The dvb-pll code should do the same thing that
alps_tdee4_stv0297_tuner_set_params() was doing. Except the dvb-pll code
will check for tuner presence when attaching, while the old code didn't.
This tuner appears to not be behind the stv0297's I2C gate but is instead
on a different I2C adapter provided by the flexcop chip. The old code
would turn the I2C gate off each time the tuner was used. I've changed it
to turn the gate off when the tuner is attached and then disable the gate
control function. This should result in the gate staying off, which should
be even better.
[[email protected]: fix compile error]
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Hans Verkuil <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The code in skystar23_samsung_tbdu18132_tuner_set_params() and
samsung_tbmu24112_tuner_set_params() is equivalent to what the dvb-pll code
does. There could be an issue because the dvb-pll code will probe to check
for the tuner, while the previous code didn't do any checks.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The code in samsung_tdtc9251dh0_calc_regs() is equivalent to what dvb-pll's
code does.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
No datasheet, data take from code in flexcop driver. That code rounded
down the divisor rather than rounding to nearest, which was probably not
intentional and the dvb-pll code will round to nearest.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
Tuner parameters determined from code in flexcop driver. That code rounded
the divisor down instead of to the nearest value. This was probably not
intentional and the dvb-pll version will round to nearest.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
Tuner parameters taken from flexcop driver. This PLL has a 17 bit divisor
while the dvb-pll driver is designed for 15 bit divisors. It's not a
problem as 15 bits is enough for the tuner's entire range. But if a larger
range was wanted, it could be done by adding additional bands with the
extra divisor bits appearing as band switch bits.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
No datasheet, tuner data comes from code in flexcop driver. This tuner is
also used on the AVerTV 771 supported by the bttv driver, but that code
uses a different tuner configuration, which is surprising.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
A recent patch didn't take into account that frontends can be compiled into
the kernel. Or that frontends compiled as modules can't be used by the
b2c2 driver if it is not a module itself.
Some frontends require multiple drivers, e.g. a demod driver and a tuner
driver. The code for the frontend support was getting added if the demod
driver was available. Change this to also require any needed tuner or SEC
drivers as well.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
There already is a "default-on" trigger but there are problems with it.
For one, it's a inefficient way to do it and requires led trigger support
to be compiled in.
But the real reason is that is produces a glitch on the LED. The GPIO is
allocate with the LED *off*, then *later* when the trigger runs it is
turned back on. If the LED was already on via the GPIO's reset default or
action of the firmware, this produces a glitch where the LED goes from on
to off to on. While normally this is fast enough that it wouldn't be
noticeable to a human observer, there are still serious problems.
One is that there may be something else on the GPIO line, like a hardware
alarm or watchdog, that is fast enough to notice the glitch.
Another is that the kernel may panic before the LED is turned back on, thus
hanging with the LED in the wrong state. This is not just speculation, but
actually happened to me with an embedded system that has an LED which
should turn off when the kernel finishes booting, which was left in the
incorrect state due to a bug in the OF LED binding code.
We also let GPIO LEDs get their initial value from whatever the current
state of the GPIO line is. On some systems the LEDs are put into some
state by the firmware or hardware before Linux boots, and it is desired to
have them keep this state which is otherwise unknown to Linux.
This requires that the underlying GPIO driver support reading the value of
output GPIOs. Some drivers support this and some do not.
The platform device binding gains a field in the platform data
"default_state" that controls this. There are three constants defined to
select from on, off, or keeping the current state. The OpenFirmware
binding uses a property named "default-state" that can be set to "on",
"off", or "keep". The default if the property isn't present is off.
Signed-off-by: Trent Piepho <[email protected]>
Acked-by: Grant Likely <[email protected]>
Acked-by: Wolfram Sang <[email protected]>
Acked-by: Sean MacLennan <[email protected]>
Signed-off-by: Richard Purdie <[email protected]>
On Fri, 12 Jun 2009, Randy Dunlap wrote:
> From: Randy Dunlap <[email protected]>
>
> Move v4l_bound_align_image() outside of an #ifdef CONFIG_I2C block
> so that it is always built. Fixes a build error:
clamp_align() should be moved as well, since it's only used by
v4l_bound_align_image(). I'm attaching an alternate version that fixes
this. Labeled the endif too.
Reported-by: Randy Dunlap <[email protected]>
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
In the TRY_FMT handler the function get_scale() is called to find what the
scaler hardware will produce for a requested size.
The problem is that get_scale(struct cx231xx *dev, ..., unsigned int *vscale,
unsigned int *hscale) saves the calculated scale values into both the
pointer arguments and into dev's hscale and vscale fields. TRY_FMT shouldn't
actually change anything in the device state.
The code to in get_scale() that writes to dev->[hv]scale can just be
deleted. In all cases when dev's fields should be modified, get_scale()
was called with get_scale(dev, ..., &dev->hscale, &dev->vscale), so dev was
getting updated anyway.
This didn't actually cause a problem because nothing ever actually made use
of the hscale and vscale fields. I changed cx231xx_resolution_set() to use
those fields rather than re-calculate them with a call to get_scale().
Updating [hv]scale in cx231xx_resolution_set() isn't necessary because
every call of cx231xx_resolution_set() was already preceded by a call to
get_scale() or setting the [hv]scale fields, so they will be always be
up-to-date w.r.t. width and height.
Removing the call to get_scale() from cx231xx_resolution_set() allowed
making get_scale() a static function, which is a good thing for something
with such a short name. There is already another function with the same
name in the em28xx driver, but that one is static.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The v4l function has a better algorithm for aligning image size.
It appears that the em2800 can only scale by 50% or 100%, i.e. the only
heights supported might be 240 and 480. In that case the old code would
set any height other than 240 to 480. Request 240 get 240, but request 239
and then you get 480. Change it to round to the nearest supported value.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The v4l function has a better algorithm for aligning image size.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The v4l function has a better algorithm for aligning image size.
The existing code was casting pointers to u32 and to unsigned int into
pointers to u16. This could mess up if someone passed in an image size
greater than 65,535 and on big-endian platforms it won't work at all.
