NetBSD/sbin/newfs/mkfs.c

1410 lines
39 KiB
C
Raw Normal View History

/* $NetBSD: mkfs.c,v 1.65 2002/09/28 20:11:07 dbj Exp $ */
1993-03-21 12:45:37 +03:00
/*
* Copyright (c) 1980, 1989, 1993
* The Regents of the University of California. All rights reserved.
1993-03-21 12:45:37 +03:00
*
* 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.
*/
1997-07-01 02:20:30 +04:00
#include <sys/cdefs.h>
1993-03-21 12:45:37 +03:00
#ifndef lint
#if 0
static char sccsid[] = "@(#)mkfs.c 8.11 (Berkeley) 5/3/95";
#else
__RCSID("$NetBSD: mkfs.c,v 1.65 2002/09/28 20:11:07 dbj Exp $");
#endif
1993-03-21 12:45:37 +03:00
#endif /* not lint */
#include <sys/param.h>
#include <sys/mman.h>
1993-03-21 12:45:37 +03:00
#include <sys/time.h>
#include <sys/resource.h>
#include <ufs/ufs/dinode.h>
#include <ufs/ufs/dir.h>
#include <ufs/ufs/ufs_bswap.h>
#include <ufs/ffs/fs.h>
#include <ufs/ffs/ffs_extern.h>
1993-03-21 12:45:37 +03:00
#include <sys/disklabel.h>
#include <err.h>
#include <errno.h>
#include <string.h>
#include <unistd.h>
1997-07-01 02:20:30 +04:00
#include <stdlib.h>
#ifndef STANDALONE
#include <stdio.h>
#endif
#include "extern.h"
1993-03-21 12:45:37 +03:00
2000-12-01 14:52:54 +03:00
static void initcg(int, time_t);
static int fsinit(time_t, mode_t, uid_t, gid_t);
2000-12-01 14:52:54 +03:00
static int makedir(struct direct *, int);
static daddr_t alloc(int, int);
static void iput(struct dinode *, ino_t);
static void rdfs(daddr_t, int, void *);
static void wtfs(daddr_t, int, void *);
static int isblock(struct fs *, unsigned char *, int);
static void clrblock(struct fs *, unsigned char *, int);
static void setblock(struct fs *, unsigned char *, int);
static int32_t calcipg(int32_t, int32_t, off_t *);
static void swap_cg(struct cg *, struct cg *);
#ifdef MFS
static void calc_memfree(void);
static void *mkfs_malloc(size_t size);
#endif
2000-12-01 14:52:54 +03:00
static int count_digits(int);
1993-03-21 12:45:37 +03:00
/*
* make file system for cylinder-group style file systems
1993-03-21 12:45:37 +03:00
*/
/*
* We limit the size of the inode map to be no more than a
* third of the cylinder group space, since we must leave at
* least an equal amount of space for the block map.
*
* N.B.: MAXIPG must be a multiple of INOPB(fs).
*/
#define MAXIPG(fs) roundup((fs)->fs_bsize * NBBY / 3, INOPB(fs))
1993-03-21 12:45:37 +03:00
#define UMASK 0755
#define MAXINOPB (MAXBSIZE / DINODE_SIZE)
#define POWEROF2(num) (((num) & ((num) - 1)) == 0)
1993-03-21 12:45:37 +03:00
union {
struct fs fs;
char pad[SBSIZE];
} fsun;
#define sblock fsun.fs
struct csum *fscs;
union {
struct cg cg;
char pad[MAXBSIZE];
} cgun;
#define acg cgun.cg
struct dinode zino[MAXBSIZE / DINODE_SIZE];
1993-03-21 12:45:37 +03:00
char writebuf[MAXBSIZE];
1993-03-21 12:45:37 +03:00
int fsi, fso;
1997-07-01 02:20:30 +04:00
void
mkfs(struct partition *pp, const char *fsys, int fi, int fo,
mode_t mfsmode, uid_t mfsuid, gid_t mfsgid)
1993-03-21 12:45:37 +03:00
{
int32_t i, mincpc, mincpg, inospercg;
int32_t cylno, rpos, blk, j, warning = 0;
int32_t used, mincpgcnt, bpcg;
off_t usedb;
int32_t mapcramped, inodecramped;
int32_t postblsize, rotblsize, totalsbsize;
1993-03-21 12:45:37 +03:00
time_t utime;
long long sizepb;
char *writebuf2; /* dynamic buffer */
int nprintcols, printcolwidth;
1993-03-21 12:45:37 +03:00
#ifndef STANDALONE
time(&utime);
#endif
#ifdef MFS
1993-03-21 12:45:37 +03:00
if (mfs) {
calc_memfree();
1993-03-21 12:45:37 +03:00
if (fssize * sectorsize > memleft)
fssize = memleft / sectorsize;
if ((membase = mkfs_malloc(fssize * sectorsize)) == 0)
1993-03-21 12:45:37 +03:00
exit(12);
}
#endif
1993-03-21 12:45:37 +03:00
fsi = fi;
fso = fo;
if (Oflag) {
sblock.fs_inodefmt = FS_42INODEFMT;
sblock.fs_maxsymlinklen = 0;
} else {
sblock.fs_inodefmt = FS_44INODEFMT;
sblock.fs_maxsymlinklen = MAXSYMLINKLEN;
}
1993-03-21 12:45:37 +03:00
/*
* Validate the given file system size.
* Verify that its last block can actually be accessed.
*/
if (fssize <= 0)
printf("preposterous size %d\n", fssize), exit(13);
wtfs(fssize - 1, sectorsize, (char *)&sblock);
if (isappleufs) {
struct appleufslabel appleufs;
ffs_appleufs_set(&appleufs,appleufs_volname,utime);
wtfs(APPLEUFS_LABEL_OFFSET/sectorsize,APPLEUFS_LABEL_SIZE,&appleufs);
}
1993-03-21 12:45:37 +03:00
/*
* collect and verify the sector and track info
*/
sblock.fs_nsect = nsectors;
sblock.fs_ntrak = ntracks;
if (sblock.fs_ntrak <= 0)
printf("preposterous ntrak %d\n", sblock.fs_ntrak), exit(14);
if (sblock.fs_nsect <= 0)
printf("preposterous nsect %d\n", sblock.fs_nsect), exit(15);
Incorporate the enhanced ffs_dirpref() by Grigoriy Orlov, as found in FreeBSD (three commits; the initial work, man page updates, and a fix to ffs_reload()), with the following differences: - Be consistent between newfs(8) and tunefs(8) as to the options which set and control the tuning parameters for this work (avgfilesize & avgfpdir) - Use u_int16_t instead of u_int8_t to keep track of the number of contiguous directories (suggested by Chuck Silvers) - Work within our FFS_EI framework - Ensure that fs->fs_maxclusters and fs->fs_contigdirs don't point to the same area of memory The new algorithm has a marked performance increase, especially when performing tasks such as untarring pkgsrc.tar.gz, etc. The original FreeBSD commit messages are attached: ===== mckusick 2001/04/10 01:39:00 PDT Directory layout preference improvements from Grigoriy Orlov <gluk@ptci.ru>. His description of the problem and solution follow. My own tests show speedups on typical filesystem intensive workloads of 5% to 12% which is very impressive considering the small amount of code change involved. ------ One day I noticed that some file operations run much faster on small file systems then on big ones. I've looked at the ffs algorithms, thought about them, and redesigned the dirpref algorithm. First I want to describe the results of my tests. These results are old and I have improved the algorithm after these tests were done. Nevertheless they show how big the perfomance speedup may be. I have done two file/directory intensive tests on a two OpenBSD systems with old and new dirpref algorithm. The first test is "tar -xzf ports.tar.gz", the second is "rm -rf ports". The ports.tar.gz file is the ports collection from the OpenBSD 2.8 release. It contains 6596 directories and 13868 files. The test systems are: 1. Celeron-450, 128Mb, two IDE drives, the system at wd0, file system for test is at wd1. Size of test file system is 8 Gb, number of cg=991, size of cg is 8m, block size = 8k, fragment size = 1k OpenBSD-current from Dec 2000 with BUFCACHEPERCENT=35 2. PIII-600, 128Mb, two IBM DTLA-307045 IDE drives at i815e, the system at wd0, file system for test is at wd1. Size of test file system is 40 Gb, number of cg=5324, size of cg is 8m, block size = 8k, fragment size = 1k OpenBSD-current from Dec 2000 with BUFCACHEPERCENT=50 You can get more info about the test systems and methods at: http://www.ptci.ru/gluk/dirpref/old/dirpref.html Test Results tar -xzf ports.tar.gz rm -rf ports mode old dirpref new dirpref speedup old dirprefnew dirpref speedup First system normal 667 472 1.41 477 331 1.44 async 285 144 1.98 130 14 9.29 sync 768 616 1.25 477 334 1.43 softdep 413 252 1.64 241 38 6.34 Second system normal 329 81 4.06 263.5 93.5 2.81 async 302 25.7 11.75 112 2.26 49.56 sync 281 57.0 4.93 263 90.5 2.9 softdep 341 40.6 8.4 284 4.76 59.66 "old dirpref" and "new dirpref" columns give a test time in seconds. speedup - speed increasement in times, ie. old dirpref / new dirpref. ------ Algorithm description The old dirpref algorithm is described in comments: /* * Find a cylinder to place a directory. * * The policy implemented by this algorithm is to select from * among those cylinder groups with above the average number of * free inodes, the one with the smallest number of directories. */ A new directory is allocated in a different cylinder groups than its parent directory resulting in a directory tree that is spreaded across all the cylinder groups. This spreading out results in a non-optimal access to the directories and files. When we have a small filesystem it is not a problem but when the filesystem is big then perfomance degradation becomes very apparent. What I mean by a big file system ? 