aio.c 45.8 KB
Newer Older
Linus Torvalds's avatar
Linus Torvalds committed
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
/*
 *	An async IO implementation for Linux
 *	Written by Benjamin LaHaise <bcrl@kvack.org>
 *
 *	Implements an efficient asynchronous io interface.
 *
 *	Copyright 2000, 2001, 2002 Red Hat, Inc.  All Rights Reserved.
 *
 *	See ../COPYING for licensing terms.
 */
#include <linux/kernel.h>
#include <linux/init.h>
#include <linux/errno.h>
#include <linux/time.h>
#include <linux/aio_abi.h>
16
#include <linux/export.h>
Linus Torvalds's avatar
Linus Torvalds committed
17
#include <linux/syscalls.h>
18
#include <linux/backing-dev.h>
19
#include <linux/uio.h>
Linus Torvalds's avatar
Linus Torvalds committed
20
21
22
23
24
25
26
27

#define DEBUG 0

#include <linux/sched.h>
#include <linux/fs.h>
#include <linux/file.h>
#include <linux/mm.h>
#include <linux/mman.h>
28
#include <linux/mmu_context.h>
Linus Torvalds's avatar
Linus Torvalds committed
29
30
31
32
33
34
#include <linux/slab.h>
#include <linux/timer.h>
#include <linux/aio.h>
#include <linux/highmem.h>
#include <linux/workqueue.h>
#include <linux/security.h>
35
#include <linux/eventfd.h>
Jeff Moyer's avatar
Jeff Moyer committed
36
#include <linux/blkdev.h>
37
#include <linux/compat.h>
Linus Torvalds's avatar
Linus Torvalds committed
38
39
40
41
42
43
44
45
46
47
48

#include <asm/kmap_types.h>
#include <asm/uaccess.h>

#if DEBUG > 1
#define dprintk		printk
#else
#define dprintk(x...)	do { ; } while (0)
#endif

/*------ sysctl variables----*/
49
50
51
static DEFINE_SPINLOCK(aio_nr_lock);
unsigned long aio_nr;		/* current system wide number of aio requests */
unsigned long aio_max_nr = 0x10000; /* system wide maximum number of aio requests */
Linus Torvalds's avatar
Linus Torvalds committed
52
53
/*----end sysctl variables---*/

54
55
static struct kmem_cache	*kiocb_cachep;
static struct kmem_cache	*kioctx_cachep;
Linus Torvalds's avatar
Linus Torvalds committed
56
57
58

static struct workqueue_struct *aio_wq;

59
static void aio_kick_handler(struct work_struct *);
60
static void aio_queue_work(struct kioctx *);
Linus Torvalds's avatar
Linus Torvalds committed
61
62
63
64
65
66
67

/* aio_setup
 *	Creates the slab caches used by the aio routines, panic on
 *	failure as this is done early during the boot sequence.
 */
static int __init aio_setup(void)
{
68
69
	kiocb_cachep = KMEM_CACHE(kiocb, SLAB_HWCACHE_ALIGN|SLAB_PANIC);
	kioctx_cachep = KMEM_CACHE(kioctx,SLAB_HWCACHE_ALIGN|SLAB_PANIC);
Linus Torvalds's avatar
Linus Torvalds committed
70

71
	aio_wq = alloc_workqueue("aio", 0, 1);	/* used to limit concurrency */
72
	BUG_ON(!aio_wq);
Linus Torvalds's avatar
Linus Torvalds committed
73
74
75
76
77

	pr_debug("aio_setup: sizeof(struct page) = %d\n", (int)sizeof(struct page));

	return 0;
}
78
__initcall(aio_setup);
Linus Torvalds's avatar
Linus Torvalds committed
79
80
81
82
83
84
85
86
87

static void aio_free_ring(struct kioctx *ctx)
{
	struct aio_ring_info *info = &ctx->ring_info;
	long i;

	for (i=0; i<info->nr_pages; i++)
		put_page(info->ring_pages[i]);

88
89
	if (info->mmap_size) {
		BUG_ON(ctx->mm != current->mm);
Al Viro's avatar
Al Viro committed
90
		vm_munmap(info->mmap_base, info->mmap_size);
91
	}
Linus Torvalds's avatar
Linus Torvalds committed
92
93
94
95
96
97
98
99
100
101
102
103

	if (info->ring_pages && info->ring_pages != info->internal_pages)
		kfree(info->ring_pages);
	info->ring_pages = NULL;
	info->nr = 0;
}

static int aio_setup_ring(struct kioctx *ctx)
{
	struct aio_ring *ring;
	struct aio_ring_info *info = &ctx->ring_info;
	unsigned nr_events = ctx->max_reqs;
104
	unsigned long size, populate;
Linus Torvalds's avatar
Linus Torvalds committed
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
	int nr_pages;

	/* Compensate for the ring buffer's head/tail overlap entry */
	nr_events += 2;	/* 1 is required, 2 for good luck */

	size = sizeof(struct aio_ring);
	size += sizeof(struct io_event) * nr_events;
	nr_pages = (size + PAGE_SIZE-1) >> PAGE_SHIFT;

	if (nr_pages < 0)
		return -EINVAL;

	nr_events = (PAGE_SIZE * nr_pages - sizeof(struct aio_ring)) / sizeof(struct io_event);

	info->nr = 0;
	info->ring_pages = info->internal_pages;
	if (nr_pages > AIO_RING_PAGES) {
122
		info->ring_pages = kcalloc(nr_pages, sizeof(struct page *), GFP_KERNEL);
Linus Torvalds's avatar
Linus Torvalds committed
123
124
125
126
127
128
129
		if (!info->ring_pages)
			return -ENOMEM;
	}

	info->mmap_size = nr_pages * PAGE_SIZE;
	dprintk("attempting mmap of %lu bytes\n", info->mmap_size);
	down_write(&ctx->mm->mmap_sem);
130
131
	info->mmap_base = do_mmap_pgoff(NULL, 0, info->mmap_size, 
					PROT_READ|PROT_WRITE,
132
133
					MAP_ANONYMOUS|MAP_PRIVATE, 0,
					&populate);
Linus Torvalds's avatar
Linus Torvalds committed
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
	if (IS_ERR((void *)info->mmap_base)) {
		up_write(&ctx->mm->mmap_sem);
		info->mmap_size = 0;
		aio_free_ring(ctx);
		return -EAGAIN;
	}

	dprintk("mmap address: 0x%08lx\n", info->mmap_base);
	info->nr_pages = get_user_pages(current, ctx->mm,
					info->mmap_base, nr_pages, 
					1, 0, info->ring_pages, NULL);
	up_write(&ctx->mm->mmap_sem);

	if (unlikely(info->nr_pages != nr_pages)) {
		aio_free_ring(ctx);
		return -EAGAIN;
	}
151
	if (populate)
152
		mm_populate(info->mmap_base, populate);
Linus Torvalds's avatar
Linus Torvalds committed
153
154
155
156
157

	ctx->user_id = info->mmap_base;

	info->nr = nr_events;		/* trusted copy */

158
	ring = kmap_atomic(info->ring_pages[0]);
Linus Torvalds's avatar
Linus Torvalds committed
159
160
161
162
163
164
165
	ring->nr = nr_events;	/* user copy */
	ring->id = ctx->user_id;
	ring->head = ring->tail = 0;
	ring->magic = AIO_RING_MAGIC;
	ring->compat_features = AIO_RING_COMPAT_FEATURES;
	ring->incompat_features = AIO_RING_INCOMPAT_FEATURES;
	ring->header_length = sizeof(struct aio_ring);
166
	kunmap_atomic(ring);
Linus Torvalds's avatar
Linus Torvalds committed
167
168
169
170
171
172

	return 0;
}


/* aio_ring_event: returns a pointer to the event at the given index from
173
 * kmap_atomic().  Release the pointer with put_aio_ring_event();
Linus Torvalds's avatar
Linus Torvalds committed
174
175
176
177
178
 */
#define AIO_EVENTS_PER_PAGE	(PAGE_SIZE / sizeof(struct io_event))
#define AIO_EVENTS_FIRST_PAGE	((PAGE_SIZE - sizeof(struct aio_ring)) / sizeof(struct io_event))
#define AIO_EVENTS_OFFSET	(AIO_EVENTS_PER_PAGE - AIO_EVENTS_FIRST_PAGE)

