Merged from http://go/wvgerrit/160797
Changed CdmResponseTypeToString param from int to CdmResponseType as per sigquit@ review comment in http://go/wvgerrit/160638.
Reorder the switch statements in CdmResponseTypeToString to follow the order in the enum declarations instead of sorting alphabetically. This should make future changes to the enum fields easier.
Test: Netflix, Play TV and Movies, Youtube
Test: adb shell dumpsys android.hardware.drm.IDrmFactory/widevine
Test: ./build_and_run_all_unit_tests.sh
Bug: 258438956
Change-Id: I3f316cff5448e8e8078da7868e3c083e6766bfa1
* changes:
Core CDM: Remove usage info as a run-time type.
Core CDM: Do not store/retrieve USAGE_INFO entries.
Core CDM: Remove usage info API from DeviceFiles.
Core CDM: Removed ability to add secure stop entry.
[ Merge of http://go/wvgerrit/160857 ]
Reboot tests were attempting to clean up usage info records when the
tests were done. However, usage info are no longer supported.
Unfortunately, the clean up code would trigger a test failure if the
usage info records could not be deleted.
Bug: 258469123
Test: OfflineLicenseTest.VariousTests
Change-Id: I17cebd82da1e45ef53853d35dbe314365a1909be
[ Merge of http://go/wvgerrit/159219 ]
Most API functions of DeviceFiles related to usage info files have
been removed. Storing and retrieving usage info files are no longer
required by the CDM. The only function remaining are the ones that
enabled detecting and deleting the remaining usage info files on the
device.
Bug: 242289743
Test: run_x86_64_tests and device_files_unittest
Change-Id: I002202b47141121a0e5adac569e47d8b8bb69b1a
[ Merge of http://go/wvgerrit/159220 ]
Usage entry info elements of type USAGE_INFO are no longer stored by
DeviceFiles. An earlier CL removed the ability to store USAGE_INFO
entries; but now tests have been updated to exclude USAGE_INFO from
their test data.
A new DeviceFiles tests verifies that USAGE_INFO entries are not
retrieved and that the |has_usage_info_entries| flag is set when
they are encountered when loading pre-usage-info-deprecation
usage table info.
Bug: 242289743
Test: run_x86_64_tests and device_files_unittest
Change-Id: Iefbfe2dd2b0304b4075bfc7c1aeffe3b6ba52624
[ Merge of http://go/wvgerrit/159221 ]
This CL modifies how usage entry info is tracked internally by the
CDM at run time. It removes the different "storage types" that
entries represent (license or usage info), and instead it contains
only the information associated with license types. The presences
of a key-set-id allows the UsageTableHeader to determine if the
entry slot is currently being used, or if it can be treated as
unoccupied.
By removing this different type, it completely prevents the CDM and
its tests from using "usage-info" type entries. This required
significant updates to the UsageTableHeader tests.
Additionally, several of the variable names within the usage table
have been simplified and shortened to reflect their new meanings.
Bug: 242289743
Test: run_x86_64_tests and usage_table_header_unittest
Change-Id: I939e479779425550a17a3c9e6c6d1bc6885e493e
[ Merge of http://go/wvgerrit/158721 ]
This CL removes support for secure stop / usage info sessions from the
CDM engine and CDM session. APIs for related to secure stop
operations will return NOT_IMPLEMENTED_ERROR.
New secure stop licenses will be rejected by the CDM when added.
Bug: 242289743
Test: run_x86_64_tests request_license_test
Change-Id: I30cd47e580d63014e001c903382a28238746f6d4
[ Merge of http://go/wvgerrit/158722 ]
The CDM will delete all usage info / secure stop files on the device
when usage table entries labeled as USAGE_INFO are detected when
restoring the usage table. DeviceFiles no longer transfers the
stored USAGE_INFO entry info into the run-time entry info vector for
the table.
This CL makes only minor changes to test data. Significant updates
to both device file and usage table header unittests are required to
remove usage info / secure stop run-time resources.
Bug: 242289743
Test: run_x86_64_tests
Change-Id: I2d1fc0d6a4c994bc10bfce47818f101f95883979
[ Merge of http://go/wvgerrit/158877 ]
The UsageTableHeader's LRU algorithm for determining which entry to
evict when full uses special considerations based on the type of
entry (offline or secure stop).
