/[path-sandbox]/trunk/ChangeLog.0
Gentoo

Contents of /trunk/ChangeLog.0

Parent Directory Parent Directory | Revision Log Revision Log


Revision 37 - (show annotations) (download)
Mon Mar 14 00:28:10 2005 UTC (9 years, 7 months ago) by azarah
Original Path: trunk/ChangeLog
File size: 16307 byte(s)
Seems -nostdlib was the problem with the constructor/destructor - remove it
from Makefile.am, and change the constructor/destructor names again.

1 # ChangeLog for Path Sandbox
2 # Copyright 1999-2004 Gentoo Foundation; Distributed under the GPL v2
3 # $Header$
4
5 14 Mar 2005; Martin Schlemmer <azarah@gentoo.org> Makefile.am, libsandbox.c:
6 Seems -nostdlib was the problem with the constructor/destructor - remove it
7 from Makefile.am, and change the constructor/destructor names again.
8
9 14 Mar 2005; Martin Schlemmer <azarah@gentoo.org> libsandbox.c:
10 Also rename the _init() and _fini() declarations.
11
12 14 Mar 2005; Martin Schlemmer <azarah@gentoo.org> libsandbox.c, sandbox.c:
13 Fixup the constructor/destructor function names again (they should be _init()
14 and _fini() it seems, and not being called caused sandbox_lib_path to be
15 unset, and thus breaking the execve() wrapper's LD_PRELOAD protection).
16 Add both the path in given SANDBOX_x variable, as well as its symlink
17 resolved path in init_env_entries(). Modify filter_path() to be able to
18 resolve paths without resolving symlinks, as well as to be able to resolve
19 symlinks. Fix a possible segfault in check_access(). Add symlink resolving
20 to check_access() resolving bug #31019. Add 'hack' for unlink, as the fix
21 for bug #31019 cause access violations if we try to remove a symlink that is
22 not in protected path, but points to a protected path. Fix a memory leak in
23 sandbox.c (sandbox_pids_file in main()). Fix the realpath() calls in main()
24 (sandbox.c) being unchecked. Fix the debug logname not having the pid in it
25 (pid_string was uninitialized). General syntax cleanups.
26
27 09 Mar 2005; Brian Harring <ferringb@gentoo.org> sandbox.c: Fixed the
28 infamous "pids file is not a regular file" w/out newline bug.
29
30 09 Mar 2005; Brian Harring <ferringb@gentoo.org> Makefile.am, configure.in:
31 Correct libc_version path detection, since it was screwing up if libdir !=
32 "/lib/".
33
34 02 Mar 2005; Martin Schlemmer <azarah@gentoo.org> libsandbox.c:
35 Hack to make sure sandboxed process cannot remove a device node, bug #79836.
36
37 02 Mar 2005; Martin Schlemmer <azarah@gentoo.org> Makefile.am:
38 Fix symbols.in not added to dist.
39
40 02 Mar 2005; Martin Schlemmer <azarah@gentoo.org> Makefile.am, canonicalize.c,
41 getcwd.c, libsandbox.c, sandbox.c, sandbox.h, sandbox_futils.c:
42 White space fixes.
43
44 02 Mar 2005; Martin Schlemmer <azarah@gentoo.org> Makefile.am, canonicalize.c,
45 configure.in, getcwd.c, libsandbox.c, symbols.in:
46 Fix inverse test logic in canonicalize.c, use a strncpy. Fix gcc warning in
47 getcwd.c. Add symbols.in and logic to Makefile.am to generate symbol versions
48 for glibc and other libc's that use this. Update libsandbox.c to use these
49 symbol versions if available. Fix exec wrapper to re-export LD_PRELOAD if the
50 process unset it.
51
52 01 Mar 2005; Brian Harring <ferringb@gentoo.org> libsandbox.c:
53 killed off _init and _fini in favor of
54 void __attribute__ ((constructor)) init_func and
55 void __attribute__ ((destructor)) closing_func. _(init|func) were deprecated.
