xref: /openbmc/linux/fs/gfs2/rgrp.c (revision 74a22e8f)
1 /*
2  * Copyright (C) Sistina Software, Inc.  1997-2003 All rights reserved.
3  * Copyright (C) 2004-2008 Red Hat, Inc.  All rights reserved.
4  *
5  * This copyrighted material is made available to anyone wishing to use,
6  * modify, copy, or redistribute it subject to the terms and conditions
7  * of the GNU General Public License version 2.
8  */
9 
10 #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
11 
12 #include <linux/slab.h>
13 #include <linux/spinlock.h>
14 #include <linux/completion.h>
15 #include <linux/buffer_head.h>
16 #include <linux/fs.h>
17 #include <linux/gfs2_ondisk.h>
18 #include <linux/prefetch.h>
19 #include <linux/blkdev.h>
20 #include <linux/rbtree.h>
21 #include <linux/random.h>
22 
23 #include "gfs2.h"
24 #include "incore.h"
25 #include "glock.h"
26 #include "glops.h"
27 #include "lops.h"
28 #include "meta_io.h"
29 #include "quota.h"
30 #include "rgrp.h"
31 #include "super.h"
32 #include "trans.h"
33 #include "util.h"
34 #include "log.h"
35 #include "inode.h"
36 #include "trace_gfs2.h"
37 #include "dir.h"
38 
39 #define BFITNOENT ((u32)~0)
40 #define NO_BLOCK ((u64)~0)
41 
42 #if BITS_PER_LONG == 32
43 #define LBITMASK   (0x55555555UL)
44 #define LBITSKIP55 (0x55555555UL)
45 #define LBITSKIP00 (0x00000000UL)
46 #else
47 #define LBITMASK   (0x5555555555555555UL)
48 #define LBITSKIP55 (0x5555555555555555UL)
49 #define LBITSKIP00 (0x0000000000000000UL)
50 #endif
51 
52 /*
53  * These routines are used by the resource group routines (rgrp.c)
54  * to keep track of block allocation.  Each block is represented by two
55  * bits.  So, each byte represents GFS2_NBBY (i.e. 4) blocks.
56  *
57  * 0 = Free
58  * 1 = Used (not metadata)
59  * 2 = Unlinked (still in use) inode
60  * 3 = Used (metadata)
61  */
62 
63 struct gfs2_extent {
64 	struct gfs2_rbm rbm;
65 	u32 len;
66 };
67 
68 static const char valid_change[16] = {
69 	        /* current */
70 	/* n */ 0, 1, 1, 1,
71 	/* e */ 1, 0, 0, 0,
72 	/* w */ 0, 0, 0, 1,
73 	        1, 0, 0, 0
74 };
75 
76 static int gfs2_rbm_find(struct gfs2_rbm *rbm, u8 state, u32 *minext,
77 			 const struct gfs2_inode *ip, bool nowrap);
78 
79 
80 /**
81  * gfs2_setbit - Set a bit in the bitmaps
82  * @rbm: The position of the bit to set
83  * @do_clone: Also set the clone bitmap, if it exists
84  * @new_state: the new state of the block
85  *
86  */
87 
88 static inline void gfs2_setbit(const struct gfs2_rbm *rbm, bool do_clone,
89 			       unsigned char new_state)
90 {
91 	unsigned char *byte1, *byte2, *end, cur_state;
92 	struct gfs2_bitmap *bi = rbm_bi(rbm);
93 	unsigned int buflen = bi->bi_bytes;
94 	const unsigned int bit = (rbm->offset % GFS2_NBBY) * GFS2_BIT_SIZE;
95 
96 	byte1 = bi->bi_bh->b_data + bi->bi_offset + (rbm->offset / GFS2_NBBY);
97 	end = bi->bi_bh->b_data + bi->bi_offset + buflen;
98 
99 	BUG_ON(byte1 >= end);
100 
101 	cur_state = (*byte1 >> bit) & GFS2_BIT_MASK;
102 
103 	if (unlikely(!valid_change[new_state * 4 + cur_state])) {
104 		struct gfs2_sbd *sdp = rbm->rgd->rd_sbd;
105 
106 		fs_warn(sdp, "buf_blk = 0x%x old_state=%d, new_state=%d\n",
107 			rbm->offset, cur_state, new_state);
108 		fs_warn(sdp, "rgrp=0x%llx bi_start=0x%x biblk: 0x%llx\n",
109 			(unsigned long long)rbm->rgd->rd_addr, bi->bi_start,
110 			(unsigned long long)bi->bi_bh->b_blocknr);
111 		fs_warn(sdp, "bi_offset=0x%x bi_bytes=0x%x block=0x%llx\n",
112 			bi->bi_offset, bi->bi_bytes,
113 			(unsigned long long)gfs2_rbm_to_block(rbm));
114 		dump_stack();
115 		gfs2_consist_rgrpd(rbm->rgd);
116 		return;
117 	}
118 	*byte1 ^= (cur_state ^ new_state) << bit;
119 
120 	if (do_clone && bi->bi_clone) {
121 		byte2 = bi->bi_clone + bi->bi_offset + (rbm->offset / GFS2_NBBY);
122 		cur_state = (*byte2 >> bit) & GFS2_BIT_MASK;
123 		*byte2 ^= (cur_state ^ new_state) << bit;
124 	}
125 }
126 
127 /**
128  * gfs2_testbit - test a bit in the bitmaps
129  * @rbm: The bit to test
130  * @use_clone: If true, test the clone bitmap, not the official bitmap.
131  *
132  * Some callers like gfs2_unaligned_extlen need to test the clone bitmaps,
133  * not the "real" bitmaps, to avoid allocating recently freed blocks.
134  *
135  * Returns: The two bit block state of the requested bit
136  */
137 
138 static inline u8 gfs2_testbit(const struct gfs2_rbm *rbm, bool use_clone)
139 {
140 	struct gfs2_bitmap *bi = rbm_bi(rbm);
141 	const u8 *buffer;
142 	const u8 *byte;
143 	unsigned int bit;
144 
145 	if (use_clone && bi->bi_clone)
146 		buffer = bi->bi_clone;
147 	else
148 		buffer = bi->bi_bh->b_data;
149 	buffer += bi->bi_offset;
150 	byte = buffer + (rbm->offset / GFS2_NBBY);
151 	bit = (rbm->offset % GFS2_NBBY) * GFS2_BIT_SIZE;
152 
153 	return (*byte >> bit) & GFS2_BIT_MASK;
154 }
155 
156 /**
157  * gfs2_bit_search
158  * @ptr: Pointer to bitmap data
159  * @mask: Mask to use (normally 0x55555.... but adjusted for search start)
160  * @state: The state we are searching for
161  *
162  * We xor the bitmap data with a patter which is the bitwise opposite
163  * of what we are looking for, this gives rise to a pattern of ones
164  * wherever there is a match. Since we have two bits per entry, we
165  * take this pattern, shift it down by one place and then and it with
166  * the original. All the even bit positions (0,2,4, etc) then represent
167  * successful matches, so we mask with 0x55555..... to remove the unwanted
168  * odd bit positions.
169  *
170  * This allows searching of a whole u64 at once (32 blocks) with a
171  * single test (on 64 bit arches).
172  */
173 
174 static inline u64 gfs2_bit_search(const __le64 *ptr, u64 mask, u8 state)
175 {
176 	u64 tmp;
177 	static const u64 search[] = {
178 		[0] = 0xffffffffffffffffULL,
179 		[1] = 0xaaaaaaaaaaaaaaaaULL,
180 		[2] = 0x5555555555555555ULL,
181 		[3] = 0x0000000000000000ULL,
182 	};
183 	tmp = le64_to_cpu(*ptr) ^ search[state];
184 	tmp &= (tmp >> 1);
185 	tmp &= mask;
186 	return tmp;
187 }
188 
189 /**
190  * rs_cmp - multi-block reservation range compare
191  * @blk: absolute file system block number of the new reservation
192  * @len: number of blocks in the new reservation
193  * @rs: existing reservation to compare against
194  *
195  * returns: 1 if the block range is beyond the reach of the reservation
196  *         -1 if the block range is before the start of the reservation
197  *          0 if the block range overlaps with the reservation
198  */
199 static inline int rs_cmp(u64 blk, u32 len, struct gfs2_blkreserv *rs)
200 {
201 	u64 startblk = gfs2_rbm_to_block(&rs->rs_rbm);
202 
203 	if (blk >= startblk + rs->rs_free)
204 		return 1;
205 	if (blk + len - 1 < startblk)
206 		return -1;
207 	return 0;
208 }
209 
210 /**
211  * gfs2_bitfit - Search an rgrp's bitmap buffer to find a bit-pair representing
212  *       a block in a given allocation state.
213  * @buf: the buffer that holds the bitmaps
214  * @len: the length (in bytes) of the buffer
215  * @goal: start search at this block's bit-pair (within @buffer)
216  * @state: GFS2_BLKST_XXX the state of the block we're looking for.
217  *
218  * Scope of @goal and returned block number is only within this bitmap buffer,
219  * not entire rgrp or filesystem.  @buffer will be offset from the actual
220  * beginning of a bitmap block buffer, skipping any header structures, but
221  * headers are always a multiple of 64 bits long so that the buffer is
222  * always aligned to a 64 bit boundary.
223  *
224  * The size of the buffer is in bytes, but is it assumed that it is
225  * always ok to read a complete multiple of 64 bits at the end
226  * of the block in case the end is no aligned to a natural boundary.
227  *
228  * Return: the block number (bitmap buffer scope) that was found
229  */
230 
231 static u32 gfs2_bitfit(const u8 *buf, const unsigned int len,
232 		       u32 goal, u8 state)
233 {
234 	u32 spoint = (goal << 1) & ((8*sizeof(u64)) - 1);
235 	const __le64 *ptr = ((__le64 *)buf) + (goal >> 5);
236 	const __le64 *end = (__le64 *)(buf + ALIGN(len, sizeof(u64)));
237 	u64 tmp;
238 	u64 mask = 0x5555555555555555ULL;
239 	u32 bit;
240 
241 	/* Mask off bits we don't care about at the start of the search */
242 	mask <<= spoint;
243 	tmp = gfs2_bit_search(ptr, mask, state);
244 	ptr++;
245 	while(tmp == 0 && ptr < end) {
246 		tmp = gfs2_bit_search(ptr, 0x5555555555555555ULL, state);
247 		ptr++;
248 	}
249 	/* Mask off any bits which are more than len bytes from the start */
250 	if (ptr == end && (len & (sizeof(u64) - 1)))
251 		tmp &= (((u64)~0) >> (64 - 8*(len & (sizeof(u64) - 1))));
252 	/* Didn't find anything, so return */
253 	if (tmp == 0)
254 		return BFITNOENT;
255 	ptr--;
256 	bit = __ffs64(tmp);
257 	bit /= 2;	/* two bits per entry in the bitmap */
258 	return (((const unsigned char *)ptr - buf) * GFS2_NBBY) + bit;
259 }
260 
261 /**
262  * gfs2_rbm_from_block - Set the rbm based upon rgd and block number
263  * @rbm: The rbm with rgd already set correctly
264  * @block: The block number (filesystem relative)
265  *
266  * This sets the bi and offset members of an rbm based on a
267  * resource group and a filesystem relative block number. The
268  * resource group must be set in the rbm on entry, the bi and
269  * offset members will be set by this function.
270  *
271  * Returns: 0 on success, or an error code
272  */
273 
274 static int gfs2_rbm_from_block(struct gfs2_rbm *rbm, u64 block)
275 {
276 	if (!rgrp_contains_block(rbm->rgd, block))
277 		return -E2BIG;
278 	rbm->bii = 0;
279 	rbm->offset = block - rbm->rgd->rd_data0;
280 	/* Check if the block is within the first block */
281 	if (rbm->offset < rbm_bi(rbm)->bi_blocks)
282 		return 0;
283 
284 	/* Adjust for the size diff between gfs2_meta_header and gfs2_rgrp */
285 	rbm->offset += (sizeof(struct gfs2_rgrp) -
286 			sizeof(struct gfs2_meta_header)) * GFS2_NBBY;
287 	rbm->bii = rbm->offset / rbm->rgd->rd_sbd->sd_blocks_per_bitmap;
288 	rbm->offset -= rbm->bii * rbm->rgd->rd_sbd->sd_blocks_per_bitmap;
289 	return 0;
290 }
291 
292 /**
293  * gfs2_rbm_incr - increment an rbm structure
294  * @rbm: The rbm with rgd already set correctly
295  *
296  * This function takes an existing rbm structure and increments it to the next
297  * viable block offset.
298  *
299  * Returns: If incrementing the offset would cause the rbm to go past the
300  *          end of the rgrp, true is returned, otherwise false.
