Recent changes (master)

PATCH Jens Axboe 1 participant 1 message Jan 23, 2020 Kbuild

lore

Share & tools

lore

Jens Axboe

author

2020-01-23 13:00 UTC #0 lore raw

The following changes since commit d9b7596a1fad5adf7f6731d067e1513c56eabe96:

Merge branch 'master' of https://github.com/bvanassche/fio (2020-01-18 09:35:25 -0700)

are available in the Git repository at:

git://git.kernel.dk/fio.git master

for you to fetch changes up to 4ee9c375b64de236356cc97843d84978edeba491:

stat: ensure we align correctly (2020-01-22 19:53:14 -0700)

----------------------------------------------------------------
Jens Axboe (2):
Merge branch 'priorityQueuing' of https://github.com/Seagate/fio
stat: ensure we align correctly

Phillip Chen (2):
Whitespace standardization
Per-command priority: Priority logging and libaio/io_uring cmdprio_percentage

Su, Friendy (1):
engines: add engine for file stat

 HOWTO                          |  43 +++++---
 Makefile                       |   2 +-
 client.c                       |  10 +-
 engines/filecreate.c           |   2 +-
 engines/filestat.c             | 116 +++++++++++++++++++++
 engines/io_uring.c             |  47 +++++++++
 engines/libaio.c               |  54 ++++++++++
 eta.c                          |   6 +-
 examples/filestat-ioengine.fio |  19 ++++
 fio.1                          |  32 ++++--
 fio.h                          |   4 +-
 init.c                         |   3 +
 io_u.c                         |   8 +-
 io_u.h                         |   3 +
 ioengines.c                    |   1 +
 iolog.c                        |   8 +-
 iolog.h                        |   1 +
 server.c                       |  13 ++-
 server.h                       |   2 +-
 stat.c                         | 229 +++++++++++++++++++++++++++++++++++------
 stat.h                         |  14 ++-
 21 files changed, 541 insertions(+), 76 deletions(-)
 create mode 100644 engines/filestat.c
 create mode 100644 examples/filestat-ioengine.fio

---

Diff of recent changes:

HOWTO +30 ~~-13~~

index 41d32c04..0a366168 100644
@@ -1996,6 +1996,11 @@ I/O engine
 			set  `filesize` so that all the accounting still occurs, but no
 			actual I/O will be done other than creating the file.
 
+		**filestat**
+			Simply do stat() and do no I/O to the file. You need to set 'filesize'
+			and 'nrfiles', so that files will be created.
+			This engine is to measure file lookup and meta data access.
+
 		**libpmem**
 			Read and write using mmap I/O to a file on a filesystem
 			mounted with DAX on a persistent memory device through the PMDK

@@ -2029,21 +2034,29 @@ In addition, there are some parameters which are only valid when a specific
 with the caveat that when used on the command line, they must come after the
 :option:`ioengine` that defines them is selected.
 
-.. option:: hipri : [io_uring]
+.. option:: cmdprio_percentage=int : [io_uring] [libaio]
 
-	If this option is set, fio will attempt to use polled IO completions.
-	Normal IO completions generate interrupts to signal the completion of
-	IO, polled completions do not. Hence they are require active reaping
-	by the application. The benefits are more efficient IO for high IOPS
-	scenarios, and lower latencies for low queue depth IO.
+    Set the percentage of I/O that will be issued with higher priority by setting
+    the priority bit. Non-read I/O is likely unaffected by ``cmdprio_percentage``.
+    This option cannot be used with the `prio` or `prioclass` options. For this
+    option to set the priority bit properly, NCQ priority must be supported and
+    enabled and :option:`direct`\=1 option must be used.
 
 .. option:: fixedbufs : [io_uring]
 
-	If fio is asked to do direct IO, then Linux will map pages for each
-	IO call, and release them when IO is done. If this option is set, the
-	pages are pre-mapped before IO is started. This eliminates the need to
-	map and release for each IO. This is more efficient, and reduces the
-	IO latency as well.
+    If fio is asked to do direct IO, then Linux will map pages for each
+    IO call, and release them when IO is done. If this option is set, the
+    pages are pre-mapped before IO is started. This eliminates the need to
+    map and release for each IO. This is more efficient, and reduces the
+    IO latency as well.
+
+.. option:: hipri : [io_uring]
+
+    If this option is set, fio will attempt to use polled IO completions.
+    Normal IO completions generate interrupts to signal the completion of
+    IO, polled completions do not. Hence they are require active reaping
+    by the application. The benefits are more efficient IO for high IOPS
+    scenarios, and lower latencies for low queue depth IO.
 
