[ Merge from http://go/wvgerrit/71443 ]
The assumption that the metrics will always outlive the CdmSession
instance appears not to always hold (at least in a non-android
multi-threaded solution). The shared_ptr ensures that the metrics
are available even in these rare race conditions.
BUG: http://b/123321465
Test: CDM unit tests. Also http://go/wvgerrit/71264 parallel tests.
Change-Id: Iaa6a8f6c0fdc46a911789759d6e1228d849aa237
[ Merge of http://go/wvgerrit/71907 ]
The client token needed to be enabled in the license request.
Bug: 123369846
Bug: 123370099
Test: WV unit/integration tests
Change-Id: I4d3e944b1d79010977c119291594878c406b00c5
[ Merge from http://go/wvgerrit/69105 ]
This adds a metric collecting decorator class around cdm engine. This
implementation uses a templated decorator. The decorator enables:
1) Wrapping the CDM Engine methods to capture timing and error
information.
2) Allows use of a mock CDM Engine for testing.
Test: Unit tests. GPlay manual testing and GTS tests.
BUG: http://b/64724336
Change-Id: I5e4a0f552974fab1939bc7ab02719a1f5849cf3f
[ Merge of http://go/wvgerrit/71326 ]
Nonce flood, frame size, session and system invalidation errors
will now bubble up to the app. OEMCrypto v15 returns
OEMCrypto_ERROR_BUFFER_TOO_LARGE, OEMCrypto_ERROR_SESSION_LOST_STATE,
OEMCrypto_ERROR_SYSTEM_INVALIDATED and a variety of nonce errors.
These will be reported to HIDL as OUTPUT_TOO_LARGE_ERROR,
ERROR_DRM_SESSION_LOST_STATE, ERROR_DRM_INVALID_STATE and
ERROR_DRM_RESOURCE_CONTENTION.
Bug: 120572706
Test: Unit/Integration tests
Change-Id: Ida177300046327ce81592a273028ef6c3a0d9fd9
This reverts commit 408509efdc.
Reason for revert: just kidding. We want to run the tests on
old devices, too.
Change-Id: I920fa825380638122f638e057bcfbe98b5f7b301
(This is a merge of http://go/wvgerrit/70666)
We need a reader-writer lock implementation in order to make use of the
new threading guarantees in OEMCrypto v15. However, we do not have
access to an STL reader-writer lock due to only being on C++11. This
patch adds a home-grown reader-writer lock, as well as tests to verify
that its behavior is sound.
Bug: 70889998
Bug: 118584039
Test: CE CDM Unit Tests
Test: Android Unit Tests
Change-Id: Iaddcefb50e72452fbd27d04879eacf775484e675
Merged from http://go/wvgerrit/69723.
The new APIs are getOfflineLicenseIds, getOfflineLicenseState and
removeOfflineLicense. These methods are currently stubbed out in
Widevine hidl service. This CL completes the implementation.
Test: unit tests - libwvdrmdrmplugin_hidl_test
Test: GTS
--test com.google.android.media.gts.MediaDrmTest#testWidevineApi29
bug: 117570686
Change-Id: I96ffb75f453e36e931effefd3664b5faa8d69d30
Merge from Widevine repo of http://go/wvgerrit/71323
When running multithreaded tests, we need to make sure there are no other
OEMCrypto calls between OEMCrypto_Initialize and the installation of the test
keybox.
bug: 123319961
test: ran juce's threading test
Change-Id: Ibd8951c764f6538da4afa9ad146a6a72064eb13a
(This is a merge of http://go/wvgerrit/70343)
Instead of using the POSIX-specific getopt(), this uses a manual
string comparison so it works on Windows.
Test: Android Unit Tests
Bug: 122953649
Change-Id: If009f8b3bfc8581d252da0f5ed55293c46d70bc6
(This is a merge of http://go/wvgerrit/71330)
The Service Certificate unit tests actually relied on the ability to
call Properties::Init() multiple times to clear previous mutable state.
