Hans Verkuil [Fri, 9 Dec 2016 14:00:49 +0000 (12:00 -0200)]
[media] cec: fix race between configuring and unconfiguring
This race was discovered by running cec-compliance -A with the cec module debug
parameter set to 2: suddenly the test would fail.
It turns out that this happens when the test configures the adapter in
non-blocking mode, then it waits for the CEC_EVENT_STATE_CHANGE event and once
the event is received it unconfigures the adapter.
What happened was that the unconfigure was executed while the configure was
still transmitting the Report Features and Report Physical Address messages.
This messed up the internal state of the cec_adapter.
The fix is to transmit those messages with the adap->lock mutex held (this will
just queue them up in the internal transmit queue, and not actually transmit
anything yet). Only unlock the mutex once everything is done. The main thread
will dequeue the messages from the internal transmit queue and transmit them
one by one, unless an unconfigure was done, and in that case any messages are
just dropped.
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Hans Verkuil [Fri, 9 Dec 2016 13:54:06 +0000 (11:54 -0200)]
[media] cec: update log_addr[] before finishing configuration
The loop that sets the unused logical addresses to INVALID should be
done before 'configured' is set to true. This ensures that cec_log_addrs
is consistent before it will be used.
Signed-off-by: Hans Verkuil <hansverk@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Hans Verkuil [Fri, 9 Dec 2016 13:14:32 +0000 (11:14 -0200)]
[media] cec: when canceling a message, don't overwrite old status info
When a pending message was canceled (e.g. due to a timeout), then the
old tx_status info was overwritten instead of ORed. The same happened
with the tx_error_cnt field. So just modify them instead of overwriting
them.
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Hans Verkuil [Tue, 6 Dec 2016 13:17:12 +0000 (11:17 -0200)]
[media] cec: fix report_current_latency
In the (very) small print of the REPORT_CURRENT_LATENCY message there is a
line that says that the last byte of the message (audio out delay) is only
present if the 'audio out compensated' value is 3.
I missed this, and so if this message was sent with a total length of 6 (i.e.
without the audio out delay byte), then it was rejected by the framework
since a minimum length of 7 was expected.
Fix this minimum length check and update the wrappers in cec-funcs.h to do
the right thing based on the message length.
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Sakari Ailus [Sat, 19 Nov 2016 21:50:10 +0000 (19:50 -0200)]
[media] smiapp: Make suspend and resume functions __maybe_unused
The smiapp_suspend() and smiapp_resume() functions will end up being unused
if CONFIG_PM is enabled but CONFIG_PM_SLEEP is disabled, causing a
compiler warning from both of the function definitions. Fix this by
marking the functions with __maybe_unused.
[media] v4l: tvp5150: Compile tvp5150_link_setup out if !CONFIG_MEDIA_CONTROLLER
The function is only referenced as a handler in the tvp5150_sd_media_ops
structure, which is only used when CONFIG_MEDIA_CONTROLLER is set. Don't
define the function and the structure when the configuration option is
unset to avoid an unused function warning.
[media] em28xx: don't store usb_device at struct em28xx
Now that we're storing usb_interface at em28xx struct,
there's no good reason to keep storing usb_device, as we can
get it from usb_interface. So, get rid of it.
Antti Palosaari [Thu, 1 Dec 2016 00:08:27 +0000 (22:08 -0200)]
[media] mn88472: fix chip id check on probe
A register used to identify chip during probe was overwritten during
firmware download and due to that later probe's for warm chip were
failing. Detect chip from the another register, which is located on
different register bank 2.
Fixes: 94d0eaa41987 ("[media] mn88472: move out of staging to media") Cc: <stable@vger.kernel.org> # v4.8+ Signed-off-by: Antti Palosaari <crope@iki.fi> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Antti Palosaari [Wed, 30 Nov 2016 21:36:14 +0000 (19:36 -0200)]
[media] mn88473: fix chip id check on probe
A register used to identify chip during probe was overwritten during
firmware download and due to that later probe's for warm chip were
failing. Detect chip from the another register, which is located on
different register bank 2.
