/[path-sandbox]/trunk/ChangeLog
Gentoo

Contents of /trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 39 - (show annotations) (download)
Mon Mar 21 08:54:02 2005 UTC (9 years, 5 months ago) by genone
File size: 16417 byte(s)
show resolved symlinks in log

1 # ChangeLog for Path Sandbox
2 # Copyright 1999-2004 Gentoo Foundation; Distributed under the GPL v2
3 # $Header$
4
5 21 Mar 2005; Marius Mauch <genone@gentoo.org> libsandbox.c:
6 Also show resolved symlink names in the log.
7
8 14 Mar 2005; Martin Schlemmer <azarah@gentoo.org> Makefile.am, libsandbox.c:
9 Seems -nostdlib was the problem with the constructor/destructor - remove it
10 from Makefile.am, and change the constructor/destructor names again.
11
12 14 Mar 2005; Martin Schlemmer <azarah@gentoo.org> libsandbox.c:
13 Also rename the _init() and _fini() declarations.
14
15 14 Mar 2005; Martin Schlemmer <azarah@gentoo.org> libsandbox.c, sandbox.c:
16 Fixup the constructor/destructor function names again (they should be _init()
17 and _fini() it seems, and not being called caused sandbox_lib_path to be
18 unset, and thus breaking the execve() wrapper's LD_PRELOAD protection).
19 Add both the path in given SANDBOX_x variable, as well as its symlink
20 resolved path in init_env_entries(). Modify filter_path() to be able to
21 resolve paths without resolving symlinks, as well as to be able to resolve
22 symlinks. Fix a possible segfault in check_access(). Add symlink resolving
23 to check_access() resolving bug #31019. Add 'hack' for unlink, as the fix
24 for bug #31019 cause access violations if we try to remove a symlink that is
25 not in protected path, but points to a protected path. Fix a memory leak in
26 sandbox.c (sandbox_pids_file in main()). Fix the realpath() calls in main()
27 (sandbox.c) being unchecked. Fix the debug logname not having the pid in it
28 (pid_string was uninitialized). General syntax cleanups.
29
30 09 Mar 2005; Brian Harring <ferringb@gentoo.org> sandbox.c: Fixed the
31 infamous "pids file is not a regular file" w/out newline bug.
32
33 09 Mar 2005; Brian Harring <ferringb@gentoo.org> Makefile.am, configure.in:
34 Correct libc_version path detection, since it was screwing up if libdir !=
35 "/lib/".
36
37 02 Mar 2005; Martin Schlemmer <azarah@gentoo.org> libsandbox.c:
38 Hack to make sure sandboxed process cannot remove a device node, bug #79836.
39
40 02 Mar 2005; Martin Schlemmer <azarah@gentoo.org> Makefile.am:
41 Fix symbols.in not added to dist.
42
43 02 Mar 2005; Martin Schlemmer <azarah@gentoo.org> Makefile.am, canonicalize.c,
44 getcwd.c, libsandbox.c, sandbox.c, sandbox.h, sandbox_futils.c:
45 White space fixes.
46
47 02 Mar 2005; Martin Schlemmer <azarah@gentoo.org> Makefile.am, canonicalize.c,
48 configure.in, getcwd.c, libsandbox.c, symbols.in:
49 Fix inverse test logic in canonicalize.c, use a strncpy. Fix gcc warning in
50 getcwd.c. Add symbols.in and logic to Makefile.am to generate symbol versions
51 for glibc and other libc's that use this. Update libsandbox.c to use these
52 symbol versions if available. Fix exec wrapper to re-export LD_PRELOAD if the
53 process unset it.
54
55 01 Mar 2005; Brian Harring <ferringb@gentoo.org> libsandbox.c:
56 killed off _init and _fini in favor of
57 void __attribute__ ((constructor)) init_func and
58 void __attribute__ ((destructor)) closing_func. _(init|func) were deprecated.
