RELEASE-NOTES revision d494bba807e74dcd1e93ae4a18b81451a57c21bd
1E2fsprogs 1.23-WIP (August 9, 2001) 2================================== 3 4Add initial support for extended attributes (EA); e2fsck will 5correctly handle a filesystem with EA's and check the EA blocks for 6corruptions. 7 8E2fsck's symlink sanity checking has been cleaned up. It now checks 9the i_size value of fast symlinks, and checks for immutable flags 10being set of symlinks, etc. 11 12E2image can now create "raw" image files, which only contain the 13filesystem metadata placed in a spare file so that e2fsck, dumpe2fs, 14debugfs, etc., can be run directly on the raw image file. 15 16Add support for the 64-bit block device patches. 17 18Fixed bugs in creating external journals with a 1k blocksize. 19 20Add initial support for external journals (so long as the external 21journal only supports a single filesystem) in e2fsck. 22 23Remove requirement for needing to run fsck on a filesystem after 24removing a journal (either internal or external). 25 26The man pages now document how to create and manage external journals. 27 28Speed up the check of an ext3 filesystems by avoiding a needless flush 29of all of the superblock and block group descriptors. 30 31Speed up creating an internal journal using tune2fs in the case where 32the filesystem has a lot blocks already allocated. 33 34Fixed a minor bug in mke2fs; if -O none is passed to mke2fs, it will 35now not set the sparse_super feature. 36 37Add support in fsck for the filesystem type "auto". 38 39E2fsck now gives an explicit warning if there filesystem still has 40errors at the end of the run. (Addresses Debian bug #104502) 41 42E2fsck will set the EXT2_ERROR_FS flag if the journal superblock 43reflects an error. 44 45E2fsck now prints the number of large files when given the -v option. 46 47Fixed minor memory leaks in e2fsck. 48 49Some minor man pages updates. (Addresses Debian bug #30833) 50 51 52E2fsprogs 1.22 (June 22, 2001) 53============================== 54 55Fixed a bug in e2fsck's handling of orphan inodes which are special 56files (i.e., block/character device files, named FIFO's, etc.). 57 58Fixed a bug accidentally introduced in e2fsprogs 1.21 which caused 59tune2fs to fail at adding a journal to a mounted filesystem. 60 61Fixed a few big-endian bugs in e2fsprogs 62 * The directory block functions were accidentally reporting 63 some directories as corrupted when they weren't. 64 * If e2fsprogs is compiled --disable-swapfs, the C language 65 equivalents weren't being included for big-endian platforms. 66 (Fixes Debian bug #101686). 67 68Fixed a Hurd compilation problem. (Addresses Debian bug #101361) 69 70Programmer's notes: 71------------------- 72 73Use platform independent method of defining the BLKFLSBUF and FDFLUSH 74ioctls, and allow them to be defined for non-i386 platforms. 75 76The uuid.h header file is now protected against multiple inclusions. 77 78E2fsprogs is now being developed using BitKeeper. Changed the test 79scripts to deal with BK's stripping CR characgters from text files, 80and changed the top-level Makefile.in to avoid including BitKeeper 81files when generating the source tarball. 82 83 84E2fsprogs 1.21 (June 15, 2001) 85============================== 86 87Added new configure flags which allow a subset e2fsprogs to be built; 88this is most useful for boot floopies, since the resulting shared 89libraries and programs are slimmed down by removing features that 90aren't necessary for a boot floppy. The new flags that were added 91are: --disable-swapfs, which removes support for byte swapping old 92legacy PPC/68k filesystems, --disable-debugfs, which removes support 93for debugfs from the libext2fs library, --disable-imager, which 94removes support for the e2image program, and --disable-resizer, which 95removes support for resize2fs. 96 97E2fsck now prints the number of mounts or days elapsed since the last 98check when e2fsck is forced to check an otherwise clean filesystem. 99 100Tune2fs now prints an informative message about how often a filesystem 101will be checked when adding a journal to the filesystem, to remind the 102user that he/she may want to adjust those parameters using tune2fs 103-c/-i. 104 105Worked around hurd brain-damage which causes e2fsck to sometimes 106believe a filesystem is the root filesystem based on device numbers 107(since Hurd doesn't have dev_t's, which is arguably a POSIX.1 108violation). 109 110Fixed a bug introduced in 1.20 which caused e2fsck to abort with an 111erroneous error with the -F option was specified. 112 113Fixed a ext3 recovery bug in the revoke handling; synchronized with 114ext3 0.7a. 115 116Fixed two bugs in e2fsck's handling of dup block handling, dealing 117with relatively uncommon edge cases: a directory with an indirect 118block which is claimed by another file, and when the last inode in the 119filesystem has blocks claimed by another file. 120 121E2fsck now checks to see if the i_size field of a fast symlink is too 122big, and offers to clear the symlink if so. 123 124E2fsck now checks to see if i_size_high of special files is non-zero, 125and offers to clear i_size_high. 126 127Fix e2fsck's handling of incompatible journal flags so that the user 128has chance to abort, and then has the option to clear out the journal 129entirely. (Addresses Debian bug #98527.) 130 131Fixed a bug in fsck which could cause it to core dump if a mix of 132standard and non-standard device names are used in /etc/fstab. 133(Debian bug #100559) 134 135Fixed a bug in debugfs which caused read errors when copying a file to 136not be noticed. 137 138The debugfs set_super_value command can now modify the s_lastcheck field. 139 140Fixed a bug in lsattr and chattr which was accidentally introduced in 1411.20 to support > 2GB files; both lsattr and chattr wasn't reading 142directories correctly because the change modified the layout of struct 143dirent to be incompatible with the libe2p shared library. 144 145Cleaned up the mke2fs manual page and included a discussion about why 146it's good to periodically check the filesystem even when journaling is 147enabled. 148 149Programmer's notes: 150------------------- 151 152Fix general gcc -Wall complaints. 153 154The types needed by the ext2 header files are now provided by 155lib/ext2fs/ext2_types.h, instead of include/asm/types.h. 156 157Integers are now preferred to longs when trying to find a 32-bit type 158in ext2_types.h. Also, if linux/types.h has already been defined, 159don't try to redefine the types. 160 161Fixed make depend script so that it automatically corrects the 162pathname cleanups performed by make -M, so I don't have to fix them up 163by hand. 164 165Fixed the d_loaddump test case to be more robust, and not depend on 166bash'isms. 167 168Removed debugfs's dependence on pread(), which was accidentally 169intrudced in e2fsprogs 1.20 170 171Fixed a performance bug in the libext2fs's icount routine; the size 172estimate of the icount array was incorrectly being calculated. 173 174Removed use of the badblocks compatibility functions in the e2fsprogs 175programs. 176 177Added paranoia code which protects against strange cases where /etc 178isn't on the root filesystem, or if /etc/mtab doesn't exist. 179 180The header file ext2_types.h is now installed. 181 182Autoconf is used to determine when we are on big-endian machines, 183instead of doing run-time tests, to save a few bytes of code. 184 185The ext2fs_mark_generic_bitmap and ext2fs_unmark_generic_bitmap 186functions are no longer inline functions, which saves space and 187doesn't really cost any real performance. 188 189The ext2fs library no longer depends on the e2p library. (What need 190there was of it --- namely, fsetflags, was coded in-line). 191 192Fixed the makefile so that lib/ext2fs/ext2_types.h is generated even 193when the user is stupid and tries compiling the package using "make 194install" as root. 195 196Miscellaneous code cleanups: 197 * Added missing files from Makefile.in's SRCS file, so that 198 their dependencies would be properly calculated. 199 * Removed redundant code 200 * Fixed comments in code 201 * Removed no-longer unneeded argsused #pragma. 202 203 204E2fsprogs 1.20 (May 20, 2001) 205============================= 206 207Add support for replaying the ext3 journal purely in user mode, 208including handling the orphaned inode list. Used code contributed by 209Andreas Dilger, with bug fixes and the orphaned inode handling done 210by Theodore Ts'o. 211 212The mke2fs and tune2fs programs can create or modify a filesystem to 213include an ext3 journal. Tune2fs also can be used to remove an ext3 214journal from a filesystem. 215 216E2fsck will now check for the existence of a linked list of orphan 217inodes in the superblock, and clear those inodes before starting the 218rest of the filesystem check (but after the journal playback). 219 220E2fsck now validates the file descriptor passed to the -C option, 221which saves against the completion bar getting written to an 222unexpected location, such as the disk being checked. (Debian 223bug/wishlist #55220) 224 225E2fsck will now bump the filesystem revision number from zero to one 226if any of the compatibility bits are set. 227 228Fixed a bug where a badly corrupted directory on a big endian system 229could cause e2fsck to die with a bus error. The 230ext2fs_read_dir_block() and ext2fs_process_dir_block() functions in 231the ext2 library now does alignment sanity checks on the rec_len field 232of the directory entry before using it. 233 234The ext2 library has been enhanced to make tune2fs safe to run on 235mounted filesystems. (Users could usually get away with using tune2fs 236on mounted filesystems before, but with the advent of ext3 and 237journaling, it became important to make tune2fs was *really* safe for 238use even when the filesystem being modified is mounted.) E2label is 239now implemented by tune2fs using an argv[0] dispatch, so that e2label 240is also now safe for use on mounted filesystems. 