X-Git-Url: https://git.kernel.dk/?p=fio.git;a=blobdiff_plain;f=stat.c;h=b723795301cc4c9bb3e71ad31450b6a6af97366b;hp=bf87917c2956eb60729fa9ea95ac93b1cbe1ed50;hb=ab02be41807ec9451c47b17129cf61457ef21db6;hpb=d5495f0b72f8fbecb97192da430720aa56f8feb9

diff --git a/stat.c b/stat.c
index bf87917c..b7237953 100644
--- a/stat.c
+++ b/stat.c
@@ -15,6 +15,7 @@
 #include "helper_thread.h"
 #include "smalloc.h"
 #include "zbd.h"
+#include "oslib/asprintf.h"
 
 #define LOG_MSEC_SLACK	1
 
@@ -158,7 +159,7 @@ unsigned int calc_clat_percentiles(uint64_t *io_u_plat, unsigned long long nr,
 	 * isn't a worry. Also note that this does not work for NaN values.
 	 */
 	if (len > 1)
-		qsort((void *)plist, len, sizeof(plist[0]), double_cmp);
+		qsort(plist, len, sizeof(plist[0]), double_cmp);
 
 	ovals = malloc(len * sizeof(*ovals));
 	if (!ovals)
@@ -258,8 +259,7 @@ static void show_clat_percentiles(uint64_t *io_u_plat, unsigned long long nr,
 	}
 
 out:
-	if (ovals)
-		free(ovals);
+	free(ovals);
 }
 
 bool calc_lat(struct io_stat *is, unsigned long long *min,
@@ -414,6 +414,18 @@ static void display_lat(const char *name, unsigned long long min,
 	free(maxp);
 }
 
+static double convert_agg_kbytes_percent(struct group_run_stats *rs, int ddir, int mean)
+{
+	double p_of_agg = 100.0;
+	if (rs && rs->agg[ddir] > 1024) {
+		p_of_agg = mean * 100 / (double) (rs->agg[ddir] / 1024.0);
+
+		if (p_of_agg > 100.0)
+			p_of_agg = 100.0;
+	}
+	return p_of_agg;
+}
+
 static void show_ddir_status(struct group_run_stats *rs, struct thread_stat *ts,
 			     int ddir, struct buf_output *out)
 {
@@ -482,21 +494,62 @@ static void show_ddir_status(struct group_run_stats *rs, struct thread_stat *ts,
 		display_lat("clat", min, max, mean, dev, out);
 	if (calc_lat(&ts->lat_stat[ddir], &min, &max, &mean, &dev))
 		display_lat(" lat", min, max, mean, dev, out);
+	if (calc_lat(&ts->clat_high_prio_stat[ddir], &min, &max, &mean, &dev)) {
+		display_lat(ts->lat_percentiles ? "high prio_lat" : "high prio_clat",
+				min, max, mean, dev, out);
+		if (calc_lat(&ts->clat_low_prio_stat[ddir], &min, &max, &mean, &dev))
+			display_lat(ts->lat_percentiles ? "low prio_lat" : "low prio_clat",
+					min, max, mean, dev, out);
+	}
+
+	if (ts->slat_percentiles && ts->slat_stat[ddir].samples > 0)
+		show_clat_percentiles(ts->io_u_plat[FIO_SLAT][ddir],
+					ts->slat_stat[ddir].samples,
+					ts->percentile_list,
+					ts->percentile_precision, "slat", out);
+	if (ts->clat_percentiles && ts->clat_stat[ddir].samples > 0)
+		show_clat_percentiles(ts->io_u_plat[FIO_CLAT][ddir],
+					ts->clat_stat[ddir].samples,
+					ts->percentile_list,
+					ts->percentile_precision, "clat", out);
+	if (ts->lat_percentiles && ts->lat_stat[ddir].samples > 0)
+		show_clat_percentiles(ts->io_u_plat[FIO_LAT][ddir],
+					ts->lat_stat[ddir].samples,
+					ts->percentile_list,
+					ts->percentile_precision, "lat", out);
 
 	if (ts->clat_percentiles || ts->lat_percentiles) {
-		const char *name = ts->clat_percentiles ? "clat" : " lat";
+		const char *name = ts->lat_percentiles ? "lat" : "clat";
+		char prio_name[32];
 		uint64_t samples;
 
-		if (ts->clat_percentiles)
-			samples = ts->clat_stat[ddir].samples;
-		else
+		if (ts->lat_percentiles)
 			samples = ts->lat_stat[ddir].samples;
+		else
+			samples = ts->clat_stat[ddir].samples;
 
-		show_clat_percentiles(ts->io_u_plat[ddir],
-					samples,
-					ts->percentile_list,
-					ts->percentile_precision, name, out);
+		/* Only print this if some high and low priority stats were collected */
+		if (ts->clat_high_prio_stat[ddir].samples > 0 &&
+			ts->clat_low_prio_stat[ddir].samples > 0)
+		{
+			sprintf(prio_name, "high prio (%.2f%%) %s",
+					100. * (double) ts->clat_high_prio_stat[ddir].samples / (double) samples,
+					name);
+			show_clat_percentiles(ts->io_u_plat_high_prio[ddir],
+						ts->clat_high_prio_stat[ddir].samples,
+						ts->percentile_list,
+						ts->percentile_precision, prio_name, out);
+
+			sprintf(prio_name, "low prio (%.2f%%) %s",
+					100. * (double) ts->clat_low_prio_stat[ddir].samples / (double) samples,
+					name);
+			show_clat_percentiles(ts->io_u_plat_low_prio[ddir],
+						ts->clat_low_prio_stat[ddir].samples,
+						ts->percentile_list,
+						ts->percentile_precision, prio_name, out);
+		}
 	}
+
 	if (calc_lat(&ts->bw_stat[ddir], &min, &max, &mean, &dev)) {
 		double p_of_agg = 100.0, fkb_base = (double)rs->kb_base;
 		const char *bw_str;
@@ -510,11 +563,7 @@ static void show_ddir_status(struct group_run_stats *rs, struct thread_stat *ts,
 		else
 			bw_str = "kB";
 
-		if (rs->agg[ddir]) {
-			p_of_agg = mean * 100 / (double) (rs->agg[ddir] / 1024);
-			if (p_of_agg > 100.0)
-				p_of_agg = 100.0;
-		}
+		p_of_agg = convert_agg_kbytes_percent(rs, ddir, mean);
 
 		if (rs->unit_base == 1) {
 			min *= 8.0;
@@ -683,7 +732,7 @@ static int calc_block_percentiles(int nr_block_infos, uint32_t *block_infos,
 	 * isn't a worry. Also note that this does not work for NaN values.
 	 */
 	if (len > 1)
-		qsort((void *)plist, len, sizeof(plist[0]), double_cmp);
+		qsort(plist, len, sizeof(plist[0]), double_cmp);
 
 	/* Start only after the uninit entries end */
 	for (nr_uninit = 0;
@@ -785,6 +834,218 @@ static void show_ss_normal(struct thread_stat *ts, struct buf_output *out)
 	free(p2);
 }
 