59
60 06 Dec 2004; Brian Harring <ferringb@gentoo.org> Makefile.am, libsandbox.c,
61 canonicalize.c, getcwd.c: Fixed compilation *again*. Hopefully cvs is done
62 having the hick-ups.
63
64 04 Dec 2004; Brian Harring <ferringb@gentoo.org> libsandbox.c, getcwd.c,
65 Makefile.am: Fixed compilation.
66
67 01 Dec 2004; Brian Harring <ferringb@gentoo.org> aclocal.m4:
68 Gutted the bugger so it stops checking for a c++ and fortran compiler.
69 Do *not* regenerate aclocal.m4 for making a release until a better
70 solution is created.
71
72 20 Nov 2004; Brian Harring <ferringb@gentoo.org> Makefile.am, sandbox_futils.c:
73 Removal of more hardcoded paths.
74
75 20 Nov 2004; Brian Harring <ferringb@gentoo.org> Makefile.am, configure.in,
76 sandbox_futils.c: tweaks to install sandbox.bashrc, and use it.
77
78 19 Nov 2004; Brian Harring <ferringb@gentoo.org>:
79 Sandbox is now autotooled, create-localdecls needs to be killed and the code
80 shifted into configure.in. Currently builds *one* libsandbox.so- if multiple
81 are desired (-m64 and -m32 for amd64), the ebuild should do it (imo).
82 To get to a point of testing, automake && autoconf; created requisite files w/
83 a(utomake|clocal)-1.8, and autoconf 2.59. Installs to /usr/, instead of
84 /lib and /usr/lib/portage/bin.
85
86 14 Nov 2004; Brian Harring <ferringb@gentoo.org> libsandbox.c, sandbox.c:
87 closing out bug #70225, potential overflow of the sandbox_pids_file var.
88
89 07 Nov 2004; Brian Harring <ferringb@gentoo.org> libsandbox.c: c99 standard,
90 (think it was at least) allows intermixing of code and data segments. bug #70351
91 should be fixed by this.
92
93 03 Nov 2004; Brian Harring <ferringb@gentoo.org> libsandbox.c, sandbox_futils.c:
94 futils fix from bug #65201 via solar, and libsandbox log path checks via #69137
95
96 02 Aug 2004; Nicholas Jones <carpaski@gentoo.org> libsandbox.c: Code from
97 Seth Robertson that tracked down all adjuct flags for read operations that
98 do not invoke a write operation.
99
100 04 Apr 2004; Nicholas Jones <carpaski@gentoo.org> libsandbox.c, sandbox.c:
101 Another fix from jstubbs regarding a free() on a stack variable for the
102 environment -- tracking now prevents extraneous free()'s segfault.
103
104 04 Apr 2004; Nicholas Jones <carpaski@gentoo.org> libsandbox.c, sandbox.c:
105 J. Stubbs tracked down a new bug where mkdir was failing to the patch on
106 the lstat in mkdir... it now only returns 0 or -1 as documented for mkdir.
107 Also remove the errno = ESUCCESS settings as documentation points out that
108 a library isn't allowed to do that.
109
110 04 Apr 2004; Nicholas Jones <carpaski@gentoo.org> libsandbox.c: Added a
111 file_security_check() function to check random potential exploits on files
112 that sandbox is to load and read -- Normally sandboxpids.tmp. This fixes
113 the 'system-crippling' exploits (bug 21923) and catches a few other
114 potential problems.
115
116 20 Mar 2004; Nicholas Jones <carpaski@gentoo.org> Makefile: Updates for
117 32/64 bit sandbox. Made CC and LD '?=' values to allow passed in CC to work.
118
119 20 Mar 2004; Nicholas Jones <carpaski@gentoo.org> libsandbox.c:
120 bug 42048 -- Fixed the lstat/errno conditions for mkdir <caleb@g.o>.