This CL removes all secure-stop-specific considerations, and instead
treats secure stop's the same as an unused entry. Secure stop entries
will always be selected for removal before offline licenses (expired
or not).
Additionally, LRU table upgrading will ignore secure-stop entries.
This has no effect in practice as DeviceFiles will not load secure
stop usage entries when UsageTableHeader is initialized on a real
file system.
Bug: 242289743
Test: run_x86_64_tests and request_license_test
Change-Id: Ib3f71f191aed94aad62951667426911e4e202068
[ Merge of http://go/wvgerrit/158872 ]
This CL updates the UsageTableHeader API such that it is no longer
possible to add a "usage info" entry. All new entries that are added
are assumed to be offline licenses.
Bug: 242289743
Test: run_x86_64_tests and request_license_test
Change-Id: I09262da0e4301356156d1715b0301c616be2251b
[ Merge of http://go/wvgerrit/160277 ]
OEMCrypto v15 did not require core messages during DRM certificate
provisioning. The CDM's certificate provisioning flow was allowing
for either case (with or without core messages) when provisioning.
Now, devices can safely assume that all provsisioning responses will
be v16 or newer; all requests/response must contain a core message.
Bug: 252670759
Test: run_x86_64_tests and request_license_test
Change-Id: I9f51e07caf642eaf646ef40bdd640b3ccfe2533c
[ Merge of http://go/wvgerrit/160000 ]
OEMCrypto v15 licenses made use of several now-obsolete API functions
of OEMCrypto (mainly LoadKeys and RefreshKeys). All license handled
by the CDM must be v16 or newer. The CDM can now rely on all license
requests/responses containing a core message, using v16 policy timers,
and requires loading using LoadLicense() / LoadRenewal().
Bug: 252670759
Test: run_x86_64_tests and policy_engine_unittest
Change-Id: I3f65a6ec0326b4c89d1919b8911e065079cb90d2
Merge from http://go/wvgerrit/158917
Use go/yamllint as reference and obfuscate portion of output to
run on an online yaml validator.
Sample output: http://go/cl/481370906
Test: Netflix, Play TV and Movies, Youtube
Test: adb shell dumpsys android.hardware.drm.IDrmFactory/widevine
Test: ./build_and_run_all_unit_tests.sh
Bug: 239462891
Change-Id: I1abf1aa50aa25b97b1f6c10995c324d6de04d056
[ CDM CL http://go/wvgerrit/156286 ]
Local test scripts can now run the CDM unit tests without using a test
keybox / test RSA key by setting an environment variable:
DISABLE_TEST_KEYBOX=yes
The default behavior will continue to be that the unit tests will load
a test keybox / test RSA key.
This is to enable testing different provisioning features when the
device uses keyboxes for its ROT. Provisioning methods like 3.0 and
4.0 were never forcibly required to load a test ROT.
This is change is required for the ability to test ECC-enabled
keyboxes.
Bug: 135283522
Test: run_x86_64_tests and build_and_run_all_unit_tests.sh
Change-Id: I2a85be541deec3f60ab8fa82452a4d4cafaed339
[ Merge of http://go/wvgerrit/154575 ]
[ Cherry-pick of http://ag/19498242 ]
There is a rare race condition experienced by some Android devices
where the a new client property set is being added while another is
being removed. The C++ stl library does not provided thread
protection by default.
This CL adds a new mutex for the client property set map which prevents
multiple threads accessing the property sets concurrently.
Bug: 235238226
Test: GtsMediaTestCases on redfin
Change-Id: I32cf11bfb1332295ba1245071102ff0adc35259d
(cherry picked from commit aaa97a5d60)
(cherry picked from commit 6109ec6d66)
[ Cherry-pick of http://ag/19216679 ]
[ Merge of http://go/wvgerrit/155370 ]
There is a rare race condition experienced by some Android devices
where the a new client property set is being added while another is
being removed. The C++ stl library does not provided thread
protection by default.
This CL adds a new mutex for the client property set map which prevents
multiple threads accessing the property sets concurrently.
Bug: 235238226
Test: GtsMediaTestCases on redfin
Change-Id: I32cf11bfb1332295ba1245071102ff0adc35259d
(cherry picked from commit aaa97a5d60)
[ Merge of http://go/wvgerrit/154593 ]
This CL updates the initial time used for reboot tests to be
within the valid range used by OPK. Also, the fake clock is
now synced with every use of the TestHost's clock function.
