RELEASE-NOTES revision 11e5292266b26c63c6ffd25b799d271088200256
1E2fsprogs 1.42.6 (September 21, 2012) 2===================================== 3 4When mke2fs creates file systems with lazy itable initialization, the 5progress updates for writing the inode table happens so quickly that 6on a serial console, the time to write the progress updates can be the 7bottleneck. So mke2fs will now only update its progress indicators 8once a second. 9 10Resize2fs will skip initializing the inode tables if the kernel 11supports lazy_itable_init, which speeds up growing off-line growth of 12uninit_bg file systems significantly. Resize2fs will now also 13correctly set the itable_unused field in the block group descriptor to 14speed up the first e2fsck after the file system is grown. 15 16Resize2fs has been fixed so that on-line resizing of meta_bg file 17systems work correctly. This is needed to grow file system which are 18larger than 16T. 19 20Resize2fs will now correctly handle resizing file systems to 16TB on 2132-bit file systems when "16TB" is specified on the command line. 22 23Fix mke2fs so that it will be careful to set the reserved blocks ratio 24larger than 50%; this can happen when creating small file systems and 25when the last block group is dropped because there are not enough 26blocks to support the metadata blocks in the last block group. 27 28Fixed spelling mistake in debugfs's help message. 29 30Fixed a potential seg fault in e2fsck when there is an I/O error while 31reading the superblock. 32 33Fixed various Debian Packaging Issues (Addresses Debian Bug #677497) 34 35Updated/fixed various man pages. (Addresses Sourceforge Bug: 36#3559210) 37 38Programmer's Notes 39------------------ 40 41The configure option --enable-relative-symlink was broken so that it 42needed to be --enable-symlink-relative-symlinks. We will support both 43for at least two years, but then the worng configure option will be 44removed. 45 46Fixed a regression introduced in 1.42.5 so the link order for 47e2fsprogs' libraries will be correct for both static and shared 48linking. (Addresses Sourceforge Bug: #3554345) 49 50Add support for e2fsprogs to be compiled using clang/LLVM. 51 52Fix portability problems on non-Linux systems: avoid compile failures 53on systems that don't have malloc.h 54 55Fix f_mmp regression test suite so that debugfs gets killed if the 56test is interrupted. 57 58 59 60E2fsprogs 1.42.5 (July 29, 2012) 61================================ 62 63Fixed a bug with mke2fs where if there is only 8 inodes per block 64group, the calculation of the number of uninitialized inodes in the 65first block group would go negative. This resulted in "mke2fs -N 256 66-t ext4 /tmp/foo.img 256m" trying to write so many blocks that /tmp 67would run out of space. (Addresses Sourceforge Bug: #3528892) 68 69Fixed a bug in how e2fsck would uniquify directory entry names. 70(AddressesSourceforge Bug: #3540545) 71 72Previously, e2fsck would only allow a mounted file system to be 73checked if it was the root file system and it was mounted read-only. 74Now it will allow any file system mounted read-only to be checked if 75the -f option is specified. This makes it easier to test how e2fsck 76handles checking file systems which are mounted without having to test 77on the root file system. 78 79Fixed a problem if e2fsck where if the root file system is mounted 80read-only, e2fsck would not clear an error indication in the journal 81superblock. Combined with a kernel bug, this would cause the e2fsck 82to check the file system after every single boot. 83 84The e4defrag program can now handle device symlinks, such as 85/dev/mapper/testvg-testlv, instead of insisting on a less 86human-friendly name such as /dev/dm-2. (Addresses Red Hat Bugzilla: 87#707209) 88 89Fixed filefrag so it will not crash with a segfault on files from a 90virtual file system such as /proc. (e.g., "filefrag 91/proc/partitions") 92 93Fixed filefrag so that it correctly reports the number of extents. 94(Addresses Red Hat Bugzilla: #840848) 95 96Fixed a file descriptor leak in logsave which could cause it to hang. 97(Addresses Debian Bug: #682592) 98 99Fixed e2fsck so that the file system is marked as containing an error 100if the user chooses not to fix the quota usage information. 101 102Fixed tune2fs so that it correctly removes the quota feature when the 103last quota inode is removed. 104 105Fix tune2fs so that after removing a quota inode, the block bitmap is 106updated; otherwise, e2fsck would complain after running 'tune2fs -O 107^quota <dev>'. 108 109Fix tune2fs so that when converting a file system from using legacy 110quota files to the new quota file system feature with hidden quota 111files, the accounting for these files is handled correctly so that 112e2fsck doesn't complain. 113 114Improved e2fsck's verbose reporting statistics, and allow the more 115verbose reporting to be enabled via /etc/e2fsck.conf. 116 117Fixed various Debian Packaging Issues (Addresses Debian Bug #678395) 118 119Updated/fixed various man pages. (Addresses Debian Bugs: #680114) 120 121 122Programmer's Notes 123------------------ 124 125Fixed portability problems on other operating systems (e.g., Hurd and 126FreeBsd) caused by the attempted inclusion of <sys/quota.h>. 127 128Make sure that shared libraries link with the shared libraries built 129in the build tree, instead of the system provided libraries. 130Previously, libraries and executables were linked with the system 131libraries if present, and possibly using static archives instead of 132shared libraries. This was also problematic since if libext2fs.so is 133linked with a static libcom_err.a from system, the build system would 134attempt to link without -lpthread. (Addresses Sourceforge Bug: 135#3542572) 136 137 138E2fsprogs 1.42.4 (June 12, 2012) 139================================ 140 141Fixed more 64-bit block number bugs (which could end up corrupting 142file systems!) in e2fsck, debugfs, and libext2fs. 143 144Fixed e2fsck's handling of the journal's s_errno field. E2fsck was 145not properly propagating the journal's s_errno field to the superblock 146field; it was not checking this field if the journal had already been 147replayed, and if the journal *was* being replayed, the "error bit" 148wasn't getting flushed out to disk. 149 150Fixed a false positive complaint by e2fsck if all of the extents in 151the last extent tree block are uninitalized and located after the end 152of the file as defined by i_size. 153 154The dumpe2fs will now display the journal's s_errno field if it is 155non-zero, and it will also display the journal's 64-bit feature flag 156if present. 157 158Fix e2fsck so that it always opens the device file in exclusive mode 159when it might need to modify the file system, and never if the -n 160option is specified. (Previously, there were a few corner cases where 161it might get things wrong in either direction.) 162 163E2fsck now correctly truncates or deallocates extent-mapped inodes on 164the orphan list. The root cause was a bug in libext2fs's block 165iterator which could end up skipping an extent when the last block in 166an extent is removed, causing the current extent to be removed from 167the extent tree. 168 169E2fsck now correctly sets the global free block and inode counts when 170truncating or removing inodes on the orphan list in preen mode. 171Previously, it would leave these values would be set incorrectly, 172which is largely a cosmetic issue since the kernel no longer pays 173attention to those fields, but it can cause spurious complaints in 174subsequent e2fsck runs. 175 176Fix i_blocks accounting when the libext2fs library needs to add or 177remove an extent tree block on bigalloc file systems. 178 179The lsattr and chattr programs now support the No_COW flag for the 180benefit of btrfs. 181 182Debugfs now interprets date strings of the form @ddd as ddd seconds 183after the beginning of the epoch. This is handy when setting an inode 184number into the d_time field when debugging orphan list handling. 185 186Fix a precedence bug with built-in quota support which might result in 187e2fsck paying attention to the quota inode field even if the built-in 188quota feature flag is not set. Fortunately, in practice that 189superblock field should be zero for non-built-in quota file systems, 190so it's unlikely this bug would have caused problems. 191 192Updated/fixed various man pages. (Addresses Debian Bugs: #674453, 193#674694) 194 195Programmer's Notes 196------------------ 197 198The regression test suite can now run the integration tests in the 199tests directory in parallel, via "make -jN check". 200 201Add new test, f_zero_extent_length which tests e2fsck's handling of 202the case where all of the extents in the last extent tree block are 203uninitialized extents after i_size. 204 205Add a new test, f_jnl_errno, which checks handling of an error 206indication set in the journal superblock. 207 208Fix the test f_jnl_64bit so that it properly checks e2fsck's handling 209of a 64-bit journal. 210 211Add two tests, f_orphan_indirect_inode and f_orphan_extent_inode which 212tests e2fsck's handling of orphan inodes in preen mode, and truncation 213of extent inodes on the orphan list. 214 215Fixed more OS X portability issues. 216 217 218E2fsprogs 1.42.3 (May 14, 2012) 219=============================== 220 221Fix a bug in the Unix I/O manager which could cause corruption of file 222systems with more than 16TB when e2fsprogs is compiled in 32-bit mode 223(i.e., when unsigned long is 32-bits). Also fix a bug which caused 224dumpe2fs to incorrectly display block numbers > 32-bits. 225 226Improve the support for integrated quota files (where quota is a first 227class supported feature using hidden files in the ext4 file system). 228Previously the quota file was getting rewritten even when it was not 229necessary, and e2fsck would erroneously try to hide quota files which 230were already hidden. 231 232Quiet complaints in e2fsck when the total free blocks or inodes are 233incorrect in the superblock after an system crash, since we don't 234update nor depend on the superblock summaries at each commit boundary. 235 236Fixed a regression introduced in 1.42.2 which would cause applications 237outside of e2fsprogs which did not pass the EXT2_FLAG_64BIT (and so 238would were still using the legacy 32-bit bitmaps) to crash. This was 239due to missing 32-bit compat code in side the function 240ext2fs_find_first_zero_generic_bmap(). (Addresses Red Hat Bugzilla: 241#808421) 242 243Fix a bug which would cause mke2fs to fail creating the journal if 244/etc/mtab and /proc/mounts are missing. (Addresses Sourceforge Bug: 245#3509398) 246 247Updated/fixed various man pages. 248 249Update Czech, Dutch, French, German, Polish, Sweedish, and Vietnamese 250translations 251 252Fixed various Debian Packaging issues. 253 254 255Programmer's Notes 256------------------ 257 258Change the nonsensically wrong types in the function signature of the 259inline function ext2fs_find_first_zero_block_bitmap2(). This was 260caused by a cut and paste error; fortunately no code in e2fsprogs used 261this inline function, and there are any users of this functions 262outside of e2fsprogs. 263 264Add support for systems which have valloc(), but which do have 265posix_memalign() nor memalign() (such as MacOS 10.5). 266 267Refactor and clean up the allocation of aligned buffers for Direct I/O 268support. Previously some allocations were requesting a greater 269alignment factor that what was strictly necessary. Also optimize 270reading and writing bitmaps using Direct I/O when the size of the 271bitmap did not fully cover the file system blocksize. 272 273Reserve the codepoints for the INCOMPAT features LARGEDATA and 274INLINEDATA. 275 276Improved the regression test suite by adding some new integration 277tests (f_jnl_32bit, f_jnl_64bit) which detect breakage of the on-disk 278jbd2 format, as well as f_eofblocks which tests the new handling of 279uninitialized and initialized blocks beyond i_size. Also add a new 280unit test which verifies 32-bit bitmaps support and the new 281find_first_zero primitives. 282 283Add a few dependencies to fix parallel (make -j) builds. 284 285Removed bash'isms which were breaking the regression test suite on 286systems where /bin/sh is not bash. 287 288The config.guess and config.sub have been updated to the 2012-02-10 289version. 290 291Fix a portability problem caused by assuming the present of mntent.h 292means that setmntent() exists. Instead, explicitly test for this in 293the configure script. 294 295If the sys/signal.h header file does not exist, don't try to include 296it, since it's not available on all systems. 297 298Add support for systems that do not support getpwuid_r() 299 300The configure script now supports a new option, 301--enable-relative-symlinks, which will install relative symlinks for 302the ELF shared library files. (Addresses Sourceforge Bug: #3520767). 303 304When building BSD shared libraries make sure the LDFLAGS variable is 305passed to the linker. Fixing this allows, for example, e2fsprogs to 306be built in 32-bit mode on Mac OS X Lion (Addresses Sourceforge Bug: 307#3517272) 308 309Fix gcc -Wall nitpicks 310 311 312E2fsprogs 1.42.2 (March 27, 2012) 313================================= 314 315The resize2fs program uses much less CPU and is much faster for very 316large file systems. (Addresses Debian Bug: #663237) 317 318The seti and freei commands in debugfs can now take an optional length 319argument to set and clear a contiguous range of inodes. 320 321E2fsck will now make explicit checks for the EOFBLOCKS_FL, since we 322plan to remove support for it from the kernel file system driver. It 323really wasn't very useful and was causing more problems than it 324solves. Since e2fsck will complain if inodes that should have 325EOFBLOCKS_FL do not have the flag set, we are going to remove this 326check from e2fsprogs first, and then only remove the flag from the 327kernel much later. 328 329The mke2fs program can now use direct I/O via "mke2fs -D". This will 330slow down the mke2fs, but it makes it more polite on a loaded server 331by limiting the amount of memory that gets dirtied by mke2fs when it 332is using buffered I/O. 333 334E2fsck was needlessly closing and re-opening the file system as a side 335effect of adding Multiple Mount Protection (MMP). This isn't 336necessary for non-MMP file systems, so drop it. 337 338Print errors returned by ext2fs_open2() and ext2fs_check_desc() so we 339can more easily diagnose memory allocation failures caused by 340insufficient memory. E2fsck will now abort if there are memory 341allocation failures when the file system is initially opened and 342during the block group descriptor checks. (Addresses Google Bug: 343#6208183) 344 345If there are incorrect block group checks, e2fsck will now report the 346incorrect and corrected checksum values. 347 348The e2fsck progam can now write log files containing the details of 349the problems that were found and fixed directly, via configuration 350parameters in /etc/e2fsck.conf. 351 352Added the ability to limit the number of messages reported by e2fsck 353for a given problem type. This avoids a potential bottleneck if there 354is a serial console which can cause a boot sequence to take a long 355time if e2fsck needs to report many, many file system errors. 356 357The dumpe2fs, debugfs, and tune2fs now use rbtree bitmaps, which cause 358them to use much less memory for large file systems. 359 360The dumpe2fs program will now print the expected block group checksum 361if it is incorrect. This helps to diagnose problems caused by 362incorrect block group checksums. 363 364E2fsck now checks for extents with a zero length, since the kernel 365will oops if it comes accross such a corrupted data structure. (See 366https://bugzilla.kernel.org/show_bug.cgi?id=42859) 367 368E2fsck has a number of bugs relating to discard that have been fixed. 369(1) Fixed a bug which could cause e2fsck to discard portions of the 370inode table which were actually in use. (2) E2fsck will now avoid 371using discard if the block device doesn't zero data on discard, since 372otherwise this could cause problems if the file system gets corrupted 373in the future. (3) E2fsck will now avoid using discard when it is run 374in read-only mode. (4) Fixed a bug which caused e2fsck to not issue 375discards in the last block group. 376 377E2fsck's CPU utilization in pass 5 has been optimized, which will 378speed up e2fsck slightly. 379 380E2image will now skip copying uninitialized bitmap and inode table 381blocks. 382 383Fixed mke2fs -S so it does not corrupt the first block group's 384information. 385 386E2fsck will now check the new sysfs interface to determine if we are 387using the battery or AC mains. (Addresses SourceForge Bug: #3439277) 388 389Updated/fixed various man pages. (Addresses Debian Bug: #665427) 390 391Fixed various Debian Packaging issues. (Addresses Debian Bug: #665885) 392 393Programmer's Notes 394------------------ 395 396Fixed various portability issues for non-Linux systems, particularly 397MacOS X, as well as Linux systems running with the just-released glibc 3982.15. 399 400Fix file descriptor leak in ext2fs_close() if the file system with 401uninit_bg is opened read/only with a backup superblock. (Addresses 402SourceForge Bug: #3444351) 403 404Fixed an invalid return in a non-void function in the quota code. 405(Addresses SourceForge Bug: #3468423) 406 407Fixed the debian rules file so that the calls to dpkg-buildflags works 408when the shell is dash. 409 410The debian package build now uses V=1 so that there is more 411information about potential build failures in debian buildd logs. 412 413If the uninit flags get cleared by functions such as 414ext2fs_new_inode() or ext2fs_new_block2(), we now make sure the 415superblock is marked dirty and the block group descriptor checksum is 416updated if necessary. 417 418The debian rules file will now try to load debian/rules.custom of it 419exists. This flie can skip various builds for speed reasons if there 420is no need for the e2fsck-static or udeb packages. Available 421customizations in the rules file includes SKIP_STATIC=yes, 422SKIP_BF=yes, and SKIP_DIETLIBC=yes. 423 424In addition, if the file misc/mke2fs.conf.custom.in exists in the 425source tree, it will be used instead of the standard misc.conf file in 426the upstream sources. This makes it easier for Debian-derived systems 427to distribute a custom mke2fs.conf file without having to worry about 428merge issues if future versions of e2fsprogs makes changes in the 429upstream default version of mke2fs.conf. 430 431 432E2fsprogs 1.42.1 (February 17, 2012) 433=================================== 434 435The mke2fs and e2fsck now use significantly less memory when creating 436or checking very large file systems. This was enabled by adding 437extent-based bitmaps which are stored using a red-block tree, since 438block and inode allocations tend to be contiguous. 439 440The command mke2fs -S is used as a last ditch recovery command to 441write new superblock and block group descriptors, but _not_ to destroy 442the inode table in hopes of recovering from a badly corrupted file 443system. So if the uninit_bg feature is enabled, mke2fs -S will now 444set the unused inodes count field to zero. Otherwise, e2fsck -fy 445after using mke2fs -S would leave the file system completely empty. 446 447Since mke2fs recognizes mke3fs in argv[0] to mean "mkfs.ext3", also 448honor "mke4fs" to work the same as "mke2fs.ext4", since RHEL5 has 449installed an mke2fs binary using that name. 450 451The usage and help messages for the -G, -t and -T options in mke2fs 452have been fixed. 453 454If e2fsck needs to use the backup group descriptors, the 455ext2fs_open2() function clears the UNINIT bits to ensure all of the 456inodes in the file systems get scanned. However, the code which reset 457the UNINIT flags did not also recalculate the checksum, which produced 458many spurious (and scary) e2fsck messages. This has been fixed by 459resetting cheksums when the UNINIT bits are cleared. 460 461Relax a check in e2fsck which required that the block bitmap to be 462initialized when the inode bitmap is in use. This will allow us to 463eventually eliminate code from the kernel which forcibly initialized 464the block bitmap when the inode bitmap is first used, requiring an 465extra journal credit and disk write. (Addresses Google Bug: #5944440) 466 467Make sure rdebugfs (which may be installed setuid or setgid disk) does 468not honor environment variables if euid != uid or egid != gid. 469 470Debugfs's ncheck command has been optimized and now is much more 471robust with faced with corrupted file systems. The ncheck command 472also now has a -c option which will verify the file type information 473in the directory entry to see if matches the inode's mode bits. This 474is extremely useful when trying to use debugsfs to determine which 475parts of the file system metadata can be trusted. 476 477E2image will try to use ftruncate64() to set the i_size for raw 478images, instead of writing a single null byte. This avoid allocating 479an extra block to the raw image, for those file systems and/or 480operating systems that support this. (Linux does.) In addition, fix 481a logic bug that caused the file to not be properly extended if the 482size of the last hole was exactly an multiple of a megabyte. 483 484Fixed a bug in resize2fs where for 1k and 2k file systems, where 485s_first_data_block is non-zero, this wasn't taken into account when 486calculate the minimum file system size for use with the -M option. 487 488Fixed the badblocks program to honor the -s flag when in read-only -t 489mode. (Addresses Debian Bug #646629) 490 491Update Czech, Dutch, French, Polish, and Sweedish translation from the 492Translation Project. 493 494Fixed various Debian Packaging issues so that dpkg-buildflags is used 495if present, which allows e2fsprogs to be built with security hardening 496flags. (Addresses Debian Bugs: #654457) 497 498Programmer's Notes 499------------------ 500 501Fix a bug in ext2fs_clear_generic_bmap() when used for 32-bit bitmaps. 502This was only an issue for programs compiled against e2fsprogs 1.41 503that manipulate bitmaps directly. (Addresses Sourceforge Bugs: 504#3451486) 505 506The libext2fs library now uses sysconf() to fetch the page size, instead 507of the deprecated getpagesize(). 508 509The ext2fs_get_pathname() function will return a partial path if an a 510directory in the path is not a directory, displaying it as an inode 511number in angle brackets instead of giving up and displaying an error. 512This is much more helpful when a user is trying to debug a corrupted 513file system. 514 515Codepoints for the RO_COMPAT_REPLICA feature has been reserved. 516 517Added a new library function, ext2fs_file_get_inode_num(), for use by 518fuse2fs. 519 520Fixed a bug in ext2fs_file_set_size2() so that when it is truncating a 521file, it actually works. 522 523The block iterator now properly honors the BLOCK_ABORT flag for 524extent-based flags. Previously, it didn't, which generally made code 525be less efficient, but it could cause bugs in ext2fs_link(), for 526example, by causing it to insert multiple directory entries. 527 528Fixed an (harmless other than causing a compiler warning) use of an 529uninitialized variable in e2fsck's MMP code. 530 531 532E2fsprogs 1.42 (November 29, 2011) 533================================== 534 535This release of e2fsprogs has support for file systems > 16TB. Online 536resize requires kernel support which will hopefully be in Linux 537version 3.2. Offline support is not yet available for > 16TB file 538systems, but will be coming. 539 540This release of e2fsprogs has support for clustered allocation. This 541reduces the number of block (now cluster) bitmaps by allocating and 542deallocating space in contiguous power-of-2 collections of blocks, 543which are called clustered. This is a file system level feature, 544called 'bigalloc', which must be enabled when the file system is 545initially formatted. It is not backwards compatible with older 546kernels. 547 548Added support for the Multi-Mount Protection (MMP) feature. 549 550E2fsck more efficiently uses scratch files for really big file 551systems. (This is a feature that has to be turned on explicitly; see 552[scratch_files] in the e2fsck.conf man page.) 553 554Fix a bug in e2fsck where if the free blocks and inodes counts are 555incorrect, e2fsck would fix them without printing an error message. 556This would cause a "*** FILE SYSTEM WAS MODIFIED ***" message without 557any explanation of what was fixed. 558 559E2fsck will no longer attempt to clone an extended attribute block in 560pass1b handling if the file system does not support extended 561attributes. 562 563E2fsck will be more careful accidentally asking the user to continue 564if the file system is mounted, so that an escape sequence won't cause 565a false positive. (Addresses Debian Bug: #619859) 566 567E2fsck now uses less cpu time in pass 5 when large portions of the 568bitmaps are uninitialized. 569 570E2fsck will no longer segault when a corrupted file system has a bad 571extent, and removing it leads to a block needing to be deallocated. 572(Addresses SourceForge Bug: #2971800) 573 574E2fsck will catch termination signals (segfaults, bus errors, sigfpe) 575and print debugging information to make it easier to find potential 576problems. 577 578E2fsck will check to see if the bad block inode looks insane, and will 579skip trying to use if it certain fields which should be zero look 580non-zero. This is to avoid a corrupted bad block inode causing e2fsck 581to get confused and do more harm than good. 582 583If e2fsck modifies certain superblock fields which the kernel doesn't 584look at, it will now mark the superblock as dirty without marking the 585file system as changed. This avoids signalling the init scripts that 586a reboot is necessary, since the kernel isn't going to look at those 587fields, so it won't care if they have been changed. 588 589Fixed a bug in the libext2fs library (in the binary search routine of 590the icount abstraction) that could (very, very rarely) cause e2fsck to 591die in the middle of pass 1 or pass 2 processing. 592 593E2fsck will not try to do a discard operation if the -n option was 594specified on the command line. 595 596E2fsck now supports an extended "discard" option which will cause 597e2fsck to attempt discard all unused blocks after a full, successful 598file system check. 599 600E2fsck will check for the bad block inode to make sure it looks sane 601before trusting it, to avoid causing more harm than good to the file 602system. 603 604E2fsck now returns additional status bits in its exit code if it 605aborts early in the e2fsck run. 606 607E2fsck now correctly calculates the maximum file size in the case of 608the huge_file file system feature enabled without extents. 609 610The mke2fs and e2fsck programs now tries to use the punch hole command 611as a "discard" when operating on normal files. 612 613The e2image program now supports the qcow2 format, which is a more 614efficient way of capturing file system dumps. 615 616Mke2fs now supports the [devices] stanza in mke2fs.conf which allows 617per-device defaults to be specified in the configuration file. 618 619Mke2fs now supports the reserved_ratio relation in the [defaults] and 620[fs_types] section in mke2fs.conf. 621 622Mke2fs now creates extent-mapped directories for the root and 623lost+found directories. 624 625Mke2fs will skip zero'ing the journal if the extended option 626"lazy_journal_init" is specified. This can save a lot of time, but it 627does add a small amount of risk if the system crashes before the 628journal is overwritten entirely once. It is epsecially useful for 629testing. 630 631Mke2fs will now create file systems that enable user namespace 632extended attributes and with time- and mount count-based file 633system checks disabled. 634 635Mke2fs will not set a stride or strip size of one block based on block 636device attributes obtained from sysfs. 637 638Mke2fs now displays a progress report during the discard process. 639 640Mke2fs now handles extreme file system parameters correctly which 641previously caused the inodes per group to drop below 8, leading to a 642segfault. (The inodes per group must be a multiple of 8, but the code 643didn't correctly deal with an inodes per group count less than 8.) 644 645Mke2fs and tune2fs previously would give an error if the user tried 646setting the stride and stripe-width parameters to zero; but this is 647necessary to disable the stride and stripe-width settings. So allow 648setting these superblock fields to zero. (Addresses Google Bug: 649#4988557) 650 651Mke2fs now gives a warning if the auto-detected block size exceeds the 652system's page size. 653 654If the enable_periodic_fsck option is false in /etc/mke2fs.conf (which 655is the default), mke2fs will now set the s_max_mnt_count superblock 656field to -1, instead of 0. Kernels older then 3.0 will print a 657spurious message on each mount then they see a s_max_mnt_count set to 6580, which will annoy users. (Addresses Debian Bug: #632637) 659 660The default mke2fs.conf now has entries for "big" and "huge", which 661are needed for very big file systems. 662 663The resize2fs program now has support for a new online resize ioctl 664that can support file systems > 16TB, once it arrives in v3.x kernels. 665 666Fixed bug which caused resize2fs to fail when shrinking an empty file 667system down to its minimal size. (Addresses Sourceforge Bug #3404051) 668 669Fixed tune2fs's mount options parsing. (Addresses Debian Bug: #641667) 670 671Allow tune2fs to remove external journals if the device can not be 672found. 673 674Debugfs's icheck will now correctly find inodes which use the 675searched-for block as an extended attribute block. 676 677Debugfs now has a new "punch" command which remove blocks from the 678middle of an inode. 679 680Debugfs now has a new "e2freefrag" command which analyzes the free 681space fragmentation of the file system, using the same code as the 682e2freefrag program. 683 684Debugfs now has a "filefrag" command which displays information about 685a file's fragmentation. 686 687Add support to build a metadata-only, read-only, stripped-down version 688of debugfs called rdebugfs. 689 690Fixed a potential stack overrun bug in debugfs. 691 692The badblocks program now correctly recovers from I/O errors when 693direct I/O is being used. The badblocks command now also supports a 694-B option which forces the use of buffered I/O, and the -v option will 695provide a more detailed breakdown of read, write, and failed 696comparison errors. 697 698Added e4defrag tool which uses the EXT4_IOC_MOVE_EXT ioctl. 699 700Added support for journals larger than 2GB. 701 702Support using both hard links and symlinks when installing e2fsprogs. 703 704Add overflow checking to tune2fs -i's fsck interval, which must fit in 705a 32-bit field. 706 707The debugfs command now has a new 'blocks' command which prints out 708data blocks of a particular inode in a format which is useful for 709scripting. 710 711Filefrag will report 0 extents correctly in verbose mode. (Addresses 712RedHat Bugzilla: #653234) 713 714Filefrag has been fixed so its -v report prints the correct expected 715block number (previously there had been an off-by-one error). In 716addition, it will now display the number of contiguous extents when -v 717is not specified. This makes it consistent with the number of extents 718printed when the -v option was specified. In addition, the number of 719contiguous extents is far more interesting/useful than the number of 720physical extents for very large files. (Addresses Debian Bug: 721#631498, #644792) 722 723Logsave's usage message has been fixed. (Addresses Debian Bug: 724#619788) 725 726Avoid an infinite loop in ext2fs_find_block_device() if there are 727symlink loops in /dev caused by a buggy udev. 728 729Added a useful "fallocate" program to the contrib directory. 730 731Fixed an ABI compatibility problem in libext2fs which broke the dump 732program. Also added back some macros which dump needed so it could 733compile against the latest version of ext2_fs.h (Addresses Debian Bug: 734#636418) 735 736Fixed parsing of MNTOPT_ options for tune2fs and debugfs (Addresses 737Debian Bug: #641667) 738 739Added internationalization support for libcom_err error table strings. 740 741Fixed various spelling mistakes found in various output strings found 742by I18N translators. 743 744Update translations: French, Chinese, Germany, Indonesian, Swedish, 745Vietnamese, Polish, Dutch, Czech. (Addresses Debian Bugs: #520985, 746#620659) 747 748Fixed various Debian Packaging issues. (Addresses Debian Bugs: 749#614662, #632169, #641838, #627535, #629355) 750 751Updated/clarified man pages. (Addresses Debian Bugs: #639411, 752#642193, #634883) 753 754Programmer's Notes 755------------------ 756 757Initial support for quota as an integrated feature, where the quota 758files are hidden system files that are automatically maintained by 759e2fsck is present, although disabled by default. It must be enabled 760by using the configure option --enable-quota. 761 762Reserved the on-disk fields for the metadata checksum and snapshot 763features, which are currently in development. 764 765The ext2fs library now has the new functions ext2fs_punch(), 766ext2fs_get_memzero() and ext2fs_file_get_inode(). 767 768The ext2fs library now has support for calculating the crc32c checksum 769(via the new functions ext2fs_crc32c_be and ext2fs_crc32c_le). 770 771The I/O manager now supports the discard operation. 772 773Reserved file system code points for new 1st class quota feature. 774 775Shortened the compilation lines by moving the autoconf defines to 776a config.h header file. 777 778Fixed a potential free of an unitialized pointer in 779ext2fs_update_bb_inode(). 780 781Fixed miscellaneous gcc -Wall and coverity warnings. 782 783Fixed portability issues for Mac OS X, Hurd, and FreeBSD. 784 785Fixed a build failure when OMIT_COM_ERR is defined. 786 787Improved error checking and fixed memory leaks caused by error return 788paths. 789 790Add ext2fs_flush2() and ext2fs_close2() which takes a flag to allow 791the fsync() to be skipped. 792 793Added a test for extent-mapped journals by mke2fs and tune2fs. 794 795Added a test for creating a large (over 4GB) journal using mke2fs. 796 797Added a test to make sure the inode size remains constant. 798 799The regression test script now prints the list of which tests failed. 800 801The regression test system now uses /tmp for its scratch files, which 802is often a tmpfs mounted file system and hence much faster. 803 804The i_e2image test was fixed so it works with valgrind. 805 806Tests that rely on debugfs are now skipped if debugfs wasn't built. 807 808Fixed the dependencies for "make check" so all required dependencies 809are built before running the gression tests. 810 811A link to com_err.h is installed in $(includedir) during a "make install". 812 813The po/*.gmo files are automatically rebuilt if they are missing or 814out of date. This helps out Debian packaging. 815 816Allow ext2fs_get_memalign() to compile on systems that don't have 817posix_memalign(). 818 819Fixed a namespace leak in libext2fs (tdb_null). 820 821 822E2fsprogs 1.41.14 (December 22, 2010) 823===================================== 824 825Fix spurious complaint in mke2fs where it would complain if the file 826system type "default" is not defined in mke2fs.conf. 827 828The resize2fs program will no longer clear the resize_inode feature 829when the number reserved group descriptor blocks reaches zero. This 830allows for subsequent shrinks of the file system to work cleanly for 831flex_bg file systems. 832 833The resize2fs program now handles devices which are exactly 16T; 834previously it would give an error saying that the file system was too 835big. 836 837E2fsck (and the libext2fs library) will not use the extended rec_len 838encoding for file systems whose block size is less than 64k, for 839consistency with the kernel. 840 841Programming notes 842----------------- 843 844E2fsprogs 1.41.13 would not compile on big-endian systems. This has 845been fixed. (Addresses Sourceforge Bug: #3138115) 846 847The ext2fs_block_iterator2() function passed an incorrect ref_offset 848to its callback function in the case of sparse files. (Addresses 849Sourceforge Bug: #3081087) 850 851Fix some type-punning warnings generated by newer versions of gcc. 852 853 854E2fsprogs 1.41.13 (December 13, 2010) 855===================================== 856 857E2fsck now supports the extended option "-E journal_only", which 858causes it to only do a journal replay. This is useful for scripts 859that want to first replay the journal and then check to see if it 860contains errors. 861 862E2fsck will now support UUID= and LABEL= specifiers for the -j option 863(which specifies where to find the external journal). (Addresses 864Debian Bug #559315) 865 866E2fsck now has support for the problems/<problem code>/force_no 867configuration option in e2fsck.conf, which forces a problem to not be 868fixed. 869 870Dumpe2fs will now avoid printing large negative offsets for the bitmap 871blocks and instead print a message which is a bit more helpful for 872flex_bg file systems. 873 874Mke2fs will now check the file system type (specified with the -t 875option) to make sure it is defined in the mke2fs.conf file; if it is 876not, it will print an error and abort. If the usage type (specified 877with the -T option) is not defined in mke2fs.conf, mke2fs will print a 878warning but will continue. (Addresses Debian Bug #594609) 879 880Clarified error message from resize2fs clarifying that on-line 881shrinking is not supported at all. (Addresses Debian Bug #599786) 882 883Fix an e2fsck bug that could cause a PROGRAMMING BUG error to be 884displayed. (Addresses Debian Bug #555456) 885 886E2fsck will open the external journal in exclusive mode, to prevent 887the journal from getting truncated while it is in use due to a user 888accidentally trying to run e2fsck on a snapshotted file system volume. 889(Addresses Debian Bug #587531) 890 891Fix a bug in e2fsck so it has the correct test for the EOFBLOCKS_FL 892flag. 893 894The tune2fs program can now set the uninit_bg feature without 895requiring an fsck. 896 897The tune2fs, dumpe2fs, and debugfs programs now support the new ext4 898default mount options settings which were added in 2.6.35. 899 900The e2fsck and dumpe2fs programs now support the new ext4 superblock 901fields which track where and when the first and most recent file 902system errors occurred. These fields are displayed by dumpe2fs and 903cleared by e2fsck. These new superblock fields were added in 2.6.36. 