This patch adds the following mount options: 'nostrictatime',
'lazytime', and 'nolazytime'.
The MS_STRICTATIME mount flag already existed, and 'nostrictatime' was
listed along with 'strictatime' in the comments of parser/mount.cc, so
this patch adds a mapping for 'nostrictatime' to clear MS_STRICTATIME.
Additionally, the Linux kernel includes the 'lazytime' option with
MS_LAZYTIME mapping to (1<<25), so this patch adds MS_LAZYTIME to
parser/mount.h and the corresponding mappings in parser/mount.cc for
'lazytime' and 'nolazytime'.
MR: https://gitlab.com/apparmor/apparmor/-/merge_requests/1005
Approved-by: John Johansen <john@jjmx.net>
Merged-by: John Johansen <john@jjmx.net>
(cherry picked from commit c37be61d17)
Signed-off-by: Jon Tourville <jon.tourville@canonical.com>
The conflicting flags value message was hard to read
conflicting flag value = lazytimenolazytime
change it to
conflicting flag values = lazytime, nolazytime
Signed-off-by: John Johansen <john.johansen@canonical.com>
(cherry picked from commit 89bc617d0d)
Signed-off-by: Jon Tourville <jon.tourville@canonical.com>
This patch adds the following mount options: 'nostrictatime',
'lazytime', and 'nolazytime'.
The MS_STRICTATIME mount flag already existed, and 'nostrictatime' was
listed along with 'strictatime' in the comments of parser/mount.cc, so
this patch adds a mapping for 'nostrictatime' to clear MS_STRICTATIME.
Additionally, the Linux kernel includes the 'lazytime' option with
MS_LAZYTIME mapping to (1<<25), so this patch adds MS_LAZYTIME to
parser/mount.h and the corresponding mappings in parser/mount.cc for
'lazytime' and 'nolazytime'.
Signed-off-by: Oliver Calder <oliver.calder@canonical.com>
Signed-off-by: John Johansen <john.johansen@canonical.com>
(cherry picked from commit bc64b824fa)
Signed-off-by: Jon Tourville <jon.tourville@canonical.com>
The mount options MS_LAZYTIME and MS_NOSYMFOLLOW were added in
kernels 4.0 and 5.10, respectively. Update the mount test script
and helper to skip testing those options if they are not available.
Signed-off-by: Jon Tourville <jon.tourville@canonical.com>
(cherry picked from commit 9a760def8d)
Signed-off-by: Jon Tourville <jon.tourville@canonical.com>
This is a partial fix for CVE-2016-1585, it address the frontend rule encoding problems particularly
- Permissions being given that shouldn't happen
- Multiple option conditionals in a single rule resulting in wider permission instead of multiple rules
- optional flags not being handled correctly
- multiple backend rules being created out of one frontend rule when they shouldn't be
it does not address the backend issue of short cut permissions not being correctly updated when deny rules carve out permissions on an allow rule that has a short cut permission in the encoding.
Thanks to the additional work by Alexander Mikhalitsyn for beating this MR into shape so we can land it
Alexander Changelog:
- rebased to an actual tree
- addressed review comments from @wbumiller and @setharnold
- fixed compiler warnings about class_mount_hdr is uninitialized
- infinite loop fix
- MS_MAKE_CMDS bitmask value fixed
- fixed condition in `gen_flag_rules` to cover cases like `mount options in (bind) /d -> /4,` when flags are empty and only opt_flags are present
- marked some tests as a FAIL case behavior was changed after `parser: add conflicting flags check for options= conditionals` commit
MR: https://gitlab.com/apparmor/apparmor/-/merge_requests/333
Approved-by: John Johansen <john@jjmx.net>
Merged-by: John Johansen <john@jjmx.net>
(cherry picked from commit c1a1a3a923)
Signed-off-by: Jon Tourville <jon.tourville@canonical.com>
MR: https://gitlab.com/apparmor/apparmor/-/merge_requests/1029
Approved-by: Georgia Garcia <georgia.garcia@canonical.com>
Merged-by: Jon Tourville <jon.tourville@canonical.com>
`unscd` is a drop-in replacement for `nscd` that uses the same binary location (`/usr/sbin/nscd`) and config file (`/etc/nscd.conf`). The `usr.sbin.nscd` profile only needs one additional permission to support it.
```
May 9 18:07:42 darkstar kernel: [ 2706.138823] audit: type=1400
audit(1683670062.580:839): apparmor="DENIED" operation="sendmsg"
profile="nscd" name="/run/systemd/notify" pid=4343 comm="nscd"
requested_mask="w" denied_mask="w" fsuid=125 ouid=0
```
MR: https://gitlab.com/apparmor/apparmor/-/merge_requests/1031
Approved-by: Christian Boltz <apparmor@cboltz.de>
Merged-by: Christian Boltz <apparmor@cboltz.de>
(cherry picked from commit dec3815f07)
bd0d401b nscd: add permission to allow supporting unscd
This is a partial fix for CVE-2016-1585, it address the frontend rule encoding problems particularly
- Permissions being given that shouldn't happen
- Multiple option conditionals in a single rule resulting in wider permission instead of multiple rules
- optional flags not being handled correctly
- multiple backend rules being created out of one frontend rule when they shouldn't be
it does not address the backend issue of short cut permissions not being correctly updated when deny rules carve out permissions on an allow rule that has a short cut permission in the encoding.
Thanks to the additional work by Alexander Mikhalitsyn for beating this MR into shape so we can land it
Alexander Changelog:
- rebased to an actual tree
- addressed review comments from @wbumiller and @setharnold
- fixed compiler warnings about class_mount_hdr is uninitialized
- infinite loop fix
- MS_MAKE_CMDS bitmask value fixed
- fixed condition in `gen_flag_rules` to cover cases like `mount options in (bind) /d -> /4,` when flags are empty and only opt_flags are present
- marked some tests as a FAIL case behavior was changed after `parser: add conflicting flags check for options= conditionals` commit
MR: https://gitlab.com/apparmor/apparmor/-/merge_requests/333
Approved-by: John Johansen <john@jjmx.net>
Merged-by: John Johansen <john@jjmx.net>
(cherry picked from commit c1a1a3a923)
Signed-off-by: Jon Tourville <jon.tourville@canonical.com>
gen_flag_rules has a boolean vs bit and case where parenthesis are
helpful to express the intended order of operations.
It also doesn't handle the case where there are no matches. Fix this
by causing that case to fail.
also improve the debug of option extraction.
Signed-off-by: John Johansen <john.johansen@canonical.com>
(cherry picked from commit f09676f5f9)
Signed-off-by: Jon Tourville <jon.tourville@canonical.com>
Now that flag processing for mount rules with single option
conditionals are fixed e-enable multiple mount conditionals on a
single mount rule. The mount conditionals are equivalent to specifying
multiple rules.
mount options=(a,b,c) options=(c,d),
is the same as
mount options=(a,b,c),
mount options=(c,d),
and
mount options in (a,b,c) options in (c,d),
is the same as
mount options in (a,b,c),
mount options in (c,d),
when multiple options= and options in are combined in a single rule
it is the same as the cross product of the options.
where
mount options=(a,b,c) options in (d,e),
is a single rule.
mount options=(a,b,c) options=(d,e) options in (f),
is equivalent to
mount options=(a,b,c) options in (f),
mount options=(d,e) options in (f),
and while it is not recommended that multiple options= and options in
conditions be used in a single rule.
mount options=(a,b,c) options=(d,e) options in (f) options in (g),
is equivalent to
mount options=(a,b,c) options in (f),
mount options=(a,b,c) options in (g),
mount options=(d,e) options in (f),
mount options=(d,e) options in (g),
Bug Link: https://bugs.launchpad.net/apparmor/+bug/1597017
Signed-off-by: John Johansen <john.johansen@canonical.com>
- rebased to bba1a023bf
- fixed infinite loop in mnt_rule::gen_policy_re
Signed-off-by: Alexander Mikhalitsyn <aleksandr.mikhalitsyn@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
(cherry picked from commit 1ec39fd437)
Signed-off-by: Jon Tourville <jon.tourville@canonical.com>
The combined optional flag and exact match flag processing is problematic
separate out the optional flag processing so it is only combined during
match string generation.
While doing so we fix the flag output so that multiple rules are
not output when they shouldn't be.
In addition we temporarily break multiple options= and 'options in'
conditionals in a single rule, which we will fix in a separate patch.
Bug Link: https://bugs.launchpad.net/apparmor/+bug/1597017
Signed-off-by: John Johansen <john.johansen@canonical.com>
- rebased to bba1a023bf
- made tests happy by changing condition in gen_policy_re()
Signed-off-by: Alexander Mikhalitsyn <aleksandr.mikhalitsyn@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
(cherry picked from commit 300889c3a4)
Signed-off-by: Jon Tourville <jon.tourville@canonical.com>
This allows regression tests to generate profiles that use rule qualifiers,
such as allow, deny, and audit. Qualifiers can be specified for a rule by
prepending 'qual=', followed by a comma-separated list of rule qualifiers,
then a ':', then the rule itself.
Signed-off-by: Jon Tourville <jon.tourville@canonical.com>
(cherry picked from commit f6bfd141bd)
Signed-off-by: Jon Tourville <jon.tourville@canonical.com>
The unix network tests are not being run on a v8 network capable kernel. Under v8 there needs to be some adjustments to the tests because unix rules get downgraded to the socket rule ```network unix,``` which does not have the same set of conditionals or fine grained permissions, meaning some tests that would fail under af_unix (like missing permission tests) will pass under v8 network rules.
Signed-off-by: John Johansen <john.johansen@canonical.com>
MR: https://gitlab.com/apparmor/apparmor/-/merge_requests/893
Approved-by: Georgia Garcia <georgia.garcia@canonical.com>
Merged-by: John Johansen <john@jjmx.net>
(cherry picked from commit 59b4109a8b)
Signed-off-by: John Johansen <john.johansen@canonical.com>
The regression tests are failing on some older kernels due to
commit 9f834ec18defc369d73ccf9e87a2790bfa05bf46 being cherry-picked
back to them without the corresponding apparmor patch
34c426acb75cc21bdf84685e106db0c1a3565057.
This means we can not rely on a simple features/flag check to determine
how the kernel is behaving with regard to mmap. Since this test is
not concerned with testing mmap, instead of adding a more complex
conditional simplify by always adding the m permission.
Fixes: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1830984
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Georgia Garcia <georgia.garcia@canonical.com>
(cherry picked from commit e6e112fba1)
Signed-off-by: John Johansen <john.johansen@canonical.com>
This is used by various applications including libreoffice etc so it may as well
be added to the base abstraction along with the existing zoneinfo DB access.
AVC apparmor="DENIED" operation="open" class="file" profile="snap.libreoffice.calc" name="/usr/share/zoneinfo-icu/44/le/zoneinfo64.res" pid=44742 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=0
Signed-off-by: Alex Murray <alex.murray@canonical.com>
MR: https://gitlab.com/apparmor/apparmor/-/merge_requests/1007
Approved-by: Georgia Garcia <georgia.garcia@canonical.com>
Merged-by: Christian Boltz <apparmor@cboltz.de>
(cherry picked from commit 8d9985ac0b)
c691b11d abstractions/base: allow reading tzdata ICU zoneinfo DB
Glibc in 2.36 and later will [1] access sysfs at
/sys/devices/system/cpu/possible when usig sysconf
for _SC_NPROCESSORS_CONF.
That will make a lot of different code, for example
anything linked against libnuma, trigger this apparmor
denial.
apparmor="DENIED" operation="open" class="file" ...
name="/sys/devices/system/cpu/possible" ...
requested_mask="r" denied_mask="r" fsuid=0 ouid=0

This entry seems rather safe, and it follows others
that are already in place. Instead of fixing each
software individually this should go into the base
profile as well.
Initially reported via
https://bugs.launchpad.net/ubuntu/+source/apparmor/+bug/1989073
Fixes: https://gitlab.com/apparmor/apparmor/-/issues/267
MR: none - ML
Signed-off-by: Christian Ehrhardt <christian.ehrhardt@canonical.com>
Signed-off-by: John Johansen <john.johansen@canonical.com>
(cherry picked from commit c159d0925a)
Signed-off-by: John Johansen <john.johansen@canonical.com>
... instead of slicing to check for prefixes and suffixes.
This change prevents a crash in aa-mergeprof - if `replacement` is empty,
trying to access `replacement[0]` causes an IndexError.
Using `.startswith()` works without crashing.
This backports parts of the severity.py changes in
commit 091c6ad59d
by Mark Grassi.
I propose this fix for 2.13 and 3.0. (3.1 and master already have this fix.)
MR: https://gitlab.com/apparmor/apparmor/-/merge_requests/931
Approved-by: Jon Tourville <jon.tourville@canonical.com>
Merged-by: Christian Boltz <apparmor@cboltz.de>
(cherry picked from commit 7ebb259610)
e1714b96 Use string startswith() and endswith() methods
Note: This was reported for /usr/libexec/libvirt_leaseshelper, but since
this is probably unrelated to the path or a path change, this commit
also adds r permissions for the previous path.
Fixes: https://bugzilla.opensuse.org/show_bug.cgi?id=1202161
I propose this patch for 3.0 and master (optionally also for 2.12 and 2.13 - please tell me if you want that after reviewing the patch, or just merge ;-)
MR: https://gitlab.com/apparmor/apparmor/-/merge_requests/905
Approved-by: John Johansen <john@jjmx.net>
Merged-by: John Johansen <john@jjmx.net>
(cherry picked from commit f51049ea2e)
c9c5208f dnsmasq: Add missing r permissions for libvirt_leaseshelper
profiles/Makefile:
Set PYTHON_DIST_BUILD_PATH using
libraries/libapparmor/swig/python/test/buildpath.py as it solves the
problem of setting the build directory generically.
(cherry picked from commit 1ff0c2c7d6,
adjusted to 2.13 which used a different python command before)
libraries/libapparmor/swig/python/test/buildpath.py:
The changes introduced in
cc7f549665
targetted a wrong setuptools version (61.2).
The change in build directory naming has been introduced with 62.1
(1c23f5e1e4).
(cherry picked from commit fda390983f)
(cherry picked from commit 47d68dac0f,
adjusted to the 2.13 branch which used slightly different python
commands. Also, utils/test/README.md doesn't exist in 2.13, therefore
drop the part that changes it)
... which will be removed in Python 3.12, and that probably won't be
used on systems running the AppArmor 2.1x branches.
This prevents CI failures on gitlab.com, which uses a new-enough python
to show
DeprecationWarning: The distutils package is deprecated and slated for removal in Python 3.12. Use setuptools or check PEP 632 for potential alternatives
For 3.0 and master, the proper fix (switching to setuptools) was done in
!813.
I propose this patch for 2.11, 2.12 and 2.13.
MR: https://gitlab.com/apparmor/apparmor/-/merge_requests/908
Approved-by: John Johansen <john@jjmx.net>
Merged-by: Christian Boltz <apparmor@cboltz.de>
... which will be removed in Python 3.12, and that probably won't be
used on systems running the AppArmor 2.1x branches.
This prevents CI failures on gitlab.com, which uses a new-enough python
to show
DeprecationWarning: The distutils package is deprecated and slated for removal in Python 3.12. Use setuptools or check PEP 632 for potential alternatives
For 3.0 and master, the proper fix (switching to setuptools) was done in
!813.
I noticed that some apps return the following errors when launched:
```
kernel: audit: type=1400 audit(1651244478.255:5501): apparmor="DENIED" operation="open" profile="some_app" name="/sys/devices/pci0000:00/0000:00:02.0/revision" pid=1877976 comm="some_app" requested_mask="r" denied_mask="r" fsuid=1000 ouid=0
kernel: audit: type=1400 audit(1651244478.255:5502): apparmor="DENIED" operation="open" profile="some_app" name="/sys/devices/pci0000:00/0000:00:02.0/config" pid=1877976 comm="some_app" requested_mask="r" denied_mask="r" fsuid=1000 ouid=0
```
Blocking the files results in the following errors when the app is executed in a terminal:
```
MESA: error: Failed to query drm device.
libGL error: failed to create dri screen
libGL error: failed to load driver: crocus
MESA: error: Failed to query drm device.
libGL error: failed to create dri screen
libGL error: failed to load driver: crocus
```
Since they have something to do with MESA, I think the mesa abstraction should
be updated to fix the issue.
MR: https://gitlab.com/apparmor/apparmor/-/merge_requests/879
Approved-by: John Johansen <john@jjmx.net>
Merged-by: John Johansen <john@jjmx.net>
Signed-off-by: John Johansen <john.johansen@canonical.com>
The snap_browsers abstraction requires more permissions
due to updates on snaps.
Some of the permissions are not required in older versions of
Ubuntu that use 2.12 and 2.13, but are introduced for unification
and ease of maintenance purposes. These include:
```
all dbus permissions,
@{PROC}/sys/kernel/random/uuid r,
owner @{PROC}/@{pid}/cgroup r,
/var/lib/snapd/sequence/{chromium,firefox,opera}.json r,
```
I also propose a cherry-pick of this commit to 2.12, 2.13 and 3.0
MR: https://gitlab.com/apparmor/apparmor/-/merge_requests/877
Approved-by: John Johansen <john@jjmx.net>
Merged-by: John Johansen <john@jjmx.net>
(cherry picked from commit bfa67b369d)
Signed-off-by: John Johansen <john.johansen@canonical.com>
similar to commit 2f9d172c64
we discovered that there was a service outage
when dovecot tried to send a usr1 signal
type=AVC msg=audit(1648024138.249:184964): apparmor="DENIED" operation="signal" profile="dovecot" pid=1690 comm="dovecot" requested_mask="send" denied_mask="send" signal=usr1 peer="dovecot-imap-login"
MR: https://gitlab.com/apparmor/apparmor/-/merge_requests/865
Approved-by: Christian Boltz <apparmor@cboltz.de>
Merged-by: Christian Boltz <apparmor@cboltz.de>
(cherry picked from commit 83685ba703)
f0919f83 Allow dovecot to use all signals
Whenever the evince deb package tries to open a snap browser which was
selected as the default, we get the following denial:
audit[2110]: AVC apparmor="DENIED" operation="exec" profile="/usr/bin/evince" name="/usr/bin/snap" pid=2110 comm="env" requested_mask="x" denied_mask="x" fsuid=1000 ouid=0
As a short-term solution, we are adding a snap-browsers profile
which restricts what snaps opened by evince can do.
The long-term solution is currently not available, but could be
accomplished by using enhanced environment variable filtering/mediation
and delegation of open fds.
Bug: https://launchpad.net/bugs/1794064
Signed-off-by: Georgia Garcia <georgia.garcia@canonical.com>
(cherry picked from commit fb3283f37e)
MR: https://gitlab.com/apparmor/apparmor/-/merge_requests/863
Signed-off-by: Georgia Garcia <georgia.garcia@canonical.com>
If apparmor_parser -N (in profiles_names_list()) fails,
aa-remove-unknown possibly gets an incomplete list of profiles in
/etc/apparmor.d/ and therefore might remove more profiles than it
should.
Replace the profiles_names_list() call with a direct apparmor_parser
call, and abort aa-remove-unknown if it exits with $? != 0
Before:
```
aa-remove-unknown -n
AppArmor parser error for /etc/apparmor.d/broken in profile /etc/apparmor.d/broken at line 1: syntax error, unexpected TOK_ID, expecting TOK_OPEN
Would remove 'delete_me'
```
After:
```
./aa-remove-unknown -n
AppArmor parser error for /etc/apparmor.d in profile /etc/apparmor.d/zbroken at line 1: syntax error, unexpected TOK_ID, expecting TOK_OPEN
apparmor_parser exited with failure, aborting.
```
And of course, after fixing the broken profile:
```
./aa-remove-unknown -n
Would remove 'delete_me'
```
(cherry picked from commit 5053a01d84)
This backports the fix in `aa-remove-unknown` from !836, but doesn't backport the cleanup in `rc.apparmor.functions`.
I propose this patch for 3.0 and all 2.x branches.
MR: https://gitlab.com/apparmor/apparmor/-/merge_requests/859
Approved-by: Georgia Garcia <georgia.garcia@canonical.com>
Merged-by: John Johansen <john@jjmx.net>
(cherry picked from commit c6324c2a3e)
Signed-off-by: John Johansen <john.johansen@canonical.com>
parser: Fix unknown state condition RLIMIT_MODEINCLDE
See merge request apparmor/apparmor!803
Acked-by: Christian Boltz for all branches
(cherry picked from commit b0bc0d5323)
dc7755e5 parser: Fix unknown state condition RLIMIT_MODEINCLDE
add a missing slash at the end of the sys rule
See merge request apparmor/apparmor!791
Acked-by: Christian Boltz <apparmor@cboltz.de> for 2.12..master
(cherry picked from commit 4fd7bcc289)
b3dcd02d add a missing slash at the end of the sys rule
Currently for directory includes the directory timestamp is ignored.
This is wrong as operations like removing a file from the dir won't
be considered in the timestamp check.
Fix this by updating the timestamp check to include the included
directories timestamp.
MR: https://gitlab.com/apparmor/apparmor/-/merge_requests/760
Signed-off-by: John Johansen <john@jjmx.net>
Acked-by: Georgia Garcia <georgia.garcia@canonical.com>
(cherry picked from commit 3d1232e640)
Currently the wutmp abstraction has the following rules:
/var/log/lastlog rwk,
/var/log/wtmp wk,
@{run}/utmp rwk,
According to what I see in my apparmor profiles, just a few apps want
to interact with the files listed above, especially with the
/var/log/wtmp . But when the apps do this, they sometimes want the
read access to this file. An example could be the last command. Is
there any reason for not having the r in the rule? The second thing
is the file /var/log/btmp (which isn't included in the
abstracion). Whenever I see an app, which wants to access the
/var/log/wtmp file, it also tries to interact with the /var/log/btmp
file, for instance lightdm/sddm or su . Most of the time they need
just wk permissions, but sometimes apps need also r on this file, an
example could be the lastb command, which is just a link to last.
Fixes: https://gitlab.com/apparmor/apparmor/-/issues/152
MR: https://gitlab.com/apparmor/apparmor/-/merge_requests/724
Signed-off-by: John Johansen <john.johansen@canonical.com>
(cherry picked from commit d4e0a94511)
Signed-off-by: John Johansen <john.johansen@canonical.com>
The backport of
855dbd4a parser: fix rule downgrade for unix rules
using the rule_t::warn_once which doesn't exist in the 2.x parser
series. Switch this the the static function warn_once.
Fixes: 3d85e123 parser: fix rule downgrade for unix rules
Signed-off-by: John Johansen <john.johansen@canonical.com>
Rule downgrades are used to provide some confinement when a feature
is only partially supported by the kernel.
Eg. On a kernel that doesn't support fine grained af_unix mediation
but does support network mediation.
unix (connect, receive, send)
type=stream
peer=(addr="@/tmp/.ICE-unix/[0-9]*"),
will be downgraded to
network unix type=stream,
Which while more permissive still provides some mediation while
allowing the appication to still function. However making the rule
a deny rule result in tightening the profile.
Eg.
deny unix (connect, receive, send)
type=stream
peer=(addr="@/tmp/.ICE-unix/[0-9]*"),
will be downgraded to
deny network unix type=stream,
and that deny rule will take priority over any allow rule. Which means
that if the profile also had unix allow rules they will get blocked by
the downgraded deny rule, because deny rules have a higher priority,
and the application will break. Even worse there is no way to add the
functionality back to the profile without deleting the offending deny
rule.
To fix this we drop deny rules that can't be downgraded in a way that
won't break the application.
Fixes: https://bugzilla.opensuse.org/show_bug.cgi?id=1180766
MR: https://gitlab.com/apparmor/apparmor/-/merge_requests/700
Signed-off-by: John Johansen <john.johansen@canonical.com>
(cherry picked from commit 855dbd4ac8)
Signed-off-by: John Johansen <john.johansen@canonical.com>
My main user account is managed by systemd-homed. When I enable
AppArmor and have nscd running, I get inconsistent behavior with my
user account - sometimes I can't log in, sometimes I can log in but
not use sudo, etc.
This is the output of getent passwd:
$ getent passwd
root❌0:0::/root:/usr/bin/zsh
bin❌1:1::/:/sbin/nologin
daemon❌2:2::/:/sbin/nologin
mail❌8:12::/var/spool/mail:/sbin/nologin
ftp❌14:11::/srv/ftp:/sbin/nologin
http❌33:33::/srv/http:/sbin/nologin
nobody❌65534:65534:Nobody:/:/sbin/nologin
dbus❌81:81:System Message Bus:/:/sbin/nologin
[...]
rose❌1000:1000:Rose Kunkel:/home/rose:/usr/bin/zsh
But getent passwd rose and getent passwd 1000 both return no output.
Stopping nscd.service fixes these problems. Checking the apparmor
logs, I noticed that nscd was denied access to
/etc/machine-id. Allowing access to that file seems to have fixed the
issue.
MR: https://gitlab.com/apparmor/apparmor/-/merge_requests/707
Fixes: https://gitlab.com/apparmor/apparmor/-/issues/145
Signed-off-by: John Johansen <john.johansen@canonical.com>
(cherry picked from commit ee5303c8a0)
Signed-off-by: John Johansen <john.johansen@canonical.com>
Ubuntu 18.04, Firefox 60.0.1+build2-0ubuntu0.18.04.1
Running firefix, then going to netflix.com and attempting to play a
movie. The widevinecdm plugin crashes, the following is found in
syslog:
Jun 15 19:13:22 xplt kernel: [301351.553043] audit: type=1400 audit(1529046802.585:246): apparmor="DENIED" operation="file_mmap" profile="/usr/lib/firefox/firefox{,*[^s][^h]}" name="/home/xav/.mozilla/firefox/wiavokxk.default-1510977878171/gmp-widevinecdm/1.4.8.1008/libwidevinecdm.so" pid=16118 comm="plugin-containe" requested_mask="m" denied_mask="m" fsuid=1000 ouid=1000
Jun 15 19:13:22 xplt kernel: [301351.553236] audit: type=1400 audit(1529046802.585:247): apparmor="DENIED" operation="ptrace" profile="/usr/lib/firefox/firefox{,*[^s][^h]}" pid=24714 comm="firefox" requested_mask="trace" denied_mask="trace" peer="/usr/lib/firefox/firefox{,*[^s][^h]}"
Jun 15 19:13:22 xplt kernel: [301351.553259] plugin-containe[16118]: segfault at 0 ip 00007fcdfdaa76af sp 00007ffc1ff03e28 error 6 in libxul.so[7fcdfb77a000+6111000]
Jun 15 19:13:22 xplt snmpd[2334]: error on subcontainer 'ia_addr' insert
...
Fixes: https://bugs.launchpad.net/ubuntu/+source/firefox/+bug/1777070
Reported-by: Xav Paice <xav.paice@canonical.com>
MR: https://gitlab.com/apparmor/apparmor/-/merge_requests/684
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Steve Beattie <steve.beattie@canonical.com>
(cherry picked from commit 656f2103ed)
Signed-off-by: John Johansen <john.johansen@canonical.com>
job scaling allows the parser to resample the number of cpus available
and increase the number of jobs that can be launched if cpu available
increases.
Unfortunately job scaling was being applied even when a fixed number
of jobs was specified. So
--jobs=2
doesn't actually clamp the compile at 2 jobs.
Instead job scaling should only be applied when --jobs=auto or when
jobs are set to a multiple of the cpus.
MR: https://gitlab.com/apparmor/apparmor/-/merge_requests/703
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Steve Beattie <steve.beattie@canonical.com>
(cherry picked from commit 65ba20b955)
This is needed to catch conflicts between uppercase and lowercase
hotkeys of the same letter, as seen with `(B)enannt` and `A(b)lehnen` in
the german utils translations.
(cherry picked from commit 07bd11390e)
Dovecot is hit with this denial on Debian 10 (buster):
```
type=AVC msg=audit(1603647096.369:24514): apparmor="DENIED"
operation="open" profile="dovecot" name="/usr/share/dovecot/dh.pem"
pid=28774 comm="doveconf" requested_mask="r" denied_mask="r" fsuid=0
ouid=0
```
This results in fatal error:
```
Oct 25 19:31:36 dovecot[28774]: doveconf: Fatal: Error in configuration
file /etc/dovecot/conf.d/10-ssl.conf line 50: ssl_dh: Can't open file
/usr/share/dovecot/dh.pem: Permission denied
```
Add rule to allow reading dh.pem.
MR: https://gitlab.com/apparmor/apparmor/-/merge_requests/671
(cherry picked from commit 9d8e111abe)
Signed-off-by: John Johansen <john.johansen@canonical.com>
On arch
make -C profiles check-logprof
fails with
*** Checking profiles from ./apparmor.d against logprof
ERROR: Can't find AppArmor profiles in /etc/apparmor.d
make: *** [Makefile:113: check-logprof] Error 1
make: Leaving directory '/build/apparmor/src/apparmor-2.13.3/profiles'
because /etc/apparmor.d/ is not available in the build environment
and aa-logprofs --dir argument, is not being passed to init_aa()
but used to update profiles_dir after the fact.
Fix this by passing profiledir as an argument to init_aa()
Fixes: https://gitlab.com/apparmor/apparmor/-/issues/36
MR: https://gitlab.com/apparmor/apparmor/-/merge_requests/663
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
(backported from commit 15dc06248c)
ab0f4ab2ed increased `AA_LIB_REVISION` and `AA_LIB_AGE`, with the result that 2.13.5 builds `libapparmor.so.0.7.3`, while 2.13.4 had `libapparmor-1.6.2`
This patch reverts the `AA_LIB_AGE` increase to fix the so name so that we'll get `libapparmor-1.6.3`.
Note: If you want to apply this fix on top of the 2.13.5 tarball, you'll need to also apply the patch to `Makefile.in`.
MR: https://gitlab.com/apparmor/apparmor/-/merge_requests/658
Acked-by: John Johansen <john.johansen@canonical.com>
ab0f4ab2ed increased AA_LIB_REVISION and
AA_LIB_AGE, with the result that 2.13.5 builds libapparmor.so.0.7.3,
while 2.13.4 had libapparmor-1.6.2
This patch reverts the AA_LIB_AGE increase to fix the so name so that
we'll get libapparmor-1.6.3.
Note: If you want to apply this fix on top of the 2.13.5 tarball, you'll
need to also apply the patch to Makefile.in.
With the backport of static caps to support caps from newer kernels
in older build environments. Builds against older kernels broke
because not all of the newer capabilities are defined in the kernel
headers, nor in apparmor.
In particular
CAP_AUDIT_READ was added to the kernel in 3.16
and
CAP_AUDIT_WRITE, CAP_AUDIT_CONTROL, CAP_SETFCAP, CAP_MAC_OVERRIDE,
CAP_MAC_ADMIN, CAP_SYSLOG, CAP_WAKE_ALARM, CAP_BLOCK_SUSPEND in 3.8
The apparmor kernel module was merge into the upstream kernel in 2.6.36.
In order to support all upstream kernels with apparmor add the set
of capabilities introduced since apparmor was merged upstream.
MR: https://gitlab.com/apparmor/apparmor/-/merge_requests/655
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Steve Beattie <steve.beattie@canonical.com>
Because of the need to be stacking LSM aware, aa_getpeercon() calls
aa_enable to ensure that apparmor is enabled. Without the permission,
aa_getpeercon() fails, causing test failures.
Signed-off-by: Steve Beattie <steve.beattie@canonical.com>
(cherry picked from commit fb773fec36)
Signed-off-by: John Johansen <john.johansen@canonical.com>
When using the in tree parser we should not be using the system
parser.conf file, as if the system apparmor is newer than the
tree being tested the parser.conf file could contain options not
understood by the in tree apparmor_parser.
Use --config-file to specify the default in tree parser.conf
MR: https://gitlab.com/apparmor/apparmor/-/merge_requests/653
Signed-off-by: John Johansen <john.johansen@canonical.com>
(cherry picked from commit 5ac368bce7)
The sed command to manipulate the known cap list (base_cap_names.h) into
a format to match the generated_cap_names.h was buggy because the
trailing '}' would never match anything, leading to failures when built
against 5.8 kernel headers, due to it not replacing the base capabilities
correctly.
Fix this by removing the trailing '}" match and instead match the third
comma-delimited field that matches a capability name, and replace that.
Signed-off-by: Steve Beattie <steve.beattie@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
MR: https://gitlab.com/apparmor/apparmor/-/merge_requests/596
(cherry picked from commit a7fc8bb500)
Signed-off-by: John Johansen <john.johansen@canonical.com>
While `include/sys/apparmor.h` makes use of `socklen_t`, it doesn't
include the `<sys/socket.h>` header to make its declaration available.
While this works on systems using glibc via transitive includes, it
breaks compilation on musl libc.
Fix the issue by including the header.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
(cherry picked from commit 47263a3a74)
Signed-off-by: John Johansen <john.johansen@canonical.com>
While `_aa_asprintf` is supposed to be of private visibility, it's used
by apparmor_parser and thus required to be visible when linking. This
commit thus adds it to the list of private symbols to make it available
for linking in apparmor_parser.
MR: https://gitlab.com/apparmor/apparmor/-/merge_requests/643
Signed-off-by: Patrick Steinhardt <ps@pks.im>
(cherry picked from commit 9a8fee6bf1)
The parser is supposed to add a rule to profiles if they are a hat
or contain hats granting write access to the kernel interfaces
used to perform the change_hat operation.
Unfortunately the check is broken and currently won't add the
rule to hats (it does add it for the parent).
MR: https://gitlab.com/apparmor/apparmor/-/merge_requests/625
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Acked-by: Steve Beattie <steve.beattie@canonical.com>
(cherry picked from commit 5b850c154f)
Refresh the postfix profiles in the 2.13 branch with (mostly) all changes in master.
The most important changes are:
* support having the `postfix/*` binaries in `/usr/lib/postfix/bin/` (like for example openSUSE has now)
* add profile names
* rename the profile files to `postfix-*`
* several "smaller" changes (especially added permissions), see the individual commits for all details
Note that some changes were not backported to the 2.13 branch:
* adding abi rules
* changing `#include` to `include`
* removal of "superfluous" rules covered by abstractions (dd4903efc6)
* removal of `peer=/usr/lib/postfix/...` rules
20/20 Revert renaming usr.lib.postfix.* to postfix.* in 2.13 branch
19/20 postfix-master: allow access to postlog socket
18/20 Allow to read icu *.dat files in postfix-related profiles
17/20 postfix/master needs to execute postfix/error
16/20 Add several permissions to the postfix.* profiles
15/20 adjust postfix profiles for openSUSE path
14/20 profiles/postfix-smtpd: Include ssl_certs, ssl_keys
13/20 profiles/postdrop: Allow reading from pickup socket
12/20 profiles/postfix-pickup: Allow reading from cleanup socket
11/20 postfix.local: Minor adjustments to make it work
10/20 postfix.*: Adapt for new queue names, and extra locking and r/w communication
9/20 postfix.tlsmgr: Connect to urandom and prng exchange
8/20 postfix.master: Change path of child processes
7/20 profiles/postfix: add locking perm to pid files
6/20 profiles: add a postfix dnsblog profile
5/20 profiles: add a postfix postscreen profile
4/20 profiles/postfix-master: grant signal+unix communication with children
3/20 profiles/postfix: use named profiles
2/20 profiles/postfix-master: use profile name instead of match pattern
1/20 allow locking /etc/aliases.db
Note: Backport Exception Requested by OpenSuse, updated profiles needed, on a 2.13 release
MR: https://gitlab.com/apparmor/apparmor/-/merge_requests/621
Acked-by: John Johansen <john.johansen@canonical.com>
On openSUSE Leap 15.1, the postfix binaries live in
/usr/lib/postfix/bin/ which was not covered in the postfix.* attachment
and mrix rules.
(cherry picked from commit f668f31bf0)
This gets us to the local process now, which comes next.
(cherry picked from commit 02528133d2)
Note: this cherry-pick doesn't include the removal of the path-based
signal and unix rules in postfix-master.
Grant the ability to communicate with the postfix named child profiles
via signals and unix sockets. Include the path-based match names as
a fallback on upgrades.
Signed-off-by: Steve Beattie <steve.beattie@canonical.com>
(cherry picked from commit 35d84895fe)
Convert all the postfix subprocesses to using named profiles instead of
path match profiles, and adjust exec paths for newer debian/ubuntu
releses. Rename profiles to match profile names.
Signed-off-by: Steve Beattie <steve.beattie@canonical.com>
(cherry picked from commit 39ca2adff6)
Convert postfix's master profile to use a named profile
(postfix-master) rather than the exec path match pattern. Adjust
postfix-common abstraction to take this into account. Rename profile
name in the profiles/apparmor/profiles/extras/ directory to match
the profile name.
Signed-off-by: Steve Beattie <steve.beattie@canonical.com>
(cherry picked from commit 01648c6a61)
The error:
type=AVC msg=audit(1585403559.846:34317577): apparmor="DENIED" operation="exec" profile="/usr/sbin/dnsmasq" name="/usr/libexec/libvirt_leaseshelper" pid=7162 comm="sh" requested_mas
k="x" denied_mask="x" fsuid=0 ouid=0
type=AVC msg=audit(1585403559.846:34317578): apparmor="DENIED" operation="open" profile="/usr/sbin/dnsmasq" name="/usr/libexec/libvirt_leaseshelper" pid=7162 comm="sh" requested_mas
k="r" denied_mask="r" fsuid=0 ouid=0
Looks like the path to libvirt_leasehelper is incorrect usr.sbin.dnsmasq, at least in gentoo. Patching the file fixes the problem:
issue: https://gitlab.com/apparmor/apparmor/-/issues/87
Signed-off-by: John Johansen <john.johansen@canonical.com>
(cherry picked from commit 28fce5f76d)
This is a partial backport of eb8f9302aa (only adding the abstraction) to the 2.13 branch.
Note that the 2.13 version of the abstraction doesn't have
* the abi rule
* the `include if exists <abstractions/hosts_access.d>`
because both are new in master and would cause interesting\[tm\] problems with the 2.13 tools.
MR: https://gitlab.com/apparmor/apparmor/-/merge_requests/612
Acked-by: John Johansen <john.johansen@canonical.com>
This is a partial backport of eb8f9302aa
(only adding the abstraction) to the 2.13 branch.
Note that the 2.13 version of the abstraction doesn't have
- the abi rule
- the `include if exists <abstractions/hosts_access.d>`
because both are new in master and would cause interesting[tm] problems
with the 2.13 tools.
Add support for CAP BPF and PERFMON
Backport from !578 the subset of patches that convert the parser to a pregenerated list of capability names instead of a dynamically generated list.
The dynamically generated list is still created and compared to the pregenerated list and the build will fail if new capabilities are added to the dynamically generated list that are not in the pregenerated list.
This enables the parser to support new capabilities like CAP BPF and PERFMON that are might not be in the kernel that the parser is being built against.
This patch series adds support CAP BPF and PERFMON in the pregenerated list of capabilities
See merge request apparmor/apparmor!594
prologue.inc:settest() in same cases invokes a sed command that uses
'~' as a pattern separator, on things that can contain filesystem
paths. However, in the debian/ubuntu world, '~' can be used in version
strings, particularly for pre-release versions, and when this happens
and the version is embedded in the path, the sed command breaks
because of the extraneous separator. Fix this by using '#' as a
separator, which has the benefit of being considered a comment if
accidentally interpreted by a shell.
Signed-off-by: Steve Beattie <steve.beattie@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
MR: https://gitlab.com/apparmor/apparmor/-/merge_requests/599
(cherry picked from commit efc6590409)
Signed-off-by: John Johansen <john.johansen@canonical.com>
There are currently two capability name tables the one that is
autogenerated and an internal hardcoded name table.
Now that the autogenerated table has been converted to a base
static table we can drop the internal static table. This
removes the chance of getting the tables getting out of sync.
Signed-off-by: John Johansen <john.johansen@canonical.com>
(cherry picked from commit fb9c5f9bcf)
The static cap sys_log declaration is no longer needed as
base_cap_names.h contains it and ensures that it will always be
present.
Signed-off-by: John Johansen <john.johansen@canonical.com>
(cherry picked from commit fdba3a571c)
The auto-generated cap_names.h has problems when the parser if the
parser is built against a kernel with a smaller capability list than
the kernel policy is being compiled for.
Moving to a pre-generated list lets us support all capabilities even
when we build against older kernels. However we don't want to only use
the pre-generated list as that would make it too easy to miss when a
new capability has been added.
Keep auto generating the caps list and compare it to the pre-generated
caps list so we can detect when new capabilities are added, and fail
the build so that the pre-generated list can be updated. We screen the
diff for only additions so that the parser can continue to build on
older kernels that don't have the full capability list without errors.
Signed-off-by: John Johansen <john.johansen@canonical.com>
(cherry picked from commit 270fb0a2b2)
Nobody told the tools that log events with operation="symlink" exist.
Add this keyword to the list of file or network operations (I don't
expect network symlinks ;-) but keeping everything in that list makes
things easier than special-casing it.)
Also add the log sample and expected result to the libapparmor tests.
Fixes https://gitlab.com/apparmor/apparmor/-/issues/107
(cherry picked from commit 98bf187323)
Signed-off-by: John Johansen <john.johansen@canonical.com>
The commit c8b6d8b393 ("profiles: Update 'make check' to select tools
based on USE_SYSTEM") set a bunch of variables but neglected to apply
them when invoking aa-logprof. This commit addresses this by:
* correcting the PYTHONPATH used with aa-logprof
* setting LD_LIBRARY_PATH when invoking aa-logprof
* adjusting LD_LIBRARY_PATH to include both the directory location
of libapparmor but also the swig libapparmor library needed for
python tools to function.
* adjusts the test for the presence of libapparmor to not use
LD_LIBRARY_PATH but instead a libapparmor specific variable
LIBAPPARMOR_PATH
Bug: https://gitlab.com/apparmor/apparmor/-/issues/98
Fixes: c8b6d8b393
Signed-off-by: Steve Beattie <steve.beattie@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
MR: https://gitlab.com/apparmor/apparmor/-/merge_requests/586
The profiles dirs make check is not always using the correct tools.
Update it to be similar to other Makefiles where the var USE_SYSTEM
make check USE_SYSTEM=1
is used to indicated that the system installed tools should be used
and
make check
is used to run the tests against the in tree tools
MR: https://gitlab.com/apparmor/apparmor/-/merge_requests/580
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
(cherry picked from commit c8b6d8b393)
This is a backport of the fix-lexer patch series for 2.13
Atm it has more patches than are strictly necessary. The open question is how much of that series do we actually want to bring back?
technically the only patch we need to keep CI working and fix the issues that have been discovered are
```
Remove TODO for half-quoted abi rule
parser: split newline and end of rule handling into separate rules
parser: update rule to process newlines to include states that eat WS
parser: add ABI_MODE to WS consumption state
```
However that would keep us from being able to catch other errors. If we want to be able to catch other potential lexer pass-through errors, at a minimum we need to include.
```
Error out on unhandled parts when parsing a profile
```
Which excludes only
```
parser: add missing states to the default rule and improve the error msg
```
While its nice to have better debug output, I am not sure it is worth backporting this patch
Summary of Discussion: The debug output is worth having
MR: https://gitlab.com/apparmor/apparmor/-/merge_requests/572
Acked-by: Steve Beattie <sbeattie@ubuntu.com>
The hashing of the featue set is wrong because it is hashing the
whole feature structure instead of just the feature string.
This results in the refcount and hash field becoming part of the
hash and the feature string not being completely hashed as the
bytes of the refcount and hash field are being counted in the
as part of the string length when the hash is taken.
MR: https://gitlab.com/apparmor/apparmor/-/merge_requests/583
Reported-by: Samuele Pedroni <samuele.pedroni@canonical.com>
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Steve Beattie <steve.beattie@canonical.com>
(cherry picked from commit b8be1c3ff8)
[2.11..2.13] fix setting apparmor.aa.profile_dir in some tests
cherry picked from commit 6fe4b5e59a -
but only the test-aa.py changes because test-profile-list.py didn't have
the affected tests in 2.13 yet.
See merge request apparmor/apparmor!574
Acked-by: Steve Beattie <steve.beattie@canonical.com> for 2.11..2.13
ABI_MODE needs to be able to consume white space, unfortunately this
was missed, and only showed up with the basckport of the patches
to stop unmatched input from being passed through.
Signed-off-by: John Johansen <john.johansen@canonical.com>
Newlines should generally be treated as whitespace. Expand the list
of states using the newline rule to include almost all rules that
eat WS.
There are two exceptions assign and comment which have special handling
of newlines.
this fixes the failures
not ok 71543 - ./simple_tests//vars/vars_simple_assignment_13.sd: quoted commas should not trigger an error
not ok 71544 - ./simple_tests//vars/vars_simple_assignment_14.sd: quoted commas should not trigger an error
found by introducing nodefault
MR: https://gitlab.com/apparmor/apparmor/-/merge_requests/569
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Steve Beattie <sbeattie@ubuntu.com>
(cherry picked from commit 21498ff9a4)
There were several states missing from the default rule which catches
unexpected input in a state.
Update the default rule to catch all input including newlines and
update its error message to include information about which state the
failure occured in. Also update the comment about what to do when
adding new states.
While the lexer now has the "nodefault" option set, it doesn't provide
as much information as the default rule does, so we prefer states
to use our provided default rule.
MR: https://gitlab.com/apparmor/apparmor/-/merge_requests/569
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Steve Beattie <sbeattie@ubuntu.com>
(cherry picked from commit 1a4288886b)
... (using `%option nodefault`) instead of echoing the unknown parts to
stdout, and ignoring the error.
This will cause the parser to error out with
flex scanner jammed
and $?=2 if a profile contains unknown/invalid parts. That's not really
a helpful error message, but still better than ignoring errors.
MR: https://gitlab.com/apparmor/apparmor/-/merge_requests/569
Signed-off-by: Christian Boltz <apparmor@cboltz.de>
Acked-by: John Johansen <john.johansen@canonical.com>
Acked-by: Steve Beattie <sbeattie@ubuntu.com>
(cherry picked from commit c01ed1d57b)
Seen on openSUSE Tumbleweed with the mail users in a mysql database.
(cherry picked from commit f7ab91f423)
Signed-off-by: John Johansen <john.johansen@canonical.com>
This is needed when using the "apparmor" plugin which means dovecot
switches to user-specific hats.
Seen on openSUSE Tumbleweed.
(backported from commit 6a388859f8)
Signed-off-by: John Johansen <john.johansen@canonical.com>
When a filesystem is mounted with the option *user*, the file selection
dialogue, e.g. in *Evince*, triggers an access of */run/mount/utab*, which
comes from *libmount* and should be allowed.
Reported-by: JrgSommer[m] in #apparmor
(cherry picked from commit cd3532f792)
Signed-off-by: John Johansen <john.johansen@canonical.com>
strip_quotes() assumed its parameter is at least one character long, and
errored out on an empty string.
It also converted a string consisting of a single quote to an empty
string because that single quote had a quote as first and last char.
This commit fixes these two bugs.
Also rewrite TestStripQuotes to use tests[], and add some test for an empty
string, a one-char path (just a slash) and a single quote.
(cherry picked from commit 373e8e23b1)
Signed-off-by: John Johansen <john.johansen@canonical.com>
... or calling is_known_rule() on events for non-existing hats.
It's the usual hasher() "fun" again - accessing a non-existing element
will create its parent.
In theory this commit might be worth a backport. In practise, it doesn't cause
any visible problem.
However, starting with the next commit, it will cause lots of test errors.
Also add a missing is_known_rule() call for dbus rules, which might have
caused similar hasher() "fun".
(Backported from 9f1b2f4014)
apparmor.vim: allow leading whitespace for alias rules
See merge request apparmor/apparmor!527
Acked-by: Steve Beattie <steve.beattie@canonical.com> for 2.11..master
(cherry picked from commit ae70ecfbaa)
c636580f apparmor.vim: allow leading whitespace for alias rules
When aa-genprof proposes a local inactive profile, it had a hardcoded
call to 'less' to display that profile.
Unsurprisingly, this doesn't work in JSON mode and breaks YaST (luckily
it's only a case of "the button doesn't work").
References: https://bugzilla.opensuse.org/show_bug.cgi?id=1171315
(cherry picked from commit 68a258b006)
(cherry picked from commit cb95e9a2568b19e2e7601c0af363e0605a6889d9)
UI_ShowFile() is more generic and can be used to display various (text)
files, not only diffs.
(cherry picked from commit bb3803b931,
adjusted for 2.13 branch)
usr.sbin.dnsmasq: update to support dnsmasq 2.81
See merge request apparmor/apparmor!475
Acked-by: Christian Boltz for 2.11..master
(cherry picked from commit acafe9de82)
88c142c6 usr.sbin.dnsmasq: allow reading @{PROC}/@{pid}/fd/ as is needed by dnsmasq 2.81
@Talkless requested xdg-open and friends be cherry-picked into 2.13
This is the set of commits (and fixes) to do that without modifying them.
