From: Jyri Sarha <jyri.sarha@xxxxxxxxx>
These are all pretty obvious fixes, but maybe the "core: Use
volume_change_safety_margin when rewinding sync-volume events"
A volume change always triggers a rewinding on the sink. This is
sometimes a problem when the flat-volume is turning HW volume down and
compensating with SW volume. The rewind gets the samples with the new
SW-volume very close to HW-pointer and we may not have enough headroom
to process rewinding on the streams and monitor source before the
first samples are played out.
With the patch I first of all rewind the volume changes differently
depending on whether volume is going up or down. The volume events are
rewound as soon as possible after DMA-buffer rewinding and immediately
after it there is a check if we should write something to HW.
Cheers,
Jyri
Jyri Sarha (5):
core: Change sematics of pa_flist_new_with_name()
core: Use volume_change_safety_margin when rewinding sync-volume
events
core: Use pa_sink_get_latency_within_thread() in sync-volume code
alsa-sink: Fix double use of string
alsa-sink: Don't assume we were able to enable hw-volume or
sync-volume
src/modules/alsa/alsa-sink.c | 53 ++++++++++++++++++++++-------------------
src/pulsecore/flist.c | 5 +++-
src/pulsecore/flist.h | 4 +-
src/pulsecore/sink.c | 33 ++++++++++++++------------
4 files changed, 52 insertions(+), 43 deletions(-)
