-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathRELEASE-NOTES
7487 lines (5397 loc) · 296 KB
/
RELEASE-NOTES
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
E2fsprogs 1.42.9 (December 28, 2013)
====================================
Mke2fs will detect an attempt to create a file system on a loop
mounted file and complain without the -FF option. Also fixed mke2fs
so it allows the creation of a file system on a mounted device with
two -F options, as documented in the man page, instead of three -F
options.
Fixed a large number of bugs in resize2fs, e2fsck, debugfs, and
libext2fs to correctly handle bigalloc and 64-bit file systems. There
were many corner cases that had not been noticed in previous uses of
these file systems, since they are not as common. Some of the bugs
could cause file system corruption or data loss, so users of 64-bit or
bigalloc file systems are strongly urged to upgrade to e2fsprogs
1.42.9.
The tune2fs program will now not allow changing the uuid on a mounted
file system which has the uninit_bg feature enabled. This avoids a
nasty race where the kernel and tune2fs are both retrying rewrite the
group descriptors at the same time, with different ideas about what
the UUID should be.
When e2fsck is rehashing an extent-mapped directory, it's possible
(although very rare) that an extent block will need to be allocated;
fix e2fsck to make sure that the block gets marked as used.
Mke2fs will now properly set the LARGE_FILE feature when creating a
journal >= 2GB --- which can happen when using 64k block size.
Fixed debugfs so that its freei command will correctly handle a request
to free a range of consecutive inodes.
Fixed 2fsck so it will not erroneously complain that an external journal
is invalid if it is exactly 2**32 blocks.
Fixed e2fsck so it won't try checking for, and adding, a missing
lost+found directory when running in read-only mode.
Fixed e2image to avoid some buffer overruns which would cause it to when
creating a "standard (non-raw, non-qcow2) image file. Standard
e2image files are actually very rarely used, so we didn't notice when
the changes to enable the qcow2 format broke this e2image mode.
Fixed mke2fs so that the extended option "-E resize=NNN" will not turn
on the resize_inode feature when the meta_bg feature is set, since
these two features can not be set at the same time.
Fixed tune2fs so that when it disables the quota feature, it updates all
of the backup superblocks.
Fixed dumpe2fs that would cause it to abort when run using an image file
when trying to print the journal information (which is not present in
an e2image created image file).
Fixed a potential integer overflow in e2reefrag.
(Addresses-Debian-Bug: #718205)
Enhance debugfs so that when copying a sparse file from a native file
system into the file system image using the "write" command, it will
create a sparse file into the destination file system.
Enhanced debugfs so it can support a command line which is up to 8k
long.
E2image will refuse (unless the -f option is specified to force the
issue) to create a raw or qcow image using a mounted file system,
unless the -f option is specified to force the issue.
E2image has been optimized for using it to efficiently copy a file
system by only copying the allocated blocks, by using the options -ra.
New options (-o and -O) have been added so that a source and
destination offset can be given. The -p option will print progress
information so the user will know how far along the copy is going.
And finally, the new option -c is useful for updating a file system on
an SSD, by avoiding unnecessary writes. E2image can also shift a file
system image by doing an in place move.
Fix a regression introduced in 1.42.8 which would cause e2fsck to
erroneously report uninitialized extents past the EOF (as determined by
i_size) to be invalid.
Fixed resize2fs so under a corner case when an inode has a complex
extent tree, it will not corrupt an interior node in the extent tree.
Fixed resize2fs which would sometimes corrupt a file system when
shrinking a file system to a minimum size using resize2fs -M.
(Addresses Debian Bug: #660793)
Fixed resize2fs so that it will relocate inode table blocks if this
becomes necessary when shrinking the file system.
Fixed resize2fs and e2fsck so they will not crash when hit a failure
while operating on a file system with the MMP feature enabled.
Fixed a bug in debugfs which caused it create an invalid inode when
trying to write a zero-length file.
E2fsck will no longer crash if it tries to delete an invalid
extent-mapped symlink.
E2fsck will no longer crash if it comes across an directory which is
larger than 2GB (which is not allowed in valid file systems).
Fixed debugfs's help texts to fully document all options and otherwise
be more helpful.
Updated/fixed various man pages. (Addresses Debian Bugs: #586218,
#669730, #698076)
Fixed various Debian Packaging Issues (#698879, #721365)
Programmer's Notes
------------------
Fix sparse, gcc -Wall and clang nits.
Update config.{guess,sub} to the latest version
Fixed various memory and file descriptor leaks on various error paths,
as well as some missing error return checks, which were found using
Coverity.
