OSDN Git Service
Nicolas Geoffray [Thu, 7 Jan 2016 09:13:24 +0000 (09:13 +0000)]
Revert "Reduce sleep duration in SuspendAllDaemonThreads"
run-test 132 randomly chokes due to this change.
Bug:
26351700
This reverts commit
adef5bde6c3392dca3dafa0f8d2c9428b59af3ef.
Change-Id: I18e007264e1e8b9027d7d241cb9f0053ad8d112a
Mathieu Chartier [Thu, 7 Jan 2016 02:28:21 +0000 (18:28 -0800)]
Reduce sleep duration in SuspendAllDaemonThreads
Previously we unconditionally slept for 200ms which caused runtime
shutdown to take 200ms longer than required.
Reduces runtime shutdown time for "am" from ~200ms to ~10ms.
Bug:
26351700
Change-Id: I7fd0059cb3b04dba0acc4a1754e76c89ec867f85
Mathieu Chartier [Thu, 7 Jan 2016 02:21:52 +0000 (02:21 +0000)]
Merge "Add systrace output to runtime shutdown"
Mathieu Chartier [Thu, 7 Jan 2016 01:23:25 +0000 (17:23 -0800)]
Add systrace output to runtime shutdown
Help debug why am command shutdown is slow.
Findings:
~200ms spent in deleting thread list on N5.
Bug:
26351700
Change-Id: Icb59390a5267ae8bade4ced12252743c48206bb1
Mathieu Chartier [Thu, 7 Jan 2016 00:23:17 +0000 (00:23 +0000)]
Merge "Address some intern table comments"
Mathieu Chartier [Wed, 6 Jan 2016 23:47:09 +0000 (15:47 -0800)]
Address some intern table comments
Change-Id: I7ffaa463272015c2924ba03e006041daee498ad4
Mathieu Chartier [Wed, 6 Jan 2016 23:26:01 +0000 (23:26 +0000)]
Merge "Remove gaps between images"
Mathieu Chartier [Wed, 6 Jan 2016 21:51:52 +0000 (13:51 -0800)]
Remove gaps between images
We now use the same logic to calculate the size as we use to create
the image header. This should mean that it always lines up correctly.
Previously the alignment rounding could have caused page off by one
errors in some cases.
Re-enable no gap check in heap.cc.
Bug:
26317072
Change-Id: I558a4e60197ca4613fc4ec4c3a618c64a83d006a
Mathieu Chartier [Wed, 6 Jan 2016 19:42:41 +0000 (19:42 +0000)]
Merge "Re-enable adding intern table to image"
Mathieu Chartier [Wed, 6 Jan 2016 19:19:22 +0000 (19:19 +0000)]
Merge "Get rid of shadow maps for images"
Mathieu Chartier [Wed, 6 Jan 2016 02:03:15 +0000 (18:03 -0800)]
Get rid of shadow maps for images
Fixed the image bitmap creation to use the End of the object section
instead of using the map size. This means that the bitmap covers at
most one word of bits past the end of the object section.
Reverts shadow map logic introduced by multi image CL:
(commit
dcdc85bbd569f0ee66c331b4219c19304a616214)
Bug:
26317072
Change-Id: I181cf3490a5b292c00c6b0f97536f3206adcd091
Mathieu Chartier [Tue, 29 Dec 2015 21:17:37 +0000 (13:17 -0800)]
Re-enable adding intern table to image
Changed intern table to have a stack of tables similarily to
ClassTable. Adding an image intern table adds to the front of the
intern table stack. Also some cleanup.
Bug:
26317072
Change-Id: I7bbf9485b5dbbbf3707fed21e29de3beccfb8705
Nicolas Geoffray [Wed, 6 Jan 2016 16:34:44 +0000 (16:34 +0000)]
Merge "Disable tests while investigating."
Nicolas Geoffray [Wed, 6 Jan 2016 16:33:41 +0000 (16:33 +0000)]
Disable tests while investigating.
