d/freebsd-dev
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
b3fe5d9322
freebsd-dev/sbin/fsck_ffs/pass2.c

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

740 lines
21 KiB
C
Raw Normal View History

General further adoption of SPDX licensing ID tags. Mainly focus on files that use BSD 3-Clause license. The Software Package Data Exchange (SPDX) group provides a specification to make it easier for automated tools to detect and summarize well known opensource licenses. We are gradually adopting the specification, noting that the tags are considered only advisory and do not, in any way, superceed or replace the license texts. Special thanks to Wind River for providing access to "The Duke of Highlander" tool: an older (2014) run over FreeBSD tree was useful as a starting point.
2017-11-20 19:49:47 +00:00
/*-
* SPDX-License-Identifier: BSD-3-Clause
*
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
* 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.
Renumber copyright clause 4 Renumber cluase 4 to 3, per what everybody else did when BSD granted them permission to remove clause 3. My insistance on keeping the same numbering for legal reasons is too pedantic, so give up on that point. Submitted by: Jan Schaumann <jschauma@stevens.edu> Pull Request: https://github.com/freebsd/freebsd/pull/96
2017-02-28 23:42:47 +00:00
* 3. Neither the name of the University nor the names of its contributors
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
* 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.
*/
Cosmetic and documentation changes brought from earlier FreeBSD versions. (e.g. RCS Id:)
1998-12-03 02:41:11 +00:00
#if 0
Replace various spelling with FALLTHROUGH which is lint()able
2002-08-25 13:10:45 +00:00
#ifndef lint
Merge from Lite2. Note that Lite2 has it's own filesystem clean check skipping code that overrides ours sooner. One should be eliminated, but for now it works.
1997-03-11 12:20:21 +00:00
static const char sccsid[] = "@(#)pass2.c 8.9 (Berkeley) 4/28/95";
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
#endif /* not lint */
Replace various spelling with FALLTHROUGH which is lint()able
2002-08-25 13:10:45 +00:00
#endif
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
#include <sys/param.h>
This update utilizes new fsck sysctl commands that allow fsck running in background mode to correct expected inconsistencies that arise during directory rename (see immediately previous update to this file for details). If run on a kernel without the new functionality, background fsck will simply ignore these inconsistencies rather than fail. Reported by: jeff
2010-01-11 20:05:38 +00:00
#include <sys/sysctl.h>
Merge from Lite2. Note that Lite2 has it's own filesystem clean check skipping code that overrides ours sooner. One should be eliminated, but for now it works.
1997-03-11 12:20:21 +00:00
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
#include <ufs/ufs/dinode.h>
#include <ufs/ufs/dir.h>
Additions to run checks on live filesystems. This change will not affect current systems until fsck is modified to use these new facilities. To try out this change, set the fsck passno to zero in /etc/fstab to cause the filesystem to be mounted without running fsck, then run `fsck_ffs -p -B <filesystem>' after the system has been brought up multiuser to run a background cleanup on <filesystem>. Note that the <filesystem> in question must have soft updates enabled.
2001-03-21 09:48:03 +00:00
#include <ufs/ffs/fs.h>
Finished (?) merging with Lite2: cleaned up #include mess.
1997-03-12 17:01:11 +00:00
Merge from Lite2. Note that Lite2 has it's own filesystem clean check skipping code that overrides ours sooner. One should be eliminated, but for now it works.
1997-03-11 12:20:21 +00:00
#include <err.h>
This update utilizes new fsck sysctl commands that allow fsck running in background mode to correct expected inconsistencies that arise during directory rename (see immediately previous update to this file for details). If run on a kernel without the new functionality, background fsck will simply ignore these inconsistencies rather than fail. Reported by: jeff
2010-01-11 20:05:38 +00:00
#include <errno.h>
Fix a bunch of format string warnings which broke the sparc64 build. Tested on: sparc64, i386
2002-07-31 12:01:14 +00:00
#include <stdint.h>
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
#include <string.h>
Merge from Lite2. Note that Lite2 has it's own filesystem clean check skipping code that overrides ours sooner. One should be eliminated, but for now it works.
1997-03-11 12:20:21 +00:00
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
#include "fsck.h"
#define MINDIRSIZE (sizeof (struct dirtemplate))
When renaming a directory it passes through several intermediate states. First its new name will be created causing it to have two names (from possibly different parents). Next, if it has different parents, its value of ".." will be changed from pointing to the old parent to pointing to the new parent. Concurrently, its old name will be removed bringing it back into a consistent state. When fsck encounters an extra name for a directory, it offers to remove the "extraneous hard link"; when it finds that the names have been changed but the update to ".." has not happened, it offers to rewrite ".." to point at the correct parent. Both of these changes were considered unexpected so would cause fsck in preen mode or fsck in background mode to fail with the need to run fsck manually to fix these problems. This update changes these errors to be expected so that in preen mode fsck will simply fix these transitional errors. For now, background fsck will note these errors, but will need additional kernel support to fix them, so will simply ignore them rather than fail. A future update will allow background fsck to fix these problems. Reported by: jeff
2010-01-11 19:52:40 +00:00
static int fix_extraneous(struct inoinfo *, struct inodesc *);
static int deleteentry(struct inodesc *);
o __P removed o ansi function prototypes o unifdef -D__STDC__ o __dead2 on usage prototype o remove now-bogus main prototype
2002-03-20 22:57:10 +00:00
static int blksort(const void *, const void *);
static int pass2check(struct inodesc *);
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
Submitted by: phk, added to by bde Fix all the warnings from `gcc -Wall'.
1995-04-02 14:52:29 +00:00
void
o __P removed o ansi function prototypes o unifdef -D__STDC__ o __dead2 on usage prototype o remove now-bogus main prototype
2002-03-20 22:57:10 +00:00
pass2(void)
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
{
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;
Remove 'register' keyword. It does not help modern compilers, and some may take some hit from it. (I also found several functions that listed *every* of its 10 local vars with "register" -- just how many free registers do people think machines have?)
2002-03-20 17:55:10 +00:00
struct inoinfo **inpp, *inp;
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
struct inoinfo **inpend;
struct inodesc curino;
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 dino;
int i;
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
char pathbuf[MAXPATHLEN + 1];
prefix UFS symbols with UFS_ to reduce namespace pollution Specifically: ROOTINO -> UFS_ROOTINO WINO -> UFS_WINO NXADDR -> UFS_NXADDR NDADDR -> UFS_NDADDR NIADDR -> UFS_NIADDR MAXSYMLINKLEN_UFS[12] -> UFS[12]_MAXSYMLINKLEN (for consistency) Also prefix ext2's and nandfs's NDADDR and NIADDR with EXT2_ and NANDFS_ Reviewed by: kib, mckusick Obtained from: NetBSD MFC after: 1 month Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D9536
2017-02-15 19:50:26 +00:00
switch (inoinfo(UFS_ROOTINO)->ino_state) {
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
case USTATE:
pfatal("ROOT INODE UNALLOCATED");
Reviewed by: dyson@freebsd.org (john Dyson), dg@root.com (david greenman) Submitted by: Kirk McKusick (mcKusick@mckusick.com) Obtained from: WHistle development tree
1998-03-08 09:59:44 +00:00
if (reply("ALLOCATE") == 0) {
ckfini(0);
Merge from Lite2. Note that Lite2 has it's own filesystem clean check skipping code that overrides ours sooner. One should be eliminated, but for now it works.
1997-03-11 12:20:21 +00:00
exit(EEXIT);
Reviewed by: dyson@freebsd.org (john Dyson), dg@root.com (david greenman) Submitted by: Kirk McKusick (mcKusick@mckusick.com) Obtained from: WHistle development tree
1998-03-08 09:59:44 +00:00
}
prefix UFS symbols with UFS_ to reduce namespace pollution Specifically: ROOTINO -> UFS_ROOTINO WINO -> UFS_WINO NXADDR -> UFS_NXADDR NDADDR -> UFS_NDADDR NIADDR -> UFS_NIADDR MAXSYMLINKLEN_UFS[12] -> UFS[12]_MAXSYMLINKLEN (for consistency) Also prefix ext2's and nandfs's NDADDR and NIADDR with EXT2_ and NANDFS_ Reviewed by: kib, mckusick Obtained from: NetBSD MFC after: 1 month Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D9536
2017-02-15 19:50:26 +00:00
if (allocdir(UFS_ROOTINO, UFS_ROOTINO, 0755) != UFS_ROOTINO)
Merge from Lite2. Note that Lite2 has it's own filesystem clean check skipping code that overrides ours sooner. One should be eliminated, but for now it works.
1997-03-11 12:20:21 +00:00
errx(EEXIT, "CANNOT ALLOCATE ROOT INODE");
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
break;
case DCLEAR:
pfatal("DUPS/BAD IN ROOT INODE");
if (reply("REALLOCATE")) {
Correct several bugs in fsck_ffs(8) triggered by corrupted filesystems. If a directory entry has an illegal inode number (less than zero or greater than the last inode in the filesystem) the entry is removed. If a directory '.' or '..' entry had an illegal inode number they were being removed. Since fsck_ffs knows what the correct value is for these two entries fix them rather deleting them. Add much more extensive cylinder group checks and use them to be more careful about rebuilding a cylinder group. Check for out-of-range block numbers before trying to free them. When a directory is deleted also remove its cache entry created in pass1 so that later passes do not try to operate on a deleted directory. Check for ctime(3) returning NULL before trying to use its return. When freeing a directory inode, do not try to interpret it as a directory. Reserve space in the inostatlist to have room to allocate a lost+found directory. If an invalid block number is found past the end of an inode simply remove it rather than clearing and removing the inode. Modernize the inoinfo structure to use queue(3) LIST rather than a handrolled linked list implementation. Reported by: Bob Prohaska, John-Mark Gurney, and Mark Millard Tested by: Peter Holm Reviewed by: Peter Holm MFC after: 2 weeks Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D38668
2023-03-07 23:12:37 +00:00
freedirino(UFS_ROOTINO, UFS_ROOTINO);
prefix UFS symbols with UFS_ to reduce namespace pollution Specifically: ROOTINO -> UFS_ROOTINO WINO -> UFS_WINO NXADDR -> UFS_NXADDR NDADDR -> UFS_NDADDR NIADDR -> UFS_NIADDR MAXSYMLINKLEN_UFS[12] -> UFS[12]_MAXSYMLINKLEN (for consistency) Also prefix ext2's and nandfs's NDADDR and NIADDR with EXT2_ and NANDFS_ Reviewed by: kib, mckusick Obtained from: NetBSD MFC after: 1 month Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D9536
2017-02-15 19:50:26 +00:00
if (allocdir(UFS_ROOTINO, UFS_ROOTINO, 0755) !=
UFS_ROOTINO)
Merge from Lite2. Note that Lite2 has it's own filesystem clean check skipping code that overrides ours sooner. One should be eliminated, but for now it works.
1997-03-11 12:20:21 +00:00
errx(EEXIT, "CANNOT ALLOCATE ROOT INODE");
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
break;
}
Reviewed by: dyson@freebsd.org (john Dyson), dg@root.com (david greenman) Submitted by: Kirk McKusick (mcKusick@mckusick.com) Obtained from: WHistle development tree
1998-03-08 09:59:44 +00:00
if (reply("CONTINUE") == 0) {
ckfini(0);
Merge from Lite2. Note that Lite2 has it's own filesystem clean check skipping code that overrides ours sooner. One should be eliminated, but for now it works.
