Guys,
As suggested, I've redesigned cx88 driver to work with video_ioctl2. The
newer tree is available at:
http://linuxtv.org/hg/~mchehab/cx88_ioctl2
For those worried about code size increasing, those are the numbers, on
i586, for the previous approach:
text data bss dec hex filename
21411 4384 12 25807 64cf cx8800.ko
33382 19012 292 52686 cdce cx88xx.ko
12439 1484 12 13935 366f cx88-blackbird.ko
20249 856 1024 22129 5671 videodev.ko
87481 25736 1340 114557 1bf7d (TTALS)
With the newer design, we have a size an overall size reduction:
text data bss dec hex filename
19445 3552 612 23609 5c39 cx8800.ko
33378 19012 292 52682 cdca cx88xx.ko
12765 1516 0 14281 37c9 cx88-blackbird.ko
21020 856 1024 22900 5974 videodev.ko
86608 24936 1928 113472 1bb40 (TTALS)
It should also be noticed that, after having all modules using
video_ioctl2, several stuff from v4l2_common can be removed, reducing
the size also for videodev.
Cheers,
Mauro.

Em Sex, 2006-11-17 12:13 -0200, Mauro Carvalho Chehab escreveu:
Guys,

As suggested, I've redesigned cx88 driver to work with video_ioctl2. The
newer tree is available at:

http://linuxtv.org/hg/~mchehab/cx88_ioctl2

I've added another branch there, at:
http://linuxtv.org/hg/~mchehab/cx88_improvements

The basic difference is a patch to make cx88 to use tuner signal level,
instead of using the stereo carrier as signal strength indicator.

This patch requires more testing and may benefit other designs where the
video decoder is not capable of measuring signal quality.

P.S.: Previously, this patch were at cx88_ioctl2, since I used the newer
cx88 design for those tests. So, for those people who cloned
cx88_ioctl2, they should re-clone it to test just cx88 conversion to
video_ioctl2.

About the conversion: the only expected effect is the memory size
reduction.

We really want testers here for cx88-based boards and cx88-blackbird
designs on cx88_ioctl2 tree.

Can we have community feedback about this?

Cheers,
Mauro.

hi,

i've hauppauge wintv go2 in side my box, working with cx88 driver.
i've cloned the cx88_ioctl2 directory yesterday.
everything worked.
recloned today cx88_iotcl2 and cloned cx88_improvements.
all working so far.

do you expect some special test's? or dmesg outputs?

regards
tom

Am Samstag, den 18.11.2006, 09:30 -0200 schrieb Mauro Carvalho Chehab:
Em Sex, 2006-11-17 *s 12:13 -0200, Mauro Carvalho Chehab escreveu:
Guys,

As suggested, I've redesigned cx88 driver to work with video_ioctl2. The
newer tree is available at:

http://linuxtv.org/hg/~mchehab/cx88_ioctl2

I've added another branch there, at:
http://linuxtv.org/hg/~mchehab/cx88_improvements

The basic difference is a patch to make cx88 to use tuner signal level,
instead of using the stereo carrier as signal strength indicator.

This patch requires more testing and may benefit other designs where the
video decoder is not capable of measuring signal quality.

P.S.: Previously, this patch were at cx88_ioctl2, since I used the newer
cx88 design for those tests. So, for those people who cloned
cx88_ioctl2, they should re-clone it to test just cx88 conversion to
video_ioctl2.

About the conversion: the only expected effect is the memory size
reduction.

We really want testers here for cx88-based boards and cx88-blackbird
designs on cx88_ioctl2 tree.

Can we have community feedback about this?

Cheers,
Mauro.

Em S, 2006-11-18 13:21 +0100, tom escreveu:
hi,

i've hauppauge wintv go2 in side my box, working with cx88 driver.
i've cloned the cx88_ioctl2 directory yesterday.
everything worked.
recloned today cx88_iotcl2 and cloned cx88_improvements.
all working so far.
Thanks for you help, Tom.

do you expect some special test's? or dmesg outputs?

It would be interesing if you can send us the dmesg and test with more
than one app. Tvtime uses V4L2 api, while xawtv/xdtv uses the legacy
V4L1 API.

Have you tested also radio?

regards
tom

With cx88_improvements, scantv should also work, even for channels that
are not transmitting in stereo (here, only one open channel sometimes
transmit in stereo, so, with the older driver, no channel were detected
by scantv). To test cx88_improvements, it would be interesting to
modprobe tuner with debug=1 parameter, before loading cx88 modules. This
will generate some interesting debug messages that may help the test.

Cheers,
Mauro.