179
#define aio_ring_event(info, nr) ({					\
Linus Torvalds's avatar
Linus Torvalds committed
180
181
182
	unsigned pos = (nr) + AIO_EVENTS_OFFSET;			\
	struct io_event *__event;					\
	__event = kmap_atomic(						\
183
			(info)->ring_pages[pos / AIO_EVENTS_PER_PAGE]); \
Linus Torvalds's avatar
Linus Torvalds committed
184
185
186
187
	__event += pos % AIO_EVENTS_PER_PAGE;				\
	__event;							\
})

188
#define put_aio_ring_event(event) do {		\
Linus Torvalds's avatar
Linus Torvalds committed
189
190
	struct io_event *__event = (event);	\
	(void)__event;				\
191
	kunmap_atomic((void *)((unsigned long)__event & PAGE_MASK)); \
Linus Torvalds's avatar
Linus Torvalds committed
192
193
} while(0)

194
195
196
197
198
static void ctx_rcu_free(struct rcu_head *head)
{
	struct kioctx *ctx = container_of(head, struct kioctx, rcu_head);
	kmem_cache_free(kioctx_cachep, ctx);
}
199
200
201
202
203
204
205

/* __put_ioctx
 *	Called when the last user of an aio context has gone away,
 *	and the struct needs to be freed.
 */
static void __put_ioctx(struct kioctx *ctx)
{
206
	unsigned nr_events = ctx->max_reqs;
207
208
	BUG_ON(ctx->reqs_active);

Al Viro's avatar
Al Viro committed
209
	cancel_delayed_work_sync(&ctx->wq);
210
211
212
	aio_free_ring(ctx);
	mmdrop(ctx->mm);
	ctx->mm = NULL;
213
214
215
216
217
218
	if (nr_events) {
		spin_lock(&aio_nr_lock);
		BUG_ON(aio_nr - nr_events > aio_nr);
		aio_nr -= nr_events;
		spin_unlock(&aio_nr_lock);
	}
219
	pr_debug("__put_ioctx: freeing %p\n", ctx);
220
	call_rcu(&ctx->rcu_head, ctx_rcu_free);
221
222
}

Nick Piggin's avatar
Nick Piggin committed
223
224
225
226
227
228
229
230
231
232
233
static inline int try_get_ioctx(struct kioctx *kioctx)
{
	return atomic_inc_not_zero(&kioctx->users);
}

static inline void put_ioctx(struct kioctx *kioctx)
{
	BUG_ON(atomic_read(&kioctx->users) <= 0);
	if (unlikely(atomic_dec_and_test(&kioctx->users)))
		__put_ioctx(kioctx);
}
234

Linus Torvalds's avatar
Linus Torvalds committed
235
236
237
238
239
240
241
/* ioctx_alloc
 *	Allocates and initializes an ioctx.  Returns an ERR_PTR if it failed.
 */
static struct kioctx *ioctx_alloc(unsigned nr_events)
{
	struct mm_struct *mm;
	struct kioctx *ctx;
242
	int err = -ENOMEM;
Linus Torvalds's avatar
Linus Torvalds committed
243
244
245
246
247
248
249
250

	/* Prevent overflows */
	if ((nr_events > (0x10000000U / sizeof(struct io_event))) ||
	    (nr_events > (0x10000000U / sizeof(struct kiocb)))) {
		pr_debug("ENOMEM: nr_events too high\n");
		return ERR_PTR(-EINVAL);
	}

251
	if (!nr_events || (unsigned long)nr_events > aio_max_nr)
Linus Torvalds's avatar
Linus Torvalds committed
252
253
		return ERR_PTR(-EAGAIN);

254
	ctx = kmem_cache_zalloc(kioctx_cachep, GFP_KERNEL);
Linus Torvalds's avatar
Linus Torvalds committed
255
256
257
258
259
260
261
	if (!ctx)
		return ERR_PTR(-ENOMEM);

	ctx->max_reqs = nr_events;
	mm = ctx->mm = current->mm;
	atomic_inc(&mm->mm_count);

Al Viro's avatar
Al Viro committed
262
	atomic_set(&ctx->users, 2);
Linus Torvalds's avatar
Linus Torvalds committed
263
264
265
266
267
268
	spin_lock_init(&ctx->ctx_lock);
	spin_lock_init(&ctx->ring_info.ring_lock);
	init_waitqueue_head(&ctx->wait);

	INIT_LIST_HEAD(&ctx->active_reqs);
	INIT_LIST_HEAD(&ctx->run_list);
269
	INIT_DELAYED_WORK(&ctx->wq, aio_kick_handler);
Linus Torvalds's avatar
Linus Torvalds committed
270
271
272
273
274

	if (aio_setup_ring(ctx) < 0)
		goto out_freectx;

	/* limit the number of system wide aios */
275
	spin_lock(&aio_nr_lock);
276
277
	if (aio_nr + nr_events > aio_max_nr ||
	    aio_nr + nr_events < aio_nr) {
278
		spin_unlock(&aio_nr_lock);
Linus Torvalds's avatar
Linus Torvalds committed
279
		goto out_cleanup;
280
281
	}
	aio_nr += ctx->max_reqs;
282
	spin_unlock(&aio_nr_lock);
Linus Torvalds's avatar
Linus Torvalds committed
283

Jeff Moyer's avatar
Jeff Moyer committed
284
	/* now link into global list. */
285
286
287
	spin_lock(&mm->ioctx_lock);
	hlist_add_head_rcu(&ctx->list, &mm->ioctx_list);
	spin_unlock(&mm->ioctx_lock);
Linus Torvalds's avatar
Linus Torvalds committed
288
289
290
291
292
293

	dprintk("aio: allocated ioctx %p[%ld]: mm=%p mask=0x%x\n",
		ctx, ctx->user_id, current->mm, ctx->ring_info.nr);
	return ctx;

out_cleanup:
294
295
	err = -EAGAIN;
	aio_free_ring(ctx);
Linus Torvalds's avatar
Linus Torvalds committed
296
297
298
out_freectx:
	mmdrop(mm);
	kmem_cache_free(kioctx_cachep, ctx);
299
300
	dprintk("aio: error allocating ioctx %d\n", err);
	return ERR_PTR(err);
Linus Torvalds's avatar
Linus Torvalds committed
301
302
}

303
/* kill_ctx
Linus Torvalds's avatar
Linus Torvalds committed
304
305
306
307
 *	Cancels all outstanding aio requests on an aio context.  Used 
 *	when the processes owning a context have all exited to encourage 
 *	the rapid destruction of the kioctx.
 */
308
static void kill_ctx(struct kioctx *ctx)
Linus Torvalds's avatar
Linus Torvalds committed
309
310
{
	int (*cancel)(struct kiocb *, struct io_event *);
311
312
	struct task_struct *tsk = current;
	DECLARE_WAITQUEUE(wait, tsk);
Linus Torvalds's avatar
Linus Torvalds committed
313
	struct io_event res;
314

Linus Torvalds's avatar
Linus Torvalds committed
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
	spin_lock_irq(&ctx->ctx_lock);
	ctx->dead = 1;
	while (!list_empty(&ctx->active_reqs)) {
		struct list_head *pos = ctx->active_reqs.next;
		struct kiocb *iocb = list_kiocb(pos);
		list_del_init(&iocb->ki_list);
		cancel = iocb->ki_cancel;
		kiocbSetCancelled(iocb);
		if (cancel) {
			iocb->ki_users++;
			spin_unlock_irq(&ctx->ctx_lock);
			cancel(iocb, &res);
			spin_lock_irq(&ctx->ctx_lock);
		}
	}

	if (!ctx->reqs_active)
332
		goto out;
Linus Torvalds's avatar
Linus Torvalds committed
333
334
335
336

	add_wait_queue(&ctx->wait, &wait);
	set_task_state(tsk, TASK_UNINTERRUPTIBLE);
	while (ctx->reqs_active) {
337
		spin_unlock_irq(&ctx->ctx_lock);
338
		io_schedule();
Linus Torvalds's avatar
Linus Torvalds committed
339
		set_task_state(tsk, TASK_UNINTERRUPTIBLE);
340
		spin_lock_irq(&ctx->ctx_lock);
Linus Torvalds's avatar
Linus Torvalds committed
341
342
343
	}
	__set_task_state(tsk, TASK_RUNNING);
	remove_wait_queue(&ctx->wait, &wait);
344
345
346

out:
	spin_unlock_irq(&ctx->ctx_lock);
Linus Torvalds's avatar
Linus Torvalds committed
347
348
349
350
351
}

