freebsd-skq/sbin/fsck/inode.c
Kirk McKusick e50342e665 Yesterday I had to fix a badly broken disk, and found that fsck kept dying:
DIR I=64512 CONNECTED. PARENT WAS I=4032
  fsck: cannot find inode 995904

fsdb found the inodes with no problem:

  fsdb (inum: 64512)> inode 995904
  current inode: directory
  I=995904 MODE=40777 SIZE=512
        MTIME=Feb 14 15:27:07 2000 [0 nsec]
        CTIME=Feb 14 15:27:07 2000 [0 nsec]
        ATIME=Feb 24 10:31:58 2000 [0 nsec]
  OWNER=nobody GRP=nobody LINKCNT=4 FLAGS=0 BLKCNT=2 GEN=38a41386
  Direct blocks: 8094568 0 0 0 0 0 0 0 0 0 0 0
  Indirect blocks:  0 0 0

The problem turns out to be a program logic error in fsck.  It stores
directory inodes internally in hash lists, using the number of
directories to form the hash key:

        inpp = &inphead[inumber % numdirs];

Elsewhere, however, it increments numdirs when it finds unattached
directories.  I've made the following fix, which solved the problem in
the case in hand.

Submitted by:	Greg Lehey <grog@lemis.com>
Reviewed by:	Matthew Dillon <dillon@apollo.backplane.com>
Approved by:	Kirk McKusick <mckusick@mckusick.com>
2000-02-28 20:02:41 +00:00

667 lines
16 KiB
C

/*
* Copyright (c) 1980, 1986, 1993
* The Regents of the University of California. All rights reserved.
*
* 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.
*/
#ifndef lint
#if 0
static const char sccsid[] = "@(#)inode.c 8.8 (Berkeley) 4/28/95";
#endif
static const char rcsid[] =
"$FreeBSD$";
#endif /* not lint */
#include <sys/param.h>
#include <sys/time.h>
#include <ufs/ufs/dinode.h>
#include <ufs/ufs/dir.h>
#include <ufs/ffs/fs.h>
#include <err.h>
#include <pwd.h>
#include <string.h>
#include "fsck.h"
static ino_t startinum;
static int iblock __P((struct inodesc *, long ilevel, quad_t isize));
int
ckinode(dp, idesc)
struct dinode *dp;
register struct inodesc *idesc;
{
ufs_daddr_t *ap;
int ret;
long n, ndb, offset;
struct dinode dino;
quad_t remsize, sizepb;
mode_t mode;
char pathbuf[MAXPATHLEN + 1];
if (idesc->id_fix != IGNORE)
idesc->id_fix = DONTKNOW;
idesc->id_entryno = 0;
idesc->id_filesize = dp->di_size;
mode = dp->di_mode & IFMT;
if (mode == IFBLK || mode == IFCHR || (mode == IFLNK &&
dp->di_size < (unsigned)sblock.fs_maxsymlinklen))
return (KEEPON);
dino = *dp;
ndb = howmany(dino.di_size, sblock.fs_bsize);
for (ap = &dino.di_db[0]; ap < &dino.di_db[NDADDR]; ap++) {
if (--ndb == 0 && (offset = blkoff(&sblock, dino.di_size)) != 0)
idesc->id_numfrags =
numfrags(&sblock, fragroundup(&sblock, offset));
else
idesc->id_numfrags = sblock.fs_frag;
if (*ap == 0) {
if (idesc->id_type == DATA && ndb >= 0) {
/* An empty block in a directory XXX */
getpathname(pathbuf, idesc->id_number,
idesc->id_number);
pfatal("DIRECTORY %s: CONTAINS EMPTY BLOCKS",
pathbuf);
if (reply("ADJUST LENGTH") == 1) {
dp = ginode(idesc->id_number);