Change-Id: I3f1e46a8bf27c962c96315d125ddd8430b7f9f0b
David Brazdil [Wed, 6 Jan 2016 12:47:07 +0000 (12:47 +0000)]
Merge "ART: Don't set initial RTI for BoundType if input untyped"
David Brazdil [Wed, 6 Jan 2016 12:45:54 +0000 (12:45 +0000)]
Merge "ART: Resolve ambiguous ArraySets"
David Brazdil [Mon, 28 Dec 2015 10:53:34 +0000 (10:53 +0000)]
ART: Don't set initial RTI for BoundType if input untyped
ReferenceTypePropagation will create a BoundType with upper bound
[Object, inexact, not null] for each if-not-null branch. The logic
setting its initial RTI will, however, set it straight to Object if
the input is untyped (loop phi or its derivate). This patch changes
the logic to leave the BoundType untyped and set it during fix-point
iteration.
Bug:
26330326
Change-Id: Ic492e2179a4c51f577908e60fbcf70d728b98a6f
David Brazdil [Wed, 16 Dec 2015 10:30:45 +0000 (10:30 +0000)]
ART: Resolve ambiguous ArraySets
Just like aget(-wide), the value operand of aput(-wide) bytecode
instructions can be both int/long and float/double. This patch builds
on the previous mechanism for resolving type of ArrayGets to type the
values of ArraySets based on the reference type of the array.
Bug:
22538329
Change-Id: Ic86abbb58de146692de04476b555010b6fcdd8b6
David Brazdil [Wed, 6 Jan 2016 10:34:16 +0000 (10:34 +0000)]
Merge "ART: Create BoundType for CheckCast early"
David Brazdil [Sun, 27 Dec 2015 13:36:12 +0000 (13:36 +0000)]
ART: Create BoundType for CheckCast early
ReferenceTypePropagation creates a BoundType for each CheckCast and
replaces all dominated uses of the casted object with it. This does
not include Phi uses on the boundary of the dominated scope, reducing
typing precision. This patch creates the BoundType in Builder, causing
SsaBuilder to replace uses of the object automatically.
Bug:
26081304
Change-Id: I083979155cccb348071ff58cb9060a896ed7d2ac
Nicolas Geoffray [Wed, 6 Jan 2016 09:52:15 +0000 (09:52 +0000)]
Merge "Update expectations after push of apache-xml and other fixes."
Nicolas Geoffray [Wed, 6 Jan 2016 09:33:56 +0000 (09:33 +0000)]
Update expectations after push of apache-xml and other fixes.
bug:
26326992
Change-Id: I7d0c0706c53b4b04a3a3038b334cac6bcda4868c
Nicolas Geoffray [Wed, 6 Jan 2016 08:43:03 +0000 (08:43 +0000)]
Merge "Revert "Add new test expectations after enso move.""
Nicolas Geoffray [Wed, 6 Jan 2016 08:34:46 +0000 (08:34 +0000)]
Revert "Add new test expectations after enso move."
Failures were due to a bogus date on devices.
Bug:
26353151
Bug:
26326992
This reverts commit
376756d0c9e23173c1d93fe32acf6b3e08f45b73.
Change-Id: I3ecdd152b27299efab37ded29b62c2f3369b5d96
Jeff Hao [Wed, 6 Jan 2016 02:09:22 +0000 (02:09 +0000)]
Merge "Fix multi-image TODOs in class linker, runtime, and oat file."
Andreas Gampe [Wed, 6 Jan 2016 00:25:40 +0000 (00:25 +0000)]
Merge "ART: Add large object allocation to ThreadStress"
Jeff Hao [Tue, 5 Jan 2016 22:13:00 +0000 (14:13 -0800)]
Fix multi-image TODOs in class linker, runtime, and oat file.
- Modified SanityCheckArtMethodPointerArray in class linker
- Put back warnings in OatFile::GetOatDexFile
- Reinstated ImageSpace VerifyImageAllocations in Runtime::Init
Bug:
26317072
(cherry-picked from commit
0dfef949bb824accde27f8cfe1b233ec9e087355)
Change-Id: I3bdb8f87d885213795c82c41e5095fec6daf00c4
Andreas Gampe [Tue, 5 Jan 2016 20:57:56 +0000 (12:57 -0800)]
ART: Add large object allocation to ThreadStress
Test interactions with large objects some more, as there are
sometimes interactions between maps and unwind (in SIGQUIT).