Unfortunately, they didn't check the return code that could have told
them their mutable state wasn't being cleared and instead proceeded to
use a pointer which — depending on compiler — could be totally valid and
allow the test to pass or could be invalid and cause a segfault. You can
read the bug for a fuller explanation of the mechanics.
The fix is twofold. First, the tests will now assert out if insertion
into the property set map fails, preventing segfaults. Second, a helper
has been added to Properties that allows tests interested in
re-initializing Properties to do so. The default behavior for most tests
remains the same: Properties can only be initialized once and subsequent
calls to Properties::Init() are ignored.
This patch also fixed a few formatting issues I noticed.
Bug: 123099779
Test: Jenkins Unit Tests w/ GCC
Test: CE CDM Unit Tests w/ GCC & Clang
Change-Id: Ifd29f3ddf5cff934933cf47b92ecd12ab0a4a938
Bug: b/121214641
Merge of http://go/wvgerrit/70768
Tests: usage_table_header_unittests (with and without hardcoded values)
DeleteEntry fails if the last entry is in use, since it cannot shrink
the table any further via swapping. The LoadEntry_LoadUsageEntry...
tests test to see that after deleting an entry x amount of times, the
usage table can load an entry. If the entry we want to load is within x
positions of the last entry of the usage table, the method cannot
succeed for the aforementioned reason.
Change-Id: Iccdaf538cb4bcbd34855f36c71aceaeb93b4a770
Merge from Widevine repo of http://go/wvgerrit/69563
This CL also disables unit tests that depend on the old test keybox when run
against OEMCrypto v13 or earlier.
Unit tests that require a particular API level have been
removed. This requirement for Android is tested in GTS tests.
Bug: 119316243
Test: unit tests
Change-Id: I9a13dbaee1693ecea906fd0184a8e2590dc5beed
[ Merge of http://go/wvgerrit/70386 ]
This allows error codes from device files to be added as sub-errors when
errors such as GET_LICENSE_ERROR are encountered.
Bug: 112357085
Bug: 115382201
Test: WV unit/integration tests
Change-Id: I505a87086ce584efc7e482984c0f132ac5329e16
(This is a merge of http://go/wvgerrit/70667)
Request ID Index generation has historically worked by incrementing a
shared variable in one place and reading it in another place and
trusting the fact that CdmLicense calls these operations in a certain
order and only once per session to give each session a unique value.
This patch cleans this up a bit, having each session store the current
Request ID Index at the same time as it stores its Request ID Base. This
guarantees that each CryptoSession will receive a unique but stable
combination of Base and ID rather than relying on the calling pattern.
Since all this generation happens during the same function, the full
Request ID can be generated up-front and stored, making
GenerateRequestId() no longer necessary.
This patch also simplifies the threading story around this shared state
by using a std::atomic<uint64_t>. Bringing the code that interacts with
the shared state together into one place and replacing it with atomic
operations will simplify locking around this code when CryptoSession
locking is revamped in a future patch.
Bug: 70889998
Bug: 118584039
Test: CE CDM Unit Tests
Test: Android Unit Tests
Change-Id: I12d2f6501f872f1973e5a9af5125ca03f23e5a56
[ Merge of http://go/wvgerrit/70543 ]
RestoreOfflineLicense and RestoreLicenseForRelease now return
CdmResponseType errors rather than a boolean. These error codes
can now be used when gathering metrics.
Bug: 115517916
Test: WV unit/integration tests
Change-Id: If4784d2cdd9825948c5dec31d3e60058ea06b61d
(This is a merge of http://go/wvgerrit/70303)
This adds a platform.h file to abstract some of the differences
between Windows and POSIX platforms. This includes ntohl, setenv,
and ssize_t.
