BTTV broken on recent kernels

BTTV broken on recent kernels

[Bob Gill]

Hi.  The last several kernel versions have led to broken bttv (up to 4 
or 5 kernel versions ago, I could watch tv on either mplayer or xawtv), 
but lately bttv is broken.  My card is an 'bt878 compatible built by ATI 
(ATI TV Wonder VE).  I'm pretty certain it worked as late as 
2.6.14-git7.  I've peeked around /Changes and didn't see anything.   I'm 
using the same build script as before, and a piece of lsmod shows
serial_core            14848  1 8250
rtc                     9524  0
tuner                  36908  0
bttv                  148564  0
video_buf              15748  1 bttv
compat_ioctl32          1152  1 bttv
i2c_algo_bit            7432  1 bttv
v4l2_common             6528  2 tuner,bttv
btcx_risc               3720  1 bttv
ir_common               7812  1 bttv
tveeprom               12304  1 bttv
i2c_core               14864  4 tuner,bttv,i2c_algo_bit,tveeprom
videodev                6912  1 bttv
snd_emu10k1            94628  2 snd_emu10k1_synth
..........also, a chunk of lspci shows:
0000:00:09.1 Input device controller: Creative Labs SB Live! MIDI/Game 
Port (rev 07)
0000:00:0b.0 Multimedia video controller: Brooktree Corporation Bt878 
Video Capture (rev 02)
0000:00:0b.1 Multimedia controller: Brooktree Corporation Bt878 Audio 
Capture (rev 02)
....it's just that I get a blank (screen blanking due to no signal) 
screen when I start a tv application.  I can try to change channels/tune 
frequencies, and it looks like the applications are trying, but nothing 
gets tuned in.   To be fair, I must mention that I *ahem* taint the 
kernel with Nvidia stuff, and recently upgraded gcc (although it has 
always worked well with tainted kernel, and it broke before I upgraded 
gcc (to gcc version 4.0.2) on Debian Sarge.  If you *really* want, I can 
revert XF86Config to use non-nvidia drivers (and revert back to the old 
version of gcc) and give a bug report from that, but I suspect things 
will remain broken.  Mplayer compiles very well with the new version of 
gcc, and the new kernel (buit with the new version of gcc) does 
everything else (sound, firewire, cd/dvd/networking, disk I/O etc.) 
without problems.
Thanks in advance,
Bob

Re: BTTV broken on recent kernels

[Lee Revell]

On Sun, 2006-01-15 at 19:05 -0700, Bob Gill wrote:
> Hi.  The last several kernel versions have led to broken bttv (up to 4 
> or 5 kernel versions ago, I could watch tv on either mplayer or xawtv), 
> but lately bttv is broken.  My card is an 'bt878 compatible built by ATI 
> (ATI TV Wonder VE).  I'm pretty certain it worked as late as 
> 2.6.14-git7.

You need to do a git bisect (check the list archives) to isolate the
change that broke it.

Lee

Re: BTTV broken on recent kernels

[Bob Gill]

Lee Revell wrote:

