splice.c 47.7 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11
/*
 * "splice": joining two ropes together by interweaving their strands.
 *
 * This is the "extended pipe" functionality, where a pipe is used as
 * an arbitrary in-memory buffer. Think of a pipe as a small kernel
 * buffer that you can use to transfer data from one end to the other.
 *
 * The traditional unix read/write is extended with a "splice()" operation
 * that transfers data buffers to or from a pipe buffer.
 *
 * Named by Larry McVoy, original implementation from Linus, extended by
12 13
 * Jens to support splicing to files, network, direct splicing, etc and
 * fixing lots of bugs.
14
 *
15
 * Copyright (C) 2005-2006 Jens Axboe <axboe@kernel.dk>
16 17
 * Copyright (C) 2005-2006 Linus Torvalds <torvalds@osdl.org>
 * Copyright (C) 2006 Ingo Molnar <mingo@elte.hu>
18 19 20 21 22
 *
 */
#include <linux/fs.h>
#include <linux/file.h>
#include <linux/pagemap.h>
23
#include <linux/splice.h>
K
KAMEZAWA Hiroyuki 已提交
24
#include <linux/memcontrol.h>
25
#include <linux/mm_inline.h>
26
#include <linux/swap.h>
27 28
#include <linux/writeback.h>
#include <linux/buffer_head.h>
J
Jeff Garzik 已提交
29
#include <linux/module.h>
30
#include <linux/syscalls.h>
31
#include <linux/uio.h>
32
#include <linux/security.h>
33

34 35 36 37 38 39
/*
 * Attempt to steal a page from a pipe buffer. This should perhaps go into
 * a vm helper function, it's already simplified quite a bit by the
 * addition of remove_mapping(). If success is returned, the caller may
 * attempt to reuse this page for another destination.
 */
40
static int page_cache_pipe_buf_steal(struct pipe_inode_info *pipe,
41 42 43
				     struct pipe_buffer *buf)
{
	struct page *page = buf->page;
44
	struct address_space *mapping;
45

46 47
	lock_page(page);

48 49 50
	mapping = page_mapping(page);
	if (mapping) {
		WARN_ON(!PageUptodate(page));
51

52 53 54 55 56 57 58 59 60
		/*
		 * At least for ext2 with nobh option, we need to wait on
		 * writeback completing on this page, since we'll remove it
		 * from the pagecache.  Otherwise truncate wont wait on the
		 * page, allowing the disk blocks to be reused by someone else
		 * before we actually wrote our data to them. fs corruption
		 * ensues.
		 */
		wait_on_page_writeback(page);
61

62 63
		if (page_has_private(page) &&
		    !try_to_release_page(page, GFP_KERNEL))
64
			goto out_unlock;
65

66 67 68 69 70 71 72 73
		/*
		 * If we succeeded in removing the mapping, set LRU flag
		 * and return good.
		 */
		if (remove_mapping(mapping, page)) {
			buf->flags |= PIPE_BUF_FLAG_LRU;
			return 0;
		}
74
	}
75

76 77 78 79
	/*
	 * Raced with truncate or failed to remove page from current
	 * address space, unlock and return failure.
	 */
80
out_unlock:
81 82
	unlock_page(page);
	return 1;
83 84
}

85
static void page_cache_pipe_buf_release(struct pipe_inode_info *pipe,
86 87 88
					struct pipe_buffer *buf)
{
	page_cache_release(buf->page);
J
Jens Axboe 已提交
89
	buf->flags &= ~PIPE_BUF_FLAG_LRU;
90 91
}

92 93 94 95
/*
 * Check whether the contents of buf is OK to access. Since the content
 * is a page cache page, IO may be in flight.
 */
96 97
static int page_cache_pipe_buf_confirm(struct pipe_inode_info *pipe,
				       struct pipe_buffer *buf)
98 99
{
	struct page *page = buf->page;
100
	int err;
101 102

	if (!PageUptodate(page)) {
103 104 105 106
		lock_page(page);

		/*
		 * Page got truncated/unhashed. This will cause a 0-byte
I
Ingo Molnar 已提交
107
		 * splice, if this is the first page.
108 109 110 111 112
		 */
		if (!page->mapping) {
			err = -ENODATA;
			goto error;
		}
113

114
		/*
I
Ingo Molnar 已提交
115
		 * Uh oh, read-error from disk.
116 117 118 119 120 121 122
		 */
		if (!PageUptodate(page)) {
			err = -EIO;
			goto error;
		}

		/*
123
		 * Page is ok afterall, we are done.
124
		 */
125 126 127
		unlock_page(page);
	}

128
	return 0;
129 130
error:
	unlock_page(page);
131
	return err;
132 133
}

134
static const struct pipe_buf_operations page_cache_pipe_buf_ops = {
135
	.can_merge = 0,
136 137
	.map = generic_pipe_buf_map,
	.unmap = generic_pipe_buf_unmap,
138
	.confirm = page_cache_pipe_buf_confirm,
139
	.release = page_cache_pipe_buf_release,
140
	.steal = page_cache_pipe_buf_steal,
141
	.get = generic_pipe_buf_get,
142 143
};

144 145 146
static int user_page_pipe_buf_steal(struct pipe_inode_info *pipe,
				    struct pipe_buffer *buf)
{
147 148 149
	if (!(buf->flags & PIPE_BUF_FLAG_GIFT))
		return 1;

J
Jens Axboe 已提交
150
	buf->flags |= PIPE_BUF_FLAG_LRU;
151
	return generic_pipe_buf_steal(pipe, buf);
152 153
}

154
static const struct pipe_buf_operations user_page_pipe_buf_ops = {
155
	.can_merge = 0,
156 157
	.map = generic_pipe_buf_map,
	.unmap = generic_pipe_buf_unmap,
158
	.confirm = generic_pipe_buf_confirm,
159 160
	.release = page_cache_pipe_buf_release,
	.steal = user_page_pipe_buf_steal,
161
	.get = generic_pipe_buf_get,
162 163
};

164 165 166 167 168 169
/**
 * splice_to_pipe - fill passed data into a pipe
 * @pipe:	pipe to fill
 * @spd:	data to fill
 *
 * Description:
R
Randy Dunlap 已提交
170
 *    @spd contains a map of pages and len/offset tuples, along with
171 172 173
 *    the struct pipe_buf_operations associated with these pages. This
 *    function will link that data to the pipe.
 *
174
 */
175 176
ssize_t splice_to_pipe(struct pipe_inode_info *pipe,
		       struct splice_pipe_desc *spd)
177
{
178
	unsigned int spd_pages = spd->nr_pages;
179
	int ret, do_wakeup, page_nr;
180 181 182

	ret = 0;
	do_wakeup = 0;
183
	page_nr = 0;
184

185
	pipe_lock(pipe);
186 187

	for (;;) {
188
		if (!pipe->readers) {
189 190 191 192 193 194
			send_sig(SIGPIPE, current, 0);
			if (!ret)
				ret = -EPIPE;
			break;
		}

195 196
		if (pipe->nrbufs < PIPE_BUFFERS) {
			int newbuf = (pipe->curbuf + pipe->nrbufs) & (PIPE_BUFFERS - 1);
197
			struct pipe_buffer *buf = pipe->bufs + newbuf;
198

199 200 201
			buf->page = spd->pages[page_nr];
			buf->offset = spd->partial[page_nr].offset;
			buf->len = spd->partial[page_nr].len;
202
			buf->private = spd->partial[page_nr].private;
203
			buf->ops = spd->ops;
204 205 206
			if (spd->flags & SPLICE_F_GIFT)
				buf->flags |= PIPE_BUF_FLAG_GIFT;

207
			pipe->nrbufs++;
208 209 210
			page_nr++;
			ret += buf->len;

211 212
			if (pipe->inode)
				do_wakeup = 1;
213

214
			if (!--spd->nr_pages)
215
				break;
216
			if (pipe->nrbufs < PIPE_BUFFERS)
217 218 219 220 221
				continue;

			break;
		}

222
		if (spd->flags & SPLICE_F_NONBLOCK) {
223 224 225 226 227
			if (!ret)
				ret = -EAGAIN;
			break;
		}

228 229 230 231 232 233 234
		if (signal_pending(current)) {
			if (!ret)
				ret = -ERESTARTSYS;
			break;
		}

		if (do_wakeup) {
235
			smp_mb();
236 237 238
			if (waitqueue_active(&pipe->wait))
				wake_up_interruptible_sync(&pipe->wait);
			kill_fasync(&pipe->fasync_readers, SIGIO, POLL_IN);
239 240 241
			do_wakeup = 0;
		}

242 243 244
		pipe->waiting_writers++;
		pipe_wait(pipe);
		pipe->waiting_writers--;
245 246
	}

