OSDN Git Service

android-x86/art.git
10 years agoMerge "ART: ReEnable check for a gap between method's arg "in" vregs and rest of...
Bill Buzbee [Sun, 27 Jul 2014 02:46:40 +0000 (02:46 +0000)]
Merge "ART: ReEnable check for a gap between method's arg "in" vregs and rest of its vregs"

10 years agoMerge "ART: Native bridge command-line parameter"
Andreas Gampe [Sun, 27 Jul 2014 01:36:55 +0000 (01:36 +0000)]
Merge "ART: Native bridge command-line parameter"

9 years agoART: Native bridge command-line parameter
Andreas Gampe [Fri, 25 Jul 2014 09:32:19 +0000 (02:32 -0700)]
ART: Native bridge command-line parameter

Add a command-line parameter for the native bridge library, slight
refactor/cleanup.

Add run-test 115 to test the native bridge interface. Currently the
tests are black-listed for the target, as the setup for the test
is too complicated in the current infrastructure.

Change-Id: I6ccf19485e8c30b96e9f2fd5425278cb1ebd403f

9 years agoMerge "ART: Allow arrays with erroneous component type"
Andreas Gampe [Sun, 27 Jul 2014 01:31:16 +0000 (01:31 +0000)]
Merge "ART: Allow arrays with erroneous component type"

9 years agoART: Allow arrays with erroneous component type
Andreas Gampe [Thu, 24 Jul 2014 03:18:36 +0000 (20:18 -0700)]
ART: Allow arrays with erroneous component type

Array classes must tolerate having component type classes that are
erroneous. Change CreateArrayClass to use LookupClass when FindClass
failed.

Bug: 16019155
Change-Id: Id4868c5498431c85c199aa3cbecd23566dce3601

9 years agoMerge "ART: Special-case cb(n)z even for in-reg constant"
Andreas Gampe [Sat, 26 Jul 2014 19:57:14 +0000 (19:57 +0000)]
Merge "ART: Special-case cb(n)z even for in-reg constant"

9 years agoART: Special-case cb(n)z even for in-reg constant
Andreas Gampe [Sat, 26 Jul 2014 08:40:39 +0000 (01:40 -0700)]
ART: Special-case cb(n)z even for in-reg constant

Call out to OpCmpImmBranch in GenCompareAndBranch if the constant
is zero and we are testing == or !=, even when zero has been loaded
to a register already. This avoids a register size mismatch on 64b
architectures when basically doing a null check, and generally
squashes a cmp + branch to a cbz or cbnz on Arm and Mips. X86 is
not degraded.

Bug: 16562601
Change-Id: I1997760f43dc186a84247ad30ae91053f71d102d

10 years agoART: ReEnable check for a gap between method's arg "in" vregs and rest of its vregs
Alexei Zavjalov [Tue, 15 Jul 2014 19:17:58 +0000 (02:17 +0700)]
ART: ReEnable check for a gap between method's arg "in" vregs and rest of its vregs

The Quick backend's runtime model includes a gap between a method's
argument ("in") vregs and the rest of its vregs.  Handling a range instruction
which spans the gap is somewhat complicated, and should not happen
in normal usage of dx. Punt to the interpreter if it happened.

This issue relates to https://code.google.com/p/android/issues/detail?id=66371
Which was fixed, but the problem returned after some changes in the verify
flags set for the range bytecodes.

Change-Id: I214f32e7b4227a450257f30607135930d0adff53
Signed-off-by: Katkov, Serguei I <serguei.i.katkov@intel.com>
Signed-off-by: Alexei Zavjalov <alexei.zavjalov@intel.com>
10 years agoMerge "Fix access to long/double stack values from debugger"
Sebastien Hertz [Thu, 24 Jul 2014 07:24:12 +0000 (07:24 +0000)]
Merge "Fix access to long/double stack values from debugger"

10 years agoMerge "Increase limit to 2600 on host clang builds for frame size."
Stephen Hines [Thu, 24 Jul 2014 02:49:36 +0000 (02:49 +0000)]
Merge "Increase limit to 2600 on host clang builds for frame size."

10 years agoIncrease limit to 2600 on host clang builds for frame size.
Stephen Hines [Thu, 24 Jul 2014 02:47:35 +0000 (19:47 -0700)]
Increase limit to 2600 on host clang builds for frame size.

Change-Id: I573e1bcdc7a935145e88077def9c90a5573d2e90

10 years agoMerge "Update frame-larger-than limit to 2440 for new clang."
Stephen Hines [Thu, 24 Jul 2014 02:14:06 +0000 (02:14 +0000)]
Merge "Update frame-larger-than limit to 2440 for new clang."

