RELEASE-NOTES revision 73c427d2a3f28c34280e51db76e580d9730c0cbf
1E2fsprogs 1.42.7 (January 1, 2013) 2================================== 3 4Add some new options to filefrag from Lustre's patches to e2fsprogs: 5 * add -k option to print extents in kB-sized units (like df -k) 6 * add -b {blocksize} to print extents in blocksize units 7 * add -e option to print extent format, even when FIBMAP is used 8 * add -X option to print extents in hexadecimal format 9 10Fix resize2fs so that it can handle off-line resizes of file systems 11with the flex_bg feature but without a resize_inode (or if we run out 12of reserved gdt blocks). This also fixes a problem where if the user 13creates a filesystem with a restricted number of reserved gdt blocks, 14an off-line resize which grows the file system could potentially 15result in file system corruption. 16 17Fix a bug in mke2fs where parsing "-E resize=NNN" will result in a 18mke2fs crash due to a divide-by-zero if the 64bit file system feature 19is enabled. 20 21Add to debugfs the functionality to corrupt a specific file system 22block via the "zap_block" command. 23 24Add to debugfs the functionality to print out a hex dump of a block in 25the file system via the "block_dump" command. 26 27Add to debugfs the functionality to manipulate the extent tree 28directly via the "extent_open" command. 29 30Fix e2fsck so it can detect and fix inconsistencies in the interior 31nodes of an inode's extent tree. 32 33Fix mke2fs's handling of the mmp_update_interval option. (Addresses 34Lustre Bug: LU-1888) 35 36If e2fsprogs is compiled without --enable-quota, make sure that 37tune2fs can not turn on the feature for new-style quota support, since 38afterwards, none of the e2fsprogs tools will be willing to touch that 39file system. 40 41Optimize e2fsck's so that it uses much less CPU for large file 42systems. This can result in significant speedups, especially on 43CPU-constrained systems. This was primarily done by optimizing 44libext2fs's bitmap functions. (Addresses Google Bug: #7534813) 45 46Fix debugfs's htree command so that all its messages are sent through 47the pager. 48 49Fixed debugfs's "dump_file" and "cat" functions work correctly 50on file systems whose block size is greater than 8k. 51 52Fix e2freefrag so it works on 64-bit file systems, and so it uses much 53less memory. (Addresses Google Bug: 7269948) 54 55 56Programmer's Notes 57------------------ 58 59Fix gcc -Wall nits. 60 61Fix a spelling typo in the libext2fs texinfo documentation. 62 63Change the output from "make check" so that tools such as emacs's "M-x 64compile" does not mistake the output as containing a compiler error. 65 66Export two new functions from libext2fs: ext2fs_extent_node_split() 67and ext2fs_extent_goto2(). 68 69The ext2fs_extents_fix_parents() was fixed so it does not modify the 70location of the extent handle if its current location is at an 71interior node, rather than a leaf node of an inode's extent tree. 72 73Add a regression test to assure that e2fsck can correctl fix an 74inconsistent interior node in an inode's extent tree. 75 76Fixed the help text in the configure script for --enable-quota. 77 78The m68k-specific bitops code has been removed since they were 79incorrectly treating bit numbers with the high bit set as signed 80integers. Furthermore, modern compilers do a good enough job 81optimizing the generic code there is no point in having the m68k 82specific asm statements. 83 84Fixed how we link the test programs so they always use the static 85libraries, so that we test using the libraries which we have just 86built. 87 88 89E2fsprogs 1.42.6 (September 21, 2012) 90===================================== 91 92When mke2fs creates file systems with lazy itable initialization, the 93progress updates for writing the inode table happens so quickly that 94on a serial console, the time to write the progress updates can be the 95bottleneck. So mke2fs will now only update its progress indicators 96once a second. 97 98Resize2fs will skip initializing the inode tables if the kernel 99supports lazy_itable_init, which speeds up growing off-line growth of 100uninit_bg file systems significantly. Resize2fs will now also 101correctly set the itable_unused field in the block group descriptor to 102speed up the first e2fsck after the file system is grown. 103 104Resize2fs has been fixed so that on-line resizing of meta_bg file 105systems work correctly. This is needed to grow file system which are 106larger than 16T. 107 108Resize2fs will now correctly handle resizing file systems to 16TB on 10932-bit file systems when "16TB" is specified on the command line. 110 111Fix mke2fs so that it will be careful to set the reserved blocks ratio 112larger than 50%; this can happen when creating small file systems and 113when the last block group is dropped because there are not enough 114blocks to support the metadata blocks in the last block group. 115 116Fixed spelling mistake in debugfs's help message. 117 118Fixed a potential seg fault in e2fsck when there is an I/O error while 119reading the superblock. 120 121Fixed various Debian Packaging Issues (Addresses Debian Bug #677497) 122 123Updated/fixed various man pages. (Addresses Sourceforge Bug: 124#3559210) 125 126Programmer's Notes 127------------------ 128 129The configure option --enable-relative-symlink was broken so that it 130needed to be --enable-symlink-relative-symlinks. We will support both 131for at least two years, but then the worng configure option will be 132removed. 133 134Fixed a regression introduced in 1.42.5 so the link order for 135e2fsprogs' libraries will be correct for both static and shared 136linking. (Addresses Sourceforge Bug: #3554345) 137 138Add support for e2fsprogs to be compiled using clang/LLVM. 139 140Fix portability problems on non-Linux systems: avoid compile failures 141on systems that don't have malloc.h 142 143Fix f_mmp regression test suite so that debugfs gets killed if the 144test is interrupted. 145 146 147 148E2fsprogs 1.42.5 (July 29, 2012) 149================================ 150 151Fixed a bug with mke2fs where if there is only 8 inodes per block 152group, the calculation of the number of uninitialized inodes in the 153first block group would go negative. This resulted in "mke2fs -N 256 154-t ext4 /tmp/foo.img 256m" trying to write so many blocks that /tmp 155would run out of space. (Addresses Sourceforge Bug: #3528892) 156 157Fixed a bug in how e2fsck would uniquify directory entry names. 158(AddressesSourceforge Bug: #3540545) 159 160Previously, e2fsck would only allow a mounted file system to be 161checked if it was the root file system and it was mounted read-only. 162Now it will allow any file system mounted read-only to be checked if 163the -f option is specified. This makes it easier to test how e2fsck 164handles checking file systems which are mounted without having to test 165on the root file system. 166 167Fixed a problem if e2fsck where if the root file system is mounted 168read-only, e2fsck would not clear an error indication in the journal 169superblock. Combined with a kernel bug, this would cause the e2fsck 170to check the file system after every single boot. 171 172The e4defrag program can now handle device symlinks, such as 173/dev/mapper/testvg-testlv, instead of insisting on a less 174human-friendly name such as /dev/dm-2. (Addresses Red Hat Bugzilla: 175#707209) 176 177Fixed filefrag so it will not crash with a segfault on files from a 178virtual file system such as /proc. (e.g., "filefrag 179/proc/partitions") 180 181Fixed filefrag so that it correctly reports the number of extents. 182(Addresses Red Hat Bugzilla: #840848) 183 184Fixed a file descriptor leak in logsave which could cause it to hang. 185(Addresses Debian Bug: #682592) 186 187Fixed e2fsck so that the file system is marked as containing an error 188if the user chooses not to fix the quota usage information. 189 190Fixed tune2fs so that it correctly removes the quota feature when the 191last quota inode is removed. 192 193Fix tune2fs so that after removing a quota inode, the block bitmap is 194updated; otherwise, e2fsck would complain after running 'tune2fs -O 195^quota <dev>'. 196 197Fix tune2fs so that when converting a file system from using legacy 198quota files to the new quota file system feature with hidden quota 199files, the accounting for these files is handled correctly so that 200e2fsck doesn't complain. 201 202Improved e2fsck's verbose reporting statistics, and allow the more 203verbose reporting to be enabled via /etc/e2fsck.conf. 204 205Fixed various Debian Packaging Issues (Addresses Debian Bug #678395) 206 207Updated/fixed various man pages. (Addresses Debian Bugs: #680114) 208 209 210Programmer's Notes 211------------------ 212 213Fixed portability problems on other operating systems (e.g., Hurd and 214FreeBsd) caused by the attempted inclusion of <sys/quota.h>. 215 216Make sure that shared libraries link with the shared libraries built 217in the build tree, instead of the system provided libraries. 218Previously, libraries and executables were linked with the system 219libraries if present, and possibly using static archives instead of 220shared libraries. This was also problematic since if libext2fs.so is 221linked with a static libcom_err.a from system, the build system would 222attempt to link without -lpthread. (Addresses Sourceforge Bug: 223#3542572) 224 225 226E2fsprogs 1.42.4 (June 12, 2012) 227================================ 228 229Fixed more 64-bit block number bugs (which could end up corrupting 230file systems!) in e2fsck, debugfs, and libext2fs. 231 232Fixed e2fsck's handling of the journal's s_errno field. E2fsck was 233not properly propagating the journal's s_errno field to the superblock 234field; it was not checking this field if the journal had already been 235replayed, and if the journal *was* being replayed, the "error bit" 236wasn't getting flushed out to disk. 237 238Fixed a false positive complaint by e2fsck if all of the extents in 239the last extent tree block are uninitalized and located after the end 240of the file as defined by i_size. 241 242The dumpe2fs will now display the journal's s_errno field if it is 243non-zero, and it will also display the journal's 64-bit feature flag 244if present. 245 246Fix e2fsck so that it always opens the device file in exclusive mode 247when it might need to modify the file system, and never if the -n 248option is specified. (Previously, there were a few corner cases where 249it might get things wrong in either direction.) 250 251E2fsck now correctly truncates or deallocates extent-mapped inodes on 252the orphan list. The root cause was a bug in libext2fs's block 253iterator which could end up skipping an extent when the last block in 254an extent is removed, causing the current extent to be removed from 255the extent tree. 256 257E2fsck now correctly sets the global free block and inode counts when 258truncating or removing inodes on the orphan list in preen mode. 259Previously, it would leave these values would be set incorrectly, 260which is largely a cosmetic issue since the kernel no longer pays 261attention to those fields, but it can cause spurious complaints in 262subsequent e2fsck runs. 263 264Fix i_blocks accounting when the libext2fs library needs to add or 265remove an extent tree block on bigalloc file systems. 266 267The lsattr and chattr programs now support the No_COW flag for the 268benefit of btrfs. 269 270Debugfs now interprets date strings of the form @ddd as ddd seconds 271after the beginning of the epoch. This is handy when setting an inode 272number into the d_time field when debugging orphan list handling. 273 274Fix a precedence bug with built-in quota support which might result in 275e2fsck paying attention to the quota inode field even if the built-in 276quota feature flag is not set. Fortunately, in practice that 277superblock field should be zero for non-built-in quota file systems, 278so it's unlikely this bug would have caused problems. 279 280Updated/fixed various man pages. (Addresses Debian Bugs: #674453, 281#674694) 282 283Programmer's Notes 284------------------ 285 286The regression test suite can now run the integration tests in the 287tests directory in parallel, via "make -jN check". 288 289Add new test, f_zero_extent_length which tests e2fsck's handling of 290the case where all of the extents in the last extent tree block are 291uninitialized extents after i_size. 292 293Add a new test, f_jnl_errno, which checks handling of an error 294indication set in the journal superblock. 295 296Fix the test f_jnl_64bit so that it properly checks e2fsck's handling 297of a 64-bit journal. 298 299Add two tests, f_orphan_indirect_inode and f_orphan_extent_inode which 300tests e2fsck's handling of orphan inodes in preen mode, and truncation 301of extent inodes on the orphan list. 302 303Fixed more OS X portability issues. 304 305 306E2fsprogs 1.42.3 (May 14, 2012) 307=============================== 308 309Fix a bug in the Unix I/O manager which could cause corruption of file 310systems with more than 16TB when e2fsprogs is compiled in 32-bit mode 311(i.e., when unsigned long is 32-bits). Also fix a bug which caused 312dumpe2fs to incorrectly display block numbers > 32-bits. 313 314Improve the support for integrated quota files (where quota is a first 315class supported feature using hidden files in the ext4 file system). 316Previously the quota file was getting rewritten even when it was not 317necessary, and e2fsck would erroneously try to hide quota files which 318were already hidden. 319 320Quiet complaints in e2fsck when the total free blocks or inodes are 321incorrect in the superblock after an system crash, since we don't 322update nor depend on the superblock summaries at each commit boundary. 323 324Fixed a regression introduced in 1.42.2 which would cause applications 325outside of e2fsprogs which did not pass the EXT2_FLAG_64BIT (and so 326would were still using the legacy 32-bit bitmaps) to crash. This was 327due to missing 32-bit compat code in side the function 328ext2fs_find_first_zero_generic_bmap(). (Addresses Red Hat Bugzilla: 329#808421) 330 331Fix a bug which would cause mke2fs to fail creating the journal if 332/etc/mtab and /proc/mounts are missing. (Addresses Sourceforge Bug: 333#3509398) 334 335Updated/fixed various man pages. 336 337Update Czech, Dutch, French, German, Polish, Sweedish, and Vietnamese 338translations 339 340Fixed various Debian Packaging issues. 341 342 343Programmer's Notes 344------------------ 345 346Change the nonsensically wrong types in the function signature of the 347inline function ext2fs_find_first_zero_block_bitmap2(). This was 348caused by a cut and paste error; fortunately no code in e2fsprogs used 349this inline function, and there are any users of this functions 350outside of e2fsprogs. 351 352Add support for systems which have valloc(), but which do have 353posix_memalign() nor memalign() (such as MacOS 10.5). 354 355Refactor and clean up the allocation of aligned buffers for Direct I/O 356support. Previously some allocations were requesting a greater 357alignment factor that what was strictly necessary. Also optimize 358reading and writing bitmaps using Direct I/O when the size of the 359bitmap did not fully cover the file system blocksize. 360 361Reserve the codepoints for the INCOMPAT features LARGEDATA and 362INLINEDATA. 363 364Improved the regression test suite by adding some new integration 365tests (f_jnl_32bit, f_jnl_64bit) which detect breakage of the on-disk 366jbd2 format, as well as f_eofblocks which tests the new handling of 367uninitialized and initialized blocks beyond i_size. Also add a new 368unit test which verifies 32-bit bitmaps support and the new 369find_first_zero primitives. 370 371Add a few dependencies to fix parallel (make -j) builds. 372 373Removed bash'isms which were breaking the regression test suite on 374systems where /bin/sh is not bash. 375 376The config.guess and config.sub have been updated to the 2012-02-10 377version. 378 379Fix a portability problem caused by assuming the present of mntent.h 380means that setmntent() exists. Instead, explicitly test for this in 381the configure script. 382 383If the sys/signal.h header file does not exist, don't try to include 384it, since it's not available on all systems. 385 386Add support for systems that do not support getpwuid_r() 387 388The configure script now supports a new option, 389--enable-relative-symlinks, which will install relative symlinks for 390the ELF shared library files. (Addresses Sourceforge Bug: #3520767). 391 392When building BSD shared libraries make sure the LDFLAGS variable is 393passed to the linker. Fixing this allows, for example, e2fsprogs to 394be built in 32-bit mode on Mac OS X Lion (Addresses Sourceforge Bug: 395#3517272) 396 397Fix gcc -Wall nitpicks 398 399 400E2fsprogs 1.42.2 (March 27, 2012) 401================================= 402 403The resize2fs program uses much less CPU and is much faster for very 404large file systems. (Addresses Debian Bug: #663237) 405 406The seti and freei commands in debugfs can now take an optional length 407argument to set and clear a contiguous range of inodes. 408 409E2fsck will now make explicit checks for the EOFBLOCKS_FL, since we 410plan to remove support for it from the kernel file system driver. It 411really wasn't very useful and was causing more problems than it 412solves. Since e2fsck will complain if inodes that should have 413EOFBLOCKS_FL do not have the flag set, we are going to remove this 414check from e2fsprogs first, and then only remove the flag from the 415kernel much later. 416 417The mke2fs program can now use direct I/O via "mke2fs -D". This will 418slow down the mke2fs, but it makes it more polite on a loaded server 419by limiting the amount of memory that gets dirtied by mke2fs when it 420is using buffered I/O. 421 422E2fsck was needlessly closing and re-opening the file system as a side 423effect of adding Multiple Mount Protection (MMP). This isn't 424necessary for non-MMP file systems, so drop it. 425 426Print errors returned by ext2fs_open2() and ext2fs_check_desc() so we 427can more easily diagnose memory allocation failures caused by 428insufficient memory. E2fsck will now abort if there are memory 429allocation failures when the file system is initially opened and 430during the block group descriptor checks. (Addresses Google Bug: 431#6208183) 432 433If there are incorrect block group checks, e2fsck will now report the 434incorrect and corrected checksum values. 435 436The e2fsck progam can now write log files containing the details of 437the problems that were found and fixed directly, via configuration 438parameters in /etc/e2fsck.conf. 439 440Added the ability to limit the number of messages reported by e2fsck 441for a given problem type. This avoids a potential bottleneck if there 442is a serial console which can cause a boot sequence to take a long 443time if e2fsck needs to report many, many file system errors. 444 445The dumpe2fs, debugfs, and tune2fs now use rbtree bitmaps, which cause 446them to use much less memory for large file systems. 447 448The dumpe2fs program will now print the expected block group checksum 449if it is incorrect. This helps to diagnose problems caused by 450incorrect block group checksums. 451 452E2fsck now checks for extents with a zero length, since the kernel 453will oops if it comes accross such a corrupted data structure. (See 454https://bugzilla.kernel.org/show_bug.cgi?id=42859) 455 456E2fsck has a number of bugs relating to discard that have been fixed. 457(1) Fixed a bug which could cause e2fsck to discard portions of the 458inode table which were actually in use. (2) E2fsck will now avoid 459using discard if the block device doesn't zero data on discard, since 460otherwise this could cause problems if the file system gets corrupted 461in the future. (3) E2fsck will now avoid using discard when it is run 462in read-only mode. (4) Fixed a bug which caused e2fsck to not issue 463discards in the last block group. 464 465E2fsck's CPU utilization in pass 5 has been optimized, which will 466speed up e2fsck slightly. 467 468E2image will now skip copying uninitialized bitmap and inode table 469blocks. 470 471Fixed mke2fs -S so it does not corrupt the first block group's 472information. 473 474E2fsck will now check the new sysfs interface to determine if we are 475using the battery or AC mains. (Addresses SourceForge Bug: #3439277) 476 477Updated/fixed various man pages. (Addresses Debian Bug: #665427) 478 479Fixed various Debian Packaging issues. (Addresses Debian Bug: #665885) 480 481Programmer's Notes 482------------------ 483 484Fixed various portability issues for non-Linux systems, particularly 485MacOS X, as well as Linux systems running with the just-released glibc 4862.15. 487 488Fix file descriptor leak in ext2fs_close() if the file system with 489uninit_bg is opened read/only with a backup superblock. (Addresses 490SourceForge Bug: #3444351) 491 492Fixed an invalid return in a non-void function in the quota code. 493(Addresses SourceForge Bug: #3468423) 494 495Fixed the debian rules file so that the calls to dpkg-buildflags works 496when the shell is dash. 497 498The debian package build now uses V=1 so that there is more 499information about potential build failures in debian buildd logs. 500 501If the uninit flags get cleared by functions such as 502ext2fs_new_inode() or ext2fs_new_block2(), we now make sure the 503superblock is marked dirty and the block group descriptor checksum is 504updated if necessary. 505 506The debian rules file will now try to load debian/rules.custom of it 507exists. This flie can skip various builds for speed reasons if there 508is no need for the e2fsck-static or udeb packages. Available 509customizations in the rules file includes SKIP_STATIC=yes, 510SKIP_BF=yes, and SKIP_DIETLIBC=yes. 511 512In addition, if the file misc/mke2fs.conf.custom.in exists in the 513source tree, it will be used instead of the standard misc.conf file in 514the upstream sources. This makes it easier for Debian-derived systems 515to distribute a custom mke2fs.conf file without having to worry about 516merge issues if future versions of e2fsprogs makes changes in the 517upstream default version of mke2fs.conf. 518 519 520E2fsprogs 1.42.1 (February 17, 2012) 521=================================== 522 523The mke2fs and e2fsck now use significantly less memory when creating 524or checking very large file systems. This was enabled by adding 525extent-based bitmaps which are stored using a red-block tree, since 526block and inode allocations tend to be contiguous. 527 528The command mke2fs -S is used as a last ditch recovery command to 529write new superblock and block group descriptors, but _not_ to destroy 530the inode table in hopes of recovering from a badly corrupted file 531system. So if the uninit_bg feature is enabled, mke2fs -S will now 532set the unused inodes count field to zero. Otherwise, e2fsck -fy 533after using mke2fs -S would leave the file system completely empty. 534 535Since mke2fs recognizes mke3fs in argv[0] to mean "mkfs.ext3", also 536honor "mke4fs" to work the same as "mke2fs.ext4", since RHEL5 has 537installed an mke2fs binary using that name. 538 539The usage and help messages for the -G, -t and -T options in mke2fs 540have been fixed. 541 542If e2fsck needs to use the backup group descriptors, the 543ext2fs_open2() function clears the UNINIT bits to ensure all of the 544inodes in the file systems get scanned. However, the code which reset 545the UNINIT flags did not also recalculate the checksum, which produced 546many spurious (and scary) e2fsck messages. This has been fixed by 547resetting cheksums when the UNINIT bits are cleared. 548 549Relax a check in e2fsck which required that the block bitmap to be 550initialized when the inode bitmap is in use. This will allow us to 551eventually eliminate code from the kernel which forcibly initialized 552the block bitmap when the inode bitmap is first used, requiring an 553extra journal credit and disk write. (Addresses Google Bug: #5944440) 554 555Make sure rdebugfs (which may be installed setuid or setgid disk) does 556not honor environment variables if euid != uid or egid != gid. 557 558Debugfs's ncheck command has been optimized and now is much more 559robust with faced with corrupted file systems. The ncheck command 560also now has a -c option which will verify the file type information 561in the directory entry to see if matches the inode's mode bits. This 562is extremely useful when trying to use debugsfs to determine which 563parts of the file system metadata can be trusted. 564 565E2image will try to use ftruncate64() to set the i_size for raw 566images, instead of writing a single null byte. This avoid allocating 567an extra block to the raw image, for those file systems and/or 568operating systems that support this. (Linux does.) In addition, fix 569a logic bug that caused the file to not be properly extended if the 570size of the last hole was exactly an multiple of a megabyte. 571 572Fixed a bug in resize2fs where for 1k and 2k file systems, where 573s_first_data_block is non-zero, this wasn't taken into account when 574calculate the minimum file system size for use with the -M option. 575 576Fixed the badblocks program to honor the -s flag when in read-only -t 577mode. (Addresses Debian Bug #646629) 578 579Update Czech, Dutch, French, Polish, and Sweedish translation from the 580Translation Project. 581 582Fixed various Debian Packaging issues so that dpkg-buildflags is used 583if present, which allows e2fsprogs to be built with security hardening 584flags. (Addresses Debian Bugs: #654457) 585 586Programmer's Notes 587------------------ 588 589Fix a bug in ext2fs_clear_generic_bmap() when used for 32-bit bitmaps. 590This was only an issue for programs compiled against e2fsprogs 1.41 591that manipulate bitmaps directly. (Addresses Sourceforge Bugs: 592#3451486) 593 594The libext2fs library now uses sysconf() to fetch the page size, instead 595of the deprecated getpagesize(). 596 597The ext2fs_get_pathname() function will return a partial path if an a 598directory in the path is not a directory, displaying it as an inode 599number in angle brackets instead of giving up and displaying an error. 600This is much more helpful when a user is trying to debug a corrupted 601file system. 602 603Codepoints for the RO_COMPAT_REPLICA feature has been reserved. 604 605Added a new library function, ext2fs_file_get_inode_num(), for use by 606fuse2fs. 607 608Fixed a bug in ext2fs_file_set_size2() so that when it is truncating a 609file, it actually works. 610 611The block iterator now properly honors the BLOCK_ABORT flag for 612extent-based flags. Previously, it didn't, which generally made code 613be less efficient, but it could cause bugs in ext2fs_link(), for 614example, by causing it to insert multiple directory entries. 615 616Fixed an (harmless other than causing a compiler warning) use of an 617uninitialized variable in e2fsck's MMP code. 618 619 620E2fsprogs 1.42 (November 29, 2011) 621================================== 622 623This release of e2fsprogs has support for file systems > 16TB. Online 624resize requires kernel support which will hopefully be in Linux 625version 3.2. Offline support is not yet available for > 16TB file 626systems, but will be coming. 627 628This release of e2fsprogs has support for clustered allocation. This 629reduces the number of block (now cluster) bitmaps by allocating and 630deallocating space in contiguous power-of-2 collections of blocks, 631which are called clustered. This is a file system level feature, 632called 'bigalloc', which must be enabled when the file system is 633initially formatted. It is not backwards compatible with older 634kernels. 635 636Added support for the Multi-Mount Protection (MMP) feature. 637 638E2fsck more efficiently uses scratch files for really big file 639systems. (This is a feature that has to be turned on explicitly; see 640[scratch_files] in the e2fsck.conf man page.) 641 642Fix a bug in e2fsck where if the free blocks and inodes counts are 643incorrect, e2fsck would fix them without printing an error message. 644This would cause a "*** FILE SYSTEM WAS MODIFIED ***" message without 645any explanation of what was fixed. 646 647E2fsck will no longer attempt to clone an extended attribute block in 648pass1b handling if the file system does not support extended 649attributes. 650 651E2fsck will be more careful accidentally asking the user to continue 652if the file system is mounted, so that an escape sequence won't cause 653a false positive. (Addresses Debian Bug: #619859) 654 655E2fsck now uses less cpu time in pass 5 when large portions of the 656bitmaps are uninitialized. 657 658E2fsck will no longer segault when a corrupted file system has a bad 659extent, and removing it leads to a block needing to be deallocated. 660(Addresses SourceForge Bug: #2971800) 661 662E2fsck will catch termination signals (segfaults, bus errors, sigfpe) 663and print debugging information to make it easier to find potential 664problems. 665 666E2fsck will check to see if the bad block inode looks insane, and will 667skip trying to use if it certain fields which should be zero look 668non-zero. This is to avoid a corrupted bad block inode causing e2fsck 669to get confused and do more harm than good. 670 671If e2fsck modifies certain superblock fields which the kernel doesn't 672look at, it will now mark the superblock as dirty without marking the 673file system as changed. This avoids signalling the init scripts that 674a reboot is necessary, since the kernel isn't going to look at those 675fields, so it won't care if they have been changed. 676 677Fixed a bug in the libext2fs library (in the binary search routine of 678the icount abstraction) that could (very, very rarely) cause e2fsck to 679die in the middle of pass 1 or pass 2 processing. 680 681E2fsck will not try to do a discard operation if the -n option was 682specified on the command line. 683 684E2fsck now supports an extended "discard" option which will cause 685e2fsck to attempt discard all unused blocks after a full, successful 686file system check. 687 688E2fsck will check for the bad block inode to make sure it looks sane 689before trusting it, to avoid causing more harm than good to the file 690system. 691 692E2fsck now returns additional status bits in its exit code if it 693aborts early in the e2fsck run. 694 695E2fsck now correctly calculates the maximum file size in the case of 696the huge_file file system feature enabled without extents. 697 698The mke2fs and e2fsck programs now tries to use the punch hole command 699as a "discard" when operating on normal files. 700 701The e2image program now supports the qcow2 format, which is a more 702efficient way of capturing file system dumps. 703 704Mke2fs now supports the [devices] stanza in mke2fs.conf which allows 705per-device defaults to be specified in the configuration file. 706 707Mke2fs now supports the reserved_ratio relation in the [defaults] and 708[fs_types] section in mke2fs.conf. 709 710Mke2fs now creates extent-mapped directories for the root and 711lost+found directories. 712 713Mke2fs will skip zero'ing the journal if the extended option 714"lazy_journal_init" is specified. This can save a lot of time, but it 715does add a small amount of risk if the system crashes before the 716journal is overwritten entirely once. It is epsecially useful for 717testing. 718 719Mke2fs will now create file systems that enable user namespace 720extended attributes and with time- and mount count-based file 721system checks disabled. 722 723Mke2fs will not set a stride or strip size of one block based on block 724device attributes obtained from sysfs. 725 726Mke2fs now displays a progress report during the discard process. 727 728Mke2fs now handles extreme file system parameters correctly which 729previously caused the inodes per group to drop below 8, leading to a 730segfault. (The inodes per group must be a multiple of 8, but the code 731didn't correctly deal with an inodes per group count less than 8.) 732 733Mke2fs and tune2fs previously would give an error if the user tried 734setting the stride and stripe-width parameters to zero; but this is 735necessary to disable the stride and stripe-width settings. So allow 736setting these superblock fields to zero. (Addresses Google Bug: 737#4988557) 738 739Mke2fs now gives a warning if the auto-detected block size exceeds the 740system's page size. 741 742If the enable_periodic_fsck option is false in /etc/mke2fs.conf (which 743is the default), mke2fs will now set the s_max_mnt_count superblock 744field to -1, instead of 0. Kernels older then 3.0 will print a 745spurious message on each mount then they see a s_max_mnt_count set to 7460, which will annoy users. (Addresses Debian Bug: #632637) 747 748The default mke2fs.conf now has entries for "big" and "huge", which 749are needed for very big file systems. 750 751The resize2fs program now has support for a new online resize ioctl 752that can support file systems > 16TB, once it arrives in v3.x kernels. 753 754Fixed bug which caused resize2fs to fail when shrinking an empty file 755system down to its minimal size. (Addresses Sourceforge Bug #3404051) 756 757Fixed tune2fs's mount options parsing. (Addresses Debian Bug: #641667) 758 759Allow tune2fs to remove external journals if the device can not be 760found. 761 762Debugfs's icheck will now correctly find inodes which use the 763searched-for block as an extended attribute block. 764 765Debugfs now has a new "punch" command which remove blocks from the 766middle of an inode. 767 768Debugfs now has a new "e2freefrag" command which analyzes the free 769space fragmentation of the file system, using the same code as the 770e2freefrag program. 771 772Debugfs now has a "filefrag" command which displays information about 773a file's fragmentation. 774 775Add support to build a metadata-only, read-only, stripped-down version 776of debugfs called rdebugfs. 777 778Fixed a potential stack overrun bug in debugfs. 779 780The badblocks program now correctly recovers from I/O errors when 781direct I/O is being used. The badblocks command now also supports a 782-B option which forces the use of buffered I/O, and the -v option will 783provide a more detailed breakdown of read, write, and failed 784comparison errors. 785 786Added e4defrag tool which uses the EXT4_IOC_MOVE_EXT ioctl. 787 788Added support for journals larger than 2GB. 789 790Support using both hard links and symlinks when installing e2fsprogs. 791 792Add overflow checking to tune2fs -i's fsck interval, which must fit in 793a 32-bit field. 794 795The debugfs command now has a new 'blocks' command which prints out 796data blocks of a particular inode in a format which is useful for 797scripting. 798 799Filefrag will report 0 extents correctly in verbose mode. (Addresses 800RedHat Bugzilla: #653234) 801 802Filefrag has been fixed so its -v report prints the correct expected 803block number (previously there had been an off-by-one error). In 804addition, it will now display the number of contiguous extents when -v 805is not specified. This makes it consistent with the number of extents 806printed when the -v option was specified. In addition, the number of 807contiguous extents is far more interesting/useful than the number of 808physical extents for very large files. (Addresses Debian Bug: 809#631498, #644792) 810 811Logsave's usage message has been fixed. (Addresses Debian Bug: 812#619788) 813 814Avoid an infinite loop in ext2fs_find_block_device() if there are 815symlink loops in /dev caused by a buggy udev. 816 817Added a useful "fallocate" program to the contrib directory. 818 819Fixed an ABI compatibility problem in libext2fs which broke the dump 820program. Also added back some macros which dump needed so it could 821compile against the latest version of ext2_fs.h (Addresses Debian Bug: 822#636418) 823 824Fixed parsing of MNTOPT_ options for tune2fs and debugfs (Addresses 825Debian Bug: #641667) 826 827Added internationalization support for libcom_err error table strings. 828 829Fixed various spelling mistakes found in various output strings found 830by I18N translators. 831 832Update translations: French, Chinese, Germany, Indonesian, Swedish, 833Vietnamese, Polish, Dutch, Czech. (Addresses Debian Bugs: #520985, 834#620659) 835 836Fixed various Debian Packaging issues. (Addresses Debian Bugs: 837#614662, #632169, #641838, #627535, #629355) 838 839Updated/clarified man pages. (Addresses Debian Bugs: #639411, 840#642193, #634883) 841 842Programmer's Notes 843------------------ 844 845Initial support for quota as an integrated feature, where the quota 846files are hidden system files that are automatically maintained by 847e2fsck is present, although disabled by default. It must be enabled 848by using the configure option --enable-quota. 849 850Reserved the on-disk fields for the metadata checksum and snapshot 851features, which are currently in development. 852 853The ext2fs library now has the new functions ext2fs_punch(), 854ext2fs_get_memzero() and ext2fs_file_get_inode(). 855 856The ext2fs library now has support for calculating the crc32c checksum 857(via the new functions ext2fs_crc32c_be and ext2fs_crc32c_le). 858 859The I/O manager now supports the discard operation. 860 861Reserved file system code points for new 1st class quota feature. 862 863Shortened the compilation lines by moving the autoconf defines to 864a config.h header file. 865 866Fixed a potential free of an unitialized pointer in 867ext2fs_update_bb_inode(). 868 869Fixed miscellaneous gcc -Wall and coverity warnings. 870 871Fixed portability issues for Mac OS X, Hurd, and FreeBSD. 872 873Fixed a build failure when OMIT_COM_ERR is defined. 874 875Improved error checking and fixed memory leaks caused by error return 876paths. 877 878Add ext2fs_flush2() and ext2fs_close2() which takes a flag to allow 879the fsync() to be skipped. 880 881Added a test for extent-mapped journals by mke2fs and tune2fs. 882 883Added a test for creating a large (over 4GB) journal using mke2fs. 884 885Added a test to make sure the inode size remains constant. 886 887The regression test script now prints the list of which tests failed. 888 889The regression test system now uses /tmp for its scratch files, which 890is often a tmpfs mounted file system and hence much faster. 891 892The i_e2image test was fixed so it works with valgrind. 893 894Tests that rely on debugfs are now skipped if debugfs wasn't built. 895 896Fixed the dependencies for "make check" so all required dependencies 897are built before running the gression tests. 898 899A link to com_err.h is installed in $(includedir) during a "make install". 900 901The po/*.gmo files are automatically rebuilt if they are missing or 902out of date. This helps out Debian packaging. 903 904Allow ext2fs_get_memalign() to compile on systems that don't have 905posix_memalign(). 906 907Fixed a namespace leak in libext2fs (tdb_null). 908 909 910E2fsprogs 1.41.14 (December 22, 2010) 911===================================== 912 913Fix spurious complaint in mke2fs where it would complain if the file 914system type "default" is not defined in mke2fs.conf. 915 916The resize2fs program will no longer clear the resize_inode feature 917when the number reserved group descriptor blocks reaches zero. This 918allows for subsequent shrinks of the file system to work cleanly for 919flex_bg file systems. 