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
{
	return bitmap->mddev ? mdname(bitmap->mddev) : "mdX";
}

/*
 * just a placeholder - calls kmalloc for bitmap pages
 */
static unsigned char *bitmap_alloc_page(struct bitmap *bitmap)
{
	unsigned char *page;

45
	page = kzalloc(PAGE_SIZE, GFP_NOIO);
46 47 48
	if (!page)
		printk("%s: bitmap_alloc_page FAILED\n", bmname(bitmap));
	else
49 50
		pr_debug("%s: bitmap_alloc_page: allocated page at %p\n",
			 bmname(bitmap), page);
51 52 53 54 55 56 57 58
	return page;
}

/*
 * for now just a placeholder -- just calls kfree for bitmap pages
 */
static void bitmap_free_page(struct bitmap *bitmap, unsigned char *page)
{
59
	pr_debug("%s: bitmap_free_page: free page %p\n", bmname(bitmap), page);
60 61 62 63 64 65 66 67 68 69 70 71 72
	kfree(page);
}

/*
 * 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
 */
73 74
static int bitmap_checkpage(struct bitmap *bitmap,
			    unsigned long page, int create)
75 76
__releases(bitmap->lock)
__acquires(bitmap->lock)
77 78 79 80
{
	unsigned char *mappage;

	if (page >= bitmap->pages) {
81 82 83 84
		/* This can happen if bitmap_start_sync goes beyond
		 * End-of-device while looking for a whole page.
		 * It is harmless.
		 */
85 86 87 88 89 90 91 92 93 94 95 96 97 98
		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 */

99 100 101 102 103
	spin_unlock_irq(&bitmap->lock);
	mappage = bitmap_alloc_page(bitmap);
	spin_lock_irq(&bitmap->lock);

	if (mappage == NULL) {
104 105
		pr_debug("%s: bitmap map page allocation failed, hijacking\n",
			 bmname(bitmap));
106 107 108 109
		/* 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;
110 111
	} else if (bitmap->bp[page].map ||
		   bitmap->bp[page].hijacked) {
112 113 114
		/* somebody beat us to getting the page */
		bitmap_free_page(bitmap, mappage);
		return 0;
115
	} else {
116

117
		/* no page was in place and we have one, so install it */
118

119 120 121
		bitmap->bp[page].map = mappage;
		bitmap->missing_pages--;
	}
122 123 124 125 126 127
	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 */
128
static void bitmap_checkfree(struct bitmap *bitmap, unsigned long page)
129 130 131 132 133 134 135 136 137 138 139
{
	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;
140 141 142 143 144 145
	} else {
		/* normal case, free the page */
		ptr = bitmap->bp[page].map;
		bitmap->bp[page].map = NULL;
		bitmap->missing_pages++;
		bitmap_free_page(bitmap, ptr);
146 147 148 149 150 151 152 153 154 155 156
	}
}

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

/*
 * basic page I/O operations
 */

157
/* IO operations when bitmap is stored near all superblocks */
158
static struct page *read_sb_page(struct mddev *mddev, loff_t offset,
159 160
				 struct page *page,
				 unsigned long index, int size)
161 162 163
{
	/* choose a good rdev and read the page from there */

164
	struct md_rdev *rdev;
165
	sector_t target;
166
	int did_alloc = 0;
167

168
	if (!page) {
169
		page = alloc_page(GFP_KERNEL);
170 171 172 173
		if (!page)
			return ERR_PTR(-ENOMEM);
		did_alloc = 1;
	}
174

N
NeilBrown 已提交
175
	rdev_for_each(rdev, mddev) {
176 177
		if (! test_bit(In_sync, &rdev->flags)
		    || test_bit(Faulty, &rdev->flags))
178 179
			continue;

J
Jonathan Brassow 已提交
180
		target = offset + index * (PAGE_SIZE/512);
181

182
		if (sync_page_io(rdev, target,
183
				 roundup(size, bdev_logical_block_size(rdev->bdev)),
J
Jonathan Brassow 已提交
184
				 page, READ, true)) {
185
			page->index = index;
186 187
			attach_page_buffers(page, NULL); /* so that free_buffer will
							  * quietly no-op */
188 189 190
			return page;
		}
	}
191 192
	if (did_alloc)
		put_page(page);
193
	return ERR_PTR(-EIO);
194 195 196

}

197
static struct md_rdev *next_active_rdev(struct md_rdev *rdev, struct mddev *mddev)
198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217
{
	/* 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) {
218
		rdev = list_entry(pos, struct md_rdev, same_set);
219 220 221 222 223 224 225 226 227 228 229 230
		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;
}

231
static int write_sb_page(struct bitmap *bitmap, struct page *page, int wait)
232
{
233
	struct md_rdev *rdev = NULL;
234
	struct block_device *bdev;
235
	struct mddev *mddev = bitmap->mddev;
236

237
	while ((rdev = next_active_rdev(rdev, mddev)) != NULL) {
238 239
		int size = PAGE_SIZE;
		loff_t offset = mddev->bitmap_info.offset;
240 241 242

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

243 244
		if (page->index == bitmap->file_pages-1)
			size = roundup(bitmap->last_page_size,
245
				       bdev_logical_block_size(bdev));
246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285
		/* 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);
286
	}
287 288

	if (wait)
289
		md_super_wait(mddev);
290
	return 0;
291 292 293

 bad_alignment:
	return -EINVAL;
294 295
}

296
static void bitmap_file_kick(struct bitmap *bitmap);
297
/*
298
 * write out a page to a file
299
 */
300
static void write_page(struct bitmap *bitmap, struct page *page, int wait)
301
{
302
	struct buffer_head *bh;
303

304 305 306 307 308
	if (bitmap->file == NULL) {
		switch (write_sb_page(bitmap, page, wait)) {
		case -EINVAL:
			bitmap->flags |= BITMAP_WRITE_ERROR;
		}
309
	} else {
310

311
		bh = page_buffers(page);
312

313 314 315 316
		while (bh && bh->b_blocknr) {
			atomic_inc(&bitmap->pending_writes);
			set_buffer_locked(bh);
			set_buffer_mapped(bh);
J
Jens Axboe 已提交
317
			submit_bh(WRITE | REQ_SYNC, bh);
318 319
			bh = bh->b_this_page;
		}
320

321
		if (wait)
322 323
			wait_event(bitmap->write_wait,
				   atomic_read(&bitmap->pending_writes)==0);
324
	}
325 326
	if (bitmap->flags & BITMAP_WRITE_ERROR)
		bitmap_file_kick(bitmap);
327 328 329 330 331 332
}

static void end_bitmap_write(struct buffer_head *bh, int uptodate)
{
	struct bitmap *bitmap = bh->b_private;
	unsigned long flags;
333

334 335 336 337
	if (!uptodate) {
		spin_lock_irqsave(&bitmap->lock, flags);
		bitmap->flags |= BITMAP_WRITE_ERROR;
		spin_unlock_irqrestore(&bitmap->lock, flags);
338
	}
339 340 341
	if (atomic_dec_and_test(&bitmap->pending_writes))
		wake_up(&bitmap->write_wait);
}
342

343 344 345 346 347 348 349 350 351 352 353
/* 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)
{
	struct buffer_head *bh = page_buffers(page);
354

355 356 357 358
	while (bh) {
		struct buffer_head *next = bh->b_this_page;
		free_buffer_head(bh);
		bh = next;
359
	}
360 361
	__clear_page_buffers(page);
	put_page(page);
362 363
}

