quota: cleanup loop in sync_dquots()
[deliverable/linux.git] / fs / gfs2 / rgrp.c
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 #include <linux/slab.h>
11 #include <linux/spinlock.h>
12 #include <linux/completion.h>
13 #include <linux/buffer_head.h>
14 #include <linux/fs.h>
15 #include <linux/gfs2_ondisk.h>
16 #include <linux/lm_interface.h>
17 #include <linux/prefetch.h>
18
19 #include "gfs2.h"
20 #include "incore.h"
21 #include "glock.h"
22 #include "glops.h"
23 #include "lops.h"
24 #include "meta_io.h"
25 #include "quota.h"
26 #include "rgrp.h"
27 #include "super.h"
28 #include "trans.h"
29 #include "util.h"
30 #include "log.h"
31 #include "inode.h"
32 #include "ops_address.h"
33
34 #define BFITNOENT ((u32)~0)
35 #define NO_BLOCK ((u64)~0)
36
37 #if BITS_PER_LONG == 32
38 #define LBITMASK (0x55555555UL)
39 #define LBITSKIP55 (0x55555555UL)
40 #define LBITSKIP00 (0x00000000UL)
41 #else
42 #define LBITMASK (0x5555555555555555UL)
43 #define LBITSKIP55 (0x5555555555555555UL)
44 #define LBITSKIP00 (0x0000000000000000UL)
45 #endif
46
47 /*
48 * These routines are used by the resource group routines (rgrp.c)
49 * to keep track of block allocation. Each block is represented by two
50 * bits. So, each byte represents GFS2_NBBY (i.e. 4) blocks.
51 *
52 * 0 = Free
53 * 1 = Used (not metadata)
54 * 2 = Unlinked (still in use) inode
55 * 3 = Used (metadata)
56 */
57
58 static const char valid_change[16] = {
59 /* current */
60 /* n */ 0, 1, 1, 1,
61 /* e */ 1, 0, 0, 0,
62 /* w */ 0, 0, 0, 1,
63 1, 0, 0, 0
64 };
65
66 static u32 rgblk_search(struct gfs2_rgrpd *rgd, u32 goal,
67 unsigned char old_state, unsigned char new_state,
68 unsigned int *n);
69
70 /**
71 * gfs2_setbit - Set a bit in the bitmaps
72 * @buffer: the buffer that holds the bitmaps
73 * @buflen: the length (in bytes) of the buffer
74 * @block: the block to set
75 * @new_state: the new state of the block
76 *
77 */
78
79 static inline void gfs2_setbit(struct gfs2_rgrpd *rgd, unsigned char *buf1,
80 unsigned char *buf2, unsigned int offset,
81 unsigned int buflen, u32 block,
82 unsigned char new_state)
83 {
84 unsigned char *byte1, *byte2, *end, cur_state;
85 const unsigned int bit = (block % GFS2_NBBY) * GFS2_BIT_SIZE;
86
87 byte1 = buf1 + offset + (block / GFS2_NBBY);
88 end = buf1 + offset + buflen;
89
90 BUG_ON(byte1 >= end);
91
92 cur_state = (*byte1 >> bit) & GFS2_BIT_MASK;
93
94 if (unlikely(!valid_change[new_state * 4 + cur_state])) {
95 gfs2_consist_rgrpd(rgd);
96 return;
97 }
98 *byte1 ^= (cur_state ^ new_state) << bit;
99
100 if (buf2) {
101 byte2 = buf2 + offset + (block / GFS2_NBBY);
102 cur_state = (*byte2 >> bit) & GFS2_BIT_MASK;
103 *byte2 ^= (cur_state ^ new_state) << bit;
104 }
105 }
106
107 /**
108 * gfs2_testbit - test a bit in the bitmaps
109 * @buffer: the buffer that holds the bitmaps
110 * @buflen: the length (in bytes) of the buffer
111 * @block: the block to read
112 *
113 */
114
115 static inline unsigned char gfs2_testbit(struct gfs2_rgrpd *rgd,
116 const unsigned char *buffer,
117 unsigned int buflen, u32 block)
118 {
119 const unsigned char *byte, *end;
120 unsigned char cur_state;
121 unsigned int bit;
122
123 byte = buffer + (block / GFS2_NBBY);
124 bit = (block % GFS2_NBBY) * GFS2_BIT_SIZE;
125 end = buffer + buflen;
126
127 gfs2_assert(rgd->rd_sbd, byte < end);
128
129 cur_state = (*byte >> bit) & GFS2_BIT_MASK;
130
131 return cur_state;
132 }
133
134 /**
135 * gfs2_bitfit - Search an rgrp's bitmap buffer to find a bit-pair representing
136 * a block in a given allocation state.
137 * @buffer: the buffer that holds the bitmaps
138 * @buflen: the length (in bytes) of the buffer
139 * @goal: start search at this block's bit-pair (within @buffer)
140 * @old_state: GFS2_BLKST_XXX the state of the block we're looking for.
141 *
142 * Scope of @goal and returned block number is only within this bitmap buffer,
143 * not entire rgrp or filesystem. @buffer will be offset from the actual
144 * beginning of a bitmap block buffer, skipping any header structures.
145 *
146 * Return: the block number (bitmap buffer scope) that was found
147 */
148
149 static u32 gfs2_bitfit(const u8 *buffer, unsigned int buflen, u32 goal,
150 u8 old_state)
151 {
152 const u8 *byte, *start, *end;
153 int bit, startbit;
154 u32 g1, g2, misaligned;
155 unsigned long *plong;
156 unsigned long lskipval;
157
158 lskipval = (old_state & GFS2_BLKST_USED) ? LBITSKIP00 : LBITSKIP55;
159 g1 = (goal / GFS2_NBBY);
160 start = buffer + g1;
161 byte = start;
162 end = buffer + buflen;
163 g2 = ALIGN(g1, sizeof(unsigned long));
164 plong = (unsigned long *)(buffer + g2);
165 startbit = bit = (goal % GFS2_NBBY) * GFS2_BIT_SIZE;
166 misaligned = g2 - g1;
167 if (!misaligned)
168 goto ulong_aligned;
169 /* parse the bitmap a byte at a time */
170 misaligned:
171 while (byte < end) {
172 if (((*byte >> bit) & GFS2_BIT_MASK) == old_state) {
173 return goal +
174 (((byte - start) * GFS2_NBBY) +
175 ((bit - startbit) >> 1));
176 }
177 bit += GFS2_BIT_SIZE;
178 if (bit >= GFS2_NBBY * GFS2_BIT_SIZE) {
179 bit = 0;
180 byte++;
181 misaligned--;
182 if (!misaligned) {
183 plong = (unsigned long *)byte;
184 goto ulong_aligned;
185 }
186 }
187 }
188 return BFITNOENT;
189
190 /* parse the bitmap a unsigned long at a time */
191 ulong_aligned:
192 /* Stop at "end - 1" or else prefetch can go past the end and segfault.
