1 tbd Sun Microsystems, Inc.
6 Description: OSS crashes frequently, e2fsck does not fix
7 Details : Add an extra check to ldiskfs extents code for the condition,
8 eh_entries = 0 & eh_depth != 0
12 Description: migrate ldiskfs - ldiskfs2 (extents on directories)
13 Details : disable preallocation for non-regular files.
16 Frequency : blocks per group < blocksize*8 and uninit_groups is enabled
18 Description: ldiskfs error: XXX blocks in bitmap, YYY in gd
19 Details : If blocks per group is less than blocksize*8, set rest of the
23 Frequency : only for filesystems larger than 8TB
25 Description: ldiskfs BUG ldiskfs_mb_use_best_found()
26 Details : The ldiskfs mballoc3 code was using a __u16 to store the group
27 number, but with 8TB+ filesystems there are more than 65536
28 groups, causing an oops.
30 Severity : enhancement
32 Description: Add a FIEMAP(FIle Extent MAP) ioctl for ldiskfs
33 Details : FIEMAP ioctl will allow an application to efficiently fetch the
34 extent information of a file. It can be used to map logical blocks
35 in a file to physical blocks in the block device.
37 -------------------------------------------------------------------------------
39 04-26-2008 Sun Microsystems, Inc.
44 Description: Kernel BUG at ... ldiskfs/ldiskfs/mballoc.c:3853
45 Details : Removing the underlying device (e.g. via a SCSI rescan) can cause a
46 kernel BUG in ldiskfs_mb_discard_inode_preallocations().
50 Description: deadlock in ext3_expand_extra_isize
51 Details : Inode expansion during ext3_mark_inode_dirty may lead to deadlock.
52 This is solved by making sure that ext3_expand_extra_isize isn't
53 called from ext3_xattr_set_handle.
57 Description: soft lockups on 1.6.2 MDS (is_subdir)
58 Details : don't add dentries with ".." to dcache and ignore such dentries in
62 Frequency : very rare, if additional xattrs are used on kernels >= 2.6.12
64 Description: files may lose file attributes in some cases
65 Details : If there are multiple extended attributes stored on the inode,
66 in particular ACLs, SELinux, or user attributes (if user_xattr
67 is specified for the client mount options) then there is a risk
68 of attribute loss. If an additional attribute is be stored
69 initially in the inode and then increase in size enough to be
70 moved to the external attribute block (e.g. ACL growing in size)
71 for the attribute to be lost.
75 Description: inode version not being initialized on new inodes
76 Details : The inode i_version field was not being initialized on disk.
77 This field is currently unused but will be needed for VBR.
79 --------------------------------------------------------------------------------
81 2008-01-11 Sun Microsystems, Inc.
86 Description: Add support for vanilla-2.6.22 kernel.
88 --------------------------------------------------------------------------------
90 2007-12-07 Cluster File Systems, Inc. <info@clusterfs.com>
95 Description: statfs speedup patches.
96 Details : The patches improve statfs performance for very large
100 Frequency : if the uninit_groups feature is enabled on ldiskfs
102 Description: e2fsck reports "invalid unused inodes count"
103 Details : If a new ldiskfs filesystem is created with the "uninit_groups"
104 feature and only a single inode is created in a group then the
105 "bg_unused_inodes" count is incorrectly updated. Creating a
106 second inode in that group would update it correctly.
108 --------------------------------------------------------------------------------
110 2007-09-27 Cluster File Systems, Inc. <info@clusterfs.com>
113 Severity : enhancement
115 Description: Add a FIEMAP(FIle Extent MAP) ioctl for ldiskfs
116 Details : FIEMAP ioctl will allow an application to efficiently fetch the
117 extent information of a file. It can be used to map logical blocks
118 in a file to physical blocks in the block device.
120 --------------------------------------------------------------------------------
122 2007-08-29 Cluster File Systems, Inc. <info@clusterfs.com>
128 Description: Multiple mount protection(MMP) support.
129 Details : This feature will protect the filesystem from being mounted more
130 than once simultaneously. It will also protect changes by e2fsprogs
131 to the filesystem if it is mounted. This assumes high importance
132 in a shared storage environment where multiple mounts can severely
133 corrupt the filesystem.
135 --------------------------------------------------------------------------------
137 2007-08-10 Cluster File Systems, Inc. <info@clusterfs.com>
139 * Initial release of ldiskfs as a separate package.
144 Description: Updated patchess for new RHEL4 kernel
145 Details : Add patch ext3-unlink-race.patch
146 Updated series file ldiskfs-2.6-rhel4.series