[fio.git] / engines / aioring.c

/*
 * aioring engine
 *
 * IO engine using the new native Linux libaio ring interface
 *
 */
#include <stdlib.h>
#include <unistd.h>
#include <errno.h>
#include <libaio.h>
#include <sys/time.h>
#include <sys/resource.h>

#include "../fio.h"
#include "../lib/pow2.h"
#include "../optgroup.h"
#include "../lib/memalign.h"

#ifndef IOCB_FLAG_HIPRI
#define IOCB_FLAG_HIPRI	(1 << 2)
#endif

/*
 * io_setup2(2) flags
 */
#ifndef IOCTX_FLAG_IOPOLL
#define IOCTX_FLAG_IOPOLL	(1 << 0)
#endif
#ifndef IOCTX_FLAG_SCQRING
#define IOCTX_FLAG_SCQRING	(1 << 1)
#endif
#ifndef IOCTX_FLAG_FIXEDBUFS
#define IOCTX_FLAG_FIXEDBUFS	(1 << 2)
#endif
#ifndef IOCTX_FLAG_SQTHREAD
#define IOCTX_FLAG_SQTHREAD	(1 << 3)
#endif
#ifndef IOCTX_FLAG_SQWQ
#define IOCTX_FLAG_SQWQ		(1 << 4)
#endif

/*
 * io_ring_enter(2) flags
 */
#ifndef IORING_FLAG_SUBMIT
#define IORING_FLAG_SUBMIT	(1 << 0)
#endif
#ifndef IORING_FLAG_GETEVENTS
#define IORING_FLAG_GETEVENTS	(1 << 1)
#endif

typedef uint64_t u64;
typedef uint32_t u32;
typedef uint16_t u16;

struct aio_sq_ring {
	union {
		struct {
			u32 head;
			u32 tail;
			u32 nr_events;
			u16 sq_thread_cpu;
			u64 iocbs;
		};
		u32 pad[16];
	};
	u32 array[0];
};

struct aio_cq_ring {
	union {
		struct {
			u32 head;
			u32 tail;
			u32 nr_events;
		};
		struct io_event pad;
	};
	struct io_event events[0];
};

struct aioring_data {
	io_context_t aio_ctx;
	struct io_u **io_us;
	struct io_u **io_u_index;

	struct aio_sq_ring *sq_ring;
	struct iocb *iocbs;

	struct aio_cq_ring *cq_ring;
	struct io_event *events;

	int queued;
	int cq_ring_off;
};

struct aioring_options {
	void *pad;
	unsigned int hipri;
	unsigned int fixedbufs;
};

static struct fio_option options[] = {
	{
		.name	= "hipri",
		.lname	= "High Priority",
		.type	= FIO_OPT_STR_SET,
		.off1	= offsetof(struct aioring_options, hipri),
		.help	= "Use polled IO completions",
		.category = FIO_OPT_C_ENGINE,
		.group	= FIO_OPT_G_LIBAIO,
	},
	{
		.name	= "fixedbufs",
		.lname	= "Fixed (pre-mapped) IO buffers",
		.type	= FIO_OPT_STR_SET,
		.off1	= offsetof(struct aioring_options, fixedbufs),
		.help	= "Pre map IO buffers",
		.category = FIO_OPT_C_ENGINE,
		.group	= FIO_OPT_G_LIBAIO,
	},
	{
		.name	= NULL,
	},
};

static int fio_aioring_commit(struct thread_data *td);

static int io_ring_enter(io_context_t ctx, unsigned int to_submit,
			 unsigned int min_complete, unsigned int flags)
{
#ifdef __NR_sys_io_ring_enter
	return syscall(__NR_sys_io_ring_enter, ctx, to_submit, min_complete,
			flags);
#else
	return -1;
#endif
}

static int fio_aioring_prep(struct thread_data *td, struct io_u *io_u)
{
	struct aioring_data *ld = td->io_ops_data;
	struct fio_file *f = io_u->file;
	struct aioring_options *o = td->eo;
	struct iocb *iocb;

	iocb = &ld->iocbs[io_u->index];