301  *
302  */
303 
304 static bool gfs2_rbm_incr(struct gfs2_rbm *rbm)
305 {
306 	if (rbm->offset + 1 < rbm_bi(rbm)->bi_blocks) { /* in the same bitmap */
307 		rbm->offset++;
308 		return false;
309 	}
310 	if (rbm->bii == rbm->rgd->rd_length - 1) /* at the last bitmap */
311 		return true;
312 
313 	rbm->offset = 0;
314 	rbm->bii++;
315 	return false;
316 }
317 
318 /**
319  * gfs2_unaligned_extlen - Look for free blocks which are not byte aligned
320  * @rbm: Position to search (value/result)
321  * @n_unaligned: Number of unaligned blocks to check
322  * @len: Decremented for each block found (terminate on zero)
323  *
324  * Returns: true if a non-free block is encountered
325  */
326 
327 static bool gfs2_unaligned_extlen(struct gfs2_rbm *rbm, u32 n_unaligned, u32 *len)
328 {
329 	u32 n;
330 	u8 res;
331 
332 	for (n = 0; n < n_unaligned; n++) {
333 		res = gfs2_testbit(rbm, true);
334 		if (res != GFS2_BLKST_FREE)
335 			return true;
336 		(*len)--;
337 		if (*len == 0)
338 			return true;
339 		if (gfs2_rbm_incr(rbm))
340 			return true;
341 	}
342 
343 	return false;
344 }
345 
346 /**
347  * gfs2_free_extlen - Return extent length of free blocks
348  * @rrbm: Starting position
349  * @len: Max length to check
350  *
351  * Starting at the block specified by the rbm, see how many free blocks
352  * there are, not reading more than len blocks ahead. This can be done
353  * using memchr_inv when the blocks are byte aligned, but has to be done
354  * on a block by block basis in case of unaligned blocks. Also this
355  * function can cope with bitmap boundaries (although it must stop on
356  * a resource group boundary)
357  *
358  * Returns: Number of free blocks in the extent
359  */
360 
361 static u32 gfs2_free_extlen(const struct gfs2_rbm *rrbm, u32 len)
362 {
363 	struct gfs2_rbm rbm = *rrbm;
364 	u32 n_unaligned = rbm.offset & 3;
365 	u32 size = len;
366 	u32 bytes;
367 	u32 chunk_size;
368 	u8 *ptr, *start, *end;
369 	u64 block;
370 	struct gfs2_bitmap *bi;
371 
372 	if (n_unaligned &&
373 	    gfs2_unaligned_extlen(&rbm, 4 - n_unaligned, &len))
374 		goto out;
375 
376 	n_unaligned = len & 3;
377 	/* Start is now byte aligned */
378 	while (len > 3) {
379 		bi = rbm_bi(&rbm);
380 		start = bi->bi_bh->b_data;
381 		if (bi->bi_clone)
382 			start = bi->bi_clone;
383 		start += bi->bi_offset;
384 		end = start + bi->bi_bytes;
385 		BUG_ON(rbm.offset & 3);
386 		start += (rbm.offset / GFS2_NBBY);
387 		bytes = min_t(u32, len / GFS2_NBBY, (end - start));
388 		ptr = memchr_inv(start, 0, bytes);
389 		chunk_size = ((ptr == NULL) ? bytes : (ptr - start));
390 		chunk_size *= GFS2_NBBY;
391 		BUG_ON(len < chunk_size);
392 		len -= chunk_size;
393 		block = gfs2_rbm_to_block(&rbm);
394 		if (gfs2_rbm_from_block(&rbm, block + chunk_size)) {
395 			n_unaligned = 0;
396 			break;
397 		}
398 		if (ptr) {
399 			n_unaligned = 3;
400 			break;
401 		}
402 		n_unaligned = len & 3;
403 	}
404 
405 	/* Deal with any bits left over at the end */
406 	if (n_unaligned)
407 		gfs2_unaligned_extlen(&rbm, n_unaligned, &len);
408 out:
409 	return size - len;
410 }
411 
412 /**
413  * gfs2_bitcount - count the number of bits in a certain state
414  * @rgd: the resource group descriptor
415  * @buffer: the buffer that holds the bitmaps
416  * @buflen: the length (in bytes) of the buffer
417  * @state: the state of the block we're looking for
418  *
419  * Returns: The number of bits
420  */
421 
422 static u32 gfs2_bitcount(struct gfs2_rgrpd *rgd, const u8 *buffer,
423 			 unsigned int buflen, u8 state)
424 {
425 	const u8 *byte = buffer;
426 	const u8 *end = buffer + buflen;
427 	const u8 state1 = state << 2;
428 	const u8 state2 = state << 4;
429 	const u8 state3 = state << 6;
430 	u32 count = 0;
431 
432 	for (; byte < end; byte++) {
433 		if (((*byte) & 0x03) == state)
434 			count++;
435 		if (((*byte) & 0x0C) == state1)
436 			count++;
437 		if (((*byte) & 0x30) == state2)
438 			count++;
439 		if (((*byte) & 0xC0) == state3)
440 			count++;
441 	}
442 
443 	return count;
444 }
445 
446 /**
447  * gfs2_rgrp_verify - Verify that a resource group is consistent
448  * @rgd: the rgrp
449  *
450  */
451 
452 void gfs2_rgrp_verify(struct gfs2_rgrpd *rgd)
453 {
454 	struct gfs2_sbd *sdp = rgd->rd_sbd;
455 	struct gfs2_bitmap *bi = NULL;
456 	u32 length = rgd->rd_length;
457 	u32 count[4], tmp;
458 	int buf, x;
459 
460 	memset(count, 0, 4 * sizeof(u32));
461 
462 	/* Count # blocks in each of 4 possible allocation states */
463 	for (buf = 0; buf < length; buf++) {
464 		bi = rgd->rd_bits + buf;
465 		for (x = 0; x < 4; x++)
466 			count[x] += gfs2_bitcount(rgd,
467 						  bi->bi_bh->b_data +
468 						  bi->bi_offset,
469 						  bi->bi_bytes, x);
470 	}
471 
472 	if (count[0] != rgd->rd_free) {
473 		if (gfs2_consist_rgrpd(rgd))
474 			fs_err(sdp, "free data mismatch:  %u != %u\n",
475 			       count[0], rgd->rd_free);
476 		return;
477 	}
478 
479 	tmp = rgd->rd_data - rgd->rd_free - rgd->rd_dinodes;
480 	if (count[1] != tmp) {
481 		if (gfs2_consist_rgrpd(rgd))
482 			fs_err(sdp, "used data mismatch:  %u != %u\n",
483 			       count[1], tmp);
484 		return;
485 	}
486 
487 	if (count[2] + count[3] != rgd->rd_dinodes) {
488 		if (gfs2_consist_rgrpd(rgd))
489 			fs_err(sdp, "used metadata mismatch:  %u != %u\n",
490 			       count[2] + count[3], rgd->rd_dinodes);
491 		return;
492 	}
493 }
494 
495 /**
496  * gfs2_blk2rgrpd - Find resource group for a given data/meta block number
497  * @sdp: The GFS2 superblock
498  * @blk: The data block number
499  * @exact: True if this needs to be an exact match
500  *
501  * The @exact argument should be set to true by most callers. The exception
502  * is when we need to match blocks which are not represented by the rgrp
503  * bitmap, but which are part of the rgrp (i.e. padding blocks) which are
504  * there for alignment purposes. Another way of looking at it is that @exact
505  * matches only valid data/metadata blocks, but with @exact false, it will
506  * match any block within the extent of the rgrp.
507  *
508  * Returns: The resource group, or NULL if not found
509  */
510 
511 struct gfs2_rgrpd *gfs2_blk2rgrpd(struct gfs2_sbd *sdp, u64 blk, bool exact)
512 {
513 	struct rb_node *n, *next;
514 	struct gfs2_rgrpd *cur;
515 
516 	spin_lock(&sdp->sd_rindex_spin);
517 	n = sdp->sd_rindex_tree.rb_node;
518 	while (n) {
519 		cur = rb_entry(n, struct gfs2_rgrpd, rd_node);
520 		next = NULL;
521 		if (blk < cur->rd_addr)
522 			next = n->rb_left;
523 		else if (blk >= cur->rd_data0 + cur->rd_data)
524 			next = n->rb_right;
525 		if (next == NULL) {
526 			spin_unlock(&sdp->sd_rindex_spin);
527 			if (exact) {
528 				if (blk < cur->rd_addr)
529 					return NULL;
530 				if (blk >= cur->rd_data0 + cur->rd_data)
531 					return NULL;
532 			}
533 			return cur;
534 		}
535 		n = next;
536 	}
537 	spin_unlock(&sdp->sd_rindex_spin);
538 
539 	return NULL;
540 }
541 
542 /**
543  * gfs2_rgrpd_get_first - get the first Resource Group in the filesystem
544  * @sdp: The GFS2 superblock
545  *
546  * Returns: The first rgrp in the filesystem
547  */
548 
549 struct gfs2_rgrpd *gfs2_rgrpd_get_first(struct gfs2_sbd *sdp)
550 {
551 	const struct rb_node *n;
552 	struct gfs2_rgrpd *rgd;
553 
554 	spin_lock(&sdp->sd_rindex_spin);
555 	n = rb_first(&sdp->sd_rindex_tree);
556 	rgd = rb_entry(n, struct gfs2_rgrpd, rd_node);
557 	spin_unlock(&sdp->sd_rindex_spin);
558 
559 	return rgd;
560 }
561 
562 /**
563  * gfs2_rgrpd_get_next - get the next RG
564  * @rgd: the resource group descriptor
565  *
566  * Returns: The next rgrp
567  */
568 
569 struct gfs2_rgrpd *gfs2_rgrpd_get_next(struct gfs2_rgrpd *rgd)
570 {
571 	struct gfs2_sbd *sdp = rgd->rd_sbd;
572 	const struct rb_node *n;
573 
574 	spin_lock(&sdp->sd_rindex_spin);
575 	n = rb_next(&rgd->rd_node);
576 	if (n == NULL)
577 		n = rb_first(&sdp->sd_rindex_tree);
578 
579 	if (unlikely(&rgd->rd_node == n)) {
580 		spin_unlock(&sdp->sd_rindex_spin);
581 		return NULL;
582 	}
583 	rgd = rb_entry(n, struct gfs2_rgrpd, rd_node);
584 	spin_unlock(&sdp->sd_rindex_spin);
585 	return rgd;
586 }
587 
588 void check_and_update_goal(struct gfs2_inode *ip)
589 {
590 	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
591 	if (!ip->i_goal || gfs2_blk2rgrpd(sdp, ip->i_goal, 1) == NULL)
592 		ip->i_goal = ip->i_no_addr;
593 }
594 
595 void gfs2_free_clones(struct gfs2_rgrpd *rgd)
596 {
597 	int x;
598 
599 	for (x = 0; x < rgd->rd_length; x++) {
600 		struct gfs2_bitmap *bi = rgd->rd_bits + x;
601 		kfree(bi->bi_clone);
602 		bi->bi_clone = NULL;
603 	}
604 }
605 
606 /**
607  * gfs2_rsqa_alloc - make sure we have a reservation assigned to the inode
608  *                 plus a quota allocations data structure, if necessary
609  * @ip: the inode for this reservation
610  */
611 int gfs2_rsqa_alloc(struct gfs2_inode *ip)
612 {
613 	return gfs2_qa_alloc(ip);
614 }
615 
616 static void dump_rs(struct seq_file *seq, const struct gfs2_blkreserv *rs)
617 {
618 	struct gfs2_inode *ip = container_of(rs, struct gfs2_inode, i_res);
619 
620 	gfs2_print_dbg(seq, "  B: n:%llu s:%llu b:%u f:%u\n",
621 		       (unsigned long long)ip->i_no_addr,
622 		       (unsigned long long)gfs2_rbm_to_block(&rs->rs_rbm),
623 		       rs->rs_rbm.offset, rs->rs_free);
624 }
625 
626 /**
627  * __rs_deltree - remove a multi-block reservation from the rgd tree
628  * @rs: The reservation to remove
629  *
630  */
631 static void __rs_deltree(struct gfs2_blkreserv *rs)
632 {
633 	struct gfs2_rgrpd *rgd;
634 
635 	if (!gfs2_rs_active(rs))
636 		return;
637 
638 	rgd = rs->rs_rbm.rgd;
639 	trace_gfs2_rs(rs, TRACE_RS_TREEDEL);
640 	rb_erase(&rs->rs_node, &rgd->rd_rstree);
641 	RB_CLEAR_NODE(&rs->rs_node);
642 
643 	if (rs->rs_free) {
644 		u64 last_block = gfs2_rbm_to_block(&rs->rs_rbm) +
645 				 rs->rs_free - 1;
646 		struct gfs2_rbm last_rbm = { .rgd = rs->rs_rbm.rgd, };
647 		struct gfs2_bitmap *start, *last;
648 
649 		/* return reserved blocks to the rgrp */
650 		BUG_ON(rs->rs_rbm.rgd->rd_reserved < rs->rs_free);
651 		rs->rs_rbm.rgd->rd_reserved -= rs->rs_free;
652 		/* The rgrp extent failure point is likely not to increase;
653 		   it will only do so if the freed blocks are somehow
654 		   contiguous with a span of free blocks that follows. Still,
655 		   it will force the number to be recalculated later. */
656 		rgd->rd_extfail_pt += rs->rs_free;
657 		rs->rs_free = 0;
658 		if (gfs2_rbm_from_block(&last_rbm, last_block))
659 			return;
660 		start = rbm_bi(&rs->rs_rbm);
661 		last = rbm_bi(&last_rbm);
662 		do
663 			clear_bit(GBF_FULL, &start->bi_flags);
664 		while (start++ != last);
665 	}
666 }
667 
668 /**
669  * gfs2_rs_deltree - remove a multi-block reservation from the rgd tree
670  * @rs: The reservation to remove
671  *
672  */
673 void gfs2_rs_deltree(struct gfs2_blkreserv *rs)
674 {
675 	struct gfs2_rgrpd *rgd;
676 
677 	rgd = rs->rs_rbm.rgd;
678 	if (rgd) {
679 		spin_lock(&rgd->rd_rsspin);
680 		__rs_deltree(rs);
681 		BUG_ON(rs->rs_free);
682 		spin_unlock(&rgd->rd_rsspin);
683 	}
684 }
685 
686 /**
687  * gfs2_rsqa_delete - delete a multi-block reservation and quota allocation
688  * @ip: The inode for this reservation
689  * @wcount: The inode's write count, or NULL
690  *
691  */
692 void gfs2_rsqa_delete(struct gfs2_inode *ip, atomic_t *wcount)
693 {
694 	down_write(&ip->i_rw_mutex);
695 	if ((wcount == NULL) || (atomic_read(wcount) <= 1))
696 		gfs2_rs_deltree(&ip->i_res);
697 	up_write(&ip->i_rw_mutex);
698 	gfs2_qa_delete(ip, wcount);
699 }
700 
701 /**
702  * return_all_reservations - return all reserved blocks back to the rgrp.
703  * @rgd: the rgrp that needs its space back
704  *
705  * We previously reserved a bunch of blocks for allocation. Now we need to
706  * give them back. This leave the reservation structures in tact, but removes
707  * all of their corresponding "no-fly zones".
