RELEASE-NOTES revision ba142c1b3b4137c3577640307579354f7c6accb4
1E2fsprogs 1.33-WIP (April 14, 2003) 2=================================== 3 4Added a new utility program, logsave, which captures the output of a 5command in a log file, even if the containing directory hasn't been 6mounted yet (in which case the it saves the output in memory until it 7can write out the logfile). This is useful for capturing the 8output of fsck during the boot sequence. 9 10Fixed some portability problems that were causing problems under 11the Cygwin32 environment. 12 13Mke2fs now interprets a negative number to the -b option as a minimum 14block size. 15 16Fixed a bug in mke2fs which was incorrectly checking the argument to 17the -g option if the default block size was used. (Addresses Debian 18bug #188319) 19 20Fsck now explicitly ignores tmpfs and devpts, and it will complain if 21it can not find filesystem checkers for jfs, reseirfs, and xfs. 22 23E2fsck now updates the global free block and inode counters from the 24block group specific counters quietly. This is needed for an 25experimental patch which eliminates locking the entire filesystem when 26allocating blocks or inodes; if the filesystem is not unmounted 27cleanly, the global counts may not be accurate. 28 29Imported bug fixes to the EVMS plugin from the EVMS 2.0 tree. (EVMS 302.0 is not yet supported; this only pulled in the bug fixes: fixed 31possible hangs caused by bugs in calling waitpid, and not setting the 32pipe to non-blocking mode; also fixed a file descriptor leak; made 33sure all functions call log entry/exit functions.) 34 35Badblocks now flushes its output file as bad blocks are discovered. 36 37The uuid library is now more paranoid about assuming the correctness 38of the /dev/random device; it mixes in a stream of bytes from 39random/srandom, just in case. 40 41Update Debian files to reflect the fact that I am now the Debian 42maintainer of e2fsprogs. Other various Debian-specific packaging 43cleanups. 44 45Move the source tarball generation functions from the top-level 46makefile to the util/gen-tarball script. 47 48Updated the Turkish .po translation file. 49 50Added heimdall and MIT krb5 extensions to the com_err library to make 51it more compatible with com_err libraries from those distributions. 52 53Changed dumpe2fs to always display the superblock fields relating to 54the journalling and/or directory indexing feature even if those 55features are not enabled. 56 57Updated and clarify copyright statement vis-a-vis alpha releases of 58e2fsprogs. 59 60The ss library will now try to dynamically link to the readline 61library and use it if it is present in the system. This means that 62the debugfs program now has line editing and history features. The 63SS_READLINE_PATH environment variable is used to find a readline or 64readline-compatible library. 65 66E2fsck now finds most duplicate filenames (all when rebuilding all 67directories via the -D option) and offers to delete or rename 68duplicate filenames/directory entries. (Addresses Debian Bug #176814). 69 70Fix bug in e2image. When writing out a raw image file, include data 71blocks from symlinks that do not store the symlink within the inode. 72 73Fix bug in resize2fs which incorrectly moved the block and inode 74bitmaps for sparse superblock filesystems and incorrectly marked 75blocks as in use. (Addresses Debian bug #174766) 76 77Added a new shared library, the blkid library, which efficiently 78allows fsck, mke2fs, e2fsck, and tune2fs to be able to look up LABEL 79and UUID filesystem specifiers without needing to search all of the 80devices in the system. Instead, the device is looked up in a cache 81file, and then verified to make sure the blkid cache is correct. 82 83Tune2fs and e2label will accept LABEL=xxx and UID=yyy specifiers for 84the device name, using the blkid library. (Addresses Debian bugs 85#166048, #179671) 86 87Fsck now supports backslash escapes in /etc/fstab so that \040 can be 88used for spaces in device labels. 89 90Removed 32-bit limitations for debugfs's dump command. 91 92If the user specifies a large number of inodes, Mke2fs will 93automatically adjust the number of blocks per group to find a valid 94set of filesystem parameters. 95 96Add workaround to detect broken MD devices where when some of the 97underlying devices are marked read-only, writes to the MD device are 98silently dropped. E2fsck will detect if there is an attempt to run 99the journal twice, and abort with an error if this is the case. 100(Addresses IBM Bugzilla bug #1226) 101 102E2fsck will print an error if more than one of the -p/-a, -n or -y 103options are specified. 104 105E2fsck will fix HTREE corruptions in preen mode, without stopping the 106boot process. This is needed because the 2.4 ext2 filesystem 107accidentally had the INDEX_FL backwards compatibility code removed. 108 109Mke2fs no longer creates filesystems with the dir_index flag set by 110default; the user has to specifically request it. 111 112Update and clarified various man pages. (Addresses Debian bugs 113#173612, #175233, #175113, and #170497, #185945, #188318) 114 115Created man page for the mk_cmds program (from the libss library). 116 117Programmer's notes: 118------------------- 119 120Fix various gcc -Wall nits. 121 122Fixed a lot of portability problems that caused e2fsprogs not to build 123successfully under Solaris and Apple/Darwin. 124 125Fixed a Makefile dependency to allow building e2fsprogs using parallel 126make jobs. 127 128Changes to create a subset distribution which consists only of the 129et, ss, uuid, and blkid libraries. The configure script and top-level 130makefile were changed to support working with a subset distribution. 131 132Removed EXT2_FEATURE_RO_COMPAT_BTREE_DIR mention of since it's not 133actually used, and might people who are looking for 134EXT2_FEATURE_COMPAT_DIR_INDEX, which is in use. 135 136Updated debian files to fix a number of Lintian warnings. 137 138Updated config.guess and config.sub with newer versions from the FSF. 139 140E2fsprogs 1.32 (Nomvember 9, 2002) 141================================== 142 143Fixed a bug in the Unix I/O routines which caused needless writebacks 144of clean blocks from the unix_io cache (they were erroneously marked 145as being dirty, so they were getting written back to disk before 146getting evicted from the disk cache). This was harmless, but it 147significantly slowed down e2fsck. 148 149Made some other minor optimizations to the Unix I/O routines to save a 150small amount of CPU time. 151 152Updated internationalization files. 153 154 155E2fsprogs 1.31 (Nomvember 8, 2002) 156================================== 157 158Update EVMS ext2fsim plugin with EVMS 1.2. (We still support 159compiling the fsim plugin with EVMS 1.0 and 1.1.) Add better error 160handling for child process that die unexpectly. Add a hack to force 161mkfs to create filesystems that won't cause problems with hardware 162that has 2k or 4k minimum blocksize requirements. Read from child 163processes in non-blocking mode, so that the GUI continues to update. 164 165Fix e2fsck so that it returns the appropriate exit code when the root 166filesystem has been changed, so that system's rc scripts will be told that 167the system needs to be rebooted. 168 169Fix a bug in ext2fs_flush/ext2fs_close; when the MASTER_SB_ONLY flag 170was set, some of the descriptor blocks that should have been written 171out were getting skipped. 172 173Changed e2fsck to force out changes to the backup copies of the 174superblock and block group descriptors when important changes are made 175to those data structures. 176 177Fix a bug where e2fsck could erroneously mark a filesystem as being 178clean if a check of dirty filesystem is interrupted with a ^C. (Bug 179introduced in e2fsprogs 1.28.) 180 181If journal debuging is enabled using --enable-jbd-debug, the debugging 182level is now set via the E2FSCK_JBD_DEBUG environment variable. 183 184If byteswapping support is disabled using configure --disable-swapfs, 185skip the tests which depend on byte-swapping, so that "make check" 186won't bomb out. 187 188Lshattr will now display the indexed directory flag. Also, some of 189the more esoteric compression flags are supressed unless compression 190support has been enabled. 191 192Update man pages. 193 194 195E2fsprogs 1.30 (October 31, 2002) 196================================= 197 198When resizing a filesystem, and inodes need to moved, resize2fs will 199update the inode's ctime field, and the mtime field of the containing 200directories, so that incremental backups using dump/restore will work 201correctly. 202 203In order to avoid spurious errors, e2fsck wil check the last 204superblock write time to determine whether or not it can safely use 205the LOW_DTIME checks to determine if there are inodes on the orphan 206inode linked list that had somehow gotten disconnected. (Addresses 207Sourceforge bug #620980) 208 209Partition in /proc/partitions that being with the string "lvm" are 210considered real partitions even if they do not end with a number. 211 212Fixed a bug in the the uuid generation function, where if /dev/urandom 213is not present, but /dev/random is, and there isn't sufficient 214entropy, the get_random_byte function could spin a loop forever. 215 216E2fsck, mke2fs, etc. will now reliably notice when image files are 217mounted using the loopback interface. (Addresses Sourceforge bug 218#619119) 219 220When flushing buffers (as is done by badblocks, for example) check to 221see if the BLKFLSBUF ioctl succeeds, and if so, avoid using the 222FDFLUSH ioctl which causes the MD device driver which causes confusing 223syslog messages about obselete ioctl messages. (Addresses Sourceforge 224bug #545832). 225 226Debugfs's write command now checks to see if the destination filename 227exists before creating it. (Addresses Sourceforge bug #478195.) 228 229When installing man pages, search for compressed man pages using all 230commonly used compression extensions (i.e., .Z, .gz, .bz2, etc.) 231 232Fixed a bug in fsck where multiple filesystems specified on the 233command were not being checked in parallel due to a logic bug 234introduced to support the FSCK_MAX_INST environment variable. 235 236We have added a new superblock field, s_mkfs_time, which indicates 237when a filesystem was created. It is set via mke2fs, and printed out 238by dumpe2fs, but is not actually touched by the kernel. 239 240Dumpe2fs has been made more robust by not aborting if there is an 241error reading the block/inode bitmaps; instead it will still print out 242the location of the block/inode bitmaps and inode table. 243 244Add support for the an alternative block group descriptor layout which 245allows for on-line resizing without needing to prepare the filesystem 246in advance. (This is the incomat feature flag meta_bg.) 247 248Add support for storing default mount options in the superblock, so 249that the filesystem can be mounted with specific mount options without 250needing to specify them on the mount command line or in the /etc/fstab 251file. 252 253Add support for a new inode flag, which is to be used for indicating 254the top of directory hierarchies for the Orlov block allocator. 255 256Fix e2fsck so that if it creates the lost+found directory, it does so 257with the more apporpriate permissions of 0700. Also change 258mklost+found so that it also creates the directory 0700. 259 260Fixed format bug in e2fsck if NLS is in use. 261 262Add a German translation for e2fsprogs's NLS support. 263 264Fixed e2fsck so that it more handles BAD_BLOCK_IN_INODE_TABLE even at 265the beginning of the inode table. This won't matter much, since if 266there is a bad block at the beginning of the inode table, the root 267directory won't be available. But at least e2fsck won't crash in this 268case. 269 270Fixed endian problems in the htree code for e2fsck and debugfs. 271 272When byte-swapping a filesystem on a PPC architecture, byte-swap the 273bitmaps since the historical big-endian ext2 variant had byte-swapped 274bitmaps, and the ext2fs library assumes this. This fixes the 275regression test suite on PPC machines. 276 277Fix e2image so that it handles a bad block in the inode table 278appropriately. 279 280E2fsck now uses a more sophisticated algorithm to salvage corrupted 281directories that recovers more information in the corrupted directory 282block. 283 284E2fsck now performs additional consistency checks on indexed (HTREE) 285directories. 286 287Fix bug where efsck might get confused about whether a completely 288empty directory block is an empty leaf block or an interior htree 289node. This could cause e2fsck to get confused into think that a valid 290indexed directory was corrupted. 291 292E2fsck no longer creates an empty directory entry at the end of a 293directory block when optimizing a directory. This would cause some 294earlier versions of the dxdir kernel code to corrupt the directory 295when attempting to split a node containing an empty directory entry. 296 297E2fsck could sometimes in rare circumstances leave the indexed flag 298set after a small directory was optimized by compressing it instead of 299indexing it. (This should never happen in real life, since 300directories that small wouldn't have been indexed, but better safe 301than sorry.) 302 303E2fsck now only updates the master superblock in all cases. This 304slightly shortens its run time. 305 306Ext2ed can deal with directory entries of length 0; previously it 307would get stuck in an infinite loop. 308 309Fsck now has support for reiserfs volumes when parsing LABEL= and UUID= 310specifiers. (Sourceforge patch #613447) 311 312Badblocks will now work correctly on read-only devices such as 313CD-ROM's. (Sourceforge patch #600451) 314 315Updated and clarified man pages. (Addresses Debian bug #167108) 316 317 318E2fsprogs 1.29 (September 24, 2002) 319=================================== 320 321Fixed a bug in e2fsck which could corrupt a directory when optimizing 322it (via the -D option) or rebuiliding the hash tree index with a 1 in 323512 probability, due to a fence post error. 