/* wait_on_sync_kiocb:
 *	Waits on the given sync kiocb to complete.
 */
352
ssize_t wait_on_sync_kiocb(struct kiocb *iocb)
Linus Torvalds's avatar
Linus Torvalds committed
353
354
355
356
357
{
	while (iocb->ki_users) {
		set_current_state(TASK_UNINTERRUPTIBLE);
		if (!iocb->ki_users)
			break;
358
		io_schedule();
Linus Torvalds's avatar
Linus Torvalds committed
359
360
361
362
	}
	__set_current_state(TASK_RUNNING);
	return iocb->ki_user_data;
}
363
EXPORT_SYMBOL(wait_on_sync_kiocb);
Linus Torvalds's avatar
Linus Torvalds committed
364
365
366
367
368
369
370
371

/* exit_aio: called when the last user of mm goes away.  At this point, 
 * there is no way for any new requests to be submited or any of the 
 * io_* syscalls to be called on the context.  However, there may be 
 * outstanding requests which hold references to the context; as they 
 * go away, they will call put_ioctx and release any pinned memory
 * associated with the request (held via struct page * references).
 */
372
void exit_aio(struct mm_struct *mm)
Linus Torvalds's avatar
Linus Torvalds committed
373
{
374
375
376
377
378
379
	struct kioctx *ctx;

	while (!hlist_empty(&mm->ioctx_list)) {
		ctx = hlist_entry(mm->ioctx_list.first, struct kioctx, list);
		hlist_del_rcu(&ctx->list);

380
		kill_ctx(ctx);
Linus Torvalds's avatar
Linus Torvalds committed
381
382
383
384
385
386

		if (1 != atomic_read(&ctx->users))
			printk(KERN_DEBUG
				"exit_aio:ioctx still alive: %d %d %d\n",
				atomic_read(&ctx->users), ctx->dead,
				ctx->reqs_active);
387
388
389
390
391
392
393
394
395
396
397
		/*
		 * We don't need to bother with munmap() here -
		 * exit_mmap(mm) is coming and it'll unmap everything.
		 * Since aio_free_ring() uses non-zero ->mmap_size
		 * as indicator that it needs to unmap the area,
		 * just set it to 0; aio_free_ring() is the only
		 * place that uses ->mmap_size, so it's safe.
		 * That way we get all munmap done to current->mm -
		 * all other callers have ctx->mm == current->mm.
		 */
		ctx->ring_info.mmap_size = 0;
Linus Torvalds's avatar
Linus Torvalds committed
398
399
400
401
402
403
404
405
406
407
408
409
410
411
		put_ioctx(ctx);
	}
}

/* aio_get_req
 *	Allocate a slot for an aio request.  Increments the users count
 * of the kioctx so that the kioctx stays around until all requests are
 * complete.  Returns NULL if no requests are free.
 *
 * Returns with kiocb->users set to 2.  The io submit code path holds
 * an extra reference while submitting the i/o.
 * This prevents races between the aio code path referencing the
 * req (after submitting it) and aio_complete() freeing the req.
 */
412
static struct kiocb *__aio_get_req(struct kioctx *ctx)
Linus Torvalds's avatar
Linus Torvalds committed
413
414
415
416
417
418
419
{
	struct kiocb *req = NULL;

	req = kmem_cache_alloc(kiocb_cachep, GFP_KERNEL);
	if (unlikely(!req))
		return NULL;

Zach Brown's avatar
Zach Brown committed
420
	req->ki_flags = 0;
Linus Torvalds's avatar
Linus Torvalds committed
421
422
423
424
425
426
427
	req->ki_users = 2;
	req->ki_key = 0;
	req->ki_ctx = ctx;
	req->ki_cancel = NULL;
	req->ki_retry = NULL;
	req->ki_dtor = NULL;
	req->private = NULL;
428
	req->ki_iovec = NULL;
Linus Torvalds's avatar
Linus Torvalds committed
429
	INIT_LIST_HEAD(&req->ki_run_list);
430
	req->ki_eventfd = NULL;
Linus Torvalds's avatar
Linus Torvalds committed
431

Jeff Moyer's avatar
Jeff Moyer committed
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
	return req;
}

/*
 * struct kiocb's are allocated in batches to reduce the number of
 * times the ctx lock is acquired and released.
 */
#define KIOCB_BATCH_SIZE	32L
struct kiocb_batch {
	struct list_head head;
	long count; /* number of requests left to allocate */
};

static void kiocb_batch_init(struct kiocb_batch *batch, long total)
{
	INIT_LIST_HEAD(&batch->head);
	batch->count = total;
}

451
static void kiocb_batch_free(struct kioctx *ctx, struct kiocb_batch *batch)
Jeff Moyer's avatar
Jeff Moyer committed
452
453
454
{
	struct kiocb *req, *n;

455
456
457
458
	if (list_empty(&batch->head))
		return;

	spin_lock_irq(&ctx->ctx_lock);
Jeff Moyer's avatar
Jeff Moyer committed
459
460
	list_for_each_entry_safe(req, n, &batch->head, ki_batch) {
		list_del(&req->ki_batch);
461
		list_del(&req->ki_list);
Jeff Moyer's avatar
Jeff Moyer committed
462
		kmem_cache_free(kiocb_cachep, req);
463
		ctx->reqs_active--;
Jeff Moyer's avatar
Jeff Moyer committed
464
	}
465
466
	if (unlikely(!ctx->reqs_active && ctx->dead))
		wake_up_all(&ctx->wait);
467
	spin_unlock_irq(&ctx->ctx_lock);
Jeff Moyer's avatar
Jeff Moyer committed
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
}

/*
 * Allocate a batch of kiocbs.  This avoids taking and dropping the
 * context lock a lot during setup.
 */
static int kiocb_batch_refill(struct kioctx *ctx, struct kiocb_batch *batch)
{
	unsigned short allocated, to_alloc;
	long avail;
	struct kiocb *req, *n;
	struct aio_ring *ring;

	to_alloc = min(batch->count, KIOCB_BATCH_SIZE);
	for (allocated = 0; allocated < to_alloc; allocated++) {
		req = __aio_get_req(ctx);
		if (!req)
			/* allocation failed, go with what we've got */
			break;
		list_add(&req->ki_batch, &batch->head);
	}

	if (allocated == 0)
		goto out;

Linus Torvalds's avatar
Linus Torvalds committed
493
	spin_lock_irq(&ctx->ctx_lock);
Jeff Moyer's avatar
Jeff Moyer committed
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
	ring = kmap_atomic(ctx->ring_info.ring_pages[0]);

	avail = aio_ring_avail(&ctx->ring_info, ring) - ctx->reqs_active;
	BUG_ON(avail < 0);
	if (avail < allocated) {
		/* Trim back the number of requests. */
		list_for_each_entry_safe(req, n, &batch->head, ki_batch) {
			list_del(&req->ki_batch);
			kmem_cache_free(kiocb_cachep, req);
			if (--allocated <= avail)
				break;
		}
	}

	batch->count -= allocated;
	list_for_each_entry(req, &batch->head, ki_batch) {
Linus Torvalds's avatar
Linus Torvalds committed
510
511
512
513
		list_add(&req->ki_list, &ctx->active_reqs);
		ctx->reqs_active++;
	}

Jeff Moyer's avatar
Jeff Moyer committed
514
515
	kunmap_atomic(ring);
	spin_unlock_irq(&ctx->ctx_lock);
Linus Torvalds's avatar
Linus Torvalds committed
516

Jeff Moyer's avatar
Jeff Moyer committed
517
518
out:
	return allocated;
Linus Torvalds's avatar
Linus Torvalds committed
519
520
}

Jeff Moyer's avatar
Jeff Moyer committed
521
522
static inline struct kiocb *aio_get_req(struct kioctx *ctx,
					struct kiocb_batch *batch)
Linus Torvalds's avatar
Linus Torvalds committed
523
524
{
	struct kiocb *req;
Jeff Moyer's avatar
Jeff Moyer committed
525
526
527
528
529
530

	if (list_empty(&batch->head))
		if (kiocb_batch_refill(ctx, batch) == 0)
			return NULL;
	req = list_first_entry(&batch->head, struct kiocb, ki_batch);
	list_del(&req->ki_batch);
Linus Torvalds's avatar
Linus Torvalds committed
531
532
533
534
535
	return req;
}

static inline void really_put_req(struct kioctx *ctx, struct kiocb *req)
{
536
537
	assert_spin_locked(&ctx->ctx_lock);

