file.c 21.4 KB
Newer Older
M
Miklos Szeredi 已提交
1 2
/*
  FUSE: Filesystem in Userspace
3
  Copyright (C) 2001-2006  Miklos Szeredi <miklos@szeredi.hu>
M
Miklos Szeredi 已提交
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>
A
Alexey Dobriyan 已提交
14
#include <linux/sched.h>
M
Miklos Szeredi 已提交
15

16
static const struct file_operations fuse_direct_io_file_operations;
17

M
Miklos Szeredi 已提交
18 19
static int fuse_send_open(struct inode *inode, struct file *file, int isdir,
			  struct fuse_open_out *outargp)
M
Miklos Szeredi 已提交
20 21 22
{
	struct fuse_conn *fc = get_fuse_conn(inode);
	struct fuse_open_in inarg;
M
Miklos Szeredi 已提交
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);
M
Miklos Szeredi 已提交
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;
M
Miklos Szeredi 已提交
34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53
	req->in.h.opcode = isdir ? FUSE_OPENDIR : FUSE_OPEN;
	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;
	req->out.args[0].size = sizeof(*outargp);
	req->out.args[0].value = outargp;
	request_send(fc, req);
	err = req->out.h.error;
	fuse_put_request(fc, req);

	return err;
}

struct fuse_file *fuse_file_alloc(void)
{
	struct fuse_file *ff;
	ff = kmalloc(sizeof(struct fuse_file), GFP_KERNEL);
	if (ff) {
54 55
		ff->reserved_req = fuse_request_alloc();
		if (!ff->reserved_req) {
M
Miklos Szeredi 已提交
56 57 58
			kfree(ff);
			ff = NULL;
		}
59
		INIT_LIST_HEAD(&ff->write_entry);
60
		atomic_set(&ff->count, 0);
M
Miklos Szeredi 已提交
61 62 63 64 65 66
	}
	return ff;
}

void fuse_file_free(struct fuse_file *ff)
{
67
	fuse_request_free(ff->reserved_req);
M
Miklos Szeredi 已提交
68 69 70
	kfree(ff);
}

71 72 73 74 75 76
static struct fuse_file *fuse_file_get(struct fuse_file *ff)
{
	atomic_inc(&ff->count);
	return ff;
}

M
Miklos Szeredi 已提交
77 78 79 80 81 82 83
static void fuse_release_end(struct fuse_conn *fc, struct fuse_req *req)
{
	dput(req->dentry);
	mntput(req->vfsmount);
	fuse_put_request(fc, req);
}

84 85 86 87 88
static void fuse_file_put(struct fuse_file *ff)
{
	if (atomic_dec_and_test(&ff->count)) {
		struct fuse_req *req = ff->reserved_req;
		struct fuse_conn *fc = get_fuse_conn(req->dentry->d_inode);
M
Miklos Szeredi 已提交
89
		req->end = fuse_release_end;
90 91 92 93 94
		request_send_background(fc, req);
		kfree(ff);
	}
}

M
Miklos Szeredi 已提交
95 96 97 98 99 100
void fuse_finish_open(struct inode *inode, struct file *file,
		      struct fuse_file *ff, struct fuse_open_out *outarg)
{
	if (outarg->open_flags & FOPEN_DIRECT_IO)
		file->f_op = &fuse_direct_io_file_operations;
	if (!(outarg->open_flags & FOPEN_KEEP_CACHE))
M
Miklos Szeredi 已提交
101
		invalidate_inode_pages2(inode->i_mapping);
M
Miklos Szeredi 已提交
102
	ff->fh = outarg->fh;
103
	file->private_data = fuse_file_get(ff);
M
Miklos Szeredi 已提交
104 105 106 107
}

int fuse_open_common(struct inode *inode, struct file *file, int isdir)
{
M
Miklos Szeredi 已提交
108 109 110 111
	struct fuse_open_out outarg;
	struct fuse_file *ff;
	int err;

M
Miklos Szeredi 已提交
112 113 114 115
	/* VFS checks this, but only _after_ ->open() */
	if (file->f_flags & O_DIRECT)
		return -EINVAL;

