file.c 48 KB
Newer Older
1
2
/*
  FUSE: Filesystem in Userspace
Miklos Szeredi's avatar
Miklos Szeredi committed
3
  Copyright (C) 2001-2008  Miklos Szeredi <miklos@szeredi.hu>
4
5
6
7
8
9
10
11
12
13

  This program can be distributed under the terms of the GNU GPL.
  See the file COPYING.
*/

#include "fuse_i.h"

#include <linux/pagemap.h>
#include <linux/slab.h>
#include <linux/kernel.h>
Alexey Dobriyan's avatar
Alexey Dobriyan committed
14
#include <linux/sched.h>
15
#include <linux/module.h>
16

17
static const struct file_operations fuse_direct_io_file_operations;
18

19
20
static int fuse_send_open(struct fuse_conn *fc, u64 nodeid, struct file *file,
			  int opcode, struct fuse_open_out *outargp)
21
22
{
	struct fuse_open_in inarg;
23
24
25
	struct fuse_req *req;
	int err;

26
27
28
	req = fuse_get_req(fc);
	if (IS_ERR(req))
		return PTR_ERR(req);
29
30

	memset(&inarg, 0, sizeof(inarg));
31
32
33
	inarg.flags = file->f_flags & ~(O_CREAT | O_EXCL | O_NOCTTY);
	if (!fc->atomic_o_trunc)
		inarg.flags &= ~O_TRUNC;
34
35
	req->in.h.opcode = opcode;
	req->in.h.nodeid = nodeid;
36
37
38
39
40
41
	req->in.numargs = 1;
	req->in.args[0].size = sizeof(inarg);
	req->in.args[0].value = &inarg;
	req->out.numargs = 1;
	req->out.args[0].size = sizeof(*outargp);
	req->out.args[0].value = outargp;
42
	fuse_request_send(fc, req);
43
44
45
46
47
48
	err = req->out.h.error;
	fuse_put_request(fc, req);

	return err;
}

Tejun Heo's avatar
Tejun Heo committed
49
struct fuse_file *fuse_file_alloc(struct fuse_conn *fc)
50
51
{
	struct fuse_file *ff;
Tejun Heo's avatar
Tejun Heo committed
52

53
	ff = kmalloc(sizeof(struct fuse_file), GFP_KERNEL);
Tejun Heo's avatar
Tejun Heo committed
54
55
56
	if (unlikely(!ff))
		return NULL;

57
	ff->fc = fc;
Tejun Heo's avatar
Tejun Heo committed
58
59
60
61
	ff->reserved_req = fuse_request_alloc();
	if (unlikely(!ff->reserved_req)) {
		kfree(ff);
		return NULL;
62
	}
Tejun Heo's avatar
Tejun Heo committed
63
64
65
66
67
68
69
70
71
72

	INIT_LIST_HEAD(&ff->write_entry);
	atomic_set(&ff->count, 0);
	RB_CLEAR_NODE(&ff->polled_node);
	init_waitqueue_head(&ff->poll_wait);

	spin_lock(&fc->lock);
	ff->kh = ++fc->khctr;
	spin_unlock(&fc->lock);

73
74
75
76
77
	return ff;
}

void fuse_file_free(struct fuse_file *ff)
{
78
	fuse_request_free(ff->reserved_req);
79
80
81
	kfree(ff);
}

82
struct fuse_file *fuse_file_get(struct fuse_file *ff)
83
84
85
86
87
{
	atomic_inc(&ff->count);
	return ff;
}

Miklos Szeredi's avatar
Miklos Szeredi committed
88
89
static void fuse_release_end(struct fuse_conn *fc, struct fuse_req *req)
{
90
	path_put(&req->misc.release.path);
Miklos Szeredi's avatar
Miklos Szeredi committed
91
92
}

93
94
95
96
static void fuse_file_put(struct fuse_file *ff)
{
	if (atomic_dec_and_test(&ff->count)) {
		struct fuse_req *req = ff->reserved_req;
97

Miklos Szeredi's avatar
Miklos Szeredi committed
98
		req->end = fuse_release_end;
99
		fuse_request_send_background(ff->fc, req);
100
101
102
103
		kfree(ff);
	}
}

104
105
int fuse_do_open(struct fuse_conn *fc, u64 nodeid, struct file *file,
		 bool isdir)
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
{
	struct fuse_open_out outarg;
	struct fuse_file *ff;
	int err;
	int opcode = isdir ? FUSE_OPENDIR : FUSE_OPEN;

	ff = fuse_file_alloc(fc);
	if (!ff)
		return -ENOMEM;

	err = fuse_send_open(fc, nodeid, file, opcode, &outarg);
	if (err) {
		fuse_file_free(ff);
		return err;
	}

	if (isdir)
		outarg.open_flags &= ~FOPEN_DIRECT_IO;

	ff->fh = outarg.fh;
	ff->nodeid = nodeid;
	ff->open_flags = outarg.open_flags;
	file->private_data = fuse_file_get(ff);

	return 0;
}
132
EXPORT_SYMBOL_GPL(fuse_do_open);
133

134
void fuse_finish_open(struct inode *inode, struct file *file)
135
{
136
137
138
	struct fuse_file *ff = file->private_data;

	if (ff->open_flags & FOPEN_DIRECT_IO)
139
		file->f_op = &fuse_direct_io_file_operations;
140
	if (!(ff->open_flags & FOPEN_KEEP_CACHE))
Miklos Szeredi's avatar
Miklos Szeredi committed
141
		invalidate_inode_pages2(inode->i_mapping);
142
	if (ff->open_flags & FOPEN_NONSEEKABLE)
Tejun Heo's avatar
Tejun Heo committed
143
		nonseekable_open(inode, file);
144
145
}

146
int fuse_open_common(struct inode *inode, struct file *file, bool isdir)
147
{
Tejun Heo's avatar
Tejun Heo committed
148
	struct fuse_conn *fc = get_fuse_conn(inode);
149
150
	int err;