920 921The resize2fs program now handles devices which are exactly 16T; 922previously it would give an error saying that the file system was too 923big. 924 925E2fsck (and the libext2fs library) will not use the extended rec_len 926encoding for file systems whose block size is less than 64k, for 927consistency with the kernel. 928 929Programming notes 930----------------- 931 932E2fsprogs 1.41.13 would not compile on big-endian systems. This has 933been fixed. (Addresses Sourceforge Bug: #3138115) 934 935The ext2fs_block_iterator2() function passed an incorrect ref_offset 936to its callback function in the case of sparse files. (Addresses 937Sourceforge Bug: #3081087) 938 939Fix some type-punning warnings generated by newer versions of gcc. 940 941 942E2fsprogs 1.41.13 (December 13, 2010) 943===================================== 944 945E2fsck now supports the extended option "-E journal_only", which 946causes it to only do a journal replay. This is useful for scripts 947that want to first replay the journal and then check to see if it 948contains errors. 949 950E2fsck will now support UUID= and LABEL= specifiers for the -j option 951(which specifies where to find the external journal). (Addresses 952Debian Bug #559315) 953 954E2fsck now has support for the problems/<problem code>/force_no 955configuration option in e2fsck.conf, which forces a problem to not be 956fixed. 957 958Dumpe2fs will now avoid printing large negative offsets for the bitmap 959blocks and instead print a message which is a bit more helpful for 960flex_bg file systems. 961 962Mke2fs will now check the file system type (specified with the -t 963option) to make sure it is defined in the mke2fs.conf file; if it is 964not, it will print an error and abort. If the usage type (specified 965with the -T option) is not defined in mke2fs.conf, mke2fs will print a 966warning but will continue. (Addresses Debian Bug #594609) 967 968Clarified error message from resize2fs clarifying that on-line 969shrinking is not supported at all. (Addresses Debian Bug #599786) 970 971Fix an e2fsck bug that could cause a PROGRAMMING BUG error to be 972displayed. (Addresses Debian Bug #555456) 973 974E2fsck will open the external journal in exclusive mode, to prevent 975the journal from getting truncated while it is in use due to a user 976accidentally trying to run e2fsck on a snapshotted file system volume. 977(Addresses Debian Bug #587531) 978 979Fix a bug in e2fsck so it has the correct test for the EOFBLOCKS_FL 980flag. 981 982The tune2fs program can now set the uninit_bg feature without 983requiring an fsck. 984 985The tune2fs, dumpe2fs, and debugfs programs now support the new ext4 986default mount options settings which were added in 2.6.35. 987 988The e2fsck and dumpe2fs programs now support the new ext4 superblock 989fields which track where and when the first and most recent file 990system errors occurred. These fields are displayed by dumpe2fs and 991cleared by e2fsck. These new superblock fields were added in 2.6.36. 992 993Debugfs now uses a more concicse format for listing extents in its 994stat command. This format also includes the interior extent tree 995blocks, which previously was missing from stat's output for 996extent-based files. 997 998Debugfs has a new option, -D, which will request Direct I/O access of 999the file system. 1000 1001Mke2fs will skip initializing the inode table if a device supports 1002discard and the discard operation will result in zero'ed blocks. 1003 1004Badblocks will now correctly backspace over UTF-8 characters when 1005displaying its progress bar. (Addresses Gentoo Bug #309909; Addresses 1006Debian Bugs #583782 and #587834) 1007 1008E2freefrag will now display the total number of free extents. 1009 1010Resize2fs -P no longer requires a freshly checked filesystem before 1011printing the minimum resize size. 1012 1013Fixed a floating point precision error in a binary tree search routine 1014that can lead to seg fault in e2fsck and resize2fs. 1015 1016Fixed a bug in e2fsck where if both the original and backup superblock 1017are invalid in some way, e2fsck will fail going back to the original 1018superblock because it didn't close the backup superblock first, and 1019the exclusive open prevented the file system from being reopened. 1020 1021Fixed a big in e2freefrag which caused getopt parsing to fail on 1022architectures with unsigned chars. (Addresses Gentoo Bug: #299386) 1023 1024Clarified an mke2fs error message so a missed common in an -E option 1025(i.e., mke2fs -t ext4 -E stride=128 stripe-width=512 /dev/sda1") 1026results in a more understandable explanation to the user. 1027 1028Mke2fs now displays the correct valid inode ratio range when 1029complaining about an invalid inode ratio specified by the user. 1030 1031Mke2fs now understands the extended option "discard" and "nodiscard", 1032and the older option -K is deprecated. The default of whether 1033discards are enabled by default can be controled by the mke2fs.conf 1034file. 1035 1036Mke2fs's handling of logical and physical sector sizes has been 1037improved to reflect the fact that there will be some SSD's with 8k and 103816k physical sectors arriving soon. Mke2fs will no longer force block 1039size to be the physical sector size, since there will be devices where 1040the physical sector size is larger than the system's page size, and 1041hence larger than the maximal supported block size. In addition, if 1042the minimal and optimal io size are not exported by the device, and 1043the physical sector size is larger than the block size, the physical 1044sector size will be used to set the Raid I/O optimization hints in the 1045superblock. 1046 1047E2fsck will now display a better, more specific error message when the 1048user enters a typo'ed device name, instead of blathering on about 1049alternate superblocks. 1050 1051Fixed various Debian Packaging Issues 1052 1053Updated/clarified man pages (Addresses Debian Bugs: #580236, #594004, 1054#589345, #591083; Addresses Launchpad Bug: #505719) 1055 1056Update the Chinese, Chzech, Dutch, French, Germany, Indonesian, 1057Polish, Swedish, and Vietnamese translations. 1058 1059 1060Programmer's Notes 1061------------------ 1062 1063Fix a dependency definition for the static and profiled blkid 1064libraries which could cause compile failures in some configurations. 1065(Addresses Debian Bug: #604629) 1066 1067Add support for Direct I/O in the Unix I/O access layer. 1068 1069Fixed a memory leak in the Unix I/O layer when changing block sizes. 1070 1071Fixed minor memory leaks in mke2fs. 1072 1073Added a new function to the ext2fs library, ext2fs_get_memalign(). 1074 1075The tst_super_size test program will check to make sure the superblock 1076fields are correctly aligned and will print them out so they can be 1077manually checked for correctness. 1078 1079Fixed some makefile dependencies for test programs in lib/ext2fs. 1080 1081Reserved the feature flags and superblock fields needed for the Next3 1082snapshot feature. 1083 1084Reserved the feature flags for EXT4_FEATURE_INCOMPAT_DIRDATA and 1085EXT4_INCOMPAT_EA_INODE. 1086 1087 1088E2fsprogs 1.41.12 (May 17, 2010) 1089================================ 1090 1091Mke2fs now gives a correct error message if the external journal 1092device is not found. (Addresses Red Hat Bug #572935) 1093 1094Resize2fs -P will now refuse to print a minimum size if the file 1095system is not clean. Previously it would go ahead and print a minimum 1096size anyway, which might not be correct, leading to user confusion. 1097 1098E2fsck now tests for extents that begin at physical block 0 and 1099rejects them as invalid. (Addresses Google Bug: #2573806) 1100 1101Fixed a bug in e2fsck which could cause it to crash when trying to 1102remove an invalid extent and the block bitmaps hadn't yet been loaded. 1103(Addresses SourceForge Bug: #2971800) 1104 1105E2fsck now will completely skip time-based checks if the system clock 1106looks insane or the option broken_system_clock is set in 1107/etc/e2fsck.conf. 1108 1109E2fsck would previously report an i_blocks corruption for a 4T file 1110created using posix_fallocate; this bug has been fixed. 1111 1112E2fsck will now correctly mark a sparse journal as invalid and will 1113delete and recreate the journal to address the problem. 1114 1115E2fsck would previously incorrectly ask the user whether she would 1116like to abort the file system check after finding a problem --- and 1117then abort regardless of the user's answer. This is annoying, and 1118has been fixed. 1119 1120E2fsck can now continue even if it fails to recreate the resize 1121inode; previously it would just abort the file system check 1122altogether. 1123 1124E2fsck could potentially remove directory entries for inodes found in 1125the unused region of the inode table; this would generally happen on 1126ext4 file systems that do not use journaling. This bug has been fixed 1127by not clearing these directory entries once it has been established 1128that bg_unused_inodes may not be trustworthy; once pass #2 has been 1129completed, e2fsck will restart the file system check from the 1130beginning, and then it will be safe to delete any directory entries 1131pointing to inodes that appear to be deleted. (Addresses Google Bug: 1132#2642165) 1133 1134E2fsck will not try to set the block group checksums if the user 1135cancels the fsck with a control-C. It's a bad idea to set the 1136checksums if e2fsck hasn't been completed, and it often results an the 1137error message, "Inode bitmap not loaded while setting block group 1138checksum info". (Addresses Launchpad Bug: #582035) 1139 1140The mke2fs program now queries the kernel for the physical as well as 1141the logical sector size, and will not allow a blocksize below the 1142logical, and will strongly encourage a blocksize at least as big as 1143the physical blocksize. This is needed for 4k sector drives that 1144emulate 512 byte sector sizes. 1145 1146Mke2fs will now allow a flex_bg size of 1. This is unusual, and 1147rarely needed, but it is a legal value. 1148 1149E2fsck will check for cases where the EOFBLOCKS_FL is set when it is 1150not needed, and offer to clear it; this is a sign of a kernel bug, but 1151more importantly, some released kernels may crash when this situation 1152is encountered on ext4 file systems. (Addresses Google Bug: #2604224) 1153 1154E2fsck will use the EOFBLOCKS_FL flag exclusively to check whether 1155i_size is correct. (Kernels starting with 2.6.34 will set 1156EOFBLOCKS_FL.) 1157 1158The com_err library will now only output ^M (a CR character) when the 1159tty is in raw mode. 1160 1161Update the Czech, Chinese, Dutch, French, Germany, Indonesian, Polish, 1162and Vietnamese translations. 1163 1164Fixed various Debian packaging issues --- see debian/changelog for 1165details. (Addresses Debian Bugs: #571247, #563487) 1166 1167 1168Programmer's Notes 1169------------------ 1170 1171The regression test suite now uses its own mke2fs.conf file, so that 1172downstream distributions want change the mke2fs.conf file which is 1173distributed in the RPM or dpkg file, without worrying about screwing 1174up the regression test results. 1175 1176Always build namei.o so that building with configure --disable-debugfs 1177works correctly. Long-term, if we care about reduced e2fsprogs 1178builds, we need a more general solution for deciding what .o files are 1179needed for a particular build. Given that install floppies are going 1180(gone?) the way the dodo bird, we probably don't care, though. 1181(Addresses Sourceforge Bug: #2911433) 1182 1183Add configure options --enable-symlink-build and 1184--enable-symlink-install, which allow e2fsprogs be built using 1185symlinks instad of hard links, and to be installed using symlinks 1186instead of hard links, respectively. It is useful when the file 1187system where the build is taking place, or the file system where 1188e2fsprogs is installed, can't handle hard links for some reason. 1189(Addresses Sourceforge Bug: #1436294) 1190 1191Fixed compile warning in mke2fs.c. 1192 1193 1194E2fsprogs 1.41.11 (March 14, 2010) 1195================================== 1196 1197E2fsck will no longer give a fatal error and abort if the physical 1198device has been resized beyond 2**32 blocks. (Addresses Launchpad 1199Bug: #521648) 1200 1201Debugfs has a bug fixed so that "logdump -b <blk>" now properly shows 1202the allocation status of the block <blk>. (Addresses Debian Bug: 1203#564084) 1204 1205E2fsck now prints a much more emphatic and hopefully scary message 1206when a file system is detected as mounted while doing a read/write 1207check of the filesystem. Hopefully this will dissuade users from 1208thinking, "surely that message doesn't apply to *me*" :-( 1209 1210E2fsck -n will now always open the file system read-only. We now 1211disallow certain combination of options which previously were manual 1212exceptions; this is bad because it causes users to think they are 1213smarter than they really are. So "-n -c", "-n -l", "-n -L", and "-n 1214-D" are no longer supported. (Addresses Launchpad Bug: #537483) 1215 1216In e2fsprogs 1.41.10, mke2fs would ask for confirmation to proceed if 1217it detected a badly aligned partition. Unfortunately, this broke some 1218distribution installation scripts, so it now just prints the warning 1219message and proceeds. (Addresses Red Hat Bug: #569021. Addresses 1220Launchpad Bug: #530071) 1221 1222Mke2fs would take a long time to create very large journal files for 1223ext4. This was caused by a bug in ext2fs_block_iterate2(), which is 1224now fixed. 1225 1226E2fsck now understands the EOFBLOCKS_FL flag which will be used in 12272.6.34 kernels to make e2fsck not complain about blocks deliberately 1228fallocated() beyond an inode's i_size. 1229 1230E2fsprogs 1.41.10 introduced a regression (in commit b71e018) where 1231e2fsck -fD can corrupt non-indexed directories when are exists one or 1232more file names which alphabetically sort before ".". This can happen 1233with ext2 filesystems or for small directories (take less than a lock) 1234which contain filenames that begin with a space or some other 1235punctuation mark. (Addresses Debian Bug: #573923, Addresses Launchpad 1236Bug: #525114) 1237 1238 1239Programmer's Notes 1240------------------ 1241 1242Add new test, f_rehash_dir, which checks to make sure e2fsck -D works 1243correctly. 1244 1245The libcom_err function now has support for Heimdal's com_right_r 1246function(). (Addresses Sourceforge Bug: #2963865, Addresses Debian 1247Bug: #558910) 1248 1249 1250E2fsprogs 1.41.10 (February 7, 2010) 1251==================================== 1252 1253Fix resize2fs bug which causes it to access invalid memory. 1254 1255Add libss support for libreadline.so.6. 1256 1257Fix e2fsck's check for extent-mapped directory with an incorrect file 1258type. 1259 1260Add new e2fsck.conf configuration option, default/broken_system_clock 1261to support systems with broken CMOS hardware clocks. Also, since too 1262many distributions seem to have broken virtualization scripts now, 1263e2fsck will by default accept dates which are off by up to 24 hours by 1264default. (Addresses Debian Bugs: #559776, #557636) 1265 1266Fix a bug where mke2fs may not use the best placement of the inode 1267table when there is only room for a single block group in the last 1268flex_bg. 1269 1270E2fsck is now smarter when it needs to allocate blocks in the course 1271of fixing file system problems. This reduces the number of spurious 1272differences found in pass #5. 1273 1274E2fsck will no longer rehash directories which fit in a single 1275directory block. 1276 1277E2fsck now correctly handles holes in extent-mapped directories (i.e., 1278sparse directories which use extents). 1279 1280Fix big-endian problems with ext2fs_bmap() and ext2fs_bmap2(). 1281 1282Fix a bug in filefrag where on platforms which can allow file systems 1283with 8k blocks, that it doesn't core dump when it sees a file system 1284with 8k block sizes. (Thanks to Mikulas Patocka for pointing this 1285out.) 1286 1287E2fsck will correctly fix directories that are have an inaccurate 1288i_size as well as other problems in a single pass, instead of 1289requiring two e2fsck runs before the file system is fully fixed. 1290 1291Fix e2fsck so it will correctly find and detect duplicate directory 1292entries for non-indexed directories. ( Addresses Sourceforge Bug: 1293#2862551) 1294 1295Mke2fs will use BLKDISCARD to pre-discard all blocks on an SSD or 1296thinly-provisioned storage device. This can be disabled using the -K 1297option. 1298 1299Enhance libext2fs so it works around bug in Linux version 2.6.19 1300and earlier where the /proc/swaps file was missing the header on 1301the first line. 1302 1303Fix bug in Linux version 2.6.19 and earlier where the /proc/swaps file 1304was missing the header on the first line. 1305 1306Fix some big-endian bugs in e2fsck and libext2fs. 1307 1308Fix resize2fs so it works correctly on file systems with external 1309journals instead of failing early with the error "Illegal inode 1310number". 1311 1312Fix libss so that it does not seg fault when using a readline library 1313which does not supply a readline_shutdown() function. 1314 1315Updated dumpe2fs's usage message so it correctly gives the right 1316arguments summary for "-o superblock=<num>" and "-o blocksize=<num>". 1317(Addresses Launchpad Bug: #448099) 1318 1319Teach libext2fs to ignore the high 32 bits of the i_blocks field 1320when huge_file file system feature is set, but the inode does not 1321have the HUGE_FILE_FL flag set. 1322 1323Fix e2fsck's handling of 64-bit i_blocks fields. 1324 1325E2fsck will now print "Illegal indirect block", "Illegal 1326double-indirect block", etc., instead of "Illegal block #-1" or 1327"Illegal block #-2", etc. This makes it easier for users to 1328understand what has gone wrong. (Addresses SourceForge Bug: #2871782) 1329 1330Mke2fs now will obtain get device topology information from blkid and 1331use it to populate the superblock stride and stripe sizes. It will 1332also warn if the block device is misaligned 1333 1334Fix file descriptor leaks in fsck and debugfs. (Addresses Novell Bug: 1335##524526) 1336 1337Fix the libext2fs library code to round up the bitmap size to a 4-byte 1338boundary, to prevent spurious seg faults caused by the x86 1339architecture. This doesn't affect Linux systems, but was a major 1340problem on a number of *BSD systems. (Addresses Sourceforge Bug: 1341#2328708) 1342 1343Fix resize2fs's minimum size required for a file system so it doesn't 1344fail when "resize2fs -M" is run. (Addresses RedHat Bugzilla: #519131) 1345 1346Dumpe2fs now prints summary information about the contents of the 1347journal. 1348 1349Avoid printing scary error messages when e2fsck starts running 1350problems on low-memory systems, as it tends to panic and mislead the 1351user. (Addresses Debian Bug: #509529) 1352 1353Fix blkid's modules.dep parser so it handles compressed (.ko.gz) 1354modules files. (Address Red Hat Bug: #518572) 1355 1356Fix tune2fs so it can add a journal when an extent-enabled file system 1357is mounted. (Addresses Launchpad bug: #416648) 1358 1359Update Czech, Indonesian, Polish and Vietnamese translations (from the 1360Translation Project). 1361 1362Update/clarify man pages. (Addresses Sourceforge Bug: #2822186) 1363 1364Fixed various Debian packaging issues --- see debian/changelog for 1365details. (Addresses Debian Bugs: #540111) 1366 1367Programmer's Notes 1368------------------ 1369 1370The configure script supports the --with-cc, --with-ccopts, and 1371--with-ldopts options. Instead, the more standard use of CC=, 1372CCFLAGS=, and LDFLAGS= in the configure command line is used instead. 1373Also, --with-ld, which never worked, was also removed. 1374(Addresses Sourceforge Bug: #2843248) 1375 1376The in-tree header files are only used if the in-tree uuid or blkid 1377libraries are used. Otherwise, use the system-provided uuid or blkid 1378header files if using the system-provided libraries. 1379 1380Fix some build failures caused by --disable-* configure options. 1381 1382Work around a bug in autoconf 2.64. 1383 1384 1385E2fsprogs 1.41.9 (August 22, 2009) 1386================================== 1387 1388Fix a bug in e2fsck routines for reallocating an inode table which 1389could cause it to loop forever on an ext4 filesystem with the FLEX_BG 1390filesystem feature with a relatively rare (and specific) filesystem 1391corruption. This fix causes e2fsck to try to find space for a new 1392portion of the inode table in the containing flex_bg, and if that 1393fails, the new portion of the inode table will be allocated in any 1394free space available in the filesystem. 1395 1396Make e2fsck less annoying by only asking for permission to relocate a 1397block group's inode table once, instead of for every overlapping 1398block. Similarly, only ask once to recompute the block group 1399checksums, instead of once for each corrupted block group's checksum. 1400 1401Fix filefrag to avoid print the extent header if the FIEMAP ioctl is 1402not present, and it needs to fall back to using the FIBMAP ioctl. 1403 1404Fix filefrag to correctly print the number of extents for zero-length 1405files. (Addresses Debian Bug: #540376) 1406 1407Filefrag now has a -B option which forces the use of the FIBMAP ioctl 1408to more easily debug the FIBMAP code. 1409 1410Fixed filefrag for non-extent based files. 1411 1412Add a new program, e2freefrag, which displays information about the 1413free space fragmentation in an ext2/3/4 filesystem. 1414 1415Fix inode resizing via tune2fs -I so that it works correctly in the 1416face of non-empty bad blocks inodes, and if the filesystem was 1417formatted using the "mke2fs -E stride=N" option for RAID arrays. 1418 1419Fix regression in ext2fs_extent_set_bmap() caused e2fsck -fD to fail 1420and corrupt large directories if the directory needs to shrink by more 1421than one block. (Addresses Debian Bug: #537510) 1422 1423Fix e2fsck's buggy_init_scritps=1 so that the if the last write and/or 1424last mount times are in the future, they are corrected even if 1425buggy_init_scripts is set. This is needed because otherwise resize2fs 1426will refuse to resize the filesystem, even after running "e2fsck -f". 1427(Addresses Launchpad bug: #373409) 1428 1429E2fsck will now print much fuller information when the last mount time 1430or last written time is in the future, since most people can't seem to 1431believe their distribution has buggy init scripts, or they have a 1432failed CMOS/RTS clock battery. 1433 1434Enhance dumpe2fs to dump the extent information via the 'stat' 1435command, and more detailed extent information via the new command 1436'dump_extents'. 1437 1438Update French, Polish, Czech, and Sweedish translation from the 1439Translation Project. 1440 1441Fixed various Debian packaging issues --- see debian/changelog for 1442details. 1443 1444Programmer's Notes 1445------------------ 1446 1447Fixed miscellaneous gcc -Wall warnings. 1448 1449Fixed memory leak in error path in ext2fs_block_iterate2() 1450 1451Fixed non-Linux build of the intl directory by adding support for the 1452E/Q/V macros. 1453 1454The bitmap read/write functions now treat uninitialized bitmaps as 1455unallocated; this fixes a number of problems in all e2fsprogs for ext4 1456filesystems when there is a need to allocate new blocks or inodes, and 1457there aren't any free blocks or inodes in the already-used block 1458groups. 1459 1460Improve ext2fs_extent_set_bmap() to avoid creating new extents which 1461get inserted into the extent tree when they are not needed. 1462 1463 1464E2fsprogs 1.41.8 (July 11, 2009) 1465================================ 1466 1467Fix resize2fs's online resizing, fixing a regression which in 1468e2fpsrogs 1.41.7. (Addresses Debian Bug: #535452) 1469 1470Fix potential filesystem corruptions caused by using resize2fs to 1471shrinking ext4 filesystems with extents enabled. (Addresses Red Hat 1472Bug: #510379) 1473 1474Optimize uuid_generate() to avoid running uuidd if it is not setuid or 1475setgid and the currently running program doesn't have write access to 1476the uuidd work directory. 1477 1478Add safety checks (for non-Linux systems) so that uuidd isn't run with 1479file descriptors 0, 1, and 2 closed; and if they are closed, uuidd 1480will be careful not to close the file descriptor for its unix domain 1481socket when it detaches itself from the controlling tty. Also add 1482safety checks so that if the unix domain socket between the uuid 1483library and uuidd program is closed for any reason, both the library 1484and the uuidd will return an appropriate error code instead of looping 1485in an infinite loop. 1486 1487The e2croncheck script, which creates an LVM snapshot and then checks 1488the ext3/4 filesystem via the LVM snapshot, has been added to the 1489contrib directory. 1490 1491Fix filefrag program for files that have more than 144 extents. 1492 1493Update French, Polish, Czech, Indonsian, and Sweedish translation from 1494the Translation Project. 1495 1496Fixed various Debian packaging issues --- see debian/changelog for 1497details. (Addresses Debian Bug #535530) 1498 1499Update/clarify man pages. 1500 1501Programmer's Notes 1502------------------ 1503 1504Fix compilation problem when configured with --disable-uuid. 1505 1506Don't build uuidgen if configured with --disable-uuid. 1507 1508Add the new library function ext2fs_test_inode_bitmap_range(), and 1509optimized ext2fs_test_block_bitmap_range(), which will be needed for 1510future optimizations for e2fsck. 1511 1512Fix makefile dependencies for libcom_err so that the Makefiles work 1513well on non-GNU make program. (Addresses Sourceforge Patches: #2813809) 1514 1515Enhance the build system so that the full set of commands executed by 1516the Makefiles are displayed, instead of the Linux kernel summary 1517output, if the build was configured with --enable-verbose-makecmds, or 1518if GNU make is in use and the V variable is non-null, i.e., via "make 1519V=1". 1520 1521 1522E2fsprogs 1.41.7 (June 29, 2009) 1523================================ 1524 1525Fix a bug in libext2fs which can cause e2fsck and resize2fs to write 1526uninitalized data into the portion of the inode beyond the first 128 1527bytes when operating on inodes mapped via extents; potentially 1528corrupting filesystems. 1529 1530Fix memory leaks in e2fsprogs, including a very large memory leak 1531which can cause e2fsck to run out of memory when checking very large 1532filesystems using extents. 1533 1534The logsave program will now filter out the ^A and ^B characters when 1535writing to the console. 1536 1537Harden ext2fs_validate_entry() so that lsdel will not read beyond the 1538end of the block even if the directory block is corrupted. 1539 1540Fix debugfs from core dumping if the logdump command fails to open the 1541output file. 1542 1543Enhance badblocks to print the currently tested block number when 1544interrupted with ^C. 1545 1546Fix lsattr to exit with a non-zero status when it encounters errors. 1547(Addresses RedHat Bugzilla #489841) 1548 1549Fix e2fsprogs to use the same encoding for rec_len in 64k (and larger) 1550filesystems as the kernel when encoding rec_len >= 64k. For 64k 1551filesystems (currently all that is supported by e2fsprogs, this is 1552only a minor corner case). 1553 1554Resize2fs will now update the journal backup fileds in the sup0erblock 1555if the journal is moved; this avoids an unnecessary full fsck after 1556resizing the filesystem. (Addresses RedHat Bugzilla: #505339) 1557 1558Fix libext2fs to properly initialize i_extra_size when creating the 1559journal and resize inodes. 1560 1561Change badblocks to allow block sizes larger than 4k. 1562 1563Fix filefrag progam so it correctly checks for errors from the fiemap ioctl. 1564 1565Update Chinese and Czech translation from the Translation Project. 1566 1567Clean up various man pages. (Addresses Debian Bug #531385 and #523063) 1568 1569 1570Programmer's Notes 1571------------------ 1572 1573Add --disable-libuuid option to configure which uses an in-system 1574installed version of libuuid; the private version is enabled by 1575default. 1576 1577Add --valgrind-leakcheck option to the test_script program which runs 1578valgrind with the appropriate options to find memory leaks. 1579 1580 1581E2fsprogs 1.41.6 (May 30, 2009) 1582=============================== 1583 1584Fix a critical regression in e2fsck introduced version 1.41.5 which 1585can undo updates to the block group descriptors after a journal replay. 1586 1587If e2fsck receives an I/O error while replaying the journal, prompt 1588the user whether they want to ignore the error or not; if e2fsck is 1589run in preen mode, force fsck to abort the preen pass. 1590 1591Fix a bug which would cause e2fsck to core dump if the filesystem 1592contains a corrupt extent header, and the user declines to clear the 1593inode. (Addresses Sourceforge Bug: #2791794) 1594 1595Fix e2fsck to restart only once in the case of multiple block groups 1596which inodes in use that are in the uninitialized part of the block 1597group's inode table. 1598 1599To reduce user confusion, if the /etc/mtab file is missing 1600ext2fs_check_mount_point and ext2fs_check_if_mounted will return a 1601new, explicit error code to indicate this case. This will cause 1602e2fsck to give a clearer error message when the user is using buggy 1603rescue CD's that don't properly set up /etc/mtab. (Addresses Debian 1604Bug: #527859) 1605 1606Fix e2fsck so that if the primary group descriptors are corrupted, but 1607the backup superblock is entirely invalid, to go back to using (and 1608fixing) the priary group descriptors instead of completely giving up 1609on the filesystem. (Addresses Debian Bug: #516820) 1610 1611Change e2fsck to not abort a preen pass if an inode is found to have 1612i_file_acl_hi non-zero. Ext3 filesystems don't care, and newer 1613kernels (post 2.6.29) will ignore this field. So let's fix it 1614automatically during the preen pass instead of forcing the user to fix 1615this up manually. (Addresses Debian Bug: #526524) 1616 1617Add resource tracking for e2fsck passes 1b through 1d. 1618 1619Speed up e2fsck by eliminating unnecessary journal checks if the 1620filesystem is already mounted and doesn't need recovery (since the 1621kernel would have run the journal when the filesystem was mounted.) 1622Also speed up e2fsck by avoiding unnecessary block group descriptor 1623updates in ext2fs_close(). 1624 1625Add support to chattr to migrate inodes from using direct/indirect 1626blocks to extents. 1627 1628Avoid corrupting the filesystem if there is an attempt to shrink a 1629filesystem using resize2fs smaller than posible by making 1630ext2fs_set_bmap() more careful not to delete the old block until the 1631new block can be inserted. In addition, fix a bug in how the minimum 1632size of the filesystem (plus a safety margin) is calculated, and 1633modify resize2fs to refuse to shrink the filesystem below that minimum 1634size without the force flag. 1635 1636Teach blkid to try to figure out DM device names using take advantage 1637of information in sysfs when running on kernels (2.6.29 and later) 1638which provide this information in /sys/block/dm-<N>/dm/name; this is 1639much faster than scanning for the devncie number in /dev/mapper. 1640 1641Fix blkid to prefer /dev/mapper/<name> device names instad of the 1642private /dev/dm-X names. 1643 1644Add an -a option to debugfs's close_filesys command which writes any 1645changes to the superblock or block group descriptors to all of the 1646backup superblock locations. 1647 1648Add support to the filefrag program to use the FIEMAP ioctl. 1649 1650Update Chinese translation from the Translation Project. 1651 1652Clean up various man pages. (Addresses Red Hat Bugzilla: #502971 and 1653Launchpad Bug: #381854) 1654 1655Fixed various Debian packaging issues --- see debian/changelog for 1656details. (Addresses Debian Bug #506064) 1657 1658 1659Programmer's Notes 1660------------------ 1661 1662Add test code to make sure e2fsck's problem.c doesn't have two problem 1663codes assigned to duplicate values. 1664 1665Avoid using a hard-coded path for /bin/rm in block's test_probe.in. 1666(Addresses Sourceforge Bug: #2780205) 1667 1668Clean up e2fsck by removing #ifdef RESOURCE_TRACK by adding an empty 1669function declaration for init_resource_track() and 1670print_resource_track() when RESOURCE_TRACK is not defined. 1671 1672The test code which is used to build the tst_csum progam has been 1673moved from from tst_csum.c into csum.c under an #ifdef DEBUG to to 1674avoid compile problems caused by not having a prototype for 1675ext2fs_group_desc_csum(). (Addresses Sourceforge Bug #2484331) 1676 1677Update the config.guess and config.sub file to the latest from the 1678FSF, to allow e2fsprogs to build on the avr32 platform. (Addresses 1679Debian Bug: #528330) 1680 1681Add a new function, ext2fs_extent_open2(), which behaves like 1682pext2fs_extent_open(), but uses the user-supplied inode structure 1683when opening an extent instead of reading the inode from disk. 1684 1685 1686E2fsprogs 1.41.5 (April 23, 2009) 1687================================= 1688 1689Fix a number of filesystem corruption bugs in resize2fs when growing 1690or shrinking ext4 filesystems off-line (i.e., when the ext4 filesystem 1691is not mounted). 1692 1693Debugfs can now set i_file_acl_high via the set_inodes_field command, 1694and print a 64-bit file acl. This is useful for debugging filesystem 1695corruptions where the high bits of i_file_acl_high are set. E2fsck 1696will detect and fix non-zero i_file_acl_high on 32-bit filesystems 1697since some Linux kernel versions pay attention to this field even when 1698they shouldn't. 1699 1700Speed up e2fsck when checking clean filesystems by avoiding 1701unnecessary block reads, and coalescing the block group descriptor 1702blocks so they are read using a single read operation. 1703 1704The libuuid library will now close all file descriptors before running 1705uuidd. This avoids problems when the calling program has open sockets 1706which then never get closed because uuidd is a long-running helper 1707daeomn. (Addresses Launchpad bug: #305057) 1708 1709In order to avoid unnecessary full filesystem checks by e2fsck after 1710an on-line resize, e2fsck will ignore the NEEDS_RECOVERY flag set on 1711the backup superblocks. (Addresses Red Hat Bugzilla: #471925) 1712 1713Mke2fs will avoid trying to create the journal when run in 1714superblock-only mode (mke2fs -S), since the left-over journal in the 1715inode table will cause mke2fs to fail. 1716 1717Fix a bug in libext2fs functions that check to see if a particular 1718device or filesystem image is mounted, which would cause these 1719functions to report that a file identical to the (relative) pathname 1720used by a pseudo-filesystem was mounted when in fact it was not. 1721 1722Update Czech translation from the Translation Project. 1723 1724Add Chinese (simplified) translation from the Translation Project. 1725 1726Fix support for external journals (which was broken in e2fsprogs 17271.41.4). 1728 1729Fix a regression in debugfs where the "stat" command when no 1730filesystem was open would cause debugfs to crash with a segmentation 1731violation. 1732 1733Starting in the 2.6.29 linux kernel, the ext4 filesystem driver can be 1734used to support filesystems without a journal. Update the blkid 1735library so it understands this. 1736 1737The blkid library will remove an entry from the blkid cache 1738representing the entire disk if partitions are found, since presuambly 1739the device previously had no partition table, but has now transitioned 1740to using a partition table. 1741 1742Add a check to mke2fs and tune2fs that the argument to the -m option 1743(which specifies the reserved ratio) must be greater than zero. 1744(Addresses Debian Bug: #517015) 1745 1746Add support for tracking the number kilobytes written to the 1747filesystem via the superblock field s_kbytes_written. It will be 1748updated by the kernel as well as by e2fsprogs programs which write to 1749the filesystem. This is useful for tracking the wear to filesystems 1750on Solid Sstate Drives. 1751 1752Fix compatibility issue in the libext2fs info file and makeinfo 1753version 4.12. (Addresses Red Hat Bugzilla: #481620) 1754 1755Update/clarify man pages. (Addresses Debian Bug: #515693, #365619) 1756 1757Fixed various Debian packaging issues --- see debian/changelog for 1758details. (Addresses Debian Bug: #506279) 1759 1760Programmer's Notes 1761------------------ 1762 1763Fix Hurd compilation problem in e2fsck and tune2fs (Addresses Debian 1764Bug: #521602) 1765 1766Fix various gcc compilation warnings and other programming cleanups. 1767(Addresses Red Hat Bugzilla: #486997) 1768 1769Add support for building the blkid command statically. 1770 1771Add support for disabling the built-in blkid library in favor of a 1772system-installed blkid implementation via the configure option 1773--disable-libblkid. 1774 1775 1776E2fsprogs 1.41.4 (January 27, 2009) 1777=================================== 1778 1779Fixed a bug which could sometimes cause blkid to return an exit value 1780of zero for a non-existent device (Addresses Debian Bug: #502541) 1781 1782Blkid will now recognize ext3 filesystems that have the test_fs flag 1783set as ext3 filesystems. 1784 1785The blkid library will now recognize btrfs filesystems and swap 1786devices currently used by user-level software suspend. 1787 1788Tune2fs now updates the block group checksums when changing the UUID 1789to avoid causing e2fsck to complain vociferously at the next reboot. 1790 1791Tune2fs's inode size resizing algorithms have been fixed so it is not 1792vastly inefficient for moderate-to-large filesystems, due to some 1793O(n**2) and O(n*m) algorithms that didn't scale well at all. 1794 1795Fix tune2fs's inode resizing algorithm so it will not corrupt 1796filesystems laid out for RAID filesystems; in addition, tune2fs will 1797refuse to change the inode size for filesystems that have the flex_bg 1798feature enabled. (This is a limitation in the current implementation 1799of tune2fs -I.) 1800 1801E2fsprogs 1.41 broke debugfs's logdump command for normal ext3/4 1802filesystems with 32-bit block numbers, when the headers for 64-bit 1803block numbers was added. This regression has been fixed. 1804 1805Debugfs's ncheck command has been fixed to avoid printing garbage 1806characters at the end of file names. 1807 1808Fix resize2fs for ext4 filesystems. Some blocks that that need moving 1809when shrinking filesystems with uninit_bg feature would not be moved. 1810In addition, blocks and inode table blocks were not being correctly 1811freed when shrinking filesystems with the flex_bg feable, which caused 1812resize2fs -M to fail. Finally, when blocks are moved, make sure the 1813uninitialized flag in extents is preserved. 1814 1815Fix bug which caused dumpe2fs to abort with an error if run on a 1816filesystem that contained an external journal. 