M
Miklos Szeredi 已提交
116 117 118 119
	err = generic_file_open(inode, file);
	if (err)
		return err;

M
Miklos Szeredi 已提交
120
	ff = fuse_file_alloc();
M
Miklos Szeredi 已提交
121
	if (!ff)
M
Miklos Szeredi 已提交
122
		return -ENOMEM;
M
Miklos Szeredi 已提交
123

M
Miklos Szeredi 已提交
124 125 126 127 128 129 130
	err = fuse_send_open(inode, file, isdir, &outarg);
	if (err)
		fuse_file_free(ff);
	else {
		if (isdir)
			outarg.open_flags &= ~FOPEN_DIRECT_IO;
		fuse_finish_open(inode, file, ff, &outarg);
M
Miklos Szeredi 已提交
131 132 133 134 135
	}

	return err;
}

136
void fuse_release_fill(struct fuse_file *ff, u64 nodeid, int flags, int opcode)
137
{
138
	struct fuse_req *req = ff->reserved_req;
M
Miklos Szeredi 已提交
139 140 141
	struct fuse_release_in *inarg = &req->misc.release_in;

	inarg->fh = ff->fh;
M
Miklos Szeredi 已提交
142
	inarg->flags = flags;
143
	req->in.h.opcode = opcode;
M
Miklos Szeredi 已提交
144
	req->in.h.nodeid = nodeid;
M
Miklos Szeredi 已提交
145 146 147
	req->in.numargs = 1;
	req->in.args[0].size = sizeof(struct fuse_release_in);
	req->in.args[0].value = inarg;
M
Miklos Szeredi 已提交
148 149 150 151 152 153
}

int fuse_release_common(struct inode *inode, struct file *file, int isdir)
{
	struct fuse_file *ff = file->private_data;
	if (ff) {
154 155
		struct fuse_conn *fc = get_fuse_conn(inode);

156 157
		fuse_release_fill(ff, get_node_id(inode), file->f_flags,
				  isdir ? FUSE_RELEASEDIR : FUSE_RELEASE);
158 159

		/* Hold vfsmount and dentry until release is finished */
160 161
		ff->reserved_req->vfsmount = mntget(file->f_path.mnt);
		ff->reserved_req->dentry = dget(file->f_path.dentry);
162 163 164 165

		spin_lock(&fc->lock);
		list_del(&ff->write_entry);
		spin_unlock(&fc->lock);
166 167 168 169 170 171
		/*
		 * 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);
M
Miklos Szeredi 已提交
172
	}
M
Miklos Szeredi 已提交
173 174 175 176 177

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

178 179 180 181 182 183 184 185 186 187
static int fuse_open(struct inode *inode, struct file *file)
{
	return fuse_open_common(inode, file, 0);
}

static int fuse_release(struct inode *inode, struct file *file)
{
	return fuse_release_common(inode, file, 0);
}

188
/*
189 190
 * Scramble the ID space with XTEA, so that the value of the files_struct
 * pointer is not exposed to userspace.
191
 */
192
static u64 fuse_lock_owner_id(struct fuse_conn *fc, fl_owner_t id)
193
{
194 195 196 197 198 199 200 201 202 203 204 205 206 207
	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);
208 209
}

210
static int fuse_flush(struct file *file, fl_owner_t id)
M
Miklos Szeredi 已提交
211
{
J
Josef Sipek 已提交
212
	struct inode *inode = file->f_path.dentry->d_inode;
M
Miklos Szeredi 已提交
213 214 215 216 217 218
	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;

219 220 221
	if (is_bad_inode(inode))
		return -EIO;

M
Miklos Szeredi 已提交
222 223 224
	if (fc->no_flush)
		return 0;