Run sparse against source files when building e2fsprogs with 'make
C=1'. If instead C=2, it configures basic ext2 types for bitwise
checking with sparse, which can help find the (many many) spots where
conversion errors are (possibly) happening.
Allow the regression test to be run in chrooted environments where
/etc/mtab might be missing.
The ext2fs_punch() function, which was introduced in 1.42, was broken
in many ways, but this was never noticed since it wasn't used for
anything significant. Some of the bugs include failing when trying to
punch a completely sparse file, failing when punching an extent-mapped
inode when the starting block was at the beginning of the inode, and
not being able to punch a single block (where start_blk == end_block).
It also didn't handle being passed an invalid, too-large ending block
number, and didn't handle properly terminate at the right place when
operating on an indirect-mapped inode.
Fixed some minor typo's in the error catalog for libext2fs.
Fixed ext2fs_file_set_size2() so that if it truncates an inode by
setting the file size, to zero the rest of the block to the end of the
file, so that if an ext4 FUSE driver tries to extended the file, that
we don't avoid stale data from being returned.
Fixed ext2fs_bmap() to disallow clients from trying to map or set
logical blocks which are larger than what an extent-mapped or indirect
block-mapped inode can allow.
If debugfs (or some userspace program using libext2fs) creates a file
which is larger than 2GB, make sure the large_file feature flag gets
set.
Fix a bug in ext2fs_link() where if there is multiple empty slots in the
directory which are large enough, the directory entry could get
inserted more than once in the directory.
If quota support is disabled (which is the default), make sure that
all traces of the quota support is removed from usage messages, man
pages, and tune2fs must not be able to enable the quota file system
feature. (Addresses Red Hat Bugzilla: #1010709)
The ext2fs_file_write() now updates i_size on a successful write,
instead of only updating i_size wen the file is closed via
ext2fs_file_close().
Added a shell script, populate-extfs.sh which uses debugfs to populate
an ext2/3/4 file system image from a given directory. It is similar
to the genext2fs program, but it supports ext3 and ext4 file system.
Add changes to the libext2fs library to support block group
descriptors larger than 64 bytes (for future compatibility).
Fixed an off-by-one bug in ext2fs_file_set_size2() so that it will not
leave an extra block in the file when truncating the file down to
size.
The html version info pages are now built using makeinfo --html
instead of the unmaintained and now-obsolete texi2html program.
E2fsprogs 1.42.8 (June 20, 2013)
================================
As a part of mke2fs's option parsing cleanup, the use of the -R option
will give a warning that it is depreated (it has been so documented
since 2005) and -E should be used instead.
Mke2fs will not give warnings about the bigalloc and quota options in
quiet mode.
If an invalid journal size is given to mke2fs, it will now complain
and exit sooner.
Debugfs was erroneously giving spurious error messages for certain
extent_inode subcommands which take arguments (split_node,
replace_node, and insert_node). This has been fixed.
Fix the parsing of the 's' (sectors) in parse_num_blocks2, which among
other programs is used by mke2fs.
Change mke2fs so that it does not set the root directory to the real
uid/gid of the mke2fs process. Add the extended option root_owner to
override this behavior.
Fix resize2fs when shrinking file systems to make sure that bitmap
blocks aren't left outside the bounds of the shrunken file system.
This could happen with flex_bg file systems that were grown using
the old online resizing algorithm.
E2fsck will now detect and repair corrupted extent trees which contain
invalid extents at the end of the extent tree leaf block.
E2fsck will now longer complain about zero length extended attribute
values.
Fix a regression introduced in e2fsprogs v1.42 which caused e2image -s
to crash.
Add safety check so tune2fs will not attempt to set the inode size to
be larger than the block size.
Fix e2fsck so it can check a read-only root file system with an
external journal. (Addresses Debian Bug: #707030
Fix off-line resizing of file systems with flex_bg && !resize_inode
(Addresses Debian Bug: #696746)
Fix e2image with large (> 32-bit) file systems (Addresses Debian Bug:
#703067)
Enhance chattr to allow clearing the extent flag if the kernel allows
migrating extent based files to use indirect blocks.
Update German translation.
Updated/fixed various man pages. (Addresses Debian Bugs: #712429,
#712430, #707609)
Fixed various Debian Packaging Issues (Addresses Debian Bug #708307)
Programmer's Notes
------------------
Use secure_getenv() in preference to __secure_getenv().
Optimize CPU utilization of ext2fs_bg_has_super().
Fix ext2fs_llseek() on 32-bit i386 systems to work correctly when
SEEK_CUR is used with large files.