	if (io_u->ddir == DDIR_READ) {
		if (o->fixedbufs) {
			iocb->aio_fildes = f->fd;
			iocb->aio_lio_opcode = IO_CMD_PREAD;
			iocb->u.c.offset = io_u->offset;
		} else {
			io_prep_pread(iocb, f->fd, io_u->xfer_buf, io_u->xfer_buflen, io_u->offset);
			if (o->hipri)
				iocb->u.c.flags |= IOCB_FLAG_HIPRI;
		}
	} else if (io_u->ddir == DDIR_WRITE) {
		if (o->fixedbufs) {
			iocb->aio_fildes = f->fd;
			iocb->aio_lio_opcode = IO_CMD_PWRITE;
			iocb->u.c.offset = io_u->offset;
		} else {
			io_prep_pwrite(iocb, f->fd, io_u->xfer_buf, io_u->xfer_buflen, io_u->offset);
			if (o->hipri)
				iocb->u.c.flags |= IOCB_FLAG_HIPRI;
		}
	} else if (ddir_sync(io_u->ddir))
		io_prep_fsync(iocb, f->fd);

	iocb->data = io_u;
	return 0;
}

static struct io_u *fio_aioring_event(struct thread_data *td, int event)
{
	struct aioring_data *ld = td->io_ops_data;
	struct io_event *ev;
	struct io_u *io_u;
	int index;

	index = event + ld->cq_ring_off;
	if (index >= ld->cq_ring->nr_events)
		index -= ld->cq_ring->nr_events;

	ev = &ld->cq_ring->events[index];
	io_u = ev->data;

	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;
}

static int fio_aioring_cqring_reap(struct thread_data *td, unsigned int events,
				   unsigned int max)
{
	struct aioring_data *ld = td->io_ops_data;
	struct aio_cq_ring *ring = ld->cq_ring;
	u32 head, reaped = 0;

	head = ring->head;
	do {
		read_barrier();
		if (head == ring->tail)
			break;
		reaped++;
		head++;
		if (head == ring->nr_events)
			head = 0;
	} while (reaped + events < max);

	ring->head = head;
	write_barrier();
	return reaped;
}

static int fio_aioring_getevents(struct thread_data *td, unsigned int min,
				 unsigned int max, const struct timespec *t)
{
	struct aioring_data *ld = td->io_ops_data;
	unsigned actual_min = td->o.iodepth_batch_complete_min == 0 ? 0 : min;
	struct aio_cq_ring *ring = ld->cq_ring;
	int r, events = 0;

	ld->cq_ring_off = ring->head;
	do {
		r = fio_aioring_cqring_reap(td, events, max);
		if (r) {
			events += r;
			continue;
		}

		r = io_ring_enter(ld->aio_ctx, 0, actual_min,
					IORING_FLAG_GETEVENTS);
		if (r < 0) {
			if (errno == EAGAIN)
				continue;
			perror("ring enter");
			break;
		}
	} while (events < min);

	return r < 0 ? r : events;
}

static enum fio_q_status fio_aioring_queue(struct thread_data *td,
					   struct io_u *io_u)
{
	struct aioring_data *ld = td->io_ops_data;
	struct aio_sq_ring *ring = ld->sq_ring;
	unsigned tail, next_tail;

	fio_ro_check(td, io_u);

	if (ld->queued == 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->queued)
			return FIO_Q_BUSY;

		do_io_u_sync(td, io_u);
		return FIO_Q_COMPLETED;
	}

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

		do_io_u_trim(td, io_u);
		io_u_mark_submit(td, 1);
		io_u_mark_complete(td, 1);
		return FIO_Q_COMPLETED;
	}

	tail = ring->tail;
	next_tail = tail + 1;
	if (next_tail == ring->nr_events)
		next_tail = 0;
	read_barrier();
	if (next_tail == ring->head)
		return FIO_Q_BUSY;

	ring->array[tail] = io_u->index;
	ring->tail = next_tail;
	write_barrier();

	ld->queued++;
	return FIO_Q_QUEUED;
}

static void fio_aioring_queued(struct thread_data *td, int start, int nr)
{
	struct aioring_data *ld = td->io_ops_data;
	struct timespec now;

	if (!fio_fill_issue_time(td))
		return;

	fio_gettime(&now, NULL);

	while (nr--) {
		int index = ld->sq_ring->array[start];
		struct io_u *io_u = io_u = ld->io_u_index[index];