 .. option:: registerfiles : [io_uring]

@@ -2687,11 +2700,15 @@ Threads, processes and job synchronization
 	Set the I/O priority value of this job. Linux limits us to a positive value
 	between 0 and 7, with 0 being the highest.  See man
 	:manpage:`ionice(1)`. Refer to an appropriate manpage for other operating
-	systems since meaning of priority may differ.
+	systems since meaning of priority may differ. For per-command priority
+	setting, see I/O engine specific `cmdprio_percentage` and `hipri_percentage`
+	options.
 
 .. option:: prioclass=int
 
-	Set the I/O priority class. See man :manpage:`ionice(1)`.
+	Set the I/O priority class. See man :manpage:`ionice(1)`. For per-command
+	priority setting, see I/O engine specific `cmdprio_percentage` and
+	`hipri_percentage` options.
 
 .. option:: cpus_allowed=str

Makefile +1 -1

index 3c5e0f5b..027b62bc 100644
@@ -45,7 +45,7 @@ SOURCE :=	$(sort $(patsubst $(SRCDIR)/%,%,$(wildcard $(SRCDIR)/crc/*.c)) \
 		pshared.c options.c \
 		smalloc.c filehash.c profile.c debug.c engines/cpu.c \
 		engines/mmap.c engines/sync.c engines/null.c engines/net.c \
-		engines/ftruncate.c engines/filecreate.c \
+		engines/ftruncate.c engines/filecreate.c engines/filestat.c \
 		server.c client.c iolog.c backend.c libfio.c flow.c cconv.c \
 		gettime-thread.c helpers.c json.c idletime.c td_error.c \
 		profiles/tiobench.c profiles/act.c io_u_queue.c filelock.c \

client.c +9 -1

index 93bca5df..4aed39e7 100644
@@ -1032,6 +1032,14 @@ static void convert_ts(struct thread_stat *dst, struct thread_stat *src)
 	dst->nr_block_infos	= le64_to_cpu(src->nr_block_infos);
 	for (i = 0; i < dst->nr_block_infos; i++)
 		dst->block_infos[i] = le32_to_cpu(src->block_infos[i]);
+	for (i = 0; i < DDIR_RWDIR_CNT; i++) {
+		for (j = 0; j < FIO_IO_U_PLAT_NR; j++) {
+			dst->io_u_plat_high_prio[i][j] = le64_to_cpu(src->io_u_plat_high_prio[i][j]);
+			dst->io_u_plat_prio[i][j] = le64_to_cpu(src->io_u_plat_prio[i][j]);
+		}
+		convert_io_stat(&dst->clat_high_prio_stat[i], &src->clat_high_prio_stat[i]);
+		convert_io_stat(&dst->clat_prio_stat[i], &src->clat_prio_stat[i]);
+	}
 
 	dst->ss_dur		= le64_to_cpu(src->ss_dur);
 	dst->ss_state		= le32_to_cpu(src->ss_state);

@@ -1693,7 +1701,7 @@ static struct cmd_iolog_pdu *convert_iolog(struct fio_net_cmd *cmd,
 
 		s->time		= le64_to_cpu(s->time);
 		s->data.val	= le64_to_cpu(s->data.val);
-		s->__ddir	= le32_to_cpu(s->__ddir);
+		s->__ddir	= __le32_to_cpu(s->__ddir);
 		s->bs		= le64_to_cpu(s->bs);
 
 		if (ret->log_offset) {

engines/filecreate.c +1 -1

index 39a29502..5fec8544 100644
@@ -49,7 +49,7 @@ static int open_file(struct thread_data *td, struct fio_file *f)
 		uint64_t nsec;
 