538
539
	if (req->ki_eventfd != NULL)
		eventfd_ctx_put(req->ki_eventfd);
Linus Torvalds's avatar
Linus Torvalds committed
540
541
	if (req->ki_dtor)
		req->ki_dtor(req);
542
543
	if (req->ki_iovec != &req->ki_inline_vec)
		kfree(req->ki_iovec);
Linus Torvalds's avatar
Linus Torvalds committed
544
545
546
547
	kmem_cache_free(kiocb_cachep, req);
	ctx->reqs_active--;

	if (unlikely(!ctx->reqs_active && ctx->dead))
548
		wake_up_all(&ctx->wait);
Linus Torvalds's avatar
Linus Torvalds committed
549
550
551
552
553
554
555
}

/* __aio_put_req
 *	Returns true if this put was the last user of the request.
 */
static int __aio_put_req(struct kioctx *ctx, struct kiocb *req)
{
Al Viro's avatar
Al Viro committed
556
557
	dprintk(KERN_DEBUG "aio_put(%p): f_count=%ld\n",
		req, atomic_long_read(&req->ki_filp->f_count));
Linus Torvalds's avatar
Linus Torvalds committed
558

559
560
	assert_spin_locked(&ctx->ctx_lock);

561
	req->ki_users--;
562
	BUG_ON(req->ki_users < 0);
Linus Torvalds's avatar
Linus Torvalds committed
563
564
565
566
567
568
	if (likely(req->ki_users))
		return 0;
	list_del(&req->ki_list);		/* remove from active_reqs */
	req->ki_cancel = NULL;
	req->ki_retry = NULL;

569
570
571
	fput(req->ki_filp);
	req->ki_filp = NULL;
	really_put_req(ctx, req);
Linus Torvalds's avatar
Linus Torvalds committed
572
573
574
575
576
577
578
	return 1;
}

/* aio_put_req
 *	Returns true if this put was the last user of the kiocb,
 *	false if the request is still in use.
 */
579
int aio_put_req(struct kiocb *req)
Linus Torvalds's avatar
Linus Torvalds committed
580
581
582
583
584
585
586
587
{
	struct kioctx *ctx = req->ki_ctx;
	int ret;
	spin_lock_irq(&ctx->ctx_lock);
	ret = __aio_put_req(ctx, req);
	spin_unlock_irq(&ctx->ctx_lock);
	return ret;
}
588
EXPORT_SYMBOL(aio_put_req);
Linus Torvalds's avatar
Linus Torvalds committed
589

590
static struct kioctx *lookup_ioctx(unsigned long ctx_id)
Linus Torvalds's avatar
Linus Torvalds committed
591
{
592
	struct mm_struct *mm = current->mm;
593
	struct kioctx *ctx, *ret = NULL;
Linus Torvalds's avatar
Linus Torvalds committed
594

595
596
	rcu_read_lock();

597
	hlist_for_each_entry_rcu(ctx, &mm->ioctx_list, list) {
Nick Piggin's avatar
Nick Piggin committed
598
599
600
601
602
603
604
		/*
		 * RCU protects us against accessing freed memory but
		 * we have to be careful not to get a reference when the
		 * reference count already dropped to 0 (ctx->dead test
		 * is unreliable because of races).
		 */
		if (ctx->user_id == ctx_id && !ctx->dead && try_get_ioctx(ctx)){
605
			ret = ctx;
Linus Torvalds's avatar
Linus Torvalds committed
606
607
			break;
		}
608
	}
Linus Torvalds's avatar
Linus Torvalds committed
609

610
	rcu_read_unlock();
611
	return ret;
Linus Torvalds's avatar
Linus Torvalds committed
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
}

/*
 * Queue up a kiocb to be retried. Assumes that the kiocb
 * has already been marked as kicked, and places it on
 * the retry run list for the corresponding ioctx, if it
 * isn't already queued. Returns 1 if it actually queued
 * the kiocb (to tell the caller to activate the work
 * queue to process it), or 0, if it found that it was
 * already queued.
 */
static inline int __queue_kicked_iocb(struct kiocb *iocb)
{
	struct kioctx *ctx = iocb->ki_ctx;

627
628
	assert_spin_locked(&ctx->ctx_lock);

Linus Torvalds's avatar
Linus Torvalds committed
629
630
631
632
633
634
635
636
637
638
639
640
641
	if (list_empty(&iocb->ki_run_list)) {
		list_add_tail(&iocb->ki_run_list,
			&ctx->run_list);
		return 1;
	}
	return 0;
}

/* aio_run_iocb
 *	This is the core aio execution routine. It is
 *	invoked both for initial i/o submission and
 *	subsequent retries via the aio_kick_handler.
 *	Expects to be invoked with iocb->ki_ctx->lock
Andreas Mohr's avatar
Andreas Mohr committed
642
 *	already held. The lock is released and reacquired
Linus Torvalds's avatar
Linus Torvalds committed
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
 *	as needed during processing.
 *
 * Calls the iocb retry method (already setup for the
 * iocb on initial submission) for operation specific
 * handling, but takes care of most of common retry
 * execution details for a given iocb. The retry method
 * needs to be non-blocking as far as possible, to avoid
 * holding up other iocbs waiting to be serviced by the
 * retry kernel thread.
 *
 * The trickier parts in this code have to do with
 * ensuring that only one retry instance is in progress
 * for a given iocb at any time. Providing that guarantee
 * simplifies the coding of individual aio operations as
 * it avoids various potential races.
 */
static ssize_t aio_run_iocb(struct kiocb *iocb)
{
	struct kioctx	*ctx = iocb->ki_ctx;
	ssize_t (*retry)(struct kiocb *);
	ssize_t ret;

	if (!(retry = iocb->ki_retry)) {
		printk("aio_run_iocb: iocb->ki_retry = NULL\n");
		return 0;
	}

	/*
	 * We don't want the next retry iteration for this
	 * operation to start until this one has returned and
	 * updated the iocb state. However, wait_queue functions
	 * can trigger a kick_iocb from interrupt context in the
	 * meantime, indicating that data is available for the next
	 * iteration. We want to remember that and enable the
	 * next retry iteration _after_ we are through with
	 * this one.
	 *
	 * So, in order to be able to register a "kick", but
	 * prevent it from being queued now, we clear the kick
	 * flag, but make the kick code *think* that the iocb is
	 * still on the run list until we are actually done.
	 * When we are done with this iteration, we check if
	 * the iocb was kicked in the meantime and if so, queue
	 * it up afresh.
	 */

	kiocbClearKicked(iocb);

	/*
	 * This is so that aio_complete knows it doesn't need to
	 * pull the iocb off the run list (We can't just call
	 * INIT_LIST_HEAD because we don't want a kick_iocb to
	 * queue this on the run list yet)
	 */
	iocb->ki_run_list.next = iocb->ki_run_list.prev = NULL;
	spin_unlock_irq(&ctx->ctx_lock);

	/* Quit retrying if the i/o has been cancelled */
	if (kiocbIsCancelled(iocb)) {
		ret = -EINTR;
		aio_complete(iocb, ret, 0);
		/* must not access the iocb after this */
		goto out;
	}

	/*
	 * Now we are all set to call the retry method in async
710
	 * context.
Linus Torvalds's avatar
Linus Torvalds committed
711
712
713
	 */
	ret = retry(iocb);

714
715
716
717
718
719
720
721
	if (ret != -EIOCBRETRY && ret != -EIOCBQUEUED) {
		/*
		 * There's no easy way to restart the syscall since other AIO's
		 * may be already running. Just fail this IO with EINTR.
		 */
		if (unlikely(ret == -ERESTARTSYS || ret == -ERESTARTNOINTR ||
			     ret == -ERESTARTNOHAND || ret == -ERESTART_RESTARTBLOCK))
			ret = -EINTR;
722
		aio_complete(iocb, ret, 0);
723
	}
Linus Torvalds's avatar
Linus Torvalds committed
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
out:
	spin_lock_irq(&ctx->ctx_lock);

	if (-EIOCBRETRY == ret) {
		/*
		 * OK, now that we are done with this iteration
		 * and know that there is more left to go,
		 * this is where we let go so that a subsequent
		 * "kick" can start the next iteration
		 */