Also, the OPK's system clock is no longer initialized to 0
because this caused some tests to set initial playback time
to 0. That broke code that assumed a time of 0 mean "never".
I also removed some log spam in the WTPI code.
Bug: 222353528
Bug: 236317198
Change-Id: Ibdbdb2440454b9cf561cd9ec65ca3e40cf5a8d93
(This is a merge of http://go/wvgerrit/153552.)
The correct way to create a FileSystem object for testing CdmEngine
depends on the platform. However, our code has historically just
instantiated a new FileSystem object using the default constructor. This
works fine for Android and has historically worked fine for the CE CDM
tests because they did not implement storage correctly. However, an
upcoming commit will fix the implementation of storage in the CE CDM
TestHost, causing tests that use a default FileSystem constructor to
fail. This patch adds a new platform-defined function that instantiates
per-origin test storage in a platform-correct way. For Android, it
continues to use the default constructor, while for CE CDM, it uses the
TestHost.
Bug: 193060708
Test: x86-64 w/ storage separated
Test: build_and_run_all_unit_tests.sh
Change-Id: I0a9b6a4d3943e053d51d6018e392b7f215f50642
(This is a merge of http://go/wvgerrit/153551.)
On CE CDM, storage is split between global and per-origin storage, and
one type of storage cannot be used to access the other. (Though, until
an upcoming commit lands, the tests will allow it.) On Android, both
types of storage access the same filesystem. This means that code may
run fine on Android but fail on CE CDM.
The Usage Table Header code normally, explicitly accesses the header
file via global storage. However, a few code paths would try to access
it inconsistently via per-origin storage. This patch updates
StoreTable() to always use the global storage, similar to how
RestoreTable() already functions.
Test: x86-64 w/ storage separated
Test: build_and_run_all_unit_tests.sh
Bug: 236400786
Bug: 192297621
Change-Id: Ie84cef43a7ad169ca8ab701d73c087294ee29705
(This is a merge of http://go/wvgerrit/153550.)
On CE CDM, storage is split between global and per-origin storage, and
one type of storage cannot be used to access the other. (Though, until
an upcoming commit lands, the tests will allow it.) On Android, both
types of storage access the same filesystem. This means that code may
run fine on Android but fail on CE CDM.
The OEM Cert in Provisioning 4.0 is a global file that should only
exist once, but it was being accessed through the per-origin storage,
which would result in a separate OEM Cert being provisioned for each app
& origin on CE CDM. This patch changes the Prov 4.0 code to access it
through the global storage, using techniques similar to how the Usage
Table Header code does this.
Test: x86-64 w/ storage separated
Test: build_and_run_all_unit_tests.sh
Bug: 236400627
Change-Id: I301d250fc9543e62949a4d9fdcbdd109bd941384
[ Merge of http://go/wvgerrit/149849 ]
With ECC based DRM cert, the session key is expected to be 32, as
compared to 16 bytes in RSA case. This CL adds supports for 32 bytes
session key.
Bug: 236317198
Test: oemcrypto_test
Change-Id: I657fdd92d17736a23375ddcd457f83efa6ca6d1f
[ Merge of http://go/wvgerrit/151311 ]
Missing a check for provision 4.0.
Note: This CL is merged out of sequence with the CDM repo, resulting
in a slightly different CL.
[ Conflicting CL http://ag/18244016 ]
Test: Native test crypto_session_unittest
Bug: 180530495
Bug: 236317198
Change-Id: I79f753ba43098e1a3422c39b9903f57e5df03af2
[ Merge of http://go/wvgerrit/150810 ]
This CL changes the call "Provision()" in some tests to use
"EnsureProvisioned()". The latter can actually call "Provision()" twice
in case of Prov 4 being used: first call to get oemcert.bin and second
call to get drm cert.bin.
Also added a file clean up during test tear-down.
Test: opk_ta with prov2 and prov4
Bug: 180530495
Bug: 236317198
Change-Id: I37c8708ead7f18e8b4d619bf6a080424e45406ee
[ Merge of http://go/wvgerrit/153290 ]
Currently OEMCrypto_Close will be called for entitled key sessions as
well upon destroying, while only OEMCrypto sessions should be called.