241 242Added a new program, e2image, which creates a backup of critical ext2 243filesystem data structures. The generated image file can be examined 244using dumpe2fs and debugfs. In the future, e2fsck will be able to use 245the image file to help recover very badly damaged filesystems. 246 247Fixed a number of LFS bugs in e2fsck; very, very large (> 2**42) files 248no longer cause e2fsck to bomb out. Also treat files > 2GB as being 249large file when deciding whether or not the filesystem has large files. 250 251Fixed lsattr and chattr so that they work correctly on large files. 252(Fixes Debian bug #72690.) 253 254Removed limitation in get_device_size() which imposed a one terrabyte 255filesystem limitation. (Most 2.2 kernels still have a signed int 256problem which cause 1 TB block device limitation. Fortunately, the 257kernel patches to fix this are much easier than fixing the 2TB 258limitation in the kernel. :-) 259 260A max_mount_count of zero is now treated as if no mount count were 261set. (Previously, no mount count was indicated by using -1, and a 262mount count of zero caused e2fsck to always be run.) 263 264Mke2fs supports two new filesystem types largefile and largefile4. 265 266Mke2fs now adds some randomness in s_max_mount_count so that multiple 267filesystems won't be all checked at the same time under normal 268operations. 269 270Fixed bug in the progress bar printing code which could cause e2fsck 271to core dump on an illegal filesystem. 272 273Fixed bug in fsck which could allow more than one instance of e2fsck 274to be printing a progress bar. (Debian bug #65267) 275 276Fsck using a UUID or a LABEL specifier will work even if devfs is 277compiled into the kernel and not mounted. If the pathnames in 278/proc/partitions are incorrect, fsck will search /dev for the correct 279device (using the new ext2fs_find_block_device library function). 280Fsck now also checks the RAID devices first so that they are properly 281found when they are in use. Support has also been added to support 282additional IDE disks and the DAC 960 device names. (Debian bug #94159) 283 284Fixed a bug in fsck which caused it not deal properly with 16 285byte long filesystem labels. 286 287Fsck's -t option has been made a lot more flexible. The semantics for 288what happens if a comma-separated list to fsck has been regularized, 289and it is now possible to filter what filesystems will get checked 290based what is in the filesystem's fstab entry's option field. (Debian 291bug #89483.) 292 293The dumpe2fs program can now print out the group description 294information in hex, and also prints the location of the superblock and 295block group descriptor for those block groups that have them. 296 297Mke2fs now clears the ext2 superblock before it starts creating a 298filesystem, so that the superblock magic number is only written if the 299filesystem creation process successfully completes. 300 301The debugfs program's stat command now pretty-prints the blocks used 302by an inode so that it's more compact and informative. 303 304The debugfs stats command now uses the same libe2p code (which is used 305by dumpe2fs) to print the superblock header information. This is more 306complete, and it avoids a bit of code duplication. 307 308Added a new debugfs command, set_super_value (ssv) which allows the 309user to set arbitrary superblock fields. 310 311Debugfs was extended to support inode numbers in hex (by prefixing 312them with 0x), and so that modify_inode can set the inode generation 313number. Also, there is now a new function command called logdump 314which will dump an ext3 journal. 315 316Fixed a bug in debugfs so that quitting out of the pager doesn't kill 317debugfs. 318 319Debugfs's dump command now stops immediately upon reporting a disk 320read error. (Fixed a bug in ext2fs_file_read library routine which 321caused debugfs not to stop.) (Debian bug #79163) 322 323On systems with /proc/mounts (mainly Linux systems), /proc/mounts is 324checked before /etc/mtab is used, since /proc/mounts is more likely to 325be accurate. 326 327Added portability fixes for Solaris and Linux/ia64. 328 329Various manual pages were clarified and cleaned up. (Fixed debian 330bugs #63442, #67446, and #87216) 331 332 333Programmer's notes: 334------------------- 335 336The e2fsck message printer now supports %Iu and %Ig, which will print 337out the inode's user and group owners, respectively. 338 339E2fsprogs now includes its own version of include/linux/ext2_fs.h, so 340that no longer dependent on the system having the correct version of 341the kernel header files. 342 343Added a new function to libext2, ext2fs_find_block_device(), which 344searches the system (i.e., /dev, /devfs, /devices) for a pathname to a 345device given its device number. 346 347Added a new function to libext2, ext2fs_sync_device, which centralizes 348all of the places which might try to use the BLKFLSBUF or FDFLUSH 349ioctls (and usually failing to define them since the system header 350files don't usually do this for us, and we're trying to avoid usage of 351kernel include files now). 352 353Added new utility programs in tests/progs: random_exercise and 354hold_inode. They aren't built by default; they're useful for 355exercising ext3 filesystem code. 356 357Added a new ext2 filesystem flag, EXT2_FLAG_SUPER_ONLY, which causes 358the filesystem close functions to only update the superblock, and to 359not touch the block group descriptors. Needed by tune2fs when 360modifying a mounted filesystem. 361 362Got rid of struct ext2fs_sb and replaced it with the standard struct 363ext2_super_block from include/linux/ext2_fs.h. Note: this may break 364source (but not binary) compatibility of some users of the ext2 365library. Those applications should just simply do a global search and 366replace of struct ext2fs_sb with struct ext2_super_block, and include 367the new header file <ext2fs/ext2_fs.h> which defines it. 368 369The ino_t type has been renamed ext2_ino_t to protect applications 370that attempt to compile -D_FILE_OFFSET_BITS=64, since this 371inexplicably changes ino_t(!?). So we use ext2_ino_t to avoid an 372unexpected ABI change. 373 374The Makefiles have been reworked so that "make check" can be run from 375the top-level Makefile. 376 377Fix general gcc -Wall complaints and removed dead code. 378 379Remove use of NOARGS, because we assume everyone does ANSI C these 380days. 381 382Added build-rpm script from sct. 383 384New functions ext2fs_image_{inode,super,bitmap}_{read,write} added 385to support e2image. 386 387New function ext2fs_flush_icache which must be called if the 388application program modifies the inode table blocks without going 389through ext2fs_write_inode() interface. 390 391New ext2fs_check_mount_point() function, which will return the mount 392point of a device if mounted. 393 394The io_channel abstraction now has an optional interface, 395io_channel_write_range, which allows specific byte ranges to be 396written. 397 398The unix_io IO channel now supports write-through caching, so that 399journal creation is more efficient. 400 401Added x86 assembly language routines to support byte swapping, to 402reduce executable size. 403 404Fixed bug in the utility program subst so that it's possible to 405replace a substitution variable with a zero-length string. 406 407Fixed numbering e2fsck pass1 problem numbers; an extra zero had 408slipped into some of the problem number. 409 410E2fsprogs 1.19 (July 13, 2000) 411============================== 412 413Release the resize2fs program since the timeout before it could 414be released under the GPL has finally expired. 415 416Add experimental support needed for the ext2 compression patches. 417This requires compiling e2fsprogs with the --enable-compression flag 418to the configure script. 419 420Added ext3 journalling support. E2fsck will run the journal (if 421necessary) by temporarily mounting the filesystem. /sbin/fsck.ext3 is 422installed as a symlink to e2fsck. Fsck has been taught about ext3, 423and treats it the same as ext2 in terms of the progress bar logic. 424Dumpe2fs will display the superblock journaling information if the 425filesystem has a journal. The ext2 library will now permit opening an 426ext3 filesystem with the recovery flag set. This is necessary for 427on-line dump's to work correctly, but there may be issues with this 428working well since ext3 is much less agressive about syncing blocks to 429the filesystem, since they're safe on the journal. 430 431Tune2fs and e2fsck have been changed to allow the mount_count check to 432be disabled by setting max_mount_count to -1. (This was already 433supported by the kernel.) 434 435Create a symbolic link for fsck.ext3, since the e2fsprogs utilities 436are used for ext3 as well. 437 438Added internationalization support for e2fsprogs; must be enabled 439by passing --enable-nls to configure. 