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

		start++;
		if (start == ld->sq_ring->nr_events)
			start = 0;
	}
}

static int fio_aioring_commit(struct thread_data *td)
{
	struct aioring_data *ld = td->io_ops_data;
	int ret;

	if (!ld->queued)
		return 0;

	do {
		int start = ld->sq_ring->head;
		long nr = ld->queued;

		ret = io_ring_enter(ld->aio_ctx, nr, 0, IORING_FLAG_SUBMIT |
						IORING_FLAG_GETEVENTS);
		if (ret == -1)
			perror("io_ring_enter");
		if (ret > 0) {
			fio_aioring_queued(td, start, ret);
			io_u_mark_submit(td, ret);

			ld->queued -= ret;
			ret = 0;
		} else if (ret == -EINTR || !ret) {
			if (!ret)
				io_u_mark_submit(td, ret);
			continue;
		} else if (ret == -EAGAIN) {
			/*
			 * If we get EAGAIN, we should break out without
			 * error and let the upper layer reap some
			 * events for us. If we have no queued IO, we
			 * must loop here. If we loop for more than 30s,
			 * just error out, something must be buggy in the
			 * IO path.
			 */
			if (ld->queued) {
				ret = 0;
				break;
			}
			usleep(1);
			continue;
		} else if (ret == -ENOMEM) {
			/*
			 * If we get -ENOMEM, reap events if we can. If
			 * we cannot, treat it as a fatal event since there's
			 * nothing we can do about it.
			 */
			if (ld->queued)
				ret = 0;
			break;
		} else
			break;
	} while (ld->queued);

	return ret;
}

static size_t aioring_cq_size(struct thread_data *td)
{
	return sizeof(struct aio_cq_ring) + 2 * td->o.iodepth * sizeof(struct io_event);
}

static size_t aioring_sq_iocb(struct thread_data *td)
{
	return sizeof(struct iocb) * td->o.iodepth;
}

static size_t aioring_sq_size(struct thread_data *td)
{
	return sizeof(struct aio_sq_ring) + td->o.iodepth * sizeof(u32);
}

static void fio_aioring_cleanup(struct thread_data *td)
{
	struct aioring_data *ld = td->io_ops_data;

	if (ld) {
		/*
		 * Work-around to avoid huge RCU stalls at exit time. If we
		 * don't do this here, then it'll be torn down by exit_aio().
		 * But for that case we can parallellize the freeing, thus
		 * speeding it up a lot.
		 */
		if (!(td->flags & TD_F_CHILD))
			io_destroy(ld->aio_ctx);
		free(ld->io_u_index);
		free(ld->io_us);
		fio_memfree(ld->sq_ring, aioring_sq_size(td), false);
		fio_memfree(ld->iocbs, aioring_sq_iocb(td), false);
		fio_memfree(ld->cq_ring, aioring_cq_size(td), false);
		free(ld);
	}
}

static int fio_aioring_queue_init(struct thread_data *td)
{
#ifdef __NR_sys_io_setup2
	struct aioring_data *ld = td->io_ops_data;
	struct aioring_options *o = td->eo;
	int flags = IOCTX_FLAG_SCQRING;
	int depth = td->o.iodepth;

	if (o->hipri)
		flags |= IOCTX_FLAG_IOPOLL;
	if (o->fixedbufs) {
		struct rlimit rlim = {
			.rlim_cur = RLIM_INFINITY,
			.rlim_max = RLIM_INFINITY,
		};

		setrlimit(RLIMIT_MEMLOCK, &rlim);
		flags |= IOCTX_FLAG_FIXEDBUFS;
	}

	return syscall(__NR_sys_io_setup2, depth, flags,
			ld->sq_ring, ld->cq_ring, &ld->aio_ctx);
#else
	return -1;
#endif
}

static int fio_aioring_post_init(struct thread_data *td)
{
	struct aioring_data *ld = td->io_ops_data;
	struct aioring_options *o = td->eo;
	struct io_u *io_u;
	struct iocb *iocb;
	int err = 0;

	if (o->fixedbufs) {
		int i;

		for (i = 0; i < td->o.iodepth; i++) {
			io_u = ld->io_u_index[i];
			iocb = &ld->iocbs[i];
			iocb->u.c.buf = io_u->buf;
			iocb->u.c.nbytes = td_max_bs(td);

			if (o->hipri)
				iocb->u.c.flags |= IOCB_FLAG_HIPRI;
		}
	}

	err = fio_aioring_queue_init(td);
	if (err) {
		td_verror(td, -err, "io_queue_init");
		return 1;
	}

	return 0;
}

static int fio_aioring_init(struct thread_data *td)
{
	struct aioring_data *ld;

	if (td->o.iodepth <= 1) {
		printf("aio-ring: needs a minimum QD of 2\n");
		return 1;
	}

	ld = calloc(1, sizeof(*ld));

