OSDN Git Service

android-x86/art.git
8 years agoMerge "Change roots view to "rooted"."
Richard Uhler [Tue, 17 Nov 2015 16:21:12 +0000 (16:21 +0000)]
Merge "Change roots view to "rooted"."

8 years agoMerge "Add ability to skip run-tests with ART_TEST_RUN_TEST_SKIP."
Alex Light [Mon, 16 Nov 2015 18:10:37 +0000 (18:10 +0000)]
Merge "Add ability to skip run-tests with ART_TEST_RUN_TEST_SKIP."

8 years agoMerge "Ignore NullConstants when computing the type for Phis."
Calin Juravle [Mon, 16 Nov 2015 17:00:23 +0000 (17:00 +0000)]
Merge "Ignore NullConstants when computing the type for Phis."

8 years agoMerge "Implemented control of "export only" feature in dexdump."
Aart Bik [Mon, 16 Nov 2015 16:41:21 +0000 (16:41 +0000)]
Merge "Implemented control of "export only" feature in dexdump."

8 years agoMerge "Mark kQuickReadBarrierForRootSlow as direct entry point on MIPS."
Roland Levillain [Mon, 16 Nov 2015 16:09:48 +0000 (16:09 +0000)]
Merge "Mark kQuickReadBarrierForRootSlow as direct entry point on MIPS."

8 years agoMark kQuickReadBarrierForRootSlow as direct entry point on MIPS.
Roland Levillain [Mon, 16 Nov 2015 15:59:08 +0000 (15:59 +0000)]
Mark kQuickReadBarrierForRootSlow as direct entry point on MIPS.

Change-Id: I35caa32c9773646692706642188ab522237a68ac

8 years agoMerge "x86/x86-64 read barrier support for concurrent GC in Optimizing."
Roland Levillain [Mon, 16 Nov 2015 15:34:28 +0000 (15:34 +0000)]
Merge "x86/x86-64 read barrier support for concurrent GC in Optimizing."

8 years agoMerge "Move the trace file out of the profiles folder."
Calin Juravle [Mon, 16 Nov 2015 11:04:13 +0000 (11:04 +0000)]
Merge "Move the trace file out of the profiles folder."

8 years agox86/x86-64 read barrier support for concurrent GC in Optimizing.
Roland Levillain [Fri, 13 Nov 2015 10:07:31 +0000 (10:07 +0000)]
x86/x86-64 read barrier support for concurrent GC in Optimizing.

This first implementation uses slow paths to instrument heap
reference loads and GC root loads for the concurrent copying
collector, respectively calling the artReadBarrierSlow and
artReadBarrierForRootSlow (new) runtime entry points.

Notes:
- This implementation does not instrument HInvokeVirtual
  nor HInvokeInterface instructions (for class reference
  loads), as the corresponding read barriers are not stricly
  required with the current concurrent copying collector.
- Intrinsics which may eventually call (on slow path) are
  disabled when read barriers are enabled, as the current
  slow path infrastructure does not support this case.
- When read barriers are enabled, the code generated for a
  HArraySet instruction always go into the array set slow
  path for object arrays (delegating the operation to the
  runtime), as we are lacking a mechanism to keep a
  temporary register live accross a runtime call (needed for
  the instrumentation of type checking code, which requires
  two successive read barriers).

Bug: 12687968
Change-Id: I14cd6107233c326389120336f93955b28ffbb329

8 years agoMerge "Fix class unloading with the CC collector."
Hiroshi Yamauchi [Sat, 14 Nov 2015 02:15:09 +0000 (02:15 +0000)]
Merge "Fix class unloading with the CC collector."

8 years agoAdd ability to skip run-tests with ART_TEST_RUN_TEST_SKIP.
Alex Light [Fri, 13 Nov 2015 18:51:10 +0000 (10:51 -0800)]
Add ability to skip run-tests with ART_TEST_RUN_TEST_SKIP.

One can skip running specific run-tests under any configuration by
setting the ART_TEST_RUN_TEST_SKIP variable to a list of the run-test
names to skip.

    $ # None of the variants of 022-interfaces will be run.
    $ ART_TEST_RUN_TEST_SKIP=022-interface mm test-art-host-run-test

Change-Id: I33f24dd7f1eb9f4d6dc407a4abc3c5919437d4fa

8 years agoFix class unloading with the CC collector.
Hiroshi Yamauchi [Thu, 12 Nov 2015 07:45:34 +0000 (23:45 -0800)]
Fix class unloading with the CC collector.