121 Added the 64/32 bit sandbox patch for AMD64 bug 32963 <brad/azarah>.
122
123 29 Feb 2004; Martin Schlemmer <azarah@gentoo.org> sandbox.c, sandbox_futils.c :
124 Fix permissions and group of pids file and logs. Permissions should be 0664
125 and group should be 'portage'. Bug #34260.
126
127 28 Feb 2004; Martin Schlemmer <azarah@gentoo.org> libsandbox.c :
128 Besides a small cleanup, redo how we replace LD_PRELOAD in the environ passed
129 to the real execve (in our execve wrapper). Seems that on some arches (sparc
130 among others) do not allow us to tamper with the readonly copy passed to
131 execve, so pass our own copy of the environment. Bug #42290.
132
133 11 Jan 2004; Nicholas Jones <carpaski@gentoo.org> create-decls:
134 Changed tail to head and added a notice about duration of glibc check.
135
136 21 Dec 2003; Nicholas Jones <carpaski@gentoo.org> create-decls:
137 Changed the glibc subversion check to use /usr/bin/* instead of /bin/sh
138 as there isn't a guarentee that it is dynamic.
139
140 02 Nov 2003; Martin Schlemmer <azarah@gentoo.org> libsandbox.c :
141 If 'file' passed to before_syscall(const char *func, const char *file) is
142 invalid, we should set errno to ENOENT, and not EINVAL. This should
143 close bug #32238.
144
145 14 Oct 2003; Martin Schlemmer <azarah@gentoo.org> libsandbox.c :
146 Fix a bug that occurs mainly on 64bit arch, where the file passed to
147 the functions we wrap, is invalid, and then cause canonicalize to pass
148 garbage to before_syscall(), thanks to great detective work from
149 Andrea Luzzardi <al@sig11.org> (bug #29846).
150
151 13 Oct 2003; Martin Schlemmer <azarah@gentoo.org> create-localdecls :
152 Add a uClibc detection patch from Peter S. Mazinger <ps.m@gmx.net>.
153
154 13 Oct 2003; Martin Schlemmer <azarah@gentoo.org> libsandbox.c :
155 Fix a bug in libsandbox.c 's checking in the rename wrapper - it basically
156 only checked the destination patch, and not the source, so we could move
157 a protected file to a unprotected directory, and then delete/modify it.
158 Thanks to Andrea Luzzardi (scox) <al@sig11.org>, bug #30992, for this fix.
159
160 12 Oct 2003; Nicholas Jones <carpaski@gentoo.org> sandbox.c :
161 Added python2.3 to the predict section/variable.
162
163 28 Sep 2003; Martin Schlemmer <azarah@gentoo.org> libsandbox.c, sandbox.c,
164 sandbox.h, sandbox_futils.c :
165 Add support to set the pids file via SANDBOX_PIDS_FILE at startup. If
166 it is not set, it will revert to its old value.
167
168 27 Sep 2003; Martin Schlemmer <azarah@gentoo.org> libsandbox.c :
169 Fix our mkdir wrapper to check if the dir exist, and return EEXIST if so,
170 rather than failing with a violation, bug #29748.
171
172 27 Jul 2003; Martin Schlemmer <azarah@gentoo.org> libsandbox.c :
173 Fix canonicalize() to ignore calls with path = "".
174
175 27 Jul 2003; Martin Schlemmer <azarah@gentoo.org> getcwd.c, libsandbox.c,
176 sandbox_futils.c, canonicalize.c :
177 Once again coreutils fails, as my systems had 2.5 kernel, the getcwd system
178 call handled strings larger than PATH_MAX (bug #21766). It however does not
179 work the same on 2.4 kernels.
