diff --git a/sound/soc/sof/intel/hda-loader.c b/sound/soc/sof/intel/hda-loader.c
index 76ac08d8d0883e..b9604b4ea27cd6 100644
--- a/sound/soc/sof/intel/hda-loader.c
+++ b/sound/soc/sof/intel/hda-loader.c
@@ -66,7 +66,8 @@ static int cl_stream_prepare(struct snd_sof_dev *sdev, unsigned int format,
 	hstream->format_val = format;
 	hstream->bufsize = size;
 
-	ret = hda_dsp_stream_hw_params(sdev, stream, dmab, NULL);
+	ret = hda_dsp_stream_hw_params(sdev, stream, dmab,
+				       NULL, SOF_HDA_CL_STREAM);
 	if (ret < 0) {
 		dev_err(sdev->dev, "error: hdac prepare failed: %x\n", ret);
 		goto error;
diff --git a/sound/soc/sof/intel/hda-pcm.c b/sound/soc/sof/intel/hda-pcm.c
index d5a09dcceeee4d..cde830311725a7 100644
--- a/sound/soc/sof/intel/hda-pcm.c
+++ b/sound/soc/sof/intel/hda-pcm.c
@@ -121,7 +121,8 @@ int hda_dsp_pcm_hw_params(struct snd_sof_dev *sdev,
 			(params->info & SNDRV_PCM_INFO_NO_PERIOD_WAKEUP) &&
 			(params->flags & SNDRV_PCM_HW_PARAMS_NO_PERIOD_WAKEUP);
 
-	ret = hda_dsp_stream_hw_params(sdev, stream, dmab, params);
+	ret = hda_dsp_stream_hw_params(sdev, stream, dmab,
+				       params, SOF_HDA_AUDIO_STREAM);
 	if (ret < 0) {
 		dev_err(sdev->dev, "error: hdac prepare failed: %x\n", ret);
 		return ret;
diff --git a/sound/soc/sof/intel/hda-stream.c b/sound/soc/sof/intel/hda-stream.c
index 42a986ce35a08f..6455d980b682e4 100644
--- a/sound/soc/sof/intel/hda-stream.c
+++ b/sound/soc/sof/intel/hda-stream.c
@@ -88,31 +88,46 @@ static int hda_setup_bdle(struct snd_sof_dev *sdev,
 /*
  * set up Buffer Descriptor List (BDL) for host memory transfer
  * BDL describes the location of the individual buffers and is little endian.
+ *
+ * The typical BDL entries after set up should looks like this:
+ *
+ * bdl[0] from the init_offset,  n entries before wrap,
+ * the e-1 full entries, and the last e entry.
+ *
+ *			     init_offset
+ *				 |
+ * |------|...........|---------S-------|------|........|--------|-------|
+ *  bdl[n] bdl[n+j]    bdl[e-1] | bdl[0] bdl[1] bdl[m]   bdl[n-2] bdl[n-1]
  */
 int hda_dsp_stream_setup_bdl(struct snd_sof_dev *sdev,
 			     struct snd_dma_buffer *dmab,
-			     struct hdac_stream *stream)
+			     struct hdac_stream *stream, u32 init_offset)
 {
 	struct sof_intel_dsp_bdl *bdl;
-	int i, offset, period_bytes, periods;
-	int remain, ioc;
+	unsigned int i, period_bytes, entries;
+	u32 bytes, offset, size = 0;
+	bool ioc, set_ioc, is_wrap = false;
 
 	period_bytes = stream->period_bytes;
 	dev_dbg(sdev->dev, "period_bytes:0x%x\n", period_bytes);
 	if (!period_bytes)
 		period_bytes = stream->bufsize;
 
-	periods = stream->bufsize / period_bytes;
+	entries = stream->bufsize / period_bytes;
 
-	dev_dbg(sdev->dev, "periods:%d\n", periods);
+	/* count the last offset fragment entry */
+	if (stream->bufsize % period_bytes)
+		entries++;
 
-	remain = stream->bufsize % period_bytes;
-	if (remain)
-		periods++;
+	/* count the last index fragment entry */
+	if (init_offset % period_bytes)
+		entries++;
+
+	dev_dbg(sdev->dev, "entries:%d\n", entries);
 