10 years agoUpdate frame-larger-than limit to 2440 for new clang.
Stephen Hines [Thu, 17 Jul 2014 08:49:18 +0000 (01:49 -0700)]
Update frame-larger-than limit to 2440 for new clang.

Bug: 16362912

The latest version of clang is generating larger frames on elf_writer_quick.cc
and dex2oat.cc than we had previously encountered. This change unblocks the
LLVM rebase temporarily.

Change-Id: I20d9f88959b91e0509e58f01c22532720225f44d

10 years agoMerge "Reduce allocations in ParallelGC test."
Mathieu Chartier [Thu, 24 Jul 2014 01:30:03 +0000 (01:30 +0000)]
Merge "Reduce allocations in ParallelGC test."

10 years agoReduce allocations in ParallelGC test.
Mathieu Chartier [Thu, 24 Jul 2014 00:25:41 +0000 (17:25 -0700)]
Reduce allocations in ParallelGC test.

Fixes GC stress test mode.

Bug: 16406852
Change-Id: I86c8fd7516c3f41075bca00c7d5b4dab2bb7b1d4

10 years agoMerge "Use a (somewhat) more documented environment variable."
Nicolas Geoffray [Wed, 23 Jul 2014 23:49:26 +0000 (23:49 +0000)]
Merge "Use a (somewhat) more documented environment variable."

10 years agoUse a (somewhat) more documented environment variable.
Nicolas Geoffray [Wed, 23 Jul 2014 23:36:48 +0000 (00:36 +0100)]
Use a (somewhat) more documented environment variable.

I believe OUT is not mandatory to define, whereas
ANDROID_PRODUCT_OUT is. Not sure our continuous tests define it.

Change-Id: I7cd23971384df265e59587b49760aba29133a59c

10 years agoMerge "x86: GenSelect utility update"
Andreas Gampe [Wed, 23 Jul 2014 22:07:30 +0000 (22:07 +0000)]
Merge "x86: GenSelect utility update"

10 years agoMerge "Fix run-test --trace option if --host is specified after."
Jeff Hao [Wed, 23 Jul 2014 20:41:11 +0000 (20:41 +0000)]
Merge "Fix run-test --trace option if --host is specified after."

10 years agoFix run-test --trace option if --host is specified after.
Jeff Hao [Wed, 23 Jul 2014 18:52:52 +0000 (11:52 -0700)]
Fix run-test --trace option if --host is specified after.

Change-Id: I3911428ba762657c4433ab23e60f1771c9ddb2fe

10 years agoMerge "ART: Fix run-test 114 ParallelGC to account for OOM"
Andreas Gampe [Fri, 25 Jul 2014 23:43:42 +0000 (23:43 +0000)]
Merge "ART: Fix run-test 114 ParallelGC to account for OOM"

9 years agoART: Fix run-test 114 ParallelGC to account for OOM
Andreas Gampe [Fri, 25 Jul 2014 23:37:09 +0000 (16:37 -0700)]
ART: Fix run-test 114 ParallelGC to account for OOM

This catches any OOMs and doesn't pollute the log.

Bug: 16406852
Change-Id: I1bc95091ccae35a8cb5f2ef0a789f8c8ce5209d0

9 years agoMerge "Fix main space memory leak and add checks."
Mathieu Chartier [Fri, 25 Jul 2014 19:06:45 +0000 (19:06 +0000)]
Merge "Fix main space memory leak and add checks."

9 years agoFix main space memory leak and add checks.
Mathieu Chartier [Fri, 25 Jul 2014 18:50:47 +0000 (11:50 -0700)]
Fix main space memory leak and add checks.

The hypothesis is that we were leaking the main space and its
bitmaps, then eventually we would run out of virtual address space,
which would cause a null bitmap (DCHECK). Finally when we tried
adding the space with a null bitmap to the heap bitmap it segfaulted.

Changed some non performance critical DCHECK -> CHECK.

Bug: 16563323
Change-Id: I08a1f873752e28ebcf63ebbd90f92d994d7ca96b

9 years agoMerge "Fix dangling pointer bug when transitioning to background."
Mathieu Chartier [Fri, 25 Jul 2014 18:42:53 +0000 (18:42 +0000)]
Merge "Fix dangling pointer bug when transitioning to background."

9 years agoFix dangling pointer bug when transitioning to background.
Mathieu Chartier [Fri, 25 Jul 2014 17:13:51 +0000 (10:13 -0700)]
Fix dangling pointer bug when transitioning to background.

Dangling pointer left behind from the old rosalloc / dlmalloc
spaces. We now avoid using this pointer by using main_space_
and non_moving_space_ as well as clear the pointer when we remove
the space.

