Tradefed needs these tests to be in the /DATA directory.
Test: cd vendor/widevine && ./build_and_run_all_unit_tests
All tests should build and pass.
make tests
unzip android-tests.zip and verify all tests located in DATA/bin
Bug: 62055647
Change-Id: I35925e29558561c4726bb2249499bfee4e54cf45
[ Merge of http://go/wvgerrit/28265 ]
A vendor specific error (10008) in response to OEMCrypto_LoadKeys
indicates that usage table corruption has occurred and that
the only way to recover is to regenerate usage tables.
Recreating usage tables will result in loss of offline licenses
and usage information. To make the app aware that this information
will be lost, a provisioning exception is generated when this error
is detected. The app can then choose to reprovision and in turn
delete and recreate usage tables.
A new OEMCrypto error has been added whose use has been reserved.
Rather than correct OEMCrypto behaviour to use the new error code,
we choose to handle this within the CDM. The fix can then be ported
to prior android releases. Also this error will not be generated
with OEMCrypto V13+.
b/33817629
Test: WV Unit, integration and GTS tests.
Change-Id: I936fc234d101b6a92d86f5735d035d19ddcf19e3
[ Merge of http://go/wvgerrit/28261 ]
Licenses (offline, secure stops) that contain provider session tokens
are handled securely using usage tables. A recent fix did not correctly
handle offline licenses that do not contain a provider session token and
are not handled by the TEE.
b/62340248
Test: WV Unit/integration tests, GtsMediaTestCases
Change-Id: Ia1331fea9deff44dd1d93219b37f5bea4b8ee168
[ Merge of http://go/wvgerrit/26421 ]
* Corrects usage_table_header lifetime management. Earlier the
UsageTableHeader class was a singleton tied to the CdmEngine lifetime.
With SPOIDs there might be multiple concurrent CdmEngine objects.
The UsageTableHeader class is now associated with OEMCrypto
lifetime. There are two UsageTableHeader objects one for each L1 and L3.
These get allocated/deallocated on OEMCrypto Initialization/Termination
respectively.
* UsageTableHeader requires OEMCrypto, file read/writes and
metric gathering to perform its required functionality. Because of the
lifetime changes, CryptoSession, DeviceFiles and MetricsGroup objects
need to passed to the methods rather than at Creation time.
* Miscellaneous fixes, when moving or deleteing entries.
* Adds usage_table_header_unittests.
* Addresses failures with request_license_test with secure stop in L3.
b/36858906
b/36855557
b/36048120
b/38341136
b/37100505
b/35946047
Test: Verified by unit and integration tests. Added new
usage_table_header_unittests
Change-Id: I20e396ab2c0afbd14372dd93b969e5b0f1ccd291
[ Merge of https://go/wvgerrit/27860 ]
The android version was updated from "O" to "8.0.0"
b/62201928
Test: Ran wv unit and integration tests.
Change-Id: I57525af2ab3bdfa0e77ad8f0efbff26cf8ff5c32
Merge from Widevine repo of http://go/wvgerrit/27462
This updates the documentation for OEMCrypto:
1) Clarification on values for the key control block verification
field. Unit tests for this change are already in the code.
2) Clarification on how Offline License Release messages are signed.
And for the Android supplement:
3) liboemcrypto.so may be installed in several places.
b/38422219
Testing: docs only change. no testing needed.
Change-Id: Iab3f5b585a41c80c5d7aa5f00aabce2de2df3a79
Merge from Widevine repo of http://go/wvgerrit/27461
In order to sign a license release message, the mac keys from the
usage entry should be used whenever keys have not been loaded.
This CL updates the reference code, the unit tests, and the level 3
oemcrypto.
b/38203566
Test: unit tests passing on bullhead.
Change-Id: Ic71fee4b4b7b45801548ab80fbbbf8f4ccab3e6e
This reverts commit 006506278a.
It needs to be reverted to fix b/38343960 on marlin.
This change is not needed since the root cause of the
resource leak was addressed in ag/2226738.
Change-Id: I93b64d39103d574b7dcde6b1477839a50b64aefd
Merge from Widevine repo of http://go/wvgerrit/27182
This function is not defined for v13, and should not be called by the
adapter.
b/38203780
Test: Ran unit tests specified in bug, on sailfish.
