OSDN Git Service

android-x86/art.git
8 years agoMerge "ART: Disambiguate access-checks mode from lock-counting" into nyc-dev
Andreas Gampe [Fri, 29 Apr 2016 23:13:53 +0000 (23:13 +0000)]
Merge "ART: Disambiguate access-checks mode from lock-counting" into nyc-dev

8 years agoMerge "ART: Fix Mac build" into nyc-dev
Andreas Gampe [Fri, 29 Apr 2016 23:04:50 +0000 (23:04 +0000)]
Merge "ART: Fix Mac build" into nyc-dev

8 years agoART: Fix Mac build
Andreas Gampe [Fri, 29 Apr 2016 22:04:49 +0000 (15:04 -0700)]
ART: Fix Mac build

Fix Mac build after fc6898769ae1ef91ec3e41c0a273401213cb82cd.

Bug: 28423466
Change-Id: I6942bc86f1b0819d84bced5499afb0a4d235a39e

8 years agoMerge "Add ability to see if a dex file is backed by an oat file." into nyc-dev
Jeff Hao [Fri, 29 Apr 2016 21:07:48 +0000 (21:07 +0000)]
Merge "Add ability to see if a dex file is backed by an oat file." into nyc-dev

8 years agoMerge "Deduplicate simple roots in hprof." into nyc-dev
Richard Uhler [Fri, 29 Apr 2016 19:51:13 +0000 (19:51 +0000)]
Merge "Deduplicate simple roots in hprof." into nyc-dev

8 years agoMerge "ART: Log all monitor operations to systrace" into nyc-dev
TreeHugger Robot [Fri, 29 Apr 2016 18:37:18 +0000 (18:37 +0000)]
Merge "ART: Log all monitor operations to systrace" into nyc-dev

8 years agoMerge "Add null checks to weak DexCache references" into nyc-dev
Brian Carlstrom [Fri, 29 Apr 2016 18:02:58 +0000 (18:02 +0000)]
Merge "Add null checks to weak DexCache references" into nyc-dev

8 years agoMerge "ART: Add missing EXPORT_PCs to x86/_64 mterp" into nyc-dev
Bill Buzbee [Fri, 29 Apr 2016 17:37:28 +0000 (17:37 +0000)]
Merge "ART: Add missing EXPORT_PCs to x86/_64 mterp" into nyc-dev

8 years agoAdd null checks to weak DexCache references
Brian Carlstrom [Fri, 29 Apr 2016 17:28:34 +0000 (10:28 -0700)]
Add null checks to weak DexCache references

Bug: 28452385
Change-Id: I88243af47b8e091c360927d1bdb867770f092475

8 years agoMerge "Add contender method for lock contention" into nyc-dev
Mathieu Chartier [Fri, 29 Apr 2016 17:24:42 +0000 (17:24 +0000)]
Merge "Add contender method for lock contention" into nyc-dev

8 years agoAdd contender method for lock contention
Mathieu Chartier [Fri, 29 Apr 2016 00:21:08 +0000 (17:21 -0700)]
Add contender method for lock contention

Now prints the method that is attempting to acquire the lock.

Bug: 28223501
Change-Id: I47c4261d14d8767a873594594b7675a176b71d70

8 years agoART: Log all monitor operations to systrace
Andreas Gampe [Tue, 26 Apr 2016 03:08:55 +0000 (20:08 -0700)]
ART: Log all monitor operations to systrace

Add a VLOG option ("-verbose:systrace-locks") to log all monitor
operations to systrace. This requires non-fastpath thread
entrypoints, and ATRACE tags for locking and unlocking.

Do a bit of cleanup to the entrypoint initialization to share
common setup.

Bug: 28423466
Change-Id: Ie67e4aa946ec15f8fcf8cb7134c5d3cff0119ab3

8 years agoART: Add missing EXPORT_PCs to x86/_64 mterp
buzbee [Thu, 28 Apr 2016 23:08:44 +0000 (16:08 -0700)]
ART: Add missing EXPORT_PCs to x86/_64 mterp

Any time there is a possiblity that execution flow will leave
the mterp environment and transition either to an alternate
interpreter or compiled code, we need to export the cached copy of
the current Dalvik PC.  One path out related to on-stack-replacement
was missed in both x86 and x86_64 versions of mterp (other targets
are OK).

Note that there is an outstanding feature request to add support
to catch missing EXPORT_PCs (Bug: 26572899), though as envisioned
it would not have caught this one.

Bug: 28249238

(cherry picked from commit 963758dbfbbbc95bbfa82d5e02258d9fd481a459)

Change-Id: I3936509665636c0b77580dda12520a82ee814fa5

8 years agoRemove invoke after inlining.
Nicolas Geoffray [Fri, 29 Apr 2016 12:56:01 +0000 (13:56 +0100)]
Remove invoke after inlining.

bug:28460178

(cherry picked from commit 1be7cbd532a41981019f1e89c8ecf404bc9af809)

Change-Id: I9a84af048600a4653af284a9b171c2ba4c4dcca5

8 years agoMerge "Fix how the ProfileSaver thread attaches to the runtime." into nyc-dev
Calin Juravle [Fri, 29 Apr 2016 11:15:33 +0000 (11:15 +0000)]
Merge "Fix how the ProfileSaver thread attaches to the runtime." into nyc-dev

8 years agoFix how the ProfileSaver thread attaches to the runtime.
Calin Juravle [Thu, 28 Apr 2016 11:59:33 +0000 (12:59 +0100)]
Fix how the ProfileSaver thread attaches to the runtime.

If the runtime is shutting down we can't attach new threads. In this
case it is expected that Runtime::Attach returns nullptr.

