/[path-sandbox]/trunk/ChangeLog
Gentoo

Contents of /trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 31 - (show annotations) (download)
Wed Mar 2 09:18:01 2005 UTC (9 years, 6 months ago) by azarah
File size: 14444 byte(s)
Fix symbols.in not added to dist.

1 # ChangeLog for Path Sandbox
2 # Copyright 1999-2004 Gentoo Foundation; Distributed under the GPL v2
3 # $Header$
4
5 02 May 2005; Martin Schlemmer <azarah@gentoo.org> Makefile.am:
6 Fix symbols.in not added to dist.
7
8 02 May 2005; Martin Schlemmer <azarah@gentoo.org> Makefile.am, canonicalize.c,
9 getcwd.c, libsandbox.c, sandbox.c, sandbox.h, sandbox_futils.c:
10 White space fixes.
11
12 02 May 2005; Martin Schlemmer <azarah@gentoo.org> Makefile.am, canonicalize.c,
13 configure.in, getcwd.c, libsandbox.c, symbols.in:
14 Fix inverse test logic in canonicalize.c, use a strncpy. Fix gcc warning in
15 getcwd.c. Add symbols.in and logic to Makefile.am to generate symbol versions
16 for glibc and other libc's that use this. Update libsandbox.c to use these
17 symbol versions if available. Fix exec wrapper to re-export LD_PRELOAD if the
18 process unset it.
19
20 01 May 2005; Brian Harring <ferringb@gentoo.org> libsandbox.c:
21 killed off _init and _fini in favor of
22 void __attribute__ ((constructor)) init_func and
23 void __attribute__ ((destructor)) closing_func. _(init|func) were deprecated.
24
25 06 Dec 2004; Brian Harring <ferringb@gentoo.org> Makefile.am, libsandbox.c,
26 canonicalize.c, getcwd.c: Fixed compilation *again*. Hopefully cvs is done
27 having the hick-ups.
28
29 04 Dec 2004; Brian Harring <ferringb@gentoo.org> libsandbox.c, getcwd.c,
30 Makefile.am: Fixed compilation.
31
32 01 Dec 2004; Brian Harring <ferringb@gentoo.org> aclocal.m4:
33 Gutted the bugger so it stops checking for a c++ and fortran compiler.
34 Do *not* regenerate aclocal.m4 for making a release until a better
35 solution is created.
36
37 20 Nov 2004; Brian Harring <ferringb@gentoo.org> Makefile.am, sandbox_futils.c:
38 Removal of more hardcoded paths.
39
40 20 Nov 2004; Brian Harring <ferringb@gentoo.org> Makefile.am, configure.in,
41 sandbox_futils.c: tweaks to install sandbox.bashrc, and use it.
42
43 19 Nov 2004; Brian Harring <ferringb@gentoo.org>:
44 Sandbox is now autotooled, create-localdecls needs to be killed and the code
45 shifted into configure.in. Currently builds *one* libsandbox.so- if multiple
46 are desired (-m64 and -m32 for amd64), the ebuild should do it (imo).
47 To get to a point of testing, automake && autoconf; created requisite files w/
48 a(utomake|clocal)-1.8, and autoconf 2.59. Installs to /usr/, instead of
49 /lib and /usr/lib/portage/bin.
50
51 14 Nov 2004; Brian Harring <ferringb@gentoo.org> libsandbox.c, sandbox.c:
52 closing out bug #70225, potential overflow of the sandbox_pids_file var.
53
54 07 Nov 2004; Brian Harring <ferringb@gentoo.org> libsandbox.c: c99 standard,
55 (think it was at least) allows intermixing of code and data segments. bug #70351
56 should be fixed by this.
57
58 03 Nov 2004; Brian Harring <ferringb@gentoo.org> libsandbox.c, sandbox_futils.c:
59 futils fix from bug #65201 via solar, and libsandbox log path checks via #69137
60
61 02 Aug 2004; Nicholas Jones <carpaski@gentoo.org> libsandbox.c: Code from
62 Seth Robertson that tracked down all adjuct flags for read operations that
63 do not invoke a write operation.
64
65 04 Apr 2004; Nicholas Jones <carpaski@gentoo.org> libsandbox.c, sandbox.c:
66 Another fix from jstubbs regarding a free() on a stack variable for the
67 environment -- tracking now prevents extraneous free()'s segfault.
