OSDN Git Service
Mathieu Chartier [Wed, 19 Apr 2017 02:41:02 +0000 (02:41 +0000)]
Merge "Call mirror::CallSite::VisitRoots"
Mathieu Chartier [Tue, 18 Apr 2017 23:14:19 +0000 (16:14 -0700)]
Call mirror::CallSite::VisitRoots
Needed to visit CallSite static class root.
Test: test/run-test --no-patchoat --relocate --host --64 --interpreter --gdb 952
Bug:
33191717
Change-Id: I1fad3df8ad810c5c7905afadfa25cdc0901ef035
Calin Juravle [Tue, 18 Apr 2017 23:20:16 +0000 (23:20 +0000)]
Merge "Invoke dex2oat explictly before dalvikvm"
Treehugger Robot [Tue, 18 Apr 2017 22:42:58 +0000 (22:42 +0000)]
Merge "Fix obsolete method change check."
Calin Juravle [Wed, 12 Apr 2017 02:04:28 +0000 (19:04 -0700)]
Invoke dex2oat explictly before dalvikvm
The goal is to stop relying on dex2oat being run when dex files are
loaded.
Test: ./out/host/linux-x86/bin/art --profile -Ximage:$PWD/out/host/linux-x86/framework/core.art -Xusejit:true -Xcompiler-option --compiler-filter=speed -Xcompiler-option --compiler-backend=Optimizing -verbose:oat -Djava.library.path=$PWD/out/x86_64/host/linux-x86/lib64 -cp $PWD/generated/benchmarks.dex benchmarks.ExoPlayerBench.java.ExoPlayerBench
Bug:
36824842
Change-Id: I9f4a1384cdc91502edea08402ee984c57ff8c37d
Alex Light [Tue, 18 Apr 2017 20:03:31 +0000 (13:03 -0700)]
Fix obsolete method change check.
We were incorrectly checking for changes in the obsolete-method map.
This could cause issues if a class was redefined multiple times.
Bug:
37475600
Test: ./test.py --host -j40
Test: cts-tradefed run cts-dev --module CtsJvmtiRedefineClassesHostTestCases
Change-Id: Icf39d5154c0e48461405b700bf1fa20830195fc1
Treehugger Robot [Tue, 18 Apr 2017 18:25:01 +0000 (18:25 +0000)]
Merge "Make ART build/test configurations use the CC collector by default."
Mathieu Chartier [Tue, 18 Apr 2017 18:24:00 +0000 (18:24 +0000)]
Merge "Differentiate between native alloc and normal background GC"
Treehugger Robot [Tue, 18 Apr 2017 18:21:41 +0000 (18:21 +0000)]
Merge "ART: Make less lock-level noise on abort"
Mathieu Chartier [Mon, 17 Apr 2017 22:24:43 +0000 (15:24 -0700)]
Differentiate between native alloc and normal background GC
Added a new GC cause kGcCauseForNativeAllocBackground.
Bug:
35872915
Test: test-art-host
Change-Id: I94e17f8bd53af29f2862b9910bd8abd2df97e229
Mathieu Chartier [Tue, 18 Apr 2017 17:14:00 +0000 (17:14 +0000)]
Merge "Use correct type for GetValueFromShadowFrame"
Roland Levillain [Tue, 18 Apr 2017 17:06:25 +0000 (17:06 +0000)]
Merge "Use WARN_UNUSED consistently in ART."
Mathieu Chartier [Tue, 18 Apr 2017 03:12:29 +0000 (20:12 -0700)]
Use correct type for GetValueFromShadowFrame
The field type is not necessarily the input type for boxed
primitives. If the field type is < 32 bits, it means there will be
partial object pointer in the JValue. If a conversion check is
later needed in GetUnboxedTypeAndValue, it will crash. The fix is
to use the PTypes.
