freebsd-dev/sbin/fsck_ffs/dir.c

928 lines
23 KiB
C
Raw Normal View History

/*-
* SPDX-License-Identifier: BSD-3-Clause
*
* 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. 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.
*/
#if 0
#ifndef lint
static const char sccsid[] = "@(#)dir.c 8.8 (Berkeley) 4/28/95";
#endif /* not lint */
#endif
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
#include <sys/param.h>
#include <sys/time.h>
#include <sys/types.h>
#include <sys/sysctl.h>
#include <ufs/ufs/dinode.h>
#include <ufs/ufs/dir.h>
#include <ufs/ffs/fs.h>
#include <err.h>
#include <string.h>
#include "fsck.h"
static struct dirtemplate emptydir = {
0, DIRBLKSIZ, DT_UNKNOWN, 0, "",
0, 0, DT_UNKNOWN, 0, ""
};
static struct dirtemplate dirhead = {
0, 12, DT_DIR, 1, ".",
0, DIRBLKSIZ - 12, DT_DIR, 2, ".."
};
static int chgino(struct inodesc *);
static int dircheck(struct inodesc *, struct bufarea *, struct direct *);
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
static int expanddir(struct inode *ip, char *name);
static void freedir(ino_t ino, ino_t parent);
static struct direct *fsck_readdir(struct inodesc *);
This commit adds basic support for the UFS2 filesystem. The UFS2 filesystem expands the inode to 256 bytes to make space for 64-bit block pointers. It also adds a file-creation time field, an ability to use jumbo blocks per inode to allow extent like pointer density, and space for extended attributes (up to twice the filesystem block size worth of attributes, e.g., on a 16K filesystem, there is space for 32K of attributes). UFS2 fully supports and runs existing UFS1 filesystems. New filesystems built using newfs can be built in either UFS1 or UFS2 format using the -O option. In this commit UFS1 is the default format, so if you want to build UFS2 format filesystems, you must specify -O 2. This default will be changed to UFS2 when UFS2 proves itself to be stable. In this commit the boot code for reading UFS2 filesystems is not compiled (see /sys/boot/common/ufsread.c) as there is insufficient space in the boot block. Once the size of the boot block is increased, this code can be defined. Things to note: the definition of SBSIZE has changed to SBLOCKSIZE. The header file <ufs/ufs/dinode.h> must be included before <ufs/ffs/fs.h> so as to get the definitions of ufs2_daddr_t and ufs_lbn_t. Still TODO: Verify that the first level bootstraps work for all the architectures. Convert the utility ffsinfo to understand UFS2 and test growfs. Add support for the extended attribute storage. Update soft updates to ensure integrity of extended attribute storage. Switch the current extended attribute interfaces to use the extended attribute storage. Add the extent like functionality (framework is there, but is currently never used). Sponsored by: DARPA & NAI Labs. Reviewed by: Poul-Henning Kamp <phk@freebsd.org>
2002-06-21 06:18:05 +00:00
static struct bufarea *getdirblk(ufs2_daddr_t blkno, long size);
static int lftempname(char *bufp, ino_t ino);
static int mkentry(struct inodesc *);
/*
* Propagate connected state through the tree.
*/
void
propagate(void)
{
struct inoinfo **inpp, *inp;
struct inoinfo **inpend;
long change;
inpend = &inpsort[inplast];
do {
change = 0;
for (inpp = inpsort; inpp < inpend; inpp++) {
inp = *inpp;
if (inp->i_parent == 0)
continue;
if (inoinfo(inp->i_parent)->ino_state == DFOUND &&
INO_IS_DUNFOUND(inp->i_number)) {
inoinfo(inp->i_number)->ino_state = DFOUND;
change++;
}
}
} while (change > 0);
}
/*
* Scan each entry in a directory block.
*/
int
dirscan(struct inodesc *idesc)
{
struct direct *dp;
struct bufarea *bp;
u_int dsize, n;
long blksiz;
char dbuf[DIRBLKSIZ];
if (idesc->id_type != DATA)
errx(EEXIT, "wrong type to dirscan %d", idesc->id_type);
if (idesc->id_entryno == 0 &&
(idesc->id_filesize & (DIRBLKSIZ - 1)) != 0)
idesc->id_filesize = roundup(idesc->id_filesize, DIRBLKSIZ);
blksiz = idesc->id_numfrags * sblock.fs_fsize;
if (chkrange(idesc->id_blkno, idesc->id_numfrags)) {
idesc->id_filesize -= blksiz;
return (SKIP);
}
idesc->id_loc = 0;
for (dp = fsck_readdir(idesc); dp != NULL; dp = fsck_readdir(idesc)) {
dsize = dp->d_reclen;
if (dsize > sizeof(dbuf))
dsize = sizeof(dbuf);
memmove(dbuf, dp, (size_t)dsize);
idesc->id_dirp = (struct direct *)dbuf;
if ((n = (*idesc->id_func)(idesc)) & ALTERED) {
bp = getdirblk(idesc->id_blkno, blksiz);
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
if (bp->b_errs != 0)
return (STOP);
memmove(bp->b_un.b_buf + idesc->id_loc - dsize, dbuf,
(size_t)dsize);
dirty(bp);
sbdirty();
}
1995-05-30 06:12:45 +00:00
if (n & STOP)
return (n);
}
return (idesc->id_filesize > 0 ? KEEPON : STOP);
}
/*
* Get and verify the next entry in a directory.
* We also verify that if there is another entry in the block that it is
* valid, so if it is not valid it can be subsumed into the current entry.
*/
static struct direct *
fsck_readdir(struct inodesc *idesc)
{
struct direct *dp, *ndp;
struct bufarea *bp;
long size, blksiz, subsume_ndp;
subsume_ndp = 0;
blksiz = idesc->id_numfrags * sblock.fs_fsize;
if (idesc->id_filesize <= 0 || idesc->id_loc >= blksiz)
return (NULL);
bp = getdirblk(idesc->id_blkno, blksiz);
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
if (bp->b_errs != 0)
return (NULL);
dp = (struct direct *)(bp->b_un.b_buf + idesc->id_loc);
/*
* Only need to check current entry if it is the first in the
* the block, as later entries will have been checked in the
* previous call to this function.
*/
if (idesc->id_loc % DIRBLKSIZ != 0 || dircheck(idesc, bp, dp) != 0) {
/*
* Current entry is good, update to point at next.
*/
idesc->id_loc += dp->d_reclen;
idesc->id_filesize -= dp->d_reclen;
/*
* If at end of directory block, just return this entry.
*/
if (idesc->id_filesize <= 0 || idesc->id_loc >= blksiz ||
idesc->id_loc % DIRBLKSIZ == 0)
return (dp);
/*
* If the next entry good, return this entry.
*/
ndp = (struct direct *)(bp->b_un.b_buf + idesc->id_loc);
if (dircheck(idesc, bp, ndp) != 0)
return (dp);
/*
* The next entry is bad, so subsume it and the remainder
* of this directory block into this entry.
*/
subsume_ndp = 1;
}
/*
* Current or next entry is bad. Zap current entry or
* subsume next entry into current entry as appropriate.
*/
size = DIRBLKSIZ - (idesc->id_loc % DIRBLKSIZ);
idesc->id_loc += size;
idesc->id_filesize -= size;
if (idesc->id_fix == IGNORE)
return (NULL);
if (subsume_ndp) {
memset(ndp, 0, size);
dp->d_reclen += size;
} else {
memset(dp, 0, size);
dp->d_reclen = size;
}
if (dofix(idesc, "DIRECTORY CORRUPTED"))
dirty(bp);
return (dp);
}
/*
* Verify that a directory entry is valid.
* This is a superset of the checks made in the kernel.
This update eliminates a kernel stack disclosure bug in UFS/FFS directory entries that is caused by uninitialized directory entry padding written to the disk. It can be viewed by any user with read access to that directory. Up to 3 bytes of kernel stack are disclosed per file entry, depending on the the amount of padding the kernel needs to pad out the entry to a 32 bit boundry. The offset in the kernel stack that is disclosed is a function of the filename size. Furthermore, if the user can create files in a directory, this 3 byte window can be expanded 3 bytes at a time to a 254 byte window with 75% of the data in that window exposed. The additional exposure is done by removing the entry, creating a new entry with a 4-byte longer name, extracting 3 more bytes by reading the directory, and repeating until a 252 byte name is created. This exploit works in part because the area of the kernel stack that is being disclosed is in an area that typically doesn't change that often (perhaps a few times a second on a lightly loaded system), and these file creates and unlinks themselves don't overwrite the area of kernel stack being disclosed. It appears that this bug originated with the creation of the Fast File System in 4.1b-BSD (Circa 1982, more than 36 years ago!), and is likely present in every Unix or Unix-like system that uses UFS/FFS. Amazingly, nobody noticed until now. This update also adds the -z flag to fsck_ffs to have it scrub the leaked information in the name padding of existing directories. It only needs to be run once on each UFS/FFS filesystem after a patched kernel is installed and running. Submitted by: David G. Lawrence <dg@dglawrence.com> Reviewed by: kib MFC after: 1 week
2019-05-03 21:54:14 +00:00
* Also optionally clears padding and unused directory space.
*
* Returns 0 if the entry is bad, 1 if the entry is good.
