Re: [alsa-devel] [PATCH 2/2] ASoC: DAPM - Make sure DAPM widget IO ops hold the component mutex.

Mark Brown wrote:

I wonder if PowerPC mutexes are less forgiving here?

I'm still debugging this (although I have higher priority tasks to worry about, unfortunately), but it appears that the mutex is already taken when that code is executed. Unfortunately, I have been unable to figure out where. One second the 'count' value is 1, and then it suddenly becomes '0', and then mutex_lock() is called.