Also, lower the serverity of the log to warning so that it does not
show up in the run-tests.

Bug: 28435649
Change-Id: I208bcd6f0f7e35ae6471e263dee00d3ab68b3b8f

8 years agoART: Disambiguate access-checks mode from lock-counting
Andreas Gampe [Thu, 28 Apr 2016 21:56:54 +0000 (14:56 -0700)]
ART: Disambiguate access-checks mode from lock-counting

Lock-counting (when structural locking verification failed) is a
special sub-mode of access-checks and must be disambiguated, because
we currently use access-checks mode class-wide when at least one
method soft-fails, but do not stop the compiler/JIT to compile
the "working" methods. So we may end up in the access-checks
interpreter for a working method through deopt without knowing
which locks are already held.

Bug: 28351535
Change-Id: I083032f064d88df8f8f0611ad8b57d1b39cd09fb

8 years agoRevert "Revert "Write conflict tables in image""
Mathieu Chartier [Thu, 28 Apr 2016 16:44:54 +0000 (09:44 -0700)]
Revert "Revert "Write conflict tables in image""

Added test.

Bug: 27906566

This reverts commit 8e2478d23e89a7022c93ddc608dcbba7b29b91e6.

Change-Id: I0894f5f7cd11af29ed9d0345b51f527fc8a41d19

8 years agoAdd ability to see if a dex file is backed by an oat file.
Jeff Hao [Fri, 22 Apr 2016 21:00:06 +0000 (14:00 -0700)]
Add ability to see if a dex file is backed by an oat file.

Native support for new DexFile.isBackedByOatFile.

Part of a multi-project change.

Bug: 26880306

(cherry-picked from commit 93a66cc779898d46c620e5d9c8bc71e2a71f6610)

Change-Id: I15e8ec44a216b13e494affd2fd3ff5f66fa79597

8 years agoMerge "Fix -Xint and -Xusejit interaction" into nyc-dev
Calin Juravle [Thu, 28 Apr 2016 15:32:33 +0000 (15:32 +0000)]
Merge "Fix -Xint and -Xusejit interaction" into nyc-dev

8 years agoMerge "ART: add exynos-m1 to a53 #835769 & #843419 erratum exception list" into nyc-dev
Vladimir Marko [Thu, 28 Apr 2016 15:20:32 +0000 (15:20 +0000)]
Merge "ART: add exynos-m1 to a53 #835769 & #843419 erratum exception list" into nyc-dev

8 years agoMerge "Decrease transition weight." into nyc-dev
TreeHugger Robot [Thu, 28 Apr 2016 14:20:59 +0000 (14:20 +0000)]
Merge "Decrease transition weight." into nyc-dev

8 years agoFix -Xint and -Xusejit interaction
Calin Juravle [Wed, 27 Apr 2016 17:50:45 +0000 (18:50 +0100)]
Fix -Xint and -Xusejit interaction

- don't allow interpret and jit compilation at the same time
- make sure we -Xsaveprofileinfo works with -Xint.

Bug: 27916886
Change-Id: Icab85e5b6f576f82374d5f88145cc81595afce37

8 years agoART: add exynos-m1 to a53 #835769 & #843419 erratum exception list
Junmo Park [Wed, 27 Apr 2016 07:51:50 +0000 (16:51 +0900)]
ART: add exynos-m1 to a53 #835769 & #843419 erratum exception list

Exynos-M1 is custom-designed 64-bit ARM CPU and does not need this A53
erratum handling.

Bug: 28412250
Signed-off-by: Junmo Park <junmoz.park@samsung.com>
(cherry picked from commit 2bd7d2d6d92daa796ee33e70bb788d3dae1f2fc9)

Change-Id: Id460e4e98baafeed1cf3540cec754b30044cda11

8 years agoDecrease transition weight.
Nicolas Geoffray [Thu, 28 Apr 2016 12:56:04 +0000 (13:56 +0100)]
Decrease transition weight.

bug:27865109
Change-Id: I551040e61f36fc675c0f5d1c3907c690112a84be

8 years agoForbid HDeoptimize instructions in OSR methods.
Nicolas Geoffray [Fri, 22 Apr 2016 12:16:14 +0000 (13:16 +0100)]
Forbid HDeoptimize instructions in OSR methods.

Otherwise dominated instructions will assume something that
isn't necessarily correct if coming from the interpreter.

bug:28335959
bug:28249238
bug:28348878
bug:28080135

Contains fix from https://android-review.googlesource.com/#/c/220661/.

(cherry picked from commit 93a18c5d4160f632ecdb92af099574e9c7098c49)

Change-Id: I86c3f9340077caa0a3e3db896e0519e7d38d91a0

8 years agoMerge "ART: Disable back edge uses for irreducible loops" into nyc-dev
David Brazdil [Thu, 28 Apr 2016 08:23:15 +0000 (08:23 +0000)]
Merge "ART: Disable back edge uses for irreducible loops" into nyc-dev

8 years agoMerge "Initialize native loader on JNI_CreateJavaVM" into nyc-dev
Dimitry Ivanov [Wed, 27 Apr 2016 23:37:10 +0000 (23:37 +0000)]
Merge "Initialize native loader on JNI_CreateJavaVM" into nyc-dev

8 years agoMerge "Revert "Write conflict tables in image"" into nyc-dev
Mathieu Chartier [Wed, 27 Apr 2016 21:20:06 +0000 (21:20 +0000)]
Merge "Revert "Write conflict tables in image"" into nyc-dev

8 years agoRevert "Write conflict tables in image"
Mathieu Chartier [Wed, 27 Apr 2016 21:03:42 +0000 (21:03 +0000)]
Revert "Write conflict tables in image"

Some strange issues on angler.