*/
static int
dircheck(struct inodesc *idesc, struct bufarea *bp, struct direct *dp)
{
size_t size;
char *cp;
u_int8_t namlen;
This update eliminates a kernel stack disclosure bug in UFS/FFS directory entries that is caused by uninitialized directory entry padding written to the disk. It can be viewed by any user with read access to that directory. Up to 3 bytes of kernel stack are disclosed per file entry, depending on the the amount of padding the kernel needs to pad out the entry to a 32 bit boundry. The offset in the kernel stack that is disclosed is a function of the filename size. Furthermore, if the user can create files in a directory, this 3 byte window can be expanded 3 bytes at a time to a 254 byte window with 75% of the data in that window exposed. The additional exposure is done by removing the entry, creating a new entry with a 4-byte longer name, extracting 3 more bytes by reading the directory, and repeating until a 252 byte name is created. This exploit works in part because the area of the kernel stack that is being disclosed is in an area that typically doesn't change that often (perhaps a few times a second on a lightly loaded system), and these file creates and unlinks themselves don't overwrite the area of kernel stack being disclosed. It appears that this bug originated with the creation of the Fast File System in 4.1b-BSD (Circa 1982, more than 36 years ago!), and is likely present in every Unix or Unix-like system that uses UFS/FFS. Amazingly, nobody noticed until now. This update also adds the -z flag to fsck_ffs to have it scrub the leaked information in the name padding of existing directories. It only needs to be run once on each UFS/FFS filesystem after a patched kernel is installed and running. Submitted by: David G. Lawrence <dg@dglawrence.com> Reviewed by: kib MFC after: 1 week
2019-05-03 21:54:14 +00:00
int spaceleft, modified, unused;
spaceleft = DIRBLKSIZ - (idesc->id_loc % DIRBLKSIZ);
size = DIRSIZ(0, dp);
if (dp->d_reclen == 0 ||
dp->d_reclen > spaceleft ||
dp->d_reclen < size ||
idesc->id_filesize < size ||
This update eliminates a kernel stack disclosure bug in UFS/FFS directory entries that is caused by uninitialized directory entry padding written to the disk. It can be viewed by any user with read access to that directory. Up to 3 bytes of kernel stack are disclosed per file entry, depending on the the amount of padding the kernel needs to pad out the entry to a 32 bit boundry. The offset in the kernel stack that is disclosed is a function of the filename size. Furthermore, if the user can create files in a directory, this 3 byte window can be expanded 3 bytes at a time to a 254 byte window with 75% of the data in that window exposed. The additional exposure is done by removing the entry, creating a new entry with a 4-byte longer name, extracting 3 more bytes by reading the directory, and repeating until a 252 byte name is created. This exploit works in part because the area of the kernel stack that is being disclosed is in an area that typically doesn't change that often (perhaps a few times a second on a lightly loaded system), and these file creates and unlinks themselves don't overwrite the area of kernel stack being disclosed. It appears that this bug originated with the creation of the Fast File System in 4.1b-BSD (Circa 1982, more than 36 years ago!), and is likely present in every Unix or Unix-like system that uses UFS/FFS. Amazingly, nobody noticed until now. This update also adds the -z flag to fsck_ffs to have it scrub the leaked information in the name padding of existing directories. It only needs to be run once on each UFS/FFS filesystem after a patched kernel is installed and running. Submitted by: David G. Lawrence <dg@dglawrence.com> Reviewed by: kib MFC after: 1 week
2019-05-03 21:54:14 +00:00
(dp->d_reclen & (DIR_ROUNDUP - 1)) != 0)
goto bad;
modified = 0;
This update eliminates a kernel stack disclosure bug in UFS/FFS directory entries that is caused by uninitialized directory entry padding written to the disk. It can be viewed by any user with read access to that directory. Up to 3 bytes of kernel stack are disclosed per file entry, depending on the the amount of padding the kernel needs to pad out the entry to a 32 bit boundry. The offset in the kernel stack that is disclosed is a function of the filename size. Furthermore, if the user can create files in a directory, this 3 byte window can be expanded 3 bytes at a time to a 254 byte window with 75% of the data in that window exposed. The additional exposure is done by removing the entry, creating a new entry with a 4-byte longer name, extracting 3 more bytes by reading the directory, and repeating until a 252 byte name is created. This exploit works in part because the area of the kernel stack that is being disclosed is in an area that typically doesn't change that often (perhaps a few times a second on a lightly loaded system), and these file creates and unlinks themselves don't overwrite the area of kernel stack being disclosed. It appears that this bug originated with the creation of the Fast File System in 4.1b-BSD (Circa 1982, more than 36 years ago!), and is likely present in every Unix or Unix-like system that uses UFS/FFS. Amazingly, nobody noticed until now. This update also adds the -z flag to fsck_ffs to have it scrub the leaked information in the name padding of existing directories. It only needs to be run once on each UFS/FFS filesystem after a patched kernel is installed and running. Submitted by: David G. Lawrence <dg@dglawrence.com> Reviewed by: kib MFC after: 1 week
2019-05-03 21:54:14 +00:00
if (dp->d_ino == 0) {
if (!zflag || fswritefd < 0)
return (1);
This update eliminates a kernel stack disclosure bug in UFS/FFS directory entries that is caused by uninitialized directory entry padding written to the disk. It can be viewed by any user with read access to that directory. Up to 3 bytes of kernel stack are disclosed per file entry, depending on the the amount of padding the kernel needs to pad out the entry to a 32 bit boundry. The offset in the kernel stack that is disclosed is a function of the filename size. Furthermore, if the user can create files in a directory, this 3 byte window can be expanded 3 bytes at a time to a 254 byte window with 75% of the data in that window exposed. The additional exposure is done by removing the entry, creating a new entry with a 4-byte longer name, extracting 3 more bytes by reading the directory, and repeating until a 252 byte name is created. This exploit works in part because the area of the kernel stack that is being disclosed is in an area that typically doesn't change that often (perhaps a few times a second on a lightly loaded system), and these file creates and unlinks themselves don't overwrite the area of kernel stack being disclosed. It appears that this bug originated with the creation of the Fast File System in 4.1b-BSD (Circa 1982, more than 36 years ago!), and is likely present in every Unix or Unix-like system that uses UFS/FFS. Amazingly, nobody noticed until now. This update also adds the -z flag to fsck_ffs to have it scrub the leaked information in the name padding of existing directories. It only needs to be run once on each UFS/FFS filesystem after a patched kernel is installed and running. Submitted by: David G. Lawrence <dg@dglawrence.com> Reviewed by: kib MFC after: 1 week
2019-05-03 21:54:14 +00:00
/*
* Special case of an unused directory entry. Normally only
* occurs at the beginning of a directory block when the block
* contains no entries. Other than the first entry in a
* directory block, the kernel coalesces unused space with
* the previous entry by extending its d_reclen. However,
* when cleaning up a directory, fsck may set d_ino to zero
* in the middle of a directory block. If we're clearing out
* directory cruft (-z flag), then make sure that all directory
* space in entries with d_ino == 0 gets fully cleared.
This update eliminates a kernel stack disclosure bug in UFS/FFS directory entries that is caused by uninitialized directory entry padding written to the disk. It can be viewed by any user with read access to that directory. Up to 3 bytes of kernel stack are disclosed per file entry, depending on the the amount of padding the kernel needs to pad out the entry to a 32 bit boundry. The offset in the kernel stack that is disclosed is a function of the filename size. Furthermore, if the user can create files in a directory, this 3 byte window can be expanded 3 bytes at a time to a 254 byte window with 75% of the data in that window exposed. The additional exposure is done by removing the entry, creating a new entry with a 4-byte longer name, extracting 3 more bytes by reading the directory, and repeating until a 252 byte name is created. This exploit works in part because the area of the kernel stack that is being disclosed is in an area that typically doesn't change that often (perhaps a few times a second on a lightly loaded system), and these file creates and unlinks themselves don't overwrite the area of kernel stack being disclosed. It appears that this bug originated with the creation of the Fast File System in 4.1b-BSD (Circa 1982, more than 36 years ago!), and is likely present in every Unix or Unix-like system that uses UFS/FFS. Amazingly, nobody noticed until now. This update also adds the -z flag to fsck_ffs to have it scrub the leaked information in the name padding of existing directories. It only needs to be run once on each UFS/FFS filesystem after a patched kernel is installed and running. Submitted by: David G. Lawrence <dg@dglawrence.com> Reviewed by: kib MFC after: 1 week
2019-05-03 21:54:14 +00:00
*/
if (dp->d_type != 0) {
dp->d_type = 0;
modified = 1;
}
if (dp->d_namlen != 0) {
dp->d_namlen = 0;
modified = 1;
}
unused = dp->d_reclen - __offsetof(struct direct, d_name);
for (cp = dp->d_name; unused > 0; unused--, cp++) {
if (*cp != '\0') {
*cp = '\0';
This update eliminates a kernel stack disclosure bug in UFS/FFS directory entries that is caused by uninitialized directory entry padding written to the disk. It can be viewed by any user with read access to that directory. Up to 3 bytes of kernel stack are disclosed per file entry, depending on the the amount of padding the kernel needs to pad out the entry to a 32 bit boundry. The offset in the kernel stack that is disclosed is a function of the filename size. Furthermore, if the user can create files in a directory, this 3 byte window can be expanded 3 bytes at a time to a 254 byte window with 75% of the data in that window exposed. The additional exposure is done by removing the entry, creating a new entry with a 4-byte longer name, extracting 3 more bytes by reading the directory, and repeating until a 252 byte name is created. This exploit works in part because the area of the kernel stack that is being disclosed is in an area that typically doesn't change that often (perhaps a few times a second on a lightly loaded system), and these file creates and unlinks themselves don't overwrite the area of kernel stack being disclosed. It appears that this bug originated with the creation of the Fast File System in 4.1b-BSD (Circa 1982, more than 36 years ago!), and is likely present in every Unix or Unix-like system that uses UFS/FFS. Amazingly, nobody noticed until now. This update also adds the -z flag to fsck_ffs to have it scrub the leaked information in the name padding of existing directories. It only needs to be run once on each UFS/FFS filesystem after a patched kernel is installed and running. Submitted by: David G. Lawrence <dg@dglawrence.com> Reviewed by: kib MFC after: 1 week
2019-05-03 21:54:14 +00:00
modified = 1;
}
}
if (modified)
dirty(bp);
return (1);
This update eliminates a kernel stack disclosure bug in UFS/FFS directory entries that is caused by uninitialized directory entry padding written to the disk. It can be viewed by any user with read access to that directory. Up to 3 bytes of kernel stack are disclosed per file entry, depending on the the amount of padding the kernel needs to pad out the entry to a 32 bit boundry. The offset in the kernel stack that is disclosed is a function of the filename size. Furthermore, if the user can create files in a directory, this 3 byte window can be expanded 3 bytes at a time to a 254 byte window with 75% of the data in that window exposed. The additional exposure is done by removing the entry, creating a new entry with a 4-byte longer name, extracting 3 more bytes by reading the directory, and repeating until a 252 byte name is created. This exploit works in part because the area of the kernel stack that is being disclosed is in an area that typically doesn't change that often (perhaps a few times a second on a lightly loaded system), and these file creates and unlinks themselves don't overwrite the area of kernel stack being disclosed. It appears that this bug originated with the creation of the Fast File System in 4.1b-BSD (Circa 1982, more than 36 years ago!), and is likely present in every Unix or Unix-like system that uses UFS/FFS. Amazingly, nobody noticed until now. This update also adds the -z flag to fsck_ffs to have it scrub the leaked information in the name padding of existing directories. It only needs to be run once on each UFS/FFS filesystem after a patched kernel is installed and running. Submitted by: David G. Lawrence <dg@dglawrence.com> Reviewed by: kib MFC after: 1 week
2019-05-03 21:54:14 +00:00
}
/*
* The d_type field should not be tested here. A bad type is an error
* in the entry itself but is not a corruption of the directory
* structure itself. So blowing away all the remaining entries in the
* directory block is inappropriate. Rather the type error should be
* checked in pass1 and fixed there.
*
* The name validation should also be done in pass1 although the
* check to see if the name is longer than fits in the space
* allocated for it (i.e., the *cp != '\0' fails after exiting the
* loop below) then it really is a structural error that requires
* the stronger action taken here.