904 905Debugfs now uses a more concicse format for listing extents in its 906stat command. This format also includes the interior extent tree 907blocks, which previously was missing from stat's output for 908extent-based files. 909 910Debugfs has a new option, -D, which will request Direct I/O access of 911the file system. 912 913Mke2fs will skip initializing the inode table if a device supports 914discard and the discard operation will result in zero'ed blocks. 915 916Badblocks will now correctly backspace over UTF-8 characters when 917displaying its progress bar. (Addresses Gentoo Bug #309909; Addresses 918Debian Bugs #583782 and #587834) 919 920E2freefrag will now display the total number of free extents. 921 922Resize2fs -P no longer requires a freshly checked filesystem before 923printing the minimum resize size. 924 925Fixed a floating point precision error in a binary tree search routine 926that can lead to seg fault in e2fsck and resize2fs. 927 928Fixed a bug in e2fsck where if both the original and backup superblock 929are invalid in some way, e2fsck will fail going back to the original 930superblock because it didn't close the backup superblock first, and 931the exclusive open prevented the file system from being reopened. 932 933Fixed a big in e2freefrag which caused getopt parsing to fail on 934architectures with unsigned chars. (Addresses Gentoo Bug: #299386) 935 936Clarified an mke2fs error message so a missed common in an -E option 937(i.e., mke2fs -t ext4 -E stride=128 stripe-width=512 /dev/sda1") 938results in a more understandable explanation to the user. 939 940Mke2fs now displays the correct valid inode ratio range when 941complaining about an invalid inode ratio specified by the user. 942 943Mke2fs now understands the extended option "discard" and "nodiscard", 944and the older option -K is deprecated. The default of whether 945discards are enabled by default can be controled by the mke2fs.conf 946file. 947 948Mke2fs's handling of logical and physical sector sizes has been 949improved to reflect the fact that there will be some SSD's with 8k and 95016k physical sectors arriving soon. Mke2fs will no longer force block 951size to be the physical sector size, since there will be devices where 952the physical sector size is larger than the system's page size, and 953hence larger than the maximal supported block size. In addition, if 954the minimal and optimal io size are not exported by the device, and 955the physical sector size is larger than the block size, the physical 956sector size will be used to set the Raid I/O optimization hints in the 957superblock. 958 959E2fsck will now display a better, more specific error message when the 960user enters a typo'ed device name, instead of blathering on about 961alternate superblocks. 962 963Fixed various Debian Packaging Issues 964 965Updated/clarified man pages (Addresses Debian Bugs: #580236, #594004, 966#589345, #591083; Addresses Launchpad Bug: #505719) 967 968Update the Chinese, Chzech, Dutch, French, Germany, Indonesian, 969Polish, Swedish, and Vietnamese translations. 970 971 972Programmer's Notes 973------------------ 974 975Fix a dependency definition for the static and profiled blkid 976libraries which could cause compile failures in some configurations. 977(Addresses Debian Bug: #604629) 978 979Add support for Direct I/O in the Unix I/O access layer. 980 981Fixed a memory leak in the Unix I/O layer when changing block sizes. 982 983Fixed minor memory leaks in mke2fs. 984 985Added a new function to the ext2fs library, ext2fs_get_memalign(). 986 987The tst_super_size test program will check to make sure the superblock 988fields are correctly aligned and will print them out so they can be 989manually checked for correctness. 990 991Fixed some makefile dependencies for test programs in lib/ext2fs. 992 993Reserved the feature flags and superblock fields needed for the Next3 994snapshot feature. 995 996Reserved the feature flags for EXT4_FEATURE_INCOMPAT_DIRDATA and 997EXT4_INCOMPAT_EA_INODE. 998 999 1000E2fsprogs 1.41.12 (May 17, 2010) 1001================================ 1002 1003Mke2fs now gives a correct error message if the external journal 1004device is not found. (Addresses Red Hat Bug #572935) 1005 1006Resize2fs -P will now refuse to print a minimum size if the file 1007system is not clean. Previously it would go ahead and print a minimum 1008size anyway, which might not be correct, leading to user confusion. 1009 1010E2fsck now tests for extents that begin at physical block 0 and 1011rejects them as invalid. (Addresses Google Bug: #2573806) 1012 1013Fixed a bug in e2fsck which could cause it to crash when trying to 1014remove an invalid extent and the block bitmaps hadn't yet been loaded. 1015(Addresses SourceForge Bug: #2971800) 1016 1017E2fsck now will completely skip time-based checks if the system clock 1018looks insane or the option broken_system_clock is set in 1019/etc/e2fsck.conf. 1020 1021E2fsck would previously report an i_blocks corruption for a 4T file 1022created using posix_fallocate; this bug has been fixed. 1023 1024E2fsck will now correctly mark a sparse journal as invalid and will 1025delete and recreate the journal to address the problem. 1026 1027E2fsck would previously incorrectly ask the user whether she would 1028like to abort the file system check after finding a problem --- and 1029then abort regardless of the user's answer. This is annoying, and 1030has been fixed. 1031 1032E2fsck can now continue even if it fails to recreate the resize 1033inode; previously it would just abort the file system check 1034altogether. 1035 1036E2fsck could potentially remove directory entries for inodes found in 1037the unused region of the inode table; this would generally happen on 1038ext4 file systems that do not use journaling. This bug has been fixed 1039by not clearing these directory entries once it has been established 1040that bg_unused_inodes may not be trustworthy; once pass #2 has been 1041completed, e2fsck will restart the file system check from the 1042beginning, and then it will be safe to delete any directory entries 1043pointing to inodes that appear to be deleted. (Addresses Google Bug: 1044#2642165) 1045 1046E2fsck will not try to set the block group checksums if the user 1047cancels the fsck with a control-C. It's a bad idea to set the 1048checksums if e2fsck hasn't been completed, and it often results an the 1049error message, "Inode bitmap not loaded while setting block group 1050checksum info". (Addresses Launchpad Bug: #582035) 1051 1052The mke2fs program now queries the kernel for the physical as well as 1053the logical sector size, and will not allow a blocksize below the 1054logical, and will strongly encourage a blocksize at least as big as 1055the physical blocksize. This is needed for 4k sector drives that 1056emulate 512 byte sector sizes. 1057 1058Mke2fs will now allow a flex_bg size of 1. This is unusual, and 1059rarely needed, but it is a legal value. 1060 1061E2fsck will check for cases where the EOFBLOCKS_FL is set when it is 1062not needed, and offer to clear it; this is a sign of a kernel bug, but 1063more importantly, some released kernels may crash when this situation 1064is encountered on ext4 file systems. (Addresses Google Bug: #2604224) 1065 1066E2fsck will use the EOFBLOCKS_FL flag exclusively to check whether 1067i_size is correct. (Kernels starting with 2.6.34 will set 1068EOFBLOCKS_FL.) 1069 1070The com_err library will now only output ^M (a CR character) when the 1071tty is in raw mode. 1072 1073Update the Czech, Chinese, Dutch, French, Germany, Indonesian, Polish, 1074and Vietnamese translations. 1075 1076Fixed various Debian packaging issues --- see debian/changelog for 1077details. (Addresses Debian Bugs: #571247, #563487) 1078 1079 1080Programmer's Notes 1081------------------ 1082 1083The regression test suite now uses its own mke2fs.conf file, so that 1084downstream distributions want change the mke2fs.conf file which is 1085distributed in the RPM or dpkg file, without worrying about screwing 1086up the regression test results. 1087 1088Always build namei.o so that building with configure --disable-debugfs 1089works correctly. Long-term, if we care about reduced e2fsprogs 1090builds, we need a more general solution for deciding what .o files are 1091needed for a particular build. Given that install floppies are going 1092(gone?) the way the dodo bird, we probably don't care, though. 1093(Addresses Sourceforge Bug: #2911433) 1094 1095Add configure options --enable-symlink-build and 1096--enable-symlink-install, which allow e2fsprogs be built using 1097symlinks instad of hard links, and to be installed using symlinks 1098instead of hard links, respectively. It is useful when the file 1099system where the build is taking place, or the file system where 1100e2fsprogs is installed, can't handle hard links for some reason. 1101(Addresses Sourceforge Bug: #1436294) 1102 1103Fixed compile warning in mke2fs.c. 1104 1105 1106E2fsprogs 1.41.11 (March 14, 2010) 1107================================== 1108 1109E2fsck will no longer give a fatal error and abort if the physical 1110device has been resized beyond 2**32 blocks. (Addresses Launchpad 1111Bug: #521648) 1112 1113Debugfs has a bug fixed so that "logdump -b <blk>" now properly shows 1114the allocation status of the block <blk>. (Addresses Debian Bug: 1115#564084) 1116 1117E2fsck now prints a much more emphatic and hopefully scary message 1118when a file system is detected as mounted while doing a read/write 1119check of the filesystem. Hopefully this will dissuade users from 1120thinking, "surely that message doesn't apply to *me*" :-( 1121 1122E2fsck -n will now always open the file system read-only. We now 1123disallow certain combination of options which previously were manual 1124exceptions; this is bad because it causes users to think they are 1125smarter than they really are. So "-n -c", "-n -l", "-n -L", and "-n 1126-D" are no longer supported. (Addresses Launchpad Bug: #537483) 1127 1128In e2fsprogs 1.41.10, mke2fs would ask for confirmation to proceed if 1129it detected a badly aligned partition. Unfortunately, this broke some 1130distribution installation scripts, so it now just prints the warning 1131message and proceeds. (Addresses Red Hat Bug: #569021. Addresses 1132Launchpad Bug: #530071) 1133 1134Mke2fs would take a long time to create very large journal files for 1135ext4. This was caused by a bug in ext2fs_block_iterate2(), which is 1136now fixed. 1137 1138E2fsck now understands the EOFBLOCKS_FL flag which will be used in 11392.6.34 kernels to make e2fsck not complain about blocks deliberately 1140fallocated() beyond an inode's i_size. 1141 1142E2fsprogs 1.41.10 introduced a regression (in commit b71e018) where 1143e2fsck -fD can corrupt non-indexed directories when are exists one or 1144more file names which alphabetically sort before ".". This can happen 1145with ext2 filesystems or for small directories (take less than a lock) 1146which contain filenames that begin with a space or some other 1147punctuation mark. (Addresses Debian Bug: #573923, Addresses Launchpad 1148Bug: #525114) 1149 1150 1151Programmer's Notes 1152------------------ 1153 1154Add new test, f_rehash_dir, which checks to make sure e2fsck -D works 1155correctly. 1156 1157The libcom_err function now has support for Heimdal's com_right_r 1158function(). (Addresses Sourceforge Bug: #2963865, Addresses Debian 1159Bug: #558910) 1160 1161 1162E2fsprogs 1.41.10 (February 7, 2010) 1163==================================== 1164 1165Fix resize2fs bug which causes it to access invalid memory. 1166 1167Add libss support for libreadline.so.6. 1168 1169Fix e2fsck's check for extent-mapped directory with an incorrect file 1170type. 1171 1172Add new e2fsck.conf configuration option, default/broken_system_clock 1173to support systems with broken CMOS hardware clocks. Also, since too 1174many distributions seem to have broken virtualization scripts now, 1175e2fsck will by default accept dates which are off by up to 24 hours by 1176default. (Addresses Debian Bugs: #559776, #557636) 1177 1178Fix a bug where mke2fs may not use the best placement of the inode 1179table when there is only room for a single block group in the last 1180flex_bg. 1181 1182E2fsck is now smarter when it needs to allocate blocks in the course 1183of fixing file system problems. This reduces the number of spurious 1184differences found in pass #5. 1185 1186E2fsck will no longer rehash directories which fit in a single 1187directory block. 1188 1189E2fsck now correctly handles holes in extent-mapped directories (i.e., 1190sparse directories which use extents). 1191 1192Fix big-endian problems with ext2fs_bmap() and ext2fs_bmap2(). 1193 1194Fix a bug in filefrag where on platforms which can allow file systems 1195with 8k blocks, that it doesn't core dump when it sees a file system 1196with 8k block sizes. (Thanks to Mikulas Patocka for pointing this 1197out.) 1198 1199E2fsck will correctly fix directories that are have an inaccurate 1200i_size as well as other problems in a single pass, instead of 1201requiring two e2fsck runs before the file system is fully fixed. 1202 1203Fix e2fsck so it will correctly find and detect duplicate directory 1204entries for non-indexed directories. ( Addresses Sourceforge Bug: 1205#2862551) 1206 1207Mke2fs will use BLKDISCARD to pre-discard all blocks on an SSD or 1208thinly-provisioned storage device. This can be disabled using the -K 1209option. 1210 1211Enhance libext2fs so it works around bug in Linux version 2.6.19 1212and earlier where the /proc/swaps file was missing the header on 1213the first line. 1214 1215Fix bug in Linux version 2.6.19 and earlier where the /proc/swaps file 1216was missing the header on the first line. 1217 1218Fix some big-endian bugs in e2fsck and libext2fs. 1219 1220Fix resize2fs so it works correctly on file systems with external 1221journals instead of failing early with the error "Illegal inode 1222number". 1223 1224Fix libss so that it does not seg fault when using a readline library 1225which does not supply a readline_shutdown() function. 1226 1227Updated dumpe2fs's usage message so it correctly gives the right 1228arguments summary for "-o superblock=<num>" and "-o blocksize=<num>". 1229(Addresses Launchpad Bug: #448099) 1230 1231Teach libext2fs to ignore the high 32 bits of the i_blocks field 1232when huge_file file system feature is set, but the inode does not 1233have the HUGE_FILE_FL flag set. 1234 1235Fix e2fsck's handling of 64-bit i_blocks fields. 1236 1237E2fsck will now print "Illegal indirect block", "Illegal 1238double-indirect block", etc., instead of "Illegal block #-1" or 1239"Illegal block #-2", etc. This makes it easier for users to 1240understand what has gone wrong. (Addresses SourceForge Bug: #2871782) 1241 1242Mke2fs now will obtain get device topology information from blkid and 1243use it to populate the superblock stride and stripe sizes. It will 1244also warn if the block device is misaligned 1245 1246Fix file descriptor leaks in fsck and debugfs. (Addresses Novell Bug: 1247##524526) 1248 1249Fix the libext2fs library code to round up the bitmap size to a 4-byte 1250boundary, to prevent spurious seg faults caused by the x86 1251architecture. This doesn't affect Linux systems, but was a major 1252problem on a number of *BSD systems. (Addresses Sourceforge Bug: 1253#2328708) 1254 1255Fix resize2fs's minimum size required for a file system so it doesn't 1256fail when "resize2fs -M" is run. (Addresses RedHat Bugzilla: #519131) 1257 1258Dumpe2fs now prints summary information about the contents of the 1259journal. 1260 1261Avoid printing scary error messages when e2fsck starts running 1262problems on low-memory systems, as it tends to panic and mislead the 1263user. (Addresses Debian Bug: #509529) 1264 1265Fix blkid's modules.dep parser so it handles compressed (.ko.gz) 1266modules files. (Address Red Hat Bug: #518572) 1267 1268Fix tune2fs so it can add a journal when an extent-enabled file system 1269is mounted. (Addresses Launchpad bug: #416648) 1270 1271Update Czech, Indonesian, Polish and Vietnamese translations (from the 1272Translation Project). 1273 1274Update/clarify man pages. (Addresses Sourceforge Bug: #2822186) 1275 1276Fixed various Debian packaging issues --- see debian/changelog for 1277details. (Addresses Debian Bugs: #540111) 1278 1279Programmer's Notes 1280------------------ 1281 1282The configure script supports the --with-cc, --with-ccopts, and 1283--with-ldopts options. Instead, the more standard use of CC=, 1284CCFLAGS=, and LDFLAGS= in the configure command line is used instead. 1285Also, --with-ld, which never worked, was also removed. 1286(Addresses Sourceforge Bug: #2843248) 1287 1288The in-tree header files are only used if the in-tree uuid or blkid 1289libraries are used. Otherwise, use the system-provided uuid or blkid 1290header files if using the system-provided libraries. 1291 1292Fix some build failures caused by --disable-* configure options. 1293 1294Work around a bug in autoconf 2.64. 1295 1296 1297E2fsprogs 1.41.9 (August 22, 2009) 1298================================== 1299 1300Fix a bug in e2fsck routines for reallocating an inode table which 1301could cause it to loop forever on an ext4 filesystem with the FLEX_BG 1302filesystem feature with a relatively rare (and specific) filesystem 1303corruption. This fix causes e2fsck to try to find space for a new 1304portion of the inode table in the containing flex_bg, and if that 1305fails, the new portion of the inode table will be allocated in any 1306free space available in the filesystem. 1307 1308Make e2fsck less annoying by only asking for permission to relocate a 1309block group's inode table once, instead of for every overlapping 1310block. Similarly, only ask once to recompute the block group 1311checksums, instead of once for each corrupted block group's checksum. 1312 1313Fix filefrag to avoid print the extent header if the FIEMAP ioctl is 1314not present, and it needs to fall back to using the FIBMAP ioctl. 1315 1316Fix filefrag to correctly print the number of extents for zero-length 1317files. (Addresses Debian Bug: #540376) 1318 1319Filefrag now has a -B option which forces the use of the FIBMAP ioctl 1320to more easily debug the FIBMAP code. 1321 1322Fixed filefrag for non-extent based files. 1323 1324Add a new program, e2freefrag, which displays information about the 1325free space fragmentation in an ext2/3/4 filesystem. 1326 1327Fix inode resizing via tune2fs -I so that it works correctly in the 1328face of non-empty bad blocks inodes, and if the filesystem was 1329formatted using the "mke2fs -E stride=N" option for RAID arrays. 1330 1331Fix regression in ext2fs_extent_set_bmap() caused e2fsck -fD to fail 1332and corrupt large directories if the directory needs to shrink by more 1333than one block. (Addresses Debian Bug: #537510) 1334 1335Fix e2fsck's buggy_init_scritps=1 so that the if the last write and/or 1336last mount times are in the future, they are corrected even if 1337buggy_init_scripts is set. This is needed because otherwise resize2fs 1338will refuse to resize the filesystem, even after running "e2fsck -f". 1339(Addresses Launchpad bug: #373409) 1340 1341E2fsck will now print much fuller information when the last mount time 1342or last written time is in the future, since most people can't seem to 1343believe their distribution has buggy init scripts, or they have a 1344failed CMOS/RTS clock battery. 1345 1346Enhance dumpe2fs to dump the extent information via the 'stat' 1347command, and more detailed extent information via the new command 1348'dump_extents'. 1349 1350Update French, Polish, Czech, and Sweedish translation from the 1351Translation Project. 1352 1353Fixed various Debian packaging issues --- see debian/changelog for 1354details. 1355 1356Programmer's Notes 1357------------------ 1358 1359Fixed miscellaneous gcc -Wall warnings. 1360 1361Fixed memory leak in error path in ext2fs_block_iterate2() 1362 1363Fixed non-Linux build of the intl directory by adding support for the 1364E/Q/V macros. 1365 1366The bitmap read/write functions now treat uninitialized bitmaps as 1367unallocated; this fixes a number of problems in all e2fsprogs for ext4 1368filesystems when there is a need to allocate new blocks or inodes, and 1369there aren't any free blocks or inodes in the already-used block 1370groups. 1371 1372Improve ext2fs_extent_set_bmap() to avoid creating new extents which 1373get inserted into the extent tree when they are not needed. 1374 1375 1376E2fsprogs 1.41.8 (July 11, 2009) 1377================================ 1378 1379Fix resize2fs's online resizing, fixing a regression which in 1380e2fpsrogs 1.41.7. (Addresses Debian Bug: #535452) 1381 1382Fix potential filesystem corruptions caused by using resize2fs to 1383shrinking ext4 filesystems with extents enabled. (Addresses Red Hat 1384Bug: #510379) 1385 1386Optimize uuid_generate() to avoid running uuidd if it is not setuid or 1387setgid and the currently running program doesn't have write access to 1388the uuidd work directory. 1389 1390Add safety checks (for non-Linux systems) so that uuidd isn't run with 1391file descriptors 0, 1, and 2 closed; and if they are closed, uuidd 1392will be careful not to close the file descriptor for its unix domain 1393socket when it detaches itself from the controlling tty. Also add 1394safety checks so that if the unix domain socket between the uuid 1395library and uuidd program is closed for any reason, both the library 1396and the uuidd will return an appropriate error code instead of looping 1397in an infinite loop. 1398 1399The e2croncheck script, which creates an LVM snapshot and then checks 1400the ext3/4 filesystem via the LVM snapshot, has been added to the 1401contrib directory. 1402 1403Fix filefrag program for files that have more than 144 extents. 1404 1405Update French, Polish, Czech, Indonsian, and Sweedish translation from 1406the Translation Project. 1407 1408Fixed various Debian packaging issues --- see debian/changelog for 1409details. (Addresses Debian Bug #535530) 1410 1411Update/clarify man pages. 1412 1413Programmer's Notes 1414------------------ 1415 1416Fix compilation problem when configured with --disable-uuid. 1417 1418Don't build uuidgen if configured with --disable-uuid. 1419 1420Add the new library function ext2fs_test_inode_bitmap_range(), and 1421optimized ext2fs_test_block_bitmap_range(), which will be needed for 1422future optimizations for e2fsck. 1423 1424Fix makefile dependencies for libcom_err so that the Makefiles work 1425well on non-GNU make program. (Addresses Sourceforge Patches: #2813809) 1426 1427Enhance the build system so that the full set of commands executed by 1428the Makefiles are displayed, instead of the Linux kernel summary 1429output, if the build was configured with --enable-verbose-makecmds, or 1430if GNU make is in use and the V variable is non-null, i.e., via "make 1431V=1". 1432 1433 1434E2fsprogs 1.41.7 (June 29, 2009) 1435================================ 1436 1437Fix a bug in libext2fs which can cause e2fsck and resize2fs to write 1438uninitalized data into the portion of the inode beyond the first 128 1439bytes when operating on inodes mapped via extents; potentially 1440corrupting filesystems. 1441 1442Fix memory leaks in e2fsprogs, including a very large memory leak 1443which can cause e2fsck to run out of memory when checking very large 1444filesystems using extents. 1445 1446The logsave program will now filter out the ^A and ^B characters when 1447writing to the console. 1448 1449Harden ext2fs_validate_entry() so that lsdel will not read beyond the 1450end of the block even if the directory block is corrupted. 1451 1452Fix debugfs from core dumping if the logdump command fails to open the 1453output file. 1454 1455Enhance badblocks to print the currently tested block number when 1456interrupted with ^C. 1457 1458Fix lsattr to exit with a non-zero status when it encounters errors. 1459(Addresses RedHat Bugzilla #489841) 1460 1461Fix e2fsprogs to use the same encoding for rec_len in 64k (and larger) 1462filesystems as the kernel when encoding rec_len >= 64k. For 64k 1463filesystems (currently all that is supported by e2fsprogs, this is 1464only a minor corner case). 1465 1466Resize2fs will now update the journal backup fileds in the sup0erblock 1467if the journal is moved; this avoids an unnecessary full fsck after 1468resizing the filesystem. (Addresses RedHat Bugzilla: #505339) 1469 1470Fix libext2fs to properly initialize i_extra_size when creating the 1471journal and resize inodes. 1472 1473Change badblocks to allow block sizes larger than 4k. 1474 1475Fix filefrag progam so it correctly checks for errors from the fiemap ioctl. 1476 1477Update Chinese and Czech translation from the Translation Project. 1478 1479Clean up various man pages. (Addresses Debian Bug #531385 and #523063) 1480 1481 1482Programmer's Notes 1483------------------ 1484 1485Add --disable-libuuid option to configure which uses an in-system 1486installed version of libuuid; the private version is enabled by 1487default. 1488 1489Add --valgrind-leakcheck option to the test_script program which runs 1490valgrind with the appropriate options to find memory leaks. 1491 1492 1493E2fsprogs 1.41.6 (May 30, 2009) 1494=============================== 1495 1496Fix a critical regression in e2fsck introduced version 1.41.5 which 1497can undo updates to the block group descriptors after a journal replay. 1498 1499If e2fsck receives an I/O error while replaying the journal, prompt 1500the user whether they want to ignore the error or not; if e2fsck is 1501run in preen mode, force fsck to abort the preen pass. 1502 1503Fix a bug which would cause e2fsck to core dump if the filesystem 1504contains a corrupt extent header, and the user declines to clear the 1505inode. (Addresses Sourceforge Bug: #2791794) 1506 1507Fix e2fsck to restart only once in the case of multiple block groups 1508which inodes in use that are in the uninitialized part of the block 1509group's inode table. 1510 1511To reduce user confusion, if the /etc/mtab file is missing 1512ext2fs_check_mount_point and ext2fs_check_if_mounted will return a 1513new, explicit error code to indicate this case. This will cause 1514e2fsck to give a clearer error message when the user is using buggy 1515rescue CD's that don't properly set up /etc/mtab. (Addresses Debian 1516Bug: #527859) 1517 1518Fix e2fsck so that if the primary group descriptors are corrupted, but 1519the backup superblock is entirely invalid, to go back to using (and 1520fixing) the priary group descriptors instead of completely giving up 1521on the filesystem. (Addresses Debian Bug: #516820) 1522 1523Change e2fsck to not abort a preen pass if an inode is found to have 1524i_file_acl_hi non-zero. Ext3 filesystems don't care, and newer 1525kernels (post 2.6.29) will ignore this field. So let's fix it 1526automatically during the preen pass instead of forcing the user to fix 1527this up manually. (Addresses Debian Bug: #526524) 1528 1529Add resource tracking for e2fsck passes 1b through 1d. 1530 1531Speed up e2fsck by eliminating unnecessary journal checks if the 1532filesystem is already mounted and doesn't need recovery (since the 1533kernel would have run the journal when the filesystem was mounted.) 1534Also speed up e2fsck by avoiding unnecessary block group descriptor 1535updates in ext2fs_close(). 1536 1537Add support to chattr to migrate inodes from using direct/indirect 1538blocks to extents. 1539 1540Avoid corrupting the filesystem if there is an attempt to shrink a 1541filesystem using resize2fs smaller than posible by making 1542ext2fs_set_bmap() more careful not to delete the old block until the 1543new block can be inserted. In addition, fix a bug in how the minimum 1544size of the filesystem (plus a safety margin) is calculated, and 1545modify resize2fs to refuse to shrink the filesystem below that minimum 1546size without the force flag. 1547 1548Teach blkid to try to figure out DM device names using take advantage 1549of information in sysfs when running on kernels (2.6.29 and later) 1550which provide this information in /sys/block/dm-<N>/dm/name; this is 1551much faster than scanning for the devncie number in /dev/mapper. 1552 1553Fix blkid to prefer /dev/mapper/<name> device names instad of the 1554private /dev/dm-X names. 1555 1556Add an -a option to debugfs's close_filesys command which writes any 1557changes to the superblock or block group descriptors to all of the 1558backup superblock locations. 1559 1560Add support to the filefrag program to use the FIEMAP ioctl. 1561 1562Update Chinese translation from the Translation Project. 1563 1564Clean up various man pages. (Addresses Red Hat Bugzilla: #502971 and 1565Launchpad Bug: #381854) 1566 1567Fixed various Debian packaging issues --- see debian/changelog for 1568details. (Addresses Debian Bug #506064) 1569 1570 1571Programmer's Notes 1572------------------ 1573 1574Add test code to make sure e2fsck's problem.c doesn't have two problem 1575codes assigned to duplicate values. 1576 1577Avoid using a hard-coded path for /bin/rm in block's test_probe.in. 1578(Addresses Sourceforge Bug: #2780205) 1579 1580Clean up e2fsck by removing #ifdef RESOURCE_TRACK by adding an empty 1581function declaration for init_resource_track() and 1582print_resource_track() when RESOURCE_TRACK is not defined. 1583 1584The test code which is used to build the tst_csum progam has been 1585moved from from tst_csum.c into csum.c under an #ifdef DEBUG to to 1586avoid compile problems caused by not having a prototype for 1587ext2fs_group_desc_csum(). (Addresses Sourceforge Bug #2484331) 1588 1589Update the config.guess and config.sub file to the latest from the 1590FSF, to allow e2fsprogs to build on the avr32 platform. (Addresses 1591Debian Bug: #528330) 1592 1593Add a new function, ext2fs_extent_open2(), which behaves like 1594pext2fs_extent_open(), but uses the user-supplied inode structure 1595when opening an extent instead of reading the inode from disk. 1596 1597 1598E2fsprogs 1.41.5 (April 23, 2009) 1599================================= 1600 1601Fix a number of filesystem corruption bugs in resize2fs when growing 1602or shrinking ext4 filesystems off-line (i.e., when the ext4 filesystem 1603is not mounted). 1604 1605Debugfs can now set i_file_acl_high via the set_inodes_field command, 1606and print a 64-bit file acl. This is useful for debugging filesystem 1607corruptions where the high bits of i_file_acl_high are set. E2fsck 1608will detect and fix non-zero i_file_acl_high on 32-bit filesystems 1609since some Linux kernel versions pay attention to this field even when 1610they shouldn't. 1611 1612Speed up e2fsck when checking clean filesystems by avoiding 1613unnecessary block reads, and coalescing the block group descriptor 1614blocks so they are read using a single read operation. 1615 1616The libuuid library will now close all file descriptors before running 1617uuidd. This avoids problems when the calling program has open sockets 1618which then never get closed because uuidd is a long-running helper 1619daeomn. (Addresses Launchpad bug: #305057) 1620 1621In order to avoid unnecessary full filesystem checks by e2fsck after 1622an on-line resize, e2fsck will ignore the NEEDS_RECOVERY flag set on 1623the backup superblocks. (Addresses Red Hat Bugzilla: #471925) 1624 1625Mke2fs will avoid trying to create the journal when run in 1626superblock-only mode (mke2fs -S), since the left-over journal in the 1627inode table will cause mke2fs to fail. 1628 1629Fix a bug in libext2fs functions that check to see if a particular 1630device or filesystem image is mounted, which would cause these 1631functions to report that a file identical to the (relative) pathname 1632used by a pseudo-filesystem was mounted when in fact it was not. 1633 1634Update Czech translation from the Translation Project. 1635 1636Add Chinese (simplified) translation from the Translation Project. 1637 1638Fix support for external journals (which was broken in e2fsprogs 16391.41.4). 1640 1641Fix a regression in debugfs where the "stat" command when no 1642filesystem was open would cause debugfs to crash with a segmentation 1643violation. 1644 1645Starting in the 2.6.29 linux kernel, the ext4 filesystem driver can be 1646used to support filesystems without a journal. Update the blkid 1647library so it understands this. 1648 1649The blkid library will remove an entry from the blkid cache 1650representing the entire disk if partitions are found, since presuambly 1651the device previously had no partition table, but has now transitioned 1652to using a partition table. 1653 1654Add a check to mke2fs and tune2fs that the argument to the -m option 1655(which specifies the reserved ratio) must be greater than zero. 1656(Addresses Debian Bug: #517015) 1657 1658Add support for tracking the number kilobytes written to the 1659filesystem via the superblock field s_kbytes_written. It will be 1660updated by the kernel as well as by e2fsprogs programs which write to 1661the filesystem. This is useful for tracking the wear to filesystems 1662on Solid Sstate Drives. 1663 1664Fix compatibility issue in the libext2fs info file and makeinfo 1665version 4.12. (Addresses Red Hat Bugzilla: #481620) 1666 1667Update/clarify man pages. (Addresses Debian Bug: #515693, #365619) 1668 1669Fixed various Debian packaging issues --- see debian/changelog for 1670details. (Addresses Debian Bug: #506279) 1671 1672Programmer's Notes 1673------------------ 1674 1675Fix Hurd compilation problem in e2fsck and tune2fs (Addresses Debian 1676Bug: #521602) 1677 1678Fix various gcc compilation warnings and other programming cleanups. 1679(Addresses Red Hat Bugzilla: #486997) 1680 1681Add support for building the blkid command statically. 1682 1683Add support for disabling the built-in blkid library in favor of a 1684system-installed blkid implementation via the configure option 1685--disable-libblkid. 1686 1687 1688E2fsprogs 1.41.4 (January 27, 2009) 1689=================================== 1690 1691Fixed a bug which could sometimes cause blkid to return an exit value 1692of zero for a non-existent device (Addresses Debian Bug: #502541) 1693 1694Blkid will now recognize ext3 filesystems that have the test_fs flag 1695set as ext3 filesystems. 1696 1697The blkid library will now recognize btrfs filesystems and swap 1698devices currently used by user-level software suspend. 1699 1700Tune2fs now updates the block group checksums when changing the UUID 1701to avoid causing e2fsck to complain vociferously at the next reboot. 1702 1703Tune2fs's inode size resizing algorithms have been fixed so it is not 1704vastly inefficient for moderate-to-large filesystems, due to some 1705O(n**2) and O(n*m) algorithms that didn't scale well at all. 1706 1707Fix tune2fs's inode resizing algorithm so it will not corrupt 1708filesystems laid out for RAID filesystems; in addition, tune2fs will 1709refuse to change the inode size for filesystems that have the flex_bg 1710feature enabled. (This is a limitation in the current implementation 1711of tune2fs -I.) 1712 1713E2fsprogs 1.41 broke debugfs's logdump command for normal ext3/4 1714filesystems with 32-bit block numbers, when the headers for 64-bit 1715block numbers was added. This regression has been fixed. 1716 1717Debugfs's ncheck command has been fixed to avoid printing garbage 1718characters at the end of file names. 1719 1720Fix resize2fs for ext4 filesystems. Some blocks that that need moving 1721when shrinking filesystems with uninit_bg feature would not be moved. 1722In addition, blocks and inode table blocks were not being correctly 1723freed when shrinking filesystems with the flex_bg feable, which caused 1724resize2fs -M to fail. Finally, when blocks are moved, make sure the 1725uninitialized flag in extents is preserved. 1726 1727Fix bug which caused dumpe2fs to abort with an error if run on a 1728filesystem that contained an external journal. 1729 1730Some distributions used "mke3fs" as an alias for "mkfs.ext3"; check 1731for this in argv[0] to provide better legacy support for these 1732distributions. This is a practice that should NOT be continued, 1733however. 1734 1735Mke2fs now has a new option -U, which allows the user to specify the 1736UUID that should be used for the new filesystem. 1737 1738Mke2fs will treat devices that are exactly 16TB as if they were 16TB 1739minus one block. This allows users who have read that ext3 supports 1740up to 16TB filesystems and who create a 16TB LVM to not get confused, 1741since the true limit is really 16TB minus one block. 1742 1743E2fsck will no longer abort an fsck run if block group has an errant 1744INODE_UNINIT flag. 1745 1746E2fsck now distinguishes between fragmented directories and fragmented 1747files in verbose mode statistics and in the fragcheck report. 1748 1749Fix a bug in e2fsck which casued it double count non-contiguous 1750extent-based inodes. 1751 1752E2fsck will leave some slack space when repacking directories to allow 1753room for a few directory entries to be added without causing leaf 1754nodes to be split right away. 1755 1756Fix a bug which caused e2fsck to crash when it comes across a 1757corrupted interior node in an extent tree with the error message: 1758"Error1: Corrupt extent header on inode XXXXXX" 1759 1760E2fsck problem descriptions involving the journal are no longer 1761referred to as "ext3" problems, since ext4 filesystems also have 1762journals. 1763 1764Fix a long-standing bug in e2fsck which would cause it to crash when 1765replying journals for filesystems with block sizes greater than 8k. 1766 1767Update Catalan translation from the Translation Project. 1768 1769Fixed various Debian packaging issues --- see debian/changelog for 1770details. (Addresses Debian Bugs: #503057, #502323, #511207) 1771 1772Programmer's Notes 1773------------------ 1774 1775Fix build of e2fsck.profiled, and add support for building profiled 1776binaries in the misc directory if configured with --enable-profile. 1777 1778The ext2fs_open() function now performs more sanity checks on the 1779superblock to avoid potential divide by zero errors by other parts of 1780library. 1781 1782The ext2fs_read_inode_full() function now has a safety check to avoid 1783a segmentation fault on corrupted filesystems. 1784 1785The ext2fs_new_inode() function now has a sanity check so that if the 1786s_first_inode field in the superblock is insane, it will return 1787EXT2_ET_INODE_ALLOC_FAIL instead of returning an invalid inode number. 1788 1789To avoid segmentation faults, ext2fs_block_alloc_stats() and 1790ext2fs_inode_alloc_stats() now validates the passed inode or block 1791number to avoid overrunning an array boundary. 