OSDN Git Service
Andreas Gampe [Wed, 29 Jul 2015 02:51:37 +0000 (19:51 -0700)]
ART: Use bionic TLS slot for thread-self
Use a private bionic TLS slot to store self instead of using
pthreads.
Change-Id: Icc86a2b7590734637366f9d5e41a5c6d18cc5772
Andreas Gampe [Tue, 28 Jul 2015 15:30:36 +0000 (15:30 +0000)]
Merge "ART: Print out hex for NewStringUTF failure"
Calin Juravle [Tue, 28 Jul 2015 11:54:11 +0000 (11:54 +0000)]
Merge "Revert "Revert "Revert "Fixes and improvements in ReferenceTypePropagation""""
Calin Juravle [Tue, 28 Jul 2015 11:53:36 +0000 (11:53 +0000)]
Merge "Revert "Revert "Revert "Use the object class as top in reference type propagation""""
Calin Juravle [Tue, 28 Jul 2015 11:52:23 +0000 (11:52 +0000)]
Revert "Revert "Revert "Fixes and improvements in ReferenceTypePropagation"""
This reverts commit
00e3b38be4b280d6d7a7e843cd336ffbd2ba4365.
Change-Id: I4dbadb2d7312a410f1c56283f063dd82156cf702
Calin Juravle [Tue, 28 Jul 2015 11:52:02 +0000 (11:52 +0000)]
Revert "Revert "Revert "Use the object class as top in reference type propagation"""
This reverts commit
80caa1478cf3df4eac1214d8a63a4da6f4fe622b.
Change-Id: I63b51ca418b19b2bfb5ede3f8444f8fbeb8a339d
Calin Juravle [Tue, 28 Jul 2015 10:44:36 +0000 (10:44 +0000)]
Merge "Revert "Revert "Use the object class as top in reference type propagation"""
Calin Juravle [Tue, 28 Jul 2015 10:44:27 +0000 (10:44 +0000)]
Merge "Revert "Revert "Fixes and improvements in ReferenceTypePropagation"""
Roland Levillain [Tue, 28 Jul 2015 10:11:04 +0000 (10:11 +0000)]
Merge "Added repe_cmpsw instruction to x86, x86_64 assemblers"
Andreas Gampe [Tue, 28 Jul 2015 04:03:25 +0000 (21:03 -0700)]
ART: Print out hex for NewStringUTF failure
Print out the input data as a hex stream in case of a NewStringUTF
input format failure.
Bug:
22773761
Change-Id: I99a275bebb89564b2d71f297c7f5b9543cf4312d
Mathieu Chartier [Mon, 27 Jul 2015 23:19:31 +0000 (23:19 +0000)]
Merge "Add remembered set functionality for native roots"
Mathieu Chartier [Mon, 27 Jul 2015 23:08:02 +0000 (16:08 -0700)]
Add remembered set functionality for native roots
Fixes GSS tests. The issue was that there could be classes which had
native roots in the target space which we need to mark and update.
Change-Id: I33e628d40bcb74f768492f740f8f084077d80edf
Mathieu Chartier [Mon, 27 Jul 2015 23:15:59 +0000 (23:15 +0000)]
Merge "Bugfix: Java char is 16 bits, can not be treated as boolean."
Fredrik Roubert [Mon, 27 Jul 2015 22:14:08 +0000 (00:14 +0200)]
Bugfix: Java char is 16 bits, can not be treated as boolean.
Using SetFieldBooleanVolatile() and SetFieldBoolean() happens to work
for char values that only use the lower 8 bits, but is a mistake that
was introduced by the "Add AccessibleObject and Field to mirror" commit:
https://android.googlesource.com/platform/art/+/
daaf326
Change-Id: Iec02ba3084c992ea239ecef688d7d29c7e21ae16
agicsaki [Mon, 27 Jul 2015 18:34:13 +0000 (11:34 -0700)]
Added repe_cmpsw instruction to x86, x86_64 assemblers
Change-Id: I7634959eebb64d607f47497db320d5c2afdef16b
Mingyao Yang [Mon, 27 Jul 2015 20:20:01 +0000 (20:20 +0000)]
Merge "HDeoptimize should hold values live in env."