Bug:
37446461
Test: test-art-host
Change-Id: I0c4b405f0c13910523b98a87ef12b9f302a5e241
Sebastien Hertz [Tue, 18 Apr 2017 16:11:55 +0000 (16:11 +0000)]
Merge "JDWP: fix Dbg::ResumeThread"
Treehugger Robot [Tue, 18 Apr 2017 15:12:58 +0000 (15:12 +0000)]
Merge "ART: Call ThreadGroup.add in Thread::FinishStartup"
Roland Levillain [Tue, 18 Apr 2017 12:37:10 +0000 (12:37 +0000)]
Merge "Fix ARM64 SystemArrayCopy intrinsic with large constant dest position."
Vladimir Marko [Tue, 18 Apr 2017 12:06:55 +0000 (12:06 +0000)]
Merge "Revert "Disable 160-read-barrier-stress temporarily""
Treehugger Robot [Tue, 18 Apr 2017 11:29:25 +0000 (11:29 +0000)]
Merge "Fix Arena allocation tracking."
Roland Levillain [Tue, 18 Apr 2017 11:01:58 +0000 (12:01 +0100)]
Use WARN_UNUSED consistently in ART.
Test: mmma art
Change-Id: I776840725f56258c45edfcfd165da61ab653f0b5
Roland Levillain [Thu, 13 Apr 2017 18:34:30 +0000 (19:34 +0100)]
Fix ARM64 SystemArrayCopy intrinsic with large constant dest position.
Make sure we do not deplete the whole VIXL scratch register pool, so
that VIXL can still use IP0 as a temporary when emitting
macro-instructions.
Test: art/test/testrunner/testrunner.py --optimizing --target --64
Bug:
37256530
Change-Id: I5da22e552297fad87db5763e2dab60ae6a7a43af
Vladimir Marko [Tue, 18 Apr 2017 09:47:12 +0000 (09:47 +0000)]
Revert "Disable 160-read-barrier-stress temporarily"
And reduce the number of fields in class ManyFields from
10000 to 5000 to reduce memory pressure on Jack. Split
the ManyFields class in a simple hierarchy to keep files
under 64KiB and allow compilation with javac.
Test: testrunner.py --host -j 48
Test: run-test --host --jvm 160-read-barrier-stress
Bug:
37335480
This reverts commit
101074999bd339c276eeaaf0538930d90a96657e.
Change-Id: I52217fd2c9d00a5b045a7faa6e4dc1438c7f93b1
Andreas Gampe [Tue, 18 Apr 2017 04:40:28 +0000 (21:40 -0700)]
ART: Make less lock-level noise on abort
The lock-level violations with the abort lock aren't really all
that interesting.
Test: m test-art-host
Change-Id: I8a5fc687009db914ec8f60d86068d87e71f8a894
Andreas Gampe [Tue, 18 Apr 2017 03:19:14 +0000 (20:19 -0700)]
ART: Call ThreadGroup.add in Thread::FinishStartup
ART should add the main thread to the main ThreadGroup. Behavior
of the Thread constructor changed.
Bug:
37444210
Test: art/test/testrunner/testrunner.py -b --host -t 051
Test: m test-art-host
Test: m build-art-host && art/tools/run-libcore-tests.sh --mode=host
Change-Id: I92cf2f9a6c5c3fdf385eb7925addc38b64fa4d98
Treehugger Robot [Tue, 18 Apr 2017 00:58:38 +0000 (00:58 +0000)]
Merge "sigchain: switch from __thread to pthread_setspecific."
Josh Gao [Mon, 17 Apr 2017 22:58:36 +0000 (15:58 -0700)]
sigchain: switch from __thread to pthread_setspecific.
__thread is implemented via emutls on Android, which will result in the
thread local variable being allocated again and leaked if it gets used
after it has been destructed already (e.g. by a later destructor
calling sigprocmask, or triggering a signal handler).
Switch to pthread_setspecific, which doesn't suffer from this problem.