(Probably should ramp up the iterations/frequency, though.)
Change-Id: I21368bfc584b9b642a39d3560391f929b5576b08
Josh Gao [Tue, 5 Jan 2016 19:38:33 +0000 (19:38 +0000)]
Merge "Remove unused reference to MALLOC_IMPL."
Andreas Gampe [Tue, 5 Jan 2016 18:28:58 +0000 (18:28 +0000)]
Merge "ART: Make opt inliner a little bit cleaner/faster"
Andreas Gampe [Thu, 24 Dec 2015 00:30:44 +0000 (16:30 -0800)]
ART: Make opt inliner a little bit cleaner/faster
Remove the double-if.
Change-Id: Ic6551e30d016d4b2c3b8dfb8841d2db90fce45cd
Richard Uhler [Tue, 5 Jan 2016 16:56:24 +0000 (16:56 +0000)]
Merge "Don't duplicate the menu in the help page for ahat."
Andreas Gampe [Tue, 5 Jan 2016 16:35:42 +0000 (16:35 +0000)]
Merge "ART: Use non-stripped core jars in gtests"
Andreas Gampe [Wed, 30 Sep 2015 18:52:02 +0000 (11:52 -0700)]
ART: Use non-stripped core jars in gtests
To support emulators and other preopted device configurations, use
a special core-libart and core-oj version that will never be stripped.
Bug:
24535627
Change-Id: Ie632bc382bbc753e9c100039f2ecc4ac1f2ab249
Andreas Gampe [Tue, 5 Jan 2016 16:26:13 +0000 (16:26 +0000)]
Merge "ART: Move ThrowStackOverflowError to common_throws"
Andreas Gampe [Mon, 4 Jan 2016 23:32:43 +0000 (15:32 -0800)]
ART: Move ThrowStackOverflowError to common_throws
This should really be in common_throws considering all the callers.
Change-Id: I410c07cbf5659f216d44597530cb85f99740f320
Narayan Kamath [Tue, 5 Jan 2016 16:01:03 +0000 (16:01 +0000)]
Merge "Add apache-xml to the device & host boot classpath."
Narayan Kamath [Tue, 5 Jan 2016 14:33:58 +0000 (14:33 +0000)]
Add apache-xml to the device & host boot classpath.
This is used by parts of OpenJdk's java.util.prefs implementation.
bug:
26326992
Change-Id: Ibc0f3043945ec6eefdc97e9bab998e2082baf95e
Nicolas Geoffray [Tue, 5 Jan 2016 12:55:28 +0000 (12:55 +0000)]
Merge "Disable test due to toybox update."
Nicolas Geoffray [Tue, 5 Jan 2016 12:46:51 +0000 (12:46 +0000)]
Disable test due to toybox update.
bug:
26395656
Change-Id: I0356bf0e5fff3cab198f6c91da097ff0de564c89
Jeff Hao [Tue, 5 Jan 2016 03:13:30 +0000 (03:13 +0000)]
Merge "Fix multi-image TODOs in patchoat."
Jeff Hao [Tue, 5 Jan 2016 01:38:06 +0000 (17:38 -0800)]
Fix multi-image TODOs in patchoat.
- Reinstated checks that were still valid (one check was removed).
- Removed unused path that patches only images.
Bug:
26317072
(cherry-picked from commit
e271fe1e2797205c57c052212c32139234f781ea)
Change-Id: I0e140cb110abbfa469c097c805657ecfdc8552d6
Josh Gao [Tue, 5 Jan 2016 01:17:46 +0000 (17:17 -0800)]
Remove unused reference to MALLOC_IMPL.
