Goal is to gain a better handle on who has access to which maps
and to allow (with bpfloader changes to create in one directory
and move into the target directory) per-map selection of
selinux context, while still having reasonable defaults for stuff
pinned directly into the target location.
BPFFS (ie. /sys/fs/bpf) labelling is as follows:
subdirectory selinux context mainline usecase / usable by
/ fs_bpf no (*) core operating system (ie. platform)
/net_private fs_bpf_net_private yes, T+ network_stack
/net_shared fs_bpf_net_shared yes, T+ network_stack & system_server
/netd_readonly fs_bpf_netd_readonly yes, T+ network_stack & system_server & r/o to netd
/netd_shared fs_bpf_netd_shared yes, T+ network_stack & system_server & netd [**]
/tethering fs_bpf_tethering yes, S+ network_stack
/vendor fs_bpf_vendor no, T+ vendor
* initial support for bpf was added back in P,
but things worked differently back then with no bpfloader,
and instead netd doing stuff by hand,
bpfloader with pinning into /sys/fs/bpf was (I believe) added in Q
(and was definitely there in R)
** additionally bpf programs are accesible to netutils_wrapper
for use by iptables xt_bpf extensions
'mainline yes' currently means shipped by the com.android.tethering apex,
but this is really another case of bad naming, as it's really
the 'networking/connectivity/tethering' apex / mainline module.
Long term the plan is to merge a few other networking mainline modules
into it (and maybe give it a saner name...).
The reason for splitting net_private vs tethering is that:
S+ must support 4.9+ kernels and S era bpfloader v0.2+
T+ must support 4.14+ kernels and T beta3 era bpfloader v0.13+
The kernel affects the intelligence of the in-kernel bpf verifier
and the available bpf helper functions. Older kernels have
a tendency to reject programs that newer kernels allow.
/ && /vendor are not shipped via mainline, so only need to work
with the bpfloader that's part of the core os.
Bug: 218408035
Test: TreeHugger, manually on cuttlefish
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: I674866ebe32aca4fc851818c1ffcbec12ac4f7d4
(cherry picked from commit 15715aea32)
... such as Cuttlefish (Cloud Android virtual device) which has a
DRM virtio-gpu based gralloc and (sometimes) DRM virtio-gpu based
rendering (when forwarding rendering commands to the host machine
with Mesa3D in the guest and virglrenderer on the host).
After this change is submitted, changes such as aosp/1997572 can
be submitted to removed sepolicy that is currently duplicated
across device/google/cuttlefish and device/linaro/dragonboard as
well.
Adds a sysfs_gpu type (existing replicated sysfs_gpu definitions
across several devices are removed in the attached topic). The
uses of `sysfs_gpu:file` comes from Mesa using libdrm's
`drmGetDevices2()` which calls into `drmParsePciDeviceInfo()` to
get vendor id, device id, version etc.
Bug: b/161819018
Test: launch_cvd
Test: launch_cvd --gpu_mode=gfxstream
Change-Id: I4f7d4b0fb90bfeef72f94396ff0c5fe44d53510c
Merged-In: I4f7d4b0fb90bfeef72f94396ff0c5fe44d53510c
The binder driver now advertises the features it supports through
individual files under /dev/binderfs/features/*. Let all domains have
access to these files to determine how to interact with the driver.
Bug: 191910201
Tested: clients are able to read feature files via libbinder
Signed-off-by: Carlos Llamas <cmllamas@google.com>
Change-Id: Ice5de9efee74e571ef0a23ce093af162fc3b276e
Who needs all those context switches?
bpfloader controls which types of vendor programs can be used.
Bug: 140330870
Bug: 162057235
Test: successfully load bpf programs from vendor
Change-Id: I36e4f6550da33fea5bad509470dfd39f301f13c8
Bug: 202785178
Test: Along with rest of topic, file
/sys/fs/bpf/prog_fuse_media_fuse_media
appears on boot with fuse-bpf in kernel
Merged-In: Ibccdf177c75fef0314c86319be3f0b0f249ce59d
Change-Id: Ibccdf177c75fef0314c86319be3f0b0f249ce59d
rss_stat will be throttled using histogram triggers and synthetic trace
events. Add genfs context labels for the synthetic tracefs files.
