/* $NetBSD: clri.c,v 1.15 2003/04/02 10:39:23 fvdl Exp $ */ /* * Copyright (c) 1990, 1993 * The Regents of the University of California. All rights reserved. * * This code is derived from software contributed to Berkeley by * Rich $alz of BBN Inc. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. All advertising materials mentioning features or use of this software * must display the following acknowledgement: * This product includes software developed by the University of * California, Berkeley and its contributors. * 4. Neither the name of the University nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. */ #include #ifndef lint __COPYRIGHT("@(#) Copyright (c) 1990, 1993\n\ The Regents of the University of California. All rights reserved.\n"); #endif /* not lint */ #ifndef lint #if 0 static char sccsid[] = "@(#)clri.c 8.3 (Berkeley) 4/28/95"; #else __RCSID("$NetBSD: clri.c,v 1.15 2003/04/02 10:39:23 fvdl Exp $"); #endif #endif /* not lint */ #include #include #include #include #include #include #include #include #include #include #include #include #include /* * Possible superblock locations ordered from most to least likely. */ static off_t sblock_try[] = SBLOCKSEARCH; off_t sblockloc; int main __P((int, char *[])); int main(argc, argv) int argc; char *argv[]; { struct fs *sbp; struct ufs1_dinode *ip1; struct ufs2_dinode *ip2; int fd; char *ibuf[MAXBSIZE]; int32_t generation; off_t offset; size_t bsize; int inonum; char *fs, sblock[SBLOCKSIZE]; int needswap = 0, is_ufs2 = 0; int i, imax; if (argc < 3) { (void)fprintf(stderr, "usage: clri filesystem inode ...\n"); exit(1); } fs = *++argv; /* get the superblock. */ if ((fd = open(fs, O_RDWR, 0)) < 0) err(1, "%s", fs); for (i = 0; i < sblock_try[i] != -1; i++) { if (lseek(fd, sblock_try[i], SEEK_SET) < 0) err(1, "%s", fs); if (read(fd, sblock, sizeof(sblock)) != sizeof(sblock)) errx(1, "%s: can't read superblock", fs); sbp = (struct fs *)sblock; switch(sbp->fs_magic) { case FS_UFS2_MAGIC: is_ufs2 = 1; /*FALLTHROUGH*/ case FS_UFS1_MAGIC: goto found; case FS_UFS2_MAGIC_SWAPPED: is_ufs2 = 1; /*FALLTHROUGH*/ case FS_UFS1_MAGIC_SWAPPED: needswap = 1; goto found; default: continue; } } errx(1, "%s: can't find superblock", fs); found: sblockloc = sblock_try[i]; /* check that inode numbers are valid */ imax = ufs_rw32(sbp->fs_ncg, needswap) * ufs_rw32(sbp->fs_ipg, needswap); for (i = 1; i < (argc - 1); i++) if (atoi(argv[i]) <= 0 || atoi(argv[i]) >= imax) errx(1, "%s is not a valid inode number", argv[i]); /* delete clean flag in the superblok */ sbp->fs_clean = ufs_rw32( ufs_rw32(sbp->fs_clean, needswap) << 1, needswap); if (lseek(fd, sblockloc, SEEK_SET) < 0) err(1, "%s", fs); if (write(fd, sblock, sizeof(sblock)) != sizeof(sblock)) errx(1, "%s: can't rewrite superblock", fs); (void)fsync(fd); if (needswap) ffs_sb_swap(sbp, sbp); bsize = sbp->fs_bsize; /* remaining arguments are inode numbers. */ while (*++argv) { /* get the inode number. */ inonum = atoi(*argv); (void)printf("clearing %d\n", inonum); /* read in the appropriate block. */ offset = ino_to_fsba(sbp, inonum); /* inode to fs blk */ offset = fsbtodb(sbp, offset); /* fs blk disk blk */ offset *= DEV_BSIZE; /* disk blk to bytes */ /* seek and read the block */ if (lseek(fd, offset, SEEK_SET) < 0) err(1, "%s", fs); if (read(fd, ibuf, bsize) != bsize) err(1, "%s", fs); /* get the inode within the block. */ if (is_ufs2) { ip2 = &((struct ufs2_dinode *)ibuf) [ino_to_fsbo(sbp, inonum)]; /* clear the inode, and bump the generation count. */ generation = ip2->di_gen + 1; memset(ip2, 0, sizeof(*ip2)); ip2->di_gen = generation; } else { ip1 = &((struct ufs1_dinode *)ibuf) [ino_to_fsbo(sbp, inonum)]; /* clear the inode, and bump the generation count. */ generation = ip1->di_gen + 1; memset(ip1, 0, sizeof(*ip1)); ip1->di_gen = generation; } /* backup and write the block */ if (lseek(fd, offset, SEEK_SET) < 0) err(1, "%s", fs); if (write(fd, ibuf, bsize) != bsize) err(1, "%s", fs); (void)fsync(fd); } (void)close(fd); exit(0); }