180
181 To fix, I added the posix implementation of getcwd() (from glibc cvs) that
182 do not need the system call. We use the default getcwd() function via a
183 wrapper (egetcwd), and then lstat the returned path. If lstat fails, it
184 means the current directory was removed, OR that the the system call for
185 getcwd failed (curious is that it do not fail and return NULL or set
186 errno, but rather just truncate the retured directory - usually from the
187 start), and if so, we use the generic getcwd() function (__egetcwd). Note
188 that we do not use the generic version all the time, as it calls lstat()
189 a great number of times, and performance degrade much.
190
191 29 Jun 2003; Martin Schlemmer <azarah@gentoo.org> create-localdecls,
192 libsandbox.c :
193 Make sure SB_PATH_MAX will not wrap. Fix two possible memory leaks.
194
195 22 Jun 2003; Martin Schlemmer <azarah@gentoo.org> libsandbox.c, canonicalize.c
196 create-localdecls :
197 When checking path names of files accessed, we need to canonicalize it, else
198 it may be a symlink in a 'write allowed' directory pointing to a file in a
199 directory we should not have write access to.
200
201 With something like coreutils-5.0, we have two problems:
202 1) One of the tests checks if getcwd() can return a path longer than
203 PATH_MAX. This test then tries to create a dir which even while
204 created local (mkdir("conftest2")), it ends up being resolved with
205 a name that is much larger than PATH_MAX. The problem now is that
206 canonicalize() have undefined behaviour when the path was too long
207 (returned wrongly truncated paths, etc), and pass the wrong path to
208 before_syscall() (causing the bogus sandbox violations).
209 2) The ecanonicalize() function we used, along with the canonicalize()
210 function did not support longer than PATH_MAX. This is partly a
211 cause for 1), but the error checking (rather lack of it) of calls
212 to erealpath() in canonicalize() was the prime reason for 1).
213
214 As we do not use this canonicalized name to call the function, we resolve this
215 by fixing canonicalize() to do better error checking, and ecanonicalize() as
216 well as all functions in libsandbox.c to use a PATH_MAX of 'PATH_MAX * 2'.
217 While they will resolve paths properly now, and can check if a write/read is
218 allowed, the functions called from the sandboxed environment will still work
219 as expected.
220
221 This should resolve bug #21766.
222
223 06 Apr 2003; Martin Schlemmer <azarah@gentoo.org> libsandbox.c :
224 For some reason sandbox fails with a 'open_wr' if you run 'locale -a' under
225 it (bug #16298).
226
227 Problem is that for some reason locale fopen's locale.alias with mode "rm".
228
229 -------------------------------------------------------
230 nosferatu root # grep fopen locale.log
231 fopen("/usr/share/locale/locale.alias", "rm"ACCESS DENIED open_wr: /usr/share/locale/locale.alias
232 nosferatu root #
233 --------------------------------------------------------
234
235 I checked the source of locale, but it have fopen with mode 'r', so
236 not sure where the "rm" mode comes from. Anyhow, changed the check in
237 before_syscall_open_char() to also see mode "rm" as readonly.
238
239 23 Feb 2003; Martin Schlemmer <azarah@gentoo.org> create-localdecls :
240
241 Add glibc-2.3 support.
242
243 22 Feb 2003; Martin Schlemmer <azarah@gentoo.org> sandbox.c :
244
245 Some /etc/ld.so.preload fixes. Just changed the #if defines to cover all
246 operations releated to preload, as well as only try to modify ld.so.preload
247 if we can. Also modify to write the pid to /tmp/sandboxpids.tmp even when
248 not using ld.so.preload. Fix to not write this instance of sandbox's pid
249 to /tmp/sandboxpids.tmp on exit if this is not the last sandbox running.