1817 1818Some distributions used "mke3fs" as an alias for "mkfs.ext3"; check 1819for this in argv[0] to provide better legacy support for these 1820distributions. This is a practice that should NOT be continued, 1821however. 1822 1823Mke2fs now has a new option -U, which allows the user to specify the 1824UUID that should be used for the new filesystem. 1825 1826Mke2fs will treat devices that are exactly 16TB as if they were 16TB 1827minus one block. This allows users who have read that ext3 supports 1828up to 16TB filesystems and who create a 16TB LVM to not get confused, 1829since the true limit is really 16TB minus one block. 1830 1831E2fsck will no longer abort an fsck run if block group has an errant 1832INODE_UNINIT flag. 1833 1834E2fsck now distinguishes between fragmented directories and fragmented 1835files in verbose mode statistics and in the fragcheck report. 1836 1837Fix a bug in e2fsck which casued it double count non-contiguous 1838extent-based inodes. 1839 1840E2fsck will leave some slack space when repacking directories to allow 1841room for a few directory entries to be added without causing leaf 1842nodes to be split right away. 1843 1844Fix a bug which caused e2fsck to crash when it comes across a 1845corrupted interior node in an extent tree with the error message: 1846"Error1: Corrupt extent header on inode XXXXXX" 1847 1848E2fsck problem descriptions involving the journal are no longer 1849referred to as "ext3" problems, since ext4 filesystems also have 1850journals. 1851 1852Fix a long-standing bug in e2fsck which would cause it to crash when 1853replying journals for filesystems with block sizes greater than 8k. 1854 1855Update Catalan translation from the Translation Project. 1856 1857Fixed various Debian packaging issues --- see debian/changelog for 1858details. (Addresses Debian Bugs: #503057, #502323, #511207) 1859 1860Programmer's Notes 1861------------------ 1862 1863Fix build of e2fsck.profiled, and add support for building profiled 1864binaries in the misc directory if configured with --enable-profile. 1865 1866The ext2fs_open() function now performs more sanity checks on the 1867superblock to avoid potential divide by zero errors by other parts of 1868library. 1869 1870The ext2fs_read_inode_full() function now has a safety check to avoid 1871a segmentation fault on corrupted filesystems. 1872 1873The ext2fs_new_inode() function now has a sanity check so that if the 1874s_first_inode field in the superblock is insane, it will return 1875EXT2_ET_INODE_ALLOC_FAIL instead of returning an invalid inode number. 1876 1877To avoid segmentation faults, ext2fs_block_alloc_stats() and 1878ext2fs_inode_alloc_stats() now validates the passed inode or block 1879number to avoid overrunning an array boundary. 1880 1881Various signed/unsigned errors for variables containing block numbers 1882have been fixed. 1883 1884Accomodations for gcc's stupidity in not realizing that constant 1885strings that do not contain a '%' character are safe to use in format 1886strings have been made so that distributions that want to compile 1887e2fsprogs with -Werror=format-security have an easier time doing so. 1888 1889Added a new 64-bit getsize interface, ext2fs_get_device_size2(). 1890 1891Added the utility make-sparse.c to the contrib directory. 1892 1893The ext2fs_block_iterate2() function now reflects errors from 1894ext2fs_extent_set_bmap() to the caller, if the callback tries to 1895change a block on an extent-based file, and ext2fs_extent_set_bmap() 1896fails for some reason (for example, there isn't enough disk space to 1897split a node and expand the extent tree. 1898 1899The ext2fs_block_iterate2() function will preserve the uninit flag in 1900extents when the callback function modifies a block in an extent-based 1901file. 1902 1903E2fsck will now flag filesystems that have an insane s_first_ino field 1904in their superblock, and attempt to use a backup superblock to repair 1905the filesystem. 1906 1907 1908E2fsprogs 1.41.3 (October 12, 2008) 1909=================================== 1910 1911E2fsck has been fixed so it prints the correct inode number for 1912uinit_bg related problems. 1913 1914E2fsck will now offer to clear the test_fs flag if the ext4 filesystem 1915is available on linux. This can be disabled via a configuration 1916option in /etc/e2fsck.conf. 1917 1918Fix a file descriptor leak in libblkid when checking to see if an ext4 1919or ext4dev module exists. 1920 1921Fix a bug in e2fsck where in preen mode, if there are disk I/O errors 1922while trying to close a filesystem can lead to infinite loops. 1923(Addresses Red Hat Bugzilla #465679) 1924 1925Fix a bug in resize2fs where passing in a bogus new size of 0 blocks 1926will cause resize2fs to drop into an infinite loop. (Addresses Red 1927Hat Bugzilla: #465984) 1928 1929Add a check in the Unix I/O functions in libext2fs so that when a 1930device is opened read/write, return an error if the device is 1931read-only using the BLKROGET ioctl. 1932 1933Fix debugfs's ncheck command so that it prints all of the names of 1934hardlinks in the same directory. 1935 1936Fix a bug in libblkid so it correctly detects whether the ext4 and 1937ext4dev filesystems are available, so that the ext4dev->ext4 1938fallback code works correctly. 1939 1940Programmer's Notes 1941------------------ 1942 1943Fix a parallel build problem by making sure util/subst is built before 1944trying to build the lib/et directory. (Addresses Sourceforge Bug: 1945#2143281) 1946 1947Updated "make depend" information for crc16.o 1948 1949 1950E2fsprogs 1.41.2 (October 2, 2008) 1951================================== 1952 1953Fix e2fsck's automatic blocksize detection. This fixes a regression 1954from e2fsprogs 1.40.7 which caused e2fsck to fail if the user 1955specifies a block number using the -b option if the blocksize option 1956isn't also specified using -B. Unfortunately, users very commonly 1957invoke e2fsck using "e2fsck -b 32768 /dev/hdXXX" to use the backup 1958superblock; in fack e2fsck will often suggest this kind of command 1959line. Oops. 1960 1961Enhance the debugfs's "ncheck" command so it will print all of the 1962pathnames for the specified inodes. (Previously, in some cases ncheck 1963might not print a pathname for an inode at all if some of the other 1964inodes had multiple hard links.) 1965 1966Enhance debugfs's "hash" command so the hash seed can be specified via 1967a command-line option. In addition, allow the hash algorithm to be 1968specified by name instead of just by number. 1969 1970Fix e2fsck so that we don't accidentally print the translation file's 1971header when asking the user a custom question so there is no prompt 1972defined for a particular problem record. For example, the question 1973"Run journal anyway" will get the PO header tacked on because e2fsck 1974erroneously passed the null string to _(). (Addresses Launchpad Bug: 1975#246892) 1976 1977Enhance badblocks so that it can test a normal file which is greater 1978than 2GB. 1979 1980Enhance the badblocks command so that it displays the time and 1981percentage complete when in verbose mode. (Addresses Debian Bug: 1982#429739) 1983 1984Fix a potential memory leak in a error handling path in debugfs's 1985ncheck function. 1986 1987Fix a potential memory corruption problem if a memory allocation fails 1988in resize2fs. 1989 1990Fix the usage message for debugfs's logdump command to be consistent 1991with its man manpage. 1992 1993Update Polish, French, Vietnamese, Dutch, Indonesian, German, Czech, 1994and Sweedish translation from the Translation Project. 1995 1996Add documentation for the file I/O functions to the libext2fs.texinfo 1997file. (Addresses Debian Bug: #484877) 1998 1999Update and clarified various man pages. (Addresses Launchpad Bug 2000#275272; Addresses Debian Bugs: #498100, #498101, #498102, #498103) 2001 2002Fixed various Debian packaging issues --- see debian/changelog for 2003details. (Addresses Debian Bug: #497619) 2004 2005Programmer's Notes 2006------------------ 2007 2008Fix a potential file descriptor leack in libcom_err by setting the 2009close-on-exec flag for a fd used for debugging. (Addresses Red Hat 2010Bugzilla #464689) 2011 2012Fix a potential race in libcom_err by using sem_post/sem_init. SuSE 2013has been carrying a patch for a long time to prevent a largely 2014theoretical race condition if a multi-threaded application adds and 2015removes error tables in multiple threads. Unfortunately SuSE's 2016approach breaks compatibility by forcing applications to link and 2017compile with the -pthread option; using pthread mutexes has 2018historically been problematic. We fix this by using sem_post/sem_init 2019instead. 2020 2021Fix e2fsprogs-libs build failure due to 'subs' target. (Addresses 2022Sourceforge Bug: #2087502) 2023 2024Avoid linking e2initrd_helper, debugfs, blkid, and fsck with unneeded 2025libraries when using ELF shared libraries. 2026 2027Fix ELF shared library when building on systems that don't already 2028have the e2fsprogs shared libraries already installed. (Addresses 2029Sourceforge Bug: #2088537) 2030 2031Fix the pkg-config files so they work correctly when linking with 2032static libraries and fix the include directory so programs don't have 2033to use #include <ext2fs/ext2fs.h>, but can use #include <ext2fs.h> 2034instead. (Addresses Sourceforge Bug: #2089537) 2035 2036Make sure ext2fs_swab64() is compiled for all platforms, and not just 2037for x86. (Addresses Debian Bug: #497515) 2038 2039Remove the unused ext2fs_find_{first,ext}_bit_set() functions for all 2040non-x86 platforms. (They had been removed for x86 earlier.) 2041 2042Fix diet libc compilation support, which had bitrotted due to lack of 2043TLC. Fixing this improves general portability. 2044 2045When installing the link library when using ELF shared libraries, 2046avoid using absolute pathnames if the link library and the shared 2047library are installed in the same directory. (Addresses Sourceforge 2048Bug: #1782913) 2049 2050Fix gen-tarball so it will work even if the top-level directory has 2051been renamed to something other than "e2fsprogs". Also make 2052gen-tarball print the size of the resulting tar.gz file. 2053 2054 2055E2fsprogs 1.41.1 (September 1, 2008) 2056==================================== 2057 2058Many people are forgetting to update their mke2fs.conf file, and this 2059causes ext3, ext4, and ext4dev filesystems won't get created with the 2060proper features enabled. We address this in two ways. First, mke2fs 2061will issue a warning if there is not definition for an ext3, ext4, or 2062ext4dev filesystem and the user is trying to create such a filesystem 2063type. Secondly, when installing from a source build, "make install" 2064will provide basic configuration file handling for /etc/mke2fs.conf. 2065If it exists, and does not mention ext4dev, it will be moved aside to 2066/etc/mke2fs.conf.e2fpsrogs-old and the new /etc/mke2fs.conf file will 2067be installed. If the existing /etc/mke2fs.conf file does mention 2068ext4dev, then "make install" will install official mke2fs.conf file as 2069/etc/mke2fs.conf.e2fsprogs-new and issue a message to the user that 2070they should look to see if any changes need to be merged. 2071 2072The mke2fs program will now create the journal in the middle of the 2073filesystem, since this minimizes seek times on average for fsync-heavy 2074workloads. In addition, mke2fs will now create journals using extents 2075for filesystems that support them. This results in a more efficient 2076encoding for the journal since it eliminates the need for using 2077indirect blocks. 2078 2079The mke2fs program will avoid allocating an extra block to the 2080journal. (Addresses Sourceforge Bug: #1483791) 2081 2082Mke2fs will correctly enforce the prohibition against features 2083(specifically read-only features) in revision 0 filesystems. (Thanks 2084to Benno Schulenberg for noticing this problem.) 2085 2086Mke2fs previously would occasionaly create some slightly non-optimally 2087placed inode tables; this bug has been fixed. 2088 2089The mke2fs and tune2fs programs now print the correct usage message 2090describing the maximum journal size. (Addresses Debian Bug: #491620) 2091 2092Add support for setting the default hash algorithm used in b-tree 2093directories in tune2fs (from a command-line option) or mke2fs (via 2094mke2fs.conf). In addition, change the default hash algorithm to 2095half_md4, since it is faster and better. 2096 2097The blkid library will now recognize MacOS hfsx filesystems, and 2098correctly extract the label and uuid for hfs, hfsx, and hfsplus 2099filesystems. (Addresses Sourceforge Feature Requests: #2060292) 2100 2101The blkid library has improved detection of JFS and HPFS filesystems. 2102(Addresses Launchpad Bug: #255255) 2103 2104The blkid library is now much more efficiently handling devicemapper 2105devices, mainly by no longer using the devicemapper library. This can 2106speed up access for systems with a large number of device mapper 2107devices. 2108 2109Blkid had a number of cache validation bugs in libblkid that have been 2110fixed. (Addresses Debian Bug: #493216) 2111 2112Resize2fs will now properly close out the "updating inode references" 2113progress bar so there is a newline printed before printing the final 2114"resize is successful" message. 2115 2116Resize2fs will now correctly handle filesystems with extents and/o 2117uninitialized block groups correctly when file/directory blocks need 2118to relocated (i.e., when shrinking a filesystem or if the resize_inode 2119is not present). To support this, the ext2fs library now supports 2120initializing inode and block bitmaps that are not yet initialized when 2121allocating them using ext2fs_new_block() and ext2fs_new_inode(). In 2122addition, e2fs_block_iterate2() can now support changing the location 2123of interior nodes of an extent tree, and ext2fs_extent_set_bmap() has 2124been optimized to avoid creating unnecessary new extents when updating 2125the location of blocks in the extent tree. This will also help out 2126e2fsck's recovery of obscurely corrupted filesystems with extents, 2127when blocks are claimed by multiple inodes. 2128 2129Add support for on-line resizing ext4 filesystem with the flex_bg 2130filesystem feature. The method for doing so is not optimal, but to do 2131a better job will require kernel support. 2132 2133E2fsprogs 1.41.0 intrduced a bug in libext2fs which casued e2image and 2134debugfs programs to not be able to read e2image files; the signed 2135vs. unsigned bug in the code which read bitmaps from the e2image has 2136been fixed. (Addresses Debian Bug: #495830) 2137 2138Resize2fs is now correctly managing the directory in-use counts when 2139shrinking filesystems and directory inodes needed to be moved from one 2140block group to another. This bug has been around since e2fsprogs 21411.26, and is largely harmless, but does cause a filesystem corruption 2142which will be flagged by e2fsck after the filesystem has been shrunk. 2143 2144E2fsck will no longer issue spurious complaints about the inode size 2145caused by very large extent-based files, and by blocks reallocated 2146using fallocate() with the FALLOC_FL_KEEP_SIZE option. (Addresses 2147Kernel Bugzilla: #11341) 2148 2149Mke2fs will now set the creation timestamp on the lost+found directory 2150and the root directory. (More generally, all new inodes created using 2151the ext2fs library will correctly set the creation timestamp.) 2152 2153E2fsck now correctly calculates ind/dind/tind statistics in the 2154presence of extent-based files. In addition, "e2fsck -v" will report 2155statistics of the depth of extent trees in the filesystem. E2fsck can 2156also give an inode fragmentation report using "e2fsck -E fragcheck" 2157which can be useful when debugging the kernel block allocation 2158routines. 2159 2160Fix support for empty directory blocks in ext4 filesystems with 64k 2161blocksize filesystems. 2162 2163E2fsck will now print the depth of corrupt htree directories. 2164 2165Debugfs's htree command now correctly understands extent-based 2166directories. It will also print out the minor hash as well as the 2167major hash. 2168 2169Debugfs has a new command which will print the supported features of 2170e2fsprogs, to enable scripts to know whether the installed version of 2171e2fsprogs can support a specific feature. 2172 2173Debugfs will now write files using extents for filesystems that 2174support them. 2175 2176The error message printed by "tune2fs -I" if the inode size was too 2177small was rather confusing, so it has been improved. Also, we won't 2178try to create an undo log until we know that command-line-specified 2179parameters such as "tune2fs -I <inode size>" are valid. 2180 2181Given some filesystems found "in the wild" that had non-zero block 2182group checksums even though the uninit_bg/gdt_sum feature was not 2183enabled, e2fsck would issue spurious error messages. Teach 2184ext2fs_group_desc_csum_verify() to ignore the block group checksum 2185entirely if the feature flag is not set. (Addresses Debian Bug: 2186#490637) 2187 2188The blkid program will now print out a user-friendly listing of all of 2189the block devices in the system and what they contain when given the 2190-L option. (Addresses Debian Bug: #490527) 2191 2192The filefrag program now has a more accurate calculation for the 2193number of ideal extents. (Addresses Debian Bug: #458306) 2194 2195The test I/O manager is now enabled by default, but its overhead is 2196only incurred when it would be enabled via the TEST_IO_FLAGS or 2197TEST_IO_BLOCK environment variables. 2198 2199Typographical errors in various program strings and usage messages 2200have been fixed; most of these were pointed out by the e2fsprogs 2201message catalog translators. (Thanks, translators!) 2202 2203Update and clarified various man pages, as well as some typographical 2204errors in the libext2fs texinfo file. 2205 2206Fixed various Debian packaging issues --- see debian/changelog for 2207details. 2208 2209Add Indonesian and update French, Polish, Dutch, German, Sweedish, 2210Czech, and Vietnamese Translations. (Addresses Debian Bugs: #313697, 2211#401092) 2212 2213Programmer's Notes 2214------------------ 2215 2216Fix portability problem with the badblocks group; for systems that 2217don't have nanosleep(), try using usleep() instead. 2218 2219The "make check" target in the e2fsck directory now sets 2220LD_LIBRARY_PATH before running the various e2fsck internal library 2221regression tests. 2222 2223The crc32 regression test in the e2fsck library is now portable to 2224greater varienty of environments, including big-endian systems and 2225when cross-building e2fsprogs for embedded systems. (Addresses 2226Sourceforge Bug: #2019287) 2227 2228The ext2fs_extent_set_bmap() had some bugs when setting the first 2229block in a file, or when replacing a single block extent. Those cases 2230fortunately were came up relatively rarely when e2fsck was checking 2231files, but caused some problems when resize2fs was shrinking 2232extent-based files. 2233 2234Fix a potential core-dumping bug in libe2p's iterate_on_dir() 2235function. 2236 2237Various ext2fs library functions --- ext2fs_block_iterate2(), 2238ext2fs_initialize() and ext2fs_extent_open() --- now correctly free 2239allocated memory to avoid memory leaks in all of their error return 2240paths. 2241 2242Ext2ed was failing to build because masix support had been removed in 2243the rest of e2fsprogs, so ext2ed no longer has masix support, either. 2244 2245The configure script now respects the LDFLAGS environment variable if 2246it is set when configure is called. (Addresses Sourceforge Feature 2247Request: #1937287) 2248 2249Libuuid is now more portable to the Windows platform. (Addresses 2250Sourceforge Feature Request: #1937287) 2251 2252The configure script now uses AC_MSG_{RESULT,WARN,ERROR} instead of 2253bare echo commands so that configure flags such as --quiet work 2254correctly. (Addresses Sourceforge Patches: #2058794) 2255 2256A few uses of sprintf have been removed from the ext2fs library to 2257make life easier for bootloaders with a limited libc environment. 2258(Addresses Sourceforge Bug: #2049120) 2259 2260The ext2fs_read_inode() checks the validity of the inode number passed 2261to it earlier, to avoid doing some needless work when it would fail 2262anyway. 2263 2264The ext2fs_open() checks the validity of the blocksize parameter 2265passed to it earlier, to avoid doing some needless work when it would 2266fail anyway. 2267 2268Disable a very annoying automatic "%.sh -> %" GNU make rule in the 2269top-level Makefile. That automatic rule is used to better support 2270SCCS, but it caused problems for a particular niche distribution which 2271likes to use configure.sh files to store the configure options used to 2272build a package. Unfortuntaely GNU make will use the configure.sh to 2273replace the configure script, resulting in a self-inflicted fork bomb 2274leading to an out-of-memory crash. 2275 2276To support old GNU C compilers don't use C99/C++ comments, but only 2277K&R style comments, and don't try to use __builtin_expect if __GNUC__ 2278is less than 3. (__builtin_expect is only supported for gcc versions 22792.96 and up, and it's tricky to check for gcc 2.95 vs gcc 2.96; since 2280this is an optimization, we only try to use __builtin_expect for gcc 3 2281and up.) 2282 2283In e2fsck's crc routines, make sure we use WORDS_BIGENDIAN instead of 2284__LITTLE_ENDIAN, which are only defined by glibc's header files and 2285hence isn't portable. 2286 2287For the convenience for some distributions that need a static tune2fs, 2288the Makefile for misc/ now has a tune2fs.static target. 2289 2290The ext2fs_block_iterate2() function now supports BLOCK_FLAG_APPEND 2291for extent-based files 2292 2293The ext2fs_bmap() function now supports BMAP_ALLOC for extent-based 2294files. 2295 2296All source files no longer have any trailing white space. 2297 2298The io_channel_read_blk64() and io_channel_write_blk64() functions are 2299now functions instead of C preprocessor macros to provide better 2300forward compatibility. 2301 2302The e2fpsrogs translation template now expands the @x abbrevation. 2303 2304Various namespace leackages in libblkid, libe2p, and libext2fs have 2305been fixed. 2306 2307Fix a parallel build problem in e2fsprogs. 2308 2309E2fsprogs is now more portable to Solaris. 2310 * blkid no longer assumes that the TIOCGSIZE and TIOCGWINSZ ioctl's 2311 are always present. 2312 * Scripts do not assume that /bin/true is always in /bin 2313 * Don't use __FUNCTION__ since Solaris's C99 doesn't support it. 2314 * Flush stdio handles before calling setbuf(), since Solaris will 2315 discard any pending output to the stream. 2316 * Define _XOPEN_SOURCE to 600 since Solaris's header files are very 2317 picky about which C compiler can beused for SUSv3 conformance. 2318 Use of C99 is not compatible with SUSv2 (_XOPEN_SOURCE=500), 2319 and C89 is not compatible with SUSv3 (_XOPEN_SOURCE=600). 2320 Since we need some SUSv3 functions, consistently use SUSv3 so 2321 that e2fsprogs will build on Solaris using c99. 2322 * Solaris C99 does not support varargs C preprocessor macros 2323 * Solaris header files pollute the C namespace if in/netinet.h 2324 is included, which conflicts with e2fsprogs' use of the kmem_cache_t 2325 typedef. 2326 * Solaris ships with a pathetically ancient shell in /bin/sh so we 2327 avoid the use of various more avanced shell constructs such as $(). 2328 2329The "make rpm" command will now take some extra configure optiosn from 2330the build environment without needing to patch the source tree. 2331 2332The ext2fs_add_dir_block() function will now grow the dblist more 2333aggressively as an optimization to avoid copying the array too often. 2334 2335The e2fsck_write_bitmaps() will write the block and inode bitmaps 2336together instead of in two passes. 2337 2338 2339E2fsprogs 1.41 (July 10, 2008) 2340============================== 2341 2342Add support for ext4 filesystem features, in particular extents, 2343uninit_bg, flex_bg, huge_file, and dir_nlink features. Also add 2344support for checking journal checksums. Debugfs will print new 2345superblock and inode fields that were defined for ext4. For example, 2346the nanosecond and i_version fields of an inode, and the 2347s_min_extra_isize and s_wanted_extra_isize fields from the superblock. 2348Note: Resize2fs doesn't currently support the combination of flex_bg 2349and !resize_inode. (Addresses Debian Bug: #388452, #425477) 2350 2351Tune2fs can support migrating a filesystem from using 128 byte inodes 2352to 256 byte inodes, so it can take advantage of the full features of 2353ext4. 2354 2355Add support for "undo" support. E2fsck and mke2fs can optionally 2356record an undo log which can replayed by the program e2undo. 2357 2358E2fsck could damage a filesystem by trying to relocate inode tables 2359due to corrupted block group descriptors, where the attempted inode 2360table relocation would do far more harm than good. E2fsck has been 2361fixed to detect this these sorts of corrupted block group descriptors 2362much earlier in e2fsck processing, so it can try to use the backup 2363superblock and block group descriptors first. This should be a much 2364better strategy for recovering these types of corrupted filesystems. 2365(Addresses Sourceforge Bug: #1840291) 2366 2367E2fsck will display a more understandable message when the last check 2368field in the superblock is in the future. (Addresses Debian Bug: 2369#446005). 2370 2371E2fsck now performs more extensive and careful checks of extended 2372attributes stored in the inode. 2373 2374Enhance mke2fs to print a more explanatory error message when 2375ext2fs_get_device_size() returns EFBIG. (Addresses Debian Bug: 2376#488663) 2377 2378Fix mke2fs to use a default block size of 4k when formatting an 2379external journal device. This is done by using a fixed filesystem 2380type list that consists only of the single filesystem type "journal" 2381when looking up configuration keys in /etc/mke2fs.conf. (Addresses 2382Debian Bug: #488663) 2383 2384Speed up how mke2fs writes the journal data blocks by writing the disk 2385blocks in larger chunks. 2386 2387Fix blkid handling of stale devices. Fix a bug which could cause a 2388core dump while garbage collecting the blkid cache, and assure that 2389blkid_find_dev_with_tag() never returns a non-existent device. Also, 2390if a filesystem is found at a new /dev location, eliminate any 2391duplicate stale entries which can not be verified. 2392(Addresses Debian Bugs: #487758, #487783) 2393 2394Add more paranoid checks for LVM volumes and swap partitions in 2395blkid's probe function, to reduce the chances of false positives. 2396 2397The mke2fs program now has a much more sophisticated system for 2398controlling configuration parameters of a newly created filesystem 2399based on a split filesystem and usage type system. The -t option to 2400mke2fs was a deprecated alias to -c; it now specifies a filesystem 2401type (ext2, ext3, ext4, etc.), while the -T option can now be a comma 2402separated usage list. The filesystem type information and type 2403information is used to extract configuration parameters from the 2404/etc/mke2fs.conf file. 2405 2406The mke2fs program will no longer complain and request the -f option 2407when the user tries to create a filesystem with greater than 2**31 2408blocks. 2409 2410When creating a filesystem for the GNU Hurd use a fs-type of Hurd and 2411adjust the mke2fs.conf file so filesystems for the Hurd are created 2412with a blocksize of 4096 and inode size of 128, which is all it knows 2413how to handle. (Addresses Debian Bug: #471977) 2414 2415Mke2fs will always make sure that lost+found always has at least 2 2416blocks, even for filesystems with very large blocksizes (i.e., 64kb). 2417 2418Resize2fs will now print the minimum needed filesystem size if given 2419the -P option, and will resize the filesystem to the smallest possible 2420size if given the -M option. 2421 2422Fix resize2fs to clean up the resize_inode if all of the reserved gdt 2423blocks are consumed during an off-line resize. 2424 2425The "ls" command in debugfs now supports the -p option, which causes 2426it to quote the filenames so that spaces or tabs in directory entries 2427are easily visible. (Addresses Red Hat Bugzilla: #149480; Addresses 2428Sourceforge Feature Request: #1201667) 2429 2430Fix a potential off-by-one buffer oveflow in the fs_device_name in an 2431e2image file. 2432 2433The chattr program will return a non-zero exit code in case of 2434failures, and error messages can be suppressed with the -f option. 2435(Addresses Red Hat Bugzilla: #180596) 2436 2437Fix a bug in badblocks which caused it to overrun an array and likely 2438crash if more than 8 test patterns are specified using the -t option. 2439(Addresses Debian Bug: #487298) 2440 2441Add support to badblocks to limit how quickly it reads from the disk 2442drive (so it can be used for background scrubbing), and so it will 2443abort after finding a given number of errors. 2444 2445Remove support for the legacy big-endian filesystem format which only 2446existed on extremely long-dead PowerPC kernels almost a decade ago. 2447 2448Remove MASIX support from e2fsprogs. 2449 2450Add I/O statistics reporting to e2fsck. 2451 2452Update Vietnamese, Polish, French, Spanish, German, Catalan, Dutch, 2453Czech translations. 2454 2455Fixed various Debian packaging issues --- see debian/changelog for 2456details. (Addresses Debian Bugs: #487443, #487675, #490003) 2457 2458Fixed spelling mistakes, typos, and otherwise clarified man pages. 2459(Addresses Debian Bugs: #393313, #487849, #440983, #440981) 2460 2461Programmer's Notes 2462------------------ 2463 2464Factor out bitmap code in preparation for adding 64-bit new-style 2465bitmaps. 2466 2467Fix gcc -Wall warnings 2468 2469Fix the pkg-config files so that private librares are specified in 2470"Libs.private:". 2471 2472Fix the libext2fs.texinfo manual so it builds with modern versions of 2473texinfo. 2474 2475Silence the makefile from showing the awk command used to build the 2476 2477Clean up the badblocks group so to make it more portable and robust. 2478 2479Avoid using predictable filenames in /tmp in blkid's regression test 2480suite. Also remove bashism's in the regression test script. 2481 2482If the configure script is given --with-diet-libc, don't use thread 2483local storage, since diet libc doesn't support TLS. (Addresses 2484Sourceforge Bug: #2000654) 2485 2486Fix the blkid regression test suite to tolerate older versions of 2487mkswap that don't support the -U option. 2488 2489A few library routines have been converted to support 64-bit block 2490numbers; in particular, the I/O manager functions, the test_io, 2491inode_io, and unix_io managers have all be converted to support 64-bit 2492operation. 2493 2494Debugfs can now be extended for use by test programs. See 2495lib/ext2fs/extents.c for an example for how it can be used. The test 2496program links against the debugfs object files, and provides 2497additional commands by defining an auxiliary libss command table. 2498 2499The lazy_bg filesystem feature, which was only used by developer's 2500testing, has been removed since it has been largely supplanted by 2501uninit_bg. This also simplifies the code. 2502 2503 2504E2fsprogs 1.40.11 (June 17, 2008) 2505================================= 2506 2507Mke2fs, tune2fs, and resize2fs now use floating point to calculate the 2508percentage of reserved blocks. (Addresses Debian Bug: #452639) 2509 2510Updated Spanish and Catalan translations. 2511 2512Fixed various Debian packaging issues --- see debian/changelog for 2513details. (Addresses Debian Bugs: #483962, #483023) 2514 2515Add detection for ZFS volumes to the libblkid library. 2516 2517Fixed spelling mistakes, typos, and otherwise clarified man pages. 2518(Addresses Debian Bug: #486463) 2519 2520Programmer's notes: 2521------------------- 2522 2523Fix marginal C code in probe_lvm2() function to the blkid library more 2524portable for older compilers. 2525 2526Fix build problems on MacOS X. (Addresses Sourceforge Bug: #1972473) 2527 2528Fix ext2fs_swap{16,32,64} functions so they can be used by external 2529applications on big-endian machines. (Addresses Debian Bug: #484879) 2530 2531 2532E2fsprogs 1.40.10 (May 21, 2008) 2533================================ 2534 2535When deciding whether or not to revalidate a blkid cache entry, if the 2536device's mtime is newer than the last time the cached entry was 2537validated, force a revalidation. 2538 2539Fix a potential data corruption bug in e2fsck in the journal replay. 2540The chances of this is happening is extremely remote, especially the 2541default data=ordered or data=writeback modes. However, if a block 2542which has been journalled starts with the first four bytes 0xc03b3998, 2543when e2fsck replays the journal, those four bytes will be replaced 2544with zero's. Fortunately, it is highly, highly unlikely for e2fsck 2545metadata to begin with those fatal 4 byte sequence, and unless 2546data=ordered mode is in use, data blocks are never journaled. 2547 2548Updated German, Dutch, Sweedish, and Vietnamese translations. 2549 2550Programmer's notes: 2551------------------- 2552 2553Fixed various Debian packaging issues --- see debian/changelog for 2554details. 2555 2556Remove default sizes of types when cross compiling, since autoconf 25572.50 can figure this out automatically now. 2558 2559 2560E2fsprogs 1.40.9 (April 27, 2008) 2561================================= 2562 2563SuSE's security team audited uuidd and came up with a few minor 2564issues. None of them are serious given that uuidd runs setuid as a 2565unprivileged user which has no special access other than libuuid 2566directory, but it's good to get them fixed. 2567 2568One additional fix in ext2fs_swap_inode_full() needed for resize2fs to 2569work correctly with in-inode extended attributes. 2570 2571Updated German, Czech, Dutch, French, Polish, Sweedish, and Vietnamese 2572translations. 2573 2574Debugfs will avoid using a pager if the standard output is not a tty. 2575 2576Fix debugfs and tune2fs to correctly handle daylight savings time when 2577parsing a time string. 2578 2579Fixed spelling mistakes, typos, and otherwise clarified man pages. 2580 2581Fix fsck completion bars when multiple filesystems were being checked 2582in parallel. (Addresses Debian Bug: #432865, Addresses Launchpad Bug: 2583#203323, Addresses Sourceforge Bug: #1926023) 2584 2585Fix fsck so that progress information is sent back correctly when 2586multiple filesystems are being check and the output of fsck is being 2587redirected to a file descriptor. Also, include the device name (w/o 2588spaces) in the progress information sent back via a file descriptor. 2589(Addresses Launchpad Bug: #203323, Addresses Sourceforge Bug: 2590#1926023) 2591 2592Teach fsck to treat "ext4" and "ext4dev" as ext* filesystems. 2593 2594If logsave receives a SIGTERM or SIGINT signal, it will now pass that 2595signal to its child process. 2596 2597Fix mke2fs's creation of are resize inode when there is a non-standard 2598s_first_data_block setting. 2599 2600Fix bug in blkid when run by an unprivileged user; most devices were 2601not reported correctly. 9Addresses Launchpad Bug: #220275) 2602 2603Mke2fs will not allow the logically incorect combination of 2604resize_inode and meta_bg, which had previously caused mke2fs to create 2605a corrupt fileystem. 2606 2607Fix fsck in German locales so that a 'j' means yes. 2608(Addresses Sourceforge Bug: #1947683) 2609 2610 2611Programmer's notes: 2612------------------- 2613 2614Fixed various Debian packaging issues --- see debian/changelog for 2615details. 2616 2617Update valgrind options in test_script to work with valgrind 3.2.3 2618 2619Update texinfo.tex to a much newer version from the FSF. 2620 2621Remove bashism for configure script and from the lib/ss Makefile. 2622Addresses Sourceforge Bug: 1921969 2623 2624Fix some silently broken tests: m_no_opt, m_meta_bg, and m_raid_opt. 2625 2626Fix build system so that if texinfo is not installed, it won't print a 2627(harmless) error message. 2628 2629 2630E2fsprogs 1.40.8 (March 13, 2008) 2631================================= 2632 2633Fixed e2image -I so it works on image files which are larger than 2GB. 2634 2635Fixed e2fsck's handling of directory inodes with a corrupt size field. If 2636the size is larger than the number of blocks found in the inode, don't 2637try to allocate extra empty blocks at the end of the directory to make 2638up the difference; there's no point to doing that. In addition, if 2639the size is not a multiple of a blocksize, always fix it. 2640 2641E2fsck handled a pass 2 "should never happen error" by not giving 2642enough information and then core dumping. Unfortunately, it was all 2643too easy to trigger the "should never happen" situation if a 2644directory's inode size was not correct. This has been fixed, but 2645e2fsck has also been taught how to handle this situation more 2646gracefully, by simply removing the inode hash tree information, so 2647that it can be rebuilt again after e2fsck's pass 3. (Addresses 2648Launchpad Bug: #129395) 2649 2650Resize2fs had a bug resizing large inodes with extended attributes 2651that was fixed in 1.40.6; unfortunately, it turned out it wasn't fixed 2652completely on big-endian systems such as PowerPC. The bug should be 2653completely fixed now. Yay for regression test suites. (Addresses Red 2654Hat Bugzilla: #434893) 2655 2656Updated German, Czech, Dutch, Polish, Sweedish, and Vietnamese 2657translations. Many thanks to Philipp Thomas from Novell for stepping 2658up to become the new German translation maintainer! (Addresses Debian 2659Bugs: #302512, #370247, #401092, #412882). 2660 2661When e2fsck is clearing a corrupt inode's HTREE directory information, 2662make it clear that it is just clearing the HTREE information, not the 2663entire inode. 2664 2665Fixed spelling mistakes, typos, and otherwise clarified man pages. 2666 2667Programmer's notes 2668------------------ 2669 2670Add new functions, ext2fs_dblist_get_last() and 2671ext2fs_dblist_drop_last(), which allows the caller to examine the last 2672directory block entry added to the list, and to drop if it necessary. 2673 2674Fixed a portability problem in libblkid with DJGPP. 2675 2676Fix an obvious typo in an "internal error" message in e2fsck. Thanks 2677to Philipp Thomas for pointing this out. 2678 2679If the info files are not built, change "make install" so it doesn't 2680fail with an error code. 2681 2682 2683E2fsprogs 1.40.7 (February 28, 2008) 2684==================================== 2685 2686Remove support for clearing the SPARSE_SUPER feature from tune2fs, and 2687depreciate the -s option, since it can result in filesystems which 2688e2fsck can't fix easily. There are very good reasons for wanting to 2689disable sparse_super; users who wants to turn off sparse_super can use 2690debugfs. (Addresses Sourceforge Bug: #1840286) 2691 2692Add missing options to mke2fs's usage message. (Addresses Sourceforge 2693Bug: #1751393) 2694 2695Fix bug in resize2fs when large (greater than 128 byte) inodes are 2696moved when a filesystem is shrunk; it was only moving the first 128 2697bytes, so extended attributes were not getting moved. (Addresses Red 2698Hat Bugzilla: #434893) 2699 2700E2fsck now prints an explicit message when the bad block inode is 2701updated, to avoid confusion about why the filesystem was modified. 