This reverts commit cda9386add68d94697449c6cb08b356747e55c21.

Change-Id: Iba3a8bac6ce40ee16d83163823356951fecb6787

8 years agoInitialize native loader on JNI_CreateJavaVM
Dimitry Ivanov [Thu, 21 Apr 2016 23:01:24 +0000 (16:01 -0700)]
Initialize native loader on JNI_CreateJavaVM

We need to initialize native loader in order to
be able to use JNI from any VM, be it dalvikvm
or VM started by zygote.

Bug: http://b/28082914
Bug: http://b/28320913
Change-Id: Ice9a33a736d7c6ec3af9a5a0fb25a378ad9d2c59
(cherry picked from commit 041169f7ca703d560a819e7a568170731d4ecdd6)

8 years agoMerge "Support to pass <uses-library> option through to dex2oat." into nyc-dev
Jeff Hao [Wed, 27 Apr 2016 19:54:02 +0000 (19:54 +0000)]
Merge "Support to pass <uses-library> option through to dex2oat." into nyc-dev

8 years agoMerge "Write conflict tables in image" into nyc-dev
Mathieu Chartier [Wed, 27 Apr 2016 18:29:20 +0000 (18:29 +0000)]
Merge "Write conflict tables in image" into nyc-dev

8 years agoSupport to pass <uses-library> option through to dex2oat.
Jeff Hao [Wed, 27 Apr 2016 18:07:41 +0000 (11:07 -0700)]
Support to pass <uses-library> option through to dex2oat.

This change takes an app's shared libraries specified by <uses-library>
and passes it through to dex2oat to be used during compilation.

Part of a multi-project change.

Includes fix from a6d46161aea07ebd1cbd6ab78b2b323f940e9c1e

Bug: 26880306

(cherry-picked from commit 26e8a2f150cd7f7195a10650ab8a5b6fa5014bc8)

Change-Id: I6bfc13693dbb835ca52fed2d03ec5346d43ec5d9

8 years agoWrite conflict tables in image
Mathieu Chartier [Thu, 14 Apr 2016 17:49:19 +0000 (10:49 -0700)]
Write conflict tables in image

Add image sections for runtime methods and conflict tables. This
makes it that we do not need to fake up a length prefixed array
for runtime methods.

Reduces .art private dirty and PSS.

System wide .art PSS goes from 32.5MB to 30.5MB after system boot.

Business card .art private dirty goes from 588K to 504K.

Increases image size by ~100K.

Bug: 27906566

Change-Id: I38cbe3150c3eeb385b8cad7cf93614e3980f4162

8 years agoART: Disable back edge uses for irreducible loops
David Brazdil [Wed, 27 Apr 2016 14:33:22 +0000 (15:33 +0100)]
ART: Disable back edge uses for irreducible loops

Algorithm adding back edge uses in liveness analysis makes assumptions
about the linear order which are not met in the presence of irreducible
loops. Disable back edges uses when the graph contains them.

This partially reverts CL I63632e8819ea3644d5c6fdfea00b66128bf22c24.

Bug: 28252747
Bug: 27615840
Bug: 27624868
Change-Id: I7ecdde0ed8a8831f7513b8e43cf7d84599b830a7
(cherry picked from commit 07b35109c64fc449a14723b259df17806fd0b16e)

8 years agoAdd command line option for jit invoke transition weight
Calin Juravle [Wed, 27 Apr 2016 13:14:58 +0000 (14:14 +0100)]
Add command line option for jit invoke transition weight

Bug: 27865109
Change-Id: If5eb53714e2c92397ad3e45bc0b4facb84329bc2

8 years agoSplit profile recording from jit compilation
Calin Juravle [Wed, 20 Apr 2016 13:22:09 +0000 (14:22 +0100)]
Split profile recording from jit compilation

We still use ProfileInfo objects to record profile information. That
gives us the flexibility to add the inline caches in the future and the
convenience of the already implemented GC.

If UseJIT is false and SaveProfilingInfo true, we will only record the
ProfileInfo and never launch compilation tasks.

Bug: 27916886
Change-Id: I6e4768dc5d58f2f85f947b276b4244aa11ce3fca

8 years agoMerge "Thumb2: Reduce memory used for fixup dependencies." into nyc-dev
Vladimir Marko [Wed, 27 Apr 2016 09:04:21 +0000 (09:04 +0000)]
Merge "Thumb2: Reduce memory used for fixup dependencies." into nyc-dev

8 years agoDeduplicate simple roots in hprof.
Richard Uhler [Tue, 26 Apr 2016 20:28:59 +0000 (13:28 -0700)]
Deduplicate simple roots in hprof.

Bug: 28311991
Change-Id: Id5ee7d7753016fafdec0c8f5cda5b1b052fa9ad1

8 years agoMerge "Test component type for errors too. With regression test." into nyc-dev
Aart Bik [Tue, 26 Apr 2016 19:55:48 +0000 (19:55 +0000)]
Merge "Test component type for errors too. With regression test." into nyc-dev

8 years agoTest component type for errors too.
Aart Bik [Mon, 25 Apr 2016 19:51:37 +0000 (12:51 -0700)]
Test component type for errors too.
With regression test.

Rationale:
Moved erroneous check in convenience method, so we
put all the same logic in one place. When testing
for erroneous T[], check both the array type
as well at the component type T for errors
(it is possible T[] is not marked erroneous
even though T is eventually).