Bug: 16567203

Change-Id: Ida9ff30783e89cd4a4d86a4d0e912701692101f1

10 years agoMerge "ART: Clean up API after change 102631"
Andreas Gampe [Fri, 25 Jul 2014 22:37:09 +0000 (22:37 +0000)]
Merge "ART: Clean up API after change 102631"

9 years agoART: Clean up API after change 102631
Andreas Gampe [Fri, 25 Jul 2014 20:36:56 +0000 (13:36 -0700)]
ART: Clean up API after change 102631

This adds an abort_on_error parameter to GetCurrentMethod, which is
by default true. This restores all previous behavior, except for
monitor installation, where it follows 101639.

Bug: 16556938
Change-Id: I9a12b9a21ccb9f558c86236bb58d15ff2fafaec0

10 years agoMerge "Fix JNI compiler frame size adjustments."
Vladimir Marko [Fri, 25 Jul 2014 14:55:49 +0000 (14:55 +0000)]
Merge "Fix JNI compiler frame size adjustments."

9 years agoFix JNI compiler frame size adjustments.
Vladimir Marko [Thu, 24 Jul 2014 16:01:58 +0000 (17:01 +0100)]
Fix JNI compiler frame size adjustments.

Bug: 16321952
Change-Id: I6f84a59c679e335e02a8e70944a5d0bc8d73f90b

10 years agoMerge "Rename openDexFileNative to openDexFile."
Narayan Kamath [Fri, 25 Jul 2014 14:26:26 +0000 (14:26 +0000)]
Merge "Rename openDexFileNative to openDexFile."

9 years agoRename openDexFileNative to openDexFile.
Calin Juravle [Wed, 16 Jul 2014 16:55:55 +0000 (17:55 +0100)]
Rename openDexFileNative to openDexFile.

We no longer need two distinct methods.

Bug: 15563230

(cherry picked from commit 8edcb9c8bc62fcbd181e136c32086f2b970306a4)

Change-Id: Ib95098af0dd26733b5946d8e24ae4dfbcdbec3ba

10 years agoMerge "Add a new stack map scheme that encodes compilation info per pc."
Nicolas Geoffray [Fri, 25 Jul 2014 13:34:39 +0000 (13:34 +0000)]
Merge "Add a new stack map scheme that encodes compilation info per pc."

10 years agoMerge "ART: Fix wrong CHECK in GetCurrentLocationForThrow"
Andreas Gampe [Fri, 25 Jul 2014 10:37:19 +0000 (10:37 +0000)]
Merge "ART: Fix wrong CHECK in GetCurrentLocationForThrow"

9 years agoART: Fix wrong CHECK in GetCurrentLocationForThrow
Andreas Gampe [Fri, 25 Jul 2014 10:07:06 +0000 (03:07 -0700)]
ART: Fix wrong CHECK in GetCurrentLocationForThrow

Under certain circumstances kDexNoIndex is actually an encoded
dex pc. Change the setup of CurrentMethodVisitor to account for that.

Bug: 16556938
Change-Id: I4671f8c71c76f3e17e02db10f6da8ecc8331c87d

10 years agoMerge "ART: Turn on ART_TEST_KEEP_GOING by default, clean up unused vars"
Andreas Gampe [Fri, 25 Jul 2014 09:38:04 +0000 (09:38 +0000)]
Merge "ART: Turn on ART_TEST_KEEP_GOING by default, clean up unused vars"

9 years agoART: Turn on ART_TEST_KEEP_GOING by default, clean up unused vars
Andreas Gampe [Fri, 25 Jul 2014 04:41:06 +0000 (21:41 -0700)]
ART: Turn on ART_TEST_KEEP_GOING by default, clean up unused vars

Change-Id: I8db60eade0338e53e76a87ef38ad661550475e3c

10 years agoMerge "Integrate ART with NativeBridge interfaces"
Andreas Gampe [Fri, 25 Jul 2014 04:01:32 +0000 (04:01 +0000)]
Merge "Integrate ART with NativeBridge interfaces"

9 years agoIntegrate ART with NativeBridge interfaces
Yong WU [Thu, 24 Jul 2014 13:32:15 +0000 (21:32 +0800)]
Integrate ART with NativeBridge interfaces

Native-bridge will provide the following interfaces to ART:

struct NativeBridgeCallbacks {
  bool  (*initialize   )(NativeBridgeArtCallbacks* vm_itf);
  void* (*loadLibrary  )(const char* libpath, int flag);
  void* (*getTrampoline)(void* handle, const char* name, const char* shorty,
                         uint32_t len);
  bool  (*isSupported  )(const char* libpath);
};

Native-bridge will expose a symbol NativeBridgeItf with the
type of NativeBridgeCallbacks to ART.