Hi Mauro,

ok i'll send my dmesg after a reboot in next 5 min.
i've installed tvtime and zapping and with both it works.
yes, gnomeradio but this doesn't work, no stations no noise

Am Samstag, den 18.11.2006, 14:43 -0200 schrieb Mauro Carvalho Chehab:
Em Sb, 2006-11-18 *s 13:21 +0100, tom escreveu:
hi,

i've hauppauge wintv go2 in side my box, working with cx88 driver.
i've cloned the cx88_ioctl2 directory yesterday.
everything worked.
recloned today cx88_iotcl2 and cloned cx88_improvements.
all working so far.
Thanks for you help, Tom.

do you expect some special test's? or dmesg outputs?

It would be interesing if you can send us the dmesg and test with more
than one app. Tvtime uses V4L2 api, while xawtv/xdtv uses the legacy
V4L1 API.

Have you tested also radio?

regards
tom

With cx88_improvements, scantv should also work, even for channels that
are not transmitting in stereo (here, only one open channel sometimes
transmit in stereo, so, with the older driver, no channel were detected
by scantv). To test cx88_improvements, it would be interesting to
modprobe tuner with debug=1 parameter, before loading cx88 modules. This
will generate some interesting debug messages that may help the test.

Cheers,
Mauro.

installed xawtv, also no prob so far.
do you expect recording test or only tv watching?

regards
tom

Am Samstag, den 18.11.2006, 14:43 -0200 schrieb Mauro Carvalho Chehab:
Em Sb, 2006-11-18 *s 13:21 +0100, tom escreveu:
hi,

i've hauppauge wintv go2 in side my box, working with cx88 driver.
i've cloned the cx88_ioctl2 directory yesterday.
everything worked.
recloned today cx88_iotcl2 and cloned cx88_improvements.
all working so far.
Thanks for you help, Tom.

do you expect some special test's? or dmesg outputs?

It would be interesing if you can send us the dmesg and test with more
than one app. Tvtime uses V4L2 api, while xawtv/xdtv uses the legacy
V4L1 API.

Have you tested also radio?

regards
tom

With cx88_improvements, scantv should also work, even for channels that
are not transmitting in stereo (here, only one open channel sometimes
transmit in stereo, so, with the older driver, no channel were detected
by scantv). To test cx88_improvements, it would be interesting to
modprobe tuner with debug=1 parameter, before loading cx88 modules. This
will generate some interesting debug messages that may help the test.

Cheers,
Mauro.

ok, attached output from /varlog/messages, lspci, lspci -n, lspci-v

regards
tom

Am Samstag, den 18.11.2006, 14:43 -0200 schrieb Mauro Carvalho Chehab:
Em Sb, 2006-11-18 *s 13:21 +0100, tom escreveu:
hi,

i've hauppauge wintv go2 in side my box, working with cx88 driver.
i've cloned the cx88_ioctl2 directory yesterday.
everything worked.
recloned today cx88_iotcl2 and cloned cx88_improvements.
all working so far.
Thanks for you help, Tom.

do you expect some special test's? or dmesg outputs?

It would be interesing if you can send us the dmesg and test with more
than one app. Tvtime uses V4L2 api, while xawtv/xdtv uses the legacy
V4L1 API.

Have you tested also radio?

regards
tom

With cx88_improvements, scantv should also work, even for channels that
are not transmitting in stereo (here, only one open channel sometimes
transmit in stereo, so, with the older driver, no channel were detected
by scantv). To test cx88_improvements, it would be interesting to
modprobe tuner with debug=1 parameter, before loading cx88 modules. This
will generate some interesting debug messages that may help the test.

Cheers,
Mauro.

btw xawtv only starts with -nodga

Am Samstag, den 18.11.2006, 18:18 +0100 schrieb tom:
installed xawtv, also no prob so far.
do you expect recording test or only tv watching?

regards
tom

Am Samstag, den 18.11.2006, 14:43 -0200 schrieb Mauro Carvalho Chehab:
Em Sb, 2006-11-18 *s 13:21 +0100, tom escreveu:
hi,

i've hauppauge wintv go2 in side my box, working with cx88 driver.
i've cloned the cx88_ioctl2 directory yesterday.
everything worked.
recloned today cx88_iotcl2 and cloned cx88_improvements.
all working so far.
Thanks for you help, Tom.

do you expect some special test's? or dmesg outputs?

It would be interesing if you can send us the dmesg and test with more
than one app. Tvtime uses V4L2 api, while xawtv/xdtv uses the legacy
V4L1 API.

Have you tested also radio?

regards
tom

With cx88_improvements, scantv should also work, even for channels that
are not transmitting in stereo (here, only one open channel sometimes
transmit in stereo, so, with the older driver, no channel were detected
by scantv). To test cx88_improvements, it would be interesting to
modprobe tuner with debug=1 parameter, before loading cx88 modules. This
will generate some interesting debug messages that may help the test.

Cheers,
Mauro.

Sat, 18 Nov 2006, tom wrote:

btw xawtv only starts with -nodga

It used to start fine without -nodga before (e.g. with master version of
cx88?).

It seems to be unrelated to the driver changes, but, instead, due to some
troubles at video adapter driver.