Bug: http://b/
26390299
Change-Id: I2e779282d8d585519c2ed1bff246d2202eee61c8
Andreas Gampe [Mon, 4 Jan 2016 19:19:30 +0000 (19:19 +0000)]
Merge changes I129fbc84,I9238ee0b
* changes:
ART: Fix JDWP GetClassLoader command
Revert "ART: Fix JDWP GetClassLoader command"
Dimitry Ivanov [Mon, 4 Jan 2016 17:15:44 +0000 (17:15 +0000)]
Merge "Revert "Temporary disable linker-namespaces""
Nicolas Geoffray [Mon, 4 Jan 2016 16:31:47 +0000 (16:31 +0000)]
Merge "Add a fence to ensure other threads see the correct data."
Nicolas Geoffray [Mon, 4 Jan 2016 16:06:51 +0000 (16:06 +0000)]
Add a fence to ensure other threads see the correct data.
Other threads may see the ProfilingInfo pointer in the ArtMethod,
but the data they have for it is bogus due to a prior load in their
cache.
bug:
26221227
Change-Id: I1fa8f088a3144df39d07a986ed57d6fc20edba18
Andreas Gampe [Mon, 4 Jan 2016 16:16:51 +0000 (16:16 +0000)]
Merge "ART: Remove core-junit from test boot classpath"
David Brazdil [Mon, 4 Jan 2016 16:16:05 +0000 (16:16 +0000)]
Merge "ART: Use Primitive::Is64BitType in SsaBuilder::TypePhiFromInputs"
David Brazdil [Mon, 4 Jan 2016 15:55:10 +0000 (15:55 +0000)]
ART: Use Primitive::Is64BitType in SsaBuilder::TypePhiFromInputs
Accidentally submitted CL Ib39f3da2b92bc5be5d76f4240a77567d82c6bebe
without this.
Bug:
26208284
Change-Id: I2e69ca2fde72171f2ce645304c4835d8d57f4e0b
Nicolas Geoffray [Mon, 4 Jan 2016 15:41:54 +0000 (15:41 +0000)]
Merge "Add 002-sleep as timing sensitive."
Nicolas Geoffray [Mon, 4 Jan 2016 15:22:00 +0000 (15:22 +0000)]
Add 002-sleep as timing sensitive.
The buildbots have chewed on it a few times.
Change-Id: Icf81cf4baa8e06573ed92bf61c0861c64825475a
David Brazdil [Mon, 4 Jan 2016 14:51:33 +0000 (14:51 +0000)]
Merge "X86: templatize GenerateTestAndBranch and friends"
Nicolas Geoffray [Mon, 4 Jan 2016 14:42:19 +0000 (14:42 +0000)]
Merge "Detect phi cycles."
Nicolas Geoffray [Mon, 4 Jan 2016 14:41:49 +0000 (14:41 +0000)]
Merge "Revert "Revert "Tweak inlining heuristics."""
Nicolas Geoffray [Fri, 18 Dec 2015 10:57:10 +0000 (10:57 +0000)]
Revert "Revert "Tweak inlining heuristics.""
This reverts commit
b17d1ccff0ac26fc22df671907ba2b4f4c656ce4.
Change-Id: I26f6f8702a448c3da12662cbc6bc0f6e562bc40b
Nicolas Geoffray [Mon, 4 Jan 2016 12:21:15 +0000 (12:21 +0000)]
Merge "MIPS32: improvements in code generation (mostly 64-bit ALU ops)"
Nicolas Geoffray [Mon, 4 Jan 2016 08:24:30 +0000 (08:24 +0000)]
Merge "Remove bogus DCHECK in induction analysis."
Dimitry Ivanov [Thu, 17 Dec 2015 19:18:27 +0000 (11:18 -0800)]
Revert "Temporary disable linker-namespaces"
This reverts commit
9aefb12cc91eb459d87392700f1199ee07580cd6.
Change-Id: Iac567c03699d65f97c380c2f32ef5a2a56d44721
Mark Mendell [Thu, 31 Dec 2015 17:28:50 +0000 (12:28 -0500)]
X86: templatize GenerateTestAndBranch and friends
Allow the use of NearLabel as well as Label. This will be used by the
HSelect patch.
Replace a couple of Label(s) with NearLabel(s) as well.
