xref: /openbmc/linux/fs/xfs/libxfs/xfs_alloc_btree.c (revision 4f3db074)
1 /*
2  * Copyright (c) 2000-2001,2005 Silicon Graphics, Inc.
3  * All Rights Reserved.
4  *
5  * This program is free software; you can redistribute it and/or
6  * modify it under the terms of the GNU General Public License as
7  * published by the Free Software Foundation.
8  *
9  * This program is distributed in the hope that it would be useful,
10  * but WITHOUT ANY WARRANTY; without even the implied warranty of
11  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
12  * GNU General Public License for more details.
13  *
14  * You should have received a copy of the GNU General Public License
15  * along with this program; if not, write the Free Software Foundation,
16  * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
17  */
18 #include "xfs.h"
19 #include "xfs_fs.h"
20 #include "xfs_shared.h"
21 #include "xfs_format.h"
22 #include "xfs_log_format.h"
23 #include "xfs_trans_resv.h"
24 #include "xfs_sb.h"
25 #include "xfs_mount.h"
26 #include "xfs_btree.h"
27 #include "xfs_alloc_btree.h"
28 #include "xfs_alloc.h"
29 #include "xfs_extent_busy.h"
30 #include "xfs_error.h"
31 #include "xfs_trace.h"
32 #include "xfs_cksum.h"
33 #include "xfs_trans.h"
34 
35 
36 STATIC struct xfs_btree_cur *
37 xfs_allocbt_dup_cursor(
38 	struct xfs_btree_cur	*cur)
39 {
40 	return xfs_allocbt_init_cursor(cur->bc_mp, cur->bc_tp,
41 			cur->bc_private.a.agbp, cur->bc_private.a.agno,
42 			cur->bc_btnum);
43 }
44 
45 STATIC void
46 xfs_allocbt_set_root(
47 	struct xfs_btree_cur	*cur,
48 	union xfs_btree_ptr	*ptr,
49 	int			inc)
50 {
51 	struct xfs_buf		*agbp = cur->bc_private.a.agbp;
52 	struct xfs_agf		*agf = XFS_BUF_TO_AGF(agbp);
53 	xfs_agnumber_t		seqno = be32_to_cpu(agf->agf_seqno);
54 	int			btnum = cur->bc_btnum;
55 	struct xfs_perag	*pag = xfs_perag_get(cur->bc_mp, seqno);
56 
57 	ASSERT(ptr->s != 0);
58 
59 	agf->agf_roots[btnum] = ptr->s;
60 	be32_add_cpu(&agf->agf_levels[btnum], inc);
61 	pag->pagf_levels[btnum] += inc;
62 	xfs_perag_put(pag);
63 
64 	xfs_alloc_log_agf(cur->bc_tp, agbp, XFS_AGF_ROOTS | XFS_AGF_LEVELS);
65 }
66 
67 STATIC int
68 xfs_allocbt_alloc_block(
69 	struct xfs_btree_cur	*cur,
70 	union xfs_btree_ptr	*start,
71 	union xfs_btree_ptr	*new,
72 	int			*stat)
73 {
74 	int			error;
75 	xfs_agblock_t		bno;
76 
77 	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
78 
79 	/* Allocate the new block from the freelist. If we can't, give up.  */
80 	error = xfs_alloc_get_freelist(cur->bc_tp, cur->bc_private.a.agbp,
81 				       &bno, 1);
82 	if (error) {
83 		XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
84 		return error;
85 	}
86 
87 	if (bno == NULLAGBLOCK) {
88 		XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
89 		*stat = 0;
90 		return 0;
91 	}
92 
93 	xfs_extent_busy_reuse(cur->bc_mp, cur->bc_private.a.agno, bno, 1, false);
94 
95 	xfs_trans_agbtree_delta(cur->bc_tp, 1);
96 	new->s = cpu_to_be32(bno);
97 
98 	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
