1 Version 0.92 (2002-03-29)
3 * Minor cleanup. Ran Lindent on the sources.
5 Version 0.92 (2002-03-27)
7 * Fixed module makefile problem. It was not compiling all the correct
9 * Removed duplicated function definition
10 * Fixed potential null pointer dereference when reporting an error
12 Version 0.91 (2002-03-26)
14 * Oy! Fixed stupid bug that would cause an unresolved symbol error.
15 Thanks to Laszlo Boszormenyi for pointing this out to me.
17 Version 0.9 (2002-03-14)
19 * Added Sergey S. Kostyliov's patch to eliminate memcpy() overhead
20 from b+tree operations. Changes the befs_read_datastream() interface.
22 * Segregated the functions that interface directly with the linux vfs
23 interface into their own file called linuxvfs.c. [WD]
25 Version 0.64 (2002-02-07)
27 * Did the string comparison really right this time (btree.c) [WD]
29 * Fixed up some places where I assumed that a long int could hold
30 a pointer value. (btree.c) [WD]
32 * Andrew Farnham <andrewfarnham@uq.net.au> pointed out that the module
33 wouldn't work on older (<2.4.10) kernels due to an unresolved symbol.
34 This is bad, since 2.4.9 is still the current RedHat kernel. I added
35 a workaround for this problem (compatibility.h) [WD]
37 * Sergey S. Kostyliov made befs_find_key() use a binary search to find
38 keys within btree nodes, rather than the linear search we were using
39 before. (btree.c) [Sergey S. Kostyliov <rathamahata@php4.ru>]
41 * Made a debian package of the source for use with kernel-package. [WD]
44 Version 0.63 (2002-01-31)
46 * Fixed bug in befs_find_brun_indirect() that would result in the wrong
47 block being read. It was introduced when adding byteswapping in
48 0.61. (datastream.c) [WD]
50 * Fixed a longstanding bug in befs_find_key() that would result in it
51 finding the first key that is a substring of the string it is searching
52 for. For example, this would cause files in the same directory with
53 names like file1 and file2 to mysteriously be duplicates of each other
54 (because they have the same inode number). Many thanks to Pavel Roskin
55 for reporting this serious bug!!!
58 * Added support for long symlinks, after Axel Dorfler explained up how
59 they work. I had forgotten all about them. (inode.c, symlink.c) [WD]
61 * Documentation improvements in source. [WD]
63 * Makefile fix for independent module when CONFIG_MODVERSION is set in
64 kernel config [Pavel Roskin <proski@gnu.org>]
66 * Compile warning fix for namei.c. [Sergey S. Kostyliov <rathamahata@php4.ru>]
71 * Fixed makefile for module install [WD]
74 Version 0.61 (2002-01-20)
76 * Made functions in endian.h to do the correct byteswapping, no matter
79 * Abbandoned silly checks for a NULL superblock pointer in debug.c. [WD]
81 * Misc code cleanups. Also cleanup of this changelog file. [WD]
83 * Added byteswapping to all metadata reads from disk.
84 Uses the functions from endian.h [WD]
86 * Remove the typedef of struct super_block to vfs_sb, as it offended
87 certain peoples' aesthetic sense. [WD]
89 * Ditto with the befs_read_block() interface. [WD]
92 Version 0.6 (2001-12-15)
94 * Cleanup of NLS functions (util.c) [WD]
96 * Make directory lookup/read use the NLS if an iocharset is provided. [WD]
98 * Fixed stupid bug where specifying the uid or gid mount options as '0'
99 would result in the filesystem using the on-disk uid and gid. [WD]
101 * Added mount option to control debug printing.
102 The option is, simply enough, 'debug'.
103 (super.c, debug.c) [WD]
105 * Removed notion of btree handle from btree.c. It was unnecessary, as the
106 linux VFS doesn't allow us to keep any state between calls. Updated
107 dir.c, namei.c befs_fs.h to account for it. [WD]
109 * Improved handleing of overflow nodes when listing directories.
110 Now works for overflow nodes hanging off of nodes other than the root
111 node. This is the cleaner solution to Brent Miszalaski's problem. [WD]
113 * Added new debug/warning/error print functions in debug.c.
114 More flexible. Will soon be controllable at mount time
117 * Rewrote datastream position lookups.
120 * Moved the TODO list to its own file.
123 Version 0.50 (2001-11-13)
125 * Added workaround for mis-understanding of the nature of the b+trees used
126 in directories. A cleaner solution will come after I've thought about it
127 for a while. Thanks to Brent Miszalaski for finding and reporting this bug.
