2 * Copyright (c) 1989, 1991, 1993
3 * The Regents of the University of California. All rights reserved.
4 * (c) UNIX System Laboratories, Inc.
5 * All or some portions of this file are derived from material licensed
6 * to the University of California by American Telephone and Telegraph
7 * Co. or Unix System Laboratories, Inc. and are reproduced herein with
8 * the permission of UNIX System Laboratories, Inc.
10 * Redistribution and use in source and binary forms, with or without
11 * modification, are permitted provided that the following conditions
13 * 1. Redistributions of source code must retain the above copyright
14 * notice, this list of conditions and the following disclaimer.
15 * 2. Redistributions in binary form must reproduce the above copyright
16 * notice, this list of conditions and the following disclaimer in the
17 * documentation and/or other materials provided with the distribution.
18 * 3. All advertising materials mentioning features or use of this software
19 * must display the following acknowledgement:
20 * This product includes software developed by the University of
21 * California, Berkeley and its contributors.
22 * 4. Neither the name of the University nor the names of its contributors
23 * may be used to endorse or promote products derived from this software
24 * without specific prior written permission.
26 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
27 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
28 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
29 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
30 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
31 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
32 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
33 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
34 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
35 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
38 * @(#)ufs_bmap.c 8.7 (Berkeley) 3/21/95
39 * $FreeBSD: src/sys/ufs/ufs/ufs_bmap.c,v 1.34.2.1 2000/03/17 10:12:14 ps Exp $
40 * $DragonFly: src/sys/vfs/ufs/ufs_bmap.c,v 1.12 2006/04/30 17:22:18 dillon Exp $
43 #include <sys/param.h>
44 #include <sys/systm.h>
47 #include <sys/vnode.h>
48 #include <sys/mount.h>
49 #include <sys/resourcevar.h>
55 #include "ufs_extern.h"
59 * Bmap converts the logical block number of a file to its physical block
60 * number on the disk. The conversion is done by using the logical block
61 * number to index into the array of block pointers described by the dinode.
63 * BMAP must return the contiguous before and after run in bytes, inclusive
64 * of the returned block.
66 * ufs_bmap(struct vnode *a_vp, off_t a_loffset, struct vnode **a_vpp,
67 * off_t *a_doffsetp, int *a_runp, int *a_runb)
70 ufs_bmap(struct vop_bmap_args *ap)
78 * Check for underlying vnode requests and ensure that logical
79 * to physical mapping is requested.
81 if (ap->a_vpp != NULL)
82 *ap->a_vpp = VTOI(ap->a_vp)->i_devvp;
83 if (ap->a_doffsetp == NULL)
86 fs = VTOI(ap->a_vp)->i_fs;
87 KKASSERT(((int)ap->a_loffset & ((1 << fs->fs_bshift) - 1)) == 0);
88 lbn = ap->a_loffset >> fs->fs_bshift;
90 error = ufs_bmaparray(ap->a_vp, lbn, &dbn, NULL, NULL,
91 ap->a_runp, ap->a_runb);
93 if (error || dbn == (ufs_daddr_t)-1) {
94 *ap->a_doffsetp = NOOFFSET;
96 *ap->a_doffsetp = dbtodoff(fs, dbn);
98 *ap->a_runp = (*ap->a_runp + 1) << fs->fs_bshift;
100 *ap->a_runb = *ap->a_runb << fs->fs_bshift;
106 * Indirect blocks are now on the vnode for the file. They are given negative
107 * logical block numbers. Indirect blocks are addressed by the negative
108 * address of the first data block to which they point. Double indirect blocks
109 * are addressed by one less than the address of the first indirect block to
110 * which they point. Triple indirect blocks are addressed by one less than
111 * the address of the first double indirect block to which they point.
113 * ufs_bmaparray does the bmap conversion, and if requested returns the
114 * array of logical blocks which must be traversed to get to a block.
115 * Each entry contains the offset into that block that gets you to the
116 * next block and the disk address of the block (if it is assigned).