1997-03-11 12:20:21 +00:00
exit(EEXIT);
Reviewed by: dyson@freebsd.org (john Dyson), dg@root.com (david greenman) Submitted by: Kirk McKusick (mcKusick@mckusick.com) Obtained from: WHistle development tree
1998-03-08 09:59:44 +00:00
}
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
break;
case FSTATE:
case FCLEAR:
Eliminate linked list used to track inodes with an initial link count of zero and instead encode this information in the inode state. Pass 4 performed a linear search of this list for each inode in the file system, which performs poorly if the list is long. Reviewed by: sam & keramida (an earlier version of the patch), mckusick MFC after: 1 month
2004-10-08 20:44:47 +00:00
case FZLINK:
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
pfatal("ROOT INODE NOT DIRECTORY");
if (reply("REALLOCATE")) {
prefix UFS symbols with UFS_ to reduce namespace pollution Specifically: ROOTINO -> UFS_ROOTINO WINO -> UFS_WINO NXADDR -> UFS_NXADDR NDADDR -> UFS_NDADDR NIADDR -> UFS_NIADDR MAXSYMLINKLEN_UFS[12] -> UFS[12]_MAXSYMLINKLEN (for consistency) Also prefix ext2's and nandfs's NDADDR and NIADDR with EXT2_ and NANDFS_ Reviewed by: kib, mckusick Obtained from: NetBSD MFC after: 1 month Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D9536
2017-02-15 19:50:26 +00:00
freeino(UFS_ROOTINO);
if (allocdir(UFS_ROOTINO, UFS_ROOTINO, 0755) !=
UFS_ROOTINO)
Merge from Lite2. Note that Lite2 has it's own filesystem clean check skipping code that overrides ours sooner. One should be eliminated, but for now it works.
1997-03-11 12:20:21 +00:00
errx(EEXIT, "CANNOT ALLOCATE ROOT INODE");
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
break;
}
Reviewed by: dyson@freebsd.org (john Dyson), dg@root.com (david greenman) Submitted by: Kirk McKusick (mcKusick@mckusick.com) Obtained from: WHistle development tree
1998-03-08 09:59:44 +00:00
if (reply("FIX") == 0) {
ckfini(0);
Merge from Lite2. Note that Lite2 has it's own filesystem clean check skipping code that overrides ours sooner. One should be eliminated, but for now it works.
1997-03-11 12:20:21 +00:00
exit(EEXIT);
Reviewed by: dyson@freebsd.org (john Dyson), dg@root.com (david greenman) Submitted by: Kirk McKusick (mcKusick@mckusick.com) Obtained from: WHistle development tree
1998-03-08 09:59:44 +00:00
}
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);
dp = ip.i_dp;
Revert r313780 (UFS_ prefix)
2018-03-17 12:59:55 +00:00
DIP_SET(dp, di_mode, DIP(dp, di_mode) & ~IFMT);
DIP_SET(dp, di_mode, DIP(dp, di_mode) | IFDIR);
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);
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
break;
case DSTATE:
Eliminate linked list used to track inodes with an initial link count of zero and instead encode this information in the inode state. Pass 4 performed a linear search of this list for each inode in the file system, which performs poorly if the list is long. Reviewed by: sam & keramida (an earlier version of the patch), mckusick MFC after: 1 month
2004-10-08 20:44:47 +00:00
case DZLINK:
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
break;
default:
Reviewed by: Don Lewis <Don.Lewis@tsc.tdk.com> Submitted by: Kirk McKusick <mckusick@McKusick.COM> Obtained from: Mckusick, BSDI and a host of others This exactly matches Kirks sources imported under the Tag MCKUSICK2. These are as supplied by kirk with one small change needed to compile under freeBSD. Some FreeBSD patches will be added back, though many have been added to Kirk's sources already.
1998-12-03 02:27:35 +00:00
errx(EEXIT, "BAD STATE %d FOR ROOT INODE",
prefix UFS symbols with UFS_ to reduce namespace pollution Specifically: ROOTINO -> UFS_ROOTINO WINO -> UFS_WINO NXADDR -> UFS_NXADDR NDADDR -> UFS_NDADDR NIADDR -> UFS_NIADDR MAXSYMLINKLEN_UFS[12] -> UFS[12]_MAXSYMLINKLEN (for consistency) Also prefix ext2's and nandfs's NDADDR and NIADDR with EXT2_ and NANDFS_ Reviewed by: kib, mckusick Obtained from: NetBSD MFC after: 1 month Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D9536
2017-02-15 19:50:26 +00:00
inoinfo(UFS_ROOTINO)->ino_state);
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
}
prefix UFS symbols with UFS_ to reduce namespace pollution Specifically: ROOTINO -> UFS_ROOTINO WINO -> UFS_WINO NXADDR -> UFS_NXADDR NDADDR -> UFS_NDADDR NIADDR -> UFS_NIADDR MAXSYMLINKLEN_UFS[12] -> UFS[12]_MAXSYMLINKLEN (for consistency) Also prefix ext2's and nandfs's NDADDR and NIADDR with EXT2_ and NANDFS_ Reviewed by: kib, mckusick Obtained from: NetBSD MFC after: 1 month Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D9536
2017-02-15 19:50:26 +00:00
inoinfo(UFS_ROOTINO)->ino_state = DFOUND;
inoinfo(UFS_WINO)->ino_state = FSTATE;
inoinfo(UFS_WINO)->ino_type = DT_WHT;
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
/*
* Sort the directory list into disk block order.
*/
qsort((char *)inpsort, (size_t)inplast, sizeof *inpsort, blksort);
/*
* Check the integrity of each directory.
*/
Merge from Lite2. Note that Lite2 has it's own filesystem clean check skipping code that overrides ours sooner. One should be eliminated, but for now it works.
1997-03-11 12:20:21 +00:00
memset(&curino, 0, sizeof(struct inodesc));
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
curino.id_type = DATA;
curino.id_func = pass2check;
inpend = &inpsort[inplast];
for (inpp = inpsort; inpp < inpend; inpp++) {
Add a simple SIGINFO handler to fsck_ffs. Shortly after receipt of a SIGINFO (normally via Ctrl-T), a line will be output indicating the current phase number and progress information relevant to the current phase. Approved by: mckusick
2000-12-15 14:23:55 +00:00
if (got_siginfo) {
Fix a bunch of format string warnings which broke the sparc64 build. Tested on: sparc64, i386
2002-07-31 12:01:14 +00:00
printf("%s: phase 2: dir %td of %d (%d%%)\n", cdevname,
Fix a large number of -Wall, -Wformat and -W compiler warnings. These were mainly missing casts or wrong format strings in printf statements, but there were also missing includes, unused variables, functions and arguments. The choice of `long' vs `int' still seems almost random in a lot of places though.
2001-11-17 23:48:21 +00:00
inpp - inpsort, (int)inplast,
(int)((inpp - inpsort) * 100 / inplast));
Add a simple SIGINFO handler to fsck_ffs. Shortly after receipt of a SIGINFO (normally via Ctrl-T), a line will be output indicating the current phase number and progress information relevant to the current phase. Approved by: mckusick
2000-12-15 14:23:55 +00:00
got_siginfo = 0;
}
In the case of a background fsck, periodically update the process title with a progress update.
2004-02-28 07:50:42 +00:00
if (got_sigalarm) {
setproctitle("%s p2 %d%%", cdevname,
(int)((inpp - inpsort) * 100 / inplast));
got_sigalarm = 0;
}
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
inp = *inpp;
if (inp->i_isize == 0)
continue;
if (inp->i_isize < MINDIRSIZE) {
direrror(inp->i_number, "DIRECTORY TOO SHORT");
inp->i_isize = roundup(MINDIRSIZE, DIRBLKSIZ);
if (reply("FIX") == 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
ginode(inp->i_number, &ip);
DIP_SET(ip.i_dp, di_size, inp->i_isize);
inodirty(&ip);
irelse(&ip);
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
}
} else if ((inp->i_isize & (DIRBLKSIZ - 1)) != 0) {
getpathname(pathbuf, inp->i_number, inp->i_number);
Reviewed by: dyson@freebsd.org (john Dyson), dg@root.com (david greenman) Submitted by: Kirk McKusick (mcKusick@mckusick.com) Obtained from: WHistle development tree
1998-03-08 09:59:44 +00:00
if (usedsoftdep)
Fix a bunch of format string warnings which broke the sparc64 build. Tested on: sparc64, i386
2002-07-31 12:01:14 +00:00
pfatal("%s %s: LENGTH %jd NOT MULTIPLE OF %d",
"DIRECTORY", pathbuf,
(intmax_t)inp->i_isize, DIRBLKSIZ);
Reviewed by: dyson@freebsd.org (john Dyson), dg@root.com (david greenman) Submitted by: Kirk McKusick (mcKusick@mckusick.com) Obtained from: WHistle development tree
1998-03-08 09:59:44 +00:00
else
Fix a bunch of format string warnings which broke the sparc64 build. Tested on: sparc64, i386
2002-07-31 12:01:14 +00:00
pwarn("%s %s: LENGTH %jd NOT MULTIPLE OF %d",
"DIRECTORY", pathbuf,
(intmax_t)inp->i_isize, DIRBLKSIZ);
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
if (preen)
printf(" (ADJUSTED)\n");
inp->i_isize = roundup(inp->i_isize, DIRBLKSIZ);
if (preen || reply("ADJUST") == 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
ginode(inp->i_number, &ip);
DIP_SET(ip.i_dp, di_size,
Create DIP_SET() and IBLK_SET() macros to fix lvalue warnings. Inspired by: kan
2004-09-01 05:48:06 +00:00
roundup(inp->i_isize, 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);
irelse(&ip);
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
}
}
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
dp = &dino;
memset(dp, 0, sizeof(struct ufs2_dinode));
Revert r313780 (UFS_ prefix)
2018-03-17 12:59:55 +00:00
DIP_SET(dp, di_mode, IFDIR);
Create DIP_SET() and IBLK_SET() macros to fix lvalue warnings. Inspired by: kan
2004-09-01 05:48:06 +00:00
DIP_SET(dp, di_size, inp->i_isize);
prefix UFS symbols with UFS_ to reduce namespace pollution Specifically: ROOTINO -> UFS_ROOTINO WINO -> UFS_WINO NXADDR -> UFS_NXADDR NDADDR -> UFS_NDADDR NIADDR -> UFS_NIADDR MAXSYMLINKLEN_UFS[12] -> UFS[12]_MAXSYMLINKLEN (for consistency) Also prefix ext2's and nandfs's NDADDR and NIADDR with EXT2_ and NANDFS_ Reviewed by: kib, mckusick Obtained from: NetBSD MFC after: 1 month Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D9536
2017-02-15 19:50:26 +00:00
for (i = 0; i < MIN(inp->i_numblks, UFS_NDADDR); i++)
Create DIP_SET() and IBLK_SET() macros to fix lvalue warnings. Inspired by: kan
2004-09-01 05:48:06 +00:00
DIP_SET(dp, di_db[i], inp->i_blks[i]);
prefix UFS symbols with UFS_ to reduce namespace pollution Specifically: ROOTINO -> UFS_ROOTINO WINO -> UFS_WINO NXADDR -> UFS_NXADDR NDADDR -> UFS_NDADDR NIADDR -> UFS_NIADDR MAXSYMLINKLEN_UFS[12] -> UFS[12]_MAXSYMLINKLEN (for consistency) Also prefix ext2's and nandfs's NDADDR and NIADDR with EXT2_ and NANDFS_ Reviewed by: kib, mckusick Obtained from: NetBSD MFC after: 1 month Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D9536
2017-02-15 19:50:26 +00:00
if (inp->i_numblks > UFS_NDADDR)
for (i = 0; i < UFS_NIADDR; i++)
DIP_SET(dp, di_ib[i],
inp->i_blks[UFS_NDADDR + i]);
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
curino.id_number = inp->i_number;
curino.id_parent = inp->i_parent;
(void)ckinode(dp, &curino);
}
/*
* Now that the parents of all directories have been found,
* make another pass to verify the value of `..'
*/
for (inpp = inpsort; inpp < inpend; inpp++) {
inp = *inpp;
if (inp->i_parent == 0 || inp->i_isize == 0)
continue;
Reviewed by: Don Lewis <Don.Lewis@tsc.tdk.com> Submitted by: Kirk McKusick <mckusick@McKusick.COM> Obtained from: Mckusick, BSDI and a host of others This exactly matches Kirks sources imported under the Tag MCKUSICK2. These are as supplied by kirk with one small change needed to compile under freeBSD. Some FreeBSD patches will be added back, though many have been added to Kirk's sources already.
