OSDN Git Service

android-x86/art.git
9 years agoMerge "Switch to C++11 style [[noreturn]]."
Ian Rogers [Sun, 12 Oct 2014 15:34:28 +0000 (15:34 +0000)]
Merge "Switch to C++11 style [[noreturn]]."

9 years agoSwitch to C++11 style [[noreturn]].
Ian Rogers [Sat, 11 Oct 2014 03:05:39 +0000 (20:05 -0700)]
Switch to C++11 style [[noreturn]].

Also remove old comment about calls to Abort getting merged. We have few
LOG(FATAL)s so merging is less of an issue. Also, recompiling with O0 and
similar will avoid the optimization and we expect that for GDB debugging.

Change-Id: I840dc6f4b8511294be0e117c634ec15b745e7be5

9 years agoMerge "ART: Add smali-based run-test"
Ian Rogers [Sat, 11 Oct 2014 02:51:08 +0000 (02:51 +0000)]
Merge "ART: Add smali-based run-test"

9 years agoART: Add smali-based run-test
Andreas Gampe [Fri, 3 Oct 2014 23:15:37 +0000 (16:15 -0700)]
ART: Add smali-based run-test

Add run-test 800 for smali-based tests. To use, drop a smali file
into the src/ directory and add a TestCase in src/Main.java.

Bug: 17814475

Change-Id: Ica9eb830689862cb3a4ffa0019fbc447c01af744

9 years agoMerge "ART Runtime: Add movb $xx,yy fault handling"
Ian Rogers [Sat, 11 Oct 2014 01:24:47 +0000 (01:24 +0000)]
Merge "ART Runtime: Add movb $xx,yy fault handling"

9 years agoMerge "Hold mutator lock in DdmSendHeapSegments for all spaces"
Mathieu Chartier [Fri, 10 Oct 2014 22:47:08 +0000 (22:47 +0000)]
Merge "Hold mutator lock in DdmSendHeapSegments for all spaces"

9 years agoHold mutator lock in DdmSendHeapSegments for all spaces
Mathieu Chartier [Fri, 10 Oct 2014 19:45:50 +0000 (12:45 -0700)]
Hold mutator lock in DdmSendHeapSegments for all spaces

Previously we were releasing the mutator lock in DdmSendHeapSegments
and only reacquiring it for RosAlloc spaces. This was causing problems
since the HeapChunkCallback access object fields through mirror.

Bug: 17950534

(cherry picked from commit d6527cf8e824d9057f32755f2ff4bdcf46c7095b)

Change-Id: Idb307fd4c01450a07e3c9621e04d2aabf2c6a0b9

9 years agoMerge "Work around ICE bugs with MIPS GCC and O1."
Ian Rogers [Fri, 10 Oct 2014 22:24:24 +0000 (22:24 +0000)]
Merge "Work around ICE bugs with MIPS GCC and O1."

9 years agoWork around ICE bugs with MIPS GCC and O1.
Ian Rogers [Fri, 10 Oct 2014 20:03:39 +0000 (13:03 -0700)]
Work around ICE bugs with MIPS GCC and O1.

Also, work around GCC warning bugs where array accesses with explicit
bounds checks are flagged as being out-of-bounds. Significantly, clean-up
the HandleScope so the array accesses don't appear out-of-bounds at compile
time.

Change-Id: I5d66567559cc1f97cd0aa02c0df8575ebadbfe3d

9 years agoMerge "Fix 2 new sets of clang compiler warnings."
Ian Rogers [Fri, 10 Oct 2014 20:23:09 +0000 (20:23 +0000)]
Merge "Fix 2 new sets of clang compiler warnings."

9 years agoMerge "Disable 114-ParallelGC for GC stress testing."
Ian Rogers [Fri, 10 Oct 2014 19:48:22 +0000 (19:48 +0000)]
Merge "Disable 114-ParallelGC for GC stress testing."

9 years agoDisable 114-ParallelGC for GC stress testing.
Ian Rogers [Fri, 10 Oct 2014 19:39:31 +0000 (12:39 -0700)]
Disable 114-ParallelGC for GC stress testing.

Bug: 17387969
Change-Id: Ic50e9642468646c36afa5cfbe4050d24e9190ae6

9 years agoFix 2 new sets of clang compiler warnings.
Ian Rogers [Fri, 10 Oct 2014 18:02:11 +0000 (11:02 -0700)]
Fix 2 new sets of clang compiler warnings.