*/
namlen = dp->d_namlen;
if (namlen == 0 || dp->d_type > 15)
goto bad;
for (cp = dp->d_name, size = 0; size < namlen; size++) {
if (*cp == '\0' || *cp++ == '/')
goto bad;
}
if (*cp != '\0')
goto bad;
This update eliminates a kernel stack disclosure bug in UFS/FFS directory entries that is caused by uninitialized directory entry padding written to the disk. It can be viewed by any user with read access to that directory. Up to 3 bytes of kernel stack are disclosed per file entry, depending on the the amount of padding the kernel needs to pad out the entry to a 32 bit boundry. The offset in the kernel stack that is disclosed is a function of the filename size. Furthermore, if the user can create files in a directory, this 3 byte window can be expanded 3 bytes at a time to a 254 byte window with 75% of the data in that window exposed. The additional exposure is done by removing the entry, creating a new entry with a 4-byte longer name, extracting 3 more bytes by reading the directory, and repeating until a 252 byte name is created. This exploit works in part because the area of the kernel stack that is being disclosed is in an area that typically doesn't change that often (perhaps a few times a second on a lightly loaded system), and these file creates and unlinks themselves don't overwrite the area of kernel stack being disclosed. It appears that this bug originated with the creation of the Fast File System in 4.1b-BSD (Circa 1982, more than 36 years ago!), and is likely present in every Unix or Unix-like system that uses UFS/FFS. Amazingly, nobody noticed until now. This update also adds the -z flag to fsck_ffs to have it scrub the leaked information in the name padding of existing directories. It only needs to be run once on each UFS/FFS filesystem after a patched kernel is installed and running. Submitted by: David G. Lawrence <dg@dglawrence.com> Reviewed by: kib MFC after: 1 week
2019-05-03 21:54:14 +00:00
if (zflag && fswritefd >= 0) {
/*
* Clear unused directory entry space, including the d_name
* padding.
*/
/* First figure the number of pad bytes. */
unused = roundup2(namlen + 1, DIR_ROUNDUP) - (namlen + 1);
/* Add in the free space to the end of the record. */
unused += dp->d_reclen - DIRSIZ(0, dp);
/*
* Now clear out the unused space, keeping track if we actually
* changed anything.
*/
for (cp = &dp->d_name[namlen + 1]; unused > 0; unused--, cp++) {
if (*cp != '\0') {
*cp = '\0';
modified = 1;
}
}
if (modified)
dirty(bp);
This update eliminates a kernel stack disclosure bug in UFS/FFS directory entries that is caused by uninitialized directory entry padding written to the disk. It can be viewed by any user with read access to that directory. Up to 3 bytes of kernel stack are disclosed per file entry, depending on the the amount of padding the kernel needs to pad out the entry to a 32 bit boundry. The offset in the kernel stack that is disclosed is a function of the filename size. Furthermore, if the user can create files in a directory, this 3 byte window can be expanded 3 bytes at a time to a 254 byte window with 75% of the data in that window exposed. The additional exposure is done by removing the entry, creating a new entry with a 4-byte longer name, extracting 3 more bytes by reading the directory, and repeating until a 252 byte name is created. This exploit works in part because the area of the kernel stack that is being disclosed is in an area that typically doesn't change that often (perhaps a few times a second on a lightly loaded system), and these file creates and unlinks themselves don't overwrite the area of kernel stack being disclosed. It appears that this bug originated with the creation of the Fast File System in 4.1b-BSD (Circa 1982, more than 36 years ago!), and is likely present in every Unix or Unix-like system that uses UFS/FFS. Amazingly, nobody noticed until now. This update also adds the -z flag to fsck_ffs to have it scrub the leaked information in the name padding of existing directories. It only needs to be run once on each UFS/FFS filesystem after a patched kernel is installed and running. Submitted by: David G. Lawrence <dg@dglawrence.com> Reviewed by: kib MFC after: 1 week
2019-05-03 21:54:14 +00:00
}
return (1);
This update eliminates a kernel stack disclosure bug in UFS/FFS directory entries that is caused by uninitialized directory entry padding written to the disk. It can be viewed by any user with read access to that directory. Up to 3 bytes of kernel stack are disclosed per file entry, depending on the the amount of padding the kernel needs to pad out the entry to a 32 bit boundry. The offset in the kernel stack that is disclosed is a function of the filename size. Furthermore, if the user can create files in a directory, this 3 byte window can be expanded 3 bytes at a time to a 254 byte window with 75% of the data in that window exposed. The additional exposure is done by removing the entry, creating a new entry with a 4-byte longer name, extracting 3 more bytes by reading the directory, and repeating until a 252 byte name is created. This exploit works in part because the area of the kernel stack that is being disclosed is in an area that typically doesn't change that often (perhaps a few times a second on a lightly loaded system), and these file creates and unlinks themselves don't overwrite the area of kernel stack being disclosed. It appears that this bug originated with the creation of the Fast File System in 4.1b-BSD (Circa 1982, more than 36 years ago!), and is likely present in every Unix or Unix-like system that uses UFS/FFS. Amazingly, nobody noticed until now. This update also adds the -z flag to fsck_ffs to have it scrub the leaked information in the name padding of existing directories. It only needs to be run once on each UFS/FFS filesystem after a patched kernel is installed and running. Submitted by: David G. Lawrence <dg@dglawrence.com> Reviewed by: kib MFC after: 1 week
2019-05-03 21:54:14 +00:00
bad:
if (debug)
printf("Bad dir: ino %d reclen %d namlen %d type %d name %s\n",
dp->d_ino, dp->d_reclen, dp->d_namlen, dp->d_type,
dp->d_name);
return (0);
}
void
direrror(ino_t ino, const char *errmesg)
{
fileerror(ino, ino, errmesg);
}
void
fileerror(ino_t cwd, ino_t ino, const char *errmesg)
{
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
struct inode ip;
This commit adds basic support for the UFS2 filesystem. The UFS2 filesystem expands the inode to 256 bytes to make space for 64-bit block pointers. It also adds a file-creation time field, an ability to use jumbo blocks per inode to allow extent like pointer density, and space for extended attributes (up to twice the filesystem block size worth of attributes, e.g., on a 16K filesystem, there is space for 32K of attributes). UFS2 fully supports and runs existing UFS1 filesystems. New filesystems built using newfs can be built in either UFS1 or UFS2 format using the -O option. In this commit UFS1 is the default format, so if you want to build UFS2 format filesystems, you must specify -O 2. This default will be changed to UFS2 when UFS2 proves itself to be stable. In this commit the boot code for reading UFS2 filesystems is not compiled (see /sys/boot/common/ufsread.c) as there is insufficient space in the boot block. Once the size of the boot block is increased, this code can be defined. Things to note: the definition of SBSIZE has changed to SBLOCKSIZE. The header file <ufs/ufs/dinode.h> must be included before <ufs/ffs/fs.h> so as to get the definitions of ufs2_daddr_t and ufs_lbn_t. Still TODO: Verify that the first level bootstraps work for all the architectures. Convert the utility ffsinfo to understand UFS2 and test growfs. Add support for the extended attribute storage. Update soft updates to ensure integrity of extended attribute storage. Switch the current extended attribute interfaces to use the extended attribute storage. Add the extent like functionality (framework is there, but is currently never used). Sponsored by: DARPA & NAI Labs. Reviewed by: Poul-Henning Kamp <phk@freebsd.org>
2002-06-21 06:18:05 +00:00
union dinode *dp;
char pathbuf[MAXPATHLEN + 1];
pwarn("%s ", errmesg);
if (ino < UFS_ROOTINO || ino > maxino) {
pfatal("out-of-range inode number %ju", (uintmax_t)ino);
return;
}
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
ginode(ino, &ip);
dp = ip.i_dp;
prtinode(&ip);
printf("\n");
getpathname(pathbuf, cwd, ino);
if (ftypeok(dp))
pfatal("%s=%s\n",
2018-03-17 12:59:55 +00:00
(DIP(dp, di_mode) & IFMT) == IFDIR ? "DIR" : "FILE",
This commit adds basic support for the UFS2 filesystem. The UFS2 filesystem expands the inode to 256 bytes to make space for 64-bit block pointers. It also adds a file-creation time field, an ability to use jumbo blocks per inode to allow extent like pointer density, and space for extended attributes (up to twice the filesystem block size worth of attributes, e.g., on a 16K filesystem, there is space for 32K of attributes). UFS2 fully supports and runs existing UFS1 filesystems. New filesystems built using newfs can be built in either UFS1 or UFS2 format using the -O option. In this commit UFS1 is the default format, so if you want to build UFS2 format filesystems, you must specify -O 2. This default will be changed to UFS2 when UFS2 proves itself to be stable. In this commit the boot code for reading UFS2 filesystems is not compiled (see /sys/boot/common/ufsread.c) as there is insufficient space in the boot block. Once the size of the boot block is increased, this code can be defined. Things to note: the definition of SBSIZE has changed to SBLOCKSIZE. The header file <ufs/ufs/dinode.h> must be included before <ufs/ffs/fs.h> so as to get the definitions of ufs2_daddr_t and ufs_lbn_t. Still TODO: Verify that the first level bootstraps work for all the architectures. Convert the utility ffsinfo to understand UFS2 and test growfs. Add support for the extended attribute storage. Update soft updates to ensure integrity of extended attribute storage. Switch the current extended attribute interfaces to use the extended attribute storage. Add the extent like functionality (framework is there, but is currently never used). Sponsored by: DARPA & NAI Labs. Reviewed by: Poul-Henning Kamp <phk@freebsd.org>
2002-06-21 06:18:05 +00:00
pathbuf);
else
pfatal("NAME=%s\n", pathbuf);
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
irelse(&ip);
}
void
adjust(struct inodesc *idesc, int lcnt)
{
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
struct inode ip;
This commit adds basic support for the UFS2 filesystem. The UFS2 filesystem expands the inode to 256 bytes to make space for 64-bit block pointers. It also adds a file-creation time field, an ability to use jumbo blocks per inode to allow extent like pointer density, and space for extended attributes (up to twice the filesystem block size worth of attributes, e.g., on a 16K filesystem, there is space for 32K of attributes). UFS2 fully supports and runs existing UFS1 filesystems. New filesystems built using newfs can be built in either UFS1 or UFS2 format using the -O option. In this commit UFS1 is the default format, so if you want to build UFS2 format filesystems, you must specify -O 2. This default will be changed to UFS2 when UFS2 proves itself to be stable. In this commit the boot code for reading UFS2 filesystems is not compiled (see /sys/boot/common/ufsread.c) as there is insufficient space in the boot block. Once the size of the boot block is increased, this code can be defined. Things to note: the definition of SBSIZE has changed to SBLOCKSIZE. The header file <ufs/ufs/dinode.h> must be included before <ufs/ffs/fs.h> so as to get the definitions of ufs2_daddr_t and ufs_lbn_t. Still TODO: Verify that the first level bootstraps work for all the architectures. Convert the utility ffsinfo to understand UFS2 and test growfs. Add support for the extended attribute storage. Update soft updates to ensure integrity of extended attribute storage. Switch the current extended attribute interfaces to use the extended attribute storage. Add the extent like functionality (framework is there, but is currently never used). Sponsored by: DARPA & NAI Labs. Reviewed by: Poul-Henning Kamp <phk@freebsd.org>
2002-06-21 06:18:05 +00:00
union dinode *dp;
int saveresolved;
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
ginode(idesc->id_number, &ip);
dp = ip.i_dp;
This commit adds basic support for the UFS2 filesystem. The UFS2 filesystem expands the inode to 256 bytes to make space for 64-bit block pointers. It also adds a file-creation time field, an ability to use jumbo blocks per inode to allow extent like pointer density, and space for extended attributes (up to twice the filesystem block size worth of attributes, e.g., on a 16K filesystem, there is space for 32K of attributes). UFS2 fully supports and runs existing UFS1 filesystems. New filesystems built using newfs can be built in either UFS1 or UFS2 format using the -O option. In this commit UFS1 is the default format, so if you want to build UFS2 format filesystems, you must specify -O 2. This default will be changed to UFS2 when UFS2 proves itself to be stable. In this commit the boot code for reading UFS2 filesystems is not compiled (see /sys/boot/common/ufsread.c) as there is insufficient space in the boot block. Once the size of the boot block is increased, this code can be defined. Things to note: the definition of SBSIZE has changed to SBLOCKSIZE. The header file <ufs/ufs/dinode.h> must be included before <ufs/ffs/fs.h> so as to get the definitions of ufs2_daddr_t and ufs_lbn_t. Still TODO: Verify that the first level bootstraps work for all the architectures. Convert the utility ffsinfo to understand UFS2 and test growfs. Add support for the extended attribute storage. Update soft updates to ensure integrity of extended attribute storage. Switch the current extended attribute interfaces to use the extended attribute storage. Add the extent like functionality (framework is there, but is currently never used). Sponsored by: DARPA & NAI Labs. Reviewed by: Poul-Henning Kamp <phk@freebsd.org>
2002-06-21 06:18:05 +00:00
if (DIP(dp, di_nlink) == lcnt) {
/*
* If we have not hit any unresolved problems, are running
* in preen mode, and are on a file system using soft updates,
* then just toss any partially allocated files.