99 	*stat = 1;
100 	return 0;
101 }
102 
103 STATIC int
104 xfs_allocbt_free_block(
105 	struct xfs_btree_cur	*cur,
106 	struct xfs_buf		*bp)
107 {
108 	struct xfs_buf		*agbp = cur->bc_private.a.agbp;
109 	struct xfs_agf		*agf = XFS_BUF_TO_AGF(agbp);
110 	xfs_agblock_t		bno;
111 	int			error;
112 
113 	bno = xfs_daddr_to_agbno(cur->bc_mp, XFS_BUF_ADDR(bp));
114 	error = xfs_alloc_put_freelist(cur->bc_tp, agbp, NULL, bno, 1);
115 	if (error)
116 		return error;
117 
118 	xfs_extent_busy_insert(cur->bc_tp, be32_to_cpu(agf->agf_seqno), bno, 1,
119 			      XFS_EXTENT_BUSY_SKIP_DISCARD);
120 	xfs_trans_agbtree_delta(cur->bc_tp, -1);
121 
122 	xfs_trans_binval(cur->bc_tp, bp);
123 	return 0;
124 }
125 
126 /*
127  * Update the longest extent in the AGF
128  */
129 STATIC void
130 xfs_allocbt_update_lastrec(
131 	struct xfs_btree_cur	*cur,
132 	struct xfs_btree_block	*block,
133 	union xfs_btree_rec	*rec,
134 	int			ptr,
135 	int			reason)
136 {
137 	struct xfs_agf		*agf = XFS_BUF_TO_AGF(cur->bc_private.a.agbp);
138 	xfs_agnumber_t		seqno = be32_to_cpu(agf->agf_seqno);
139 	struct xfs_perag	*pag;
140 	__be32			len;
141 	int			numrecs;
142 
143 	ASSERT(cur->bc_btnum == XFS_BTNUM_CNT);
144 
145 	switch (reason) {
146 	case LASTREC_UPDATE:
147 		/*
148 		 * If this is the last leaf block and it's the last record,
149 		 * then update the size of the longest extent in the AG.
150 		 */
151 		if (ptr != xfs_btree_get_numrecs(block))
152 			return;
153 		len = rec->alloc.ar_blockcount;
154 		break;
155 	case LASTREC_INSREC:
156 		if (be32_to_cpu(rec->alloc.ar_blockcount) <=
157 		    be32_to_cpu(agf->agf_longest))
158 			return;
159 		len = rec->alloc.ar_blockcount;
160 		break;
161 	case LASTREC_DELREC:
162 		numrecs = xfs_btree_get_numrecs(block);
163 		if (ptr <= numrecs)
164 			return;
165 		ASSERT(ptr == numrecs + 1);
166 
167 		if (numrecs) {
168 			xfs_alloc_rec_t *rrp;
169 
170 			rrp = XFS_ALLOC_REC_ADDR(cur->bc_mp, block, numrecs);
171 			len = rrp->ar_blockcount;
172 		} else {
173 			len = 0;
174 		}
175 
176 		break;
177 	default:
178 		ASSERT(0);
179 		return;
180 	}
181 
182 	agf->agf_longest = len;
183 	pag = xfs_perag_get(cur->bc_mp, seqno);
184 	pag->pagf_longest = be32_to_cpu(len);
185 	xfs_perag_put(pag);
186 	xfs_alloc_log_agf(cur->bc_tp, cur->bc_private.a.agbp, XFS_AGF_LONGEST);
187 }
188 
189 STATIC int
190 xfs_allocbt_get_minrecs(
191 	struct xfs_btree_cur	*cur,
192 	int			level)
193 {
194 	return cur->bc_mp->m_alloc_mnr[level != 0];
195 }
196 
197 STATIC int
198 xfs_allocbt_get_maxrecs(
199 	struct xfs_btree_cur	*cur,
200 	int			level)
201 {
202 	return cur->bc_mp->m_alloc_mxr[level != 0];
203 }
204 
205 STATIC void
206 xfs_allocbt_init_key_from_rec(
207 	union xfs_btree_key	*key,
208 	union xfs_btree_rec	*rec)
209 {
210 	ASSERT(rec->alloc.ar_startblock != 0);
211 
212 	key->alloc.ar_startblock = rec->alloc.ar_startblock;
213 	key->alloc.ar_blockcount = rec->alloc.ar_blockcount;
214 }
215 
216 STATIC void