225
	req = fuse_get_req_nofail(fc, file);
M
Miklos Szeredi 已提交
226 227
	memset(&inarg, 0, sizeof(inarg));
	inarg.fh = ff->fh;
228
	inarg.lock_owner = fuse_lock_owner_id(fc, id);
M
Miklos Szeredi 已提交
229 230 231 232 233
	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;
234
	req->force = 1;
235
	request_send(fc, req);
M
Miklos Szeredi 已提交
236 237 238 239 240 241 242 243 244
	err = req->out.h.error;
	fuse_put_request(fc, req);
	if (err == -ENOSYS) {
		fc->no_flush = 1;
		err = 0;
	}
	return err;
}

245 246
int fuse_fsync_common(struct file *file, struct dentry *de, int datasync,
		      int isdir)
M
Miklos Szeredi 已提交
247 248 249 250 251 252 253 254
{
	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;

255 256 257
	if (is_bad_inode(inode))
		return -EIO;

258
	if ((!isdir && fc->no_fsync) || (isdir && fc->no_fsyncdir))
M
Miklos Szeredi 已提交
259 260
		return 0;

261 262 263
	req = fuse_get_req(fc);
	if (IS_ERR(req))
		return PTR_ERR(req);
M
Miklos Szeredi 已提交
264 265 266 267

	memset(&inarg, 0, sizeof(inarg));
	inarg.fh = ff->fh;
	inarg.fsync_flags = datasync ? 1 : 0;
268
	req->in.h.opcode = isdir ? FUSE_FSYNCDIR : FUSE_FSYNC;
M
Miklos Szeredi 已提交
269 270 271 272 273 274 275 276
	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;
	request_send(fc, req);
	err = req->out.h.error;
	fuse_put_request(fc, req);
	if (err == -ENOSYS) {
277 278 279 280
		if (isdir)
			fc->no_fsyncdir = 1;
		else
			fc->no_fsync = 1;
M
Miklos Szeredi 已提交
281 282 283 284 285
		err = 0;
	}
	return err;
}

286 287 288 289 290
static int fuse_fsync(struct file *file, struct dentry *de, int datasync)
{
	return fuse_fsync_common(file, de, datasync, 0);
}

291
void fuse_read_fill(struct fuse_req *req, struct fuse_file *ff,
292
		    struct inode *inode, loff_t pos, size_t count, int opcode)
M
Miklos Szeredi 已提交
293
{
294
	struct fuse_read_in *inarg = &req->misc.read_in;
M
Miklos Szeredi 已提交
295

296 297 298 299
	inarg->fh = ff->fh;
	inarg->offset = pos;
	inarg->size = count;
	req->in.h.opcode = opcode;
M
Miklos Szeredi 已提交
300 301 302
	req->in.h.nodeid = get_node_id(inode);
	req->in.numargs = 1;
	req->in.args[0].size = sizeof(struct fuse_read_in);
303
	req->in.args[0].value = inarg;
M
Miklos Szeredi 已提交
304 305 306 307 308 309
	req->out.argpages = 1;
	req->out.argvar = 1;
	req->out.numargs = 1;
	req->out.args[0].size = count;
}

310 311
static size_t fuse_send_read(struct fuse_req *req, struct file *file,
			     struct inode *inode, loff_t pos, size_t count)
312
{
313
	struct fuse_conn *fc = get_fuse_conn(inode);
314 315
	struct fuse_file *ff = file->private_data;
	fuse_read_fill(req, ff, inode, pos, count, FUSE_READ);
316 317
	request_send(fc, req);
	return req->out.args[0].size;
318 319
}

M
Miklos Szeredi 已提交
320 321 322 323
static int fuse_readpage(struct file *file, struct page *page)
{
	struct inode *inode = page->mapping->host;
	struct fuse_conn *fc = get_fuse_conn(inode);
324 325 326 327 328 329 330
	struct fuse_req *req;
	int err;

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

331 332 333
	req = fuse_get_req(fc);
	err = PTR_ERR(req);
	if (IS_ERR(req))
M
Miklos Szeredi 已提交
334 335 336 337 338
		goto out;

	req->out.page_zeroing = 1;
	req->num_pages = 1;
	req->pages[0] = page;
339
	fuse_send_read(req, file, inode, page_offset(page), PAGE_CACHE_SIZE);
M
Miklos Szeredi 已提交
340 341 342 343
	err = req->out.h.error;
	fuse_put_request(fc, req);
	if (!err)
		SetPageUptodate(page);
344
	fuse_invalidate_attr(inode); /* atime changed */
M
Miklos Szeredi 已提交
345 346 347 348 349
 out:
	unlock_page(page);
	return err;
}