1792 1793Various signed/unsigned errors for variables containing block numbers 1794have been fixed. 1795 1796Accomodations for gcc's stupidity in not realizing that constant 1797strings that do not contain a '%' character are safe to use in format 1798strings have been made so that distributions that want to compile 1799e2fsprogs with -Werror=format-security have an easier time doing so. 1800 1801Added a new 64-bit getsize interface, ext2fs_get_device_size2(). 1802 1803Added the utility make-sparse.c to the contrib directory. 1804 1805The ext2fs_block_iterate2() function now reflects errors from 1806ext2fs_extent_set_bmap() to the caller, if the callback tries to 1807change a block on an extent-based file, and ext2fs_extent_set_bmap() 1808fails for some reason (for example, there isn't enough disk space to 1809split a node and expand the extent tree. 1810 1811The ext2fs_block_iterate2() function will preserve the uninit flag in 1812extents when the callback function modifies a block in an extent-based 1813file. 1814 1815E2fsck will now flag filesystems that have an insane s_first_ino field 1816in their superblock, and attempt to use a backup superblock to repair 1817the filesystem. 1818 1819 1820E2fsprogs 1.41.3 (October 12, 2008) 1821=================================== 1822 1823E2fsck has been fixed so it prints the correct inode number for 1824uinit_bg related problems. 1825 1826E2fsck will now offer to clear the test_fs flag if the ext4 filesystem 1827is available on linux. This can be disabled via a configuration 1828option in /etc/e2fsck.conf. 1829 1830Fix a file descriptor leak in libblkid when checking to see if an ext4 1831or ext4dev module exists. 1832 1833Fix a bug in e2fsck where in preen mode, if there are disk I/O errors 1834while trying to close a filesystem can lead to infinite loops. 1835(Addresses Red Hat Bugzilla #465679) 1836 1837Fix a bug in resize2fs where passing in a bogus new size of 0 blocks 1838will cause resize2fs to drop into an infinite loop. (Addresses Red 1839Hat Bugzilla: #465984) 1840 1841Add a check in the Unix I/O functions in libext2fs so that when a 1842device is opened read/write, return an error if the device is 1843read-only using the BLKROGET ioctl. 1844 1845Fix debugfs's ncheck command so that it prints all of the names of 1846hardlinks in the same directory. 1847 1848Fix a bug in libblkid so it correctly detects whether the ext4 and 1849ext4dev filesystems are available, so that the ext4dev->ext4 1850fallback code works correctly. 1851 1852Programmer's Notes 1853------------------ 1854 1855Fix a parallel build problem by making sure util/subst is built before 1856trying to build the lib/et directory. (Addresses Sourceforge Bug: 1857#2143281) 1858 1859Updated "make depend" information for crc16.o 1860 1861 1862E2fsprogs 1.41.2 (October 2, 2008) 1863================================== 1864 1865Fix e2fsck's automatic blocksize detection. This fixes a regression 1866from e2fsprogs 1.40.7 which caused e2fsck to fail if the user 1867specifies a block number using the -b option if the blocksize option 1868isn't also specified using -B. Unfortunately, users very commonly 1869invoke e2fsck using "e2fsck -b 32768 /dev/hdXXX" to use the backup 1870superblock; in fack e2fsck will often suggest this kind of command 1871line. Oops. 1872 1873Enhance the debugfs's "ncheck" command so it will print all of the 1874pathnames for the specified inodes. (Previously, in some cases ncheck 1875might not print a pathname for an inode at all if some of the other 1876inodes had multiple hard links.) 1877 1878Enhance debugfs's "hash" command so the hash seed can be specified via 1879a command-line option. In addition, allow the hash algorithm to be 1880specified by name instead of just by number. 1881 1882Fix e2fsck so that we don't accidentally print the translation file's 1883header when asking the user a custom question so there is no prompt 1884defined for a particular problem record. For example, the question 1885"Run journal anyway" will get the PO header tacked on because e2fsck 1886erroneously passed the null string to _(). (Addresses Launchpad Bug: 1887#246892) 1888 1889Enhance badblocks so that it can test a normal file which is greater 1890than 2GB. 1891 1892Enhance the badblocks command so that it displays the time and 1893percentage complete when in verbose mode. (Addresses Debian Bug: 1894#429739) 1895 1896Fix a potential memory leak in a error handling path in debugfs's 1897ncheck function. 1898 1899Fix a potential memory corruption problem if a memory allocation fails 1900in resize2fs. 1901 1902Fix the usage message for debugfs's logdump command to be consistent 1903with its man manpage. 1904 1905Update Polish, French, Vietnamese, Dutch, Indonesian, German, Czech, 1906and Sweedish translation from the Translation Project. 1907 1908Add documentation for the file I/O functions to the libext2fs.texinfo 1909file. (Addresses Debian Bug: #484877) 1910 1911Update and clarified various man pages. (Addresses Launchpad Bug 1912#275272; Addresses Debian Bugs: #498100, #498101, #498102, #498103) 1913 1914Fixed various Debian packaging issues --- see debian/changelog for 1915details. (Addresses Debian Bug: #497619) 1916 1917Programmer's Notes 1918------------------ 1919 1920Fix a potential file descriptor leack in libcom_err by setting the 1921close-on-exec flag for a fd used for debugging. (Addresses Red Hat 1922Bugzilla #464689) 1923 1924Fix a potential race in libcom_err by using sem_post/sem_init. SuSE 1925has been carrying a patch for a long time to prevent a largely 1926theoretical race condition if a multi-threaded application adds and 1927removes error tables in multiple threads. Unfortunately SuSE's 1928approach breaks compatibility by forcing applications to link and 1929compile with the -pthread option; using pthread mutexes has 1930historically been problematic. We fix this by using sem_post/sem_init 1931instead. 1932 1933Fix e2fsprogs-libs build failure due to 'subs' target. (Addresses 1934Sourceforge Bug: #2087502) 1935 1936Avoid linking e2initrd_helper, debugfs, blkid, and fsck with unneeded 1937libraries when using ELF shared libraries. 1938 1939Fix ELF shared library when building on systems that don't already 1940have the e2fsprogs shared libraries already installed. (Addresses 1941Sourceforge Bug: #2088537) 1942 1943Fix the pkg-config files so they work correctly when linking with 1944static libraries and fix the include directory so programs don't have 1945to use #include <ext2fs/ext2fs.h>, but can use #include <ext2fs.h> 1946instead. (Addresses Sourceforge Bug: #2089537) 1947 1948Make sure ext2fs_swab64() is compiled for all platforms, and not just 1949for x86. (Addresses Debian Bug: #497515) 1950 1951Remove the unused ext2fs_find_{first,ext}_bit_set() functions for all 1952non-x86 platforms. (They had been removed for x86 earlier.) 1953 1954Fix diet libc compilation support, which had bitrotted due to lack of 1955TLC. Fixing this improves general portability. 1956 1957When installing the link library when using ELF shared libraries, 1958avoid using absolute pathnames if the link library and the shared 1959library are installed in the same directory. (Addresses Sourceforge 1960Bug: #1782913) 1961 1962Fix gen-tarball so it will work even if the top-level directory has 1963been renamed to something other than "e2fsprogs". Also make 1964gen-tarball print the size of the resulting tar.gz file. 1965 1966 1967E2fsprogs 1.41.1 (September 1, 2008) 1968==================================== 1969 1970Many people are forgetting to update their mke2fs.conf file, and this 1971causes ext3, ext4, and ext4dev filesystems won't get created with the 1972proper features enabled. We address this in two ways. First, mke2fs 1973will issue a warning if there is not definition for an ext3, ext4, or 1974ext4dev filesystem and the user is trying to create such a filesystem 1975type. Secondly, when installing from a source build, "make install" 1976will provide basic configuration file handling for /etc/mke2fs.conf. 1977If it exists, and does not mention ext4dev, it will be moved aside to 1978/etc/mke2fs.conf.e2fpsrogs-old and the new /etc/mke2fs.conf file will 1979be installed. If the existing /etc/mke2fs.conf file does mention 1980ext4dev, then "make install" will install official mke2fs.conf file as 1981/etc/mke2fs.conf.e2fsprogs-new and issue a message to the user that 1982they should look to see if any changes need to be merged. 1983 1984The mke2fs program will now create the journal in the middle of the 1985filesystem, since this minimizes seek times on average for fsync-heavy 1986workloads. In addition, mke2fs will now create journals using extents 1987for filesystems that support them. This results in a more efficient 1988encoding for the journal since it eliminates the need for using 1989indirect blocks. 1990 1991The mke2fs program will avoid allocating an extra block to the 1992journal. (Addresses Sourceforge Bug: #1483791) 1993 1994Mke2fs will correctly enforce the prohibition against features 1995(specifically read-only features) in revision 0 filesystems. (Thanks 1996to Benno Schulenberg for noticing this problem.) 1997 1998Mke2fs previously would occasionaly create some slightly non-optimally 1999placed inode tables; this bug has been fixed. 2000 2001The mke2fs and tune2fs programs now print the correct usage message 2002describing the maximum journal size. (Addresses Debian Bug: #491620) 2003 2004Add support for setting the default hash algorithm used in b-tree 2005directories in tune2fs (from a command-line option) or mke2fs (via 2006mke2fs.conf). In addition, change the default hash algorithm to 2007half_md4, since it is faster and better. 2008 2009The blkid library will now recognize MacOS hfsx filesystems, and 2010correctly extract the label and uuid for hfs, hfsx, and hfsplus 2011filesystems. (Addresses Sourceforge Feature Requests: #2060292) 2012 2013The blkid library has improved detection of JFS and HPFS filesystems. 2014(Addresses Launchpad Bug: #255255) 2015 2016The blkid library is now much more efficiently handling devicemapper 2017devices, mainly by no longer using the devicemapper library. This can 2018speed up access for systems with a large number of device mapper 2019devices. 2020 2021Blkid had a number of cache validation bugs in libblkid that have been 2022fixed. (Addresses Debian Bug: #493216) 2023 2024Resize2fs will now properly close out the "updating inode references" 2025progress bar so there is a newline printed before printing the final 2026"resize is successful" message. 2027 2028Resize2fs will now correctly handle filesystems with extents and/o 2029uninitialized block groups correctly when file/directory blocks need 2030to relocated (i.e., when shrinking a filesystem or if the resize_inode 2031is not present). To support this, the ext2fs library now supports 2032initializing inode and block bitmaps that are not yet initialized when 2033allocating them using ext2fs_new_block() and ext2fs_new_inode(). In 2034addition, e2fs_block_iterate2() can now support changing the location 2035of interior nodes of an extent tree, and ext2fs_extent_set_bmap() has 2036been optimized to avoid creating unnecessary new extents when updating 2037the location of blocks in the extent tree. This will also help out 2038e2fsck's recovery of obscurely corrupted filesystems with extents, 2039when blocks are claimed by multiple inodes. 2040 2041Add support for on-line resizing ext4 filesystem with the flex_bg 2042filesystem feature. The method for doing so is not optimal, but to do 2043a better job will require kernel support. 2044 2045E2fsprogs 1.41.0 intrduced a bug in libext2fs which casued e2image and 2046debugfs programs to not be able to read e2image files; the signed 2047vs. unsigned bug in the code which read bitmaps from the e2image has 2048been fixed. (Addresses Debian Bug: #495830) 2049 2050Resize2fs is now correctly managing the directory in-use counts when 2051shrinking filesystems and directory inodes needed to be moved from one 2052block group to another. This bug has been around since e2fsprogs 20531.26, and is largely harmless, but does cause a filesystem corruption 2054which will be flagged by e2fsck after the filesystem has been shrunk. 2055 2056E2fsck will no longer issue spurious complaints about the inode size 2057caused by very large extent-based files, and by blocks reallocated 2058using fallocate() with the FALLOC_FL_KEEP_SIZE option. (Addresses 2059Kernel Bugzilla: #11341) 2060 2061Mke2fs will now set the creation timestamp on the lost+found directory 2062and the root directory. (More generally, all new inodes created using 2063the ext2fs library will correctly set the creation timestamp.) 2064 2065E2fsck now correctly calculates ind/dind/tind statistics in the 2066presence of extent-based files. In addition, "e2fsck -v" will report 2067statistics of the depth of extent trees in the filesystem. E2fsck can 2068also give an inode fragmentation report using "e2fsck -E fragcheck" 2069which can be useful when debugging the kernel block allocation 2070routines. 2071 2072Fix support for empty directory blocks in ext4 filesystems with 64k 2073blocksize filesystems. 2074 2075E2fsck will now print the depth of corrupt htree directories. 2076 2077Debugfs's htree command now correctly understands extent-based 2078directories. It will also print out the minor hash as well as the 2079major hash. 2080 2081Debugfs has a new command which will print the supported features of 2082e2fsprogs, to enable scripts to know whether the installed version of 2083e2fsprogs can support a specific feature. 2084 2085Debugfs will now write files using extents for filesystems that 2086support them. 2087 2088The error message printed by "tune2fs -I" if the inode size was too 2089small was rather confusing, so it has been improved. Also, we won't 2090try to create an undo log until we know that command-line-specified 2091parameters such as "tune2fs -I <inode size>" are valid. 2092 2093Given some filesystems found "in the wild" that had non-zero block 2094group checksums even though the uninit_bg/gdt_sum feature was not 2095enabled, e2fsck would issue spurious error messages. Teach 2096ext2fs_group_desc_csum_verify() to ignore the block group checksum 2097entirely if the feature flag is not set. (Addresses Debian Bug: 2098#490637) 2099 2100The blkid program will now print out a user-friendly listing of all of 2101the block devices in the system and what they contain when given the 2102-L option. (Addresses Debian Bug: #490527) 2103 2104The filefrag program now has a more accurate calculation for the 2105number of ideal extents. (Addresses Debian Bug: #458306) 2106 2107The test I/O manager is now enabled by default, but its overhead is 2108only incurred when it would be enabled via the TEST_IO_FLAGS or 2109TEST_IO_BLOCK environment variables. 2110 2111Typographical errors in various program strings and usage messages 2112have been fixed; most of these were pointed out by the e2fsprogs 2113message catalog translators. (Thanks, translators!) 2114 2115Update and clarified various man pages, as well as some typographical 2116errors in the libext2fs texinfo file. 2117 2118Fixed various Debian packaging issues --- see debian/changelog for 2119details. 2120 2121Add Indonesian and update French, Polish, Dutch, German, Sweedish, 2122Czech, and Vietnamese Translations. (Addresses Debian Bugs: #313697, 2123#401092) 2124 2125Programmer's Notes 2126------------------ 2127 2128Fix portability problem with the badblocks group; for systems that 2129don't have nanosleep(), try using usleep() instead. 2130 2131The "make check" target in the e2fsck directory now sets 2132LD_LIBRARY_PATH before running the various e2fsck internal library 2133regression tests. 2134 2135The crc32 regression test in the e2fsck library is now portable to 2136greater varienty of environments, including big-endian systems and 2137when cross-building e2fsprogs for embedded systems. (Addresses 2138Sourceforge Bug: #2019287) 2139 2140The ext2fs_extent_set_bmap() had some bugs when setting the first 2141block in a file, or when replacing a single block extent. Those cases 2142fortunately were came up relatively rarely when e2fsck was checking 2143files, but caused some problems when resize2fs was shrinking 2144extent-based files. 2145 2146Fix a potential core-dumping bug in libe2p's iterate_on_dir() 2147function. 2148 2149Various ext2fs library functions --- ext2fs_block_iterate2(), 2150ext2fs_initialize() and ext2fs_extent_open() --- now correctly free 2151allocated memory to avoid memory leaks in all of their error return 2152paths. 2153 2154Ext2ed was failing to build because masix support had been removed in 2155the rest of e2fsprogs, so ext2ed no longer has masix support, either. 2156 2157The configure script now respects the LDFLAGS environment variable if 2158it is set when configure is called. (Addresses Sourceforge Feature 2159Request: #1937287) 2160 2161Libuuid is now more portable to the Windows platform. (Addresses 2162Sourceforge Feature Request: #1937287) 2163 2164The configure script now uses AC_MSG_{RESULT,WARN,ERROR} instead of 2165bare echo commands so that configure flags such as --quiet work 2166correctly. (Addresses Sourceforge Patches: #2058794) 2167 2168A few uses of sprintf have been removed from the ext2fs library to 2169make life easier for bootloaders with a limited libc environment. 2170(Addresses Sourceforge Bug: #2049120) 2171 2172The ext2fs_read_inode() checks the validity of the inode number passed 2173to it earlier, to avoid doing some needless work when it would fail 2174anyway. 2175 2176The ext2fs_open() checks the validity of the blocksize parameter 2177passed to it earlier, to avoid doing some needless work when it would 2178fail anyway. 2179 2180Disable a very annoying automatic "%.sh -> %" GNU make rule in the 2181top-level Makefile. That automatic rule is used to better support 2182SCCS, but it caused problems for a particular niche distribution which 2183likes to use configure.sh files to store the configure options used to 2184build a package. Unfortuntaely GNU make will use the configure.sh to 2185replace the configure script, resulting in a self-inflicted fork bomb 2186leading to an out-of-memory crash. 2187 2188To support old GNU C compilers don't use C99/C++ comments, but only 2189K&R style comments, and don't try to use __builtin_expect if __GNUC__ 2190is less than 3. (__builtin_expect is only supported for gcc versions 21912.96 and up, and it's tricky to check for gcc 2.95 vs gcc 2.96; since 2192this is an optimization, we only try to use __builtin_expect for gcc 3 2193and up.) 2194 2195In e2fsck's crc routines, make sure we use WORDS_BIGENDIAN instead of 2196__LITTLE_ENDIAN, which are only defined by glibc's header files and 2197hence isn't portable. 2198 2199For the convenience for some distributions that need a static tune2fs, 2200the Makefile for misc/ now has a tune2fs.static target. 2201 2202The ext2fs_block_iterate2() function now supports BLOCK_FLAG_APPEND 2203for extent-based files 2204 2205The ext2fs_bmap() function now supports BMAP_ALLOC for extent-based 2206files. 2207 2208All source files no longer have any trailing white space. 2209 2210The io_channel_read_blk64() and io_channel_write_blk64() functions are 2211now functions instead of C preprocessor macros to provide better 2212forward compatibility. 2213 2214The e2fpsrogs translation template now expands the @x abbrevation. 2215 2216Various namespace leackages in libblkid, libe2p, and libext2fs have 2217been fixed. 2218 2219Fix a parallel build problem in e2fsprogs. 2220 2221E2fsprogs is now more portable to Solaris. 2222 * blkid no longer assumes that the TIOCGSIZE and TIOCGWINSZ ioctl's 2223 are always present. 2224 * Scripts do not assume that /bin/true is always in /bin 2225 * Don't use __FUNCTION__ since Solaris's C99 doesn't support it. 2226 * Flush stdio handles before calling setbuf(), since Solaris will 2227 discard any pending output to the stream. 2228 * Define _XOPEN_SOURCE to 600 since Solaris's header files are very 2229 picky about which C compiler can beused for SUSv3 conformance. 2230 Use of C99 is not compatible with SUSv2 (_XOPEN_SOURCE=500), 2231 and C89 is not compatible with SUSv3 (_XOPEN_SOURCE=600). 2232 Since we need some SUSv3 functions, consistently use SUSv3 so 2233 that e2fsprogs will build on Solaris using c99. 2234 * Solaris C99 does not support varargs C preprocessor macros 2235 * Solaris header files pollute the C namespace if in/netinet.h 2236 is included, which conflicts with e2fsprogs' use of the kmem_cache_t 2237 typedef. 2238 * Solaris ships with a pathetically ancient shell in /bin/sh so we 2239 avoid the use of various more avanced shell constructs such as $(). 2240 2241The "make rpm" command will now take some extra configure optiosn from 2242the build environment without needing to patch the source tree. 2243 2244The ext2fs_add_dir_block() function will now grow the dblist more 2245aggressively as an optimization to avoid copying the array too often. 2246 2247The e2fsck_write_bitmaps() will write the block and inode bitmaps 2248together instead of in two passes. 2249 2250 2251E2fsprogs 1.41 (July 10, 2008) 2252============================== 2253 2254Add support for ext4 filesystem features, in particular extents, 2255uninit_bg, flex_bg, huge_file, and dir_nlink features. Also add 2256support for checking journal checksums. Debugfs will print new 2257superblock and inode fields that were defined for ext4. For example, 2258the nanosecond and i_version fields of an inode, and the 2259s_min_extra_isize and s_wanted_extra_isize fields from the superblock. 2260Note: Resize2fs doesn't currently support the combination of flex_bg 2261and !resize_inode. (Addresses Debian Bug: #388452, #425477) 2262 2263Tune2fs can support migrating a filesystem from using 128 byte inodes 2264to 256 byte inodes, so it can take advantage of the full features of 2265ext4. 2266 2267Add support for "undo" support. E2fsck and mke2fs can optionally 2268record an undo log which can replayed by the program e2undo. 2269 2270E2fsck could damage a filesystem by trying to relocate inode tables 2271due to corrupted block group descriptors, where the attempted inode 2272table relocation would do far more harm than good. E2fsck has been 2273fixed to detect this these sorts of corrupted block group descriptors 2274much earlier in e2fsck processing, so it can try to use the backup 2275superblock and block group descriptors first. This should be a much 2276better strategy for recovering these types of corrupted filesystems. 2277(Addresses Sourceforge Bug: #1840291) 2278 2279E2fsck will display a more understandable message when the last check 2280field in the superblock is in the future. (Addresses Debian Bug: 2281#446005). 2282 2283E2fsck now performs more extensive and careful checks of extended 2284attributes stored in the inode. 2285 2286Enhance mke2fs to print a more explanatory error message when 2287ext2fs_get_device_size() returns EFBIG. (Addresses Debian Bug: 2288#488663) 2289 2290Fix mke2fs to use a default block size of 4k when formatting an 2291external journal device. This is done by using a fixed filesystem 2292type list that consists only of the single filesystem type "journal" 2293when looking up configuration keys in /etc/mke2fs.conf. (Addresses 2294Debian Bug: #488663) 2295 2296Speed up how mke2fs writes the journal data blocks by writing the disk 2297blocks in larger chunks. 2298 2299Fix blkid handling of stale devices. Fix a bug which could cause a 2300core dump while garbage collecting the blkid cache, and assure that 2301blkid_find_dev_with_tag() never returns a non-existent device. Also, 2302if a filesystem is found at a new /dev location, eliminate any 2303duplicate stale entries which can not be verified. 2304(Addresses Debian Bugs: #487758, #487783) 2305 2306Add more paranoid checks for LVM volumes and swap partitions in 2307blkid's probe function, to reduce the chances of false positives. 2308 2309The mke2fs program now has a much more sophisticated system for 2310controlling configuration parameters of a newly created filesystem 2311based on a split filesystem and usage type system. The -t option to 2312mke2fs was a deprecated alias to -c; it now specifies a filesystem 2313type (ext2, ext3, ext4, etc.), while the -T option can now be a comma 2314separated usage list. The filesystem type information and type 2315information is used to extract configuration parameters from the 2316/etc/mke2fs.conf file. 2317 2318The mke2fs program will no longer complain and request the -f option 2319when the user tries to create a filesystem with greater than 2**31 2320blocks. 2321 2322When creating a filesystem for the GNU Hurd use a fs-type of Hurd and 2323adjust the mke2fs.conf file so filesystems for the Hurd are created 2324with a blocksize of 4096 and inode size of 128, which is all it knows 2325how to handle. (Addresses Debian Bug: #471977) 2326 2327Mke2fs will always make sure that lost+found always has at least 2 2328blocks, even for filesystems with very large blocksizes (i.e., 64kb). 2329 2330Resize2fs will now print the minimum needed filesystem size if given 2331the -P option, and will resize the filesystem to the smallest possible 2332size if given the -M option. 2333 2334Fix resize2fs to clean up the resize_inode if all of the reserved gdt 2335blocks are consumed during an off-line resize. 2336 2337The "ls" command in debugfs now supports the -p option, which causes 2338it to quote the filenames so that spaces or tabs in directory entries 2339are easily visible. (Addresses Red Hat Bugzilla: #149480; Addresses 2340Sourceforge Feature Request: #1201667) 2341 2342Fix a potential off-by-one buffer oveflow in the fs_device_name in an 2343e2image file. 2344 2345The chattr program will return a non-zero exit code in case of 2346failures, and error messages can be suppressed with the -f option. 2347(Addresses Red Hat Bugzilla: #180596) 2348 2349Fix a bug in badblocks which caused it to overrun an array and likely 2350crash if more than 8 test patterns are specified using the -t option. 2351(Addresses Debian Bug: #487298) 2352 2353Add support to badblocks to limit how quickly it reads from the disk 2354drive (so it can be used for background scrubbing), and so it will 2355abort after finding a given number of errors. 2356 2357Remove support for the legacy big-endian filesystem format which only 2358existed on extremely long-dead PowerPC kernels almost a decade ago. 2359 2360Remove MASIX support from e2fsprogs. 2361 2362Add I/O statistics reporting to e2fsck. 2363 2364Update Vietnamese, Polish, French, Spanish, German, Catalan, Dutch, 2365Czech translations. 2366 2367Fixed various Debian packaging issues --- see debian/changelog for 2368details. (Addresses Debian Bugs: #487443, #487675, #490003) 2369 2370Fixed spelling mistakes, typos, and otherwise clarified man pages. 2371(Addresses Debian Bugs: #393313, #487849, #440983, #440981) 2372 2373Programmer's Notes 2374------------------ 2375 2376Factor out bitmap code in preparation for adding 64-bit new-style 2377bitmaps. 2378 2379Fix gcc -Wall warnings 2380 2381Fix the pkg-config files so that private librares are specified in 2382"Libs.private:". 2383 2384Fix the libext2fs.texinfo manual so it builds with modern versions of 2385texinfo. 2386 2387Silence the makefile from showing the awk command used to build the 2388 2389Clean up the badblocks group so to make it more portable and robust. 2390 2391Avoid using predictable filenames in /tmp in blkid's regression test 2392suite. Also remove bashism's in the regression test script. 2393 2394If the configure script is given --with-diet-libc, don't use thread 2395local storage, since diet libc doesn't support TLS. (Addresses 2396Sourceforge Bug: #2000654) 2397 2398Fix the blkid regression test suite to tolerate older versions of 2399mkswap that don't support the -U option. 2400 2401A few library routines have been converted to support 64-bit block 2402numbers; in particular, the I/O manager functions, the test_io, 2403inode_io, and unix_io managers have all be converted to support 64-bit 2404operation. 2405 2406Debugfs can now be extended for use by test programs. See 2407lib/ext2fs/extents.c for an example for how it can be used. The test 2408program links against the debugfs object files, and provides 2409additional commands by defining an auxiliary libss command table. 2410 2411The lazy_bg filesystem feature, which was only used by developer's 2412testing, has been removed since it has been largely supplanted by 2413uninit_bg. This also simplifies the code. 2414 2415 2416E2fsprogs 1.40.11 (June 17, 2008) 2417================================= 2418 2419Mke2fs, tune2fs, and resize2fs now use floating point to calculate the 2420percentage of reserved blocks. (Addresses Debian Bug: #452639) 2421 2422Updated Spanish and Catalan translations. 2423 2424Fixed various Debian packaging issues --- see debian/changelog for 2425details. (Addresses Debian Bugs: #483962, #483023) 2426 2427Add detection for ZFS volumes to the libblkid library. 2428 2429Fixed spelling mistakes, typos, and otherwise clarified man pages. 2430(Addresses Debian Bug: #486463) 2431 2432Programmer's notes: 2433------------------- 2434 2435Fix marginal C code in probe_lvm2() function to the blkid library more 2436portable for older compilers. 2437 2438Fix build problems on MacOS X. (Addresses Sourceforge Bug: #1972473) 2439 2440Fix ext2fs_swap{16,32,64} functions so they can be used by external 2441applications on big-endian machines. (Addresses Debian Bug: #484879) 2442 2443 2444E2fsprogs 1.40.10 (May 21, 2008) 2445================================ 2446 2447When deciding whether or not to revalidate a blkid cache entry, if the 2448device's mtime is newer than the last time the cached entry was 2449validated, force a revalidation. 2450 2451Fix a potential data corruption bug in e2fsck in the journal replay. 2452The chances of this is happening is extremely remote, especially the 2453default data=ordered or data=writeback modes. However, if a block 2454which has been journalled starts with the first four bytes 0xc03b3998, 2455when e2fsck replays the journal, those four bytes will be replaced 2456with zero's. Fortunately, it is highly, highly unlikely for e2fsck 2457metadata to begin with those fatal 4 byte sequence, and unless 2458data=ordered mode is in use, data blocks are never journaled. 2459 2460Updated German, Dutch, Sweedish, and Vietnamese translations. 2461 2462Programmer's notes: 2463------------------- 2464 2465Fixed various Debian packaging issues --- see debian/changelog for 2466details. 2467 2468Remove default sizes of types when cross compiling, since autoconf 24692.50 can figure this out automatically now. 2470 2471 2472E2fsprogs 1.40.9 (April 27, 2008) 2473================================= 2474 2475SuSE's security team audited uuidd and came up with a few minor 2476issues. None of them are serious given that uuidd runs setuid as a 2477unprivileged user which has no special access other than libuuid 2478directory, but it's good to get them fixed. 2479 2480One additional fix in ext2fs_swap_inode_full() needed for resize2fs to 2481work correctly with in-inode extended attributes. 2482 2483Updated German, Czech, Dutch, French, Polish, Sweedish, and Vietnamese 2484translations. 2485 2486Debugfs will avoid using a pager if the standard output is not a tty. 2487 2488Fix debugfs and tune2fs to correctly handle daylight savings time when 2489parsing a time string. 2490 2491Fixed spelling mistakes, typos, and otherwise clarified man pages. 2492 2493Fix fsck completion bars when multiple filesystems were being checked 2494in parallel. (Addresses Debian Bug: #432865, Addresses Launchpad Bug: 2495#203323, Addresses Sourceforge Bug: #1926023) 2496 2497Fix fsck so that progress information is sent back correctly when 2498multiple filesystems are being check and the output of fsck is being 2499redirected to a file descriptor. Also, include the device name (w/o 2500spaces) in the progress information sent back via a file descriptor. 2501(Addresses Launchpad Bug: #203323, Addresses Sourceforge Bug: 2502#1926023) 2503 2504Teach fsck to treat "ext4" and "ext4dev" as ext* filesystems. 2505 2506If logsave receives a SIGTERM or SIGINT signal, it will now pass that 2507signal to its child process. 2508 2509Fix mke2fs's creation of are resize inode when there is a non-standard 2510s_first_data_block setting. 2511 2512Fix bug in blkid when run by an unprivileged user; most devices were 2513not reported correctly. 9Addresses Launchpad Bug: #220275) 2514 2515Mke2fs will not allow the logically incorect combination of 2516resize_inode and meta_bg, which had previously caused mke2fs to create 2517a corrupt fileystem. 2518 2519Fix fsck in German locales so that a 'j' means yes. 2520(Addresses Sourceforge Bug: #1947683) 2521 2522 2523Programmer's notes: 2524------------------- 2525 2526Fixed various Debian packaging issues --- see debian/changelog for 2527details. 2528 2529Update valgrind options in test_script to work with valgrind 3.2.3 2530 2531Update texinfo.tex to a much newer version from the FSF. 2532 2533Remove bashism for configure script and from the lib/ss Makefile. 2534Addresses Sourceforge Bug: 1921969 2535 2536Fix some silently broken tests: m_no_opt, m_meta_bg, and m_raid_opt. 2537 2538Fix build system so that if texinfo is not installed, it won't print a 2539(harmless) error message. 2540 2541 2542E2fsprogs 1.40.8 (March 13, 2008) 2543================================= 2544 2545Fixed e2image -I so it works on image files which are larger than 2GB. 2546 2547Fixed e2fsck's handling of directory inodes with a corrupt size field. If 2548the size is larger than the number of blocks found in the inode, don't 2549try to allocate extra empty blocks at the end of the directory to make 2550up the difference; there's no point to doing that. In addition, if 2551the size is not a multiple of a blocksize, always fix it. 2552 2553E2fsck handled a pass 2 "should never happen error" by not giving 2554enough information and then core dumping. Unfortunately, it was all 2555too easy to trigger the "should never happen" situation if a 2556directory's inode size was not correct. This has been fixed, but 2557e2fsck has also been taught how to handle this situation more 2558gracefully, by simply removing the inode hash tree information, so 2559that it can be rebuilt again after e2fsck's pass 3. (Addresses 2560Launchpad Bug: #129395) 2561 2562Resize2fs had a bug resizing large inodes with extended attributes 2563that was fixed in 1.40.6; unfortunately, it turned out it wasn't fixed 2564completely on big-endian systems such as PowerPC. The bug should be 2565completely fixed now. Yay for regression test suites. (Addresses Red 2566Hat Bugzilla: #434893) 2567 2568Updated German, Czech, Dutch, Polish, Sweedish, and Vietnamese 2569translations. Many thanks to Philipp Thomas from Novell for stepping 2570up to become the new German translation maintainer! (Addresses Debian 2571Bugs: #302512, #370247, #401092, #412882). 2572 2573When e2fsck is clearing a corrupt inode's HTREE directory information, 2574make it clear that it is just clearing the HTREE information, not the 2575entire inode. 2576 2577Fixed spelling mistakes, typos, and otherwise clarified man pages. 2578 2579Programmer's notes 2580------------------ 2581 2582Add new functions, ext2fs_dblist_get_last() and 2583ext2fs_dblist_drop_last(), which allows the caller to examine the last 2584directory block entry added to the list, and to drop if it necessary. 2585 2586Fixed a portability problem in libblkid with DJGPP. 2587 2588Fix an obvious typo in an "internal error" message in e2fsck. Thanks 2589to Philipp Thomas for pointing this out. 2590 2591If the info files are not built, change "make install" so it doesn't 2592fail with an error code. 2593 2594 2595E2fsprogs 1.40.7 (February 28, 2008) 2596==================================== 2597 2598Remove support for clearing the SPARSE_SUPER feature from tune2fs, and 2599depreciate the -s option, since it can result in filesystems which 2600e2fsck can't fix easily. There are very good reasons for wanting to 2601disable sparse_super; users who wants to turn off sparse_super can use 2602debugfs. (Addresses Sourceforge Bug: #1840286) 2603 2604Add missing options to mke2fs's usage message. (Addresses Sourceforge 2605Bug: #1751393) 2606 2607Fix bug in resize2fs when large (greater than 128 byte) inodes are 2608moved when a filesystem is shrunk; it was only moving the first 128 2609bytes, so extended attributes were not getting moved. (Addresses Red 2610Hat Bugzilla: #434893) 2611 2612E2fsck now prints an explicit message when the bad block inode is 2613updated, to avoid confusion about why the filesystem was modified. 2614(Addresses Sourceforge Bug: #756460) 2615 2616Allow mke2fs and tune2fs manipulate the large_file feature. 2617Previously we just let the kernel and e2fsck do this automatically, 2618but e2fsck will no longer automatically clear the large_file feature. 2619It still isn't really necessary to worry about this feature flag 2620explicitly, but some users seem to care. (Addresses Red Hat Bugzilla: 2621#258381) 2622 2623Suppress message about an old-style fstab if the fstab file is empty. 2624(Addresses Debian Bug: #468176) 2625 2626Fix (really minor) bug in debugfs's find_free_block so it avoids 2627reporting a free block more than once if there are too few free blocks 2628in the filesystem. (Addresses Sourceforge Bug: #1096315) 2629 2630Change e2fsck to no longer clear the LARGE_FILES feature flag 2631automatically, when there are no more > 2GB files in the filesystem. 2632It's been almost a decade since there have been kernels that don't 2633support this flag, and e2fsck clears it quietly without telling the 2634user why the filesystem has been changed. 2635 2636Fix bug which could cause libblkid to seg fault if a device mapper 2637volume disappears while it is being probed. (Addresses RedHat 2638Bugzilla: #433857) 2639 2640Inhance e2fsck's reporting of unsupported filesystem feature flags. 2641(Addresses Sourceforge Feature Request: #1175808) 2642 2643Fix option syntax in dumpe2fs for explicit superblock and blocksize 2644parameters. What was currently documented in the man page has been 2645broken for some time, due to getopt() implementation changes. The 2646option sytax has been changed to one which is can be more portable 2647supported and which is consistent with the format for extended options 2648in mke2fs and tune2fs. (Addresses Sourceforge Bug: #1830994) 2649 2650Add support to tune2fs to clear the resize_inode feature. This 2651requires an fsck afterwards. (Addresses Red Hat Bugzilla: #167816) 2652 2653Teach blkid to detect LVM2 physical volumes. (Addresses Red Hat 2654Bugzilla: #409321) 2655 2656Add support for setting RAID stride and stripe-width via mke2fs and 2657tune2fs. Teach dumpe2fs to print the RAID parameters. 