dp->di_size = (ap - &dino.di_db[0]) *
sblock.fs_bsize;
printf(
"YOU MUST RERUN FSCK AFTERWARDS\n");
rerun = 1;
inodirty();
}
}
continue;
}
idesc->id_blkno = *ap;
if (idesc->id_type == ADDR)
ret = (*idesc->id_func)(idesc);
else
ret = dirscan(idesc);
if (ret & STOP)
return (ret);
}
idesc->id_numfrags = sblock.fs_frag;
remsize = dino.di_size - sblock.fs_bsize * NDADDR;
sizepb = sblock.fs_bsize;
for (ap = &dino.di_ib[0], n = 1; n <= NIADDR; ap++, n++) {
if (*ap) {
idesc->id_blkno = *ap;
ret = iblock(idesc, n, remsize);
if (ret & STOP)
return (ret);
} else {
if (idesc->id_type == DATA && remsize > 0) {
/* An empty block in a directory XXX */
getpathname(pathbuf, idesc->id_number,
idesc->id_number);
pfatal("DIRECTORY %s: CONTAINS EMPTY BLOCKS",
pathbuf);
if (reply("ADJUST LENGTH") == 1) {
dp = ginode(idesc->id_number);
dp->di_size -= remsize;
remsize = 0;
printf(
"YOU MUST RERUN FSCK AFTERWARDS\n");
rerun = 1;
inodirty();
break;
}
}
}
sizepb *= NINDIR(&sblock);
remsize -= sizepb;
}
return (KEEPON);
}
static int
iblock(idesc, ilevel, isize)
struct inodesc *idesc;
long ilevel;
quad_t isize;
{
ufs_daddr_t *ap;
ufs_daddr_t *aplim;
struct bufarea *bp;
int i, n, (*func)(), nif;
quad_t sizepb;
char buf[BUFSIZ];
char pathbuf[MAXPATHLEN + 1];
struct dinode *dp;
if (idesc->id_type == ADDR) {
func = idesc->id_func;
if (((n = (*func)(idesc)) & KEEPON) == 0)
return (n);
} else
func = dirscan;
if (chkrange(idesc->id_blkno, idesc->id_numfrags))
return (SKIP);
bp = getdatablk(idesc->id_blkno, sblock.fs_bsize);
ilevel--;
for (sizepb = sblock.fs_bsize, i = 0; i < ilevel; i++)
sizepb *= NINDIR(&sblock);
nif = howmany(isize , sizepb);
if (nif > NINDIR(&sblock))
nif = NINDIR(&sblock);
if (idesc->id_func == pass1check && nif < NINDIR(&sblock)) {
aplim = &bp->b_un.b_indir[NINDIR(&sblock)];
for (ap = &bp->b_un.b_indir[nif]; ap < aplim; ap++) {
if (*ap == 0)
continue;
(void)sprintf(buf, "PARTIALLY TRUNCATED INODE I=%lu",
(u_long)idesc->id_number);
if (dofix(idesc, buf)) {
*ap = 0;
dirty(bp);
}
}
flush(fswritefd, bp);
}
aplim = &bp->b_un.b_indir[nif];
for (ap = bp->b_un.b_indir; ap < aplim; ap++) {
if (*ap) {
idesc->id_blkno = *ap;
if (ilevel == 0)
n = (*func)(idesc);
else
n = iblock(idesc, ilevel, isize);
if (n & STOP) {
bp->b_flags &= ~B_INUSE;
return (n);
}
} else {
if (idesc->id_type == DATA && isize > 0) {
/* An empty block in a directory XXX */
getpathname(pathbuf, idesc->id_number,
idesc->id_number);
pfatal("DIRECTORY %s: CONTAINS EMPTY BLOCKS",
pathbuf);
if (reply("ADJUST LENGTH") == 1) {
dp = ginode(idesc->id_number);
dp->di_size -= isize;
isize = 0;
printf(
"YOU MUST RERUN FSCK AFTERWARDS\n");
rerun = 1;
inodirty();
bp->b_flags &= ~B_INUSE;
return(STOP);
}
}
}
isize -= sizepb;
}
bp->b_flags &= ~B_INUSE;
return (KEEPON);
}
/*
* Check that a block in a legal block number.
* Return 0 if in range, 1 if out of range.