Bug: http://b/
36871013
Test: ran dalvikvm on a class that spins creating threads
Change-Id: Ie5deab453be387490ba30a0010e12f60d736c8ad
Treehugger Robot [Mon, 17 Apr 2017 21:45:18 +0000 (21:45 +0000)]
Merge "Enable some redefinition run-tests in CTS"
Treehugger Robot [Mon, 17 Apr 2017 21:42:46 +0000 (21:42 +0000)]
Merge "ART: Change run-test 913 array testing"
Treehugger Robot [Mon, 17 Apr 2017 21:35:48 +0000 (21:35 +0000)]
Merge "Disable read barrier thunks for heap poisoning"
Calin Juravle [Mon, 17 Apr 2017 21:26:47 +0000 (21:26 +0000)]
Merge "Print runtime isa when invoking 'dalvikvm --showversion'"
Alex Light [Fri, 14 Apr 2017 23:20:53 +0000 (16:20 -0700)]
Enable some redefinition run-tests in CTS
We added the tests to run-test-jvmti-java-library
target and made some small changes to tests to
ensure that everything works without linking to
libart.so.
Bug:
32072923
Test: ./test.py --host -j40
Change-Id: I8fdad180c77b125c22c043cd61305cee79801cc7
Andreas Gampe [Mon, 17 Apr 2017 19:46:36 +0000 (12:46 -0700)]
ART: Change run-test 913 array testing
Use an explicit array in the test. Ignore all arrays that are
untagged.
Bug:
32072923
Bug:
36727422
Test: m
Test: art/test.py --host -r -t 913
Change-Id: Idf7bc03ae81e3e645935c8809bf59636cfa3784a
Mathieu Chartier [Mon, 17 Apr 2017 19:50:45 +0000 (12:50 -0700)]
Disable read barrier thunks for heap poisoning
Logic to unpoison is not yet implemented, this causes SIGSEGVs in
various places.
Bug:
29516974
Bug:
30126666
Bug:
36141117
Test: ./test/testrunner/testrunner.py -j4 --optimizing --debuggable --ndebuggable --target --verbose
Change-Id: I8317a142d07af36090d5f05ce46100dfa07c17e7
Treehugger Robot [Mon, 17 Apr 2017 18:59:40 +0000 (18:59 +0000)]
Merge "Use standard loading code in many tests"
Treehugger Robot [Mon, 17 Apr 2017 17:19:05 +0000 (17:19 +0000)]
Merge "Remove native code from test 944 for CTS"
Alex Light [Mon, 17 Apr 2017 17:06:44 +0000 (10:06 -0700)]
Use standard loading code in many tests
In order to prepare for use in CTS tests make tests 902, 914, 915,
917, 919, 926, 930, 932, 940, 942, 945, 981 & 982 use standard OnLoad
functions. Final configuration is done during test initialization.
Bug:
32072923
Test: ./test/testrunner/testrunner.py --host -j40
Change-Id: I84b2e17d0e024255d2dc49452f3d74dfd64abd22
Aart Bik [Mon, 17 Apr 2017 16:34:52 +0000 (16:34 +0000)]
Merge "Fixed bug on pending environment use of termination condition. With regression test."
David Sehr [Mon, 17 Apr 2017 15:49:41 +0000 (15:49 +0000)]
Merge "Add art_defaults to dexlayout"
Alex Light [Mon, 17 Apr 2017 15:36:14 +0000 (08:36 -0700)]
Remove native code from test 944 for CTS
Replaced native code in test 944 with java reflection based code.
Bug:
32072923
Test: ./test/testrunner/testrunner.py --host -j40 -t 944
Change-Id: I678d805936009a33ce8fe6b8570bfb788a8da72d
David Sehr [Sat, 15 Apr 2017 17:06:21 +0000 (10:06 -0700)]
Add art_defaults to dexlayout
The build rule defaults for dexlayout targets didn't include defaults
from the ART project. Because of this ART_TARGET_* weren't defined,
among other problems
Bug:
35800981
Test: make -j 40 test-art-host
Change-Id: Ic528bfc6b622fa9b4f5c0ddf883f15912641b720
Treehugger Robot [Sat, 15 Apr 2017 06:35:54 +0000 (06:35 +0000)]
Merge "ART: Tentative fix for stack count issue in test 913"
Andreas Gampe [Sat, 15 Apr 2017 04:49:33 +0000 (21:49 -0700)]
ART: Tentative fix for stack count issue in test 913
Use kCountTransition == kNo to also skip runtime methods in the
cur_depth_ count of a StackVisitor.
