Hans Verkuil [Fri, 1 Jul 2016 10:33:10 +0000 (07:33 -0300)]
[media] cec: fix Kconfig dependency problems
- Use IS_REACHABLE(RC_CORE) instead of IS_ENABLED: if cec is built-in and
RC_CORE is a module, then CEC can't reach the RC symbols.
- Both cec and cec-edid should be bool and use the same build 'mode' as
MEDIA_SUPPORT (just as is done for the media controller code).
- Add a note to staging that this should be changed once the cec framework
is moved out of staging.
Arnd Bergmann [Wed, 29 Jun 2016 14:26:34 +0000 (11:26 -0300)]
[media] s5p_cec: mark suspend/resume as __maybe_unused
The suspend/resume functions in the s5p-cec driver are only
referenced when CONFIG_PM is enabled, so we get a warning
about unused functions otherwise:
drivers/staging/media/s5p-cec/s5p_cec.c:260:12: error: 's5p_cec_resume' defined but not used [-Werror=unused-function]
static int s5p_cec_resume(struct device *dev)
^~~~~~~~~~~~~~
drivers/staging/media/s5p-cec/s5p_cec.c:253:12: error: 's5p_cec_suspend' defined but not used [-Werror=unused-function]
static int s5p_cec_suspend(struct device *dev)
This marks them as __maybe_unused to avoid the warning without
having to introduce an extra #ifdef.
Hans Verkuil [Thu, 30 Jun 2016 10:08:53 +0000 (07:08 -0300)]
[media] cec-adap: prevent write to out-of-bounds array index
CEC_MSG_REPORT_PHYSICAL_ADDR can theoretically be received from
an unregistered device, but in that case the code should not attempt
to write the received physical address to the phys_addrs array.
That would be pointless since there can be multiple unregistered
devices that report a physical address. We just ignore those.
While at it, improve the dprintk since it would attempt to read
from that array as well with the same out-of-bounds problem.
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Reported-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Hans Verkuil [Sun, 26 Jun 2016 09:57:56 +0000 (06:57 -0300)]
[media] cec-adap: on reply, restore the tx_status value from the transmit
When a reply to an earlier transmit is received, the tx_status of that
transmit needs to be restored. Otherwise the status that the transmit
was successful would be lost.
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
* tag 'v4.7-rc6': (1245 commits)
Linux 4.7-rc6
ovl: warn instead of error if d_type is not supported
MIPS: Fix possible corruption of cache mode by mprotect.
locks: use file_inode()
usb: dwc3: st: Use explicit reset_control_get_exclusive() API
phy: phy-stih407-usb: Use explicit reset_control_get_exclusive() API
phy: miphy28lp: Inform the reset framework that our reset line may be shared
namespace: update event counter when umounting a deleted dentry
9p: use file_dentry()
lockd: unregister notifier blocks if the service fails to come up completely
ACPI,PCI,IRQ: correct operator precedence
fuse: serialize dirops by default
drm/i915: Fix missing unlock on error in i915_ppgtt_info()
powerpc: Initialise pci_io_base as early as possible
mfd: da9053: Fix compiler warning message for uninitialised variable
mfd: max77620: Fix FPS switch statements
phy: phy-stih407-usb: Inform the reset framework that our reset line may be shared
usb: dwc3: st: Inform the reset framework that our reset line may be shared
usb: host: ehci-st: Inform the reset framework that our reset line may be shared
usb: host: ohci-st: Inform the reset framework that our reset line may be shared
...
Heiner Kallweit [Fri, 24 Jun 2016 05:38:54 +0000 (02:38 -0300)]
[media] media: rc: nuvoton: remove interrupt handling for wakeup
There is a rudimentary ISR for handling interrupts generated by the
wakeup part of the chip. Interrupts however have never been enabled
in the wakeup part.
Also, now that we can set the wakeup sequence via sysfs, there's in
general no need to enable interrupts in the wakeup part.
To detect a rx fifo overrun it's checked whether the number of elements
in the chip fifo exceeds the fifo size. This check can never return true
and is wrong.
Instead we should generate an interrupt if the fifo overruns.
