Originally landed as I5dbb01fc23abd5e0108c565f96d25e62c77fc16d and
previously reverted in I6265f8de3db31f07506906cee82a91fe3baac0bc due to
timeline constraints.
Add a test presenting from multiple threads. Split up execute() so that
it can be called for a single display. Update MultiThreadedPresent
(test) to offload presentDisplay to a separate thread, similar to how
Ib9d074671e32c95875ef7e0791dd95d6e595e47a does it, as described in
go/multi-threaded-present.
Bug: 259132483
Bug: 284156408
Fixes: 251842321
Fixes: 295841597
Test: VtsHalGraphicsComposer3_TargetTest
Change-Id: If975ee9bb0b9c6f64ef50401e2aee32f934e3f08
We deliberately use a deprecated API to test that it is not used. Ignore
warnings from using it in this test.
In PresentDisplayNoLayerStateChanges, remove the check for
SKIP_VALIDATE, removing the other warning. As written, the test would be
skipped without SKIP_VALIDATE support. But devices should *not* claim
Capability::SKIP_VALIDATE, meaning the test would always be skipped. The
docs state that the behavior previously indicated by SKIP_VALIDATE is
now the default, so we should always be running the test.
Bug: NA
Test: atest VtsHalGraphicsComposer3_TargetTest (no warnings)
Change-Id: Icc89a82b4850428c2f0940be0f78f5ebff1881b3
... as a
static vendor::mapper::IMapperProvider<IMPL> provider;
may be destructed during a clean process exit (during __cxa_finalize())
but the underlying IMPL may still be in use on other threads.
Bug: b/304343434
Test: boot Cuttlefish
Change-Id: I80687494c61210d527908b5a67375950ca45c8ea
Right now this only makes sure that extended sRGB buffers aren't
corrupted by the time it hits the display. This is especially important
for Ultra HDR, because display compensation is defined by the spec.
Bug: 276331687
Bug: 259144781
Test: VtsHalGraphicsComposer3_ReadbackTest
Change-Id: Ib6c39858289d5559ae76280ca1a0ef756f2daef0
The nit information was used when we set exact nit values rather than a
relative brightness per layer. But we only need nit values for the
renderengine interface, which isn't tied to any hwc or panel
capabilities.
Bug: 301261125
Test: VtsHalGraphicsComposer3_TargetTest
(cherry picked from https://googleplex-android-review.googlesource.com/q/commit:712b3d9880b6aff51ab17d539bfcbdf3785cc6ca)
Merged-In: I770dc5620648df2eab608e030c5e76cf190f315d
Change-Id: I770dc5620648df2eab608e030c5e76cf190f315d
The nit information was used when we set exact nit values rather than a
relative brightness per layer. But we only need nit values for the
renderengine interface, which isn't tied to any hwc or panel
capabilities.
Bug: 301261125
Test: VtsHalGraphicsComposer3_TargetTest
Change-Id: I770dc5620648df2eab608e030c5e76cf190f315d
This reverts commit f850701be0.
Reason for revert: Delaying rename to avoid churn for external partner work.
Change-Id: If96786d074a61ac0f781540e98d020a6f7fd673b
Color management is now turned on by default, so there's no reason to
explicitly enable it in renderengine.
Bug: 295966830
Test: builds
Change-Id: I42c8afb30950081cd57072680da06895dcc02f2c
Some devices shipped with broken validation for HDR metadata. Relax the
test for vendor freeze purposes.
Bug: 276843519
Test: VtsHalGraphicsMapperV4_0TargetTest
Change-Id: I564f4bd12145f7831faa1e5a0f2d2abe074e4769
Test getDisplayConfigurations with vrr parameters
Test notifyExpectedPresent does not return error
Test: atest VtsHalGraphicsComposer3_TargetTest
BUG: 287517352
BUG: 284881912
Change-Id: I49d6876922e9fde47e531200a29413495458b8a5