The existing bounding code would shrink an image if it was too big, but
returned ERANGE if it was too small. The code will not shrink or expand as
necessary.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The v4l function has a better algorithm for aligning image size.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The v4l function has a better algorithm for aligning image size.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The v4l function has a better algorithm for aligning image size.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The v4l function has a better algorithm for aligning image size.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The v4l function has a better algorithm for aligning image size.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The v4l function has a better algorithm for aligning image size.
For instance the old code would change 159x243 into 156x240 to meet the
alignment requirements. The new function will use 160x243, which is a lot
closer to what was asked for originally.
Cc: Robert Jarzmik <[email protected]>
Cc: Guennadi Liakhovetski <[email protected]>
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
Most hardware has limits on minimum and maximum image dimensions and also
requirements about alignment. For example, image width must be even or a
multiple of four. Some hardware has requirements that the total image size
(width * height) be a multiple of some power of two.
v4l_bound_align_image() will enforce min and max width and height, power of
two alignment on width and height, and power of two alignment on total
image size.
It uses an efficient algorithm that will try to find the "closest" image
size that meets the requirements.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The frontend attachment code didn't handle cases where the frontend
partially failed to attach. For instance, when the demod was attached
successfully but the tuner driver wasn't compiled or fails to init for some
reason. In these cases we try to clean up the partial attachment and fail
instead of proceeding with a broken frontend.
If frontend registration fails, clean up with dvb_frontend_detach() rather
than just calling the frontend's main release method. The former does some
additional stuff, like release an attached tuner and take care of putting
symbols when dynamic binding is used.
In skystar2_rev23_attach() it's not necessary to set fc->dev_type, that
gets set before skystar2_rev23_attach() is called.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Patrick Boettcher <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
Some ioctls have structs that are a different size depending on what type
of buffer is being used. If the buffer type leaves a field unused or has
padding space at the end, this space should be zeroed out.
The problems with S_FMT and REQBUFS were original identified and patched by
Marton Nemeth <[email protected]>.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
For a number of different ioctls, the v4l2-ioctl code checks that the
passed buffer type is supported by the driver. It did this by checking
that the driver defined a method for the try_fmt handler for that buffer
type. However, try_fmt is optional and a driver might not provide it even
though it does support that type. So use g_fmt instead, since that isn't
optional.
This should fix a problem with VBI capture with saa7146.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
If someone requests a format at fmt->index == (unsigned)-1 and the first
format in the array doesn't have the requested type then num will still be
-1 when it's compared to fmt->index and there will appear to be a match.
Restructure the loop so this can't happen. It's simpler this way too. The
unnecessary check for (unsigned)fmt->index < 0 found by Roel Kluin
<[email protected]> is removed this way too.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The ->vidioc_(s|g|try|enum)_fmt_vid_cap() methods are only called on
VIDEO_CAPTURE buffers. Thus, there is no need to check or set the buffer's
'type' field since it must already be set to VIDEO_CAPTURE. Checking the
buffer type can be removed from zr364xx_vidioc_(s|g|try|enum)_fmt_vid_cap().
The v4l2 core code in v4l2_ioctl will zero out the structure the driver is
supposed to fill in for read-only ioctls. For read/write ioctls, all the
fields which aren't supplied from userspace will be zeroed out.
Zeroing code can be removed from zr364xx_vidioc_querycap(),
zr364xx_vidioc_enum_input(), zr364xx_vidioc_enum_fmt_vid_cap(), and
zr364xx_vidioc_g_fmt_vid_cap().
Cc: Antoine Jacquet <[email protected]>
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The v4l2-ioctl core only allows buffer types for which the corresponding
->vidioc_try_fmt_xxx() methods are defined to be used with
vidioc_(q|dq|query)buf() and vidioc_reqbufs().
Since this driver only defines ->vidioc_try_fmt_vid_cap() the checks can be
removed from vidioc_reqbufs(), vidioc_qbuf(), and vidioc_dqbuf().
The ->vidioc_(s|g|try|enum)_fmt_vid_cap() methods are only called on
VIDEO_CAPTURE buffers. Thus, there is no need to check or set the buffer's
'type' field since it must already be set to VIDEO_CAPTURE. So setting the
buffer type in vidioc_enum_fmt_vid_cap() can be removed.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The ->vidioc_(s|g|try|enum)_fmt_vid_cap() methods are only called on
VIDEO_CAPTURE buffers. Thus, there is no need to check or set the buffer's
'type' field since it must already be set to VIDEO_CAPTURE.
The v4l2-ioctl core only allows buffer types for which the corresponding
->vidioc_try_fmt_xxx() methods are defined to be used with
vidioc_(q|dq|query)bufs() and vidioc_reqbufs().
Since this driver only defines ->vidioc_try_fmt_vid_cap() the checks can be
removed from vidioc_reqbufs(), vidioc_querybuf(), vidioc_qbuf(), and
vidioc_dqbuf(). Also, the buffer index is unsigned so it's not necessary
to check if it is less than zero.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The ->vidioc_(s|g|try|enum)_fmt_vid_cap() methods are only called on
VIDEO_CAPTURE buffers. Thus, there is no need to check or set the 'type'
field since it must already be set to VIDEO_CAPTURE.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The v4l2 core code in v4l2_ioctl will zero out the structure the driver is
supposed to fill in for read-only ioctls. For read/write ioctls, all the
fields which aren't supplied from userspace will be zeroed out.
Zeroing code is removed from enum_input, g_tuner, g_frequency, querycap,
enum_fmt_vid_cap, g_fmt_vid_cap, and try_fmt_vid_cap.
The ->vidioc_(s|g|try|enum)_fmt_vid_cap() methods are only called on
VIDEO_CAPTURE buffers. Thus, there is no need to check or set the buffer's
'type' field since it must already be set to VIDEO_CAPTURE.
There also appeared to be a copy and paste error in
vidioc_try_fmt_vid_cap() that would set f->fmt.pix.sizeimage to zero.
Note that the s_fmt_vid_cap method doesn't appear to actually do anything.
