1 2010-04-30 Oracle, Inc.
6 Description: Reduce group prealloc size and skip groups with little free space.
10 Description: Fix issue with proc_remove.
12 -------------------------------------------------------------------------------
14 2010-04-30 Oracle, Inc.
18 -------------------------------------------------------------------------------
20 2010-01-29 Sun Microsystems, Inc.
24 -------------------------------------------------------------------------------
26 2009-11-05 Sun Microsystems, Inc.
30 Frequency : in recovery
32 Description: ext3-corrupted-orphans-2.6.patch added.
33 Details : Backport of ext3 patch to handle corrupted orphans.
35 -------------------------------------------------------------------------------
37 2009-07-31 Sun Microsystems, Inc.
41 Frequency : in recovery
43 Description: don't mix llog inodes with normal.
44 Details : allocate inodes for log in last inode group
46 -------------------------------------------------------------------------------
48 2009-04-20 Sun Microsystems, Inc.
53 Description: minor fixes and cleanups
54 Details : use EXT_UNSET_BLOCK to avoid confusion with EXT_MAX_BLOCK.
55 Initialize 'ix' variable in extents patch to stop compiler warning.
59 Description: update FIEMAP ioctl to match upstream kernel version
60 Details : the FIEMAP block-mapping ioctl had a prototype version in
61 ldiskfs 3.0.7 but this release updates it to match the
62 interface in the upstream kernel, with a new ioctl number.
65 Frequency : only if MMP is active and detects filesystem is in use
67 Description: if MMP startup fails, an oops is triggered
68 Details : if ldiskfs mounting doesn't succeed the error handling doesn't
69 clean up the MMP data correctly, causing an oops.
71 -------------------------------------------------------------------------------
73 2009-04-06 Sun Microsystems, Inc.
79 Description: fix racy locking of mballoc block bitmaps causing BUG
80 Details : The locking of the mballoc buddy bitmap and the in-memory
81 block bitmap was using two different spin locks in some
82 cases. This made it possible to incorrectly access the
83 mballoc bitmap while another process was modifying it,
84 causing a sanity assertion to fail. While no on-disk corruption
85 was reported, there was some risk of this happening.
87 -------------------------------------------------------------------------------
89 2009-02-07 Sun Microsystems, Inc.
92 Severity : enhancement
94 Description: Get RAID stripe size from superblock
95 Details : RAID striping parameters are now saved in the superblock itself,
96 so we should use these parameters instead of having to specify
97 a mount option each time.
100 Frequency : only if server is running on unsupported big-endian machine
102 Description: Disable big-endian ldiskfs server support.
103 Details : The ldiskfs code is not tested on big-endian machines, and
104 there are known compatibility problems in the extents code
105 when running for most of the kernels. Print an error message
106 and refuse to mount, in case anyone tests this. For existing
107 filesystems that might have been created in this way it is
108 possible to mount with the "bigendian_extents" option to
112 Frequency : only with software RAID-5
114 Description: MMP block reads may fail to detect another user of the filesystem
115 Details : with software RAID it is possible that the RAID driver will
116 cache the MMP block and not refetch it from disk. Force the
117 read to invalidate the RAID page cache and go directly to disk.
119 -------------------------------------------------------------------------------
121 2008-08-31 Sun Microsystems, Inc.
124 Severity : enhancement
126 Description: Interoperability at server side (Disk interoperability)
127 Details : Exported some ldiskfs functions which are required for iop
131 Description: OSS crashes frequently, e2fsck does not fix
132 Details : Add an extra check to ldiskfs extents code for the condition,
133 eh_entries = 0 & eh_depth != 0
137 Description: migrate ldiskfs - ldiskfs2 (extents on directories)
138 Details : disable preallocation for non-regular files.
141 Frequency : blocks per group < blocksize*8 and uninit_groups is enabled
143 Description: ldiskfs error: XXX blocks in bitmap, YYY in gd
144 Details : If blocks per group is less than blocksize*8, set rest of the
148 Frequency : only for filesystems larger than 8TB
150 Description: ldiskfs BUG ldiskfs_mb_use_best_found()
151 Details : The ldiskfs mballoc3 code was using a __u16 to store the group
152 number, but with 8TB+ filesystems there are more than 65536
153 groups, causing an oops.