The ext2fs_read_inode_full() function will no longer use
fs->read_inode() if the caller has requested more than the base 128
byte inode structure and the inode size is greater than 128 bytes.
Fix build failure with --enable-jbd-debug.
Clean up filtering of outputs for the regression tests by using a
common sed script.
Fix gcc -Wall and clang nits.
E2fsprogs 1.42.7 (January 21, 2013)
===================================
Add warnings to mke2fs, resize2fs, and tune2fs that the bigalloc and
quota features are still under development. For more information
please see:
* https://ext4.wiki.kernel.org/index.php/Bigalloc
* https://ext4.wiki.kernel.org/index.php/Quota
Add some new options to filefrag from Lustre's patches to e2fsprogs:
* add -k option to print extents in kB-sized units (like df -k)
* add -b {blocksize} to print extents in blocksize units
* add -e option to print extent format, even when FIBMAP is used
* add -X option to print extents in hexadecimal format
Fix resize2fs so that it can handle off-line resizes of file systems
with the flex_bg feature but without a resize_inode (or if we run out
of reserved gdt blocks). This also fixes a problem where if the user
creates a filesystem with a restricted number of reserved gdt blocks,
an off-line resize which grows the file system could potentially
result in file system corruption.
Fix a resize2fs bug which could cause it to corrupt bigalloc file
systems when doing an off-line resize.
Further optimize resize2fs so it doesn't use quite as much CPU when
resizing very large file systems.
Fixed 32-bit overflow bugs which could cause resize2fs to fail and
possibly corrupt the file system while resizing 64-bit file systems.
Fix a bug in mke2fs where parsing "-E resize=NNN" will result in a
mke2fs crash due to a divide-by-zero if the 64bit file system feature
is enabled.
Add better error checking to mke2fs to check for invalid parameters
when creating bigalloc file system.
When creating bigalloc filesystems, the -g option to mke2fs will now
specify the number of clusters per block group.
Add to debugfs the functionality to corrupt a specific file system
block via the "zap_block" command.
Add to debugfs the functionality to print out a hex dump of a block in
the file system via the "block_dump" command.
Add to debugfs the functionality to manipulate the extent tree
directly via the "extent_open" command.
Fixed debugfs's mknod command so that it updates the block group
statistics.
Fix e2fsck so it can detect and fix inconsistencies in the interior
nodes of an inode's extent tree.
Fix a potential memory corruption failure in e2fsck's error path if
the call to ext2fs_open2() fails.
Fix e2fsck if its logging function is enabled in e2fsck.conf, and the
resulting file name for the log file is longer than 100 bytes, that it
properly handles this situation instead of crashing.
E2fsck will now report the amount of memory that it attempted to
allocate when a memory allocation request fails, to make it easier to
track down the problem.
Fix mke2fs's handling of the mmp_update_interval option. (Addresses
Lustre Bug: LU-1888)
E2image can now include all data blocks in the e2image output file
when the user specifies the -a option.
If e2fsprogs is compiled without --enable-quota, make sure that
tune2fs can not turn on the feature for new-style quota support, since
afterwards, none of the e2fsprogs tools will be willing to touch that
file system.
Optimize e2fsck's so that it uses much less CPU for large file
systems. This can result in significant speedups, especially on
CPU-constrained systems. This was primarily done by optimizing
libext2fs's bitmap functions. (Addresses Google Bug: #7534813)
Fix debugfs's htree command so that all its messages are sent through
the pager.
Fixed debugfs's "dump_file" and "cat" functions work correctly
on file systems whose block size is greater than 8k.
Fix e2freefrag so it works on 64-bit file systems, and so it uses much
less memory. (Addresses Google Bug: 7269948)
Update the spd_readdir.c file in the contrib directory to include some
additions which were made in 2008 that didn't get folded into the
version which we checked into the e2fsprogs source tree. These
enhancements include thread safety, support for readdir64_r(), and
safe_getenv() support in case spd_readdir.so is used for setuid binaries.
Update Czech, Dutch, French, German, Polish, Sweedish, and Vietnamese
translations
Add a command to debugfs to create symlinks.
Document the bigalloc feature in the mke2fs man page. (Addresses
Debian Bug: #669730)
Programmer's Notes
------------------
Fix gcc -Wall nits.
Fix a spelling typo in the libext2fs texinfo documentation.
Change the output from "make check" so that tools such as emacs's "M-x
compile" does not mistake the output as containing a compiler error.
Export two new functions from libext2fs: ext2fs_extent_node_split()
and ext2fs_extent_goto2().