708  */
709 static void return_all_reservations(struct gfs2_rgrpd *rgd)
710 {
711 	struct rb_node *n;
712 	struct gfs2_blkreserv *rs;
713 
714 	spin_lock(&rgd->rd_rsspin);
715 	while ((n = rb_first(&rgd->rd_rstree))) {
716 		rs = rb_entry(n, struct gfs2_blkreserv, rs_node);
717 		__rs_deltree(rs);
718 	}
719 	spin_unlock(&rgd->rd_rsspin);
720 }
721 
722 void gfs2_clear_rgrpd(struct gfs2_sbd *sdp)
723 {
724 	struct rb_node *n;
725 	struct gfs2_rgrpd *rgd;
726 	struct gfs2_glock *gl;
727 
728 	while ((n = rb_first(&sdp->sd_rindex_tree))) {
729 		rgd = rb_entry(n, struct gfs2_rgrpd, rd_node);
730 		gl = rgd->rd_gl;
731 
732 		rb_erase(n, &sdp->sd_rindex_tree);
733 
734 		if (gl) {
735 			glock_clear_object(gl, rgd);
736 			gfs2_rgrp_brelse(rgd);
737 			gfs2_glock_put(gl);
738 		}
739 
740 		gfs2_free_clones(rgd);
741 		kfree(rgd->rd_bits);
742 		rgd->rd_bits = NULL;
743 		return_all_reservations(rgd);
744 		kmem_cache_free(gfs2_rgrpd_cachep, rgd);
745 	}
746 }
747 
748 static void gfs2_rindex_print(const struct gfs2_rgrpd *rgd)
749 {
750 	struct gfs2_sbd *sdp = rgd->rd_sbd;
751 
752 	fs_info(sdp, "ri_addr = %llu\n", (unsigned long long)rgd->rd_addr);
753 	fs_info(sdp, "ri_length = %u\n", rgd->rd_length);
754 	fs_info(sdp, "ri_data0 = %llu\n", (unsigned long long)rgd->rd_data0);
755 	fs_info(sdp, "ri_data = %u\n", rgd->rd_data);
756 	fs_info(sdp, "ri_bitbytes = %u\n", rgd->rd_bitbytes);
757 }
758 
759 /**
760  * gfs2_compute_bitstructs - Compute the bitmap sizes
761  * @rgd: The resource group descriptor
762  *
763  * Calculates bitmap descriptors, one for each block that contains bitmap data
764  *
765  * Returns: errno
766  */
767 
768 static int compute_bitstructs(struct gfs2_rgrpd *rgd)
769 {
770 	struct gfs2_sbd *sdp = rgd->rd_sbd;
771 	struct gfs2_bitmap *bi;
772 	u32 length = rgd->rd_length; /* # blocks in hdr & bitmap */
773 	u32 bytes_left, bytes;
774 	int x;
775 
776 	if (!length)
777 		return -EINVAL;
778 
779 	rgd->rd_bits = kcalloc(length, sizeof(struct gfs2_bitmap), GFP_NOFS);
780 	if (!rgd->rd_bits)
781 		return -ENOMEM;
782 
783 	bytes_left = rgd->rd_bitbytes;
784 
785 	for (x = 0; x < length; x++) {
786 		bi = rgd->rd_bits + x;
787 
788 		bi->bi_flags = 0;
789 		/* small rgrp; bitmap stored completely in header block */
790 		if (length == 1) {
791 			bytes = bytes_left;
792 			bi->bi_offset = sizeof(struct gfs2_rgrp);
793 			bi->bi_start = 0;
794 			bi->bi_bytes = bytes;
795 			bi->bi_blocks = bytes * GFS2_NBBY;
796 		/* header block */
797 		} else if (x == 0) {
798 			bytes = sdp->sd_sb.sb_bsize - sizeof(struct gfs2_rgrp);
799 			bi->bi_offset = sizeof(struct gfs2_rgrp);
800 			bi->bi_start = 0;
801 			bi->bi_bytes = bytes;
802 			bi->bi_blocks = bytes * GFS2_NBBY;
803 		/* last block */
804 		} else if (x + 1 == length) {
805 			bytes = bytes_left;
806 			bi->bi_offset = sizeof(struct gfs2_meta_header);
807 			bi->bi_start = rgd->rd_bitbytes - bytes_left;
808 			bi->bi_bytes = bytes;
809 			bi->bi_blocks = bytes * GFS2_NBBY;
810 		/* other blocks */
811 		} else {
812 			bytes = sdp->sd_sb.sb_bsize -
813 				sizeof(struct gfs2_meta_header);
814 			bi->bi_offset = sizeof(struct gfs2_meta_header);
815 			bi->bi_start = rgd->rd_bitbytes - bytes_left;
816 			bi->bi_bytes = bytes;
817 			bi->bi_blocks = bytes * GFS2_NBBY;
818 		}
819 
820 		bytes_left -= bytes;
821 	}
822 
823 	if (bytes_left) {
824 		gfs2_consist_rgrpd(rgd);
825 		return -EIO;
826 	}
827 	bi = rgd->rd_bits + (length - 1);
828 	if ((bi->bi_start + bi->bi_bytes) * GFS2_NBBY != rgd->rd_data) {
829 		if (gfs2_consist_rgrpd(rgd)) {
830 			gfs2_rindex_print(rgd);
831 			fs_err(sdp, "start=%u len=%u offset=%u\n",
832 			       bi->bi_start, bi->bi_bytes, bi->bi_offset);
833 		}
834 		return -EIO;
835 	}
836 
837 	return 0;
838 }
839 
840 /**
841  * gfs2_ri_total - Total up the file system space, according to the rindex.
842  * @sdp: the filesystem
843  *
844  */
845 u64 gfs2_ri_total(struct gfs2_sbd *sdp)
846 {
847 	u64 total_data = 0;
848 	struct inode *inode = sdp->sd_rindex;
849 	struct gfs2_inode *ip = GFS2_I(inode);
850 	char buf[sizeof(struct gfs2_rindex)];
851 	int error, rgrps;
852 
853 	for (rgrps = 0;; rgrps++) {
854 		loff_t pos = rgrps * sizeof(struct gfs2_rindex);
855 
856 		if (pos + sizeof(struct gfs2_rindex) > i_size_read(inode))
857 			break;
858 		error = gfs2_internal_read(ip, buf, &pos,
859 					   sizeof(struct gfs2_rindex));
860 		if (error != sizeof(struct gfs2_rindex))
861 			break;
862 		total_data += be32_to_cpu(((struct gfs2_rindex *)buf)->ri_data);
863 	}
864 	return total_data;
865 }
866 
867 static int rgd_insert(struct gfs2_rgrpd *rgd)
868 {
869 	struct gfs2_sbd *sdp = rgd->rd_sbd;
870 	struct rb_node **newn = &sdp->sd_rindex_tree.rb_node, *parent = NULL;
871 
872 	/* Figure out where to put new node */
873 	while (*newn) {
874 		struct gfs2_rgrpd *cur = rb_entry(*newn, struct gfs2_rgrpd,
875 						  rd_node);
876 
877 		parent = *newn;
878 		if (rgd->rd_addr < cur->rd_addr)
879 			newn = &((*newn)->rb_left);
880 		else if (rgd->rd_addr > cur->rd_addr)
881 			newn = &((*newn)->rb_right);
882 		else
883 			return -EEXIST;
884 	}
885 
886 	rb_link_node(&rgd->rd_node, parent, newn);
887 	rb_insert_color(&rgd->rd_node, &sdp->sd_rindex_tree);
888 	sdp->sd_rgrps++;
889 	return 0;
890 }
891 
892 /**
893  * read_rindex_entry - Pull in a new resource index entry from the disk
894  * @ip: Pointer to the rindex inode
895  *
896  * Returns: 0 on success, > 0 on EOF, error code otherwise
897  */
898 
899 static int read_rindex_entry(struct gfs2_inode *ip)
900 {
901 	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
902 	const unsigned bsize = sdp->sd_sb.sb_bsize;
903 	loff_t pos = sdp->sd_rgrps * sizeof(struct gfs2_rindex);
904 	struct gfs2_rindex buf;
905 	int error;
906 	struct gfs2_rgrpd *rgd;
907 
908 	if (pos >= i_size_read(&ip->i_inode))
909 		return 1;
910 
911 	error = gfs2_internal_read(ip, (char *)&buf, &pos,
912 				   sizeof(struct gfs2_rindex));
913 
914 	if (error != sizeof(struct gfs2_rindex))
915 		return (error == 0) ? 1 : error;
916 
917 	rgd = kmem_cache_zalloc(gfs2_rgrpd_cachep, GFP_NOFS);
918 	error = -ENOMEM;
919 	if (!rgd)
920 		return error;
921 
922 	rgd->rd_sbd = sdp;
923 	rgd->rd_addr = be64_to_cpu(buf.ri_addr);
924 	rgd->rd_length = be32_to_cpu(buf.ri_length);
925 	rgd->rd_data0 = be64_to_cpu(buf.ri_data0);
926 	rgd->rd_data = be32_to_cpu(buf.ri_data);
927 	rgd->rd_bitbytes = be32_to_cpu(buf.ri_bitbytes);
928 	spin_lock_init(&rgd->rd_rsspin);
929 
930 	error = compute_bitstructs(rgd);
931 	if (error)
932 		goto fail;
933 
934 	error = gfs2_glock_get(sdp, rgd->rd_addr,
935 			       &gfs2_rgrp_glops, CREATE, &rgd->rd_gl);
936 	if (error)
937 		goto fail;
938 
939 	rgd->rd_rgl = (struct gfs2_rgrp_lvb *)rgd->rd_gl->gl_lksb.sb_lvbptr;
940 	rgd->rd_flags &= ~(GFS2_RDF_UPTODATE | GFS2_RDF_PREFERRED);
941 	if (rgd->rd_data > sdp->sd_max_rg_data)
942 		sdp->sd_max_rg_data = rgd->rd_data;
943 	spin_lock(&sdp->sd_rindex_spin);
944 	error = rgd_insert(rgd);
945 	spin_unlock(&sdp->sd_rindex_spin);
946 	if (!error) {
947 		glock_set_object(rgd->rd_gl, rgd);
948 		rgd->rd_gl->gl_vm.start = (rgd->rd_addr * bsize) & PAGE_MASK;
949 		rgd->rd_gl->gl_vm.end = PAGE_ALIGN((rgd->rd_addr +
950 						    rgd->rd_length) * bsize) - 1;
951 		return 0;
952 	}
953 
954 	error = 0; /* someone else read in the rgrp; free it and ignore it */
955 	gfs2_glock_put(rgd->rd_gl);
956 
957 fail:
958 	kfree(rgd->rd_bits);
959 	rgd->rd_bits = NULL;
960 	kmem_cache_free(gfs2_rgrpd_cachep, rgd);
961 	return error;
962 }
963 
964 /**
965  * set_rgrp_preferences - Run all the rgrps, selecting some we prefer to use
966  * @sdp: the GFS2 superblock
967  *
968  * The purpose of this function is to select a subset of the resource groups
969  * and mark them as PREFERRED. We do it in such a way that each node prefers
970  * to use a unique set of rgrps to minimize glock contention.
971  */
972 static void set_rgrp_preferences(struct gfs2_sbd *sdp)
973 {
974 	struct gfs2_rgrpd *rgd, *first;
975 	int i;
976 
977 	/* Skip an initial number of rgrps, based on this node's journal ID.
978 	   That should start each node out on its own set. */
979 	rgd = gfs2_rgrpd_get_first(sdp);
980 	for (i = 0; i < sdp->sd_lockstruct.ls_jid; i++)
981 		rgd = gfs2_rgrpd_get_next(rgd);
982 	first = rgd;
983 
984 	do {
985 		rgd->rd_flags |= GFS2_RDF_PREFERRED;
986 		for (i = 0; i < sdp->sd_journals; i++) {
987 			rgd = gfs2_rgrpd_get_next(rgd);
988 			if (!rgd || rgd == first)
989 				break;
990 		}
991 	} while (rgd && rgd != first);
992 }
993 
994 /**
995  * gfs2_ri_update - Pull in a new resource index from the disk
996  * @ip: pointer to the rindex inode
997  *
998  * Returns: 0 on successful update, error code otherwise
999  */
1000 
1001 static int gfs2_ri_update(struct gfs2_inode *ip)
1002 {
1003 	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
1004 	int error;
1005 
1006 	do {
1007 		error = read_rindex_entry(ip);
1008 	} while (error == 0);
1009 
1010 	if (error < 0)
1011 		return error;
1012 
1013 	set_rgrp_preferences(sdp);
1014 
1015 	sdp->sd_rindex_uptodate = 1;
1016 	return 0;
1017 }
1018 
1019 /**
1020  * gfs2_rindex_update - Update the rindex if required
1021  * @sdp: The GFS2 superblock
1022  *
1023  * We grab a lock on the rindex inode to make sure that it doesn't
1024  * change whilst we are performing an operation. We keep this lock
1025  * for quite long periods of time compared to other locks. This
1026  * doesn't matter, since it is shared and it is very, very rarely
1027  * accessed in the exclusive mode (i.e. only when expanding the filesystem).
1028  *
1029  * This makes sure that we're using the latest copy of the resource index
1030  * special file, which might have been updated if someone expanded the
1031  * filesystem (via gfs2_grow utility), which adds new resource groups.