1998-12-03 02:27:35 +00:00
if (inoinfo(inp->i_parent)->ino_state == DFOUND &&
Improvement in UFS/FFS directory placement when doing mkdir(2). The algorithm for laying out new directories was devised in the 1980s and markedly improved the performance of the filesystem. In those days large disks had at most 100 cylinder groups and often as few as 10-20. Modern multi-terrabyte disks have thousands of cylinder groups. The original algorithm does not handle these large sizes well. This change attempts to expand the scope of the original algorithm to work well with these much larger disks while still retaining the properties of the original algorithm for small disks. The filesystem implementation is divided into policy routines and implementation routines. The policy routines can be changed in any way desired without risk of corrupting the filesystem. The policy requests are handled by the implementation layer. If the policy asks for an available resource, it is granted. But if it asks for an already in-use resource, then the implementation will provide an available one nearby the request. Thus it is impossible for a policy to double allocate. This change is limited to the policy implementation. This change updates the ffs_dirpref() routine which is responsible for selecting the cylinder group into which a new directory should be placed. If we are near the root of the filesystem we aim to spread them out as much as possible. As we descend deeper from the root we cluster them closer together around their parent as we expect them to be more closely interactive. Higher-level directories like usr/src/sys and usr/src/bin should be separated while the directories in these areas are more likely to be accessed together so should be closer. And directories within commands or kernel subsystems should be closer still. We pick a range of cylinder groups around the cylinder group of the directory in which we are being created. The size of the range for our search is based on our depth from the root of our filesystem. We then probe that range based on how many directories are already present. The first new directory is at 1/2 (middle) of the range; the second is in the first 1/4 of the range, then at 3/4, 1/8, 3/8, 5/8, 7/8, 1/16, 3/16, 5/16, etc. It is desirable to store the depth of a directory in its on-disk inode so that it is available when we need it. We add a new field di_dirdepth to track the depth of each directory. Because there are few spare fields left in the inode, we choose to share an existing field in the inode rather than having one of our own. Specifically we create a union with the di_freelink field. The di_freelink field is used to track inodes that have been unlinked but remain referenced. It is not needed until a rmdir(2) operation has been done on a directory. At that point, the directory has no contents and even if it is kept active as a current directory is no longer able to have any new directories or files created in it. Thus the use of di_dirdepth and di_freelink will never coincide. Reported by: Timo Voelker Reviewed by: kib Tested by: Peter Holm MFC after: 2 weeks Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D39246
2023-03-30 04:09:39 +00:00
INO_IS_DUNFOUND(inp->i_number)) {
Reviewed by: Don Lewis <Don.Lewis@tsc.tdk.com> Submitted by: Kirk McKusick <mckusick@McKusick.COM> Obtained from: Mckusick, BSDI and a host of others This exactly matches Kirks sources imported under the Tag MCKUSICK2. These are as supplied by kirk with one small change needed to compile under freeBSD. Some FreeBSD patches will be added back, though many have been added to Kirk's sources already.
1998-12-03 02:27:35 +00:00
inoinfo(inp->i_number)->ino_state = DFOUND;
Improvement in UFS/FFS directory placement when doing mkdir(2). The algorithm for laying out new directories was devised in the 1980s and markedly improved the performance of the filesystem. In those days large disks had at most 100 cylinder groups and often as few as 10-20. Modern multi-terrabyte disks have thousands of cylinder groups. The original algorithm does not handle these large sizes well. This change attempts to expand the scope of the original algorithm to work well with these much larger disks while still retaining the properties of the original algorithm for small disks. The filesystem implementation is divided into policy routines and implementation routines. The policy routines can be changed in any way desired without risk of corrupting the filesystem. The policy requests are handled by the implementation layer. If the policy asks for an available resource, it is granted. But if it asks for an already in-use resource, then the implementation will provide an available one nearby the request. Thus it is impossible for a policy to double allocate. This change is limited to the policy implementation. This change updates the ffs_dirpref() routine which is responsible for selecting the cylinder group into which a new directory should be placed. If we are near the root of the filesystem we aim to spread them out as much as possible. As we descend deeper from the root we cluster them closer together around their parent as we expect them to be more closely interactive. Higher-level directories like usr/src/sys and usr/src/bin should be separated while the directories in these areas are more likely to be accessed together so should be closer. And directories within commands or kernel subsystems should be closer still. We pick a range of cylinder groups around the cylinder group of the directory in which we are being created. The size of the range for our search is based on our depth from the root of our filesystem. We then probe that range based on how many directories are already present. The first new directory is at 1/2 (middle) of the range; the second is in the first 1/4 of the range, then at 3/4, 1/8, 3/8, 5/8, 7/8, 1/16, 3/16, 5/16, etc. It is desirable to store the depth of a directory in its on-disk inode so that it is available when we need it. We add a new field di_dirdepth to track the depth of each directory. Because there are few spare fields left in the inode, we choose to share an existing field in the inode rather than having one of our own. Specifically we create a union with the di_freelink field. The di_freelink field is used to track inodes that have been unlinked but remain referenced. It is not needed until a rmdir(2) operation has been done on a directory. At that point, the directory has no contents and even if it is kept active as a current directory is no longer able to have any new directories or files created in it. Thus the use of di_dirdepth and di_freelink will never coincide. Reported by: Timo Voelker Reviewed by: kib Tested by: Peter Holm MFC after: 2 weeks Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D39246
2023-03-30 04:09:39 +00:00
check_dirdepth(inp);
}
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
if (inp->i_dotdot == inp->i_parent ||
inp->i_dotdot == (ino_t)-1)
continue;
if (inp->i_dotdot == 0) {
inp->i_dotdot = inp->i_parent;
Improve handling of missing '.' and '..' in UFS directories. The UFS filesystem expects to find '.' and '..' as the first two entries in a directory. The kernel's UFS name cache can become quite confused when these two entries are not present as the first two entries. Prior to this change, when the fsck_ffs(8) utility detected that '.' and/or '..' were missing, it would report them, but only offered to replace them if the space at the beginning of the directory was available. Otherwise it was left to the system administrator to move the offending file(s) out of the way and then rerun fsck_ffs(8) to create the '.' and '..' entries. With this change, fsck_ffs(8) will always be able to create the '.' and/or '..' entries. It moves any files in the way elsewhere in the directory block. If there is no room in the directory block to which to move them, they are placed in the lost+found directory. Reported by: Peter Holm Sponsored by: The FreeBSD Foundation
2022-08-17 21:19:59 +00:00
if (debug)
fileerror(inp->i_parent, inp->i_number,
"DEFERRED MISSING '..' FIX");
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
(void)makeentry(inp->i_number, inp->i_parent, "..");
Reviewed by: Don Lewis <Don.Lewis@tsc.tdk.com> Submitted by: Kirk McKusick <mckusick@McKusick.COM> Obtained from: Mckusick, BSDI and a host of others This exactly matches Kirks sources imported under the Tag MCKUSICK2. These are as supplied by kirk with one small change needed to compile under freeBSD. Some FreeBSD patches will be added back, though many have been added to Kirk's sources already.
1998-12-03 02:27:35 +00:00
inoinfo(inp->i_parent)->ino_linkcnt--;
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
continue;
}
This update utilizes new fsck sysctl commands that allow fsck running in background mode to correct expected inconsistencies that arise during directory rename (see immediately previous update to this file for details). If run on a kernel without the new functionality, background fsck will simply ignore these inconsistencies rather than fail. Reported by: jeff
2010-01-11 20:05:38 +00:00
/*
* Here we have:
* inp->i_number is directory with bad ".." in it.
* inp->i_dotdot is current value of "..".
* inp->i_parent is directory to which ".." should point.
*/
getpathname(pathbuf, inp->i_parent, inp->i_number);
Fix fsck_ffs build with a 64-bit ino_t. Original code by: Gleb Kurtsou
2012-09-27 23:30:58 +00:00
printf("BAD INODE NUMBER FOR '..' in DIR I=%ju (%s)\n",
(uintmax_t)inp->i_number, pathbuf);
This update utilizes new fsck sysctl commands that allow fsck running in background mode to correct expected inconsistencies that arise during directory rename (see immediately previous update to this file for details). If run on a kernel without the new functionality, background fsck will simply ignore these inconsistencies rather than fail. Reported by: jeff
2010-01-11 20:05:38 +00:00
getpathname(pathbuf, inp->i_dotdot, inp->i_dotdot);
Fix fsck_ffs build with a 64-bit ino_t. Original code by: Gleb Kurtsou
2012-09-27 23:30:58 +00:00
printf("CURRENTLY POINTS TO I=%ju (%s), ",
(uintmax_t)inp->i_dotdot, pathbuf);
This update utilizes new fsck sysctl commands that allow fsck running in background mode to correct expected inconsistencies that arise during directory rename (see immediately previous update to this file for details). If run on a kernel without the new functionality, background fsck will simply ignore these inconsistencies rather than fail. Reported by: jeff
2010-01-11 20:05:38 +00:00
getpathname(pathbuf, inp->i_parent, inp->i_parent);
Fix fsck_ffs build with a 64-bit ino_t. Original code by: Gleb Kurtsou
2012-09-27 23:30:58 +00:00
printf("SHOULD POINT TO I=%ju (%s)",
(uintmax_t)inp->i_parent, pathbuf);
This update utilizes new fsck sysctl commands that allow fsck running in background mode to correct expected inconsistencies that arise during directory rename (see immediately previous update to this file for details). If run on a kernel without the new functionality, background fsck will simply ignore these inconsistencies rather than fail. Reported by: jeff
2010-01-11 20:05:38 +00:00
if (cursnapshot != 0) {
/*
* We need to:
* setcwd(inp->i_number);
* setdotdot(inp->i_dotdot, inp->i_parent);
*/
cmd.value = inp->i_number;
if (sysctlbyname("vfs.ffs.setcwd", 0, 0,
&cmd, sizeof cmd) == -1) {
/* kernel lacks support for these functions */
printf(" (IGNORED)\n");
continue;
}
cmd.value = inp->i_dotdot; /* verify same value */
cmd.size = inp->i_parent; /* new parent */
if (sysctlbyname("vfs.ffs.setdotdot", 0, 0,
&cmd, sizeof cmd) == -1) {
printf(" (FIX FAILED: %s)\n", strerror(errno));
continue;
}
printf(" (FIXED)\n");
inoinfo(inp->i_parent)->ino_linkcnt--;
inp->i_dotdot = inp->i_parent;
continue;
}
if (preen)
printf(" (FIXED)\n");
else if (reply("FIX") == 0)
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
continue;
Reviewed by: Don Lewis <Don.Lewis@tsc.tdk.com> Submitted by: Kirk McKusick <mckusick@McKusick.COM> Obtained from: Mckusick, BSDI and a host of others This exactly matches Kirks sources imported under the Tag MCKUSICK2. These are as supplied by kirk with one small change needed to compile under freeBSD. Some FreeBSD patches will be added back, though many have been added to Kirk's sources already.