247
	pipe_unlock(pipe);
248

249 250 251 252 253
	if (do_wakeup) {
		smp_mb();
		if (waitqueue_active(&pipe->wait))
			wake_up_interruptible(&pipe->wait);
		kill_fasync(&pipe->fasync_readers, SIGIO, POLL_IN);
254 255
	}

256
	while (page_nr < spd_pages)
257
		spd->spd_release(spd, page_nr++);
258 259 260 261

	return ret;
}

262 263 264 265 266
static void spd_release_page(struct splice_pipe_desc *spd, unsigned int i)
{
	page_cache_release(spd->pages[i]);
}

267
static int
268 269 270
__generic_file_splice_read(struct file *in, loff_t *ppos,
			   struct pipe_inode_info *pipe, size_t len,
			   unsigned int flags)
271 272
{
	struct address_space *mapping = in->f_mapping;
F
Fengguang Wu 已提交
273
	unsigned int loff, nr_pages, req_pages;
274
	struct page *pages[PIPE_BUFFERS];
275
	struct partial_page partial[PIPE_BUFFERS];
276
	struct page *page;
277 278
	pgoff_t index, end_index;
	loff_t isize;
279
	int error, page_nr;
280 281 282 283 284
	struct splice_pipe_desc spd = {
		.pages = pages,
		.partial = partial,
		.flags = flags,
		.ops = &page_cache_pipe_buf_ops,
285
		.spd_release = spd_release_page,
286
	};
287

288
	index = *ppos >> PAGE_CACHE_SHIFT;
289
	loff = *ppos & ~PAGE_CACHE_MASK;
F
Fengguang Wu 已提交
290 291
	req_pages = (len + loff + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
	nr_pages = min(req_pages, (unsigned)PIPE_BUFFERS);
292

293 294 295 296
	/*
	 * Lookup the (hopefully) full range of pages we need.
	 */
	spd.nr_pages = find_get_pages_contig(mapping, index, nr_pages, pages);
297
	index += spd.nr_pages;
298

299 300
	/*
	 * If find_get_pages_contig() returned fewer pages than we needed,
301
	 * readahead/allocate the rest and fill in the holes.
302
	 */
303
	if (spd.nr_pages < nr_pages)
304 305
		page_cache_sync_readahead(mapping, &in->f_ra, in,
				index, req_pages - spd.nr_pages);
306

307
	error = 0;
308
	while (spd.nr_pages < nr_pages) {
309
		/*
310 311
		 * Page could be there, find_get_pages_contig() breaks on
		 * the first hole.
312
		 */
313 314 315
		page = find_get_page(mapping, index);
		if (!page) {
			/*
316
			 * page didn't exist, allocate one.
317 318 319 320 321 322
			 */
			page = page_cache_alloc_cold(mapping);
			if (!page)
				break;

			error = add_to_page_cache_lru(page, mapping, index,
H
Hugh Dickins 已提交
323
						mapping_gfp_mask(mapping));
324 325
			if (unlikely(error)) {
				page_cache_release(page);
326 327
				if (error == -EEXIST)
					continue;
328 329
				break;
			}
330 331 332 333 334
			/*
			 * add_to_page_cache() locks the page, unlock it
			 * to avoid convoluting the logic below even more.
			 */
			unlock_page(page);
335 336
		}

337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359
		pages[spd.nr_pages++] = page;
		index++;
	}

	/*
	 * Now loop over the map and see if we need to start IO on any
	 * pages, fill in the partial map, etc.
	 */
	index = *ppos >> PAGE_CACHE_SHIFT;
	nr_pages = spd.nr_pages;
	spd.nr_pages = 0;
	for (page_nr = 0; page_nr < nr_pages; page_nr++) {
		unsigned int this_len;

		if (!len)
			break;

		/*
		 * this_len is the max we'll use from this page
		 */
		this_len = min_t(unsigned long, len, PAGE_CACHE_SIZE - loff);
		page = pages[page_nr];

360
		if (PageReadahead(page))
361
			page_cache_async_readahead(mapping, &in->f_ra, in,
F
Fengguang Wu 已提交
362
					page, index, req_pages - page_nr);
363

364 365 366 367
		/*
		 * If the page isn't uptodate, we may need to start io on it
		 */
		if (!PageUptodate(page)) {
368 369 370 371
			/*
			 * If in nonblock mode then dont block on waiting
			 * for an in-flight io page
			 */
372
			if (flags & SPLICE_F_NONBLOCK) {
N
Nick Piggin 已提交
373
				if (!trylock_page(page)) {
374
					error = -EAGAIN;
375
					break;
376
				}
377 378
			} else
				lock_page(page);
379 380

			/*
381 382 383 384
			 * Page was truncated, or invalidated by the
			 * filesystem.  Redo the find/create, but this time the
			 * page is kept locked, so there's no chance of another
			 * race with truncate/invalidate.
385 386 387
			 */
			if (!page->mapping) {
				unlock_page(page);
388 389 390 391 392 393 394 395 396
				page = find_or_create_page(mapping, index,
						mapping_gfp_mask(mapping));

				if (!page) {
					error = -ENOMEM;
					break;
				}
				page_cache_release(pages[page_nr]);
				pages[page_nr] = page;
397 398 399 400 401 402 403 404
			}
			/*
			 * page was already under io and is now done, great
			 */
			if (PageUptodate(page)) {
				unlock_page(page);
				goto fill_it;
			}
405

406 407 408 409
			/*
			 * need to read in the page
			 */
			error = mapping->a_ops->readpage(in, page);
410
			if (unlikely(error)) {
411 412 413 414 415 416
				/*
				 * We really should re-lookup the page here,
				 * but it complicates things a lot. Instead
				 * lets just do what we already stored, and
				 * we'll get it the next time we are called.
				 */
417
				if (error == AOP_TRUNCATED_PAGE)
418 419
					error = 0;

420 421
				break;
			}
422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437
		}
fill_it:
		/*
		 * i_size must be checked after PageUptodate.
		 */
		isize = i_size_read(mapping->host);
		end_index = (isize - 1) >> PAGE_CACHE_SHIFT;
		if (unlikely(!isize || index > end_index))
			break;

		/*
		 * if this is the last page, see if we need to shrink
		 * the length and stop
		 */
		if (end_index == index) {
			unsigned int plen;
438 439

			/*
440
			 * max good bytes in this page
441
			 */
442 443
			plen = ((isize - 1) & ~PAGE_CACHE_MASK) + 1;
			if (plen <= loff)
444 445 446
				break;

			/*
447
			 * force quit after adding this page
448
			 */
449 450
			this_len = min(this_len, plen - loff);
			len = this_len;
451
		}
452

453 454
		partial[page_nr].offset = loff;
		partial[page_nr].len = this_len;
455
		len -= this_len;
456
		loff = 0;
457 458
		spd.nr_pages++;
		index++;
459 460
	}

461
	/*
462
	 * Release any pages at the end, if we quit early. 'page_nr' is how far
463 464 465 466
	 * we got, 'nr_pages' is how many pages are in the map.
	 */
	while (page_nr < nr_pages)
		page_cache_release(pages[page_nr++]);
467
	in->f_ra.prev_pos = (loff_t)index << PAGE_CACHE_SHIFT;
468

469
	if (spd.nr_pages)
470
		return splice_to_pipe(pipe, &spd);
471

472
	return error;
473 474
}

475 476 477
/**
 * generic_file_splice_read - splice data from file to a pipe
 * @in:		file to splice from
478
 * @ppos:	position in @in
479 480 481 482
 * @pipe:	pipe to splice to
 * @len:	number of bytes to splice
 * @flags:	splice modifier flags
 *
483 484 485 486 487
 * Description:
 *    Will read pages from given file and fill them into a pipe. Can be
 *    used as long as the address_space operations for the source implements
 *    a readpage() hook.
 *
488
 */
489 490 491
ssize_t generic_file_splice_read(struct file *in, loff_t *ppos,
				 struct pipe_inode_info *pipe, size_t len,
				 unsigned int flags)
492
{
493
	loff_t isize, left;
494
	int ret;
495 496 497 498 499 500 501 502

	isize = i_size_read(in->f_mapping->host);
	if (unlikely(*ppos >= isize))
		return 0;

	left = isize - *ppos;
	if (unlikely(left < len))
		len = left;
503

504 505
	ret = __generic_file_splice_read(in, ppos, pipe, len, flags);
	if (ret > 0)
506
		*ppos += ret;
507 508 509

	return ret;
}
510 511
EXPORT_SYMBOL(generic_file_splice_read);

