| Age | Commit message (Collapse) | Author |
|---|
|
This reverts commit 02a80d00ffd5eaa1985e18becd542579576720fc.
Reason for revert: Fix dump_trace.cc to parse timestamps as 64-bit unsigned lebs and not 32-bit. These were always 64-bit fields but the bug didn't impact because timestamp diff did fit in 32-bits.
When using ns, it no longer fits exposing the bug.
Change-Id: If979c0bf891ae5bd0fdf278915fcda8ed09bfa9f
|
|
This reverts commit 84c60905008131ff35b93b9f8b64923e267f92a6.
Reason for revert: Reverting because of failures on qemu.arm64.
https://ci.chromium.org/ui/p/art/builders/ci/qemu.arm.64/202/overview
Change-Id: I96eeb4b675eb36850f64e4f3102e55f456747762
|
|
This reverts commit e8ea3a17d66636e6cc6a395b967e22aac8761e67.
Reason for revert: Upload after potential fix for the failures.
We used to use 32-bit entries for storing threadcpu time. With
nanosecond precision this overflows in about 4 seconds. On gcstress
configuration this was potentially overflowing causing the failures.
The fix now uses 64-bit entries for threadcpu time. Wall clock time was
already using 64-bit entries.
Change-Id: I819fe45e0a465e5a5b037248a4cd7107acb5130d
|
|
This reverts commit 1a9a8abf19e7641149fcaed3971bf0aa99f4dfb8.
Reason for revert: causes flakes on art-jit-on-first-use-gcstress
Change-Id: I38161f65e2fc584e5e6221f85982035fb66c13ae
|
|
Update the precision of timestamps to be nanoseconds and use nanosecond
precision in method trace format v2. For format v1 we use microsecond
precision for backwards compatability.
Bug: 259258187
Test: art/test.py -t 2246
Change-Id: I099e21f5d143c5444193e7886b5281b32f87d5d3
|
|
For proxy methods we need to get corresponding interface method. This
line was dropped accidentally in aosp/3434019. This CL re-introduces the
line.
Bug: 352518093
Test: art/test.py --tracing
Change-Id: If53bd6b7219c6ec81889b16cfb6677adf0cbb3ab
|
|
This CL extends low overhead tracing to support collecting long running
methods. This feature is disabled by default and is implemented behind
always_enable_profiling_code flag. Turning on the flag enables the
support but doesn't start tracing. Tracing has to be started via the
StartTraceLongRunningMethods interface in VMDebug. If a trace was
running the collected data is included in the data generated by the
SIGQUIT handlers. This would help when analyzing ANR reports.
The trace can also be dumped using the API provided in VMDebug.
Bug: 352518093
Test: art/test.py
Change-Id: I407560c45cd31bfa06906bb2286c455eef7ceec8
|