	/* io_u index */
	ld->io_u_index = calloc(td->o.iodepth, sizeof(struct io_u *));
	ld->io_us = calloc(td->o.iodepth, sizeof(struct io_u *));

	ld->iocbs = fio_memalign(page_size, aioring_sq_iocb(td), false);
	memset(ld->iocbs, 0, aioring_sq_iocb(td));

	ld->sq_ring = fio_memalign(page_size, aioring_sq_size(td), false);
	memset(ld->sq_ring, 0, aioring_sq_size(td));
	ld->sq_ring->nr_events = td->o.iodepth;
	ld->sq_ring->iocbs = (u64) ld->iocbs;

	ld->cq_ring = fio_memalign(page_size, aioring_cq_size(td), false);
	memset(ld->cq_ring, 0, aioring_cq_size(td));
	ld->cq_ring->nr_events = td->o.iodepth * 2;

	td->io_ops_data = ld;
	return 0;
}

static int fio_aioring_io_u_init(struct thread_data *td, struct io_u *io_u)
{
	struct aioring_data *ld = td->io_ops_data;

	ld->io_u_index[io_u->index] = io_u;
	return 0;
}

static struct ioengine_ops ioengine = {
	.name			= "aio-ring",
	.version		= FIO_IOOPS_VERSION,
	.init			= fio_aioring_init,
	.post_init		= fio_aioring_post_init,
	.io_u_init		= fio_aioring_io_u_init,
	.prep			= fio_aioring_prep,
	.queue			= fio_aioring_queue,
	.commit			= fio_aioring_commit,
	.getevents		= fio_aioring_getevents,
	.event			= fio_aioring_event,
	.cleanup		= fio_aioring_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 aioring_options),
};

static void fio_init fio_aioring_register(void)
{
#ifdef __NR_sys_io_setup2
#ifdef __NR_sys_io_ring_enter
	register_ioengine(&ioengine);
#endif
#endif
}