Whatever parameters were requested are just silently ignored. Was this
intentional? Who knows, as the commit log entry for the driver just says,
"Add generic cx23417 hardware encoder support." There are no docs. A
comment like "this driver totally ignores the v4l2 spec w.r.t.
VIDIOC_S_FMT because ..." would have gone a long way.
Cc: Steven Toth <[email protected]>
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The v4l2-ioctl core only allows buffer types for which the corresponding
->vidioc_try_fmt_xxx() methods are defined to be used with
vidioc_(q|dq|query)bufs() and vidioc_reqbufs().
This driver only defines ->vidioc_try_fmt_vid_cap() so only VIDEO_CAPTURE
buffers are allowed to be used with cafe_vidioc_reqbufs(),
cafe_vidioc_querybuf(), cafe_vidioc_qbuf(), and cafe_vidioc_dqbuf().
The ->vidioc_(s|g|try|enum)_fmt_vid_cap() methods are only called on
VIDEO_CAPTURE buffers. Thus, there is no need to check or set the buffer's
'type' field since it must already be set to VIDEO_CAPTURE. So the
check in cafe_vidioc_enum_fmt_vid_cap() can be removed.
The 'index' field of v4l2_buffer is unsigned so the checks for it being
less than zero can be removed too.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The v4l2-ioctl core only allows buffer types for which the corresponding
->vidioc_try_fmt_xxx() methods are defined to be used in vidioc_(g|s)_parm,
vidioc_(q|dq|query)buf, and vidioc_reqbufs.
Remove buffer type checking from vino_g_parm(), vino_s_parm(),
vino_reqbufs(), vino_querybuf(), vino_qbuf(), and vino_dqbuf(). This
reduced the indent level of the code so a few lines can be wrapped better.
Also fixed the C++ type comments.
The v4l2-ioctl core also provides structs that have been pre-zeroed for all
fields that driver is supposed to fill in, so remove zeroing code from
vino_enum_fmt_vid_cap(). Also, the format index is unsigned so it's not
necessary to check if it's less than zero.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The v4l2-ioctl core only allows buffer types for which the corresponding
->vidioc_try_fmt_xxx() methods are defined to be used with
vidioc_(q|dq|query)bufs(), vidioc_reqbufs() and now vidioc_(s|g)_parm.
This driver only defines ->vidioc_try_fmt_vid_cap() so only VIDEO_CAPTURE
buffers are allowed to be used with vidioc_g_parm(), vidioc_qbuf(),
vidioc_dqbuf(), and vidioc_reqbufs().
Cc: Jaime Velasco Juan <[email protected]>
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The v4l2-ioctl core now only allows buffer types for which the corresponding
->vidioc_try_fmt_xxx() methods are defined to be used in vidioc_(g|s)_parm.
This driver only defines ->vidioc_try_fmt_vid_cap() so only VIDEO_CAPTURE
buffers are allowed to be used with vidioc_s_parm() and vidioc_g_parm().
Cc: Sakari Ailus <[email protected]>
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The v4l2-ioctl core checks the buffer type now by only allowing buffer
types for which the corresponding ->vidioc_try_fmt_xxx() methods are
defined.
This driver only defines ->vidioc_try_fmt_vid_cap() so only VIDEO_CAPTURE
buffers are allowed to be used with vidioc_g_parm. Also,
->vidioc_enum_fmt_vid_cap() is only called for VIDEO_CAPTURE buffers.
There is no need to set the buffer type since it must already be the
correct value.
The struct which ->vidioc_querycap() is supposed to fill in is already
zeroed so it's not necessary to call memset on it.
Cc: Jean-Francois Moine <[email protected]>
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The v4l2-ioctl core only allows buffer types for which the corresponding
->vidioc_try_fmt_xxx() methods are defined to be used with
vidioc_(q|dq|query)bufs(), vidioc_reqbufs() and now vidioc_(s|g)_parm.
The driver was only allowing VIDEO_CAPTURE buffers for g_parm, but since
the driver defines ->vidioc_try_fmt_vid_overlay() and
->vidioc_try_fmt_vbi_cap() it will now allow VIDEO_OVERLAY and VBI_CAPTURE
buffers as well. This should be fine as the driver only fills in the frame
rate field, which is just as valid for video overlay and vbi capture as it
is for video capture.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The v4l2-ioctl core now only allows buffer types for which the
corresponding ->vidioc_try_fmt_xxx() methods are defined to be used with
vidioc_(g|s)_parm.
The driver was only allowing VIDEO_CAPTURE buffers for g_parm, but since
the driver defines ->vidioc_try_fmt_vid_overlay() it will now allow
VIDEO_OVERLAY buffers as well. This should be fine as the fields the
driver fills in, readbuffers and frame rate, aren't wrong for VIDEO_OVERLAY
buffers.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
Return EINVAL if VIDIOC_S/G_PARM is called for a buffer type that the
driver doesn't define a ->vidioc_try_fmt_XXX() method for. Several other
ioctls, like QUERYBUF, QBUF, and DQBUF, etc. do this too. It saves each
driver from having to check if the buffer type is one that it supports.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
It reduces the size of the driver over all, and the function names in
strings need to be manually kept up to date while __func__ doesn't.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
Many functions had a struct file pointer argument, but all they wants is
the struct zoran_fh pointer from the file's private data. Since every
caller of those functions already has the zoran_fh, just pass the that
instead. This saves a dereference in each function change.
While I'm at it, change the code formatting of affected functions to be
kernel standard style.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The zoran driver had two kinds of buffer descriptors, one for jpg buffers
and one for raw buffers. They were mostly the same with only a couple
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
It was a struct file *, but all that function wants is the struct zoran_fh
from the file's private data. Since every caller already has this, just
pass the zoran_fh instead.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The v4l2 core code in v4l2_ioctl will zero out the structure the driver is
supposed to fill in for read-only ioctls. For read/write ioctls, all the
fields which aren't supplied from userspace will be zeroed out.
Zeroing code is removed from querycap, enum_input, enum_fmt_vid_cap, and
g_fmt_vid_cap.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The v4l2 core code in v4l2_ioctl will zero out the structure the driver is
supposed to fill in for read-only ioctls. For read/write ioctls, all the
fields which aren't supplied from userspace will be zeroed out.
