Age | Commit message (Collapse) | Author |
|
Change-Id: Ieb3ea88d65d9f7682e136f9cc701031ed8875d83
|
|
Original change: https://googleplex-android-review.googlesource.com/c/platform/system/extras/+/12128008
Change-Id: Ifde03a5f1d91740c4b353d7ee6a71f48eaa5e45d
|
|
Execute the commands with the API version where the command was
introduced. This makes it possible to run the v1.0 HAL commands on
a device where v1.1 HAL implementation is not available.
Without this patch, on devices that only support bootcontrol v1.0 API,
bootctl would just fail with error message.
Bug: 160405575
Test: Execute command on device that only implements v1.0 API
adb shell bootctl hal-info
Change-Id: I56e9a81bcae49b6859b04394d571b5151a235675
Signed-off-by: Håkan Kvist <hakan.kvist@sony.com>
Signed-off-by: Alessio Balsini <balsini@google.com>
|
|
Change-Id: I4ff28ec77ff10c4d778190d878e580d8156f4896
|
|
Original change: https://googleplex-android-review.googlesource.com/c/platform/system/extras/+/11829463
Change-Id: I8dce95af33f6704e287a72849c6846f37dfce346
|
|
Bug:158702468
Test: Run bootanalyze and see there is no error when the
event time is 0
Change-Id: If3efa4c33a70cf6168a777f0f442fe1557b600b4
|
|
Change-Id: Ifb227a7e5d31efedd6e22a60ba21780135d4b3d4
|
|
counters. am: 756a972f99
Original change: https://googleplex-android-review.googlesource.com/c/platform/system/extras/+/11735563
Change-Id: Ibf73c7aca5e698ce98b4e7a3a955ab746e8d1121
|
|
Some Socs like Broadcom STB will not be available when
disable jtag(non-invasive debug), so omit tests requiring
hw counters on them.
Bug: 158095412
Bug: 154157678
Test: run cts -m CtsSimpleperfTestCases -t record_cmd#generate_samples_by_hw_counters
Test: hw counters.
Change-Id: Iddfb34bfa1d66e37398bb162f8e1cfa645d756f4
Merged-In: Iddfb34bfa1d66e37398bb162f8e1cfa645d756f4
(cherry picked from commit 572c226db410b2c3b20e3a63cc6c5e9f0aeed65c)
|
|
Change-Id: Iae4936ef7832dd61d98da14aba894b608906dc22
|
|
Change-Id: I53fc8b07ab20aea0378553f1e523e3d2abfa67d8
|
|
It's possible that the an image doesn't contain a hashtree
descriptor. And the current raises a segfault when we attempt
to validate and copy a non existing descriptor.
Bug: 157778739
Test: call fec_open on dtbo image.
Change-Id: Ieb5acf016ceeb547f2b23de43d4276ea4d013a6f
(cherry picked from commit 90e7155bdd10039448bbc5ee4636f14c0119718e)
|
|
Change-Id: Ic5c4516194f3541bd9d1210c4c1338580cca21c6
|
|
Change-Id: Iea606c456f5c57043069c0d39a6d8efbf39a075a
|
|
Bug: 144046242
Test: atest libfscrypt_unit_test
Test: Change fstab on Cuttlefish with patched kernel, check dir policy
Cherry-Picked-From: 476ce0a30615c1249969bedba0f1fb084d8ef193
Merged-In: I362f9a55fa28bb25afe992c8b0bae48546fc9ab0
Change-Id: I362f9a55fa28bb25afe992c8b0bae48546fc9ab0
|
|
Change-Id: I6d5349105212651644f71503540844bea5803c83
|
|
Change-Id: I8d4abf72e13a709deb28debf93ed60dd95c0fbbe
|
|
Change-Id: If6529f8519e710ca35f48694ae93c59834751aef
|
|
Change-Id: Ib33f81c7962ab0a61f191a3795abfcc01b2486f3
|
|
Currently, simpleperf gets branch_addr from instruction decoder.
But OpenCSD is changed to cache InstrRange elements. Thus when
simpleperf gets an InstrRange element, the instruction decoder
may point to an instruction executed later.
Fix it by using the start_addr of the next InstrRange element
as the branch_addr of the current InstrRange element.
Also add a test.
Also use recorded binary path in output to help test.
Bug: 153039105
Test: run simpleperf manually so the inject result doesn't
Test: change for a 10s system wide recording data.
Test: run simpleperf_unit_test.
Change-Id: I6fdf32d3bac18ed3762c944c282ec881d09395b4
Merged-In: I6fdf32d3bac18ed3762c944c282ec881d09395b4
(cherry picked from commit 2c294914c1eeb30754a8a2449795b296bd2176e9)
|
|
Add MapLocator as a packet callback. It is then used
to find maps in MemAccess and InstrRangeParser.
MapLocator uses a separate cache for each trace id.
It also fix a bug that let a kernel map override
following user maps.
Bug: 153039105
Test: run simpleperf_unit_test.