And ART will provide the interfaces below to native-bridge:

struct NativeBridgeArtCallbacks {
  int   (*logger               )(int prio, const char* tag, const char* fmt, ...);
  const char* (*getMethodShorty)(JNIEnv* env, jmethodID mid);
  int   (*getNativeMethodCount )(JNIEnv* env, jclass clazz);
  int   (*getNativeMethods     )(JNIEnv* env, jclass clazz, JNINativeMethod* methods,
                                 uint32_t method_count);
};

Based on the interfaces, if an ART call to dlopen fails to open a native library,
it queries the native bridge by using NativeBridgeCallbacks::isSupported(). If the
native library is supported by native-bridge, ART can load the native library
using NativeBridgeCallbacks::loadLibrary() and get a trampoline for a specific
native method using NativeBridgeCallbacks::getTrampoline(). ART can then call
the native method using the normal signature and the address of the trampoline.

On the other side, in the case of a native method calling JNI native function
CallXXXXMethodY(), native-bridge calls back to Art for the shorty of the method
using NativeBridgeArtCallbacks::getMethodShorty() so that it can prepare based
on host calling convention.

In case of JNI function RegisterNatives()/UnregisterNatives(), native bridge can
call back to NativeBridgeArtCallbacks::getNativeMethodCount() and NativeBridgeArtCallbacks
::getNativeMethods() to get all native methods of specified class so that all
corresponding trampolines can be prepared/destroyed.

Class NativeBridge is created to encapsulate the function pointers of
NativeBridgeCallbacks and provides better abstraction to ART.

Note: functionality is turned off in native_bridge.cc at the moment.

Change-Id: I652755044957a7960254648652b538cce70dd011

10 years agoMerge "Fix memory leak in RemoveRememberedSet."
Mathieu Chartier [Thu, 24 Jul 2014 20:47:09 +0000 (20:47 +0000)]
Merge "Fix memory leak in RemoveRememberedSet."

10 years agoFix memory leak in RemoveRememberedSet.
Mathieu Chartier [Thu, 24 Jul 2014 18:11:05 +0000 (11:11 -0700)]
Fix memory leak in RemoveRememberedSet.

RemoveRememberedSet now deletes the remembered set.

Bug: 16532086
Change-Id: I01092931cc20cd0688dd42eed3dde9ad140889b2

10 years agoMerge "Fix stale remembered sets error."
Mathieu Chartier [Thu, 24 Jul 2014 17:53:11 +0000 (17:53 +0000)]
Merge "Fix stale remembered sets error."

10 years agoFix stale remembered sets error.
Mathieu Chartier [Thu, 24 Jul 2014 01:45:17 +0000 (18:45 -0700)]
Fix stale remembered sets error.

We were forgetting to remove the remembered set in transition to
background. This resulted in remembered sets being added for
spaces which no longer existed. This finally caused an error when
a new space happened to have the same address as the old space,
resulting in a CHECK failure.

Also tuned the number of ParallelGC to prevent spurrious failures
and removed the ParallelGC from broken tests in the make file.

Bug: 16532086
Bug: 16406852

Change-Id: I00bbcbd7daa03c867732d165be62b72e6c43bce1

10 years agoMerge "ART: Rewrite ParallelGC run-test"
Andreas Gampe [Sat, 26 Jul 2014 09:17:40 +0000 (09:17 +0000)]
Merge "ART: Rewrite ParallelGC run-test"

9 years agoART: Rewrite ParallelGC run-test
Andreas Gampe [Sat, 26 Jul 2014 08:13:13 +0000 (01:13 -0700)]
ART: Rewrite ParallelGC run-test

To better test what ParallelGC is intended to do, write a version
that will more fairly allocate objects between the threads, and
will let each thread OOM only once. Use a barrier to wait for
completion of all threads, and force a final GC by allocating some
more objects.

Bug: 16406852
Change-Id: I019ddc08515b9610c18d55280cd46cbe3ae3e5ac

10 years agoMerge "ART: Relax CurrentMethodVisitor requirements on GetDexPC"
Andreas Gampe [Fri, 25 Jul 2014 09:23:04 +0000 (09:23 +0000)]
Merge "ART: Relax CurrentMethodVisitor requirements on GetDexPC"

9 years agoART: Relax CurrentMethodVisitor requirements on GetDexPC
Andreas Gampe [Thu, 17 Jul 2014 05:20:31 +0000 (22:20 -0700)]
ART: Relax CurrentMethodVisitor requirements on GetDexPC

In case we want to dump a Java stack after an unhandled fault, in
case we hold a thinlocked monitor, that monitor might get inflated.
That can cause an abort as we may not have enough/correct information
for the state at the bottom-most call.