324 325Fixed a bug in the LVM support code which caused LABEL='xxx' not to 326work correctly. 327 328Mke2fs now enables the directory indexing flag by default. (Since 329this is a compatible feature flag, it's safe to do so.) 330 331Tune2fs will support setting the directory indexing feature flag. It 332will automatically set up the default hash algorithm and hash seed 333fields in the superblock. 334 335If the bone-headed user enters the root filesystem twice in 336/etc/fstab, the -R option which skips the root filesystem will skip 337all of them. (Addresses Debian bug #159423). Note! This is not a 338precedent for dealing intelligently with any other kind of doubled 339entry in /etc/fstab! 340 341 342Programmer's notes: 343------------------- 344 345Removed perror declaration in lib/et/internal.h. All modern systems 346can be expected to define perror() these days. Besides, the lib/et 347code wasn't using perror at all anyway. :-) 348 349 350E2fsprogs 1.28 (August 31, 2002) 351================================ 352 353Add support for the Hashed-Tree Directory Indexing to e2fsck. Support 354for setting the htree flag is not included yet, although it can be 355manually turned on via the debugfs program. 356 357Clarified e2fsck error message which is printed when it cannot find 358sufficient contiguous block when relcating filesystem metadata. 359 360Added support for building an EVMS plugin module for ext2/3. This 361module is substantially the same as the EVMS module shipping with EVMS 3621.1, with one or two bugfixes. E2fsprogs can also build this plugin 363for use with EVMS 1.0 (which did not include the ext2 plugin module), 364if the configure --enable-old-evms flag is given. 365 366Fsck will search through EVMS volumes when trying to resolve 367filesystem specifications such as LABEL=xxx or UUID=xxx. 368 369Added a new utility program, /sbin/findfs, which will return 370filesystem specifications such as LABEL=xxx or UUID=xxx, and prints 371the device name. 372 373Update and clarified various man pages. (addresses Debian Bug #145044, 374#146437, #131350, #151990, #144621, #141938) 375 376If there are no filesystems specified on fsck's command line, fsck now 377treat that as if the -As options were given. Previously it would 378simply do nothing. (Addresses Debian Bug #153102) 379 380Mke2fs no longer treats a failure to be able to clear the MD signature 381at the end of the filesystem as a fatal error. (Addresses Debian Bug 382#155007) 383 384The e2p library functions (which are used by lsattr and chattr) now 385double check to make sure the file is a regular file or directory 386before attempting to use the ext2 ioctls. Some device drivers 387unfortunately respond to the ext2 ioctl's with unknown behaviour. 388(Addresses Debian Bug #152029). 389 390The extended attribute handling has been updated to correspond with 391the latest V2 bestbits ACL code. 392 393Fixed bug in e2fsck which caused it to not clear the dtime field when 394processing truncated inodes on the orphan list. This could cause data 395loss(!) if a filesystem is rebooted before a truncate has been 396committed. 397 398E2fsck now uses red/black trees in pass1b, which removes some O(n**2) 399algorithms. This makes e2fsck much faster in the case of severely 400corrupted filesystems where a large number of blocks are claimed by a 401large number of inodes. (Thanks to the 2.5 IDE device driver for 402inspiring this work.) 403 404Resize2fs has been significantly sped up when shrinking and expanding 405a filesystem by a very small number of blocks (for example, when EVMS 406is converting a partition to be an EVMS legacy volume). 407 408Added a new option to e2fsck, -D, which will optimize or compress all 409of the directories in the filesystem. 410 411E2fsck now catches SIGINT and SIGTERM to make sure it will can 412properly clean up and only exit at safe points. Fsck will pass 413SIGINT/SIGTERM to its child processes, and wait until they have all 414exited before it exits. 415 416The uuid parsing code in the uuid library now properly complains when 417an illegally formated uuid is presented to it. (Addresses Debian bug 418#152891) 419 420Restrict use of the 2.4 setrlimit ABI f*ckup to kernels between 2.4.10 421and 2.4.17, since the workaround can cause problems when using a 2.4 422kernel with an old version of glibc built with the 2.2 headers. 423 424Fixed a bug in mke2fs where it wasn't properly clearing the initial 425superblock used by other filesystems. (Addresses Debian bug #147256.) 426 427Added support for the synchronous directory feature written by Andrew 428Morton. 429 430The debugfs program can delete directories using the rmdir command. 431 432Add support for 8k and 16k filesystems (for systems with page sizes 433that are greater or equal to 8k or 16k, respectively). Note that 434these filesystems can not be mounted on x86 systems, or other systems 435with only 4k page sizes, due to limitations in the current Linux VM 436code. 437 438Resize2fs requires that the filesystem state be valid and have no 439errors; otherwise, e2fsck -f must be run first. (Previously it simply 440required that the last fsck time be greater than the last mount time.) 441 442Configure now defaults the man pages directory to /usr/share/man on 443Linux systems. 444 445E2fsck now offers to truncate inodes which contain too many blocks (so 446that i_blocks would overflow. Also fixed handling of large sparse 447files. 448 449E2fsck now more completely checks for symlink validity, including 450requiring NULL termination and length checks. 451 452E2fsck will offer to try forcing a disk write to remap a bad block 453after finding a read error when reading a filesystem metadata block. 454 455Fixed a bug in debugfs which caused the -b and -s options to crash 456debugfs, as well as breaking the testb, setb, and clearb functions. 457 458Added a bmap command to debugfs which calculates the logical to 459physical block mapping for a particular inode. 460 461Fixed a bug in code which checked to see if a device was mounted which 462sometimes (rarely) failed in the case of a plain file. 463 464Fixed a bug in resize2fs where when it reported an error, it would 465print a message erroneously indicating that the filesystem had been 466resized before it aborted. 467 468When resizing a plain file which is smaller than the requested size, 469resize2fs will attempt to extended the file so that programs like 470e2fsck will not complain that the file is too small. 471 472Resize2fs will print the actual new size of the filesystem when it is 473finished resizing. 474 475Fixed a bug in debugfs where "ls -l" would report incorrectl file type 476information on big-endian systems. 477 478 479Programmer's notes: 480------------------- 481 482Fixed collisions with C++ reserved words. 483 484Added portability fixes for building e2fsprogs on the HURD and AIX. 485 486Added the ext2ed program for creation of test cases. (ext2ed has many 487limitations and bugs which make it unsuitable for production use.) 488 489The ext2fs_read_dir_block2 and ext2fs_write_dir_block now take a new 490flag, EXT2_DIRBLOCK_V2_STRUCT, which will will reverse when the 491name_len field is byte swampped on big-endian machines, since in the 492V2 structure, name_len is a char field which is doesn't need to be 493byte swapped --- except if an old-style kernel had byte-swapped the 494name_len field as part of the V1 structure. 495 496 497E2fsprogs 1.27 (March 8, 2002) 498============================== 499 500The warning messages for mke2fs now go to standard error. 501 502Fixed to make sure "make check" runs all of the test programs with 503LD_LIBRARY_PATH set, so that we test the libraries in the build tree. 504 505The mke2fs program checks the boot sector for the BSD disk label, and 506avoids erasing it if it is there. 507 508Fixed a bug in e2fsck which caused it to core dump if the journal 509inode was missing when it was supposed to be there. 510 511E2fsck now prints ranges in pass 5 when printing deltas for the block 512and inode bitmaps. 513 514Debugfs's "ls -l" command now will print out the file type information 515in the directory entry. 516 517Create man pages and hard links for fsck.ext3 and mkfs.ext3. If 518mke2fs is invoked as mkfs.ext3, create the filesystem with a journal. 519 520Debugfs can now examine the experimental directory indexing 521information. 522 523Fixed bug in debugfs which caused it to core dump if modify_inode is 524called without an open filesystem. 525 526The debugfs lsdel command now runs its output through a pager first. 527 528When installing manual pages, remove the compressed manual pages first. 529 530Synchronized with Debian's packaging information for e2fsprogs-1.26-1. 531 532Fix the 2.4 resource limitation workaround introduced in 1.26 which 533actually broke things on mips32, sparc32, and Alpha platforms. 534 535Updated the I18N code so that calls to setlocate(LC_CTYPE, 0) are made 536(which is required by the newer libintl libraries). 537 538Programmer's notes: 539------------------- 540 541Fixed various gcc -Wall complaints. 542 543Fixed a few memory leaks in the e2fsck journaling code, and in the 544ismounted code checking for a swap device. 545 546Add new inode I/O abstraction interface which exports an inode as 547an I/O object. 548 549Exported ext2_file_flush as a public interface. 550 551In ext2_file_write, we now mark the buffer void * argument as a const, since ext2_file_write doesn't modify the buffer. 552 553Lots of small random portability fixes to make e2fsprogs build under 554AIX --- even without the 5L compatibility toolkit, and even using the 555uber-crippled AIX native C compiler. 556 557 558E2fsprogs 1.26 (February 3, 2002) 559================================= 560 561Dumpe2fs will keep going now if the bad block inode can't be read. 562(Previously it stopped with a fatal error.) 563 564E2fsck will now give an opportunity to run the journal even if the 565recovery flag is not set. This is the default behaviour if e2fsck is 566using a backup superblock, since the needs_recovery flag will never be 567set in the backup superblock copies. 568 569E2fsck now automatically finds the backup superblock/group descriptors 570even when the primary superblock is completely destroyed for 2k and 4k 571filesystems. (Previously it just guessed/assumed that we were dealing 572with a 1k filesystem in that case, and users had to manually specify 573the backup superblock number.) 574 575Fixed a bug in e2fsck where it dereferences a null pointer when there 576is a problem opening a filesystem in preen mode. 577 578E2fsck now handles external journals correctly; previously it trashed 579the external journal device if the journal needed to be replayed. 580 581Work around ulimit incompatibility problem caused by recent 2.4 582kernels; the unix IO module will automatically try to set any resource 583limits to be infinite on startup. 584 585Fixed a bug in e2fsck where it wasn't allocating a new block or inode 586bitmap if it wasn't present and the blocksize was 2k or 4k. 587(Addresses Debian Bug #116975) 588 589E2fsck will check and fix botched ext3 V1 to V2 superblock updates by 590clearing the new V2 fields if they do not make sense or if the ext3 591superblock is version 1 superblock. 592 593E2fsck will automatically relocate the ext3 journal from a visible 594file (i.e., /.journal) to an hidden inode if the filesystem has been 595opened read/write. This allows the users to add a journal while the 596filesystem is mounted, but the next time the system is rebooted, the 597journal file will disappear. This avoids problems with backups, 598stupid operators with superuser bits, etc. 599 600Fix a bug in fsck where it would not support a filesystem type of 601"auto" if the device was specified in terms of LABEL=xxx or UUID=xxx. 602 603Fsck now supports fstab entries of the form "ext3,ext2". It will also 604automatically identify reiserfs filesystems. 605 606The number of processes spawned by fsck can now be limited using the 607FSCK_MAX_INST environment variable. 608 609Fsck now searchs the LVM proc hierarchy to find logical volumes which 610should be searched for UUID and label information. 611 612Work around a bug in 2.4.10+ kernels by trying to unset the filesize 613limit if at all possible, when opening a block device. (The filesize 614limit shouldn't be applied against writes to a block device.) 615 616In mke2fs and e2fsck, specifying the -c option twice will now do 617a read/write test on the disk. Update the man pages to encourage 618using the -c option, and to discouraging running badblocks separately, 619since users tend to forget to set the blocksize when running 620badblocks. 621 622Mke2fs now automatically clears the immutable attribute on a 623pre-existing .journal file left over from a previous failed attempt to 624add a journal to an alreadyy-mounted filesystem. 625 626Fixed mke2fs's exit codes to consistently indicate when the mke2fs 627operation failed. 628 629Mke2fs now creates the lost+found directory with permissions of 0700, 630so that files that had previously lived in protected directory are 631safe if they get dropped in lost+found and the system administrator 632doesn't deal with immediately. (Addresses Debian #bug 118443) 633 634Mke2fs and e2fsck (and all programs that use the 635ext2fs_check_if_mounted function) will now properly identify that a 636device is mounted, even in cases where devfs has confused things with 637multiple devices names with the same identity, or if a dim-witted 638system administrator has done something stupid like creating device 639file aliases in their home directory. Also now checks for swap devices 640by using /proc/swaps. (Addresses Debian bug #106622) 641 642Added a new option (-T) to tune2fs which sets the time a filesystem 643was last checked. 644 645Speed up e2image when creating sparse raw image files by optimizing 646away excess lseek() system calls. 