512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537
static const struct pipe_buf_operations default_pipe_buf_ops = {
	.can_merge = 0,
	.map = generic_pipe_buf_map,
	.unmap = generic_pipe_buf_unmap,
	.confirm = generic_pipe_buf_confirm,
	.release = generic_pipe_buf_release,
	.steal = generic_pipe_buf_steal,
	.get = generic_pipe_buf_get,
};

static ssize_t kernel_readv(struct file *file, const struct iovec *vec,
			    unsigned long vlen, loff_t offset)
{
	mm_segment_t old_fs;
	loff_t pos = offset;
	ssize_t res;

	old_fs = get_fs();
	set_fs(get_ds());
	/* The cast to a user pointer is valid due to the set_fs() */
	res = vfs_readv(file, (const struct iovec __user *)vec, vlen, &pos);
	set_fs(old_fs);

	return res;
}

538 539 540 541 542 543 544 545 546 547 548 549 550 551 552
static ssize_t kernel_writev(struct file *file, const struct iovec *vec,
			    unsigned long vlen, loff_t *ppos)
{
	mm_segment_t old_fs;
	ssize_t res;

	old_fs = get_fs();
	set_fs(get_ds());
	/* The cast to a user pointer is valid due to the set_fs() */
	res = vfs_writev(file, (const struct iovec __user *)vec, vlen, ppos);
	set_fs(old_fs);

	return res;
}

553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582
ssize_t default_file_splice_read(struct file *in, loff_t *ppos,
				 struct pipe_inode_info *pipe, size_t len,
				 unsigned int flags)
{
	unsigned int nr_pages;
	unsigned int nr_freed;
	size_t offset;
	struct page *pages[PIPE_BUFFERS];
	struct partial_page partial[PIPE_BUFFERS];
	struct iovec vec[PIPE_BUFFERS];
	pgoff_t index;
	ssize_t res;
	size_t this_len;
	int error;
	int i;
	struct splice_pipe_desc spd = {
		.pages = pages,
		.partial = partial,
		.flags = flags,
		.ops = &default_pipe_buf_ops,
		.spd_release = spd_release_page,
	};

	index = *ppos >> PAGE_CACHE_SHIFT;
	offset = *ppos & ~PAGE_CACHE_MASK;
	nr_pages = (len + offset + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;

	for (i = 0; i < nr_pages && i < PIPE_BUFFERS && len; i++) {
		struct page *page;

583
		page = alloc_page(GFP_USER);
584 585 586 587 588
		error = -ENOMEM;
		if (!page)
			goto err;

		this_len = min_t(size_t, len, PAGE_CACHE_SIZE - offset);
589
		vec[i].iov_base = (void __user *) page_address(page);
590 591 592 593 594 595 596 597
		vec[i].iov_len = this_len;
		pages[i] = page;
		spd.nr_pages++;
		len -= this_len;
		offset = 0;
	}

	res = kernel_readv(in, vec, spd.nr_pages, *ppos);
A
Andrew Morton 已提交
598 599
	if (res < 0) {
		error = res;
600
		goto err;
A
Andrew Morton 已提交
601
	}
602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627

	error = 0;
	if (!res)
		goto err;

	nr_freed = 0;
	for (i = 0; i < spd.nr_pages; i++) {
		this_len = min_t(size_t, vec[i].iov_len, res);
		partial[i].offset = 0;
		partial[i].len = this_len;
		if (!this_len) {
			__free_page(pages[i]);
			pages[i] = NULL;
			nr_freed++;
		}
		res -= this_len;
	}
	spd.nr_pages -= nr_freed;

	res = splice_to_pipe(pipe, &spd);
	if (res > 0)
		*ppos += res;

	return res;

err:
628
	for (i = 0; i < spd.nr_pages; i++)
629
		__free_page(pages[i]);
630

631 632 633 634
	return error;
}
EXPORT_SYMBOL(default_file_splice_read);

635
/*
636
 * Send 'sd->len' bytes to socket from 'sd->file' at position 'sd->pos'
J
Jens Axboe 已提交
637
 * using sendpage(). Return the number of bytes sent.
638
 */
639
static int pipe_to_sendpage(struct pipe_inode_info *pipe,
640 641
			    struct pipe_buffer *buf, struct splice_desc *sd)
{
642
	struct file *file = sd->u.file;
643
	loff_t pos = sd->pos;
644
	int ret, more;
645

646
	ret = buf->ops->confirm(pipe, buf);
647 648
	if (!ret) {
		more = (sd->flags & SPLICE_F_MORE) || sd->len < sd->total_len;
649

650 651 652
		ret = file->f_op->sendpage(file, buf->page, buf->offset,
					   sd->len, &pos, more);
	}
653

J
Jens Axboe 已提交
654
	return ret;
655 656 657 658 659 660 661 662 663 664 665 666 667 668 669
}

/*
 * This is a little more tricky than the file -> pipe splicing. There are
 * basically three cases:
 *
 *	- Destination page already exists in the address space and there
 *	  are users of it. For that case we have no other option that
 *	  copying the data. Tough luck.
 *	- Destination page already exists in the address space, but there
 *	  are no users of it. Make sure it's uptodate, then drop it. Fall
 *	  through to last case.
 *	- Destination page does not exist, we can add the pipe page to
 *	  the page cache and avoid the copy.
 *
670 671 672 673 674 675
 * If asked to move pages to the output file (SPLICE_F_MOVE is set in
 * sd->flags), we attempt to migrate pages from the pipe to the output
 * file address space page cache. This is possible if no one else has
 * the pipe page referenced outside of the pipe and page cache. If
 * SPLICE_F_MOVE isn't set, or we cannot move the page, we simply create
 * a new page in the output file page cache and fill/dirty that.
676
 */
677 678
int pipe_to_file(struct pipe_inode_info *pipe, struct pipe_buffer *buf,
		 struct splice_desc *sd)
679
{
680
	struct file *file = sd->u.file;
681
	struct address_space *mapping = file->f_mapping;
J
Jens Axboe 已提交
682
	unsigned int offset, this_len;
683
	struct page *page;
684
	void *fsdata;
685
	int ret;
686 687

	/*
688
	 * make sure the data in this buffer is uptodate
689
	 */
690
	ret = buf->ops->confirm(pipe, buf);
691 692
	if (unlikely(ret))
		return ret;
693 694 695

	offset = sd->pos & ~PAGE_CACHE_MASK;

J
Jens Axboe 已提交
696 697 698 699
	this_len = sd->len;
	if (this_len + offset > PAGE_CACHE_SIZE)
		this_len = PAGE_CACHE_SIZE - offset;

700 701 702 703
	ret = pagecache_write_begin(file, mapping, sd->pos, this_len,
				AOP_FLAG_UNINTERRUPTIBLE, &page, &fsdata);
	if (unlikely(ret))
		goto out;
704

705
	if (buf->page != page) {
706 707 708
		/*
		 * Careful, ->map() uses KM_USER0!
		 */
709
		char *src = buf->ops->map(pipe, buf, 1);
710
		char *dst = kmap_atomic(page, KM_USER1);
711

J
Jens Axboe 已提交
712
		memcpy(dst + offset, src + buf->offset, this_len);
713
		flush_dcache_page(page);
714
		kunmap_atomic(dst, KM_USER1);
715
		buf->ops->unmap(pipe, buf, src);
716
	}
717 718
	ret = pagecache_write_end(file, mapping, sd->pos, this_len, this_len,
				page, fsdata);
719 720 721
out:
	return ret;
}
722
EXPORT_SYMBOL(pipe_to_file);
723

724 725 726 727 728 729 730 731
static void wakeup_pipe_writers(struct pipe_inode_info *pipe)
{
	smp_mb();
	if (waitqueue_active(&pipe->wait))
		wake_up_interruptible(&pipe->wait);
	kill_fasync(&pipe->fasync_writers, SIGIO, POLL_OUT);
}

732
/**
733
 * splice_from_pipe_feed - feed available data from a pipe to a file
734 735 736 737 738
 * @pipe:	pipe to splice from
 * @sd:		information to @actor
 * @actor:	handler that splices the data
 *
 * Description:
739 740 741 742 743 744 745
 *    This function loops over the pipe and calls @actor to do the
 *    actual moving of a single struct pipe_buffer to the desired
 *    destination.  It returns when there's no more buffers left in
 *    the pipe or if the requested number of bytes (@sd->total_len)
 *    have been copied.  It returns a positive number (one) if the
 *    pipe needs to be filled with more data, zero if the required
 *    number of bytes have been copied and -errno on error.
746
 *
747 748 749 750
 *    This, together with splice_from_pipe_{begin,end,next}, may be
 *    used to implement the functionality of __splice_from_pipe() when
 *    locking is required around copying the pipe buffers to the
 *    destination.
751
 */
752 753
int splice_from_pipe_feed(struct pipe_inode_info *pipe, struct splice_desc *sd,
			  splice_actor *actor)
754
{
755
	int ret;
756