The ext2fs_extents_fix_parents() was fixed so it does not modify the
location of the extent handle if its current location is at an
interior node, rather than a leaf node of an inode's extent tree.
Add a regression test to assure that e2fsck can correctly fix an
inconsistent interior node in an inode's extent tree.
The ext2fs_{mark,unmark,test}_block_bitmap_range2() functions now
correctly support bigalloc file systems which use store block usage
information in units of clusters.
Fixed the help text in the configure script for --enable-quota.
The m68k-specific bitops code has been removed since they were
incorrectly treating bit numbers with the high bit set as signed
integers. Furthermore, modern compilers do a good enough job
optimizing the generic code there is no point in having the m68k
specific asm statements.
Fixed how we link the test programs so they always use the static
libraries, so that we test using the libraries which we have just
built.
Update config.guess and config.sub to the latest versions from the GNU
project.
Fixed the com_err.texinfo file so that it can produce a valid
postscript/pdf printed output.
Add a regression test which checks debugfs's ability to create
symlinks, named FIFO's, and device nodes.
Add a function ext2fs_symlink(), which creates symlinks to the
libext2fs library.
Add debugging code so we can test old kernel interfaces for online
resize to resize2fs. This backwards compatibility checking is keyed
off of the RESIZE2FS_KERNEL_VERSION, which is designed to allow us to
test the functionality of the kernel's older resize ioctls without
needing to install an old version of resize2fs, and to also test a
modern resize2fs's ability to work with older kernels without having
to install an older version of the kernel.
E2fsprogs 1.42.6 (September 21, 2012)
=====================================
When mke2fs creates file systems with lazy itable initialization, the
progress updates for writing the inode table happens so quickly that
on a serial console, the time to write the progress updates can be the
bottleneck. So mke2fs will now only update its progress indicators
once a second.
Resize2fs will skip initializing the inode tables if the kernel
supports lazy_itable_init, which speeds up growing off-line growth of
uninit_bg file systems significantly. Resize2fs will now also
correctly set the itable_unused field in the block group descriptor to
speed up the first e2fsck after the file system is grown.
Resize2fs has been fixed so that on-line resizing of meta_bg file
systems work correctly. This is needed to grow file system which are
larger than 16T.
Resize2fs will now correctly handle resizing file systems to 16TB on
32-bit file systems when "16TB" is specified on the command line.
Fix mke2fs so that it will be careful to set the reserved blocks ratio
larger than 50%; this can happen when creating small file systems and
when the last block group is dropped because there are not enough
blocks to support the metadata blocks in the last block group.
Fixed spelling mistake in debugfs's help message.
Fixed a potential seg fault in e2fsck when there is an I/O error while
reading the superblock.
Fixed various Debian Packaging Issues (Addresses Debian Bug #677497)
Updated/fixed various man pages. (Addresses Sourceforge Bug:
#3559210)
Programmer's Notes
------------------
The configure option --enable-relative-symlink was broken so that it
needed to be --enable-symlink-relative-symlinks. We will support both
for at least two years, but then the worng configure option will be
removed.
Fixed a regression introduced in 1.42.5 so the link order for
e2fsprogs' libraries will be correct for both static and shared
linking. (Addresses Sourceforge Bug: #3554345)
Add support for e2fsprogs to be compiled using clang/LLVM.
Fix portability problems on non-Linux systems: avoid compile failures
on systems that don't have malloc.h
Fix f_mmp regression test suite so that debugfs gets killed if the
test is interrupted.
E2fsprogs 1.42.5 (July 29, 2012)
================================
Fixed a bug with mke2fs where if there is only 8 inodes per block
group, the calculation of the number of uninitialized inodes in the
first block group would go negative. This resulted in "mke2fs -N 256
-t ext4 /tmp/foo.img 256m" trying to write so many blocks that /tmp
would run out of space. (Addresses Sourceforge Bug: #3528892)
Fixed a bug in how e2fsck would uniquify directory entry names.
(AddressesSourceforge Bug: #3540545)
Previously, e2fsck would only allow a mounted file system to be
checked if it was the root file system and it was mounted read-only.
Now it will allow any file system mounted read-only to be checked if
the -f option is specified. This makes it easier to test how e2fsck
handles checking file systems which are mounted without having to test
on the root file system.
Fixed a problem if e2fsck where if the root file system is mounted
read-only, e2fsck would not clear an error indication in the journal
superblock. Combined with a kernel bug, this would cause the e2fsck
to check the file system after every single boot.
