bitmap.c 55.8 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
/*
 * bitmap.c two-level bitmap (C) Peter T. Breuer (ptb@ot.uc3m.es) 2003
 *
 * bitmap_create  - sets up the bitmap structure
 * bitmap_destroy - destroys the bitmap structure
 *
 * additions, Copyright (C) 2003-2004, Paul Clements, SteelEye Technology, Inc.:
 * - added disk storage for bitmap
 * - changes to allow various bitmap chunk sizes
 */

/*
 * Still to do:
 *
 * flush after percent set rather than just time based. (maybe both).
 */

18
#include <linux/blkdev.h>
19 20 21 22 23 24 25 26 27 28
#include <linux/module.h>
#include <linux/errno.h>
#include <linux/slab.h>
#include <linux/init.h>
#include <linux/timer.h>
#include <linux/sched.h>
#include <linux/list.h>
#include <linux/file.h>
#include <linux/mount.h>
#include <linux/buffer_head.h>
29
#include <linux/seq_file.h>
30
#include "md.h"
31
#include "bitmap.h"
32

33
static inline char *bmname(struct bitmap *bitmap)
34 35 36 37 38 39 40 41 42 43 44 45 46 47
{
	return bitmap->mddev ? mdname(bitmap->mddev) : "mdX";
}

/*
 * check a page and, if necessary, allocate it (or hijack it if the alloc fails)
 *
 * 1) check to see if this page is allocated, if it's not then try to alloc
 * 2) if the alloc fails, set the page's hijacked flag so we'll use the
 *    page pointer directly as a counter
 *
 * if we find our page, we increment the page's refcount so that it stays
 * allocated while we're using it
 */
48 49
static int bitmap_checkpage(struct bitmap *bitmap,
			    unsigned long page, int create)
50 51
__releases(bitmap->lock)
__acquires(bitmap->lock)
52 53 54 55
{
	unsigned char *mappage;

	if (page >= bitmap->pages) {
56 57 58 59
		/* This can happen if bitmap_start_sync goes beyond
		 * End-of-device while looking for a whole page.
		 * It is harmless.
		 */
60 61 62 63 64 65 66 67 68 69 70 71 72 73
		return -EINVAL;
	}

	if (bitmap->bp[page].hijacked) /* it's hijacked, don't try to alloc */
		return 0;

	if (bitmap->bp[page].map) /* page is already allocated, just return */
		return 0;

	if (!create)
		return -ENOENT;

	/* this page has not been allocated yet */

74
	spin_unlock_irq(&bitmap->lock);
75
	mappage = kzalloc(PAGE_SIZE, GFP_NOIO);
76 77 78
	spin_lock_irq(&bitmap->lock);

	if (mappage == NULL) {
79 80
		pr_debug("%s: bitmap map page allocation failed, hijacking\n",
			 bmname(bitmap));
81 82 83 84
		/* failed - set the hijacked flag so that we can use the
		 * pointer as a counter */
		if (!bitmap->bp[page].map)
			bitmap->bp[page].hijacked = 1;
85 86
	} else if (bitmap->bp[page].map ||
		   bitmap->bp[page].hijacked) {
87
		/* somebody beat us to getting the page */
88
		kfree(mappage);
89
		return 0;
90
	} else {
91

92
		/* no page was in place and we have one, so install it */
93

94 95 96
		bitmap->bp[page].map = mappage;
		bitmap->missing_pages--;
	}
97 98 99 100 101 102
	return 0;
}

/* if page is completely empty, put it back on the free list, or dealloc it */
/* if page was hijacked, unmark the flag so it might get alloced next time */
/* Note: lock should be held when calling this */
103
static void bitmap_checkfree(struct bitmap *bitmap, unsigned long page)
104 105 106 107 108 109 110 111 112 113 114
{
	char *ptr;

	if (bitmap->bp[page].count) /* page is still busy */
		return;

	/* page is no longer in use, it can be released */

	if (bitmap->bp[page].hijacked) { /* page was hijacked, undo this now */
		bitmap->bp[page].hijacked = 0;
		bitmap->bp[page].map = NULL;
115 116 117 118 119
	} else {
		/* normal case, free the page */
		ptr = bitmap->bp[page].map;
		bitmap->bp[page].map = NULL;
		bitmap->missing_pages++;
120
		kfree(ptr);
121 122 123 124 125 126 127 128 129 130 131
	}
}

/*
 * bitmap file handling - read and write the bitmap file and its superblock
 */

/*
 * basic page I/O operations
 */

132
/* IO operations when bitmap is stored near all superblocks */
133 134 135
static int read_sb_page(struct mddev *mddev, loff_t offset,
			struct page *page,
			unsigned long index, int size)
136 137 138
{
	/* choose a good rdev and read the page from there */

139
	struct md_rdev *rdev;
140 141
	sector_t target;

N
NeilBrown 已提交
142
	rdev_for_each(rdev, mddev) {
143 144
		if (! test_bit(In_sync, &rdev->flags)
		    || test_bit(Faulty, &rdev->flags))
145 146
			continue;

J
Jonathan Brassow 已提交
147
		target = offset + index * (PAGE_SIZE/512);
148

149
		if (sync_page_io(rdev, target,
150
				 roundup(size, bdev_logical_block_size(rdev->bdev)),
J
Jonathan Brassow 已提交
151
				 page, READ, true)) {
152
			page->index = index;
153
			return 0;
154 155
		}
	}
156
	return -EIO;
157 158
}

159
static struct md_rdev *next_active_rdev(struct md_rdev *rdev, struct mddev *mddev)
160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179
{
	/* Iterate the disks of an mddev, using rcu to protect access to the
	 * linked list, and raising the refcount of devices we return to ensure
	 * they don't disappear while in use.
	 * As devices are only added or removed when raid_disk is < 0 and
	 * nr_pending is 0 and In_sync is clear, the entries we return will
	 * still be in the same position on the list when we re-enter
	 * list_for_each_continue_rcu.
	 */
	struct list_head *pos;
	rcu_read_lock();
	if (rdev == NULL)
		/* start at the beginning */
		pos = &mddev->disks;
	else {
		/* release the previous rdev and start from there. */
		rdev_dec_pending(rdev, mddev);
		pos = &rdev->same_set;
	}
	list_for_each_continue_rcu(pos, &mddev->disks) {
180
		rdev = list_entry(pos, struct md_rdev, same_set);
181 182 183 184 185 186 187 188 189 190 191 192
		if (rdev->raid_disk >= 0 &&
		    !test_bit(Faulty, &rdev->flags)) {
			/* this is a usable devices */
			atomic_inc(&rdev->nr_pending);
			rcu_read_unlock();
			return rdev;
		}
	}
	rcu_read_unlock();
	return NULL;
}

193
static int write_sb_page(struct bitmap *bitmap, struct page *page, int wait)
194
{
195
	struct md_rdev *rdev = NULL;
196
	struct block_device *bdev;
197
	struct mddev *mddev = bitmap->mddev;
198
	struct bitmap_storage *store = &bitmap->storage;
199

200
	while ((rdev = next_active_rdev(rdev, mddev)) != NULL) {
201 202
		int size = PAGE_SIZE;
		loff_t offset = mddev->bitmap_info.offset;
203 204 205

		bdev = (rdev->meta_bdev) ? rdev->meta_bdev : rdev->bdev;

206 207 208 209 210
		if (page->index == store->file_pages-1) {
			int last_page_size = store->bytes & (PAGE_SIZE-1);
			if (last_page_size == 0)
				last_page_size = PAGE_SIZE;
			size = roundup(last_page_size,
211
				       bdev_logical_block_size(bdev));
212
		}
213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252
		/* Just make sure we aren't corrupting data or
		 * metadata
		 */
		if (mddev->external) {
			/* Bitmap could be anywhere. */
			if (rdev->sb_start + offset + (page->index
						       * (PAGE_SIZE/512))
			    > rdev->data_offset
			    &&
			    rdev->sb_start + offset
			    < (rdev->data_offset + mddev->dev_sectors
			     + (PAGE_SIZE/512)))
				goto bad_alignment;
		} else if (offset < 0) {
			/* DATA  BITMAP METADATA  */
			if (offset
			    + (long)(page->index * (PAGE_SIZE/512))
			    + size/512 > 0)
				/* bitmap runs in to metadata */
				goto bad_alignment;
			if (rdev->data_offset + mddev->dev_sectors
			    > rdev->sb_start + offset)
				/* data runs in to bitmap */
				goto bad_alignment;
		} else if (rdev->sb_start < rdev->data_offset) {
			/* METADATA BITMAP DATA */
			if (rdev->sb_start
			    + offset
			    + page->index*(PAGE_SIZE/512) + size/512
			    > rdev->data_offset)
				/* bitmap runs in to data */
				goto bad_alignment;
		} else {
			/* DATA METADATA BITMAP - no problems */
		}
		md_super_write(mddev, rdev,
			       rdev->sb_start + offset
			       + page->index * (PAGE_SIZE/512),
			       size,
			       page);
253
	}
254 255

	if (wait)
256
		md_super_wait(mddev);
257
	return 0;
258 259 260

 bad_alignment:
	return -EINVAL;
261 262
}

263
static void bitmap_file_kick(struct bitmap *bitmap);
264
/*
265
 * write out a page to a file
266
 */
267
static void write_page(struct bitmap *bitmap, struct page *page, int wait)
268
{
269
	struct buffer_head *bh;
270

271
	if (bitmap->storage.file == NULL) {
272 273 274 275
		switch (write_sb_page(bitmap, page, wait)) {
		case -EINVAL:
			bitmap->flags |= BITMAP_WRITE_ERROR;
		}
276
	} else {
277

278
		bh = page_buffers(page);
279

280 281 282 283
		while (bh && bh->b_blocknr) {
			atomic_inc(&bitmap->pending_writes);
			set_buffer_locked(bh);
			set_buffer_mapped(bh);
J
Jens Axboe 已提交
284
			submit_bh(WRITE | REQ_SYNC, bh);
285 286
			bh = bh->b_this_page;
		}
287

288
		if (wait)
289 290
			wait_event(bitmap->write_wait,
				   atomic_read(&bitmap->pending_writes)==0);
291
	}
292 293
	if (bitmap->flags & BITMAP_WRITE_ERROR)
		bitmap_file_kick(bitmap);
294 295 296 297 298 299
}

static void end_bitmap_write(struct buffer_head *bh, int uptodate)
{
	struct bitmap *bitmap = bh->b_private;
	unsigned long flags;
300