68
69 04 Apr 2004; Nicholas Jones <carpaski@gentoo.org> libsandbox.c, sandbox.c:
70 J. Stubbs tracked down a new bug where mkdir was failing to the patch on
71 the lstat in mkdir... it now only returns 0 or -1 as documented for mkdir.
72 Also remove the errno = ESUCCESS settings as documentation points out that
73 a library isn't allowed to do that.
74
75 04 Apr 2004; Nicholas Jones <carpaski@gentoo.org> libsandbox.c: Added a
76 file_security_check() function to check random potential exploits on files
77 that sandbox is to load and read -- Normally sandboxpids.tmp. This fixes
78 the 'system-crippling' exploits (bug 21923) and catches a few other
79 potential problems.
80
81 20 Mar 2004; Nicholas Jones <carpaski@gentoo.org> Makefile: Updates for
82 32/64 bit sandbox. Made CC and LD '?=' values to allow passed in CC to work.
83
84 20 Mar 2004; Nicholas Jones <carpaski@gentoo.org> libsandbox.c:
85 bug 42048 -- Fixed the lstat/errno conditions for mkdir <caleb@g.o>.
86 Added the 64/32 bit sandbox patch for AMD64 bug 32963 <brad/azarah>.
87
88 29 Feb 2004; Martin Schlemmer <azarah@gentoo.org> sandbox.c, sandbox_futils.c :
89 Fix permissions and group of pids file and logs. Permissions should be 0664
90 and group should be 'portage'. Bug #34260.
91
92 28 Feb 2004; Martin Schlemmer <azarah@gentoo.org> libsandbox.c :
93 Besides a small cleanup, redo how we replace LD_PRELOAD in the environ passed
94 to the real execve (in our execve wrapper). Seems that on some arches (sparc
95 among others) do not allow us to tamper with the readonly copy passed to
96 execve, so pass our own copy of the environment. Bug #42290.
97
98 11 Jan 2004; Nicholas Jones <carpaski@gentoo.org> create-decls:
99 Changed tail to head and added a notice about duration of glibc check.
100
101 21 Dec 2003; Nicholas Jones <carpaski@gentoo.org> create-decls:
102 Changed the glibc subversion check to use /usr/bin/* instead of /bin/sh
103 as there isn't a guarentee that it is dynamic.
104
105 02 Nov 2003; Martin Schlemmer <azarah@gentoo.org> libsandbox.c :
106 If 'file' passed to before_syscall(const char *func, const char *file) is
107 invalid, we should set errno to ENOENT, and not EINVAL. This should
108 close bug #32238.
109
110 14 Oct 2003; Martin Schlemmer <azarah@gentoo.org> libsandbox.c :
111 Fix a bug that occurs mainly on 64bit arch, where the file passed to
112 the functions we wrap, is invalid, and then cause canonicalize to pass
113 garbage to before_syscall(), thanks to great detective work from
114 Andrea Luzzardi <al@sig11.org> (bug #29846).
115
116 13 Oct 2003; Martin Schlemmer <azarah@gentoo.org> create-localdecls :
117 Add a uClibc detection patch from Peter S. Mazinger <ps.m@gmx.net>.
118
119 13 Oct 2003; Martin Schlemmer <azarah@gentoo.org> libsandbox.c :
120 Fix a bug in libsandbox.c 's checking in the rename wrapper - it basically
121 only checked the destination patch, and not the source, so we could move
122 a protected file to a unprotected directory, and then delete/modify it.
123 Thanks to Andrea Luzzardi (scox) <al@sig11.org>, bug #30992, for this fix.
124
125 12 Oct 2003; Nicholas Jones <carpaski@gentoo.org> sandbox.c :
126 Added python2.3 to the predict section/variable.
127
128 28 Sep 2003; Martin Schlemmer <azarah@gentoo.org> libsandbox.c, sandbox.c,
129 sandbox.h, sandbox_futils.c :
130 Add support to set the pids file via SANDBOX_PIDS_FILE at startup. If
131 it is not set, it will revert to its old value.
132
133 27 Sep 2003; Martin Schlemmer <azarah@gentoo.org> libsandbox.c :
134 Fix our mkdir wrapper to check if the dir exist, and return EEXIST if so,
135 rather than failing with a violation, bug #29748.