Relax GetDexPc in the CurrentMethodVisitor to not abort when it cannot
find a dex pc. Instead, let the caller handle such a case. This CL
allows the locking_dex_pc_ in Monitor to be DexFile::kDexNoIndex,
which avoids the above abort.

Bug: 1635280216556938
Change-Id: I3adf89b2d8f018a0c3e3abdd26e542f46ee59eef

10 years agoMerge "ART: Account for multidex location strings in VMClassLoader"
Andreas Gampe [Fri, 25 Jul 2014 08:46:44 +0000 (08:46 +0000)]
Merge "ART: Account for multidex location strings in VMClassLoader"

9 years agoART: Account for multidex location strings in VMClassLoader
Andreas Gampe [Thu, 24 Jul 2014 22:35:50 +0000 (15:35 -0700)]
ART: Account for multidex location strings in VMClassLoader

To look up resources, look in the unadorned location.

Bug: 16530747
Change-Id: Ia97e39366444f6666a78ade7298d3c22b4b79f8a

10 years agoMerge "Tweaks to patchoat and other related things."
Andreas Gampe [Sat, 26 Jul 2014 09:07:18 +0000 (09:07 +0000)]
Merge "Tweaks to patchoat and other related things."

9 years agoTweaks to patchoat and other related things.
Alex Light [Thu, 24 Jul 2014 18:29:14 +0000 (11:29 -0700)]
Tweaks to patchoat and other related things.

Removed some flags from patchoat that were poorly specified and fixed
some other issues with the relocation system.

Bug: 15358152

Change-Id: Ia6d47b1a008f02373307d833ba45f00ea408d76f

10 years agox86: GenSelect utility update
Serguei Katkov [Thu, 17 Jul 2014 07:39:03 +0000 (14:39 +0700)]
x86: GenSelect utility update

The is follow-up https://android-review.googlesource.com/#/c/101396/
to make x86 GenSelectConst32 implementation complete.

Change-Id: I69f318e18093f9a5b00f8f00f0f1c2e4ff7a9ab2
Signed-off-by: Serguei Katkov <serguei.i.katkov@intel.com>
10 years agoMerge "Add test-art-*-run-test-###-test-name32/64 test targets."
Andreas Gampe [Wed, 23 Jul 2014 18:07:50 +0000 (18:07 +0000)]
Merge "Add test-art-*-run-test-###-test-name32/64 test targets."

10 years agoAdd test-art-*-run-test-###-test-name32/64 test targets.
Alex Light [Wed, 23 Jul 2014 17:31:34 +0000 (10:31 -0700)]
Add test-art-*-run-test-###-test-name32/64 test targets.

This allows one to run only a single architecture for one of the run-tests.

Change-Id: Icb06c2b36074cbec7f7808a2fcaba97ec5501578

10 years agoMerge "ART: Fix x86_64 instrumentation_exit, also movsd -> movq"
Andreas Gampe [Wed, 23 Jul 2014 18:12:08 +0000 (18:12 +0000)]
Merge "ART: Fix x86_64 instrumentation_exit, also movsd -> movq"

10 years agoART: Fix x86_64 instrumentation_exit, also movsd -> movq
Andreas Gampe [Wed, 23 Jul 2014 17:05:02 +0000 (10:05 -0700)]
ART: Fix x86_64 instrumentation_exit, also movsd -> movq

Change movd/movsd to movq.

Bug: 16386215
Change-Id: Icca71ca2aeeb2917aff46043051d6046f04395d4

10 years agoMerge "ART: Fix erroneous output"
Andreas Gampe [Wed, 23 Jul 2014 17:51:24 +0000 (17:51 +0000)]
Merge "ART: Fix erroneous output"

10 years agoART: Fix erroneous output
Andreas Gampe [Wed, 23 Jul 2014 17:47:54 +0000 (10:47 -0700)]
ART: Fix erroneous output

Make run-tests quiet.

Change-Id: I80da9ac04fa95a4094dee874c8703ea563acc1ee

10 years agoMerge "Fix x86 instrumentation exit entrypoint and trace size limit."
Jeff Hao [Wed, 23 Jul 2014 01:50:09 +0000 (01:50 +0000)]
Merge "Fix x86 instrumentation exit entrypoint and trace size limit."

10 years agoFix x86 instrumentation exit entrypoint and trace size limit.
Jeff Hao [Wed, 23 Jul 2014 01:38:42 +0000 (18:38 -0700)]
Fix x86 instrumentation exit entrypoint and trace size limit.

