[fio.git] / engines / libaio.c

/*
 * libaio engine
 *
 * IO engine using the Linux native aio interface.
 *
 */
#include <stdio.h>
#include <stdlib.h>
#include <unistd.h>
#include <errno.h>
#include <assert.h>

#include "../fio.h"

#ifdef FIO_HAVE_LIBAIO

struct libaio_data {
	io_context_t aio_ctx;
	struct io_event *aio_events;
	struct iocb **iocbs;
	struct io_u **io_us;
	int iocbs_nr;
};

struct libaio_options {
	struct thread_data *td;
	unsigned int userspace_reap;
};

static struct fio_option options[] = {
	{
		.name	= "userspace_reap",
		.type	= FIO_OPT_STR_SET,
		.off1	= offsetof(struct libaio_options, userspace_reap),
		.help	= "Use alternative user-space reap implementation",
	},
	{
		.name	= NULL,
	},
};

static int fio_libaio_prep(struct thread_data fio_unused *td, struct io_u *io_u)
{
	struct fio_file *f = io_u->file;

	if (io_u->ddir == DDIR_READ)
		io_prep_pread(&io_u->iocb, f->fd, io_u->xfer_buf, io_u->xfer_buflen, io_u->offset);
	else if (io_u->ddir == DDIR_WRITE)
		io_prep_pwrite(&io_u->iocb, f->fd, io_u->xfer_buf, io_u->xfer_buflen, io_u->offset);
	else if (ddir_sync(io_u->ddir))
		io_prep_fsync(&io_u->iocb, f->fd);

	return 0;
}

static struct io_u *fio_libaio_event(struct thread_data *td, int event)
{
	struct libaio_data *ld = td->io_ops->data;
	struct io_event *ev;
	struct io_u *io_u;

	ev = ld->aio_events + event;
	io_u = container_of(ev->obj, struct io_u, iocb);

	if (ev->res != io_u->xfer_buflen) {
		if (ev->res > io_u->xfer_buflen)
			io_u->error = -ev->res;
		else
			io_u->resid = io_u->xfer_buflen - ev->res;
	} else
		io_u->error = 0;

	return io_u;
}

struct aio_ring {
	unsigned id;		 /** kernel internal index number */
	unsigned nr;		 /** number of io_events */
	unsigned head;
	unsigned tail;

	unsigned magic;
	unsigned compat_features;
	unsigned incompat_features;
	unsigned header_length;	/** size of aio_ring */

	struct io_event events[0];
};

#define AIO_RING_MAGIC	0xa10a10a1

static int user_io_getevents(io_context_t aio_ctx, unsigned int max,
			     struct io_event *events)
{
	long i = 0;
	unsigned head;
	struct aio_ring *ring = (struct aio_ring*) aio_ctx;

	while (i < max) {
		head = ring->head;

		if (head == ring->tail) {
			/* There are no more completions */
			break;
		} else {
			/* There is another completion to reap */
			events[i] = ring->events[head];
			read_barrier();
			ring->head = (head + 1) % ring->nr;
			i++;
		}
	}

	return i;
}

static int fio_libaio_getevents(struct thread_data *td, unsigned int min,
				unsigned int max, struct timespec *t)
{
	struct libaio_data *ld = td->io_ops->data;
	struct libaio_options *o = td->eo;
	unsigned actual_min = td->o.iodepth_batch_complete == 0 ? 0 : min;
	int r, events = 0;

	do {
		if (o->userspace_reap == 1
		    && actual_min == 0
		    && ((struct aio_ring *)(ld->aio_ctx))->magic
				== AIO_RING_MAGIC) {
			r = user_io_getevents(ld->aio_ctx, max,
				ld->aio_events + events);
		} else {
			r = io_getevents(ld->aio_ctx, actual_min,
				max, ld->aio_events + events, t);
		}
		if (r >= 0)
			events += r;
		else if (r == -EAGAIN)
			usleep(100);
	} while (events < min);

	return r < 0 ? r : events;
}

static int fio_libaio_queue(struct thread_data *td, struct io_u *io_u)
{
	struct libaio_data *ld = td->io_ops->data;

	fio_ro_check(td, io_u);

	if (ld->iocbs_nr == (int) td->o.iodepth)
		return FIO_Q_BUSY;