static void fio_exit fio_aioring_unregister(void)
{
#ifdef __NR_sys_io_setup2
#ifdef __NR_sys_io_ring_enter
	unregister_ioengine(&ioengine);
#endif
#endif
}
Commit	Line	Data
52885fa2 JA	1	/*
	2	* aioring engine
	3	*
	4	* IO engine using the new native Linux libaio ring interface
	5	*
	6	*/
	7	#include <stdlib.h>
	8	#include <unistd.h>
	9	#include <errno.h>
	10	#include <libaio.h>
	11	#include <sys/time.h>
	12	#include <sys/resource.h>
	13
	14	#include "../fio.h"
	15	#include "../lib/pow2.h"
	16	#include "../optgroup.h"
	17	#include "../lib/memalign.h"
	18
	19	#ifndef IOCB_FLAG_HIPRI
	20	#define IOCB_FLAG_HIPRI (1 << 2)
	21	#endif
	22
	23	/*
	24	* io_setup2(2) flags
	25	*/
	26	#ifndef IOCTX_FLAG_IOPOLL
	27	#define IOCTX_FLAG_IOPOLL (1 << 0)
	28	#endif
	29	#ifndef IOCTX_FLAG_SCQRING
	30	#define IOCTX_FLAG_SCQRING (1 << 1)
	31	#endif
	32	#ifndef IOCTX_FLAG_FIXEDBUFS
	33	#define IOCTX_FLAG_FIXEDBUFS (1 << 2)
	34	#endif
	35	#ifndef IOCTX_FLAG_SQTHREAD
	36	#define IOCTX_FLAG_SQTHREAD (1 << 3)
	37	#endif
	38	#ifndef IOCTX_FLAG_SQWQ
	39	#define IOCTX_FLAG_SQWQ (1 << 4)
	40	#endif
	41
	42	/*
	43	* io_ring_enter(2) flags
	44	*/
	45	#ifndef IORING_FLAG_SUBMIT
	46	#define IORING_FLAG_SUBMIT (1 << 0)
	47	#endif
	48	#ifndef IORING_FLAG_GETEVENTS
	49	#define IORING_FLAG_GETEVENTS (1 << 1)
	50	#endif
	51
	52	typedef uint64_t u64;
	53	typedef uint32_t u32;
	54	typedef uint16_t u16;
	55
	56	struct aio_sq_ring {
	57	union {
	58	struct {
	59	u32 head;
	60	u32 tail;
	61	u32 nr_events;
	62	u16 sq_thread_cpu;
	63	u64 iocbs;
	64	};
65	u32 pad[16];
66	};
67	u32 array[0];
68	};
69
70	struct aio_cq_ring {
71	union {
72	struct {
73	u32 head;
74	u32 tail;
75	u32 nr_events;
76	};
77	struct io_event pad;
78	};
79	struct io_event events[0];
80	};
81
82	struct aioring_data {
83	io_context_t aio_ctx;
84	struct io_u **io_us;
85	struct io_u **io_u_index;
86
87	struct aio_sq_ring *sq_ring;
88	struct iocb *iocbs;
89
90	struct aio_cq_ring *cq_ring;
91	struct io_event *events;
92
93	int queued;
94	int cq_ring_off;
95	};
96
97	struct aioring_options {
98	void *pad;
99	unsigned int hipri;
100	unsigned int fixedbufs;
101	};
102
103	static struct fio_option options[] = {
104	{
105	.name = "hipri",
106	.lname = "High Priority",
107	.type = FIO_OPT_STR_SET,
108	.off1 = offsetof(struct aioring_options, hipri),
109	.help = "Use polled IO completions",
110	.category = FIO_OPT_C_ENGINE,
111	.group = FIO_OPT_G_LIBAIO,
112	},
113	{
114	.name = "fixedbufs",
115	.lname = "Fixed (pre-mapped) IO buffers",
116	.type = FIO_OPT_STR_SET,
117	.off1 = offsetof(struct aioring_options, fixedbufs),
118	.help = "Pre map IO buffers",
119	.category = FIO_OPT_C_ENGINE,
120	.group = FIO_OPT_G_LIBAIO,
121	},
122	{
123	.name = NULL,
124	},
125	};
126
127	static int fio_aioring_commit(struct thread_data *td);
128
129	static int io_ring_enter(io_context_t ctx, unsigned int to_submit,
130	unsigned int min_complete, unsigned int flags)
131	{
132	#ifdef __NR_sys_io_ring_enter
133	return syscall(__NR_sys_io_ring_enter, ctx, to_submit, min_complete,
134	flags);
135	#else
136	return -1;
137	#endif
138	}
139
140	static int fio_aioring_prep(struct thread_data td, struct io_u io_u)
141	{
142	struct aioring_data *ld = td->io_ops_data;
143	struct fio_file *f = io_u->file;
144	struct aioring_options *o = td->eo;
145	struct iocb *iocb;
146
147	iocb = &ld->iocbs[io_u->index];
148
149	if (io_u->ddir == DDIR_READ) {
150	if (o->fixedbufs) {
