This CL modifies the canonical interface for reusable burst executions:
- Add new method IBurst::createExecution
The reusable burst execution will not fallback to another execution path
if sending request packet fails. The behavior of single-time burst
execution remains unchanged.
Additionally, this CL enables pointer -> shared memory conversion in
1.2/1.3 burst implementation.
Bug: 184073769
Test: NNT_static
Test: neuralnetworks_utils_hal_1_0_test
Test: neuralnetworks_utils_hal_1_1_test
Test: neuralnetworks_utils_hal_1_2_test
Test: neuralnetworks_utils_hal_1_3_test
Test: neuralnetworks_utils_hal_common_test
Change-Id: Iaac81668d247c2cb76d70e6abbd10f00b397b19f
Merged-In: Iaac81668d247c2cb76d70e6abbd10f00b397b19f
(cherry picked from commit ead6d37ae9)
This CL modifies the canonical interface for reusable executions:
- Add new interface: IExecution with compute and computeFenced methods
- Add new method IPreparedModel::createExecution
In NNAPI runtime, the new interface IExecution is used to
memoize request-specific execution resources (e.g. converted HAL
request). The expected usage is that, IPreparedModel::createExecution
will be invoked in the first computation of a reusable NDK ANNExecution
object, and IExecution::compute* will be invoked repeatedly.
The IPreparedModel::execute* methods are preserved to avoid redundant
object creation and memoization overhead for a single-time
(non-reusable) execution.
For a vendor implementing the canonical interfaces, only the
IPreparedModel::execute* methods will be called because there is
currently no reusable execution at HAL interface. A DefaultExecution
implementation is provided to reduce the work needed on the vendor side.
Bug: 184073769
Test: NNT_static
Test: neuralnetworks_utils_hal_1_0_test
Test: neuralnetworks_utils_hal_1_1_test
Test: neuralnetworks_utils_hal_1_2_test
Test: neuralnetworks_utils_hal_1_3_test
Test: neuralnetworks_utils_hal_common_test
Test: neuralnetworks_utils_hal_aidl_test
Change-Id: I91790bb5ccf5ae648687fe603f88ffda2c9fd2b2
Merged-In: I91790bb5ccf5ae648687fe603f88ffda2c9fd2b2
(cherry picked from commit 727a7b2104)
Currently, vendors are asked to duplicate old events to onEvent()
and onEvent_1_1(). This is not the typically way other modules
would do. Thus, remove the requirement of duplicate events,
and replace it with only firing onEvent_1_1 when feasible.
Note no VTS is added currently, from vendor feedback, there is
no reliable way to guarantee the callbacks will be called in the
test.
Test: hidl-gen -L hash -r android.hardware:hardware/interfaces \
android.hardware.tetheroffload.control@1.1
Bug: 149467454
Merged-In: I75c1a7788e5d48fb3e2836d54b4fd854c0247214
Change-Id: I75c1a7788e5d48fb3e2836d54b4fd854c0247214
(cherry-picked from ag/14408065)
Commit 40eb8f53ea ("KeyMint AIDL tweaks") removed references to
getKeyCharacteristics() in docs, as it was a KeyMaster entrypoint
that wasn't present on the KeyMint HAL.
Commit dadb18dd29 ("Add getKeyCharacteristics method to KeyMint")
added the getKeyCharacteristics() entrypoint to KeyMint, as it
turned out it was needed after all.
This commit restores references to getKeyCharacteristics() in the
Tag.aidl documentation.
Test: VtsAidlKeyMintTargetTest
Change-Id: I860479554b85f4adfeddd4eee70a09cf5265c938
If GenerateKey() with user-provide key_blob, it needs to be specified in
the following begin() operations as well. Update the test case just to
take key_blob from private member instead of creating a local one.
Note:
- Remove redudent TAG_NO_AUTH_REQUIRED in DeviceUniqueAttestationTest
Change-Id: I81860294e1e7e01a57e66e08e75507a8292ec0c3
Keymaster is now optional because KeyMint can be used instead. This
CL makes this change retroactively back to compatibility matrix v3,
per discussion with the Treble team.
