Bjørn Mork [Sun, 14 Nov 2010 17:24:36 +0000 (14:24 -0300)]
[media] Mantis: use dvb_attach to avoid double dereferencing on module removal
Convert the driver to use the dvb_attach macro to avoid the hard dependency
on the frontend drivers. The hard dependecy will result in loading a
number of unused frontends, and unwanted automatic dereferencing.
This fixes a bug where unloading the mantis driver will derefence any
attached frontend twice, which will cause an oops if the same frontend is
used by another driver.
Signed-off-by: Bjørn Mork <bjorn@mork.no> Signed-off-by: Manu Abraham <manu@linuxtv.org> Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
Marko Ristola [Sun, 14 Nov 2010 17:09:04 +0000 (14:09 -0300)]
[media] Mantis: append tasklet maintenance for DVB stream delivery
After dvb-core has called mantis-fe->stop_feed(dvbdmxfeed)
the last time (count to zero), no data should ever be copied
with dvb_dmx_swfilter() by a tasklet: the target structure
might be in an unusable state. Caller of mantis_fe->stop_feed()
assumes that feeding is stopped after stop_feed() has been
called, ie. dvb_dmx_swfilter() isn't running, and won't be called.
There is a risk that dvb_dmx_swfilter() references freed resources
(memory or spinlocks or ???) causing instabilities. Thus
tasklet_disable(&mantis->tasklet) must be called inside of
mantis-fe->stop_feed(dvbdmxfeed) when necessary.
Signed-off-by: Marko Ristola <marko.ristola@kolumbus.fi> Signed-off-by: Manu Abraham <manu@linuxtv.org> Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
Wolfram Sang [Wed, 10 Nov 2010 12:28:19 +0000 (09:28 -0300)]
[media] i2c: Remove obsolete cleanup for clientdata
A few new i2c-drivers came into the kernel which clear the clientdata-pointer
on exit. This is obsolete meanwhile, so fix it and hope the word will spread.
Jesper Juhl [Mon, 8 Nov 2010 23:08:41 +0000 (20:08 -0300)]
[media] cx231xx-417: Remove unnecessary casts of void ptr returning alloc function return values
Hi,
The [vk][cmz]alloc(_node) family of functions return void pointers which
it's completely unnecessary/pointless to cast to other pointer types since
that happens implicitly.
Randy Dunlap [Wed, 24 Nov 2010 19:41:09 +0000 (16:41 -0300)]
[media] timblogiw: fix kconfig & build error
timblogiw uses dma() interfaces and it selects TIMB_DMA for that
support. However, drivers/dma/ is not built unless
CONFIG_DMA_ENGINE is enabled, so select/enable that symbol also.
drivers/built-in.o: In function `timblogiw_close':
timblogiw.c:(.text+0x4419fe): undefined reference to `dma_release_channel'
drivers/built-in.o: In function `buffer_release':
timblogiw.c:(.text+0x441a8d): undefined reference to `dma_sync_wait'
drivers/built-in.o: In function `timblogiw_open':
timblogiw.c:(.text+0x44212b): undefined reference to `__dma_request_channel'
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
Richard Röjfors [Mon, 8 Nov 2010 13:45:44 +0000 (10:45 -0300)]
[media] media: Add timberdale video-in driver
This patch adds the timberdale video-in driver.
The video IP of timberdale delivers the video data via DMA.
The driver uses the DMA api to handle DMA transfers, and make use
of the V4L2 video buffers to handle buffers against user space.
If available the driver uses an encoder to get/set the video standard
Signed-off-by: Richard Röjfors <richard.rojfors@pelagicore.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
Joe Perches [Sun, 7 Nov 2010 20:48:21 +0000 (17:48 -0300)]
[media] drivers/staging/cx25821: Use pr_fmt and pr_<level>
Fix several defects with bad line continuation uses that
introduce whitespace.
Fix several defects with lines missing "\n".
Standardize prefixes via pr_fmt.
Remove internal cx25821 prefixes.
Standardize :%s():" ... __func__ uses.