Change-Id: I8e674c89e691bcdbccf4a5cdc07ad13b29ec21dd
Signed-off-by: Mark Mendell <mark.p.mendell@intel.com>
Nicolas Geoffray [Tue, 29 Dec 2015 16:12:27 +0000 (16:12 +0000)]
Remove bogus DCHECK in induction analysis.
Method is called during BCE, which may change the graph in non-SSA
form temporarily.
Change-Id: I84bed7d3370c6871dc2d3b883d0aca90b3a37696
David Brazdil [Thu, 31 Dec 2015 11:06:56 +0000 (11:06 +0000)]
Merge "ART: Set RTI of HArm64IntermediateAddress"
David Brazdil [Thu, 31 Dec 2015 11:06:00 +0000 (11:06 +0000)]
ART: Set RTI of HArm64IntermediateAddress
Change-Id: I2145bc249cc940d7b133fd6cbbd133cc62fee187
David Brazdil [Thu, 31 Dec 2015 10:29:06 +0000 (10:29 +0000)]
Merge "ART: Refactor SsaBuilder for more precise typing info"
David Brazdil [Wed, 16 Dec 2015 10:37:39 +0000 (10:37 +0000)]
ART: Refactor SsaBuilder for more precise typing info
This reverts commit
68289a531484d26214e09f1eadd9833531a3bc3c.
Now uses Primitive::Is64BitType instead of Primitive::ComponentSize
because it was incorrectly optimized by GCC.
Bug:
26208284
Bug:
24252151
Bug:
24252100
Bug:
22538329
Bug:
25786318
Change-Id: Ib39f3da2b92bc5be5d76f4240a77567d82c6bebe
Andreas Gampe [Thu, 31 Dec 2015 03:33:49 +0000 (19:33 -0800)]
ART: Fix JDWP GetClassLoader command
The command is spec-ed to take a type, and return the type's
classloader. Decode the input as a class, just like all the
other ReferenceType commands already do.
Bug:
26349019
Change-Id: I129fbc844f529fa9234742ffe128d7ac41ce6d7e
Andreas Gampe [Thu, 31 Dec 2015 03:31:57 +0000 (03:31 +0000)]
Revert "ART: Fix JDWP GetClassLoader command"
This reverts commit
40144cc022d321e2ec28d7fac2f5a24ba6714aff.
The implementation should not DCHECK.
Bug:
26349019
Change-Id: I9238ee0b05f8a150f55773e78301ded9e8b38334
Aart Bik [Wed, 30 Dec 2015 23:47:56 +0000 (23:47 +0000)]
Merge "Temporary put -DAG back to fix broken test."
Aart Bik [Wed, 30 Dec 2015 23:46:27 +0000 (15:46 -0800)]
Temporary put -DAG back to fix broken test.
Change-Id: I1fb159e8e207f515afd7020e188b0af55dbc2bc9
Aart Bik [Wed, 30 Dec 2015 22:50:09 +0000 (22:50 +0000)]
Merge "Improved side-effects/can-throw information on intrinsics."
Andreas Gampe [Wed, 30 Dec 2015 20:46:18 +0000 (20:46 +0000)]
Merge "Revert "Revert "ART: Fix up some multi-image cases"""
Andreas Gampe [Wed, 30 Dec 2015 19:03:17 +0000 (19:03 +0000)]
Revert "Revert "ART: Fix up some multi-image cases""
This reverts commit
de38b797c3e5ba3ee44c480db7093386975c51eb.
Fix up imgdiag for std::string and multi-image.
Bug:
26317072
Bug:
26320300
Change-Id: I94ce9528e9fea6fb3231a70c32db02d567143db9
Aart Bik [Mon, 14 Dec 2015 19:57:01 +0000 (11:57 -0800)]
Improved side-effects/can-throw information on intrinsics.
Rationale: improved side effect and exception analysis gives
many more opportunities for GVN/LICM/BCE.
Change-Id: I8aa9b757d77c7bd9d58271204a657c2c525195b5
Nicolas Geoffray [Wed, 30 Dec 2015 14:50:40 +0000 (14:50 +0000)]
Merge "Revert "ART: Fix up some multi-image cases""
Nicolas Geoffray [Wed, 30 Dec 2015 14:50:12 +0000 (14:50 +0000)]
Revert "ART: Fix up some multi-image cases"
Fails imgdiag_test on device.