Sean Young [Sat, 26 Nov 2016 21:31:24 +0000 (19:31 -0200)]
[media] lirc: fix error paths in lirc_cdev_add()
"c77d17c0 [media] lirc: use-after free" introduces two problems:
cdev_del() can be called with a NULL argument, and the kobject_put()
path will cause a double free.
Reported-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Sean Young <sean@mess.org> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Marek Szyprowski [Wed, 16 Nov 2016 09:04:57 +0000 (07:04 -0200)]
[media] s5p-mfc: Rework clock handling
This patch changes the code for handling clocks. Now clocks are defined
per each device variant, what is a preparation for adding support for
Exynos 5433 MFC V8, which has more clocks than all previous versions.
Also use devm_clk_get() to simplify cleanup path.
Marek Szyprowski [Wed, 16 Nov 2016 09:04:56 +0000 (07:04 -0200)]
[media] s5p-mfc: Don't keep clock prepared all the time
This patch moves preparation of clocks from s5p_mfc_init_pm()
(driver probe) to s5p_mfc_power_on() (start of device operation).
This change will allow to use runtime power usage optimization
on newer Samsung Exynos platforms (for example Exynos 5433).
Marek Szyprowski [Wed, 16 Nov 2016 09:04:55 +0000 (07:04 -0200)]
[media] s5p-mfc: Kill all IS_ERR_OR_NULL in clocks management code
After commit "s5p-mfc: Fix clock management in s5p_mfc_release function"
all clocks related functions are called only when MFC device is really
available, so there is no additional check needed for NULL
gate clocks. This patch simplifies the code and kills IS_ERR_OR_NULL
macro usage.
Marek Szyprowski [Wed, 16 Nov 2016 09:04:53 +0000 (07:04 -0200)]
[media] s5p-mfc: Ensure that clock is disabled before turning power off
Move clock disabling before turning power off. This will enable later
to add calls to clk_prepare/unprepare in the s5p_mfc_power_off() function
to avoid keeping clocks prepared all the time when driver is bound.
Marek Szyprowski [Wed, 16 Nov 2016 09:04:52 +0000 (07:04 -0200)]
[media] s5p-mfc: Remove special clock rate management
The maximum rate of special clock depends on SoC variant and should
be set in device tree via assigned-clock-rates property, so remove
the code which forces special clock to 200MHz.
Marek Szyprowski [Wed, 16 Nov 2016 09:04:51 +0000 (07:04 -0200)]
[media] s5p-mfc: Use printk_ratelimited for reporting ioctl errors
Some applications don't check error codes from QBUF/DQBUF ioctls,
so don't spam kernel log with errors if they fall into endless loop
trying to queue next buffer after a failure.
Douglas Anderson [Wed, 16 Nov 2016 09:04:50 +0000 (07:04 -0200)]
[media] s5p-mfc: Set DMA_ATTR_ALLOC_SINGLE_PAGES
We do video allocation all the time and we need it to be fast. Plus TLB
efficiency isn't terribly important for video.
That means we want to set DMA_ATTR_ALLOC_SINGLE_PAGES.
See also the previous change (commit 14d3ae2efeed "ARM: dma-mapping: Use
DMA_ATTR_ALLOC_SINGLE_PAGES hint to optimize allocation").
[m.szyprowski: rebased patch onto v4.9-rc1 and adapted changes
to latest videbuf2 changes, this simplifies code changes to
only set proper dma attribute flag and comment the reason
for it, added commit id of arch/arm/mm patch]
Signed-off-by: Douglas Anderson <dianders@chromium.org> Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com> Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Ricardo Ribalda [Tue, 15 Nov 2016 12:06:24 +0000 (10:06 -0200)]
[media] v4l2-tpg: Init hv_enc field with a valid value
Zero is not a valid value for hsv_enc. Set the field to a valid
initial value.
This is not a problem for vivid, because it sets the field to 180 via
tpg_s_hsv_enc() on the control initialization, but it might be a source
of errors for other drivers that use this code.
Dan Carpenter [Fri, 25 Nov 2016 10:28:35 +0000 (08:28 -0200)]
[media] uvcvideo: freeing an error pointer
A recent cleanup introduced a potential dereference of -EFAULT when we
call kfree(map->menu_info).