Bug: 145972256
Test: Check log cat for avc denials
Change-Id: I7e183aa930bb6ee79613d011bed7174d553f9c1a
The root init.rc does "write /proc/cpu/alignment 4", but we don't
actually allow this write in core sepolicy. This seems to be a 32-bit
ARM only proc file.
Noticed when booting 32-bit ARM Cuttlefish.
Bug: 145371497
Change-Id: Ic099395708f7236bcc2fc5c561809a7e129786de
Stop using these SELinux attributes since the apexd and init SELinux
policies no longer rely on these attributes.
The difference between the previous versions of this patch and the
current patch is that the current patch does not remove any SELinux
attributes. See also
https://android-review.googlesource.com/c/platform/system/sepolicy/+/1850656.
See also
https://android-review.googlesource.com/c/platform/system/sepolicy/+/1862919.
This patch includes a revert of commit 8b2b951349 ("Restore permission
for shell to list /sys/class/block"). That commit is no longer necessary
since it was a bug fix for the introduction of the sysfs_block type.
Bug: 202520796
Test: source build/envsetup.sh && lunch aosp_x86_64 && m && launch_cvd
Change-Id: I73e1133af8146c154af95d4b96132e49dbec730c
Signed-off-by: Bart Van Assche <bvanassche@google.com>
Remove these SELinux attributes since the apexd and init SELinux policies
no longer rely on these attributes.
The only difference between a previous version of this patch and the
current patch is that the current patch moves these attributes to the
'compat' policy. See also
https://android-review.googlesource.com/c/platform/system/sepolicy/+/1850656.
This patch includes a revert of commit 8b2b951349 ("Restore permission
for shell to list /sys/class/block"). That commit is no longer necessary
since it was a bug fix for the introduction of the sysfs_block type.
Bug: 202520796
Test: source build/envsetup.sh && lunch aosp_x86_64 && m && launch_cvd && adb -e shell dmesg | grep avc
Change-Id: Id7d32a914e48bc74da63d87ce6a09f11e323c186
Signed-off-by: Bart Van Assche <bvanassche@google.com>
Revert "Remove the bdev_type and sysfs_block_type SELinux attributes"
Revert "Remove the bdev_type and sysfs_block_type SELinux attributes"
Revert submission 1850578-remove-selinux-bdev-type
Reason for revert: DroidMonitor-triggered revert due to breakage, bug b/203480787
BUG: 203480787
Reverted Changes:
I263bce9c4:Remove the bdev_type and sysfs_block_type SELinux ...
Ibc9039f96:Revert "Add the 'bdev_type' attribute to all block...
Ic6ae83576:Remove the bdev_type and sysfs_block_type SELinux ...
Ie493022a8:Remove the bdev_type and sysfs_block_type SELinux ...
I1f1ca439b:Revert "Add the 'bdev_type' attribute to all block...
I283f8676b:Revert "Add the 'bdev_type' attribute to all block...
I7c5c242c5:Revert "Add the 'bdev_type' attribute to all block...
Id78d8f7dc:Remove the bdev_type and sysfs_block_type SELinux ...
I9c4b2c48b:Remove the bdev_type and sysfs_block_type SELinux ...
I51e9d384a:Remove the bdev_type and sysfs_block_type SELinux ...
I2c414de3b:Remove the sysfs_block_type SELinux attribute
Change-Id: I55609803d530772d507d9dca8ba202a96daf24b7
Remove these SELinux attributes since adding these attributes introduces
a depencency from vendor SELinux policies on the generic SELinux policy,
something that is not allowed. This patch includes a revert of commit
8b2b951349 ("Restore permission for shell to list /sys/class/block").
That commit is no longer necessary since it was a bug fix for the
introduction of the sysfs_block type.