1032  *
1033  * Returns: 0 on succeess, error code otherwise
1034  */
1035 
1036 int gfs2_rindex_update(struct gfs2_sbd *sdp)
1037 {
1038 	struct gfs2_inode *ip = GFS2_I(sdp->sd_rindex);
1039 	struct gfs2_glock *gl = ip->i_gl;
1040 	struct gfs2_holder ri_gh;
1041 	int error = 0;
1042 	int unlock_required = 0;
1043 
1044 	/* Read new copy from disk if we don't have the latest */
1045 	if (!sdp->sd_rindex_uptodate) {
1046 		if (!gfs2_glock_is_locked_by_me(gl)) {
1047 			error = gfs2_glock_nq_init(gl, LM_ST_SHARED, 0, &ri_gh);
1048 			if (error)
1049 				return error;
1050 			unlock_required = 1;
1051 		}
1052 		if (!sdp->sd_rindex_uptodate)
1053 			error = gfs2_ri_update(ip);
1054 		if (unlock_required)
1055 			gfs2_glock_dq_uninit(&ri_gh);
1056 	}
1057 
1058 	return error;
1059 }
1060 
1061 static void gfs2_rgrp_in(struct gfs2_rgrpd *rgd, const void *buf)
1062 {
1063 	const struct gfs2_rgrp *str = buf;
1064 	u32 rg_flags;
1065 
1066 	rg_flags = be32_to_cpu(str->rg_flags);
1067 	rg_flags &= ~GFS2_RDF_MASK;
1068 	rgd->rd_flags &= GFS2_RDF_MASK;
1069 	rgd->rd_flags |= rg_flags;
1070 	rgd->rd_free = be32_to_cpu(str->rg_free);
1071 	rgd->rd_dinodes = be32_to_cpu(str->rg_dinodes);
1072 	rgd->rd_igeneration = be64_to_cpu(str->rg_igeneration);
1073 	/* rd_data0, rd_data and rd_bitbytes already set from rindex */
1074 }
1075 
1076 static void gfs2_rgrp_ondisk2lvb(struct gfs2_rgrp_lvb *rgl, const void *buf)
1077 {
1078 	const struct gfs2_rgrp *str = buf;
1079 
1080 	rgl->rl_magic = cpu_to_be32(GFS2_MAGIC);
1081 	rgl->rl_flags = str->rg_flags;
1082 	rgl->rl_free = str->rg_free;
1083 	rgl->rl_dinodes = str->rg_dinodes;
1084 	rgl->rl_igeneration = str->rg_igeneration;
1085 	rgl->__pad = 0UL;
1086 }
1087 
1088 static void gfs2_rgrp_out(struct gfs2_rgrpd *rgd, void *buf)
1089 {
1090 	struct gfs2_rgrpd *next = gfs2_rgrpd_get_next(rgd);
1091 	struct gfs2_rgrp *str = buf;
1092 	u32 crc;
1093 
1094 	str->rg_flags = cpu_to_be32(rgd->rd_flags & ~GFS2_RDF_MASK);
1095 	str->rg_free = cpu_to_be32(rgd->rd_free);
1096 	str->rg_dinodes = cpu_to_be32(rgd->rd_dinodes);
1097 	if (next == NULL)
1098 		str->rg_skip = 0;
1099 	else if (next->rd_addr > rgd->rd_addr)
1100 		str->rg_skip = cpu_to_be32(next->rd_addr - rgd->rd_addr);
1101 	str->rg_igeneration = cpu_to_be64(rgd->rd_igeneration);
1102 	str->rg_data0 = cpu_to_be64(rgd->rd_data0);
1103 	str->rg_data = cpu_to_be32(rgd->rd_data);
1104 	str->rg_bitbytes = cpu_to_be32(rgd->rd_bitbytes);
1105 	str->rg_crc = 0;
1106 	crc = gfs2_disk_hash(buf, sizeof(struct gfs2_rgrp));
1107 	str->rg_crc = cpu_to_be32(crc);
1108 
1109 	memset(&str->rg_reserved, 0, sizeof(str->rg_reserved));
1110 	gfs2_rgrp_ondisk2lvb(rgd->rd_rgl, buf);
1111 }
1112 
1113 static int gfs2_rgrp_lvb_valid(struct gfs2_rgrpd *rgd)
1114 {
1115 	struct gfs2_rgrp_lvb *rgl = rgd->rd_rgl;
1116 	struct gfs2_rgrp *str = (struct gfs2_rgrp *)rgd->rd_bits[0].bi_bh->b_data;
1117 	int valid = 1;
1118 
1119 	if (rgl->rl_flags != str->rg_flags) {
1120 		printk(KERN_WARNING "GFS2: rgd: %llu lvb flag mismatch %u/%u",
1121 		       (unsigned long long)rgd->rd_addr,
1122 		       be32_to_cpu(rgl->rl_flags), be32_to_cpu(str->rg_flags));
1123 		valid = 0;
1124 	}
1125 	if (rgl->rl_free != str->rg_free) {
1126 		printk(KERN_WARNING "GFS2: rgd: %llu lvb free mismatch %u/%u",
1127 		       (unsigned long long)rgd->rd_addr,
1128 		       be32_to_cpu(rgl->rl_free), be32_to_cpu(str->rg_free));
1129 		valid = 0;
1130 	}
1131 	if (rgl->rl_dinodes != str->rg_dinodes) {
1132 		printk(KERN_WARNING "GFS2: rgd: %llu lvb dinode mismatch %u/%u",
1133 		       (unsigned long long)rgd->rd_addr,
1134 		       be32_to_cpu(rgl->rl_dinodes),
1135 		       be32_to_cpu(str->rg_dinodes));
1136 		valid = 0;
1137 	}
1138 	if (rgl->rl_igeneration != str->rg_igeneration) {
1139 		printk(KERN_WARNING "GFS2: rgd: %llu lvb igen mismatch "
1140 		       "%llu/%llu", (unsigned long long)rgd->rd_addr,
1141 		       (unsigned long long)be64_to_cpu(rgl->rl_igeneration),
1142 		       (unsigned long long)be64_to_cpu(str->rg_igeneration));
1143 		valid = 0;
1144 	}
1145 	return valid;
1146 }
1147 
1148 static u32 count_unlinked(struct gfs2_rgrpd *rgd)
1149 {
1150 	struct gfs2_bitmap *bi;
1151 	const u32 length = rgd->rd_length;
1152 	const u8 *buffer = NULL;
1153 	u32 i, goal, count = 0;
1154 
1155 	for (i = 0, bi = rgd->rd_bits; i < length; i++, bi++) {
1156 		goal = 0;
1157 		buffer = bi->bi_bh->b_data + bi->bi_offset;
1158 		WARN_ON(!buffer_uptodate(bi->bi_bh));
1159 		while (goal < bi->bi_blocks) {
1160 			goal = gfs2_bitfit(buffer, bi->bi_bytes, goal,
1161 					   GFS2_BLKST_UNLINKED);
1162 			if (goal == BFITNOENT)
1163 				break;
1164 			count++;
1165 			goal++;
1166 		}
1167 	}
1168 
1169 	return count;
1170 }
1171 
1172 
1173 /**
1174  * gfs2_rgrp_bh_get - Read in a RG's header and bitmaps
1175  * @rgd: the struct gfs2_rgrpd describing the RG to read in
1176  *
1177  * Read in all of a Resource Group's header and bitmap blocks.
1178  * Caller must eventually call gfs2_rgrp_brelse() to free the bitmaps.
1179  *
1180  * Returns: errno
1181  */
1182 
1183 static int gfs2_rgrp_bh_get(struct gfs2_rgrpd *rgd)
1184 {
1185 	struct gfs2_sbd *sdp = rgd->rd_sbd;
1186 	struct gfs2_glock *gl = rgd->rd_gl;
1187 	unsigned int length = rgd->rd_length;
1188 	struct gfs2_bitmap *bi;
1189 	unsigned int x, y;
1190 	int error;
1191 
1192 	if (rgd->rd_bits[0].bi_bh != NULL)
1193 		return 0;
1194 
1195 	for (x = 0; x < length; x++) {
1196 		bi = rgd->rd_bits + x;
1197 		error = gfs2_meta_read(gl, rgd->rd_addr + x, 0, 0, &bi->bi_bh);
1198 		if (error)
1199 			goto fail;
1200 	}
1201 
1202 	for (y = length; y--;) {
1203 		bi = rgd->rd_bits + y;
1204 		error = gfs2_meta_wait(sdp, bi->bi_bh);
1205 		if (error)
1206 			goto fail;
1207 		if (gfs2_metatype_check(sdp, bi->bi_bh, y ? GFS2_METATYPE_RB :
1208 					      GFS2_METATYPE_RG)) {
1209 			error = -EIO;
1210 			goto fail;
1211 		}
1212 	}
1213 
1214 	if (!(rgd->rd_flags & GFS2_RDF_UPTODATE)) {
1215 		for (x = 0; x < length; x++)
1216 			clear_bit(GBF_FULL, &rgd->rd_bits[x].bi_flags);
1217 		gfs2_rgrp_in(rgd, (rgd->rd_bits[0].bi_bh)->b_data);
1218 		rgd->rd_flags |= (GFS2_RDF_UPTODATE | GFS2_RDF_CHECK);
1219 		rgd->rd_free_clone = rgd->rd_free;
1220 		/* max out the rgrp allocation failure point */
1221 		rgd->rd_extfail_pt = rgd->rd_free;
1222 	}
1223 	if (cpu_to_be32(GFS2_MAGIC) != rgd->rd_rgl->rl_magic) {
1224 		rgd->rd_rgl->rl_unlinked = cpu_to_be32(count_unlinked(rgd));
1225 		gfs2_rgrp_ondisk2lvb(rgd->rd_rgl,
1226 				     rgd->rd_bits[0].bi_bh->b_data);
1227 	}
1228 	else if (sdp->sd_args.ar_rgrplvb) {
1229 		if (!gfs2_rgrp_lvb_valid(rgd)){
1230 			gfs2_consist_rgrpd(rgd);
1231 			error = -EIO;
1232 			goto fail;
1233 		}
1234 		if (rgd->rd_rgl->rl_unlinked == 0)
1235 			rgd->rd_flags &= ~GFS2_RDF_CHECK;
1236 	}
1237 	return 0;
1238 
1239 fail:
1240 	while (x--) {
1241 		bi = rgd->rd_bits + x;
1242 		brelse(bi->bi_bh);
1243 		bi->bi_bh = NULL;
1244 		gfs2_assert_warn(sdp, !bi->bi_clone);
1245 	}
1246 
1247 	return error;
1248 }
1249 
1250 static int update_rgrp_lvb(struct gfs2_rgrpd *rgd)
1251 {
1252 	u32 rl_flags;
1253 
1254 	if (rgd->rd_flags & GFS2_RDF_UPTODATE)
1255 		return 0;
1256 
1257 	if (cpu_to_be32(GFS2_MAGIC) != rgd->rd_rgl->rl_magic)
1258 		return gfs2_rgrp_bh_get(rgd);
1259 
1260 	rl_flags = be32_to_cpu(rgd->rd_rgl->rl_flags);
1261 	rl_flags &= ~GFS2_RDF_MASK;
1262 	rgd->rd_flags &= GFS2_RDF_MASK;
1263 	rgd->rd_flags |= (rl_flags | GFS2_RDF_CHECK);
1264 	if (rgd->rd_rgl->rl_unlinked == 0)
1265 		rgd->rd_flags &= ~GFS2_RDF_CHECK;
1266 	rgd->rd_free = be32_to_cpu(rgd->rd_rgl->rl_free);
1267 	rgd->rd_free_clone = rgd->rd_free;
1268 	rgd->rd_dinodes = be32_to_cpu(rgd->rd_rgl->rl_dinodes);
1269 	rgd->rd_igeneration = be64_to_cpu(rgd->rd_rgl->rl_igeneration);
1270 	return 0;
1271 }
1272 
1273 int gfs2_rgrp_go_lock(struct gfs2_holder *gh)
1274 {
1275 	struct gfs2_rgrpd *rgd = gh->gh_gl->gl_object;
1276 	struct gfs2_sbd *sdp = rgd->rd_sbd;
1277 
1278 	if (gh->gh_flags & GL_SKIP && sdp->sd_args.ar_rgrplvb)
1279 		return 0;
1280 	return gfs2_rgrp_bh_get(rgd);
1281 }
1282 
1283 /**
1284  * gfs2_rgrp_brelse - Release RG bitmaps read in with gfs2_rgrp_bh_get()
1285  * @rgd: The resource group
1286  *
1287  */
1288 
1289 void gfs2_rgrp_brelse(struct gfs2_rgrpd *rgd)
1290 {
1291 	int x, length = rgd->rd_length;
1292 
1293 	for (x = 0; x < length; x++) {
1294 		struct gfs2_bitmap *bi = rgd->rd_bits + x;
1295 		if (bi->bi_bh) {
1296 			brelse(bi->bi_bh);
1297 			bi->bi_bh = NULL;
1298 		}
1299 	}
1300 
1301 }
1302 
1303 /**
1304  * gfs2_rgrp_go_unlock - Unlock a rgrp glock
1305  * @gh: The glock holder for the resource group
1306  *
1307  */
1308 
1309 void gfs2_rgrp_go_unlock(struct gfs2_holder *gh)
1310 {
1311 	struct gfs2_rgrpd *rgd = gh->gh_gl->gl_object;
1312 	int demote_requested = test_bit(GLF_DEMOTE, &gh->gh_gl->gl_flags) |
1313 		test_bit(GLF_PENDING_DEMOTE, &gh->gh_gl->gl_flags);
1314 
1315 	if (rgd && demote_requested)
1316 		gfs2_rgrp_brelse(rgd);
1317 }
1318 
1319 int gfs2_rgrp_send_discards(struct gfs2_sbd *sdp, u64 offset,
1320 			     struct buffer_head *bh,
1321 			     const struct gfs2_bitmap *bi, unsigned minlen, u64 *ptrimmed)
1322 {
1323 	struct super_block *sb = sdp->sd_vfs;
1324 	u64 blk;
1325 	sector_t start = 0;
1326 	sector_t nr_blks = 0;
1327 	int rv;
1328 	unsigned int x;
1329 	u32 trimmed = 0;
1330 	u8 diff;
1331 
1332 	for (x = 0; x < bi->bi_bytes; x++) {
1333 		const u8 *clone = bi->bi_clone ? bi->bi_clone : bi->bi_bh->b_data;
1334 		clone += bi->bi_offset;
1335 		clone += x;
1336 		if (bh) {
1337 			const u8 *orig = bh->b_data + bi->bi_offset + x;
1338 			diff = ~(*orig | (*orig >> 1)) & (*clone | (*clone >> 1));
1339 		} else {
1340 			diff = ~(*clone | (*clone >> 1));
1341 		}
1342 		diff &= 0x55;
1343 		if (diff == 0)
1344 			continue;
1345 		blk = offset + ((bi->bi_start + x) * GFS2_NBBY);
1346 		while(diff) {
1347 			if (diff & 1) {
1348 				if (nr_blks == 0)
1349 					goto start_new_extent;
1350 				if ((start + nr_blks) != blk) {
1351 					if (nr_blks >= minlen) {
1352 						rv = sb_issue_discard(sb,
1353 							start, nr_blks,
1354 							GFP_NOFS, 0);
1355 						if (rv)
1356 							goto fail;
1357 						trimmed += nr_blks;
1358 					}
1359 					nr_blks = 0;
1360 start_new_extent:
1361 					start = blk;
1362 				}
1363 				nr_blks++;
1364 			}
1365 			diff >>= 2;
1366 			blk++;
1367 		}
1368 	}
1369 	if (nr_blks >= minlen) {
1370 		rv = sb_issue_discard(sb, start, nr_blks, GFP_NOFS, 0);
1371 		if (rv)
1372 			goto fail;
1373 		trimmed += nr_blks;
1374 	}
1375 	if (ptrimmed)
1376 		*ptrimmed = trimmed;
1377 	return 0;
1378 
1379 fail:
1380 	if (sdp->sd_args.ar_discard)
1381 		fs_warn(sdp, "error %d on discard request, turning discards off for this filesystem\n", rv);
1382 	sdp->sd_args.ar_discard = 0;