We could drop backporting dbus-strict by modifying both the adding missing .d dirs, and add xdg-open and friends patches.
This series does not currently include the make check test and its fixes for the .d directories, as they were not required but we may want to include them to catch any potential errors.
PR: https://gitlab.com/apparmor/apparmor/-/merge_requests/471
Signed-off-by: John Johansen <john.johansen@canonical.com>
Implement set of abstractions to handle opening uris via xdg-open and similar helpers used on different desktop environments.
Abstractions are intended to be included into child profile, together with bundle abstractions such as ubuntu-browsers, ubuntu-email and others, for fine-grained control on what confined application can actually open via xdg-open and similar helpers.
PR: https://gitlab.com/apparmor/apparmor/-/merge_requests/404
Acked-by: John Johansen <john.johansen@canonical.com>
(cherry picked from commit d257afd309)
Signed-off-by: John Johansen <john.johansen@canonical.com>
Some applications queries network configuration (using QNetworkConfigurationManager class in Qt and similar), and that produces DBus denials under AppArmor confinement when NetworkManager backend is used.
Add abstraction that allows most common read-only DBus queries for getting current network configuration from NetworkManager backend.
PR: https://gitlab.com/apparmor/apparmor/merge_requests/409
Acked-by: John Johansen <john.johansen@canonical.com>
(cherry picked from commit a10fa57fb6)
Signed-off-by: John Johansen <john.johansen@canonical.com>
In focal users of mdns get denials in apparmor confined applications.
An exampel can be found in the original bug below.
It seems it is a common pattern, see
https://github.com/lathiat/nss-mdns#etcmdnsallow
Therefore I'm asking to add
/etc/mdns.allow r,
to the file
/etc/apparmor.d/abstractions/mdns"
by default.
--- original bug ---
Many repetitions of
audit: type=1400 audit(1585517168.705:63): apparmor="DENIED" operation="open" profile="/usr/sbin/chronyd" name="/etc/mdns.allow" pid=1983815 comm="chronyd" requested_mask="r" denied_mask="r" fsuid=123 ouid=0
in log. I use libnss-mdns for .local name resolution, so /etc/nsswitch.conf contains
hosts: files mdns [NOTFOUND=return] myhostname dns
and /etc/mnds.allow contains the domains to resolve with mDNS (in may case, "local." and "local"; see /usr/share/doc/libnss-mdns/README.html.)
Presumably cronyd calls a gethostbyX() somewhere, thus eventually trickling down through the name service switch and opening /etc/mdns.allow, which the AppArmor profile in the chrony package does not allow.
Fixes: https://bugs.launchpad.net/ubuntu/+source/apparmor/+bug/1869629
Signed-off-by: John Johansen <john.johansen@canonical.com>
(cherry picked from commit eeac8c11c9)
abstractions/mesa: allow checking if the kernel supports the i915 perf interface
See merge request apparmor/apparmor!464
Acked-by: Vincas Dargis <vindrg@gmail.com>
Acked-by: Christian Boltz <apparmor@cboltz.de> for master and 2.13
(cherry picked from commit f56bab3f75)
61571da1 abstractions/mesa: allow checking if the kernel supports the i915 perf interface
This way we could generate the capabilities in a way that works with
every version of make.
Changes to list_capabilities are intended to exactly replicate the old
behavior.
(cherry picked from commit e92da079ca)
This change updates parser/Makefile to respect target dependencies and
not rebuild apparmor_parser if nothing's changed. The goal is to allow
cross-compiled tests #17 to run on a target system without the tests
attempting to rebuild the parser.
Two changes were made:
* Generate af_names.h in a script so the script timestamp is compared.
* Use FORCE instead of PHONY for libapparmor_re/libapparmor_re.a
Changes to list_af_names are intended to exactly replicate the old
behavior.
Signed-off-by: Eric Chiang <ericchiang@google.com>
(cherry picked from commit cb8c3377ba)
abstractions/base: allow read access to /run/uuidd/request
See merge request apparmor/apparmor!445
Acked-by: John Johansen <john.johansen@canonical.com> for 2.11..master
Acked-by: Christian Boltz <apparmor@cboltz.de> for 2.11..master
(cherry picked from commit 80bf920929)
45fffc12 abstractions/base: allow read access to /run/uuidd/request
abstractions/base: allow read access to top-level ecryptfs directories
See merge request apparmor/apparmor!443
Acked-by: John Johansen <john.johansen@canonical.com> for 2.11..master
Acked-by: Christian Boltz <apparmor@cboltz.de> for 2.11..master
(cherry picked from commit 24895ea302)
fbd8981e abstractions/base: allow read access to top-level ecryptfs directories
Allow /usr/etc/ in abstractions/authentication
openSUSE (and hopefully some other distributions) work on moving shipped
config files from /etc/ to /usr/etc/ so that /etc/ only contains files
written by the admin of each system.
See https://en.opensuse.org/openSUSE:Packaging_UsrEtc for details and
the first moved files.
Updating abstractions/authentication is the first step, and also fixes
bugzilla.opensuse.org/show_bug.cgi?id=1153162
See merge request apparmor/apparmor!426
Acked-by: John Johansen <john.johansen@canonical.com> for 2.12..master
(cherry picked from commit 1cfd4d4bbc)
ee7194a7 Allow /usr/etc/ in abstractions/authentication
abstractions/kerberosclient: allow reading /etc/krb5.conf.d/
See merge request apparmor/apparmor!425
Acked-by: Steve Beattie <steve@nxnw.org> for 2.10..master
Acked-by: John Johansen <john.johansen@canonical.com> for 2.10..master
(cherry picked from commit 663546c284)
dffed831 abstractions/kerberosclient: allow reading /etc/krb5.conf.d/
Drop 'localinclude' in parse_profile_data() and ProfileStorage
See merge request apparmor/apparmor!427
Acked-by: John Johansen <john.johansen@canonical.com> for 2.12..master
Acked-by: Steve Beattie <steve@nxnw.org> for 2.12..master
(cherry picked from commit b017f8f8a9)
001ea9e3 Drop 'localinclude' in parse_profile_data() and ProfileStorage
879531b36ec3dfc7f9b72475c68c30e4f4b7b6af changed access for
@{HOME}/.{,cache/}fontconfig/** to include 'w'rite. Fontconfig has been
a source of CVEs. Confined applications should absolutely have read
access, but write access could lead to breaking out of the sandbox if a
confined application can write a malformed font cache file since
unconfined applications could then pick them up and be controlled via
the malformed cache. The breakout is dependent on the fontconfig
vulnerability, but this is the sort of thing AppArmor is meant to help
guard against.
(cherry picked from commit c5968c70d0)
PR: https://gitlab.com/apparmor/apparmor/merge_requests/420
Signed-off-by: John Johansen <john.johansen@canonical.com>
v2:
- parse partial log line broken at \n
- add testcase_dbus_10.* for partial log line
- remove quotes from testcasw_dbus_09.profile
The following log format has been seen in the wild, and currently results
in a RECORD_INVALID
[4835959.046111] audit: type=1107 audit(1561053426.749:186): pid=640 uid=103 auid=4294967295 ses=4294967295 msg='apparmor="ALLOWED" operation="dbus_method_call" bus="system" path="/org/freedesktop/systemd1" interface="org.freedesktop.systemd1.Manager" member="LookupDynamicUserByName" mask="send" name="org.freedesktop.systemd1" pid=20596 label="/usr/sbin/sshd" peer_pid=1 peer_label="unconfined"
exe="/usr/bin/dbus-daemon" sauid=103 hostname=? addr=? terminal=?'
Test parsing the above message with and without the \n embedded between
peer_label= and exec=
PR: https://gitlab.com/apparmor/apparmor/merge_requests/395
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Signed-off-by: John Johansen <john.johansen@canonical.com>
(cherry picked from commit 0349cf2d0a)
The default path is /etc/certbot/archive/{some domain}/{file name}.pem
See merge request apparmor/apparmor!397
This is a manual cherry-pick of 4d275bab69
and 3016ffb336
The mult_mount test creates a small disk image, formats it, and mounts
it in multiple locations in preparation for the tests. However, the
created raw file (80KB) is too small to make a working file system if
4K blocks are used by mkfs. In Ubuntu 19.10, the default was recently
changed for mkfs to default to always using 4K blocks, causing the
script to fail.
We could force mkfs to use 1K blocks, but instead, in case some future
version of mkfs decides not to support 1K blocks at all, we bump up the
size of the disk image to 512KB; large enough to work with 4K blocks
yet small enough to be workable in small scale test environments.
Signed-off-by: Steve Beattie <steve.beattie@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Bug: https://bugs.launchpad.net/ubuntu/+source/apparmor/+bug/1834192
MR: https://gitlab.com/apparmor/apparmor/merge_requests/396
(cherry picked from commit 7c7a4bc531)
Signed-off-by: Steve Beattie <steve.beattie@canonical.com>
This is especially handy if your distro doesn't split sbin and bin
and only wants to install into bin (so that the sbin directory doesn't
clash with the sbin -> bin symlink)
[Per feedback, added USR_SBINDIR as a toggle for the install location
of aa-teardown -- @smb]
Signed-off-by: Rasmus Thomsen <cogitri@exherbo.org>
Signed-off-by: Steve Beattie <steve.beattie@canonical.com>
PR: https://gitlab.com/apparmor/apparmor/merge_requests/111/
(cherry picked from commit 7c86a2acaf)
Found this path is used by gtk_compose_hash_get_cache_path() in
gtkcomposetable.c.
(cherry picked from commit 6da7ed2a78)
Signed-off-by: John Johansen <john.johansen@canonical.com>
Add file rule to allow reading application profiles for NVIDIA
Linux graphics driver.
(cherry picked from commit f2e0fdc72b)
Signed-off-by: John Johansen <john.johansen@canonical.com>
When dc010bc034 was
backported to the apparmor-2.13 branch (in commit
75236d62e2), it did not take into
account cb8c3377ba, which creates the
common/list_af_names.sh script as used in the test case, was not also
backported to the apparmor-2.13 branch.
Change the test case to get the list of network AF names via the same
make invocation taken by the utils/vim/create-apparmor.vim.py script
before the common/list_af_names.sh existed.
PR: https://gitlab.com/apparmor/apparmor/merge_requests/391
Signed-off-by: Steve Beattie <steve.beattie@canonical.com>
Signed-off-by: John Johansen <john.johansen@canonical.com>
Omnibus collection of translations updates.
Signed-off-by: Steve Beattie <steve.beattie@canonical.com>
(cherry picked from merge commit 3ee468864d
plus following translation fix.)
The translated action character for Deny conflicted with the
untranslated action character for Finish in the Swedish translation.
Remote it, and hope for more action translations.
Signed-off-by: Steve Beattie <steve.beattie@canonical.com>
(cherry picked from commit 03c08cf989)
Signed-off-by: Steve Beattie <steve.beattie@canonical.com>
The parser currently skips the cache if optimizations are specified
because it can not determine if the cached policy was compiled
with the specified optimization. However this causes cache misses
even if policy is cached with those options, and distros are setting
some optimizations by default.
Instead of skipping reading the cache if optimizations are set, users
can force overwriting the cache if needed, until the parser can
store aditional meta info in the cache.
PR: https://gitlab.com/apparmor/apparmor/merge_requests/385
BugLink: http://bugs.launchpad.net/bugs/1820068
Signed-off-by: John Johansen <john.johansen@canonical.com>
(cherry picked from commit f6cd5c01c1)
This is a partial backport of bc36daa264
(without the abstractions/nameservice removal in
usr.lib.dovecot.pop3-login)
Original commmit message:
dovecot: align {pop3,managesieve}-login to imap-login
Those 3 login daemons should have similiar needs and thus similar
profiles. IMAP is likely the most tested one so let's align the
other 2 with it. Unix and TCP sockets rules were added to pop3-login
after the removal of abstractions/nameservice that included them
implicitly.
Signed-off-by: Simon Deziel <simon@sdeziel.info>
Even if we don't backport the latest postfix profiles (in extras) to
2.12 and 2.13, making the abstraction compatible with them (by adding
peer=postfix-master rules in addition to the path-based ones) makes
things much easier for people who want to use the latest profiles.
looping through the first 16 loop devices to find a free device will
fail if those mount devices are taken, and unfortunately there are
now services that use an excessive amount of loop devices causing
the regression test to fail.
PR: https://gitlab.com/apparmor/apparmor/merge_requests/379
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
(cherry picked from commit ab0f2af1da)
Choosing "Deny" when handle_children() asks about adding a hat should
not add that hat, but it should _not_ leave the function because that
means all the other log events that were not migrated to prelog yet will
be lost.
Change "return" to "continue" to fix this.
Add several libapparmor/swig/ruby files to gitignore
See merge request apparmor/apparmor!366
(cherry picked from commit 9c11ce37c6)
7ed1a16a Add several libapparmor/swig/ruby files to gitignore
Fix error 'KeyError: 'logfiles'' when no logprof.conf exists
See merge request apparmor/apparmor!365
Acked-by: Christian Boltz <apparmor@cboltz.de> for 2.12..master
(cherry picked from commit cece787182)
455c4413 aa.py: Ensure there is always a fallback falue for the logfile location
3c7e1668 aa.py: Indicate permission error if log file is found but cannot be opened
Since !345 the set of permissions that are granted (get_file_perms_2)
or suggested (propose_file_rules) has changed. These new sets are
expected due to the changes brought by this MR, so let's adjust
the test suite accordingly.
(cherry picked from commit 0170e98f9c)
PR: https://gitlab.com/apparmor/apparmor/merge_requests/358
Signed-off-by: John Johansen <john.johansen@canonical.com>
For example, VirtualBox guests have /usr/lib/VBoxOGL.so.
Without this changes, in a VirtualBox VM with VBoxVGA graphics,
at least one Qt5 application (OnionShare) won't start and display:
ImportError: libGL.so.1: failed to map segment from shared object
… and the system logs have:
apparmor="DENIED" operation="file_mmap" profile="/usr/bin/onionshare-gui" name="/usr/lib/VBoxOGL.so" pid=11415 comm="onionshare-gui" requested_mask="m" denied_mask="m" fsuid=1000 ouid=0
While this works fine with VBoxSVGA and VMSVGA when 3D acceleration is enabled.
So let's not assume all libraries have a name that starts with "lib".
PR: https://gitlab.com/apparmor/apparmor/merge_requests/345
(cherry picked from commit 5cbb7df95e)
Signed-off-by: John Johansen <john.johansen@canonical.com>
update network keyword list in utils and add test
See merge request apparmor/apparmor!350
Acked-by: Eric Chiang <ericchiang@google.com> for 2.12..master
(cherry picked from commit dc010bc034)
49849ed7 update network keyword list in utils and add test
apparmor.d manpage: update list of network domain keywords
See merge request apparmor/apparmor!349
Acked-by: Eric Chiang <ericchiang@google.com> for 2.12..master
(cherry picked from commit 6416ccebf6)
6b276563 apparmor.d manpage: update list of network domain keywords
The calling function and the merge() macro both make use of a variable
named "i" but this causes problems when merging. The aa_policy_cache.sh
test script experiences "bad file descriptor" failures due to the merged
list containing invalid fd values (-1).
Fix this by renaming merge()'s index variables from i and j to y and z
to hopefully prevent future overlap. The better fix here would be to
convert merge() to a function but that's a more intrusive change and I
don't have an easy way to test the overlay feature.
PR: https://gitlab.com/apparmor/apparmor/merge_requests/348
(cherry picked from commit d7ac07afc4)
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Signed-off-by: John Johansen <john.johansen@canonical.com>
qsort()'s _size_ parameter is used to indicate the size of the elements
in the _base_ array parameter. Adjust the third argument to qsort() to
indicate that we're dealing with an array of struct dirent pointers
rather than an array of struct dirent.
PR: https://gitlab.com/apparmor/apparmor/merge_requests/348
(cherry picked from commit 8b21871820)
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Signed-off-by: John Johansen <john.johansen@canonical.com>
Adjust aa_policy_cache.sh to handle the additional layer in the
directory hierarchy when determining where the policy cache binaries are
stored. This is needed due to the multicache changes that allow multiple
policy caches to exist on a single system.
Differentiate between the cache location (the top level directory
containing all caches) and the cache directory (the directory used to
store the cached policies).
Use the libapparmor wrapper to get the cache directory for the given
cache location and the features of the currently running kernel.
PR: https://gitlab.com/apparmor/apparmor/merge_requests/348
(cherry picked from commit f31457b26e)
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Signed-off-by: John Johansen <john.johansen@canonical.com>
The length of a xmatch is used to prioritize multiple profiles that
match the same path, with the intent that the more specific match wins.
Currently, the length of a xmatch is computed by the position of the
first regex character.
While trying to work around issues with no_new_privs by combining
profiles, we noticed that the xmatch length computation doesn't work as
expected for multiple regexs. Consider the following two profiles:
profile all /** { }
profile bins /{,usr/,usr/local/}bin/** { }
xmatch_len is currently computed as "1" for both profiles, even though
"bins" is clearly more specific.
When determining the length of a regex, compute the smallest possible
match and use that for xmatch priority instead of the position of the
first regex character.
(cherry picked from commit cc09794fbd)
Signed-off-by: John Johansen <john.johansen@canonical.com>
Using /usr/{bin,sbin}/dbsmasq as both the profile name and attachment break compatibility with the peer=/usr/sbin/dnsmasq rule, in the libvirtd profile.
Instead specify the profile has the old name of /usr/sbin/dnsmasq and specify the attachment separately. This looks funny but it avoids breaking libvirtd and still provides the broader attachment need by distros that have merged sbin into bin.
References: https://bugzilla.opensuse.org/show_bug.cgi?id=1127073
I propose this patch for 2.12 and 2.13. (Older versions didn't get the alternation, master has a profile name added.)
PR: https://gitlab.com/apparmor/apparmor/merge_requests/346
Signed-off-by: John Johansen <john.johansen@canonical.com>
Even if we expected it to stay compatible with peer=/usr/sbin/dnsmasq in
the libvirtd profile, practise shows that we were wrong.
This patch adds a tricky workaround so that the profile name doesn't
change, but the profile will attach even on distributions with merged
bin and sbin. (Credits for this crazy idea go to John ;-)
It also re-adds the libvirtd peer name /usr/sbin/libvirtd to avoid
breaking libvirtd.
References: https://bugzilla.opensuse.org/show_bug.cgi?id=1127073
and the discussion in https://gitlab.com/apparmor/apparmor/merge_requests/346
tunables/share: fix buggy syntax that broke the ~/.local/share part of the @{user_share_dirs} tunable
See merge request apparmor/apparmor!344
(cherry picked from commit 2ed3763a2f)
cdeb6185 tunables/share: fix buggy syntax that broke the ~/.local/share part of the...
misc dovecot fixes (take #2)
See merge request apparmor/apparmor!336
Acked-by: Christian Boltz <apparmor@cboltz.de> for master..2.10
(cherry picked from commit e68beb988a)
a57f01d8 dovecot: allow FD passing between dovecot and dovecot's anvil
d0aa863f dovecot: allow chroot'ing the auth processes
9afeb225 dovecot: let dovecot/anvil rw the auth-penalty socket
17db8f38 dovecot: auth processes need to read from postfix auth socket
6a7c49b1 dovecot: add abstractions/ssl_certs to lmtp
Debian and Ubuntu have releases coming out with usr-merge in place. For
these systems, /bin and /sbin are symlinks to their respective /usr
directories. This breaks a few tests in the python utils and in the
regression tests. This patch series fixes them, mostly by performing
realpath() calls when necessary. For the ptrace regression test,
it copies the called /bin/true binary into the created temporary
directory and executes it from there. (Good for other reasons, too.)
(cherry picked from commit b4ab8476e4)
Signed-off-by: Steve Beattie <steve.beattie@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
MR: https://gitlab.com/apparmor/apparmor/merge_requests/331
aa-logprof errors out if it hits a log event for a non-existing profile
while a profile file with the default filename for that non-existing
profile exists. This can for example happen after adding a profile name
to a profile if audit.log still contains events for the attachment-based
profile name.
Since we ignore log events for non-existing profiles in general, drop
the code for the special case "but a file matching the default filename
for that non-existing profile exists" and also silently ignore events
for this very special non-existing profile.
References: https://bugzilla.opensuse.org/show_bug.cgi?id=1120472
(cherry picked from commit 03ea5b82b7)
Signed-off-by: John Johansen <john.johansen@canonical.com>
This updates the mysqld to what I use on my servers nowadays.
Note: my profile also has capability sys_resource,, but I'm not sure
why I had to add this and therefore didn't include it in this merge
request.
Speaking about "why I had added $whatever" - these changes were
collected over the last years and of course ;-) I don't remember any
details.
PR: https://gitlab.com/apparmor/apparmor/merge_requests/310
(cherry picked from commit 0199edf8e7)
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Signed-off-by: John Johansen <john.johansen@canonical.com>
[2.10..2.13] Backport updated font paths
Update fonts for Debian and openSUSE
- Allow to read conf-avail dir itself.
- Add various openSUSE-specific font config directories.
See merge request !96 (merged) for details.
PR: https://gitlab.com/apparmor/apparmor/merge_requests/309
(cherry picked from commit 7bd3029f)
Update fonts for Debian and openSUSE
See merge request apparmor/apparmor!96
(cherry picked from commit 7bd3029f25)
b902d250 Update fonts for Debian and openSUSE
dnsmasq: allow peer=libvirtd to support named profile
See merge request apparmor/apparmor!304
Acked-by: Eric Chiang <ericchiang@google.com> for 2.12..master
(cherry picked from commit 5d384d9625)
20fe099c dnsmasq: allow peer=libvirtd to support named profile
New abstraction: lsb_release (sub-profile).
See merge request apparmor/apparmor!154
(cherry picked from commit 47e38944f3)
f7351405 New profile: lsb_release (no attachment path)
99e45b59 lsb_release: added permissions needed by openSUSE implementation.
parser/apparmor.systemd: fix minor issues detected by shellcheck
See merge request apparmor/apparmor!293
Acked-by: Christian Boltz <apparmor@cboltz.de> for master and 2.13
(cherry picked from commit a772ee0f8b)
b3937d19 parser/apparmor.systemd: fix minor issues detected by shellcheck
abstractions/ssl_{certs,keys}: dehydrated uses /var/lib on Debian
See merge request apparmor/apparmor!299
Acked-by: Christian Boltz <apparmor@cboltz.de> for 2.10..master
(cherry picked from commit 1f53de174d)
1306f9a6 abstractions/ssl_{certs,keys}: dehydrated uses /var/lib on Debian
c5a89d5d abstractions/ssl_{certs,keys}: sort the alternation for dehydrated and drop...
04b2842e abstractions/ssl_{certs,keys}: allow reading ocsp.der maintained by dehydrated for OCSP stapling
[2.12+2.13] revert naming the dnsmasq profile
Changing to "profile dnsmasq /..." broke the peer=/usr/sbin/dnsmasq in the libvirtd profile. Revert adding the name to avoid breaking the libvirtd profile in stable branches.
See also https://bugzilla.opensuse.org/show_bug.cgi?id=1118952 which is a request to update the libvirtd profile to allow both peer=dnsmasq and peer=/usr/sbin/dnsmasq
I propose this revert for 2.12 and 2.13 (older branches didn't get the named profile)
PR: https://gitlab.com/apparmor/apparmor/merge_requests/290
Acked-by: John Johansen <john.johansen@canonical.com>
Changing to "profile dnsmasq /..." broke the peer=/usr/sbin/dnsmasq in
the libvirtd profile. Revert adding the name to avoid breaking the
libvirtd profile in stable branches.
See also https://bugzilla.opensuse.org/show_bug.cgi?id=1118952
which is a request to update the libvirtd profile to allow both
peer=dnsmasq and peer=/usr/sbin/dnsmasq
[2.12+2.13] Replace "existing_profiles" & fix minitools for named profiles
(This is the 2.13 version of !249 (merged) which had a few merge conflicts in the 2.13 branch, and needs a little change (last commit) on top)
This patchset introduces the ProfileList class which replaces "existing_profiles" in aa.py and fixes some bugs in aa-complain and the other minitools:
* aa-complain etc. never found profiles that have a profile name (the attachment wasn't checked)
* even if the profile name was given as parameter to aa-complain, it first did "which $parameter" so it never matched on named profiles
* profile names with alternations (without attachment specification) also never matched because the old code didn't use AARE.
References: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=882047#92 (search for "As usual" ;-)
See the individual commit messages for details.
All changes survived my tests (both manually and unittests), but as always when doing bigger changes to aa.py, more manual testing is always welcome ;-)
I propose this patch for 2.12 and 2.13.
Acked-by: John Johansen <john.johansen@canonical.com>
PR: https://gitlab.com/apparmor/apparmor/merge_requests/268
i.e. move '*' from beginning to before suffix.
Commit 025c7dc6 ("dnsmasq: Add permission to open log files") added
pattern, which is not compatible with SELinux. As this pattern has been
in SELinux since 2011 (with recent change to accept '.log' suffix +
logrotate patterns which are not relevant to AppArmor) IMHO it's better
to adjust our profile.
Fixes: 025c7dc6 ("dnsmasq: Add permission to open log files")
PR: PR: https://gitlab.com/apparmor/apparmor/merge_requests/288
Signed-off-by: Petr Vorel <pvorel@suse.cz>
(cherry picked from commit 3ef8df6ac0)
Signed-off-by: John Johansen <john.johansen@canonical.com>
Add /etc/letsencrypt/archive to ssl_key abstraction
See merge request apparmor/apparmor!283
Acked-by: Christian Boltz <apparmor@cboltz.de> for 2.10..master
(cherry picked from commit 0a666b8e48)
cb468786 Add /etc/letsencrypt stuff to ssl_keys/ssl_certs abstraction
aa-genprof checks if one of the profiles in the extra profile dir
matches the binary, and proposes to use that profile as a starting
point.
Since 4d722f1839 the "(V)iew profile"
option to display the proposed profile was broken.
The easiest fix is to remember the filename in the extras directory, and
display the file from there.
Sidenote: when choosing to use the extra profile, it gets written to
disk without any problems, so this bug really only affected "(V)iew
profile" to preview the proposed extra profile.
(cherry picked from commit 8b4e76a7d5)
'lastline' gets merged into 'line' (and reset to None) when reading the
next line. If 'lastline' isn't empty after reading the whole profile,
this means there's something unparseable at the end of the profile,
therefore parse_profile_data() should error out.
Also remove some simple_tests testcases from the 'exception_not_raised'
list - they only didn't raise the exception because the invalid rule was
the last line in the affected profile.
Thanks to Eric Chiang for accidently (and maybe even unnoticedly ;-)
discovering this bug while adding some xattr testcases that surprisingly
didn't fail in the tools.
PR: https://gitlab.com/apparmor/apparmor/merge_requests/271
(cherry picked from commit 4efff35bf8)
Signed-off-by: John Johansen <john.johansen@canonical.com>
In the 2.13 branch (and older), 'options' is not always a dict, but can
also be None or an empty string.
Adjust the if condition in serialize_profile() so that "View changes
between clean profiles" doesn't error out.
Technical stuff first:
Replace existing_profiles (a dict with the filenames for both active and
inactive profiles) with active_profiles and extra_profiles which are
ProfileList()s and store the active profiles and those in the extra
directory separately. Thanks to ProfileList, now also the relation
between attachments and filenames is easily available.
Also replace all usage of existing_profiles with active_profiles and
extra_profiles, and adjust it to the ProfileList syntax everywhere.
With this change, several bugs in aa-complain and the other minitools
get fixed:
- aa-complain etc. never found profiles that have a profile name
(the attachment wasn't checked)
- even if the profile name was given as parameter to aa-complain, it
first did "which $parameter" so it never matched on named profiles
- profile names with alternations (without attachment specification)
also never matched because the old code didn't use AARE.
References: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=882047#92
(search for "As usual" ;-)
Just for completeness - the matching still doesn't honor/expand
variables in the profile name.
(cherry picked from commit 4d722f1839)
ProfileList is meant to store the list of profiles (both name and
attachment) and in which files they live.
Also add unittests to make sure everything works as expected.
(cherry picked from commit 789c4658e2)
parse_profile_data() returns the parsed profiles, but writes to
existing_profiles directly.
read_profiles() calls parse_profile_data() and already handles adding
the parsed profiles to aa, original_aa or extras, which means updating
existing_profiles there is a much better place.
This commit also includes a hidden change: Previously, when parsing
include files, they were also added to existing_profiles. This is
superfluous, only real profiles need to be stored there.
(cherry picked from commit 8809218ac8)
Split get_profile_filename() into
- get_profile_filename_from_profile_name() (parameter: a profile name)
- get_profile_filename_from_attachment() (parameter: an attachment)
Currently both functions call get_profile_filename_orig() (formerly
get_profile_filename()) so the behaviour doesn't change yet.
The most important part of this commit is changing all
get_profile_filename() calls to use one of the new functions to make
clear if they specify a profile or an attachment/executable as
parameter.
As promised, the is_attachment parameter starts to get used in this
patch ;-)
Note: The get_new parameter (which I'll explain in the patch actually
using it) is set to True in all calls to the new functions.
The long term plan is to get rid of it in most cases (hence defaulting
to False), but that will need more testing.
(cherry picked from commit ec741424f8)
The minitools call write_profile(), write_profile_feedback_ui() and
serialize_profile() with the _attachment_ as parameter.
However, aa-logprof etc. call them with the _profile name_ as parameter.
This patch adds an is_attachment parameter to write_profile() and
write_profile_feedback_ui(). It also passes it through to
serialize_profile() via the options parameter.
If is_attachment is True, the parameter will be handled as attachment,
otherwise it is expected to be a profile name.
tools.py gets changed to set is_attachment to True when calling the
functions listed above to make clear that the parameter is an attachment.
Note: This patch only adds the is_attachment parameter/option, but
doesn't change any behaviour. That will happen in the next patch.
(cherry picked from commit bc783372b8)
... which is unused since the last commit.
Note: unlike 0eb12a8cbd, this commit does
_not_ delete several write_* function that were only used by this
function. Verifying that these functions are really unused is not worth
the effort in the 2.13 branch.
(cherry picked from commit 0eb12a8cbd -
but only apply partially)
Commit aa06528790 made @{sys} tunable
available by default.
Update profiles and abstractions to actually use @{sys} tunable for
better confinement in the future (when @{sys} becomes kernel var).
Closes LP#1728551
disable abi/ok_10 and abi/ok_12 tests
See merge request apparmor/apparmor!259
(cherry picked from commit 608af94dff)
a3305b51 disable abi/ok_10 and abi/ok_12 tests
Split the features file into compile features and kernel features
which is needed for policy versioning and the new caching scheme.
A new flag --kernel-features was added to set the kernel features but
unfortunately -M, --features-file was setup to only specify the
compile features, when it used to effectively specify both the
compile and kernel features.
This broke existing uses of -M.
Fix this by having -M specify both the compile and kernel features,
and a new flag --compile-features that can be used to specify the
compile fature set separate from the kernel feature set.
sbeattie> fixed up error message to refer to compile features when
--compile-features argument fails.
Backport-requested-by: intrigeri <intrigeri@debian.org>
PR: https://gitlab.com/apparmor/apparmor/merge_requests/104
(cherry picked from commit e83fa67edf)
Fixes: 9e48a5da5e ("parser: split kernel features from compile features.")
Signed-off-by: John Johansen <john.johansen@canonical.com>
Signed-off-by: Steve Beattie <steve.beattie@canonical.com>
Add profile names to all profiles with {bin,sbin} attachment
See merge request apparmor/apparmor!242
Acked-by: intrigeri <intrigeri@debian.org>
(cherry picked from commit fd68a5eb64)
b77116e6 Add profile names to all profiles with {bin,sbin} attachment
profiles/Makefile: test abstractions against apparmor_parser
See merge request apparmor/apparmor!237
Acked-by: Christian Boltz <apparmor@cboltz.de> for trunk and 2.13.
Pre-acked for 2.10..2.12 after removing the --config-file option which is not supported in these branches.
(cherry picked from commit 2863e20f37)
dc7ae28d profiles/Makefile: test abstractions against apparmor_parser
AppArmor 3.0 requires policy to use a feature abi rule for access to
new features. However some policy may start using abi rules even if
they don't have rules that require new features. This is especially
true for out of tree policy being shipped in other packages.
Add enough support to older releases that the parser will ignore the
abi rule and warn that it is falling back to the apparmor 2.x
technique of using the system abi.
If the profile contains rules that the older parser does not
understand it will fail policy compilation at the unknown rule instead
of the abi rule.
PR: https://gitlab.com/apparmor/apparmor/merge_requests/196
(backported form commit 83df7c4747)
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Fix aa-mergeprof crash caused by accidentially initialzed hat
See merge request apparmor/apparmor!234
Acked-by: John Johansen <john.johansen@canonical.com>
(cherry picked from commit 93445ca02d)
bc492533 Fix aa-mergeprof crash caused by accidentially initialzed hat
commit 94dfe15b28 attempted to remove
LD_RUN_PATH unfortunately
But all it actually does is cause the Makefile.perl to embed the rpath
"" instead. Which is still an rpath, only I guess an even worse one.
--
Eli Schwartz
Arch Linux Bug Wrangler and Trusted User
This is because it cleared the setting of the variable LD_RUN_PATH
which was expanded in the command
$(INST_DYNAMIC) : $(OBJECT) $(MYEXTLIB) $(INST_ARCHAUTODIR)$(DFSEP).exists $(EXPORT_LIST) $(PERL_ARCHIVEDEP) $(PERL_ARCHIVE_AFTER) $(INST_DYNAMIC_DEP)
$(RM_F) $@
LD_RUN_PATH="$(LD_RUN_PATH)" $(LD) $(LDDLFLAGS) $(LDFROM) $(OTHERLDFLAGS) -o $@ $(MYEXTLIB) \
$(PERL_ARCHIVE) $(LDLOADLIBS) $(PERL_ARCHIVE_AFTER) $(EXPORT_LIST) \
$(INST_DYNAMIC_FIX)
$(CHMOD) $(PERM_RWX) $@
resulting in LD_RUN_PATH="" being passed to the command.
Finish removing LD_RUN_PATH from Makefile.perl by removing it from
the command invocation if it is present.
Note: we use \x24 instead of $ in the regex as there seems to be a bug
and no level of escaping $ would allow it to be used.
PR: https://gitlab.com/apparmor/apparmor/merge_requests/207
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
(cherry picked from commit 958cc28876)
--log-facility option needs to have permission to open files.
Use '*' to allow using more files (for using more dnsmasq instances).
Signed-off-by: Petr Vorel <pvorel@suse.cz>
Signed-off-by: Jamie Strandboge <jamie@canonical.com>
(cherry picked from commit 025c7dc6a1)
Signed-off-by: Steve Beattie <steve.beattie@canonical.com>
parser: fix Makefile hardcoded paths to flex and bison
Closes#4
See merge request apparmor/apparmor!224
Acked-by: Christian Boltz <apparmor@cboltz.de> for 2.10..master
(cherry picked from commit 34cf085036)
17e059a2 parser: fix Makefile hardcoded paths to flex and bison
Commit 8f9bd5b0e3 rightfully removed PUx
transition into nvidia-modprobe executable due to security concerns. To
overcome this, commit 327420b151 added
named nvidia_modprobe profile, which allows to use this abstraction
without requiring additional rules to make OpenCL work with NVIDIA
drivers.
Add rule to allow Px transition into nvidia_modprobe profile for
nvidia-modprobe executable.
https://gitlab.com/apparmor/apparmor/merge_requests/219
(cherry picked from commit e4b1cadf63)
Signed-off-by: John Johansen <john.johansen@canonical.com>
Add missing paths to usr.sbin.nmbd, usr.sbin.smbd and abstractions/samba
See merge request apparmor/apparmor!210
Acked-by: Christian Boltz <apparmor@cboltz.de> for 2.10..master
(cherry picked from commit f76a718f28)
80e98f2d Update usr.sbin.nmbd & usr.sbin.smbd
2.13: Add basic support for abi rules to the tools
Add basic "understand and keep" support for abi rules, where
"understand" means to not error out when seeing an abi rule, and "keep"
simply means to keep the original abi rule when serializing a profile.
On the long term, abi rules should be parsed (similar to include rules),
but for now, this patch is the smallest possible changeset and easy to
backport.
Note that the only added test is via cleanprof_test.* which is used by
minitools_test.py - and does not run if you do a 'make check'.
Oh, and of course the simple_tests/abi/ files also get parsed by
test-parser-simple-tests.py.
BTW: Even serialize_profile_from_old_profile() can handle abi rules :-)
This is a backport of 072d3e04 / !202 (merged) to
2.13 (with some adjustments because that commit didn't appy cleanly)
I propose this patch for 2.10..2.13
PR: https://gitlab.com/apparmor/apparmor/merge_requests/216
Acked-by: John Johansen <john.johansen@canonical.com>
Add basic "understand and keep" support for abi rules, where
"understand" means to not error out when seeing an abi rule, and "keep"
simply means to keep the original abi rule when serializing a profile.
On the long term, abi rules should be parsed (similar to include rules),
but for now, this patch is the smallest possible changeset and easy to
backport.
Note that the only added test is via cleanprof_test.* which is used by
minitools_test.py - and does _not_ run if you do a 'make check'.
Oh, and of course the simple_tests/abi/ files also get parsed by
test-parser-simple-tests.py.
BTW: Even serialize_profile_from_old_profile() can handle abi rules :-)
This is a backport of 072d3e0451 / !202 to
2.13 (with some adjustments because that commit didn't appy cleanly)
Qt GUI applications that uses "platforminputcontexts"-class of plugins
might need reading and/or writing compose cache. Add read-only rule in
qt5 abstraction and create new writing dedicated for compose cache
writing.
PR: https://gitlab.com/apparmor/apparmor/merge_requests/159
(cherry picked from commit 67816c42cf)
Signed-off-by: John Johansen <john.johansen@canonical.com>
Qt-based applications stores QFileDialog (latest browsed directory) and
other shared user settings inside ~/.config/QtProject.conf. Currently
available qt abstraction only allows to read it (by design), so this
patch introduces abstraction that grants permissions for writing.
PR: https://gitlab.com/apparmor/apparmor/merge_requests/159
(cherry picked from commit 69c4cabb93)
Signed-off-by: John Johansen <john.johansen@canonical.com>
The function was messing up its use of fds, it could get away with
it because the cb_dirfd passed to fdopendir was still valid until
closedir was called but if code was moved around, or fdopendir
code changed behavior it could easily break.
Also the check for dup failing was wrong fix it.
Reference: coverity #187003
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Time-out
(cherry picked from commit 63cb46d20a)
Harden abstractions part ii
- abstractions/private-files: disallow access to the dirs of private files
- private-files{,-strict}: disallow writes to parent dirs too
- user-files: disallow writes to parents dirs
PR: https://gitlab.com/apparmor/apparmor/merge_requests/206
Signed-off-by: John Johansen <john.johansen@canonical.com>
Harden abstractions
Harden abstractions
remove antiquated abstractions/launchpad-integration
abstractions/opencl-nvidia: don't allow PUx on nvidia-modprobe
abstractions/private-files-strict: disallow access to the dirs of private files
abstractions/private-files: disallow writes to thumbnailer dir (LP: #1788929)
ubuntu-browsers.d/user-files: disallow access to the dirs of private files
Nominating launchpad-integration and opencl-nvidia for 2.13. Nominating private-files-strict, private-files and user-files for 2.10 and higher
PR: https://gitlab.com/apparmor/apparmor/merge_requests/203
Signed-off-by: John Johansen <john.johansen@canonical.com>
Also add /usr/share/dnsmasq/, DNSSEC trust anchors are kept there.
(cherry picked from commit 5bc7a9fbd6)
Signed-off-by: John Johansen <john.johansen@canonical.com>
The compiler is spitting out the warning
parser_main.c:1291:16: warning: ISO C++ forbids converting a string constant to 'char*' [-Wwrite-strings]
char *tmp = "/var/cache/apparmor";
fix this by constifying the cacheloc array.
Signed-off-by: John Johansen <john.johansen@canonical.com>
(cherry picked from commit e7949d09fa)
Signed-off-by: John Johansen <john.johansen@canonical.com>
Let's not store a bunch of automatically generated binary files in /etc.
AppArmor 3.0 will store the cache in /var/cache and most distros
(openSUSE, Debian, and soon Ubuntu) moved it there already.
Bug-Debian: https://bugs.debian.org/904637
(cherry picked from commit 3d21cf0e32)
Signed-off-by: John Johansen <john.johansen@canonical.com>
Requiring --config-file to be first in the option list is not user
friendly fix the option parsing so that --config-file can be specified
anywhere in the option list.
This also fixes a bug where even when the --config-file option is
first the option parsing fails because the detection logic is broken
for some option cases.
PR: https://gitlab.com/apparmor/apparmor/merge_requests/175
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Signed-off-by: John Johansen <john.johansen@canonical.com>
(cherry picked from commit af1818c053)
The parser config file can affect the parsers behavior during tests.
Allow overriding the default location with the option
--config-file=
the option must be the first option in the commands argument list.
Also provile a
--print-config-file
option to display what the parser is using for a config file.
BugLink: http://bugs.launchpad.net/bugs/1277711
Acked-by: Christian Boltz <apparmor@cboltz.de>
Signed-off-by: John Johansen <john.johansen@canonical.com>
(cherry picked from commit b1967c892a)
The version of --config-file that landed in apparmor-2.13 has bugs
and the upstream version evolved before it was committed (it is
not just commits on top of the 2.13 patch).
So to backport the newer version with fixes,
revert commit 56b8e16698.
Signed-off-by: John Johansen <john.johansen@canonical.com>
Add mesa abstraction to allow writing to the Mesa-specific cache
locations and listing devices. Abstraction is needed for applications
utilizing OpenGL API with Mesa implementation available on the system.
When apparmor-2.13 was branched from master the branch name was not
updated in the Makefile. Fix it.
Signed-off-by: John Johansen <john.johansen@canonical.com>
The URL redirect ends up at a page in the new wiki that doesn't exist.
We have to link directly to the gitlab URL here since the current URL
redirect doesn't let us use a wiki.apparmor.net URL and still reach the
expected Profiles page.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
The open-coded readdirfd fn used to replace scandirat skipped
checks for memory allocation failures and cleaning on faulures,
fix this.
Acked-by: Tyler Hicks <tyhicks@canonical.com>
Signed-off-by: John Johansen <john.johansen@canonical.com>
(cherry picked from commit 25f98537db)
The `scandirat` function is a nonstandard GNU extension, which opens a
directory relative to a file descriptor. musl libc does not implement
that function and thus cannot be used to compile libapparmor.
All our uses of `scandirat` directly scan the directory the file
descriptor is referring to, not any directory beneath the FD. Implement
a function `readdirfd()`, which gets as arguments the directory FD, the
location where to put the list of directory entries as well as a
function pointer to a comparing function. `readdirfd` will then scan all
directory entries except "." and ".." and return them via an allocated
array. The array is sorted in case the comparing function is set.
PR: https://gitlab.com/apparmor/apparmor/merge_requests/107
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: John Johansen <john.johansen@canonical.com>
(cherry picked from commit 259a4bad50)
Using stdin with --write-cache set results in
# apparmor_parser --show-cache --write-cache
Cache: added primary location '/var/cache/apparmor'
Warnung aus stdin (Zeile 1): Cache: added readonly location '/usr/share/apparmor/cache'
Warnung aus stdin (Zeile 1): apparmor_parser: cannot use or update cache, disable, or force-complain via stdin
Cache miss: stdin
Wrote cache: /var/cache/apparmor/9b2cd0d0.0/(null)
The "Wrote cache:" message is referencing a null value and should not
be displayed.
BugLink: http://bugs.launchpad.net/bugs/1787717
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Timeout
2.13: prevent that aa-complain etc. overwrites flags in child profiles if they differ from the main profile
See merge request apparmor/apparmor!185
Acked-by: John Johansen <john.johansen@canonical.com>
... instead of overwriting them with the flags of the main profile.
This fixes a longstanding issue with aa-complain, aa-enforce and
aa-audit which broke the flags of child profiles and hats if they
differed from the main profile.
It also fixes several issues documented in the tests (which obviously
need adjustment to match the fixed behaviour).
Also change the "no profile found" cases to AppArmorException - errors
in a profile are not worth triggering AppArmorBug ;-)
(cherry picked from commit b00aab0843)
All callers call change_profile_flags(), so it makes sense to test this
function instead of set_profile_flags().
Besides that, set_profile_flags() will be merged into
change_profile_flags() in the next commit ;-)
Note that this commit adds some '# XXX' notes to the tests. These will
be addressed in later commits.
(cherry picked from commit abd124c00d)
If the old flags are given as str (or None), call split_flags() to
convert them to a list.
This allows to simplify change_profile_flags() which now doesn't need to
call split_flags() on its own.
Also add some tests with a str for the old flags
(cherry picked from commit e80caa130a +
conflict resolution)
... and change change_profile_flags() to use it instead of doing it
itsself
Also add some tests for split_flags()
Cherry-picked from ce7ea062c5 + conflict
resolution
... instead of set_profile_flags() to keep possibly existing flags like
attach_disconnected.
Note that this function is unused (meant to be used with the
no-longer-existing profile repo), therefore nobody noticed that
set_profile_flags() was called with the wrong number of parameters ;-)
KIconLoader uses ~/.cache/icon-cache.kcache, and it is opened in
read-write mode. Because access to it does not seem to be critical, and
read-only mode is not used, rules for accessing this cache is added to
it's own new "write" abstraction, instead of making kde abstraction more
permissive by default.
(cherry picked from commit 94014c09f0)
Signed-off-by: John Johansen <john.johansen@canonical.com>
Currently, kde abstraction only allows reading
~/.config/klanguageoverridesrc file (by design). Some KDE applications
has option to change language for it's interface, and this needs write
access. This is fixed by introducing new abstraction.
(cherry picked from commit 7345f61e9c)
Signed-off-by: John Johansen <john.johansen@canonical.com>
Currently, kde abstraction only allows reading ~/.config/kdeglobals (by
design), though some applications might need to update it's contents
such as KFileDialog settings. This patch fixes it by introducing new
abstraction.
(cherry picked from commit fae93f1b6c)
Signed-off-by: John Johansen <john.johansen@canonical.com>
Each coverity command writes its debugging output to
cov-int/build-log.txt, which means that multiple runs of cov-build
overwrite previous logs, resulting in only the last invocation's output
remaining at the end of the build, making debugging why failures to
capture coverity output difficult. Fix this by renaming the build-log to
per-directory log files.
(This would still be an issue even if we had a single build command
for the entire tree, as capturing python and other interpreted
files requires a second invocation of cov-build to scan for those
file types.)
Signed-off-by: Steve Beattie <steve.beattie@canonical.com>
PR: https://gitlab.com/apparmor/apparmor/merge_requests/145
(cherry picked from commit fed101920b)
Signed-off-by: John Johansen <john.johansen@canonical.com>
When creating a new profile with aa-genprof, get_profile() searches for
an inactive ("extra") profile and, if it finds one, removes the filename
from that profile so that it gets stored in /etc/apparmor.d/ later.
However, it used .pop() to remove the filename, which explodes since
ProfileStorage is a class now.
This patch fixes this (tested manually).
PR: !140
(cherry picked from commit 73b33bdf36)
Signed-off-by: John Johansen <john.johansen@canonical.com>
The parser config file can affect the parsers behavior during tests.
Allow overriding the default location with the option
--config-file=
the option must be the first option in the commands argument list.
Also provile a
--print-config-file
option to display what the parser is using for a config file.
BugLink: http://bugs.launchpad.net/bugs/1277711
Signed-off-by: John Johansen <john.johansen@canonical.com>
Allow /usr/local/lib/python3/dist-packages in abstractions/python
See merge request apparmor/apparmor!160
Acked-by: John Johansen <john.johansen@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de> for 2.10..master
(cherry picked from commit 763a6787d8)
6a10f076 Allow /usr/local/lib/python3/dist-packages in abstractions/python
As Simon McVittie wrote, "if a specification or library creates extra caches, or
has .desktop files in a subdirectory, or anything like that, then I don't see
why we wouldn't want to allow reading those too".
As Simon McVittie <smcv@collabora.com> wrote on
https://bugs.debian.org/865206 and on the AppArmor mailing list:
"Anything in /var/lib/flatpak/exports/share or
~/.local/share/flatpak/exports/share is essentially equivalent to
the corresponding path in /usr/{local/,}share, and is something
that has deliberately been "exported" to the rest of the system by a
Flatpak-confined app.
The only reason to prevent reading those directories would be if you do
not want the AppArmor-confined app to be able to enumerate the other
software you have installed on your system, as an anti-fingerprinting
mechanism.".
