OSDN Git Service
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
Hiroshi Yamauchi [Mon, 21 Jul 2014 18:01:22 +0000 (18:01 +0000)]
Merge "Add 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
Ian Rogers [Mon, 21 Jul 2014 16:37:48 +0000 (16:37 +0000)]
Merge "Fix 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
Nicolas Geoffray [Mon, 21 Jul 2014 12:40:53 +0000 (12:40 +0000)]
Merge "Fix sizeof -> strlen to get the test to run on build servers."
Nicolas Geoffray [Mon, 21 Jul 2014 10:51:32 +0000 (11:51 +0100)]
Fix sizeof -> strlen to get the test to run on build servers.
Change-Id: I8501d02a9bc5f84a769a80d46d0ec5754c972e5d
Nicolas Geoffray [Mon, 21 Jul 2014 10:14:31 +0000 (10:14 +0000)]
Merge "Add write barriers to optimizing compiler."
Nicolas Geoffray [Thu, 17 Jul 2014 14:15:34 +0000 (15:15 +0100)]
Add write barriers to optimizing compiler.
Change-Id: I43a40954757f51d49782e70bc28f7c314d6dbe17
Nicolas Geoffray [Mon, 21 Jul 2014 09:10:54 +0000 (09:10 +0000)]
Merge "Add assembly operations with constants in optimizing compiler."
Nicolas Geoffray [Fri, 11 Jul 2014 09:57:49 +0000 (10:57 +0100)]
Add assembly operations with constants in optimizing compiler.
Change-Id: I5bcc35ab50d4457186effef5592a75d7f4e5b65f
Hiroshi Yamauchi [Fri, 18 Jul 2014 00:16:30 +0000 (00:16 +0000)]
Merge "Fix two read barrier bugs introduced in CL 97882."
Hiroshi Yamauchi [Thu, 17 Jul 2014 23:16:31 +0000 (16:16 -0700)]
Fix two read barrier bugs introduced in CL 97882.
Bug:
12687968
Change-Id: Ib28da4c33977cc58b09913ef5b738dec75365714
Ian Rogers [Sat, 19 Jul 2014 05:41:18 +0000 (05:41 +0000)]
Merge "Fix the name of a black listed run-test."
Ian Rogers [Sat, 19 Jul 2014 05:38:56 +0000 (22:38 -0700)]
Fix the name of a black listed run-test.
Change-Id: I1ac4d28bd959b33066cce32812613c2e27d40f7e
Ian Rogers [Sat, 19 Jul 2014 03:56:27 +0000 (20:56 -0700)]
Black list timing sensitive tests from gcverify and gcstress
but only when run on build servers.
Change-Id: Icb3b125142d75bd4cc140c9f8ef7abf1d6451c5a
Ian Rogers [Sat, 19 Jul 2014 02:52:07 +0000 (19:52 -0700)]
Black list more trace tests.
Bug:
16386215
Change-Id: Ieda8306f9419669935f5452a1765c09ff898ff7f
Ian Rogers [Sat, 19 Jul 2014 03:59:07 +0000 (03:59 +0000)]
Merge "Black list timing sensitive tests from gcverify and gcstress"
Ian Rogers [Fri, 18 Jul 2014 23:09:06 +0000 (23:09 +0000)]
Merge "Run-test fix."
Ian Rogers [Fri, 18 Jul 2014 22:55:59 +0000 (15:55 -0700)]
Run-test fix.
Change-Id: I601d425582a69f77786a76d625079f3ca5974428
Ian Rogers [Fri, 18 Jul 2014 22:13:37 +0000 (22:13 +0000)]
Merge "Add a GC stress run that uses a moving collector and small heap sizes."
Ian Rogers [Fri, 18 Jul 2014 22:06:53 +0000 (15:06 -0700)]
Add a GC stress run that uses a moving collector and small heap sizes.
Disable 3 run-tests for this configuration that fail for reasons relating to
OOME handling.
Add 012-math to the failing trace tests, it fails due to hitting a file size
limit when running in the interpreter.
