3 4.2 28 FEB 2011 XZ compression, and compression options support
5 1. Filesystem improvements:
7 1.1 Added XZ compression
8 1.2 Added compression options support
10 2. Miscellaneous improvements/bug fixes
12 1.1 Add missing NO_XATTR filesystem flag to indicate no-xattrs
13 option was specified and no xattrs should be stored when
15 1.2 Add suppport in Unquashfs -stat option for displaying
17 1.3 Remove checkdata entry from Unsquashfs -stat option if a 4.0
18 filesystem - checkdata is no longer supported.
19 1.4 Fix appending bug when appending to an empty filesystem - this
20 would be incorrectly treated as an error.
21 1.5 Use glibc sys/xattr.h include rather than using attr/xattr.h
22 which isn't present by default on some distributions.
23 1.6 Unsquashfs, fix block calculation error with regular files when
24 file size is between 2^32-block_size+1 and 2^32-1.
25 1.7 Unsquashfs, fix sparse file writing when holes are larger than
27 1.8 Add external CFLAGS and LDFLAGS support to Makefile, and allow
28 build options to be specified on command line. Also don't
29 over-write passed in CFLAGS definition.
32 4.1 19 SEPT 2010 Major filesystem and tools improvements
34 1. Filesystem improvements:
36 1.1 Extended attribute support
37 1.2 New compression framework
38 1.3 Support for LZO compression
39 1.4 Support for LZMA compression (not yet in mainline)
41 2. Mksquashfs improvements:
43 1.1 Enhanced pseudo file support
44 1.2 New options for choosing compression algorithm used
45 1.3 New options for controlling extended attributes
46 1.4 Fix misalignment issues with memcpy etc. seen on ARM
47 1.5 Fix floating point error in progress_bar when max == 0
48 1.6 Removed use of get_nproc() call unavailable in ulibc
49 1.7 Reorganised help text
51 3. Unsquashfs improvements:
53 1.1 New options for controlling extended attributes
54 1.2 Fix misalignment issues with memcpy etc. seen on ARM
55 1.3 Fix floating point error in progress_bar when max == 0
56 1.4 Removed use of get_nproc() call unavailable in ulibc
59 4.0 5 APR 2009 Major filesystems improvements
61 1. Kernel code improvements:
63 1.1 Fixed little endian layout adopted. All swapping macros
64 removed, and in-line swapping added for big-endian
66 1.2 Kernel code substantially improved and restructured.
67 1.3 Kernel code split into separate files along functional lines.
68 1.4 Vmalloc usage removed, and code changed to use separately
71 2. Unsquashfs improvements:
73 2.1 Support for 4.0 filesystems added.
74 2.2 Swapping macros rewritten.
75 2.3 Unsquashfs code restructured and split into separate files.
77 3. Mksquashfs improvements:
79 3.1 Swapping macros rewritten. Fixed little-endian layout allows
80 code to be optimised and only added at compile time for
82 3.2 Support for pseudo files added.
84 3.4 26 AUG 2008 Performance improvements to Unsquashfs, Mksquashfs
85 and the kernel code. Plus many small bug fixes.
87 1. Kernel code improvements:
89 1.1 Internal Squashfs kernel metadata and fragment cache
90 implementations have been merged and optimised. Spinlocks are
91 now used, locks are held for smaller periods and wakeups have
92 been minimised. Small race condition fixed where if two or
93 more processes tried to read the same cache block
94 simultaneously they would both read and decompress it. 10-20%+
95 speed improvement has been seen on tests.
96 1.2 NFS export code rewritten following VFS changes in
98 1.3 New patches for linux-2.6.25, linux-2.6.26, and linux-2.6.27.
99 Fixed patch for linux-2.6.24.
100 1.4 Fixed small buffer_head leak in squashfs_read_data when
101 handling badly corrupted filesystems.
102 1.5 Fixed bug in get_dir_index_using_offset.
104 2. Unsquashfs improvements:
106 2.1 Unsquashfs has been parallelised. Filesystem reading, writing
107 and decompression is now multi-threaded. Up to 40% speed
108 improvement seen on tests.