*/
if (resolved && (preen || bkgrdflag) && usedsoftdep) {
clri(idesc, "UNREF", 1);
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
irelse(&ip);
return;
} else {
/*
* The file system can be marked clean even if
* a file is not linked up, but is cleared.
* Hence, resolved should not be cleared when
* linkup is answered no, but clri is answered yes.
*/
saveresolved = resolved;
if (linkup(idesc->id_number, (ino_t)0, NULL) == 0) {
resolved = saveresolved;
clri(idesc, "UNREF", 0);
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
irelse(&ip);
return;
}
/*
* Account for the new reference created by linkup().
*/
lcnt--;
}
}
if (lcnt != 0) {
pwarn("LINK COUNT %s", (lfdir == idesc->id_number) ? lfname :
2018-03-17 12:59:55 +00:00
((DIP(dp, di_mode) & IFMT) == IFDIR ? "DIR" : "FILE"));
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
prtinode(&ip);
printf(" COUNT %d SHOULD BE %d",
This commit adds basic support for the UFS2 filesystem. The UFS2 filesystem expands the inode to 256 bytes to make space for 64-bit block pointers. It also adds a file-creation time field, an ability to use jumbo blocks per inode to allow extent like pointer density, and space for extended attributes (up to twice the filesystem block size worth of attributes, e.g., on a 16K filesystem, there is space for 32K of attributes). UFS2 fully supports and runs existing UFS1 filesystems. New filesystems built using newfs can be built in either UFS1 or UFS2 format using the -O option. In this commit UFS1 is the default format, so if you want to build UFS2 format filesystems, you must specify -O 2. This default will be changed to UFS2 when UFS2 proves itself to be stable. In this commit the boot code for reading UFS2 filesystems is not compiled (see /sys/boot/common/ufsread.c) as there is insufficient space in the boot block. Once the size of the boot block is increased, this code can be defined. Things to note: the definition of SBSIZE has changed to SBLOCKSIZE. The header file <ufs/ufs/dinode.h> must be included before <ufs/ffs/fs.h> so as to get the definitions of ufs2_daddr_t and ufs_lbn_t. Still TODO: Verify that the first level bootstraps work for all the architectures. Convert the utility ffsinfo to understand UFS2 and test growfs. Add support for the extended attribute storage. Update soft updates to ensure integrity of extended attribute storage. Switch the current extended attribute interfaces to use the extended attribute storage. Add the extent like functionality (framework is there, but is currently never used). Sponsored by: DARPA & NAI Labs. Reviewed by: Poul-Henning Kamp <phk@freebsd.org>
2002-06-21 06:18:05 +00:00
DIP(dp, di_nlink), DIP(dp, di_nlink) - lcnt);
if (preen || usedsoftdep) {
if (lcnt < 0) {
printf("\n");
pfatal("LINK COUNT INCREASING");
}
if (preen)
printf(" (ADJUSTED)\n");
}
if (preen || reply("ADJUST") == 1) {
if (bkgrdflag == 0) {
DIP_SET(dp, di_nlink, DIP(dp, di_nlink) - lcnt);
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
inodirty(&ip);
} else {
cmd.value = idesc->id_number;
cmd.size = -lcnt;
if (debug)
printf("adjrefcnt ino %ld amt %lld\n",
(long)cmd.value,
(long long)cmd.size);
if (sysctl(adjrefcnt, MIBSIZE, 0, 0,
&cmd, sizeof cmd) == -1)
rwerror("ADJUST INODE", cmd.value);
}
}
}
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
irelse(&ip);
}
static int
mkentry(struct inodesc *idesc)
{
struct direct *dirp = idesc->id_dirp;
struct direct newent;
int newlen, oldlen;
newent.d_namlen = strlen(idesc->id_name);
newlen = DIRSIZ(0, &newent);
if (dirp->d_ino != 0)
oldlen = DIRSIZ(0, dirp);
else
oldlen = 0;
if (dirp->d_reclen - oldlen < newlen)
return (KEEPON);
newent.d_reclen = dirp->d_reclen - oldlen;
dirp->d_reclen = oldlen;
dirp = (struct direct *)(((char *)dirp) + oldlen);
dirp->d_ino = idesc->id_parent; /* ino to be entered is in id_parent */
dirp->d_reclen = newent.d_reclen;
dirp->d_type = inoinfo(idesc->id_parent)->ino_type;
dirp->d_namlen = newent.d_namlen;
memmove(dirp->d_name, idesc->id_name, (size_t)newent.d_namlen + 1);
return (ALTERED|STOP);
}
static int
chgino(struct inodesc *idesc)
{
struct direct *dirp = idesc->id_dirp;
if (memcmp(dirp->d_name, idesc->id_name, (int)dirp->d_namlen + 1))
return (KEEPON);
dirp->d_ino = idesc->id_parent;
dirp->d_type = inoinfo(idesc->id_parent)->ino_type;
return (ALTERED|STOP);
}
int
linkup(ino_t orphan, ino_t parentdir, char *name)
{
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
struct inode ip;
This commit adds basic support for the UFS2 filesystem. The UFS2 filesystem expands the inode to 256 bytes to make space for 64-bit block pointers. It also adds a file-creation time field, an ability to use jumbo blocks per inode to allow extent like pointer density, and space for extended attributes (up to twice the filesystem block size worth of attributes, e.g., on a 16K filesystem, there is space for 32K of attributes). UFS2 fully supports and runs existing UFS1 filesystems. New filesystems built using newfs can be built in either UFS1 or UFS2 format using the -O option. In this commit UFS1 is the default format, so if you want to build UFS2 format filesystems, you must specify -O 2. This default will be changed to UFS2 when UFS2 proves itself to be stable. In this commit the boot code for reading UFS2 filesystems is not compiled (see /sys/boot/common/ufsread.c) as there is insufficient space in the boot block. Once the size of the boot block is increased, this code can be defined. Things to note: the definition of SBSIZE has changed to SBLOCKSIZE. The header file <ufs/ufs/dinode.h> must be included before <ufs/ffs/fs.h> so as to get the definitions of ufs2_daddr_t and ufs_lbn_t. Still TODO: Verify that the first level bootstraps work for all the architectures. Convert the utility ffsinfo to understand UFS2 and test growfs. Add support for the extended attribute storage. Update soft updates to ensure integrity of extended attribute storage. Switch the current extended attribute interfaces to use the extended attribute storage. Add the extent like functionality (framework is there, but is currently never used). Sponsored by: DARPA & NAI Labs. Reviewed by: Poul-Henning Kamp <phk@freebsd.org>
2002-06-21 06:18:05 +00:00
union dinode *dp;
int lostdir;
ino_t oldlfdir;
struct inoinfo *inp;
struct inodesc idesc;
char tempname[BUFSIZ];
memset(&idesc, 0, sizeof(struct inodesc));
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
ginode(orphan, &ip);
dp = ip.i_dp;
2018-03-17 12:59:55 +00:00
lostdir = (DIP(dp, di_mode) & IFMT) == IFDIR;
pwarn("UNREF %s ", lostdir ? "DIR" : "FILE");
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
prtinode(&ip);
printf("\n");
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
if (preen && DIP(dp, di_size) == 0) {
irelse(&ip);
return (0);
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
}
irelse(&ip);
if (cursnapshot != 0) {
pfatal("FILE LINKUP IN SNAPSHOT");
return (0);
}
if (preen)
printf(" (RECONNECTED)\n");
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
else if (reply("RECONNECT") == 0)
return (0);
if (lfdir == 0) {
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
ginode(UFS_ROOTINO, &ip);
idesc.id_name = strdup(lfname);
idesc.id_type = DATA;
idesc.id_func = findino;
idesc.id_number = UFS_ROOTINO;
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
if ((ckinode(ip.i_dp, &idesc) & FOUND) != 0) {
lfdir = idesc.id_parent;
} else {
pwarn("NO lost+found DIRECTORY");
if (preen || reply("CREATE")) {
lfdir = allocdir(UFS_ROOTINO, (ino_t)0, lfmode);
if (lfdir != 0) {
if (makeentry(UFS_ROOTINO, lfdir,
lfname) != 0) {
numdirs++;
if (preen)
printf(" (CREATED)\n");
} else {
freedir(lfdir, UFS_ROOTINO);
lfdir = 0;
if (preen)
printf("\n");
}
}
}
}
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
irelse(&ip);
if (lfdir == 0) {
pfatal("SORRY. CANNOT CREATE lost+found DIRECTORY");
printf("\n\n");
return (0);
}
}
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
ginode(lfdir, &ip);
dp = ip.i_dp;
2018-03-17 12:59:55 +00:00
if ((DIP(dp, di_mode) & IFMT) != IFDIR) {
pfatal("lost+found IS NOT A DIRECTORY");
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
if (reply("REALLOCATE") == 0) {
irelse(&ip);
return (0);
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
}
oldlfdir = lfdir;
if ((lfdir = allocdir(UFS_ROOTINO, (ino_t)0, lfmode)) == 0) {
pfatal("SORRY. CANNOT CREATE lost+found DIRECTORY\n\n");
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
irelse(&ip);
return (0);
}
if ((changeino(UFS_ROOTINO, lfname, lfdir) & ALTERED) == 0) {
pfatal("SORRY. CANNOT CREATE lost+found DIRECTORY\n\n");
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
irelse(&ip);
return (0);
}
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
idesc.id_type = inoinfo(oldlfdir)->ino_idtype;
idesc.id_func = freeblock;
idesc.id_number = oldlfdir;
adjust(&idesc, inoinfo(oldlfdir)->ino_linkcnt + 1);
inoinfo(oldlfdir)->ino_linkcnt = 0;
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
inodirty(&ip);
irelse(&ip);
ginode(lfdir, &ip);
dp = ip.i_dp;
}
if (inoinfo(lfdir)->ino_state != DFOUND) {
pfatal("SORRY. NO lost+found DIRECTORY\n\n");
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
irelse(&ip);
return (0);
}
(void)lftempname(tempname, orphan);
if (makeentry(lfdir, orphan, (name ? name : tempname)) == 0) {
pfatal("SORRY. NO SPACE IN lost+found DIRECTORY");
printf("\n\n");
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
irelse(&ip);
return (0);
}
inoinfo(orphan)->ino_linkcnt--;
if (lostdir) {
if ((changeino(orphan, "..", lfdir) & ALTERED) == 0 &&
parentdir != (ino_t)-1)
(void)makeentry(orphan, lfdir, "..");
DIP_SET(dp, di_nlink, DIP(dp, di_nlink) + 1);
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
inodirty(&ip);
inoinfo(lfdir)->ino_linkcnt++;
pwarn("DIR I=%lu CONNECTED. ", (u_long)orphan);
inp = getinoinfo(parentdir);
if (parentdir != (ino_t)-1 && inp != NULL &&
(inp->i_flags & INFO_NEW) == 0) {
1998-06-28 19:23:03 +00:00
printf("PARENT WAS I=%lu\n", (u_long)parentdir);
/*
* If the parent directory did not have to
* be replaced then because of the ordering
* guarantees, has had the link count incremented
* for the child, but no entry was made. This
* fixes the parent link count so that fsck does
* not need to be rerun.
