1 /* $OpenBSD: ufs_bmap.c,v 1.25 2007/06/01 23:47:57 deraadt Exp $ */
2 /* $NetBSD: ufs_bmap.c,v 1.3 1996/02/09 22:36:00 christos Exp $ */
3
4 /*
5 * Copyright (c) 1989, 1991, 1993
6 * The Regents of the University of California. All rights reserved.
7 * (c) UNIX System Laboratories, Inc.
8 * All or some portions of this file are derived from material licensed
9 * to the University of California by American Telephone and Telegraph
10 * Co. or Unix System Laboratories, Inc. and are reproduced herein with
11 * the permission of UNIX System Laboratories, Inc.
12 *
13 * Redistribution and use in source and binary forms, with or without
14 * modification, are permitted provided that the following conditions
15 * are met:
16 * 1. Redistributions of source code must retain the above copyright
17 * notice, this list of conditions and the following disclaimer.
18 * 2. Redistributions in binary form must reproduce the above copyright
19 * notice, this list of conditions and the following disclaimer in the
20 * documentation and/or other materials provided with the distribution.
21 * 3. Neither the name of the University nor the names of its contributors
22 * may be used to endorse or promote products derived from this software
23 * without specific prior written permission.
24 *
25 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
26 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
27 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
28 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
29 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
30 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
31 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
32 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
33 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
34 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
35 * SUCH DAMAGE.
36 *
37 * @(#)ufs_bmap.c 8.6 (Berkeley) 1/21/94
38 */
39
40 #include <sys/param.h>
41 #include <sys/systm.h>
42 #include <sys/buf.h>
43 #include <sys/proc.h>
44 #include <sys/vnode.h>
45 #include <sys/mount.h>
46 #include <sys/resourcevar.h>
47
48 #include <miscfs/specfs/specdev.h>
49
50 #include <ufs/ufs/quota.h>
51 #include <ufs/ufs/inode.h>
52 #include <ufs/ufs/ufsmount.h>
53 #include <ufs/ufs/ufs_extern.h>
54
55 /*
56 * Bmap converts a the logical block number of a file to its physical block
57 * number on the disk. The conversion is done by using the logical block
58 * number to index into the array of block pointers described by the dinode.
59 */
60 int
61 ufs_bmap(void *v)
62 {
63 struct vop_bmap_args *ap = v;
64
65 /*
66 * Check for underlying vnode requests and ensure that logical
67 * to physical mapping is requested.
68 */
69 if (ap->a_vpp != NULL)
70 *ap->a_vpp = VTOI(ap->a_vp)->i_devvp;
71 if (ap->a_bnp == NULL)
72 return (0);
73
74 return (ufs_bmaparray(ap->a_vp, ap->a_bn, ap->a_bnp, NULL, NULL,
75 ap->a_runp));
76 }
77
78 /*
79 * Indirect blocks are now on the vnode for the file. They are given negative
80 * logical block numbers. Indirect blocks are addressed by the negative
81 * address of the first data block to which they point. Double indirect blocks
82 * are addressed by one less than the address of the first indirect block to
83 * which they point. Triple indirect blocks are addressed by one less than
84 * the address of the first double indirect block to which they point.
85 *
86 * ufs_bmaparray does the bmap conversion, and if requested returns the
87 * array of logical blocks which must be traversed to get to a block.
88 * Each entry contains the offset into that block that gets you to the
89 * next block and the disk address of the block (if it is assigned).
90 */
91 int
92 ufs_bmaparray(struct vnode *vp, daddr_t bn, daddr64_t *bnp, struct indir *ap,
93 int *nump, int *runp)
94 {
95 struct inode *ip;
96 struct buf *bp;
97 struct ufsmount *ump;
98 struct mount *mp;
99 struct vnode *devvp;
100 struct indir a[NIADDR+1], *xap;
101 daddr64_t daddr;
102 long metalbn;
103 int error, maxrun = 0, num;
104
105 ip = VTOI(vp);
106 mp = vp->v_mount;
107 ump = VFSTOUFS(mp);
108 #ifdef DIAGNOSTIC
109 if ((ap != NULL && nump == NULL) || (ap == NULL && nump != NULL))
110 panic("ufs_bmaparray: invalid arguments");
111 #endif
112
113 if (runp) {
114 /*
115 * XXX
116 * If MAXBSIZE is the largest transfer the disks can handle,
117 * we probably want maxrun to be 1 block less so that we
118 * don't create a block larger than the device can handle.
119 */
120 *runp = 0;
121 maxrun = MAXBSIZE / mp->mnt_stat.f_iosize - 1;
122 }
123
124 xap = ap == NULL ? a : ap;
125 if (!nump)
126 nump = #
127 if ((error = ufs_getlbns(vp, bn, xap, nump)) != 0)
128 return (error);
129
130 num = *nump;
131 if (num == 0) {
132 *bnp = blkptrtodb(ump, DIP(ip, db[bn]));
133 if (*bnp == 0)
134 *bnp = -1;
135 else if (runp)
136 for (++bn; bn < NDADDR && *runp < maxrun &&
137 is_sequential(ump, DIP(ip, db[bn - 1]),
138 DIP(ip, db[bn]));
139 ++bn, ++*runp);
140 return (0);
141 }
142
143
144 /* Get disk address out of indirect block array */
145 daddr = DIP(ip, ib[xap->in_off]);
146
147 devvp = VFSTOUFS(vp->v_mount)->um_devvp;
148 for (bp = NULL, ++xap; --num; ++xap) {
149 /*
150 * Exit the loop if there is no disk address assigned yet and
151 * the indirect block isn't in the cache, or if we were
152 * looking for an indirect block and we've found it.