757 758 759
	while (pipe->nrbufs) {
		struct pipe_buffer *buf = pipe->bufs + pipe->curbuf;
		const struct pipe_buf_operations *ops = buf->ops;
760

761 762 763
		sd->len = buf->len;
		if (sd->len > sd->total_len)
			sd->len = sd->total_len;
764

765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786
		ret = actor(pipe, buf, sd);
		if (ret <= 0) {
			if (ret == -ENODATA)
				ret = 0;
			return ret;
		}
		buf->offset += ret;
		buf->len -= ret;

		sd->num_spliced += ret;
		sd->len -= ret;
		sd->pos += ret;
		sd->total_len -= ret;

		if (!buf->len) {
			buf->ops = NULL;
			ops->release(pipe, buf);
			pipe->curbuf = (pipe->curbuf + 1) & (PIPE_BUFFERS - 1);
			pipe->nrbufs--;
			if (pipe->inode)
				sd->need_wakeup = true;
		}
787

788 789 790
		if (!sd->total_len)
			return 0;
	}
791

792 793 794
	return 1;
}
EXPORT_SYMBOL(splice_from_pipe_feed);
795

796 797 798 799 800 801 802 803 804 805 806 807 808 809 810
/**
 * splice_from_pipe_next - wait for some data to splice from
 * @pipe:	pipe to splice from
 * @sd:		information about the splice operation
 *
 * Description:
 *    This function will wait for some data and return a positive
 *    value (one) if pipe buffers are available.  It will return zero
 *    or -errno if no more data needs to be spliced.
 */
int splice_from_pipe_next(struct pipe_inode_info *pipe, struct splice_desc *sd)
{
	while (!pipe->nrbufs) {
		if (!pipe->writers)
			return 0;
J
Jens Axboe 已提交
811

812 813
		if (!pipe->waiting_writers && sd->num_spliced)
			return 0;
I
Ingo Molnar 已提交
814

815 816
		if (sd->flags & SPLICE_F_NONBLOCK)
			return -EAGAIN;
817

818 819
		if (signal_pending(current))
			return -ERESTARTSYS;
820

821 822 823
		if (sd->need_wakeup) {
			wakeup_pipe_writers(pipe);
			sd->need_wakeup = false;
824 825
		}

826 827
		pipe_wait(pipe);
	}
828

829 830 831
	return 1;
}
EXPORT_SYMBOL(splice_from_pipe_next);
832

833 834
/**
 * splice_from_pipe_begin - start splicing from pipe
835
 * @sd:		information about the splice operation
836 837 838 839 840 841 842 843 844 845 846 847
 *
 * Description:
 *    This function should be called before a loop containing
 *    splice_from_pipe_next() and splice_from_pipe_feed() to
 *    initialize the necessary fields of @sd.
 */
void splice_from_pipe_begin(struct splice_desc *sd)
{
	sd->num_spliced = 0;
	sd->need_wakeup = false;
}
EXPORT_SYMBOL(splice_from_pipe_begin);
848

849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864
/**
 * splice_from_pipe_end - finish splicing from pipe
 * @pipe:	pipe to splice from
 * @sd:		information about the splice operation
 *
 * Description:
 *    This function will wake up pipe writers if necessary.  It should
 *    be called after a loop containing splice_from_pipe_next() and
 *    splice_from_pipe_feed().
 */
void splice_from_pipe_end(struct pipe_inode_info *pipe, struct splice_desc *sd)
{
	if (sd->need_wakeup)
		wakeup_pipe_writers(pipe);
}
EXPORT_SYMBOL(splice_from_pipe_end);
865

866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882
/**
 * __splice_from_pipe - splice data from a pipe to given actor
 * @pipe:	pipe to splice from
 * @sd:		information to @actor
 * @actor:	handler that splices the data
 *
 * Description:
 *    This function does little more than loop over the pipe and call
 *    @actor to do the actual moving of a single struct pipe_buffer to
 *    the desired destination. See pipe_to_file, pipe_to_sendpage, or
 *    pipe_to_user.
 *
 */
ssize_t __splice_from_pipe(struct pipe_inode_info *pipe, struct splice_desc *sd,
			   splice_actor *actor)
{
	int ret;
883

884 885 886 887 888 889 890 891 892
	splice_from_pipe_begin(sd);
	do {
		ret = splice_from_pipe_next(pipe, sd);
		if (ret > 0)
			ret = splice_from_pipe_feed(pipe, sd, actor);
	} while (ret > 0);
	splice_from_pipe_end(pipe, sd);

	return sd->num_spliced ? sd->num_spliced : ret;
893
}
M
Mark Fasheh 已提交
894
EXPORT_SYMBOL(__splice_from_pipe);
895

896 897 898 899 900 901 902 903 904 905
/**
 * splice_from_pipe - splice data from a pipe to a file
 * @pipe:	pipe to splice from
 * @out:	file to splice to
 * @ppos:	position in @out
 * @len:	how many bytes to splice
 * @flags:	splice modifier flags
 * @actor:	handler that splices the data
 *
 * Description:
906
 *    See __splice_from_pipe. This function locks the pipe inode,
907 908 909
 *    otherwise it's identical to __splice_from_pipe().
 *
 */
910 911 912 913 914
ssize_t splice_from_pipe(struct pipe_inode_info *pipe, struct file *out,
			 loff_t *ppos, size_t len, unsigned int flags,
			 splice_actor *actor)
{
	ssize_t ret;
J
Jens Axboe 已提交
915 916 917 918
	struct splice_desc sd = {
		.total_len = len,
		.flags = flags,
		.pos = *ppos,
919
		.u.file = out,
J
Jens Axboe 已提交
920
	};
921

922
	pipe_lock(pipe);
J
Jens Axboe 已提交
923
	ret = __splice_from_pipe(pipe, &sd, actor);
924
	pipe_unlock(pipe);
925 926 927 928

	return ret;
}

929 930
/**
 * generic_file_splice_write - splice data from a pipe to a file
931
 * @pipe:	pipe info
932
 * @out:	file to write to
933
 * @ppos:	position in @out
934 935 936
 * @len:	number of bytes to splice
 * @flags:	splice modifier flags
 *
937 938 939
 * Description:
 *    Will either move or copy pages (determined by @flags options) from
 *    the given pipe inode to the given file.
940 941
 *
 */
942 943
ssize_t
generic_file_splice_write(struct pipe_inode_info *pipe, struct file *out,
944
			  loff_t *ppos, size_t len, unsigned int flags)
945
{
946
	struct address_space *mapping = out->f_mapping;
947
	struct inode *inode = mapping->host;
948 949 950 951 952 953
	struct splice_desc sd = {
		.total_len = len,
		.flags = flags,
		.pos = *ppos,
		.u.file = out,
	};
954 955
	ssize_t ret;

956
	pipe_lock(pipe);
957 958 959 960 961 962 963 964 965 966 967 968 969 970 971

	splice_from_pipe_begin(&sd);
	do {
		ret = splice_from_pipe_next(pipe, &sd);
		if (ret <= 0)
			break;

		mutex_lock_nested(&inode->i_mutex, I_MUTEX_CHILD);
		ret = file_remove_suid(out);
		if (!ret)
			ret = splice_from_pipe_feed(pipe, &sd, pipe_to_file);
		mutex_unlock(&inode->i_mutex);
	} while (ret > 0);
	splice_from_pipe_end(pipe, &sd);

972
	pipe_unlock(pipe);
973 974 975 976

	if (sd.num_spliced)
		ret = sd.num_spliced;

J
Jens Axboe 已提交
977
	if (ret > 0) {
978 979
		unsigned long nr_pages;

J
Jens Axboe 已提交
980
		*ppos += ret;
981
		nr_pages = (ret + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
J
Jens Axboe 已提交
982 983 984 985 986 987

		/*
		 * If file or inode is SYNC and we actually wrote some data,
		 * sync it.
		 */
		if (unlikely((out->f_flags & O_SYNC) || IS_SYNC(inode))) {
988 989
			int err;

J
Jens Axboe 已提交
990 991 992 993
			mutex_lock(&inode->i_mutex);
			err = generic_osync_inode(inode, mapping,
						  OSYNC_METADATA|OSYNC_DATA);
			mutex_unlock(&inode->i_mutex);
994

J
Jens Axboe 已提交
995 996 997
			if (err)
				ret = err;
		}
998
		balance_dirty_pages_ratelimited_nr(mapping, nr_pages);
999 1000 1001
	}

	return ret;
1002 1003
}

1004 1005
EXPORT_SYMBOL(generic_file_splice_write);