Fix issues that are flagged by -Wfloat-equal and -Wmissing-noreturn.
In the case of -Wfloat-equal the current cases in regular code are deliberate,
so the change is to silence the warning. For gtest code the appropriate fix is
to switch from EXPECT_EQ to EXPECT_(FLOAT|DOUBLE)_EQ.
The -Wmissing-noreturn warning isn't enabled due to a missing noreturn in
gtest. This issue has been reported to gtest.

Change-Id: Id84c70c21c542716c9ee0c41492e8ff8788c4ef8

9 years agoMerge "Cleanup baseline register allocator."
Nicolas Geoffray [Fri, 10 Oct 2014 11:36:33 +0000 (11:36 +0000)]
Merge "Cleanup baseline register allocator."

9 years agoCleanup baseline register allocator.
Nicolas Geoffray [Thu, 9 Oct 2014 21:13:55 +0000 (22:13 +0100)]
Cleanup baseline register allocator.

- Use three arrays for blocking regsters instead of
  one and computing offsets in that array.]
- Don't pass blocked_registers_ to methods, just use the field.

Change-Id: Ib698564c31127c59b5a64c80f4262394b8394dc6

9 years agoMerge "Fix code generation of materialized conditions."
Nicolas Geoffray [Fri, 10 Oct 2014 11:22:50 +0000 (11:22 +0000)]
Merge "Fix code generation of materialized conditions."

9 years agoMerge "Only watch location for BREAKPOINT event"
Sebastien Hertz [Fri, 10 Oct 2014 08:22:18 +0000 (08:22 +0000)]
Merge "Only watch location for BREAKPOINT event"

9 years agoMerge "Enable -Wimplicit-fallthrough."
Ian Rogers [Fri, 10 Oct 2014 05:23:50 +0000 (05:23 +0000)]
Merge "Enable -Wimplicit-fallthrough."

9 years agoEnable -Wimplicit-fallthrough.
Ian Rogers [Fri, 10 Oct 2014 04:56:44 +0000 (21:56 -0700)]
Enable -Wimplicit-fallthrough.

Falling through switch cases on a clang build must now annotate the fallthrough
with the FALLTHROUGH_INTENDED macro.
Bug: 17731372

Change-Id: I836451cd5f96b01d1ababdbf9eef677fe8fa8324

9 years agoMerge "Fix object header size for baker read barrier."
Ian Rogers [Thu, 9 Oct 2014 23:48:06 +0000 (23:48 +0000)]
Merge "Fix object header size for baker read barrier."

9 years agoFix object header size for baker read barrier.
Ian Rogers [Thu, 9 Oct 2014 23:29:40 +0000 (16:29 -0700)]
Fix object header size for baker read barrier.

Change-Id: I2b4a61488b9017d22210d79a07f485f9afef3b42

9 years agoMerge "stdint types all the way!"
Ian Rogers [Thu, 9 Oct 2014 23:07:50 +0000 (23:07 +0000)]
Merge "stdint types all the way!"

9 years agostdint types all the way!
Ian Rogers [Wed, 8 Oct 2014 19:43:28 +0000 (12:43 -0700)]
stdint types all the way!

Change-Id: I4e4ef3a2002fc59ebd9097087f150eaf3f2a7e08

9 years agoMerge "Simplify instance reference offsets."
Ian Rogers [Thu, 9 Oct 2014 22:42:24 +0000 (22:42 +0000)]
Merge "Simplify instance reference offsets."

9 years agoSimplify instance reference offsets.
Ian Rogers [Thu, 9 Oct 2014 20:21:38 +0000 (13:21 -0700)]
Simplify instance reference offsets.

Don't encode Object's class. Use trailing rather than leading zeroes to give
offset position.

Change-Id: I1ae74e7a01f63696429644adf81cdf6ee58832fe

9 years agoMerge "Fix race in thread list destructor"
Mathieu Chartier [Thu, 9 Oct 2014 20:46:32 +0000 (20:46 +0000)]
Merge "Fix race in thread list destructor"

9 years agoFix race in thread list destructor
Mathieu Chartier [Thu, 9 Oct 2014 19:57:58 +0000 (12:57 -0700)]
Fix race in thread list destructor

The ThreadList::Contains wasn't being guarded by the thread list
lock, which could cause a corrupted thread list if another thread
unregistered itself while std::find was searching the thread list.