		/* will make __queue_kicked_iocb succeed from here on */
		INIT_LIST_HEAD(&iocb->ki_run_list);
		/* we must queue the next iteration ourselves, if it
		 * has already been kicked */
		if (kiocbIsKicked(iocb)) {
			__queue_kicked_iocb(iocb);
741
742
743
744
745
746
747
748

			/*
			 * __queue_kicked_iocb will always return 1 here, because
			 * iocb->ki_run_list is empty at this point so it should
			 * be safe to unconditionally queue the context into the
			 * work queue.
			 */
			aio_queue_work(ctx);
Linus Torvalds's avatar
Linus Torvalds committed
749
750
751
752
753
754
755
756
757
758
		}
	}
	return ret;
}

/*
 * __aio_run_iocbs:
 * 	Process all pending retries queued on the ioctx
 * 	run list.
 * Assumes it is operating within the aio issuer's mm
759
 * context.
Linus Torvalds's avatar
Linus Torvalds committed
760
761
762
763
 */
static int __aio_run_iocbs(struct kioctx *ctx)
{
	struct kiocb *iocb;
764
	struct list_head run_list;
Linus Torvalds's avatar
Linus Torvalds committed
765

766
767
	assert_spin_locked(&ctx->ctx_lock);

768
	list_replace_init(&ctx->run_list, &run_list);
Linus Torvalds's avatar
Linus Torvalds committed
769
770
771
772
773
774
775
776
777
	while (!list_empty(&run_list)) {
		iocb = list_entry(run_list.next, struct kiocb,
			ki_run_list);
		list_del(&iocb->ki_run_list);
		/*
		 * Hold an extra reference while retrying i/o.
		 */
		iocb->ki_users++;       /* grab extra reference */
		aio_run_iocb(iocb);
778
		__aio_put_req(ctx, iocb);
Linus Torvalds's avatar
Linus Torvalds committed
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
 	}
	if (!list_empty(&ctx->run_list))
		return 1;
	return 0;
}

static void aio_queue_work(struct kioctx * ctx)
{
	unsigned long timeout;
	/*
	 * if someone is waiting, get the work started right
	 * away, otherwise, use a longer delay
	 */
	smp_mb();
	if (waitqueue_active(&ctx->wait))
		timeout = 1;
	else
		timeout = HZ/10;
	queue_delayed_work(aio_wq, &ctx->wq, timeout);
}

/*
801
802
803
804
805
 * aio_run_all_iocbs:
 *	Process all pending retries queued on the ioctx
 *	run list, and keep running them until the list
 *	stays empty.
 * Assumes it is operating within the aio issuer's mm context.
Linus Torvalds's avatar
Linus Torvalds committed
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
 */
static inline void aio_run_all_iocbs(struct kioctx *ctx)
{
	spin_lock_irq(&ctx->ctx_lock);
	while (__aio_run_iocbs(ctx))
		;
	spin_unlock_irq(&ctx->ctx_lock);
}

/*
 * aio_kick_handler:
 * 	Work queue handler triggered to process pending
 * 	retries on an ioctx. Takes on the aio issuer's
 *	mm context before running the iocbs, so that
 *	copy_xxx_user operates on the issuer's address
 *      space.
 * Run on aiod's context.
 */
824
static void aio_kick_handler(struct work_struct *work)
Linus Torvalds's avatar
Linus Torvalds committed
825
{
826
	struct kioctx *ctx = container_of(work, struct kioctx, wq.work);
Linus Torvalds's avatar
Linus Torvalds committed
827
	mm_segment_t oldfs = get_fs();
828
	struct mm_struct *mm;
Linus Torvalds's avatar
Linus Torvalds committed
829
830
831
832
833
834
	int requeue;

	set_fs(USER_DS);
	use_mm(ctx->mm);
	spin_lock_irq(&ctx->ctx_lock);
	requeue =__aio_run_iocbs(ctx);
835
	mm = ctx->mm;
Linus Torvalds's avatar
Linus Torvalds committed
836
	spin_unlock_irq(&ctx->ctx_lock);
837
 	unuse_mm(mm);
Linus Torvalds's avatar
Linus Torvalds committed
838
839
	set_fs(oldfs);
	/*
840
	 * we're in a worker thread already; no point using non-zero delay
Linus Torvalds's avatar
Linus Torvalds committed
841
842
	 */
	if (requeue)
843
		queue_delayed_work(aio_wq, &ctx->wq, 0);
Linus Torvalds's avatar
Linus Torvalds committed
844
845
846
847
848
849
850
851
}


/*
 * Called by kick_iocb to queue the kiocb for retry
 * and if required activate the aio work queue to process
 * it
 */
852
static void try_queue_kicked_iocb(struct kiocb *iocb)
Linus Torvalds's avatar
Linus Torvalds committed
853
854
855
856
857
858
{
 	struct kioctx	*ctx = iocb->ki_ctx;
	unsigned long flags;
	int run = 0;

	spin_lock_irqsave(&ctx->ctx_lock, flags);
859
860
861
862
	/* set this inside the lock so that we can't race with aio_run_iocb()
	 * testing it and putting the iocb on the run list under the lock */
	if (!kiocbTryKick(iocb))
		run = __queue_kicked_iocb(iocb);
Linus Torvalds's avatar
Linus Torvalds committed
863
	spin_unlock_irqrestore(&ctx->ctx_lock, flags);
Ken Chen's avatar
Ken Chen committed
864
	if (run)
Linus Torvalds's avatar
Linus Torvalds committed
865
866
867
868
869
870
		aio_queue_work(ctx);
}

/*
 * kick_iocb:
 *      Called typically from a wait queue callback context
Shaohua Li's avatar
Shaohua Li committed
871
 *      to trigger a retry of the iocb.
Linus Torvalds's avatar
Linus Torvalds committed
872
873
874
 *      The retry is usually executed by aio workqueue
 *      threads (See aio_kick_handler).
 */
875
void kick_iocb(struct kiocb *iocb)
Linus Torvalds's avatar
Linus Torvalds committed
876
877
878
879
880
881
882
883
884
{
	/* sync iocbs are easy: they can only ever be executing from a 
	 * single context. */
	if (is_sync_kiocb(iocb)) {
		kiocbSetKicked(iocb);
	        wake_up_process(iocb->ki_obj.tsk);
		return;
	}

885
	try_queue_kicked_iocb(iocb);
Linus Torvalds's avatar
Linus Torvalds committed
886
887
888
889
890
891
892
893
}
EXPORT_SYMBOL(kick_iocb);

/* aio_complete
 *	Called when the io request on the given iocb is complete.
 *	Returns true if this is the last user of the request.  The 
 *	only other user of the request can be the cancellation code.
 */
894
int aio_complete(struct kiocb *iocb, long res, long res2)
Linus Torvalds's avatar
Linus Torvalds committed
895
896
897
898
899
900
901
902
903
{
	struct kioctx	*ctx = iocb->ki_ctx;
	struct aio_ring_info	*info;
	struct aio_ring	*ring;
	struct io_event	*event;
	unsigned long	flags;
	unsigned long	tail;
	int		ret;

904
905
906
907
908
909
	/*
	 * Special case handling for sync iocbs:
	 *  - events go directly into the iocb for fast handling
	 *  - the sync task with the iocb in its stack holds the single iocb
	 *    ref, no other paths have a way to get another ref
	 *  - the sync task helpfully left a reference to itself in the iocb
Linus Torvalds's avatar
Linus Torvalds committed
910
911
	 */
	if (is_sync_kiocb(iocb)) {
912
		BUG_ON(iocb->ki_users != 1);
Linus Torvalds's avatar
Linus Torvalds committed
913
		iocb->ki_user_data = res;
914
		iocb->ki_users = 0;
Linus Torvalds's avatar
Linus Torvalds committed
915
		wake_up_process(iocb->ki_obj.tsk);
916
		return 1;
Linus Torvalds's avatar
Linus Torvalds committed
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
	}

	info = &ctx->ring_info;

	/* add a completion event to the ring buffer.
	 * must be done holding ctx->ctx_lock to prevent
	 * other code from messing with the tail
	 * pointer since we might be called from irq
	 * context.
	 */
	spin_lock_irqsave(&ctx->ctx_lock, flags);

	if (iocb->ki_run_list.prev && !list_empty(&iocb->ki_run_list))
		list_del_init(&iocb->ki_run_list);