2658 2659Add support for setting new superblock fields to debugfs's 2660set_super_value. 2661 2662Add support for printing "mostly-printable" extended attributes in 2663Debugfs. 2664 2665Add support for the -M option to fsck, which causes it to ignore 2666mounted filesystem. 2667 2668Fix uuidd so that it creates the pid file with the correct pid number. 2669(Addresses Sourceforge Bug: #1893244) 2670 2671Fix various gcc -Wall warnings. 2672 2673Update Czech, Dutch, Polish, Sweedish, and Vietnamese translations 2674 2675Fixed spelling mistakes, typos, and otherwise clarified man pages. 2676(Addresses Sourceforge Patch: #1399325) 2677 2678 2679Programmer's notes: 2680------------------- 2681 2682Fixed various Debian packaging issues --- see debian/changelog for 2683details. (Addresses Debian Bug: #466929) 2684 2685Add new flag EXT2_FLAG_NONFREE_ON_ERROR ext2fs_open2() which returns a 2686partially filled-in filesystem object on an error, so that e2fsck can 2687print more intelligent error messages. 2688 2689Add a new function e2p_edit_feature2() which allows the caller to 2690specify which feature flags are OK to set or clear, and which returns 2691more specific information about feature flags which are not allowed to 2692be set/cleared. 2693 2694Set the C locale in the test_script driver since it uses [A-Za-z]. 2695(Addresses Sourceforge Bug: #1890526) 2696 2697Use fcntl locking instead of lockf in libuuid since Cygwin doesn't 2698support lockf(). 2699 2700Change configure.in to avoid using the 'dc' command unless it is 2701absolutely needed. (i.e., when using parsing a WIP-style version 2702number) (Addresses Sourceforge Bug: #1893024) 2703 2704Add portability checks to support compilation under DJGPP. 2705 2706Update to the latest samba tdb code before the LGPLv3 change, which 2707fixes a realloc() leak on failure. 2708 2709Fix memory leak in ext2fs_alloc_block(). 2710 2711Fix makefile dependency issues for various install targets. 2712(Addresses-Sourceforge-Patches: #1903484, #1903466, #1903456) 2713 2714Improve descriptions for the r_move_itable and r_resize_inode tests. 2715 2716 2717E2fsprogs 1.40.6 (February 9, 2008) 2718=================================== 2719 2720Add support for returning labels for UDF filesystems in the blkid 2721library. 2722 2723Fix bug in the blkid library where cached filesystems was not being 2724flushed when opening USB devices returned the error ENOMEDIUM. 2725(Addresses Debian Bug: #463787) 2726 2727Added logic to the blkid library to automatically choose whether a 2728filesystem should be mounted as ext4 or ext4dev, as appropriate. 2729 2730Allow tune2fs to set and clear the test_fs flag on ext4 filesystems. 2731 2732Fix a bug in e2fsck which caused it to core dump if e2fsprogs had been 2733configured with --enable-jbd-debug. 2734 2735Document the BLKID_FILE environment variable in the libblkid man page 2736 2737Programmer's Notes: 2738------------------- 2739 2740Update e2fsprogs translation template and Vietnamese and Czech translations 2741 2742Fixed various Debian packaging issues --- see debian/changelog for 2743details. (Addresses Debian Bugs: #436058) 2744 2745Don't try to create $DESTDIR/etc/init.d as part of make install as we 2746don't install the init.d script (and it's not the recommended way to 2747start uuidd anyway). (Addresses Sourceforge Bug: #1885085) 2748 2749Use thread local storage to fix a theoretical race condition if two 2750threads try to format an unknown error code at the same time in the 2751com_err library. 2752 2753 2754E2fsprogs 1.40.5 (January 27, 2008) 2755=================================== 2756 2757Fix a potential overflow big in e2image if the device name is too 2758long. 2759 2760Mke2fs will now create new filesystems with 256 byte inodes and the 2761ext_attr feature flag by default. This allows for much better future 2762compatibity with ext4 and speeds up extended attributes even on ext3 2763filesystems. 2764 2765Teach e2fsck to ignore certain "safe" filesystem features which are 2766set automatically by the kernel. Having the kernel set features 2767automagically behind the user's back is a bad idea, and we should try 2768to break the kernel of this habit, especially for the newer ext4 2769feature flags. But for now, we will try to avoid needless full checks 2770which can annoy users who are doing fresh installs. 2771 2772Add support in tune2fs and mke2fs for making a filesystem as being "ok 2773to be used with test kernel code". This will be needed for using test 2774filesystems with the latest ext4 kernel code. 2775 2776Change e2fsck -fD so that it sorts non-htree directories by inode 2777numbers instead of by name, since that optimizes performances much 2778more significantly. (Addresses-Sourceforge-Feature-Request: #532439) 2779 2780If e2image fills the disk, fix it so it exits right away instead of 2781spewing large numbers of error messages. 2782(Addresses-Sourceforge-Feature-Request: #606508) 2783 2784If ftruncate64() is not available for resize2fs, let it use ftrucate() 2785instead, but have it check to see if the size would get truncated, and 2786skip calling ftruncate in that case. 2787 2788Add support for detecting HFS+ filesystems in the blkid library. 2789 2790Add supprt in the blkid library for ext4/ext4dev filesystems. 2791 2792Fix a bug in blkid where it could die on a floating point exception 2793when presented with a corrupt reiserfs image. 2794 2795Fix blkid's handling of ntfs UUID's so that leading zeros are printed 2796such that UUID string is a fixed length. 2797 2798Add sample python bindings for the uuid library to the contrib 2799directory. (Addresses-Sourceforge-Patches: #778817) 2800 2801Fix debugfs's 'lsdel' command so it uses ext2fs_block_iterate2 so it 2802will work with large files. (Addresses Sourceforge Feature Request: 2803#1257500 and Sourceforge Support Request: #1253511) 2804 2805Allow the debugfs 'undel' command to undelete an inode without linking 2806it to a specific destination directory, since this might require 2807allocating blocks that could overwrite some yet-to-be-recovered 2808deleted files. (Addresses-Sourceforge-Feature-Request: #967141) 2809 2810Update Swedish translations from the Translation Project. 2811 2812Programmer's Notes: 2813------------------- 2814 2815Fix configure handling of --sbindir (which should rarely be used, but 2816someone did complain, so let's fix it). (Addresses Sourceforge Bug: 2817#498381) 2818 2819Updated e2fsprogs.spec file to include a new uuidd package 2820 2821Use pkg-config to determine where to find the devmapper library so we 2822can find out where it is located on different distributions. 2823 2824Fix Makefile race so that "make -j3 distclean" works correctly 2825 2826Fix portability problems on non-Linux/non-Hurd/non-Masix systems, 2827especially on MacOS X systems. (Addresses Sourceforge Bugs: #1861633, 2828#1819034, #1863819) 2829 2830Fixed spelling mistakes, typos, and otherwise clarified man pages. 2831 2832Fixed various Debian packaging issues --- see debian/changelog for 2833details. (Addresses Debian Bugs: #459403, #459475, #459614) 2834 2835Remove the --enable-dynamic-static configure option, and build e2fsck 2836dynamically by default. If the user wants e2fsck.static, he/she will 2837need to build it via "make -C e2fsck e2fsck.static" 2838 2839Fix various build warnings due to missing prototypes. 2840(Addresses Sourceforge Patch: #1861663, #1861659) 2841 2842 2843E2fsprogs 1.40.4 (December 31, 2007) 2844==================================== 2845 2846Improve time-based UUID generation. A new daemon uuidd, is started 2847automatically by libuuid if necessary. This daemon is setuid to allow 2848updates to /var/lib/libuuid, so the clock sequence number can be 2849stored and so if the clock is set backwards, it can be detected. 2850(Addresses Sourceforge Bug: #1529672, Addresses Red Hat Bugzilla: 2851#233471) 2852 2853Filter out the NEEDS_RECOVERY feature flag when writing out the backup 2854superblocks. This avoids e2fsck from concluding that a full 2855filesystem check is required before backing up the superblock due to 2856changes in the feature flags. (Addresses Debian Bug: #454926) 2857 2858Fix fsck to only treat the '#' character as a comment at the beginning 2859of the line in /etc/fstab. Otherwise fstabs for the fuse filesystem 2860will cause fsck to issue an bogus warning message. 2861(Addresses Gentoo bug: #195405, Addresses Sourceforge bug: #1826147) 2862 2863Format control characters and characters with the high eighth bit set 2864when printing the contents of the blkid cache, to prevent filesystems 2865with garbage labels from sending escape sequences to the user's screen 2866that might, for example place it in graphics mode. (Addresses Ubuntu 2867Bug: #78087) 2868 2869Fix sign-extension problem on 64-bit systems in in the com_err 2870library. (Addresses Sourceforge Bug: #1809658) 2871 2872Avoid division by zero error when probing an invalid FAT filesystem in 2873the blkid library. (Addresses Sourceforge Bug: #1831627) 2874 2875Update Dutch, Polish, and Vietnamese translations from the Translation 2876Project. Remove the Rwandan translation upon advice of the 2877Translation Project. 2878 2879Programmer's Notes: 2880------------------- 2881 2882Fix the libss "make check" regression test so that it works if the 2883current directory is not in the user's path or if the libss shared 2884library is not installed. (Addresses Sourceforge Bug: #1848974) 2885 2886Fixed spelling mistakes, typos, and otherwise clarified man pages. 2887(Addresses Debian Bugs: #444883, #441872) 2888 2889Fixed various Debian packaging issues --- see debian/changelog for 2890details. (Addresses Debian Bugs: #437720, #451172, #458017) 2891 2892Fix build failure on non-Linux/non-Hurd/non-Masix systems. 2893(Addresses Sourceforge Bug: #1859778) 2894 2895Fix Hurd portability issues. (Addresses Debian Bug: #437720) 2896 2897 2898E2fsprogs 1.40.3 (December 5, 2007) 2899=================================== 2900 2901Fix a potential security vulnerability where an untrusted filesystem 2902can be corrupted in such a way that a program using libext2fs will 2903allocate a buffer which is far too small. This can lead to either a 2904crash or potentially a heap-based buffer overflow crash. No known 2905exploits exist, but main concern is where an untrusted user who 2906possesses privileged access in a guest Xen environment could corrupt a 2907filesystem which is then accessed by the pygrub program, running as 2908root in the dom0 host environment, thus allowing the untrusted user to 2909gain privileged access in the host OS. Thanks to the McAfee AVERT 2910Research group for reporting this issue. (Addresses CVE-2007-5497.) 2911 2912Fix hueristics in blkid which could cause a disk without partitions to 2913be incorrectly skipped when a loopback device is present. (Addresses 2914Red Hat Bugzilla #400321.) 2915 2916Fix e2image so that in raw mode it does not create an image file which 2917is one byte too large. 2918 2919Change mke2fs's usage message so it recommends the preferred -E option 2920instead of the deprecated -R option. 2921 2922Enhance the blkid library so it will recognize squashfs filesystems. 2923(Addresses Red Hat Bugzilla #305151.) 2924 2925Enhance e2fsck so it will force the backup superblocks to be backed up 2926if the filesystem is consistent and key constants have been changed 2927(i.e., by an on-line resize) or by e2fsck in the course of its 2928operations. 2929 2930Enhance blkid's detection of FAT filesystems; so that USB disks with 2931only a single bootable partition will not get missed. 2932 2933E2fsck will no longer mark a filesystem as invalid if it has time 2934errors (i.e., if superblock mount time or last write time is in the 2935future) and the user refuses to fix the problem. 2936 2937The Ubuntu init scripts don't properly set the system time correctly 2938from hardware clock if the hardware clock is configured to tick local 2939time instead of GMT time. Work around this as best as we can by 2940providing an option, buggy_init_scripts, in /etc/e2fsck.conf which can 2941be set on Ubuntu systems. (Addresses Debian Bug #441093, and Ubuntu 2942Bug #131201.) 2943 2944Fix fsck to ignore /etc/fstab entries for bind mounts. (Addresses Red 2945Hat Bugzilla #151533.) 2946 2947Fix e2fsck so that if the superblock is corrupt, but still looks 2948vaguely like an ext2/3/4 superblock, that it automatically tries to 2949fall back to the backup superblock, instead of failing with a hard 2950error. 2951 2952Make the e2fsprogs program more robust so that they will not crash 2953when opening a corrupt filesystem where s_inode_size is zero. 2954 2955Change e2fsck so it uses sscanf() instead of atoi() so it non-numeric 2956arguments are detected as such and the parse error is reported to the 2957user. (Addresses Debian Bug #435381.) 2958 2959Change e2fsck so it will not complain if a file has blocks reallocated 2960up to the next multiple of a system's page size. 2961 2962Fix bug in ext2fs_check_desc() which will cause e2fsck to complain 2963about (valid) filesystems where the inode table extends to the last 2964block of the block group. (Addresses Red Hat Bugzilla #214765.) 2965 2966Fix a bug in ext2fs_initialize() which causes mke2fs to fail while 2967allocating inode tables for some relatively rare odd disk sizes. 2968(Addresses Red Hat Bugzilla #241767.) 2969 2970Add Catalan translation and update Dutch and Swedish translations 2971from the Translation Project. 2972 2973Fix big-endian byte-swapping bug in ext2fs_swap_inode_full(). We 2974still had an issue when trying to figure out whether we need to 2975byte-swap fast symlinks that contained extended attributes. 2976 2977Fixed spelling mistakes, typos, and otherwise clarified man pages. 2978(Addresses SourceForge Bug #1821333.) 2979 2980 2981Programmer's Notes: 2982------------------- 2983 2984Fix mke2fs tests to avoid needing any significant ^M (CR) characters 2985 2986Add "make check" to the RPM spec file 2987 2988Fix "make install" and 'make unstall" in misc/Makefile.in so that it 2989works correctly when the prefix is not the root directory. 2990 2991Fix the resize2fs tests, r_move_itable and r_resize_inode, so they 2992clena up after themselves by deleting the test.img temporary file 2993after completing the test. 2994 2995Fixed a corner case bug ext2fs_unlink() when trying to delete the 2996first directory entry in a directory block and the last directory 2997entry in the previous directory block is not in use. Fortunately 2998ext2fs_unlink() is only used by debugfs and e2fsck, and in e2fsck in 2999places where it is extremely unlikely to run into this corner case. 3000 3001Fix missing dependency which would cuase parallel builds to fail. 3002(Addresses Sourceforge Bug #1842331.) 3003 3004Fix a build error on newer gcc caused by lib/ext2fs/ismounted.c 3005calling open(O_CREATE) with a missing mode parameter. 3006 3007Fix the test_ss.c in lib/ss so it can be used as an example 3008application program for the library as well as a regression test 3009suite. 3010 3011Fix ext2fs_dblist_dir_iterate() so that error codes and abort codes 3012are properly passed back up through the call stack. 3013 3014E2fsprogs 1.40.2 (July 12, 2007) 3015================================ 3016 3017A recent change to e2fsck_add_dir_info() to use tdb files to check 3018filesystems with a very large number of filesystems had a typo which 3019caused us to resize the wrong data structure. This would cause a 3020array overrun leading to malloc pointer corruptions and segfaults. 3021Since we normally can very accurately predict how big the the dirinfo 3022array needs to be, this bug only got triggered on very badly corrupted 3023filesystems. 3024 3025Fix a bug in e2fsck which caused it to incorrectly salvange 3026directories when the last entry's rec_len is bogusly too big. This 3027resulted in a nonsense filesystem corruption to be reported, and 3028required a second run of e2fsck to fully fix up the directory. 3029 3030Update tune2fs man page to include more discussion of reserved blocks 3031(Addresses Launchpad bug #47817) 3032 3033Update Turkish, Polish, Dutch, and Vietnamese PO files from the 3034Translation Project 3035 3036 3037E2fsprogs 1.40.1 (July 7, 2007) 3038=============================== 3039 3040Fix bug which could cause libblkid to loop forever. When revalidating 3041a partition where there is obsolete information in /etc/blkid.tab, we 3042end up freeing a the type tag without clearing dev->bid_type, causing 3043blkid_verify() to loop forever. (Addresses Debian Bug: #432052) 3044 3045The Turkish translation has a bug in it where it has the translation 3046of "E@e '%Dn' in %p (%i)" to "E@E". This causes @E to be expanded at 3047@E, recursively, forever, until the stack fills up and e2fsck core 3048dumps. We fix this by making e2fsck stop @-expansions after a 3049recursive depth of 10, which is far more than we need. 3050(Addresses Sourceforge Bug: #1646081) 3051 3052Compile the default mke2fs.conf into mke2fs program. People are 3053getting surprised by mke2fs creating filesystems with different 3054defaults than earlier versions of mke2fs if mke2fs.conf is not 3055present. So we now create a built in version of mke2fs.conf file 3056which is used by mke2fs if the /etc/mke2fs.conf is not present. 3057(Addresses SourceforgeBug: #1745818) 3058 3059Improve the config/parse_types.sh helper script. Fix a potential 3060security problem if e2fsprogs is built as root (as Gentoo does!). In 3061addition fix the script and how it is called from the configure script 3062so that it does the right thing when cross-compiling. (Fixes Gentoo 3063bug: #146903) 3064 3065Update Vietnamese, French, and Dutch PO files from the Translation 3066Project. Also created a new e2fsprogs.pot file for translator. 3067 3068Fix bogus strip permission errors when building under Debian. When 3069building the e2fsprogs dpkg's, the dh_strip command emits a large 3070number of error messages caused by the permissions not being right. 3071So run dh_fixperms before running dh_strip. 3072 3073Programmer's Notes: 3074------------------- 3075 3076Add new function: profile_set_default(). This function sets the value 3077of the pseudo file "<default>". If the file "<default>" had 3078previously been passed to profile_init(), then def_string parameter 3079will be parsed and used as the profile information for the "<default>" 3080file. 3081 3082Fix mk_cmds's error reporting so that it is unambiguous that it is the 3083mk_cmds script which is generating the error. (Obviates Gentoo patch: 3084e2fsprogs-1.32-mk_cmds-cosmetic.patch) 3085 3086Fix the test suite to use LC_ALL instead of LANG. LC_ALL is the "high 3087priority" environment variable that overrides all others, where as 3088LANG is the lowest priorty environment variable. If LC_ALL is set, it 3089doesn't matter whether LANG, LANGUAGE, LC_COLLATE, LC_MESSAGES, and 3090the all the rest are set. This will assure that the locale when 3091running the test suites is the "C" locale. (Obviates Gentoo patch: 3092e2fsprogs-1.38-tests-locale.patch) 3093 3094 3095E2fsprogs 1.40 (June 29, 2007) 3096============================== 3097 3098Fix divide by zero error in blkid's NTFS probing logic. 3099 3100Add new blkid -g option which causes the blkid cache to be garbage 3101collected. 3102 3103Fix a bug in libblkid which could cause the internal field bid_type to 3104become corrupted. Fortunately bid_type isn't used much, and bid_label 3105and bid_uuid is only used by debugging code, so the impact of this bug 3106was very minor. 3107 3108Mke2fs will now store the RAID stride value when a filesystem is 3109created with a requested RAID stride, and then use it automatically in 3110resize2fs. 3111 3112Mke2fs has a sanity check added to make sure (inode_size * num_inodes) 3113isn't too big. In some cases Lustre users have tried specifying an 3114inode size of 4096 bytes, while keeping an inode ratio of one inode 3115per 4096 bytes. 3116 3117Improve sanity check in e2fsck's algorithm for finding a backup 3118superblock, so that it won't accidentally find a superblock that was 3119located in the journal, and then later reject it as being not a valid 3120backup superblock. 3121 3122Fix e2fsck get_size logic so that it will work with the Linux floppy 3123driver. The Linux floppy driver is a bit different from the other 3124block device drivers, in that if the device has been opened with 3125O_EXCL, it disallows another open(), even if the second open() does 3126not have the O_EXCL flag. (Addresses Debian Bug: #410569) 3127 3128Fix error checking of badblock's last-block and start-block arguments. 3129(Addresses Debian Bug: #416477) 3130 3131Fix e2fsck so that it doesn't overwrite the backup superblocks when 3132recovering a journal until the master superblock has been confirmed as 3133being sane. 3134 3135Change the blkid library to be much more paranoid about concluding 3136that a partition contains an NTFS filesystem, and fetch the UUID and 3137LABEL information from NTFS filesystems. (Addresses Launchpad Bug: 3138#110138) 3139 3140Factor out the code which sets the default journal size and move it 3141into libext2fs. 3142 3143Enhance e2fsck so it will recreate the ext3 journal if the original 3144journal inode was cleared to due it being corrupt after finishing the 3145filesystem check. 3146 3147Fix e2fsck so that it updates the journal inode if it is corrupted and 3148the backup journal information from the superblock was successfully 3149used to recover the filesystem. 3150 3151Fix e2fsck so that it checks all of the blocks in the journal inode 3152for validity. The original code only checked the direct blocks to 3153make sure the journal inode was sane. Unfortunately, if some or all 3154of the indirect or doubly indirect blocks were corrupted, this would 3155not be caught. 3156 3157Add support in blkid to detect LUKS encrypted partitions. 3158 3159Add extra sanity checks for extended attributes in the case where the 3160size is zero but the offset is very large. 3161 3162Fix byte-swapping issues for large inodes in ext2fs_read_inode_full() 3163and ext2fs_get_next_inode_full(). 3164 3165Clarify the copyright licenses used by the various libraries in 3166the top-level COPYING file (Red Hat Bugzilla: 166058) 3167 3168Make mke2fs's defaults when /etc/mke2fs.conf doesn't exist more sane. 3169 3170Fix mke2fs and debugfs to support large (> 16 bit) uid's and gid's. 3171 3172Remove check in e2fsck which requires EA's in inodes to be sorted; 3173they don't need to be sorted, and e2fsck was previously wrongly 3174clearing unsorted EA's stored in the inode structure. 3175 3176Allow mke2fs or tune2fs to create a substantially larger journal (up 3177to 10,240,000 blocks). 3178 3179Fix MD superblock detection, and make sure the correct UUID is 3180reported from the MD superblock. 3181 3182Fix a signed vs. unsigned bug in debugfs. 3183 3184Enhance debugfs's date parser so that it accepts integer values. 3185 3186Fix e2fsck's pass1c accounting so it doesn't terminate too early if a 3187file with multiply claimed blocks is hard linked. or not at all if the 3188root directory contains shared blocks 3189 3190Enhance debugfs so it can modify the block group descriptors using the 3191command set_block_group_descriptor. 3192 3193Improve e2fsck's reporting of I/O errors so it's clearer what it was 3194trying to do when an error happens 3195 3196Fix a bug in in how e2fsprogs byte swaps inodes containing fast 3197symlinks that have extended attributes. (Addresses Red Hat Bugzilla: 3198#232663 and LTC Bugzilla: #27634) 3199 3200Fix potential file descriptor leak in ext2fs_get_device_size() in an 3201error case. 3202 3203Add libreadline.so.5 support to libss. 3204 3205Impove badblocks -n/-w exclusive usage message. 3206 3207Fix dump_unused segfault in debugfs when a filesystem is not open 3208 3209Fix memory leak in blkid library. (Addresses Debian Bug: #413661) 3210 3211Allow the debugfs lcd command to work w/o a filesystem being open. 3212(Addresses LTC Bugzilla #27513) 3213 3214Fix e2fsck to clear i_size for special devices with a bogus i_blocks 3215field on the first pass. 3216 3217Fix e2fsck to set the file type of the '..' entry when connecting 3218a directory to lost+found. (Addresses Lustre Bug: #11645) 3219 3220Enhance e2fsck to recover directories whose modes field were corrupted 3221to look like special files. This is probably only useful in 3222artificial test cases, but it will be useful if we ever do the "inodes 3223in directory" idea for ext4. 3224 3225Allow debugfs to dump (and rdump) > 2GB files. (Addresses Debian Bug: 3226#412614) 3227 3228Fix resize2fs parsing of size parameter (in sector units). This was 3229actually a bug in libe2p's parse_num_blocks() function. When handling 3230the 's' suffix, it was ignoring the blocksize information passed in 3231from the caller and always interpreting the number in terms of a 1k 3232blocksize. (Addresses Debian Bug: #408298) 3233 3234There was a floating point precision error which could cause e2fsck to 3235loop forever on really big filesystems with a large inode count. 3236(Addresses Debian Bug: #411838) 3237 3238Fix memory leak in ext2fs_write_new_inode() 3239 3240Add support for using a scratch files directory to reduce e2fsck's 3241memory utilization on really big filesystems. This uses the TDB 3242library. See the [scratch_files] section of the e2fsck.conf man page 3243for more details. 3244 3245Fixed type-punning bug which caused dumpe2fs to crash on the Arm 3246platform (Addresses Debian Bug: #397044) 3247 3248Add explanatory message to badblocks that -n and -w are mutually exclusive 3249(Addresses Debian Bug: #371869) 3250 3251Allow debugfs and dumpe2fs to support fs features under development. 3252 3253Add support for the new flag EXT2_FLAG_SOFTSUPP_FEATURES flag to 3254ext2fs_open() , which allows application to open filesystes with features 3255which are currently only partially supported by e2fsprogs. 3256 3257Allow unix_io to support offsets greater than 2G (Addresses 3258SourceForge Bug: #1547922) 3259 3260Fixed overflow and signed/unsigned problems caused by the number of 3261blocks or inodes exceeding 2**31 or being close to 2**32-1. 3262 3263Add support for unsigned directory hash calculations with hints in the 3264superblock to fix cross-architectural portability for htree 3265directories with filenames where the high 8th bit is set. (Addresses 3266Debian: #389772) 3267 3268Fix resize2fs so that it gives user-intelligible error messages if the 3269filesystem or the kernel does not support on-line resizing. 3270(Addresses Debian Bug: #380548) 3271 3272Require mke2fs -F -F for really dangerous operations, since -F is 3273needed for less dangerous operations such as creating filesystems 3274images in regular files, or creating filesystems on whole block 3275devices. These relatively innocuous usages should NOT be confused 3276with running mke2fs on an apparently-mounted or in-use filesystem. 3277 3278Allow the default inode size to be specified into the mke2fs.conf 3279file. 3280 3281Make the smallest default journal size is big enough so that on-line 3282resizing should always work. 3283 3284Fix silly spelling error in e2fsck. (Addresses SourceForge bug: 3285#1531372) 3286 3287Fix debugfs coredump when lsdel is run without an open filesystem 3288(Addresses Debian Bug: #378335) 3289 3290Fix debugfs display bug us that bytes that have the high bit set are 3291displayed as "ec" instead of "ffffffec". 3292 3293Add support in lsattr so it will display the EXT4_EXTENTS_FL flag. 3294 3295Device mapper scanning wasn't working in the blkid library because the 3296pathnames had an extra "/dev" when they were being probed. 3297 3298Add GFS/GFS2 support to the blkid library. 3299 3300Fix blkid support of empty FAT filesystem labels. 3301 3302Avoid recursing forever (or for a long time) when the blkid library 3303searches for a device and there are symlinks to directories in /dev. 3304 3305Avoid unaligned halfword access in blkid when accessing FAT 3306superblocks, as this will cause Sparc/Solaris systems to throw a 3307SIGBUS error. 3308 3309The latest devmapper libraries requires pthreads, add -lpthreads to 3310the static link libraries for e2fsck.static if devmapper is enabled. 3311(Addresses Debian bug: #388718) 3312 3313Improve the (non-installed, for experts only) findsuper program by 3314printing the uuid and label from the superblocks, as well as the 3315starting and ending offsets of the filesystem given the information in 3316the superblock. Omit by default printing superblocks that are likely 3317found in located in an ext3 journal unless an explicit -j option is 3318given. 3319 3320Updated Spanish, French and Dutch translations and added Catalan 3321translation. (Addresses Debian bug: #411562) 3322 3323Use FreeBSD's DIOCGMEDIASIZE and DIOCGDINFO ioctls if available when 3324determining a partition's size, since binary searching to determine 3325the device doesn't work on FreeBSD. 3326 3327Documentation about UUID's is available in enough places, and it's 3328awkward to deal with debian-legal's insanities. So I'm caving in the 3329"more-lunatic-than-RMS" wing of Debian by removing RFC-4122 so we 3330don't have do the dfsg tarball. Also remove the rule that only tried 3331to install RFC-4122 on Ubuntu, since Ubuntu seems to want to fetch 3332e2fsprogs exclusively from Debian. (Addresses Debian Bug: #407107) 3333 3334Fix the info-dir line so that the menu name does not contain a .info 3335prefix. First of all, it's ugly, secondly, it causes the install-info 3336command to fail to remove the com_err info file from the 3337/usr/share/info/dir file when the comerr-dev package is removed and 3338purged. (Addresses Debian Bug: #401711) 3339 3340Fixed spelling mistakes, typos, and otherwise clarified man pages. 3341(Addresses Debian Bug: #369761, #373004, #379695) 3342 3343Fixed various Debian packaging issues --- see debian/changelog for 3344details. (Addresses Debian Bugs #389554, #390664, #413208, #419605, 3345#408352, #415560, #399155) 3346 3347 3348Programmer's notes: 3349------------------- 3350 3351E2fsck now supports the %It expansion when printing a problem report. 3352It will print the type of the inode in the problem context. 3353 3354Fix misc/Makefile.in so that it builds even if e2fsck hasn't been built yet 3355(Addresses Sourceforge Bug: #1565561) 3356 3357Remove unused variables and other lint/gcc -Wall cleanups 3358 3359Add check to ext2fs_get_device_size() so it will return EFBIG for for 3360filesystems contained in regular files where the filesystem image size 3361is returned by stat64(). 3362 3363Set local environment variables to C so mk_cmds and compile_et always 3364work. (Addresses SourceForge Bug: #1532177) 3365 3366Added the 64-bit byte swapping function ext2fs_swab64(). 3367 3368Added two new helper functions to prevent 2**31/2**32-1 overflow 3369problems: ext2fs_div_ceil() and e2p_percent(). 3370 3371Create new ext2fs library inline functions ext2fs_group_first_block() 3372and ext2fs_group_last_block() in order to calculate the starting and 3373ending blocks in a block group. 3374 3375Create the generated files read-only to remind developers not to edit them. 3376 3377Add support for autoconf 2.60 (with backwards compatibility for older 3378versions of autoconf). 3379 3380Added an "make rpm" target to top-level Makefile 3381 3382Added various FreeBSD portability fixes. 3383 3384Exclude mercurial files from the RPM build tree to speed up copy/build. 3385 3386Use root_sysconfdir to define the locations of mke2fs.conf and 3387e2fsck.conf instead of using a hard-coded /etc pathname. 3388 3389Prevent e2fsck.h and ext2_ext_attr.h from getting included multiple times. 3390 3391Fixed "make clean" in blkid's Makefile.in file from removing tst_*.c files. 3392 3393If diff -u is supported, use it to report test failures. 3394 3395Updates/improvements to RPM spec file 3396 3397Add on-disk format definitions for the following new features: 3398EXT4_FEATURE_RO_COMPAT_HUGE_FILE, EXT4_FEATURE_RO_COMPAT_GDT_CSUM, 3399EXT4_FEATURE_RO_COMPAT_DIR_NLINK, EXT4_FEATURE_INCOMPAT_64BIT, 3400EXT4_FEATURE_RO_COMPAT_EXTRA_ISIZE 3401 3402Add a new make target "checked-failed" in the tests directory which 3403reruns any failed tests 3404 3405Update draft-leach-uuids-guids-01.txt with rfc4122.txt 3406 3407Fix miscellaneous bugs reported by Coverity: Dead code, potential nul 3408pointer dereferences, memory leaks, etc. None were security-criticial 3409problems. 3410 3411Fix up usage and decrement error messages in the test_icount program 3412 3413Add debugging code to the com_err library; if the environment variable 3414COMERR_DEBUG is set to 1, print out debugging messages as error tables 3415are added and removed from the com_err library. If the 3416COMERR_DEBUG_FILE environment variable is set (and the process is not 3417setuid) the debugging messages may be redirected to a file. 3418 3419Change all of the e2fsprogs programs to use the newer add_error_table() 3420and remove_error_table() interfaces instead of the much older 3421initialize_*_error_table() function. 3422 3423Add TDB support into the ext2fs library. This allows us to have a 3424guaranteed library we can count on always being present so we can 3425store data in an on-disk database. 3426 3427Add support for using TDB to store the icount data, so we don't run out 3428of memory when checking really large filesystems. 3429 3430Change the regression test suite so that it skips empty test directories. 3431 3432Define the l_i_iversion field in ext2_inode. The l_i_version field is 3433now defined from the old l_i_reserved1 field in the ext2 inode. This 3434field will be used to store high 32 bits of the 64-bit inode version 3435number. 3436 3437Add Makefile production rule for e2fsprogs.spec in case it gets 3438deleted. 3439 3440Add new function profile_get_uint() to allow for a clean way to fetch 3441unsigned integers from the context. 3442 3443Add test to make sure the ext2 superblock structure is 1024 bytes. 3444 3445Fix typo in name of f_dup4 regression test 3446 3447Add new function blkid_gc_cache() which performs a garbage collection 3448pass on the /etc/blkid.tab file. 3449 3450The ext2fs_open() function now sets EXT2_FLAG_MASTER_SB_ONLY. In 3451general, only e2fsck (after the filesystem is clean), tune2fs, and 3452resize2fs should change the backup superblocks by default. Most 3453callers of ext2fs_open() should not be touching any superblock fields 3454which require the backups to be touched. 3455 3456Add new function to libext2fs, ext2fs_default_journal_size(), which 3457returns the default journal size. 3458 3459 3460E2fsprogs 1.39 (May 29, 2006) 3461============================= 3462 3463Fix 32-bit cleanliness in e2fsprogs so that we can support filesystems 3464between 2**31 and 2**32 blocks. 3465 3466Change mke2fs to use /etc/mke2fs.conf as a configuration file to 3467configure the filesystem features, blocksize, and inode_ratio for 3468different filesystem types. 3469 3470Mke2fs will now create filesystems hash trees and on-line resizing 3471enabled by default, based on the new /etc/mke2fs.conf file. 3472 3473The e2fsprogs tools (resize2fs, e2fsck, mke2fs) will open the 3474filesystem device node in exclusive mode to prevent accidents by 3475system administrators. In the case of resize2fs and mke2fs, it will 3476only use exclusive mode if the filesystem is not mounted. 3477 3478Fixed a bug in mke2fs which caused it to to fail when creating the 3479resize inode for large filesystems. (Addresses Debian Bug #346580) 3480 3481When allocating space for the RAID filesystems with the stride 3482parameter, mke2fs will now place each portion of the group's inode 3483table right up after the superblock (if present) in order to minimize 3484fragmentation of the freespace. 3485 3486Speed up mke2fs and e2fsck by writing inode and block bitmaps more 3487efficiently by writing the inode and block bitmaps in one pass, thus 3488reducing the number of disk seeks required. 3489 3490Add support for on-line resizing to resize2fs. 3491 3492Fix blkid library so that logic to determine whether or not a device's 3493cached information in /etc/blkid.tab needs to be verified or not 3494doesn't get confused by a system clock which is insane (for example, 3495if the battery is dead on a Macintosh running PPC Linux. (Addresses 3496Red Hat Bug: #182188) 3497 3498The blkid library will now store the UUID of the external journal used 3499by ext3 filesystems, so that in the future, the userspace mount binary 3500can use this to find the location of the external journal and pass 3501this information to the kernel. 3502 3503E2fsck will now consult a configuration file, /etc/e2fsck.conf to 3504control how various options should be handled. See the e2fsck.conf 3505man page for more details. (Addresses Debian Bug: #150295) 3506 3507E2fsck now prints an explanatory message when delaying a filesystem 3508check when the system is running on battery. (Addresses Debian Bug: 3509#350306) 3510 3511E2fsck will detect if the superblock's last mount field or last write 3512field is in the future, and offer to fix if so. (Addresses Debian Bug 3513#327580) These problems will be fixed automatically in preen mode 3514since Debian's boot sequence bogusly doesn't set the time correctly 3515until potentially very late in the bootup process, and this can cause 3516false positives which will cause users' systems to fail to boot. 3517(Addresses Debian Bugs #343662 and #343645) 3518 3519E2fsck now checks to see if the superblock hint for the location of 3520the external journal is incorrect, and if so, offer to update it. 3521(Addresses Debian Bug: #355644) 3522 3523Fix e2fsck from segfaulting on disconnected inodes that contain one or 3524more extended attributes. (Addresses Debian Bug: #316736, #318463) 3525 3526E2fsck will stop and print a warning if the user tries running a 3527read/write badblocks test on a read-only mounted root filesystem. 3528 3529Fix a memory leak in e2fsck's error paths. (Thanks to Michael 3530C. Thompson for pointing these out; they were originally found using 3531Coverity.) 3532 3533When resizing a file containing a filesystem, resize2fs will expand or 3534truncate a file as necessary. (Addresses Debian Bug: #271607) 3535 3536Resize2fs will now automatically determine the RAID stride parameter that 3537had been used to create the filesystem, and use that for newly created 3538block groups. The RAID stride parameter may also be manually specified 3539on the command line using the new -S option to resize2fs. 3540 3541Fix mke2fs so that it correctly creates external journals on 3542big-endian machines (such as a S/390). 