151	iocb->aio_fildes = f->fd;
152	iocb->aio_lio_opcode = IO_CMD_PREAD;
153	iocb->u.c.offset = io_u->offset;
154	} else {
155	io_prep_pread(iocb, f->fd, io_u->xfer_buf, io_u->xfer_buflen, io_u->offset);
156	if (o->hipri)
157	iocb->u.c.flags \|= IOCB_FLAG_HIPRI;
158	}
159	} else if (io_u->ddir == DDIR_WRITE) {
160	if (o->fixedbufs) {
161	iocb->aio_fildes = f->fd;
162	iocb->aio_lio_opcode = IO_CMD_PWRITE;
163	iocb->u.c.offset = io_u->offset;
164	} else {
165	io_prep_pwrite(iocb, f->fd, io_u->xfer_buf, io_u->xfer_buflen, io_u->offset);
166	if (o->hipri)
167	iocb->u.c.flags \|= IOCB_FLAG_HIPRI;
168	}
169	} else if (ddir_sync(io_u->ddir))
170	io_prep_fsync(iocb, f->fd);
171
172	iocb->data = io_u;
173	return 0;
174	}
175
176	static struct io_u fio_aioring_event(struct thread_data td, int event)
177	{
178	struct aioring_data *ld = td->io_ops_data;
179	struct io_event *ev;
180	struct io_u *io_u;
181	int index;
182
183	index = event + ld->cq_ring_off;
184	if (index >= ld->cq_ring->nr_events)
185	index -= ld->cq_ring->nr_events;
186
187	ev = &ld->cq_ring->events[index];
188	io_u = ev->data;
189
190	if (ev->res != io_u->xfer_buflen) {
191	if (ev->res > io_u->xfer_buflen)
192	io_u->error = -ev->res;
193	else
194	io_u->resid = io_u->xfer_buflen - ev->res;
195	} else
196	io_u->error = 0;
197
198	return io_u;
199	}
200
201	static int fio_aioring_cqring_reap(struct thread_data *td, unsigned int events,
202	unsigned int max)
203	{
204	struct aioring_data *ld = td->io_ops_data;
205	struct aio_cq_ring *ring = ld->cq_ring;
206	u32 head, reaped = 0;
207
208	head = ring->head;
209	do {
210	read_barrier();
211	if (head == ring->tail)
212	break;
213	reaped++;
214	head++;
215	if (head == ring->nr_events)
216	head = 0;
217	} while (reaped + events < max);
218
219	ring->head = head;
220	write_barrier();
221	return reaped;
222	}
223
224	static int fio_aioring_getevents(struct thread_data *td, unsigned int min,
225	unsigned int max, const struct timespec *t)
226	{
227	struct aioring_data *ld = td->io_ops_data;
228	unsigned actual_min = td->o.iodepth_batch_complete_min == 0 ? 0 : min;
229	struct aio_cq_ring *ring = ld->cq_ring;
230	int r, events = 0;
231
232	ld->cq_ring_off = ring->head;
233	do {
234	r = fio_aioring_cqring_reap(td, events, max);
235	if (r) {
236	events += r;
237	continue;
238	}
239
240	r = io_ring_enter(ld->aio_ctx, 0, actual_min,
241	IORING_FLAG_GETEVENTS);
242	if (r < 0) {
243	if (errno == EAGAIN)
244	continue;
245	perror("ring enter");
246	break;
247	}
248	} while (events < min);
249
250	return r < 0 ? r : events;
251	}
252
253	static enum fio_q_status fio_aioring_queue(struct thread_data *td,
254	struct io_u *io_u)
255	{
256	struct aioring_data *ld = td->io_ops_data;
257	struct aio_sq_ring *ring = ld->sq_ring;
258	unsigned tail, next_tail;
259
260	fio_ro_check(td, io_u);
261
262	if (ld->queued == td->o.iodepth)
263	return FIO_Q_BUSY;
264
265	/*
266	* fsync is tricky, since it can fail and we need to do it
267	* serialized with other io. the reason is that linux doesn't
268	* support aio fsync yet. So return busy for the case where we
269	* have pending io, to let fio complete those first.
270	*/
271	if (ddir_sync(io_u->ddir)) {
272	if (ld->queued)
273	return FIO_Q_BUSY;
274
275	do_io_u_sync(td, io_u);
276	return FIO_Q_COMPLETED;
277	}
278
279	if (io_u->ddir == DDIR_TRIM) {
280	if (ld->queued)
281	return FIO_Q_BUSY;
282
283	do_io_u_trim(td, io_u);