Coalesce long formats.
Add KERN_<level> prefixes via pr_<level> to printks.
Signed-off-by: Joe Perches <joe@perches.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
It seems that ix2505v driver ignores a i2c error in ix2505v_read_status_reg.
This looks like a typing error using (ret = 1) instead of correct (ret == 1).
Jean Delvare [Sun, 7 Nov 2010 15:54:39 +0000 (12:54 -0300)]
[media] TM6000: Drop unused macro
Signed-off-by: Jean Delvare <khali@linux-fr.org> Cc: Michel Ludwig <michel.ludwig@gmail.com> Cc: Stefan Ringel <stefan.ringel@arcor.de> Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
Jean Delvare [Sun, 7 Nov 2010 15:53:44 +0000 (12:53 -0300)]
[media] TM6000: Clean-up i2c initialization
Usage of templates for large structures is a bad idea, as it wastes a
lot of space. Manually initializing the few fields we need is way more
efficient.
Also set the algorithm data const, use strlcpy instead of strcpy, fix
a small race (device data must always be set before registering said
device) and properly return error on adapter registration failure.
Signed-off-by: Jean Delvare <khali@linux-fr.org> Cc: Michel Ludwig <michel.ludwig@gmail.com> Cc: Stefan Ringel <stefan.ringel@arcor.de> Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
Now, just one old bttv board uses the old RC5 raw decoding routines.
Its conversion to rc-core requires the generation of IRQ data for both
positive and negative transitions at the IRQ line. I'm not sure if
bttv driver supports it or if the transitions will be reliable enough.
So, due to the lack of hardware for testing, the better for now is to
just move the legacy routines to bttv driver, and wait for someone with
a Nebula Digi could help to port it to use also rc-core raw decoders.
Acked-by: Hermann Pitton <hermann-pitton@arcor.de> Acked-by: Jarod Wilson <jarod@redhat.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
[media] saa7134: Remove legacy IR decoding logic inside the module
The only IR left still using the old raw decoders on saa7134 is ENCORE
FM 5.3. As it is now using the standard rc-core raw decoders, lots
of old code can be removed from saa7134.
Acked-by: Jarod Wilson <jarod@redhat.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
Hans de Goede [Wed, 27 Oct 2010 12:12:30 +0000 (09:12 -0300)]
[media] gspca-stv06xx: support bandwidth changing
stv06xx devices have only one altsetting, but the actual used
bandwidth can be programmed through a register. We were already
setting this register lower then the max packetsize of the altsetting
indicates. This patch makes the gspca-stv06xx update the usb descriptor
for the alt setting to reflect the actual packetsize in use, so that
the usb subsystem uses the correct information for scheduling usb transfers.
This patch also tries to fallback to lower speeds in case a ENOSPC error
is received when submitting urbs, but currently this is only supported
with stv06xx cams with the pb0100 sensor, as this is the only one for
which we know how to change the framerate.
This patch is based on an initial incomplete patch by
Lee Jones <lee.jones@canonical.com>
Signed-off-by: Lee Jones <lee.jones@canonical.com> Signed-off-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
Dan Carpenter [Tue, 2 Nov 2010 03:22:23 +0000 (00:22 -0300)]
[media] saa7164: make buffer smaller
This isn't a runtime bug, it's just to make static checkers happy.
In vidioc_querycap() we copy a saa7164_dev ->name driver array into a
v4l2_capability -> driver array. The ->driver array is only 16 chars
long so ->name also can't be more than 16 characters.
The ->name gets set in v4l2_capability() and it always is less than 16
characters so we can easily make the buffer smaller.
Signed-off-by: Dan Carpenter <error27@gmail.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
VDR User [Sat, 30 Oct 2010 18:49:49 +0000 (15:49 -0300)]
[media] dvb-usb-gp8psk: get firmware and fpga versions
This patch adds retrieval of firmware and FPGA versions of Genpix devices.
That information is useful for users who experience performance differences
with the various firmware versions, and may want to use a specific firmware
that best suits their needs.
