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. Neither the name of the University nor the names of its contributors
19 * may be used to endorse or promote products derived from this software
20 * without specific prior written permission.
22 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
23 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
24 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
25 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
26 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
27 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
28 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
29 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
30 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
31 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
34 * @(#)ufs_bmap.c 8.7 (Berkeley) 3/21/95
35 * $FreeBSD: src/sys/ufs/ufs/ufs_bmap.c,v 1.34.2.1 2000/03/17 10:12:14 ps Exp $
38 #include <sys/param.h>
39 #include <sys/systm.h>
42 #include <sys/vnode.h>
43 #include <sys/mount.h>
44 #include <sys/resourcevar.h>
50 #include "ext2_fs_sb.h"
51 #include "ext2_mount.h"
52 #include "ext2_extern.h"
55 static int ext2_bmaparray(struct vnode *vp, ext2_daddr_t bn,
56 ext2_daddr_t *bnp, struct indir *ap, int *nump,
57 int *runp, int *runb);
60 * Bmap converts the logical block number of a file to its physical block
61 * number on the disk. The conversion is done by using the logical block
62 * number to index into the array of block pointers described by the dinode.
64 * BMAP must return the contiguous before and after run in bytes, inclusive
65 * of the returned block.
67 * ext2_bmap(struct vnode *a_vp, off_t a_loffset,
68 * off_t *a_doffsetp, int *a_runp, int *a_runb)
71 ext2_bmap(struct vop_bmap_args *ap)
73 struct ext2_sb_info *fs;
79 * Check for underlying vnode requests and ensure that logical
80 * to physical mapping is requested.
82 if (ap->a_doffsetp == NULL)
85 fs = VTOI(ap->a_vp)->i_e2fs;
86 KKASSERT(((int)ap->a_loffset & ((1 << fs->s_bshift) - 1)) == 0);
87 lbn = ap->a_loffset >> fs->s_bshift;
89 error = ext2_bmaparray(ap->a_vp, lbn, &dbn, NULL, NULL,
90 ap->a_runp, ap->a_runb);
92 if (error || dbn == (ext2_daddr_t)-1) {
93 *ap->a_doffsetp = NOOFFSET;
95 *ap->a_doffsetp = dbtodoff(fs, dbn);
97 *ap->a_runp = (*ap->a_runp + 1) << fs->s_bshift;
99 *ap->a_runb = *ap->a_runb << fs->s_bshift;
105 * Indirect blocks are now on the vnode for the file. They are given negative
106 * logical block numbers. Indirect blocks are addressed by the negative
107 * address of the first data block to which they point. Double indirect blocks
108 * are addressed by one less than the address of the first indirect block to
109 * which they point. Triple indirect blocks are addressed by one less than
110 * the address of the first double indirect block to which they point.
112 * ext2_bmaparray does the bmap conversion, and if requested returns the
113 * array of logical blocks which must be traversed to get to a block.
114 * Each entry contains the offset into that block that gets you to the
115 * next block and the disk address of the block (if it is assigned).
119 ext2_bmaparray(struct vnode *vp, ext2_daddr_t bn, ext2_daddr_t *bnp,
120 struct indir *ap, int *nump, int *runp, int *runb)
124 struct ext2_mount *ump;
126 struct ext2_sb_info *fs;
127 struct indir a[NIADDR+1], *xap;
130 int error, maxrun, num;
137 if ((ap != NULL && nump == NULL) || (ap == NULL && nump != NULL))
138 panic("ext2_bmaparray: invalid arguments");
149 maxrun = mp->mnt_iosize_max / mp->mnt_stat.f_iosize - 1;
151 xap = ap == NULL ? a : ap;
154 error = ext2_getlbns(vp, bn, xap, nump);
160 *bnp = blkptrtodb(ump, ip->i_db[bn]);
165 for (++bn; bn < NDADDR && *runp < maxrun &&
166 is_sequential(ump, ip->i_db[bn - 1], ip->i_db[bn]);
169 if (runb && (bn > 0)) {
170 for (--bn; (bn >= 0) && (*runb < maxrun) &&
171 is_sequential(ump, ip->i_db[bn],
180 /* Get disk address out of indirect block array */
181 daddr = ip->i_ib[xap->in_off];
183 for (bp = NULL, ++xap; --num; ++xap) {
185 * Exit the loop if there is no disk address assigned yet and
186 * the indirect block isn't in the cache, or if we were
187 * looking for an indirect block and we've found it.