1383 	return -EIO;
1384 }
1385 
1386 /**
1387  * gfs2_fitrim - Generate discard requests for unused bits of the filesystem
1388  * @filp: Any file on the filesystem
1389  * @argp: Pointer to the arguments (also used to pass result)
1390  *
1391  * Returns: 0 on success, otherwise error code
1392  */
1393 
1394 int gfs2_fitrim(struct file *filp, void __user *argp)
1395 {
1396 	struct inode *inode = file_inode(filp);
1397 	struct gfs2_sbd *sdp = GFS2_SB(inode);
1398 	struct request_queue *q = bdev_get_queue(sdp->sd_vfs->s_bdev);
1399 	struct buffer_head *bh;
1400 	struct gfs2_rgrpd *rgd;
1401 	struct gfs2_rgrpd *rgd_end;
1402 	struct gfs2_holder gh;
1403 	struct fstrim_range r;
1404 	int ret = 0;
1405 	u64 amt;
1406 	u64 trimmed = 0;
1407 	u64 start, end, minlen;
1408 	unsigned int x;
1409 	unsigned bs_shift = sdp->sd_sb.sb_bsize_shift;
1410 
1411 	if (!capable(CAP_SYS_ADMIN))
1412 		return -EPERM;
1413 
1414 	if (!blk_queue_discard(q))
1415 		return -EOPNOTSUPP;
1416 
1417 	if (copy_from_user(&r, argp, sizeof(r)))
1418 		return -EFAULT;
1419 
1420 	ret = gfs2_rindex_update(sdp);
1421 	if (ret)
1422 		return ret;
1423 
1424 	start = r.start >> bs_shift;
1425 	end = start + (r.len >> bs_shift);
1426 	minlen = max_t(u64, r.minlen,
1427 		       q->limits.discard_granularity) >> bs_shift;
1428 
1429 	if (end <= start || minlen > sdp->sd_max_rg_data)
1430 		return -EINVAL;
1431 
1432 	rgd = gfs2_blk2rgrpd(sdp, start, 0);
1433 	rgd_end = gfs2_blk2rgrpd(sdp, end, 0);
1434 
1435 	if ((gfs2_rgrpd_get_first(sdp) == gfs2_rgrpd_get_next(rgd_end))
1436 	    && (start > rgd_end->rd_data0 + rgd_end->rd_data))
1437 		return -EINVAL; /* start is beyond the end of the fs */
1438 
1439 	while (1) {
1440 
1441 		ret = gfs2_glock_nq_init(rgd->rd_gl, LM_ST_EXCLUSIVE, 0, &gh);
1442 		if (ret)
1443 			goto out;
1444 
1445 		if (!(rgd->rd_flags & GFS2_RGF_TRIMMED)) {
1446 			/* Trim each bitmap in the rgrp */
1447 			for (x = 0; x < rgd->rd_length; x++) {
1448 				struct gfs2_bitmap *bi = rgd->rd_bits + x;
1449 				ret = gfs2_rgrp_send_discards(sdp,
1450 						rgd->rd_data0, NULL, bi, minlen,
1451 						&amt);
1452 				if (ret) {
1453 					gfs2_glock_dq_uninit(&gh);
1454 					goto out;
1455 				}
1456 				trimmed += amt;
1457 			}
1458 
1459 			/* Mark rgrp as having been trimmed */
1460 			ret = gfs2_trans_begin(sdp, RES_RG_HDR, 0);
1461 			if (ret == 0) {
1462 				bh = rgd->rd_bits[0].bi_bh;
1463 				rgd->rd_flags |= GFS2_RGF_TRIMMED;
1464 				gfs2_trans_add_meta(rgd->rd_gl, bh);
1465 				gfs2_rgrp_out(rgd, bh->b_data);
1466 				gfs2_trans_end(sdp);
1467 			}
1468 		}
1469 		gfs2_glock_dq_uninit(&gh);
1470 
1471 		if (rgd == rgd_end)
1472 			break;
1473 
1474 		rgd = gfs2_rgrpd_get_next(rgd);
1475 	}
1476 
1477 out:
1478 	r.len = trimmed << bs_shift;
1479 	if (copy_to_user(argp, &r, sizeof(r)))
1480 		return -EFAULT;
1481 
1482 	return ret;
1483 }
1484 
1485 /**
1486  * rs_insert - insert a new multi-block reservation into the rgrp's rb_tree
1487  * @ip: the inode structure
1488  *
1489  */
1490 static void rs_insert(struct gfs2_inode *ip)
1491 {
1492 	struct rb_node **newn, *parent = NULL;
1493 	int rc;
1494 	struct gfs2_blkreserv *rs = &ip->i_res;
1495 	struct gfs2_rgrpd *rgd = rs->rs_rbm.rgd;
1496 	u64 fsblock = gfs2_rbm_to_block(&rs->rs_rbm);
1497 
1498 	BUG_ON(gfs2_rs_active(rs));
1499 
1500 	spin_lock(&rgd->rd_rsspin);
1501 	newn = &rgd->rd_rstree.rb_node;
1502 	while (*newn) {
1503 		struct gfs2_blkreserv *cur =
1504 			rb_entry(*newn, struct gfs2_blkreserv, rs_node);
1505 
1506 		parent = *newn;
1507 		rc = rs_cmp(fsblock, rs->rs_free, cur);
1508 		if (rc > 0)
1509 			newn = &((*newn)->rb_right);
1510 		else if (rc < 0)
1511 			newn = &((*newn)->rb_left);
1512 		else {
1513 			spin_unlock(&rgd->rd_rsspin);
1514 			WARN_ON(1);
1515 			return;
1516 		}
1517 	}
1518 
1519 	rb_link_node(&rs->rs_node, parent, newn);
1520 	rb_insert_color(&rs->rs_node, &rgd->rd_rstree);
1521 
1522 	/* Do our rgrp accounting for the reservation */
1523 	rgd->rd_reserved += rs->rs_free; /* blocks reserved */
1524 	spin_unlock(&rgd->rd_rsspin);
1525 	trace_gfs2_rs(rs, TRACE_RS_INSERT);
1526 }
1527 
1528 /**
1529  * rgd_free - return the number of free blocks we can allocate.
1530  * @rgd: the resource group
1531  *
1532  * This function returns the number of free blocks for an rgrp.
1533  * That's the clone-free blocks (blocks that are free, not including those
1534  * still being used for unlinked files that haven't been deleted.)
1535  *
1536  * It also subtracts any blocks reserved by someone else, but does not
1537  * include free blocks that are still part of our current reservation,
1538  * because obviously we can (and will) allocate them.
1539  */
1540 static inline u32 rgd_free(struct gfs2_rgrpd *rgd, struct gfs2_blkreserv *rs)
1541 {
1542 	u32 tot_reserved, tot_free;
1543 
1544 	if (WARN_ON_ONCE(rgd->rd_reserved < rs->rs_free))
1545 		return 0;
1546 	tot_reserved = rgd->rd_reserved - rs->rs_free;
1547 
1548 	if (rgd->rd_free_clone < tot_reserved)
1549 		tot_reserved = 0;
1550 
1551 	tot_free = rgd->rd_free_clone - tot_reserved;
1552 
1553 	return tot_free;
1554 }
1555 
1556 /**
1557  * rg_mblk_search - find a group of multiple free blocks to form a reservation
1558  * @rgd: the resource group descriptor
1559  * @ip: pointer to the inode for which we're reserving blocks
1560  * @ap: the allocation parameters
1561  *
1562  */
1563 
1564 static void rg_mblk_search(struct gfs2_rgrpd *rgd, struct gfs2_inode *ip,
1565 			   const struct gfs2_alloc_parms *ap)
1566 {
1567 	struct gfs2_rbm rbm = { .rgd = rgd, };
1568 	u64 goal;
1569 	struct gfs2_blkreserv *rs = &ip->i_res;
1570 	u32 extlen;
1571 	u32 free_blocks = rgd_free(rgd, rs);
1572 	int ret;
1573 	struct inode *inode = &ip->i_inode;
1574 
1575 	if (S_ISDIR(inode->i_mode))
1576 		extlen = 1;
1577 	else {
1578 		extlen = max_t(u32, atomic_read(&ip->i_sizehint), ap->target);
1579 		extlen = clamp(extlen, (u32)RGRP_RSRV_MINBLKS, free_blocks);
1580 	}
1581 	if ((rgd->rd_free_clone < rgd->rd_reserved) || (free_blocks < extlen))
1582 		return;
1583 
1584 	/* Find bitmap block that contains bits for goal block */
1585 	if (rgrp_contains_block(rgd, ip->i_goal))
1586 		goal = ip->i_goal;
1587 	else
1588 		goal = rgd->rd_last_alloc + rgd->rd_data0;
1589 
1590 	if (WARN_ON(gfs2_rbm_from_block(&rbm, goal)))
1591 		return;
1592 
1593 	ret = gfs2_rbm_find(&rbm, GFS2_BLKST_FREE, &extlen, ip, true);
1594 	if (ret == 0) {
1595 		rs->rs_rbm = rbm;
1596 		rs->rs_free = extlen;
1597 		rs_insert(ip);
1598 	} else {
1599 		if (goal == rgd->rd_last_alloc + rgd->rd_data0)
1600 			rgd->rd_last_alloc = 0;
1601 	}
1602 }
1603 
1604 /**
1605  * gfs2_next_unreserved_block - Return next block that is not reserved
1606  * @rgd: The resource group
1607  * @block: The starting block
1608  * @length: The required length
1609  * @ip: Ignore any reservations for this inode
1610  *
1611  * If the block does not appear in any reservation, then return the
1612  * block number unchanged. If it does appear in the reservation, then
1613  * keep looking through the tree of reservations in order to find the
1614  * first block number which is not reserved.
1615  */
1616 
1617 static u64 gfs2_next_unreserved_block(struct gfs2_rgrpd *rgd, u64 block,
1618 				      u32 length,
1619 				      const struct gfs2_inode *ip)
1620 {
1621 	struct gfs2_blkreserv *rs;
1622 	struct rb_node *n;
1623 	int rc;
1624 
1625 	spin_lock(&rgd->rd_rsspin);
1626 	n = rgd->rd_rstree.rb_node;
1627 	while (n) {
1628 		rs = rb_entry(n, struct gfs2_blkreserv, rs_node);
1629 		rc = rs_cmp(block, length, rs);
1630 		if (rc < 0)
1631 			n = n->rb_left;
1632 		else if (rc > 0)
1633 			n = n->rb_right;
1634 		else
1635 			break;
1636 	}
1637 
1638 	if (n) {
1639 		while ((rs_cmp(block, length, rs) == 0) && (&ip->i_res != rs)) {
1640 			block = gfs2_rbm_to_block(&rs->rs_rbm) + rs->rs_free;
1641 			n = n->rb_right;
1642 			if (n == NULL)
1643 				break;
1644 			rs = rb_entry(n, struct gfs2_blkreserv, rs_node);
1645 		}
1646 	}
1647 
1648 	spin_unlock(&rgd->rd_rsspin);
1649 	return block;
1650 }
1651 
1652 /**
1653  * gfs2_reservation_check_and_update - Check for reservations during block alloc
1654  * @rbm: The current position in the resource group
1655  * @ip: The inode for which we are searching for blocks
1656  * @minext: The minimum extent length
1657  * @maxext: A pointer to the maximum extent structure
1658  *
1659  * This checks the current position in the rgrp to see whether there is
1660  * a reservation covering this block. If not then this function is a
1661  * no-op. If there is, then the position is moved to the end of the
1662  * contiguous reservation(s) so that we are pointing at the first
1663  * non-reserved block.
1664  *
1665  * Returns: 0 if no reservation, 1 if @rbm has changed, otherwise an error
1666  */
1667 
1668 static int gfs2_reservation_check_and_update(struct gfs2_rbm *rbm,
1669 					     const struct gfs2_inode *ip,
1670 					     u32 minext,
1671 					     struct gfs2_extent *maxext)
1672 {
1673 	u64 block = gfs2_rbm_to_block(rbm);
1674 	u32 extlen = 1;
1675 	u64 nblock;
1676 	int ret;
1677 
1678 	/*
1679 	 * If we have a minimum extent length, then skip over any extent
1680 	 * which is less than the min extent length in size.
1681 	 */
1682 	if (minext) {
1683 		extlen = gfs2_free_extlen(rbm, minext);
1684 		if (extlen <= maxext->len)
1685 			goto fail;
1686 	}
1687 
1688 	/*
1689 	 * Check the extent which has been found against the reservations
1690 	 * and skip if parts of it are already reserved
1691 	 */
1692 	nblock = gfs2_next_unreserved_block(rbm->rgd, block, extlen, ip);
1693 	if (nblock == block) {
1694 		if (!minext || extlen >= minext)
1695 			return 0;
1696 
1697 		if (extlen > maxext->len) {
1698 			maxext->len = extlen;
1699 			maxext->rbm = *rbm;
1700 		}
1701 fail:
1702 		nblock = block + extlen;
1703 	}
1704 	ret = gfs2_rbm_from_block(rbm, nblock);
1705 	if (ret < 0)
1706 		return ret;
1707 	return 1;
1708 }
1709 
1710 /**
1711  * gfs2_rbm_find - Look for blocks of a particular state
1712  * @rbm: Value/result starting position and final position
1713  * @state: The state which we want to find
1714  * @minext: Pointer to the requested extent length (NULL for a single block)
1715  *          This is updated to be the actual reservation size.
1716  * @ip: If set, check for reservations
1717  * @nowrap: Stop looking at the end of the rgrp, rather than wrapping
1718  *          around until we've reached the starting point.
1719  *
1720  * Side effects:
1721  * - If looking for free blocks, we set GBF_FULL on each bitmap which
1722  *   has no free blocks in it.
1723  * - If looking for free blocks, we set rd_extfail_pt on each rgrp which
1724  *   has come up short on a free block search.