153 */
154
155 metalbn = xap->in_lbn;
156 if ((daddr == 0 && !incore(vp, metalbn)) || metalbn == bn)
157 break;
158 /*
159 * If we get here, we've either got the block in the cache
160 * or we have a disk address for it, go fetch it.
161 */
162 if (bp)
163 brelse(bp);
164
165 xap->in_exists = 1;
166 bp = getblk(vp, metalbn, mp->mnt_stat.f_iosize, 0, 0);
167 if (bp->b_flags & (B_DONE | B_DELWRI)) {
168 ;
169 }
170 #ifdef DIAGNOSTIC
171 else if (!daddr)
172 panic("ufs_bmaparray: indirect block not in cache");
173 #endif
174 else {
175 bp->b_blkno = blkptrtodb(ump, daddr);
176 bp->b_flags |= B_READ;
177 VOP_STRATEGY(bp);
178 curproc->p_stats->p_ru.ru_inblock++; /* XXX */
179 if ((error = biowait(bp)) != 0) {
180 brelse(bp);
181 return (error);
182 }
183 }
184
185 #ifdef FFS2
186 if (ip->i_ump->um_fstype == UM_UFS2) {
187 daddr = ((daddr64_t *)bp->b_data)[xap->in_off];
188 if (num == 1 && daddr && runp)
189 for (bn = xap->in_off + 1;
190 bn < MNINDIR(ump) && *runp < maxrun &&
191 is_sequential(ump,
192 ((daddr64_t *)bp->b_data)[bn - 1],
193 ((daddr64_t *)bp->b_data)[bn]);
194 ++bn, ++*runp);
195
196 continue;
197 }
198
199 #endif /* FFS2 */
200
201 daddr = ((daddr_t *)bp->b_data)[xap->in_off];
202 if (num == 1 && daddr && runp)
203 for (bn = xap->in_off + 1;
204 bn < MNINDIR(ump) && *runp < maxrun &&
205 is_sequential(ump,
206 ((daddr_t *)bp->b_data)[bn - 1],
207 ((daddr_t *)bp->b_data)[bn]);
208 ++bn, ++*runp);
209 }
210 if (bp)
211 brelse(bp);
212
213 daddr = blkptrtodb(ump, daddr);
214 *bnp = daddr == 0 ? -1 : daddr;
215 return (0);
216 }
217
218 /*
219 * Create an array of logical block number/offset pairs which represent the
220 * path of indirect blocks required to access a data block. The first "pair"
221 * contains the logical block number of the appropriate single, double or
222 * triple indirect block and the offset into the inode indirect block array.
223 * Note, the logical block number of the inode single/double/triple indirect
224 * block appears twice in the array, once with the offset into the i_ffs_ib and
225 * once with the offset into the page itself.
226 */
227 int
228 ufs_getlbns(struct vnode *vp, daddr_t bn, struct indir *ap, int *nump)
229 {
230 long metalbn, realbn;
231 struct ufsmount *ump;
232 int64_t blockcnt;
233 int i, numlevels, off;
234
235 ump = VFSTOUFS(vp->v_mount);
236 if (nump)
237 *nump = 0;
238 numlevels = 0;
239 realbn = bn;
240 if ((long)bn < 0)
241 bn = -(long)bn;
242
243 #ifdef DIAGNOSTIC
244 if (realbn < 0 && realbn > -NDADDR) {
245 panic ("ufs_getlbns: Invalid indirect block %d specified",
246 realbn);
247 }
248 #endif
249
250 /* The first NDADDR blocks are direct blocks. */
251 if (bn < NDADDR)
252 return (0);
253
254 /*
255 * Determine the number of levels of indirection. After this loop
256 * is done, blockcnt indicates the number of data blocks possible
257 * at the given level of indirection, and NIADDR - i is the number
258 * of levels of indirection needed to locate the requested block.
259 */
260 for (blockcnt = 1, i = NIADDR, bn -= NDADDR;; i--, bn -= blockcnt) {
261 if (i == 0)
262 return (EFBIG);
263 blockcnt *= MNINDIR(ump);
264 if (bn < blockcnt)
265 break;
266 }
267
268 /* Calculate the address of the first meta-block. */
269 if (realbn >= 0)
270 metalbn = -(realbn - bn + NIADDR - i);
271 else
272 metalbn = -(-realbn - bn + NIADDR - i);
273
274 /*
275 * At each iteration, off is the offset into the bap array which is
276 * an array of disk addresses at the current level of indirection.
277 * The logical block number and the offset in that block are stored
278 * into the argument array.
279 */
280 ap->in_lbn = metalbn;
281 ap->in_off = off = NIADDR - i;
282 ap->in_exists = 0;
283 ap++;
284 for (++numlevels; i <= NIADDR; i++) {
285 /* If searching for a meta-data block, quit when found. */
286 if (metalbn == realbn)
287 break;
288
289 blockcnt /= MNINDIR(ump);
290 off = (bn / blockcnt) % MNINDIR(ump);
291
292 ++numlevels;
293 ap->in_lbn = metalbn;
294 ap->in_off = off;
295 ap->in_exists = 0;
296 ++ap;
297
298 metalbn -= -1 + off * blockcnt;
299 }
300 #ifdef DIAGNOSTIC
301 if (realbn < 0 && metalbn != realbn) {
302 panic("ufs_getlbns: indirect block %d not found", realbn);
303 }
304 #endif
305 if (nump)
306 *nump = numlevels;
307 return (0);
308 }