440 441Always use the provided ext2fs header files to insulate ourselves from 442kernel version changes. Which include files are used by e2fsprogs 443have also been cleaned up to improve portability. 444 445Limit the number of times that e2fsck updates the progress bar so that 446people who are booting using a 9600 baud console don't get swampped by 447too many updates. 448 449Improved the loop detection algorithm in e2sck's pass #3 so that it is 450much, much faster for large filesystems with a large number of 451directories. 452 453The memory footprint for e2fsck is now slightly smaller than before. 454 455E2fsck now checks if special devices have a non-zero size, and offers 456to clear the size field if it finds such an inode. 457 458E2fsck now checks if special devices have the append-only flag set, 459and offers to clear the inode. 460 461E2fsck now properly handles some "should never fail" cases during a 462bitmap copy in pass5. 463 464E2fsck now properly prints control characters in filenames as ^A .. ^Z. 465 466E2fsck now calculates the correct location of the backup superblock in 467the case of filesystem blocksizes > 1k. 468 469Fixed a bug in e2fsck's calculation of the number of inodes_per_block 470which normally didn't cause problems under most filesystem parameters, 471but could cause a valid superblock to be rejected in extreme cases. 472Other checks for validating superblock values were made more 473stringent. 474 475Added non-destructive write testing to the badblocks program, courtesy 476of David Beattie. The badblocks also now has an option to input the 477current set of bad blocks, so that known bad blocks are skipped to 478speed up the badblocks test. There is also a persistent rescan 479feature which causes badblocks to run until it has completed some 480number of passes without discovering any new bad blocks. 481 482Badblocks now checks to see if the device is mounted and refuses to do 483the tests involving writing to the device if it is mounted. Also, 484badblocks now allows the number of blocks to be checked to be 485defaulted to the size of the partition. 486 487Fixed a bug in fsck which didn't allow non-root users to be able to 488check filesystems if there were any LABEL= or UUID= entries in 489/etc/fstab. 490 491The Hurd doesn't support the filetype filesystem feature. The mke2fs 492program now makes sure that for the Hurd, the filestype feature is 493turned off. E2fsck will check to see if the filetype feature is 494turned on for Hurd filesystems, and offer to turn off the feature. 495 496Mke2fs now has a safety check to make sure the number of blocks do not 497exceed 32 bits even on a 64 bit platform. 498 499Really fixed a bug in fsck to allow "fsck -As" to run interactive 500fsck's. (For those people who like to do interactive fsck's in the 501/etc/rc scripts!?!) 502 503Debugfs has a few new features: the rdump command, which will do a 504recursive dump of a directory and all of its contents, and the lcd 505command which does a local chdir (much like the ftp command of the 506same name). In addition, the debugfs program and the open_filesystem 507command now takes three new options: -b and -s, which allows the 508blocksize and superblock location to be specified, and the -c option 509which is used in catastrophic situations where the block group 510descriptors are corrupt. If the -c option is specified, debugfs will 511skip trying to read in the block and inode bitmaps. 512 513Debufs's lsdel command was fixed to handle bad blocks in the inode 514table. 515 516A Y2K bug in debugfs's "ls -l" handling was fixed by switching to use 5174 digit years. 518 519General improvements in error messages 520 521 - Mke2fs prints a sane error message if the partition size is zero 522 (usually because the partition table wasn't reread by the 523 kernel due to the partition being busy), instead of "invalid 524 argument passed to ext2 library while initializing superblock". 525 526 - Fsck now prints more self-explanatory message if an invalid UUID= 527 or LABEL= specification is passed to it. 528 529UUID library changed to use the LGPL. 530 531Fixed a bug in the UUID library where very rapid calls to the 532time-based UUID generator could cause duplicate UUID's to be returned. 533This was not a problem for e2fsprogs, but it could be a problem for 534other users of the library. 535 536Make the UUID library more robust in the face of missing or an 537improper /dev/urandom or /dev/random files. 538 539Added some random portability fixes for Solaris. 540 541Some minor man page updates. 542 543Fixed a memory leak in the ss library. 544 545 546Programmer's notes: 547------------------- 548 549We now try to use lseek64 and open64 from the LFS if possible. 550 551The 3rd parameter in e2p's print_flags is now a flags word, instead of 552a boolean option. 553 554The mark and unmark bitmap functions now return the previous state of 555the bit that was being changed, which is useful for some speed 556optimizations. 557 558The following functions have been added to enhance the badblocks list 559handling in libext2fs: ext2fs_write_bb_FILE, ext2fs_read_bb_FILE2, and 560ext2fs_badblocks_equal. 561 562The ext2 header files now have the latest journalling fields to the 563superblock. 564 565The ext2fs_mkdir function in libext2fs now properly backs out of error 566conditions robustly. 567 568Cleaned up makefiles: 569 - to cleanly compile with the -j flag. 570 - so distclean removes all generated files. 571 - so in case of an error while installing header files, the make aborts. 572 573Fix test_script so that it works correctly when compiling in the 574source directory. 575 576The random UUID generation routine has been made slightly better in 577the case where /dev/random doesn't exist. (Use of randomly-based UUID 578is still not recommended if /dev/random doesn't exist, however; it's 579better to use the time/ethernet MAC address UUID in this case.) 580 581Clean up the build process so it's more friendly in case of missing 582directories. 583 584The ext2fs header file can now be #include'd into C++ programs. 585 586The e2p.h header file is now installed. 587 588Added workaround to a gawk 3.0.5 bug in lib/ss/mk_cmds. 589 590 591 592E2fsprogs 1.18 (November 10, 1999) 593================================== 594 595Fix a core dumping bug in e2fsck if an imagic inode is present or 596(more rarely) if the filesystem is badly corrupted enough that e2fsck 597has to restart pass 1 processing. E2fsck now closes the filesystem 598before freeing a large number of its data structures, so in the case 599of future memory faults, at least the fixed filesystem will be fully 600written out. 601 602If a filesystem doesn't support imagic inodes, and e2fsck discovers an 603imagic inode, it will offer to clear the imagic flag. 604 605E2fsck will now offer to clear the immutable flag on special files 606(device/socket/fifos) when running it in non-preen mode. 607 608E2fsck will now set the filetype when creating /lost+found, and when 609connected orphaned inodes to /lost+found. 610 611Debugfs's ncheck and icheck commands now handles the case where there 612are bad blocks in the inode table without bombing out. 613 614The badblocks list processing code has been made more efficiently for 615appending a large number of (ordered) badblocks to the badblocks list. 616 617Some minor man page updates. 618 619Fsck now allows interactive e2fsck's when using fsck -As (not a common 620mode, but some people like to do this in boot scripts for silly reasons). 621 622Programmer's notes: 623------------------- 624 625The internal e2fsck problem code for PR_2_SPLIT_DOT was fixed to meet 626with the problem code convention. 627 628The badblocks list regression test program has been updated to work 629with previously made API name changes. 630 631The ext2fs_free() command now uses the new badblocks API to avoid 632using the compatibility layer. 633 634Added new regression test cases; the run_e2fsck test script now 635supports the ability for a test case to run a prepratory command 636before running e2fsck. 637 638E2fsprogs 1.17 (October 26, 1999) 639================================= 640 641Fixed nasty typo in fsck which caused parallelized fsck's to go into an 642infinite loop. 643 644Fixed a bug in fsck where it used strncmp to compare a binary UUID, 645thus potentially causing problems if a binary UUID contained a NULL 646character. 647 648E2fsck now uses stricter checks for directory entries in pass 2: 649zero-length filenames are not allowed; neither are 8 byte long 650directory entries. 651 652The debugfs "dirty" command now clears the filesystem valid bit. 653(Previously this just set the dirty-as-in-needs-writing-out-to-disk 654bit in the in-core superblock image. The new functionality is more 655what the user expects, and is more useful.) 656 657Added a debugging hook to test parallel fsck; if the environment 658variable FSCK_FORCE_ALL_PARALLEL, then filesystems on the same drive 659will be checked in parallel when they normally would not be. 660 661Programmer's notes: 662------------------- 663 664Fixed some #ifdef's for compilation under the Hurd OS. 665 666Fixed minor W2K compatibility problems. 667 668Fixed some miscellaneous GCC warnings. 669 670 671E2fsprogs 1.16 (October 22, 1999) 672================================= 673 674Fixed a race condition bug in fsck; when printing a progress bar, if 675checking multiple filesystems in parallel, it was possible for fsck to 676send e2fsck a SIGUSR1 signal before e2fsck had installed its signal 677handler, which would cause it to terminate with a signal 10. 