Currently this is only used in the VisitRoots function of Thread.
If the additional load is too expensive, we can make this dependent
on the requested precision (and thus revert "regular" visits to
the old scheme of also counting transitions between managed stack
frames and shadow frames, which is only an increment of a field).
Bug:
31385354
Bug:
32072923
Test: m test-art-host
Change-Id: Ief951585407dd6e03f8050774fe612e7b204247e
Treehugger Robot [Sat, 15 Apr 2017 02:43:14 +0000 (02:43 +0000)]
Merge "ART: Add debug code to run-test 913"
Alex Light [Sat, 15 Apr 2017 02:20:48 +0000 (02:20 +0000)]
Merge changes I5b05e23a,I9350b250,I1d66b42a
* changes:
More refactor for JVMTI redefinition run tests (3/3)
More refactor for JVMTI redefinition run tests (2/3)
More refactor for JVMTI redefinition run tests (1/3)
Treehugger Robot [Sat, 15 Apr 2017 01:03:50 +0000 (01:03 +0000)]
Merge "Add check that TypeIndex is valid in StringByTypeIdx."
Jeff Hao [Fri, 14 Apr 2017 21:33:52 +0000 (14:33 -0700)]
Add check that TypeIndex is valid in StringByTypeIdx.
StringByTypeIdx should fail gracefully if given a bad TypeIndex. This
adds a check that the TypeIndex is valid before getting its TypeId.
This fixes a regression that removed this check when it was refactored
in this CL: https://android-review.googlesource.com/#/c/243493/
Bug:
37287051
Test: mm -j31 test-art-host-gtest-dex_file_test
Change-Id: Ib68cb8135011f5f30335251583e181b089982754
Calin Juravle [Thu, 13 Apr 2017 23:35:42 +0000 (16:35 -0700)]
Print runtime isa when invoking 'dalvikvm --showversion'
The runtime ISA is appended at the end, after the version number.
Test: dalvikvm --showversion
Bug:
36824842
Change-Id: I09bbe3bf06054fe237543b5f96e9480631ea0838
Treehugger Robot [Fri, 14 Apr 2017 22:47:55 +0000 (22:47 +0000)]
Merge "Remove test skips associated with b/
37240685"
Alex Light [Fri, 14 Apr 2017 20:22:22 +0000 (13:22 -0700)]
More refactor for JVMTI redefinition run tests (3/3)
This adds in the Main.java files for the modified tests.
By doing the 3 CLs we keep git from getting confused about the history
of the files.
Bug:
32072923
Test: ./test.py --host -j40
Change-Id: I5b05e23aaf9e90101817ca01d1928a30f14cb7e5
Andreas Gampe [Fri, 14 Apr 2017 21:43:01 +0000 (14:43 -0700)]
ART: Add debug code to run-test 913
Try to figure out why we get wrong depth reporting by aborting in
a deep frame.
Bug:
32072923
Test: m
Test: art/test/testrunner/testrunner.py --host -t 913
Change-Id: Icea02ebdee0e5395a5935c58c80e6d7c82eb6c6d
Alex Light [Fri, 14 Apr 2017 20:20:12 +0000 (13:20 -0700)]
More refactor for JVMTI redefinition run tests (2/3)
Move refactored tests into their final places.
NOTE: BROKEN WITHOUT A FOLLOW-UP CL TO ADD NEW MAIN.JAVA FILES.
Bug:
32072923
Test: None
Change-Id: I9350b2505815c5d21816df4189482aee4868c91c
Alex Light [Fri, 14 Apr 2017 20:17:26 +0000 (13:17 -0700)]
More refactor for JVMTI redefinition run tests (1/3)
Does as much as possible without making git get confused about the
provenance of files. Follow up CLs will move the files into the
appropriate places.
NOTE: BROKEN WITHOUT A FOLLOW-UP CL TO CHANGE FILE NAMES!
