dir.c 39.7 KB
Newer Older
1 2
/*
  FUSE: Filesystem in Userspace
M
Miklos Szeredi 已提交
3
  Copyright (C) 2001-2008  Miklos Szeredi <miklos@szeredi.hu>
4 5 6 7 8 9 10 11 12 13 14 15 16

  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/file.h>
#include <linux/gfp.h>
#include <linux/sched.h>
#include <linux/namei.h>

M
Miklos Szeredi 已提交
17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43
#if BITS_PER_LONG >= 64
static inline void fuse_dentry_settime(struct dentry *entry, u64 time)
{
	entry->d_time = time;
}

static inline u64 fuse_dentry_time(struct dentry *entry)
{
	return entry->d_time;
}
#else
/*
 * On 32 bit archs store the high 32 bits of time in d_fsdata
 */
static void fuse_dentry_settime(struct dentry *entry, u64 time)
{
	entry->d_time = time;
	entry->d_fsdata = (void *) (unsigned long) (time >> 32);
}

static u64 fuse_dentry_time(struct dentry *entry)
{
	return (u64) entry->d_time +
		((u64) (unsigned long) entry->d_fsdata << 32);
}
#endif

44 45 46 47 48 49 50 51 52
/*
 * FUSE caches dentries and attributes with separate timeout.  The
 * time in jiffies until the dentry/attributes are valid is stored in
 * dentry->d_time and fuse_inode->i_time respectively.
 */

/*
 * Calculate the time in jiffies until a dentry/attributes are valid
 */
M
Miklos Szeredi 已提交
53
static u64 time_to_jiffies(unsigned long sec, unsigned long nsec)
54
{
M
Miklos Szeredi 已提交
55 56
	if (sec || nsec) {
		struct timespec ts = {sec, nsec};
M
Miklos Szeredi 已提交
57
		return get_jiffies_64() + timespec_to_jiffies(&ts);
M
Miklos Szeredi 已提交
58
	} else
M
Miklos Szeredi 已提交
59
		return 0;
60 61
}

62 63 64 65
/*
 * Set dentry and possibly attribute timeouts from the lookup/mk*
 * replies
 */
66 67
static void fuse_change_entry_timeout(struct dentry *entry,
				      struct fuse_entry_out *o)
68
{
M
Miklos Szeredi 已提交
69 70
	fuse_dentry_settime(entry,
		time_to_jiffies(o->entry_valid, o->entry_valid_nsec));
71 72 73 74 75 76 77 78 79 80
}

static u64 attr_timeout(struct fuse_attr_out *o)
{
	return time_to_jiffies(o->attr_valid, o->attr_valid_nsec);
}

static u64 entry_attr_timeout(struct fuse_entry_out *o)
{
	return time_to_jiffies(o->attr_valid, o->attr_valid_nsec);
81 82
}

83 84 85 86
/*
 * Mark the attributes as stale, so that at the next call to
 * ->getattr() they will be fetched from userspace
 */
87 88
void fuse_invalidate_attr(struct inode *inode)
{
M
Miklos Szeredi 已提交
89
	get_fuse_inode(inode)->i_time = 0;
90 91
}

92 93 94 95 96 97 98 99
/*
 * Just mark the entry as stale, so that a next attempt to look it up
 * will result in a new lookup call to userspace
 *
 * This is called when a dentry is about to become negative and the
 * timeout is unknown (unlink, rmdir, rename and in some cases
 * lookup)
 */
M
Miklos Szeredi 已提交
100
void fuse_invalidate_entry_cache(struct dentry *entry)
101
{
M
Miklos Szeredi 已提交
102
	fuse_dentry_settime(entry, 0);
103 104
}

105 106 107 108
/*
 * Same as fuse_invalidate_entry_cache(), but also try to remove the
 * dentry from the hash
 */
109 110 111 112
static void fuse_invalidate_entry(struct dentry *entry)
{
	d_invalidate(entry);
	fuse_invalidate_entry_cache(entry);
113 114
}

115 116
static void fuse_lookup_init(struct fuse_conn *fc, struct fuse_req *req,
			     u64 nodeid, struct qstr *name,
117 118
			     struct fuse_entry_out *outarg)
{
119
	memset(outarg, 0, sizeof(struct fuse_entry_out));
120
	req->in.h.opcode = FUSE_LOOKUP;
121
	req->in.h.nodeid = nodeid;
122
	req->in.numargs = 1;
123 124
	req->in.args[0].size = name->len + 1;
	req->in.args[0].value = name->name;
125
	req->out.numargs = 1;
126 127 128 129
	if (fc->minor < 9)
		req->out.args[0].size = FUSE_COMPAT_ENTRY_OUT_SIZE;
	else
		req->out.args[0].size = sizeof(struct fuse_entry_out);
130 131 132
	req->out.args[0].value = outarg;
}

133
u64 fuse_get_attr_version(struct fuse_conn *fc)
134 135 136 137 138 139 140 141 142 143 144 145 146 147
{
	u64 curr_version;

	/*
	 * The spin lock isn't actually needed on 64bit archs, but we
	 * don't yet care too much about such optimizations.
	 */
	spin_lock(&fc->lock);
	curr_version = fc->attr_version;
	spin_unlock(&fc->lock);

	return curr_version;
}

148 149 150 151 152 153 154 155 156
/*
 * Check whether the dentry is still valid
 *
 * If the entry validity timeout has expired and the dentry is
 * positive, try to redo the lookup.  If the lookup results in a
 * different inode, then let the VFS invalidate the dentry and redo
 * the lookup once more.  If the lookup results in the same inode,
 * then refresh the attributes, timeouts and mark the dentry valid.
 */
157 158
static int fuse_dentry_revalidate(struct dentry *entry, struct nameidata *nd)
{
159
	struct inode *inode;
160

161 162 163 164
	if (nd->flags & LOOKUP_RCU)
		return -ECHILD;

	inode = entry->d_inode;
165
	if (inode && is_bad_inode(inode))
166
		return 0;
M
Miklos Szeredi 已提交
167
	else if (fuse_dentry_time(entry) < get_jiffies_64()) {
168 169
		int err;
		struct fuse_entry_out outarg;
170 171
		struct fuse_conn *fc;
		struct fuse_req *req;
M
Miklos Szeredi 已提交
172
		struct fuse_req *forget_req;
173
		struct dentry *parent;
174
		u64 attr_version;
175

176
		/* For negative dentries, always do a fresh lookup */
177 178 179 180
		if (!inode)
			return 0;

		fc = get_fuse_conn(inode);
181 182
		req = fuse_get_req(fc);
		if (IS_ERR(req))
183 184
			return 0;

M
Miklos Szeredi 已提交
185 186 187 188 189 190
		forget_req = fuse_get_req(fc);
		if (IS_ERR(forget_req)) {
			fuse_put_request(fc, req);
			return 0;
		}

191
		attr_version = fuse_get_attr_version(fc);
192

193
		parent = dget_parent(entry);
194 195
		fuse_lookup_init(fc, req, get_node_id(parent->d_inode),
				 &entry->d_name, &outarg);
196
		fuse_request_send(fc, req);
197
		dput(parent);
198
		err = req->out.h.error;
M
Miklos Szeredi 已提交
199
		fuse_put_request(fc, req);
200 201 202
		/* Zero nodeid is same as -ENOENT */
		if (!err && !outarg.nodeid)
			err = -ENOENT;
203
		if (!err) {
204
			struct fuse_inode *fi = get_fuse_inode(inode);
205
			if (outarg.nodeid != get_node_id(inode)) {
M
Miklos Szeredi 已提交
206 207
				fuse_send_forget(fc, forget_req,
						 outarg.nodeid, 1);
208 209
				return 0;
			}
210
			spin_lock(&fc->lock);
M
Miklos Szeredi 已提交
211
			fi->nlookup++;
212
			spin_unlock(&fc->lock);
213
		}
M
Miklos Szeredi 已提交
214
		fuse_put_request(fc, forget_req);
215
		if (err || (outarg.attr.mode ^ inode->i_mode) & S_IFMT)
216 217
			return 0;