3543 3544Fix a bug in the e2p library which could cause dumpe2fs to (rarely) 3545fail to print out the journal or hash seed UUID. (Thanks to Guillaume 3546Chambraud for pointing this out.) 3547 3548Dumpe2fs will now print the size of the journal (if present). 3549 3550Fix debugfs's set_inode_field command so it can properly set the frag, 3551fsize, uid_high, gid_high, and author fields in the inode instead of 3552silently failing, and so that setting the i_size actually sets i_size 3553correctly. 3554 3555Add a new debugfs command, set_current_time, which sets fs->now so 3556that regression test suites can repeatedly modify the filesystem's 3557last_write fields. 3558 3559Fix a bug in debugfs's icheck which would incorrectly report the owner 3560of an extended attribute block. 3561 3562Fix the debugfs commands htree_dump, dx_hash, and list_dir so they print a 3563print a usage message when an illegal option character is given. 3564 3565Fix debugfs's dump_unsued command on filesystems with a 64k blocksize 3566so it won't core dump. (Addresses SourceForge bug #1424311) 3567 3568Fix mklost+found so that it creates a full-sized directory on 3569filesystems with larger block sizes. 3570 3571Fix a file descriptor leak in blkid library. 3572 3573Fix a display bug in "badblocks -sv" so that the done message properly 3574clears the block number at the end of the test. (Addresses Debian Bug 3575#322231) 3576 3577Allow fractional percentages to the -m option in mke2fs and tune2fs 3578(Addresses Debian Bug: #80205) 3579 3580Use fstat/fstat64 in getsize.c if the the target is a regular file, 3581instead of attempting to do a binary search. Fix some fd leaks in 3582error cases. 3583 3584Add support for device mapper library to the blkid library to ensure 3585that the "best" (i.e., leaf) device is probed by the blkid library. 3586 3587Fix the blkid library so that it notices when an ext2 filesystem is 3588upgraded to ext3. 3589 3590Improve the blkid's library VFAT/FAT detection; it now understands 3591labels stored in the root directory, and is more paranoid about 3592checking the FAT superblock values. 3593 3594Fixed a fd leak in the uuid library which was causing problems for the 3595LVM tools. (Addresses Debian Bug: #345832) 3596 3597Add support for the reiser4 and software suspend partitions to the 3598blkid library. Also add support for extract the label from iso9660 3599filesystems. 3600 3601Fix a compile_et bug which miscount the number of error messages if 3602continuations are used in the .et file. 3603 3604Add extra sanity checks to protect users from unusual cirucmstances 3605where /etc/mtab may not be sane, by checking to see if the device is 3606reported busy (works on Linux 2.6) kernels. (Addresses Debian Bug 3607#319002) 3608 3609Updated French, Dutch, Polish, and Swedish translations. (Addresses Debian 3610Bug: #343149, #341911, #300871, #316604, #316782, #330789) 3611 3612Fix use-after-free bug in e2fsck when finishing up the use of the 3613e2fsck context structure. 3614 3615Fixed spelling mistakes, typos, and otherwise clarified man pages and 3616documentation. (Addresses Debian Bugs: #329859, #322188, #316811, 3617#312515, #351268, #357951, #347295, #316040, #368392, #368393, #368394, 3618#368179) 3619 3620Fixed various Debian packaging issues --- see debian/changelog for 3621details. (Addresses Debian Bugs #317862, #320389, #290429, #310950, 3622#310428, #330737, #330736, #329074, #356293, #360046, #366017, #364516, 3623#362544, #362970) 3624 3625 3626Programmer's notes: 3627------------------- 3628 3629Update config.guess and config.sub to latest version (2006-02-23) from 3630FSF. 3631 3632Fix asm_types.h type conflicts on AMD 64 platforms. (Addresses Debian 3633Bugs: #360661, #360317) 3634 3635Fixed the Makefile so that they work correctly on newer versions of 3636GNU make (i.e., 3.81). 3637 3638Add valgrind support to the regression test suites, and eliminate 3639false positives from valgrind. 3640 3641Add a regression test suite for the blkid library. 3642 3643Fix a fencepost error in resize2fs caught by valgrind. 3644 3645Fix compiler warnings about missing memcpy prototypes. 3646 3647We no longer have the sparc assembly code in the header file any more, 3648so we shouldn't set _EXT2_HAVE_HAS_BITOPS_ for the sparc. This would 3649break compiles on the sparc architectures when using gcc. 3650 3651In the libext2fs library, add the new field fs->now which if non-zero 3652is used instead of the system time when setting various filesystem 3653fields (last modified time, last write time, etc.) 3654 3655Fix gcc 4.01 complaints by adding a missing #include <string.h> to 3656ext2fs.h which is needed since the inline functions use memcpy(). 3657(Addresses Sourceforge Bug #1251062) 3658 3659Use BUILD_CFLAGS and BUILD_LDFLAGS instead of CFLAGS and LDFLAGS in 3660the build system when building files in the util directory which are 3661needed during the build process. This avoids potential problems when 3662cross-compiling and some of the options specified in CFLAGS or LDFLAGS 3663are not recognized as valid by the host compiler. (Addresses 3664Sourceforge Bug #1261547) 3665 3666Clean up the blkid library by making the superblock and generic i/o 3667functions to be more generic. Clean up interface to the probe 3668function, and fix memory leak. Finallly, remove an unneeded reference 3669to probe.h in the lib/blkid/resolve.c 3670 3671Add an ext2fs_read_bb_FILE regression test to confirm proper detection 3672of invalid block #'s. 3673 3674The x86 asm constraints for ext2fs_{set/clear}_bit have been fixed to 3675indicate that the the function read/writes the memory location. 3676 3677Fix various gcc -Wall complaints. 3678 3679Add a dependency to make sure that the subdirectories are created 3680before creating all of the object files, in order to address parallel 3681build problem in the library Makefiles. (Addresses Sourceforge Bug: 3682#1261553) 3683 3684Add $(LDFLAGS) to the command line argument when generating an ELF or 3685Solaris shared library, to allow cross-compile and other builds that 3686might need to specify -L paths to needed libraries. (Addresses 3687Sourceforge Bug #1261549) 3688 3689Add a new feature, EXT2_FEATURE_COMPAT_LAZY_BG, which is initially 3690intended for testing purposes. It allows an ext2/ext3 developer to 3691create very large filesystems using sparse files where most of the 3692block groups are not initialized and so do not require much disk 3693space. Eventually it could be used as a way of speeding up mke2fs and 3694e2fsck for large filesystem, but that would be best done by adding an 3695RO_COMPAT extension to the filesystem to allow the inode table to be 3696lazily initialized on a per-block basis, instead of being entirely 3697initialized or entirely unused on a per-blockgroup basis. 3698 3699Fix backwards compatibility so e2fsprogs will better compile on Linux 37002.0.35 systems. 3701 3702Make test scripts more robust against locale-related environment variables 3703 3704Fix type warning problem with sizeof() in ext2fs_open2(). 3705 3706Fix type warning problem with time_t in debugfs. 3707 3708 3709E2fsprogs 1.38 (June 30, 2005) 3710============================== 3711 3712Fix blkid's test programs (built with "make check") compile correctly 3713even without "configure --enable-blkid-debug". 3714 3715Fix ia64 core dump bug caused by e2fsprogs running afoul of C99 strict 3716type aliasing rules on newer gcc compilers. (Addresses Red Hat 3717Bugzilla ##161183.) 3718 3719Fix com_err library to make it more compatible with recent changes 3720made to the com_err library in MIT Kerberos V5 version 1.4. 3721(Addresses Sourcefroge Bug #1150146) 3722 3723General cleanup of messages printed by e2fsprogs programs for grammar, 3724consistency, and to make life easier for translators. Fixed a few 3725strings containing English that had not been marked as needing 3726translations. Removed strings that do not need to be translated, to 3727make life easier for translators. 3728 3729Mke2fs and badblocks will take advantage of a feature in Linux 2.6 to 3730test to see if a device appears to be in use instead of just relying 3731on /proc/mounts and /etc/mtab. (Addresses Debian Bug #308594). 3732 3733Fix portability problem in the filefrag program affecting platforms 3734where the size of an integer is smaller than the size of a long. 3735(Addresses Debian Bug #309655) 3736 3737Mke2fs will now use a larger journal by default for filesystems 3738greater than 4GB. (128 MB instead of 32MB). 3739 3740Mke2fs will refuse to create filesystems greater than 2**31-1 blocks, 3741unless forced. This is to avoid signed vs. unsigned kernel bugs in 3742block numbers that still need to be fixed. 3743 3744The blkid program has a new option which will more efficiently search 3745for device when it is known (or expected) that only one matching 3746device will be found in the system, such as when doing a lookup by 3747UUID. 3748 3749Debian's e2fsprogs-specific initrd fragment will avoid including 3750unnecessary libraries into the initrd ramdisk by unsetting LD_PRELOAD 3751and LD_LIBRARY_PATH, and filtering out libraries found in 3752/etc/ld.so.preload. (Addresses Debian Bug: #304003) 3753 3754Fixed a potential portability issue in the blkid programs for 3755architectures where the char type is unsigned. (Addresses Sourceforge 3756Bug: #1180585) 3757 3758Fix a bug in filefrag so that it doesn't falsely count an extra 3759discontinuity when the first block found is an indirect block. 3760(Addresses Debian Bug #307607). 3761 3762Fix blkid's recognition of cramfs filesystems, and enhance it to be 3763able to handle cramfs labels. 3764 3765Fix debugfs's stat command to not core dump when a filesystem is not 3766open. 3767 3768Fix e2fsck's handling of error conditions caused by the resize inode 3769claiming blocks that are also used by other inodes, a filesystem 3770corruption which was commonly caused by a bug in Fedora Core 3's 3771resize2fs program. 3772 3773Fixed bug in filefrag which caused it to fail on non-ext2/3 3774filesystems. (Addresses Debian Bug: #303509) 3775 3776If the superblock last mount time indicates that the system clock may 3777not be accurate, then e2fsck will omit checking inodes' deletion time 3778field for indications of a potential corrupted orphaned inode list. 3779(Previously e2fsck only ommited these LOW_DTIME checks when the 3780superblock's last write time looked insane.) 3781 3782Fixed a IA64 core dump bug in the e2p library which affected dumpe2fs. 3783(Addresses Debian bug #302200) 3784 3785Make the blkid library more paranoid about being run from setgid 3786programs, and to use __secure_getenv() from libc if it is available. 3787 3788Fixed spelling mistakes, typos, and otherwise clarified man pages. 3789(Addresses Debian Bugs: #304591, #304592, #304594, #304597, #304593 3790and Sourceforge Bug: #1189803) 3791 3792Updated and fixed translations. 3793 3794Fixed various Debian packaging issues --- see debian/changelog for 3795details. 3796 3797Programmer's notes: 3798------------------- 3799 3800Ext2fs_set_bit(), ext2fs_clear_bit(), and ext2fs_test_bit( have been 3801changed to take an unsigned int for the bit number. Negative bit 3802numbers were never allowed (and didn't make any sense), so this should 3803be a safe change. This is needed to allow safe use of block numbers 3804greater than or equal to 2**31. 3805 3806The compile_et program will avoid recreating generated foo_err.c and 3807foo_err.h files if no changes are necessary. The compile_et program 3808will also atomically replace these files to avoid a potential parallel 3809build race problem on SMP systems. (Addresses Sourceforge Bug: 3810#1157933) 3811 3812Added a new function to the blkid library, blkid_probe_all_new(), 3813which only probes newly added disk devices, and change 3814blkid_find_dev_with_tag() to use this function so that when a 3815requested tag is not found, devices that were previously not checked 3816are searched before searching all devices in the system. 3817 3818Added new functions to the blkid library, blkid_dev_set_search() and 3819blkid_dev_has_tag(). 3820 3821E2fsck's problem strings can now use @m and @n as abbrevations for 3822"multiply-claimed" and "invalid", respectively. 3823 3824The e2fsprog.pot file now has an explanation of how the @-expansion 3825and %-expansion works, and strings in e2fsck/problem.c which contain @ 3826characters now have comments in e2fsprogs.pot with the @-expansion to 3827make life easier for translators. 3828 3829Fixed missing return values in the ext2fs library which could cause it 3830to return random garbage in certain error conditions. 3831 3832Allow the current time to be overriden via the E2FSCK_TIME environment 3833variable for use in regression tests. 3834 3835The test scrpit driver program now exits with a non-zero status if 3836there any of its test that it ran failed. 3837 3838Fixed problems with parabuilds on SMP systems. (Addresses Sourceforge 3839Bug: #1157933) 3840 3841Fixed "make check" so that it compiles correctly even when e2fsprogs' 3842header files have not be installed in the system include directories. 3843(Addresses Sourceforge Bug: #1180572) 3844 3845Fixed gcc -Wall nits. 3846 3847 3848E2fsprogs 1.37 (March 21, 2005) 3849=============================== 3850 3851Add support for checking the validity of Extended Attributes stored in 3852inodes to e2fsck. 3853 3854Add support for dumping the contents of large inodes to debugfs, 3855including the extended attributes stored in inodes. 3856 3857Fix mke2fs, e2fsck, debugfs, and the ext2fs_mkdir function so that 3858when we create a new inode we make sure that the extra information in 3859the inode (any extra fields in a large inode and any ea-in-inode 3860information) is initialized correctly. This can take place when 3861mke2fs creates the root and lost+found directory, when e2fsck creates 3862a new root inode or a new lost+found directory, and when the user uses 3863the debugfs write, mknod, or mkdir commands. Otherwise, the newly 3864create inode could inherit garbage (or old EA information) from a 3865previously deleted inode. 3866 3867Fixed a bug in e2fsck so it would notice if a file with an extended 3868attribute block was exactly 2**32 blocks, such that i_blocks wrapped 3869to zero. 3870 3871Added support to filefrag to detect files which are using the new 3872experimental file extents format, and use the non-ext2 algorithm in 3873that case. Fixed a bug to avoid reporting a false discontinuity if 3874there is one or more unallocated blocks at the beginning of a file. 3875 3876Duplicated a check for noticing whether or not the number of blocks 3877(given a certain blocksize) is greater than 2**32 when the 3878BLKGETSIZE64 ioctl is not available to ext2fs_get_device_size(). This 3879allows mke2fs to automatically use a larger blocksize when creating a 3880filesystem on a very large device when run on systems that do not 3881support BLKGETSIZE64. 3882 3883Fix the I18N build which was broken in e2fsprogs 1.36 because the 3884build system had been switched to treat the .gmo files as shipped 3885files (for backwards compatibility with systems that have older GNU 3886I18N tools installed), but the gen_tarball.in script was still 3887removing the .gmo files from the official source distribution. 3888 3889Fixed various Debian packaging issues --- see debian/changelog for 3890details. (Addresses Debian Bugs ##296769, #299341) 3891 3892Programmer's notes: 3893------------------- 3894 3895Added new functions to the e2p library which convert between a string 3896and os_type: e2p_os2string() and e2p_string2os(), and used them to 3897make the generated binaries more compact. 3898 3899Fixed a compile-time error on Darwin systems. 3900 3901Cleaned up the lib/ext2fs Makefile slightly. 3902 3903 3904E2fsprogs 1.36 (February 4, 2005) 3905================================= 3906 3907All of the patches that were applied to Fedore Core 3's 3908e2fsprogs-1.35-11.2 have been integrated, although sometimes with a 3909lot of bug fixes first. Users of Fedora Core 3 are strongly 3910encouraged to upgrade to e2fsprogs 1.36 as soon as possible. 3911 3912Add support for filesystem with the online resizing via resize inode 3913feature. Fixed numerous bugs from the Fedora patches. The Fedora 3914patches also didn't bother to do any consistency checking on the 3915resize inode, or add any tests to the regression test suite. The "-R 3916resize=4g" option to mke2fs was a no-op in the Fedora patches, despite 3917being listed in mke2fs's usage message. All of these shortcomings 3918have been corrected. 3919 3920E2fsck can also also fix filesystems trashed by Fedora's resize2fs 3921program. In order to do this, the user must run the commands: 3922 3923 debugfs -w /dev/hdXXX -R "features ^resize_inode" 3924 e2fsck -f /dev/hdXXX 3925 3926Optionally, the ext2prepare command can be used to re-enable online 3927resizing after the filesystem has been fixed. 3928 3929The fsck program will now accept an optional filedescriptor argument 3930to the -C option. (The Fedora version of this patch would sometimes 3931cause fsck to ignore a parameter on fsck's command line in some rare 3932cases, sigh.) 3933 3934Make sure e2fsprogs doesn't write garbage into the reserved portion of 3935large inodes. 3936 3937Make sure resize2fs releases the blocks belonging to the old inode 3938table blocks when moving the inode table. (Addresses Debian Bug: 3939#290894) 3940 3941Skip the r_resize_inode test if resize2fs is not compiled (due to 3942configure --disable-resizer) 3943 3944E2fsck now checks the summary filesystem accounting information, and 3945if any of the information is obviously wrong, it will force a full 3946filesystem check. (Addresses Debian Bug #291571) 3947 3948Fix e2fsck to not complain when the resize_inode feature is enabled, 3949s_reserved_gdt_blocks is zero, and there is no DIND block allocated in 3950the resize inode. 3951 3952Fix e2fsck to note delete symlinks that contain an extended attribute 3953after the ext_attr feature flag has been cleared. (Addresses Red Hat 3954Bugzilla #146284). 3955 3956Add new utility program, copy_sparse.c, which is very useful 3957for dealing with large sparse files (such as e2image files). 3958 3959Add support for jnl_blocks[] for debugfs's set_super_value. 3960 3961Fix filefrag so that it works correctly with sparse files. 3962 3963Filefrag -v will print first and last blocks. 3964 3965Add interpretation of OS Creator values for FreeBSD and Lites in mke2fs 3966and dumpe2fs. 3967 3968Add mke2fs support so that it can support filesystems larger than 4TB 3969automatically, by retrying with a 4k blocksize if the device size is 3970too big to be expressed using a 1k blocksize. (Addresses Sourceforge 3971bug #1106631) 3972 3973Change blkid to test for NTFS first because Windows sometimes doesn't 3974clear enough of the parition to confuse the probing routines into 3975thinking the old filesystem type is still valid. (Addresses Debian 3976Bug #291990) 3977 3978Add support for swap partition label and uuid's in the blkid library. 3979 3980Add support to the blkid library to recognize Oracle ASM volumes. 3981 3982Make blkid -t display all devices that match the specified criteria, 3983not just the first one, and work more consistently when the blkid 3984cache file is not available or set to /dev/null. (Addresses Debian 3985Bug #290530 and #292425) 3986 3987Badblocks will now correctly display block numbers greater than 3988999,999,999 in its progress display. 3989 3990The tune2fs program will not allow the user from setting a ridiculous 3991number of reserved blocks which would cause e2fsck to assume the 3992superblock was corrupt. E2fsck's standards for what is a ridiculous 3993number of reserved block has also been relaxed to 50% of the blocks in 3994the filesystem. 3995 3996The blkid library will return vfat in preference to msdos, and ext3 in 3997preference to ext2 (if the journalling flag is set) so that mount will 3998do the right thing. (Addresses Debian bug #287455) 3999 4000Mke2fs will now use the -E option for extended options; the old -R 4001(raid options) option is still accepted for backwards compatibility. 4002Fix a double-free problem in resize2fs. (Red Hat Bugzilla #132707) 4003 4004Mke2fs will now accept a size in megabytes, gigabytes, and other units 4005(via "32m" or "4g" on the command line) if the user finds this more 4006convenient than specifying a block count. 4007 4008Fix an obscure, hard-to find bug in "e2fsck -S" caused by an inode 4009cache conherency problem. 4010 4011Debugfs now supports a new command, set_inode_field, which allows a 4012user to manually set a specific inode field more conveniently, as well 4013as set entries in the indirect block map. 4014 4015Debugfs's set_super_value command has been enhanced so that the user 4016can set most superblock fields, including the date/time fields and 4017some of the more newsly added superblock fields. 4018 4019E2fsprogs programs now accept an offset to be passed to the file 4020specifiers, via the syntax: "/tmp/test.img?offset=1024". 4021 4022E2fsprogs programs will now accept blocksizes up to 65536; kernel 4023support on the x86 doesn't exist for now, but it can be useful on 4024other architectures with page sizes greater than 4k. There are 2.6 4025kernel patches out there which enable this, but they are of this 4026writing still experimental. 4027 4028The e2image command now takes the -s option which will scramble 4029directory entries for raw image files. 4030 4031Fix a file descriptor leak in the filefrag program. 4032 4033Make sure e2fsck doesn't crash when /proc/acpi/ac_adapter is not 4034present. 4035 4036Fix bug in debugfs where kill_file would lead to errors when deleting 4037devices and symlinks. (Sourceforge Bugs #954741 and #957244) 4038 4039Fix bug in the blkid library when detecting the ocfs1 filesystem 4040 4041Remove obsolete EVMS 1.x and a.out DLL support. 4042 4043E2fsck will attempt to recover from a journal containing illegal blocks. 4044 4045Fixed two potential ordering constraint problems in e2fsck which might 4046cause the filesystem to be corrupted if e2fsck is interrupted during a 4047(extremely narrow) race window. Thanks to Junfeng Yang from the 4048Stanford Metacompilation group for pointing this out. 4049 4050Fixed bug in e2fsck where it would not accurately detect whether or 4051not the system is running on adaptor if the ACPI device representing 4052the AC adapter didn't correspond to the what was used on IBM 4053Thinkpads. 4054 4055Change e2fsck to accept directories greater than 32MB. 4056 4057Fix e2fsck so that a checkinterval of zero disables a time-based check 4058of the filesystem. 4059 4060Debugfs will check the DEBUGFS_PAGER enviroment variable in preference 4061to the PAGER environment variable. (Addresses Debian Bug #239547) 4062 4063Tune2fs will not mark rewrite the superblock if the feature bitmasks 4064are not modified. 4065 4066The debugfs program will set the filetype information when creating a 4067link. 4068 4069Add debugfs -d option to use a separate source of data blocks when 4070reading from an e2image file. 4071 4072Add e2image -I option which allows the e2image metadata to be 4073installed into a filesystem. 4074 4075Fixed bug in the badblocks program which caused "done" to always 4076appear in english even when a translation was available. (Addresses 4077Debian Bug #252836) 4078 4079The blkid program has a new option -o which controls the output format 4080of the blkid program; this is makes blkid more convenient to use in 4081shell scripts. 4082 4083Fix a minor bug in uuid library, which was not using the full 14 bits 4084of clock sequence when generating UUID's. 4085 4086Fix a Y8.8888K problem in the uuid library. 4087 4088Logsave now creates a new session id for itself to avoid getting 4089killed by init whan transitioning between init levels. 4090 4091Change the licensing of the UUID library to be the 3-clause BSD-style 4092license; this allows Apple to use the uuid library in Darwin. 4093 4094Add ocfs and ocfs2 probe support into the blkid library. 4095 4096Fix a memory and file descriptor leak in the blkid library. 4097 4098The blkid library will revalidate the device if the system time is 4099earlier than last verification time of the device, since that 4100indicates that the system time is probably nottrustworthy. 4101 4102The blkid library will override the default location of the blkid.tab 4103file by the BLKID_FILE environment variable, if it is available. 4104 4105Change the getsize functions to use the BLKGETSIZE64 ioctl on Linux 2.6. 4106 4107Add various portability fixes for lame new versions of glibc, Darwin 4108and GNU/KFreeBSD, as well as removing XSI:ism's. (Addresses Debian 4109Bugs #239934, #264630, #269044, #255589, #289133) 4110 4111Add support for Windows 9x/NT under Cygwin. 4112 4113Updated and clarified various man pages. (Addresses Debian Bugs #236383, 4114 #241940, #238741, #242995, #256669, #268148, #256760, #273679) 4115 4116Updated and fixed translations. (Addresses Debian bugs #244105, #262836) 4117 4118Update the rpm spec files so that it works better with Fedora core 2 4119and RH9. 4120 4121Fixed various Debian packaging issues (see debian/changelog). In 4122particular, fixed the Debian initrd scripts. (#241183, #248050, 4123#253595, #247775) 4124 4125 4126Programmer's notes: 4127------------------- 4128 4129Fixed various gcc -Wall warnings. 4130 4131The uuid library now has new functions uuid_unparse_upper() and 4132uuid_unparse_lower() which forces the case of the hex digits to be 4133upper case, or lower case. 4134 4135The build process has been speeded up by enhancing the subst program 4136to update the modtime on the generated files even when the generated 4137file hasn't changed. 4138 4139The uuid library now uses C99 stdint.h types instead of custom types. 4140 4141Updated config.guess and config.sub with newer versions from the FSF. 4142 4143Removed out of date .cvsignore files from the source distribution. 4144 4145The ext2fs_unlink() function will return an error if both the name and 4146inode number are unspecified, to avoid doing something surprising 4147(such as unconditionally deleting the first directory entry). 4148Directory entries are now deleted by coalescing them with the previous 4149directory entry if possible, to avoid directory fragmentation. This 4150is not an issue with the e2fsprogs suite, but may be a problem for 4151some of the users of libext2fs, such as e2tools. 4152 4153Add support for version numbers of the form "1.36-rc1". 4154 4155Fix build of mke2fs.static. 4156 4157Add basic ext2fs library support for large (EA in inode) inodes. 4158 4159The test_io mechanism can now abort after n reads or writes to a 4160particular block. The block is specified by TEST_IO_BLOCK environment 4161variable, and the read/write count by the TEST_IO_READ_ABORT and 4162TEST_IO_WRITE_ABORT environment variables. The block data is now only 4163dumped if the 0x10 bit is set in TEST_IO_FLAGS. 4164 4165UUID_DEFINE() in the uuid library now creates a static variable, with 4166__attribute__ ((unused)) if we are using GCC, so that UUID_DEFINE can 4167be used in header files. 4168 4169Add support for the install-strip and install-shlibs-strip targets, as 4170suggested by the GNU coding guielines. "make install" no longer 4171strips the binaries which are installed. 4172 4173Remove support for the --enable-old-bitops configure option which was 4174only for very old sparc systems. 4175 4176Remove support for --enable-clear-htree; this was only needed during 4177the early development of the htree patch. 4178 4179Use Linux-kernel-style makefile output so it is easier to see compiler 4180warnings. 4181 4182Update gettext files to version 0.14.1. 4183 4184Update to use autoconf 2.5x. 4185 4186Improved support for compiling e2fsprogs under dietlibc. 4187 4188Make e2fsprogs portable to Solaris and FreeBSD systems. 4189 4190Add blkid_verify(), blkid_get_library_version(), and 4191blkid_parse_version_string() functions to the blkid library. 4192 4193Add pkg-config files for e2fsprogs's libraries. 4194 4195Fix "make uninstall" to so that it removes everything that is installed. 4196 4197Add a configure --enable-maintainer-mode option which enables the 4198makefile rules to rebuild the configure script from configure.in, and 4199to reubuild the .gmo files in po directory. 4200 4201Drop the sparc assembly bitwise operations; it's less efficient 4202than the GCC 3.4 compile code and triggers compiler warnings on 4203sparc64. Thanks to Matthias Andree for his analysis and suggestions. 4204(Addresses Debian Bug #232326) 4205 4206 4207E2fsprogs 1.35 (February 28, 2004) 4208================================== 4209 4210E2fsck has a new -k option, which in conjunction with the -c options, 4211preserves the existing badblocks list. 4212 4213Cleaned up e2fsck's preen-mode messages during the passes 1b, 1c, and 1d. 4214 4215E2fsprogs will now deal correctly with symlinks that contain 4216extended attribute information, which can be created using SE Linux. 4217(Addresses Debian Bug #232328) 4218 4219Remove a double longjump into an invalid stack frame bug in e2fsck. 4220(This was during an abort sequence, which normally worked on Linux and 4221caused a core dump on other operating systems.) 4222 4223Fix NLS bug in e2fsck, by avoiding trying to expand an empty string 4224(the NLS library will replace "" with the .po header information). 4225 4226Fix a bug in mke2fs which caused -T largefile or -T largefile4 to core 4227dump due to a division by zero error. (Addresses Debian bug #207082) 4228 4229Fixed a bug in e2fsck which caused it to incorrectly fix a filesystem 4230when reconnecting a directory requires creating a lost+found 4231directory. (Addresses Debian bug #219640). 4232 4233Fixed a bug where e2fsck would bomb out if a journal needed to be 4234replayed when using an alternate superblock. 4235 4236E2fsck will give an extra grace period before actually forcing a check 4237if the laptop is running on battery. The next time fsck runs while 4238the system is on the AC mains, or after the grace period is exceeded, 4239the filesystem will be checked. (Addresses Debian bug #205177) 4240 4241E2fsck will inform the user when there are 5 or fewer mounts before a 4242filesystem check will be forced. (Addresses Debian bug #157194) 4243 4244Fix e2fsck's handling of corrupted indirect blocks in the bad block. 4245We now correctly handle the case where there is an overlap between a 4246block group descriptor or a superblock and a bad block indirect block. 4247In the case where the indirect block is corrupted, we now suggest 4248"e2fsck -c". 4249 4250Fix byte swap bugs in e2fsck that caused the journal backup location 4251in the superblock and symlinks created by SE Linux to be cleared 4252by e2fsck on big-endian machines. (Addresses Debian bug #228723) 4253 4254E2fsck -c now replaces the current list of bad blocks with the ones 4255found by badblocks. 4256 4257Fix bugs in e2fsck and tune2fs which could cause a core dump if a 4258non-existent LABEL or UUID specifier is to e2fsck or tune2fs. 4259 4260Fix a potential bug in e2fsck which could cause it to core dump when 4261trying to print the location of the backup superblock. 4262 4263Protect against a potential core dump in e2fsck when printing a 4264message about backup superblocks. 4265 4266Add support for backing up the journal inode location in the 4267superblock. E2fsck will automatically save the journal information in 4268the superblock if it is not there already, and will use it if the 4269journal inode appears to be corrupted. ext2fs_add_journal_inode() 4270will also save the backup information, so that new filesystems created 4271by mke2fs and filesystems that have journals added via tune2fs will 4272also have journal location written to the superblock as well. 4273Debugfs's logdump command has been enhanced so that it can use the 4274journal information in the superblock. 4275 4276E2fsck will now update all superblocks when moving the journal inode. 4277 4278Shrink the size of the e2fsck executable by moving some initialized 4279variables to the BSS segment. 4280 4281E2fsck will avoid printing the ^A and ^B characters which bracket the 4282progress bar when stdout and stdin are a tty device instead of a pipe 4283to another program. (Addresses Debian bug #204137) 4284 4285Debugfs's mkdir command will automatically expand the directory if 4286necessary. (Addresses Debian Bug: #217892) 4287 4288Fixed a bug in debugfs so that copying a file from /dev/null uses the 4289correct mode bits. (Addresses Debian Bug: #217456) 4290 4291If the environment variables DEBUFS_PAGER and PAGER are not set, 4292debugfs now searches for the appropriate pager to use, beginning with 4293/usr/bin/pager, and then falling back to 'more' and 'less'. 4294(Addresses Debian bug #221977) 4295 4296Debugfs will now support 2.6 device numbers where the major or minor 4297number may be larger than 255. (Addresses Sourceforge bug #865289) 4298 4299Fix debugging printf in resize2fs. (Addresses Debian Bug #271605) 4300 4301Chattr now stops processing options when it sees '--'. (Addresses 4302Debian bug #225188) 4303 4304Fix regression tests so they work correctly when e2fsprogs is compiled 4305with configure --disable-htree. 4306 4307Fix bug in uuid library when there is no network card and the library 4308is generating a time-based uuid. The random MAC address was not 4309correctly generated to be a multicast address. 4310 4311Add compile_et extensions from Heimdall that were missed the first time 4312around. 4313 4314Fix bug in badblocks when using O_DIRECT; we need to make sure that 4315we're reading from an offset which is page aligned. For read-only and 4316read-write tests, we try to recover after an error so that we can 4317continue reading on page-aligned boundaries. (Addresses Debian Bug 4318#203713) 4319 4320Badblocks now checks 64 blocks at a time instead of 16. (Addresses 4321Debian bug #232240) 4322 4323Updated and clarified various man pages. (Addresses Debian Bug 4324#206845, #222606, #214920, #232406) 4325 4326Updated and fixed translations. (Addresses Debian bugs #200086, #214633) 4327 4328Fixed various Debian packaging issues (see debian/changelog). 4329 4330Programmer's notes: 4331------------------- 4332 4333Fixed a build problem so that e2fsprogs would compile with the 4334--enable-profile option to configure selected. (Addresses Sourceforge 4335bug #811408) 4336 4337Fixed C++ problems with the ext2fs.h header. (Addresses Red Hat 4338Bugzilla Bug #112448) 4339 4340Centralize code which calculates the location of the superblock 4341and block group descriptors so that it is in a single library routine. 4342 4343Added two new functions, ext2fs_file_open2() and 4344ext2fs_inode_io_intern2() which take a pointer to an inode structure. 4345 4346Fix compile_et to output the correct prototype for 4347initialize_xxx_err_table_r() in the header file. (Addresses Debian 4348bug #204332) 4349 4350In the lib/et makefile, make sure com_err.info is deleted on "make clean". 4351 4352Fix 64-bit warnings in e2fsprogs pass1b by using inttypes.h if 4353present. This is for when we try stuffing an int into void * pointer. 4354 4355Fix type-punning which can cause gcc 3.x to miscompile code by getting 4356confused about pointer aliasing. ext2fs_getmem(), ext2fs_free_mem(), 4357and ext2fs_resize_mem() all now take a 'void *' instead of a 'void 4358**'. The EVMS code uses an ugly union approach since we don't want to 4359modify the EVMS interfaces. 4360 4361Make sure all Makefiles use $(MAKE) rather than hardcoded "make", to 4362aid build process on systems can use invoke GNU make as "gmake". 4363 4364Added regression testing for mke2fs. 4365 4366Fixed gcc -Wall nitpicks. 4367 4368Fixed various compiler warnings. 4369 4370Add portability fixes for FreeBSD and for using fsctl under Darwin to 4371support ext2 ioctl's. 4372 4373 4374E2fsprogs 1.34 (July 25, 2003) 4375=============================== 4376 4377Fixed a bug introduced in E2fsprogs 1.30 which caused fsck to spin in 4378a tight loop while waiting for a child fsck to exit in some cases. 4379This burns CPU times which slows down the low-level filesystem check. 4380 4381Added code to mke2fs to assure that the default block size for a 4382filesystem is at least as big as the sector size of the device, if it 4383can be determined. 4384 4385Changed mke2fs and resize2fs to round the default size of a filesystem 4386to be an even multiple of the VM pagesize in order to avoid a Linux 4387kernel bug introduced when the storage of the buffer cache was moved 4388into the page cache. 4389 4390Mke2fs will warn the user when creating a filesystem with journaling 4391and a blocksize greater than 4096. (Addresses Debian bug #193773) 4392 4393Fixed a bug in resize2fs which caused it to fail on filesystems with a 4394non-empty bad block list. Resize2fs now discards any blocks on the 4395badblock list which are no longer part of the filesystem as the result 4396of a filesystem shrink. (Note: this means that shrinking and then 4397enlarging a filesystem is no longer a reversible operation; 4398information about bad blocks in the part of the filesystem which is to 4399be chopped off will be lost.) 4400 4401Changed resize2fs so the user can use prefixes to specify the units of 4402the new filesystem size (sectors, kilobytes, megabytes, or gigabytes), 4403and to make the error and informational messages explicitly display 4404the blocksize used by the filesystem, in order to avoid confusion. 4405(Addresses Debian bug: #189814) 4406 4407Added a new debugfs command, dump_unused, which dumps the contents of 4408all unused blocks to stdout. (Useful as an emergency try-to-find 4409deleted data command.) 4410 4411Added a new debugfs command, imap, which prints the location of a 4412specified inode in the inode table. 4413 4414Fixed a bug in the badblocks program which caused it to use one bit of 4415randomness in its non-destructive read/write test, instead of using a 4416full 8 bits of randomness. 4417 4418Added a new option (-t) to badblocks, which allows the user to control 4419the test pattern(s) used when checking a disk. 4420 4421The blkid probe function now more correctly detects UDF filesystems. 4422 4423Fixed a bug in the blkid library which caused it to not update its 4424cache if a filesystem changes from having a label to not having a 4425label. 4426 4427Fixed a bug in the blkid library wihch could avoid an infinite loop 4428in blkid_find_dev_with_tag() if /proc is not mounted and there the 4429/etc/blkid.tab file has not yet been created. 