1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 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 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121
static struct pipe_buffer *nth_pipe_buf(struct pipe_inode_info *pipe, int n)
{
	return &pipe->bufs[(pipe->curbuf + n) % PIPE_BUFFERS];
}

static ssize_t default_file_splice_write(struct pipe_inode_info *pipe,
					 struct file *out, loff_t *ppos,
					 size_t len, unsigned int flags)
{
	ssize_t ret = 0;
	ssize_t total_len = 0;
	int do_wakeup = 0;

	pipe_lock(pipe);
	while (len) {
		struct pipe_buffer *buf;
		void *data[PIPE_BUFFERS];
		struct iovec vec[PIPE_BUFFERS];
		unsigned int nr_pages = 0;
		unsigned int write_len = 0;
		unsigned int now_len = len;
		unsigned int this_len;
		int i;

		BUG_ON(pipe->nrbufs > PIPE_BUFFERS);
		for (i = 0; i < pipe->nrbufs && now_len; i++) {
			buf = nth_pipe_buf(pipe, i);

			ret = buf->ops->confirm(pipe, buf);
			if (ret)
				break;

			data[i] = buf->ops->map(pipe, buf, 0);
			this_len = min(buf->len, now_len);
			vec[i].iov_base = (void __user *) data[i] + buf->offset;
			vec[i].iov_len = this_len;
			now_len -= this_len;
			write_len += this_len;
			nr_pages++;
		}

		if (nr_pages) {
			ret = kernel_writev(out, vec, nr_pages, ppos);
			if (ret == 0)
				ret = -EIO;
			if (ret > 0) {
				len -= ret;
				total_len += ret;
			}
		}

		for (i = 0; i < nr_pages; i++) {
			buf = nth_pipe_buf(pipe, i);
			buf->ops->unmap(pipe, buf, data[i]);

			if (ret > 0) {
				this_len = min_t(unsigned, vec[i].iov_len, ret);
				buf->offset += this_len;
				buf->len -= this_len;
				ret -= this_len;
			}
		}

		if (ret < 0)
			break;

		while (pipe->nrbufs) {
			const struct pipe_buf_operations *ops;

			buf = nth_pipe_buf(pipe, 0);
			if (buf->len)
				break;

			ops = buf->ops;
			buf->ops = NULL;
			ops->release(pipe, buf);
			pipe->curbuf = (pipe->curbuf + 1) % PIPE_BUFFERS;
			pipe->nrbufs--;
			if (pipe->inode)
				do_wakeup = 1;
		}

		if (pipe->nrbufs)
			continue;
		if (!pipe->writers)
			break;
		if (!pipe->waiting_writers) {
			if (total_len)
				break;
		}

		if (flags & SPLICE_F_NONBLOCK) {
			ret = -EAGAIN;
			break;
		}

		if (signal_pending(current)) {
			ret = -ERESTARTSYS;
			break;
		}

		if (do_wakeup) {
			wakeup_pipe_writers(pipe);
			do_wakeup = 0;
		}

		pipe_wait(pipe);
	}
	pipe_unlock(pipe);

	if (do_wakeup)
		wakeup_pipe_writers(pipe);

	return total_len ? total_len : ret;
}

1122 1123
/**
 * generic_splice_sendpage - splice data from a pipe to a socket
1124
 * @pipe:	pipe to splice from
1125
 * @out:	socket to write to
1126
 * @ppos:	position in @out
1127 1128 1129
 * @len:	number of bytes to splice
 * @flags:	splice modifier flags
 *
1130 1131 1132
 * Description:
 *    Will send @len bytes from the pipe to a network socket. No data copying
 *    is involved.
1133 1134
 *
 */
1135
ssize_t generic_splice_sendpage(struct pipe_inode_info *pipe, struct file *out,
1136
				loff_t *ppos, size_t len, unsigned int flags)
1137
{
1138
	return splice_from_pipe(pipe, out, ppos, len, flags, pipe_to_sendpage);
1139 1140
}

1141
EXPORT_SYMBOL(generic_splice_sendpage);
J
Jeff Garzik 已提交
1142

1143 1144 1145
/*
 * Attempt to initiate a splice from pipe to file.
 */
1146
static long do_splice_from(struct pipe_inode_info *pipe, struct file *out,
1147
			   loff_t *ppos, size_t len, unsigned int flags)
1148
{
1149 1150
	ssize_t (*splice_write)(struct pipe_inode_info *, struct file *,
				loff_t *, size_t, unsigned int);
1151 1152
	int ret;

1153
	if (unlikely(!(out->f_mode & FMODE_WRITE)))
1154 1155
		return -EBADF;

1156 1157 1158
	if (unlikely(out->f_flags & O_APPEND))
		return -EINVAL;

1159
	ret = rw_verify_area(WRITE, out, ppos, len);
1160 1161 1162
	if (unlikely(ret < 0))
		return ret;

1163 1164 1165 1166 1167
	splice_write = out->f_op->splice_write;
	if (!splice_write)
		splice_write = default_file_splice_write;

	return splice_write(pipe, out, ppos, len, flags);
1168 1169
}

1170 1171 1172
/*
 * Attempt to initiate a splice from a file to a pipe.
 */
1173 1174 1175
static long do_splice_to(struct file *in, loff_t *ppos,
			 struct pipe_inode_info *pipe, size_t len,
			 unsigned int flags)
1176
{
1177 1178
	ssize_t (*splice_read)(struct file *, loff_t *,
			       struct pipe_inode_info *, size_t, unsigned int);
1179 1180
	int ret;

1181
	if (unlikely(!(in->f_mode & FMODE_READ)))
1182 1183
		return -EBADF;

1184
	ret = rw_verify_area(READ, in, ppos, len);
1185 1186 1187
	if (unlikely(ret < 0))
		return ret;

1188 1189 1190 1191 1192
	splice_read = in->f_op->splice_read;
	if (!splice_read)
		splice_read = default_file_splice_read;

	return splice_read(in, ppos, pipe, len, flags);
1193 1194
}

1195 1196 1197 1198 1199 1200 1201 1202 1203
/**
 * splice_direct_to_actor - splices data directly between two non-pipes
 * @in:		file to splice from
 * @sd:		actor information on where to splice to
 * @actor:	handles the data splicing
 *
 * Description:
 *    This is a special case helper to splice directly between two
 *    points, without requiring an explicit pipe. Internally an allocated
R
Randy Dunlap 已提交
1204
 *    pipe is cached in the process, and reused during the lifetime of
1205 1206
 *    that process.
 *
J
Jens Axboe 已提交
1207 1208 1209
 */
ssize_t splice_direct_to_actor(struct file *in, struct splice_desc *sd,
			       splice_direct_actor *actor)
1210 1211 1212 1213
{
	struct pipe_inode_info *pipe;
	long ret, bytes;
	umode_t i_mode;
J
Jens Axboe 已提交
1214 1215
	size_t len;
	int i, flags;
1216 1217 1218 1219 1220 1221

	/*
	 * We require the input being a regular file, as we don't want to
	 * randomly drop data for eg socket -> socket splicing. Use the
	 * piped splicing for that!
	 */
1222
	i_mode = in->f_path.dentry->d_inode->i_mode;
1223 1224 1225 1226 1227 1228 1229 1230
	if (unlikely(!S_ISREG(i_mode) && !S_ISBLK(i_mode)))
		return -EINVAL;

	/*
	 * neither in nor out is a pipe, setup an internal pipe attached to
	 * 'out' and transfer the wanted data from 'in' to 'out' through that
	 */
	pipe = current->splice_pipe;
1231
	if (unlikely(!pipe)) {
1232 1233 1234 1235 1236 1237
		pipe = alloc_pipe_info(NULL);
		if (!pipe)
			return -ENOMEM;

		/*
		 * We don't have an immediate reader, but we'll read the stuff
1238
		 * out of the pipe right after the splice_to_pipe(). So set
1239 1240 1241 1242 1243 1244 1245 1246
		 * PIPE_READERS appropriately.
		 */
		pipe->readers = 1;

		current->splice_pipe = pipe;
	}

	/*
I
Ingo Molnar 已提交
1247
	 * Do the splice.
1248 1249 1250
	 */
	ret = 0;
	bytes = 0;
J
Jens Axboe 已提交
1251 1252 1253 1254 1255 1256 1257
	len = sd->total_len;
	flags = sd->flags;

	/*
	 * Don't block on output, we have to drain the direct pipe.
	 */
	sd->flags &= ~SPLICE_F_NONBLOCK;
1258 1259

	while (len) {
1260
		size_t read_len;
1261
		loff_t pos = sd->pos, prev_pos = pos;
1262

1263
		ret = do_splice_to(in, &pos, pipe, len, flags);
1264
		if (unlikely(ret <= 0))
1265 1266 1267
			goto out_release;

		read_len = ret;
J
Jens Axboe 已提交
1268
		sd->total_len = read_len;
1269 1270 1271 1272 1273 1274

