hw/arm/virt: formatting: memory map
[qemu/ar7.git] / block / linux-aio.c
blob7ac7e8c99c72f6e6e2c7b8fb176835e8c6a8e95d
1 /*
2 * Linux native AIO support.
4 * Copyright (C) 2009 IBM, Corp.
5 * Copyright (C) 2009 Red Hat, Inc.
7 * This work is licensed under the terms of the GNU GPL, version 2 or later.
8 * See the COPYING file in the top-level directory.
9 */
10 #include "qemu-common.h"
11 #include "block/aio.h"
12 #include "qemu/queue.h"
13 #include "block/raw-aio.h"
14 #include "qemu/event_notifier.h"
16 #include <libaio.h>
19 * Queue size (per-device).
21 * XXX: eventually we need to communicate this to the guest and/or make it
22 * tunable by the guest. If we get more outstanding requests at a time
23 * than this we will get EAGAIN from io_submit which is communicated to
24 * the guest as an I/O error.
26 #define MAX_EVENTS 128
28 #define MAX_QUEUED_IO 128
30 struct qemu_laiocb {
31 BlockDriverAIOCB common;
32 struct qemu_laio_state *ctx;
33 struct iocb iocb;
34 ssize_t ret;
35 size_t nbytes;
36 QEMUIOVector *qiov;
37 bool is_read;
38 QLIST_ENTRY(qemu_laiocb) node;
41 typedef struct {
42 struct iocb *iocbs[MAX_QUEUED_IO];
43 int plugged;
44 unsigned int size;
45 unsigned int idx;
46 } LaioQueue;
48 struct qemu_laio_state {
49 io_context_t ctx;
50 EventNotifier e;
52 /* io queue for submit at batch */
53 LaioQueue io_q;
56 static inline ssize_t io_event_ret(struct io_event *ev)
58 return (ssize_t)(((uint64_t)ev->res2 << 32) | ev->res);
62 * Completes an AIO request (calls the callback and frees the ACB).
64 static void qemu_laio_process_completion(struct qemu_laio_state *s,
65 struct qemu_laiocb *laiocb)
67 int ret;
69 ret = laiocb->ret;
70 if (ret != -ECANCELED) {
71 if (ret == laiocb->nbytes) {
72 ret = 0;
73 } else if (ret >= 0) {
74 /* Short reads mean EOF, pad with zeros. */
75 if (laiocb->is_read) {
76 qemu_iovec_memset(laiocb->qiov, ret, 0,
77 laiocb->qiov->size - ret);
78 } else {
79 ret = -EINVAL;
83 laiocb->common.cb(laiocb->common.opaque, ret);
86 qemu_aio_release(laiocb);
89 static void qemu_laio_completion_cb(EventNotifier *e)
91 struct qemu_laio_state *s = container_of(e, struct qemu_laio_state, e);
93 while (event_notifier_test_and_clear(&s->e)) {
94 struct io_event events[MAX_EVENTS];
95 struct timespec ts = { 0 };
96 int nevents, i;
98 do {
99 nevents = io_getevents(s->ctx, MAX_EVENTS, MAX_EVENTS, events, &ts);
100 } while (nevents == -EINTR);
102 for (i = 0; i < nevents; i++) {
103 struct iocb *iocb = events[i].obj;
104 struct qemu_laiocb *laiocb =
105 container_of(iocb, struct qemu_laiocb, iocb);
107 laiocb->ret = io_event_ret(&events[i]);
108 qemu_laio_process_completion(s, laiocb);
113 static void laio_cancel(BlockDriverAIOCB *blockacb)
115 struct qemu_laiocb *laiocb = (struct qemu_laiocb *)blockacb;
116 struct io_event event;
117 int ret;
119 if (laiocb->ret != -EINPROGRESS)
120 return;
123 * Note that as of Linux 2.6.31 neither the block device code nor any
124 * filesystem implements cancellation of AIO request.
125 * Thus the polling loop below is the normal code path.
127 ret = io_cancel(laiocb->ctx->ctx, &laiocb->iocb, &event);
128 if (ret == 0) {
129 laiocb->ret = -ECANCELED;
130 return;
134 * We have to wait for the iocb to finish.
136 * The only way to get the iocb status update is by polling the io context.
137 * We might be able to do this slightly more optimal by removing the
138 * O_NONBLOCK flag.