364 365 366 367 368 369 370
/* 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.
 */
371
static struct page *read_page(struct file *file, unsigned long index,
372 373
			      struct bitmap *bitmap,
			      unsigned long count)
374 375
{
	struct page *page = NULL;
J
Josef Sipek 已提交
376
	struct inode *inode = file->f_path.dentry->d_inode;
377 378
	struct buffer_head *bh;
	sector_t block;
379

380 381
	pr_debug("read bitmap file (%dB @ %llu)\n", (int)PAGE_SIZE,
		 (unsigned long long)index << PAGE_SHIFT);
382

383 384 385
	page = alloc_page(GFP_KERNEL);
	if (!page)
		page = ERR_PTR(-ENOMEM);
386 387
	if (IS_ERR(page))
		goto out;
388 389 390

	bh = alloc_page_buffers(page, 1<<inode->i_blkbits, 0);
	if (!bh) {
391
		put_page(page);
392
		page = ERR_PTR(-ENOMEM);
393 394
		goto out;
	}
395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415
	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! */
				free_buffers(page);
				page = ERR_PTR(-EINVAL);
				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;
416 417 418 419
			atomic_inc(&bitmap->pending_writes);
			set_buffer_locked(bh);
			set_buffer_mapped(bh);
			submit_bh(READ, bh);
420 421 422 423 424
		}
		block++;
		bh = bh->b_this_page;
	}
	page->index = index;
425 426 427 428 429 430 431

	wait_event(bitmap->write_wait,
		   atomic_read(&bitmap->pending_writes)==0);
	if (bitmap->flags & BITMAP_WRITE_ERROR) {
		free_buffers(page);
		page = ERR_PTR(-EIO);
	}
432 433
out:
	if (IS_ERR(page))
434
		printk(KERN_ALERT "md: bitmap read error: (%dB @ %llu): %ld\n",
435 436
			(int)PAGE_SIZE,
			(unsigned long long)index << PAGE_SHIFT,
437 438 439 440 441 442 443 444 445
			PTR_ERR(page));
	return page;
}

/*
 * bitmap file superblock operations
 */

/* update the event counter and sync the superblock to disk */
446
void bitmap_update_sb(struct bitmap *bitmap)
447 448 449 450 451
{
	bitmap_super_t *sb;
	unsigned long flags;

	if (!bitmap || !bitmap->mddev) /* no bitmap for this array */
452
		return;
453 454
	if (bitmap->mddev->bitmap_info.external)
		return;
455 456 457
	spin_lock_irqsave(&bitmap->lock, flags);
	if (!bitmap->sb_page) { /* no superblock */
		spin_unlock_irqrestore(&bitmap->lock, flags);
458
		return;
459 460
	}
	spin_unlock_irqrestore(&bitmap->lock, flags);
461
	sb = kmap_atomic(bitmap->sb_page, KM_USER0);
462
	sb->events = cpu_to_le64(bitmap->mddev->events);
463
	if (bitmap->mddev->events < bitmap->events_cleared)
464 465
		/* rocking back to read-only */
		bitmap->events_cleared = bitmap->mddev->events;
466 467
	sb->events_cleared = cpu_to_le64(bitmap->events_cleared);
	sb->state = cpu_to_le32(bitmap->flags);
468 469 470
	/* 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);
471
	kunmap_atomic(sb, KM_USER0);
472
	write_page(bitmap, bitmap->sb_page, 1);
473 474 475 476 477 478 479 480 481
}

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

	if (!bitmap || !bitmap->sb_page)
		return;
482
	sb = kmap_atomic(bitmap->sb_page, KM_USER0);
483
	printk(KERN_DEBUG "%s: bitmap file superblock:\n", bmname(bitmap));
484 485 486
	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",
487 488 489 490
					*(__u32 *)(sb->uuid+0),
					*(__u32 *)(sb->uuid+4),
					*(__u32 *)(sb->uuid+8),
					*(__u32 *)(sb->uuid+12));
491
	printk(KERN_DEBUG "        events: %llu\n",
492
			(unsigned long long) le64_to_cpu(sb->events));
493
	printk(KERN_DEBUG "events cleared: %llu\n",
494
			(unsigned long long) le64_to_cpu(sb->events_cleared));
495 496 497 498 499
	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);
500
	printk(KERN_DEBUG "max write behind: %d\n", le32_to_cpu(sb->write_behind));
501
	kunmap_atomic(sb, KM_USER0);
502 503
}

504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579
/*
 * 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;

	bitmap->sb_page = alloc_page(GFP_KERNEL);
	if (IS_ERR(bitmap->sb_page)) {
		err = PTR_ERR(bitmap->sb_page);
		bitmap->sb_page = NULL;
		return err;
	}
	bitmap->sb_page->index = 0;

	sb = kmap_atomic(bitmap->sb_page, KM_USER0);

	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)) {
		kunmap_atomic(sb, KM_USER0);
		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;
	sb->state |= cpu_to_le32(BITMAP_STALE);
	bitmap->events_cleared = bitmap->mddev->events;
	sb->events_cleared = cpu_to_le64(bitmap->mddev->events);

	bitmap->flags |= BITMAP_HOSTENDIAN;
	sb->version = cpu_to_le32(BITMAP_MAJOR_HOSTENDIAN);

	kunmap_atomic(sb, KM_USER0);

	return 0;
}

580 581 582 583 584
/* 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;
585
	unsigned long chunksize, daemon_sleep, write_behind;
586 587 588 589
	unsigned long long events;
	int err = -EINVAL;

	/* page 0 is the superblock, read it... */
590 591 592 593 594 595
	if (bitmap->file) {
		loff_t isize = i_size_read(bitmap->file->f_mapping->host);
		int bytes = isize > PAGE_SIZE ? PAGE_SIZE : isize;

		bitmap->sb_page = read_page(bitmap->file, 0, bitmap, bytes);
	} else {
596 597
		bitmap->sb_page = read_sb_page(bitmap->mddev,
					       bitmap->mddev->bitmap_info.offset,
598 599
					       NULL,
					       0, sizeof(bitmap_super_t));
600
	}
601 602 603 604 605 606
	if (IS_ERR(bitmap->sb_page)) {
		err = PTR_ERR(bitmap->sb_page);
		bitmap->sb_page = NULL;
		return err;
	}

607
	sb = kmap_atomic(bitmap->sb_page, KM_USER0);
608 609

	chunksize = le32_to_cpu(sb->chunksize);
610
	daemon_sleep = le32_to_cpu(sb->daemon_sleep) * HZ;
611
	write_behind = le32_to_cpu(sb->write_behind);
612 613 614 615

	/* verify that the bitmap-specific fields are valid */
	if (sb->magic != cpu_to_le32(BITMAP_MAGIC))
		reason = "bad magic";
616 617
	else if (le32_to_cpu(sb->version) < BITMAP_MAJOR_LO ||
		 le32_to_cpu(sb->version) > BITMAP_MAJOR_HI)
618
		reason = "unrecognized superblock version";
619
	else if (chunksize < 512)
620
		reason = "bitmap chunksize too small";
J
Jonathan Brassow 已提交
621
	else if (!is_power_of_2(chunksize))
622
		reason = "bitmap chunksize not a power of 2";
623
	else if (daemon_sleep < 1 || daemon_sleep > MAX_SCHEDULE_TIMEOUT)
624
		reason = "daemon sleep period out of range";
625 626
	else if (write_behind > COUNTER_MAX)
		reason = "write-behind limit out of range (0 - 16383)";
627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652
	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);

	if (!bitmap->mddev->persistent)
		goto success;

