Hi all, ext3-0.0.5c is now up at: ftp.uk.linux.org:/pub/linux/sct/fs/jfs/ and ftp.*.kernel.org:/pub/linux/kernel/people/sct/ext3/ The most important part of this release is the e2fsprogs: e2fsck now supports the journal changes for metadata-only journaling. Ted, I've changed around the use of jfs* include files in e2fsprogs quite heavily here. In each build directory --- the lib/ext2fs dir where mkjournal.c resides, and e2fsck/ itself --- there is a "jfs_user.h" which needs to include the necessary header files. e2fsprogs/include/linux/jfs.h is now a verbatim copy of the kernel's own jfs.h, and likewise recovery.c and revoke.c are shared. mke2fs in this release can only create V2 superblock journals, so filesystems created by the mke2fs will not work on ext3-0.0.3b kernels. You can still use the old technique of creating the journal via kernel mount options if you want to create an older format filesystem, but I want to deprecate the V1 superblock as quickly as possible. e2fsck will understand both V1 and V2 format superblocks, though. Ted, we should sync up on e2fsprogs soon if we can --- do you want to do an official WIP of this stuff shortly?>From the readme:Changes in this release ----------------------- 0.0.5c: Fix a major kernel recovery bug which could prevent recovery if you crashed too soon after first mouting a filesystem. Fixed up all of the unnecessary recovery and journal warnings at boot time. Merged in Andreas's flags-handling cleanups. Made the V2 data mode default to ordered, as documented (5b actually defaulted to writeback). Add e2fsprogs RPMs and SRPMS, including e2fsck support for recovering the V2 superblocks. Cheers Stephen
> > Hi all, > > ext3-0.0.5c is now up at: > > ftp.uk.linux.org:/pub/linux/sct/fs/jfs/ > and ftp.*.kernel.org:/pub/linux/kernel/people/sct/ext3/ >And what about rpm's on redhat? Regards, Henri J. Schlereth -- "Never play chicken with someone who has nothing to lose." --The Cynic's Book of Wisdom
Stephen,
I noticed a couple small bugs in the COMPAT fixes that are in 0.0.5c that
were fixed after I posted my patches. One is only a warning message that
is printed, but since it mentions e2fsck, people may worry about it. The
other is also a very rare case - if you are creating a > 2GB file for the
first time on a rev 0 filesystem.
As I write my network is down since last night, so I don't know when this
will get out. It may have been fixed by you already.
I'm just in the process of seeing what else is different between 2.2.18-ext2
and 0.0.5c ext3 to sync them up again.
Cheers, Andreas
==========================================================================---
linux/fs/ext3/super.c.orig Tue Dec 12 15:52:23 2000
+++ linux/fs/ext3/super.c Tue Dec 12 16:00:21 2000
@@ -682,7 +682,7 @@
MOD_DEC_USE_COUNT;
return NULL;
}
- if (le32_to_cpu(es->s_rev_level) > EXT3_GOOD_OLD_REV &&
+ if (le32_to_cpu(es->s_rev_level) == EXT3_GOOD_OLD_REV &&
(EXT3_HAS_COMPAT_FEATURE(sb, ~0U) ||
EXT3_HAS_RO_COMPAT_FEATURE(sb, ~0U) ||
EXT3_HAS_INCOMPAT_FEATURE(sb, ~0U)))
--- linux/fs/ext3/file.c.orig Tue Dec 12 15:52:23 2000
+++ linux/fs/ext3/file.c Tue Dec 12 16:00:21 2000
@@ -270,6 +270,7 @@
/* If this is the first large file created, add a flag
to the superblock */
journal_get_write_access(handle, sb->u.ext3_sb.s_sbh);
+ ext3_update_fs_rev(sb);
EXT3_SET_RO_COMPAT_FEATURE(sb,
EXT3_FEATURE_RO_COMPAT_LARGE_FILE);
journal_dirty_metadata(handle, sb->u.ext3_sb.s_sbh);
--
Andreas Dilger \ "If a man ate a pound of pasta and a pound of antipasto,
\ would they cancel out, leaving him still hungry?"
http://www-mddsp.enel.ucalgary.ca/People/adilger/ -- Dogbert