301 302 303 304
	if (!uptodate) {
		spin_lock_irqsave(&bitmap->lock, flags);
		bitmap->flags |= BITMAP_WRITE_ERROR;
		spin_unlock_irqrestore(&bitmap->lock, flags);
305
	}
306 307 308
	if (atomic_dec_and_test(&bitmap->pending_writes))
		wake_up(&bitmap->write_wait);
}
309

310 311 312 313 314 315 316 317 318 319
/* copied from buffer.c */
static void
__clear_page_buffers(struct page *page)
{
	ClearPagePrivate(page);
	set_page_private(page, 0);
	page_cache_release(page);
}
static void free_buffers(struct page *page)
{
320
	struct buffer_head *bh;
321

322 323 324 325
	if (!PagePrivate(page))
		return;

	bh = page_buffers(page);
326 327 328 329
	while (bh) {
		struct buffer_head *next = bh->b_this_page;
		free_buffer_head(bh);
		bh = next;
330
	}
331 332
	__clear_page_buffers(page);
	put_page(page);
333 334
}

335 336 337 338 339 340 341
/* read a page from a file.
 * We both read the page, and attach buffers to the page to record the
 * address of each block (using bmap).  These addresses will be used
 * to write the block later, completely bypassing the filesystem.
 * This usage is similar to how swap files are handled, and allows us
 * to write to a file with no concerns of memory allocation failing.
 */
342 343 344 345
static int read_page(struct file *file, unsigned long index,
		     struct bitmap *bitmap,
		     unsigned long count,
		     struct page *page)
346
{
347
	int ret = 0;
J
Josef Sipek 已提交
348
	struct inode *inode = file->f_path.dentry->d_inode;
349 350
	struct buffer_head *bh;
	sector_t block;
351

352 353
	pr_debug("read bitmap file (%dB @ %llu)\n", (int)PAGE_SIZE,
		 (unsigned long long)index << PAGE_SHIFT);
354

355 356
	bh = alloc_page_buffers(page, 1<<inode->i_blkbits, 0);
	if (!bh) {
357
		ret = -ENOMEM;
358 359
		goto out;
	}
360 361 362 363 364 365 366 367 368
	attach_page_buffers(page, bh);
	block = index << (PAGE_SHIFT - inode->i_blkbits);
	while (bh) {
		if (count == 0)
			bh->b_blocknr = 0;
		else {
			bh->b_blocknr = bmap(inode, block);
			if (bh->b_blocknr == 0) {
				/* Cannot use this file! */
369
				ret = -EINVAL;
370 371 372 373 374 375 376 377 378 379
				goto out;
			}
			bh->b_bdev = inode->i_sb->s_bdev;
			if (count < (1<<inode->i_blkbits))
				count = 0;
			else
				count -= (1<<inode->i_blkbits);

			bh->b_end_io = end_bitmap_write;
			bh->b_private = bitmap;
380 381 382 383
			atomic_inc(&bitmap->pending_writes);
			set_buffer_locked(bh);
			set_buffer_mapped(bh);
			submit_bh(READ, bh);
384 385 386 387 388
		}
		block++;
		bh = bh->b_this_page;
	}
	page->index = index;
389 390 391

	wait_event(bitmap->write_wait,
		   atomic_read(&bitmap->pending_writes)==0);
392 393
	if (bitmap->flags & BITMAP_WRITE_ERROR)
		ret = -EIO;
394
out:
395 396
	if (ret)
		printk(KERN_ALERT "md: bitmap read error: (%dB @ %llu): %d\n",
397 398
			(int)PAGE_SIZE,
			(unsigned long long)index << PAGE_SHIFT,
399 400
			ret);
	return ret;
401 402 403 404 405 406 407
}

/*
 * bitmap file superblock operations
 */

/* update the event counter and sync the superblock to disk */
408
void bitmap_update_sb(struct bitmap *bitmap)
409 410 411 412
{
	bitmap_super_t *sb;

	if (!bitmap || !bitmap->mddev) /* no bitmap for this array */
413
		return;
414 415
	if (bitmap->mddev->bitmap_info.external)
		return;
416
	if (!bitmap->storage.sb_page) /* no superblock */
417
		return;
418
	sb = kmap_atomic(bitmap->storage.sb_page);
419
	sb->events = cpu_to_le64(bitmap->mddev->events);
420
	if (bitmap->mddev->events < bitmap->events_cleared)
421 422
		/* rocking back to read-only */
		bitmap->events_cleared = bitmap->mddev->events;
423 424
	sb->events_cleared = cpu_to_le64(bitmap->events_cleared);
	sb->state = cpu_to_le32(bitmap->flags);
425 426 427
	/* Just in case these have been changed via sysfs: */
	sb->daemon_sleep = cpu_to_le32(bitmap->mddev->bitmap_info.daemon_sleep/HZ);
	sb->write_behind = cpu_to_le32(bitmap->mddev->bitmap_info.max_write_behind);
428
	kunmap_atomic(sb);
429
	write_page(bitmap, bitmap->storage.sb_page, 1);
430 431 432 433 434 435 436
}

/* print out the bitmap file superblock */
void bitmap_print_sb(struct bitmap *bitmap)
{
	bitmap_super_t *sb;

437
	if (!bitmap || !bitmap->storage.sb_page)
438
		return;
439
	sb = kmap_atomic(bitmap->storage.sb_page);
440
	printk(KERN_DEBUG "%s: bitmap file superblock:\n", bmname(bitmap));
441 442 443
	printk(KERN_DEBUG "         magic: %08x\n", le32_to_cpu(sb->magic));
	printk(KERN_DEBUG "       version: %d\n", le32_to_cpu(sb->version));
	printk(KERN_DEBUG "          uuid: %08x.%08x.%08x.%08x\n",
444 445 446 447
					*(__u32 *)(sb->uuid+0),
					*(__u32 *)(sb->uuid+4),
					*(__u32 *)(sb->uuid+8),
					*(__u32 *)(sb->uuid+12));
448
	printk(KERN_DEBUG "        events: %llu\n",
449
			(unsigned long long) le64_to_cpu(sb->events));
450
	printk(KERN_DEBUG "events cleared: %llu\n",
451
			(unsigned long long) le64_to_cpu(sb->events_cleared));
452 453 454 455 456
	printk(KERN_DEBUG "         state: %08x\n", le32_to_cpu(sb->state));
	printk(KERN_DEBUG "     chunksize: %d B\n", le32_to_cpu(sb->chunksize));
	printk(KERN_DEBUG "  daemon sleep: %ds\n", le32_to_cpu(sb->daemon_sleep));
	printk(KERN_DEBUG "     sync size: %llu KB\n",
			(unsigned long long)le64_to_cpu(sb->sync_size)/2);
457
	printk(KERN_DEBUG "max write behind: %d\n", le32_to_cpu(sb->write_behind));
458
	kunmap_atomic(sb);
459 460
}

461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477
/*
 * bitmap_new_disk_sb
 * @bitmap
 *
 * This function is somewhat the reverse of bitmap_read_sb.  bitmap_read_sb
 * reads and verifies the on-disk bitmap superblock and populates bitmap_info.
 * This function verifies 'bitmap_info' and populates the on-disk bitmap
 * structure, which is to be written to disk.
 *
 * Returns: 0 on success, -Exxx on error
 */
static int bitmap_new_disk_sb(struct bitmap *bitmap)
{
	bitmap_super_t *sb;
	unsigned long chunksize, daemon_sleep, write_behind;
	int err = -EINVAL;

478 479 480 481
	bitmap->storage.sb_page = alloc_page(GFP_KERNEL);
	if (IS_ERR(bitmap->storage.sb_page)) {
		err = PTR_ERR(bitmap->storage.sb_page);
		bitmap->storage.sb_page = NULL;
482 483
		return err;
	}
484
	bitmap->storage.sb_page->index = 0;
485

486
	sb = kmap_atomic(bitmap->storage.sb_page);
487 488 489 490 491 492 493

	sb->magic = cpu_to_le32(BITMAP_MAGIC);
	sb->version = cpu_to_le32(BITMAP_MAJOR_HI);

	chunksize = bitmap->mddev->bitmap_info.chunksize;
	BUG_ON(!chunksize);
	if (!is_power_of_2(chunksize)) {
494
		kunmap_atomic(sb);
495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524
		printk(KERN_ERR "bitmap chunksize not a power of 2\n");
		return -EINVAL;
	}
	sb->chunksize = cpu_to_le32(chunksize);

	daemon_sleep = bitmap->mddev->bitmap_info.daemon_sleep;
	if (!daemon_sleep ||
	    (daemon_sleep < 1) || (daemon_sleep > MAX_SCHEDULE_TIMEOUT)) {
		printk(KERN_INFO "Choosing daemon_sleep default (5 sec)\n");
		daemon_sleep = 5 * HZ;
	}
	sb->daemon_sleep = cpu_to_le32(daemon_sleep);
	bitmap->mddev->bitmap_info.daemon_sleep = daemon_sleep;

	/*
	 * FIXME: write_behind for RAID1.  If not specified, what
	 * is a good choice?  We choose COUNTER_MAX / 2 arbitrarily.
	 */
	write_behind = bitmap->mddev->bitmap_info.max_write_behind;
	if (write_behind > COUNTER_MAX)
		write_behind = COUNTER_MAX / 2;
	sb->write_behind = cpu_to_le32(write_behind);
	bitmap->mddev->bitmap_info.max_write_behind = write_behind;

	/* keep the array size field of the bitmap superblock up to date */
	sb->sync_size = cpu_to_le64(bitmap->mddev->resync_max_sectors);

	memcpy(sb->uuid, bitmap->mddev->uuid, 16);

	bitmap->flags |= BITMAP_STALE;
525
	sb->state = cpu_to_le32(bitmap->flags);
526 527 528
	bitmap->events_cleared = bitmap->mddev->events;
	sb->events_cleared = cpu_to_le64(bitmap->mddev->events);

529
	kunmap_atomic(sb);
530 531 532 533

	return 0;
}

534 535 536 537 538
/* read the superblock from the bitmap file and initialize some bitmap fields */
static int bitmap_read_sb(struct bitmap *bitmap)
{
	char *reason = NULL;
	bitmap_super_t *sb;
539
	unsigned long chunksize, daemon_sleep, write_behind;
540 541
	unsigned long long events;
	int err = -EINVAL;
542
	struct page *sb_page;
543