218 219 220 221
		fuse_change_attributes(inode, &outarg.attr,
				       entry_attr_timeout(&outarg),
				       attr_version);
		fuse_change_entry_timeout(entry, &outarg);
222 223 224 225
	}
	return 1;
}

226
static int invalid_nodeid(u64 nodeid)
227 228 229 230
{
	return !nodeid || nodeid == FUSE_ROOT_ID;
}

A
Al Viro 已提交
231
const struct dentry_operations fuse_dentry_operations = {
232 233 234
	.d_revalidate	= fuse_dentry_revalidate,
};

235
int fuse_valid_type(int m)
236 237 238 239 240
{
	return S_ISREG(m) || S_ISDIR(m) || S_ISLNK(m) || S_ISCHR(m) ||
		S_ISBLK(m) || S_ISFIFO(m) || S_ISSOCK(m);
}

241 242 243 244
/*
 * Add a directory inode to a dentry, ensuring that no other dentry
 * refers to this inode.  Called with fc->inst_mutex.
 */
245 246
static struct dentry *fuse_d_add_directory(struct dentry *entry,
					   struct inode *inode)
247 248
{
	struct dentry *alias = d_find_alias(inode);
249
	if (alias && !(alias->d_flags & DCACHE_DISCONNECTED)) {
250 251 252 253
		/* This tries to shrink the subtree below alias */
		fuse_invalidate_entry(alias);
		dput(alias);
		if (!list_empty(&inode->i_dentry))
254 255 256
			return ERR_PTR(-EBUSY);
	} else {
		dput(alias);
257
	}
258
	return d_splice_alias(inode, entry);
259 260
}

261 262
int fuse_lookup_name(struct super_block *sb, u64 nodeid, struct qstr *name,
		     struct fuse_entry_out *outarg, struct inode **inode)
263
{
264
	struct fuse_conn *fc = get_fuse_conn_super(sb);
265
	struct fuse_req *req;
M
Miklos Szeredi 已提交
266
	struct fuse_req *forget_req;
267
	u64 attr_version;
268
	int err;
269

270 271 272 273
	*inode = NULL;
	err = -ENAMETOOLONG;
	if (name->len > FUSE_NAME_MAX)
		goto out;
274

275
	req = fuse_get_req(fc);
276
	err = PTR_ERR(req);
277
	if (IS_ERR(req))
278
		goto out;
279

M
Miklos Szeredi 已提交
280
	forget_req = fuse_get_req(fc);
281
	err = PTR_ERR(forget_req);
M
Miklos Szeredi 已提交
282 283
	if (IS_ERR(forget_req)) {
		fuse_put_request(fc, req);
284
		goto out;
M
Miklos Szeredi 已提交
285 286
	}

287
	attr_version = fuse_get_attr_version(fc);
288

289
	fuse_lookup_init(fc, req, nodeid, name, outarg);
290
	fuse_request_send(fc, req);
291
	err = req->out.h.error;
M
Miklos Szeredi 已提交
292
	fuse_put_request(fc, req);
293
	/* Zero nodeid is same as -ENOENT, but with valid timeout */
294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309
	if (err || !outarg->nodeid)
		goto out_put_forget;

	err = -EIO;
	if (!outarg->nodeid)
		goto out_put_forget;
	if (!fuse_valid_type(outarg->attr.mode))
		goto out_put_forget;

	*inode = fuse_iget(sb, outarg->nodeid, outarg->generation,
			   &outarg->attr, entry_attr_timeout(outarg),
			   attr_version);
	err = -ENOMEM;
	if (!*inode) {
		fuse_send_forget(fc, forget_req, outarg->nodeid, 1);
		goto out;
310
	}
311 312 313
	err = 0;

 out_put_forget:
M
Miklos Szeredi 已提交
314
	fuse_put_request(fc, forget_req);
315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340
 out:
	return err;
}

static struct dentry *fuse_lookup(struct inode *dir, struct dentry *entry,
				  struct nameidata *nd)
{
	int err;
	struct fuse_entry_out outarg;
	struct inode *inode;
	struct dentry *newent;
	struct fuse_conn *fc = get_fuse_conn(dir);
	bool outarg_valid = true;

	err = fuse_lookup_name(dir->i_sb, get_node_id(dir), &entry->d_name,
			       &outarg, &inode);
	if (err == -ENOENT) {
		outarg_valid = false;
		err = 0;
	}
	if (err)
		goto out_err;

	err = -EIO;
	if (inode && get_node_id(inode) == FUSE_ROOT_ID)
		goto out_iput;
341

342 343
	if (inode && S_ISDIR(inode->i_mode)) {
		mutex_lock(&fc->inst_mutex);
344
		newent = fuse_d_add_directory(entry, inode);
345
		mutex_unlock(&fc->inst_mutex);
346 347 348 349
		err = PTR_ERR(newent);
		if (IS_ERR(newent))
			goto out_iput;
	} else {
350
		newent = d_splice_alias(inode, entry);
351
	}
352

353
	entry = newent ? newent : entry;
354
	d_set_d_op(entry, &fuse_dentry_operations);
355
	if (outarg_valid)
356
		fuse_change_entry_timeout(entry, &outarg);
357 358
	else
		fuse_invalidate_entry_cache(entry);
359

360
	return newent;
361 362 363 364 365

 out_iput:
	iput(inode);
 out_err:
	return ERR_PTR(err);
366 367
}

368 369 370 371 372 373
/*
 * Atomic create+open operation
 *
 * If the filesystem doesn't support this, then fall back to separate
 * 'mknod' + 'open' requests.
 */
M
Miklos Szeredi 已提交
374 375 376 377 378 379 380
static int fuse_create_open(struct inode *dir, struct dentry *entry, int mode,
			    struct nameidata *nd)
{
	int err;
	struct inode *inode;
	struct fuse_conn *fc = get_fuse_conn(dir);
	struct fuse_req *req;
381
	struct fuse_req *forget_req;
382
	struct fuse_create_in inarg;
M
Miklos Szeredi 已提交
383 384 385 386 387 388 389
	struct fuse_open_out outopen;
	struct fuse_entry_out outentry;
	struct fuse_file *ff;
	struct file *file;
	int flags = nd->intent.open.flags - 1;

	if (fc->no_create)
390
		return -ENOSYS;
M
Miklos Szeredi 已提交
391

392 393 394
	if (flags & O_DIRECT)
		return -EINVAL;

395 396 397 398
	forget_req = fuse_get_req(fc);
	if (IS_ERR(forget_req))
		return PTR_ERR(forget_req);

399
	req = fuse_get_req(fc);
400
	err = PTR_ERR(req);
401
	if (IS_ERR(req))
402
		goto out_put_forget_req;
M
Miklos Szeredi 已提交
403

404
	err = -ENOMEM;
T
Tejun Heo 已提交
405
	ff = fuse_file_alloc(fc);
M
Miklos Szeredi 已提交
406 407 408
	if (!ff)
		goto out_put_request;

409 410 411
	if (!fc->dont_mask)
		mode &= ~current_umask();