Zeroing code is removed from g_audio, enum_input, g_parm and gmbuf.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The v4l2 core code in v4l2_ioctl will zero out the structure the driver is
supposed to fill in for read-only ioctls. For read/write ioctls, all the
fields which aren't supplied from userspace will be zeroed out.
Zeroing code is removed from g_audio, queryctrl, and enum_fmt_vid_cap.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The v4l2 core code in v4l2_ioctl will zero out the structure the driver is
supposed to fill in for read-only ioctls. For read/write ioctls, all the
fields which aren't supplied from userspace will be zeroed out.
Zeroing code is removed from enum_fmt_vid_cap, g_fmt_vid_cap, and g_parm.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The v4l2 core code in v4l2_ioctl will zero out the structure the driver is
supposed to fill in for read-only ioctls. For read/write ioctls, all the
fields which aren't supplied from userspace will be zeroed out.
Zeroing code is removed from enum_input and g_tuner.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The v4l2 core code in v4l2_ioctl will zero out the structure the driver is
supposed to fill in for read-only ioctls. For read/write ioctls, all the
fields which aren't supplied from userspace will be zeroed out.
Zeroing code is removed from enum_input and g_tuner.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
vidioc_enum_fmt_vid_overlay() did nothing but call
vidioc_enum_fmt_vid_cap(), so just make
saa7146_video_ioctl_ops.vidioc_enum_fmt_vid_overlay point to
vidioc_enum_fmt_vid_cap() and get ride of vidioc_enum_fmt_vid_overlay().
Have gparm use v4l2_video_std_frame_period to fill in the frame period
instead of just assuming PAL.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
Some code was calling v4l2_video_std_construct() when all it cared about
was the frame period. So make a function that just returns that and have
v4l2_video_std_construct() use it.
At this point there are no users of v4l2_video_std_construct() left outside
of v4l2-ioctl, so it could be un-exported and made static.
Change v4l2_video_std_construct() so that it doesn't zero out the struct
v4l2_standard passed in. It's already been zeroed out in the common ioctl
code.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
If an ioctl is read-only then the driver fills in all the fields. Lots of
times drivers only care about some fields so it's best if video_ioctl2
takes care of zeroing out the entire structure before handing it to the
driver. This saves code in each driver to do it and driver authors often
forget.
The existing memset code in some of the read-only ioctl handlers
can be deleted.
Convert a case statement to a single if statement.
Deleted a debug line from ENUMAUDOUT that was copy-and-pasted to G_AUDOUT
by mistake.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
There are many RW ioctls() in v4l2 where userspace only supplies one or two
of the first fields in the structure passed to the ioctl. The driver then
fills in the rest of the fields.
Instead of copying the entire structure from userspace to the kernel we
only need to copy those fields that userspace is actually supposed to
supply.
What's more, the fields that are meant to be only be output from the driver
can be zeroed out in the videodev code, in case the driver doesn't fill
them all in. Many of the ioctl handlers in v4l2_ioctl do this already, but
my patch does this at one common point and so all the memsets for each
ioctl can be deleted.
For VIDIOC_G_SLICED_VBI_CAP, which has one input field ('type') and other
output-only fields, the input field is near the end of the structure
instead of at the beginning. So there is still a memset in it's ioctl
handler to zero out the beginning of the struct.
There were a couple mistakes with the existing code:
For VIDIOC_G_AUDIO the index field was preserved, but G_AUDIO is a read
only ioctl so nothing is copied from userspace to preserve.
For VIDIOC_G_FREQUENCY the tuner field was not preserved like it should
have been. This would be a problem if there was any hardware with more
than one tuner/modulator.
For VIDIOC_ENUM_FRAMESIZES and VIDIOC_ENUM_FRAMEINTERVALS, none of the
fields were preserved even though each ioctl has several field that are
supposed to be inputs to the driver! Obviously these ioctls don't get
used much. The index field is needed if the driver has multiple
discrete sizes/rates and other fields can be used too, e.g. if the size
depends on pixel format or frame rate depends on image size for
example.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
For VIDIOC_QUERYBUF only the first two fields, size and type, are used as
input. The rest can be filled in by the driver as output. Most drivers do
not actually use all the field and unused ones should be zeroed out. Some
drivers have code to do this and some drivers should but don't. So put
some zero out code in v4l2_ioctl so that all drivers using that system get
it.
The drivers that have zeroing code get that code removed.
Some drivers checked that the type field was valid, but v4l2_ioctl already
does this so those checks can be removed as well.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
A recent patch added a forward declaration of cx88_core right before the
main definition of that structure, which is obviously unneeded.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
video_register_device_index() checks if it was passed a NULL video_device
pointer (which isn't allowed) _after_ it has already dereferenced it with
video_get_drvdata(vdev). The checks are clearly pointless and can be
removed, as the function would have crashed before reaching them if vdev
ever was NULL.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The bttv driver had static array of structures for up to 16 possible bttv
devices, even though few users have more than one or two. The structures
were quite large and this resulted in a huge BSS segment.
Change the driver to allocate the bttv device data dynamically, which
changes "struct bttv bttvs[BTTV_MAX]" to "struct bttv *bttvs[BTTV_MAX]".
It would be nice to get ride of "bttvs" entirely but there are some
complications with gpio access from the audio & mpeg drivers.
To help bttvs removal along anyway, I changed the open() methods use the
video device's drvdata to get the driver data instead of looking it up in
the bttvs array. This is also more efficient. Some WARN_ON()s are added
in cases the device node exists by the bttv device doesn't, which I don't
think should be possible.
The gpio access functions need to check if bttvs[card] is NULL now. Though
calling them on a non-existent card in the first place is wrong, but hard
to solve given the fundamental problems in how the gpio access code works.
This patch reduces the bss size by 66560 bytes on ia32. Overall change is a
reduction of 66398 bytes, as the WARN_ON()s add some 198 bytes.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
Over half of the card database was used to store muxsel data. 64 bytes
were used to store one 32 bit word for each of up to 16 inputs.