2702(Addresses Sourceforge Bug: #756460) 2703 2704Allow mke2fs and tune2fs manipulate the large_file feature. 2705Previously we just let the kernel and e2fsck do this automatically, 2706but e2fsck will no longer automatically clear the large_file feature. 2707It still isn't really necessary to worry about this feature flag 2708explicitly, but some users seem to care. (Addresses Red Hat Bugzilla: 2709#258381) 2710 2711Suppress message about an old-style fstab if the fstab file is empty. 2712(Addresses Debian Bug: #468176) 2713 2714Fix (really minor) bug in debugfs's find_free_block so it avoids 2715reporting a free block more than once if there are too few free blocks 2716in the filesystem. (Addresses Sourceforge Bug: #1096315) 2717 2718Change e2fsck to no longer clear the LARGE_FILES feature flag 2719automatically, when there are no more > 2GB files in the filesystem. 2720It's been almost a decade since there have been kernels that don't 2721support this flag, and e2fsck clears it quietly without telling the 2722user why the filesystem has been changed. 2723 2724Fix bug which could cause libblkid to seg fault if a device mapper 2725volume disappears while it is being probed. (Addresses RedHat 2726Bugzilla: #433857) 2727 2728Inhance e2fsck's reporting of unsupported filesystem feature flags. 2729(Addresses Sourceforge Feature Request: #1175808) 2730 2731Fix option syntax in dumpe2fs for explicit superblock and blocksize 2732parameters. What was currently documented in the man page has been 2733broken for some time, due to getopt() implementation changes. The 2734option sytax has been changed to one which is can be more portable 2735supported and which is consistent with the format for extended options 2736in mke2fs and tune2fs. (Addresses Sourceforge Bug: #1830994) 2737 2738Add support to tune2fs to clear the resize_inode feature. This 2739requires an fsck afterwards. (Addresses Red Hat Bugzilla: #167816) 2740 2741Teach blkid to detect LVM2 physical volumes. (Addresses Red Hat 2742Bugzilla: #409321) 2743 2744Add support for setting RAID stride and stripe-width via mke2fs and 2745tune2fs. Teach dumpe2fs to print the RAID parameters. 2746 2747Add support for setting new superblock fields to debugfs's 2748set_super_value. 2749 2750Add support for printing "mostly-printable" extended attributes in 2751Debugfs. 2752 2753Add support for the -M option to fsck, which causes it to ignore 2754mounted filesystem. 2755 2756Fix uuidd so that it creates the pid file with the correct pid number. 2757(Addresses Sourceforge Bug: #1893244) 2758 2759Fix various gcc -Wall warnings. 2760 2761Update Czech, Dutch, Polish, Sweedish, and Vietnamese translations 2762 2763Fixed spelling mistakes, typos, and otherwise clarified man pages. 2764(Addresses Sourceforge Patch: #1399325) 2765 2766 2767Programmer's notes: 2768------------------- 2769 2770Fixed various Debian packaging issues --- see debian/changelog for 2771details. (Addresses Debian Bug: #466929) 2772 2773Add new flag EXT2_FLAG_NONFREE_ON_ERROR ext2fs_open2() which returns a 2774partially filled-in filesystem object on an error, so that e2fsck can 2775print more intelligent error messages. 2776 2777Add a new function e2p_edit_feature2() which allows the caller to 2778specify which feature flags are OK to set or clear, and which returns 2779more specific information about feature flags which are not allowed to 2780be set/cleared. 2781 2782Set the C locale in the test_script driver since it uses [A-Za-z]. 2783(Addresses Sourceforge Bug: #1890526) 2784 2785Use fcntl locking instead of lockf in libuuid since Cygwin doesn't 2786support lockf(). 2787 2788Change configure.in to avoid using the 'dc' command unless it is 2789absolutely needed. (i.e., when using parsing a WIP-style version 2790number) (Addresses Sourceforge Bug: #1893024) 2791 2792Add portability checks to support compilation under DJGPP. 2793 2794Update to the latest samba tdb code before the LGPLv3 change, which 2795fixes a realloc() leak on failure. 2796 2797Fix memory leak in ext2fs_alloc_block(). 2798 2799Fix makefile dependency issues for various install targets. 2800(Addresses-Sourceforge-Patches: #1903484, #1903466, #1903456) 2801 2802Improve descriptions for the r_move_itable and r_resize_inode tests. 2803 2804 2805E2fsprogs 1.40.6 (February 9, 2008) 2806=================================== 2807 2808Add support for returning labels for UDF filesystems in the blkid 2809library. 2810 2811Fix bug in the blkid library where cached filesystems was not being 2812flushed when opening USB devices returned the error ENOMEDIUM. 2813(Addresses Debian Bug: #463787) 2814 2815Added logic to the blkid library to automatically choose whether a 2816filesystem should be mounted as ext4 or ext4dev, as appropriate. 2817 2818Allow tune2fs to set and clear the test_fs flag on ext4 filesystems. 2819 2820Fix a bug in e2fsck which caused it to core dump if e2fsprogs had been 2821configured with --enable-jbd-debug. 2822 2823Document the BLKID_FILE environment variable in the libblkid man page 2824 2825Programmer's Notes: 2826------------------- 2827 2828Update e2fsprogs translation template and Vietnamese and Czech translations 2829 2830Fixed various Debian packaging issues --- see debian/changelog for 2831details. (Addresses Debian Bugs: #436058) 2832 2833Don't try to create $DESTDIR/etc/init.d as part of make install as we 2834don't install the init.d script (and it's not the recommended way to 2835start uuidd anyway). (Addresses Sourceforge Bug: #1885085) 2836 2837Use thread local storage to fix a theoretical race condition if two 2838threads try to format an unknown error code at the same time in the 2839com_err library. 2840 2841 2842E2fsprogs 1.40.5 (January 27, 2008) 2843=================================== 2844 2845Fix a potential overflow big in e2image if the device name is too 2846long. 2847 2848Mke2fs will now create new filesystems with 256 byte inodes and the 2849ext_attr feature flag by default. This allows for much better future 2850compatibity with ext4 and speeds up extended attributes even on ext3 2851filesystems. 2852 2853Teach e2fsck to ignore certain "safe" filesystem features which are 2854set automatically by the kernel. Having the kernel set features 2855automagically behind the user's back is a bad idea, and we should try 2856to break the kernel of this habit, especially for the newer ext4 2857feature flags. But for now, we will try to avoid needless full checks 2858which can annoy users who are doing fresh installs. 2859 2860Add support in tune2fs and mke2fs for making a filesystem as being "ok 2861to be used with test kernel code". This will be needed for using test 2862filesystems with the latest ext4 kernel code. 2863 2864Change e2fsck -fD so that it sorts non-htree directories by inode 2865numbers instead of by name, since that optimizes performances much 2866more significantly. (Addresses-Sourceforge-Feature-Request: #532439) 2867 2868If e2image fills the disk, fix it so it exits right away instead of 2869spewing large numbers of error messages. 2870(Addresses-Sourceforge-Feature-Request: #606508) 2871 2872If ftruncate64() is not available for resize2fs, let it use ftrucate() 2873instead, but have it check to see if the size would get truncated, and 2874skip calling ftruncate in that case. 2875 2876Add support for detecting HFS+ filesystems in the blkid library. 2877 2878Add supprt in the blkid library for ext4/ext4dev filesystems. 2879 2880Fix a bug in blkid where it could die on a floating point exception 2881when presented with a corrupt reiserfs image. 2882 2883Fix blkid's handling of ntfs UUID's so that leading zeros are printed 2884such that UUID string is a fixed length. 2885 2886Add sample python bindings for the uuid library to the contrib 2887directory. (Addresses-Sourceforge-Patches: #778817) 2888 2889Fix debugfs's 'lsdel' command so it uses ext2fs_block_iterate2 so it 2890will work with large files. (Addresses Sourceforge Feature Request: 2891#1257500 and Sourceforge Support Request: #1253511) 2892 2893Allow the debugfs 'undel' command to undelete an inode without linking 2894it to a specific destination directory, since this might require 2895allocating blocks that could overwrite some yet-to-be-recovered 2896deleted files. (Addresses-Sourceforge-Feature-Request: #967141) 2897 2898Update Swedish translations from the Translation Project. 2899 2900Programmer's Notes: 2901------------------- 2902 2903Fix configure handling of --sbindir (which should rarely be used, but 2904someone did complain, so let's fix it). (Addresses Sourceforge Bug: 2905#498381) 2906 2907Updated e2fsprogs.spec file to include a new uuidd package 2908 2909Use pkg-config to determine where to find the devmapper library so we 2910can find out where it is located on different distributions. 2911 2912Fix Makefile race so that "make -j3 distclean" works correctly 2913 2914Fix portability problems on non-Linux/non-Hurd/non-Masix systems, 2915especially on MacOS X systems. (Addresses Sourceforge Bugs: #1861633, 2916#1819034, #1863819) 2917 2918Fixed spelling mistakes, typos, and otherwise clarified man pages. 2919 2920Fixed various Debian packaging issues --- see debian/changelog for 2921details. (Addresses Debian Bugs: #459403, #459475, #459614) 2922 2923Remove the --enable-dynamic-static configure option, and build e2fsck 2924dynamically by default. If the user wants e2fsck.static, he/she will 2925need to build it via "make -C e2fsck e2fsck.static" 2926 2927Fix various build warnings due to missing prototypes. 2928(Addresses Sourceforge Patch: #1861663, #1861659) 2929 2930 2931E2fsprogs 1.40.4 (December 31, 2007) 2932==================================== 2933 2934Improve time-based UUID generation. A new daemon uuidd, is started 2935automatically by libuuid if necessary. This daemon is setuid to allow 2936updates to /var/lib/libuuid, so the clock sequence number can be 2937stored and so if the clock is set backwards, it can be detected. 2938(Addresses Sourceforge Bug: #1529672, Addresses Red Hat Bugzilla: 2939#233471) 2940 2941Filter out the NEEDS_RECOVERY feature flag when writing out the backup 2942superblocks. This avoids e2fsck from concluding that a full 2943filesystem check is required before backing up the superblock due to 2944changes in the feature flags. (Addresses Debian Bug: #454926) 2945 2946Fix fsck to only treat the '#' character as a comment at the beginning 2947of the line in /etc/fstab. Otherwise fstabs for the fuse filesystem 2948will cause fsck to issue an bogus warning message. 2949(Addresses Gentoo bug: #195405, Addresses Sourceforge bug: #1826147) 2950 2951Format control characters and characters with the high eighth bit set 2952when printing the contents of the blkid cache, to prevent filesystems 2953with garbage labels from sending escape sequences to the user's screen 2954that might, for example place it in graphics mode. (Addresses Ubuntu 2955Bug: #78087) 2956 2957Fix sign-extension problem on 64-bit systems in in the com_err 2958library. (Addresses Sourceforge Bug: #1809658) 2959 2960Avoid division by zero error when probing an invalid FAT filesystem in 2961the blkid library. (Addresses Sourceforge Bug: #1831627) 2962 2963Update Dutch, Polish, and Vietnamese translations from the Translation 2964Project. Remove the Rwandan translation upon advice of the 2965Translation Project. 2966 2967Programmer's Notes: 2968------------------- 2969 2970Fix the libss "make check" regression test so that it works if the 2971current directory is not in the user's path or if the libss shared 2972library is not installed. (Addresses Sourceforge Bug: #1848974) 2973 2974Fixed spelling mistakes, typos, and otherwise clarified man pages. 2975(Addresses Debian Bugs: #444883, #441872) 2976 2977Fixed various Debian packaging issues --- see debian/changelog for 2978details. (Addresses Debian Bugs: #437720, #451172, #458017) 2979 2980Fix build failure on non-Linux/non-Hurd/non-Masix systems. 2981(Addresses Sourceforge Bug: #1859778) 2982 2983Fix Hurd portability issues. (Addresses Debian Bug: #437720) 2984 2985 2986E2fsprogs 1.40.3 (December 5, 2007) 2987=================================== 2988 2989Fix a potential security vulnerability where an untrusted filesystem 2990can be corrupted in such a way that a program using libext2fs will 2991allocate a buffer which is far too small. This can lead to either a 2992crash or potentially a heap-based buffer overflow crash. No known 2993exploits exist, but main concern is where an untrusted user who 2994possesses privileged access in a guest Xen environment could corrupt a 2995filesystem which is then accessed by the pygrub program, running as 2996root in the dom0 host environment, thus allowing the untrusted user to 2997gain privileged access in the host OS. Thanks to the McAfee AVERT 2998Research group for reporting this issue. (Addresses CVE-2007-5497.) 2999 3000Fix hueristics in blkid which could cause a disk without partitions to 3001be incorrectly skipped when a loopback device is present. (Addresses 3002Red Hat Bugzilla #400321.) 3003 3004Fix e2image so that in raw mode it does not create an image file which 3005is one byte too large. 3006 3007Change mke2fs's usage message so it recommends the preferred -E option 3008instead of the deprecated -R option. 3009 3010Enhance the blkid library so it will recognize squashfs filesystems. 3011(Addresses Red Hat Bugzilla #305151.) 3012 3013Enhance e2fsck so it will force the backup superblocks to be backed up 3014if the filesystem is consistent and key constants have been changed 3015(i.e., by an on-line resize) or by e2fsck in the course of its 3016operations. 3017 3018Enhance blkid's detection of FAT filesystems; so that USB disks with 3019only a single bootable partition will not get missed. 3020 3021E2fsck will no longer mark a filesystem as invalid if it has time 3022errors (i.e., if superblock mount time or last write time is in the 3023future) and the user refuses to fix the problem. 3024 3025The Ubuntu init scripts don't properly set the system time correctly 3026from hardware clock if the hardware clock is configured to tick local 3027time instead of GMT time. Work around this as best as we can by 3028providing an option, buggy_init_scripts, in /etc/e2fsck.conf which can 3029be set on Ubuntu systems. (Addresses Debian Bug #441093, and Ubuntu 3030Bug #131201.) 3031 3032Fix fsck to ignore /etc/fstab entries for bind mounts. (Addresses Red 3033Hat Bugzilla #151533.) 3034 3035Fix e2fsck so that if the superblock is corrupt, but still looks 3036vaguely like an ext2/3/4 superblock, that it automatically tries to 3037fall back to the backup superblock, instead of failing with a hard 3038error. 3039 3040Make the e2fsprogs program more robust so that they will not crash 3041when opening a corrupt filesystem where s_inode_size is zero. 3042 3043Change e2fsck so it uses sscanf() instead of atoi() so it non-numeric 3044arguments are detected as such and the parse error is reported to the 3045user. (Addresses Debian Bug #435381.) 3046 3047Change e2fsck so it will not complain if a file has blocks reallocated 3048up to the next multiple of a system's page size. 3049 3050Fix bug in ext2fs_check_desc() which will cause e2fsck to complain 3051about (valid) filesystems where the inode table extends to the last 3052block of the block group. (Addresses Red Hat Bugzilla #214765.) 3053 3054Fix a bug in ext2fs_initialize() which causes mke2fs to fail while 3055allocating inode tables for some relatively rare odd disk sizes. 3056(Addresses Red Hat Bugzilla #241767.) 3057 3058Add Catalan translation and update Dutch and Swedish translations 3059from the Translation Project. 3060 3061Fix big-endian byte-swapping bug in ext2fs_swap_inode_full(). We 3062still had an issue when trying to figure out whether we need to 3063byte-swap fast symlinks that contained extended attributes. 3064 3065Fixed spelling mistakes, typos, and otherwise clarified man pages. 3066(Addresses SourceForge Bug #1821333.) 3067 3068 3069Programmer's Notes: 3070------------------- 3071 3072Fix mke2fs tests to avoid needing any significant ^M (CR) characters 3073 3074Add "make check" to the RPM spec file 3075 3076Fix "make install" and 'make unstall" in misc/Makefile.in so that it 3077works correctly when the prefix is not the root directory. 3078 3079Fix the resize2fs tests, r_move_itable and r_resize_inode, so they 3080clena up after themselves by deleting the test.img temporary file 3081after completing the test. 3082 3083Fixed a corner case bug ext2fs_unlink() when trying to delete the 3084first directory entry in a directory block and the last directory 3085entry in the previous directory block is not in use. Fortunately 3086ext2fs_unlink() is only used by debugfs and e2fsck, and in e2fsck in 3087places where it is extremely unlikely to run into this corner case. 3088 3089Fix missing dependency which would cuase parallel builds to fail. 3090(Addresses Sourceforge Bug #1842331.) 3091 3092Fix a build error on newer gcc caused by lib/ext2fs/ismounted.c 3093calling open(O_CREATE) with a missing mode parameter. 3094 3095Fix the test_ss.c in lib/ss so it can be used as an example 3096application program for the library as well as a regression test 3097suite. 3098 3099Fix ext2fs_dblist_dir_iterate() so that error codes and abort codes 3100are properly passed back up through the call stack. 3101 3102E2fsprogs 1.40.2 (July 12, 2007) 3103================================ 3104 3105A recent change to e2fsck_add_dir_info() to use tdb files to check 3106filesystems with a very large number of filesystems had a typo which 3107caused us to resize the wrong data structure. This would cause a 3108array overrun leading to malloc pointer corruptions and segfaults. 3109Since we normally can very accurately predict how big the the dirinfo 3110array needs to be, this bug only got triggered on very badly corrupted 3111filesystems. 3112 3113Fix a bug in e2fsck which caused it to incorrectly salvange 3114directories when the last entry's rec_len is bogusly too big. This 3115resulted in a nonsense filesystem corruption to be reported, and 3116required a second run of e2fsck to fully fix up the directory. 3117 3118Update tune2fs man page to include more discussion of reserved blocks 3119(Addresses Launchpad bug #47817) 3120 3121Update Turkish, Polish, Dutch, and Vietnamese PO files from the 3122Translation Project 3123 3124 3125E2fsprogs 1.40.1 (July 7, 2007) 3126=============================== 3127 3128Fix bug which could cause libblkid to loop forever. When revalidating 3129a partition where there is obsolete information in /etc/blkid.tab, we 3130end up freeing a the type tag without clearing dev->bid_type, causing 3131blkid_verify() to loop forever. (Addresses Debian Bug: #432052) 3132 3133The Turkish translation has a bug in it where it has the translation 3134of "E@e '%Dn' in %p (%i)" to "E@E". This causes @E to be expanded at 3135@E, recursively, forever, until the stack fills up and e2fsck core 3136dumps. We fix this by making e2fsck stop @-expansions after a 3137recursive depth of 10, which is far more than we need. 3138(Addresses Sourceforge Bug: #1646081) 3139 3140Compile the default mke2fs.conf into mke2fs program. People are 3141getting surprised by mke2fs creating filesystems with different 3142defaults than earlier versions of mke2fs if mke2fs.conf is not 3143present. So we now create a built in version of mke2fs.conf file 3144which is used by mke2fs if the /etc/mke2fs.conf is not present. 3145(Addresses SourceforgeBug: #1745818) 3146 3147Improve the config/parse_types.sh helper script. Fix a potential 3148security problem if e2fsprogs is built as root (as Gentoo does!). In 3149addition fix the script and how it is called from the configure script 3150so that it does the right thing when cross-compiling. (Fixes Gentoo 3151bug: #146903) 3152 3153Update Vietnamese, French, and Dutch PO files from the Translation 3154Project. Also created a new e2fsprogs.pot file for translator. 3155 3156Fix bogus strip permission errors when building under Debian. When 3157building the e2fsprogs dpkg's, the dh_strip command emits a large 3158number of error messages caused by the permissions not being right. 3159So run dh_fixperms before running dh_strip. 3160 3161Programmer's Notes: 3162------------------- 3163 3164Add new function: profile_set_default(). This function sets the value 3165of the pseudo file "<default>". If the file "<default>" had 3166previously been passed to profile_init(), then def_string parameter 3167will be parsed and used as the profile information for the "<default>" 3168file. 3169 3170Fix mk_cmds's error reporting so that it is unambiguous that it is the 3171mk_cmds script which is generating the error. (Obviates Gentoo patch: 3172e2fsprogs-1.32-mk_cmds-cosmetic.patch) 3173 3174Fix the test suite to use LC_ALL instead of LANG. LC_ALL is the "high 3175priority" environment variable that overrides all others, where as 3176LANG is the lowest priorty environment variable. If LC_ALL is set, it 3177doesn't matter whether LANG, LANGUAGE, LC_COLLATE, LC_MESSAGES, and 3178the all the rest are set. This will assure that the locale when 3179running the test suites is the "C" locale. (Obviates Gentoo patch: 3180e2fsprogs-1.38-tests-locale.patch) 3181 3182 3183E2fsprogs 1.40 (June 29, 2007) 3184============================== 3185 3186Fix divide by zero error in blkid's NTFS probing logic. 3187 3188Add new blkid -g option which causes the blkid cache to be garbage 3189collected. 3190 3191Fix a bug in libblkid which could cause the internal field bid_type to 3192become corrupted. Fortunately bid_type isn't used much, and bid_label 3193and bid_uuid is only used by debugging code, so the impact of this bug 3194was very minor. 3195 3196Mke2fs will now store the RAID stride value when a filesystem is 3197created with a requested RAID stride, and then use it automatically in 3198resize2fs. 3199 3200Mke2fs has a sanity check added to make sure (inode_size * num_inodes) 3201isn't too big. In some cases Lustre users have tried specifying an 3202inode size of 4096 bytes, while keeping an inode ratio of one inode 3203per 4096 bytes. 3204 3205Improve sanity check in e2fsck's algorithm for finding a backup 3206superblock, so that it won't accidentally find a superblock that was 3207located in the journal, and then later reject it as being not a valid 3208backup superblock. 3209 3210Fix e2fsck get_size logic so that it will work with the Linux floppy 3211driver. The Linux floppy driver is a bit different from the other 3212block device drivers, in that if the device has been opened with 3213O_EXCL, it disallows another open(), even if the second open() does 3214not have the O_EXCL flag. (Addresses Debian Bug: #410569) 3215 3216Fix error checking of badblock's last-block and start-block arguments. 3217(Addresses Debian Bug: #416477) 3218 3219Fix e2fsck so that it doesn't overwrite the backup superblocks when 3220recovering a journal until the master superblock has been confirmed as 3221being sane. 3222 3223Change the blkid library to be much more paranoid about concluding 3224that a partition contains an NTFS filesystem, and fetch the UUID and 3225LABEL information from NTFS filesystems. (Addresses Launchpad Bug: 3226#110138) 3227 3228Factor out the code which sets the default journal size and move it 3229into libext2fs. 3230 3231Enhance e2fsck so it will recreate the ext3 journal if the original 3232journal inode was cleared to due it being corrupt after finishing the 3233filesystem check. 3234 3235Fix e2fsck so that it updates the journal inode if it is corrupted and 3236the backup journal information from the superblock was successfully 3237used to recover the filesystem. 3238 3239Fix e2fsck so that it checks all of the blocks in the journal inode 3240for validity. The original code only checked the direct blocks to 3241make sure the journal inode was sane. Unfortunately, if some or all 3242of the indirect or doubly indirect blocks were corrupted, this would 3243not be caught. 3244 3245Add support in blkid to detect LUKS encrypted partitions. 3246 3247Add extra sanity checks for extended attributes in the case where the 3248size is zero but the offset is very large. 3249 3250Fix byte-swapping issues for large inodes in ext2fs_read_inode_full() 3251and ext2fs_get_next_inode_full(). 3252 3253Clarify the copyright licenses used by the various libraries in 3254the top-level COPYING file (Red Hat Bugzilla: 166058) 3255 3256Make mke2fs's defaults when /etc/mke2fs.conf doesn't exist more sane. 3257 3258Fix mke2fs and debugfs to support large (> 16 bit) uid's and gid's. 3259 3260Remove check in e2fsck which requires EA's in inodes to be sorted; 3261they don't need to be sorted, and e2fsck was previously wrongly 3262clearing unsorted EA's stored in the inode structure. 3263 3264Allow mke2fs or tune2fs to create a substantially larger journal (up 3265to 10,240,000 blocks). 3266 3267Fix MD superblock detection, and make sure the correct UUID is 3268reported from the MD superblock. 3269 3270Fix a signed vs. unsigned bug in debugfs. 3271 3272Enhance debugfs's date parser so that it accepts integer values. 3273 3274Fix e2fsck's pass1c accounting so it doesn't terminate too early if a 3275file with multiply claimed blocks is hard linked. or not at all if the 3276root directory contains shared blocks 3277 3278Enhance debugfs so it can modify the block group descriptors using the 3279command set_block_group_descriptor. 3280 3281Improve e2fsck's reporting of I/O errors so it's clearer what it was 3282trying to do when an error happens 3283 3284Fix a bug in in how e2fsprogs byte swaps inodes containing fast 3285symlinks that have extended attributes. (Addresses Red Hat Bugzilla: 3286#232663 and LTC Bugzilla: #27634) 3287 3288Fix potential file descriptor leak in ext2fs_get_device_size() in an 3289error case. 3290 3291Add libreadline.so.5 support to libss. 3292 3293Impove badblocks -n/-w exclusive usage message. 3294 3295Fix dump_unused segfault in debugfs when a filesystem is not open 3296 3297Fix memory leak in blkid library. (Addresses Debian Bug: #413661) 3298 3299Allow the debugfs lcd command to work w/o a filesystem being open. 3300(Addresses LTC Bugzilla #27513) 3301 3302Fix e2fsck to clear i_size for special devices with a bogus i_blocks 3303field on the first pass. 3304 3305Fix e2fsck to set the file type of the '..' entry when connecting 3306a directory to lost+found. (Addresses Lustre Bug: #11645) 3307 3308Enhance e2fsck to recover directories whose modes field were corrupted 3309to look like special files. This is probably only useful in 3310artificial test cases, but it will be useful if we ever do the "inodes 3311in directory" idea for ext4. 3312 3313Allow debugfs to dump (and rdump) > 2GB files. (Addresses Debian Bug: 3314#412614) 3315 3316Fix resize2fs parsing of size parameter (in sector units). This was 3317actually a bug in libe2p's parse_num_blocks() function. When handling 3318the 's' suffix, it was ignoring the blocksize information passed in 3319from the caller and always interpreting the number in terms of a 1k 3320blocksize. (Addresses Debian Bug: #408298) 3321 3322There was a floating point precision error which could cause e2fsck to 3323loop forever on really big filesystems with a large inode count. 3324(Addresses Debian Bug: #411838) 3325 3326Fix memory leak in ext2fs_write_new_inode() 3327 3328Add support for using a scratch files directory to reduce e2fsck's 3329memory utilization on really big filesystems. This uses the TDB 3330library. See the [scratch_files] section of the e2fsck.conf man page 3331for more details. 3332 3333Fixed type-punning bug which caused dumpe2fs to crash on the Arm 3334platform (Addresses Debian Bug: #397044) 3335 3336Add explanatory message to badblocks that -n and -w are mutually exclusive 3337(Addresses Debian Bug: #371869) 3338 3339Allow debugfs and dumpe2fs to support fs features under development. 3340 3341Add support for the new flag EXT2_FLAG_SOFTSUPP_FEATURES flag to 3342ext2fs_open() , which allows application to open filesystes with features 3343which are currently only partially supported by e2fsprogs. 3344 3345Allow unix_io to support offsets greater than 2G (Addresses 3346SourceForge Bug: #1547922) 3347 3348Fixed overflow and signed/unsigned problems caused by the number of 3349blocks or inodes exceeding 2**31 or being close to 2**32-1. 3350 3351Add support for unsigned directory hash calculations with hints in the 3352superblock to fix cross-architectural portability for htree 3353directories with filenames where the high 8th bit is set. (Addresses 3354Debian: #389772) 3355 3356Fix resize2fs so that it gives user-intelligible error messages if the 3357filesystem or the kernel does not support on-line resizing. 3358(Addresses Debian Bug: #380548) 3359 3360Require mke2fs -F -F for really dangerous operations, since -F is 3361needed for less dangerous operations such as creating filesystems 3362images in regular files, or creating filesystems on whole block 3363devices. These relatively innocuous usages should NOT be confused 3364with running mke2fs on an apparently-mounted or in-use filesystem. 3365 3366Allow the default inode size to be specified into the mke2fs.conf 3367file. 3368 3369Make the smallest default journal size is big enough so that on-line 3370resizing should always work. 3371 3372Fix silly spelling error in e2fsck. (Addresses SourceForge bug: 3373#1531372) 3374 3375Fix debugfs coredump when lsdel is run without an open filesystem 3376(Addresses Debian Bug: #378335) 3377 3378Fix debugfs display bug us that bytes that have the high bit set are 3379displayed as "ec" instead of "ffffffec". 3380 3381Add support in lsattr so it will display the EXT4_EXTENTS_FL flag. 3382 3383Device mapper scanning wasn't working in the blkid library because the 3384pathnames had an extra "/dev" when they were being probed. 3385 3386Add GFS/GFS2 support to the blkid library. 3387 3388Fix blkid support of empty FAT filesystem labels. 3389 3390Avoid recursing forever (or for a long time) when the blkid library 3391searches for a device and there are symlinks to directories in /dev. 3392 3393Avoid unaligned halfword access in blkid when accessing FAT 3394superblocks, as this will cause Sparc/Solaris systems to throw a 3395SIGBUS error. 3396 3397The latest devmapper libraries requires pthreads, add -lpthreads to 3398the static link libraries for e2fsck.static if devmapper is enabled. 3399(Addresses Debian bug: #388718) 3400 3401Improve the (non-installed, for experts only) findsuper program by 3402printing the uuid and label from the superblocks, as well as the 3403starting and ending offsets of the filesystem given the information in 3404the superblock. Omit by default printing superblocks that are likely 3405found in located in an ext3 journal unless an explicit -j option is 3406given. 3407 3408Updated Spanish, French and Dutch translations and added Catalan 3409translation. (Addresses Debian bug: #411562) 3410 3411Use FreeBSD's DIOCGMEDIASIZE and DIOCGDINFO ioctls if available when 3412determining a partition's size, since binary searching to determine 3413the device doesn't work on FreeBSD. 3414 3415Documentation about UUID's is available in enough places, and it's 3416awkward to deal with debian-legal's insanities. So I'm caving in the 3417"more-lunatic-than-RMS" wing of Debian by removing RFC-4122 so we 3418don't have do the dfsg tarball. Also remove the rule that only tried 3419to install RFC-4122 on Ubuntu, since Ubuntu seems to want to fetch 3420e2fsprogs exclusively from Debian. (Addresses Debian Bug: #407107) 3421 3422Fix the info-dir line so that the menu name does not contain a .info 3423prefix. First of all, it's ugly, secondly, it causes the install-info 3424command to fail to remove the com_err info file from the 3425/usr/share/info/dir file when the comerr-dev package is removed and 3426purged. (Addresses Debian Bug: #401711) 3427 3428Fixed spelling mistakes, typos, and otherwise clarified man pages. 3429(Addresses Debian Bug: #369761, #373004, #379695) 3430 3431Fixed various Debian packaging issues --- see debian/changelog for 3432details. (Addresses Debian Bugs #389554, #390664, #413208, #419605, 3433#408352, #415560, #399155) 3434 3435 3436Programmer's notes: 3437------------------- 3438 3439E2fsck now supports the %It expansion when printing a problem report. 3440It will print the type of the inode in the problem context. 3441 3442Fix misc/Makefile.in so that it builds even if e2fsck hasn't been built yet 3443(Addresses Sourceforge Bug: #1565561) 3444 3445Remove unused variables and other lint/gcc -Wall cleanups 3446 3447Add check to ext2fs_get_device_size() so it will return EFBIG for for 3448filesystems contained in regular files where the filesystem image size 3449is returned by stat64(). 3450 3451Set local environment variables to C so mk_cmds and compile_et always 3452work. (Addresses SourceForge Bug: #1532177) 3453 3454Added the 64-bit byte swapping function ext2fs_swab64(). 3455 3456Added two new helper functions to prevent 2**31/2**32-1 overflow 3457problems: ext2fs_div_ceil() and e2p_percent(). 3458 3459Create new ext2fs library inline functions ext2fs_group_first_block() 3460and ext2fs_group_last_block() in order to calculate the starting and 3461ending blocks in a block group. 3462 3463Create the generated files read-only to remind developers not to edit them. 3464 3465Add support for autoconf 2.60 (with backwards compatibility for older 3466versions of autoconf). 3467 3468Added an "make rpm" target to top-level Makefile 3469 3470Added various FreeBSD portability fixes. 3471 3472Exclude mercurial files from the RPM build tree to speed up copy/build. 3473 3474Use root_sysconfdir to define the locations of mke2fs.conf and 3475e2fsck.conf instead of using a hard-coded /etc pathname. 3476 3477Prevent e2fsck.h and ext2_ext_attr.h from getting included multiple times. 3478 3479Fixed "make clean" in blkid's Makefile.in file from removing tst_*.c files. 3480 3481If diff -u is supported, use it to report test failures. 3482 3483Updates/improvements to RPM spec file 3484 3485Add on-disk format definitions for the following new features: 3486EXT4_FEATURE_RO_COMPAT_HUGE_FILE, EXT4_FEATURE_RO_COMPAT_GDT_CSUM, 3487EXT4_FEATURE_RO_COMPAT_DIR_NLINK, EXT4_FEATURE_INCOMPAT_64BIT, 3488EXT4_FEATURE_RO_COMPAT_EXTRA_ISIZE 3489 3490Add a new make target "checked-failed" in the tests directory which 3491reruns any failed tests 3492 3493Update draft-leach-uuids-guids-01.txt with rfc4122.txt 3494 3495Fix miscellaneous bugs reported by Coverity: Dead code, potential nul 3496pointer dereferences, memory leaks, etc. None were security-criticial 3497problems. 3498 3499Fix up usage and decrement error messages in the test_icount program 3500 3501Add debugging code to the com_err library; if the environment variable 3502COMERR_DEBUG is set to 1, print out debugging messages as error tables 3503are added and removed from the com_err library. If the 3504COMERR_DEBUG_FILE environment variable is set (and the process is not 3505setuid) the debugging messages may be redirected to a file. 3506 3507Change all of the e2fsprogs programs to use the newer add_error_table() 3508and remove_error_table() interfaces instead of the much older 3509initialize_*_error_table() function. 3510 3511Add TDB support into the ext2fs library. This allows us to have a 3512guaranteed library we can count on always being present so we can 3513store data in an on-disk database. 3514 3515Add support for using TDB to store the icount data, so we don't run out 3516of memory when checking really large filesystems. 3517 3518Change the regression test suite so that it skips empty test directories. 3519 3520Define the l_i_iversion field in ext2_inode. The l_i_version field is 3521now defined from the old l_i_reserved1 field in the ext2 inode. This 3522field will be used to store high 32 bits of the 64-bit inode version 3523number. 3524 3525Add Makefile production rule for e2fsprogs.spec in case it gets 3526deleted. 3527 3528Add new function profile_get_uint() to allow for a clean way to fetch 3529unsigned integers from the context. 3530 3531Add test to make sure the ext2 superblock structure is 1024 bytes. 3532 3533Fix typo in name of f_dup4 regression test 3534 3535Add new function blkid_gc_cache() which performs a garbage collection 3536pass on the /etc/blkid.tab file. 3537 3538The ext2fs_open() function now sets EXT2_FLAG_MASTER_SB_ONLY. In 3539general, only e2fsck (after the filesystem is clean), tune2fs, and 3540resize2fs should change the backup superblocks by default. Most 3541callers of ext2fs_open() should not be touching any superblock fields 3542which require the backups to be touched. 3543 3544Add new function to libext2fs, ext2fs_default_journal_size(), which 3545returns the default journal size. 3546 3547 3548E2fsprogs 1.39 (May 29, 2006) 3549============================= 3550 3551Fix 32-bit cleanliness in e2fsprogs so that we can support filesystems 3552between 2**31 and 2**32 blocks. 3553 3554Change mke2fs to use /etc/mke2fs.conf as a configuration file to 3555configure the filesystem features, blocksize, and inode_ratio for 3556different filesystem types. 3557 3558Mke2fs will now create filesystems hash trees and on-line resizing 3559enabled by default, based on the new /etc/mke2fs.conf file. 3560 3561The e2fsprogs tools (resize2fs, e2fsck, mke2fs) will open the 3562filesystem device node in exclusive mode to prevent accidents by 3563system administrators. In the case of resize2fs and mke2fs, it will 3564only use exclusive mode if the filesystem is not mounted. 3565 3566Fixed a bug in mke2fs which caused it to to fail when creating the 3567resize inode for large filesystems. (Addresses Debian Bug #346580) 3568 3569When allocating space for the RAID filesystems with the stride 3570parameter, mke2fs will now place each portion of the group's inode 3571table right up after the superblock (if present) in order to minimize 3572fragmentation of the freespace. 3573 3574Speed up mke2fs and e2fsck by writing inode and block bitmaps more 3575efficiently by writing the inode and block bitmaps in one pass, thus 3576reducing the number of disk seeks required. 3577 3578Add support for on-line resizing to resize2fs. 3579 3580Fix blkid library so that logic to determine whether or not a device's 3581cached information in /etc/blkid.tab needs to be verified or not 3582doesn't get confused by a system clock which is insane (for example, 3583if the battery is dead on a Macintosh running PPC Linux. (Addresses 3584Red Hat Bug: #182188) 3585 3586The blkid library will now store the UUID of the external journal used 3587by ext3 filesystems, so that in the future, the userspace mount binary 3588can use this to find the location of the external journal and pass 3589this information to the kernel. 