	/*
	 * cancelled requests don't get events, userland was given one
	 * when the event got cancelled.
	 */
	if (kiocbIsCancelled(iocb))
		goto put_rq;

939
	ring = kmap_atomic(info->ring_pages[0]);
Linus Torvalds's avatar
Linus Torvalds committed
940
941

	tail = info->tail;
942
	event = aio_ring_event(info, tail);
943
944
	if (++tail >= info->nr)
		tail = 0;
Linus Torvalds's avatar
Linus Torvalds committed
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962

	event->obj = (u64)(unsigned long)iocb->ki_obj.user;
	event->data = iocb->ki_user_data;
	event->res = res;
	event->res2 = res2;

	dprintk("aio_complete: %p[%lu]: %p: %p %Lx %lx %lx\n",
		ctx, tail, iocb, iocb->ki_obj.user, iocb->ki_user_data,
		res, res2);

	/* after flagging the request as done, we
	 * must never even look at it again
	 */
	smp_wmb();	/* make event visible before updating tail */

	info->tail = tail;
	ring->tail = tail;

963
964
	put_aio_ring_event(event);
	kunmap_atomic(ring);
Linus Torvalds's avatar
Linus Torvalds committed
965
966

	pr_debug("added to ring %p at [%lu]\n", iocb, tail);
Davide Libenzi's avatar
Davide Libenzi committed
967
968
969
970
971
972

	/*
	 * Check if the user asked us to deliver the result through an
	 * eventfd. The eventfd_signal() function is safe to be called
	 * from IRQ context.
	 */
973
	if (iocb->ki_eventfd != NULL)
Davide Libenzi's avatar
Davide Libenzi committed
974
975
		eventfd_signal(iocb->ki_eventfd, 1);

Linus Torvalds's avatar
Linus Torvalds committed
976
977
978
979
put_rq:
	/* everything turned out well, dispose of the aiocb. */
	ret = __aio_put_req(ctx, iocb);

980
981
982
983
984
985
986
987
	/*
	 * We have to order our ring_info tail store above and test
	 * of the wait list below outside the wait lock.  This is
	 * like in wake_up_bit() where clearing a bit has to be
	 * ordered with the unlocked test.
	 */
	smp_mb();

Linus Torvalds's avatar
Linus Torvalds committed
988
989
990
	if (waitqueue_active(&ctx->wait))
		wake_up(&ctx->wait);

991
	spin_unlock_irqrestore(&ctx->ctx_lock, flags);
Linus Torvalds's avatar
Linus Torvalds committed
992
993
	return ret;
}
994
EXPORT_SYMBOL(aio_complete);
Linus Torvalds's avatar
Linus Torvalds committed
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008

/* aio_read_evt
 *	Pull an event off of the ioctx's event ring.  Returns the number of 
 *	events fetched (0 or 1 ;-)
 *	FIXME: make this use cmpxchg.
 *	TODO: make the ringbuffer user mmap()able (requires FIXME).
 */
static int aio_read_evt(struct kioctx *ioctx, struct io_event *ent)
{
	struct aio_ring_info *info = &ioctx->ring_info;
	struct aio_ring *ring;
	unsigned long head;
	int ret = 0;

1009
	ring = kmap_atomic(info->ring_pages[0]);
Linus Torvalds's avatar
Linus Torvalds committed
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
	dprintk("in aio_read_evt h%lu t%lu m%lu\n",
		 (unsigned long)ring->head, (unsigned long)ring->tail,
		 (unsigned long)ring->nr);

	if (ring->head == ring->tail)
		goto out;

	spin_lock(&info->ring_lock);

	head = ring->head % info->nr;
	if (head != ring->tail) {
1021
		struct io_event *evp = aio_ring_event(info, head);
Linus Torvalds's avatar
Linus Torvalds committed
1022
1023
1024
1025
1026
		*ent = *evp;
		head = (head + 1) % info->nr;
		smp_mb(); /* finish reading the event before updatng the head */
		ring->head = head;
		ret = 1;
1027
		put_aio_ring_event(evp);
Linus Torvalds's avatar
Linus Torvalds committed
1028
1029
1030
1031
1032
1033
	}
	spin_unlock(&info->ring_lock);

out:
	dprintk("leaving aio_read_evt: %d  h%lu t%lu\n", ret,
		 (unsigned long)ring->head, (unsigned long)ring->tail);
1034
	kunmap_atomic(ring);
Linus Torvalds's avatar
Linus Torvalds committed
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
	return ret;
}

struct aio_timeout {
	struct timer_list	timer;
	int			timed_out;
	struct task_struct	*p;
};

static void timeout_func(unsigned long data)
{
	struct aio_timeout *to = (struct aio_timeout *)data;

	to->timed_out = 1;
	wake_up_process(to->p);
}

static inline void init_timeout(struct aio_timeout *to)
{
1054
	setup_timer_on_stack(&to->timer, timeout_func, (unsigned long) to);
Linus Torvalds's avatar
Linus Torvalds committed
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
	to->timed_out = 0;
	to->p = current;
}

static inline void set_timeout(long start_jiffies, struct aio_timeout *to,
			       const struct timespec *ts)
{
	to->timer.expires = start_jiffies + timespec_to_jiffies(ts);
	if (time_after(to->timer.expires, jiffies))
		add_timer(&to->timer);
	else
		to->timed_out = 1;
}

static inline void clear_timeout(struct aio_timeout *to)
{
	del_singleshot_timer_sync(&to->timer);
}

static int read_events(struct kioctx *ctx,
			long min_nr, long nr,
			struct io_event __user *event,
			struct timespec __user *timeout)
{
	long			start_jiffies = jiffies;
	struct task_struct	*tsk = current;
	DECLARE_WAITQUEUE(wait, tsk);
	int			ret;
	int			i = 0;
	struct io_event		ent;
	struct aio_timeout	to;
	int			retry = 0;

	/* needed to zero any padding within an entry (there shouldn't be 
	 * any, but C is fun!
	 */
	memset(&ent, 0, sizeof(ent));
retry:
	ret = 0;
	while (likely(i < nr)) {
		ret = aio_read_evt(ctx, &ent);
		if (unlikely(ret <= 0))
			break;

		dprintk("read event: %Lx %Lx %Lx %Lx\n",
			ent.data, ent.obj, ent.res, ent.res2);

		/* Could we split the check in two? */
		ret = -EFAULT;
		if (unlikely(copy_to_user(event, &ent, sizeof(ent)))) {
			dprintk("aio: lost an event due to EFAULT.\n");
			break;
		}
		ret = 0;

		/* Good, event copied to userland, update counts. */
		event ++;
		i ++;
	}

	if (min_nr <= i)
		return i;
	if (ret)
		return ret;

	/* End fast path */

	/* racey check, but it gets redone */
	if (!retry && unlikely(!list_empty(&ctx->run_list))) {
		retry = 1;
		aio_run_all_iocbs(ctx);
		goto retry;
	}

	init_timeout(&to);
	if (timeout) {
		struct timespec	ts;
		ret = -EFAULT;
		if (unlikely(copy_from_user(&ts, timeout, sizeof(ts))))
			goto out;

		set_timeout(start_jiffies, &to, &ts);
	}

	while (likely(i < nr)) {
		add_wait_queue_exclusive(&ctx->wait, &wait);
		do {
			set_task_state(tsk, TASK_INTERRUPTIBLE);
			ret = aio_read_evt(ctx, &ent);
			if (ret)
				break;
			if (min_nr <= i)
				break;
1148
1149
1150
1151
			if (unlikely(ctx->dead)) {
				ret = -EINVAL;
				break;
			}
Linus Torvalds's avatar
Linus Torvalds committed
1152
1153
			if (to.timed_out)	/* Only check after read evt */
				break;
1154
1155
1156
1157
1158
1159
			/* Try to only show up in io wait if there are ops
			 *  in flight */
			if (ctx->reqs_active)
				io_schedule();
			else
				schedule();
Linus Torvalds's avatar
Linus Torvalds committed
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
			if (signal_pending(tsk)) {
				ret = -EINTR;
				break;
			}
			/*ret = aio_read_evt(ctx, &ent);*/
		} while (1) ;

		set_task_state(tsk, TASK_RUNNING);
		remove_wait_queue(&ctx->wait, &wait);

		if (unlikely(ret <= 0))
			break;

		ret = -EFAULT;
		if (unlikely(copy_to_user(event, &ent, sizeof(ent)))) {
			dprintk("aio: lost an event due to EFAULT.\n");
			break;
		}

