diff --git a/hal/audio_hw.c b/hal/audio_hw.c
index d83002a..9ccbcc0 100644
--- a/hal/audio_hw.c
+++ b/hal/audio_hw.c
@@ -4358,6 +4358,8 @@
     struct audio_device *adev = out->dev;
     ssize_t ret = 0;
     int channels = 0;
+    const size_t frame_size = audio_stream_out_frame_size(stream);
+    const size_t frames = (frame_size != 0) ? bytes / frame_size : bytes;
 
     ATRACE_BEGIN("out_write");
     lock_output_stream(out);
@@ -4563,8 +4565,35 @@
         return ret;
     } else {
         if (out->pcm) {
+            size_t bytes_to_write = bytes;
             if (out->muted)
                 memset((void *)buffer, 0, bytes);
+            ALOGV("%s: frames=%zu, frame_size=%zu, bytes_to_write=%zu",
+                     __func__, frames, frame_size, bytes_to_write);
+
+            if (out->usecase == USECASE_INCALL_MUSIC_UPLINK) {
+                size_t channel_count = audio_channel_count_from_out_mask(out->channel_mask);
+                int16_t *src = (int16_t *)buffer;
+                int16_t *dst = (int16_t *)buffer;
+
+                LOG_ALWAYS_FATAL_IF(out->config.channels != 1 || channel_count != 2 ||
+                                    out->format != AUDIO_FORMAT_PCM_16_BIT,
+                                    "out_write called for incall music use case with wrong properties");
+
+                /*
+                 * FIXME: this can be removed once audio flinger mixer supports
+                 * mono output
+                 */
+
+                /*
+                 * Code below goes over each frame in the buffer and adds both
+                 * L and R samples and then divides by 2 to convert to mono
+                 */
+                for (size_t i = 0; i < frames ; i++, dst++, src += 2) {
+                    *dst = (int16_t)(((int32_t)src[0] + (int32_t)src[1]) >> 1);
+                }
+                bytes_to_write /= 2;
+            }
 
             ALOGVV("%s: writing buffer (%zu bytes) to pcm device", __func__, bytes);
 
@@ -4574,12 +4603,11 @@
                 ns = pcm_bytes_to_frames(out->pcm, bytes)*1000000000LL/
                                                      out->config.rate;
 
+            request_out_focus(out, ns);
             bool use_mmap = is_mmap_usecase(out->usecase) || out->realtime;
 
-            request_out_focus(out, ns);
-
             if (use_mmap)
-                ret = pcm_mmap_write(out->pcm, (void *)buffer, bytes);
+                ret = pcm_mmap_write(out->pcm, (void *)buffer, bytes_to_write);
             else if (out->hal_op_format != out->hal_ip_format &&
                        out->convert_buffer != NULL) {
 
@@ -4613,7 +4641,7 @@
                            out_get_sample_rate(&out->stream.common));
                     ret = 0;
                 } else
-                    ret = pcm_write(out->pcm, (void *)buffer, bytes);
+                    ret = pcm_write(out->pcm, (void *)buffer, bytes_to_write);
             }
 
             release_out_focus(out);
@@ -6119,13 +6147,55 @@
         create_offload_callback_thread(out);
 