217 xfs_allocbt_init_rec_from_key(
218 	union xfs_btree_key	*key,
219 	union xfs_btree_rec	*rec)
220 {
221 	ASSERT(key->alloc.ar_startblock != 0);
222 
223 	rec->alloc.ar_startblock = key->alloc.ar_startblock;
224 	rec->alloc.ar_blockcount = key->alloc.ar_blockcount;
225 }
226 
227 STATIC void
228 xfs_allocbt_init_rec_from_cur(
229 	struct xfs_btree_cur	*cur,
230 	union xfs_btree_rec	*rec)
231 {
232 	ASSERT(cur->bc_rec.a.ar_startblock != 0);
233 
234 	rec->alloc.ar_startblock = cpu_to_be32(cur->bc_rec.a.ar_startblock);
235 	rec->alloc.ar_blockcount = cpu_to_be32(cur->bc_rec.a.ar_blockcount);
236 }
237 
238 STATIC void
239 xfs_allocbt_init_ptr_from_cur(
240 	struct xfs_btree_cur	*cur,
241 	union xfs_btree_ptr	*ptr)
242 {
243 	struct xfs_agf		*agf = XFS_BUF_TO_AGF(cur->bc_private.a.agbp);
244 
245 	ASSERT(cur->bc_private.a.agno == be32_to_cpu(agf->agf_seqno));
246 	ASSERT(agf->agf_roots[cur->bc_btnum] != 0);
247 
248 	ptr->s = agf->agf_roots[cur->bc_btnum];
249 }
250 
251 STATIC __int64_t
252 xfs_allocbt_key_diff(
253 	struct xfs_btree_cur	*cur,
254 	union xfs_btree_key	*key)
255 {
256 	xfs_alloc_rec_incore_t	*rec = &cur->bc_rec.a;
257 	xfs_alloc_key_t		*kp = &key->alloc;
258 	__int64_t		diff;
259 
260 	if (cur->bc_btnum == XFS_BTNUM_BNO) {
261 		return (__int64_t)be32_to_cpu(kp->ar_startblock) -
262 				rec->ar_startblock;
263 	}
264 
265 	diff = (__int64_t)be32_to_cpu(kp->ar_blockcount) - rec->ar_blockcount;
266 	if (diff)
267 		return diff;
268 
269 	return (__int64_t)be32_to_cpu(kp->ar_startblock) - rec->ar_startblock;
270 }
271 
272 static bool
273 xfs_allocbt_verify(
274 	struct xfs_buf		*bp)
275 {
276 	struct xfs_mount	*mp = bp->b_target->bt_mount;
277 	struct xfs_btree_block	*block = XFS_BUF_TO_BLOCK(bp);
278 	struct xfs_perag	*pag = bp->b_pag;
279 	unsigned int		level;
280 
281 	/*
282 	 * magic number and level verification
283 	 *
284 	 * During growfs operations, we can't verify the exact level or owner as
285 	 * the perag is not fully initialised and hence not attached to the
286 	 * buffer.  In this case, check against the maximum tree depth.
287 	 *
288 	 * Similarly, during log recovery we will have a perag structure
289 	 * attached, but the agf information will not yet have been initialised
290 	 * from the on disk AGF. Again, we can only check against maximum limits
291 	 * in this case.
292 	 */
293 	level = be16_to_cpu(block->bb_level);
294 	switch (block->bb_magic) {
295 	case cpu_to_be32(XFS_ABTB_CRC_MAGIC):
296 		if (!xfs_sb_version_hascrc(&mp->m_sb))
297 			return false;
298 		if (!uuid_equal(&block->bb_u.s.bb_uuid, &mp->m_sb.sb_uuid))
299 			return false;
300 		if (block->bb_u.s.bb_blkno != cpu_to_be64(bp->b_bn))
301 			return false;
302 		if (pag &&
303 		    be32_to_cpu(block->bb_u.s.bb_owner) != pag->pag_agno)
304 			return false;
305 		/* fall through */
306 	case cpu_to_be32(XFS_ABTB_MAGIC):
307 		if (pag && pag->pagf_init) {
308 			if (level >= pag->pagf_levels[XFS_BTNUM_BNOi])
309 				return false;
310 		} else if (level >= mp->m_ag_maxlevels)