132 * Added test for "impossible" condition of empty internal nodes in
133 seekleaf() in btree.c [WD]
135 * Implemented the abstracted read_block() in io.c [WD]
137 * Cleaned up the inode validation in inode.c [WD]
139 * Anton Altaparmakov figured out (by asking Linus :) ) what was causing the
140 hanging disk io problem. It turns out you need to have the sync_pages
141 callback defined in your address_space_ops, even if it just uses the
142 default linux-supplied implementation. Fixed. Works now.
145 * Anton Altaparmakov and Christoph Hellwig alerted me to the fact that
146 filesystem code should be using GFP_NOFS instead of GFP_KERNEL as the
147 priority parameter to kmalloc(). Fixed.
148 (datastream.c, btree.c super.c inode.c) [WD]
150 * Anton also told me that the blocksize is not allowed to be larger than
151 the page size in linux, which is 4k i386. Oops. Added a test for
152 (blocksize > PAGE_SIZE), and refuse to mount in that case. What this
153 practically means is that 8k blocksize volumes won't work without a major
154 restructuring of the driver (or an alpha or other 64bit hardware). [WD]
156 * Cleaned up the befs_count_blocks() function. Much smarter now.
157 And somewhat smaller too. [WD]
159 * Made inode allocations use a slab cache
160 (super.c inode.c) [WD]
162 * Moved the freeing of the private inode section from put_inode() to
163 clear_inode(). This fixes a potential free twice type bug. Put_inode()
164 can be called multiple times for each inode struct. [WD]
166 * Converted all non vfs-callback functions to use befs_sb_info as the
167 superblock type, rather than struct super_block. This is for
170 * Fixed a couple of compile warnings due to use of malloc.h, when slab.h
171 is the new way. (inode.c, super.c) [WD]
173 * Fixed erronous includes of linux/befs_fs_i.h and linux/befs_fs_sb.h
176 Version 0.45 (2001-10-29)
178 * Added functions to get the private superblock and inode structures from
179 their enclosing public structures. Switched all references to the
180 private portions to use them. (many files) [WD]
182 * Made read_super and read_inode allocate the private portions of those
183 structures into the generic pointer fields of the public structures
184 with kmalloc(). put_super and put_inode free them. This allows us not
185 to have to touch the definitions of the public structures in
186 include/linux/fs.h. Also, befs_inode_info is huge (because of the
187 symlink string). (super.c, inode.c, befs_fs.h) [WD]
189 * Fixed a thinko that was corrupting file reads after the first block_run
190 is done being read. (datastream.c) [WD]
192 * Removed fsync() hooks, since a read-only filesystem doesn't need them.
195 * Fixed befs_readlink() (symlink.c) [Christoph Hellwig].
197 * Removed all the Read-Write stuff. I'll redo it when it is time to add
198 write support (various files) [WD].
200 * Removed prototypes for functions who's definitions have been removed
204 Version 0.4 (2001-10-28)
206 * Made it an option to use the old non-pagecache befs_file_read() for
207 testing purposes. (fs/Config.in)
209 * Fixed unused variable warnings when compiling without debugging.
211 * Fixed a bug where the inode and super_block didn't get their blockbits
212 fields set (inode.c and super.c).
214 * Release patch version 11. AKA befs-driver version 0.4.
216 * Thats right. New versioning scheme.
217 I've done some serious testing on it now (on my box anyhow), and it
218 seems stable and not outragously slow. Existing features are more-or-less
219 correct (see TODO list). But it isn't 1.0 yet. I think 0.4 gives me some
220 headroom before the big 1.0.
225 * Fixed date format in this file. Was I smoking crack?
227 * Removed old datastream code from file.c, since it is nolonger used.
229 * Generic_read_file() is now used to read regular file data.
230 It doesn't chew up the buffer cache (it does page io instead), and seems
231 to be about as fast (even though it has to look up each file block
232 indivdualy). And it knows about doing readahead, which is a major plus.
233 So it does i/o in much larger chunks. It is the correct linux way. It
234 uses befs_get_block() by way of befs_readpage() to find the disk offsets
235 of blocks, which in turn calls befs_fpos2brun() in datastream.c to do
236 the hard work of finding the disk block number.
238 * Changed method of checking for a dirty filesystem in befs_read_super
239 (super.c). Now we check to see if log_start and log_end differ. If so,
240 the journal needs to be replayed, and the filesystem cannot be mounted.
242 * Fixed an extra instance of MOD_DEC_USE_COUNT in super.c
244 * Fixed a problem with reading the superblock on devices with large sector
245 sizes (such as cdroms) on linux 2.4.10 and up.
249 * Fix nasty bug in converting block numbers to struct befs_inode_addr.