 		nsec = ntime_since_now(&start);
-		add_clat_sample(td, data->stat_ddir, nsec, 0, 0);
+		add_clat_sample(td, data->stat_ddir, nsec, 0, 0, 0);
 	}
 
 	return 0;

engines/filestat.c new file +116 -0

116 lines not on this page · raw message

engines/io_uring.c +47 -0

88 lines not on this page · raw message

engines/libaio.c +54 -0

103 lines not on this page · raw message

eta.c +3 -3

index 9950ef30..13f61ba4 100644
@@ -509,9 +509,9 @@ bool calc_thread_status(struct jobs_eta *je, int force)
 		calc_rate(unified_rw_rep, rate_time, io_bytes, rate_io_bytes,
 				je->rate);
 		memcpy(&rate_prev_time, &now, sizeof(now));
-		add_agg_sample(sample_val(je->rate[DDIR_READ]), DDIR_READ, 0);
-		add_agg_sample(sample_val(je->rate[DDIR_WRITE]), DDIR_WRITE, 0);
-		add_agg_sample(sample_val(je->rate[DDIR_TRIM]), DDIR_TRIM, 0);
+		add_agg_sample(sample_val(je->rate[DDIR_READ]), DDIR_READ, 0, 0);
+		add_agg_sample(sample_val(je->rate[DDIR_WRITE]), DDIR_WRITE, 0, 0);
+		add_agg_sample(sample_val(je->rate[DDIR_TRIM]), DDIR_TRIM, 0, 0);
 	}
 
 	disp_time = mtime_since(&disp_prev_time, &now);

examples/filestat-ioengine.fio new file +19 -0

new file mode 100644
index 00000000..932fced8
@@ -0,0 +1,19 @@
+# Example filestat job
+
+# 'filestat' engine only do 'stat(filename)', file will not be open().
+# 'filesize' must be set, then files will be created at setup stage.
+
+[global]
+ioengine=filestat
+numjobs=1
+filesize=4k
+nrfiles=200
+thread
+
+[t0]
+[t1]
+[t2]
+[t3]
+[t4]
+[t5]
+

fio.1 +24 -8

index cf5dd853..05896e61 100644
@@ -1760,6 +1760,11 @@ Simply create the files and do no I/O to them.  You still need to set
 \fBfilesize\fR so that all the accounting still occurs, but no actual I/O will be
 done other than creating the file.
 .TP
+.B filestat
+Simply do stat() and do no I/O to the file. You need to set 'filesize'
+and 'nrfiles', so that files will be created.
+This engine is to measure file lookup and meta data access.
+.TP
 .B libpmem
 Read and write using mmap I/O to a file on a filesystem
 mounted with DAX on a persistent memory device through the PMDK

@@ -1790,12 +1795,12 @@ In addition, there are some parameters which are only valid when a specific
 with the caveat that when used on the command line, they must come after the
 \fBioengine\fR that defines them is selected.
 .TP
-.BI (io_uring)hipri
-If this option is set, fio will attempt to use polled IO completions. Normal IO
-completions generate interrupts to signal the completion of IO, polled
-completions do not. Hence they are require active reaping by the application.
-The benefits are more efficient IO for high IOPS scenarios, and lower latencies
-for low queue depth IO.
+.BI (io_uring, libaio)cmdprio_percentage \fR=\fPint
+Set the percentage of I/O that will be issued with higher priority by setting
+the priority bit. Non-read I/O is likely unaffected by ``cmdprio_percentage``.
+This option cannot be used with the `prio` or `prioclass` options. For this
+option to set the priority bit properly, NCQ priority must be supported and
+enabled and `direct=1' option must be used.
 .TP
 .BI (io_uring)fixedbufs
 If fio is asked to do direct IO, then Linux will map pages for each IO call, and

@@ -1803,6 +1808,13 @@ release them when IO is done. If this option is set, the pages are pre-mapped
 before IO is started. This eliminates the need to map and release for each IO.
 This is more efficient, and reduces the IO latency as well.
 .TP
+.BI (io_uring)hipri
+If this option is set, fio will attempt to use polled IO completions. Normal IO
+completions generate interrupts to signal the completion of IO, polled
+completions do not. Hence they are require active reaping by the application.
+The benefits are more efficient IO for high IOPS scenarios, and lower latencies
+for low queue depth IO.
+.TP
 .BI (io_uring)registerfiles
 With this option, fio registers the set of files being used with the kernel.
 This avoids the overhead of managing file counts in the kernel, making the