	/*
	 * fsync is tricky, since it can fail and we need to do it
	 * serialized with other io. the reason is that linux doesn't
	 * support aio fsync yet. So return busy for the case where we
	 * have pending io, to let fio complete those first.
	 */
	if (ddir_sync(io_u->ddir)) {
		if (ld->iocbs_nr)
			return FIO_Q_BUSY;

		do_io_u_sync(td, io_u);
		return FIO_Q_COMPLETED;
	}

	if (io_u->ddir == DDIR_TRIM) {
		if (ld->iocbs_nr)
			return FIO_Q_BUSY;

		do_io_u_trim(td, io_u);
		return FIO_Q_COMPLETED;
	}

	ld->iocbs[ld->iocbs_nr] = &io_u->iocb;
	ld->io_us[ld->iocbs_nr] = io_u;
	ld->iocbs_nr++;
	return FIO_Q_QUEUED;
}

static void fio_libaio_queued(struct thread_data *td, struct io_u **io_us,
			      unsigned int nr)
{
	struct timeval now;
	unsigned int i;

	if (!fio_fill_issue_time(td))
		return;

	fio_gettime(&now, NULL);

	for (i = 0; i < nr; i++) {
		struct io_u *io_u = io_us[i];

		memcpy(&io_u->issue_time, &now, sizeof(now));
		io_u_queued(td, io_u);
	}
}

static int fio_libaio_commit(struct thread_data *td)
{
	struct libaio_data *ld = td->io_ops->data;
	struct iocb **iocbs;
	struct io_u **io_us;
	int ret;

	if (!ld->iocbs_nr)
		return 0;

	io_us = ld->io_us;
	iocbs = ld->iocbs;
	do {
		ret = io_submit(ld->aio_ctx, ld->iocbs_nr, iocbs);
		if (ret > 0) {
			fio_libaio_queued(td, io_us, ret);
			io_u_mark_submit(td, ret);
			ld->iocbs_nr -= ret;
			io_us += ret;
			iocbs += ret;
			ret = 0;
		} else if (!ret || ret == -EAGAIN || ret == -EINTR) {
			if (!ret)
				io_u_mark_submit(td, ret);
			continue;
		} else
			break;
	} while (ld->iocbs_nr);

	return ret;
}

static int fio_libaio_cancel(struct thread_data *td, struct io_u *io_u)
{
	struct libaio_data *ld = td->io_ops->data;

	return io_cancel(ld->aio_ctx, &io_u->iocb, ld->aio_events);
}

static void fio_libaio_cleanup(struct thread_data *td)
{
	struct libaio_data *ld = td->io_ops->data;

	if (ld) {
		io_destroy(ld->aio_ctx);
		free(ld->aio_events);
		free(ld->iocbs);
		free(ld->io_us);
		free(ld);
	}
}

static int fio_libaio_init(struct thread_data *td)
{
	struct libaio_data *ld = malloc(sizeof(*ld));
	struct libaio_options *o = td->eo;
	int err = 0;

	memset(ld, 0, sizeof(*ld));

	/*
	 * First try passing in 0 for queue depth, since we don't
	 * care about the user ring. If that fails, the kernel is too old
	 * and we need the right depth.
	 */
	if (!o->userspace_reap)
		err = io_queue_init(INT_MAX, &ld->aio_ctx);
	if (o->userspace_reap || err == -EINVAL)
		err = io_queue_init(td->o.iodepth, &ld->aio_ctx);
	if (err) {
		td_verror(td, -err, "io_queue_init");
		log_err("fio: check /proc/sys/fs/aio-max-nr\n");
		free(ld);
		return 1;
	}

	ld->aio_events = malloc(td->o.iodepth * sizeof(struct io_event));
	memset(ld->aio_events, 0, td->o.iodepth * sizeof(struct io_event));
	ld->iocbs = malloc(td->o.iodepth * sizeof(struct iocb *));
	memset(ld->iocbs, 0, sizeof(struct iocb *));
	ld->io_us = malloc(td->o.iodepth * sizeof(struct io_u *));
	memset(ld->io_us, 0, td->o.iodepth * sizeof(struct io_u *));
	ld->iocbs_nr = 0;

	td->io_ops->data = ld;
	return 0;
}

static struct ioengine_ops ioengine = {
	.name			= "libaio",
	.version		= FIO_IOOPS_VERSION,
	.init			= fio_libaio_init,
	.prep			= fio_libaio_prep,
	.queue			= fio_libaio_queue,
	.commit			= fio_libaio_commit,
	.cancel			= fio_libaio_cancel,
	.getevents		= fio_libaio_getevents,
	.event			= fio_libaio_event,
	.cleanup		= fio_libaio_cleanup,
	.open_file		= generic_open_file,
	.close_file		= generic_close_file,
	.get_file_size		= generic_get_file_size,
	.options		= options,
	.option_struct_size	= sizeof(struct libaio_options),
};

