drivers/staging/media/airspy/airspy.c: In function 'airspy_stop_streaming':
drivers/staging/media/airspy/airspy.c:569:6: warning: variable 'ret' set but not used [-Wunused-but-set-variable]
int ret;
^
Cc: Antti Palosaari <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
V4L2_SDR_FMT_CS14LE is complex signed 14-bit sample format, used
for software defined radio devices.
[[email protected]: remove V4L2_FL_USE_FH_PRIO setting to avoid
compilation breakage, as this flag doesn't exist anymore]
Signed-off-by: Antti Palosaari <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
V4L2_SDR_FMT_CS8 is complex signed 8-bit sample format, used for
software defined radio devices.
Signed-off-by: Antti Palosaari <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
AirSpy SDR driver.
Thanks to Youssef Touil and Benjamin Vernoux for support, help and
hardware!
http://airspy.com/
Signed-off-by: Antti Palosaari <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
Document V4L2_SDR_FMT_RU12LE format. It is real unsigned 12-bit
little endian sample inside 16-bit space. Used by software defined
radio devices.
Signed-off-by: Antti Palosaari <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
Implement unlocked variants of v4l2_ctrl_s_ctrl() and
v4l2_ctrl_s_ctrl_int64(). As drivers need to set controls as they access
driver internal state elsewhere than in the control framework unlocked
variants of these functions become handy.
Signed-off-by: Sakari Ailus <[email protected]>
Signed-off-by: Hans Verkuil <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
Drivers may use the v4l2_ctrl_modify_range() internally as part of other
operations that need to be both serialised using a driver's lock which can
also be used to serialise access to the control handler. Provide an unlocked
version of the function, __v4l2_ctrl_modify_range() which then may be used
by drivers for the purpose.
Signed-off-by: Sakari Ailus <[email protected]>
Signed-off-by: Hans Verkuil <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The v4l2_ctrl_{,un}lock will be needed elsewhere. Define them before the
functions that perform operations on controls.
Signed-off-by: Sakari Ailus <[email protected]>
Signed-off-by: Hans Verkuil <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
Fall-out from the recent struct v4l2_framebuffer change.
drivers/media/common/saa7146/saa7146_fops.c: In function ‘saa7146_vv_init’:
drivers/media/common/saa7146/saa7146_fops.c:536:6: warning: assignment from incompatible pointer type [enabled by default]
fmt = &vv->ov_fb.fmt;
^
Signed-off-by: Hans Verkuil <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
[linux-3.16-rc5/drivers/staging/media/davinci_vpfe/dm365_ipipeif.c:210]:
(style) Checking if unsigned variable 'val' is less than zero.
val = get_oneshot_mode(ipipeif->input);
if (val < 0) {
pr_err("ipipeif: links setup required");
return -EINVAL;
}
but
static int get_oneshot_mode(enum ipipeif_input_entity input)
Introduced temporary variable for negativity check.
"val" is afterwards used in a lot of bitwise operations, so changing its type
to signed is not safe, according to CERT C Secure Coding Standards chapter
INT13-C: "Use bitwise operators only on unsigned operands"
https://www.securecoding.cert.org/confluence/display/seccode/INT13-C.+Use+bitwise+operators+only+on+unsigned+operands
Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=80521
Reported-by: David Binderman <[email protected]>
Signed-off-by: Andrey Utkin <[email protected]>
Signed-off-by: Hans Verkuil <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
This driver moved over to v4l2_async_unregister_subdev()
but still retained a call to v4l2_unregister_subdev(). Remove.
Signed-off-by: Ian Molton <[email protected]>
Signed-off-by: Ben Dooks <[email protected]>
Acked-by: Hans Verkuil <[email protected]>
Signed-off-by: Hans Verkuil <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The CPB size is limited by the hardware. Add this limit to the s_fmt.
Signed-off-by: panpan liu <[email protected]>
Signed-off-by: Kamil Debski <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
OVL and BTP IRAM buffers are never used, setup the bits for
for DBK/BIT/IP usage depending on CODA version in one place.
Also, use a simple allocator function and group IRAM addresses
and size in a coda_aux_buf structure.