678 679E2fsck now properly handles filesystems that have the 680INCOMPAT_FILETYPE feature turned on. It can be used to convert a 681filesystem into using or not using FILETYPE feature. 682 683E2fsck now properly handles filesystems that have the IMAGIC feature 684turned on (this is used on Linux AFS servers). 685 686The mke2fs program now creates filesystems that have the filetype and 687sparse_superblock features enabled by default, unless it is run on a 688pre-2.2 kernel. These features are not supported by a pre-2.2 kernel, 689so there is now a new flag -O which allows the user to specify with 690which features she would like to create the filesystem; "mke2fs -O 691none" will create a filesystsem compatible with 2.0 kernels. 692 693The tune2fs program now has a -O option which allows the user to set 694and reset "safe" filesystem features. Currently, the only ones which 695allows to be modified are the filetype and sparse_superblock features. 696Note setting or clearing either feature will require running e2fsck on 697the filesystem afterwards. (n.b. Clearing the sparse_superblock feature 698requires that there is enough free space on the filesystem for the 699extra superblocks which will be created by e2fsck.) 700 701Debugfs can now set and print filesystem features in the superblock 702using the "features" command. Dumpe2fs will print out the complete 703set of features when listing the superblock. 704 705Dumpe2fs has new options -f (force) and -h (header-only). 706 707Fixed a bug in e2fsck which could cause the PROGRAMMING ERROR/bonehead 708message to come up. This could happen when decrementing or 709incrementing a link count could result in an overflow. 710 711Fixed a bug in e2fsck where the block count on the lost+found 712directory would not be properly incremented when the directory was 713expanded to the point where an indirect block needed to be allocated. 714 715E2fsck now makes some additional sanity checks on the superblock to 716avoid crashing or giving a memory allocation error if some of the 717values in the superblock are unresonable (but the superblock otherwise 718looks valid). 719 720Fixed a bug in e2fsck where a very badly corrupted filesystem might 721require two passes to completely fix the filesystem. This happened if 722an inode claimed blocks that was part of the filesystem metadata 723(typically, when garbage was written into an inode table or indirect 724block, since this kind of filesystem corruption normally doesn't 725happen otherwise). 726 727On the Alpha, glibc declares st_flags although it isn't actually used; 728the configure script was improved to detect this case so that 729e2fsprogs can avoid using the non-functional stat field. 730 731The manual pages were updated to use a more consistent formatting 732style consistent with standard Unix man pages. Mke2fs's man page 733added documentation for a few previously undocumented options. 734 735Fixed minor display bugs in tune2fs and mke2fs. 736 737Programmer's notes: 738------------------- 739 740Improved portability of e2fsprogs to non-Unix systems (in particular, NT). 741 742Added features to parse and print feature strings into the e2p library. 743(e2p_feature2string, e2p_string2feature, e2p_edit_feature). 744 745ext2fs_mkdir() and ext2fs_new_dir_block() now creates directories 746whose directory entries contain proper filetype information if the 747filesystem supports it. 748 749ext2fs_link() now uses the low 3 bits of its flags parameter to pass 750the directory entry filetype information. This is used to set the 751directory entry filetype information if the filesystem supports it. 752 753Fixed a bug in ext2fs_expand_dir() where the block count in a 754directory's inode would not be properly incremented when the directory 755was expanded to the point where an indirect block needed to be 756allocated. 757 758 759E2fsprogs 1.15 (July 18, 1999) 760============================== 761 762Add configuration checks so that e2fsprogs will compile cleanly on 763Linux 2.3 kernels that have renamed i_version to i_generation. 764 765E2fsck now prints a progress/completion bar (and not just a simple 766spinner) if the -C0 option is requested or if it receives a SIGUSR1 767signal. Fsck will automatically manage the (potentially muliple) 768e2fsck processes to print completion bars if it is given a -C option, 769with the right thing happening if multiple filesystems are being 770checked in parallel. 771 772Mke2fs now has better automatic hueristics to determine the filesystem 773parameters to be used for a particular filesystem. Added a new option 774-T which allows the user to specify how the filesystem is to be used, 775which helps mke2fs do a better job selecting the filesystem parameters. 776 777Mke2fs now creates revision 1 filesystems by default, and with the 778sparse superblock feature enabled. The sparse superblock feature is 779not understood by Linux 2.0 kernels, so they will only allow read-only 780mounts of filesystems with this sparse superblocks. 781 782Fix bug where if /dev/null couldn't be opened (should never happen), 783e2fsck would hang in a tight loop. 784 785Make e2fsck handle the case where /lost+found isn't a directory. 786 787E2fsck now uses mallinfo if it exists to get accurate statistics about 788its memory usage. 789 790Fix bug in e2fsck where it wouldn't check to see if a disconnected 791inode had any problems before connecting it to /lost+found. 792 793Add check to e2fsck so it makes sure that total number of inodes in 794the filesystem is a sane number. 795 796Fix fencepost error when clearing an the end of the block bitmap which 797caused the last block in the bitmap not to get cleared. 798 799Cleaned up a number of messages in e2fsck: 800 * The message "Group's #'s copy of the group descriptor..." 801 was fixed so that the correct number would be displayed. 802 * Added missing space in the "disk write-protected" error messsage 803 * Cleaned up the error message printed when a non-interactive 804 e2fsck needs to abort a check because the filesystem 805 appears to be mounted. 806 807Added a new command-line utility, uuidgen, which will create and print 808a UUID. 809 810Make debugfs's icheck command more robust by checking to make sure an 811inode has valid blocks before interarting over the inode's blocks. 812 813UUID generation now uses a random-based scheme whenever possible to 814prevent potential privacy problems. 815 816Man pages for all of the UUID functions in the lirbary were added. 817 818Fixed bug in fsck so it won't coredump if a filesystem not in 819/etc/fstab is given to it. 820 821Fsck now understands the UUID=xxxx and LABEL=yyyy forms in /etc/fstab 822that most of the other mount utilities understands. 823 824Mke2fs will make a filesystem even if it appears mounted if the force 825option is given. 826 827Dumpe2fs has new command-line options which allow a filesystem expert 828to specify the superblock and blocksize when opening a filesystem. 829This is mainly useful when examining the remains of a toasted 830filesystem. 831 832The badblocks program has been updated to display correctly on disks 833with large block numbers. 834 835The badblocks program no longer gives spurious errors when errors 836occur on non-block boundaries, which is common if the blocksize is 837larger than 1k. 838 839Mke2fs will sync the disk device every MKE2FS_SYNC block groups if the 840MKE2FS_SYNC environment variable is set. This is to work around a VM 841bug in the 2.0 kernel. I've heard a report that a RAID user was able 842to trigger it even using a 2.2 kernel, but hopefully it will not be 843needed for most Linux 2.2 users. 844 845Fixed miscellaneous documentation and man pages. 846 847Programmer's notes: 848------------------- 849 850Cleaned up functions such as pass1_get_blocks, pass1_read_inode which 851in e2fsck's pass1.c really should have been static. 852 853The return value of the uuid_compare() function was changed to make it 854match with the convetions used by strcmp, memcmp, and Paul Leach's 855UUID sample document. 856 857The "make depend" process has now been made more automated; it now 858automatically word-wraps the dependencies, and only replaces source 859Makefile.in if there has been a change in the dependencies. Also, a 860top-level "make depend" now recurses through all the subdirectories 861automatically. 862 863The Makefile in .../util has been changed so that subst is built using 864the native C compiler during a cross-compilation, since the subst 865program is only used during the build process. Also add an explicit 866rule to build util/subst by cd'ing to the correct directory and 867running Makefile. 868 869The man directories are defined in terms mandir, so that the configure 870script can override the location of the manual pages. 871 872The config files have been updated to recognize new machine types for 873both the i386 and alpha families. 874 875Fsck has been modified so that it will accurately create an 876fsck_instance even when the noexecute flag is set. This allows for 877accurate debugging of the fsck pass structure. Also, when the verbose 878flag is given twice, fsck will print debugging information about when 879fsck is waiting for jobs to finish. 