audioconvert: handle odd writes in delay

Add some padding to the delay buffer. If we wrap around, copy the
spilled samples to the front of the buffer. This makes it possible to
use the more optimized sse delay function in more cases.

spa/plugins/audioconvert/channelmix-ops.h

@@ -59,7 +59,7 @@ struct channelmix {
 	uint32_t hilbert_taps;				/* to phase shift, 0 disabled */
 	struct lr4 lr4[SPA_AUDIO_MAX_CHANNELS];
 
-	float buffer[2][BUFFER_SIZE*2];
+	float buffer[2][BUFFER_SIZE*2 + 16];
 	uint32_t pos[2];
 	uint32_t delay;
 	float taps[MAX_TAPS];