Bug-Debian: https://bugs.debian.org/865206
parse_profile_start(): Error out on nested child profiles
See merge request apparmor/apparmor!136
Acked-by: John Johansen <john.johansen@canonical.com> for 2.10..master
(cherry picked from commit b7a4f37cbb)
8462c39b parse_profile_start(): Error out on nested child profiles
The `secure_getenv` function is a non-POSIX compliant extension of
glibc. In contrast to the POSIX `getenv`, `secure_getenv` will return
`NULL` for all environment variables when the program is run with
escalated privileges due to an SUID or SGID bit. Some strictly
POSIX-compliant libc libraries, most notably musl libc, do not have this
function and do not wish to implement it. Thus, AppArmor cannot be
compiled on such systems.
In libapparmor, `secure_getenv` is only used to determine whether the
environment variable DEBUG_ENV_VAR has been set to enable debugging. In
case an unprivileged user runs a SUID/SGID executable linked against
libapparmor, we do not want that user to be able to get additional
information via debug output.
The fix here is to produce an error only in case where debug output is
enabled by defining ENABLE_DEBUG_OUTPUT. Otherwise, we simply define
`secure_getenv` to `NULL` to completely disable the debug output.
(cherry picked from commit 778176b9d8)
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Acked-by: Steve Beattie <steve@nxnw.org>
PR: https://gitlab.com/apparmor/apparmor/merge_requests/107
While the parser makes use of the `aa_policy_cache_add_ro_dir` function,
it is not being declared as a global function in the libapparmor.map
file. Due to this, dynamic linking of apparmor_parser with
libapparmor.so is not possible.
[Fixed up to use 2.13.1 symbol section as when the
`aa_policy_cache_add_ro_dir` was introduced -- @smb]
(cherry picked from commit 1506f2cf0e)
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Steve Beattie <steve.beattie@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
PR: https://gitlab.com/apparmor/apparmor/merge_requests/107
Writing a "link subset" rule missed a space, which resulted in something
like
link subset/foo -> /bar,
Also add a test rule to tests/cleanprof.* to ensure this doesn't break
again.
(cherry picked from commit 514535608f)
Acked-by: Steve Beattie <steve@nxnw.org>
PR: https://gitlab.com/apparmor/apparmor/merge_requests/117
Commit 63b7cb0660 (libapparmor: convert
multicache from using djb2 hashing to murmur3 hash) mistakenly added
PmurHash.h to the list of files generated by the build process and thus
should be removed when the 'maintainerclean' make target is invoked.
This fixes the issue by removing PmurHash.h from the list of generated
files.
(cherry picked from commit 9f2959482f)
Signed-off-by: Steve Beattie <steve.beattie@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
PR: https://gitlab.com/apparmor/apparmor/merge_requests/112
The recently added overlay cache directory support added to libapparmor
makes use of reallocarray(3) to resize memory allocations; however,
reallocarray() was only included in glibc 2.26. This commit adds a
configure check for reallocarray() and if it's not available, provides
it as a wrapper around realloc(3).
PR: https://gitlab.com/apparmor/apparmor/merge_requests/100
Signed-off-by: Steve Beattie <steve.beattie@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
The versioning previous patches was inconsistent because the multicache
patch have been in development for a while and the target version has
changed.
Cleanup libapparmor .map file to use the 2.13 release version
Signed-off-by: John Johansen <john.johansen@canonical.com>
Christian Boltz <apparmor@cboltz.de>
There are several small mistakes/typos in the previous patches. Just
fix them all here.
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
Allow the parser to use cache overlays by extending the --cache-loc
flag to support multiple locations via a comma separated list.
eg.
--cache-loc=/var/cache/apparmor/,/etc/apparmor.d/cache.d/
The overlayed cache directories are searched in the order
specified. So in the above example /var/cache/apparmor is searched
before /etc/apparmor.d/
Time stamps are ignored in the search, the first match found wins
regardless if there exists a matching cache file with a newer timestamp
in a directory is later in the search.
Cache writes will only occur to the first dir in the list. So
/var/cache/apparmor/ in the above example.
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
The feature set needs to be split, the kernel features set determines
the cache location and controls features down grades to ensure
policy generates a policy that is usable on a given kernel.
The compile featurs set governs the feature set supported by policy
and primarily determines how policy is parsed and compiled.
Taking the intersection of the two feature sets to determine rule
downgrades for a specific kernel is left to a separate patch.
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
For some longer pathname fstat is returning a bad size resulting in
the path being truncated. Fix this by detecting a potential truncation
and re-doing the readlink.
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
Add the support to have the cache be able to search multiple locations
so that the policy cache can be split into multiple locations and
that there can be a local cache that can override preshipped caches.
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
This adds the fn aa_policy_cache_add_ro_dir() to the library allowing
for readonly layers to be added to the policy cache. It does not
make those additional layers functional. Which requires the ability
to create and search an overlay of directories.
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
Make the internal cache dir tracking use a fixed array and update
all references to the internal dirfd to index the array.
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
A policy cache is for a specific set of kernel features so there is no
need to keep these separate.
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
Adjust the cache directory name from
<cache_loc>/<feature_id>
to
<cache_loc>/<feature_id>.<n>
where <n> is 0 for the first cache created for a given feature_id.
If there is a feature_id collision then <n> will be incremented to
the next number.
The .features file within each cache directory is used to disambiguate
which feature_id cache dir belongs to which feature set.
Cache collisions and missing caches cause a slow path that searches
existing cache dirs that fit the cache_name pattern, to ensure the
proper dir is chosen.
TODO: add regression tests
create cache dir check it
copy different feature set to it
create cache dir again, check it, check that it incremented...
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
It is possible that a given feature set will hash to the same cache
directory as a different feature set. This will be a problem if binary
caches are required, eg. early boot with systemd doing the cache load.
Detect cache collisions and fail. This is a precursor to handling
collision resolution and should not be committed without the follow
up patch to properly handle collisions.
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
Murmur3 hash is a better hash that djb2 and has a lower chance of
hash collisions, so switch over to using it.
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
Carry the changes made in
libapparmor: Preserve errno across aa_*_unref() functions
into the multicache patcheset
Signed-off-by: John Johansen <john.johansen@canonical.com>
Christian Boltz <apparmor@cboltz.de>
Move the policy cache directory from <cacheloc>/cache/ to
<cacheloc>/cache.d/<features_id>/ where <features_id> is a unique
identifier for a set of aa_features. This allows for multiple AppArmor
policy caches exist on a system. Each policy cache will uniquely
correspond to a specific set of AppArmor kernel features. This means
that a system can reboot into a number of different kernels and the
parser will select the existing policy cache that matches each kernel's
set of AppArmor features.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
Add and export aa_features_id() which can be used to get a unique
identifier for an aa_features object. Internally, this is a djb2 hash of
the features string. The hash function used and even the makeup of the
features ID can be easily changed in the future since external consumers
must use this function to fetch the features ID.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
Store a hash value that can be used to represent the aa_features
object. This will be useful when storing multiple AppArmor policy cache
directories, each based on a kernel feature set.
The hash algorithm used is currently djb2. It was simple to add for
testing purposes, but may eventually need to be changed to something
that is resilient against collisions since there is no handling of
features file hash collisions.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
Use the new --print-cache-dir parser option to construct the policy
cache dir when testing the policy caching functionality.
The majority of the required changes involve fully initializing
self.cmd_prefix prior to calling self.get_cache_dir() since that
function requires self.cmd_prefix to be initialized.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
The --print-cache-dir option can be used to have the parser print the
value of the cache directory that is specific to the features used (from
the current kernel, the --match-string option, or the --features-file
option). After printing the path, apparmor_parser will exit. This is
helpful because the final component in the path will become
unpredictable because it will be based on arbitrary hash function
output.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
Add and export aa_policy_cache_dir_path_preview() which allows the
parser to know exactly where the policy cache binaries, for the
specified aa_policy_cache and aa_features objects, would be stored. This
function may be useful to preview the policy cache dir without having
sufficient permissions or desires to create a policy cache dir.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
Add and export aa_policy_cache_dir_path() which allows the parser to
know exactly where the policy cache binaries, for the current
aa_policy_cache and aa_features objects, will be stored. The parser
previously assumed that it was <cacheloc>/cache/ but it will soon be
<cacheloc>/cache.d/<features_id>/.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
Since the latest change, calling {get,set}_profile_flags() with the
profile name failed when attachment was specified ("profile foo /bar").
Catched by the unittests.
Also fix a whitespace issue.
Getting and Setting profile represented by a glob does not work correctly
because they are checked for equality. Use a glob match to check for them.
Also, add a warning stating that the profile being set represents multiple programs.
traceroute is an example whose profile name is represented as
/usr/{sbin/traceroute,bin/traceroute.db} and exhibits the issue:
Setting /usr/sbin/traceroute to enforce mode.
ERROR: /etc/apparmor.d/usr.sbin.traceroute contains no profile
Signed-off-by: Goldwyn <goldwyn@fiona.lan>
The fix for issue #3/merge !86 in commit f0876ea9 contained a syntax
error that prevented libapparmor from building successfully. This
commit addresses the issue.
Signed-off-by: Steve Beattie <steve.beattie@canonical.com>
... and the apparmor.systemd wrapper.
Also add a new 'install-systemd' target to the Makefile to install these
systemd-related files on (open)SUSE by default. Other distributions can
follow by adding a dependency on 'install-systemd' on their
'install-$DISTRO' target.
Note that apparmor.service has ExecStop=/bin/true to avoid that running
processes get unconfined if someone accidently types
systemctl restart apparmor (instead of using "reload")
Use aa-teardown if you really want to unload all profiles.
The files in this commit are used in openSUSE since a while, and also in
Arch Linux.
BTW: The condition on var-lib.mount is because openSUSE uses
/var/lib/apparmor/cache/ - but with the changed btrfs layout on
openSUSE, maybe I'll change that to /var/cache/apparmor/ which is
a) used by Debian and b) more sane
use_group is only honored if it is defined.
The "real" permission check is reading the logfile - the group check
in aa-notify is just an annoying additional check, and the default
"admin" only works on Ubuntu (other distributions typically use
"wheel").
This commit comments out use_group in the default config, which allows
everybody to use aa-notify. Permissions for reading the log file are of
course still needed.
References: https://bugzilla.opensuse.org/show_bug.cgi?id=1058787
By default, it stays at the "calling directory" instead of the directory
of the Makefile, which breaks "make -C profiles check".
Explicitely set it in the Makefile to get the right directory.
Also adjust the install-suse make target to
- make 'rcapparmor' a symlink to 'service'
- no longer create the 'rcsubdomain' symlink
(open)SUSE does this in apparmor.spec since several releases, so this
commit upstreams the changes the spec did after running make install.
The environ.sh test fails with the following fatal error:
Fatal Error (environ): Unable to run test sub-executable
The reason is due to the tests which use the env_check.sh helper see
unexpected output in the helper's output:
/bin/bash: warning: setlocale: LC_ALL: cannot change locale (C.UTF-8)
I see a number of locale related denials:
apparmor="DENIED" operation="open" profile="/aa/tests/regression/apparmor/env_check.sh" name="/usr/lib/locale/locale-archive" pid=738 comm="env_check.sh" requested_mask="r" denied_mask="r" fsuid=0 ouid=0
apparmor="DENIED" operation="open" profile="/aa/tests/regression/apparmor/env_check.sh" name="/etc/locale.alias" pid=738 comm="env_check.sh" requested_mask="r" denied_mask="r" fsuid=0 ouid=0
apparmor="DENIED" operation="open" profile="/aa/tests/regression/apparmor/env_check.sh" name="/usr/lib/locale/C.UTF-8/LC_IDENTIFICATION" pid=738 comm="env_check.sh" requested_mask="r" denied_mask="r" fsuid=0 ouid=0
apparmor="DENIED" operation="open" profile="/aa/tests/regression/apparmor/env_check.sh" name="/usr/lib/locale/C.UTF-8/LC_CTYPE" pid=738 comm="env_check.sh" requested_mask="r" denied_mask="r" fsuid=0 ouid=0
apparmor="DENIED" operation="open" profile="/aa/tests/regression/apparmor/env_check.sh" name="/usr/lib/locale/C.UTF-8/LC_COLLATE" pid=738 comm="env_check.sh" requested_mask="r" denied_mask="r" fsuid=0 ouid=0
apparmor="DENIED" operation="open" profile="/aa/tests/regression/apparmor/env_check.sh" name="/usr/lib/locale/C.UTF-8/LC_MESSAGES/" pid=738 comm="env_check.sh" requested_mask="r" denied_mask="r" fsuid=0 ouid=0
apparmor="DENIED" operation="open" profile="/aa/tests/regression/apparmor/env_check.sh" name="/usr/lib/locale/C.UTF-8/LC_NUMERIC" pid=738 comm="env_check.sh" requested_mask="r" denied_mask="r" fsuid=0 ouid=0
apparmor="DENIED" operation="open" profile="/aa/tests/regression/apparmor/env_check.sh" name="/usr/lib/locale/C.UTF-8/LC_TIME" pid=738 comm="env_check.sh" requested_mask="r" denied_mask="r" fsuid=0 ouid=0
Allowing everything under /usr/lib/locale/ to be read by the helper
results in the environ.sh test passing.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
libaalogparse uses (unsigned long) -1 to indicate that a log entry does
not contain ouid and/or fsuid fields. The utils logparser was
incorrectly using 2^64 - 1 to detect such a condition but that wasn't
sufficient for 32 bit environments.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Coverity now supports scanning python (and other languages). Apply the
fs-capture-search option to the libapparmor and utils directpries to
capture the python source.
Signed-off-by: Steve Beattie <steve.beattie@canonical.com>
Signed-off-by: John Johansen <john.johansen@canonical.com>
Right now, if you have a named profile with regular expressions to
match binaries, the profile will be shown in aa-status under the
"process list", which doesn't make sense. Instead, show the actual
executable name, and if the profile name differs, report it at the
end (or as a separate field in the json output mode).
Signed-off-by: Kees Cook <keescook@chromium.org>
Update nvidia abstraction to allow creating NVIDIA-specific user directories in
case it is missing (due to fresh $HOME or if manually removed for any reason).
Update base abstraction for ld.so.conf and friends.
See merge request apparmor/apparmor!62
Acked-by: Christian Boltz <apparmor@cboltz.de> for 2.9..master
Add new dri-common abstraction to contain basic DRI-specific rules.
This refactoring is based on a decision to have set of dri-* abstractions for
fine grained control on case-by-case basis. While dri-common is included in X
abstraction by default, additional DRI-related abstractions can be introduced
(such as for enumerating graphics devices) while keeping them logically together
with same dri- prefix.
apparmor: fix regression in network mediation when using feature pinning
When the 4.14-rc6 and earlier kernels are used with an upstream 4.13
or earlier pinned feature set, there is a regression in network
mediation where policy is not being correctly enforced, because the
compilation is completely dropping the af mediation table as expected
by pre 4.14 kernels but the 4.14 kernel is not accounting for this.
Resulting in network denials that can not be fixed by policy.
Signed-off-by: John Johansen <john.johansen@canonical.com>
apparmor: fix regression in network mediation when using feature pinning
When the 4.14-rc6 and earlier kernels are used with an upstream 4.13
or earlier pinned feature set, there is a regression in network
mediation where policy is not being correctly enforced, because the
compilation is completely dropping the af mediation table as expected
by pre 4.14 kernels but the 4.14 kernel is not accounting for this.
Resulting in network denials that can not be fixed by policy.
Signed-off-by: John Johansen <john.johansen@canonical.com>
write_include() was the only user of write_single(), and write_single()
had some include-specific code. This patch gets rid of write_single().
write_include() gets a simplified version of the code instead of calling
write_single().
The tools don't support having multiple rules in one line (they expect
\n after each rule), therefore mark some of the bare_include_tests as
known failures.
console_select_and_upload_profiles() and set_profiles_local_only() both
use a local variable named 'profs'. Rename it to 'profiles'.
This is the first baby step for rewriting how aa.py stores the profiles
internally. I plan to use 'profs' as variable name instead of 'aa', and
this commit gets the result for "grep -r profs" down to 0.
Some of the regression tests are missing conditionals or have the
wrong conditionals so that they fail on current upstream kernels.
Fix this by adding and changing conditionals and requires where
appropriate. With the patches the tests report passing on 4.14 and
4.15 kernels.
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Time out
This is a minimal patch to add conditional includes to the profile
language.
The syntax for conditional includes is similar to regular includes
except with the addition of "if exists" after "include"
include if exists <foo/bar>
include if exists "foo/bar"
include if exists "/foo/bar"
include if exists foo/bar
Note: The patch is designed to be backportable with minimum
effort. Cleanups and code refactoring are planned for follow up
patches that won't be back ported.
Signed-off-by: John Johansen <john.johansen@canonical.com>
On current Debian sid it needs to read
/usr/share/dovecot/protocols.d/imapd.protocol, which is not surprising given it
already needed read access to /usr/share/dovecot/protocols.d/.
I could not find the reason why the upstream Makefile has been installing it
with permissions 555: this predates the migration from SVN.
Regardless, at least on Debian and derivatives, dh_fixperms has been
changing these permissions to 755 forever so it was causing problems,
likely we would know about it by now.
The initial motivation for this change is supporting rootless builds on Debian
and derivatives, also known as "Rules-Requires-Root: no":
- /usr/share/doc/dpkg-dev/rootless-builds.txt* on a Debian system
with a sufficiently recent dpkg-dev installed
- https://nthykier.wordpress.com/2017/10/29/building-packages-without-fakeroot/
- https://lists.debian.org/debian-devel/2017/10/msg00520.html
With this change applied upstream, Debian-based downstreams don't need to adjust
their debian/rules to make this work with "Rules-Requires-Root: no":
chrpath -d $(CURDIR)/debian/tmp/lib/security/pam_apparmor.so
This is needed by new versions of notify-send, as found on openSUSE
Tumbleweed. Without this, desktop notifications don't work anymore, and
notify-send starts to eat up CPU.
If DBUS_SESSION_BUS_ADDRESS is already set, it won't be changed.
This is a minimal patch so that it can be backported to 2.11 and 2.10
which reverts the abort on error failure when the cache can not be
created and write-cache is set.
This is meant as a temporary fix for
https://bugzilla.suse.com/show_bug.cgi?id=1069906https://bugzilla.opensuse.org/show_bug.cgi?id=1074429
where the cache location is being mounted readonly and the cache
creation failure is causing policy to not be loaded. And the
thrown parser error to cause issues for openQA.
Note: A cache failure warning will be reported after the policy load.
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Christian Boltz apparmor@cboltz.de
(cherry picked from commit 42b68b65fe1861609ffe31e05be02a007d11ca1c)
This patch supports rolling a tarball for a release, as well as doing
'make tag'. Only stuff that's been committed should get incorporated
into the tarball.
Signed-off-by: John Johansen <john.johansen@canonical.com>
Denying it means netstat -p (actually tested with -tulpen) can't find
out the program name.
sys_ptrace is "only" needed for tracing processes that run under a
different uid.
Also add ptrace (read), for systems that support ptrace rules.
For now we only allow quoted absolute paths without spaces in the name
due to:
- 1738877: include rules don't handle files with spaces in the name
- 1738879: include rules don't handle absolute paths without quotes in
some versions of parser
- 1738880: include rules don't handle relative paths in some versions of
the parser
.gitignore additions for libapparmor tests, binutils, and the vim syntax highlighting files
See merge request apparmor/apparmor!43
Acked-by: Christian Boltz <apparmor@cboltz.de>
- dict needs abstractions/openssl (seen with dovecot 2.2.31 since
using openssl 1.1)
- imap needs to write tempfiles (seen with dovecot 2.2.31)
- managesieve-login needs access to the login-master-notify socket
(seen with dovecot 2.2.33)
- pop3-login needs access to the anvil socket (reported by pfak on
IRC some months ago)
- extend available_buttons() to display an "owner permissions on/off"
button if the rule supports it
- extend ask_the_questions() to handle these buttons
- add some tests to test-translations.py to avoid hotkey conflicts with
the newly added buttons
- move the code of set_options_audit_mode() to a new function
set_options_mode() and make set_options_audit_mode() a wrapper for it.
- add set_options_owner_mode() as another wrapper for set_options_mode()
and add code to switch the owner flag to set_options_mode()
- add tests for set_options_owner_mode()
This flag defines if the "Owner permissions on/off" button gets
displayed in aa-logprof.
False by default for all rule types (most of them don't support the
owner conditional). Also false for non-owner FileRule.
True only for FileRule if owner=True.
Several log examples result in rules where the 'owner' conditional
should be added. With logparser.py fixed to handle owner-only events, we
need to add the owner conditional to several test_multi/*.profile files.
I verified all log files for the changed profiles and made sure that
- the log line contains fsuid= and ouid=
- fsuid == ouid
I also did a quick check on all log events containing ouid= and for
those with fsuid == ouid, I checked that the profile has the owner
conditional.
Acked-by: Seth Arnold <seth.arnold@canonical.com> for trunk and 2.11
(see mail from 2017-07-31)
logparser.py failed to notice if file events are owner-only in modern
audit.log (using fsuid=... and ouid=...).
This patch adds a comparison of fsuid and ouid and marks file events
as 'owner' if they match.
Note that log events without fsuid=... or ouid=... will have
18446744073709551615 as fsuid / ouid value (that's 2^64 - 1).
'None' would clearly be better ;-)
References: https://bugs.launchpad.net/apparmor/+bug/1538340
Exit rather than returning from shell snippets in Makefiles. It is
reported that returning causes the following error message with bash:
/bin/sh: line 4: return: can only `return' from a function or sourced script
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Reported-by: Christian Boltz <apparmor@cboltz.de>
The utils have tests that rely on the in-tree parser to be built so it
should be documented that the parser should be built first.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
The test-aa-easyprof.py script relies on the parser to be built so the
check target of the utils/test/Makefile should detect if the parser
exists before running any tests.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Reported-by: Christian Boltz <apparmor@cboltz.de>
Since some kernel versions, inherit (ix) needs mmap permissions. Instead
of annoying the user with an avoidable question after adding an ix rule,
always add m permissions.
Together with the already existing code, this means newly added inherit
rules will now have 'mrix' permissions.
autodep() calls read_inactive_profiles() each time it's called (= for
each binary). The result is a "Conflicting profile" error (showing the
same filename twice) if autodep() runs more than once. This can easily
happen when using "aa-autodep /usr/bin/*".
This patch adds an attribute to read_inactive_profiles() that lets the
function return without doing anything if was called before.
check_po.pl lists lots of false positives saying that
msgstr ""
does not have the (h)otkey translated.
This patch whitelists those untranslated strings.
I also tested (by manually "breaking" a translation) that missing
hotkeys still get noticed.
This bug probably exists since forever, therefore I propose this patch
for 2.9..trunk. (OTOH, nobody noticed it, so maybe trunk is enough ;-)
Note: I still get a few false positives for ru.po (no idea why, similar
texts in the other languages don't cause this) - ideas and fixes welcome.
* Alter paths to allow Java version 8 and up.
* Add file rules to fix IcedTea browser plugin.
* Refactor to keep path consistensy against parent and child profile,
reduce repetitive rules.
Allow to read pulseaudio config subdirectories
See merge request apparmor/apparmor!12
Acked-by: Christian Boltz <apparmor@cboltz.de> for 2.9, 2.10, 2.11 and trunk
After using "view changes", the selection got reset to the first changed
profile. This could mislead the user into saving the wrong profile.
This patch ensures the selection is kept.
I propose this patch for trunk and 2.11.
(2.11 will need different indentation again.)
I'm not sure if we should also apply this in 2.10 and 2.9 - they have
the same behaviour, but OTOH I'm not sure if changing behaviour (even if
it's an improvement) in those old releases is a good idea.
Opinions?
The last change in save_profiles() sorted() the order in which the
changed profiles get displayed. However, it did not honor the sorting
when displaying changes or saving the selected profile, leading to the
wrong profile displayed or saved.
This patch fixes picking the selected profile, and at the same time
replaces the duplicated code for doing this with a single instance.
I propose this patch for trunk and 2.11.
Note that the 2.11 branch needs a slightly different patch (different
indentation).
Also note that this regression made it into 2.11.1, so distributions
shipping 2.11.1 should add this patch.
The RETURN VALUE section contained two typos where "kernel_features" was
used instead of "kernel_interface".
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Callers of aa_features_unref(), aa_kernel_interface_unref(), and
aa_policy_cache_unref() had to store off errno and restore it after
calling those functions in error paths. This patch preserves errno
across those *_unref() functions so that callers don't have to.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Ignoring cscope.* files allows users of cscope to not be bothered by
`git status` reporting that an unknown file is in the source tree.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
We never did a release with the JSON code, and YaST (the only known user
of the JSON interface) will work with the added 'changes' dialog type
from r3721 without needing changes.
Also add a better comment/reason why a response for 'changes' is
expected, but gets ignored.
Reviewed-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
Acked-by: Steve Beattie <steve@nxnw.org>
The python setup tools script is set to rewrite the shbang line of
scripts installed in ${PREFIX}/bin/ if the PYTHON environment variable
is set. Unfortunately, this (a) only covers the aa-easyprof script
as the rest are installed in ${PREFIX}/sbin/, and (b) we've deprecated
python 2 support, and hardcoded python3 as the interpreter for all of
the python scripts in the utils/ directory.
The only use for this feature would be if for some reason the utils did
not work properly with the default python3 interpreter and a specific
version was needed to be set, but I don't think that warrants keeping
the extra bit of code complexity around (and indeed, the snippet that
does this is forcibly disabled in Debian/Ubuntu).
Therefore, drop the shbang rewriting entirely.
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: John Johansen <john.johansen@canonical.com>
On 64bit systems, /proc/sys/kernel/pid_max can be set to PID_MAX_LIMIT,
(2^22), which results in seven digit pids. Adjust the @{PID} variable in
tunables/global to accept this.
Acked-by: intrigeri <intrigeri@boum.org>
Acked-by: Steve Beattie <steve@nxnw.org>
Provides the filename in the json format, which can be
directly read by Yast. Increased the protocol version; perhaps
it should go in the next release.
Signed-off-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
[cboltz] fix "unused variable" warning and add a comment about ignoring
the JSON response
Acked-by: Christian Boltz <apparmor@cboltz.de>
This is a preparation patch to use for JSON mode of conveying
diff filename. In this patch we move diff generation functions to UI.
In the process, I have cleaned up the code to reduce code and enable reuse.
Remove unused function get_profile_diff().
Signed-off-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
[cboltz] Also adjust aa-mergeprof to the new function name/location
Acked-by: Seth Arnold <seth.arnold@canonical.com>
The crash was caused by the more strict ProfileStorage in bzr trunk
(older versions use hasher() which is more forgiving, but also very
"useful" to hide quite some bugs)
Acked-by: Christian Boltz <apparmor@cboltz.de>
The added testcase for a ptrace target with an empty string
(ptrace_garbage_lp1689667_1.in) was causing the swig python test script
to fail. The generated python swig record for libapparmor ends up
setting a number of fields to None or other values that indicate the
value is unset, and the test script was checking if the value in the
field didn't evaluate to False in a python 'if' test.
Unfortunately, python evaluates the empty string '' as False in 'if'
tests, resulting in the specific field that contained the empty string
to be dropped from the returned record. This commit fixes that by
special case checking for the empty string.
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: John Johansen <john.johansen@canonical.com>
YaST has two issues in the "save changed profiles" dialog:
- when using "save selected", the list of profiles doesn't get updated.
Update q.options inside the loop to fix this.
- the list of profiles is displayed as "["/usr/bin/foo", true]" instead
of just "/usr/bin/foo". Use changed.keys() instead of changed to fix
this. (text-mode aa-logprof doesn't change, it always displayed
"/usr/bin/foo" and continues to do so.)
References: https://bugzilla.opensuse.org/show_bug.cgi?id=1062667 part a)
Acked-by: Seth Arnold <seth.arnold@canonical.com> for trunk and 2.11.
Note that 2.11 needs a slightly different patch (whitespace diff).
Merge fixes from Christian to address conflicting apparmor-utils
hotkeys in the Indonesian translation. Plus the usual lp timestamp
update.
Acked-by: Steve Beattie <steve@nxnw.org>
Updates to the following translations:
* binutils - add and update an entry to de.po
* utils
- de.po: add several entries
- en_GB.po: add many entries
- es.po: add non-existing(?) entry
- id.po: add many entries
- sv.po: update and add correct a number of entries
All other changes are the usual nonsense of launchpad updating
timestamps and export information.
Note one use of dbus is left because it is represnative of a unix
socket name used for communication with dbus
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
The macro `PATH_MAX` macro is typically defined in the <limits.h>
header by the system's libc implementation. While we do not
include it right now, glibc indirectly includes it via other
headers already and thus compilation of the file succeeds. For
other libc implementations this may not be the case, which would
then lead to a compilation error. This is the case for musl libc.
Explicitly include <limits.h> to fix this.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
The define `RLIMIT_OFILE` is a historic macro originating from
the BSDs, which is nowadays an alias for `RLIMIT_NOFILE`. On some
implementations, it has thus been dropped in favor of the new
define, but we still assume it will always be defined in our
rlimit keywords table. Wrap it in an `ifdef` to fix compilation
on systems where it does not exist.
For the second macro `RLIMIT_RTTIME`, we do check for its
existence in our keywords table, but then forgot to do so in the
YACC rules. Wrap it into an `ifdef`, as well.
Both patches serve the goal to fix compilation on musl libc.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
The macros __BEGIN_DECLS and __END_DECLS are not conforming to
any standard, but are a custom extension of the glibc library. As
such, it may not be available in other libc implementations, with
one example being musl libc. So compiling libapparmor won't work
with a strictly standards-conforming library.
These macros are typically used for header files which might be
included in a C++ project. Depending on whether the header is
seen by a C or C++ compiler, it will hint that functions have C
linkage. The macros themselves are rather simple:
#ifdef __cplusplus
# define __BEGIN_DECLS extern "C" {
# define __END_DECLS }
#else
# define __BEGIN_DECLS
# define __END_DECLS
#endif
To fix compilation with musl libc, simply expand those macros to
explicitly use `extern "C"`. This is already used in other parts
of apparmor and should thus be safe to use.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
The old out of tree patchseries has been completely dropped. v4.13
has most of the newer apparmor 3.x code in it. v4.14 has the rest except
the af_unix mediation which is included as the last patch
Not all kernels support writing the path_max kernel parameter after
boot. Detect if it can be written and run the long_path tests only
if it can be.
Signed-off-by: Seth Forshee <seth.forshee@canonical.com>
Signed-off-by: John Johansen <john.johansen@canonical.com>
so that policy will work on kernels that support network socket controls
but not the extended af_unix rules
however this is currently broken if the socket type is left unspecified
(initialized to -1), resulting in denials for kernels that don't support
the extended af_unix rules.
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: timeout
'smc' seems to be new in kernel 4.12.
Note that the 2.10 apparmor.d manpage also misses the 'kcm' keyword, so
the patch also adds it there.
Acked-by: Seth Arnold <seth.arnold@canonical.com> for trunk, 2.11 and 2.10.
The Samba package used by the INVIS server (based on openSUSE) needs
some additional Samba permissions for the added ActiveDirectory /
Kerberos support.
As discussed with Seth, add /var/lib/sss/mc/initgroups read permissions
to abstractions/nameservice instead of only to the smbd profile because
it's probably needed by more than just Samba if someone uses sss.
Acked-by: Seth Arnold <seth.arnold@canonical.com> for 2.9, 2.10, 2.11 and trunk.
This parameter is always [], so we can simplify the ReadLog __init__()
parameters.
Note that some tests handed over '' instead of []. This was a bug, but
didn't matter because those tests only use a small portion of ReadLog.
Acked-by: Seth Arnold <seth.arnold@canonical.com>
'log' is only used in do_logprof_pass, and reset to [] at the beginning
of the function. Therefore it doesn't need to be a global variable.
Also, do_logprof_pass() initializes log = [], which gets then handed
over to ReadLog and overwritten by the read_log() call in the next line.
To make clear that [] gets handed over to ReadLog, replace log with []
and drop the now superfluous initialization with [].
Acked-by: Seth Arnold <seth.arnold@canonical.com>
- change abstractions/postfix-common to allow /etc/postfix/*.db k
- add several permissions to postfix/error, postfix/lmtp and postfix/pipe
- remove superfluous abstractions/kerberosclient from all postfix
profiles - it's included via abstractions/nameservice
Acked-by: Seth Arnold <seth.arnold@canonical.com> for 2.9, 2.10, 2.11 and trunk
In http://bazaar.launchpad.net/~apparmor-dev/apparmor/master/revision/3659,
a testcase was added that where the expected output file did not match
the input source name, cause libapparmor's regression tests to fail:
Output doesn't match expected data:
--- ./test_multi/ptrace_no_denied_mask.out 2017-08-18 16:35:30.000000000 -0700
+++ ./test_multi/out/ptrace_no_denied_mask.out 2017-08-18 16:35:38.985863094 -0700
@@ -1,5 +1,5 @@
START
-File: ptrace_1.in
+File: ptrace_no_denied_mask.in
Event type: AA_RECORD_DENIED
Audit ID: 1495217772.047:4471
Operation: ptrace
FAIL: ptrace_no_denied_mask
This patch corrects the issue.
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Merge from Vincas Dargis, approved by intrigeri
Fix user-write and user-download abstractions for non-latin file names.
Acked-by: Steve Beattie <steve@nxnw.org>
The updated rule covers the old-style /usr/lib/firefox/firefox.sh
wrapper and the current /usr/lib/firefox{,-esr}/firefox{,-esr} paths.
It is a tiny bit wide but let's lean on the side of compatibility with
whatever similar paths are used in the future. It doesn't grant access
to anything we don't want on a current Debian sid system.
The updated rule covers the old-style /usr/lib/firefox/firefox.sh
wrapper and the current /usr/lib/firefox{,-esr}/firefox{,-esr} paths.
It is a tiny bit wide but let's lean on the side of compatibility with
whatever similar paths are used in the future. It doesn't grant access
to anything we don't want on a current Debian sid system.
- allow reading @{PROC}/@{pid}/net/netstat and @{PROC}/@{pid}/net/snmp
- drop owner conditional - /proc/*/net/* is always owned by root, and
the owner conditional means breaking netstat for non-root users
- drop "@{PROC}/@{pids}/fd r," - /proc/*/fd is a directory, so this rule
would never apply
Acked-by: Steve Beattie <steve@nxnw.org>
Addition by Steve Beattie:
- also allow @{PROC}/@{pid}/net/udplite and @{PROC}/@{pid}/net/udplit6
Acked-by: Christian Boltz <apparmor@cboltz.de>
get_file_perms() and propose_file_rules() happily collect all file
permissions. This could lead to proposing 'wa' permissions in
aa-logprof, which then errored out because of conflicting permissions.
This patch adds a check to both functions that removes 'a' if 'w' is
present, and extends the tests to check this.
Acked-by: Seth Arnold <seth.arnold@canonical.com> for trunk and 2.11.
Note: Both functions (including this bug) were introduced together with
FileRule, so older releases are not affected.
When creating a new child profile, handle_children() did only copy over
include and path rules. While this was correct in the past, path rules
got changed to FileRule in the meantime and were therefore lost.
(In practise, this means the "$binary mr," rule wasn't added to the new
child profile, causing a "superfluous" question in aa-logprof.)
This patch changes handle_children() to carry over the complete new
child profile instead of only cherry-picking include and path rules.
Acked-by: Steve Beattie <steve@nxnw.org> for trunk and 2.11.
Older versions (with path as hasher) are not affected.
Create an EXIT STATUS header and place the BUGS section after the EXIT
STATUS section to match the style in aa-enabled.pod.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-By: Jamie Strandboge <jamie@canonical.com>
Make the possible exit status values bold to match the style used in
aa-status.pod as of r3680.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-By: Jamie Strandboge <jamie@canonical.com>
This option exists in several aa-* tools since 2.9, but isn't mentioned
in the manpage.
Also drop some trailing whitespace in the manpages.
Acked-by: John Johansen <john.johansen@canonical.com>
for 2.9, 2.10, 2.11 and trunk.
While performing genprof, The message to start and scan the program
is mentioned in a separate important message, while it can be
presented as a part of the explanation of the PromptQuestion.
While this will not change the output of text mode, this will help
json clients like yast be more expressive.
Signed-off-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
Also make 'ruletypes' a dict pointing to the *Ruleset class, and change
ProfileStorage __init__() to iterate over 'ruleset'.
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Move ProfileStorage() from aa.py to the new profile_storage.py and make
it a class. The variable name in __init__() changes (profile -> self.data),
but the content stays the same.
The ProfileStorage class acts like a dict(), but has some additional
checks for unknown keys in place.
Also add some tests to make sure unknown keys really raise an exception.
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Network rules are the only rule type that had this safety net - if
profile_data[profile][hat] really isn't initialized (which shouldn't
happen), things will break at lots of other places ;-)
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Based on Cameron Norman's initial work
(http://bazaar.launchpad.net/~cameronnemo/apparmor/gnome-abstraction/revision/3111) with the following changes:
* don't include GTK+ 3.0 configuration: already done earlier
* generalize to future GLib versions
* support /usr/local
* allow reading the parent directory as well, following the lead
of usr.lib.telepathy: this is harmless and could be needed in some cases.
Description: adjust the multiarch alternation rule in the perl abstraction for
modern Debian and Ubuntu systems which store some modules under the
architecture-specific perl-base directory instead of perl or perl5.
Signed-Off-By: Jamie Strandboge <jamie@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
In parse_event_for_tree(), map_log_type() never gets called. Also,
aamode is never 'UNKNOWN'.
Proof for both: I have a local patch that raises an exception for both
cases since two years ;-)
This patch drops the call to map_log_type() and the function itsself.
It also adds a safety check for 'UNKNOWN' - instead of silently ignoring
it, raise an exception (which will most probably never happen).
Acked-by: Seth Arnold <seth.arnold@canonical.com>
tools.py act() is only used by aa-cleanprof, therefore the else branch
(self.name != cleanprof) never gets used.
This patch drops the dead code and renames act() to cleanprof_act() to
make it clear that only aa-cleanprof calls this function.
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Description: adjust the multiarch alternation rule in the perl abstraction for
modern Debian and Ubuntu systems which store some modules under the
architecture-specific perl-base directory instead of perl or perl5.
Signed-Off-By: Jamie Strandboge <jamie@canonical.com>
From: Goldwyn Rodrigues <rgoldwyn@suse.com>
Provides json support to tools in order to interact with other
utilities such as Yast.
The JSON output is one per line, in order to differentiate between
multiple records. Each JSON record has a "dialog" entry which defines
the type of message passed. A response must contain the "dialog"
entry. "info" message does not require a response.
"apparmor-json-version" added in order to identify the communication
protocol version for future updates.
This is based on work done by Christian Boltz.
Signed-off-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
From: Goldwyn Rodrigues <rgoldwyn@suse.com>
This is the yast cleanup from the utils code. All yast communication
should be done with JSON interface now.
Signed-off-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
This patch makes the profile_storage() data structure more strict. It
- initializes everything inside a profile with proper values
- makes the profile storage a dict() instead of a hasher(), which means
it will complain loudly when trying to access non-existing elements
(hasher() was more forgiving, but this also meant hiding bugs)
The patch also fixes a minor issue related to the more strict 'repo'
profile property in serialize_profile().
Acked-by: Seth Arnold <seth.arnold@canonical.com>
- parser/libapparmor_re/parse.cc is autogenerated during build
- parser/tst_lib gets compiled during "make check"
Both files get deleted by make clean.
Acked-by: John Johansen <john.johansen@canonical.com> for trunk and 2.11.
(garbage) ptrace events like
... apparmor="DENIED" operation="ptrace" profile="/bin/netstat" pid=1962 comm="netstat" target=""
cause an empty name2 field, which leads to a crash in the tools.
This patch lets logparser.py ignore such garbage log events, which also
avoids the crash.
As usual, add some testcases.
test-libapparmor-test_multi.py needs some special handling to ignore the
empty name2 field in one of the testcases.
References: https://bugs.launchpad.net/apparmor/+bug/1689667
Acked-by: Seth Arnold <seth.arnold@canonical.com> for trunk and 2.11.
Older releases can't handle ptrace log events and therefore can't crash ;-)
The base abstraction already allows write access to
/run/systemd/journal/dev-log but journald offers both:
- a native journal API at /run/systemd/journal/socket (see sd_journal_print(4))
- /run/systemd/journal/stdout for connecting a program's output to the journal
(see systemd-cat(1)).
In addition to systemd-cat, the stdout access is required for nested container
(eg, LXD) logs to show up in the host. Interestingly, systemd-cat and LXD
containers require 'r' in addtion to 'w' to work. journald does not allow
reading log entries from this socket so the access is deemed safe.
Signed-off-by: Jamie Strandboge <jamie@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
/run/systemd/journal/dev-log but journald offers both:
- a native journal API at /run/systemd/journal/socket (see sd_journal_print(4))
- /run/systemd/journal/stdout for connecting a program's output to the journal
(see systemd-cat(1)).
In addition to systemd-cat, the stdout access is required for nested container
(eg, LXD) logs to show up in the host. Interestingly, systemd-cat and LXD
containers require 'r' in addtion to 'w' to work. journald does not allow
reading log entries from this socket so the access is deemed safe.
Signed-off-by: Jamie Strandboge <jamie@canonical.com>
Error messages should only show up in build logs when the error has been
encountered. This patch silences these shell commands from being printed
before they're interpreted.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
A multi job `make check` command could fail due to check-local running
before the check-DEJAGNU target, which is automatically generated by
automake, would complete. This would result in a build failure due to
libaalogparse.log not yet existing.
Fix the issue by depending on the check-DEJAGNU target.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
Since r3634, the tools allow any order of dbus conditionals.
Quoting the r3634 patch description:
This patch eases the restriction on the ordering at the expense of the
utils no longer being able to detect and reject a single attribute that
is repeated multiple times. In that situation, only the last occurrence
of the attribute will be honored by the utils.
It seems nobody tested with all test profiles generated ;-) so we have to
add some exceptions to the "does not raise an exception" list now.
Acked-by <timeout> for trunk and 2.11
critical urgency notifications result in a notification that must be explictly
clicked to dismiss (ie, they don't time out) and gnome-shell does not honor --
expire-time with (at least) critical urgency. In other popular DEs critical
urgency notifications time out. This patch updates the urgency to 'normal' to
obtain intended behavior across DEs.
Signed-off-by: Jamie Strandboge <jamie@canonical.com>
Acked-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
glibc implements this by doing a readdir() and filtering.
We already allowed sysconf(_SC_NPROCESSORS_ONLN), which is
basically a read from /sys/devices/system/cpu/online.
Signed-off-by: Simon McVittie <smcv@collabora.com>
dovecot-lda needs
- the attach_disconnected flags
- read access to /usr/share/dovecot/protocols.d/
- rw for /run/dovecot/auth-userdb
References: https://bugs.launchpad.net/bugs/1650827
Acked-by: Steve Beattie <steve@nxnw.org> for 2.9, 2.10 and trunk.
In commit 3649, Colin King fixed the readdir test build issue where
aarch64 only supports getdetns64(), not getdents(). Realistically,
however, we want to ensure mediation occurs on both syscalls where
they exist. This patch changes the test to attempt performing both
versions of getdents(). Because we want to catch the situation where
the result of getdents differs from getdents64, we now pass in the
expected result.
Also add a test to verify that having write access does not grant
the ability to read a directory's contents.
Bug: https://bugs.launchpad.net/bugs/1674245
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Acked-by: Tyler Hicks <tyhicks@canonical.com>
https://launchpad.net/bugs/1674245
arm64 build of the tests breaks because getdents is not available.
Where available, use gendents64 as the preferred choice.
Fixes:
cc -g -O0 -Wall -Wstrict-prototypes readdir.c -lapparmor -o readdir
readdir.c: In function ‘main’:
readdir.c:45:14: error: ‘SYS_getdents’ undeclared (first use in this function)
if (syscall(SYS_getdents, fd, &dir, sizeof(struct dirent)) == -1){
^~~~~~~~~~~~
readdir.c:45:14: note: each undeclared identifier is reported only once for each function it appears in
<builtin>: recipe for target 'readdir' failed
make: *** [readdir] Error 1
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
https://launchpad.net/bugs/1668892
This patch creates a new utility, with the code previously used in the
init script 'restart' action, that removes unknown profiles which are
not found in /etc/apparmor.d/. The functionality was removed from the
common init script code in the fix for CVE-2017-6507.
The new utility prints a message containing the name of each unknown
profile before the profiles are removed. It also supports a dry run mode
so that an administrator can check which profiles will be removed before
unloading any unknown profiles.
If you backport this utility with the fix for CVE-2017-6507 to an
apparmor 2.10 release and your backported aa-remove-unknown utility is
sourcing the upstream rc.apparmor.functions file, you'll want to include
the following bug fix to prevent the aa-remove-unknown utility from
removing child profiles that it shouldn't remove:
r3440 - Fix: parser: incorrect output of child profile names
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
CVE-2017-6507
https://launchpad.net/bugs/1668892
The common AppArmor 'restart' code used by some init scripts, upstart
jobs, and/or systemd units contained functionality that is no longer
appropriate to retain. Any profiles not found /etc/apparmor.d/ were
assumed to be obsolete and were unloaded. That behavior became
problematic now that there's a growing number of projects that maintain
their own internal set of AppArmor profiles outside of /etc/apparmor.d/.
It resulted in the AppArmor 'restart' code leaving some important
processes running unconfined. A couple examples are profiles managed by
LXD and Docker.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
With the init_aa() patch series commited, minitools_test.py showed
several test failures - which effectively means the -d option of
aa-complain, aa-cleanprof etc. was broken.
These failures were caused by
- calling init_aa() too late in tools.py - _after_ setting the
profiledir, which then got overwritten by init_aa()
- calling init_aa() twice (because apparmor.aa gets imported in two
modules used by aa-cleanprof), which overwrote the manually set values
on the second run
This patch fixes the call order in tools.py and adds a check to
init_aa() so that it only runs once and ignores additional calls.
Acked-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
FileRule understands leading permissions, so the reason to skip those
(generated) test profiles in test-parser-simple-tests.py is gone.
However, the gen-xtrans.pl script generates profiles with a not-so-valid
mix of uppercase and lowercase, for example "Pux" and "Cux". The parser
accepts this, but the tools complain about such rules. Therefore add the
affected profiles to the exception list.
In total, this means we now test 319 of the 380 generated_perms_leading
test profiles.
The patch also moves some lines around to get the \-escaped profiles
out of the mixed uppercase/lowercase exec rule section.
Acked-by: Seth Arnold <seth.arnold@canonical.com>
The test-aa-easyprof.py script was attempting to do its own special
setup to import the in-tree easyprof module. However, this proved to be
very flaky and resulted in the test periodically failing due to an
AttributeError the first time easyprof.parse_args() was called.
This patch removes the flakiness by trusting that PYTHONPATH is set up
appropriately before the test script is ran. PYTHONPATH is already
initialized appropriately by utils/test/Makefile according to the
USE_SYSTEM make variable.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
if USE_SYSTEM is not set, the utils make check target will instruct
test-aa-easyprof.py to provide the path of the in-tree parser executable
to aa-easyprof.
If USE_SYSTEM is set, the default parser path (/sbin/apparmor_parser or
the result of `which apparmor_parser`) is used.
The test-aa-easyprof.py script receives the parser path by checking the
__AA_PARSER environment variable. This environment variable is strictly
used by the test script and not any user-facing code so two leading
underscores were used.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
When testing against a clean system without the apparmor_parser binary
installed, the test-aa-easyprof.py script ends up skipping profile
verification because it can't find the parser binary. This even causes a
test failure due to the test_genpolicy_invalid_template_policy test.
Adding a --parser option to aa-easyprof is the first step in addressing
this problem.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
If USE_SYSTEM is not set, the utils make check target will instruct
test-aa-easyprof.py to provide the path of the in-tree
profiles/apparmor.d directory to aa-easyprof as the parser base
directory.
If USE_SYSTEM is set, the default base directory (/etc/apparmor.d) is
used.
The test-aa-easyprof.py script receives the base path by checking the
__AA_BASEDIR environment variable. This environment variable is strictly
used by the test script and not any user-facing code so two leading
underscores were used.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Bug: https://launchpad.net/bugs/1538306
https://launchpad.net/bugs/1521031
aa-easyprof accepts a list of abstractions to include and, by default,
execs apparmor_parser to verify the generated profile including any
abstractions. However, aa-easyprof didn't provide the same flexibility
as apparmor_parser when it came to where in the filesystem the
abstraction files could exist.