880 881 882E2fsprogs 1.14 (January 9, 1999) 883================================ 884 885Fix the fstab parsing code so that it can handle blank lines and 886comment characters. Also, missing pass numbers need to be treated as 887zero. 888 889Fixed a bug in e2fsck where under some circumstances (when e2fsck 890needs to restart processing after fixing an egregious inconsistency) 891it would try to access already freed memory. 892 893E2fsck now prints non-printable characters in directory entries and 894pathnames using '^' and 'M-' notation. 895 896Fixed chattr so that it will ignore symbolic links when doing 897recursive descent traversals. For both chattr and lsattr, no longer 898print the version string unless the -V option is given. 899 900Allow the system administrator to directly specify the number of 901inodes desired in the filesystem, for some special cases where this is 902necessary. 903 904Fix portability problems so that e2fsprogs can be compiled under Linux 9051.2 systems and Solaris systems. 906 907Update the config.guess file with a more recent version that will 908identify newer Linux platforms. 909 910Programmer's notes 911------------------ 912 913Ext2fs_read_inode and ext2fs_write_inode will now return an error if 914an inode number of zero is passed to them. 915 916E2fsprogs 1.13 (December 15, 1998) 917================================== 918 919Fixed a bug in debugfs where an error messages weren't getting printed 920when the ext2 library routines to read inodes returned errors in the 921stat, cmri and rm commands. 922 923Fixed a bug in mke2fs so that if a ridiculous inode ratio parameter is 924provided, it won't create an inode table smaller than the minimum 925number of inodes required for a proper ext2 filesystem. 926 927Fsck now parses the /etc/fstab file directly (instead of using 928getmntent()), so that it can distinguish between a missing pass number 929field and pass number field of zero. This caused problems for 930diskless workstations where all of the filesystems in /etc/fstab have 931an explicit pass number of zero, and fsck could not distinguish this 932from a /etc/fstab file with missing pass numbers. 933 934E2fsck will create a /lost+found directory if there isn't one in the 935filesystem, since it's safer to create the lost+found directory before 936it's needed. 937 938Fixed e2fsck so that it would detect bogus immutable inodes which 939happen to be sockets and FIFO files, and offer to clear them. 940 941If a filesystem has multiple reasons why it needs to be checked, and 942one of the reasons is that it is uncleanly mounted, e2fsck will print 943that as the reason why the filesystem is being checked. 944 945Cleaned up the output routines of mke2fs so that it doesn't overflow 946an 80 column display when formating really big filesystems. 947 948Added a sanity check to e2fsck to make sure that file descriptors 0, 9491, 2 are open before opening the hard disk. This avoids a problem 950where a broken program might exec e2fsck with those file descriptors 951closed, which would cause disastrous results if the kernel returns a 952file descriptor for the block device which is also used by FILE * 953stdout. 954 955Fixed up the e2fsck progress reporting functions so that the values 956reliably reach 100% at the completion of all of the e2fsck passes. 957 958Fixed minor documentation bugs in man pages and usage messages. 959 960Programmer's notes: 961------------------- 962 963Fixed a number of lint warnings in the ext2fs library and potential 964portability problems from other OS's header files that might define 965CPP macros for names like "max" and "min". 966 967ext2fs_badblocks_list_add() has been made more efficient when it needs 968to grow the bad blocks list. 969 970Fixed a bug in e2fsck which caused it to dereference a freed pointer 971just before exiting. 972 973Fixed the substition process for generating the mk_cmds and compile_et 974scripts so that they will work outside of the build tree. 975 976Add sanity check to e2fsck so that if an internal routine 977(ext2fs_get_dir_info) returns NULL, avoid dereferencing the pointer 978and causing a core dump. This should never happen, but... 979 980E2fsprogs 1.12 (July 9, 1998) 981================================== 982 983E2fsprogs now works with glibc (at least with the version shipped wtih 984RedHat 5.0). The ext2fs_llseek() function should now work even with 985i386 ELF shared libraries and if llseek() is not present. We also 986explicitly do a configure test to see if (a) llseek is in libc, and 987(b) if llseek is declared in the system header files. (See standard 988complaints about libc developers don't understand the concept of 989compatibility with previous versions of libc.) 990 991The ext2fs library now writes out the block group number in each of 992the superblock copies. This makes it easier to automatically 993determine the starting block group of the filesystem when the block 994group information is trashed. 995 996Added support for the EXT2_FEATURE_INCOMPAT_FILETYPE feature, 997which means that e2fsprogs will ignore the high 8 bits of the 998directory entry's name_len field, so that it can be used for other 999purposes. 1000 1001Added support for the EXT2_FEATURE_RO_COMPAT_LARGE_FILE feature. 1002E2fsprogs will now support filesystems with 64-bit sized files. 1003 1004Added support for the EXT2_FEATURE_COMPAT_DIR_PREALLOC feature. 1005 1006Added new program "e2label", contributed by Andries Brouwer. E2label 1007provides an easy-to-use interface to modify the filesystem label. 1008 1009Fixed bug so that lsattr -v works instead of producing a core dump. 1010 1011Fixed a minor bug in mke2fs so that all groups with bad superblock 1012backup blocks are printed (not just the first one). 1013 1014Mke2fs will check the size of the device, and if the user specifies a 1015filesystem size larger than the apparent size of the device it will 1016print a warning message and ask if the user wants to proceed. 1017 1018E2fsck has a new option -C, which sends completion information to the 1019specified file descriptor. For the most part, this is intended for 1020programs to use, although -C 0 will print a spinning character to the 1021stdout device, which may be useful for users who want to see something 1022happening while e2fsck goes about its business. 1023 1024Fixed a bug in e2fsck which could cause a core dump when it needs to 1025expand the /lost+found directory, and sometimes the bitmaps haven't 1026been merged in. Also fixed a related bug where ext2fs_write_dir_block 1027was used to write out a non-directory block. (Which would be bad on a 1028non-Intel platform with byte swapping going on.) 1029 1030Fixed bug in e2fsck where it would print a "programming error" message 1031instead of correctly identifying where a bad block was in used when 1032the bad block was in a non-primary superblock or block group 1033descriptor. Also fixed a related bug when sparse superblocks are in 1034use and there is a bad block where a superblock or block group 1035descriptor would have been in a group that doesn't include a 1036superblock. 1037 1038Fixed a bug in e2fsck (really in libext2fs's dblist function) where if 1039the block group descriptor table is corrupt, it was possible to try to 1040allocate a huge array, fail, and then abort e2fsck. 1041ext2fs_get_num_dirs() now sanity checks the block group descriptor, 1042and subsitutes reasonable values if the descriptors are obviously bogus. 1043 1044If e2fsck finds a device file which has the immutable flag set and the 1045i_blocks beyond the normal device number are non-zero, e2fsck will 1046offer to remove it, since it's probably caused by garbage in the inode 1047table. 1048 1049When opening a filesystem, e2fsck specially checks for the EROFS error 1050code, and prints a specific error message to the user which is more 1051user friendly. 1052 1053If the filesystem revision is too high, change the e2fsck to print 1054that this is either because e2fsck is out of date, or because the 1055superblock is corrupt. 1056 1057E2fsck now checks for directories that have duplicate '.' and '..' 1058entries, and fixes this corruption. 1059 1060E2fsck no longer forces a sync of the filesystem (with attendant sleep 1061calls) at all times. The ext2fs_flush() function now performs a sync 1062only if it needed to write data blocks to disk. 1063 1064Fixed a minor bug in e2fsck's pass1b's file cloning function, where 1065certain errors would not be properly reported. 1066 1067Updated and expanded a few points in the man pages which users 1068complained wheren't explicit enough. 1069 1070Added special case byte-swapping code if compiling on the PowerPC, to 1071accomodate the strange big-endian variant of the ext2 filesystem that 1072was previously used on the PowerPC port. 1073 1074 1075Programmer's notes: 1076------------------- 1077 1078Removed C++ keywords from the ext2fs libraries so that it could be 1079compiled with C++. 1080 1081E2fsck's internal organization has now been massively reorganized so 1082that pass*.c don't have any printf statements. Instead, all problems 1083are reported through the fix_problem() abstraction interface. E2fsck 1084has also been revamped so that it can be called as a library from a 1085application. 1086 1087Added new fileio primitives in libext2fs for reading and writing 1088files on an unmounted ext2 filesystem. This interface is now used by 1089debugfs. 1090 1091Added a new libext2fs function for mapping logical block numbers of 1092a file to a physical block number. 1093 1094Added a new libext2fs function, ext2fs_alloc_block(), which allocates 1095a block, zeros it, and updates the filesystem accounting records 1096appropriately. 