Change-Id: I38b1cc8493dd22da724f415f28d94f47d6d9c942
These tests verify whether a L3 device respects the HDCP policy set in the
license request. L3 device should only play the clear lead, and stop at
encrypted content. MeidaDrm should throw an ERROR_INSUFFICIENT_OUTPUT_PROTECTION
exception. This is because L3 device always returns current HDCP connection
status as HDCP_NONE.
This CL modifies the policy engine CanDecryptContent method to return
CdmResponseType instead of a boolean, so the app can generate the correct
error response if HDCP constriants are not met.
Test: GTS tests
ANDROID_BUILD_TOP= ./android-gts/tooadefed run gts -m GtsMediaTestCases --test
com.google.android.media.gts.WidevineDashPolicyTests#testL3PlayHDCPV*Required
Test: unit tests
adb shell /data/app/policy_engine_unittest
adb shell /data/app/policy_engine_constraints_unittest
Test: Play Movies
bug: 34258607
Change-Id: I11fc9da1e077e18e38f34159daae9d8ebcd948b6
[ Merge from http://go/wvgerrit/27261/ ]
Sessions created to release keys are periodically cleaned up if the
key release operation does not complete within a specific
amount of time. If other sessions are open, they will be released
through the timer thread. This would result in deadlock as a mutex
was taken twice.
Test: Verified by cdm_extended_duration_test
(AutomatedOfflineSessionReleaseOnOpenSession and
AutomatedOfflineSessionReleaseOnTimerEvent tests)
b/37546078
Change-Id: I7d45f939bdce77e5db461a401364da4f42c1c034
A strong pointer cycle was preventing the drm plugin
from being released. This change converts the listener
to a weak pointer to break the cycle.
bug:36408047
test: manual testing to verify leaks are fixed
Change-Id: I2e2af392c2b263b7b0943d61dcaee5e94089ce27
Merge from Widevine repo of http://go/wvgerrit/26943
b/37987506
Test: built and run unit tests on sailfish -- no new failures. Ran
Netflix on sailfish.
Change-Id: I96f8bb7a5c060fefd116b165406ce6bfad29f59d
Merge from Widevine repo of http://go/wvgerrit/26940
b/38014227
b/38016711
Test: built sdk_phone_mips
Change-Id: Ibcdbf49f96e976f51c38db25a944268a86a1671e
Merge from Widevine repo of http://go/wvgerrit/26780
and http://go/wvgerrit/26740
Previously, in oemcrypto level 3, a session's current key's duration
was not updated until the next call to SelectKey. This caused
problems with license that only used one key. This CL fixes that.
arm64/libwvlevel3.a Level3 Library 7283 May 2 2017 13:16:21
arm/libwvlevel3.a Level3 Library 4445 May 2 2017 11:49:34
x86_64/libwvlevel3.a Level3 Library 7284 May 2 2017 12:09:21
x86/libwvlevel3.a Level3 Library 4464 May 2 2017 11:53:46
Test: unit tests run on emulator.
b/37481239
b/37523523
Change-Id: Ife90a3358b6620c8fb81324ec2331d3775a38191
Merge from Widevine repo of http://go/wvgerrit/26202
This is a workaround to fix a broken mips build.
Test: Built for sdk_phone_mips.
b/37523523
Change-Id: I56ce65919303113f712d422d33e335b594292d42
[ Merge of http://go/wvgerrit/26201 ]
Race conditions arose when a session was closed while data was
still queued for decryption in MediaCodec buffers. If a session
is closed while data is still queued for decryption, subsequent
decryption requests will be rejected with a CryptoException
ERROR_SESSION_NOT_OPENED.
Test: Verified by wv unit/integration test and
WvCdmExtendedDurationTest.DecryptionCloseSessionConcurrencyTest
b/36747801
Change-Id: I044d1d6b9fc886a1c353d20b9c6365319aa71e80
(This is a merge of http://go/wvgerrit/26221)
We have agreed on API 26 for O, so these comments can be removed without
making any further changes.
Bug: 34548395
Test: This is a comment-only change. But I verified it builds.
Change-Id: I490befba4f47902a84b60e562a7a5eb162a9f3be
(This is a merge of wvgerrit/25583)
Devices that use Provisioning 3.0 did not have a Provisioning-Unique ID
defined. Attempting to retrieve it would result in an error.
Devices that use SPOIDs with keyboxes would expose the keybox's real
Provisioning-Unique ID when asked. This is a security flaw.