Bug: 202520796
Test: source build/envsetup.sh && lunch aosp_x86_64 && m && launch_cvd
Change-Id: Ic6ae835768212648ca09fd5c83c39180103c3b1b
Signed-off-by: Bart Van Assche <bvanassche@google.com>
extra_free_kbytes.sh is used by init to set /sys/vm/watermark_scale_factor
value. Allow init to execute extra_free_kbytes.sh and the script to access
/proc/sys/vm/watermark_scale_factor and /proc/sys/vm/extra_free_kbytes
files.
Bug: 109664768
Signed-off-by: Suren Baghdasaryan <surenb@google.com>
Change-Id: I55ec07e12a1cc5322cfdd4a48d0bdc607f45d832
Addressing b/194450129 requires configuring the I/O scheduler and the
queue depth of loop devices. Doing this in a generic way requires
iterating over the block devices under /sys/class/block and also to
examine the properties of the boot device (/dev/sda). Hence this patch
that allows 'init' and 'apexd' to read the properties of all block
devices. The patch that configures the queue depth is available at
https://android-review.googlesource.com/c/platform/system/core/+/1783847.
Test: Built Android images, installed these on an Android device and verified that modified init and apexd processes do not trigger any SELinux complaints.
Change-Id: Icb62449fe0d21b3790198768a2bb8e808c7b968e
Signed-off-by: Bart Van Assche <bvanassche@google.com>
Previously vendor_sched is put under product area which will be replaced
by GSI. To solve it, move it to system/sepolicy.
Bug: 194656257
Test: build pass
Change-Id: Ia0b855e3a876a58b58f79b4fba09293419797b47
Create contexts for /sys/kernel/tracing/instances/bootreceiver
Allow read access to files in this dir for system_server.
Bug: 172316664
Bug: 181778620
Test: manual runs with KFENCE enabled
Signed-off-by: Alexander Potapenko <glider@google.com>
Change-Id: I7021a9f32b1392b9afb77294a1fd0a1be232b1f2
Revert submission 1572240-kernel_bootreceiver
Reason for revert: DroidMonitor: Potential culprit for Bug 181778620 - verifying through Forrest before revert submission. This is part of the standard investigation process, and does not mean your CL will be reverted.
Reverted Changes:
Ic1c49a695:init.rc: set up a tracing instance for BootReceive...
I828666ec3:Selinux policy for bootreceiver tracing instance
Change-Id: I9a8da7ae501a4b7c3d6cb5bf365458cfd1bef906
Create contexts for /sys/kernel/tracing/instances/bootreceiver
Allow read access to files in this dir for system_server.
Bug: 172316664
Test: manual runs with KFENCE enabled
Signed-off-by: Alexander Potapenko <glider@google.com>
Change-Id: I828666ec3154aadf138cfa552832a66ad8f4a201
Vendor boot hal, init, and vold processes all require permission.
Test: build and boot aosp_cf_x86_64_phone
Bug: 173815685
Change-Id: I15692dcd39dfc9c3a3b7d8c12d03eff0a7c96f72
This primarily affects perfetto's traced_probes and shell-invoked
binaries like atrace, but also anyone with access to "debugfs_tracing".
These tracepoints are being actively collected in internal tracing, so
we would like to also make them available on release builds, as they
should be a source of useful system information there as well.
The ones we definitely need:
* sched_waking, sched_wakeup_new: both are similar to the
already-allowed sched_wakeup. The first differs in which exact process
context it occurs in, and the latter is the wakeup events of only the
fresh tasks.
* oom/mark_victim: contains only the pid of the victim. Useful for
memory-related tracing and analysis.
The other events in this patch are of lesser importance, but also are
fairly straightforward - clocks and priority for frequency/power tracing.
Small extra change: sched_process_free was only relabeled in the tracefs
block, so I've added it to debugfs to keep them in sync. (I wonder whether
debugfs is even necessary at this point... but that's outside of scope
here.)
See the attached bug for a longer explanation. There will also be a
separate patch for system/frameworks/native/atrace/atrace.rc for the
Unix file permissions of these files.