This is done in preparation for CODA960 support.
Signed-off-by: Philipp Zabel <[email protected]>
Signed-off-by: Kamil Debski <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
If bitrate is not set, the encoder is running in VBR mode, with the
I- and P-frame quantization parameters configured from userspace.
For the quantization parameters, 0 is a valid value.
Signed-off-by: Philipp Zabel <[email protected]>
Signed-off-by: Kamil Debski <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
Currently the rotator unit is used to copy decoded frames out into buffers
provided by videobuf2. Since the CODA reports the I/P/B frame type of the
last decoded frame, and this frame will be copied out in a later device_run,
depending on display order, we have to store the frame type until such time.
This patch also adds the B-frame type.
Signed-off-by: Philipp Zabel <[email protected]>
Signed-off-by: Kamil Debski <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
fc906b6d "Remove special clock usage in driver" removed
initialization of MFC special clock, arguing that there's
no need to do it explicitly, since it's one of MFC gate clock's
dependencies and gets enabled along with it. However, there's
no promise of keeping this hierarchy across Exynos SoC
releases, therefore this approach fails to provide a stable,
portable solution.
Out of all MFC versions, only v6 doesn't use special clock at all.
For other versions log a message only in case clk_get fails,
as not all the devices with the same MFC version require
initializing the clock explicitly.
Signed-off-by: Mateusz Zalega <[email protected]>
Signed-off-by: Seung-Woo Kim <[email protected]>
Signed-off-by: Jacek Anaszewski <[email protected]>
Signed-off-by: Kyungmin Park <[email protected]>
Signed-off-by: Kamil Debski <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
In the s5p_mfc_alloc_firmware() function there are some mistakes
where the code checks whether the DMA memory is properly allocated or
not.
First of all dma_alloc_coherent() returns NULL in case of error.
The code also checked two times fw_virt_addr, ignoring
the bank2_virt pointer.
Signed-off-by: Maurizio Lombardi <[email protected]>
Signed-off-by: Kamil Debski <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The ZR050_MO_COMP define is cut and pasted twice so we can delete the
second instance.
Signed-off-by: Dan Carpenter <[email protected]>
Signed-off-by: Hans Verkuil <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
This patch removes the null test on dvb. dvb is initialized at the
beginning of the function to &port->dvb. Since port is dereferenced
prior to the null test, port must be a valid pointer, and
&port->dvb cannot be null.
The following Coccinelle script is used for detecting the change:
@r@
expression e,f;
identifier g,y;
statement S1,S2;
@@
*e = &f->g
<+...
f->y
...+>
*if (e != NULL || ...)
S1 else S2
Signed-off-by: Himangi Saraogi <[email protected]>
Acked-by: Julia Lawall <[email protected]>
Signed-off-by: Hans Verkuil <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
Fix the following warnings reported in radio-bcm2048.c:
WARNING: else is not generally useful after a break or return
374: FILE: radio-bcm2048.c:374:
+ return 0;
+ } else {
WARNING: else is not generally useful after a break or return
728: FILE: radio-bcm2048.c:728:
+ return BCM2048_DE_EMPHASIS_75us;
+ else
WARNING: unchecked sscanf return value
1974: FILE: radio-bcm2048.c:1974:
+ sscanf(buf, mask, &value); \
WARNING: Missing a blank line after declarations
2245: FILE: radio-bcm2048.c:2245:
+ unsigned char tmpbuf[3];
+ tmpbuf[i] = bdev->rds_info.radio_text[bdev->rd_index+i+2];
WARNING: Possible unnecessary 'out of memory' message
2601: FILE: radio-bcm2048.c:2601:
+ if (!bdev) {
+ dev_dbg(&client->dev, "Failed to alloc video device.\n");
The following error was left since it seems to be a false positive:
ERROR: Macros with complex values should be enclosed in parenthesis
+#define DEFINE_SYSFS_PROPERTY(prop, signal, size, mask, check) \
+property_write(prop, signal size, mask, check) \
+property_read(prop, size, mask)
Signed-off-by: Luke Hart <[email protected]>
Signed-off-by: Hans Verkuil <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The issue was discovered by static analysis. It turns out that code is
somewhat insane, being
if (x) {...} else { if (x) {...} }
Edited it to do the only reasonable thing, which is to log the
information about the failed call. The most descriptive logging commands
set is taken from original code.
Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=79801
Reported-by: David Binderman <[email protected]>
Signed-off-by: Andrey Utkin <[email protected]>
Acked-by: Mike Isely <[email protected]>
Signed-off-by: Hans Verkuil <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
Previously, it was needed to reopen device to update GOP size and
quantization parameter. Now we update device registers with new values
immediately.
Signed-off-by: Andrey Utkin <[email protected]>
Signed-off-by: Ismael Luceno <[email protected]>
Signed-off-by: Hans Verkuil <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
solo6*10 boards have configurable quantization parameter which takes
values from 0 to 31, inclusively.
This change enables setting it with ioctl VIDIOC_S_CTRL with id
V4L2_CID_MPEG_VIDEO_H264_MIN_QP.
Signed-off-by: Andrey Utkin <[email protected]>
Signed-off-by: Ismael Luceno <[email protected]>
Signed-off-by: Hans Verkuil <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The colorspace for HDTV is REC709, not SMPTE240M.
Signed-off-by: Hans Verkuil <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
Otherwise you cannot get the current clip and bitmap information from
an overlay.
Signed-off-by: Hans Verkuil <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
If the vidioc_cropcap op is implemented by the driver then the v4l2
core will call that directly.
If g_selection is available, then the core cropcap implementation
uses g_selection to fill in the bounds and defrect and it sets the
pixelaspect to 1x1.
But if both are available, then I would like to use g_selection to
fill in defrect and bounds before calling cropcap. That way the
driver's cropcap implementation doesn't have to set defrect or
bounds.
Signed-off-by: Hans Verkuil <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The edid field is the last field of the struct, so there is nothing to clear.
Signed-off-by: Hans Verkuil <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
Some applications poll a lot, so prevent the poll message from flooding
the log.
Signed-off-by: Hans Verkuil <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
Add a macro to test if the field consists of a single top
or bottom field. Anyone who needs to work with fields as opposed to
frame will need this.
Signed-off-by: Hans Verkuil <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The VIDIOC_STREAMON/OFF ioctls are not valid for radio devices, just
like the other streaming I/O ioctls. Add the streamon/off ioctls
to the other streaming I/O ioctls.
Signed-off-by: Hans Verkuil <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
If the driver supports enum_freq_bands, but only for certain device
nodes, then it may return -ENOTTY. But in that case the code should
fall into the fall-back case where the current tuner/modulator range
is returned.
Signed-off-by: Hans Verkuil <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The other format ioctls (g/s/try_fmt) all check if the passed buffer type
makes sense for the device node's vfl_type. E.g. it makes no sense for a
VBI buffer type to be passed through a video node instead of a vbi node.
But this check was missing in ENUM_FMT which can cause a problem if you
have both video and sdr device nodes.
Signed-off-by: Hans Verkuil <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The text referred to videodev.h when videodev2.h was meant. Fixed.
Signed-off-by: Hans Verkuil <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The v4l2_event union has a new src_change field, but that was never
added to the VIDIOC_DQEVENT documentation of that union. Fixed.
Signed-off-by: Hans Verkuil <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
There shouldn't be any spaces after <constant> or before </constant>.
This leads to ugly results like: 'image size set by VIDIOC_S_FMT .'
Signed-off-by: Hans Verkuil <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
Update the OMAP Image Signal Processor entry to cover both the OMAP3 ISP
and OMAP4 ISS.
Signed-off-by: Laurent Pinchart <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
When a fatal error occurs in the pipeline signal it to the vb2 queue
with a call to vb2_queue_error(). The queue will then take care to
return -EIO when preparing buffers, remove the driver-specific code that
now duplicates that check.
Signed-off-by: Laurent Pinchart <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The parameter is used to initialize the video node debug field and
activate the V4L debug infrastructure.
Signed-off-by: Laurent Pinchart <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
Initialize the spin lock once only when initializing the video object.
Signed-off-by: Laurent Pinchart <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