544
	if (!bitmap->storage.file && !bitmap->mddev->bitmap_info.offset) {
545 546 547 548 549 550 551
		chunksize = 128 * 1024 * 1024;
		daemon_sleep = 5 * HZ;
		write_behind = 0;
		bitmap->flags = BITMAP_STALE;
		err = 0;
		goto out_no_sb;
	}
552
	/* page 0 is the superblock, read it... */
553 554 555
	sb_page = alloc_page(GFP_KERNEL);
	if (!sb_page)
		return -ENOMEM;
556
	bitmap->storage.sb_page = sb_page;
557

558 559
	if (bitmap->storage.file) {
		loff_t isize = i_size_read(bitmap->storage.file->f_mapping->host);
560 561
		int bytes = isize > PAGE_SIZE ? PAGE_SIZE : isize;

562
		err = read_page(bitmap->storage.file, 0,
563
				bitmap, bytes, sb_page);
564
	} else {
565 566 567 568
		err = read_sb_page(bitmap->mddev,
				   bitmap->mddev->bitmap_info.offset,
				   sb_page,
				   0, sizeof(bitmap_super_t));
569
	}
570
	if (err)
571 572
		return err;

573
	sb = kmap_atomic(sb_page);
574 575

	chunksize = le32_to_cpu(sb->chunksize);
576
	daemon_sleep = le32_to_cpu(sb->daemon_sleep) * HZ;
577
	write_behind = le32_to_cpu(sb->write_behind);
578 579 580 581

	/* verify that the bitmap-specific fields are valid */
	if (sb->magic != cpu_to_le32(BITMAP_MAGIC))
		reason = "bad magic";
582 583
	else if (le32_to_cpu(sb->version) < BITMAP_MAJOR_LO ||
		 le32_to_cpu(sb->version) > BITMAP_MAJOR_HI)
584
		reason = "unrecognized superblock version";
585
	else if (chunksize < 512)
586
		reason = "bitmap chunksize too small";
J
Jonathan Brassow 已提交
587
	else if (!is_power_of_2(chunksize))
588
		reason = "bitmap chunksize not a power of 2";
589
	else if (daemon_sleep < 1 || daemon_sleep > MAX_SCHEDULE_TIMEOUT)
590
		reason = "daemon sleep period out of range";
591 592
	else if (write_behind > COUNTER_MAX)
		reason = "write-behind limit out of range (0 - 16383)";
593 594 595 596 597 598 599 600 601
	if (reason) {
		printk(KERN_INFO "%s: invalid bitmap file superblock: %s\n",
			bmname(bitmap), reason);
		goto out;
	}

	/* keep the array size field of the bitmap superblock up to date */
	sb->sync_size = cpu_to_le64(bitmap->mddev->resync_max_sectors);

602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619
	if (bitmap->mddev->persistent) {
		/*
		 * We have a persistent array superblock, so compare the
		 * bitmap's UUID and event counter to the mddev's
		 */
		if (memcmp(sb->uuid, bitmap->mddev->uuid, 16)) {
			printk(KERN_INFO
			       "%s: bitmap superblock UUID mismatch\n",
			       bmname(bitmap));
			goto out;
		}
		events = le64_to_cpu(sb->events);
		if (events < bitmap->mddev->events) {
			printk(KERN_INFO
			       "%s: bitmap file is out of date (%llu < %llu) "
			       "-- forcing full recovery\n",
			       bmname(bitmap), events,
			       (unsigned long long) bitmap->mddev->events);
620
			bitmap->flags |= BITMAP_STALE;
621
		}
622
	}
623

624
	/* assign fields using values from superblock */
625
	bitmap->flags |= le32_to_cpu(sb->state);
626 627
	if (le32_to_cpu(sb->version) == BITMAP_MAJOR_HOSTENDIAN)
		bitmap->flags |= BITMAP_HOSTENDIAN;
628 629 630
	bitmap->events_cleared = le64_to_cpu(sb->events_cleared);
	err = 0;
out:
631
	kunmap_atomic(sb);
632 633 634 635 636 637
out_no_sb:
	if (bitmap->flags & BITMAP_STALE)
		bitmap->events_cleared = bitmap->mddev->events;
	bitmap->mddev->bitmap_info.chunksize = chunksize;
	bitmap->mddev->bitmap_info.daemon_sleep = daemon_sleep;
	bitmap->mddev->bitmap_info.max_write_behind = write_behind;
638 639 640 641 642 643 644 645 646
	if (err)
		bitmap_print_sb(bitmap);
	return err;
}

/*
 * general bitmap file operations
 */

647 648 649 650 651 652
/*
 * on-disk bitmap:
 *
 * Use one bit per "chunk" (block set). We do the disk I/O on the bitmap
 * file a page at a time. There's a superblock at the start of the file.
 */
653
/* calculate the index of the page that contains this bit */
654 655
static inline unsigned long file_page_index(struct bitmap_storage *store,
					    unsigned long chunk)
656
{
657
	if (store->sb_page)
658 659
		chunk += sizeof(bitmap_super_t) << 3;
	return chunk >> PAGE_BIT_SHIFT;
660 661 662
}

/* calculate the (bit) offset of this bit within a page */
663 664
static inline unsigned long file_page_offset(struct bitmap_storage *store,
					     unsigned long chunk)
665
{
666
	if (store->sb_page)
667 668
		chunk += sizeof(bitmap_super_t) << 3;
	return chunk & (PAGE_BITS - 1);
669 670 671 672 673 674 675 676 677
}

/*
 * return a pointer to the page in the filemap that contains the given bit
 *
 * this lookup is complicated by the fact that the bitmap sb might be exactly
 * 1 page (e.g., x86) or less than 1 page -- so the bitmap might start on page
 * 0 or page 1
 */
678
static inline struct page *filemap_get_page(struct bitmap_storage *store,
679
					    unsigned long chunk)
680
{
681
	if (file_page_index(store, chunk) >= store->file_pages)
682
		return NULL;
683 684
	return store->filemap[file_page_index(store, chunk)
			      - file_page_index(store, 0)];
685 686
}

687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738
static int bitmap_storage_alloc(struct bitmap_storage *store,
				unsigned long chunks, int with_super)
{
	int pnum;
	unsigned long num_pages;
	unsigned long bytes;

	bytes = DIV_ROUND_UP(chunks, 8);
	if (with_super)
		bytes += sizeof(bitmap_super_t);

	num_pages = DIV_ROUND_UP(bytes, PAGE_SIZE);

	store->filemap = kmalloc(sizeof(struct page *)
				 * num_pages, GFP_KERNEL);
	if (!store->filemap)
		return -ENOMEM;

	if (with_super && !store->sb_page) {
		store->sb_page = alloc_page(GFP_KERNEL);
		if (store->sb_page == NULL)
			return -ENOMEM;
		store->sb_page->index = 0;
	}
	pnum = 0;
	if (store->sb_page) {
		store->filemap[0] = store->sb_page;
		pnum = 1;
	}
	for ( ; pnum < num_pages; pnum++) {
		store->filemap[pnum] = alloc_page(GFP_KERNEL);
		if (!store->filemap[pnum]) {
			store->file_pages = pnum;
			return -ENOMEM;
		}
		store->filemap[pnum]->index = pnum;
	}
	store->file_pages = pnum;

	/* We need 4 bits per page, rounded up to a multiple
	 * of sizeof(unsigned long) */
	store->filemap_attr = kzalloc(
		roundup(DIV_ROUND_UP(num_pages*4, 8), sizeof(unsigned long)),
		GFP_KERNEL);
	if (!store->filemap_attr)
		return -ENOMEM;

	store->bytes = bytes;

	return 0;
}

739 740 741 742 743 744
static void bitmap_file_unmap(struct bitmap *bitmap)
{
	struct page **map, *sb_page;
	unsigned long *attr;
	int pages;
	unsigned long flags;
745
	struct bitmap_storage *store = &bitmap->storage;
746 747

	spin_lock_irqsave(&bitmap->lock, flags);
748 749 750 751 752 753 754 755
	map = store->filemap;
	store->filemap = NULL;
	attr = store->filemap_attr;
	store->filemap_attr = NULL;
	pages = store->file_pages;
	store->file_pages = 0;
	sb_page = store->sb_page;
	store->sb_page = NULL;
756 757 758
	spin_unlock_irqrestore(&bitmap->lock, flags);

	while (pages--)
759
		if (map[pages] != sb_page) /* 0 is sb_page, release it below */
760
			free_buffers(map[pages]);
761 762 763
	kfree(map);
	kfree(attr);

764 765
	if (sb_page)
		free_buffers(sb_page);
766 767 768 769 770 771 772 773
}

static void bitmap_file_put(struct bitmap *bitmap)
{
	struct file *file;
	unsigned long flags;

	spin_lock_irqsave(&bitmap->lock, flags);
774 775
	file = bitmap->storage.file;
	bitmap->storage.file = NULL;
776 777
	spin_unlock_irqrestore(&bitmap->lock, flags);

778 779 780
	if (file)
		wait_event(bitmap->write_wait,
			   atomic_read(&bitmap->pending_writes)==0);
781 782
	bitmap_file_unmap(bitmap);

783
	if (file) {
J
Josef Sipek 已提交
784
		struct inode *inode = file->f_path.dentry->d_inode;
785
		invalidate_mapping_pages(inode->i_mapping, 0, -1);
786
		fput(file);
787
	}
788 789 790 791 792 793 794 795 796 797 798
}

/*
 * bitmap_file_kick - if an error occurs while manipulating the bitmap file
 * then it is no longer reliable, so we stop using it and we mark the file
 * as failed in the superblock
 */
static void bitmap_file_kick(struct bitmap *bitmap)
{
	char *path, *ptr = NULL;

N
NeilBrown 已提交
799 800
	if (!(bitmap->flags & BITMAP_STALE)) {
		bitmap->flags |= BITMAP_STALE;
801
		bitmap_update_sb(bitmap);
802

803
		if (bitmap->storage.file) {
804 805
			path = kmalloc(PAGE_SIZE, GFP_KERNEL);
			if (path)
806 807
				ptr = d_path(&bitmap->storage.file->f_path,
					     path, PAGE_SIZE);
C
Christoph Hellwig 已提交
808

809 810
			printk(KERN_ALERT
			      "%s: kicking failed bitmap file %s from array!\n",
C
Christoph Hellwig 已提交
811
			      bmname(bitmap), IS_ERR(ptr) ? "" : ptr);
812