Miklos Szeredi's avatar
Miklos Szeredi committed
151
152
153
154
	/* VFS checks this, but only _after_ ->open() */
	if (file->f_flags & O_DIRECT)
		return -EINVAL;

155
156
157
158
	err = generic_file_open(inode, file);
	if (err)
		return err;

159
	err = fuse_do_open(fc, get_node_id(inode), file, isdir);
160
	if (err)
161
		return err;
162

163
164
165
	fuse_finish_open(inode, file);

	return 0;
166
167
}

168
static void fuse_prepare_release(struct fuse_file *ff, int flags, int opcode)
169
{
170
	struct fuse_conn *fc = ff->fc;
171
	struct fuse_req *req = ff->reserved_req;
172
	struct fuse_release_in *inarg = &req->misc.release.in;
173

174
175
176
177
178
179
180
181
	spin_lock(&fc->lock);
	list_del(&ff->write_entry);
	if (!RB_EMPTY_NODE(&ff->polled_node))
		rb_erase(&ff->polled_node, &fc->polled_files);
	spin_unlock(&fc->lock);

	wake_up_interruptible_sync(&ff->poll_wait);

182
	inarg->fh = ff->fh;
183
	inarg->flags = flags;
184
	req->in.h.opcode = opcode;
185
	req->in.h.nodeid = ff->nodeid;
186
187
188
	req->in.numargs = 1;
	req->in.args[0].size = sizeof(struct fuse_release_in);
	req->in.args[0].value = inarg;
189
190
}

191
void fuse_release_common(struct file *file, int opcode)
192
{
Tejun Heo's avatar
Tejun Heo committed
193
194
	struct fuse_file *ff;
	struct fuse_req *req;
195

Tejun Heo's avatar
Tejun Heo committed
196
197
	ff = file->private_data;
	if (unlikely(!ff))
198
		return;
Tejun Heo's avatar
Tejun Heo committed
199
200

	req = ff->reserved_req;
201
	fuse_prepare_release(ff, file->f_flags, opcode);
Tejun Heo's avatar
Tejun Heo committed
202
203

	/* Hold vfsmount and dentry until release is finished */
204
205
	path_get(&file->f_path);
	req->misc.release.path = file->f_path;
Tejun Heo's avatar
Tejun Heo committed
206
207
208
209
210
211
212

	/*
	 * Normally this will send the RELEASE request, however if
	 * some asynchronous READ or WRITE requests are outstanding,
	 * the sending will be delayed.
	 */
	fuse_file_put(ff);
213
214
}

215
216
static int fuse_open(struct inode *inode, struct file *file)
{
217
	return fuse_open_common(inode, file, false);
218
219
220
221
}

static int fuse_release(struct inode *inode, struct file *file)
{
222
223
224
225
226
227
228
229
230
231
232
233
234
235
	fuse_release_common(file, FUSE_RELEASE);

	/* return value is ignored by VFS */
	return 0;
}

void fuse_sync_release(struct fuse_file *ff, int flags)
{
	WARN_ON(atomic_read(&ff->count) > 1);
	fuse_prepare_release(ff, flags, FUSE_RELEASE);
	ff->reserved_req->force = 1;
	fuse_request_send(ff->fc, ff->reserved_req);
	fuse_put_request(ff->fc, ff->reserved_req);
	kfree(ff);
236
}
237
EXPORT_SYMBOL_GPL(fuse_sync_release);
238

239
/*
240
241
 * Scramble the ID space with XTEA, so that the value of the files_struct
 * pointer is not exposed to userspace.
242
 */
243
u64 fuse_lock_owner_id(struct fuse_conn *fc, fl_owner_t id)
244
{
245
246
247
248
249
250
251
252
253
254
255
256
257
258
	u32 *k = fc->scramble_key;
	u64 v = (unsigned long) id;
	u32 v0 = v;
	u32 v1 = v >> 32;
	u32 sum = 0;
	int i;

	for (i = 0; i < 32; i++) {
		v0 += ((v1 << 4 ^ v1 >> 5) + v1) ^ (sum + k[sum & 3]);
		sum += 0x9E3779B9;
		v1 += ((v0 << 4 ^ v0 >> 5) + v0) ^ (sum + k[sum>>11 & 3]);
	}

	return (u64) v0 + ((u64) v1 << 32);
259
260
}

Miklos Szeredi's avatar
Miklos Szeredi committed
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
/*
 * Check if page is under writeback
 *
 * This is currently done by walking the list of writepage requests
 * for the inode, which can be pretty inefficient.
 */
static bool fuse_page_is_writeback(struct inode *inode, pgoff_t index)
{
	struct fuse_conn *fc = get_fuse_conn(inode);
	struct fuse_inode *fi = get_fuse_inode(inode);
	struct fuse_req *req;
	bool found = false;

	spin_lock(&fc->lock);
	list_for_each_entry(req, &fi->writepages, writepages_entry) {
		pgoff_t curr_index;

		BUG_ON(req->inode != inode);
		curr_index = req->misc.write.in.offset >> PAGE_CACHE_SHIFT;
		if (curr_index == index) {
			found = true;
			break;
		}
	}
	spin_unlock(&fc->lock);

	return found;
}

/*
 * Wait for page writeback to be completed.
 *
 * Since fuse doesn't rely on the VM writeback tracking, this has to
 * use some other means.
 */
static int fuse_wait_on_page_writeback(struct inode *inode, pgoff_t index)
{
	struct fuse_inode *fi = get_fuse_inode(inode);

	wait_event(fi->page_waitq, !fuse_page_is_writeback(inode, index));
	return 0;
}

304
static int fuse_flush(struct file *file, fl_owner_t id)
305
{
306
	struct inode *inode = file->f_path.dentry->d_inode;
307
308
309
310
311
312
	struct fuse_conn *fc = get_fuse_conn(inode);
	struct fuse_file *ff = file->private_data;
	struct fuse_req *req;
	struct fuse_flush_in inarg;
	int err;