*/
inoinfo(parentdir)->ino_linkcnt++;
}
if (preen == 0)
printf("\n");
}
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
irelse(&ip);
return (1);
}
/*
* fix an entry in a directory.
*/
int
changeino(ino_t dir, const char *name, ino_t newnum)
{
struct inodesc idesc;
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
struct inode ip;
int error;
memset(&idesc, 0, sizeof(struct inodesc));
idesc.id_type = DATA;
idesc.id_func = chgino;
idesc.id_number = dir;
idesc.id_fix = DONTKNOW;
idesc.id_name = strdup(name);
idesc.id_parent = newnum; /* new value for name */
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
ginode(dir, &ip);
error = ckinode(ip.i_dp, &idesc);
irelse(&ip);
return (error);
}
/*
* make an entry in a directory
*/
int
makeentry(ino_t parent, ino_t ino, const char *name)
{
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
struct inode ip;
This commit adds basic support for the UFS2 filesystem. The UFS2 filesystem expands the inode to 256 bytes to make space for 64-bit block pointers. It also adds a file-creation time field, an ability to use jumbo blocks per inode to allow extent like pointer density, and space for extended attributes (up to twice the filesystem block size worth of attributes, e.g., on a 16K filesystem, there is space for 32K of attributes). UFS2 fully supports and runs existing UFS1 filesystems. New filesystems built using newfs can be built in either UFS1 or UFS2 format using the -O option. In this commit UFS1 is the default format, so if you want to build UFS2 format filesystems, you must specify -O 2. This default will be changed to UFS2 when UFS2 proves itself to be stable. In this commit the boot code for reading UFS2 filesystems is not compiled (see /sys/boot/common/ufsread.c) as there is insufficient space in the boot block. Once the size of the boot block is increased, this code can be defined. Things to note: the definition of SBSIZE has changed to SBLOCKSIZE. The header file <ufs/ufs/dinode.h> must be included before <ufs/ffs/fs.h> so as to get the definitions of ufs2_daddr_t and ufs_lbn_t. Still TODO: Verify that the first level bootstraps work for all the architectures. Convert the utility ffsinfo to understand UFS2 and test growfs. Add support for the extended attribute storage. Update soft updates to ensure integrity of extended attribute storage. Switch the current extended attribute interfaces to use the extended attribute storage. Add the extent like functionality (framework is there, but is currently never used). Sponsored by: DARPA & NAI Labs. Reviewed by: Poul-Henning Kamp <phk@freebsd.org>
2002-06-21 06:18:05 +00:00
union dinode *dp;
struct inodesc idesc;
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
int retval;
char pathbuf[MAXPATHLEN + 1];
1995-05-30 06:12:45 +00:00
if (parent < UFS_ROOTINO || parent >= maxino ||
ino < UFS_ROOTINO || ino >= maxino)
return (0);
memset(&idesc, 0, sizeof(struct inodesc));
idesc.id_type = DATA;
idesc.id_func = mkentry;
idesc.id_number = parent;
idesc.id_parent = ino; /* this is the inode to enter */
idesc.id_fix = DONTKNOW;
idesc.id_name = strdup(name);
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
ginode(parent, &ip);
dp = ip.i_dp;
This commit adds basic support for the UFS2 filesystem. The UFS2 filesystem expands the inode to 256 bytes to make space for 64-bit block pointers. It also adds a file-creation time field, an ability to use jumbo blocks per inode to allow extent like pointer density, and space for extended attributes (up to twice the filesystem block size worth of attributes, e.g., on a 16K filesystem, there is space for 32K of attributes). UFS2 fully supports and runs existing UFS1 filesystems. New filesystems built using newfs can be built in either UFS1 or UFS2 format using the -O option. In this commit UFS1 is the default format, so if you want to build UFS2 format filesystems, you must specify -O 2. This default will be changed to UFS2 when UFS2 proves itself to be stable. In this commit the boot code for reading UFS2 filesystems is not compiled (see /sys/boot/common/ufsread.c) as there is insufficient space in the boot block. Once the size of the boot block is increased, this code can be defined. Things to note: the definition of SBSIZE has changed to SBLOCKSIZE. The header file <ufs/ufs/dinode.h> must be included before <ufs/ffs/fs.h> so as to get the definitions of ufs2_daddr_t and ufs_lbn_t. Still TODO: Verify that the first level bootstraps work for all the architectures. Convert the utility ffsinfo to understand UFS2 and test growfs. Add support for the extended attribute storage. Update soft updates to ensure integrity of extended attribute storage. Switch the current extended attribute interfaces to use the extended attribute storage. Add the extent like functionality (framework is there, but is currently never used). Sponsored by: DARPA & NAI Labs. Reviewed by: Poul-Henning Kamp <phk@freebsd.org>
2002-06-21 06:18:05 +00:00
if (DIP(dp, di_size) % DIRBLKSIZ) {
DIP_SET(dp, di_size, roundup(DIP(dp, di_size), DIRBLKSIZ));
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
inodirty(&ip);
}
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
if ((ckinode(dp, &idesc) & ALTERED) != 0) {
irelse(&ip);
return (1);
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
}
getpathname(pathbuf, parent, parent);
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
if (expanddir(&ip, pathbuf) == 0) {
irelse(&ip);
return (0);
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
}
retval = ckinode(dp, &idesc) & ALTERED;
irelse(&ip);
return (retval);
}
/*
* Attempt to expand the size of a directory
*/
static int
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
expanddir(struct inode *ip, char *name)
{
ufs2_daddr_t lastlbn, oldblk, newblk, indirblk;
size_t filesize, lastlbnsize;
struct bufarea *bp, *nbp;
struct inodesc idesc;
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
union dinode *dp;
Add support for managing UFS/FFS snapshots to fsck_ffs(8). The kernel handles the managment of UFS/FFS snapshots. Since UFS/FFS updates filesystem data (rather than always writing changes to new locations like ZFS), the kernel must check every filesystem write to see if the block being written is part of a snapshot. If it is part of a snapshot, then the kernel must make a copy of the old block value into a newly allocated block for the snapshot before allowing the write to be done. Similarly, if a block is being freed, the kernel must check to see if it is part of a snapshot and let the snapshot claim the block rather than freeing it for future use. When a snapshot is freed, its blocks need to be offered to older snapshots and freed only if no older snapshots wish to claim them. When snapshots were added to UFS/FFS they were integrated into soft updates and just a small part of the management of snapshots needed to be added to fsck_ffs(8) as soft updates minimized the set of snapshot changes that might need correction. When journaling was added to soft updates a much more complete knowledge of snapshots needed to be added to fsck_ffs(8) for it to be able to properly handle the filesystem changes that a journal rollback needs to do (specifically the freeing and allocation of blocks). Since this functionality was unavailable, the use of snapshots was disabled when running with journaled soft updates. This set of changes imports the kernel code for the management of snapshots to fsck_ffs(8). With this code in place it will become possible to enable snapshots when running with journalled soft updates. The most immediate benefit will be the ability to use snapshots to take consistent filesystem dumps on live filesystems. Future work will be done to update fsck_ffs(8) to be able to use snapshots to run in background on live filesystems running with journaled soft updates. Reviewed by: kib Tested by: Peter Holm Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D36491
2022-11-09 18:44:03 +00:00
long cg, indiralloced;
char *cp;
nbp = NULL;
indiralloced = newblk = indirblk = 0;
Add support for managing UFS/FFS snapshots to fsck_ffs(8). The kernel handles the managment of UFS/FFS snapshots. Since UFS/FFS updates filesystem data (rather than always writing changes to new locations like ZFS), the kernel must check every filesystem write to see if the block being written is part of a snapshot. If it is part of a snapshot, then the kernel must make a copy of the old block value into a newly allocated block for the snapshot before allowing the write to be done. Similarly, if a block is being freed, the kernel must check to see if it is part of a snapshot and let the snapshot claim the block rather than freeing it for future use. When a snapshot is freed, its blocks need to be offered to older snapshots and freed only if no older snapshots wish to claim them. When snapshots were added to UFS/FFS they were integrated into soft updates and just a small part of the management of snapshots needed to be added to fsck_ffs(8) as soft updates minimized the set of snapshot changes that might need correction. When journaling was added to soft updates a much more complete knowledge of snapshots needed to be added to fsck_ffs(8) for it to be able to properly handle the filesystem changes that a journal rollback needs to do (specifically the freeing and allocation of blocks). Since this functionality was unavailable, the use of snapshots was disabled when running with journaled soft updates. This set of changes imports the kernel code for the management of snapshots to fsck_ffs(8). With this code in place it will become possible to enable snapshots when running with journalled soft updates. The most immediate benefit will be the ability to use snapshots to take consistent filesystem dumps on live filesystems. Future work will be done to update fsck_ffs(8) to be able to use snapshots to run in background on live filesystems running with journaled soft updates. Reviewed by: kib Tested by: Peter Holm Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D36491
2022-11-09 18:44:03 +00:00
memset(&idesc, 0, sizeof(struct inodesc));
idesc.id_type = ADDR;
pwarn("NO SPACE LEFT IN %s", name);
if (!preen && reply("EXPAND") == 0)
return (0);
Add support for managing UFS/FFS snapshots to fsck_ffs(8). The kernel handles the managment of UFS/FFS snapshots. Since UFS/FFS updates filesystem data (rather than always writing changes to new locations like ZFS), the kernel must check every filesystem write to see if the block being written is part of a snapshot. If it is part of a snapshot, then the kernel must make a copy of the old block value into a newly allocated block for the snapshot before allowing the write to be done. Similarly, if a block is being freed, the kernel must check to see if it is part of a snapshot and let the snapshot claim the block rather than freeing it for future use. When a snapshot is freed, its blocks need to be offered to older snapshots and freed only if no older snapshots wish to claim them. When snapshots were added to UFS/FFS they were integrated into soft updates and just a small part of the management of snapshots needed to be added to fsck_ffs(8) as soft updates minimized the set of snapshot changes that might need correction. When journaling was added to soft updates a much more complete knowledge of snapshots needed to be added to fsck_ffs(8) for it to be able to properly handle the filesystem changes that a journal rollback needs to do (specifically the freeing and allocation of blocks). Since this functionality was unavailable, the use of snapshots was disabled when running with journaled soft updates. This set of changes imports the kernel code for the management of snapshots to fsck_ffs(8). With this code in place it will become possible to enable snapshots when running with journalled soft updates. The most immediate benefit will be the ability to use snapshots to take consistent filesystem dumps on live filesystems. Future work will be done to update fsck_ffs(8) to be able to use snapshots to run in background on live filesystems running with journaled soft updates. Reviewed by: kib Tested by: Peter Holm Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D36491
2022-11-09 18:44:03 +00:00
cg = ino_to_cg(&sblock, ip->i_number);
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
dp = ip->i_dp;
filesize = DIP(dp, di_size);
lastlbn = lblkno(&sblock, filesize);
/*
* We only expand lost+found to a single indirect block.