284	io_u_mark_submit(td, 1);
285	io_u_mark_complete(td, 1);
286	return FIO_Q_COMPLETED;
287	}
288
289	tail = ring->tail;
290	next_tail = tail + 1;
291	if (next_tail == ring->nr_events)
292	next_tail = 0;
293	read_barrier();
294	if (next_tail == ring->head)
295	return FIO_Q_BUSY;
296
297	ring->array[tail] = io_u->index;
298	ring->tail = next_tail;
299	write_barrier();
300
301	ld->queued++;
302	return FIO_Q_QUEUED;
303	}
304
305	static void fio_aioring_queued(struct thread_data *td, int start, int nr)
306	{
307	struct aioring_data *ld = td->io_ops_data;
308	struct timespec now;
309
310	if (!fio_fill_issue_time(td))
311	return;
312
313	fio_gettime(&now, NULL);
314
315	while (nr--) {
316	int index = ld->sq_ring->array[start];
317	struct io_u *io_u = io_u = ld->io_u_index[index];
318
319	memcpy(&io_u->issue_time, &now, sizeof(now));
320	io_u_queued(td, io_u);
321
322	start++;
323	if (start == ld->sq_ring->nr_events)
324	start = 0;
325	}
326	}
327
328	static int fio_aioring_commit(struct thread_data *td)
329	{
330	struct aioring_data *ld = td->io_ops_data;
331	int ret;
332
333	if (!ld->queued)
334	return 0;
335
336	do {
337	int start = ld->sq_ring->head;
338	long nr = ld->queued;
339
340	ret = io_ring_enter(ld->aio_ctx, nr, 0, IORING_FLAG_SUBMIT \|
341	IORING_FLAG_GETEVENTS);
342	if (ret == -1)
343	perror("io_ring_enter");
344	if (ret > 0) {
345	fio_aioring_queued(td, start, ret);
346	io_u_mark_submit(td, ret);
347
348	ld->queued -= ret;
349	ret = 0;
350	} else if (ret == -EINTR \|\| !ret) {
351	if (!ret)
352	io_u_mark_submit(td, ret);
353	continue;
354	} else if (ret == -EAGAIN) {
355	/*
356	* If we get EAGAIN, we should break out without
357	* error and let the upper layer reap some
358	* events for us. If we have no queued IO, we
359	* must loop here. If we loop for more than 30s,
360	* just error out, something must be buggy in the
361	* IO path.
362	*/
363	if (ld->queued) {
364	ret = 0;
365	break;
366	}
367	usleep(1);
368	continue;
369	} else if (ret == -ENOMEM) {
370	/*
371	* If we get -ENOMEM, reap events if we can. If
372	* we cannot, treat it as a fatal event since there's
373	* nothing we can do about it.
374	*/
375	if (ld->queued)
376	ret = 0;
377	break;
378	} else
379	break;
380	} while (ld->queued);
381
382	return ret;
383	}
384
385	static size_t aioring_cq_size(struct thread_data *td)
386	{
387	return sizeof(struct aio_cq_ring) + 2 * td->o.iodepth * sizeof(struct io_event);
388	}
389
390	static size_t aioring_sq_iocb(struct thread_data *td)
391	{
392	return sizeof(struct iocb) * td->o.iodepth;
393	}
394
395	static size_t aioring_sq_size(struct thread_data *td)
396	{
397	return sizeof(struct aio_sq_ring) + td->o.iodepth * sizeof(u32);
398	}
399
400	static void fio_aioring_cleanup(struct thread_data *td)
401	{
402	struct aioring_data *ld = td->io_ops_data;
403
404	if (ld) {
405	/*
406	* Work-around to avoid huge RCU stalls at exit time. If we
407	* don't do this here, then it'll be torn down by exit_aio().
408	* But for that case we can parallellize the freeing, thus
409	* speeding it up a lot.
410	*/
411	if (!(td->flags & TD_F_CHILD))
412	io_destroy(ld->aio_ctx);
413	free(ld->io_u_index);
414	free(ld->io_us);
415	fio_memfree(ld->sq_ring, aioring_sq_size(td), false);
416	fio_memfree(ld->iocbs, aioring_sq_iocb(td), false);
417	fio_memfree(ld->cq_ring, aioring_cq_size(td), false);
418	free(ld);
419	}
420	}
421
422	static int fio_aioring_queue_init(struct thread_data *td)