313
314
315
	if (is_bad_inode(inode))
		return -EIO;

316
317
318
	if (fc->no_flush)
		return 0;

319
	req = fuse_get_req_nofail(fc, file);
320
321
	memset(&inarg, 0, sizeof(inarg));
	inarg.fh = ff->fh;
322
	inarg.lock_owner = fuse_lock_owner_id(fc, id);
323
324
325
326
327
	req->in.h.opcode = FUSE_FLUSH;
	req->in.h.nodeid = get_node_id(inode);
	req->in.numargs = 1;
	req->in.args[0].size = sizeof(inarg);
	req->in.args[0].value = &inarg;
328
	req->force = 1;
329
	fuse_request_send(fc, req);
330
331
332
333
334
335
336
337
338
	err = req->out.h.error;
	fuse_put_request(fc, req);
	if (err == -ENOSYS) {
		fc->no_flush = 1;
		err = 0;
	}
	return err;
}

Miklos Szeredi's avatar
Miklos Szeredi committed
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
/*
 * Wait for all pending writepages on the inode to finish.
 *
 * This is currently done by blocking further writes with FUSE_NOWRITE
 * and waiting for all sent writes to complete.
 *
 * This must be called under i_mutex, otherwise the FUSE_NOWRITE usage
 * could conflict with truncation.
 */
static void fuse_sync_writes(struct inode *inode)
{
	fuse_set_nowrite(inode);
	fuse_release_nowrite(inode);
}

354
355
int fuse_fsync_common(struct file *file, struct dentry *de, int datasync,
		      int isdir)
356
357
358
359
360
361
362
363
{
	struct inode *inode = de->d_inode;
	struct fuse_conn *fc = get_fuse_conn(inode);
	struct fuse_file *ff = file->private_data;
	struct fuse_req *req;
	struct fuse_fsync_in inarg;
	int err;

364
365
366
	if (is_bad_inode(inode))
		return -EIO;

367
	if ((!isdir && fc->no_fsync) || (isdir && fc->no_fsyncdir))
368
369
		return 0;

Miklos Szeredi's avatar
Miklos Szeredi committed
370
371
372
373
374
375
376
377
378
379
380
	/*
	 * Start writeback against all dirty pages of the inode, then
	 * wait for all outstanding writes, before sending the FSYNC
	 * request.
	 */
	err = write_inode_now(inode, 0);
	if (err)
		return err;

	fuse_sync_writes(inode);

381
382
383
	req = fuse_get_req(fc);
	if (IS_ERR(req))
		return PTR_ERR(req);
384
385
386
387

	memset(&inarg, 0, sizeof(inarg));
	inarg.fh = ff->fh;
	inarg.fsync_flags = datasync ? 1 : 0;
388
	req->in.h.opcode = isdir ? FUSE_FSYNCDIR : FUSE_FSYNC;
389
390
391
392
	req->in.h.nodeid = get_node_id(inode);
	req->in.numargs = 1;
	req->in.args[0].size = sizeof(inarg);
	req->in.args[0].value = &inarg;
393
	fuse_request_send(fc, req);
394
395
396
	err = req->out.h.error;
	fuse_put_request(fc, req);
	if (err == -ENOSYS) {
397
398
399
400
		if (isdir)
			fc->no_fsyncdir = 1;
		else
			fc->no_fsync = 1;
401
402
403
404
405
		err = 0;
	}
	return err;
}

406
407
408
409
410
static int fuse_fsync(struct file *file, struct dentry *de, int datasync)
{
	return fuse_fsync_common(file, de, datasync, 0);
}

411
412
void fuse_read_fill(struct fuse_req *req, struct file *file, loff_t pos,
		    size_t count, int opcode)
413
{
414
	struct fuse_read_in *inarg = &req->misc.read.in;
415
	struct fuse_file *ff = file->private_data;
416

417
418
419
	inarg->fh = ff->fh;
	inarg->offset = pos;
	inarg->size = count;
420
	inarg->flags = file->f_flags;
421
	req->in.h.opcode = opcode;
422
	req->in.h.nodeid = ff->nodeid;
423
424
	req->in.numargs = 1;
	req->in.args[0].size = sizeof(struct fuse_read_in);
425
	req->in.args[0].value = inarg;
426
427
428
429
430
	req->out.argvar = 1;
	req->out.numargs = 1;
	req->out.args[0].size = count;
}

431
static size_t fuse_send_read(struct fuse_req *req, struct file *file,
432
			     loff_t pos, size_t count, fl_owner_t owner)
433
{
434
435
	struct fuse_file *ff = file->private_data;
	struct fuse_conn *fc = ff->fc;
436

437
	fuse_read_fill(req, file, pos, count, FUSE_READ);
438
	if (owner != NULL) {
439
		struct fuse_read_in *inarg = &req->misc.read.in;
440
441
442
443

		inarg->read_flags |= FUSE_READ_LOCKOWNER;
		inarg->lock_owner = fuse_lock_owner_id(fc, owner);
	}
444
	fuse_request_send(fc, req);
445
	return req->out.args[0].size;
446
447
}

448
449
450
451
452
453
454
455
456
457
458
459
460
461
static void fuse_read_update_size(struct inode *inode, loff_t size,
				  u64 attr_ver)
{
	struct fuse_conn *fc = get_fuse_conn(inode);
	struct fuse_inode *fi = get_fuse_inode(inode);

	spin_lock(&fc->lock);
	if (attr_ver == fi->attr_version && size < inode->i_size) {
		fi->attr_version = ++fc->attr_version;
		i_size_write(inode, size);
	}
	spin_unlock(&fc->lock);
}

462
463
464
465
static int fuse_readpage(struct file *file, struct page *page)
{
	struct inode *inode = page->mapping->host;
	struct fuse_conn *fc = get_fuse_conn(inode);
466
	struct fuse_req *req;
467
468
469
470
	size_t num_read;
	loff_t pos = page_offset(page);
	size_t count = PAGE_CACHE_SIZE;
	u64 attr_ver;
471
472
473
474
475
476
	int err;

	err = -EIO;
	if (is_bad_inode(inode))
		goto out;