1998-12-03 02:27:35 +00:00
inoinfo(inp->i_dotdot)->ino_linkcnt++;
inoinfo(inp->i_parent)->ino_linkcnt--;
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
inp->i_dotdot = inp->i_parent;
Improvement in UFS/FFS directory placement when doing mkdir(2). The algorithm for laying out new directories was devised in the 1980s and markedly improved the performance of the filesystem. In those days large disks had at most 100 cylinder groups and often as few as 10-20. Modern multi-terrabyte disks have thousands of cylinder groups. The original algorithm does not handle these large sizes well. This change attempts to expand the scope of the original algorithm to work well with these much larger disks while still retaining the properties of the original algorithm for small disks. The filesystem implementation is divided into policy routines and implementation routines. The policy routines can be changed in any way desired without risk of corrupting the filesystem. The policy requests are handled by the implementation layer. If the policy asks for an available resource, it is granted. But if it asks for an already in-use resource, then the implementation will provide an available one nearby the request. Thus it is impossible for a policy to double allocate. This change is limited to the policy implementation. This change updates the ffs_dirpref() routine which is responsible for selecting the cylinder group into which a new directory should be placed. If we are near the root of the filesystem we aim to spread them out as much as possible. As we descend deeper from the root we cluster them closer together around their parent as we expect them to be more closely interactive. Higher-level directories like usr/src/sys and usr/src/bin should be separated while the directories in these areas are more likely to be accessed together so should be closer. And directories within commands or kernel subsystems should be closer still. We pick a range of cylinder groups around the cylinder group of the directory in which we are being created. The size of the range for our search is based on our depth from the root of our filesystem. We then probe that range based on how many directories are already present. The first new directory is at 1/2 (middle) of the range; the second is in the first 1/4 of the range, then at 3/4, 1/8, 3/8, 5/8, 7/8, 1/16, 3/16, 5/16, etc. It is desirable to store the depth of a directory in its on-disk inode so that it is available when we need it. We add a new field di_dirdepth to track the depth of each directory. Because there are few spare fields left in the inode, we choose to share an existing field in the inode rather than having one of our own. Specifically we create a union with the di_freelink field. The di_freelink field is used to track inodes that have been unlinked but remain referenced. It is not needed until a rmdir(2) operation has been done on a directory. At that point, the directory has no contents and even if it is kept active as a current directory is no longer able to have any new directories or files created in it. Thus the use of di_dirdepth and di_freelink will never coincide. Reported by: Timo Voelker Reviewed by: kib Tested by: Peter Holm MFC after: 2 weeks Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D39246
2023-03-30 04:09:39 +00:00
(void)changeino(inp->i_number, "..", inp->i_parent,
getinoinfo(inp->i_parent)->i_depth + 1);
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
}
/*
* Mark all the directories that can be found from the root.
*/
propagate();
}
Merge from Lite2. Note that Lite2 has it's own filesystem clean check skipping code that overrides ours sooner. One should be eliminated, but for now it works.
1997-03-11 12:20:21 +00:00
static int
o __P removed o ansi function prototypes o unifdef -D__STDC__ o __dead2 on usage prototype o remove now-bogus main prototype
2002-03-20 22:57:10 +00:00
pass2check(struct inodesc *idesc)
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
{
Remove 'register' keyword. It does not help modern compilers, and some may take some hit from it. (I also found several functions that listed *every* of its 10 local vars with "register" -- just how many free registers do people think machines have?)
2002-03-20 17:55:10 +00:00
struct direct *dirp = idesc->id_dirp;
Add support to background fsck to delete zero-length directories.
2010-05-20 06:05:40 +00:00
char dirname[MAXPATHLEN + 1];
Remove 'register' keyword. It does not help modern compilers, and some may take some hit from it. (I also found several functions that listed *every* of its 10 local vars with "register" -- just how many free registers do people think machines have?)
2002-03-20 17:55:10 +00:00
struct inoinfo *inp;
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
int n, entrysize, ret = 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
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;
Warning cleanup. Format changes by peter
2002-07-30 13:01:25 +00:00
const char *errmsg;
Improve handling of missing '.' and '..' in UFS directories. The UFS filesystem expects to find '.' and '..' as the first two entries in a directory. The kernel's UFS name cache can become quite confused when these two entries are not present as the first two entries. Prior to this change, when the fsck_ffs(8) utility detected that '.' and/or '..' were missing, it would report them, but only offered to replace them if the space at the beginning of the directory was available. Otherwise it was left to the system administrator to move the offending file(s) out of the way and then rerun fsck_ffs(8) to create the '.' and '..' entries. With this change, fsck_ffs(8) will always be able to create the '.' and/or '..' entries. It moves any files in the way elsewhere in the directory block. If there is no room in the directory block to which to move them, they are placed in the lost+found directory. Reported by: Peter Holm Sponsored by: The FreeBSD Foundation
2022-08-17 21:19:59 +00:00
struct direct proto, *newdirp;
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
Remove trailing whitespace.
1995-05-30 06:12:45 +00:00
/*
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
* check for "."
*/
if (idesc->id_entryno != 0)
goto chk1;
if (dirp->d_ino != 0 && strcmp(dirp->d_name, ".") == 0) {
if (dirp->d_ino != idesc->id_number) {
direrror(idesc->id_number, "BAD INODE NUMBER FOR '.'");
Improve handling of missing '.' and '..' in UFS directories. The UFS filesystem expects to find '.' and '..' as the first two entries in a directory. The kernel's UFS name cache can become quite confused when these two entries are not present as the first two entries. Prior to this change, when the fsck_ffs(8) utility detected that '.' and/or '..' were missing, it would report them, but only offered to replace them if the space at the beginning of the directory was available. Otherwise it was left to the system administrator to move the offending file(s) out of the way and then rerun fsck_ffs(8) to create the '.' and '..' entries. With this change, fsck_ffs(8) will always be able to create the '.' and/or '..' entries. It moves any files in the way elsewhere in the directory block. If there is no room in the directory block to which to move them, they are placed in the lost+found directory. Reported by: Peter Holm Sponsored by: The FreeBSD Foundation
2022-08-17 21:19:59 +00:00
if (reply("FIX") == 1) {
dirp->d_ino = idesc->id_number;
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
ret |= ALTERED;
Improve handling of missing '.' and '..' in UFS directories. The UFS filesystem expects to find '.' and '..' as the first two entries in a directory. The kernel's UFS name cache can become quite confused when these two entries are not present as the first two entries. Prior to this change, when the fsck_ffs(8) utility detected that '.' and/or '..' were missing, it would report them, but only offered to replace them if the space at the beginning of the directory was available. Otherwise it was left to the system administrator to move the offending file(s) out of the way and then rerun fsck_ffs(8) to create the '.' and '..' entries. With this change, fsck_ffs(8) will always be able to create the '.' and/or '..' entries. It moves any files in the way elsewhere in the directory block. If there is no room in the directory block to which to move them, they are placed in the lost+found directory. Reported by: Peter Holm Sponsored by: The FreeBSD Foundation
2022-08-17 21:19:59 +00:00
}
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
}
UFS2 preparation commit: Remove support for converting old FFS formats to newer. Submitted by: mckusick Sponspored by: DARPA & NAI Labs.
2002-05-12 23:44:15 +00:00
if (dirp->d_type != DT_DIR) {
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
direrror(idesc->id_number, "BAD TYPE VALUE FOR '.'");
Improve handling of missing '.' and '..' in UFS directories. The UFS filesystem expects to find '.' and '..' as the first two entries in a directory. The kernel's UFS name cache can become quite confused when these two entries are not present as the first two entries. Prior to this change, when the fsck_ffs(8) utility detected that '.' and/or '..' were missing, it would report them, but only offered to replace them if the space at the beginning of the directory was available. Otherwise it was left to the system administrator to move the offending file(s) out of the way and then rerun fsck_ffs(8) to create the '.' and '..' entries. With this change, fsck_ffs(8) will always be able to create the '.' and/or '..' entries. It moves any files in the way elsewhere in the directory block. If there is no room in the directory block to which to move them, they are placed in the lost+found directory. Reported by: Peter Holm Sponsored by: The FreeBSD Foundation
2022-08-17 21:19:59 +00:00
if (reply("FIX") == 1) {
dirp->d_type = DT_DIR;
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
ret |= ALTERED;
Improve handling of missing '.' and '..' in UFS directories. The UFS filesystem expects to find '.' and '..' as the first two entries in a directory. The kernel's UFS name cache can become quite confused when these two entries are not present as the first two entries. Prior to this change, when the fsck_ffs(8) utility detected that '.' and/or '..' were missing, it would report them, but only offered to replace them if the space at the beginning of the directory was available. Otherwise it was left to the system administrator to move the offending file(s) out of the way and then rerun fsck_ffs(8) to create the '.' and '..' entries. With this change, fsck_ffs(8) will always be able to create the '.' and/or '..' entries. It moves any files in the way elsewhere in the directory block. If there is no room in the directory block to which to move them, they are placed in the lost+found directory. Reported by: Peter Holm Sponsored by: The FreeBSD Foundation
2022-08-17 21:19:59 +00:00
}
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
}
goto chk1;
}
proto.d_ino = idesc->id_number;
UFS2 preparation commit: Remove support for converting old FFS formats to newer. Submitted by: mckusick Sponspored by: DARPA & NAI Labs.
2002-05-12 23:44:15 +00:00
proto.d_type = DT_DIR;
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
proto.d_namlen = 1;
(void)strcpy(proto.d_name, ".");
entrysize = DIRSIZ(0, &proto);
Improve handling of missing '.' and '..' in UFS directories. The UFS filesystem expects to find '.' and '..' as the first two entries in a directory. The kernel's UFS name cache can become quite confused when these two entries are not present as the first two entries. Prior to this change, when the fsck_ffs(8) utility detected that '.' and/or '..' were missing, it would report them, but only offered to replace them if the space at the beginning of the directory was available. Otherwise it was left to the system administrator to move the offending file(s) out of the way and then rerun fsck_ffs(8) to create the '.' and '..' entries. With this change, fsck_ffs(8) will always be able to create the '.' and/or '..' entries. It moves any files in the way elsewhere in the directory block. If there is no room in the directory block to which to move them, they are placed in the lost+found directory. Reported by: Peter Holm Sponsored by: The FreeBSD Foundation
2022-08-17 21:19:59 +00:00
direrror(idesc->id_number, "MISSING '.'");
errmsg = "ADD '.' ENTRY";
if (dirp->d_reclen < entrysize + DIRSIZ(0, dirp)) {
/* Not enough space to add '.', replace first entry with '.' */
if (dirp->d_ino != 0) {
pwarn("\nFIRST ENTRY IN DIRECTORY CONTAINS %s\n",
dirp->d_name);
errmsg = "REPLACE WITH '.'";
}
if (reply(errmsg) == 0)
goto chk1;
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
proto.d_reclen = dirp->d_reclen;
Merge from Lite2. Note that Lite2 has it's own filesystem clean check skipping code that overrides ours sooner. One should be eliminated, but for now it works.
1997-03-11 12:20:21 +00:00
memmove(dirp, &proto, (size_t)entrysize);
Improve handling of missing '.' and '..' in UFS directories. The UFS filesystem expects to find '.' and '..' as the first two entries in a directory. The kernel's UFS name cache can become quite confused when these two entries are not present as the first two entries. Prior to this change, when the fsck_ffs(8) utility detected that '.' and/or '..' were missing, it would report them, but only offered to replace them if the space at the beginning of the directory was available. Otherwise it was left to the system administrator to move the offending file(s) out of the way and then rerun fsck_ffs(8) to create the '.' and '..' entries. With this change, fsck_ffs(8) will always be able to create the '.' and/or '..' entries. It moves any files in the way elsewhere in the directory block. If there is no room in the directory block to which to move them, they are placed in the lost+found directory. Reported by: Peter Holm Sponsored by: The FreeBSD Foundation
2022-08-17 21:19:59 +00:00
ret |= ALTERED;
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
} else {
Improve handling of missing '.' and '..' in UFS directories. The UFS filesystem expects to find '.' and '..' as the first two entries in a directory. The kernel's UFS name cache can become quite confused when these two entries are not present as the first two entries. Prior to this change, when the fsck_ffs(8) utility detected that '.' and/or '..' were missing, it would report them, but only offered to replace them if the space at the beginning of the directory was available. Otherwise it was left to the system administrator to move the offending file(s) out of the way and then rerun fsck_ffs(8) to create the '.' and '..' entries. With this change, fsck_ffs(8) will always be able to create the '.' and/or '..' entries. It moves any files in the way elsewhere in the directory block. If there is no room in the directory block to which to move them, they are placed in the lost+found directory. Reported by: Peter Holm Sponsored by: The FreeBSD Foundation
2022-08-17 21:19:59 +00:00
/* Move over first entry and add '.' entry */
if (reply(errmsg) == 0)
goto chk1;
newdirp = (struct direct *)((char *)(dirp) + entrysize);
dirp->d_reclen -= entrysize;
memmove(newdirp, dirp, dirp->d_reclen);
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
proto.d_reclen = entrysize;
Merge from Lite2. Note that Lite2 has it's own filesystem clean check skipping code that overrides ours sooner. One should be eliminated, but for now it works.