Bug:
32072923
Test: None
Change-Id: I1d66b42a5041d4ea1a0e8eaa61d55422a507e26f
Aart Bik [Fri, 14 Apr 2017 19:00:15 +0000 (12:00 -0700)]
Fixed bug on pending environment use of termination condition.
With regression test.
Test: test-art-host
Bug:
37247891
Change-Id: I55b06939d465d3ddb736d1ba659b1df179a5c390
Alex Light [Thu, 13 Apr 2017 17:49:45 +0000 (10:49 -0700)]
Remove test skips associated with b/
37240685
It's been fixed.
Test: ./test.py --host --jvmti-stress -j40
Change-Id: Ic1bb74dc62a080f54f93cec6fd75d748a9941653
Treehugger Robot [Fri, 14 Apr 2017 16:17:10 +0000 (16:17 +0000)]
Merge "ART: Refactor run-test 913"
Treehugger Robot [Fri, 14 Apr 2017 13:51:03 +0000 (13:51 +0000)]
Merge "Disable 160-read-barrier-stress temporarily"
Sebastien Hertz [Fri, 14 Apr 2017 12:18:36 +0000 (14:18 +0200)]
JDWP: fix Dbg::ResumeThread
A thread should be resumed if and only if it has been suspended by
the debugger. Therefore, we must check the debug suspend count, not
the suspend count (which includes debug suspend count). Otherwise we
could end up resuming a thread that has been suspended for another
reason (like a GC) but not by the debugger itself.
Bug:
27385848
Test: art/tools/run-jdwp-tests.sh --mode=host --variant=X64
Change-Id: I70ed6dfc233501e92332634650babf21ba911048
Sebastien Hertz [Fri, 14 Apr 2017 10:16:06 +0000 (12:16 +0200)]
Disable 160-read-barrier-stress temporarily
This test is causing an OOME with Jack compiler. This CL disables it
temporarily until we find a fix.
Bug:
37335480
Test: make -j test-art-host-run-test
Change-Id: I580fc99fbec0b0c94c25256fe20b957e07106205
Andreas Gampe [Thu, 13 Apr 2017 23:52:23 +0000 (16:52 -0700)]
ART: Refactor run-test 913
Remove references to ART. Filter roots from the JIT. Canonicalize
some thread IDs. Move the test to its own thread, and filter
stack-locals of other threads.
These changes ensure that the test is less dependent on the main
thread and its environment, which is required to use it in CTS.
Also fix a reporting issue for roots.
Bug:
32072923
Test: art/test/testrunner/testrunner.py --host -t 913
Change-Id: I8480ba7751fb6420c256db87cba11b8a65e25ea5
Treehugger Robot [Fri, 14 Apr 2017 00:10:36 +0000 (00:10 +0000)]
Merge "ART: Use _exit in dex2oat"
Treehugger Robot [Thu, 13 Apr 2017 23:54:34 +0000 (23:54 +0000)]
Merge "Remove native printing from test 980 and reliance on print internals"
Treehugger Robot [Thu, 13 Apr 2017 23:24:21 +0000 (23:24 +0000)]
Merge "Add check that map offset is sane to DexFile object init."
Mathieu Chartier [Thu, 13 Apr 2017 23:21:54 +0000 (23:21 +0000)]
Merge "Various debug ART GC performance improvements"
Treehugger Robot [Thu, 13 Apr 2017 23:16:11 +0000 (23:16 +0000)]
Merge "Add check that dexlayout can open output file."
Alex Light [Thu, 13 Apr 2017 19:58:06 +0000 (12:58 -0700)]
Remove native printing from test 980 and reliance on print internals
In preparation for a making a CTS test out of this make the test not
rely on internal allocation patterns of the 'System.out' object and
not make use of std::cout.
Test: ./test.py --host -j40
Change-Id: Ib7e874aaec71f93e834cf94ac5fe96663536691a
Jeff Hao [Thu, 13 Apr 2017 21:36:29 +0000 (14:36 -0700)]
Add check that map offset is sane to DexFile object init.
The DexFile constructor calls InitializeSectionsFromMapList, which
uses the header's map_offset before any checks are done on the file.