The x86 instruction movd only copies a 32-bit value when used with a
memory location. movsd properly copies 64-bits.

Bug: 16386215

Change-Id: Ia59b2c2af72ce5202c7b62413687aedb48cfd25e

10 years agoMerge "Recycle mem-maps for collector transitions."
Mathieu Chartier [Wed, 23 Jul 2014 01:01:30 +0000 (01:01 +0000)]
Merge "Recycle mem-maps for collector transitions."

10 years agoRecycle mem-maps for collector transitions.
Mathieu Chartier [Wed, 16 Jul 2014 20:28:58 +0000 (13:28 -0700)]
Recycle mem-maps for collector transitions.

We now create spaces when we need them for collector transitions or
homogeneous compaction by recycling mem maps. Change the bump
pointer space size to be as large as the heap capacity instead of
1/2 heap capacity like it used to be. For GSS, bump pointer spaces
are set to 32MB currently.

Changed GSS to have main space == non moving space since we don't
need to copy from the main space.

Fixes GC stress tests 074, 096.
Fixed test 080 oom throw with -Xmx2m for GC stress test, this was
broken since it was allocating a 4 MB array before starting the
OOM process.

Bug: 14059466
Bug: 16406852
Change-Id: I62877cfa24ec944a6f34ffac30334f454a8002fd

10 years agoMerge "ART: Make run tests out of oat tests"
Andreas Gampe [Wed, 23 Jul 2014 17:03:23 +0000 (17:03 +0000)]
Merge "ART: Make run tests out of oat tests"

10 years agoART: Make run tests out of oat tests
Andreas Gampe [Wed, 23 Jul 2014 01:52:29 +0000 (18:52 -0700)]
ART: Make run tests out of oat tests

Transforms all former oat tests into run tests.

Change-Id: I190dd39456454c36e5538a2c044d993965a67533

10 years agoMerge "Changed default non-full tests to be prebuild."
Andreas Gampe [Wed, 23 Jul 2014 01:43:30 +0000 (01:43 +0000)]
Merge "Changed default non-full tests to be prebuild."

10 years agoChanged default non-full tests to be prebuild.
Alex Light [Wed, 23 Jul 2014 01:07:12 +0000 (18:07 -0700)]
Changed default non-full tests to be prebuild.

Also added ability to turn off non-prebuild manually.

Change-Id: I0315ac43703a9c7e79812658523013799fb5d5dd

10 years agoMerge "Fix build, missing spaces around =/<."
Mingyao Yang [Wed, 23 Jul 2014 00:35:01 +0000 (00:35 +0000)]
Merge "Fix build, missing spaces around =/<."

10 years agoFix build, missing spaces around =/<.
Mingyao Yang [Wed, 23 Jul 2014 00:33:25 +0000 (17:33 -0700)]
Fix build, missing spaces around =/<.

Change-Id: I2e7824075626a07eccb0a5eb77ef157214fe70fb

10 years agoMerge "Put oat test output on stderr."
Ian Rogers [Tue, 22 Jul 2014 22:05:57 +0000 (22:05 +0000)]
Merge "Put oat test output on stderr."

10 years agoPut oat test output on stderr.
Ian Rogers [Tue, 22 Jul 2014 21:22:22 +0000 (14:22 -0700)]
Put oat test output on stderr.

Change-Id: Iad318f63263dc5d264d8a84eacedf5065a4e5248

10 years agoMerge "ART: Fix checks for relocation delta"
Andreas Gampe [Tue, 22 Jul 2014 21:09:07 +0000 (21:09 +0000)]
Merge "ART: Fix checks for relocation delta"

10 years agoART: Fix checks for relocation delta
Andreas Gampe [Tue, 22 Jul 2014 21:04:34 +0000 (14:04 -0700)]
ART: Fix checks for relocation delta

Change < to <= and > to >=.

Change-Id: I7579e3c13dc1b33801ace3a0f8d85cadaee41e6e

10 years agoMerge "ART: Blacklist relocate option in oat tests"
Andreas Gampe [Tue, 22 Jul 2014 20:57:39 +0000 (20:57 +0000)]
Merge "ART: Blacklist relocate option in oat tests"

10 years agoART: Blacklist relocate option in oat tests
Andreas Gampe [Tue, 22 Jul 2014 18:46:28 +0000 (11:46 -0700)]
ART: Blacklist relocate option in oat tests

To avoid overwriting issues when running the relocate option for
oat tests, blacklist them.

Change-Id: I37685a761fd0949502da134690d88ba2a399ea44

10 years agoMerge "Use the TMPDIR environment variable for test directory if set."
Ian Rogers [Wed, 23 Jul 2014 15:18:54 +0000 (15:18 +0000)]
Merge "Use the TMPDIR environment variable for test directory if set."