3590 3591E2fsck will now consult a configuration file, /etc/e2fsck.conf to 3592control how various options should be handled. See the e2fsck.conf 3593man page for more details. (Addresses Debian Bug: #150295) 3594 3595E2fsck now prints an explanatory message when delaying a filesystem 3596check when the system is running on battery. (Addresses Debian Bug: 3597#350306) 3598 3599E2fsck will detect if the superblock's last mount field or last write 3600field is in the future, and offer to fix if so. (Addresses Debian Bug 3601#327580) These problems will be fixed automatically in preen mode 3602since Debian's boot sequence bogusly doesn't set the time correctly 3603until potentially very late in the bootup process, and this can cause 3604false positives which will cause users' systems to fail to boot. 3605(Addresses Debian Bugs #343662 and #343645) 3606 3607E2fsck now checks to see if the superblock hint for the location of 3608the external journal is incorrect, and if so, offer to update it. 3609(Addresses Debian Bug: #355644) 3610 3611Fix e2fsck from segfaulting on disconnected inodes that contain one or 3612more extended attributes. (Addresses Debian Bug: #316736, #318463) 3613 3614E2fsck will stop and print a warning if the user tries running a 3615read/write badblocks test on a read-only mounted root filesystem. 3616 3617Fix a memory leak in e2fsck's error paths. (Thanks to Michael 3618C. Thompson for pointing these out; they were originally found using 3619Coverity.) 3620 3621When resizing a file containing a filesystem, resize2fs will expand or 3622truncate a file as necessary. (Addresses Debian Bug: #271607) 3623 3624Resize2fs will now automatically determine the RAID stride parameter that 3625had been used to create the filesystem, and use that for newly created 3626block groups. The RAID stride parameter may also be manually specified 3627on the command line using the new -S option to resize2fs. 3628 3629Fix mke2fs so that it correctly creates external journals on 3630big-endian machines (such as a S/390). 3631 3632Fix a bug in the e2p library which could cause dumpe2fs to (rarely) 3633fail to print out the journal or hash seed UUID. (Thanks to Guillaume 3634Chambraud for pointing this out.) 3635 3636Dumpe2fs will now print the size of the journal (if present). 3637 3638Fix debugfs's set_inode_field command so it can properly set the frag, 3639fsize, uid_high, gid_high, and author fields in the inode instead of 3640silently failing, and so that setting the i_size actually sets i_size 3641correctly. 3642 3643Add a new debugfs command, set_current_time, which sets fs->now so 3644that regression test suites can repeatedly modify the filesystem's 3645last_write fields. 3646 3647Fix a bug in debugfs's icheck which would incorrectly report the owner 3648of an extended attribute block. 3649 3650Fix the debugfs commands htree_dump, dx_hash, and list_dir so they print a 3651print a usage message when an illegal option character is given. 3652 3653Fix debugfs's dump_unsued command on filesystems with a 64k blocksize 3654so it won't core dump. (Addresses SourceForge bug #1424311) 3655 3656Fix mklost+found so that it creates a full-sized directory on 3657filesystems with larger block sizes. 3658 3659Fix a file descriptor leak in blkid library. 3660 3661Fix a display bug in "badblocks -sv" so that the done message properly 3662clears the block number at the end of the test. (Addresses Debian Bug 3663#322231) 3664 3665Allow fractional percentages to the -m option in mke2fs and tune2fs 3666(Addresses Debian Bug: #80205) 3667 3668Use fstat/fstat64 in getsize.c if the the target is a regular file, 3669instead of attempting to do a binary search. Fix some fd leaks in 3670error cases. 3671 3672Add support for device mapper library to the blkid library to ensure 3673that the "best" (i.e., leaf) device is probed by the blkid library. 3674 3675Fix the blkid library so that it notices when an ext2 filesystem is 3676upgraded to ext3. 3677 3678Improve the blkid's library VFAT/FAT detection; it now understands 3679labels stored in the root directory, and is more paranoid about 3680checking the FAT superblock values. 3681 3682Fixed a fd leak in the uuid library which was causing problems for the 3683LVM tools. (Addresses Debian Bug: #345832) 3684 3685Add support for the reiser4 and software suspend partitions to the 3686blkid library. Also add support for extract the label from iso9660 3687filesystems. 3688 3689Fix a compile_et bug which miscount the number of error messages if 3690continuations are used in the .et file. 3691 3692Add extra sanity checks to protect users from unusual cirucmstances 3693where /etc/mtab may not be sane, by checking to see if the device is 3694reported busy (works on Linux 2.6) kernels. (Addresses Debian Bug 3695#319002) 3696 3697Updated French, Dutch, Polish, and Swedish translations. (Addresses Debian 3698Bug: #343149, #341911, #300871, #316604, #316782, #330789) 3699 3700Fix use-after-free bug in e2fsck when finishing up the use of the 3701e2fsck context structure. 3702 3703Fixed spelling mistakes, typos, and otherwise clarified man pages and 3704documentation. (Addresses Debian Bugs: #329859, #322188, #316811, 3705#312515, #351268, #357951, #347295, #316040, #368392, #368393, #368394, 3706#368179) 3707 3708Fixed various Debian packaging issues --- see debian/changelog for 3709details. (Addresses Debian Bugs #317862, #320389, #290429, #310950, 3710#310428, #330737, #330736, #329074, #356293, #360046, #366017, #364516, 3711#362544, #362970) 3712 3713 3714Programmer's notes: 3715------------------- 3716 3717Update config.guess and config.sub to latest version (2006-02-23) from 3718FSF. 3719 3720Fix asm_types.h type conflicts on AMD 64 platforms. (Addresses Debian 3721Bugs: #360661, #360317) 3722 3723Fixed the Makefile so that they work correctly on newer versions of 3724GNU make (i.e., 3.81). 3725 3726Add valgrind support to the regression test suites, and eliminate 3727false positives from valgrind. 3728 3729Add a regression test suite for the blkid library. 3730 3731Fix a fencepost error in resize2fs caught by valgrind. 3732 3733Fix compiler warnings about missing memcpy prototypes. 3734 3735We no longer have the sparc assembly code in the header file any more, 3736so we shouldn't set _EXT2_HAVE_HAS_BITOPS_ for the sparc. This would 3737break compiles on the sparc architectures when using gcc. 3738 3739In the libext2fs library, add the new field fs->now which if non-zero 3740is used instead of the system time when setting various filesystem 3741fields (last modified time, last write time, etc.) 3742 3743Fix gcc 4.01 complaints by adding a missing #include <string.h> to 3744ext2fs.h which is needed since the inline functions use memcpy(). 3745(Addresses Sourceforge Bug #1251062) 3746 3747Use BUILD_CFLAGS and BUILD_LDFLAGS instead of CFLAGS and LDFLAGS in 3748the build system when building files in the util directory which are 3749needed during the build process. This avoids potential problems when 3750cross-compiling and some of the options specified in CFLAGS or LDFLAGS 3751are not recognized as valid by the host compiler. (Addresses 3752Sourceforge Bug #1261547) 3753 3754Clean up the blkid library by making the superblock and generic i/o 3755functions to be more generic. Clean up interface to the probe 3756function, and fix memory leak. Finallly, remove an unneeded reference 3757to probe.h in the lib/blkid/resolve.c 3758 3759Add an ext2fs_read_bb_FILE regression test to confirm proper detection 3760of invalid block #'s. 3761 3762The x86 asm constraints for ext2fs_{set/clear}_bit have been fixed to 3763indicate that the the function read/writes the memory location. 3764 3765Fix various gcc -Wall complaints. 3766 3767Add a dependency to make sure that the subdirectories are created 3768before creating all of the object files, in order to address parallel 3769build problem in the library Makefiles. (Addresses Sourceforge Bug: 3770#1261553) 3771 3772Add $(LDFLAGS) to the command line argument when generating an ELF or 3773Solaris shared library, to allow cross-compile and other builds that 3774might need to specify -L paths to needed libraries. (Addresses 3775Sourceforge Bug #1261549) 3776 3777Add a new feature, EXT2_FEATURE_COMPAT_LAZY_BG, which is initially 3778intended for testing purposes. It allows an ext2/ext3 developer to 3779create very large filesystems using sparse files where most of the 3780block groups are not initialized and so do not require much disk 3781space. Eventually it could be used as a way of speeding up mke2fs and 3782e2fsck for large filesystem, but that would be best done by adding an 3783RO_COMPAT extension to the filesystem to allow the inode table to be 3784lazily initialized on a per-block basis, instead of being entirely 3785initialized or entirely unused on a per-blockgroup basis. 3786 3787Fix backwards compatibility so e2fsprogs will better compile on Linux 37882.0.35 systems. 3789 3790Make test scripts more robust against locale-related environment variables 3791 3792Fix type warning problem with sizeof() in ext2fs_open2(). 3793 3794Fix type warning problem with time_t in debugfs. 3795 3796 3797E2fsprogs 1.38 (June 30, 2005) 3798============================== 3799 3800Fix blkid's test programs (built with "make check") compile correctly 3801even without "configure --enable-blkid-debug". 3802 3803Fix ia64 core dump bug caused by e2fsprogs running afoul of C99 strict 3804type aliasing rules on newer gcc compilers. (Addresses Red Hat 3805Bugzilla ##161183.) 3806 3807Fix com_err library to make it more compatible with recent changes 3808made to the com_err library in MIT Kerberos V5 version 1.4. 3809(Addresses Sourcefroge Bug #1150146) 3810 3811General cleanup of messages printed by e2fsprogs programs for grammar, 3812consistency, and to make life easier for translators. Fixed a few 3813strings containing English that had not been marked as needing 3814translations. Removed strings that do not need to be translated, to 3815make life easier for translators. 3816 3817Mke2fs and badblocks will take advantage of a feature in Linux 2.6 to 3818test to see if a device appears to be in use instead of just relying 3819on /proc/mounts and /etc/mtab. (Addresses Debian Bug #308594). 3820 3821Fix portability problem in the filefrag program affecting platforms 3822where the size of an integer is smaller than the size of a long. 3823(Addresses Debian Bug #309655) 3824 3825Mke2fs will now use a larger journal by default for filesystems 3826greater than 4GB. (128 MB instead of 32MB). 3827 3828Mke2fs will refuse to create filesystems greater than 2**31-1 blocks, 3829unless forced. This is to avoid signed vs. unsigned kernel bugs in 3830block numbers that still need to be fixed. 3831 3832The blkid program has a new option which will more efficiently search 3833for device when it is known (or expected) that only one matching 3834device will be found in the system, such as when doing a lookup by 3835UUID. 3836 3837Debian's e2fsprogs-specific initrd fragment will avoid including 3838unnecessary libraries into the initrd ramdisk by unsetting LD_PRELOAD 3839and LD_LIBRARY_PATH, and filtering out libraries found in 3840/etc/ld.so.preload. (Addresses Debian Bug: #304003) 3841 3842Fixed a potential portability issue in the blkid programs for 3843architectures where the char type is unsigned. (Addresses Sourceforge 3844Bug: #1180585) 3845 3846Fix a bug in filefrag so that it doesn't falsely count an extra 3847discontinuity when the first block found is an indirect block. 3848(Addresses Debian Bug #307607). 3849 3850Fix blkid's recognition of cramfs filesystems, and enhance it to be 3851able to handle cramfs labels. 3852 3853Fix debugfs's stat command to not core dump when a filesystem is not 3854open. 3855 3856Fix e2fsck's handling of error conditions caused by the resize inode 3857claiming blocks that are also used by other inodes, a filesystem 3858corruption which was commonly caused by a bug in Fedora Core 3's 3859resize2fs program. 3860 3861Fixed bug in filefrag which caused it to fail on non-ext2/3 3862filesystems. (Addresses Debian Bug: #303509) 3863 3864If the superblock last mount time indicates that the system clock may 3865not be accurate, then e2fsck will omit checking inodes' deletion time 3866field for indications of a potential corrupted orphaned inode list. 3867(Previously e2fsck only ommited these LOW_DTIME checks when the 3868superblock's last write time looked insane.) 3869 3870Fixed a IA64 core dump bug in the e2p library which affected dumpe2fs. 3871(Addresses Debian bug #302200) 3872 3873Make the blkid library more paranoid about being run from setgid 3874programs, and to use __secure_getenv() from libc if it is available. 3875 3876Fixed spelling mistakes, typos, and otherwise clarified man pages. 3877(Addresses Debian Bugs: #304591, #304592, #304594, #304597, #304593 3878and Sourceforge Bug: #1189803) 3879 3880Updated and fixed translations. 3881 3882Fixed various Debian packaging issues --- see debian/changelog for 3883details. 3884 3885Programmer's notes: 3886------------------- 3887 3888Ext2fs_set_bit(), ext2fs_clear_bit(), and ext2fs_test_bit( have been 3889changed to take an unsigned int for the bit number. Negative bit 3890numbers were never allowed (and didn't make any sense), so this should 3891be a safe change. This is needed to allow safe use of block numbers 3892greater than or equal to 2**31. 3893 3894The compile_et program will avoid recreating generated foo_err.c and 3895foo_err.h files if no changes are necessary. The compile_et program 3896will also atomically replace these files to avoid a potential parallel 3897build race problem on SMP systems. (Addresses Sourceforge Bug: 3898#1157933) 3899 3900Added a new function to the blkid library, blkid_probe_all_new(), 3901which only probes newly added disk devices, and change 3902blkid_find_dev_with_tag() to use this function so that when a 3903requested tag is not found, devices that were previously not checked 3904are searched before searching all devices in the system. 3905 3906Added new functions to the blkid library, blkid_dev_set_search() and 3907blkid_dev_has_tag(). 3908 3909E2fsck's problem strings can now use @m and @n as abbrevations for 3910"multiply-claimed" and "invalid", respectively. 3911 3912The e2fsprog.pot file now has an explanation of how the @-expansion 3913and %-expansion works, and strings in e2fsck/problem.c which contain @ 3914characters now have comments in e2fsprogs.pot with the @-expansion to 3915make life easier for translators. 3916 3917Fixed missing return values in the ext2fs library which could cause it 3918to return random garbage in certain error conditions. 3919 3920Allow the current time to be overriden via the E2FSCK_TIME environment 3921variable for use in regression tests. 3922 3923The test scrpit driver program now exits with a non-zero status if 3924there any of its test that it ran failed. 3925 3926Fixed problems with parabuilds on SMP systems. (Addresses Sourceforge 3927Bug: #1157933) 3928 3929Fixed "make check" so that it compiles correctly even when e2fsprogs' 3930header files have not be installed in the system include directories. 3931(Addresses Sourceforge Bug: #1180572) 3932 3933Fixed gcc -Wall nits. 3934 3935 3936E2fsprogs 1.37 (March 21, 2005) 3937=============================== 3938 3939Add support for checking the validity of Extended Attributes stored in 3940inodes to e2fsck. 3941 3942Add support for dumping the contents of large inodes to debugfs, 3943including the extended attributes stored in inodes. 3944 3945Fix mke2fs, e2fsck, debugfs, and the ext2fs_mkdir function so that 3946when we create a new inode we make sure that the extra information in 3947the inode (any extra fields in a large inode and any ea-in-inode 3948information) is initialized correctly. This can take place when 3949mke2fs creates the root and lost+found directory, when e2fsck creates 3950a new root inode or a new lost+found directory, and when the user uses 3951the debugfs write, mknod, or mkdir commands. Otherwise, the newly 3952create inode could inherit garbage (or old EA information) from a 3953previously deleted inode. 3954 3955Fixed a bug in e2fsck so it would notice if a file with an extended 3956attribute block was exactly 2**32 blocks, such that i_blocks wrapped 3957to zero. 3958 3959Added support to filefrag to detect files which are using the new 3960experimental file extents format, and use the non-ext2 algorithm in 3961that case. Fixed a bug to avoid reporting a false discontinuity if 3962there is one or more unallocated blocks at the beginning of a file. 3963 3964Duplicated a check for noticing whether or not the number of blocks 3965(given a certain blocksize) is greater than 2**32 when the 3966BLKGETSIZE64 ioctl is not available to ext2fs_get_device_size(). This 3967allows mke2fs to automatically use a larger blocksize when creating a 3968filesystem on a very large device when run on systems that do not 3969support BLKGETSIZE64. 3970 3971Fix the I18N build which was broken in e2fsprogs 1.36 because the 3972build system had been switched to treat the .gmo files as shipped 3973files (for backwards compatibility with systems that have older GNU 3974I18N tools installed), but the gen_tarball.in script was still 3975removing the .gmo files from the official source distribution. 3976 3977Fixed various Debian packaging issues --- see debian/changelog for 3978details. (Addresses Debian Bugs ##296769, #299341) 3979 3980Programmer's notes: 3981------------------- 3982 3983Added new functions to the e2p library which convert between a string 3984and os_type: e2p_os2string() and e2p_string2os(), and used them to 3985make the generated binaries more compact. 3986 3987Fixed a compile-time error on Darwin systems. 3988 3989Cleaned up the lib/ext2fs Makefile slightly. 3990 3991 3992E2fsprogs 1.36 (February 4, 2005) 3993================================= 3994 3995All of the patches that were applied to Fedore Core 3's 3996e2fsprogs-1.35-11.2 have been integrated, although sometimes with a 3997lot of bug fixes first. Users of Fedora Core 3 are strongly 3998encouraged to upgrade to e2fsprogs 1.36 as soon as possible. 3999 4000Add support for filesystem with the online resizing via resize inode 4001feature. Fixed numerous bugs from the Fedora patches. The Fedora 4002patches also didn't bother to do any consistency checking on the 4003resize inode, or add any tests to the regression test suite. The "-R 4004resize=4g" option to mke2fs was a no-op in the Fedora patches, despite 4005being listed in mke2fs's usage message. All of these shortcomings 4006have been corrected. 4007 4008E2fsck can also also fix filesystems trashed by Fedora's resize2fs 4009program. In order to do this, the user must run the commands: 4010 4011 debugfs -w /dev/hdXXX -R "features ^resize_inode" 4012 e2fsck -f /dev/hdXXX 4013 4014Optionally, the ext2prepare command can be used to re-enable online 4015resizing after the filesystem has been fixed. 4016 4017The fsck program will now accept an optional filedescriptor argument 4018to the -C option. (The Fedora version of this patch would sometimes 4019cause fsck to ignore a parameter on fsck's command line in some rare 4020cases, sigh.) 4021 4022Make sure e2fsprogs doesn't write garbage into the reserved portion of 4023large inodes. 4024 4025Make sure resize2fs releases the blocks belonging to the old inode 4026table blocks when moving the inode table. (Addresses Debian Bug: 4027#290894) 4028 4029Skip the r_resize_inode test if resize2fs is not compiled (due to 4030configure --disable-resizer) 4031 4032E2fsck now checks the summary filesystem accounting information, and 4033if any of the information is obviously wrong, it will force a full 4034filesystem check. (Addresses Debian Bug #291571) 4035 4036Fix e2fsck to not complain when the resize_inode feature is enabled, 4037s_reserved_gdt_blocks is zero, and there is no DIND block allocated in 4038the resize inode. 4039 4040Fix e2fsck to note delete symlinks that contain an extended attribute 4041after the ext_attr feature flag has been cleared. (Addresses Red Hat 4042Bugzilla #146284). 4043 4044Add new utility program, copy_sparse.c, which is very useful 4045for dealing with large sparse files (such as e2image files). 4046 4047Add support for jnl_blocks[] for debugfs's set_super_value. 4048 4049Fix filefrag so that it works correctly with sparse files. 4050 4051Filefrag -v will print first and last blocks. 4052 4053Add interpretation of OS Creator values for FreeBSD and Lites in mke2fs 4054and dumpe2fs. 4055 4056Add mke2fs support so that it can support filesystems larger than 4TB 4057automatically, by retrying with a 4k blocksize if the device size is 4058too big to be expressed using a 1k blocksize. (Addresses Sourceforge 4059bug #1106631) 4060 4061Change blkid to test for NTFS first because Windows sometimes doesn't 4062clear enough of the parition to confuse the probing routines into 4063thinking the old filesystem type is still valid. (Addresses Debian 4064Bug #291990) 4065 4066Add support for swap partition label and uuid's in the blkid library. 4067 4068Add support to the blkid library to recognize Oracle ASM volumes. 4069 4070Make blkid -t display all devices that match the specified criteria, 4071not just the first one, and work more consistently when the blkid 4072cache file is not available or set to /dev/null. (Addresses Debian 4073Bug #290530 and #292425) 4074 4075Badblocks will now correctly display block numbers greater than 4076999,999,999 in its progress display. 4077 4078The tune2fs program will not allow the user from setting a ridiculous 4079number of reserved blocks which would cause e2fsck to assume the 4080superblock was corrupt. E2fsck's standards for what is a ridiculous 4081number of reserved block has also been relaxed to 50% of the blocks in 4082the filesystem. 4083 4084The blkid library will return vfat in preference to msdos, and ext3 in 4085preference to ext2 (if the journalling flag is set) so that mount will 4086do the right thing. (Addresses Debian bug #287455) 4087 4088Mke2fs will now use the -E option for extended options; the old -R 4089(raid options) option is still accepted for backwards compatibility. 4090Fix a double-free problem in resize2fs. (Red Hat Bugzilla #132707) 4091 4092Mke2fs will now accept a size in megabytes, gigabytes, and other units 4093(via "32m" or "4g" on the command line) if the user finds this more 4094convenient than specifying a block count. 4095 4096Fix an obscure, hard-to find bug in "e2fsck -S" caused by an inode 4097cache conherency problem. 4098 4099Debugfs now supports a new command, set_inode_field, which allows a 4100user to manually set a specific inode field more conveniently, as well 4101as set entries in the indirect block map. 4102 4103Debugfs's set_super_value command has been enhanced so that the user 4104can set most superblock fields, including the date/time fields and 4105some of the more newsly added superblock fields. 4106 4107E2fsprogs programs now accept an offset to be passed to the file 4108specifiers, via the syntax: "/tmp/test.img?offset=1024". 4109 4110E2fsprogs programs will now accept blocksizes up to 65536; kernel 4111support on the x86 doesn't exist for now, but it can be useful on 4112other architectures with page sizes greater than 4k. There are 2.6 4113kernel patches out there which enable this, but they are of this 4114writing still experimental. 4115 4116The e2image command now takes the -s option which will scramble 4117directory entries for raw image files. 4118 4119Fix a file descriptor leak in the filefrag program. 4120 4121Make sure e2fsck doesn't crash when /proc/acpi/ac_adapter is not 4122present. 4123 4124Fix bug in debugfs where kill_file would lead to errors when deleting 4125devices and symlinks. (Sourceforge Bugs #954741 and #957244) 4126 4127Fix bug in the blkid library when detecting the ocfs1 filesystem 4128 4129Remove obsolete EVMS 1.x and a.out DLL support. 4130 4131E2fsck will attempt to recover from a journal containing illegal blocks. 4132 4133Fixed two potential ordering constraint problems in e2fsck which might 4134cause the filesystem to be corrupted if e2fsck is interrupted during a 4135(extremely narrow) race window. Thanks to Junfeng Yang from the 4136Stanford Metacompilation group for pointing this out. 4137 4138Fixed bug in e2fsck where it would not accurately detect whether or 4139not the system is running on adaptor if the ACPI device representing 4140the AC adapter didn't correspond to the what was used on IBM 4141Thinkpads. 4142 4143Change e2fsck to accept directories greater than 32MB. 4144 4145Fix e2fsck so that a checkinterval of zero disables a time-based check 4146of the filesystem. 4147 4148Debugfs will check the DEBUGFS_PAGER enviroment variable in preference 4149to the PAGER environment variable. (Addresses Debian Bug #239547) 4150 4151Tune2fs will not mark rewrite the superblock if the feature bitmasks 4152are not modified. 4153 4154The debugfs program will set the filetype information when creating a 4155link. 4156 4157Add debugfs -d option to use a separate source of data blocks when 4158reading from an e2image file. 4159 4160Add e2image -I option which allows the e2image metadata to be 4161installed into a filesystem. 4162 4163Fixed bug in the badblocks program which caused "done" to always 4164appear in english even when a translation was available. (Addresses 4165Debian Bug #252836) 4166 4167The blkid program has a new option -o which controls the output format 4168of the blkid program; this is makes blkid more convenient to use in 4169shell scripts. 4170 4171Fix a minor bug in uuid library, which was not using the full 14 bits 4172of clock sequence when generating UUID's. 4173 4174Fix a Y8.8888K problem in the uuid library. 4175 4176Logsave now creates a new session id for itself to avoid getting 4177killed by init whan transitioning between init levels. 4178 4179Change the licensing of the UUID library to be the 3-clause BSD-style 4180license; this allows Apple to use the uuid library in Darwin. 4181 4182Add ocfs and ocfs2 probe support into the blkid library. 4183 4184Fix a memory and file descriptor leak in the blkid library. 4185 4186The blkid library will revalidate the device if the system time is 4187earlier than last verification time of the device, since that 4188indicates that the system time is probably nottrustworthy. 4189 4190The blkid library will override the default location of the blkid.tab 4191file by the BLKID_FILE environment variable, if it is available. 4192 4193Change the getsize functions to use the BLKGETSIZE64 ioctl on Linux 2.6. 4194 4195Add various portability fixes for lame new versions of glibc, Darwin 4196and GNU/KFreeBSD, as well as removing XSI:ism's. (Addresses Debian 4197Bugs #239934, #264630, #269044, #255589, #289133) 4198 4199Add support for Windows 9x/NT under Cygwin. 4200 4201Updated and clarified various man pages. (Addresses Debian Bugs #236383, 4202 #241940, #238741, #242995, #256669, #268148, #256760, #273679) 4203 4204Updated and fixed translations. (Addresses Debian bugs #244105, #262836) 4205 4206Update the rpm spec files so that it works better with Fedora core 2 4207and RH9. 4208 4209Fixed various Debian packaging issues (see debian/changelog). In 4210particular, fixed the Debian initrd scripts. (#241183, #248050, 4211#253595, #247775) 4212 4213 4214Programmer's notes: 4215------------------- 4216 4217Fixed various gcc -Wall warnings. 4218 4219The uuid library now has new functions uuid_unparse_upper() and 4220uuid_unparse_lower() which forces the case of the hex digits to be 4221upper case, or lower case. 4222 4223The build process has been speeded up by enhancing the subst program 4224to update the modtime on the generated files even when the generated 4225file hasn't changed. 4226 4227The uuid library now uses C99 stdint.h types instead of custom types. 4228 4229Updated config.guess and config.sub with newer versions from the FSF. 4230 4231Removed out of date .cvsignore files from the source distribution. 4232 4233The ext2fs_unlink() function will return an error if both the name and 4234inode number are unspecified, to avoid doing something surprising 4235(such as unconditionally deleting the first directory entry). 4236Directory entries are now deleted by coalescing them with the previous 4237directory entry if possible, to avoid directory fragmentation. This 4238is not an issue with the e2fsprogs suite, but may be a problem for 4239some of the users of libext2fs, such as e2tools. 4240 4241Add support for version numbers of the form "1.36-rc1". 4242 4243Fix build of mke2fs.static. 4244 4245Add basic ext2fs library support for large (EA in inode) inodes. 4246 4247The test_io mechanism can now abort after n reads or writes to a 4248particular block. The block is specified by TEST_IO_BLOCK environment 4249variable, and the read/write count by the TEST_IO_READ_ABORT and 4250TEST_IO_WRITE_ABORT environment variables. The block data is now only 4251dumped if the 0x10 bit is set in TEST_IO_FLAGS. 4252 4253UUID_DEFINE() in the uuid library now creates a static variable, with 4254__attribute__ ((unused)) if we are using GCC, so that UUID_DEFINE can 4255be used in header files. 4256 4257Add support for the install-strip and install-shlibs-strip targets, as 4258suggested by the GNU coding guielines. "make install" no longer 4259strips the binaries which are installed. 4260 4261Remove support for the --enable-old-bitops configure option which was 4262only for very old sparc systems. 4263 4264Remove support for --enable-clear-htree; this was only needed during 4265the early development of the htree patch. 4266 4267Use Linux-kernel-style makefile output so it is easier to see compiler 4268warnings. 4269 4270Update gettext files to version 0.14.1. 4271 4272Update to use autoconf 2.5x. 4273 4274Improved support for compiling e2fsprogs under dietlibc. 4275 4276Make e2fsprogs portable to Solaris and FreeBSD systems. 4277 4278Add blkid_verify(), blkid_get_library_version(), and 4279blkid_parse_version_string() functions to the blkid library. 4280 4281Add pkg-config files for e2fsprogs's libraries. 4282 4283Fix "make uninstall" to so that it removes everything that is installed. 4284 4285Add a configure --enable-maintainer-mode option which enables the 4286makefile rules to rebuild the configure script from configure.in, and 4287to reubuild the .gmo files in po directory. 4288 4289Drop the sparc assembly bitwise operations; it's less efficient 4290than the GCC 3.4 compile code and triggers compiler warnings on 4291sparc64. Thanks to Matthias Andree for his analysis and suggestions. 4292(Addresses Debian Bug #232326) 4293 4294 4295E2fsprogs 1.35 (February 28, 2004) 4296================================== 4297 4298E2fsck has a new -k option, which in conjunction with the -c options, 4299preserves the existing badblocks list. 4300 4301Cleaned up e2fsck's preen-mode messages during the passes 1b, 1c, and 1d. 4302 4303E2fsprogs will now deal correctly with symlinks that contain 4304extended attribute information, which can be created using SE Linux. 4305(Addresses Debian Bug #232328) 4306 4307Remove a double longjump into an invalid stack frame bug in e2fsck. 4308(This was during an abort sequence, which normally worked on Linux and 4309caused a core dump on other operating systems.) 4310 4311Fix NLS bug in e2fsck, by avoiding trying to expand an empty string 4312(the NLS library will replace "" with the .po header information). 4313 4314Fix a bug in mke2fs which caused -T largefile or -T largefile4 to core 4315dump due to a division by zero error. (Addresses Debian bug #207082) 4316 4317Fixed a bug in e2fsck which caused it to incorrectly fix a filesystem 4318when reconnecting a directory requires creating a lost+found 4319directory. (Addresses Debian bug #219640). 4320 4321Fixed a bug where e2fsck would bomb out if a journal needed to be 4322replayed when using an alternate superblock. 4323 4324E2fsck will give an extra grace period before actually forcing a check 4325if the laptop is running on battery. The next time fsck runs while 4326the system is on the AC mains, or after the grace period is exceeded, 4327the filesystem will be checked. (Addresses Debian bug #205177) 4328 4329E2fsck will inform the user when there are 5 or fewer mounts before a 4330filesystem check will be forced. (Addresses Debian bug #157194) 4331 4332Fix e2fsck's handling of corrupted indirect blocks in the bad block. 4333We now correctly handle the case where there is an overlap between a 4334block group descriptor or a superblock and a bad block indirect block. 4335In the case where the indirect block is corrupted, we now suggest 4336"e2fsck -c". 4337 4338Fix byte swap bugs in e2fsck that caused the journal backup location 4339in the superblock and symlinks created by SE Linux to be cleared 4340by e2fsck on big-endian machines. (Addresses Debian bug #228723) 4341 4342E2fsck -c now replaces the current list of bad blocks with the ones 4343found by badblocks. 4344 4345Fix bugs in e2fsck and tune2fs which could cause a core dump if a 4346non-existent LABEL or UUID specifier is to e2fsck or tune2fs. 4347 4348Fix a potential bug in e2fsck which could cause it to core dump when 4349trying to print the location of the backup superblock. 4350 4351Protect against a potential core dump in e2fsck when printing a 4352message about backup superblocks. 4353 4354Add support for backing up the journal inode location in the 4355superblock. E2fsck will automatically save the journal information in 4356the superblock if it is not there already, and will use it if the 4357journal inode appears to be corrupted. ext2fs_add_journal_inode() 4358will also save the backup information, so that new filesystems created 4359by mke2fs and filesystems that have journals added via tune2fs will 4360also have journal location written to the superblock as well. 4361Debugfs's logdump command has been enhanced so that it can use the 4362journal information in the superblock. 4363 4364E2fsck will now update all superblocks when moving the journal inode. 4365 4366Shrink the size of the e2fsck executable by moving some initialized 4367variables to the BSS segment. 4368 4369E2fsck will avoid printing the ^A and ^B characters which bracket the 4370progress bar when stdout and stdin are a tty device instead of a pipe 4371to another program. (Addresses Debian bug #204137) 4372 4373Debugfs's mkdir command will automatically expand the directory if 4374necessary. (Addresses Debian Bug: #217892) 4375 4376Fixed a bug in debugfs so that copying a file from /dev/null uses the 4377correct mode bits. (Addresses Debian Bug: #217456) 4378 4379If the environment variables DEBUFS_PAGER and PAGER are not set, 4380debugfs now searches for the appropriate pager to use, beginning with 4381/usr/bin/pager, and then falling back to 'more' and 'less'. 4382(Addresses Debian bug #221977) 4383 4384Debugfs will now support 2.6 device numbers where the major or minor 4385number may be larger than 255. (Addresses Sourceforge bug #865289) 4386 4387Fix debugging printf in resize2fs. (Addresses Debian Bug #271605) 4388 4389Chattr now stops processing options when it sees '--'. (Addresses 4390Debian bug #225188) 4391 4392Fix regression tests so they work correctly when e2fsprogs is compiled 4393with configure --disable-htree. 4394 4395Fix bug in uuid library when there is no network card and the library 4396is generating a time-based uuid. The random MAC address was not 4397correctly generated to be a multicast address. 4398 4399Add compile_et extensions from Heimdall that were missed the first time 4400around. 4401 4402Fix bug in badblocks when using O_DIRECT; we need to make sure that 4403we're reading from an offset which is page aligned. For read-only and 4404read-write tests, we try to recover after an error so that we can 4405continue reading on page-aligned boundaries. (Addresses Debian Bug 4406#203713) 4407 4408Badblocks now checks 64 blocks at a time instead of 16. (Addresses 4409Debian bug #232240) 4410 4411Updated and clarified various man pages. (Addresses Debian Bug 4412#206845, #222606, #214920, #232406) 4413 4414Updated and fixed translations. (Addresses Debian bugs #200086, #214633) 4415 4416Fixed various Debian packaging issues (see debian/changelog). 4417 4418Programmer's notes: 4419------------------- 4420 4421Fixed a build problem so that e2fsprogs would compile with the 4422--enable-profile option to configure selected. (Addresses Sourceforge 4423bug #811408) 4424 4425Fixed C++ problems with the ext2fs.h header. (Addresses Red Hat 4426Bugzilla Bug #112448) 4427 4428Centralize code which calculates the location of the superblock 4429and block group descriptors so that it is in a single library routine. 4430 4431Added two new functions, ext2fs_file_open2() and 4432ext2fs_inode_io_intern2() which take a pointer to an inode structure. 4433 4434Fix compile_et to output the correct prototype for 4435initialize_xxx_err_table_r() in the header file. (Addresses Debian 4436bug #204332) 4437 4438In the lib/et makefile, make sure com_err.info is deleted on "make clean". 4439 4440Fix 64-bit warnings in e2fsprogs pass1b by using inttypes.h if 4441present. This is for when we try stuffing an int into void * pointer. 4442 4443Fix type-punning which can cause gcc 3.x to miscompile code by getting 4444confused about pointer aliasing. ext2fs_getmem(), ext2fs_free_mem(), 4445and ext2fs_resize_mem() all now take a 'void *' instead of a 'void 4446**'. The EVMS code uses an ugly union approach since we don't want to 4447modify the EVMS interfaces. 4448 4449Make sure all Makefiles use $(MAKE) rather than hardcoded "make", to 4450aid build process on systems can use invoke GNU make as "gmake". 4451 4452Added regression testing for mke2fs. 4453 4454Fixed gcc -Wall nitpicks. 4455 4456Fixed various compiler warnings. 4457 4458Add portability fixes for FreeBSD and for using fsctl under Darwin to 4459support ext2 ioctl's. 4460 4461 4462E2fsprogs 1.34 (July 25, 2003) 4463=============================== 4464 4465Fixed a bug introduced in E2fsprogs 1.30 which caused fsck to spin in 4466a tight loop while waiting for a child fsck to exit in some cases. 4467This burns CPU times which slows down the low-level filesystem check. 4468 4469Added code to mke2fs to assure that the default block size for a 4470filesystem is at least as big as the sector size of the device, if it 4471can be determined. 4472 4473Changed mke2fs and resize2fs to round the default size of a filesystem 4474to be an even multiple of the VM pagesize in order to avoid a Linux 4475kernel bug introduced when the storage of the buffer cache was moved 4476into the page cache. 4477 4478Mke2fs will warn the user when creating a filesystem with journaling 4479and a blocksize greater than 4096. (Addresses Debian bug #193773) 4480 4481Fixed a bug in resize2fs which caused it to fail on filesystems with a 4482non-empty bad block list. Resize2fs now discards any blocks on the 4483badblock list which are no longer part of the filesystem as the result 4484of a filesystem shrink. (Note: this means that shrinking and then 4485enlarging a filesystem is no longer a reversible operation; 4486information about bad blocks in the part of the filesystem which is to 4487be chopped off will be lost.) 4488 4489Changed resize2fs so the user can use prefixes to specify the units of 4490the new filesystem size (sectors, kilobytes, megabytes, or gigabytes), 4491and to make the error and informational messages explicitly display 4492the blocksize used by the filesystem, in order to avoid confusion. 