Change-Id: Ia58f4c245d9d2d14caf5f2f2f3dcbfa40d4ccadb
Andreas Gampe [Fri, 18 Jul 2014 19:17:01 +0000 (19:17 +0000)]
Merge "ART: Improve overflow detection in dex file verifier"
Andreas Gampe [Fri, 18 Jul 2014 04:49:24 +0000 (21:49 -0700)]
ART: Improve overflow detection in dex file verifier
Overflows were hidden by the design of the checks. Push all range
checks as lists, so we can explicitly check against the count.
Bug:
16017886
Change-Id: I0083f83006ef1e55518b0919dff319004b66dcb8
Ian Rogers [Fri, 18 Jul 2014 18:50:37 +0000 (18:50 +0000)]
Merge "Add GC verification test rules."
Ian Rogers [Fri, 18 Jul 2014 18:38:13 +0000 (11:38 -0700)]
Add GC verification test rules.
Change-Id: I7469ecaf98f680e7e642035f3536b00bdd97b486
Ian Rogers [Fri, 18 Jul 2014 18:02:19 +0000 (11:02 -0700)]
Allow optimizing and trace tests to be not run.
Make it so that optimizing and trace tests can be chosen to be not run.
mm -j32 test-art
will run all optimizing and trace tests.
mm -j32 test-art ART_TEST_FULL=false
will not run optimizing and trace tests.
mm -j32 test-art ART_TEST_FULL=false ART_TEST_OPTIMIZING=true
will not run trace tests but will run optimizing tests.
Change-Id: I9c04d9489f08fb837da5561d9496c97aee7c4a4a
Ian Rogers [Fri, 18 Jul 2014 18:17:41 +0000 (18:17 +0000)]
Merge "Allow optimizing and trace tests to be not run."
Ian Rogers [Fri, 18 Jul 2014 15:39:41 +0000 (15:39 +0000)]
Merge "Improved implementation of inline of String.indexOf"
nikolay serdjuk [Fri, 18 Jul 2014 05:35:46 +0000 (12:35 +0700)]
Improved implementation of inline of String.indexOf
This version pushes EDI only once and only in 32-bit mode.
Change-Id: I4e871d3531ac539536f8f53ec09ffb664409c9cc
Ian Rogers [Fri, 18 Jul 2014 15:21:35 +0000 (15:21 +0000)]
Merge "Add optimized assembler implementation of __memcmp16 for x86."
Ian Rogers [Sat, 19 Jul 2014 03:12:33 +0000 (03:12 +0000)]
Merge "Black list more trace tests."
Andreas Gampe [Fri, 18 Jul 2014 23:40:02 +0000 (23:40 +0000)]
Merge "ART: Also accept java.lang.Throwable as a catch-all handler."
Andreas Gampe [Fri, 18 Jul 2014 22:41:00 +0000 (15:41 -0700)]
ART: Also accept java.lang.Throwable as a catch-all handler.
Accept catch handlers with catch type of java.lang.Throwable as
catch-all handlers.
Bug:
16308310
Change-Id: Ie9b9582ee71c94c82b7695dc6f9c2c6df3d869d8
Andreas Gampe [Fri, 18 Jul 2014 17:39:47 +0000 (17:39 +0000)]
Merge "ART: Do not attempt to compile a method with too many registers"
Andreas Gampe [Fri, 18 Jul 2014 01:04:32 +0000 (18:04 -0700)]
ART: Do not attempt to compile a method with too many registers
The mir_graph has a hard limit on how many ssa registers it can
handle, and will silently fail or fatal out in debug builds when
more SSA registers are necessary. This adds a check in the frontend
that tries to catch some of these cases, and makes the DCHECK a
CHECK to not silently fail.
Bug:
16018097
Change-Id: Iad82178945a8f8c34943cc7573fd73d2b4decade
Ian Rogers [Fri, 18 Jul 2014 15:08:14 +0000 (15:08 +0000)]
Merge "ART: inline Math.Max/Min (float and double)"
Calin Juravle [Fri, 18 Jul 2014 13:25:27 +0000 (13:25 +0000)]
Merge "Fix bad comment in class_linker"
Vladimir Marko [Fri, 18 Jul 2014 09:52:54 +0000 (09:52 +0000)]
Merge "Fix GVN to handle normal paths leading to catch entry."