140 while (laiocb->ret == -EINPROGRESS) {
141 qemu_laio_completion_cb(&laiocb->ctx->e);
145 static const AIOCBInfo laio_aiocb_info = {
146 .aiocb_size = sizeof(struct qemu_laiocb),
147 .cancel = laio_cancel,
150 static void ioq_init(LaioQueue *io_q)
152 io_q->size = MAX_QUEUED_IO;
153 io_q->idx = 0;
154 io_q->plugged = 0;
157 static int ioq_submit(struct qemu_laio_state *s)
159 int ret, i = 0;
160 int len = s->io_q.idx;
162 do {
163 ret = io_submit(s->ctx, len, s->io_q.iocbs);
164 } while (i++ < 3 && ret == -EAGAIN);
166 /* empty io queue */
167 s->io_q.idx = 0;
169 if (ret < 0) {
170 i = 0;
171 } else {
172 i = ret;
175 for (; i < len; i++) {
176 struct qemu_laiocb *laiocb =
177 container_of(s->io_q.iocbs[i], struct qemu_laiocb, iocb);
179 laiocb->ret = (ret < 0) ? ret : -EIO;
180 qemu_laio_process_completion(s, laiocb);
182 return ret;
185 static void ioq_enqueue(struct qemu_laio_state *s, struct iocb *iocb)
187 unsigned int idx = s->io_q.idx;
189 s->io_q.iocbs[idx++] = iocb;
190 s->io_q.idx = idx;
192 /* submit immediately if queue is full */
193 if (idx == s->io_q.size) {
194 ioq_submit(s);
198 void laio_io_plug(BlockDriverState *bs, void *aio_ctx)
200 struct qemu_laio_state *s = aio_ctx;
202 s->io_q.plugged++;
205 int laio_io_unplug(BlockDriverState *bs, void *aio_ctx, bool unplug)
207 struct qemu_laio_state *s = aio_ctx;
208 int ret = 0;
210 assert(s->io_q.plugged > 0 || !unplug);
212 if (unplug && --s->io_q.plugged > 0) {
213 return 0;
216 if (s->io_q.idx > 0) {
217 ret = ioq_submit(s);
220 return ret;
223 BlockDriverAIOCB *laio_submit(BlockDriverState *bs, void *aio_ctx, int fd,
224 int64_t sector_num, QEMUIOVector *qiov, int nb_sectors,
225 BlockDriverCompletionFunc *cb, void *opaque, int type)
227 struct qemu_laio_state *s = aio_ctx;
228 struct qemu_laiocb *laiocb;
229 struct iocb *iocbs;
230 off_t offset = sector_num * 512;
232 laiocb = qemu_aio_get(&laio_aiocb_info, bs, cb, opaque);
233 laiocb->nbytes = nb_sectors * 512;
234 laiocb->ctx = s;
235 laiocb->ret = -EINPROGRESS;
236 laiocb->is_read = (type == QEMU_AIO_READ);
237 laiocb->qiov = qiov;
239 iocbs = &laiocb->iocb;
241 switch (type) {
242 case QEMU_AIO_WRITE:
243 io_prep_pwritev(iocbs, fd, qiov->iov, qiov->niov, offset);
244 break;
245 case QEMU_AIO_READ:
246 io_prep_preadv(iocbs, fd, qiov->iov, qiov->niov, offset);
247 break;
248 /* Currently Linux kernel does not support other operations */
249 default:
250 fprintf(stderr, "%s: invalid AIO request type 0x%x.\n",
251 __func__, type);
252 goto out_free_aiocb;
254 io_set_eventfd(&laiocb->iocb, event_notifier_get_fd(&s->e));
256 if (!s->io_q.plugged) {
257 if (io_submit(s->ctx, 1, &iocbs) < 0) {
258 goto out_free_aiocb;
260 } else {
261 ioq_enqueue(s, iocbs);
263 return &laiocb->common;
265 out_free_aiocb:
266 qemu_aio_release(laiocb);
267 return NULL;
270 void laio_detach_aio_context(void *s_, AioContext *old_context)
272 struct qemu_laio_state *s = s_;
274 aio_set_event_notifier(old_context, &s->e, NULL);
277 void laio_attach_aio_context(void *s_, AioContext *new_context)
279 struct qemu_laio_state *s = s_;
281 aio_set_event_notifier(new_context, &s->e, qemu_laio_completion_cb);
284 void *laio_init(void)
286 struct qemu_laio_state *s;
288 s = g_malloc0(sizeof(*s));
289 if (event_notifier_init(&s->e, false) < 0) {
290 goto out_free_state;
293 if (io_setup(MAX_EVENTS, &s->ctx) != 0) {
294 goto out_close_efd;
297 ioq_init(&s->io_q);
299 return s;
301 out_close_efd:
302 event_notifier_cleanup(&s->e);
303 out_free_state:
304 g_free(s);
305 return NULL;
308 void laio_cleanup(void *s_)
310 struct qemu_laio_state *s = s_;
312 event_notifier_cleanup(&s->e);
314 if (io_destroy(s->ctx) != 0) {
315 fprintf(stderr, "%s: destroy AIO context %p failed\n",
316 __func__, &s->ctx);
318 g_free(s);