Bug:
37235346
Test: mm test-art-host-gtest-dex_file_test
Change-Id: I4fb8fcb57f9ef7e0182965b7ce663424b953abcb
Mathieu Chartier [Thu, 13 Apr 2017 18:47:53 +0000 (11:47 -0700)]
Various debug ART GC performance improvements
Removed unnecessary read barriers in SizeOf and a few other places.
Disabled disallow read barrier check.
Before:
GC time: 15.817s
Real 0m26.113s
user 1m16.780s
sys 0m3.152s
After:
GC time: 9.212s
real 0m19.875s
user 1m9.916s
sys 0m1.916s
Bug:
35644369
Test: test-art-host
Change-Id: I79a65259deff2a478a96e02ae69b14730b6dcbe6
Treehugger Robot [Thu, 13 Apr 2017 19:55:19 +0000 (19:55 +0000)]
Merge changes from topic 'art_run_test_910_cts'
* changes:
ART: Refactor run-test 910
ART: Fix GetArgumentsSize
Treehugger Robot [Thu, 13 Apr 2017 19:31:02 +0000 (19:31 +0000)]
Merge "MIPS32: java.lang.System.arraycopy() for copying char[] to char[]."
Andreas Gampe [Thu, 13 Apr 2017 19:20:42 +0000 (12:20 -0700)]
ART: Use _exit in dex2oat
That's what we really wanted to call. Seems to save about 50ms when
compiling something small-ish (out of .8s wall-clock time).
Test: m test-art-host
Change-Id: I9a20451adeaf63f535a7e42b46b4db78ee1d2bea
Andreas Gampe [Thu, 13 Apr 2017 03:49:38 +0000 (20:49 -0700)]
ART: Refactor run-test 910
Ensure a stable name for the proxy to be tested.
Bug:
32072923
Test: art/test/testrunner/testrunner.py --host -t 910
Change-Id: I34aedcdfae55f0bd0563d8aff0b8a8fdeb08de99
Andreas Gampe [Thu, 13 Apr 2017 04:11:28 +0000 (21:11 -0700)]
ART: Fix GetArgumentsSize
Fix the copy-paste mistake in GetArgumentsSize for abstract and
proxy methods. Use the usual computation derived from the shorty.
Bug:
34163329
Test: art/test/testrunner/testrunner.py --host -t 910
Change-Id: I71eaab70063303bea8eaa391ea067ebf8027820f
Jeff Hao [Wed, 12 Apr 2017 23:14:54 +0000 (16:14 -0700)]
Add check that dexlayout can open output file.
Test: mm test-art-host
Change-Id: I49b37b3e3696c8c002e22533c2e86e47a0e13e51
Treehugger Robot [Thu, 13 Apr 2017 17:16:12 +0000 (17:16 +0000)]
Merge "ART: Refactor run-test 904"
David Sehr [Thu, 13 Apr 2017 16:39:30 +0000 (16:39 +0000)]
Merge "Add OAT file mapping to dexdiag"
Chris Larsen [Thu, 23 Mar 2017 22:37:03 +0000 (15:37 -0700)]
MIPS32: java.lang.System.arraycopy() for copying char[] to char[].
Test: run-test --no-prebuild --optimizing 011-array-copy
Test: run-test --no-prebuild 011-array-copy
Test: run-test --optimizing 011-array-copy
Test: run-test 011-array-copy
Test: mma -j2 ART_TEST_OPTIMIZING=true test-art-target-run-test
Test: booted MIPS32R2 emulator.
Note: Tested against both the MIPS32R2, and MIPS64R6 emulators.
Change-Id: I0fc59fe8a242692b52cf8e8413d28397e210bb65
Roland Levillain [Thu, 13 Apr 2017 13:21:01 +0000 (14:21 +0100)]
Make ART build/test configurations use the CC collector by default.
As the concurrent copying (CC) garbage collector is currently the
default GC in ART, set `ART_USE_READ_BARRIER` to `true` on most
build/test configurations, with the exception of configurations meant
to explicitly exercise a specific collector, GC configuration, heap
poisoning, or a build using Valgrind.
