If the qtaguid proc file is no longer exist, the device is running new
eBPF module to do traffic accounting. So the NetworkStatsFactory need to
use the proc/net/dev interface to get the per interface traffic stats
summary. Also, adding a test to verify the helper function work properly
Bug: 30950746
Test: run NetworkStatsFactoryTest
Change-Id: Ia36808bf02f1637dd41a3e7c50917b91b1a77524
Add the native method used to read the detail information of network
stats from bpf maps. The native method of NetworkStatsFactory should
choose the correct implementation to get the stats detail depending on
the kernel version. Currently the bpf result is printed as a reference
and the actual behavior of NetworkStatsFactory should not change.
Test: NetworkStatsFactory related cts test should not fail.
Bug: 30950746
Change-Id: I4715a23559b5b2306bd556cea0431f0ed172a993
Add the native implementation for NetworkStatsService to read the
uidStatsMap and tagStatsMap to get the network traffic information.
Currently the implementation only try to get the data from bpf for a
reference but the final value returned is still from old xt_qtaguid
module. So the result will not be effected. The bpf implementation is
depending on libbpf inside netd.
Bug: 30950746
Test: CtsNetTestCases -t android.net.cts.TrafficStatsTest
Change-Id: Ia85e4d4d602f6bd536739d89085e6dc37ad82e15
This change adds an additional check in CheckIpsecConfig to prevent
users from using the same SPI twice. This allows for a more granular
error message.
Bug: 70642141
Test: Tests added in IpSecServiceParameterizedTest
Change-Id: I9621fb05c6b162bd8ae8db4ac1e64feaa9d0ac73
The only caller is telephony code, and it's being fixed in this
topic.
Bug: 35142602
Test: builds, boots
Change-Id: Ia77a1c2c297037f311c2355610c206cd8e3192c4
CreateTransportModeTransform and DeleteTransportModeTransform are both
agnostic as far as which mode of transform it creates/deletes. As such,
to facilitate the implementation of tunnel mode, this patch renames them
to CreateTransform and DeleteTransform, along with all test names.
Bug: 63588681
Test: frameworks/base unit tests and CTS tests run, passed
Change-Id: I1f015eb7ad0e85fca966658a9402485ca2b44091
This change adds the defaultNetwork to the ident based on current
system networking state, and persists that value to flash.
Bug: 35142602
Test: runtest frameworks-net
Test: downloaded files using default/non-default networks while watching dumpsys netstats
Change-Id: Id8776d149b2977548a9eb455ad4926af55e25aba
This allows us to maintain NetworkStats entries that track
whether the traffic was on the default network.
At the moment, the stats collection code always passes in
DEFAULT_NETWORK_NO. However, this value is a no-op, since it is
not persisted to disk. Only the ident, the uid/set/tag, and the
packet/byte/operation counters are persisted.
A future change will add defaultNetwork to the ident and start
persisting it.
Bug: 35142602
Test: runtest frameworks-net
Change-Id: Ifa291c62c0fa389b88e5561086a29dcd7cee2253
This will allow NetworkStatsService to treat traffic on these
networks differently from traffic where the app selects a network
that is not the default.
Bug: 35142602
Test: runtest frameworks-net
Change-Id: I5ea9d200d9fb153490c6108bb9390bf152f297da
Add a new interface and a new management object,
IpSecTunnelInterface to the IpSecManager surface.
This object will be used to control IPsec tunnels.
-Add IpSecTunnelInterface object
-Add methods to create and use an IpSecTunnelInterface
-Update the IpSecTransform builder to create Tunnel
mode IpSecTransform objects (usable with an IpSecTunnel)
Bug: 36033193
Test: compilation
Change-Id: Ib6948b12c15c93674234dc36288058ae44435b90
Added some extra comments on reference counting and moved a few methods
around. No significant logical changes made in this CL
Bug: 63409385
Test: CTS, Unit tests (both frameworks-base and netd) and binder tests
all pass
Change-Id: I89f1f4a021db48ae406fefefa6aca7406045736c
Because IpSecTransforms are now unidirectional,
and because the only mechanism for removing Transforms
removes it from both directions, the API can no longer
use the Transform parameter to meaningfully validate
that the caller had applied a transform. Since that
functionality was as-yet unimplemented and is now
infeasible, the transform parameter is removed.
