X-Git-Url: https://git.kernel.dk/?p=fio.git;a=blobdiff_plain;f=backend.c;h=936203dcb1c4858b1b3f7ddea23ab5f965e04bb4;hp=f541676c1d71607e8f81a315d3a2d505bb7d0c53;hb=HEAD;hpb=da8f124f8cf55eedab9cd2d3bd8afc0cd4971295

diff --git a/backend.c b/backend.c
index f541676c..fe03eab3 100644
--- a/backend.c
+++ b/backend.c
@@ -49,6 +49,7 @@
 #include "helper_thread.h"
 #include "pshared.h"
 #include "zone-dist.h"
+#include "fio_time.h"
 
 static struct fio_sem *startup_sem;
 static struct flist_head *cgroup_list;
@@ -466,7 +467,7 @@ int io_queue_event(struct thread_data *td, struct io_u *io_u, int *ret,
 				if (!from_verify)
 					unlog_io_piece(td, io_u);
 				td_verror(td, EIO, "full resid");
-				put_io_u(td, io_u);
+				clear_io_u(td, io_u);
 				break;
 			}
 
@@ -897,7 +898,16 @@ static void handle_thinktime(struct thread_data *td, enum fio_ddir ddir,
 	if (left)
 		total = usec_spin(left);
 
-	left = td->o.thinktime - total;
+	/*
+	 * usec_spin() might run for slightly longer than intended in a VM
+	 * where the vCPU could get descheduled or the hypervisor could steal
+	 * CPU time. Ensure "left" doesn't become negative.
+	 */
+	if (total < td->o.thinktime)
+		left = td->o.thinktime - total;
+	else
+		left = 0;
+
 	if (td->o.timeout) {
 		runtime_left = td->o.timeout - utime_since_now(&td->epoch);
 		if (runtime_left < (unsigned long long)left)
@@ -967,6 +977,11 @@ static void do_io(struct thread_data *td, uint64_t *bytes_done)
 	*/
 	if (td_write(td) && td_random(td) && td->o.norandommap)
 		total_bytes = max(total_bytes, (uint64_t) td->o.io_size);
+
+	/* Don't break too early if io_size > size */
+	if (td_rw(td) && !td_random(td))
+		total_bytes = max(total_bytes, (uint64_t)td->o.io_size);
+
 	/*
 	 * If verify_backlog is enabled, we'll run the verify in this
 	 * handler as well. For that case, we may need up to twice the
@@ -1124,6 +1139,9 @@ reap:
 		if (ret < 0)
 			break;
 
+		if (ddir_rw(ddir) && td->o.thinkcycles)
+			cycles_spin(td->o.thinkcycles);
+
 		if (ddir_rw(ddir) && td->o.thinktime)
 			handle_thinktime(td, ddir, &comp_time);
 
@@ -1320,7 +1338,7 @@ static int init_io_u(struct thread_data *td)
 int init_io_u_buffers(struct thread_data *td)
 {
 	struct io_u *io_u;
-	unsigned long long max_bs, min_write;
+	unsigned long long max_bs, min_write, trim_bs = 0;
 	int i, max_units;
 	int data_xfer = 1;
 	char *p;
@@ -1331,7 +1349,18 @@ int init_io_u_buffers(struct thread_data *td)
 	td->orig_buffer_size = (unsigned long long) max_bs
 					* (unsigned long long) max_units;
 
-	if (td_ioengine_flagged(td, FIO_NOIO) || !(td_read(td) || td_write(td)))
+	if (td_trim(td) && td->o.num_range > 1) {
+		trim_bs = td->o.num_range * sizeof(struct trim_range);
+		td->orig_buffer_size = trim_bs
+					* (unsigned long long) max_units;
+	}
+
+	/*
+	 * For reads, writes, and multi-range trim operations we need a
+	 * data buffer
+	 */
+	if (td_ioengine_flagged(td, FIO_NOIO) ||
+	    !(td_read(td) || td_write(td) || (td_trim(td) && td->o.num_range > 1)))
 		data_xfer = 0;
 