+static void show_agg_stats(struct disk_util_agg *agg, int terse,
+			   struct buf_output *out)
+{
+	if (!agg->slavecount)
+		return;
+
+	if (!terse) {
+		log_buf(out, ", aggrios=%llu/%llu, aggrmerge=%llu/%llu, "
+			 "aggrticks=%llu/%llu, aggrin_queue=%llu, "
+			 "aggrutil=%3.2f%%",
+			(unsigned long long) agg->ios[0] / agg->slavecount,
+			(unsigned long long) agg->ios[1] / agg->slavecount,
+			(unsigned long long) agg->merges[0] / agg->slavecount,
+			(unsigned long long) agg->merges[1] / agg->slavecount,
+			(unsigned long long) agg->ticks[0] / agg->slavecount,
+			(unsigned long long) agg->ticks[1] / agg->slavecount,
+			(unsigned long long) agg->time_in_queue / agg->slavecount,
+			agg->max_util.u.f);
+	} else {
+		log_buf(out, ";slaves;%llu;%llu;%llu;%llu;%llu;%llu;%llu;%3.2f%%",
+			(unsigned long long) agg->ios[0] / agg->slavecount,
+			(unsigned long long) agg->ios[1] / agg->slavecount,
+			(unsigned long long) agg->merges[0] / agg->slavecount,
+			(unsigned long long) agg->merges[1] / agg->slavecount,
+			(unsigned long long) agg->ticks[0] / agg->slavecount,
+			(unsigned long long) agg->ticks[1] / agg->slavecount,
+			(unsigned long long) agg->time_in_queue / agg->slavecount,
+			agg->max_util.u.f);
+	}
+}
+
+static void aggregate_slaves_stats(struct disk_util *masterdu)
+{
+	struct disk_util_agg *agg = &masterdu->agg;
+	struct disk_util_stat *dus;
+	struct flist_head *entry;
+	struct disk_util *slavedu;
+	double util;
+
+	flist_for_each(entry, &masterdu->slaves) {
+		slavedu = flist_entry(entry, struct disk_util, slavelist);
+		dus = &slavedu->dus;
+		agg->ios[0] += dus->s.ios[0];
+		agg->ios[1] += dus->s.ios[1];
+		agg->merges[0] += dus->s.merges[0];
+		agg->merges[1] += dus->s.merges[1];
+		agg->sectors[0] += dus->s.sectors[0];
+		agg->sectors[1] += dus->s.sectors[1];
+		agg->ticks[0] += dus->s.ticks[0];
+		agg->ticks[1] += dus->s.ticks[1];
+		agg->time_in_queue += dus->s.time_in_queue;
+		agg->slavecount++;
+
+		util = (double) (100 * dus->s.io_ticks / (double) slavedu->dus.s.msec);
+		/* System utilization is the utilization of the
+		 * component with the highest utilization.
+		 */
+		if (util > agg->max_util.u.f)
+			agg->max_util.u.f = util;
+
+	}
+
+	if (agg->max_util.u.f > 100.0)
+		agg->max_util.u.f = 100.0;
+}
+
+void print_disk_util(struct disk_util_stat *dus, struct disk_util_agg *agg,
+		     int terse, struct buf_output *out)
+{
+	double util = 0;
+
+	if (dus->s.msec)
+		util = (double) 100 * dus->s.io_ticks / (double) dus->s.msec;
+	if (util > 100.0)
+		util = 100.0;
+
+	if (!terse) {
+		if (agg->slavecount)
+			log_buf(out, "  ");
+
+		log_buf(out, "  %s: ios=%llu/%llu, merge=%llu/%llu, "
+			 "ticks=%llu/%llu, in_queue=%llu, util=%3.2f%%",
+				dus->name,
+				(unsigned long long) dus->s.ios[0],
+				(unsigned long long) dus->s.ios[1],
+				(unsigned long long) dus->s.merges[0],
+				(unsigned long long) dus->s.merges[1],
+				(unsigned long long) dus->s.ticks[0],
+				(unsigned long long) dus->s.ticks[1],
+				(unsigned long long) dus->s.time_in_queue,
+				util);
+	} else {
+		log_buf(out, ";%s;%llu;%llu;%llu;%llu;%llu;%llu;%llu;%3.2f%%",
+				dus->name,
+				(unsigned long long) dus->s.ios[0],
+				(unsigned long long) dus->s.ios[1],
+				(unsigned long long) dus->s.merges[0],
+				(unsigned long long) dus->s.merges[1],
+				(unsigned long long) dus->s.ticks[0],
+				(unsigned long long) dus->s.ticks[1],
+				(unsigned long long) dus->s.time_in_queue,
+				util);
+	}
+
+	/*
+	 * If the device has slaves, aggregate the stats for
+	 * those slave devices also.
+	 */
+	show_agg_stats(agg, terse, out);
+
+	if (!terse)
+		log_buf(out, "\n");
+}
+
+void json_array_add_disk_util(struct disk_util_stat *dus,
+		struct disk_util_agg *agg, struct json_array *array)
+{
+	struct json_object *obj;
+	double util = 0;
+
+	if (dus->s.msec)
+		util = (double) 100 * dus->s.io_ticks / (double) dus->s.msec;
+	if (util > 100.0)
+		util = 100.0;
+
+	obj = json_create_object();
+	json_array_add_value_object(array, obj);
+
+	json_object_add_value_string(obj, "name", (const char *)dus->name);
+	json_object_add_value_int(obj, "read_ios", dus->s.ios[0]);
+	json_object_add_value_int(obj, "write_ios", dus->s.ios[1]);
+	json_object_add_value_int(obj, "read_merges", dus->s.merges[0]);
+	json_object_add_value_int(obj, "write_merges", dus->s.merges[1]);
+	json_object_add_value_int(obj, "read_ticks", dus->s.ticks[0]);
+	json_object_add_value_int(obj, "write_ticks", dus->s.ticks[1]);
+	json_object_add_value_int(obj, "in_queue", dus->s.time_in_queue);
+	json_object_add_value_float(obj, "util", util);
+
+	/*
+	 * If the device has slaves, aggregate the stats for
+	 * those slave devices also.
+	 */
+	if (!agg->slavecount)
+		return;
+	json_object_add_value_int(obj, "aggr_read_ios",
+				agg->ios[0] / agg->slavecount);
+	json_object_add_value_int(obj, "aggr_write_ios",
+				agg->ios[1] / agg->slavecount);
+	json_object_add_value_int(obj, "aggr_read_merges",
+				agg->merges[0] / agg->slavecount);
+	json_object_add_value_int(obj, "aggr_write_merge",
+				agg->merges[1] / agg->slavecount);
+	json_object_add_value_int(obj, "aggr_read_ticks",
+				agg->ticks[0] / agg->slavecount);
+	json_object_add_value_int(obj, "aggr_write_ticks",
+				agg->ticks[1] / agg->slavecount);
+	json_object_add_value_int(obj, "aggr_in_queue",
+				agg->time_in_queue / agg->slavecount);
+	json_object_add_value_float(obj, "aggr_util", agg->max_util.u.f);
+}
+
+static void json_object_add_disk_utils(struct json_object *obj,
+				       struct flist_head *head)
+{
+	struct json_array *array = json_create_array();
+	struct flist_head *entry;
+	struct disk_util *du;
+
+	json_object_add_value_array(obj, "disk_util", array);
+
+	flist_for_each(entry, head) {
+		du = flist_entry(entry, struct disk_util, list);
+
+		aggregate_slaves_stats(du);
+		json_array_add_disk_util(&du->dus, &du->agg, array);
+	}
+}
+
+void show_disk_util(int terse, struct json_object *parent,
+		    struct buf_output *out)
+{
+	struct flist_head *entry;
+	struct disk_util *du;
+	bool do_json;
+
+	if (!is_running_backend())
+		return;
+
+	if (flist_empty(&disk_list)) {
+		return;
+	}
+
+	if ((output_format & FIO_OUTPUT_JSON) && parent)
+		do_json = true;
+	else
+		do_json = false;
+
+	if (!terse && !do_json)
+		log_buf(out, "\nDisk stats (read/write):\n");
+
+	if (do_json)
+		json_object_add_disk_utils(parent, &disk_list);
+	else if (output_format & ~(FIO_OUTPUT_JSON | FIO_OUTPUT_JSON_PLUS)) {
+		flist_for_each(entry, &disk_list) {
+			du = flist_entry(entry, struct disk_util, list);
+
+			aggregate_slaves_stats(du);
+			print_disk_util(&du->dus, &du->agg, terse, out);
+		}
+	}
+}
+
 static void show_thread_status_normal(struct thread_stat *ts,
 				      struct group_run_stats *rs,
 				      struct buf_output *out)
@@ -797,7 +1058,7 @@ static void show_thread_status_normal(struct thread_stat *ts,
 