1997-03-11 12:20:21 +00:00
memmove(dirp, &proto, (size_t)entrysize);
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
idesc->id_entryno++;
Improve handling of missing '.' and '..' in UFS directories. The UFS filesystem expects to find '.' and '..' as the first two entries in a directory. The kernel's UFS name cache can become quite confused when these two entries are not present as the first two entries. Prior to this change, when the fsck_ffs(8) utility detected that '.' and/or '..' were missing, it would report them, but only offered to replace them if the space at the beginning of the directory was available. Otherwise it was left to the system administrator to move the offending file(s) out of the way and then rerun fsck_ffs(8) to create the '.' and '..' entries. With this change, fsck_ffs(8) will always be able to create the '.' and/or '..' entries. It moves any files in the way elsewhere in the directory block. If there is no room in the directory block to which to move them, they are placed in the lost+found directory. Reported by: Peter Holm Sponsored by: The FreeBSD Foundation
2022-08-17 21:19:59 +00:00
inoinfo(idesc->id_number)->ino_linkcnt--;
dirp = newdirp;
ret |= ALTERED;
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
}
chk1:
if (idesc->id_entryno > 1)
goto chk2;
inp = getinoinfo(idesc->id_number);
proto.d_ino = inp->i_parent;
UFS2 preparation commit: Remove support for converting old FFS formats to newer. Submitted by: mckusick Sponspored by: DARPA & NAI Labs.
2002-05-12 23:44:15 +00:00
proto.d_type = DT_DIR;
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
proto.d_namlen = 2;
(void)strcpy(proto.d_name, "..");
entrysize = DIRSIZ(0, &proto);
if (idesc->id_entryno == 0) {
n = DIRSIZ(0, dirp);
if (dirp->d_reclen < n + entrysize)
goto chk2;
proto.d_reclen = dirp->d_reclen - n;
dirp->d_reclen = n;
idesc->id_entryno++;
Reviewed by: Don Lewis <Don.Lewis@tsc.tdk.com> Submitted by: Kirk McKusick <mckusick@McKusick.COM> Obtained from: Mckusick, BSDI and a host of others This exactly matches Kirks sources imported under the Tag MCKUSICK2. These are as supplied by kirk with one small change needed to compile under freeBSD. Some FreeBSD patches will be added back, though many have been added to Kirk's sources already.
1998-12-03 02:27:35 +00:00
inoinfo(dirp->d_ino)->ino_linkcnt--;
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
dirp = (struct direct *)((char *)(dirp) + n);
Merge from Lite2. Note that Lite2 has it's own filesystem clean check skipping code that overrides ours sooner. One should be eliminated, but for now it works.
1997-03-11 12:20:21 +00:00
memset(dirp, 0, (size_t)proto.d_reclen);
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
dirp->d_reclen = proto.d_reclen;
}
if (dirp->d_ino != 0 && strcmp(dirp->d_name, "..") == 0) {
Correct several bugs in fsck_ffs(8) triggered by corrupted filesystems. If a directory entry has an illegal inode number (less than zero or greater than the last inode in the filesystem) the entry is removed. If a directory '.' or '..' entry had an illegal inode number they were being removed. Since fsck_ffs knows what the correct value is for these two entries fix them rather deleting them. Add much more extensive cylinder group checks and use them to be more careful about rebuilding a cylinder group. Check for out-of-range block numbers before trying to free them. When a directory is deleted also remove its cache entry created in pass1 so that later passes do not try to operate on a deleted directory. Check for ctime(3) returning NULL before trying to use its return. When freeing a directory inode, do not try to interpret it as a directory. Reserve space in the inostatlist to have room to allocate a lost+found directory. If an invalid block number is found past the end of an inode simply remove it rather than clearing and removing the inode. Modernize the inoinfo structure to use queue(3) LIST rather than a handrolled linked list implementation. Reported by: Bob Prohaska, John-Mark Gurney, and Mark Millard Tested by: Peter Holm Reviewed by: Peter Holm MFC after: 2 weeks Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D38668
2023-03-07 23:12:37 +00:00
if (dirp->d_ino > maxino) {
direrror(idesc->id_number, "BAD INODE NUMBER FOR '..'");
/*
* If we know parent set it now, otherwise let it
* point to the root inode and it will get cleaned
* up later if that is not correct.
*/
if (inp->i_parent != 0)
dirp->d_ino = inp->i_parent;
else
dirp->d_ino = UFS_ROOTINO;
if (reply("FIX") == 1)
ret |= ALTERED;
}
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
inp->i_dotdot = dirp->d_ino;
UFS2 preparation commit: Remove support for converting old FFS formats to newer. Submitted by: mckusick Sponspored by: DARPA & NAI Labs.
2002-05-12 23:44:15 +00:00
if (dirp->d_type != DT_DIR) {
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
direrror(idesc->id_number, "BAD TYPE VALUE FOR '..'");
dirp->d_type = DT_DIR;
if (reply("FIX") == 1)
ret |= ALTERED;
}
goto chk2;
}
Improve handling of missing '.' and '..' in UFS directories. The UFS filesystem expects to find '.' and '..' as the first two entries in a directory. The kernel's UFS name cache can become quite confused when these two entries are not present as the first two entries. Prior to this change, when the fsck_ffs(8) utility detected that '.' and/or '..' were missing, it would report them, but only offered to replace them if the space at the beginning of the directory was available. Otherwise it was left to the system administrator to move the offending file(s) out of the way and then rerun fsck_ffs(8) to create the '.' and '..' entries. With this change, fsck_ffs(8) will always be able to create the '.' and/or '..' entries. It moves any files in the way elsewhere in the directory block. If there is no room in the directory block to which to move them, they are placed in the lost+found directory. Reported by: Peter Holm Sponsored by: The FreeBSD Foundation
2022-08-17 21:19:59 +00:00
fileerror(inp->i_parent != 0 ? inp->i_parent : idesc->id_number,
idesc->id_number, "MISSING '..'");
errmsg = "ADD '..' ENTRY";
if (dirp->d_reclen < entrysize + DIRSIZ(0, dirp)) {
/* No space to add '..', replace second entry with '..' */
if (dirp->d_ino != 0) {
pfatal("SECOND ENTRY IN DIRECTORY CONTAINS %s\n",
dirp->d_name);
errmsg = "REPLACE WITH '..'";
}
if (reply(errmsg) == 0) {
inp->i_dotdot = (ino_t)-1;
goto chk2;
}
if (proto.d_ino == 0) {
/* Defer processing until parent known */
idesc->id_entryno++;
if (debug)
printf("(FIX DEFERRED)\n");
}
inp->i_dotdot = proto.d_ino;
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
proto.d_reclen = dirp->d_reclen;
Merge from Lite2. Note that Lite2 has it's own filesystem clean check skipping code that overrides ours sooner. One should be eliminated, but for now it works.
1997-03-11 12:20:21 +00:00
memmove(dirp, &proto, (size_t)entrysize);
Improve handling of missing '.' and '..' in UFS directories. The UFS filesystem expects to find '.' and '..' as the first two entries in a directory. The kernel's UFS name cache can become quite confused when these two entries are not present as the first two entries. Prior to this change, when the fsck_ffs(8) utility detected that '.' and/or '..' were missing, it would report them, but only offered to replace them if the space at the beginning of the directory was available. Otherwise it was left to the system administrator to move the offending file(s) out of the way and then rerun fsck_ffs(8) to create the '.' and '..' entries. With this change, fsck_ffs(8) will always be able to create the '.' and/or '..' entries. It moves any files in the way elsewhere in the directory block. If there is no room in the directory block to which to move them, they are placed in the lost+found directory. Reported by: Peter Holm Sponsored by: The FreeBSD Foundation
2022-08-17 21:19:59 +00:00
ret |= ALTERED;
} else {
/* Move over second entry and add '..' entry */
if (reply(errmsg) == 0) {
inp->i_dotdot = (ino_t)-1;
goto chk2;
}
if (proto.d_ino == 0) {
/* Defer processing until parent known */
idesc->id_entryno++;
if (debug)
printf("(FIX DEFERRED)\n");
}
inp->i_dotdot = proto.d_ino;
if (dirp->d_ino == 0) {
proto.d_reclen = dirp->d_reclen;
memmove(dirp, &proto, (size_t)entrysize);
} else {
newdirp = (struct direct *)((char *)(dirp) + entrysize);
dirp->d_reclen -= entrysize;
memmove(newdirp, dirp, dirp->d_reclen);
proto.d_reclen = entrysize;
memmove(dirp, &proto, (size_t)entrysize);
if (dirp->d_ino != 0) {
idesc->id_entryno++;
inoinfo(dirp->d_ino)->ino_linkcnt--;
}
dirp = newdirp;
}
ret |= ALTERED;
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
}
chk2:
if (dirp->d_ino == 0)
return (ret|KEEPON);
if (dirp->d_namlen <= 2 &&
dirp->d_name[0] == '.' &&
idesc->id_entryno >= 2) {
if (dirp->d_namlen == 1) {
direrror(idesc->id_number, "EXTRA '.' ENTRY");
dirp->d_ino = 0;
if (reply("FIX") == 1)
ret |= ALTERED;
return (KEEPON | ret);
}
if (dirp->d_name[1] == '.') {
direrror(idesc->id_number, "EXTRA '..' ENTRY");
dirp->d_ino = 0;
if (reply("FIX") == 1)
ret |= ALTERED;
return (KEEPON | ret);
}
}
idesc->id_entryno++;
n = 0;
if (dirp->d_ino > maxino) {
fileerror(idesc->id_number, dirp->d_ino, "I OUT OF RANGE");
n = reply("REMOVE");
prefix UFS symbols with UFS_ to reduce namespace pollution Specifically: ROOTINO -> UFS_ROOTINO WINO -> UFS_WINO NXADDR -> UFS_NXADDR NDADDR -> UFS_NDADDR NIADDR -> UFS_NIADDR MAXSYMLINKLEN_UFS[12] -> UFS[12]_MAXSYMLINKLEN (for consistency) Also prefix ext2's and nandfs's NDADDR and NIADDR with EXT2_ and NANDFS_ Reviewed by: kib, mckusick Obtained from: NetBSD MFC after: 1 month Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D9536
2017-02-15 19:50:26 +00:00
} else if (((dirp->d_ino == UFS_WINO && dirp->d_type != DT_WHT) ||
(dirp->d_ino != UFS_WINO && dirp->d_type == DT_WHT))) {
Merge from Lite2. Note that Lite2 has it's own filesystem clean check skipping code that overrides ours sooner. One should be eliminated, but for now it works.
1997-03-11 12:20:21 +00:00
fileerror(idesc->id_number, dirp->d_ino, "BAD WHITEOUT ENTRY");
prefix UFS symbols with UFS_ to reduce namespace pollution Specifically: ROOTINO -> UFS_ROOTINO WINO -> UFS_WINO NXADDR -> UFS_NXADDR NDADDR -> UFS_NDADDR NIADDR -> UFS_NIADDR MAXSYMLINKLEN_UFS[12] -> UFS[12]_MAXSYMLINKLEN (for consistency) Also prefix ext2's and nandfs's NDADDR and NIADDR with EXT2_ and NANDFS_ Reviewed by: kib, mckusick Obtained from: NetBSD MFC after: 1 month Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D9536
2017-02-15 19:50:26 +00:00
dirp->d_ino = UFS_WINO;
Merge from Lite2. Note that Lite2 has it's own filesystem clean check skipping code that overrides ours sooner. One should be eliminated, but for now it works.
