Comment 24 for bug 348353

pl bossart in Pulseaudio email list discussion said:
"
NAK. This does not work for me. You need to rewind also in the thread context:

diff --git a/src/modules/alsa/alsa-sink.c b/src/modules/alsa/alsa-sink.c
index 2995c3c..4059a9b 100644
--- a/src/modules/alsa/alsa-sink.c
+++ b/src/modules/alsa/alsa-sink.c
@@ -953,7 +953,12 @@ static int update_sw_params(struct userdata *u) {
    }

    pa_sink_set_max_request_within_thread(u->sink, u->hwbuf_size -
u->hwbuf_unused);
- pa_sink_set_max_rewind_within_thread(u->sink, u->hwbuf_size);
+ if (pa_alsa_pcm_is_hw(u->pcm_handle))
+ pa_sink_set_max_rewind_within_thread(u->sink, u->hwbuf_size);
+ else {
+ pa_log("Disabling rewind_within_thread for device %s", u->device_name);
+ pa_sink_set_max_rewind_within_thread(u->sink, 0);
+ }

    return 0;
}
@@ -1906,7 +1911,12 @@ pa_sink *pa_alsa_sink_new(pa_module *m,
pa_modargs *ma, const char*driver, pa_ca
                (double) pa_bytes_to_usec(u->hwbuf_size, &ss) /
PA_USEC_PER_MSEC);

    pa_sink_set_max_request(u->sink, u->hwbuf_size);
- pa_sink_set_max_rewind(u->sink, u->hwbuf_size);
+ if (pa_alsa_pcm_is_hw(u->pcm_handle))
+ pa_sink_set_max_rewind(u->sink, u->hwbuf_size);
+ else {
+ pa_log("Disabling rewind for device %s", u->device_name);
+ pa_sink_set_max_rewind(u->sink, 0);
+ }

    if (u->use_tsched) {
        u->tsched_watermark =
pa_usec_to_bytes_round_up(pa_bytes_to_usec_round_up(tsched_watermark,
&requested_ss), &u->sink->sample_spec);

pa_pcm_is_hw detects a52 but allows rewind for iec958. Looks like a
better fix than a hard-coded string compare.
"

I didn't have a line with pa_sink_set_max_rewind_within_thread but I added it anyway with no ill effect. Probably my Pulseaudio is older.

Danny, can you update the patch with pl bossart's code?