 	if (!ddir_rw_sum(ts->io_bytes) && !ddir_rw_sum(ts->total_io_u))
 		return;
-		
+
 	memset(time_buf, 0, sizeof(time_buf));
 
 	time(&time_p);
@@ -817,12 +1078,10 @@ static void show_thread_status_normal(struct thread_stat *ts,
 	if (strlen(ts->description))
 		log_buf(out, "  Description  : [%s]\n", ts->description);
 
-	if (ts->io_bytes[DDIR_READ])
-		show_ddir_status(rs, ts, DDIR_READ, out);
-	if (ts->io_bytes[DDIR_WRITE])
-		show_ddir_status(rs, ts, DDIR_WRITE, out);
-	if (ts->io_bytes[DDIR_TRIM])
-		show_ddir_status(rs, ts, DDIR_TRIM, out);
+	for_each_rw_ddir(ddir) {
+		if (ts->io_bytes[ddir])
+			show_ddir_status(rs, ts, ddir, out);
+	}
 
 	show_latencies(ts, out);
 
@@ -934,12 +1193,17 @@ static void show_ddir_status_terse(struct thread_stat *ts,
 	else
 		log_buf(out, ";%llu;%llu;%f;%f", 0ULL, 0ULL, 0.0, 0.0);
 
-	if (ts->clat_percentiles || ts->lat_percentiles) {
-		len = calc_clat_percentiles(ts->io_u_plat[ddir],
+	if (ts->lat_percentiles)
+		len = calc_clat_percentiles(ts->io_u_plat[FIO_LAT][ddir],
+					ts->lat_stat[ddir].samples,
+					ts->percentile_list, &ovals, &maxv,
+					&minv);
+	else if (ts->clat_percentiles)
+		len = calc_clat_percentiles(ts->io_u_plat[FIO_CLAT][ddir],
 					ts->clat_stat[ddir].samples,
 					ts->percentile_list, &ovals, &maxv,
 					&minv);
-	} else
+	else
 		len = 0;
 
 	for (i = 0; i < FIO_IO_U_LIST_MAX_LEN; i++) {
@@ -955,8 +1219,7 @@ static void show_ddir_status_terse(struct thread_stat *ts,
 	else
 		log_buf(out, ";%llu;%llu;%f;%f", 0ULL, 0ULL, 0.0, 0.0);
 
-	if (ovals)
-		free(ovals);
+	free(ovals);
 
 	bw_stat = calc_lat(&ts->bw_stat[ddir], &min, &max, &mean, &dev);
 	if (bw_stat) {
@@ -986,17 +1249,63 @@ static void show_ddir_status_terse(struct thread_stat *ts,
 	}
 }
 
+static struct json_object *add_ddir_lat_json(struct thread_stat *ts, uint32_t percentiles,
+		struct io_stat *lat_stat, uint64_t *io_u_plat)
+{
+	char buf[120];
+	double mean, dev;
+	unsigned int i, len;
+	struct json_object *lat_object, *percentile_object, *clat_bins_object;
+	unsigned long long min, max, maxv, minv, *ovals = NULL;
+
+	if (!calc_lat(lat_stat, &min, &max, &mean, &dev)) {
+		min = max = 0;
+		mean = dev = 0.0;
+	}
+	lat_object = json_create_object();
+	json_object_add_value_int(lat_object, "min", min);
+	json_object_add_value_int(lat_object, "max", max);
+	json_object_add_value_float(lat_object, "mean", mean);
+	json_object_add_value_float(lat_object, "stddev", dev);
+	json_object_add_value_int(lat_object, "N", lat_stat->samples);
+
+	if (percentiles && lat_stat->samples) {
+		len = calc_clat_percentiles(io_u_plat, lat_stat->samples,
+				ts->percentile_list, &ovals, &maxv, &minv);
+
+		if (len > FIO_IO_U_LIST_MAX_LEN)
+			len = FIO_IO_U_LIST_MAX_LEN;
+
+		percentile_object = json_create_object();
+		json_object_add_value_object(lat_object, "percentile", percentile_object);
+		for (i = 0; i < len; i++) {
+			snprintf(buf, sizeof(buf), "%f", ts->percentile_list[i].u.f);
+			json_object_add_value_int(percentile_object, buf, ovals[i]);
+		}
+		free(ovals);
+
+		if (output_format & FIO_OUTPUT_JSON_PLUS) {
+			clat_bins_object = json_create_object();
+			json_object_add_value_object(lat_object, "bins", clat_bins_object);
+
+			for(i = 0; i < FIO_IO_U_PLAT_NR; i++)
+				if (io_u_plat[i]) {
+					snprintf(buf, sizeof(buf), "%llu", plat_idx_to_val(i));
+					json_object_add_value_int(clat_bins_object, buf, io_u_plat[i]);
+				}
+		}
+	}
+
+	return lat_object;
+}
+
 static void add_ddir_status_json(struct thread_stat *ts,
 		struct group_run_stats *rs, int ddir, struct json_object *parent)
 {
-	unsigned long long min, max, minv, maxv;
+	unsigned long long min, max;
 	unsigned long long bw_bytes, bw;
-	unsigned long long *ovals = NULL;
 	double mean, dev, iops;
-	unsigned int len;
-	int i;
-	struct json_object *dir_object, *tmp_object, *percentile_object, *clat_bins_object = NULL;
-	char buf[120];
+	struct json_object *dir_object, *tmp_object;
 	double p_of_agg = 100.0;
 
 	assert(ddir_rw(ddir) || ddir_sync(ddir));
@@ -1030,125 +1339,55 @@ static void add_ddir_status_json(struct thread_stat *ts,
 		json_object_add_value_int(dir_object, "short_ios", ts->short_io_u[ddir]);
 		json_object_add_value_int(dir_object, "drop_ios", ts->drop_io_u[ddir]);
 