109 2.2 Unsquashfs now has a progress bar. Use -no-progress to
111 2.3 Fixed small bug where unistd.h wasn't being included on
112 some distributions, leading to lseek being used rather than
113 lseek64 - which meant on these distributions Unsquashfs
114 couldn't unsquash filesystems larger than 4GB.
116 3. Mksquashfs improvements:
118 3.1 Removed some small remaining parallelisation bottlenecks.
119 Depending on source filesystem, up to 10%+ speed improvement.
120 3.2 Progress bar improved, and moved to separate thread.
121 3.3 Sparse file handling bug in Mksquashfs 3.3 fixed.
122 3.4 Two rare appending restore bugs fixed (when ^C hit twice).
125 3.3 1 NOV 2007 Increase in block size, sparse file support,
126 Mksquashfs and Unsquashfs extended to use
127 pattern matching in exclude/extract files, plus
128 many more improvements and bug fixes.
130 1. Filesystem improvements:
132 1.1. Maximum block size has been increased to 1Mbyte, and the
133 default block size has been increased to 128 Kbytes.
134 This improves compression.
136 1.2. Sparse files are now supported. Sparse files are files
137 which have large areas of unallocated data commonly called
138 holes. These files are now detected by Squashfs and stored
139 more efficiently. This improves compression and read
140 performance for sparse files.
142 2. Mksquashfs improvements:
144 2.1. Exclude files have been extended to use wildcard pattern
145 matching and regular expressions. Support has also been
146 added for non-anchored excludes, which means it is
147 now possible to specify excludes which match anywhere
148 in the filesystem (i.e. leaf files), rather than always
149 having to specify exclude files starting from the root
150 directory (anchored excludes).
152 2.2. Recovery files are now created when appending to existing
153 Squashfs filesystems. This allows the original filesystem
154 to be recovered if Mksquashfs aborts unexpectedly
155 (i.e. power failure).
157 3. Unsquashfs improvements:
159 3.1. Multiple extract files can now be specified on the
160 command line, and the files/directories to be extracted can
161 now also be given in a file.
163 3.2. Extract files have been extended to use wildcard pattern
164 matching and regular expressions.
166 3.3. Filename printing has been enhanced and Unquashfs can
167 now display filenames with file attributes
168 ('ls -l' style output).
170 3.4. A -stat option has been added which displays the filesystem
171 superblock information.
173 3.5. Unsquashfs now supports 1.x filesystems.
175 4. Miscellaneous improvements/bug fixes:
177 4.1. Squashfs kernel code improved to use SetPageError in
178 squashfs_readpage() if I/O error occurs.
180 4.2. Fixed Squashfs kernel code bug preventing file
183 4.3. Mksquashfs now detects file size changes between
184 first phase directory scan and second phase filesystem create.
185 It also deals better with file I/O errors.
188 3.2-r2 15 JAN 2007 Kernel patch update and progress bar bug fix
190 1. Kernel patches 2.6.19/2.6.20 have been updated to use
191 const structures and mutexes rather than older semaphores.
192 2. Minor SMP bug fixes.
193 3. Progress bar broken on x86-64. Fixed.
195 3.2 2 JAN 2007 NFS support, improvements to the Squashfs-tools, major
196 bug fixes, lots of small improvements/bug fixes, and new
201 1. Squashfs filesystems can now be exported via NFS.
202 2. Unsquashfs now supports 2.x filesystems.
203 3. Mksquashfs now displays a progress bar.
204 4. Squashfs kernel code has been hardened against accidently or
205 maliciously corrupted Squashfs filesystems.
209 5. Race condition occurring on S390 in readpage() fixed.
210 6. Odd behaviour of MIPS memcpy in read_data() routine worked-around.
211 7. Missing cache_flush in Squashfs symlink_readpage() added.
214 3.1-r2 30 AUG 2006 Mksquashfs -sort bug fix
216 A code optimisation after testing unfortunately
217 broke sorting in Mksquashfs. This has been fixed.
219 3.1 19 AUG 2006 This release has some major improvements to
220 the squashfs-tools, a couple of major bug
221 fixes, lots of small improvements/bug fixes,
222 and new kernel patches.