Are you using nvidia or ati proprietary drivers?
--
Am Samstag, den 18.11.2006, 18:18 +0100 schrieb tom:
>installed xawtv, also no prob so far.
>do you expect recording test or only tv watching?
>>
>regards
>tom
>>
>Am Samstag, den 18.11.2006, 14:43 -0200 schrieb Mauro Carvalho Chehab:
Em Sb, 2006-11-18 *s 13:21 +0100, tom escreveu:
hi,

i've hauppauge wintv go2 in side my box, working with cx88 driver.
i've cloned the cx88_ioctl2 directory yesterday.
everything worked.
recloned today cx88_iotcl2 and cloned cx88_improvements.
all working so far.
Thanks for you help, Tom.

do you expect some special test's? or dmesg outputs?

It would be interesing if you can send us the dmesg and test with more
than one app. Tvtime uses V4L2 api, while xawtv/xdtv uses the legacy
V4L1 API.

Have you tested also radio?

regards
tom

With cx88_improvements, scantv should also work, even for channels that
are not transmitting in stereo (here, only one open channel sometimes
transmit in stereo, so, with the older driver, no channel were detected
by scantv). To test cx88_improvements, it would be interesting to
modprobe tuner with debug=1 parameter, before loading cx88 modules. This
will generate some interesting debug messages that may help the test.

Cheers,
Mauro.

--
video4linux-list mailing list
Unsubscribe @redhat.com?subject=unsubscribe

>>
>--
>video4linux-list mailing list
>Unsubscribe @redhat.com?subject=unsubscribe
>
>

Sat, 18 Nov 2006, tom wrote:
installed xawtv, also no prob so far.
do you expect recording test or only tv watching?
It would also be interesting to test recording.
Cheers,
Mauro Carvalho Chehab
http://linuxtv.org
mchehab (AT) infradead (DOT) org

striiped my dmesg output just the relevant lines:

[17179586.112000] cx2388x v4l2 driver version 0.0.6 loaded
[17179586.112000] ACPI: PCI Interrupt 0000:04:01.0[A] -GSI 19 (level,
low) -IRQ 217
[17179586.112000] CRE cx88[0]: subsystem: 0070:3401, board: Hauppauge
WinTV 34xxx models [card=1,autodetected]
[17179586.112000] TV tuner -1 at 0x1fe, Radio tuner -1 at 0x1fe
[17179586.136000] cx2388x alsa driver version 0.0.6 loaded
[17179586.152000] nvidia: module license 'NVIDIA' taints kernel.
[17179586.404000] Initializing USB Mass Storage driver
[17179586.404000] scsi2 : SCSI emulation for USB Mass Storage devices
[17179586.404000] usbcore: registered new driver usb-storage
[17179586.404000] USB Mass Storage support registered.
[17179586.420000] i2c-algo-bit.o: cx88[0] passed test.
[17179586.456000] tveeprom 0-0050: Hauppauge model 34704, rev J1B3,
Serial# 8521833
[17179586.456000] tveeprom 0-0050: tuner model is TCL M2523_3DB_E (idx
113, type 55)
[17179586.456000] tveeprom 0-0050: TV standards PAL(B/G) PAL(D/D1/K)
(eeprom 0x44)
[17179586.456000] tveeprom 0-0050: audio processor is CX881 (idx 31)
Nov 18 12:40:56 atlantis kernel: [17179586.456000] tveeprom 0-0050: has
radio
[17179586.456000] cx88[0]: warning: unknown hauppauge model #34704
Nov 18 12:40:56 atlantis kernel: [17179586.456000] cx88[0]: hauppauge
eeprom: model=34704
[17179586.456000] usbcore: registered new driver hiddev
[17179586.456000] input: cx88 IR (Hauppauge WinTV 34xxx
as /class/input/input2
[17179586.456000] cx88[0]/0: found at 0000:04:01.0, rev: 5, irq: 217,
latency: 32, mmio: 0xe5000000
[17179586.472000] input: Logitech USB-PS/2 Mouse
as /class/input/input3
[17179586.472000] input: USB HID v1.10 Mouse [Logitech USB-PS/2
Mouse] on usb-0000:00:1d.3-1
[17179586.472000] usbcore: registered new driver usbhid
[17179586.472000] drivers/usb/input/hid-core.c: v2.6:USB HID core driver
[17179586.500000] tuner 0-0061: chip found @ 0xc2 (cx88[0])
[17179586.500000] tuner 0-0061: type set to 55 (TCL 2002MB)
[17179586.512000] cx88[0]/0: registered device video0 [v4l2]
[17179586.512000] cx88[0]/0: registered device vbi0
[17179586.512000] cx88[0]/0: registered device radio0
[17179586.512000] ACPI: PCI Interrupt 0000:04:01.1[A] -GSI 19 (level,
low) -IRQ 217
[17179586.512000] cx88[0]/1: CX88x/0: ALSA support for cx2388x boards
[17179586.512000] ACPI: PCI Interrupt 0000:01:00.0[A] -GSI 16 (level,
low) -IRQ 169

regards
tom

Am Samstag, den 18.11.2006, 18:23 +0100 schrieb tom:
ok, attached output from /varlog/messages, lspci, lspci -n, lspci-v

regards
tom

Am Samstag, den 18.11.2006, 14:43 -0200 schrieb Mauro Carvalho Chehab:
Em Sb, 2006-11-18 *s 13:21 +0100, tom escreveu:
hi,

i've hauppauge wintv go2 in side my box, working with cx88 driver.
i've cloned the cx88_ioctl2 directory yesterday.
everything worked.
recloned today cx88_iotcl2 and cloned cx88_improvements.
all working so far.
Thanks for you help, Tom.