*/
int
chkrange(blk, cnt)
ufs_daddr_t blk;
int cnt;
{
register int c;
if (cnt <= 0 || blk <= 0 || blk > maxfsblock ||
cnt - 1 > maxfsblock - blk)
return (1);
if (cnt > sblock.fs_frag ||
fragnum(&sblock, blk) + cnt > sblock.fs_frag) {
if (debug)
printf("bad size: blk %ld, offset %ld, size %ld\n",
blk, fragnum(&sblock, blk), cnt);
return (1);
}
c = dtog(&sblock, blk);
if (blk < cgdmin(&sblock, c)) {
if ((blk + cnt) > cgsblock(&sblock, c)) {
if (debug) {
printf("blk %ld < cgdmin %ld;",
(long)blk, (long)cgdmin(&sblock, c));
printf(" blk + cnt %ld > cgsbase %ld\n",
(long)(blk + cnt),
(long)cgsblock(&sblock, c));
}
return (1);
}
} else {
if ((blk + cnt) > cgbase(&sblock, c+1)) {
if (debug) {
printf("blk %ld >= cgdmin %ld;",
(long)blk, (long)cgdmin(&sblock, c));
printf(" blk + cnt %ld > sblock.fs_fpg %ld\n",
(long)(blk + cnt), (long)sblock.fs_fpg);
}
return (1);
}
}
return (0);
}
/*
* General purpose interface for reading inodes.
*/
struct dinode *
ginode(inumber)
ino_t inumber;
{
ufs_daddr_t iblk;
if (inumber < ROOTINO || inumber > maxino)
errx(EEXIT, "bad inode number %d to ginode", inumber);
if (startinum == 0 ||
inumber < startinum || inumber >= startinum + INOPB(&sblock)) {
iblk = ino_to_fsba(&sblock, inumber);
if (pbp != 0)
pbp->b_flags &= ~B_INUSE;
pbp = getdatablk(iblk, sblock.fs_bsize);
startinum = (inumber / INOPB(&sblock)) * INOPB(&sblock);
}
return (&pbp->b_un.b_dinode[inumber % INOPB(&sblock)]);
}
/*
* Special purpose version of ginode used to optimize first pass
* over all the inodes in numerical order.
*/
ino_t nextino, lastinum;
long readcnt, readpercg, fullcnt, inobufsize, partialcnt, partialsize;
struct dinode *inodebuf;
struct dinode *
getnextinode(inumber)
ino_t inumber;
{
long size;
ufs_daddr_t dblk;
static struct dinode *dp;
if (inumber != nextino++ || inumber > maxino)
errx(EEXIT, "bad inode number %d to nextinode", inumber);
if (inumber >= lastinum) {
readcnt++;
dblk = fsbtodb(&sblock, ino_to_fsba(&sblock, lastinum));
if (readcnt % readpercg == 0) {
size = partialsize;
lastinum += partialcnt;
} else {
size = inobufsize;
lastinum += fullcnt;
}
/*
* If bread returns an error, it will already have zeroed
* out the buffer, so we do not need to do so here.
*/
(void)bread(fsreadfd, (char *)inodebuf, dblk, size);
dp = inodebuf;
}
return (dp++);
}
void
setinodebuf(inum)
ino_t inum;
{
if (inum % sblock.fs_ipg != 0)
errx(EEXIT, "bad inode number %d to setinodebuf", inum);
startinum = 0;
nextino = inum;
lastinum = inum;
readcnt = 0;
if (inodebuf != NULL)
return;
inobufsize = blkroundup(&sblock, INOBUFSIZE);
fullcnt = inobufsize / sizeof(struct dinode);
readpercg = sblock.fs_ipg / fullcnt;
partialcnt = sblock.fs_ipg % fullcnt;
partialsize = partialcnt * sizeof(struct dinode);
if (partialcnt != 0) {
readpercg++;
} else {
partialcnt = fullcnt;
partialsize = inobufsize;
}
if ((inodebuf = (struct dinode *)malloc((unsigned)inobufsize)) == NULL)
errx(EEXIT, "cannot allocate space for inode buffer");
}
void
freeinodebuf()
{
if (inodebuf != NULL)
free((char *)inodebuf);
inodebuf = NULL;
}
/*
* Routines to maintain information about directory inodes.