350
static void fuse_readpages_end(struct fuse_conn *fc, struct fuse_req *req)
351
{
352 353 354 355
	int i;

	fuse_invalidate_attr(req->pages[0]->mapping->host); /* atime changed */

356 357 358 359
	for (i = 0; i < req->num_pages; i++) {
		struct page *page = req->pages[i];
		if (!req->out.h.error)
			SetPageUptodate(page);
360 361
		else
			SetPageError(page);
362 363
		unlock_page(page);
	}
364 365
	if (req->ff)
		fuse_file_put(req->ff);
366 367 368
	fuse_put_request(fc, req);
}

369
static void fuse_send_readpages(struct fuse_req *req, struct fuse_file *ff,
370 371 372 373 374 375
				struct inode *inode)
{
	struct fuse_conn *fc = get_fuse_conn(inode);
	loff_t pos = page_offset(req->pages[0]);
	size_t count = req->num_pages << PAGE_CACHE_SHIFT;
	req->out.page_zeroing = 1;
376
	fuse_read_fill(req, ff, inode, pos, count, FUSE_READ);
377
	if (fc->async_read) {
378
		req->ff = fuse_file_get(ff);
379 380 381 382 383 384
		req->end = fuse_readpages_end;
		request_send_background(fc, req);
	} else {
		request_send(fc, req);
		fuse_readpages_end(fc, req);
	}
385 386
}

387
struct fuse_fill_data {
388
	struct fuse_req *req;
389
	struct fuse_file *ff;
390 391 392 393 394
	struct inode *inode;
};

static int fuse_readpages_fill(void *_data, struct page *page)
{
395
	struct fuse_fill_data *data = _data;
396 397 398 399 400 401 402 403
	struct fuse_req *req = data->req;
	struct inode *inode = data->inode;
	struct fuse_conn *fc = get_fuse_conn(inode);

	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)) {
404
		fuse_send_readpages(req, data->ff, inode);
405 406
		data->req = req = fuse_get_req(fc);
		if (IS_ERR(req)) {
407
			unlock_page(page);
408
			return PTR_ERR(req);
409 410 411 412 413 414 415 416 417 418 419 420
		}
	}
	req->pages[req->num_pages] = page;
	req->num_pages ++;
	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);
421
	struct fuse_fill_data data;
422
	int err;
423

424
	err = -EIO;
425
	if (is_bad_inode(inode))
426
		goto out;
427

428
	data.ff = file->private_data;
429
	data.inode = inode;
430
	data.req = fuse_get_req(fc);
431
	err = PTR_ERR(data.req);
432
	if (IS_ERR(data.req))
433
		goto out;
434 435

	err = read_cache_pages(mapping, pages, fuse_readpages_fill, &data);
436 437
	if (!err) {
		if (data.req->num_pages)
438
			fuse_send_readpages(data.req, data.ff, inode);
439 440 441
		else
			fuse_put_request(fc, data.req);
	}
442
out:
443
	return err;
444 445
}

446 447
static size_t fuse_send_write(struct fuse_req *req, struct file *file,
			      struct inode *inode, loff_t pos, size_t count)
M
Miklos Szeredi 已提交
448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467
{
	struct fuse_conn *fc = get_fuse_conn(inode);
	struct fuse_file *ff = file->private_data;
	struct fuse_write_in inarg;
	struct fuse_write_out outarg;

	memset(&inarg, 0, sizeof(struct fuse_write_in));
	inarg.fh = ff->fh;
	inarg.offset = pos;
	inarg.size = count;
	req->in.h.opcode = FUSE_WRITE;
	req->in.h.nodeid = get_node_id(inode);
	req->in.argpages = 1;
	req->in.numargs = 2;
	req->in.args[0].size = sizeof(struct fuse_write_in);
	req->in.args[0].value = &inarg;
	req->in.args[1].size = count;
	req->out.numargs = 1;
	req->out.args[0].size = sizeof(struct fuse_write_out);
	req->out.args[0].value = &outarg;
468
	request_send(fc, req);
M
Miklos Szeredi 已提交
469 470 471
	return outarg.size;
}