Example dmesg output:
gp8psk: FW Version = 2.09.4 (0x20904) Build 2009/04/02
gp8psk: FPGA Version = 1
Signed-off-by: Derek Kelly <user.vdr@gmail.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
There was a bug at cx231xx-input, where it were registering the remote
controls twice, one via ir-kbd-i2c and another directly.
Also, the patch that added rc_register_device() broke compilation for it.
This patch fixes cx231xx-input by fixing the depends on, to point to the
new symbol, and initializing the scanmask via platform_data.
While here, also fix Kconfig symbol change for IR core dependencies.
[media] ir-kbd-i2c: add rc_dev as a parameter to the driver
There are several fields on rc_dev that drivers can benefit. Allow drivers
to pass it as a parameter to the driver.
For now, the rc_dev parameter is optional. If drivers don't pass it, create
them internally. However, the best is to create rc_dev inside the drivers,
in order to fill other fields, like open(), close(), driver_name, etc.
So, a latter patch making it mandatory and changing the caller drivers is
welcome.
David Härdeman [Fri, 29 Oct 2010 19:08:23 +0000 (16:08 -0300)]
[media] ir-core: make struct rc_dev the primary interface
This patch merges the ir_input_dev and ir_dev_props structs into a single
struct called rc_dev. The drivers and various functions in rc-core used
by the drivers are also changed to use rc_dev as the primary interface
when dealing with rc-core.
This means that the input_dev is abstracted away from the drivers which
is necessary if we ever want to support multiple input devs per rc device.
The new API is similar to what the input subsystem uses, i.e:
rc_device_alloc()
rc_device_free()
rc_device_register()
rc_device_unregister()
[mchehab@redhat.com: Fix compilation on mceusb and cx231xx, due to merge conflicts] Signed-off-by: David Härdeman <david@hardeman.nu> Acked-by: Jarod Wilson <jarod@redhat.com> Tested-by: Jarod Wilson <jarod@redhat.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
David Härdeman [Wed, 10 Nov 2010 14:04:19 +0000 (11:04 -0300)]
[media] rc-core: Code cleanup after merging rc-sysfs and rc-map into rc-main
[mchehab@redhat.com: this patch were originally bundled with some renaming
stuff and with the file merges, as seen at:
https://patchwork.kernel.org/patch/291092/.
Instead of using the original approach, I wrote the rename patches and the
code merge as separate changes, then applied the difference on this patch.
This way, it is easier to see the real changes at the code, and will be easier
to merge upstream, especially if some conflict rises on the renaming patches]
Signed-off-by: David Härdeman <david@hardeman.nu> Acked-by: Jarod Wilson <jarod@redhat.com> Tested-by: Jarod Wilson <jarod@redhat.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
David Härdeman [Fri, 29 Oct 2010 19:08:07 +0000 (16:08 -0300)]
[media] ir-core: remove remaining users of the ir-functions keyhandlers
This patch removes the remaining usages of the ir_input_nokey() and
ir_input_keydown() functions provided by drivers/media/IR/ir-functions.c
by using the corresponding functionality in ir-core instead.
Signed-off-by: David Härdeman <david@hardeman.nu> Acked-by: Jarod Wilson <jarod@redhat.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
David Härdeman [Fri, 29 Oct 2010 19:08:02 +0000 (16:08 -0300)]
[media] ir-core: convert drivers/media/video/cx88 to ir-core
This patch converts the cx88 driver (for sampling hw) to use the
decoders provided by ir-core instead of the separate ones provided
by ir-functions (and gets rid of those).
The value for MO_DDS_IO had a comment saying it corresponded to
a 4kHz samplerate. That comment was unfortunately misleading. The
actual samplerate was something like 3250Hz.
The current value has been derived by analyzing the elapsed time
between interrupts for different values (knowing that each interrupt
corresponds to 32 samples).
Thanks to Mariusz Bialonczyk <manio@skyboo.net> for testing my patches
(about one a day for two weeks!) on actual hardware.