BUG=28358598

(cherry picked from commit f417ff44d1eb111854d7a213f106912b3dd9e3d4)

Change-Id: Ieba66aa4b55d8e7ebddf200239c7e4095dfd4678

8 years agoThumb2: Reduce memory used for fixup dependencies.
Vladimir Marko [Mon, 25 Apr 2016 12:43:50 +0000 (12:43 +0000)]
Thumb2: Reduce memory used for fixup dependencies.

Mark fixups for far backward branches as 32-bit from the
beginning and ignore their dependencies since they cannot
be expanded anymore. This reduces the memory used by the
fixup_dependents_ when compiling methods with a lot of
slow paths.

Bug: 28256882

This is a cherry-pick of the fixed revert-revert
    Revert "Revert "Thumb2: Reduce memory used for fixup dependencies.""
    (cherry picked from commit a64f249add4a0a9e93aa6f246045ca79f5916dcd)

Change-Id: I9ca766b73d9f5a1da5dc10ad070145ac88fbcd53

8 years agoReduce memory lost by ArenaAllocator for large allocations.
Vladimir Marko [Fri, 22 Apr 2016 17:07:13 +0000 (18:07 +0100)]
Reduce memory lost by ArenaAllocator for large allocations.

When allocating from a new arena, check if the old arena has
more remaining space than the new one after the current
allocation. If so, keep using the old arena to reduce the
amount of "lost" arena memory. This can happen when we try
to allocate more than half the default arena size. If the
allocation exceeds the default arena size, it's very likely
to happen even though the ArenaPool could still provide some
much larger previously allocated arena.

Also avoid artithmetic overflow when checking if the
request can be satisfied from the current arena.
And abort immediately if calloc() fails.

Bug: 28173563
Bug: 28256882

In addition to the initial CL
    (cherry picked from commit 3e0e7173c0cdfc57dba39fe781e30d187d50fa9c)
this contains a squashed subsequent fix
    Fix valgrind tests: mark allocated space as defined.
    (cherry picked from commit 3f84f2cb3cadc25d75e1e3e2c1bc26c1a671f336)

Change-Id: Id80d5601874e8e28d930c0dd47a51c73c4810094

8 years agoMerge "ART: Change x86 from modify_ldt to set_thread_data" into nyc-dev
Andreas Gampe [Tue, 26 Apr 2016 00:32:38 +0000 (00:32 +0000)]
Merge "ART: Change x86 from modify_ldt to set_thread_data" into nyc-dev

8 years agoMerge "Make a fake libart for misbehaving apps." into nyc-dev
TreeHugger Robot [Mon, 25 Apr 2016 21:53:29 +0000 (21:53 +0000)]
Merge "Make a fake libart for misbehaving apps." into nyc-dev

8 years agoMake sure the referring class is in the dex cache.
Nicolas Geoffray [Mon, 25 Apr 2016 13:58:06 +0000 (14:58 +0100)]
Make sure the referring class is in the dex cache.

The method CanAccessResolvedMethod expects the referring class
in the dex file is already in the dex cache, which is true during AOT,
but not necessarilly during JIT.

bug:28295348

(cherry picked from commit 393fdb8b4822d80bbbd6347b088e28c03a72289e)

Change-Id: I9665d377070278639eb4b5a6eeced85d656e6cb6

8 years agoART: Fix ROR Checker test
David Brazdil [Mon, 25 Apr 2016 16:47:43 +0000 (17:47 +0100)]
ART: Fix ROR Checker test

Recent update of Jack started encoding "x << -2" as "x << 30". Update
the corresponding Checker test to accept both values.

Bug: 28244190
Change-Id: I65833afa416a7eb8fa05d381cff547a58dcc0bea
(cherry picked from commit 22c1f543ed75114d8ee5219598ba766e7dc674c0)

8 years agoART: Change x86 from modify_ldt to set_thread_data
Andreas Gampe [Tue, 16 Feb 2016 23:58:20 +0000 (15:58 -0800)]
ART: Change x86 from modify_ldt to set_thread_data

Do not use modify_ldt, so it can be deprecated.

Bug: 27199066
Change-Id: Ib5fb40d4f720b697d786c5c971638fd57681d308
(cherry picked from commit 8a2c62c62b3398afbac00b1cb8772ae5b53b62a3)

8 years agoAdd flag to ArtMethod to skip compilation.
Nicolas Geoffray [Wed, 20 Apr 2016 15:27:53 +0000 (16:27 +0100)]
Add flag to ArtMethod to skip compilation.

To ensure even the JIT will not try to compile methods with
soft failures a runtime_throw.

bug:28293819
bug:28313047

(cherry picked from commit 250a378d5a2152662e0fa820f2b38f794ddd3596)

Change-Id: Ic6d019bc2dd24e35e1377a6c3f8530348c1049df

8 years agoMerge "ART: Naive NullCheck elimination in InstructionBuilder" into nyc-dev
David Brazdil [Mon, 25 Apr 2016 08:22:00 +0000 (08:22 +0000)]
Merge "ART: Naive NullCheck elimination in InstructionBuilder" into nyc-dev

8 years agoMerge "Add support for eagerly calculating conflict tables" into nyc-dev
Mathieu Chartier [Fri, 22 Apr 2016 20:34:17 +0000 (20:34 +0000)]
Merge "Add support for eagerly calculating conflict tables" into nyc-dev

8 years agoAdd support for eagerly calculating conflict tables
Mathieu Chartier [Thu, 14 Apr 2016 17:49:19 +0000 (10:49 -0700)]
Add support for eagerly calculating conflict tables

Will be used to put them in the image by having the compiler eagerly
calculate them.

Enabled for debug builds (non compiler). Support for having conflict
tables written in the image will come in the next CL.

