The function OEMCrypto_LoadKeys should accept a null pointer for the
enc_mac_keys pointer. This indicates that the mac_keys are not being
updated. This CL just updates the documentation to clarify this behavior.
bug: 9549308
Change-Id: I1073e72f9c8d0fa712ad02e4c2d517202b3defab
When the video packets are clear, they are being
combined into a full frame between MediaExtractor or
DecryptCallback (where packets are never > 64k) and
MediaCrypto/DecryptVideo (where we are seeing clear
packets > 64k).
The max block size to be passed to DecryptVideo is
specified as 64K, handling of packets larger than 64K
is undefined and OEM implementation-dependent. In the
Nexus 10 case, it generates a SEG. fault.
Solution:
Add mClearSizes vector to keep track of each clear
packet size for android_media_MediaExtractor_getSampleCryptoInfo()
to process.
In android_media_MediaExtractor_getSampleCryptoInfo(), if it
does not see kKeyEncryptedSizes meta data, it will not
process the clear vector that contains the actual packet
sizes. Add a kKeyEncryptedSizes meta data that contains a zero
length size packet to trick the JNI code to process the
clear packets.
related-to-bug: 9261447
Change-Id: Ib0b655a95e099856babaf649f4a95fc7f9c17d41
Only the first 64K of the movie is pulled when sniffing,
if the widevine version metadata is not present in the
sniff buffer then a "format unsupported" error would occur.
This change increases the sniff buffer size to 128K.
bug: 9351294
Change-Id: If162cbea6915bf2b70122afd30e556e9206e8e43
Adds a property that allows applications to get the provisioning-unique serial
number.
Bug: 9175567
Also fixes some missing mutexes that were causing intermittent failures in
calls to OEMCrypto due to concurrency issues.
Bug: 9175583
Merge of https://widevine-internal-review.googlesource.com/#/c/5831/
from the Widevine CDM repository
Change-Id: I1d7e3ca9f3b06da345022f5f0d64e0c17a5cedca
The field provisioning code generates a randum number to use as the
device id, and then restricts the data to alphanumeric characters.
Previously, it could have also included a 0 byte in the data. This CL
corrects that.
bug: 9073146
Merge of https://widevine-internal-review.googlesource.com/#/c/5740/
from the Widevine CDM repository
Change-Id: Iaf3e9e733f7c66c19d4168178a8e25ee0ba7e936
The company_name was hardcoded in the CDM as "Google" for all devices.
On Android, it needs to come from the ro.product.manufacturer system
property.
bug: 9074091
This is a merge of https://widevine-internal-review.googlesource.com/#/c/5730/
from the Widevine CDM repository.
Change-Id: Ia3ae82abf350c32ba8b4d05b59e95361927dea40
Signing and encryption keys are not correctly setup in OEMCrypto, when
an offline license is restored, before generating a key release message.
This results in key release failures. Playing back the license response
causes keys to be derived and allows the key release message to be constructed.
b/9016545
Merge of https://widevine-internal-review.googlesource.com/#/c/5682/
from the Widevine CDM repository
Change-Id: Ica9f13acc7c87e3125fa706f3a56e95b77a14a3c
Since CryptoSession::RewrapDeviceRSAKey takes a const std::string& nonce, the calling code
implicitly constructa temporary std::string from a character array with undefined termination
and sometimes bogus value when nonce.data() is passed in.
Bug: 9016828
Merge of https://widevine-internal-review.googlesource.com/#/c/5690/
from the Widevine CDM repository
Change-Id: Ia3beaa5f1a96924a8cc53a719f25af52217b8cc5
Because we do not want to accidentally install a test keybox on a
production device, most of the oemcrypto unit tests are being disabled
by default.
If you wish to run these tests, you can override this choice, by
running:
adb shell /system/bin/oemcrypto_test --gtest_also_run_disabled_tests
This change prompts for and requires positive confirmation before
running the disabled tests on a device that has a non-test keybox
already installed.
Bug: 8907626
Merge of https://widevine-internal-review.googlesource.com/#/c/5531/
from the Widevine CDM repository
Change-Id: Ib8e3605129ebf0861b4af15d04676f7a06cc5b78
There is an ambiguity in the use of the signature length in
OEMCrypto_GenerateRSASignature. If the pointer to the vector is null,
the function should set the length parameter to be the size of the
buffer needed.
This ambiguity has been clarified in the documentation: the
length returned should be the exact length of the buffer. It
will be corrected in a future release of Qualcomm's L1. However,
we are putting this change in both as a stop-gap and as a
belt-and-suspenders fix.
Bug: 8878371
Merge of https://widevine-internal-review.googlesource.com/#/c/5494/
from the Widevine CDM repository
Change-Id: I7574874884ca10da68d15674c971a565d015767d
In the android media DRM api test, provisioning is indicated, when a
generate license request command is issued after a switch between L1 and
L3. This is as expected as oemcrypto is unable to decrypt
the key wrapped earlier (bad padding). Subsequent provisioning request
and storage of wrapped keys complete successfully. If the same session
is used to reissue a generate license request command, the wrapped keys
and cert used are those present in memory from the initial retrival,
rather than rereading the new ones from persistent storage.
This results in a cycle of successful provisioning attempts followed
by generation of license requests commands which return a provisioning needed
error.
A change has been added to reinitialize the session and reload the
wrapped keys.
b/8878324
Merge of https://widevine-internal-review.googlesource.com/#/c/5600/
from the Widevine CDM repository.
Change-Id: Iaf47d15d104fd681706df5f64be583af24186abe
In a renewal message, in order to extend the duration, the server sdk
sends back a single key in the key container. Earlier the key was of type
CONTENT and only contained a key control block. No id, key data or
iv information were supplied. After a recent server sdk update
the key type in the renewal message has been changed to KEY_CONTROL.
This change enables the client to process the updated renewal message.
bug: 8736545
Merge of https://widevine-internal-review.googlesource.com/#/c/5434/
from Widevine CDM repository
Change-Id: Ia9f3620b86460278285fddee57ee923e269de7c1
Bug: 8770327
1. Allow provisioning retries
Allow multiple provisioning request messages to be generated without
requiring an equal number of HandleProvisioningResponse's. This is to
allow for lost messages.
2. Properly deletes cdm and crypto sessions created for cert provisioning.
The CleanupProvisioningSession() has not been deleting the cdm and crypto sessions
created for certificate provisioning properly. The lives of these sessions are
short and therefore, not added to the CdmSessionMap. We need to explicitly delete
these objects when error occurs or when we are done with provisioning.
3. Fixes provisioning responses that contain multiple chunks.
When we make multiple provisioning requests during testing, Apiary
sends response that contains more than one chunk. The test app.
needs to parse the response and concatenates the chunk data.
Otherwise, the size for each chunk is treated as base64 encoded data,
which will generate error when we try to deserialize the response
message.
Merge of https://widevine-internal-review.googlesource.com/#/c/5451/
from the Widevine CDM repository
Change-Id: I5b0ed982849c12628a3949f8d51515fcf6ce5a5f
The android timer class was not generating timer events correctly. This caused
renewal and expiration events not to be sent. A strong pointer to the
timer thread was not held and this caused the android util timer
thread to exit after firing once. This is now addressed.
Bug: 8736545
Merge of https://widevine-internal-review.googlesource.com/#/c/5353/
from the Widevine CDM repository.
Change-Id: I2d904e55d4d10eacc1a51f1c6b5c1a267c92c8d8