The Bt8x8 only has two bits to control its mux, so muxsel data for 16
inputs will fit into a single 32 bit word. There were a couple cards that
had special muxsel data that didn't fit in two bits, but I cleaned them up
in earlier patches.
Unfortunately, C doesn't allow us to have an array of bit fields. This
makes initializing the structure more of a pain. But with some cpp magic,
we can do it by changing:
.muxsel = { 2, 3, 0, 1 },
.muxsel = { 2, 2, 2, 2, 3, 3, 3, 3, 1, 1 },
Into:
.muxsel = MUXSEL(2, 3, 0, 1),
.muxsel = MUXSEL(2, 2, 2, 2, 3, 3, 3, 3, 1, 1),
That's not so bad. MUXSEL is a fancy macro that packs the arguments (of
which there can be one to sixteen!) into a single word two bits at a time.
It's a compile time constant (a variadic function wouldn't be) so we can
use it to initialize the structure. It's important the the arguments to
the macro only be plain decimal integers. Stuff like "0x01", "(2)", or
"MUX3" won't work properly.
I also created an accessor function, bttv_muxsel(btv, input), that gets the
mux bits for the selected input. It makes it cleaner to change the way the
muxsel data is stored.
This patch doesn't change the code size and decreases the datasegment by
9440 bytes.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
This card apparently uses an external mux and the Bt878's mux should always
be set to MUX2. The values for the external mux control bits were stored
in the muxsel field. This meant that when changing inputs the driver would
switch the Bt878's mux to whatever value the external mux was supposed to
be set to, then eagle_muxsel() would switch it back to MUX2 and program the
external mux. This creates an unnecessary switch of the Bt878's mux.
So change muxsel to be 2 for each input. The external mux bits are just
"input&3" so they don't really need to be stored anywhere. This also
eliminates the last non-standard use of the muxsel data.
Cc: M G Berberich <[email protected]>
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
Old versions of the bttv driver would use the high nibble of an input's
muxsel value to program the GPIO lines enabled via gpiomask2. Apparently
this was supposed to be for switching external audio muxes. Anyway, the
code that did this was removed sometime in the pre-git 2.6 series.
The RemoteVision MX board used this feature to control an external video
mux and I guess no one noticed when they removed the code.
Move the extra gpio mux data out of the high nibble of muxsel and to
rv605_muxsel(), then have that function set the gpio lines with it.
From looking at the CD22M3494E datasheet, it seems like the mdelay(1) is a
much longer delay than necessary. It looks like only around 20 ns is
necessary.
Cc: Miguel Freitas <[email protected]>
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
Old versions of the bttv driver would use the high nibble of an input's
muxsel value to program the GPIO lines enabled via gpiomask2. Apparently
this was supposed to be for switching external audio muxes. Anyway, the
code that did this was removed sometime in the pre-git 2.6 series.
These phytec boards used this feature to control an external video mux and
I guess no one noticed when they removed the code.
So add a muxsel_hook for these boards that does the necessary gpio setting.
BTW, I doubt the needs_tvaudio setting for these cards is correct.
Cc: Dirk Heer <[email protected]>
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The card data for BTTV_BOARD_IVC120 set muxsel to a bunch of bogus values
(1 to 16), which the common mux code would use to set the Bt878's mux to
some random value. Then the custom code in ivc120_muxsel() would change
the Bt878's mux to the right value (always MUX0).
Better to just make the muxsel data correct (all zeros, easy!) and get the
mux right to begin with. Then the extra Bt878 mux setting code in
ivc120_muxsel() can be eliminated (the rest of the code for the IVC-120G's
external mux is still there of course).
This will help me clean up muxsel for some other changes. It should also
get rid of an unnecessary mux switch when changing from certain inputs to
certain other inputs on the IVC-120G.
Cc: Alan Garfield <[email protected]>
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The code was using a muxsel value of -1U to indicate a digital input. A
couple places in were checking of muxsel < 0 to detect this, which doesn't
work of course because muxsel is unsigned and can't be negative.
Only a couple cards had digital inputs and it was always the last one, so
for the card database create a one bit field that indicates the last input
is digital. On init, this is used to set a new field in the bttv struct to
the digital input's number or UNSET for none. This makes it easier to
check if the current input is digital.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The bttv card database is quite large and the data structure used to store
it wasn't very efficient. Most of the field are only used at card
initialization time so it doesn't matter if they aren't efficient to
access.
Overall the changes reduce code size by 60 bytes in ia32. The data size is
decreased by 5024 byes. It is probably even more for 64-bit kernels.
Move the fields in the struct around to be sorted from largest to smallest.
This saves on padding space used for alignment.
Get rid of the unused digital_mode field. Leave the setting as a comment
in the few cards entries that set it, in case someone ever writes the code.
Get rid of the unused audio_inputs field. Leave the values in the card
entries in case someone ever writes code that might use it.
Get ride of the unused radio_addr field. No card entries even set it to
anything interesting so it's not left as comments. All the code that used
it was removed in commit v2.6.14-3466-g291d1d7 from Nov 8th 2005.
Reduce video_inputs to u8 as no card has more than 255 inputs (the most is
16).
Change tuner_addr to u8. I2C addresses are only seven bits and 255 means
ADDR_UNSET, so everything fits.
Make has_radio a one bit flag.
Make the pll setting a two bit field.
Reduce svhs to four bits as no card has an s-video input above 9. Change
the value for no s-video input from UNSET (which is -1U and out of range of
four bits) to NO_SVHS (which is now 15).
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The bttv card database structure had a "tuner" field that was the input
number of the tuner input or UNSET for no tuner. However, the only values
it could ever be are 0 and UNSET. Having a tuner on an input other than 0
didn't work and was never used.
There is also a "tuner_type" field that can be set to TUNER_ABSENT to
indicate no tuner, which makes "tuner = UNSET" redundant. In many cases,
tuner_type was set to UNSET when there was no tuner, which isn't quite
correct. tuner_type == UNSET is supposed to mean the tuner type isn't yet
known.
So, I changed cards where "tuner == UNSET" to always have tuner_type of
TUNER_ABSENT. At this point the tuner field is redundant, so I deleted it.