Bug: 122953649
Test: Android Unit Tests
Change-Id: I3235f3f284b53d24d7365ff3f4a06dcd9b403697
(This is a merge of http://go/wvgerrit/70384)
It turns out the reason the binary provisioning unit test was failing is
because the test base class was setting the property to turn on binary
provisioning before calling Init(), however all current Init()
implementations overwrite the value of that field. As such, the tests
weren't actually using binary provisioning. With that fixed, everything
passes; the binary provisioning flow doesn't appear to actually be
broken.
This is the exact same commit as was previously reverted,
cf5464d7a2fbecd1938ae0700199145b7b61c3c3. However, a pending patch fixes
the multiple runs of Properties::Init() that lead to the inconsistent
failures of the previous patch. (See http://go/wvgerrit/70383)
Bug: 112046733
Test: CE CDM Unit Tests
Test: Android Unit Tests
Change-Id: I79fc2c8d4d24505b46cab0e21eef85a37d66748d
Bug: b/121214641
Merge of http://go/wvgerrit/69703
Test: Android + Linux unit tests
LoadEntry attempts to handle INSUFFICIENT_RESOURCES by deleting an entry
and retrying, but it's possible that the randomly-generated number of
the entry to be deleted might match the entry we want to load. In this
case, we have wasted a retry, since the code just continues on to the
next iteration. This is changed to generate a number different from the
entry to load. Furthermore, if the number of usage entries is 1, we
break since there are no more entries to delete besides the one we want
to load. The code is also changed to call srand in the creation of the
usage_table_header, since without it, rand() would produce the same
values, and similarly, our random generation is changed to use a simple
mod. Tests are modified to reflect these changes.
Change-Id: I95e125b8adbd85d0189f9d40ca15f3fe69e6d6b9
(The matching reversion in the Widevine Repository is
http://go/wvgerrit/69923)
This reverts commit c207ef8ea6.
Reason for revert: Widevine Buildbot failures
Change-Id: Ied9e36f9eaa94bb7e9cd13b8dff7a2cbdca07a07
(This is a merge of http://go/wvgerrit/69843)
It turns out the reason the binary provisioning unit test was failing is
because the test base class was setting the property to turn on binary
provisioning before calling Init(), however all current Init()
implementations overwrite the value of that field. As such, the tests
weren't actually using binary provisioning. With that fixed, everything
passes; the binary provisioning flow doesn't appear to actually be
broken.
Bug: 112046733
Test: CE CDM Unit Tests
Test: Android Unit Tests
Change-Id: I7413f7fb2227e596fb610d6ddc5b95cda2f406b8
[ Merge of http://go/wvgerrit/68083 ]
Add ability to query decrypt hash support, set a hash computed over a frame
and retrieve the last error at a later point.
Bug: 34080802
Test: WV unit/integration tests. New tests added to cdm_engine_test,
libwvdrmdrmplugin_hidl_test and request_license_test.
Change-Id: I7548c8798c873a6af3e1cfc0df57c117e1e474a6
Merge from: http://go/wvgerrit/69211
Cleanup unaligned code while working on b/117570686 and remove
a duplicate break statement.
Test: unit tests
./run_all_unit_tests.sh
Bug: 121971420
Change-Id: I0ca1dc0fe54e7fe3a87aa7c8dac7ef9da04e90b2
Import of http://go/wvgerrit/68188
This adds an attribute to metrics indicating if the license was online
or offline.
Also, added a unit test for CdmEngineMetricsImpl.
Test: Unit tests. GPlay manual. GTS tests.
Bug: 115523917
Change-Id: Id315c643048914a2c51904451f9665987bc87eb7
[ Merge of http://go/wvgerrit/67324 ]
This introduces the ability to query resource rating tier information
through the plugin and CDM. Resource rating tiers are also
sent in the client identification portion of the license request.
Bug: 117112392
Test: WV unit/integration tests
Change-Id: I68ac6dfc4362f61150af822bd526e346b5cc4bf7
[ Merge of http://go/wvgerrit/63682 and http://go/wvgerrit/4977904 ]
Enforce OEMCrypto insufficient resources error reporting in CDM tests.