Calin Juravle [Thu, 16 Jul 2015 15:51:30 +0000 (16:51 +0100)]
Revert "Revert "Use the object class as top in reference type propagation""
This reverts commit
7733bd644ac71f86d4b30a319624b23343882e53.
Change-Id: I7d393a808c01c084c18d632a54e0554b4b455f2c
Calin Juravle [Wed, 15 Jul 2015 13:41:29 +0000 (14:41 +0100)]
Revert "Revert "Fixes and improvements in ReferenceTypePropagation""
This reverts commit
9b0096ba77e7e61bc2dcbbf954831dcae54a6c27.
Change-Id: I824f16e800ca32e646577d5e1e0d593887ccead1
Mingyao Yang [Wed, 22 Jul 2015 22:56:34 +0000 (15:56 -0700)]
HDeoptimize should hold values live in env.
Values that are not live in compiled code anymore may still be needed in
interpreter, due to code motion, etc.
Bug:
22665511
Change-Id: I8b85833c5c462f8fe36f86d6026a51b07563995a
Aart Bik [Mon, 27 Jul 2015 16:34:39 +0000 (16:34 +0000)]
Merge "Add checker ability to 525-arrays-and-fields tests."
Roland Levillain [Mon, 27 Jul 2015 15:31:51 +0000 (15:31 +0000)]
Merge "Exercise constant folding of int & long mul, div and rem."
Roland Levillain [Mon, 27 Jul 2015 15:29:58 +0000 (15:29 +0000)]
Merge "Cosmetic changes in ART run-tests."
Roland Levillain [Mon, 27 Jul 2015 15:28:20 +0000 (15:28 +0000)]
Merge "Fix & add Checker assertions in 442-checker-constant-folding."
Roland Levillain [Wed, 22 Jul 2015 13:12:01 +0000 (14:12 +0100)]
Exercise constant folding of int & long mul, div and rem.
Change-Id: Ic70b56dbcb4a660a4c8ca04f7b2b5201ad29952a
Roland Levillain [Thu, 23 Jul 2015 11:15:01 +0000 (12:15 +0100)]
Cosmetic changes in ART run-tests.
Perform a copyright headers clean up mostly.
Change-Id: I26c855b2f54d0887676d9b9b8281b14b7978ab29
Roland Levillain [Thu, 23 Jul 2015 11:11:42 +0000 (12:11 +0100)]
Fix & add Checker assertions in 442-checker-constant-folding.
Change-Id: Id4165b14816c04961d5c274f24861d0e7856d95e
Calin Juravle [Mon, 27 Jul 2015 10:39:13 +0000 (10:39 +0000)]
Merge "Enable art host static builds only for checkbuilds."
Andreas Gampe [Sun, 26 Jul 2015 17:31:59 +0000 (17:31 +0000)]
Merge "ART: Abbreviate interpreter-access-checks"
Andreas Gampe [Sun, 26 Jul 2015 17:26:03 +0000 (10:26 -0700)]
ART: Abbreviate interpreter-access-checks
Tag is too long.
Change-Id: I49febb97cec1172bcf06438fe8e0605f266e3fde
Mathieu Chartier [Fri, 24 Jul 2015 23:17:16 +0000 (23:17 +0000)]
Merge "Fix gcstress tests"
Mathieu Chartier [Fri, 24 Jul 2015 23:05:55 +0000 (16:05 -0700)]
Fix gcstress tests
We may visit the native roots multiple times (e.g. temp classes).
This used to not be handled by the SS GC but is now handle by
MarkObjectIfNotInToSpace.