Fixes: 4cc5bed1caeb ("[media] uvcvideo: Use memdup_user() rather than duplicating its implementation") Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Arnd Bergmann [Tue, 22 Nov 2016 20:52:17 +0000 (18:52 -0200)]
[media] DaVinci-VPFE-Capture: fix error handling
A recent cleanup had the right idea to remove the initialization
of the error variable, but missed the actual benefit of that,
which is that we get warnings if there is a bug in it. Now
we get a warning about a bug that was introduced by this cleanup:
drivers/media/platform/davinci/vpfe_capture.c: In function 'vpfe_probe':
drivers/media/platform/davinci/vpfe_capture.c:1992:9: error: 'ret' may be used uninitialized in this function [-Werror=maybe-uninitialized]
This adds the missing initialization that the warning is about,
and another one that was preexisting and that we did not get
a warning for. That second bug has existed since the driver
was first added.
Fixes: efb74461f5a6 ("[media] DaVinci-VPFE-Capture: Delete an unnecessary variable initialisation in vpfe_probe()") Fixes: 7da8a6cb3e5b ("V4L/DVB (12248): v4l: vpfe capture bridge driver for DM355 and DM6446")
[mchehab@s-opensource.com: fix a merge conflict] Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Usually, I don't like fixing coding style issues on non-staging
drivers, as it could be a mess pretty easy, and could become like
a snow ball. That's the case of recent changes on two changesets:
they disalign some statements. Yet, a care a lot with cx88 driver,
as it was the first driver I touched at the Kernel, and I've been
maintaining it since 2005. So, several of the coding style issues
were due to my code.
Per Andrey's suggestion, I ran checkpatch.pl in strict mode, with
fixed several other issues, did some function alinments, but broke
other alinments.
So, I had to manually apply another round of manual fixes to make
sure that everything is ok, and to make checkpatch happy with
this patch.
With this patch, checkpatch.pl is now happy when called with:
./scripts/checkpatch.pl -f --max-line-length=998 --ignore PREFER_PR_LEVEL
Also, the 80-cols violations that made sense were fixed.
Checkpatch would be happier if we convert it to use dev_foo(),
but this is a more complex change.
NOTE: there are some places with msleep(1). As this driver was
written at the time that the default was to sleep at least 10ms
on such calls (e. g. CONFIG_HZ=100), I replaced those calls by
usleep_range(10000, 20000), with should be safe to avoid breakages.
Fixes: 65bc2fe86e66 ("[media] cx88: convert it to use pr_foo() macros") Fixes: 7b61ba8ff838 ("[media] cx88: make checkpatch happier") Suggested-by: Andrey Utkin <andrey_utkin@fastmail.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com> Reviewed-by: Andrey Utkin <andrey_utkin@fastmail.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Arnd Bergmann [Fri, 18 Nov 2016 16:16:05 +0000 (14:16 -0200)]
[media] v4l: rcar_fdp1: add FCP dependency
Commit 4710b752e029 ("[media] v4l: Add Renesas R-Car FDP1 Driver") in the
v4l-dvb tree adds CONFIG_VIDEO_RENESAS_FDP1.
It calls into the FCP driver, but when there is no dependency, FCP might
be a module while FDP1 is built-in.
We have the same logic in VIDEO_RENESAS_VSP1, which also depends on
FCP not being a module when it is built-in itself.
drivers/media/platform/rcar_fdp1.o: In function `fdp1_pm_runtime_resume':
rcar_fdp1.c:(.text.fdp1_pm_runtime_resume+0x78): undefined reference to `rcar_fcp_enable'
drivers/media/platform/rcar_fdp1.o: In function `fdp1_pm_runtime_suspend':
rcar_fdp1.c:(.text.fdp1_pm_runtime_suspend+0x14): undefined reference to `rcar_fcp_disable'
drivers/media/platform/rcar_fdp1.o: In function `fdp1_probe':
rcar_fdp1.c:(.text.fdp1_probe+0x15c): undefined reference to `rcar_fcp_get'
Arnd Bergmann [Fri, 18 Nov 2016 16:16:04 +0000 (14:16 -0200)]
[media] v4l: rcar_fdp1: mark PM functions as __maybe_unused
The new driver produces a warning when CONFIG_PM is disabled:
platform/rcar_fdp1.c:2408:12: error: 'fdp1_pm_runtime_resume' defined but not used [-Werror=unused-function]
platform/rcar_fdp1.c:2399:12: error: 'fdp1_pm_runtime_suspend' defined but not used [-Werror=unused-function]
Shuah Khan [Wed, 16 Nov 2016 20:49:50 +0000 (18:49 -0200)]
[media] media: remove obsolete Media Device Managed resource interfaces
Remove obsolete media_device_get_devres(), media_device_find_devres(),
and media_device_release_devres() interfaces. These interfaces are now
obsolete.