Antti Palosaari [Wed, 29 Jun 2016 23:40:56 +0000 (20:40 -0300)]
[media] rtl2830: move statistics to read_status()
Move statistics polling to read_status() in order to avoid use of
kernel work. Also replace home made sign extension used for
statistics with kernel sign_extend32().
Signed-off-by: Antti Palosaari <crope@iki.fi> Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Antti Palosaari [Wed, 29 Jun 2016 23:39:48 +0000 (20:39 -0300)]
[media] m88ds3103: refactor firmware download
* remove some unneeded variable initialization
* rename variables
* use min() macro to calc max i2c xfer len
* change bad firmware error code from EFAULT to EINVAL
Signed-off-by: Antti Palosaari <crope@iki.fi> Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Antti Palosaari [Wed, 29 Jun 2016 23:39:47 +0000 (20:39 -0300)]
[media] m88ds3103: use Hz instead of kHz on calculations
There was some calculations where was kHz used in order to keep
calculation withing 32-bit. Convert all to Hz and use 64-bit
division helpers where needed.
Signed-off-by: Antti Palosaari <crope@iki.fi> Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Antti Palosaari [Wed, 29 Jun 2016 23:39:45 +0000 (20:39 -0300)]
[media] m88ds3103: calculate DiSEqC message sending time
DiSEqC message sending takes 13.5 ms per byte, which is 54 ms total
when typical 4 byte message is sent. Don't hard-code time limit to
54 ms, but calculate it. Time limit is only used to determine when to
start poll "DiSEqC Tx ready" status from the chip.
Signed-off-by: Antti Palosaari <crope@iki.fi> Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Antti Palosaari [Wed, 29 Jun 2016 23:38:17 +0000 (20:38 -0300)]
[media] si2168: add support for newer firmwares
Si2168-B40 firmware API has changed somewhere between 4.0-11 and
4.0-19 so that sleep will lose firmware upgrade from the chip. Due
to that firmware re-upload is needed when newer firmwares are used.
Rewrote firmware handling logic partly at the same.
Signed-off-by: Antti Palosaari <crope@iki.fi> Cc: Olli Salonen <olli.salonen@iki.fi> Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Hans Verkuil [Fri, 1 Jul 2016 14:37:27 +0000 (11:37 -0300)]
[media] tw686x: make const structs static
Fix sparse warnings:
tw686x-video.c:148:29: warning: symbol 'memcpy_dma_ops' was not declared. Should it be static?
tw686x-video.c:195:29: warning: symbol 'contig_dma_ops' was not declared. Should it be static?
tw686x-video.c:361:29: warning: symbol 'sg_dma_ops' was not declared. Should it be static?
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Dragos Bogdan [Wed, 22 Jun 2016 11:30:42 +0000 (08:30 -0300)]
[media] adv7604: Add support for hardware reset
The part can be reset by a low pulse on the RESET pin (i.e. a hardware
reset) with a minimum width of 5 ms. It is recommended to wait 5 ms
after the low pulse before an I2C write is performed to the part.
For safety reasons, the delays will be between 5 and 10 ms.
The RESET pin can be tied high, so the GPIO is optional.
Arnd Bergmann [Fri, 17 Jun 2016 20:46:28 +0000 (17:46 -0300)]
[media] dvb: use ktime_t for internal timeout
The dvb demuxer code uses a 'struct timespec' to pass a timeout
as absolute time. This will cause problems on 32-bit architectures
in 2038 when time_t overflows, and it is racy with a concurrent
settimeofday() call.
This patch changes the code to use ktime_get() instead, using
the monotonic time base to avoid both the race and the overflow.
Lubomir Rintel [Wed, 1 Jun 2016 13:04:07 +0000 (10:04 -0300)]
[media] usbtv: improve a comment
Patrick Keshishian improved the explanation of the protocol when porting
the driver to OpenBSD. Given it's a reverse engineering one and there's
no documetnation it might be helpful to whoever hacks on the driver.