* This is built during the first pass and used during the
* second and third passes.
*
* Enter inodes into the cache.
*/
void
cacheino(dp, inumber)
register struct dinode *dp;
ino_t inumber;
{
register struct inoinfo *inp;
struct inoinfo **inpp;
int blks;
blks = howmany(dp->di_size, sblock.fs_bsize);
if (blks > NDADDR)
blks = NDADDR + NIADDR;
inp = (struct inoinfo *)
malloc(sizeof(*inp) + (blks - 1) * sizeof(ufs_daddr_t));
if (inp == NULL)
errx(EEXIT, "cannot increase directory list");
inpp = &inphead[inumber % dirhash];
inp->i_nexthash = *inpp;
*inpp = inp;
inp->i_parent = inumber == ROOTINO ? ROOTINO : (ino_t)0;
inp->i_dotdot = (ino_t)0;
inp->i_number = inumber;
inp->i_isize = dp->di_size;
inp->i_numblks = blks * sizeof(ufs_daddr_t);
memmove(&inp->i_blks[0], &dp->di_db[0], (size_t)inp->i_numblks);
if (inplast == listmax) {
listmax += 100;
inpsort = (struct inoinfo **)realloc((char *)inpsort,
(unsigned)listmax * sizeof(struct inoinfo *));
if (inpsort == NULL)
errx(EEXIT, "cannot increase directory list");
}
inpsort[inplast++] = inp;
}
/*
* Look up an inode cache structure.
*/
struct inoinfo *
getinoinfo(inumber)
ino_t inumber;
{
register struct inoinfo *inp;
for (inp = inphead[inumber % dirhash]; inp; inp = inp->i_nexthash) {
if (inp->i_number != inumber)
continue;
return (inp);
}
errx(EEXIT, "cannot find inode %d", inumber);
return ((struct inoinfo *)0);
}
/*
* Clean up all the inode cache structure.
*/
void
inocleanup()
{
register struct inoinfo **inpp;
if (inphead == NULL)
return;
for (inpp = &inpsort[inplast - 1]; inpp >= inpsort; inpp--)
free((char *)(*inpp));
free((char *)inphead);
free((char *)inpsort);
inphead = inpsort = NULL;
}
void
inodirty()
{
dirty(pbp);
}
void
clri(idesc, type, flag)
register struct inodesc *idesc;
char *type;
int flag;
{
register struct dinode *dp;
dp = ginode(idesc->id_number);
if (flag == 1) {
pwarn("%s %s", type,
(dp->di_mode & IFMT) == IFDIR ? "DIR" : "FILE");
pinode(idesc->id_number);
}
if (preen || reply("CLEAR") == 1) {
if (preen)
printf(" (CLEARED)\n");
n_files--;
(void)ckinode(dp, idesc);
clearinode(dp);
inoinfo(idesc->id_number)->ino_state = USTATE;
inodirty();
}
}
int
findname(idesc)
struct inodesc *idesc;
{
register struct direct *dirp = idesc->id_dirp;
if (dirp->d_ino != idesc->id_parent || idesc->id_entryno < 2) {
idesc->id_entryno++;
return (KEEPON);
}
memmove(idesc->id_name, dirp->d_name, (size_t)dirp->d_namlen + 1);
return (STOP|FOUND);
}
int
findino(idesc)
struct inodesc *idesc;
{
register struct direct *dirp = idesc->id_dirp;
if (dirp->d_ino == 0)
return (KEEPON);
if (strcmp(dirp->d_name, idesc->id_name) == 0 &&
dirp->d_ino >= ROOTINO && dirp->d_ino <= maxino) {
idesc->id_parent = dirp->d_ino;
return (STOP|FOUND);
}
return (KEEPON);
}
int