*/
if ((DIP(dp, di_mode) & IFMT) != IFDIR || filesize == 0 ||
lastlbn >= UFS_NDADDR + NINDIR(&sblock))
goto bad;
/*
* If last block is a fragment, expand it to a full size block.
*/
lastlbnsize = sblksize(&sblock, filesize, lastlbn);
if (lastlbnsize > 0 && lastlbnsize < sblock.fs_bsize) {
oldblk = DIP(dp, di_db[lastlbn]);
bp = getdirblk(oldblk, lastlbnsize);
if (bp->b_errs)
goto bad;
Add support for managing UFS/FFS snapshots to fsck_ffs(8). The kernel handles the managment of UFS/FFS snapshots. Since UFS/FFS updates filesystem data (rather than always writing changes to new locations like ZFS), the kernel must check every filesystem write to see if the block being written is part of a snapshot. If it is part of a snapshot, then the kernel must make a copy of the old block value into a newly allocated block for the snapshot before allowing the write to be done. Similarly, if a block is being freed, the kernel must check to see if it is part of a snapshot and let the snapshot claim the block rather than freeing it for future use. When a snapshot is freed, its blocks need to be offered to older snapshots and freed only if no older snapshots wish to claim them. When snapshots were added to UFS/FFS they were integrated into soft updates and just a small part of the management of snapshots needed to be added to fsck_ffs(8) as soft updates minimized the set of snapshot changes that might need correction. When journaling was added to soft updates a much more complete knowledge of snapshots needed to be added to fsck_ffs(8) for it to be able to properly handle the filesystem changes that a journal rollback needs to do (specifically the freeing and allocation of blocks). Since this functionality was unavailable, the use of snapshots was disabled when running with journaled soft updates. This set of changes imports the kernel code for the management of snapshots to fsck_ffs(8). With this code in place it will become possible to enable snapshots when running with journalled soft updates. The most immediate benefit will be the ability to use snapshots to take consistent filesystem dumps on live filesystems. Future work will be done to update fsck_ffs(8) to be able to use snapshots to run in background on live filesystems running with journaled soft updates. Reviewed by: kib Tested by: Peter Holm Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D36491
2022-11-09 18:44:03 +00:00
newblk = allocblk(cg, sblock.fs_frag, std_checkblkavail);
if (newblk == 0)
goto bad;
nbp = getdatablk(newblk, sblock.fs_bsize, BT_DIRDATA);
if (nbp->b_errs)
goto bad;
DIP_SET(dp, di_db[lastlbn], newblk);
DIP_SET(dp, di_size, filesize + sblock.fs_bsize - lastlbnsize);
DIP_SET(dp, di_blocks, DIP(dp, di_blocks) +
btodb(sblock.fs_bsize - lastlbnsize));
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
inodirty(ip);
memmove(nbp->b_un.b_buf, bp->b_un.b_buf, lastlbnsize);
memset(&nbp->b_un.b_buf[lastlbnsize], 0,
sblock.fs_bsize - lastlbnsize);
for (cp = &nbp->b_un.b_buf[lastlbnsize];
cp < &nbp->b_un.b_buf[sblock.fs_bsize];
cp += DIRBLKSIZ)
memmove(cp, &emptydir, sizeof emptydir);
dirty(nbp);
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
brelse(nbp);
Add support for managing UFS/FFS snapshots to fsck_ffs(8). The kernel handles the managment of UFS/FFS snapshots. Since UFS/FFS updates filesystem data (rather than always writing changes to new locations like ZFS), the kernel must check every filesystem write to see if the block being written is part of a snapshot. If it is part of a snapshot, then the kernel must make a copy of the old block value into a newly allocated block for the snapshot before allowing the write to be done. Similarly, if a block is being freed, the kernel must check to see if it is part of a snapshot and let the snapshot claim the block rather than freeing it for future use. When a snapshot is freed, its blocks need to be offered to older snapshots and freed only if no older snapshots wish to claim them. When snapshots were added to UFS/FFS they were integrated into soft updates and just a small part of the management of snapshots needed to be added to fsck_ffs(8) as soft updates minimized the set of snapshot changes that might need correction. When journaling was added to soft updates a much more complete knowledge of snapshots needed to be added to fsck_ffs(8) for it to be able to properly handle the filesystem changes that a journal rollback needs to do (specifically the freeing and allocation of blocks). Since this functionality was unavailable, the use of snapshots was disabled when running with journaled soft updates. This set of changes imports the kernel code for the management of snapshots to fsck_ffs(8). With this code in place it will become possible to enable snapshots when running with journalled soft updates. The most immediate benefit will be the ability to use snapshots to take consistent filesystem dumps on live filesystems. Future work will be done to update fsck_ffs(8) to be able to use snapshots to run in background on live filesystems running with journaled soft updates. Reviewed by: kib Tested by: Peter Holm Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D36491
2022-11-09 18:44:03 +00:00
binval(bp);
idesc.id_blkno = oldblk;
idesc.id_numfrags = numfrags(&sblock, lastlbnsize);
(void)freeblock(&idesc);
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
if (preen)
printf(" (EXPANDED)\n");
return (1);
}
Add support for managing UFS/FFS snapshots to fsck_ffs(8). The kernel handles the managment of UFS/FFS snapshots. Since UFS/FFS updates filesystem data (rather than always writing changes to new locations like ZFS), the kernel must check every filesystem write to see if the block being written is part of a snapshot. If it is part of a snapshot, then the kernel must make a copy of the old block value into a newly allocated block for the snapshot before allowing the write to be done. Similarly, if a block is being freed, the kernel must check to see if it is part of a snapshot and let the snapshot claim the block rather than freeing it for future use. When a snapshot is freed, its blocks need to be offered to older snapshots and freed only if no older snapshots wish to claim them. When snapshots were added to UFS/FFS they were integrated into soft updates and just a small part of the management of snapshots needed to be added to fsck_ffs(8) as soft updates minimized the set of snapshot changes that might need correction. When journaling was added to soft updates a much more complete knowledge of snapshots needed to be added to fsck_ffs(8) for it to be able to properly handle the filesystem changes that a journal rollback needs to do (specifically the freeing and allocation of blocks). Since this functionality was unavailable, the use of snapshots was disabled when running with journaled soft updates. This set of changes imports the kernel code for the management of snapshots to fsck_ffs(8). With this code in place it will become possible to enable snapshots when running with journalled soft updates. The most immediate benefit will be the ability to use snapshots to take consistent filesystem dumps on live filesystems. Future work will be done to update fsck_ffs(8) to be able to use snapshots to run in background on live filesystems running with journaled soft updates. Reviewed by: kib Tested by: Peter Holm Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D36491
2022-11-09 18:44:03 +00:00
if ((newblk = allocblk(cg, sblock.fs_frag, std_checkblkavail)) == 0)
goto bad;
bp = getdirblk(newblk, sblock.fs_bsize);
if (bp->b_errs)
goto bad;
memset(bp->b_un.b_buf, 0, sblock.fs_bsize);
for (cp = bp->b_un.b_buf;
cp < &bp->b_un.b_buf[sblock.fs_bsize];
cp += DIRBLKSIZ)
memmove(cp, &emptydir, sizeof emptydir);
dirty(bp);
if (lastlbn < UFS_NDADDR) {
DIP_SET(dp, di_db[lastlbn], newblk);
} else {
/*
* Allocate indirect block if needed.
