voice processing: support several sessions per input stream
Bug: 128419018
Test: CTS AudioPreProcessingTest, manual tests for VoIP calls and
camcorder.
Change-Id: I5fe39451c30889e8abd56a16ddbb6bdf378ef2f2
(cherry picked from commit f69b0d0f1b2f3ea236961ca2c15ef8bec8087f26)
diff --git a/voice_processing/voice_processing.c b/voice_processing/voice_processing.c
index 366adf7..50cb7af 100644
--- a/voice_processing/voice_processing.c
+++ b/voice_processing/voice_processing.c
@@ -414,7 +414,7 @@
list_for_each(node, &session_list) {
session = node_to_item(node, struct session_s, node);
- if (session->io == ioId) {
+ if (session->id == sessionId) {
if (session->created_msk & (1 << id)) {
ALOGV("get_session() effect %d already created", id);
return NULL;