clearentry(idesc)
struct inodesc *idesc;
{
register struct direct *dirp = idesc->id_dirp;
if (dirp->d_ino != idesc->id_parent || idesc->id_entryno < 2) {
idesc->id_entryno++;
return (KEEPON);
}
dirp->d_ino = 0;
return (STOP|FOUND|ALTERED);
}
void
pinode(ino)
ino_t ino;
{
register struct dinode *dp;
register char *p;
struct passwd *pw;
time_t t;
printf(" I=%lu ", (u_long)ino);
if (ino < ROOTINO || ino > maxino)
return;
dp = ginode(ino);
printf(" OWNER=");
if ((pw = getpwuid((int)dp->di_uid)) != 0)
printf("%s ", pw->pw_name);
else
printf("%u ", (unsigned)dp->di_uid);
printf("MODE=%o\n", dp->di_mode);
if (preen)
printf("%s: ", cdevname);
printf("SIZE=%qu ", dp->di_size);
t = dp->di_mtime;
p = ctime(&t);
printf("MTIME=%12.12s %4.4s ", &p[4], &p[20]);
}
void
blkerror(ino, type, blk)
ino_t ino;
char *type;
ufs_daddr_t blk;
{
pfatal("%ld %s I=%lu", blk, type, ino);
printf("\n");
switch (inoinfo(ino)->ino_state) {
case FSTATE:
inoinfo(ino)->ino_state = FCLEAR;
return;
case DSTATE:
inoinfo(ino)->ino_state = DCLEAR;
return;
case FCLEAR:
case DCLEAR:
return;
default:
errx(EEXIT, "BAD STATE %d TO BLKERR", inoinfo(ino)->ino_state);
/* NOTREACHED */
}
}
/*
* allocate an unused inode
*/
ino_t
allocino(request, type)
ino_t request;
int type;
{
register ino_t ino;
register struct dinode *dp;
struct cg *cgp = &cgrp;
int cg;
if (request == 0)
request = ROOTINO;
else if (inoinfo(request)->ino_state != USTATE)
return (0);
for (ino = request; ino < maxino; ino++)
if (inoinfo(ino)->ino_state == USTATE)
break;
if (ino == maxino)
return (0);
cg = ino_to_cg(&sblock, ino);
getblk(&cgblk, cgtod(&sblock, cg), sblock.fs_cgsize);
if (!cg_chkmagic(cgp))
pfatal("CG %d: BAD MAGIC NUMBER\n", cg);
setbit(cg_inosused(cgp), ino % sblock.fs_ipg);
cgp->cg_cs.cs_nifree--;
switch (type & IFMT) {
case IFDIR:
inoinfo(ino)->ino_state = DSTATE;
cgp->cg_cs.cs_ndir++;
break;
case IFREG:
case IFLNK:
inoinfo(ino)->ino_state = FSTATE;
break;
default:
return (0);
}
cgdirty();
dp = ginode(ino);
dp->di_db[0] = allocblk((long)1);
if (dp->di_db[0] == 0) {
inoinfo(ino)->ino_state = USTATE;
return (0);
}
dp->di_mode = type;
dp->di_flags = 0;
dp->di_atime = time(NULL);
dp->di_mtime = dp->di_ctime = dp->di_atime;
dp->di_mtimensec = dp->di_ctimensec = dp->di_atimensec = 0;
dp->di_size = sblock.fs_fsize;
dp->di_blocks = btodb(sblock.fs_fsize);
n_files++;
inodirty();
if (newinofmt)
inoinfo(ino)->ino_type = IFTODT(type);
return (ino);
}
/*
* deallocate an inode
*/
void
freeino(ino)
ino_t ino;
{
struct inodesc idesc;
struct dinode *dp;
memset(&idesc, 0, sizeof(struct inodesc));
idesc.id_type = ADDR;
idesc.id_func = pass4check;
idesc.id_number = ino;
dp = ginode(ino);
(void)ckinode(dp, &idesc);
clearinode(dp);
inodirty();
inoinfo(ino)->ino_state = USTATE;
n_files--;
}