250
251 22 Feb 2003; Nicholas Jones <carpaski@gentoo.org> Makefile :
252
253 Changed the LD to CC for hppa.
254
255 22 Feb 2003; Nicholas Jones <carpaski@gentoo.org> create-localdecls :
256
257 Killed the previous changes I made.
258
259 17 Feb 2003; Nicholas Jones <carpaski@gentoo.org> create-localdecls :
260
261 Added parisc to BROKEN_RTLD_ARCHLIST to see if it we can fix the relocation probs.
262
263 09 Jan 2003; J Robert Ray <jrray@gentoo.org> sandbox.c :
264
265 Don't segfault if $HOME isn't set, set $HOME to "/" instead. Fixes bug 10868.
266
267 16 Dec 2002; Martin Schlemmer <azarah@gentoo.org> create-localdecls :
268
269 Fix memory leak for mips, bug #12236. Thanks to Torgeir Hansen <torgeir@trenger.ro>
270 for this fix.
271
272 4 Dec 2002; J Robert Ray <jrray@gentoo.org> sandbox.h sandbox_futils.c :
273
274 sandbox_futils defined a dirname() function that was masking the same
275 function in glibc and was broken (e.g.: SANDBOX_DIR was being set to
276 '/usr/lib/portage/bi/'). Fixed function to return expected results and
277 renamed it to sb_dirname() to no longer mask the glibc function. Closes bug
278 11231.
279
280 4 Dec 2002; Martin Schlemmer <azarah@gentoo.org> :
281
282 Fix a segfault in libsandbox.c if canonicalize() was called with
283 first parameter = NULL.
284
285 1 Sep 2002; Martin Schlemmer <azarah@gentoo.org> :
286
287 Fix my braindead 'return 1;' in a void function. Updated sandbox.c,
288 cleanup() for this.
289
290 Change cleanup() in sandbox.c not to exit with fail status if
291 the pidsfile is missing. We really should still display sandbox
292 violations if they occured.
293
294 31 Aug 2002; Martin Schlemmer <azarah@gentoo.org> :
295
296 Update cleanup() in sandbox.c to remove the PIDSFILE if this is
297 the last sandbox running.
298
299 25 Aug 2002; Martin Schlemmer <azarah@gentoo.org> :
300
301 Major cleanups to mainly libsandbox.c again.
302
303 22 Aug 2002; Martin Schlemmer <azarah@gentoo.org> :
304
305 Add copyrights to sandbox.h and sandbox_futils.h. If wrong, the
306 parties involved should please contact me so that we can fix it.
307
308 Add opendir wrapper to libsandbox.c.
309
310 21 Aug 2002; Martin Schlemmer <azarah@gentoo.org> :
311
312 Do some more cleanups to ecanonicalize(), as it dropped filenames in
313 rare cases (after my symlink cleanups), and caused glibc to bork.
314 These fixes went into canonicalize.c.
315
316 20 Aug 2002; Martin Schlemmer <azarah@gentoo.org> :
317
318 Fix spawn_shell() and main() in sandbox.c to properly return fail
319 status.
320
321 19 Aug 2002; Martin Schlemmer <azarah@gentoo.org> :
322
323 The new canonicalize() function in libsandbox.c also resolved symlinks,
324 which caused on cleaning sandbox errors if the symlink pointed to a
325 file in the live root. Ripped out canonicalize() and realpath() from
326 glibc; removed the symlink stuff, and changed them to ecanonicalize()
327 and erealpath().
328
329 18 Aug 2002; Martin Schlemmer <azarah@gentoo.org> :
330
331 Ripped out all the wrappers, and implemented those of InstallWatch.
332 Losts of cleanups and bugfixes. Implement a execve that forces
333 $LIBSANDBOX in $LD_PRELOAD. We can now thus do away with the feared
334 /etc/ld.so.preload (*g*) ... Made the needed changes to sandbox.c,
335 sandbox.h and sandbox_futils.c. Rewrote the Makefile for most
336 parts; it now have an install target.
337
338 Reformat the whole thing to look somewhat like the reworked sandbox.c
339 and new sandbox.h and sandbox_futils.c from:
340
341 Brad House <brad@mainstreetsoftworks.com>.
342
343 Additional Copyrights now due to the InstallWatch code:
344
345 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