Bug:
26317072
Bug:
26320300
This reverts commit
288b1e9a0dddfb91e85067fe81de55174f313c7c.
Change-Id: Iccd05827b0630281b6f959331eaa4202526df78e
Nicolas Geoffray [Wed, 30 Dec 2015 14:40:35 +0000 (14:40 +0000)]
Merge "Add new test expectations after enso move."
Nicolas Geoffray [Wed, 30 Dec 2015 14:39:15 +0000 (14:39 +0000)]
Add new test expectations after enso move.
bug:
26353151
bug:
26326992
Change-Id: I7e172fceb265938435cae8eb4c0ef79ba6209348
Nicolas Geoffray [Wed, 30 Dec 2015 14:26:57 +0000 (14:26 +0000)]
Merge "On x64, cmpl can never take a int64 immediate."
Nicolas Geoffray [Wed, 30 Dec 2015 14:10:13 +0000 (14:10 +0000)]
On x64, cmpl can never take a int64 immediate.
Fix a wrong type widening in x64 code generator and add
CHECKs in the assembler.
Change-Id: Id35f5d47c6cf78ed07e73ab783db09712d3c437f
Nicolas Geoffray [Wed, 30 Dec 2015 14:20:48 +0000 (14:20 +0000)]
Merge "Fix code generation for String.<init> on x64."
Narayan Kamath [Wed, 30 Dec 2015 13:15:35 +0000 (13:15 +0000)]
Merge "Fix bogus logging statement."
Narayan Kamath [Thu, 3 Dec 2015 14:11:46 +0000 (14:11 +0000)]
Fix bogus logging statement.
We were always logging an empty string.
Change-Id: I9ebc89ce2df8cb664aaf2640a16dcd85bdfa036e
Andreas Gampe [Wed, 30 Dec 2015 00:54:55 +0000 (00:54 +0000)]
Merge "ART: Fix JDWP GetClassLoader command"
Andreas Gampe [Wed, 30 Dec 2015 00:23:20 +0000 (16:23 -0800)]
ART: Fix JDWP GetClassLoader command
The command is spec-ed to take a type, and return the type's
classloader. The input is thus already a Class, not a generic Object.
Fix the implementation to not always return null.
Bug:
26349019
Change-Id: I4aa075902f4a112624edc644d3540a59eea057c8
Andreas Gampe [Tue, 29 Dec 2015 18:31:55 +0000 (18:31 +0000)]
Merge "ART: Fix up some multi-image cases"
Nicolas Geoffray [Tue, 29 Dec 2015 16:20:52 +0000 (16:20 +0000)]
Fix code generation for String.<init> on x64.
The ArtMethod is a 64bit pointer so should be loaded with movq.
Change-Id: I80803046a9144776d7f069e8baee61e39ae289d5
Andreas Gampe [Mon, 28 Dec 2015 18:41:49 +0000 (10:41 -0800)]
ART: Fix up some multi-image cases
Change the auto-generated multi-image names to include the path
components from the first image, as well as prefix them with the
first image's name to disambiguate. This fixes vogar-style usage.
Fix an out-of-bounds issue in dex2oat when dex files are missing.
Forbid generating or patching multi-image parts when loading images.
Instead just fail loading them.
Remember ImageSpace instances that have been added while trying to
load a multi-image set. Remove all loaded instances when the overall
loading process fails.
Refactor the dex location adaptation into ImageSpace. Reuse the code
in the Runtime path for fallback, so that all dex files can be found
correctly.
Fix an out-of-bounds access in OatFileAssistant in fallback mode.
Partially reverts
d895961d07a1d320b29f2045a48bc5a1944a4d3c. Push an
actual image name, that is, something with an art extension, to
the vogar scripts.
Partially reverts
c525604b313bb77a2077e1fec43dfab76cb1b9b1. Test
119-noimage-patchoat works again.