*/
if ((indirblk = DIP(dp, di_ib[0])) == 0) {
Add support for managing UFS/FFS snapshots to fsck_ffs(8). The kernel handles the managment of UFS/FFS snapshots. Since UFS/FFS updates filesystem data (rather than always writing changes to new locations like ZFS), the kernel must check every filesystem write to see if the block being written is part of a snapshot. If it is part of a snapshot, then the kernel must make a copy of the old block value into a newly allocated block for the snapshot before allowing the write to be done. Similarly, if a block is being freed, the kernel must check to see if it is part of a snapshot and let the snapshot claim the block rather than freeing it for future use. When a snapshot is freed, its blocks need to be offered to older snapshots and freed only if no older snapshots wish to claim them. When snapshots were added to UFS/FFS they were integrated into soft updates and just a small part of the management of snapshots needed to be added to fsck_ffs(8) as soft updates minimized the set of snapshot changes that might need correction. When journaling was added to soft updates a much more complete knowledge of snapshots needed to be added to fsck_ffs(8) for it to be able to properly handle the filesystem changes that a journal rollback needs to do (specifically the freeing and allocation of blocks). Since this functionality was unavailable, the use of snapshots was disabled when running with journaled soft updates. This set of changes imports the kernel code for the management of snapshots to fsck_ffs(8). With this code in place it will become possible to enable snapshots when running with journalled soft updates. The most immediate benefit will be the ability to use snapshots to take consistent filesystem dumps on live filesystems. Future work will be done to update fsck_ffs(8) to be able to use snapshots to run in background on live filesystems running with journaled soft updates. Reviewed by: kib Tested by: Peter Holm Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D36491
2022-11-09 18:44:03 +00:00
indirblk = allocblk(cg, sblock.fs_frag,
std_checkblkavail);
if (indirblk == 0) {
binval(bp);
goto bad;
Add support for managing UFS/FFS snapshots to fsck_ffs(8). The kernel handles the managment of UFS/FFS snapshots. Since UFS/FFS updates filesystem data (rather than always writing changes to new locations like ZFS), the kernel must check every filesystem write to see if the block being written is part of a snapshot. If it is part of a snapshot, then the kernel must make a copy of the old block value into a newly allocated block for the snapshot before allowing the write to be done. Similarly, if a block is being freed, the kernel must check to see if it is part of a snapshot and let the snapshot claim the block rather than freeing it for future use. When a snapshot is freed, its blocks need to be offered to older snapshots and freed only if no older snapshots wish to claim them. When snapshots were added to UFS/FFS they were integrated into soft updates and just a small part of the management of snapshots needed to be added to fsck_ffs(8) as soft updates minimized the set of snapshot changes that might need correction. When journaling was added to soft updates a much more complete knowledge of snapshots needed to be added to fsck_ffs(8) for it to be able to properly handle the filesystem changes that a journal rollback needs to do (specifically the freeing and allocation of blocks). Since this functionality was unavailable, the use of snapshots was disabled when running with journaled soft updates. This set of changes imports the kernel code for the management of snapshots to fsck_ffs(8). With this code in place it will become possible to enable snapshots when running with journalled soft updates. The most immediate benefit will be the ability to use snapshots to take consistent filesystem dumps on live filesystems. Future work will be done to update fsck_ffs(8) to be able to use snapshots to run in background on live filesystems running with journaled soft updates. Reviewed by: kib Tested by: Peter Holm Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D36491
2022-11-09 18:44:03 +00:00
}
indiralloced = 1;
}
nbp = getdatablk(indirblk, sblock.fs_bsize, BT_LEVEL1);
if (nbp->b_errs)
goto bad;
if (indiralloced) {
memset(nbp->b_un.b_buf, 0, sblock.fs_bsize);
DIP_SET(dp, di_ib[0], indirblk);
DIP_SET(dp, di_blocks,
DIP(dp, di_blocks) + btodb(sblock.fs_bsize));
}
IBLK_SET(nbp, lastlbn - UFS_NDADDR, newblk);
dirty(nbp);
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
brelse(nbp);
}
DIP_SET(dp, di_size, filesize + sblock.fs_bsize);
DIP_SET(dp, di_blocks, DIP(dp, di_blocks) + btodb(sblock.fs_bsize));
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
inodirty(ip);
if (preen)
printf(" (EXPANDED)\n");
return (1);
bad:
pfatal(" (EXPANSION FAILED)\n");
Add support for managing UFS/FFS snapshots to fsck_ffs(8). The kernel handles the managment of UFS/FFS snapshots. Since UFS/FFS updates filesystem data (rather than always writing changes to new locations like ZFS), the kernel must check every filesystem write to see if the block being written is part of a snapshot. If it is part of a snapshot, then the kernel must make a copy of the old block value into a newly allocated block for the snapshot before allowing the write to be done. Similarly, if a block is being freed, the kernel must check to see if it is part of a snapshot and let the snapshot claim the block rather than freeing it for future use. When a snapshot is freed, its blocks need to be offered to older snapshots and freed only if no older snapshots wish to claim them. When snapshots were added to UFS/FFS they were integrated into soft updates and just a small part of the management of snapshots needed to be added to fsck_ffs(8) as soft updates minimized the set of snapshot changes that might need correction. When journaling was added to soft updates a much more complete knowledge of snapshots needed to be added to fsck_ffs(8) for it to be able to properly handle the filesystem changes that a journal rollback needs to do (specifically the freeing and allocation of blocks). Since this functionality was unavailable, the use of snapshots was disabled when running with journaled soft updates. This set of changes imports the kernel code for the management of snapshots to fsck_ffs(8). With this code in place it will become possible to enable snapshots when running with journalled soft updates. The most immediate benefit will be the ability to use snapshots to take consistent filesystem dumps on live filesystems. Future work will be done to update fsck_ffs(8) to be able to use snapshots to run in background on live filesystems running with journaled soft updates. Reviewed by: kib Tested by: Peter Holm Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D36491
2022-11-09 18:44:03 +00:00
if (nbp != NULL) {
binval(bp);
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
brelse(nbp);
Add support for managing UFS/FFS snapshots to fsck_ffs(8). The kernel handles the managment of UFS/FFS snapshots. Since UFS/FFS updates filesystem data (rather than always writing changes to new locations like ZFS), the kernel must check every filesystem write to see if the block being written is part of a snapshot. If it is part of a snapshot, then the kernel must make a copy of the old block value into a newly allocated block for the snapshot before allowing the write to be done. Similarly, if a block is being freed, the kernel must check to see if it is part of a snapshot and let the snapshot claim the block rather than freeing it for future use. When a snapshot is freed, its blocks need to be offered to older snapshots and freed only if no older snapshots wish to claim them. When snapshots were added to UFS/FFS they were integrated into soft updates and just a small part of the management of snapshots needed to be added to fsck_ffs(8) as soft updates minimized the set of snapshot changes that might need correction. When journaling was added to soft updates a much more complete knowledge of snapshots needed to be added to fsck_ffs(8) for it to be able to properly handle the filesystem changes that a journal rollback needs to do (specifically the freeing and allocation of blocks). Since this functionality was unavailable, the use of snapshots was disabled when running with journaled soft updates. This set of changes imports the kernel code for the management of snapshots to fsck_ffs(8). With this code in place it will become possible to enable snapshots when running with journalled soft updates. The most immediate benefit will be the ability to use snapshots to take consistent filesystem dumps on live filesystems. Future work will be done to update fsck_ffs(8) to be able to use snapshots to run in background on live filesystems running with journaled soft updates. Reviewed by: kib Tested by: Peter Holm Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D36491
2022-11-09 18:44:03 +00:00
}
if (newblk != 0) {
idesc.id_blkno = newblk;
idesc.id_numfrags = sblock.fs_frag;
(void)freeblock(&idesc);
}
if (indiralloced) {
idesc.id_blkno = indirblk;
idesc.id_numfrags = sblock.fs_frag;
(void)freeblock(&idesc);
}
return (0);
}
/*
* allocate a new directory
*/
ino_t
allocdir(ino_t parent, ino_t request, int mode)
{
ino_t ino;
char *cp;
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
struct inode ip;
This commit adds basic support for the UFS2 filesystem. The UFS2 filesystem expands the inode to 256 bytes to make space for 64-bit block pointers. It also adds a file-creation time field, an ability to use jumbo blocks per inode to allow extent like pointer density, and space for extended attributes (up to twice the filesystem block size worth of attributes, e.g., on a 16K filesystem, there is space for 32K of attributes). UFS2 fully supports and runs existing UFS1 filesystems. New filesystems built using newfs can be built in either UFS1 or UFS2 format using the -O option. In this commit UFS1 is the default format, so if you want to build UFS2 format filesystems, you must specify -O 2. This default will be changed to UFS2 when UFS2 proves itself to be stable. In this commit the boot code for reading UFS2 filesystems is not compiled (see /sys/boot/common/ufsread.c) as there is insufficient space in the boot block. Once the size of the boot block is increased, this code can be defined. Things to note: the definition of SBSIZE has changed to SBLOCKSIZE. The header file <ufs/ufs/dinode.h> must be included before <ufs/ffs/fs.h> so as to get the definitions of ufs2_daddr_t and ufs_lbn_t. Still TODO: Verify that the first level bootstraps work for all the architectures. Convert the utility ffsinfo to understand UFS2 and test growfs. Add support for the extended attribute storage. Update soft updates to ensure integrity of extended attribute storage. Switch the current extended attribute interfaces to use the extended attribute storage. Add the extent like functionality (framework is there, but is currently never used). Sponsored by: DARPA & NAI Labs. Reviewed by: Poul-Henning Kamp <phk@freebsd.org>
2002-06-21 06:18:05 +00:00
union dinode *dp;
struct bufarea *bp;
struct inoinfo *inp;
struct dirtemplate *dirp;
2018-03-17 12:59:55 +00:00
ino = allocino(request, IFDIR|mode);
if (ino == 0)
return (0);
dirp = &dirhead;
dirp->dot_ino = ino;
dirp->dotdot_ino = parent;
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
ginode(ino, &ip);
dp = ip.i_dp;