N
Nick Piggin 已提交
472 473 474
static int fuse_write_begin(struct file *file, struct address_space *mapping,
			loff_t pos, unsigned len, unsigned flags,
			struct page **pagep, void **fsdata)
M
Miklos Szeredi 已提交
475
{
N
Nick Piggin 已提交
476 477 478 479 480
	pgoff_t index = pos >> PAGE_CACHE_SHIFT;

	*pagep = __grab_cache_page(mapping, index);
	if (!*pagep)
		return -ENOMEM;
M
Miklos Szeredi 已提交
481 482 483
	return 0;
}

N
Nick Piggin 已提交
484 485
static int fuse_buffered_write(struct file *file, struct inode *inode,
			       loff_t pos, unsigned count, struct page *page)
M
Miklos Szeredi 已提交
486 487
{
	int err;
488
	size_t nres;
M
Miklos Szeredi 已提交
489
	struct fuse_conn *fc = get_fuse_conn(inode);
490
	struct fuse_inode *fi = get_fuse_inode(inode);
N
Nick Piggin 已提交
491
	unsigned offset = pos & (PAGE_CACHE_SIZE - 1);
492 493 494 495 496
	struct fuse_req *req;

	if (is_bad_inode(inode))
		return -EIO;

497 498 499
	req = fuse_get_req(fc);
	if (IS_ERR(req))
		return PTR_ERR(req);
M
Miklos Szeredi 已提交
500 501 502 503 504 505 506

	req->num_pages = 1;
	req->pages[0] = page;
	req->page_offset = offset;
	nres = fuse_send_write(req, file, inode, pos, count);
	err = req->out.h.error;
	fuse_put_request(fc, req);
N
Nick Piggin 已提交
507
	if (!err && !nres)
M
Miklos Szeredi 已提交
508 509
		err = -EIO;
	if (!err) {
N
Nick Piggin 已提交
510
		pos += nres;
M
Miklos Szeredi 已提交
511
		spin_lock(&fc->lock);
512
		fi->attr_version = ++fc->attr_version;
M
Miklos Szeredi 已提交
513
		if (pos > inode->i_size)
M
Miklos Szeredi 已提交
514
			i_size_write(inode, pos);
M
Miklos Szeredi 已提交
515
		spin_unlock(&fc->lock);
M
Miklos Szeredi 已提交
516

N
Nick Piggin 已提交
517
		if (count == PAGE_CACHE_SIZE)
M
Miklos Szeredi 已提交
518
			SetPageUptodate(page);
519 520
	}
	fuse_invalidate_attr(inode);
N
Nick Piggin 已提交
521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536
	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;
M
Miklos Szeredi 已提交
537 538
}

M
Miklos Szeredi 已提交
539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563
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,
			       unsigned nbytes, int write)
{
	unsigned long user_addr = (unsigned long) buf;
	unsigned offset = user_addr & ~PAGE_MASK;
	int npages;

	/* This doesn't work with nfsd */
	if (!current->mm)
		return -EPERM;

	nbytes = min(nbytes, (unsigned) FUSE_MAX_PAGES_PER_REQ << PAGE_SHIFT);
	npages = (nbytes + offset + PAGE_SIZE - 1) >> PAGE_SHIFT;
564
	npages = min(max(npages, 1), FUSE_MAX_PAGES_PER_REQ);
M
Miklos Szeredi 已提交
565 566 567 568 569 570 571 572 573 574 575 576 577 578 579
	down_read(&current->mm->mmap_sem);
	npages = get_user_pages(current, current->mm, user_addr, npages, write,
				0, req->pages, NULL);
	up_read(&current->mm->mmap_sem);
	if (npages < 0)
		return npages;