225 1. Mksquashfs has been rewritten to be multi-threaded. It
226 has the following improvements
228 1.1. Parallel compression. By default as many compression and
229 fragment compression threads are created as there are available
230 processors. This significantly speeds up performance on SMP
232 1.2. File input and filesystem output is peformed in parallel on
233 separate threads to maximise I/O performance. Even on single
234 processor systems this speeds up performance by at least 10%.
235 1.3. Appending has been significantly improved, and files within the
236 filesystem being appended to are no longer scanned and
237 checksummed. This significantly improves append time for large
239 1.4. File duplicate checking has been optimised, and split into two
240 separate phases. Only files which are considered possible
241 duplicates after the first phase are checksummed and cached in
243 1.5 The use of swap memory was found to significantly impact
244 performance. The amount of memory used to cache files is now a
245 command line option, by default this is 512 Mbytes.
247 2. Unsquashfs has the following improvements
249 2.1 Unsquashfs now allows you to specify the filename or the
250 directory within the Squashfs filesystem that is to be
251 extracted, rather than always extracting the entire filesystem.
252 2.2 A new -force option has been added which forces Unsquashfs to
253 output to the destination directory even if files and directories
254 already exist in the destination directory. This allows you to
255 update an already existing directory tree, or to Unsquashfs to
256 a partially filled directory tree. Without the -force option
257 Unsquashfs will refuse to output.
259 3. The following major bug fixes have been made
261 3.1 A fragment table rounding bug has been fixed in Mksquashfs.
262 Previously if the number of fragments in the filesystem
263 were a multiple of 512, Mksquashfs would generate an
264 incorrect filesystem.
265 3.2 A rare SMP bug which occurred when simultaneously acccessing
266 multiply mounted Squashfs filesystems has been fixed.
268 4. Miscellaneous improvements/bug fixes
270 4.1 Kernel code stack usage has been reduced. This is to ensure
271 Squashfs works with 4K stacks.
272 4.2 Readdir (Squashfs kernel code) has been fixed to always
273 return 0, rather than the number of directories read. Squashfs
274 should now interact better with NFS.
275 4.3 Lseek bug in Mksquashfs when appending to larger than 4GB
277 4.4 Squashfs 2.x initrds can now been mounted.
278 4.5 Unsquashfs exit status fixed.
279 4.6 New patches for linux-2.6.18 and linux-2.4.33.
282 3.0 15 MAR 2006 Major filesystem improvements
284 1. Filesystems are no longer limited to 4 GB. In
285 theory 2^64 or 4 exabytes is now supported.
286 2. Files are no longer limited to 4 GB. In theory the maximum
287 file size is 4 exabytes.
288 3. Metadata (inode table and directory tables) are no longer
289 restricted to 16 Mbytes.
290 4. Hardlinks are now suppported.
291 5. Nlink counts are now supported.
292 6. Readdir now returns '.' and '..' entries.
293 7. Special support for files larger than 256 MB has been added to
294 the Squashfs kernel code for faster read access.
295 8. Inode numbers are now stored within the inode rather than being
296 computed from inode location on disk (this is not so much an
297 improvement, but a change forced by the previously listed
300 2.2-r2 8 SEPT 2005 Second release of 2.2, this release fixes a couple
301 of small bugs, a couple of small documentation
302 mistakes, and adds a patch for kernel 2.6.13.
304 1. Mksquashfs now deletes the output filesystem image file if an
305 error occurs whilst generating the filesystem. Previously on
306 error the image file was left empty or partially written.
307 2. Updated mksquashfs so that it doesn't allow you to generate
308 filesystems with block sizes smaller than 4K. Squashfs hasn't
309 supported block sizes less than 4K since 2.0-alpha.
310 3. Mksquashfs now ignores missing files/directories in sort files.
311 This was the original behaviour before 2.2.
312 4. Fixed small mistake in fs/Kconfig where the version was still
314 5. Updated ACKNOWLEDGEMENTS file.
317 2.2 3 JUL 2005 This release has some small improvements, bug fixes
318 and patches for new kernels.
320 1. Sort routine re-worked and debugged from release 2.1. It now allows
321 you to give Mkisofs style sort files and checks for filenames that
322 don't match anything. Sort priority has also been changed to
323 conform to Mkisofs usage, highest priority files are now placed
324 at the start of the filesystem (this means they will be on the
325 inside of a CD or DVD).