@@ -2381,10 +2393,14 @@ priority class.
 Set the I/O priority value of this job. Linux limits us to a positive value
 between 0 and 7, with 0 being the highest. See man
 \fBionice\fR\|(1). Refer to an appropriate manpage for other operating
-systems since meaning of priority may differ.
+systems since meaning of priority may differ. For per-command priority
+setting, see I/O engine specific `cmdprio_percentage` and `hipri_percentage`
+options.
 .TP
 .BI prioclass \fR=\fPint
-Set the I/O priority class. See man \fBionice\fR\|(1).
+Set the I/O priority class. See man \fBionice\fR\|(1). For per-command
+priority setting, see I/O engine specific `cmdprio_percentage` and `hipri_percent`
+options.
 .TP
 .BI cpus_allowed \fR=\fPstr
 Controls the same options as \fBcpumask\fR, but accepts a textual

fio.h +3 -1

index 6a5ead4d..2a9eef45 100644
@@ -139,6 +139,7 @@ enum {
 	FIO_RAND_ZONE_OFF,
 	FIO_RAND_POISSON2_OFF,
 	FIO_RAND_POISSON3_OFF,
+	FIO_RAND_PRIO_CMDS,
 	FIO_RAND_NR_OFFS,
 };

@@ -258,6 +259,7 @@ struct thread_data {
 	struct frand_state buf_state_prev;
 	struct frand_state dedupe_state;
 	struct frand_state zone_state;
+	struct frand_state prio_state;
 
 	struct zone_split_index **zone_state_index;

@@ -460,7 +462,7 @@ struct thread_data {
 	CUdevice  cu_dev;
 	CUcontext cu_ctx;
 	CUdeviceptr dev_mem_ptr;
-#endif	
+#endif
 
 };

init.c +3 -0

index 2f64726c..dca44bca 100644
@@ -1042,6 +1042,7 @@ static void td_fill_rand_seeds_internal(struct thread_data *td, bool use64)
 	init_rand_seed(&td->poisson_state[2], td->rand_seeds[FIO_RAND_POISSON3_OFF], 0);
 	init_rand_seed(&td->dedupe_state, td->rand_seeds[FIO_DEDUPE_OFF], false);
 	init_rand_seed(&td->zone_state, td->rand_seeds[FIO_RAND_ZONE_OFF], false);
+	init_rand_seed(&td->prio_state, td->rand_seeds[FIO_RAND_PRIO_CMDS], false);
 
 	if (!td_random(td))
 		return;

@@ -1518,6 +1519,8 @@ static int add_job(struct thread_data *td, const char *jobname, int job_add_num,
 		td->ts.lat_stat[i].min_val = ULONG_MAX;
 		td->ts.bw_stat[i].min_val = ULONG_MAX;
 		td->ts.iops_stat[i].min_val = ULONG_MAX;
+		td->ts.clat_high_prio_stat[i].min_val = ULONG_MAX;
+		td->ts.clat_prio_stat[i].min_val = ULONG_MAX;
 	}
 	td->ts.sync_stat.min_val = ULONG_MAX;
 	td->ddir_seq_nr = o->ddir_seq_nr;

io_u.c +4 -4

index 03f5c21f..bcb893c5 100644
@@ -1541,7 +1541,7 @@ again:
 		assert(io_u->flags & IO_U_F_FREE);
 		io_u_clear(td, io_u, IO_U_F_FREE | IO_U_F_NO_FILE_PUT |
 				 IO_U_F_TRIMMED | IO_U_F_BARRIER |
-				 IO_U_F_VER_LIST);
+				 IO_U_F_VER_LIST | IO_U_F_PRIORITY);
 
 		io_u->error = 0;
 		io_u->acct_ddir = -1;