#else /* FIO_HAVE_LIBAIO */

/*
 * When we have a proper configure system in place, we simply wont build
 * and install this io engine. For now install a crippled version that
 * just complains and fails to load.
 */
static int fio_libaio_init(struct thread_data fio_unused *td)
{
	log_err("fio: libaio not available\n");
	return 1;
}

static struct ioengine_ops ioengine = {
	.name		= "libaio",
	.version	= FIO_IOOPS_VERSION,
	.init		= fio_libaio_init,
};

#endif

static void fio_init fio_libaio_register(void)
{
	register_ioengine(&ioengine);
}

static void fio_exit fio_libaio_unregister(void)
{
	unregister_ioengine(&ioengine);
}
Commit	Line	Data
	1	/*
	2	* libaio engine
	3	*
	4	* IO engine using the Linux native aio interface.
	5	*
	6	*/
	7	#include <stdio.h>
	8	#include <stdlib.h>
	9	#include <unistd.h>
	10	#include <errno.h>
	11	#include <assert.h>
	12
	13	#include "../fio.h"
	14
	15	#ifdef FIO_HAVE_LIBAIO
	16
	17	struct libaio_data {
	18	io_context_t aio_ctx;
	19	struct io_event *aio_events;
	20	struct iocb **iocbs;
	21	struct io_u **io_us;
	22	int iocbs_nr;
	23	};
	24
	25	struct libaio_options {
	26	struct thread_data *td;
	27	unsigned int userspace_reap;
	28	};
	29
	30	static struct fio_option options[] = {
	31	{
	32	.name = "userspace_reap",
	33	.type = FIO_OPT_STR_SET,
	34	.off1 = offsetof(struct libaio_options, userspace_reap),
	35	.help = "Use alternative user-space reap implementation",
	36	},
	37	{
	38	.name = NULL,
	39	},
	40	};
	41
	42	static int fio_libaio_prep(struct thread_data fio_unused td, struct io_u io_u)
	43	{
	44	struct fio_file *f = io_u->file;
	45
	46	if (io_u->ddir == DDIR_READ)
	47	io_prep_pread(&io_u->iocb, f->fd, io_u->xfer_buf, io_u->xfer_buflen, io_u->offset);
	48	else if (io_u->ddir == DDIR_WRITE)
	49	io_prep_pwrite(&io_u->iocb, f->fd, io_u->xfer_buf, io_u->xfer_buflen, io_u->offset);
	50	else if (ddir_sync(io_u->ddir))
	51	io_prep_fsync(&io_u->iocb, f->fd);
	52
	53	return 0;
	54	}
	55
	56	static struct io_u fio_libaio_event(struct thread_data td, int event)
	57	{
	58	struct libaio_data *ld = td->io_ops->data;
	59	struct io_event *ev;
	60	struct io_u *io_u;
	61
	62	ev = ld->aio_events + event;
	63	io_u = container_of(ev->obj, struct io_u, iocb);
	64
	65	if (ev->res != io_u->xfer_buflen) {
	66	if (ev->res > io_u->xfer_buflen)
	67	io_u->error = -ev->res;
	68	else
	69	io_u->resid = io_u->xfer_buflen - ev->res;
	70	} else
	71	io_u->error = 0;
	72
	73	return io_u;
	74	}
	75
	76	struct aio_ring {
	77	unsigned id; /** kernel internal index number */
	78	unsigned nr; /** number of io_events */
	79	unsigned head;
	80	unsigned tail;
	81
	82	unsigned magic;
	83	unsigned compat_features;
	84	unsigned incompat_features;
	85	unsigned header_length; /** size of aio_ring */
	86
	87	struct io_event events[0];
	88	};
	89
	90	#define AIO_RING_MAGIC 0xa10a10a1
	91
	92	static int user_io_getevents(io_context_t aio_ctx, unsigned int max,
	93	struct io_event *events)
	94	{
	95	long i = 0;
	96	unsigned head;
	97	struct aio_ring ring = (struct aio_ring) aio_ctx;
	98
	99	while (i < max) {
	100	head = ring->head;
	101
	102	if (head == ring->tail) {