56
57 06 Dec 2004; Brian Harring <ferringb@gentoo.org> Makefile.am, libsandbox.c,
58 canonicalize.c, getcwd.c: Fixed compilation *again*. Hopefully cvs is done
59 having the hick-ups.
60
61 04 Dec 2004; Brian Harring <ferringb@gentoo.org> libsandbox.c, getcwd.c,
62 Makefile.am: Fixed compilation.
63
64 01 Dec 2004; Brian Harring <ferringb@gentoo.org> aclocal.m4:
65 Gutted the bugger so it stops checking for a c++ and fortran compiler.
66 Do *not* regenerate aclocal.m4 for making a release until a better
67 solution is created.
68
69 20 Nov 2004; Brian Harring <ferringb@gentoo.org> Makefile.am, sandbox_futils.c:
70 Removal of more hardcoded paths.
71
72 20 Nov 2004; Brian Harring <ferringb@gentoo.org> Makefile.am, configure.in,
73 sandbox_futils.c: tweaks to install sandbox.bashrc, and use it.
74
75 19 Nov 2004; Brian Harring <ferringb@gentoo.org>:
76 Sandbox is now autotooled, create-localdecls needs to be killed and the code
77 shifted into configure.in. Currently builds *one* libsandbox.so- if multiple
78 are desired (-m64 and -m32 for amd64), the ebuild should do it (imo).
79 To get to a point of testing, automake && autoconf; created requisite files w/
80 a(utomake|clocal)-1.8, and autoconf 2.59. Installs to /usr/, instead of
81 /lib and /usr/lib/portage/bin.
82
83 14 Nov 2004; Brian Harring <ferringb@gentoo.org> libsandbox.c, sandbox.c:
84 closing out bug #70225, potential overflow of the sandbox_pids_file var.
85
86 07 Nov 2004; Brian Harring <ferringb@gentoo.org> libsandbox.c: c99 standard,
87 (think it was at least) allows intermixing of code and data segments. bug #70351
88 should be fixed by this.
89
90 03 Nov 2004; Brian Harring <ferringb@gentoo.org> libsandbox.c, sandbox_futils.c:
91 futils fix from bug #65201 via solar, and libsandbox log path checks via #69137
92
93 02 Aug 2004; Nicholas Jones <carpaski@gentoo.org> libsandbox.c: Code from
94 Seth Robertson that tracked down all adjuct flags for read operations that
95 do not invoke a write operation.
96
97 04 Apr 2004; Nicholas Jones <carpaski@gentoo.org> libsandbox.c, sandbox.c:
98 Another fix from jstubbs regarding a free() on a stack variable for the
99 environment -- tracking now prevents extraneous free()'s segfault.
100
101 04 Apr 2004; Nicholas Jones <carpaski@gentoo.org> libsandbox.c, sandbox.c:
102 J. Stubbs tracked down a new bug where mkdir was failing to the patch on
103 the lstat in mkdir... it now only returns 0 or -1 as documented for mkdir.
104 Also remove the errno = ESUCCESS settings as documentation points out that
105 a library isn't allowed to do that.
106
107 04 Apr 2004; Nicholas Jones <carpaski@gentoo.org> libsandbox.c: Added a
108 file_security_check() function to check random potential exploits on files
109 that sandbox is to load and read -- Normally sandboxpids.tmp. This fixes
110 the 'system-crippling' exploits (bug 21923) and catches a few other
111 potential problems.
112
113 20 Mar 2004; Nicholas Jones <carpaski@gentoo.org> Makefile: Updates for
114 32/64 bit sandbox. Made CC and LD '?=' values to allow passed in CC to work.
115
116 20 Mar 2004; Nicholas Jones <carpaski@gentoo.org> libsandbox.c:
117 bug 42048 -- Fixed the lstat/errno conditions for mkdir <caleb@g.o>.