647 648Fix support of large (> 2GB) files when used as a filesystem in 649mke2fs, tune2fs, debugfs, and findsuper. 650 651Debugfs's stat and icheck commands now properly deals with large (> 6522GB) files. 653 654Debugfs's set_super_value command now prints out the list of valid 655superblock fields which can be set using the command. 656 657Debugfs's rm and kill_file command now updates the superblock free block 658and inode counts, thus keeping the filesystem consistent. 659 660Debugfs's lsdel command now takes an optional argument which allows 661the user to only see the most recently deleted files. 662 663A new command (undel) was added to debugfs which automates 664undeleting a deleted inode and linking it back to a directory. 665 666Debugfs's ls command now takes a new option, -d, which lists 667deleted directory entries. 668 669Debugfs's testb, freeb, setb, and find_free_block commands now take an 670optional count argument. 671 672Add support for a new ext2 file attribute, EXT2_NOTAIL_FL, which will 673be used to signal that a particular inode should not be eligible for 674tail-merging --- this is needed for programs like LILO. 675 676The findsuper (an unreleased, uninstalled utility program) has been 677improved to add extra validity checks and to add a progress meter. 678(It is still an unsupported program; the officially supported way to 679recover from a trashed partition table is to use gpart; findsuper is 680for wizards only. :-) 681 682Debugfs was fixed to compile with "configure --disable-swapfs". 683 684Cleaned up various manual pages. (Addresses Debian bug #119624, #120171) 685 686Added new translation file for Turkish. 687 688Programmer's notes: 689------------------- 690 691Fix general gcc -Wall complaints. 692 693E2fsprogs (mostly) works with the dietlibc. 694 695The programming texinfo file has been expanded to include description 696of additional libext2fs functions. (Still not compeltely done, but a 697lot of the more important functions have been documented.) 698 699Added a umask structure to struct_ext2_filsys, which currently only 700modifies the behaviour of ext2fs_mkdir(), but if we add any file 701creation functions to libext2fs, we should also make sure they respect 702the umask setting. 703 704The build-rpm script was fixed to be a bit more safe. 705 706The tests' Makefile now has a way of automating test case creation 707for e2fsck, using "make testnew". 708 709Created a new function, ext2fs_dir_iterate2 which passes more 710information to the callback function (identical to the one used by 711ext2fs_dblist_dir_iterate). The directory iterator functions take a 712new flag, DIRENT_FLAG_INCLUDE_REMOVED, which will return deleted 713directory entries. If the directory entry is deleted, the callback 714function will be called with the entry paraemter set to 715DIRENT_DELETED_FILE. 716 717Added new functions, ext2fs_inode_alloc_stats and 718ext2fs_block_alloc_stats, which takes updates block/inode allocation 719statistics in the bitmaps, block group descriptors, and superblock 720when inodes or blocks are allocated or deallocated. 721 722 723E2fsprogs 1.25 (September 20, 2001) 724=================================== 725 726This is primarily a bug-fix release; no new features were added, but 727there are a number of embarassing bug fixes and cleanups applied. 728 729Fix a bug mke2fs which causes stack garbage to be written to disk when 730zapping disk sectors. (This bug could cause mke2fs to core dump for 731some kernels, I suspect with security enhancement patches.) 732 733Remove unneeded #include of <linux/config.h> which was breaking 734building e2fsprogs on the Hurd. (Addresses Debian bug #112414.) 735 736Updated tune2fs man page to reflect the fact that adding or removing a 737journal doesn't require running e2fsck. 738 739Remove use of AC_REQUIRE from autoconf which had been used to prevent 740AC_CANONICAL_HOST from being called twice; unfortunately this causes 741recent autoconf to bomb out since they don't allow AC_REQUIRE to be 742used outside of autoconf macros. Fortunately, it doesn't seem to be 743necessary any more, anyway. 744 745E2fsck will now not fall back to an alternate superblock if the user 746specifies the superblock location explicitly on the command-line. 747This allows easier recovery from a situation where the primary 748superblock and block groups are slightly corrupted, but the backup 749superblocks are completely broken. 750 751Fix problem which caused compile_et and mk_cmds to blow up if 752e2fsprogs was in a directory with a pathname that contained spaces. 753 754We are now more paranoid about checking the entry in /etc/mtab to make 755sure the filesystem is really mounted, since some broken distributions 756(read: Slackware) don't initialize /etc/mtab before checking non-root 757filesystems. (Slackware also doesn't check the root filesystem 758separately, and reboot if the root filesystem had changes applied to 759it, which is dangerous and broken, but there's nothing I can do about 760that.) 761 762Make UUID library C++ friendly by adding appropriate extern "C" 763declarations and using const in the function declarations. 764 765Fix up the com_err texinfo file so that it can product a valid info 766file (previously, it could only be used to generate paper 767documentation using texinfo.tex). 768 769 770E2fsprogs 1.24a (September 2, 2001) 771=================================== 772 773Fix brown-paper bug in mke2fs which caused it to segfault when 774printing the version string. 775 776 777E2fsprogs 1.24 (August 30, 2001) 778================================ 779 780Revert the BLKGETSIZE64 support, since for some distributions, this 781ioctl has been used by an unofficial kernel patch to update the last 782sector on the disk, and this was causing disk corruption problems as a 783result. 784 785Mke2fs is now more careful about zapping swap space signatures and 786other filesystem/raid superblock magic values so that programs like 787mount who try to use hueristics to determine the filesystem type of a 788partition are less likely to be confused. 789 790E2fsck will no longer complain if the the mode of EXT2_RESIZE_INO is a 791regular file (since Andreas Dilger's on-line resizing tools will set 792its mode bits). 793 794Fixed some minor texinfo, man pages nits for spelling errors, 795texinfo warnings, etc. (Addresses Debian bug #110621.) 796 797E2fsprogs program no longer print the filesystem version number 798(i.e. 0.5b), since it only confuses people and doesn't serve any real 799purpose. 800 801E2fsck will once again compile under libc5, since it will supply its 802own version of strnlen if necesssary. 803 804mke2fs and tune2fs will allow the use of UUID= or LABEL= specifiers 805when specifying the external journal device. tune2fs will also search 806devices looking for the externla journal debice when removing. 807 808E2fsprogs 1.23 (August 15, 2001) 809================================ 810 811Add initial support for extended attributes (EA); e2fsck will 812correctly handle a filesystem with EA's and check the EA blocks for 813corruptions. 814 815E2fsck's symlink sanity checking has been cleaned up. It now checks 816the i_size value of fast symlinks, and checks for immutable flags 817being set of symlinks, etc. 818 819E2fsck now offers to clear inodes which are special files that have a 820non-zero i_blocks or i_size field. (The i_size field check was in the 821previous version of the code, but due to a bug it didn't offer to 822clear the inode unless i_size and i_size_high were both non-zero.) 823 824E2image can now create "raw" image files, which only contain the 825filesystem metadata placed in a spare file so that e2fsck, dumpe2fs, 826debugfs, etc., can be run directly on the raw image file. 827 828Add support for the 64-bit block device patches. 829 830Fixed bugs in creating external journals with a 1k blocksize. 831 832Add initial support for external journals (so long as the external 833journal only supports a single filesystem) in e2fsck. 834 835Remove requirement for needing to run fsck on a filesystem after 836removing a journal (either internal or external). 837 838The man pages now document how to create and manage external journals. 839 840Speed up the check of an ext3 filesystems by avoiding a needless flush 841of all of the superblock and block group descriptors. 842 843Speed up creating an internal journal using tune2fs in the case where 844the filesystem has a lot blocks already allocated. 845 846Tune2fs has been fixed to make sure that only error messages go to 847stderr, and normal message go to stdout. (Addresses Debian bug #108555) 848 849Fixed a minor bug in mke2fs; if -O none is passed to mke2fs, it will 850now not set the sparse_super feature. (Addresses Debian bug #108165) 851 852Add support in fsck for the filesystem type "auto". 853 854Fsck -A will not try to interpret device names for filesystems which 855have a pass number is 0. (Addresses Debian bug #106696). 856 857Fsck prints a warning message if now valid filesystems are passed to 858it. (Addresses Debian Bug #107458.) 859 860E2fsck now gives an explicit warning if there filesystem still has 861errors at the end of the run. (Addresses Debian bug #104502) 862 863E2fsck will set the EXT2_ERROR_FS flag if the journal superblock 864reflects an error. E2fsck will also not run the orphan list if the 865filesystem contains errors, since the orphan list might be corrupted. 866 867E2fsck now prints the number of large files when given the -v option. 868 869Fixed minor memory leaks in e2fsck. 870 871Some minor man pages updates. (Addresses Debian bug #30833, #108174) 872 873 874E2fsprogs 1.22 (June 22, 2001) 875============================== 876 877Fixed a bug in e2fsck's handling of orphan inodes which are special 878files (i.e., block/character device files, named FIFO's, etc.). 879 880Fixed a bug accidentally introduced in e2fsprogs 1.21 which caused 881tune2fs to fail at adding a journal to a mounted filesystem. 882 883Fixed a few big-endian bugs in e2fsprogs 884 * The directory block functions were accidentally reporting 885 some directories as corrupted when they weren't. 886 * If e2fsprogs is compiled --disable-swapfs, the C language 887 equivalents weren't being included for big-endian platforms. 888 (Fixes Debian bug #101686). 889 890Fixed a Hurd compilation problem. (Addresses Debian bug #101361) 891 892Programmer's notes: 893------------------- 894 895Use platform independent method of defining the BLKFLSBUF and FDFLUSH 896ioctls, and allow them to be defined for non-i386 platforms. 897 898The uuid.h header file is now protected against multiple inclusions. 899 900E2fsprogs is now being developed using BitKeeper. Changed the test 901scripts to deal with BK's stripping CR characgters from text files, 902and changed the top-level Makefile.in to avoid including BitKeeper 903files when generating the source tarball. 904 905 906E2fsprogs 1.21 (June 15, 2001) 907============================== 908 909Added new configure flags which allow a subset e2fsprogs to be built; 910this is most useful for boot floopies, since the resulting shared 911libraries and programs are slimmed down by removing features that 912aren't necessary for a boot floppy. The new flags that were added 913are: --disable-swapfs, which removes support for byte swapping old 914legacy PPC/68k filesystems, --disable-debugfs, which removes support 915for debugfs from the libext2fs library, --disable-imager, which 916removes support for the e2image program, and --disable-resizer, which 917removes support for resize2fs. 918 919E2fsck now prints the number of mounts or days elapsed since the last 920check when e2fsck is forced to check an otherwise clean filesystem. 921 922Tune2fs now prints an informative message about how often a filesystem 923will be checked when adding a journal to the filesystem, to remind the 924user that he/she may want to adjust those parameters using tune2fs 925-c/-i. 926 927Worked around hurd brain-damage which causes e2fsck to sometimes 928believe a filesystem is the root filesystem based on device numbers 929(since Hurd doesn't have dev_t's, which is arguably a POSIX.1 930violation). 931 932Fixed a bug introduced in 1.20 which caused e2fsck to abort with an 933erroneous error with the -F option was specified. 934 935Fixed a ext3 recovery bug in the revoke handling; synchronized with 936ext3 0.7a. 937 938Fixed two bugs in e2fsck's handling of dup block handling, dealing 939with relatively uncommon edge cases: a directory with an indirect 940block which is claimed by another file, and when the last inode in the 941filesystem has blocks claimed by another file. 942 943E2fsck now checks to see if the i_size field of a fast symlink is too 944big, and offers to clear the symlink if so. 945 946E2fsck now checks to see if i_size_high of special files is non-zero, 947and offers to clear i_size_high. 948 949Fix e2fsck's handling of incompatible journal flags so that the user 950has chance to abort, and then has the option to clear out the journal 951entirely. (Addresses Debian bug #98527.) 952 953Fixed a bug in fsck which could cause it to core dump if a mix of 954standard and non-standard device names are used in /etc/fstab. 955(Debian bug #100559) 956 957Fixed a bug in debugfs which caused read errors when copying a file to 958not be noticed. 959 960The debugfs set_super_value command can now modify the s_lastcheck field. 961 962Fixed a bug in lsattr and chattr which was accidentally introduced in 9631.20 to support > 2GB files; both lsattr and chattr wasn't reading 964directories correctly because the change modified the layout of struct 965dirent to be incompatible with the libe2p shared library. 966 967Cleaned up the mke2fs manual page and included a discussion about why 968it's good to periodically check the filesystem even when journaling is 969enabled. 970 971Programmer's notes: 972------------------- 973 974Fix general gcc -Wall complaints. 