Nicolas Geoffray [Fri, 18 Jul 2014 09:32:24 +0000 (09:32 +0000)]
Merge "We can also run the linear scan register allocator on thumb."
Nicolas Geoffray [Fri, 18 Jul 2014 09:23:59 +0000 (10:23 +0100)]
We can also run the linear scan register allocator on thumb.
Change-Id: I5d21b5cbcdd93ff36342111de4ebcaab172034dd
Andreas Gampe [Fri, 18 Jul 2014 03:53:31 +0000 (03:53 +0000)]
Merge "ART: Fail dex file verification on invalid access flags"
Andreas Gampe [Fri, 18 Jul 2014 02:26:50 +0000 (19:26 -0700)]
ART: Fail dex file verification on invalid access flags
Only allow standard java access flags in class-def items.
Bug:
16018234
Change-Id: I54e788817d5974faf7b3841cd01b9ca4568249c1
Ian Rogers [Fri, 18 Jul 2014 02:12:18 +0000 (02:12 +0000)]
Merge "Avoid race in single thread suspension."
Ian Rogers [Fri, 18 Jul 2014 01:52:42 +0000 (18:52 -0700)]
Avoid race in single thread suspension.
Don't allow more than one concurrent single thread suspension to avoid
potential cycles and deadlocks where threads try to suspend each other.
Bug:
16364458,
16354227
Change-Id: I907f1d5591a6aa5c241d37d6b4a34f968f98df77
Ian Rogers [Fri, 18 Jul 2014 16:50:27 +0000 (16:50 +0000)]
Merge "x86_64: Fix GenDalvikArgsRange for 64-bit ref"
Serguei Katkov [Tue, 15 Jul 2014 05:01:00 +0000 (12:01 +0700)]
x86_64: Fix GenDalvikArgsRange for 64-bit ref
32-bit virtual register can be in 64-bit solo register.
So we should not compute the size of virtual register
basing on size of phyical register.
Change-Id: I4e11be13df8469be63808d0ce9d1ca6f80bef483
Signed-off-by: Serguei Katkov <serguei.i.katkov@intel.com>
Andreas Gampe [Fri, 18 Jul 2014 01:32:01 +0000 (01:32 +0000)]
Merge "ART: Make a soft failure in method verifier hard, add check"
Andreas Gampe [Thu, 17 Jul 2014 23:13:35 +0000 (16:13 -0700)]
ART: Make a soft failure in method verifier hard, add check
Make the soft error of using a primitive type in an instanceof a
hard verifier error.
Add a check when doing check_cast peep-hole optimization to check
whether the cast type is valid.
Bug:
16020300
Change-Id: Icaea560d17804f24a13dc09c115e9ec7039cb472
Andreas Gampe [Fri, 18 Jul 2014 01:31:36 +0000 (01:31 +0000)]
Merge "ART: Use array-inl.h in reference_table_test"
Andreas Gampe [Thu, 17 Jul 2014 21:52:06 +0000 (14:52 -0700)]
ART: Use array-inl.h in reference_table_test
The test allocates arrays, which is an inline definition.
Change-Id: I1423c5419949d7c352ed0e614d4f9c5920831deb
Calin Juravle [Thu, 17 Jul 2014 20:14:23 +0000 (21:14 +0100)]
Fix bad comment in class_linker
Change-Id: Ic8c90e9d880c5aae7e7bdb6c2a0c854fdaa5c357
Calin Juravle [Thu, 17 Jul 2014 20:08:08 +0000 (20:08 +0000)]
Merge "Fix identifier size in the heap dump."
Calin Juravle [Fri, 4 Jul 2014 15:24:03 +0000 (16:24 +0100)]
Fix identifier size in the heap dump.
We're using addresses as IDs and previously we would have written
sizeof(void*) as the identifier size in the header. Heap references
are actually uint32_t so writting 4 bytes should be enough.
Bug:
16016862
Change-Id: I785abf5030681b6009370b7b2d7552e8b4b54e67
Vladimir Marko [Thu, 17 Jul 2014 19:50:07 +0000 (20:50 +0100)]
Fix GVN to handle normal paths leading to catch entry.
