Remove the usage of TPCircularBuffer and multiple atomic variables that start
to make this code way too complicated. Replace it by os_unfair_lock and a block
chain.

os_unfair_lock is a safe spinlock that waits in the kernel in case of thread
contention.

Fallback to pthread_mutex_t if os_unfair_lock is not availaible (before macOS
10.12 / iOS 10.0).

The unfairness of this new lock is not an issue here since both locking threads
(the render callback and the VLC DecoderThread calling aout_DecPlay) will be
automatically paced (and will let the other thread take the lock). Indeed, the
render thread need a sample every 22 or 88ms, and the DecoderThread will wait
for the decoder, wait in the decoder lock, or wait from the aout if the FIFO is
full.