This CL adds a session type to the oemcrypto adaptor.
Bug: 232225911
Bug: 236317198
Test: request_license_test
Change-Id: I323fff80139ce949f801cf0df5aa2bd3171dfb51
[ Merge of http://go/wvgerrit/152650 ]
There are two LOGEs in device_files.cpp that would frequently trigger
during testing despite the condition they were logging being totally
normal. The files were not present because the device had never had to
write them yet. This patch downgrades these bengin errors to warnings
to reduce log spam.
Test: build_and_run_all_unit_tests.sh
Bug: 236317198
Change-Id: I56f9d477f8a6b26635f0eb4a6b7c636e866381d3
[ Merge of http://go/wvgerrit/148729 ]
A recent merge from google3 of the ODK code broke builds on the
oemcrypto-v17 branch because it depended on a license_protocol.proto
change that had not been brought over yet. This patch brings over the
missing line to fix the build.
Test: build_and_run_all_unit_tests.sh
Bug: 236317198
Change-Id: Ia8133696f3156c16ef4bd1382358be5da6159820
(This is a merge of http://go/wvgerrit/152969.)
C++ makes absolutely no guarantees about the order of initialization of
global variables in different compilation units. The class-scope static
WvCdmTestBase::default_config_ in test_base.cpp invokes the
ConfigTestEnv constructor on creation, which depends on the prior
initialization of several file-scope static variables in
config_test_env.cpp. Since those are different compilation units, there
is no guarantee that they will initialize in the correct order to avoid
referencing uninitialized memory. This is one of the reasons Google
Style really encourages people not to have global-scope variables with
complex types.
As it happens, on all our internal platforms, these files get linked in
such a way that the variables get initialized in the right order and
there is no crash. But that's not guaranteed, and some partners have
reported crashes here. In at least one case, the "right" linker order
was platform-dependent, and the partner ended up having to maintain
separate linker orders for separate platforms.
This patch defers default_config_ initialization until
WvCdmTestBase::Initialize() is called. By that time, all static
variables will be initialized, so it will be safe to reference them.
Bug: 173252165
Test: x86-64
Test: build_and_run_all_unit_tests.sh
Change-Id: If31128a999c7d6945f47293ca57f08e43d8274de
[ Merge of http://go/wvgerrit/153489 ]
OEMCrypto does not provide an API for retrieving the system ID when
the TEE uses a built-in DRM certificate (provisioning 1.0). New OEMs
and Android devices do not use prov 1.0; however, the Zimperium CDM
(at least the tests) use a built-in certificate and are failing
certain tests because of the missing system ID. To address this
failure; the CDM SystemIdExtractor has been updated to return a null
system ID.
Bug: 235879962
Test: system_id_extractor_unittest
Change-Id: Ib4c2bd75a7825967b0aa9e31e144184ae18fe8fb
(This is a merge of http://go/wvgerrit/151930.)
While grepping the code to respond to some CR feedback, I noticed a few
places where we had sprinkled some unnecessary "const" specifiers
amongst constexpr declarations. This patch cleans them up. There should
be little semantic difference in the code after this patch, as it only
removes specifiers that were redundant. The only exception is where
"constexpr const char* X" was converted to "constexpr char X[]", which
has slightly different semantics in edge cases we don't use.
Test: x86-64
Bug: 231439638
Change-Id: I0b33777f8d3b718a3410f6d802c51b1220508d34
(This is a merge of http://go/wvgerrit/151891.)
A previous patch changed how we skip padding when extracting keys from
key containers in license.cpp. Unfortunately, this broke generic
signing when an ODK core message is not in use:
1) "Content" keys for signing are 32 bytes long, but content keys were
assumed to be 16 bytes long.
2) When an ODK core message IS in use, the result of the extraction in
license.cpp is ignored.
The only way to know the correct length of a content key container in
License Protocol 2.1 is to leverage the knowledge that it will always be
padded by exactly 16 bytes. This will have to change if we ever
implement support for License Protocol 2.2, as all key containers are
unpadded in that version.
Bug: 231439638
Bug: 114159862
Test: oemcrypto_dynamic_v15
Change-Id: I1d6c24b3a922247b970fd1517c6f23aded570adf