1. A big filesystem is a filesystem which occupy 20-30 or more percent of total drive space, i.e. first and last cylinder are physically located relatively far from each other. 2. It has a relatively large number of cylinder groups, for example more cylinder groups than 50% of the buffers in the buffer cache. The first results in long access times, while the second results in many buffers being used by metadata operations. Such operations use cylinder group blocks and on-disk inode blocks. The cylinder group block (fs->fs_cblkno) contains struct cg, inode and block bit maps. It is 2k in size for the default filesystem parameters. If new and parent directories are located in different cylinder groups then the system performs more input/output operations and uses more buffers. On filesystems with many cylinder groups, lots of cache buffers are used for metadata operations. My solution for this problem is very simple. I allocate many directories in one cylinder group. I also do some things, so that the new allocation method does not cause excessive fragmentation and all directory inodes will not be located at a location far from its file's inodes and data. The algorithm is: /* * Find a cylinder group to place a directory. * * The policy implemented by this algorithm is to allocate a * directory inode in the same cylinder group as its parent * directory, but also to reserve space for its files inodes * and data. Restrict the number of directories which may be * allocated one after another in the same cylinder group * without intervening allocation of files. * * If we allocate a first level directory then force allocation * in another cylinder group. */ My early versions of dirpref give me a good results for a wide range of file operations and different filesystem capacities except one case: those applications that create their entire directory structure first and only later fill this structure with files. My solution for such and similar cases is to limit a number of directories which may be created one after another in the same cylinder group without intervening file creations. For this purpose, I allocate an array of counters at mount time. This array is linked to the superblock fs->fs_contigdirs[cg]. Each time a directory is created the counter increases and each time a file is created the counter decreases. A 60Gb filesystem with 8mb/cg requires 10kb of memory for the counters array. The maxcontigdirs is a maximum number of directories which may be created without an intervening file creation. I found in my tests that the best performance occurs when I restrict the number of directories in one cylinder group such that all its files may be located in the same cylinder group. There may be some deterioration in performance if all the file inodes are in the same cylinder group as its containing directory, but their data partially resides in a different cylinder group. The maxcontigdirs value is calculated to try to prevent this condition. Since there is no way to know how many files and directories will be allocated later I added two optimization parameters in superblock/tunefs. They are: int32_t fs_avgfilesize; /* expected average file size */ int32_t fs_avgfpdir; /* expected # of files per directory */ These parameters have reasonable defaults but may be tweeked for special uses of a filesystem. They are only necessary in rare cases like better tuning a filesystem being used to store a squid cache. I have been using this algorithm for about 3 months. I have done a lot of testing on filesystems with different capacities, average filesize, average number of files per directory, and so on. I think this algorithm has no negative impact on filesystem perfomance. It works better than the default one in all cases. The new dirpref will greatly improve untarring/removing/coping of big directories, decrease load on cvs servers and much more. The new dirpref doesn't speedup a compilation process, but also doesn't slow it down. Obtained from: Grigoriy Orlov <gluk@ptci.ru> ===== ===== iedowse 2001/04/23 17:37:17 PDT Pre-dirpref versions of fsck may zero out the new superblock fields fs_contigdirs, fs_avgfilesize and fs_avgfpdir. This could cause panics if these fields were zeroed while a filesystem was mounted read-only, and then remounted read-write. Add code to ffs_reload() which copies the fs_contigdirs pointer from the previous superblock, and reinitialises fs_avgf* if necessary. Reviewed by: mckusick ===== ===== nik 2001/04/10 03:36:44 PDT Add information about the new options to newfs and tunefs which set the expected average file size and number of files per directory. Could do with some fleshing out. =====
2001-09-06 06:16:00 +04:00
/*
* collect and verify the filesystem density info
*/
sblock.fs_avgfilesize = avgfilesize;
sblock.fs_avgfpdir = avgfpdir;
if (sblock.fs_avgfilesize <= 0)
printf("illegal expected average file size %d\n",
sblock.fs_avgfilesize), exit(14);
if (sblock.fs_avgfpdir <= 0)
printf("illegal expected number of files per directory %d\n",
sblock.fs_avgfpdir), exit(15);
1993-03-21 12:45:37 +03:00
/*
* collect and verify the block and fragment sizes
*/
sblock.fs_bsize = bsize;
sblock.fs_fsize = fsize;
if (!POWEROF2(sblock.fs_bsize)) {
printf("block size must be a power of 2, not %d\n",
sblock.fs_bsize);
exit(16);
}
if (!POWEROF2(sblock.fs_fsize)) {
printf("fragment size must be a power of 2, not %d\n",
sblock.fs_fsize);
exit(17);
}
if (sblock.fs_fsize < sectorsize) {
printf("fragment size %d is too small, minimum is %d\n",
sblock.fs_fsize, sectorsize);
exit(18);
}
if (sblock.fs_bsize < MINBSIZE) {
printf("block size %d is too small, minimum is %d\n",
sblock.fs_bsize, MINBSIZE);
exit(19);
}
2001-12-13 09:31:31 +03:00
if (sblock.fs_bsize > MAXBSIZE) {
printf("block size %d is too large, maximum is %d\n",
sblock.fs_bsize, MAXBSIZE);
exit(19);
}
1993-03-21 12:45:37 +03:00
if (sblock.fs_bsize < sblock.fs_fsize) {
printf("block size (%d) cannot be smaller than fragment size (%d)\n",
sblock.fs_bsize, sblock.fs_fsize);
exit(20);
}
sblock.fs_bmask = ~(sblock.fs_bsize - 1);
sblock.fs_fmask = ~(sblock.fs_fsize - 1);
sblock.fs_qbmask = ~sblock.fs_bmask;
sblock.fs_qfmask = ~sblock.fs_fmask;
1993-03-21 12:45:37 +03:00
for (sblock.fs_bshift = 0, i = sblock.fs_bsize; i > 1; i >>= 1)
sblock.fs_bshift++;
for (sblock.fs_fshift = 0, i = sblock.fs_fsize; i > 1; i >>= 1)
sblock.fs_fshift++;
sblock.fs_frag = numfrags(&sblock, sblock.fs_bsize);
for (sblock.fs_fragshift = 0, i = sblock.fs_frag; i > 1; i >>= 1)
sblock.fs_fragshift++;
if (sblock.fs_frag > MAXFRAG) {
printf("fragment size %d is too small, "
"minimum with block size %d is %d\n",
1993-03-21 12:45:37 +03:00
sblock.fs_fsize, sblock.fs_bsize,
sblock.fs_bsize / MAXFRAG);
exit(21);
}
sblock.fs_nrpos = nrpos;
sblock.fs_nindir = sblock.fs_bsize / sizeof(daddr_t);
sblock.fs_inopb = sblock.fs_bsize / DINODE_SIZE;
1993-03-21 12:45:37 +03:00
sblock.fs_nspf = sblock.fs_fsize / sectorsize;
for (sblock.fs_fsbtodb = 0, i = NSPF(&sblock); i > 1; i >>= 1)
sblock.fs_fsbtodb++;
sblock.fs_sblkno =
roundup(howmany(bbsize + sbsize, sblock.fs_fsize), sblock.fs_frag);
sblock.fs_cblkno = (daddr_t)(sblock.fs_sblkno +
roundup(howmany(sbsize, sblock.fs_fsize), sblock.fs_frag));
sblock.fs_iblkno = sblock.fs_cblkno + sblock.fs_frag;
sblock.fs_cgoffset = roundup(
howmany(sblock.fs_nsect, NSPF(&sblock)), sblock.fs_frag);
for (sblock.fs_cgmask = 0xffffffff, i = sblock.fs_ntrak; i > 1; i >>= 1)
sblock.fs_cgmask <<= 1;
if (!POWEROF2(sblock.fs_ntrak))
sblock.fs_cgmask <<= 1;
sblock.fs_maxfilesize = sblock.fs_bsize * NDADDR - 1;
for (sizepb = sblock.fs_bsize, i = 0; i < NIADDR; i++) {
sizepb *= NINDIR(&sblock);
sblock.fs_maxfilesize += sizepb;
}
1993-03-21 12:45:37 +03:00
/*
* Validate specified/determined secpercyl
* and calculate minimum cylinders per group.