Miklos Szeredi's avatar
Miklos Szeredi committed
477
478
479
480
481
482
483
	/*
	 * Page writeback can extend beyond the liftime of the
	 * page-cache page, so make sure we read a properly synced
	 * page.
	 */
	fuse_wait_on_page_writeback(inode, page->index);

484
485
486
	req = fuse_get_req(fc);
	err = PTR_ERR(req);
	if (IS_ERR(req))
487
488
		goto out;

489
490
	attr_ver = fuse_get_attr_version(fc);

491
	req->out.page_zeroing = 1;
492
	req->out.argpages = 1;
493
494
	req->num_pages = 1;
	req->pages[0] = page;
495
	num_read = fuse_send_read(req, file, pos, count, NULL);
496
497
	err = req->out.h.error;
	fuse_put_request(fc, req);
498
499
500
501
502
503
504
505

	if (!err) {
		/*
		 * Short read means EOF.  If file size is larger, truncate it
		 */
		if (num_read < count)
			fuse_read_update_size(inode, pos + num_read, attr_ver);

506
		SetPageUptodate(page);
507
508
	}

509
	fuse_invalidate_attr(inode); /* atime changed */
510
511
512
513
514
 out:
	unlock_page(page);
	return err;
}

515
static void fuse_readpages_end(struct fuse_conn *fc, struct fuse_req *req)
516
{
517
	int i;
518
519
520
	size_t count = req->misc.read.in.size;
	size_t num_read = req->out.args[0].size;
	struct inode *inode = req->pages[0]->mapping->host;
521

522
523
524
525
526
527
528
529
530
	/*
	 * Short read means EOF.  If file size is larger, truncate it
	 */
	if (!req->out.h.error && num_read < count) {
		loff_t pos = page_offset(req->pages[0]) + num_read;
		fuse_read_update_size(inode, pos, req->misc.read.attr_ver);
	}

	fuse_invalidate_attr(inode); /* atime changed */
531

532
533
534
535
	for (i = 0; i < req->num_pages; i++) {
		struct page *page = req->pages[i];
		if (!req->out.h.error)
			SetPageUptodate(page);
536
537
		else
			SetPageError(page);
538
539
		unlock_page(page);
	}
540
541
	if (req->ff)
		fuse_file_put(req->ff);
542
543
}

544
static void fuse_send_readpages(struct fuse_req *req, struct file *file)
545
{
546
547
	struct fuse_file *ff = file->private_data;
	struct fuse_conn *fc = ff->fc;
548
549
	loff_t pos = page_offset(req->pages[0]);
	size_t count = req->num_pages << PAGE_CACHE_SHIFT;
550
551

	req->out.argpages = 1;
552
	req->out.page_zeroing = 1;
553
	fuse_read_fill(req, file, pos, count, FUSE_READ);
554
	req->misc.read.attr_ver = fuse_get_attr_version(fc);
555
	if (fc->async_read) {
556
		req->ff = fuse_file_get(ff);
557
		req->end = fuse_readpages_end;
558
		fuse_request_send_background(fc, req);
559
	} else {
560
		fuse_request_send(fc, req);
561
		fuse_readpages_end(fc, req);
562
		fuse_put_request(fc, req);
563
	}
564
565
}

566
struct fuse_fill_data {
567
	struct fuse_req *req;
568
	struct file *file;
569
570
571
572
573
	struct inode *inode;
};

static int fuse_readpages_fill(void *_data, struct page *page)
{
574
	struct fuse_fill_data *data = _data;
575
576
577
578
	struct fuse_req *req = data->req;
	struct inode *inode = data->inode;
	struct fuse_conn *fc = get_fuse_conn(inode);

Miklos Szeredi's avatar
Miklos Szeredi committed
579
580
	fuse_wait_on_page_writeback(inode, page->index);

581
582
583
584
	if (req->num_pages &&
	    (req->num_pages == FUSE_MAX_PAGES_PER_REQ ||
	     (req->num_pages + 1) * PAGE_CACHE_SIZE > fc->max_read ||
	     req->pages[req->num_pages - 1]->index + 1 != page->index)) {
585
		fuse_send_readpages(req, data->file);
586
587
		data->req = req = fuse_get_req(fc);
		if (IS_ERR(req)) {
588
			unlock_page(page);
589
			return PTR_ERR(req);
590
591
592
		}
	}
	req->pages[req->num_pages] = page;
Miklos Szeredi's avatar
Miklos Szeredi committed
593
	req->num_pages++;
594
595
596
597
598
599
600
601
	return 0;
}

static int fuse_readpages(struct file *file, struct address_space *mapping,
			  struct list_head *pages, unsigned nr_pages)
{
	struct inode *inode = mapping->host;
	struct fuse_conn *fc = get_fuse_conn(inode);
602
	struct fuse_fill_data data;
603
	int err;
604

605
	err = -EIO;
606
	if (is_bad_inode(inode))
607
		goto out;
608

609
	data.file = file;
610
	data.inode = inode;
611
	data.req = fuse_get_req(fc);
612
	err = PTR_ERR(data.req);
613
	if (IS_ERR(data.req))
614
		goto out;
615
616

	err = read_cache_pages(mapping, pages, fuse_readpages_fill, &data);
617
618
	if (!err) {
		if (data.req->num_pages)
619
			fuse_send_readpages(data.req, file);
620
621
622
		else
			fuse_put_request(fc, data.req);
	}
623
out:
624
	return err;
625
626
}