193 We could "if" it but we'd lose some of the performance gained.
194 This way will only slow down searching the very last 4/8 bytes
195 depending on architecture. I've experimented with several ways
196 of writing this section such as using an else before the goto
197 but this one seems to be the fastest. */
198 while ((unsigned char *)plong < end - sizeof(unsigned long)) {
199 prefetch(plong + 1);
200 if (((*plong) & LBITMASK) != lskipval)
201 break;
202 plong++;
203 }
204 if ((unsigned char *)plong < end) {
205 byte = (const u8 *)plong;
206 misaligned += sizeof(unsigned long) - 1;
207 goto misaligned;
208 }
209 return BFITNOENT;
210 }
211
212 /**
213 * gfs2_bitcount - count the number of bits in a certain state
214 * @buffer: the buffer that holds the bitmaps
215 * @buflen: the length (in bytes) of the buffer
216 * @state: the state of the block we're looking for
217 *
218 * Returns: The number of bits
219 */
220
221 static u32 gfs2_bitcount(struct gfs2_rgrpd *rgd, const u8 *buffer,
222 unsigned int buflen, u8 state)
223 {
224 const u8 *byte = buffer;
225 const u8 *end = buffer + buflen;
226 const u8 state1 = state << 2;
227 const u8 state2 = state << 4;
228 const u8 state3 = state << 6;
229 u32 count = 0;
230
231 for (; byte < end; byte++) {
232 if (((*byte) & 0x03) == state)
233 count++;
234 if (((*byte) & 0x0C) == state1)
235 count++;
236 if (((*byte) & 0x30) == state2)
237 count++;
238 if (((*byte) & 0xC0) == state3)
239 count++;
240 }
241
242 return count;
243 }
244
245 /**
246 * gfs2_rgrp_verify - Verify that a resource group is consistent
247 * @sdp: the filesystem
248 * @rgd: the rgrp
249 *
250 */
251
252 void gfs2_rgrp_verify(struct gfs2_rgrpd *rgd)
253 {
254 struct gfs2_sbd *sdp = rgd->rd_sbd;
255 struct gfs2_bitmap *bi = NULL;
256 u32 length = rgd->rd_length;
257 u32 count[4], tmp;
258 int buf, x;
259
260 memset(count, 0, 4 * sizeof(u32));
261
262 /* Count # blocks in each of 4 possible allocation states */
263 for (buf = 0; buf < length; buf++) {
264 bi = rgd->rd_bits + buf;
265 for (x = 0; x < 4; x++)
266 count[x] += gfs2_bitcount(rgd,
267 bi->bi_bh->b_data +
268 bi->bi_offset,
269 bi->bi_len, x);
270 }
271
272 if (count[0] != rgd->rd_rg.rg_free) {
273 if (gfs2_consist_rgrpd(rgd))
274 fs_err(sdp, "free data mismatch: %u != %u\n",
275 count[0], rgd->rd_rg.rg_free);
276 return;
277 }
278
279 tmp = rgd->rd_data -
280 rgd->rd_rg.rg_free -
281 rgd->rd_rg.rg_dinodes;
282 if (count[1] + count[2] != tmp) {
283 if (gfs2_consist_rgrpd(rgd))
284 fs_err(sdp, "used data mismatch: %u != %u\n",
285 count[1], tmp);
286 return;
287 }
288
289 if (count[3] != rgd->rd_rg.rg_dinodes) {
290 if (gfs2_consist_rgrpd(rgd))
291 fs_err(sdp, "used metadata mismatch: %u != %u\n",
292 count[3], rgd->rd_rg.rg_dinodes);
293 return;
294 }
295
296 if (count[2] > count[3]) {
297 if (gfs2_consist_rgrpd(rgd))
298 fs_err(sdp, "unlinked inodes > inodes: %u\n",
299 count[2]);
300 return;
301 }
302
303 }
304
305 static inline int rgrp_contains_block(struct gfs2_rgrpd *rgd, u64 block)
306 {
307 u64 first = rgd->rd_data0;
308 u64 last = first + rgd->rd_data;
309 return first <= block && block < last;
310 }
311
312 /**
313 * gfs2_blk2rgrpd - Find resource group for a given data/meta block number
314 * @sdp: The GFS2 superblock
315 * @n: The data block number
316 *
317 * Returns: The resource group, or NULL if not found
318 */
319
320 struct gfs2_rgrpd *gfs2_blk2rgrpd(struct gfs2_sbd *sdp, u64 blk)
321 {
322 struct gfs2_rgrpd *rgd;
323
324 spin_lock(&sdp->sd_rindex_spin);
325
326 list_for_each_entry(rgd, &sdp->sd_rindex_mru_list, rd_list_mru) {
327 if (rgrp_contains_block(rgd, blk)) {
328 list_move(&rgd->rd_list_mru, &sdp->sd_rindex_mru_list);
329 spin_unlock(&sdp->sd_rindex_spin);
330 return rgd;
331 }
332 }
333
334 spin_unlock(&sdp->sd_rindex_spin);
335
336 return NULL;
337 }
338
339 /**
340 * gfs2_rgrpd_get_first - get the first Resource Group in the filesystem
341 * @sdp: The GFS2 superblock
342 *
343 * Returns: The first rgrp in the filesystem
344 */
345
346 struct gfs2_rgrpd *gfs2_rgrpd_get_first(struct gfs2_sbd *sdp)
347 {
348 gfs2_assert(sdp, !list_empty(&sdp->sd_rindex_list));
349 return list_entry(sdp->sd_rindex_list.next, struct gfs2_rgrpd, rd_list);
350 }
351
352 /**
353 * gfs2_rgrpd_get_next - get the next RG
354 * @rgd: A RG
355 *
356 * Returns: The next rgrp
357 */
358
359 struct gfs2_rgrpd *gfs2_rgrpd_get_next(struct gfs2_rgrpd *rgd)
360 {
361 if (rgd->rd_list.next == &rgd->rd_sbd->sd_rindex_list)
362 return NULL;
363 return list_entry(rgd->rd_list.next, struct gfs2_rgrpd, rd_list);
364 }
365
366 static void clear_rgrpdi(struct gfs2_sbd *sdp)
367 {
368 struct list_head *head;
369 struct gfs2_rgrpd *rgd;
370 struct gfs2_glock *gl;
371
372 spin_lock(&sdp->sd_rindex_spin);
373 sdp->sd_rindex_forward = NULL;
374 spin_unlock(&sdp->sd_rindex_spin);
375
376 head = &sdp->sd_rindex_list;
377 while (!list_empty(head)) {
378 rgd = list_entry(head->next, struct gfs2_rgrpd, rd_list);
379 gl = rgd->rd_gl;
380
381 list_del(&rgd->rd_list);
382 list_del(&rgd->rd_list_mru);
383
384 if (gl) {
385 gl->gl_object = NULL;
386 gfs2_glock_put(gl);
387 }
388
389 kfree(rgd->rd_bits);
390 kmem_cache_free(gfs2_rgrpd_cachep, rgd);
391 }
392 }
393
394 void gfs2_clear_rgrpd(struct gfs2_sbd *sdp)
395 {
396 mutex_lock(&sdp->sd_rindex_mutex);
397 clear_rgrpdi(sdp);
398 mutex_unlock(&sdp->sd_rindex_mutex);
399 }
400
401 static void gfs2_rindex_print(const struct gfs2_rgrpd *rgd)
402 {
403 printk(KERN_INFO " ri_addr = %llu\n", (unsigned long long)rgd->rd_addr);
404 printk(KERN_INFO " ri_length = %u\n", rgd->rd_length);
405 printk(KERN_INFO " ri_data0 = %llu\n", (unsigned long long)rgd->rd_data0);
406 printk(KERN_INFO " ri_data = %u\n", rgd->rd_data);
407 printk(KERN_INFO " ri_bitbytes = %u\n", rgd->rd_bitbytes);
408 }
409
410 /**
411 * gfs2_compute_bitstructs - Compute the bitmap sizes
412 * @rgd: The resource group descriptor
413 *
414 * Calculates bitmap descriptors, one for each block that contains bitmap data
415 *
416 * Returns: errno
417 */
418
419 static int compute_bitstructs(struct gfs2_rgrpd *rgd)
420 {
421 struct gfs2_sbd *sdp = rgd->rd_sbd;
422 struct gfs2_bitmap *bi;
423 u32 length = rgd->rd_length; /* # blocks in hdr & bitmap */
424 u32 bytes_left, bytes;
425 int x;
426
427 if (!length)
428 return -EINVAL;
429
430 rgd->rd_bits = kcalloc(length, sizeof(struct gfs2_bitmap), GFP_NOFS);
431 if (!rgd->rd_bits)
432 return -ENOMEM;
433
434 bytes_left = rgd->rd_bitbytes;
435
436 for (x = 0; x < length; x++) {
437 bi = rgd->rd_bits + x;
438
439 /* small rgrp; bitmap stored completely in header block */
440 if (length == 1) {
441 bytes = bytes_left;
442 bi->bi_offset = sizeof(struct gfs2_rgrp);
443 bi->bi_start = 0;
444 bi->bi_len = bytes;
445 /* header block */
446 } else if (x == 0) {
447 bytes = sdp->sd_sb.sb_bsize - sizeof(struct gfs2_rgrp);
448 bi->bi_offset = sizeof(struct gfs2_rgrp);
449 bi->bi_start = 0;
450 bi->bi_len = bytes;
451 /* last block */
452 } else if (x + 1 == length) {
453 bytes = bytes_left;
454 bi->bi_offset = sizeof(struct gfs2_meta_header);
455 bi->bi_start = rgd->rd_bitbytes - bytes_left;
456 bi->bi_len = bytes;
457 /* other blocks */
458 } else {
459 bytes = sdp->sd_sb.sb_bsize -
460 sizeof(struct gfs2_meta_header);
461 bi->bi_offset = sizeof(struct gfs2_meta_header);
462 bi->bi_start = rgd->rd_bitbytes - bytes_left;
463 bi->bi_len = bytes;
464 }
465
466 bytes_left -= bytes;
467 }
468
469 if (bytes_left) {
470 gfs2_consist_rgrpd(rgd);
471 return -EIO;
472 }
473 bi = rgd->rd_bits + (length - 1);
474 if ((bi->bi_start + bi->bi_len) * GFS2_NBBY != rgd->rd_data) {
475 if (gfs2_consist_rgrpd(rgd)) {
476 gfs2_rindex_print(rgd);
477 fs_err(sdp, "start=%u len=%u offset=%u\n",
478 bi->bi_start, bi->bi_len, bi->bi_offset);
479 }
480 return -EIO;
481 }
482
483 return 0;
484 }
485
486 /**
487 * gfs2_ri_total - Total up the file system space, according to the rindex.