975 976The types needed by the ext2 header files are now provided by 977lib/ext2fs/ext2_types.h, instead of include/asm/types.h. 978 979Integers are now preferred to longs when trying to find a 32-bit type 980in ext2_types.h. Also, if linux/types.h has already been defined, 981don't try to redefine the types. 982 983Fixed make depend script so that it automatically corrects the 984pathname cleanups performed by make -M, so I don't have to fix them up 985by hand. 986 987Fixed the d_loaddump test case to be more robust, and not depend on 988bash'isms. 989 990Removed debugfs's dependence on pread(), which was accidentally 991intrudced in e2fsprogs 1.20 992 993Fixed a performance bug in the libext2fs's icount routine; the size 994estimate of the icount array was incorrectly being calculated. 995 996Removed use of the badblocks compatibility functions in the e2fsprogs 997programs. 998 999Added paranoia code which protects against strange cases where /etc 1000isn't on the root filesystem, or if /etc/mtab doesn't exist. 1001 1002The header file ext2_types.h is now installed. 1003 1004Autoconf is used to determine when we are on big-endian machines, 1005instead of doing run-time tests, to save a few bytes of code. 1006 1007The ext2fs_mark_generic_bitmap and ext2fs_unmark_generic_bitmap 1008functions are no longer inline functions, which saves space and 1009doesn't really cost any real performance. 1010 1011The ext2fs library no longer depends on the e2p library. (What need 1012there was of it --- namely, fsetflags, was coded in-line). 1013 1014Fixed the makefile so that lib/ext2fs/ext2_types.h is generated even 1015when the user is stupid and tries compiling the package using "make 1016install" as root. 1017 1018Miscellaneous code cleanups: 1019 * Added missing files from Makefile.in's SRCS file, so that 1020 their dependencies would be properly calculated. 1021 * Removed redundant code 1022 * Fixed comments in code 1023 * Removed no-longer unneeded argsused #pragma. 1024 1025 1026E2fsprogs 1.20 (May 20, 2001) 1027============================= 1028 1029Add support for replaying the ext3 journal purely in user mode, 1030including handling the orphaned inode list. Used code contributed by 1031Andreas Dilger, with bug fixes and the orphaned inode handling done 1032by Theodore Ts'o. 1033 1034The mke2fs and tune2fs programs can create or modify a filesystem to 1035include an ext3 journal. Tune2fs also can be used to remove an ext3 1036journal from a filesystem. 1037 1038E2fsck will now check for the existence of a linked list of orphan 1039inodes in the superblock, and clear those inodes before starting the 1040rest of the filesystem check (but after the journal playback). 1041 1042E2fsck now validates the file descriptor passed to the -C option, 1043which saves against the completion bar getting written to an 1044unexpected location, such as the disk being checked. (Debian 1045bug/wishlist #55220) 1046 1047E2fsck will now bump the filesystem revision number from zero to one 1048if any of the compatibility bits are set. 1049 1050Fixed a bug where a badly corrupted directory on a big endian system 1051could cause e2fsck to die with a bus error. The 1052ext2fs_read_dir_block() and ext2fs_process_dir_block() functions in 1053the ext2 library now does alignment sanity checks on the rec_len field 1054of the directory entry before using it. 1055 1056The ext2 library has been enhanced to make tune2fs safe to run on 1057mounted filesystems. (Users could usually get away with using tune2fs 1058on mounted filesystems before, but with the advent of ext3 and 1059journaling, it became important to make tune2fs was *really* safe for 1060use even when the filesystem being modified is mounted.) E2label is 1061now implemented by tune2fs using an argv[0] dispatch, so that e2label 1062is also now safe for use on mounted filesystems. 1063 1064Added a new program, e2image, which creates a backup of critical ext2 1065filesystem data structures. The generated image file can be examined 1066using dumpe2fs and debugfs. In the future, e2fsck will be able to use 1067the image file to help recover very badly damaged filesystems. 1068 1069Fixed a number of LFS bugs in e2fsck; very, very large (> 2**42) files 1070no longer cause e2fsck to bomb out. Also treat files > 2GB as being 1071large file when deciding whether or not the filesystem has large files. 1072 1073Fixed lsattr and chattr so that they work correctly on large files. 1074(Fixes Debian bug #72690.) 1075 1076Removed limitation in get_device_size() which imposed a one terrabyte 1077filesystem limitation. (Most 2.2 kernels still have a signed int 1078problem which cause 1 TB block device limitation. Fortunately, the 1079kernel patches to fix this are much easier than fixing the 2TB 1080limitation in the kernel. :-) 1081 1082A max_mount_count of zero is now treated as if no mount count were 1083set. (Previously, no mount count was indicated by using -1, and a 1084mount count of zero caused e2fsck to always be run.) 1085 1086Mke2fs supports two new filesystem types largefile and largefile4. 1087 1088Mke2fs now adds some randomness in s_max_mount_count so that multiple 1089filesystems won't be all checked at the same time under normal 1090operations. 1091 1092Fixed bug in the progress bar printing code which could cause e2fsck 1093to core dump on an illegal filesystem. 1094 1095Fixed bug in fsck which could allow more than one instance of e2fsck 1096to be printing a progress bar. (Debian bug #65267) 1097 1098Fsck using a UUID or a LABEL specifier will work even if devfs is 1099compiled into the kernel and not mounted. If the pathnames in 1100/proc/partitions are incorrect, fsck will search /dev for the correct 1101device (using the new ext2fs_find_block_device library function). 1102Fsck now also checks the RAID devices first so that they are properly 1103found when they are in use. Support has also been added to support 1104additional IDE disks and the DAC 960 device names. (Debian bug #94159) 1105 1106Fixed a bug in fsck which caused it not deal properly with 16 1107byte long filesystem labels. 1108 1109Fsck's -t option has been made a lot more flexible. The semantics for 1110what happens if a comma-separated list to fsck has been regularized, 1111and it is now possible to filter what filesystems will get checked 1112based what is in the filesystem's fstab entry's option field. (Debian 1113bug #89483.) 1114 1115The dumpe2fs program can now print out the group description 1116information in hex, and also prints the location of the superblock and 1117block group descriptor for those block groups that have them. 1118 1119Mke2fs now clears the ext2 superblock before it starts creating a 1120filesystem, so that the superblock magic number is only written if the 1121filesystem creation process successfully completes. 1122 1123The debugfs program's stat command now pretty-prints the blocks used 1124by an inode so that it's more compact and informative. 1125 1126The debugfs stats command now uses the same libe2p code (which is used 1127by dumpe2fs) to print the superblock header information. This is more 1128complete, and it avoids a bit of code duplication. 1129 1130Added a new debugfs command, set_super_value (ssv) which allows the 1131user to set arbitrary superblock fields. 1132 1133Debugfs was extended to support inode numbers in hex (by prefixing 1134them with 0x), and so that modify_inode can set the inode generation 1135number. Also, there is now a new function command called logdump 1136which will dump an ext3 journal. 1137 1138Fixed a bug in debugfs so that quitting out of the pager doesn't kill 1139debugfs. 1140 1141Debugfs's dump command now stops immediately upon reporting a disk 1142read error. (Fixed a bug in ext2fs_file_read library routine which 1143caused debugfs not to stop.) (Debian bug #79163) 1144 1145On systems with /proc/mounts (mainly Linux systems), /proc/mounts is 1146checked before /etc/mtab is used, since /proc/mounts is more likely to 1147be accurate. 1148 1149Added portability fixes for Solaris and Linux/ia64. 1150 1151Various manual pages were clarified and cleaned up. (Fixed debian 1152bugs #63442, #67446, and #87216) 1153 1154 1155Programmer's notes: 1156------------------- 1157 1158The e2fsck message printer now supports %Iu and %Ig, which will print 1159out the inode's user and group owners, respectively. 1160 1161E2fsprogs now includes its own version of include/linux/ext2_fs.h, so 1162that no longer dependent on the system having the correct version of 1163the kernel header files. 1164 1165Added a new function to libext2, ext2fs_find_block_device(), which 1166searches the system (i.e., /dev, /devfs, /devices) for a pathname to a 1167device given its device number. 1168 1169Added a new function to libext2, ext2fs_sync_device, which centralizes 1170all of the places which might try to use the BLKFLSBUF or FDFLUSH 1171ioctls (and usually failing to define them since the system header 1172files don't usually do this for us, and we're trying to avoid usage of 1173kernel include files now). 1174 1175Added new utility programs in tests/progs: random_exercise and 1176hold_inode. They aren't built by default; they're useful for 1177exercising ext3 filesystem code. 1178 1179Added a new ext2 filesystem flag, EXT2_FLAG_SUPER_ONLY, which causes 1180the filesystem close functions to only update the superblock, and to 1181not touch the block group descriptors. Needed by tune2fs when 1182modifying a mounted filesystem. 1183 1184Got rid of struct ext2fs_sb and replaced it with the standard struct 1185ext2_super_block from include/linux/ext2_fs.h. Note: this may break 1186source (but not binary) compatibility of some users of the ext2 1187library. Those applications should just simply do a global search and 1188replace of struct ext2fs_sb with struct ext2_super_block, and include 1189the new header file <ext2fs/ext2_fs.h> which defines it. 1190 1191The ino_t type has been renamed ext2_ino_t to protect applications 1192that attempt to compile -D_FILE_OFFSET_BITS=64, since this 1193inexplicably changes ino_t(!?). So we use ext2_ino_t to avoid an 1194unexpected ABI change. 1195 1196The Makefiles have been reworked so that "make check" can be run from 1197the top-level Makefile. 1198 1199Fix general gcc -Wall complaints and removed dead code. 1200 1201Remove use of NOARGS, because we assume everyone does ANSI C these 1202days. 1203 1204Added build-rpm script from sct. 1205 1206New functions ext2fs_image_{inode,super,bitmap}_{read,write} added 1207to support e2image. 1208 1209New function ext2fs_flush_icache which must be called if the 1210application program modifies the inode table blocks without going 1211through ext2fs_write_inode() interface. 1212 1213New ext2fs_check_mount_point() function, which will return the mount 1214point of a device if mounted. 1215 1216The io_channel abstraction now has an optional interface, 1217io_channel_write_range, which allows specific byte ranges to be 1218written. 1219 1220The unix_io IO channel now supports write-through caching, so that 1221journal creation is more efficient. 1222 1223Added x86 assembly language routines to support byte swapping, to 1224reduce executable size. 1225 1226Fixed bug in the utility program subst so that it's possible to 1227replace a substitution variable with a zero-length string. 1228 1229Fixed numbering e2fsck pass1 problem numbers; an extra zero had 1230slipped into some of the problem number. 1231 1232E2fsprogs 1.19 (July 13, 2000) 1233============================== 1234 1235Release the resize2fs program since the timeout before it could 1236be released under the GPL has finally expired. 1237 1238Add experimental support needed for the ext2 compression patches. 1239This requires compiling e2fsprogs with the --enable-compression flag 1240to the configure script. 1241 1242Added ext3 journalling support. E2fsck will run the journal (if 1243necessary) by temporarily mounting the filesystem. /sbin/fsck.ext3 is 1244installed as a symlink to e2fsck. Fsck has been taught about ext3, 1245and treats it the same as ext2 in terms of the progress bar logic. 1246Dumpe2fs will display the superblock journaling information if the 1247filesystem has a journal. The ext2 library will now permit opening an 1248ext3 filesystem with the recovery flag set. This is necessary for 1249on-line dump's to work correctly, but there may be issues with this 1250working well since ext3 is much less agressive about syncing blocks to 1251the filesystem, since they're safe on the journal. 1252 1253Tune2fs and e2fsck have been changed to allow the mount_count check to 1254be disabled by setting max_mount_count to -1. (This was already 1255supported by the kernel.) 1256 1257Create a symbolic link for fsck.ext3, since the e2fsprogs utilities 1258are used for ext3 as well. 1259 1260Added internationalization support for e2fsprogs; must be enabled 1261by passing --enable-nls to configure. 1262 1263Always use the provided ext2fs header files to insulate ourselves from 1264kernel version changes. Which include files are used by e2fsprogs 1265have also been cleaned up to improve portability. 1266 1267Limit the number of times that e2fsck updates the progress bar so that 1268people who are booting using a 9600 baud console don't get swampped by 1269too many updates. 1270 1271Improved the loop detection algorithm in e2sck's pass #3 so that it is 1272much, much faster for large filesystems with a large number of 1273directories. 1274 1275The memory footprint for e2fsck is now slightly smaller than before. 1276 1277E2fsck now checks if special devices have a non-zero size, and offers 1278to clear the size field if it finds such an inode. 1279 1280E2fsck now checks if special devices have the append-only flag set, 1281and offers to clear the inode. 1282 1283E2fsck now properly handles some "should never fail" cases during a 1284bitmap copy in pass5. 1285 1286E2fsck now properly prints control characters in filenames as ^A .. ^Z. 1287 1288E2fsck now calculates the correct location of the backup superblock in 1289the case of filesystem blocksizes > 1k. 1290 1291Fixed a bug in e2fsck's calculation of the number of inodes_per_block 1292which normally didn't cause problems under most filesystem parameters, 1293but could cause a valid superblock to be rejected in extreme cases. 