I have the card setup code set the card's tuner_type (not the card type's
tuner_type!) to TUNER_ABSENT if it hasn't yet been set at the end of the
setup code. Various places that check if the card has a tuner will now
look for this instead of checking the card type's "tuner" field.
Also autoload the tuner module before issuing the TUNER_SET_TYPE_ADDR I2C
client call instead of after issuing it.
Overall, on ia32 this decreases compiled code size by about 24 bytes and
reduces the data size by 640 bytes.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The code to set the norm for the TDA9880 analog demod was comparing
btv->norm, an index into the bttv driver's norm array, to V4L2_STD_NTSC,
which is a bit flag that's part of the V4L2 API. This doesn't work of
course and results in the PAL path always being taken.
What's more, it modified the bttv_tvcards[] entries for cards using the
TDA9880. This is wrong because changing the norm on one card will also
affect other cards of the same type. Writing to bttv_tvcards is also bad
because it should be read-only or even devinitdata.
Changing the norm would also cause the audio to become unmuted.
Have the code get called for both norm setting and audio input setting
(which where the gpios are set) to avoid needed to modify bttv_tvcards.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The norm value in the driver is an index into an array and the the driver
doesn't allow it to be negative or otherwise invalid. It should be
unsigned but wasn't in all places.
Fix some structs and functions to have the norm be unsigned. Get rid of
useless checks for "< 0". Most of the driver code can't handle a norm
value that's out of range, so change some ">= BTTV_TVNORMS" checks to
BUG_ON(). There's no point in silently ignoring invalid driver state just
to crash because of it later.
Reported-by: Roel Kluin <[email protected]>
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
A complete re-implementation of fakephp is necessary if it is to
present its former interface (pre-2.6.27, when it broke). The
reason is that PCI hotplug drivers call pci_hp_register(), which
enforces the rule that only one /sys/bus/pci/slots/ file may be
created per physical slot.
The change breaks the old fakephp's assumption that it could
create a file per function. So we re-implement fakephp to avoid
using the standard PCI hotplug API so that we can restore the old
fakephp user interface.
It puts entries in /sys/bus/pci/slots with the names of all PCI
devices/functions, exactly symmetrical to what is shown in
/sys/bus/pci/devices. Each slots/ entry has a "power" attribute,
which works the same way as the fakephp driver's power attribute
has worked.
There are a few improvements over old fakephp, which couldn't handle
PCI devices being added or removed via a means outside of
fakephp's knowledge. If a device was added another way, old fakephp
didn't notice and didn't create the fake slot for it. If a
device was removed another way, old fakephp didn't delete the fake
slot for it (and accessing the stale slot caused an oops).
The new implementation overcomes these limitations. As a
consequence, removing a bridge with other devices behind it now
works as well, which is something else old fakephp couldn't do
previously.
This duplicates a tiny bit of the code in the PCI core that does
this same function. Re-using that code ends up being more
complex than duplicating it, and it makes code in the PCI core
more ugly just to support this legacy fakephp interface
compatibility layer.
Reviewed-by: James Cameron <[email protected]>
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Alex Chiang <[email protected]>
Signed-off-by: Jesse Barnes <[email protected]>
pci_scan_slot() has been rewritten to be less complex and will now
return the number of *new* devices found.
Existing callers need not worry because they already assume that
they can't call pci_scan_slot() on an already-scanned slot.
Thus, there is no semantic change for existing callers: returning
newly found devices (this patch) is exactly equal to returning all
found devices (before this patch).
This patch adds some more groundwork to allow us to rescan the
PCI bus during runtime to discover newly added devices.
Signed-off-by: Trent Piepho <[email protected]>
Reviewed-by: Alex Chiang <[email protected]>
Signed-off-by: Alex Chiang <[email protected]>
Signed-off-by: Jesse Barnes <[email protected]>
pci_scan_single_device is supposed to add newly discovered
devices to pci_bus->devices, but doesn't check to see if the
device has already been added. This can cause problems if we ever
want to use this interface to rescan the PCI bus.
If the device is already added, just return it.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Alex Chiang <[email protected]>
Signed-off-by: Jesse Barnes <[email protected]>
Ronald Bultje hasn't been maintaining the zoran driver for some time.
Re-direct people to the mailing lists and web pages.
MAINTAINERS | 6 +++---
Signed-off-by: Trent Piepho <[email protected]>
Acked-by: Jean Delvare <[email protected]>
Acked-by: Ronald S. Bultje <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The zoran driver does a module_get/put of THIS_MODULE on device open/close.
This isn't necessary as the kernel does this automatically.
Clean up the failure path of zoran_open() somewhat.
Make the dprintk()s on open/close a higher debug level and make the user
count printed take the current open/close into account.
Signed-off-by: Trent Piepho <[email protected]>
Acked-by: Jean Delvare <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The driver should only use the kernel mapped io address, zr36057_mem, and
not the PCI bus address, zr36057_adr. Since the latter is only printed out
once, there is no need to save it in the driver data structure.
There was some old code that looked like it was for the Alpha architecture
which would use the PCI bus address. It probably no longer applies to
modern kernels.
Signed-off-by: Trent Piepho <[email protected]>
Acked-by: Jean Delvare <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
Instead of using custom code, just let the device layer look it up for us
from the pci device table. This requires extending the pci device table to
list each known card, plus a catch-all entry for the cards that don't have
sub-system vendor/device data.
Improve some of the info and error messages too.
Signed-off-by: Trent Piepho <[email protected]>
Acked-by: Jean Delvare <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The driver was keeping a global array with an entry for each zoran device
probed. It was a leftover from when the driver didn't dynamically allocate
the driver data for each device.
There was only one use left, in the video device's ->open() method, looking
up the struct zoran for the opened device from the minor number. This can
be done better with video_get_drvdata().
Since zoran_num is now only used in the pci driver's ->probe() method, it
doesn't need to be an atomic_t and be static. There is a race if multiple
zoran cards could be probed at the same time, but currently the probe
method for a given driver is single threaded.