		/* Good, event copied to userland, update counts. */
		event ++;
		i ++;
	}

	if (timeout)
		clear_timeout(&to);
out:
1187
	destroy_timer_on_stack(&to.timer);
Linus Torvalds's avatar
Linus Torvalds committed
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
	return i ? i : ret;
}

/* Take an ioctx and remove it from the list of ioctx's.  Protects 
 * against races with itself via ->dead.
 */
static void io_destroy(struct kioctx *ioctx)
{
	struct mm_struct *mm = current->mm;
	int was_dead;

	/* delete the entry from the list is someone else hasn't already */
1200
	spin_lock(&mm->ioctx_lock);
Linus Torvalds's avatar
Linus Torvalds committed
1201
1202
	was_dead = ioctx->dead;
	ioctx->dead = 1;
1203
1204
	hlist_del_rcu(&ioctx->list);
	spin_unlock(&mm->ioctx_lock);
Linus Torvalds's avatar
Linus Torvalds committed
1205
1206
1207
1208
1209

	dprintk("aio_release(%p)\n", ioctx);
	if (likely(!was_dead))
		put_ioctx(ioctx);	/* twice for the list */

1210
	kill_ctx(ioctx);
1211
1212
1213
1214
1215
1216

	/*
	 * Wake up any waiters.  The setting of ctx->dead must be seen
	 * by other CPUs at this point.  Right now, we rely on the
	 * locking done by the above calls to ensure this consistency.
	 */
1217
	wake_up_all(&ioctx->wait);
Linus Torvalds's avatar
Linus Torvalds committed
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
}

/* sys_io_setup:
 *	Create an aio_context capable of receiving at least nr_events.
 *	ctxp must not point to an aio_context that already exists, and
 *	must be initialized to 0 prior to the call.  On successful
 *	creation of the aio_context, *ctxp is filled in with the resulting 
 *	handle.  May fail with -EINVAL if *ctxp is not initialized,
 *	if the specified nr_events exceeds internal limits.  May fail 
 *	with -EAGAIN if the specified nr_events exceeds the user's limit 
 *	of available events.  May fail with -ENOMEM if insufficient kernel
 *	resources are available.  May fail with -EFAULT if an invalid
 *	pointer is passed for ctxp.  Will fail with -ENOSYS if not
 *	implemented.
 */
1233
SYSCALL_DEFINE2(io_setup, unsigned, nr_events, aio_context_t __user *, ctxp)
Linus Torvalds's avatar
Linus Torvalds committed
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
{
	struct kioctx *ioctx = NULL;
	unsigned long ctx;
	long ret;

	ret = get_user(ctx, ctxp);
	if (unlikely(ret))
		goto out;

	ret = -EINVAL;
1244
1245
1246
	if (unlikely(ctx || nr_events == 0)) {
		pr_debug("EINVAL: io_setup: ctx %lu nr_events %u\n",
		         ctx, nr_events);
Linus Torvalds's avatar
Linus Torvalds committed
1247
1248
1249
1250
1251
1252
1253
		goto out;
	}

	ioctx = ioctx_alloc(nr_events);
	ret = PTR_ERR(ioctx);
	if (!IS_ERR(ioctx)) {
		ret = put_user(ioctx->user_id, ctxp);
1254
1255
1256
		if (ret)
			io_destroy(ioctx);
		put_ioctx(ioctx);
Linus Torvalds's avatar
Linus Torvalds committed
1257
1258
1259
1260
1261
1262
1263
1264
1265
	}

out:
	return ret;
}

/* sys_io_destroy:
 *	Destroy the aio_context specified.  May cancel any outstanding 
 *	AIOs and block on completion.  Will fail with -ENOSYS if not
1266
 *	implemented.  May fail with -EINVAL if the context pointed to
Linus Torvalds's avatar
Linus Torvalds committed
1267
1268
 *	is invalid.
 */
1269
SYSCALL_DEFINE1(io_destroy, aio_context_t, ctx)
Linus Torvalds's avatar
Linus Torvalds committed
1270
1271
1272
1273
{
	struct kioctx *ioctx = lookup_ioctx(ctx);
	if (likely(NULL != ioctx)) {
		io_destroy(ioctx);
1274
		put_ioctx(ioctx);
Linus Torvalds's avatar
Linus Torvalds committed
1275
1276
1277
1278
1279
1280
		return 0;
	}
	pr_debug("EINVAL: io_destroy: invalid context id\n");
	return -EINVAL;
}

1281
static void aio_advance_iovec(struct kiocb *iocb, ssize_t ret)
Linus Torvalds's avatar
Linus Torvalds committed
1282
{
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
	struct iovec *iov = &iocb->ki_iovec[iocb->ki_cur_seg];

	BUG_ON(ret <= 0);

	while (iocb->ki_cur_seg < iocb->ki_nr_segs && ret > 0) {
		ssize_t this = min((ssize_t)iov->iov_len, ret);
		iov->iov_base += this;
		iov->iov_len -= this;
		iocb->ki_left -= this;
		ret -= this;
		if (iov->iov_len == 0) {
			iocb->ki_cur_seg++;
			iov++;
1296
		}
1297
	}
Linus Torvalds's avatar
Linus Torvalds committed
1298

1299
1300
1301
	/* the caller should not have done more io than what fit in
	 * the remaining iovecs */
	BUG_ON(ret > 0 && iocb->ki_left == 0);
Linus Torvalds's avatar
Linus Torvalds committed
1302
1303
}

1304
static ssize_t aio_rw_vect_retry(struct kiocb *iocb)
Linus Torvalds's avatar
Linus Torvalds committed
1305
1306
{
	struct file *file = iocb->ki_filp;
1307
1308
1309
1310
	struct address_space *mapping = file->f_mapping;
	struct inode *inode = mapping->host;
	ssize_t (*rw_op)(struct kiocb *, const struct iovec *,
			 unsigned long, loff_t);
Linus Torvalds's avatar
Linus Torvalds committed
1311
	ssize_t ret = 0;
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
	unsigned short opcode;

	if ((iocb->ki_opcode == IOCB_CMD_PREADV) ||
		(iocb->ki_opcode == IOCB_CMD_PREAD)) {
		rw_op = file->f_op->aio_read;
		opcode = IOCB_CMD_PREADV;
	} else {
		rw_op = file->f_op->aio_write;
		opcode = IOCB_CMD_PWRITEV;
	}
Linus Torvalds's avatar
Linus Torvalds committed
1322

1323
1324
1325
1326
	/* This matches the pread()/pwrite() logic */
	if (iocb->ki_pos < 0)
		return -EINVAL;

1327
1328
	if (opcode == IOCB_CMD_PWRITEV)
		file_start_write(file);
1329
	do {
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
		ret = rw_op(iocb, &iocb->ki_iovec[iocb->ki_cur_seg],
			    iocb->ki_nr_segs - iocb->ki_cur_seg,
			    iocb->ki_pos);
		if (ret > 0)
			aio_advance_iovec(iocb, ret);

	/* retry all partial writes.  retry partial reads as long as its a
	 * regular file. */
	} while (ret > 0 && iocb->ki_left > 0 &&
		 (opcode == IOCB_CMD_PWRITEV ||
		  (!S_ISFIFO(inode->i_mode) && !S_ISSOCK(inode->i_mode))));
1341
1342
	if (opcode == IOCB_CMD_PWRITEV)
		file_end_write(file);
Linus Torvalds's avatar
Linus Torvalds committed
1343

1344
1345
	/* This means we must have transferred all that we could */
	/* No need to retry anymore */
Linus Torvalds's avatar
Linus Torvalds committed
1346
1347
1348
	if ((ret == 0) || (iocb->ki_left == 0))
		ret = iocb->ki_nbytes - iocb->ki_left;

1349
1350
1351
1352
1353
1354
1355
	/* If we managed to write some out we return that, rather than
	 * the eventual error. */
	if (opcode == IOCB_CMD_PWRITEV
	    && ret < 0 && ret != -EIOCBQUEUED && ret != -EIOCBRETRY
	    && iocb->ki_nbytes - iocb->ki_left)
		ret = iocb->ki_nbytes - iocb->ki_left;