4430 4431Fixed the badblocks program so that the destructive read/write test 4432honors the -c option, and to use O_DIRECT when possible to avoid 4433thrashing the system block buffer cache. 4434 4435Fixed various NLS issues. 4436 - Added Czech and Swedish translations 4437 - Removed testing NYC translation 4438 - Fixed NLS support for message abbrevations in e2fsck 4439 - Remove de-utf.po, since we shouldn't have two versions using different 4440 charset encodings. 4441 - Used ngettext() (abbreivated with the macro P_(str1, str2, n)) to 4442 simplify the statistics reporting in e2fsck. 4443 4444Changed configure.in so that its defaults for *BSD systems no longer 4445build an fsck wrapper, and not to install in /usr/local by default. 4446 4447Fixed some minor spelling errors/typo's in e2fsck and the configure 4448script. 4449 4450Fixed various Debian packaging issues (see debian/changelog). 4451 4452Updated and clarified man pages. (Addresses Debian Bug #195616) 4453 4454Programmer's notes: 4455------------------- 4456 4457Fix gcc -Wall nitpicks. 4458 4459Updated gettext implementation used by e2fsprogs to 0.11.5, and enable 4460NLS support by default. (Added partial workaround for gettext/Darwin 4461incompatibility problems.) 4462 4463Added full MIT KRB5 and Himdall compaibility support to the com_err 4464library and the compile_et program. (Addresses Debian bug #191900) 4465 4466Added the blkid_known_fstype() function to the blkid library, which 4467returns true if it is passed a filesystem type which is recognized by 4468the blkid probing functions. 4469 4470Improved the documentation for the blkid library. 4471 4472Added the ext2fs_get_device_sectsize() function the the ext2fs library, which 4473returns the hardware sector size of a device, if it is available. 4474 4475Added a dependency in the blkid library's .so file to the uuid 4476library, since the former uses the latter. (Addresses Debian bug 4477#194094) 4478 4479Added --with-diet-libc and --disable-evms to the configure script. 4480 4481Fixed a minor memory leak in the badblocks program. 4482 4483Fixed a portability problem in tune2fs --- not all systems have strptime(). 4484 4485Fixed a portability problem in debugfs with the use of getopt() more 4486than once. Old-style BSD, new-style BSD, and Linux C libraries all do 4487things differently. 4488 4489Add support Windows support to ext2fs_get_device_size(). 4490 4491Added (normally disabled) debugging code to the Unix I/O manager which 4492causes it to disable all userspace caching if the NO_IO_CACHE is 4493defined. 4494 4495Changed the test I/O manager so it can always be linked into e2fsck, 4496mke2fs, and tune2fs if enabled via --enable-test-io-debug to the 4497configure script. The test I/O manager will only print any debugging 4498information if the TEST_IO_FLAGS or TEST_IO_BLOCK environment 4499variables are set, which specifies which I/O operations are logged and 4500a block number to watch, respectively. The log messages are sent to 4501stderr by default, unless a filename is specified via the 4502TEST_IO_LOGFILE environment variable. 4503 4504 4505E2fsprogs 1.33 (April 21, 2003) 4506=============================== 4507 4508Added a new utility program, logsave, which captures the output of a 4509command in a log file, even if the containing directory hasn't been 4510mounted yet (in which case the it saves the output in memory until it 4511can write out the logfile). This is useful for capturing the 4512output of fsck during the boot sequence. 4513 4514Fixed some portability problems that were causing problems under 4515the Cygwin32 environment. 4516 4517Mke2fs now interprets a negative number to the -b option as a minimum 4518block size. 4519 4520Fixed a bug in mke2fs which was incorrectly checking the argument to 4521the -g option if the default block size was used. (Addresses Debian 4522bug #188319) 4523 4524Fsck now explicitly ignores tmpfs and devpts, and it will complain if 4525it can not find filesystem checkers for jfs, reseirfs, and xfs. 4526 4527E2fsck now updates the global free block and inode counters from the 4528block group specific counters quietly. This is needed for an 4529experimental patch which eliminates locking the entire filesystem when 4530allocating blocks or inodes; if the filesystem is not unmounted 4531cleanly, the global counts may not be accurate. 4532 4533Imported bug fixes to the EVMS plugin from the EVMS 2.0 tree. (EVMS 45342.0 is not yet supported; this only pulled in the bug fixes: fixed 4535possible hangs caused by bugs in calling waitpid, and not setting the 4536pipe to non-blocking mode; also fixed a file descriptor leak; made 4537sure all functions call log entry/exit functions.) 4538 4539Badblocks now flushes its output file as bad blocks are discovered. 4540 4541The uuid library is now more paranoid about assuming the correctness 4542of the /dev/random device; it mixes in a stream of bytes from 4543random/srandom, just in case. 4544 4545Update Debian files to reflect the fact that I am now the Debian 4546maintainer of e2fsprogs. Other various Debian-specific packaging 4547cleanups. 4548 4549Move the source tarball generation functions from the top-level 4550makefile to the util/gen-tarball script. 4551 4552Updated the Turkish .po translation file. 4553 4554Added Heimdal and MIT krb5 extensions to the com_err library to make 4555it more compatible with com_err libraries from those distributions. 4556 4557Changed dumpe2fs to always display the superblock fields relating to 4558the journalling and/or directory indexing feature even if those 4559features are not enabled. 4560 4561Updated and clarify copyright statement vis-a-vis alpha releases of 4562e2fsprogs. 4563 4564The ss library will now try to dynamically link to the readline 4565library and use it if it is present in the system. This means that 4566the debugfs program now has line editing and history features. The 4567SS_READLINE_PATH environment variable is used to find a readline or 4568readline-compatible library. 4569 4570E2fsck now finds most duplicate filenames (all when rebuilding all 4571directories via the -D option) and offers to delete or rename 4572duplicate filenames/directory entries. (Addresses Debian Bug #176814). 4573 4574Fix bug in e2image. When writing out a raw image file, include data 4575blocks from symlinks that do not store the symlink within the inode. 4576 4577Fix bug in resize2fs which incorrectly moved the block and inode 4578bitmaps for sparse superblock filesystems and incorrectly marked 4579blocks as in use. (Addresses Debian bug #174766) 4580 4581Added a new shared library, the blkid library, which efficiently 4582allows fsck, mke2fs, e2fsck, and tune2fs to be able to look up LABEL 4583and UUID filesystem specifiers without needing to search all of the 4584devices in the system. Instead, the device is looked up in a cache 4585file, and then verified to make sure the blkid cache is correct. 4586 4587Tune2fs and e2label will accept LABEL=xxx and UID=yyy specifiers for 4588the device name, using the blkid library. (Addresses Debian bugs 4589#166048, #179671) 4590 4591Fsck now supports backslash escapes in /etc/fstab so that \040 can be 4592used for spaces in device labels. 4593 4594Removed 32-bit limitations for debugfs's dump command. 4595 4596If the user specifies a large number of inodes, Mke2fs will 4597automatically adjust the number of blocks per group to find a valid 4598set of filesystem parameters. 4599 4600Add workaround to detect broken MD devices where when some of the 4601underlying devices are marked read-only, writes to the MD device are 4602silently dropped. E2fsck will detect if there is an attempt to run 4603the journal twice, and abort with an error if this is the case. 4604(Addresses IBM Bugzilla bug #1226) 4605 4606E2fsck will print an error if more than one of the -p/-a, -n or -y 4607options are specified. 4608 4609E2fsck will fix HTREE corruptions in preen mode, without stopping the 4610boot process. This is needed because the 2.4 ext2 filesystem 4611accidentally had the INDEX_FL backwards compatibility code removed. 4612 4613Mke2fs no longer creates filesystems with the dir_index flag set by 4614default; the user has to specifically request it. 4615 4616Update and clarified various man pages. (Addresses Debian bugs 4617#173612, #175233, #175113, and #170497, #185945, #188318) 4618 4619Created man page for the mk_cmds program (from the libss library). 4620 4621Programmer's notes: 4622------------------- 4623 4624Fix various gcc -Wall nits. 4625 4626Fixed a lot of portability problems that caused e2fsprogs not to build 4627successfully under Solaris and Apple/Darwin. 4628 4629Fixed a Makefile dependency to allow building e2fsprogs using parallel 4630make jobs. 4631 4632Changes to create a subset distribution which consists only of the 4633et, ss, uuid, and blkid libraries. The configure script and top-level 4634makefile were changed to support working with a subset distribution. 4635 4636Removed EXT2_FEATURE_RO_COMPAT_BTREE_DIR mention of since it's not 4637actually used, and might people who are looking for 4638EXT2_FEATURE_COMPAT_DIR_INDEX, which is in use. 4639 4640Updated debian files to fix a number of Lintian warnings. 4641 4642Updated config.guess and config.sub with newer versions from the FSF. 4643 4644Removed unnecessary libraries from being linked into the fsck, lsattr, 4645chattr, and blkid executables. 4646 4647E2fsprogs 1.32 (Nomvember 9, 2002) 4648================================== 4649 4650Fixed a bug in the Unix I/O routines which caused needless writebacks 4651of clean blocks from the unix_io cache (they were erroneously marked 4652as being dirty, so they were getting written back to disk before 4653getting evicted from the disk cache). This was harmless, but it 4654significantly slowed down e2fsck. 4655 4656Made some other minor optimizations to the Unix I/O routines to save a 4657small amount of CPU time. 4658 4659Updated internationalization files. 4660 4661 4662E2fsprogs 1.31 (Nomvember 8, 2002) 4663================================== 4664 4665Update EVMS ext2fsim plugin with EVMS 1.2. (We still support 4666compiling the fsim plugin with EVMS 1.0 and 1.1.) Add better error 4667handling for child process that die unexpectly. Add a hack to force 4668mkfs to create filesystems that won't cause problems with hardware 4669that has 2k or 4k minimum blocksize requirements. Read from child 4670processes in non-blocking mode, so that the GUI continues to update. 4671 4672Fix e2fsck so that it returns the appropriate exit code when the root 4673filesystem has been changed, so that system's rc scripts will be told that 4674the system needs to be rebooted. 4675 4676Fix a bug in ext2fs_flush/ext2fs_close; when the MASTER_SB_ONLY flag 4677was set, some of the descriptor blocks that should have been written 4678out were getting skipped. 4679 4680Changed e2fsck to force out changes to the backup copies of the 4681superblock and block group descriptors when important changes are made 4682to those data structures. 4683 4684Fix a bug where e2fsck could erroneously mark a filesystem as being 4685clean if a check of dirty filesystem is interrupted with a ^C. (Bug 4686introduced in e2fsprogs 1.28.) 4687 4688If journal debuging is enabled using --enable-jbd-debug, the debugging 4689level is now set via the E2FSCK_JBD_DEBUG environment variable. 4690 4691If byteswapping support is disabled using configure --disable-swapfs, 4692skip the tests which depend on byte-swapping, so that "make check" 4693won't bomb out. 4694 4695Lshattr will now display the indexed directory flag. Also, some of 4696the more esoteric compression flags are supressed unless compression 4697support has been enabled. 4698 4699Update man pages. 4700 4701 4702E2fsprogs 1.30 (October 31, 2002) 4703================================= 4704 4705When resizing a filesystem, and inodes need to moved, resize2fs will 4706update the inode's ctime field, and the mtime field of the containing 4707directories, so that incremental backups using dump/restore will work 4708correctly. 4709 4710In order to avoid spurious errors, e2fsck wil check the last 4711superblock write time to determine whether or not it can safely use 4712the LOW_DTIME checks to determine if there are inodes on the orphan 4713inode linked list that had somehow gotten disconnected. (Addresses 4714Sourceforge bug #620980) 4715 4716Partition in /proc/partitions that being with the string "lvm" are 4717considered real partitions even if they do not end with a number. 4718 4719Fixed a bug in the the uuid generation function, where if /dev/urandom 4720is not present, but /dev/random is, and there isn't sufficient 4721entropy, the get_random_byte function could spin a loop forever. 4722 4723E2fsck, mke2fs, etc. will now reliably notice when image files are 4724mounted using the loopback interface. (Addresses Sourceforge bug 4725#619119) 4726 4727When flushing buffers (as is done by badblocks, for example) check to 4728see if the BLKFLSBUF ioctl succeeds, and if so, avoid using the 4729FDFLUSH ioctl which causes the MD device driver which causes confusing 4730syslog messages about obselete ioctl messages. (Addresses Sourceforge 4731bug #545832). 4732 4733Debugfs's write command now checks to see if the destination filename 4734exists before creating it. (Addresses Sourceforge bug #478195.) 4735 4736When installing man pages, search for compressed man pages using all 4737commonly used compression extensions (i.e., .Z, .gz, .bz2, etc.) 4738 4739Fixed a bug in fsck where multiple filesystems specified on the 4740command were not being checked in parallel due to a logic bug 4741introduced to support the FSCK_MAX_INST environment variable. 4742 4743We have added a new superblock field, s_mkfs_time, which indicates 4744when a filesystem was created. It is set via mke2fs, and printed out 4745by dumpe2fs, but is not actually touched by the kernel. 4746 4747Dumpe2fs has been made more robust by not aborting if there is an 4748error reading the block/inode bitmaps; instead it will still print out 4749the location of the block/inode bitmaps and inode table. 4750 4751Add support for the an alternative block group descriptor layout which 4752allows for on-line resizing without needing to prepare the filesystem 4753in advance. (This is the incomat feature flag meta_bg.) 4754 4755Add support for storing default mount options in the superblock, so 4756that the filesystem can be mounted with specific mount options without 4757needing to specify them on the mount command line or in the /etc/fstab 4758file. 4759 4760Add support for a new inode flag, which is to be used for indicating 4761the top of directory hierarchies for the Orlov block allocator. 4762 4763Fix e2fsck so that if it creates the lost+found directory, it does so 4764with the more apporpriate permissions of 0700. Also change 4765mklost+found so that it also creates the directory 0700. 4766 4767Fixed format bug in e2fsck if NLS is in use. 4768 4769Add a German translation for e2fsprogs's NLS support. 4770 4771Fixed e2fsck so that it more handles BAD_BLOCK_IN_INODE_TABLE even at 4772the beginning of the inode table. This won't matter much, since if 4773there is a bad block at the beginning of the inode table, the root 4774directory won't be available. But at least e2fsck won't crash in this 4775case. 4776 4777Fixed endian problems in the htree code for e2fsck and debugfs. 4778 4779When byte-swapping a filesystem on a PPC architecture, byte-swap the 4780bitmaps since the historical big-endian ext2 variant had byte-swapped 4781bitmaps, and the ext2fs library assumes this. This fixes the 4782regression test suite on PPC machines. 4783 4784Fix e2image so that it handles a bad block in the inode table 4785appropriately. 4786 4787E2fsck now uses a more sophisticated algorithm to salvage corrupted 4788directories that recovers more information in the corrupted directory 4789block. 4790 4791E2fsck now performs additional consistency checks on indexed (HTREE) 4792directories. 4793 4794Fix bug where efsck might get confused about whether a completely 4795empty directory block is an empty leaf block or an interior htree 4796node. This could cause e2fsck to get confused into think that a valid 4797indexed directory was corrupted. 4798 4799E2fsck no longer creates an empty directory entry at the end of a 4800directory block when optimizing a directory. This would cause some 4801earlier versions of the dxdir kernel code to corrupt the directory 4802when attempting to split a node containing an empty directory entry. 4803 4804E2fsck could sometimes in rare circumstances leave the indexed flag 4805set after a small directory was optimized by compressing it instead of 4806indexing it. (This should never happen in real life, since 4807directories that small wouldn't have been indexed, but better safe 4808than sorry.) 4809 4810E2fsck now only updates the master superblock in all cases. This 4811slightly shortens its run time. 4812 4813Ext2ed can deal with directory entries of length 0; previously it 4814would get stuck in an infinite loop. 4815 4816Fsck now has support for reiserfs volumes when parsing LABEL= and UUID= 4817specifiers. (Sourceforge patch #613447) 4818 4819Badblocks will now work correctly on read-only devices such as 4820CD-ROM's. (Sourceforge patch #600451) 4821 4822Updated and clarified man pages. (Addresses Debian bug #167108) 4823 4824 4825E2fsprogs 1.29 (September 24, 2002) 4826=================================== 4827 4828Fixed a bug in e2fsck which could corrupt a directory when optimizing 4829it (via the -D option) or rebuiliding the hash tree index with a 1 in 4830512 probability, due to a fence post error. 4831 4832Fixed a bug in the LVM support code which caused LABEL='xxx' not to 4833work correctly. 4834 4835Mke2fs now enables the directory indexing flag by default. (Since 4836this is a compatible feature flag, it's safe to do so.) 4837 4838Tune2fs will support setting the directory indexing feature flag. It 4839will automatically set up the default hash algorithm and hash seed 4840fields in the superblock. 4841 4842If the bone-headed user enters the root filesystem twice in 4843/etc/fstab, the -R option which skips the root filesystem will skip 4844all of them. (Addresses Debian bug #159423). Note! This is not a 4845precedent for dealing intelligently with any other kind of doubled 4846entry in /etc/fstab! 4847 4848 4849Programmer's notes: 4850------------------- 4851 4852Removed perror declaration in lib/et/internal.h. All modern systems 4853can be expected to define perror() these days. Besides, the lib/et 4854code wasn't using perror at all anyway. :-) 4855 4856 4857E2fsprogs 1.28 (August 31, 2002) 4858================================ 4859 4860Add support for the Hashed-Tree Directory Indexing to e2fsck. Support 4861for setting the htree flag is not included yet, although it can be 4862manually turned on via the debugfs program. 4863 4864Clarified e2fsck error message which is printed when it cannot find 4865sufficient contiguous block when relcating filesystem metadata. 4866 4867Added support for building an EVMS plugin module for ext2/3. This 4868module is substantially the same as the EVMS module shipping with EVMS 48691.1, with one or two bugfixes. E2fsprogs can also build this plugin 4870for use with EVMS 1.0 (which did not include the ext2 plugin module), 4871if the configure --enable-old-evms flag is given. 4872 4873Fsck will search through EVMS volumes when trying to resolve 4874filesystem specifications such as LABEL=xxx or UUID=xxx. 4875 4876Added a new utility program, /sbin/findfs, which will return 4877filesystem specifications such as LABEL=xxx or UUID=xxx, and prints 4878the device name. 4879 4880Update and clarified various man pages. (addresses Debian Bug #145044, 4881#146437, #131350, #151990, #144621, #141938) 4882 4883If there are no filesystems specified on fsck's command line, fsck now 4884treat that as if the -As options were given. Previously it would 4885simply do nothing. (Addresses Debian Bug #153102) 4886 4887Mke2fs no longer treats a failure to be able to clear the MD signature 4888at the end of the filesystem as a fatal error. (Addresses Debian Bug 4889#155007) 4890 4891The e2p library functions (which are used by lsattr and chattr) now 4892double check to make sure the file is a regular file or directory 4893before attempting to use the ext2 ioctls. Some device drivers 4894unfortunately respond to the ext2 ioctl's with unknown behaviour. 4895(Addresses Debian Bug #152029). 4896 4897The extended attribute handling has been updated to correspond with 4898the latest V2 bestbits ACL code. 4899 4900Fixed bug in e2fsck which caused it to not clear the dtime field when 4901processing truncated inodes on the orphan list. This could cause data 4902loss(!) if a filesystem is rebooted before a truncate has been 4903committed. 4904 4905E2fsck now uses red/black trees in pass1b, which removes some O(n**2) 4906algorithms. This makes e2fsck much faster in the case of severely 4907corrupted filesystems where a large number of blocks are claimed by a 4908large number of inodes. (Thanks to the 2.5 IDE device driver for 4909inspiring this work.) 4910 4911Resize2fs has been significantly sped up when shrinking and expanding 4912a filesystem by a very small number of blocks (for example, when EVMS 4913is converting a partition to be an EVMS legacy volume). 4914 4915Added a new option to e2fsck, -D, which will optimize or compress all 4916of the directories in the filesystem. 4917 4918E2fsck now catches SIGINT and SIGTERM to make sure it will can 4919properly clean up and only exit at safe points. Fsck will pass 4920SIGINT/SIGTERM to its child processes, and wait until they have all 4921exited before it exits. 4922 4923The uuid parsing code in the uuid library now properly complains when 4924an illegally formated uuid is presented to it. (Addresses Debian bug 4925#152891) 4926 4927Restrict use of the 2.4 setrlimit ABI f*ckup to kernels between 2.4.10 4928and 2.4.17, since the workaround can cause problems when using a 2.4 4929kernel with an old version of glibc built with the 2.2 headers. 4930 4931Fixed a bug in mke2fs where it wasn't properly clearing the initial 4932superblock used by other filesystems. (Addresses Debian bug #147256.) 4933 4934Added support for the synchronous directory feature written by Andrew 4935Morton. 4936 4937The debugfs program can delete directories using the rmdir command. 4938 4939Add support for 8k and 16k filesystems (for systems with page sizes 4940that are greater or equal to 8k or 16k, respectively). Note that 4941these filesystems can not be mounted on x86 systems, or other systems 4942with only 4k page sizes, due to limitations in the current Linux VM 4943code. 4944 4945Resize2fs requires that the filesystem state be valid and have no 4946errors; otherwise, e2fsck -f must be run first. (Previously it simply 4947required that the last fsck time be greater than the last mount time.) 4948 4949Configure now defaults the man pages directory to /usr/share/man on 4950Linux systems. 4951 4952E2fsck now offers to truncate inodes which contain too many blocks (so 4953that i_blocks would overflow. Also fixed handling of large sparse 4954files. 4955 4956E2fsck now more completely checks for symlink validity, including 4957requiring NULL termination and length checks. 4958 4959E2fsck will offer to try forcing a disk write to remap a bad block 4960after finding a read error when reading a filesystem metadata block. 4961 4962Fixed a bug in debugfs which caused the -b and -s options to crash 4963debugfs, as well as breaking the testb, setb, and clearb functions. 4964 4965Added a bmap command to debugfs which calculates the logical to 4966physical block mapping for a particular inode. 4967 4968Fixed a bug in code which checked to see if a device was mounted which 4969sometimes (rarely) failed in the case of a plain file. 4970 4971Fixed a bug in resize2fs where when it reported an error, it would 4972print a message erroneously indicating that the filesystem had been 4973resized before it aborted. 4974 4975When resizing a plain file which is smaller than the requested size, 4976resize2fs will attempt to extended the file so that programs like 4977e2fsck will not complain that the file is too small. 4978 4979Resize2fs will print the actual new size of the filesystem when it is 4980finished resizing. 4981 4982Fixed a bug in debugfs where "ls -l" would report incorrectl file type 4983information on big-endian systems. 4984 4985 4986Programmer's notes: 4987------------------- 4988 4989Fixed collisions with C++ reserved words. 4990 4991Added portability fixes for building e2fsprogs on the HURD and AIX. 4992 4993Added the ext2ed program for creation of test cases. (ext2ed has many 4994limitations and bugs which make it unsuitable for production use.) 4995 4996The ext2fs_read_dir_block2 and ext2fs_write_dir_block now take a new 4997flag, EXT2_DIRBLOCK_V2_STRUCT, which will will reverse when the 4998name_len field is byte swampped on big-endian machines, since in the 4999V2 structure, name_len is a char field which is doesn't need to be 5000byte swapped --- except if an old-style kernel had byte-swapped the 5001name_len field as part of the V1 structure. 5002 5003 5004E2fsprogs 1.27 (March 8, 2002) 5005============================== 5006 5007The warning messages for mke2fs now go to standard error. 5008 5009Fixed to make sure "make check" runs all of the test programs with 5010LD_LIBRARY_PATH set, so that we test the libraries in the build tree. 5011 5012The mke2fs program checks the boot sector for the BSD disk label, and 5013avoids erasing it if it is there. 5014 5015Fixed a bug in e2fsck which caused it to core dump if the journal 5016inode was missing when it was supposed to be there. 5017 5018E2fsck now prints ranges in pass 5 when printing deltas for the block 5019and inode bitmaps. 5020 5021Debugfs's "ls -l" command now will print out the file type information 5022in the directory entry. 5023 5024Create man pages and hard links for fsck.ext3 and mkfs.ext3. If 5025mke2fs is invoked as mkfs.ext3, create the filesystem with a journal. 5026 5027Debugfs can now examine the experimental directory indexing 5028information. 5029 5030Fixed bug in debugfs which caused it to core dump if modify_inode is 5031called without an open filesystem. 5032 5033The debugfs lsdel command now runs its output through a pager first. 5034 5035When installing manual pages, remove the compressed manual pages first. 5036 5037Synchronized with Debian's packaging information for e2fsprogs-1.26-1. 5038 5039Fix the 2.4 resource limitation workaround introduced in 1.26 which 5040actually broke things on mips32, sparc32, and Alpha platforms. 5041 5042Updated the I18N code so that calls to setlocate(LC_CTYPE, 0) are made 5043(which is required by the newer libintl libraries). 5044 5045Programmer's notes: 5046------------------- 5047 5048Fixed various gcc -Wall complaints. 5049 5050Fixed a few memory leaks in the e2fsck journaling code, and in the 5051ismounted code checking for a swap device. 5052 5053Add new inode I/O abstraction interface which exports an inode as 5054an I/O object. 5055 5056Exported ext2_file_flush as a public interface. 5057 5058In ext2_file_write, we now mark the buffer void * argument as a const, since ext2_file_write doesn't modify the buffer. 5059 5060Lots of small random portability fixes to make e2fsprogs build under 5061AIX --- even without the 5L compatibility toolkit, and even using the 5062uber-crippled AIX native C compiler. 5063 5064 5065E2fsprogs 1.26 (February 3, 2002) 5066================================= 5067 5068Dumpe2fs will keep going now if the bad block inode can't be read. 5069(Previously it stopped with a fatal error.) 5070 5071E2fsck will now give an opportunity to run the journal even if the 5072recovery flag is not set. This is the default behaviour if e2fsck is 5073using a backup superblock, since the needs_recovery flag will never be 5074set in the backup superblock copies. 5075 5076E2fsck now automatically finds the backup superblock/group descriptors 5077even when the primary superblock is completely destroyed for 2k and 4k 5078filesystems. (Previously it just guessed/assumed that we were dealing 5079with a 1k filesystem in that case, and users had to manually specify 5080the backup superblock number.) 5081 5082Fixed a bug in e2fsck where it dereferences a null pointer when there 5083is a problem opening a filesystem in preen mode. 5084 5085E2fsck now handles external journals correctly; previously it trashed 5086the external journal device if the journal needed to be replayed. 5087 5088Work around ulimit incompatibility problem caused by recent 2.4 5089kernels; the unix IO module will automatically try to set any resource 5090limits to be infinite on startup. 5091 5092Fixed a bug in e2fsck where it wasn't allocating a new block or inode 5093bitmap if it wasn't present and the blocksize was 2k or 4k. 5094 (Addresses Debian Bug #116975) 5095 5096E2fsck will check and fix botched ext3 V1 to V2 superblock updates by 5097clearing the new V2 fields if they do not make sense or if the ext3 5098superblock is version 1 superblock. 5099 5100E2fsck will automatically relocate the ext3 journal from a visible 5101file (i.e., /.journal) to an hidden inode if the filesystem has been 5102opened read/write. This allows the users to add a journal while the 5103filesystem is mounted, but the next time the system is rebooted, the 5104journal file will disappear. This avoids problems with backups, 5105stupid operators with superuser bits, etc. 5106 5107Fix a bug in fsck where it would not support a filesystem type of 5108"auto" if the device was specified in terms of LABEL=xxx or UUID=xxx. 5109 5110Fsck now supports fstab entries of the form "ext3,ext2". It will also 5111automatically identify reiserfs filesystems. 5112 5113The number of processes spawned by fsck can now be limited using the 5114FSCK_MAX_INST environment variable. 5115 5116Fsck now searchs the LVM proc hierarchy to find logical volumes which 5117should be searched for UUID and label information. 5118 5119Work around a bug in 2.4.10+ kernels by trying to unset the filesize 5120limit if at all possible, when opening a block device. (The filesize 5121limit shouldn't be applied against writes to a block device.) 5122 5123In mke2fs and e2fsck, specifying the -c option twice will now do 5124a read/write test on the disk. Update the man pages to encourage 5125using the -c option, and to discouraging running badblocks separately, 5126since users tend to forget to set the blocksize when running 5127badblocks. 5128 5129Mke2fs now automatically clears the immutable attribute on a 5130pre-existing .journal file left over from a previous failed attempt to 5131add a journal to an alreadyy-mounted filesystem. 5132 5133Fixed mke2fs's exit codes to consistently indicate when the mke2fs 5134operation failed. 5135 5136Mke2fs now creates the lost+found directory with permissions of 0700, 5137so that files that had previously lived in protected directory are 5138safe if they get dropped in lost+found and the system administrator 5139doesn't deal with immediately. (Addresses Debian #bug 118443) 5140 5141Mke2fs and e2fsck (and all programs that use the 5142ext2fs_check_if_mounted function) will now properly identify that a 5143device is mounted, even in cases where devfs has confused things with 5144multiple devices names with the same identity, or if a dim-witted 5145system administrator has done something stupid like creating device 5146file aliases in their home directory. Also now checks for swap devices 5147by using /proc/swaps. (Addresses Debian bug #106622) 5148 5149Added a new option (-T) to tune2fs which sets the time a filesystem 5150was last checked. 5151 5152Speed up e2image when creating sparse raw image files by optimizing 5153away excess lseek() system calls. 5154 5155Fix support of large (> 2GB) files when used as a filesystem in 5156mke2fs, tune2fs, debugfs, and findsuper. 5157 5158Debugfs's stat and icheck commands now properly deals with large (> 51592GB) files. 5160 5161Debugfs's set_super_value command now prints out the list of valid 5162superblock fields which can be set using the command. 5163 5164Debugfs's rm and kill_file command now updates the superblock free block 5165and inode counts, thus keeping the filesystem consistent. 5166 5167Debugfs's lsdel command now takes an optional argument which allows 5168the user to only see the most recently deleted files. 5169 5170A new command (undel) was added to debugfs which automates 5171undeleting a deleted inode and linking it back to a directory. 5172 5173Debugfs's ls command now takes a new option, -d, which lists 5174deleted directory entries. 5175 5176Debugfs's testb, freeb, setb, and find_free_block commands now take an 5177optional count argument. 5178 5179Add support for a new ext2 file attribute, EXT2_NOTAIL_FL, which will 5180be used to signal that a particular inode should not be eligible for 5181tail-merging --- this is needed for programs like LILO. 5182 5183The findsuper (an unreleased, uninstalled utility program) has been 5184improved to add extra validity checks and to add a progress meter. 5185(It is still an unsupported program; the officially supported way to 5186recover from a trashed partition table is to use gpart; findsuper is 5187for wizards only. :-) 5188 5189Debugfs was fixed to compile with "configure --disable-swapfs". 5190 5191Cleaned up various manual pages. (Addresses Debian bug #119624, #120171) 5192 5193Added new translation file for Turkish. 5194 5195Programmer's notes: 5196------------------- 5197 5198Fix general gcc -Wall complaints. 5199 5200E2fsprogs (mostly) works with the dietlibc. 5201 5202The programming texinfo file has been expanded to include description 5203of additional libext2fs functions. (Still not compeltely done, but a 5204lot of the more important functions have been documented.) 5205 5206Added a umask structure to struct_ext2_filsys, which currently only 5207modifies the behaviour of ext2fs_mkdir(), but if we add any file 5208creation functions to libext2fs, we should also make sure they respect 5209the umask setting. 5210 5211The build-rpm script was fixed to be a bit more safe. 5212 5213The tests' Makefile now has a way of automating test case creation 5214for e2fsck, using "make testnew". 5215 5216Created a new function, ext2fs_dir_iterate2 which passes more 5217information to the callback function (identical to the one used by 5218ext2fs_dblist_dir_iterate). The directory iterator functions take a 5219new flag, DIRENT_FLAG_INCLUDE_REMOVED, which will return deleted 5220directory entries. If the directory entry is deleted, the callback 5221function will be called with the entry paraemter set to 5222DIRENT_DELETED_FILE. 5223 5224Added new functions, ext2fs_inode_alloc_stats and 5225ext2fs_block_alloc_stats, which takes updates block/inode allocation 5226statistics in the bitmaps, block group descriptors, and superblock 5227when inodes or blocks are allocated or deallocated. 5228 5229 5230E2fsprogs 1.25 (September 20, 2001) 5231=================================== 5232 5233This is primarily a bug-fix release; no new features were added, but 5234there are a number of embarassing bug fixes and cleanups applied. 5235 5236Fix a bug mke2fs which causes stack garbage to be written to disk when 5237zapping disk sectors. (This bug could cause mke2fs to core dump for 5238some kernels, I suspect with security enhancement patches.) 5239 5240Remove unneeded #include of <linux/config.h> which was breaking 5241building e2fsprogs on the Hurd. (Addresses Debian bug #112414.) 5242 5243Updated tune2fs man page to reflect the fact that adding or removing a 5244journal doesn't require running e2fsck. 5245 5246Remove use of AC_REQUIRE from autoconf which had been used to prevent 5247AC_CANONICAL_HOST from being called twice; unfortunately this causes 5248recent autoconf to bomb out since they don't allow AC_REQUIRE to be 5249used outside of autoconf macros. Fortunately, it doesn't seem to be 5250necessary any more, anyway. 5251 5252E2fsck will now not fall back to an alternate superblock if the user 5253specifies the superblock location explicitly on the command-line. 5254This allows easier recovery from a situation where the primary 5255superblock and block groups are slightly corrupted, but the backup 5256superblocks are completely broken. 5257 5258Fix problem which caused compile_et and mk_cmds to blow up if 5259e2fsprogs was in a directory with a pathname that contained spaces. 5260 5261We are now more paranoid about checking the entry in /etc/mtab to make 5262sure the filesystem is really mounted, since some broken distributions 5263(read: Slackware) don't initialize /etc/mtab before checking non-root 5264filesystems. (Slackware also doesn't check the root filesystem 5265separately, and reboot if the root filesystem had changes applied to 5266it, which is dangerous and broken, but there's nothing I can do about 5267that.) 5268 5269Make UUID library C++ friendly by adding appropriate extern "C" 5270declarations and using const in the function declarations. 5271 5272Fix up the com_err texinfo file so that it can product a valid info 5273file (previously, it could only be used to generate paper 5274documentation using texinfo.tex). 5275 5276 5277E2fsprogs 1.24a (September 2, 2001) 5278=================================== 5279 5280Fix brown-paper bug in mke2fs which caused it to segfault when 5281printing the version string. 5282 5283 5284E2fsprogs 1.24 (August 30, 2001) 5285================================ 5286 5287Revert the BLKGETSIZE64 support, since for some distributions, this 5288ioctl has been used by an unofficial kernel patch to update the last 5289sector on the disk, and this was causing disk corruption problems as a 5290result. 5291 5292Mke2fs is now more careful about zapping swap space signatures and 5293other filesystem/raid superblock magic values so that programs like 5294mount who try to use hueristics to determine the filesystem type of a 5295partition are less likely to be confused. 5296 5297E2fsck will no longer complain if the the mode of EXT2_RESIZE_INO is a 5298regular file (since Andreas Dilger's on-line resizing tools will set 5299its mode bits). 5300 5301Fixed some minor texinfo, man pages nits for spelling errors, 5302texinfo warnings, etc. (Addresses Debian bug #110621.) 5303 5304E2fsprogs program no longer print the filesystem version number 5305(i.e. 0.5b), since it only confuses people and doesn't serve any real 5306purpose. 5307 5308E2fsck will once again compile under libc5, since it will supply its 5309own version of strnlen if necesssary. 5310 5311mke2fs and tune2fs will allow the use of UUID= or LABEL= specifiers 5312when specifying the external journal device. tune2fs will also search 5313devices looking for the externla journal debice when removing. 