118 Added the 64/32 bit sandbox patch for AMD64 bug 32963 <brad/azarah>.
119
120 29 Feb 2004; Martin Schlemmer <azarah@gentoo.org> sandbox.c, sandbox_futils.c :
121 Fix permissions and group of pids file and logs. Permissions should be 0664
122 and group should be 'portage'. Bug #34260.
123
124 28 Feb 2004; Martin Schlemmer <azarah@gentoo.org> libsandbox.c :
125 Besides a small cleanup, redo how we replace LD_PRELOAD in the environ passed
126 to the real execve (in our execve wrapper). Seems that on some arches (sparc
127 among others) do not allow us to tamper with the readonly copy passed to
128 execve, so pass our own copy of the environment. Bug #42290.
129
130 11 Jan 2004; Nicholas Jones <carpaski@gentoo.org> create-decls:
131 Changed tail to head and added a notice about duration of glibc check.
132
133 21 Dec 2003; Nicholas Jones <carpaski@gentoo.org> create-decls:
134 Changed the glibc subversion check to use /usr/bin/* instead of /bin/sh
135 as there isn't a guarentee that it is dynamic.
136
137 02 Nov 2003; Martin Schlemmer <azarah@gentoo.org> libsandbox.c :
138 If 'file' passed to before_syscall(const char *func, const char *file) is
139 invalid, we should set errno to ENOENT, and not EINVAL. This should
140 close bug #32238.
141
142 14 Oct 2003; Martin Schlemmer <azarah@gentoo.org> libsandbox.c :
143 Fix a bug that occurs mainly on 64bit arch, where the file passed to
144 the functions we wrap, is invalid, and then cause canonicalize to pass
145 garbage to before_syscall(), thanks to great detective work from
146 Andrea Luzzardi <al@sig11.org> (bug #29846).
147
148 13 Oct 2003; Martin Schlemmer <azarah@gentoo.org> create-localdecls :
149 Add a uClibc detection patch from Peter S. Mazinger <ps.m@gmx.net>.
150
151 13 Oct 2003; Martin Schlemmer <azarah@gentoo.org> libsandbox.c :
152 Fix a bug in libsandbox.c 's checking in the rename wrapper - it basically
153 only checked the destination patch, and not the source, so we could move
154 a protected file to a unprotected directory, and then delete/modify it.
155 Thanks to Andrea Luzzardi (scox) <al@sig11.org>, bug #30992, for this fix.
156
157 12 Oct 2003; Nicholas Jones <carpaski@gentoo.org> sandbox.c :
158 Added python2.3 to the predict section/variable.
159
160 28 Sep 2003; Martin Schlemmer <azarah@gentoo.org> libsandbox.c, sandbox.c,
161 sandbox.h, sandbox_futils.c :
162 Add support to set the pids file via SANDBOX_PIDS_FILE at startup. If
163 it is not set, it will revert to its old value.
164
165 27 Sep 2003; Martin Schlemmer <azarah@gentoo.org> libsandbox.c :
166 Fix our mkdir wrapper to check if the dir exist, and return EEXIST if so,
167 rather than failing with a violation, bug #29748.
168
169 27 Jul 2003; Martin Schlemmer <azarah@gentoo.org> libsandbox.c :
170 Fix canonicalize() to ignore calls with path = "".
171
172 27 Jul 2003; Martin Schlemmer <azarah@gentoo.org> getcwd.c, libsandbox.c,
173 sandbox_futils.c, canonicalize.c :
174 Once again coreutils fails, as my systems had 2.5 kernel, the getcwd system
175 call handled strings larger than PATH_MAX (bug #21766). It however does not
176 work the same on 2.4 kernels.