Miklos Szeredi's avatar
Miklos Szeredi committed
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
static ssize_t fuse_file_aio_read(struct kiocb *iocb, const struct iovec *iov,
				  unsigned long nr_segs, loff_t pos)
{
	struct inode *inode = iocb->ki_filp->f_mapping->host;

	if (pos + iov_length(iov, nr_segs) > i_size_read(inode)) {
		int err;
		/*
		 * If trying to read past EOF, make sure the i_size
		 * attribute is up-to-date.
		 */
		err = fuse_update_attributes(inode, NULL, iocb->ki_filp, NULL);
		if (err)
			return err;
	}

	return generic_file_aio_read(iocb, iov, nr_segs, pos);
}

646
static void fuse_write_fill(struct fuse_req *req, struct fuse_file *ff,
647
			    loff_t pos, size_t count)
648
{
649
650
	struct fuse_write_in *inarg = &req->misc.write.in;
	struct fuse_write_out *outarg = &req->misc.write.out;
651

652
653
654
	inarg->fh = ff->fh;
	inarg->offset = pos;
	inarg->size = count;
655
	req->in.h.opcode = FUSE_WRITE;
656
	req->in.h.nodeid = ff->nodeid;
657
	req->in.numargs = 2;
658
	if (ff->fc->minor < 9)
659
660
661
		req->in.args[0].size = FUSE_COMPAT_WRITE_IN_SIZE;
	else
		req->in.args[0].size = sizeof(struct fuse_write_in);
662
	req->in.args[0].value = inarg;
663
664
665
	req->in.args[1].size = count;
	req->out.numargs = 1;
	req->out.args[0].size = sizeof(struct fuse_write_out);
666
667
668
669
	req->out.args[0].value = outarg;
}

static size_t fuse_send_write(struct fuse_req *req, struct file *file,
670
			      loff_t pos, size_t count, fl_owner_t owner)
671
{
672
673
	struct fuse_file *ff = file->private_data;
	struct fuse_conn *fc = ff->fc;
674
675
	struct fuse_write_in *inarg = &req->misc.write.in;

676
	fuse_write_fill(req, ff, pos, count);
677
	inarg->flags = file->f_flags;
678
679
680
681
	if (owner != NULL) {
		inarg->write_flags |= FUSE_WRITE_LOCKOWNER;
		inarg->lock_owner = fuse_lock_owner_id(fc, owner);
	}
682
	fuse_request_send(fc, req);
683
	return req->misc.write.out.size;
684
685
}

Nick Piggin's avatar
Nick Piggin committed
686
687
688
static int fuse_write_begin(struct file *file, struct address_space *mapping,
			loff_t pos, unsigned len, unsigned flags,
			struct page **pagep, void **fsdata)
689
{
Nick Piggin's avatar
Nick Piggin committed
690
691
	pgoff_t index = pos >> PAGE_CACHE_SHIFT;

692
	*pagep = grab_cache_page_write_begin(mapping, index, flags);
Nick Piggin's avatar
Nick Piggin committed
693
694
	if (!*pagep)
		return -ENOMEM;
695
696
697
	return 0;
}

698
699
700
701
702
703
704
705
706
707
708
709
static void fuse_write_update_size(struct inode *inode, loff_t pos)
{
	struct fuse_conn *fc = get_fuse_conn(inode);
	struct fuse_inode *fi = get_fuse_inode(inode);

	spin_lock(&fc->lock);
	fi->attr_version = ++fc->attr_version;
	if (pos > inode->i_size)
		i_size_write(inode, pos);
	spin_unlock(&fc->lock);
}

Nick Piggin's avatar
Nick Piggin committed
710
711
static int fuse_buffered_write(struct file *file, struct inode *inode,
			       loff_t pos, unsigned count, struct page *page)
712
713
{
	int err;
714
	size_t nres;
715
	struct fuse_conn *fc = get_fuse_conn(inode);
Nick Piggin's avatar
Nick Piggin committed
716
	unsigned offset = pos & (PAGE_CACHE_SIZE - 1);
717
718
719
720
721
	struct fuse_req *req;

	if (is_bad_inode(inode))
		return -EIO;

Miklos Szeredi's avatar
Miklos Szeredi committed
722
723
724
725
726
727
	/*
	 * Make sure writepages on the same page are not mixed up with
	 * plain writes.
	 */
	fuse_wait_on_page_writeback(inode, page->index);

728
729
730
	req = fuse_get_req(fc);
	if (IS_ERR(req))
		return PTR_ERR(req);
731

732
	req->in.argpages = 1;
733
734
735
	req->num_pages = 1;
	req->pages[0] = page;
	req->page_offset = offset;
736
	nres = fuse_send_write(req, file, pos, count, NULL);
737
738
	err = req->out.h.error;
	fuse_put_request(fc, req);
Nick Piggin's avatar
Nick Piggin committed
739
	if (!err && !nres)
740
741
		err = -EIO;
	if (!err) {
Nick Piggin's avatar
Nick Piggin committed
742
		pos += nres;
743
		fuse_write_update_size(inode, pos);
Nick Piggin's avatar
Nick Piggin committed
744
		if (count == PAGE_CACHE_SIZE)
745
			SetPageUptodate(page);
746
747
	}
	fuse_invalidate_attr(inode);
Nick Piggin's avatar
Nick Piggin committed
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
	return err ? err : nres;
}

static int fuse_write_end(struct file *file, struct address_space *mapping,
			loff_t pos, unsigned len, unsigned copied,
			struct page *page, void *fsdata)
{
	struct inode *inode = mapping->host;
	int res = 0;

	if (copied)
		res = fuse_buffered_write(file, inode, pos, copied, page);

	unlock_page(page);
	page_cache_release(page);
	return res;
764
765
}

Nick Piggin's avatar
Nick Piggin committed
766
767
768
769
770
771
772
773
774
775
776
static size_t fuse_send_write_pages(struct fuse_req *req, struct file *file,
				    struct inode *inode, loff_t pos,
				    size_t count)
{
	size_t res;
	unsigned offset;
	unsigned i;

	for (i = 0; i < req->num_pages; i++)
		fuse_wait_on_page_writeback(inode, req->pages[i]->index);