Bug: 72079356
Test: cts - IpSecManagerTest; runtest frameworks-net
Change-Id: If19b0d34bdc6daf31a40d6d62bff326dcbca08c0
The version of applyTransportModeTransform() and
removeTransportModeTransform() that accepted
Socket and DatagramSocket were closing the underlying
FDs upon return. It's unclear whether this is due to
a behavior change elsewhere in ParcelFileDescriptor,
but either way, converting to using getFileDescriptor$
and then calling dup() explicitly rather than relying
on ParcelFileDescriptor seems like a better idea anyway.
Bug: 72047396
Test: CTS - IpSecManagerTest.testCreateTransform()
Change-Id: Ia2f02564e1289f25bf113dbb861fcfd2240537a7
Convert the IpSecTransform from being a bi-directional
pair of SAs to a unidirectional single SA. This CL
also removes the concept of "direction from SAs meaning
that a IpSecTransform may now be applied to a socket
in either direction.
-Make transforms unidirectional
-Add Convert allocateSpi() to use destination rather
than direction and remote address
-Remove directionality from builders for IpSecTransform
-Change applyTransportModeTransform() to take a
direction in which to apply the transform object.
-Additional minor naming updates
-Restrict IpSecConfig to only print keys on eng builds
-Move DIRECTION constants to IpSecManager
-Add sourceAddress parameter to IpSecTransform to provide
additional guarantees about the source address of data;
(explicit failure rather than implicit failure).
-Move SPI to the build() method of IpSecTransform
Bug: 71717213
Test: runtest frameworks-net, CTS - IpSecManagerTest
Change-Id: I0824b37f443f4b8c62536d9801238c63ed8f2a1c
Adds checks to ensure that users can only set the correct types of
algorithms for the Authentication, Encryption and Authenticated
Encryption algorithms.
Bug: 65223935
Test: Added tests in IpSecConfigTest, and passed on aosp_marlin-eng
Change-Id: I462c77d9eb5710b8d03a48866453649d3b6fc6bf
The mNextResourceId variable is only accessed within synchronized
blocks, so there is no need to use an atomic integer to synchronize
it. This eliminates the misleading notion that the variable is
accessed outside of guarded blocks, which it is not.
Bug: 62279167
Test: cts
Change-Id: I815835622659f54d2d2d33b349b17c632ebced8d
To facilitate error checking we should invalidate
the resource IDs of all objects when we close() them.
Today, the resource ID is invalidated on the Transform
object but not on the SPI or Encap Socket. This CL
unifies the behavior.
Bug: 70641274
Test: cts - IpSecManagerTest
Change-Id: I28caec3e913902c748c6a50b4ef742ccef8b1b09
CTS test should fail if IPV6 Sockets is used on IPV4 address.
Unfortunately its been passing fine due to unhandled catch
exception. This is fixed in this CL.
Bug: 64435904
Test: Ran CTS test to verify ./run_cts.sh
Change-Id: I369c1b5baaa64b64feaa3f298e61f46d1200ea17
Symptom:
System_server crashed due to unsolved "netd" service.
Root cause:
My previous patch was not enough for avoiding this crash.
a00d4b0 Abort connecting to native daemon during shutdown
listenToSocket loop can be finished without throwing exception.
Solution:
It's better to test shutdown condition before reconnection even
though listenToSocket is finished without exception.
Bug: 68742327
Change-Id: Ifdfe43d52bef891f55bd7b07cc1aa0fa248a6030
Move the native implementation of TrafficStats to NetworkStatsService
and apps need to get the NetworkStatsService binder interface from
system_server in order to get the network usage stats since boot. This
implementation can hide the detail of retrieving network traffic
information from apps and the NetworkStatsService can choose which
methoed it use to get the data depending on the kernel version of the
device.
Bug: 30950746
Test: CtsNetTestCases -t android.net.cts.TrafficStatsTest
Change-Id: I53bbefd19aa0b783b9b4b42ea4d76db3e9ec07a3
Encap sockets are currently created as the system server, and should be
fchown'd to the user for whom it was created on behalf of.
Bug: 62994731
Test: New tests added and run to IpSecService
Change-Id: Icc49e709ae588981e69765fdb77537d7ffbac5fe
Added calls to tag encap sockets to that of the UID for which the encap
socket is being created on behalf of. This ensures that all data
accounting generated for the UDP-encap-ESP socket is correctly billed to
the right UID.
Bug: 62994731
Test: New tests added to IpSecServiceTest.java, passing
Change-Id: I15365ea9c982fd7b4e3cdeff314ddfba2289c86e
This enables app A to create a socket, pass it to app B, and have
app B accept blame for the traffic performed on that socket.
Also adds helpful public APIs for tagging raw FileDescriptor sockets
instead of making developers go through shady SocketImpl wrappers.