Bug: 17896374

(cherry picked from commit b90132cb5132eaeb4dbfca1e63d79b4005dacec5)

Change-Id: I8f6f979f365ca00ac0655e04eb26020d2ad7d6ee

9 years agoMerge changes Ia0044523,Ieefe83cf
Ian Rogers [Thu, 9 Oct 2014 18:51:11 +0000 (18:51 +0000)]
Merge changes Ia0044523,Ieefe83cf

* changes:
  Fix ARM64 build.
  ARM64: Update code after the VIXL 1.6 release.

9 years agoFix ARM64 build.
Ian Rogers [Thu, 9 Oct 2014 18:50:09 +0000 (11:50 -0700)]
Fix ARM64 build.

Change-Id: Ia0044523473d7a93ef3f0eed8e6c28602e767c17

9 years agoMerge "Fix arm64 build: move macro usage after definition."
Vladimir Marko [Thu, 9 Oct 2014 17:39:46 +0000 (17:39 +0000)]
Merge "Fix arm64 build: move macro usage after definition."

9 years agoMerge "ART: X86: GenLongArith should handle overlapped VRs"
Bill Buzbee [Thu, 9 Oct 2014 17:37:21 +0000 (17:37 +0000)]
Merge "ART: X86: GenLongArith should handle overlapped VRs"

9 years agoFix arm64 build: move macro usage after definition.
Vladimir Marko [Thu, 9 Oct 2014 17:34:28 +0000 (18:34 +0100)]
Fix arm64 build: move macro usage after definition.

The build was broken by
    https://android-review.googlesource.com/110341

Change-Id: I137b4bc694af15465bf0db3ffe337950064d2a53

9 years agoMerge "Restore valgrind-test-art-host"
Brian Carlstrom [Thu, 9 Oct 2014 17:29:06 +0000 (17:29 +0000)]
Merge "Restore valgrind-test-art-host"

9 years agoRestore valgrind-test-art-host
Brian Carlstrom [Thu, 9 Oct 2014 17:17:07 +0000 (10:17 -0700)]
Restore valgrind-test-art-host

Change-Id: I2586e16664dabfc543bc9d0dab471ec8d3810f34

9 years agoART: X86: GenLongArith should handle overlapped VRs
Alexei Zavjalov [Wed, 8 Oct 2014 08:51:59 +0000 (15:51 +0700)]
ART: X86: GenLongArith should handle overlapped VRs

In a case, when src and dest VRs are overlapped when we called
GenLongArith it may cause the incorrect use of regs.

The solution is to map src to an physical reg and work with this
reg instead of mem.

Renamed BadOverlap() to PartiallyIntersects() for consistency.

Change-Id: Ia3fc7f741f0a92556e1b2a1b084506662ef04c9d
Signed-off-by: Katkov, Serguei I <serguei.i.katkov@intel.com>
Signed-off-by: Alexei Zavjalov <alexei.zavjalov@intel.com>
9 years agoMerge "Makefile cleanups."
Nicolas Geoffray [Thu, 9 Oct 2014 17:03:53 +0000 (17:03 +0000)]
Merge "Makefile cleanups."

9 years agoMakefile cleanups.
Nicolas Geoffray [Thu, 9 Oct 2014 16:34:45 +0000 (17:34 +0100)]
Makefile cleanups.

Also add the DEX2OAT_FLAGS to the generation of core.oat/core.art.

Change-Id: I5d113fad460b2e14fc652816069adba29ca744b8

9 years agoOnly watch location for BREAKPOINT event
Sebastien Hertz [Wed, 8 Oct 2014 11:54:55 +0000 (13:54 +0200)]
Only watch location for BREAKPOINT event

This CL ensures we watch a location for BREAKPOINT event only. Other
events (single-step, method entry/exit, ...) are handled differently
and LocationOnly modifier is used as an event filter in this case.

This prevents from failing a check when we need to deoptimize for
non-breakpoint event.

Bug: 17908144
Change-Id: Ib413d62fa31480fec8d750543c0605ba52188350

9 years agoARM64: Update code after the VIXL 1.6 release.
Alexandre Rames [Wed, 8 Oct 2014 17:41:21 +0000 (18:41 +0100)]
ARM64: Update code after the VIXL 1.6 release.