813 814 815 816 817
			kfree(path);
		} else
			printk(KERN_ALERT
			       "%s: disabling internal bitmap due to errors\n",
			       bmname(bitmap));
818
	}
819 820 821 822 823 824 825

	bitmap_file_put(bitmap);

	return;
}

enum bitmap_page_attr {
826
	BITMAP_PAGE_DIRTY = 0,     /* there are set bits that need to be synced */
827 828
	BITMAP_PAGE_PENDING = 1,   /* there are bits that are being cleaned.
				    * i.e. counter is 1 or 2. */
829
	BITMAP_PAGE_NEEDWRITE = 2, /* there are cleared bits that need to be synced */
830 831
};

832 833
static inline void set_page_attr(struct bitmap *bitmap, int pnum,
				 enum bitmap_page_attr attr)
834
{
835
	__set_bit((pnum<<2) + attr, bitmap->storage.filemap_attr);
836 837
}

838 839
static inline void clear_page_attr(struct bitmap *bitmap, int pnum,
				   enum bitmap_page_attr attr)
840
{
841
	__clear_bit((pnum<<2) + attr, bitmap->storage.filemap_attr);
842 843
}

844
static inline unsigned long test_page_attr(struct bitmap *bitmap, int pnum,
845
					   enum bitmap_page_attr attr)
846
{
847
	return test_bit((pnum<<2) + attr, bitmap->storage.filemap_attr);
848 849 850 851 852 853 854 855 856 857 858 859
}

/*
 * bitmap_file_set_bit -- called before performing a write to the md device
 * to set (and eventually sync) a particular bit in the bitmap file
 *
 * we set the bit immediately, then we record the page number so that
 * when an unplug occurs, we can flush the dirty pages out to disk
 */
static void bitmap_file_set_bit(struct bitmap *bitmap, sector_t block)
{
	unsigned long bit;
860
	struct page *page;
861
	void *kaddr;
862
	unsigned long chunk = block >> bitmap->chunkshift;
863

864
	page = filemap_get_page(&bitmap->storage, chunk);
865 866
	if (!page)
		return;
867
	bit = file_page_offset(&bitmap->storage, chunk);
868

869
	/* set the bit */
870
	kaddr = kmap_atomic(page);
871 872 873 874
	if (bitmap->flags & BITMAP_HOSTENDIAN)
		set_bit(bit, kaddr);
	else
		__set_bit_le(bit, kaddr);
875
	kunmap_atomic(kaddr);
876
	pr_debug("set file bit %lu page %lu\n", bit, page->index);
877
	/* record page number so it gets flushed to disk when unplug occurs */
878
	set_page_attr(bitmap, page->index, BITMAP_PAGE_DIRTY);
879 880
}

881 882 883 884 885 886 887
static void bitmap_file_clear_bit(struct bitmap *bitmap, sector_t block)
{
	unsigned long bit;
	struct page *page;
	void *paddr;
	unsigned long chunk = block >> bitmap->chunkshift;

888
	page = filemap_get_page(&bitmap->storage, chunk);
889 890
	if (!page)
		return;
891
	bit = file_page_offset(&bitmap->storage, chunk);
892 893 894 895 896 897
	paddr = kmap_atomic(page);
	if (bitmap->flags & BITMAP_HOSTENDIAN)
		clear_bit(bit, paddr);
	else
		__clear_bit_le(bit, paddr);
	kunmap_atomic(paddr);
898 899
	if (!test_page_attr(bitmap, page->index, BITMAP_PAGE_NEEDWRITE)) {
		set_page_attr(bitmap, page->index, BITMAP_PAGE_PENDING);
900 901 902 903
		bitmap->allclean = 0;
	}
}

904 905 906
/* this gets called when the md device is ready to unplug its underlying
 * (slave) device queues -- before we let any writes go down, we need to
 * sync the dirty pages of the bitmap file to disk */
907
void bitmap_unplug(struct bitmap *bitmap)
908
{
909 910
	unsigned long i, flags;
	int dirty, need_write;
911 912
	int wait = 0;

913
	if (!bitmap || !bitmap->storage.filemap)
914
		return;
915 916 917

	/* look at each page to see if there are any set bits that need to be
	 * flushed out to disk */
918
	for (i = 0; i < bitmap->storage.file_pages; i++) {
919
		spin_lock_irqsave(&bitmap->lock, flags);
920
		if (!bitmap->storage.filemap) {
921
			spin_unlock_irqrestore(&bitmap->lock, flags);
922
			return;
923
		}
924 925 926 927
		dirty = test_page_attr(bitmap, i, BITMAP_PAGE_DIRTY);
		need_write = test_page_attr(bitmap, i, BITMAP_PAGE_NEEDWRITE);
		clear_page_attr(bitmap, i, BITMAP_PAGE_DIRTY);
		clear_page_attr(bitmap, i, BITMAP_PAGE_NEEDWRITE);
928
		if (dirty || need_write)
929
			clear_page_attr(bitmap, i, BITMAP_PAGE_PENDING);
930
		if (dirty)
931 932 933
			wait = 1;
		spin_unlock_irqrestore(&bitmap->lock, flags);

934
		if (dirty || need_write)
935
			write_page(bitmap, bitmap->storage.filemap[i], 0);
936 937
	}
	if (wait) { /* if any writes were performed, we need to wait on them */
938
		if (bitmap->storage.file)
939 940
			wait_event(bitmap->write_wait,
				   atomic_read(&bitmap->pending_writes)==0);
941
		else
942
			md_super_wait(bitmap->mddev);
943
	}
944 945
	if (bitmap->flags & BITMAP_WRITE_ERROR)
		bitmap_file_kick(bitmap);
946
}
947
EXPORT_SYMBOL(bitmap_unplug);
948

949
static void bitmap_set_memory_bits(struct bitmap *bitmap, sector_t offset, int needed);
950 951 952 953 954 955 956
/* * bitmap_init_from_disk -- called at bitmap_create time to initialize
 * the in-memory bitmap from the on-disk bitmap -- also, sets up the
 * memory mapping of the bitmap file
 * Special cases:
 *   if there's no bitmap file, or if the bitmap file had been
 *   previously kicked from the array, we mark all the bits as
 *   1's in order to cause a full resync.
957 958 959
 *
 * We ignore all bits for sectors that end earlier than 'start'.
 * This is used when reading an out-of-date bitmap...
960
 */
961
static int bitmap_init_from_disk(struct bitmap *bitmap, sector_t start)
962 963
{
	unsigned long i, chunks, index, oldindex, bit;
964
	struct page *page = NULL;
965
	unsigned long bit_cnt = 0;
966
	struct file *file;
967
	unsigned long offset;
968 969
	int outofdate;
	int ret = -ENOSPC;
970
	void *paddr;
971
	struct bitmap_storage *store = &bitmap->storage;
972 973

	chunks = bitmap->chunks;
974
	file = store->file;
975

976 977
	if (!file && !bitmap->mddev->bitmap_info.offset) {
		/* No permanent bitmap - fill with '1s'. */
978 979
		store->filemap = NULL;
		store->file_pages = 0;
980 981 982 983 984 985 986 987 988 989
		for (i = 0; i < chunks ; i++) {
			/* if the disk bit is set, set the memory bit */
			int needed = ((sector_t)(i+1) << (bitmap->chunkshift)
				      >= start);
			bitmap_set_memory_bits(bitmap,
					       (sector_t)i << bitmap->chunkshift,
					       needed);
		}
		return 0;
	}
990 991 992 993 994 995

	outofdate = bitmap->flags & BITMAP_STALE;
	if (outofdate)
		printk(KERN_INFO "%s: bitmap file is out of date, doing full "
			"recovery\n", bmname(bitmap));

996
	if (file && i_size_read(file->f_mapping->host) < store->bytes) {
997
		printk(KERN_INFO "%s: bitmap file too short %lu < %lu\n",
998 999 1000
		       bmname(bitmap),
		       (unsigned long) i_size_read(file->f_mapping->host),
		       store->bytes);
1001
		goto err;
1002
	}
1003

1004
	oldindex = ~0L;
1005
	offset = 0;
1006
	if (!bitmap->mddev->bitmap_info.external)
1007
		offset = sizeof(bitmap_super_t);
1008 1009

	for (i = 0; i < chunks; i++) {
1010
		int b;
1011 1012
		index = file_page_index(&bitmap->storage, i);
		bit = file_page_offset(&bitmap->storage, i);
1013
		if (index != oldindex) { /* this is a new page, read it in */
1014
			int count;
1015
			/* unmap the old page, we're done with it */
1016 1017
			if (index == store->file_pages-1)
				count = store->bytes - index * PAGE_SIZE;
1018 1019
			else
				count = PAGE_SIZE;
1020
			page = store->filemap[index];
1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031
			if (file)
				ret = read_page(file, index, bitmap,
						count, page);
			else
				ret = read_sb_page(
					bitmap->mddev,
					bitmap->mddev->bitmap_info.offset,
					page,
					index, count);

			if (ret)
1032
				goto err;
1033

1034 1035 1036 1037 1038
			oldindex = index;

			if (outofdate) {
				/*
				 * if bitmap is out of date, dirty the
1039
				 * whole page and write it out
1040
				 */
1041
				paddr = kmap_atomic(page);
1042
				memset(paddr + offset, 0xff,
1043
				       PAGE_SIZE - offset);
1044
				kunmap_atomic(paddr);
1045 1046 1047
				write_page(bitmap, page, 1);

				ret = -EIO;
1048
				if (bitmap->flags & BITMAP_WRITE_ERROR)
1049
					goto err;
1050 1051
			}
		}
1052
		paddr = kmap_atomic(page);
1053
		if (bitmap->flags & BITMAP_HOSTENDIAN)
1054
			b = test_bit(bit, paddr);
1055
		else
A
Akinobu Mita 已提交
1056
			b = test_bit_le(bit, paddr);
1057
		kunmap_atomic(paddr);
1058
		if (b) {
1059
			/* if the disk bit is set, set the memory bit */
1060
			int needed = ((sector_t)(i+1) << bitmap->chunkshift
1061 1062
				      >= start);
			bitmap_set_memory_bits(bitmap,
1063
					       (sector_t)i << bitmap->chunkshift,
1064
					       needed);
1065 1066
			bit_cnt++;
		}
1067
		offset = 0;
1068 1069 1070
	}

	printk(KERN_INFO "%s: bitmap initialized from disk: "
1071
	       "read %lu pages, set %lu of %lu bits\n",
1072
	       bmname(bitmap), store->file_pages,
1073
	       bit_cnt, chunks);
1074 1075

	return 0;
1076

1077 1078 1079
 err:
	printk(KERN_INFO "%s: bitmap initialisation failed: %d\n",
	       bmname(bitmap), ret);
1080 1081 1082
	return ret;
}