Bug: 27906566

Change-Id: I03d1671a4b49317aaab5a741bbeaed7957cd6229

8 years agoMerge "ART: Disable unstarted runtime pow test for valgrind" into nyc-dev
TreeHugger Robot [Fri, 22 Apr 2016 19:32:44 +0000 (19:32 +0000)]
Merge "ART: Disable unstarted runtime pow test for valgrind" into nyc-dev

8 years agoMerge "ART: Slightly change InitializeClass flow" into nyc-dev
TreeHugger Robot [Fri, 22 Apr 2016 18:36:16 +0000 (18:36 +0000)]
Merge "ART: Slightly change InitializeClass flow" into nyc-dev

8 years agoART: Disable unstarted runtime pow test for valgrind
Andreas Gampe [Fri, 22 Apr 2016 00:23:31 +0000 (17:23 -0700)]
ART: Disable unstarted runtime pow test for valgrind

Valgrind seems to give the wrong result. Disable test for now.

Bug: 27265238

(cherry picked from commit b6795157fd010375ca646ea95e972b616b4aeeaf)

Change-Id: I2b136abc107b53d89b70ccb20dd378a7b704d077

8 years agoART: Naive NullCheck elimination in InstructionBuilder
David Brazdil [Fri, 22 Apr 2016 15:57:00 +0000 (16:57 +0100)]
ART: Naive NullCheck elimination in InstructionBuilder

Save a little bit of memory by not generating trivially redundant
HNullCheck instructions. This patch builds on the fact that the
InstructionBuilder now directly generates SSA form and looks at the
input of the NullCheck. For obvious cases, such as NewInstance/Array
or `this`, the NullCheck generation is completely avoided.

Bug: 28173563
Change-Id: I1fdf3b096c7a939f7b8586c02a2a6b44dfa43443
(cherry picked from commit c120bbe26fe116d3c9d8322f44bb9e330e07f745)

8 years agoART: Slightly change InitializeClass flow
Andreas Gampe [Thu, 21 Apr 2016 19:21:55 +0000 (12:21 -0700)]
ART: Slightly change InitializeClass flow

Since 884f3b83ed6b2a378535ac6b2be57d6b2e22de09, verification isn't
run completely under a class' lock. This means it is possible to
race from unverified to initialized in InitializeClass. So check
the class state after VerifyClass, and handle new success and
failure cases.

Bug: 28254258
Change-Id: I22a6121477e409987281bc81c28b6c942f1bd319

8 years agoMerge "Make dex file loading failure more explicit about failure." into nyc-dev
TreeHugger Robot [Thu, 21 Apr 2016 23:14:05 +0000 (23:14 +0000)]
Merge "Make dex file loading failure more explicit about failure." into nyc-dev

8 years agoMerge "Fix interaction between instrumentation and jit lock." into nyc-dev
TreeHugger Robot [Thu, 21 Apr 2016 22:16:57 +0000 (22:16 +0000)]
Merge "Fix interaction between instrumentation and jit lock." into nyc-dev

8 years agoMake dex file loading failure more explicit about failure.
Alex Light [Wed, 20 Apr 2016 21:26:34 +0000 (14:26 -0700)]
Make dex file loading failure more explicit about failure.

We make sure that if we fail to load a dex file due to having an
unsupported dex version number we will say so explicitly in the
exceptions.

Bug: 28269864
Change-Id: I3c61d088cd0d04cf2a02246e69a4161180e3f2d8
(cherry picked from commit 3045b66613404fa973aafc8c2aae3728e9c12d9a)

8 years agoFix interaction between instrumentation and jit lock.
Nicolas Geoffray [Mon, 18 Apr 2016 11:09:30 +0000 (12:09 +0100)]
Fix interaction between instrumentation and jit lock.

The jit lock needs to have higher priority than the
deoptimized methods lock.

bug:28236735

(cherry picked from commit 480d5108fa62f28cbc2e7072610f974953ff73a8)

Change-Id: I6e5f8597c4ab88993f6a32a204383b766ff03e33

8 years agoMerge "Fix heap check." into nyc-dev
Mathieu Chartier [Thu, 21 Apr 2016 19:01:32 +0000 (19:01 +0000)]
Merge "Fix heap check." into nyc-dev

8 years agoFix heap check.
Mathieu Chartier [Thu, 21 Apr 2016 17:23:16 +0000 (10:23 -0700)]
Fix heap check.

Previously, we only checked the image we were patching. This did not
work if the referent was in another image. Fixed the DCHECK to check
all the boot images.

Bug: 28286961
Change-Id: I7195314e531d0a950d495296d1e2f953e7b66659

8 years agoMerge "Use dex cache from compilation unit in RTP." into nyc-dev
Vladimir Marko [Thu, 21 Apr 2016 16:52:32 +0000 (16:52 +0000)]
Merge "Use dex cache from compilation unit in RTP." into nyc-dev

8 years agoMerge "Thumb2: Fix disassembly of the b.w offset." into nyc-dev
Vladimir Marko [Thu, 21 Apr 2016 16:52:08 +0000 (16:52 +0000)]
Merge "Thumb2: Fix disassembly of the b.w offset." into nyc-dev

8 years agoART: Address late comments on a GVN memory-saving CL
David Brazdil [Thu, 21 Apr 2016 13:00:15 +0000 (14:00 +0100)]
ART: Address late comments on a GVN memory-saving CL

Added extra comments and removed redundant code as requested.

Bug: 28173563
Bug: 28287086

Change-Id: If6aff68c4c30427a86a27ffba5df1ae135edd294
(cherry picked from commit 94408d3144061bd6efc74b3d884d38169969c63f)

8 years agoThumb2: Fix disassembly of the b.w offset.
Vladimir Marko [Thu, 21 Apr 2016 13:52:03 +0000 (14:52 +0100)]
Thumb2: Fix disassembly of the b.w offset.