Avoid unnecessarily decoding dex cache and class loader weak roots,
which would trigger read barriers.

Re-enable 141-class-unload with the CC collector.

Bug: 12687968
Bug: 24468364
Change-Id: Ib4c19f25000873cab0e06047040442d135285745

8 years agoMerge "Disable test 964 with gcstress since it often times out."
Alex Light [Fri, 13 Nov 2015 19:07:37 +0000 (19:07 +0000)]
Merge "Disable test 964 with gcstress since it often times out."

8 years agoDisable test 964 with gcstress since it often times out.
Alex Light [Fri, 13 Nov 2015 18:39:34 +0000 (10:39 -0800)]
Disable test 964 with gcstress since it often times out.

Change-Id: Ice9b25d373b849d4415f9189fe94d8ccd6d05cbe

8 years agoMerge "Revert "Revert "Add missing null check to String::ToCharArray"""
Mathieu Chartier [Fri, 13 Nov 2015 18:33:54 +0000 (18:33 +0000)]
Merge "Revert "Revert "Add missing null check to String::ToCharArray"""

8 years agoImplemented control of "export only" feature in dexdump.
Aart Bik [Fri, 13 Nov 2015 18:06:01 +0000 (10:06 -0800)]
Implemented control of "export only" feature in dexdump.

Rationale:
Export only is now under explicit control of a new -e flag, and
can be used to control both -l xml and -l plain. The result is
that, by default, xml output now shows all items (old behavior
can be forced with -e -l xml). This solution feels a bit cleaner
than adding a flag to override export only when -l xml is given
(old dexdump forced this tight relation).

Bug: 25673756 (feature request).

Change-Id: I7cfd48bf91195d2a81d134936719d267fda399f5

8 years agoRevert "Revert "Add missing null check to String::ToCharArray""
Mathieu Chartier [Fri, 13 Nov 2015 16:36:59 +0000 (08:36 -0800)]
Revert "Revert "Add missing null check to String::ToCharArray""

Fix flaky test.

Bug: 25641543

This reverts commit b52a39c76e0cb0638c92cdcf6b5c5aa551420e1a.

Change-Id: I6b0d3087dba3680252e65af554530e2769cde55c

8 years agoMerge "Stream DWARF debug section directly to disk."
David Srbecky [Fri, 13 Nov 2015 15:53:30 +0000 (15:53 +0000)]
Merge "Stream DWARF debug section directly to disk."

8 years agoMerge "Revert "Add missing null check to String::ToCharArray""
Nicolas Geoffray [Fri, 13 Nov 2015 08:16:55 +0000 (08:16 +0000)]
Merge "Revert "Add missing null check to String::ToCharArray""

8 years agoRevert "Add missing null check to String::ToCharArray"
Nicolas Geoffray [Fri, 13 Nov 2015 08:16:23 +0000 (08:16 +0000)]
Revert "Add missing null check to String::ToCharArray"

Does not work with the interpreter.

Bug: 25641543

This reverts commit 952d608062eec2d7f47f9b45dba935ad8b4d23e5.

Change-Id: Ic112fa69c7ddd119cbccc5b65007b5ee4dfccd09

8 years agoMerge "Add missing null check to String::ToCharArray"
Mathieu Chartier [Fri, 13 Nov 2015 01:32:44 +0000 (01:32 +0000)]
Merge "Add missing null check to String::ToCharArray"

8 years agoAdd missing null check to String::ToCharArray
Mathieu Chartier [Fri, 13 Nov 2015 00:53:41 +0000 (16:53 -0800)]
Add missing null check to String::ToCharArray

Added test.

Bug: 25641543
Change-Id: Ic9a21ce8bc530dbedf14334ad47f5faa90ae4ddc

8 years agoMerge "Fix bug with test 964."
Alex Light [Fri, 13 Nov 2015 00:50:11 +0000 (00:50 +0000)]
Merge "Fix bug with test 964."

8 years agoFix bug with test 964.
Alex Light [Thu, 12 Nov 2015 23:38:24 +0000 (15:38 -0800)]
Fix bug with test 964.

An indentation error meant that some test cases would not be
generated. Also fixed run-test script since fixing this causes the
output to be larger than the default ulimit.