do you expect some special test's? or dmesg outputs?

It would be interesing if you can send us the dmesg and test with more
than one app. Tvtime uses V4L2 api, while xawtv/xdtv uses the legacy
V4L1 API.

Have you tested also radio?

regards
tom

With cx88_improvements, scantv should also work, even for channels that
are not transmitting in stereo (here, only one open channel sometimes
transmit in stereo, so, with the older driver, no channel were detected
by scantv). To test cx88_improvements, it would be interesting to
modprobe tuner with debug=1 parameter, before loading cx88 modules. This
will generate some interesting debug messages that may help the test.

Cheers,
Mauro.

nope, never worked with ubuntu edgy xawtv, maybe an old release of xawtv
so i don't think is driver related

i'll use nvidia open source driver, with xorg aiglx and beryl

regards
tom

Am Samstag, den 18.11.2006, 18:23 +0000 schrieb Mauro Carvalho Chehab:
Sat, 18 Nov 2006, tom wrote:

btw xawtv only starts with -nodga

It used to start fine without -nodga before (e.g. with master version of
cx88?).

It seems to be unrelated to the driver changes, but, instead, due to some
troubles at video adapter driver.

Are you using nvidia or ati proprietary drivers?
--
Am Samstag, den 18.11.2006, 18:18 +0100 schrieb tom:
>installed xawtv, also no prob so far.
>do you expect recording test or only tv watching?
>>
>regards
>tom
>>
>Am Samstag, den 18.11.2006, 14:43 -0200 schrieb Mauro Carvalho Chehab:
Em Sb, 2006-11-18 *s 13:21 +0100, tom escreveu:
hi,

i've hauppauge wintv go2 in side my box, working with cx88 driver.
i've cloned the cx88_ioctl2 directory yesterday.
everything worked.
recloned today cx88_iotcl2 and cloned cx88_improvements.
all working so far.
Thanks for you help, Tom.

do you expect some special test's? or dmesg outputs?

It would be interesing if you can send us the dmesg and test with more
than one app. Tvtime uses V4L2 api, while xawtv/xdtv uses the legacy
V4L1 API.

Have you tested also radio?

regards
tom

With cx88_improvements, scantv should also work, even for channels that
are not transmitting in stereo (here, only one open channel sometimes
transmit in stereo, so, with the older driver, no channel were detected
by scantv). To test cx88_improvements, it would be interesting to
modprobe tuner with debug=1 parameter, before loading cx88 modules. This
will generate some interesting debug messages that may help the test.

Cheers,
Mauro.

ok no prob, in zapping it's mostly working, i've sometimes trouble with
sound not in sync or it's sometimes crapy but i don't know if it's
zapping related or not (had this befor also with original edgy drivers).

so do you propose a test applic for recording? maybe mencoder?

mencoder tv:// -tv
-nosound -ovc lavc -o sample.avi

and what format should i use for testing?

regards
tom

Am Samstag, den 18.11.2006, 18:24 +0000 schrieb Mauro Carvalho Chehab:
Sat, 18 Nov 2006, tom wrote:

installed xawtv, also no prob so far.
do you expect recording test or only tv watching?

It would also be interesting to test recording.

Cheers,
Mauro Carvalho Chehab
http://linuxtv.org
mchehab (AT) infradead (DOT) org

i've a funny think with zapping, every time i start recording my line in
volume is resetted to 56, changes with alsamixer or sudo aslactl store
doens't work.
i don't know if this is zapping or driver related, i also don't know if
this also happen with original driver or originla repo zapping since
edgy repo zapping allways crash :-)

regards
tom

Am Samstag, den 18.11.2006, 18:24 +0000 schrieb Mauro Carvalho Chehab:
Sat, 18 Nov 2006, tom wrote:

installed xawtv, also no prob so far.
do you expect recording test or only tv watching?

It would also be interesting to test recording.

Cheers,
Mauro Carvalho Chehab
http://linuxtv.org
mchehab (AT) infradead (DOT) org

Em S, 2006-11-18 19:32 +0100, tom escreveu:
ok no prob, in zapping it's mostly working, i've sometimes trouble
with
sound not in sync or it's sometimes crapy but i don't know if it's
zapping related or not (had this befor also with original edgy
drivers).
Yes, this seems to be the normal behaviour.

so do you propose a test applic for recording? maybe mencoder?

mencoder tv:// -tv
-nosound -ovc lavc -o sample.avi
This should work.