-		if (!calc_lat(&ts->slat_stat[ddir], &min, &max, &mean, &dev)) {
-			min = max = 0;
-			mean = dev = 0.0;
-		}
-		tmp_object = json_create_object();
+		tmp_object = add_ddir_lat_json(ts, ts->slat_percentiles,
+				&ts->slat_stat[ddir], ts->io_u_plat[FIO_SLAT][ddir]);
 		json_object_add_value_object(dir_object, "slat_ns", tmp_object);
-		json_object_add_value_int(tmp_object, "min", min);
-		json_object_add_value_int(tmp_object, "max", max);
-		json_object_add_value_float(tmp_object, "mean", mean);
-		json_object_add_value_float(tmp_object, "stddev", dev);
-
-		if (!calc_lat(&ts->clat_stat[ddir], &min, &max, &mean, &dev)) {
-			min = max = 0;
-			mean = dev = 0.0;
-		}
-		tmp_object = json_create_object();
+
+		tmp_object = add_ddir_lat_json(ts, ts->clat_percentiles,
+				&ts->clat_stat[ddir], ts->io_u_plat[FIO_CLAT][ddir]);
 		json_object_add_value_object(dir_object, "clat_ns", tmp_object);
-		json_object_add_value_int(tmp_object, "min", min);
-		json_object_add_value_int(tmp_object, "max", max);
-		json_object_add_value_float(tmp_object, "mean", mean);
-		json_object_add_value_float(tmp_object, "stddev", dev);
-	} else {
-		if (!calc_lat(&ts->sync_stat, &min, &max, &mean, &dev)) {
-			min = max = 0;
-			mean = dev = 0.0;
-		}
 
-		tmp_object = json_create_object();
+		tmp_object = add_ddir_lat_json(ts, ts->lat_percentiles,
+				&ts->lat_stat[ddir], ts->io_u_plat[FIO_LAT][ddir]);
 		json_object_add_value_object(dir_object, "lat_ns", tmp_object);
+	} else {
 		json_object_add_value_int(dir_object, "total_ios", ts->total_io_u[DDIR_SYNC]);
-		json_object_add_value_int(tmp_object, "min", min);
-		json_object_add_value_int(tmp_object, "max", max);
-		json_object_add_value_float(tmp_object, "mean", mean);
-		json_object_add_value_float(tmp_object, "stddev", dev);
+		tmp_object = add_ddir_lat_json(ts, ts->lat_percentiles | ts->clat_percentiles,
+				&ts->sync_stat, ts->io_u_sync_plat);
+		json_object_add_value_object(dir_object, "lat_ns", tmp_object);
 	}
 
-	if (ts->clat_percentiles || ts->lat_percentiles) {
-		if (ddir_rw(ddir)) {
-			uint64_t samples;
+	if (!ddir_rw(ddir))
+		return;
 
-			if (ts->clat_percentiles)
-				samples = ts->clat_stat[ddir].samples;
-			else
-				samples = ts->lat_stat[ddir].samples;
+	/* Only print PRIO latencies if some high priority samples were gathered */
+	if (ts->clat_high_prio_stat[ddir].samples > 0) {
+		const char *high, *low;
 
-			len = calc_clat_percentiles(ts->io_u_plat[ddir],
-					samples, ts->percentile_list, &ovals,
-					&maxv, &minv);
+		if (ts->lat_percentiles) {
+			high = "lat_high_prio";
+			low = "lat_low_prio";
 		} else {
-			len = calc_clat_percentiles(ts->io_u_sync_plat,
-					ts->sync_stat.samples,
-					ts->percentile_list, &ovals, &maxv,
-					&minv);
-		}
-
-		if (len > FIO_IO_U_LIST_MAX_LEN)
-			len = FIO_IO_U_LIST_MAX_LEN;
-	} else
-		len = 0;
-
-	percentile_object = json_create_object();
-	if (ts->clat_percentiles)
-		json_object_add_value_object(tmp_object, "percentile", percentile_object);
-	for (i = 0; i < len; i++) {
-		snprintf(buf, sizeof(buf), "%f", ts->percentile_list[i].u.f);
-		json_object_add_value_int(percentile_object, (const char *)buf, ovals[i]);
-	}
-
-	if (output_format & FIO_OUTPUT_JSON_PLUS) {
-		clat_bins_object = json_create_object();
-		if (ts->clat_percentiles)
-			json_object_add_value_object(tmp_object, "bins", clat_bins_object);
-
-		for(i = 0; i < FIO_IO_U_PLAT_NR; i++) {
-			if (ddir_rw(ddir)) {
-				if (ts->io_u_plat[ddir][i]) {
-					snprintf(buf, sizeof(buf), "%llu", plat_idx_to_val(i));
-					json_object_add_value_int(clat_bins_object, (const char *)buf, ts->io_u_plat[ddir][i]);
-				}
-			} else {
-				if (ts->io_u_sync_plat[i]) {
-					snprintf(buf, sizeof(buf), "%llu", plat_idx_to_val(i));
-					json_object_add_value_int(clat_bins_object, (const char *)buf, ts->io_u_sync_plat[i]);
-				}
-			}
+			high = "clat_high_prio";
+			low = "clat_low_prio";
 		}
-	}
 
-	if (!ddir_rw(ddir))
-		return;
+		tmp_object = add_ddir_lat_json(ts, ts->clat_percentiles | ts->lat_percentiles,
+				&ts->clat_high_prio_stat[ddir], ts->io_u_plat_high_prio[ddir]);
+		json_object_add_value_object(dir_object, high, tmp_object);
 
-	if (!calc_lat(&ts->lat_stat[ddir], &min, &max, &mean, &dev)) {
-		min = max = 0;
-		mean = dev = 0.0;
+		tmp_object = add_ddir_lat_json(ts, ts->clat_percentiles | ts->lat_percentiles,
+				&ts->clat_low_prio_stat[ddir], ts->io_u_plat_low_prio[ddir]);
+		json_object_add_value_object(dir_object, low, tmp_object);
 	}
-	tmp_object = json_create_object();
-	json_object_add_value_object(dir_object, "lat_ns", tmp_object);
-	json_object_add_value_int(tmp_object, "min", min);
-	json_object_add_value_int(tmp_object, "max", max);
-	json_object_add_value_float(tmp_object, "mean", mean);
-	json_object_add_value_float(tmp_object, "stddev", dev);
-	if (ts->lat_percentiles)
-		json_object_add_value_object(tmp_object, "percentile", percentile_object);
-	if (output_format & FIO_OUTPUT_JSON_PLUS && ts->lat_percentiles)
-		json_object_add_value_object(tmp_object, "bins", clat_bins_object);
-
-	if (ovals)
-		free(ovals);
 
 	if (calc_lat(&ts->bw_stat[ddir], &min, &max, &mean, &dev)) {
-		if (rs->agg[ddir]) {
-			p_of_agg = mean * 100 / (double) (rs->agg[ddir] / 1024);
-			if (p_of_agg > 100.0)
-				p_of_agg = 100.0;
-		}
+		p_of_agg = convert_agg_kbytes_percent(rs, ddir, mean);
 	} else {
 		min = max = 0;
 		p_of_agg = mean = dev = 0.0;
 	}
+
 	json_object_add_value_int(dir_object, "bw_min", min);
 	json_object_add_value_int(dir_object, "bw_max", max);
 	json_object_add_value_float(dir_object, "bw_agg", p_of_agg);
@@ -1269,12 +1508,8 @@ static void json_add_job_opts(struct json_object *root, const char *name,
 	json_object_add_value_object(root, name, dir_object);
 
 	flist_for_each(entry, opt_list) {
-		const char *pos = "";
-
 		p = flist_entry(entry, struct print_option, list);
-		if (p->value)
-			pos = p->value;
-		json_object_add_value_string(dir_object, p->name, pos);
+		json_object_add_value_string(dir_object, p->name, p->value);
 	}
 }
 