We now leave the assembler buffer management to VIXL.

Change-Id: Ieefe83cf5cf5e1ab8c924b0e7dc03af6a55053ae

9 years agoMerge "Tidy ELF builder."
Ian Rogers [Thu, 9 Oct 2014 15:52:50 +0000 (15:52 +0000)]
Merge "Tidy ELF builder."

9 years agoMerge "Quick: Reduce memory usage and improve compile time."
Vladimir Marko [Thu, 9 Oct 2014 15:30:26 +0000 (15:30 +0000)]
Merge "Quick: Reduce memory usage and improve compile time."

9 years agoMerge "Quick: Separate null check elimination and type inference."
Vladimir Marko [Thu, 9 Oct 2014 15:28:22 +0000 (15:28 +0000)]
Merge "Quick: Separate null check elimination and type inference."

9 years agoTidy ELF builder.
Ian Rogers [Thu, 9 Oct 2014 00:27:48 +0000 (17:27 -0700)]
Tidy ELF builder.

Don't do "if (ptr)". Use const. Use DISALLOW_COPY_AND_ASSIGN. Avoid public
member variables.
Move ValueObject to base and use in ELF builder.
Tidy VectorOutputStream to not use non-const reference arguments.

Change-Id: I2c727c3fc61769c3726de7cfb68b2d6eb4477e53

9 years agoQuick: Reduce memory usage and improve compile time.
Vladimir Marko [Thu, 9 Oct 2014 13:57:59 +0000 (14:57 +0100)]
Quick: Reduce memory usage and improve compile time.

Move the def-block-matrix from Arena to ScopedArena. Remove
BasicBlockDataFlow::ending_check_v and use a temporary bit
matrix instead. Remove unused BasicBlockDataFlow::phi_v.
Avoid some BitVector::Copy() at the end of null and clinit
check elimination passes when the contents of the source
BitVector is no longer needed.

Change-Id: I8111b2f8a51e63075aa124b528d61b79b6933274

9 years agoQuick: Separate null check elimination and type inference.
Vladimir Marko [Thu, 9 Oct 2014 11:26:10 +0000 (12:26 +0100)]
Quick: Separate null check elimination and type inference.

Change-Id: I4566ae9354c91ca935481cb4f5b729bba05c1592

9 years agoFix code generation of materialized conditions.
Nicolas Geoffray [Wed, 8 Oct 2014 20:07:48 +0000 (21:07 +0100)]
Fix code generation of materialized conditions.

Move the logic for knowing if a condition needs to be materialized
in an optimization pass (so that the information does not change
as a side effect of another optimization).

Also clean-up arm and x86_64 codegen:
- arm: ldr and str are for power-users when a constant is
  in play. We should use LoadFromOffset and StoreToOffset.
- x86_64: fix misuses of movq instead of movl.

Change-Id: I01a03b91803624be2281a344a13ad5efbf4f3ef3

9 years agoMerge "Stop converting from Location to ManagedRegister."
Nicolas Geoffray [Thu, 9 Oct 2014 13:55:05 +0000 (13:55 +0000)]
Merge "Stop converting from Location to ManagedRegister."

9 years agoStop converting from Location to ManagedRegister.
Nicolas Geoffray [Thu, 9 Oct 2014 10:47:51 +0000 (11:47 +0100)]
Stop converting from Location to ManagedRegister.

Now the source of truth is the Location object that knows
which register (core, pair, fpu) it needs to refer to.

Change-Id: I62401343d7479ecfb24b5ed161ec7829cda5a0b1

9 years agoMerge "Disable kEnableJavaStackTraceHandler"
Brian Carlstrom [Thu, 9 Oct 2014 04:11:22 +0000 (04:11 +0000)]
Merge "Disable kEnableJavaStackTraceHandler"

9 years agoDisable kEnableJavaStackTraceHandler
Brian Carlstrom [Wed, 8 Oct 2014 20:56:25 +0000 (13:56 -0700)]
Disable kEnableJavaStackTraceHandler

Bug: 17890306

(cherry picked from commit b6404fe0bb782cbbccda88903901fe0096a741d5)

Change-Id: Ie13d6af1d384c6ebf87e8f055120609e2bff3338

9 years agoMerge "Don't copy fill array data to quick literal pool."
Ian Rogers [Wed, 8 Oct 2014 23:25:54 +0000 (23:25 +0000)]
Merge "Don't copy fill array data to quick literal pool."