	/*
	 * if we have a persistent array superblock, 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);
653
		sb->state |= cpu_to_le32(BITMAP_STALE);
654 655 656
	}
success:
	/* assign fields using values from superblock */
657 658 659
	bitmap->mddev->bitmap_info.chunksize = chunksize;
	bitmap->mddev->bitmap_info.daemon_sleep = daemon_sleep;
	bitmap->mddev->bitmap_info.max_write_behind = write_behind;
660
	bitmap->flags |= le32_to_cpu(sb->state);
661 662
	if (le32_to_cpu(sb->version) == BITMAP_MAJOR_HOSTENDIAN)
		bitmap->flags |= BITMAP_HOSTENDIAN;
663
	bitmap->events_cleared = le64_to_cpu(sb->events_cleared);
664
	if (bitmap->flags & BITMAP_STALE)
665
		bitmap->events_cleared = bitmap->mddev->events;
666 667
	err = 0;
out:
668
	kunmap_atomic(sb, KM_USER0);
669 670 671 672 673 674 675 676 677 678
	if (err)
		bitmap_print_sb(bitmap);
	return err;
}

enum bitmap_mask_op {
	MASK_SET,
	MASK_UNSET
};

679 680 681
/* record the state of the bitmap in the superblock.  Return the old value */
static int bitmap_mask_state(struct bitmap *bitmap, enum bitmap_state bits,
			     enum bitmap_mask_op op)
682 683 684
{
	bitmap_super_t *sb;
	unsigned long flags;
685
	int old;
686 687

	spin_lock_irqsave(&bitmap->lock, flags);
688
	if (!bitmap->sb_page) { /* can't set the state */
689
		spin_unlock_irqrestore(&bitmap->lock, flags);
690
		return 0;
691 692
	}
	spin_unlock_irqrestore(&bitmap->lock, flags);
693
	sb = kmap_atomic(bitmap->sb_page, KM_USER0);
694
	old = le32_to_cpu(sb->state) & bits;
695
	switch (op) {
696 697
	case MASK_SET:
		sb->state |= cpu_to_le32(bits);
698
		bitmap->flags |= bits;
699 700 701
		break;
	case MASK_UNSET:
		sb->state &= cpu_to_le32(~bits);
702
		bitmap->flags &= ~bits;
703 704 705
		break;
	default:
		BUG();
706
	}
707
	kunmap_atomic(sb, KM_USER0);
708
	return old;
709 710 711 712 713 714
}

/*
 * general bitmap file operations
 */

715 716 717 718 719 720
/*
 * 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.
 */
721
/* calculate the index of the page that contains this bit */
722
static inline unsigned long file_page_index(struct bitmap *bitmap, unsigned long chunk)
723
{
724 725 726
	if (!bitmap->mddev->bitmap_info.external)
		chunk += sizeof(bitmap_super_t) << 3;
	return chunk >> PAGE_BIT_SHIFT;
727 728 729
}

/* calculate the (bit) offset of this bit within a page */
730
static inline unsigned long file_page_offset(struct bitmap *bitmap, unsigned long chunk)
731
{
732 733 734
	if (!bitmap->mddev->bitmap_info.external)
		chunk += sizeof(bitmap_super_t) << 3;
	return chunk & (PAGE_BITS - 1);
735 736 737 738 739 740 741 742 743 744
}

/*
 * 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
 */
static inline struct page *filemap_get_page(struct bitmap *bitmap,
745
					    unsigned long chunk)
746
{
747 748
	if (file_page_index(bitmap, chunk) >= bitmap->file_pages)
		return NULL;
749 750
	return bitmap->filemap[file_page_index(bitmap, chunk)
			       - file_page_index(bitmap, 0)];
751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771
}

static void bitmap_file_unmap(struct bitmap *bitmap)
{
	struct page **map, *sb_page;
	unsigned long *attr;
	int pages;
	unsigned long flags;

	spin_lock_irqsave(&bitmap->lock, flags);
	map = bitmap->filemap;
	bitmap->filemap = NULL;
	attr = bitmap->filemap_attr;
	bitmap->filemap_attr = NULL;
	pages = bitmap->file_pages;
	bitmap->file_pages = 0;
	sb_page = bitmap->sb_page;
	bitmap->sb_page = NULL;
	spin_unlock_irqrestore(&bitmap->lock, flags);

	while (pages--)
772
		if (map[pages] != sb_page) /* 0 is sb_page, release it below */
773
			free_buffers(map[pages]);
774 775 776
	kfree(map);
	kfree(attr);

777 778
	if (sb_page)
		free_buffers(sb_page);
779 780 781 782 783 784 785 786 787 788 789 790
}

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

	spin_lock_irqsave(&bitmap->lock, flags);
	file = bitmap->file;
	bitmap->file = NULL;
	spin_unlock_irqrestore(&bitmap->lock, flags);

791 792 793
	if (file)
		wait_event(bitmap->write_wait,
			   atomic_read(&bitmap->pending_writes)==0);
794 795
	bitmap_file_unmap(bitmap);

796
	if (file) {
J
Josef Sipek 已提交
797
		struct inode *inode = file->f_path.dentry->d_inode;
798
		invalidate_mapping_pages(inode->i_mapping, 0, -1);
799
		fput(file);
800
	}
801 802 803 804 805 806 807 808 809 810 811
}

/*
 * 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;

812 813
	if (bitmap_mask_state(bitmap, BITMAP_STALE, MASK_SET) == 0) {
		bitmap_update_sb(bitmap);
814

815 816 817
		if (bitmap->file) {
			path = kmalloc(PAGE_SIZE, GFP_KERNEL);
			if (path)
C
Christoph Hellwig 已提交
818 819 820
				ptr = d_path(&bitmap->file->f_path, path,
					     PAGE_SIZE);

821 822
			printk(KERN_ALERT
			      "%s: kicking failed bitmap file %s from array!\n",
C
Christoph Hellwig 已提交
823
			      bmname(bitmap), IS_ERR(ptr) ? "" : ptr);
824

825 826 827 828 829
			kfree(path);
		} else
			printk(KERN_ALERT
			       "%s: disabling internal bitmap due to errors\n",
			       bmname(bitmap));
830
	}
831 832 833 834 835 836 837

	bitmap_file_put(bitmap);

	return;
}

enum bitmap_page_attr {
838
	BITMAP_PAGE_DIRTY = 0,     /* there are set bits that need to be synced */
839 840
	BITMAP_PAGE_PENDING = 1,   /* there are bits that are being cleaned.
				    * i.e. counter is 1 or 2. */
841
	BITMAP_PAGE_NEEDWRITE = 2, /* there are cleared bits that need to be synced */
842 843 844 845 846
};

static inline void set_page_attr(struct bitmap *bitmap, struct page *page,
				enum bitmap_page_attr attr)
{
847
	__set_bit((page->index<<2) + attr, bitmap->filemap_attr);
848 849 850 851 852
}

static inline void clear_page_attr(struct bitmap *bitmap, struct page *page,
				enum bitmap_page_attr attr)
{
853
	__clear_bit((page->index<<2) + attr, bitmap->filemap_attr);
854 855
}

856 857
static inline unsigned long test_page_attr(struct bitmap *bitmap, struct page *page,
					   enum bitmap_page_attr attr)
858
{
859
	return test_bit((page->index<<2) + attr, bitmap->filemap_attr);
860 861 862 863 864 865 866 867 868 869 870 871
}

/*
 * 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;
872
	struct page *page;
873 874 875
	void *kaddr;
	unsigned long chunk = block >> CHUNK_BLOCK_SHIFT(bitmap);

876 877
	if (!bitmap->filemap)
		return;
878

879 880 881 882
	page = filemap_get_page(bitmap, chunk);
	if (!page)
		return;
	bit = file_page_offset(bitmap, chunk);
883