Test: Replace Keymaster with KeyMint & build
Change-Id: Id26d5bac326c494e1810516582fdd56a42c04e0d
Tests for:
- non-prime RSA exponent (fails with CF KeyMint)
- RSA exponent value of 3
- key size > 512 for `STRONGBOX`
- unknown tag inclusion
- CBC input size not block size multiple
- challenge omitted for attestation (fails with CF KeyMint)
- import RSA key with implicit params
- vestigial upgradeKey test
- importWrappedKey errors
- importWrappedKey sids ignored
- duplicate/missing params on begin()
- more tests for incompatible params on begin()
- HMAC size not multiple of 8 (fails with CF KeyMint)
- wrong size caller IV for 3DES rejected
- too large MIN_MAC_LENGTH for HMAC
- invalid AES-GCM minimum MAC length values
- check failed updateAad() cancels operation
- check that auto-generated nonces are distinct
- (DISABLED_) invoke destroyAttestationIds()
- omitting optional RSA keygen tags
Also add commenting to illustrate the ASN.1 structure of hex data.
Test: VtsKeyMintAidlTargetTest
Change-Id: I4663c42671cbb094ffe8d603e0352ffa9f1dbf2e
Add tests for:
- Too much entropy should be rejected with INVALID_INPUT_LENGTH
- All authorization lists should include a vendor and boot patchlevel.
These requirements are in both the KeyMint and the KeyMaster 4.0 AIDL
specificications, but have never been policed before.
Currently disabled with a command-line flag because CF does not have
the patchlevels and so fails lots of tests.
Test: VtsKeyMintAidlTargetTest
Change-Id: Ic9622ef3f1b80e013a34059218e3e029f392eb72
- Describe 3DES requirements
- Remove references to non-existent getKeyCharacteristics() entrypoint
- Remove references to non-existent attestKey() entrypoint
- Remove mention of asymmetric verification operations
- Remove text describing incomplete update()
- Remove text discussing padding for ECDSA
- EC keys can't encrypt/decrypt
- Force use of curve not key size for EC keys
- Drop TODO re: CBOR encoding of wrapped key
- Indicate caller-provided nonces must have correct size
- Add more reference to IKeyMintOperation
- Align parameter names with signatures
- Correct MIN_MAC_LENGTH error code
- Make CREATION_DATETIME consistently optional
- Describe use of MGF1 digest option
- Fix typos
Test: VtsKeyMintAidlTargetTest (CF)
Change-Id: Ie01b99d33be7336e872bf24a7761a3ac5b1d584c
China DTMB digital TV.
See SJ/T11368-2006
“Specification for multichannel digital audio coding technology”.
Test: atest AudioFormatTest
Bug: 185950819
Merged-In: I941c4991e598e864babfb888a8827d1625d95990
Change-Id: I941c4991e598e864babfb888a8827d1625d95990
* changes:
aidl: Set explicit default values
aidl: Set explicit default values
aidl: Set explicit default values for enum-type fields
aidl: Set explicit defaults to enum-type fields
Somehow the attestation format docs got dropped from KeyMint in the
transition from Keymaster. This replaces them, and also clarifies
that KeyMint StrongBox should not support device attestation.
Test: VtsAidlKeyMintTargetTest
Change-Id: I2334e99b4797c7a0e2e59727ffa730cf7504df31
This fixes warnings in AIDL files(-Wenum-explicit-default).
No semantic changes. (Explicit values are all zero enumerators.)
Fixes: 179853367
Test: mma in hardware/interfaces/keymaster
Change-Id: If5a70da3efd05a344c39ef1d2e73b7ec2b894f33
out-param arrays are considered harmful since the Java backend requires
them to be correctly allocated at caller site. But the only user of
android.hardware.identity is in C++, we can just ignore them.
Fixes: 179853403
Test: mma in hardware/interfaces/identity
Change-Id: Ib9b7ad895b3827c0888869a5b154696e3aff5624