@@ -1830,7 +1830,7 @@ static void account_io_completion(struct thread_data *td, struct io_u *io_u,
 		unsigned long long tnsec;
 
 		tnsec = ntime_since(&io_u->start_time, &icd->time);
-		add_lat_sample(td, idx, tnsec, bytes, io_u->offset);
+		add_lat_sample(td, idx, tnsec, bytes, io_u->offset, io_u_is_prio(io_u));
 
 		if (td->flags & TD_F_PROFILE_OPS) {
 			struct prof_io_ops *ops = &td->prof_io_ops;

@@ -1849,7 +1849,7 @@ static void account_io_completion(struct thread_data *td, struct io_u *io_u,
 
 	if (ddir_rw(idx)) {
 		if (!td->o.disable_clat) {
-			add_clat_sample(td, idx, llnsec, bytes, io_u->offset);
+			add_clat_sample(td, idx, llnsec, bytes, io_u->offset, io_u_is_prio(io_u));
 			io_u_mark_latency(td, llnsec);
 		}

@@ -2091,7 +2091,7 @@ void io_u_queued(struct thread_data *td, struct io_u *io_u)
 			td = td->parent;
 
 		add_slat_sample(td, io_u->ddir, slat_time, io_u->xfer_buflen,
-				io_u->offset);
+				io_u->offset, io_u_is_prio(io_u));
 	}
 }

io_u.h +3 -0

index e75993bd..0f63cdd0 100644
@@ -24,6 +24,7 @@ enum {
 	IO_U_F_TRIMMED		= 1 << 5,
 	IO_U_F_BARRIER		= 1 << 6,
 	IO_U_F_VER_LIST		= 1 << 7,
+	IO_U_F_PRIORITY		= 1 << 8,
 };
 
 /*

@@ -193,5 +194,7 @@ static inline enum fio_ddir acct_ddir(struct io_u *io_u)
 	td_flags_clear((td), &(io_u->flags), (val))
 #define io_u_set(td, io_u, val)		\
 	td_flags_set((td), &(io_u)->flags, (val))
+#define io_u_is_prio(io_u)	\
+	(io_u->flags & (unsigned int) IO_U_F_PRIORITY) != 0
 
 #endif

ioengines.c +1 -0

index b9200ba9..2c7a0df9 100644
@@ -318,6 +318,7 @@ enum fio_q_status td_io_queue(struct thread_data *td, struct io_u *io_u)
 					sizeof(io_u->issue_time));
 	}
 
+
 	if (ddir_rw(ddir)) {
 		if (!(io_u->flags & IO_U_F_VER_LIST)) {
 			td->io_issues[ddir]++;

iolog.c +4 -4

index b72dcf97..917a446c 100644
@@ -896,18 +896,18 @@ void flush_samples(FILE *f, void *samples, uint64_t sample_size)
 		s = __get_sample(samples, log_offset, i);
 
 		if (!log_offset) {
-			fprintf(f, "%lu, %" PRId64 ", %u, %llu\n",
+			fprintf(f, "%lu, %" PRId64 ", %u, %llu, %u\n",
 					(unsigned long) s->time,
 					s->data.val,
-					io_sample_ddir(s), (unsigned long long) s->bs);
+					io_sample_ddir(s), (unsigned long long) s->bs, s->priority_bit);
 		} else {
 			struct io_sample_offset *so = (void *) s;
 
-			fprintf(f, "%lu, %" PRId64 ", %u, %llu, %llu\n",
+			fprintf(f, "%lu, %" PRId64 ", %u, %llu, %llu, %u\n",
 					(unsigned long) s->time,
 					s->data.val,
 					io_sample_ddir(s), (unsigned long long) s->bs,
-					(unsigned long long) so->offset);
+					(unsigned long long) so->offset, s->priority_bit);
 		}
 	}
 }

iolog.h +1 -0

index 17be908f..981081f9 100644
@@ -42,6 +42,7 @@ struct io_sample {
 	uint64_t time;
 	union io_sample_data data;
 	uint32_t __ddir;
+	uint8_t priority_bit;
 	uint64_t bs;
 };

server.c +11 -2

index 1a070e56..a5af5e74 100644
@@ -1262,7 +1262,7 @@ static int handle_connection(struct sk_out *sk_out)
 	_exit(ret);
 }
 