4493(Addresses Debian bug: #189814) 4494 4495Added a new debugfs command, dump_unused, which dumps the contents of 4496all unused blocks to stdout. (Useful as an emergency try-to-find 4497deleted data command.) 4498 4499Added a new debugfs command, imap, which prints the location of a 4500specified inode in the inode table. 4501 4502Fixed a bug in the badblocks program which caused it to use one bit of 4503randomness in its non-destructive read/write test, instead of using a 4504full 8 bits of randomness. 4505 4506Added a new option (-t) to badblocks, which allows the user to control 4507the test pattern(s) used when checking a disk. 4508 4509The blkid probe function now more correctly detects UDF filesystems. 4510 4511Fixed a bug in the blkid library which caused it to not update its 4512cache if a filesystem changes from having a label to not having a 4513label. 4514 4515Fixed a bug in the blkid library wihch could avoid an infinite loop 4516in blkid_find_dev_with_tag() if /proc is not mounted and there the 4517/etc/blkid.tab file has not yet been created. 4518 4519Fixed the badblocks program so that the destructive read/write test 4520honors the -c option, and to use O_DIRECT when possible to avoid 4521thrashing the system block buffer cache. 4522 4523Fixed various NLS issues. 4524 - Added Czech and Swedish translations 4525 - Removed testing NYC translation 4526 - Fixed NLS support for message abbrevations in e2fsck 4527 - Remove de-utf.po, since we shouldn't have two versions using different 4528 charset encodings. 4529 - Used ngettext() (abbreivated with the macro P_(str1, str2, n)) to 4530 simplify the statistics reporting in e2fsck. 4531 4532Changed configure.in so that its defaults for *BSD systems no longer 4533build an fsck wrapper, and not to install in /usr/local by default. 4534 4535Fixed some minor spelling errors/typo's in e2fsck and the configure 4536script. 4537 4538Fixed various Debian packaging issues (see debian/changelog). 4539 4540Updated and clarified man pages. (Addresses Debian Bug #195616) 4541 4542Programmer's notes: 4543------------------- 4544 4545Fix gcc -Wall nitpicks. 4546 4547Updated gettext implementation used by e2fsprogs to 0.11.5, and enable 4548NLS support by default. (Added partial workaround for gettext/Darwin 4549incompatibility problems.) 4550 4551Added full MIT KRB5 and Himdall compaibility support to the com_err 4552library and the compile_et program. (Addresses Debian bug #191900) 4553 4554Added the blkid_known_fstype() function to the blkid library, which 4555returns true if it is passed a filesystem type which is recognized by 4556the blkid probing functions. 4557 4558Improved the documentation for the blkid library. 4559 4560Added the ext2fs_get_device_sectsize() function the the ext2fs library, which 4561returns the hardware sector size of a device, if it is available. 4562 4563Added a dependency in the blkid library's .so file to the uuid 4564library, since the former uses the latter. (Addresses Debian bug 4565#194094) 4566 4567Added --with-diet-libc and --disable-evms to the configure script. 4568 4569Fixed a minor memory leak in the badblocks program. 4570 4571Fixed a portability problem in tune2fs --- not all systems have strptime(). 4572 4573Fixed a portability problem in debugfs with the use of getopt() more 4574than once. Old-style BSD, new-style BSD, and Linux C libraries all do 4575things differently. 4576 4577Add support Windows support to ext2fs_get_device_size(). 4578 4579Added (normally disabled) debugging code to the Unix I/O manager which 4580causes it to disable all userspace caching if the NO_IO_CACHE is 4581defined. 4582 4583Changed the test I/O manager so it can always be linked into e2fsck, 4584mke2fs, and tune2fs if enabled via --enable-test-io-debug to the 4585configure script. The test I/O manager will only print any debugging 4586information if the TEST_IO_FLAGS or TEST_IO_BLOCK environment 4587variables are set, which specifies which I/O operations are logged and 4588a block number to watch, respectively. The log messages are sent to 4589stderr by default, unless a filename is specified via the 4590TEST_IO_LOGFILE environment variable. 4591 4592 4593E2fsprogs 1.33 (April 21, 2003) 4594=============================== 4595 4596Added a new utility program, logsave, which captures the output of a 4597command in a log file, even if the containing directory hasn't been 4598mounted yet (in which case the it saves the output in memory until it 4599can write out the logfile). This is useful for capturing the 4600output of fsck during the boot sequence. 4601 4602Fixed some portability problems that were causing problems under 4603the Cygwin32 environment. 4604 4605Mke2fs now interprets a negative number to the -b option as a minimum 4606block size. 4607 4608Fixed a bug in mke2fs which was incorrectly checking the argument to 4609the -g option if the default block size was used. (Addresses Debian 4610bug #188319) 4611 4612Fsck now explicitly ignores tmpfs and devpts, and it will complain if 4613it can not find filesystem checkers for jfs, reseirfs, and xfs. 4614 4615E2fsck now updates the global free block and inode counters from the 4616block group specific counters quietly. This is needed for an 4617experimental patch which eliminates locking the entire filesystem when 4618allocating blocks or inodes; if the filesystem is not unmounted 4619cleanly, the global counts may not be accurate. 4620 4621Imported bug fixes to the EVMS plugin from the EVMS 2.0 tree. (EVMS 46222.0 is not yet supported; this only pulled in the bug fixes: fixed 4623possible hangs caused by bugs in calling waitpid, and not setting the 4624pipe to non-blocking mode; also fixed a file descriptor leak; made 4625sure all functions call log entry/exit functions.) 4626 4627Badblocks now flushes its output file as bad blocks are discovered. 4628 4629The uuid library is now more paranoid about assuming the correctness 4630of the /dev/random device; it mixes in a stream of bytes from 4631random/srandom, just in case. 4632 4633Update Debian files to reflect the fact that I am now the Debian 4634maintainer of e2fsprogs. Other various Debian-specific packaging 4635cleanups. 4636 4637Move the source tarball generation functions from the top-level 4638makefile to the util/gen-tarball script. 4639 4640Updated the Turkish .po translation file. 4641 4642Added Heimdal and MIT krb5 extensions to the com_err library to make 4643it more compatible with com_err libraries from those distributions. 4644 4645Changed dumpe2fs to always display the superblock fields relating to 4646the journalling and/or directory indexing feature even if those 4647features are not enabled. 4648 4649Updated and clarify copyright statement vis-a-vis alpha releases of 4650e2fsprogs. 4651 4652The ss library will now try to dynamically link to the readline 4653library and use it if it is present in the system. This means that 4654the debugfs program now has line editing and history features. The 4655SS_READLINE_PATH environment variable is used to find a readline or 4656readline-compatible library. 4657 4658E2fsck now finds most duplicate filenames (all when rebuilding all 4659directories via the -D option) and offers to delete or rename 4660duplicate filenames/directory entries. (Addresses Debian Bug #176814). 4661 4662Fix bug in e2image. When writing out a raw image file, include data 4663blocks from symlinks that do not store the symlink within the inode. 4664 4665Fix bug in resize2fs which incorrectly moved the block and inode 4666bitmaps for sparse superblock filesystems and incorrectly marked 4667blocks as in use. (Addresses Debian bug #174766) 4668 4669Added a new shared library, the blkid library, which efficiently 4670allows fsck, mke2fs, e2fsck, and tune2fs to be able to look up LABEL 4671and UUID filesystem specifiers without needing to search all of the 4672devices in the system. Instead, the device is looked up in a cache 4673file, and then verified to make sure the blkid cache is correct. 4674 4675Tune2fs and e2label will accept LABEL=xxx and UID=yyy specifiers for 4676the device name, using the blkid library. (Addresses Debian bugs 4677#166048, #179671) 4678 4679Fsck now supports backslash escapes in /etc/fstab so that \040 can be 4680used for spaces in device labels. 4681 4682Removed 32-bit limitations for debugfs's dump command. 4683 4684If the user specifies a large number of inodes, Mke2fs will 4685automatically adjust the number of blocks per group to find a valid 4686set of filesystem parameters. 4687 4688Add workaround to detect broken MD devices where when some of the 4689underlying devices are marked read-only, writes to the MD device are 4690silently dropped. E2fsck will detect if there is an attempt to run 4691the journal twice, and abort with an error if this is the case. 4692(Addresses IBM Bugzilla bug #1226) 4693 4694E2fsck will print an error if more than one of the -p/-a, -n or -y 4695options are specified. 4696 4697E2fsck will fix HTREE corruptions in preen mode, without stopping the 4698boot process. This is needed because the 2.4 ext2 filesystem 4699accidentally had the INDEX_FL backwards compatibility code removed. 4700 4701Mke2fs no longer creates filesystems with the dir_index flag set by 4702default; the user has to specifically request it. 4703 4704Update and clarified various man pages. (Addresses Debian bugs 4705#173612, #175233, #175113, and #170497, #185945, #188318) 4706 4707Created man page for the mk_cmds program (from the libss library). 4708 4709Programmer's notes: 4710------------------- 4711 4712Fix various gcc -Wall nits. 4713 4714Fixed a lot of portability problems that caused e2fsprogs not to build 4715successfully under Solaris and Apple/Darwin. 4716 4717Fixed a Makefile dependency to allow building e2fsprogs using parallel 4718make jobs. 4719 4720Changes to create a subset distribution which consists only of the 4721et, ss, uuid, and blkid libraries. The configure script and top-level 4722makefile were changed to support working with a subset distribution. 4723 4724Removed EXT2_FEATURE_RO_COMPAT_BTREE_DIR mention of since it's not 4725actually used, and might people who are looking for 4726EXT2_FEATURE_COMPAT_DIR_INDEX, which is in use. 4727 4728Updated debian files to fix a number of Lintian warnings. 4729 4730Updated config.guess and config.sub with newer versions from the FSF. 4731 4732Removed unnecessary libraries from being linked into the fsck, lsattr, 4733chattr, and blkid executables. 4734 4735E2fsprogs 1.32 (Nomvember 9, 2002) 4736================================== 4737 4738Fixed a bug in the Unix I/O routines which caused needless writebacks 4739of clean blocks from the unix_io cache (they were erroneously marked 4740as being dirty, so they were getting written back to disk before 4741getting evicted from the disk cache). This was harmless, but it 4742significantly slowed down e2fsck. 4743 4744Made some other minor optimizations to the Unix I/O routines to save a 4745small amount of CPU time. 4746 4747Updated internationalization files. 4748 4749 4750E2fsprogs 1.31 (Nomvember 8, 2002) 4751================================== 4752 4753Update EVMS ext2fsim plugin with EVMS 1.2. (We still support 4754compiling the fsim plugin with EVMS 1.0 and 1.1.) Add better error 4755handling for child process that die unexpectly. Add a hack to force 4756mkfs to create filesystems that won't cause problems with hardware 4757that has 2k or 4k minimum blocksize requirements. Read from child 4758processes in non-blocking mode, so that the GUI continues to update. 4759 4760Fix e2fsck so that it returns the appropriate exit code when the root 4761filesystem has been changed, so that system's rc scripts will be told that 4762the system needs to be rebooted. 4763 4764Fix a bug in ext2fs_flush/ext2fs_close; when the MASTER_SB_ONLY flag 4765was set, some of the descriptor blocks that should have been written 4766out were getting skipped. 4767 4768Changed e2fsck to force out changes to the backup copies of the 4769superblock and block group descriptors when important changes are made 4770to those data structures. 4771 4772Fix a bug where e2fsck could erroneously mark a filesystem as being 4773clean if a check of dirty filesystem is interrupted with a ^C. (Bug 4774introduced in e2fsprogs 1.28.) 4775 4776If journal debuging is enabled using --enable-jbd-debug, the debugging 4777level is now set via the E2FSCK_JBD_DEBUG environment variable. 4778 4779If byteswapping support is disabled using configure --disable-swapfs, 4780skip the tests which depend on byte-swapping, so that "make check" 4781won't bomb out. 4782 4783Lshattr will now display the indexed directory flag. Also, some of 4784the more esoteric compression flags are supressed unless compression 4785support has been enabled. 4786 4787Update man pages. 4788 4789 4790E2fsprogs 1.30 (October 31, 2002) 4791================================= 4792 4793When resizing a filesystem, and inodes need to moved, resize2fs will 4794update the inode's ctime field, and the mtime field of the containing 4795directories, so that incremental backups using dump/restore will work 4796correctly. 4797 4798In order to avoid spurious errors, e2fsck wil check the last 4799superblock write time to determine whether or not it can safely use 4800the LOW_DTIME checks to determine if there are inodes on the orphan 4801inode linked list that had somehow gotten disconnected. (Addresses 4802Sourceforge bug #620980) 4803 4804Partition in /proc/partitions that being with the string "lvm" are 4805considered real partitions even if they do not end with a number. 4806 4807Fixed a bug in the the uuid generation function, where if /dev/urandom 4808is not present, but /dev/random is, and there isn't sufficient 4809entropy, the get_random_byte function could spin a loop forever. 4810 4811E2fsck, mke2fs, etc. will now reliably notice when image files are 4812mounted using the loopback interface. (Addresses Sourceforge bug 4813#619119) 4814 4815When flushing buffers (as is done by badblocks, for example) check to 4816see if the BLKFLSBUF ioctl succeeds, and if so, avoid using the 4817FDFLUSH ioctl which causes the MD device driver which causes confusing 4818syslog messages about obselete ioctl messages. (Addresses Sourceforge 4819bug #545832). 4820 4821Debugfs's write command now checks to see if the destination filename 4822exists before creating it. (Addresses Sourceforge bug #478195.) 4823 4824When installing man pages, search for compressed man pages using all 4825commonly used compression extensions (i.e., .Z, .gz, .bz2, etc.) 4826 4827Fixed a bug in fsck where multiple filesystems specified on the 4828command were not being checked in parallel due to a logic bug 4829introduced to support the FSCK_MAX_INST environment variable. 4830 4831We have added a new superblock field, s_mkfs_time, which indicates 4832when a filesystem was created. It is set via mke2fs, and printed out 4833by dumpe2fs, but is not actually touched by the kernel. 4834 4835Dumpe2fs has been made more robust by not aborting if there is an 4836error reading the block/inode bitmaps; instead it will still print out 4837the location of the block/inode bitmaps and inode table. 4838 4839Add support for the an alternative block group descriptor layout which 4840allows for on-line resizing without needing to prepare the filesystem 4841in advance. (This is the incomat feature flag meta_bg.) 4842 4843Add support for storing default mount options in the superblock, so 4844that the filesystem can be mounted with specific mount options without 4845needing to specify them on the mount command line or in the /etc/fstab 4846file. 4847 4848Add support for a new inode flag, which is to be used for indicating 4849the top of directory hierarchies for the Orlov block allocator. 4850 4851Fix e2fsck so that if it creates the lost+found directory, it does so 4852with the more apporpriate permissions of 0700. Also change 4853mklost+found so that it also creates the directory 0700. 4854 4855Fixed format bug in e2fsck if NLS is in use. 4856 4857Add a German translation for e2fsprogs's NLS support. 4858 4859Fixed e2fsck so that it more handles BAD_BLOCK_IN_INODE_TABLE even at 4860the beginning of the inode table. This won't matter much, since if 4861there is a bad block at the beginning of the inode table, the root 4862directory won't be available. But at least e2fsck won't crash in this 4863case. 4864 4865Fixed endian problems in the htree code for e2fsck and debugfs. 4866 4867When byte-swapping a filesystem on a PPC architecture, byte-swap the 4868bitmaps since the historical big-endian ext2 variant had byte-swapped 4869bitmaps, and the ext2fs library assumes this. This fixes the 4870regression test suite on PPC machines. 4871 4872Fix e2image so that it handles a bad block in the inode table 4873appropriately. 4874 4875E2fsck now uses a more sophisticated algorithm to salvage corrupted 4876directories that recovers more information in the corrupted directory 4877block. 4878 4879E2fsck now performs additional consistency checks on indexed (HTREE) 4880directories. 4881 4882Fix bug where efsck might get confused about whether a completely 4883empty directory block is an empty leaf block or an interior htree 4884node. This could cause e2fsck to get confused into think that a valid 4885indexed directory was corrupted. 4886 4887E2fsck no longer creates an empty directory entry at the end of a 4888directory block when optimizing a directory. This would cause some 4889earlier versions of the dxdir kernel code to corrupt the directory 4890when attempting to split a node containing an empty directory entry. 4891 4892E2fsck could sometimes in rare circumstances leave the indexed flag 4893set after a small directory was optimized by compressing it instead of 4894indexing it. (This should never happen in real life, since 4895directories that small wouldn't have been indexed, but better safe 4896than sorry.) 4897 4898E2fsck now only updates the master superblock in all cases. This 4899slightly shortens its run time. 4900 4901Ext2ed can deal with directory entries of length 0; previously it 4902would get stuck in an infinite loop. 4903 4904Fsck now has support for reiserfs volumes when parsing LABEL= and UUID= 4905specifiers. (Sourceforge patch #613447) 4906 4907Badblocks will now work correctly on read-only devices such as 4908CD-ROM's. (Sourceforge patch #600451) 4909 4910Updated and clarified man pages. (Addresses Debian bug #167108) 4911 4912 4913E2fsprogs 1.29 (September 24, 2002) 4914=================================== 4915 4916Fixed a bug in e2fsck which could corrupt a directory when optimizing 4917it (via the -D option) or rebuiliding the hash tree index with a 1 in 4918512 probability, due to a fence post error. 4919 4920Fixed a bug in the LVM support code which caused LABEL='xxx' not to 4921work correctly. 4922 4923Mke2fs now enables the directory indexing flag by default. (Since 4924this is a compatible feature flag, it's safe to do so.) 4925 4926Tune2fs will support setting the directory indexing feature flag. It 4927will automatically set up the default hash algorithm and hash seed 4928fields in the superblock. 4929 4930If the bone-headed user enters the root filesystem twice in 4931/etc/fstab, the -R option which skips the root filesystem will skip 4932all of them. (Addresses Debian bug #159423). Note! This is not a 4933precedent for dealing intelligently with any other kind of doubled 4934entry in /etc/fstab! 4935 4936 4937Programmer's notes: 4938------------------- 4939 4940Removed perror declaration in lib/et/internal.h. All modern systems 4941can be expected to define perror() these days. Besides, the lib/et 4942code wasn't using perror at all anyway. :-) 4943 4944 4945E2fsprogs 1.28 (August 31, 2002) 4946================================ 4947 4948Add support for the Hashed-Tree Directory Indexing to e2fsck. Support 4949for setting the htree flag is not included yet, although it can be 4950manually turned on via the debugfs program. 4951 4952Clarified e2fsck error message which is printed when it cannot find 4953sufficient contiguous block when relcating filesystem metadata. 4954 4955Added support for building an EVMS plugin module for ext2/3. This 4956module is substantially the same as the EVMS module shipping with EVMS 49571.1, with one or two bugfixes. E2fsprogs can also build this plugin 4958for use with EVMS 1.0 (which did not include the ext2 plugin module), 4959if the configure --enable-old-evms flag is given. 4960 4961Fsck will search through EVMS volumes when trying to resolve 4962filesystem specifications such as LABEL=xxx or UUID=xxx. 4963 4964Added a new utility program, /sbin/findfs, which will return 4965filesystem specifications such as LABEL=xxx or UUID=xxx, and prints 4966the device name. 4967 4968Update and clarified various man pages. (addresses Debian Bug #145044, 4969#146437, #131350, #151990, #144621, #141938) 4970 4971If there are no filesystems specified on fsck's command line, fsck now 4972treat that as if the -As options were given. Previously it would 4973simply do nothing. (Addresses Debian Bug #153102) 4974 4975Mke2fs no longer treats a failure to be able to clear the MD signature 4976at the end of the filesystem as a fatal error. (Addresses Debian Bug 4977#155007) 4978 4979The e2p library functions (which are used by lsattr and chattr) now 4980double check to make sure the file is a regular file or directory 4981before attempting to use the ext2 ioctls. Some device drivers 4982unfortunately respond to the ext2 ioctl's with unknown behaviour. 4983(Addresses Debian Bug #152029). 4984 4985The extended attribute handling has been updated to correspond with 4986the latest V2 bestbits ACL code. 4987 4988Fixed bug in e2fsck which caused it to not clear the dtime field when 4989processing truncated inodes on the orphan list. This could cause data 4990loss(!) if a filesystem is rebooted before a truncate has been 4991committed. 4992 4993E2fsck now uses red/black trees in pass1b, which removes some O(n**2) 4994algorithms. This makes e2fsck much faster in the case of severely 4995corrupted filesystems where a large number of blocks are claimed by a 4996large number of inodes. (Thanks to the 2.5 IDE device driver for 4997inspiring this work.) 4998 4999Resize2fs has been significantly sped up when shrinking and expanding 5000a filesystem by a very small number of blocks (for example, when EVMS 5001is converting a partition to be an EVMS legacy volume). 5002 5003Added a new option to e2fsck, -D, which will optimize or compress all 5004of the directories in the filesystem. 5005 5006E2fsck now catches SIGINT and SIGTERM to make sure it will can 5007properly clean up and only exit at safe points. Fsck will pass 5008SIGINT/SIGTERM to its child processes, and wait until they have all 5009exited before it exits. 5010 5011The uuid parsing code in the uuid library now properly complains when 5012an illegally formated uuid is presented to it. (Addresses Debian bug 5013#152891) 5014 5015Restrict use of the 2.4 setrlimit ABI f*ckup to kernels between 2.4.10 5016and 2.4.17, since the workaround can cause problems when using a 2.4 5017kernel with an old version of glibc built with the 2.2 headers. 5018 5019Fixed a bug in mke2fs where it wasn't properly clearing the initial 5020superblock used by other filesystems. (Addresses Debian bug #147256.) 5021 5022Added support for the synchronous directory feature written by Andrew 5023Morton. 5024 5025The debugfs program can delete directories using the rmdir command. 5026 5027Add support for 8k and 16k filesystems (for systems with page sizes 5028that are greater or equal to 8k or 16k, respectively). Note that 5029these filesystems can not be mounted on x86 systems, or other systems 5030with only 4k page sizes, due to limitations in the current Linux VM 5031code. 5032 5033Resize2fs requires that the filesystem state be valid and have no 5034errors; otherwise, e2fsck -f must be run first. (Previously it simply 5035required that the last fsck time be greater than the last mount time.) 5036 5037Configure now defaults the man pages directory to /usr/share/man on 5038Linux systems. 5039 5040E2fsck now offers to truncate inodes which contain too many blocks (so 5041that i_blocks would overflow. Also fixed handling of large sparse 5042files. 5043 5044E2fsck now more completely checks for symlink validity, including 5045requiring NULL termination and length checks. 5046 5047E2fsck will offer to try forcing a disk write to remap a bad block 5048after finding a read error when reading a filesystem metadata block. 5049 5050Fixed a bug in debugfs which caused the -b and -s options to crash 5051debugfs, as well as breaking the testb, setb, and clearb functions. 5052 5053Added a bmap command to debugfs which calculates the logical to 5054physical block mapping for a particular inode. 5055 5056Fixed a bug in code which checked to see if a device was mounted which 5057sometimes (rarely) failed in the case of a plain file. 5058 5059Fixed a bug in resize2fs where when it reported an error, it would 5060print a message erroneously indicating that the filesystem had been 5061resized before it aborted. 5062 5063When resizing a plain file which is smaller than the requested size, 5064resize2fs will attempt to extended the file so that programs like 5065e2fsck will not complain that the file is too small. 5066 5067Resize2fs will print the actual new size of the filesystem when it is 5068finished resizing. 5069 5070Fixed a bug in debugfs where "ls -l" would report incorrectl file type 5071information on big-endian systems. 5072 5073 5074Programmer's notes: 5075------------------- 5076 5077Fixed collisions with C++ reserved words. 5078 5079Added portability fixes for building e2fsprogs on the HURD and AIX. 5080 5081Added the ext2ed program for creation of test cases. (ext2ed has many 5082limitations and bugs which make it unsuitable for production use.) 5083 5084The ext2fs_read_dir_block2 and ext2fs_write_dir_block now take a new 5085flag, EXT2_DIRBLOCK_V2_STRUCT, which will will reverse when the 5086name_len field is byte swampped on big-endian machines, since in the 5087V2 structure, name_len is a char field which is doesn't need to be 5088byte swapped --- except if an old-style kernel had byte-swapped the 5089name_len field as part of the V1 structure. 5090 5091 5092E2fsprogs 1.27 (March 8, 2002) 5093============================== 5094 5095The warning messages for mke2fs now go to standard error. 5096 5097Fixed to make sure "make check" runs all of the test programs with 5098LD_LIBRARY_PATH set, so that we test the libraries in the build tree. 5099 5100The mke2fs program checks the boot sector for the BSD disk label, and 5101avoids erasing it if it is there. 5102 5103Fixed a bug in e2fsck which caused it to core dump if the journal 5104inode was missing when it was supposed to be there. 5105 5106E2fsck now prints ranges in pass 5 when printing deltas for the block 5107and inode bitmaps. 5108 5109Debugfs's "ls -l" command now will print out the file type information 5110in the directory entry. 5111 5112Create man pages and hard links for fsck.ext3 and mkfs.ext3. If 5113mke2fs is invoked as mkfs.ext3, create the filesystem with a journal. 5114 5115Debugfs can now examine the experimental directory indexing 5116information. 5117 5118Fixed bug in debugfs which caused it to core dump if modify_inode is 5119called without an open filesystem. 5120 5121The debugfs lsdel command now runs its output through a pager first. 5122 5123When installing manual pages, remove the compressed manual pages first. 5124 5125Synchronized with Debian's packaging information for e2fsprogs-1.26-1. 5126 5127Fix the 2.4 resource limitation workaround introduced in 1.26 which 5128actually broke things on mips32, sparc32, and Alpha platforms. 5129 5130Updated the I18N code so that calls to setlocate(LC_CTYPE, 0) are made 5131(which is required by the newer libintl libraries). 5132 5133Programmer's notes: 5134------------------- 5135 5136Fixed various gcc -Wall complaints. 5137 5138Fixed a few memory leaks in the e2fsck journaling code, and in the 5139ismounted code checking for a swap device. 5140 5141Add new inode I/O abstraction interface which exports an inode as 5142an I/O object. 5143 5144Exported ext2_file_flush as a public interface. 5145 5146In ext2_file_write, we now mark the buffer void * argument as a const, since ext2_file_write doesn't modify the buffer. 5147 5148Lots of small random portability fixes to make e2fsprogs build under 5149AIX --- even without the 5L compatibility toolkit, and even using the 5150uber-crippled AIX native C compiler. 5151 5152 5153E2fsprogs 1.26 (February 3, 2002) 5154================================= 5155 5156Dumpe2fs will keep going now if the bad block inode can't be read. 5157(Previously it stopped with a fatal error.) 5158 5159E2fsck will now give an opportunity to run the journal even if the 5160recovery flag is not set. This is the default behaviour if e2fsck is 5161using a backup superblock, since the needs_recovery flag will never be 5162set in the backup superblock copies. 5163 5164E2fsck now automatically finds the backup superblock/group descriptors 5165even when the primary superblock is completely destroyed for 2k and 4k 5166filesystems. (Previously it just guessed/assumed that we were dealing 5167with a 1k filesystem in that case, and users had to manually specify 5168the backup superblock number.) 5169 5170Fixed a bug in e2fsck where it dereferences a null pointer when there 5171is a problem opening a filesystem in preen mode. 5172 5173E2fsck now handles external journals correctly; previously it trashed 5174the external journal device if the journal needed to be replayed. 5175 5176Work around ulimit incompatibility problem caused by recent 2.4 5177kernels; the unix IO module will automatically try to set any resource 5178limits to be infinite on startup. 5179 5180Fixed a bug in e2fsck where it wasn't allocating a new block or inode 5181bitmap if it wasn't present and the blocksize was 2k or 4k. 5182 (Addresses Debian Bug #116975) 5183 5184E2fsck will check and fix botched ext3 V1 to V2 superblock updates by 5185clearing the new V2 fields if they do not make sense or if the ext3 5186superblock is version 1 superblock. 5187 5188E2fsck will automatically relocate the ext3 journal from a visible 5189file (i.e., /.journal) to an hidden inode if the filesystem has been 5190opened read/write. This allows the users to add a journal while the 5191filesystem is mounted, but the next time the system is rebooted, the 5192journal file will disappear. This avoids problems with backups, 5193stupid operators with superuser bits, etc. 5194 5195Fix a bug in fsck where it would not support a filesystem type of 5196"auto" if the device was specified in terms of LABEL=xxx or UUID=xxx. 5197 5198Fsck now supports fstab entries of the form "ext3,ext2". It will also 5199automatically identify reiserfs filesystems. 5200 5201The number of processes spawned by fsck can now be limited using the 5202FSCK_MAX_INST environment variable. 5203 5204Fsck now searchs the LVM proc hierarchy to find logical volumes which 5205should be searched for UUID and label information. 5206 5207Work around a bug in 2.4.10+ kernels by trying to unset the filesize 5208limit if at all possible, when opening a block device. (The filesize 5209limit shouldn't be applied against writes to a block device.) 5210 5211In mke2fs and e2fsck, specifying the -c option twice will now do 5212a read/write test on the disk. Update the man pages to encourage 5213using the -c option, and to discouraging running badblocks separately, 5214since users tend to forget to set the blocksize when running 5215badblocks. 5216 5217Mke2fs now automatically clears the immutable attribute on a 5218pre-existing .journal file left over from a previous failed attempt to 5219add a journal to an alreadyy-mounted filesystem. 5220 5221Fixed mke2fs's exit codes to consistently indicate when the mke2fs 5222operation failed. 5223 5224Mke2fs now creates the lost+found directory with permissions of 0700, 5225so that files that had previously lived in protected directory are 5226safe if they get dropped in lost+found and the system administrator 5227doesn't deal with immediately. (Addresses Debian #bug 118443) 5228 5229Mke2fs and e2fsck (and all programs that use the 5230ext2fs_check_if_mounted function) will now properly identify that a 5231device is mounted, even in cases where devfs has confused things with 5232multiple devices names with the same identity, or if a dim-witted 5233system administrator has done something stupid like creating device 5234file aliases in their home directory. Also now checks for swap devices 5235by using /proc/swaps. (Addresses Debian bug #106622) 5236 5237Added a new option (-T) to tune2fs which sets the time a filesystem 5238was last checked. 5239 5240Speed up e2image when creating sparse raw image files by optimizing 5241away excess lseek() system calls. 5242 5243Fix support of large (> 2GB) files when used as a filesystem in 5244mke2fs, tune2fs, debugfs, and findsuper. 5245 5246Debugfs's stat and icheck commands now properly deals with large (> 52472GB) files. 5248 5249Debugfs's set_super_value command now prints out the list of valid 5250superblock fields which can be set using the command. 5251 5252Debugfs's rm and kill_file command now updates the superblock free block 5253and inode counts, thus keeping the filesystem consistent. 5254 5255Debugfs's lsdel command now takes an optional argument which allows 5256the user to only see the most recently deleted files. 5257 5258A new command (undel) was added to debugfs which automates 5259undeleting a deleted inode and linking it back to a directory. 5260 5261Debugfs's ls command now takes a new option, -d, which lists 5262deleted directory entries. 5263 5264Debugfs's testb, freeb, setb, and find_free_block commands now take an 5265optional count argument. 5266 5267Add support for a new ext2 file attribute, EXT2_NOTAIL_FL, which will 5268be used to signal that a particular inode should not be eligible for 5269tail-merging --- this is needed for programs like LILO. 5270 5271The findsuper (an unreleased, uninstalled utility program) has been 5272improved to add extra validity checks and to add a progress meter. 5273(It is still an unsupported program; the officially supported way to 5274recover from a trashed partition table is to use gpart; findsuper is 5275for wizards only. :-) 5276 5277Debugfs was fixed to compile with "configure --disable-swapfs". 5278 5279Cleaned up various manual pages. (Addresses Debian bug #119624, #120171) 5280 5281Added new translation file for Turkish. 5282 5283Programmer's notes: 5284------------------- 5285 5286Fix general gcc -Wall complaints. 5287 5288E2fsprogs (mostly) works with the dietlibc. 5289 5290The programming texinfo file has been expanded to include description 5291of additional libext2fs functions. (Still not compeltely done, but a 5292lot of the more important functions have been documented.) 5293 5294Added a umask structure to struct_ext2_filsys, which currently only 5295modifies the behaviour of ext2fs_mkdir(), but if we add any file 5296creation functions to libext2fs, we should also make sure they respect 5297the umask setting. 5298 5299The build-rpm script was fixed to be a bit more safe. 5300 5301The tests' Makefile now has a way of automating test case creation 5302for e2fsck, using "make testnew". 5303 5304Created a new function, ext2fs_dir_iterate2 which passes more 5305information to the callback function (identical to the one used by 5306ext2fs_dblist_dir_iterate). The directory iterator functions take a 5307new flag, DIRENT_FLAG_INCLUDE_REMOVED, which will return deleted 5308directory entries. If the directory entry is deleted, the callback 5309function will be called with the entry paraemter set to 5310DIRENT_DELETED_FILE. 5311 5312Added new functions, ext2fs_inode_alloc_stats and 5313ext2fs_block_alloc_stats, which takes updates block/inode allocation 5314statistics in the bitmaps, block group descriptors, and superblock 5315when inodes or blocks are allocated or deallocated. 5316 5317 5318E2fsprogs 1.25 (September 20, 2001) 5319=================================== 5320 5321This is primarily a bug-fix release; no new features were added, but 5322there are a number of embarassing bug fixes and cleanups applied. 5323 5324Fix a bug mke2fs which causes stack garbage to be written to disk when 5325zapping disk sectors. (This bug could cause mke2fs to core dump for 5326some kernels, I suspect with security enhancement patches.) 5327 5328Remove unneeded #include of <linux/config.h> which was breaking 5329building e2fsprogs on the Hurd. (Addresses Debian bug #112414.) 5330 5331Updated tune2fs man page to reflect the fact that adding or removing a 5332journal doesn't require running e2fsck. 5333 5334Remove use of AC_REQUIRE from autoconf which had been used to prevent 5335AC_CANONICAL_HOST from being called twice; unfortunately this causes 5336recent autoconf to bomb out since they don't allow AC_REQUIRE to be 5337used outside of autoconf macros. Fortunately, it doesn't seem to be 5338necessary any more, anyway. 5339 5340E2fsck will now not fall back to an alternate superblock if the user 5341specifies the superblock location explicitly on the command-line. 5342This allows easier recovery from a situation where the primary 5343superblock and block groups are slightly corrupted, but the backup 5344superblocks are completely broken. 5345 5346Fix problem which caused compile_et and mk_cmds to blow up if 5347e2fsprogs was in a directory with a pathname that contained spaces. 5348 5349We are now more paranoid about checking the entry in /etc/mtab to make 5350sure the filesystem is really mounted, since some broken distributions 5351(read: Slackware) don't initialize /etc/mtab before checking non-root 5352filesystems. (Slackware also doesn't check the root filesystem 5353separately, and reboot if the root filesystem had changes applied to 5354it, which is dangerous and broken, but there's nothing I can do about 5355that.) 5356 5357Make UUID library C++ friendly by adding appropriate extern "C" 5358declarations and using const in the function declarations. 5359 5360Fix up the com_err texinfo file so that it can product a valid info 5361file (previously, it could only be used to generate paper 5362documentation using texinfo.tex). 5363 5364 5365E2fsprogs 1.24a (September 2, 2001) 5366=================================== 5367 5368Fix brown-paper bug in mke2fs which caused it to segfault when 5369printing the version string. 5370 5371 5372E2fsprogs 1.24 (August 30, 2001) 5373================================ 5374 5375Revert the BLKGETSIZE64 support, since for some distributions, this 5376ioctl has been used by an unofficial kernel patch to update the last 5377sector on the disk, and this was causing disk corruption problems as a 5378result. 