5314 5315E2fsprogs 1.23 (August 15, 2001) 5316================================ 5317 5318Add initial support for extended attributes (EA); e2fsck will 5319correctly handle a filesystem with EA's and check the EA blocks for 5320corruptions. 5321 5322E2fsck's symlink sanity checking has been cleaned up. It now checks 5323the i_size value of fast symlinks, and checks for immutable flags 5324being set of symlinks, etc. 5325 5326E2fsck now offers to clear inodes which are special files that have a 5327non-zero i_blocks or i_size field. (The i_size field check was in the 5328previous version of the code, but due to a bug it didn't offer to 5329clear the inode unless i_size and i_size_high were both non-zero.) 5330 5331E2image can now create "raw" image files, which only contain the 5332filesystem metadata placed in a spare file so that e2fsck, dumpe2fs, 5333debugfs, etc., can be run directly on the raw image file. 5334 5335Add support for the 64-bit block device patches. 5336 5337Fixed bugs in creating external journals with a 1k blocksize. 5338 5339Add initial support for external journals (so long as the external 5340journal only supports a single filesystem) in e2fsck. 5341 5342Remove requirement for needing to run fsck on a filesystem after 5343removing a journal (either internal or external). 5344 5345The man pages now document how to create and manage external journals. 5346 5347Speed up the check of an ext3 filesystems by avoiding a needless flush 5348of all of the superblock and block group descriptors. 5349 5350Speed up creating an internal journal using tune2fs in the case where 5351the filesystem has a lot blocks already allocated. 5352 5353Tune2fs has been fixed to make sure that only error messages go to 5354stderr, and normal message go to stdout. (Addresses Debian bug #108555) 5355 5356Fixed a minor bug in mke2fs; if -O none is passed to mke2fs, it will 5357now not set the sparse_super feature. (Addresses Debian bug #108165) 5358 5359Add support in fsck for the filesystem type "auto". 5360 5361Fsck -A will not try to interpret device names for filesystems which 5362have a pass number is 0. (Addresses Debian bug #106696). 5363 5364Fsck prints a warning message if now valid filesystems are passed to 5365it. (Addresses Debian Bug #107458.) 5366 5367E2fsck now gives an explicit warning if there filesystem still has 5368errors at the end of the run. (Addresses Debian bug #104502) 5369 5370E2fsck will set the EXT2_ERROR_FS flag if the journal superblock 5371reflects an error. E2fsck will also not run the orphan list if the 5372filesystem contains errors, since the orphan list might be corrupted. 5373 5374E2fsck now prints the number of large files when given the -v option. 5375 5376Fixed minor memory leaks in e2fsck. 5377 5378Some minor man pages updates. (Addresses Debian bug #30833, #108174) 5379 5380 5381E2fsprogs 1.22 (June 22, 2001) 5382============================== 5383 5384Fixed a bug in e2fsck's handling of orphan inodes which are special 5385files (i.e., block/character device files, named FIFO's, etc.). 5386 5387Fixed a bug accidentally introduced in e2fsprogs 1.21 which caused 5388tune2fs to fail at adding a journal to a mounted filesystem. 5389 5390Fixed a few big-endian bugs in e2fsprogs 5391 * The directory block functions were accidentally reporting 5392 some directories as corrupted when they weren't. 5393 * If e2fsprogs is compiled --disable-swapfs, the C language 5394 equivalents weren't being included for big-endian platforms. 5395 (Fixes Debian bug #101686). 5396 5397Fixed a Hurd compilation problem. (Addresses Debian bug #101361) 5398 5399Programmer's notes: 5400------------------- 5401 5402Use platform independent method of defining the BLKFLSBUF and FDFLUSH 5403ioctls, and allow them to be defined for non-i386 platforms. 5404 5405The uuid.h header file is now protected against multiple inclusions. 5406 5407E2fsprogs is now being developed using BitKeeper. Changed the test 5408scripts to deal with BK's stripping CR characgters from text files, 5409and changed the top-level Makefile.in to avoid including BitKeeper 5410files when generating the source tarball. 5411 5412 5413E2fsprogs 1.21 (June 15, 2001) 5414============================== 5415 5416Added new configure flags which allow a subset e2fsprogs to be built; 5417this is most useful for boot floopies, since the resulting shared 5418libraries and programs are slimmed down by removing features that 5419aren't necessary for a boot floppy. The new flags that were added 5420are: --disable-swapfs, which removes support for byte swapping old 5421legacy PPC/68k filesystems, --disable-debugfs, which removes support 5422for debugfs from the libext2fs library, --disable-imager, which 5423removes support for the e2image program, and --disable-resizer, which 5424removes support for resize2fs. 5425 5426E2fsck now prints the number of mounts or days elapsed since the last 5427check when e2fsck is forced to check an otherwise clean filesystem. 5428 5429Tune2fs now prints an informative message about how often a filesystem 5430will be checked when adding a journal to the filesystem, to remind the 5431user that he/she may want to adjust those parameters using tune2fs 5432-c/-i. 5433 5434Worked around hurd brain-damage which causes e2fsck to sometimes 5435believe a filesystem is the root filesystem based on device numbers 5436(since Hurd doesn't have dev_t's, which is arguably a POSIX.1 5437violation). 5438 5439Fixed a bug introduced in 1.20 which caused e2fsck to abort with an 5440erroneous error with the -F option was specified. 5441 5442Fixed a ext3 recovery bug in the revoke handling; synchronized with 5443ext3 0.7a. 5444 5445Fixed two bugs in e2fsck's handling of dup block handling, dealing 5446with relatively uncommon edge cases: a directory with an indirect 5447block which is claimed by another file, and when the last inode in the 5448filesystem has blocks claimed by another file. 5449 5450E2fsck now checks to see if the i_size field of a fast symlink is too 5451big, and offers to clear the symlink if so. 5452 5453E2fsck now checks to see if i_size_high of special files is non-zero, 5454and offers to clear i_size_high. 5455 5456Fix e2fsck's handling of incompatible journal flags so that the user 5457has chance to abort, and then has the option to clear out the journal 5458entirely. (Addresses Debian bug #98527.) 5459 5460Fixed a bug in fsck which could cause it to core dump if a mix of 5461standard and non-standard device names are used in /etc/fstab. 5462(Debian bug #100559) 5463 5464Fixed a bug in debugfs which caused read errors when copying a file to 5465not be noticed. 5466 5467The debugfs set_super_value command can now modify the s_lastcheck field. 5468 5469Fixed a bug in lsattr and chattr which was accidentally introduced in 54701.20 to support > 2GB files; both lsattr and chattr wasn't reading 5471directories correctly because the change modified the layout of struct 5472dirent to be incompatible with the libe2p shared library. 5473 5474Cleaned up the mke2fs manual page and included a discussion about why 5475it's good to periodically check the filesystem even when journaling is 5476enabled. 5477 5478Programmer's notes: 5479------------------- 5480 5481Fix general gcc -Wall complaints. 5482 5483The types needed by the ext2 header files are now provided by 5484lib/ext2fs/ext2_types.h, instead of include/asm/types.h. 5485 5486Integers are now preferred to longs when trying to find a 32-bit type 5487in ext2_types.h. Also, if linux/types.h has already been defined, 5488don't try to redefine the types. 5489 5490Fixed make depend script so that it automatically corrects the 5491pathname cleanups performed by make -M, so I don't have to fix them up 5492by hand. 5493 5494Fixed the d_loaddump test case to be more robust, and not depend on 5495bash'isms. 5496 5497Removed debugfs's dependence on pread(), which was accidentally 5498intrudced in e2fsprogs 1.20 5499 5500Fixed a performance bug in the libext2fs's icount routine; the size 5501estimate of the icount array was incorrectly being calculated. 5502 5503Removed use of the badblocks compatibility functions in the e2fsprogs 5504programs. 5505 5506Added paranoia code which protects against strange cases where /etc 5507isn't on the root filesystem, or if /etc/mtab doesn't exist. 5508 5509The header file ext2_types.h is now installed. 5510 5511Autoconf is used to determine when we are on big-endian machines, 5512instead of doing run-time tests, to save a few bytes of code. 5513 5514The ext2fs_mark_generic_bitmap and ext2fs_unmark_generic_bitmap 5515functions are no longer inline functions, which saves space and 5516doesn't really cost any real performance. 5517 5518The ext2fs library no longer depends on the e2p library. (What need 5519there was of it --- namely, fsetflags, was coded in-line). 5520 5521Fixed the makefile so that lib/ext2fs/ext2_types.h is generated even 5522when the user is stupid and tries compiling the package using "make 5523install" as root. 5524 5525Miscellaneous code cleanups: 5526 * Added missing files from Makefile.in's SRCS file, so that 5527 their dependencies would be properly calculated. 5528 * Removed redundant code 5529 * Fixed comments in code 5530 * Removed no-longer unneeded argsused #pragma. 5531 5532 5533E2fsprogs 1.20 (May 20, 2001) 5534============================= 5535 5536Add support for replaying the ext3 journal purely in user mode, 5537including handling the orphaned inode list. Used code contributed by 5538Andreas Dilger, with bug fixes and the orphaned inode handling done 5539by Theodore Ts'o. 5540 5541The mke2fs and tune2fs programs can create or modify a filesystem to 5542include an ext3 journal. Tune2fs also can be used to remove an ext3 5543journal from a filesystem. 5544 5545E2fsck will now check for the existence of a linked list of orphan 5546inodes in the superblock, and clear those inodes before starting the 5547rest of the filesystem check (but after the journal playback). 5548 5549E2fsck now validates the file descriptor passed to the -C option, 5550which saves against the completion bar getting written to an 5551unexpected location, such as the disk being checked. (Debian 5552bug/wishlist #55220) 5553 5554E2fsck will now bump the filesystem revision number from zero to one 5555if any of the compatibility bits are set. 5556 5557Fixed a bug where a badly corrupted directory on a big endian system 5558could cause e2fsck to die with a bus error. The 5559ext2fs_read_dir_block() and ext2fs_process_dir_block() functions in 5560the ext2 library now does alignment sanity checks on the rec_len field 5561of the directory entry before using it. 5562 5563The ext2 library has been enhanced to make tune2fs safe to run on 5564mounted filesystems. (Users could usually get away with using tune2fs 5565on mounted filesystems before, but with the advent of ext3 and 5566journaling, it became important to make tune2fs was *really* safe for 5567use even when the filesystem being modified is mounted.) E2label is 5568now implemented by tune2fs using an argv[0] dispatch, so that e2label 5569is also now safe for use on mounted filesystems. 5570 5571Added a new program, e2image, which creates a backup of critical ext2 5572filesystem data structures. The generated image file can be examined 5573using dumpe2fs and debugfs. In the future, e2fsck will be able to use 5574the image file to help recover very badly damaged filesystems. 5575 5576Fixed a number of LFS bugs in e2fsck; very, very large (> 2**42) files 5577no longer cause e2fsck to bomb out. Also treat files > 2GB as being 5578large file when deciding whether or not the filesystem has large files. 5579 5580Fixed lsattr and chattr so that they work correctly on large files. 5581(Fixes Debian bug #72690.) 5582 5583Removed limitation in get_device_size() which imposed a one terrabyte 5584filesystem limitation. (Most 2.2 kernels still have a signed int 5585problem which cause 1 TB block device limitation. Fortunately, the 5586kernel patches to fix this are much easier than fixing the 2TB 5587limitation in the kernel. :-) 5588 5589A max_mount_count of zero is now treated as if no mount count were 5590set. (Previously, no mount count was indicated by using -1, and a 5591mount count of zero caused e2fsck to always be run.) 5592 5593Mke2fs supports two new filesystem types largefile and largefile4. 5594 5595Mke2fs now adds some randomness in s_max_mount_count so that multiple 5596filesystems won't be all checked at the same time under normal 5597operations. 5598 5599Fixed bug in the progress bar printing code which could cause e2fsck 5600to core dump on an illegal filesystem. 5601 5602Fixed bug in fsck which could allow more than one instance of e2fsck 5603to be printing a progress bar. (Debian bug #65267) 5604 5605Fsck using a UUID or a LABEL specifier will work even if devfs is 5606compiled into the kernel and not mounted. If the pathnames in 5607/proc/partitions are incorrect, fsck will search /dev for the correct 5608device (using the new ext2fs_find_block_device library function). 5609Fsck now also checks the RAID devices first so that they are properly 5610found when they are in use. Support has also been added to support 5611additional IDE disks and the DAC 960 device names. (Debian bug #94159) 5612 5613Fixed a bug in fsck which caused it not deal properly with 16 5614byte long filesystem labels. 5615 5616Fsck's -t option has been made a lot more flexible. The semantics for 5617what happens if a comma-separated list to fsck has been regularized, 5618and it is now possible to filter what filesystems will get checked 5619based what is in the filesystem's fstab entry's option field. (Debian 5620bug #89483.) 5621 5622The dumpe2fs program can now print out the group description 5623information in hex, and also prints the location of the superblock and 5624block group descriptor for those block groups that have them. 5625 5626Mke2fs now clears the ext2 superblock before it starts creating a 5627filesystem, so that the superblock magic number is only written if the 5628filesystem creation process successfully completes. 5629 5630The debugfs program's stat command now pretty-prints the blocks used 5631by an inode so that it's more compact and informative. 5632 5633The debugfs stats command now uses the same libe2p code (which is used 5634by dumpe2fs) to print the superblock header information. This is more 5635complete, and it avoids a bit of code duplication. 5636 5637Added a new debugfs command, set_super_value (ssv) which allows the 5638user to set arbitrary superblock fields. 5639 5640Debugfs was extended to support inode numbers in hex (by prefixing 5641them with 0x), and so that modify_inode can set the inode generation 5642number. Also, there is now a new function command called logdump 5643which will dump an ext3 journal. 5644 5645Fixed a bug in debugfs so that quitting out of the pager doesn't kill 5646debugfs. 5647 5648Debugfs's dump command now stops immediately upon reporting a disk 5649read error. (Fixed a bug in ext2fs_file_read library routine which 5650caused debugfs not to stop.) (Debian bug #79163) 5651 5652On systems with /proc/mounts (mainly Linux systems), /proc/mounts is 5653checked before /etc/mtab is used, since /proc/mounts is more likely to 5654be accurate. 5655 5656Added portability fixes for Solaris and Linux/ia64. 5657 5658Various manual pages were clarified and cleaned up. (Fixed debian 5659bugs #63442, #67446, and #87216) 5660 5661 5662Programmer's notes: 5663------------------- 5664 5665The e2fsck message printer now supports %Iu and %Ig, which will print 5666out the inode's user and group owners, respectively. 5667 5668E2fsprogs now includes its own version of include/linux/ext2_fs.h, so 5669that no longer dependent on the system having the correct version of 5670the kernel header files. 5671 5672Added a new function to libext2, ext2fs_find_block_device(), which 5673searches the system (i.e., /dev, /devfs, /devices) for a pathname to a 5674device given its device number. 5675 5676Added a new function to libext2, ext2fs_sync_device, which centralizes 5677all of the places which might try to use the BLKFLSBUF or FDFLUSH 5678ioctls (and usually failing to define them since the system header 5679files don't usually do this for us, and we're trying to avoid usage of 5680kernel include files now). 5681 5682Added new utility programs in tests/progs: random_exercise and 5683hold_inode. They aren't built by default; they're useful for 5684exercising ext3 filesystem code. 5685 5686Added a new ext2 filesystem flag, EXT2_FLAG_SUPER_ONLY, which causes 5687the filesystem close functions to only update the superblock, and to 5688not touch the block group descriptors. Needed by tune2fs when 5689modifying a mounted filesystem. 5690 5691Got rid of struct ext2fs_sb and replaced it with the standard struct 5692ext2_super_block from include/linux/ext2_fs.h. Note: this may break 5693source (but not binary) compatibility of some users of the ext2 5694library. Those applications should just simply do a global search and 5695replace of struct ext2fs_sb with struct ext2_super_block, and include 5696the new header file <ext2fs/ext2_fs.h> which defines it. 5697 5698The ino_t type has been renamed ext2_ino_t to protect applications 5699that attempt to compile -D_FILE_OFFSET_BITS=64, since this 5700inexplicably changes ino_t(!?). So we use ext2_ino_t to avoid an 5701unexpected ABI change. 5702 5703The Makefiles have been reworked so that "make check" can be run from 5704the top-level Makefile. 5705 5706Fix general gcc -Wall complaints and removed dead code. 5707 5708Remove use of NOARGS, because we assume everyone does ANSI C these 5709days. 5710 5711Added build-rpm script from sct. 5712 5713New functions ext2fs_image_{inode,super,bitmap}_{read,write} added 5714to support e2image. 5715 5716New function ext2fs_flush_icache which must be called if the 5717application program modifies the inode table blocks without going 5718through ext2fs_write_inode() interface. 5719 5720New ext2fs_check_mount_point() function, which will return the mount 5721point of a device if mounted. 5722 5723The io_channel abstraction now has an optional interface, 5724io_channel_write_range, which allows specific byte ranges to be 5725written. 5726 5727The unix_io IO channel now supports write-through caching, so that 5728journal creation is more efficient. 5729 5730Added x86 assembly language routines to support byte swapping, to 5731reduce executable size. 5732 5733Fixed bug in the utility program subst so that it's possible to 5734replace a substitution variable with a zero-length string. 5735 5736Fixed numbering e2fsck pass1 problem numbers; an extra zero had 5737slipped into some of the problem number. 5738 5739E2fsprogs 1.19 (July 13, 2000) 5740============================== 5741 5742Release the resize2fs program since the timeout before it could 5743be released under the GPL has finally expired. 5744 5745Add experimental support needed for the ext2 compression patches. 5746This requires compiling e2fsprogs with the --enable-compression flag 5747to the configure script. 5748 5749Added ext3 journalling support. E2fsck will run the journal (if 5750necessary) by temporarily mounting the filesystem. /sbin/fsck.ext3 is 5751installed as a symlink to e2fsck. Fsck has been taught about ext3, 5752and treats it the same as ext2 in terms of the progress bar logic. 5753Dumpe2fs will display the superblock journaling information if the 5754filesystem has a journal. The ext2 library will now permit opening an 5755ext3 filesystem with the recovery flag set. This is necessary for 5756on-line dump's to work correctly, but there may be issues with this 5757working well since ext3 is much less agressive about syncing blocks to 5758the filesystem, since they're safe on the journal. 5759 5760Tune2fs and e2fsck have been changed to allow the mount_count check to 5761be disabled by setting max_mount_count to -1. (This was already 5762supported by the kernel.) 5763 5764Create a symbolic link for fsck.ext3, since the e2fsprogs utilities 5765are used for ext3 as well. 5766 5767Added internationalization support for e2fsprogs; must be enabled 5768by passing --enable-nls to configure. 5769 5770Always use the provided ext2fs header files to insulate ourselves from 5771kernel version changes. Which include files are used by e2fsprogs 5772have also been cleaned up to improve portability. 5773 5774Limit the number of times that e2fsck updates the progress bar so that 5775people who are booting using a 9600 baud console don't get swampped by 5776too many updates. 5777 5778Improved the loop detection algorithm in e2sck's pass #3 so that it is 5779much, much faster for large filesystems with a large number of 5780directories. 5781 5782The memory footprint for e2fsck is now slightly smaller than before. 5783 5784E2fsck now checks if special devices have a non-zero size, and offers 5785to clear the size field if it finds such an inode. 5786 5787E2fsck now checks if special devices have the append-only flag set, 5788and offers to clear the inode. 5789 5790E2fsck now properly handles some "should never fail" cases during a 5791bitmap copy in pass5. 5792 5793E2fsck now properly prints control characters in filenames as ^A .. ^Z. 5794 5795E2fsck now calculates the correct location of the backup superblock in 5796the case of filesystem blocksizes > 1k. 5797 5798Fixed a bug in e2fsck's calculation of the number of inodes_per_block 5799which normally didn't cause problems under most filesystem parameters, 5800but could cause a valid superblock to be rejected in extreme cases. 5801Other checks for validating superblock values were made more 5802stringent. 5803 5804Added non-destructive write testing to the badblocks program, courtesy 5805of David Beattie. The badblocks also now has an option to input the 5806current set of bad blocks, so that known bad blocks are skipped to 5807speed up the badblocks test. There is also a persistent rescan 5808feature which causes badblocks to run until it has completed some 5809number of passes without discovering any new bad blocks. 5810 5811Badblocks now checks to see if the device is mounted and refuses to do 5812the tests involving writing to the device if it is mounted. Also, 5813badblocks now allows the number of blocks to be checked to be 5814defaulted to the size of the partition. 5815 5816Fixed a bug in fsck which didn't allow non-root users to be able to 5817check filesystems if there were any LABEL= or UUID= entries in 5818/etc/fstab. 5819 5820The Hurd doesn't support the filetype filesystem feature. The mke2fs 5821program now makes sure that for the Hurd, the filestype feature is 5822turned off. E2fsck will check to see if the filetype feature is 5823turned on for Hurd filesystems, and offer to turn off the feature. 5824 5825Mke2fs now has a safety check to make sure the number of blocks do not 5826exceed 32 bits even on a 64 bit platform. 5827 5828Really fixed a bug in fsck to allow "fsck -As" to run interactive 5829fsck's. (For those people who like to do interactive fsck's in the 5830/etc/rc scripts!?!) 5831 5832Debugfs has a few new features: the rdump command, which will do a 5833recursive dump of a directory and all of its contents, and the lcd 5834command which does a local chdir (much like the ftp command of the 5835same name). In addition, the debugfs program and the open_filesystem 5836command now takes three new options: -b and -s, which allows the 5837blocksize and superblock location to be specified, and the -c option 5838which is used in catastrophic situations where the block group 5839descriptors are corrupt. If the -c option is specified, debugfs will 5840skip trying to read in the block and inode bitmaps. 5841 5842Debufs's lsdel command was fixed to handle bad blocks in the inode 5843table. 5844 5845A Y2K bug in debugfs's "ls -l" handling was fixed by switching to use 58464 digit years. 5847 5848General improvements in error messages 5849 5850 - Mke2fs prints a sane error message if the partition size is zero 5851 (usually because the partition table wasn't reread by the 5852 kernel due to the partition being busy), instead of "invalid 5853 argument passed to ext2 library while initializing superblock". 5854 5855 - Fsck now prints more self-explanatory message if an invalid UUID= 5856 or LABEL= specification is passed to it. 5857 5858UUID library changed to use the LGPL. 5859 5860Fixed a bug in the UUID library where very rapid calls to the 5861time-based UUID generator could cause duplicate UUID's to be returned. 5862This was not a problem for e2fsprogs, but it could be a problem for 5863other users of the library. 5864 5865Make the UUID library more robust in the face of missing or an 5866improper /dev/urandom or /dev/random files. 5867 5868Added some random portability fixes for Solaris. 5869 5870Some minor man page updates. 5871 5872Fixed a memory leak in the ss library. 5873 5874 5875Programmer's notes: 5876------------------- 5877 5878We now try to use lseek64 and open64 from the LFS if possible. 5879 5880The 3rd parameter in e2p's print_flags is now a flags word, instead of 5881a boolean option. 5882 5883The mark and unmark bitmap functions now return the previous state of 5884the bit that was being changed, which is useful for some speed 5885optimizations. 5886 5887The following functions have been added to enhance the badblocks list 5888handling in libext2fs: ext2fs_write_bb_FILE, ext2fs_read_bb_FILE2, and 5889ext2fs_badblocks_equal. 5890 5891The ext2 header files now have the latest journalling fields to the 5892superblock. 5893 5894The ext2fs_mkdir function in libext2fs now properly backs out of error 5895conditions robustly. 5896 5897Cleaned up makefiles: 5898 - to cleanly compile with the -j flag. 5899 - so distclean removes all generated files. 5900 - so in case of an error while installing header files, the make aborts. 5901 5902Fix test_script so that it works correctly when compiling in the 5903source directory. 5904 5905The random UUID generation routine has been made slightly better in 5906the case where /dev/random doesn't exist. (Use of randomly-based UUID 5907is still not recommended if /dev/random doesn't exist, however; it's 5908better to use the time/ethernet MAC address UUID in this case.) 5909 5910Clean up the build process so it's more friendly in case of missing 5911directories. 5912 5913The ext2fs header file can now be #include'd into C++ programs. 5914 5915The e2p.h header file is now installed. 5916 5917Added workaround to a gawk 3.0.5 bug in lib/ss/mk_cmds. 5918 5919 5920 5921E2fsprogs 1.18 (November 10, 1999) 5922================================== 5923 5924Fix a core dumping bug in e2fsck if an imagic inode is present or 5925(more rarely) if the filesystem is badly corrupted enough that e2fsck 5926has to restart pass 1 processing. E2fsck now closes the filesystem 5927before freeing a large number of its data structures, so in the case 5928of future memory faults, at least the fixed filesystem will be fully 5929written out. 5930 5931If a filesystem doesn't support imagic inodes, and e2fsck discovers an 5932imagic inode, it will offer to clear the imagic flag. 5933 5934E2fsck will now offer to clear the immutable flag on special files 5935(device/socket/fifos) when running it in non-preen mode. 5936 5937E2fsck will now set the filetype when creating /lost+found, and when 5938connected orphaned inodes to /lost+found. 5939 5940Debugfs's ncheck and icheck commands now handles the case where there 5941are bad blocks in the inode table without bombing out. 5942 5943The badblocks list processing code has been made more efficiently for 5944appending a large number of (ordered) badblocks to the badblocks list. 5945 5946Some minor man page updates. 5947 5948Fsck now allows interactive e2fsck's when using fsck -As (not a common 5949mode, but some people like to do this in boot scripts for silly reasons). 5950 5951Programmer's notes: 5952------------------- 5953 5954The internal e2fsck problem code for PR_2_SPLIT_DOT was fixed to meet 5955with the problem code convention. 5956 5957The badblocks list regression test program has been updated to work 5958with previously made API name changes. 5959 5960The ext2fs_free() command now uses the new badblocks API to avoid 5961using the compatibility layer. 5962 5963Added new regression test cases; the run_e2fsck test script now 5964supports the ability for a test case to run a prepratory command 5965before running e2fsck. 5966 5967E2fsprogs 1.17 (October 26, 1999) 5968================================= 5969 5970Fixed nasty typo in fsck which caused parallelized fsck's to go into an 5971infinite loop. 5972 5973Fixed a bug in fsck where it used strncmp to compare a binary UUID, 5974thus potentially causing problems if a binary UUID contained a NULL 5975character. 5976 5977E2fsck now uses stricter checks for directory entries in pass 2: 5978zero-length filenames are not allowed; neither are 8 byte long 5979directory entries. 5980 5981The debugfs "dirty" command now clears the filesystem valid bit. 5982(Previously this just set the dirty-as-in-needs-writing-out-to-disk 5983bit in the in-core superblock image. The new functionality is more 5984what the user expects, and is more useful.) 5985 5986Added a debugging hook to test parallel fsck; if the environment 5987variable FSCK_FORCE_ALL_PARALLEL, then filesystems on the same drive 5988will be checked in parallel when they normally would not be. 5989 5990Programmer's notes: 5991------------------- 5992 5993Fixed some #ifdef's for compilation under the Hurd OS. 5994 5995Fixed minor W2K compatibility problems. 5996 5997Fixed some miscellaneous GCC warnings. 5998 5999 6000E2fsprogs 1.16 (October 22, 1999) 6001================================= 6002 6003Fixed a race condition bug in fsck; when printing a progress bar, if 6004checking multiple filesystems in parallel, it was possible for fsck to 6005send e2fsck a SIGUSR1 signal before e2fsck had installed its signal 6006handler, which would cause it to terminate with a signal 10. 6007 6008E2fsck now properly handles filesystems that have the 6009INCOMPAT_FILETYPE feature turned on. It can be used to convert a 6010filesystem into using or not using FILETYPE feature. 6011 6012E2fsck now properly handles filesystems that have the IMAGIC feature 6013turned on (this is used on Linux AFS servers). 6014 6015The mke2fs program now creates filesystems that have the filetype and 6016sparse_superblock features enabled by default, unless it is run on a 6017pre-2.2 kernel. These features are not supported by a pre-2.2 kernel, 6018so there is now a new flag -O which allows the user to specify with 6019which features she would like to create the filesystem; "mke2fs -O 6020none" will create a filesystsem compatible with 2.0 kernels. 6021 6022The tune2fs program now has a -O option which allows the user to set 6023and reset "safe" filesystem features. Currently, the only ones which 6024allows to be modified are the filetype and sparse_superblock features. 6025Note setting or clearing either feature will require running e2fsck on 6026the filesystem afterwards. (n.b. Clearing the sparse_superblock feature 6027requires that there is enough free space on the filesystem for the 6028extra superblocks which will be created by e2fsck.) 6029 6030Debugfs can now set and print filesystem features in the superblock 6031using the "features" command. Dumpe2fs will print out the complete 6032set of features when listing the superblock. 6033 6034Dumpe2fs has new options -f (force) and -h (header-only). 6035 6036Fixed a bug in e2fsck which could cause the PROGRAMMING ERROR/bonehead 6037message to come up. This could happen when decrementing or 6038incrementing a link count could result in an overflow. 6039 6040Fixed a bug in e2fsck where the block count on the lost+found 6041directory would not be properly incremented when the directory was 6042expanded to the point where an indirect block needed to be allocated. 6043 6044E2fsck now makes some additional sanity checks on the superblock to 6045avoid crashing or giving a memory allocation error if some of the 6046values in the superblock are unresonable (but the superblock otherwise 6047looks valid). 6048 6049Fixed a bug in e2fsck where a very badly corrupted filesystem might 6050require two passes to completely fix the filesystem. This happened if 6051an inode claimed blocks that was part of the filesystem metadata 6052(typically, when garbage was written into an inode table or indirect 6053block, since this kind of filesystem corruption normally doesn't 6054happen otherwise). 6055 6056On the Alpha, glibc declares st_flags although it isn't actually used; 6057the configure script was improved to detect this case so that 6058e2fsprogs can avoid using the non-functional stat field. 6059 6060The manual pages were updated to use a more consistent formatting 6061style consistent with standard Unix man pages. Mke2fs's man page 6062added documentation for a few previously undocumented options. 6063 6064Fixed minor display bugs in tune2fs and mke2fs. 6065 6066Programmer's notes: 6067------------------- 6068 6069Improved portability of e2fsprogs to non-Unix systems (in particular, NT). 6070 6071Added features to parse and print feature strings into the e2p library. 6072(e2p_feature2string, e2p_string2feature, e2p_edit_feature). 6073 6074ext2fs_mkdir() and ext2fs_new_dir_block() now creates directories 6075whose directory entries contain proper filetype information if the 6076filesystem supports it. 6077 6078ext2fs_link() now uses the low 3 bits of its flags parameter to pass 6079the directory entry filetype information. This is used to set the 6080directory entry filetype information if the filesystem supports it. 6081 6082Fixed a bug in ext2fs_expand_dir() where the block count in a 6083directory's inode would not be properly incremented when the directory 6084was expanded to the point where an indirect block needed to be 6085allocated. 6086 6087 6088E2fsprogs 1.15 (July 18, 1999) 6089============================== 6090 6091Add configuration checks so that e2fsprogs will compile cleanly on 6092Linux 2.3 kernels that have renamed i_version to i_generation. 6093 6094E2fsck now prints a progress/completion bar (and not just a simple 6095spinner) if the -C0 option is requested or if it receives a SIGUSR1 6096signal. Fsck will automatically manage the (potentially muliple) 6097e2fsck processes to print completion bars if it is given a -C option, 6098with the right thing happening if multiple filesystems are being 6099checked in parallel. 6100 6101Mke2fs now has better automatic hueristics to determine the filesystem 6102parameters to be used for a particular filesystem. Added a new option 6103-T which allows the user to specify how the filesystem is to be used, 6104which helps mke2fs do a better job selecting the filesystem parameters. 6105 6106Mke2fs now creates revision 1 filesystems by default, and with the 6107sparse superblock feature enabled. The sparse superblock feature is 6108not understood by Linux 2.0 kernels, so they will only allow read-only 6109mounts of filesystems with this sparse superblocks. 6110 6111Fix bug where if /dev/null couldn't be opened (should never happen), 6112e2fsck would hang in a tight loop. 6113 6114Make e2fsck handle the case where /lost+found isn't a directory. 6115 6116E2fsck now uses mallinfo if it exists to get accurate statistics about 6117its memory usage. 6118 6119Fix bug in e2fsck where it wouldn't check to see if a disconnected 6120inode had any problems before connecting it to /lost+found. 6121 6122Add check to e2fsck so it makes sure that total number of inodes in 6123the filesystem is a sane number. 6124 6125Fix fencepost error when clearing an the end of the block bitmap which 6126caused the last block in the bitmap not to get cleared. 6127 6128Cleaned up a number of messages in e2fsck: 6129 * The message "Group's #'s copy of the group descriptor..." 6130 was fixed so that the correct number would be displayed. 6131 * Added missing space in the "disk write-protected" error messsage 6132 * Cleaned up the error message printed when a non-interactive 6133 e2fsck needs to abort a check because the filesystem 6134 appears to be mounted. 6135 6136Added a new command-line utility, uuidgen, which will create and print 6137a UUID. 6138 6139Make debugfs's icheck command more robust by checking to make sure an 6140inode has valid blocks before interarting over the inode's blocks. 6141 6142UUID generation now uses a random-based scheme whenever possible to 6143prevent potential privacy problems. 6144 6145Man pages for all of the UUID functions in the lirbary were added. 6146 6147Fixed bug in fsck so it won't coredump if a filesystem not in 6148/etc/fstab is given to it. 6149 6150Fsck now understands the UUID=xxxx and LABEL=yyyy forms in /etc/fstab 6151that most of the other mount utilities understands. 6152 6153Mke2fs will make a filesystem even if it appears mounted if the force 6154option is given. 6155 6156Dumpe2fs has new command-line options which allow a filesystem expert 6157to specify the superblock and blocksize when opening a filesystem. 6158This is mainly useful when examining the remains of a toasted 6159filesystem. 6160 6161The badblocks program has been updated to display correctly on disks 6162with large block numbers. 6163 6164The badblocks program no longer gives spurious errors when errors 6165occur on non-block boundaries, which is common if the blocksize is 6166larger than 1k. 6167 6168Mke2fs will sync the disk device every MKE2FS_SYNC block groups if the 6169MKE2FS_SYNC environment variable is set. This is to work around a VM 6170bug in the 2.0 kernel. I've heard a report that a RAID user was able 6171to trigger it even using a 2.2 kernel, but hopefully it will not be 6172needed for most Linux 2.2 users. 6173 6174Fixed miscellaneous documentation and man pages. 6175 6176Programmer's notes: 6177------------------- 6178 6179Cleaned up functions such as pass1_get_blocks, pass1_read_inode which 6180in e2fsck's pass1.c really should have been static. 6181 6182The return value of the uuid_compare() function was changed to make it 6183match with the convetions used by strcmp, memcmp, and Paul Leach's 6184UUID sample document. 6185 6186The "make depend" process has now been made more automated; it now 6187automatically word-wraps the dependencies, and only replaces source 6188Makefile.in if there has been a change in the dependencies. Also, a 6189top-level "make depend" now recurses through all the subdirectories 6190automatically. 6191 6192The Makefile in .../util has been changed so that subst is built using 6193the native C compiler during a cross-compilation, since the subst 6194program is only used during the build process. Also add an explicit 6195rule to build util/subst by cd'ing to the correct directory and 6196running Makefile. 