Signed-off-by: Patrick Keshishian <sidster@boxsoft.com> Signed-off-by: Lubomir Rintel <lkundrak@v3.sk> Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Florian Echtler [Tue, 31 May 2016 20:15:33 +0000 (17:15 -0300)]
[media] sur40: fix occasional oopses on device close
Closing the V4L2 device sometimes triggers a kernel oops.
Present patch fixes this.
Signed-off-by: Martin Kaltenbrunner <modin@yuri.at> Signed-off-by: Florian Echtler <floe@butterbrot.org> Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Cc: <stable@vger.kernel.org> # for v4.2 and up Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Florian Echtler [Tue, 31 May 2016 20:15:32 +0000 (17:15 -0300)]
[media] sur40: lower poll interval to fix occasional FPS drops to ~56 FPS
The framerate sometimes drops below 60 Hz if the poll interval is too high.
Lowering it to the minimum of 1 ms fixes this.
Signed-off-by: Martin Kaltenbrunner <modin@yuri.at> Signed-off-by: Florian Echtler <floe@butterbrot.org> Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Cc: <stable@vger.kernel.org> # for v4.2 and up Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
[media] DocBook: mention the memory type to be set for all streaming I/O
The DMA buffer importing streaming I/O section mentions the memory type
value that has to be set when calling the VIDIOC_REQBUFS ioctl but this
isn't mentioned in neither the Memory Mapping nor User Pointer sections.
A user can know the values by looking at the examples but it's better
to explicitly mention in the documentation, and also makes all sections
about streaming I/O methods more consistent.
Signed-off-by: Javier Martinez Canillas <javier@osg.samsung.com> Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
[media] DocBook: add dmabuf as streaming I/O in VIDIOC_REQBUFS description
Commit 4b9c1cb641c46 ("[media] Documentation: media: description of DMABUF
importing in V4L2") documented the V4L2 dma-buf importing support but did
not update the VIDIOC_REQBUFS description, so only the Memory Mapping and
User Pointer I/O methods are mentioned.
Signed-off-by: Javier Martinez Canillas <javier@osg.samsung.com> Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
It is responsibility of a caller of fops->open(),
to make sure an owner of the fops is available until file is closed.
So, there is no need to lock THIS_MODULE explicitly.
Found by Linux Driver Verification project (linuxtesting.org).
Hans Verkuil [Sun, 26 Jun 2016 10:44:56 +0000 (07:44 -0300)]
[media] rc-main: fix kernel oops after unloading keymap module
When the rc_map table is created the char pointer of the name of the keymap
is copied to the rc_map->name field. However, this pointer points to memory
from the keymap module itself.
Since these keymap modules are not refcounted, that means anyone can call
rmmod to unload that module. Which is not a big deal because the contents of
the map is all copied to rc_map, except for the keymap name.
So after a keymap module is unloaded the name pointer has become stale. Unloading
the rc-core module will now cause a kernel oops in rc_dev_uevent().
The solution is to kstrdup the name so there are no more references to the
keymap module remaining.
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Hans Verkuil [Fri, 8 Jul 2016 19:11:18 +0000 (16:11 -0300)]
[media] mtk-vcodec: convert driver to use the new vb2_queue dev field
The patch dropping the vb2_dma_contig_init_ctx() and _cleanup_ctx()
functions was already applied before this driver was added. So convert
this driver as well.
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
When CONFIG_INPUT is disabled, multiple gspca backend drivers
print compile-time warnings about unused variables:
media/usb/gspca/cpia1.c: In function 'sd_stopN':
media/usb/gspca/cpia1.c:1627:13: error: unused variable 'sd' [-Werror=unused-variable]
media/usb/gspca/konica.c: In function 'sd_stopN':
media/usb/gspca/konica.c:246:13: error: unused variable 'sd' [-Werror=unused-variable]
This annotates the variables as __maybe_unused, to let the compiler
know that they are declared intentionally.
Fixes: ee186fd96a5f ("[media] gscpa_t613: Add support for the camera button") Fixes: c2f644aeeba3 ("[media] gspca_cpia1: Add support for button") Fixes: b517af722860 ("V4L/DVB: gspca_konica: New gspca subdriver for konica chipset using cams") Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Hans Verkuil [Mon, 4 Jul 2016 13:18:10 +0000 (10:18 -0300)]
[media] MAINTAINERS: change maintainer for gscpa/pwc/radio-shark
Hans de Goede has no more time to work on those, so I'll take over.