119 ufs_bmaparray(struct vnode *vp, ufs_daddr_t bn, ufs_daddr_t *bnp,
120 struct indir *ap, int *nump, int *runp, int *runb)
124 struct ufsmount *ump;
128 struct indir a[NIADDR+1], *xap;
131 int error, maxrun, num;
136 devvp = ump->um_devvp;
139 if ((ap != NULL && nump == NULL) || (ap == NULL && nump != NULL))
140 panic("ufs_bmaparray: invalid arguments");
151 maxrun = mp->mnt_iosize_max / mp->mnt_stat.f_iosize - 1;
153 xap = ap == NULL ? a : ap;
156 error = ufs_getlbns(vp, bn, xap, nump);
162 *bnp = blkptrtodb(ump, ip->i_db[bn]);
167 for (++bn; bn < NDADDR && *runp < maxrun &&
168 is_sequential(ump, ip->i_db[bn - 1], ip->i_db[bn]);
171 if (runb && (bn > 0)) {
172 for (--bn; (bn >= 0) && (*runb < maxrun) &&
173 is_sequential(ump, ip->i_db[bn],
182 /* Get disk address out of indirect block array */
183 daddr = ip->i_ib[xap->in_off];
185 for (bp = NULL, ++xap; --num; ++xap) {
187 * Exit the loop if there is no disk address assigned yet and
188 * the indirect block isn't in the cache, or if we were
189 * looking for an indirect block and we've found it.
192 metalbn = xap->in_lbn;
193 if ((daddr == 0 && !findblk(vp, dbtodoff(fs, metalbn))) || metalbn == bn)
196 * If we get here, we've either got the block in the cache
197 * or we have a disk address for it, go fetch it.
203 bp = getblk(vp, lblktodoff(fs, metalbn),
204 mp->mnt_stat.f_iosize, 0, 0);
205 if ((bp->b_flags & B_CACHE) == 0) {
208 panic("ufs_bmaparray: indirect block not in cache");
210 bp->b_bio2.bio_offset = fsbtodoff(fs, daddr);
211 bp->b_flags &= ~(B_INVAL|B_ERROR);
212 bp->b_cmd = BUF_CMD_READ;
213 vfs_busy_pages(bp->b_vp, bp);
214 vn_strategy(bp->b_vp, &bp->b_bio1);
222 daddr = ((ufs_daddr_t *)bp->b_data)[xap->in_off];
223 if (num == 1 && daddr && runp) {
224 for (bn = xap->in_off + 1;
225 bn < MNINDIR(ump) && *runp < maxrun &&
227 ((ufs_daddr_t *)bp->b_data)[bn - 1],
228 ((ufs_daddr_t *)bp->b_data)[bn]);
232 for(--bn; bn >= 0 && *runb < maxrun &&
233 is_sequential(ump, ((daddr_t *)bp->b_data)[bn],
234 ((daddr_t *)bp->b_data)[bn+1]);
242 daddr = blkptrtodb(ump, daddr);
243 *bnp = daddr == 0 ? -1 : daddr;
248 * Create an array of logical block number/offset pairs which represent the
249 * path of indirect blocks required to access a data block. The first "pair"
250 * contains the logical block number of the appropriate single, double or
251 * triple indirect block and the offset into the inode indirect block array.
252 * Note, the logical block number of the inode single/double/triple indirect
253 * block appears twice in the array, once with the offset into the i_ib and
254 * once with the offset into the page itself.
257 ufs_getlbns(struct vnode *vp, ufs_daddr_t bn, struct indir *ap, int *nump)
259 long blockcnt, metalbn, realbn;
260 struct ufsmount *ump;
261 int i, numlevels, off;
264 ump = VFSTOUFS(vp->v_mount);
272 /* The first NDADDR blocks are direct blocks. */
277 * Determine the number of levels of indirection. After this loop
278 * is done, blockcnt indicates the number of data blocks possible
279 * at the previous level of indirection, and NIADDR - i is the number
280 * of levels of indirection needed to locate the requested block.
282 for (blockcnt = 1, i = NIADDR, bn -= NDADDR;; i--, bn -= blockcnt) {
286 * Use int64_t's here to avoid overflow for triple indirect
287 * blocks when longs have 32 bits and the block size is more
290 qblockcnt = (int64_t)blockcnt * MNINDIR(ump);
293 blockcnt = qblockcnt;
296 /* Calculate the address of the first meta-block. */
298 metalbn = -(realbn - bn + NIADDR - i);
300 metalbn = -(-realbn - bn + NIADDR - i);
303 * At each iteration, off is the offset into the bap array which is
304 * an array of disk addresses at the current level of indirection.
305 * The logical block number and the offset in that block are stored
306 * into the argument array.
308 ap->in_lbn = metalbn;
309 ap->in_off = off = NIADDR - i;
312 for (++numlevels; i <= NIADDR; i++) {
313 /* If searching for a meta-data block, quit when found. */
314 if (metalbn == realbn)
317 off = (bn / blockcnt) % MNINDIR(ump);
320 ap->in_lbn = metalbn;
325 metalbn -= -1 + off * blockcnt;
326 blockcnt /= MNINDIR(ump);