311 			return false;
312 		break;
313 	case cpu_to_be32(XFS_ABTC_CRC_MAGIC):
314 		if (!xfs_sb_version_hascrc(&mp->m_sb))
315 			return false;
316 		if (!uuid_equal(&block->bb_u.s.bb_uuid, &mp->m_sb.sb_uuid))
317 			return false;
318 		if (block->bb_u.s.bb_blkno != cpu_to_be64(bp->b_bn))
319 			return false;
320 		if (pag &&
321 		    be32_to_cpu(block->bb_u.s.bb_owner) != pag->pag_agno)
322 			return false;
323 		/* fall through */
324 	case cpu_to_be32(XFS_ABTC_MAGIC):
325 		if (pag && pag->pagf_init) {
326 			if (level >= pag->pagf_levels[XFS_BTNUM_CNTi])
327 				return false;
328 		} else if (level >= mp->m_ag_maxlevels)
329 			return false;
330 		break;
331 	default:
332 		return false;
333 	}
334 
335 	/* numrecs verification */
336 	if (be16_to_cpu(block->bb_numrecs) > mp->m_alloc_mxr[level != 0])
337 		return false;
338 
339 	/* sibling pointer verification */
340 	if (!block->bb_u.s.bb_leftsib ||
341 	    (be32_to_cpu(block->bb_u.s.bb_leftsib) >= mp->m_sb.sb_agblocks &&
342 	     block->bb_u.s.bb_leftsib != cpu_to_be32(NULLAGBLOCK)))
343 		return false;
344 	if (!block->bb_u.s.bb_rightsib ||
345 	    (be32_to_cpu(block->bb_u.s.bb_rightsib) >= mp->m_sb.sb_agblocks &&
346 	     block->bb_u.s.bb_rightsib != cpu_to_be32(NULLAGBLOCK)))
347 		return false;
348 
349 	return true;
350 }
351 
352 static void
353 xfs_allocbt_read_verify(
354 	struct xfs_buf	*bp)
355 {
356 	if (!xfs_btree_sblock_verify_crc(bp))
357 		xfs_buf_ioerror(bp, -EFSBADCRC);
358 	else if (!xfs_allocbt_verify(bp))
359 		xfs_buf_ioerror(bp, -EFSCORRUPTED);
360 
361 	if (bp->b_error) {
362 		trace_xfs_btree_corrupt(bp, _RET_IP_);
363 		xfs_verifier_error(bp);
364 	}
365 }
366 
367 static void
368 xfs_allocbt_write_verify(
369 	struct xfs_buf	*bp)
370 {
371 	if (!xfs_allocbt_verify(bp)) {
372 		trace_xfs_btree_corrupt(bp, _RET_IP_);
373 		xfs_buf_ioerror(bp, -EFSCORRUPTED);
374 		xfs_verifier_error(bp);
375 		return;
376 	}
377 	xfs_btree_sblock_calc_crc(bp);
378 
379 }
380 
381 const struct xfs_buf_ops xfs_allocbt_buf_ops = {
382 	.verify_read = xfs_allocbt_read_verify,
383 	.verify_write = xfs_allocbt_write_verify,
384 };
385 
386 
387 #if defined(DEBUG) || defined(XFS_WARN)
388 STATIC int
389 xfs_allocbt_keys_inorder(
390 	struct xfs_btree_cur	*cur,
391 	union xfs_btree_key	*k1,
392 	union xfs_btree_key	*k2)
393 {
394 	if (cur->bc_btnum == XFS_BTNUM_BNO) {
395 		return be32_to_cpu(k1->alloc.ar_startblock) <
396 		       be32_to_cpu(k2->alloc.ar_startblock);
397 	} else {
398 		return be32_to_cpu(k1->alloc.ar_blockcount) <
399 			be32_to_cpu(k2->alloc.ar_blockcount) ||
400 			(k1->alloc.ar_blockcount == k2->alloc.ar_blockcount &&
401 			 be32_to_cpu(k1->alloc.ar_startblock) <
402 			 be32_to_cpu(k2->alloc.ar_startblock));
403 	}
404 }
405 
406 STATIC int
407 xfs_allocbt_recs_inorder(
408 	struct xfs_btree_cur	*cur,
409 	union xfs_btree_rec	*r1,
410 	union xfs_btree_rec	*r2)
411 {
412 	if (cur->bc_btnum == XFS_BTNUM_BNO) {
413 		return be32_to_cpu(r1->alloc.ar_startblock) +