[media] dvb_net: prepare to split a very complex function
The dvb_net code has a really complex function, meant to handle
DVB network packages: it is long, has several loops and ifs
inside, and even cause warnings with gcc5.
Prepare it to be split into smaller functions by storing all
arguments and internal vars inside a struct.
When building with "make W=1", we get multiple harmless build warnings
for the vpfe driver:
drivers/staging/media/davinci_vpfe/dm365_resizer.c:241:1: error: 'static' is not at beginning of declaration [-Werror=old-style-declaration]
drivers/staging/media/davinci_vpfe/dm365_resizer.c: In function 'resizer_set_defualt_configuration':
drivers/staging/media/davinci_vpfe/dm365_resizer.c:831:16: error: initialized field overwritten [-Werror=override-init]
drivers/staging/media/davinci_vpfe/dm365_resizer.c:831:16: note: (near initialization for 'rsz_default_config.rsz_rsc_param[0].h_typ_c')
drivers/staging/media/davinci_vpfe/dm365_resizer.c:849:16: error: initialized field overwritten [-Werror=override-init]
drivers/staging/media/davinci_vpfe/dm365_resizer.c:849:16: note: (near initialization for 'rsz_default_config.rsz_rsc_param[1].h_typ_c')
All of them are trivial to fix without changing the behavior of the
driver, as "static const" is interpreted the same as "const static",
and VPFE_RSZ_INTP_CUBIC is defined as zero, so the initializations
are not really needed.
It has never been possible to actually build this driver as
a loadable module, only built-in because the Makefile attempts
to build each file into its own module and fails:
It took a long time to catch this bug with randconfig builds
because at least 14 other Kconfig symbols have to be enabled in
order to configure this one, and it was clearly only ever tested
as built-in with mainline kernels, if at all.
The solution is really easy: this patch changes the Makefile to
link all files into one module. As discussed previously, the
driver has never before used successfully as a loadable module,
but there is no reason to prevent that configuration.
Shuah Khan [Mon, 14 Mar 2016 22:24:25 +0000 (19:24 -0300)]
[media] media: Update documentation for media_entity_notify
Update documentation for media_entity_notify to clearly state the usage
restrictions. This handler is intended for creating links between exiting
entities and should not used to create and register entities.
Arnd Bergmann [Tue, 26 Jan 2016 14:09:59 +0000 (12:09 -0200)]
[media] em28xx: only use mt9v011 if camera support is enabled
In randconfig builds that select VIDEO_EM28XX_V4L2 and
MEDIA_SUBDRV_AUTOSELECT, but not MEDIA_CAMERA_SUPPORT, we get
a Kconfig warning:
warning: (VIDEO_EM28XX_V4L2) selects VIDEO_MT9V011 which has unmet direct dependencies (MEDIA_SUPPORT && I2C && VIDEO_V4L2 && MEDIA_CAMERA_SUPPORT)
This avoids the warning by making that 'select' conditional on
MEDIA_CAMERA_SUPPORT. Alternatively we could mark EM28XX as
'depends on MEDIA_CAMERA_SUPPORT', but it does not seem to
have any real dependency on that itself.
If MEDIA_SUBDRV_AUTOSELECT and VIDEO_GO7007 are both set, we
automatically select VIDEO_OV7640, but that depends on MEDIA_CAMERA_SUPPORT,
so we get a Kconfig warning if that is disabled:
warning: (VIDEO_GO7007) selects VIDEO_OV7640 which has unmet direct dependencies (MEDIA_SUPPORT && I2C && VIDEO_V4L2 && MEDIA_CAMERA_SUPPORT)
This adds another dependency so we don't accidentally select
it when it is unavailable.
