This reverts commit 9899568f6c.
Reason for revert: Reports of high numbers of SELinux denials
showing up on the SELinux dashboard.
Bug: 110043362
Change-Id: Id8fc260c47ffd269ac2f15ff7dab668c959e3ab0
What changed:
- Removed cgroup access from untrusted and priv apps.
- Settings app writes to /dev/stune/foreground/tasks, so system_app domain
retains access to cgroup.
- libcutils exports API to /dev/{cpuset, stune}/*. This API seems to be used
abundantly in native code. So added a blanket allow rule for (coredomain - apps)
to access cgroups.
- For now, only audit cgroup access from vendor domains. Ultimately, we want to
either constrain vendor access to individual domains or, even better, remove
vendor access and have platform manage cgroups exclusively.
Bug: 110043362
Test: adb shell setprop ro.config.per_app_memcg true, device correctly populates
/dev/memcg on a per app basis on a device that supports that.
Test: aosp_sailfish, wahoo boot without cgroup denials
Change-Id: I9e441b26792f1edb1663c660bcff422ec7a6332b
Access is deprecated for apps with targetSdkVersion=26+.
Test: build (neverallow rules are build time assertions)
Change-Id: I36480c38d45cf6bfb75f4988ffcefefc6b62d4b1
DropboxManager may pass FDs to any app with the READ_LOGS
permission which is available to all apps as a development
permission.
Test: atest CtsIncidentHostTestCases
Fixes: 111856304
Change-Id: I329e3125dab83de948b860061df9d232e31cb23e
Bug: 110887137
Test: Flash new system policy onto a device with vendor policy that uses
untrusted_app_visible_* attributes, and check that old and new attributes
are applied to exactly same types.
Change-Id: Ibee0ec645878fcc8c93cd0fbd169a8d45129d79e
Merged-In: Ibee0ec645878fcc8c93cd0fbd169a8d45129d79e
(cherry picked from commit 7abca51d19)
Currently, both untrusted apps and priv-apps use the SELinux file label
"app_data_file" for files in their /data/data directory. This is
problematic, as we really want different rules for such files. For
example, we may want to allow untrusted apps to load executable code
from priv-app directories, but disallow untrusted apps from loading
executable code from their own home directories.
This change adds a new file type "privapp_data_file". For compatibility,
we adjust the policy to support access privapp_data_files almost
everywhere we were previously granting access to app_data_files
(adbd and run-as being exceptions). Additional future tightening is
possible here by removing some of these newly added rules.
This label will start getting used in a followup change to
system/sepolicy/private/seapp_contexts, similar to:
-user=_app isPrivApp=true domain=priv_app type=app_data_file levelFrom=user
+user=_app isPrivApp=true domain=priv_app type=privapp_data_file levelFrom=user
For now, this newly introduced label has no usage, so this change
is essentially a no-op.
Test: Factory reset and boot - no problems on fresh install.
Test: Upgrade to new version and test. No compatibility problems on
filesystem upgrade.
Change-Id: I9618b7d91d1c2bcb5837cdabc949f0cf741a2837
This hwservice isn't registered with hwservicemanager but rather passed
to the thermal hal, so it doesn't need sepolicy associated with it to
do so.
Test: manual: boot, inspect logs
Test: VtsHalThermalV1_1TargetTest
Bug: 109802374
Change-Id: Ifb727572bf8eebddc58deba6c0ce513008e01861
Merged-In: Ifb727572bf8eebddc58deba6c0ce513008e01861
Grant access to icmp_socket to netdomain. This was previously
labeled as rawip_socket which apps are allowed to use. Neverallow
all other new socket types for apps.
Kernels versions > 4.9 redefine ICMP sockets from rawip_socket
to icmp_socket. To pass neverallow tests, we need to define
which IOCTLs are allowed (and disallowed).
Note that this does not change behavior on devices with
kernel versions <=4.9. However, it is necessary (although not
sufficient) to pass CTS on kernel version 4.14.
Bug: 110520616
Test: Grant icmp_socket in net.te and build.
Change-Id: I5c7cb6867d1a4cd1554a8da0d55daa8e06daf803
(breaks vendor blobs, will have to be regenerated
after this CL)
This moves mediacodec to vendor so it is replaced with
hal_omx_server. The main benefit of this is that someone
can create their own implementation of mediacodec without
having to alter the one in the tree. mediacodec is still
seccomp enforced by CTS tests.
Fixes: 36375899
Test: (sanity) YouTube
Test: (sanity) camera pics + video
Test: check for denials
Change-Id: I31f91b7ad6cd0a891a1681ff3b9af82ab400ce5e
Create a new label for /data/system/dropbox, and neverallow direct
access to anything other than init and system_server.
While all apps may write to the dropbox service, only apps with
android.permission.READ_LOGS, a signature|privileged|development
permission, may read them. Grant access to priv_app, system_app,
and platform_app, and neverallow access to all untrusted_apps.