488 *
489 */
490 u64 gfs2_ri_total(struct gfs2_sbd *sdp)
491 {
492 u64 total_data = 0;
493 struct inode *inode = sdp->sd_rindex;
494 struct gfs2_inode *ip = GFS2_I(inode);
495 char buf[sizeof(struct gfs2_rindex)];
496 struct file_ra_state ra_state;
497 int error, rgrps;
498
499 mutex_lock(&sdp->sd_rindex_mutex);
500 file_ra_state_init(&ra_state, inode->i_mapping);
501 for (rgrps = 0;; rgrps++) {
502 loff_t pos = rgrps * sizeof(struct gfs2_rindex);
503
504 if (pos + sizeof(struct gfs2_rindex) >= ip->i_di.di_size)
505 break;
506 error = gfs2_internal_read(ip, &ra_state, buf, &pos,
507 sizeof(struct gfs2_rindex));
508 if (error != sizeof(struct gfs2_rindex))
509 break;
510 total_data += be32_to_cpu(((struct gfs2_rindex *)buf)->ri_data);
511 }
512 mutex_unlock(&sdp->sd_rindex_mutex);
513 return total_data;
514 }
515
516 static void gfs2_rindex_in(struct gfs2_rgrpd *rgd, const void *buf)
517 {
518 const struct gfs2_rindex *str = buf;
519
520 rgd->rd_addr = be64_to_cpu(str->ri_addr);
521 rgd->rd_length = be32_to_cpu(str->ri_length);
522 rgd->rd_data0 = be64_to_cpu(str->ri_data0);
523 rgd->rd_data = be32_to_cpu(str->ri_data);
524 rgd->rd_bitbytes = be32_to_cpu(str->ri_bitbytes);
525 }
526
527 /**
528 * read_rindex_entry - Pull in a new resource index entry from the disk
529 * @gl: The glock covering the rindex inode
530 *
531 * Returns: 0 on success, error code otherwise
532 */
533
534 static int read_rindex_entry(struct gfs2_inode *ip,
535 struct file_ra_state *ra_state)
536 {
537 struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
538 loff_t pos = sdp->sd_rgrps * sizeof(struct gfs2_rindex);
539 char buf[sizeof(struct gfs2_rindex)];
540 int error;
541 struct gfs2_rgrpd *rgd;
542
543 error = gfs2_internal_read(ip, ra_state, buf, &pos,
544 sizeof(struct gfs2_rindex));
545 if (!error)
546 return 0;
547 if (error != sizeof(struct gfs2_rindex)) {
548 if (error > 0)
549 error = -EIO;
550 return error;
551 }
552
553 rgd = kmem_cache_zalloc(gfs2_rgrpd_cachep, GFP_NOFS);
554 error = -ENOMEM;
555 if (!rgd)
556 return error;
557
558 mutex_init(&rgd->rd_mutex);
559 lops_init_le(&rgd->rd_le, &gfs2_rg_lops);
560 rgd->rd_sbd = sdp;
561
562 list_add_tail(&rgd->rd_list, &sdp->sd_rindex_list);
563 list_add_tail(&rgd->rd_list_mru, &sdp->sd_rindex_mru_list);
564
565 gfs2_rindex_in(rgd, buf);
566 error = compute_bitstructs(rgd);
567 if (error)
568 return error;
569
570 error = gfs2_glock_get(sdp, rgd->rd_addr,
571 &gfs2_rgrp_glops, CREATE, &rgd->rd_gl);
572 if (error)
573 return error;
574
575 rgd->rd_gl->gl_object = rgd;
576 rgd->rd_flags &= ~GFS2_RDF_UPTODATE;
577 rgd->rd_flags |= GFS2_RDF_CHECK;
578 return error;
579 }
580
581 /**
582 * gfs2_ri_update - Pull in a new resource index from the disk
583 * @ip: pointer to the rindex inode
584 *
585 * Returns: 0 on successful update, error code otherwise
586 */
587
588 static int gfs2_ri_update(struct gfs2_inode *ip)
589 {
590 struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
591 struct inode *inode = &ip->i_inode;
592 struct file_ra_state ra_state;
593 u64 rgrp_count = ip->i_di.di_size;
594 int error;
595
596 if (do_div(rgrp_count, sizeof(struct gfs2_rindex))) {
597 gfs2_consist_inode(ip);
598 return -EIO;
599 }
600
601 clear_rgrpdi(sdp);
602
603 file_ra_state_init(&ra_state, inode->i_mapping);
604 for (sdp->sd_rgrps = 0; sdp->sd_rgrps < rgrp_count; sdp->sd_rgrps++) {
605 error = read_rindex_entry(ip, &ra_state);
606 if (error) {
607 clear_rgrpdi(sdp);
608 return error;
609 }
610 }
611
612 sdp->sd_rindex_uptodate = 1;
613 return 0;
614 }
615
616 /**
617 * gfs2_ri_update_special - Pull in a new resource index from the disk
618 *
619 * This is a special version that's safe to call from gfs2_inplace_reserve_i.
620 * In this case we know that we don't have any resource groups in memory yet.
621 *
622 * @ip: pointer to the rindex inode
623 *
624 * Returns: 0 on successful update, error code otherwise
625 */
626 static int gfs2_ri_update_special(struct gfs2_inode *ip)
627 {
628 struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
629 struct inode *inode = &ip->i_inode;
630 struct file_ra_state ra_state;
631 int error;
632
633 file_ra_state_init(&ra_state, inode->i_mapping);
634 for (sdp->sd_rgrps = 0;; sdp->sd_rgrps++) {
635 /* Ignore partials */
636 if ((sdp->sd_rgrps + 1) * sizeof(struct gfs2_rindex) >
637 ip->i_di.di_size)
638 break;
639 error = read_rindex_entry(ip, &ra_state);
640 if (error) {
641 clear_rgrpdi(sdp);
642 return error;
643 }
644 }
645
646 sdp->sd_rindex_uptodate = 1;
647 return 0;
648 }
649
650 /**
651 * gfs2_rindex_hold - Grab a lock on the rindex
652 * @sdp: The GFS2 superblock
653 * @ri_gh: the glock holder
654 *
655 * We grab a lock on the rindex inode to make sure that it doesn't
656 * change whilst we are performing an operation. We keep this lock
657 * for quite long periods of time compared to other locks. This
658 * doesn't matter, since it is shared and it is very, very rarely
659 * accessed in the exclusive mode (i.e. only when expanding the filesystem).