Change-Id: I454f89078f8307f400b66627fdab13feaf492512

8 years agoMerge "Add class loader argument to openDexFileNative"
Mathieu Chartier [Thu, 12 Nov 2015 19:19:36 +0000 (19:19 +0000)]
Merge "Add class loader argument to openDexFileNative"

8 years agoAdd class loader argument to openDexFileNative
Mathieu Chartier [Thu, 12 Nov 2015 18:00:58 +0000 (10:00 -0800)]
Add class loader argument to openDexFileNative

Bug: 22858531
Change-Id: I315919d91822db0c73cf16b21d660d5870d5746f

8 years agoMerge "Optimizing/X86: PC-relative dex cache array addressing."
Vladimir Marko [Thu, 12 Nov 2015 15:20:08 +0000 (15:20 +0000)]
Merge "Optimizing/X86: PC-relative dex cache array addressing."

8 years agoMerge "Optimizing: Clean up constant location handling."
Vladimir Marko [Thu, 12 Nov 2015 11:56:22 +0000 (11:56 +0000)]
Merge "Optimizing: Clean up constant location handling."

8 years agoMerge "Revert "Mark __jit_debug_register_code as a unique symbol""
Tamas Berghammer [Thu, 12 Nov 2015 11:06:38 +0000 (11:06 +0000)]
Merge "Revert "Mark __jit_debug_register_code as a unique symbol""

8 years agoRevert "Mark __jit_debug_register_code as a unique symbol"
Tamas Berghammer [Thu, 12 Nov 2015 11:05:50 +0000 (11:05 +0000)]
Revert "Mark __jit_debug_register_code as a unique symbol"

This reverts commit e7132a9a5d98103fa52a750287476b063fa88863.

We are still using ld on some targets instead of gold when compiling for ararch64. Revert until we figure out why.

Change-Id: If09fac8d72485dc8adbc81e44a2af84ed91dcd91

8 years agoMerge "JDWP: default implementation of Method.IsObsolete"
Sebastien Hertz [Thu, 12 Nov 2015 11:00:38 +0000 (11:00 +0000)]
Merge "JDWP: default implementation of Method.IsObsolete"

8 years agoOptimizing/X86: PC-relative dex cache array addressing.
Vladimir Marko [Mon, 2 Nov 2015 14:36:43 +0000 (14:36 +0000)]
Optimizing/X86: PC-relative dex cache array addressing.

Add PC-relative dex cache array addressing for X86 and use
it for better invoke-static/-direct dispatch. Also delay
the initialization to the PC-relative base until needed.

Change-Id: Ib8634d5edce4920cd70172fd13211809cf6948d1

8 years agoMerge "MIPS64: fix calling conventions in LoadClass and field accesses"
Roland Levillain [Thu, 12 Nov 2015 10:43:30 +0000 (10:43 +0000)]
Merge "MIPS64: fix calling conventions in LoadClass and field accesses"

8 years agoMerge "Mark __jit_debug_register_code as a unique symbol"
Tamas Berghammer [Thu, 12 Nov 2015 10:40:23 +0000 (10:40 +0000)]
Merge "Mark __jit_debug_register_code as a unique symbol"

8 years agoMerge "ART: Add comments to Scoped-/ArenaAllocator."
Vladimir Marko [Thu, 12 Nov 2015 10:16:29 +0000 (10:16 +0000)]
Merge "ART: Add comments to Scoped-/ArenaAllocator."

8 years agoMerge "Image space cleanup for app images"
Mathieu Chartier [Thu, 12 Nov 2015 03:23:54 +0000 (03:23 +0000)]
Merge "Image space cleanup for app images"

8 years agoImage space cleanup for app images
Mathieu Chartier [Tue, 10 Nov 2015 22:13:23 +0000 (14:13 -0800)]
Image space cleanup for app images

Removed Heap::GetImageSpace, added Heap::GetBootImageSpace.

Generalized some logic in the class linker for image spaces.

Bug: 22858531

Change-Id: Ib3e12bb061a247e232d3dc93b0d6b35eb3a34d25

8 years agoMerge "Get class loader to remember interrupts"
Hans Boehm [Thu, 12 Nov 2015 02:51:20 +0000 (02:51 +0000)]
Merge "Get class loader to remember interrupts"

8 years agoGet class loader to remember interrupts
Hans Boehm [Thu, 12 Nov 2015 00:13:57 +0000 (16:13 -0800)]
Get class loader to remember interrupts

Bug: 25606036

Monitor::Wait should not reset the interrupt flag unless it actually
throws the exception.

