- 21 Oct, 2010 40 commits
-
-
Mauro Carvalho Chehab authored
Instead of doing its own lock, use core-assisted one. As a bonus, it will do the proper unlock during queue wait events. This fixes a long-standing bug where softwares like tvtime would hang if you try to use cx231xx-alsa. Signed-off-by:Mauro Carvalho Chehab <mchehab@redhat.com>
-
Mauro Carvalho Chehab authored
Signed-off-by: -
Mauro Carvalho Chehab authored
This file came originally from cx23885 driver. Some functions aren't used. Now that they are declared as static, we have those errors: drivers/media/video/cx231xx/cx231xx-417.c:615: warning: ‘mc417_gpio_set’ defined but not used drivers/media/video/cx231xx/cx231xx-417.c:625: warning: ‘mc417_gpio_clear’ defined but not used drivers/media/video/cx231xx/cx231xx-417.c:635: warning: ‘mc417_gpio_enable’ defined but not used As they're not used, just remove them. If needed, they can be restored from the git logs or from the cx23885 driver. Signed-off-by: -
Mauro Carvalho Chehab authored
drivers/media/video/cx23885/built-in.o: In function `mc417_memory_write': /home/v4l/v4l/patchwork/drivers/media/video/cx23885/cx23885-417.c:482: multiple definition of `mc417_memory_write' drivers/media/video/cx231xx/built-in.o:/home/v4l/v4l/patchwork/drivers/media/video/cx231xx/cx231xx-417.c:477: first defined here drivers/media/video/cx23885/built-in.o: In function `mc417_gpio_set': /home/v4l/v4l/patchwork/drivers/media/video/cx23885/cx23885-417.c:636: multiple definition of `mc417_gpio_set' drivers/media/video/cx231xx/built-in.o:/home/v4l/v4l/patchwork/drivers/media/video/cx231xx/cx231xx-417.c:615: first defined here drivers/media/video/cx23885/built-in.o: In function `mc417_gpio_enable': /home/v4l/v4l/patchwork/drivers/media/video/cx23885/cx23885-417.c:656: multiple definition of `mc417_gpio_enable' drivers/media/video/cx231xx/built-in.o:/home/v4l/v4l/patchwork/drivers/media/video/cx231xx/cx231xx-417.c:635: first defined here drivers/media/video/cx23885/built-in.o: In function `mc417_memory_read': /home/v4l/v4l/patchwork/drivers/media/video/cx23885/cx23885-417.c:546: multiple definition of `mc417_memory_read' drivers/media/video/cx231xx/built-in.o:/home/v4l/v4l/patchwork/drivers/media/video/cx231xx/cx231xx-417.c:541: first defined here drivers/media/video/cx23885/built-in.o: In function `mc417_gpio_clear': /home/v4l/v4l/patchwork/drivers/media/video/cx23885/cx23885-417.c:646: multiple definition of `mc417_gpio_clear' drivers/media/video/cx231xx/built-in.o:/home/v4l/v4l/patchwork/drivers/media/video/cx231xx/cx231xx-417.c:625: first defined here drivers/media/video/cx23885/built-in.o: In function `mc417_register_read': /home/v4l/v4l/patchwork/drivers/media/video/cx23885/cx23885-417.c:388: multiple definition of `mc417_register_read' drivers/media/video/cx231xx/built-in.o:/home/v4l/v4l/patchwork/drivers/media/video/cx231xx/cx231xx-417.c:401: first defined here drivers/media/video/cx23885/built-in.o: In function `mc417_register_write': /home/v4l/v4l/patchwork/drivers/media/video/cx23885/cx23885-417.c:324: multiple definition of `mc417_register_write' drivers/media/video/cx231xx/built-in.o:/home/v4l/v4l/patchwork/drivers/media/video/cx231xx/cx231xx-417.c:343: first defined here Signed-off-by: -
Mauro Carvalho Chehab authored
gcc didn't like to have i++ inside a complex operation: drivers/media/video/cx231xx/cx231xx-417.c: In function ‘cx231xx_load_firmware’: drivers/media/video/cx231xx/cx231xx-417.c:1059: warning: operation on ‘i’ may be undefined drivers/media/video/cx231xx/cx231xx-417.c:1061: warning: operation on ‘i’ may be undefined drivers/media/video/cx231xx/cx231xx-417.c:1063: warning: operation on ‘i’ may be undefined Btw, I agree with gcc, as we're using i and i++ at the same operation and, depending on how optimization may occur, it may produce a wrong code. While here, fix CodingStyle issues on the changed code. Acked-by:
Sri Deevi <Srinivasa.Deevi@conexant.com> Signed-off-by:
-
Mauro Carvalho Chehab authored
The patches received from the vendor contained a lot of CodingStyle issues. Cleans the style issues reported by checkpatch.pl on those drivers. It is better to do such style fixes when merging a big set of changes than latter. Of course, the better is to receive patches already cleaned ;) Acked-by:
-
Mauro Carvalho Chehab authored
Acked-by:
-
Mauro Carvalho Chehab authored
By default, tda18271 tries to optimize I2C bus by updating all registers at the same time. Unfortunately, some devices doesn't support it. The current logic has a problem when small_i2c is equal to 8, since there are some transfers using 11 + 1 bytes. Fix the problem by enforcing the max size at the right place, and allows reducing it to max = 3 + 1. Acked-by:
Michael Krufky <mkrufky@kernellabs.com> Acked-by:
-
Mauro Carvalho Chehab authored
Acked-by:
-
Mauro Carvalho Chehab authored
Improves the logic, for it to be clearer and to avoid having board-dependent config there. Acked-by:
-
Mauro Carvalho Chehab authored
The driver has a field to indicate what bus is used by tuner and by demod. However, this field were never used. On Pixelview, it uses I2C 2 for tuner, instead of I2C 1. drivers/media/video/cx231xx/cx231xx-cards.c Acked-by:
-
Mauro Carvalho Chehab authored
This allows usage of a tool at v4l-utils tree to parse URB messages and display them to userspace. The tool is available at: http://git.linuxtv.org/v4l-utils.git?a=blob;f=contrib/cx231xx/parse_cx231xx.pl;hb=HEADAcked-by:
-
Mauro Carvalho Chehab authored
ERROR: "cx2341x_mpeg_ctrls" [drivers/media/video/cx231xx/cx231xx.ko] undefined! ERROR: "cx2341x_fill_defaults" [drivers/media/video/cx231xx/cx231xx.ko] undefined! ERROR: "cx2341x_log_status" [drivers/media/video/cx231xx/cx231xx.ko] undefined! ERROR: "cx2341x_ctrl_get_menu" [drivers/media/video/cx231xx/cx231xx.ko] undefined! ERROR: "cx2341x_update" [drivers/media/video/cx231xx/cx231xx.ko] undefined! ERROR: "cx2341x_ctrl_query" [drivers/media/video/cx231xx/cx231xx.ko] undefined! ERROR: "cx2341x_ext_ctrls" [drivers/media/video/cx231xx/cx231xx.ko] undefined! Acked-by:
-
Mauro Carvalho Chehab authored
drivers/media/video/cx231xx/cx231xx-avcore.c:1608: warning: format ‘%d’ expects type ‘int’, but argument 3 has type ‘long unsigned int’ drivers/media/video/cx231xx/cx231xx-417.c:1047: warning: format ‘%d’ expects type ‘int’, but argument 3 has type ‘size_t’ Signed-off-by: -
Devin Heitmueller authored
Adjust the vblank and vactive counts so that they don't throw an error in cx25840's set_std call (we did an equivalent change for NTSC when we got the scaler working). Signed-off-by:
Devin Heitmueller <dheitmueller@hauppauge.com> Signed-off-by:
-
Devin Heitmueller authored
The cx231xx_do_mode_ctrl_overrides() function was not touching the vactive line count for PAL/SECAM modes, which in some use cases results in it being left in the chip default state of 480 (NTSC). Explicitly set the values, as is already done for NTSC. Signed-off-by:
-
Devin Heitmueller authored
Fix some indentation problems and remove an "if (1)" from the Colibri setup function. Signed-off-by:
-
Devin Heitmueller authored
The cx231xx_set_Colibri_For_LowIF() function is only ever called if the tuner has a DIF (see vidioc_s_frequency() in cx231xx-video.c). Hence, we do not need to do a board specific check in this function. Signed-off-by:
-
Devin Heitmueller authored
The current code was deciding whether a DIF was present based on the board profile. However, this is just another thing for someone to get wrong when adding new boards. Make the decision based on the tuner instead, so that a developer adding new boards only needs to specify which tuner the device has. Of course, the first time somebody adds a board with a tuner other than xc5000 or tda18271, he/she will need to add another line for that tuner. But we provide a friendly message in dmesg to let them know that. Signed-off-by:
-
Devin Heitmueller authored
Nobody is ever going to implement an i2c based IR controller on a bridge that has an onboard universal IR receiver. This stuff was all copied from em28xx, which has old enough versions of the chip that some didn't have onboard IR. Remove the stubs related to i2c based IR (keeping the cx231xx-input code). Signed-off-by:
-
Devin Heitmueller authored
Move a printk() message which refers to enabling the cx23417 so that it only shows up on a board that has the cx23417. Signed-off-by:
-
Devin Heitmueller authored
The video grabber reference design (Veyron) does not have a tuner input, so do not have it defined in the board profile. Signed-off-by:
-
Devin Heitmueller authored
Extend the board profile structure to allow configuration of the output mode. Right now they are all doing VIP 1.1 format, but we have a board that needs ITU656 format (which hasn't been checked in yet). Signed-off-by:
-
Devin Heitmueller authored
Add initial support for the Hauppauge USBLive 2 (2040:c200). Note that I had to copy a bunch of the case statements used for the Conexant video grabber reference design (which also doesn't have a tuner). This will likely need to be refactored out into the board profile. Signed-off-by:
-
Devin Heitmueller authored