*/
sblock.fs_spc = secpercyl;
for (sblock.fs_cpc = NSPB(&sblock), i = sblock.fs_spc;
sblock.fs_cpc > 1 && (i & 1) == 0;
sblock.fs_cpc >>= 1, i >>= 1)
/* void */;
mincpc = sblock.fs_cpc;
bpcg = sblock.fs_spc * sectorsize;
inospercg = roundup(bpcg / DINODE_SIZE, INOPB(&sblock));
1993-03-21 12:45:37 +03:00
if (inospercg > MAXIPG(&sblock))
inospercg = MAXIPG(&sblock);
used = (sblock.fs_iblkno + inospercg / INOPF(&sblock)) * NSPF(&sblock);
mincpgcnt = howmany(sblock.fs_cgoffset * (~sblock.fs_cgmask) + used,
sblock.fs_spc);
mincpg = roundup(mincpgcnt, mincpc);
/*
* Ensure that cylinder group with mincpg has enough space
* for block maps.
1993-03-21 12:45:37 +03:00
*/
sblock.fs_cpg = mincpg;
sblock.fs_ipg = inospercg;
if (maxcontig > 1)
sblock.fs_contigsumsize = MIN(maxcontig, FS_MAXCONTIG);
1993-03-21 12:45:37 +03:00
mapcramped = 0;
while (CGSIZE(&sblock) > sblock.fs_bsize) {
mapcramped = 1;
if (sblock.fs_bsize < MAXBSIZE) {
sblock.fs_bsize <<= 1;
if ((i & 1) == 0) {
i >>= 1;
} else {
sblock.fs_cpc <<= 1;
mincpc <<= 1;
mincpg = roundup(mincpgcnt, mincpc);
sblock.fs_cpg = mincpg;
}
sblock.fs_frag <<= 1;
sblock.fs_fragshift += 1;
if (sblock.fs_frag <= MAXFRAG)
continue;
}
if (sblock.fs_fsize == sblock.fs_bsize) {
printf("There is no block size that");
printf(" can support this disk\n");
exit(22);
}
sblock.fs_frag >>= 1;
sblock.fs_fragshift -= 1;
sblock.fs_fsize <<= 1;
sblock.fs_nspf <<= 1;
}
/*
* Ensure that cylinder group with mincpg has enough space for inodes.
1993-03-21 12:45:37 +03:00
*/
inodecramped = 0;
inospercg = calcipg(mincpg, bpcg, &usedb);
1993-03-21 12:45:37 +03:00
sblock.fs_ipg = inospercg;
while (inospercg > MAXIPG(&sblock)) {
inodecramped = 1;
if (mincpc == 1 || sblock.fs_frag == 1 ||
sblock.fs_bsize == MINBSIZE)
break;
printf("With a block size of %d %s %d\n", sblock.fs_bsize,
"minimum bytes per inode is",
(int)((mincpg * (off_t)bpcg - usedb)
/ MAXIPG(&sblock) + 1));
1993-03-21 12:45:37 +03:00
sblock.fs_bsize >>= 1;
sblock.fs_frag >>= 1;
sblock.fs_fragshift -= 1;
mincpc >>= 1;
sblock.fs_cpg = roundup(mincpgcnt, mincpc);
if (CGSIZE(&sblock) > sblock.fs_bsize) {
sblock.fs_bsize <<= 1;
break;
}
mincpg = sblock.fs_cpg;
inospercg = calcipg(mincpg, bpcg, &usedb);
1993-03-21 12:45:37 +03:00
sblock.fs_ipg = inospercg;
}
if (inodecramped) {
if (inospercg > MAXIPG(&sblock)) {
printf("Minimum bytes per inode is %d\n",
(int)((mincpg * (off_t)bpcg - usedb)
/ MAXIPG(&sblock) + 1));
1993-03-21 12:45:37 +03:00
} else if (!mapcramped) {
printf("With %d bytes per inode, ", density);
printf("minimum cylinders per group is %d\n", mincpg);
1993-03-21 12:45:37 +03:00
}
}
if (mapcramped) {
printf("With %d sectors per cylinder, ", sblock.fs_spc);
printf("minimum cylinders per group is %d\n", mincpg);
1993-03-21 12:45:37 +03:00
}
if (inodecramped || mapcramped) {
if (sblock.fs_bsize != bsize)
printf("%s to be changed from %d to %d\n",
"This requires the block size",
bsize, sblock.fs_bsize);
if (sblock.fs_fsize != fsize)
printf("\t%s to be changed from %d to %d\n",
"and the fragment size",
fsize, sblock.fs_fsize);
exit(23);
}
/*
* Calculate the number of cylinders per group
*/
sblock.fs_cpg = cpg;
if (sblock.fs_cpg % mincpc != 0) {
printf("%s groups must have a multiple of %d cylinders\n",
1993-03-21 12:45:37 +03:00
cpgflg ? "Cylinder" : "Warning: cylinder", mincpc);
sblock.fs_cpg = roundup(sblock.fs_cpg, mincpc);
if (!cpgflg)
cpg = sblock.fs_cpg;
}
/*
* Must ensure there is enough space for inodes.
1993-03-21 12:45:37 +03:00
*/
sblock.fs_ipg = calcipg(sblock.fs_cpg, bpcg, &usedb);
1993-03-21 12:45:37 +03:00
while (sblock.fs_ipg > MAXIPG(&sblock)) {
inodecramped = 1;
sblock.fs_cpg -= mincpc;
sblock.fs_ipg = calcipg(sblock.fs_cpg, bpcg, &usedb);
1993-03-21 12:45:37 +03:00
}
/*
* Must ensure there is enough space to hold block map.
1993-03-21 12:45:37 +03:00
*/
while (CGSIZE(&sblock) > sblock.fs_bsize) {
mapcramped = 1;
sblock.fs_cpg -= mincpc;
sblock.fs_ipg = calcipg(sblock.fs_cpg, bpcg, &usedb);
1993-03-21 12:45:37 +03:00
}
sblock.fs_fpg = (sblock.fs_cpg * sblock.fs_spc) / NSPF(&sblock);
if ((sblock.fs_cpg * sblock.fs_spc) % NSPB(&sblock) != 0) {
1997-07-01 02:20:30 +04:00
printf("panic (fs_cpg * fs_spc) %% NSPF != 0");
1993-03-21 12:45:37 +03:00
exit(24);
}
if (sblock.fs_cpg < mincpg) {
printf("cylinder groups must have at least %d cylinders\n",
1993-03-21 12:45:37 +03:00
mincpg);
exit(25);
} else if (sblock.fs_cpg != cpg && cpgflg) {
if (!mapcramped && !inodecramped)
1993-03-21 12:45:37 +03:00
exit(26);
if (mapcramped && inodecramped)
printf("Block size and bytes per inode restrict");
else if (mapcramped)
printf("Block size restricts");
else
printf("Bytes per inode restrict");
printf(" cylinders per group to %d.\n", sblock.fs_cpg);
exit(27);
1993-03-21 12:45:37 +03:00
}
sblock.fs_cgsize = fragroundup(&sblock, CGSIZE(&sblock));
/*
* Now have size for file system and nsect and ntrak.
* Determine number of cylinders and blocks in the file system.
*/
sblock.fs_size = fssize = dbtofsb(&sblock, fssize);
sblock.fs_ncyl = fssize * NSPF(&sblock) / sblock.fs_spc;
if (fssize * NSPF(&sblock) > sblock.fs_ncyl * sblock.fs_spc) {
sblock.fs_ncyl++;
warning = 1;
1993-03-21 12:45:37 +03:00
}
if (sblock.fs_ncyl < 1) {
printf("file systems must have at least one cylinder\n");
exit(28);
}
/*
* Determine feasability/values of rotational layout tables.
*
* The size of the rotational layout tables is limited by the
* size of the superblock, SBSIZE. The amount of space available
* for tables is calculated as (SBSIZE - sizeof (struct fs)).