Test: cts-tradefed run commandAndExit cts-dev -m CtsAppSecurityHostTestCases -t android.appsecurity.cts.AppSecurityTests#testAppFailAccessPrivateData
Bug: 63932076
Change-Id: I08925c843974675fc82e4080cec2eaab9ab7cd41
Merged-In: I08925c843974675fc82e4080cec2eaab9ab7cd41
(cherry picked from commit 4e164f9e94)
This is part 2 of 2 of the refcounting refactor for IpSecService
resources.
Switched ManagedResources to use RefcountedResource structure for
managing reference counts and eventual cleanup. Further, resource arrays
and quota management have been aggregated into a UserRecord for better
isolation. UID access checking has been similarly moved into the
UserRecordTracker, and resourceId checking has been rolled into
RefcountedResourceArray's accessor methods.
Bug: 63409385
Test: CTS, all unit tests run on aosp_marlin-eng, new tests added
Change-Id: Iee52dd1c9d2583bb6bfaf65be87569e9d50a5b63
This patch adds (but does not enable the usage of) RefcountedResource
objects to IpSecService, with tests to ensure correct function. This is
patch 1 of a series of patches to refactor the resource management
systems in IpSecService.
RefcountedResource objects allow for management of acyclical dependency
trees, ensuring eventual cleanup when resources are no longer used. This
cleanup may be triggered by binder death or by explicit user action.
Bug: 63409385
Test: New tests written in IpSecServiceRefcountedResourceTest,
explicitly testing the RefcountedResource class
Change-Id: Ib5be7482b2ef5f1c8dec9be68f15e90d8b3aba6d
Throughout the IPsec code (API, system server, netd) we use "reserve"
SPI and "allocate" SPI interchangeably. This renames to use "allocate"
everywhere for self-consistency and consistency with the kernel
(ALLOCSPI).
In javadoc, I am leaving the word "reserve" in several places because it
is still an accurate description of how the method behaves.
Bug: 69128142
Test: TreeHugger should be enough
Change-Id: I8ea603b4612303b0393beef04032671fa53d2106
Improve the Validation of IpSecAlgorithm by
explicitly checking the length in addition to
the truncation length (previously an oversight).
In addition, we now check the lengths during
un-parceling, which will catch someone maliciously
manually building a parcel and passing it, bypassing
the checks in the constructor.
Bug: 68780091
Test: runtest -x IpSecAlgorithmTest.java
Change-Id: I8172762617264d34f47d5144336464510f07a701
This changes only comments and annotations, which results
in minor API changes.
Bug: 63777932
Test: `make update-api` and visually inspect generated HTML.
Change-Id: I7a7a9244eade6cc55aca17a47e0a838cd6f84cc9
The API linter on goog/master noticed that this method is
listed in the public API even though it's protected. The
change is probably related to a signature change from the
internal finalize method which throws a Throwable. Fix the
method in IpSecManager to throw Throwable, which should
fix the current.txt and resolve the lint error.
Bug: 69006767
Test: compilation, make update-api
Change-Id: I173d014baaa505c365b7916fcb52f2a8b4af9373
Added notes that keymat length must include 32 bits of salt.
Bug: 68672051
Test: Frameworks-net unit tests & IpSecManager CTS tests run
Change-Id: I0ae0c5be8a45b2374783b3bd1fa8bf930f15e687
The "roaming" state of a network really belongs on NetworkCapabilities
instead of being published through NetworkInfo.isRoaming(). One major
reason is to support developers creating NetworkRequests for a
non-roaming network.
Watch for any capability changes that network statistics are
interested in (either metered or roaming) and notify it to perform
an update pass; fixes bug where we previously only triggered on
roaming changes.
Fix bug in VPNs where metered/roaming capabilities of underlying
networks weren't being propagated; this was probably preventing
some jobs from running over unmetered networks, and causing other
jobs to run over roaming networks! Also passes along link bandwidth
information from underlying networks, and propegates any changes
to underlying networks.
Fix race condition by reading prevNc inside lock. Utility methods
correctly calculate min/max link bandwidth values.
Test: bit FrameworksNetTests:android.net.,com.android.server.net.,com.android.server.connectivity.,com.android.server.ConnectivityServiceTest
Bug: 68397798, 16207332
Change-Id: I3e1a6544c902bf3a79356b72d3616af1fd2b0f49
Replaced null with empty byte array in createTransportModeTransform
for null auth.getKey, crypt.getKey() and authCrypt.getKey()
Test: Ran CTS test
Bug: 68052730
Change-Id: I5110b1297b4bfbb5766e8ecfd3d64f8110b52945