For gspca/pwc I'll do 'Odd Fixes', for radio-shark I'll be a
full maintainer.
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Acked-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
The workqueue "workqueue" is involved in transmitting hdpvr buffers.
It has a single work item(&dev->worker) and hence doesn't require
ordering. Also, it is not being used on a memory reclaim path. Hence,
the singlethreaded workqueue has been replaced with the use of system_wq.
System workqueues have been able to handle high level of concurrency
for a long time now and hence it's not required to have a singlethreaded
workqueue just to gain concurrency. Unlike a dedicated per-cpu workqueue
created with create_singlethread_workqueue(), system_wq allows multiple
work items to overlap executions even on the same CPU; however, a
per-cpu workqueue doesn't have any CPU locality or global ordering
guarantee unless the target CPU is explicitly specified and thus the
increase of local concurrency shouldn't make any difference.
Work item has been flushed in hdpvr_device_release() to ensure
that there are no pending tasks while disconnecting the driver.
The workqueue "work_queues" enables hotplugging.
It has a single work item(&state->delayed_work_enable_hotplug) and hence
doesn't require ordering. Also, it is not being used on a memory
reclaim path. Hence, the singlethreaded workqueue has been replaced with
the use of system_wq.
System workqueues have been able to handle high level of concurrency
for a long time now and hence it's not required to have a singlethreaded
workqueue just to gain concurrency. Unlike a dedicated per-cpu workqueue
created with create_singlethread_workqueue(), system_wq allows multiple
work items to overlap executions even on the same CPU; however, a
per-cpu workqueue doesn't have any CPU locality or global ordering
guarantee unless the target CPU is explicitly specified and thus the
increase of local concurrency shouldn't make any difference.
Work item has been sync cancelled in adv76xx_remove() to ensure
that there are no pending tasks while disconnecting the driver.
The workqueue "work_queues" enables hotplugging.
It has a single work item(&state->delayed_work_enable_hotplug) and hence
doesn't require ordering. Also, it is not being used on a memory
reclaim path. Hence, the singlethreaded workqueue has been replaced with
the use of system_wq.
System workqueues have been able to handle high level of concurrency
for a long time now and hence it's not required to have a singlethreaded
workqueue just to gain concurrency. Unlike a dedicated per-cpu workqueue
created with create_singlethread_workqueue(), system_wq allows multiple
work items to overlap executions even on the same CPU; however, a
per-cpu workqueue doesn't have any CPU locality or global ordering
guarantee unless the target CPU is explicitly specified and thus the
increase of local concurrency shouldn't make any difference.
Work item has been sync cancelled in tc358743_remove() to ensure
that there are no pending tasks while disconnecting the driver.
The workqueue "work_queues" enables hotplugging.
It has a single work item(&state->delayed_work_enable_hotplug) and hence
doesn't require ordering. Also, it is not being used on a memory
reclaim path. Hence, the singlethreaded workqueue has been replaced with
the use of system_wq.
System workqueues have been able to handle high level of concurrency
for a long time now and hence it's not required to have a singlethreaded
workqueue just to gain concurrency. Unlike a dedicated per-cpu workqueue
created with create_singlethread_workqueue(), system_wq allows multiple
work items to overlap executions even on the same CPU; however, a
per-cpu workqueue doesn't have any CPU locality or global ordering
guarantee unless the target CPU is explicitly specified and thus the
increase of local concurrency shouldn't make any difference.
Work item has been sync cancelled in adv7842_remove() to ensure
that there are no pending tasks while disconnecting the driver.
The workqueue "work_thread" is involved in JPEG quality update.
It has a single work item(&sd->work) and hence doesn't require ordering.
Also, it is not being used on a memory reclaim path. Hence, the
singlethreaded workqueue has been replaced with the use of system_wq.