190 metalbn = xap->in_lbn;
192 !findblk(vp, dbtodoff(fs, metalbn), FINDBLK_TEST)) ||
197 * If we get here, we've either got the block in the cache
198 * 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("ext2_bmaparray: indirect block not in cache");
211 * This runs through ext2_strategy using bio2 to
212 * cache the disk offset, then comes back through
213 * bio1. So we want to wait on bio1
215 bp->b_bio1.bio_done = biodone_sync;
216 bp->b_bio1.bio_flags |= BIO_SYNC;
217 bp->b_bio2.bio_offset = fsbtodoff(fs, daddr);
218 bp->b_flags &= ~(B_INVAL|B_ERROR);
219 bp->b_cmd = BUF_CMD_READ;
220 vfs_busy_pages(bp->b_vp, bp);
221 vn_strategy(bp->b_vp, &bp->b_bio1);
222 error = biowait(&bp->b_bio1, "biord");
229 daddr = ((ext2_daddr_t *)bp->b_data)[xap->in_off];
230 if (num == 1 && daddr && runp) {
231 for (bn = xap->in_off + 1;
232 bn < MNINDIR(ump) && *runp < maxrun &&
234 ((ext2_daddr_t *)bp->b_data)[bn - 1],
235 ((ext2_daddr_t *)bp->b_data)[bn]);
239 for(--bn; bn >= 0 && *runb < maxrun &&
240 is_sequential(ump, ((daddr_t *)bp->b_data)[bn],
241 ((daddr_t *)bp->b_data)[bn+1]);
249 daddr = blkptrtodb(ump, daddr);
250 *bnp = daddr == 0 ? -1 : daddr;
255 * Create an array of logical block number/offset pairs which represent the
256 * path of indirect blocks required to access a data block. The first "pair"
257 * contains the logical block number of the appropriate single, double or
258 * triple indirect block and the offset into the inode indirect block array.
259 * Note, the logical block number of the inode single/double/triple indirect
260 * block appears twice in the array, once with the offset into the i_ib and
261 * once with the offset into the page itself.
264 ext2_getlbns(struct vnode *vp, ext2_daddr_t bn, struct indir *ap, int *nump)
266 long blockcnt, metalbn, realbn;
267 struct ext2_mount *ump;
268 int i, numlevels, off;
271 ump = VFSTOEXT2(vp->v_mount);
279 /* The first NDADDR blocks are direct blocks. */
284 * Determine the number of levels of indirection. After this loop
285 * is done, blockcnt indicates the number of data blocks possible
286 * at the previous level of indirection, and NIADDR - i is the number
287 * of levels of indirection needed to locate the requested block.
289 for (blockcnt = 1, i = NIADDR, bn -= NDADDR;; i--, bn -= blockcnt) {
293 * Use int64_t's here to avoid overflow for triple indirect
294 * blocks when longs have 32 bits and the block size is more
297 qblockcnt = (int64_t)blockcnt * MNINDIR(ump);
300 blockcnt = qblockcnt;
303 /* Calculate the address of the first meta-block. */
305 metalbn = -(realbn - bn + NIADDR - i);
307 metalbn = -(-realbn - bn + NIADDR - i);
310 * At each iteration, off is the offset into the bap array which is
311 * an array of disk addresses at the current level of indirection.
312 * The logical block number and the offset in that block are stored
313 * into the argument array.
315 ap->in_lbn = metalbn;
316 ap->in_off = off = NIADDR - i;
318 for (++numlevels; i <= NIADDR; i++) {
319 /* If searching for a meta-data block, quit when found. */
320 if (metalbn == realbn)
323 off = (bn / blockcnt) % MNINDIR(ump);
326 ap->in_lbn = metalbn;
330 metalbn -= -1 + off * blockcnt;
331 blockcnt /= MNINDIR(ump);