1294Other checks for validating superblock values were made more 1295stringent. 1296 1297Added non-destructive write testing to the badblocks program, courtesy 1298of David Beattie. The badblocks also now has an option to input the 1299current set of bad blocks, so that known bad blocks are skipped to 1300speed up the badblocks test. There is also a persistent rescan 1301feature which causes badblocks to run until it has completed some 1302number of passes without discovering any new bad blocks. 1303 1304Badblocks now checks to see if the device is mounted and refuses to do 1305the tests involving writing to the device if it is mounted. Also, 1306badblocks now allows the number of blocks to be checked to be 1307defaulted to the size of the partition. 1308 1309Fixed a bug in fsck which didn't allow non-root users to be able to 1310check filesystems if there were any LABEL= or UUID= entries in 1311/etc/fstab. 1312 1313The Hurd doesn't support the filetype filesystem feature. The mke2fs 1314program now makes sure that for the Hurd, the filestype feature is 1315turned off. E2fsck will check to see if the filetype feature is 1316turned on for Hurd filesystems, and offer to turn off the feature. 1317 1318Mke2fs now has a safety check to make sure the number of blocks do not 1319exceed 32 bits even on a 64 bit platform. 1320 1321Really fixed a bug in fsck to allow "fsck -As" to run interactive 1322fsck's. (For those people who like to do interactive fsck's in the 1323/etc/rc scripts!?!) 1324 1325Debugfs has a few new features: the rdump command, which will do a 1326recursive dump of a directory and all of its contents, and the lcd 1327command which does a local chdir (much like the ftp command of the 1328same name). In addition, the debugfs program and the open_filesystem 1329command now takes three new options: -b and -s, which allows the 1330blocksize and superblock location to be specified, and the -c option 1331which is used in catastrophic situations where the block group 1332descriptors are corrupt. If the -c option is specified, debugfs will 1333skip trying to read in the block and inode bitmaps. 1334 1335Debufs's lsdel command was fixed to handle bad blocks in the inode 1336table. 1337 1338A Y2K bug in debugfs's "ls -l" handling was fixed by switching to use 13394 digit years. 1340 1341General improvements in error messages 1342 1343 - Mke2fs prints a sane error message if the partition size is zero 1344 (usually because the partition table wasn't reread by the 1345 kernel due to the partition being busy), instead of "invalid 1346 argument passed to ext2 library while initializing superblock". 1347 1348 - Fsck now prints more self-explanatory message if an invalid UUID= 1349 or LABEL= specification is passed to it. 1350 1351UUID library changed to use the LGPL. 1352 1353Fixed a bug in the UUID library where very rapid calls to the 1354time-based UUID generator could cause duplicate UUID's to be returned. 1355This was not a problem for e2fsprogs, but it could be a problem for 1356other users of the library. 1357 1358Make the UUID library more robust in the face of missing or an 1359improper /dev/urandom or /dev/random files. 1360 1361Added some random portability fixes for Solaris. 1362 1363Some minor man page updates. 1364 1365Fixed a memory leak in the ss library. 1366 1367 1368Programmer's notes: 1369------------------- 1370 1371We now try to use lseek64 and open64 from the LFS if possible. 1372 1373The 3rd parameter in e2p's print_flags is now a flags word, instead of 1374a boolean option. 1375 1376The mark and unmark bitmap functions now return the previous state of 1377the bit that was being changed, which is useful for some speed 1378optimizations. 1379 1380The following functions have been added to enhance the badblocks list 1381handling in libext2fs: ext2fs_write_bb_FILE, ext2fs_read_bb_FILE2, and 1382ext2fs_badblocks_equal. 1383 1384The ext2 header files now have the latest journalling fields to the 1385superblock. 1386 1387The ext2fs_mkdir function in libext2fs now properly backs out of error 1388conditions robustly. 1389 1390Cleaned up makefiles: 1391 - to cleanly compile with the -j flag. 1392 - so distclean removes all generated files. 1393 - so in case of an error while installing header files, the make aborts. 1394 1395Fix test_script so that it works correctly when compiling in the 1396source directory. 1397 1398The random UUID generation routine has been made slightly better in 1399the case where /dev/random doesn't exist. (Use of randomly-based UUID 1400is still not recommended if /dev/random doesn't exist, however; it's 1401better to use the time/ethernet MAC address UUID in this case.) 1402 1403Clean up the build process so it's more friendly in case of missing 1404directories. 1405 1406The ext2fs header file can now be #include'd into C++ programs. 1407 1408The e2p.h header file is now installed. 1409 1410Added workaround to a gawk 3.0.5 bug in lib/ss/mk_cmds. 1411 1412 1413 1414E2fsprogs 1.18 (November 10, 1999) 1415================================== 1416 1417Fix a core dumping bug in e2fsck if an imagic inode is present or 1418(more rarely) if the filesystem is badly corrupted enough that e2fsck 1419has to restart pass 1 processing. E2fsck now closes the filesystem 1420before freeing a large number of its data structures, so in the case 1421of future memory faults, at least the fixed filesystem will be fully 1422written out. 1423 1424If a filesystem doesn't support imagic inodes, and e2fsck discovers an 1425imagic inode, it will offer to clear the imagic flag. 1426 1427E2fsck will now offer to clear the immutable flag on special files 1428(device/socket/fifos) when running it in non-preen mode. 1429 1430E2fsck will now set the filetype when creating /lost+found, and when 1431connected orphaned inodes to /lost+found. 1432 1433Debugfs's ncheck and icheck commands now handles the case where there 1434are bad blocks in the inode table without bombing out. 1435 1436The badblocks list processing code has been made more efficiently for 1437appending a large number of (ordered) badblocks to the badblocks list. 1438 1439Some minor man page updates. 1440 1441Fsck now allows interactive e2fsck's when using fsck -As (not a common 1442mode, but some people like to do this in boot scripts for silly reasons). 1443 1444Programmer's notes: 1445------------------- 1446 1447The internal e2fsck problem code for PR_2_SPLIT_DOT was fixed to meet 1448with the problem code convention. 1449 1450The badblocks list regression test program has been updated to work 1451with previously made API name changes. 1452 1453The ext2fs_free() command now uses the new badblocks API to avoid 1454using the compatibility layer. 1455 1456Added new regression test cases; the run_e2fsck test script now 1457supports the ability for a test case to run a prepratory command 1458before running e2fsck. 1459 1460E2fsprogs 1.17 (October 26, 1999) 1461================================= 1462 1463Fixed nasty typo in fsck which caused parallelized fsck's to go into an 1464infinite loop. 1465 1466Fixed a bug in fsck where it used strncmp to compare a binary UUID, 1467thus potentially causing problems if a binary UUID contained a NULL 1468character. 1469 1470E2fsck now uses stricter checks for directory entries in pass 2: 1471zero-length filenames are not allowed; neither are 8 byte long 1472directory entries. 1473 1474The debugfs "dirty" command now clears the filesystem valid bit. 1475(Previously this just set the dirty-as-in-needs-writing-out-to-disk 1476bit in the in-core superblock image. The new functionality is more 1477what the user expects, and is more useful.) 1478 1479Added a debugging hook to test parallel fsck; if the environment 1480variable FSCK_FORCE_ALL_PARALLEL, then filesystems on the same drive 1481will be checked in parallel when they normally would not be. 1482 1483Programmer's notes: 1484------------------- 1485 1486Fixed some #ifdef's for compilation under the Hurd OS. 1487 1488Fixed minor W2K compatibility problems. 1489 1490Fixed some miscellaneous GCC warnings. 1491 1492 1493E2fsprogs 1.16 (October 22, 1999) 1494================================= 1495 1496Fixed a race condition bug in fsck; when printing a progress bar, if 1497checking multiple filesystems in parallel, it was possible for fsck to 1498send e2fsck a SIGUSR1 signal before e2fsck had installed its signal 1499handler, which would cause it to terminate with a signal 10. 1500 1501E2fsck now properly handles filesystems that have the 1502INCOMPAT_FILETYPE feature turned on. It can be used to convert a 1503filesystem into using or not using FILETYPE feature. 1504 1505E2fsck now properly handles filesystems that have the IMAGIC feature 1506turned on (this is used on Linux AFS servers). 1507 1508The mke2fs program now creates filesystems that have the filetype and 1509sparse_superblock features enabled by default, unless it is run on a 1510pre-2.2 kernel. These features are not supported by a pre-2.2 kernel, 1511so there is now a new flag -O which allows the user to specify with 1512which features she would like to create the filesystem; "mke2fs -O 1513none" will create a filesystsem compatible with 2.0 kernels. 1514 1515The tune2fs program now has a -O option which allows the user to set 1516and reset "safe" filesystem features. Currently, the only ones which 1517allows to be modified are the filetype and sparse_superblock features. 1518Note setting or clearing either feature will require running e2fsck on 1519the filesystem afterwards. (n.b. Clearing the sparse_superblock feature 1520requires that there is enough free space on the filesystem for the 1521extra superblocks which will be created by e2fsck.) 1522 1523Debugfs can now set and print filesystem features in the superblock 1524using the "features" command. Dumpe2fs will print out the complete 1525set of features when listing the superblock. 1526 1527Dumpe2fs has new options -f (force) and -h (header-only). 1528 1529Fixed a bug in e2fsck which could cause the PROGRAMMING ERROR/bonehead 1530message to come up. This could happen when decrementing or 1531incrementing a link count could result in an overflow. 1532 1533Fixed a bug in e2fsck where the block count on the lost+found 1534directory would not be properly incremented when the directory was 1535expanded to the point where an indirect block needed to be allocated. 1536 1537E2fsck now makes some additional sanity checks on the superblock to 1538avoid crashing or giving a memory allocation error if some of the 1539values in the superblock are unresonable (but the superblock otherwise 1540looks valid). 1541 1542Fixed a bug in e2fsck where a very badly corrupted filesystem might 1543require two passes to completely fix the filesystem. This happened if 1544an inode claimed blocks that was part of the filesystem metadata 1545(typically, when garbage was written into an inode table or indirect 1546block, since this kind of filesystem corruption normally doesn't 1547happen otherwise). 1548 1549On the Alpha, glibc declares st_flags although it isn't actually used; 1550the configure script was improved to detect this case so that 1551e2fsprogs can avoid using the non-functional stat field. 1552 1553The manual pages were updated to use a more consistent formatting 1554style consistent with standard Unix man pages. Mke2fs's man page 1555added documentation for a few previously undocumented options. 1556 1557Fixed minor display bugs in tune2fs and mke2fs. 1558 1559Programmer's notes: 1560------------------- 1561 1562Improved portability of e2fsprogs to non-Unix systems (in particular, NT). 1563 1564Added features to parse and print feature strings into the e2p library. 1565(e2p_feature2string, e2p_string2feature, e2p_edit_feature). 1566 1567ext2fs_mkdir() and ext2fs_new_dir_block() now creates directories 1568whose directory entries contain proper filetype information if the 1569filesystem supports it. 1570 1571ext2fs_link() now uses the low 3 bits of its flags parameter to pass 1572the directory entry filetype information. This is used to set the 1573directory entry filetype information if the filesystem supports it. 1574 1575Fixed a bug in ext2fs_expand_dir() where the block count in a 1576directory's inode would not be properly incremented when the directory 1577was expanded to the point where an indirect block needed to be 1578allocated. 1579 1580 1581E2fsprogs 1.15 (July 18, 1999) 1582============================== 1583 1584Add configuration checks so that e2fsprogs will compile cleanly on 1585Linux 2.3 kernels that have renamed i_version to i_generation. 1586 1587E2fsck now prints a progress/completion bar (and not just a simple 1588spinner) if the -C0 option is requested or if it receives a SIGUSR1 1589signal. Fsck will automatically manage the (potentially muliple) 1590e2fsck processes to print completion bars if it is given a -C option, 1591with the right thing happening if multiple filesystems are being 1592checked in parallel. 1593 1594Mke2fs now has better automatic hueristics to determine the filesystem 1595parameters to be used for a particular filesystem. Added a new option 1596-T which allows the user to specify how the filesystem is to be used, 1597which helps mke2fs do a better job selecting the filesystem parameters. 1598 1599Mke2fs now creates revision 1 filesystems by default, and with the 1600sparse superblock feature enabled. The sparse superblock feature is 1601not understood by Linux 2.0 kernels, so they will only allow read-only 1602mounts of filesystems with this sparse superblocks. 1603 1604Fix bug where if /dev/null couldn't be opened (should never happen), 1605e2fsck would hang in a tight loop. 1606 1607Make e2fsck handle the case where /lost+found isn't a directory. 1608 1609E2fsck now uses mallinfo if it exists to get accurate statistics about 1610its memory usage. 