/[path-sandbox]/trunk/ChangeLog
Gentoo

Contents of /trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 29 - (show annotations) (download)
Wed Mar 2 09:01:36 2005 UTC (9 years, 6 months ago) by azarah
File size: 14170 byte(s)
Fix inverse test logic in canonicalize.c, use a strncpy.  Fix gcc warning in
getcwd.c.  Add symbols.in and logic to Makefile.am to generate symbol versions
for glibc and other libc's that use this.  Update libsandbox.c to use these
symbol versions if available.  Fix exec wrapper to re-export LD_PRELOAD if the
process unset it.

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