777
	res = fuse_send_write(req, file, pos, count, NULL);
Nick Piggin's avatar
Nick Piggin committed
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808

	offset = req->page_offset;
	count = res;
	for (i = 0; i < req->num_pages; i++) {
		struct page *page = req->pages[i];

		if (!req->out.h.error && !offset && count >= PAGE_CACHE_SIZE)
			SetPageUptodate(page);

		if (count > PAGE_CACHE_SIZE - offset)
			count -= PAGE_CACHE_SIZE - offset;
		else
			count = 0;
		offset = 0;

		unlock_page(page);
		page_cache_release(page);
	}

	return res;
}

static ssize_t fuse_fill_write_pages(struct fuse_req *req,
			       struct address_space *mapping,
			       struct iov_iter *ii, loff_t pos)
{
	struct fuse_conn *fc = get_fuse_conn(mapping->host);
	unsigned offset = pos & (PAGE_CACHE_SIZE - 1);
	size_t count = 0;
	int err;

809
	req->in.argpages = 1;
Nick Piggin's avatar
Nick Piggin committed
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
	req->page_offset = offset;

	do {
		size_t tmp;
		struct page *page;
		pgoff_t index = pos >> PAGE_CACHE_SHIFT;
		size_t bytes = min_t(size_t, PAGE_CACHE_SIZE - offset,
				     iov_iter_count(ii));

		bytes = min_t(size_t, bytes, fc->max_write - count);

 again:
		err = -EFAULT;
		if (iov_iter_fault_in_readable(ii, bytes))
			break;

		err = -ENOMEM;
827
		page = grab_cache_page_write_begin(mapping, index, 0);
Nick Piggin's avatar
Nick Piggin committed
828
829
830
		if (!page)
			break;

831
832
833
		if (mapping_writably_mapped(mapping))
			flush_dcache_page(page);

Nick Piggin's avatar
Nick Piggin committed
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
		pagefault_disable();
		tmp = iov_iter_copy_from_user_atomic(page, ii, offset, bytes);
		pagefault_enable();
		flush_dcache_page(page);

		if (!tmp) {
			unlock_page(page);
			page_cache_release(page);
			bytes = min(bytes, iov_iter_single_seg_count(ii));
			goto again;
		}

		err = 0;
		req->pages[req->num_pages] = page;
		req->num_pages++;

		iov_iter_advance(ii, tmp);
		count += tmp;
		pos += tmp;
		offset += tmp;
		if (offset == PAGE_CACHE_SIZE)
			offset = 0;

857
858
		if (!fc->big_writes)
			break;
Nick Piggin's avatar
Nick Piggin committed
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
	} while (iov_iter_count(ii) && count < fc->max_write &&
		 req->num_pages < FUSE_MAX_PAGES_PER_REQ && offset == 0);

	return count > 0 ? count : err;
}

static ssize_t fuse_perform_write(struct file *file,
				  struct address_space *mapping,
				  struct iov_iter *ii, loff_t pos)
{
	struct inode *inode = mapping->host;
	struct fuse_conn *fc = get_fuse_conn(inode);
	int err = 0;
	ssize_t res = 0;

	if (is_bad_inode(inode))
		return -EIO;

	do {
		struct fuse_req *req;
		ssize_t count;

		req = fuse_get_req(fc);
		if (IS_ERR(req)) {
			err = PTR_ERR(req);
			break;
		}

		count = fuse_fill_write_pages(req, mapping, ii, pos);
		if (count <= 0) {
			err = count;
		} else {
			size_t num_written;

			num_written = fuse_send_write_pages(req, file, inode,
							    pos, count);
			err = req->out.h.error;
			if (!err) {
				res += num_written;
				pos += num_written;

				/* break out of the loop on short write */
				if (num_written != count)
					err = -EIO;
			}
		}
		fuse_put_request(fc, req);
	} while (!err && iov_iter_count(ii));

	if (res > 0)
		fuse_write_update_size(inode, pos);

	fuse_invalidate_attr(inode);

	return res > 0 ? res : err;
}

static ssize_t fuse_file_aio_write(struct kiocb *iocb, const struct iovec *iov,
				   unsigned long nr_segs, loff_t pos)
{
	struct file *file = iocb->ki_filp;
	struct address_space *mapping = file->f_mapping;
	size_t count = 0;
	ssize_t written = 0;
	struct inode *inode = mapping->host;
	ssize_t err;
	struct iov_iter i;

	WARN_ON(iocb->ki_pos != pos);

	err = generic_segment_checks(iov, &nr_segs, &count, VERIFY_READ);
	if (err)
		return err;

	mutex_lock(&inode->i_mutex);
	vfs_check_frozen(inode->i_sb, SB_FREEZE_WRITE);

	/* We can write back this queue in page reclaim */
	current->backing_dev_info = mapping->backing_dev_info;

	err = generic_write_checks(file, &pos, &count, S_ISBLK(inode->i_mode));
	if (err)
		goto out;

	if (count == 0)
		goto out;

946
	err = file_remove_suid(file);
Nick Piggin's avatar
Nick Piggin committed
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
	if (err)
		goto out;

	file_update_time(file);

	iov_iter_init(&i, iov, nr_segs, count, 0);
	written = fuse_perform_write(file, mapping, &i, pos);
	if (written >= 0)
		iocb->ki_pos = pos + written;

out:
	current->backing_dev_info = NULL;
	mutex_unlock(&inode->i_mutex);

	return written ? written : err;
}

Miklos Szeredi's avatar
Miklos Szeredi committed
964
965
966
967
968
969
970
971
972
973
974
975
976
static void fuse_release_user_pages(struct fuse_req *req, int write)
{
	unsigned i;

	for (i = 0; i < req->num_pages; i++) {
		struct page *page = req->pages[i];
		if (write)
			set_page_dirty_lock(page);
		put_page(page);
	}
}

static int fuse_get_user_pages(struct fuse_req *req, const char __user *buf,
977
			       size_t *nbytesp, int write)
Miklos Szeredi's avatar
Miklos Szeredi committed
978
{
979
	size_t nbytes = *nbytesp;
Miklos Szeredi's avatar
Miklos Szeredi committed
980
981
982
983
	unsigned long user_addr = (unsigned long) buf;
	unsigned offset = user_addr & ~PAGE_MASK;
	int npages;