The parser supports --base (defaulting to /etc/apparmor.d) and --Include
(defaulting to unset) options to specify the search paths for
abstraction files. This patch adds the same options to aa-easyprof to
aide in two different situations:
1) Some Ubuntu packages use aa-easyprof to generate AppArmor profiles
at build time. Something that has been previously needed is a way
for those packages to ship their own abstractions file(s) that are
#included in the easyprof-generated profile. That's not been
possible since the abstraction file(s) have not yet been installed
during the package build.
2) The test-aa-easyprof.py script contains some tests that specify
abstractions that should be #included. Without the ability to
specify a different --base or --Include directory, the abstractions
were required to be present in /etc/apparmor.d/abstractions/ or the
tests would fail. This prevents the Python utils from being able to
strictly test against in-tree code/profiles/etc.
I don't like the names of the command line options --base and --Include.
They're not particularly descriptive and the capital 'I' is not user
friendly. However, I decided to preserve the name of the options from
apparmor_parser.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Introduce an apparmor.aa.init_aa() method and move the initialization
code of the apparmor.aa module into it. Note that this change will break
any external users of apparmor.aa because global variables that were
previously initialized when importing apparmor.aa will not be
initialized unless a call to the new apparmor.aa.init_aa() method is
made.
The main purpose of this change is to allow the utils tests to be able
to set a non-default location for configuration files. Instead of
hard-coding the location of logprof.conf and other utils related
configuration files to /etc/apparmor/, this patch allows it to be
configured by calling apparmor.aa.init_aa(confdir=PATH).
This allows for the make check target to use the in-tree config file,
profiles, and parser by default. A helper method, setup_aa(), is added
to common_test.py that checks for an environment variable containing a
non-default configuration directory path prior to calling
apparmor.aa.init_aa(). All test scripts that use apparmor.aa are updated
to call setup_aa().
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Suggested-by: Christian Boltz <apparmor@cboltz.de>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
The utils tests should make use of the logprof.conf that resides in
utils/test/ when testing against the in-tree parser and profiles. When
testing against the system, it the utils tests should continue to use
the system logprof.conf.
This patch updates the parser and profiles paths to point to the in-tree
paths. Another patch is needed to get aa.py to honor a non-hardcoded
search path for logprof.conf and other configuration files.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
When aa.py is imported, it looks for a set of profiles and it also looks
for the parser. Both of these paths are configured by logprof.conf but
it isn't always obvious which logprof.conf file was used and, therefore,
it isn't always obvious where aa.py is looking. This patch includes the
paths in the error messages.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
https://launchpad.net/bugs/1628286
The utils were enforcing that the dbus rule attributes were strictly
ordered in the following fashion:
bus -> path -> interface -> member -> peer
However, the parser has always accepted the attributes in any order. If
the system contained a profile which did not use the strict ordering
enforced by the utils, the utils would refuse to operate at all.
This patch eases the restriction on the ordering at the expense of the
utils no longer being able to detect and reject a single attribute that
is repeated multiple times. In that situation, only the last occurrence
of the attribute will be honored by the utils.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
The merged /usr patches to the policy broke some utils tests due to a
change in the expected output.
Fixes: r3600 update lots of profiles for usrMerge
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
That's much better than crashing aa-logprof ;-) (use the log line in
the added testcase if you want to see the crash)
Reported by pfak on IRC.
Acked-by: Seth Arnold <seth.arnold@canonical.com> for trunk, 2.10 and 2.9.
Starting with python 3.6, the re.LOCALE flag can only be used with byte
patterns, and errors out if used with str. This patch removes the flag
in get_translated_hotkey().
References: https://bugs.launchpad.net/apparmor/+bug/1661766
Acked-by: Steve Beattie <steve@nxnw.org> for trunk, 2.10 and 2.9
In the environ regression test, when the exec() of the child process
fails, we don't report FAIL to stdout, so the regression tests consider
it an error rather than a failure and abort, short-circuiting the
test script.
This commit fixes this by emitting the FAIL message when the result
from the wait() syscall indicates the child process did not succeed.
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
aa.py has a global variable "pid", but it also has several functions
that use "pid" as a local variable name. do_logprof_pass() even uses
both - first, it passes the global variable to ReadLog, and then it
creates a local variable in the "for pid in ..." loop.
This patch renames the global variable to log_pid to get rid of the
confusion.
Note that the global variable is only handed over to ReadLog, and the
only case where its previous content _might_ be used is aa-genprof which
does multipe do_logprof_pass() runs.
Maybe we could even get rid of this variable in aa.py and make it local
to the ReadLog class, but I'm not sure if that would affect aa-genprof
in interesting[tm] ways.
Acked-by: John Johansen <john.johansen@canonical.com>
Some of the /usr/lib/dovecot/* rules already have mrPx permissions,
while others don't.
With a more recent kernel, I noticed that at least auth, config, dict,
lmtp, pop3 and ssl-params need mrPx instead of just Px (confirmed by the
audit.log and actual breakage caused by the missing mr permissions).
The mr additions for anvil, log and managesieve are just a wild guess,
but I would be very surprised if they don't need mr.
Acked-by: Seth Arnold <seth.arnold@canonical.com> for trunk, 2.10 and 2.9.
Add several permissions to the dovecot profiles that are needed on ubuntu
(surprisingly not on openSUSE, maybe it depends on the dovecot config?)
As discussed some weeks ago, the added permissions use only /run/
instead of /{var/,}run/ (which is hopefully superfluous nowadays).
References: https://bugs.launchpad.net/apparmor/+bug/1512131
Acked-by: Seth Arnold <seth.arnold@canonical.com> for trunk, 2.10 and 2.9.
Grepping through the code shows that running_under_genprof,
unimplemented_warning, ALL, t, seen and skip are unused, so drop them.
Acked-by: Steve Beattie <steve@nxnw.org>
Also drop a '# t = hasher()" comment, as noticed by Steve.
Replace most of aa-mergeprof ask_merge_questions() with a call to
aa.py ask_the_questions() (which is, besides some small exceptions that
are not relevant for aa-mergeprof, in sync with the dropped code).
The remaining part gets renamed to ask_merge_questions() to avoid
confusion with the function name in aa.py. Also drop the (now
superfluous) parameter.
aa.py ask_the_questions() needs to allow 'merge' as aamode.
While on it, replace the fatal_error() call for unknown aamode with
raising an AppArmorBug.
Acked-by: Seth Arnold <seth.arnold@canonical.com>
This allows to hand over any source instead of using the global variable.
Now that the function expects its input as parameter, get rid of the
global log_dict, which means
- change collapse_log() to initialize log_dict as local variable and
return it
- change do_logprof_pass() to catch collapse_log()'s return value and
hand it over to ask_the_questions()
- drop all references to the global log_dict variable
- update test-libapparmor-test_multi to follow the changes
Also fix an if condition that would fail if aa[profile][hat] does not
exist - get() defaults to None if the requested item doesn't exist, and
None.get('file') will raise an Exception.
Acked-by: Seth Arnold <seth.arnold@canonical.com>
The function is an exact copy of the code in aa-mergeprof (except
removing the 'self' function parameter and changing the whitespace
level)
Also add a ask_conflict_mode() call to aa.py ask_the_questions().
This is needed for aa-mergeprof, and won't hurt in aa-logprof mode
because handle_children() already handles all exec events.
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Bug: https://launchpad.net/bugs/1522938
Everything below "if aamode == 'merge':" is an exact copy of the code in
aa-mergeprof (with whitespace changed).
aa-logprof and aa-mergeprof will continue to ignore events from unknown
hats and subprofiles.
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Set log_dict['merge'] = other.aa and aamode = 'merge', and use
log_dict[aamode] everywhere.
This brings aa-mergeprof ask_the_questions() closer to the code in aa.py.
Acked-by: Seth Arnold <seth.arnold@canonical.com>
3-way-merge was never really implemented.
This patch drops all traces of it to make the code more readable and
easier to maintain.
Acked-by: Seth Arnold <seth.arnold@canonical.com>
The local/ include in the sshd profile in extras causes some trouble:
- it breaks "make check" because the parser can't find the local/ file
- it results in a broken profile if someone uses this profile as
starting point, but doesn't notice it needs the local include
Acked-by: Steve Beattie <steve@nxnw.org>
Thanks to Daniel Curtis for working on this!
Acked-by: Seth Arnold <seth.arnold@canonical.com> for whichever branches
it makes sense for
-> trunk (includes 2.11) only - if we want it in 2.10 and 2.9, we'll
also need to backport the usrMerge changes
ldd exits with $? == 1 if a file is 'not a dynamic executable'.
This is correct behaviour of ldd, so we should handle it instead of
raising an exception ;-)
Also extend fake_ldd and add a test to test-aa.py to cover this.
Note that 2.10 and 2.9 don't have tests for get_reqs() nor fake_ldd,
so those branches will only get the aa.py changes.
Acked-by: John Johansen <john.johansen@canonical.com> for trunk, 2.10 and 2.9.
This patch allows a user to specify a specific location for ss or
netstat in the invocations of get_pids_ss() or get_pids_netstat().
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: Christian Boltz <apparmor@cboltz.de>
This patch adjusts aa-unconfined to avoid using cat(1) to read
/proc/PID/cmdline entries, and instead opens them for reading directly.
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: John Johansen <john.johansen@caanonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
It was reported that converting the netstat command to examine
processes bound to ipv6 addresses broke on OpenSUSE due to the version
of nettools not supporting the short -4 -6 arguments.
This patch switches to use the ss(8) utility from iproute2 by default
(if ss is found) as netstat/net-tools is deprecated. Unfortunately,
ss's '--family' argument does not accept multiple families, nor
does passing '--family' multiple times with different arguments work
either, so aa-unconfined invokes ss multiple times to gather the
different socket families.
It also fixes the invocation of netstat to use the "--protocol
inet,inet6" arguments instead, which should return the same results
as the short options.
This patch provides command line arguments to manually switch using
one tool or the other, as well as converting the invocations of ss
and netstat to not use a shell, and documents these options in the
aa-unconfined man page.
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: Christian Boltz <apparmor@cboltz.de>
Acked-by: John Johansen <john.johansen@canonical.com>
nmbd needs some additional permissions:
- k for /var/cache/samba/lck/* (via abstractions/samba)
- rw for /var/cache/samba/msg/ (the log only mentioned r, but that
directory needs to be created first)
- w for /var/cache/samba/msg/* (the log didn't indicate any read access)
Reported by FLD on IRC, audit log on https://paste.debian.net/902010/
Acked-by: Steve Beattie <steve@nxnw.org> for trunk, 2.10 and 2.9
Rename th odt files to no longer contain spaces in their names, as
make(1) does not work well with such files.
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
The latex based techdoc in the parser/ tree adds a number of build
dependencies for downstreams to create it; it also is the primary
element to make the builds unrepeatable. Creating the techdoc and other
documentation when generating a tarball for distribution avoids all
that.
* Makefile: build documentation as part of the tarball creation. Skip
the libraries/libapparmor directory as it needs to have configure run
before the manpages can be made.
* changehat/mod_apparmor/Makefile, changehat/mod_apparmor/Makefile,
utils/Makefile, profiles/Makefile: create separate docs target,
some of them dummies.
* parser/Makefile: pull the techdoc out of the default build target, add
an extra_docs target to create it.
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: John Johansen <john.johansen@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
The snapshot/tarball builds use some shell constructs that end
up causing failures at various stages to be ignored. This commit
addresses that.
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: John Johansen <john.johansen@canonical.com>
Merge lp:~intrigeri/apparmor/usrMerge:
intrigeri@boum.org 2016-12-07 Adjust white-space back to "tabular style" and make one merged-/usr related rule look like the others.
intrigeri@boum.org 2016-12-03 abstractions/base: drop 'ix' for ld-*.so and friends.
intrigeri@boum.org 2016-12-03 abstractions/base: revert ix→Pix.
intrigeri@boum.org 2016-12-03 abstractions/base: turn remaining ix rules into Pix.
intrigeri@boum.org 2016-12-03 abstractions/base: turn merged-/usr-enabled ix rules into Pix, to avoid conflicts with other profiles.
intrigeri@boum.org 2016-12-03 abstractions/base: drop obsolete rule, supersede by @{multiarch} a while ago.
intrigeri@boum.org 2016-12-03 Make policy compatible with merged-/usr.
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
Additionally, I did some whitespace fixes in the dhclient and procmail
profile before commiting the merge.
openSUSE uses "php7" (not just "php") in several paths, so also allow that.
Acked-by: John Johansen <john.johansen@canonical.com> for trunk, 2.10 and 2.9.
Sometimes network events come with an operation keyword looking like
file_perm which makes them look like file events. Instead of ignoring
these events (which was a hotfix to avoid crashes), improve the type
detection.
In detail, this means:
- replace OPERATION_TYPES (which was basically a list of network event
keywords) with OP_TYPE_FILE_OR_NET (which is a list of keywords for
file and network events)
- change op_type() parameters to expect the whole event, not only the
operation keyword, and rebuild the type detection based on the event
details
- as a side effect, this simplifies the detection for file event
operations in parse_event_for_tree()
- remove workaround code from parse_event_for_tree()
Also add 4 new testcases with log messages that were ignored before.
References:
a) various bugreports about crashes caused by unexpected operation keywords:
https://bugs.launchpad.net/apparmor/+bug/1466812https://bugs.launchpad.net/apparmor/+bug/1509030https://bugs.launchpad.net/apparmor/+bug/1540562https://bugs.launchpad.net/apparmor/+bug/1577051https://bugs.launchpad.net/apparmor/+bug/1582374
b) the summary bug for this patch
https://bugs.launchpad.net/apparmor/+bug/1613061
Acked-by: Steve Beattie <steve@nxnw.org> for trunk and 2.10.
This should solve the "overlapping rules with conflicting 'x'
modifiers" problem (introduced with r3594) entirely.
The other options I could think of were:
* ix → Pix, adjust all profiles that do 'ix' accordingly, and leave
alone those that do Pix already; downsides: requires updating quite
a few profiles all around the place, and breaks a mere "file," rule;
* ix → Pix, adjust all profiles that do 'ix' accordingly, and change
the "file," rule semantics to imply Pix; downside: very intrusive,
and likely to break random existing policy in ways that are hard
to predict;
* stick to ix, and adjust all profiles that do anything else with
overlapping rules, to do ix instead; downside: in some cases this means
removing the 'P' modifier, which can cause regressions in how we confine
stuff.
I've looked up in the bzr history to understand why execution rights
would be needed, and… the answer predates the move to bzr.
Looking into the SVN history, if it's even available anywhere, is
a bit too much for me, so I've tested this change and the few
applications I've tried did not complain. Of course, more testing will
be needed.
Having consistent x modifiers in this abstraction is needed
to allow profiles including abstractions/base to apply x rules
overlapping with several of the rules from the base abstraction.
E.g. one may need to have rules applying to /**, for example because
a mere "file," conflicts with the ix→Pix change I did in r3596.
netstat -nlp46 output:
raw6 0 0 :::58 :::* 7 1326/NetworkManager
which when asking netstat to display name resolution ends up being:
raw6 0 0 [::]:ipv6-icmp [::]:* 7 1326/NetworkManager
Of course, aa-unconfined doesn't show this, the following patch adds
that, by adding the raw keyword as an alternative to tcp|udp and
accepting a number as an alternative to LISTEN.
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: John Johansen <john.johansen@canonical.com>
On servers with not too much memory ("only" 16 GB), dovecot logins fail:
Nov 25 21:35:15 server dovecot[28737]: master: Fatal: setrlimit(RLIMIT_DATA, 268435456): Permission denied
Nov 25 21:35:15 server dovecot[28731]: master: Error: service(auth): command startup failed, throttling for 2 secs
Nov 25 21:35:15 server dovecot[28737]: auth: Fatal: master: service(auth): child 25976 returned error 89 (Fatal failure)
audit.log messages are:
... apparmor="DENIED" operation="capable" profile="/usr/sbin/dovecot" pid=25000 comm="dovecot" capability=24 capname="sys_resource"
... apparmor="DENIED" operation="setrlimit" profile="/usr/sbin/dovecot" pid=25000 comm="dovecot" rlimit=data value=268435456
After allowing capability sys_resource, dovecot can increase the limit
and works again.
Acked-by: Steve Beattie <steve@nxnw.org> for trunk, 2.10 and 2.9
Storing these event details depending on the operation type only makes
things more difficult because it's hard to differenciate between file
and network events.
Note that this happens at the first log parsing stage (libapparmor log
event -> temporary python array) and therefore doesn't add a serious
memory footprint. The event tree will still only contain the elements
relevant for the actual event type.
This change means that lots of testcases now get 3 more fields (all
None) when testing parse_event(), so update all affected testcases.
(test-network doesn't need a change for probably obvious reasons.)
Also rename a misnamed test in test-change_profile.
Acked-by: Seth Arnold <seth.arnold@canonical.com> for trunk and 2.10.
The latest glibc (including nscd) in openSUSE Tumbleweed comes with
glibc-2.3.3-nscd-db-path.diff: Move persistent nscd databases to
/var/lib/nscd
This needs updates (adding /var/lib/nscd/) to abstractions/nameservice
and the nscd profile.
Acked-by: Seth Arnold <seth.arnold@canonical.com> for trunk, 2.10 and 2.9
nmbd, winbindd (and most probably also smbd - but it has a more
permissive profile that already allows this) need rw access to
/var/cache/samba/lck/* on Debian 8.6.
Reported by FLD on IRC.
Acked-by: Seth Arnold <seth.arnold@canonical.com> for trunk, 2.10 and 2.9
The apparmor.d description about alias rules was broken in multiple
ways. The manpage
- didn't include the alias keyword
- listed alias rules in the "COMMA RULES" section - while that's correct
for the comma requirement, it's also wrong because COMMA RULES is
meant to be inside a profile
- didn't list alias rules in the PREAMBLE section
This patch fixes this.
It also moves the definition of VARIABLE, VARIABLE ASSIGNMENT (both
unchanged) and ALIAS RULE next to PREAMBLE.
Acked-by: Seth Arnold <seth.arnold@canonical.com> for trunk and 2.10
Fix import errors with swig > 3.0.8 with the libapparmor python
bindings. Do this by removing the code to rename the generated
LibAppArmor.py, and instead use a stub __init__.py that automatically
imports everything from LibAppArmor.py. Also adjust bzrignore to
compensate for the autogenerated file name changing.
Bug: https://bugzilla.opensuse.org/show_bug.cgi?id=987607
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: Christian Boltz <apparmor@cboltz.de>
Newer kernels need m permissions for the binary the profile covers,
so add it before someone hits this problem in the wild ;-)
Also add a note that the mlmmj-recieve profile is probably superfluous
because upstream renamed the misspelled binary.
Acked-by: Seth Arnold <seth.arnold@canonical.com> for trunk, 2.10 and 2.9
This patch updates the mlmmj profiles in the extras directory to the
profiles that are used on lists.opensuse.org now. Besides adding lots
of trailing slashes for directories, several permissions were added.
Also, usr.bin.mlmmj-receive gets added - it seems upstream renamed
mlmmj-recieve to fix a typo.
These profiles were provided by Per Jessen.
References: https://bugzilla.opensuse.org/show_bug.cgi?id=1000201
Acked-by: Seth Arnold <seth.arnold@canonical.com> for trunk, 2.10 and 2.9.
Some conditions in RlimitRule can never be hit under normal
circumstances, so this patch adds some "pragma: no cover" and
"pragma: no branch" comments to beautify the coverage report.
Acked-by: Seth Arnold <seth.arnold@canonical.com>
The log line (with a different profile=...) was sitting around on my
disk since a year, so let's do something useful with it ;-)
Acked-by: Seth Arnold <seth.arnold@canonical.com>
This patch adds profiles for all log sniplets that are expected to
result in a profile rule.
This also means some changes in test-libapparmor-test_multi.py are
needed:
- split off log_to_profile_skip from log_to_profile_known_failures to
- only skip tests in log_to_profile_skip (causing a crash or requiring
user interaction)
- run tests in log_to_profile_known_failures, but expect a non-equal
result (caused by not added rules etc.)
- add quite some tests to log_to_profile_known_failures - they were
skipped before because they didn't have a *.profile file.
- add handling for hats to shorten list of known failures
This fixes testcase24 and testcase33 (after adjusting the profiles)
and lots of the new *.profile files.
- since we now have *.profile files for all log events that should result
in a profile rule, no longer ignore FileNotFoundError
Acked-by: Seth Arnold <seth.arnold@canonical.com>
This patch adds TestLogToProfile to test-libapparmor-test_multi.py which
"translates" the test_multi log sniplets to a profile, and checks if it
matches the expected profile.
The expected profile for one log event will obviously contain only one
rule, and gets added as *.profile to the test_multi directory.
This patch includes 33 test_multi profiles - which means 83 more need to
be created. Whenever you have some time, add one or two! (Please write
those test_multi profiles manually, without using the tools.)
I know some parts of the test code looks complicated. Unfortunately this
is how things work - compare it with do_logprof_pass() in aa.py...
While on it, set tests = 'invalid' which ensures a failure in case
parse_test_profiles() doesn't set the tests array, and move printing
the test name out of parse_test_profiles() to avoid printing it twice.
A nice side effect of this patch is increased test coverage:
- 30% -> 40% in aa.py (= 250 more lines)
- 52% -> 78% in aamode.py (= 23 more lines)
- 26% -> 68% in logparser.py (= 120 more lines)
- total coverage increases from 57% to 62%
Acked-by: Seth Arnold <seth.arnold@canonical.com>
I already did this in the python code a month ago, and now realized that
we should also update the apparmor.d manpage ;-)
Acked-by: Seth Arnold <seth.arnold@canonical.com> for trunk and 2.10.
seen_events is a global variable in aa.py that gets increased at several
places, but isn't used (read or printed) anywhere. Since I can't imagine
how it could become useful, simply drop it.
Also drop an outdated comment in handle_children that lived next to a
seen_events line.
Acked-by: Seth Arnold <seth.arnold@canonical.com>
These files are needed for disk-based buffering (added in syslog-ng 3.8).
This was reported to me by Peter Czanik, one of the syslog-ng developers.
Note: I'm not sure about adding @{CHROOT_BASE} to this rule, so for now
I prefer not to do it - adding it later is easy, but finding out if it
could be removed is hard ;-)
Acked-by: John Johansen <john.johansen@canonical.com> for trunk, 2.10 and 2.9.
This little change means that the tests will run as part of 'make check'.
This commit is only a 'bzr mv utils/test/config_test.py utils/test/test-config.py'
without any changes in the file content.
Acked-by: Steve Beattie <steve@nxnw.org>
aa_test.py doesn't run in 'make check' because its filename doesn't
match the 'test-*.py' pattern, so this move means the tests now actually
get run.
While on it, migrate test-aamode.py to use the AATest base class, and
migrate the str_to_mode() tests to a tests[] array.
After this move, aa_test.py doesn't do anything anymore, so delete it.
Acked-by: Steve Beattie <steve@nxnw.org>.
Also add another test proposed by Steve:
(None, set()),
aa_test.py doesn't run in 'make check' because its filename doesn't
match the 'test-*.py' pattern.
mode_to_str() was dropped as part of the FileRule series, so it's
pointless to keep its tests. (The replacement is totally different and
has full test coverage already.)
loadincludes() still exists, but only testing if the function runs
without errors is not really helpful, so drop this test.
Also drop unused imports and add an explicit import for apparmor.aamode.
Acked-by: Steve Beattie <steve@nxnw.org>
aa_test.py doesn't run in 'make check' because its filename doesn't
match the 'test-*.py' pattern.
Move tests for globbing ("plain" globbing and globbing with ext) to
test-aare.py to make sure those tests actually run.
Note: This isn't an exact move - I adjusted some of the tests to make
them more useful, and added some more tests.
Also, glob_path() and glob_path_withext() no longer exist in aa.py.
They moved to the AARE class as part of the FileRule patch series.
Acked-by: Steve Beattie <steve@nxnw.org>
Add a testcase with exec-only permissions (which get ignored by
get_perms_for_path()) to increase FileRule test coverage to 100%.
Acked-by: Steve Beattie <steve@nxnw.org>
- dovecot/auth: allow to read stats-user
- dovecot/config: allow to read /usr/share/dovecot/**
- dovecot/imap: allow to ix doveconf, read /etc/dovecot/ and
/usr/share/dovecot/**
These things were reported by Félix Sipma in Debian Bug#835826
(with some help from sarnold on IRC)
References: https://bugs.debian.org/835826
Acked-by: Seth Arnold <seth.arnold@canonical.com> for trunk, 2.10 and 2.9.
Also allow reading ~/.dovecot.svbin (that's the default filename in the
dovecot config) in dovecot/lmtp profile.
(*.svbin files can probably also appear inside @{DOVECOT_MAILSTORE}, but
that's already covered by the existing rules.)
References: https://bugs.debian.org/835826 (again)
Acked-by: John Johansen <john.johansen@canonical.com> for trunk, 2.10 and 2.9
exec_stack picked up a fix to address a semantic change introduced in
4.8 kernels.
However older kernels don't need the extra permission and the exec_stack
test is the only test we currently have that caught the semantic change.
Keep exec_stack to the minimum set of permissions needed for a given
kernel. Which allows us to use exec_stack as a test to detect the
semantic change showing up in unexpected place until we have a test
specifically designed for this.
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Acked-by: Tyler Hicks <tyhicks@canonical.com>
The latest version of pyflakes (1.3.0 / python 3.5) complains that
CMD_CONTINUE is defined twice in ui.py (with different texts).
Funnily CMD_CONTINUE isn't used anywhere, so we can just drop both.
Acked-by: Seth Arnold <seth.arnold@canonical.com> for trunk, 2.10 and 2.9
As discussed a while ago, switch the utils (including their tests) to
use python3 by default. While on it, drop usage of "env" to always get
the system python3 instead of a random one that happens to live
somewhere in $PATH.
In practise, this patch doesn't change much - AFAIK openSUSE, Debian and
Ubuntu already patch aa-* to use python3.
Also add a note to README to officially deprecate Python 2.x.
(I won't break Python 2.x support intentionally - unless some future
change gives me a very good reason to finally drop Python 2.x support.)
Acked-by: Seth Arnold <seth.arnold@canonical.com>
(since 2016-08-23, but the commit had to wait for the FileRule series
because it touches test-file.py)
After looking at matchliteral(), I found out that it's only user is
rematchfrag(), which is only called in a) an "if False:" block and
b) match_include_to_path() - and that is only called by the also unused
match_prof_incs_to_path() function.
This patch drops some dead code (like the mentioned "if False:" block)
and the now unused functions
- matchliteral()
- rematchfrag()
- match_include_to_path()
- match_prof_incs_to_path()
This patch is also THE ANSWER to the question when I'll finally consider
this patch series complete.
42. It can't become better than that! ;-)
Acked-by: Steve Beattie <steve@nxnw.org>
If a merged profile contains additional hats or subprofiles, the "old"
aa-mergeprof silently created them as additional hasher elements (partly
buggy, because subprofiles would end up as '^/subprofile' instead of
'profile /subprofile'). After switching to FileRule, aa-mergeprof crashes
on new hats or subprofiles.
This patch adds code to ask the user if the new hat or subprofile should
be added - which means this patch replaces two bugs (crash + silently
adding subprofiles and hats) with a new feature ;-)
The new questions also add a new text CMD_ADDSUBPROFILE in ui.py.
Finally, the new "button" combinations get added to test-translations.py.
If you want to test, try to aa-mergeprof this profile (the subprofile
and hat are dummies, nothing ping would really require):
#include <tunables/global>
/{usr/,}bin/ping {
#include <abstractions/base>
#include <abstractions/consoles>
#include <abstractions/nameservice>
capability net_raw,
capability setuid,
network inet raw,
network inet6 raw,
/{,usr/}bin/ping mixr,
/etc/modules.conf r,
^hat {
/bin/hat r,
/bin/bash px,
}
profile /subprofile {
/bin/subprofile r,
/bin/bash px,
}
# Site-specific additions and overrides. See local/README for details.
#include <local/bin.ping>
}
Note that this patch is not covered by unittests, but it passed all my
manual tests.
Acked-by: Steve Beattie <steve@nxnw.org>
Bug: https://launchpad.net/bugs/1507469
aa-mergeprof empties 'includes' when running reset_aa(). The result is
KeyError: 'abstractions/newly_added_abstraction'
if an include file gets added because it isn't part of 'includes' at
this time. Note that you'll need to add another rule after adding the
include to trigger checking the includes for superfluous rules.
This fixes the regression found by Steve - which isn't really a
regression, "just" one more thing that got more visible with the new
code. Before, it was just an ill-addressed hasher that didn't complain ;-)
Acked-by: Steve Beattie <steve@nxnw.org>
The switch to FileRule made some bugs visible that survived unnoticed
with hasher for years.
If aa-logprof sees an exec event for a non-existing profile _and_ a
profile file matching the expected profile filename exists in
/etc/apparmor.d/, it asks for the exec mode nevertheless (instead of
being silent). In the old code, this created a superfluous entry
somewhere in the aa hasher, and caused the existing profile to be
rewritten (without changes).
However, with FileRule it causes a crash saying
File ".../utils/apparmor/aa.py", line 1335, in handle_children
aa[profile][hat]['file'].add(FileRule(exec_target, file_perm, exec_mode, rule_to_name, owner=False, log_event=True))
AttributeError: 'collections.defaultdict' object has no attribute 'add'
This patch makes sure exec events for unknown profiles get ignored.
Reproducer:
python3 aa-logprof -f <(echo 'type=AVC msg=audit(1407865079.883:215): apparmor="ALLOWED" operation="exec" profile="/sbin/klogd" name="/does/not/exist" pid=11832 comm="foo" requested_mask="x" denied_mask="x" fsuid=1000 ouid=0 target="/sbin/klogd//null-1"')
This causes a crash without this patch because
/etc/apparmor.d/sbin.klogd exists, but has
profile klogd /{usr/,}sbin/klogd {
References: https://bugs.launchpad.net/bugs/1379874
Acked-by: Steve Beattie <steve@nxnw.org> for trunk, 2.10 and 2.9
FileRule uses RE_PROFILE_FILE_ENTRY, which also means
RE_PROFILE_PATH_ENTRY, RE_PROFILE_BARE_FILE_ENTRY and RE_OWNER are now
unused.
This patch drops these regexes and their tests in test-regex_matches.py.
Acked-by: Steve Beattie <steve@nxnw.org>
rank() in severity.py is a dispatcher that calls the needed function
(rank_path(), rank_capability()) based on the parameter. Since all
calling code knows what rule type it is handling, this dispatcher is
superfluous - the calling code can call rank_path() or rank_capability()
directly.
This patch drops rank() and switches the remaining users of rank() to
call the rank_*() functions directly. For the tests, this means to drop
the CAP_ prefix because rank_capability doesn't expect this prefix.
Acked-by: Steve Beattie <steve@nxnw.org>
After switching to FileRule, several functions in aamode.py are no
longer used and can be deleted:
- print_mode()
- sub_mode_to_str()
- is_user_mode()
- split_mode()
- mode_to_str()
- flatten_mode()
- owner_flatten_mode()
- mode_to_str_user()
- log_str_to_mode()
The AA_EXEC_TYPE and ALL_AA_EXEC_TYPE variables are also unused now.
Acked-by: Steve Beattie <steve@nxnw.org>
When an user adds a new rule to a profile, cleanup / delete existing
rules that are covered by the new rule, and report the number of deleted
rules.
Acked-by: Steve Beattie <steve@nxnw.org>
Adding a rule to *Ruleset means it simply gets added. This also means
that then-superfluous rules will be kept.
This patch adds an optional cleanup flag to add(). If set, rules covered
by the new rule will be deleted. The difference to delete_duplicates()
is that cleanup only deletes rules that are covered by the new rule, but
keeps other, unrelated superfluous rules.
Also return the number of deleted rules to give the UI a chance to
report this number.
Finally, adjust the existing tests for FileRuleset to ensure default
mode (without cleanup) doesn't delete any rules, and add a test using
the cleanup flag.
Acked-by: Steve Beattie <steve@nxnw.org>
Replace the old (hasher-based) conflict_mode() with the new
(FileRule-based) ask_conflict_mode() function. If it detects conflicting
exec rules, it asks the user which one to keep.
Also call ask_conflict_mode() from ask_the_questions() so that it is
actually used.
Note: This patch isn't covered by unittests, but I did some manual
testing to make sure it works as expected.
Acked-by: Steve Beattie <steve@nxnw.org>
get_exec_rules_for_path() returns a FileRuleset with all rules matching
the given path.
get_exec_conflict_rules() returns a FileRuleset with all exec rules that
conflict with the given oldrule. This will be used by aa-mergeprof to
ask the user which rule he wants to keep.
Also add tests for both functions.
Acked-by: Steve Beattie <steve@nxnw.org>
The clear_common() call was disabled because it crashed in
delete_path_duplicates(). With the switch to FileRule, this function
no longer exists and therefore it can't crash ;-)
This patch re-enables the clear_common() call to avoid asking
superfluous questions.
References: https://bugs.launchpad.net/apparmor/+bug/1382236
Acked-by: Steve Beattie <steve@nxnw.org>
This is the correct way of doing AARE matches. However, this check is
more strict when matching against an AARE containing wildcards etc.
(which can "by luck" match when doing str matching)
To avoid breaking DbusRule, PtraceRule and SignalRule (especially their
tests), introduce _is_covered_aare_compat() which keeps the previous
behaviour of doing str matching, and use it in these classes.
On the long term, _is_covered_aare_compat() needs to go away, but doing
the changes needed in DbusRule, PtraceRule and SignalRule (or ideally
just in AARE) are out of scope for the FileRule patch series.
Acked-by: Steve Beattie <steve@nxnw.org>
When matching an AARE against another AARE, most AARE objects don't
contain orig_regex (only AARE instances originating from a log event
contain orig_regex).
In this case, match() will use is_equal() to error out on the safe side.
Unfortunately this also means that there are lots of false negative
cases where match() returns False errornously.
With this patch, match() checks the given AARE regex and, if it doesn't
contain any special characters (wildcards, alternations or variables),
handles it as plain path. This avoids most of the false negatives.
Also extend the AARE tests to check a bunch of plain path regexes using
AARE matching instead of only str matching.
Acked-by: Steve Beattie <steve@nxnw.org>
Merge the existing and requested permissions into a nice set of headers
that can be displayed by aa-logprof. This will look like:
Path: /foo
Old Mode: r + owner w
New Mode: rw
Also split off a _join_given_perms() function off _joint_perms() so that
we can use the permission string merging for things not stored in self.*.
Finally add some tests for logprof_header().
Acked-by: Steve Beattie <steve@nxnw.org>
If the audit.log contains an event for a non-existing profile (this can
happen when running with a foreign log or if the user manually deleted a
profile or hat), propose_file_rules() crashes because rule_obj is None
instead of a profile_storage() struct.
This patch adds a check that skips events for non-existing profiles and
hats.
Note: I'm quite sure this happens only for file events (because the
other rule types don't have something similar to propose_file_rules()),
therefore no backport to older versions is needed.
Acked-by: Steve Beattie <steve@nxnw.org>
Add set_options_audit_mode() to switch the audit mode in all options
offered by aa-logprof and aa-mergeprof, not only the "original" rule
(in aa-logprof, this means the non-globbed rule_obj).
As usual, add some tests to ensure the function works as expected.
Acked-by: Steve Beattie <steve@nxnw.org>
aa.py:
- add propose_file_rules() - will propose matching paths from existing
rules in the profile or one of the includes
- save user_globs if user selects '(N)ew' (will be re-used when
proposing rules)
- change user_globs to a dict so that it can carry the human-readable
path and an AARE object for it
- change order_globs() to ensure the original path (given as parameter)
is always the last item in the resulting list
- add a ruletype switch to ask_the_questions() so that it uses
propose_file_rules() for file events (I don't like this
ruletype-specific solution too much, but everything else would make
things even more complicated)
Also keep aa-mergeprof ask_the_questions() in sync with aa.py.
In FileRule, add original_perms (might be set by propose_file_rules())
Finally, add some tests to ensure propose_file_rules() does what it promises.
Acked-by: Steve Beattie <steve@nxnw.org>
get_file_perms() collects the existing permissions for a file from
various rules (exact matches, wildcards) in the main profile and the
included abstractions.
It will be used to get displaying the current permissions back, and
also to propose rules with merged permissions (next patch).
Also add some tests to make sure it does what it promises ;-)
Acked-by: Steve Beattie <steve@nxnw.org>
- get_rules_for_path() returns all rules matching the given path
(both exact matches and AARE matches)
- get_perms_for_path() returns the merged permissions for the given
path and a list of paths used in the matching rules
Also add tests for these two functions.
Acked-by: Steve Beattie <steve@nxnw.org>
Also add a rank_path() function to severity.py and change rank() to call
rank_path() for paths.
Long-term goal: get rid of the type "guessing" in rank()
Finally add some tests, mostly based on test-severity.py SeverityTest
Acked-by: Steve Beattie <steve@nxnw.org>
This brings back the edit option for the path of file rules.
Also add it to aa-mergeprof to keep ask_the_questions() in sync.
Note: aa-mergeprof will ask about path mismatchs basically always.
That's because AARE is too careful on the matching - something to be
fixed in a later patch.
Acked-by: Steve Beattie <steve@nxnw.org>
This means adding
- self.can_edit - True if editing via '(N)ew' should be possible (will
be False for bare file rules)
- edit_header() - returns the prompt text and the current path
- validate_edit() - checks if the new path matches the original one
- store_edit() - changes the path to the new one (even if it doesn't
match the old one)
self.can_edit and the 3 functions are also added to BaseRule:
- can_edit is False by default
- the functions raise a NotImplementedError
Also add tests for the added code.
Acked-by: Steve Beattie <steve@nxnw.org>
This change also needs some other changes in ask_the_questions():
- set q.options and q.selected inside the loop (because glob() and
glob_ext() add another option)
- set 'selection' outside the if block to avoid doing it in nearly every
if branch
- make sure to add the selected rule, not just rule_obj (which doesn't
contain a modified, for example globbed, rule)
- skip 'deny' if an #include is selected
- re-add handling for CMD_GLOB and CMD_GLOB_EXT (was lost when switching
to FileRule)
- add selection_to_rule_obj() helper function
- add glob and glob with ext buttons in available_buttons() if
rule_obj.can_glob or rule_obj.can_glob_ext
Also apply the changes in ask_the_questions() to aa-mergeprof to keep it
in sync with aa.py, and disable the old path handling in aa-mergeprof.
Note: in its current state, aa-mergeprof will ask for some "superfluous"
file permissions, and doesn't check for 'x' conflicts. One of the
following patches will fix that.
Acked-by: Steve Beattie <steve@nxnw.org>
Add the glob() and glob_ext() functions to FileRule, and set
self.can_glob and self.can_glob_ext. Also add some tests (just enough to
make sure the FileRule integration works - the globbing is handled
inside AARE,and the AARE tests contain more testcases).
Note that the implementation differs from the original plan (which was
to have globbing in *Ruleset). Therefore add can_glob and can_glob_ext
to BaseRule (both default to False), and add a comment to BaseRuleset
that globbing needs to be removed from all *Ruleset classes.
Acked-by: Steve Beattie <steve@nxnw.org>
As discussed, I added a pointer to the test-aare.py globbing tests in
test-file.py.
glob_path() and glob_path_ext() modify a (path) regex, so move them to
AARE. Also change them to use self.regex instead of the newpath
parameter, and to return a new AARE object.
While on it, also add several tests to test-aare.py.
Note: There are still glob_path() and glob_path_ext() calls in aa.py,
but those calls are in a (since the middle of this patch series) dead
code section. pyflakes will complain about them nevertheless ;-)
Acked-by: Steve Beattie <steve@nxnw.org>
This patch changes handle_children() (which asks about exec events) and
ask_the_questions() (which asks everything else) to FileRule. This
solves the "brain split" introduced by the previous patch.
This means aa-logprof and aa-genprof ask useful questions again, and
store the answers at the right place.
In detail, this means (with '-' line number from the diff)
- (391) handle_binfmt(): use FileRule. Also avoid breakage if glob_common()
returns an empty result.
- (484) profile_storage(): drop profile['allow']['path'] and
profile['deny']['path']
- (510) create_new_profile(): switch to FileRule
- (1190..1432) lots of changes in handle_children():
- drop escaping (done in FileRule)
- don't add events with 'x' perms to prelog
- use is_known_rule() instead of profile_known_exec()
- replace several regexes for the selected CMD_* with more readable
'in' clauses. While on it, drop unused parts of the regex.
- use plain 'ix', 'px' (as str) instead of str_to_mode() format
- call handle_binfmt() for the interpreter in ix, Pix and Cix rules
- (1652) ask_the_questions(): disable the old file-specific code
(not dropped because some features aren't ported to FileRule yet)
- (2336) collapse_log():
- convert file log events to FileRule (and add some workarounds and
TODOs for logparser.py behaviour that needs to change)
- disable the old file-specific code (not dropped because merging of
existing permissions isn't ported to FileRule yet)
- (2403) drop now unused validate_profile_mode() and the regexes it used
- (3374) drop now unused profile_known_exec()
Test changes:
- adjust fake_ldd to handle /bin/bash
- change test-aa.py AaTest_create_new_profile to expect FileRule instead
of a path hasher. Also copy the profiles to the tempdir and load the
abstractions that are needed by the test.
(These tests get skipped on py2 because changing
apparmor.aa.cfg['settings']['ldd'] doesn't work for some unknown reason)
Important: Some nice-to-have features are not yet implemented for
FileRule:
- globbing
- (N)ew (allowing the user to enter a custom path)
- displaying and merging of permissions already existing in the profile
This means: aa-logprof works, but it's not as user-friendly as before.
The next patches will fix that ;-)
Also note that pyflakes will fail for ask_the_questions_OLD_FILE_CODE()
because of undefined symbols (aamode, profile, hat). This will be fixed
when the old code gets dropped in one of the later patches.
Acked-by: Steve Beattie <steve@nxnw.org>
Bug: https://launchpad.net/bugs/1569316
Change aa.py to use FileRule and FileRuleset for parsing and saving
profiles.
In detail, this means:
- add 'file' to the list of rule classes to enable it at various places
- store file rules in aa[profile][hat]['file'] (not 'path' as before)
to be consistent with the FileRule name
- drop the no longer needed delete_path_duplicates() - this is now
handled by FileRuleset like in all other rule classes.
(same change in cleanprofile.py)
- replace usage of RE_PROFILE_BARE_FILE_ENTRY and RE_PROFILE_PATH_ENTRY
with FileRule.match()
- drop write_path_rules() and write_paths() and replace them with the
new write_file() function.
- adjust several code sections to use write_file() and 'file' instead of
'path'
FileRule doesn't drop optional keywords ('allow' and 'file'), therefore
adjust cleanprof_test.out to the changed behaviour. (If someone insists
on dropping optional keywords in aa-cleanprof, that's something for a
future patch.)
Also adjust the list of known failures in test-parser-simple-tests.py -
switching to FileRule avoids several test failures (and introduces a few
new ones ;-)
IMPORTANT:
This patch introduces a "brain split" which means
- parsing and writing the profile and aa-cleanprof use the new location
(aa[profile][hat]['file'])
- aa-logprof and aa-genprof still save data to the old location
(aa[profile][hat]['allow']['path']) and probably ask superfluous
questions because there are no rules existing in the old location
TL;DR: don't try aa-logprof or aa-genprof with only this patch applied.
I know this isn't ideal, but still better than an even bigger and
totally unreadable patch ;-)
Acked-by: Steve Beattie <steve@nxnw.org>
aa-logprof needs to check if an exec rule for a given path exists.
This patch adds a __FileAnyExec class to FileRule, as well as ANY_EXEC
(which should be used externally, similar to ALL), and adjusts several
checks to allow it as a special execute mode.
This will allow to use is_covered() (or aa.py is_known_rule()) to find
out if execute is permitted, which replaces aa.py profile_known_exec()
in one of the following patches.
As usual, also add some tests.
Acked-by: Steve Beattie <steve@nxnw.org>
Note: as discussed, I adjusted the comment for 'pass' around line 240.
Patch 14 will drop the RE_PROFILE_PATH_ENTRY and
RE_PROFILE_BARE_FILE_ENTRY import from apparmor.aa.
This would break test-regex_matches.py, therefore
import these regexes from apparmor.regex.
Acked-by: Steve Beattie <steve@nxnw.org>
The parser accepts duplicated execute permissions as long as they don't
conflict. For example,
/bin/foo pxpxpxpx,
is a valid rule.
This patch changes FileRule to also accept those duplicated permissions,
even if it's unlikely to hit them outside of the parser tests ;-)
Also add some tests to make sure the parsing works as expected.
Acked-by: Steve Beattie <steve@nxnw.org>
RE_PATH expected (simplified) '/.+', however this excludes a plain '/'
that can appear in path rules.
This patch changes the regex so that it also matches '/'.
Acked-by: Seth Arnold <seth.arnold@canonical.com>
After dropping the dead code in handle_children(), there's only one use
of contains() left in log_str_to_mode().
This patch changes log_str_to_mode to use mode_contains() and drops the
now unused contains() function.
Acked-by: Seth Arnold <seth.arnold@canonical.com>
The 'exec' handling in handle_children starts with
if do_execute:
if profile_known_exec(...)
continue
which means if profile_known_exec() returns True, the rest of the loop
will be skipped. profile_known_exec() will return True if it finds an
exec rule in the profile or an include (independent of the exec type,
and (thanks to rematchfrag()) even if the path is globbed.
Later in the loop, there are checks for various exec modes - but those
checks can only be reached without an existing x rule, so they'll never
be hit.
This patch removes the dead code in the handle_children() / 'exec' / 'no
existing x rule found' section.
I confirmed that this code is really dead by
a) reading the code and, after being confused
b) two manual aa-logprof runs with coverage enabled - in one of them, I
added some ix, Px and Cx rules, and in the second one, no more exec
rules were needed/asked.
After dropping the dead code, combinedmode and combinedaudit are no
longer used, so we can also drop the code that sets those variables.
Sidenote: this patch drops 2% of the lines in aa.py ;-)
Acked-by: Seth Arnold <seth.arnold@canonical.com>
These classes handle file rules, including file rules with leading
perms, and are meant to replace lots of file rule code in aa.py and
aa-mergeprof.
Note: get_glob() and logprof_header_localvars() don't even look
finalized and will be changed in a later patch. (Some other things will
also be changed or added with later patches - but you probably won't
notice them while reviewing this patch.)
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Acked-by: Kshitij Gupta <kgupta8592@gmail.com> (with some suggestions for a follow-up patch)
v1.1: remove 'and not deny' from a condition in split_perms() to get
more helpful error messages for rules like "deny /foo pix,"
Acked-by: Steve Beattie <steve@nxnw.org>
_is_covered_list() has a sanity check that raises an exception if both
other_value and other_all evaluate to False. This breaks when using
_is_covered_list() for FileRule.perms which can be empty if exec_perms
are specified.
This patch adds an optional parameter that allows to skip the sanity
check.
Acked-by: Seth Arnold <seth.arnold@canonical.com>
For now, use an additional regex RE_PROFILE_FILE_ENTRY to avoid
breakage of the existing code by the added match groups.
The regex includes support for file rules with leading and trailing
permissions as well as bare file rules.
Note: even with the restriction to the permission letters we actually
use, it's in theory still possible that a future additional rule type or
permission letter might lead to additional matches for other rule types.
Therefore the parsing code should check for all other rule types before
matching for file rules.
Acked-by: Seth Arnold <seth.arnold@canonical.com>
File permissions can be an empty list (if only exec permissions are
specified). This patch adds the optional allow_empty_list parameter so
that the function can handle this case.
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
File rules contain some optional details (like leading permissions and
the file keyword) which should be ignored in non-strict mode.
This patch passes through the 'strict' parameter to is_equal_localvars
and adds it as function parameter in all existing rule classes.
It also adjusts test-baserule.py to test with the additional parameter.
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
The features_struct.size variable is used to hold a buffer size and it
is also passed in as the size parameter to read(). It should be a size_t
instead of an int.
A new helper function, features_buffer_remaining(), is created to handle
the two places where the remaining bytes in the features buffer are
calculated.
This patch also changes the size parameter of load_features_dir() to a
size_t to match the same parameter of load_features_file() as well as
the features_struct.size change described above.
Two casts were needed when comparing signed types to unsigned types.
These casts are safe because the signed value is checked for "< 0"
immediately before the casts.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
The load_features_file() function returned an int but calculated the
value by subtracting two pointers. On 64 bit systems, that results in a
64 bit value being represented as a 32 bit type.
Coverity CID #55992
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
In 2011 (r1803), the traceroute profile was changed to also match
/usr/bin/traceroute.db:
/usr/{sbin/traceroute,bin/traceroute.db} {
However, permissions for /usr/bin/traceroute.db were never added.
This patch fixes this.
While on it, also change the /usr/sbin/traceroute permissions from
rmix to the less confusing mrix.