		/*
		 * NOTE: nonblocking mode only applies to the input. We
		 * must not do the output in nonblocking mode as then we
		 * could get stuck data in the internal pipe:
		 */
J
Jens Axboe 已提交
1275
		ret = actor(pipe, sd);
1276 1277
		if (unlikely(ret <= 0)) {
			sd->pos = prev_pos;
1278
			goto out_release;
1279
		}
1280 1281 1282

		bytes += ret;
		len -= ret;
1283
		sd->pos = pos;
1284

1285 1286
		if (ret < read_len) {
			sd->pos = prev_pos + ret;
1287
			goto out_release;
1288
		}
1289 1290
	}

1291
done:
1292
	pipe->nrbufs = pipe->curbuf = 0;
1293
	file_accessed(in);
1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309
	return bytes;

out_release:
	/*
	 * If we did an incomplete transfer we must release
	 * the pipe buffers in question:
	 */
	for (i = 0; i < PIPE_BUFFERS; i++) {
		struct pipe_buffer *buf = pipe->bufs + i;

		if (buf->ops) {
			buf->ops->release(pipe, buf);
			buf->ops = NULL;
		}
	}

1310 1311
	if (!bytes)
		bytes = ret;
J
Jens Axboe 已提交
1312

1313
	goto done;
J
Jens Axboe 已提交
1314 1315 1316 1317 1318 1319
}
EXPORT_SYMBOL(splice_direct_to_actor);

static int direct_splice_actor(struct pipe_inode_info *pipe,
			       struct splice_desc *sd)
{
1320
	struct file *file = sd->u.file;
J
Jens Axboe 已提交
1321 1322 1323 1324

	return do_splice_from(pipe, file, &sd->pos, sd->total_len, sd->flags);
}

1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339
/**
 * do_splice_direct - splices data directly between two files
 * @in:		file to splice from
 * @ppos:	input file offset
 * @out:	file to splice to
 * @len:	number of bytes to splice
 * @flags:	splice modifier flags
 *
 * Description:
 *    For use by do_sendfile(). splice can easily emulate sendfile, but
 *    doing it in the application would incur an extra system call
 *    (splice in + splice out, as compared to just sendfile()). So this helper
 *    can splice directly through a process-private pipe.
 *
 */
J
Jens Axboe 已提交
1340 1341 1342 1343 1344 1345 1346 1347
long do_splice_direct(struct file *in, loff_t *ppos, struct file *out,
		      size_t len, unsigned int flags)
{
	struct splice_desc sd = {
		.len		= len,
		.total_len	= len,
		.flags		= flags,
		.pos		= *ppos,
1348
		.u.file		= out,
J
Jens Axboe 已提交
1349
	};
1350
	long ret;
J
Jens Axboe 已提交
1351 1352

	ret = splice_direct_to_actor(in, &sd, direct_splice_actor);
1353
	if (ret > 0)
1354
		*ppos = sd.pos;
1355

J
Jens Axboe 已提交
1356
	return ret;
1357 1358
}

1359 1360 1361
static int splice_pipe_to_pipe(struct pipe_inode_info *ipipe,
			       struct pipe_inode_info *opipe,
			       size_t len, unsigned int flags);
1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374
/*
 * After the inode slimming patch, i_pipe/i_bdev/i_cdev share the same
 * location, so checking ->i_pipe is not enough to verify that this is a
 * pipe.
 */
static inline struct pipe_inode_info *pipe_info(struct inode *inode)
{
	if (S_ISFIFO(inode->i_mode))
		return inode->i_pipe;

	return NULL;
}

1375 1376 1377
/*
 * Determine where to splice to/from.
 */
1378 1379 1380
static long do_splice(struct file *in, loff_t __user *off_in,
		      struct file *out, loff_t __user *off_out,
		      size_t len, unsigned int flags)
1381
{
1382 1383
	struct pipe_inode_info *ipipe;
	struct pipe_inode_info *opipe;
1384
	loff_t offset, *off;
J
Jens Axboe 已提交
1385
	long ret;
1386

1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407
	ipipe = pipe_info(in->f_path.dentry->d_inode);
	opipe = pipe_info(out->f_path.dentry->d_inode);

	if (ipipe && opipe) {
		if (off_in || off_out)
			return -ESPIPE;

		if (!(in->f_mode & FMODE_READ))
			return -EBADF;

		if (!(out->f_mode & FMODE_WRITE))
			return -EBADF;

		/* Splicing to self would be fun, but... */
		if (ipipe == opipe)
			return -EINVAL;

		return splice_pipe_to_pipe(ipipe, opipe, len, flags);
	}

	if (ipipe) {
1408 1409
		if (off_in)
			return -ESPIPE;
1410 1411 1412
		if (off_out) {
			if (out->f_op->llseek == no_llseek)
				return -EINVAL;
1413
			if (copy_from_user(&offset, off_out, sizeof(loff_t)))
1414
				return -EFAULT;
1415 1416 1417
			off = &offset;
		} else
			off = &out->f_pos;
1418

1419
		ret = do_splice_from(ipipe, out, off, len, flags);
J
Jens Axboe 已提交
1420 1421 1422 1423 1424

		if (off_out && copy_to_user(off_out, off, sizeof(loff_t)))
			ret = -EFAULT;

		return ret;
1425
	}
1426

1427
	if (opipe) {
1428 1429
		if (off_out)
			return -ESPIPE;
1430 1431 1432
		if (off_in) {
			if (in->f_op->llseek == no_llseek)
				return -EINVAL;
1433
			if (copy_from_user(&offset, off_in, sizeof(loff_t)))
1434
				return -EFAULT;
1435 1436 1437
			off = &offset;
		} else
			off = &in->f_pos;
1438

1439
		ret = do_splice_to(in, off, opipe, len, flags);
J
Jens Axboe 已提交
1440 1441 1442 1443 1444

		if (off_in && copy_to_user(off_in, off, sizeof(loff_t)))
			ret = -EFAULT;

		return ret;
1445
	}
1446 1447 1448 1449

	return -EINVAL;
}

1450 1451 1452 1453 1454 1455 1456 1457 1458
/*
 * Map an iov into an array of pages and offset/length tupples. With the
 * partial_page structure, we can map several non-contiguous ranges into
 * our ones pages[] map instead of splitting that operation into pieces.
 * Could easily be exported as a generic helper for other users, in which
 * case one would probably want to add a 'max_nr_pages' parameter as well.
 */
static int get_iovec_page_array(const struct iovec __user *iov,
				unsigned int nr_vecs, struct page **pages,
1459
				struct partial_page *partial, int aligned)
1460 1461 1462 1463 1464
{
	int buffers = 0, error = 0;

	while (nr_vecs) {
		unsigned long off, npages;
1465
		struct iovec entry;
1466 1467 1468 1469
		void __user *base;
		size_t len;
		int i;

1470
		error = -EFAULT;
N
Nick Piggin 已提交
1471
		if (copy_from_user(&entry, iov, sizeof(entry)))
1472 1473
			break;

1474 1475 1476
		base = entry.iov_base;
		len = entry.iov_len;

1477 1478 1479
		/*
		 * Sanity check this iovec. 0 read succeeds.
		 */
1480
		error = 0;
1481 1482 1483
		if (unlikely(!len))
			break;
		error = -EFAULT;
1484
		if (!access_ok(VERIFY_READ, base, len))
1485 1486 1487 1488 1489 1490 1491
			break;

		/*
		 * Get this base offset and number of pages, then map
		 * in the user pages.
		 */
		off = (unsigned long) base & ~PAGE_MASK;
1492 1493 1494 1495 1496 1497 1498 1499 1500

		/*
		 * If asked for alignment, the offset must be zero and the
		 * length a multiple of the PAGE_SIZE.
		 */
		error = -EINVAL;
		if (aligned && (off || len & ~PAGE_MASK))
			break;

1501 1502 1503 1504
		npages = (off + len + PAGE_SIZE - 1) >> PAGE_SHIFT;
		if (npages > PIPE_BUFFERS - buffers)
			npages = PIPE_BUFFERS - buffers;

N
Nick Piggin 已提交
1505 1506
		error = get_user_pages_fast((unsigned long)base, npages,
					0, &pages[buffers]);
1507 1508 1509 1510 1511 1512 1513 1514

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

		/*
		 * Fill this contiguous range into the partial page map.
		 */
		for (i = 0; i < error; i++) {
1515
			const int plen = min_t(size_t, len, PAGE_SIZE - off);
1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550

			partial[buffers].offset = off;
			partial[buffers].len = plen;

			off = 0;
			len -= plen;
			buffers++;
		}

		/*
		 * We didn't complete this iov, stop here since it probably
		 * means we have to move some of this into a pipe to
		 * be able to continue.
		 */
		if (len)
			break;