This commit adds basic support for the UFS2 filesystem. The UFS2 filesystem expands the inode to 256 bytes to make space for 64-bit block pointers. It also adds a file-creation time field, an ability to use jumbo blocks per inode to allow extent like pointer density, and space for extended attributes (up to twice the filesystem block size worth of attributes, e.g., on a 16K filesystem, there is space for 32K of attributes). UFS2 fully supports and runs existing UFS1 filesystems. New filesystems built using newfs can be built in either UFS1 or UFS2 format using the -O option. In this commit UFS1 is the default format, so if you want to build UFS2 format filesystems, you must specify -O 2. This default will be changed to UFS2 when UFS2 proves itself to be stable. In this commit the boot code for reading UFS2 filesystems is not compiled (see /sys/boot/common/ufsread.c) as there is insufficient space in the boot block. Once the size of the boot block is increased, this code can be defined. Things to note: the definition of SBSIZE has changed to SBLOCKSIZE. The header file <ufs/ufs/dinode.h> must be included before <ufs/ffs/fs.h> so as to get the definitions of ufs2_daddr_t and ufs_lbn_t. Still TODO: Verify that the first level bootstraps work for all the architectures. Convert the utility ffsinfo to understand UFS2 and test growfs. Add support for the extended attribute storage. Update soft updates to ensure integrity of extended attribute storage. Switch the current extended attribute interfaces to use the extended attribute storage. Add the extent like functionality (framework is there, but is currently never used). Sponsored by: DARPA & NAI Labs. Reviewed by: Poul-Henning Kamp <phk@freebsd.org>
2002-06-21 06:18:05 +00:00
bp = getdirblk(DIP(dp, di_db[0]), sblock.fs_fsize);
if (bp->b_errs) {
freeino(ino);
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
irelse(&ip);
return (0);
}
memmove(bp->b_un.b_buf, dirp, sizeof(struct dirtemplate));
for (cp = &bp->b_un.b_buf[DIRBLKSIZ];
cp < &bp->b_un.b_buf[sblock.fs_fsize];
cp += DIRBLKSIZ)
memmove(cp, &emptydir, sizeof emptydir);
dirty(bp);
DIP_SET(dp, di_nlink, 2);
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
inodirty(&ip);
if (ino == UFS_ROOTINO) {
This commit adds basic support for the UFS2 filesystem. The UFS2 filesystem expands the inode to 256 bytes to make space for 64-bit block pointers. It also adds a file-creation time field, an ability to use jumbo blocks per inode to allow extent like pointer density, and space for extended attributes (up to twice the filesystem block size worth of attributes, e.g., on a 16K filesystem, there is space for 32K of attributes). UFS2 fully supports and runs existing UFS1 filesystems. New filesystems built using newfs can be built in either UFS1 or UFS2 format using the -O option. In this commit UFS1 is the default format, so if you want to build UFS2 format filesystems, you must specify -O 2. This default will be changed to UFS2 when UFS2 proves itself to be stable. In this commit the boot code for reading UFS2 filesystems is not compiled (see /sys/boot/common/ufsread.c) as there is insufficient space in the boot block. Once the size of the boot block is increased, this code can be defined. Things to note: the definition of SBSIZE has changed to SBLOCKSIZE. The header file <ufs/ufs/dinode.h> must be included before <ufs/ffs/fs.h> so as to get the definitions of ufs2_daddr_t and ufs_lbn_t. Still TODO: Verify that the first level bootstraps work for all the architectures. Convert the utility ffsinfo to understand UFS2 and test growfs. Add support for the extended attribute storage. Update soft updates to ensure integrity of extended attribute storage. Switch the current extended attribute interfaces to use the extended attribute storage. Add the extent like functionality (framework is there, but is currently never used). Sponsored by: DARPA & NAI Labs. Reviewed by: Poul-Henning Kamp <phk@freebsd.org>
2002-06-21 06:18:05 +00:00
inoinfo(ino)->ino_linkcnt = DIP(dp, di_nlink);
if ((inp = getinoinfo(ino)) == NULL)
inp = cacheino(dp, ino);
else
inp->i_flags = INFO_NEW;
inp->i_parent = parent;
inp->i_dotdot = parent;
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
irelse(&ip);
return(ino);
}
if (!INO_IS_DVALID(parent)) {
freeino(ino);
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
irelse(&ip);
return (0);
}
inp = cacheino(dp, ino);
inp->i_parent = parent;
inp->i_dotdot = parent;
inoinfo(ino)->ino_state = inoinfo(parent)->ino_state;
if (inoinfo(ino)->ino_state == DSTATE) {
This commit adds basic support for the UFS2 filesystem. The UFS2 filesystem expands the inode to 256 bytes to make space for 64-bit block pointers. It also adds a file-creation time field, an ability to use jumbo blocks per inode to allow extent like pointer density, and space for extended attributes (up to twice the filesystem block size worth of attributes, e.g., on a 16K filesystem, there is space for 32K of attributes). UFS2 fully supports and runs existing UFS1 filesystems. New filesystems built using newfs can be built in either UFS1 or UFS2 format using the -O option. In this commit UFS1 is the default format, so if you want to build UFS2 format filesystems, you must specify -O 2. This default will be changed to UFS2 when UFS2 proves itself to be stable. In this commit the boot code for reading UFS2 filesystems is not compiled (see /sys/boot/common/ufsread.c) as there is insufficient space in the boot block. Once the size of the boot block is increased, this code can be defined. Things to note: the definition of SBSIZE has changed to SBLOCKSIZE. The header file <ufs/ufs/dinode.h> must be included before <ufs/ffs/fs.h> so as to get the definitions of ufs2_daddr_t and ufs_lbn_t. Still TODO: Verify that the first level bootstraps work for all the architectures. Convert the utility ffsinfo to understand UFS2 and test growfs. Add support for the extended attribute storage. Update soft updates to ensure integrity of extended attribute storage. Switch the current extended attribute interfaces to use the extended attribute storage. Add the extent like functionality (framework is there, but is currently never used). Sponsored by: DARPA & NAI Labs. Reviewed by: Poul-Henning Kamp <phk@freebsd.org>
2002-06-21 06:18:05 +00:00
inoinfo(ino)->ino_linkcnt = DIP(dp, di_nlink);
inoinfo(parent)->ino_linkcnt++;
}
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
irelse(&ip);
ginode(parent, &ip);
dp = ip.i_dp;
DIP_SET(dp, di_nlink, DIP(dp, di_nlink) + 1);
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
inodirty(&ip);
irelse(&ip);
return (ino);
}
/*
* free a directory inode
*/
static void
freedir(ino_t ino, ino_t parent)
{
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
struct inode ip;
This commit adds basic support for the UFS2 filesystem. The UFS2 filesystem expands the inode to 256 bytes to make space for 64-bit block pointers. It also adds a file-creation time field, an ability to use jumbo blocks per inode to allow extent like pointer density, and space for extended attributes (up to twice the filesystem block size worth of attributes, e.g., on a 16K filesystem, there is space for 32K of attributes). UFS2 fully supports and runs existing UFS1 filesystems. New filesystems built using newfs can be built in either UFS1 or UFS2 format using the -O option. In this commit UFS1 is the default format, so if you want to build UFS2 format filesystems, you must specify -O 2. This default will be changed to UFS2 when UFS2 proves itself to be stable. In this commit the boot code for reading UFS2 filesystems is not compiled (see /sys/boot/common/ufsread.c) as there is insufficient space in the boot block. Once the size of the boot block is increased, this code can be defined. Things to note: the definition of SBSIZE has changed to SBLOCKSIZE. The header file <ufs/ufs/dinode.h> must be included before <ufs/ffs/fs.h> so as to get the definitions of ufs2_daddr_t and ufs_lbn_t. Still TODO: Verify that the first level bootstraps work for all the architectures. Convert the utility ffsinfo to understand UFS2 and test growfs. Add support for the extended attribute storage. Update soft updates to ensure integrity of extended attribute storage. Switch the current extended attribute interfaces to use the extended attribute storage. Add the extent like functionality (framework is there, but is currently never used). Sponsored by: DARPA & NAI Labs. Reviewed by: Poul-Henning Kamp <phk@freebsd.org>
2002-06-21 06:18:05 +00:00
union dinode *dp;
if (ino != parent) {
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
ginode(parent, &ip);
dp = ip.i_dp;
DIP_SET(dp, di_nlink, DIP(dp, di_nlink) - 1);
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
inodirty(&ip);
irelse(&ip);
}
freeino(ino);
}
/*
* generate a temporary name for the lost+found directory.
*/
static int
lftempname(char *bufp, ino_t ino)
{
ino_t in;
char *cp;
int namlen;
cp = bufp + 2;
for (in = maxino; in > 0; in /= 10)
cp++;
*--cp = 0;
namlen = cp - bufp;
in = ino;
while (cp > bufp) {
*--cp = (in % 10) + '0';
in /= 10;
}
*cp = '#';
return (namlen);
}
/*
* Get a directory block.
* Insure that it is held until another is requested.
*/
static struct bufarea *
This commit adds basic support for the UFS2 filesystem. The UFS2 filesystem expands the inode to 256 bytes to make space for 64-bit block pointers. It also adds a file-creation time field, an ability to use jumbo blocks per inode to allow extent like pointer density, and space for extended attributes (up to twice the filesystem block size worth of attributes, e.g., on a 16K filesystem, there is space for 32K of attributes). UFS2 fully supports and runs existing UFS1 filesystems. New filesystems built using newfs can be built in either UFS1 or UFS2 format using the -O option. In this commit UFS1 is the default format, so if you want to build UFS2 format filesystems, you must specify -O 2. This default will be changed to UFS2 when UFS2 proves itself to be stable. In this commit the boot code for reading UFS2 filesystems is not compiled (see /sys/boot/common/ufsread.c) as there is insufficient space in the boot block. Once the size of the boot block is increased, this code can be defined. Things to note: the definition of SBSIZE has changed to SBLOCKSIZE. The header file <ufs/ufs/dinode.h> must be included before <ufs/ffs/fs.h> so as to get the definitions of ufs2_daddr_t and ufs_lbn_t. Still TODO: Verify that the first level bootstraps work for all the architectures. Convert the utility ffsinfo to understand UFS2 and test growfs. Add support for the extended attribute storage. Update soft updates to ensure integrity of extended attribute storage. Switch the current extended attribute interfaces to use the extended attribute storage. Add the extent like functionality (framework is there, but is currently never used). Sponsored by: DARPA & NAI Labs. Reviewed by: Poul-Henning Kamp <phk@freebsd.org>
2002-06-21 06:18:05 +00:00
getdirblk(ufs2_daddr_t blkno, long size)
{
Rewrite the disk I/O management system in fsck_ffs(8). Other than making fsck_ffs(8) run faster, there should be no functional change. The original fsck_ffs(8) had its own disk I/O management system. When gjournal(8) was added to FreeBSD 7, code was added to fsck_ffs(8) to do the necessary gjournal rollback. Rather than use the existing fsck_ffs(8) disk I/O system, it wrote its own from scratch. Similarly when journalled soft updates were added in FreeBSD 9, code was added to fsck_ffs(8) to do the necessary journal rollback. And once again, rather than using either of the existing fsck_ffs(8) disk I/O systems, it wrote its own from scratch. Lastly the fsdb(8) utility uses the fsck_ffs(8) disk I/O management system. In preparation for making the changes necessary to enable snapshots to be taken when using journalled soft updates, it was necessary to have a single disk I/O system used by all the various subsystems in fsck_ffs(8). This commit merges the functionality required by all the different subsystems into a single disk I/O system that supports all of their needs. In so doing it picks up optimizations from each of them with the results that each of the subsystems does fewer reads and writes than it did with its own customized I/O system. It also greatly simplifies making changes to fsck_ffs(8) since everything goes through a single place. For example the ginode() function fetches an inode from the disk. When inode check hashes were added, they previously had to be checked in the code implementing inode fetch in each of the three different disk I/O systems. Now they need only be checked in ginode(). Tested by: Peter Holm Sponsored by: Netflix
2021-01-07 01:37:08 +00:00
if (pdirbp != NULL && pdirbp->b_errs == 0)
brelse(pdirbp);
pdirbp = getdatablk(blkno, size, BT_DIRDATA);
return (pdirbp);
}