1083 1084 1085 1086 1087
void bitmap_write_all(struct bitmap *bitmap)
{
	/* We don't actually write all bitmap blocks here,
	 * just flag them as needing to be written
	 */
1088
	int i;
1089

1090
	if (!bitmap || !bitmap->storage.filemap)
1091
		return;
1092
	if (bitmap->storage.file)
1093 1094 1095
		/* Only one copy, so nothing needed */
		return;

1096
	spin_lock_irq(&bitmap->lock);
1097
	for (i = 0; i < bitmap->storage.file_pages; i++)
1098
		set_page_attr(bitmap, i,
1099
			      BITMAP_PAGE_NEEDWRITE);
1100
	bitmap->allclean = 0;
1101
	spin_unlock_irq(&bitmap->lock);
1102 1103
}

1104 1105
static void bitmap_count_page(struct bitmap *bitmap, sector_t offset, int inc)
{
1106
	sector_t chunk = offset >> bitmap->chunkshift;
1107 1108 1109 1110
	unsigned long page = chunk >> PAGE_COUNTER_SHIFT;
	bitmap->bp[page].count += inc;
	bitmap_checkfree(bitmap, page);
}
1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121

static void bitmap_set_pending(struct bitmap *bitmap, sector_t offset)
{
	sector_t chunk = offset >> bitmap->chunkshift;
	unsigned long page = chunk >> PAGE_COUNTER_SHIFT;
	struct bitmap_page *bp = &bitmap->bp[page];

	if (!bp->pending)
		bp->pending = 1;
}

1122
static bitmap_counter_t *bitmap_get_counter(struct bitmap *bitmap,
N
NeilBrown 已提交
1123
					    sector_t offset, sector_t *blocks,
1124 1125 1126 1127 1128 1129 1130
					    int create);

/*
 * bitmap daemon -- periodically wakes up to clean bits and flush pages
 *			out to disk
 */

1131
void bitmap_daemon_work(struct mddev *mddev)
1132
{
1133
	struct bitmap *bitmap;
1134
	unsigned long j;
1135
	unsigned long nextpage;
1136
	unsigned long flags;
N
NeilBrown 已提交
1137
	sector_t blocks;
1138

1139 1140 1141
	/* Use a mutex to guard daemon_work against
	 * bitmap_destroy.
	 */
1142
	mutex_lock(&mddev->bitmap_info.mutex);
1143 1144
	bitmap = mddev->bitmap;
	if (bitmap == NULL) {
1145
		mutex_unlock(&mddev->bitmap_info.mutex);
1146
		return;
1147
	}
1148
	if (time_before(jiffies, bitmap->daemon_lastrun
N
NeilBrown 已提交
1149
			+ mddev->bitmap_info.daemon_sleep))
1150 1151
		goto done;

1152
	bitmap->daemon_lastrun = jiffies;
1153
	if (bitmap->allclean) {
N
NeilBrown 已提交
1154
		mddev->thread->timeout = MAX_SCHEDULE_TIMEOUT;
1155
		goto done;
1156 1157
	}
	bitmap->allclean = 1;
1158

1159 1160 1161 1162
	/* Any file-page which is PENDING now needs to be written.
	 * So set NEEDWRITE now, then after we make any last-minute changes
	 * we will write it.
	 */
1163
	spin_lock_irqsave(&bitmap->lock, flags);
1164
	for (j = 0; j < bitmap->storage.file_pages; j++)
1165
		if (test_page_attr(bitmap, j,
1166
				   BITMAP_PAGE_PENDING)) {
1167
			set_page_attr(bitmap, j,
1168
				      BITMAP_PAGE_NEEDWRITE);
1169
			clear_page_attr(bitmap, j,
1170 1171 1172 1173 1174 1175 1176 1177 1178
					BITMAP_PAGE_PENDING);
		}

	if (bitmap->need_sync &&
	    mddev->bitmap_info.external == 0) {
		/* Arrange for superblock update as well as
		 * other changes */
		bitmap_super_t *sb;
		bitmap->need_sync = 0;
1179 1180
		if (bitmap->storage.filemap) {
			sb = kmap_atomic(bitmap->storage.sb_page);
1181 1182 1183
			sb->events_cleared =
				cpu_to_le64(bitmap->events_cleared);
			kunmap_atomic(sb);
1184
			set_page_attr(bitmap, 0,
1185 1186
				      BITMAP_PAGE_NEEDWRITE);
		}
1187 1188 1189 1190 1191
	}
	/* Now look at the bitmap counters and if any are '2' or '1',
	 * decrement and handle accordingly.
	 */
	nextpage = 0;
1192 1193
	for (j = 0; j < bitmap->chunks; j++) {
		bitmap_counter_t *bmc;
1194
		sector_t  block = (sector_t)j << bitmap->chunkshift;
1195

1196 1197 1198 1199
		if (j == nextpage) {
			nextpage += PAGE_COUNTER_RATIO;
			if (!bitmap->bp[j >> PAGE_COUNTER_SHIFT].pending) {
				j |= PAGE_COUNTER_MASK;
1200 1201
				continue;
			}
1202
			bitmap->bp[j >> PAGE_COUNTER_SHIFT].pending = 0;
1203
		}
1204
		bmc = bitmap_get_counter(bitmap,
1205
					 block,
1206
					 &blocks, 0);
1207 1208

		if (!bmc) {
1209
			j |= PAGE_COUNTER_MASK;
1210 1211 1212 1213 1214
			continue;
		}
		if (*bmc == 1 && !bitmap->need_sync) {
			/* We can clear the bit */
			*bmc = 0;
1215 1216
			bitmap_count_page(bitmap, block, -1);
			bitmap_file_clear_bit(bitmap, block);
1217 1218
		} else if (*bmc && *bmc <= 2) {
			*bmc = 1;
1219
			bitmap_set_pending(bitmap, block);
1220
			bitmap->allclean = 0;
1221
		}
1222 1223
	}

1224 1225 1226 1227 1228 1229 1230 1231
	/* Now start writeout on any page in NEEDWRITE that isn't DIRTY.
	 * DIRTY pages need to be written by bitmap_unplug so it can wait
	 * for them.
	 * If we find any DIRTY page we stop there and let bitmap_unplug
	 * handle all the rest.  This is important in the case where
	 * the first blocking holds the superblock and it has been updated.
	 * We mustn't write any other blocks before the superblock.
	 */
1232
	for (j = 0; j < bitmap->storage.file_pages; j++) {
1233

1234
		if (test_page_attr(bitmap, j,
1235 1236 1237
				   BITMAP_PAGE_DIRTY))
			/* bitmap_unplug will handle the rest */
			break;
1238
		if (test_page_attr(bitmap, j,
1239
				   BITMAP_PAGE_NEEDWRITE)) {
1240
			clear_page_attr(bitmap, j,
1241
					BITMAP_PAGE_NEEDWRITE);
1242
			spin_unlock_irqrestore(&bitmap->lock, flags);
1243
			write_page(bitmap, bitmap->storage.filemap[j], 0);
1244
			spin_lock_irqsave(&bitmap->lock, flags);
1245
			if (!bitmap->storage.filemap)
1246
				break;
1247 1248
		}
	}
1249
	spin_unlock_irqrestore(&bitmap->lock, flags);
1250

1251
 done:
1252
	if (bitmap->allclean == 0)
N
NeilBrown 已提交
1253 1254
		mddev->thread->timeout =
			mddev->bitmap_info.daemon_sleep;
1255
	mutex_unlock(&mddev->bitmap_info.mutex);
1256 1257 1258
}

static bitmap_counter_t *bitmap_get_counter(struct bitmap *bitmap,
N
NeilBrown 已提交
1259
					    sector_t offset, sector_t *blocks,
1260
					    int create)
1261 1262
__releases(bitmap->lock)
__acquires(bitmap->lock)
1263 1264 1265 1266 1267
{
	/* If 'create', we might release the lock and reclaim it.
	 * The lock must have been taken with interrupts enabled.
	 * If !create, we don't release the lock.
	 */
1268
	sector_t chunk = offset >> bitmap->chunkshift;
1269 1270 1271
	unsigned long page = chunk >> PAGE_COUNTER_SHIFT;
	unsigned long pageoff = (chunk & PAGE_COUNTER_MASK) << COUNTER_BYTE_SHIFT;
	sector_t csize;
1272
	int err;
1273

1274 1275 1276 1277
	err = bitmap_checkpage(bitmap, page, create);

	if (bitmap->bp[page].hijacked ||
	    bitmap->bp[page].map == NULL)
1278
		csize = ((sector_t)1) << (bitmap->chunkshift +
1279 1280
					  PAGE_COUNTER_SHIFT - 1);
	else
1281
		csize = ((sector_t)1) << bitmap->chunkshift;
1282 1283 1284
	*blocks = csize - (offset & (csize - 1));

	if (err < 0)
1285
		return NULL;
1286

1287 1288 1289 1290 1291 1292 1293 1294
	/* now locked ... */

	if (bitmap->bp[page].hijacked) { /* hijacked pointer */
		/* should we use the first or second counter field
		 * of the hijacked pointer? */
		int hi = (pageoff > PAGE_COUNTER_MASK);
		return  &((bitmap_counter_t *)
			  &bitmap->bp[page].map)[hi];
1295
	} else /* page is allocated */
1296 1297 1298 1299
		return (bitmap_counter_t *)
			&(bitmap->bp[page].map[pageoff]);
}

1300
int bitmap_startwrite(struct bitmap *bitmap, sector_t offset, unsigned long sectors, int behind)
1301
{
1302 1303
	if (!bitmap)
		return 0;
1304 1305

	if (behind) {
1306
		int bw;
1307
		atomic_inc(&bitmap->behind_writes);
1308 1309 1310 1311
		bw = atomic_read(&bitmap->behind_writes);
		if (bw > bitmap->behind_writes_used)
			bitmap->behind_writes_used = bw;

1312 1313
		pr_debug("inc write-behind count %d/%lu\n",
			 bw, bitmap->mddev->bitmap_info.max_write_behind);
1314 1315
	}

1316
	while (sectors) {
N
NeilBrown 已提交
1317
		sector_t blocks;
1318 1319 1320 1321 1322 1323 1324 1325 1326
		bitmap_counter_t *bmc;

		spin_lock_irq(&bitmap->lock);
		bmc = bitmap_get_counter(bitmap, offset, &blocks, 1);
		if (!bmc) {
			spin_unlock_irq(&bitmap->lock);
			return 0;
		}