884 885 886 887 888 889 890
	/* set the bit */
	kaddr = kmap_atomic(page, KM_USER0);
	if (bitmap->flags & BITMAP_HOSTENDIAN)
		set_bit(bit, kaddr);
	else
		__set_bit_le(bit, kaddr);
	kunmap_atomic(kaddr, KM_USER0);
891
	pr_debug("set file bit %lu page %lu\n", bit, page->index);
892 893 894 895 896 897 898
	/* record page number so it gets flushed to disk when unplug occurs */
	set_page_attr(bitmap, page, BITMAP_PAGE_DIRTY);
}

/* 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 */
899
void bitmap_unplug(struct bitmap *bitmap)
900
{
901 902
	unsigned long i, flags;
	int dirty, need_write;
903 904 905 906
	struct page *page;
	int wait = 0;

	if (!bitmap)
907
		return;
908 909 910 911 912

	/* look at each page to see if there are any set bits that need to be
	 * flushed out to disk */
	for (i = 0; i < bitmap->file_pages; i++) {
		spin_lock_irqsave(&bitmap->lock, flags);
913
		if (!bitmap->filemap) {
914
			spin_unlock_irqrestore(&bitmap->lock, flags);
915
			return;
916 917
		}
		page = bitmap->filemap[i];
918 919
		dirty = test_page_attr(bitmap, page, BITMAP_PAGE_DIRTY);
		need_write = test_page_attr(bitmap, page, BITMAP_PAGE_NEEDWRITE);
920 921
		clear_page_attr(bitmap, page, BITMAP_PAGE_DIRTY);
		clear_page_attr(bitmap, page, BITMAP_PAGE_NEEDWRITE);
922
		if (dirty)
923 924 925
			wait = 1;
		spin_unlock_irqrestore(&bitmap->lock, flags);

926
		if (dirty || need_write)
927
			write_page(bitmap, page, 0);
928 929
	}
	if (wait) { /* if any writes were performed, we need to wait on them */
930
		if (bitmap->file)
931 932
			wait_event(bitmap->write_wait,
				   atomic_read(&bitmap->pending_writes)==0);
933
		else
934
			md_super_wait(bitmap->mddev);
935
	}
936 937
	if (bitmap->flags & BITMAP_WRITE_ERROR)
		bitmap_file_kick(bitmap);
938
}
939
EXPORT_SYMBOL(bitmap_unplug);
940

941
static void bitmap_set_memory_bits(struct bitmap *bitmap, sector_t offset, int needed);
942 943 944 945 946 947 948
/* * 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.
949 950 951
 *
 * We ignore all bits for sectors that end earlier than 'start'.
 * This is used when reading an out-of-date bitmap...
952
 */
953
static int bitmap_init_from_disk(struct bitmap *bitmap, sector_t start)
954 955 956 957 958
{
	unsigned long i, chunks, index, oldindex, bit;
	struct page *page = NULL, *oldpage = NULL;
	unsigned long num_pages, bit_cnt = 0;
	struct file *file;
959
	unsigned long bytes, offset;
960 961
	int outofdate;
	int ret = -ENOSPC;
962
	void *paddr;
963 964 965 966

	chunks = bitmap->chunks;
	file = bitmap->file;

967
	BUG_ON(!file && !bitmap->mddev->bitmap_info.offset);
968 969 970 971 972 973

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

974
	bytes = DIV_ROUND_UP(bitmap->chunks, 8);
975 976
	if (!bitmap->mddev->bitmap_info.external)
		bytes += sizeof(bitmap_super_t);
977

978
	num_pages = DIV_ROUND_UP(bytes, PAGE_SIZE);
979

980
	if (file && i_size_read(file->f_mapping->host) < bytes) {
981 982 983
		printk(KERN_INFO "%s: bitmap file too short %lu < %lu\n",
			bmname(bitmap),
			(unsigned long) i_size_read(file->f_mapping->host),
984
			bytes);
985
		goto err;
986
	}
987 988 989

	ret = -ENOMEM;

990
	bitmap->filemap = kmalloc(sizeof(struct page *) * num_pages, GFP_KERNEL);
991
	if (!bitmap->filemap)
992
		goto err;
993

994 995
	/* We need 4 bits per page, rounded up to a multiple of sizeof(unsigned long) */
	bitmap->filemap_attr = kzalloc(
996
		roundup(DIV_ROUND_UP(num_pages*4, 8), sizeof(unsigned long)),
997
		GFP_KERNEL);
998
	if (!bitmap->filemap_attr)
999
		goto err;
1000 1001 1002 1003

	oldindex = ~0L;

	for (i = 0; i < chunks; i++) {
1004
		int b;
1005 1006
		index = file_page_index(bitmap, i);
		bit = file_page_offset(bitmap, i);
1007
		if (index != oldindex) { /* this is a new page, read it in */
1008
			int count;
1009
			/* unmap the old page, we're done with it */
1010
			if (index == num_pages-1)
1011
				count = bytes - index * PAGE_SIZE;
1012 1013
			else
				count = PAGE_SIZE;
1014
			if (index == 0 && bitmap->sb_page) {
1015 1016 1017 1018 1019 1020 1021
				/*
				 * if we're here then the superblock page
				 * contains some bits (PAGE_SIZE != sizeof sb)
				 * we've already read it in, so just use it
				 */
				page = bitmap->sb_page;
				offset = sizeof(bitmap_super_t);
N
NeilBrown 已提交
1022
				if (!file)
1023 1024 1025 1026 1027
					page = read_sb_page(
						bitmap->mddev,
						bitmap->mddev->bitmap_info.offset,
						page,
						index, count);
1028
			} else if (file) {
1029
				page = read_page(file, index, bitmap, count);
1030 1031
				offset = 0;
			} else {
1032 1033
				page = read_sb_page(bitmap->mddev,
						    bitmap->mddev->bitmap_info.offset,
1034 1035
						    NULL,
						    index, count);
1036 1037
				offset = 0;
			}
1038 1039
			if (IS_ERR(page)) { /* read error */
				ret = PTR_ERR(page);
1040
				goto err;
1041 1042
			}

1043 1044 1045
			oldindex = index;
			oldpage = page;

1046 1047 1048
			bitmap->filemap[bitmap->file_pages++] = page;
			bitmap->last_page_size = count;

1049 1050 1051
			if (outofdate) {
				/*
				 * if bitmap is out of date, dirty the
1052
				 * whole page and write it out
1053
				 */
1054 1055
				paddr = kmap_atomic(page, KM_USER0);
				memset(paddr + offset, 0xff,
1056
				       PAGE_SIZE - offset);
1057
				kunmap_atomic(paddr, KM_USER0);
1058 1059 1060
				write_page(bitmap, page, 1);

				ret = -EIO;
1061
				if (bitmap->flags & BITMAP_WRITE_ERROR)
1062
					goto err;
1063 1064
			}
		}
1065
		paddr = kmap_atomic(page, KM_USER0);
1066
		if (bitmap->flags & BITMAP_HOSTENDIAN)
1067
			b = test_bit(bit, paddr);
1068
		else
A
Akinobu Mita 已提交
1069
			b = test_bit_le(bit, paddr);
1070
		kunmap_atomic(paddr, KM_USER0);
1071
		if (b) {
1072
			/* if the disk bit is set, set the memory bit */
1073 1074 1075 1076 1077
			int needed = ((sector_t)(i+1) << (CHUNK_BLOCK_SHIFT(bitmap))
				      >= start);
			bitmap_set_memory_bits(bitmap,
					       (sector_t)i << CHUNK_BLOCK_SHIFT(bitmap),
					       needed);
1078 1079 1080 1081
			bit_cnt++;
		}
	}