Also, I've added a 'really small script' (tm) at v4l2-apps/util. There,
you will find other parameters for mencoder (using v4l2 api and also
recording audio) and for ffmpeg.

and what format should i use for testing?
Any output format should be ok.

Cheers,
Mauro.

Em S, 2006-11-18 19:40 +0100, tom escreveu:
i've a funny think with zapping, every time i start recording my line in
volume is resetted to 56, changes with alsamixer or sudo aslactl store
doens't work.
i don't know if this is zapping or driver related, i also don't know if
this also happen with original driver or originla repo zapping since
edgy repo zapping allways crash :-)

The original driver is available at http://linuxtv.org/hg/v4l-dvb. It
seems that this is zapping related. "Line in" volume is not related to
V4L api.

regards
tom

Am Samstag, den 18.11.2006, 18:24 +0000 schrieb Mauro Carvalho Chehab:
Sat, 18 Nov 2006, tom wrote:

installed xawtv, also no prob so far.
do you expect recording test or only tv watching?

It would also be interesting to test recording.

Cheers,
Mauro Carvalho Chehab
http://linuxtv.org
mchehab (AT) infradead (DOT) org

yep, tested with original drivers, seems to be zapping related, thanks

do some testing with mencoder, also do some test with your script

regards
tom

Am Samstag, den 18.11.2006, 16:45 -0200 schrieb Mauro Carvalho Chehab:
Em Sb, 2006-11-18 *s 19:40 +0100, tom escreveu:
i've a funny think with zapping, every time i start recording my line in
volume is resetted to 56, changes with alsamixer or sudo aslactl store
doens't work.
i don't know if this is zapping or driver related, i also don't know if
this also happen with original driver or originla repo zapping since
edgy repo zapping allways crash :-)

The original driver is available at http://linuxtv.org/hg/v4l-dvb. It
seems that this is zapping related. "Line in" volume is not related to
V4L api.

regards
tom

Am Samstag, den 18.11.2006, 18:24 +0000 schrieb Mauro Carvalho Chehab:
Sat, 18 Nov 2006, tom wrote:

installed xawtv, also no prob so far.
do you expect recording test or only tv watching?

It would also be interesting to test recording.

Cheers,
Mauro Carvalho Chehab
http://linuxtv.org
mchehab (AT) infradead (DOT) org

Em S, 2006-11-18 19:25 +0100, tom escreveu:
striiped my dmesg output just the relevant lines:

Seems ok to me ;)

[17179586.112000] cx2388x v4l2 driver version 0.0.6 loaded
[17179586.112000] ACPI: PCI Interrupt 0000:04:01.0[A] -GSI 19 (level,
low) -IRQ 217
[17179586.112000] CRE cx88[0]: subsystem: 0070:3401, board: Hauppauge
WinTV 34xxx models [card=1,autodetected]
[17179586.112000] TV tuner -1 at 0x1fe, Radio tuner -1 at 0x1fe
[17179586.136000] cx2388x alsa driver version 0.0.6 loaded
[17179586.152000] nvidia: module license 'NVIDIA' taints kernel.
[17179586.404000] Initializing USB Mass Storage driver
[17179586.404000] scsi2 : SCSI emulation for USB Mass Storage devices
[17179586.404000] usbcore: registered new driver usb-storage
[17179586.404000] USB Mass Storage support registered.
[17179586.420000] i2c-algo-bit.o: cx88[0] passed test.
[17179586.456000] tveeprom 0-0050: Hauppauge model 34704, rev J1B3,
Serial# 8521833
[17179586.456000] tveeprom 0-0050: tuner model is TCL M2523_3DB_E (idx
113, type 55)
[17179586.456000] tveeprom 0-0050: TV standards PAL(B/G) PAL(D/D1/K)
(eeprom 0x44)
[17179586.456000] tveeprom 0-0050: audio processor is CX881 (idx 31)
Nov 18 12:40:56 atlantis kernel: [17179586.456000] tveeprom 0-0050: has
radio
[17179586.456000] cx88[0]: warning: unknown hauppauge model #34704
Nov 18 12:40:56 atlantis kernel: [17179586.456000] cx88[0]: hauppauge
eeprom: model=34704
[17179586.456000] usbcore: registered new driver hiddev
[17179586.456000] input: cx88 IR (Hauppauge WinTV 34xxx
as /class/input/input2
[17179586.456000] cx88[0]/0: found at 0000:04:01.0, rev: 5, irq: 217,
latency: 32, mmio: 0xe5000000
[17179586.472000] input: Logitech USB-PS/2 Mouse
as /class/input/input3
[17179586.472000] input: USB HID v1.10 Mouse [Logitech USB-PS/2
Mouse] on usb-0000:00:1d.3-1
[17179586.472000] usbcore: registered new driver usbhid
[17179586.472000] drivers/usb/input/hid-core.c: v2.6:USB HID core driver
[17179586.500000] tuner 0-0061: chip found @ 0xc2 (cx88[0])
[17179586.500000] tuner 0-0061: type set to 55 (TCL 2002MB)
[17179586.512000] cx88[0]/0: registered device video0 [v4l2]
[17179586.512000] cx88[0]/0: registered device vbi0
[17179586.512000] cx88[0]/0: registered device radio0
[17179586.512000] ACPI: PCI Interrupt 0000:04:01.1[A] -GSI 19 (level,
low) -IRQ 217
[17179586.512000] cx88[0]/1: CX88x/0: ALSA support for cx2388x boards
[17179586.512000] ACPI: PCI Interrupt 0000:01:00.0[A] -GSI 16 (level,
low) -IRQ 169