System workqueues have been able to handle high level of concurrency
for a long time now and hence it's not required to have a singlethreaded
workqueue just to gain concurrency. Unlike a dedicated per-cpu workqueue
created with create_singlethread_workqueue(), system_wq allows multiple
work items to overlap executions even on the same CPU; however, a
per-cpu workqueue doesn't have any CPU locality or global ordering
guarantee unless the target CPU is explicitly specified and thus the
increase of local concurrency shouldn't make any difference.
Work item has been flushed in sd_stop0() to ensure that there are no
pending tasks while disconnecting the driver.
When setting the frame time, gspca uses v4l2_get_timestamp() which uses
ktime_get_ts() which uses ktime_get_ts64() which returns a monotonic
timestamp, so it's safe to initialize the buffer flags to
V4L2_BUF_FLAG_TIMESTAMP_MONOTONIC to fix the failure.
Signed-off-by: Antonio Ospite <ao2@ao2.it> Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Antonio Ospite [Wed, 9 Mar 2016 16:03:18 +0000 (13:03 -0300)]
[media] gspca: fix a v4l2-compliance failure about VIDIOC_ENUM_FRAMEINTERVALS
According to v4l2-compliance VIDIOC_ENUM_FRAMEINTERVALS should fail for
unsupported frame sizes, but gspca is too tolerant and tries to find
the frame intervals for the frame size nearest to the requested one.
This makes v4l2-compliance fail with this message:
fail: v4l2-test-formats.cpp(123): \
found frame intervals for invalid size 321x240
test VIDIOC_ENUM_FMT/FRAMESIZES/FRAMEINTERVALS: FAIL
Fix this by using an exact match for the frame size when enumerating
frame intervals, and retuning an error if the frame size for which the
frame intervals have been asked is not supported.
Signed-off-by: Antonio Ospite <ao2@ao2.it> Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Antonio Ospite [Wed, 9 Mar 2016 16:03:15 +0000 (13:03 -0300)]
[media] gspca: ov534/topro: use a define for the default framerate
When writing the change in commit dcc7fdbec53a ("[media] gspca:
ov534/topro: prevent a division by 0") I used magic numbers for the
default framerate to minimize the code footprint to make it easier to
backport the patch to the stable trees.
However it's better if the default framerate has its own define to avoid
risking using different values in different places, and for readability.
While at it also remove some trivial comments about the framerates which
don't add much to the code anymore.
Signed-off-by: Antonio Ospite <ao2@ao2.it> Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Arnd Bergmann [Tue, 26 Jan 2016 16:17:24 +0000 (14:17 -0200)]
[media] pwc: hide unused label
The pwc driver causes a warning when CONFIG_USB_PWC_INPUT_EVDEV is unset:
drivers/media/usb/pwc/pwc-if.c: In function 'usb_pwc_probe':
drivers/media/usb/pwc/pwc-if.c:1115:1: warning: label 'err_video_unreg' defined but not used [-Wunused-label]
This moves the unused label and code inside another #ifdef to
get rid of the warning.
Hans Verkuil [Fri, 15 Apr 2016 12:15:05 +0000 (09:15 -0300)]
[media] vb2: replace void *alloc_ctxs by struct device *alloc_devs
Make this a proper typed array. Drop the old allocate context code since
that is no longer used.
Note that the memops functions now get a struct device pointer instead of
the struct device ** that was there initially (actually a void pointer to
a struct containing only a struct device pointer).
This code is now a lot cleaner.
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Cc: Sakari Ailus <sakari.ailus@iki.fi> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Hans Verkuil [Wed, 16 Dec 2015 10:30:43 +0000 (08:30 -0200)]
[media] vb2: add a dev field to use for the default allocation context
The allocation context is nothing more than a per-plane device pointer
to use when allocating buffers. So just provide a dev pointer in vb2_queue
for that purpose and drivers can skip allocating/releasing/filling in
the allocation context unless they require different per-plane device
pointers as used by some Samsung SoCs.