984
985
986
987
988
989
990
991
992
	/* Special case for kernel I/O: can copy directly into the buffer */
	if (segment_eq(get_fs(), KERNEL_DS)) {
		if (write)
			req->in.args[1].value = (void *) user_addr;
		else
			req->out.args[0].value = (void *) user_addr;

		return 0;
	}
Miklos Szeredi's avatar
Miklos Szeredi committed
993

994
	nbytes = min_t(size_t, nbytes, FUSE_MAX_PAGES_PER_REQ << PAGE_SHIFT);
Miklos Szeredi's avatar
Miklos Szeredi committed
995
	npages = (nbytes + offset + PAGE_SIZE - 1) >> PAGE_SHIFT;
996
	npages = clamp(npages, 1, FUSE_MAX_PAGES_PER_REQ);
997
	npages = get_user_pages_fast(user_addr, npages, !write, req->pages);
Miklos Szeredi's avatar
Miklos Szeredi committed
998
999
1000
1001
1002
	if (npages < 0)
		return npages;

	req->num_pages = npages;
	req->page_offset = offset;
1003
1004
1005
1006
1007
1008
1009
1010
1011

	if (write)
		req->in.argpages = 1;
	else
		req->out.argpages = 1;

	nbytes = (req->num_pages << PAGE_SHIFT) - req->page_offset;
	*nbytesp = min(*nbytesp, nbytes);

Miklos Szeredi's avatar
Miklos Szeredi committed
1012
1013
1014
	return 0;
}

1015
1016
ssize_t fuse_direct_io(struct file *file, const char __user *buf,
		       size_t count, loff_t *ppos, int write)
Miklos Szeredi's avatar
Miklos Szeredi committed
1017
{
1018
1019
	struct fuse_file *ff = file->private_data;
	struct fuse_conn *fc = ff->fc;
Miklos Szeredi's avatar
Miklos Szeredi committed
1020
1021
1022
	size_t nmax = write ? fc->max_write : fc->max_read;
	loff_t pos = *ppos;
	ssize_t res = 0;
1023
1024
	struct fuse_req *req;

1025
1026
1027
	req = fuse_get_req(fc);
	if (IS_ERR(req))
		return PTR_ERR(req);
Miklos Szeredi's avatar
Miklos Szeredi committed
1028
1029
1030

	while (count) {
		size_t nres;
1031
		fl_owner_t owner = current->files;
1032
1033
		size_t nbytes = min(count, nmax);
		int err = fuse_get_user_pages(req, buf, &nbytes, write);
Miklos Szeredi's avatar
Miklos Szeredi committed
1034
1035
1036
1037
		if (err) {
			res = err;
			break;
		}
1038

Miklos Szeredi's avatar
Miklos Szeredi committed
1039
		if (write)
1040
			nres = fuse_send_write(req, file, pos, nbytes, owner);
Miklos Szeredi's avatar
Miklos Szeredi committed
1041
		else
1042
1043
			nres = fuse_send_read(req, file, pos, nbytes, owner);

Miklos Szeredi's avatar
Miklos Szeredi committed
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
		fuse_release_user_pages(req, !write);
		if (req->out.h.error) {
			if (!res)
				res = req->out.h.error;
			break;
		} else if (nres > nbytes) {
			res = -EIO;
			break;
		}
		count -= nres;
		res += nres;
		pos += nres;
		buf += nres;
		if (nres != nbytes)
			break;
1059
1060
1061
1062
1063
1064
		if (count) {
			fuse_put_request(fc, req);
			req = fuse_get_req(fc);
			if (IS_ERR(req))
				break;
		}
Miklos Szeredi's avatar
Miklos Szeredi committed
1065
	}
1066
1067
	if (!IS_ERR(req))
		fuse_put_request(fc, req);
1068
	if (res > 0)
Miklos Szeredi's avatar
Miklos Szeredi committed
1069
1070
1071
1072
		*ppos = pos;

	return res;
}
1073
EXPORT_SYMBOL_GPL(fuse_direct_io);
Miklos Szeredi's avatar
Miklos Szeredi committed
1074
1075
1076
1077

static ssize_t fuse_direct_read(struct file *file, char __user *buf,
				     size_t count, loff_t *ppos)
{
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
	ssize_t res;
	struct inode *inode = file->f_path.dentry->d_inode;

	if (is_bad_inode(inode))
		return -EIO;

	res = fuse_direct_io(file, buf, count, ppos, 0);

	fuse_invalidate_attr(inode);

	return res;
Miklos Szeredi's avatar
Miklos Szeredi committed
1089
1090
1091
1092
1093
}

static ssize_t fuse_direct_write(struct file *file, const char __user *buf,
				 size_t count, loff_t *ppos)
{
1094
	struct inode *inode = file->f_path.dentry->d_inode;
Miklos Szeredi's avatar
Miklos Szeredi committed
1095
	ssize_t res;
1096
1097
1098
1099

	if (is_bad_inode(inode))
		return -EIO;

Miklos Szeredi's avatar
Miklos Szeredi committed
1100
	/* Don't allow parallel writes to the same file */
1101
	mutex_lock(&inode->i_mutex);
1102
	res = generic_write_checks(file, ppos, &count, 0);
1103
	if (!res) {
1104
		res = fuse_direct_io(file, buf, count, ppos, 1);
1105
1106
1107
		if (res > 0)
			fuse_write_update_size(inode, *ppos);
	}
1108
	mutex_unlock(&inode->i_mutex);
1109
1110
1111

	fuse_invalidate_attr(inode);