Test: Rely on ab/master-art-host builds.
Bug:
12687968
Change-Id: I6ef7d3ad668aa2d608f7e70794fb110aa65375ea
Treehugger Robot [Thu, 13 Apr 2017 14:38:54 +0000 (14:38 +0000)]
Merge "ARM/VIXL: Fix TypeConversion of large long constants to int."
Vladimir Marko [Thu, 13 Apr 2017 10:50:14 +0000 (11:50 +0100)]
ARM/VIXL: Fix TypeConversion of large long constants to int.
The ARM VIXL backend was using Int32ConstantFrom() to get
the value of a HLongConstant for TypeConversion to int.
That function has DCHECK()s that the constant fits into a
32-bit int which is bogus in this particular case.
This was out of sync with the old ARM backend, so bring the
two in sync by using the approach from the old backend.
Bug:
37236345
Test: 646-checker-lon-const-to-int
Change-Id: I675ebd7bbaa14d5dfc3f280f62867d342cbab683
Vladimir Marko [Thu, 13 Apr 2017 12:54:22 +0000 (12:54 +0000)]
Merge "ARM64: Use link-time generated thunks for Baker CC read barrier."
Vladimir Marko [Thu, 13 Apr 2017 12:43:43 +0000 (13:43 +0100)]
Fix Arena allocation tracking.
Add missing allocation kind name and try to avoid
bit-rot in future by explicitly instantiating
ArenaAllocatorStatsImpl<true> in debug builds.
Bug:
34083438
Test: Rely on TreeHugger.
Change-Id: Ibe78a9f15d04b2ca2f03143fb4f49c01ae18471e
Treehugger Robot [Thu, 13 Apr 2017 11:17:06 +0000 (11:17 +0000)]
Merge "MIPS32: Add direct calls to math intrinsics."
Andreas Gampe [Thu, 13 Apr 2017 04:58:19 +0000 (04:58 +0000)]
Merge "ART: Fix test dependencies"
Andreas Gampe [Thu, 13 Apr 2017 04:53:55 +0000 (21:53 -0700)]
ART: Fix test dependencies
Follow-up to commit
8f2c6d479aeda3165c4f54c1076ec7ac335fccc7.
Test: m test-art-host
Change-Id: I67bf92067509ddf658fc22e333430fa5ee44f6ea
Andreas Gampe [Thu, 13 Apr 2017 02:06:12 +0000 (19:06 -0700)]
ART: Refactor run-test 904
Refactor run-test so it does not depend on printf output. That way
we can run it in CTS.
Bug:
32072923
Test: art/test/testrunner/testrunner.py -b --host -t 904
Change-Id: Ia0d2e01dba4f46b81e78c88d9c50aafbc69f679e
Treehugger Robot [Thu, 13 Apr 2017 01:32:51 +0000 (01:32 +0000)]
Merge changes I49c02b92,I898e5cff
* changes:
Ensure one can call DisposeEnvironment during event callbacks.
run-tests with jvmti-stress configuration
Alex Light [Wed, 12 Apr 2017 23:13:33 +0000 (16:13 -0700)]
Ensure one can call DisposeEnvironment during event callbacks.
Previously calling DisposeEnvironment during an event callback would
cause a currently in-use iterator to be invalidated. This could cause
undefined behavior.
Bug:
37283252
Test: ./test/testrunner/testrunner.py --host --jvmti-stress -j40
Change-Id: I49c02b925f0da5d4f66965f0fca21bf8dc83132a
Alex Light [Mon, 10 Apr 2017 23:27:35 +0000 (16:27 -0700)]
run-tests with jvmti-stress configuration
Adds a jvmti-stress mode to our run-tests. This mode runs our test
suite with every class-load being intercepted and undergoing
redefinition.
Test: DEXTER_BINARY="/path/to/built/dexter" \
./testrunner/testrunner.py --host --jvmti-stress -j40
Change-Id: I898e5cffea42a203aa4c45981b48baf8dc64fbb6
David Sehr [Wed, 5 Apr 2017 21:43:38 +0000 (14:43 -0700)]
Add OAT file mapping to dexdiag
Adds the ability to visualize the executable mapped sections of an OAT
file.