	req->num_pages = npages;
	req->page_offset = offset;
	return 0;
}

static ssize_t fuse_direct_io(struct file *file, const char __user *buf,
			      size_t count, loff_t *ppos, int write)
{
J
Josef Sipek 已提交
580
	struct inode *inode = file->f_path.dentry->d_inode;
M
Miklos Szeredi 已提交
581 582 583 584
	struct fuse_conn *fc = get_fuse_conn(inode);
	size_t nmax = write ? fc->max_write : fc->max_read;
	loff_t pos = *ppos;
	ssize_t res = 0;
585 586 587 588 589
	struct fuse_req *req;

	if (is_bad_inode(inode))
		return -EIO;

590 591 592
	req = fuse_get_req(fc);
	if (IS_ERR(req))
		return PTR_ERR(req);
M
Miklos Szeredi 已提交
593 594 595 596 597 598 599 600 601

	while (count) {
		size_t nres;
		size_t nbytes = min(count, nmax);
		int err = fuse_get_user_pages(req, buf, nbytes, !write);
		if (err) {
			res = err;
			break;
		}
602 603
		nbytes = (req->num_pages << PAGE_SHIFT) - req->page_offset;
		nbytes = min(count, nbytes);
M
Miklos Szeredi 已提交
604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622
		if (write)
			nres = fuse_send_write(req, file, inode, pos, nbytes);
		else
			nres = fuse_send_read(req, file, inode, pos, nbytes);
		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;
623 624 625 626 627 628
		if (count) {
			fuse_put_request(fc, req);
			req = fuse_get_req(fc);
			if (IS_ERR(req))
				break;
		}
M
Miklos Szeredi 已提交
629 630 631
	}
	fuse_put_request(fc, req);
	if (res > 0) {
M
Miklos Szeredi 已提交
632 633 634 635 636 637
		if (write) {
			spin_lock(&fc->lock);
			if (pos > inode->i_size)
				i_size_write(inode, pos);
			spin_unlock(&fc->lock);
		}
M
Miklos Szeredi 已提交
638
		*ppos = pos;
639 640
	}
	fuse_invalidate_attr(inode);
M
Miklos Szeredi 已提交
641 642 643 644 645 646 647 648 649 650 651 652 653

	return res;
}

static ssize_t fuse_direct_read(struct file *file, char __user *buf,
				     size_t count, loff_t *ppos)
{
	return fuse_direct_io(file, buf, count, ppos, 0);
}

static ssize_t fuse_direct_write(struct file *file, const char __user *buf,
				 size_t count, loff_t *ppos)
{
J
Josef Sipek 已提交
654
	struct inode *inode = file->f_path.dentry->d_inode;
M
Miklos Szeredi 已提交
655 656
	ssize_t res;
	/* Don't allow parallel writes to the same file */
657
	mutex_lock(&inode->i_mutex);
658 659 660
	res = generic_write_checks(file, ppos, &count, 0);
	if (!res)
		res = fuse_direct_io(file, buf, count, ppos, 1);
661
	mutex_unlock(&inode->i_mutex);
M
Miklos Szeredi 已提交
662 663 664
	return res;
}

M
Miklos Szeredi 已提交
665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682
static int fuse_file_mmap(struct file *file, struct vm_area_struct *vma)
{
	if ((vma->vm_flags & VM_SHARED)) {
		if ((vma->vm_flags & VM_WRITE))
			return -ENODEV;
		else
			vma->vm_flags &= ~VM_MAYWRITE;
	}
	return generic_file_mmap(file, vma);
}

static int fuse_set_page_dirty(struct page *page)
{
	printk("fuse_set_page_dirty: should not happen\n");
	dump_stack();
	return 0;
}

