platform/upstream/llvm.git
10 years agoAST: Cleanup __uuidof related code
David Majnemer [Mon, 14 Jul 2014 23:12:54 +0000 (23:12 +0000)]
AST: Cleanup __uuidof related code

Switch some things to use range-based for loops.
Change CXXUuidofExpr::GetUuidAttrOfType's return type to be const.
No functionality changed.

llvm-svn: 213009

10 years agoRevert "Move clang feature flags settings out of LLVM core and into cfe"
Alp Toker [Mon, 14 Jul 2014 23:10:07 +0000 (23:10 +0000)]
Revert "Move clang feature flags settings out of LLVM core and into cfe"

Prospectively revert to get one of the many build configurations back working
(presumably the side-by-side+modular CMake config).

This reverts commit r212997.

llvm-svn: 213008

10 years agoIf Process::Finalize() has been called, don't track process state changes.
Greg Clayton [Mon, 14 Jul 2014 23:09:29 +0000 (23:09 +0000)]
If Process::Finalize() has been called, don't track process state changes.

<rdar://problem/17540766>

llvm-svn: 213007

10 years agoCodeGen: Stick constant pool entries in COMDAT sections for WinCOFF
David Majnemer [Mon, 14 Jul 2014 22:57:27 +0000 (22:57 +0000)]
CodeGen: Stick constant pool entries in COMDAT sections for WinCOFF

COFF lacks a feature that other object file formats support: mergeable
sections.

To work around this, MSVC sticks constant pool entries in special COMDAT
sections so that each constant is in it's own section.  This permits
unused constants to be dropped and it also allows duplicate constants in
different translation units to get merged together.

This fixes PR20262.

Differential Revision: http://reviews.llvm.org/D4482

llvm-svn: 213006

10 years agoRevert "Move clang feature flags settings out of LLVM core and into cfe"
Alp Toker [Mon, 14 Jul 2014 22:54:22 +0000 (22:54 +0000)]
Revert "Move clang feature flags settings out of LLVM core and into cfe"

This broke one of the builds, presumably side-by-side modular CMake.
Investigating.

This reverts commit r212998.

llvm-svn: 213005

10 years agolldb.LLDB_ARCH_DEFAULT now works correctly on Macs with haswell enabled kernels.
Greg Clayton [Mon, 14 Jul 2014 22:53:02 +0000 (22:53 +0000)]
lldb.LLDB_ARCH_DEFAULT now works correctly on Macs with haswell enabled kernels.

<rdar://problem/17604133>

llvm-svn: 213004

10 years agoContinue parsing an expression list even after an error is encountered.
Kaelyn Takata [Mon, 14 Jul 2014 22:48:10 +0000 (22:48 +0000)]
Continue parsing an expression list even after an error is encountered.

Otherwise, multiple errors such as having unknown identifiers for two
arguments won't be diagnosed properly (e.g. only the first one would
have a diagnostic message if typo correction fails even though both
would be diagnosed if typo correction suggests a replacement).

llvm-svn: 213003

10 years agoFix a -Wunused-local-typedefs warning
Alp Toker [Mon, 14 Jul 2014 22:46:45 +0000 (22:46 +0000)]
Fix a -Wunused-local-typedefs warning

llvm-svn: 213002

10 years ago[DAGCombiner] Add more rules to combine shuffle vector dag nodes.
Andrea Di Biagio [Mon, 14 Jul 2014 22:46:26 +0000 (22:46 +0000)]
[DAGCombiner] Add more rules to combine shuffle vector dag nodes.

This patch teaches the DAGCombiner how to fold a pair of shuffles
according to rules:
  1.  shuffle(shuffle A, B, M0), B, M1) -> shuffle(A, B, M2)
  2.  shuffle(shuffle A, B, M0), A, M1) -> shuffle(A, B, M3)

The new rules would only trigger if the resulting shuffle has legal type and
legal mask.

Added test 'combine-vec-shuffle-3.ll' to verify that DAGCombiner correctly
folds shuffles on x86 when the resulting mask is legal. Also added some negative
cases to verify that we avoid introducing illegal shuffles.

llvm-svn: 213001

10 years agoLook through addrspacecast in IsConstantOffsetFromGlobal
Matt Arsenault [Mon, 14 Jul 2014 22:39:26 +0000 (22:39 +0000)]
Look through addrspacecast in IsConstantOffsetFromGlobal

llvm-svn: 213000

10 years agoLook through addrspacecast in GetPointerBaseWithConstantOffset
Matt Arsenault [Mon, 14 Jul 2014 22:39:22 +0000 (22:39 +0000)]
Look through addrspacecast in GetPointerBaseWithConstantOffset

llvm-svn: 212999

10 years agoMove clang feature flags settings out of LLVM core and into cfe
Alp Toker [Mon, 14 Jul 2014 22:19:04 +0000 (22:19 +0000)]
Move clang feature flags settings out of LLVM core and into cfe

clang r212997 incorporated these settings into its own build system.  They no
longer need to be set from LLVM.

llvm-svn: 212998

10 years agoMove clang feature flags settings out of LLVM core and into cfe
Alp Toker [Mon, 14 Jul 2014 22:17:26 +0000 (22:17 +0000)]
Move clang feature flags settings out of LLVM core and into cfe

llvm-svn: 212997

10 years agoFix build with various feature flag combinations
Alp Toker [Mon, 14 Jul 2014 22:17:22 +0000 (22:17 +0000)]
Fix build with various feature flag combinations

llvm-svn: 212996

10 years agoFix the !CLANG_ENABLE_ARCMT build
Alp Toker [Mon, 14 Jul 2014 22:17:16 +0000 (22:17 +0000)]
Fix the !CLANG_ENABLE_ARCMT build

llvm-svn: 212995

10 years agoFix build when various clang feature flags are enabled/disabled
Alp Toker [Mon, 14 Jul 2014 22:15:29 +0000 (22:15 +0000)]
Fix build when various clang feature flags are enabled/disabled

llvm-svn: 212994

10 years agoCodeGen: Add a getSectionKind method to MachineConstantPoolEntry
David Majnemer [Mon, 14 Jul 2014 22:06:29 +0000 (22:06 +0000)]
CodeGen: Add a getSectionKind method to MachineConstantPoolEntry