Change-Id: Id6a35c888160043831f30055f9bd39e7f0440439

8 years agoMIPS64: fix calling conventions in LoadClass and field accesses
Alexey Frunze [Wed, 11 Nov 2015 21:31:12 +0000 (13:31 -0800)]
MIPS64: fix calling conventions in LoadClass and field accesses

This fixes failures in test 529-checker-unresolved

Change-Id: I6170c22059e9711b2fcc965b00d6e34edd839539

8 years agoMerge "Make test 968's name shorter"
Alex Light [Wed, 11 Nov 2015 23:36:15 +0000 (23:36 +0000)]
Merge "Make test 968's name shorter"

8 years agoMake test 968's name shorter
Alex Light [Wed, 11 Nov 2015 23:32:24 +0000 (15:32 -0800)]
Make test 968's name shorter

It was causing intermittent CI breaks due to the long name of the
generated oat file.

Change-Id: I71b4b3aa1a13df2696b276998cafe40006d54b66

8 years agoMerge "Revert "lambda: unit test for lambda expressions at Java Language source level.""
Aart Bik [Wed, 11 Nov 2015 22:38:10 +0000 (22:38 +0000)]
Merge "Revert "lambda: unit test for lambda expressions at Java Language source level.""

8 years agoRevert "lambda: unit test for lambda expressions at Java Language source level."
Aart Bik [Wed, 11 Nov 2015 22:36:19 +0000 (22:36 +0000)]
Revert "lambda: unit test for lambda expressions at Java Language source level."

REASON: Buildbots are having jack issues

This reverts commit 152ee552652691ecaec2db420d468cb178647952.

Change-Id: I650fe1a94f1b97c14a689599703498506dd70b23

8 years agoMerge "Move .debug_line before .debug_info."
David Srbecky [Wed, 11 Nov 2015 21:36:22 +0000 (21:36 +0000)]
Merge "Move .debug_line before .debug_info."

8 years agoStream DWARF debug section directly to disk.
David Srbecky [Wed, 11 Nov 2015 20:19:38 +0000 (20:19 +0000)]
Stream DWARF debug section directly to disk.

Change the structure so that the debug sections are written to
disk as we go.  There are still some temporary buffers, however,
we no longer hold all of the data in memory before writing it.
We can not avoid buffering of some things (e.g. .debug_str).

Change-Id: Id4940cf10ae5b6f3cac7fb8d20197f0304079b92

8 years agoMove .debug_line before .debug_info.
David Srbecky [Wed, 11 Nov 2015 20:38:16 +0000 (20:38 +0000)]
Move .debug_line before .debug_info.

Swap the order since .debug_info depends on .debug_line, but
not vice versa.  This is for disk streaming convenience.

Change-Id: Iabe18fea97aa5527aedfc55c9bdf66e5587568bc

8 years agoART: Add comments to Scoped-/ArenaAllocator.
Vladimir Marko [Wed, 11 Nov 2015 18:45:50 +0000 (18:45 +0000)]
ART: Add comments to Scoped-/ArenaAllocator.

Change-Id: Iba7350114fb998ac9ee1466ef12f4a5024a18e8e

8 years agoMerge "ART: Fix arena allocation for valgrind."
Vladimir Marko [Wed, 11 Nov 2015 18:44:25 +0000 (18:44 +0000)]
Merge "ART: Fix arena allocation for valgrind."

8 years agoMerge "lambda: unit test for lambda expressions at Java Language source level."
Aart Bik [Wed, 11 Nov 2015 18:30:36 +0000 (18:30 +0000)]
Merge "lambda: unit test for lambda expressions at Java Language source level."

8 years agolambda: unit test for lambda expressions at Java Language source level.
Aart Bik [Wed, 11 Nov 2015 00:49:45 +0000 (16:49 -0800)]
lambda: unit test for lambda expressions at Java Language source level.

Rationale: having a checkers-flavored unit test will enable testing
           both if the correct HIR is generated using the checker
           as well as the correctness of execution (interpreter for now,
           but native code later).

Change-Id: I588e9955efaaf53e5c5e3398399e009041cd65f0

8 years agoMove the trace file out of the profiles folder.
Calin Juravle [Wed, 4 Nov 2015 16:16:47 +0000 (16:16 +0000)]
Move the trace file out of the profiles folder.

/dalvik-cache/profiles/ is going away. Create the trace file under
/data/misc/trace.

This enable tracing without forcing SElinux into permissive mode.