1611 1612Fix bug in e2fsck where it wouldn't check to see if a disconnected 1613inode had any problems before connecting it to /lost+found. 1614 1615Add check to e2fsck so it makes sure that total number of inodes in 1616the filesystem is a sane number. 1617 1618Fix fencepost error when clearing an the end of the block bitmap which 1619caused the last block in the bitmap not to get cleared. 1620 1621Cleaned up a number of messages in e2fsck: 1622 * The message "Group's #'s copy of the group descriptor..." 1623 was fixed so that the correct number would be displayed. 1624 * Added missing space in the "disk write-protected" error messsage 1625 * Cleaned up the error message printed when a non-interactive 1626 e2fsck needs to abort a check because the filesystem 1627 appears to be mounted. 1628 1629Added a new command-line utility, uuidgen, which will create and print 1630a UUID. 1631 1632Make debugfs's icheck command more robust by checking to make sure an 1633inode has valid blocks before interarting over the inode's blocks. 1634 1635UUID generation now uses a random-based scheme whenever possible to 1636prevent potential privacy problems. 1637 1638Man pages for all of the UUID functions in the lirbary were added. 1639 1640Fixed bug in fsck so it won't coredump if a filesystem not in 1641/etc/fstab is given to it. 1642 1643Fsck now understands the UUID=xxxx and LABEL=yyyy forms in /etc/fstab 1644that most of the other mount utilities understands. 1645 1646Mke2fs will make a filesystem even if it appears mounted if the force 1647option is given. 1648 1649Dumpe2fs has new command-line options which allow a filesystem expert 1650to specify the superblock and blocksize when opening a filesystem. 1651This is mainly useful when examining the remains of a toasted 1652filesystem. 1653 1654The badblocks program has been updated to display correctly on disks 1655with large block numbers. 1656 1657The badblocks program no longer gives spurious errors when errors 1658occur on non-block boundaries, which is common if the blocksize is 1659larger than 1k. 1660 1661Mke2fs will sync the disk device every MKE2FS_SYNC block groups if the 1662MKE2FS_SYNC environment variable is set. This is to work around a VM 1663bug in the 2.0 kernel. I've heard a report that a RAID user was able 1664to trigger it even using a 2.2 kernel, but hopefully it will not be 1665needed for most Linux 2.2 users. 1666 1667Fixed miscellaneous documentation and man pages. 1668 1669Programmer's notes: 1670------------------- 1671 1672Cleaned up functions such as pass1_get_blocks, pass1_read_inode which 1673in e2fsck's pass1.c really should have been static. 1674 1675The return value of the uuid_compare() function was changed to make it 1676match with the convetions used by strcmp, memcmp, and Paul Leach's 1677UUID sample document. 1678 1679The "make depend" process has now been made more automated; it now 1680automatically word-wraps the dependencies, and only replaces source 1681Makefile.in if there has been a change in the dependencies. Also, a 1682top-level "make depend" now recurses through all the subdirectories 1683automatically. 1684 1685The Makefile in .../util has been changed so that subst is built using 1686the native C compiler during a cross-compilation, since the subst 1687program is only used during the build process. Also add an explicit 1688rule to build util/subst by cd'ing to the correct directory and 1689running Makefile. 1690 1691The man directories are defined in terms mandir, so that the configure 1692script can override the location of the manual pages. 1693 1694The config files have been updated to recognize new machine types for 1695both the i386 and alpha families. 1696 1697Fsck has been modified so that it will accurately create an 1698fsck_instance even when the noexecute flag is set. This allows for 1699accurate debugging of the fsck pass structure. Also, when the verbose 1700flag is given twice, fsck will print debugging information about when 1701fsck is waiting for jobs to finish. 1702 1703 1704E2fsprogs 1.14 (January 9, 1999) 1705================================ 1706 1707Fix the fstab parsing code so that it can handle blank lines and 1708comment characters. Also, missing pass numbers need to be treated as 1709zero. 1710 1711Fixed a bug in e2fsck where under some circumstances (when e2fsck 1712needs to restart processing after fixing an egregious inconsistency) 1713it would try to access already freed memory. 1714 1715E2fsck now prints non-printable characters in directory entries and 1716pathnames using '^' and 'M-' notation. 1717 1718Fixed chattr so that it will ignore symbolic links when doing 1719recursive descent traversals. For both chattr and lsattr, no longer 1720print the version string unless the -V option is given. 1721 1722Allow the system administrator to directly specify the number of 1723inodes desired in the filesystem, for some special cases where this is 1724necessary. 1725 1726Fix portability problems so that e2fsprogs can be compiled under Linux 17271.2 systems and Solaris systems. 1728 1729Update the config.guess file with a more recent version that will 1730identify newer Linux platforms. 1731 1732Programmer's notes 1733------------------ 1734 1735Ext2fs_read_inode and ext2fs_write_inode will now return an error if 1736an inode number of zero is passed to them. 1737 1738E2fsprogs 1.13 (December 15, 1998) 1739================================== 1740 1741Fixed a bug in debugfs where an error messages weren't getting printed 1742when the ext2 library routines to read inodes returned errors in the 1743stat, cmri and rm commands. 1744 1745Fixed a bug in mke2fs so that if a ridiculous inode ratio parameter is 1746provided, it won't create an inode table smaller than the minimum 1747number of inodes required for a proper ext2 filesystem. 1748 1749Fsck now parses the /etc/fstab file directly (instead of using 1750getmntent()), so that it can distinguish between a missing pass number 1751field and pass number field of zero. This caused problems for 1752diskless workstations where all of the filesystems in /etc/fstab have 1753an explicit pass number of zero, and fsck could not distinguish this 1754from a /etc/fstab file with missing pass numbers. 1755 1756E2fsck will create a /lost+found directory if there isn't one in the 1757filesystem, since it's safer to create the lost+found directory before 1758it's needed. 1759 1760Fixed e2fsck so that it would detect bogus immutable inodes which 1761happen to be sockets and FIFO files, and offer to clear them. 1762 1763If a filesystem has multiple reasons why it needs to be checked, and 1764one of the reasons is that it is uncleanly mounted, e2fsck will print 1765that as the reason why the filesystem is being checked. 1766 1767Cleaned up the output routines of mke2fs so that it doesn't overflow 1768an 80 column display when formating really big filesystems. 1769 1770Added a sanity check to e2fsck to make sure that file descriptors 0, 17711, 2 are open before opening the hard disk. This avoids a problem 1772where a broken program might exec e2fsck with those file descriptors 1773closed, which would cause disastrous results if the kernel returns a 1774file descriptor for the block device which is also used by FILE * 1775stdout. 1776 1777Fixed up the e2fsck progress reporting functions so that the values 1778reliably reach 100% at the completion of all of the e2fsck passes. 1779 1780Fixed minor documentation bugs in man pages and usage messages. 1781 1782Programmer's notes: 1783------------------- 1784 1785Fixed a number of lint warnings in the ext2fs library and potential 1786portability problems from other OS's header files that might define 1787CPP macros for names like "max" and "min". 1788 1789ext2fs_badblocks_list_add() has been made more efficient when it needs 1790to grow the bad blocks list. 1791 1792Fixed a bug in e2fsck which caused it to dereference a freed pointer 1793just before exiting. 1794 1795Fixed the substition process for generating the mk_cmds and compile_et 1796scripts so that they will work outside of the build tree. 1797 1798Add sanity check to e2fsck so that if an internal routine 1799(ext2fs_get_dir_info) returns NULL, avoid dereferencing the pointer 1800and causing a core dump. This should never happen, but... 1801 1802E2fsprogs 1.12 (July 9, 1998) 1803================================== 1804 1805E2fsprogs now works with glibc (at least with the version shipped wtih 1806RedHat 5.0). The ext2fs_llseek() function should now work even with 1807i386 ELF shared libraries and if llseek() is not present. We also 1808explicitly do a configure test to see if (a) llseek is in libc, and 1809(b) if llseek is declared in the system header files. (See standard 1810complaints about libc developers don't understand the concept of 1811compatibility with previous versions of libc.) 1812 1813The ext2fs library now writes out the block group number in each of 1814the superblock copies. This makes it easier to automatically 1815determine the starting block group of the filesystem when the block 1816group information is trashed. 1817 1818Added support for the EXT2_FEATURE_INCOMPAT_FILETYPE feature, 1819which means that e2fsprogs will ignore the high 8 bits of the 1820directory entry's name_len field, so that it can be used for other 1821purposes. 1822 1823Added support for the EXT2_FEATURE_RO_COMPAT_LARGE_FILE feature. 1824E2fsprogs will now support filesystems with 64-bit sized files. 1825 1826Added support for the EXT2_FEATURE_COMPAT_DIR_PREALLOC feature. 1827 1828Added new program "e2label", contributed by Andries Brouwer. E2label 1829provides an easy-to-use interface to modify the filesystem label. 1830 1831Fixed bug so that lsattr -v works instead of producing a core dump. 1832 1833Fixed a minor bug in mke2fs so that all groups with bad superblock 1834backup blocks are printed (not just the first one). 1835 1836Mke2fs will check the size of the device, and if the user specifies a 1837filesystem size larger than the apparent size of the device it will 1838print a warning message and ask if the user wants to proceed. 1839 1840E2fsck has a new option -C, which sends completion information to the 1841specified file descriptor. For the most part, this is intended for 1842programs to use, although -C 0 will print a spinning character to the 1843stdout device, which may be useful for users who want to see something 1844happening while e2fsck goes about its business. 1845 1846Fixed a bug in e2fsck which could cause a core dump when it needs to 1847expand the /lost+found directory, and sometimes the bitmaps haven't 1848been merged in. Also fixed a related bug where ext2fs_write_dir_block 1849was used to write out a non-directory block. (Which would be bad on a 1850non-Intel platform with byte swapping going on.) 1851 1852Fixed bug in e2fsck where it would print a "programming error" message 1853instead of correctly identifying where a bad block was in used when 1854the bad block was in a non-primary superblock or block group 1855descriptor. Also fixed a related bug when sparse superblocks are in 1856use and there is a bad block where a superblock or block group 1857descriptor would have been in a group that doesn't include a 1858superblock. 1859 1860Fixed a bug in e2fsck (really in libext2fs's dblist function) where if 1861the block group descriptor table is corrupt, it was possible to try to 1862allocate a huge array, fail, and then abort e2fsck. 1863ext2fs_get_num_dirs() now sanity checks the block group descriptor, 1864and subsitutes reasonable values if the descriptors are obviously bogus. 1865 1866If e2fsck finds a device file which has the immutable flag set and the 1867i_blocks beyond the normal device number are non-zero, e2fsck will 1868offer to remove it, since it's probably caused by garbage in the inode 1869table. 1870 1871When opening a filesystem, e2fsck specially checks for the EROFS error 1872code, and prints a specific error message to the user which is more 1873user friendly. 1874 1875If the filesystem revision is too high, change the e2fsck to print 1876that this is either because e2fsck is out of date, or because the 1877superblock is corrupt. 1878 1879E2fsck now checks for directories that have duplicate '.' and '..' 1880entries, and fixes this corruption. 1881 1882E2fsck no longer forces a sync of the filesystem (with attendant sleep 1883calls) at all times. The ext2fs_flush() function now performs a sync 1884only if it needed to write data blocks to disk. 1885 1886Fixed a minor bug in e2fsck's pass1b's file cloning function, where 1887certain errors would not be properly reported. 1888 1889Updated and expanded a few points in the man pages which users 1890complained wheren't explicit enough. 1891 1892Added special case byte-swapping code if compiling on the PowerPC, to 1893accomodate the strange big-endian variant of the ext2 filesystem that 1894was previously used on the PowerPC port. 1895 1896 1897Programmer's notes: 1898------------------- 1899 1900Removed C++ keywords from the ext2fs libraries so that it could be 1901compiled with C++. 1902 1903E2fsck's internal organization has now been massively reorganized so 1904that pass*.c don't have any printf statements. Instead, all problems 1905are reported through the fix_problem() abstraction interface. E2fsck 1906has also been revamped so that it can be called as a library from a 1907application. 1908 1909Added new fileio primitives in libext2fs for reading and writing 1910files on an unmounted ext2 filesystem. This interface is now used by 1911debugfs. 1912 1913Added a new libext2fs function for mapping logical block numbers of 1914a file to a physical block number. 1915 1916Added a new libext2fs function, ext2fs_alloc_block(), which allocates 1917a block, zeros it, and updates the filesystem accounting records 1918appropriately. 1919 1920Added a new libext2fs function, ext2fs_set_bitmap_padding(), which 1921sets the padding of the bitmap to be all one's. Used by e2fsck pass 5. 