flags") added documentation for new V4L2 defines, but
it forgot to update videodev2.h.rst.exceptions to point to where
the documentation for those new values will be inside the book,
causing those warnings:
Documentation/output/videodev2.h.rst:6: WARNING: undefined label: v4l2-dv-bt-std-sdi (if the link has no caption the label must precede a section header)
Documentation/output/videodev2.h.rst:6: WARNING: undefined label: v4l2-dv-fl-first-field-extra-line (if the link has no caption the label must precede a section header)
Documentation/output/videodev2.h.rst:6: WARNING: undefined label: v4l2-in-st-no-v-lock (if the link has no caption the label must precede a section header)
Documentation/output/videodev2.h.rst:6: WARNING: undefined label: v4l2-in-st-no-std-lock (if the link has no caption the label must precede a section header)
Fixes: ab6343956f9c ("[media] V4L2: Add documentation for SDI timings and related flags") Cc: Charles-Antoine Couret <charles-antoine.couret@nexvision.fr> Cc: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Building with -Wmaybe-uninitialized reveals the use on an uninitialized
variable containing the physical address of the device whenever
firmware before version 2 is used:
drivers/staging/media/pulse8-cec/pulse8-cec.c: In function 'pulse8_connect':
drivers/staging/media/pulse8-cec/pulse8-cec.c:447:2: error: 'pa' may be used uninitialized in this function [-Werror=maybe-uninitialized]
This sets the address to CEC_PHYS_ADDR_INVALID in this case, so we don't
try to write back the uninitialized data to the device.
Jean Delvare [Thu, 8 Sep 2016 15:25:16 +0000 (12:25 -0300)]
[media] cec: fix Kconfig help text
MEDIA_CEC is no longer a tristate option, so the user can't actually
choose M. Whether the code is built-in or built as a module is
decided somewhere else.
In addition to setting the ACK mask, also set the logical address mask
setting in the dongle. This (and not the ACK mask) is persisted for
use in autonomous mode.
The logical address mask to use is deduced from the primary device type
in adap->log_addrs.
Signed-off-by: Johan Fjeldtvedt <jaffe1@gmail.com> Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
trivial fix to spelling mistake in pr_debug message
Signed-off-by: Colin Ian King <colin.king@canonical.com> Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Hans Verkuil [Sun, 11 Sep 2016 08:57:30 +0000 (05:57 -0300)]
[media] pulse8-cec: fix compiler warning
pulse8-cec.c: In function 'pulse8_connect':
pulse8-cec.c:447:2: warning: 'pa' may be used uninitialized in this function [-Wmaybe-uninitialized]
cec_s_phys_addr(pulse8->adap, pa, false);
^
pulse8-cec.c:609:6: note: 'pa' was declared here
u16 pa;
^
As far as I can tell, this can't actually happen. Still, it is better to just
initialize it.
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Hans Verkuil [Sun, 11 Sep 2016 08:51:54 +0000 (05:51 -0300)]
[media] pxa_camera: merge soc_mediabus.c into pxa_camera.c
Linking soc_mediabus into this driver causes multiple definition linker warnings
if soc_camera is also enabled:
drivers/media/platform/soc_camera/built-in.o:(___ksymtab+soc_mbus_image_size+0x0): multiple definition of `__ksymtab_soc_mbus_image_size'
drivers/media/platform/soc_camera/soc_mediabus.o:(___ksymtab+soc_mbus_image_size+0x0): first defined here
>> drivers/media/platform/soc_camera/built-in.o:(___ksymtab+soc_mbus_samples_per_pixel+0x0): multiple definition of `__ksymtab_soc_mbus_samples_per_pixel'
drivers/media/platform/soc_camera/soc_mediabus.o:(___ksymtab+soc_mbus_samples_per_pixel+0x0): first defined here
drivers/media/platform/soc_camera/built-in.o: In function `soc_mbus_config_compatible':
(.text+0x3840): multiple definition of `soc_mbus_config_compatible'
drivers/media/platform/soc_camera/soc_mediabus.o:(.text+0x134): first defined here
Since we really don't want to have to use any of the soc-camera code this patch
copies the relevant code and data structures from soc_mediabus and renames it to pxa_mbus_*.
The large table of formats has been culled a bit, removing formats that are not supported
by this driver.
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Cc: Robert Jarzmik <robert.jarzmik@free.fr> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
[media] v4l: vsp1: Report device model and rev through media device information
Instead of hardcoding the media device model and hardware revision to
"VSP1" and 0 respectively, report the actual hardware device model and
IP version number.
Kieran Bingham [Thu, 9 Jun 2016 17:06:43 +0000 (14:06 -0300)]
[media] v4l: rcar-fcp: Extend compatible list to support the FDP
The FCP must be powered up for the FDP1 to function, even when the FDP1
does not make use of the FCNL features. Extend the compatible list
to allow us to use the power domain and runtime-pm support.