5379 5380Mke2fs is now more careful about zapping swap space signatures and 5381other filesystem/raid superblock magic values so that programs like 5382mount who try to use hueristics to determine the filesystem type of a 5383partition are less likely to be confused. 5384 5385E2fsck will no longer complain if the the mode of EXT2_RESIZE_INO is a 5386regular file (since Andreas Dilger's on-line resizing tools will set 5387its mode bits). 5388 5389Fixed some minor texinfo, man pages nits for spelling errors, 5390texinfo warnings, etc. (Addresses Debian bug #110621.) 5391 5392E2fsprogs program no longer print the filesystem version number 5393(i.e. 0.5b), since it only confuses people and doesn't serve any real 5394purpose. 5395 5396E2fsck will once again compile under libc5, since it will supply its 5397own version of strnlen if necesssary. 5398 5399mke2fs and tune2fs will allow the use of UUID= or LABEL= specifiers 5400when specifying the external journal device. tune2fs will also search 5401devices looking for the externla journal debice when removing. 5402 5403E2fsprogs 1.23 (August 15, 2001) 5404================================ 5405 5406Add initial support for extended attributes (EA); e2fsck will 5407correctly handle a filesystem with EA's and check the EA blocks for 5408corruptions. 5409 5410E2fsck's symlink sanity checking has been cleaned up. It now checks 5411the i_size value of fast symlinks, and checks for immutable flags 5412being set of symlinks, etc. 5413 5414E2fsck now offers to clear inodes which are special files that have a 5415non-zero i_blocks or i_size field. (The i_size field check was in the 5416previous version of the code, but due to a bug it didn't offer to 5417clear the inode unless i_size and i_size_high were both non-zero.) 5418 5419E2image can now create "raw" image files, which only contain the 5420filesystem metadata placed in a spare file so that e2fsck, dumpe2fs, 5421debugfs, etc., can be run directly on the raw image file. 5422 5423Add support for the 64-bit block device patches. 5424 5425Fixed bugs in creating external journals with a 1k blocksize. 5426 5427Add initial support for external journals (so long as the external 5428journal only supports a single filesystem) in e2fsck. 5429 5430Remove requirement for needing to run fsck on a filesystem after 5431removing a journal (either internal or external). 5432 5433The man pages now document how to create and manage external journals. 5434 5435Speed up the check of an ext3 filesystems by avoiding a needless flush 5436of all of the superblock and block group descriptors. 5437 5438Speed up creating an internal journal using tune2fs in the case where 5439the filesystem has a lot blocks already allocated. 5440 5441Tune2fs has been fixed to make sure that only error messages go to 5442stderr, and normal message go to stdout. (Addresses Debian bug #108555) 5443 5444Fixed a minor bug in mke2fs; if -O none is passed to mke2fs, it will 5445now not set the sparse_super feature. (Addresses Debian bug #108165) 5446 5447Add support in fsck for the filesystem type "auto". 5448 5449Fsck -A will not try to interpret device names for filesystems which 5450have a pass number is 0. (Addresses Debian bug #106696). 5451 5452Fsck prints a warning message if now valid filesystems are passed to 5453it. (Addresses Debian Bug #107458.) 5454 5455E2fsck now gives an explicit warning if there filesystem still has 5456errors at the end of the run. (Addresses Debian bug #104502) 5457 5458E2fsck will set the EXT2_ERROR_FS flag if the journal superblock 5459reflects an error. E2fsck will also not run the orphan list if the 5460filesystem contains errors, since the orphan list might be corrupted. 5461 5462E2fsck now prints the number of large files when given the -v option. 5463 5464Fixed minor memory leaks in e2fsck. 5465 5466Some minor man pages updates. (Addresses Debian bug #30833, #108174) 5467 5468 5469E2fsprogs 1.22 (June 22, 2001) 5470============================== 5471 5472Fixed a bug in e2fsck's handling of orphan inodes which are special 5473files (i.e., block/character device files, named FIFO's, etc.). 5474 5475Fixed a bug accidentally introduced in e2fsprogs 1.21 which caused 5476tune2fs to fail at adding a journal to a mounted filesystem. 5477 5478Fixed a few big-endian bugs in e2fsprogs 5479 * The directory block functions were accidentally reporting 5480 some directories as corrupted when they weren't. 5481 * If e2fsprogs is compiled --disable-swapfs, the C language 5482 equivalents weren't being included for big-endian platforms. 5483 (Fixes Debian bug #101686). 5484 5485Fixed a Hurd compilation problem. (Addresses Debian bug #101361) 5486 5487Programmer's notes: 5488------------------- 5489 5490Use platform independent method of defining the BLKFLSBUF and FDFLUSH 5491ioctls, and allow them to be defined for non-i386 platforms. 5492 5493The uuid.h header file is now protected against multiple inclusions. 5494 5495E2fsprogs is now being developed using BitKeeper. Changed the test 5496scripts to deal with BK's stripping CR characgters from text files, 5497and changed the top-level Makefile.in to avoid including BitKeeper 5498files when generating the source tarball. 5499 5500 5501E2fsprogs 1.21 (June 15, 2001) 5502============================== 5503 5504Added new configure flags which allow a subset e2fsprogs to be built; 5505this is most useful for boot floopies, since the resulting shared 5506libraries and programs are slimmed down by removing features that 5507aren't necessary for a boot floppy. The new flags that were added 5508are: --disable-swapfs, which removes support for byte swapping old 5509legacy PPC/68k filesystems, --disable-debugfs, which removes support 5510for debugfs from the libext2fs library, --disable-imager, which 5511removes support for the e2image program, and --disable-resizer, which 5512removes support for resize2fs. 5513 5514E2fsck now prints the number of mounts or days elapsed since the last 5515check when e2fsck is forced to check an otherwise clean filesystem. 5516 5517Tune2fs now prints an informative message about how often a filesystem 5518will be checked when adding a journal to the filesystem, to remind the 5519user that he/she may want to adjust those parameters using tune2fs 5520-c/-i. 5521 5522Worked around hurd brain-damage which causes e2fsck to sometimes 5523believe a filesystem is the root filesystem based on device numbers 5524(since Hurd doesn't have dev_t's, which is arguably a POSIX.1 5525violation). 5526 5527Fixed a bug introduced in 1.20 which caused e2fsck to abort with an 5528erroneous error with the -F option was specified. 5529 5530Fixed a ext3 recovery bug in the revoke handling; synchronized with 5531ext3 0.7a. 5532 5533Fixed two bugs in e2fsck's handling of dup block handling, dealing 5534with relatively uncommon edge cases: a directory with an indirect 5535block which is claimed by another file, and when the last inode in the 5536filesystem has blocks claimed by another file. 5537 5538E2fsck now checks to see if the i_size field of a fast symlink is too 5539big, and offers to clear the symlink if so. 5540 5541E2fsck now checks to see if i_size_high of special files is non-zero, 5542and offers to clear i_size_high. 5543 5544Fix e2fsck's handling of incompatible journal flags so that the user 5545has chance to abort, and then has the option to clear out the journal 5546entirely. (Addresses Debian bug #98527.) 5547 5548Fixed a bug in fsck which could cause it to core dump if a mix of 5549standard and non-standard device names are used in /etc/fstab. 5550(Debian bug #100559) 5551 5552Fixed a bug in debugfs which caused read errors when copying a file to 5553not be noticed. 5554 5555The debugfs set_super_value command can now modify the s_lastcheck field. 5556 5557Fixed a bug in lsattr and chattr which was accidentally introduced in 55581.20 to support > 2GB files; both lsattr and chattr wasn't reading 5559directories correctly because the change modified the layout of struct 5560dirent to be incompatible with the libe2p shared library. 5561 5562Cleaned up the mke2fs manual page and included a discussion about why 5563it's good to periodically check the filesystem even when journaling is 5564enabled. 5565 5566Programmer's notes: 5567------------------- 5568 5569Fix general gcc -Wall complaints. 5570 5571The types needed by the ext2 header files are now provided by 5572lib/ext2fs/ext2_types.h, instead of include/asm/types.h. 5573 5574Integers are now preferred to longs when trying to find a 32-bit type 5575in ext2_types.h. Also, if linux/types.h has already been defined, 5576don't try to redefine the types. 5577 5578Fixed make depend script so that it automatically corrects the 5579pathname cleanups performed by make -M, so I don't have to fix them up 5580by hand. 5581 5582Fixed the d_loaddump test case to be more robust, and not depend on 5583bash'isms. 5584 5585Removed debugfs's dependence on pread(), which was accidentally 5586intrudced in e2fsprogs 1.20 5587 5588Fixed a performance bug in the libext2fs's icount routine; the size 5589estimate of the icount array was incorrectly being calculated. 5590 5591Removed use of the badblocks compatibility functions in the e2fsprogs 5592programs. 5593 5594Added paranoia code which protects against strange cases where /etc 5595isn't on the root filesystem, or if /etc/mtab doesn't exist. 5596 5597The header file ext2_types.h is now installed. 5598 5599Autoconf is used to determine when we are on big-endian machines, 5600instead of doing run-time tests, to save a few bytes of code. 5601 5602The ext2fs_mark_generic_bitmap and ext2fs_unmark_generic_bitmap 5603functions are no longer inline functions, which saves space and 5604doesn't really cost any real performance. 5605 5606The ext2fs library no longer depends on the e2p library. (What need 5607there was of it --- namely, fsetflags, was coded in-line). 5608 5609Fixed the makefile so that lib/ext2fs/ext2_types.h is generated even 5610when the user is stupid and tries compiling the package using "make 5611install" as root. 5612 5613Miscellaneous code cleanups: 5614 * Added missing files from Makefile.in's SRCS file, so that 5615 their dependencies would be properly calculated. 5616 * Removed redundant code 5617 * Fixed comments in code 5618 * Removed no-longer unneeded argsused #pragma. 5619 5620 5621E2fsprogs 1.20 (May 20, 2001) 5622============================= 5623 5624Add support for replaying the ext3 journal purely in user mode, 5625including handling the orphaned inode list. Used code contributed by 5626Andreas Dilger, with bug fixes and the orphaned inode handling done 5627by Theodore Ts'o. 5628 5629The mke2fs and tune2fs programs can create or modify a filesystem to 5630include an ext3 journal. Tune2fs also can be used to remove an ext3 5631journal from a filesystem. 5632 5633E2fsck will now check for the existence of a linked list of orphan 5634inodes in the superblock, and clear those inodes before starting the 5635rest of the filesystem check (but after the journal playback). 5636 5637E2fsck now validates the file descriptor passed to the -C option, 5638which saves against the completion bar getting written to an 5639unexpected location, such as the disk being checked. (Debian 5640bug/wishlist #55220) 5641 5642E2fsck will now bump the filesystem revision number from zero to one 5643if any of the compatibility bits are set. 5644 5645Fixed a bug where a badly corrupted directory on a big endian system 5646could cause e2fsck to die with a bus error. The 5647ext2fs_read_dir_block() and ext2fs_process_dir_block() functions in 5648the ext2 library now does alignment sanity checks on the rec_len field 5649of the directory entry before using it. 5650 5651The ext2 library has been enhanced to make tune2fs safe to run on 5652mounted filesystems. (Users could usually get away with using tune2fs 5653on mounted filesystems before, but with the advent of ext3 and 5654journaling, it became important to make tune2fs was *really* safe for 5655use even when the filesystem being modified is mounted.) E2label is 5656now implemented by tune2fs using an argv[0] dispatch, so that e2label 5657is also now safe for use on mounted filesystems. 5658 5659Added a new program, e2image, which creates a backup of critical ext2 5660filesystem data structures. The generated image file can be examined 5661using dumpe2fs and debugfs. In the future, e2fsck will be able to use 5662the image file to help recover very badly damaged filesystems. 5663 5664Fixed a number of LFS bugs in e2fsck; very, very large (> 2**42) files 5665no longer cause e2fsck to bomb out. Also treat files > 2GB as being 5666large file when deciding whether or not the filesystem has large files. 5667 5668Fixed lsattr and chattr so that they work correctly on large files. 5669(Fixes Debian bug #72690.) 5670 5671Removed limitation in get_device_size() which imposed a one terrabyte 5672filesystem limitation. (Most 2.2 kernels still have a signed int 5673problem which cause 1 TB block device limitation. Fortunately, the 5674kernel patches to fix this are much easier than fixing the 2TB 5675limitation in the kernel. :-) 5676 5677A max_mount_count of zero is now treated as if no mount count were 5678set. (Previously, no mount count was indicated by using -1, and a 5679mount count of zero caused e2fsck to always be run.) 5680 5681Mke2fs supports two new filesystem types largefile and largefile4. 5682 5683Mke2fs now adds some randomness in s_max_mount_count so that multiple 5684filesystems won't be all checked at the same time under normal 5685operations. 5686 5687Fixed bug in the progress bar printing code which could cause e2fsck 5688to core dump on an illegal filesystem. 5689 5690Fixed bug in fsck which could allow more than one instance of e2fsck 5691to be printing a progress bar. (Debian bug #65267) 5692 5693Fsck using a UUID or a LABEL specifier will work even if devfs is 5694compiled into the kernel and not mounted. If the pathnames in 5695/proc/partitions are incorrect, fsck will search /dev for the correct 5696device (using the new ext2fs_find_block_device library function). 5697Fsck now also checks the RAID devices first so that they are properly 5698found when they are in use. Support has also been added to support 5699additional IDE disks and the DAC 960 device names. (Debian bug #94159) 5700 5701Fixed a bug in fsck which caused it not deal properly with 16 5702byte long filesystem labels. 5703 5704Fsck's -t option has been made a lot more flexible. The semantics for 5705what happens if a comma-separated list to fsck has been regularized, 5706and it is now possible to filter what filesystems will get checked 5707based what is in the filesystem's fstab entry's option field. (Debian 5708bug #89483.) 5709 5710The dumpe2fs program can now print out the group description 5711information in hex, and also prints the location of the superblock and 5712block group descriptor for those block groups that have them. 5713 5714Mke2fs now clears the ext2 superblock before it starts creating a 5715filesystem, so that the superblock magic number is only written if the 5716filesystem creation process successfully completes. 5717 5718The debugfs program's stat command now pretty-prints the blocks used 5719by an inode so that it's more compact and informative. 5720 5721The debugfs stats command now uses the same libe2p code (which is used 5722by dumpe2fs) to print the superblock header information. This is more 5723complete, and it avoids a bit of code duplication. 5724 5725Added a new debugfs command, set_super_value (ssv) which allows the 5726user to set arbitrary superblock fields. 5727 5728Debugfs was extended to support inode numbers in hex (by prefixing 5729them with 0x), and so that modify_inode can set the inode generation 5730number. Also, there is now a new function command called logdump 5731which will dump an ext3 journal. 5732 5733Fixed a bug in debugfs so that quitting out of the pager doesn't kill 5734debugfs. 5735 5736Debugfs's dump command now stops immediately upon reporting a disk 5737read error. (Fixed a bug in ext2fs_file_read library routine which 5738caused debugfs not to stop.) (Debian bug #79163) 5739 5740On systems with /proc/mounts (mainly Linux systems), /proc/mounts is 5741checked before /etc/mtab is used, since /proc/mounts is more likely to 5742be accurate. 5743 5744Added portability fixes for Solaris and Linux/ia64. 5745 5746Various manual pages were clarified and cleaned up. (Fixed debian 5747bugs #63442, #67446, and #87216) 5748 5749 5750Programmer's notes: 5751------------------- 5752 5753The e2fsck message printer now supports %Iu and %Ig, which will print 5754out the inode's user and group owners, respectively. 5755 5756E2fsprogs now includes its own version of include/linux/ext2_fs.h, so 5757that no longer dependent on the system having the correct version of 5758the kernel header files. 5759 5760Added a new function to libext2, ext2fs_find_block_device(), which 5761searches the system (i.e., /dev, /devfs, /devices) for a pathname to a 5762device given its device number. 5763 5764Added a new function to libext2, ext2fs_sync_device, which centralizes 5765all of the places which might try to use the BLKFLSBUF or FDFLUSH 5766ioctls (and usually failing to define them since the system header 5767files don't usually do this for us, and we're trying to avoid usage of 5768kernel include files now). 5769 5770Added new utility programs in tests/progs: random_exercise and 5771hold_inode. They aren't built by default; they're useful for 5772exercising ext3 filesystem code. 5773 5774Added a new ext2 filesystem flag, EXT2_FLAG_SUPER_ONLY, which causes 5775the filesystem close functions to only update the superblock, and to 5776not touch the block group descriptors. Needed by tune2fs when 5777modifying a mounted filesystem. 5778 5779Got rid of struct ext2fs_sb and replaced it with the standard struct 5780ext2_super_block from include/linux/ext2_fs.h. Note: this may break 5781source (but not binary) compatibility of some users of the ext2 5782library. Those applications should just simply do a global search and 5783replace of struct ext2fs_sb with struct ext2_super_block, and include 5784the new header file <ext2fs/ext2_fs.h> which defines it. 5785 5786The ino_t type has been renamed ext2_ino_t to protect applications 5787that attempt to compile -D_FILE_OFFSET_BITS=64, since this 5788inexplicably changes ino_t(!?). So we use ext2_ino_t to avoid an 5789unexpected ABI change. 5790 5791The Makefiles have been reworked so that "make check" can be run from 5792the top-level Makefile. 5793 5794Fix general gcc -Wall complaints and removed dead code. 5795 5796Remove use of NOARGS, because we assume everyone does ANSI C these 5797days. 5798 5799Added build-rpm script from sct. 5800 5801New functions ext2fs_image_{inode,super,bitmap}_{read,write} added 5802to support e2image. 5803 5804New function ext2fs_flush_icache which must be called if the 5805application program modifies the inode table blocks without going 5806through ext2fs_write_inode() interface. 5807 5808New ext2fs_check_mount_point() function, which will return the mount 5809point of a device if mounted. 5810 5811The io_channel abstraction now has an optional interface, 5812io_channel_write_range, which allows specific byte ranges to be 5813written. 5814 5815The unix_io IO channel now supports write-through caching, so that 5816journal creation is more efficient. 5817 5818Added x86 assembly language routines to support byte swapping, to 5819reduce executable size. 5820 5821Fixed bug in the utility program subst so that it's possible to 5822replace a substitution variable with a zero-length string. 5823 5824Fixed numbering e2fsck pass1 problem numbers; an extra zero had 5825slipped into some of the problem number. 5826 5827E2fsprogs 1.19 (July 13, 2000) 5828============================== 5829 5830Release the resize2fs program since the timeout before it could 5831be released under the GPL has finally expired. 5832 5833Add experimental support needed for the ext2 compression patches. 5834This requires compiling e2fsprogs with the --enable-compression flag 5835to the configure script. 5836 5837Added ext3 journalling support. E2fsck will run the journal (if 5838necessary) by temporarily mounting the filesystem. /sbin/fsck.ext3 is 5839installed as a symlink to e2fsck. Fsck has been taught about ext3, 5840and treats it the same as ext2 in terms of the progress bar logic. 5841Dumpe2fs will display the superblock journaling information if the 5842filesystem has a journal. The ext2 library will now permit opening an 5843ext3 filesystem with the recovery flag set. This is necessary for 5844on-line dump's to work correctly, but there may be issues with this 5845working well since ext3 is much less agressive about syncing blocks to 5846the filesystem, since they're safe on the journal. 5847 5848Tune2fs and e2fsck have been changed to allow the mount_count check to 5849be disabled by setting max_mount_count to -1. (This was already 5850supported by the kernel.) 5851 5852Create a symbolic link for fsck.ext3, since the e2fsprogs utilities 5853are used for ext3 as well. 5854 5855Added internationalization support for e2fsprogs; must be enabled 5856by passing --enable-nls to configure. 5857 5858Always use the provided ext2fs header files to insulate ourselves from 5859kernel version changes. Which include files are used by e2fsprogs 5860have also been cleaned up to improve portability. 5861 5862Limit the number of times that e2fsck updates the progress bar so that 5863people who are booting using a 9600 baud console don't get swampped by 5864too many updates. 5865 5866Improved the loop detection algorithm in e2sck's pass #3 so that it is 5867much, much faster for large filesystems with a large number of 5868directories. 5869 5870The memory footprint for e2fsck is now slightly smaller than before. 5871 5872E2fsck now checks if special devices have a non-zero size, and offers 5873to clear the size field if it finds such an inode. 5874 5875E2fsck now checks if special devices have the append-only flag set, 5876and offers to clear the inode. 5877 5878E2fsck now properly handles some "should never fail" cases during a 5879bitmap copy in pass5. 5880 5881E2fsck now properly prints control characters in filenames as ^A .. ^Z. 5882 5883E2fsck now calculates the correct location of the backup superblock in 5884the case of filesystem blocksizes > 1k. 5885 5886Fixed a bug in e2fsck's calculation of the number of inodes_per_block 5887which normally didn't cause problems under most filesystem parameters, 5888but could cause a valid superblock to be rejected in extreme cases. 5889Other checks for validating superblock values were made more 5890stringent. 5891 5892Added non-destructive write testing to the badblocks program, courtesy 5893of David Beattie. The badblocks also now has an option to input the 5894current set of bad blocks, so that known bad blocks are skipped to 5895speed up the badblocks test. There is also a persistent rescan 5896feature which causes badblocks to run until it has completed some 5897number of passes without discovering any new bad blocks. 5898 5899Badblocks now checks to see if the device is mounted and refuses to do 5900the tests involving writing to the device if it is mounted. Also, 5901badblocks now allows the number of blocks to be checked to be 5902defaulted to the size of the partition. 5903 5904Fixed a bug in fsck which didn't allow non-root users to be able to 5905check filesystems if there were any LABEL= or UUID= entries in 5906/etc/fstab. 5907 5908The Hurd doesn't support the filetype filesystem feature. The mke2fs 5909program now makes sure that for the Hurd, the filestype feature is 5910turned off. E2fsck will check to see if the filetype feature is 5911turned on for Hurd filesystems, and offer to turn off the feature. 5912 5913Mke2fs now has a safety check to make sure the number of blocks do not 5914exceed 32 bits even on a 64 bit platform. 5915 5916Really fixed a bug in fsck to allow "fsck -As" to run interactive 5917fsck's. (For those people who like to do interactive fsck's in the 5918/etc/rc scripts!?!) 5919 5920Debugfs has a few new features: the rdump command, which will do a 5921recursive dump of a directory and all of its contents, and the lcd 5922command which does a local chdir (much like the ftp command of the 5923same name). In addition, the debugfs program and the open_filesystem 5924command now takes three new options: -b and -s, which allows the 5925blocksize and superblock location to be specified, and the -c option 5926which is used in catastrophic situations where the block group 5927descriptors are corrupt. If the -c option is specified, debugfs will 5928skip trying to read in the block and inode bitmaps. 5929 5930Debufs's lsdel command was fixed to handle bad blocks in the inode 5931table. 5932 5933A Y2K bug in debugfs's "ls -l" handling was fixed by switching to use 59344 digit years. 5935 5936General improvements in error messages 5937 5938 - Mke2fs prints a sane error message if the partition size is zero 5939 (usually because the partition table wasn't reread by the 5940 kernel due to the partition being busy), instead of "invalid 5941 argument passed to ext2 library while initializing superblock". 5942 5943 - Fsck now prints more self-explanatory message if an invalid UUID= 5944 or LABEL= specification is passed to it. 5945 5946UUID library changed to use the LGPL. 5947 5948Fixed a bug in the UUID library where very rapid calls to the 5949time-based UUID generator could cause duplicate UUID's to be returned. 5950This was not a problem for e2fsprogs, but it could be a problem for 5951other users of the library. 5952 5953Make the UUID library more robust in the face of missing or an 5954improper /dev/urandom or /dev/random files. 5955 5956Added some random portability fixes for Solaris. 5957 5958Some minor man page updates. 5959 5960Fixed a memory leak in the ss library. 5961 5962 5963Programmer's notes: 5964------------------- 5965 5966We now try to use lseek64 and open64 from the LFS if possible. 5967 5968The 3rd parameter in e2p's print_flags is now a flags word, instead of 5969a boolean option. 5970 5971The mark and unmark bitmap functions now return the previous state of 5972the bit that was being changed, which is useful for some speed 5973optimizations. 5974 5975The following functions have been added to enhance the badblocks list 5976handling in libext2fs: ext2fs_write_bb_FILE, ext2fs_read_bb_FILE2, and 5977ext2fs_badblocks_equal. 5978 5979The ext2 header files now have the latest journalling fields to the 5980superblock. 5981 5982The ext2fs_mkdir function in libext2fs now properly backs out of error 5983conditions robustly. 5984 5985Cleaned up makefiles: 5986 - to cleanly compile with the -j flag. 5987 - so distclean removes all generated files. 5988 - so in case of an error while installing header files, the make aborts. 5989 5990Fix test_script so that it works correctly when compiling in the 5991source directory. 5992 5993The random UUID generation routine has been made slightly better in 5994the case where /dev/random doesn't exist. (Use of randomly-based UUID 5995is still not recommended if /dev/random doesn't exist, however; it's 5996better to use the time/ethernet MAC address UUID in this case.) 5997 5998Clean up the build process so it's more friendly in case of missing 5999directories. 6000 6001The ext2fs header file can now be #include'd into C++ programs. 6002 6003The e2p.h header file is now installed. 6004 6005Added workaround to a gawk 3.0.5 bug in lib/ss/mk_cmds. 6006 6007 6008 6009E2fsprogs 1.18 (November 10, 1999) 6010================================== 6011 6012Fix a core dumping bug in e2fsck if an imagic inode is present or 6013(more rarely) if the filesystem is badly corrupted enough that e2fsck 6014has to restart pass 1 processing. E2fsck now closes the filesystem 6015before freeing a large number of its data structures, so in the case 6016of future memory faults, at least the fixed filesystem will be fully 6017written out. 6018 6019If a filesystem doesn't support imagic inodes, and e2fsck discovers an 6020imagic inode, it will offer to clear the imagic flag. 6021 6022E2fsck will now offer to clear the immutable flag on special files 6023(device/socket/fifos) when running it in non-preen mode. 6024 6025E2fsck will now set the filetype when creating /lost+found, and when 6026connected orphaned inodes to /lost+found. 6027 6028Debugfs's ncheck and icheck commands now handles the case where there 6029are bad blocks in the inode table without bombing out. 6030 6031The badblocks list processing code has been made more efficiently for 6032appending a large number of (ordered) badblocks to the badblocks list. 6033 6034Some minor man page updates. 6035 6036Fsck now allows interactive e2fsck's when using fsck -As (not a common 6037mode, but some people like to do this in boot scripts for silly reasons). 6038 6039Programmer's notes: 6040------------------- 6041 6042The internal e2fsck problem code for PR_2_SPLIT_DOT was fixed to meet 6043with the problem code convention. 6044 6045The badblocks list regression test program has been updated to work 6046with previously made API name changes. 6047 6048The ext2fs_free() command now uses the new badblocks API to avoid 6049using the compatibility layer. 6050 6051Added new regression test cases; the run_e2fsck test script now 6052supports the ability for a test case to run a prepratory command 6053before running e2fsck. 6054 6055E2fsprogs 1.17 (October 26, 1999) 6056================================= 6057 6058Fixed nasty typo in fsck which caused parallelized fsck's to go into an 6059infinite loop. 6060 6061Fixed a bug in fsck where it used strncmp to compare a binary UUID, 6062thus potentially causing problems if a binary UUID contained a NULL 6063character. 6064 6065E2fsck now uses stricter checks for directory entries in pass 2: 6066zero-length filenames are not allowed; neither are 8 byte long 6067directory entries. 6068 6069The debugfs "dirty" command now clears the filesystem valid bit. 6070(Previously this just set the dirty-as-in-needs-writing-out-to-disk 6071bit in the in-core superblock image. The new functionality is more 6072what the user expects, and is more useful.) 6073 6074Added a debugging hook to test parallel fsck; if the environment 6075variable FSCK_FORCE_ALL_PARALLEL, then filesystems on the same drive 6076will be checked in parallel when they normally would not be. 6077 6078Programmer's notes: 6079------------------- 6080 6081Fixed some #ifdef's for compilation under the Hurd OS. 6082 6083Fixed minor W2K compatibility problems. 6084 6085Fixed some miscellaneous GCC warnings. 6086 6087 6088E2fsprogs 1.16 (October 22, 1999) 6089================================= 6090 6091Fixed a race condition bug in fsck; when printing a progress bar, if 6092checking multiple filesystems in parallel, it was possible for fsck to 6093send e2fsck a SIGUSR1 signal before e2fsck had installed its signal 6094handler, which would cause it to terminate with a signal 10. 6095 6096E2fsck now properly handles filesystems that have the 6097INCOMPAT_FILETYPE feature turned on. It can be used to convert a 6098filesystem into using or not using FILETYPE feature. 6099 6100E2fsck now properly handles filesystems that have the IMAGIC feature 6101turned on (this is used on Linux AFS servers). 6102 6103The mke2fs program now creates filesystems that have the filetype and 6104sparse_superblock features enabled by default, unless it is run on a 6105pre-2.2 kernel. These features are not supported by a pre-2.2 kernel, 6106so there is now a new flag -O which allows the user to specify with 6107which features she would like to create the filesystem; "mke2fs -O 6108none" will create a filesystsem compatible with 2.0 kernels. 6109 6110The tune2fs program now has a -O option which allows the user to set 6111and reset "safe" filesystem features. Currently, the only ones which 6112allows to be modified are the filetype and sparse_superblock features. 6113Note setting or clearing either feature will require running e2fsck on 6114the filesystem afterwards. (n.b. Clearing the sparse_superblock feature 6115requires that there is enough free space on the filesystem for the 6116extra superblocks which will be created by e2fsck.) 6117 6118Debugfs can now set and print filesystem features in the superblock 6119using the "features" command. Dumpe2fs will print out the complete 6120set of features when listing the superblock. 6121 6122Dumpe2fs has new options -f (force) and -h (header-only). 6123 6124Fixed a bug in e2fsck which could cause the PROGRAMMING ERROR/bonehead 6125message to come up. This could happen when decrementing or 6126incrementing a link count could result in an overflow. 6127 6128Fixed a bug in e2fsck where the block count on the lost+found 6129directory would not be properly incremented when the directory was 6130expanded to the point where an indirect block needed to be allocated. 6131 6132E2fsck now makes some additional sanity checks on the superblock to 6133avoid crashing or giving a memory allocation error if some of the 6134values in the superblock are unresonable (but the superblock otherwise 6135looks valid). 6136 6137Fixed a bug in e2fsck where a very badly corrupted filesystem might 6138require two passes to completely fix the filesystem. This happened if 6139an inode claimed blocks that was part of the filesystem metadata 6140(typically, when garbage was written into an inode table or indirect 6141block, since this kind of filesystem corruption normally doesn't 6142happen otherwise). 6143 6144On the Alpha, glibc declares st_flags although it isn't actually used; 6145the configure script was improved to detect this case so that 6146e2fsprogs can avoid using the non-functional stat field. 6147 6148The manual pages were updated to use a more consistent formatting 6149style consistent with standard Unix man pages. Mke2fs's man page 6150added documentation for a few previously undocumented options. 6151 6152Fixed minor display bugs in tune2fs and mke2fs. 6153 6154Programmer's notes: 6155------------------- 6156 6157Improved portability of e2fsprogs to non-Unix systems (in particular, NT). 6158 6159Added features to parse and print feature strings into the e2p library. 6160(e2p_feature2string, e2p_string2feature, e2p_edit_feature). 6161 6162ext2fs_mkdir() and ext2fs_new_dir_block() now creates directories 6163whose directory entries contain proper filetype information if the 6164filesystem supports it. 6165 6166ext2fs_link() now uses the low 3 bits of its flags parameter to pass 6167the directory entry filetype information. This is used to set the 6168directory entry filetype information if the filesystem supports it. 6169 6170Fixed a bug in ext2fs_expand_dir() where the block count in a 6171directory's inode would not be properly incremented when the directory 6172was expanded to the point where an indirect block needed to be 6173allocated. 6174 6175 6176E2fsprogs 1.15 (July 18, 1999) 6177============================== 6178 6179Add configuration checks so that e2fsprogs will compile cleanly on 6180Linux 2.3 kernels that have renamed i_version to i_generation. 6181 6182E2fsck now prints a progress/completion bar (and not just a simple 6183spinner) if the -C0 option is requested or if it receives a SIGUSR1 6184signal. Fsck will automatically manage the (potentially muliple) 6185e2fsck processes to print completion bars if it is given a -C option, 6186with the right thing happening if multiple filesystems are being 6187checked in parallel. 6188 6189Mke2fs now has better automatic hueristics to determine the filesystem 6190parameters to be used for a particular filesystem. Added a new option 6191-T which allows the user to specify how the filesystem is to be used, 6192which helps mke2fs do a better job selecting the filesystem parameters. 6193 6194Mke2fs now creates revision 1 filesystems by default, and with the 6195sparse superblock feature enabled. The sparse superblock feature is 6196not understood by Linux 2.0 kernels, so they will only allow read-only 6197mounts of filesystems with this sparse superblocks. 6198 6199Fix bug where if /dev/null couldn't be opened (should never happen), 6200e2fsck would hang in a tight loop. 6201 6202Make e2fsck handle the case where /lost+found isn't a directory. 6203 6204E2fsck now uses mallinfo if it exists to get accurate statistics about 6205its memory usage. 6206 6207Fix bug in e2fsck where it wouldn't check to see if a disconnected 6208inode had any problems before connecting it to /lost+found. 6209 6210Add check to e2fsck so it makes sure that total number of inodes in 6211the filesystem is a sane number. 6212 6213Fix fencepost error when clearing an the end of the block bitmap which 6214caused the last block in the bitmap not to get cleared. 6215 6216Cleaned up a number of messages in e2fsck: 6217 * The message "Group's #'s copy of the group descriptor..." 6218 was fixed so that the correct number would be displayed. 6219 * Added missing space in the "disk write-protected" error messsage 6220 * Cleaned up the error message printed when a non-interactive 6221 e2fsck needs to abort a check because the filesystem 6222 appears to be mounted. 6223 6224Added a new command-line utility, uuidgen, which will create and print 6225a UUID. 6226 6227Make debugfs's icheck command more robust by checking to make sure an 6228inode has valid blocks before interarting over the inode's blocks. 6229 6230UUID generation now uses a random-based scheme whenever possible to 6231prevent potential privacy problems. 6232 6233Man pages for all of the UUID functions in the lirbary were added. 6234 6235Fixed bug in fsck so it won't coredump if a filesystem not in 6236/etc/fstab is given to it. 6237 6238Fsck now understands the UUID=xxxx and LABEL=yyyy forms in /etc/fstab 6239that most of the other mount utilities understands. 6240 6241Mke2fs will make a filesystem even if it appears mounted if the force 6242option is given. 6243 6244Dumpe2fs has new command-line options which allow a filesystem expert 6245to specify the superblock and blocksize when opening a filesystem. 6246This is mainly useful when examining the remains of a toasted 6247filesystem. 6248 6249The badblocks program has been updated to display correctly on disks 6250with large block numbers. 6251 6252The badblocks program no longer gives spurious errors when errors 6253occur on non-block boundaries, which is common if the blocksize is 6254larger than 1k. 6255 6256Mke2fs will sync the disk device every MKE2FS_SYNC block groups if the 6257MKE2FS_SYNC environment variable is set. This is to work around a VM 6258bug in the 2.0 kernel. I've heard a report that a RAID user was able 6259to trigger it even using a 2.2 kernel, but hopefully it will not be 6260needed for most Linux 2.2 users. 6261 6262Fixed miscellaneous documentation and man pages. 6263 6264Programmer's notes: 6265------------------- 6266 6267Cleaned up functions such as pass1_get_blocks, pass1_read_inode which 6268in e2fsck's pass1.c really should have been static. 6269 6270The return value of the uuid_compare() function was changed to make it 6271match with the convetions used by strcmp, memcmp, and Paul Leach's 6272UUID sample document. 