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