1097 1098Added a new libext2fs function, ext2fs_set_bitmap_padding(), which 1099sets the padding of the bitmap to be all one's. Used by e2fsck pass 5. 1100 1101The libext2fs functions now use a set of memory allocation wrapper 1102functions: ext2fs_get_mem, ext2fs_free_mem, and ext2fs_resize_mem, 1103instead of malloc, free, and resize. This makes it easier for us to 1104be ported to strange environments where malloc, et. al. aren't 1105necessarily available. 1106 1107Change the libext2fs fucntion to return ext2-specific error codes 1108(EXT2_DIR_EXISTS and EXT2_DB_NOT_FOUND, for example) instead of using 1109and depending on the existence of system error codes (such as EEXIST 1110and ENOENT). 1111 1112Renamed io.h to ext2_io.h to avoid collision with other OS's header 1113files. 1114 1115Add protection against ext2_io.h and ext2fs.h being included multiple 1116times. 1117 1118The types used for memory lengths, etc. have been made more portable. 1119In generla, the code has been made 16-bit safe. Added Mark 1120Habersack's contributed DOS disk i/o routines. 1121 1122Miscellaneous portability fixes, including not depending on char's 1123being signed. 1124 1125The io_channel structure has a new element, app_data, which is 1126initialized by the ext2fs routines to contain a copy of the filesystem 1127handle. 1128 1129ext2fs_check_directory()'s callback function may now return the error 1130EXT2_ET_CALLBACK_NOTHANDLED if it wishes ext2fs_check_directory() to 1131really do the checking, despite the presence of the callback function. 1132 1133 1134E2fsprosg 1.11 (June 17, 1997) 1135============================== 1136 1137Fixed e2fsck to detect (previously ignored) conflicts between the 1138superblock or block group descriptors and block bitmaps, inode 1139bitmaps, and inode tables. 1140 1141Fixed bug in e2fsck so that when the message printed out when a block 1142or inode bitmap conflicts with other data, it has the correct group 1143number. 1144 1145Fixed bug in e2fsck and mke2fs where the blocksize wasn't being passed 1146to badblocks. This meant that not all of the filesystem was being 1147tested for bad blocks! 1148 1149Fixed an array boundary overrun case which cropped up in 1150ext2fs_badblocks_list_test when a user tried running "mke2fs -c 1151-b 4096". 1152 1153Adjusted the number of columns printed by mke2fs when displaying the 1154superblock backups to avoid running over 80 columns when making a 1155really big filesystem. 1156 1157Fixed up the man pages for e2fsck, debugfs, badblocks, chattr, 1158dumpe2fs, fsck, mke2fs, and tune2fs (typos and other minor grammar 1159fixes), thanks to some suggestions from Bill Hawes (whawes@star.net). 1160 1161Programmer's notes: 1162------------------- 1163 1164Fixed install rule in lib/ss so that ss_err.h is actually getting 1165installed. 1166 1167Fixed bug in ext2fs_copy_bitmap; the destination bitmap wasn't getting 1168bassed back to the caller. 1169 1170Fixed bug in ext2fs_inode_scan_goto_blockgroup; it had not been 1171setting the current inode number (which meant this function wasn't 1172working at all). 1173 1174Fixed bug in ext2fs_resize_generic_bitmap; it had not be zeroing all 1175blocks in the bitmap when increasing the size of the bitmap. 1176 1177Changed the initial number of blocks allocated by ext2fs_init_dblist() 1178to be more realistic. 1179 1180Added a new function ext2fs_allocate_group_table, which sets up the 1181group descriptor information (and allocates inode and block bitmaps, 1182and inode tables for a particular group). The function was created by 1183factoring out code form ext2fs_allocate_tables(). 1184 1185Added a new function ext2fs_move_blocks which takes a bitmap of the 1186blocks to be moved, and moves them to another location on the 1187boardboard. 1188 1189Make the unix_io channel's io_channel_flush implementation calls sync() 1190to to flush the kernel buffers to disk. 1191 1192Added a new function ext2fs_dblist_count returns the number of 1193directory blocks in dblist. 1194 1195 1196E2fsprogs 1.10 (April 24, 1997) 1197=============================== 1198 1199Mke2fs once again defaults to creating revision #0 filesystems, since 1200people were complaining about breaking compatibility with 1.2 kernels. 1201Warning messages were added to the mke2fs and tune2fs man pages that 1202the sparse superblock option isn't supported by most kernels yet (1.2 1203and 2.0 both don't support parse superblocks.) 1204 1205Added new flag to mke2fs, -R <raid options>, which allows the user to 1206tell mke2fs about the RAID configuration of the filesystem. Currently 1207the only supported raid option is "stride" which specifies the width 1208of the RAID stripe. 1209 1210Fixed bug in e2fsck where pass1b would bomb out if there were any 1211blocks marked bad in the inode table. 1212 1213Fixed rare bug in mke2fs where if the user had a very unlucky number 1214of blocks in a filesystem (probability less than .002) the resulting 1215filesystem would be corrupt in the last block group. 1216 1217Fixed bug where if e2fsck tried to allocate a block to fix a 1218filesystem corruption problem and the filesystem had no free blocks, 1219ext2fs_new_block() would loop forever. 1220 1221The configure script now checks explicitly to see if "-static" works, 1222since that can't be assumed to be true --- RedHat doesn't install 1223libc-static by default. 1224 1225Fixed bug in libext2's block iterator functions where under some 1226cirmcustances, file with holes would cause the bcount parameter to the 1227callback function to be incorrect. This bug didn't affect any of 1228e2fsprogs programs, but it was discovered by Paul Mackerras, the 1229author of the PPC boot loader. 1230 1231Removed use of static variables to store the inode cache in libext2fs. 1232This caused problems if more than one filesystem was accessed via 1233libext2fs (static variables in libraries are generally a bad idea). 1234Again, this didn't affect e2fsprogs programs, but it was discovered by 1235Paul Mackerras. 1236 1237Fixed minor bugs and version code drift to assure that e2fsprogs 1.10 1238will compile cleanly with 1.2.13 kernels (even with a.out shared 1239libraries!) 1240 1241Programmer's notes: 1242------------------- 1243 1244Added new functions to duplicate an ext2 filesystem handle, and its 1245associated substructure. New functions: ext2fs_dup_handle(), 1246ext2fs_copy_dblist(), ext2fs_badblocks_copy(), ext2fs_copy_bitmap(). 1247Other structures, such as the io_channel and the inode_cache, now have 1248a ref count so that they only get freed when they are no longer used 1249by any filesystem handle. (These functions were added as part of the 1250development effort for an ext2 resizer). 1251 1252E2fsprogs 1.09 (April 14, 1997) 1253=============================== 1254 1255Fixed bug in mke2fs (really in lib/ext2fs/initialize.c) which was 1256accidentally introduced in the 1.08 release. The overhead calculation 1257was accidentally removed, which caused ext2fs_initialize() to not 1258notice when the filesystem size needed to be adjusted down because 1259there wasn't enough space in the last block group. 1260 1261Fixed bug in version parsing library routine; it was always parsing 1262the library version string, instead of using the passed-in string. 1263 1264Clarified chattr man page. 1265 1266E2fsprogs 1.08 (April 10, 1997) 1267=============================== 1268 1269E2fsck 1.07 was very slow when checking very large filesystems with a 1270lot of files that had hard links (i.e., news spools). This was fixed 1271by seriously revamping the icount abstraction. Added a formal test 1272suite for the icount abstraction. 1273 1274Debugfs now has a "-l" option to the "ls" command, which lists the 1275inode number, permissions, owner, group, size, and name of the files 1276in the directory. 1277 1278Fix a bug in e2fsck where when a directory had its blocks moved to 1279another location during the pass 1b processing, the directory block 1280list wasn't updated, so pass 2 wouldn't check (and correct) the 1281correct directory block. 1282 1283E2fsck will now treat inodes which contain blocks which are claimed by 1284the filesystem metadata by treating them as multiply claimed blocks. 1285This way, the data in those blocks can be copied to a new block during 1286the pass 1b--1d processing. 1287 1288E2fsck will attempt to determine the correct superblock number and 1289display it in the diagnostic and warning messages if possible. 1290 1291Add support for a new (incompatible) feature, "sparse_super". This 1292feature reduces the number of blocks which contain copies of backup 1293superblocks and block group descriptors. (It is only an incompatible 1294feature because of a bug in ext2_free_blocks.) mke2fs and tune2fs now 1295support a new -s option; e2fsck will recognize filesystems built with 1296this feature turned on. 1297 1298E2fsck now checks the library to make sure is the correct version, 1299using new library functions. (This helps to diagnose incorrectly 1300installed e2fsprogs distributions.) 1301 1302Dumpe2fs now prints more information; its now prints the the 1303filesystem revision number, the filesystem sparse_super feature (if 1304present), the block ranges for each block group, and the offset from 1305the beginning of the block group. 1306 1307Mke2fs now distributes the inode and block bitmap blok so that the 1308won't be concentrated in one or two disks in RAID/striping setups. 