Change-Id: I18226e20b36ace41c155ba4c3f54eac2244f524c
Andreas Gampe [Fri, 24 Jul 2015 22:56:43 +0000 (22:56 +0000)]
Merge "ART: Make warnings into errors in InitZygote"
Andreas Gampe [Fri, 24 Jul 2015 22:41:20 +0000 (15:41 -0700)]
ART: Make warnings into errors in InitZygote
These warnings are actual errors and lead to failure to start the
runtime. Print them accordingly.
Change-Id: I907e32a3c5da5364ecc9678bb9a11604234d669c
Andreas Gampe [Fri, 24 Jul 2015 22:25:01 +0000 (22:25 +0000)]
Merge "ART: Change the stack dump format to be in line with debuggerd"
Andreas Gampe [Thu, 23 Jul 2015 21:37:18 +0000 (14:37 -0700)]
ART: Change the stack dump format to be in line with debuggerd
Make offsets 16 digits on 64-bit platforms.
Add the ability to provide the build fingerprint, and print it in
the traces output.
Bug:
22693991
(cherry picked from commit
242ae94dc39be53e519a6163ea1db494ceb75954)
Change-Id: Ibd3d3b3bd65dce84acfb97a487807d6f280a9508
Igor Murashkin [Fri, 24 Jul 2015 22:01:21 +0000 (22:01 +0000)]
Merge "Revert "runtest: Have gcstress tests skip 955-lambda-smali""
Igor Murashkin [Fri, 24 Jul 2015 21:43:26 +0000 (14:43 -0700)]
Revert "runtest: Have gcstress tests skip 955-lambda-smali"
This reverts commit
ef3119e6c75c431d9cf54e92dfe5b805597b1657.
Mathieu Chartier [Fri, 24 Jul 2015 18:58:57 +0000 (18:58 +0000)]
Merge "Visit class native roots from VisitReferences"
Mathieu Chartier [Thu, 23 Jul 2015 23:01:26 +0000 (16:01 -0700)]
Visit class native roots from VisitReferences
Visit class roots when we call Class::VisitReferences instead of in
the class linker. This makes it easier to implement class unloading
since unmarked classes won't have their roots visited by the class
linker.
Bug:
22181835
Change-Id: I63f31e5ebef7b2a0b764b3ba3cb038b3f561b379
Aart Bik [Thu, 23 Jul 2015 21:15:22 +0000 (14:15 -0700)]
Add checker ability to 525-arrays-and-fields tests.
Rationale:
Good to verify hoisting occurs and even some
not-hoisting (although the latter is already
reflected in the computed results).
Change-Id: I23575e4fd480836b436b1f27f67e561c7b40419d
Calin Juravle [Fri, 24 Jul 2015 15:45:26 +0000 (16:45 +0100)]
Enable art host static builds only for checkbuilds.
Change-Id: I42cc850089a5ecbcc201699f0b996da4fd959bb3
Mathieu Chartier [Thu, 23 Jul 2015 21:56:01 +0000 (21:56 +0000)]
Merge "Add some more uninterruptible annotations"
Mathieu Chartier [Thu, 23 Jul 2015 21:11:16 +0000 (14:11 -0700)]
Add some more uninterruptible annotations
Added to mirror sites, class linker, and intern table.
Bug:
20072211
Change-Id: I2224bcefefaab05d325f14844b1514e3d719e649
Mathieu Chartier [Thu, 23 Jul 2015 18:59:17 +0000 (18:59 +0000)]
Merge "Fix alignments in quick_entrypoints_x86_64.S"
Man Cao [Thu, 23 Jul 2015 18:18:49 +0000 (11:18 -0700)]
Fix alignments in quick_entrypoints_x86_64.S
Places calling artIsAssignableFromCode() were not 16-byte aligned.
Change-Id: I86ff4f73a942ede09c0206e76614eb826dd896c2
Vladimir Marko [Thu, 23 Jul 2015 16:53:02 +0000 (16:53 +0000)]
Merge "ART: Fix Quick/Optimizing suspend check assumption mismatch."