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
static int convert_fuse_file_lock(const struct fuse_file_lock *ffl,
				  struct file_lock *fl)
{
	switch (ffl->type) {
	case F_UNLCK:
		break;

	case F_RDLCK:
	case F_WRLCK:
		if (ffl->start > OFFSET_MAX || ffl->end > OFFSET_MAX ||
		    ffl->end < ffl->start)
			return -EIO;

		fl->fl_start = ffl->start;
		fl->fl_end = ffl->end;
		fl->fl_pid = ffl->pid;
		break;

	default:
		return -EIO;
	}
	fl->fl_type = ffl->type;
	return 0;
}

static void fuse_lk_fill(struct fuse_req *req, struct file *file,
709 710
			 const struct file_lock *fl, int opcode, pid_t pid,
			 int flock)
711
{
J
Josef Sipek 已提交
712
	struct inode *inode = file->f_path.dentry->d_inode;
713
	struct fuse_conn *fc = get_fuse_conn(inode);
714 715 716 717
	struct fuse_file *ff = file->private_data;
	struct fuse_lk_in *arg = &req->misc.lk_in;

	arg->fh = ff->fh;
718
	arg->owner = fuse_lock_owner_id(fc, fl->fl_owner);
719 720 721 722
	arg->lk.start = fl->fl_start;
	arg->lk.end = fl->fl_end;
	arg->lk.type = fl->fl_type;
	arg->lk.pid = pid;
723 724
	if (flock)
		arg->lk_flags |= FUSE_LK_FLOCK;
725 726 727 728 729 730 731 732 733
	req->in.h.opcode = opcode;
	req->in.h.nodeid = get_node_id(inode);
	req->in.numargs = 1;
	req->in.args[0].size = sizeof(*arg);
	req->in.args[0].value = arg;
}

static int fuse_getlk(struct file *file, struct file_lock *fl)
{
J
Josef Sipek 已提交
734
	struct inode *inode = file->f_path.dentry->d_inode;
735 736 737 738 739 740 741 742 743
	struct fuse_conn *fc = get_fuse_conn(inode);
	struct fuse_req *req;
	struct fuse_lk_out outarg;
	int err;

	req = fuse_get_req(fc);
	if (IS_ERR(req))
		return PTR_ERR(req);

744
	fuse_lk_fill(req, file, fl, FUSE_GETLK, 0, 0);
745 746 747 748 749 750 751 752 753 754 755 756
	req->out.numargs = 1;
	req->out.args[0].size = sizeof(outarg);
	req->out.args[0].value = &outarg;
	request_send(fc, req);
	err = req->out.h.error;
	fuse_put_request(fc, req);
	if (!err)
		err = convert_fuse_file_lock(&outarg.lk, fl);

	return err;
}

757
static int fuse_setlk(struct file *file, struct file_lock *fl, int flock)
758
{
J
Josef Sipek 已提交
759
	struct inode *inode = file->f_path.dentry->d_inode;
760 761 762 763 764 765 766 767 768 769 770 771 772 773
	struct fuse_conn *fc = get_fuse_conn(inode);
	struct fuse_req *req;
	int opcode = (fl->fl_flags & FL_SLEEP) ? FUSE_SETLKW : FUSE_SETLK;
	pid_t pid = fl->fl_type != F_UNLCK ? current->tgid : 0;
	int err;

	/* Unlock on close is handled by the flush method */
	if (fl->fl_flags & FL_CLOSE)
		return 0;

	req = fuse_get_req(fc);
	if (IS_ERR(req))
		return PTR_ERR(req);

774
	fuse_lk_fill(req, file, fl, opcode, pid, flock);
775 776
	request_send(fc, req);
	err = req->out.h.error;
777 778 779
	/* locking is restartable */
	if (err == -EINTR)
		err = -ERESTARTSYS;
780 781 782 783 784 785
	fuse_put_request(fc, req);
	return err;
}

static int fuse_file_lock(struct file *file, int cmd, struct file_lock *fl)
{
J
Josef Sipek 已提交
786
	struct inode *inode = file->f_path.dentry->d_inode;
787 788 789 790 791
	struct fuse_conn *fc = get_fuse_conn(inode);
	int err;

	if (cmd == F_GETLK) {
		if (fc->no_lock) {
792
			posix_test_lock(file, fl);
793 794 795 796 797 798 799
			err = 0;
		} else
			err = fuse_getlk(file, fl);
	} else {
		if (fc->no_lock)
			err = posix_lock_file_wait(file, fl);
		else
800
			err = fuse_setlk(file, fl, 0);
801 802 803 804
	}
	return err;
}