M
Miklos Szeredi 已提交
412 413
	flags &= ~O_NOCTTY;
	memset(&inarg, 0, sizeof(inarg));
414
	memset(&outentry, 0, sizeof(outentry));
M
Miklos Szeredi 已提交
415 416
	inarg.flags = flags;
	inarg.mode = mode;
417
	inarg.umask = current_umask();
M
Miklos Szeredi 已提交
418 419 420
	req->in.h.opcode = FUSE_CREATE;
	req->in.h.nodeid = get_node_id(dir);
	req->in.numargs = 2;
421 422
	req->in.args[0].size = fc->minor < 12 ? sizeof(struct fuse_open_in) :
						sizeof(inarg);
M
Miklos Szeredi 已提交
423 424 425 426
	req->in.args[0].value = &inarg;
	req->in.args[1].size = entry->d_name.len + 1;
	req->in.args[1].value = entry->d_name.name;
	req->out.numargs = 2;
427 428 429 430
	if (fc->minor < 9)
		req->out.args[0].size = FUSE_COMPAT_ENTRY_OUT_SIZE;
	else
		req->out.args[0].size = sizeof(outentry);
M
Miklos Szeredi 已提交
431 432 433
	req->out.args[0].value = &outentry;
	req->out.args[1].size = sizeof(outopen);
	req->out.args[1].value = &outopen;
434
	fuse_request_send(fc, req);
M
Miklos Szeredi 已提交
435 436 437 438 439 440 441 442
	err = req->out.h.error;
	if (err) {
		if (err == -ENOSYS)
			fc->no_create = 1;
		goto out_free_ff;
	}

	err = -EIO;
443
	if (!S_ISREG(outentry.attr.mode) || invalid_nodeid(outentry.nodeid))
M
Miklos Szeredi 已提交
444 445
		goto out_free_ff;

446
	fuse_put_request(fc, req);
447 448 449
	ff->fh = outopen.fh;
	ff->nodeid = outentry.nodeid;
	ff->open_flags = outopen.open_flags;
M
Miklos Szeredi 已提交
450
	inode = fuse_iget(dir->i_sb, outentry.nodeid, outentry.generation,
451
			  &outentry.attr, entry_attr_timeout(&outentry), 0);
M
Miklos Szeredi 已提交
452 453
	if (!inode) {
		flags &= ~(O_CREAT | O_EXCL | O_TRUNC);
454
		fuse_sync_release(ff, flags);
455 456
		fuse_send_forget(fc, forget_req, outentry.nodeid, 1);
		return -ENOMEM;
M
Miklos Szeredi 已提交
457
	}
458
	fuse_put_request(fc, forget_req);
M
Miklos Szeredi 已提交
459
	d_instantiate(entry, inode);
460
	fuse_change_entry_timeout(entry, &outentry);
461
	fuse_invalidate_attr(dir);
M
Miklos Szeredi 已提交
462 463
	file = lookup_instantiate_filp(nd, entry, generic_file_open);
	if (IS_ERR(file)) {
464
		fuse_sync_release(ff, flags);
M
Miklos Szeredi 已提交
465 466
		return PTR_ERR(file);
	}
467 468
	file->private_data = fuse_file_get(ff);
	fuse_finish_open(inode, file);
M
Miklos Szeredi 已提交
469 470 471 472 473 474
	return 0;

 out_free_ff:
	fuse_file_free(ff);
 out_put_request:
	fuse_put_request(fc, req);
475 476
 out_put_forget_req:
	fuse_put_request(fc, forget_req);
M
Miklos Szeredi 已提交
477 478 479
	return err;
}

480 481 482
/*
 * Code shared between mknod, mkdir, symlink and link
 */
483 484 485 486 487 488 489
static int create_new_entry(struct fuse_conn *fc, struct fuse_req *req,
			    struct inode *dir, struct dentry *entry,
			    int mode)
{
	struct fuse_entry_out outarg;
	struct inode *inode;
	int err;
M
Miklos Szeredi 已提交
490 491 492 493 494 495 496
	struct fuse_req *forget_req;

	forget_req = fuse_get_req(fc);
	if (IS_ERR(forget_req)) {
		fuse_put_request(fc, req);
		return PTR_ERR(forget_req);
	}
497

498
	memset(&outarg, 0, sizeof(outarg));
499 500
	req->in.h.nodeid = get_node_id(dir);
	req->out.numargs = 1;
501 502 503 504
	if (fc->minor < 9)
		req->out.args[0].size = FUSE_COMPAT_ENTRY_OUT_SIZE;
	else
		req->out.args[0].size = sizeof(outarg);
505
	req->out.args[0].value = &outarg;
506
	fuse_request_send(fc, req);
507
	err = req->out.h.error;
M
Miklos Szeredi 已提交
508 509 510 511
	fuse_put_request(fc, req);
	if (err)
		goto out_put_forget_req;

512 513
	err = -EIO;
	if (invalid_nodeid(outarg.nodeid))
M
Miklos Szeredi 已提交
514
		goto out_put_forget_req;
515 516

	if ((outarg.attr.mode ^ mode) & S_IFMT)
M
Miklos Szeredi 已提交
517
		goto out_put_forget_req;
518

519
	inode = fuse_iget(dir->i_sb, outarg.nodeid, outarg.generation,
520
			  &outarg.attr, entry_attr_timeout(&outarg), 0);
521
	if (!inode) {
M
Miklos Szeredi 已提交
522
		fuse_send_forget(fc, forget_req, outarg.nodeid, 1);
523 524
		return -ENOMEM;
	}
M
Miklos Szeredi 已提交
525
	fuse_put_request(fc, forget_req);
526

527 528 529 530 531 532 533 534 535 536 537 538 539 540 541
	if (S_ISDIR(inode->i_mode)) {
		struct dentry *alias;
		mutex_lock(&fc->inst_mutex);
		alias = d_find_alias(inode);
		if (alias) {
			/* New directory must have moved since mkdir */
			mutex_unlock(&fc->inst_mutex);
			dput(alias);
			iput(inode);
			return -EBUSY;
		}
		d_instantiate(entry, inode);
		mutex_unlock(&fc->inst_mutex);
	} else
		d_instantiate(entry, inode);
542

543
	fuse_change_entry_timeout(entry, &outarg);
544 545
	fuse_invalidate_attr(dir);
	return 0;
546

M
Miklos Szeredi 已提交
547 548
 out_put_forget_req:
	fuse_put_request(fc, forget_req);
549
	return err;
550 551 552 553 554 555 556
}

static int fuse_mknod(struct inode *dir, struct dentry *entry, int mode,
		      dev_t rdev)
{
	struct fuse_mknod_in inarg;
	struct fuse_conn *fc = get_fuse_conn(dir);
557 558 559
	struct fuse_req *req = fuse_get_req(fc);
	if (IS_ERR(req))
		return PTR_ERR(req);
560

561 562 563
	if (!fc->dont_mask)
		mode &= ~current_umask();

564 565 566
	memset(&inarg, 0, sizeof(inarg));
	inarg.mode = mode;
	inarg.rdev = new_encode_dev(rdev);
567
	inarg.umask = current_umask();
568 569
	req->in.h.opcode = FUSE_MKNOD;
	req->in.numargs = 2;
570 571
	req->in.args[0].size = fc->minor < 12 ? FUSE_COMPAT_MKNOD_IN_SIZE :
						sizeof(inarg);
572 573 574 575 576 577 578 579 580
	req->in.args[0].value = &inarg;
	req->in.args[1].size = entry->d_name.len + 1;
	req->in.args[1].value = entry->d_name.name;
	return create_new_entry(fc, req, dir, entry, mode);
}

static int fuse_create(struct inode *dir, struct dentry *entry, int mode,
		       struct nameidata *nd)
{
M
Miklos Szeredi 已提交
581
	if (nd && (nd->flags & LOOKUP_OPEN)) {
M
Miklos Szeredi 已提交
582 583 584 585 586
		int err = fuse_create_open(dir, entry, mode, nd);
		if (err != -ENOSYS)
			return err;
		/* Fall back on mknod */
	}
587 588 589 590 591 592 593
	return fuse_mknod(dir, entry, mode, 0);
}

static int fuse_mkdir(struct inode *dir, struct dentry *entry, int mode)
{
	struct fuse_mkdir_in inarg;
	struct fuse_conn *fc = get_fuse_conn(dir);
594 595 596
	struct fuse_req *req = fuse_get_req(fc);
	if (IS_ERR(req))
		return PTR_ERR(req);
597

