]> git.wh0rd.org - dump.git/blob - CHANGES
Buffer overflow in dump, as reported from Bugtraq
[dump.git] / CHANGES
1 $Id: CHANGES,v 1.45 2000/03/01 10:16:05 stelian Exp $
2
3 Changes between versions 0.4b14 and 0.4b15 (released ?????????????????)
4 =======================================================================
5
6 1. Added a prompt command in interactive restore mode. Thanks
7 to Andreas Dilger <adilger@home.com> for the patch.
8
9 2. Fixed a buffer overflow problem in dump (caused by
10 not checking the size of the filesystem parameter).
11 Thanks to Kim Yong-jun <loveyou@hackerslab.org> for
12 reporting this on Bugtraq (and to several dump users
13 who forwarded me his mail).
14
15 Changes between versions 0.4b13 and 0.4b14 (released February 10, 2000)
16 =======================================================================
17
18 1. Fixed a bug in dump which may have caused invalid deleted
19 directories to be dumped out if they were deleted after the
20 mapping pass. This could occure on active filesystem and lead
21 to heap corruption (causing dump malfunction in many possible ways).
22 Thanks to Jan Sanislo <oystr@cs.washington.edu> for finding this
23 bug and submitting the patch.
24
25 2. Fixed the handling of the filesystem parameter in dump. A
26 '/mnt/disk' parameter caused the disk contents to be dumped,
27 but a '/mnt/disk/' parameter caused the mountpoint directory
28 to be dumped (generally an empty directory).
29
30 3. Improved the output of dump in order to tell which directory
31 it is currently dumping (when dumping a subtree).
32
33 4. Added the '-e' exclude inode option to dump. Thanks to
34 Isaac Chuang <ike@isl.stanford.edu> for contributing with the patch.
35
36 5. Added a REPORTING-BUGS file in order to provide a guide
37 on how to correctly report a bug in dump/restore.
38
39 6. Fixed a restore problem when restoring a hard link to an inode
40 having the immutable or append-only attribute set. Thanks to
41 Ambrose Li <acli@mingpaoxpress.com> for submitting the patch.
42
43 7. Fixed a compatibility problem between dumps made on little
44 endian machines (the format was correct) and big endian
45 machines (the format was incorrect). This fix break the
46 compatibility with the older dumps made on big endian
47 machines (sparc, m86k, ppc etc). For the first time in
48 linux dump's history (I believe), the dumps made by *BSD,
49 Linux/alpha, Linux/sparc and Linux/x86 are compatible,
50 so interchangeable. Thanks to Rob Cermak
51 <cermak@ahab.rutgers.edu> for submitting the bug and
52 helping me test the fix.
53
54 8. Fixed the way dump reports the remaining percent/time, if
55 the number of blocks actually dumped exceeds the estimated
56 number of blocks. Thanks to Jean-Paul van der Jagt
57 <jeanpaul@dutepp0.et.tudelft.nl> for reporting the bug.
58
59 Changes between versions 0.4b12 and 0.4b13 (released January 21, 2000)
60 ======================================================================
61
62 1. Small Y2K fix in the man pages :). Thanks to Bernhard Sadlowski
63 <sadlowsk@Mathematik.Uni-Bielefeld.DE> for reporting the bug.
64
65 2. Removed the requirement to build the RPM as root from the
66 spec file. Thanks to Christian Weisgerber
67 <naddy@mips.rhein-neckar.de> for submitting this.
68
69 3. Fixed a bug in dump related to the 'filetype' feature of ext2fs,
70 causing dump to block when dumping really huge filesystems.
71 Many thanks to Patrik Schindler <poc@pocnet.net> for
72 helping me find this bug.
73
74 4. Fixed the treatment for an interrupt signal when dump access
75 the remote tape through RSH. Thanks to Christian Weisgerber
76 <naddy@mips.rhein-neckar.de> for providing the patch.
77
78 5. Fixed a bug which was causing dump/restore to display
79 garbage characters instead of the remote host name.
80
81 Changes between versions 0.4b11 and 0.4b12 (released January 8, 2000)
82 =====================================================================
83
84 1. Small fix in the dump man page. Thanks to Thorsten Kukuk
85 <kukuk@suse.de> for submitting the patch.
86
87 2. Fix for the exit code when using the size estimate option of
88 dump. Thanks to Matti Taskinen <mkt@rni.helsinki.fi> for
89 submitting the patch.