The systime function uses struct timespec, which we want to stop
using in the kernel because it overflows in 2038. Fortunately,
this use in dibx000_common is in a function that is never called,
so we can just remove it.
Joerg Riechardt [Sat, 29 Aug 2015 15:02:57 +0000 (12:02 -0300)]
[media] stv090x: use lookup tables for carrier/noise ratio
The stv090x driver uses the lookup table for signal strength already,
with this patch we use the lookup tables for carrier/noise ratio as well.
This has the advantage, that values for DVB-S and DVB-S2 are now
corresponding, while before they were way off. The values are now
proportional to real carrier/noise ratio, while before they were
corresponding to register values. So now applications are able to give
the user real carrier/noise ratio.
Because the output has to be within 0x0000...0xFFFF the three negative
values for DVB-S2 are omitted. This is no significant loss, because
reception is lost at 7.5 dB already (TT S2-1600, Cine S2), so the
negative values are not really important, and also for DVB-S they don´t
exist.
Dan Carpenter [Wed, 16 Nov 2016 11:11:54 +0000 (09:11 -0200)]
[media] stk-webcam: fix an endian bug in stk_camera_read_reg()
We pass an int pointer to stk_camera_read_reg() but only write to the
highest byte. It's a bug on big endian systems and generally a nasty
thing to do and doesn't match the write function either.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Implement DVBv5 statistics support for DVB-T, DVB-T2 and DVB-C. All
information was taken from the LinuxTV wiki, where Benjamin Larsson has
documented all registers:
https://www.linuxtv.org/wiki/index.php/Panasonic_MN88472
Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com> Signed-off-by: Antti Palosaari <crope@iki.fi> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Sakari Ailus [Tue, 8 Nov 2016 16:06:44 +0000 (14:06 -0200)]
[media] v4l: compat: Prevent allocating excessive amounts of memory
get_v4l2_ext_controls32() is used to convert the 32-bit compat struct into
native 64-bit representation. The function multiplies the array length by
the entry length before validating size. Perform the size validation
first.
Also use unsigned values for size computation.
Make similar changes to get_v4l2_buffer32() for multi-plane buffers.
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
When compiled on i386, it produces several warnings:
./arch/x86/include/asm/bitops.h:457:22: warning: asm output is not an lvalue
./arch/x86/include/asm/bitops.h:457:22: warning: asm output is not an lvalue
./arch/x86/include/asm/bitops.h:457:22: warning: asm output is not an lvalue
./arch/x86/include/asm/bitops.h:457:22: warning: asm output is not an lvalue
./arch/x86/include/asm/bitops.h:457:22: warning: asm output is not an lvalue
./arch/x86/include/asm/bitops.h:457:22: warning: asm output is not an lvalue
I suspect that some gcc optimization could be causing the asm code to be
incorrectly generated. Splitting it into two macro calls fix the issues
and gets us rid of 6 smatch warnings, with is a good thing. As it should
not cause any troubles, as we're basically doing the same thing, let's
apply such change to vpe.c.
Despite being exported, there's no prototype for it at the
headers, as warned by sparse:
Fixes this sparse warning:
drivers/media/platform/ti-vpe/vpdma.c:1000:6: warning: no previous prototype for 'vpdma_enable_list_notify_irq' [-Wmissing-prototypes]
void vpdma_enable_list_notify_irq(struct vpdma_data *vpdma, int irq_num,
^~~~~~~~~~~~~~~~~~~~~~~~~~~~
Worse than that, it is not even used, as making it static it
would produce:
drivers/media/platform/ti-vpe/vpdma.c:1000:13: warning: 'vpdma_enable_list_notify_irq' defined but not used [-Wunused-function]
static void vpdma_enable_list_notify_irq(struct vpdma_data *vpdma, int irq_num,
^~~~~~~~~~~~~~~~~~~~~~~~~~~~
So, let's just get rid of the dead code. If needed in the future,
someone could re-add it.