* The size of these tables is inversely proportional to the block
* size of the file system. The size increases if sectors per track
* are not powers of two, because more cylinders must be described
* by the tables before the rotational pattern repeats (fs_cpc).
*/
sblock.fs_interleave = interleave;
sblock.fs_trackskew = trackskew;
sblock.fs_npsect = nphyssectors;
sblock.fs_postblformat = FS_DYNAMICPOSTBLFMT;
sblock.fs_sbsize = fragroundup(&sblock, sizeof(struct fs));
if (sblock.fs_ntrak == 1) {
sblock.fs_cpc = 0;
goto next;
}
1995-03-21 04:28:05 +03:00
postblsize = sblock.fs_nrpos * sblock.fs_cpc * sizeof(int16_t);
1993-03-21 12:45:37 +03:00
rotblsize = sblock.fs_cpc * sblock.fs_spc / NSPB(&sblock);
totalsbsize = sizeof(struct fs) + rotblsize;
if (sblock.fs_nrpos == 8 && sblock.fs_cpc <= 16) {
/* use old static table space */
sblock.fs_postbloff = (char *)(&sblock.fs_opostbl[0][0]) -
1994-12-18 19:00:21 +03:00
(char *)(&sblock.fs_firstfield);
1993-03-21 12:45:37 +03:00
sblock.fs_rotbloff = &sblock.fs_space[0] -
1994-12-18 19:00:21 +03:00
(u_char *)(&sblock.fs_firstfield);
1993-03-21 12:45:37 +03:00
} else {
/* use dynamic table space */
sblock.fs_postbloff = &sblock.fs_space[0] -
1994-12-18 19:00:21 +03:00
(u_char *)(&sblock.fs_firstfield);
1993-03-21 12:45:37 +03:00
sblock.fs_rotbloff = sblock.fs_postbloff + postblsize;
totalsbsize += postblsize;
}
if (totalsbsize > SBSIZE ||
sblock.fs_nsect > (1 << NBBY) * NSPB(&sblock)) {
printf("%s %s %d %s %d.%s",
"Warning: insufficient space in super block for\n",
"rotational layout tables with nsect", sblock.fs_nsect,
"and ntrak", sblock.fs_ntrak,
"\nFile system performance may be impaired.\n");
sblock.fs_cpc = 0;
goto next;
}
sblock.fs_sbsize = fragroundup(&sblock, totalsbsize);
/*
* calculate the available blocks for each rotational position
*/
for (cylno = 0; cylno < sblock.fs_cpc; cylno++)
for (rpos = 0; rpos < sblock.fs_nrpos; rpos++)
fs_postbl(&sblock, cylno)[rpos] = -1;
for (i = (rotblsize - 1) << sblock.fs_fragshift;
1993-03-21 12:45:37 +03:00
i >= 0; i -= sblock.fs_frag) {
cylno = cbtocylno(&sblock, i);
rpos = cbtorpos(&sblock, i);
blk = fragstoblks(&sblock, i);
if (fs_postbl(&sblock, cylno)[rpos] == -1)
fs_rotbl(&sblock)[blk] = 0;
else
fs_rotbl(&sblock)[blk] = fs_postbl(&sblock, cylno)[rpos] - blk;
1993-03-21 12:45:37 +03:00
fs_postbl(&sblock, cylno)[rpos] = blk;
}
next:
/*
* Compute/validate number of cylinder groups.
*/
sblock.fs_ncg = sblock.fs_ncyl / sblock.fs_cpg;
if (sblock.fs_ncyl % sblock.fs_cpg)
sblock.fs_ncg++;
sblock.fs_dblkno = sblock.fs_iblkno + sblock.fs_ipg / INOPF(&sblock);
i = MIN(~sblock.fs_cgmask, sblock.fs_ncg - 1);
if (cgdmin(&sblock, i) - cgbase(&sblock, i) >= sblock.fs_fpg) {
printf("inode blocks/cyl group (%d) >= data blocks (%d)\n",
cgdmin(&sblock, i) -
(cgbase(&sblock, i) >> sblock.fs_fragshift),
sblock.fs_fpg >> sblock.fs_fragshift);
1993-03-21 12:45:37 +03:00
printf("number of cylinders per cylinder group (%d) %s.\n",
sblock.fs_cpg, "must be increased");
exit(29);
}
j = sblock.fs_ncg - 1;
if ((i = fssize - j * sblock.fs_fpg) < sblock.fs_fpg &&
cgdmin(&sblock, j) - cgbase(&sblock, j) > i) {
if (j == 0) {
2001-07-27 10:41:02 +04:00
printf("File system must have at least %d sectors\n",
1993-03-21 12:45:37 +03:00
NSPF(&sblock) *
(cgdmin(&sblock, 0) + (3 << sblock.fs_fragshift)));
1993-03-21 12:45:37 +03:00
exit(30);
}
printf("Warning: inode blocks/cyl group (%d) >= "
"data blocks (%d) in last\n",
(cgdmin(&sblock, j) -
cgbase(&sblock, j)) >> sblock.fs_fragshift,
i >> sblock.fs_fragshift);
printf(" cylinder group. This implies %d sector(s) "
"cannot be allocated.\n",
1993-03-21 12:45:37 +03:00
i * NSPF(&sblock));
sblock.fs_ncg--;
sblock.fs_ncyl -= sblock.fs_ncyl % sblock.fs_cpg;
sblock.fs_size = fssize = sblock.fs_ncyl * sblock.fs_spc /
NSPF(&sblock);
warning = 0;
1993-03-21 12:45:37 +03:00
}
if (warning && !mfs) {
1993-03-21 12:45:37 +03:00
printf("Warning: %d sector(s) in last cylinder unallocated\n",
sblock.fs_spc -
(fssize * NSPF(&sblock) - (sblock.fs_ncyl - 1)
* sblock.fs_spc));
}
/*
* fill in remaining fields of the super block
*/
sblock.fs_csaddr = cgdmin(&sblock, 0);
sblock.fs_cssize =
fragroundup(&sblock, sblock.fs_ncg * sizeof(struct csum));
/*
* The superblock fields 'fs_csmask' and 'fs_csshift' are no
* longer used. However, we still initialise them so that the
* filesystem remains compatible with old kernels.
*/
1993-03-21 12:45:37 +03:00
i = sblock.fs_bsize / sizeof(struct csum);
sblock.fs_csmask = ~(i - 1);
for (sblock.fs_csshift = 0; i > 1; i >>= 1)
sblock.fs_csshift++;
fscs = (struct csum *)calloc(1, sblock.fs_cssize);
if (fscs == NULL)
exit(39);
1993-03-21 12:45:37 +03:00
sblock.fs_magic = FS_MAGIC;
sblock.fs_rotdelay = rotdelay;
sblock.fs_minfree = minfree;
sblock.fs_maxcontig = maxcontig;
sblock.fs_maxbpg = maxbpg;
sblock.fs_rps = rpm / 60;
sblock.fs_optim = opt;
sblock.fs_cgrotor = 0;
sblock.fs_cstotal.cs_ndir = 0;
sblock.fs_cstotal.cs_nbfree = 0;
sblock.fs_cstotal.cs_nifree = 0;
sblock.fs_cstotal.cs_nffree = 0;
sblock.fs_fmod = 0;
1995-04-13 01:24:31 +04:00
sblock.fs_clean = FS_ISCLEAN;
1993-03-21 12:45:37 +03:00
sblock.fs_ronly = 0;
/*
* Dump out summary information about file system.
*/
if (!mfs) {
printf("%s:\t%d sectors in %d %s of %d tracks, %d sectors\n",
fsys, sblock.fs_size * NSPF(&sblock), sblock.fs_ncyl,
"cylinders", sblock.fs_ntrak, sblock.fs_nsect);
#define B2MBFACTOR (1 / (1024.0 * 1024.0))
1993-03-21 12:45:37 +03:00
printf("\t%.1fMB in %d cyl groups (%d c/g, %.2fMB/g, %d i/g)\n",
(float)sblock.fs_size * sblock.fs_fsize * B2MBFACTOR,
1993-03-21 12:45:37 +03:00
sblock.fs_ncg, sblock.fs_cpg,
(float)sblock.fs_fpg * sblock.fs_fsize * B2MBFACTOR,
1993-03-21 12:45:37 +03:00
sblock.fs_ipg);
#undef B2MBFACTOR
1993-03-21 12:45:37 +03:00
}
/*
* Now determine how wide each column will be, and calculate how
* many columns will fit in a 76 char line. 76 is the width of the
* subwindows in sysinst.