1082
	/* everything went OK */
1083 1084 1085 1086 1087 1088 1089 1090 1091
	ret = 0;
	bitmap_mask_state(bitmap, BITMAP_STALE, MASK_UNSET);

	if (bit_cnt) { /* Kick recovery if any bits were set */
		set_bit(MD_RECOVERY_NEEDED, &bitmap->mddev->recovery);
		md_wakeup_thread(bitmap->mddev->thread);
	}

	printk(KERN_INFO "%s: bitmap initialized from disk: "
1092 1093
	       "read %lu/%lu pages, set %lu of %lu bits\n",
	       bmname(bitmap), bitmap->file_pages, num_pages, bit_cnt, chunks);
1094 1095

	return 0;
1096

1097 1098 1099
 err:
	printk(KERN_INFO "%s: bitmap initialisation failed: %d\n",
	       bmname(bitmap), ret);
1100 1101 1102
	return ret;
}

1103 1104 1105 1106 1107
void bitmap_write_all(struct bitmap *bitmap)
{
	/* We don't actually write all bitmap blocks here,
	 * just flag them as needing to be written
	 */
1108
	int i;
1109

1110
	spin_lock_irq(&bitmap->lock);
1111
	for (i = 0; i < bitmap->file_pages; i++)
1112 1113
		set_page_attr(bitmap, bitmap->filemap[i],
			      BITMAP_PAGE_NEEDWRITE);
1114
	bitmap->allclean = 0;
1115
	spin_unlock_irq(&bitmap->lock);
1116 1117
}

1118 1119 1120 1121 1122 1123 1124 1125
static void bitmap_count_page(struct bitmap *bitmap, sector_t offset, int inc)
{
	sector_t chunk = offset >> CHUNK_BLOCK_SHIFT(bitmap);
	unsigned long page = chunk >> PAGE_COUNTER_SHIFT;
	bitmap->bp[page].count += inc;
	bitmap_checkfree(bitmap, page);
}
static bitmap_counter_t *bitmap_get_counter(struct bitmap *bitmap,
N
NeilBrown 已提交
1126
					    sector_t offset, sector_t *blocks,
1127 1128 1129 1130 1131 1132 1133
					    int create);

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

1134
void bitmap_daemon_work(struct mddev *mddev)
1135
{
1136
	struct bitmap *bitmap;
1137
	unsigned long j;
1138 1139
	unsigned long flags;
	struct page *page = NULL, *lastpage = NULL;
N
NeilBrown 已提交
1140
	sector_t blocks;
1141
	void *paddr;
1142

1143 1144 1145
	/* Use a mutex to guard daemon_work against
	 * bitmap_destroy.
	 */
1146
	mutex_lock(&mddev->bitmap_info.mutex);
1147 1148
	bitmap = mddev->bitmap;
	if (bitmap == NULL) {
1149
		mutex_unlock(&mddev->bitmap_info.mutex);
1150
		return;
1151
	}
1152
	if (time_before(jiffies, bitmap->daemon_lastrun
N
NeilBrown 已提交
1153
			+ mddev->bitmap_info.daemon_sleep))
1154 1155
		goto done;

1156
	bitmap->daemon_lastrun = jiffies;
1157
	if (bitmap->allclean) {
N
NeilBrown 已提交
1158
		mddev->thread->timeout = MAX_SCHEDULE_TIMEOUT;
1159
		goto done;
1160 1161
	}
	bitmap->allclean = 1;
1162

1163
	spin_lock_irqsave(&bitmap->lock, flags);
1164 1165
	for (j = 0; j < bitmap->chunks; j++) {
		bitmap_counter_t *bmc;
1166 1167 1168 1169 1170
		if (!bitmap->filemap)
			/* error or shutdown */
			break;

		page = filemap_get_page(bitmap, j);
1171 1172

		if (page != lastpage) {
1173
			/* skip this page unless it's marked as needing cleaning */
1174
			if (!test_page_attr(bitmap, page, BITMAP_PAGE_PENDING)) {
1175 1176
				int need_write = test_page_attr(bitmap, page,
								BITMAP_PAGE_NEEDWRITE);
1177
				if (need_write)
1178
					clear_page_attr(bitmap, page, BITMAP_PAGE_NEEDWRITE);
1179

1180
				spin_unlock_irqrestore(&bitmap->lock, flags);
1181
				if (need_write)
1182
					write_page(bitmap, page, 0);
1183 1184
				spin_lock_irqsave(&bitmap->lock, flags);
				j |= (PAGE_BITS - 1);
1185 1186 1187
				continue;
			}

1188 1189
			/* grab the new page, sync and release the old */
			if (lastpage != NULL) {
1190 1191 1192 1193
				if (test_page_attr(bitmap, lastpage,
						   BITMAP_PAGE_NEEDWRITE)) {
					clear_page_attr(bitmap, lastpage,
							BITMAP_PAGE_NEEDWRITE);
1194
					spin_unlock_irqrestore(&bitmap->lock, flags);
1195
					write_page(bitmap, lastpage, 0);
1196
				} else {
1197 1198 1199
					set_page_attr(bitmap, lastpage,
						      BITMAP_PAGE_NEEDWRITE);
					bitmap->allclean = 0;
1200 1201 1202 1203 1204
					spin_unlock_irqrestore(&bitmap->lock, flags);
				}
			} else
				spin_unlock_irqrestore(&bitmap->lock, flags);
			lastpage = page;
1205 1206 1207 1208

			/* We are possibly going to clear some bits, so make
			 * sure that events_cleared is up-to-date.
			 */
1209
			if (bitmap->need_sync &&
N
NeilBrown 已提交
1210
			    mddev->bitmap_info.external == 0) {
1211 1212 1213 1214 1215 1216 1217 1218
				bitmap_super_t *sb;
				bitmap->need_sync = 0;
				sb = kmap_atomic(bitmap->sb_page, KM_USER0);
				sb->events_cleared =
					cpu_to_le64(bitmap->events_cleared);
				kunmap_atomic(sb, KM_USER0);
				write_page(bitmap, bitmap->sb_page, 1);
			}
1219
			spin_lock_irqsave(&bitmap->lock, flags);
1220
			if (!bitmap->need_sync)
1221
				clear_page_attr(bitmap, page, BITMAP_PAGE_PENDING);
1222 1223
			else
				bitmap->allclean = 0;
1224
		}
1225 1226 1227
		bmc = bitmap_get_counter(bitmap,
					 (sector_t)j << CHUNK_BLOCK_SHIFT(bitmap),
					 &blocks, 0);
1228 1229 1230 1231
		if (!bmc)
			j |= PAGE_COUNTER_MASK;
		else if (*bmc) {
			if (*bmc == 1 && !bitmap->need_sync) {
1232 1233
				/* we can clear the bit */
				*bmc = 0;
1234 1235
				bitmap_count_page(bitmap,
						  (sector_t)j << CHUNK_BLOCK_SHIFT(bitmap),
1236 1237 1238
						  -1);

				/* clear the bit */
1239 1240 1241 1242 1243 1244
				paddr = kmap_atomic(page, KM_USER0);
				if (bitmap->flags & BITMAP_HOSTENDIAN)
					clear_bit(file_page_offset(bitmap, j),
						  paddr);
				else
					__clear_bit_le(
1245 1246 1247
						file_page_offset(bitmap,
								 j),
						paddr);
1248
				kunmap_atomic(paddr, KM_USER0);
1249 1250 1251
			} else if (*bmc <= 2) {
				*bmc = 1; /* maybe clear the bit next time */
				set_page_attr(bitmap, page, BITMAP_PAGE_PENDING);
1252
				bitmap->allclean = 0;
1253
			}
1254
		}
1255
	}
1256
	spin_unlock_irqrestore(&bitmap->lock, flags);
1257 1258