     } else if (out->flags & AUDIO_OUTPUT_FLAG_INCALL_MUSIC) {
+          switch (config->sample_rate) {
+            case 0:
+                out->sample_rate = DEFAULT_OUTPUT_SAMPLING_RATE;
+                break;
+            case 8000:
+            case 16000:
+            case 48000:
+                out->sample_rate = config->sample_rate;
+                break;
+            default:
+                ALOGE("%s: Unsupported sampling rate %d for Incall Music", __func__,
+                      config->sample_rate);
+                config->sample_rate = DEFAULT_OUTPUT_SAMPLING_RATE;
+                ret = -EINVAL;
+                goto error_open;
+        }
+        //FIXME: add support for MONO stream configuration when audioflinger mixer supports it
+        switch (config->channel_mask) {
+            case AUDIO_CHANNEL_NONE:
+            case AUDIO_CHANNEL_OUT_STEREO:
+                out->channel_mask = AUDIO_CHANNEL_OUT_STEREO;
+                break;
+            default:
+                ALOGE("%s: Unsupported channel mask %#x for Incall Music", __func__,
+                      config->channel_mask);
+                config->channel_mask = AUDIO_CHANNEL_OUT_STEREO;
+                ret = -EINVAL;
+                goto error_open;
+        }
+        switch (config->format) {
+            case AUDIO_FORMAT_DEFAULT:
+            case AUDIO_FORMAT_PCM_16_BIT:
+                out->format = AUDIO_FORMAT_PCM_16_BIT;
+                break;
+            default:
+                ALOGE("%s: Unsupported format %#x for Incall Music", __func__,
+                      config->format);
+                config->format = AUDIO_FORMAT_PCM_16_BIT;
+                ret = -EINVAL;
+                goto error_open;
+        }
+
         ret = voice_extn_check_and_set_incall_music_usecase(adev, out);
         if (ret != 0) {
             ALOGE("%s: Incall music delivery usecase cannot be set error:%d",
-                  __func__, ret);
+                __func__, ret);
             goto error_open;
         }
-    } else  if (out->devices == AUDIO_DEVICE_OUT_TELEPHONY_TX) {
+    } else if (out->devices == AUDIO_DEVICE_OUT_TELEPHONY_TX) {
         if (config->sample_rate == 0)
             config->sample_rate = AFE_PROXY_SAMPLING_RATE;
         if (config->sample_rate != 48000 && config->sample_rate != 16000 &&
diff --git a/hal/voice_extn/voice_extn.c b/hal/voice_extn/voice_extn.c
index 93653ca..ec85259 100644
--- a/hal/voice_extn/voice_extn.c
+++ b/hal/voice_extn/voice_extn.c
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2013-2017, The Linux Foundation. All rights reserved.
+ * Copyright (c) 2013-2018, The Linux Foundation. All rights reserved.
  * Not a contribution.
  *
  * Copyright (C) 2013 The Android Open Source Project
@@ -594,27 +594,16 @@
     voice_extn_compress_voip_in_get_parameters(in, query, reply);
 }
 
-#ifdef INCALL_MUSIC_ENABLED
 int voice_extn_check_and_set_incall_music_usecase(struct audio_device *adev,
                                                   struct stream_out *out)
 {
-    uint32_t session_id = 0;
-
-    session_id = get_session_id_with_state(adev, CALL_LOCAL_HOLD);
-    if (session_id == VOICE_VSID) {
-        out->usecase = USECASE_INCALL_MUSIC_UPLINK;
-    } else if (session_id == VOICE2_VSID) {
-        out->usecase = USECASE_INCALL_MUSIC_UPLINK2;
-    } else {
-        ALOGE("%s: Invalid session id %x", __func__, session_id);
-        return -EINVAL;
-    }
-
+    out->usecase = USECASE_INCALL_MUSIC_UPLINK;
     out->config = pcm_config_incall_music;
-    out->supported_channel_masks[0] = AUDIO_CHANNEL_OUT_MONO;
-    out->channel_mask = AUDIO_CHANNEL_OUT_MONO;
+    //FIXME: add support for MONO stream configuration when audioflinger mixer supports it
+    out->supported_channel_masks[0] = AUDIO_CHANNEL_OUT_STEREO;
+    out->channel_mask = AUDIO_CHANNEL_OUT_STEREO;
+    out->config.rate = out->sample_rate;
 
+    ALOGV("%s: mode=%d, usecase id=%d", __func__, adev->mode, out->usecase);
     return 0;
 }
-#endif
-
diff --git a/hal/voice_extn/voice_extn.h b/hal/voice_extn/voice_extn.h
index f35344f..5d1cac3 100644
--- a/hal/voice_extn/voice_extn.h
+++ b/hal/voice_extn/voice_extn.h
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2013-2014, 2016-2017, The Linux Foundation. All rights reserved.
+ * Copyright (c) 2013-2014, 2016-2018, The Linux Foundation. All rights reserved.
  * Not a contribution.
  *
  * Copyright (C) 2013 The Android Open Source Project
@@ -101,13 +101,8 @@
 }
 #endif
 
-#ifdef INCALL_MUSIC_ENABLED
 int voice_extn_check_and_set_incall_music_usecase(struct audio_device *adev,
                                                   struct stream_out *out);
-#else
-#define voice_extn_check_and_set_incall_music_usecase(adev, out) -ENOSYS
-#endif
-
 #ifdef COMPRESS_VOIP_ENABLED
 int voice_extn_compress_voip_close_output_stream(struct audio_stream *stream);
 int voice_extn_compress_voip_open_output_stream(struct stream_out *out);