*/
printcolwidth = count_digits(
fsbtodb(&sblock, cgsblock(&sblock, sblock.fs_ncg -1)));
nprintcols = 76 / (printcolwidth + 2);
1993-03-21 12:45:37 +03:00
/*
* Now build the cylinders group blocks and
* then print out indices of cylinder groups.
*/
if (!mfs)
printf("super-block backups (for fsck -b #) at:");
for (cylno = 0; cylno < sblock.fs_ncg; cylno++) {
initcg(cylno, utime);
if (mfs)
continue;
if (cylno % nprintcols == 0)
1993-03-21 12:45:37 +03:00
printf("\n");
printf(" %*d,", printcolwidth,
fsbtodb(&sblock, cgsblock(&sblock, cylno)));
fflush(stdout);
1993-03-21 12:45:37 +03:00
}
if (!mfs)
printf("\n");
if (Nflag && !mfs)
exit(0);
/*
* Now construct the initial file system,
* then write out the super-block.
*/
if (fsinit(utime, mfsmode, mfsuid, mfsgid) == 0 && mfs)
errx(1, "Error making filesystem");
1993-03-21 12:45:37 +03:00
sblock.fs_time = utime;
memcpy(writebuf, &sblock, sbsize);
if (needswap)
ffs_sb_swap(&sblock, (struct fs*)writebuf);
wtfs((int)SBOFF / sectorsize, sbsize, writebuf);
1993-03-21 12:45:37 +03:00
/*
* Write out the duplicate super blocks
*/
for (cylno = 0; cylno < sblock.fs_ncg; cylno++)
wtfs(fsbtodb(&sblock, cgsblock(&sblock, cylno)),
sbsize, writebuf);
/*
* if we need to swap, create a buffer for the cylinder summaries
* to get swapped to.
*/
if (needswap) {
if ((writebuf2 = malloc(sblock.fs_cssize)) == NULL)
exit(12);
ffs_csum_swap(fscs, (struct csum*)writebuf2, sblock.fs_cssize);
} else
writebuf2 = (char *)fscs;
for (i = 0; i < sblock.fs_cssize; i += sblock.fs_bsize)
wtfs(fsbtodb(&sblock, sblock.fs_csaddr + numfrags(&sblock, i)),
sblock.fs_cssize - i < sblock.fs_bsize ?
sblock.fs_cssize - i : sblock.fs_bsize,
((char *)writebuf2) + i);
if (writebuf2 != (char *)fscs)
free(writebuf2);
1993-03-21 12:45:37 +03:00
/*
* Update information about this partion in pack
* label, to that it may be updated on disk.
*/
if (isappleufs)
pp->p_fstype = FS_APPLEUFS;
else
pp->p_fstype = FS_BSDFFS;
1993-03-21 12:45:37 +03:00
pp->p_fsize = sblock.fs_fsize;
pp->p_frag = sblock.fs_frag;
pp->p_cpg = sblock.fs_cpg;
}
/*
* Initialize a cylinder group.
*/
1997-07-01 02:20:30 +04:00
void
2000-12-01 14:52:54 +03:00
initcg(int cylno, time_t utime)
1993-03-21 12:45:37 +03:00
{
daddr_t cbase, d, dlower, dupper, dmax, blkno;
int32_t i;
1997-07-01 02:20:30 +04:00
struct csum *cs;
2002-04-10 21:28:13 +04:00
int cn;
1993-03-21 12:45:37 +03:00
/*
* Determine block bounds for cylinder group.
* Allow space for super block summary information in first
* cylinder group.
*/
cbase = cgbase(&sblock, cylno);
dmax = cbase + sblock.fs_fpg;
if (dmax > sblock.fs_size)
dmax = sblock.fs_size;
dlower = cgsblock(&sblock, cylno) - cbase;
dupper = cgdmin(&sblock, cylno) - cbase;
if (cylno == 0)
dupper += howmany(sblock.fs_cssize, sblock.fs_fsize);
cs = fscs + cylno;
memset(&acg, 0, sblock.fs_cgsize);
1993-03-21 12:45:37 +03:00
acg.cg_time = utime;
acg.cg_magic = CG_MAGIC;
acg.cg_cgx = cylno;
if (cylno == sblock.fs_ncg - 1)
acg.cg_ncyl = sblock.fs_ncyl % sblock.fs_cpg;
else
acg.cg_ncyl = sblock.fs_cpg;
acg.cg_niblk = sblock.fs_ipg;
acg.cg_ndblk = dmax - cbase;
if (sblock.fs_contigsumsize > 0)
acg.cg_nclusterblks = acg.cg_ndblk >> sblock.fs_fragshift;
1994-12-18 19:00:21 +03:00
acg.cg_btotoff = &acg.cg_space[0] - (u_char *)(&acg.cg_firstfield);
acg.cg_boff = acg.cg_btotoff + sblock.fs_cpg * sizeof(int32_t);
1993-03-21 12:45:37 +03:00
acg.cg_iusedoff = acg.cg_boff +
1995-03-21 04:28:05 +03:00
sblock.fs_cpg * sblock.fs_nrpos * sizeof(int16_t);
1993-03-21 12:45:37 +03:00
acg.cg_freeoff = acg.cg_iusedoff + howmany(sblock.fs_ipg, NBBY);
if (sblock.fs_contigsumsize <= 0) {
acg.cg_nextfreeoff = acg.cg_freeoff +
howmany(sblock.fs_fpg, NBBY);
} else {
acg.cg_clustersumoff = acg.cg_freeoff +
howmany(sblock.fs_fpg, NBBY) - sizeof(int32_t);
if (isappleufs) {
/* Apple PR2216969 gives rationale for this change.
* I believe they were mistaken, but we need to
* duplicate it for compatibility. -- dbj@netbsd.org
*/
acg.cg_clustersumoff += sizeof(int32_t);
}
acg.cg_clustersumoff =
roundup(acg.cg_clustersumoff, sizeof(int32_t));
acg.cg_clusteroff = acg.cg_clustersumoff +
(sblock.fs_contigsumsize + 1) * sizeof(int32_t);
acg.cg_nextfreeoff = acg.cg_clusteroff +
howmany(fragstoblks(&sblock, sblock.fs_fpg), NBBY);
}
if (acg.cg_nextfreeoff > sblock.fs_cgsize) {
printf("Panic: cylinder group too big\n");
exit(37);
1993-03-21 12:45:37 +03:00
}
acg.cg_cs.cs_nifree += sblock.fs_ipg;
if (cylno == 0)
for (i = 0; i < ROOTINO; i++) {
setbit(cg_inosused(&acg, 0), i);
1993-03-21 12:45:37 +03:00
acg.cg_cs.cs_nifree--;
}
for (i = 0; i < sblock.fs_ipg / INOPF(&sblock); i += sblock.fs_frag)
wtfs(fsbtodb(&sblock, cgimin(&sblock, cylno) + i),
sblock.fs_bsize, (char *)zino);
if (cylno > 0) {
/*
* In cylno 0, beginning space is reserved
* for boot and super blocks.