1725  *
1726  * Returns: 0 on success, -ENOSPC if there is no block of the requested state
1727  */
1728 
1729 static int gfs2_rbm_find(struct gfs2_rbm *rbm, u8 state, u32 *minext,
1730 			 const struct gfs2_inode *ip, bool nowrap)
1731 {
1732 	bool scan_from_start = rbm->bii == 0 && rbm->offset == 0;
1733 	struct buffer_head *bh;
1734 	int last_bii;
1735 	u32 offset;
1736 	u8 *buffer;
1737 	bool wrapped = false;
1738 	int ret;
1739 	struct gfs2_bitmap *bi;
1740 	struct gfs2_extent maxext = { .rbm.rgd = rbm->rgd, };
1741 
1742 	/*
1743 	 * Determine the last bitmap to search.  If we're not starting at the
1744 	 * beginning of a bitmap, we need to search that bitmap twice to scan
1745 	 * the entire resource group.
1746 	 */
1747 	last_bii = rbm->bii - (rbm->offset == 0);
1748 
1749 	while(1) {
1750 		bi = rbm_bi(rbm);
1751 		if ((ip == NULL || !gfs2_rs_active(&ip->i_res)) &&
1752 		    test_bit(GBF_FULL, &bi->bi_flags) &&
1753 		    (state == GFS2_BLKST_FREE))
1754 			goto next_bitmap;
1755 
1756 		bh = bi->bi_bh;
1757 		buffer = bh->b_data + bi->bi_offset;
1758 		WARN_ON(!buffer_uptodate(bh));
1759 		if (state != GFS2_BLKST_UNLINKED && bi->bi_clone)
1760 			buffer = bi->bi_clone + bi->bi_offset;
1761 		offset = gfs2_bitfit(buffer, bi->bi_bytes, rbm->offset, state);
1762 		if (offset == BFITNOENT) {
1763 			if (state == GFS2_BLKST_FREE && rbm->offset == 0)
1764 				set_bit(GBF_FULL, &bi->bi_flags);
1765 			goto next_bitmap;
1766 		}
1767 		rbm->offset = offset;
1768 		if (ip == NULL)
1769 			return 0;
1770 
1771 		ret = gfs2_reservation_check_and_update(rbm, ip,
1772 							minext ? *minext : 0,
1773 							&maxext);
1774 		if (ret == 0)
1775 			return 0;
1776 		if (ret > 0)
1777 			goto next_iter;
1778 		if (ret == -E2BIG) {
1779 			rbm->bii = 0;
1780 			rbm->offset = 0;
1781 			goto res_covered_end_of_rgrp;
1782 		}
1783 		return ret;
1784 
1785 next_bitmap:	/* Find next bitmap in the rgrp */
1786 		rbm->offset = 0;
1787 		rbm->bii++;
1788 		if (rbm->bii == rbm->rgd->rd_length)
1789 			rbm->bii = 0;
1790 res_covered_end_of_rgrp:
1791 		if (rbm->bii == 0) {
1792 			if (wrapped)
1793 				break;
1794 			wrapped = true;
1795 			if (nowrap)
1796 				break;
1797 		}
1798 next_iter:
1799 		/* Have we scanned the entire resource group? */
1800 		if (wrapped && rbm->bii > last_bii)
1801 			break;
1802 	}
1803 
1804 	if (minext == NULL || state != GFS2_BLKST_FREE)
1805 		return -ENOSPC;
1806 
1807 	/* If the extent was too small, and it's smaller than the smallest
1808 	   to have failed before, remember for future reference that it's
1809 	   useless to search this rgrp again for this amount or more. */
1810 	if (wrapped && (scan_from_start || rbm->bii > last_bii) &&
1811 	    *minext < rbm->rgd->rd_extfail_pt)
1812 		rbm->rgd->rd_extfail_pt = *minext;
1813 
1814 	/* If the maximum extent we found is big enough to fulfill the
1815 	   minimum requirements, use it anyway. */
1816 	if (maxext.len) {
1817 		*rbm = maxext.rbm;
1818 		*minext = maxext.len;
1819 		return 0;
1820 	}
1821 
1822 	return -ENOSPC;
1823 }
1824 
1825 /**
1826  * try_rgrp_unlink - Look for any unlinked, allocated, but unused inodes
1827  * @rgd: The rgrp
1828  * @last_unlinked: block address of the last dinode we unlinked
1829  * @skip: block address we should explicitly not unlink
1830  *
1831  * Returns: 0 if no error
1832  *          The inode, if one has been found, in inode.
1833  */
1834 
1835 static void try_rgrp_unlink(struct gfs2_rgrpd *rgd, u64 *last_unlinked, u64 skip)
1836 {
1837 	u64 block;
1838 	struct gfs2_sbd *sdp = rgd->rd_sbd;
1839 	struct gfs2_glock *gl;
1840 	struct gfs2_inode *ip;
1841 	int error;
1842 	int found = 0;
1843 	struct gfs2_rbm rbm = { .rgd = rgd, .bii = 0, .offset = 0 };
1844 
1845 	while (1) {
1846 		down_write(&sdp->sd_log_flush_lock);
1847 		error = gfs2_rbm_find(&rbm, GFS2_BLKST_UNLINKED, NULL, NULL,
1848 				      true);
1849 		up_write(&sdp->sd_log_flush_lock);
1850 		if (error == -ENOSPC)
1851 			break;
1852 		if (WARN_ON_ONCE(error))
1853 			break;
1854 
1855 		block = gfs2_rbm_to_block(&rbm);
1856 		if (gfs2_rbm_from_block(&rbm, block + 1))
1857 			break;
1858 		if (*last_unlinked != NO_BLOCK && block <= *last_unlinked)
1859 			continue;
1860 		if (block == skip)
1861 			continue;
1862 		*last_unlinked = block;
1863 
1864 		error = gfs2_glock_get(sdp, block, &gfs2_iopen_glops, CREATE, &gl);
1865 		if (error)
1866 			continue;
1867 
1868 		/* If the inode is already in cache, we can ignore it here
1869 		 * because the existing inode disposal code will deal with
1870 		 * it when all refs have gone away. Accessing gl_object like
1871 		 * this is not safe in general. Here it is ok because we do
1872 		 * not dereference the pointer, and we only need an approx
1873 		 * answer to whether it is NULL or not.
1874 		 */
1875 		ip = gl->gl_object;
1876 
1877 		if (ip || queue_work(gfs2_delete_workqueue, &gl->gl_delete) == 0)
1878 			gfs2_glock_put(gl);
1879 		else
1880 			found++;
1881 
1882 		/* Limit reclaim to sensible number of tasks */
1883 		if (found > NR_CPUS)
1884 			return;
1885 	}
1886 
1887 	rgd->rd_flags &= ~GFS2_RDF_CHECK;
1888 	return;
1889 }
1890 
1891 /**
1892  * gfs2_rgrp_congested - Use stats to figure out whether an rgrp is congested
1893  * @rgd: The rgrp in question
1894  * @loops: An indication of how picky we can be (0=very, 1=less so)
1895  *
1896  * This function uses the recently added glock statistics in order to
1897  * figure out whether a parciular resource group is suffering from
1898  * contention from multiple nodes. This is done purely on the basis
1899  * of timings, since this is the only data we have to work with and
1900  * our aim here is to reject a resource group which is highly contended
1901  * but (very important) not to do this too often in order to ensure that
1902  * we do not land up introducing fragmentation by changing resource
1903  * groups when not actually required.
1904  *
1905  * The calculation is fairly simple, we want to know whether the SRTTB
1906  * (i.e. smoothed round trip time for blocking operations) to acquire
1907  * the lock for this rgrp's glock is significantly greater than the
1908  * time taken for resource groups on average. We introduce a margin in
1909  * the form of the variable @var which is computed as the sum of the two
1910  * respective variences, and multiplied by a factor depending on @loops
1911  * and whether we have a lot of data to base the decision on. This is
1912  * then tested against the square difference of the means in order to
1913  * decide whether the result is statistically significant or not.
1914  *
1915  * Returns: A boolean verdict on the congestion status
1916  */
1917 
1918 static bool gfs2_rgrp_congested(const struct gfs2_rgrpd *rgd, int loops)
1919 {
1920 	const struct gfs2_glock *gl = rgd->rd_gl;
1921 	const struct gfs2_sbd *sdp = gl->gl_name.ln_sbd;
1922 	struct gfs2_lkstats *st;
1923 	u64 r_dcount, l_dcount;
1924 	u64 l_srttb, a_srttb = 0;
1925 	s64 srttb_diff;
1926 	u64 sqr_diff;
1927 	u64 var;
1928 	int cpu, nonzero = 0;
1929 
1930 	preempt_disable();
1931 	for_each_present_cpu(cpu) {
1932 		st = &per_cpu_ptr(sdp->sd_lkstats, cpu)->lkstats[LM_TYPE_RGRP];
1933 		if (st->stats[GFS2_LKS_SRTTB]) {
1934 			a_srttb += st->stats[GFS2_LKS_SRTTB];
1935 			nonzero++;
1936 		}
1937 	}
1938 	st = &this_cpu_ptr(sdp->sd_lkstats)->lkstats[LM_TYPE_RGRP];
1939 	if (nonzero)
1940 		do_div(a_srttb, nonzero);
1941 	r_dcount = st->stats[GFS2_LKS_DCOUNT];
1942 	var = st->stats[GFS2_LKS_SRTTVARB] +
1943 	      gl->gl_stats.stats[GFS2_LKS_SRTTVARB];
1944 	preempt_enable();
1945 
1946 	l_srttb = gl->gl_stats.stats[GFS2_LKS_SRTTB];
1947 	l_dcount = gl->gl_stats.stats[GFS2_LKS_DCOUNT];
1948 
1949 	if ((l_dcount < 1) || (r_dcount < 1) || (a_srttb == 0))
1950 		return false;
1951 
1952 	srttb_diff = a_srttb - l_srttb;
1953 	sqr_diff = srttb_diff * srttb_diff;
1954 
1955 	var *= 2;
1956 	if (l_dcount < 8 || r_dcount < 8)
1957 		var *= 2;
1958 	if (loops == 1)
1959 		var *= 2;
1960 
1961 	return ((srttb_diff < 0) && (sqr_diff > var));
1962 }
1963 
1964 /**
1965  * gfs2_rgrp_used_recently
1966  * @rs: The block reservation with the rgrp to test
1967  * @msecs: The time limit in milliseconds
1968  *
1969  * Returns: True if the rgrp glock has been used within the time limit
1970  */
1971 static bool gfs2_rgrp_used_recently(const struct gfs2_blkreserv *rs,
1972 				    u64 msecs)
1973 {
1974 	u64 tdiff;
1975 
1976 	tdiff = ktime_to_ns(ktime_sub(ktime_get_real(),
1977                             rs->rs_rbm.rgd->rd_gl->gl_dstamp));
1978 
1979 	return tdiff > (msecs * 1000 * 1000);
1980 }
1981 
1982 static u32 gfs2_orlov_skip(const struct gfs2_inode *ip)
1983 {
1984 	const struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
1985 	u32 skip;
1986 
1987 	get_random_bytes(&skip, sizeof(skip));
1988 	return skip % sdp->sd_rgrps;
1989 }
1990 
1991 static bool gfs2_select_rgrp(struct gfs2_rgrpd **pos, const struct gfs2_rgrpd *begin)
1992 {
1993 	struct gfs2_rgrpd *rgd = *pos;
1994 	struct gfs2_sbd *sdp = rgd->rd_sbd;
1995 
1996 	rgd = gfs2_rgrpd_get_next(rgd);
1997 	if (rgd == NULL)
1998 		rgd = gfs2_rgrpd_get_first(sdp);
1999 	*pos = rgd;
2000 	if (rgd != begin) /* If we didn't wrap */
2001 		return true;
2002 	return false;
2003 }
2004 
2005 /**
2006  * fast_to_acquire - determine if a resource group will be fast to acquire
2007  *
2008  * If this is one of our preferred rgrps, it should be quicker to acquire,
2009  * because we tried to set ourselves up as dlm lock master.
2010  */
2011 static inline int fast_to_acquire(struct gfs2_rgrpd *rgd)
2012 {
2013 	struct gfs2_glock *gl = rgd->rd_gl;
2014 
2015 	if (gl->gl_state != LM_ST_UNLOCKED && list_empty(&gl->gl_holders) &&
2016 	    !test_bit(GLF_DEMOTE_IN_PROGRESS, &gl->gl_flags) &&
2017 	    !test_bit(GLF_DEMOTE, &gl->gl_flags))
2018 		return 1;
2019 	if (rgd->rd_flags & GFS2_RDF_PREFERRED)
2020 		return 1;
2021 	return 0;
2022 }
2023 
2024 /**
2025  * gfs2_inplace_reserve - Reserve space in the filesystem
2026  * @ip: the inode to reserve space for
2027  * @ap: the allocation parameters
2028  *
2029  * We try our best to find an rgrp that has at least ap->target blocks
2030  * available. After a couple of passes (loops == 2), the prospects of finding
2031  * such an rgrp diminish. At this stage, we return the first rgrp that has
2032  * at least ap->min_target blocks available. Either way, we set ap->allowed to
2033  * the number of blocks available in the chosen rgrp.