To solve both cases, an alternative Provisioning-Unique ID is used,
consisting of the Device-Unique ID bitwise-inverted.
Bug: 36065223
Test: run_all_unit_tests.sh
Change-Id: I32512a3e11403e679939187e156904a57a9e24ef
(This is a merge of wvgerrit/25582)
Provisioning 3.0 devices that do not use SPOIDs have been returning
their full OEM Public Certificate as their device ID. While this is not
a security concern, (it is a PUBLIC cert) the cert is many times larger
than applications are likely expecting. (several kilobytes vs. just a
few bytes) This patch hashes the OEM Public Certificate to produce a
smaller value, but only when it is being provided out of the CDM to a
caller.
Bug: 34716264
Test: run_all_unit_tests.sh
Change-Id: Ib82cf7a174a8bf02ff606edd0394ada13842224c
(This is a merge of http://go/wvgerrit/25900 and http://go/wvgerrit/25920)
This commit contains the latest Haystack prebuilt binaries based on
the latest OEMCrypto code in the Widevine tree. This is to pick up the
change "Report error OEMCrypto_ERROR_ENTRY_IN_USE", which fixes
several failing unit tests. This CL also adjusts the L3 library so
that it stores its data in the directory specified by the layer above.
level3/x86_64/libwvlevel3.a Level3 Library 7284 Apr 13 2017 16:58:07
level3/arm/libwvlevel3.a Level3 Library 4445 Apr 13 2017 16:48:19
level3/mips64/libwvlevel3.a Level3 Library 7285 Apr 13 2017 23:52:51
level3/mips/libwvlevel3.a Level3 Library 4465 Apr 13 2017 23:16:02
level3/arm64/libwvlevel3.a Level3 Library 7283 Apr 13 2017 17:05:12
level3/x86/libwvlevel3.a Level3 Library 4464 Apr 13 2017 16:36:25
Bug: 36656178
Bug: 37210088
Test: build_and_run_all_unit_tests.sh and GTS tests
Change-Id: Ia5b272cd945f0c9f2fe37ecd0f3ccd2f2ff398bc
[ Merge of http://go/wvgerrit/25987 ]
Test: Added Cdm/WvCdmSessionSharingNoKeyTest to request_license_test.
Verified by wv unit and integration tests.
b/28182987
Change-Id: I9039bc1fca83e2c1044c3af2f47f31dc4fb5c910
[ Merge of http://go/wvgerrit/25983 ]
Earlier versions of android returned CryptoException with
error code ERROR_NO_KEY, when a decrypt call was received before keys were
loaded. Changes to O resulted in ERROR_SESSION_NOT_OPENED being returned
instead. This CL reverts the behaviour.
Also a change to correct CDM error code numbering in comments.
Test: Verified by unit and integration tests
b/37219830
Change-Id: I43758cd29cf9d1945f878ac352a5f26538b48cdb
(This is a merge of http://go/wvgerrit/25581)
To prevent dead DRM Plugins from being instantiated when there are no
sessions available, the time at which the device ID is queried has been
moved from instantiation-time to SPOID-calculation-time.
SPOIDs can now fail to be generated. Which means anything that depends
on a SPOID may fail because of this. However, this is a more actionable
problem for apps than having them receive a dead or unusable DRM Plugin
object.
Bug: 36660726
Test: libwvdrmdrmplugin_hidl_test
Change-Id: Ice6a8eabfee8d48bf2af02e2e7169aa95af9f2e4
(This is a merge of http://go/wvgerrit/25580)
While writing fixes for b/36660726, b/34716264, and b/36065223, it
became clear that having the logic that checks whether the device
supports SPOIDs embedded inside WVDrmPlugin was complicating its code
and inhibiting testing of the class. By moving this check into the code
that instantiates WVDrmPlugin, the result of the calculation can be
independently tested while the tests for WVDrmPlugin can put it in
whatever state they need for the sake of unit testing.
As a consequence of this, the check on retrieving the "deviceUniqueId"
byte array property, which was removed when SPOIDs were implemented, can
be reinstated.
Bug: 36660726
Bug: 34716264
Bug: 36065223
Test: libwvdrmdrmplugin_hidl_test & libwvdrmengine_hidl_test
Change-Id: I961d2ee42bbdc42f0c324e36d9a74ac92205a437