136
137 27 Jul 2003; Martin Schlemmer <azarah@gentoo.org> libsandbox.c :
138 Fix canonicalize() to ignore calls with path = "".
139
140 27 Jul 2003; Martin Schlemmer <azarah@gentoo.org> getcwd.c, libsandbox.c,
141 sandbox_futils.c, canonicalize.c :
142 Once again coreutils fails, as my systems had 2.5 kernel, the getcwd system
143 call handled strings larger than PATH_MAX (bug #21766). It however does not
144 work the same on 2.4 kernels.
145
146 To fix, I added the posix implementation of getcwd() (from glibc cvs) that
147 do not need the system call. We use the default getcwd() function via a
148 wrapper (egetcwd), and then lstat the returned path. If lstat fails, it
149 means the current directory was removed, OR that the the system call for
150 getcwd failed (curious is that it do not fail and return NULL or set
151 errno, but rather just truncate the retured directory - usually from the
152 start), and if so, we use the generic getcwd() function (__egetcwd). Note
153 that we do not use the generic version all the time, as it calls lstat()
154 a great number of times, and performance degrade much.
155
156 29 Jun 2003; Martin Schlemmer <azarah@gentoo.org> create-localdecls,
157 libsandbox.c :
158 Make sure SB_PATH_MAX will not wrap. Fix two possible memory leaks.
159
160 22 Jun 2003; Martin Schlemmer <azarah@gentoo.org> libsandbox.c, canonicalize.c
161 create-localdecls :
162 When checking path names of files accessed, we need to canonicalize it, else
163 it may be a symlink in a 'write allowed' directory pointing to a file in a
164 directory we should not have write access to.
165
166 With something like coreutils-5.0, we have two problems:
167 1) One of the tests checks if getcwd() can return a path longer than
168 PATH_MAX. This test then tries to create a dir which even while
169 created local (mkdir("conftest2")), it ends up being resolved with
170 a name that is much larger than PATH_MAX. The problem now is that
171 canonicalize() have undefined behaviour when the path was too long
172 (returned wrongly truncated paths, etc), and pass the wrong path to
173 before_syscall() (causing the bogus sandbox violations).
174 2) The ecanonicalize() function we used, along with the canonicalize()
175 function did not support longer than PATH_MAX. This is partly a
176 cause for 1), but the error checking (rather lack of it) of calls
177 to erealpath() in canonicalize() was the prime reason for 1).
178
179 As we do not use this canonicalized name to call the function, we resolve this
180 by fixing canonicalize() to do better error checking, and ecanonicalize() as
181 well as all functions in libsandbox.c to use a PATH_MAX of 'PATH_MAX * 2'.
182 While they will resolve paths properly now, and can check if a write/read is
183 allowed, the functions called from the sandboxed environment will still work
184 as expected.
185
186 This should resolve bug #21766.
187
188 06 Apr 2003; Martin Schlemmer <azarah@gentoo.org> libsandbox.c :
189 For some reason sandbox fails with a 'open_wr' if you run 'locale -a' under
190 it (bug #16298).
191
192 Problem is that for some reason locale fopen's locale.alias with mode "rm".
193
194 -------------------------------------------------------
195 nosferatu root # grep fopen locale.log
196 fopen("/usr/share/locale/locale.alias", "rm"ACCESS DENIED open_wr: /usr/share/locale/locale.alias
197 nosferatu root #
198 --------------------------------------------------------
199
200 I checked the source of locale, but it have fopen with mode 'r', so
201 not sure where the "rm" mode comes from. Anyhow, changed the check in
202 before_syscall_open_char() to also see mode "rm" as readonly.
203
204 23 Feb 2003; Martin Schlemmer <azarah@gentoo.org> create-localdecls :
205
206 Add glibc-2.3 support.
207
208 22 Feb 2003; Martin Schlemmer <azarah@gentoo.org> sandbox.c :
209
210 Some /etc/ld.so.preload fixes. Just changed the #if defines to cover all
211 operations releated to preload, as well as only try to modify ld.so.preload
212 if we can. Also modify to write the pid to /tmp/sandboxpids.tmp even when
213 not using ld.so.preload. Fix to not write this instance of sandbox's pid
214 to /tmp/sandboxpids.tmp on exit if this is not the last sandbox running.