Bug: 24698874
Bug: 25612377
Change-Id: I78d09ef629bebe635c3baabaf88f01996d3333e5

8 years agoMerge "Correct exception behavior for default methods"
Alex Light [Wed, 11 Nov 2015 18:07:14 +0000 (18:07 +0000)]
Merge "Correct exception behavior for default methods"

8 years agoART: Fix arena allocation for valgrind.
Vladimir Marko [Tue, 10 Nov 2015 20:54:22 +0000 (20:54 +0000)]
ART: Fix arena allocation for valgrind.

Move the zero-initialization check after marking the newly
allocated chunk as defined and check only the allocated
space without the red zone. Also mark unallocated space as
inaccessible instead of just undefined.

Change-Id: I74fc65f5b53acb74cec4e5a0146f41dacf4a1470

8 years agoChange roots view to "rooted".
Richard Uhler [Wed, 11 Nov 2015 17:13:23 +0000 (09:13 -0800)]
Change roots view to "rooted".

Instead of just showing roots in the roots view, show all objects
whose immediate dominator is the SENTINEL_ROOT.

Bug: 24613815
Change-Id: I96429d75395edfe109222e88d31cdc0bd87a7767

8 years agoIgnore NullConstants when computing the type for Phis.
Calin Juravle [Wed, 11 Nov 2015 14:59:46 +0000 (14:59 +0000)]
Ignore NullConstants when computing the type for Phis.

NullConstants are always typed as Object and they widen the
type of Phis unnecessarily.

Change-Id: I2b29e67617a22120642cb23d2bc6c1ff75b4e7c2

8 years agoOptimizing: Clean up constant location handling.
Vladimir Marko [Wed, 11 Nov 2015 12:37:46 +0000 (12:37 +0000)]
Optimizing: Clean up constant location handling.

Locations builder should use ConstantLocation() when the
code generator relies on a location to be constant. Code
generator should interrogate locations, not inputs, about
being const.

Change-Id: Ic35bb84aa9f83e0977b151a0430aca6c88f19cf0

8 years agoMark __jit_debug_register_code as a unique symbol
Tamas Berghammer [Wed, 2 Sep 2015 10:24:33 +0000 (11:24 +0100)]
Mark __jit_debug_register_code as a unique symbol

This symbol is used to notify the debugger about JIT library load/unload
events. Because of ICF, several functions with the same body put to the
same adress by the linker what causes the debugger to get notifictaion
about JIT load/unload events in several case where it isn't really happened.

This change disable ICF for the __jit_debug_register_code symbol to prevent
this behavior with keeping the symbol unique in the object file.

This is the recommit of 5caf652d7bfb43bdd2f3addfe8021a5fb8e3f796 after
fixing the compilation issue on sdk_phone_armv7-sdk (mac) target.

Change-Id: I027d2c7964098d47ed350ee6b7c74d82c88a5d1f

8 years agoMerge "Revert "Revert "Run type propagation after inliner only when needed."""
Calin Juravle [Wed, 11 Nov 2015 13:55:39 +0000 (13:55 +0000)]
Merge "Revert "Revert "Run type propagation after inliner only when needed."""

8 years agoRevert "Revert "Run type propagation after inliner only when needed.""
Calin Juravle [Mon, 26 Oct 2015 14:05:01 +0000 (14:05 +0000)]
Revert "Revert "Run type propagation after inliner only when needed.""

This reverts commit 271743601650308c7ac5c7a3ec35025d8130a298.

Change-Id: I173e27a0a4d7d54f90ca459eb48d280d1d40ab70

8 years agoMerge "Ignore failures of WeakHashMapTest#test_keySet in libcore tests."
Roland Levillain [Wed, 11 Nov 2015 12:40:26 +0000 (12:40 +0000)]
Merge "Ignore failures of WeakHashMapTest#test_keySet in libcore tests."

8 years agoIgnore failures of WeakHashMapTest#test_keySet in libcore tests.
Roland Levillain [Wed, 11 Nov 2015 12:16:52 +0000 (12:16 +0000)]
Ignore failures of WeakHashMapTest#test_keySet in libcore tests.

The outcome of this test depends on a weak reference being
collected before a timeout, which sometimes fails.  This
test used to pass, but recent changes have made this
non-determinism more frequent.  Until we rewrite this
properly (if ever), just ignore these failures.