1922 1923The libext2fs functions now use a set of memory allocation wrapper 1924functions: ext2fs_get_mem, ext2fs_free_mem, and ext2fs_resize_mem, 1925instead of malloc, free, and resize. This makes it easier for us to 1926be ported to strange environments where malloc, et. al. aren't 1927necessarily available. 1928 1929Change the libext2fs fucntion to return ext2-specific error codes 1930(EXT2_DIR_EXISTS and EXT2_DB_NOT_FOUND, for example) instead of using 1931and depending on the existence of system error codes (such as EEXIST 1932and ENOENT). 1933 1934Renamed io.h to ext2_io.h to avoid collision with other OS's header 1935files. 1936 1937Add protection against ext2_io.h and ext2fs.h being included multiple 1938times. 1939 1940The types used for memory lengths, etc. have been made more portable. 1941In generla, the code has been made 16-bit safe. Added Mark 1942Habersack's contributed DOS disk i/o routines. 1943 1944Miscellaneous portability fixes, including not depending on char's 1945being signed. 1946 1947The io_channel structure has a new element, app_data, which is 1948initialized by the ext2fs routines to contain a copy of the filesystem 1949handle. 1950 1951ext2fs_check_directory()'s callback function may now return the error 1952EXT2_ET_CALLBACK_NOTHANDLED if it wishes ext2fs_check_directory() to 1953really do the checking, despite the presence of the callback function. 1954 1955 1956E2fsprosg 1.11 (June 17, 1997) 1957============================== 1958 1959Fixed e2fsck to detect (previously ignored) conflicts between the 1960superblock or block group descriptors and block bitmaps, inode 1961bitmaps, and inode tables. 1962 1963Fixed bug in e2fsck so that when the message printed out when a block 1964or inode bitmap conflicts with other data, it has the correct group 1965number. 1966 1967Fixed bug in e2fsck and mke2fs where the blocksize wasn't being passed 1968to badblocks. This meant that not all of the filesystem was being 1969tested for bad blocks! 1970 1971Fixed an array boundary overrun case which cropped up in 1972ext2fs_badblocks_list_test when a user tried running "mke2fs -c 1973-b 4096". 1974 1975Adjusted the number of columns printed by mke2fs when displaying the 1976superblock backups to avoid running over 80 columns when making a 1977really big filesystem. 1978 1979Fixed up the man pages for e2fsck, debugfs, badblocks, chattr, 1980dumpe2fs, fsck, mke2fs, and tune2fs (typos and other minor grammar 1981fixes), thanks to some suggestions from Bill Hawes (whawes@star.net). 1982 1983Programmer's notes: 1984------------------- 1985 1986Fixed install rule in lib/ss so that ss_err.h is actually getting 1987installed. 1988 1989Fixed bug in ext2fs_copy_bitmap; the destination bitmap wasn't getting 1990bassed back to the caller. 1991 1992Fixed bug in ext2fs_inode_scan_goto_blockgroup; it had not been 1993setting the current inode number (which meant this function wasn't 1994working at all). 1995 1996Fixed bug in ext2fs_resize_generic_bitmap; it had not be zeroing all 1997blocks in the bitmap when increasing the size of the bitmap. 1998 1999Changed the initial number of blocks allocated by ext2fs_init_dblist() 2000to be more realistic. 2001 2002Added a new function ext2fs_allocate_group_table, which sets up the 2003group descriptor information (and allocates inode and block bitmaps, 2004and inode tables for a particular group). The function was created by 2005factoring out code form ext2fs_allocate_tables(). 2006 2007Added a new function ext2fs_move_blocks which takes a bitmap of the 2008blocks to be moved, and moves them to another location on the 2009boardboard. 2010 2011Make the unix_io channel's io_channel_flush implementation calls sync() 2012to to flush the kernel buffers to disk. 2013 2014Added a new function ext2fs_dblist_count returns the number of 2015directory blocks in dblist. 2016 2017 2018E2fsprogs 1.10 (April 24, 1997) 2019=============================== 2020 2021Mke2fs once again defaults to creating revision #0 filesystems, since 2022people were complaining about breaking compatibility with 1.2 kernels. 2023Warning messages were added to the mke2fs and tune2fs man pages that 2024the sparse superblock option isn't supported by most kernels yet (1.2 2025and 2.0 both don't support parse superblocks.) 2026 2027Added new flag to mke2fs, -R <raid options>, which allows the user to 2028tell mke2fs about the RAID configuration of the filesystem. Currently 2029the only supported raid option is "stride" which specifies the width 2030of the RAID stripe. 2031 2032Fixed bug in e2fsck where pass1b would bomb out if there were any 2033blocks marked bad in the inode table. 2034 2035Fixed rare bug in mke2fs where if the user had a very unlucky number 2036of blocks in a filesystem (probability less than .002) the resulting 2037filesystem would be corrupt in the last block group. 2038 2039Fixed bug where if e2fsck tried to allocate a block to fix a 2040filesystem corruption problem and the filesystem had no free blocks, 2041ext2fs_new_block() would loop forever. 2042 2043The configure script now checks explicitly to see if "-static" works, 2044since that can't be assumed to be true --- RedHat doesn't install 2045libc-static by default. 2046 2047Fixed bug in libext2's block iterator functions where under some 2048cirmcustances, file with holes would cause the bcount parameter to the 2049callback function to be incorrect. This bug didn't affect any of 2050e2fsprogs programs, but it was discovered by Paul Mackerras, the 2051author of the PPC boot loader. 2052 2053Removed use of static variables to store the inode cache in libext2fs. 2054This caused problems if more than one filesystem was accessed via 2055libext2fs (static variables in libraries are generally a bad idea). 2056Again, this didn't affect e2fsprogs programs, but it was discovered by 2057Paul Mackerras. 2058 2059Fixed minor bugs and version code drift to assure that e2fsprogs 1.10 2060will compile cleanly with 1.2.13 kernels (even with a.out shared 2061libraries!) 2062 2063Programmer's notes: 2064------------------- 2065 2066Added new functions to duplicate an ext2 filesystem handle, and its 2067associated substructure. New functions: ext2fs_dup_handle(), 2068ext2fs_copy_dblist(), ext2fs_badblocks_copy(), ext2fs_copy_bitmap(). 2069Other structures, such as the io_channel and the inode_cache, now have 2070a ref count so that they only get freed when they are no longer used 2071by any filesystem handle. (These functions were added as part of the 2072development effort for an ext2 resizer). 2073 2074E2fsprogs 1.09 (April 14, 1997) 2075=============================== 2076 2077Fixed bug in mke2fs (really in lib/ext2fs/initialize.c) which was 2078accidentally introduced in the 1.08 release. The overhead calculation 2079was accidentally removed, which caused ext2fs_initialize() to not 2080notice when the filesystem size needed to be adjusted down because 2081there wasn't enough space in the last block group. 2082 2083Fixed bug in version parsing library routine; it was always parsing 2084the library version string, instead of using the passed-in string. 2085 2086Clarified chattr man page. 2087 2088E2fsprogs 1.08 (April 10, 1997) 2089=============================== 2090 2091E2fsck 1.07 was very slow when checking very large filesystems with a 2092lot of files that had hard links (i.e., news spools). This was fixed 2093by seriously revamping the icount abstraction. Added a formal test 2094suite for the icount abstraction. 2095 2096Debugfs now has a "-l" option to the "ls" command, which lists the 2097inode number, permissions, owner, group, size, and name of the files 2098in the directory. 2099 2100Fix a bug in e2fsck where when a directory had its blocks moved to 2101another location during the pass 1b processing, the directory block 2102list wasn't updated, so pass 2 wouldn't check (and correct) the 2103correct directory block. 2104 2105E2fsck will now treat inodes which contain blocks which are claimed by 2106the filesystem metadata by treating them as multiply claimed blocks. 2107This way, the data in those blocks can be copied to a new block during 2108the pass 1b--1d processing. 2109 2110E2fsck will attempt to determine the correct superblock number and 2111display it in the diagnostic and warning messages if possible. 2112 2113Add support for a new (incompatible) feature, "sparse_super". This 2114feature reduces the number of blocks which contain copies of backup 2115superblocks and block group descriptors. (It is only an incompatible 2116feature because of a bug in ext2_free_blocks.) mke2fs and tune2fs now 2117support a new -s option; e2fsck will recognize filesystems built with 2118this feature turned on. 2119 2120E2fsck now checks the library to make sure is the correct version, 2121using new library functions. (This helps to diagnose incorrectly 2122installed e2fsprogs distributions.) 2123 2124Dumpe2fs now prints more information; its now prints the the 2125filesystem revision number, the filesystem sparse_super feature (if 2126present), the block ranges for each block group, and the offset from 2127the beginning of the block group. 2128 2129Mke2fs now distributes the inode and block bitmap blok so that the 2130won't be concentrated in one or two disks in RAID/striping setups. 2131Also, if the user chooses a 2k or 4k block group, mke2fs will try to 2132choose the largest blocks per group that be chosen. (For 2k blocks, 2133you can have up to 16384 blocks/group; for 4k blocks, you can have up 2134to 32768 blocks/group.) Previously mke2fs would not allow 2135specification of more than 8192 blocks per group, even if you were 2136using a 2k or 4k block group. 2137 2138Programmer's notes: 2139------------------- 2140 2141Added a new function ext2fs_create_icount2() which takes a "hint" 2142argument. This hint argument presets the icount array with the list 2143of inodes which actually need to be in the icount array. This really 2144helps to speed up e2fsck. 2145 2146Added a new function ext2fs_icount_validate() which checks the rep 2147invariant for the icount structure. This is used mostly for testing. 2148 2149The error mesasage given when a bad inode number is passed to 2150test_generic_bitmap to reflect EXT2FS_TEST_ERROR (instead of 2151EXT2FS_UNMARK_ERROR). 2152 2153Added a new function ext2fs_set_dir_block which sets the block of a 2154dblist entry, given the directory inode and blockcnt. 2155 2156Added a new function ext2fs_get_library_version() which returns the 2157current library version, and ext2fs_parse_version_string() which 2158returns a version number based on a e2fsprogs version string. 2159 2160The icount functions will return EINVAL if the passed in inode number 2161is out of bounds. 2162 2163E2fsprogs 1.07 (March 9, 1997) 2164============================== 2165 2166E2fsck is now uses much less memory when checking really large 2167filesystems (or rather, filesystems with a large number of inodes). 2168Previously a filesystem with 1 million inodes required 4 megabytes of 2169memory to store inode count statistics; that storage requirement has 2170now been reduced to roughly half a megabyte. 2171 2172E2fsck can now properly deal with bad blocks appearing inside the 2173inode table. Instead of trying to relocate the inode table (which 2174often failed because there wasn't enough space), the inodes in the bad 2175block are marked as in use. 2176 2177E2fsck will automatically try to use the backup superblocks if the 2178primary superblocks have a bad magic number or have missing meta-data 2179blocks (or meta-data blocks which are out of range). 2180 2181E2fsck's pass 3 has been made more efficient; most noticeable on 2182filesystems with a very large number of directories. 2183 2184Completely revamped e2fsck's system of printing problem reports. It 2185is now table driven, to make them more easily customizeable and 2186extendable. Error messages which can be printed out during preen mode 2187are now one line long. 2188 2189Fixed e2fsck's filesystem swapping code so that it won't try to swap 2190fast symbolic links or deleted files. 2191 2192Fixed e2fsck core dumping when fixing a filesystem which has no 2193directories (not even a root directory). 2194 2195Added a check to e2fsck to make sure that the length of every 2196directory entry is a multiple of 4 (since the kernel complains if it 2197isn't). 2198 2199Added a check to e2fsck to make sure that a directory entry isn't a 2200link to the root directory, since that isn't allowed. 2201 2202Added a check to e2fsk to now make sure the '.' and '..' directory 2203entries are null terminated, since the 2.0 kernel requires it. 2204 2205Added check to write_bitmaps() to make sure the superblock doesn't get 2206trashed if the inode or block bitmap is marked as being block zero. 2207 2208Added checking of the new feature set fields in the superblock, to 2209avoid dealing with new filesystem features that this package wasn't 2210set up to handle. 2211 2212Fixed a fencepost error in ext2fs_new_block() which would occasionally 2213try to allocate a block beyond the end of a filesystem. 2214 2215When the UUID library picks a random IEEE 802 address (because it 2216can't find one from a network card), it sets the multicast bit, to 2217avoid conflicting with a legitimate IEEE 802 address. 2218 2219Mke2fs now sets the root directory's owner to be the real uid of the 2220user running mke2fs. If the real uid is non-zero, it also sets 2221the group ownership of the root directory to be the real group-id of 2222the user running mke2fs. 2223 2224Mke2fs now has more intelligent error checking when it is given a 2225non-existent device. 2226 2227When badblocks is given the -vv option, it now updates the block that 2228it is currently testing on every block. 2229 2230Fixed a bug in fsck where it wouldn't modify the PATH envirnoment 2231currently correctly if PATH wasn't already set. 2232 2233Shared libraries now built with dependencies. This allows the shared 2234library files to be used with dlopen(); it also makes the transition 2235to libc 6 easier, since ld.so can tell which libc a particular shared 2236library expects to use. 2237 2238Programmer's notes: 2239------------------- 2240 2241Added new abstraction (defined in dblist.c) for maintaining a list of 2242blocks which belongs to directories. This is used in e2fsck and other 2243programs which need to iterate over all directories. 