regards
tom

Am Samstag, den 18.11.2006, 18:23 +0100 schrieb tom:
ok, attached output from /varlog/messages, lspci, lspci -n, lspci-v

regards
tom

Am Samstag, den 18.11.2006, 14:43 -0200 schrieb Mauro Carvalho Chehab:
Em S, 2006-11-18 13:21 +0100, tom escreveu:
hi,

i've hauppauge wintv go2 in side my box, working with cx88 driver.
i've cloned the cx88_ioctl2 directory yesterday.
everything worked.
recloned today cx88_iotcl2 and cloned cx88_improvements.
all working so far.
Thanks for you help, Tom.

do you expect some special test's? or dmesg outputs?

It would be interesing if you can send us the dmesg and test with more
than one app. Tvtime uses V4L2 api, while xawtv/xdtv uses the legacy
V4L1 API.

Have you tested also radio?

regards
tom

With cx88_improvements, scantv should also work, even for channels that
are not transmitting in stereo (here, only one open channel sometimes
transmit in stereo, so, with the older driver, no channel were detected
by scantv). To test cx88_improvements, it would be interesting to
modprobe tuner with debug=1 parameter, before loading cx88 modules. This
will generate some interesting debug messages that may help the test.

Cheers,
Mauro.

Hi,

Am Samstag, den 18.11.2006, 19:50 +0100 schrieb tom:
yep, tested with original drivers, seems to be zapping related, thanks

do some testing with mencoder, also do some test with your script

regards
tom

Am Samstag, den 18.11.2006, 16:45 -0200 schrieb Mauro Carvalho Chehab:
Em Sb, 2006-11-18 *s 19:40 +0100, tom escreveu:
i've a funny think with zapping, every time i start recording my line in
volume is resetted to 56, changes with alsamixer or sudo aslactl store
doens't work.
i don't know if this is zapping or driver related, i also don't know if
this also happen with original driver or originla repo zapping since
edgy repo zapping allways crash :-)

The original driver is available at http://linuxtv.org/hg/v4l-dvb. It
seems that this is zapping related. "Line in" volume is not related to
V4L api.

regards
tom

thanks for helping Mauro with testing.

Just some notes.
As far as I know, all WinTV G models don't have radio tuners.
You get a fake radio device registered, since the cx88 card=1 entry is
also used by cards with radio support. do you have an extra FM
antenna connector on the tuner? Don't think so.

Here is a TCL M2523-1/DB from this series.

You _are_ using a recent NVIDIA binary driver and your kernel is tainted.
The -nodga issue exists since about one year with these drivers when they
stopped to support VERLAY with direct access to the graphics card memory.
(there are several mobo chipsets out, which don't play nice with it)

mplayer/mencoder driver=v4l2 might be better, except for the case
Mauro wants to have the v4l1-compat layer tested too.

Cheers,
Hermann

Am Samstag, den 18.11.2006, 18:24 +0000 schrieb Mauro Carvalho Chehab:
Sat, 18 Nov 2006, tom wrote:

installed xawtv, also no prob so far.
do you expect recording test or only tv watching?

It would also be interesting to test recording.

Cheers,
Mauro Carvalho Chehab
http://linuxtv.org
mchehab (AT) infradead (DOT) org

Em S, 2006-11-18 22:04 +0100, hermann pitton escreveu:
Hi,
Just some notes.
As far as I know, all WinTV G models don't have radio tuners.
You get a fake radio device registered, since the cx88 card=1 entry is
also used by cards with radio support. do you have an extra FM
antenna connector on the tuner? Don't think so.

It seems that some models have FM:

mplayer/mencoder driver=v4l2 might be better, except for the case
Mauro wants to have the v4l1-compat layer tested too.
Both tests are interesting, since v4l1-compat calls several v4l2 ioctls
for some api calls. The script I've pointed uses driver=v4l2.

Thanks for helping me with those tests!