>On Sun, 2006-01-15 at 19:05 -0700, Bob Gill wrote:
>  
>
>>Hi.  The last several kernel versions have led to broken bttv (up to 4 
>>or 5 kernel versions ago, I could watch tv on either mplayer or xawtv), 
>>but lately bttv is broken.  My card is an 'bt878 compatible built by ATI 
>>(ATI TV Wonder VE).  I'm pretty certain it worked as late as 
>>2.6.14-git7.
>>    
>>
>
>You need to do a git bisect (check the list archives) to isolate the
>change that broke it.
>
>Lee
>
>
>  
>
OK.  I will, but I will add a tiny bit of additional info first.
dmesg shows bttv loading up like this:
Linux video capture interface: v1.00
bttv: driver version 0.9.16 loaded
bttv: using 8 buffers with 2080k (520 pages) each for capture
bttv: Bt8xx card found (0).
ACPI: PCI Interrupt Link [LNKD] enabled at IRQ 5
PCI: setting IRQ 5 as level-triggered
ACPI: PCI Interrupt 0000:00:0b.0[A] -> Link [LNKD] -> GSI 5 (level, low) 
-> IRQ
5
bttv0: Bt878 (rev 2) at 0000:00:0b.0, irq: 5, latency: 32, mmio: 0xeb003000
bttv0: detected: ATI TV Wonder/VE [card=64], PCI subsystem ID is 1002:0003
bttv0: using: ATI TV-Wonder VE [card=64,autodetected]
bttv0: gpio: en=00000000, out=00000000 in=00ffffff [init]
bttv0: using tuner=19
bttv0: i2c: checking for TDA9875 @ 0xb0... not found
bttv0: i2c: checking for TDA7432 @ 0x8a... not found
bttv0: i2c: checking for TDA9887 @ 0x86... not found
tuner 0-0060: All bytes are equal. It is not a TEA5767
tuner 0-0060: chip found @ 0xc0 (bt878 #0 [sw])
tuner 0-0060: type set to 19 (Temic PAL* auto (4006 FN5))
bttv0: registered device video0
bttv0: registered device vbi0
bttv0: PLL: 28636363 => 35468950 .. ok
Real Time Clock Driver v1.12ac
Serial: 8250/16550 driver $Revision: 1.90 $ 4 ports, IRQ sharing disabled
.....and when trying to tune audio comes in, but dmsg shows video giving:
agpgart: Putting AGP V2 device at 0000:01:00.0 into 4x mode
bttv0: PLL can sleep, using XTAL (28636363).
bttv0: PLL: 28636363 => 35468950 .. ok
bttv0: PLL can sleep, using XTAL (28636363).
bttv0: PLL: 28636363 => 35468950 .. ok
bttv0: PLL can sleep, using XTAL (28636363).
bttv0: SCERR @ 348ec014,bits: HSYNC OFLOW FBUS SCERR*
bttv0: SCERR @ 348ec000,bits: OFLOW FBUS SCERR*
.... I will do the git bisect and reply in a bit.
Thanks for your reply,
Bob

Re: BTTV broken on recent kernels

[Alistair John Strachan]

On Monday 16 January 2006 02:05, Bob Gill wrote:
> Hi.  The last several kernel versions have led to broken bttv (up to 4
> or 5 kernel versions ago, I could watch tv on either mplayer or xawtv),
> but lately bttv is broken.  My card is an 'bt878 compatible built by ATI
> (ATI TV Wonder VE).  I'm pretty certain it worked as late as
> 2.6.14-git7.  I've peeked around /Changes and didn't see anything.   I'm
> using the same build script as before, and a piece of lsmod shows
> serial_core            14848  1 8250
> rtc                     9524  0
> tuner                  36908  0
> bttv                  148564  0
> video_buf              15748  1 bttv
> compat_ioctl32          1152  1 bttv
> i2c_algo_bit            7432  1 bttv
> v4l2_common             6528  2 tuner,bttv
> btcx_risc               3720  1 bttv
> ir_common               7812  1 bttv
> tveeprom               12304  1 bttv
> i2c_core               14864  4 tuner,bttv,i2c_algo_bit,tveeprom
> videodev                6912  1 bttv
> snd_emu10k1            94628  2 snd_emu10k1_synth
> ..........also, a chunk of lspci shows:
> 0000:00:09.1 Input device controller: Creative Labs SB Live! MIDI/Game
> Port (rev 07)
> 0000:00:0b.0 Multimedia video controller: Brooktree Corporation Bt878
> Video Capture (rev 02)
> 0000:00:0b.1 Multimedia controller: Brooktree Corporation Bt878 Audio
> Capture (rev 02)
> ....it's just that I get a blank (screen blanking due to no signal)
> screen when I start a tv application.  I can try to change channels/tune
> frequencies, and it looks like the applications are trying, but nothing
> gets tuned in.   To be fair, I must mention that I *ahem* taint the
> kernel with Nvidia stuff, and recently upgraded gcc (although it has
> always worked well with tainted kernel, and it broke before I upgraded
> gcc (to gcc version 4.0.2) on Debian Sarge. 

This problem sounds suspiciously like an overlay bug, known in the binary 
NVIDIA driver. Please try changing it to "nv" in XF86Config, then restarting 
your TV application..

> If you *really* want, I can 
> revert XF86Config to use non-nvidia drivers (and revert back to the old
> version of gcc) and give a bug report from that, but I suspect things
> will remain broken.  Mplayer compiles very well with the new version of
> gcc, and the new kernel (buit with the new version of gcc) does
> everything else (sound, firewire, cd/dvd/networking, disk I/O etc.)
> without problems.
> Thanks in advance,
> Bob

-- 
Cheers,
Alistair.