9 years agoDon't copy fill array data to quick literal pool.
Ian Rogers [Wed, 8 Oct 2014 22:35:22 +0000 (15:35 -0700)]
Don't copy fill array data to quick literal pool.

Currently quick copies the fill array data from the dex file to the literal
pool. It then has to go through hoops to pass this PC relative address down
to out-of-line code. Instead, pass the offset of the table to the out-of-line
code and use the CodeItem data associated with the ArtMethod. This reduces
the size of oat code while greatly simplifying it.
Unify the FillArrayData implementation in quick, portable and the interpreters.

Change-Id: I9c6971cf46285fbf197856627368c0185fdc98ca

9 years agoMerge "Dump mutator lock for thread suspend timeout"
Mathieu Chartier [Wed, 8 Oct 2014 22:41:05 +0000 (22:41 +0000)]
Merge "Dump mutator lock for thread suspend timeout"

9 years agoDump mutator lock for thread suspend timeout
Mathieu Chartier [Wed, 8 Oct 2014 21:26:23 +0000 (14:26 -0700)]
Dump mutator lock for thread suspend timeout

Should help us see if thread suspend timeouts are due to someone
holding on the mutator lock for too long or if it is a missed
futex wake.

(cherry picked from commit f924d2381a5ea53967ba1e279766d601b9be05ea)

Bug: 17837911
Change-Id: I3417156a78f574076d5da17c86836c5a7c76084a

9 years agoMerge "Really fix --gdb when running tests on target."
Nicolas Geoffray [Wed, 8 Oct 2014 20:13:53 +0000 (20:13 +0000)]
Merge "Really fix --gdb when running tests on target."

9 years agoMerge "Header file clean up."
Ian Rogers [Wed, 8 Oct 2014 19:15:21 +0000 (19:15 +0000)]
Merge "Header file clean up."

9 years agoHeader file clean up.
Ian Rogers [Wed, 8 Oct 2014 18:47:24 +0000 (11:47 -0700)]
Header file clean up.

Remove runtime.h from object.h.
Move TypeStaticIf to its own header file to avoid bringing utils.h into
allocator.h.
Move Array::DataOffset into -inl.h as it now has a utils.h dependency.
Fix include issues arising from this.

Change-Id: I4605b1aa4ff5f8dc15706a0132e15df03c7c8ba0

9 years agoMerge "ART: fix NullCheckElimination to preserve MIR_IGNORE_NULL_CHECK"
Bill Buzbee [Wed, 8 Oct 2014 17:20:31 +0000 (17:20 +0000)]
Merge "ART: fix NullCheckElimination to preserve MIR_IGNORE_NULL_CHECK"

9 years agoMerge "ART: Fix SelectKind to work with nullptr"
Bill Buzbee [Wed, 8 Oct 2014 17:20:22 +0000 (17:20 +0000)]
Merge "ART: Fix SelectKind to work with nullptr"

9 years agoMerge "ART: Fix issues with SplitBlock"
Bill Buzbee [Wed, 8 Oct 2014 17:19:59 +0000 (17:19 +0000)]
Merge "ART: Fix issues with SplitBlock"

9 years agoMerge "Fix to MirGraph::GetSSANameWithConst()"
Bill Buzbee [Wed, 8 Oct 2014 17:19:51 +0000 (17:19 +0000)]
Merge "Fix to MirGraph::GetSSANameWithConst()"

9 years agoMerge "Fix and improve static evaluation of constant expressions."
Roland Levillain [Wed, 8 Oct 2014 16:46:56 +0000 (16:46 +0000)]
Merge "Fix and improve static evaluation of constant expressions."

9 years agoFix and improve static evaluation of constant expressions.
Roland Levillain [Mon, 6 Oct 2014 18:24:02 +0000 (19:24 +0100)]
Fix and improve static evaluation of constant expressions.

- Fix the definition of art::HSub::Evaluate.
- Qualify Evaluate methods as OVERRIDE.
- Evaluate comparisons in a deterministic way: if a
  comparison is true, always return 1 (instead of letting
  the compiler return any non-null value).
- Better exercise static evaluation of constant expressions
  in compiler/optimizing/constant_propagation_test.cc.