326 2. New Configure options for embedded systems (memory constrained
327 systems). See INSTALL file for further details.
328 3. Directory index bug fixed where chars were treated as signed on
329 some architectures. A file would not be found in the rare case
330 that the filename started with a chracter greater than 127.
331 4. Bug introduced into the read_data() routine when sped up to use data
332 block queueing fixed. If the second or later block resulted in an
333 I/O error this was not checked.
334 5. Append bug introduced in 2.1 fixed. The code to compute the new
335 compressed and uncompressed directory parts after appending was
337 6. Metadata block length read routine altered to not perform a
338 misaligned short read. This was to fix reading on an ARM7 running
339 uCLinux without a misaligned read interrupt handler.
340 7. Checkdata bug introduced in 2.1 fixed.
343 2.1-r2 15 DEC 2004 Code changed so it can be compiled with gcc 2.x
345 1. In some of the code added for release 2.1 I unknowingly used some
346 gcc extensions only supported by 3.x compilers. I have received
347 a couple of reports that the 2.1 release doesn't build on 2.x and so
348 people are clearly still using gcc 2.x. The code has been
349 rewritten to remove these extensions.
351 2.1 10 DEC 2004 Significantly improved directory handling plus numerous
352 other smaller improvements
354 1. Fast indexed directories implemented. These speed up directory
355 operations (ls, file lookup etc.) significantly for directories
357 2. All directories are now sorted in alphabetical order. This again
358 speeds up directory operations, and in some cases it also results in
359 a small compression improvement (greater data similarity between
360 files with alphabetically similar names).
361 3. Maximum directory size increased from 512 KB to 128 MB.
362 4. Duplicate fragment checking and appending optimised in mksquashfs,
363 depending on filesystem, this is now up to 25% faster.
364 5. Mksquashfs help information reformatted and reorganised.
365 6. The Squashfs version and release date is now printed at kernel
366 boot-time or module insertion. This addition will hopefully help
367 to reduce the growing problem where the Squashfs version supported
368 by a kernel is unknown and the kernel source is unavailable.
369 7. New PERFORMANCE.README file.
370 8. New -2.0 mksquashfs option.
371 9. CHANGES file reorganised.
372 10. README file reorganised, clarified and updated to include the 2.0
374 11. New patch for Linux 2.6.9.
375 12. New patch for Linux 2.4.28.
377 2.0r2 29 AUG 2004 Workaround for kernel bug in kernels 2.6.8 and newer
380 1. New patch for kernel 2.6.8.1. This includes a workaround for a
381 kernel bug introduced in 2.6.7bk14, which is present in all later
382 versions of the kernel.
384 If you're using a 2.6.8 kernel or later then you must use this
385 2.6.8.1 patch. If you've experienced hangs or oopses using Squashfs
386 with a 2.6.8 or later kernel then you've hit this bug, and this
389 It is worth mentioning that this kernel bug potentially affects
390 other filesystems. If you receive odd results with other
391 filesystems you may be experiencing this bug with that filesystem.
392 I submitted a patch but this has not yet gone into the
393 kernel, hopefully the bug will be fixed in later kernels.
395 2.0 13 JULY 2004 A couple of new options, and some bug fixes
397 1. New mksquashfs -all-root, -root-owned, -force-uid, and -force-gid
398 options. These allow the uids/gids of files in the generated
399 filesystem to be specified, overriding the uids/gids in the
401 2. Initrds are now supported for kernels 2.6.x.
402 3. amd64 bug fixes. If you use an amd64, please read the README-AMD64
404 4. Check-data and gid bug fixes. With 2.0-alpha when mounting 1.x
405 filesystems in certain cases file gids were corrupted.
406 5. New patch for Linux 2.6.7.
408 2.0-ALPHA 21 MAY 2004 Filesystem changes and compression improvements
410 1. Squashfs 2.0 has added the concept of fragment blocks.
411 Files smaller than the file block size and optionally the
412 remainder of files that do not fit fully into a block (i.e. the
413 last 32K in a 96K file) are packed into shared fragments and
414 compressed together. This achieves on average 5 - 20% better
415 compression than Squashfs 1.x.