Bug:
26317072
Bug:
26320300
Change-Id: I3f05fa77f22a2b9ca54c3105ffc53646c1928604
Richard Uhler [Mon, 21 Dec 2015 22:34:59 +0000 (14:34 -0800)]
Don't duplicate the menu in the help page for ahat.
Future changes to the menu will need to be made in only a single
location.
Change-Id: I24ffcf17d5564b98bd3552d4ba9284eb6924d52b
Andreas Gampe [Mon, 28 Dec 2015 22:31:06 +0000 (14:31 -0800)]
ART: Remove core-junit from test boot classpath
Remove core-junit, which isn't used. Follow-up to commit
0ad65a3c794b18101def5096d8ed151382c9d5aa.
Change-Id: Ibac1c16e8b72688405dc59289dd22c1d9de69bb3
Andreas Gampe [Mon, 28 Dec 2015 17:59:43 +0000 (17:59 +0000)]
Merge changes I00dc6cfc,Iae8f1c88
* changes:
ART: Disable Math.round intrinsics
Revert "Make Math.round consistent on arm64."
Andreas Gampe [Mon, 28 Dec 2015 17:54:29 +0000 (09:54 -0800)]
ART: Disable Math.round intrinsics
The move to OpenJDK means that Android has caught up with the
definition change of Math.round. Disable intrinsics.
Bug:
26327751
Change-Id: I00dc6cfca12bd7c95e56a4ab76ffee707d3822dc
Andreas Gampe [Mon, 28 Dec 2015 17:06:29 +0000 (09:06 -0800)]
Revert "Make Math.round consistent on arm64."
This reverts commit
40041c9a38e3961d8675d117517719458a115520.
Needs to be generalized to all platforms.
Bug:
26327751
Change-Id: Iae8f1c8846d120d8e3e99b6eb87f3760bf793ec5
Nicolas Geoffray [Sun, 27 Dec 2015 23:02:37 +0000 (23:02 +0000)]
Merge "Remove duplicate entry in expectations."
Nicolas Geoffray [Sun, 27 Dec 2015 23:01:28 +0000 (23:01 +0000)]
Remove duplicate entry in expectations.
Change-Id: Ia265787f132ae8a5685fa0a699fa133f6f7f0554
Nicolas Geoffray [Sun, 27 Dec 2015 16:26:40 +0000 (16:26 +0000)]
Merge "Make Math.round consistent on arm64."
Nicolas Geoffray [Sun, 27 Dec 2015 16:20:14 +0000 (16:20 +0000)]
Make Math.round consistent on arm64.
OpenJDK seems to have a different rounding implementation than
libcore. Temporarily disable the intrinsic.
Test that fails:
Assert.assertEquals(StrictMath.round(0.
49999999999999994d), 1l);
Assert.assertEquals(Math.round(0.
49999999999999994d), 1l);
bug:
26327751
Change-Id: Iad2fb847e4a553b8c1f5031f772c81e7e4db9f4c
Nicolas Geoffray [Sun, 27 Dec 2015 15:50:52 +0000 (15:50 +0000)]
Merge "Tweaks to get vogar runs working after enso move."
Nicolas Geoffray [Sun, 27 Dec 2015 15:46:35 +0000 (15:46 +0000)]
Tweaks to get vogar runs working after enso move.
- Use ART_TEST_ANDROID_ROOT for the dex locations, so that
ART can find resources.
- Use the pre-compiled boot image to avoid b/
26320300.
- Add new failures to failures file.
bug:
26320300
Change-Id: I08653b563f837a7ac3a014424951417f6bdd3294
Nicolas Geoffray [Sat, 26 Dec 2015 19:44:09 +0000 (19:44 +0000)]
Merge "Tweaks to get run-test working after enso."
Nicolas Geoffray [Sat, 26 Dec 2015 19:41:37 +0000 (19:41 +0000)]
Tweaks to get run-test working after enso.
- test 119 temporarily disabled
- set java.io.tmpdir in the command line for target
bug:
26320300
Change-Id: I9ddeac4f23f065eea62511f9fa392ea4764d0435