2034  *
2035  * Returns: 0 on success,
2036  *          -ENOMEM if a suitable rgrp can't be found
2037  *          errno otherwise
2038  */
2039 
2040 int gfs2_inplace_reserve(struct gfs2_inode *ip, struct gfs2_alloc_parms *ap)
2041 {
2042 	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
2043 	struct gfs2_rgrpd *begin = NULL;
2044 	struct gfs2_blkreserv *rs = &ip->i_res;
2045 	int error = 0, rg_locked, flags = 0;
2046 	u64 last_unlinked = NO_BLOCK;
2047 	int loops = 0;
2048 	u32 free_blocks, skip = 0;
2049 
2050 	if (sdp->sd_args.ar_rgrplvb)
2051 		flags |= GL_SKIP;
2052 	if (gfs2_assert_warn(sdp, ap->target))
2053 		return -EINVAL;
2054 	if (gfs2_rs_active(rs)) {
2055 		begin = rs->rs_rbm.rgd;
2056 	} else if (rs->rs_rbm.rgd &&
2057 		   rgrp_contains_block(rs->rs_rbm.rgd, ip->i_goal)) {
2058 		begin = rs->rs_rbm.rgd;
2059 	} else {
2060 		check_and_update_goal(ip);
2061 		rs->rs_rbm.rgd = begin = gfs2_blk2rgrpd(sdp, ip->i_goal, 1);
2062 	}
2063 	if (S_ISDIR(ip->i_inode.i_mode) && (ap->aflags & GFS2_AF_ORLOV))
2064 		skip = gfs2_orlov_skip(ip);
2065 	if (rs->rs_rbm.rgd == NULL)
2066 		return -EBADSLT;
2067 
2068 	while (loops < 3) {
2069 		rg_locked = 1;
2070 
2071 		if (!gfs2_glock_is_locked_by_me(rs->rs_rbm.rgd->rd_gl)) {
2072 			rg_locked = 0;
2073 			if (skip && skip--)
2074 				goto next_rgrp;
2075 			if (!gfs2_rs_active(rs)) {
2076 				if (loops == 0 &&
2077 				    !fast_to_acquire(rs->rs_rbm.rgd))
2078 					goto next_rgrp;
2079 				if ((loops < 2) &&
2080 				    gfs2_rgrp_used_recently(rs, 1000) &&
2081 				    gfs2_rgrp_congested(rs->rs_rbm.rgd, loops))
2082 					goto next_rgrp;
2083 			}
2084 			error = gfs2_glock_nq_init(rs->rs_rbm.rgd->rd_gl,
2085 						   LM_ST_EXCLUSIVE, flags,
2086 						   &ip->i_rgd_gh);
2087 			if (unlikely(error))
2088 				return error;
2089 			if (!gfs2_rs_active(rs) && (loops < 2) &&
2090 			    gfs2_rgrp_congested(rs->rs_rbm.rgd, loops))
2091 				goto skip_rgrp;
2092 			if (sdp->sd_args.ar_rgrplvb) {
2093 				error = update_rgrp_lvb(rs->rs_rbm.rgd);
2094 				if (unlikely(error)) {
2095 					gfs2_glock_dq_uninit(&ip->i_rgd_gh);
2096 					return error;
2097 				}
2098 			}
2099 		}
2100 
2101 		/* Skip unusable resource groups */
2102 		if ((rs->rs_rbm.rgd->rd_flags & (GFS2_RGF_NOALLOC |
2103 						 GFS2_RDF_ERROR)) ||
2104 		    (loops == 0 && ap->target > rs->rs_rbm.rgd->rd_extfail_pt))
2105 			goto skip_rgrp;
2106 
2107 		if (sdp->sd_args.ar_rgrplvb)
2108 			gfs2_rgrp_bh_get(rs->rs_rbm.rgd);
2109 
2110 		/* Get a reservation if we don't already have one */
2111 		if (!gfs2_rs_active(rs))
2112 			rg_mblk_search(rs->rs_rbm.rgd, ip, ap);
2113 
2114 		/* Skip rgrps when we can't get a reservation on first pass */
2115 		if (!gfs2_rs_active(rs) && (loops < 1))
2116 			goto check_rgrp;
2117 
2118 		/* If rgrp has enough free space, use it */
2119 		free_blocks = rgd_free(rs->rs_rbm.rgd, rs);
2120 		if (free_blocks >= ap->target ||
2121 		    (loops == 2 && ap->min_target &&
2122 		     free_blocks >= ap->min_target)) {
2123 			ap->allowed = free_blocks;
2124 			return 0;
2125 		}
2126 check_rgrp:
2127 		/* Check for unlinked inodes which can be reclaimed */
2128 		if (rs->rs_rbm.rgd->rd_flags & GFS2_RDF_CHECK)
2129 			try_rgrp_unlink(rs->rs_rbm.rgd, &last_unlinked,
2130 					ip->i_no_addr);
2131 skip_rgrp:
2132 		/* Drop reservation, if we couldn't use reserved rgrp */
2133 		if (gfs2_rs_active(rs))
2134 			gfs2_rs_deltree(rs);
2135 
2136 		/* Unlock rgrp if required */
2137 		if (!rg_locked)
2138 			gfs2_glock_dq_uninit(&ip->i_rgd_gh);
2139 next_rgrp:
2140 		/* Find the next rgrp, and continue looking */
2141 		if (gfs2_select_rgrp(&rs->rs_rbm.rgd, begin))
2142 			continue;
2143 		if (skip)
2144 			continue;
2145 
2146 		/* If we've scanned all the rgrps, but found no free blocks
2147 		 * then this checks for some less likely conditions before
2148 		 * trying again.
2149 		 */
2150 		loops++;
2151 		/* Check that fs hasn't grown if writing to rindex */
2152 		if (ip == GFS2_I(sdp->sd_rindex) && !sdp->sd_rindex_uptodate) {
2153 			error = gfs2_ri_update(ip);
2154 			if (error)
2155 				return error;
2156 		}
2157 		/* Flushing the log may release space */
2158 		if (loops == 2)
2159 			gfs2_log_flush(sdp, NULL, GFS2_LOG_HEAD_FLUSH_NORMAL |
2160 				       GFS2_LFC_INPLACE_RESERVE);
2161 	}
2162 
2163 	return -ENOSPC;
2164 }
2165 
2166 /**
2167  * gfs2_inplace_release - release an inplace reservation
2168  * @ip: the inode the reservation was taken out on
2169  *
2170  * Release a reservation made by gfs2_inplace_reserve().
2171  */
2172 
2173 void gfs2_inplace_release(struct gfs2_inode *ip)
2174 {
2175 	if (gfs2_holder_initialized(&ip->i_rgd_gh))
2176 		gfs2_glock_dq_uninit(&ip->i_rgd_gh);
2177 }
2178 
2179 /**
2180  * gfs2_alloc_extent - allocate an extent from a given bitmap
2181  * @rbm: the resource group information
2182  * @dinode: TRUE if the first block we allocate is for a dinode
2183  * @n: The extent length (value/result)
2184  *
2185  * Add the bitmap buffer to the transaction.
2186  * Set the found bits to @new_state to change block's allocation state.
2187  */
2188 static void gfs2_alloc_extent(const struct gfs2_rbm *rbm, bool dinode,
2189 			     unsigned int *n)
2190 {
2191 	struct gfs2_rbm pos = { .rgd = rbm->rgd, };
2192 	const unsigned int elen = *n;
2193 	u64 block;
2194 	int ret;
2195 
2196 	*n = 1;
2197 	block = gfs2_rbm_to_block(rbm);
2198 	gfs2_trans_add_meta(rbm->rgd->rd_gl, rbm_bi(rbm)->bi_bh);
2199 	gfs2_setbit(rbm, true, dinode ? GFS2_BLKST_DINODE : GFS2_BLKST_USED);
2200 	block++;
2201 	while (*n < elen) {
2202 		ret = gfs2_rbm_from_block(&pos, block);
2203 		if (ret || gfs2_testbit(&pos, true) != GFS2_BLKST_FREE)
2204 			break;
2205 		gfs2_trans_add_meta(pos.rgd->rd_gl, rbm_bi(&pos)->bi_bh);
2206 		gfs2_setbit(&pos, true, GFS2_BLKST_USED);
2207 		(*n)++;
2208 		block++;
2209 	}
2210 }
2211 
2212 /**
2213  * rgblk_free - Change alloc state of given block(s)
2214  * @sdp: the filesystem
2215  * @rgd: the resource group the blocks are in
2216  * @bstart: the start of a run of blocks to free
2217  * @blen: the length of the block run (all must lie within ONE RG!)
2218  * @new_state: GFS2_BLKST_XXX the after-allocation block state
2219  */
2220 
2221 static void rgblk_free(struct gfs2_sbd *sdp, struct gfs2_rgrpd *rgd,
2222 		       u64 bstart, u32 blen, unsigned char new_state)
2223 {
2224 	struct gfs2_rbm rbm;
2225 	struct gfs2_bitmap *bi, *bi_prev = NULL;
2226 
2227 	rbm.rgd = rgd;
2228 	if (WARN_ON_ONCE(gfs2_rbm_from_block(&rbm, bstart)))
2229 		return;
2230 	while (blen--) {
2231 		bi = rbm_bi(&rbm);
2232 		if (bi != bi_prev) {
2233 			if (!bi->bi_clone) {
2234 				bi->bi_clone = kmalloc(bi->bi_bh->b_size,
2235 						      GFP_NOFS | __GFP_NOFAIL);
2236 				memcpy(bi->bi_clone + bi->bi_offset,
2237 				       bi->bi_bh->b_data + bi->bi_offset,
2238 				       bi->bi_bytes);
2239 			}
2240 			gfs2_trans_add_meta(rbm.rgd->rd_gl, bi->bi_bh);
2241 			bi_prev = bi;
2242 		}
2243 		gfs2_setbit(&rbm, false, new_state);
2244 		gfs2_rbm_incr(&rbm);
2245 	}
2246 }
2247 
2248 /**
2249  * gfs2_rgrp_dump - print out an rgrp
2250  * @seq: The iterator
2251  * @gl: The glock in question
2252  *
2253  */
2254 
2255 void gfs2_rgrp_dump(struct seq_file *seq, struct gfs2_glock *gl)
2256 {
2257 	struct gfs2_rgrpd *rgd = gl->gl_object;
2258 	struct gfs2_blkreserv *trs;
2259 	const struct rb_node *n;
2260 
2261 	if (rgd == NULL)
2262 		return;
2263 	gfs2_print_dbg(seq, " R: n:%llu f:%02x b:%u/%u i:%u r:%u e:%u\n",
2264 		       (unsigned long long)rgd->rd_addr, rgd->rd_flags,
2265 		       rgd->rd_free, rgd->rd_free_clone, rgd->rd_dinodes,
2266 		       rgd->rd_reserved, rgd->rd_extfail_pt);
2267 	if (rgd->rd_sbd->sd_args.ar_rgrplvb) {
2268 		struct gfs2_rgrp_lvb *rgl = rgd->rd_rgl;
2269 
2270 		gfs2_print_dbg(seq, "  L: f:%02x b:%u i:%u\n",
2271 			       be32_to_cpu(rgl->rl_flags),
2272 			       be32_to_cpu(rgl->rl_free),
2273 			       be32_to_cpu(rgl->rl_dinodes));
2274 	}
2275 	spin_lock(&rgd->rd_rsspin);
2276 	for (n = rb_first(&rgd->rd_rstree); n; n = rb_next(&trs->rs_node)) {
2277 		trs = rb_entry(n, struct gfs2_blkreserv, rs_node);
2278 		dump_rs(seq, trs);
2279 	}
2280 	spin_unlock(&rgd->rd_rsspin);
2281 }
2282 
2283 static void gfs2_rgrp_error(struct gfs2_rgrpd *rgd)
2284 {
2285 	struct gfs2_sbd *sdp = rgd->rd_sbd;
2286 	fs_warn(sdp, "rgrp %llu has an error, marking it readonly until umount\n",
2287 		(unsigned long long)rgd->rd_addr);
2288 	fs_warn(sdp, "umount on all nodes and run fsck.gfs2 to fix the error\n");
2289 	gfs2_rgrp_dump(NULL, rgd->rd_gl);
2290 	rgd->rd_flags |= GFS2_RDF_ERROR;
2291 }
2292 
2293 /**
2294  * gfs2_adjust_reservation - Adjust (or remove) a reservation after allocation
2295  * @ip: The inode we have just allocated blocks for
2296  * @rbm: The start of the allocated blocks
2297  * @len: The extent length
2298  *
2299  * Adjusts a reservation after an allocation has taken place. If the
2300  * reservation does not match the allocation, or if it is now empty
2301  * then it is removed.
2302  */
2303 
2304 static void gfs2_adjust_reservation(struct gfs2_inode *ip,
2305 				    const struct gfs2_rbm *rbm, unsigned len)
2306 {
2307 	struct gfs2_blkreserv *rs = &ip->i_res;
2308 	struct gfs2_rgrpd *rgd = rbm->rgd;
2309 	unsigned rlen;
2310 	u64 block;
2311 	int ret;
2312 
2313 	spin_lock(&rgd->rd_rsspin);
2314 	if (gfs2_rs_active(rs)) {
2315 		if (gfs2_rbm_eq(&rs->rs_rbm, rbm)) {
2316 			block = gfs2_rbm_to_block(rbm);
2317 			ret = gfs2_rbm_from_block(&rs->rs_rbm, block + len);
2318 			rlen = min(rs->rs_free, len);
2319 			rs->rs_free -= rlen;
2320 			rgd->rd_reserved -= rlen;
2321 			trace_gfs2_rs(rs, TRACE_RS_CLAIM);
2322 			if (rs->rs_free && !ret)
2323 				goto out;
2324 			/* We used up our block reservation, so we should
2325 			   reserve more blocks next time. */
2326 			atomic_add(RGRP_RSRV_ADDBLKS, &ip->i_sizehint);
2327 		}
2328 		__rs_deltree(rs);
2329 	}
2330 out:
2331 	spin_unlock(&rgd->rd_rsspin);
2332 }
2333 
2334 /**
2335  * gfs2_set_alloc_start - Set starting point for block allocation
2336  * @rbm: The rbm which will be set to the required location
2337  * @ip: The gfs2 inode
2338  * @dinode: Flag to say if allocation includes a new inode
2339  *
2340  * This sets the starting point from the reservation if one is active
2341  * otherwise it falls back to guessing a start point based on the
2342  * inode's goal block or the last allocation point in the rgrp.