Acked-by: Seth Arnold <seth.arnold@canonical.com> for trunk, 2.10 and 2.9.
https://launchpad.net/bugs/1628745
The following upstream kernel commit changed the semantics of the exec
permission check in the 4.8 kernel:
commit 9f834ec18defc369d73ccf9e87a2790bfa05bf46
Author: Linus Torvalds <torvalds@linux-foundation.org>
Date: Mon Aug 22 16:41:46 2016 -0700
binfmt_elf: switch to new creds when switching to new mm
That change means that the target profile of an exec transition must
have permission to map the binary being executed. This patch fixes
regression test failures while the exec_stack.sh test is running against
4.8 and newer kernels by granting mapping permission to the target
profile.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
For reasons that aren't entirely clear, the action to set
apparmor.aa.cfg['settings']['ldd'] to './fake_ldd' does not actually
work on python2.7, get_reqs() tries to use /usr/bin/ldd anyway (printing
out the contents of apparmor.aa.cfg['settings']['ldd'] after the set
operation shows it to still contain '/usr/bin/ldd' o.O). Therefore, skip
these two tests when running under python2.7.
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: Christian Boltz <apparmor@cboltz.de>
Bug: https://launchpad.net/bugs/1522938
Fixes build error when attempting to build and test the 2.10.95 release
on Ubuntu 14.04:
$ (cd libraries/libapparmor/ && ./autogen.sh && ./configure && \
make && make check) > /dev/null
...
libtool: Version mismatch error. This is libtool 2.4.6 Debian-2.4.6-0.1, but the
libtool: definition of this LT_INIT comes from libtool 2.4.2.
libtool: You should recreate aclocal.m4 with macros from libtool 2.4.6 Debian-2.4.6-0.1
libtool: and run autoconf again.
make[2]: *** [grammar.lo] Error 63
make[1]: *** [all] Error 2
make: *** [all-recursive] Error 1
The --force option is needed to regenerate the libtool file in
libraries/libapparmor/.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
This is the least invasive solution to the problem I'm trying to solve
right now (Evince not starting in GNOME on Wayland, and probably
similar issues for other GNOME applications I suppose).
At some point, we will probably want to source the wayland abstraction
from other desktop environments' abstractions, or simply from the
X one. Let's come back to it once people using these other desktop
environments on Wayland with AppArmor enabled tell us what policy
change is needed to make it work for them.
profile_A//&:ns1://unconfined (mixed)
this is confusing and can even break some trusted helpers. The unconfined
profile has been special cased and now will report enforce when stacking
with unconfined
profile_A//&:ns1://unconfined (enforce)
This patch fixes the regression tests to work with this change
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Bug: https://launchpad.net/bugs/1521400
In Debian, gnome-session (3.20.1-2)'s changelog reads:
If /etc/gnome/defaults.list was modified by the system administrator,
the file is moved to /etc/xdg/gnome-mimeapps.list during the upgrade.
So we want to at least support /etc/xdg/gnome-mimeapps.list. And
while we're at it, let's support *-mimeapps.list instead of just
gnome-mimeapps.list, in case other desktop environments or derivatives
need such customizations.
In Debian, gnome-session (3.20.1-2)'s changelog reads:
If /etc/gnome/defaults.list was modified by the system administrator,
the file is moved to /etc/xdg/gnome-mimeapps.list during the upgrade.
So we want to at least support /etc/xdg/gnome-mimeapps.list. And while
we're at it, let's support *-mimeapps.list instead of just gnome-mimeapps.list,
in case other desktop environments or derivatives need such customizations.
This turned out to be a simple case of misinterpreting the promptUser()
result - it returns the answer and the selected option, and
"surprisingly" something like
('CMD_ADDHAT', 0)
never matched
'CMD_ADDHAT'
;-)
I also noticed that the new hat doesn't get initialized as
profile_storage(), and that the changed profile doesn't get marked as
changed. This is also fixed by this patch.
References: https://bugs.launchpad.net/apparmor/+bug/1538306
Acked-by: Steve Beattie <steve@nxnw.org> for trunk, 2.10 and 2.9
pyflakes3 doesn't check sys.version and therefore complains about
'unicode' being undefined.
This patch defines unicode as alias of str to make pyflakes3 happy, and
as a side effect, simplifies type_is_str().
Acked-by: Seth Arnold <seth.arnold@canonical.com> for trunk and 2.10.
By calling self.delete() inside the delete_duplicates() loop, the
self.rules list was modified. This resulted in some rules not being
checked and therefore (some, not all) superfluous rules not being
removed.
This patch switches to a temporary variable to loop over, and rebuilds
self.rules with the rules that are not superfluous.
This also fixes some strange issues already marked with a "Huh?" comment
in the tests.
Acked-by: Seth Arnold <seth.arnold@canonical.com> for trunk and 2.10.
Note that in 2.10 cleanprof_test.* doesn't contain a ptrace rule,
therefore the cleanprof_test.out change doesn't make sense for 2.10.
Network events can come with an operation= that looks like a file event.
Nevertheless, if the event has a typical network parameter (like
net_protocol) set, make sure to store the network-related flags in ev.
This fixes the test failure introduced in my last commit.
Acked-by: Kshitij Gupta <kgupta8592@gmail.com> for trunk, 2.10 and 2.9
We already ignore network events that look like file events (based on
the operation keyword) if they have a request_mask of 'send' or
'receive' to avoid aa-logprof crashes because of "unknown" permissions.
It turned out that both can happen at once, so we should also ignore
this case.
Also add the now-ignored log event as test_multi testcase.
References: https://bugs.launchpad.net/apparmor/+bug/1577051#13
Acked-by: Tyler Hicks <tyhicks@canonical.com> for trunk, 2.10 and 2.9.
patches
0001-0022 are backports of fixes from the 4.8 pull-request
0023-0025 are the out of tree feature patches
Signed-off-by: John Johansen <john.johansen@canonical.com>
https://launchpad.net/bugs/1584069
This patch adds support for the safe and unsafe exec modes for
change_profile rules. The logic is pretty simple at this point because
the kernel's default for exec modes changed in newer versions.
Therefore, this patch simply retains any specified exec mode in parsed
rules. If an exec mode is not specified in a rule, there is no attempt
to force the usage of "safe" because older kernels do not support it.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
The onexec.sh test has periodically exhibited unexplicable failures that
are possibly due to race conditions when onexec.sh is verifying the
/proc/PID/attr/{current,exec} values of the process under test. This
patch attempts to solve the flaky test failures by removing the need for
IPC to coordinate between the test script and the test program.
The old onexec test program is removed and the transition test program
is used instead. This allows for the test script to tell the transition
test program what its current and exec procattr labels should be via
command line options.
Since IPC is no longer needed, the signal:ALL allow rule can be dropped
from the test profile. A new allow rule is needed to grant reading of
/proc/*/attr/{current,exec} since transition must verify the contents of
these files.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Add optional command line parameters to the transition test program that
can be used to verify a certain label and/or mode that should be found
in /proc/self/attr/exec.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Date: Tue, 21 Jun 2016 18:18:45 +0100
Subject: abstractions/nameservice: also support ConnMan-managed resolv.conf
Follow the same logic we already did for NetworkManager,
resolvconf and systemd-resolved. The wonderful thing about
standards is that there are so many to choose from.
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Create a set of strict and non-strict abstractions, much like the
existing dbus abstractions, for connecting to the fcitx bus.
Signed-off-by: Jamie Strandboge <jamie@ubuntu.com>
[tyhicks: Wrote commit message]
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
BugLink: https://launchpad.net/bugs/1588069
Currently
change_profile /** -> A,
change_profile unsafe /** -> A,
do not conflict because the safe rules only set the change_profile
permission where the unsafe set unsafe exec. To fix this we have the
safe version set exec bits as well with out setting unsafe exec.
This allows the exec conflict logic to detect any conflicts.
This is safe to do even for older kernels as the exec bits off of the
2nd term encoding in the change_onexec rules are unused.
Test files
tst/simple_tests/change_profile/onx_no_conflict_safe1.sd
tst/simple_tests/change_profile/onx_no_conflict_safe2.sd
by Christian Boltz <apparmor@cboltz.de>
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Tyler Hicks <tyhicks@canonical.com>
Thanks to reading the wrong directory in read_inactive_profiles()
(profile_dir instead of extra_profile_dir), aa-genprof never asked about
using a profile from the extra_profile_dir.
Sounds like an easy fix, right? ;-)
After fixing this (last chunk), several other errors popped up, one
after the other:
- get_profile() missed a required parameter in a serialize_profile() call
- when saving the profile, it was written to extra_profile_dir, not to
profile_dir where it (as a now-active profile) should be. This is
fixed by removing the filename from existing_profiles{} so that it can
pick up the default name.
- CMD_FINISHED (when asking if the extra profile should be used or a new
one) behaved exactly like CMD_CREATE_PROFILE, but this is surprising
for the user. Remove it to avoid confusion.
- displaying the extra profile was only implemented in YaST mode
- get_pager() returned None, not an actual pager. Since we have 'less'
hardcoded at several places, also return it in get_pager()
Finally, also remove CMD_FINISHED from the get_profile() test in
test-translations.py.
(test-translations.py is only in trunk, therefore this part of the patch
is obviously trunk-only.)
Acked-by: Seth Arnold <seth.arnold@canonical.com> for trunk
Acked-by: John Johansen <john.johansen@canonical.com> for trunk + a 50% ACK for 2.10 and 2.9
Acked-by: Kshitij Gupta <kgupta8592@gmail.com> for trunk, 2.10 and 2.9
The opt_unsafe token was being used to represent 'safe' and 'unsafe' so
it is renamed to opt_exec_mode. Create helpfully named macros to compare
opt_exec_mode's value against instead of hard-coded '0', '1', and '2'
values.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
Add logic to the at_secure.sh test script to verifies that the parser is
new enough to support change_profile exec modes and determine what the
kernel's support for change_profile exec modes before verifying that
AT_SECURE is set correctly after various exec transitions.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
The gen_change_profile() function must be changed to allow the extra
condition in change_profiles rules.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
Simple tests that validate the parser's ability to handle change_profile
rules containing an exec mode.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
https://launchpad.net/bugs/1584069
This patch allows policy authors to specify how exec transitions should
be handled with respect to setting AT_SECURE in the new process'
auxiliary vector and, ultimately, having libc scrub (or not scrub) the
environment.
An exec mode of 'safe' means that the environment will be scrubbed and
this is the default in kernels that support AppArmor profile stacking.
An exec mode of 'unsafe' means that the environment will not be scrubbed
and this is the default and only supported change_profile exec mode in
kernels that do not support AppArmor profile stacking.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
Quote $@ so that the for loop doesn't iterate on the space-delimited
version of the rule(s) under test. This allows more complex rules such
as "change_profile foo -> bar," to be tested where, before this patch,
only "change_profile," could be tested.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
Make future modifications to the change_profile grammar rules easier by
simplifying things. First, the change_profile rule handling is collapsed
into a single grammar rule. The inputs to the grammar rule are given
helpful variable names to make it harder to mix up which variable we're
dealing with. Finally, the two separate calls to new_entry() are unified
into a single call.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
The AT_SECURE value in the kernel's per-process auxiliary vector is what
signals to libc that the process' environment should be scrubbed. This
new set of regression tests checks the AT_SECURE value after performing
the various types of exec transitions that AppArmor supports (file rules
with different exec access modes and change_profile rules).
Different versions of the kernel handle AT_SECURE differently with
respect to change_profile rules. This change in behavior was introduced
in the AppArmor profile stacking kernel support and the tests are
conditionalized to account for this change.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
Extend the transition test program to allow for changing to a new
profile. This change will be useful in test scripts that need to test
operations across profile stacks and/or profile changes.
The calls to aa_stack_onexec() and aa_stack_profile() are build-time
conditionalized on whether or not the libapparmor being used has
implemented those functions.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
This test will soon be made to do more than just stack a new profile.
It will be extended to allow for changing to a new profile and,
therefore, should be renamed.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
This patch includes several changes and fixes in change_profile highlighting:
- allow audit and deny keywords
- allow bare change_profile rules
- allow change_profile rules without '-> ...' part
- allow usage of the new 'safe' and 'unsafe' keywords
- ensure the exec condition starts with / or @
Acked-by: Seth Arnold <seth.arnold@canonical.com>
This commit touches up the .po files that generate warnings
when msgfmt processes them to create .mo files, at least with gettext
0.19.7-2ubuntu3 in Ubuntu 16.04 LTS. Example warning types cleaned up
include:
ce.po:7: warning: header field 'Last-Translator' still has the initial default value
ce.po:7: warning: header field 'Language' missing in header
de.po:6: warning: header field 'Language-Team' still has the initial default value
This commit also fixes up po files where the Report-Msgid-Bugs-To:
field had not been updated, setting it with the email address
'AppArmor list <apparmor@lists.ubuntu.com>'
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
In detail, this means:
- handle ptrace events in logparser.py
- "translate" those events in aa.py - from log (logparser.py readlog())
to prelog (handle_children()) to log_dict (collapse_log()))
- finally ask the user about the ptrace in ask_the_questions()
(no code change needed there)
Note that these changes are not covered by tests, however they worked in
a manual test with the log examples in the libapparmor testsuite.
Unfortunately there's no example log for eavesdrop, so it might be a
good idea to a) add such a log line and b) test with it
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Note: as discussed on #apparmor, I changed the mapping of peer_profile so
that it ends up in peer=(label=...) instead of the wrong peer=(name=...).
"Everywhere" means aa-mergeprof and aa-cleanprof. In theory also
aa-logprof, but that needs some code that parses dbus log events ;-)
Also add some dbus rules to the aa-cleanprof test profiles to ensure
superfluous dbus rules get deleted.
Acked-by: Seth Arnold <seth.arnold@canonical.com>
DBUS_Rule (in rules.py) was added in r2424 as a "this is how it should
look like" proof of concept, but was never used.
We have a "real" class for dbus rules now, so we can drop the proof of
concept class.
Also remove a commented, old version of RE_DBUS_ENTRY from aa.py
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Change aa.py to use DbusRule and DbusRuleset in profile_storage,
parse_profile_data() and write_dbus. This also means we can drop the
now unused parse_dbus_rule() and write_dbus_rules() functions.
Raw_DBUS_Rule in rules.py is now also unused and can be dropped.
Also shorten the list of known-failing tests in
test-parser-simple-tests.py. Even if the list of removals doesn't look
too long, the generated_dbus/* removals mean 1989 tests now cause the
expected failures.
OTOH, I had to add 4 tests to the known-failing list:
- 3 tests with a "wrong" order of the conditionals which the parser
accepts (which is slightly surprising, because usually we enforce the
order of rule parts)
- one test fails because the path in the path= conditional doesn't start
with / or a variable. Instead, it starts with an alternation, which
wouldn't be allowed in file rules.
Those 4 failures need more investigation, but shouldn't block this
patchset.
Finally, adjust test-regex_matches.py to import RE_PROFILE_DBUS from
apparmor.regex instead of apparmor.aa.
Acked-by: Seth Arnold <seth.arnold@canonical.com>
The tests include the two tests from test-dbus_parse.py, therefore
delete this file.
As usual, we have 100% coverage :-)
Also addd an explicit str() conversion to common_test.py to avoid
TypeError: not all arguments converted during string formatting
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Those classes will be used to parse and handle dbus rules.
They understand the syntax of dbus rules.
Note that get_clean() doesn't output superfluos things, so
dbus ( send ),
will become
dbus send,
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Some dbus rule conditionals come with optional parenthesis. Instead of
making the regex even more complicated, use a small function to strip
those parenthesis.
Also add some tests for strip_parenthesis() to test-regex.py.
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
As a preparation for the DbusRule class, add a <details> match group
to RE_PROFILE_DBUS.
Also adjust test-regex_matches.py for the added group.
Note: RE_PROFILE_DBUS is only used in aa.py, and only matches[0..2]
are used. 0 and 1 are audit and allow/deny and 2 is and stays the whole
rule (except audit and allow/deny). Therefore no aa.py changes are
needed.
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
The latest iputils merged ping and ping6 into a single binary that does
both IPv4 and IPv6 pings (by default, it really does both).
This means we need to allow network inet6 raw in the ping profile.
References: https://bugzilla.opensuse.org/show_bug.cgi?id=980596
(contains more details and example output)
Acked-by: Steve Beattie <steve@nxnw.org> for trunk, 2.10 and 2.9
From: Simon McVittie <simon.mcvittie@collabora.co.uk>
Date: Wed, 4 May 2016 13:48:36 +0100
Subject: dbus-session-strict: allow access to the user bus socket
If dbus is configured with --enable-user-bus (for example in the
dbus-user-session package in Debian and its derivatives), and the user
session is started with systemd, then the "dbus-daemon --session" will be
started by "systemd --user" and listen on $XDG_RUNTIME_DIR/bus. Similarly,
on systems where dbus-daemon has been replaced with kdbus, the
bridge/proxy used to provide compatibility with the traditional D-Bus
protocol listens on that same socket.
In practice, $XDG_RUNTIME_DIR is /run/user/$uid on all systemd systems,
where $uid represents the numeric uid. I have not used /{var/,}run here,
because systemd does not support configurations where /var/run and /run
are distinct; in practice, /var/run is a symbolic link.
Based on a patch by Sjoerd Simons, which originally used the historical
path /run/user/*/dbus/user_bus_socket. That path was popularized by the
user-session-units git repository, but has never been used in a released
version of dbus and should be considered unsupported.
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
From: Simon McVittie <simon.mcvittie@collabora.co.uk>
Date: Wed, 11 May 2016 13:52:56 +0100
Subject: syscall_sysctl test: correctly skip if CONFIG_SYSCTL_SYSCALL=n
This test attempts to auto-skip the sysctl() part if that syscall
was not compiled into the current kernel, via
CONFIG_SYSCTL_SYSCALL=n. Unfortunately, this didn't actually work,
for two reasons:
* Because "${test} ro" wasn't in "&&", "||", a pipeline or an "if",
and it had nonzero exit status, the trap on ERR was triggered,
causing execution of the error_handler() shell function, which
aborts the test with a failed status. The rules for ERR are the
same as for "set -e", so we can circumvent it in the same ways.
* Because sysctl_syscall.c prints its diagnostic message to stderr,
but the $() operator only captures stdout, it never matched
in the string comparison. This is easily solved by redirecting
its stderr to stdout.
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Variables can be used in several rule types (from the existing *Rule
classes: change_profile, dbus, ptrace, signal). It seems nobody uses
variables with those rules, otherwise we'd have received a bugreport ;-)
I noticed this while working on FileRule, where usage of variables is
more common. The file code in bzr (not using a *Rule class) already
loads the variables, so old versions don't need changes for file rule
handling.
However, 2.10 already has ChangeProfileRule and therefore also needs
this fix.
Acked-by: Seth Arnold <seth.arnold@canonical.com> for trunk and 2.10.
While running test-translations.py with the fixed german translations,
I noticed that I still get errors about hotkey conflicts
It turned out that test-translations.py reads the system-wide
apparmor-utils.mo in addition to the in-tree translations.
(I have the 2.11 beta1 translations installed, which contain hotkey
conflicts for the german translations).
This is surprising because test-translations.py explicitely sets the
locale path. Interestingly, this happens only 4 times (checked with a
temp profile with audit for those files) while test-translations.py has
9 tests).
(Any idea if this behaviour is normal or a bug?)
This patch adds LC_ALL=C to the make check and make coverage commandline
so that the system-wide translations don't get used.
I checked with a modified de.po that in-tree hotkey conflicts still get
detected.
Acked-by: Seth Arnold <seth.arnold@canonical.com>
This test builds and installs the apparmor-utils translations into a
tempdir, and then checks if there's any hotkey conflict in one of the
languages. This is based on a manually maintained list of "buttons" that
are displayed at the same time.
To make things a bit easier to test, add CMD_CANCEL to ui.py CMDS[].
Also replace hardcoded usage of '(Y)es', '(N)o' and '(C)ancel' with
CMDS['CMD_YES'], CMDS['CMD_NO'] and CMDS['CMD_CANCEL'].
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Matthew Dawson explained why:
> sshd doesn't actually require the net_admin capability. libpam-systemd tries
> to use it if available to set the send/receive buffers size, but will fall
> back to a non-privileged version if it fails.
https://lists.ubuntu.com/archives/apparmor/2016-April/009586.html
BugLink: http://bugs.launchpad.net/bugs/1551950
The apparmor_parser is incorrectly outputting the names of child profiles
and hats, by adding a : between the parent and the child profile name
Eg.
/usr/sbin/httpd{,2}-prefork
/usr/sbin/httpd{,2}-prefork://DEFAULT_URI
/usr/sbin/httpd{,2}-prefork://HANDLING_UNTRUSTED_INPUT
instead of what it should be
/usr/sbin/httpd{,2}-prefork
/usr/sbin/httpd{,2}-prefork//DEFAULT_URI
/usr/sbin/httpd{,2}-prefork//HANDLING_UNTRUSTED_INPUT
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
https://launchpad.net/bugs/1569316
When Ubuntu made the jump from network-manager 1.0.4 to 1.1.93, the
dnsmasq process spawned from network-manager started hitting a
disconnected path denial:
audit: type=1400 audit(1460463960.943:31702): apparmor="ALLOWED"
operation="connect" info="Failed name lookup - disconnected path"
error=-13 profile="/usr/sbin/dnsmasq"
name="run/dbus/system_bus_socket" pid=3448 comm="dnsmasq"
requested_mask="wr" denied_mask="wr" fsuid=65534 ouid=0
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
BugLink: http://bugs.launchpad.net/bugs/1566490
This patch enables to parser to scale the max jobs if new resources are
being brought online by the scheduler.
It only enables the scaling check if there is a difference between the
maximum number of cpus (CONF) and the number of online (ONLN) cpus.
Instead of checking for more resources regardless, of whether the online
cpu count is increasing it limits its checking to a maximum of
MAX CPUS + 1 - ONLN cpus times. With each check coming after fork spawns a
new work unit, giving the scheduler a chance to bring new cpus online
before the next check. The +1 ensures the checks will be done at least
once after the scheduling task sleeps waiting for its children giving
the scheduler an extra chance to bring cpus online.
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Since the latest openSUSE Tumbleweed update (dovecot 2.2.21 -> 2.2.22),
dovecot/auth writes to /var/run/dovecot/stats-user.
Acked-by: Seth Arnold <seth.arnold@canonical.com> for trunk, 2.10 and 2.9.
parser 'make install' failed if 'make' wasn't run before. This patch
adds the missing dependency 'install-indep: indep'.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
binutils 'make install' failed if 'make' wasn't run before.
This patch adds the missing dependency 'install-indep: indep'
Acked-by: Seth Arnold <seth.arnold@canonical.com>
The binutils (aa-enabled and aa-exec) get installed into /usr/bin/ and
are meant to be used by non-root users. Therefore the manpages should be
in section 1 instead of 8 (which is for sysadmin commands).
Acked-by: Seth Arnold <seth.arnold@canonical.com>
The stacking tests worked fine when using in-tree programs and libraries
but the tests unexpectedly failed when USE_SYSTEM=1 was specified. This
patch makes use of the addimage:$test argument to mkprofile.pl to
generate the correct file permissions needed to use the system binaries.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
exec choices are stored in transitions[], but that's never used
(and I don't see a need for it), therefore stop storing it.
Note: hat choices (CMD_ADDHAT, CMD_USEDEFAULT and CMD_DENY) get still
stored in transitions[], and that information is used if the same hat
name appears again.
Acked-by: Steve Beattie <steve@nxnw.org>
acmetool is an alternative client for Let's Encrypt.
(https://github.com/hlandau/acme/)
It stores the certificates etc. in the following directory layout:
/var/lib/acme/live/<domain> -> ../certs/<hash>
/var/lib/acme/certs/<hash>/cert
/var/lib/acme/certs/<hash>/chain
/var/lib/acme/certs/<hash>/privkey -> ../../keys/<hash>/privkey
/var/lib/acme/certs/<hash>/url
/var/lib/acme/certs/<hash>/fullchain
/var/lib/acme/keys/<hash>/privkey
This patch adds the needed permissions to the ssl_certs and ssl_keys
abstractions so that the certificates can be used.
Acked-by: Seth Arnold <seth.arnold@canonical.com> for trunk, 2.10 and 2.9.
I configured the stacking test binary to only be built when libapparmor
2.11 is present. The versioning of the 2.11 Beta 1 release (2.10.95)
causes that check to fail and the stacking tests to not be used.
This patch adjusts the libapparmor version check to be aware of the 2.11
Beta 1 versioning.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
Automated infrastructure management tools, such as Chef, Puppet, and so
on, could use a way to check AppArmor status that is both high-level
(meaning it does not rely on kernel interfaces in /proc) and machine-
readable (meaning it does not require the complexity of parsing output
of tools originally intended for human consumption).
Adding a JSON variant of the standard aa-status output achieves both.
In /etc/nscd.conf there is an option allowing to restart nscd after a
certain time. However, this requires reading /proc/self/cmdline -
otherwise nscd will disable paranoia mode.
References: https://bugzilla.opensuse.org/show_bug.cgi?id=971790
Acked-By: Jamie Strandboge <jamie@canonical.com> for trunk, 2.10 and 2.9
The debugging code for profile entries contains a check to ensure that
it's not NULL, but the list iterator macro already ensures that the
iteration will stop if the item is NULL, making the check redundant.
Coverity CID #55983
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: Tyler Hicks <tyhicks@canonical.com>
It's possible to end up unreferencing a kernel_interface object that
has ->dirfd set to -1. This patch avoids calling close(2) on that fd.
(close(-1) will just return EBADF anyway.)
Coverity CIDs #55996 and #55997
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: Tyler Hicks <tyhicks@canonical.com>
This makes some of the references to functions in the aa_query_label(2)
manpage more consistent and fixes a couple of grammar issues. It also
tries to make the qualifying statements in apparmor.d(5) more distinct,
and also fixes some typos there as well.
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: Tyler Hicks <tyhicks@canonical.com>
This adds support to the profile generator script for change_profile
rules, giving the ability to write the 3 factor version of the rule
(e.g. "change_profile /t -> A_PROFILE") which was significantly more
difficult using straight raw rules, which is why we don't have any 3
factor rule tests.
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: Tyler Hicks <tyhicks@canonical.com>
Now that the onexec test program notices that it failed to send SIGSTOP
to itself, causing a whole bunch of tests to be detected as failing,
grant the ability to send and receive signals to the onexec tests.
(The onexec tests are not tests intended to verify signal mediation.)
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: Tyler Hicks <tyhicks@canonical.com>
The onexec test was ignoring errors from the kill() call, so it didn't
notice when it had failed to send SIGSTOP to itself.
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: Tyler Hicks <tyhicks@canonical.com>
Based on a patch by John Johansen <john.johansen@canonical.com>
(I converted the check to look for the process directory in /proc
rather than sending signal 0 to the task, as John had done in a patch
sent to me, to prevent failures in signal delivery from blocking the
check from working correctly.)
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: Tyler Hicks <tyhicks@canonical.com>
Based on a patch by John Johansen <john.johansen@canonical.com>
Add more details to the checks in the regression tests onexec tests,
to make debugging failures easier. Also, use more local variables
to indicate what and how many arguments are expected to the onexec
check_* functions.
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: Tyler Hicks <tyhicks@canonical.com>
The aa_stack_profile() and aa_stack_onexec() functions were added to
libapparmor since 2.10.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
The audit_read capability, mpls address family, and profile stacking are
all new features advertised by the latest AppArmor kernel features file.
Without this change, the parser tests will fail because parsing profiles
that utilize stacking results in an error when the features file
indicates that stacking is not supported by the kernel.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
The stacking test binary links against libapparmor for
aa_stack_profile() and aa_stack_onexec(), which will be present in 2.11.
This means that regression test builds using the system libapparmor
should not build the stacking test binary unless the libapparmor 2.11 or
newer is present.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
Policy namespaces are not well supported in older parsers and kernels.
This is a case where the kernel support doesn't seem to be working.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
Stacking is a complex feature and, in order to sufficiently test all
aspects of stacking, a relatively complex test program is needed.
This patch adds a program that can call
aa_stack_onexec()/aa_stack_profile(), perform file IO on a given file
path, verify that the current confinement context is what it is expected
to be, and/or execute itself or another program.
The confinement context verification can handle stacked labels with any
ordering.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
Some tidying up is needed in order to reuse do_open(). This patch
eliminates the chance of returning 0 due to errno being not set. It also
adjusts the file string to be const.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
The idea is that the $test profile grants $file access and the
$othertest profile grants $subfile access. Both profiles grant
$stacktest access. The tests verify that after changing to the stacked
$othertest//&$test profile, only $stacktest can be accessed.
Similar tests are also added for stacking with a namespaced profile.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
The kernel patches that implement AppArmor profile stacking made changes
that allow the the backed for change_profile to detect if the target
profile does not exist prior to checking if the current profile allows
the change_profile.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
Check if the current kernel supports stacking. If not, ensure that named
transitions (exec, change_profile, etc.) do not attempt to stack their
targets.
Also, set up the change_profile vector according to whether or not the
kernel supports stacking. Earlier kernels expect the policy namespace to
be in its own NUL-terminated vector element rather than passing the
entire label (namespace and profile name) as a single string to the
kernel.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Allow for a leading '&' character to be present in the named transition
target strings to indicate that the transition should stack the current
profile with the specified profile.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
The parser was splitting up the namespace and profile name from named
transition targets only to rejoin it later when creating the binary
policy. This complicated the changes needed to support the stacking
identifier '&' in named transition targets.
To keep the stacking support simple, this patch keeps the entire named
transition target string intact from initial profile parsing to writing
out the binary.
All of these changes are straightforward except the hunk that removes
the namespace string addition to the vector in the process_dfa_entry()
function. After speaking with John, kernels with stacking have support
for consuming the namespace with the profile name.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
This patch separates the label parsing functionality from the program
termination and memory allocation duties of parse_label(). This will
ultimately help in creating simple helper functions that simply need to
check if a label contains a namespace.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
The add_named_transition function was written in a way that is difficult
to understand while attempting to read the function. This patch attempts
to clean it up.
First, this patch removes this confusing code flow issue:
if (!entry->ns) { ... }
if (entry->ns) { ... } else { ... }
It then unifies the way that the ns and nt_name strings of the cod_entry
struct are handled prior to calling add_entry_to_x_table() and/or
returning. ns and nt_name are now guaranteed to be NULL before
performing either of those actions.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
The copy_cod_entry() function was not copying the nt_name field of the
cod_entry struct.
This was discovered during code review and I'm not certain if it causes
any real world bugs.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
Modeled after the aa_change_profile(2) man page, this profile defines
the libapparmor and kernel interfaces for the in-progress profile
stacking feature.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
Bug: https://bugs.launchpad.net/ubuntu/+source/apparmor/+bug/1480492
If python3-apparmor is not installed, aa-status aborts due to the added
import to handle fancier exception handling failing. This patch makes
aa-status(8) work even in that case, falling back to normal python
exceptions, to keep its required dependencies as small as possible.
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
write_prof_data[hat] is correct (it only contains one profile, see bug 1528139),
write_prof_data[profile][hat] is not and returns an empty (sub)hasher.
This affects RE_PROFILE_START and RE_PROFILE_BARE_FILE_ENTRY.
Acked-by: Kshitij Gupta <kgupta8592@gmail.com> for trunk, 2.9 and 2.10
Instead of reusing opt_named_transition and be forced to reconstruct the
target path when is looks like ":odd:target", create simpler grammer
rules that have nothing to do with named transitions and namespaces.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
a) change log_dict to profile_storage()
Change collapse_log() to initialize log_dict[aamode][profile][hat]
as profile_storage() instead of a hasher().
This also means path events need to go into
log_dict[aamode][profile][hat]['allow']['path']
instead of
log_dict[aamode][profile][hat]['path']
to match the profile_storage() layout.
b) Simplify log translation
The translation from logparser.py's output to *Rule events was more ugly
than needed. This patch removes one step.
Instead of translating log_dict to log_obj in ask_the_questions(), add
*Rule objects to log_dict and adjust ask_the_questions() to use log_dict
instead of log_obj.
This also means log_obj in ask_the_questions() is now superfluous and
can be removed.
c) Other small changes:
- use is_known_rule() instead of .is_covered() for capability events,
which means included files are also checked now.
- remove the "if rule_obj.log_event != aamode:" check, because
a) it depends on the content of *Rule.log_event (which means it
ignores events with log_event != 'ALLOWING' or 'REJECTING'
b) it's superfluous because the whole code section is wrapped in a
"for aamode in sorted(log.dict.keys())" which means we have
separate loops for enforce and complain mode already
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
If the program specified as get_output param isn't executable or doesn't
exist at all, get_output() returns with ret = -1.
Raising an exception looks like a better option, especially because
other possible exec failures already raise an exception ("Unable to
fork").
Note: get_output is only used by get_reqs() which also does the
os.access() check for x permissions (and raises an exception), so in
practise raising an exception in get_output() doesn't change anything.
This change also allows to rewrite and simplify get_output() quite a bit.
Another minor change (and fix) is in the removal of the last line. The
old code removed the last line if output contained at least two items.
This had two not-so-nice effects:
- an empty output resulted in [''] instead of []
- if a command didn't add a \n on the last line, this line was deleted
nevertheless
The patch changes that to always remove the last line if it is empty,
which fixes both issues mentioned above.
Also add a test to ensure the exception is really raised, and adjust the
test that expects an empty stdout.
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
To make these tests independent from the underlaying system, add a
fake_ldd script that provides hardcoded ldd output for the "known"
executables and libraries.
To avoid interferences with the real system (especially symlinks), all
paths in fake_ldd have '/AATest' prepended.
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
To ensure aa-cleanprof works as expected (and writing the rules works
as expected), add some rules for every rule class to the cleanprof.in
and cleanprof.out test profiles.
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
According to a discussion with John on IRC, denied_mask="x" can only
happen for 'exec' log events. This patch raises an exception if John
is wrong ;-)
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
This should happen rarely, but nevertheless it can happen - and since
AppArmor needs the symlink target in the profile, we have to resolve all
symlinks.
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
If a profile file contains multiple profiles and one of those profiles
contains a rule managed by a *Ruleset class,
serialize_profile_from_old_profile() crashes with an AttributeError.
This happens because profile_data / write_prof_data contain only one
profile with its hats, which explodes if a file contains multiple
profiles, as reported in lp#1528139
Fixing this would need lots of
write_prof_data[hat] -> write_prof_data[profile][hat]
changes (and of course also a change in the calling code) or, better
option, a full rewrite of serialize_profile_from_old_profile().
Unfortunately I don't have the time to do the rewrite at the moment (I
have other things on my TODO list), and changing write_prof_data[hat] ->
write_prof_data[profile][hat] is something that might introduce more
breakage, so I'm not too keen to do that.
Therefore this patch wraps the serialize_profile_from_old_profile() call
in try/except. If it fails, the diff will include an error message and
recommend to use 'View Changes b/w (C)lean profiles' instead, which is
known to work.
Note: I know using an error message as 'newprofile' isn't an usual way
to display an error message, but I found it more intuitive than
displaying it as a warning (without $PAGER).
References: https://bugs.launchpad.net/ubuntu/+source/apparmor/+bug/1528139
Acked-by: Seth Arnold <seth.arnold@canonical.com> for trunk and 2.10
dovecot-lda needs to read and write /tmp/dovecot.lda.*.
It also needs to be able to execute sendmail to send sieve vacation
mails.
For now, I'm using a child profile for sendmail to avoid introducing a
new profile with possible regressions. This child profile is based on
the usr.sbin.sendmail profile in extras and should cover both postfix'
and sendmail's sendmail.
I also mixed in some bits that were needed for (postfix) sendmail on my
servers, and dropped some rules that were obsolete (directory rules not
ending with a /) or covered by an abstraction.
In the future, we might want to provide a stand-alone profile for
sendmail (based on this child profile) and change the rule in the
dovecot-lda profile to Px.
References: https://bugzilla.opensuse.org/show_bug.cgi?id=954959https://bugzilla.opensuse.org/show_bug.cgi?id=954958
Acked-by: Seth Arnold <seth.arnold@canonical.com> for trunk, 2.10 and 2.9.
parser/tst/simple_tests/profile/profile_ns_bad8.sd was added in r3376
(trunk) / r3312 (2.10 branch) and contains the profile name ':ns/t'
which misses the terminating ':' for the namespace.
Unfortunately the tools don't understand namespaces yet and just use the
full profile name. This also means this test doesn't fail as expected
when tested against the utils code.
This patch adds profile_ns_bad8.sd to the exception list of
test-parser-simple-tests.py.
Acked-by: Steve Beattie <steve@nxnw.org> for trunk and 2.10.
https://launchpad.net/bugs/1544387
Don't split namespaces from profile names using YACC grammar. Instead,
treat the entire string as a label in the grammer. The label can then be
split into a namespace and a profile name using the new parse_label()
function.
This fixes a bug that caused the profile keyword to not be used with a
label containing a namespace in the profile declaration.
Fixing this bug uncovered a bad parser test case at
simple_tests/profile/profile_ns_ok1.sd. The test case mistakenly
included two definitions of the :foo:unattached profile despite being
marked as expected to pass. I've adjusted the name of one of the
profiles to :foo:unattached2.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
https://launchpad.net/bugs/1546455
Don't filter out AF_UNSPEC from the list of valid protocol families so
that the parser will accept rules such as 'network unspec,'.
There are certain syscalls, such as socket(2), where the LSM hooks are
called before the protocol family is validated. In these cases, AppArmor
was emitting denials even though socket(2) will eventually fail. There
may be cases where AF_UNSPEC sockets are accepted and we need to make
sure that we're mediating those appropriately.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Suggested-by: Steve Beattie <steve@nxnw.org>
Acked-by: John Johansen <john.johansen@canonical.com>
[cboltz: Add 'unspec' to the network domain keywords of the utils]
If a profile file contains multiple profiles, aa-mergeprof crashes on
saving in write_profile() because the second profile in the file is not
listed in 'changed'. (This happens only if the second profile didn't
change.)
This patch first checks if 'changed' contains the profile before
pop()ing it.
Reproducer: copy utils/test/cleanprof_test.in to your profile directory
and run aa-mergeprof utils/test/cleanprof_test.out. Then just press
's' to save the profile.
Acked-by: Kshitij Gupta <kgupta8592@gmail.com> for trunk, 2.10 and 2.9
If autodep() is called with a pname starting with / (which can happen
for (N)amed exec depending on the user input), this pname is mapped to
bin_name.
This might look like a good idea, however if the given pname doesn't
exist as file on-disk, autodep() returns None instead of a (mostly
empty) profile. (Reproducer: choose (N)amed, enter "/foo/bar")
Further down the road, this results in two things:
a) the None result gets written as empty profile file (with only a "Last
modified" line)
b) a crash if someone chooses to add an abstraction to the None, because
None doesn't support the delete_duplicates() method for obvious
reasons ;-)
Unfortunately this patch also introduces a regression - aa-logprof now
fails to follow the exec and doesn't ask about the log events for the
exec target anymore. However this doesn't really matter because of a) -
asking and saving to /dev/null vs. not asking isn't a real difference ;-)
Actually the patch slightly improves things - it creates a profile for
the exec target, but only with the depmod() defaults (abstractions/base)
and always in complain mode.
I'd prefer a patch that also creates a complete profile for the exec
target, but that isn't as easy as fixing the issues mentioned above and
therefore is something for a future fix. To avoid we forget it, I opened
https://bugs.launchpad.net/apparmor/+bug/1545155
Note: 2.9 "only" writes an empty file and doesn't crash - but writing
an empty profile is still an improvement.
Acked-by: Kshitij Gupta <kgupta8592@gmail.com> for trunk, 2.10 and 2.9
deny rules don't allow ix, Px, Ux etc. - only 'deny /foo x,' is allowed.
Acked-by: Seth Arnold <seth.arnold@canonical.com> for trunk and 2.10
Note: Seth mentioned in the mail that he doesn't like the 'deny x'
section too much, but we didn't find a better solution when discussing
it on IRC. Therefore I keep the patch unchanged, but will happily
review a follow-up patch if someone sends one ;-)
This test causes `make check` to fail but it is known bug so mark it as
a TODO test.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
According to the discussion with John on IRC, exec log events for
directories should never happen, therefore let handle_children()
raise an exception.
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
Most probably-file log events can also be network events. Therefore
check for request_mask in all events, not only file_perm, file_inherit
and (from the latest bugreport) file_receive.
References: https://bugs.launchpad.net/apparmor/+bug/1540562
Acked-by: Kshitij Gupta <kgupta8592@gmail.com> for trunk, 2.10 and 2.9.
support systems that use libnl-3-200 via libnss-gw-name.
Patch initially proposed by Simon McVittie <smcv@debian.org>.
Bug-Debian: #810888
Acked-By: Jamie Strandboge <jamie@canonical.com>
support systems with NetworkManager but no resolvconf where /etc/resolv.conf is
a symlink to /var/run/NetworkManager/resolv.conf
Patch proposed by Simon McVittie <smcv@debian.org>.
Bug-Debian: #813835
Acked-By: Jamie Strandboge <jamie@canonical.com>
If reading /dev/urandom failed, the corresponding file descriptor was
leaked through the error path.
Coverity CID #56012
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
The variable was only referenced by commented section of code so move
the declaration into the comment.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
Add a target that uses cov-build, which must be found in $PATH, to
generate an intermediate Coverity directory called cov-int. The
intermediate Coverity directory will be based on a clean snapshot of the
last commit in the bzr tree. Finally, the intermediate directory is
converted to a compressed tarball, stored in
apparmor-<SNAPSHOT_VERSION>-cov-int.tar.gz, and is suitable for
uploading to scan.coverity.com.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
Turn REPO_VERSION and SNAPSHOT_DIR into make variables that may be
reused by future targets that specify the snapshot target as a
prerequisite. This prevents us from having to repeatedly call out to
potentially slow commands on bound bzr branches, such as the bzr
version-info command stored in the REPO_VERSION_CMD make variable.
The new REPO_VERSION make variable is turned into a "simply expanded"
variable as to not require a callout to bzr each time it is expanded.
The SNAPSHOT_DIR shell variable is renamed to SNAPSHOT_NAME as a make
variable. The new name may be slightly more descriptive in the future as
the variable will be reused in other ways besides a simple directory
name.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
bzr version-info supports directly printing the bare revno to stdout so
we should use that instead of parsing the default verbose output.
This change simplifies the shell snippet used to assign the
REPO_VERSION_CMD make variable. It was also tested to work with the bzr
present in Ubuntu 12.04.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
Acked-by: Steve Beattie <steve@nxnw.org>
Order the DIRS variable according to build order. This allows the DIRS
variable to be iterated over to build libapparmor, binutils, parser,
utils, etc., without having to reorder the list.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
Acked-by: Christian Boltz <apparmor@cboltz.de>
I suspect that the incorrect description of EPERM was copied from
the aa_change_hat man page, where it is possible to see EPERM if the
application is not confined by AppArmor.
This patch corrects the description by documenting that the only
possible way to see EPERM is if a confined application has the
no_new_privs bit set.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Reported-by: Seth Arnold <seth.arnold@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
It is possible that file descriptors will be revalidated after an
aa_change_profile() but there is a lot of complexity involved that
doesn't need to be spelled out in the man page. Instead, mention that
revalidation is possible but the only way to ensure that file
descriptors are not passed on is to close them.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Reported-by: Seth Arnold <seth.arnold@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
The statement was meant to convey the difference between aa_change_hat()
and aa_change_profile(). Unfortunately, it read as if there was
something preventing a program from using aa_change_profile() twice to
move from profile A to profile B and back to profile A, even if profiles
A and B contained the necessary rules.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Reported-by: Seth Arnold <seth.arnold@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
On Debian and Ubuntu it's possible to have multiple ruby interpreters
installed, and the default to use is handled by the ruby-defaults
package, which includes a symlink from /usr/bin/ruby to the versioned
ruby interpreter.
This patch makes aa.py:get_interpreter_and_abstraction() take that into
account by using a regex to match possible versions of ruby. Testcases
are included. (I noticed this lack of support because on Ubuntu the ruby
test was failing because get_interpreter_and_abstraction() would get the
complete path, which on my 16.04 laptop would get /usr/bin/ruby2.2.)
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
This patch frees some leaked memory that occur when errors are
detected while adding variables to the parser's symbol table. While not
a significant issue currently due to the parser exiting on failures, as
the process of library-ifying the parser continues, these need to be
addressed. It also makes it easier to use tools like Address Sanitizer
on the parser against our test suite.
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: Tyler Hicks <tyhicks@canonical.com>
The rule classes have lots of
if self.all_foo:
foo_txt = _('ALL')
else:
foo_txt = self.foo
in logprof_header_localvars().
To avoid repeating this over and over, split it off to a
logprof_value_or_all() function.
This function can handle
- str (will be returned unmodified
- AARE (.regex will be used)
- sets/lists/tuples (will be ' '.join()ed and sorted)
Other types are returned unmodified.
Acked-by: Steve Beattie <steve@nxnw.org>
When hitting an unknown line while parsing a profile, it's a good idea
to include that line in the error message ;-)
Note: 2.9 would print a literal \n because it doesn't have apparmor.fail,
so it will get a slightly different patch with spaces instead of \n.
Acked-by: Steve Beattie <steve@nxnw.org> for trunk, 2.10 and 2.9.
Checking if two AARE objects are equal is not hard, but also not a
one-liner.
Since we need to do this more than once (and even more often in other
outstanding rule classes), split that code into an _is_equal_aare()
function and change PtraceRule and SignalRule to use it.
To make things even more easier, the parameters to use match the
_is_covered_aare() syntax.
Acked-by: Steve Beattie <steve@nxnw.org>
If a *Ruleset is empty, let __repr__() print/return
<FooRuleset (empty) />
instead of
<FooRuleset>
</FooRuleset>
Acked-by: Steve Beattie <steve@nxnw.org> for trunk and 2.10.
PtraceRule 'access' and SignalRule 'access' and 'signal' can contain
more than one value. Therefore adjust is_covered_localvars() in both
to use the list (subset) instead of the plain (exactly equal) check.
Also add a testcase for each to ensure the list/subset check works as
expected.
Acked-by: Steve Beattie <steve@nxnw.org>
is_covered_localvars() in the rule classes need the same set of checks
again and again. This patch adds the helper functions _is_covered_list(),
_is_covered_aare() and _is_covered_plain() to check against lists, AARE
and plain variables like str.
The helpers check if the values from the other rule are valid (either
ALL or the value need to be set) and then check if the value is covered
by the other rule's values.
This results in replacing 7 lines with 2 in the rule classes and avoids
repeating code over and over.
Note that the helper functions depend on the *Rule.rule_name variable in
the exception message, therefore rule_name gets added to all rule
classes.
Acked-by: Steve Beattie <steve@nxnw.org>
The first entry in the grouping_count array is never initialized to 0;
subsequent depths are. This patch initializes the whole array.
Issue found with valgrind.
Signed-off-by: Steve Beattie <steve@nxnw.org> (with improvement from Seth)
Acked-by: Seth Arnold <seth.arnold@canonical.com>
'!' is a reserved symbol and needs to be escaped in AARE.
Note: aare.py only exists in trunk, therefore this part is trunk-only.
Acked-by: Seth Arnold <seth.arnold@canonical.com> for trunk, 2.10 and 2.9 as needed.
The capnames list missed a comma, which lead to the funny
"mac_overridesyslog" capability name.
__debug_capabilities() seems to be the only user of capnames, which
might explain why this bug wasn't noticed earlier.
Acked-by: Seth Arnold <seth.arnold@canonical.com> for trunk, 2.10 and 2.9.
BugLink: http://bugs.launchpad.net/bugs/1534405
Patch -r 2952 switched over to using the library kernel interface, and
added a kernel_interface parameter to the dir_cb struct, that is
used to process directories.
Unfortunately kernel_interface parameter of the dir_cb struct is not being
properly initialized resulting in odd failures and sefaults when the parser
is processing directories.
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
This adds a basic support for parallel compiles. It uses a fork()/wait
model due to the parsers current dependence on global variables and
structures. It has been setup in a similar manner to how cilk handles
multithreading to make it easy to port to a managed thread model once
the parser removes the dependence on global compute structures in the
backend.
This patch adds two new command line flags
-j <n> or --jobs <n>
which follows the make syntax of specifying parallel jobs currently
defaults to -jauto
-j8 or --jobs=8 allows for 8 parallel jobs
-jauto or --jobs=auto sets the jobs to the # of cpus
-jx4 or --jobs=x4 sets the jobs to # of cpus * 4
-jx1 is equivalent to -jauto
Note: unlike make -j must be accompanied by an option
--max-jobs=<n>
allows setting hard cap on the number of jobs that can be specified
by --jobs. It defaults to the number of processors in the system * 8.