598 599 600
	if (!fc->dont_mask)
		mode &= ~current_umask();

601 602
	memset(&inarg, 0, sizeof(inarg));
	inarg.mode = mode;
603
	inarg.umask = current_umask();
604 605 606 607 608 609 610 611 612 613 614 615 616 617
	req->in.h.opcode = FUSE_MKDIR;
	req->in.numargs = 2;
	req->in.args[0].size = sizeof(inarg);
	req->in.args[0].value = &inarg;
	req->in.args[1].size = entry->d_name.len + 1;
	req->in.args[1].value = entry->d_name.name;
	return create_new_entry(fc, req, dir, entry, S_IFDIR);
}

static int fuse_symlink(struct inode *dir, struct dentry *entry,
			const char *link)
{
	struct fuse_conn *fc = get_fuse_conn(dir);
	unsigned len = strlen(link) + 1;
618 619 620
	struct fuse_req *req = fuse_get_req(fc);
	if (IS_ERR(req))
		return PTR_ERR(req);
621 622 623 624 625 626 627 628 629 630 631 632 633 634

	req->in.h.opcode = FUSE_SYMLINK;
	req->in.numargs = 2;
	req->in.args[0].size = entry->d_name.len + 1;
	req->in.args[0].value = entry->d_name.name;
	req->in.args[1].size = len;
	req->in.args[1].value = link;
	return create_new_entry(fc, req, dir, entry, S_IFLNK);
}

static int fuse_unlink(struct inode *dir, struct dentry *entry)
{
	int err;
	struct fuse_conn *fc = get_fuse_conn(dir);
635 636 637
	struct fuse_req *req = fuse_get_req(fc);
	if (IS_ERR(req))
		return PTR_ERR(req);
638 639 640 641 642 643

	req->in.h.opcode = FUSE_UNLINK;
	req->in.h.nodeid = get_node_id(dir);
	req->in.numargs = 1;
	req->in.args[0].size = entry->d_name.len + 1;
	req->in.args[0].value = entry->d_name.name;
644
	fuse_request_send(fc, req);
645 646 647 648 649
	err = req->out.h.error;
	fuse_put_request(fc, req);
	if (!err) {
		struct inode *inode = entry->d_inode;

M
Miklos Szeredi 已提交
650 651 652 653 654
		/*
		 * Set nlink to zero so the inode can be cleared, if the inode
		 * does have more links this will be discovered at the next
		 * lookup/getattr.
		 */
655
		clear_nlink(inode);
656 657
		fuse_invalidate_attr(inode);
		fuse_invalidate_attr(dir);
658
		fuse_invalidate_entry_cache(entry);
659 660 661 662 663 664 665 666 667
	} else if (err == -EINTR)
		fuse_invalidate_entry(entry);
	return err;
}

static int fuse_rmdir(struct inode *dir, struct dentry *entry)
{
	int err;
	struct fuse_conn *fc = get_fuse_conn(dir);
668 669 670
	struct fuse_req *req = fuse_get_req(fc);
	if (IS_ERR(req))
		return PTR_ERR(req);
671 672 673 674 675 676

	req->in.h.opcode = FUSE_RMDIR;
	req->in.h.nodeid = get_node_id(dir);
	req->in.numargs = 1;
	req->in.args[0].size = entry->d_name.len + 1;
	req->in.args[0].value = entry->d_name.name;
677
	fuse_request_send(fc, req);
678 679 680
	err = req->out.h.error;
	fuse_put_request(fc, req);
	if (!err) {
681
		clear_nlink(entry->d_inode);
682
		fuse_invalidate_attr(dir);
683
		fuse_invalidate_entry_cache(entry);
684 685 686 687 688 689 690 691 692 693 694
	} else if (err == -EINTR)
		fuse_invalidate_entry(entry);
	return err;
}

static int fuse_rename(struct inode *olddir, struct dentry *oldent,
		       struct inode *newdir, struct dentry *newent)
{
	int err;
	struct fuse_rename_in inarg;
	struct fuse_conn *fc = get_fuse_conn(olddir);
695 696 697
	struct fuse_req *req = fuse_get_req(fc);
	if (IS_ERR(req))
		return PTR_ERR(req);
698 699 700 701 702 703 704 705 706 707 708 709

	memset(&inarg, 0, sizeof(inarg));
	inarg.newdir = get_node_id(newdir);
	req->in.h.opcode = FUSE_RENAME;
	req->in.h.nodeid = get_node_id(olddir);
	req->in.numargs = 3;
	req->in.args[0].size = sizeof(inarg);
	req->in.args[0].value = &inarg;
	req->in.args[1].size = oldent->d_name.len + 1;
	req->in.args[1].value = oldent->d_name.name;
	req->in.args[2].size = newent->d_name.len + 1;
	req->in.args[2].value = newent->d_name.name;
710
	fuse_request_send(fc, req);
711 712 713
	err = req->out.h.error;
	fuse_put_request(fc, req);
	if (!err) {
714 715 716
		/* ctime changes */
		fuse_invalidate_attr(oldent->d_inode);

717 718 719
		fuse_invalidate_attr(olddir);
		if (olddir != newdir)
			fuse_invalidate_attr(newdir);
720 721

		/* newent will end up negative */
M
Miklos Szeredi 已提交
722 723
		if (newent->d_inode) {
			fuse_invalidate_attr(newent->d_inode);
724
			fuse_invalidate_entry_cache(newent);
M
Miklos Szeredi 已提交
725
		}
726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746
	} else if (err == -EINTR) {
		/* If request was interrupted, DEITY only knows if the
		   rename actually took place.  If the invalidation
		   fails (e.g. some process has CWD under the renamed
		   directory), then there can be inconsistency between
		   the dcache and the real filesystem.  Tough luck. */
		fuse_invalidate_entry(oldent);
		if (newent->d_inode)
			fuse_invalidate_entry(newent);
	}

	return err;
}

static int fuse_link(struct dentry *entry, struct inode *newdir,
		     struct dentry *newent)
{
	int err;
	struct fuse_link_in inarg;
	struct inode *inode = entry->d_inode;
	struct fuse_conn *fc = get_fuse_conn(inode);
747 748 749
	struct fuse_req *req = fuse_get_req(fc);
	if (IS_ERR(req))
		return PTR_ERR(req);
750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770

	memset(&inarg, 0, sizeof(inarg));
	inarg.oldnodeid = get_node_id(inode);
	req->in.h.opcode = FUSE_LINK;
	req->in.numargs = 2;
	req->in.args[0].size = sizeof(inarg);
	req->in.args[0].value = &inarg;
	req->in.args[1].size = newent->d_name.len + 1;
	req->in.args[1].value = newent->d_name.name;
	err = create_new_entry(fc, req, newdir, newent, inode->i_mode);
	/* Contrary to "normal" filesystems it can happen that link
	   makes two "logical" inodes point to the same "physical"
	   inode.  We invalidate the attributes of the old one, so it
	   will reflect changes in the backing inode (link count,
	   etc.)
	*/
	if (!err || err == -EINTR)
		fuse_invalidate_attr(inode);
	return err;
}