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Cc: Sakari Ailus <sakari.ailus@iki.fi> Cc: Florian Echtler <floe@butterbrot.org> Cc: Federico Vaga <federico.vaga@gmail.com> Cc: "Lad, Prabhakar" <prabhakar.csengg@gmail.com> Cc: Scott Jiang <scott.jiang.linux@gmail.com> Acked-by: Philipp Zabel <p.zabel@pengutronix.de> Cc: Fabien Dessenne <fabien.dessenne@st.com> Acked-by: Benoit Parrot <bparrot@ti.com> Cc: Mikhail Ulyanov <mikhail.ulyanov@cogentembedded.com> Cc: Guennadi Liakhovetski <g.liakhovetski@gmx.de> Cc: Jonathan Corbet <corbet@lwn.net> Cc: Ludovic Desroches <ludovic.desroches@atmel.com> Cc: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com> Cc: Kyungmin Park <kyungmin.park@samsung.com> Cc: Sylwester Nawrocki <s.nawrocki@samsung.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Hans Verkuil [Fri, 15 Apr 2016 12:15:05 +0000 (09:15 -0300)]
[media] vb2: move dma_attrs to vb2_queue
Make the dma attributes struct part of vb2_queue. This greatly simplifies
the remainder of the patch series since the dma_contig alloc context is
now (as before) just a struct device pointer.
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com> Cc: Sakari Ailus <sakari.ailus@iki.fi> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
[media] exynos4-is: Fix buffer release issue on fimc m2m video nodes
This fixes dropping ownership of buffers in the driver's stop_streaming
callback, so buffers on the memory-to-memory video nodes are properly
released, also in case when the driver has a buffer only on one of the
queues (OUTPUT, CAPTURE) before the video node close.
The issue was being reported by videobuf2 with a following warning while
checking q->owned_by_drv_count:
[ 2498.310766] WARNING: CPU: 0 PID: 9358 at drivers/media/v4l2-core/videobuf2-core.c:1818 __vb2_queue_cancel+0xe8/0x14c
[ 2498.320258] Modules linked in:
[ 2498.323212] CPU: 0 PID: 9358 Comm: v4l2_decode Not tainted 4.7.0-rc4-next-20160627 #1210
[ 2498.331284] Hardware name: SAMSUNG EXYNOS (Flattened Device Tree)
[ 2498.331327] [<c010d738>] (unwind_backtrace) from [<c010a4b0>] (show_stack+0x10/0x14)
[ 2498.331344] [<c010a4b0>] (show_stack) from [<c031a4ac>] (dump_stack+0x74/0x94)
[ 2498.331358] [<c031a4ac>] (dump_stack) from [<c011a52c>] (__warn+0xd4/0x100)
[ 2498.331369] [<c011a52c>] (__warn) from [<c011a578>] (warn_slowpath_null+0x20/0x28)
[ 2498.331381] [<c011a578>] (warn_slowpath_null) from [<c04ed420>] (__vb2_queue_cancel+0xe8/0x14c)
[ 2498.331395] [<c04ed420>] (__vb2_queue_cancel) from [<c04ee10c>] (vb2_core_queue_release+0x18/0x38)
[ 2498.331406] [<c04ee10c>] (vb2_core_queue_release) from [<c04eab50>] (v4l2_m2m_ctx_release+0x1c/0x28)
[ 2498.331420] [<c04eab50>] (v4l2_m2m_ctx_release) from [<c04fe184>] (fimc_m2m_release+0x24/0x78)
[ 2498.331437] [<c04fe184>] (fimc_m2m_release) from [<c04d76c8>] (v4l2_release+0x34/0x74)
[ 2498.331455] [<c04d76c8>] (v4l2_release) from [<c01dc8d4>] (__fput+0x80/0x1bc)
[ 2498.331469] [<c01dc8d4>] (__fput) from [<c0132edc>] (task_work_run+0xc0/0xe4)
[ 2498.331482] [<c0132edc>] (task_work_run) from [<c011d460>] (do_exit+0x304/0xa24)
[ 2498.331493] [<c011d460>] (do_exit) from [<c011dccc>] (do_group_exit+0x3c/0xbc)
[ 2498.331505] [<c011dccc>] (do_group_exit) from [<c0126cac>] (get_signal+0x200/0x65c)
[ 2498.331517] [<c0126cac>] (get_signal) from [<c010e928>] (do_signal+0x84/0x3c4)
[ 2498.331532] [<c010e928>] (do_signal) from [<c010a0ec>] (do_work_pending+0xa4/0xb4)
[ 2498.331545] [<c010a0ec>] (do_work_pending) from [<c0107954>] (slow_work_pending+0xc/0x20)
[media] s5p-mfc: improve v4l2_capability driver and card fields
According to the V4L2 documentation the driver and card fields should be
used to identify the driver and the device but the s5p-mfc driver fills
those field using the platform device name, which in turn is the name of
the device DT node.