90
91 3. Handle EINTR in atomical reads/writes in dump, which was causing
92 dump to fail on some systems. Thanks to Eric Jergensen
93 <eric@dvns.com> for reporting the bug and submitting the patch.
94
95 4. Handle more than 16 characters for the device names in dumpdates.
96 (up to 255 now). Thanks to Rainer Clasen <bj@ncc.cicely.de> for
97 tracking down the problem and proposing the solution.
98
99 5. Fixed a bug in dump which prevented the creation of the
100 dumpdates file when doing a 0-level dump without already
101 having a dumpdates file. Thanks to Patrik Schindler
102 <poc@pocnet.net> for reporting the bug.
103
104 6. Changed the way dump 'S' flag reports the size estimate
105 from number of blocks into bytes (making it compatible
106 with the Solaris version, and simplifying things for
107 amanda users). Thanks to Jason L Tibbitts III
108 <tibbs@math.uh.edu> for reporting the bug.
109
110 7. Fixed a compatibility problem in linux/alpha dump tape format.
111 Now the linux/alpha dump are (again) compatible with the
112 other dump formats. But this breaks compatibility with
113 older dumps made on alpha. Thanks to Mike Tibor
114 <tibor@lib.uaa.alaska.edu> for helping me in finding this bug.
115
116 Changes between versions 0.4b10 and 0.4b11 (released December 5, 1999)
117 ======================================================================
118
119 1. Added a '--enable-kerberos' to configure.
120
121 2. Added a 'S' option to dump which determines the amount of space
122 that is needed to perform the dump without actually doing it, similar
123 to the Sun's ufsdump 'S' option. Patch contributed by Rob Cermak
124 <cermak@ahab.rutgers.edu>.
125
126 3. Added a 'M' multi-volume option to dump and restore which enables
127 dumping to multiple files (useful when dumping to an ext2
128 partition to make several dump files in order to bypass the 2GB
129 file size limitation). The 'f' argument is treated as a prefix and
130 the output files will be named <prefix>001, <prefix>002 etc. With
131 the 'M' flag, restore automatically selects the right file without
132 asking to enter a new tape each time.
133
134 4. Fixed a memory leak which was causing dump to grow very big
135 (270MB when dumping a 10GB filesystem...). Thanks to Jason
136 Fearon <jasonf@netrider.org.au> for reporting the bug.
137
138 Changes between versions 0.4b9 and 0.4b10 (released November 21, 1999)
139 ======================================================================
140
141 1. Make configure test if the system glob routines support
142 extended syntax (ALTDIRFUNC). If not, use the internal glob
143 routines instead of system ones. Thanks to Bernhard Sadlowski
144 <sadlowsk@Mathematik.Uni-Bielefeld.DE> for reporting the bug
145 and helping me resolve this and other minor libc5 compiling
146 glitches.
147
148 2. Fix a problem when dumping a ext2fs with the 'filetype'
149 feature enabled. Thanks to Patrick J. LoPresti
150 <patl@cag.lcs.mit.edu> for reporting the bug and to
151 Theodore Y. Ts'o <tytso@mit.edu> for providing the patch.
152
153 3. Made the nodump flag work on directories. A directory which
154 has the nodump flag gets never dumped, regardless of its
155 contents.
156
157 4. Integrate a patch from Jeremy Fitzhardinge <jeremy@goop.org>
158 which allows dump on an active ext3 filesystem. However, this
159 is a "quick and dirty" patch which enables backup of an ext3
160 filesystem through the ext2 compatibility (by ignoring the
161 NEEDS_RECOVERY bit). The journal file is not recognized and
162 it is dumped (it should not).
163
164 5. Test the superblock compatibility flags when dumping, in order
165 to be sure that we know how to deal with specific features.
166
167 Changes between versions 0.4b8 and 0.4b9 (released November 5, 1999)
168 ====================================================================
169
170 1. Use lchown instead of chown, fixing a possible security problem
171 when restoring symlinks (a malicious user could use this
172 to deliberately corrupt the ownership of important system files).
173 Thanks to Chris Siebenmann <cks@utcc.utoronto.ca> for detecting
174 this and providing the patch.
175
176 Changes between versions 0.4b7 and 0.4b8 (released November 3, 1999)
177 ====================================================================
178
179 1. Put dump sources under CVS, added Id tags in all files so
180 one can use 'ident' on binary files.
181
182 2. Added the dump/restore version in the usage text so one can
183 easily verify the version he is using.