2244 2245Added new functions which test to see if a contiguous range of blocks 2246(or inodes) are available. (ext2fs_*_bitmap_range). 2247 2248Added new function (ext2_inode_has_valid_blocks) which returns true if 2249an inode has valid blocks. (moved from e2fsck code). 2250 2251Added new function (ext2fs_allocate_tables) which allocates the 2252meta-data blocks as part of initializing a filesystem. (moved from 2253mke2fs code). 2254 2255Added a new I/O manager for testing purposes. It will either allow a 2256program to intercept I/O requests, or print debugging messages to 2257trace the activity of a program using the I/O manager. 2258 2259The badblocks_list functions now store the bad blocks in a sorted 2260order, and use a binary search to speed up badblocks_list_test. 2261 2262The inode scan function ext2fs_get_next_inode() may now return a soft 2263error returns: MISSING_INODE_TABLE and BAD_BLOCK_IN_INODE_TABLE in 2264those cases where part of an inode table is missing or there is a bad 2265block in the inode table. 2266 2267Added a new function (ext2fs_block_iterate2) which adds new arguments to 2268the callback function to return a pointer (block and offset) to the 2269reference of the block. 2270 2271Added new function (ext2fs_inode_scan_goto_blockgroup) which allows an 2272application to jump to a particular block group while doing an inode 2273scan. 2274 2275The badblocks list functions were renamed from badblocks_* to 2276ext2fs_badblocks_*. Backwards compatibility functions are available 2277for now, but programs should be modified to use the new interface. 2278 2279Some of the library functions were reorganized into separate files to 2280reduce the size of some programs which statically link against the 2281ext2 library. 2282 2283Put in some miscellaneous fixes for the Alpha platform. 2284 2285 2286E2fsprogs 1.06 (October 7, 1996) 2287================================ 2288 2289Fixed serious bug in e2fsck: if the block descriptors are bad, don't 2290smash the backup copies in ext2fs_close(). (The problem was that when 2291e2fsck -p discovered the problem, while it was closing the filesystem 2292and exiting, it was also blowing away the backup superblocks on the 2293disk, which was less than friendly.) We now make it the case that we 2294only write out the backup superblock and the back block descriptors if 2295the filesystem is completely free from problems. 2296 2297Fixed a bug in block_interate in the lib/ext2fs library which caused 2298e2fsck to fail on GNU Hurd-created filesystems. 2299 2300Add support for Linux/FT's bootloader, which actually uses 2301EXT2_BOOT_LOADER, and sets its mode bits which caused e2fsck to want 2302to clear the inode. 2303 2304Add support for the "A" (no atime update) attribute. (Note: this 2305attribute is not yet in production kernels.) 2306 2307The test suite is not automatically run when doing a "make all" from 2308the top level directory. Users should manually run "make check" if 2309they wish to run the test suite. 2310 2311Upon a preenhalt(), make the printed message more explicit that 2312running e2fsck "MANAULLY" means without the -p or -a options. 2313 2314In e2fsck, if a disconnected inode is zero-length, offer to clear it 2315instead of offering to connect it to lost+found. 2316 2317In e2fsck, if a filesystem was just unmounted uncleanly, and needs 2318e2fsck to be run over it, change e2fsck to explicitly display this 2319fact. 2320 2321For dumpe2fs and e2fsck, cause the -V option to print out which 2322version of the ext2fs library is actually getting used. (This will 2323help detect mismatches of using a 1.06 utility with a 1.05 library, 2324etc.) 2325 2326Programmers' notes: 2327------------------- 2328 2329EXT2_SWAP_BYTES was changed to EXT2_FLAG_SWAP_BYTES, which better fits 2330the naming convention. 2331 2332In ext2fs_initialize(), make sure the description for the inode bitmap 2333is correctly initialize. 2334 2335Fixed minor type typo in ext2fs_allocate_generic_bitmap(); 2336 2337E2fsprogs 1.05 (September 7, 1996) 2338================================== 2339 2340Add support for new fields in the ext2 superblock --- volume name, 2341volume UUID, and last mounted field. Dumpe2fs displays these fields, 2342tune2fs and mke2fs allows you to set them. E2fsck will automatically 2343generate a UUID for those volumes that don't have them. 2344 2345Put in support for e2fsck to recognize HURD specific ext2 features --- 2346most notably, the translator block. The e2fsprogs tools will now use 2347the creator_os field in the superblock to correctly handle different 2348OS-specific variants of the ext2 filesystem. 2349 2350E2fsck now fixes inodes which have a the deletion time set, but which 2351have a non-zero i_link_count field by offering to clear the deletion 2352time. Previously e2fsck assumed that the inode was deleted (per 0.3c 2353ext2 kernel behavior) and offered to unlink the file. 2354 2355If e2fsck sets the clean bit, but nothing else, set the exit code 2356FSCK_NONDESTRUCT. After all, e2fsck did fix a filesystem error --- it 2357set the filesystem valid bit when it was previously cleared. :-) This 2358was needed to make the HURD fsck driver happy. 2359 2360If the user refuses to attach an unattached inode, e2fsck will no 2361longer set the inode's link count. Otherwise, the inode would end up 2362getting marked as unused, which might cause loss of data later. 2363 2364Make the message issued by e2fsck when the superblock is corrupt less 2365confusing for users. It now mentions that another reason for the 2366"corrupt superblock" message might be that the partition might not be 2367an ext2 filesystem at all (it might swap, msdos filesystem, ufs, etc.) 2368 2369Make the libext2 library more robuest so that e2fsck won't coredump on 2370an illegal superblock where the blocksize is zero. (f_crashdisk is 2371the test case). 2372 2373By default, create filesystems where the default checkinterval is 6 2374months (180 days). Linux servers can be robust enough that 20 reboots 2375can be a long, long time. 2376 2377Added configure flag --enable-old-bitops, which forces the bitops to 2378use the old (native) bitmask operations. By default on the sparc 2379platform, the standard ext2 bit ordering is now used. 2380 2381Added a new feature to e2fsck to byte-swap filesystems; this can be 2382used to convert old m68k filesystems to use the standard byte-order 2383storage for the superblock, inodes, and directory blocks. This 2384function is invoked by using the '-s' option to e2fsck. 2385 2386Debugfs's "dump" command has been enhanced so that it writes out the 2387exact size of the file so that the nulls at the end of the file are 2388eliminated. The command also accept a new "-p" option which will 2389attempt preserve to preserve the ownernship, permissions, and 2390file modification/access times. 2391 2392Debugfs has two new options, -f and -R. The -R option allows the user 2393to execute a single debugfs command from the command line. The -f 2394option allows the user to specify a "command file" containing debugfs 2395commands which will get executed. 2396 2397Dumpe2fs now pretty prints the check interval, instead of just 2398printing the check interval as a number of seconds. 2399 2400Fix bugs in debugfs: the params command when no filesystem is opened 2401no longer causes a core dump. It is now possible to unlink a file 2402when a pathame containing a '/' is specified. 2403 2404Tune2fs has a new -C option which sets the number of times the 2405filesystem has been mounted. 2406 2407Fix the chattr '-v' option so that it actually works. Chattr was 2408being buggy about the -v option parsing. 2409 2410Programmers' notes: 2411------------------- 2412 2413The directory lib/uuid contains a set of library routines to generate 2414DCE compatible UUIDs. 2415 2416Extended ext2fs_namei() to handle symbolic links. Added new function 2417ext2fs_nami_follow() which will follow last symbolic link in the case 2418where the pathname points to a sym link. 2419 2420The ext2fs_block_iterate function will now return the HURD translator 2421block, if present. The new flag BLOCK_FLAG_DATA_ONLY will cause the 2422iterator to return data blocks only. The ext2fs.h file now defines 2423constants BLOCK_COUNT_IND, BLOCK_COUNT_DIND, BLOCK_COUNT_TIND, and 2424BLOCK_COUNT_TRANSLATOR, which are the magic values passed in the block 2425count field of the iterator callback function. 2426 2427The test script driver now takes an optional second argument, which is 2428the test case to be run. This allows you to run a test case without 2429needing to run the entire test suite. 2430 2431On Linux ELF systems, install the .so files in the correct places 2432(/usr/lib). The .so files must be stored in the same directory as the 2433.a files. 2434 2435Fixed miscellaneous HURD compilation issues with header file being 2436included in the right order. 2437 2438Fixed debugfs so that it resets optind to zero, not one, since setting 2439optind to zero is more correct. 2440 2441 2442E2fsprogs 1.04 (May 16, 1996) 2443============================= 2444 2445First "official" (1.03 was a limited release only) to support building 2446e2fsprogs under Linux 2.0 kernels (as well as late model 1.3 and 1.99 2447kernels). 2448 2449This package includes a RPM specs file, that it can be built using the 2450RedHat Package Manager. 2451 2452E2fsck now prints a hint that if there are lots of incorrectly located 2453inode bitmaps, block bitmaps, and inode table blocks, the user might 2454want to try using e2fsck -b 8193 first, to see if that fares any 2455better. 2456 2457For ext2 filesystem written with the hurd, debugfs will now print out 2458the translator field when printing an inode structure. 2459 2460Lots of miscellaneous linking/installation cleanups: 2461 2462 Libraries are now linked using a relative pathname, instead of 2463 relying on -L working correct. It doesn't, in many cases, including 2464 current versions of GNU ld. This guarantees that the build tree is 2465 linking with the right libraries, instead of the ones installed in 2466 /usr/lib. 2467 2468 Header files, man pages, and the et/ss shell scripts are now 2469 generated using a custom substitution script, instead of relying on 2470 the configure script. This prevents needless recompilation of 2471 files; in addition, the custom substitution script is much faster. 2472 2473 e2fsck may now be linked dynamically, by using the 2474 --enable-dynamic-e2fsck flag to configure. This is not recommended, 2475 since it increases e2fsck's dependence on other files, but some 2476 people need to save disk space, and other critical programs on their 2477 systems were being linked dynamically anyway. 2478 2479 Programs such as fsck which didn't need to be linked against 2480 libext2fs (or mke2fs which didn't need to be linked against libe2p) 2481 only link against libraries they actually need. Otherwise, those 2482 programs would require the presense of libraries that otherwise 2483 could be removed from a rescuse diskette. 2484 2485 The ss include files are now installed correctly so they can 2486 actually be used by another package. 2487 2488 If the profiling libraries are built, they are now installed on a 2489 "make install-libs". 2490 2491 2492E2fsprogs 1.03 (March 27, 1996) 2493=============================== 2494 2495Change the m68k bit numbering for bitmasks to match the bit numbering 2496used by all other ext2 implementations. (This change was requested by 2497the m68k kernel development team.) 2498 2499Support (in-development) filesystem format revision which supports 2500(among other things) dynamically sized inodes. 2501 2502Fixed a bug in the ext2 library so that an intelligent error is 2503returned if mke2fs is run with a ridiculously small number of blocks 2504for a partition. 2505 2506Fixed a bug in the ext2 library which required that the device be 2507openable in read/write mode in order to determine its size. This 2508caused e2fsck -n to require read/write access when it was not 2509previously necessary. 2510 2511Fixed a bug in e2fsck which casued it to occasionally fail the test 2512suite depending on which version of the floating point library it was 2513using. 2514 2515Fixed a bug in e2fsck so that it now halts with a fatal error when 2516certain superblock consistency checks fail. Previously it continued 2517running e2fsck, with some potential confusing/damaging consequences. 2518 2519Added new flag to fsck which allows the root to be checked in parallel 2520with other filesytems. This is not the safest thing in the world to 2521do, but some system administrators really wanted it. 2522 2523Fixed -Wall flames in lib/ss. 2524 2525 2526E2fsprogs 1.02 (January 16, 1996) 2527================================= 2528 2529Fix to allow e2fsprogs to be compiled on recent 1.3 (pl45+) kernels. 2530 2531Change e2fsck to print statistics of how many non-contiguous files are 2532on the system. Note that a file which is larger than 8k blocks, it is 2533guaranteed to be non-contiguous. 2534 2535In mke2fs, print a warning message if a user tries to format a whole 2536disk (/dev/hda versus /dev/hda1). If a user really wants to format a 2537whole disk, the -F (force) option forces mke2fs to format a whole disk 2538as a filesytem. 2539 2540Fix a bug in fsck where in some cases it might start checking 2541partitions in the next pass before it finishes checking partitions in 2542the current pass. This still won't cause two partitions on the same 2543disk will be checked, so it's rarely a problem in real life. 2544 2545Patch lsattr so that it won't hang when checking a named pipe. 2546 2547Minor compilation fixes: 2548 * Fix the order of libraries that were linked in debugfs. 2549 * Allow the sources to be compiled with -ansi turned on. 2550