771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791
static void fuse_fillattr(struct inode *inode, struct fuse_attr *attr,
			  struct kstat *stat)
{
	stat->dev = inode->i_sb->s_dev;
	stat->ino = attr->ino;
	stat->mode = (inode->i_mode & S_IFMT) | (attr->mode & 07777);
	stat->nlink = attr->nlink;
	stat->uid = attr->uid;
	stat->gid = attr->gid;
	stat->rdev = inode->i_rdev;
	stat->atime.tv_sec = attr->atime;
	stat->atime.tv_nsec = attr->atimensec;
	stat->mtime.tv_sec = attr->mtime;
	stat->mtime.tv_nsec = attr->mtimensec;
	stat->ctime.tv_sec = attr->ctime;
	stat->ctime.tv_nsec = attr->ctimensec;
	stat->size = attr->size;
	stat->blocks = attr->blocks;
	stat->blksize = (1 << inode->i_blkbits);
}

792 793
static int fuse_do_getattr(struct inode *inode, struct kstat *stat,
			   struct file *file)
794 795
{
	int err;
796 797
	struct fuse_getattr_in inarg;
	struct fuse_attr_out outarg;
798
	struct fuse_conn *fc = get_fuse_conn(inode);
799 800 801 802
	struct fuse_req *req;
	u64 attr_version;

	req = fuse_get_req(fc);
803 804
	if (IS_ERR(req))
		return PTR_ERR(req);
805

806
	attr_version = fuse_get_attr_version(fc);
807

808
	memset(&inarg, 0, sizeof(inarg));
809
	memset(&outarg, 0, sizeof(outarg));
810 811 812 813 814 815 816
	/* Directories have separate file-handle space */
	if (file && S_ISREG(inode->i_mode)) {
		struct fuse_file *ff = file->private_data;

		inarg.getattr_flags |= FUSE_GETATTR_FH;
		inarg.fh = ff->fh;
	}
817 818
	req->in.h.opcode = FUSE_GETATTR;
	req->in.h.nodeid = get_node_id(inode);
819 820 821
	req->in.numargs = 1;
	req->in.args[0].size = sizeof(inarg);
	req->in.args[0].value = &inarg;
822
	req->out.numargs = 1;
823 824 825 826
	if (fc->minor < 9)
		req->out.args[0].size = FUSE_COMPAT_ATTR_OUT_SIZE;
	else
		req->out.args[0].size = sizeof(outarg);
827
	req->out.args[0].value = &outarg;
828
	fuse_request_send(fc, req);
829 830 831
	err = req->out.h.error;
	fuse_put_request(fc, req);
	if (!err) {
832
		if ((inode->i_mode ^ outarg.attr.mode) & S_IFMT) {
833 834 835
			make_bad_inode(inode);
			err = -EIO;
		} else {
836 837
			fuse_change_attributes(inode, &outarg.attr,
					       attr_timeout(&outarg),
838 839
					       attr_version);
			if (stat)
840
				fuse_fillattr(inode, &outarg.attr, stat);
841 842 843 844 845
		}
	}
	return err;
}

M
Miklos Szeredi 已提交
846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870
int fuse_update_attributes(struct inode *inode, struct kstat *stat,
			   struct file *file, bool *refreshed)
{
	struct fuse_inode *fi = get_fuse_inode(inode);
	int err;
	bool r;

	if (fi->i_time < get_jiffies_64()) {
		r = true;
		err = fuse_do_getattr(inode, stat, file);
	} else {
		r = false;
		err = 0;
		if (stat) {
			generic_fillattr(inode, stat);
			stat->mode = fi->orig_i_mode;
		}
	}

	if (refreshed != NULL)
		*refreshed = r;

	return err;
}

J
John Muir 已提交
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
int fuse_reverse_inval_entry(struct super_block *sb, u64 parent_nodeid,
			     struct qstr *name)
{
	int err = -ENOTDIR;
	struct inode *parent;
	struct dentry *dir;
	struct dentry *entry;

	parent = ilookup5(sb, parent_nodeid, fuse_inode_eq, &parent_nodeid);
	if (!parent)
		return -ENOENT;

	mutex_lock(&parent->i_mutex);
	if (!S_ISDIR(parent->i_mode))
		goto unlock;

	err = -ENOENT;
	dir = d_find_alias(parent);
	if (!dir)
		goto unlock;

	entry = d_lookup(dir, name);
	dput(dir);
	if (!entry)
		goto unlock;

	fuse_invalidate_attr(parent);
	fuse_invalidate_entry(entry);
	dput(entry);
	err = 0;

 unlock:
	mutex_unlock(&parent->i_mutex);
	iput(parent);
	return err;
}

908 909 910 911 912 913 914 915 916 917 918 919 920
/*
 * Calling into a user-controlled filesystem gives the filesystem
 * daemon ptrace-like capabilities over the requester process.  This
 * means, that the filesystem daemon is able to record the exact
 * filesystem operations performed, and can also control the behavior
 * of the requester process in otherwise impossible ways.  For example
 * it can delay the operation for arbitrary length of time allowing
 * DoS against the requester.
 *
 * For this reason only those processes can call into the filesystem,
 * for which the owner of the mount has ptrace privilege.  This
 * excludes processes started by other users, suid or sgid processes.
 */
M
Miklos Szeredi 已提交
921
int fuse_allow_task(struct fuse_conn *fc, struct task_struct *task)
922
{
923 924
	const struct cred *cred;
	int ret;
925

926
	if (fc->flags & FUSE_ALLOW_OTHER)
927 928
		return 1;

929 930 931 932 933 934 935 936 937 938 939 940 941
	rcu_read_lock();
	ret = 0;
	cred = __task_cred(task);
	if (cred->euid == fc->user_id &&
	    cred->suid == fc->user_id &&
	    cred->uid  == fc->user_id &&
	    cred->egid == fc->group_id &&
	    cred->sgid == fc->group_id &&
	    cred->gid  == fc->group_id)
		ret = 1;
	rcu_read_unlock();

	return ret;
942 943
}

M
Miklos Szeredi 已提交
944 945 946 947 948 949 950 951 952 953
static int fuse_access(struct inode *inode, int mask)
{
	struct fuse_conn *fc = get_fuse_conn(inode);
	struct fuse_req *req;
	struct fuse_access_in inarg;
	int err;

	if (fc->no_access)
		return 0;

954 955 956
	req = fuse_get_req(fc);
	if (IS_ERR(req))
		return PTR_ERR(req);
M
Miklos Szeredi 已提交
957 958

	memset(&inarg, 0, sizeof(inarg));
959
	inarg.mask = mask & (MAY_READ | MAY_WRITE | MAY_EXEC);
M
Miklos Szeredi 已提交
960 961 962 963 964
	req->in.h.opcode = FUSE_ACCESS;
	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;
965
	fuse_request_send(fc, req);
M
Miklos Szeredi 已提交
966 967 968 969 970 971 972 973 974
	err = req->out.h.error;
	fuse_put_request(fc, req);
	if (err == -ENOSYS) {
		fc->no_access = 1;
		err = 0;
	}
	return err;
}

975 976 977 978 979 980 981 982 983 984 985 986 987
/*
 * Check permission.  The two basic access models of FUSE are:
 *
 * 1) Local access checking ('default_permissions' mount option) based
 * on file mode.  This is the plain old disk filesystem permission
 * modell.
 *
 * 2) "Remote" access checking, where server is responsible for
 * checking permission in each inode operation.  An exception to this
 * is if ->permission() was invoked from sys_access() in which case an
 * access request is sent.  Execute permission is still checked
 * locally based on file mode.
 */
988
static int fuse_permission(struct inode *inode, int mask)
989 990
{
	struct fuse_conn *fc = get_fuse_conn(inode);
991 992
	bool refreshed = false;
	int err = 0;
993