@@ -1448,7 +1683,7 @@ static struct json_object *show_thread_status_json(struct thread_stat *ts,
 				snprintf(buf, sizeof(buf), "%f",
 					 ts->percentile_list[i].u.f);
 				json_object_add_value_int(percentile_object,
-							  (const char *)buf,
+							  buf,
 							  percentiles[i]);
 			}
 
@@ -1637,11 +1872,13 @@ void sum_group_stats(struct group_run_stats *dst, struct group_run_stats *src)
 void sum_thread_stats(struct thread_stat *dst, struct thread_stat *src,
 		      bool first)
 {
-	int l, k;
+	int k, l, m;
 
 	for (l = 0; l < DDIR_RWDIR_CNT; l++) {
 		if (!dst->unified_rw_rep) {
 			sum_stat(&dst->clat_stat[l], &src->clat_stat[l], first, false);
+			sum_stat(&dst->clat_high_prio_stat[l], &src->clat_high_prio_stat[l], first, false);
+			sum_stat(&dst->clat_low_prio_stat[l], &src->clat_low_prio_stat[l], first, false);
 			sum_stat(&dst->slat_stat[l], &src->slat_stat[l], first, false);
 			sum_stat(&dst->lat_stat[l], &src->lat_stat[l], first, false);
 			sum_stat(&dst->bw_stat[l], &src->bw_stat[l], first, true);
@@ -1653,6 +1890,8 @@ void sum_thread_stats(struct thread_stat *dst, struct thread_stat *src,
 				dst->runtime[l] = src->runtime[l];
 		} else {
 			sum_stat(&dst->clat_stat[0], &src->clat_stat[l], first, false);
+			sum_stat(&dst->clat_high_prio_stat[0], &src->clat_high_prio_stat[l], first, false);
+			sum_stat(&dst->clat_low_prio_stat[0], &src->clat_low_prio_stat[l], first, false);
 			sum_stat(&dst->slat_stat[0], &src->slat_stat[l], first, false);
 			sum_stat(&dst->lat_stat[0], &src->lat_stat[l], first, false);
 			sum_stat(&dst->bw_stat[0], &src->bw_stat[l], first, true);
@@ -1691,9 +1930,6 @@ void sum_thread_stats(struct thread_stat *dst, struct thread_stat *src,
 	for (k = 0; k < FIO_IO_U_LAT_M_NR; k++)
 		dst->io_u_lat_m[k] += src->io_u_lat_m[k];
 
-	for (k = 0; k < FIO_IO_U_PLAT_NR; k++)
-		dst->io_u_sync_plat[k] += src->io_u_sync_plat[k];
-
 	for (k = 0; k < DDIR_RWDIR_CNT; k++) {
 		if (!dst->unified_rw_rep) {
 			dst->total_io_u[k] += src->total_io_u[k];
@@ -1708,14 +1944,27 @@ void sum_thread_stats(struct thread_stat *dst, struct thread_stat *src,
 
 	dst->total_io_u[DDIR_SYNC] += src->total_io_u[DDIR_SYNC];
 
-	for (k = 0; k < DDIR_RWDIR_CNT; k++) {
-		int m;
+	for (k = 0; k < FIO_LAT_CNT; k++)
+		for (l = 0; l < DDIR_RWDIR_CNT; l++)
+			for (m = 0; m < FIO_IO_U_PLAT_NR; m++)
+				if (!dst->unified_rw_rep)
+					dst->io_u_plat[k][l][m] += src->io_u_plat[k][l][m];
+				else
+					dst->io_u_plat[k][0][m] += src->io_u_plat[k][l][m];
 
+	for (k = 0; k < FIO_IO_U_PLAT_NR; k++)
+		dst->io_u_sync_plat[k] += src->io_u_sync_plat[k];
+
+	for (k = 0; k < DDIR_RWDIR_CNT; k++) {
 		for (m = 0; m < FIO_IO_U_PLAT_NR; m++) {
-			if (!dst->unified_rw_rep)
-				dst->io_u_plat[k][m] += src->io_u_plat[k][m];
-			else
-				dst->io_u_plat[0][m] += src->io_u_plat[k][m];
+			if (!dst->unified_rw_rep) {
+				dst->io_u_plat_high_prio[k][m] += src->io_u_plat_high_prio[k][m];
+				dst->io_u_plat_low_prio[k][m] += src->io_u_plat_low_prio[k][m];
+			} else {
+				dst->io_u_plat_high_prio[0][m] += src->io_u_plat_high_prio[k][m];
+				dst->io_u_plat_low_prio[0][m] += src->io_u_plat_low_prio[k][m];
+			}
+
 		}
 	}
 
@@ -1748,6 +1997,8 @@ void init_thread_stat(struct thread_stat *ts)
 		ts->slat_stat[j].min_val = -1UL;
 		ts->bw_stat[j].min_val = -1UL;
 		ts->iops_stat[j].min_val = -1UL;
+		ts->clat_high_prio_stat[j].min_val = -1UL;
+		ts->clat_low_prio_stat[j].min_val = -1UL;
 	}
 	ts->sync_stat.min_val = -1UL;
 	ts->groupid = -1;
@@ -1817,6 +2068,7 @@ void __show_run_stats(void)
 
 		ts->clat_percentiles = td->o.clat_percentiles;
 		ts->lat_percentiles = td->o.lat_percentiles;
+		ts->slat_percentiles = td->o.slat_percentiles;
 		ts->percentile_precision = td->o.percentile_precision;
 		memcpy(ts->percentile_list, td->o.percentile_list, sizeof(td->o.percentile_list));
 		opt_lists[j] = &td->opt_list;
@@ -1828,10 +2080,11 @@ void __show_run_stats(void)
 			/*
 			 * These are per-group shared already
 			 */
-			strncpy(ts->name, td->o.name, FIO_JOBNAME_SIZE - 1);
+			snprintf(ts->name, sizeof(ts->name), "%s", td->o.name);
 			if (td->o.description)
-				strncpy(ts->description, td->o.description,
-						FIO_JOBDESC_SIZE - 1);
+				snprintf(ts->description,
+					 sizeof(ts->description), "%s",
+					 td->o.description);
 			else
 				memset(ts->description, 0, FIO_JOBDESC_SIZE);
 