The e4defrag program can now handle device symlinks, such as
/dev/mapper/testvg-testlv, instead of insisting on a less
human-friendly name such as /dev/dm-2. (Addresses Red Hat Bugzilla:
#707209)
Fixed filefrag so it will not crash with a segfault on files from a
virtual file system such as /proc. (e.g., "filefrag
/proc/partitions")
Fixed filefrag so that it correctly reports the number of extents.
(Addresses Red Hat Bugzilla: #840848)
Fixed a file descriptor leak in logsave which could cause it to hang.
(Addresses Debian Bug: #682592)
Fixed e2fsck so that the file system is marked as containing an error
if the user chooses not to fix the quota usage information.
Fixed tune2fs so that it correctly removes the quota feature when the
last quota inode is removed.
Fix tune2fs so that after removing a quota inode, the block bitmap is
updated; otherwise, e2fsck would complain after running 'tune2fs -O
^quota <dev>'.
Fix tune2fs so that when converting a file system from using legacy
quota files to the new quota file system feature with hidden quota
files, the accounting for these files is handled correctly so that
e2fsck doesn't complain.
Improved e2fsck's verbose reporting statistics, and allow the more
verbose reporting to be enabled via /etc/e2fsck.conf.
Fixed various Debian Packaging Issues (Addresses Debian Bug #678395)
Updated/fixed various man pages. (Addresses Debian Bugs: #680114)
Programmer's Notes
------------------
Fixed portability problems on other operating systems (e.g., Hurd and
FreeBsd) caused by the attempted inclusion of <sys/quota.h>.
Make sure that shared libraries link with the shared libraries built
in the build tree, instead of the system provided libraries.
Previously, libraries and executables were linked with the system
libraries if present, and possibly using static archives instead of
shared libraries. This was also problematic since if libext2fs.so is
linked with a static libcom_err.a from system, the build system would
attempt to link without -lpthread. (Addresses Sourceforge Bug:
#3542572)
E2fsprogs 1.42.4 (June 12, 2012)
================================
Fixed more 64-bit block number bugs (which could end up corrupting
file systems!) in e2fsck, debugfs, and libext2fs.
Fixed e2fsck's handling of the journal's s_errno field. E2fsck was
not properly propagating the journal's s_errno field to the superblock
field; it was not checking this field if the journal had already been
replayed, and if the journal *was* being replayed, the "error bit"
wasn't getting flushed out to disk.
Fixed a false positive complaint by e2fsck if all of the extents in
the last extent tree block are uninitalized and located after the end
of the file as defined by i_size.
The dumpe2fs will now display the journal's s_errno field if it is
non-zero, and it will also display the journal's 64-bit feature flag
if present.
Fix e2fsck so that it always opens the device file in exclusive mode
when it might need to modify the file system, and never if the -n
option is specified. (Previously, there were a few corner cases where
it might get things wrong in either direction.)
E2fsck now correctly truncates or deallocates extent-mapped inodes on
the orphan list. The root cause was a bug in libext2fs's block
iterator which could end up skipping an extent when the last block in
an extent is removed, causing the current extent to be removed from
the extent tree.
E2fsck now correctly sets the global free block and inode counts when
truncating or removing inodes on the orphan list in preen mode.
Previously, it would leave these values would be set incorrectly,
which is largely a cosmetic issue since the kernel no longer pays
attention to those fields, but it can cause spurious complaints in
subsequent e2fsck runs.
Fix i_blocks accounting when the libext2fs library needs to add or
remove an extent tree block on bigalloc file systems.
The lsattr and chattr programs now support the No_COW flag for the
benefit of btrfs.
Debugfs now interprets date strings of the form @ddd as ddd seconds
after the beginning of the epoch. This is handy when setting an inode
number into the d_time field when debugging orphan list handling.
Fix a precedence bug with built-in quota support which might result in
e2fsck paying attention to the quota inode field even if the built-in
quota feature flag is not set. Fortunately, in practice that
superblock field should be zero for non-built-in quota file systems,
so it's unlikely this bug would have caused problems.
Updated/fixed various man pages. (Addresses Debian Bugs: #674453,
#674694)
Programmer's Notes
------------------
The regression test suite can now run the integration tests in the
tests directory in parallel, via "make -jN check".
Add new test, f_zero_extent_length which tests e2fsck's handling of
the case where all of the extents in the last extent tree block are
uninitialized extents after i_size.
Add a new test, f_jnl_errno, which checks handling of an error
indication set in the journal superblock.
Fix the test f_jnl_64bit so that it properly checks e2fsck's handling
of a 64-bit journal.
Add two tests, f_orphan_indirect_inode and f_orphan_extent_inode which
tests e2fsck's handling of orphan inodes in preen mode, and truncation
of extent inodes on the orphan list.