1309Also, if the user chooses a 2k or 4k block group, mke2fs will try to 1310choose the largest blocks per group that be chosen. (For 2k blocks, 1311you can have up to 16384 blocks/group; for 4k blocks, you can have up 1312to 32768 blocks/group.) Previously mke2fs would not allow 1313specification of more than 8192 blocks per group, even if you were 1314using a 2k or 4k block group. 1315 1316Programmer's notes: 1317------------------- 1318 1319Added a new function ext2fs_create_icount2() which takes a "hint" 1320argument. This hint argument presets the icount array with the list 1321of inodes which actually need to be in the icount array. This really 1322helps to speed up e2fsck. 1323 1324Added a new function ext2fs_icount_validate() which checks the rep 1325invariant for the icount structure. This is used mostly for testing. 1326 1327The error mesasage given when a bad inode number is passed to 1328test_generic_bitmap to reflect EXT2FS_TEST_ERROR (instead of 1329EXT2FS_UNMARK_ERROR). 1330 1331Added a new function ext2fs_set_dir_block which sets the block of a 1332dblist entry, given the directory inode and blockcnt. 1333 1334Added a new function ext2fs_get_library_version() which returns the 1335current library version, and ext2fs_parse_version_string() which 1336returns a version number based on a e2fsprogs version string. 1337 1338The icount functions will return EINVAL if the passed in inode number 1339is out of bounds. 1340 1341E2fsprogs 1.07 (March 9, 1997) 1342============================== 1343 1344E2fsck is now uses much less memory when checking really large 1345filesystems (or rather, filesystems with a large number of inodes). 1346Previously a filesystem with 1 million inodes required 4 megabytes of 1347memory to store inode count statistics; that storage requirement has 1348now been reduced to roughly half a megabyte. 1349 1350E2fsck can now properly deal with bad blocks appearing inside the 1351inode table. Instead of trying to relocate the inode table (which 1352often failed because there wasn't enough space), the inodes in the bad 1353block are marked as in use. 1354 1355E2fsck will automatically try to use the backup superblocks if the 1356primary superblocks have a bad magic number or have missing meta-data 1357blocks (or meta-data blocks which are out of range). 1358 1359E2fsck's pass 3 has been made more efficient; most noticeable on 1360filesystems with a very large number of directories. 1361 1362Completely revamped e2fsck's system of printing problem reports. It 1363is now table driven, to make them more easily customizeable and 1364extendable. Error messages which can be printed out during preen mode 1365are now one line long. 1366 1367Fixed e2fsck's filesystem swapping code so that it won't try to swap 1368fast symbolic links or deleted files. 1369 1370Fixed e2fsck core dumping when fixing a filesystem which has no 1371directories (not even a root directory). 1372 1373Added a check to e2fsck to make sure that the length of every 1374directory entry is a multiple of 4 (since the kernel complains if it 1375isn't). 1376 1377Added a check to e2fsck to make sure that a directory entry isn't a 1378link to the root directory, since that isn't allowed. 1379 1380Added a check to e2fsk to now make sure the '.' and '..' directory 1381entries are null terminated, since the 2.0 kernel requires it. 1382 1383Added check to write_bitmaps() to make sure the superblock doesn't get 1384trashed if the inode or block bitmap is marked as being block zero. 1385 1386Added checking of the new feature set fields in the superblock, to 1387avoid dealing with new filesystem features that this package wasn't 1388set up to handle. 1389 1390Fixed a fencepost error in ext2fs_new_block() which would occasionally 1391try to allocate a block beyond the end of a filesystem. 1392 1393When the UUID library picks a random IEEE 802 address (because it 1394can't find one from a network card), it sets the multicast bit, to 1395avoid conflicting with a legitimate IEEE 802 address. 1396 1397Mke2fs now sets the root directory's owner to be the real uid of the 1398user running mke2fs. If the real uid is non-zero, it also sets 1399the group ownership of the root directory to be the real group-id of 1400the user running mke2fs. 1401 1402Mke2fs now has more intelligent error checking when it is given a 1403non-existent device. 1404 1405When badblocks is given the -vv option, it now updates the block that 1406it is currently testing on every block. 1407 1408Fixed a bug in fsck where it wouldn't modify the PATH envirnoment 1409currently correctly if PATH wasn't already set. 1410 1411Shared libraries now built with dependencies. This allows the shared 1412library files to be used with dlopen(); it also makes the transition 1413to libc 6 easier, since ld.so can tell which libc a particular shared 1414library expects to use. 1415 1416Programmer's notes: 1417------------------- 1418 1419Added new abstraction (defined in dblist.c) for maintaining a list of 1420blocks which belongs to directories. This is used in e2fsck and other 1421programs which need to iterate over all directories. 1422 1423Added new functions which test to see if a contiguous range of blocks 1424(or inodes) are available. (ext2fs_*_bitmap_range). 1425 1426Added new function (ext2_inode_has_valid_blocks) which returns true if 1427an inode has valid blocks. (moved from e2fsck code). 1428 1429Added new function (ext2fs_allocate_tables) which allocates the 1430meta-data blocks as part of initializing a filesystem. (moved from 1431mke2fs code). 1432 1433Added a new I/O manager for testing purposes. It will either allow a 1434program to intercept I/O requests, or print debugging messages to 1435trace the activity of a program using the I/O manager. 1436 1437The badblocks_list functions now store the bad blocks in a sorted 1438order, and use a binary search to speed up badblocks_list_test. 1439 1440The inode scan function ext2fs_get_next_inode() may now return a soft 1441error returns: MISSING_INODE_TABLE and BAD_BLOCK_IN_INODE_TABLE in 1442those cases where part of an inode table is missing or there is a bad 1443block in the inode table. 1444 1445Added a new function (ext2fs_block_iterate2) which adds new arguments to 1446the callback function to return a pointer (block and offset) to the 1447reference of the block. 1448 1449Added new function (ext2fs_inode_scan_goto_blockgroup) which allows an 1450application to jump to a particular block group while doing an inode 1451scan. 1452 1453The badblocks list functions were renamed from badblocks_* to 1454ext2fs_badblocks_*. Backwards compatibility functions are available 1455for now, but programs should be modified to use the new interface. 1456 1457Some of the library functions were reorganized into separate files to 1458reduce the size of some programs which statically link against the 1459ext2 library. 1460 1461Put in some miscellaneous fixes for the Alpha platform. 1462 1463 1464E2fsprogs 1.06 (October 7, 1996) 1465================================ 1466 1467Fixed serious bug in e2fsck: if the block descriptors are bad, don't 1468smash the backup copies in ext2fs_close(). (The problem was that when 1469e2fsck -p discovered the problem, while it was closing the filesystem 1470and exiting, it was also blowing away the backup superblocks on the 1471disk, which was less than friendly.) We now make it the case that we 1472only write out the backup superblock and the back block descriptors if 1473the filesystem is completely free from problems. 1474 1475Fixed a bug in block_interate in the lib/ext2fs library which caused 1476e2fsck to fail on GNU Hurd-created filesystems. 1477 1478Add support for Linux/FT's bootloader, which actually uses 1479EXT2_BOOT_LOADER, and sets its mode bits which caused e2fsck to want 1480to clear the inode. 1481 1482Add support for the "A" (no atime update) attribute. (Note: this 1483attribute is not yet in production kernels.) 1484 1485The test suite is not automatically run when doing a "make all" from 1486the top level directory. Users should manually run "make check" if 1487they wish to run the test suite. 1488 1489Upon a preenhalt(), make the printed message more explicit that 1490running e2fsck "MANAULLY" means without the -p or -a options. 1491 1492In e2fsck, if a disconnected inode is zero-length, offer to clear it 1493instead of offering to connect it to lost+found. 1494 1495In e2fsck, if a filesystem was just unmounted uncleanly, and needs 1496e2fsck to be run over it, change e2fsck to explicitly display this 1497fact. 1498 1499For dumpe2fs and e2fsck, cause the -V option to print out which 1500version of the ext2fs library is actually getting used. (This will 1501help detect mismatches of using a 1.06 utility with a 1.05 library, 1502etc.) 1503 1504Programmers' notes: 1505------------------- 1506 1507EXT2_SWAP_BYTES was changed to EXT2_FLAG_SWAP_BYTES, which better fits 1508the naming convention. 1509 1510In ext2fs_initialize(), make sure the description for the inode bitmap 1511is correctly initialize. 1512 1513Fixed minor type typo in ext2fs_allocate_generic_bitmap(); 1514 1515E2fsprogs 1.05 (September 7, 1996) 1516================================== 1517 1518Add support for new fields in the ext2 superblock --- volume name, 1519volume UUID, and last mounted field. Dumpe2fs displays these fields, 1520tune2fs and mke2fs allows you to set them. E2fsck will automatically 1521generate a UUID for those volumes that don't have them. 1522 1523Put in support for e2fsck to recognize HURD specific ext2 features --- 1524most notably, the translator block. The e2fsprogs tools will now use 1525the creator_os field in the superblock to correctly handle different 1526OS-specific variants of the ext2 filesystem. 