So not only the filled information isn't correct but also the same values
are used in all the fields for both the encoder and decoder video devices.
Before this patch:
Driver Info (not using libv4l2):
Driver name : 11000000.codec
Card type : 11000000.codec
Bus info : platform:11000000.codec
Driver version: 4.7.0
Driver Info (not using libv4l2):
Driver name : 11000000.codec
Card type : 11000000.codec
Bus info : platform:11000000.codec
Driver version: 4.7.0
After this patch:
Driver Info (not using libv4l2):
Driver name : s5p-mfc
Card type : s5p-mfc-dec
Bus info : platform:11000000.codec
Driver version: 4.7.0
Driver Info (not using libv4l2):
Driver name : s5p-mfc
Card type : s5p-mfc-enc
Bus info : platform:11000000.codec
Driver version: 4.7.0
Signed-off-by: Javier Martinez Canillas <javier@osg.samsung.com> Acked-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Kamil Debski <k.debski@samsung.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
[media] s5p-mfc: use vb2_is_streaming() to check vb2 queue status
The streaming field in struct vb2_queue is meant to be private and should
not be used by drivers directly, instead the vb2_is_streaming() function
should be used to check the videobuf2 queue streaming status.
[media] s5p-mfc: don't print errors on VIDIOC_REQBUFS unsupported mem type
The V4L2 documentation says that applications must call the VIDIOC_REQBUFS
ioctl to determine if a memory mapped, user pointer or DMABUF based I/O is
supported by the driver.
For example GStreamer does this by first calling VIDIOC_REQBUFS with count
zero for all the possible streaming I/O methods and then finally doing the
real VIDIOC_REQBUFS with count N using a known to be supported memory type.
But the driver prints an error on VIDIOC_REQBUFS if the memory type is not
supported which leads to the following errors that can confuse the users:
[ 178.704390] vidioc_reqbufs:575: Only V4L2_MEMORY_MMAP is supported
[ 178.704666] vidioc_reqbufs:575: Only V4L2_MEMORY_MMAP is supported
[ 178.714956] vidioc_reqbufs:575: Only V4L2_MEMORY_MMAP is supported
[ 178.715229] vidioc_reqbufs:575: Only V4L2_MEMORY_MMAP is supported
Shuah Khan [Tue, 28 Jun 2016 19:17:18 +0000 (16:17 -0300)]
[media] s5p-mfc: fix null pointer deference in clk_core_enable()
Fix null pointer deference in clk_core_enable() when driver unbind is run
when there is an application has an active pipeline playing.
s5p_mfc_release() gets called after s5p_mfc_final_pm() disables and does
clk_put() and s5p_mfc_release() attempts to enable clock and runs into
null pointer deference accessing invalid pointer.
Shuah Khan [Tue, 28 Jun 2016 19:17:16 +0000 (16:17 -0300)]
[media] s5p-mfc: fix video device release double release in probe error path
Fix Decoder and encoder video device double release in probe error path.
video_device_release(dev->vfd_dec) get called twice if decoder register
fails. Also, video_device_release(dev->vfd_enc) get called twice if encoder
register fails.
Tiffany Lin [Tue, 3 May 2016 10:11:24 +0000 (07:11 -0300)]
[media] vcodec: mediatek: Add Mediatek V4L2 Video Encoder Driver
Add v4l2 layer encoder driver for MT8173
Signed-off-by: Tiffany Lin <tiffany.lin@mediatek.com>
[hans.verkuil@cisco.com: drop unnecessary ARM || ARM64 dependency] Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
projects / karo-tx-linux.git / log