When the catch block is empty, the catch entry is actually
the normal path block after the try block. Fix the LVN
merge for catch entries that didn't expect it during GVN.
Bug:
16360024
Change-Id: I9adfc3445245d3fa3c4809d4df1b7b76fbef5ff2
Alexei Zavjalov [Tue, 15 Jul 2014 17:54:13 +0000 (00:54 +0700)]
ART: inline Math.Max/Min (float and double)
This implements the inlined version of Math.Max/Min intrinsics.
Change-Id: I2db8fa7603db3cdf01016ec26811a96f91b1e6ed
Signed-off-by: Alexei Zavjalov <alexei.zavjalov@intel.com>
Signed-off-by: Shou, Yixin <yixin.shou@intel.com>
Ian Rogers [Thu, 17 Jul 2014 18:49:18 +0000 (18:49 +0000)]
Merge "Avoid direct use of QuasiAtomic by using Atomic."
Ian Rogers [Thu, 17 Jul 2014 18:25:30 +0000 (11:25 -0700)]
Avoid direct use of QuasiAtomic by using Atomic.
Change-Id: I92d0f822cddb2085de7cbb4c44c3a429081892a2
Alexander Ivchenko [Mon, 14 Jul 2014 14:54:34 +0000 (18:54 +0400)]
Add optimized assembler implementation of __memcmp16 for x86.
Change-Id: Ia442dd749f8113244ea580f97d574ebaa73e4989
Signed-off-by: Alexander Ivchenko <alexander.ivchenko@intel.com>
Nicolas Geoffray [Thu, 17 Jul 2014 08:17:12 +0000 (08:17 +0000)]
Merge "Fix lint errors."
Nicolas Geoffray [Thu, 17 Jul 2014 08:16:19 +0000 (09:16 +0100)]
Fix lint errors.
Change-Id: I7bbaad590d84db97acfa7f8c4d093bc40e64684a
Nicolas Geoffray [Thu, 17 Jul 2014 08:00:17 +0000 (08:00 +0000)]
Merge "Use the thumb2 assembler for the optimizing compiler."
Andreas Gampe [Thu, 17 Jul 2014 17:03:43 +0000 (17:03 +0000)]
Merge "ART: Refactor GenSelect, refactor gen_common accordingly"
Andreas Gampe [Wed, 16 Jul 2014 06:02:11 +0000 (23:02 -0700)]
ART: Refactor GenSelect, refactor gen_common accordingly
This adds a GenSelect method meant for selection of constants. The
general-purpose GenInstanceof code is refactored to take advantage of
this. This cleans up code and squashes a branch-over on ARM64 to a
cset.
Also add a slow-path for type initialization in GenInstanceof.
Change-Id: Ie4494858bb8c26d386cf2e628172b81bba911ae5
Andreas Gampe [Thu, 17 Jul 2014 15:25:04 +0000 (15:25 +0000)]
Merge "ART: Lock an arg reg in GenInlinedGet"
Andreas Gampe [Thu, 17 Jul 2014 07:12:32 +0000 (00:12 -0700)]
ART: Lock an arg reg in GenInlinedGet
Must lock an explicit target register so that it doesn't get used
for a temp and clobbered.
Bug:
16352802
Change-Id: I9e4ac76f92460cdbe59392989d12f8d0321cc624
Ian Rogers [Thu, 17 Jul 2014 16:03:17 +0000 (16:03 +0000)]
Merge "Remove some to-done TODOs."
Ian Rogers [Thu, 17 Jul 2014 15:34:21 +0000 (08:34 -0700)]
Remove some to-done TODOs.
Change-Id: I976d35b1aa79f4dded86502a18bc679ba232d664
Stephen Hines [Wed, 16 Jul 2014 23:50:39 +0000 (23:50 +0000)]
Merge "Add back a deleted check related to verification."
Ian Rogers [Thu, 17 Jul 2014 05:49:49 +0000 (05:49 +0000)]
Merge "Fix test 304-method-tracing failure for Mips."
Douglas Leung [Thu, 17 Jul 2014 00:32:41 +0000 (17:32 -0700)]
Fix test 304-method-tracing failure for Mips.