10 years agoUse the TMPDIR environment variable for test directory if set.
Ian Rogers [Wed, 23 Jul 2014 15:03:36 +0000 (08:03 -0700)]
Use the TMPDIR environment variable for test directory if set.

Bug: 16499668

Change-Id: I7da5559c13597d7bca4e4037a1e9335f0b6d8230

10 years agoMerge "Fix implicit stack overflow check on optimizing/x86."
Nicolas Geoffray [Wed, 23 Jul 2014 12:26:17 +0000 (12:26 +0000)]
Merge "Fix implicit stack overflow check on optimizing/x86."

10 years agoFix implicit stack overflow check on optimizing/x86.
Nicolas Geoffray [Wed, 23 Jul 2014 11:57:19 +0000 (12:57 +0100)]
Fix implicit stack overflow check on optimizing/x86.

They need to happen before changing ESP, and require a suspend point.

Change-Id: Id41aa9c99714f7ab8591367ea5cb9ca105b17ce8

10 years agoMerge "Rewrite topological sort order and improve GVN."
Vladimir Marko [Wed, 23 Jul 2014 12:12:24 +0000 (12:12 +0000)]
Merge "Rewrite topological sort order and improve GVN."

10 years agoRewrite topological sort order and improve GVN.
Vladimir Marko [Thu, 10 Jul 2014 11:42:52 +0000 (12:42 +0100)]
Rewrite topological sort order and improve GVN.

Rewrite the topological sort order to include a full loop
before the blocks that go after the loop. Add a new iterator
class LoopRepeatingTopologicalSortIterator that differs from
the RepeatingTopologicalSortIterator by repeating only loops
and repeating them early. It returns to the loop head if the
head needs recalculation when we reach the end of the loop.

In GVN, use the new loop-repeating topological sort iterator
and for a loop head merge only the preceding blocks' LVNs
if we're not currently recalculating this loop.

Also fix LocalValueNumbering::InPlaceIntersectMaps() which
was keeping only the last element of the intersection, avoid
some unnecessary processing during LVN merge and add some
missing braces to MIRGraph::InferTypeAndSize().

Bug: 16398693
Change-Id: I4e10d4acb626a5b8a28ec0de106a7b37f9cbca32

10 years agoMerge "Use the correct HOST_ARCH for oat files."
Nicolas Geoffray [Wed, 23 Jul 2014 12:07:12 +0000 (12:07 +0000)]
Merge "Use the correct HOST_ARCH for oat files."

10 years agoUse the correct HOST_ARCH for oat files.
Nicolas Geoffray [Wed, 23 Jul 2014 11:37:10 +0000 (12:37 +0100)]
Use the correct HOST_ARCH for oat files.

This typo explains why test-art-host-oat-optimizing-SignalTest32
would work on a 32bit host build and fail on a 64bit host build:
because dex2oat does not generate the code in the expected path
(x86_64 instead of x86), dalvikvm would re-generate it, with
the default settings (that is with quick compiler), and hence the
test would not fail.

Change-Id: I08b5ec316deb4d669fe76066d33a91fc3ed9e3f9

10 years agoMerge "Fix null pointer check elimination for catch entries."
Vladimir Marko [Wed, 23 Jul 2014 12:05:26 +0000 (12:05 +0000)]
Merge "Fix null pointer check elimination for catch entries."

10 years agoMerge "Rules depend on DEX2OATD, not DEX2OATD_DEPENDENCY."
Nicolas Geoffray [Wed, 23 Jul 2014 09:12:11 +0000 (09:12 +0000)]
Merge "Rules depend on DEX2OATD, not DEX2OATD_DEPENDENCY."

10 years agoRules depend on DEX2OATD, not DEX2OATD_DEPENDENCY.
Nicolas Geoffray [Wed, 23 Jul 2014 08:59:21 +0000 (09:59 +0100)]
Rules depend on DEX2OATD, not DEX2OATD_DEPENDENCY.

If dex2oatd changes, we need to re-run .oat/.art file generation.

Change-Id: I43487e024745b22b1681f4ff558bac6fd07d9881

10 years agoFix null pointer check elimination for catch entries.
Vladimir Marko [Fri, 11 Jul 2014 13:44:36 +0000 (14:44 +0100)]
Fix null pointer check elimination for catch entries.

Remove the special treatment of catch blocks for null
pointer check elimination and class initialization check
elimination. In both cases this can help optimizing
previously missed cases. In the null check case, this
avoids incorrect optimization as exposed by the new test.