David Brazdil [Thu, 23 Jul 2015 16:18:59 +0000 (16:18 +0000)]
Merge "Disable Checker for tests that don't invoke the compiler."
Vladimir Marko [Wed, 22 Jul 2015 16:50:37 +0000 (17:50 +0100)]
ART: Fix Quick/Optimizing suspend check assumption mismatch.
Quick's SuspendCheckElimination (SCE) expects that every
method contains a suspend check and it eliminates suspend
checks in loops containing an invoke. Optimizing eliminates
the suspend check from leaf methods, so the combination of
a Quick-compiled loop calling an Optimizing-compiled leaf
method can lead to missing suspend checks and potentially
leading to ANRs.
Enable Quick's kLeafOptimization flag to remove suspend
checks from leaf methods and disable Quick's SCE. This
aligns the suspend check placement for the two backends
and avoids the broken combination.
Currently, all methods containing a try-catch are compiled
with Quick, so it's relatively easy to create a regression
test. However, this test will not be valid when Optimizing
starts supporting try-catch.
Bug:
22657404
Change-Id: I3bc40bf3f5c1e7d18704d1547b139e939950b770
David Brazdil [Thu, 23 Jul 2015 10:53:42 +0000 (11:53 +0100)]
Disable Checker for tests that don't invoke the compiler.
Running tests in no-prebuild mode will resort to running the
interpreter when either dex2oat or patchoat are not available, thus
not producing the CFG dump tested by Checker. This patch disables
Checker under such configurations.
Bug:
22552692
Change-Id: I23c97aa0aa627675861ef84f8c570303b614d9af
Roland Levillain [Thu, 23 Jul 2015 10:26:14 +0000 (10:26 +0000)]
Merge "Fix and extend the documentation of art::Thread's fields offsets."
Andreas Gampe [Thu, 23 Jul 2015 08:14:14 +0000 (08:14 +0000)]
Merge "ART: Fix missing role definition"
Andreas Gampe [Thu, 23 Jul 2015 06:44:55 +0000 (23:44 -0700)]
ART: Fix missing role definition
Add missing static member definition. Follow-up to
4e2cb098017bf073335ebb02b1bc0a36828cd720
Bug:
20072211
Change-Id: I392da3724ded7c455b9b57add6fceb70191f9ff6
Mathieu Chartier [Thu, 23 Jul 2015 03:14:15 +0000 (03:14 +0000)]
Merge "Fix hash set memory leaks"
Mathieu Chartier [Thu, 23 Jul 2015 02:39:56 +0000 (19:39 -0700)]
Fix hash set memory leaks
There was a memory leak if ShrinkToMaximumLoad caused the set to
become 0 size. Fixes valgrind tests.
Change-Id: Ibdd033768d622443fb34e7a5e5e1c00e082a1a4a
Mathieu Chartier [Thu, 23 Jul 2015 00:49:51 +0000 (00:49 +0000)]
Merge "Add uninterruptible role"
Mathieu Chartier [Wed, 22 Jul 2015 23:17:51 +0000 (16:17 -0700)]
Add uninterruptible role
Example error:
cannot call function 'CopyOf' while mutex 'uninterruptible_' is held
TODO: Add annotation to more locations.
Bug:
20072211
Change-Id: I1bbf5a77e3deeafa5898df529cb7cb53a6d010d2
Mathieu Chartier [Wed, 22 Jul 2015 22:19:19 +0000 (22:19 +0000)]
Merge "Move to newer clang annotations"
Igor Murashkin [Wed, 22 Jul 2015 22:18:42 +0000 (22:18 +0000)]
Merge "runtest: Have gcstress tests skip 955-lambda-smali"
Igor Murashkin [Wed, 22 Jul 2015 22:15:06 +0000 (15:15 -0700)]
runtest: Have gcstress tests skip 955-lambda-smali
This prevents the test from failing due to a timeout (temporarily).