416 2. The maximum block size has been increased to 64K (in the ALPHA
417 version of Squashfs 2.0).
418 3. The maximum number of UIDs has been increased to 256 (from 48 in
420 4. The maximum number of GIDs has been increased to 256 (from 15 in
422 5. Removal of sleep_on() function call in 2.6.x patch, to allow Squashfs
423 to work on the Fedora rc2 kernel.
424 6. Numerous small bug fixes have been made.
426 1.3r3 18 JAN 2004 Third release of 1.3, this adds a new mksquashfs option,
427 some bug fixes, and extra patches for new kernels
429 1. New mksquashfs -ef exclude option. This option reads the exclude
430 dirs/files from an exclude file, one exclude dir/file per line. This
431 avoids the command line size limit when using the -e exclude option,
432 2. When appending to existing filesystems, if mksquashfs experiences a
433 fatal error (e.g. out of space when adding to the destination), the
434 original filesystem is restored,
435 3. Mksquashfs now builds standalone, without the kernel needing to be
437 4. Bug fix in the kernel squashfs filesystem, where the pages being
438 filled were not kmapped. This seems to only have caused problems
440 5. New patch for Linux 2.4.24,
442 6. New patch for Linux 2.6.1, this replaces the patch for 2.6.0-test7.
444 1.3r2 14 OCT 2003 Second release of 1.3, bug fixes and extra patches for
447 1. Bug fix in routine that adds files to the filesystem being
448 generated in mksquashfs. This bug was introduced in 1.3
449 (not enough testing...) when I rewrote it to handle files larger
450 than available memory. This bug caused a SEGV, so if you've ever
451 got that, it is now fixed,
452 2. Long running bug where ls -s and du reported wrong block size
453 fixed. I'm pretty sure this used to work many kernel versions ago
454 (2.4.7) but it broke somewhere along the line since then,
455 3. New patch for Linux 2.4.22,
456 4. New patch for 2.6.0-test7, this replaces the patch for 2.6.0-test1.
458 1.3 29 JUL 2003 FIFO/Socket support added plus optimisations and
461 1. FIFOs and Socket inodes are now supported,
462 2. Mksquashfs can now compress files larger than available
464 3. File duplicate check routine optimised,
465 4. Exit codes fixed in Mksquashfs,
466 5. Patch for Linux 2.4.21,
467 6. Patch for Linux 2.6.0-test1. Hopefully, this will work for
468 the next few releases of 2.6.0-testx, otherwise, I'll be
469 releasing a lot of updates to the 2.6.0 patch...
471 1.2 13 MAR 2003 Append feature and new mksquashfs options added
473 Mksquashfs can now add to existing squashfs filesystems. Three extra
474 options "-noappend", "-keep-as-directory", and "root-becomes"
477 The append option with file duplicate detection, means squashfs can be
478 used as a simple versioning archiving filesystem. A squashfs
479 filesystem can be created with for example the linux-2.4.19 source.
480 Appending the linux-2.4.20 source will create a filesystem with the
481 two source trees, but only the changed files will take extra room,
482 the unchanged files will be detected as duplicates.
484 See the README file for usage changes.
486 1.1b 16 JAN 2003 Bug fix release
488 Fixed readpage deadlock bug. This was a rare deadlock bug that
489 happened when pushing pages into the page cache when using greater
490 than 4K blocks. I never got this bug when I tested the filesystem,
491 but two people emailed me on the same day about the problem!
492 I fixed it by using a page cache function that wasn't there when
493 I originally did the work, which was nice :-)
495 1.1 8 JAN 2003 Added features
497 1. Kernel squashfs can now mount different byte order filesystems.
498 2. Additional features added to mksquashfs. Mksquashfs now supports
499 exclude files and multiple source files/directories can be
500 specified. A nopad option has also been added, which
501 informs mksquashfs not to pad filesystems to a multiple of 4K.
502 See README for mksquashfs usage changes.
503 3. Greater than 2GB filesystems bug fix. Filesystems greater than 2GB
506 1.0c 14 NOV 2002 Bug fix release
508 Fixed bugs with initrds and device nodes
510 1.0 23 OCT 2002 Initial release