@@ -1868,12 +2121,12 @@ void __show_run_stats(void)
 			if (!td->error && td->o.continue_on_error &&
 			    td->first_error) {
 				ts->error = td->first_error;
-				ts->verror[sizeof(ts->verror) - 1] = '\0';
-				strncpy(ts->verror, td->verror, sizeof(ts->verror) - 1);
+				snprintf(ts->verror, sizeof(ts->verror), "%s",
+					 td->verror);
 			} else  if (td->error) {
 				ts->error = td->error;
-				ts->verror[sizeof(ts->verror) - 1] = '\0';
-				strncpy(ts->verror, td->verror, sizeof(ts->verror) - 1);
+				snprintf(ts->verror, sizeof(ts->verror), "%s",
+					 td->verror);
 			}
 		}
 
@@ -2046,7 +2299,7 @@ void __show_run_stats(void)
 	free(opt_lists);
 }
 
-void __show_running_run_stats(void)
+int __show_running_run_stats(void)
 {
 	struct thread_data *td;
 	unsigned long long *rt;
@@ -2060,9 +2313,9 @@ void __show_running_run_stats(void)
 
 	for_each_td(td, i) {
 		td->update_rusage = 1;
-		td->ts.io_bytes[DDIR_READ] = td->io_bytes[DDIR_READ];
-		td->ts.io_bytes[DDIR_WRITE] = td->io_bytes[DDIR_WRITE];
-		td->ts.io_bytes[DDIR_TRIM] = td->io_bytes[DDIR_TRIM];
+		for_each_rw_ddir(ddir) {
+			td->ts.io_bytes[ddir] = td->io_bytes[ddir];
+		}
 		td->ts.total_run_time = mtime_since(&td->epoch, &ts);
 
 		rt[i] = mtime_since(&td->start, &ts);
@@ -2097,10 +2350,10 @@ void __show_running_run_stats(void)
 
 	free(rt);
 	fio_sem_up(stat_sem);
+
+	return 0;
 }
 
-static bool status_interval_init;
-static struct timespec status_time;
 static bool status_file_disabled;
 
 #define FIO_STATUS_FILE		"fio-dump-status"
@@ -2122,6 +2375,9 @@ static int check_status_file(void)
 	}
 	if (temp_dir == NULL)
 		temp_dir = "/tmp";
+#ifdef __COVERITY__
+	__coverity_tainted_data_sanitize__(temp_dir);
+#endif
 
 	snprintf(fio_status_file_path, sizeof(fio_status_file_path), "%s/%s", temp_dir, FIO_STATUS_FILE);
 
@@ -2140,16 +2396,6 @@ static int check_status_file(void)
 
 void check_for_running_stats(void)
 {
-	if (status_interval) {
-		if (!status_interval_init) {
-			fio_gettime(&status_time, NULL);
-			status_interval_init = true;
-		} else if (mtime_since_now(&status_time) >= status_interval) {
-			show_running_run_stats();
-			fio_gettime(&status_time, NULL);
-			return;
-		}
-	}
 	if (check_status_file()) {
 		show_running_run_stats();
 		return;
@@ -2290,6 +2536,14 @@ void regrow_logs(struct thread_data *td)
 	td->flags &= ~TD_F_REGROW_LOGS;
 }
 
+void regrow_agg_logs(void)
+{
+	enum fio_ddir ddir;
+
+	for (ddir = 0; ddir < DDIR_RWDIR_CNT; ddir++)
+		regrow_log(agg_io_log[ddir]);
+}
+
 static struct io_logs *get_cur_log(struct io_log *iolog)
 {
 	struct io_logs *cur_log;
@@ -2324,7 +2578,7 @@ static struct io_logs *get_cur_log(struct io_log *iolog)
 
 static void __add_log_sample(struct io_log *iolog, union io_sample_data data,
 			     enum fio_ddir ddir, unsigned long long bs,
-			     unsigned long t, uint64_t offset)
+			     unsigned long t, uint64_t offset, uint8_t priority_bit)
 {
 	struct io_logs *cur_log;
 
@@ -2343,6 +2597,7 @@ static void __add_log_sample(struct io_log *iolog, union io_sample_data data,
 		s->time = t + (iolog->td ? iolog->td->unix_epoch : 0);
 		io_sample_set_ddir(iolog, s, ddir);
 		s->bs = bs;
+		s->priority_bit = priority_bit;
 
 		if (iolog->log_offset) {
 			struct io_sample_offset *so = (void *) s;
@@ -2367,9 +2622,11 @@ static inline void reset_io_stat(struct io_stat *ios)
 void reset_io_stats(struct thread_data *td)
 {
 	struct thread_stat *ts = &td->ts;
-	int i, j;
+	int i, j, k;
 
 	for (i = 0; i < DDIR_RWDIR_CNT; i++) {
+		reset_io_stat(&ts->clat_high_prio_stat[i]);
+		reset_io_stat(&ts->clat_low_prio_stat[i]);
 		reset_io_stat(&ts->clat_stat[i]);
 		reset_io_stat(&ts->slat_stat[i]);
 		reset_io_stat(&ts->lat_stat[i]);
@@ -2383,12 +2640,18 @@ void reset_io_stats(struct thread_data *td)
 		ts->drop_io_u[i] = 0;
 
 		for (j = 0; j < FIO_IO_U_PLAT_NR; j++) {
-			ts->io_u_plat[i][j] = 0;
+			ts->io_u_plat_high_prio[i][j] = 0;
+			ts->io_u_plat_low_prio[i][j] = 0;
 			if (!i)
 				ts->io_u_sync_plat[j] = 0;
 		}
 	}
 
+	for (i = 0; i < FIO_LAT_CNT; i++)
+		for (j = 0; j < DDIR_RWDIR_CNT; j++)
+			for (k = 0; k < FIO_IO_U_PLAT_NR; k++)
+				ts->io_u_plat[i][j][k] = 0;
+
 	ts->total_io_u[DDIR_SYNC] = 0;
 
 	for (i = 0; i < FIO_IO_U_MAP_NR; i++) {
@@ -2411,7 +2674,7 @@ void reset_io_stats(struct thread_data *td)
 }
 
 static void __add_stat_to_log(struct io_log *iolog, enum fio_ddir ddir,
-			      unsigned long elapsed, bool log_max)
+			      unsigned long elapsed, bool log_max, uint8_t priority_bit)
 {
 	/*
 	 * Note an entry in the log. Use the mean from the logged samples,
@@ -2426,26 +2689,26 @@ static void __add_stat_to_log(struct io_log *iolog, enum fio_ddir ddir,
 		else
 			data.val = iolog->avg_window[ddir].mean.u.f + 0.50;
 
-		__add_log_sample(iolog, data, ddir, 0, elapsed, 0);
+		__add_log_sample(iolog, data, ddir, 0, elapsed, 0, priority_bit);
 	}
 
 	reset_io_stat(&iolog->avg_window[ddir]);
 }
 
 static void _add_stat_to_log(struct io_log *iolog, unsigned long elapsed,
-			     bool log_max)
+			     bool log_max, uint8_t priority_bit)
 {
 	int ddir;
 
 	for (ddir = 0; ddir < DDIR_RWDIR_CNT; ddir++)
-		__add_stat_to_log(iolog, ddir, elapsed, log_max);
+		__add_stat_to_log(iolog, ddir, elapsed, log_max, priority_bit);
 }
 
 static unsigned long add_log_sample(struct thread_data *td,
 				    struct io_log *iolog,
 				    union io_sample_data data,
 				    enum fio_ddir ddir, unsigned long long bs,
-				    uint64_t offset)
+				    uint64_t offset, uint8_t priority_bit)
 {
 	unsigned long elapsed, this_window;
 