Bug:
22665783
Change-Id: I978786ba2f94de8f13e32718a16e6909820dab51
Mathieu Chartier [Fri, 17 Jul 2015 03:32:27 +0000 (20:32 -0700)]
Move to newer clang annotations
Also enable -Wthread-safety-negative.
Changes:
Switch to capabilities and negative capabilities.
Future work:
Use capabilities to implement uninterruptible annotations to work
with AssertNoThreadSuspension.
Bug:
20072211
Change-Id: I42fcbe0300d98a831c89d1eff3ecd5a7e99ebf33
Andreas Gampe [Wed, 22 Jul 2015 21:45:44 +0000 (21:45 +0000)]
Merge "ART: Add VerboseMethods to optimizing compiler"
Andreas Gampe [Wed, 22 Jul 2015 19:10:13 +0000 (12:10 -0700)]
ART: Add VerboseMethods to optimizing compiler
Bring the optimizing compiler up to Quick levels for debuggability.
Change-Id: Ib2adce7a31cc6a60c37c63d0df5cb6a35a85e790
Richard Uhler [Wed, 22 Jul 2015 19:24:28 +0000 (19:24 +0000)]
Merge "More diagnostics for flaky oat file assistant test."
Aart Bik [Wed, 22 Jul 2015 18:33:17 +0000 (18:33 +0000)]
Merge "Replaced CHECK with EXPECT in tests."
Aart Bik [Wed, 22 Jul 2015 18:27:39 +0000 (11:27 -0700)]
Replaced CHECK with EXPECT in tests.
Doh!
Change-Id: Ibef45031146215034647c7eaab4065c766d6fb86
Calin Juravle [Wed, 22 Jul 2015 17:20:08 +0000 (17:20 +0000)]
Merge "Revert "Fixes and improvements in ReferenceTypePropagation""
Calin Juravle [Wed, 22 Jul 2015 17:19:20 +0000 (17:19 +0000)]
Merge "Revert "Use the object class as top in reference type propagation""
Calin Juravle [Wed, 22 Jul 2015 17:14:50 +0000 (17:14 +0000)]
Revert "Use the object class as top in reference type propagation"
This reverts commit
3fabec7a25d151b26ba7de13615bbead0dd615a6.
Change-Id: Id8614f6b6e3e0e4c9caeb9f771e4c145d9fec64f
Calin Juravle [Wed, 22 Jul 2015 17:14:32 +0000 (17:14 +0000)]
Revert "Fixes and improvements in ReferenceTypePropagation"
This reverts commit
b0d5fc0ac139da4aaa1440263416b9bde05630b0.
Change-Id: Iea8adfc0bd4cb7ee2b292278b8bac80a259acbd1
Vladimir Marko [Wed, 22 Jul 2015 16:32:16 +0000 (16:32 +0000)]
Merge "ART: Fix Thumb2 literal fixup."
Roland Levillain [Wed, 22 Jul 2015 16:19:28 +0000 (17:19 +0100)]
Fix and extend the documentation of art::Thread's fields offsets.
Change-Id: Ie13c71d1b734ec0778c388b9bcb7773dcf319f16
Aart Bik [Wed, 22 Jul 2015 16:13:43 +0000 (16:13 +0000)]
Merge "Added unit tests to LICM."
Richard Uhler [Wed, 22 Jul 2015 15:25:11 +0000 (08:25 -0700)]
More diagnostics for flaky oat file assistant test.
When we generate an odex file for test, verify it is generated properly, and
if it is not supposed to be generated pic, verify it needs relocation right
after generating the odex file.
Bug:
22599792
Change-Id: I75fbbaafc7a425e20fe3495657e1c38ae7b15d56
Roland Levillain [Wed, 22 Jul 2015 15:18:42 +0000 (15:18 +0000)]
Merge "When a Checker assertion cannot be parsed, display it."
Roland Levillain [Wed, 22 Jul 2015 14:42:54 +0000 (14:42 +0000)]
Merge "Delete extraneous prefix `SideEffects::` in `nodes.h`."