	103	/* There are no more completions */
	104	break;
	105	} else {
	106	/* There is another completion to reap */
	107	events[i] = ring->events[head];
	108	read_barrier();
	109	ring->head = (head + 1) % ring->nr;
	110	i++;
	111	}
	112	}
	113
	114	return i;
	115	}
	116
	117	static int fio_libaio_getevents(struct thread_data *td, unsigned int min,
	118	unsigned int max, struct timespec *t)
	119	{
	120	struct libaio_data *ld = td->io_ops->data;
	121	struct libaio_options *o = td->eo;
	122	unsigned actual_min = td->o.iodepth_batch_complete == 0 ? 0 : min;
	123	int r, events = 0;
	124
	125	do {
	126	if (o->userspace_reap == 1
	127	&& actual_min == 0
	128	&& ((struct aio_ring *)(ld->aio_ctx))->magic
	129	== AIO_RING_MAGIC) {
	130	r = user_io_getevents(ld->aio_ctx, max,
	131	ld->aio_events + events);
	132	} else {
	133	r = io_getevents(ld->aio_ctx, actual_min,
	134	max, ld->aio_events + events, t);
	135	}
	136	if (r >= 0)
	137	events += r;
	138	else if (r == -EAGAIN)
	139	usleep(100);
	140	} while (events < min);
	141
	142	return r < 0 ? r : events;
	143	}
	144
	145	static int fio_libaio_queue(struct thread_data td, struct io_u io_u)
	146	{
	147	struct libaio_data *ld = td->io_ops->data;
	148
	149	fio_ro_check(td, io_u);
	150
	151	if (ld->iocbs_nr == (int) td->o.iodepth)
	152	return FIO_Q_BUSY;
	153
	154	/*
	155	* fsync is tricky, since it can fail and we need to do it
	156	* serialized with other io. the reason is that linux doesn't
	157	* support aio fsync yet. So return busy for the case where we
	158	* have pending io, to let fio complete those first.
	159	*/
	160	if (ddir_sync(io_u->ddir)) {
	161	if (ld->iocbs_nr)
	162	return FIO_Q_BUSY;
	163
	164	do_io_u_sync(td, io_u);
	165	return FIO_Q_COMPLETED;
	166	}
	167
	168	if (io_u->ddir == DDIR_TRIM) {
	169	if (ld->iocbs_nr)
	170	return FIO_Q_BUSY;
	171
	172	do_io_u_trim(td, io_u);
	173	return FIO_Q_COMPLETED;
	174	}
	175
	176	ld->iocbs[ld->iocbs_nr] = &io_u->iocb;
	177	ld->io_us[ld->iocbs_nr] = io_u;
	178	ld->iocbs_nr++;
	179	return FIO_Q_QUEUED;
	180	}
	181
	182	static void fio_libaio_queued(struct thread_data td, struct io_u *io_us,
	183	unsigned int nr)
	184	{
	185	struct timeval now;
	186	unsigned int i;
	187
	188	if (!fio_fill_issue_time(td))
	189	return;
	190
	191	fio_gettime(&now, NULL);
	192
	193	for (i = 0; i < nr; i++) {
	194	struct io_u *io_u = io_us[i];
	195
	196	memcpy(&io_u->issue_time, &now, sizeof(now));
	197	io_u_queued(td, io_u);
	198	}
	199	}
	200
	201	static int fio_libaio_commit(struct thread_data *td)
	202	{
	203	struct libaio_data *ld = td->io_ops->data;
	204	struct iocb **iocbs;
	205	struct io_u **io_us;
	206	int ret;
	207
	208	if (!ld->iocbs_nr)
	209	return 0;
	210
	211	io_us = ld->io_us;
	212	iocbs = ld->iocbs;
	213	do {
	214	ret = io_submit(ld->aio_ctx, ld->iocbs_nr, iocbs);
	215	if (ret > 0) {
	216	fio_libaio_queued(td, io_us, ret);
	217	io_u_mark_submit(td, ret);
	218	ld->iocbs_nr -= ret;
	219	io_us += ret;
	220	iocbs += ret;
	221	ret = 0;
	222	} else if (!ret \|\| ret == -EAGAIN \|\| ret == -EINTR) {
	223	if (!ret)
	224	io_u_mark_submit(td, ret);
	225	continue;
	226	} else
	227	break;