1327
		if (unlikely(COUNTER(*bmc) == COUNTER_MAX)) {
1328 1329 1330 1331 1332 1333 1334 1335
			DEFINE_WAIT(__wait);
			/* note that it is safe to do the prepare_to_wait
			 * after the test as long as we do it before dropping
			 * the spinlock.
			 */
			prepare_to_wait(&bitmap->overflow_wait, &__wait,
					TASK_UNINTERRUPTIBLE);
			spin_unlock_irq(&bitmap->lock);
J
Jens Axboe 已提交
1336
			io_schedule();
1337 1338 1339 1340
			finish_wait(&bitmap->overflow_wait, &__wait);
			continue;
		}

1341
		switch (*bmc) {
1342 1343
		case 0:
			bitmap_file_set_bit(bitmap, offset);
1344
			bitmap_count_page(bitmap, offset, 1);
1345 1346 1347 1348
			/* fall through */
		case 1:
			*bmc = 2;
		}
1349

1350 1351 1352 1353 1354 1355 1356
		(*bmc)++;

		spin_unlock_irq(&bitmap->lock);

		offset += blocks;
		if (sectors > blocks)
			sectors -= blocks;
1357 1358
		else
			sectors = 0;
1359 1360 1361
	}
	return 0;
}
1362
EXPORT_SYMBOL(bitmap_startwrite);
1363 1364

void bitmap_endwrite(struct bitmap *bitmap, sector_t offset, unsigned long sectors,
1365
		     int success, int behind)
1366
{
1367 1368
	if (!bitmap)
		return;
1369
	if (behind) {
1370 1371
		if (atomic_dec_and_test(&bitmap->behind_writes))
			wake_up(&bitmap->behind_wait);
1372 1373 1374
		pr_debug("dec write-behind count %d/%lu\n",
			 atomic_read(&bitmap->behind_writes),
			 bitmap->mddev->bitmap_info.max_write_behind);
1375 1376
	}

1377
	while (sectors) {
N
NeilBrown 已提交
1378
		sector_t blocks;
1379 1380 1381 1382 1383 1384 1385 1386 1387 1388
		unsigned long flags;
		bitmap_counter_t *bmc;

		spin_lock_irqsave(&bitmap->lock, flags);
		bmc = bitmap_get_counter(bitmap, offset, &blocks, 0);
		if (!bmc) {
			spin_unlock_irqrestore(&bitmap->lock, flags);
			return;
		}

1389
		if (success && !bitmap->mddev->degraded &&
1390 1391 1392
		    bitmap->events_cleared < bitmap->mddev->events) {
			bitmap->events_cleared = bitmap->mddev->events;
			bitmap->need_sync = 1;
1393
			sysfs_notify_dirent_safe(bitmap->sysfs_can_clear);
1394 1395
		}

1396
		if (!success && !NEEDED(*bmc))
1397 1398
			*bmc |= NEEDED_MASK;

1399
		if (COUNTER(*bmc) == COUNTER_MAX)
1400 1401
			wake_up(&bitmap->overflow_wait);

1402
		(*bmc)--;
1403
		if (*bmc <= 2) {
1404
			bitmap_set_pending(bitmap, offset);
1405 1406
			bitmap->allclean = 0;
		}
1407 1408 1409 1410
		spin_unlock_irqrestore(&bitmap->lock, flags);
		offset += blocks;
		if (sectors > blocks)
			sectors -= blocks;
1411 1412
		else
			sectors = 0;
1413 1414
	}
}
1415
EXPORT_SYMBOL(bitmap_endwrite);
1416

N
NeilBrown 已提交
1417
static int __bitmap_start_sync(struct bitmap *bitmap, sector_t offset, sector_t *blocks,
1418
			       int degraded)
1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434
{
	bitmap_counter_t *bmc;
	int rv;
	if (bitmap == NULL) {/* FIXME or bitmap set as 'failed' */
		*blocks = 1024;
		return 1; /* always resync if no bitmap */
	}
	spin_lock_irq(&bitmap->lock);
	bmc = bitmap_get_counter(bitmap, offset, blocks, 0);
	rv = 0;
	if (bmc) {
		/* locked */
		if (RESYNC(*bmc))
			rv = 1;
		else if (NEEDED(*bmc)) {
			rv = 1;
1435 1436 1437 1438
			if (!degraded) { /* don't set/clear bits if degraded */
				*bmc |= RESYNC_MASK;
				*bmc &= ~NEEDED_MASK;
			}
1439 1440 1441 1442 1443 1444
		}
	}
	spin_unlock_irq(&bitmap->lock);
	return rv;
}

N
NeilBrown 已提交
1445
int bitmap_start_sync(struct bitmap *bitmap, sector_t offset, sector_t *blocks,
1446 1447 1448 1449 1450 1451 1452 1453 1454 1455
		      int degraded)
{
	/* bitmap_start_sync must always report on multiples of whole
	 * pages, otherwise resync (which is very PAGE_SIZE based) will
	 * get confused.
	 * So call __bitmap_start_sync repeatedly (if needed) until
	 * At least PAGE_SIZE>>9 blocks are covered.
	 * Return the 'or' of the result.
	 */
	int rv = 0;
N
NeilBrown 已提交
1456
	sector_t blocks1;
1457 1458 1459 1460 1461 1462 1463 1464 1465 1466

	*blocks = 0;
	while (*blocks < (PAGE_SIZE>>9)) {
		rv |= __bitmap_start_sync(bitmap, offset,
					  &blocks1, degraded);
		offset += blocks1;
		*blocks += blocks1;
	}
	return rv;
}
1467
EXPORT_SYMBOL(bitmap_start_sync);
1468

N
NeilBrown 已提交
1469
void bitmap_end_sync(struct bitmap *bitmap, sector_t offset, sector_t *blocks, int aborted)
1470 1471 1472
{
	bitmap_counter_t *bmc;
	unsigned long flags;
1473 1474

	if (bitmap == NULL) {
1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488
		*blocks = 1024;
		return;
	}
	spin_lock_irqsave(&bitmap->lock, flags);
	bmc = bitmap_get_counter(bitmap, offset, blocks, 0);
	if (bmc == NULL)
		goto unlock;
	/* locked */
	if (RESYNC(*bmc)) {
		*bmc &= ~RESYNC_MASK;

		if (!NEEDED(*bmc) && aborted)
			*bmc |= NEEDED_MASK;
		else {
1489
			if (*bmc <= 2) {
1490
				bitmap_set_pending(bitmap, offset);
1491 1492
				bitmap->allclean = 0;
			}
1493 1494 1495 1496 1497
		}
	}
 unlock:
	spin_unlock_irqrestore(&bitmap->lock, flags);
}
1498
EXPORT_SYMBOL(bitmap_end_sync);
1499 1500 1501 1502 1503 1504 1505 1506

void bitmap_close_sync(struct bitmap *bitmap)
{
	/* Sync has finished, and any bitmap chunks that weren't synced
	 * properly have been aborted.  It remains to us to clear the
	 * RESYNC bit wherever it is still on
	 */
	sector_t sector = 0;
N
NeilBrown 已提交
1507
	sector_t blocks;
N
NeilBrown 已提交
1508 1509
	if (!bitmap)
		return;
1510 1511
	while (sector < bitmap->mddev->resync_max_sectors) {
		bitmap_end_sync(bitmap, sector, &blocks, 0);
N
NeilBrown 已提交
1512 1513 1514
		sector += blocks;
	}
}
1515
EXPORT_SYMBOL(bitmap_close_sync);
N
NeilBrown 已提交
1516 1517 1518 1519

void bitmap_cond_end_sync(struct bitmap *bitmap, sector_t sector)
{
	sector_t s = 0;
N
NeilBrown 已提交
1520
	sector_t blocks;
N
NeilBrown 已提交
1521 1522 1523 1524 1525 1526 1527 1528

	if (!bitmap)
		return;
	if (sector == 0) {
		bitmap->last_end_sync = jiffies;
		return;
	}
	if (time_before(jiffies, (bitmap->last_end_sync
1529
				  + bitmap->mddev->bitmap_info.daemon_sleep)))
N
NeilBrown 已提交
1530 1531 1532 1533
		return;
	wait_event(bitmap->mddev->recovery_wait,
		   atomic_read(&bitmap->mddev->recovery_active) == 0);

1534
	bitmap->mddev->curr_resync_completed = sector;
1535
	set_bit(MD_CHANGE_CLEAN, &bitmap->mddev->flags);
1536
	sector &= ~((1ULL << bitmap->chunkshift) - 1);
N
NeilBrown 已提交
1537 1538 1539 1540
	s = 0;
	while (s < sector && s < bitmap->mddev->resync_max_sectors) {
		bitmap_end_sync(bitmap, s, &blocks, 0);
		s += blocks;
1541
	}
N
NeilBrown 已提交
1542
	bitmap->last_end_sync = jiffies;
1543
	sysfs_notify(&bitmap->mddev->kobj, NULL, "sync_completed");
1544
}
1545
EXPORT_SYMBOL(bitmap_cond_end_sync);
1546

1547
static void bitmap_set_memory_bits(struct bitmap *bitmap, sector_t offset, int needed)
1548 1549
{
	/* For each chunk covered by any of these sectors, set the
1550
	 * counter to 2 and possibly set resync_needed.  They should all
1551 1552
	 * be 0 at this point
	 */
1553

N
NeilBrown 已提交
1554
	sector_t secs;
1555 1556 1557 1558
	bitmap_counter_t *bmc;
	spin_lock_irq(&bitmap->lock);
	bmc = bitmap_get_counter(bitmap, offset, &secs, 1);
	if (!bmc) {
1559
		spin_unlock_irq(&bitmap->lock);
1560
		return;
1561
	}
1562
	if (!*bmc) {
1563
		*bmc = 2 | (needed ? NEEDED_MASK : 0);
1564
		bitmap_count_page(bitmap, offset, 1);
1565
		bitmap_set_pending(bitmap, offset);
1566
		bitmap->allclean = 0;
1567 1568
	}
	spin_unlock_irq(&bitmap->lock);
1569 1570
}

1571 1572 1573 1574 1575 1576
/* dirty the memory and file bits for bitmap chunks "s" to "e" */
void bitmap_dirty_bits(struct bitmap *bitmap, unsigned long s, unsigned long e)
{
	unsigned long chunk;

	for (chunk = s; chunk <= e; chunk++) {
1577
		sector_t sec = (sector_t)chunk << bitmap->chunkshift;
1578
		bitmap_set_memory_bits(bitmap, sec, 1);
1579
		spin_lock_irq(&bitmap->lock);
1580
		bitmap_file_set_bit(bitmap, sec);
1581
		spin_unlock_irq(&bitmap->lock);
1582 1583 1584 1585 1586 1587
		if (sec < bitmap->mddev->recovery_cp)
			/* We are asserting that the array is dirty,
			 * so move the recovery_cp address back so
			 * that it is obvious that it is dirty
			 */
			bitmap->mddev->recovery_cp = sec;
1588 1589 1590
	}
}