Signed-off-by: David Härdeman <david@hardeman.nu> Acked-by: Jarod Wilson <jarod@redhat.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
Jarkko Nikula [Fri, 29 Oct 2010 14:31:39 +0000 (11:31 -0300)]
[media] radio-si4713: Add regulator framework support
Convert the driver to use regulator framework instead of set_power callback.
This with gpio_reset platform data provide cleaner way to manage chip VIO,
VDD and reset signal inside the driver.
[media] au0828: enable VBI timeout when calling read() without streamon()
Because some clients (e.g. tvtime) will just open the filehandle and call
read() instead of explicitly calling streamon first, we need to make sure the
VBI timeout routine gets enabled in this case.
The first pass of fixes to cause the driver to continue streaming even when
there is no video arriving over the ITU656 bus were not adequate. Continue
the work from the previous patch.
[media] au0828: continue video streaming even when no ITU-656 coming in
We need the au0828 to continue delivering frames even when the device is not
delivering video, or else applications such as tvtime will block indefinitely.
Unfortunately, the au8522 doesn't have any sort of free-running mode or "blue
screen on no video" like some other decoders.
[media] au8522: cleanup code which disables audio decoder
Cleanup the routine which disables the audio decoder. In particular, the i2s
output was being shut off but the audio processor itself was not, resulting
in more power being consumed than needed.
[media] au0828: Fix field alignment for video frames delivered by driver
Fix the alignment of fields being delivered by the driver, which was
resulting in jerky video when there was horizontal motion (since the bottom
field of the frame was being sent with the top field of the previous frame)
[media] au0828: set max packets per URB to match Windows driver
We had the max packets per URB set to an incredibly low "10" as opposed to
the Windows driver which has it as 128. Set the value to match the
Windows driver, which will reduce interrupts considerably.
The chip's default value for the brightness didn't match what we were sending
back in the queryctrl ioctl(), so if the application actually set the
brightness to the "default", it would actually end up being way too bright.
Hans de Goede [Tue, 26 Oct 2010 14:18:59 +0000 (11:18 -0300)]
[media] gspca_xirlink_cit: various usb bandwidth allocation improvements / fixes
The following usb bandwidth allocation changes were made to the ibm netcam
pro code:
- Don't restart negotiation at max packet size on stop0, as that gets called
by gspca_main during negotiation. Move this to sd_isoc_init.
- Don't ask for full bandwidth when running at 160x120, that does not need
full bandwidth
- Make minimum acceptable bandwidth depend upon resolution
[mchehab@redhat.com: Fix CodingStyle problems at switch statements] Signed-off-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
Hans de Goede [Tue, 26 Oct 2010 08:16:32 +0000 (05:16 -0300)]
[media] gspca: only set gspca->int_urb if submitting it succeeds
Currently alloc_and_submit_int_urb() is setting gspca->int_urb
as soon as the allocation has succeeded, but if the subsequent
submit fails, the urb gets destroyed. And then later will
get destroyed again in gspca_input_destroy_urb() because
gspca->int_urb is set, leading to a double free.
Signed-off-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
Hans de Goede [Tue, 26 Oct 2010 08:15:46 +0000 (05:15 -0300)]
[media] gspca: submit interrupt urbs *after* isoc urbs
Currently gspca supported usb-1.1 webcams for which we support the input
button through an interrupt endpoint won't stream (not enough bandwidth
error) when used through an USB-2.0 hub.
After much debugging I've found out that the cause for this is that the
ehci-sched.c schedeling code does not like it when there are already urb's
scheduled when (large) isoc urbs are queued. By moving the submission
of the interrupt urbs to after submitting the isoc urbs the camera
starts working again through usb-2.0 hubs.
Note that this does not fix isoc. streaming through a usb-hub while another
1.1 usb device (like the microphone of the same cam) is also active
at the same time :(
I've spend a long time analyzing the linux kernel ehci scheduler code,
resulting in this (long) mail:
http://www.spinics.net/lists/linux-usb/msg37982.html
The conclusion of the following mail thread is that yes there are several
issues when using usb-1.1 devices through a usb-2.0 hub, but these are not
easily fixable in the current code. Fixing this in ehci-sched.c requires
an almost full rewrite, which is not bound to happen anytime soon.