Linus Torvalds's avatar
Linus Torvalds committed
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
	return ret;
}

static ssize_t aio_fdsync(struct kiocb *iocb)
{
	struct file *file = iocb->ki_filp;
	ssize_t ret = -EINVAL;

	if (file->f_op->aio_fsync)
		ret = file->f_op->aio_fsync(iocb, 1);
	return ret;
}

static ssize_t aio_fsync(struct kiocb *iocb)
{
	struct file *file = iocb->ki_filp;
	ssize_t ret = -EINVAL;

	if (file->f_op->aio_fsync)
		ret = file->f_op->aio_fsync(iocb, 0);
	return ret;
}

1379
static ssize_t aio_setup_vectored_rw(int type, struct kiocb *kiocb, bool compat)
1380
1381
1382
{
	ssize_t ret;

1383
1384
1385
1386
1387
#ifdef CONFIG_COMPAT
	if (compat)
		ret = compat_rw_copy_check_uvector(type,
				(struct compat_iovec __user *)kiocb->ki_buf,
				kiocb->ki_nbytes, 1, &kiocb->ki_inline_vec,
1388
				&kiocb->ki_iovec);
1389
1390
1391
1392
1393
	else
#endif
		ret = rw_copy_check_uvector(type,
				(struct iovec __user *)kiocb->ki_buf,
				kiocb->ki_nbytes, 1, &kiocb->ki_inline_vec,
1394
				&kiocb->ki_iovec);
1395
1396
1397
	if (ret < 0)
		goto out;

1398
1399
1400
1401
	ret = rw_verify_area(type, kiocb->ki_filp, &kiocb->ki_pos, ret);
	if (ret < 0)
		goto out;

1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
	kiocb->ki_nr_segs = kiocb->ki_nbytes;
	kiocb->ki_cur_seg = 0;
	/* ki_nbytes/left now reflect bytes instead of segs */
	kiocb->ki_nbytes = ret;
	kiocb->ki_left = ret;

	ret = 0;
out:
	return ret;
}

1413
static ssize_t aio_setup_single_vector(int type, struct file * file, struct kiocb *kiocb)
1414
{
1415
1416
1417
1418
1419
1420
	int bytes;

	bytes = rw_verify_area(type, file, &kiocb->ki_pos, kiocb->ki_left);
	if (bytes < 0)
		return bytes;

1421
1422
	kiocb->ki_iovec = &kiocb->ki_inline_vec;
	kiocb->ki_iovec->iov_base = kiocb->ki_buf;
1423
	kiocb->ki_iovec->iov_len = bytes;
1424
1425
1426
1427
1428
	kiocb->ki_nr_segs = 1;
	kiocb->ki_cur_seg = 0;
	return 0;
}

Linus Torvalds's avatar
Linus Torvalds committed
1429
1430
1431
1432
1433
/*
 * aio_setup_iocb:
 *	Performs the initial checks and aio retry method
 *	setup for the kiocb at the time of io submission.
 */
1434
static ssize_t aio_setup_iocb(struct kiocb *kiocb, bool compat)
Linus Torvalds's avatar
Linus Torvalds committed
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
{
	struct file *file = kiocb->ki_filp;
	ssize_t ret = 0;

	switch (kiocb->ki_opcode) {
	case IOCB_CMD_PREAD:
		ret = -EBADF;
		if (unlikely(!(file->f_mode & FMODE_READ)))
			break;
		ret = -EFAULT;
		if (unlikely(!access_ok(VERIFY_WRITE, kiocb->ki_buf,
			kiocb->ki_left)))
			break;
1448
		ret = aio_setup_single_vector(READ, file, kiocb);
1449
1450
		if (ret)
			break;
Linus Torvalds's avatar
Linus Torvalds committed
1451
1452
		ret = -EINVAL;
		if (file->f_op->aio_read)
1453
			kiocb->ki_retry = aio_rw_vect_retry;
Linus Torvalds's avatar
Linus Torvalds committed
1454
1455
1456
1457
1458
1459
1460
1461
1462
		break;
	case IOCB_CMD_PWRITE:
		ret = -EBADF;
		if (unlikely(!(file->f_mode & FMODE_WRITE)))
			break;
		ret = -EFAULT;
		if (unlikely(!access_ok(VERIFY_READ, kiocb->ki_buf,
			kiocb->ki_left)))
			break;
1463
		ret = aio_setup_single_vector(WRITE, file, kiocb);
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
		if (ret)
			break;
		ret = -EINVAL;
		if (file->f_op->aio_write)
			kiocb->ki_retry = aio_rw_vect_retry;
		break;
	case IOCB_CMD_PREADV:
		ret = -EBADF;
		if (unlikely(!(file->f_mode & FMODE_READ)))
			break;
1474
		ret = aio_setup_vectored_rw(READ, kiocb, compat);
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
		if (ret)
			break;
		ret = -EINVAL;
		if (file->f_op->aio_read)
			kiocb->ki_retry = aio_rw_vect_retry;
		break;
	case IOCB_CMD_PWRITEV:
		ret = -EBADF;
		if (unlikely(!(file->f_mode & FMODE_WRITE)))
			break;
1485
		ret = aio_setup_vectored_rw(WRITE, kiocb, compat);
1486
1487
		if (ret)
			break;
Linus Torvalds's avatar
Linus Torvalds committed
1488
1489
		ret = -EINVAL;
		if (file->f_op->aio_write)
1490
			kiocb->ki_retry = aio_rw_vect_retry;
Linus Torvalds's avatar
Linus Torvalds committed
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
		break;
	case IOCB_CMD_FDSYNC:
		ret = -EINVAL;
		if (file->f_op->aio_fsync)
			kiocb->ki_retry = aio_fdsync;
		break;
	case IOCB_CMD_FSYNC:
		ret = -EINVAL;
		if (file->f_op->aio_fsync)
			kiocb->ki_retry = aio_fsync;
		break;
	default:
		dprintk("EINVAL: io_submit: no operation provided\n");
		ret = -EINVAL;
	}

	if (!kiocb->ki_retry)
		return ret;

	return 0;
}

1513
static int io_submit_one(struct kioctx *ctx, struct iocb __user *user_iocb,
Jeff Moyer's avatar
Jeff Moyer committed
1514
1515
			 struct iocb *iocb, struct kiocb_batch *batch,
			 bool compat)
Linus Torvalds's avatar
Linus Torvalds committed
1516
1517
1518
1519
1520
1521
{
	struct kiocb *req;
	struct file *file;
	ssize_t ret;

	/* enforce forwards compatibility on users */
1522
	if (unlikely(iocb->aio_reserved1 || iocb->aio_reserved2)) {
Linus Torvalds's avatar
Linus Torvalds committed
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
		pr_debug("EINVAL: io_submit: reserve field set\n");
		return -EINVAL;
	}

	/* prevent overflows */
	if (unlikely(
	    (iocb->aio_buf != (unsigned long)iocb->aio_buf) ||
	    (iocb->aio_nbytes != (size_t)iocb->aio_nbytes) ||
	    ((ssize_t)iocb->aio_nbytes < 0)
	   )) {
		pr_debug("EINVAL: io_submit: overflow check\n");
		return -EINVAL;
	}

	file = fget(iocb->aio_fildes);
	if (unlikely(!file))
		return -EBADF;

Jeff Moyer's avatar
Jeff Moyer committed
1541
	req = aio_get_req(ctx, batch);  /* returns with 2 references to req */
Linus Torvalds's avatar
Linus Torvalds committed
1542
1543
1544
1545
	if (unlikely(!req)) {
		fput(file);
		return -EAGAIN;
	}
1546
	req->ki_filp = file;
1547
1548
1549
1550
1551
1552
1553
	if (iocb->aio_flags & IOCB_FLAG_RESFD) {
		/*
		 * If the IOCB_FLAG_RESFD flag of aio_flags is set, get an
		 * instance of the file* now. The file descriptor must be
		 * an eventfd() fd, and will be signaled for each completed
		 * event using the eventfd_signal() function.
		 */
1554
		req->ki_eventfd = eventfd_ctx_fdget((int) iocb->aio_resfd);
1555
		if (IS_ERR(req->ki_eventfd)) {
1556
			ret = PTR_ERR(req->ki_eventfd);
1557
			req->ki_eventfd = NULL;
1558
1559
1560
			goto out_put_req;
		}
	}
Linus Torvalds's avatar
Linus Torvalds committed
1561