1E2fsprogs 1.30 (October 31, 2002) 2================================= 3 4When resizing a filesystem, and inodes need to moved, resize2fs will 5update the inode's ctime field, and the mtime field of the containing 6directories, so that incremental backups using dump/restore will work 7correctly. 8 9In order to avoid spurious errors, e2fsck will check the last 10superblock write time to determine whether or not it can safely use 11the LOW_DTIME checks to determine if there are inodes on the orphan 12inode linked list that had somehow gotten disconnected. (Addresses 13Sourceforge bug #620980) 14 15Partition in /proc/partitions that being with the string "lvm" are 16considered real partitions even if they do not end with a number. 17 18Fixed a bug in the the uuid generation function, where if /dev/urandom 19is not present, but /dev/random is, and there isn't sufficient 20entropy, the get_random_byte function could spin a loop forever. 21 22E2fsck, mke2fs, etc. will now reliably notice when image files are 23mounted using the loopback interface. (Addresses Sourceforge bug 24#619119) 25 26When flushing buffers (as is done by badblocks, for example) check to 27see if the BLKFLSBUF ioctl succeeds, and if so, avoid using the 28FDFLUSH ioctl which causes the MD device driver which causes confusing 29syslog messages about obsolete ioctl messages. (Addresses Sourceforge 30bug #545832). 31 32Debugfs's write command now checks to see if the destination filename 33exists before creating it. (Addresses Sourceforge bug #478195.) 34 35When installing man pages, search for compressed man pages using all 36commonly used compression extensions (i.e., .Z, .gz, .bz2, etc.) 37 38Fixed a bug in fsck where multiple filesystems specified on the 39command were not being checked in parallel due to a logic bug 40introduced to support the FSCK_MAX_INST environment variable. 41 42We have added a new superblock field, s_mkfs_time, which indicates 43when a filesystem was created. It is set via mke2fs, and printed out 44by dumpe2fs, but is not actually touched by the kernel. 45 46Dumpe2fs has been made more robust by not aborting if there is an 47error reading the block/inode bitmaps; instead it will still print out 48the location of the block/inode bitmaps and inode table. 49 50Add support for the an alternative block group descriptor layout which 51allows for on-line resizing without needing to prepare the filesystem 52in advance. (This is the incompat feature flag meta_bg.) 53 54Add support for storing default mount options in the superblock, so 55that the filesystem can be mounted with specific mount options without 56needing to specify them on the mount command line or in the /etc/fstab 57file. 58 59Add support for a new inode flag, which is to be used for indicating 60the top of directory hierarchies for the Orlov block allocator. 61 62Fix e2fsck so that if it creates the lost+found directory, it does so 63with the more appropriate permissions of 0700. Also change 64mklost+found so that it also creates the directory 0700. 65 66Fixed format bug in e2fsck if NLS is in use. 67 68Add a German translation for e2fsprogs's NLS support. 69 70Fixed e2fsck so that it more handles BAD_BLOCK_IN_INODE_TABLE even at 71the beginning of the inode table. This won't matter much, since if 72there is a bad block at the beginning of the inode table, the root 73directory won't be available. But at least e2fsck won't crash in this 74case. 75 76Fixed endian problems in the htree code for e2fsck and debugfs. 77 78When byte-swapping a filesystem on a PPC architecture, byte-swap the 79bitmaps since the historical big-endian ext2 variant had byte-swapped 80bitmaps, and the ext2fs library assumes this. This fixes the 81regression test suite on PPC machines. 82 83Fix e2image so that it handles a bad block in the inode table 84appropriately. 85 86E2fsck now uses a more sophisticated algorithm to salvage corrupted 87directories that recovers more information in the corrupted directory 88block. 89 90E2fsck now performs additional consistency checks on indexed (HTREE) 91directories. 92 93Fix a bug where e2fsck might get confused about whether a completely 94empty directory block is an empty leaf block or an interior htree 95node. This could cause e2fsck to get confused into think that a valid 96indexed directory was corrupted. 97 98E2fsck no longer creates an empty directory entry at the end of a 99directory block when optimizing a directory. This would cause some 100earlier versions of the dxdir kernel code to corrupt the directory 101when attempting to split a node containing an empty directory entry. 102 103E2fsck could sometimes in rare circumstances leave the indexed flag 104set after a small directory was optimized by compressing it instead of 105indexing it. (This should never happen in real life, since 106directories that small wouldn't have been indexed, but better safe 107than sorry.) 108 109E2fsck now only updates the master superblock in all cases. This 110slightly shortens its run time. 111 112Ext2ed can deal with directory entries of length 0; previously it 113would get stuck in an infinite loop. 114 115Fsck now has support for reiserfs volumes when parsing LABEL= and UUID= 116specifiers. (Sourceforge patch #613447) 117 118Badblocks will now work correctly on read-only devices such as 119CD-ROM's. (Sourceforge patch #600451) 120 121Updated and clarified man pages. (Addresses Debian bug #167108) 122 123