Bug: 31681871
Test: atest CtsStatsdHostTestCases
Test: atest DropBoxTest
Test: atest ErrorsTests
Change-Id: Ice302b74b13c4d66e07b069c1cdac55954d9f5df
We're adding support for OEMs to ship exFAT, which behaves identical
to vfat. Some rules have been manually enumerating labels related
to these "public" volumes, so unify them all behind "sdcard_type".
Test: atest
Bug: 67822822
Change-Id: I09157fd1fc666ec5d98082c6e2cefce7c8d3ae56
This is a partial cherry pick of commit 6231b4d9
'Enforce per-app data protections for targetSdk 28+'.
Untrusted_app_27 remains unreachable, but it's existence
prevents future merge conflicts.
Bug: 63897054
Test: build/boot aosp_walleye-userdebug
Change-Id: I64b013874fe87b55f47e817a1279e76ecf86b7c0
Merged-In: I64b013874fe87b55f47e817a1279e76ecf86b7c0
(cherry picked from commit 6231b4d9fc)
Access to these files was removed in Oreo. Enforce that access is not
granted by partners via neverallow rule.
Also disallow most untrusted app access to net.dns.* properties.
Bug: 77225170
Test: system/sepolicy/tools/build_policies.sh
Change-Id: I85b634af509203393dd2d9311ab5d30c65f157c1
This is an experimental feature only on userdebug and eng build.
Test: play MP4 file. install & uninstall media update apk.
Bug: 67908547
Change-Id: I513cdbfda962f00079e886b7a42f9928e81f6474
The NeuralNetworks runtime is a library that communicates with
NeuralNetworks HIDL services and is linked by applications. To enable
the NN runtime to use these services, applications must have explicit
sepolicy permissions to find the NN services and communicate across
binder.
This CL relaxes neverallow rules for hal_neuralnetworks_*.
Because it is affecting pre-existing neverallow rules, this CL requires
a CTS rebuild.
Bug: 70340780
Test: mm
Test: ran neuralnetworks vts and cts binaries
Change-Id: I84f73ac77486681f91d1f8687268c0fa22a7ba0b
(cherry picked from commit 598870bebc4bb34542df81799b46f3cdcfb6723b)
Do not let apps read uid_concurrent_active_time and
uid_concurrent_policy_time.
b/68399339
Test: Check that they can't be read from the shell
without root permissions and system_server was able
to read them
Change-Id: I6f09ef608607cb9f4084ba403a1e7254b8c49a06
Renamed this type:
proc_asound_cards -> proc_asound
Labeled /proc/asound/devices as proc_asound.
We now use proc_asound type to label files under /proc/asound which we
want to expose to system components.
Bug: 66988327
Test: Pixel 2 boots, can play sound with or without headphones, and
selinux denials to proc_asound are not seen.
Change-Id: I453d9bfdd70eb80931ec9e80f17c8fd0629db3d0
Same-process HALs are forbidden except for very specific HALs that have
been provided and whitelisted by AOSP. As a result, a vendor extension
HAL may have a need to be accessed by untrusted_app. This is still
discouraged, and the existing AOSP hwservices are still forbidden, but
remove the blanket prohibition. Also indicate that this is temporary,
and that partners should expect to get exceptions to the rule into AOSP
in the future.
Bug: 62806062
Test: neverallow-only change builds. Verify new attribute is in policy.
Change-Id: I6d3e659147d509a3503c2c9e0b6bb9016cc75832
This was accidentally omitted from all_untrusted_app
While I'm here, split across mutiple lines and alphabetize.
Test: policy compiles.
Change-Id: I7fe1d1d0a4ef2ed3ab010931ee2ba15637c2be51
The fuse_device neverallow rules are too aggressive and are inhibiting
certain vendor customizations. Relax the /dev/fuse neverallow rules so
that they better reflect the security invariants we want to uphold.
Bug: 37496487
Test: policy compiles.
Change-Id: Ie73b0ba7c76446afc2a7a23ebed1275c977d932d
This adds neverallow rules which enforce the prohibition on
communication between framework and vendor components over VendorBinder.
This prohibition is similar in spirit to the one for Binder
communications.
Most changes consist of adding neverallow rules, which do not affect
runtime behavior. The only change which does affect runtime behavior
is the change which takes away the right of servicemanager domain to
transfer Binder tokens to hwservicemanager and vndservicemanager. This
grant was there by accident (because it was overly broad) and is not
expected to be needed: servicemanager, hwservicemanager, and
vndservicemanager are not supposed to be communicating with each
other.
P. S. The new neverallow rules in app_neverallows.te are covered by
the new rules in domain.te. The rules were nevertheless added to
app_neverallows.te for consistency with other *Binder rules there.