2343  */
2344 
2345 static void gfs2_set_alloc_start(struct gfs2_rbm *rbm,
2346 				 const struct gfs2_inode *ip, bool dinode)
2347 {
2348 	u64 goal;
2349 
2350 	if (gfs2_rs_active(&ip->i_res)) {
2351 		*rbm = ip->i_res.rs_rbm;
2352 		return;
2353 	}
2354 
2355 	if (!dinode && rgrp_contains_block(rbm->rgd, ip->i_goal))
2356 		goal = ip->i_goal;
2357 	else
2358 		goal = rbm->rgd->rd_last_alloc + rbm->rgd->rd_data0;
2359 
2360 	if (WARN_ON_ONCE(gfs2_rbm_from_block(rbm, goal))) {
2361 		rbm->bii = 0;
2362 		rbm->offset = 0;
2363 	}
2364 }
2365 
2366 /**
2367  * gfs2_alloc_blocks - Allocate one or more blocks of data and/or a dinode
2368  * @ip: the inode to allocate the block for
2369  * @bn: Used to return the starting block number
2370  * @nblocks: requested number of blocks/extent length (value/result)
2371  * @dinode: 1 if we're allocating a dinode block, else 0
2372  * @generation: the generation number of the inode
2373  *
2374  * Returns: 0 or error
2375  */
2376 
2377 int gfs2_alloc_blocks(struct gfs2_inode *ip, u64 *bn, unsigned int *nblocks,
2378 		      bool dinode, u64 *generation)
2379 {
2380 	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
2381 	struct buffer_head *dibh;
2382 	struct gfs2_rbm rbm = { .rgd = ip->i_res.rs_rbm.rgd, };
2383 	unsigned int ndata;
2384 	u64 block; /* block, within the file system scope */
2385 	int error;
2386 
2387 	gfs2_set_alloc_start(&rbm, ip, dinode);
2388 	error = gfs2_rbm_find(&rbm, GFS2_BLKST_FREE, NULL, ip, false);
2389 
2390 	if (error == -ENOSPC) {
2391 		gfs2_set_alloc_start(&rbm, ip, dinode);
2392 		error = gfs2_rbm_find(&rbm, GFS2_BLKST_FREE, NULL, NULL, false);
2393 	}
2394 
2395 	/* Since all blocks are reserved in advance, this shouldn't happen */
2396 	if (error) {
2397 		fs_warn(sdp, "inum=%llu error=%d, nblocks=%u, full=%d fail_pt=%d\n",
2398 			(unsigned long long)ip->i_no_addr, error, *nblocks,
2399 			test_bit(GBF_FULL, &rbm.rgd->rd_bits->bi_flags),
2400 			rbm.rgd->rd_extfail_pt);
2401 		goto rgrp_error;
2402 	}
2403 
2404 	gfs2_alloc_extent(&rbm, dinode, nblocks);
2405 	block = gfs2_rbm_to_block(&rbm);
2406 	rbm.rgd->rd_last_alloc = block - rbm.rgd->rd_data0;
2407 	if (gfs2_rs_active(&ip->i_res))
2408 		gfs2_adjust_reservation(ip, &rbm, *nblocks);
2409 	ndata = *nblocks;
2410 	if (dinode)
2411 		ndata--;
2412 
2413 	if (!dinode) {
2414 		ip->i_goal = block + ndata - 1;
2415 		error = gfs2_meta_inode_buffer(ip, &dibh);
2416 		if (error == 0) {
2417 			struct gfs2_dinode *di =
2418 				(struct gfs2_dinode *)dibh->b_data;
2419 			gfs2_trans_add_meta(ip->i_gl, dibh);
2420 			di->di_goal_meta = di->di_goal_data =
2421 				cpu_to_be64(ip->i_goal);
2422 			brelse(dibh);
2423 		}
2424 	}
2425 	if (rbm.rgd->rd_free < *nblocks) {
2426 		fs_warn(sdp, "nblocks=%u\n", *nblocks);
2427 		goto rgrp_error;
2428 	}
2429 
2430 	rbm.rgd->rd_free -= *nblocks;
2431 	if (dinode) {
2432 		rbm.rgd->rd_dinodes++;
2433 		*generation = rbm.rgd->rd_igeneration++;
2434 		if (*generation == 0)
2435 			*generation = rbm.rgd->rd_igeneration++;
2436 	}
2437 
2438 	gfs2_trans_add_meta(rbm.rgd->rd_gl, rbm.rgd->rd_bits[0].bi_bh);
2439 	gfs2_rgrp_out(rbm.rgd, rbm.rgd->rd_bits[0].bi_bh->b_data);
2440 
2441 	gfs2_statfs_change(sdp, 0, -(s64)*nblocks, dinode ? 1 : 0);
2442 	if (dinode)
2443 		gfs2_trans_remove_revoke(sdp, block, *nblocks);
2444 
2445 	gfs2_quota_change(ip, *nblocks, ip->i_inode.i_uid, ip->i_inode.i_gid);
2446 
2447 	rbm.rgd->rd_free_clone -= *nblocks;
2448 	trace_gfs2_block_alloc(ip, rbm.rgd, block, *nblocks,
2449 			       dinode ? GFS2_BLKST_DINODE : GFS2_BLKST_USED);
2450 	*bn = block;
2451 	return 0;
2452 
2453 rgrp_error:
2454 	gfs2_rgrp_error(rbm.rgd);
2455 	return -EIO;
2456 }
2457 
2458 /**
2459  * __gfs2_free_blocks - free a contiguous run of block(s)
2460  * @ip: the inode these blocks are being freed from
2461  * @rgd: the resource group the blocks are in
2462  * @bstart: first block of a run of contiguous blocks
2463  * @blen: the length of the block run
2464  * @meta: 1 if the blocks represent metadata
2465  *
2466  */
2467 
2468 void __gfs2_free_blocks(struct gfs2_inode *ip, struct gfs2_rgrpd *rgd,
2469 			u64 bstart, u32 blen, int meta)
2470 {
2471 	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
2472 
2473 	rgblk_free(sdp, rgd, bstart, blen, GFS2_BLKST_FREE);
2474 	trace_gfs2_block_alloc(ip, rgd, bstart, blen, GFS2_BLKST_FREE);
2475 	rgd->rd_free += blen;
2476 	rgd->rd_flags &= ~GFS2_RGF_TRIMMED;
2477 	gfs2_trans_add_meta(rgd->rd_gl, rgd->rd_bits[0].bi_bh);
2478 	gfs2_rgrp_out(rgd, rgd->rd_bits[0].bi_bh->b_data);
2479 
2480 	/* Directories keep their data in the metadata address space */
2481 	if (meta || ip->i_depth)
2482 		gfs2_meta_wipe(ip, bstart, blen);
2483 }
2484 
2485 /**
2486  * gfs2_free_meta - free a contiguous run of data block(s)
2487  * @ip: the inode these blocks are being freed from
2488  * @rgd: the resource group the blocks are in
2489  * @bstart: first block of a run of contiguous blocks
2490  * @blen: the length of the block run
2491  *
2492  */
2493 
2494 void gfs2_free_meta(struct gfs2_inode *ip, struct gfs2_rgrpd *rgd,
2495 		    u64 bstart, u32 blen)
2496 {
2497 	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
2498 
2499 	__gfs2_free_blocks(ip, rgd, bstart, blen, 1);
2500 	gfs2_statfs_change(sdp, 0, +blen, 0);
2501 	gfs2_quota_change(ip, -(s64)blen, ip->i_inode.i_uid, ip->i_inode.i_gid);
2502 }
2503 
2504 void gfs2_unlink_di(struct inode *inode)
2505 {
2506 	struct gfs2_inode *ip = GFS2_I(inode);
2507 	struct gfs2_sbd *sdp = GFS2_SB(inode);
2508 	struct gfs2_rgrpd *rgd;
2509 	u64 blkno = ip->i_no_addr;
2510 
2511 	rgd = gfs2_blk2rgrpd(sdp, blkno, true);
2512 	if (!rgd)
2513 		return;
2514 	rgblk_free(sdp, rgd, blkno, 1, GFS2_BLKST_UNLINKED);
2515 	trace_gfs2_block_alloc(ip, rgd, blkno, 1, GFS2_BLKST_UNLINKED);
2516 	gfs2_trans_add_meta(rgd->rd_gl, rgd->rd_bits[0].bi_bh);
2517 	gfs2_rgrp_out(rgd, rgd->rd_bits[0].bi_bh->b_data);
2518 	be32_add_cpu(&rgd->rd_rgl->rl_unlinked, 1);
2519 }
2520 
2521 void gfs2_free_di(struct gfs2_rgrpd *rgd, struct gfs2_inode *ip)
2522 {
2523 	struct gfs2_sbd *sdp = rgd->rd_sbd;
2524 
2525 	rgblk_free(sdp, rgd, ip->i_no_addr, 1, GFS2_BLKST_FREE);
2526 	if (!rgd->rd_dinodes)
2527 		gfs2_consist_rgrpd(rgd);
2528 	rgd->rd_dinodes--;
2529 	rgd->rd_free++;
2530 
2531 	gfs2_trans_add_meta(rgd->rd_gl, rgd->rd_bits[0].bi_bh);
2532 	gfs2_rgrp_out(rgd, rgd->rd_bits[0].bi_bh->b_data);
2533 	be32_add_cpu(&rgd->rd_rgl->rl_unlinked, -1);
2534 
2535 	gfs2_statfs_change(sdp, 0, +1, -1);
2536 	trace_gfs2_block_alloc(ip, rgd, ip->i_no_addr, 1, GFS2_BLKST_FREE);
2537 	gfs2_quota_change(ip, -1, ip->i_inode.i_uid, ip->i_inode.i_gid);
2538 	gfs2_meta_wipe(ip, ip->i_no_addr, 1);
2539 }
2540 
2541 /**
2542  * gfs2_check_blk_type - Check the type of a block
2543  * @sdp: The superblock
2544  * @no_addr: The block number to check
2545  * @type: The block type we are looking for
2546  *
2547  * Returns: 0 if the block type matches the expected type
2548  *          -ESTALE if it doesn't match
2549  *          or -ve errno if something went wrong while checking
2550  */
2551 
2552 int gfs2_check_blk_type(struct gfs2_sbd *sdp, u64 no_addr, unsigned int type)
2553 {
2554 	struct gfs2_rgrpd *rgd;
2555 	struct gfs2_holder rgd_gh;
2556 	struct gfs2_rbm rbm;
2557 	int error = -EINVAL;
2558 
2559 	rgd = gfs2_blk2rgrpd(sdp, no_addr, 1);
2560 	if (!rgd)
2561 		goto fail;
2562 
2563 	error = gfs2_glock_nq_init(rgd->rd_gl, LM_ST_SHARED, 0, &rgd_gh);
2564 	if (error)
2565 		goto fail;
2566 
2567 	rbm.rgd = rgd;
2568 	error = gfs2_rbm_from_block(&rbm, no_addr);
2569 	if (WARN_ON_ONCE(error))
2570 		goto fail;
2571 
2572 	if (gfs2_testbit(&rbm, false) != type)
2573 		error = -ESTALE;
2574 
2575 	gfs2_glock_dq_uninit(&rgd_gh);
2576 fail:
2577 	return error;
2578 }
2579 
2580 /**
2581  * gfs2_rlist_add - add a RG to a list of RGs
2582  * @ip: the inode
2583  * @rlist: the list of resource groups
2584  * @block: the block
2585  *
2586  * Figure out what RG a block belongs to and add that RG to the list
2587  *
2588  * FIXME: Don't use NOFAIL
2589  *
2590  */
2591 
2592 void gfs2_rlist_add(struct gfs2_inode *ip, struct gfs2_rgrp_list *rlist,
2593 		    u64 block)
2594 {
2595 	struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
2596 	struct gfs2_rgrpd *rgd;
2597 	struct gfs2_rgrpd **tmp;
2598 	unsigned int new_space;
2599 	unsigned int x;
2600 
2601 	if (gfs2_assert_warn(sdp, !rlist->rl_ghs))
2602 		return;
2603 
2604 	/*
2605 	 * The resource group last accessed is kept in the last position.
2606 	 */
2607 
2608 	if (rlist->rl_rgrps) {
2609 		rgd = rlist->rl_rgd[rlist->rl_rgrps - 1];
2610 		if (rgrp_contains_block(rgd, block))
2611 			return;
2612 		rgd = gfs2_blk2rgrpd(sdp, block, 1);
2613 	} else {
2614 		rgd = ip->i_res.rs_rbm.rgd;
2615 		if (!rgd || !rgrp_contains_block(rgd, block))
2616 			rgd = gfs2_blk2rgrpd(sdp, block, 1);
2617 	}
2618 
2619 	if (!rgd) {
2620 		fs_err(sdp, "rlist_add: no rgrp for block %llu\n",
2621 		       (unsigned long long)block);
2622 		return;
2623 	}
2624 
2625 	for (x = 0; x < rlist->rl_rgrps; x++) {
2626 		if (rlist->rl_rgd[x] == rgd) {
2627 			swap(rlist->rl_rgd[x],
2628 			     rlist->rl_rgd[rlist->rl_rgrps - 1]);
2629 			return;
2630 		}
2631 	}
2632 
2633 	if (rlist->rl_rgrps == rlist->rl_space) {
2634 		new_space = rlist->rl_space + 10;
2635 
2636 		tmp = kcalloc(new_space, sizeof(struct gfs2_rgrpd *),
2637 			      GFP_NOFS | __GFP_NOFAIL);
2638 
2639 		if (rlist->rl_rgd) {
2640 			memcpy(tmp, rlist->rl_rgd,
2641 			       rlist->rl_space * sizeof(struct gfs2_rgrpd *));
2642 			kfree(rlist->rl_rgd);
2643 		}
2644 
2645 		rlist->rl_space = new_space;
2646 		rlist->rl_rgd = tmp;
2647 	}
2648 
2649 	rlist->rl_rgd[rlist->rl_rgrps++] = rgd;
2650 }
2651 
2652 /**
2653  * gfs2_rlist_alloc - all RGs have been added to the rlist, now allocate
2654  *      and initialize an array of glock holders for them
2655  * @rlist: the list of resource groups
2656  *
2657  * FIXME: Don't use NOFAIL
2658  *
2659  */
2660 
2661 void gfs2_rlist_alloc(struct gfs2_rgrp_list *rlist)
2662 {
2663 	unsigned int x;
2664 
2665 	rlist->rl_ghs = kmalloc_array(rlist->rl_rgrps,
2666 				      sizeof(struct gfs2_holder),
2667 				      GFP_NOFS | __GFP_NOFAIL);
2668 	for (x = 0; x < rlist->rl_rgrps; x++)
2669 		gfs2_holder_init(rlist->rl_rgd[x]->rd_gl,
2670 				LM_ST_EXCLUSIVE, 0,
2671 				&rlist->rl_ghs[x]);
2672 }
2673 
2674 /**
2675  * gfs2_rlist_free - free a resource group list
2676  * @rlist: the list of resource groups
2677  *
2678  */
2679 
2680 void gfs2_rlist_free(struct gfs2_rgrp_list *rlist)
2681 {
2682 	unsigned int x;
2683 
2684 	kfree(rlist->rl_rgd);
2685 
2686 	if (rlist->rl_ghs) {
2687 		for (x = 0; x < rlist->rl_rgrps; x++)
2688 			gfs2_holder_uninit(&rlist->rl_ghs[x]);
2689 		kfree(rlist->rl_ghs);
2690 		rlist->rl_ghs = NULL;
2691 	}
2692 }
2693 
2694