Roland Levillain [Wed, 22 Jul 2015 12:37:27 +0000 (13:37 +0100)]
When a Checker assertion cannot be parsed, display it.
Change-Id: Ib667c7aa24b2753f0963ed6b1f9f4227a9e586f4
Alexandre Rames [Wed, 22 Jul 2015 10:32:58 +0000 (11:32 +0100)]
Delete extraneous prefix `SideEffects::` in `nodes.h`.
Change-Id: Ic0a8442d20323df0d9db9e6a1d26c07bd903a13d
Vladimir Marko [Wed, 22 Jul 2015 10:28:14 +0000 (11:28 +0100)]
ART: Fix Thumb2 literal fixup.
When processing a load literal during fixup, we use the
current code size to determine whether we need to add the
padding before literals. However, this may change by the end
of the fixup, yet we didn't recalculate to see if the load
literal is pushed out of range.
Instead of making the load literal fixup also depend on all
preceding fixups, add an extra pass over literals when we
need the padding and repeat the fixup loop if needed.
Change-Id: Ia21d660486167a2dcb1ad4afe8acc669b4af669d
Calin Juravle [Wed, 22 Jul 2015 11:20:14 +0000 (11:20 +0000)]
Merge "Use the object class as top in reference type propagation"
Calin Juravle [Wed, 22 Jul 2015 11:20:01 +0000 (11:20 +0000)]
Merge "Fixes and improvements in ReferenceTypePropagation"
Aart Bik [Wed, 22 Jul 2015 01:30:18 +0000 (18:30 -0700)]
Added unit tests to LICM.
Rationale:
LICM did not have its own unit test yet, and it was a good
time to ensure that the improved side effect analysis does
what it is supposed to. Also, a nice way for me to get
familiar with the testing infrastructure.
Change-Id: I16471e5fd99615c04a59b09c4a69126d7a94684f
Igor Murashkin [Wed, 22 Jul 2015 01:00:59 +0000 (01:00 +0000)]
Merge changes from topic 'lambda_experimental'
* changes:
interpreter: Add tests for move-result after invoke-lambda
verifier: Skip verification of methods when seeing experimental opcodes
Igor Murashkin [Wed, 22 Jul 2015 00:04:35 +0000 (17:04 -0700)]
interpreter: Add tests for move-result after invoke-lambda
Bug:
22638098
Change-Id: I9c428055eef0bc95f4f7aacd03be56547df812d9
Igor Murashkin [Wed, 22 Jul 2015 00:03:36 +0000 (17:03 -0700)]
verifier: Skip verification of methods when seeing experimental opcodes
Bug:
22638098
Change-Id: I9f172f3e0e7ad2aa8873e4036415702fee6bf2eb
Ying Wang [Tue, 21 Jul 2015 23:45:21 +0000 (23:45 +0000)]
Merge "Prefer 32-bit host executables on Mac."
Richard Uhler [Tue, 21 Jul 2015 20:27:18 +0000 (20:27 +0000)]
Merge "Add more checks to diagnose flaky oat file assistant test."
Ying Wang [Tue, 21 Jul 2015 19:05:30 +0000 (12:05 -0700)]
Prefer 32-bit host executables on Mac.
Bug:
22527295
Change-Id: I2cb1d3086b65b8227e6bdac9c0485fdc452f1e9c
Andreas Gampe [Tue, 21 Jul 2015 17:56:40 +0000 (17:56 +0000)]
Merge "Refine the canary pattern in RedZone"
Andreas Gampe [Tue, 21 Jul 2015 16:32:22 +0000 (16:32 +0000)]
Merge "ART: Fix UninitializedReference handling"
Andreas Gampe [Tue, 14 Jul 2015 04:12:43 +0000 (21:12 -0700)]
ART: Fix UninitializedReference handling
The merge rules in the verifier allowed Object to be successfully
merged with uninitialized references. This is invalid and should
result in a conflict. Fix by moving UninitializedReference rules
earlier.