Laurent Pinchart [Wed, 17 Aug 2016 12:57:37 +0000 (09:57 -0300)]
[media] v4l: rcar-fcp: Don't force users to check for disabled FCP support
The rcar_fcp_enable() function immediately returns successfully when the
FCP device pointer is NULL to avoid forcing the users to check the FCP
device manually before every call. However, the stub version of the
function used when the FCP driver is disabled returns -ENOSYS
unconditionally, resulting in a different API contract for the two
versions of the function.
As a user that requires FCP support will fail at probe time when calling
rcar_fcp_get() if the FCP driver is disabled, the stub version of the
rcar_fcp_enable() function will only be called with a NULL FCP device.
We can thus return 0 unconditionally to align the behaviour with the
normal version of the function.
Laurent Pinchart [Tue, 28 Jun 2016 13:20:10 +0000 (10:20 -0300)]
[media] v4l: ioctl: Clear the v4l2_pix_format_mplane reserved field
The S_FMT and TRY_FMT handlers in multiplane mode attempt at clearing
the reserved fields of the v4l2_format structure after the pix_mp
member. However, the reserved fields are inside pix_mp, not after it.
The power domain must be specified to bring the device out of module
standby. Document this in the bindings provided, so that new additions
are not missed.
Sakari Ailus [Thu, 7 Jul 2016 06:10:49 +0000 (03:10 -0300)]
[media] doc-rst: 16-bit BGGR is always 16 bits
The V4L2_PIX_FMT_SBGGR16 format is documented to contain samples of fewer
than 16 bits. However, we do have specific definitions for smaller sample
sizes. Therefore, this note is redundant from the API point of view.
Currently only two drivers, am437x and davinci, use the
V4L2_PIX_FMT_SBGGR16 pixelformat currently. The sampling precision is
understood to be 16 bits in all current cases.
Sakari Ailus [Fri, 5 Aug 2016 10:20:21 +0000 (07:20 -0300)]
[media] doc-rst: Unify documentation of the 8-bit bayer formats
The other raw bayer formats had a single sample depth dependent definition
whereas the 8-bit formats had one page for each. Unify the documentation
of the 8-bit formats.
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Sakari Ailus [Mon, 20 Jun 2016 15:53:55 +0000 (12:53 -0300)]
[media] doc-rst: Clean up raw bayer pixel format definitions
- Explicitly state that the most significant n bits are zeroed on 10 and
12 bpp formats.
- Remove extra comma from the last entry of the format list
- Add a missing colon before a list
- Use figures versus word numerals consistently
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Acked-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Sakari Ailus [Fri, 11 Mar 2016 22:47:36 +0000 (19:47 -0300)]
[media] doc-rst: Correct the ordering of LSBs of the 10-bit raw packed formats
The 10-bit packed raw bayer format documented that the data of the first
pixel of a four-pixel group was found in the first byte and the two
highest bits of the fifth byte. This was not entirely correct. The two
bits in the fifth byte are the two lowest bits. The second pixel occupies
the second byte and third and fourth least significant bits and so on.
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Acked-by: Aviv Greenberg <aviv.d.greenberg@intel.com> Acked-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
[media] pxa_camera: remove an unused structure pointer
As reported by smatch:
drivers/media/platform/pxa_camera.c: In function 'pxa_dma_start_channels':
drivers/media/platform/pxa_camera.c:457:21: warning: variable 'active' set but not used [-Wunused-but-set-variable]
struct pxa_buffer *active;
^~~~~~
Instead of the legacy behavior where it was required to wait for all
video buffers to be finished by the hardware, use a cancel like strategy
: as soon as the stop_streaming() call is done, abort all DMA transfers,
report the already buffers as failed and return.
This makes stop_streaming() more a "cancel capture" than a "wait for end
of capture" semantic.
Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr> Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Robert Jarzmik [Tue, 6 Sep 2016 09:04:20 +0000 (06:04 -0300)]
[media] media: platform: pxa_camera: make a standalone v4l2 device
This patch removes the soc_camera API dependency from pxa_camera.
In the current status :
- all previously captures are working the same on pxa270
- the s_crop() call was removed, judged not working
(see what happens soc_camera_s_crop() when get_crop() == NULL)
- if the pixel clock is provided by then sensor, ie. not MCLK, the dual
stage change is not handled yet.
=> there is no in-tree user of this, so I'll let it that way
- the MCLK is not yet finished, it's as in the legacy way,
ie. activated at video device opening and closed at video device
closing.
In a subsequence patch pxa_camera_mclk_ops should be used, and
platform data MCLK ignored. It will be the sensor's duty to request
the clock and enable it, which will end in pxa_camera_mclk_ops.
Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr> Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
This is to be seen as a regression as the set_selection (former
set_crop) function is removed. This is a temporary situation in the v4l2
porting, and will have to be added later.
Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr> Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Introduce sensor_call(), which will be used for all sensor invocations.
This is a preparation move to v4l2 device conversion, ie. soc_camera
adherence removal.
Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr> Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Robert Jarzmik [Tue, 6 Sep 2016 09:04:15 +0000 (06:04 -0300)]
[media] media: platform: pxa_camera: trivial move of functions
Move the functions in the file to be regrouped into meaningful blocks :
1. pxa camera core handling functions, manipulating the herdware
2. videobuf2 functions, dealing with video buffers
3. video ioctl (vidioc) related functions
4. driver probing, removal, suspend and resume
This patch doesn't modify a single line of code.
Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr> Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Robert Jarzmik [Tue, 6 Sep 2016 09:04:11 +0000 (06:04 -0300)]
[media] media: mt9m111: make a standalone v4l2 subdevice
Remove the soc_camera adherence. Mostly the change removes the power
manipulation provided by soc_camera, and instead :
- powers on the sensor when the s_power control is activated
- powers on the sensor in initial probe
- enables and disables the MCLK provided to it in power on/off
This patch also drops support for inverters on synchronisation and clock
lines. It is assumed, if any board ever needs such inverters, support
for them can be added in the future
Acked-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de> Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr> Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
This patch adds the H.264 video encoding capability in the V4L2 HVA
video encoder driver for STMicroelectronics SoC (hva-h264.c).
The main supported features are:
- profile: baseline, main, high, stereo high
- level: up to 4.2
- bitrate mode: CBR, VBR
- entropy mode: CABAC, CAVLC
- video aspect: 1x1 only
[media] st-hva: multi-format video encoder V4L2 driver
This patch adds V4L2 HVA (Hardware Video Accelerator) video encoder
driver for STMicroelectronics SoC. It uses the V4L2 mem2mem framework.
This patch only contains the core parts of the driver:
- the V4L2 interface with the userland (hva-v4l2.c)
- the hardware services (hva-hw.c)
- the memory management utilities (hva-mem.c)
This patch doesn't include the support of specific codec (e.g. H.264)
video encoding: this support is part of subsequent patches.
[media] videobuf2-v4l2: document two helper functions
Document vb2_ops_wait_prepare() and vb2_ops_wait_finish(),
in order to fix those two warnings:
Documentation/media/kapi/v4l2-dev.rst:166: WARNING: c:func reference target not found: vb2_ops_wait_prepare
Documentation/media/kapi/v4l2-dev.rst:166: WARNING: c:func reference target not found: vb2_ops_wait_finish
[media] videobuf2-core.h: move function descriptions from c file
There are several functions that were documented at the .c
file. As we only include the headers, we need to move them to
there, in order to have documentation.
[media] mc-core.rst: fix a warning about an internal routine
Fix this warning:
Documentation/media/kapi/mc-core.rst:97: WARNING: c:func reference target not found: media_devnode_release
The media_device_release() is a function internal to media-devnode.c,
and not exported elsewhere. So, we can't cross-reference it here.
Make it explicit at the documentation.
[media] media-ioc-g-topology.rst: fix a c domain reference
One reference there is still using :ref:. Fix it, to solve this
warning:
Documentation/media/uapi/mediactl/media-ioc-g-topology.rst:236: WARNING: undefined label: media-v2-intf-devnode (if the link has no caption the label must precede a section header)
[media] docs-rst: fix cross-references for videodev2.h
There are several broken references there, due to the conversion to
C domain. Fix them using this shell script and manually adjust what's
broken:
# funcs is a file with the broken functions/references
for i in $(cat funcs|sort|uniq|perl -ne 'print "$1\n" if (m/(\S+)$/)'); do
i=${i//-/_}
echo $i
j=${i//_/-}
for k in $(git grep -l "_$j:" Documentation/); do
sed s,\_$j\:,"c\:type\:\: $i", <$k >a && mv a $k
done
for k in $(git grep -l "$j" Documentation/media/*.exceptions); do
sed s,$j,":c\:type\:\`$i\`", <$k >a && mv a $k
done
for k in $(git grep -l "$j" Documentation/); do
sed "s,:ref:\`$i <$j>\`,:c:type:\`$i\`," <$k >a && mv a $k
sed "s,:ref:\`$j\`,:c:type:\`$i\`," <$k >a && mv a $k
sed -E "s,:ref:\`(.*)<$j>\`,:c:type:\`\1<$i>\`," <$k >a && mv a $k
done
for k in $(git grep -l "<$j>" include/media); do
sed -E "s,:ref:\`(.*)<$j>\`,enum \&$i," <$k >a && mv a $k
done
done
Documentation/media/uapi/cec/cec-ioc-dqevent.rst:124: WARNING: c:func reference target not found: clock_gettime(2)
By replacing it with the right function name, using this shell script:
for i in `find Documentation/media -type f`; do sed 's,clock_gettime(2),clock_gettime,' <$i >a && mv a $i; done
Please notice that this will make the nitpick mode to shut up
complaining about that, becasue clock_gettime is on its exclude list,
but the cross reference will be undefined until someone documents
this function at the core documentation.
Trivially fix those broken references, by copying the structs
fron the header, just like other API documentation at the
DVB side.
This doesn't have the level of quality used at the V4L2 side
of the API, but, as this documents a deprecated API, used
only by av7110 driver, it doesn't make much sense to invest
time making it better.
projects / linux-beck.git / log