'No sense being pessimistic, it probably wouldn't work anyway.'
Third year Computer Science undergraduate.
1F2 55 South Clerk Street, Edinburgh, UK.

Re: BTTV broken on recent kernels

[Bob Gill]

Alistair John Strachan wrote:

>On Monday 16 January 2006 02:05, Bob Gill wrote:
>  
>
>>Hi.  The last several kernel versions have led to broken bttv (up to 4
>>or 5 kernel versions ago, I could watch tv on either mplayer or xawtv),
>>but lately bttv is broken.  My card is an 'bt878 compatible built by ATI
>>(ATI TV Wonder VE).  I'm pretty certain it worked as late as
>>2.6.14-git7.  I've peeked around /Changes and didn't see anything.   I'm
>>using the same build script as before, and a piece of lsmod shows
>>serial_core            14848  1 8250
>>rtc                     9524  0
>>tuner                  36908  0
>>bttv                  148564  0
>>video_buf              15748  1 bttv
>>compat_ioctl32          1152  1 bttv
>>i2c_algo_bit            7432  1 bttv
>>v4l2_common             6528  2 tuner,bttv
>>btcx_risc               3720  1 bttv
>>ir_common               7812  1 bttv
>>tveeprom               12304  1 bttv
>>i2c_core               14864  4 tuner,bttv,i2c_algo_bit,tveeprom
>>videodev                6912  1 bttv
>>snd_emu10k1            94628  2 snd_emu10k1_synth
>>..........also, a chunk of lspci shows:
>>0000:00:09.1 Input device controller: Creative Labs SB Live! MIDI/Game
>>Port (rev 07)
>>0000:00:0b.0 Multimedia video controller: Brooktree Corporation Bt878
>>Video Capture (rev 02)
>>0000:00:0b.1 Multimedia controller: Brooktree Corporation Bt878 Audio
>>Capture (rev 02)
>>....it's just that I get a blank (screen blanking due to no signal)
>>screen when I start a tv application.  I can try to change channels/tune
>>frequencies, and it looks like the applications are trying, but nothing
>>gets tuned in.   To be fair, I must mention that I *ahem* taint the
>>kernel with Nvidia stuff, and recently upgraded gcc (although it has
>>always worked well with tainted kernel, and it broke before I upgraded
>>gcc (to gcc version 4.0.2) on Debian Sarge. 
>>    
>>
>
>This problem sounds suspiciously like an overlay bug, known in the binary 
>NVIDIA driver. Please try changing it to "nv" in XF86Config, then restarting 
>your TV application..
>
>  
>
>>If you *really* want, I can 
>>revert XF86Config to use non-nvidia drivers (and revert back to the old
>>version of gcc) and give a bug report from that, but I suspect things
>>will remain broken.  Mplayer compiles very well with the new version of
>>gcc, and the new kernel (buit with the new version of gcc) does
>>everything else (sound, firewire, cd/dvd/networking, disk I/O etc.)
>>without problems.
>>Thanks in advance,
>>Bob
>>    
>>
>
>  
>
Alistair, Lee: thank you for your replies.  Sadly the bug is in the 
Nvidia binary.  Changing the driver to nv from nvidia did resolve the 
issue.  Now I have to decide whether I like accelerated 3d more, or 
using the tv tuner more (till Nvidia fixes the driver).  Thank you for 
your time.  I won't waste any more LKML bandwidth on this (Nvidia bug).  
Thanks again.
Bob

Re: BTTV broken on recent kernels

[Michael Krufky]

Bob Gill wrote:

> Alistair John Strachan wrote:
>
>> On Monday 16 January 2006 02:05, Bob Gill wrote:
>>  
>>
>>> Hi.  The last several kernel versions have led to broken bttv (up to 4
>>> or 5 kernel versions ago, I could watch tv on either mplayer or xawtv),
>>> but lately bttv is broken.  My card is an 'bt878 compatible built by 
>>> ATI
>>> (ATI TV Wonder VE).  I'm pretty certain it worked as late as
>>> 2.6.14-git7.  I've peeked around /Changes and didn't see anything.   
>>> I'm
>>> using the same build script as before, and a piece of lsmod shows
>>> serial_core            14848  1 8250
>>> rtc                     9524  0
>>> tuner                  36908  0
>>> bttv                  148564  0
>>> video_buf              15748  1 bttv
>>> compat_ioctl32          1152  1 bttv
>>> i2c_algo_bit            7432  1 bttv
>>> v4l2_common             6528  2 tuner,bttv
>>> btcx_risc               3720  1 bttv
>>> ir_common               7812  1 bttv
>>> tveeprom               12304  1 bttv
>>> i2c_core               14864  4 tuner,bttv,i2c_algo_bit,tveeprom
>>> videodev                6912  1 bttv
>>> snd_emu10k1            94628  2 snd_emu10k1_synth
>>> ..........also, a chunk of lspci shows:
>>> 0000:00:09.1 Input device controller: Creative Labs SB Live! MIDI/Game
>>> Port (rev 07)
>>> 0000:00:0b.0 Multimedia video controller: Brooktree Corporation Bt878
>>> Video Capture (rev 02)
>>> 0000:00:0b.1 Multimedia controller: Brooktree Corporation Bt878 Audio
>>> Capture (rev 02)
>>> ....it's just that I get a blank (screen blanking due to no signal)
>>> screen when I start a tv application.  I can try to change 
>>> channels/tune
>>> frequencies, and it looks like the applications are trying, but nothing
>>> gets tuned in.   To be fair, I must mention that I *ahem* taint the
>>> kernel with Nvidia stuff, and recently upgraded gcc (although it has
>>> always worked well with tainted kernel, and it broke before I upgraded
>>> gcc (to gcc version 4.0.2) on Debian Sarge.   
>>
>>
>> This problem sounds suspiciously like an overlay bug, known in the 
>> binary NVIDIA driver. Please try changing it to "nv" in XF86Config, 
>> then restarting your TV application..
>>
>>  
>>
>>> If you *really* want, I can revert XF86Config to use non-nvidia 
>>> drivers (and revert back to the old
>>> version of gcc) and give a bug report from that, but I suspect things
>>> will remain broken.  Mplayer compiles very well with the new version of
>>> gcc, and the new kernel (buit with the new version of gcc) does
>>> everything else (sound, firewire, cd/dvd/networking, disk I/O etc.)
>>> without problems.
>>> Thanks in advance,
>>> Bob
>>>   
>>
>>
>>  
>>
> Alistair, Lee: thank you for your replies.  Sadly the bug is in the 
> Nvidia binary.  Changing the driver to nv from nvidia did resolve the 
> issue.  Now I have to decide whether I like accelerated 3d more, or 
> using the tv tuner more (till Nvidia fixes the driver).  Thank you for 
> your time.  I won't waste any more LKML bandwidth on this (Nvidia 
> bug).  Thanks again.
> Bob 

Not that I would recommend the binary nvidia driver, but:

You can probably get bttv to work despite this overlay bug by using 
"grabdisplay" mode instead.  There is an option for it in xawtv.  The 
difference is, in overlay mode, the capture card is writing directly to 
videoram.  Using grabdisplay mode eliminates this optimization, and 
would be a decent workaround for you.

Hope this helps,

Michael

-- 
Michael Krufky

Re: BTTV broken on recent kernels

[Bob Gill]

Michael Krufky wrote:

>
>> Alistair, Lee: thank you for your replies.  Sadly the bug is in the 
>> Nvidia binary.  Changing the driver to nv from nvidia did resolve the 
>> issue.  Now I have to decide whether I like accelerated 3d more, or 
>> using the tv tuner more (till Nvidia fixes the driver).  Thank you 
>> for your time.  I won't waste any more LKML bandwidth on this (Nvidia 
>> bug).  Thanks again.
>> Bob 
>
>
> Not that I would recommend the binary nvidia driver, but:
>
> You can probably get bttv to work despite this overlay bug by using 
> "grabdisplay" mode instead.  There is an option for it in xawtv.  The 
> difference is, in overlay mode, the capture card is writing directly 
> to videoram.  Using grabdisplay mode eliminates this optimization, and 
> would be a decent workaround for you.
>
> Hope this helps,
>
> Michael
>
I haven't tried using nv with grabdisplay instead of overlay, but I have 
recently run a large (nvidia...and no, I don't work for them) patch 
against the 81.78 driver (about 5 small patches all in one) and it works 
successfully on 2.6.16-rc1-git3 (in both grabdisplay and overlay 
modes).  The only thing I noted is that the old nvidia binaries wanted a 
channel frequency offset of 108, but with the (heavily patched) 81.78 
driver, you don't need an channel frequency offset (both xawtv and 
mplayer).  Your test is simple to try though, and if it doesn't work, 
I'll reply and mention it.   Thanks for your reply.
Bob