In addition, when CreateNew/LoadUsageEntry fails with OEMCrypto
insufficient resources error, delete a random usage entry
and retry. The current behavior evicts low index entries
from the usage table, which was a crude attempt to emulate a LRU.
This was deficient as, on occasion, it will result in the deletion
of a recently added usage entry.
Bug: 111260263
Bug: 113828866
Bug: 120433165
Test: Widevine OEMCrypto tests, integration tests. GtsMediaTestCases.
Play movies and Netflix playback tests.
GtsMediaTestCases MediaDrmTest#testUsageTableCapacity
Change-Id: I63340f76d1e2af3c6834b98ad816e11eea18fc7f
(This is a merge of http://go/wvgerrit/66625)
Google C++ Style dictates that methods which override base class or
interface methods should be declared "override" but not "virtual". Since
our codebase has not had access to "override" until now, many of our
classes do not follow this rule. I've updated as many places as I could
find to follow Google C++ Style, which should hopefully help us catch
errors better in the future.
Bug: 111851141
Test: CE CDM Unit Tests
Test: Android Unit Tests
Change-Id: Ic23e2e482e967256da306791532b5fec7b81b2f2
(This is a merge of http://go/wvgerrit/66643)
The sub-license feature has been removed from the server and packager.
So that we do not have to continue maintaining the code that supports
this feature that never shipped, I am removing it from the CDM as well.
Bug: 113165466
Test: CE CDM Unit Tests
Test: Android Unit Tests
Change-Id: I5d25844b161e74aa19adf19a29c56e4881aa7304
Bug: b/119276649
Merge from: http://go/wvgerrit/66367
Test: Android, CE CDM, Linux unit tests
The FileSystem interface as it exists expects an Open for a file and
then a Close when finished. However, the Close doesn't delete the file
itself and depending on the platform, the underlying impl_ as well,
leading to a memory leak. To fix this leak as well as harden against
future memory issues, this change refactors the interface to shift away
from raw pointers and towards smart pointers.
Change-Id: I7a7132ea95cd3775796a540f510b698f4f27dd24
[ Merge of http://go/wvgerrit/65442 ]
This also requires the removal of sub session related code as references
were removed from the proto.
Bug: 119077124
Test: WV unit/integration tests
Change-Id: Ida1a591afc267ec97344e5bba00bbf401887a202
Merge from master branch of Widevine repo of http://go/wvgerrit/66073
Merge from oemcrypto-v15 branch of Widevine repo of http://go/wvgerrit/64083
As part of the update to v15, LoadKeys, RefreshKeys, and
LoadEntitledContentKeys should all use offsets and lengths into the
message rather than a pointer for its parameters. The CDM, tests,
adapters, and OEMCrypto implementations are changed to reflect this.
Test: tested as part of http://go/ag/5501993
Bug: 115874964
Change-Id: I981fa322dec7c565066fd163ca5775dbff71fccf
Merge from master branch of Widevine repo of http://go/wvgerrit/66066
Merge from oemcrypto-v15 branch of Widevine repo of http://go/wvgerrit/63628
The error code OEMCrypto_KEY_NOT_LOADED is redundant with
OEMCrypto_ERROR_NO_CONTENT_KEY and OEMCrypto_KEY_NOT_ENTITLED. The
function LoadEntitledContentKey should return KEY_NOT_ENTITLED if it
does not find the corresponding entitlement key in its key table. All
other functions that do not find a key id in the key table should
return OEMCrypto_ERROR_NO_CONTENT_KEY. This includes QueryKeyControl,
SelectKey, and RefreshKeys.
Test: unit tests
Test: tested as part of http://go/ag/5501993
Bug: 115574797
Change-Id: Ida2111f32e331b99f3f0c77fa404a42654d0870c
