GTA02 ASoC capture kernel oops

[laforge]

When trying to use the GTA02 to capture PCM data, the stable-tracking kernel oopses.

The following steps can be used to reproduce this:

$ wget ​http://wildsau.enemy.org/~moko/voice-recording.state
$ alsactl -f voice-recording.state restore
$ arecord -v -f cd -c 2 foo.wav
Segmentation fault

I will attach the kernel oops

[laforge]

dmesg output of asoc kernel oops

[laforge]

the voice-recording.state file used

[laforge]

there seem to be actually two independent problems

1) the fact that there are no matching rates " WM8753 HiFi? <-> s3c24xx-i2s No matching rates"

2) the fact that this causes a kernel oops in the
neo1973_gta02_hifi_hw_free() function. This is caused by dai_ops.set_pll() == NULL, which in turn seems to be caused by the fact that snd_pcm_open_substream() is not finished (due to the 'no matching rate' bug above) and calls snd_pcm_release_substream() for a not fully opened substream.

[erl]

I think this is a dupe of bug #2073.

It would be great if someone could raise the priority of this bug. It would be useful to debug microphone problems, hack VoIP, speech recognition and general audio recording stuff.

[andy]

I pinged Graeme about it yesterday if he's not able to look at it somebody else will.

[marek]

duplicate of #2073

[joerg]

made 2073 a dup of this

[joerg]

an oops. closed again. sorry for the noise