Hans Verkuil [Mon, 21 Nov 2016 13:59:20 +0000 (11:59 -0200)]
[media] vpfe_capture: fix compiler warning
davinci/vpfe_capture.c: In function 'vpfe_probe':
davinci/vpfe_capture.c:1992:9: warning: 'ret' may be used uninitialized
in this function [-Wmaybe-uninitialized]
return ret;
^~~
This is indeed correct, so if the kmalloc fails set ret to -ENOMEM.
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
Benoit Parrot [Fri, 18 Nov 2016 23:20:45 +0000 (21:20 -0200)]
[media] media: ti-vpe: vpe: Add proper support single and multi-plane buffer
The VPE was restricting the number of plane per buffer based on
the fact that if a particular format had color separation it was
meant to need 2 planes.
However NV12/NV16 are color separate format which are meant to be
presented in a single contiguous buffer/plane.
It could also be presented in a multi-plane as well if need be.
So we must support both modes for more flexibility.
The number of plane requested by user space was previously ignored
and was therefore always overwritten.
The driver now use the requested num plane as hint to calculate needed
offset when required.
Benoit Parrot [Fri, 18 Nov 2016 23:20:42 +0000 (21:20 -0200)]
[media] media: ti-vpe: vpdma: Add RAW8 and RAW16 data types
Add RAW8 and RAW16 data type to VPDMA.
To handle RAW format we are re-using the YUV CBY422
vpdma data type so that we use the vpdma to re-order
the incoming bytes, as the VIP parser assumes that the
first byte presented on the bus is the MSB of a 2
bytes value.
RAW8 handles from 1 to 8 bits.
RAW16 handles from 9 to 16 bits.
Benoit Parrot [Fri, 18 Nov 2016 23:20:41 +0000 (21:20 -0200)]
[media] media: ti-vpe: vpe: Make sure frame size dont exceed scaler capacity
When scaler is to be used we need to make sure that the input and
output frame size do not exceed the maximum frame sizes that the
scaler h/w can handle otherwise streaming stall as the scaler
cannot proceed.
The scaler buffer is limited to 2047 pixels (i.e. 11 bits) when
attempting anything larger (2048 for example) the scaler stalls.
Realistically in an mem2mem device we can only check for this type
of issue when start_streaming is called. We can't do it during the
try_fmt/s_fmt because we do not have all of the info needed at that
point. So instead when start_streaming is called we need to check
that the input and output frames size do not exceed the scaler's
capability. The only time larger frame size are allowed is when
the input frame szie is the same as the output frame size.
Now in the case where we need to fail, start_streaming must return
all previously queued buffer back otherwise the vb2 framework
will issue kernel WARN messages.
In this case we also give an error message.
Benoit Parrot [Fri, 18 Nov 2016 23:20:39 +0000 (21:20 -0200)]
[media] media: ti-vpe: Make scaler library into its own module
In preparation to add scaler support into VIP we need to
turn sc.c into its own kernel module.
Add support for multiple SC memory block as VIP contains
2 scaler instances.
This is done by passing the resource name to sc_create() and
modify the vpe invocation accordingly.
[media] media: ti-vpe: vpe: Fix line stride for output motion vector
For deinterlacing operation, VPE hardware uses motion vectors.
MV calculated in the previous iteration are used for next interation.
Therefore driver allocates two motion vectors in ping-pong fashion.
For every transaction, one MV is DMAed in and one is DMAed out.
All the outbound DMAs (DMA to memory) use output parameters, but as
the motion vectors is generated purely out of input fields, it should
use the input parameters for DMA.
Fix the add_out_dtd to use source q_data for creating descriptor.
If the output size is greater than input stride, without this change,
MV DMA may overwrite the buffer causing memory corruption.
This CRITICAL fix ensures that the motion vector DMA descriptor is
created based on the attributes with which the buffer was allocated.
[media] media: ti-vpe: vpdma: Use bidirectional cached buffers
VPDMA buffer will be used by CPU as well as by the VPDMA.
CPU will write/update the VPDMA descriptors containing data
about the video buffer DMA addresses.
VPDMA will write the "write descriptor" containing the
data about the DMA operation.
When mapping/unmapping the buffer, driver has to take care of
WriteBack and invalidation of the cache so that all the
coherency is maintained from both directions.
Use DMA_BIDIRECTIONAL to maintain coherency between CPU and VPDMA.