660 *
661 * This makes sure that we're using the latest copy of the resource index
662 * special file, which might have been updated if someone expanded the
663 * filesystem (via gfs2_grow utility), which adds new resource groups.
664 *
665 * Returns: 0 on success, error code otherwise
666 */
667
668 int gfs2_rindex_hold(struct gfs2_sbd *sdp, struct gfs2_holder *ri_gh)
669 {
670 struct gfs2_inode *ip = GFS2_I(sdp->sd_rindex);
671 struct gfs2_glock *gl = ip->i_gl;
672 int error;
673
674 error = gfs2_glock_nq_init(gl, LM_ST_SHARED, 0, ri_gh);
675 if (error)
676 return error;
677
678 /* Read new copy from disk if we don't have the latest */
679 if (!sdp->sd_rindex_uptodate) {
680 mutex_lock(&sdp->sd_rindex_mutex);
681 if (!sdp->sd_rindex_uptodate) {
682 error = gfs2_ri_update(ip);
683 if (error)
684 gfs2_glock_dq_uninit(ri_gh);
685 }
686 mutex_unlock(&sdp->sd_rindex_mutex);
687 }
688
689 return error;
690 }
691
692 static void gfs2_rgrp_in(struct gfs2_rgrpd *rgd, const void *buf)
693 {
694 const struct gfs2_rgrp *str = buf;
695 struct gfs2_rgrp_host *rg = &rgd->rd_rg;
696 u32 rg_flags;
697
698 rg_flags = be32_to_cpu(str->rg_flags);
699 if (rg_flags & GFS2_RGF_NOALLOC)
700 rgd->rd_flags |= GFS2_RDF_NOALLOC;
701 else
702 rgd->rd_flags &= ~GFS2_RDF_NOALLOC;
703 rg->rg_free = be32_to_cpu(str->rg_free);
704 rg->rg_dinodes = be32_to_cpu(str->rg_dinodes);
705 rg->rg_igeneration = be64_to_cpu(str->rg_igeneration);
706 }
707
708 static void gfs2_rgrp_out(struct gfs2_rgrpd *rgd, void *buf)
709 {
710 struct gfs2_rgrp *str = buf;
711 struct gfs2_rgrp_host *rg = &rgd->rd_rg;
712 u32 rg_flags = 0;
713
714 if (rgd->rd_flags & GFS2_RDF_NOALLOC)
715 rg_flags |= GFS2_RGF_NOALLOC;
716 str->rg_flags = cpu_to_be32(rg_flags);
717 str->rg_free = cpu_to_be32(rg->rg_free);
718 str->rg_dinodes = cpu_to_be32(rg->rg_dinodes);
719 str->__pad = cpu_to_be32(0);
720 str->rg_igeneration = cpu_to_be64(rg->rg_igeneration);
721 memset(&str->rg_reserved, 0, sizeof(str->rg_reserved));
722 }
723
724 /**
725 * gfs2_rgrp_bh_get - Read in a RG's header and bitmaps
726 * @rgd: the struct gfs2_rgrpd describing the RG to read in
727 *
728 * Read in all of a Resource Group's header and bitmap blocks.
729 * Caller must eventually call gfs2_rgrp_relse() to free the bitmaps.
730 *
731 * Returns: errno
732 */
733
734 int gfs2_rgrp_bh_get(struct gfs2_rgrpd *rgd)
735 {
736 struct gfs2_sbd *sdp = rgd->rd_sbd;
737 struct gfs2_glock *gl = rgd->rd_gl;
738 unsigned int length = rgd->rd_length;
739 struct gfs2_bitmap *bi;
740 unsigned int x, y;
741 int error;
742
743 mutex_lock(&rgd->rd_mutex);
744
745 spin_lock(&sdp->sd_rindex_spin);
746 if (rgd->rd_bh_count) {
747 rgd->rd_bh_count++;
748 spin_unlock(&sdp->sd_rindex_spin);
749 mutex_unlock(&rgd->rd_mutex);
750 return 0;
751 }
752 spin_unlock(&sdp->sd_rindex_spin);
753
754 for (x = 0; x < length; x++) {
755 bi = rgd->rd_bits + x;
756 error = gfs2_meta_read(gl, rgd->rd_addr + x, 0, &bi->bi_bh);
757 if (error)
758 goto fail;
759 }
760
761 for (y = length; y--;) {
762 bi = rgd->rd_bits + y;
763 error = gfs2_meta_wait(sdp, bi->bi_bh);
764 if (error)
765 goto fail;
766 if (gfs2_metatype_check(sdp, bi->bi_bh, y ? GFS2_METATYPE_RB :
767 GFS2_METATYPE_RG)) {
768 error = -EIO;
769 goto fail;
770 }
771 }
772
773 if (!(rgd->rd_flags & GFS2_RDF_UPTODATE)) {
774 gfs2_rgrp_in(rgd, (rgd->rd_bits[0].bi_bh)->b_data);
775 rgd->rd_flags |= GFS2_RDF_UPTODATE;
776 }
777
778 spin_lock(&sdp->sd_rindex_spin);
779 rgd->rd_free_clone = rgd->rd_rg.rg_free;
780 rgd->rd_bh_count++;
781 spin_unlock(&sdp->sd_rindex_spin);
782
783 mutex_unlock(&rgd->rd_mutex);
784
785 return 0;
786
787 fail:
788 while (x--) {
789 bi = rgd->rd_bits + x;
790 brelse(bi->bi_bh);
791 bi->bi_bh = NULL;
792 gfs2_assert_warn(sdp, !bi->bi_clone);
793 }
794 mutex_unlock(&rgd->rd_mutex);
795
796 return error;
797 }
798
799 void gfs2_rgrp_bh_hold(struct gfs2_rgrpd *rgd)
800 {
801 struct gfs2_sbd *sdp = rgd->rd_sbd;
802
803 spin_lock(&sdp->sd_rindex_spin);
804 gfs2_assert_warn(rgd->rd_sbd, rgd->rd_bh_count);
805 rgd->rd_bh_count++;
806 spin_unlock(&sdp->sd_rindex_spin);
807 }
808
809 /**
810 * gfs2_rgrp_bh_put - Release RG bitmaps read in with gfs2_rgrp_bh_get()
811 * @rgd: the struct gfs2_rgrpd describing the RG to read in
812 *
813 */
814
815 void gfs2_rgrp_bh_put(struct gfs2_rgrpd *rgd)
816 {
817 struct gfs2_sbd *sdp = rgd->rd_sbd;
818 int x, length = rgd->rd_length;
819
820 spin_lock(&sdp->sd_rindex_spin);
821 gfs2_assert_warn(rgd->rd_sbd, rgd->rd_bh_count);
822 if (--rgd->rd_bh_count) {
823 spin_unlock(&sdp->sd_rindex_spin);
824 return;
825 }
826
827 for (x = 0; x < length; x++) {
828 struct gfs2_bitmap *bi = rgd->rd_bits + x;
829 kfree(bi->bi_clone);
830 bi->bi_clone = NULL;
831 brelse(bi->bi_bh);
832 bi->bi_bh = NULL;
833 }
834
835 spin_unlock(&sdp->sd_rindex_spin);
836 }
837
838 void gfs2_rgrp_repolish_clones(struct gfs2_rgrpd *rgd)
839 {
840 struct gfs2_sbd *sdp = rgd->rd_sbd;
841 unsigned int length = rgd->rd_length;
842 unsigned int x;
843
844 for (x = 0; x < length; x++) {
845 struct gfs2_bitmap *bi = rgd->rd_bits + x;
846 if (!bi->bi_clone)
847 continue;
848 memcpy(bi->bi_clone + bi->bi_offset,
849 bi->bi_bh->b_data + bi->bi_offset, bi->bi_len);
850 }
851
852 spin_lock(&sdp->sd_rindex_spin);
853 rgd->rd_free_clone = rgd->rd_rg.rg_free;
854 spin_unlock(&sdp->sd_rindex_spin);
855 }
856
857 /**
858 * gfs2_alloc_get - get the struct gfs2_alloc structure for an inode
859 * @ip: the incore GFS2 inode structure
860 *
861 * Returns: the struct gfs2_alloc
862 */
863
864 struct gfs2_alloc *gfs2_alloc_get(struct gfs2_inode *ip)
865 {
866 BUG_ON(ip->i_alloc != NULL);
867 ip->i_alloc = kzalloc(sizeof(struct gfs2_alloc), GFP_KERNEL);
868 return ip->i_alloc;
869 }
870
871 /**
872 * try_rgrp_fit - See if a given reservation will fit in a given RG
873 * @rgd: the RG data
874 * @al: the struct gfs2_alloc structure describing the reservation
875 *
876 * If there's room for the requested blocks to be allocated from the RG:
877 * Sets the $al_rgd field in @al.