 	/* program the initial BDL entries */
 	bdl = (struct sof_intel_dsp_bdl *)stream->bdl.area;
-	offset = 0;
+	offset = init_offset;
 	stream->frags = 0;
 
 	/*
@@ -122,19 +137,28 @@ int hda_dsp_stream_setup_bdl(struct snd_sof_dev *sdev,
 	ioc = sdev->hda->no_ipc_position ?
 	      !stream->no_period_wakeup : 0;
 
-	for (i = 0; i < periods; i++) {
-		if (i == (periods - 1) && remain)
-			/* set the last small entry */
-			offset = hda_setup_bdle(sdev, dmab,
-						stream, &bdl, offset,
-						remain, 0);
+	for (i = 0; i < entries; i++) {
+		bytes = period_bytes - offset % period_bytes;
+		if (is_wrap)
+			bytes = min(bytes, init_offset - offset);
 		else
-			offset = hda_setup_bdle(sdev, dmab,
-						stream, &bdl, offset,
-						period_bytes, ioc);
+			bytes = min(bytes, stream->bufsize - offset);
+
+		/* only set ioc at period bundary */
+		set_ioc = (offset + bytes) % period_bytes ? 0 : 1;
+		offset = hda_setup_bdle(sdev, dmab,
+					stream, &bdl, offset,
+					bytes, set_ioc && ioc);
+
+		if (offset == stream->bufsize) {
+			offset = 0;
+			is_wrap = true;
+		}
+
+		size += bytes;
 	}
 
-	return offset;
+	return size;
 }
 
 int hda_dsp_stream_spib_config(struct snd_sof_dev *sdev,
@@ -344,19 +368,21 @@ int hda_dsp_stream_trigger(struct snd_sof_dev *sdev,
 }
 
 /*
- * prepare for common hdac registers settings, for both code loader
+ * prepare for common hdac registers settings, for code loader, dma trace,
  * and normal stream.
  */
 int hda_dsp_stream_hw_params(struct snd_sof_dev *sdev,
 			     struct hdac_ext_stream *stream,
 			     struct snd_dma_buffer *dmab,
-			     struct snd_pcm_hw_params *params)
+			     struct snd_pcm_hw_params *params,
+			     enum sof_hda_tream_type stream_type)
 {
 	struct hdac_bus *bus = sof_to_bus(sdev);
 	struct hdac_stream *hstream = &stream->hstream;
 	int sd_offset = SOF_STREAM_SD_OFFSET(hstream);
 	int ret, timeout = HDA_DSP_STREAM_RESET_TIMEOUT;
 	u32 val, mask;
+	u32 init_offset = 0;
 
 	if (!stream) {
 		dev_err(sdev->dev, "error: no stream available\n");
@@ -436,10 +462,13 @@ int hda_dsp_stream_hw_params(struct snd_sof_dev *sdev,
 
 	hstream->frags = 0;
 
-	ret = hda_dsp_stream_setup_bdl(sdev, dmab, hstream);
-	if (ret < 0) {
-		dev_err(sdev->dev, "error: set up of BDL failed\n");
-		return ret;
+	if (stream_type == SOF_HDA_TRACE_STREAM)
+		init_offset = sdev->trace_init_offset;
+
+	ret = hda_dsp_stream_setup_bdl(sdev, dmab, hstream, init_offset);
+	if (ret != hstream->bufsize) {
+		dev_err(sdev->dev, "error: set up of BDL failed, bufsize:0x%x, only 0x%x set\n", hstream->bufsize, ret);
+		return -EFAULT;
 	}
 
 	/* set up stream descriptor for DMA */
diff --git a/sound/soc/sof/intel/hda-trace.c b/sound/soc/sof/intel/hda-trace.c
index a1c9f1b45dca14..a158ae069c180c 100644
--- a/sound/soc/sof/intel/hda-trace.c
+++ b/sound/soc/sof/intel/hda-trace.c
@@ -43,7 +43,8 @@ static int hda_dsp_trace_prepare(struct snd_sof_dev *sdev)
 
 	hstream->bufsize = sdev->dmatb.bytes;
 