	/* now sync the final page */
1259
	if (lastpage != NULL) {
1260
		spin_lock_irqsave(&bitmap->lock, flags);
1261
		if (test_page_attr(bitmap, lastpage, BITMAP_PAGE_NEEDWRITE)) {
1262 1263
			clear_page_attr(bitmap, lastpage, BITMAP_PAGE_NEEDWRITE);
			spin_unlock_irqrestore(&bitmap->lock, flags);
1264
			write_page(bitmap, lastpage, 0);
1265 1266
		} else {
			set_page_attr(bitmap, lastpage, BITMAP_PAGE_NEEDWRITE);
1267
			bitmap->allclean = 0;
1268 1269 1270 1271
			spin_unlock_irqrestore(&bitmap->lock, flags);
		}
	}

1272
 done:
1273
	if (bitmap->allclean == 0)
N
NeilBrown 已提交
1274 1275
		mddev->thread->timeout =
			mddev->bitmap_info.daemon_sleep;
1276
	mutex_unlock(&mddev->bitmap_info.mutex);
1277 1278 1279
}

static bitmap_counter_t *bitmap_get_counter(struct bitmap *bitmap,
N
NeilBrown 已提交
1280
					    sector_t offset, sector_t *blocks,
1281
					    int create)
1282 1283
__releases(bitmap->lock)
__acquires(bitmap->lock)
1284 1285 1286 1287 1288 1289 1290 1291 1292
{
	/* 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.
	 */
	sector_t chunk = offset >> CHUNK_BLOCK_SHIFT(bitmap);
	unsigned long page = chunk >> PAGE_COUNTER_SHIFT;
	unsigned long pageoff = (chunk & PAGE_COUNTER_MASK) << COUNTER_BYTE_SHIFT;
	sector_t csize;
1293
	int err;
1294

1295 1296 1297 1298 1299 1300 1301
	err = bitmap_checkpage(bitmap, page, create);

	if (bitmap->bp[page].hijacked ||
	    bitmap->bp[page].map == NULL)
		csize = ((sector_t)1) << (CHUNK_BLOCK_SHIFT(bitmap) +
					  PAGE_COUNTER_SHIFT - 1);
	else
1302
		csize = ((sector_t)1) << (CHUNK_BLOCK_SHIFT(bitmap));
1303 1304 1305
	*blocks = csize - (offset & (csize - 1));

	if (err < 0)
1306
		return NULL;
1307

1308 1309 1310 1311 1312 1313 1314 1315
	/* 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];
1316
	} else /* page is allocated */
1317 1318 1319 1320
		return (bitmap_counter_t *)
			&(bitmap->bp[page].map[pageoff]);
}

1321
int bitmap_startwrite(struct bitmap *bitmap, sector_t offset, unsigned long sectors, int behind)
1322
{
1323 1324
	if (!bitmap)
		return 0;
1325 1326

	if (behind) {
1327
		int bw;
1328
		atomic_inc(&bitmap->behind_writes);
1329 1330 1331 1332
		bw = atomic_read(&bitmap->behind_writes);
		if (bw > bitmap->behind_writes_used)
			bitmap->behind_writes_used = bw;

1333 1334
		pr_debug("inc write-behind count %d/%lu\n",
			 bw, bitmap->mddev->bitmap_info.max_write_behind);
1335 1336
	}

1337
	while (sectors) {
N
NeilBrown 已提交
1338
		sector_t blocks;
1339 1340 1341 1342 1343 1344 1345 1346 1347
		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;
		}

1348
		if (unlikely(COUNTER(*bmc) == COUNTER_MAX)) {
1349 1350 1351 1352 1353 1354 1355 1356
			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
1357
			io_schedule();
1358 1359 1360 1361
			finish_wait(&bitmap->overflow_wait, &__wait);
			continue;
		}

1362
		switch (*bmc) {
1363 1364
		case 0:
			bitmap_file_set_bit(bitmap, offset);
1365
			bitmap_count_page(bitmap, offset, 1);
1366 1367 1368 1369
			/* fall through */
		case 1:
			*bmc = 2;
		}
1370

1371 1372 1373 1374 1375 1376 1377
		(*bmc)++;

		spin_unlock_irq(&bitmap->lock);

		offset += blocks;
		if (sectors > blocks)
			sectors -= blocks;
1378 1379
		else
			sectors = 0;
1380 1381 1382
	}
	return 0;
}
1383
EXPORT_SYMBOL(bitmap_startwrite);
1384 1385

void bitmap_endwrite(struct bitmap *bitmap, sector_t offset, unsigned long sectors,
1386
		     int success, int behind)
1387
{
1388 1389
	if (!bitmap)
		return;
1390
	if (behind) {
1391 1392
		if (atomic_dec_and_test(&bitmap->behind_writes))
			wake_up(&bitmap->behind_wait);
1393 1394 1395
		pr_debug("dec write-behind count %d/%lu\n",
			 atomic_read(&bitmap->behind_writes),
			 bitmap->mddev->bitmap_info.max_write_behind);
1396 1397
	}

1398
	while (sectors) {
N
NeilBrown 已提交
1399
		sector_t blocks;
1400 1401 1402 1403 1404 1405 1406 1407 1408 1409
		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;
		}

1410
		if (success && !bitmap->mddev->degraded &&
1411 1412 1413
		    bitmap->events_cleared < bitmap->mddev->events) {
			bitmap->events_cleared = bitmap->mddev->events;
			bitmap->need_sync = 1;
1414
			sysfs_notify_dirent_safe(bitmap->sysfs_can_clear);
1415 1416
		}

1417
		if (!success && !NEEDED(*bmc))
1418 1419
			*bmc |= NEEDED_MASK;

1420
		if (COUNTER(*bmc) == COUNTER_MAX)
1421 1422
			wake_up(&bitmap->overflow_wait);

1423
		(*bmc)--;
1424
		if (*bmc <= 2) {
1425
			set_page_attr(bitmap,
1426 1427 1428
				      filemap_get_page(
					      bitmap,
					      offset >> CHUNK_BLOCK_SHIFT(bitmap)),
1429
				      BITMAP_PAGE_PENDING);
1430 1431
			bitmap->allclean = 0;
		}
1432 1433 1434 1435
		spin_unlock_irqrestore(&bitmap->lock, flags);
		offset += blocks;
		if (sectors > blocks)
			sectors -= blocks;
1436 1437
		else
			sectors = 0;
1438 1439
	}
}
1440
EXPORT_SYMBOL(bitmap_endwrite);
1441

N
NeilBrown 已提交
1442
static int __bitmap_start_sync(struct bitmap *bitmap, sector_t offset, sector_t *blocks,
1443
			       int degraded)
1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459
{
	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;
1460 1461 1462 1463
			if (!degraded) { /* don't set/clear bits if degraded */
				*bmc |= RESYNC_MASK;
				*bmc &= ~NEEDED_MASK;
			}
1464 1465 1466 1467 1468 1469
		}
	}
	spin_unlock_irq(&bitmap->lock);
	return rv;
}

N
NeilBrown 已提交
1470
int bitmap_start_sync(struct bitmap *bitmap, sector_t offset, sector_t *blocks,
1471 1472 1473 1474 1475 1476 1477 1478 1479 1480
		      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
1481
	sector_t blocks1;
1482 1483 1484 1485 1486 1487 1488 1489 1490 1491

