[ANNOUNCEMENT] Journaled File System (JFS) release 1.0.6

[Steve Best]

Release 1.0.6 of JFS was made available today.

Drop 44 on September 28, 2001 (jfs-2.4-1.0.6-patch.tar.gz
and jfsutils-1.0.6.tar.gz) includes fixes to the file system
and utilities.

Function and Fixes in release 1.0.6

- If fsck rebuilds root directory, reset di_next_index to 2
- fsck needs to process, validate inodes with size=0
- Create jfs_incore.h which merges linux/jfs_fs.h, linux/jfs_fs_i.h
  , and jfs_fs_sb.h (Thanks Christoph Hellwig)
- Create a configuration option to handle JFS_DEBUG define (Thanks
  Christoph Hellwig)
- Fixed a few cases where positive error codes were returned to
  the VFS (Thanks Christoph Hellwig).
- Replace jfs_dir_read by generic_read_dir (Thanks Christoph Hellwig).
- jfs_fsync_inode is only called by jfs_fsync_file, merge the two
  and rename to jfs_fsync (Thanks Christoph Hellwig).
- Add a bunch of missing externs (Thanks Christoph Hellwig).
- jfs_rwlock_lock is unused, nuke it (Thanks Christoph Hellwig).
- Always use atomic set/test_bit operations to protect
  jfs_ip->cflag (Thanks Christoph Hellwig)
- Combine jfs_ip->flag with jfs_ip->cflag (Thanks Christoph Hellwig).
- Fixed minor format errors reported by fsck
- cflags should be long so bitops always works correctly (Thanks
  Christoph Hellwig)
- Use GFP_NOFS for runtime memory allocations (Thanks Christoph Hellwig)
- Support  VM changes in 2.4.10 of the kernel
- Remove ifdefs supporting older 2.4 kernels. JFS now requires
  at least 2.4.3 or 2.4.2-ac2
- Simplify and remove one use of IWRITE_TRYLOCK
- jfs_truncate was not passing tid to xtTruncate
- removed obsolete extent_page workaround
- correct recovery from failed diAlloc call (disk full)
- In write_metapage, don't call commit_write if prepare_write failed

For more details about JFS, please see the README or changelog.jfs.

Steve
JFS for Linux http://oss.software.ibm.com/jfs