Change-Id: Id5ce9fd3c880dd568b11a1deea7e42ea587934ba
Signed-off-by: Douglas Leung <douglas@mips.com>
Stephen Hines [Wed, 16 Jul 2014 23:50:34 +0000 (23:50 +0000)]
Merge "Remove errors/warnings from tautological comparisons."
Stephen Hines [Wed, 16 Jul 2014 18:03:43 +0000 (11:03 -0700)]
Add back a deleted check related to verification.
Change-Id: Ib939d1a7395a4b6cc9a7256cfb0a3498ba296744
Nicolas Geoffray [Wed, 16 Jul 2014 15:23:40 +0000 (16:23 +0100)]
Use the thumb2 assembler for the optimizing compiler.
Change-Id: I2b058f4433504dc3299c06f5cb0b5ab12f34aa82
Jeff Hao [Wed, 16 Jul 2014 23:28:05 +0000 (23:28 +0000)]
Merge "Fix art test failures for Mips."
Dave Allison [Wed, 16 Jul 2014 22:00:14 +0000 (22:00 +0000)]
Merge "Revert "Revert "Revert "Revert "Add implicit null and stack checks for x86"""""
Dave Allison [Fri, 11 Jul 2014 17:11:58 +0000 (17:11 +0000)]
Revert "Revert "Revert "Revert "Add implicit null and stack checks for x86""""
This reverts commit
0025a86411145eb7cd4971f9234fc21c7b4aced1.
Bug:
16256184
Change-Id: Ie0760a0c293aa3b62e2885398a8c512b7a946a73
Mathieu Chartier [Wed, 16 Jul 2014 19:55:08 +0000 (19:55 +0000)]
Merge "Disable adding main and non moving spaces to immune region in GSS"
Mathieu Chartier [Mon, 14 Jul 2014 21:57:16 +0000 (14:57 -0700)]
Disable adding main and non moving spaces to immune region in GSS
Disabled adding the main and non moving space to the immune region.
This will enable us to recycle bump pointer spaces for malloc space
-> malloc space compaction as well as collector transitions.
Also added logic for falling back to the non moving space, we may
copy objects there.
Refactored mod union table logic into MarkReachableObjects.
No measurable performance benefit or regression.
Bug:
14059466
Bug:
16291259
Change-Id: If663d9fdbde943b988173b7f6ac844e5f78a0327
Stephen Hines [Wed, 16 Jul 2014 04:41:06 +0000 (21:41 -0700)]
Remove errors/warnings from tautological comparisons.
Change-Id: I2728b08d55abccd14c111c6c9da93068d4e14275
Ian Rogers [Wed, 16 Jul 2014 21:51:58 +0000 (21:51 +0000)]
Merge "Add another test to the trace test black list."
Ian Rogers [Wed, 16 Jul 2014 21:49:34 +0000 (14:49 -0700)]
Add another test to the trace test black list.
Change-Id: Ib2266b4fa2684e477791fe317a598c22d79e3400
Douglas Leung [Wed, 9 Jul 2014 21:28:35 +0000 (14:28 -0700)]
Fix art test failures for Mips.
This patch fixes the following art test failures for Mips:
003-omnibus-opcodes
030-bad-finalizer
041-narrowing
059-finalizer-throw
Change-Id: I4e0e9ff75f949c92059dd6b8d579450dc15f4467
Signed-off-by: Douglas Leung <douglas@mips.com>
Mathieu Chartier [Wed, 16 Jul 2014 01:44:49 +0000 (01:44 +0000)]
Merge "Use sched_yield in Monitor::MonitorEnter."
Mathieu Chartier [Wed, 16 Jul 2014 01:10:25 +0000 (18:10 -0700)]
Use sched_yield in Monitor::MonitorEnter.
Previously we used NanoSleep(1000), but this was unreliable. It could
result in waiting for >= 40ms instead of 1us. Since this was in a loop
it was especially bad if the GC was trying to suspend all the
threads when we were sleeping. This resulted in thread suspension
occasionally taking longer than a second.
Results on the provided picasso-sample app on Nexus 5:
Longest GC pause before: ~1.5s.
Longest GC pause after: <5ms.
Also added a warning if thread suspension takes longer than a
threshold (currently 5ms).