@@ -2458,7 +2721,7 @@ static unsigned long add_log_sample(struct thread_data *td,
 	 * If no time averaging, just add the log sample.
 	 */
 	if (!iolog->avg_msec) {
-		__add_log_sample(iolog, data, ddir, bs, elapsed, offset);
+		__add_log_sample(iolog, data, ddir, bs, elapsed, offset, priority_bit);
 		return 0;
 	}
 
@@ -2482,9 +2745,10 @@ static unsigned long add_log_sample(struct thread_data *td,
 			return diff;
 	}
 
-	__add_stat_to_log(iolog, ddir, elapsed, td->o.log_max != 0);
+	__add_stat_to_log(iolog, ddir, elapsed, td->o.log_max != 0, priority_bit);
+
+	iolog->avg_last[ddir] = elapsed - (elapsed % iolog->avg_msec);
 
-	iolog->avg_last[ddir] = elapsed - (this_window - iolog->avg_msec);
 	return iolog->avg_msec;
 }
 
@@ -2495,18 +2759,19 @@ void finalize_logs(struct thread_data *td, bool unit_logs)
 	elapsed = mtime_since_now(&td->epoch);
 
 	if (td->clat_log && unit_logs)
-		_add_stat_to_log(td->clat_log, elapsed, td->o.log_max != 0);
+		_add_stat_to_log(td->clat_log, elapsed, td->o.log_max != 0, 0);
 	if (td->slat_log && unit_logs)
-		_add_stat_to_log(td->slat_log, elapsed, td->o.log_max != 0);
+		_add_stat_to_log(td->slat_log, elapsed, td->o.log_max != 0, 0);
 	if (td->lat_log && unit_logs)
-		_add_stat_to_log(td->lat_log, elapsed, td->o.log_max != 0);
+		_add_stat_to_log(td->lat_log, elapsed, td->o.log_max != 0, 0);
 	if (td->bw_log && (unit_logs == per_unit_log(td->bw_log)))
-		_add_stat_to_log(td->bw_log, elapsed, td->o.log_max != 0);
+		_add_stat_to_log(td->bw_log, elapsed, td->o.log_max != 0, 0);
 	if (td->iops_log && (unit_logs == per_unit_log(td->iops_log)))
-		_add_stat_to_log(td->iops_log, elapsed, td->o.log_max != 0);
+		_add_stat_to_log(td->iops_log, elapsed, td->o.log_max != 0, 0);
 }
 
-void add_agg_sample(union io_sample_data data, enum fio_ddir ddir, unsigned long long bs)
+void add_agg_sample(union io_sample_data data, enum fio_ddir ddir, unsigned long long bs,
+					uint8_t priority_bit)
 {
 	struct io_log *iolog;
 
@@ -2514,7 +2779,7 @@ void add_agg_sample(union io_sample_data data, enum fio_ddir ddir, unsigned long
 		return;
 
 	iolog = agg_io_log[ddir];
-	__add_log_sample(iolog, data, ddir, bs, mtime_since_genesis(), 0);
+	__add_log_sample(iolog, data, ddir, bs, mtime_since_genesis(), 0, priority_bit);
 }
 
 void add_sync_clat_sample(struct thread_stat *ts, unsigned long long nsec)
@@ -2526,18 +2791,32 @@ void add_sync_clat_sample(struct thread_stat *ts, unsigned long long nsec)
 	add_stat_sample(&ts->sync_stat, nsec);
 }
 
-static void add_clat_percentile_sample(struct thread_stat *ts,
-				unsigned long long nsec, enum fio_ddir ddir)
+static void add_lat_percentile_sample_noprio(struct thread_stat *ts,
+				unsigned long long nsec, enum fio_ddir ddir, enum fio_lat lat)
 {
 	unsigned int idx = plat_val_to_idx(nsec);
 	assert(idx < FIO_IO_U_PLAT_NR);
 
-	ts->io_u_plat[ddir][idx]++;
+	ts->io_u_plat[lat][ddir][idx]++;
+}
+
+static void add_lat_percentile_sample(struct thread_stat *ts,
+				unsigned long long nsec, enum fio_ddir ddir, uint8_t priority_bit,
+				enum fio_lat lat)
+{
+	unsigned int idx = plat_val_to_idx(nsec);
+
+	add_lat_percentile_sample_noprio(ts, nsec, ddir, lat);
+
+	if (!priority_bit)
+		ts->io_u_plat_low_prio[ddir][idx]++;
+	else
+		ts->io_u_plat_high_prio[ddir][idx]++;
 }
 
 void add_clat_sample(struct thread_data *td, enum fio_ddir ddir,
 		     unsigned long long nsec, unsigned long long bs,
-		     uint64_t offset)
+		     uint64_t offset, uint8_t priority_bit)
 {
 	const bool needs_lock = td_async_processing(td);
 	unsigned long elapsed, this_window;
@@ -2549,12 +2828,23 @@ void add_clat_sample(struct thread_data *td, enum fio_ddir ddir,
 
 	add_stat_sample(&ts->clat_stat[ddir], nsec);
 
+	if (!ts->lat_percentiles) {
+		if (priority_bit)
+			add_stat_sample(&ts->clat_high_prio_stat[ddir], nsec);
+		else
+			add_stat_sample(&ts->clat_low_prio_stat[ddir], nsec);
+	}
+
 	if (td->clat_log)
 		add_log_sample(td, td->clat_log, sample_val(nsec), ddir, bs,
-			       offset);
+			       offset, priority_bit);
 
-	if (ts->clat_percentiles)
-		add_clat_percentile_sample(ts, nsec, ddir);
+	if (ts->clat_percentiles) {
+		if (ts->lat_percentiles)
+			add_lat_percentile_sample_noprio(ts, nsec, ddir, FIO_CLAT);
+		else
+			add_lat_percentile_sample(ts, nsec, ddir, priority_bit, FIO_CLAT);
+	}
 
 	if (iolog && iolog->hist_msec) {
 		struct io_hist *hw = &iolog->hist_window[ddir];
@@ -2564,7 +2854,7 @@ void add_clat_sample(struct thread_data *td, enum fio_ddir ddir,
 		if (!hw->hist_last)
 			hw->hist_last = elapsed;
 		this_window = elapsed - hw->hist_last;
-		
+
 		if (this_window >= iolog->hist_msec) {
 			uint64_t *io_u_plat;
 			struct io_u_plat_entry *dst;
@@ -2576,13 +2866,13 @@ void add_clat_sample(struct thread_data *td, enum fio_ddir ddir,
 			 * located in iolog.c after printing this sample to the
 			 * log file.
 			 */
-			io_u_plat = (uint64_t *) td->ts.io_u_plat[ddir];
+			io_u_plat = (uint64_t *) td->ts.io_u_plat[FIO_CLAT][ddir];
 			dst = malloc(sizeof(struct io_u_plat_entry));
 			memcpy(&(dst->io_u_plat), io_u_plat,
-				FIO_IO_U_PLAT_NR * sizeof(unsigned int));
+				FIO_IO_U_PLAT_NR * sizeof(uint64_t));
 			flist_add(&dst->list, &hw->list);
 			__add_log_sample(iolog, sample_plat(dst), ddir, bs,
-						elapsed, offset);
+						elapsed, offset, priority_bit);
 