Bug: 25437292
Change-Id: I1f041f6dfc1318026bce966575477486207f5351

8 years agoMerge "Optimizing/ARM: Improve shifts of long values by a constant."
Vladimir Marko [Wed, 11 Nov 2015 11:27:03 +0000 (11:27 +0000)]
Merge "Optimizing/ARM: Improve shifts of long values by a constant."

8 years agoOptimizing/ARM: Improve shifts of long values by a constant.
Vladimir Marko [Tue, 10 Nov 2015 19:31:26 +0000 (19:31 +0000)]
Optimizing/ARM: Improve shifts of long values by a constant.

Change-Id: Id66ef8cdb9e64306f2be547370b90cc100a3e086

8 years agoCorrect exception behavior for default methods
Alex Light [Fri, 9 Oct 2015 22:59:48 +0000 (15:59 -0700)]
Correct exception behavior for default methods

Default methods are defined to throw an IncompatibleClassChangeError
(ICCE) when they are called and there is no "best" implementation.
Previously we would simply throw an ICCE during class loading as soon
as we noticed that this would happen if called. This makes us wait
until we actually attempt to execute the method. Furthermore, this
allows us to use other, non-conflicting, methods on the object as
normal.

Furthermore, this makes us correctly throw AbstractMethodErrors in
cases where all default implementations of a method are overridden by
abstract declarations.

Adds 3 tests for this new behavior.

Bug: 24618811

Change-Id: Id891958a81f9b3862b2ce5919636aabef7d3422e

8 years agoMerge "Don't use fixed addresses in mem map test"
Mathieu Chartier [Tue, 10 Nov 2015 22:09:40 +0000 (22:09 +0000)]
Merge "Don't use fixed addresses in mem map test"

8 years agoDon't use fixed addresses in mem map test
Mathieu Chartier [Tue, 10 Nov 2015 18:32:52 +0000 (10:32 -0800)]
Don't use fixed addresses in mem map test

We obtain a valid address instead of assuming there won't be a map
there.

These checks used to be OK when we didn't start the runtime.

Change-Id: Ie636ed3773d2c0c62d87e57623ee4ce380c64db0

8 years agoMerge "ART: Fix leak in Heap"
Andreas Gampe [Tue, 10 Nov 2015 21:55:08 +0000 (21:55 +0000)]
Merge "ART: Fix leak in Heap"

8 years agoART: Fix leak in Heap
Andreas Gampe [Tue, 10 Nov 2015 21:34:17 +0000 (13:34 -0800)]
ART: Fix leak in Heap

Delete a mutex in the destructor.

Change-Id: I7fb8f3f9e7a30baf352f384f38520eb0dd4906eb

8 years agoMerge "Optimizing: Clean up Equal/NotEqual constant folding for nulls."
Vladimir Marko [Tue, 10 Nov 2015 20:04:08 +0000 (20:04 +0000)]
Merge "Optimizing: Clean up Equal/NotEqual constant folding for nulls."

8 years agoMerge "Rosalloc fast path in assembly for x86_64."
Hiroshi Yamauchi [Tue, 10 Nov 2015 19:55:05 +0000 (19:55 +0000)]
Merge "Rosalloc fast path in assembly for x86_64."

8 years agoOptimizing: Clean up Equal/NotEqual constant folding for nulls.
Vladimir Marko [Tue, 10 Nov 2015 19:53:01 +0000 (19:53 +0000)]
Optimizing: Clean up Equal/NotEqual constant folding for nulls.

Change-Id: I17766395092ec61df61ef0b9ae4c37fd38164a3b

8 years agoMerge "ART: Remove unnecessary helper method from GraphChecker"
David Brazdil [Tue, 10 Nov 2015 19:04:28 +0000 (19:04 +0000)]
Merge "ART: Remove unnecessary helper method from GraphChecker"

8 years agoART: Remove unnecessary helper method from GraphChecker
David Brazdil [Tue, 10 Nov 2015 18:49:28 +0000 (18:49 +0000)]
ART: Remove unnecessary helper method from GraphChecker

Change-Id: I59f50a559da560033f4c9407d83b68fc5d64bd3c

8 years agoMerge "Optimizing: Improve constant folding + DCE for inlining."
Vladimir Marko [Tue, 10 Nov 2015 18:31:55 +0000 (18:31 +0000)]
Merge "Optimizing: Improve constant folding + DCE for inlining."