994
	if (!fuse_allow_task(fc, current))
995
		return -EACCES;
996 997

	/*
998
	 * If attributes are needed, refresh them before proceeding
999
	 */
1000 1001
	if ((fc->flags & FUSE_DEFAULT_PERMISSIONS) ||
	    ((mask & MAY_EXEC) && S_ISREG(inode->i_mode))) {
M
Miklos Szeredi 已提交
1002 1003 1004
		err = fuse_update_attributes(inode, NULL, NULL, &refreshed);
		if (err)
			return err;
1005 1006 1007
	}

	if (fc->flags & FUSE_DEFAULT_PERMISSIONS) {
M
Miklos Szeredi 已提交
1008
		err = generic_permission(inode, mask, NULL);
M
Miklos Szeredi 已提交
1009 1010 1011 1012

		/* If permission is denied, try to refresh file
		   attributes.  This is also needed, because the root
		   node will at first have no permissions */
1013
		if (err == -EACCES && !refreshed) {
1014
			err = fuse_do_getattr(inode, NULL, NULL);
M
Miklos Szeredi 已提交
1015 1016 1017 1018
			if (!err)
				err = generic_permission(inode, mask, NULL);
		}

1019 1020 1021 1022
		/* Note: the opposite of the above test does not
		   exist.  So if permissions are revoked this won't be
		   noticed immediately, only after the attribute
		   timeout has expired */
E
Eric Paris 已提交
1023
	} else if (mask & (MAY_ACCESS | MAY_CHDIR)) {
1024 1025 1026 1027 1028 1029
		err = fuse_access(inode, mask);
	} else if ((mask & MAY_EXEC) && S_ISREG(inode->i_mode)) {
		if (!(inode->i_mode & S_IXUGO)) {
			if (refreshed)
				return -EACCES;

1030
			err = fuse_do_getattr(inode, NULL, NULL);
1031 1032 1033
			if (!err && !(inode->i_mode & S_IXUGO))
				return -EACCES;
		}
1034
	}
1035
	return err;
1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062
}

static int parse_dirfile(char *buf, size_t nbytes, struct file *file,
			 void *dstbuf, filldir_t filldir)
{
	while (nbytes >= FUSE_NAME_OFFSET) {
		struct fuse_dirent *dirent = (struct fuse_dirent *) buf;
		size_t reclen = FUSE_DIRENT_SIZE(dirent);
		int over;
		if (!dirent->namelen || dirent->namelen > FUSE_NAME_MAX)
			return -EIO;
		if (reclen > nbytes)
			break;

		over = filldir(dstbuf, dirent->name, dirent->namelen,
			       file->f_pos, dirent->ino, dirent->type);
		if (over)
			break;

		buf += reclen;
		nbytes -= reclen;
		file->f_pos = dirent->off;
	}

	return 0;
}

1063
static int fuse_readdir(struct file *file, void *dstbuf, filldir_t filldir)
1064
{
1065 1066 1067
	int err;
	size_t nbytes;
	struct page *page;
J
Josef Sipek 已提交
1068
	struct inode *inode = file->f_path.dentry->d_inode;
1069
	struct fuse_conn *fc = get_fuse_conn(inode);
1070 1071 1072 1073 1074
	struct fuse_req *req;

	if (is_bad_inode(inode))
		return -EIO;

1075 1076 1077
	req = fuse_get_req(fc);
	if (IS_ERR(req))
		return PTR_ERR(req);
1078

1079 1080 1081 1082 1083
	page = alloc_page(GFP_KERNEL);
	if (!page) {
		fuse_put_request(fc, req);
		return -ENOMEM;
	}
1084
	req->out.argpages = 1;
1085 1086
	req->num_pages = 1;
	req->pages[0] = page;
1087
	fuse_read_fill(req, file, file->f_pos, PAGE_SIZE, FUSE_READDIR);
1088
	fuse_request_send(fc, req);
1089
	nbytes = req->out.args[0].size;
1090 1091 1092
	err = req->out.h.error;
	fuse_put_request(fc, req);
	if (!err)
1093 1094
		err = parse_dirfile(page_address(page), nbytes, file, dstbuf,
				    filldir);
1095

1096
	__free_page(page);
1097
	fuse_invalidate_attr(inode); /* atime changed */
1098
	return err;
1099 1100 1101 1102 1103 1104
}

static char *read_link(struct dentry *dentry)
{
	struct inode *inode = dentry->d_inode;
	struct fuse_conn *fc = get_fuse_conn(inode);
1105
	struct fuse_req *req = fuse_get_req(fc);
1106 1107
	char *link;

1108
	if (IS_ERR(req))
1109
		return ERR_CAST(req);
1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121

	link = (char *) __get_free_page(GFP_KERNEL);
	if (!link) {
		link = ERR_PTR(-ENOMEM);
		goto out;
	}
	req->in.h.opcode = FUSE_READLINK;
	req->in.h.nodeid = get_node_id(inode);
	req->out.argvar = 1;
	req->out.numargs = 1;
	req->out.args[0].size = PAGE_SIZE - 1;
	req->out.args[0].value = link;
1122
	fuse_request_send(fc, req);
1123 1124 1125 1126 1127 1128 1129
	if (req->out.h.error) {
		free_page((unsigned long) link);
		link = ERR_PTR(req->out.h.error);
	} else
		link[req->out.args[0].size] = '\0';
 out:
	fuse_put_request(fc, req);
1130
	fuse_invalidate_attr(inode); /* atime changed */
1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152
	return link;
}

static void free_link(char *link)
{
	if (!IS_ERR(link))
		free_page((unsigned long) link);
}

static void *fuse_follow_link(struct dentry *dentry, struct nameidata *nd)
{
	nd_set_link(nd, read_link(dentry));
	return NULL;
}

static void fuse_put_link(struct dentry *dentry, struct nameidata *nd, void *c)
{
	free_link(nd_get_link(nd));
}

static int fuse_dir_open(struct inode *inode, struct file *file)
{
1153
	return fuse_open_common(inode, file, true);
1154 1155 1156 1157
}

static int fuse_dir_release(struct inode *inode, struct file *file)
{
1158 1159 1160
	fuse_release_common(file, FUSE_RELEASEDIR);

	return 0;
1161 1162
}

1163
static int fuse_dir_fsync(struct file *file, int datasync)
1164
{
1165
	return fuse_fsync_common(file, datasync, 1);
1166 1167
}

M
Miklos Szeredi 已提交
1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181
static bool update_mtime(unsigned ivalid)
{
	/* Always update if mtime is explicitly set  */
	if (ivalid & ATTR_MTIME_SET)
		return true;

	/* If it's an open(O_TRUNC) or an ftruncate(), don't update */
	if ((ivalid & ATTR_SIZE) && (ivalid & (ATTR_OPEN | ATTR_FILE)))
		return false;

	/* In all other cases update */
	return true;
}

1182
static void iattr_to_fattr(struct iattr *iattr, struct fuse_setattr_in *arg)
1183 1184 1185 1186
{
	unsigned ivalid = iattr->ia_valid;

	if (ivalid & ATTR_MODE)
1187
		arg->valid |= FATTR_MODE,   arg->mode = iattr->ia_mode;
1188
	if (ivalid & ATTR_UID)
1189
		arg->valid |= FATTR_UID,    arg->uid = iattr->ia_uid;
1190
	if (ivalid & ATTR_GID)
1191
		arg->valid |= FATTR_GID,    arg->gid = iattr->ia_gid;
1192
	if (ivalid & ATTR_SIZE)
1193
		arg->valid |= FATTR_SIZE,   arg->size = iattr->ia_size;
M
Miklos Szeredi 已提交
1194 1195
	if (ivalid & ATTR_ATIME) {
		arg->valid |= FATTR_ATIME;
1196
		arg->atime = iattr->ia_atime.tv_sec;
M
Miklos Szeredi 已提交
1197 1198 1199 1200 1201 1202
		arg->atimensec = iattr->ia_atime.tv_nsec;
		if (!(ivalid & ATTR_ATIME_SET))
			arg->valid |= FATTR_ATIME_NOW;
	}
	if ((ivalid & ATTR_MTIME) && update_mtime(ivalid)) {
		arg->valid |= FATTR_MTIME;
1203
		arg->mtime = iattr->ia_mtime.tv_sec;
M
Miklos Szeredi 已提交
1204 1205 1206
		arg->mtimensec = iattr->ia_mtime.tv_nsec;
		if (!(ivalid & ATTR_MTIME_SET))
			arg->valid |= FATTR_MTIME_NOW;
1207
	}
1208 1209
}