Fixed more OS X portability issues.
E2fsprogs 1.42.3 (May 14, 2012)
===============================
Fix a bug in the Unix I/O manager which could cause corruption of file
systems with more than 16TB when e2fsprogs is compiled in 32-bit mode
(i.e., when unsigned long is 32-bits). Also fix a bug which caused
dumpe2fs to incorrectly display block numbers > 32-bits.
Improve the support for integrated quota files (where quota is a first
class supported feature using hidden files in the ext4 file system).
Previously the quota file was getting rewritten even when it was not
necessary, and e2fsck would erroneously try to hide quota files which
were already hidden.
Quiet complaints in e2fsck when the total free blocks or inodes are
incorrect in the superblock after an system crash, since we don't
update nor depend on the superblock summaries at each commit boundary.
Fixed a regression introduced in 1.42.2 which would cause applications
outside of e2fsprogs which did not pass the EXT2_FLAG_64BIT (and so
would were still using the legacy 32-bit bitmaps) to crash. This was
due to missing 32-bit compat code in side the function
ext2fs_find_first_zero_generic_bmap(). (Addresses Red Hat Bugzilla:
#808421)
Fix a bug which would cause mke2fs to fail creating the journal if
/etc/mtab and /proc/mounts are missing. (Addresses Sourceforge Bug:
#3509398)
Updated/fixed various man pages.
Update Czech, Dutch, French, German, Polish, Sweedish, and Vietnamese
translations
Fixed various Debian Packaging issues.
Programmer's Notes
------------------
Change the nonsensically wrong types in the function signature of the
inline function ext2fs_find_first_zero_block_bitmap2(). This was
caused by a cut and paste error; fortunately no code in e2fsprogs used
this inline function, and there are any users of this functions
outside of e2fsprogs.
Add support for systems which have valloc(), but which do have
posix_memalign() nor memalign() (such as MacOS 10.5).
Refactor and clean up the allocation of aligned buffers for Direct I/O
support. Previously some allocations were requesting a greater
alignment factor that what was strictly necessary. Also optimize
reading and writing bitmaps using Direct I/O when the size of the
bitmap did not fully cover the file system blocksize.
Reserve the codepoints for the INCOMPAT features LARGEDATA and
INLINEDATA.
Improved the regression test suite by adding some new integration
tests (f_jnl_32bit, f_jnl_64bit) which detect breakage of the on-disk
jbd2 format, as well as f_eofblocks which tests the new handling of
uninitialized and initialized blocks beyond i_size. Also add a new
unit test which verifies 32-bit bitmaps support and the new
find_first_zero primitives.
Add a few dependencies to fix parallel (make -j) builds.
Removed bash'isms which were breaking the regression test suite on
systems where /bin/sh is not bash.
The config.guess and config.sub have been updated to the 2012-02-10
version.
Fix a portability problem caused by assuming the present of mntent.h
means that setmntent() exists. Instead, explicitly test for this in
the configure script.
If the sys/signal.h header file does not exist, don't try to include
it, since it's not available on all systems.
Add support for systems that do not support getpwuid_r()
The configure script now supports a new option,
--enable-relative-symlinks, which will install relative symlinks for
the ELF shared library files. (Addresses Sourceforge Bug: #3520767).
When building BSD shared libraries make sure the LDFLAGS variable is
passed to the linker. Fixing this allows, for example, e2fsprogs to
be built in 32-bit mode on Mac OS X Lion (Addresses Sourceforge Bug:
#3517272)
Fix gcc -Wall nitpicks
E2fsprogs 1.42.2 (March 27, 2012)
=================================
The resize2fs program uses much less CPU and is much faster for very
large file systems. (Addresses Debian Bug: #663237)
The seti and freei commands in debugfs can now take an optional length
argument to set and clear a contiguous range of inodes.
E2fsck will now make explicit checks for the EOFBLOCKS_FL, since we
plan to remove support for it from the kernel file system driver. It
really wasn't very useful and was causing more problems than it
solves. Since e2fsck will complain if inodes that should have
EOFBLOCKS_FL do not have the flag set, we are going to remove this
check from e2fsprogs first, and then only remove the flag from the
kernel much later.
The mke2fs program can now use direct I/O via "mke2fs -D". This will
slow down the mke2fs, but it makes it more polite on a loaded server
by limiting the amount of memory that gets dirtied by mke2fs when it
is using buffered I/O.
E2fsck was needlessly closing and re-opening the file system as a side
effect of adding Multiple Mount Protection (MMP). This isn't
necessary for non-MMP file systems, so drop it.