414 			be32_to_cpu(r1->alloc.ar_blockcount) <=
415 			be32_to_cpu(r2->alloc.ar_startblock);
416 	} else {
417 		return be32_to_cpu(r1->alloc.ar_blockcount) <
418 			be32_to_cpu(r2->alloc.ar_blockcount) ||
419 			(r1->alloc.ar_blockcount == r2->alloc.ar_blockcount &&
420 			 be32_to_cpu(r1->alloc.ar_startblock) <
421 			 be32_to_cpu(r2->alloc.ar_startblock));
422 	}
423 }
424 #endif	/* DEBUG */
425 
426 static const struct xfs_btree_ops xfs_allocbt_ops = {
427 	.rec_len		= sizeof(xfs_alloc_rec_t),
428 	.key_len		= sizeof(xfs_alloc_key_t),
429 
430 	.dup_cursor		= xfs_allocbt_dup_cursor,
431 	.set_root		= xfs_allocbt_set_root,
432 	.alloc_block		= xfs_allocbt_alloc_block,
433 	.free_block		= xfs_allocbt_free_block,
434 	.update_lastrec		= xfs_allocbt_update_lastrec,
435 	.get_minrecs		= xfs_allocbt_get_minrecs,
436 	.get_maxrecs		= xfs_allocbt_get_maxrecs,
437 	.init_key_from_rec	= xfs_allocbt_init_key_from_rec,
438 	.init_rec_from_key	= xfs_allocbt_init_rec_from_key,
439 	.init_rec_from_cur	= xfs_allocbt_init_rec_from_cur,
440 	.init_ptr_from_cur	= xfs_allocbt_init_ptr_from_cur,
441 	.key_diff		= xfs_allocbt_key_diff,
442 	.buf_ops		= &xfs_allocbt_buf_ops,
443 #if defined(DEBUG) || defined(XFS_WARN)
444 	.keys_inorder		= xfs_allocbt_keys_inorder,
445 	.recs_inorder		= xfs_allocbt_recs_inorder,
446 #endif
447 };
448 
449 /*
450  * Allocate a new allocation btree cursor.
451  */
452 struct xfs_btree_cur *			/* new alloc btree cursor */
453 xfs_allocbt_init_cursor(
454 	struct xfs_mount	*mp,		/* file system mount point */
455 	struct xfs_trans	*tp,		/* transaction pointer */
456 	struct xfs_buf		*agbp,		/* buffer for agf structure */
457 	xfs_agnumber_t		agno,		/* allocation group number */
458 	xfs_btnum_t		btnum)		/* btree identifier */
459 {
460 	struct xfs_agf		*agf = XFS_BUF_TO_AGF(agbp);
461 	struct xfs_btree_cur	*cur;
462 
463 	ASSERT(btnum == XFS_BTNUM_BNO || btnum == XFS_BTNUM_CNT);
464 
465 	cur = kmem_zone_zalloc(xfs_btree_cur_zone, KM_SLEEP);
466 
467 	cur->bc_tp = tp;
468 	cur->bc_mp = mp;
469 	cur->bc_btnum = btnum;
470 	cur->bc_blocklog = mp->m_sb.sb_blocklog;
471 	cur->bc_ops = &xfs_allocbt_ops;
472 
473 	if (btnum == XFS_BTNUM_CNT) {
474 		cur->bc_nlevels = be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNT]);
475 		cur->bc_flags = XFS_BTREE_LASTREC_UPDATE;
476 	} else {
477 		cur->bc_nlevels = be32_to_cpu(agf->agf_levels[XFS_BTNUM_BNO]);
478 	}
479 
480 	cur->bc_private.a.agbp = agbp;
481 	cur->bc_private.a.agno = agno;
482 
483 	if (xfs_sb_version_hascrc(&mp->m_sb))
484 		cur->bc_flags |= XFS_BTREE_CRC_BLOCKS;
485 
486 	return cur;
487 }
488 
489 /*
490  * Calculate number of records in an alloc btree block.
491  */
492 int
493 xfs_allocbt_maxrecs(
494 	struct xfs_mount	*mp,
495 	int			blocklen,
496 	int			leaf)
497 {
498 	blocklen -= XFS_ALLOC_BLOCK_LEN(mp);
499 
500 	if (leaf)
501 		return blocklen / sizeof(xfs_alloc_rec_t);
502 	return blocklen / (sizeof(xfs_alloc_key_t) + sizeof(xfs_alloc_ptr_t));
503 }
504