1997-03-11 12:20:21 +00:00
dirp->d_type = DT_WHT;
if (reply("FIX") == 1)
ret |= ALTERED;
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
} else {
again:
Reviewed by: Don Lewis <Don.Lewis@tsc.tdk.com> Submitted by: Kirk McKusick <mckusick@McKusick.COM> Obtained from: Mckusick, BSDI and a host of others This exactly matches Kirks sources imported under the Tag MCKUSICK2. These are as supplied by kirk with one small change needed to compile under freeBSD. Some FreeBSD patches will be added back, though many have been added to Kirk's sources already.
1998-12-03 02:27:35 +00:00
switch (inoinfo(dirp->d_ino)->ino_state) {
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
case USTATE:
if (idesc->id_entryno <= 2)
break;
fileerror(idesc->id_number, dirp->d_ino, "UNALLOCATED");
n = reply("REMOVE");
break;
case DCLEAR:
case FCLEAR:
if (idesc->id_entryno <= 2)
break;
Reviewed by: Don Lewis <Don.Lewis@tsc.tdk.com> Submitted by: Kirk McKusick <mckusick@McKusick.COM> Obtained from: Mckusick, BSDI and a host of others This exactly matches Kirks sources imported under the Tag MCKUSICK2. These are as supplied by kirk with one small change needed to compile under freeBSD. Some FreeBSD patches will be added back, though many have been added to Kirk's sources already.
1998-12-03 02:27:35 +00:00
if (inoinfo(dirp->d_ino)->ino_state == FCLEAR)
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
errmsg = "DUP/BAD";
Reviewed by: dyson@freebsd.org (john Dyson), dg@root.com (david greenman) Submitted by: Kirk McKusick (mcKusick@mckusick.com) Obtained from: WHistle development tree
1998-03-08 09:59:44 +00:00
else if (!preen && !usedsoftdep)
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
errmsg = "ZERO LENGTH DIRECTORY";
Add support to background fsck to delete zero-length directories.
2010-05-20 06:05:40 +00:00
else if (cursnapshot == 0) {
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
n = 1;
break;
Add support to background fsck to delete zero-length directories.
2010-05-20 06:05:40 +00:00
} else {
getpathname(dirname, idesc->id_number,
dirp->d_ino);
Fix fsck_ffs build with a 64-bit ino_t. Original code by: Gleb Kurtsou
2012-09-27 23:30:58 +00:00
pwarn("ZERO LENGTH DIRECTORY %s I=%ju",
dirname, (uintmax_t)dirp->d_ino);
Add support to background fsck to delete zero-length directories.
2010-05-20 06:05:40 +00:00
/*
* We need to:
* setcwd(idesc->id_parent);
* rmdir(dirp->d_name);
*/
cmd.value = idesc->id_number;
if (sysctlbyname("vfs.ffs.setcwd", 0, 0,
&cmd, sizeof cmd) == -1) {
/* kernel lacks support */
printf(" (IGNORED)\n");
n = 1;
break;
}
if (rmdir(dirp->d_name) == -1) {
printf(" (REMOVAL FAILED: %s)\n",
strerror(errno));
n = 1;
break;
}
/* ".." reference to parent is removed */
inoinfo(idesc->id_number)->ino_linkcnt--;
printf(" (REMOVED)\n");
break;
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
}
fileerror(idesc->id_number, dirp->d_ino, errmsg);
if ((n = reply("REMOVE")) == 1)
break;
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(dirp->d_ino, &ip);
dp = ip.i_dp;
Reviewed by: Don Lewis <Don.Lewis@tsc.tdk.com> Submitted by: Kirk McKusick <mckusick@McKusick.COM> Obtained from: Mckusick, BSDI and a host of others This exactly matches Kirks sources imported under the Tag MCKUSICK2. These are as supplied by kirk with one small change needed to compile under freeBSD. Some FreeBSD patches will be added back, though many have been added to Kirk's sources already.
1998-12-03 02:27:35 +00:00
inoinfo(dirp->d_ino)->ino_state =
Revert r313780 (UFS_ prefix)
2018-03-17 12:59:55 +00:00
(DIP(dp, di_mode) & IFMT) == IFDIR ? DSTATE : FSTATE;
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(dirp->d_ino)->ino_linkcnt = DIP(dp, di_nlink);
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);
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
goto again;
case DSTATE:
Eliminate linked list used to track inodes with an initial link count of zero and instead encode this information in the inode state. Pass 4 performed a linear search of this list for each inode in the file system, which performs poorly if the list is long. Reviewed by: sam & keramida (an earlier version of the patch), mckusick MFC after: 1 month
2004-10-08 20:44:47 +00:00
case DZLINK:
Reviewed by: Don Lewis <Don.Lewis@tsc.tdk.com> Submitted by: Kirk McKusick <mckusick@McKusick.COM> Obtained from: Mckusick, BSDI and a host of others This exactly matches Kirks sources imported under the Tag MCKUSICK2. These are as supplied by kirk with one small change needed to compile under freeBSD. Some FreeBSD patches will be added back, though many have been added to Kirk's sources already.
1998-12-03 02:27:35 +00:00
if (inoinfo(idesc->id_number)->ino_state == DFOUND)
inoinfo(dirp->d_ino)->ino_state = DFOUND;
Replace various spelling with FALLTHROUGH which is lint()able
2002-08-25 13:10:45 +00:00
/* FALLTHROUGH */
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
case DFOUND:
inp = getinoinfo(dirp->d_ino);
When renaming a directory it passes through several intermediate states. First its new name will be created causing it to have two names (from possibly different parents). Next, if it has different parents, its value of ".." will be changed from pointing to the old parent to pointing to the new parent. Concurrently, its old name will be removed bringing it back into a consistent state. When fsck encounters an extra name for a directory, it offers to remove the "extraneous hard link"; when it finds that the names have been changed but the update to ".." has not happened, it offers to rewrite ".." to point at the correct parent. Both of these changes were considered unexpected so would cause fsck in preen mode or fsck in background mode to fail with the need to run fsck manually to fix these problems. This update changes these errors to be expected so that in preen mode fsck will simply fix these transitional errors. For now, background fsck will note these errors, but will need additional kernel support to fix them, so will simply ignore them rather than fail. A future update will allow background fsck to fix these problems. Reported by: jeff
2010-01-11 19:52:40 +00:00
if (idesc->id_entryno > 2) {
Improvement in UFS/FFS directory placement when doing mkdir(2). The algorithm for laying out new directories was devised in the 1980s and markedly improved the performance of the filesystem. In those days large disks had at most 100 cylinder groups and often as few as 10-20. Modern multi-terrabyte disks have thousands of cylinder groups. The original algorithm does not handle these large sizes well. This change attempts to expand the scope of the original algorithm to work well with these much larger disks while still retaining the properties of the original algorithm for small disks. The filesystem implementation is divided into policy routines and implementation routines. The policy routines can be changed in any way desired without risk of corrupting the filesystem. The policy requests are handled by the implementation layer. If the policy asks for an available resource, it is granted. But if it asks for an already in-use resource, then the implementation will provide an available one nearby the request. Thus it is impossible for a policy to double allocate. This change is limited to the policy implementation. This change updates the ffs_dirpref() routine which is responsible for selecting the cylinder group into which a new directory should be placed. If we are near the root of the filesystem we aim to spread them out as much as possible. As we descend deeper from the root we cluster them closer together around their parent as we expect them to be more closely interactive. Higher-level directories like usr/src/sys and usr/src/bin should be separated while the directories in these areas are more likely to be accessed together so should be closer. And directories within commands or kernel subsystems should be closer still. We pick a range of cylinder groups around the cylinder group of the directory in which we are being created. The size of the range for our search is based on our depth from the root of our filesystem. We then probe that range based on how many directories are already present. The first new directory is at 1/2 (middle) of the range; the second is in the first 1/4 of the range, then at 3/4, 1/8, 3/8, 5/8, 7/8, 1/16, 3/16, 5/16, etc. It is desirable to store the depth of a directory in its on-disk inode so that it is available when we need it. We add a new field di_dirdepth to track the depth of each directory. Because there are few spare fields left in the inode, we choose to share an existing field in the inode rather than having one of our own. Specifically we create a union with the di_freelink field. The di_freelink field is used to track inodes that have been unlinked but remain referenced. It is not needed until a rmdir(2) operation has been done on a directory. At that point, the directory has no contents and even if it is kept active as a current directory is no longer able to have any new directories or files created in it. Thus the use of di_dirdepth and di_freelink will never coincide. Reported by: Timo Voelker Reviewed by: kib Tested by: Peter Holm MFC after: 2 weeks Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D39246
2023-03-30 04:09:39 +00:00
if (inp->i_parent == 0) {
When renaming a directory it passes through several intermediate states. First its new name will be created causing it to have two names (from possibly different parents). Next, if it has different parents, its value of ".." will be changed from pointing to the old parent to pointing to the new parent. Concurrently, its old name will be removed bringing it back into a consistent state. When fsck encounters an extra name for a directory, it offers to remove the "extraneous hard link"; when it finds that the names have been changed but the update to ".." has not happened, it offers to rewrite ".." to point at the correct parent. Both of these changes were considered unexpected so would cause fsck in preen mode or fsck in background mode to fail with the need to run fsck manually to fix these problems. This update changes these errors to be expected so that in preen mode fsck will simply fix these transitional errors. For now, background fsck will note these errors, but will need additional kernel support to fix them, so will simply ignore them rather than fail. A future update will allow background fsck to fix these problems. Reported by: jeff
2010-01-11 19:52:40 +00:00
inp->i_parent = idesc->id_number;
Improvement in UFS/FFS directory placement when doing mkdir(2). The algorithm for laying out new directories was devised in the 1980s and markedly improved the performance of the filesystem. In those days large disks had at most 100 cylinder groups and often as few as 10-20. Modern multi-terrabyte disks have thousands of cylinder groups. The original algorithm does not handle these large sizes well. This change attempts to expand the scope of the original algorithm to work well with these much larger disks while still retaining the properties of the original algorithm for small disks. The filesystem implementation is divided into policy routines and implementation routines. The policy routines can be changed in any way desired without risk of corrupting the filesystem. The policy requests are handled by the implementation layer. If the policy asks for an available resource, it is granted. But if it asks for an already in-use resource, then the implementation will provide an available one nearby the request. Thus it is impossible for a policy to double allocate. This change is limited to the policy implementation. This change updates the ffs_dirpref() routine which is responsible for selecting the cylinder group into which a new directory should be placed. If we are near the root of the filesystem we aim to spread them out as much as possible. As we descend deeper from the root we cluster them closer together around their parent as we expect them to be more closely interactive. Higher-level directories like usr/src/sys and usr/src/bin should be separated while the directories in these areas are more likely to be accessed together so should be closer. And directories within commands or kernel subsystems should be closer still. We pick a range of cylinder groups around the cylinder group of the directory in which we are being created. The size of the range for our search is based on our depth from the root of our filesystem. We then probe that range based on how many directories are already present. The first new directory is at 1/2 (middle) of the range; the second is in the first 1/4 of the range, then at 3/4, 1/8, 3/8, 5/8, 7/8, 1/16, 3/16, 5/16, etc. It is desirable to store the depth of a directory in its on-disk inode so that it is available when we need it. We add a new field di_dirdepth to track the depth of each directory. Because there are few spare fields left in the inode, we choose to share an existing field in the inode rather than having one of our own. Specifically we create a union with the di_freelink field. The di_freelink field is used to track inodes that have been unlinked but remain referenced. It is not needed until a rmdir(2) operation has been done on a directory. At that point, the directory has no contents and even if it is kept active as a current directory is no longer able to have any new directories or files created in it. Thus the use of di_dirdepth and di_freelink will never coincide. Reported by: Timo Voelker Reviewed by: kib Tested by: Peter Holm MFC after: 2 weeks Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D39246
2023-03-30 04:09:39 +00:00
check_dirdepth(inp);
} else if ((n = fix_extraneous(inp, idesc))) {
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
break;
Improvement in UFS/FFS directory placement when doing mkdir(2). The algorithm for laying out new directories was devised in the 1980s and markedly improved the performance of the filesystem. In those days large disks had at most 100 cylinder groups and often as few as 10-20. Modern multi-terrabyte disks have thousands of cylinder groups. The original algorithm does not handle these large sizes well. This change attempts to expand the scope of the original algorithm to work well with these much larger disks while still retaining the properties of the original algorithm for small disks. The filesystem implementation is divided into policy routines and implementation routines. The policy routines can be changed in any way desired without risk of corrupting the filesystem. The policy requests are handled by the implementation layer. If the policy asks for an available resource, it is granted. But if it asks for an already in-use resource, then the implementation will provide an available one nearby the request. Thus it is impossible for a policy to double allocate. This change is limited to the policy implementation. This change updates the ffs_dirpref() routine which is responsible for selecting the cylinder group into which a new directory should be placed. If we are near the root of the filesystem we aim to spread them out as much as possible. As we descend deeper from the root we cluster them closer together around their parent as we expect them to be more closely interactive. Higher-level directories like usr/src/sys and usr/src/bin should be separated while the directories in these areas are more likely to be accessed together so should be closer. And directories within commands or kernel subsystems should be closer still. We pick a range of cylinder groups around the cylinder group of the directory in which we are being created. The size of the range for our search is based on our depth from the root of our filesystem. We then probe that range based on how many directories are already present. The first new directory is at 1/2 (middle) of the range; the second is in the first 1/4 of the range, then at 3/4, 1/8, 3/8, 5/8, 7/8, 1/16, 3/16, 5/16, etc. It is desirable to store the depth of a directory in its on-disk inode so that it is available when we need it. We add a new field di_dirdepth to track the depth of each directory. Because there are few spare fields left in the inode, we choose to share an existing field in the inode rather than having one of our own. Specifically we create a union with the di_freelink field. The di_freelink field is used to track inodes that have been unlinked but remain referenced. It is not needed until a rmdir(2) operation has been done on a directory. At that point, the directory has no contents and even if it is kept active as a current directory is no longer able to have any new directories or files created in it. Thus the use of di_dirdepth and di_freelink will never coincide. Reported by: Timo Voelker Reviewed by: kib Tested by: Peter Holm MFC after: 2 weeks Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D39246
2023-03-30 04:09:39 +00:00
}
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
}
Replace various spelling with FALLTHROUGH which is lint()able
2002-08-25 13:10:45 +00:00
/* FALLTHROUGH */
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
case FSTATE:
Eliminate linked list used to track inodes with an initial link count of zero and instead encode this information in the inode state. Pass 4 performed a linear search of this list for each inode in the file system, which performs poorly if the list is long. Reviewed by: sam & keramida (an earlier version of the patch), mckusick MFC after: 1 month
2004-10-08 20:44:47 +00:00
case FZLINK:
UFS2 preparation commit: Remove support for converting old FFS formats to newer. Submitted by: mckusick Sponspored by: DARPA & NAI Labs.