8 years agoMerge "ART: Refactor iteration over normal/exceptional successors"
David Brazdil [Tue, 10 Nov 2015 18:30:18 +0000 (18:30 +0000)]
Merge "ART: Refactor iteration over normal/exceptional successors"

8 years agoART: Refactor iteration over normal/exceptional successors
David Brazdil [Tue, 10 Nov 2015 11:07:31 +0000 (11:07 +0000)]
ART: Refactor iteration over normal/exceptional successors

Add helper methods on HBasicBlock which return ArrayRef with the
suitable sub-array of the `successors_` list.

Change-Id: I66c83bb56f2984d7550bf77c48110af4087515a8

8 years agoOptimizing: Improve constant folding + DCE for inlining.
Vladimir Marko [Tue, 10 Nov 2015 17:14:35 +0000 (17:14 +0000)]
Optimizing: Improve constant folding + DCE for inlining.

Run constant folding before DCE in inliner to eliminate more
code that can prevent inlining. Improve the constant folding
to evaluate Equals and NotEquals for null inputs.

Change-Id: I876ffb903ef39484370b6c8793f0f8467a977362

8 years agoMerge "ART: Improve comment about inlining into try/catch"
David Brazdil [Tue, 10 Nov 2015 13:10:08 +0000 (13:10 +0000)]
Merge "ART: Improve comment about inlining into try/catch"

8 years agoART: Improve comment about inlining into try/catch
David Brazdil [Tue, 10 Nov 2015 13:04:30 +0000 (13:04 +0000)]
ART: Improve comment about inlining into try/catch

Change-Id: I66a4fd3206847c8d5bb57b1678d9d3dc94331294

8 years agoMerge "ART: Fix critical edge checks in GraphChecker"
David Brazdil [Tue, 10 Nov 2015 13:01:50 +0000 (13:01 +0000)]
Merge "ART: Fix critical edge checks in GraphChecker"

8 years agoMerge "Revert "Revert "ART: Update DCE to work with try/catch"""
David Brazdil [Tue, 10 Nov 2015 13:01:28 +0000 (13:01 +0000)]
Merge "Revert "Revert "ART: Update DCE to work with try/catch"""

8 years agoART: Fix critical edge checks in GraphChecker
David Brazdil [Tue, 10 Nov 2015 10:12:41 +0000 (10:12 +0000)]
ART: Fix critical edge checks in GraphChecker

Previous CL I5a13b8bb74509b48f5d628906f7158af007f99ae fixed logic for
splitting critical edges. The same logic is used in GraphChecker when
testing critical edges weren't introduced by a pass. This patch updates
it too.

Bug: 25493695
Bug: 25454012
Change-Id: I56226c82324ee55ae2fbe7262608dd2868d930f1

8 years agoMerge "ART: Add clarifying comment"
David Brazdil [Tue, 10 Nov 2015 09:46:41 +0000 (09:46 +0000)]
Merge "ART: Add clarifying comment"

8 years agoART: Add clarifying comment
David Brazdil [Tue, 10 Nov 2015 09:43:28 +0000 (09:43 +0000)]
ART: Add clarifying comment

Change-Id: I189ec4cae0aa1a5245a79e86d1ec0592e38eac4a

8 years agoRevert "Revert "ART: Update DCE to work with try/catch""
David Brazdil [Mon, 2 Nov 2015 20:24:55 +0000 (20:24 +0000)]
Revert "Revert "ART: Update DCE to work with try/catch""

The previous CL failed because it did not update inputs of catch phis.
Since phi input indices cannot be easily mapped back to throwing
instructions, this new implementation at least removes catch phi uses
of values defined in the removed blocks to preserve graph consistency.

This reverts commit fb552d7061746f7a90fdd5002696e255e2e15c35.

Change-Id: I63d95915d1ef50e71d3bcf0cd10aaded554035b4

8 years agoMerge "Delete flaky DCHECK"
Mathieu Chartier [Tue, 10 Nov 2015 03:02:58 +0000 (03:02 +0000)]
Merge "Delete flaky DCHECK"

8 years agoDelete flaky DCHECK
Mathieu Chartier [Tue, 10 Nov 2015 03:01:00 +0000 (19:01 -0800)]
Delete flaky DCHECK

Fixes mips build

Change-Id: I738b41f09f3c926f21640e358cf0bd5557167b0e

8 years agoMerge "Fix mac build"
Mathieu Chartier [Tue, 10 Nov 2015 02:51:37 +0000 (02:51 +0000)]
Merge "Fix mac build"

8 years agoFix mac build
Mathieu Chartier [Tue, 10 Nov 2015 02:50:49 +0000 (18:50 -0800)]
Fix mac build

Change-Id: I22bf169f15483e54b97d77ff0887cd1fde428af9

8 years agoRosalloc fast path in assembly for x86_64.
Hiroshi Yamauchi [Thu, 29 Oct 2015 22:34:54 +0000 (15:34 -0700)]
Rosalloc fast path in assembly for x86_64.