It supports the "auto" and "max" keywords, and using x<n> for a
multiple of the available cpus.
additionally the -d flag has been modified to take an optional parameter
and
--debug=jobs
will output debug information for the job control logic.
In light testing on one machine the job control logic provides a nice
performance boost. On an x86 test machine with 60 profiles in the
/etc/apparmor.d/ directory, for the command
time apparmor_parser -QT /etc/apparmor.d/
old (equiv of -j1):
real 0m10.968s
user 0m10.888s
sys 0m0.088s
ubuntu parallel load using xargs:
real 0m8.003s
user 0m21.680s
sys 0m0.216s
-j:
real 0m6.547s
user 0m17.900s
sys 0m0.132s
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
We need to check a rule part if it is *Rule.ALL or a string at various
places. Therefore split off the checks in PtraceRule's and SignalRule's
__init__() to the new _aare_or_alll() function in BaseRule.
This also makes the *Rule __init__() much more readable because we now
have one line to set self.foo and self.all_foo instead of 10 lines of
nested if conditions.
Acked-by: Steve Beattie <steve@nxnw.org>.
If parse_event_for_tree() raises an AppArmorException (for example
because of an invalid/unknown request_mask), catch it in read_log() and
re-raise it together with the log line causing the Exception.
Acked-by: Steve Beattie <steve@nxnw.org> for trunk, 2.10 and 2.9.
Swap aa_query_link_path_len() and aa_query_link_path() to match the
order of aa_query_file_path() and aa_query_file_path_len().
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
Doing manual line wraps resulted in an unreadable SYNOPSIS section.
Allow man to handle line wrapping the function prototypes itself.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
aa_query_file_path, aa_query_file_path_len, aa_query_link_path, and
aa_query_link_path_len were omitted from the NAME section.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
handle_children() has some special code for handling link events with
denied_mask = 'l'. Unfortunately this special code depends on a regex
that matches the old, obsolete log format - in a not really parsed
format ("^from .* to .*$").
The result was that aa-logprof did not ask about events containing 'l'
in denied_mask.
Fortunately the fix is easy - delete the code with the special handling
for 'l' events, and the remaining code that handles other file
permissions will handle it :-)
References: Bugreport by pfak on IRC
Testcase (with hand-tuned log event):
aa-logprof -f <( echo 'Jan 7 03:11:24 mail kernel: [191223.562261] type=1400 audit(1452136284.727:344): apparmor="ALLOWED" operation="link" profile="/usr/sbin/smbd" name="/foo" pid=10262 comm=616D617669736420286368362D3130 requested_mask="l" denied_mask="l" fsuid=110 ouid=110 target="/bar"')
should ask to add '/foo l,' to the profile.
Acked-by: Seth Arnold <seth.arnold@canonical.com> for trunk, 2.10 and 2.9.
from intrigery:
dnsmasq profile: extract confinement of libvirt_leaseshelper into a dedicated sub-profile.
Acked-by: Christian Boltz <apparmor@cboltz.de>
Thanks to http://bugs.python.org/issue10076, we need to implement this
ourself :-/
Also add some tests to ensure __deepcopy__() works as expected.
I found this bug while testing the dbus patch series, which crashed
aa-cleanprof with
TypeError: cannot deepcopy this pattern object
Acked-by: John Johansen <john.johansen@canonical.com>
collapse_log() creates temporary SignalRule etc. objects which are then
checked against the existing profile content.
These temporary objects are based on log events, therefore flag them as
such. This will ensure proper handling and escaping by the AARE class.
Acked-by: John Johansen <john.johansen@canonical.com>
In detail, this means:
- handle ptrace events in logparser.py
- "translate" those events in aa.py - from log (logparser.py readlog())
to prelog (handle_children()) to log_dict (collapse_log()) to
log_obj (ask_the_questions())
(yes, really! :-/ - needless to say that this is ugly...)
- finally ask the user about the ptrace in ask_the_questions()
Also add a logparser test to test-ptrace.py to ensure the logparser step
works as expected.
Note that the aa.py changes are not covered by tests, however they
worked in a manual test.
If you want to test manually, try this (faked) log line:
msg=audit(1409700683.304:547661): apparmor="DENIED" operation="ptrace" profile="/usr/sbin/smbd" pid=22465 comm="ptrace" requested_mask="trace" denied_mask="trace" peer="/foo/bar"
Acked-by: John Johansen <john.johansen@canonical.com>
"Everywhere" means aa-mergeprof and aa-cleanprof. In theory also
aa-logprof, but that needs some code that parses ptrace log events ;-)
Acked-by: John Johansen <john.johansen@canonical.com>
Change aa.py to use PtraceRule and PtraceRuleset in profile_storage(),
parse_profile_data() and write_ptrace(). This also means we can drop the
now unused parse_ptrace_rule() and write_ptrace_rules() functions.
Raw_Ptrace_Rule in rules.py is now also unused and can be dropped.
Also adjust logparser.py to include the peer in the result, and shorten
the list of known-failing tests in test-parser-simple-tests.py.
Acked-by: John Johansen <john.johansen@canonical.com>
As usual, we have 100% test coverage :-)
Those tests include all tests from test-ptrace_parse.py, therefore
delete this file.
Acked-by: John Johansen <john.johansen@canonical.com>
The tests in test-ptrace_parse.py used aa.parse_ptrace_rule(), which is
based on Raw_Ptrace_Rule (= regex check + "just store it").
This patch changes the tests to test against PtraceRule.get_clean().
Since get_clean does some cleanups, the expected result slightly differs
from the original rule.
Finally switch to the AATest class and setup_all_loops() we use in most
tests.
Also change test-regex_matches.py to import RE_PROFILE_SIGNAL directly
from apparmor.regex instead of apparmor.aa (where it will vanish soon).
Acked-by: John Johansen <john.johansen@canonical.com>
Those classes will be used to parse and handle ptrace rules.
They understand the syntax of ptrace rules.
Note that get_clean() doesn't output superfluos things, so
ptrace ( trace ),
will become
ptrace trace,
Acked-by: John Johansen <john.johansen@canonical.com>
As a preparation for the PtraceRule class, add a <details> match group
to RE_PROFILE_PTRACE.
Also adjust test-regex_matches.py for the added group.
Note: RE_PROFILE_PTRACE is only used in aa.py, and only matches[0..2]
are used. 0 and 1 are audit and allow/deny and 2 is and stays the whole
rule (except audit and allow/deny). Therefore no aa.py changes are
needed.
Acked-by: John Johansen <john.johansen@canonical.com>
Remove the Perl aa-exec implementation, move the aa-exec(8) man page to
binutils/, and point the regression test to the C based aa-exec in
binutils/.
Note that the new C aa-exec does not implement the --file option which
was present in the Perl aa-exec. It encouraged running programs as root,
since root privileges were required to load the specified profile.
All other features of the Perl aa-exec are present in the C aa-exec.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
Call aa_change_profile(), instead of aa_change_onexec(), when
--immediate is passed in.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
Create a simple aa-exec implementation, written in C, matching the
--help, --debug, --verbose, and --profile options present in the current
Perl implementation.
The new aa-exec sources reside in the binutils/ directory.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
aa-enabled should live in /usr/bin, rather than /sbin, since it is not
used in early boot and requires no root privileges.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
r2637 added support for parsing unix rules, but forgot to add write
support. The result was that a profile lost its unix rules when it was
saved.
This patch adds the write_unix_rules() and write_unix() functions (based
on the write_pivot_root() and write_pivot_root_rules() functions) and
makes sure they get called at the right place.
The cleanprof testcase gets an unix rule added to ensure it's not
deleted when writing the profile. (Note that minitools_test.py is not
part of the default "make check", however I always run it.)
References: https://bugs.launchpad.net/apparmor/+bug/1522938https://bugzilla.opensuse.org/show_bug.cgi?id=954104
Acked-by: Tyler Hicks <tyhicks@canonical.com> for trunk, 2.10 and 2.9.
This means:
- expect unicode (instead of str) when reading from a file in py2
- convert keys() result to a set to avoid test failures because of
dict_keys type
After this change, all tests work for both py2 and py3.
Acked-by: Tyler Hicks <tyhicks@canonical.com> for trunk and 2.10.
python 3 uses only the 'str' type, while python 2 also uses 'unicode'.
This patch adds a type_is_str() function to common.py - depending on the
python version, it checks for both. This helper function is used to keep
the complexity outside of the rule classes.
The rule classes get adjusted to use type_is_str() instead of checking
for type(x) == str, which means they support both python versions.
As pointed out by Tyler, there are also some type(...) == str checks in
aare.py and rule/__init__.py which should get the same change.
Finally, add test-common.py with some tests for type_is_str().
References: https://bugs.launchpad.net/apparmor/+bug/1513880
Acked-by: Tyler Hicks <tyhicks@canonical.com> for trunk and 2.10
Note: 2.10 doesn't contain SignalRule, therefore it doesn't get that
part of the patch.
Add regression tests for the --profile, --namespace, and --immediate
options of aa-exec.
A new variable is added to uservars.inc to point to the in-tree or
system aa-exec depending on the presence of the USE_SYSTEM=1 make
variable at build time.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
Clean up the Makefile by removing distro-related install targets. These
should not be needed.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
The new aa-enabled program can be used as a barebones replacement for
`aa-status --enabled`. It is written in C, rather than Python, which
keeps its dependencies to a minimum.
By default, aa-enabled prints a human-readable status of AppArmor's
availability to stdout. It supports a --quiet option which allows for
functionality equivalent to `aa-status --enabled`, which does not print
any messages.
The aa-enabled exit statuses mimic the behavior documented in the
aa-status(8) man page.
Signed-off-by: John Johansen <john.johansen@canonical.com>
[tyhicks: Incorporated feedback from the code review process]
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Don't catch AppArmorExceptions in aa-easyprof any longer and rely on
apparmor.fail to print the exception to stderr.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
https://launchpad.net/bugs/1526085
Revno 2934 'Add fns to handle profile removal to the kernel interface'
introduced a regression in the parser's namespace support by causing the
--namespace-string option to be ignored. This resulted in the profile(s)
being loaded into the global namespace rather than the namespace
specified on the command line.
This patch fixes the bug by setting the Profile object's ns member, if
the --namespace-string option was specified, immediately after the
Profile object is allocated.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
'change_hat' events have the target profile in 'name2', not in 'name'
(which is None and therefore causes a crash when checking if it contains
'//')
Also add the log event causing this crash to the libapparmor testsuite.
References: https://bugs.launchpad.net/apparmor/+bug/1523297
Acked-by: John Johansen <john.johansen@canonical.com> for trunk, 2.10 and 2.9.
Parsing variables was broken in several ways:
- empty quotes (representing an intentionally empty value) were lost,
causing parser failures
- items consisting of only one letter were lost due to a bug in RE_VARS
- RE_VARS didn't start with ^, which means leading garbage (= syntax
errors) was ignored
- trailing garbage was also ignored
This patch fixes those issues in separate_vars() and changes
var_transform() to write out empty quotes (instead of nothing) for empty
values.
Also add some tests for separate_vars() with empty quotes and adjust
several tests with invalid syntax to expect an AppArmorException.
var_transform() gets some tests added.
Finally, remove 3 testcases from the "fails to raise an exception" list
in test-parser-simple-tests.py.
Acked-by: John Johansen <john.johansen@canonical.com> for trunk and 2.9
(which also implies 2.10)
Note: 2.9 doesn't have test-parser-simple-tests.py, therefore it won't
get that part of the patch.
This patch adds a check-local target to libapparmor/testsuite/Makefile.am
that checks the logfile generated by the test_multi tests
(libaalogparse.log) and errors out if
- the logfile doesn't exist (which might mean that dejagnu isn't installed
- the logfile contains 'ERROR'
This isn't the best solution I can imagine, but it's the only/easiest
way I found that doesn't need changing of autogenerated files.
Also extend clean-local to delete libaalogparse.{log,sum}
Finally, add test_multi/testcase_syslog_read.err (empty file) to avoid
make check fails.
Acked-by: John Johansen <john.johansen@canonical.com>
Also adjust test-signal.py for AARE (it needs a change in _compare_obj())
and enable the regex-based tests.
Acked-by: John Johansen <john.johansen@canonical.com>
The AARE class is meant to handle the internals of path AppArmor regexes
at various places / rule types (filename, signal peer etc.). The goal is
to use it in rule classes to hide all regex magic, so that the rule
class can just use the match() method.
If log_event is given (which means handing over a raw path, not a regex),
the given path is converted to a regex in convert_expression_to_aare().
(Also, the raw path is used in match().)
BTW: The reason for delaying re.compile to match() is performance - I'd
guess a logprof run calls match() only for profiles with existing log
events, so we can save 90% of the re.compile() calls.
The patch also includes several tests.
Acked-by: John Johansen <john.johansen@canonical.com>
Having a list of rule types/classes at several places is annoying and
error-prone. This patch centralizes the list in aa.py.
This also means ask_the_question() in aa.py will now (in theory) support
'change_profile' and 'rlimit'. In practise, that doesn't change anything
because logparser.py doesn't support change_profile events yet - and
rlimit doesn't cause any log events.
Also add some long overdue copyright headers.
Acked-by: Seth Arnold <seth.arnold@canonical.com>
In detail, this means:
- handle signal events in logparser.py
- "translate" those events in aa.py - from log (logparser.py readlog())
to prelog (handle_children()) to log_dict (collapse_log()) to
log_obj (ask_the_questions())
(yes, really! :-/ - needless to say that this is ugly...)
- finally ask the user about the signal in ask_the_questions()
Also add a logparser test to test-signal.py to ensure the logparser step
works as expected.
Note that the aa.py changes are not covered by tests, however they
worked in a manual test.
Acked-by: John Johansen <john.johansen@canonical.com>
As Kshitij mentioned, abstract methods should use NotImplementedError
instead of AppArmorBug.
While changing this, I noticed that __repr__() needs to be robust against
NotImplementedError because get_raw() is not available in BaseRule.
Therefore the patch changes __repr__() to catch NotImplementedError.
Of course the change to NotImplementedError also needs several
adjustments in the tests.
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
(long before branching off 2.10, therefore I'll also commit to 2.10)
It's pointless to keep a separate file for those tests - they integrate
well in test-signal.py.
After the move, test-signal_parse.py is empty and will be deleted.
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
Besides 'signal', also 'change_profile' and 'rlimit' cleanup was missing
for the main profile.
In aa.py delete_duplicates() (used to check includes), only 'signal' was
missing.
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
This means:
- import the classes instead of RE_PROFILE_SIGNAL
- simplify signal rule parsing a lot
- drop the (now unused) functions parse_signal_rule() and write_signal_rules()
- change write_signal() to use the SignalRuleset class
Also drop the now unused Raw_Signal_Rule from rules.py.
Finally, drop most parser signal tests from the "known wrong results"
blacklist in test-parser-simple-tests.py because those tests succeed
with SignalRule.
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
The tests in test-signal_parse.py used aa.parse_signal_rule(), which is
based on Raw_Signal_Rule (= regex check + "just store it").
This patch changes the tests to test against SignalRule.get_clean().
Since get_clean() does some cleanups, the expected result slightly
differs from the original rule.
Finally switch to the AATest class and setup_all_loops() we use in most
tests.
Also change test-regex_matches.py to import RE_PROFILE_SIGNAL directly
from apparmor.regex instead of apparmor.aa (where it will vanish soon).
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
Those classes will be used to parse and handle signal rules.
They understand the (surprisingly complex) syntax of signal rules.
Note that get_clean() doesn't output superfluos things, so
signal ( send ) set = ( int ),
will become
signal send set=int,
Also add a set of tests (100% coverage :-) to make sure everything works
as expected.
This is a merged commit of the following patches:
- 07-add-SignalRule-and-SignalRuleset.diff
- 13-test-signal-compare_obj.diff
- 17-signal-rule-cleanup.diff
- 21-test-signal-rename-tests.diff
- 22-signal-rule-adjustments.diff
- 24-signal-rule-fix-error-message.diff
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
(all patches in this commit)
As a preparation for the SignalRule class, add a <details> match group
to RE_PROFILE_SIGNAL.
Also adjust test-regex_matches.py for the added group.
Note: RE_PROFILE_SIGNAL is only used in aa.py, and only matches[0..2]
are used. 0 and 1 are audit and allow/deny and 2 is and stays the whole
rule (except audit and allow/deny). Therefore no aa.py changes are
needed.
Acked-by: John Johansen <john.johansen@canonical.com> for trunk and 2.10
Creating a file is in theory covered by the 'a' permission, however
discussion on IRC brought up that depending on the open flags it might
not be enough (real-world example: creating the apache pid file).
Therefore change the mapping to 'w' permissions. That might allow more
than needed in some cases, but makes sure the profile always works.
Acked-by: Kshitij Gupta <kgupta8592@gmail.com> for 2.9, 2.10 and trunk
For debugging, it's helpful to know which part of the code initialized a
profile_storage and for which profile and hat this was done.
This patch adds an 'info' array with that information, adds the
corresponding parameters to profile_storage() and changes the callers to
deliver some useful content.
Acked-by: John Johansen <john.johansen@canonical.com> for trunk and 2.10
We replaced parse_audit_allow() with parse_modifiers() in r2833, but
overlooked that parse_modifiers() returns allow/deny as boolean. This
resulted in storing bare file rules in aa[profile][hat]['path'][False]
instead of aa[profile][hat]['path']['allow'] (or True instead of 'deny'
for 'deny file,' rules), with the user-visible result of loosing bare
file rules when saving the profile.
This patch converts the boolean value from parse_modifiers back to a
string.
Note: 2.9 is not affected because the old parse_audit_allow() returns
'allow' or 'deny' as string, not as boolean.
Acked-by: Kshitij Gupta <kgupta8592@gmail.com> for trunk and 2.10
The last utils/test/Makefile change switched to using the in-tree
libapparmor by default (unless USE_SYSTEM=1 is given). However, I missed
to add the swig/python parts of libapparmor to PYTHONPATH, so the
system-wide LibAppArmor/__init__.py was always used.
This patch adds the in-tree libapparmor python module to PYTHONPATH.
I'm sorry for the interesting[tm] way to find out that path, but
a) I don't know a better / less ugly way and
b) a similar monster already works in libapparmor/swig/python/test/ ;-)
Acked-by: John Johansen <john.johansen@canonical.com> for 2.9 and trunk
(that also implies 2.10 ;-)
To make things more interesting, /usr/bin/python and /usr/bin/python[23]
are symlinks to /usr/bin/python[23].[0-9], so we have to explicitely
list several versions.
Acked-by: John Johansen <john.johansen@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com> for 2.9, 2.10 and trunk
2.9.x and 2.10 had some time stamp bugs around cache handling that
result in the cache getting a wrong time stamp, and then not getting
correctly updated when policy changes.
Force cache recompiles for these versions by bumping the parser abi
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
syslog-ng needs to access both the permanent /var/log/journal/ and the
non-permanent /run/journal/.
I also included /var/run/journal/ to stay consistent with supporting
both /run/ and /var/run/.
Acked-by: Steve Beattie <steve@nxnw.org> for trunk and 2.9.
Some packages like libvirt-bin and lxc drop conf snippets in /etc/dnsmasq.d-available
and make them available through symlinks in /etc/dnsmasq.d created during postinst.
This makes print()ing a class object much more helpful - instead of
<apparmor.rule.network.NetworkRule object at 0x7f416b239e48>
we now get something like
<NetworkRule> network inet stream,
(based on get_raw())
A NetworkRuleset will be printed as (also based on get_raw())
<NetworkRuleset>
network inet stream,
allow network inet stream, # comment
</NetworkRuleset>
Also add tests to test-network.py to ensure that __repr__() works as
expected.
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
aa-logprof is able to parse all profiles, so there is no longer a
reason to skip this test.
This patch reverts r2097 and r2098 from 2013-01-02.
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
Acked-by: John Johansen <john.johansen@canonical.com>
(and now that the tests work even if logprof.conf doesn't exist,
Steve's NACK is no longer valid)
This patch checks if the cfg object is empty (happens if logprof.conf
doesn't exist). If so, it adds some empty sections to prevent various
failures in code that expects those sections to exist.
Another source of failures was using cfg['section']['setting']. The
patch changes various places to cfg['section'].get('setting') to prevent
those failures. (Those places all have a 'or ...' fallback.)
Finally, find_first_file() in config.py crashed if file_list was Null.
This is fixed by adding an "if file_list:" check before trying to
split() it.
With all those changes applied, 'make check' will work even if
/etc/apparmor/logprof.conf doesn't exist.
The patch also fixes the default value for inactive_profiledir
(I missed aa.py when I changed it to /usr/share/apparmor/extra-profiles/)
References: https://bugs.launchpad.net/apparmor/+bug/1393979
Acked-by: John Johansen <john.johansen@canonical.com>
Both create_new_profile() and handle_children() check if the given exec
target is a script and add permissions for the interpreter and a
matching abstraction.
This patch merges that into the get_interpreter_and_abstraction()
function and changes create_new_profile() and handle_children() to use
this function.
A nice side effect is that handle_children() now knows more abstractions
(its original list was incomplete).
The behaviour of create_new_profile() doesn't change.
Also add tests for get_interpreter_and_abstraction() to make sure it
does what we expect.
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
Bug: https://launchpad.net/bugs/1505775
These tests ensure that create_new_profile() sets the expected basic
permissions for scripts and non-script files.
Acked-by: John Johansen <john.johansen@canonical.com>
oftc_ftw reported on IRC that Arch Linux has a symlink /bin -> /usr/bin.
This means we have to update paths for /bin/ in several profiles to also
allow /usr/bin/
Acked-by: John Johansen <john.johansen@canonical.com> for trunk and 2.9
Also add support for the USE_SYSTEM variable, which means:
- test against the in-tree libapparmor and python modules by default
- test against the system libapparmor and python modules if USE_SYSTEM
is set
The old behaviour was a mix of both - it always used the in-tree python
modules and the system libapparmor.
For obvious reasons, you'll need to build libapparmor before running the
tests (unless you specify USE_SYSTEM=1 as parameter to make check).
Acked-by: John Johansen <john.johansen@canonical.com> for trunk and 2.9
Add a testcase that parses all tests in the parser/tst/simple_tests/
directory with parse_profile_data() to ensure that everything with valid
syntax is accepted, and that all tests marked as FAIL raise an
exception.
This already resulted in
- several patches to fix low-hanging fruits (including some bugs in the
parser simple_tests itsself)
- a list of tests that don't behave as expected. Those files get their
expected result reverted to make sure we notice any change in the
tools behaviour, especially changing to the really expected resulted.
This method also makes sure that the testcase doesn't report any of
the known failures.
- a 5% improvement in test coverage - mostly caused by nearly completely
covering parse_profile_data.
- addition of some missing testcased (as noticed by missing coverage),
for example several "rule outside of a profile" testcases.
As indicated above, the tools don't work as expected on all test
profiles - most of the failures happen on expected-to-fail tests that
pass parse_profile_data() without raising an exception. There are also
some tests failing despite valid syntax, often with rarely used syntax
like if conditions and qualifier blocks.
Most of the failing (generated) tests are caused by features not
implemented in the tools yet:
- validating dbus rules (currently we just store them without any parsing)
- checks for conflicting x permissions
- permissions before path ("r /foo,")
- 'safe' and 'unsafe' keywords for *x rules
- 'Pux' and 'Cux' permissions (which actually mean PUx and CUx, and get
rejected by the tools - ideally the generator script should create
PUx and CUx tests instead)
skip_startswith excludes several generated tests from being run. I know
that skip_startswith also excludes tests that would not fail, but the
generated filenames (especially generated_x/exact-*) don't have a
pattern that I could easily use to exclude less tests - and I'm not too
keen to add a list with 1000 single filenames ;-)
Acked-by: John Johansen <john.johansen@canonical.com>
The global variable 'logger' in aa.py is only used by aa-genprof.
This patch changes aa_genprof to use the (new) logger_path() function,
and moves the code for finding the logger path to that function.
Also make the error message more helpful if logger can't be found.
Acked-by: John Johansen <john.johansen@canonical.com>
The 'ldd' variable in aa.py is only used by get_reqs(), therefore move
setting it (based on the configfile) into the function.
get_reqs() doesn't run too often (only called by create_new_profile(),
which means aa-genprof or when adding a Px or Cx rule to a non-existing
profile). This might even lead to a minor performance win - on average,
I'd guess not every aa-logprof run will lead to a completely new profile
or child profile. And, more important, we get rid of a global variable.
Acked-by: John Johansen <john.johansen@canonical.com>
create_new_profile() didn't init missing required_hats as
profile_storage(), which might lead to crashes when creating a profile
for an application listed in the required_hats config option (= in very
rare cases).
This patch adds the missing profile_storage() call.
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
This also means the duplicate detection can use the hat's filename instead
of the (possibly wrong) main profile's filename.
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
This patch is based on a SLE12 patch to allow executing the
--dhcp-script. We already have most parts of that patch since r2841,
except /dev/tty rw which is needed for the shell's stdout and stderr.
References: https://bugzilla.opensuse.org/show_bug.cgi?id=940749 (non-public)
Acked by Seth Arnold on IRC (with "owner" added)
With this addition, all globbing styles (as documented in apparmor.d(5))
are covered in the convert_regexp() tests.
Acked-by: Seth Arnold <seth.arnold@canonical.com>
load_include() used a custom os.listdir call instead of
include_dir_filelist() for directory includes, which means it also read
skippable files like *.rpmnew or README. (It seems nobody created a
README inside an included directory, otherwise we'd have seen a
bugreport ;-)
This patch changes load_include() to use include_dir_filelist(). This
function is used in some more places already and removes skippable files
from the file list.
Acked-by <timeout>
load_include() has a "if not incdata:" block which would be entered if
parse_profile_data() returns None. However, parse_profile_data() always
returns a hasher with [incfile][incfile] = profile_storage(), so that
"if not incdata:" never matches.
Acked-by <timeout>
The "already loaded?" check in load_include() was done at the beginning
of the function, before entering the loop and before the individual
files of directory includes were added to the filelist. This resulted in
a (wrong) "Conflicting profiles" error for directory includes.
This patch moves the "alreay loaded?" check inside the loop, so that
it's executed for all files, including those of directory includes.
Acked-by <timeout>
TL;DR: aa-genprof crashes with a wrong 'Conflicting profiles' error.
aa-genprof uses autodep() to create a basic profile, which is then
stored in aa and original_aa. After that, read_profiles() is called,
which reads all profiles (including the new one) from disk, causing a
(wrong) 'Conflicting profiles' error in attach_profile_data() because
the autodep()-generated profile is already there.
Therefore this patch resets aa and original_aa in read_profiles() to
avoid that problem.
Acked-by <timeout>
The tests for convert_regexp() were hidden in common_test.py, where they
were never executed.
This patch moves them to the new file test-aare.py and also converts the
regex_tests.ini to a tests[] array to have the test data inside the test
file. (All tests from regex_tests.ini are in test-aare.py, and two tests
with prepended and appended path segments were added.)
Also add some tests that check the raw behaviour of convert_regexp() -
the tests "by example" are probably more useful and for sure more
readable ;-) but I want to have some examples of the converted regexes
available.
Acked-by <timeout>
logparser.py does a regex check on log lines as performance improvement
so that it only hands over lines that look like AppArmor events to
LibAppArmor parsing. Those regexes were incomplete and didn't cover all
log formats LibAppArmor accepts, with the end result of "overlooking"
events.
This patch splits off common parts of the regex, adds more regexes for
several log types and finally merges everything into one regex.
test-libapparmor-test_multi.py now also checks all test_multi log lines
against the regex to ensure logparser.py doesn't silently ignore events.
test-logparser.py gets adjusted to the merged RE_LOG_ALL regex.
Finally, add a new test that was posted on IRC to the test_multi set.
As already threatened nearly a month ago,
Acked by <timeout> for trunk and 2.9
This patch is based on a SLE12 patch to allow executing the
--dhcp-script. We already have most parts of that patch since r2841,
however the SLE bugreport indicates that /bin/sh is executed (which is
usually a symlink to /bin/bash or /bin/dash), so we should also allow
/bin/sh
References: https://bugzilla.opensuse.org/show_bug.cgi?id=940749 (non-public)
Acked-by: Seth Arnold <seth.arnold@canonicalc.com> for trunk and 2.9
Add some permissions that I need on my system:
- execute nm-dhcp-helper
- read and write /var/lib/dhcp6/dhclient.leases
- read /var/lib/NetworkManager/dhclient-*.conf
- read and write /var/lib/NetworkManager/dhclient-*.conf
Looks-good-by: Steve Beattie <steve@nxnw.org>
Acked-by: <timeout> for trunk and 2.9
This testcase will parse all libraries/libapparmor/testsuite/test_multi
tests and compare the result with the *.out files.
It also include a "ToDo list" of keywords that are not yet supported in
the python code - those are typically related to rule types not
supported in the tools yet (dbus, signal etc.).
An interesting special case are exec events with network details:
testcase01.in, testcase12.in, testcase13.in
which might be hand-made, invalid logs, but nobody remembers ;-)
Acked-by <timeout>
Drop the reference to the libapparmor policy_cache pseudo object when
the parser is done.
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Fix memory leaks when parsing dmesg timestamps as well as when handling
message the library does not understand.
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: John Johansen <john.johansen@canonical.com>
In recent commits, Tyler fixed some problems with the caching behavior
of the parser, as well as adjusting and improving the caching test
script to verify these behaviors.
In doing so, the test script adjusts the mtime of various
files and ensures that the written files have the expected mtime
timestamp. Unfortunately, the os.utime() function used to adjust mtime
in python 3.2 (as included in Ubuntu 12.04 LTS) does not update with
nanosecond precision, even though the timestamps returned by os.stat()
do have precision to nanoseconds. This causes the tests to fail when
running under python 3.2 with errors like the following:
======================================================================
FAIL: test_abstraction_newer_rewrites_cache (__main__.AAParserCachingTests)
----------------------------------------------------------------------
Traceback (most recent call last):
File "/��PKGBUILDDIR��/parser/tst/testlib.py", line 50, in new_unittest_func
return unittest_func(self)
File "./caching.py", line 424, in test_abstraction_newer_rewrites_cache
self._set_mtime(self.abstraction, abstraction_mtime)
File "./caching.py", line 238, in _set_mtime
self.assertEquals(os.stat(path).st_mtime, mtime)
AssertionError: 1440337039.40212 != 1440337039.4021206
The following patch creates a new time stamp equality assertion
function that detects if it's running on python 3.2 or earlier, and
loosens the equality bounds when comparing the passed timestamps. On
python 3.3 and newer, where writing timestamps with nanosecond precision
is supported, the strict equality assertion is used.
(Note: I did not convert all time stamp comparisons, just ones where
the timestamp written and checked could be based on a timestamp
derived from os.stat().)
Reference: https://bugs.python.org/issue12904
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
/usr/share/locale-bundle/ contains translations packaged in
bundle-lang-* packages in openSUSE.
Acked-by: Steve Beattie <steve@nxnw.org> for trunk and 2.9
The contents of the policy cache files varies based on kernel feature
support found in apparmorfs but the caching tests are mostly about
whether or not a cache file was generated and with the right timestamps.
This patch makes it so that the tests are not entirely skipped when
apparmorfs is not available. Instead, a flat features file will be used
in most cases and only the specific tests that require apparmorfs will
be skipped.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
This makes several improvements to the parser caching tests to verify
that the parser is properly consuming the mtime of profiles and
abstractions when dealing with the policy cache.
It introduces a simple abstraction file and tests the mtime handling by
changing the mtime on the profile, abstraction, apparmor_parser, and
cache file in various combinations to check the parser's behavior.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
This patch fixes a regression in setting the cache file's timestamp
handling that was introduced in r3079:
Set cache file tstamp to the mtime of most recent policy file tstamp
The previously used utimes(2) syscall requires a two element timeval
array. The first element in the array is the atime to be used and the
second element is the mtime. The equivalent of a one element timeval
array was being passed to it, resulting in garbage being used for the
mtime value. The utimes(2) syscall either failed, due to the invalid
input, or set mtime to an unexpected value. The return code wasn't being
checked so the failure went unknown.
This patch switches to utimensat(2) for a couple reasons. The UTIME_OMIT
special value allows us to preserve the inode's atime without calling
stat(2) to fetch the atime. It also allows for nanosecond precision
which better aligns with what stat(2) returns on the input profile and
abstraction files. That means that we can have the exact same mtime on
the input profile or abstraction file and the output cache file.
https://launchpad.net/bugs/1484178
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
... and add a few mostly innocuous permissions in there, that are not
strictly needed for a seemingly functional setup, but the lack thereof
triggers denial logs, that could indicate that the software falls back
to some degraded operation mode.
In testing against the 4.1 kernel, the syscall_sysctl testcase started
failing even in the unconfined case. What the test program does is
attempt to adjust the kernel.threads-max sysctl to be slightly larger
and see if the operation succeeds by reading the value back out. It
also attempts to save the original value and restore it. The test
was failing because (in VMs at least) the default value chosen by
the kernel for the kernel.threads-max setting was high enough that
attempts to increase it would be ignored (likely to prevent too much
use of kernel memory by threads), helpfully without any message being
report to dmesg. Thus, the initial read of the current value would
succeed, the write of that value + 1024 would appear to succeed,
but then reading the value back out and comparing it to the expected
value would fail, as it would still be the original value, not the
expected new value.
This patch attempts to address this by first attempting to raise
the value, and if that does not appear to work, to then attempt
to lower it. It also refactors the code a bit by creating helper
functions to perform the actual sysctl(2) calls to make the code a
bit easier to read.
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Having two profiles for the same binary is "technically allowed", but it
leads to interesting[tm] behaviour because one of them "wins" depending
on the load order. To make things even more interesting, the kernel load
order can be different from the tools load order, leading to even more
fun.
Short version: you do _not_ want that situation ;-)
This patch adds a duplicate check to attach_profile_data() so that it
errors out if it finds duplicate profiles or hats, and lists the profile
files that contain them.
Acked-by: Kshitij Gupta <kgupta8592@gmail.com> for both trunk and 2.9.
In some cases, the return value of name_to_prof_filename() is undefined.
This happens when deleting the to-be-confined binary while running
aa-genprof and leads to a not-too-helpful
File "/usr/lib/python3/dist-packages/apparmor/aa.py", line 265, in enforce
prof_filename, name = name_to_prof_filename(path)
TypeError: 'NoneType' object is not iterable
(reported by maslen on IRC)
This patch makes sure name_to_prof_filename() always returns None, None
(instead of undefined aka just None) so that at least the caller can
successfully split it into two None values.
For the exotic aa-genprof usecase given above, this at least improves
the error message to
Can't find $binary_name
(raised by enforce() via fatal_error())
The patch also changes fatal_error() to display the traceback first, and
the human-readable message at the end, which makes it more likely that
the user actually notices the human-readable message.
Acked-by: Kshitij Gupta <kgupta8592@gmail.com> for both trunk and 2.9.
Profile name and attachment can contain variables, so the
RE_PROFILE_START regex should accept it.
(Note: the variable content isn't checked.)
Also add some tests with variables.
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
add_event_to_tree() is a hard-to-test function because it hands over its
result to add_to_tree().
This patch converts add_event_to_tree() to a simple wrapper function and
moves the main code into parse_event_for_tree() and map_log_type(). These
two new functions return their results and are therefore easier to test.
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
"rcapparmor kill" results in a funny error message:
/lib/apparmor/rc.apparmor.functions: line 441: return: -v: invalid option
return: usage: return [n]
SLE12 includes a patch that prevents this error message, but also
prevents that $? is handed over correctly to rc_status. This means that
"rcapparmor kill" will happily display "done" even with a compiled-in
apparmor module that can't be unloaded.
This patch is the improved version - it adds a small helper function to
set $? (as handed over to aa_log_end_msg()) and then calls rc_status -v.
This means that "rcapparmor kill" now shows "failed" because it's
impossible to unload something that is compiled directly into the
kernel.
References: https://bugzilla.opensuse.org/show_bug.cgi?id=862170 (non-public)
Acked-by: Seth Arnold <seth.arnold@canonical.com> for 2.9 and trunk
dconf abstraction: allow reading /etc/dconf/**.
That's needed e.g. for Totem on current Debian Jessie.
Acked-By: Jamie Strandboge <jamie@canonical.com>
The '#!/usr/bin/env python' line in apparmor/rule/*.py is superfluous
and causes "non-executable script" rpmlint warnings on openSUSE.
Acked-by: Tyler Hicks <tyhicks@canonical.com>
TL;DR: the answer is "yes" ;-)
(see the patch for the question...)
Long version:
When creating a new child profile with aa-logprof or aa-genprof, the
child profile wasn't properly initialized in handle_children(), which
lead to a crash in delete_duplicates() later because capability etc.
was not set to a CapabilityRuleset etc. class and therefore
profile['capability'] didn't have a .delete_duplicates() method.
Funnily there was already a comment "do we need to init the profile here?"
This patch replaces the question in the comment with the answer.
Acked-by: Steve Beattie <steve@nxnw.org>
The local defines in the link_subset test collide and result in build
warnings. Replace the defines with a naming that won't collide and
makes it clear a local define for the test is being used.
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
1. The test is using the wrong defines: It is using the defines from the
parser for the packed dfa permissions. This set of permissions is not
meant to be exposed to the outside world
2. The kernel is using the wrong mapping function for the permissions
in the file class. This results in partially exposing the packed
permissions, but even then it doesn't fully line up with the packed
permissions, and is not correct for several of the potential permissions.
Attached is a patch that fixes the test, and moves the two tests that
fail due to the kernel to xpass.
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Tyler Hicks <tyhicks@canonical.com>
In the commit "Rev 3169: regression tests: have
ptrace use PTRACE_GETREGSET by default", I created
some ifdef magic to use the per arch general purpose
register data structures for various architectures,
including arm64. Unfortunately, in the upstream glibc commit
7d05a8168b
<bits/ptrace.h> is no longer included in the arm64 specific user.h,
which defined the structure as 'struct user_pt_regs'; instead user.h
was converted to define 'struct user_regs_struct'. Because of this, the
ptrace test fails to compile on arm64 when glibc is 2.20 or newer.
This patch adjusts the ptrace test to use the newer structure on arm64
if it's detected that a newer glibc is detected and reverts to using
the older one for older glibcs. It also adds an error when compiling
on architectures that haven't been incorporated yet.
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: John Johansen <john.johansen@canonical.com>
Besides adding this feature, this also fixes a crash in tools.py __init__():
AttributeError: 'Namespace' object has no attribute 'do_reload'
Acked-by: Steve Beattie <steve@nxnw.org>
create_new_profile() created a wrong structure for local_profile, which
resulted in an aa-genprof crash directly at startup (in the autodep
phase).
This patch fixes it to use the correct structure.
Acked-by: Steve Beattie <steve@nxnw.org>
Some of the newly added simple_tests contain lines like
profile foo@{FOO} { }
which are not supported by the tools because the '}' is in the same line,
while the tools expect \n as rule separator.
This patch changes those tests to
profile foo@{FOO} {
}
Acked-by: John Johansen <john.johansen@canonical.com>
cux and CUx are valid exec permissions, so they should be accepted
by validate_profile_mode() ;-)
Acked-by: John Johansen <john.johansen@canonical.com> for trunk and 2.9
Some of the include files added to simple_tests recently don't live in
one of the main include directories (includes/, includes-preamble/ or
include_tests/) which lets test-parser-simple-tests.py fail because
those files don't contain EXRESULT.
Instead of adding more exceptions to test-parser-simple-tests.py, this
patch adds DESCRIPTION and EXRESULT to those include files.
Acked-by: John Johansen <john.johansen@canonical.com>
- allow only a specific set of time units
- optionally allow whitespace between rlimit value and unit
- move check for invalid time units to time_to_int()
Also update the tests:
- add several tests with whitespace between value and unit
- change a test that used the (now invalid) "1m" to "1min"
- change the time_to_int() tests to use 'us' as default unit, and add
a test with 'seconds' as default unit
Acked-by: Steve Beattie <steve@nxnw.org>
currently the parser supports ambiguous units like m for time,
which could mean minutes or milliseconds. Fix this and refactor the
time parsing into a single routine.
Signed-off-by: John Johansen <john.johansen@canonical.com>
Signed-off-by: Steve Beattie <steve@nxnw.org>
When @{profile_name} is used within a rule matching expression any
aare expressions should be matched literally and not be interpreted as
aare.
That is
profile /foo/** { }
needs /foo/** to expand into a regular expression for its attachment
but, /foo/** is also the profiles literal name. And when trying to
match @{profile_name} in a rule, eg.
ptrace @{profile_name},
the variable needs to be expaned to
ptrace /foo/\*\*,
not
ptrace /foo/**,
that is currently happening.
BugLink: http://bugs.launchpad.net/bugs/1317555
equality tests by
Tyler Hicks <tyhicks@canonical.com>
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
The @{profile_name} is incorrectly expanded as a fully qualified path
including its namespace if one was specified in the profile declaration.
ie.
profile :ns://a {
ptrace @{profile_name},
# expands to
# ptrace :ns://a,
}
This is wrong however because within a profile if a rule refers
to a namespace it will be wrt a sub-namespace. That is in the above
example the ptrace rule is refering to a profile in a subnamespace
"ns".
Or from the current profile declaration scope
:ns//ns://a
Instead @{profile_name} should expand into the hname (hierarchical name),
which is the profile hierarchy specification within the namespace the
profile is part of.
In this case
a
or for a child profile case
profile :ns://a {
profile b {
ptrace @{profile_name},
}
}
the hname expansion would be
a//b
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
allow
@{FOO}=bar
/foo@{FOO} { }
to be expanded into
/foobar { }
and
@{FOO}=bar baz
/foo@{FOO} { }
to be expanded into
/foo{bar,baz} { }
which is used as a regular expression for attachment purposes
Further allow variable expansion in attachment specifications
profile foo /foo@{FOO} { }
profile name (if begun with profile keyword) and attachments to begin
with a variable
profile @{FOO} { }
profile /foo @{FOO} { }
profile @{FOO} @{BAR} {}
hats
^@{FOO}
hat @{FOO}
and for subprofiles as well
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
aa-logprof raises an exception if
- an include file contains a hat
- that file is included in a profile and
- aa-logprof hits an audit log entry for this profile
Reproducer ("works" on 2.9 and trunk):
python3 aa-logprof -f <(echo 'Jun 19 11:50:36 piorun kernel: [4474496.458789] audit: type=1400 audit(1434707436.696:153): apparmor="DENIED" operation="open" profile="/usr/sbin/apache2" name="/etc/gai.conf" pid=2910 comm="apache2" requested_mask="r" denied_mask="r" fsuid=0 ouid=0') -d ../profiles/apparmor.d/
This happens because profiles/apparmor.d/apache2.d/phpsysinfo was
already read when pre-loading the include files.
This patch changes aa.py parse_profile_data() to only raise the
exception if it is not handling includes currently.
Acked-by: Steve Beattie <steve@nxnw.org> for both trunk and 2.9.
Fix the regression that caused using 'include' instead of '#include' for
includes to stop working.
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
is_known_rule() ignored directory includes, which resulted in asking for
and adding superfluous rules that are already covered by a file in the
included directory.
This patch looks bigger than it is because it moves quite some lines
into the "else:" branch. Everything inside the "else:" just got an
additional whitespace level.
References: https://bugs.launchpad.net/apparmor/+bug/1471425
(however, trunk didn't crash, it "just" ignored directory includes)
Acked-by: Steve Beattie <steve@nxnw.org>
is_known_rule() in aa.py checked only direct includes, but not includes
in the included files. As a result, aa-logprof asked about things that
are already covered by an indirect include.
For example, the dovecot/auth profile includes abstractions/nameservice,
and abstractions/nameservice includes abstractions/nis, which contains
"capability net_bind_service,".
Nevertheless, aa-logprof asked to add capability net_bind_service.
Reproducer: (asks for net_bind_service without this patch, should not
ask for anything after applying the patch):
python3 aa-logprof -d ../profiles/apparmor.d/ -f <(echo 'type=AVC msg=audit(1415403814.628:662): apparmor="ALLOWED" operation="capable" profile="/usr/lib/dovecot/auth" pid=15454 comm="auth" capability=13 capname="net_bind_service"')
The patch adds code to check include files included by other include
files. Note that python doesn't allow to change a list while looping
over it, therefore we have to use "while includelist" as workaround.
This fixes a regression for network rules (this patch is based on the
old match_net_include() code). Funnily it "only" fixes capability rule
handling (without the "regression" part) because the old
match_cap_include() didn't do the recursive include handling.
Acked-by: Steve Beattie <steve@nxnw.org>
For some (not yet known) reason, we get file_perm events without
request_mask set, which causes an aa-logprof crash.
Reproducer log entry:
Jun 19 12:00:55 piorun kernel: [4475115.459952] audit: type=1400 audit(1434708055.676:19629): apparmor="ALLOWED" operation="file_perm" profile="/usr/sbin/apache2" pid=3512 comm="apache2" laddr=::ffff:193.0.236.159 lport=80 faddr=::ffff:192.168.103.80 fport=61985 family="inet6" sock_type="stream" protocol=6
This patch changes logparser.py to ignore those events.
References: https://bugs.launchpad.net/apparmor/+bug/1466812/
Acked-by: Steve Beattie <steve@nxnw.org> for trunk and 2.9
According to the parser test profiles (which are the only
"documentation" I found about this), definition of boolean variables
is only allowed outside profiles, not inside them.
parse_profile_data() got it the wrong way round, therefore this patch
fixes the condition and updates the error message.
Acked-by: Steve Beattie <steve@nxnw.org> for both trunk and 2.9.
Thanks to a bug in the apparmor.d manpage, NetworkRule rejected rules
that contained only TYPE (for example "network stream,"). A bugreport on
IRC and some testing with the parser showed that this is actually
allowed, so NetworkRule should of course allow it.
Note: not strip()ing rule_details is the easiest way to ensure we have
whitespace in front of the TYPE in TYPE-only rules, which is needed by
the RE_NETWORK_DETAILS regex.
Also adjust the tests to the correct behaviour.
Acked-by: Steve Beattie <steve@nxnw.org>
Instead of always showing a backtrace,
- for AppArmorException (used for profile syntax errors etc.), print only
the exceptions value because a backtrace is superfluous and would
confuse users.
- for other (unexpected) exceptions, print backtrace and save detailed
information in a file in /tmp/ (including variable content etc.) to
make debugging easier.
This is done by adding the apparmor.fail module which contains a custom
exception handler (using cgitb, except for AppArmorException).
Also change all python aa-* tools to use the new exception handler.
Note: aa-audit did show backtraces only if the --trace option was given.
This is superfluous with the improved exception handling, therefore this
patch removes the --trace option. (The other aa-* tools never had this
option.)
If you want to test the behaviour of the new exception handler, you can
use this script:
#!/usr/bin/python
from apparmor.common import AppArmorException, AppArmorBug
from apparmor.fail import enable_aa_exception_handler
enable_aa_exception_handler()
# choose one ;-)
raise AppArmorException('Harmless example failure')
#raise AppArmorBug('b\xe4d bug!')
#raise Exception('something is broken!')
Acked-by: Seth Arnold <seth.arnold@canonical.com>
As shown in parser/tst/simple_tests/profile/flags/flags_ok_whitespace.sd,
the parser is quite tolerant to additional or missing whitespace around
flags=, while the tools are more strict.
This patch updates the RE_PROFILE_START regex to follow this tolerance.
Acked-by: Steve Beattie <steve@nxnw.org>.
The only difference between PROFILE_MODE_RE and PROFILE_MODE_NT_RE
was that the latter one additionally allowed 'x', which looks wrong.
(Standalone 'x' is ok for deny rules, but those are handled by
PROFILE_MODE_DENY_RE.)
This patch completely drops PROFILE_MODE_NT_RE and the related code in
validate_profile_mode().
Also wrap the two remaining regexes in '^(...)+$' instead of doing it
inside validate_profile_mode(). This makes the code more readable and
also results in a 2% performance improvement when parsing profiles.
Acked-by: Steve Beattie <steve@nxnw.org> for trunk and 2.9.
Add the missing "pux" to PROFILE_MODE_RE and PROFILE_MODE_NT_RE.
Also move those regexes and PROFILE_MODE_DENY_RE directly above
validate_profile_mode() which is the only user.
Acked-by: Steve Beattie <steve@nxnw.org> for trunk and 2.9
Parsing of boolean assignments failed with
TypeError: '_sre.SRE_Match' object is not subscriptable
because of a missing ".groups()"
Acked-by: Steve Beattie <steve@nxnw.org> for trunk and 2.9
Errors include typos ("DESCRIPT__ON"), missing value after #=EXRESULT
and #=EXRESULT=PASS (= instead of space).