[media] media: ti-vpe: vpdma: Fix race condition for firmware loading
vpdma_create API is supposed to allocated the struct vpdma_data and
return it to the driver. Also, it would call the callback function
when the VPDMA firmware is loaded.
Typically, VPE driver have following function call:
dev->vpdma = vpdma_create(pdev, firmware_load_callback);
And the callback implementation would continue the probe further.
Also, the dev->vpdma is accessed from the callback implementation.
This may lead to race condition between assignment of dev->vpdma
and the callback function being triggered.
This would lead to kernel crash because of NULL pointer access.
Fix this by passing a driver wrapped &vpdma_data instead of allocating
inside vpdma_create.
Change the vpdma_create prototype accordingly and fix return paths.
Also, update the VPE driver to use the updated API and
initialize the dev->vpdma before hand so that the race condition
is avoided.
When stop_streaming is called we need to cleanup the queued
vb2 buffers properly.
This was not previously being done which caused kernel
warning when the application using the resources was killed.
Kernel warnings were also generated on successful completion
of a de-interlacing case as well as upon aborting a
conversion.
Make sure every vb2 buffers is properly handled in all cases.
Current scaler library implementation of sc_set_hs_coeffs and
sc_set_vs_coeffs tries to return immediately if the calculated
coefficient index is already being used.
As the same scaler block is going to be used for all the VPE contexts,
even if the calculated index is same, the parameters have to be
reconfigured for each of the context.
Because of this, when multiple contexts use the same coefficients,
all other contexts would have zero scaling coefficients.
Fix this and also remove the unnecessary hs_index and vs_index fields.
Benoit Parrot [Fri, 18 Nov 2016 23:20:32 +0000 (21:20 -0200)]
[media] media: ti-vpe: vpdma: RGB data type yield inverted data
The VPDMA RGB data type definition have been updated
to match with Errata i839.
But some of the ARGB definition appeared to be wrong
in the document also. As they would yield RGBA instead.
They have been corrected based on experimentation.
Benoit Parrot [Fri, 18 Nov 2016 23:20:31 +0000 (21:20 -0200)]
[media] media: ti-vpe: vpdma: Corrected YUV422 data type label
The YUV data type definition below are taken from
both the TRM and i839 Errata information.
Use the correct data type considering byte
reordering of components.
Added the 2 missing YUV422 variant.
Also since the single use of "C" in the 422 case
to mean "Cr" (i.e. V component). It was decided
to explicitly label them CR to remove any confusion.
Bear in mind that the type label refer to the memory
packed order (LSB - MSB).
[media] media: ti-vpe: vpdma: allocate and maintain hwlist
VPDMA block used in ti-vip and ti-vpe modules have support for
up to 8 hardware descriptor lists. A descriptor list can be
submitted to any of the 8 lists (as long as it's not busy).
When multiple clients want to transfer data in parallel, its easier
to allocate one list per client and let it use it. This way, the
list numbers need not be hard-coded into the driver.
Add support for allocating hwlist and maintain them with a priv data.
[media] media: ti-vpe: vpe: Add RGB565 and RGB5551 support
VPE hardware can generate output in RGB565 or in RGB5551 format.
Add these formats in the supported format list for CAPTURE stream.
Also, for RGB5551 format, the alpha component is not processed,
so the alpha value is taken from the default color.
Set the default color to make alpha component full when the dst
format is of RGB color space.
[media] media: ti-vpe: vpe: Post next descriptor only for list complete IRQ
vpe_irq checks for the possible interrupt sources and prints the
errors for the DEI_ERROR and DS_UV interrupts. But it also post the
next descriptor list irrespective of whichever interrupt has occurred.
Because of this, driver may release the buffers even before DMA is
complete and also schedule next descriptor list.
Fix this by _actually_ handling the IRQ only when ListComplete IRQ
occurs.
[media] media: ti-vpe: vpe: Setup srcdst parameters in start_streaming
For deinterlacing operation, each operation needs 2 fields in the
history. This is achieved by holding three buffers in
ctx->src_vbs[0,1,2] (f,f-1,f-2)
This is achieved by using the ctx->sequence which gets reset via the
s_fmt ioctl.
These buffers are dequeued in stream OFF by calling free_vbs()
But the corresponding references aren't removed anywhere.