Bug: 16230771
Change-Id: I834b7a1835d9ca8572f4f8d8516d93913c701ad1

10 years agoMerge "Set vtable in class object to null after linking."
Mingyao Yang [Tue, 22 Jul 2014 22:05:50 +0000 (22:05 +0000)]
Merge "Set vtable in class object to null after linking."

10 years agoSet vtable in class object to null after linking.
Mingyao Yang [Wed, 16 Jul 2014 17:44:41 +0000 (10:44 -0700)]
Set vtable in class object to null after linking.

This is follow-up work of embedding imt and vtable for
faster interface/virtual call dispatching.
Once vtable becomes embedded, the original vtable is nulled.

Change-Id: I307696657d1e283654169dbecb8f7815c42bbabc

10 years agoMerge "Made sync-test actually wait for first thread to start."
Andreas Gampe [Tue, 22 Jul 2014 21:40:47 +0000 (21:40 +0000)]
Merge "Made sync-test actually wait for first thread to start."

10 years agoMade sync-test actually wait for first thread to start.
Alex Light [Tue, 22 Jul 2014 21:04:31 +0000 (14:04 -0700)]
Made sync-test actually wait for first thread to start.

This makes sure that the threads execute in the order we expect.

Change-Id: I0c4976e72ef2eb9e5fd66184eb1cd4821ea1e42b

10 years agoMerge "Runtime can now be set to require relocation"
Andreas Gampe [Tue, 22 Jul 2014 15:25:28 +0000 (15:25 +0000)]
Merge "Runtime can now be set to require relocation"

10 years agoRuntime can now be set to require relocation
Alex Light [Wed, 2 Jul 2014 23:28:08 +0000 (16:28 -0700)]
Runtime can now be set to require relocation

Add a pair of runtime flags -Xrelocate and -Xnorelocate that can force
the runtime to require that all files that are run are relocated, to
prevent attacks based on the known art base address.

Add support for running patchoat on oat files compiled without an image.

Change run-test to have new --prebuild and --relocate flags.

Bug: 15358152

Change-Id: I91166c62dd1ab80e5cbcb7883a2cd0d56afca32d

10 years agoMerge "Add read barriers for the GC roots in Instrumentation."
Hiroshi Yamauchi [Mon, 21 Jul 2014 18:01:22 +0000 (18:01 +0000)]
Merge "Add read barriers for the GC roots in Instrumentation."

10 years agoAdd read barriers for the GC roots in Instrumentation.
Hiroshi Yamauchi [Fri, 18 Jul 2014 22:38:17 +0000 (15:38 -0700)]
Add read barriers for the GC roots in Instrumentation.

Bug: 12687968
Change-Id: I324e2f950ce4500b0e00722044af3a9c82487b23

10 years agoAdd a new stack map scheme that encodes compilation info per pc.
Nicolas Geoffray [Wed, 2 Jul 2014 14:08:17 +0000 (15:08 +0100)]
Add a new stack map scheme that encodes compilation info per pc.

Encodes stack mask, register mask, dex register values, and inlining
info. The encoding is currently very straightforward: there is no
clever encoding, nor compression.

Change-Id: I5fd9ae28189a5513cd9e3c8d52c648463207643d

10 years agoMerge "Fix erroneous behaviors with OOME present."
Ian Rogers [Mon, 21 Jul 2014 16:37:48 +0000 (16:37 +0000)]
Merge "Fix erroneous behaviors with OOME present."

10 years agoFix erroneous behaviors with OOME present.
Ian Rogers [Mon, 21 Jul 2014 15:55:01 +0000 (08:55 -0700)]
Fix erroneous behaviors with OOME present.

Bug: 16454510
Change-Id: I757088a7b82ff73f58aba8d357080028b56442e6

10 years agoMerge "Make unit test tell if a method is a leaf."
Nicolas Geoffray [Tue, 22 Jul 2014 16:57:58 +0000 (16:57 +0000)]
Merge "Make unit test tell if a method is a leaf."

10 years agoMake unit test tell if a method is a leaf.
Nicolas Geoffray [Tue, 22 Jul 2014 16:47:56 +0000 (17:47 +0100)]
Make unit test tell if a method is a leaf.

The runtime is not initialized completely in gtests, so we
cannot run code (such as explicit stack overflow checks) that
look at tls values.

Change-Id: I74a4449b01eb203f1b411dda700e9459878d0d55

10 years agoMerge "Stack overflow checks and NPE checks for optimizing."
Nicolas Geoffray [Tue, 22 Jul 2014 15:12:37 +0000 (15:12 +0000)]
Merge "Stack overflow checks and NPE checks for optimizing."