184
185 3. Small patch from Nuno Oliveira <nuno@eq.uc.pt> which fixes
186 a va_start/va_end problem on linux-ppc (always call va_start
187 va_end in pairs each time we use a vararg function).
188
189 4. Added again the DT_* constants because old libc does not
190 contain them :(. Thanks to Eric Maisonobe <virnet@nat.fr>
191 for submitting the bug report.
192
193 5. Use ext2fs_llseek instead of llseek. With recent e2fsprogs
194 this should enable dumping big (huge) filesystems.
195
196 6. Added the RSH environment variable in order to be able to
197 use a rsh replacement like ssh when doing remote backups (and
198 bypass the security limitations of rcmd). Now you can do remote
199 backups without being root (or making dump setuid root).
200
201 7. Modified again the way dumpdates works. For incremental dumps,
202 we need to read dumpdates even if we are not using 'u' option.
203 Thanks to Bdale Garbee <bdale@gag.com> for his ideas on how
204 this should work.
205
206 Changes between versions 0.4b6 and 0.4b7 (released October 8, 1999)
207 ===================================================================
208
209 1. Removed the 'k' flag from the restore 'about' text if kerberos
210 was not compiled in.
211
212 2. Prototyped (f)setflags from e2fsprogs and corrected the calls
213 to them (fsetflags takes a char*, setflags an open fd!).
214
215 3. (f)setflags is called only if the flags aren't empty. If the
216 file is a special file, a warning is printed, because changing
217 flags implies opening the device. Normally, a special file
218 should not have any flag... (Debian bug #29775, patch provided
219 by Abhijit Dasgupta <abhijit@ans.net>).
220
221 4. Made possible to dump a file system not mentioned in /etc/fstab.
222 (Debian bug #11904, patch provided by Eirik Fuller <eirik@netcom.com>).
223
224 5. Changed the default behaviour to not create dumpdates
225 unless 'u' option is specified. Removed the old "debian-patch"
226 which provided the same thing. (Debian bug #38136, #33818).
227
228 6. Removed all those dump*announce, since they were getting old...
229
230 7. Added warning messages if dumpdates does not exist and
231 when an inferior level dump does not exist (except for a level 0
232 dump).
233
234 8. Debugged the glob calls in interactive mode: restore used a
235 dirent struct which was different from the /usr/include/dirent.h
236 one (this used to work, is it a glibc2 change?), so none of the
237 compat glob (which used /usr/include/dirent.h) or the system glob
238 worked. Restore use now the system dirent (and the system
239 DT_* constants), which are compatible with BSD ones.
240
241 9. Added a configure flag (--with-dumpdatespath) to specify
242 the location of dumpdates file. By default, it is
243 /etc/dumpdates.
244
245 10. Added the "AUTHOR" and "AVAILABILITY" sections and
246 included the current date/version in man pages.
247
248 11. Corrected the estimation of remaining time when
249 the operator doesn't change the tapes quickly enough. This
250 was an old bug, I thought I corrected it, and discovered
251 that in fact it was corrected in two different places, so
252 the results canceled each other...
253
254 Changes between versions 0.4b5 and 0.4b6 (released October 1, 1999)
255 ===================================================================
256
257 1. Integrated multiple patches from RedHat, Debian and SuSE:
258
259 - tweak dump/itime.c to not try to read dumpdates if the 'u' option
260 isn't specified.
261 - several fixes in the man pages.
262 - update the default tape device to /dev/st0.
263 - many updates for Linux Alpha (byte ordering, size_t etc).
264 - buffer overruns.
265 - use environment variable for TMPDIR (instead of /tmp).
266 - use sigjmp_buf instead of jmp_buf (RedHat bug #3260).
267 - workaround egcs bug (RedHat bugs #4281 and #2989).
268 - wire $(OPT) throughout Makefile's.
269
270 2. Upgrade the dump revision to 1, making possible to dump filesystems
271 made with e2fsprogs-1.15 or newer. Nothing seems to break...
272
273 3. Fix some compile warnings, prototype all functions.
274
275 4. Use glibc err/glob instead of internal compatibility
276 routines (only if available).
277
278 5. Fix a compile error on Linux 2.2.7 / libc5 (5.4.44) (patch provided
279 by Bernhard Sadlowski <sadlowsk@mathematik.uni-bielefeld.de>).