When application tries to stream ON and OFF continuously, s_fmt ioctl
won't be called and it won't setup the srcdst parameters.
Setting source/destination parameters in stream ON ioctl would make
sure that the context is re-initialized before it is being used by
the driver.
[media] media: ti-vpe: vpe: configure line mode separately
Current driver configures the line mode of the DEI clients
from the open function directly. Even if the newly created context
is not yet scheduled, it updates some of the VPDMA registers.
This causes a problem in multi instance use case where just opening
the m2m device second time causes the running job to stall. This
happens especially if the source buffers used are NV12.
While all other configuration is being written to context specific
shadow registers, only line mode configuration is happening directly.
As there is no shadow register for line mode configuration, it's better
to separate the config_mode setting and line_mode setting. Call the
new "set_line_modes" functions only when actually loading the mmrs.
This makes sure that no non-running job will write to the registers
directly.
[media] media: ti-vpe: vpdma: Clear IRQs for individual lists
VPDMA IRQs are registered for multiple lists
When clearing an IRQ for a list interrupt, all the
IRQs for the individual lists are to be cleared separately.
[media] media: ti-vpe: vpdma: Make list post atomic operation
Writing to the "VPDMA list attribute" register is considered as a list
post. This informs the VPDMA firmware to load the list from the address
which should be taken from the "VPDMA list address" register.
As these two register writes are dependent, it is important that the two
writes happen in atomic manner. This ensures multiple slices (which share
same VPDMA) can post lists asynchronously and all of them point to the
correct addresses.
Slightly modified to implementation for the original patch to use
spin_lock instead of mutex as the list post is also called from
interrupt context.
Whenever VPDMA processes a data descriptor of a list, it processes it
and sets up the channel for the DMA transaction. List manager holds the
descriptor in the list until the DMA is complete. If sync_on_channel
descriptor, or another descriptor for the same channel is present in
the FIFO, list manager keeps them until the current channel is free.
When the capture stream is closed suddenly while there are pending
descriptors in the FIFO (streamON failed, application killed), it would
keep the VPDMA in a busy state. Any further list post would fail with
EBUSY.
To avoid this, drivers need to stop the current processing list and
cleanup all the resources VPDMA has taken and also clear the internal FSM
of list manager. The state machine is cleared by issuing channel specific
abort descriptor.
Therefore, the vpdma_list_cleanup accepts an array of channels for which
abort_channel descriptors should be posted. It is driver's responsibility
to post for all the channels or the channels which were used in the last
context.
[media] media: ti-vpe: vpe: Return NULL for invalid buffer type
get_q_data can be called with different values for type
e.g. vpe_try_crop calls it with the buffer type which gets passed
from user space
Framework doesn't check wheather its correct type or not
If user space passes wrong type, kernel should not crash.
Return NULL when the passed type is invalid.
[media] media: ti-vpe: Add support for SEQ_TB buffers
The video source can generate the data in the SEQ_TB buffer format.
In the case of TI SoC, the IVA_HD can generate the interlaced content in
the SEQ_TB buffer format. This is the format where the top and bottom field
data can be contained in a single buffer. For example, for NV12, interlaced
format, the data in Y buffer will be arranged as Y-top followed by
Y-bottom. And likewise for UV plane.
Also, queuing one buffer of SEQ_TB is equivalent to queuing two different
buffers for top and bottom fields. Driver needs to take care of this when
handling source buffer lists.
[media] media: ti-vpe: vpe: Do not perform job transaction atomically
Current VPE driver does not start the job until all the buffers for
a transaction are queued. When running in multiple context, this might
increase the processing latency.
Alternate solution would be to try to continue the same context as long as
buffers for the transaction are ready; else switch the context. This may
increase number of context switches but it reduces latency significantly.
In this approach, the job_ready always succeeds as long as there are
buffers on the CAPTURE and OUTPUT stream. Processing may start immediately
as the first 2 iterations don't need extra source buffers. Shift all the
source buffers after each iteration and remove the oldest buffer.
Also, with this removes the constraint of pre buffering 3 buffers before
call to STREAMON in case of de-interlacing.
projects / karo-tx-linux.git / log