215
216 22 Feb 2003; Nicholas Jones <carpaski@gentoo.org> Makefile :
217
218 Changed the LD to CC for hppa.
219
220 22 Feb 2003; Nicholas Jones <carpaski@gentoo.org> create-localdecls :
221
222 Killed the previous changes I made.
223
224 17 Feb 2003; Nicholas Jones <carpaski@gentoo.org> create-localdecls :
225
226 Added parisc to BROKEN_RTLD_ARCHLIST to see if it we can fix the relocation probs.
227
228 09 Jan 2003; J Robert Ray <jrray@gentoo.org> sandbox.c :
229
230 Don't segfault if $HOME isn't set, set $HOME to "/" instead. Fixes bug 10868.
231
232 16 Dec 2002; Martin Schlemmer <azarah@gentoo.org> create-localdecls :
233
234 Fix memory leak for mips, bug #12236. Thanks to Torgeir Hansen <torgeir@trenger.ro>
235 for this fix.
236
237 4 Dec 2002; J Robert Ray <jrray@gentoo.org> sandbox.h sandbox_futils.c :
238
239 sandbox_futils defined a dirname() function that was masking the same
240 function in glibc and was broken (e.g.: SANDBOX_DIR was being set to
241 '/usr/lib/portage/bi/'). Fixed function to return expected results and
242 renamed it to sb_dirname() to no longer mask the glibc function. Closes bug
243 11231.
244
245 4 Dec 2002; Martin Schlemmer <azarah@gentoo.org> :
246
247 Fix a segfault in libsandbox.c if canonicalize() was called with
248 first parameter = NULL.
249
250 1 Sep 2002; Martin Schlemmer <azarah@gentoo.org> :
251
252 Fix my braindead 'return 1;' in a void function. Updated sandbox.c,
253 cleanup() for this.
254
255 Change cleanup() in sandbox.c not to exit with fail status if
256 the pidsfile is missing. We really should still display sandbox
257 violations if they occured.
258
259 31 Aug 2002; Martin Schlemmer <azarah@gentoo.org> :
260
261 Update cleanup() in sandbox.c to remove the PIDSFILE if this is
262 the last sandbox running.
263
264 25 Aug 2002; Martin Schlemmer <azarah@gentoo.org> :
265
266 Major cleanups to mainly libsandbox.c again.
267
268 22 Aug 2002; Martin Schlemmer <azarah@gentoo.org> :
269
270 Add copyrights to sandbox.h and sandbox_futils.h. If wrong, the
271 parties involved should please contact me so that we can fix it.
272
273 Add opendir wrapper to libsandbox.c.
274
275 21 Aug 2002; Martin Schlemmer <azarah@gentoo.org> :
276
277 Do some more cleanups to ecanonicalize(), as it dropped filenames in
278 rare cases (after my symlink cleanups), and caused glibc to bork.
279 These fixes went into canonicalize.c.
280
281 20 Aug 2002; Martin Schlemmer <azarah@gentoo.org> :
282
283 Fix spawn_shell() and main() in sandbox.c to properly return fail
284 status.
285
286 19 Aug 2002; Martin Schlemmer <azarah@gentoo.org> :
287
288 The new canonicalize() function in libsandbox.c also resolved symlinks,
289 which caused on cleaning sandbox errors if the symlink pointed to a
290 file in the live root. Ripped out canonicalize() and realpath() from
291 glibc; removed the symlink stuff, and changed them to ecanonicalize()
292 and erealpath().
293
294 18 Aug 2002; Martin Schlemmer <azarah@gentoo.org> :
295
296 Ripped out all the wrappers, and implemented those of InstallWatch.
297 Losts of cleanups and bugfixes. Implement a execve that forces
298 $LIBSANDBOX in $LD_PRELOAD. We can now thus do away with the feared
299 /etc/ld.so.preload (*g*) ... Made the needed changes to sandbox.c,
300 sandbox.h and sandbox_futils.c. Rewrote the Makefile for most
301 parts; it now have an install target.
302
303 Reformat the whole thing to look somewhat like the reworked sandbox.c
304 and new sandbox.h and sandbox_futils.c from:
305
306 Brad House <brad@mainstreetsoftworks.com>.
307
308 Additional Copyrights now due to the InstallWatch code:
309
310 Copyright (C) 1998-9 Pancrazio `Ezio' de Mauro <p@demauro.net>

Properties

Name Value
svn:eol-style native
svn:keywords Author Date Id Revision

  ViewVC Help
Powered by ViewVC 1.1.20