1591 1592 1593
/*
 * flush out any pending updates
 */
1594
void bitmap_flush(struct mddev *mddev)
1595 1596
{
	struct bitmap *bitmap = mddev->bitmap;
1597
	long sleep;
1598 1599 1600 1601 1602 1603 1604

	if (!bitmap) /* there was no bitmap */
		return;

	/* run the daemon_work three time to ensure everything is flushed
	 * that can be
	 */
1605
	sleep = mddev->bitmap_info.daemon_sleep * 2;
1606
	bitmap->daemon_lastrun -= sleep;
1607
	bitmap_daemon_work(mddev);
1608
	bitmap->daemon_lastrun -= sleep;
1609
	bitmap_daemon_work(mddev);
1610
	bitmap->daemon_lastrun -= sleep;
1611
	bitmap_daemon_work(mddev);
1612 1613 1614
	bitmap_update_sb(bitmap);
}

1615 1616 1617
/*
 * free memory that was allocated
 */
1618
static void bitmap_free(struct bitmap *bitmap)
1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640
{
	unsigned long k, pages;
	struct bitmap_page *bp;

	if (!bitmap) /* there was no bitmap */
		return;

	/* release the bitmap file and kill the daemon */
	bitmap_file_put(bitmap);

	bp = bitmap->bp;
	pages = bitmap->pages;

	/* free all allocated memory */

	if (bp) /* deallocate the page memory */
		for (k = 0; k < pages; k++)
			if (bp[k].map && !bp[k].hijacked)
				kfree(bp[k].map);
	kfree(bp);
	kfree(bitmap);
}
1641

1642
void bitmap_destroy(struct mddev *mddev)
1643 1644 1645 1646 1647 1648
{
	struct bitmap *bitmap = mddev->bitmap;

	if (!bitmap) /* there was no bitmap */
		return;

1649
	mutex_lock(&mddev->bitmap_info.mutex);
1650
	mddev->bitmap = NULL; /* disconnect from the md device */
1651
	mutex_unlock(&mddev->bitmap_info.mutex);
1652 1653
	if (mddev->thread)
		mddev->thread->timeout = MAX_SCHEDULE_TIMEOUT;
1654

1655 1656 1657
	if (bitmap->sysfs_can_clear)
		sysfs_put(bitmap->sysfs_can_clear);

1658 1659
	bitmap_free(bitmap);
}
1660 1661 1662 1663 1664

/*
 * initialize the bitmap structure
 * if this returns an error, bitmap_destroy must be called to do clean up
 */
1665
int bitmap_create(struct mddev *mddev)
1666 1667
{
	struct bitmap *bitmap;
1668
	sector_t blocks = mddev->resync_max_sectors;
1669 1670
	unsigned long chunks;
	unsigned long pages;
1671
	struct file *file = mddev->bitmap_info.file;
1672
	int err;
1673
	struct sysfs_dirent *bm = NULL;
1674

A
Alexey Dobriyan 已提交
1675
	BUILD_BUG_ON(sizeof(bitmap_super_t) != 256);
1676

1677
	BUG_ON(file && mddev->bitmap_info.offset);
1678

1679
	bitmap = kzalloc(sizeof(*bitmap), GFP_KERNEL);
1680 1681 1682 1683
	if (!bitmap)
		return -ENOMEM;

	spin_lock_init(&bitmap->lock);
1684 1685
	atomic_set(&bitmap->pending_writes, 0);
	init_waitqueue_head(&bitmap->write_wait);
1686
	init_waitqueue_head(&bitmap->overflow_wait);
1687
	init_waitqueue_head(&bitmap->behind_wait);
1688

1689 1690
	bitmap->mddev = mddev;

1691 1692
	if (mddev->kobj.sd)
		bm = sysfs_get_dirent(mddev->kobj.sd, NULL, "bitmap");
1693
	if (bm) {
1694
		bitmap->sysfs_can_clear = sysfs_get_dirent(bm, NULL, "can_clear");
1695 1696 1697 1698
		sysfs_put(bm);
	} else
		bitmap->sysfs_can_clear = NULL;

1699
	bitmap->storage.file = file;
1700 1701
	if (file) {
		get_file(file);
1702 1703 1704 1705
		/* As future accesses to this file will use bmap,
		 * and bypass the page cache, we must sync the file
		 * first.
		 */
1706
		vfs_fsync(file, 1);
1707
	}
1708
	/* read superblock from bitmap file (this sets mddev->bitmap_info.chunksize) */
1709 1710 1711 1712 1713 1714 1715 1716 1717 1718
	if (!mddev->bitmap_info.external) {
		/*
		 * If 'MD_ARRAY_FIRST_USE' is set, then device-mapper is
		 * instructing us to create a new on-disk bitmap instance.
		 */
		if (test_and_clear_bit(MD_ARRAY_FIRST_USE, &mddev->flags))
			err = bitmap_new_disk_sb(bitmap);
		else
			err = bitmap_read_sb(bitmap);
	} else {
1719 1720 1721 1722 1723 1724 1725
		err = 0;
		if (mddev->bitmap_info.chunksize == 0 ||
		    mddev->bitmap_info.daemon_sleep == 0)
			/* chunksize and time_base need to be
			 * set first. */
			err = -EINVAL;
	}
1726
	if (err)
1727
		goto error;
1728

1729
	bitmap->daemon_lastrun = jiffies;
1730 1731
	bitmap->chunkshift = (ffz(~mddev->bitmap_info.chunksize)
			      - BITMAP_BLOCK_SHIFT);
1732

1733
	chunks = (blocks + (1 << bitmap->chunkshift) - 1) >>
1734
			bitmap->chunkshift;
1735
	pages = (chunks + PAGE_COUNTER_RATIO - 1) / PAGE_COUNTER_RATIO;
1736 1737 1738 1739 1740 1741 1742

	BUG_ON(!pages);

	bitmap->chunks = chunks;
	bitmap->pages = pages;
	bitmap->missing_pages = pages;

1743
	bitmap->bp = kzalloc(pages * sizeof(*bitmap->bp), GFP_KERNEL);
1744

1745
	err = -ENOMEM;
1746
	if (!bitmap->bp)
1747
		goto error;
1748

1749 1750 1751 1752 1753 1754
	if (file || mddev->bitmap_info.offset) {
		err = bitmap_storage_alloc(&bitmap->storage, bitmap->chunks,
					   !mddev->bitmap_info.external);
		if (err)
			goto error;
	}
1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767
	printk(KERN_INFO "created bitmap (%lu pages) for device %s\n",
		pages, bmname(bitmap));

	mddev->bitmap = bitmap;


	return (bitmap->flags & BITMAP_WRITE_ERROR) ? -EIO : 0;

 error:
	bitmap_free(bitmap);
	return err;
}

1768
int bitmap_load(struct mddev *mddev)
1769 1770
{
	int err = 0;
1771
	sector_t start = 0;
1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783
	sector_t sector = 0;
	struct bitmap *bitmap = mddev->bitmap;

	if (!bitmap)
		goto out;

	/* Clear out old bitmap info first:  Either there is none, or we
	 * are resuming after someone else has possibly changed things,
	 * so we should forget old cached info.
	 * All chunks should be clean, but some might need_sync.
	 */
	while (sector < mddev->resync_max_sectors) {
N
NeilBrown 已提交
1784
		sector_t blocks;
1785 1786 1787 1788 1789
		bitmap_start_sync(bitmap, sector, &blocks, 0);
		sector += blocks;
	}
	bitmap_close_sync(bitmap);

1790 1791 1792 1793 1794 1795
	if (mddev->degraded == 0
	    || bitmap->events_cleared == mddev->events)
		/* no need to keep dirty bits to optimise a
		 * re-add of a missing device */
		start = mddev->recovery_cp;

1796
	mutex_lock(&mddev->bitmap_info.mutex);
1797
	err = bitmap_init_from_disk(bitmap, start);
1798
	mutex_unlock(&mddev->bitmap_info.mutex);
1799

1800
	if (err)
1801
		goto out;
N
NeilBrown 已提交
1802
	bitmap->flags &= ~BITMAP_STALE;
1803 1804 1805

	/* Kick recovery in case any bits were set */
	set_bit(MD_RECOVERY_NEEDED, &bitmap->mddev->recovery);
1806

1807
	mddev->thread->timeout = mddev->bitmap_info.daemon_sleep;
1808
	md_wakeup_thread(mddev->thread);
1809

1810 1811
	bitmap_update_sb(bitmap);

1812 1813 1814
	if (bitmap->flags & BITMAP_WRITE_ERROR)
		err = -EIO;
out:
1815
	return err;
1816
}
1817
EXPORT_SYMBOL_GPL(bitmap_load);
1818