423	{
424	#ifdef __NR_sys_io_setup2
425	struct aioring_data *ld = td->io_ops_data;
426	struct aioring_options *o = td->eo;
427	int flags = IOCTX_FLAG_SCQRING;
428	int depth = td->o.iodepth;
429
430	if (o->hipri)
431	flags \|= IOCTX_FLAG_IOPOLL;
432	if (o->fixedbufs) {
433	struct rlimit rlim = {
434	.rlim_cur = RLIM_INFINITY,
435	.rlim_max = RLIM_INFINITY,
436	};
437
438	setrlimit(RLIMIT_MEMLOCK, &rlim);
439	flags \|= IOCTX_FLAG_FIXEDBUFS;
440	}
441
442	return syscall(__NR_sys_io_setup2, depth, flags,
443	ld->sq_ring, ld->cq_ring, &ld->aio_ctx);
444	#else
445	return -1;
446	#endif
447	}
448
449	static int fio_aioring_post_init(struct thread_data *td)
450	{
451	struct aioring_data *ld = td->io_ops_data;
452	struct aioring_options *o = td->eo;
453	struct io_u *io_u;
454	struct iocb *iocb;
455	int err = 0;
456
457	if (o->fixedbufs) {
458	int i;
459
460	for (i = 0; i < td->o.iodepth; i++) {
461	io_u = ld->io_u_index[i];
462	iocb = &ld->iocbs[i];
463	iocb->u.c.buf = io_u->buf;
464	iocb->u.c.nbytes = td_max_bs(td);
465
466	if (o->hipri)
467	iocb->u.c.flags \|= IOCB_FLAG_HIPRI;
468	}
469	}
470
471	err = fio_aioring_queue_init(td);
472	if (err) {
473	td_verror(td, -err, "io_queue_init");
474	return 1;
475	}
476
477	return 0;
478	}
479
480	static int fio_aioring_init(struct thread_data *td)
481	{
482	struct aioring_data *ld;
483
484	if (td->o.iodepth <= 1) {
485	printf("aio-ring: needs a minimum QD of 2\n");
486	return 1;
487	}
488
489	ld = calloc(1, sizeof(*ld));
490
491	/* io_u index */
492	ld->io_u_index = calloc(td->o.iodepth, sizeof(struct io_u *));
493	ld->io_us = calloc(td->o.iodepth, sizeof(struct io_u *));
494
495	ld->iocbs = fio_memalign(page_size, aioring_sq_iocb(td), false);
496	memset(ld->iocbs, 0, aioring_sq_iocb(td));
497
498	ld->sq_ring = fio_memalign(page_size, aioring_sq_size(td), false);
499	memset(ld->sq_ring, 0, aioring_sq_size(td));
500	ld->sq_ring->nr_events = td->o.iodepth;
501	ld->sq_ring->iocbs = (u64) ld->iocbs;
502
503	ld->cq_ring = fio_memalign(page_size, aioring_cq_size(td), false);
504	memset(ld->cq_ring, 0, aioring_cq_size(td));
505	ld->cq_ring->nr_events = td->o.iodepth * 2;
506
507	td->io_ops_data = ld;
508	return 0;
509	}
510
511	static int fio_aioring_io_u_init(struct thread_data td, struct io_u io_u)
512	{
513	struct aioring_data *ld = td->io_ops_data;
514
515	ld->io_u_index[io_u->index] = io_u;
516	return 0;
517	}
518
519	static struct ioengine_ops ioengine = {
520	.name = "aio-ring",
521	.version = FIO_IOOPS_VERSION,
522	.init = fio_aioring_init,
523	.post_init = fio_aioring_post_init,
524	.io_u_init = fio_aioring_io_u_init,
525	.prep = fio_aioring_prep,
526	.queue = fio_aioring_queue,
527	.commit = fio_aioring_commit,
528	.getevents = fio_aioring_getevents,
529	.event = fio_aioring_event,
530	.cleanup = fio_aioring_cleanup,
531	.open_file = generic_open_file,
532	.close_file = generic_close_file,
533	.get_file_size = generic_get_file_size,
534	.options = options,
535	.option_struct_size = sizeof(struct aioring_options),
536	};
537
538	static void fio_init fio_aioring_register(void)
539	{
540	#ifdef __NR_sys_io_setup2
541	#ifdef __NR_sys_io_ring_enter
542	register_ioengine(&ioengine);
543	#endif
544	#endif
545	}
546
547	static void fio_exit fio_aioring_unregister(void)
548	{
549	#ifdef __NR_sys_io_setup2
550	#ifdef __NR_sys_io_ring_enter
551	unregister_ioengine(&ioengine);
552	#endif
553	#endif
554	}