*/
2002-04-10 21:28:13 +04:00
for (d = 0, blkno = 0; d < dlower; ) {
setblock(&sblock, cg_blksfree(&acg, 0), blkno);
if (sblock.fs_contigsumsize > 0)
setbit(cg_clustersfree(&acg, 0), blkno);
1993-03-21 12:45:37 +03:00
acg.cg_cs.cs_nbfree++;
2002-04-10 21:28:13 +04:00
cn = cbtocylno(&sblock, d);
cg_blktot(&acg, 0)[cn]++;
cg_blks(&sblock, &acg, cn, 0)[cbtorpos(&sblock, d)]++;
d += sblock.fs_frag;
blkno++;
1993-03-21 12:45:37 +03:00
}
sblock.fs_dsize += dlower;
}
sblock.fs_dsize += acg.cg_ndblk - dupper;
if ((i = (dupper & (sblock.fs_frag - 1))) != 0) {
1993-03-21 12:45:37 +03:00
acg.cg_frsum[sblock.fs_frag - i]++;
for (d = dupper + sblock.fs_frag - i; dupper < d; dupper++) {
setbit(cg_blksfree(&acg, 0), dupper);
1993-03-21 12:45:37 +03:00
acg.cg_cs.cs_nffree++;
}
}
2002-04-10 21:28:13 +04:00
for (d = dupper, blkno = dupper >> sblock.fs_fragshift;
d + sblock.fs_frag <= dmax - cbase; ) {
setblock(&sblock, cg_blksfree(&acg, 0), blkno);
if (sblock.fs_contigsumsize > 0)
setbit(cg_clustersfree(&acg, 0), blkno);
1993-03-21 12:45:37 +03:00
acg.cg_cs.cs_nbfree++;
2002-04-10 21:28:13 +04:00
cn = cbtocylno(&sblock, d);
cg_blktot(&acg, 0)[cn]++;
cg_blks(&sblock, &acg, cn, 0)[cbtorpos(&sblock, d)]++;
1993-03-21 12:45:37 +03:00
d += sblock.fs_frag;
2002-04-10 21:28:13 +04:00
blkno++;
1993-03-21 12:45:37 +03:00
}
if (d < dmax - cbase) {
acg.cg_frsum[dmax - cbase - d]++;
for (; d < dmax - cbase; d++) {
setbit(cg_blksfree(&acg, 0), d);
1993-03-21 12:45:37 +03:00
acg.cg_cs.cs_nffree++;
}
}
if (sblock.fs_contigsumsize > 0) {
int32_t *sump = cg_clustersum(&acg, 0);
u_char *mapp = cg_clustersfree(&acg, 0);
int map = *mapp++;
int bit = 1;
int run = 0;
for (i = 0; i < acg.cg_nclusterblks; i++) {
if ((map & bit) != 0) {
run++;
} else if (run != 0) {
if (run > sblock.fs_contigsumsize)
run = sblock.fs_contigsumsize;
sump[run]++;
run = 0;
}
if ((i & (NBBY - 1)) != (NBBY - 1)) {
bit <<= 1;
} else {
map = *mapp++;
bit = 1;
}
}
if (run != 0) {
if (run > sblock.fs_contigsumsize)
run = sblock.fs_contigsumsize;
sump[run]++;
}
}
1993-03-21 12:45:37 +03:00
sblock.fs_cstotal.cs_ndir += acg.cg_cs.cs_ndir;
sblock.fs_cstotal.cs_nffree += acg.cg_cs.cs_nffree;
sblock.fs_cstotal.cs_nbfree += acg.cg_cs.cs_nbfree;
sblock.fs_cstotal.cs_nifree += acg.cg_cs.cs_nifree;
*cs = acg.cg_cs;
memcpy(writebuf, &acg, sblock.fs_bsize);
if (needswap)
swap_cg(&acg, (struct cg*)writebuf);
1993-03-21 12:45:37 +03:00
wtfs(fsbtodb(&sblock, cgtod(&sblock, cylno)),
sblock.fs_bsize, writebuf);
1993-03-21 12:45:37 +03:00
}
/*
* initialize the file system
*/
struct dinode node;
#ifdef LOSTDIR
#define PREDEFDIR 3
1993-03-21 12:45:37 +03:00
#else
#define PREDEFDIR 2
1993-03-21 12:45:37 +03:00
#endif
struct direct root_dir[] = {
{ ROOTINO, sizeof(struct direct), DT_DIR, 1, "." },
{ ROOTINO, sizeof(struct direct), DT_DIR, 2, ".." },
#ifdef LOSTDIR
{ LOSTFOUNDINO, sizeof(struct direct), DT_DIR, 10, "lost+found" },
#endif
};
struct odirect {
u_int32_t d_ino;
u_int16_t d_reclen;
u_int16_t d_namlen;
u_char d_name[MAXNAMLEN + 1];
} oroot_dir[] = {
1993-03-21 12:45:37 +03:00
{ ROOTINO, sizeof(struct direct), 1, "." },
{ ROOTINO, sizeof(struct direct), 2, ".." },
#ifdef LOSTDIR
{ LOSTFOUNDINO, sizeof(struct direct), 10, "lost+found" },
#endif
};
#ifdef LOSTDIR
struct direct lost_found_dir[] = {
{ LOSTFOUNDINO, sizeof(struct direct), DT_DIR, 1, "." },
{ ROOTINO, sizeof(struct direct), DT_DIR, 2, ".." },
{ 0, DIRBLKSIZ, 0, 0, 0 },
};
struct odirect olost_found_dir[] = {
1993-03-21 12:45:37 +03:00
{ LOSTFOUNDINO, sizeof(struct direct), 1, "." },
{ ROOTINO, sizeof(struct direct), 2, ".." },
{ 0, DIRBLKSIZ, 0, 0 },
};
#endif
char buf[MAXBSIZE];
2000-12-01 14:52:54 +03:00
static void copy_dir(struct direct *, struct direct *);
1993-03-21 12:45:37 +03:00
int
fsinit(time_t utime, mode_t mfsmode, uid_t mfsuid, gid_t mfsgid)
1993-03-21 12:45:37 +03:00
{
1997-07-01 02:20:30 +04:00
#ifdef LOSTDIR
1993-03-21 12:45:37 +03:00
int i;
int dirblksiz = DIRBLKSIZ;
if (isappleufs)
dirblksiz = APPLEUFS_DIRBLKSIZ;
1997-07-01 02:20:30 +04:00
#endif
1993-03-21 12:45:37 +03:00
/*
* initialize the node
*/
memset(&node, 0, sizeof(node));
1995-06-19 01:35:33 +04:00
node.di_atime = utime;
node.di_mtime = utime;
node.di_ctime = utime;
1993-03-21 12:45:37 +03:00
#ifdef LOSTDIR
/*
* create the lost+found directory
*/
if (Oflag) {
(void)makedir((struct direct *)olost_found_dir, 2);
for (i = dirblksiz; i < sblock.fs_bsize; i += dirblksiz)
copy_dir((struct direct*)&olost_found_dir[2],
(struct direct*)&buf[i]);
} else {
(void)makedir(lost_found_dir, 2);
for (i = dirblksiz; i < sblock.fs_bsize; i += dirblksiz)
copy_dir(&lost_found_dir[2], (struct direct*)&buf[i]);
}
1993-03-21 12:45:37 +03:00
node.di_mode = IFDIR | UMASK;
node.di_nlink = 2;
node.di_size = sblock.fs_bsize;
node.di_db[0] = alloc(node.di_size, node.di_mode);
node.di_blocks = btodb(fragroundup(&sblock, node.di_size));
node.di_uid = geteuid();
node.di_gid = getegid();
1993-03-21 12:45:37 +03:00
wtfs(fsbtodb(&sblock, node.di_db[0]), node.di_size, buf);
iput(&node, LOSTFOUNDINO);
#endif
/*
* create the root directory
*/
if (mfs) {
node.di_mode = IFDIR | mfsmode;
node.di_uid = mfsuid;
node.di_gid = mfsgid;
} else {
1993-03-21 12:45:37 +03:00
node.di_mode = IFDIR | UMASK;
node.di_uid = geteuid();
node.di_gid = getegid();
}
1993-03-21 12:45:37 +03:00
node.di_nlink = PREDEFDIR;
if (Oflag)
node.di_size = makedir((struct direct *)oroot_dir, PREDEFDIR);
else
node.di_size = makedir(root_dir, PREDEFDIR);
1993-03-21 12:45:37 +03:00
node.di_db[0] = alloc(sblock.fs_fsize, node.di_mode);
if (node.di_db[0] == 0)
return (0);
1993-03-21 12:45:37 +03:00
node.di_blocks = btodb(fragroundup(&sblock, node.di_size));
wtfs(fsbtodb(&sblock, node.di_db[0]), sblock.fs_fsize, buf);
iput(&node, ROOTINO);
return (1);
1993-03-21 12:45:37 +03:00
}
/*
* construct a set of directory entries in "buf".
* return size of directory.