878 *
879 * Returns: 1 on success (it fits), 0 on failure (it doesn't fit)
880 */
881
882 static int try_rgrp_fit(struct gfs2_rgrpd *rgd, struct gfs2_alloc *al)
883 {
884 struct gfs2_sbd *sdp = rgd->rd_sbd;
885 int ret = 0;
886
887 if (rgd->rd_flags & GFS2_RDF_NOALLOC)
888 return 0;
889
890 spin_lock(&sdp->sd_rindex_spin);
891 if (rgd->rd_free_clone >= al->al_requested) {
892 al->al_rgd = rgd;
893 ret = 1;
894 }
895 spin_unlock(&sdp->sd_rindex_spin);
896
897 return ret;
898 }
899
900 /**
901 * try_rgrp_unlink - Look for any unlinked, allocated, but unused inodes
902 * @rgd: The rgrp
903 *
904 * Returns: The inode, if one has been found
905 */
906
907 static struct inode *try_rgrp_unlink(struct gfs2_rgrpd *rgd, u64 *last_unlinked)
908 {
909 struct inode *inode;
910 u32 goal = 0, block;
911 u64 no_addr;
912 struct gfs2_sbd *sdp = rgd->rd_sbd;
913 unsigned int n;
914
915 for(;;) {
916 if (goal >= rgd->rd_data)
917 break;
918 down_write(&sdp->sd_log_flush_lock);
919 n = 1;
920 block = rgblk_search(rgd, goal, GFS2_BLKST_UNLINKED,
921 GFS2_BLKST_UNLINKED, &n);
922 up_write(&sdp->sd_log_flush_lock);
923 if (block == BFITNOENT)
924 break;
925 /* rgblk_search can return a block < goal, so we need to
926 keep it marching forward. */
927 no_addr = block + rgd->rd_data0;
928 goal++;
929 if (*last_unlinked != NO_BLOCK && no_addr <= *last_unlinked)
930 continue;
931 *last_unlinked = no_addr;
932 inode = gfs2_inode_lookup(rgd->rd_sbd->sd_vfs, DT_UNKNOWN,
933 no_addr, -1, 1);
934 if (!IS_ERR(inode))
935 return inode;
936 }
937
938 rgd->rd_flags &= ~GFS2_RDF_CHECK;
939 return NULL;
940 }
941
942 /**
943 * recent_rgrp_next - get next RG from "recent" list
944 * @cur_rgd: current rgrp
945 *
946 * Returns: The next rgrp in the recent list
947 */
948
949 static struct gfs2_rgrpd *recent_rgrp_next(struct gfs2_rgrpd *cur_rgd)
950 {
951 struct gfs2_sbd *sdp = cur_rgd->rd_sbd;
952 struct list_head *head;
953 struct gfs2_rgrpd *rgd;
954
955 spin_lock(&sdp->sd_rindex_spin);
956 head = &sdp->sd_rindex_mru_list;
957 if (unlikely(cur_rgd->rd_list_mru.next == head)) {
958 spin_unlock(&sdp->sd_rindex_spin);
959 return NULL;
960 }
961 rgd = list_entry(cur_rgd->rd_list_mru.next, struct gfs2_rgrpd, rd_list_mru);
962 spin_unlock(&sdp->sd_rindex_spin);
963 return rgd;
964 }
965
966 /**
967 * forward_rgrp_get - get an rgrp to try next from full list
968 * @sdp: The GFS2 superblock
969 *
970 * Returns: The rgrp to try next
971 */
972
973 static struct gfs2_rgrpd *forward_rgrp_get(struct gfs2_sbd *sdp)
974 {
975 struct gfs2_rgrpd *rgd;
976 unsigned int journals = gfs2_jindex_size(sdp);
977 unsigned int rg = 0, x;
978
979 spin_lock(&sdp->sd_rindex_spin);
980
981 rgd = sdp->sd_rindex_forward;
982 if (!rgd) {
983 if (sdp->sd_rgrps >= journals)
984 rg = sdp->sd_rgrps * sdp->sd_jdesc->jd_jid / journals;
985
986 for (x = 0, rgd = gfs2_rgrpd_get_first(sdp); x < rg;
987 x++, rgd = gfs2_rgrpd_get_next(rgd))
988 /* Do Nothing */;
989
990 sdp->sd_rindex_forward = rgd;
991 }
992
993 spin_unlock(&sdp->sd_rindex_spin);
994
995 return rgd;
996 }
997
998 /**
999 * forward_rgrp_set - set the forward rgrp pointer
1000 * @sdp: the filesystem
1001 * @rgd: The new forward rgrp
1002 *
1003 */
1004
1005 static void forward_rgrp_set(struct gfs2_sbd *sdp, struct gfs2_rgrpd *rgd)
1006 {
1007 spin_lock(&sdp->sd_rindex_spin);
1008 sdp->sd_rindex_forward = rgd;
1009 spin_unlock(&sdp->sd_rindex_spin);
1010 }
1011
1012 /**
1013 * get_local_rgrp - Choose and lock a rgrp for allocation
1014 * @ip: the inode to reserve space for
1015 * @rgp: the chosen and locked rgrp
1016 *
1017 * Try to acquire rgrp in way which avoids contending with others.