Bug:
37217968
Test: test-art-target-gtest-dexdiag_test
Change-Id: Id9f33885ee38a3753e781f955127eb8e1b83fe4d
Treehugger Robot [Wed, 12 Apr 2017 18:39:45 +0000 (18:39 +0000)]
Merge "MIPS64: Add direct calls to math intrinsics."
Chris Larsen [Fri, 24 Mar 2017 19:11:54 +0000 (12:11 -0700)]
MIPS32: Add direct calls to math intrinsics.
- cos - sin - acos - asin
- atan - atan2 - cbrt - cosh
- exp - expm1 - hypot - log
- log10 - nextAfter - sinh - tan
- tanh
Test: run-test --no-prebuild --optimizing 123-inline-execute2
Test: run-test --no-prebuild 123-inline-execute2
Test: run-test --optimizing 123-inline-execute2
Test: run-test 123-inline-execute2
Test: Booted MIPS32R2 and MIPS64R6 emulators.
Note: Ran tests under both emulators, and on CI20 hardware.
Change-Id: I4c1782d609a985ae49737f47c43caa5b0b6a10e3
Treehugger Robot [Wed, 12 Apr 2017 17:56:13 +0000 (17:56 +0000)]
Merge "Fix broken DCHECK."
Mathieu Chartier [Wed, 12 Apr 2017 17:28:38 +0000 (17:28 +0000)]
Merge "Clear zygote space cards for CC"
Mathieu Chartier [Wed, 12 Apr 2017 17:27:39 +0000 (17:27 +0000)]
Merge "Fix -XX:DumpGCPerformanceOnShutdown for debug builds"
Vladimir Marko [Mon, 20 Mar 2017 18:26:59 +0000 (18:26 +0000)]
ARM64: Use link-time generated thunks for Baker CC read barrier.
Remaining work for follow-up CLs:
- array loads,
- volatile field loads,
- use implicit null check in field thunk.
Test: Added tests to relative_patcher_arm64
Test: New run-test 160-read-barrier-stress
Test: m test-art-target-gtest on Nexus 6P.
Test: testrunner.py --target on Nexus 6P.
Bug:
29516974
Bug:
30126666
Bug:
36141117
Change-Id: Id68ff171c55a3f1bf1ac1b657f480531aa7b3710
Alex Light [Wed, 12 Apr 2017 16:01:47 +0000 (09:01 -0700)]
Fix broken DCHECK.
The DCHECK in ti_class_definition was incorrectly checking that the
original dex file was a primitive long instead of a boxed long.
Bug:
37271822
Test: ./test/testrunner/testrunner.py --host --jvmti-stress -j40
Change-Id: I901c6e2f6f93dca584faff70134cee957d8462c3
Aart Bik [Wed, 12 Apr 2017 15:52:08 +0000 (15:52 +0000)]
Merge changes I1d4db176,Ifb931a99
* changes:
ARM64: Support vectorization for double and long.
ARM64: Support 128-bit registers for SIMD.
Treehugger Robot [Wed, 12 Apr 2017 12:17:44 +0000 (12:17 +0000)]
Merge "MIPS: Use Lsa/Dlsa when possible."
Richard Uhler [Wed, 12 Apr 2017 07:44:46 +0000 (07:44 +0000)]
Merge "Add an 'art-preopt' test target configuration."
Andreas Gampe [Wed, 12 Apr 2017 04:07:43 +0000 (04:07 +0000)]
Merge changes Ic7d75e1e,Iaf828034
* changes:
ART: Fix GetSystemProperty (2)
ART: Fix GetSystemProperty
Mathieu Chartier [Wed, 12 Apr 2017 00:50:00 +0000 (17:50 -0700)]
Fix -XX:DumpGCPerformanceOnShutdown for debug builds
There was a failing DCHECK from CheckUnattachedThread. The fix is
to dump after attaching the shutdown thread.
Bug:
35644369
Test: test-art-host
Change-Id: I3d927e380888418167c101b2f09d1e547fe728cf