1527 1528E2fsck now fixes inodes which have a the deletion time set, but which 1529have a non-zero i_link_count field by offering to clear the deletion 1530time. Previously e2fsck assumed that the inode was deleted (per 0.3c 1531ext2 kernel behavior) and offered to unlink the file. 1532 1533If e2fsck sets the clean bit, but nothing else, set the exit code 1534FSCK_NONDESTRUCT. After all, e2fsck did fix a filesystem error --- it 1535set the filesystem valid bit when it was previously cleared. :-) This 1536was needed to make the HURD fsck driver happy. 1537 1538If the user refuses to attach an unattached inode, e2fsck will no 1539longer set the inode's link count. Otherwise, the inode would end up 1540getting marked as unused, which might cause loss of data later. 1541 1542Make the message issued by e2fsck when the superblock is corrupt less 1543confusing for users. It now mentions that another reason for the 1544"corrupt superblock" message might be that the partition might not be 1545an ext2 filesystem at all (it might swap, msdos filesystem, ufs, etc.) 1546 1547Make the libext2 library more robuest so that e2fsck won't coredump on 1548an illegal superblock where the blocksize is zero. (f_crashdisk is 1549the test case). 1550 1551By default, create filesystems where the default checkinterval is 6 1552months (180 days). Linux servers can be robust enough that 20 reboots 1553can be a long, long time. 1554 1555Added configure flag --enable-old-bitops, which forces the bitops to 1556use the old (native) bitmask operations. By default on the sparc 1557platform, the standard ext2 bit ordering is now used. 1558 1559Added a new feature to e2fsck to byte-swap filesystems; this can be 1560used to convert old m68k filesystems to use the standard byte-order 1561storage for the superblock, inodes, and directory blocks. This 1562function is invoked by using the '-s' option to e2fsck. 1563 1564Debugfs's "dump" command has been enhanced so that it writes out the 1565exact size of the file so that the nulls at the end of the file are 1566eliminated. The command also accept a new "-p" option which will 1567attempt preserve to preserve the ownernship, permissions, and 1568file modification/access times. 1569 1570Debugfs has two new options, -f and -R. The -R option allows the user 1571to execute a single debugfs command from the command line. The -f 1572option allows the user to specify a "command file" containing debugfs 1573commands which will get executed. 1574 1575Dumpe2fs now pretty prints the check interval, instead of just 1576printing the check interval as a number of seconds. 1577 1578Fix bugs in debugfs: the params command when no filesystem is opened 1579no longer causes a core dump. It is now possible to unlink a file 1580when a pathame containing a '/' is specified. 1581 1582Tune2fs has a new -C option which sets the number of times the 1583filesystem has been mounted. 1584 1585Fix the chattr '-v' option so that it actually works. Chattr was 1586being buggy about the -v option parsing. 1587 1588Programmers' notes: 1589------------------- 1590 1591The directory lib/uuid contains a set of library routines to generate 1592DCE compatible UUIDs. 1593 1594Extended ext2fs_namei() to handle symbolic links. Added new function 1595ext2fs_nami_follow() which will follow last symbolic link in the case 1596where the pathname points to a sym link. 1597 1598The ext2fs_block_iterate function will now return the HURD translator 1599block, if present. The new flag BLOCK_FLAG_DATA_ONLY will cause the 1600iterator to return data blocks only. The ext2fs.h file now defines 1601constants BLOCK_COUNT_IND, BLOCK_COUNT_DIND, BLOCK_COUNT_TIND, and 1602BLOCK_COUNT_TRANSLATOR, which are the magic values passed in the block 1603count field of the iterator callback function. 1604 1605The test script driver now takes an optional second argument, which is 1606the test case to be run. This allows you to run a test case without 1607needing to run the entire test suite. 1608 1609On Linux ELF systems, install the .so files in the correct places 1610(/usr/lib). The .so files must be stored in the same directory as the 1611.a files. 1612 1613Fixed miscellaneous HURD compilation issues with header file being 1614included in the right order. 1615 1616Fixed debugfs so that it resets optind to zero, not one, since setting 1617optind to zero is more correct. 1618 1619 1620E2fsprogs 1.04 (May 16, 1996) 1621============================= 1622 1623First "official" (1.03 was a limited release only) to support building 1624e2fsprogs under Linux 2.0 kernels (as well as late model 1.3 and 1.99 1625kernels). 1626 1627This package includes a RPM specs file, that it can be built using the 1628RedHat Package Manager. 1629 1630E2fsck now prints a hint that if there are lots of incorrectly located 1631inode bitmaps, block bitmaps, and inode table blocks, the user might 1632want to try using e2fsck -b 8193 first, to see if that fares any 1633better. 1634 1635For ext2 filesystem written with the hurd, debugfs will now print out 1636the translator field when printing an inode structure. 1637 1638Lots of miscellaneous linking/installation cleanups: 1639 1640 Libraries are now linked using a relative pathname, instead of 1641 relying on -L working correct. It doesn't, in many cases, including 1642 current versions of GNU ld. This guarantees that the build tree is 1643 linking with the right libraries, instead of the ones installed in 1644 /usr/lib. 1645 1646 Header files, man pages, and the et/ss shell scripts are now 1647 generated using a custom substitution script, instead of relying on 1648 the configure script. This prevents needless recompilation of 1649 files; in addition, the custom substitution script is much faster. 1650 1651 e2fsck may now be linked dynamically, by using the 1652 --enable-dynamic-e2fsck flag to configure. This is not recommended, 1653 since it increases e2fsck's dependence on other files, but some 1654 people need to save disk space, and other critical programs on their 1655 systems were being linked dynamically anyway. 1656 1657 Programs such as fsck which didn't need to be linked against 1658 libext2fs (or mke2fs which didn't need to be linked against libe2p) 1659 only link against libraries they actually need. Otherwise, those 1660 programs would require the presense of libraries that otherwise 1661 could be removed from a rescuse diskette. 1662 1663 The ss include files are now installed correctly so they can 1664 actually be used by another package. 1665 1666 If the profiling libraries are built, they are now installed on a 1667 "make install-libs". 1668 1669 1670E2fsprogs 1.03 (March 27, 1996) 1671=============================== 1672 1673Change the m68k bit numbering for bitmasks to match the bit numbering 1674used by all other ext2 implementations. (This change was requested by 1675the m68k kernel development team.) 1676 1677Support (in-development) filesystem format revision which supports 1678(among other things) dynamically sized inodes. 1679 1680Fixed a bug in the ext2 library so that an intelligent error is 1681returned if mke2fs is run with a ridiculously small number of blocks 1682for a partition. 1683 1684Fixed a bug in the ext2 library which required that the device be 1685openable in read/write mode in order to determine its size. This 1686caused e2fsck -n to require read/write access when it was not 1687previously necessary. 1688 1689Fixed a bug in e2fsck which casued it to occasionally fail the test 1690suite depending on which version of the floating point library it was 1691using. 1692 1693Fixed a bug in e2fsck so that it now halts with a fatal error when 1694certain superblock consistency checks fail. Previously it continued 1695running e2fsck, with some potential confusing/damaging consequences. 1696 1697Added new flag to fsck which allows the root to be checked in parallel 1698with other filesytems. This is not the safest thing in the world to 1699do, but some system administrators really wanted it. 1700 1701Fixed -Wall flames in lib/ss. 1702 1703 1704E2fsprogs 1.02 (January 16, 1996) 1705================================= 1706 1707Fix to allow e2fsprogs to be compiled on recent 1.3 (pl45+) kernels. 1708 1709Change e2fsck to print statistics of how many non-contiguous files are 1710on the system. Note that a file which is larger than 8k blocks, it is 1711guaranteed to be non-contiguous. 1712 1713In mke2fs, print a warning message if a user tries to format a whole 1714disk (/dev/hda versus /dev/hda1). If a user really wants to format a 1715whole disk, the -F (force) option forces mke2fs to format a whole disk 1716as a filesytem. 1717 1718Fix a bug in fsck where in some cases it might start checking 1719partitions in the next pass before it finishes checking partitions in 1720the current pass. This still won't cause two partitions on the same 1721disk will be checked, so it's rarely a problem in real life. 1722 1723Patch lsattr so that it won't hang when checking a named pipe. 1724 1725Minor compilation fixes: 1726 * Fix the order of libraries that were linked in debugfs. 1727 * Allow the sources to be compiled with -ansi turned on. 1728