*/
1997-07-01 02:20:30 +04:00
int
2000-12-01 14:52:54 +03:00
makedir(struct direct *protodir, int entries)
1993-03-21 12:45:37 +03:00
{
char *cp;
int i, spcleft;
int dirblksiz = DIRBLKSIZ;
if (isappleufs)
dirblksiz = APPLEUFS_DIRBLKSIZ;
1993-03-21 12:45:37 +03:00
spcleft = dirblksiz;
1993-03-21 12:45:37 +03:00
for (cp = buf, i = 0; i < entries - 1; i++) {
protodir[i].d_reclen = DIRSIZ(Oflag, &protodir[i], 0);
copy_dir(&protodir[i], (struct direct*)cp);
1993-03-21 12:45:37 +03:00
cp += protodir[i].d_reclen;
spcleft -= protodir[i].d_reclen;
}
protodir[i].d_reclen = spcleft;
copy_dir(&protodir[i], (struct direct*)cp);
return (dirblksiz);
1993-03-21 12:45:37 +03:00
}
/*
* allocate a block or frag
*/
daddr_t
2000-12-01 14:52:54 +03:00
alloc(int size, int mode)
1993-03-21 12:45:37 +03:00
{
int i, frag;
daddr_t d, blkno;
1993-03-21 12:45:37 +03:00
1997-07-01 02:20:30 +04:00
rdfs(fsbtodb(&sblock, cgtod(&sblock, 0)), sblock.fs_cgsize, &acg);
/* fs -> host byte order */
if (needswap)
swap_cg(&acg, &acg);
1993-03-21 12:45:37 +03:00
if (acg.cg_magic != CG_MAGIC) {
printf("cg 0: bad magic number\n");
return (0);
}
if (acg.cg_cs.cs_nbfree == 0) {
printf("first cylinder group ran out of space\n");
return (0);
}
for (d = 0; d < acg.cg_ndblk; d += sblock.fs_frag)
if (isblock(&sblock, cg_blksfree(&acg, 0),
d >> sblock.fs_fragshift))
1993-03-21 12:45:37 +03:00
goto goth;
printf("internal error: can't find block in cyl 0\n");
return (0);
goth:
blkno = fragstoblks(&sblock, d);
clrblock(&sblock, cg_blksfree(&acg, 0), blkno);
if (sblock.fs_contigsumsize > 0)
clrbit(cg_clustersfree(&acg, 0), blkno);
1993-03-21 12:45:37 +03:00
acg.cg_cs.cs_nbfree--;
sblock.fs_cstotal.cs_nbfree--;
fscs[0].cs_nbfree--;
if (mode & IFDIR) {
acg.cg_cs.cs_ndir++;
sblock.fs_cstotal.cs_ndir++;
fscs[0].cs_ndir++;
}
cg_blktot(&acg, 0)[cbtocylno(&sblock, d)]--;
cg_blks(&sblock, &acg, cbtocylno(&sblock, d), 0)[cbtorpos(&sblock, d)]--;
1993-03-21 12:45:37 +03:00
if (size != sblock.fs_bsize) {
frag = howmany(size, sblock.fs_fsize);
fscs[0].cs_nffree += sblock.fs_frag - frag;
sblock.fs_cstotal.cs_nffree += sblock.fs_frag - frag;
acg.cg_cs.cs_nffree += sblock.fs_frag - frag;
acg.cg_frsum[sblock.fs_frag - frag]++;
for (i = frag; i < sblock.fs_frag; i++)
setbit(cg_blksfree(&acg, 0), d + i);
1993-03-21 12:45:37 +03:00
}
/* host -> fs byte order */
if (needswap)
swap_cg(&acg, &acg);
1993-03-21 12:45:37 +03:00
wtfs(fsbtodb(&sblock, cgtod(&sblock, 0)), sblock.fs_cgsize,
(char *)&acg);
return (d);
}
/*
* Calculate number of inodes per group.
*/
int32_t
calcipg(int32_t cylpg, int32_t bpcg, off_t *usedbp)
{
int i;
int32_t ipg, new_ipg, ncg, ncyl;
off_t usedb;
/*
* Prepare to scale by fssize / (number of sectors in cylinder groups).
2001-07-27 10:41:02 +04:00
* Note that fssize is still in sectors, not file system blocks.
*/
ncyl = howmany(fssize, secpercyl);
ncg = howmany(ncyl, cylpg);
/*
* Iterate a few times to allow for ipg depending on itself.
*/
ipg = 0;
for (i = 0; i < 10; i++) {
usedb = (sblock.fs_iblkno + ipg / INOPF(&sblock))
* NSPF(&sblock) * (off_t)sectorsize;
new_ipg = (cylpg * (long long)bpcg - usedb) /
(long long)density * fssize / (ncg * secpercyl * cylpg);
if (new_ipg <= 0)
new_ipg = 1; /* ensure ipg > 0 */
new_ipg = roundup(new_ipg, INOPB(&sblock));
if (new_ipg == ipg)
break;
ipg = new_ipg;
}
*usedbp = usedb;
return (ipg);
}
1993-03-21 12:45:37 +03:00
/*
* Allocate an inode on the disk
*/
1997-07-01 02:20:30 +04:00
static void
2000-12-01 14:52:54 +03:00
iput(struct dinode *ip, ino_t ino)
1993-03-21 12:45:37 +03:00
{
struct dinode ibuf[MAXINOPB];
1993-03-21 12:45:37 +03:00
daddr_t d;
int c, i;
1993-03-21 12:45:37 +03:00
c = ino_to_cg(&sblock, ino);
1997-07-01 02:20:30 +04:00
rdfs(fsbtodb(&sblock, cgtod(&sblock, 0)), sblock.fs_cgsize, &acg);
/* fs -> host byte order */
if (needswap)
swap_cg(&acg, &acg);
1993-03-21 12:45:37 +03:00
if (acg.cg_magic != CG_MAGIC) {
printf("cg 0: bad magic number\n");
exit(31);
}
acg.cg_cs.cs_nifree--;
setbit(cg_inosused(&acg, 0), ino);
/* host -> fs byte order */
if (needswap)
swap_cg(&acg, &acg);
1993-03-21 12:45:37 +03:00
wtfs(fsbtodb(&sblock, cgtod(&sblock, 0)), sblock.fs_cgsize,
(char *)&acg);
sblock.fs_cstotal.cs_nifree--;
fscs[0].cs_nifree--;
if (ino >= sblock.fs_ipg * sblock.fs_ncg) {
printf("fsinit: inode value out of range (%d).\n", ino);
exit(32);
}
d = fsbtodb(&sblock, ino_to_fsba(&sblock, ino));
rdfs(d, sblock.fs_bsize, ibuf);
if (needswap) {
ffs_dinode_swap(ip, &ibuf[ino_to_fsbo(&sblock, ino)]);
/* ffs_dinode_swap() doesn't swap blocks addrs */
for (i=0; i<NDADDR + NIADDR; i++)
(&ibuf[ino_to_fsbo(&sblock, ino)])->di_db[i] =
bswap32(ip->di_db[i]);
} else
ibuf[ino_to_fsbo(&sblock, ino)] = *ip;
wtfs(d, sblock.fs_bsize, ibuf);
1993-03-21 12:45:37 +03:00
}
/*
* read a block from the file system
*/
1997-07-01 02:20:30 +04:00
void
2000-12-01 14:52:54 +03:00
rdfs(daddr_t bno, int size, void *bf)
1993-03-21 12:45:37 +03:00
{
int n;
off_t offset;
1993-03-21 12:45:37 +03:00
#ifdef MFS
1993-03-21 12:45:37 +03:00
if (mfs) {
memmove(bf, membase + bno * sectorsize, size);
1993-03-21 12:45:37 +03:00
return;
}
#endif
offset = bno;
offset *= sectorsize;
if (lseek(fsi, offset, SEEK_SET) < 0) {
printf("rdfs: seek error for sector %d: %s\n",
bno, strerror(errno));
1993-03-21 12:45:37 +03:00
exit(33);
}
n = read(fsi, bf, size);
if (n != size) {
printf("rdfs: read error for sector %d: %s\n",
bno, strerror(errno));
1993-03-21 12:45:37 +03:00
exit(34);
}
}
/*
* write a block to the file system
*/
1997-07-01 02:20:30 +04:00
void
2000-12-01 14:52:54 +03:00
wtfs(daddr_t bno, int size, void *bf)
1993-03-21 12:45:37 +03:00
{
int n;
off_t offset;
1993-03-21 12:45:37 +03:00
#ifdef MFS
1993-03-21 12:45:37 +03:00
if (mfs) {
memmove(membase + bno * sectorsize, bf, size);
1993-03-21 12:45:37 +03:00
return;
}
#endif
1993-03-21 12:45:37 +03:00
if (Nflag)
return;
offset = bno;
offset *= sectorsize;
if (lseek(fso, offset, SEEK_SET) < 0) {
printf("wtfs: seek error for sector %d: %s\n",
bno, strerror(errno));
1993-03-21 12:45:37 +03:00
exit(35);
}
n = write(fso, bf, size);
if (n != size) {
printf("wtfs: write error for sector %d: %s\n",
bno, strerror(errno));
1993-03-21 12:45:37 +03:00
exit(36);
}
}
/*
* check if a block is available
*/
1997-07-01 02:20:30 +04:00
int