Bug: 179788446
Tested: I did not have access to a "user" build, but I've manually
checked the labels of events/.../enable tracefs files via ls -Z,
and strace'd traced_probes on a hacky debug build where I
commented out its SELinux allow-rule for debugfs_tracing_debug.
Change-Id: I15a9cb33950718757e3ecbd7c71de23b25f85f1d
We want to label /sys/fs/bpf/tethering/... with a new label distinct
from /sys/fs/bpf, as this will allow locking down the programs/maps
tighter then is currently possible with the existing system.
These programs and maps are provided via the tethering mainline module,
and as such their number, names, key/value types, etc. are all prone to
be changed by a tethering mainline module update.
Test: atest, TreeHugger
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: Ifc4108d76a1106a936b941a3dda1abc5a65c05b0
These permissions are required for dumpstate to read the DMA-BUF sysfs
stats present at /sys/kernel/dmabuf/buffers
Bug: 167709539
Test: adb shell am bug-report
Change-Id: I1c00843775452b7a7aa39b059e1d77d77aed1e9c
IncFS in S adds a bunch of new ioctls, and requires the users
to read its features in sysfs directory. This change adds
all the features, maps them into the processes that need to
call into them, and allows any incfs user to query the features
Bug: 170231230
Test: incremental unit tests
Change-Id: Ieea6dca38ae9829230bc17d0c73f50c93c407d35
Access to /proc/locks is necessary to activity manager to determine
wheter a process holds a lock or not prior freezing it.
Test: verified access of /proc/locks while testing other CLs in the same
topic.
Bug: 176928302
Change-Id: I14a65da126ff26c6528edae137d3ee85d3611509
Bug: 169279846
Test: atest bpf-time-in-state-tests
Test: verified that the time-in-state BPF prog still loads into
the kernel with no errors and gets attached without errors
Change-Id: If74632ae6f72e0371fea844d4ba7bef9260d1bdb
Test: ls -lZ /sys/kernel/tracing/printk_formats
[...] u:object_r:debugfs_tracing_printk_formats:s0 [...]
Test: setenforce 0;
runcon u:r:system_server:s0 cat /sys/kernel/tracing/printk_formats
logcat complains about /sys/kernel/tracing/printk_formats
Test: setenforce 0;
runcon u:r:traced_probes:s0 cat /sys/kernel/tracing/printk_formats
logcat does not complain about /sys/kernel/tracing/printk_formats
(need to setenforce 0, because otherwise the exec of ls is denied).
Bug: 70292203
Change-Id: I15ddef686f979c59daaba5263fa99aca3cd139e5
This CL allows the traced_probes service to temporarily
lower kptr_restrict and read /proc/kallsyms.
This is allowed only on userdebug/eng builds.
The lowering of kptr_restrict is done via an init
property because the kernel checks that the kptr_restrict
writer is CAP_SYS_ADMIN, regardless of the /proc file ACLs [1].
[1] 4cbffc461e/kernel/sysctl.c (L2254)
Bug: 136133013
Design doc: go/perfetto-kallsyms
Test: perfetto_integrationtests --gtest_filter=PerfettoTest.KernelAddressSymbolization in r.android.com/1454882
Change-Id: Ic06e7a9a74c0f3e42fa63f7f41decc385c9fea2c
irq and ipi tracepoint are required for Traceur's irq category.
Bug: 153486528
Test: manually captured a Traceur trace on user builds, verified that the
tracepoints were present.
Change-Id: I2f42fc293e87e2a3567ab55f7533fe6a5d7852bf
cgroup v2 is going to be used for freezer v2 support. The cgroup v2 hiearchy
will be mounted by init under /sys/fs/cgroup hence proper access rights
are necessary for sysfs. After mounting, the cgroup v2 kernfs will use
the label cgroup_v2 and system_manager will handle the freezer
Bug: 154548692
Test: verified that files undes sysfs and cgroup v2 kernfs are accessed
as required to allow proper functioning for the freezer.
Change-Id: Idfb3f6e77b60dad032d1e306d2f9b58cd5775960