805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821
static int fuse_file_flock(struct file *file, int cmd, struct file_lock *fl)
{
	struct inode *inode = file->f_path.dentry->d_inode;
	struct fuse_conn *fc = get_fuse_conn(inode);
	int err;

	if (fc->no_lock) {
		err = flock_lock_file_wait(file, fl);
	} else {
		/* emulate flock with POSIX locks */
		fl->fl_owner = (fl_owner_t) file;
		err = fuse_setlk(file, fl, 1);
	}

	return err;
}

M
Miklos Szeredi 已提交
822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857
static sector_t fuse_bmap(struct address_space *mapping, sector_t block)
{
	struct inode *inode = mapping->host;
	struct fuse_conn *fc = get_fuse_conn(inode);
	struct fuse_req *req;
	struct fuse_bmap_in inarg;
	struct fuse_bmap_out outarg;
	int err;

	if (!inode->i_sb->s_bdev || fc->no_bmap)
		return 0;

	req = fuse_get_req(fc);
	if (IS_ERR(req))
		return 0;

	memset(&inarg, 0, sizeof(inarg));
	inarg.block = block;
	inarg.blocksize = inode->i_sb->s_blocksize;
	req->in.h.opcode = FUSE_BMAP;
	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;
	req->out.args[0].size = sizeof(outarg);
	req->out.args[0].value = &outarg;
	request_send(fc, req);
	err = req->out.h.error;
	fuse_put_request(fc, req);
	if (err == -ENOSYS)
		fc->no_bmap = 1;

	return err ? 0 : outarg.block;
}

858
static const struct file_operations fuse_file_operations = {
M
Miklos Szeredi 已提交
859
	.llseek		= generic_file_llseek,
860 861 862 863
	.read		= do_sync_read,
	.aio_read	= generic_file_aio_read,
	.write		= do_sync_write,
	.aio_write	= generic_file_aio_write,
M
Miklos Szeredi 已提交
864 865 866 867 868
	.mmap		= fuse_file_mmap,
	.open		= fuse_open,
	.flush		= fuse_flush,
	.release	= fuse_release,
	.fsync		= fuse_fsync,
869
	.lock		= fuse_file_lock,
870
	.flock		= fuse_file_flock,
871
	.splice_read	= generic_file_splice_read,
M
Miklos Szeredi 已提交
872 873
};

874
static const struct file_operations fuse_direct_io_file_operations = {
M
Miklos Szeredi 已提交
875 876 877 878 879 880 881
	.llseek		= generic_file_llseek,
	.read		= fuse_direct_read,
	.write		= fuse_direct_write,
	.open		= fuse_open,
	.flush		= fuse_flush,
	.release	= fuse_release,
	.fsync		= fuse_fsync,
882
	.lock		= fuse_file_lock,
883
	.flock		= fuse_file_flock,
884
	/* no mmap and splice_read */
M
Miklos Szeredi 已提交
885 886
};

887
static const struct address_space_operations fuse_file_aops  = {
M
Miklos Szeredi 已提交
888
	.readpage	= fuse_readpage,
N
Nick Piggin 已提交
889 890
	.write_begin	= fuse_write_begin,
	.write_end	= fuse_write_end,
891
	.readpages	= fuse_readpages,
M
Miklos Szeredi 已提交
892
	.set_page_dirty	= fuse_set_page_dirty,
M
Miklos Szeredi 已提交
893
	.bmap		= fuse_bmap,
M
Miklos Szeredi 已提交
894 895 896 897
};

void fuse_init_file_inode(struct inode *inode)
{
898 899
	inode->i_fop = &fuse_file_operations;
	inode->i_data.a_ops = &fuse_file_aops;
M
Miklos Szeredi 已提交
900
}