Print errors returned by ext2fs_open2() and ext2fs_check_desc() so we
can more easily diagnose memory allocation failures caused by
insufficient memory. E2fsck will now abort if there are memory
allocation failures when the file system is initially opened and
during the block group descriptor checks. (Addresses Google Bug:
#6208183)
If there are incorrect block group checks, e2fsck will now report the
incorrect and corrected checksum values.
The e2fsck progam can now write log files containing the details of
the problems that were found and fixed directly, via configuration
parameters in /etc/e2fsck.conf.
Added the ability to limit the number of messages reported by e2fsck
for a given problem type. This avoids a potential bottleneck if there
is a serial console which can cause a boot sequence to take a long
time if e2fsck needs to report many, many file system errors.
The dumpe2fs, debugfs, and tune2fs now use rbtree bitmaps, which cause
them to use much less memory for large file systems.
The dumpe2fs program will now print the expected block group checksum
if it is incorrect. This helps to diagnose problems caused by
incorrect block group checksums.
E2fsck now checks for extents with a zero length, since the kernel
will oops if it comes accross such a corrupted data structure. (See
https://bugzilla.kernel.org/show_bug.cgi?id=42859)
E2fsck has a number of bugs relating to discard that have been fixed.
(1) Fixed a bug which could cause e2fsck to discard portions of the
inode table which were actually in use. (2) E2fsck will now avoid
using discard if the block device doesn't zero data on discard, since
otherwise this could cause problems if the file system gets corrupted
in the future. (3) E2fsck will now avoid using discard when it is run
in read-only mode. (4) Fixed a bug which caused e2fsck to not issue
discards in the last block group.
E2fsck's CPU utilization in pass 5 has been optimized, which will
speed up e2fsck slightly.
E2image will now skip copying uninitialized bitmap and inode table
blocks.
Fixed mke2fs -S so it does not corrupt the first block group's
information.
E2fsck will now check the new sysfs interface to determine if we are
using the battery or AC mains. (Addresses SourceForge Bug: #3439277)
Updated/fixed various man pages. (Addresses Debian Bug: #665427)
Fixed various Debian Packaging issues. (Addresses Debian Bug: #665885)
Programmer's Notes
------------------
Fixed various portability issues for non-Linux systems, particularly
MacOS X, as well as Linux systems running with the just-released glibc
2.15.
Fix file descriptor leak in ext2fs_close() if the file system with
uninit_bg is opened read/only with a backup superblock. (Addresses
SourceForge Bug: #3444351)
Fixed an invalid return in a non-void function in the quota code.
(Addresses SourceForge Bug: #3468423)
Fixed the debian rules file so that the calls to dpkg-buildflags works
when the shell is dash.
The debian package build now uses V=1 so that there is more
information about potential build failures in debian buildd logs.
If the uninit flags get cleared by functions such as
ext2fs_new_inode() or ext2fs_new_block2(), we now make sure the
superblock is marked dirty and the block group descriptor checksum is
updated if necessary.
The debian rules file will now try to load debian/rules.custom of it
exists. This flie can skip various builds for speed reasons if there
is no need for the e2fsck-static or udeb packages. Available
customizations in the rules file includes SKIP_STATIC=yes,
SKIP_BF=yes, and SKIP_DIETLIBC=yes.
In addition, if the file misc/mke2fs.conf.custom.in exists in the
source tree, it will be used instead of the standard misc.conf file in
the upstream sources. This makes it easier for Debian-derived systems
to distribute a custom mke2fs.conf file without having to worry about
merge issues if future versions of e2fsprogs makes changes in the
upstream default version of mke2fs.conf.
E2fsprogs 1.42.1 (February 17, 2012)
===================================
The mke2fs and e2fsck now use significantly less memory when creating
or checking very large file systems. This was enabled by adding
extent-based bitmaps which are stored using a red-block tree, since
block and inode allocations tend to be contiguous.
The command mke2fs -S is used as a last ditch recovery command to
write new superblock and block group descriptors, but _not_ to destroy
the inode table in hopes of recovering from a badly corrupted file
system. So if the uninit_bg feature is enabled, mke2fs -S will now
set the unused inodes count field to zero. Otherwise, e2fsck -fy
after using mke2fs -S would leave the file system completely empty.
Since mke2fs recognizes mke3fs in argv[0] to mean "mkfs.ext3", also
honor "mke4fs" to work the same as "mke2fs.ext4", since RHEL5 has
installed an mke2fs binary using that name.
The usage and help messages for the -G, -t and -T options in mke2fs
have been fixed.