M
Miklos Szeredi 已提交
1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253
/*
 * Prevent concurrent writepages on inode
 *
 * This is done by adding a negative bias to the inode write counter
 * and waiting for all pending writes to finish.
 */
void fuse_set_nowrite(struct inode *inode)
{
	struct fuse_conn *fc = get_fuse_conn(inode);
	struct fuse_inode *fi = get_fuse_inode(inode);

	BUG_ON(!mutex_is_locked(&inode->i_mutex));

	spin_lock(&fc->lock);
	BUG_ON(fi->writectr < 0);
	fi->writectr += FUSE_NOWRITE;
	spin_unlock(&fc->lock);
	wait_event(fi->page_waitq, fi->writectr == FUSE_NOWRITE);
}

/*
 * Allow writepages on inode
 *
 * Remove the bias from the writecounter and send any queued
 * writepages.
 */
static void __fuse_release_nowrite(struct inode *inode)
{
	struct fuse_inode *fi = get_fuse_inode(inode);

	BUG_ON(fi->writectr != FUSE_NOWRITE);
	fi->writectr = 0;
	fuse_flush_writepages(inode);
}

void fuse_release_nowrite(struct inode *inode)
{
	struct fuse_conn *fc = get_fuse_conn(inode);

	spin_lock(&fc->lock);
	__fuse_release_nowrite(inode);
	spin_unlock(&fc->lock);
}

1254 1255 1256 1257 1258
/*
 * Set attributes, and at the same time refresh them.
 *
 * Truncation is slightly complicated, because the 'truncate' request
 * may fail, in which case we don't want to touch the mapping.
M
Miklos Szeredi 已提交
1259 1260
 * vmtruncate() doesn't allow for this case, so do the rlimit checking
 * and the actual truncation by hand.
1261
 */
1262 1263
static int fuse_do_setattr(struct dentry *entry, struct iattr *attr,
			   struct file *file)
1264 1265 1266 1267 1268 1269
{
	struct inode *inode = entry->d_inode;
	struct fuse_conn *fc = get_fuse_conn(inode);
	struct fuse_req *req;
	struct fuse_setattr_in inarg;
	struct fuse_attr_out outarg;
M
Miklos Szeredi 已提交
1270 1271
	bool is_truncate = false;
	loff_t oldsize;
1272 1273
	int err;

M
Miklos Szeredi 已提交
1274 1275 1276
	if (!fuse_allow_task(fc, current))
		return -EACCES;

1277 1278 1279 1280 1281 1282
	if (!(fc->flags & FUSE_DEFAULT_PERMISSIONS))
		attr->ia_valid |= ATTR_FORCE;

	err = inode_change_ok(inode, attr);
	if (err)
		return err;
M
Miklos Szeredi 已提交
1283

1284 1285 1286
	if ((attr->ia_valid & ATTR_OPEN) && fc->atomic_o_trunc)
		return 0;

1287
	if (attr->ia_valid & ATTR_SIZE)
M
Miklos Szeredi 已提交
1288
		is_truncate = true;
1289

1290 1291 1292
	req = fuse_get_req(fc);
	if (IS_ERR(req))
		return PTR_ERR(req);
1293

M
Miklos Szeredi 已提交
1294 1295 1296
	if (is_truncate)
		fuse_set_nowrite(inode);

1297
	memset(&inarg, 0, sizeof(inarg));
1298
	memset(&outarg, 0, sizeof(outarg));
1299
	iattr_to_fattr(attr, &inarg);
1300 1301 1302 1303 1304
	if (file) {
		struct fuse_file *ff = file->private_data;
		inarg.valid |= FATTR_FH;
		inarg.fh = ff->fh;
	}
1305 1306 1307 1308 1309
	if (attr->ia_valid & ATTR_SIZE) {
		/* For mandatory locking in truncate */
		inarg.valid |= FATTR_LOCKOWNER;
		inarg.lock_owner = fuse_lock_owner_id(fc, current->files);
	}
1310 1311 1312 1313 1314 1315
	req->in.h.opcode = FUSE_SETATTR;
	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;
	req->out.numargs = 1;
1316 1317 1318 1319
	if (fc->minor < 9)
		req->out.args[0].size = FUSE_COMPAT_ATTR_OUT_SIZE;
	else
		req->out.args[0].size = sizeof(outarg);
1320
	req->out.args[0].value = &outarg;
1321
	fuse_request_send(fc, req);
1322 1323
	err = req->out.h.error;
	fuse_put_request(fc, req);
1324 1325 1326
	if (err) {
		if (err == -EINTR)
			fuse_invalidate_attr(inode);
M
Miklos Szeredi 已提交
1327
		goto error;
1328
	}
1329

1330 1331
	if ((inode->i_mode ^ outarg.attr.mode) & S_IFMT) {
		make_bad_inode(inode);
M
Miklos Szeredi 已提交
1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352
		err = -EIO;
		goto error;
	}

	spin_lock(&fc->lock);
	fuse_change_attributes_common(inode, &outarg.attr,
				      attr_timeout(&outarg));
	oldsize = inode->i_size;
	i_size_write(inode, outarg.attr.size);

	if (is_truncate) {
		/* NOTE: this may release/reacquire fc->lock */
		__fuse_release_nowrite(inode);
	}
	spin_unlock(&fc->lock);

	/*
	 * Only call invalidate_inode_pages2() after removing
	 * FUSE_NOWRITE, otherwise fuse_launder_page() would deadlock.
	 */
	if (S_ISREG(inode->i_mode) && oldsize != outarg.attr.size) {
N
npiggin@suse.de 已提交
1353
		truncate_pagecache(inode, oldsize, outarg.attr.size);
M
Miklos Szeredi 已提交
1354
		invalidate_inode_pages2(inode->i_mapping);
1355 1356 1357
	}

	return 0;
M
Miklos Szeredi 已提交
1358 1359 1360 1361 1362 1363

error:
	if (is_truncate)
		fuse_release_nowrite(inode);

	return err;
1364 1365
}

1366 1367 1368 1369 1370 1371 1372 1373
static int fuse_setattr(struct dentry *entry, struct iattr *attr)
{
	if (attr->ia_valid & ATTR_FILE)
		return fuse_do_setattr(entry, attr, attr->ia_file);
	else
		return fuse_do_setattr(entry, attr, NULL);
}

1374 1375 1376 1377
static int fuse_getattr(struct vfsmount *mnt, struct dentry *entry,
			struct kstat *stat)
{
	struct inode *inode = entry->d_inode;
1378 1379 1380 1381 1382
	struct fuse_conn *fc = get_fuse_conn(inode);

	if (!fuse_allow_task(fc, current))
		return -EACCES;

M
Miklos Szeredi 已提交
1383
	return fuse_update_attributes(inode, stat, NULL, NULL);
1384 1385
}

1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397
static int fuse_setxattr(struct dentry *entry, const char *name,
			 const void *value, size_t size, int flags)
{
	struct inode *inode = entry->d_inode;
	struct fuse_conn *fc = get_fuse_conn(inode);
	struct fuse_req *req;
	struct fuse_setxattr_in inarg;
	int err;

	if (fc->no_setxattr)
		return -EOPNOTSUPP;