177
178 To fix, I added the posix implementation of getcwd() (from glibc cvs) that
179 do not need the system call. We use the default getcwd() function via a
180 wrapper (egetcwd), and then lstat the returned path. If lstat fails, it
181 means the current directory was removed, OR that the the system call for
182 getcwd failed (curious is that it do not fail and return NULL or set
183 errno, but rather just truncate the retured directory - usually from the
184 start), and if so, we use the generic getcwd() function (__egetcwd). Note
185 that we do not use the generic version all the time, as it calls lstat()
186 a great number of times, and performance degrade much.
187
188 29 Jun 2003; Martin Schlemmer <azarah@gentoo.org> create-localdecls,
189 libsandbox.c :
190 Make sure SB_PATH_MAX will not wrap. Fix two possible memory leaks.
191
192 22 Jun 2003; Martin Schlemmer <azarah@gentoo.org> libsandbox.c, canonicalize.c
193 create-localdecls :
194 When checking path names of files accessed, we need to canonicalize it, else
195 it may be a symlink in a 'write allowed' directory pointing to a file in a
196 directory we should not have write access to.
197
198 With something like coreutils-5.0, we have two problems:
199 1) One of the tests checks if getcwd() can return a path longer than
200 PATH_MAX. This test then tries to create a dir which even while
201 created local (mkdir("conftest2")), it ends up being resolved with
202 a name that is much larger than PATH_MAX. The problem now is that
203 canonicalize() have undefined behaviour when the path was too long
204 (returned wrongly truncated paths, etc), and pass the wrong path to
205 before_syscall() (causing the bogus sandbox violations).
206 2) The ecanonicalize() function we used, along with the canonicalize()
207 function did not support longer than PATH_MAX. This is partly a
208 cause for 1), but the error checking (rather lack of it) of calls
209 to erealpath() in canonicalize() was the prime reason for 1).
210
211 As we do not use this canonicalized name to call the function, we resolve this
212 by fixing canonicalize() to do better error checking, and ecanonicalize() as
213 well as all functions in libsandbox.c to use a PATH_MAX of 'PATH_MAX * 2'.
214 While they will resolve paths properly now, and can check if a write/read is
215 allowed, the functions called from the sandboxed environment will still work
216 as expected.
217
218 This should resolve bug #21766.
219
220 06 Apr 2003; Martin Schlemmer <azarah@gentoo.org> libsandbox.c :
221 For some reason sandbox fails with a 'open_wr' if you run 'locale -a' under
222 it (bug #16298).
223
224 Problem is that for some reason locale fopen's locale.alias with mode "rm".
225
226 -------------------------------------------------------
227 nosferatu root # grep fopen locale.log
228 fopen("/usr/share/locale/locale.alias", "rm"ACCESS DENIED open_wr: /usr/share/locale/locale.alias
229 nosferatu root #
230 --------------------------------------------------------
231
232 I checked the source of locale, but it have fopen with mode 'r', so
233 not sure where the "rm" mode comes from. Anyhow, changed the check in
234 before_syscall_open_char() to also see mode "rm" as readonly.
235
236 23 Feb 2003; Martin Schlemmer <azarah@gentoo.org> create-localdecls :
237
238 Add glibc-2.3 support.
239
240 22 Feb 2003; Martin Schlemmer <azarah@gentoo.org> sandbox.c :
241
242 Some /etc/ld.so.preload fixes. Just changed the #if defines to cover all
243 operations releated to preload, as well as only try to modify ld.so.preload
244 if we can. Also modify to write the pid to /tmp/sandboxpids.tmp even when
245 not using ld.so.preload. Fix to not write this instance of sandbox's pid
246 to /tmp/sandboxpids.tmp on exit if this is not the last sandbox running.