Acked-by: Steve Beattie <steve@nxnw.org> for trunk and 2.9
Bug: https://bugs.launchpad.net/apparmor/+bug/1470985
The ptrace regression test fails to compile on the arm64 platform,
because it uses PTRACE_GETREGS and not the newer PTRACE_GETREGSET
interface for getting access to arch-specific register information[0].
However, fixing it is complicated by the fact that the struct name
for for the general purpose registers is not named consistently
across architectures. This patch attempts to address those issues,
and compiles at least on i386, amd64, arm64, arm (armhf), ppc64,
and ppc64el. The test is verified to continue to function correctly
on i386 and amd64.
[0] https://sourceware.org/ml/archer/2010-q3/msg00193.html
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: John Johansen <john.johansen@canonical.com>
RlimitRule accidently used 'ms' (milliseconds) as default unit for
rttime rules, but rttime without unit means 'us' (microseconds). This
patch fixes this.
Also add some tests with 'us' as unit, and two more to cover terribly
invalid corner cases (and to improve test coverage by 2 lines ;-)
Acked-by: John Johansen <john.johansen@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
Change minitools tests to use AATest and work inside a tmpdir.
This results in lots of changes ('./profiles' -> self.profile_dir,
local_profilename -> self.local_profilename etc.) and also moves some
code from the global area to AASetup().
Also drop the no longer needed clean_profile_dir() and add linebreaks
in assert* calls with a long error message specified.
Acked-by: Steve Beattie <steve@nxnw.org>
It's allowed to only specify a TYPE without specifying a DOMAIN.
Also add a missing "]" for QUALIFIERS.
Acked-by: Seth Arnold <seth.arnold@canonical.com>
The current rule simplification algorithm has issues that need to be
addressed in a rewrite, but it is still often a win, especially for
larger profiles.
However doing rule simplification as a single pass limits what it can
do. We default to right simplification first because this has historically
shown the most benefits. For two reasons
1. It allowed better grouping of the split out accept nodes that we
used to do (changed in previous patches)
2. because trailing regexes like
/foo/**,
/foo/**.txt,
can be combined and they are the largest source of node set
explosion.
However the move to unique node sets, eliminates 1, and forces 2 to
work within only the single unique permission set on the right side
factoring pass, but it still incures the penalty of walking the whole
tree looking for potential nodes to factor.
Moving tree simplification into the construction phases gets rid of
the need for the right side factoring pass to walk other node sets
that will never combine, and since we are doing simplification we can
do it before the cat and permission nodes are added reducing the
set of nodes to look at by another two.
We do loose the ability to combine nodes from different sets during
the left factoring pass, but experimentation shows that doing
simplification only within the unique permission sets achieve most of
the factoring that a single global pass would achieve.
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
Currently rules are added to the expression tree in order, and then
tree simplification and factoring is done. This forces simplification
to "search" through the tree to find rules with the same permissions
during right factoring, which dependent on ordering of factoring may
not be able to group all rules of the same permissions.
Instead of having tree factoring do the work to regroup rules with the
same permissions, pregroup them as part of the expr tree construction.
And only build the full tree when the dfa is constructed.
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
accept nodes per perm bit where done from the very begining in a
false belief that they would help produce minimized dfas because
a nfa states could share partial overlapping permissions.
In reality they make tree factoring harder, reduce in longer nfa
state sets during dfa construction and do not result in a minimized
dfa.
Moving to unique permission sets, allows us to minimize the number
of nodes sets, and helps reduce recreating each set type multiple
times during the dfa construction.
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
parser_regex.c includes libapparmor_re/aare_rules.h and thus it should
depend on it in the Makefile.
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
LSMs, such as AppArmor, aren't consulted when a program calls access(2).
This can result in access(2) returning 0 but a subsequent open(2)
failing.
The aa-status utility was doing the access() -> open() sequence and we
became aware of a large number of tracebacks due to open() failing for
lack of permissions. This patch catches any IOError exceptions thrown by
open(). It continues to print the same error message as before when
access() failed but also prints that error message when AppArmor blocks
the open of the apparmorfs profiles file.
https://launchpad.net/bugs/1466768
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
The function is basically a wrapper around a regex, so regex.py is a
much better home.
While on it, rename the regex to RE_INCLUDE, change it to named matches,
use RE_EOL to handle comments and compile it outside the function, which
should result in a (small) performance improvement.
Also rewrite re_match_include(), let it check for empty include
filenames ("#include <>") and let it raise AppArmorException in that
case.
Finally, adjust code calling it to the new location, and add some tests
for re_match_include()
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
profile_storage() returns an empty, properly initialized profile.
It doesn't explicitly init all keys (yet) and will be extended over
time, with the final goal to get rid of hasher().
Also change various places in aa.py to use it (instead of an empty
hasher or sub-hasher), and remove various "init rule class (if not done
yet)" cases.
This also avoids a crash in aa-cleanprof remove_duplicate_rules().
Hats weren't properly initialized in aa.py parse_profile_data()
(especially rule classes were missing), which caused a crash because
hasher doesn't support the delete_duplicates() method.
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
Change hat declarations ("^hat,") are no longer supported (see previous
patch for details). Therefore remove support for writing them.
This also means to completely remove the 'declared' flag, which was only
needed for hat declarations, and was (after the previous patch) always
set to False.
Also add a hat to the cleanprof_test.{in,out} test profile to make sure
aa-cleanprof doesn't break hats, and a hat declaration with the same
name to make sure it gets removed and doesn't break the "real" hat.
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
Hat declarations ("^hat,") were added in 2.3 for declaring external
hats, but in the meantime aren't supported by the parser anymore (tested
with 2.9.2 parser).
Additionally, if a profile contains both a hat declaration and the hat
("^hat { ...}"), the hat declaration can overwrite the content of the
hat on a "last one wins" base.
This is caused by setting 'declared' to True, which means write_piece()
will only write the "^hat," line, but not the "^hat { ... }" block.
Therefore no longer set 'declared' to True, print a warning that hat
declarations are no longer supported, and ignore the rule. This also
means that running aa-cleanprof can make the profile valid again :-)
Also no longer change 'hat' when hitting a profile declaration, which
also looks wrong.
Note: This change removes the only usage of 'declared'. A follow-up
patch (trunk only) will completely remove the 'declared' handling.
Reproducer profile (run aa-cleanprof on it):
(will crash in remove_duplicate_rules() 80% of the time - if so, try
multiple times. One of the next patches will fix that. Or just try 2.9,
which doesn't have the crash in remove_duplicate_rules().)
/usr/bin/true {
^FOO {
capability setgid,
}
# deletes the content of ^FOO when saving the profile! (last one wins)
# additionally, the parser says this is invalid syntax
^FOO,
}
See also the "Hat declarations" thread on the ML,
https://lists.ubuntu.com/archives/apparmor/2015-June/008107.html
Acked-by: Kshitij Gupta <kgupta8592@gmail.com> for both 2.9 and trunk.
Change aa.py to use RlimitRule and RlimitRuleset instead of a sub-hasher
to store and write rlimit rules. In detail:
- drop all rlimit rule parsing from parse_profile_data() and
serialize_profile_from_old_profile() - instead, just call
RlimitRule.parse()
- change write_rlimits() to use RlimitRuleset
- add removal of superfluous/duplicate change_profile rules (the old
code didn't do this)
- update the comment about aa[profile][hat] usage - rlimit and
change_profile are no longer dicts.
Also cleanup RE_PROFILE_RLIMIT in regex.py - the parenthesis around
'<=' are no longer needed.
Note: This patch is quite small because aa-logprof doesn't ask for
rlimit rules.
I tested all changes manually with aa-cleanprof and aa-logprof (adding
some file rules, rlimit rules kept unchanged)
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
check-logprof in profiles/Makefile needs the local/* files.
Add a dependency to make sure they are generated.
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
Only use the special %exception directive for functions that return a
negative int and set errno upon error.
This prevents, for example, _aa_is_blacklisted() from raising an
exception when it returns -1. This is important because it doesn't set
errno so an exception based on the value of errno would be
unpredictable.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
When is_blacklisted() was internal to the parser, it would print an
error message when encountering some file names. If the path parameter
was non-null, the error message would include the file path instead of
the file name.
Now that the function has been moved to libapparmor, callers are
expected to print the appropriate error message if _aa_is_blacklisted()
returns -1. Since the error message printing no longer occurs inside of
_aa_is_blacklisted(), the path parameter can be removed.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
Ignore README files when performing an operation on a list of files.
This matches the behavior of the is_skipped_file() function in aa.py.
The hope is that is_skippable_file() can reuse _aa_is_blacklisted().
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
It looks odd to access the first character of a string before checking
to see if the string's length is zero. This is actually fine, in
practice, since strlen() looks at the first character of the string for
the presence of '\0' which means this is entirely a cosmetic change.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Prepend the function prototypes with extern to match the style of the
existing prototypes.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
The errno values libapparmor's aa_policy_cache_new() uses to indicate
when the cache directory does not exist and when an existing, invalid
cache already exists needed to be separated out. They were both ENOENT
but now the latter situation uses EEXIST.
libapparmor also needed to be updated to not print an error message to
the syslog from aa_policy_cache_new() when the max_caches parameter is
0, indicating that a new cache should not be created, and the cache
directory does not exist. This is an error situation but a debug message
is more appropriate.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
Create a section 3 man page for the aa_policy_cache family of functions.
Additionally, update the in-code descriptions to match the descriptions
in the man page.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
Create a section 3 man page for the aa_kernel_interface family of
functions. Additionally, update the in-code descriptions to match the
descriptions in the man page.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
Create a section 3 man page for the aa_features family of functions.
Additionally, update the in-code descriptions to match the descriptions
in the man page.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johanse@canonical.com>
The aa_policy_cache_new() and aa_policy_cache_remove() functions are
changed to accept a dirfd parameter.
The cache dirfd (by default, /etc/apparmor.d/cache) is opened earlier in
aa_policy_cache_new(). Previously, the directory wasn't accessed until
later in the following call chain:
aa_policy_cache_new() -> init_cache_features() -> create_cache()
Because of this change, the logic to create the cache dir must be moved
from create_cache() to aa_policy_cache_new().
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
Instead of only accepting a path in the aa_features API, accept a
directory file descriptor and a path like then openat() family of
syscalls. This type of interface is better since it can operate exactly
like a path-only interface, by passing AT_FDCWD or -1 as the dirfd.
However, using the dirfd/path combination, it can eliminate string
allocations needed to open files in subdirectories along with the
even more important benefits mentioned in the open(2) man page.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Make the function prototype for reading a features directory the same
as the function prototype for reading a features file.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Two different implementations were in use for reading features files.
One for reading a single file and another for reading a single file
after walking a directory. This patch creates a single function that is
used in both cases.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
The _aa_dirat_for_each() function used the DIR * type for its first
parameter. It then switched back and forth between the directory file
descriptors, retrieved with dirfd(), and directory streams, retrieved
with fdopendir(), when making syscalls and calling the call back
function.
This patch greatly simplifies the function by simply using directory
file descriptors. No functionality is lost since callers can still
easily use the function after calling dirfd() to retrieve the underlying
file descriptor.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
The most common case when creating an aa_kernel_interface object will be
to do so while using the current kernel's feature set for the
kernel_features parameter. Rather than have callers instantiate their
own aa_features object in this situation, aa_kernel_interface_new()
should do it for them if they specify NULL for the kernel_features
parameter.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
The most common case when creating an aa_policy_cache object will be to
do so while using the current kernel's feature set for the
kernel_features parameter. Rather than have callers instantiate their
own aa_features object in this situation, aa_policy_cache_new() should
do it for them if they specify NULL for the kernel_features parameter.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
The aa_features object that is passed to aa_policy_cache_new() does not
have to represent the currently running kernel. It may represent a
different kernel, such as a kernel that was just installed, that is not
currently running.
This patch adjusts the function comments to remove mentions of
"... the currently running kernel".
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
This patch changes the aa_policy_cache_new() prototype and gets rid of
aa_policy_cache_is_valid() and aa_policy_cache_create().
The create bool of aa_policy_cache_new() is replaced with a 16 bit
unsigned int used to specify the maximum number of caches that should be
present in the specified cache directory. If the number is exceeded, the
old cache directories are reaped. The definition of "old" is private to
libapparmor and only 1 cache directory is currently supported. However,
that will change in the near future and multiple cache directories will
be supported.
If 0 is specified for the max_caches parameter, no new caches can be
created and only an existing, valid cache can be used. An error is
returned if no valid caches exist in that case.
If UINT16_MAX is specified, an unlimited amount of caches can be created
and reaping is disabled.
This means that 0 to (2^16)-2, or infinite, caches will be supported in
the future.
This change allows for the parser to continue to support the
--skip-bad-cache (by passing 0 for max_caches) and the --write-cache
option (by passing 1 or more for max_caches) without confusing
libapparmor users with the aa_policy_cache_{is_valid,create}()
functions.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
The default change_onexec id is slightly wrong, it allows matching
'/' as an executable but it really should be anything under /
This results in the equality tests for change_profile failing as it
is different than what specifying /** in a rule does.
We could define rules need to be {/,}** to be equivalent but since
/ can not be an executable change the default value to match what
/** is converted in to.
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
bison isn't properly handling the 3 options of
TOK_CHANGE_PROFILE opt_id TOK_END_OF_RULE
TOK_CHANGE_PROFILE opt_id TOK_ARROW TOK_ID TOK_END_OF_RULE
TOK_CHANGE_PROFILE opt_id TOK_ARROW TOK_COLON TOK_ID TOK_COLON TOK_END_OF_RULE
specifying
change_profile /exec,
results in an unexpected TOK_ID error
refactor so that they share the 3 options share a common head which fixes
the problem.
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
While change_profile rules are always created separately from file
rules. The merge phase can result in change_profile rules merging
with file rules, resulting in the change_profile permission being
set when a file rule is created.
Make sure to screen off the change_profile permission, when creating
a file rule.
Note: the proper long term fix is to split file, link and change_profile
rules into their own classes.
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
Note: this patch currently overlays onexec with link_name to take
advantage of code already being used on link_name. Ideally what needs
to happen is entry needs to be split into file, link and change_profile
entry classes.
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
Add two variable references (aa and changed) in aa-mergeprof
ask_the_questions() so that the code can use the short name and be more
in sync with aa.py ask_the_questions().
With this patch applied, the "for ruletype in ['capability', 'network']:"
block is in sync, with the exception of the sections that intentionally
differ:
- the check for the profile mode
- the default button selection based on profile mode
- the seen_events counter
The patch also includes some minor whitespace fixes.
Acked-by: Steve Beattie <steve@nxnw.org>
Applying patches often creates *.orig files, and those files are quite
annoying in the "bzr status" output and also in the "unknown" file list
when commiting.
Note: I intentionally don't want to add *.rej files - while those files
should never end up in bzr, they are important enough to be listed in
bzr status output.
Acked-by: Steve Beattie <steve@nxnw.org>
flags_bad.sd contains multiple failures. Split the file into multiple
files with one failure in each and, while on it, using more helpful
filenames.
Acked-by: Steve Beattie <steve@nxnw.org>
The following patch:
- removes re import
- uses apparmor.re_match_include instead of the regex
which also means to use the correct regex instead of
the slightly wrong one cleanprofile.py had
Acked-by: Christian Boltz <apparmor@cboltz.de>
The cleanprofile.py has an apparmor import, this patch modifies the import to make it consistent with the rest of modules.
Acked-by: Christian Boltz <apparmor@cboltz.de>
The following patch:
- Brings the return to the correct indentation
- Adds a sorted call over the set keys of hat in the profile
Acked-by: Christian Boltz <apparmor@cboltz.de> for trunk and 2.9.
After switching to winbindd as test profile, comments about the ntpd
profile don't make sense anymore ;-)
The patch also includes some whitespace fixes.
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
Acked-by: Steve Beattie <steve@nxnw.org>
This time we only have 98% coverage (some missing and partial) because
I didn't find corner cases that raise some exceptions ;-)
Acked-by: Steve Beattie <steve@nxnw.org>
The class comes with the usual set of features, so I'll only mention a
special feature: the is_covered() and is_equal() functions can even
compare limits with different units (for example they recognize that
2minutes == 120seconds).
Also change RE_PROFILE_RLIMIT:
- make it a bit more strict (the old one accepted any chars, including
spaces, for rlimit and value)
- convert it to named matches
- '<=' isn't optional - remove the '?' (but keep the parenthesis to
avoid breaking parsing in aa.py)
- allow rules with no spaces around '<='
Acked-by: Steve Beattie <steve@nxnw.org>
aa-cleanprof (actually clean_profile() in tools.py) used reload_base()
from aa.py which sends the parser output to /dev/null. This had two
effects:
- aa-cleanprof ignored the --no-reload parameter
- there was no error message because reload_base() /dev/null's the
parser output
This patch changes clean_profile() to use reload_profile() from tools.py
(which honors the --no-reload option).
Also add a TODO note to aa.py reload_base(), the (AFAIK only) winner of
the 'useless use of cat' award in the AppArmor code.
We should really change it to use reload_profile(), even if that means
moving the function from tools.py to aa.py or common.py. And it should
not /dev/null the apparmor_parser output. ;-)
References: https://bugs.launchpad.net/apparmor/+bug/1443637
Acked-by: Steve Beattie <steve@nxnw.org>
aa-complain is part of the enforce/complain/disable triple. Therefore
I expect it to actually load a profile in complain mode.
To do this, it has to delete the 'disable' symlink, but set_complain()
in aa.py didn't do this (and therefore kept the profile disabled).
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
Users might expect that setting a profile into audit mode also activates
it (which shouldn't happen IMHO because the audit flag is not part of
the enforce/complain/disable triple), so we should at least tell them.
References: https://bugs.launchpad.net/apparmor/+bug/1429448
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
aa-complain, aa-enforce, aa-disable and aa-audit refused to change
profiles for non-existing binaries. This patch also allows paths
starting with /. This also makes it possible to use
aa-complain '/{usr/,}bin/ping'
and
aa-complain /etc/apparmor.d/bin.ping
This patch fixes https://bugs.launchpad.net/apparmor/+bug/1416346
Well, mostly - we still need to decide how we handle wildcards in
profile names:
aa-complain ping
aa-complain /usr/bin/ping
will still error out with "Profile not found" because it isn't an exact
match (and matching the wildcard would change more than the user wants).
Oh, and this patch also fixes the last failure in minitools_test.py.
Acked-by: Steve Beattie <steve@nxnw.org> for trunk and 2.9
Change minitools_test.py to use the winbind instead of the ntpd profile
for testing. The tests broke because the ntpd profile has the
attach_disconnected flag set now, and therefore didn't match the
expected flags anymore.
Also replace the usage of filecmp.cmp() in the cleanprof test with
reading the file and using assertEqual - this has the advantage that we
get a full diff instead of just "files differ".
Note: The aa-cleanprof test is still failing because of a bug in
tools.py, but will be fixed by the next patch.
See https://bugs.launchpad.net/apparmor/+bug/1416346 for details.
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
This allows to run minitools_test.py as non-root user.
Also add a check that only creates the force-complain directory if it
doesn't exist yet.
Note: With this patch applied, there are still 4 failing tests, probably
caused by changes in the profiles that are used in the tests.
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
Add a --no-reload parameter to aa-audit, aa-cleanprof, aa-complain,
aa-disable and aa-enforce. This makes it possible to change the
profile flags without reloading the profile.
Also change tools.py to honor the --no-reload parameter.
References: https://bugs.launchpad.net/apparmor/+bug/1458480
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
Acked-by: Steve Beattie <steve@nxnw.org> for trunk and 2.9.
--fixes lp:1458480
The function will return the 'Exec Condition' and the 'Target Profile'
as nice list to use in aa-logprof (once we have support for
change_profile in logparser.py) and aa-mergeprof.
Also add some tests to ensure the correct result.
Acked-by: Steve Beattie <steve@nxnw.org>
This allows to drop the "apparmor.aa." prefix in ask_the_question() to
get the code more in sync with aa.py ask_the_question().
Acked-by: Steve Beattie <steve@nxnw.org>
Replace the code in aa.py ask_the_questions() that handles network rules
with the ask_the_questions() code initially copied from aa-mergeprof.
This means to convert the network/netdomain log events to a
NetworkRuleset stored in the log_obj hasher, and then let the code from
aa-mergeprof operate on this hasher.
The user interface is mostly unchanged, with two exceptions:
- options always displayed, even if there is only one option
- some slightly changed texts
If you didn't understand why there's a need for the previous patch, this
one should explain it :-)
This also ends up fixing at least one bug where the 'audit' keyword
wasn't listed as a separate qualifier, but instead showed up smooshed
into the Network Family header.
Acked-by: Steve Beattie <steve@nxnw.org>
Replace the code in aa.py ask_the_questions() that handles capabilities
with the ask_the_questions() code from aa-mergeprof.
This means to convert the capability log events to a CapabilityRuleset
stored in the (new) log_obj hasher, and then let the code from
aa-mergeprof operate on this hasher.
Most of the code after the "aa-mergeprof also has this code" comment is
a direct copy of the aa-mergeprof code, with the following changes:
- filter for profile mode (enforce/complain)
- set default button (allow or deny) based on profile mode
- keep seen_events counter happy (even if it isn't displayed anywhere)
- replace apparmor.aa.foo with just foo
The user interface is mostly unchanged, with two exceptions:
- options always displayed, even if there is only one option
- some slightly changed texts
Acked-by: Steve Beattie <steve@nxnw.org>
When switching the audit flag for network events in aa-logprof
(technically, it happens in aa.py ask_the_question()), the "(I)gnore"
button gets "lost".
This patch fixes the list of available buttons.
I propose this patch for trunk and 2.9.
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
Acked-by: Steve Beattie <steve@nxnw.org> for trunk and 2.9
Move the code to set q.headers, q.functions and q.default for network
and capability rules inside the "while not done" loop. This ensures to
always have valid headers (for example, after changing the audit
qualifier, the severity was "lost" before) and avoids some duplicated
code.
Also drop a useless "if True:" condition and change the whitespace of
the following lines.
Acked-by: Steve Beattie <steve@nxnw.org>
Now that the handling for capability and network rules is the same,
wrap the former network rule-only code with
for ruletype in ['capability', 'network']:
and delete the superfluous ;-) capabiltiy code block.
Needless to say that future updates for other rule types will be
quite easy ;-)
Acked-by: Steve Beattie <steve@nxnw.org>
BaseRule:
- add logprof_header() - sets the 'Qualifier' (audit, allow/deny) header
if a qualifier is specified, calls logprof_header_localvars() and then
returns an array of headers to display in aa-logprof and aa-mergeprof
- add logprof_header_localvars() - dummy function that needs to be
implemented in the child classes
NetworkRule: add logprof_header_localvars() - adds 'Network Family'
and 'Socket Type' to the headers
CapabilityRule: add logprof_header_localvars() - adds 'Capability' to
the headers
Also change aa-mergeprof to use rule_obj.logprof_header() for network
and capability rules. This means deleting lots of lines (that moved to
the *Rule classes) and also deleting the last differences between
capabiltiy and network rules.
Finally add tests for the newly added functions.
Acked-by: Steve Beattie <steve@nxnw.org>
This means:
a) for capability rules:
- move audit and deny to a new "Qualifier" header (only displayed if
non-empty)
- always display options, even if only one is available
- use available_buttons(), which means to add the CMD_AUDIT_* button
- add handling for CMD_AUDIT_* button
- CMD_ALLOW: only add rule_obj if the user didn't select a #include
- move around some code to get it in sync with network rule handling
b) for network rules
- move audit and deny to a new "Qualifier" header (only displayed if
non-empty)
- call rule_obj.severity() (not implemented for network rules, does
nothing)
- change messages to generic 'Adding %s to profile.'
- move around some code to get it in sync with capability rule
handling
The only remaining difference is in q.headers[] and the variables
feeding it:
- capability rules show "Capability: foo"
- network rules show "Network Family: foo" and "Socket type: bar"
Acked-by: Steve Beattie <steve@nxnw.org>
Note: the != sev_db.NOT_IMPLEMENTED: check in aa-mergeprof is
superfluous for capabilities, but will become useful once this code
block is used for other rule types.
Acked-by: Steve Beattie <steve@nxnw.org>
Also implement handling for the special capability value '__ALL__' in
severity.py, which is used for 'capability,' rules (aa-mergeprof might
need to display the severity for such a rule).
Finally, add some tests for severity() in test-capability.py and a test
for '__ALL__' in test-severity.py.
Acked-by: Steve Beattie <steve@nxnw.org>
severity() will, surprise!, return the severity of a rule, or
sev_db.NOT_IMPLEMENTED if a *Rule class doesn't implement the severity()
function.
Also add the NOT_IMPLEMENTED constant to severity.py, and a test to
test-baserule.py that checks the return value in BaseRule.
Acked-by: Steve Beattie <steve@nxnw.org>
allow specifying the change_profile keyword
change_profile,
to grant all permissions change_profile permissions
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Tyler Hicks <tyhicks@canonical.com>
The parser currently is still using the old permission layout, the kernel
uses a newer layout that allows for more permission bits. The newer
newer permission layout is needed by the library to query the kernel,
however that causes some of the permission bits to be redefined.
Rename the permission bits that cause redefination warnings to use
AA_OLD_MAY_XXX
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Tyler Hicks <tyhicks@canonical.com>
Currently the cache file has its mtime set at creation time, but this
can lead to cache issues when a policy file is updated separately from
the cache. This makes it possible for an update to ship a policy file
that is newer than the what the cache file was generated from, but
result in a cache hit because the cache file was local compiled after
the policy file was package into an update (this requires the update
to set the mtime of the file when locally installed to the mtime of
the file in its update archive but this is commonly done, especially
in image based updates).
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
When caching was converted to use mtime instead of ctime, the cache
file timestamp did not get switched over. This means we are comparing
the cache file's ctime against the policy file's mtime. Which can make
the cache look newer than it really is.
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
- missing formatting code prefixes, usually I for BNFish arguments
- added blank lines before preformatted sections as the html formatter
wasn't treating them as seperate from the preceding text (also, they
generated podchecker warnings)
- fixed a grammar issue
- fixed link description text block that was mistakenly indented and
thus treated as preformatted text
- moved the "Qualifier Blocks" subsection out of the =over/=back as
all the pod tools did not like this and it caused podchecker to exit
with an error, breaking builds that ran make check on the parser
tree.
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Rename require_features to require_kernel_features and
have_features to kernel_features
to indicate they are tests for kernel features, as now there are tests
for parser features and in the future there might be library features
as well.
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Tyler Hicks <tyhicks@canonical.com>
has not been updated. The issue is that the regression tests detect the
kernel features set and generate policy that the parser may not be able
to compile.
Augment the regressions tests with a couple simple functions to test what
is supported by the parser, and update the test conditionals to use them.
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
This option was previously only documented in the --help output.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
I decided to use a "small" solution for now, which basically means
s/unittest.TestCase/AATest/, cleanup of some setUp() and renaming the
remaining setUp() functions to AASetup().
This doesn't mean an instant win (like in test-severity.py), but allows
to add tests with a tests[] array.
Acked-by: Steve Beattie <steve@nxnw.org>
To be able to distinguish between severity 10 and unknown severity,
change AASetup to specify 'unknown' as default rank, and change the
expected result to 'unknown' where it's expected.
Also change the "expected rank %d" to "%s" because it can be a string
now, and add a test that contains directories with different severity
in one variable.
After these changes, handle_variable_rank() errors out with
TypeError: unorderable types: str() > int()
so fix it by
- initializing rank with the default rank (instead of none)
- explicitely check that rank and rank_new are != the default rank before
doing a comparison
A side effect is another bugfix - '@{HOME}/sys/@{PROC}/overcommit_memory'
is severity 4, not 10 or unknown (confirmed by reading severity.db).
Acked-by: Steve Beattie <steve@nxnw.org>
This simplifies test-severity.py a lot:
- lots of test functions are replaced with tests[] arrays
- tempdir handling and cleanup is now done automagically
Even if test-severity.py shrunk by 65 lines, all tests are still there.
There's even an addition - SeverityTestCap now additionally verifies the
result of rank_capability().
Acked-by: Steve Beattie <steve@nxnw.org>
Change rank_capability() so that it doesn't expect the CAP_ prefix.
This makes usage easier because callers can simply hand over the
capability name.
Also change rank() to call rank_capability() without the CAP_ prefix.
Acked-by: Steve Beattie <steve@nxnw.org>
Replace rule-specific names with generic names:
- s/'capability'/ruletype/
- s/cap_obj/rule_obj/
- s/'network'/ruletype/
- s/net_obj/rule_obj/
Also set ruletype at the beginning of each block.
The long-term goal is to have
for ruletype in ['capability', 'network', ...]:
with common code to handle all rule types, and having common names makes
it easier to compare the blocks.
Acked-by: Steve Beattie <steve@nxnw.org>
aa-mergeprof has some sections where it first resets the 'deleted'
variable, and then overwrites it again a line or two later.
This patch removes the superfluous variable resets.
Acked-by: Steve Beattie <steve@nxnw.org>
Add a check to parse_profile_data() to detect if a file contains two
profiles with the same name.
Note: Two profiles with the same name, but in different files, won't be
detected by this check.
Also add basic tests to ensure that a valid profile gets parsed, and two
profiles with the same name inside the same file raise an exception.
(Sidenote: these simple tests improve aa.py coverage from 9% to 12%,
which also confirms the function is too long ;-)
Acked-by: Steve Beattie <steve@nxnw.org>
Add writeTmpfile() to AATest to write a file into the tmpdir. If no
tmpdir exists yet, automatically create one.
createTmpdir() is a separate function so that it's possible to manually
create the tmpdir (for example, if a test needs an empty tmpdir).
Also add a tearDown() function to delete the tmpdir again. This function
calls self.AATeardown() to avoid the need for super() in child classes.
Finally, simplify AaTestWithTempdir in test-aa.py to use createTmpdir()
and add an example for AATeardown() to test-example.py.
Acked-by: Steve Beattie <steve@nxnw.org>
aa-mergeprof no longer calls match_net_includes(), which means the
function can be dropped.
After that, match_net_include() is also unused, so also drop it.
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
aa-mergeprof still used the old aa[profile][hat][allow]['netdomain']
which no longer gets populated. This resulted in not asking for merging
any network rules.
This patch changes ask_the_question() to the NetworkRule(set) layout.
Besides that,
- don't ask for network rules that are already covered.
Using is_known_rule() also fixes
https://bugs.launchpad.net/apparmor/+bug/1382241
- include the audit keyword in the "Network Family" headline
(I'd prefer to just use the get_clean() rule, but that's another topic)
- hide "(A)llow" when merging a deny rule
- as a side effect of using NetworkRule, fix crashes for 'network,' and
'network foo,' rules
To avoid having to repeat the list of available "buttons" and the logic
to update that list, add a available_buttons() function that returns the
list of available buttons depending on rule_obj.deny and rule_obj.audit
to aa.py, and import it into mergeprof.
I tested all changes manually.
Acked-by: Steve Beattie <steve@nxnw.org>
aa-mergeprof still used the old aa[profile][hat][allow]['capability']
which no longer gets populated - which resulted in not asking for
merging any capabilities.
Actually (and funnily),
- if other.aa[profile][hat].get(allow, False):
- continue
resulted in never merging capability rules even before the change to
CapabilityRule(set) - this was meant as optimization, but a "not" was
missing in the condition ;-) so it always skipped capability rules.
The patch changes ask_the_question to the CapabilityRule(set) layout.
Besides that,
- include the audit and deny keywords in the "Capability" headline
(I'd prefer to just use the get_clean() rule, but that's another topic)
- hide "(A)llow" when merging a deny rule
- don't ask for capabilities that are already covered
Also delete match_cap_includes() from aa.py, which is no longer used.
Acked-by: Steve Beattie <steve@nxnw.org>
Bug: https://launchpad.net/bugs/1382241
Also rename RE_PROFILE_CHANGE_PROFILE_2 to RE_PROFILE_CHANGE_PROFILE
and update apparmor/rule/change_profile.py to use the changed name.
Acked-by: Steve Beattie <steve@nxnw.org>
Change aa.py to use ChangeProfileRule and ChangeProfileRuleset instead
of a sub-hasher to store and write change_profile rules. In detail:
- drop all the change_profile rule parsing from parse_profile_data() and
serialize_profile_from_old_profile() - instead, just call
ChangeProfileRule.parse()
- change write_change_profile to use ChangeProfileRuleset
- add removal of superfluous/duplicate change_profile rules (the old
code didn't do this)
Note that this patch is much smaller than the NetworkRule and
CapabilityRule patches because aa-logprof doesn't ask for adding
change_profile rules - adding that is something for a later patch.
Acked-by: Steve Beattie <steve@nxnw.org>
Add utils/apparmor/rule/change_profile.py with the ChangeProfileRule and
ChangeProfileRuleset classes. These classes are meant to handle
change_profile rules.
In comparison to the current code in aa.py, ChangeProfileRule has some
added features:
- support for audit and allow/deny keywords (for which John promised a
parser patch really soon)
- support for change_profile rules with an exec condition
Also add the improved regex RE_PROFILE_CHANGE_PROFILE_2 to regex.py.
Acked-by: Steve Beattie <steve@nxnw.org>
It did this in the old 2.8 code, but didn't in 2.9.x (first there was a
broken hat regex, then I commented out the hat handling to avoid
breakage caused by the broken regex).
This patch makes sure the hat flags get set when setting the flags for
the main profile.
Also change RE_PROFILE_HAT_DEF to use more named matches
(leadingwhitespace and hat_keyword). Luckily all code that uses the
regex uses named matches already, which means adding another (...) pair
doesn't hurt.
Finally adjust the tests:
- change _test_set_flags to accept another optional parameter
expected_more_rules (used to specify the expected hat definition)
- add tests for hats (with '^foobar' and 'hat foobar' syntax)
- add tests for child profiles, one of them commented out (see below)
Remaining known issues (also added as TODO notes):
- The hat and child profile flags are *overwritten* with the flags used
for the main profile. (That's well-known behaviour from 2.8 :-/ but we
have more flags now, which makes this more annoying.)
The correct behaviour would be to add or remove the specified flag,
while keeping other flags unchanged.
- Child profiles are not handled/changed if you specify the 'program'
parameter. This means:
- 'aa-complain smbldap-useradd' or 'aa-complain /usr/sbin/smbldap-useradd'
_will not_ change the flags for the nscd child profile
- 'aa-complain /etc/apparmor.d/usr.sbin.smbldap-useradd' _will_ change
the flags for the nscd child profile (and any other profile and
child profile in that file)
Even with those remaining issues (which need bigger changes in
set_profile_flags() and maybe also in the whole flags handling), the
patch improves things and fixes the regression from the 2.8 code.
Acked-by: Steve Beattie <steve@nxnw.org> for trunk and 2.9
The test program was querying its own profile. Adjust the profile
generation so that a separate profile is generated and have query_label
query the separate profile.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
Adjust the internal splitcon() function to strip a single trailing
newline character when the bool strip_newline argument is true.
aa_getprocattr_raw(2) needs to set strip_newline to true since the
kernel appends a newline character to the end of the AppArmor contexts
read from /proc/>PID>/attr/current.
aa_splitcon(3) also sets strip_newline to true since it is unknown
whether the context is originated from a location that appends a newline
or not.
aa_getpeercon_raw(2) does not set strip_newline to true since it is
unexpected for the kernel to append a newline to the the buffer returned
from getsockopt(2).
This patch also creates tests specifically for splitcon() and updates
the aa_splitcon(3) man page.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Test confinement context splitting, using aa_splitcon(3), with and
without a valid mode pointer.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Create a new libapparmor public function that allows external code to
split an AppArmor confinement context.
This is immediately useful for code that retrieves a D-Bus peer's
AppArmor confinement context using the
org.freedesktop.DBus.GetConnectionCredentials bus method.
https://launchpad.net/bugs/1430532
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
The parse_confinement_mode() function returned NULL when a confinement
mode was not present (unconfined) and when it could not properly parse
the confinement context. The two situations should be differentiated
since the latter should be treated as an error.
This patch reworks parse_confinement_mode() to split a confinement
context and, optionally, assign the mode string. If a parsing error is
encountered, NULL is returned to indicate error.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Use the passed in confinement context string size to improve the
comparison by only doing the string comparison if the size matches and
removing the possibility of reading past the end of the buffer.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
When passing the size of the confinement context to
parse_confinement_mode(), don't include the NUL terminator byte in the
size.
It is confusing to count the NUL terminator as part of the string's
length. This change makes it so that, after a few additional changes,
parse_confinement_mode() can be exposed as part of libapparmor's public
API.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
This patch modifies the socketpair.c test to verify the return value of
aa_getpeercon() based upon the expected label and expected mode lengths.
The test had to be changed slightly so that the returned mode, from
aa_getpeercon(), was preserved. It was being overwritten with the
special NO_MODE value.
This change helps to make sure that future changes to the code behind
aa_getpeercon() does not unintentionally change the function's return
value.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
Samba 4.2 needs some more permissions for nmbd and winbindd.
To avoid overcomplicated profiles, change abstractions/samba to allow
/var/lib/samba/** rwk, (instead of **.tdb rwk) - this change already
fixes the nmbd profile.
winbindd additionally needs some more write permissions in /etc/samba/
(and also in /var/lib/samba/, which is covered by the abstractions/samba
change and also results in some profile cleanup)
References: https://bugzilla.opensuse.org/show_bug.cgi?id=921098 and
https://bugzilla.opensuse.org/show_bug.cgi?id=923201
Acked-by: Seth Arnold <seth.arnold@canonical.com>
I noticed "disconnected path" (run/nscd/*) events for ntpd while
updating to the latest openSUSE Tumbleweed.
Acked-by: Seth Arnold <seth.arnold@canonical.com> for trunk and 2.9.
aa-mergeprof failed to fail ;-) when it should raise an AppArmorException.
Instead, it failed with
AttributeError: 'module' object has no attribute 'AppArmorException'
I confirmed this bug in trunk and 2.9.
Acked-by: Steve Beattie <steve@nxnw.org> for trunk and 2.9.
quote_if_needed() will be used by the upcoming ChangeProfileRule class,
which means it must be moved out of aa.py to avoid an import loop.
rule/__init__.py looks like a better place.
Also re-import quote_if_needed() into aa.py because it's still needed
there by various functions.
Acked-by: Seth Arnold <seth.arnold@canonical.com>
(might get re-used later ;-)
Also add two tests for profile names not starting with / - the quoted
version wasn't catched as invalid before, so this change is actually
also a bugfix.
Acked-by: Steve Beattie <steve@nxnw.org> for trunk and 2.9.
Change aa-notify parse_message() to also honor complain mode log events.
This affects both modes - desktop notifications and the summary report.
Acked-by: Steve Beattie <steve@nxnw.org>
Add setUp() to AATest that sets "self.maxDiff = None" (unlimited).
This gives us unlimited array diffs everywhere where AATest is used.
Also rename several setUp() functions in test-regex_matches.py to
AASetup() to avoid that the shiny new AATest setUp() gets overwritten.
Acked-by: Steve Beattie <steve@nxnw.org>
As requested by Steve, also add an example AASetup() to test-example.py.
This ignores the sniplets generated by profiles/Makefile, but doesn't
ignore local/README because it doesn't have a dot in its name.
Acked-by: John Johansen <john.johansen@canonical.com>
Add several missing network DOMAINs to the apparmor.d manpage.
The list is based on the list that utils/vim/Makefile generates.
Acked-by: John Johansen <john.johansen@canonical.com>
reported by darix on IRC. This is needed if you have a bigger setup with
dovecot on a different (or multiple) machines
Acked-by: Steve Beattie <steve@nxnw.org> for trunk and 2.9
Replace usage of RE_PROFILE_CAP and RE_PROFILE_NETWORK with
CapabilityRule.match() and NetworkRule.match() calls.
This also means aa.py doesn't need to import those regexes anymore.
As a side effect of this change, test-regex_matches.py needs a small
fix because it imported RE_PROFILE_CAP from apparmor.aa instead of
apparmor.regex.
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Add match() and _match() class methods to rule classes:
- _match() returns a regex match object for the given raw_rule
- match() converts the _match() result to True or False
The primary usage is to get an answer to the question "is this raw_rule
your job?". (For a moment, I thought about naming the function
*Rule.myjob() instead of *Rule.match() ;-)
My next patch will change aa.py to use *Rule.match() instead of directly
using RE_*, which will make the import list much shorter and hide
another implementation detail inside the rule classes.
Also change _parse() to use _match() instead of the regex, and add some
tests for match() and _match().
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Change aa.py to use NetworkRule and NetworkRuleset instead of a
sub-hasher to store, check and write network rules. In detail:
- drop profile_known_network() and use is_known_rule() instead
- replace match_net_includes() usage with match_includes() calls
- drop delete_net_duplicates(), use the code in NetworkRule and
NetworkRuleset instead
- make match_net_includes() (still used by aa-mergeprof) a wrapper for
match_includes()
- drop all the network rule parsing from parse_profile_data() and
serialize_profile_from_old_profile() - instead, just call
NetworkRule.parse()
- now that write_net_rules() got fixed, drop it ;-)
- change write_netdomain to use NetworkRuleset
- drop netrules_access_check() - that's is_covered() now
- use 'network' instead of 'netdomain' as storage keyword (log events
still use 'netdomain')
Also update cleanprofile.py to use the NetworkRuleset class.
This also means to delete the (now superfluous) delete_net_duplicates()
function.
Finally, there are some changes in regex.py:
- change RE_PROFILE_NETWORK in regex.py to named matches and to use
RE_COMMA_EOL (not only RE_EOL)
- drop the no longer needed RE_NETWORK_FAMILY and RE_NETWORK_FAMILY_TYPE
(rule/network.py has regexes that check against the list of available
keywords)
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
Add utils/test/test-network.py with tests for NetworkRule and
NetworkRuleset.
The tests are hopefully self-explaining, so let me just mention the most
important things:
- I started to play with namedtuple, which looks very useful (see "exp")
- the test loops make the tests much more readable (compare with
test-capability.py!) and make it easy to add some more tests
- 100% coverage :-)
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
Add utils/apparmor/rule/network.py with the NetworkRule and
NetworkRuleset classes. These classes are meant to handle network rules.
In comparison to the existing code in aa.py, relevant news are:
- the keywords are checked against a list of allowed domains, types and
protocols (these lists are based on what the utils/vim/Makefile
generates - on the long term an autogenerated file with the keywords
for all rule types would be nice ;-)
- there are variables for domain and type_or_protocol instead of
first_param and second_param. (If someone is bored enough to map the
protocol "shortcuts" to their expanded meaning, that shouldn't be too
hard.)
- (obviously) more readable code because we have everything at one place
now
- some bugs are fixed along the way (for example, "network foo," will now
be kept, not "network foo bar," - see my last mail about
write_net_rules() for details)
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
CleanProf.remove_duplicate_rules() didn't call
$profile['capability'].delete_duplicates()
because aa-cleanprof sets same_file=True.
Fix this by calling delete_duplicates(None) so that it
only checks the profile against itsself.
Note: this is only needed if the to-be-cleaned profile doesn't
contain any include rules - with includes present, the
"for inc in includes:" block already called delete_duplicates()
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
Implement in-profile de-duplication in BaseRuleset (currently affects
"only" CapabilityRuleset, but will also work for all future *Ruleset
classes).
Also change 'deleted' to be a simple counter and add some tests that
verify the in-profile deduplication.
Acked-by: Seth Arnold <seth.arnold@canonical.com>
test_parse_modifiers_invalid() uses a hand-broken ;-) regex to parse
only the allow/deny/audit keywords. This test applies to all rule types
and doesn't contain anything specific to capability or other rules,
therefore it should live in test-baserule.py
Moving that test also means to move the imports for parse_modifiers and
re around (nothing else in test-capability.py needs them).
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
Add some tests for the Baserule class to cover the 3 functions that must
be re-implemented in each rule class. This means we finally get 100%
test coverage for apparmor/rule/__init__.py ;-)
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
Ensure nosetests sees all tests in the tests[] tuples. This requires
some name changes because nosetests thinks all function names containing
"test" are tests. (A "not a test" docorator would be an alternative, but
that would require some try/except magic to avoid a dependency on nose.)
To avoid nosetests thinks the functions are a test,
- rename setup_all_tests() to setup_all_loops()
- rename regex_test() to _regex_test() (in test-regex_matches.py)
Also add the module_name as parameter to setup_all_loops and always run
it (not only if __name__ == '__main__').
Known issue: nosetests errors out with
ValueError: no such test method in <class ...>: stub_test
when trying to run a single test generated out of tests[].
(debugging hint: stub_test is the name used in setup_test_loop().)
But that's still an improvement over not seeing those tests at all ;-)
Acked-by: Steve Beattie <steve@nxnw.org> for trunk and 2.9.
Assume you have a profile like
/bin/foo {
/etc/ r,
network,
/usr/ r,
}
(important: there must be be a non-path rule between the two path blocks)
Then run aa-logprof and add another path event. When choosing (V)iew changes,
it will crash with a misleading
File ".../utils/apparmor/aamode.py", line 205, in split_mode
other = mode - user
TypeError: unsupported operand type(s) for -: 'collections.defaultdict' and 'set'
The reason for this is our beloved hasher, which is playing funny games
another time.
The patch wraps the hasher usage with a check for the parent element to
avoid auto-creation of empty childs, which then lead to the above crash.
BTW: This is another issue uncovered by the LibreOffice profile ;-)
Acked-by: Steve Beattie <steve@nxnw.org> for trunk and 2.9.
Update the postfix-common abstraction to cope with signal and unix
socket mediation, update the access to the sasl library locations
in a multiarch compliant way, and allow access to limited bits
of the filesystem paths under which postfix chroots itself to
(/var/spool/postfix/ on Ubuntu).
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: Tyler Hicks <tyhicks@canonical.com>
serialize_profile_from_old_profiles() calls store_list_var() with an
empty hasher. This fails for "+=" because in this case store_list_var()
expects a non-empty hasher with the variable already defined, and raises
an exception because of the empty hasher.
This patch sets "correct = False" if a "+=" operation appears, which
means the variable will be written in "clean" mode instead.
Adding proper support for "add to variable" needs big changes (like
storing a variable's "history" - where it was initially defined and what
got added where).
Acked-by: Steve Beattie <steve@nxnw.org> for trunk and 2.9.
the LibreOffice profile uncovered that handling of @{var} += is broken:
File ".../utils/apparmor/aa.py", line 3272, in store_list_var
var[list_var] = set(var[list_var] + vlist)
TypeError: unsupported operand type(s) for +: 'set' and 'list'
This patch fixes it:
- change separate_vars() to use and return a set instead of a list
(FYI: separate_vars() is only called by store_list_var())
- adoptstore_list_var() to expect a set
- remove some old comments in these functions
- explain the less-intuitive parameters of store_list_var()
Also add some tests for separate_vars() and store_list_var().
The tests were developed based on the old code, but not all of them
succeed with the old code.
As usual, the tests uncovered some interesting[tm] behaviour in
separate_vars() (see the XXX comments and tell me what the really
expected behaviour is ;-)
Acked-by: Steve Beattie <steve@nxnw.org> for trunk and 2.9
Move the code that does the c -> a and d -> w replacement in denied_mask
and requested_mask so that it only runs for path and exec events, but not
for other events (like dbus and ptrace). The validate_log_mode() and
log_str_to_mode() calls are also moved.
Technically, this means moving code from parse_event() to the path
and exec sections in add_event_to_tree().
This also means aa-logprof no longer crashes if it hits a ptrace or
dbus event in the log.
The "if dmask:" and "if rmask:" checks are removed - if a path event
doesn't have these two, it is totally broken and worth a aa-logprof
crash ;-)
Also adjust the parse_event() tests to expect the "raw" mask instead of
a set.
This patch fixes
https://bugs.launchpad.net/apparmor/+bug/1426651 and
https://bugs.launchpad.net/apparmor/+bug/1243932
I manually tested that
- c and d log events are still converted to a and w
- aa-logprof handles exec events correctly
- ptrace events no longer crash aa-logprof
Note: add_event_to_tree() is not covered by tests.
Acked-by: Steve Beattie <steve@nxnw.org> for trunk and 2.9
"capability foo".is_covered("deny capability foo") should return False
even if check_allow_deny is False.
Also add some tests with check_allow_deny=False.
Acked-by: Steve Beattie <steve@nxnw.org>
Also add libraries/libapparmor/swig/perl/Makefile.perle (noticed and
proposed by Steve)
With these changes, "bzr status" is clean again after "make distclean"
Acked-by: Steve Beattie <steve@nxnw.org>.
Acked-by: Tyler Hicks <tyhicks@canonical.com>
Thanks to the used data structure, write_net_rules() replaces bare
'network,' rules with the invalid 'network all,' when saving a profile.
This patch makes sure a correct 'network,' rule is written.
Also reset 'audit' to avoid all (remaining) rules get the audit flag
after writing an audit network rule.