-	ret = hda_dsp_stream_hw_params(sdev, stream, dmab, NULL);
+	ret = hda_dsp_stream_hw_params(sdev, stream, dmab,
+				       NULL, SOF_HDA_TRACE_STREAM);
 	if (ret < 0)
 		dev_err(sdev->dev, "error: hdac prepare failed: %x\n", ret);
 
@@ -75,7 +76,6 @@ int hda_dsp_trace_release(struct snd_sof_dev *sdev)
 
 	if (sdev->hda->dtrace_stream) {
 		hstream = &sdev->hda->dtrace_stream->hstream;
-		hstream->opened = false;
 		hda_dsp_stream_put_cstream(sdev,
 					   hstream->stream_tag);
 		sdev->hda->dtrace_stream = NULL;
diff --git a/sound/soc/sof/intel/hda.h b/sound/soc/sof/intel/hda.h
index c9c3e52508f00f..17962601dbf5de 100644
--- a/sound/soc/sof/intel/hda.h
+++ b/sound/soc/sof/intel/hda.h
@@ -343,6 +343,12 @@ struct sof_intel_dsp_desc {
 	struct snd_sof_dsp_ops *ops;
 };
 
+enum sof_hda_tream_type {
+	SOF_HDA_AUDIO_STREAM = 0,	/* normal playback/capture stream */
+	SOF_HDA_CL_STREAM,		/* code loader stream */
+	SOF_HDA_TRACE_STREAM,		/* dma trace capture stream */
+};
+
 #define SOF_HDA_PLAYBACK_STREAMS	16
 #define SOF_HDA_CAPTURE_STREAMS		16
 #define SOF_HDA_PLAYBACK		0
@@ -442,14 +448,15 @@ void hda_dsp_stream_free(struct snd_sof_dev *sdev);
 int hda_dsp_stream_hw_params(struct snd_sof_dev *sdev,
 			     struct hdac_ext_stream *stream,
 			     struct snd_dma_buffer *dmab,
-			     struct snd_pcm_hw_params *params);
+			     struct snd_pcm_hw_params *params,
+			     enum sof_hda_tream_type stream_type);
 int hda_dsp_stream_trigger(struct snd_sof_dev *sdev,
 			   struct hdac_ext_stream *stream, int cmd);
 irqreturn_t hda_dsp_stream_interrupt(int irq, void *context);
 irqreturn_t hda_dsp_stream_threaded_handler(int irq, void *context);
 int hda_dsp_stream_setup_bdl(struct snd_sof_dev *sdev,
 			     struct snd_dma_buffer *dmab,
-			     struct hdac_stream *stream);
+			     struct hdac_stream *stream, u32 init_offset);
 
 struct hdac_ext_stream *
 	hda_dsp_stream_get(struct snd_sof_dev *sdev, int direction);
diff --git a/sound/soc/sof/sof-priv.h b/sound/soc/sof/sof-priv.h
index 0f7ddf19bf8a15..90ac1c1192497a 100644
--- a/sound/soc/sof/sof-priv.h
+++ b/sound/soc/sof/sof-priv.h
@@ -363,7 +363,8 @@ struct snd_sof_dev {
 	struct snd_dma_buffer dmatp;
 	int dma_trace_pages;
 	wait_queue_head_t trace_sleep;
-	u32 host_offset;
+	u32 trace_init_offset; /* restore trace offset upon resume */
+	u32 trace_offset;
 	bool dtrace_is_enabled;
 	bool dtrace_error;
 
diff --git a/sound/soc/sof/trace.c b/sound/soc/sof/trace.c
index 5aeb6ea3c58b2b..c08bc7b98ea131 100644
--- a/sound/soc/sof/trace.c
+++ b/sound/soc/sof/trace.c
@@ -30,18 +30,23 @@ static size_t sof_wait_trace_avail(struct snd_sof_dev *sdev,
 				   loff_t pos, size_t buffer_size)
 {
 	wait_queue_entry_t wait;
+	u32 host_offset;
+
+	/* count the initial offset to get the buffer offset */
+	host_offset = (sdev->trace_init_offset + sdev->trace_offset) %
+		      buffer_size;
 