Also add a test that forward merging is correctly allowed, both
with a valid result as well as a conflict.
Also add tests that backwards branches have the expected behavior.
Bug:
22411633
Change-Id: If837376c15f0b3550d6ce1721a3cde5901c80c7f
Richard Uhler [Tue, 21 Jul 2015 16:08:10 +0000 (16:08 +0000)]
Merge "Fix bug in OatFileAssistant::GetBestOatFile."
Andreas Gampe [Tue, 21 Jul 2015 15:56:21 +0000 (15:56 +0000)]
Merge "ART: Remove some of the Mac craziness"
Andreas Gampe [Wed, 15 Jul 2015 06:34:44 +0000 (23:34 -0700)]
ART: Remove some of the Mac craziness
We rely on new-enough Clang/LLVM builds nowadays. The integrated
assembler supports named parameters. Throw away most of the
old duplication (effectively cutting support for older Clang
versions). The only required duplications are:
1) Clang as does not support .altmacro. However, the Clang
preprocessor works different wrt/ the GCC preprocessor
and does not give us trouble with inserted spaces.
2) On the Mac, symbols are prefixed with an underscore.
This should help to avoid breaking the Mac build when changing
the assembly code, and prepare for a complete Clang-only build
for x86 and x86-64. Switching to the integrated assembler for
the host build may be done in a follow-up CL.
Bug:
17443165
Change-Id: I1a077d4b612abc2b1b851c1bdabb5008a52e5aa6
Calin Juravle [Thu, 16 Jul 2015 15:51:30 +0000 (16:51 +0100)]
Use the object class as top in reference type propagation
This properly types all instructions, making it safe to query the type
at any time.
This also moves a few functions from class.h to class-inl.h to please
gcc linker when compiling for target.
Change-Id: I6b7ce965c10834c994b95529ab65a548515b4406
Calin Juravle [Wed, 15 Jul 2015 13:41:29 +0000 (14:41 +0100)]
Fixes and improvements in ReferenceTypePropagation
- Bound object types after a CheckCast. This increases the precision of
(inlining) generic operations.
- Make sure that the BoundType is exact when the class is final.
- Make sure that we don't duplicate BoundTypes when we run the analysis
more than once.
Change-Id: Ic22b610766fae101f942c0d753ddcac32ac1844a
Mathieu Chartier [Tue, 21 Jul 2015 01:20:16 +0000 (01:20 +0000)]
Merge "Don't check code pointer for proxies in ArtMethod::GetQuickFrameInfo"
Aart Bik [Tue, 21 Jul 2015 01:01:36 +0000 (01:01 +0000)]
Merge "Fix broken tests."
Aart Bik [Mon, 20 Jul 2015 21:08:59 +0000 (14:08 -0700)]
Fix broken tests.
Rationale:
(1) volatile field write/read need to apply to all
to comply with Java memory model
(2) clinit only needs only the write
(3) added conservative assumptions to memory barrier
(nothing broke, but this seems better)
Change-Id: I37787ec8f3f2c8d6166a94c57193fa4544ad3372
Igor Murashkin [Mon, 20 Jul 2015 22:36:04 +0000 (22:36 +0000)]
Merge "runtime: Add lambda box/unbox object equality"
Igor Murashkin [Fri, 10 Jul 2015 20:49:08 +0000 (13:49 -0700)]
runtime: Add lambda box/unbox object equality
A lambda that is boxed with box-lambda is now stored as a weak reference
in a global runtime table (lambda::BoxTable). Repeatedly boxing the same
lambda closure value will always return the same java.lang.Object back.
Since there is no way to observe the address of an object, a GC can
happen and clean up the table of any dead boxed lambdas, which can also
shrink the table to prevent the memory use from growing too much.
(Note that a lambda closure is immutable, so hashing over it is
guaranteed safe.)
Change-Id: I786c1323ff14eed937936b303d511875f9642524