Change-Id: I13d0862e5f4eba1275016fb8c3c17e9aff54408b

9 years agoART: Fix issues with SplitBlock
Mathew Zaleski [Mon, 15 Sep 2014 13:44:14 +0000 (09:44 -0400)]
ART: Fix issues with SplitBlock

The SplitBlock is a public interface from MIRGraph to be able to create two
blocks from one. This interface can be called not just during the control
flow graph building phase, but later. At that later point, extended instructions
may have already been inserted.

For example, SpecialMethodInliner can insert kMirOpNullCheck and kMirOpNop,
and then another pass tries to split block that contains this. The DCHECK
will fail even though the split is semantically correct.

Signed-off-by: Mathew Zaleski <mathew.zaleski@intel.com>
Conflicts:

compiler/dex/mir_graph.cc

Change-Id: I8bb3551407907909cda444a3379da2eb9331b9ff

9 years agoMerge "Implement method calls using relative BL on ARM64."
Vladimir Marko [Wed, 8 Oct 2014 12:20:28 +0000 (12:20 +0000)]
Merge "Implement method calls using relative BL on ARM64."

9 years agoReally fix --gdb when running tests on target.
Nicolas Geoffray [Wed, 8 Oct 2014 08:56:45 +0000 (09:56 +0100)]
Really fix --gdb when running tests on target.

Change-Id: Ic96edb066d345839a4af371e7c10327713a045c7

9 years agoMerge "Add trivial register hints to the register allocator."
Nicolas Geoffray [Tue, 7 Oct 2014 20:27:28 +0000 (20:27 +0000)]
Merge "Add trivial register hints to the register allocator."

9 years agoAdd trivial register hints to the register allocator.
Nicolas Geoffray [Wed, 1 Oct 2014 10:32:17 +0000 (11:32 +0100)]
Add trivial register hints to the register allocator.

- Add hints for phis, same as first input, and expected registers.
- Make the if instruction accept non-condition instructions.

Change-Id: I34fa68393f0d0c19c68128f017b7a05be556fbe5

9 years agoMerge "Add support for floats and doubles."
Nicolas Geoffray [Tue, 7 Oct 2014 19:33:07 +0000 (19:33 +0000)]
Merge "Add support for floats and doubles."

9 years agoAdd support for floats and doubles.
Nicolas Geoffray [Mon, 6 Oct 2014 08:12:41 +0000 (09:12 +0100)]
Add support for floats and doubles.

- Follows Quick conventions.
- Currently only works with baseline register allocator.

Change-Id: Ie4b8e298f4f5e1cd82364da83e4344d4fc3621a3

9 years agoMerge "Inserting a node must also update its inputs users."
Nicolas Geoffray [Tue, 7 Oct 2014 15:32:24 +0000 (15:32 +0000)]
Merge "Inserting a node must also update its inputs users."

9 years agoInserting a node must also update its inputs users.
Nicolas Geoffray [Tue, 7 Oct 2014 13:14:27 +0000 (14:14 +0100)]
Inserting a node must also update its inputs users.

Change-Id: I55357564b81efcc0cf52fffdf23289696fe27dd1

9 years agoMerge "Propagate optimization flags to kMirOpCheck."
Vladimir Marko [Tue, 7 Oct 2014 14:26:47 +0000 (14:26 +0000)]
Merge "Propagate optimization flags to kMirOpCheck."

9 years agoMerge "Fix movw on x86/x86_64 to accept any 16bits immediate."
Nicolas Geoffray [Tue, 7 Oct 2014 14:14:14 +0000 (14:14 +0000)]
Merge "Fix movw on x86/x86_64 to accept any 16bits immediate."

9 years agoFix movw on x86/x86_64 to accept any 16bits immediate.
Nicolas Geoffray [Tue, 7 Oct 2014 13:54:48 +0000 (14:54 +0100)]
Fix movw on x86/x86_64 to accept any 16bits immediate.

Change-Id: I282eece0cd497431f207cec61852b4585ed3655c

9 years agoMerge "Do not use kDiesAtEntry when inputs must be in specific reg."
Nicolas Geoffray [Tue, 7 Oct 2014 11:29:10 +0000 (11:29 +0000)]
Merge "Do not use kDiesAtEntry when inputs must be in specific reg."