Bug: 28311085

(cherry picked from commit cb55b29e466939d222b83f0593a1141221f2425c)

Change-Id: I36ec3ef951ef9d2e5a9e60c0709cfeaf1210ddd0

8 years agoUse dex cache from compilation unit in RTP.
Vladimir Marko [Tue, 19 Apr 2016 14:12:13 +0000 (14:12 +0000)]
Use dex cache from compilation unit in RTP.

Avoid calling the costly ClassLinker::FindDexCache() from
reference type propagation when the dex cache from the
compilation unit will do, i.e. almost always. Compiling
the Nexus 5 boot image on host under perf(1) shows that
the FindDexCache() hits drop from about 0.2% to almost
nothing, though enabling inlining for the boot image will
increase it a bit to 0.03% due to unavoidable calls from
the inliner.

Also clean up the ScopedObjectAccess usage a bit.

Bug: 28173563

Cherry-picked the "revert-revert"
    (cherry picked from commit 456307a47336e3d6576ed6d8563b67573a4238d3)
and squashed two subsequent fixes
    Fix RTP to hold mutator lock while using raw mirror pointers.
    (cherry picked from commit 62977ff198deb673a6990202a2fb8b993217c57c)
    Fix reference_type_propagation_test.
    (cherry picked from commit 5eed0c5d27f091c952704f652cd77c4e3833ad88)

Change-Id: Ia944452d7ab26aed963832a9346df363743a419f

8 years agoReduce memory usage in GVN
David Brazdil [Wed, 20 Apr 2016 13:24:12 +0000 (14:24 +0100)]
Reduce memory usage in GVN

Implement recycling of ValueSet data structures which the GVN
algorithm will not access any more.

Savings depend on the shape of the graph, but can be as high as 93%.
Peak memory usage for GSA drops from 32MB to 26MB, compile times seem
unaffected.

Bug: 28173563
Bug: 28287086

Change-Id: If227177449bc90ad24fa68c37b0c2615924af1ed
(cherry picked from commit cc857cfbe4a179dfa7935b7334f1efbb21f2ac76)

8 years agoMerge "Thumb2: Fix EmitJumpTables() to extend buffer only if needed." into nyc-dev
Vladimir Marko [Thu, 21 Apr 2016 07:54:30 +0000 (07:54 +0000)]
Merge "Thumb2: Fix EmitJumpTables() to extend buffer only if needed." into nyc-dev

8 years agoMerge "Fix HInstruction::ReplaceInput(), allow no-op." into nyc-dev
Vladimir Marko [Thu, 21 Apr 2016 07:53:47 +0000 (07:53 +0000)]
Merge "Fix HInstruction::ReplaceInput(), allow no-op." into nyc-dev

8 years agoMake a fake libart for misbehaving apps.
Alex Light [Fri, 15 Apr 2016 19:41:09 +0000 (12:41 -0700)]
Make a fake libart for misbehaving apps.

Bug: 27775991
Change-Id: I22864aa22d4864d85e055a5b1d3ff1faf122011b

8 years agoMerge "ART: Add logging for CHECK failure" into nyc-dev
TreeHugger Robot [Wed, 20 Apr 2016 19:03:49 +0000 (19:03 +0000)]
Merge "ART: Add logging for CHECK failure" into nyc-dev

8 years agoFix HInstruction::ReplaceInput(), allow no-op.
Vladimir Marko [Wed, 20 Apr 2016 17:45:25 +0000 (18:45 +0100)]
Fix HInstruction::ReplaceInput(), allow no-op.

Allow HInstruction::ReplaceInput() to be called with
a `replacement` being the same as the old input and
do nothing in that case.

This is a follow-up to
    https://android-review.googlesource.com/216923
where I erroneously assumed that it never happens.

Also adhere to the standard C++ std::forward_list<>
semantics in the single-element overload of
`IntrusiveForwardList<>::splice_after()`.

Bug: 28173563

(cherry picked from commit c6b5627c25ff5653e97ccff8c5ccf6ac967b6f83)

Change-Id: I66b7d4b48f629284d0bcb1d31100519e02a872e5

8 years agoMerge changes I0360a078,I3e267519 into nyc-dev
Andreas Gampe [Wed, 20 Apr 2016 17:09:24 +0000 (17:09 +0000)]
Merge changes I0360a078,I3e267519 into nyc-dev

* changes:
  ART: Add Math cutouts
  ART: Add some reflection cutouts

8 years agoART: Add logging for CHECK failure
Andreas Gampe [Wed, 20 Apr 2016 16:55:25 +0000 (09:55 -0700)]
ART: Add logging for CHECK failure

Add more logging around which thread raced this one.

Bug: 28254258
Change-Id: I136a8074295527a4908825f424fc7fe4b13fbae4

8 years agoMerge "Avoid constructing types with errors." into nyc-dev
Aart Bik [Wed, 20 Apr 2016 16:29:45 +0000 (16:29 +0000)]
Merge "Avoid constructing types with errors." into nyc-dev

8 years agoMerge changes Iddd42d75,Ifb2d7b35 into nyc-dev
Vladimir Marko [Wed, 20 Apr 2016 16:17:49 +0000 (16:17 +0000)]
Merge changes Iddd42d75,Ifb2d7b35 into nyc-dev

* changes:
  Reuse HUseListNode<>s when replacing instruction or input.
  Use iterators "before" the use node in HUserRecord<>.