		/*
		 * Don't continue if we mapped fewer pages than we asked for,
		 * or if we mapped the max number of pages that we have
		 * room for.
		 */
		if (error < npages || buffers == PIPE_BUFFERS)
			break;

		nr_vecs--;
		iov++;
	}

	if (buffers)
		return buffers;

	return error;
}

1551 1552 1553 1554 1555 1556
static int pipe_to_user(struct pipe_inode_info *pipe, struct pipe_buffer *buf,
			struct splice_desc *sd)
{
	char *src;
	int ret;

1557
	ret = buf->ops->confirm(pipe, buf);
1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584
	if (unlikely(ret))
		return ret;

	/*
	 * See if we can use the atomic maps, by prefaulting in the
	 * pages and doing an atomic copy
	 */
	if (!fault_in_pages_writeable(sd->u.userptr, sd->len)) {
		src = buf->ops->map(pipe, buf, 1);
		ret = __copy_to_user_inatomic(sd->u.userptr, src + buf->offset,
							sd->len);
		buf->ops->unmap(pipe, buf, src);
		if (!ret) {
			ret = sd->len;
			goto out;
		}
	}

	/*
	 * No dice, use slow non-atomic map and copy
 	 */
	src = buf->ops->map(pipe, buf, 0);

	ret = sd->len;
	if (copy_to_user(sd->u.userptr, src + buf->offset, sd->len))
		ret = -EFAULT;

1585
	buf->ops->unmap(pipe, buf, src);
1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608
out:
	if (ret > 0)
		sd->u.userptr += ret;
	return ret;
}

/*
 * For lack of a better implementation, implement vmsplice() to userspace
 * as a simple copy of the pipes pages to the user iov.
 */
static long vmsplice_to_user(struct file *file, const struct iovec __user *iov,
			     unsigned long nr_segs, unsigned int flags)
{
	struct pipe_inode_info *pipe;
	struct splice_desc sd;
	ssize_t size;
	int error;
	long ret;

	pipe = pipe_info(file->f_path.dentry->d_inode);
	if (!pipe)
		return -EBADF;

1609
	pipe_lock(pipe);
1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635

	error = ret = 0;
	while (nr_segs) {
		void __user *base;
		size_t len;

		/*
		 * Get user address base and length for this iovec.
		 */
		error = get_user(base, &iov->iov_base);
		if (unlikely(error))
			break;
		error = get_user(len, &iov->iov_len);
		if (unlikely(error))
			break;

		/*
		 * Sanity check this iovec. 0 read succeeds.
		 */
		if (unlikely(!len))
			break;
		if (unlikely(!base)) {
			error = -EFAULT;
			break;
		}

1636 1637 1638 1639 1640
		if (unlikely(!access_ok(VERIFY_WRITE, base, len))) {
			error = -EFAULT;
			break;
		}

1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663
		sd.len = 0;
		sd.total_len = len;
		sd.flags = flags;
		sd.u.userptr = base;
		sd.pos = 0;

		size = __splice_from_pipe(pipe, &sd, pipe_to_user);
		if (size < 0) {
			if (!ret)
				ret = size;

			break;
		}

		ret += size;

		if (size < len)
			break;

		nr_segs--;
		iov++;
	}

1664
	pipe_unlock(pipe);
1665 1666 1667 1668 1669 1670 1671

	if (!ret)
		ret = error;

	return ret;
}

1672 1673 1674 1675 1676
/*
 * vmsplice splices a user address range into a pipe. It can be thought of
 * as splice-from-memory, where the regular splice is splice-from-file (or
 * to file). In both cases the output is a pipe, naturally.
 */
1677 1678
static long vmsplice_to_pipe(struct file *file, const struct iovec __user *iov,
			     unsigned long nr_segs, unsigned int flags)
1679
{
1680
	struct pipe_inode_info *pipe;
1681 1682 1683 1684 1685 1686 1687
	struct page *pages[PIPE_BUFFERS];
	struct partial_page partial[PIPE_BUFFERS];
	struct splice_pipe_desc spd = {
		.pages = pages,
		.partial = partial,
		.flags = flags,
		.ops = &user_page_pipe_buf_ops,
1688
		.spd_release = spd_release_page,
1689 1690
	};

1691
	pipe = pipe_info(file->f_path.dentry->d_inode);
1692
	if (!pipe)
1693 1694
		return -EBADF;

1695 1696
	spd.nr_pages = get_iovec_page_array(iov, nr_segs, pages, partial,
					    flags & SPLICE_F_GIFT);
1697 1698 1699
	if (spd.nr_pages <= 0)
		return spd.nr_pages;

1700
	return splice_to_pipe(pipe, &spd);
1701 1702
}

1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718
/*
 * Note that vmsplice only really supports true splicing _from_ user memory
 * to a pipe, not the other way around. Splicing from user memory is a simple
 * operation that can be supported without any funky alignment restrictions
 * or nasty vm tricks. We simply map in the user memory and fill them into
 * a pipe. The reverse isn't quite as easy, though. There are two possible
 * solutions for that:
 *
 *	- memcpy() the data internally, at which point we might as well just
 *	  do a regular read() on the buffer anyway.
 *	- Lots of nasty vm tricks, that are neither fast nor flexible (it
 *	  has restriction limitations on both ends of the pipe).
 *
 * Currently we punt and implement it as a normal copy, see pipe_to_user().
 *
 */
1719 1720
SYSCALL_DEFINE4(vmsplice, int, fd, const struct iovec __user *, iov,
		unsigned long, nr_segs, unsigned int, flags)
1721 1722 1723 1724 1725
{
	struct file *file;
	long error;
	int fput;

1726 1727 1728 1729 1730
	if (unlikely(nr_segs > UIO_MAXIOV))
		return -EINVAL;
	else if (unlikely(!nr_segs))
		return 0;

1731 1732 1733 1734
	error = -EBADF;
	file = fget_light(fd, &fput);
	if (file) {
		if (file->f_mode & FMODE_WRITE)
1735 1736 1737
			error = vmsplice_to_pipe(file, iov, nr_segs, flags);
		else if (file->f_mode & FMODE_READ)
			error = vmsplice_to_user(file, iov, nr_segs, flags);
1738 1739 1740 1741 1742 1743 1744

		fput_light(file, fput);
	}

	return error;
}

1745 1746 1747
SYSCALL_DEFINE6(splice, int, fd_in, loff_t __user *, off_in,
		int, fd_out, loff_t __user *, off_out,
		size_t, len, unsigned int, flags)
1748 1749 1750 1751 1752 1753 1754 1755 1756
{
	long error;
	struct file *in, *out;
	int fput_in, fput_out;

	if (unlikely(!len))
		return 0;

	error = -EBADF;
1757
	in = fget_light(fd_in, &fput_in);
1758 1759
	if (in) {
		if (in->f_mode & FMODE_READ) {
1760
			out = fget_light(fd_out, &fput_out);
1761 1762
			if (out) {
				if (out->f_mode & FMODE_WRITE)
1763 1764 1765
					error = do_splice(in, off_in,
							  out, off_out,
							  len, flags);
1766 1767 1768 1769 1770 1771 1772 1773 1774
				fput_light(out, fput_out);
			}
		}

		fput_light(in, fput_in);
	}

	return error;
}
1775

1776 1777 1778 1779
/*
 * Make sure there's data to read. Wait for input if we can, otherwise
 * return an appropriate error.
 */
1780
static int ipipe_prep(struct pipe_inode_info *pipe, unsigned int flags)
1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791
{
	int ret;

	/*
	 * Check ->nrbufs without the inode lock first. This function
	 * is speculative anyways, so missing one is ok.
	 */
	if (pipe->nrbufs)
		return 0;

	ret = 0;
1792
	pipe_lock(pipe);
1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809

	while (!pipe->nrbufs) {
		if (signal_pending(current)) {
			ret = -ERESTARTSYS;
			break;
		}
		if (!pipe->writers)
			break;
		if (!pipe->waiting_writers) {
			if (flags & SPLICE_F_NONBLOCK) {
				ret = -EAGAIN;
				break;
			}
		}
		pipe_wait(pipe);
	}

1810
	pipe_unlock(pipe);
1811 1812 1813 1814 1815 1816 1817
	return ret;
}

/*
 * Make sure there's writeable room. Wait for room if we can, otherwise
 * return an appropriate error.
 */
1818
static int opipe_prep(struct pipe_inode_info *pipe, unsigned int flags)
1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829
{
	int ret;