Note: The first section of the function (that claims to be responsible
for bare 'network,' rules) is probably never hit - but I'm not too keen
to remove it and try it out ;-)
Acked-by: Steve Beattie <steve@nxnw.org> for trunk and 2.9.
When parsing a profile with named exec rules, the exec target included
the arrow. This resulted in two arrows when writing the profile (and one
more each time the profile was updated).
Fix this by using the match group that only contains the exec target
without the arrow in parse_profile_data() and
serialize_profile_from_old_profile().
References: https://bugs.launchpad.net/apparmor/+bug/1437901
Acked-by: Steve Beattie <steve@nxnw.org> for trunk and 2.9.
When evince opens a dvi file, it updates the user fonts using
texlive commands in /usr/share/texlive/texmf-dist/web2c/ (or possibly
/usr/share/texlive/texmf/web2c/ in older releases). This patch adjusts
the sanitized_helper profile to allow these tools to run.
Bug: https://bugs.launchpad.net/ubuntu/+source/apparmor/+bug/1010909
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-By: Jamie Strandboge <jamie@canonical.com>
write_net_rules() doesn't add a space after 'audit' in two of three
cases, leading to invalid network rules.
This patch adds the missing spaces.
Acked-by: Seth Arnold <seth.arnold@canonical.com>
for both trunk and 2.9
write_net_rules() creates invalid rules for network rules with one
parameter (for example "network bluetooth").
Add a trailing comma to create valid rules.
Acked-by: Seth Arnold <seth.arnold@canonical.com>
for both trunk and 2.9.
Change serialize_parse_profile_start() to use parse_profile_start()
instead of using duplicated code.
The behaviour is mostly kept, with the exception that the function is
more strict now and raises exceptions instead of ignoring errors.
In practise, this won't change anything because the profiles are parsed
with parse_profile() (which calls parse_profile_start()) - and that
already errors out.
The tests are updated to match the more strict behaviour.
The next step would be to drop serialize_parse_profile_start()
completely, but this isn't urgent and can/should be done when we have
test coverage for serialize_profile_from_old_profile() one day ;-)
Acked-by: Steve Beattie <steve@nxnw.org>
Fix is_skippable_dir() - the regex also matched things like
/etc/apparmor.d/dont_disable, while it should match on the full
directory name.
Also add some tests based on a real-world aa-logprof run (with "print (path)"
in is_skippable_dir()) and some additional "funny"[tm] dirs.
Needless to say that the tests
('dont_disable', False),
('/etc/apparmor.d/cache_foo', False),
will fail with the old is_skippable_dir().
Acked-by: Steve Beattie <steve@nxnw.org>
Replace RE_PROFILE_START with RE_PROFILE_START_2 and adjust all
code sections that used RE_PROFILE_START_2.
The only real change is that test_get_flags_invalid_01 and
test_get_flags_invalid_02 now expect AppArmorException instead of
AppArmorBug.
Acked-by: Steve Beattie <steve@nxnw.org> for trunk
This patch implements attachment handling - aa-logprof now works with
profiles that have an attachment defined, instead of ignoring audit.log
entries for those profiles.
Changes:
- parse_profile_start_line(): remove workaround that merged the
attachment into the profile name
- parse_profile_data(): store attachment when parsing a profile
- update test_parse_profile_start_03, test_serialize_parse_profile_start_03,
test_set_flags_nochange_09 and some parse_profile_start_line() tests -
they now expect correct attachment handling
Acked-by: Steve Beattie <steve@nxnw.org>
this patch makes set_profile_flags more strict:
- raise AppArmorBug if newflags contains only whitespace
- raise AppArmorBug if the file doesn't contain the specified profile or
no profile at all
The tests are adjusted to expect AppArmorBug instead of a silent
failure. Also, some tests are added for profile=None, which means to
change the flags for all profiles in a file.
- test_set_flags_08 is now test_set_flags_invalid_04
- test_set_flags_invalid_03 is changed to only contain one reason for a
failure, not two ;-)
Acked-by: Steve Beattie <steve@nxnw.org>
Changes in set_profile_flags():
- rewrite set_profile_flags to use parse_profile_start_line() and
write_header().
- replace the silent failure for non-existing files with a proper
exception (using lazy programming - the check is done by removing the
"if os.path.isfile()" check, open_file_read then raises the
exception ;-)
- comment out regex_hat_flag and the code that was supposed to handle
hat flags, which were totally broken. We'll need another patch to fix
it, and we also need to decide if we want to do that because it
introduces a behaviour change (currently, aa-complain etc. don't
change hat flags).
The tests for set_profile_flags() are also updated:
- prepend a space to comments because write_header always adds a space
between '{' and the comment
- remove a test with superfluous quotes that are no longer kept (that's
just a profile cleanup, so dropping that test is the easiest way)
- update test_set_flags_10 and test_set_flags_12 to use the correct
profile name
- enable the tests for invalid (empty) flags
- update the test for a non-existing file
Note: test_set_flags_10, test_set_flags_12 and test_set_flags_nochange_09
will fail with this patch applied. The next patch will fix that.
Acked-by: Steve Beattie <steve@nxnw.org>
The Makefiles don't create/need the 'common' symlinks since some time,
which also means we no longer need to have them in .bzrignore.
Acked-by: Steve Beattie <steve@nxnw.org>
if 3/2 == 1:
print("python2 inside")
Add "from __future__ import division" so that python2 returns the
correct result (if needed, as float)
On related news: At least python3 knows how to calculate correctly.
Acked-by: Steve Beattie <steve@nxnw.org> for trunk and 2.9
aa-logprof doesn't ask anything for
type=AVC msg=audit(1427633461.202:281): apparmor="DENIED" operation="chmod" profile="/usr/lib64/firefox/plugin-container" name="/home/cb/.config/ibus/bus/" pid=7779 comm="plugin-containe" requested_mask="w" denied_mask="w" fsuid=1000 ouid=1000
This patch fixes this by adding 'chmod' to the list of file operation
types in logparser.py.
Acked-by: Seth Arnold <seth.arnold@canonical.com>
for both trunk and 2.9.
Rewrite parse_profile_start() in aa.py to a more readable version.
The behaviour remains unchanged (and is covered by tests).
The patch also updates the comment about the internal struct of
aa[profile][hat] - initial_comment was missing.
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
Change the write_header tests so that the 'profile_keyword' and
'header_comment' parameters can be (and are) tested:
- add a None for both to the existing tests
- add some tests that come with the profile keyword and/or a comment
Acked-by: Steve Beattie <steve@nxnw.org>
- add support for prof_data['header_comment'] (comment after '{')
and prof_data['profile_keyword'] (to force the 'profile' keyword, even
if it isn't needed) to write_header().
(set_profile_flags() will be the only user of these two for now)
- fix a crash if depth is not an integer - for example,
len(' ')/2 # 3 spaces = 1.5
would cause a crash.
Also add a test for 1.5 and 1.3 spaces.
- rewrite the handling of flags to avoid we have to maintain two
different template lines.
- update the tests to set 'profile_keyword' and 'header_comment' to None.
This avoids big changes in the test code. I'll send another patch that
makes sure profile_keyword and header_comment are tested ;-)
Acked-by: Steve Beattie <steve@nxnw.org>
Add the attachment to the parse_profile_start() and
serialize_parse_profile_start() return values, and adjust the functions
calling the *parse_profile_start() functions to save the attachment in
the "attachment" variable (which isn't used yet).
Also adjust the tests for the added return value.
(Sorry for not getting the resultset right from the beginning!)
Acked-by: Steve Beattie <steve@nxnw.org>
Also fix a little bug that added the profile keyword if the path needed
quotes (profile "/foo bar" - but "/foo bar" is enough). This was caused
by a regex that always matched on quoted paths (hint: "/ matches
^[^/] ;-)
Also add some tests with attachments and update the test for the bugfix
mentioned above.
Now the remaining part is to make sure that prof_data['attachment'] gets
set when parsing the profiles :-)
Acked-by: Steve Beattie <steve@nxnw.org>
Also add loop support to test-aa.py.
BTW: In case you wonder - the need to replace unittest.TestCase with
AATest is intentional. It might look annoying, but it makes sure that
a test-*.py file doesn't contain a test class where tests = [...] is
ignored because it's still unittest.TestCase.
(Technically, setup_all_tests() will error out if a test class doesn't
contain tests = [...] - either explicit or via its parent AATest.)
Acked-by: Steve Beattie <steve@nxnw.org>
Add various tests for set_profile_flags, and document various
interesting[tm] things I discovered while writing the tests (see
the inline comments for details).
Also adds a read_file() function to common_test.py.
The most interesting[tm] thing I found is:
regex_hat_flag = re.compile('^([a-z]*)\s+([A-Z]*)\s*(#.*)?$')
which matches various unexpected things - but not a hat :-/
(see mailinglist for all funny details)
Acked-by: Steve Beattie <steve@nxnw.org>
Convert serialize_parse_profile_start() to use
parse_profile_start_line(), and adjust a test to expect an AppArmorBug
instead of an AttributeError exception.
Also add two tests (they succeed with the old and the new code).
Note that these tests document interesting[tm] behaviour - I tend to
think that those cases should raise an exception, but I'm not sure about
this because serialize_profile_from_old_profile() is a good example for
interesting[tm] code :-/
I couldn't come up with a real-world test profile that would hit those
cases without erroring out aa-logprof earlier - maybe the (more
sane-looking) parse_profiles() / serialize_parse_profile_start()
protects us from hitting this interesting[tm] behaviour.
Acked-by: Steve Beattie <steve@nxnw.org>
The commit message for r2976 says:
[...]
The patch also adds test-example.py, which is
- a demo of the code added to common_test.py
- a template file that we can copy for future test-*.py
Acked-by: Steve Beattie <steve@nxnw.org>
but I forgot to add test-example.py to bzr, which I hereby do.
The previous patch slightly changed the behaviour of parse_profile_start()
and get_profile_flags() - they raise AppArmorBug instead of
AppArmorException when specifying a line that is not the start of a
profile and therefore doesn't match RE_PROFILE_START_2.
This patch updates test-aa.py to expect the correct exceptions, and adds
another test with quoted profile name to ensure that stripping the
quotes works as expected.
Acked-by: Steve Beattie <steve@nxnw.org>
Add the parse_profile_start_line() function to regex.py, which is a
wrapper for RE_PROFILE_START_2 and returns an array with named matches.
Also change some places in aa.py from using RE_PROFILE_START to the
parse_profile_start_line() function.
Notes:
- until everything is migrated to the new function, I'll keep the old
RE_PROFILE_START unchanged - that's the reason to add the new regex
as RE_PROFILE_START_2
- the patch changes only aa.py sections that are covered by tests already
(which means some users of RE_PROFILE_START are remaining)
- parse_profile_start_line() merges 'profile' and 'attachment' into
'profile' (aka the old, broken behaviour) until aa.py can handle the
attachment properly. The alternative would be to ignore 'attachment',
which would be worse.
Acked-by: Steve Beattie <steve@nxnw.org>
Add better support for looping over a tests[] array to common_test.py:
- class AATest - a base class we can use for all tests, and that will
probably get more features in the future (for example tempdir
handling)
- setup_all_tests() - a function that iterates over all classes in the
given file and calls setup_test_loops() for each of them
- setup_tests_loop() - a function that creates tests based on tests[]
in the given class. Those tests call the class' _run_test() method for
each test specified in tests[] (inspired by setup_regex_tests() ;-)
This means we can get rid of the manually maintained tests list in
test-regex_matches.py and just need to call setup_all_tests() once in
each file.
The patch also adds test-example.py, which is
- a demo of the code added to common_test.py
- a template file that we can copy for future test-*.py
Acked-by: Steve Beattie <steve@nxnw.org>
The following patch addresses two issues on older releases:
1) In trunk commit 2911, the line 'undefine VERBOSE' was added to
parser/tst/Makefile so that the equality tests would not generate
verbose output when $VERBOSE != 1. Unfortunately, the 'undefine'
keyword was not introduced in GNU Make until version 3.82. On
distro releases like Ubuntu 12.04 LTS that include versions of Make
older than that, make check and make clean abort when VERBOSE is
not set to 1. The patch fixes that by setting VERBOSE to a zero
length string if does not already equal 1.
2) In trunk commit 2923, a workaround for systemd as init was added
to the pivot_root regression test. The workaround included a
call to ps(1) to determine if systemd is pid 1. Unfortunately,
in older versions of the procps package (such as the version in
Ubuntu 12.04 LTS), 'ps -hp1' emits the warning
Warning: bad ps syntax, perhaps a bogus '-'? See http://procps.sf.net/faq.html
The patch below converts the ps call to 'ps hp1' which does not
generate the warning.
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: John Johansen <john.johansen@canonical.com>
Merge from Cameron Norman <camerontnorman@gmail.com> based on a patch
from Christian Boltz <apparmor@cboltz.de>.
This patch allows /var/lib/misc/dnsmasq.*.leases rw and
/{,var/}run/lxc/dnsmasq.pid rw for LXC networking setup.
Acked-by: Steve Beattie <steve@nxnw.org>
The two internal aa_features objects weren't being unreferenced when the
aa_policy_cache object was being freed.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
The aa_features and aa_kernel_interface APIs get a little bit of
testing, as well.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
realloc() returns NULL when it fails. Using the same pointer to specify
the buffer to reallocate *and* to store realloc()'s return value will
result in a leak of the previously allocated buffer upon error.
These issues were discovered by cppcheck.
Note that 'buffer' in write_policy_fd_to_iface() has the autofree
attribute so it must not be manually freed if the realloc(3) fails as
it'll be automatically freed.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
The libapparmor library is built with gcc, while the parser is built
with g++. The parser code needs to cast pointers returned from the
malloc(3) family of calls. However, code removed from the parser to
libapparmor can drop the casts.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
Creates a libapparmor function, _aa_asprintf(), which sets the *strp to
NULL on error. This is needed for all of the users of the _aa_autofree
cleanup attribute because the value of *strp is undefined when
asprintf() fails and that could result in _aa_autofree() being passed a
pointer value that it should not free.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
The parser no longer has a need for the atomic operations since all
callers have been moved to libapparmor.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
With create_cache() headed for libapparmor, we can't use the show_cache
or write_cache globals.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
As code is moved from the parser to libapparmor, the libapparmor code
base will need to have the "unused" macro defined. This macro will need
to be duplicated in the parser and libapparmor due to it being a
compiler-specific macro that shouldn't be exported from libapparmor.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
The existing kernel_interface.c file collides with the expected file
name of the implementation of the aa_kernel_interface API.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Remove the use of the "_" macro, which translates into gettext(3), from
code that will be used from the parser to libapparmor since libapparmor
will not support gettext(3) for debug messages and syslog messages.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
The parser's copy of the two atomic operations will be removed once the
new API's (aa_features, aa_policy_cache, aa_kernel_interface) are moved
from the parser to libapparmor.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
The function names must be prepended with "_aa_" since they're going to
be exported from libapparmor. The code bases using the _aa_autofree(),
_aa_autoclose(), and _aa_autofclose() will need to internally alias
those functions to the previously used autofree, autoclose, and
autofclose names.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
This patch adds equivalents of the parser's PDEBUG() and PERROR()
functions to libapparmor.
It does not add gettext(3) support to libapparmor since these are
messages that only developers will see (debug builds with
LIBAPPARMOR_DEBUG=1) or messages that go to the syslog.
PDEBUG() does nothing unless libapparmor is built with --enable-debug.
It prints to stderr if libapparmor is built with --enable-debug and the
LIBAPPARMOR_DEBUG environment variable is set.
PERROR() uses syslog(LOG_ERR, ...) by default. The message is sent to
the syslog and to stderr if libapparmor is built with --enable-debug and
the LIBAPPARMOR_DEBUG environment variable is set.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
This is needed for some of the parser functionality that will be moved
to libapparmor. In the short term, only the 'bool' type is needed but it
makes sense to simply require a C99 compliant compiler for libapparmor
since the parser is being rewritten in C++. The use of C99 will reduce
future headaches when moving code between the two code bases.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
This patch creates a private API in libapparmor in which upstream
provides no guarantees in regards to ABI stability.
A new header file, <sys/apparmor_private.h>, is created. The "_aa"
prefix will be used for symbols belonging to the private API.
To kick things off, a library friendly version of is_blacklisted() is
moved into libapparmor.
The purpose of a private libapparmor API is to prevent duplicated code
between the parser and libapparmor. This becomes an issue as we prepare
to move chunks of the parser into libapparmor.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
This may be useful for something like an init daemon that simply wants
to load all cached binaries without worrying about any sort of policy
compilation.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
Create new, ref, and unref functions for aa_kernel_interface. The "new"
function allows for the caller to pass in an aa_features object that is
then used to check if the kernel supports set load operations.
Additionally, the "new" function allows for the apparmorfs path to be
discovered once instead of during every policy load.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
__sd_serialize_profile() had a duplicated implementation for writing to
apparmorfs interface files after a profile compilation. This patch
migrates it to the new aa_kernel_interface API.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
This is the start of the kernel_interface API that allows callers to
specify a buffer, a file path, or a file descriptor that should be
copied to the proper kernel interface for loading, replacing, or
removing in-kernel policies.
Support exists for reading from a file path or file descriptor into a
buffer and then writing that buffer to the appropriate apparmorfs
interface file.
An aa_kernel_interface_write_policy() function is also provided for
callers that want to route a buffer to an arbitrary file descriptor
instead of to an apparmorfs file. This is useful when an admin instructs
apparmor_parser to write to stdout or a file.
Additionally, it removes some parser-specific globals from the
kernel_interface.c file, such as OPTION_{ADD,REPLACE,REMOVE}, in
preparation for moving the code into a library.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
This function allows for a policy cache to be removed without having a
previously instatiated aa_policy_cache object. It simply works off of a
path.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
This API has the same look-and-feel of the previous aa_features API.
The cache setup code was heavily dependent on globals set by CLI
options. Options such as "skip the read cache", or "skip the write
cache", or "don't clear the cache if it isn't valid", won't be useful
for all aa_policy_cache API users so some of that logic was lifted out
of the API. The constructor function still provides a bool parameter
that specifies if the cache should be created or not.
If the policy cache is invalid (currently meaning that the cache
features file doesn't match the kernel features file), then a new
aa_policy_cache object is still created but a call to
aa_policy_cache_is_valid() will return false. The caller can then decide
what to do (create a new valid cache, stop, etc.)
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
This option adds unneeded complexity to the parser CLI and the upcoming
aa_policy_cache API. Get rid of it and simply create the cache dir if
--write-cache is specified.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
This patch removes the final dependency on callers needing access to the
features string so aa_features_get_string() can go away.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
Defines a function that can be called to test features support. It is
string based which allows the support tests to work with new kernel
features without any changes.
The use of global variables in the parser to store and check features
support is still preserved. The parser should probably move over to
passing the aa_features object around but that's left for later.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
This is a simple aa_features equality test. Placing it behind a function
call allows us to do something more complex than a simple string
comparison later.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
The aa_features_new_*() functions create an aa_features object. They can
be thought of as the constructor of aa_features objects. A number of
constructors are available depending on whether the features are coming
from a file in the policy cache, a string specified on the command line,
or from apparmorfs.
The aa_features_ref() and aa_features_unref() functions are used to grab
and give up references to an aa_features. When the ref count hits zero,
all allocated memory is freed. Like with free(), aa_features_unref() can
be called with a NULL pointer for convenience.
Pre-processor macros are hidden behind functions so that they don't
become part of our ABI when we move this code into libapparmor later on.
A temporary convenience function, aa_features_get_string(), is provided
while code that uses aa_features is migrated from expecting raw features
string access to something more abstract. The function will be removed
in an upcoming patch.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
snprintf_buffer() needed to be modified in order to properly return error
conditions up the stack, instead of exiting, but there were some other
cleanups that it could use.
It was obviously implemented with the features_struct in mind so this
patch simplifies the input parameters by directly accepting a
features_struct pointer. Also, the name is changed to reflect that it is
intended to work on a features_struct instead of an arbritrary buffer.
A quick sanity check is added to make sure that the features_struct.pos
value isn't pointing past the end of the buffer.
The printf(3) family of functions can return a negative value upon error
so a check of the return value of vsnprintf(3) is added.
Finally, the return values of the function are simplified to 0 on
success or -1, with errno set, on error. This is possible since
features_struct.pos can be internally updated after a successful
vsnprintf(3).
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
These operations will be used for grabbing and releasing references to
objects. They leverage the GCC builtins for atomic operations.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
Require the caller of setup_cache() to pass in a valid cache location
string. This removes the use of the basedir global from the
policy_cache.c file.
Additionally, it is no longer necessary to return the "cache dir" path
from setup_cache() since it will always be the same as the input path.
The return value is changed to an int so an error code can be returned
instead of using exit().
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
Modify setup_cache() to accept the user-supplied cacheloc and return the
validated or created cache directory. The caller must then track that
variable and pass it into any parser/policy_cache.c functions that need
it.
The main reason for this change is that the cache location and the cache
directory will soon be two different paths. The cache location will
typically be the parent of the cache directory.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
This patch moves the logic that sets up the policy into a new function
in policy_cache.c
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
Signed-off-by: John Johansen <john.johansen@canonical.com>
[tyhicks: Fixed build failures]
[tyhicks: Fixed bug where a warning was being printed when it shouldn't]
[tyhicks: Forward ported to trunk]
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Signed-off-by: John Johansen <john.johansen@canonical.com>
[tyhicks: Forward ported patch to trunk]
[tyhicks: remove commented out code]
[tyhicks: fix use after free]
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Signed-off-by: John Johansen <john.johansen@canonical.com>
[tyhicks: Handle inverted return from find_subdomainfs_mountpoint()]
[tyhicks: Link test progs to libapparmor to fix make check build fail]
[tyhicks: Migrate from opendir() to open() for opening apparmorfs]
[tyhicks: Make some of the split out functions static]
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
While some of these allocations will go away as we convert to C++,
some of these need to stay C as the are going to be moved into a
library to support loading cache from init daemons etc.
For the bits that will eventually be C++ this helps clean things up,
in the interim.
TODO: apply to libapparmor as well
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Tyler Hicks <tyhicks@canonical.com>
While some of these allocations will go away as we convert to C++,
some of these need to stay C as the are going to be moved into a
library to support loading cache from init daemons etc.
For the bits that will eventually be C++ this helps clean things up,
in the interim.
TODO: apply to libapparmor as well
Signed-off-by: John Johansen <john.johansen@canonical.com>
Currently the cache tracks the most recent timestamp of parsed files
and then compares that to the cache timestamp. This unfortunately
prevents the parser from being able to know which files caused the
cache check failure.
Rework the cache check so that there is a debug option, and that
the cache file timestamp is set first so that we can output
a deug message for each file that causes a cache check failure.
Signed-off-by: John Johansen <john.johansen@canonical.com>
[tyhicks: Forward ported to trunk and minor cleanups]
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
Update the file rule pattern to show it is possible to specify a bare
file rule. Eg.
file,
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
Refactor FILEGLOB so that it means both quoted and unquoted file globs.
Also
FILEGLOB was uncorrectly referenced in a few places where it should have
allowed for quoting.
There were also a few places that provided a parameter description with
FILEGLOB without defining that that is full equivalent to FILEGLOB.
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
Add miss ix and ux fallback permission modes, named profile transitions.
Also fix the file access modes and rule pattern to properly reflect
what is allowed.
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
Consolidate and update the qualifier information in the man page.
Most of the rule qualifiers where duplicated instead of being pulled
into a common section.
Also the rule qualifiers where missing the 'allow' qualifier.
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Christian Boltz <apparmor@cboltz.de>
- verify audit and audit allow is equal
- verify audit differs from deny and audit deny
- verify deny differs from audit deny
- make the verbose text a little more useful for some cases
- correct overlap exec tests to substitute in looped perms
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
- make the verbose output of equality.sh honor whether or not
the environment variable VERBOSE is set
- thereby making the output verbose when 'make check V=1' or 'make
check VERBOSE=1' is given from within the parser/ directory. This
will make distribution packagers happy when diagnosing build
failures caused by test failures.
- if verbose output is not emitted and the tests were successful, emit
a newline before printing PASS.
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: John Johansen <john.johansen@canonical.com>
This adds several new equality tests and turned up a couple of more
bugs
https://launchpad.net/bugs/1433829https://launchpad.net/bugs/1434018
- add link/link subset tests
- add pix, Pix, cix, Cix, pux, Pux, cux, Cux and specified profile
transitions (/f px -> b ...)
- test equality of leading and trailing permission file rules
ie. /foo rw, == rw /foo,
- test that specific x match overrides generic x rule. ie.
/** ix, /foo px, is different than /** ix, /foo ix,
- test that deny removes permission
/f[abc] r, deny /fb r, is differnt than /f[abc] r,
In addition to adding the new tests, it changes the output of the
equality tests, so that if the $verbose variable is not set successful
tests only output a period, with failed tests outputing the full
info. If verbose is set the full test info is output as before.
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
BugLink: http://bugs.launchpad.net/bugs/1433829
The apparmor_parser fails to compile deny rules with only link
permissions.
Eg.
deny /f l,
deny l /f,
deny link /f -> /d,
Will all fail to compile with the following assert
apparmor_parser: aare_rules.cc:99: Node* convert_file_perms(int, uint32_t, uint32_t, bool): Assertion `perms != 0' failed.
NOTE: this is a minimal patch a bigger patch that cleans-up and separates
and reorganizes file, link, exec, and change_profile rules is needed
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
Darix' guess is that this is needed by libpq because he uses a postgresql
database with dovecot and has ssl enabled in postgresql.
Acked-by: Seth Arnold <seth.arnold@canonical.com> for trunk and 2.9
This patch fixes the equality test script and the valgrind wrapper
script to make the parser under test use the features.all features file
from the features_files/ subdirectory. Otherwise, the equality tests
will fail on systems where the not all of the current language features
are supported. The equality fix does so in a way to make the script work
correctly regardless of the directory it is run from.
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: John Johansen <john.johansen@canonical.com>
The fix to prevent the compiler from SEGV'ing when dumping network
rules in commit 2888 introduced the following compiler warning:
network.c: In function ‘const char* net_find_af_name(unsigned int)’:
network.c:331:16: warning: comparison between signed and unsigned integer expressions [-Wsign-compare]
for (i = 0; i < sizeof(network_mappings) / sizeof(*network_mappings); i++) {
The problem is that the counter i is an int, but sizeof returns size_t
which is unsigned. The following patch fixes the issue by converting the
type of i to size_t.
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: John Johansen <john.johansen@canonical.com>
Previously, we only had the ability to test that binary policy files
were equal. This patch allows for the testing of binary policy files
that are not equal.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
This fixes the incorrect compilation of audit modifiers for exec and
pivot_root as detailed in
https://launchpad.net/bugs/1431717https://launchpad.net/bugs/1432045
The permission accumulation routine on the backend was incorrectly setting
the audit mask based off of the exec type bits (info about the exec) and
not the actual exec permission.
This bug could have also caused permissions issues around overlapping exec
generic and exact match exec rules, except the encoding of EXEC_MODIFIERS
ensured that the
exact_match_allow & AA_USER/OTHER_EXEC_TYPE
test would never fail for a permission accumulation with the exec permission
set.
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
serialize_profile_from_old_profile() in aa.py, as a preparation to add
tests and then switch to the upcoming RE_PROFILE_START wrapper function.
Besides moving the code, I replaced write_prof_data[profile][hat]['profile']
and write_prof_data[profile][hat]['external'] with function parameters
to avoid that I have to pass around the full write_prof_data.
Note: The "lineno" parameter is technically superfluous - I kept it to
have the parameters as close to parse_profile_start() as possible and
hope that I can merge those functions later (when we have test coverage).
Acked-by: Steve Beattie <steve@nxnw.org> for trunk and 2.9.
The parser.conf example statement for Include statements used
/etc/apparmor.d/abstractions which is unlikely to make anyone enabling
it happy as our shipped and example policies all include the
'abstractions/' directory in the relative paths. This patch adjusts the
example and provides a second example, based on an enabled entry as
shipped in Ubuntu.
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: Christian Boltz <apparmor@cboltz.de>
and change the code to use them.
Also add a comment to act() that it's only used by aa-cleanprof.
Note: The new functions add the --base parameter to the apparmor_parser
calls, which also means the disable directory inside the given profile
dir (and not always /etc/apparmor.d/disable) is now honored.
Acked-by: Steve Beattie <steve@nxnw.org> for trunk and 2.9.
logparser.py / add_event_to_tree() has 5 places to handle 'path' events.
This patch merges most if conditions to reduce that to 2 places.
It also makes the matching a bit more strict - instead of using 'in',
'xattr' has to be an exact match and 'file_' is matched with startswith().
Also, 'getattr' is added to the list of file events.
Acked-by: Steve Beattie <steve@nxnw.org>
---------- trunk only, unclear for 2.9 --------------
Without it, aa-disable
- didn't error out when hitting a broken profile directory
- didn't find a profile if it doesn't use the default naming scheme
(for example /bin/true profile hiding in bin.false)
Acked-by: Steve Beattie <steve@nxnw.org> for trunk and 2.9
As mir has come into use in Ubuntu touch and is available for testing on
Ubuntu desktop, confined apps need access to a few mir specific things.
This patch adds a mir abstraction.
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: Tyler Hicks <tyhicks@canonical.com>
aa-status was crashing when parsing through /proc/mounts looking to see
if and where the securityfs synthetic file system is mounted if there
was a mount point that contained characters outside of the charset in
use in the environment of aa-status. This patch fixes the issue by
converting the read of /proc/mounts into a binary read and then uses
decode on the elements.
Patch by Alain BENEDETTI.
Acked-by: Steve Beattie <steve@nxnw.org>
The error path was being taken when openat() return 0 but openat()
returns -1 on error.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
As a follow-up to the logparser.py change that converts disconnected
path events to an error, add a testcase to test-logparser.py.
Acked-by: Steve Beattie <steve@nxnw.org> for both trunk and 2.9.
Parts of the regression tests that use the do_open() inline function
from changehat.h fail to build under gcc-5 like so:
cc -g -O0 -Wall -Wstrict-prototypes changeprofile.c -lapparmor -o changeprofile
/tmp/ccT6GE6k.o: In function `main':
/home/ubuntu/bzr/apparmor/tests/regression/apparmor/changeprofile.c:43: undefined reference to `do_open'
collect2: error: ld returned 1 exit status
<builtin>: recipe for target 'changeprofile' failed
This patch converts the do_open function declaration to be static
inline, which apparently keeps gcc-5 from getting confused.
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
The upcoming function parse_profile_start() (which is a wrapper around
the updated RE_PROFILE_START, and will live in regex.py) needs
strip_profile(), but importing it from aa.py fails with an import loop.
Therefore this patch moves strip_quotes() from aa.py to regex.py and
re-imports it into aa.py.
As a bonus, the patch also adds some tests for strip_quotes() ;-)
Also add TestStripQuotes to the test_suite list because it won't run
otherwise.
Acked-by: Steve Beattie <steve@nxnw.org> for both trunk and 2.9
Move the code for parsing the profile start ("/foo {") from aa.py
parse_profile_data() to a separate function parse_profile_start().
Most of the changes are just moving around code, with some small
exceptions:
- instead of handing over profile_data to parse_profile_start() to
modify it, it sets two variables (pps_set_profile and
pps_set_hat_external) as part of its return value, which are then
used in parse_profile_data() to set the flags in profile_data.
- existing_profiles[profile] = file is executed later, which means
it used the strip_quotes() version of profile now
- whitespace / tab level changes
The patch also adds some tests for the parse_profile_start() function.
Acked-by: Steve Beattie <steve@nxnw.org>
flags_bad5.sd contains tests to ensure the debug flag is no longer
accepted.
However, the file contains multiple expected failures, which means that
it will still fail as long as at least one of them fails. This patch
splits each test into its own file to ensure each of them fails.
Acked-by: Steve Beattie <steve@nxnw.org>
Also adds a check to get_profile_flags() to catch an invalid syntax:
/foo ( ) {
was accepted by get_profile_flags, while
/foo () {
failed.
When testing with the parser, both result in a syntax error, therefore
the patch makes sure it also fails in get_profile_flags().
Acked-by: Steve Beattie <steve@nxnw.org>
Seth pointed out that dirat_for_each() didn't correctly handle the
return value from readdir_r(). On error, it directly returns a positive
errno value. This would have resulted in that positive errno value being
returned, with an undefined errno value set, from dirat_for_each().
However, the dirat_for_each() documentation states that -1 is returned,
with errno set, on error.
This patch results in readdir_r()'s return value being handled
appropriately. In addition, it ensures that 0 is always returned on
success.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Reported-by: Seth Arnold <seth.arnold@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
The smbd profile contains /{,var/}run/cups/cups.sock rw, which is
covered by abstractions/cups-client and therefore superfluous.
Acked-by: Steve Beattie <steve@nxnw.org>
This means that aa-logprof will ignore the event instead of crashing with
AppArmorException: 'Unexpected rank input: var/run/nscd/passwd'
Note that I made the check as specific as possible to be sure it doesn't
hide other events.
References: https://bugzilla.opensuse.org/show_bug.cgi?id=918787
Acked-by: Steve Beattie <steve@nxnw.org>
Also update test-capability.py - it contains a test that needs
'error_code': 0,
added to avoid a failure.
Patch by: Steve Beattie <steve@nxnw.org>
Acked-by: Christian Boltz <apparmor@cboltz.de>
From: Felix Geyer <debfx@ubuntu.com>
At least Debian/Ubuntu started shipping some aspell files in
/usr/share/aspell/.
For example:
/usr/share/aspell/iso-8859-1.cmap
/usr/share/aspell/iso-8859-1.cset
The abstraction should allow read access to these files.
Acked-by: Steve Beattie <steve@nxnw.org>
Remove the check if the disable directory exists. If it's really
missing, it will be auto-created by create_symlink(), so we
automagically fix things instead of annoying the user with an
error message ;-)
Acked-by: Steve Beattie <steve@nxnw.org> for both trunk and 2.9.
Like net_find_af_name, this assumed that AF_* values were consecutive.
[smcv: split out from a larger patch, added commit message,
removed dead declaration]
Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
The network_families array is automatically built from AF_NAMES, which is
extracted from the defines in <bits/socket.h>. The code assumes that
network_families is indexed by the AF defines. However, since the
defines are sparse, and the gaps in the array are not packed with
zeroes, the array is shorter than expected, and the indexing is wrong.
When this function was written, the network families that were
covered might well have been consecutive, but this is no longer true:
there's a gap between AF_LLC (26) and AF_CAN (29). In addition,
the code that parses <sys/socket.h> does not recognise AF_DECnet (12)
due to the lower-case letters, leading to a gap betwen AF_ROSE (11)
and AF_NETBEUI (13).
This assumption caused a crash in our testing while parsing the rule
"network raw".
[smcv: split out from a larger patch, added commit message]
Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Don't pass an ostream reference into another ostream via <<.
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Make sure most tools (for example aa-complain) don't error out if
no logfile can be found. (For obvious reasons, aa-logprof and
aa-genprof will still require a logfile ;-)
This is done by moving code from the global area in aa.py to the new
function set_logfile(), which is called by aa-logprof and aa-genprof.
While on it,
- rename apparmor.filename to apparmor.logfile
- move the error handling for user-specified logfile from aa-genprof
and aa-logprof to aa.py set_logfile()
Note: I'd have prefered to hand over the logfile as parameter to
do_logprof_pass(), but that would break last_audit_entry_time() in
aa-genprof which requires the log filename before do_logprof_pass()
is called.
References: https://bugs.launchpad.net/apparmor/+bug/1423702
Acked-by: Seth Arnold <seth.arnold@canonical.com>
Abstract af_unix socket names can contain a null character, however the
aare to pcre conversion explicitly disallows null characters because they
are not valid characters for pathnames. Fix this so that they type of
globbing is selectable.
this is a partial fix for
Bug: http://bugs.launchpad.net/bugs/1413410
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
The lexer front end currently incorrectly processes the \000 \x00 \d00 escape sequence resulting in a null character being embedded in the processed string, this results in the string not being full processed later.
The aare to pcre regex conversion fn also incorrectly strips out the \00, and any other escape sequence it doesn't know about, resulting in incorrect strings being passed to the backend. Fix this by passing through any valid escape sequence that is not handled by the fn.
this is a partial fix for
Bug: http://bugs.launchpad.net/bugs/1413410
Signed-off-by: John Johansen <john.johansen@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
Adjust the libapparmor function prototypes, variable names, and comments
that incorrectly used the name "con" when referring to the label.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
The correct usage of the terms context and label is not clear in the
aa_getcon(2) man page. The aa_getcon(2) family of functions are also
prototyped incorrectly since the *con parameter represents a label and
not a context.
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
Acked-by: John Johansen <john.johansen@canonical.com>
libapparmor _aa_is_blacklisted() - some extensions were missing in the
python code.
Also make the code more readable and add some testcases.
Notes:
- the original code additionally ignored *.swp. I didn't include that -
*.swp looks like vim swap files which are also dot files
- the python code ignores README files, but the C code doesn't
(do we need to add README in the C code?)
Acked-by: Kshitij Gupta <kgupta8592@gmail.com> for 2.9 and trunk
Acked-by: Steve Beattie <steve@nxnw.org>
string or if a mode_char is not in MODE_HASH.
Also update the testcase for "asdf42" (which raises AppArmorBug now)
and add a test that simulates MODE_HASH and MODE_MAP_SET getting out
of sync (tests the second part of the if condition).
Acked-by: Steve Beattie <steve@nxnw.org>
Since the Makefile cleanup, the _clean target is only used to delete
manpages etc. generated from *.pod files.
This patch renames the _clean target to pod_clean to make it obvious
what it does.
Acked-by: John Johansen <john.johansen@canonical.com>
The lexer front end currently incorrectly processes the \000 \x00 \d00 escape sequence resulting in a null character being embedded in the processed string, this results in the string not being full processed later.
The aare to pcre regex conversion fn also incorrectly strips out the \00, and any other escape sequence it doesn't know about, resulting in incorrect strings being passed to the backend. Fix this by passing through any valid escape sequence that is not handled by the fn.
this is a partial fix for
Bug: http://bugs.launchpad.net/bugs/1413410
Signed-off-by: John Johansen <john.johansen@canonical.com>
Get rid of the relics in libapparmor's Makefile.am for generating
tarballs from svn, which is no longer relevant. Also clean generated
manpages during make clean rather than just make maintainer-clean.
This patch removes a bunch of the per-directory tarball and rpm
generation cruft that is no longer needed now that we've been
distributing a unified tarball in our releases.
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: Christian Boltz <apparmor@cboltz.de>
- drop the symlink magic of the common/ directory, and just include
files directly from there.
- update comments indicating required steps to take when including
common/Make.rules
- drop make clean steps that refer to no longer generated tarballs,
specfiles, and symlinks to the common directory/Make.rules.
- don't silence clean steps if VERBOSE is set
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: Christian "Ghostbuster" Boltz <apparmor@cboltz.de>
make sure nothing accidently hits the first-best target (well,
first-not-so-good would better describe the rpm target ;-)
Also add a dummy "all:" target to the toplevel Makefile with a short
hint towards README.
(see "[patch] fun with the toplevel Makefile") on the ML for the fun
that lead to this patch)
Acked-by: Steve Beattie <steve@nxnw.org>
journal socket. On Debian and Ubuntu systems, /dev/log is a symlink to
/run/systemd/journal/dev-log, so this access is now required in the base
abstraction to maintain current behavior.
Bug: https://bugs.launchpad.net/apparmor/+bug/1413232
Acked-By: Jamie Strandboge <jamie@canonical.com>
Acked-by: Steve Beattie <steve@nxnw.org>
Split is_covered() in capability.py into
- is_covered_localparts() for rule-specific code
- is_covered() for common code - located in __init__.py
The object type comparison now uses type(self) and a slightly different
error message to make it usable everywhere.
Also rename rule_obj to other_rule which is more self-explaining
(inspired by the parameter name in the is_covered() dummy in __init__.py).
v2:
- remove check_allow_deny and check_audit parameters from
is_covered_localvars()
Acked-by: Steve Beattie <steve@nxnw.org>
If one of the testcases fail, this goes unnoticed in "make coverage".
This patch changes the Makefile so that test failures let
"make coverage" fail.
You can use make COVERAGE_IGNORE_FAILURES=true coverage to build
coverage data even if some tests fail.
Signed-off-by: Steve Beattie <steve@nxnw.org>
(which was most probably meant as an Acked-by)
Also Acked-by: <timeout> ;-)
For reasons that are unclear, python's setuptools doesn't install
recursively from a directory, meaning that on make install, the new
Rules/Ruleset classes were not being installed. This patch causes
the rule subdirectory to be included.
Bug: https://bugs.launchpad.net/bugs/1407437
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: Tyler Hicks <tyhicks@canonical.com>
adds some tests for severity.py and improves the test coverage to
nearly 100% (only 3 partial left).
Added tests and details (all in SeverityVarsTest):
- move writing the tunables file from setUp() into _init_tunables() for
more flexibility (allows to specify other file content)
- test adding to a variable (+=)
- test #include
- make sure double definition of a variable fails
- make sure redefinition of non-existing variable fails
BTW: even the comment added to VARIABLE_DEFINITIONS contributes to
the coverage ;-)
severity.py passes all added tests, however I should note that including
a non-existing file is silently ignored.
Acked-by: Kshitij Gupta <kgupta8592@gmail.com>
Adds #include <abstractions/dovecot-common> to the usr.sbin.dovecot
profile. Effectively this adds "deny capability block_suspend," which
is the only missing part from
https://bugs.launchpad.net/apparmor/+bug/1296667/
It also removes "capability setgid," (covered by
abstractions/dovecot-common) and "@{PROC}/filesystems r," (part of
abstractions/base).
Acked-by: John Johansen <john.johansen@canonical.com>
This patch hides raw_rule within the BaseRule class by making parse() be
a class method for all the rule types, implemented via a rule-specific
abstract method _parse() that returns a parsed Rule object.
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: Christian Boltz <apparmor@cboltz.de>
This patch integrated the new capability rule class into aa.py and
cleanprof.py.
Patch changes:
v6:
- fix logic around same_file in cleanprofile.py that was causing
capabilities to be deleted when they weren't covered by an
abstraction.
v5:
- merge my changes into Christian's original patches
- use CapabilityRule.parse() for parsing raw capability rules and
getting a CapabilityRule instance back
- cope with move of parse_modifiers back into rule/__init__.py.
Originally-by: Christian Boltz <apparmor@cboltz.de>
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: Christian Boltz <apparmor@cboltz.de>
Patch changes:
v5:
- merge my changes into Christian's original patches
- update to use CapabilityRule.parse() as the entry point for
parsing raw rules and getting a CapabilityRule instance in
return.
Originally-by: Christian Boltz <apparmor@cboltz.de>
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: Christian Boltz <apparmor@cboltz.de>
This patch adds four classes - two "base" classes and two specific for
capabilities:
utils/apparmor/rule/__init__.py:
class base_rule(object):
Base class to handle and store a single rule
class base_rules(object):
Base class to handle and store a collection of rules
utils/apparmor/rule/capability.py:
class capability_rule(base_rule):
Class to handle and store a single capability rule
class capability_rules(base_rules):
Class to handle and store a collection of capability rules
Changes:
v5:
- flattened my changes into Christian's patches
- pull parse_modifiers into rule/__init__.py
- pull parse_capability into rule/capability.py
- make CapabiltyRule.parse() be the class/static method for parsing
raw capability rules.
- parse_capability: renamed inlinecomment and rawrule to comment
and raw_rule to be consistent with CapabilityRule fields.
Originally-by: Christian Boltz <apparmor@cboltz.de>
Signed-off-by: Steve Beattie <steve@nxnw.org>
Acked-by: Christian Boltz <apparmor@cboltz.de>
[](https://gitlab.com/apparmor/apparmor/pipelines)
[](https://bestpractices.coreinfrastructure.org/projects/1699)
------------
Introduction
------------
@@ -17,9 +23,45 @@ library, available under the LGPL license, which allows change_hat(2)
and change_profile(2) to be used by non-GPL binaries).
For more information, you can read the techdoc.pdf (available after
building the parser) and by visiting the http://apparmor.net/ web
building the parser) and by visiting the https://apparmor.net/ web
site.
----------------
Getting in Touch
----------------
Please send all complaints, feature requests, rants about the software,
Security issues can be filed as security bugs on launchpad
or directed to `security@apparmor.net`. Additional details can be found
in the [wiki](https://gitlab.com/apparmor/apparmor/wikis/home#reporting-security-vulnerabilities).
--------------
Privacy Policy
--------------
The AppArmor security project respects users privacy and data and does not collect data from or on its users beyond what is required for a given component to function.
The AppArmor kernel security module will log violations to the audit subsystem, and those will be logged/forwarded/recorded on the user's system(s) according to how the administrator has logging configured. Again this is not forwarded to or collected by the AppArmor project.
The AppArmor userspace tools do not collect information on the system user beyond the logs and information needed to interact with the user. This is not forwarded to, nor collected by the AppArmor project.
Users may submit information as part of an email, bug report or merge request, etc. and that will be recorded as part of the mailing list, bug/issue tracker, or code repository but only as part of a user initiated action.
The AppArmor project does not collect information from contributors beyond their interactions with the AppArmor project, code, and community. However contributors are subject to the terms and conditions and privacy policy of the individual platforms (currently GitLab and LaunchPad) should they choose to contribute through those platforms. And those platforms may collect data on the user that the AppArmor project does not.
Currently both GitLab an LaunchPad require a user account to submit patches or report bugs and issues. If a contributor does not wish to create an account for these platforms the mailing list is available. Membership in the list is not required. Content from non-list members will be sent to moderation, to ensure that it is on topic, so there may be a delay in choosing to interact in this way.
-------------
Source Layout
@@ -27,6 +69,8 @@ Source Layout
AppArmor consists of several different parts:
```
binutils/ source for basic utilities written in compiled languages
changehat/ source for using changehat with Apache, PAM and Tomcat
common/ common makefile rules
desktop/ empty
@@ -36,6 +80,7 @@ parser/ source for parser/loader and corresponding documentation
profiles/ configuration files, reference profiles and abstractions
tests/ regression and stress testsuites
utils/ high-level utilities for working with AppArmor
```
--------------------------------------
Important note on AppArmor kernel code
@@ -56,55 +101,88 @@ Building and Installing AppArmor Userspace
------------------------------------------
To build and install AppArmor userspace on your system, build and install in
the following order.
the following order. Some systems may need to export various python-related
environment variables to complete the build. For example, before building
anything on these systems, use something along the lines of:
@@ -138,7 +138,7 @@ my $ratelimit_saved = sysctl_read($ratelimit_sysctl);
END { sysctl_write($ratelimit_sysctl, $ratelimit_saved); }
sysctl_write($ratelimit_sysctl, 0);
UI_Info(gettext("\nBefore you begin, you may wish to check if a\nprofile already exists for the application you\nwish to confine. See the following wiki page for\nmore information:\nhttp://wiki.apparmor.net/index.php/Profiles"));
UI_Info(gettext("\nBefore you begin, you may wish to check if a\nprofile already exists for the application you\nwish to confine. See the following wiki page for\nmore information:\nhttps://gitlab.com/apparmor/apparmor/wikis/Profiles"));
UI_Important(gettext("Please start the application to be profiled in \nanother window and exercise its functionality now.\n\nOnce completed, select the \"Scan\" button below in \norder to scan the system logs for AppArmor events. \n\nFor each AppArmor event, you will be given the \nopportunity to choose whether the access should be \nallowed or denied."));
@@ -195,7 +195,7 @@ for my $p (sort keys %helpers) {
}
UI_Info(gettext("Reloaded AppArmor profiles in enforce mode."));
UI_Info(gettext("\nPlease consider contributing your new profile! See\nthe following wiki page for more information:\nhttp://wiki.apparmor.net/index.php/Profiles\n"));
UI_Info(gettext("\nPlease consider contributing your new profile! See\nthe following wiki page for more information:\nhttps://gitlab.com/apparmor/apparmor/wikis/Profiles\n"));
UI_Info(sprintf(gettext('Finished generating profile for %s.'), $fqdbin));
/* only support the "current" and "exec" process attributes */
- return -EINVAL;
+ goto fail;
if (!error)
error = size;
+out:
+ kfree(largs);
return error;
fail:
@@ -588,9 +590,9 @@ fail:
aad.profile = aa_current_profile();
aad.op = OP_SETPROCATTR;
aad.info = name;
- aad.error = -EINVAL;
+ aad.error = error = -EINVAL;
aa_audit_msg(AUDIT_APPARMOR_DENIED, &sa, NULL);
- return -EINVAL;
+ goto out;
}
static int apparmor_task_setrlimit(struct task_struct *task,
--
2.7.4
Some files were not shown because too many files have changed in this diff
Show More
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.