1018 *
1019 * Returns: errno
1020 */
1021
1022 static struct inode *get_local_rgrp(struct gfs2_inode *ip, u64 *last_unlinked)
1023 {
1024 struct inode *inode = NULL;
1025 struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
1026 struct gfs2_rgrpd *rgd, *begin = NULL;
1027 struct gfs2_alloc *al = ip->i_alloc;
1028 int flags = LM_FLAG_TRY;
1029 int skipped = 0;
1030 int loops = 0;
1031 int error, rg_locked;
1032
1033 rgd = gfs2_blk2rgrpd(sdp, ip->i_goal);
1034
1035 while (rgd) {
1036 rg_locked = 0;
1037
1038 if (gfs2_glock_is_locked_by_me(rgd->rd_gl)) {
1039 rg_locked = 1;
1040 error = 0;
1041 } else {
1042 error = gfs2_glock_nq_init(rgd->rd_gl, LM_ST_EXCLUSIVE,
1043 LM_FLAG_TRY, &al->al_rgd_gh);
1044 }
1045 switch (error) {
1046 case 0:
1047 if (try_rgrp_fit(rgd, al))
1048 goto out;
1049 if (rgd->rd_flags & GFS2_RDF_CHECK)
1050 inode = try_rgrp_unlink(rgd, last_unlinked);
1051 if (!rg_locked)
1052 gfs2_glock_dq_uninit(&al->al_rgd_gh);
1053 if (inode)
1054 return inode;
1055 /* fall through */
1056 case GLR_TRYFAILED:
1057 rgd = recent_rgrp_next(rgd);
1058 break;
1059
1060 default:
1061 return ERR_PTR(error);
1062 }
1063 }
1064
1065 /* Go through full list of rgrps */
1066
1067 begin = rgd = forward_rgrp_get(sdp);
1068
1069 for (;;) {
1070 rg_locked = 0;
1071
1072 if (gfs2_glock_is_locked_by_me(rgd->rd_gl)) {
1073 rg_locked = 1;
1074 error = 0;
1075 } else {
1076 error = gfs2_glock_nq_init(rgd->rd_gl, LM_ST_EXCLUSIVE, flags,
1077 &al->al_rgd_gh);
1078 }
1079 switch (error) {
1080 case 0:
1081 if (try_rgrp_fit(rgd, al))
1082 goto out;
1083 if (rgd->rd_flags & GFS2_RDF_CHECK)
1084 inode = try_rgrp_unlink(rgd, last_unlinked);
1085 if (!rg_locked)
1086 gfs2_glock_dq_uninit(&al->al_rgd_gh);
1087 if (inode)
1088 return inode;
1089 break;
1090
1091 case GLR_TRYFAILED:
1092 skipped++;
1093 break;
1094
1095 default:
1096 return ERR_PTR(error);
1097 }
1098
1099 rgd = gfs2_rgrpd_get_next(rgd);
1100 if (!rgd)
1101 rgd = gfs2_rgrpd_get_first(sdp);
1102
1103 if (rgd == begin) {
1104 if (++loops >= 3)
1105 return ERR_PTR(-ENOSPC);
1106 if (!skipped)
1107 loops++;
1108 flags = 0;
1109 if (loops == 2)
1110 gfs2_log_flush(sdp, NULL);
1111 }
1112 }
1113
1114 out:
1115 if (begin) {
1116 spin_lock(&sdp->sd_rindex_spin);
1117 list_move(&rgd->rd_list_mru, &sdp->sd_rindex_mru_list);
1118 spin_unlock(&sdp->sd_rindex_spin);
1119 rgd = gfs2_rgrpd_get_next(rgd);
1120 if (!rgd)
1121 rgd = gfs2_rgrpd_get_first(sdp);
1122 forward_rgrp_set(sdp, rgd);
1123 }
1124
1125 return NULL;
1126 }
1127
1128 /**
1129 * gfs2_inplace_reserve_i - Reserve space in the filesystem
1130 * @ip: the inode to reserve space for
1131 *
1132 * Returns: errno
1133 */
1134
1135 int gfs2_inplace_reserve_i(struct gfs2_inode *ip, char *file, unsigned int line)
1136 {
1137 struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
1138 struct gfs2_alloc *al = ip->i_alloc;
1139 struct inode *inode;
1140 int error = 0;
1141 u64 last_unlinked = NO_BLOCK;
1142
1143 if (gfs2_assert_warn(sdp, al->al_requested))
1144 return -EINVAL;
1145
1146 try_again:
1147 /* We need to hold the rindex unless the inode we're using is
1148 the rindex itself, in which case it's already held. */
1149 if (ip != GFS2_I(sdp->sd_rindex))
1150 error = gfs2_rindex_hold(sdp, &al->al_ri_gh);
1151 else if (!sdp->sd_rgrps) /* We may not have the rindex read in, so: */
1152 error = gfs2_ri_update_special(ip);
1153
1154 if (error)
1155 return error;
1156
1157 inode = get_local_rgrp(ip, &last_unlinked);
1158 if (inode) {
1159 if (ip != GFS2_I(sdp->sd_rindex))
1160 gfs2_glock_dq_uninit(&al->al_ri_gh);
1161 if (IS_ERR(inode))
1162 return PTR_ERR(inode);
1163 iput(inode);
1164 gfs2_log_flush(sdp, NULL);
1165 goto try_again;
1166 }
1167
1168 al->al_file = file;
1169 al->al_line = line;
1170
1171 return 0;
1172 }
1173
1174 /**
1175 * gfs2_inplace_release - release an inplace reservation
1176 * @ip: the inode the reservation was taken out on
1177 *
1178 * Release a reservation made by gfs2_inplace_reserve().
1179 */
1180
1181 void gfs2_inplace_release(struct gfs2_inode *ip)
1182 {
1183 struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
1184 struct gfs2_alloc *al = ip->i_alloc;
1185
1186 if (gfs2_assert_warn(sdp, al->al_alloced <= al->al_requested) == -1)
1187 fs_warn(sdp, "al_alloced = %u, al_requested = %u "
1188 "al_file = %s, al_line = %u\n",
1189 al->al_alloced, al->al_requested, al->al_file,
1190 al->al_line);
1191
1192 al->al_rgd = NULL;
1193 if (al->al_rgd_gh.gh_gl)
1194 gfs2_glock_dq_uninit(&al->al_rgd_gh);
1195 if (ip != GFS2_I(sdp->sd_rindex))
1196 gfs2_glock_dq_uninit(&al->al_ri_gh);
1197 }
1198
1199 /**
1200 * gfs2_get_block_type - Check a block in a RG is of given type
1201 * @rgd: the resource group holding the block
1202 * @block: the block number
1203 *
1204 * Returns: The block type (GFS2_BLKST_*)
1205 */
1206
1207 unsigned char gfs2_get_block_type(struct gfs2_rgrpd *rgd, u64 block)
1208 {
1209 struct gfs2_bitmap *bi = NULL;
1210 u32 length, rgrp_block, buf_block;
1211 unsigned int buf;
1212 unsigned char type;
1213
1214 length = rgd->rd_length;
1215 rgrp_block = block - rgd->rd_data0;
1216
1217 for (buf = 0; buf < length; buf++) {
1218 bi = rgd->rd_bits + buf;
1219 if (rgrp_block < (bi->bi_start + bi->bi_len) * GFS2_NBBY)
1220 break;
1221 }
1222
1223 gfs2_assert(rgd->rd_sbd, buf < length);
1224 buf_block = rgrp_block - bi->bi_start * GFS2_NBBY;
1225
1226 type = gfs2_testbit(rgd, bi->bi_bh->b_data + bi->bi_offset,
1227 bi->bi_len, buf_block);
1228
1229 return type;
1230 }
1231
1232 /**
1233 * rgblk_search - find a block in @old_state, change allocation
1234 * state to @new_state
1235 * @rgd: the resource group descriptor
1236 * @goal: the goal block within the RG (start here to search for avail block)
1237 * @old_state: GFS2_BLKST_XXX the before-allocation state to find
1238 * @new_state: GFS2_BLKST_XXX the after-allocation block state
1239 * @n: The extent length
1240 *
1241 * Walk rgrp's bitmap to find bits that represent a block in @old_state.
1242 * Add the found bitmap buffer to the transaction.
1243 * Set the found bits to @new_state to change block's allocation state.
1244 *
1245 * This function never fails, because we wouldn't call it unless we
1246 * know (from reservation results, etc.) that a block is available.
1247 *
1248 * Scope of @goal and returned block is just within rgrp, not the whole
1249 * filesystem.
1250 *
1251 * Returns: the block number allocated
1252 */
1253
1254 static u32 rgblk_search(struct gfs2_rgrpd *rgd, u32 goal,
1255 unsigned char old_state, unsigned char new_state,
1256 unsigned int *n)
1257 {
1258 struct gfs2_bitmap *bi = NULL;
1259 const u32 length = rgd->rd_length;
1260 u32 blk = 0;
1261 unsigned int buf, x;
1262 const unsigned int elen = *n;
1263 const u8 *buffer;
1264
1265 *n = 0;
1266 /* Find bitmap block that contains bits for goal block */
1267 for (buf = 0; buf < length; buf++) {
1268 bi = rgd->rd_bits + buf;
1269 if (goal < (bi->bi_start + bi->bi_len) * GFS2_NBBY)
1270 break;
1271 }
1272
1273 gfs2_assert(rgd->rd_sbd, buf < length);
1274
1275 /* Convert scope of "goal" from rgrp-wide to within found bit block */
1276 goal -= bi->bi_start * GFS2_NBBY;
1277
1278 /* Search (up to entire) bitmap in this rgrp for allocatable block.