155 Severity : enhancement
157 Description: Add a FIEMAP(FIle Extent MAP) ioctl for ldiskfs
158 Details : FIEMAP ioctl will allow an application to efficiently fetch the
159 extent information of a file. It can be used to map logical blocks
160 in a file to physical blocks in the block device.
164 Description: mkfs.lustre: Unable to mount /dev/cciss/c0d1:Cannot allocate memory
165 Details : correctly handle device paths using a subdirectory in /dev when
166 creating the per-device procfs directory under /proc/fs/ldiskfs.
168 -------------------------------------------------------------------------------
170 2008-04-26 Sun Microsystems, Inc.
175 Description: Kernel BUG at ... ldiskfs/ldiskfs/mballoc.c:3853
176 Details : Removing the underlying device (e.g. via a SCSI rescan) can cause a
177 kernel BUG in ldiskfs_mb_discard_inode_preallocations().
181 Description: deadlock in ext3_expand_extra_isize
182 Details : Inode expansion during ext3_mark_inode_dirty may lead to deadlock.
183 This is solved by making sure that ext3_expand_extra_isize isn't
184 called from ext3_xattr_set_handle.
188 Description: soft lockups on 1.6.2 MDS (is_subdir)
189 Details : don't add dentries with ".." to dcache and ignore such dentries in
193 Frequency : very rare, if additional xattrs are used on kernels >= 2.6.12
195 Description: files may lose file attributes in some cases
196 Details : If there are multiple extended attributes stored on the inode,
197 in particular ACLs, SELinux, or user attributes (if user_xattr
198 is specified for the client mount options) then there is a risk
199 of attribute loss. If an additional attribute is be stored
200 initially in the inode and then increase in size enough to be
201 moved to the external attribute block (e.g. ACL growing in size)
202 for the attribute to be lost.
206 Description: inode version not being initialized on new inodes
207 Details : The inode i_version field was not being initialized on disk.
208 This field is currently unused but will be needed for VBR.
210 --------------------------------------------------------------------------------
212 2008-01-11 Sun Microsystems, Inc.
217 Description: Add support for vanilla-2.6.22 kernel.
219 --------------------------------------------------------------------------------
221 2007-12-07 Cluster File Systems, Inc. <info@clusterfs.com>
226 Description: statfs speedup patches.
227 Details : The patches improve statfs performance for very large
231 Frequency : if the uninit_groups feature is enabled on ldiskfs
233 Description: e2fsck reports "invalid unused inodes count"
234 Details : If a new ldiskfs filesystem is created with the "uninit_groups"
235 feature and only a single inode is created in a group then the
236 "bg_unused_inodes" count is incorrectly updated. Creating a
237 second inode in that group would update it correctly.
239 --------------------------------------------------------------------------------
241 2007-09-27 Cluster File Systems, Inc. <info@clusterfs.com>
244 Severity : enhancement
246 Description: Add a FIEMAP(FIle Extent MAP) ioctl for ldiskfs
247 Details : FIEMAP ioctl will allow an application to efficiently fetch the
248 extent information of a file. It can be used to map logical blocks
249 in a file to physical blocks in the block device.
251 --------------------------------------------------------------------------------
253 2007-08-29 Cluster File Systems, Inc. <info@clusterfs.com>
259 Description: Multiple mount protection(MMP) support.
260 Details : This feature will protect the filesystem from being mounted more
261 than once simultaneously. It will also protect changes by e2fsprogs
262 to the filesystem if it is mounted. This assumes high importance
263 in a shared storage environment where multiple mounts can severely
264 corrupt the filesystem.
266 --------------------------------------------------------------------------------
268 2007-08-10 Cluster File Systems, Inc. <info@clusterfs.com>
270 * Initial release of ldiskfs as a separate package.
275 Description: Updated patchess for new RHEL4 kernel
276 Details : Add patch ext3-unlink-race.patch
277 Updated series file ldiskfs-2.6-rhel4.series