8 years agoReuse HUseListNode<>s when replacing instruction or input.
Vladimir Marko [Tue, 19 Apr 2016 13:36:35 +0000 (14:36 +0100)]
Reuse HUseListNode<>s when replacing instruction or input.

Compiling the Nexus 5 boot image with the 64-bit dex2oat
on host this CL reduces the memory used for compiling the
most hungry method, BatteryStats.dumpLocked(), by ~5.6MiB:

Before:
  MEM: used: 44393040, allocated: 45361248, lost: 968208
  Number of arenas allocated: 319,
  Number of allocations: 815492, avg size: 54
  ...
  UseListNode    10308480
  ...
After:
  MEM: used: 38554536, allocated: 39463008, lost: 908472
  Number of arenas allocated: 274,
  Number of allocations: 572221, avg size: 67
  ...
  UseListNode     4469976
  ...

With 32-bit dex2oat, the UseListNode would be 2/3 of the
values for 64-bit dex2oat (both before and after).

Bug: 28173563
Bug: 27856014

(cherry picked from commit 3c19d3e029a9fcc123d2c6fd1e5e13867d2cfe1f)

Change-Id: Iddd42d7545e4f97a13da63590b33711a5bbdd43b

8 years agoUse iterators "before" the use node in HUserRecord<>.
Vladimir Marko [Tue, 29 Mar 2016 11:21:58 +0000 (12:21 +0100)]
Use iterators "before" the use node in HUserRecord<>.

Create a new template class IntrusiveForwardList<> that
mimicks std::forward_list<> except that all allocations
are handled externally. This is essentially the same as
boost::intrusive::slist<> but since we're not using Boost
we have to reinvent the wheel.

Use the new container to replace the HUseList and use the
iterators to "before" use nodes in HUserRecord<> to avoid
the extra pointer to the previous node which was used
exclusively for removing nodes from the list. This reduces
the size of the HUseListNode by 25%, 32B to 24B in 64-bit
compiler, 16B to 12B in 32-bit compiler. This translates
directly to overall memory savings for the 64-bit compiler
but due to rounding up of the arena allocations to 8B, we
do not get any improvement in the 32-bit compiler.

Compiling the Nexus 5 boot image with the 64-bit dex2oat
on host this CL reduces the memory used for compiling the
most hungry method, BatteryStats.dumpLocked(), by ~3.3MiB:

Before:
  MEM: used: 47829200, allocated: 48769120, lost: 939920
  Number of arenas allocated: 345,
  Number of allocations: 815492, avg size: 58
  ...
  UseListNode    13744640
  ...
After:
  MEM: used: 44393040, allocated: 45361248, lost: 968208
  Number of arenas allocated: 319,
  Number of allocations: 815492, avg size: 54
  ...
  UseListNode    10308480
  ...

Note that while we do not ship the 64-bit dex2oat to the
device, the JIT compilation for 64-bit processes is using
the 64-bit libart-compiler.

Bug: 28173563
Bug: 27856014

(cherry picked from commit 46817b876ab00d6b78905b80ed12b4344c522b6c)

Change-Id: Ifb2d7b357064b003244e92c0d601d81a05e56a7b

8 years agoMerge "ProfileSaver: query profiling_infos instead of the code_map" into nyc-dev
TreeHugger Robot [Wed, 20 Apr 2016 14:40:26 +0000 (14:40 +0000)]
Merge "ProfileSaver: query profiling_infos instead of the code_map" into nyc-dev

8 years agoThumb2: Fix EmitJumpTables() to extend buffer only if needed.
Vladimir Marko [Wed, 20 Apr 2016 13:39:47 +0000 (14:39 +0100)]
Thumb2: Fix EmitJumpTables() to extend buffer only if needed.

Bug: 28256882

(cherry picked from commit 9152fed693f5d823ef29c373d658adc67fa92fe7)

Change-Id: I804e45c8c6a7bc97fcec5a380ccc1bdf70de8b00

8 years agoProfileSaver: query profiling_infos instead of the code_map
Calin Juravle [Tue, 19 Apr 2016 15:33:46 +0000 (16:33 +0100)]
ProfileSaver: query profiling_infos instead of the code_map

This is the first step in being able to record profiles without jit
being active.

Bug: 27916886
Change-Id: I2fcbd03560e109e6866ad6d82c79b902a889e620

8 years agoMerge changes I5044e94b,I149273b7 into nyc-dev
Vladimir Marko [Wed, 20 Apr 2016 09:07:45 +0000 (09:07 +0000)]
Merge changes I5044e94b,I149273b7 into nyc-dev

* changes:
  Fix 570-checker-osr test for non-debuggable mode.
  Fix inlining loops in OSR mode.

8 years agoMerge "ART: Fast-fail empty other handlers in fault handling" into nyc-dev
Andreas Gampe [Wed, 20 Apr 2016 03:53:51 +0000 (03:53 +0000)]
Merge "ART: Fast-fail empty other handlers in fault handling" into nyc-dev

8 years agoAvoid constructing types with errors.
Aart Bik [Wed, 6 Apr 2016 18:22:12 +0000 (11:22 -0700)]
Avoid constructing types with errors.

BUG=27626735

Rationale:
Do not construct classes with a link error. Without this,
the error type thought it was Object (mirror's method
IsObjectClass() returns true if there is no superclass).

(cherry picked from commit 8b3f9b246d5bdbf67faeb2b872b75b8d72777bc0)
(also contains follow-up commit 31244b4cde9156632a08103a8bf1cbff4cbae3cc)

Change-Id: I4443779dda47c320115975c1c71b22e118bd8252

8 years agoDo less work holding thread list lock
Mathieu Chartier [Tue, 19 Apr 2016 20:46:03 +0000 (13:46 -0700)]
Do less work holding thread list lock

Avoid doing stuff that requires access java heap like PrettyMethod.
Fixes lock violation.