Cheers,
Mauro.

hi Hermann,

since im a user of a cx88 card im willing to help as far as i can so no
need to say thank you.

hmm i think you're right, i don't have a extra fm antenna connector,
must be a fake device

ok i also could use v4l2 as driver for mencoder/mplayer, mauro any
suggestions about witch driver i should use for recording test's?

regards
tom

btw i've to say thank you since i'm not a developer

Am Samstag, den 18.11.2006, 22:04 +0100 schrieb hermann pitton:
Hi,

Am Samstag, den 18.11.2006, 19:50 +0100 schrieb tom:
yep, tested with original drivers, seems to be zapping related, thanks

do some testing with mencoder, also do some test with your script

regards
tom

Am Samstag, den 18.11.2006, 16:45 -0200 schrieb Mauro Carvalho Chehab:
Em Sb, 2006-11-18 *s 19:40 +0100, tom escreveu:
i've a funny think with zapping, every time i start recording my line in
volume is resetted to 56, changes with alsamixer or sudo aslactl store
doens't work.
i don't know if this is zapping or driver related, i also don't know if
this also happen with original driver or originla repo zapping since
edgy repo zapping allways crash :-)

The original driver is available at http://linuxtv.org/hg/v4l-dvb. It
seems that this is zapping related. "Line in" volume is not related to
V4L api.

regards
tom

thanks for helping Mauro with testing.

Just some notes.
As far as I know, all WinTV G models don't have radio tuners.
You get a fake radio device registered, since the cx88 card=1 entry is
also used by cards with radio support. do you have an extra FM
antenna connector on the tuner? Don't think so.

Here is a TCL M2523-1/DB from this series.

You _are_ using a recent NVIDIA binary driver and your kernel is tainted.
The -nodga issue exists since about one year with these drivers when they
stopped to support VERLAY with direct access to the graphics card memory.
(there are several mobo chipsets out, which don't play nice with it)

mplayer/mencoder driver=v4l2 might be better, except for the case
Mauro wants to have the v4l1-compat layer tested too.

Cheers,
Hermann

Am Samstag, den 18.11.2006, 18:24 +0000 schrieb Mauro Carvalho Chehab:
Sat, 18 Nov 2006, tom wrote:

installed xawtv, also no prob so far.
do you expect recording test or only tv watching?

It would also be interesting to test recording.

Cheers,
Mauro Carvalho Chehab
http://linuxtv.org
mchehab (AT) infradead (DOT) org

yes hard to say, i haven't any booklet, card was assembled by a no name
manufacture here in switzerland but on my bill is a Hauppauge WinTV Go2
(TV/Radio) but only with 1 Antenna input
the card also have connectors to line in/line out and remote

Am Samstag, den 18.11.2006, 19:09 -0200 schrieb Mauro Carvalho Chehab:
Em Sb, 2006-11-18 *s 22:04 +0100, hermann pitton escreveu:
Hi,
Just some notes.
As far as I know, all WinTV G models don't have radio tuners.
You get a fake radio device registered, since the cx88 card=1 entry is
also used by cards with radio support. do you have an extra FM
antenna connector on the tuner? Don't think so.

It seems that some models have FM:

mplayer/mencoder driver=v4l2 might be better, except for the case
Mauro wants to have the v4l1-compat layer tested too.
Both tests are interesting, since v4l1-compat calls several v4l2 ioctls
for some api calls. The script I've pointed uses driver=v4l2.

Thanks for helping me with those tests!

Cheers,
Mauro.

Mauro,

that's exactly my card i've inside my box

thnaks
tom

Am Samstag, den 18.11.2006, 19:09 -0200 schrieb Mauro Carvalho Chehab:
Em Sb, 2006-11-18 *s 22:04 +0100, hermann pitton escreveu:
Hi,
Just some notes.
As far as I know, all WinTV G models don't have radio tuners.
You get a fake radio device registered, since the cx88 card=1 entry is
also used by cards with radio support. do you have an extra FM
antenna connector on the tuner? Don't think so.

It seems that some models have FM:

mplayer/mencoder driver=v4l2 might be better, except for the case
Mauro wants to have the v4l1-compat layer tested too.
Both tests are interesting, since v4l1-compat calls several v4l2 ioctls
for some api calls. The script I've pointed uses driver=v4l2.

Thanks for helping me with those tests!

Cheers,
Mauro.

Am Samstag, den 18.11.2006, 19:09 -0200 schrieb Mauro Carvalho Chehab:
Em Sb, 2006-11-18 *s 22:04 +0100, hermann pitton escreveu:
Hi,
Just some notes.
As far as I know, all WinTV G models don't have radio tuners.
You get a fake radio device registered, since the cx88 card=1 entry is
also used by cards with radio support. do you have an extra FM
antenna connector on the tuner? Don't think so.

It seems that some models have FM:

Hi Mauro,

Hauppauge never announces FM on them so far.

See the customer review below on your link ;)
-quote-
First, this card does not have FM radio. This looks like it could be
Amazon's fault, although the delivery note says it is the right model.