Signed-off-by: Trent Piepho <[email protected]>
Acked-by: Jean Delvare <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
When initializing a module parameter that is a per-card array, use
"{ [0 ... (BUZ_MAX-1)] = -1 }" instead of "{ -1, -1, -1, -1 }". This way
all of the entries will be correctly set to -1 if someone changes BUZ_MAX
to a value other than 4.
Adjust some of the parameter help text too.
Signed-off-by: Trent Piepho <[email protected]>
Acked-by: Jean Delvare <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
This is a really old and crufty driver that wasn't using the long
established pci driver framework.
Signed-off-by: Trent Piepho <[email protected]>
Acked-by: Jean Delvare <[email protected]>
[[email protected]: Cleaned up a few CodingStyle issues]
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
PCI devices have three settable boolean attributes, enable,
broken_parity_status, and msi_bus.
The store functions for these would silently interpret "0x01" as false,
"1llogical" as true, and "true" would be (silently!) ignored and do
nothing.
This is inconsistent with typical sysfs handling of settable attributes,
and just plain doesn't make much sense.
So, use strict_strtoul(), which was created for this purpose. The store
functions will treat a value of 0 as false, non-zero as true, and return
-EINVAL for a parse failure.
Additionally, is_enabled_store() and msi_bus_store() return -EPERM if
CAP_SYS_ADMIN is lacking, rather than silently doing nothing. This is more
typical behavior for sysfs attributes that need a capability.
And msi_bus_store() will only print the "forced subordinate bus ..."
warning if the MSI flag was actually forced to a different value.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Jesse Barnes <[email protected]>
The last gpio belonging to a chip is chip->base + chip->ngpios - 1. Some
places in the code, but not all, forgot the critical minus one.
Signed-off-by: Trent Piepho <[email protected]>
Acked-by: David Brownell <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
The ttpci Kconfig file has bugs that cause it to fail in certain Kconfig
situations. The basic problem is that it selects certain drivers, but does
not depend on the dependencies of those drivers.
See http://article.gmane.org/gmane.comp.video.video4linux/35072
Using the Kconfig file also has some annoyances. For instance one can't turn
off AV7110 support unless you go down several options and first turn off
budget-patch support. Normally user selectable drivers are not forced on like
this.
The "AV7110 cards with Budget Patch" option is disabled if "Budget cards"
isn't on. Normally a driver appears nested under a driver it depends on, but
since drivers that don't depend on "Budget cards" are between the two options,
the config programs can't display the tree correctly.
The Makefile has an issue too. Some modules, ttpci-eeprom and budget-core,
appear in the Makefile under several different config symbols. If more than
one of these symbols is on, they will get added the to list of objects
multiple times. The normal convention is to have a config symbol just the
common object(s) and have the users of the that object either depend on or
select that config symbol.
This patch fixes all these issues. ttpci-eepom is under a new config symbol,
and so is the budget-core module. The four different budget card types appear
as sub-drivers under a main "SAA7146 DVB cards" option. Turning on
budget-patch doesn't force AV7110. Drivers using SAA7146_VV have the
necessary VIDEO_DEV dependency, so that it isn't possible to select SAA7146_VV
without V4L being on.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
Someone wasn't using the v4l-dvb commit scripts and so didn't run the
automatic whitespace cleaner on their code.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
This em28xx-video.c uses functions from this header, but doesn't include it.
It depends on some v4l headers included two levels down including poll.h,
which includes mm.h.
These v4l headers might change, so it's best to include the headers needed
directly.
It also causes problems for the out of core build system's backward
compatibility with older kernels, which is the real reason I bothered to
create a patch for something that would otherwise be so minor that it would
hardly be worth the trouble.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The V4L1 spec says that the image size should be with with VIDIOCSWIN before
requesting buffers with VIDIOCGMBUF and capturing into them with
VIDIOCMCAPTURE.
But it seems that many apps don't do this. They set the size using the fields
in the VIDIOCMCAPTURE ioctl. The driver doesn't know what size to capture
until it actually starts to capture. In particular, it doesn't know what size
to capture until it has already mmap the captured buffers. Which is quite
stupid. Why V4L1 has size and format fields for VIDIOCMCAPTURE I have no idea.
Many drivers don't support this, including those using v4l1-compat.
The bttv does, which is probably the only reason such broken software is so
prevalent.
But, the driver doesn't adjust its idea of what size is being captured when it
is set this way. If you try to query the driver's current setting with
v4l2-ctl, it won't be correct.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
The bttv driver instists that the depth specified in the call to VIDIOCSPICT
match the pixel format specified in the same call.
vlc doesn't set the depth field, which makes the SPICT ioctl always fail.
The V4L1 standard is not clear on how most operation are supposed to work, and
this is no exception. The depth field would appear to be entirely redundant,
as the pixel format specifies a specific depth. It could be that this field
was only meant for output from the *G*PICT ioctl and should be ignored in
*S*PICT. This is in fact what the v4l1-compat wrapper does.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
If the load_all option was turned on all cards would be treated as unknown,
even those which are in the database. Of course, if the card is in the
database there is no reason to use the load_all option. It's there to force
loading when the card isn't in the database. But there are out of date wikis
that say to do this and some distros might turn this option on by default.
So, we keep the load_all option from turning known cards into unknown cards.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Takashi Iwai <[email protected]>
Signed-off-by: Jaroslav Kysela <[email protected]>
Someone added a new case without adding a break to the one before it.
Thanks to Margus <b-berski at mbox200 dot swipnet dot se> for spotting this.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
When a cx8802 sub-driver was unregistered, the struct cx8802_driver, which was
kmalloc()ed by cx8802_register_driver(), was deleted from the list of drivers,
but never freed.
Signed-off-by: Trent Piepho <[email protected]>
Reviewed-by: Michael Krufky <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
Less code and more efficient.
Got ride of a variable that counted the number of devices in
cx8802_unregister_driver() but was never used. Looked leftover from a
cut&paste.
Signed-off-by: Trent Piepho <[email protected]>
Reviewed-by: Michael Krufky <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
It was a struct cx8802_driver for no apparent reason. Nothing uses a
cx8802_driver in the cx8802_dev struct. The only field that was used was
devlist, a list_head.
