Re: [Alsa-user] bt87x audio recording - broke between 2.6.14 and 2.6.17.4

[Takashi I. <ti...@su...>]

At Sun, 10 Sep 2006 20:04:30 -0400,
Dan Pritts wrote:
> 
> Hi -
> 
> I have a Centos 4.3 system with a wintv pci card with FM tuner:
> 
> 00:0b.0 Multimedia video controller: Brooktree Corporation Bt878 Video Capture (rev 02)
> 00:0b.1 Multimedia controller: Brooktree Corporation Bt878 Audio Capture (rev 02)
> 
> As of kernel 2.6.14, I was able to use this card to record FM radio.
> 
> Updating to 2.6.17.4, it broke (but I didn't notice for a few weeks until
> I went to listen to recordings).  Updating to 2.6.17.13 does not fix it.
> 
> All I get out of the card now is soft static.
> 
> I have checked with alsamixer and the card is not muted.  
> 
> I can reboot with the old kernel and it works fine.  
> 
> Here's my /etc/modprobe.conf (nothing relevant):
> 
>   alias eth0 e100
>   alias char-major-89 i2c-dev
>   alias eth1 e100
>   alias char-major-81 bttv
> 
> In case it's not obvious, I have a kernel.org kernel, not the
> centos/redhat kernel, on this box.
> 
> The video4linux interface to this card appears to be working - I can use
> the FM tuner applications to control the tuner, and it indicates signal
> quality, stereo (or not), etc., as i would expect.
> 
> I use this to do the recording:
> 
>   fm 89.1 33000 
>   sox -t ossdsp -c 2 -r 44100 /dev/dsp1 -t wav
> 
> I just tried today to get it working with arecord, which I had never
> used before.  I wasn't successful even with 2.6.14 but I didn't try
> very hard to divine the appropriate syntax.
> 
> I diffed the bt87x.c files between the kernel versions and noticed lots
> of changes, mostly calling "struct" explicitly and new hardware
> sense info.  However, I did notice:
> 
>   < static int snd_bt87x_capture_source_info(struct snd_kcontrol *kcontrol,
>   <                                        struct snd_ctl_elem_info *info)
> 
>   > static int snd_bt87x_capture_source_info(snd_kcontrol_t *kcontrol, snd_ctl_elem_info_t *info)
> 
> snd_kcontrol replaces snd_kcontrol_t.  Looks like there are no references
> to snd_kcontrol_t anywhere in the new kernel's source tree so this is
> probably a simple global change but i thought i'd mention it.
> 
> help!

You need to check whether it's a regression in ALSA part or not.
One option is to try the later alsa-driver package on 2.6.14 kernel.
If it works, the problem is in other parts.  If not, you can try to
earlier version of ALSA package in a bisect manner.


Takashi