2002-05-12 23:44:15 +00:00
if (dirp->d_type != inoinfo(dirp->d_ino)->ino_type) {
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
fileerror(idesc->id_number, dirp->d_ino,
"BAD TYPE VALUE");
Reviewed by: Don Lewis <Don.Lewis@tsc.tdk.com> Submitted by: Kirk McKusick <mckusick@McKusick.COM> Obtained from: Mckusick, BSDI and a host of others This exactly matches Kirks sources imported under the Tag MCKUSICK2. These are as supplied by kirk with one small change needed to compile under freeBSD. Some FreeBSD patches will be added back, though many have been added to Kirk's sources already.
1998-12-03 02:27:35 +00:00
dirp->d_type = inoinfo(dirp->d_ino)->ino_type;
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
if (reply("FIX") == 1)
ret |= ALTERED;
}
Reviewed by: Don Lewis <Don.Lewis@tsc.tdk.com> Submitted by: Kirk McKusick <mckusick@McKusick.COM> Obtained from: Mckusick, BSDI and a host of others This exactly matches Kirks sources imported under the Tag MCKUSICK2. These are as supplied by kirk with one small change needed to compile under freeBSD. Some FreeBSD patches will be added back, though many have been added to Kirk's sources already.
1998-12-03 02:27:35 +00:00
inoinfo(dirp->d_ino)->ino_linkcnt--;
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
break;
default:
Fix fsck_ffs build with a 64-bit ino_t. Original code by: Gleb Kurtsou
2012-09-27 23:30:58 +00:00
errx(EEXIT, "BAD STATE %d FOR INODE I=%ju",
inoinfo(dirp->d_ino)->ino_state,
(uintmax_t)dirp->d_ino);
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
}
}
if (n == 0)
return (ret|KEEPON);
dirp->d_ino = 0;
return (ret|KEEPON|ALTERED);
}
When renaming a directory it passes through several intermediate states. First its new name will be created causing it to have two names (from possibly different parents). Next, if it has different parents, its value of ".." will be changed from pointing to the old parent to pointing to the new parent. Concurrently, its old name will be removed bringing it back into a consistent state. When fsck encounters an extra name for a directory, it offers to remove the "extraneous hard link"; when it finds that the names have been changed but the update to ".." has not happened, it offers to rewrite ".." to point at the correct parent. Both of these changes were considered unexpected so would cause fsck in preen mode or fsck in background mode to fail with the need to run fsck manually to fix these problems. This update changes these errors to be expected so that in preen mode fsck will simply fix these transitional errors. For now, background fsck will note these errors, but will need additional kernel support to fix them, so will simply ignore them rather than fail. A future update will allow background fsck to fix these problems. Reported by: jeff
2010-01-11 19:52:40 +00:00
static int
fix_extraneous(struct inoinfo *inp, struct inodesc *idesc)
{
This update utilizes new fsck sysctl commands that allow fsck running in background mode to correct expected inconsistencies that arise during directory rename (see immediately previous update to this file for details). If run on a kernel without the new functionality, background fsck will simply ignore these inconsistencies rather than fail. Reported by: jeff
2010-01-11 20:05:38 +00:00
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;
When renaming a directory it passes through several intermediate states. First its new name will be created causing it to have two names (from possibly different parents). Next, if it has different parents, its value of ".." will be changed from pointing to the old parent to pointing to the new parent. Concurrently, its old name will be removed bringing it back into a consistent state. When fsck encounters an extra name for a directory, it offers to remove the "extraneous hard link"; when it finds that the names have been changed but the update to ".." has not happened, it offers to rewrite ".." to point at the correct parent. Both of these changes were considered unexpected so would cause fsck in preen mode or fsck in background mode to fail with the need to run fsck manually to fix these problems. This update changes these errors to be expected so that in preen mode fsck will simply fix these transitional errors. For now, background fsck will note these errors, but will need additional kernel support to fix them, so will simply ignore them rather than fail. A future update will allow background fsck to fix these problems. Reported by: jeff
2010-01-11 19:52:40 +00:00
struct inodesc dotdesc;
char oldname[MAXPATHLEN + 1];
char newname[MAXPATHLEN + 1];
Mechanical whitespace cleanup. MFC after: 3 weeks
2011-04-27 02:55:03 +00:00
When renaming a directory it passes through several intermediate states. First its new name will be created causing it to have two names (from possibly different parents). Next, if it has different parents, its value of ".." will be changed from pointing to the old parent to pointing to the new parent. Concurrently, its old name will be removed bringing it back into a consistent state. When fsck encounters an extra name for a directory, it offers to remove the "extraneous hard link"; when it finds that the names have been changed but the update to ".." has not happened, it offers to rewrite ".." to point at the correct parent. Both of these changes were considered unexpected so would cause fsck in preen mode or fsck in background mode to fail with the need to run fsck manually to fix these problems. This update changes these errors to be expected so that in preen mode fsck will simply fix these transitional errors. For now, background fsck will note these errors, but will need additional kernel support to fix them, so will simply ignore them rather than fail. A future update will allow background fsck to fix these problems. Reported by: jeff
2010-01-11 19:52:40 +00:00
/*
* If we have not yet found "..", look it up now so we know
* which inode the directory itself believes is its parent.
*/
if (inp->i_dotdot == 0) {
memset(&dotdesc, 0, sizeof(struct inodesc));
dotdesc.id_type = DATA;
dotdesc.id_number = idesc->id_dirp->d_ino;
dotdesc.id_func = findino;
dotdesc.id_name = strdup("..");
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(dotdesc.id_number, &ip);
if ((ckinode(ip.i_dp, &dotdesc) & FOUND))
When renaming a directory it passes through several intermediate states. First its new name will be created causing it to have two names (from possibly different parents). Next, if it has different parents, its value of ".." will be changed from pointing to the old parent to pointing to the new parent. Concurrently, its old name will be removed bringing it back into a consistent state. When fsck encounters an extra name for a directory, it offers to remove the "extraneous hard link"; when it finds that the names have been changed but the update to ".." has not happened, it offers to rewrite ".." to point at the correct parent. Both of these changes were considered unexpected so would cause fsck in preen mode or fsck in background mode to fail with the need to run fsck manually to fix these problems. This update changes these errors to be expected so that in preen mode fsck will simply fix these transitional errors. For now, background fsck will note these errors, but will need additional kernel support to fix them, so will simply ignore them rather than fail. A future update will allow background fsck to fix these problems. Reported by: jeff
2010-01-11 19:52:40 +00:00
inp->i_dotdot = dotdesc.id_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);
Rewrite function definitions with identifier lists. A few functions snuck in with K&R style definitions. Also add some missing memory frees. MFC after: 1 week
2023-03-22 22:57:26 +00:00
free(dotdesc.id_name);
When renaming a directory it passes through several intermediate states. First its new name will be created causing it to have two names (from possibly different parents). Next, if it has different parents, its value of ".." will be changed from pointing to the old parent to pointing to the new parent. Concurrently, its old name will be removed bringing it back into a consistent state. When fsck encounters an extra name for a directory, it offers to remove the "extraneous hard link"; when it finds that the names have been changed but the update to ".." has not happened, it offers to rewrite ".." to point at the correct parent. Both of these changes were considered unexpected so would cause fsck in preen mode or fsck in background mode to fail with the need to run fsck manually to fix these problems. This update changes these errors to be expected so that in preen mode fsck will simply fix these transitional errors. For now, background fsck will note these errors, but will need additional kernel support to fix them, so will simply ignore them rather than fail. A future update will allow background fsck to fix these problems. Reported by: jeff
2010-01-11 19:52:40 +00:00
}
/*
* We have the previously found old name (inp->i_parent) and the
* just found new name (idesc->id_number). We have five cases:
* 1) ".." is missing - can remove either name, choose to delete
* new one and let fsck create ".." pointing to old name.
* 2) Both new and old are in same directory, choose to delete
* the new name and let fsck fix ".." if it is wrong.
* 3) ".." does not point to the new name, so delete it and let
* fsck fix ".." to point to the old one if it is wrong.
* 4) ".." points to the old name only, so delete the new one.
* 5) ".." points to the new name only, so delete the old one.
*
* For cases 1-4 we eliminate the new name;
* for case 5 we eliminate the old name.