	/*
	 * Check ->nrbufs without the inode lock first. This function
	 * is speculative anyways, so missing one is ok.
	 */
	if (pipe->nrbufs < PIPE_BUFFERS)
		return 0;

	ret = 0;
1830
	pipe_lock(pipe);
1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850

	while (pipe->nrbufs >= PIPE_BUFFERS) {
		if (!pipe->readers) {
			send_sig(SIGPIPE, current, 0);
			ret = -EPIPE;
			break;
		}
		if (flags & SPLICE_F_NONBLOCK) {
			ret = -EAGAIN;
			break;
		}
		if (signal_pending(current)) {
			ret = -ERESTARTSYS;
			break;
		}
		pipe->waiting_writers++;
		pipe_wait(pipe);
		pipe->waiting_writers--;
	}

1851
	pipe_unlock(pipe);
1852 1853 1854
	return ret;
}

1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972
/*
 * Splice contents of ipipe to opipe.
 */
static int splice_pipe_to_pipe(struct pipe_inode_info *ipipe,
			       struct pipe_inode_info *opipe,
			       size_t len, unsigned int flags)
{
	struct pipe_buffer *ibuf, *obuf;
	int ret = 0, nbuf;
	bool input_wakeup = false;


retry:
	ret = ipipe_prep(ipipe, flags);
	if (ret)
		return ret;

	ret = opipe_prep(opipe, flags);
	if (ret)
		return ret;

	/*
	 * Potential ABBA deadlock, work around it by ordering lock
	 * grabbing by pipe info address. Otherwise two different processes
	 * could deadlock (one doing tee from A -> B, the other from B -> A).
	 */
	pipe_double_lock(ipipe, opipe);

	do {
		if (!opipe->readers) {
			send_sig(SIGPIPE, current, 0);
			if (!ret)
				ret = -EPIPE;
			break;
		}

		if (!ipipe->nrbufs && !ipipe->writers)
			break;

		/*
		 * Cannot make any progress, because either the input
		 * pipe is empty or the output pipe is full.
		 */
		if (!ipipe->nrbufs || opipe->nrbufs >= PIPE_BUFFERS) {
			/* Already processed some buffers, break */
			if (ret)
				break;

			if (flags & SPLICE_F_NONBLOCK) {
				ret = -EAGAIN;
				break;
			}

			/*
			 * We raced with another reader/writer and haven't
			 * managed to process any buffers.  A zero return
			 * value means EOF, so retry instead.
			 */
			pipe_unlock(ipipe);
			pipe_unlock(opipe);
			goto retry;
		}

		ibuf = ipipe->bufs + ipipe->curbuf;
		nbuf = (opipe->curbuf + opipe->nrbufs) % PIPE_BUFFERS;
		obuf = opipe->bufs + nbuf;

		if (len >= ibuf->len) {
			/*
			 * Simply move the whole buffer from ipipe to opipe
			 */
			*obuf = *ibuf;
			ibuf->ops = NULL;
			opipe->nrbufs++;
			ipipe->curbuf = (ipipe->curbuf + 1) % PIPE_BUFFERS;
			ipipe->nrbufs--;
			input_wakeup = true;
		} else {
			/*
			 * Get a reference to this pipe buffer,
			 * so we can copy the contents over.
			 */
			ibuf->ops->get(ipipe, ibuf);
			*obuf = *ibuf;

			/*
			 * Don't inherit the gift flag, we need to
			 * prevent multiple steals of this page.
			 */
			obuf->flags &= ~PIPE_BUF_FLAG_GIFT;

			obuf->len = len;
			opipe->nrbufs++;
			ibuf->offset += obuf->len;
			ibuf->len -= obuf->len;
		}
		ret += obuf->len;
		len -= obuf->len;
	} while (len);

	pipe_unlock(ipipe);
	pipe_unlock(opipe);

	/*
	 * If we put data in the output pipe, wakeup any potential readers.
	 */
	if (ret > 0) {
		smp_mb();
		if (waitqueue_active(&opipe->wait))
			wake_up_interruptible(&opipe->wait);
		kill_fasync(&opipe->fasync_readers, SIGIO, POLL_IN);
	}
	if (input_wakeup)
		wakeup_pipe_writers(ipipe);

	return ret;
}

1973 1974 1975 1976 1977 1978 1979 1980
/*
 * Link contents of ipipe to opipe.
 */
static int link_pipe(struct pipe_inode_info *ipipe,
		     struct pipe_inode_info *opipe,
		     size_t len, unsigned int flags)
{
	struct pipe_buffer *ibuf, *obuf;
1981
	int ret = 0, i = 0, nbuf;
1982 1983 1984

	/*
	 * Potential ABBA deadlock, work around it by ordering lock
1985
	 * grabbing by pipe info address. Otherwise two different processes
1986 1987
	 * could deadlock (one doing tee from A -> B, the other from B -> A).
	 */
1988
	pipe_double_lock(ipipe, opipe);
1989

1990
	do {
1991 1992 1993 1994 1995 1996 1997
		if (!opipe->readers) {
			send_sig(SIGPIPE, current, 0);
			if (!ret)
				ret = -EPIPE;
			break;
		}

1998 1999 2000 2001 2002 2003
		/*
		 * If we have iterated all input buffers or ran out of
		 * output room, break.
		 */
		if (i >= ipipe->nrbufs || opipe->nrbufs >= PIPE_BUFFERS)
			break;
2004

2005 2006
		ibuf = ipipe->bufs + ((ipipe->curbuf + i) & (PIPE_BUFFERS - 1));
		nbuf = (opipe->curbuf + opipe->nrbufs) & (PIPE_BUFFERS - 1);
2007 2008

		/*
2009 2010
		 * Get a reference to this pipe buffer,
		 * so we can copy the contents over.
2011
		 */
2012 2013 2014 2015 2016
		ibuf->ops->get(ipipe, ibuf);

		obuf = opipe->bufs + nbuf;
		*obuf = *ibuf;

2017
		/*
2018 2019
		 * Don't inherit the gift flag, we need to
		 * prevent multiple steals of this page.
2020
		 */
2021
		obuf->flags &= ~PIPE_BUF_FLAG_GIFT;
2022

2023 2024
		if (obuf->len > len)
			obuf->len = len;
2025

2026 2027 2028 2029 2030
		opipe->nrbufs++;
		ret += obuf->len;
		len -= obuf->len;
		i++;
	} while (len);
2031

2032 2033 2034 2035 2036 2037 2038
	/*
	 * return EAGAIN if we have the potential of some data in the
	 * future, otherwise just return 0
	 */
	if (!ret && ipipe->waiting_writers && (flags & SPLICE_F_NONBLOCK))
		ret = -EAGAIN;

2039 2040
	pipe_unlock(ipipe);
	pipe_unlock(opipe);
2041

2042 2043 2044 2045
	/*
	 * If we put data in the output pipe, wakeup any potential readers.
	 */
	if (ret > 0) {
2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063
		smp_mb();
		if (waitqueue_active(&opipe->wait))
			wake_up_interruptible(&opipe->wait);
		kill_fasync(&opipe->fasync_readers, SIGIO, POLL_IN);
	}

	return ret;
}

/*
 * This is a tee(1) implementation that works on pipes. It doesn't copy
 * any data, it simply references the 'in' pages on the 'out' pipe.
 * The 'flags' used are the SPLICE_F_* variants, currently the only
 * applicable one is SPLICE_F_NONBLOCK.
 */
static long do_tee(struct file *in, struct file *out, size_t len,
		   unsigned int flags)
{
2064 2065
	struct pipe_inode_info *ipipe = pipe_info(in->f_path.dentry->d_inode);
	struct pipe_inode_info *opipe = pipe_info(out->f_path.dentry->d_inode);
2066
	int ret = -EINVAL;
2067 2068

	/*
2069 2070
	 * Duplicate the contents of ipipe to opipe without actually
	 * copying the data.
2071
	 */
2072 2073 2074 2075 2076
	if (ipipe && opipe && ipipe != opipe) {
		/*
		 * Keep going, unless we encounter an error. The ipipe/opipe
		 * ordering doesn't really matter.
		 */
2077
		ret = ipipe_prep(ipipe, flags);
2078
		if (!ret) {
2079
			ret = opipe_prep(opipe, flags);
2080
			if (!ret)
2081 2082 2083
				ret = link_pipe(ipipe, opipe, len, flags);
		}
	}
2084

2085
	return ret;
2086 2087
}

2088
SYSCALL_DEFINE4(tee, int, fdin, int, fdout, size_t, len, unsigned int, flags)
2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113
{
	struct file *in;
	int error, fput_in;

	if (unlikely(!len))
		return 0;

	error = -EBADF;
	in = fget_light(fdin, &fput_in);
	if (in) {
		if (in->f_mode & FMODE_READ) {
			int fput_out;
			struct file *out = fget_light(fdout, &fput_out);

			if (out) {
				if (out->f_mode & FMODE_WRITE)
					error = do_tee(in, out, len, flags);
				fput_light(out, fput_out);
			}
		}
 		fput_light(in, fput_in);
 	}

	return error;
}