The code in cx8802_remove() that removed any loaded sub-drivers was broken.
It would delete the current list entry, but didn't use list_for_each_safe. It
also called list_del() on the list _head_ inside the list_for_each loop? It
would crash if it was run, which I don't think can ever happen.
Since the cx8802 sub-drivers use the cx8802 driver, they have to be unloaded
first. So there isn't any way for a sub-driver to still be loaded when
cx8802_remove() is called... Except maybe with PCI hot-plug, if one removes
the PCI card while the drivers are loaded?
So I left some code in to handle that if it's actually possible. It will
remove the sub-drivers from the device cx8802_remove() was called on, and only
that device. If one has two DVB cards and unplugs one, there is no reason to
unload the DVB drivers for both cards. I have no way to test this, but it
can't be worse than what was there before.
cx8802_get_driver() is passed a cx8802_dev pointer and looks for the requested
driver on that device. It first loops over the cx8802 device list looking for
the device it was passed, which is pointless. It doesn't need to find the
device pointer in the list, as it already has the pointer.
The list_head in the cx8802_driver struct, which joins all the _drivers_
attached to a device, was named devlist. Changed that to drvlist, since the
devlist is used for a list of _devices_ in other cx8802 structs.
Signed-off-by: Trent Piepho <[email protected]>
Reviewed-by: Michael Krufky <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
card_priv was only used to store a pointer to the vp3054 state struct.
There's no need to use a void * since it doesn't have multiple types.
Make the field conditional on VP3045 support. It was already conditional on
DVB support, but it's only used if VP3045 support is on, so that makes for a
better option to check.
Signed-off-by: Trent Piepho <[email protected]>
Reviewed-by: Michael Krufky <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
Add some ifdefs around fields only used for blackbird support, similar to the
way the dvb fields are only included with dvb support.
Signed-off-by: Trent Piepho <[email protected]>
Reviewed-by: Michael Krufky <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
Lets mixer apps display a dB range for the volume control.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
Add two mute controls. One mutes everything, the other just mutes the analog
pass-through output.
Rename the existing volume control. The controls are now:
Playback Volume
Playback Switch
Capture Switch
These names might seem odd, but I believe they are more correct. The previous
"Capture Volume" control didn't actually effect the volume of the captured
audio. Instead it controls the volume of the analog pass-thought output. It
appears that pass-through controls like this are usually considered to be in
the playback direction, not capture. For example, "CAPTURE feedback Playback
Volume" is the name used for a control that appears to have the same effect in
the ca0106 driver. We only have one volume control, so we can omit the
"CAPTURE feedback" part.
If someone where to add PCM playback support to the driver, then this would be
the volume control.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Mauro Carvalho Chehab <[email protected]>
When part of build an external module tree, modpost first reads in the
kernel's and then the external tree's Module.symvers files. From these files
it establishes a symbol => module mapping. When it later reads in each module
built and processes the symbols it finds, it discovers the symbol=>module
mapping from Module.symvers and leaves it as it is.
The problem comes with a module has been re-named or a symbol has moved from
one module to another, since the Module.symvers file was generated. modpost
does not update the symbol=>module mapping when it finds the new location of
the symbol when scanning the newly built modules. This results in the module
containing incorrect dependency information and the new Module.symvers file
written by modpost will also contain the incorrect mappings, perpetuating the
problem to the next build, and so on.
When building the out of kernel development tree for kernel subsystem, like
v4l-dvb or ALSA, deleting the external Module.symvers file before building
(which the kernel build system doesn't do and shouldn't be necessary anyway),
won't fix the problem. modpost still reads the kernel's Module.symvers, and
since we a building a kernel subsystem, it will define the same symbols as the
external modules.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Sam Ravnborg <[email protected]>
The polling loop to check for ACI to go down was more convoluted than it
needed to be. New loop should be more efficient and it is a lot simpler. The
old loop checked for a timeout before checking for ACI down, which could
result in an erroneous timeout. It's only a failure if the timeout expires
_and_ ACI is still high. There is nothing wrong with the timeout expiring
while the task is sleeping if ACI went low.
A polling loop to check for the device to leaving INIT mode is removed. The
device must have already left init for the previous ACI loop to have finished.
Acked-by: Rene Herman <[email protected]>
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Takashi Iwai <[email protected]>
Signed-off-by: Jaroslav Kysela <[email protected]>
Sets a bit to power down the Bt87x's internal audio ADC when the ALSA device
isn't open, or when it is in 'digital mode' using an external ADC.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Takashi Iwai <[email protected]>
Signed-off-by: Jaroslav Kysela <[email protected]>
Different cards have different audio configurations, but the driver didn't
support this. The only setting it had was the digital rate.
This patch adds a board configuration list. Currently, configurable items are
the digital rate and the digital data format (for cards with an external ADC),
a flag for the absence of an external ADC, and a flag for no connection to the
Bt87x internal ADC.
This allows cards that don't use the internal ADC to omit the ALSA 'Bt87x
analog' device and related controls. Cards without an external ADC can omit
the 'Bt87x digital' device.
In order to support the CS5331A ADC used on the Osprey 440 and 2x0 cards, the
digital format needs to be different than the default.
Support could be added for defining:
The connections or lack of them to the Bt87x's internal ADC mux
Multiple sample rates for an external ADC (e.g. Osprey)
Control of an external mux for an external ADC (e.g. Osprey)
The card definitions for cards other than the Ospreys are kept equivalent to
their old values. This is likely inaccurate for most cards, as it is doubtful
that both an external and the internal ADC would be used. Lacking information
on those cards, the behavior is left unchanged.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Takashi Iwai <[email protected]>
Signed-off-by: Jaroslav Kysela <[email protected]>
The init sequence set a number of registers more than once to different
values. It's only necessary to set them once to their final values.
It also never actually updated the digital attenuation settings.
Signed-off-by: Trent Piepho <[email protected]>
Signed-off-by: Takashi Iwai <[email protected]>
Signed-off-by: Jaroslav Kysela <[email protected]>
Trent Piepho