6197 6198The man directories are defined in terms mandir, so that the configure 6199script can override the location of the manual pages. 6200 6201The config files have been updated to recognize new machine types for 6202both the i386 and alpha families. 6203 6204Fsck has been modified so that it will accurately create an 6205fsck_instance even when the noexecute flag is set. This allows for 6206accurate debugging of the fsck pass structure. Also, when the verbose 6207flag is given twice, fsck will print debugging information about when 6208fsck is waiting for jobs to finish. 6209 6210 6211E2fsprogs 1.14 (January 9, 1999) 6212================================ 6213 6214Fix the fstab parsing code so that it can handle blank lines and 6215comment characters. Also, missing pass numbers need to be treated as 6216zero. 6217 6218Fixed a bug in e2fsck where under some circumstances (when e2fsck 6219needs to restart processing after fixing an egregious inconsistency) 6220it would try to access already freed memory. 6221 6222E2fsck now prints non-printable characters in directory entries and 6223pathnames using '^' and 'M-' notation. 6224 6225Fixed chattr so that it will ignore symbolic links when doing 6226recursive descent traversals. For both chattr and lsattr, no longer 6227print the version string unless the -V option is given. 6228 6229Allow the system administrator to directly specify the number of 6230inodes desired in the filesystem, for some special cases where this is 6231necessary. 6232 6233Fix portability problems so that e2fsprogs can be compiled under Linux 62341.2 systems and Solaris systems. 6235 6236Update the config.guess file with a more recent version that will 6237identify newer Linux platforms. 6238 6239Programmer's notes 6240------------------ 6241 6242Ext2fs_read_inode and ext2fs_write_inode will now return an error if 6243an inode number of zero is passed to them. 6244 6245E2fsprogs 1.13 (December 15, 1998) 6246================================== 6247 6248Fixed a bug in debugfs where an error messages weren't getting printed 6249when the ext2 library routines to read inodes returned errors in the 6250stat, cmri and rm commands. 6251 6252Fixed a bug in mke2fs so that if a ridiculous inode ratio parameter is 6253provided, it won't create an inode table smaller than the minimum 6254number of inodes required for a proper ext2 filesystem. 6255 6256Fsck now parses the /etc/fstab file directly (instead of using 6257getmntent()), so that it can distinguish between a missing pass number 6258field and pass number field of zero. This caused problems for 6259diskless workstations where all of the filesystems in /etc/fstab have 6260an explicit pass number of zero, and fsck could not distinguish this 6261from a /etc/fstab file with missing pass numbers. 6262 6263E2fsck will create a /lost+found directory if there isn't one in the 6264filesystem, since it's safer to create the lost+found directory before 6265it's needed. 6266 6267Fixed e2fsck so that it would detect bogus immutable inodes which 6268happen to be sockets and FIFO files, and offer to clear them. 6269 6270If a filesystem has multiple reasons why it needs to be checked, and 6271one of the reasons is that it is uncleanly mounted, e2fsck will print 6272that as the reason why the filesystem is being checked. 6273 6274Cleaned up the output routines of mke2fs so that it doesn't overflow 6275an 80 column display when formating really big filesystems. 6276 6277Added a sanity check to e2fsck to make sure that file descriptors 0, 62781, 2 are open before opening the hard disk. This avoids a problem 6279where a broken program might exec e2fsck with those file descriptors 6280closed, which would cause disastrous results if the kernel returns a 6281file descriptor for the block device which is also used by FILE * 6282stdout. 6283 6284Fixed up the e2fsck progress reporting functions so that the values 6285reliably reach 100% at the completion of all of the e2fsck passes. 6286 6287Fixed minor documentation bugs in man pages and usage messages. 6288 6289Programmer's notes: 6290------------------- 6291 6292Fixed a number of lint warnings in the ext2fs library and potential 6293portability problems from other OS's header files that might define 6294CPP macros for names like "max" and "min". 6295 6296ext2fs_badblocks_list_add() has been made more efficient when it needs 6297to grow the bad blocks list. 6298 6299Fixed a bug in e2fsck which caused it to dereference a freed pointer 6300just before exiting. 6301 6302Fixed the substition process for generating the mk_cmds and compile_et 6303scripts so that they will work outside of the build tree. 6304 6305Add sanity check to e2fsck so that if an internal routine 6306(ext2fs_get_dir_info) returns NULL, avoid dereferencing the pointer 6307and causing a core dump. This should never happen, but... 6308 6309E2fsprogs 1.12 (July 9, 1998) 6310================================== 6311 6312E2fsprogs now works with glibc (at least with the version shipped wtih 6313RedHat 5.0). The ext2fs_llseek() function should now work even with 6314i386 ELF shared libraries and if llseek() is not present. We also 6315explicitly do a configure test to see if (a) llseek is in libc, and 6316(b) if llseek is declared in the system header files. (See standard 6317complaints about libc developers don't understand the concept of 6318compatibility with previous versions of libc.) 6319 6320The ext2fs library now writes out the block group number in each of 6321the superblock copies. This makes it easier to automatically 6322determine the starting block group of the filesystem when the block 6323group information is trashed. 6324 6325Added support for the EXT2_FEATURE_INCOMPAT_FILETYPE feature, 6326which means that e2fsprogs will ignore the high 8 bits of the 6327directory entry's name_len field, so that it can be used for other 6328purposes. 6329 6330Added support for the EXT2_FEATURE_RO_COMPAT_LARGE_FILE feature. 6331E2fsprogs will now support filesystems with 64-bit sized files. 6332 6333Added support for the EXT2_FEATURE_COMPAT_DIR_PREALLOC feature. 6334 6335Added new program "e2label", contributed by Andries Brouwer. E2label 6336provides an easy-to-use interface to modify the filesystem label. 6337 6338Fixed bug so that lsattr -v works instead of producing a core dump. 6339 6340Fixed a minor bug in mke2fs so that all groups with bad superblock 6341backup blocks are printed (not just the first one). 6342 6343Mke2fs will check the size of the device, and if the user specifies a 6344filesystem size larger than the apparent size of the device it will 6345print a warning message and ask if the user wants to proceed. 6346 6347E2fsck has a new option -C, which sends completion information to the 6348specified file descriptor. For the most part, this is intended for 6349programs to use, although -C 0 will print a spinning character to the 6350stdout device, which may be useful for users who want to see something 6351happening while e2fsck goes about its business. 6352 6353Fixed a bug in e2fsck which could cause a core dump when it needs to 6354expand the /lost+found directory, and sometimes the bitmaps haven't 6355been merged in. Also fixed a related bug where ext2fs_write_dir_block 6356was used to write out a non-directory block. (Which would be bad on a 6357non-Intel platform with byte swapping going on.) 6358 6359Fixed bug in e2fsck where it would print a "programming error" message 6360instead of correctly identifying where a bad block was in used when 6361the bad block was in a non-primary superblock or block group 6362descriptor. Also fixed a related bug when sparse superblocks are in 6363use and there is a bad block where a superblock or block group 6364descriptor would have been in a group that doesn't include a 6365superblock. 6366 6367Fixed a bug in e2fsck (really in libext2fs's dblist function) where if 6368the block group descriptor table is corrupt, it was possible to try to 6369allocate a huge array, fail, and then abort e2fsck. 6370ext2fs_get_num_dirs() now sanity checks the block group descriptor, 6371and subsitutes reasonable values if the descriptors are obviously bogus. 6372 6373If e2fsck finds a device file which has the immutable flag set and the 6374i_blocks beyond the normal device number are non-zero, e2fsck will 6375offer to remove it, since it's probably caused by garbage in the inode 6376table. 6377 6378When opening a filesystem, e2fsck specially checks for the EROFS error 6379code, and prints a specific error message to the user which is more 6380user friendly. 6381 6382If the filesystem revision is too high, change the e2fsck to print 6383that this is either because e2fsck is out of date, or because the 6384superblock is corrupt. 6385 6386E2fsck now checks for directories that have duplicate '.' and '..' 6387entries, and fixes this corruption. 6388 6389E2fsck no longer forces a sync of the filesystem (with attendant sleep 6390calls) at all times. The ext2fs_flush() function now performs a sync 6391only if it needed to write data blocks to disk. 6392 6393Fixed a minor bug in e2fsck's pass1b's file cloning function, where 6394certain errors would not be properly reported. 6395 6396Updated and expanded a few points in the man pages which users 6397complained wheren't explicit enough. 6398 6399Added special case byte-swapping code if compiling on the PowerPC, to 6400accomodate the strange big-endian variant of the ext2 filesystem that 6401was previously used on the PowerPC port. 6402 6403 6404Programmer's notes: 6405------------------- 6406 6407Removed C++ keywords from the ext2fs libraries so that it could be 6408compiled with C++. 6409 6410E2fsck's internal organization has now been massively reorganized so 6411that pass*.c don't have any printf statements. Instead, all problems 6412are reported through the fix_problem() abstraction interface. E2fsck 6413has also been revamped so that it can be called as a library from a 6414application. 6415 6416Added new fileio primitives in libext2fs for reading and writing 6417files on an unmounted ext2 filesystem. This interface is now used by 6418debugfs. 6419 6420Added a new libext2fs function for mapping logical block numbers of 6421a file to a physical block number. 6422 6423Added a new libext2fs function, ext2fs_alloc_block(), which allocates 6424a block, zeros it, and updates the filesystem accounting records 6425appropriately. 6426 6427Added a new libext2fs function, ext2fs_set_bitmap_padding(), which 6428sets the padding of the bitmap to be all one's. Used by e2fsck pass 5. 6429 6430The libext2fs functions now use a set of memory allocation wrapper 6431functions: ext2fs_get_mem, ext2fs_free_mem, and ext2fs_resize_mem, 6432instead of malloc, free, and resize. This makes it easier for us to 6433be ported to strange environments where malloc, et. al. aren't 6434necessarily available. 6435 6436Change the libext2fs fucntion to return ext2-specific error codes 6437(EXT2_DIR_EXISTS and EXT2_DB_NOT_FOUND, for example) instead of using 6438and depending on the existence of system error codes (such as EEXIST 6439and ENOENT). 6440 6441Renamed io.h to ext2_io.h to avoid collision with other OS's header 6442files. 6443 6444Add protection against ext2_io.h and ext2fs.h being included multiple 6445times. 6446 6447The types used for memory lengths, etc. have been made more portable. 6448In generla, the code has been made 16-bit safe. Added Mark 6449Habersack's contributed DOS disk i/o routines. 6450 6451Miscellaneous portability fixes, including not depending on char's 6452being signed. 6453 6454The io_channel structure has a new element, app_data, which is 6455initialized by the ext2fs routines to contain a copy of the filesystem 6456handle. 6457 6458ext2fs_check_directory()'s callback function may now return the error 6459EXT2_ET_CALLBACK_NOTHANDLED if it wishes ext2fs_check_directory() to 6460really do the checking, despite the presence of the callback function. 6461 6462 6463E2fsprosg 1.11 (June 17, 1997) 6464============================== 6465 6466Fixed e2fsck to detect (previously ignored) conflicts between the 6467superblock or block group descriptors and block bitmaps, inode 6468bitmaps, and inode tables. 6469 6470Fixed bug in e2fsck so that when the message printed out when a block 6471or inode bitmap conflicts with other data, it has the correct group 6472number. 6473 6474Fixed bug in e2fsck and mke2fs where the blocksize wasn't being passed 6475to badblocks. This meant that not all of the filesystem was being 6476tested for bad blocks! 6477 6478Fixed an array boundary overrun case which cropped up in 6479ext2fs_badblocks_list_test when a user tried running "mke2fs -c 6480-b 4096". 6481 6482Adjusted the number of columns printed by mke2fs when displaying the 6483superblock backups to avoid running over 80 columns when making a 6484really big filesystem. 6485 6486Fixed up the man pages for e2fsck, debugfs, badblocks, chattr, 6487dumpe2fs, fsck, mke2fs, and tune2fs (typos and other minor grammar 6488fixes), thanks to some suggestions from Bill Hawes (whawes@star.net). 6489 6490Programmer's notes: 6491------------------- 6492 6493Fixed install rule in lib/ss so that ss_err.h is actually getting 6494installed. 6495 6496Fixed bug in ext2fs_copy_bitmap; the destination bitmap wasn't getting 6497bassed back to the caller. 6498 6499Fixed bug in ext2fs_inode_scan_goto_blockgroup; it had not been 6500setting the current inode number (which meant this function wasn't 6501working at all). 6502 6503Fixed bug in ext2fs_resize_generic_bitmap; it had not be zeroing all 6504blocks in the bitmap when increasing the size of the bitmap. 6505 6506Changed the initial number of blocks allocated by ext2fs_init_dblist() 6507to be more realistic. 6508 6509Added a new function ext2fs_allocate_group_table, which sets up the 6510group descriptor information (and allocates inode and block bitmaps, 6511and inode tables for a particular group). The function was created by 6512factoring out code form ext2fs_allocate_tables(). 6513 6514Added a new function ext2fs_move_blocks which takes a bitmap of the 6515blocks to be moved, and moves them to another location on the 6516boardboard. 6517 6518Make the unix_io channel's io_channel_flush implementation calls sync() 6519to to flush the kernel buffers to disk. 6520 6521Added a new function ext2fs_dblist_count returns the number of 6522directory blocks in dblist. 6523 6524 6525E2fsprogs 1.10 (April 24, 1997) 6526=============================== 6527 6528Mke2fs once again defaults to creating revision #0 filesystems, since 6529people were complaining about breaking compatibility with 1.2 kernels. 6530Warning messages were added to the mke2fs and tune2fs man pages that 6531the sparse superblock option isn't supported by most kernels yet (1.2 6532and 2.0 both don't support parse superblocks.) 6533 6534Added new flag to mke2fs, -R <raid options>, which allows the user to 6535tell mke2fs about the RAID configuration of the filesystem. Currently 6536the only supported raid option is "stride" which specifies the width 6537of the RAID stripe. 6538 6539Fixed bug in e2fsck where pass1b would bomb out if there were any 6540blocks marked bad in the inode table. 6541 6542Fixed rare bug in mke2fs where if the user had a very unlucky number 6543of blocks in a filesystem (probability less than .002) the resulting 6544filesystem would be corrupt in the last block group. 6545 6546Fixed bug where if e2fsck tried to allocate a block to fix a 6547filesystem corruption problem and the filesystem had no free blocks, 6548ext2fs_new_block() would loop forever. 6549 6550The configure script now checks explicitly to see if "-static" works, 6551since that can't be assumed to be true --- RedHat doesn't install 6552libc-static by default. 6553 6554Fixed bug in libext2's block iterator functions where under some 6555cirmcustances, file with holes would cause the bcount parameter to the 6556callback function to be incorrect. This bug didn't affect any of 6557e2fsprogs programs, but it was discovered by Paul Mackerras, the 6558author of the PPC boot loader. 6559 6560Removed use of static variables to store the inode cache in libext2fs. 6561This caused problems if more than one filesystem was accessed via 6562libext2fs (static variables in libraries are generally a bad idea). 6563Again, this didn't affect e2fsprogs programs, but it was discovered by 6564Paul Mackerras. 6565 6566Fixed minor bugs and version code drift to assure that e2fsprogs 1.10 6567will compile cleanly with 1.2.13 kernels (even with a.out shared 6568libraries!) 6569 6570Programmer's notes: 6571------------------- 6572 6573Added new functions to duplicate an ext2 filesystem handle, and its 6574associated substructure. New functions: ext2fs_dup_handle(), 6575ext2fs_copy_dblist(), ext2fs_badblocks_copy(), ext2fs_copy_bitmap(). 6576Other structures, such as the io_channel and the inode_cache, now have 6577a ref count so that they only get freed when they are no longer used 6578by any filesystem handle. (These functions were added as part of the 6579development effort for an ext2 resizer). 6580 6581E2fsprogs 1.09 (April 14, 1997) 6582=============================== 6583 6584Fixed bug in mke2fs (really in lib/ext2fs/initialize.c) which was 6585accidentally introduced in the 1.08 release. The overhead calculation 6586was accidentally removed, which caused ext2fs_initialize() to not 6587notice when the filesystem size needed to be adjusted down because 6588there wasn't enough space in the last block group. 6589 6590Fixed bug in version parsing library routine; it was always parsing 6591the library version string, instead of using the passed-in string. 6592 6593Clarified chattr man page. 6594 6595E2fsprogs 1.08 (April 10, 1997) 6596=============================== 6597 6598E2fsck 1.07 was very slow when checking very large filesystems with a 6599lot of files that had hard links (i.e., news spools). This was fixed 6600by seriously revamping the icount abstraction. Added a formal test 6601suite for the icount abstraction. 6602 6603Debugfs now has a "-l" option to the "ls" command, which lists the 6604inode number, permissions, owner, group, size, and name of the files 6605in the directory. 6606 6607Fix a bug in e2fsck where when a directory had its blocks moved to 6608another location during the pass 1b processing, the directory block 6609list wasn't updated, so pass 2 wouldn't check (and correct) the 6610correct directory block. 6611 6612E2fsck will now treat inodes which contain blocks which are claimed by 6613the filesystem metadata by treating them as multiply claimed blocks. 6614This way, the data in those blocks can be copied to a new block during 6615the pass 1b--1d processing. 6616 6617E2fsck will attempt to determine the correct superblock number and 6618display it in the diagnostic and warning messages if possible. 6619 6620Add support for a new (incompatible) feature, "sparse_super". This 6621feature reduces the number of blocks which contain copies of backup 6622superblocks and block group descriptors. (It is only an incompatible 6623feature because of a bug in ext2_free_blocks.) mke2fs and tune2fs now 6624support a new -s option; e2fsck will recognize filesystems built with 6625this feature turned on. 6626 6627E2fsck now checks the library to make sure is the correct version, 6628using new library functions. (This helps to diagnose incorrectly 6629installed e2fsprogs distributions.) 6630 6631Dumpe2fs now prints more information; its now prints the the 6632filesystem revision number, the filesystem sparse_super feature (if 6633present), the block ranges for each block group, and the offset from 6634the beginning of the block group. 6635 6636Mke2fs now distributes the inode and block bitmap blok so that the 6637won't be concentrated in one or two disks in RAID/striping setups. 6638Also, if the user chooses a 2k or 4k block group, mke2fs will try to 6639choose the largest blocks per group that be chosen. (For 2k blocks, 6640you can have up to 16384 blocks/group; for 4k blocks, you can have up 6641to 32768 blocks/group.) Previously mke2fs would not allow 6642specification of more than 8192 blocks per group, even if you were 6643using a 2k or 4k block group. 6644 6645Programmer's notes: 6646------------------- 6647 6648Added a new function ext2fs_create_icount2() which takes a "hint" 6649argument. This hint argument presets the icount array with the list 6650of inodes which actually need to be in the icount array. This really 6651helps to speed up e2fsck. 6652 6653Added a new function ext2fs_icount_validate() which checks the rep 6654invariant for the icount structure. This is used mostly for testing. 6655 6656The error mesasage given when a bad inode number is passed to 6657test_generic_bitmap to reflect EXT2FS_TEST_ERROR (instead of 6658EXT2FS_UNMARK_ERROR). 6659 6660Added a new function ext2fs_set_dir_block which sets the block of a 6661dblist entry, given the directory inode and blockcnt. 6662 6663Added a new function ext2fs_get_library_version() which returns the 6664current library version, and ext2fs_parse_version_string() which 6665returns a version number based on a e2fsprogs version string. 6666 6667The icount functions will return EINVAL if the passed in inode number 6668is out of bounds. 6669 6670E2fsprogs 1.07 (March 9, 1997) 6671============================== 6672 6673E2fsck is now uses much less memory when checking really large 6674filesystems (or rather, filesystems with a large number of inodes). 6675Previously a filesystem with 1 million inodes required 4 megabytes of 6676memory to store inode count statistics; that storage requirement has 6677now been reduced to roughly half a megabyte. 6678 6679E2fsck can now properly deal with bad blocks appearing inside the 6680inode table. Instead of trying to relocate the inode table (which 6681often failed because there wasn't enough space), the inodes in the bad 6682block are marked as in use. 6683 6684E2fsck will automatically try to use the backup superblocks if the 6685primary superblocks have a bad magic number or have missing meta-data 6686blocks (or meta-data blocks which are out of range). 6687 6688E2fsck's pass 3 has been made more efficient; most noticeable on 6689filesystems with a very large number of directories. 6690 6691Completely revamped e2fsck's system of printing problem reports. It 6692is now table driven, to make them more easily customizeable and 6693extendable. Error messages which can be printed out during preen mode 6694are now one line long. 6695 6696Fixed e2fsck's filesystem swapping code so that it won't try to swap 6697fast symbolic links or deleted files. 6698 6699Fixed e2fsck core dumping when fixing a filesystem which has no 6700directories (not even a root directory). 6701 6702Added a check to e2fsck to make sure that the length of every 6703directory entry is a multiple of 4 (since the kernel complains if it 6704isn't). 6705 6706Added a check to e2fsck to make sure that a directory entry isn't a 6707link to the root directory, since that isn't allowed. 6708 6709Added a check to e2fsk to now make sure the '.' and '..' directory 6710entries are null terminated, since the 2.0 kernel requires it. 6711 6712Added check to write_bitmaps() to make sure the superblock doesn't get 6713trashed if the inode or block bitmap is marked as being block zero. 6714 6715Added checking of the new feature set fields in the superblock, to 6716avoid dealing with new filesystem features that this package wasn't 6717set up to handle. 6718 6719Fixed a fencepost error in ext2fs_new_block() which would occasionally 6720try to allocate a block beyond the end of a filesystem. 6721 6722When the UUID library picks a random IEEE 802 address (because it 6723can't find one from a network card), it sets the multicast bit, to 6724avoid conflicting with a legitimate IEEE 802 address. 6725 6726Mke2fs now sets the root directory's owner to be the real uid of the 6727user running mke2fs. If the real uid is non-zero, it also sets 6728the group ownership of the root directory to be the real group-id of 6729the user running mke2fs. 6730 6731Mke2fs now has more intelligent error checking when it is given a 6732non-existent device. 6733 6734When badblocks is given the -vv option, it now updates the block that 6735it is currently testing on every block. 6736 6737Fixed a bug in fsck where it wouldn't modify the PATH envirnoment 6738currently correctly if PATH wasn't already set. 6739 6740Shared libraries now built with dependencies. This allows the shared 6741library files to be used with dlopen(); it also makes the transition 6742to libc 6 easier, since ld.so can tell which libc a particular shared 6743library expects to use. 6744 6745Programmer's notes: 6746------------------- 6747 6748Added new abstraction (defined in dblist.c) for maintaining a list of 6749blocks which belongs to directories. This is used in e2fsck and other 6750programs which need to iterate over all directories. 6751 6752Added new functions which test to see if a contiguous range of blocks 6753(or inodes) are available. (ext2fs_*_bitmap_range). 6754 6755Added new function (ext2_inode_has_valid_blocks) which returns true if 6756an inode has valid blocks. (moved from e2fsck code). 6757 6758Added new function (ext2fs_allocate_tables) which allocates the 6759meta-data blocks as part of initializing a filesystem. (moved from 6760mke2fs code). 6761 6762Added a new I/O manager for testing purposes. It will either allow a 6763program to intercept I/O requests, or print debugging messages to 6764trace the activity of a program using the I/O manager. 6765 6766The badblocks_list functions now store the bad blocks in a sorted 6767order, and use a binary search to speed up badblocks_list_test. 6768 6769The inode scan function ext2fs_get_next_inode() may now return a soft 6770error returns: MISSING_INODE_TABLE and BAD_BLOCK_IN_INODE_TABLE in 6771those cases where part of an inode table is missing or there is a bad 6772block in the inode table. 6773 6774Added a new function (ext2fs_block_iterate2) which adds new arguments to 6775the callback function to return a pointer (block and offset) to the 6776reference of the block. 6777 6778Added new function (ext2fs_inode_scan_goto_blockgroup) which allows an 6779application to jump to a particular block group while doing an inode 6780scan. 6781 6782The badblocks list functions were renamed from badblocks_* to 6783ext2fs_badblocks_*. Backwards compatibility functions are available 6784for now, but programs should be modified to use the new interface. 6785 6786Some of the library functions were reorganized into separate files to 6787reduce the size of some programs which statically link against the 6788ext2 library. 6789 6790Put in some miscellaneous fixes for the Alpha platform. 6791 6792 6793E2fsprogs 1.06 (October 7, 1996) 6794================================ 6795 6796Fixed serious bug in e2fsck: if the block descriptors are bad, don't 6797smash the backup copies in ext2fs_close(). (The problem was that when 6798e2fsck -p discovered the problem, while it was closing the filesystem 6799and exiting, it was also blowing away the backup superblocks on the 6800disk, which was less than friendly.) We now make it the case that we 6801only write out the backup superblock and the back block descriptors if 6802the filesystem is completely free from problems. 6803 6804Fixed a bug in block_interate in the lib/ext2fs library which caused 6805e2fsck to fail on GNU Hurd-created filesystems. 6806 6807Add support for Linux/FT's bootloader, which actually uses 6808EXT2_BOOT_LOADER, and sets its mode bits which caused e2fsck to want 6809to clear the inode. 6810 6811Add support for the "A" (no atime update) attribute. (Note: this 6812attribute is not yet in production kernels.) 6813 6814The test suite is not automatically run when doing a "make all" from 6815the top level directory. Users should manually run "make check" if 6816they wish to run the test suite. 6817 6818Upon a preenhalt(), make the printed message more explicit that 6819running e2fsck "MANAULLY" means without the -p or -a options. 6820 6821In e2fsck, if a disconnected inode is zero-length, offer to clear it 6822instead of offering to connect it to lost+found. 6823 6824In e2fsck, if a filesystem was just unmounted uncleanly, and needs 6825e2fsck to be run over it, change e2fsck to explicitly display this 6826fact. 6827 6828For dumpe2fs and e2fsck, cause the -V option to print out which 6829version of the ext2fs library is actually getting used. (This will 6830help detect mismatches of using a 1.06 utility with a 1.05 library, 6831etc.) 6832 6833Programmers' notes: 6834------------------- 6835 6836EXT2_SWAP_BYTES was changed to EXT2_FLAG_SWAP_BYTES, which better fits 6837the naming convention. 6838 6839In ext2fs_initialize(), make sure the description for the inode bitmap 6840is correctly initialize. 6841 6842Fixed minor type typo in ext2fs_allocate_generic_bitmap(); 6843 6844E2fsprogs 1.05 (September 7, 1996) 6845================================== 6846 6847Add support for new fields in the ext2 superblock --- volume name, 6848volume UUID, and last mounted field. Dumpe2fs displays these fields, 6849tune2fs and mke2fs allows you to set them. E2fsck will automatically 6850generate a UUID for those volumes that don't have them. 6851 6852Put in support for e2fsck to recognize HURD specific ext2 features --- 6853most notably, the translator block. The e2fsprogs tools will now use 6854the creator_os field in the superblock to correctly handle different 6855OS-specific variants of the ext2 filesystem. 6856 6857E2fsck now fixes inodes which have a the deletion time set, but which 6858have a non-zero i_link_count field by offering to clear the deletion 6859time. Previously e2fsck assumed that the inode was deleted (per 0.3c 6860ext2 kernel behavior) and offered to unlink the file. 6861 6862If e2fsck sets the clean bit, but nothing else, set the exit code 6863FSCK_NONDESTRUCT. After all, e2fsck did fix a filesystem error --- it 6864set the filesystem valid bit when it was previously cleared. :-) This 6865was needed to make the HURD fsck driver happy. 6866 6867If the user refuses to attach an unattached inode, e2fsck will no 6868longer set the inode's link count. Otherwise, the inode would end up 6869getting marked as unused, which might cause loss of data later. 6870 6871Make the message issued by e2fsck when the superblock is corrupt less 6872confusing for users. It now mentions that another reason for the 6873"corrupt superblock" message might be that the partition might not be 6874an ext2 filesystem at all (it might swap, msdos filesystem, ufs, etc.) 6875 6876Make the libext2 library more robuest so that e2fsck won't coredump on 6877an illegal superblock where the blocksize is zero. (f_crashdisk is 6878the test case). 6879 6880By default, create filesystems where the default checkinterval is 6 6881months (180 days). Linux servers can be robust enough that 20 reboots 6882can be a long, long time. 6883 6884Added configure flag --enable-old-bitops, which forces the bitops to 6885use the old (native) bitmask operations. By default on the sparc 6886platform, the standard ext2 bit ordering is now used. 6887 6888Added a new feature to e2fsck to byte-swap filesystems; this can be 6889used to convert old m68k filesystems to use the standard byte-order 6890storage for the superblock, inodes, and directory blocks. This 6891function is invoked by using the '-s' option to e2fsck. 6892 6893Debugfs's "dump" command has been enhanced so that it writes out the 6894exact size of the file so that the nulls at the end of the file are 6895eliminated. The command also accept a new "-p" option which will 6896attempt preserve to preserve the ownernship, permissions, and 6897file modification/access times. 6898 6899Debugfs has two new options, -f and -R. The -R option allows the user 6900to execute a single debugfs command from the command line. The -f 6901option allows the user to specify a "command file" containing debugfs 6902commands which will get executed. 6903 6904Dumpe2fs now pretty prints the check interval, instead of just 6905printing the check interval as a number of seconds. 6906 6907Fix bugs in debugfs: the params command when no filesystem is opened 6908no longer causes a core dump. It is now possible to unlink a file 6909when a pathame containing a '/' is specified. 6910 6911Tune2fs has a new -C option which sets the number of times the 6912filesystem has been mounted. 6913 6914Fix the chattr '-v' option so that it actually works. Chattr was 6915being buggy about the -v option parsing. 6916 6917Programmers' notes: 6918------------------- 6919 6920The directory lib/uuid contains a set of library routines to generate 6921DCE compatible UUIDs. 6922 6923Extended ext2fs_namei() to handle symbolic links. Added new function 6924ext2fs_nami_follow() which will follow last symbolic link in the case 6925where the pathname points to a sym link. 6926 6927The ext2fs_block_iterate function will now return the HURD translator 6928block, if present. The new flag BLOCK_FLAG_DATA_ONLY will cause the 6929iterator to return data blocks only. The ext2fs.h file now defines 6930constants BLOCK_COUNT_IND, BLOCK_COUNT_DIND, BLOCK_COUNT_TIND, and 6931BLOCK_COUNT_TRANSLATOR, which are the magic values passed in the block 6932count field of the iterator callback function. 6933 6934The test script driver now takes an optional second argument, which is 6935the test case to be run. This allows you to run a test case without 6936needing to run the entire test suite. 6937 6938On Linux ELF systems, install the .so files in the correct places 6939(/usr/lib). The .so files must be stored in the same directory as the 6940.a files. 6941 6942Fixed miscellaneous HURD compilation issues with header file being 6943included in the right order. 6944 6945Fixed debugfs so that it resets optind to zero, not one, since setting 6946optind to zero is more correct. 6947 6948 6949E2fsprogs 1.04 (May 16, 1996) 6950============================= 6951 6952First "official" (1.03 was a limited release only) to support building 6953e2fsprogs under Linux 2.0 kernels (as well as late model 1.3 and 1.99 6954kernels). 6955 6956This package includes a RPM specs file, that it can be built using the 6957RedHat Package Manager. 6958 6959E2fsck now prints a hint that if there are lots of incorrectly located 6960inode bitmaps, block bitmaps, and inode table blocks, the user might 6961want to try using e2fsck -b 8193 first, to see if that fares any 6962better. 6963 6964For ext2 filesystem written with the hurd, debugfs will now print out 6965the translator field when printing an inode structure. 6966 6967Lots of miscellaneous linking/installation cleanups: 6968 6969 Libraries are now linked using a relative pathname, instead of 6970 relying on -L working correct. It doesn't, in many cases, including 6971 current versions of GNU ld. This guarantees that the build tree is 6972 linking with the right libraries, instead of the ones installed in 6973 /usr/lib. 6974 6975 Header files, man pages, and the et/ss shell scripts are now 6976 generated using a custom substitution script, instead of relying on 6977 the configure script. This prevents needless recompilation of 6978 files; in addition, the custom substitution script is much faster. 6979 6980 e2fsck may now be linked dynamically, by using the 6981 --enable-dynamic-e2fsck flag to configure. This is not recommended, 6982 since it increases e2fsck's dependence on other files, but some 6983 people need to save disk space, and other critical programs on their 6984 systems were being linked dynamically anyway. 6985 6986 Programs such as fsck which didn't need to be linked against 6987 libext2fs (or mke2fs which didn't need to be linked against libe2p) 6988 only link against libraries they actually need. Otherwise, those 6989 programs would require the presense of libraries that otherwise 6990 could be removed from a rescuse diskette. 6991 6992 The ss include files are now installed correctly so they can 6993 actually be used by another package. 6994 6995 If the profiling libraries are built, they are now installed on a 6996 "make install-libs". 6997 6998 6999E2fsprogs 1.03 (March 27, 1996) 7000=============================== 7001 7002Change the m68k bit numbering for bitmasks to match the bit numbering 7003used by all other ext2 implementations. (This change was requested by 7004the m68k kernel development team.) 7005 7006Support (in-development) filesystem format revision which supports 7007(among other things) dynamically sized inodes. 7008 7009Fixed a bug in the ext2 library so that an intelligent error is 7010returned if mke2fs is run with a ridiculously small number of blocks 7011for a partition. 7012 7013Fixed a bug in the ext2 library which required that the device be 7014openable in read/write mode in order to determine its size. This 7015caused e2fsck -n to require read/write access when it was not 7016previously necessary. 7017 7018Fixed a bug in e2fsck which casued it to occasionally fail the test 7019suite depending on which version of the floating point library it was 7020using. 7021 7022Fixed a bug in e2fsck so that it now halts with a fatal error when 7023certain superblock consistency checks fail. Previously it continued 7024running e2fsck, with some potential confusing/damaging consequences. 7025 7026Added new flag to fsck which allows the root to be checked in parallel 7027with other filesytems. This is not the safest thing in the world to 7028do, but some system administrators really wanted it. 7029 7030Fixed -Wall flames in lib/ss. 7031 7032 7033E2fsprogs 1.02 (January 16, 1996) 7034================================= 7035 7036Fix to allow e2fsprogs to be compiled on recent 1.3 (pl45+) kernels. 7037 7038Change e2fsck to print statistics of how many non-contiguous files are 7039on the system. Note that a file which is larger than 8k blocks, it is 7040guaranteed to be non-contiguous. 7041 7042In mke2fs, print a warning message if a user tries to format a whole 7043disk (/dev/hda versus /dev/hda1). If a user really wants to format a 7044whole disk, the -F (force) option forces mke2fs to format a whole disk 7045as a filesytem. 7046 7047Fix a bug in fsck where in some cases it might start checking 7048partitions in the next pass before it finishes checking partitions in 7049the current pass. This still won't cause two partitions on the same 7050disk will be checked, so it's rarely a problem in real life. 7051 7052Patch lsattr so that it won't hang when checking a named pipe. 7053 7054Minor compilation fixes: 7055 * Fix the order of libraries that were linked in debugfs. 7056 * Allow the sources to be compiled with -ansi turned on. 7057