Change-Id: Ic285f81ac05c3e65b481fb97a5355282639692e4
Merged-In: Ic285f81ac05c3e65b481fb97a5355282639692e4
(cherry picked from commit 418ba0d37570172c4ee5e24fc452b3a5ef9ae665)
|
|
This is to keep file mapping memory buffer. And support reading
more than one content in one open instance in the future.
It also supports embedded elf files in apks as normal elf files.
Bug: 153039105
Test: run simpleperf_unit_test.
Change-Id: Ia424926d112cbcd9970f11ffa56d047ff6df7872
Merged-In: Ia424926d112cbcd9970f11ffa56d047ff6df7872
(cherry picked from commit 02e2033f9ef34be0189e151bf6fef60b1285afd9)
|
|
Change-Id: I539306d244e239fb67121b6ded45262def58bba9
|
|
Change-Id: Ie6271dccf7788d7ecdcbe134f73bc755a750f63d
|
|
|
|
Change-Id: I8b2492e8b2f6ecf98da007249e7420e70138634e
|
|
Change-Id: Ibc1f192780f0d15720c06a9c1d77dbd66a8dca68
|
|
Fix test apk by always using internal cache dir.
Bug: 153605370
Test: run simpleperf_unit_test.
Change-Id: Ie38242e2034f0f9888da380f7cee696bf3877818
(cherry picked from commit 42ce297dff6df9c6a54366b9090a1bab7f52e638)
|
|
With b/152655547, all aidl_interface modules are considered as stable
unless it is explicitly with "unstable: true". This change marks the
aidl_interface that are not used across updatable module bounraries
as unstable, so that the build system does not run the API
dumping/checking on them.
Bug: 152655547
Test: m
Change-Id: I8291dca96fcca0c45a2501ab88970ace360a7ac7
|
|
Change-Id: I468d2b85c66eb688c6d7a716d38331b99b3d99a5
|
|
ae7f2edaf4
Change-Id: I1e84bc082295a7fcd8a44f1d1acc8e08ed66b737
|
|
|
|
Change-Id: I2fc6ddd6b563b3014dda165bfac7387826f4f057
|
|
Change-Id: I15ae6a865847c3c6667404cce8234c8c41fe922e
|
|
|
|
Change-Id: I47fda596e7f8d31e2587671cc7a38677bf0ca157
|
|
The tests are only needed in vts10 suite, as they are part of CTS
package and no longer needed in vts 11.
Bug: 151896491
Test: local build
Exempt-From-Owner-Approval: This CL removes vts10 tests from vts suite.
It won't change test logic or behavior.
Change-Id: Idadd1f99ba3620c36a7aca4f648a013da40bd3a0
Merged-In: Idadd1f99ba3620c36a7aca4f648a013da40bd3a0
|
|
Change-Id: I728bfe8faf614637f1024b459eb78bfefdd0394e
|
|
sys.init.perf_lsm_hooks isn't accessible in app context, thus
checking it in CanRecordRawData() breaks recording app with
--trace-offcpu option.
Fix it by not recording raw data in non-root users.
Bug: 153381808
Test: run simpleperf_unit_test.
Test: run test.py *TraceOffCpu*.
Change-Id: Id87c73d25b19b265c022c4fa6ebb8e7d3a84152b
(cherry picked from commit c98cb1193d73c1fa616d0d366a221751fde38f67)
|
|
Change-Id: I9fb247ba2ba53d2c8d1b23331f19334feef022e2
|
|
We may grab the by-name path for /data's block device. Use realpath to
ensure we grab the correct name to access the sysfs entries.
Bug: 153049163
Test: Run /vendor/bin/checkpoint_gc
Change-Id: I297c7b60105d1966eb1304222a9727794035d3fc
|
|
Change-Id: I5c9480298504af21f86d580d1ce2806a54e3037d
|
|
Change-Id: I1c27a6b4bac6cfbc2780f163145734398fe57fe2
|
|
Change-Id: I5fa5fa68185b7d1ee226b441761a4f9d32b4af7d
|
|
For task-clock, it uses software counter, so multiplexing will not
happen. And it isn't proper to scale comment value in per-core report.
For rate computation involving runtime, the current way to find runtime
is via FindRunningTimeForSummary() function. But that function doesn't
take care of thread and core match. And it seems more proper to use
the sum of runtime given by perf counters. So switch to it.
Also add some tests for comment string.
Also adjust comment title in report.
Bug: 152911261
Test: run simpleperf_unit_test.
Change-Id: I475fae32952ed2680888c6a76b43cfb10bc9b06e
(cherry picked from commit 6922c076ce73fd4bca357aeea0d20797e3321749)
|
|
This is to test it in later CLs.
Bug: 152911261
Test: run simpleperf_unit_test.
Change-Id: I0cdad299654eb36db4a64363335bfa03e54ee7b1
(cherry picked from commit 5271aa719b7cae4cddc9f67803648f17f9b24018)
|
|
Change-Id: Ieec7288bf4cd912dcfb9deb2a91922946b400c7d
|
|
Change-Id: I1dc0da8ead55bf0fba2d028e07c80bbea8d7f7d0
|
|
Change-Id: I80619690dda6934bc61d2e2879a57e8d44ad97e6
|
|
Change-Id: I31123383f61ae782c3998123c73554114dc4954a
|