summaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
authorManuel Bachmann <manuel.bachmann@iot.bzh>2016-06-10 14:08:06 +0200
committerYannick Gicquel <yannick.gicquel@iot.bzh>2016-10-11 17:09:06 +0200
commitf8bc0b1decdd464e62e2a1497e3287ead96f40d0 (patch)
tree619f79c269a3f61d97e81af618bd4abb6694f1a6
parent6c2aef81c4182695487fabbd19c5c0a84409b49f (diff)
Fix invalid sink input treatment after module-loopback
When loading module-loopback, it may create a new sink input replicating the original one, but with no client attached. We don't want to re-pass this one through the whole logic (with NULL sink, etc), so just consider it legit. Change-Id: I218c459e1685621c1f59e1d8722e0f1cff88d225 Signed-off-by: Manuel Bachmann <manuel.bachmann@iot.bzh>
-rw-r--r--discover.c4
1 files changed, 4 insertions, 0 deletions
diff --git a/discover.c b/discover.c
index e6e8243..eea6986 100644
--- a/discover.c
+++ b/discover.c
@@ -586,6 +586,10 @@ bool pa_discover_preroute_sink_input(struct userdata *u, pa_sink_input_new_data
pa_assert_se (nodeset = u->nodeset);
pa_assert_se (pl = data->proplist);
+ /* is this a valid sink input ? */
+ if (!data->client)
+ return true;
+
/* is there an existing matching node ? */
node = agl_node_get_from_data (u, agl_input, data);