Age | Commit message (Collapse) | Author |
|
|
|
App profile directories are stored in DE storage, not CE. Run the
profile dir chown migration at the correct timing.
Test: manual
Bug: 217762162
Change-Id: If725e70f8ff12d9f374ca38fac1ef870a98708a9
|
|
|
|
|
|
Test: TreeHugger, m libgpumem gpuservice_unittest
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: If52eabe9430f70d3b4962b97a5b955ebba5bca37
|
|
Test: TreeHugger, m libtimeinstate libtimeinstate_test
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: I260f224316ae47df0e7c22e76cd07fd4efa9afc2
|
|
|
|
Generate BTF debug info when building gpu_mem.o. bpfloader will load
map info alongside gpu_mem_total_map, which allows pretty-printing via
the "cat" command for easier debugging.
Test: m gpu_mem.o
Test: gpuservice_unittest --gtest_filter=GpuMem*
Test: cat /sys/fs/bpf/map_gpu_mem_gpu_mem_total_map
Bug: 203823368
Signed-off-by: Connor O'Brien <connoro@google.com>
Change-Id: I453766c0f76e1aba5720400a0ec2e2f871967a91
|
|
|
|
As per b/32470528#comment31, there are no longer plans to go forward
with this.
Bug: 32470528
Test: atest dumpstate_smoke_test
Change-Id: I5d064aff73e19de573404c5a5b6f1376a1538db4
|
|
|
|
|
|
The async onBufferReleased callback can trigger another
onBufferReleased which will end up deadlocking the
async worker thread. Fix this by executing the callbacks
outside the lock.
Test: atest android.media.cts.MediaSyncTest#testPlaybackRateDouble --rerun-util-failure 100
Bug: 201604213
Change-Id: I40d163c3644c6a0128936cf41e8bf8969766d9da
Merged-In: I40d163c3644c6a0128936cf41e8bf8969766d9da
|
|
|
|
* changes:
libbinder: Status writeOverParcel
libbinder: format tests bp
|
|
Test: atest dumpstate_smoke_test
Change-Id: Iaf3bac5b3d80252a006a03bcc503e18200366771
|
|
|
|
There's some occurences of a transaction just under 1 MB into
system_server, which causes system_server to reject other transactions,
and sometimes throw DeadSystemException.
Until we have better backend infrastructure in place, at least log the
source and some data of really large transactions, which might give us
some sort of clue.
Bug: 198380036
Bug: 213349547
Test: N/A
Change-Id: Ifccf4a2dea32cd43421b61d3003d2a2bdc86ee23
|
|
For errors written to the parcel late.
Bug: 211823183
Test: aidl_integration_test
Change-Id: I036a016e143408bf1c3cb4f67f622c7e0dac91bf
|
|
Is in PREUPLOAD.cfg.
Bug: N/A
Test: N/A
Change-Id: Iefe28e2ef927f9f298ab5ac5e26875bc5d2336d5
|
|
|
|
This can help binder clients know when a service has been registered.
Potentially helpful, in situations where some HALs take some seconds to
come up and register with service manager.
Bug: 196432585
Test: manual
Test: binderLibTest
Change-Id: Ib1d9b967a7cffa99f43dc26f9025caf1f59ab77f
Signed-off-by: Jayant Chowdhary <jchowdhary@google.com>
|
|
surfaceflinger will crash."
|
|
|
|
|
|
* changes:
libtimeinstate: add tests to presubmit
libtimeinstate: skip tests on non-supporting devices
|
|
Run libtimeinstate_test in presubmit for libtimeinstate changes.
Test: atest --test-mapping frameworks/native/libs/cputimeinstate;
confirm that libtimeinstate_test runs & passes
Bug: 138317993
Change-Id: Ide638de1062c3b5b7fc7c4e969f3030e4264a846
Signed-off-by: Connor O'Brien <connoro@google.com>
|
|
Devices that can't support BPF time in state tracking (e.g. upgrading
devices without kernel support for perf event selinux hooks) currently
fail libtimeinstate_test without this failure indicating a bug or
incorrect behavior. To make test failures meaningful, skip the tests
on such devices instead of failing.
isTrackingUidTimesSupported() is now exercised by the SetUp()
function, making the IsTrackingSupported test redundant, so remove
that test.
Test: run libtimeinstate_test on device without time_in_state.o and
confirm all tests are skipped
Change-Id: I11bc5cc6b2e32f56c40852846fddd6790d6643b3
Signed-off-by: Connor O'Brien <connoro@google.com>
|
|
|
|
|
|
musl libc uses the posix definition of msg_iovlen as int. Change the
niovs argument to an int, make sure it is not negative, and cast it to
size_t if necessary.
Bug: 190084016
Test: m USE_HOST_MUSL=true host-native
Change-Id: I6ff9206e9e7396f2f89622735f790b4fac18f76c
|
|
|
|
|
|
|
|
|
|
* changes:
SF: adjust kNonExactMatchingPenalty
SF: give a higher score to frame rates which exact matches
Tune RefreshRateConfigs for fractional refresh rates
|
|
|
|
We still weren't getting coverage here, so manual corpus.
Here is the layout of a special transaction to get the root object, and
a transaction on the root object, respectively:
/ .... init ... \/ conn init \/ RpcWireHeader \/ RpcTransactionBody \/ ctrl \
000000F0000000000000000000000000636369000000000000000000280000001111111111111111000000000000000000000000000000000000000000000000000000000000000000000000000000008d4a8d4a
000000F0000000000000000000000000636369000000000000000000280000001111111111111111030000000000000000000000000000000000000000000000000000000000000000000000000000008d4a8d4a
The test case which is added is done by joining these two transactions
(the root object must be retrieved before we can transact on it).
echo "000000F00000000000000000000000006363690000000000000000002800000011111111111111110000000000000000000000000000000000000000000000000000000000000000000000000000000000000000280000001111111111111111030000000000000000000000000000000000000000000000000000000000000000000000000000008d4a8d4a" | xxd -ps -r
Hopefully this will be enough for the fuzzer to find the transaction
operations we have setup there.
Bug: 199324691
Test: rpc_binder_fuzzer tests/rpc_fuzzer/corpus/transact_on_binder (w/ log showing coverage)
Change-Id: I4b9b62525ec27994db261d7ee354049a21168602
|
|
Bug: 216357072
Bug: 216358770
Change-Id: Iddf81605d2cc4f7d7ec79538fb2fc51384a5a30d
|
|
Fixes: 205301887
Test: Compare binary locations
Change-Id: Id9a3730ac703c1c840dc3b9a13251a61ec346411
|
|
Sometimes bugreports finish slightly before the 30 second
threshold, causing test failures. This change reduces this
threshold to 20 seconds and fixes some misleading failure
logs.
Test: atest dumpstate_smoke_test
Bug: 211004508
Bug: 211488888
Change-Id: I37e7707df3237f39d099484a26f195a85f16767a
|
|
|
|
|
|
|
|
|
|
Bug: 215092166
Test: manual test
Change-Id: I616bfa4994b2b20c8545834d7d746987baef9ca1
|
|
|
|
|
|
|
|
Bug: 214231981
Test: run microdroid
Change-Id: I036441537e00aae8bae58200007a545d6ce9905e
|