Miklos Szeredi's avatar
Miklos Szeredi committed
1112
1113
1114
	return res;
}

Miklos Szeredi's avatar
Miklos Szeredi committed
1115
static void fuse_writepage_free(struct fuse_conn *fc, struct fuse_req *req)
1116
{
Miklos Szeredi's avatar
Miklos Szeredi committed
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
	__free_page(req->pages[0]);
	fuse_file_put(req->ff);
}

static void fuse_writepage_finish(struct fuse_conn *fc, struct fuse_req *req)
{
	struct inode *inode = req->inode;
	struct fuse_inode *fi = get_fuse_inode(inode);
	struct backing_dev_info *bdi = inode->i_mapping->backing_dev_info;

	list_del(&req->writepages_entry);
	dec_bdi_stat(bdi, BDI_WRITEBACK);
	dec_zone_page_state(req->pages[0], NR_WRITEBACK_TEMP);
	bdi_writeout_inc(bdi);
	wake_up(&fi->page_waitq);
}

/* Called under fc->lock, may release and reacquire it */
static void fuse_send_writepage(struct fuse_conn *fc, struct fuse_req *req)
1136
1137
__releases(&fc->lock)
__acquires(&fc->lock)
Miklos Szeredi's avatar
Miklos Szeredi committed
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
{
	struct fuse_inode *fi = get_fuse_inode(req->inode);
	loff_t size = i_size_read(req->inode);
	struct fuse_write_in *inarg = &req->misc.write.in;

	if (!fc->connected)
		goto out_free;

	if (inarg->offset + PAGE_CACHE_SIZE <= size) {
		inarg->size = PAGE_CACHE_SIZE;
	} else if (inarg->offset < size) {
		inarg->size = size & (PAGE_CACHE_SIZE - 1);
	} else {
		/* Got truncated off completely */
		goto out_free;
1153
	}
Miklos Szeredi's avatar
Miklos Szeredi committed
1154
1155
1156

	req->in.args[1].size = inarg->size;
	fi->writectr++;
1157
	fuse_request_send_background_locked(fc, req);
Miklos Szeredi's avatar
Miklos Szeredi committed
1158
1159
1160
1161
1162
1163
	return;

 out_free:
	fuse_writepage_finish(fc, req);
	spin_unlock(&fc->lock);
	fuse_writepage_free(fc, req);
1164
	fuse_put_request(fc, req);
Miklos Szeredi's avatar
Miklos Szeredi committed
1165
	spin_lock(&fc->lock);
1166
1167
}

Miklos Szeredi's avatar
Miklos Szeredi committed
1168
1169
1170
1171
1172
1173
1174
/*
 * If fi->writectr is positive (no truncate or fsync going on) send
 * all queued writepage requests.
 *
 * Called with fc->lock
 */
void fuse_flush_writepages(struct inode *inode)
1175
1176
__releases(&fc->lock)
__acquires(&fc->lock)
1177
{
Miklos Szeredi's avatar
Miklos Szeredi committed
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
	struct fuse_conn *fc = get_fuse_conn(inode);
	struct fuse_inode *fi = get_fuse_inode(inode);
	struct fuse_req *req;

	while (fi->writectr >= 0 && !list_empty(&fi->queued_writes)) {
		req = list_entry(fi->queued_writes.next, struct fuse_req, list);
		list_del_init(&req->list);
		fuse_send_writepage(fc, req);
	}
}

static void fuse_writepage_end(struct fuse_conn *fc, struct fuse_req *req)
{
	struct inode *inode = req->inode;
	struct fuse_inode *fi = get_fuse_inode(inode);

	mapping_set_error(inode->i_mapping, req->out.h.error);
	spin_lock(&fc->lock);
	fi->writectr--;
	fuse_writepage_finish(fc, req);
	spin_unlock(&fc->lock);
	fuse_writepage_free(fc, req);
}

static int fuse_writepage_locked(struct page *page)
{
	struct address_space *mapping = page->mapping;
	struct inode *inode = mapping->host;
	struct fuse_conn *fc = get_fuse_conn(inode);
	struct fuse_inode *fi = get_fuse_inode(inode);
	struct fuse_req *req;
	struct fuse_file *ff;
	struct page *tmp_page;

	set_page_writeback(page);

	req = fuse_request_alloc_nofs();
	if (!req)
		goto err;

	tmp_page = alloc_page(GFP_NOFS | __GFP_HIGHMEM);
	if (!tmp_page)
		goto err_free;

	spin_lock(&fc->lock);
	BUG_ON(list_empty(&fi->write_files));
	ff = list_entry(fi->write_files.next, struct fuse_file, write_entry);
	req->ff = fuse_file_get(ff);
	spin_unlock(&fc->lock);

1228
	fuse_write_fill(req, ff, page_offset(page), 0);
Miklos Szeredi's avatar
Miklos Szeredi committed
1229
1230

	copy_highpage(tmp_page, page);
1231
	req->misc.write.in.write_flags |= FUSE_WRITE_CACHE;
1232
	req->in.argpages = 1;
Miklos Szeredi's avatar
Miklos Szeredi committed
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
	req->num_pages = 1;
	req->pages[0] = tmp_page;
	req->page_offset = 0;
	req->end = fuse_writepage_end;
	req->inode = inode;

	inc_bdi_stat(mapping->backing_dev_info, BDI_WRITEBACK);
	inc_zone_page_state(tmp_page, NR_WRITEBACK_TEMP);
	end_page_writeback(page);

	spin_lock(&fc->lock);
	list_add(&req->writepages_entry, &fi->writepages);
	list_add_tail(&req->list, &fi->queued_writes);
	fuse_flush_writepages(inode);
	spin_unlock(&fc->lock);

	return 0;

err_free:
	fuse_request_free(req);
err:
	end_page_writeback(page);
	return -ENOMEM;
}

static int fuse_writepage(struct page *page, struct writeback_control *wbc)
{
	int err;

	err = fuse_writepage_locked(page);
	unlock_page(page);

	return err;
}

static int fuse_launder_page(struct page *page)
{
	int err = 0;
	if (clear_page_dirty_for_io(page)) {
		struct inode *inode = page->mapping->host;
		err = fuse_writepage_locked(page);
		if (!err)
			fuse_wait_on_page_writeback(inode, page->index);
	}