-/* get the address on this host bound by the input socket, 
+/* get the address on this host bound by the input socket,
  * whether it is ipv6 or ipv4 */
 
 static int get_my_addr_str(int sk)

@@ -1574,6 +1574,15 @@ void fio_server_send_ts(struct thread_stat *ts, struct group_run_stats *rs)
 	p.ts.cachehit		= cpu_to_le64(ts->cachehit);
 	p.ts.cachemiss		= cpu_to_le64(ts->cachemiss);
 
+	for (i = 0; i < DDIR_RWDIR_CNT; i++) {
+		for (j = 0; j < FIO_IO_U_PLAT_NR; j++) {
+			p.ts.io_u_plat_high_prio[i][j] = cpu_to_le64(ts->io_u_plat_high_prio[i][j]);
+			p.ts.io_u_plat_prio[i][j] = cpu_to_le64(ts->io_u_plat_prio[i][j]);
+		}
+		convert_io_stat(&p.ts.clat_high_prio_stat[i], &ts->clat_high_prio_stat[i]);
+		convert_io_stat(&p.ts.clat_prio_stat[i], &ts->clat_prio_stat[i]);
+	}
+
 	convert_gs(&p.rs, rs);
 
 	dprint(FD_NET, "ts->ss_state = %d\n", ts->ss_state);

@@ -1998,7 +2007,7 @@ int fio_send_iolog(struct thread_data *td, struct io_log *log, const char *name)
 
 			s->time		= cpu_to_le64(s->time);
 			s->data.val	= cpu_to_le64(s->data.val);
-			s->__ddir	= cpu_to_le32(s->__ddir);
+			s->__ddir	= __cpu_to_le32(s->__ddir);
 			s->bs		= cpu_to_le64(s->bs);
 
 			if (log->log_offset) {

server.h +1 -1

index de1d7f9b..6ac75366 100644
@@ -48,7 +48,7 @@ struct fio_net_cmd_reply {
 };
 
 enum {
-	FIO_SERVER_VER			= 80,
+	FIO_SERVER_VER			= 81,
 
 	FIO_SERVER_MAX_FRAGMENT_PDU	= 1024,
 	FIO_SERVER_MAX_CMD_MB		= 2048,

stat.c +196 ~~-33~~

444 lines not on this page · raw message

stat.h +10 -4

index 2ce91ff0..9320c6bd 100644
@@ -239,6 +239,11 @@ struct thread_stat {
 	fio_fp64_t ss_deviation;
 	fio_fp64_t ss_criterion;
 
+	uint64_t io_u_plat_high_prio[DDIR_RWDIR_CNT][FIO_IO_U_PLAT_NR] __attribute__((aligned(8)));;
+	uint64_t io_u_plat_prio[DDIR_RWDIR_CNT][FIO_IO_U_PLAT_NR];
+	struct io_stat clat_high_prio_stat[DDIR_RWDIR_CNT] __attribute__((aligned(8)));
+	struct io_stat clat_prio_stat[DDIR_RWDIR_CNT];
+
 	union {
 		uint64_t *ss_iops_data;
 		uint64_t pad4;

@@ -323,12 +328,13 @@ extern void update_rusage_stat(struct thread_data *);
 extern void clear_rusage_stat(struct thread_data *);
 
 extern void add_lat_sample(struct thread_data *, enum fio_ddir, unsigned long long,
-				unsigned long long, uint64_t);
+				unsigned long long, uint64_t, uint8_t);
 extern void add_clat_sample(struct thread_data *, enum fio_ddir, unsigned long long,
-				unsigned long long, uint64_t);
+				unsigned long long, uint64_t, uint8_t);
 extern void add_slat_sample(struct thread_data *, enum fio_ddir, unsigned long,
-				unsigned long long, uint64_t);
-extern void add_agg_sample(union io_sample_data, enum fio_ddir, unsigned long long);
+				unsigned long long, uint64_t, uint8_t);
+extern void add_agg_sample(union io_sample_data, enum fio_ddir, unsigned long long bs,
+				uint8_t priority_bit);
 extern void add_iops_sample(struct thread_data *, struct io_u *,
 				unsigned int);
 extern void add_bw_sample(struct thread_data *, struct io_u *,