280
281 Changes between versions 0.4b4 and 0.4b5 (released September 22, 1999)
282 ======================================================================
283
284 1. Integrated the changes from FreeBSD-3.1-RELEASE
285 (mostly bug fixes, buffer overruns, dump has now an "automatic
286 tape length calculation" flag, dump/restore can use kerberos now
287 (this is NOT tested), use environment variables for TAPE and
288 RMT etc.).
289
290 2. Integrated three RedHat patches ("glibc", "kernel" and "bread" patches)
291
292 3. Corrected a bug in restore when using 'C' option with multi-volumes
293 tapes (files splited accros two tapes give "size changed" errors
294 when comparing).
295
296 4. Corrected the long standing bug when dumping multiple tapes.
297 This works for me, needs further testing.
298
299 Changes between versions 0.4b3 and 0.4b4 (released January 17, 1997)
300 ====================================================================
301
302 1. Dump now runs correctly on kernels 2.1.x
303 Fix made by Gerald Peters <gapeters@worldnet.att.net>
304
305 Changes between versions 0.4b2 and 0.4b3
306 ========================================
307
308 1. Use realpath() if available
309
310 2. Report statistics
311
312 Changes between versions 0.4b1 and 0.4b2
313 ========================================
314
315 1. Fixed the bug fix from Greg Lutz (I had made a mistake when integrating
316 the patch)
317
318 2. Fixed restore to make it able to read FreeBSD 2.x dumps again
319
320 3. Fixed configure.in to correctly handle --enable-rmt
321
322 Changes between versions 0.3 and 0.4b1
323 ======================================
324
325 1. Integrated the changes from 4.4BSD-Lite2
326
327 2. Integrated the patches from Debian and Red Hat
328
329 3. Portability changes: use the __u32, __u16, __s32, and __s16 types
330
331 4. Changed dump to use the Ext2fs library to get block addresses. This
332 should solve the endianness problem on SparcLinux.
333
334 5. Created a configure.in file (shamelessly stolen from the e2fsprogs
335 distribution's one) to use autoconf
336
337 6. Fixed a few minor bugs
338
339 Changes between versions 0.2e and 0.2f
340 ======================================
341
342 1. Added the creation of named pipes (fifos) in restore.
343
344 2. Added the -N flag in the restore manual page.
345
346 3. Added the file kernel.patch which contains the llseek() optimization
347 patch for 1.2.x kernels.
348
349 4. Fixed a bug in the restoration of symbolic links: owner and group were
350 not restored.
351
352 5. Integrated some changes from FreeBSD 2.2.
353
354 6. Added a call to ftruncate() after restoring each file to restore
355 correctly files ending by a hole.
356
357 Changes between versions 0.2d and 0.2e
358 ======================================
359
360 1. Fixed a bug in the "set owner/mode" process. Every file was restored
361 with owner = root (0) and group = root/wheel/whatever (0).
362
363 Changes between versions 0.2c and 0.2d
364 ======================================
365
366 1. Dump is now able to backup 2GB+ filesystems.
367
368 2. Dump and restore can now be linked as static binaries.
369
370 Changes between versions 0.2b and 0.2c
371 ======================================
372
373 1. Fixed a bug when dumping ``slow'' (i.e. normal) symbolic links.
374
375 Changes between versions 0.2a and 0.2b
376 ======================================
377
378 1. Really fixed the bug that I should have corrected in 0.2a.
379
380 2. Enabled optimization again.
381
382 Changes between versions 0.2 and 0.2a
383 =====================================
384
385 1. Disabled the optimization during compilation.
386
387 Changes between versions 0.1 and 0.2
388 ====================================
389
390 1. Fixed a bug in fstab.c which caused a null pointer to be stored in
391 the fs_type field (actually, I modified the file fstab.c to make it
392 use the mntent functions).
393
394 2. Dump and restore now use a 4.3 BSD compatible dump format. Backups
395 made by dump should be readable by the BSD restore and backups made
396 by the BSD dump should be readable by restore. Unfortunately, this
397 means that the dump format has changed between version 0.1 and version
398 0.2 :-(
399
400 3. Dump is now able to backup a subtree, it is no longer limited to whole
401 filesystems like the BSD version.
402
403 4. Dump now uses ext2_llseek() so it is able to backup filesystems bigger
404 than 2 GB.
405
406 Changes between versions 0.0 and 0.1
407 ====================================
408
409 1. Now create links rdump and rrestore during the `make install' step.
410
411 2. Linux port specific bugs added to the manual pages
412
413 3. Incorrect estimation of the number of tapes blocks fixed when doing
414 incremental backups.
415
416 4. Better ls-like format in restore in interactive mode.