250 Subtle, because the old version was only sometimes wrong.
251 Probably responsible for lots of problems. (inode.c)
253 * Fix bug with reading an empty directory. (btree.c and dir.c)
255 * This one looks good. Release patch version 10
259 * Added btree searching function.
261 * Use befs_btree_find in befs_lookup (namei.c)
263 * Additional comments in btree.c
267 * Added B+tree reading functions (in btree.c).
268 Made befs_readdir() use them them instead of the cruft in index.c.
272 * Converted befs_read_file() to use the new datastream code.
274 * Finally updated the README file.
276 * Added many comments.
280 * Removed byte-order conversion code.
281 I have no intention of supporting it, and it was very ugly.
282 Flow control with #ifdef (ugh). Maybe I'll redo it once
283 native byteorder works 100%.
287 * Finished implementing read_datastream()
289 * made befs_read_brun() more general
290 Supports an offset to start at and a max bytes to read
291 Added a wrapper function to give the old call
295 * Discovered that the datastream handleing code in file.c is quite deficient
296 in several respects. For one thing, it doesn't deal with indirect blocks
298 * Rewrote datastream handleing.
300 * Created io.c, for io related functions.
301 Previously, the befs_bread() funtions lived in file.c
302 Created the befs_read_brun() function.
307 * Made a function to actually count the number of fs blocks used by a file.
308 And helper functions.
313 * Fixed a misunderstanding of the inode fields.
314 This fixed the problmem with wrong file sizes from du and others.
315 The i_blocks field of the inode struct is not the number of blocks for the
316 inode, it is the number of blocks for the file. Also, i_blksize is not
317 necessarily the size of the inode, although in practice it works out.
318 Changed to blocksize of filesystem.
321 * Permanently removed code that had been provisionally ifdefed out of befs_fs.h
323 * Since we don't support access time, make that field zero, instead of
327 * Added sanity check for inode reading
328 Make sure inode we got was the one we asked for.
332 Local pointers to commonly used structures in inode.c.
333 Got rid of abominations befs_iaddr2inode() and befs_inode2ino().
334 Replaced with single function iaddr2blockno().
335 (fs/befs/super.c) (fs/befs/inode.c)
339 * Fixed the problem with statfs where it would always claim the disk was
340 half full, due to improper understanding of the statfs fields.
343 * Posted verion 4 of the patch
347 * Changed the macros in befs_fs.h to inline functions.
348 More readable. Typesafe. Better
349 (include/linux/befs_fs.h)
351 * Moved type definitions from befs_fs.h to a new file, befs_fs_types.h
352 Because befs_fs_i.h and befs_fs_sb.h were including befs_fs.h for the
353 typedefs, and they are inlcuded in <linux/fs.h>, which has definitions
354 that I want the inline functions in befs_fs.h to be able to see. Nasty
356 (include/linux/befs_fs.h)
360 * Cleaned up some wording.
362 * Added additional consitency checks on mount
363 Check block_size agrees with block_shift
364 Check flags == BEFS_CLEAN
367 * Tell the kernel to only mount befs read-only.
368 By setting the SB_RDONLY flag in befs_read_super().
369 Not that it was possible to write before. But now the kernel won't even try.
372 * Got rid of kernel warning on mount.
373 The kernel doesn't like it if you call set_blocksize() on a device when
374 you have some of its blocks open. Moved the second set_blocksize() to the
375 very end of befs_read_super(), after we are done with the disk superblock.
378 * Fixed wrong number of args bug in befs_dump_inode
381 * Solved lots of type mismatches in kprint()s
386 * Cleaned up the fs/Config.in entries a bit, now slightly more descriptive.
388 * BeFS depends on NLS, so I made activating BeFS enable the NLS questions
391 * Added Configure.help entries for CONFIG_BEFS_FS and CONFIG_DEBUG_BEFS
392 (currently at fs/befs/Kconfig)
396 * Removed superblock locking calls in befs_read_super(). In 2.4, the VFS
397 hands us a super_block struct that is already locked.
401 * Will Dyson <will_dyson@pobox.com> is now attempting to maintain this module
402 Makoto Kato <m_kato@ga2.so-net.ne.jp> is original author.Daniel Berlin
403 also did some work on it (fixing it up for the later 2.3.x kernels, IIRC).
405 * Fixed compile errors on 2.4.1 kernel (WD)
406 Resolve rejected patches
407 Accommodate changed NLS interface (util.h)
408 Needed to include <linux/slab.h> in most files
412 * Tried to niceify the code using the ext2 fs as a guide
413 Declare befs_fs_type using the DECLARE_FSTYPE_DEV() macro
415 * Made it a configure option to turn on debugging (fs/Config.in)