If e2fsck needs to use the backup group descriptors, the
ext2fs_open2() function clears the UNINIT bits to ensure all of the
inodes in the file systems get scanned. However, the code which reset
the UNINIT flags did not also recalculate the checksum, which produced
many spurious (and scary) e2fsck messages. This has been fixed by
resetting cheksums when the UNINIT bits are cleared.
Relax a check in e2fsck which required that the block bitmap to be
initialized when the inode bitmap is in use. This will allow us to
eventually eliminate code from the kernel which forcibly initialized
the block bitmap when the inode bitmap is first used, requiring an
extra journal credit and disk write. (Addresses Google Bug: #5944440)
Make sure rdebugfs (which may be installed setuid or setgid disk) does
not honor environment variables if euid != uid or egid != gid.
Debugfs's ncheck command has been optimized and now is much more
robust with faced with corrupted file systems. The ncheck command
also now has a -c option which will verify the file type information
in the directory entry to see if matches the inode's mode bits. This
is extremely useful when trying to use debugsfs to determine which
parts of the file system metadata can be trusted.
E2image will try to use ftruncate64() to set the i_size for raw
images, instead of writing a single null byte. This avoid allocating
an extra block to the raw image, for those file systems and/or
operating systems that support this. (Linux does.) In addition, fix
a logic bug that caused the file to not be properly extended if the
size of the last hole was exactly an multiple of a megabyte.
Fixed a bug in resize2fs where for 1k and 2k file systems, where
s_first_data_block is non-zero, this wasn't taken into account when
calculate the minimum file system size for use with the -M option.
Fixed the badblocks program to honor the -s flag when in read-only -t
mode. (Addresses Debian Bug #646629)
Update Czech, Dutch, French, Polish, and Sweedish translation from the
Translation Project.
Fixed various Debian Packaging issues so that dpkg-buildflags is used
if present, which allows e2fsprogs to be built with security hardening
flags. (Addresses Debian Bugs: #654457)
Programmer's Notes
------------------
Fix a bug in ext2fs_clear_generic_bmap() when used for 32-bit bitmaps.
This was only an issue for programs compiled against e2fsprogs 1.41
that manipulate bitmaps directly. (Addresses Sourceforge Bugs:
#3451486)
The libext2fs library now uses sysconf() to fetch the page size, instead
of the deprecated getpagesize().
The ext2fs_get_pathname() function will return a partial path if an a
directory in the path is not a directory, displaying it as an inode
number in angle brackets instead of giving up and displaying an error.
This is much more helpful when a user is trying to debug a corrupted
file system.
Codepoints for the RO_COMPAT_REPLICA feature has been reserved.
Added a new library function, ext2fs_file_get_inode_num(), for use by
fuse2fs.
Fixed a bug in ext2fs_file_set_size2() so that when it is truncating a
file, it actually works.
The block iterator now properly honors the BLOCK_ABORT flag for
extent-based flags. Previously, it didn't, which generally made code
be less efficient, but it could cause bugs in ext2fs_link(), for
example, by causing it to insert multiple directory entries.
Fixed an (harmless other than causing a compiler warning) use of an
uninitialized variable in e2fsck's MMP code.
E2fsprogs 1.42 (November 29, 2011)
==================================
This release of e2fsprogs has support for file systems > 16TB. Online
resize requires kernel support which will hopefully be in Linux
version 3.2. Offline support is not yet available for > 16TB file
systems, but will be coming.
This release of e2fsprogs has support for clustered allocation. This
reduces the number of block (now cluster) bitmaps by allocating and
deallocating space in contiguous power-of-2 collections of blocks,
which are called clustered. This is a file system level feature,
called 'bigalloc', which must be enabled when the file system is
initially formatted. It is not backwards compatible with older
kernels.
Added support for the Multi-Mount Protection (MMP) feature.
E2fsck more efficiently uses scratch files for really big file
systems. (This is a feature that has to be turned on explicitly; see
[scratch_files] in the e2fsck.conf man page.)
Fix a bug in e2fsck where if the free blocks and inodes counts are
incorrect, e2fsck would fix them without printing an error message.
This would cause a "*** FILE SYSTEM WAS MODIFIED ***" message without
any explanation of what was fixed.
E2fsck will no longer attempt to clone an extended attribute block in
pass1b handling if the file system does not support extended
attributes.
E2fsck will be more careful accidentally asking the user to continue
if the file system is mounted, so that an escape sequence won't cause
a false positive. (Addresses Debian Bug: #619859)
E2fsck now uses less cpu time in pass 5 when large portions of the
bitmaps are uninitialized.