9 years agoDo not use kDiesAtEntry when inputs must be in specific reg.
Nicolas Geoffray [Tue, 7 Oct 2014 11:24:52 +0000 (12:24 +0100)]
Do not use kDiesAtEntry when inputs must be in specific reg.

The way the register allocator blocks registers currently
does not handle these cases. Since it only applies to x86 for now,
just ensure such requests cannot happen.

Change-Id: Idfa25532b9b4996a192d05800f56c6e44edd3a8a

9 years agoMerge "Add a prepare for register allocation pass."
Nicolas Geoffray [Tue, 7 Oct 2014 10:20:01 +0000 (10:20 +0000)]
Merge "Add a prepare for register allocation pass."

9 years agoMerge "Make which dex2oat version to use configurable."
Nicolas Geoffray [Tue, 7 Oct 2014 09:38:37 +0000 (09:38 +0000)]
Merge "Make which dex2oat version to use configurable."

9 years agoART: fix NullCheckElimination to preserve MIR_IGNORE_NULL_CHECK
Yevgeny Rouban [Wed, 1 Oct 2014 11:25:28 +0000 (18:25 +0700)]
ART: fix NullCheckElimination to preserve MIR_IGNORE_NULL_CHECK

If the MIRGraph::EliminateNullChecksAndInferTypes() function managed
to prove that some regs are non-null then it sets the flag
MIR_IGNORE_NULL_CHECK and resets this flag for all the other regs.
If some previous optimizations have already set MIR_IGNORE_NULL_CHECK
then it can be reset by EliminateNullChecksAndInferTypes. This way
NullCheckElimination discards some optimization efforts.
Optimization passes should not reset MIR_IGNORE_NULL_CHECK unless
they 100% sure NullCheck is needed.

This patch makes the NCE_TypeInference pass be conservative in
resetting MIR_IGNORE_NULL_CHECK.

Change-Id: I4ea74020968b5c5bd8e3af48211ffd4c6afd7f80
Signed-off-by: Yevgeny Rouban <yevgeny.y.rouban@intel.com>
9 years agoMerge "Fix a merge error in cl 108991."
Hiroshi Yamauchi [Mon, 6 Oct 2014 22:06:33 +0000 (22:06 +0000)]
Merge "Fix a merge error in cl 108991."

9 years agoMerge "Add way to warn about missing JNI_ABORT"
Mathieu Chartier [Mon, 6 Oct 2014 21:46:26 +0000 (21:46 +0000)]
Merge "Add way to warn about missing JNI_ABORT"

9 years agoFix a merge error in cl 108991.
Hiroshi Yamauchi [Mon, 6 Oct 2014 21:24:36 +0000 (14:24 -0700)]
Fix a merge error in cl 108991.

Bug: 12687968
Change-Id: I37603c154b252321b59a5ea02c6a92f5a9300667

9 years agoAdd way to warn about missing JNI_ABORT
Mathieu Chartier [Mon, 6 Oct 2014 20:41:33 +0000 (13:41 -0700)]
Add way to warn about missing JNI_ABORT

Bug: 16858794
Change-Id: I6794a14ee323ef95569cc7646619e6869771c7c6

9 years agoMerge "Merge host-run-test-jar and push-and-run-test-jar."
Nicolas Geoffray [Mon, 6 Oct 2014 20:01:08 +0000 (20:01 +0000)]
Merge "Merge host-run-test-jar and push-and-run-test-jar."

9 years agoMerge host-run-test-jar and push-and-run-test-jar.
Nicolas Geoffray [Mon, 6 Oct 2014 11:23:04 +0000 (12:23 +0100)]
Merge host-run-test-jar and push-and-run-test-jar.

- Too many code duplication between the two files.
- Also fix --zygote and --gdb invocations.

Change-Id: I1f99320e63a7a0f19ad1ad2b12693901b8c6bb43

9 years agoMerge "Handlerize methods across some GC points."
Hiroshi Yamauchi [Mon, 6 Oct 2014 18:22:05 +0000 (18:22 +0000)]
Merge "Handlerize methods across some GC points."

9 years agoAdd a prepare for register allocation pass.
Nicolas Geoffray [Tue, 30 Sep 2014 12:54:09 +0000 (13:54 +0100)]
Add a prepare for register allocation pass.

- Currently the pass just changes the uses of checks to the
  actual values.
- Also optimize array access, now that inputs can be constants.
- And fix another bug in the register allocator reveiled by
  this change.