6273 6274The "make depend" process has now been made more automated; it now 6275automatically word-wraps the dependencies, and only replaces source 6276Makefile.in if there has been a change in the dependencies. Also, a 6277top-level "make depend" now recurses through all the subdirectories 6278automatically. 6279 6280The Makefile in .../util has been changed so that subst is built using 6281the native C compiler during a cross-compilation, since the subst 6282program is only used during the build process. Also add an explicit 6283rule to build util/subst by cd'ing to the correct directory and 6284running Makefile. 6285 6286The man directories are defined in terms mandir, so that the configure 6287script can override the location of the manual pages. 6288 6289The config files have been updated to recognize new machine types for 6290both the i386 and alpha families. 6291 6292Fsck has been modified so that it will accurately create an 6293fsck_instance even when the noexecute flag is set. This allows for 6294accurate debugging of the fsck pass structure. Also, when the verbose 6295flag is given twice, fsck will print debugging information about when 6296fsck is waiting for jobs to finish. 6297 6298 6299E2fsprogs 1.14 (January 9, 1999) 6300================================ 6301 6302Fix the fstab parsing code so that it can handle blank lines and 6303comment characters. Also, missing pass numbers need to be treated as 6304zero. 6305 6306Fixed a bug in e2fsck where under some circumstances (when e2fsck 6307needs to restart processing after fixing an egregious inconsistency) 6308it would try to access already freed memory. 6309 6310E2fsck now prints non-printable characters in directory entries and 6311pathnames using '^' and 'M-' notation. 6312 6313Fixed chattr so that it will ignore symbolic links when doing 6314recursive descent traversals. For both chattr and lsattr, no longer 6315print the version string unless the -V option is given. 6316 6317Allow the system administrator to directly specify the number of 6318inodes desired in the filesystem, for some special cases where this is 6319necessary. 6320 6321Fix portability problems so that e2fsprogs can be compiled under Linux 63221.2 systems and Solaris systems. 6323 6324Update the config.guess file with a more recent version that will 6325identify newer Linux platforms. 6326 6327Programmer's notes 6328------------------ 6329 6330Ext2fs_read_inode and ext2fs_write_inode will now return an error if 6331an inode number of zero is passed to them. 6332 6333E2fsprogs 1.13 (December 15, 1998) 6334================================== 6335 6336Fixed a bug in debugfs where an error messages weren't getting printed 6337when the ext2 library routines to read inodes returned errors in the 6338stat, cmri and rm commands. 6339 6340Fixed a bug in mke2fs so that if a ridiculous inode ratio parameter is 6341provided, it won't create an inode table smaller than the minimum 6342number of inodes required for a proper ext2 filesystem. 6343 6344Fsck now parses the /etc/fstab file directly (instead of using 6345getmntent()), so that it can distinguish between a missing pass number 6346field and pass number field of zero. This caused problems for 6347diskless workstations where all of the filesystems in /etc/fstab have 6348an explicit pass number of zero, and fsck could not distinguish this 6349from a /etc/fstab file with missing pass numbers. 6350 6351E2fsck will create a /lost+found directory if there isn't one in the 6352filesystem, since it's safer to create the lost+found directory before 6353it's needed. 6354 6355Fixed e2fsck so that it would detect bogus immutable inodes which 6356happen to be sockets and FIFO files, and offer to clear them. 6357 6358If a filesystem has multiple reasons why it needs to be checked, and 6359one of the reasons is that it is uncleanly mounted, e2fsck will print 6360that as the reason why the filesystem is being checked. 6361 6362Cleaned up the output routines of mke2fs so that it doesn't overflow 6363an 80 column display when formating really big filesystems. 6364 6365Added a sanity check to e2fsck to make sure that file descriptors 0, 63661, 2 are open before opening the hard disk. This avoids a problem 6367where a broken program might exec e2fsck with those file descriptors 6368closed, which would cause disastrous results if the kernel returns a 6369file descriptor for the block device which is also used by FILE * 6370stdout. 6371 6372Fixed up the e2fsck progress reporting functions so that the values 6373reliably reach 100% at the completion of all of the e2fsck passes. 6374 6375Fixed minor documentation bugs in man pages and usage messages. 6376 6377Programmer's notes: 6378------------------- 6379 6380Fixed a number of lint warnings in the ext2fs library and potential 6381portability problems from other OS's header files that might define 6382CPP macros for names like "max" and "min". 6383 6384ext2fs_badblocks_list_add() has been made more efficient when it needs 6385to grow the bad blocks list. 6386 6387Fixed a bug in e2fsck which caused it to dereference a freed pointer 6388just before exiting. 6389 6390Fixed the substition process for generating the mk_cmds and compile_et 6391scripts so that they will work outside of the build tree. 6392 6393Add sanity check to e2fsck so that if an internal routine 6394(ext2fs_get_dir_info) returns NULL, avoid dereferencing the pointer 6395and causing a core dump. This should never happen, but... 6396 6397E2fsprogs 1.12 (July 9, 1998) 6398================================== 6399 6400E2fsprogs now works with glibc (at least with the version shipped wtih 6401RedHat 5.0). The ext2fs_llseek() function should now work even with 6402i386 ELF shared libraries and if llseek() is not present. We also 6403explicitly do a configure test to see if (a) llseek is in libc, and 6404(b) if llseek is declared in the system header files. (See standard 6405complaints about libc developers don't understand the concept of 6406compatibility with previous versions of libc.) 6407 6408The ext2fs library now writes out the block group number in each of 6409the superblock copies. This makes it easier to automatically 6410determine the starting block group of the filesystem when the block 6411group information is trashed. 6412 6413Added support for the EXT2_FEATURE_INCOMPAT_FILETYPE feature, 6414which means that e2fsprogs will ignore the high 8 bits of the 6415directory entry's name_len field, so that it can be used for other 6416purposes. 6417 6418Added support for the EXT2_FEATURE_RO_COMPAT_LARGE_FILE feature. 6419E2fsprogs will now support filesystems with 64-bit sized files. 6420 6421Added support for the EXT2_FEATURE_COMPAT_DIR_PREALLOC feature. 6422 6423Added new program "e2label", contributed by Andries Brouwer. E2label 6424provides an easy-to-use interface to modify the filesystem label. 6425 6426Fixed bug so that lsattr -v works instead of producing a core dump. 6427 6428Fixed a minor bug in mke2fs so that all groups with bad superblock 6429backup blocks are printed (not just the first one). 6430 6431Mke2fs will check the size of the device, and if the user specifies a 6432filesystem size larger than the apparent size of the device it will 6433print a warning message and ask if the user wants to proceed. 6434 6435E2fsck has a new option -C, which sends completion information to the 6436specified file descriptor. For the most part, this is intended for 6437programs to use, although -C 0 will print a spinning character to the 6438stdout device, which may be useful for users who want to see something 6439happening while e2fsck goes about its business. 6440 6441Fixed a bug in e2fsck which could cause a core dump when it needs to 6442expand the /lost+found directory, and sometimes the bitmaps haven't 6443been merged in. Also fixed a related bug where ext2fs_write_dir_block 6444was used to write out a non-directory block. (Which would be bad on a 6445non-Intel platform with byte swapping going on.) 6446 6447Fixed bug in e2fsck where it would print a "programming error" message 6448instead of correctly identifying where a bad block was in used when 6449the bad block was in a non-primary superblock or block group 6450descriptor. Also fixed a related bug when sparse superblocks are in 6451use and there is a bad block where a superblock or block group 6452descriptor would have been in a group that doesn't include a 6453superblock. 6454 6455Fixed a bug in e2fsck (really in libext2fs's dblist function) where if 6456the block group descriptor table is corrupt, it was possible to try to 6457allocate a huge array, fail, and then abort e2fsck. 6458ext2fs_get_num_dirs() now sanity checks the block group descriptor, 6459and subsitutes reasonable values if the descriptors are obviously bogus. 6460 6461If e2fsck finds a device file which has the immutable flag set and the 6462i_blocks beyond the normal device number are non-zero, e2fsck will 6463offer to remove it, since it's probably caused by garbage in the inode 6464table. 6465 6466When opening a filesystem, e2fsck specially checks for the EROFS error 6467code, and prints a specific error message to the user which is more 6468user friendly. 6469 6470If the filesystem revision is too high, change the e2fsck to print 6471that this is either because e2fsck is out of date, or because the 6472superblock is corrupt. 6473 6474E2fsck now checks for directories that have duplicate '.' and '..' 6475entries, and fixes this corruption. 6476 6477E2fsck no longer forces a sync of the filesystem (with attendant sleep 6478calls) at all times. The ext2fs_flush() function now performs a sync 6479only if it needed to write data blocks to disk. 6480 6481Fixed a minor bug in e2fsck's pass1b's file cloning function, where 6482certain errors would not be properly reported. 6483 6484Updated and expanded a few points in the man pages which users 6485complained wheren't explicit enough. 6486 6487Added special case byte-swapping code if compiling on the PowerPC, to 6488accomodate the strange big-endian variant of the ext2 filesystem that 6489was previously used on the PowerPC port. 6490 6491 6492Programmer's notes: 6493------------------- 6494 6495Removed C++ keywords from the ext2fs libraries so that it could be 6496compiled with C++. 6497 6498E2fsck's internal organization has now been massively reorganized so 6499that pass*.c don't have any printf statements. Instead, all problems 6500are reported through the fix_problem() abstraction interface. E2fsck 6501has also been revamped so that it can be called as a library from a 6502application. 6503 6504Added new fileio primitives in libext2fs for reading and writing 6505files on an unmounted ext2 filesystem. This interface is now used by 6506debugfs. 6507 6508Added a new libext2fs function for mapping logical block numbers of 6509a file to a physical block number. 6510 6511Added a new libext2fs function, ext2fs_alloc_block(), which allocates 6512a block, zeros it, and updates the filesystem accounting records 6513appropriately. 6514 6515Added a new libext2fs function, ext2fs_set_bitmap_padding(), which 6516sets the padding of the bitmap to be all one's. Used by e2fsck pass 5. 6517 6518The libext2fs functions now use a set of memory allocation wrapper 6519functions: ext2fs_get_mem, ext2fs_free_mem, and ext2fs_resize_mem, 6520instead of malloc, free, and resize. This makes it easier for us to 6521be ported to strange environments where malloc, et. al. aren't 6522necessarily available. 6523 6524Change the libext2fs fucntion to return ext2-specific error codes 6525(EXT2_DIR_EXISTS and EXT2_DB_NOT_FOUND, for example) instead of using 6526and depending on the existence of system error codes (such as EEXIST 6527and ENOENT). 6528 6529Renamed io.h to ext2_io.h to avoid collision with other OS's header 6530files. 6531 6532Add protection against ext2_io.h and ext2fs.h being included multiple 6533times. 6534 6535The types used for memory lengths, etc. have been made more portable. 6536In generla, the code has been made 16-bit safe. Added Mark 6537Habersack's contributed DOS disk i/o routines. 6538 6539Miscellaneous portability fixes, including not depending on char's 6540being signed. 6541 6542The io_channel structure has a new element, app_data, which is 6543initialized by the ext2fs routines to contain a copy of the filesystem 6544handle. 6545 6546ext2fs_check_directory()'s callback function may now return the error 6547EXT2_ET_CALLBACK_NOTHANDLED if it wishes ext2fs_check_directory() to 6548really do the checking, despite the presence of the callback function. 6549 6550 6551E2fsprosg 1.11 (June 17, 1997) 6552============================== 6553 6554Fixed e2fsck to detect (previously ignored) conflicts between the 6555superblock or block group descriptors and block bitmaps, inode 6556bitmaps, and inode tables. 6557 6558Fixed bug in e2fsck so that when the message printed out when a block 6559or inode bitmap conflicts with other data, it has the correct group 6560number. 6561 6562Fixed bug in e2fsck and mke2fs where the blocksize wasn't being passed 6563to badblocks. This meant that not all of the filesystem was being 6564tested for bad blocks! 6565 6566Fixed an array boundary overrun case which cropped up in 6567ext2fs_badblocks_list_test when a user tried running "mke2fs -c 6568-b 4096". 6569 6570Adjusted the number of columns printed by mke2fs when displaying the 6571superblock backups to avoid running over 80 columns when making a 6572really big filesystem. 6573 6574Fixed up the man pages for e2fsck, debugfs, badblocks, chattr, 6575dumpe2fs, fsck, mke2fs, and tune2fs (typos and other minor grammar 6576fixes), thanks to some suggestions from Bill Hawes (whawes@star.net). 6577 6578Programmer's notes: 6579------------------- 6580 6581Fixed install rule in lib/ss so that ss_err.h is actually getting 6582installed. 6583 6584Fixed bug in ext2fs_copy_bitmap; the destination bitmap wasn't getting 6585bassed back to the caller. 6586 6587Fixed bug in ext2fs_inode_scan_goto_blockgroup; it had not been 6588setting the current inode number (which meant this function wasn't 6589working at all). 6590 6591Fixed bug in ext2fs_resize_generic_bitmap; it had not be zeroing all 6592blocks in the bitmap when increasing the size of the bitmap. 6593 6594Changed the initial number of blocks allocated by ext2fs_init_dblist() 6595to be more realistic. 6596 6597Added a new function ext2fs_allocate_group_table, which sets up the 6598group descriptor information (and allocates inode and block bitmaps, 6599and inode tables for a particular group). The function was created by 6600factoring out code form ext2fs_allocate_tables(). 6601 6602Added a new function ext2fs_move_blocks which takes a bitmap of the 6603blocks to be moved, and moves them to another location on the 6604boardboard. 6605 6606Make the unix_io channel's io_channel_flush implementation calls sync() 6607to to flush the kernel buffers to disk. 6608 6609Added a new function ext2fs_dblist_count returns the number of 6610directory blocks in dblist. 6611 6612 6613E2fsprogs 1.10 (April 24, 1997) 6614=============================== 6615 6616Mke2fs once again defaults to creating revision #0 filesystems, since 6617people were complaining about breaking compatibility with 1.2 kernels. 6618Warning messages were added to the mke2fs and tune2fs man pages that 6619the sparse superblock option isn't supported by most kernels yet (1.2 6620and 2.0 both don't support parse superblocks.) 6621 6622Added new flag to mke2fs, -R <raid options>, which allows the user to 6623tell mke2fs about the RAID configuration of the filesystem. Currently 6624the only supported raid option is "stride" which specifies the width 6625of the RAID stripe. 6626 6627Fixed bug in e2fsck where pass1b would bomb out if there were any 6628blocks marked bad in the inode table. 6629 6630Fixed rare bug in mke2fs where if the user had a very unlucky number 6631of blocks in a filesystem (probability less than .002) the resulting 6632filesystem would be corrupt in the last block group. 6633 6634Fixed bug where if e2fsck tried to allocate a block to fix a 6635filesystem corruption problem and the filesystem had no free blocks, 6636ext2fs_new_block() would loop forever. 6637 6638The configure script now checks explicitly to see if "-static" works, 6639since that can't be assumed to be true --- RedHat doesn't install 6640libc-static by default. 6641 6642Fixed bug in libext2's block iterator functions where under some 6643cirmcustances, file with holes would cause the bcount parameter to the 6644callback function to be incorrect. This bug didn't affect any of 6645e2fsprogs programs, but it was discovered by Paul Mackerras, the 6646author of the PPC boot loader. 6647 6648Removed use of static variables to store the inode cache in libext2fs. 6649This caused problems if more than one filesystem was accessed via 6650libext2fs (static variables in libraries are generally a bad idea). 6651Again, this didn't affect e2fsprogs programs, but it was discovered by 6652Paul Mackerras. 6653 6654Fixed minor bugs and version code drift to assure that e2fsprogs 1.10 6655will compile cleanly with 1.2.13 kernels (even with a.out shared 6656libraries!) 6657 6658Programmer's notes: 6659------------------- 6660 6661Added new functions to duplicate an ext2 filesystem handle, and its 6662associated substructure. New functions: ext2fs_dup_handle(), 6663ext2fs_copy_dblist(), ext2fs_badblocks_copy(), ext2fs_copy_bitmap(). 6664Other structures, such as the io_channel and the inode_cache, now have 6665a ref count so that they only get freed when they are no longer used 6666by any filesystem handle. (These functions were added as part of the 6667development effort for an ext2 resizer). 6668 6669E2fsprogs 1.09 (April 14, 1997) 6670=============================== 6671 6672Fixed bug in mke2fs (really in lib/ext2fs/initialize.c) which was 6673accidentally introduced in the 1.08 release. The overhead calculation 6674was accidentally removed, which caused ext2fs_initialize() to not 6675notice when the filesystem size needed to be adjusted down because 6676there wasn't enough space in the last block group. 6677 6678Fixed bug in version parsing library routine; it was always parsing 6679the library version string, instead of using the passed-in string. 6680 6681Clarified chattr man page. 6682 6683E2fsprogs 1.08 (April 10, 1997) 6684=============================== 6685 6686E2fsck 1.07 was very slow when checking very large filesystems with a 6687lot of files that had hard links (i.e., news spools). This was fixed 6688by seriously revamping the icount abstraction. Added a formal test 6689suite for the icount abstraction. 6690 6691Debugfs now has a "-l" option to the "ls" command, which lists the 6692inode number, permissions, owner, group, size, and name of the files 6693in the directory. 6694 6695Fix a bug in e2fsck where when a directory had its blocks moved to 6696another location during the pass 1b processing, the directory block 6697list wasn't updated, so pass 2 wouldn't check (and correct) the 6698correct directory block. 6699 6700E2fsck will now treat inodes which contain blocks which are claimed by 6701the filesystem metadata by treating them as multiply claimed blocks. 6702This way, the data in those blocks can be copied to a new block during 6703the pass 1b--1d processing. 6704 6705E2fsck will attempt to determine the correct superblock number and 6706display it in the diagnostic and warning messages if possible. 6707 6708Add support for a new (incompatible) feature, "sparse_super". This 6709feature reduces the number of blocks which contain copies of backup 6710superblocks and block group descriptors. (It is only an incompatible 6711feature because of a bug in ext2_free_blocks.) mke2fs and tune2fs now 6712support a new -s option; e2fsck will recognize filesystems built with 6713this feature turned on. 6714 6715E2fsck now checks the library to make sure is the correct version, 6716using new library functions. (This helps to diagnose incorrectly 6717installed e2fsprogs distributions.) 6718 6719Dumpe2fs now prints more information; its now prints the the 6720filesystem revision number, the filesystem sparse_super feature (if 6721present), the block ranges for each block group, and the offset from 6722the beginning of the block group. 6723 6724Mke2fs now distributes the inode and block bitmap blok so that the 6725won't be concentrated in one or two disks in RAID/striping setups. 6726Also, if the user chooses a 2k or 4k block group, mke2fs will try to 6727choose the largest blocks per group that be chosen. (For 2k blocks, 6728you can have up to 16384 blocks/group; for 4k blocks, you can have up 6729to 32768 blocks/group.) Previously mke2fs would not allow 6730specification of more than 8192 blocks per group, even if you were 6731using a 2k or 4k block group. 6732 6733Programmer's notes: 6734------------------- 6735 6736Added a new function ext2fs_create_icount2() which takes a "hint" 6737argument. This hint argument presets the icount array with the list 6738of inodes which actually need to be in the icount array. This really 6739helps to speed up e2fsck. 6740 6741Added a new function ext2fs_icount_validate() which checks the rep 6742invariant for the icount structure. This is used mostly for testing. 6743 6744The error mesasage given when a bad inode number is passed to 6745test_generic_bitmap to reflect EXT2FS_TEST_ERROR (instead of 6746EXT2FS_UNMARK_ERROR). 6747 6748Added a new function ext2fs_set_dir_block which sets the block of a 6749dblist entry, given the directory inode and blockcnt. 6750 6751Added a new function ext2fs_get_library_version() which returns the 6752current library version, and ext2fs_parse_version_string() which 6753returns a version number based on a e2fsprogs version string. 6754 6755The icount functions will return EINVAL if the passed in inode number 6756is out of bounds. 6757 6758E2fsprogs 1.07 (March 9, 1997) 6759============================== 6760 6761E2fsck is now uses much less memory when checking really large 6762filesystems (or rather, filesystems with a large number of inodes). 6763Previously a filesystem with 1 million inodes required 4 megabytes of 6764memory to store inode count statistics; that storage requirement has 6765now been reduced to roughly half a megabyte. 6766 6767E2fsck can now properly deal with bad blocks appearing inside the 6768inode table. Instead of trying to relocate the inode table (which 6769often failed because there wasn't enough space), the inodes in the bad 6770block are marked as in use. 6771 6772E2fsck will automatically try to use the backup superblocks if the 6773primary superblocks have a bad magic number or have missing meta-data 6774blocks (or meta-data blocks which are out of range). 6775 6776E2fsck's pass 3 has been made more efficient; most noticeable on 6777filesystems with a very large number of directories. 6778 6779Completely revamped e2fsck's system of printing problem reports. It 6780is now table driven, to make them more easily customizeable and 6781extendable. Error messages which can be printed out during preen mode 6782are now one line long. 6783 6784Fixed e2fsck's filesystem swapping code so that it won't try to swap 6785fast symbolic links or deleted files. 6786 6787Fixed e2fsck core dumping when fixing a filesystem which has no 6788directories (not even a root directory). 6789 6790Added a check to e2fsck to make sure that the length of every 6791directory entry is a multiple of 4 (since the kernel complains if it 6792isn't). 6793 6794Added a check to e2fsck to make sure that a directory entry isn't a 6795link to the root directory, since that isn't allowed. 6796 6797Added a check to e2fsk to now make sure the '.' and '..' directory 6798entries are null terminated, since the 2.0 kernel requires it. 6799 6800Added check to write_bitmaps() to make sure the superblock doesn't get 6801trashed if the inode or block bitmap is marked as being block zero. 6802 6803Added checking of the new feature set fields in the superblock, to 6804avoid dealing with new filesystem features that this package wasn't 6805set up to handle. 6806 6807Fixed a fencepost error in ext2fs_new_block() which would occasionally 6808try to allocate a block beyond the end of a filesystem. 6809 6810When the UUID library picks a random IEEE 802 address (because it 6811can't find one from a network card), it sets the multicast bit, to 6812avoid conflicting with a legitimate IEEE 802 address. 6813 6814Mke2fs now sets the root directory's owner to be the real uid of the 6815user running mke2fs. If the real uid is non-zero, it also sets 6816the group ownership of the root directory to be the real group-id of 6817the user running mke2fs. 6818 6819Mke2fs now has more intelligent error checking when it is given a 6820non-existent device. 6821 6822When badblocks is given the -vv option, it now updates the block that 6823it is currently testing on every block. 6824 6825Fixed a bug in fsck where it wouldn't modify the PATH envirnoment 6826currently correctly if PATH wasn't already set. 6827 6828Shared libraries now built with dependencies. This allows the shared 6829library files to be used with dlopen(); it also makes the transition 6830to libc 6 easier, since ld.so can tell which libc a particular shared 6831library expects to use. 6832 6833Programmer's notes: 6834------------------- 6835 6836Added new abstraction (defined in dblist.c) for maintaining a list of 6837blocks which belongs to directories. This is used in e2fsck and other 6838programs which need to iterate over all directories. 6839 6840Added new functions which test to see if a contiguous range of blocks 6841(or inodes) are available. (ext2fs_*_bitmap_range). 6842 6843Added new function (ext2_inode_has_valid_blocks) which returns true if 6844an inode has valid blocks. (moved from e2fsck code). 6845 6846Added new function (ext2fs_allocate_tables) which allocates the 6847meta-data blocks as part of initializing a filesystem. (moved from 6848mke2fs code). 6849 6850Added a new I/O manager for testing purposes. It will either allow a 6851program to intercept I/O requests, or print debugging messages to 6852trace the activity of a program using the I/O manager. 6853 6854The badblocks_list functions now store the bad blocks in a sorted 6855order, and use a binary search to speed up badblocks_list_test. 6856 6857The inode scan function ext2fs_get_next_inode() may now return a soft 6858error returns: MISSING_INODE_TABLE and BAD_BLOCK_IN_INODE_TABLE in 6859those cases where part of an inode table is missing or there is a bad 6860block in the inode table. 6861 6862Added a new function (ext2fs_block_iterate2) which adds new arguments to 6863the callback function to return a pointer (block and offset) to the 6864reference of the block. 6865 6866Added new function (ext2fs_inode_scan_goto_blockgroup) which allows an 6867application to jump to a particular block group while doing an inode 6868scan. 6869 6870The badblocks list functions were renamed from badblocks_* to 6871ext2fs_badblocks_*. Backwards compatibility functions are available 6872for now, but programs should be modified to use the new interface. 6873 6874Some of the library functions were reorganized into separate files to 6875reduce the size of some programs which statically link against the 6876ext2 library. 6877 6878Put in some miscellaneous fixes for the Alpha platform. 6879 6880 6881E2fsprogs 1.06 (October 7, 1996) 6882================================ 6883 6884Fixed serious bug in e2fsck: if the block descriptors are bad, don't 6885smash the backup copies in ext2fs_close(). (The problem was that when 6886e2fsck -p discovered the problem, while it was closing the filesystem 6887and exiting, it was also blowing away the backup superblocks on the 6888disk, which was less than friendly.) We now make it the case that we 6889only write out the backup superblock and the back block descriptors if 6890the filesystem is completely free from problems. 6891 6892Fixed a bug in block_interate in the lib/ext2fs library which caused 6893e2fsck to fail on GNU Hurd-created filesystems. 6894 6895Add support for Linux/FT's bootloader, which actually uses 6896EXT2_BOOT_LOADER, and sets its mode bits which caused e2fsck to want 6897to clear the inode. 6898 6899Add support for the "A" (no atime update) attribute. (Note: this 6900attribute is not yet in production kernels.) 6901 6902The test suite is not automatically run when doing a "make all" from 6903the top level directory. Users should manually run "make check" if 6904they wish to run the test suite. 6905 6906Upon a preenhalt(), make the printed message more explicit that 6907running e2fsck "MANAULLY" means without the -p or -a options. 6908 6909In e2fsck, if a disconnected inode is zero-length, offer to clear it 6910instead of offering to connect it to lost+found. 6911 6912In e2fsck, if a filesystem was just unmounted uncleanly, and needs 6913e2fsck to be run over it, change e2fsck to explicitly display this 6914fact. 6915 6916For dumpe2fs and e2fsck, cause the -V option to print out which 6917version of the ext2fs library is actually getting used. (This will 6918help detect mismatches of using a 1.06 utility with a 1.05 library, 6919etc.) 6920 6921Programmers' notes: 6922------------------- 6923 6924EXT2_SWAP_BYTES was changed to EXT2_FLAG_SWAP_BYTES, which better fits 6925the naming convention. 6926 6927In ext2fs_initialize(), make sure the description for the inode bitmap 6928is correctly initialize. 6929 6930Fixed minor type typo in ext2fs_allocate_generic_bitmap(); 6931 6932E2fsprogs 1.05 (September 7, 1996) 6933================================== 6934 6935Add support for new fields in the ext2 superblock --- volume name, 6936volume UUID, and last mounted field. Dumpe2fs displays these fields, 6937tune2fs and mke2fs allows you to set them. E2fsck will automatically 6938generate a UUID for those volumes that don't have them. 6939 6940Put in support for e2fsck to recognize HURD specific ext2 features --- 6941most notably, the translator block. The e2fsprogs tools will now use 6942the creator_os field in the superblock to correctly handle different 6943OS-specific variants of the ext2 filesystem. 6944 6945E2fsck now fixes inodes which have a the deletion time set, but which 6946have a non-zero i_link_count field by offering to clear the deletion 6947time. Previously e2fsck assumed that the inode was deleted (per 0.3c 6948ext2 kernel behavior) and offered to unlink the file. 6949 6950If e2fsck sets the clean bit, but nothing else, set the exit code 6951FSCK_NONDESTRUCT. After all, e2fsck did fix a filesystem error --- it 6952set the filesystem valid bit when it was previously cleared. :-) This 6953was needed to make the HURD fsck driver happy. 6954 6955If the user refuses to attach an unattached inode, e2fsck will no 6956longer set the inode's link count. Otherwise, the inode would end up 6957getting marked as unused, which might cause loss of data later. 6958 6959Make the message issued by e2fsck when the superblock is corrupt less 6960confusing for users. It now mentions that another reason for the 6961"corrupt superblock" message might be that the partition might not be 6962an ext2 filesystem at all (it might swap, msdos filesystem, ufs, etc.) 6963 6964Make the libext2 library more robuest so that e2fsck won't coredump on 6965an illegal superblock where the blocksize is zero. (f_crashdisk is 6966the test case). 6967 6968By default, create filesystems where the default checkinterval is 6 6969months (180 days). Linux servers can be robust enough that 20 reboots 6970can be a long, long time. 6971 6972Added configure flag --enable-old-bitops, which forces the bitops to 6973use the old (native) bitmask operations. By default on the sparc 6974platform, the standard ext2 bit ordering is now used. 6975 6976Added a new feature to e2fsck to byte-swap filesystems; this can be 6977used to convert old m68k filesystems to use the standard byte-order 6978storage for the superblock, inodes, and directory blocks. This 6979function is invoked by using the '-s' option to e2fsck. 6980 6981Debugfs's "dump" command has been enhanced so that it writes out the 6982exact size of the file so that the nulls at the end of the file are 6983eliminated. The command also accept a new "-p" option which will 6984attempt preserve to preserve the ownernship, permissions, and 6985file modification/access times. 6986 6987Debugfs has two new options, -f and -R. The -R option allows the user 6988to execute a single debugfs command from the command line. The -f 6989option allows the user to specify a "command file" containing debugfs 6990commands which will get executed. 6991 6992Dumpe2fs now pretty prints the check interval, instead of just 6993printing the check interval as a number of seconds. 6994 6995Fix bugs in debugfs: the params command when no filesystem is opened 6996no longer causes a core dump. It is now possible to unlink a file 6997when a pathame containing a '/' is specified. 6998 6999Tune2fs has a new -C option which sets the number of times the 7000filesystem has been mounted. 7001 7002Fix the chattr '-v' option so that it actually works. Chattr was 7003being buggy about the -v option parsing. 7004 7005Programmers' notes: 7006------------------- 7007 7008The directory lib/uuid contains a set of library routines to generate 7009DCE compatible UUIDs. 7010 7011Extended ext2fs_namei() to handle symbolic links. Added new function 7012ext2fs_nami_follow() which will follow last symbolic link in the case 7013where the pathname points to a sym link. 7014 7015The ext2fs_block_iterate function will now return the HURD translator 7016block, if present. The new flag BLOCK_FLAG_DATA_ONLY will cause the 7017iterator to return data blocks only. The ext2fs.h file now defines 7018constants BLOCK_COUNT_IND, BLOCK_COUNT_DIND, BLOCK_COUNT_TIND, and 7019BLOCK_COUNT_TRANSLATOR, which are the magic values passed in the block 7020count field of the iterator callback function. 7021 7022The test script driver now takes an optional second argument, which is 7023the test case to be run. This allows you to run a test case without 7024needing to run the entire test suite. 7025 7026On Linux ELF systems, install the .so files in the correct places 7027(/usr/lib). The .so files must be stored in the same directory as the 7028.a files. 7029 7030Fixed miscellaneous HURD compilation issues with header file being 7031included in the right order. 7032 7033Fixed debugfs so that it resets optind to zero, not one, since setting 7034optind to zero is more correct. 7035 7036 7037E2fsprogs 1.04 (May 16, 1996) 7038============================= 7039 7040First "official" (1.03 was a limited release only) to support building 7041e2fsprogs under Linux 2.0 kernels (as well as late model 1.3 and 1.99 7042kernels). 7043 7044This package includes a RPM specs file, that it can be built using the 7045RedHat Package Manager. 7046 7047E2fsck now prints a hint that if there are lots of incorrectly located 7048inode bitmaps, block bitmaps, and inode table blocks, the user might 7049want to try using e2fsck -b 8193 first, to see if that fares any 7050better. 7051 7052For ext2 filesystem written with the hurd, debugfs will now print out 7053the translator field when printing an inode structure. 7054 7055Lots of miscellaneous linking/installation cleanups: 7056 7057 Libraries are now linked using a relative pathname, instead of 7058 relying on -L working correct. It doesn't, in many cases, including 7059 current versions of GNU ld. This guarantees that the build tree is 7060 linking with the right libraries, instead of the ones installed in 7061 /usr/lib. 7062 7063 Header files, man pages, and the et/ss shell scripts are now 7064 generated using a custom substitution script, instead of relying on 7065 the configure script. This prevents needless recompilation of 7066 files; in addition, the custom substitution script is much faster. 7067 7068 e2fsck may now be linked dynamically, by using the 7069 --enable-dynamic-e2fsck flag to configure. This is not recommended, 7070 since it increases e2fsck's dependence on other files, but some 7071 people need to save disk space, and other critical programs on their 7072 systems were being linked dynamically anyway. 7073 7074 Programs such as fsck which didn't need to be linked against 7075 libext2fs (or mke2fs which didn't need to be linked against libe2p) 7076 only link against libraries they actually need. Otherwise, those 7077 programs would require the presense of libraries that otherwise 7078 could be removed from a rescuse diskette. 7079 7080 The ss include files are now installed correctly so they can 7081 actually be used by another package. 7082 7083 If the profiling libraries are built, they are now installed on a 7084 "make install-libs". 7085 7086 7087E2fsprogs 1.03 (March 27, 1996) 7088=============================== 7089 7090Change the m68k bit numbering for bitmasks to match the bit numbering 7091used by all other ext2 implementations. (This change was requested by 7092the m68k kernel development team.) 7093 7094Support (in-development) filesystem format revision which supports 7095(among other things) dynamically sized inodes. 7096 7097Fixed a bug in the ext2 library so that an intelligent error is 7098returned if mke2fs is run with a ridiculously small number of blocks 7099for a partition. 7100 7101Fixed a bug in the ext2 library which required that the device be 7102openable in read/write mode in order to determine its size. This 7103caused e2fsck -n to require read/write access when it was not 7104previously necessary. 7105 7106Fixed a bug in e2fsck which casued it to occasionally fail the test 7107suite depending on which version of the floating point library it was 7108using. 7109 7110Fixed a bug in e2fsck so that it now halts with a fatal error when 7111certain superblock consistency checks fail. Previously it continued 7112running e2fsck, with some potential confusing/damaging consequences. 7113 7114Added new flag to fsck which allows the root to be checked in parallel 7115with other filesytems. This is not the safest thing in the world to 7116do, but some system administrators really wanted it. 7117 7118Fixed -Wall flames in lib/ss. 7119 7120 7121E2fsprogs 1.02 (January 16, 1996) 7122================================= 7123 7124Fix to allow e2fsprogs to be compiled on recent 1.3 (pl45+) kernels. 7125 7126Change e2fsck to print statistics of how many non-contiguous files are 7127on the system. Note that a file which is larger than 8k blocks, it is 7128guaranteed to be non-contiguous. 7129 7130In mke2fs, print a warning message if a user tries to format a whole 7131disk (/dev/hda versus /dev/hda1). If a user really wants to format a 7132whole disk, the -F (force) option forces mke2fs to format a whole disk 7133as a filesytem. 7134 7135Fix a bug in fsck where in some cases it might start checking 7136partitions in the next pass before it finishes checking partitions in 7137the current pass. This still won't cause two partitions on the same 7138disk will be checked, so it's rarely a problem in real life. 7139 7140Patch lsattr so that it won't hang when checking a named pipe. 7141 7142Minor compilation fixes: 7143 * Fix the order of libraries that were linked in debugfs. 7144 * Allow the sources to be compiled with -ansi turned on. 7145