1279 "x <= length", instead of "x < length", because we typically start
1280 the search in the middle of a bit block, but if we can't find an
1281 allocatable block anywhere else, we want to be able wrap around and
1282 search in the first part of our first-searched bit block. */
1283 for (x = 0; x <= length; x++) {
1284 /* The GFS2_BLKST_UNLINKED state doesn't apply to the clone
1285 bitmaps, so we must search the originals for that. */
1286 buffer = bi->bi_bh->b_data + bi->bi_offset;
1287 if (old_state != GFS2_BLKST_UNLINKED && bi->bi_clone)
1288 buffer = bi->bi_clone + bi->bi_offset;
1289
1290 blk = gfs2_bitfit(buffer, bi->bi_len, goal, old_state);
1291 if (blk != BFITNOENT)
1292 break;
1293
1294 /* Try next bitmap block (wrap back to rgrp header if at end) */
1295 buf = (buf + 1) % length;
1296 bi = rgd->rd_bits + buf;
1297 goal = 0;
1298 }
1299
1300 if (blk != BFITNOENT && old_state != new_state) {
1301 *n = 1;
1302 gfs2_trans_add_bh(rgd->rd_gl, bi->bi_bh, 1);
1303 gfs2_setbit(rgd, bi->bi_bh->b_data, bi->bi_clone, bi->bi_offset,
1304 bi->bi_len, blk, new_state);
1305 goal = blk;
1306 while (*n < elen) {
1307 goal++;
1308 if (goal >= (bi->bi_len * GFS2_NBBY))
1309 break;
1310 if (gfs2_testbit(rgd, buffer, bi->bi_len, goal) !=
1311 GFS2_BLKST_FREE)
1312 break;
1313 gfs2_setbit(rgd, bi->bi_bh->b_data, bi->bi_clone,
1314 bi->bi_offset, bi->bi_len, goal,
1315 new_state);
1316 (*n)++;
1317 }
1318 }
1319
1320 return (blk == BFITNOENT) ? blk : (bi->bi_start * GFS2_NBBY) + blk;
1321 }
1322
1323 /**
1324 * rgblk_free - Change alloc state of given block(s)
1325 * @sdp: the filesystem
1326 * @bstart: the start of a run of blocks to free
1327 * @blen: the length of the block run (all must lie within ONE RG!)
1328 * @new_state: GFS2_BLKST_XXX the after-allocation block state
1329 *
1330 * Returns: Resource group containing the block(s)
1331 */
1332
1333 static struct gfs2_rgrpd *rgblk_free(struct gfs2_sbd *sdp, u64 bstart,
1334 u32 blen, unsigned char new_state)
1335 {
1336 struct gfs2_rgrpd *rgd;
1337 struct gfs2_bitmap *bi = NULL;
1338 u32 length, rgrp_blk, buf_blk;
1339 unsigned int buf;
1340
1341 rgd = gfs2_blk2rgrpd(sdp, bstart);
1342 if (!rgd) {
1343 if (gfs2_consist(sdp))
1344 fs_err(sdp, "block = %llu\n", (unsigned long long)bstart);
1345 return NULL;
1346 }
1347
1348 length = rgd->rd_length;
1349
1350 rgrp_blk = bstart - rgd->rd_data0;
1351
1352 while (blen--) {
1353 for (buf = 0; buf < length; buf++) {
1354 bi = rgd->rd_bits + buf;
1355 if (rgrp_blk < (bi->bi_start + bi->bi_len) * GFS2_NBBY)
1356 break;
1357 }
1358
1359 gfs2_assert(rgd->rd_sbd, buf < length);
1360
1361 buf_blk = rgrp_blk - bi->bi_start * GFS2_NBBY;
1362 rgrp_blk++;
1363
1364 if (!bi->bi_clone) {
1365 bi->bi_clone = kmalloc(bi->bi_bh->b_size,
1366 GFP_NOFS | __GFP_NOFAIL);
1367 memcpy(bi->bi_clone + bi->bi_offset,
1368 bi->bi_bh->b_data + bi->bi_offset,
1369 bi->bi_len);
1370 }
1371 gfs2_trans_add_bh(rgd->rd_gl, bi->bi_bh, 1);
1372 gfs2_setbit(rgd, bi->bi_bh->b_data, NULL, bi->bi_offset,
1373 bi->bi_len, buf_blk, new_state);
1374 }
1375
1376 return rgd;
1377 }
1378
1379 /**
1380 * gfs2_alloc_block - Allocate a block
1381 * @ip: the inode to allocate the block for
1382 *
1383 * Returns: the allocated block
1384 */
1385
1386 u64 gfs2_alloc_block(struct gfs2_inode *ip, unsigned int *n)
1387 {
1388 struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
1389 struct gfs2_alloc *al = ip->i_alloc;
1390 struct gfs2_rgrpd *rgd = al->al_rgd;
1391 u32 goal, blk;
1392 u64 block;
1393
1394 if (rgrp_contains_block(rgd, ip->i_goal))
1395 goal = ip->i_goal - rgd->rd_data0;
1396 else
1397 goal = rgd->rd_last_alloc;
1398
1399 blk = rgblk_search(rgd, goal, GFS2_BLKST_FREE, GFS2_BLKST_USED, n);
1400 BUG_ON(blk == BFITNOENT);
1401
1402 rgd->rd_last_alloc = blk;
1403 block = rgd->rd_data0 + blk;
1404 ip->i_goal = block;
1405
1406 gfs2_assert_withdraw(sdp, rgd->rd_rg.rg_free >= *n);
1407 rgd->rd_rg.rg_free -= *n;
1408
1409 gfs2_trans_add_bh(rgd->rd_gl, rgd->rd_bits[0].bi_bh, 1);
1410 gfs2_rgrp_out(rgd, rgd->rd_bits[0].bi_bh->b_data);
1411
1412 al->al_alloced += *n;
1413
1414 gfs2_statfs_change(sdp, 0, -(s64)*n, 0);
1415 gfs2_quota_change(ip, *n, ip->i_inode.i_uid, ip->i_inode.i_gid);
1416
1417 spin_lock(&sdp->sd_rindex_spin);
1418 rgd->rd_free_clone -= *n;
1419 spin_unlock(&sdp->sd_rindex_spin);
1420
1421 return block;
1422 }
1423
1424 /**
1425 * gfs2_alloc_di - Allocate a dinode
1426 * @dip: the directory that the inode is going in
1427 *
1428 * Returns: the block allocated
1429 */
1430
1431 u64 gfs2_alloc_di(struct gfs2_inode *dip, u64 *generation)
1432 {
1433 struct gfs2_sbd *sdp = GFS2_SB(&dip->i_inode);
1434 struct gfs2_alloc *al = dip->i_alloc;
1435 struct gfs2_rgrpd *rgd = al->al_rgd;
1436 u32 blk;
1437 u64 block;
1438 unsigned int n = 1;
1439
1440 blk = rgblk_search(rgd, rgd->rd_last_alloc,
1441 GFS2_BLKST_FREE, GFS2_BLKST_DINODE, &n);
1442 BUG_ON(blk == BFITNOENT);
1443
1444 rgd->rd_last_alloc = blk;
1445
1446 block = rgd->rd_data0 + blk;
1447
1448 gfs2_assert_withdraw(sdp, rgd->rd_rg.rg_free);
1449 rgd->rd_rg.rg_free--;
1450 rgd->rd_rg.rg_dinodes++;
1451 *generation = rgd->rd_rg.rg_igeneration++;
1452 gfs2_trans_add_bh(rgd->rd_gl, rgd->rd_bits[0].bi_bh, 1);
1453 gfs2_rgrp_out(rgd, rgd->rd_bits[0].bi_bh->b_data);
1454
1455 al->al_alloced++;
1456
1457 gfs2_statfs_change(sdp, 0, -1, +1);
1458 gfs2_trans_add_unrevoke(sdp, block, 1);
1459
1460 spin_lock(&sdp->sd_rindex_spin);
1461 rgd->rd_free_clone--;
1462 spin_unlock(&sdp->sd_rindex_spin);
1463
1464 return block;
1465 }
1466
1467 /**