Cheers,
Hermann

Em S, 2006-11-18 22:20 +0100, hermann pitton escreveu:
Am Samstag, den 18.11.2006, 19:09 -0200 schrieb Mauro Carvalho Chehab:
Em S, 2006-11-18 22:04 +0100, hermann pitton escreveu:
Hi,
Just some notes.
As far as I know, all WinTV G models don't have radio tuners.
You get a fake radio device registered, since the cx88 card=1 entry is
also used by cards with radio support. do you have an extra FM
antenna connector on the tuner? Don't think so.

It seems that some models have FM:

Hi Mauro,

Hauppauge never announces FM on them so far.

See the customer review below on your link ;)
Yes weird to announce it as FM ;)

Cheers,
Mauro.

Mauro Carvalho Chehab wrote:
Em Sex, 2006-11-17 12:13 -0200, Mauro Carvalho Chehab escreveu:
>Guys,
>>
>As suggested, I've redesigned cx88 driver to work with video_ioctl2. The
>newer tree is available at:
>>
>http://linuxtv.org/hg/~mchehab/cx88_ioctl2
>>
I've added another branch there, at:
http://linuxtv.org/hg/~mchehab/cx88_improvements

The basic difference is a patch to make cx88 to use tuner signal level,
instead of using the stereo carrier as signal strength indicator.

This patch requires more testing and may benefit other designs where the
video decoder is not capable of measuring signal quality.

P.S.: Previously, this patch were at cx88_ioctl2, since I used the newer
cx88 design for those tests. So, for those people who cloned
cx88_ioctl2, they should re-clone it to test just cx88 conversion to
video_ioctl2.

About the conversion: the only expected effect is the memory size
reduction.

We really want testers here for cx88-based boards and cx88-blackbird
designs on cx88_ioctl2 tree.

Can we have community feedback about this?

Mauro,

I have reviewed and tested your changes. This seems to be a great improvement
to me! I absolutely love how those huge foo_do_ioctl switchcase blacks have
been replaced by a struct of function pointers for the new video_ioctl2.

However, there seems to be a bug in the cx88-blackbird mpeg encoder changes.
For some reason, the tvnorm autodetection has been broken for NTSC :-(

I have seen this problem in the past Perhaps a regression was caused
somewhere. I will spend some more time digging here and see if I can fix this.

, I really like this change, but I'd like to fix this tvnorm detection
problem in the mpeg encoder before we merge this in.

Regards,

Mike

See the customer review below on your link ;)
Yes weird to announce it as FM ;)
In trust we stay.
Hermann

Em S, 2006-11-18 19:16 -0500, Michael Krufky escreveu:
Mauro Carvalho Chehab wrote:

I have reviewed and tested your changes. This seems to be a great improvement
to me! I absolutely love how those huge foo_do_ioctl switchcase blacks have
been replaced by a struct of function pointers for the new video_ioctl2.
Yes. The previous code were really ugly ;)

However, there seems to be a bug in the cx88-blackbird mpeg encoder changes.
For some reason, the tvnorm autodetection has been broken for NTSC :-(
Can you send me the dmesg with debug turned on? Probably, it would be
easy to fix. Now, the code becamed much more simpler than before.

a quick review, I noticed that the current norm is missing at
video_device array on cx88-blackbird. Try to add this just
bellow .tvnorms there:
.current_norm = V4L2_STD_PAL_BG,
(this is the same default I'm using for cx88-video).

I have seen this problem in the past Perhaps a regression was caused
somewhere. I will spend some more time digging here and see if I can fix this.

, I really like this change, but I'd like to fix this tvnorm detection
problem in the mpeg encoder before we merge this in.
Yes.

Regards,

Mike
Cheers,
Mauro.

Em S, 2006-11-18 23:05 -0200, Mauro Carvalho Chehab escreveu:
However, there seems to be a bug in the cx88-blackbird mpeg encoder changes.
For some reason, the tvnorm autodetection has been broken for NTSC :-(
Can you send me the dmesg with debug turned on? Probably, it would be
easy to fix. Now, the code becamed much more simpler than before.

a quick review, I noticed that the current norm is missing at
video_device array on cx88-blackbird. Try to add this just
bellow .tvnorms there:
.current_norm = V4L2_STD_PAL_BG,
(this is the same default I'm using for cx88-video).

, I've added this small patch to the tree.

Cheers,
Mauro.

Hi Mauro,

tested recording with zapping, mencoder (v4l and v4l2) with mpeg and
avi.
till now no problems with the new cx88 drivers.

regards
tom

Am Samstag, den 18.11.2006, 18:24 +0000 schrieb Mauro Carvalho Chehab:
Sat, 18 Nov 2006, tom wrote:

installed xawtv, also no prob so far.
do you expect recording test or only tv watching?

It would also be interesting to test recording.

Cheers,
Mauro Carvalho Chehab
http://linuxtv.org
mchehab (AT) infradead (DOT) org