1398 1399 1400
	req = fuse_get_req(fc);
	if (IS_ERR(req))
		return PTR_ERR(req);
1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413

	memset(&inarg, 0, sizeof(inarg));
	inarg.size = size;
	inarg.flags = flags;
	req->in.h.opcode = FUSE_SETXATTR;
	req->in.h.nodeid = get_node_id(inode);
	req->in.numargs = 3;
	req->in.args[0].size = sizeof(inarg);
	req->in.args[0].value = &inarg;
	req->in.args[1].size = strlen(name) + 1;
	req->in.args[1].value = name;
	req->in.args[2].size = size;
	req->in.args[2].value = value;
1414
	fuse_request_send(fc, req);
1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436
	err = req->out.h.error;
	fuse_put_request(fc, req);
	if (err == -ENOSYS) {
		fc->no_setxattr = 1;
		err = -EOPNOTSUPP;
	}
	return err;
}

static ssize_t fuse_getxattr(struct dentry *entry, const char *name,
			     void *value, size_t size)
{
	struct inode *inode = entry->d_inode;
	struct fuse_conn *fc = get_fuse_conn(inode);
	struct fuse_req *req;
	struct fuse_getxattr_in inarg;
	struct fuse_getxattr_out outarg;
	ssize_t ret;

	if (fc->no_getxattr)
		return -EOPNOTSUPP;

1437 1438 1439
	req = fuse_get_req(fc);
	if (IS_ERR(req))
		return PTR_ERR(req);
1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459

	memset(&inarg, 0, sizeof(inarg));
	inarg.size = size;
	req->in.h.opcode = FUSE_GETXATTR;
	req->in.h.nodeid = get_node_id(inode);
	req->in.numargs = 2;
	req->in.args[0].size = sizeof(inarg);
	req->in.args[0].value = &inarg;
	req->in.args[1].size = strlen(name) + 1;
	req->in.args[1].value = name;
	/* This is really two different operations rolled into one */
	req->out.numargs = 1;
	if (size) {
		req->out.argvar = 1;
		req->out.args[0].size = size;
		req->out.args[0].value = value;
	} else {
		req->out.args[0].size = sizeof(outarg);
		req->out.args[0].value = &outarg;
	}
1460
	fuse_request_send(fc, req);
1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482
	ret = req->out.h.error;
	if (!ret)
		ret = size ? req->out.args[0].size : outarg.size;
	else {
		if (ret == -ENOSYS) {
			fc->no_getxattr = 1;
			ret = -EOPNOTSUPP;
		}
	}
	fuse_put_request(fc, req);
	return ret;
}

static ssize_t fuse_listxattr(struct dentry *entry, char *list, size_t size)
{
	struct inode *inode = entry->d_inode;
	struct fuse_conn *fc = get_fuse_conn(inode);
	struct fuse_req *req;
	struct fuse_getxattr_in inarg;
	struct fuse_getxattr_out outarg;
	ssize_t ret;

M
Miklos Szeredi 已提交
1483 1484 1485
	if (!fuse_allow_task(fc, current))
		return -EACCES;

1486 1487 1488
	if (fc->no_listxattr)
		return -EOPNOTSUPP;

1489 1490 1491
	req = fuse_get_req(fc);
	if (IS_ERR(req))
		return PTR_ERR(req);
1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509

	memset(&inarg, 0, sizeof(inarg));
	inarg.size = size;
	req->in.h.opcode = FUSE_LISTXATTR;
	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;
	/* This is really two different operations rolled into one */
	req->out.numargs = 1;
	if (size) {
		req->out.argvar = 1;
		req->out.args[0].size = size;
		req->out.args[0].value = list;
	} else {
		req->out.args[0].size = sizeof(outarg);
		req->out.args[0].value = &outarg;
	}
1510
	fuse_request_send(fc, req);
1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533
	ret = req->out.h.error;
	if (!ret)
		ret = size ? req->out.args[0].size : outarg.size;
	else {
		if (ret == -ENOSYS) {
			fc->no_listxattr = 1;
			ret = -EOPNOTSUPP;
		}
	}
	fuse_put_request(fc, req);
	return ret;
}

static int fuse_removexattr(struct dentry *entry, const char *name)
{
	struct inode *inode = entry->d_inode;
	struct fuse_conn *fc = get_fuse_conn(inode);
	struct fuse_req *req;
	int err;

	if (fc->no_removexattr)
		return -EOPNOTSUPP;

1534 1535 1536
	req = fuse_get_req(fc);
	if (IS_ERR(req))
		return PTR_ERR(req);
1537 1538 1539 1540 1541 1542

	req->in.h.opcode = FUSE_REMOVEXATTR;
	req->in.h.nodeid = get_node_id(inode);
	req->in.numargs = 1;
	req->in.args[0].size = strlen(name) + 1;
	req->in.args[0].value = name;
1543
	fuse_request_send(fc, req);
1544 1545 1546 1547 1548 1549 1550 1551 1552
	err = req->out.h.error;
	fuse_put_request(fc, req);
	if (err == -ENOSYS) {
		fc->no_removexattr = 1;
		err = -EOPNOTSUPP;
	}
	return err;
}

1553
static const struct inode_operations fuse_dir_inode_operations = {
1554
	.lookup		= fuse_lookup,
1555 1556 1557 1558 1559 1560 1561 1562 1563
	.mkdir		= fuse_mkdir,
	.symlink	= fuse_symlink,
	.unlink		= fuse_unlink,
	.rmdir		= fuse_rmdir,
	.rename		= fuse_rename,
	.link		= fuse_link,
	.setattr	= fuse_setattr,
	.create		= fuse_create,
	.mknod		= fuse_mknod,
1564 1565
	.permission	= fuse_permission,
	.getattr	= fuse_getattr,
1566 1567 1568 1569
	.setxattr	= fuse_setxattr,
	.getxattr	= fuse_getxattr,
	.listxattr	= fuse_listxattr,
	.removexattr	= fuse_removexattr,
1570 1571
};

1572
static const struct file_operations fuse_dir_operations = {
M
Miklos Szeredi 已提交
1573
	.llseek		= generic_file_llseek,
1574 1575 1576 1577
	.read		= generic_read_dir,
	.readdir	= fuse_readdir,
	.open		= fuse_dir_open,
	.release	= fuse_dir_release,
1578
	.fsync		= fuse_dir_fsync,
1579 1580
};

1581
static const struct inode_operations fuse_common_inode_operations = {
1582
	.setattr	= fuse_setattr,
1583 1584
	.permission	= fuse_permission,
	.getattr	= fuse_getattr,
1585 1586 1587 1588
	.setxattr	= fuse_setxattr,
	.getxattr	= fuse_getxattr,
	.listxattr	= fuse_listxattr,
	.removexattr	= fuse_removexattr,
1589 1590
};

1591
static const struct inode_operations fuse_symlink_inode_operations = {
1592
	.setattr	= fuse_setattr,
1593 1594 1595 1596
	.follow_link	= fuse_follow_link,
	.put_link	= fuse_put_link,
	.readlink	= generic_readlink,
	.getattr	= fuse_getattr,
1597 1598 1599 1600
	.setxattr	= fuse_setxattr,
	.getxattr	= fuse_getxattr,
	.listxattr	= fuse_listxattr,
	.removexattr	= fuse_removexattr,
1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617
};

void fuse_init_common(struct inode *inode)
{
	inode->i_op = &fuse_common_inode_operations;
}

void fuse_init_dir(struct inode *inode)
{
	inode->i_op = &fuse_dir_inode_operations;
	inode->i_fop = &fuse_dir_operations;
}

void fuse_init_symlink(struct inode *inode)
{
	inode->i_op = &fuse_symlink_inode_operations;
}