Measurements (host, ms)
    BinaryTrees:   324 ->  360 (+11%)
    BinaryTrees with 64 MB alloc stack + 1 GB heap:
                   299 ->  275  (-8%)
    MemAllocTest:  414 ->  368 (-11%)

Interestingly, BinaryTrees gets slower with default settings due to more
blocking (gc-for-alloc) collections. It seems because allocations are
faster, the allocation stack size and the heap size become the
bottleneck (note both an allocation stack overflow as well as heap
exhaustion cause gc-for-alloc collections). With a larger allocation
stack and heap where no blocking collections are observed, BinaryTrees
gets faster.

Bug: 9986565
Change-Id: I642b9fecd0a583cc133998c2f3932de815c4a757

8 years agoMerge "Add low_4gb support to MapFile and MapFileAtAddress"
Mathieu Chartier [Tue, 10 Nov 2015 02:07:04 +0000 (02:07 +0000)]
Merge "Add low_4gb support to MapFile and MapFileAtAddress"

8 years agoMerge "Some cleanup for tests 960-965 build scripts."
Alex Light [Tue, 10 Nov 2015 01:36:31 +0000 (01:36 +0000)]
Merge "Some cleanup for tests 960-965 build scripts."

8 years agoAdd low_4gb support to MapFile and MapFileAtAddress
Mathieu Chartier [Mon, 9 Nov 2015 23:16:56 +0000 (15:16 -0800)]
Add low_4gb support to MapFile and MapFileAtAddress

Motivation is to use this for loading app images in low 4GB at a
non fixed address.

Added test.

Bug: 22858531
Change-Id: I0f79a4a7bfbfbdfc112e41b25c8682b1fb932ab7

8 years agoSome cleanup for tests 960-965 build scripts.
Alex Light [Mon, 9 Nov 2015 22:26:35 +0000 (14:26 -0800)]
Some cleanup for tests 960-965 build scripts.

Make all 96* build scripts consistent and update 965 to build and run
with jack.

Change-Id: If27ee9c92880c75307a18b054ba926670877073a

8 years agoMerge "ART: Rerun the verifier for compile-time failures"
Andreas Gampe [Mon, 9 Nov 2015 21:25:47 +0000 (21:25 +0000)]
Merge "ART: Rerun the verifier for compile-time failures"

8 years agoMerge "Class Linker: Implement default interface verification ordering."
Alex Light [Mon, 9 Nov 2015 19:30:10 +0000 (19:30 +0000)]
Merge "Class Linker: Implement default interface verification ordering."

8 years agoClass Linker: Implement default interface verification ordering.
Alex Light [Wed, 7 Oct 2015 23:08:36 +0000 (16:08 -0700)]
Class Linker: Implement default interface verification ordering.

We ensure that default interface verification failure is propagated to
the implementing class so that the compiler is able to properly handle
verification failures. Furthermore we ensure that default interfaces
are verified at the same time their implementing classes are.

This does not add additional support for default methods to the verifier.

This adds a basic test of this functionality.

Previously assumptions made about the order of verification in our
code could cause a check failure during compilation. Our old code
assumed that all code that would be run during a classes
initialization had been verified before any of it was run. The default
interfaces broke this assumption. Now we will simply verify default
interfaces with any classes that implement them.

Bug: 24618811

Change-Id: Ia11a50a28889dbd7ca22ffa730ef443e39f3cb5b

8 years agoART: Rerun the verifier for compile-time failures
Andreas Gampe [Fri, 6 Nov 2015 04:39:56 +0000 (20:39 -0800)]
ART: Rerun the verifier for compile-time failures

To aid app failure diagnosis, by default re-run the verifier at
runtime to compute a better VerifyError message.

Rewrite the verifier driver code to pass the last actual low-level
verifier message.

Bug: 25432718
Change-Id: Ib8e6dd1ce8121045c0d38f54969100094c3dde6e

8 years agoMerge "Add more image writer comments and clean up"
Mathieu Chartier [Mon, 9 Nov 2015 18:07:01 +0000 (18:07 +0000)]
Merge "Add more image writer comments and clean up"