 	/*
@@ -1383,7 +1412,10 @@ int init_io_u_buffers(struct thread_data *td)
 				fill_verify_pattern(td, io_u->buf, max_bs, io_u, 0, 0);
 			}
 		}
-		p += max_bs;
+		if (td_trim(td) && td->o.num_range > 1)
+			p += trim_bs;
+		else
+			p += max_bs;
 	}
 
 	return 0;
@@ -1633,7 +1665,7 @@ static void *thread_main(void *data)
 	uint64_t bytes_done[DDIR_RWDIR_CNT];
 	int deadlock_loop_cnt;
 	bool clear_state;
-	int res, ret;
+	int ret;
 
 	sk_out_assign(sk_out);
 	free(fd);
@@ -1790,13 +1822,16 @@ static void *thread_main(void *data)
 
 	/* ioprio_set() has to be done before td_io_init() */
 	if (fio_option_is_set(o, ioprio) ||
-	    fio_option_is_set(o, ioprio_class)) {
-		ret = ioprio_set(IOPRIO_WHO_PROCESS, 0, o->ioprio_class, o->ioprio);
+	    fio_option_is_set(o, ioprio_class) ||
+	    fio_option_is_set(o, ioprio_hint)) {
+		ret = ioprio_set(IOPRIO_WHO_PROCESS, 0, o->ioprio_class,
+				 o->ioprio, o->ioprio_hint);
 		if (ret == -1) {
 			td_verror(td, errno, "ioprio_set");
 			goto err;
 		}
-		td->ioprio = ioprio_value(o->ioprio_class, o->ioprio);
+		td->ioprio = ioprio_value(o->ioprio_class, o->ioprio,
+					  o->ioprio_hint);
 		td->ts.ioprio = td->ioprio;
 	}
 
@@ -1846,7 +1881,7 @@ static void *thread_main(void *data)
 	if (rate_submit_init(td, sk_out))
 		goto err;
 
-	set_epoch_time(td, o->log_unix_epoch | o->log_alternate_epoch, o->log_alternate_epoch_clock_id);
+	set_epoch_time(td, o->log_alternate_epoch_clock_id, o->job_start_clock_id);
 	fio_getrusage(&td->ru_start);
 	memcpy(&td->bw_sample_time, &td->epoch, sizeof(td->epoch));
 	memcpy(&td->iops_sample_time, &td->epoch, sizeof(td->epoch));
@@ -1974,13 +2009,23 @@ static void *thread_main(void *data)
 	 * another thread is checking its io_u's for overlap
 	 */
 	if (td_offload_overlap(td)) {
-		int res = pthread_mutex_lock(&overlap_check);
-		assert(res == 0);
+		int res;
+
+		res = pthread_mutex_lock(&overlap_check);
+		if (res) {
+			td->error = errno;
+			goto err;
+		}
 	}
 	td_set_runstate(td, TD_FINISHING);
 	if (td_offload_overlap(td)) {
+		int res;
+
 		res = pthread_mutex_unlock(&overlap_check);
-		assert(res == 0);
+		if (res) {
+			td->error = errno;
+			goto err;
+		}
 	}
 
 	update_rusage_stat(td);
@@ -2054,14 +2099,14 @@ static void reap_threads(unsigned int *nr_running, uint64_t *t_rate,
 			 uint64_t *m_rate)
 {
 	unsigned int cputhreads, realthreads, pending;
-	int status, ret;
+	int ret;
 
 	/*
 	 * reap exited threads (TD_EXITED -> TD_REAPED)
 	 */
 	realthreads = pending = cputhreads = 0;
 	for_each_td(td) {
-		int flags = 0;
+		int flags = 0, status;
 
 		if (!strcmp(td->o.ioengine, "cpuio"))
 			cputhreads++;