	*blocks = 0;
	while (*blocks < (PAGE_SIZE>>9)) {
		rv |= __bitmap_start_sync(bitmap, offset,
					  &blocks1, degraded);
		offset += blocks1;
		*blocks += blocks1;
	}
	return rv;
}
1492
EXPORT_SYMBOL(bitmap_start_sync);
1493

N
NeilBrown 已提交
1494
void bitmap_end_sync(struct bitmap *bitmap, sector_t offset, sector_t *blocks, int aborted)
1495 1496 1497
{
	bitmap_counter_t *bmc;
	unsigned long flags;
1498 1499

	if (bitmap == NULL) {
1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513
		*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 {
1514
			if (*bmc <= 2) {
1515 1516
				set_page_attr(bitmap,
					      filemap_get_page(bitmap, offset >> CHUNK_BLOCK_SHIFT(bitmap)),
1517
					      BITMAP_PAGE_PENDING);
1518 1519
				bitmap->allclean = 0;
			}
1520 1521 1522 1523 1524
		}
	}
 unlock:
	spin_unlock_irqrestore(&bitmap->lock, flags);
}
1525
EXPORT_SYMBOL(bitmap_end_sync);
1526 1527 1528 1529 1530 1531 1532 1533

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
1534
	sector_t blocks;
N
NeilBrown 已提交
1535 1536
	if (!bitmap)
		return;
1537 1538
	while (sector < bitmap->mddev->resync_max_sectors) {
		bitmap_end_sync(bitmap, sector, &blocks, 0);
N
NeilBrown 已提交
1539 1540 1541
		sector += blocks;
	}
}
1542
EXPORT_SYMBOL(bitmap_close_sync);
N
NeilBrown 已提交
1543 1544 1545 1546

void bitmap_cond_end_sync(struct bitmap *bitmap, sector_t sector)
{
	sector_t s = 0;
N
NeilBrown 已提交
1547
	sector_t blocks;
N
NeilBrown 已提交
1548 1549 1550 1551 1552 1553 1554 1555

	if (!bitmap)
		return;
	if (sector == 0) {
		bitmap->last_end_sync = jiffies;
		return;
	}
	if (time_before(jiffies, (bitmap->last_end_sync
1556
				  + bitmap->mddev->bitmap_info.daemon_sleep)))
N
NeilBrown 已提交
1557 1558 1559 1560
		return;
	wait_event(bitmap->mddev->recovery_wait,
		   atomic_read(&bitmap->mddev->recovery_active) == 0);

1561
	bitmap->mddev->curr_resync_completed = sector;
1562
	set_bit(MD_CHANGE_CLEAN, &bitmap->mddev->flags);
N
NeilBrown 已提交
1563 1564 1565 1566 1567
	sector &= ~((1ULL << CHUNK_BLOCK_SHIFT(bitmap)) - 1);
	s = 0;
	while (s < sector && s < bitmap->mddev->resync_max_sectors) {
		bitmap_end_sync(bitmap, s, &blocks, 0);
		s += blocks;
1568
	}
N
NeilBrown 已提交
1569
	bitmap->last_end_sync = jiffies;
1570
	sysfs_notify(&bitmap->mddev->kobj, NULL, "sync_completed");
1571
}
1572
EXPORT_SYMBOL(bitmap_cond_end_sync);
1573

1574
static void bitmap_set_memory_bits(struct bitmap *bitmap, sector_t offset, int needed)
1575 1576
{
	/* For each chunk covered by any of these sectors, set the
1577
	 * counter to 1 and set resync_needed.  They should all
1578 1579
	 * be 0 at this point
	 */
1580

N
NeilBrown 已提交
1581
	sector_t secs;
1582 1583 1584 1585
	bitmap_counter_t *bmc;
	spin_lock_irq(&bitmap->lock);
	bmc = bitmap_get_counter(bitmap, offset, &secs, 1);
	if (!bmc) {
1586
		spin_unlock_irq(&bitmap->lock);
1587
		return;
1588
	}
1589
	if (!*bmc) {
1590
		struct page *page;
1591
		*bmc = 2 | (needed ? NEEDED_MASK : 0);
1592 1593
		bitmap_count_page(bitmap, offset, 1);
		page = filemap_get_page(bitmap, offset >> CHUNK_BLOCK_SHIFT(bitmap));
1594
		set_page_attr(bitmap, page, BITMAP_PAGE_PENDING);
1595
		bitmap->allclean = 0;
1596 1597
	}
	spin_unlock_irq(&bitmap->lock);
1598 1599
}

1600 1601 1602 1603 1604 1605
/* 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++) {
1606
		sector_t sec = (sector_t)chunk << CHUNK_BLOCK_SHIFT(bitmap);
1607
		bitmap_set_memory_bits(bitmap, sec, 1);
1608
		spin_lock_irq(&bitmap->lock);
1609
		bitmap_file_set_bit(bitmap, sec);
1610
		spin_unlock_irq(&bitmap->lock);
1611 1612 1613 1614 1615 1616
		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;
1617 1618 1619
	}
}

1620 1621 1622
/*
 * flush out any pending updates
 */
1623
void bitmap_flush(struct mddev *mddev)
1624 1625
{
	struct bitmap *bitmap = mddev->bitmap;
1626
	long sleep;
1627 1628 1629 1630 1631 1632 1633

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

	/* run the daemon_work three time to ensure everything is flushed
	 * that can be
	 */
1634
	sleep = mddev->bitmap_info.daemon_sleep * 2;
1635
	bitmap->daemon_lastrun -= sleep;
1636
	bitmap_daemon_work(mddev);
1637
	bitmap->daemon_lastrun -= sleep;
1638
	bitmap_daemon_work(mddev);
1639
	bitmap->daemon_lastrun -= sleep;
1640
	bitmap_daemon_work(mddev);
1641 1642 1643
	bitmap_update_sb(bitmap);
}

1644 1645 1646
/*
 * free memory that was allocated
 */
1647
static void bitmap_free(struct bitmap *bitmap)
1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669
{
	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);
}
1670

1671
void bitmap_destroy(struct mddev *mddev)
1672 1673 1674 1675 1676 1677
{
	struct bitmap *bitmap = mddev->bitmap;

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

1678
	mutex_lock(&mddev->bitmap_info.mutex);
1679
	mddev->bitmap = NULL; /* disconnect from the md device */
1680
	mutex_unlock(&mddev->bitmap_info.mutex);
1681 1682
	if (mddev->thread)
		mddev->thread->timeout = MAX_SCHEDULE_TIMEOUT;
1683

1684 1685 1686
	if (bitmap->sysfs_can_clear)
		sysfs_put(bitmap->sysfs_can_clear);

1687 1688
	bitmap_free(bitmap);
}
1689 1690 1691 1692 1693

/*
 * initialize the bitmap structure
 * if this returns an error, bitmap_destroy must be called to do clean up
 */
1694
int bitmap_create(struct mddev *mddev)
1695 1696
{
	struct bitmap *bitmap;
1697
	sector_t blocks = mddev->resync_max_sectors;
1698 1699
	unsigned long chunks;
	unsigned long pages;
1700
	struct file *file = mddev->bitmap_info.file;
1701
	int err;
1702
	struct sysfs_dirent *bm = NULL;
1703

A
Alexey Dobriyan 已提交
1704
	BUILD_BUG_ON(sizeof(bitmap_super_t) != 256);
1705

1706
	if (!file
1707
	    && !mddev->bitmap_info.offset) /* bitmap disabled, nothing to do */
1708 1709
		return 0;