1468 * gfs2_free_data - free a contiguous run of data block(s)
1469 * @ip: the inode these blocks are being freed from
1470 * @bstart: first block of a run of contiguous blocks
1471 * @blen: the length of the block run
1472 *
1473 */
1474
1475 void gfs2_free_data(struct gfs2_inode *ip, u64 bstart, u32 blen)
1476 {
1477 struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
1478 struct gfs2_rgrpd *rgd;
1479
1480 rgd = rgblk_free(sdp, bstart, blen, GFS2_BLKST_FREE);
1481 if (!rgd)
1482 return;
1483
1484 rgd->rd_rg.rg_free += blen;
1485
1486 gfs2_trans_add_bh(rgd->rd_gl, rgd->rd_bits[0].bi_bh, 1);
1487 gfs2_rgrp_out(rgd, rgd->rd_bits[0].bi_bh->b_data);
1488
1489 gfs2_trans_add_rg(rgd);
1490
1491 gfs2_statfs_change(sdp, 0, +blen, 0);
1492 gfs2_quota_change(ip, -(s64)blen, ip->i_inode.i_uid, ip->i_inode.i_gid);
1493 }
1494
1495 /**
1496 * gfs2_free_meta - free a contiguous run of data block(s)
1497 * @ip: the inode these blocks are being freed from
1498 * @bstart: first block of a run of contiguous blocks
1499 * @blen: the length of the block run
1500 *
1501 */
1502
1503 void gfs2_free_meta(struct gfs2_inode *ip, u64 bstart, u32 blen)
1504 {
1505 struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
1506 struct gfs2_rgrpd *rgd;
1507
1508 rgd = rgblk_free(sdp, bstart, blen, GFS2_BLKST_FREE);
1509 if (!rgd)
1510 return;
1511
1512 rgd->rd_rg.rg_free += blen;
1513
1514 gfs2_trans_add_bh(rgd->rd_gl, rgd->rd_bits[0].bi_bh, 1);
1515 gfs2_rgrp_out(rgd, rgd->rd_bits[0].bi_bh->b_data);
1516
1517 gfs2_trans_add_rg(rgd);
1518
1519 gfs2_statfs_change(sdp, 0, +blen, 0);
1520 gfs2_quota_change(ip, -(s64)blen, ip->i_inode.i_uid, ip->i_inode.i_gid);
1521 gfs2_meta_wipe(ip, bstart, blen);
1522 }
1523
1524 void gfs2_unlink_di(struct inode *inode)
1525 {
1526 struct gfs2_inode *ip = GFS2_I(inode);
1527 struct gfs2_sbd *sdp = GFS2_SB(inode);
1528 struct gfs2_rgrpd *rgd;
1529 u64 blkno = ip->i_no_addr;
1530
1531 rgd = rgblk_free(sdp, blkno, 1, GFS2_BLKST_UNLINKED);
1532 if (!rgd)
1533 return;
1534 gfs2_trans_add_bh(rgd->rd_gl, rgd->rd_bits[0].bi_bh, 1);
1535 gfs2_rgrp_out(rgd, rgd->rd_bits[0].bi_bh->b_data);
1536 gfs2_trans_add_rg(rgd);
1537 }
1538
1539 static void gfs2_free_uninit_di(struct gfs2_rgrpd *rgd, u64 blkno)
1540 {
1541 struct gfs2_sbd *sdp = rgd->rd_sbd;
1542 struct gfs2_rgrpd *tmp_rgd;
1543
1544 tmp_rgd = rgblk_free(sdp, blkno, 1, GFS2_BLKST_FREE);
1545 if (!tmp_rgd)
1546 return;
1547 gfs2_assert_withdraw(sdp, rgd == tmp_rgd);
1548
1549 if (!rgd->rd_rg.rg_dinodes)
1550 gfs2_consist_rgrpd(rgd);
1551 rgd->rd_rg.rg_dinodes--;
1552 rgd->rd_rg.rg_free++;
1553
1554 gfs2_trans_add_bh(rgd->rd_gl, rgd->rd_bits[0].bi_bh, 1);
1555 gfs2_rgrp_out(rgd, rgd->rd_bits[0].bi_bh->b_data);
1556
1557 gfs2_statfs_change(sdp, 0, +1, -1);
1558 gfs2_trans_add_rg(rgd);
1559 }
1560
1561
1562 void gfs2_free_di(struct gfs2_rgrpd *rgd, struct gfs2_inode *ip)
1563 {
1564 gfs2_free_uninit_di(rgd, ip->i_no_addr);
1565 gfs2_quota_change(ip, -1, ip->i_inode.i_uid, ip->i_inode.i_gid);
1566 gfs2_meta_wipe(ip, ip->i_no_addr, 1);
1567 }
1568
1569 /**
1570 * gfs2_rlist_add - add a RG to a list of RGs
1571 * @sdp: the filesystem
1572 * @rlist: the list of resource groups
1573 * @block: the block
1574 *
1575 * Figure out what RG a block belongs to and add that RG to the list
1576 *
1577 * FIXME: Don't use NOFAIL
1578 *
1579 */
1580
1581 void gfs2_rlist_add(struct gfs2_sbd *sdp, struct gfs2_rgrp_list *rlist,
1582 u64 block)
1583 {
1584 struct gfs2_rgrpd *rgd;
1585 struct gfs2_rgrpd **tmp;
1586 unsigned int new_space;
1587 unsigned int x;
1588
1589 if (gfs2_assert_warn(sdp, !rlist->rl_ghs))
1590 return;
1591
1592 rgd = gfs2_blk2rgrpd(sdp, block);
1593 if (!rgd) {
1594 if (gfs2_consist(sdp))
1595 fs_err(sdp, "block = %llu\n", (unsigned long long)block);
1596 return;
1597 }
1598
1599 for (x = 0; x < rlist->rl_rgrps; x++)
1600 if (rlist->rl_rgd[x] == rgd)
1601 return;
1602
1603 if (rlist->rl_rgrps == rlist->rl_space) {
1604 new_space = rlist->rl_space + 10;
1605
1606 tmp = kcalloc(new_space, sizeof(struct gfs2_rgrpd *),
1607 GFP_NOFS | __GFP_NOFAIL);
1608
1609 if (rlist->rl_rgd) {
1610 memcpy(tmp, rlist->rl_rgd,
1611 rlist->rl_space * sizeof(struct gfs2_rgrpd *));
1612 kfree(rlist->rl_rgd);
1613 }
1614
1615 rlist->rl_space = new_space;
1616 rlist->rl_rgd = tmp;
1617 }
1618
1619 rlist->rl_rgd[rlist->rl_rgrps++] = rgd;
1620 }
1621
1622 /**
1623 * gfs2_rlist_alloc - all RGs have been added to the rlist, now allocate
1624 * and initialize an array of glock holders for them
1625 * @rlist: the list of resource groups
1626 * @state: the lock state to acquire the RG lock in
1627 * @flags: the modifier flags for the holder structures
1628 *
1629 * FIXME: Don't use NOFAIL
1630 *
1631 */
1632
1633 void gfs2_rlist_alloc(struct gfs2_rgrp_list *rlist, unsigned int state)
1634 {
1635 unsigned int x;
1636
1637 rlist->rl_ghs = kcalloc(rlist->rl_rgrps, sizeof(struct gfs2_holder),
1638 GFP_NOFS | __GFP_NOFAIL);
1639 for (x = 0; x < rlist->rl_rgrps; x++)
1640 gfs2_holder_init(rlist->rl_rgd[x]->rd_gl,
1641 state, 0,
1642 &rlist->rl_ghs[x]);
1643 }
1644
1645 /**
1646 * gfs2_rlist_free - free a resource group list
1647 * @list: the list of resource groups
1648 *
1649 */
1650
1651 void gfs2_rlist_free(struct gfs2_rgrp_list *rlist)
1652 {
1653 unsigned int x;
1654
1655 kfree(rlist->rl_rgd);
1656
1657 if (rlist->rl_ghs) {
1658 for (x = 0; x < rlist->rl_rgrps; x++)
1659 gfs2_holder_uninit(&rlist->rl_ghs[x]);
1660 kfree(rlist->rl_ghs);
1661 }
1662 }
1663
This page took 0.162016 seconds and 5 git commands to generate.