1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836
void bitmap_status(struct seq_file *seq, struct bitmap *bitmap)
{
	unsigned long chunk_kb;
	unsigned long flags;

	if (!bitmap)
		return;

	spin_lock_irqsave(&bitmap->lock, flags);
	chunk_kb = bitmap->mddev->bitmap_info.chunksize >> 10;
	seq_printf(seq, "bitmap: %lu/%lu pages [%luKB], "
		   "%lu%s chunk",
		   bitmap->pages - bitmap->missing_pages,
		   bitmap->pages,
		   (bitmap->pages - bitmap->missing_pages)
		   << (PAGE_SHIFT - 10),
		   chunk_kb ? chunk_kb : bitmap->mddev->bitmap_info.chunksize,
		   chunk_kb ? "KB" : "B");
1837
	if (bitmap->storage.file) {
1838
		seq_printf(seq, ", file: ");
1839
		seq_path(seq, &bitmap->storage.file->f_path, " \t\n");
1840 1841 1842 1843 1844 1845
	}

	seq_printf(seq, "\n");
	spin_unlock_irqrestore(&bitmap->lock, flags);
}

1846
static ssize_t
1847
location_show(struct mddev *mddev, char *page)
1848 1849
{
	ssize_t len;
1850
	if (mddev->bitmap_info.file)
1851
		len = sprintf(page, "file");
1852
	else if (mddev->bitmap_info.offset)
1853
		len = sprintf(page, "%+lld", (long long)mddev->bitmap_info.offset);
1854
	else
1855 1856 1857 1858 1859 1860
		len = sprintf(page, "none");
	len += sprintf(page+len, "\n");
	return len;
}

static ssize_t
1861
location_store(struct mddev *mddev, const char *buf, size_t len)
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
{

	if (mddev->pers) {
		if (!mddev->pers->quiesce)
			return -EBUSY;
		if (mddev->recovery || mddev->sync_thread)
			return -EBUSY;
	}

	if (mddev->bitmap || mddev->bitmap_info.file ||
	    mddev->bitmap_info.offset) {
		/* bitmap already configured.  Only option is to clear it */
		if (strncmp(buf, "none", 4) != 0)
			return -EBUSY;
		if (mddev->pers) {
			mddev->pers->quiesce(mddev, 1);
			bitmap_destroy(mddev);
			mddev->pers->quiesce(mddev, 0);
		}
		mddev->bitmap_info.offset = 0;
		if (mddev->bitmap_info.file) {
			struct file *f = mddev->bitmap_info.file;
			mddev->bitmap_info.file = NULL;
			restore_bitmap_write_access(f);
			fput(f);
		}
	} else {
		/* No bitmap, OK to set a location */
		long long offset;
		if (strncmp(buf, "none", 4) == 0)
			/* nothing to be done */;
		else if (strncmp(buf, "file:", 5) == 0) {
			/* Not supported yet */
			return -EINVAL;
		} else {
			int rv;
			if (buf[0] == '+')
				rv = strict_strtoll(buf+1, 10, &offset);
			else
				rv = strict_strtoll(buf, 10, &offset);
			if (rv)
				return rv;
			if (offset == 0)
				return -EINVAL;
1906 1907
			if (mddev->bitmap_info.external == 0 &&
			    mddev->major_version == 0 &&
1908 1909 1910 1911 1912 1913
			    offset != mddev->bitmap_info.default_offset)
				return -EINVAL;
			mddev->bitmap_info.offset = offset;
			if (mddev->pers) {
				mddev->pers->quiesce(mddev, 1);
				rv = bitmap_create(mddev);
1914 1915
				if (!rv)
					rv = bitmap_load(mddev);
1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938
				if (rv) {
					bitmap_destroy(mddev);
					mddev->bitmap_info.offset = 0;
				}
				mddev->pers->quiesce(mddev, 0);
				if (rv)
					return rv;
			}
		}
	}
	if (!mddev->external) {
		/* Ensure new bitmap info is stored in
		 * metadata promptly.
		 */
		set_bit(MD_CHANGE_DEVS, &mddev->flags);
		md_wakeup_thread(mddev->thread);
	}
	return len;
}

static struct md_sysfs_entry bitmap_location =
__ATTR(location, S_IRUGO|S_IWUSR, location_show, location_store);

1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962
/* 'bitmap/space' is the space available at 'location' for the
 * bitmap.  This allows the kernel to know when it is safe to
 * resize the bitmap to match a resized array.
 */
static ssize_t
space_show(struct mddev *mddev, char *page)
{
	return sprintf(page, "%lu\n", mddev->bitmap_info.space);
}

static ssize_t
space_store(struct mddev *mddev, const char *buf, size_t len)
{
	unsigned long sectors;
	int rv;

	rv = kstrtoul(buf, 10, &sectors);
	if (rv)
		return rv;

	if (sectors == 0)
		return -EINVAL;

	if (mddev->bitmap &&
1963
	    sectors < (mddev->bitmap->storage.bytes + 511) >> 9)
1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975
		return -EFBIG; /* Bitmap is too big for this small space */

	/* could make sure it isn't too big, but that isn't really
	 * needed - user-space should be careful.
	 */
	mddev->bitmap_info.space = sectors;
	return len;
}

static struct md_sysfs_entry bitmap_space =
__ATTR(space, S_IRUGO|S_IWUSR, space_show, space_store);

1976
static ssize_t
1977
timeout_show(struct mddev *mddev, char *page)
1978 1979 1980 1981
{
	ssize_t len;
	unsigned long secs = mddev->bitmap_info.daemon_sleep / HZ;
	unsigned long jifs = mddev->bitmap_info.daemon_sleep % HZ;
1982

1983 1984 1985 1986 1987 1988 1989 1990
	len = sprintf(page, "%lu", secs);
	if (jifs)
		len += sprintf(page+len, ".%03u", jiffies_to_msecs(jifs));
	len += sprintf(page+len, "\n");
	return len;
}

static ssize_t
1991
timeout_store(struct mddev *mddev, const char *buf, size_t len)
1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026
{
	/* timeout can be set at any time */
	unsigned long timeout;
	int rv = strict_strtoul_scaled(buf, &timeout, 4);
	if (rv)
		return rv;

	/* just to make sure we don't overflow... */
	if (timeout >= LONG_MAX / HZ)
		return -EINVAL;

	timeout = timeout * HZ / 10000;

	if (timeout >= MAX_SCHEDULE_TIMEOUT)
		timeout = MAX_SCHEDULE_TIMEOUT-1;
	if (timeout < 1)
		timeout = 1;
	mddev->bitmap_info.daemon_sleep = timeout;
	if (mddev->thread) {
		/* if thread->timeout is MAX_SCHEDULE_TIMEOUT, then
		 * the bitmap is all clean and we don't need to
		 * adjust the timeout right now
		 */
		if (mddev->thread->timeout < MAX_SCHEDULE_TIMEOUT) {
			mddev->thread->timeout = timeout;
			md_wakeup_thread(mddev->thread);
		}
	}
	return len;
}

static struct md_sysfs_entry bitmap_timeout =
__ATTR(time_base, S_IRUGO|S_IWUSR, timeout_show, timeout_store);

static ssize_t
2027
backlog_show(struct mddev *mddev, char *page)
2028 2029 2030 2031 2032
{
	return sprintf(page, "%lu\n", mddev->bitmap_info.max_write_behind);
}

static ssize_t
2033
backlog_store(struct mddev *mddev, const char *buf, size_t len)
2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048
{
	unsigned long backlog;
	int rv = strict_strtoul(buf, 10, &backlog);
	if (rv)
		return rv;
	if (backlog > COUNTER_MAX)
		return -EINVAL;
	mddev->bitmap_info.max_write_behind = backlog;
	return len;
}

static struct md_sysfs_entry bitmap_backlog =
__ATTR(backlog, S_IRUGO|S_IWUSR, backlog_show, backlog_store);

static ssize_t
2049
chunksize_show(struct mddev *mddev, char *page)
2050 2051 2052 2053 2054
{
	return sprintf(page, "%lu\n", mddev->bitmap_info.chunksize);
}

static ssize_t
2055
chunksize_store(struct mddev *mddev, const char *buf, size_t len)
2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074
{
	/* Can only be changed when no bitmap is active */
	int rv;
	unsigned long csize;
	if (mddev->bitmap)
		return -EBUSY;
	rv = strict_strtoul(buf, 10, &csize);
	if (rv)
		return rv;
	if (csize < 512 ||
	    !is_power_of_2(csize))
		return -EINVAL;
	mddev->bitmap_info.chunksize = csize;
	return len;
}

static struct md_sysfs_entry bitmap_chunksize =
__ATTR(chunksize, S_IRUGO|S_IWUSR, chunksize_show, chunksize_store);

2075
static ssize_t metadata_show(struct mddev *mddev, char *page)
2076 2077 2078 2079 2080
{
	return sprintf(page, "%s\n", (mddev->bitmap_info.external
				      ? "external" : "internal"));
}

2081
static ssize_t metadata_store(struct mddev *mddev, const char *buf, size_t len)
2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098
{
	if (mddev->bitmap ||
	    mddev->bitmap_info.file ||
	    mddev->bitmap_info.offset)
		return -EBUSY;
	if (strncmp(buf, "external", 8) == 0)
		mddev->bitmap_info.external = 1;
	else if (strncmp(buf, "internal", 8) == 0)
		mddev->bitmap_info.external = 0;
	else
		return -EINVAL;
	return len;
}

static struct md_sysfs_entry bitmap_metadata =
__ATTR(metadata, S_IRUGO|S_IWUSR, metadata_show, metadata_store);

2099
static ssize_t can_clear_show(struct mddev *mddev, char *page)
2100 2101 2102 2103 2104 2105 2106 2107 2108 2109
{
	int len;
	if (mddev->bitmap)
		len = sprintf(page, "%s\n", (mddev->bitmap->need_sync ?
					     "false" : "true"));
	else
		len = sprintf(page, "\n");
	return len;
}

2110
static ssize_t can_clear_store(struct mddev *mddev, const char *buf, size_t len)
2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127
{
	if (mddev->bitmap == NULL)
		return -ENOENT;
	if (strncmp(buf, "false", 5) == 0)
		mddev->bitmap->need_sync = 1;
	else if (strncmp(buf, "true", 4) == 0) {
		if (mddev->degraded)
			return -EBUSY;
		mddev->bitmap->need_sync = 0;
	} else
		return -EINVAL;
	return len;
}

static struct md_sysfs_entry bitmap_can_clear =
__ATTR(can_clear, S_IRUGO|S_IWUSR, can_clear_show, can_clear_store);

2128
static ssize_t
2129
behind_writes_used_show(struct mddev *mddev, char *page)
2130 2131 2132 2133 2134 2135 2136 2137
{
	if (mddev->bitmap == NULL)
		return sprintf(page, "0\n");
	return sprintf(page, "%lu\n",
		       mddev->bitmap->behind_writes_used);
}

static ssize_t
2138
behind_writes_used_reset(struct mddev *mddev, const char *buf, size_t len)
2139 2140 2141 2142 2143 2144 2145 2146 2147 2148
{
	if (mddev->bitmap)
		mddev->bitmap->behind_writes_used = 0;
	return len;
}

static struct md_sysfs_entry max_backlog_used =
__ATTR(max_backlog_used, S_IRUGO | S_IWUSR,
       behind_writes_used_show, behind_writes_used_reset);

2149 2150
static struct attribute *md_bitmap_attrs[] = {
	&bitmap_location.attr,
2151
	&bitmap_space.attr,
2152 2153 2154
	&bitmap_timeout.attr,
	&bitmap_backlog.attr,
	&bitmap_chunksize.attr,
2155 2156
	&bitmap_metadata.attr,
	&bitmap_can_clear.attr,
2157
	&max_backlog_used.attr,
2158 2159 2160 2161 2162 2163 2164
	NULL
};
struct attribute_group md_bitmap_group = {
	.name = "bitmap",
	.attrs = md_bitmap_attrs,
};