	228	} while (ld->iocbs_nr);
	229
	230	return ret;
	231	}
	232
	233	static int fio_libaio_cancel(struct thread_data td, struct io_u io_u)
	234	{
	235	struct libaio_data *ld = td->io_ops->data;
	236
	237	return io_cancel(ld->aio_ctx, &io_u->iocb, ld->aio_events);
	238	}
	239
	240	static void fio_libaio_cleanup(struct thread_data *td)
	241	{
	242	struct libaio_data *ld = td->io_ops->data;
	243
	244	if (ld) {
	245	io_destroy(ld->aio_ctx);
	246	free(ld->aio_events);
	247	free(ld->iocbs);
	248	free(ld->io_us);
	249	free(ld);
	250	}
	251	}
	252
	253	static int fio_libaio_init(struct thread_data *td)
	254	{
	255	struct libaio_data ld = malloc(sizeof(ld));
	256	struct libaio_options *o = td->eo;
	257	int err = 0;
	258
	259	memset(ld, 0, sizeof(*ld));
	260
	261	/*
	262	* First try passing in 0 for queue depth, since we don't
	263	* care about the user ring. If that fails, the kernel is too old
	264	* and we need the right depth.
	265	*/
	266	if (!o->userspace_reap)
	267	err = io_queue_init(INT_MAX, &ld->aio_ctx);
	268	if (o->userspace_reap \|\| err == -EINVAL)
	269	err = io_queue_init(td->o.iodepth, &ld->aio_ctx);
	270	if (err) {
	271	td_verror(td, -err, "io_queue_init");
	272	log_err("fio: check /proc/sys/fs/aio-max-nr\n");
	273	free(ld);
	274	return 1;
	275	}
	276
	277	ld->aio_events = malloc(td->o.iodepth * sizeof(struct io_event));
	278	memset(ld->aio_events, 0, td->o.iodepth * sizeof(struct io_event));
	279	ld->iocbs = malloc(td->o.iodepth * sizeof(struct iocb *));
	280	memset(ld->iocbs, 0, sizeof(struct iocb *));
	281	ld->io_us = malloc(td->o.iodepth * sizeof(struct io_u *));
	282	memset(ld->io_us, 0, td->o.iodepth * sizeof(struct io_u *));
	283	ld->iocbs_nr = 0;
	284
	285	td->io_ops->data = ld;
	286	return 0;
	287	}
	288
	289	static struct ioengine_ops ioengine = {
	290	.name = "libaio",
	291	.version = FIO_IOOPS_VERSION,
	292	.init = fio_libaio_init,
	293	.prep = fio_libaio_prep,
	294	.queue = fio_libaio_queue,
	295	.commit = fio_libaio_commit,
	296	.cancel = fio_libaio_cancel,
	297	.getevents = fio_libaio_getevents,
	298	.event = fio_libaio_event,
	299	.cleanup = fio_libaio_cleanup,
	300	.open_file = generic_open_file,
	301	.close_file = generic_close_file,
	302	.get_file_size = generic_get_file_size,
	303	.options = options,
	304	.option_struct_size = sizeof(struct libaio_options),
	305	};
	306
	307	#else /* FIO_HAVE_LIBAIO */
	308
	309	/*
	310	* When we have a proper configure system in place, we simply wont build
	311	* and install this io engine. For now install a crippled version that
	312	* just complains and fails to load.
	313	*/
	314	static int fio_libaio_init(struct thread_data fio_unused *td)
	315	{
	316	log_err("fio: libaio not available\n");
	317	return 1;
	318	}
	319
	320	static struct ioengine_ops ioengine = {
	321	.name = "libaio",
	322	.version = FIO_IOOPS_VERSION,
	323	.init = fio_libaio_init,
	324	};
	325
	326	#endif
	327
	328	static void fio_init fio_libaio_register(void)
	329	{
	330	register_ioengine(&ioengine);
	331	}
	332
	333	static void fio_exit fio_libaio_unregister(void)
	334	{
	335	unregister_ioengine(&ioengine);
	336	}