2000-12-01 14:52:54 +03:00
isblock(struct fs *fs, unsigned char *cp, int h)
1993-03-21 12:45:37 +03:00
{
unsigned char mask;
switch (fs->fs_fragshift) {
case 3:
1993-03-21 12:45:37 +03:00
return (cp[h] == 0xff);
case 2:
1993-03-21 12:45:37 +03:00
mask = 0x0f << ((h & 0x1) << 2);
return ((cp[h >> 1] & mask) == mask);
case 1:
1993-03-21 12:45:37 +03:00
mask = 0x03 << ((h & 0x3) << 1);
return ((cp[h >> 2] & mask) == mask);
case 0:
1993-03-21 12:45:37 +03:00
mask = 0x01 << (h & 0x7);
return ((cp[h >> 3] & mask) == mask);
default:
#ifdef STANDALONE
printf("isblock bad fs_fragshift %d\n", fs->fs_fragshift);
1993-03-21 12:45:37 +03:00
#else
fprintf(stderr, "isblock bad fs_fragshift %d\n",
fs->fs_fragshift);
1993-03-21 12:45:37 +03:00
#endif
return (0);
}
}
/*
* take a block out of the map
*/
1997-07-01 02:20:30 +04:00
void
2000-12-01 14:52:54 +03:00
clrblock(struct fs *fs, unsigned char *cp, int h)
1993-03-21 12:45:37 +03:00
{
switch ((fs)->fs_fragshift) {
case 3:
1993-03-21 12:45:37 +03:00
cp[h] = 0;
return;
case 2:
1993-03-21 12:45:37 +03:00
cp[h >> 1] &= ~(0x0f << ((h & 0x1) << 2));
return;
case 1:
1993-03-21 12:45:37 +03:00
cp[h >> 2] &= ~(0x03 << ((h & 0x3) << 1));
return;
case 0:
1993-03-21 12:45:37 +03:00
cp[h >> 3] &= ~(0x01 << (h & 0x7));
return;
default:
#ifdef STANDALONE
printf("clrblock bad fs_fragshift %d\n", fs->fs_fragshift);
1993-03-21 12:45:37 +03:00
#else
fprintf(stderr, "clrblock bad fs_fragshift %d\n",
fs->fs_fragshift);
1993-03-21 12:45:37 +03:00
#endif
return;
}
}
/*
* put a block into the map
*/
1997-07-01 02:20:30 +04:00
void
2000-12-01 14:52:54 +03:00
setblock(struct fs *fs, unsigned char *cp, int h)
1993-03-21 12:45:37 +03:00
{
switch (fs->fs_fragshift) {
case 3:
1993-03-21 12:45:37 +03:00
cp[h] = 0xff;
return;
case 2:
1993-03-21 12:45:37 +03:00
cp[h >> 1] |= (0x0f << ((h & 0x1) << 2));
return;
case 1:
1993-03-21 12:45:37 +03:00
cp[h >> 2] |= (0x03 << ((h & 0x3) << 1));
return;
case 0:
1993-03-21 12:45:37 +03:00
cp[h >> 3] |= (0x01 << (h & 0x7));
return;
default:
#ifdef STANDALONE
printf("setblock bad fs_frag %d\n", fs->fs_fragshift);
1993-03-21 12:45:37 +03:00
#else
fprintf(stderr, "setblock bad fs_fragshift %d\n",
fs->fs_fragshift);
1993-03-21 12:45:37 +03:00
#endif
return;
}
}
/* swap byte order of cylinder group */
static void
2000-12-01 14:52:54 +03:00
swap_cg(struct cg *o, struct cg *n)
{
int i, btotsize, fbsize;
u_int32_t *n32, *o32;
u_int16_t *n16, *o16;
n->cg_firstfield = bswap32(o->cg_firstfield);
n->cg_magic = bswap32(o->cg_magic);
n->cg_time = bswap32(o->cg_time);
n->cg_cgx = bswap32(o->cg_cgx);
n->cg_ncyl = bswap16(o->cg_ncyl);
n->cg_niblk = bswap16(o->cg_niblk);
n->cg_ndblk = bswap32(o->cg_ndblk);
n->cg_cs.cs_ndir = bswap32(o->cg_cs.cs_ndir);
n->cg_cs.cs_nbfree = bswap32(o->cg_cs.cs_nbfree);
n->cg_cs.cs_nifree = bswap32(o->cg_cs.cs_nifree);
n->cg_cs.cs_nffree = bswap32(o->cg_cs.cs_nffree);
n->cg_rotor = bswap32(o->cg_rotor);
n->cg_frotor = bswap32(o->cg_frotor);
n->cg_irotor = bswap32(o->cg_irotor);
n->cg_btotoff = bswap32(o->cg_btotoff);
n->cg_boff = bswap32(o->cg_boff);
n->cg_iusedoff = bswap32(o->cg_iusedoff);
n->cg_freeoff = bswap32(o->cg_freeoff);
n->cg_nextfreeoff = bswap32(o->cg_nextfreeoff);
n->cg_clustersumoff = bswap32(o->cg_clustersumoff);
n->cg_clusteroff = bswap32(o->cg_clusteroff);
n->cg_nclusterblks = bswap32(o->cg_nclusterblks);
for (i=0; i < MAXFRAG; i++)
n->cg_frsum[i] = bswap32(o->cg_frsum[i]);
/* alays new format */
if (n->cg_magic == CG_MAGIC) {
btotsize = n->cg_boff - n->cg_btotoff;
fbsize = n->cg_iusedoff - n->cg_boff;
n32 = (u_int32_t*)((u_int8_t*)n + n->cg_btotoff);
o32 = (u_int32_t*)((u_int8_t*)o + n->cg_btotoff);
n16 = (u_int16_t*)((u_int8_t*)n + n->cg_boff);
o16 = (u_int16_t*)((u_int8_t*)o + n->cg_boff);
} else {
btotsize = bswap32(n->cg_boff) - bswap32(n->cg_btotoff);
fbsize = bswap32(n->cg_iusedoff) - bswap32(n->cg_boff);
n32 = (u_int32_t*)((u_int8_t*)n + bswap32(n->cg_btotoff));
o32 = (u_int32_t*)((u_int8_t*)o + bswap32(n->cg_btotoff));
n16 = (u_int16_t*)((u_int8_t*)n + bswap32(n->cg_boff));
o16 = (u_int16_t*)((u_int8_t*)o + bswap32(n->cg_boff));
}
for (i=0; i < btotsize / sizeof(u_int32_t); i++)
n32[i] = bswap32(o32[i]);
for (i=0; i < fbsize/sizeof(u_int16_t); i++)
n16[i] = bswap16(o16[i]);
if (n->cg_magic == CG_MAGIC) {
n32 = (u_int32_t*)((u_int8_t*)n + n->cg_clustersumoff);
o32 = (u_int32_t*)((u_int8_t*)o + n->cg_clustersumoff);
} else {
n32 = (u_int32_t*)((u_int8_t*)n + bswap32(n->cg_clustersumoff));
o32 = (u_int32_t*)((u_int8_t*)o + bswap32(n->cg_clustersumoff));
}
for (i = 1; i < sblock.fs_contigsumsize + 1; i++)
n32[i] = bswap32(o32[i]);
}
/* copy a direntry to a buffer, in fs byte order */
static void
2000-12-01 14:52:54 +03:00
copy_dir(struct direct *dir, struct direct *dbuf)
{
memcpy(dbuf, dir, DIRSIZ(Oflag, dir, 0));
if (needswap) {
dbuf->d_ino = bswap32(dir->d_ino);
dbuf->d_reclen = bswap16(dir->d_reclen);
if (Oflag)
((struct odirect*)dbuf)->d_namlen =
bswap16(((struct odirect*)dir)->d_namlen);
}
}
/* Determine how many digits are needed to print a given integer */
static int
2000-12-01 14:52:54 +03:00
count_digits(int num)
{
int ndig;
for(ndig = 1; num > 9; num /=10, ndig++);
return (ndig);
}
#ifdef MFS
/*
* XXX!
* Attempt to guess how much more space is available for process data. The
* heuristic we use is
*
* max_data_limit - (sbrk(0) - etext) - 128kB
*
* etext approximates that start address of the data segment, and the 128kB
* allows some slop for both segment gap between text and data, and for other
* (libc) malloc usage.
*/
static void
calc_memfree(void)
{
extern char etext;
struct rlimit rlp;
u_long base;
base = (u_long)sbrk(0) - (u_long)&etext;
if (getrlimit(RLIMIT_DATA, &rlp) < 0)
perror("getrlimit");
rlp.rlim_cur = rlp.rlim_max;
if (setrlimit(RLIMIT_DATA, &rlp) < 0)
perror("setrlimit");
memleft = rlp.rlim_max - base - (128 * 1024);
}
/*
* Internal version of malloc that trims the requested size if not enough
* memory is available.
*/
static void *
mkfs_malloc(size_t size)
{
u_long pgsz;
if (size == 0)
return (NULL);
if (memleft == 0)
calc_memfree();
pgsz = getpagesize() - 1;
size = (size + pgsz) &~ pgsz;
if (size > memleft)
size = memleft;
memleft -= size;
return (mmap(0, size, PROT_READ|PROT_WRITE, MAP_ANON|MAP_PRIVATE,
-1, 0));
}
#endif /* MFS */