*/
if (inp->i_dotdot == 0 || /* Case 1 */
idesc->id_number == inp->i_parent || /* Case 2 */
inp->i_dotdot != idesc->id_number || /* Case 3 */
inp->i_dotdot == inp->i_parent) { /* Case 4 */
getpathname(newname, idesc->id_number, idesc->id_number);
if (strcmp(newname, "/") != 0)
strcat (newname, "/");
strcat(newname, idesc->id_dirp->d_name);
getpathname(oldname, inp->i_number, inp->i_number);
This update utilizes new fsck sysctl commands that allow fsck running in background mode to correct expected inconsistencies that arise during directory rename (see immediately previous update to this file for details). If run on a kernel without the new functionality, background fsck will simply ignore these inconsistencies rather than fail. Reported by: jeff
2010-01-11 20:05:38 +00:00
pwarn("%s IS AN EXTRANEOUS HARD LINK TO DIRECTORY %s",
When renaming a directory it passes through several intermediate states. First its new name will be created causing it to have two names (from possibly different parents). Next, if it has different parents, its value of ".." will be changed from pointing to the old parent to pointing to the new parent. Concurrently, its old name will be removed bringing it back into a consistent state. When fsck encounters an extra name for a directory, it offers to remove the "extraneous hard link"; when it finds that the names have been changed but the update to ".." has not happened, it offers to rewrite ".." to point at the correct parent. Both of these changes were considered unexpected so would cause fsck in preen mode or fsck in background mode to fail with the need to run fsck manually to fix these problems. This update changes these errors to be expected so that in preen mode fsck will simply fix these transitional errors. For now, background fsck will note these errors, but will need additional kernel support to fix them, so will simply ignore them rather than fail. A future update will allow background fsck to fix these problems. Reported by: jeff
2010-01-11 19:52:40 +00:00
newname, oldname);
if (cursnapshot != 0) {
/*
* We need to
* setcwd(idesc->id_number);
* unlink(idesc->id_dirp->d_name);
*/
This update utilizes new fsck sysctl commands that allow fsck running in background mode to correct expected inconsistencies that arise during directory rename (see immediately previous update to this file for details). If run on a kernel without the new functionality, background fsck will simply ignore these inconsistencies rather than fail. Reported by: jeff
2010-01-11 20:05:38 +00:00
cmd.value = idesc->id_number;
if (sysctlbyname("vfs.ffs.setcwd", 0, 0,
&cmd, sizeof cmd) == -1) {
printf(" (IGNORED)\n");
return (0);
}
Cast 64-bit quantity to intptr_t rather than int so as to work properly with 64-bit architectures (such as amd64). Reported by: Xin LI and Josh Paetzel
2010-01-11 23:33:30 +00:00
cmd.value = (intptr_t)idesc->id_dirp->d_name;
This update utilizes new fsck sysctl commands that allow fsck running in background mode to correct expected inconsistencies that arise during directory rename (see immediately previous update to this file for details). If run on a kernel without the new functionality, background fsck will simply ignore these inconsistencies rather than fail. Reported by: jeff
2010-01-11 20:05:38 +00:00
cmd.size = inp->i_number; /* verify same name */
if (sysctlbyname("vfs.ffs.unlink", 0, 0,
&cmd, sizeof cmd) == -1) {
printf(" (UNLINK FAILED: %s)\n",
strerror(errno));
return (0);
}
printf(" (REMOVED)\n");
When renaming a directory it passes through several intermediate states. First its new name will be created causing it to have two names (from possibly different parents). Next, if it has different parents, its value of ".." will be changed from pointing to the old parent to pointing to the new parent. Concurrently, its old name will be removed bringing it back into a consistent state. When fsck encounters an extra name for a directory, it offers to remove the "extraneous hard link"; when it finds that the names have been changed but the update to ".." has not happened, it offers to rewrite ".." to point at the correct parent. Both of these changes were considered unexpected so would cause fsck in preen mode or fsck in background mode to fail with the need to run fsck manually to fix these problems. This update changes these errors to be expected so that in preen mode fsck will simply fix these transitional errors. For now, background fsck will note these errors, but will need additional kernel support to fix them, so will simply ignore them rather than fail. A future update will allow background fsck to fix these problems. Reported by: jeff
2010-01-11 19:52:40 +00:00
return (0);
}
if (preen) {
printf(" (REMOVED)\n");
return (1);
}
return (reply("REMOVE"));
}
/*
* None of the first four cases above, so must be case (5).
* Eliminate the old name and make the new the name the parent.
*/
getpathname(oldname, inp->i_parent, inp->i_number);
getpathname(newname, inp->i_number, inp->i_number);
This update utilizes new fsck sysctl commands that allow fsck running in background mode to correct expected inconsistencies that arise during directory rename (see immediately previous update to this file for details). If run on a kernel without the new functionality, background fsck will simply ignore these inconsistencies rather than fail. Reported by: jeff
2010-01-11 20:05:38 +00:00
pwarn("%s IS AN EXTRANEOUS HARD LINK TO DIRECTORY %s", oldname,
When renaming a directory it passes through several intermediate states. First its new name will be created causing it to have two names (from possibly different parents). Next, if it has different parents, its value of ".." will be changed from pointing to the old parent to pointing to the new parent. Concurrently, its old name will be removed bringing it back into a consistent state. When fsck encounters an extra name for a directory, it offers to remove the "extraneous hard link"; when it finds that the names have been changed but the update to ".." has not happened, it offers to rewrite ".." to point at the correct parent. Both of these changes were considered unexpected so would cause fsck in preen mode or fsck in background mode to fail with the need to run fsck manually to fix these problems. This update changes these errors to be expected so that in preen mode fsck will simply fix these transitional errors. For now, background fsck will note these errors, but will need additional kernel support to fix them, so will simply ignore them rather than fail. A future update will allow background fsck to fix these problems. Reported by: jeff
2010-01-11 19:52:40 +00:00
newname);
if (cursnapshot != 0) {
/*
* We need to
* setcwd(inp->i_parent);
* unlink(last component of oldname pathname);
*/
This update utilizes new fsck sysctl commands that allow fsck running in background mode to correct expected inconsistencies that arise during directory rename (see immediately previous update to this file for details). If run on a kernel without the new functionality, background fsck will simply ignore these inconsistencies rather than fail. Reported by: jeff
2010-01-11 20:05:38 +00:00
cmd.value = inp->i_parent;
if (sysctlbyname("vfs.ffs.setcwd", 0, 0,
&cmd, sizeof cmd) == -1) {
printf(" (IGNORED)\n");
return (0);
}
Replace index() and rindex() calls with strchr() and strrchr(). The index() and rindex() functions were marked LEGACY in the 2001 revision of POSIX and were subsequently removed from the 2008 revision. The strchr() and strrchr() functions are part of the C standard. This makes the source code a lot more consistent, as most of these C files also call into other str*() routines. In fact, about a dozen already perform strchr() calls.
2012-01-03 18:51:58 +00:00
if ((cp = strchr(oldname, '/')) == NULL) {
This update utilizes new fsck sysctl commands that allow fsck running in background mode to correct expected inconsistencies that arise during directory rename (see immediately previous update to this file for details). If run on a kernel without the new functionality, background fsck will simply ignore these inconsistencies rather than fail. Reported by: jeff
2010-01-11 20:05:38 +00:00
printf(" (IGNORED)\n");
return (0);
}
Cast 64-bit quantity to intptr_t rather than int so as to work properly with 64-bit architectures (such as amd64). Reported by: Xin LI and Josh Paetzel
2010-01-11 23:33:30 +00:00
cmd.value = (intptr_t)(cp + 1);
This update utilizes new fsck sysctl commands that allow fsck running in background mode to correct expected inconsistencies that arise during directory rename (see immediately previous update to this file for details). If run on a kernel without the new functionality, background fsck will simply ignore these inconsistencies rather than fail. Reported by: jeff
2010-01-11 20:05:38 +00:00
cmd.size = inp->i_number; /* verify same name */
if (sysctlbyname("vfs.ffs.unlink", 0, 0,
&cmd, sizeof cmd) == -1) {
printf(" (UNLINK FAILED: %s)\n",
strerror(errno));
return (0);
}
printf(" (REMOVED)\n");
inp->i_parent = idesc->id_number; /* reparent to correct dir */
When renaming a directory it passes through several intermediate states. First its new name will be created causing it to have two names (from possibly different parents). Next, if it has different parents, its value of ".." will be changed from pointing to the old parent to pointing to the new parent. Concurrently, its old name will be removed bringing it back into a consistent state. When fsck encounters an extra name for a directory, it offers to remove the "extraneous hard link"; when it finds that the names have been changed but the update to ".." has not happened, it offers to rewrite ".." to point at the correct parent. Both of these changes were considered unexpected so would cause fsck in preen mode or fsck in background mode to fail with the need to run fsck manually to fix these problems. This update changes these errors to be expected so that in preen mode fsck will simply fix these transitional errors. For now, background fsck will note these errors, but will need additional kernel support to fix them, so will simply ignore them rather than fail. A future update will allow background fsck to fix these problems. Reported by: jeff
2010-01-11 19:52:40 +00:00
return (0);
}
if (!preen && !reply("REMOVE"))
return (0);
memset(&dotdesc, 0, sizeof(struct inodesc));
dotdesc.id_type = DATA;
dotdesc.id_number = inp->i_parent; /* directory in which name appears */
dotdesc.id_parent = inp->i_number; /* inode number in entry to delete */
dotdesc.id_func = deleteentry;
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(dotdesc.id_number, &ip);
if ((ckinode(ip.i_dp, &dotdesc) & FOUND) && preen)
When renaming a directory it passes through several intermediate states. First its new name will be created causing it to have two names (from possibly different parents). Next, if it has different parents, its value of ".." will be changed from pointing to the old parent to pointing to the new parent. Concurrently, its old name will be removed bringing it back into a consistent state. When fsck encounters an extra name for a directory, it offers to remove the "extraneous hard link"; when it finds that the names have been changed but the update to ".." has not happened, it offers to rewrite ".." to point at the correct parent. Both of these changes were considered unexpected so would cause fsck in preen mode or fsck in background mode to fail with the need to run fsck manually to fix these problems. This update changes these errors to be expected so that in preen mode fsck will simply fix these transitional errors. For now, background fsck will note these errors, but will need additional kernel support to fix them, so will simply ignore them rather than fail. A future update will allow background fsck to fix these problems. Reported by: jeff
2010-01-11 19:52:40 +00:00
printf(" (REMOVED)\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);
When renaming a directory it passes through several intermediate states. First its new name will be created causing it to have two names (from possibly different parents). Next, if it has different parents, its value of ".." will be changed from pointing to the old parent to pointing to the new parent. Concurrently, its old name will be removed bringing it back into a consistent state. When fsck encounters an extra name for a directory, it offers to remove the "extraneous hard link"; when it finds that the names have been changed but the update to ".." has not happened, it offers to rewrite ".." to point at the correct parent. Both of these changes were considered unexpected so would cause fsck in preen mode or fsck in background mode to fail with the need to run fsck manually to fix these problems. This update changes these errors to be expected so that in preen mode fsck will simply fix these transitional errors. For now, background fsck will note these errors, but will need additional kernel support to fix them, so will simply ignore them rather than fail. A future update will allow background fsck to fix these problems. Reported by: jeff
2010-01-11 19:52:40 +00:00
inp->i_parent = idesc->id_number; /* reparent to correct directory */
inoinfo(inp->i_number)->ino_linkcnt++; /* name gone, return reference */
return (0);
}
static int
deleteentry(struct inodesc *idesc)
{
struct direct *dirp = idesc->id_dirp;
if (idesc->id_entryno++ < 2 || dirp->d_ino != idesc->id_parent)
return (KEEPON);
dirp->d_ino = 0;
return (ALTERED|STOP|FOUND);
}
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
/*
* Routine to sort disk blocks.
*/
Merge from Lite2. Note that Lite2 has it's own filesystem clean check skipping code that overrides ours sooner. One should be eliminated, but for now it works.
1997-03-11 12:20:21 +00:00
static int
o __P removed o ansi function prototypes o unifdef -D__STDC__ o __dead2 on usage prototype o remove now-bogus main prototype
2002-03-20 22:57:10 +00:00
blksort(const void *arg1, const void *arg2)
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
{
Warning cleanup. Format changes by peter
2002-07-30 13:01:25 +00:00
return ((*(struct inoinfo * const *)arg1)->i_blks[0] -
(*(struct inoinfo * const *)arg2)->i_blks[0]);
BSD 4.4 Lite sbin Sources Note: XNSrouted and routed NOT imported here, they shall be imported with usr.sbin.
1994-05-26 06:35:07 +00:00
}
Reference in New Issue Copy Permalink