Bug: 28268478
Change-Id: Ie090879690df7a6db15a9c8b7e82f3809450d653

8 years agoFix 570-checker-osr test for non-debuggable mode.
Vladimir Marko [Tue, 19 Apr 2016 16:26:54 +0000 (17:26 +0100)]
Fix 570-checker-osr test for non-debuggable mode.

For non-debuggable mode, we need to create profiling info.

Bug: 28210356

(cherry picked from commit b9d338b96c804a077dc2ed60b36b7ab64dc2c78c)

Change-Id: I5044e94bf78bd982007c8e7a3a9e5ee56f861e9c

8 years agoFix comparison in profile saver
Calin Juravle [Tue, 19 Apr 2016 17:17:41 +0000 (18:17 +0100)]
Fix comparison in profile saver

Bug: 27893309
Change-Id: I076a20413cffc69ca9e562fb4fc75615fead7690

8 years agoMerge "Fix tracking foreign dex files" into nyc-dev
Calin Juravle [Tue, 19 Apr 2016 17:03:28 +0000 (17:03 +0000)]
Merge "Fix tracking foreign dex files" into nyc-dev

8 years agoMerge "Skip profiling if the dex file is fully compiled" into nyc-dev
TreeHugger Robot [Tue, 19 Apr 2016 16:56:18 +0000 (16:56 +0000)]
Merge "Skip profiling if the dex file is fully compiled" into nyc-dev

8 years agoMerge "ART: Remove Security initialization cutout" into nyc-dev
Andreas Gampe [Tue, 19 Apr 2016 16:49:03 +0000 (16:49 +0000)]
Merge "ART: Remove Security initialization cutout" into nyc-dev

8 years agoMerge changes Ia55ae13e,Ib3e2e6d2 into nyc-dev
Richard Uhler [Tue, 19 Apr 2016 16:24:12 +0000 (16:24 +0000)]
Merge changes Ia55ae13e,Ib3e2e6d2 into nyc-dev

* changes:
  Disable stripped-dex test for no-dex2oat.
  Disable stripped-dex test on target.

8 years agoMerge "Use dex cache from compilation unit in HInstructionBuilder." into nyc-dev
TreeHugger Robot [Tue, 19 Apr 2016 15:50:39 +0000 (15:50 +0000)]
Merge "Use dex cache from compilation unit in HInstructionBuilder." into nyc-dev

8 years agoART: Fast-fail empty other handlers in fault handling
Andreas Gampe [Tue, 19 Apr 2016 00:15:42 +0000 (17:15 -0700)]
ART: Fast-fail empty other handlers in fault handling

Don't try to set up the nested handler environment if there are
no other handlers installed.

Stop-gap for the current setup without other handlers. There are
a myriad more issues before we can safely re-enable other handlers.

Bug: 28245957

(cherry picked from commit d14b73d9d9027539c1dfeff93f784a58c93f4e32)

Change-Id: I526aef82aa1ca2b6c4b1959858fe76231bc067d9

8 years agoART: Fix IsLinearOrderWellFormed in liveness analysis
David Brazdil [Tue, 19 Apr 2016 09:10:17 +0000 (10:10 +0100)]
ART: Fix IsLinearOrderWellFormed in liveness analysis

Graph may contain nullptr in list of basic blocks when
the basic block is removed. So we must handle nullptr explicitly.

Bug: 28252537
Bug: 27615840
Bug: 27624868

Change-Id: If430939b8fe793e4a9c974936180d06c1a9f90c6
Signed-off-by: Serguei Katkov <serguei.i.katkov@intel.com>
(cherry picked from commit fa3091ef031da6c18478fdede12d7762634ee271)

8 years agoFix tracking foreign dex files
Calin Juravle [Mon, 18 Apr 2016 17:59:22 +0000 (18:59 +0100)]
Fix tracking foreign dex files

Apps which share the same VM will have different application data
directories. We used to store only the first one registered and miss
subsequent registrations.

Bug: 28012567
Change-Id: I27c2aa2efa938b072be7c73bb778db82f80fcf1e

8 years agoSkip profiling if the dex file is fully compiled
Calin Juravle [Mon, 18 Apr 2016 15:38:27 +0000 (16:38 +0100)]
Skip profiling if the dex file is fully compiled

The OatFile does not store the original dex location that was used to
compile the file. So in order to get the oat file for a given code path
we need to iterate over the inner oat dex files and check against them.

Also, we used to pass the unfiltered code_paths set to the saver which
would make the filtering useless. This fixes the parameter passing as
well.

Bug: 27893309
Change-Id: Ia95d61a82a9e74769ebf3329bb1f9c714bcff27f

8 years agoUse dex cache from compilation unit in HInstructionBuilder.
Vladimir Marko [Wed, 13 Apr 2016 18:29:26 +0000 (19:29 +0100)]
Use dex cache from compilation unit in HInstructionBuilder.

Avoid calling costly ClassLinker::FindDexCache() from
HInstructionBuilder, the dex cache is already available in
the compilation unit. Compiling Nexus 5 boot image on host
under perf(1) shows that the time spent in FindDexCache()
is reduced from about 2% to well under 0.2%, 90% of the
remaining hits coming from ReferenceTypePropagation which
doesn't have access to the compilation unit.

Bug: 28173563

(cherry picked from commit 3cd50df11b3076b801954018236c366fd9b97948)

Change-Id: Ife0dba2dd8f49bd52f86ddadf06ee787bad03d66