247
248 22 Feb 2003; Nicholas Jones <carpaski@gentoo.org> Makefile :
249
250 Changed the LD to CC for hppa.
251
252 22 Feb 2003; Nicholas Jones <carpaski@gentoo.org> create-localdecls :
253
254 Killed the previous changes I made.
255
256 17 Feb 2003; Nicholas Jones <carpaski@gentoo.org> create-localdecls :
257
258 Added parisc to BROKEN_RTLD_ARCHLIST to see if it we can fix the relocation probs.
259
260 09 Jan 2003; J Robert Ray <jrray@gentoo.org> sandbox.c :
261
262 Don't segfault if $HOME isn't set, set $HOME to "/" instead. Fixes bug 10868.
263
264 16 Dec 2002; Martin Schlemmer <azarah@gentoo.org> create-localdecls :
265
266 Fix memory leak for mips, bug #12236. Thanks to Torgeir Hansen <torgeir@trenger.ro>
267 for this fix.
268
269 4 Dec 2002; J Robert Ray <jrray@gentoo.org> sandbox.h sandbox_futils.c :
270
271 sandbox_futils defined a dirname() function that was masking the same
272 function in glibc and was broken (e.g.: SANDBOX_DIR was being set to
273 '/usr/lib/portage/bi/'). Fixed function to return expected results and
274 renamed it to sb_dirname() to no longer mask the glibc function. Closes bug
275 11231.
276
277 4 Dec 2002; Martin Schlemmer <azarah@gentoo.org> :
278
279 Fix a segfault in libsandbox.c if canonicalize() was called with
280 first parameter = NULL.
281
282 1 Sep 2002; Martin Schlemmer <azarah@gentoo.org> :
283
284 Fix my braindead 'return 1;' in a void function. Updated sandbox.c,
285 cleanup() for this.
286
287 Change cleanup() in sandbox.c not to exit with fail status if
288 the pidsfile is missing. We really should still display sandbox
289 violations if they occured.
290
291 31 Aug 2002; Martin Schlemmer <azarah@gentoo.org> :
292
293 Update cleanup() in sandbox.c to remove the PIDSFILE if this is
294 the last sandbox running.
295
296 25 Aug 2002; Martin Schlemmer <azarah@gentoo.org> :
297
298 Major cleanups to mainly libsandbox.c again.
299
300 22 Aug 2002; Martin Schlemmer <azarah@gentoo.org> :
301
302 Add copyrights to sandbox.h and sandbox_futils.h. If wrong, the
303 parties involved should please contact me so that we can fix it.
304
305 Add opendir wrapper to libsandbox.c.
306
307 21 Aug 2002; Martin Schlemmer <azarah@gentoo.org> :
308
309 Do some more cleanups to ecanonicalize(), as it dropped filenames in
310 rare cases (after my symlink cleanups), and caused glibc to bork.
311 These fixes went into canonicalize.c.
312
313 20 Aug 2002; Martin Schlemmer <azarah@gentoo.org> :
314
315 Fix spawn_shell() and main() in sandbox.c to properly return fail
316 status.
317
318 19 Aug 2002; Martin Schlemmer <azarah@gentoo.org> :
319
320 The new canonicalize() function in libsandbox.c also resolved symlinks,
321 which caused on cleaning sandbox errors if the symlink pointed to a
322 file in the live root. Ripped out canonicalize() and realpath() from
323 glibc; removed the symlink stuff, and changed them to ecanonicalize()
324 and erealpath().
325
326 18 Aug 2002; Martin Schlemmer <azarah@gentoo.org> :
327
328 Ripped out all the wrappers, and implemented those of InstallWatch.
329 Losts of cleanups and bugfixes. Implement a execve that forces
330 $LIBSANDBOX in $LD_PRELOAD. We can now thus do away with the feared
331 /etc/ld.so.preload (*g*) ... Made the needed changes to sandbox.c,
332 sandbox.h and sandbox_futils.c. Rewrote the Makefile for most
333 parts; it now have an install target.
334
335 Reformat the whole thing to look somewhat like the reworked sandbox.c
336 and new sandbox.h and sandbox_futils.c from:
337
338 Brad House <brad@mainstreetsoftworks.com>.
339
340 Additional Copyrights now due to the InstallWatch code:
341
342 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