The tuner is on i2c port 3 just like all the reference designs. Having it improperly set to port #1 causes the tuner initialization to fail when enabling the device. Signed-off-by:
-
Devin Heitmueller authored
If switching to a tuner input, reset the standard and tune to the last known frequency. We need to do this in particular for this bridge since the tuner gets powered down when captuing on the composite or s-video inputs. Signed-off-by:
-
Devin Heitmueller authored
Remove a hack which was jammed into s_input to force the device into bulk mode at runtime (an abuse of the API). If this sort of functionality is actually needed (aside from the existing "transfer_mode" modprobe variable), a patch can be submitted which makes use of a private control. Signed-off-by:
-
Devin Heitmueller authored
The Exeter hardware design does not use GPIOs to manage whether its in digital mode or analog mode, but we need to setup the power control properly. For that board, setup power control and remove the mux select call. It is highly likely that this change could be used by other boards as well, which would make power management cleaner (fixing known issues transitioning between analog and digital mode). Signed-off-by:
-
Devin Heitmueller authored
You cannot call the video_mux routine from within the isoc setup, because that code is shared with the digital isoc handler. This was causing the GPIOs and power control to be put into the wrong state when starting up digital mode. Signed-off-by:
-
Devin Heitmueller authored
We need to clear out the field before setting individual bits, or else we end up with a union of whatever was there and what we are trying to set. For example, switching to digital mode ends up being 0x30 instead of 0x10 if we were previously in analog tv mode. Signed-off-by:
-
Devin Heitmueller authored
Move the responsibility for setting up the horizontal and vertical scalers entirely to the cx25840 driver. The cx231xx-avcore was actually programming garbage into the HSCALE_CTRL and VSCALE_CTRL registers (because of differences in how the em28xx driver worked, which the cx231xx driver was derived from). The net effect is that the scaler now works properly (tested with both PAL and NTSC under mplayer and tvtime). This patch also gets rid of cx25840 errors showing up in dmesg which say "720x480 is not a valid size" (since we now properly setup the size of the active video area). Signed-off-by:
-
Devin Heitmueller authored
Change the log level from info to debug for some log events that occur frequently and should never need to be seen in normal operation. Signed-off-by:
-
Devin Heitmueller authored
There is no need for a switch statement here. Use the contents of the board profile to dictate the tuner driver and i2c address. Eventually if a board ever comes around which has a different i2c bus than #1, well that should be a field in the board profile as well. Signed-off-by:
-
Devin Heitmueller authored
The current code was sending one videobuf per field (despite having specified V4L2_FIELD_SEQ_TB during setup). As a result, application which used the read() interface would work, except they would sometimes have the fields reversed (depending on the luck of which field the device was on when the application started VBI capture). The net effect was that CC decoding would only work about 50% of the time. Restructure the VBI code a bit so that works like all the other drivers, such that both fields are delivered in a single videobuf buffer, which ensures that they are always received in a predictable order. Signed-off-by:
-
Devin Heitmueller authored
The VBI sampling rate and offset were incorrectly specified, which resulted in CC data not being rendered under zvbi or tvtime. Set the fields correctly. Signed-off-by:
-
Devin Heitmueller authored
Change a %x to a %p since the variable is a pointer Signed-off-by:
-
Devin Heitmueller authored
Get rid of warning about unused variable Signed-off-by:
-
Devin Heitmueller authored
Run Lindent and fix a few spacing issues. This patch makes no functional change to the driver. Signed-off-by:
-
Devin Heitmueller authored
Configure the tda18271 to use a smaller transaction size by default, which works around some sort of i2c bug in the Polaris driver (which needs to be debugged). This should be safe for other boards (being in tuner-core means it will be enabled by default), although testing needs to be done. Signed-off-by:
-
Devin Heitmueller authored
Fix case where analog calls come in while the DVB side of the board is still initializing. This patch is actually just an exact port of the same patch made by Mauro to em28xx in hg rev 14762. Signed-off-by:
-