Test: mmm system/sepolicy
Bug: 37663632
Change-Id: I7c2ae23924bf0f2fed3f1e3a8d4d603129286329
This adds a neverallow rules which checks that SELinux app domains
which host arbitrary code are not allowed to access hwservicemanager
operations other than "find" operation for which there already are
strict neverallow rules in the policy.
Test: mmm system/sepolicy -- neverallow-only change
Bug: 34454312
Change-Id: I3b80c6ae2c254495704e0409e0c5c88f6ce3a6a7
App domains which host arbitrary code must not have access to
arbitrary HwBinder services. Such access unnecessarily increases the
attack surface. The reason is twofold:
1. HwBinder servers do not perform client authentication because HIDL
currently does not expose caller UID information and, even if it
did, many HwBinder services either operate at a layer below that of
apps (e.g., HALs) or must not rely on app identity for
authorization. Thus, to be safe, the default assumption is that
a HwBinder service treats all its clients as equally authorized to
perform operations offered by the service.
2. HAL servers (a subset of HwBinder services) contain code with
higher incidence rate of security issues than system/core
components and have access to lower layes of the stack (all the way
down to hardware) thus increasing opportunities for bypassing the
Android security model.
HwBinder services offered by core components (as opposed to vendor
components) are considered safer because of point #2 above.
Always same-process aka always-passthrough HwBinder services are
considered safe for access by these apps. This is because these HALs
by definition do not offer any additional access beyond what its
client already as, because these services run in the process of the
client.
This commit thus introduces these two categories of HwBinder services
in neverallow rules.
Test: mmm system/sepolicy -- this does not change on-device policy
Bug: 34454312
Change-Id: I4f5f4dd10b3fc3bb9d262dda532d4a23dcdf061d
MediaProvider requires permissions that diverge from those
of a typical priv_app. This create a new domain and removes
Mtp related permissions from priv_app.
Bug: 33574909
Test: Connect with MTP, download apps and files, select ringtones
Test: DownloadProvider instrument tests, CtsProviderTestCases
Change-Id: I950dc11f21048c34af639cb3ab81873d2a6730a9
Prevent app domains (processes spawned by zygote) from acquiring
locks on files in /system. In particular, /system/etc/xtables.lock
must never be lockable by applications, as it will block future
iptables commands from running.
Test: device boots and no obvious problems.
Change-Id: Ifd8dc7b117cf4a622b30fd4fffbcab1b76c4421b
We simplified the way we track whether or not a dex file is used by
other apps. DexManager in the framework keeps track of the data and we
no longer need file markers on disk.
Test: device boots, foreign dex markers are not created anymore
Bug: 32871170
Change-Id: I464ed6b09439cf0342020ee07596f9aa8ae53b62
untrusted_v2_app is basically a refinement of untrusted_app with legacy
capabilities removed and potentially backwards incompatible changes.
This is not currently hooked up to anything.
Bug: 33350220
Test: builds
Change-Id: Ic9fad57476bc2b6022b1eaca8667bf6d844753c2
Motivation:
Provide the ability to phase in new security policies by
applying them to apps with a minimum targetSdkVersion.
Place untrusted apps with targetSdkVersion<=25 into the
untrustd_app_25 domain. Apps with targetSdkVersion>=26 are placed
into the untrusted_app domain. Common rules are included in the
untrusted_app_all attribute. Apps with a more recent targetSdkVersion
are granted fewer permissions.
Test: Marlin builds and boots. Apps targeting targetSdkVersion<=25
run in untrusted_app_25 domain. Apps targeting the current development
build >=26 run in the untrusted_app domain with fewer permissions. No
new denials observed during testing.
Bug: 34115651
Bug: 35323421
Change-Id: Ie6a015566fac07c44ea06c963c40793fcdc9a083
The implementation for NETLINK_FIREWALL and NETLINK_IP6_FW protocols
was removed from the kernel in commit
d16cf20e2f2f13411eece7f7fb72c17d141c4a84 ("netfilter: remove ip_queue
support") circa Linux 3.5. Unless we need to retain compatibility
for kernels < 3.5, we can drop these classes from the policy altogether.
Possibly the neverallow rule in app.te should be augmented to include
the newer netlink security classes, similar to webview_zygote, but
that can be a separate change.
Test: policy builds
Change-Id: Iab9389eb59c96772e5fa87c71d0afc86fe99bb6b
Signed-off-by: Stephen Smalley <sds@tycho.nsa.gov>
The neverallows in untrusted_app will all apply equally to ephemeral app
and any other untrusted app domains we may add, so this moves them to a
dedicated separate file.
This also removes the duplicate rules from isolated_app.te and ensures
that all the untrusted_app neverallows also apply to isolated_app.
Test: builds
Change-Id: Ib38e136216ccbe5c94daab732b7ee6acfad25d0b