 	/*
 	 * If host offset is less than local pos, it means write pointer of
 	 * host DMA buffer has been wrapped. We should output the trace data
 	 * at the end of host DMA buffer at first.
 	 */
-	if (sdev->host_offset < pos)
+	if (host_offset < pos)
 		return buffer_size - pos;
 
 	/* If there is available trace data now, it is unnecessary to wait. */
-	if (sdev->host_offset > pos)
-		return sdev->host_offset - pos;
+	if (host_offset > pos)
+		return host_offset - pos;
 
 	/* wait for available trace data from FW */
 	init_waitqueue_entry(&wait, current);
@@ -58,11 +63,15 @@ static size_t sof_wait_trace_avail(struct snd_sof_dev *sdev,
 	remove_wait_queue(&sdev->trace_sleep, &wait);
 
 out:
+	/* update buffer offset after wait done */
+	host_offset = (sdev->trace_init_offset + sdev->trace_offset) %
+		      buffer_size;
+
 	/* return bytes available for copy */
-	if (sdev->host_offset < pos)
+	if (host_offset < pos)
 		return buffer_size - pos;
 	else
-		return sdev->host_offset - pos;
+		return host_offset - pos;
 }
 
 static ssize_t sof_dfsentry_trace_read(struct file *file, char __user *buffer,
@@ -154,16 +163,24 @@ int snd_sof_init_trace_ipc(struct snd_sof_dev *sdev)
 	if (sdev->dtrace_is_enabled)
 		return -EINVAL;
 
+	/* update init_offset, and reinitial offset to 0 */
+	sdev->trace_init_offset = (sdev->trace_init_offset +
+				   sdev->trace_offset) %
+				   sdev->dmatb.bytes;
+	sdev->trace_offset = 0;
+
 	/* set IPC parameters */
 	params.hdr.size = sizeof(params);
 	params.hdr.cmd = SOF_IPC_GLB_TRACE_MSG | SOF_IPC_TRACE_DMA_PARAMS;
 	params.buffer.phy_addr = sdev->dmatp.addr;
 	params.buffer.size = sdev->dmatb.bytes;
-	params.buffer.offset = 0;
 	params.buffer.pages = sdev->dma_trace_pages;
 
-	init_waitqueue_head(&sdev->trace_sleep);
-	sdev->host_offset = 0;
+	/*
+	 * send offset to FW to require copying new logs
+	 * start from new offset if possible.
+	 */
+	params.buffer.offset = sdev->trace_init_offset;
 
 	ret = snd_sof_dma_trace_init(sdev, &params.stream_tag);
 	if (ret < 0) {
@@ -201,6 +218,8 @@ int snd_sof_init_trace(struct snd_sof_dev *sdev)
 
 	/* set false before start initialization */
 	sdev->dtrace_is_enabled = false;
+	sdev->trace_offset = 0;
+	sdev->trace_init_offset = 0;
 
 	/* allocate trace page table buffer */
 	ret = snd_dma_alloc_pages(SNDRV_DMA_TYPE_DEV, sdev->parent,
@@ -235,6 +254,10 @@ int snd_sof_init_trace(struct snd_sof_dev *sdev)
 			goto table_err;
 	}
 
+	init_waitqueue_head(&sdev->trace_sleep);
+
+	init_waitqueue_head(&sdev->trace_sleep);
+
 	ret = snd_sof_init_trace_ipc(sdev);
 	if (ret < 0)
 		goto table_err;
@@ -251,8 +274,9 @@ EXPORT_SYMBOL(snd_sof_init_trace);
 int snd_sof_trace_update_pos(struct snd_sof_dev *sdev,
 			     struct sof_ipc_dma_trace_posn *posn)
 {
-	if (sdev->dtrace_is_enabled && sdev->host_offset != posn->host_offset) {
-		sdev->host_offset = posn->host_offset;
+	if (sdev->dtrace_is_enabled &&
+	    sdev->trace_offset != posn->host_offset) {
+		sdev->trace_offset = posn->host_offset;
 		wake_up(&sdev->trace_sleep);
 	}