 			/*
 			 * Update the last time we recorded as being now, minus
@@ -2599,7 +2889,8 @@ void add_clat_sample(struct thread_data *td, enum fio_ddir ddir,
 }
 
 void add_slat_sample(struct thread_data *td, enum fio_ddir ddir,
-		     unsigned long usec, unsigned long long bs, uint64_t offset)
+			unsigned long long nsec, unsigned long long bs, uint64_t offset,
+			uint8_t priority_bit)
 {
 	const bool needs_lock = td_async_processing(td);
 	struct thread_stat *ts = &td->ts;
@@ -2610,10 +2901,14 @@ void add_slat_sample(struct thread_data *td, enum fio_ddir ddir,
 	if (needs_lock)
 		__td_io_u_lock(td);
 
-	add_stat_sample(&ts->slat_stat[ddir], usec);
+	add_stat_sample(&ts->slat_stat[ddir], nsec);
 
 	if (td->slat_log)
-		add_log_sample(td, td->slat_log, sample_val(usec), ddir, bs, offset);
+		add_log_sample(td, td->slat_log, sample_val(nsec), ddir, bs, offset,
+			priority_bit);
+
+	if (ts->slat_percentiles)
+		add_lat_percentile_sample_noprio(ts, nsec, ddir, FIO_SLAT);
 
 	if (needs_lock)
 		__td_io_u_unlock(td);
@@ -2621,7 +2916,7 @@ void add_slat_sample(struct thread_data *td, enum fio_ddir ddir,
 
 void add_lat_sample(struct thread_data *td, enum fio_ddir ddir,
 		    unsigned long long nsec, unsigned long long bs,
-		    uint64_t offset)
+		    uint64_t offset, uint8_t priority_bit)
 {
 	const bool needs_lock = td_async_processing(td);
 	struct thread_stat *ts = &td->ts;
@@ -2636,11 +2931,16 @@ void add_lat_sample(struct thread_data *td, enum fio_ddir ddir,
 
 	if (td->lat_log)
 		add_log_sample(td, td->lat_log, sample_val(nsec), ddir, bs,
-			       offset);
+			       offset, priority_bit);
 
-	if (ts->lat_percentiles)
-		add_clat_percentile_sample(ts, nsec, ddir);
+	if (ts->lat_percentiles) {
+		add_lat_percentile_sample(ts, nsec, ddir, priority_bit, FIO_LAT);
+		if (priority_bit)
+			add_stat_sample(&ts->clat_high_prio_stat[ddir], nsec);
+		else
+			add_stat_sample(&ts->clat_low_prio_stat[ddir], nsec);
 
+	}
 	if (needs_lock)
 		__td_io_u_unlock(td);
 }
@@ -2664,7 +2964,7 @@ void add_bw_sample(struct thread_data *td, struct io_u *io_u,
 
 	if (td->bw_log)
 		add_log_sample(td, td->bw_log, sample_val(rate), io_u->ddir,
-			       bytes, io_u->offset);
+			       bytes, io_u->offset, io_u_is_prio(io_u));
 
 	td->stat_io_bytes[io_u->ddir] = td->this_io_bytes[io_u->ddir];
 
@@ -2686,7 +2986,7 @@ static int __add_samples(struct thread_data *td, struct timespec *parent_tv,
 	next_log = avg_time;
 
 	spent = mtime_since(parent_tv, t);
-	if (spent < avg_time && avg_time - spent >= LOG_MSEC_SLACK)
+	if (spent < avg_time && avg_time - spent > LOG_MSEC_SLACK)
 		return avg_time - spent;
 
 	if (needs_lock)
@@ -2718,14 +3018,14 @@ static int __add_samples(struct thread_data *td, struct timespec *parent_tv,
 			if (td->o.min_bs[ddir] == td->o.max_bs[ddir])
 				bs = td->o.min_bs[ddir];
 
-			next = add_log_sample(td, log, sample_val(rate), ddir, bs, 0);
+			next = add_log_sample(td, log, sample_val(rate), ddir, bs, 0, 0);
 			next_log = min(next_log, next);
 		}
 
 		stat_io_bytes[ddir] = this_io_bytes[ddir];
 	}
 
-	timespec_add_msec(parent_tv, avg_time);
+	*parent_tv = *t;
 
 	if (needs_lock)
 		__td_io_u_unlock(td);
@@ -2758,7 +3058,7 @@ void add_iops_sample(struct thread_data *td, struct io_u *io_u,
 
 	if (td->iops_log)
 		add_log_sample(td, td->iops_log, sample_val(1), io_u->ddir,
-			       bytes, io_u->offset);
+			       bytes, io_u->offset, io_u_is_prio(io_u));
 
 	td->stat_io_blocks[io_u->ddir] = td->this_io_blocks[io_u->ddir];
 
@@ -2779,13 +3079,16 @@ static int add_iops_samples(struct thread_data *td, struct timespec *t)
 int calc_log_samples(void)
 {
 	struct thread_data *td;
-	unsigned int next = ~0U, tmp;
+	unsigned int next = ~0U, tmp = 0, next_mod = 0, log_avg_msec_min = -1U;
 	struct timespec now;
 	int i;
+	long elapsed_time = 0;
 
 	fio_gettime(&now, NULL);
 
 	for_each_td(td, i) {
+		elapsed_time = mtime_since_now(&td->epoch);
+
 		if (!td->o.stats)
 			continue;
 		if (in_ramp_time(td) ||
@@ -2796,17 +3099,34 @@ int calc_log_samples(void)
 		if (!td->bw_log ||
 			(td->bw_log && !per_unit_log(td->bw_log))) {
 			tmp = add_bw_samples(td, &now);
-			if (tmp < next)
-				next = tmp;
+
+			if (td->bw_log)
+				log_avg_msec_min = min(log_avg_msec_min, (unsigned int)td->bw_log->avg_msec);
 		}
 		if (!td->iops_log ||
 			(td->iops_log && !per_unit_log(td->iops_log))) {
 			tmp = add_iops_samples(td, &now);
-			if (tmp < next)
-				next = tmp;
+
+			if (td->iops_log)
+				log_avg_msec_min = min(log_avg_msec_min, (unsigned int)td->iops_log->avg_msec);
 		}
+
+		if (tmp < next)
+			next = tmp;
 	}
 
+	/* if log_avg_msec_min has not been changed, set it to 0 */
+	if (log_avg_msec_min == -1U)
+		log_avg_msec_min = 0;
+
+	if (log_avg_msec_min == 0)
+		next_mod = elapsed_time;
+	else
+		next_mod = elapsed_time % log_avg_msec_min;
+
+	/* correction to keep the time on the log avg msec boundary */
+	next = min(next, (log_avg_msec_min - next_mod));
+
 	return next == ~0U ? 0 : next;
 }
 
@@ -2843,3 +3163,4 @@ uint32_t *io_u_block_info(struct thread_data *td, struct io_u *io_u)
 	assert(idx < td->ts.nr_block_infos);
 	return info;
 }
+