1710
	BUG_ON(file && mddev->bitmap_info.offset);
1711

1712
	bitmap = kzalloc(sizeof(*bitmap), GFP_KERNEL);
1713 1714 1715 1716
	if (!bitmap)
		return -ENOMEM;

	spin_lock_init(&bitmap->lock);
1717 1718
	atomic_set(&bitmap->pending_writes, 0);
	init_waitqueue_head(&bitmap->write_wait);
1719
	init_waitqueue_head(&bitmap->overflow_wait);
1720
	init_waitqueue_head(&bitmap->behind_wait);
1721

1722 1723
	bitmap->mddev = mddev;

1724 1725
	if (mddev->kobj.sd)
		bm = sysfs_get_dirent(mddev->kobj.sd, NULL, "bitmap");
1726
	if (bm) {
1727
		bitmap->sysfs_can_clear = sysfs_get_dirent(bm, NULL, "can_clear");
1728 1729 1730 1731
		sysfs_put(bm);
	} else
		bitmap->sysfs_can_clear = NULL;

1732
	bitmap->file = file;
1733 1734
	if (file) {
		get_file(file);
1735 1736 1737 1738
		/* As future accesses to this file will use bmap,
		 * and bypass the page cache, we must sync the file
		 * first.
		 */
1739
		vfs_fsync(file, 1);
1740
	}
1741
	/* read superblock from bitmap file (this sets mddev->bitmap_info.chunksize) */
1742 1743 1744 1745 1746 1747 1748 1749 1750 1751
	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 {
1752 1753 1754 1755 1756 1757 1758
		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;
	}
1759
	if (err)
1760
		goto error;
1761

1762
	bitmap->daemon_lastrun = jiffies;
1763
	bitmap->chunkshift = ffz(~mddev->bitmap_info.chunksize);
1764 1765

	/* now that chunksize and chunkshift are set, we can use these macros */
1766
	chunks = (blocks + CHUNK_BLOCK_RATIO(bitmap) - 1) >>
1767
			CHUNK_BLOCK_SHIFT(bitmap);
1768
	pages = (chunks + PAGE_COUNTER_RATIO - 1) / PAGE_COUNTER_RATIO;
1769 1770 1771 1772 1773 1774 1775

	BUG_ON(!pages);

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

1776
	bitmap->bp = kzalloc(pages * sizeof(*bitmap->bp), GFP_KERNEL);
1777

1778
	err = -ENOMEM;
1779
	if (!bitmap->bp)
1780
		goto error;
1781

1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794
	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;
}

1795
int bitmap_load(struct mddev *mddev)
1796 1797
{
	int err = 0;
1798
	sector_t start = 0;
1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810
	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
1811
		sector_t blocks;
1812 1813 1814 1815 1816
		bitmap_start_sync(bitmap, sector, &blocks, 0);
		sector += blocks;
	}
	bitmap_close_sync(bitmap);

1817 1818 1819 1820 1821 1822 1823 1824
	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;

	err = bitmap_init_from_disk(bitmap, start);

1825
	if (err)
1826
		goto out;
1827

1828
	mddev->thread->timeout = mddev->bitmap_info.daemon_sleep;
1829
	md_wakeup_thread(mddev->thread);
1830

1831 1832
	bitmap_update_sb(bitmap);

1833 1834 1835
	if (bitmap->flags & BITMAP_WRITE_ERROR)
		err = -EIO;
out:
1836
	return err;
1837
}
1838
EXPORT_SYMBOL_GPL(bitmap_load);
1839

1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866
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");
	if (bitmap->file) {
		seq_printf(seq, ", file: ");
		seq_path(seq, &bitmap->file->f_path, " \t\n");
	}

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

1867
static ssize_t
1868
location_show(struct mddev *mddev, char *page)
1869 1870
{
	ssize_t len;
1871
	if (mddev->bitmap_info.file)
1872
		len = sprintf(page, "file");
1873
	else if (mddev->bitmap_info.offset)
1874
		len = sprintf(page, "%+lld", (long long)mddev->bitmap_info.offset);
1875
	else
1876 1877 1878 1879 1880 1881
		len = sprintf(page, "none");
	len += sprintf(page+len, "\n");
	return len;
}

static ssize_t
1882
location_store(struct mddev *mddev, const char *buf, size_t len)
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
{

	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;
1927 1928
			if (mddev->bitmap_info.external == 0 &&
			    mddev->major_version == 0 &&
1929 1930 1931 1932 1933 1934
			    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);
1935 1936
				if (!rv)
					rv = bitmap_load(mddev);
1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960
				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);

static ssize_t
1961
timeout_show(struct mddev *mddev, char *page)
1962 1963 1964 1965
{
	ssize_t len;
	unsigned long secs = mddev->bitmap_info.daemon_sleep / HZ;
	unsigned long jifs = mddev->bitmap_info.daemon_sleep % HZ;
1966

1967 1968 1969 1970 1971 1972 1973 1974
	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
1975
timeout_store(struct mddev *mddev, const char *buf, size_t len)
1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
{
	/* 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
2011
backlog_show(struct mddev *mddev, char *page)
2012 2013 2014 2015 2016
{
	return sprintf(page, "%lu\n", mddev->bitmap_info.max_write_behind);
}

static ssize_t
2017
backlog_store(struct mddev *mddev, const char *buf, size_t len)
2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032
{
	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
2033
chunksize_show(struct mddev *mddev, char *page)
2034 2035 2036 2037 2038
{
	return sprintf(page, "%lu\n", mddev->bitmap_info.chunksize);
}

static ssize_t
2039
chunksize_store(struct mddev *mddev, const char *buf, size_t len)
2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058
{
	/* 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);

2059
static ssize_t metadata_show(struct mddev *mddev, char *page)
2060 2061 2062 2063 2064
{
	return sprintf(page, "%s\n", (mddev->bitmap_info.external
				      ? "external" : "internal"));
}

2065
static ssize_t metadata_store(struct mddev *mddev, const char *buf, size_t len)
2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082
{
	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);

2083
static ssize_t can_clear_show(struct mddev *mddev, char *page)
2084 2085 2086 2087 2088 2089 2090 2091 2092 2093
{
	int len;
	if (mddev->bitmap)
		len = sprintf(page, "%s\n", (mddev->bitmap->need_sync ?
					     "false" : "true"));
	else
		len = sprintf(page, "\n");
	return len;
}

2094
static ssize_t can_clear_store(struct mddev *mddev, const char *buf, size_t len)
2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111
{
	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);

2112
static ssize_t
2113
behind_writes_used_show(struct mddev *mddev, char *page)
2114 2115 2116 2117 2118 2119 2120 2121
{
	if (mddev->bitmap == NULL)
		return sprintf(page, "0\n");
	return sprintf(page, "%lu\n",
		       mddev->bitmap->behind_writes_used);
}

static ssize_t
2122
behind_writes_used_reset(struct mddev *mddev, const char *buf, size_t len)
2123 2124 2125 2126 2127 2128 2129 2130 2131 2132
{
	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);

2133 2134 2135 2136 2137
static struct attribute *md_bitmap_attrs[] = {
	&bitmap_location.attr,
	&bitmap_timeout.attr,
	&bitmap_backlog.attr,
	&bitmap_chunksize.attr,
2138 2139
	&bitmap_metadata.attr,
	&bitmap_can_clear.attr,
2140
	&max_backlog_used.attr,
2141 2142 2143 2144 2145 2146 2147
	NULL
};
struct attribute_group md_bitmap_group = {
	.name = "bitmap",
	.attrs = md_bitmap_attrs,
};