Bug:
16307460
External bug: https://code.google.com/p/android-developer-preview/issues/detail?id=367
Change-Id: I3c2a9636357e255f38634615101eff8ca84e632f
Ian Rogers [Wed, 16 Jul 2014 18:23:06 +0000 (18:23 +0000)]
Merge "Add --trace variants of the run-test testing rules."
Ian Rogers [Wed, 16 Jul 2014 18:18:03 +0000 (11:18 -0700)]
Add --trace variants of the run-test testing rules.
Only for host and the default compiler at the moment.
Also, fix test-art-host32 and test-art-host64.
Change-Id: Ic13190e766aca522de95d2b5b12926c906ba8fb8
Ian Rogers [Wed, 16 Jul 2014 19:12:11 +0000 (19:12 +0000)]
Merge "Fix x86 build."
Ian Rogers [Wed, 16 Jul 2014 19:06:35 +0000 (12:06 -0700)]
Fix x86 build.
Also fix attributes/annotalysis on entrypoint_utils functions now we have
clang that is smarter wrt warnings than GCC.
Change-Id: I69257b4ad9a27d07acbc973d21a1cfa4260a8ed6
Sebastien Hertz [Wed, 16 Jul 2014 19:03:36 +0000 (19:03 +0000)]
Merge "Fix class initialization checks in interpreter"
Sebastien Hertz [Wed, 16 Jul 2014 18:00:11 +0000 (20:00 +0200)]
Fix class initialization checks in interpreter
Check field's class initialization after handling null pointer exception.
Bug:
16324235
Change-Id: I33d537168e068b7fb51ddf97bc5aadee9dd65f67
Ian Rogers [Wed, 16 Jul 2014 15:27:47 +0000 (15:27 +0000)]
Merge "Remove object_utils.h."
Ian Rogers [Wed, 16 Jul 2014 05:23:51 +0000 (22:23 -0700)]
Remove object_utils.h.
Break into object_lock, field_helper and method_helper.
Clean up header files following this.
Also tidy some of the Handle code in response to compiler errors when resolving
the changes in this CL.
Change-Id: I73e63015a0f02a754d0866bfaf58208aebcaa295
Calin Juravle [Wed, 16 Jul 2014 15:04:20 +0000 (15:04 +0000)]
Merge "Make ART fail gracefully when it can't update the desired code."
Calin Juravle [Wed, 16 Jul 2014 13:45:03 +0000 (14:45 +0100)]
Make ART fail gracefully when it can't update the desired code.
ART was exiting with a fatal error when it couldn't clean an obsolete
file. Relaxing this and failing gracefully preserves the behaviour that
Dalvik had.
Bug:
15313272
Change-Id: I8d0d6d374c90d2a434909dd4ae56f0799f30134d
Andreas Gampe [Tue, 15 Jul 2014 17:44:31 +0000 (17:44 +0000)]
Merge "ART: Squash a cmp w/ zero and b.ls to cbz (ARM/ARM64)"
Andreas Gampe [Tue, 15 Jul 2014 03:16:59 +0000 (20:16 -0700)]
ART: Squash a cmp w/ zero and b.ls to cbz (ARM/ARM64)
In case of array bounds checks at constant index 0 we generate a
compare and a branch. Squash into a cbz.
Change-Id: I1c6a6e37a7a2356b2c4580a3387cedb55436e251
Mathieu Chartier [Wed, 16 Jul 2014 00:59:09 +0000 (00:59 +0000)]
Merge "Revert "Revert "Revert "Revert "Add intrinsic for Reference.get()"""""
Fred Shih [Fri, 11 Jul 2014 16:59:27 +0000 (09:59 -0700)]
Revert "Revert "Revert "Revert "Add intrinsic for Reference.get()""""
Fixed TargetReg issue causing build failure for x86.
This reverts commit
9e82bd3f0ce9e5f5777bea2f752ff3e251d32f9f.
Change-Id: I7e6a526954467aaf68deeed999880dfe9aa5f06e
Andreas Gampe [Tue, 15 Jul 2014 18:52:36 +0000 (18:52 +0000)]
Merge "AArch64: improve usage of TargetReg() and friends."