So with this patch gspca driven usb-1.1 webcams will atleast work when
connected through an usb-2.0 hub when the microphone is not used.
As an added bonus this patch avoids extra destroy/create input urb cycles
when we end up falling back to a lower speed alt setting because of bandwidth
limitations.
Signed-off-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
Daniel Drake [Wed, 27 Oct 2010 13:55:00 +0000 (10:55 -0300)]
[media] cafe_ccic: fix colorspace corruption on resume
If you suspend and resume during video capture, the video colours
are corrupted on resume. This is because the sensor is being unconditionally
powered off during the resume path.
Only power down during resume if the camera is not in use, and correctly
reconfigure the sensor during resume.
Fixes http://dev.laptop.org/ticket/10190
Signed-off-by: Daniel Drake <dsd@laptop.org> Acked-by: Jonathan Corbet <corbet@lwn.net> Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
Although cx231xx has a very good IR support, already supported by
mceusb driver, some designs decided to add a separate I2C
microcontroller chip in order to handle IR.
Due to that, add a glue to ir-kbd-i2c is needed, in order to support
those devices.
By comparing the traffic between Pixelview (cx23102-based and Kworld
(saa7134-based), the only difference is at register 0xd5. After some
tests, It seems that it is used to change mode between serial and parallel.
This adds the trivial bits to mb86a20s. As the driver won't touch
at the channel/layer parameters, this may not be enough for
receiving all channels, especially ISDB-Tsb, but the driver worked
properly for receiving video channels on my tests.
On our tests with Leadership ISDBT, the s921 frontend were not work. As its
design contained some weird things, it ended to be easier to just re-write
it, getting another frontend as an example (cx24123).
As the old s921 driver weren't used, there's no regression. Some info from
the old frontend were used as a way to double check the behavior that were
noticed on the USB dumps retrieved from Leadership driver.
Signed-off-by: Douglas Schilling Landgraf <dougsland@redhat.com> Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
Linus Torvalds [Tue, 28 Dec 2010 23:54:24 +0000 (15:54 -0800)]
Merge branch 'perf-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip
* 'perf-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
ring_buffer: Off-by-one and duplicate events in ring_buffer_read_page
Linus Torvalds [Tue, 28 Dec 2010 23:53:54 +0000 (15:53 -0800)]
Merge branch 'x86-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip
* 'x86-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
x86/microcode: Fix double vfree() and remove redundant pointer checks before vfree()
Imre Kaloz [Mon, 27 Dec 2010 21:59:57 +0000 (22:59 +0100)]
ARM: fix IXP4xx build failure
arm: export dma_set_coherent_mask
While a regression was fixed with commit 710224fa2750cf (arm: fix
"arm: fix pci_set_consistent_dma_mask for dmabounce devices"), a
new one was introduced as dma_set_coherent_mask wasn't exported
for modules. This patch takes care for this issue.
Signed-off-by: Imre Kaloz <kaloz@openwrt.org> Signed-off-by: Krzysztof Hałasa <khc@pm.waw.pl> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Ahmed S. Darwish [Sat, 25 Dec 2010 09:57:09 +0000 (11:57 +0200)]
RAMOOPS: Don't overflow over non-allocated regions
The current code mis-calculates the ramoops header size, leading to an
overflow over the next record at best, or over a non-allocated region at
worst. Fix that calculation.
Signed-off-by: Ahmed S. Darwish <darwish.07@gmail.com> Acked-by: Marco Stornelli <marco.stornelli@gmail.com> Cc: stable@kernel.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Linus Torvalds [Mon, 27 Dec 2010 18:36:27 +0000 (10:36 -0800)]
Merge branch 'nommu-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/lethal/nommu-2.6
* 'nommu-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/lethal/nommu-2.6:
nommu: Provide stubbed alloc/free_vm_area() implementation.
nommu: Fix up vmalloc_node() symbol export regression.