Change-Id: I43be0dbde9330ee5c8f9d678de11361292d8bd98

9 years agoHandlerize methods across some GC points.
Hiroshi Yamauchi [Mon, 6 Oct 2014 17:46:14 +0000 (10:46 -0700)]
Handlerize methods across some GC points.

Bug: 12687968
Change-Id: I0d5b0a78488ba76db4d25991d8db95b24bb624e9

9 years agoMerge "Add support for inputs dying at entry of instructions."
Nicolas Geoffray [Mon, 6 Oct 2014 15:37:52 +0000 (15:37 +0000)]
Merge "Add support for inputs dying at entry of instructions."

9 years agoAdd support for inputs dying at entry of instructions.
Nicolas Geoffray [Tue, 30 Sep 2014 21:40:23 +0000 (22:40 +0100)]
Add support for inputs dying at entry of instructions.

- Start using it in places where it makes sense.
- Also improve suspend check on arm to use subs directly.

Change-Id: I09ac0589f5ccb9b850ee757c76dcbcf35ee8cd01

9 years agoMake which dex2oat version to use configurable.
Nicolas Geoffray [Mon, 6 Oct 2014 13:55:06 +0000 (14:55 +0100)]
Make which dex2oat version to use configurable.

Change-Id: I86503a51538c0b48ad96e8624b81461c2ba3279a

9 years agoFix to MirGraph::GetSSANameWithConst()
Mark Mendell [Mon, 6 Oct 2014 14:58:54 +0000 (10:58 -0400)]
Fix to MirGraph::GetSSANameWithConst()

Don't call ConstantValueWide() for a wide constant if high_word is set,
as it will DCHECK if you try to get the value for the high word.

Change-Id: I046ee3e6833ceb556a3c2dbc95699882d30d65f9
Signed-off-by: Mark Mendell <mark.p.mendell@intel.com>
9 years agoPropagate optimization flags to kMirOpCheck.
Vladimir Marko [Mon, 6 Oct 2014 09:52:20 +0000 (10:52 +0100)]
Propagate optimization flags to kMirOpCheck.

Bug: 17751474
Change-Id: I9561b641d956e21dac686c7cd414fa0c6a176147

9 years agoMerge "Respect USE_CLANG_PLATFORM_BUILD for art builds"
Brian Carlstrom [Sat, 4 Oct 2014 20:52:26 +0000 (20:52 +0000)]
Merge "Respect USE_CLANG_PLATFORM_BUILD for art builds"

9 years agoRespect USE_CLANG_PLATFORM_BUILD for art builds
Bernhard Rosenkraenzer [Sat, 4 Oct 2014 17:02:06 +0000 (19:02 +0200)]
Respect USE_CLANG_PLATFORM_BUILD for art builds

Without this patch, ART_TARGET_CLANG is set to false unconditionally.
It should be set to whatever USE_CLANG_PLATFORM_BUILD requests, if the
latter is set.

Change-Id: I29c4397c308a1549c2097aad2c997fe732be27ac
Signed-off-by: Bernhard Rosenkraenzer <Bernhard.Rosenkranzer@linaro.org>
9 years agoMerge "Make host-run-test-ar take --no-prebuild."
Nicolas Geoffray [Sat, 4 Oct 2014 09:42:17 +0000 (09:42 +0000)]
Merge "Make host-run-test-ar take --no-prebuild."

9 years agoMake host-run-test-ar take --no-prebuild.
Nicolas Geoffray [Sat, 4 Oct 2014 09:40:54 +0000 (10:40 +0100)]
Make host-run-test-ar take --no-prebuild.

Forgot to update this file in last commit.

Change-Id: I7f7c6ecbabecaaabdeffa3fb5a50a9e9e78d03c0

9 years agoMerge "Merge push-and-run-{prebuilt}-test-jar files."
Nicolas Geoffray [Sat, 4 Oct 2014 09:04:45 +0000 (09:04 +0000)]
Merge "Merge push-and-run-{prebuilt}-test-jar files."

9 years agoMerge "Add some missing TEMP_FAILURE_RETRYs around unlink(2)"
Brian Carlstrom [Sat, 4 Oct 2014 00:49:29 +0000 (00:49 +0000)]
Merge "Add some missing TEMP_FAILURE_RETRYs around unlink(2)"