This is just a helper routine, no functionality has changed.

llvm-svn: 212993

10 years agoConvert test to FileCheck
Matt Arsenault [Mon, 14 Jul 2014 21:59:26 +0000 (21:59 +0000)]
Convert test to FileCheck

llvm-svn: 212992

10 years agoADT: Surface LowerCase argument for utohexstr
David Majnemer [Mon, 14 Jul 2014 21:56:54 +0000 (21:56 +0000)]
ADT: Surface LowerCase argument for utohexstr

The underlying function. utohex_buffer, already supports an argument for
deciding if the hex characters should be upper or lower case.  Expose an
identical argument for utohexstr.

llvm-svn: 212991

10 years agoremoved circular definitions in comments
Sanjay Patel [Mon, 14 Jul 2014 21:51:59 +0000 (21:51 +0000)]
removed circular definitions in comments

llvm-svn: 212990

10 years agoTry harder to supress the test from r212975 on Windows
Ben Langmuir [Mon, 14 Jul 2014 21:14:28 +0000 (21:14 +0000)]
Try harder to supress the test from r212975 on Windows

The attempt in r212980 was broken because we might not fail if
LLVM_ON_UNIX is enabled for cross compiling to Windows, and it didn't
consider mingw either.

llvm-svn: 212989

10 years agoDrop separate UIntMaxType and just derive it from IntMaxType.
Joerg Sonnenberger [Mon, 14 Jul 2014 21:00:38 +0000 (21:00 +0000)]
Drop separate UIntMaxType and just derive it from IntMaxType.

llvm-svn: 212987

10 years agoSupport: Fix option handling when using cl::Required with aliasopt
Justin Bogner [Mon, 14 Jul 2014 20:53:57 +0000 (20:53 +0000)]
Support: Fix option handling when using cl::Required with aliasopt

Until now, attempting to create an alias of a required option would
complain if the user supplied the alias, because the required option
didn't have a value. Similarly, if you said the alias was required,
then using the base option would complain that the alias wasn't
supplied. Lastly, if you put required on both, *neither* option would
work.

By changning alias to overload addOccurrence and setting cl::Required
on the original option, we can get this to behave in a more useful
way. I've also added a test and updated a user that was getting this
wrong.

llvm-svn: 212986

10 years agoFix a test broken in r212981
David Majnemer [Mon, 14 Jul 2014 20:46:04 +0000 (20:46 +0000)]
Fix a test broken in r212981

@icmp_sdiv_neg1 should have referred to %a instead of %call, it was
renamed at the last second.

llvm-svn: 212983

10 years agoIntroduce getCorrespondingUnsignedType() in TargetInfo to work like the
Joerg Sonnenberger [Mon, 14 Jul 2014 20:40:56 +0000 (20:40 +0000)]
Introduce getCorrespondingUnsignedType() in TargetInfo to work like the
corresponding AST context function, only restricted to basic integer
types. Use this to ensure getUIntPtrType() gives types consistent with
getIntPtrType(). Fix NVPTX backend to give signed intptr_t.

llvm-svn: 212982

10 years agoInstSimplify: Correct sdiv x / -1
David Majnemer [Mon, 14 Jul 2014 20:38:45 +0000 (20:38 +0000)]
InstSimplify: Correct sdiv x / -1

Determining the bounds of x/ -1 would start off with us dividing it by
INT_MIN.  Suffice to say, this would not work very well.

Instead, handle it upfront by checking for -1 and mapping it to the
range: [INT_MIN + 1, INT_MAX.  This means that the result of our
division can be any value other than INT_MIN.

llvm-svn: 212981

10 years agoXFAIL test that doesn't work on Windows yet
Ben Langmuir [Mon, 14 Jul 2014 20:36:40 +0000 (20:36 +0000)]
XFAIL test that doesn't work on Windows yet

llvm-svn: 212980

10 years agoAdd support for intercepting thunks of the form
Ehsan Akhgari [Mon, 14 Jul 2014 20:28:21 +0000 (20:28 +0000)]
Add support for intercepting thunks of the form

llvm-svn: 212979

10 years agoIn C++98, if an rvalue reference binds to a function lvalue (or an xvalue or an
Richard Smith [Mon, 14 Jul 2014 19:54:05 +0000 (19:54 +0000)]
In C++98, if an rvalue reference binds to a function lvalue (or an xvalue or an
array prvalue), treat that as a direct binding. Only the class prvalue case
needs to be excluded here; the rest are extensions anyway, so we can treat them
as we would in C++11.

llvm-svn: 212978

10 years agofixed link
Sanjay Patel [Mon, 14 Jul 2014 19:52:36 +0000 (19:52 +0000)]
fixed link

llvm-svn: 212977

10 years agoInstSimplify: The upper bound of X / C was missing a rounding step
David Majnemer [Mon, 14 Jul 2014 19:49:57 +0000 (19:49 +0000)]
InstSimplify: The upper bound of X / C was missing a rounding step

Summary:
When calculating the upper bound of X / -8589934592, we would perform
the following calculation: Floor[INT_MAX / 8589934592]

However, flooring the result would make us wrongly come to the
conclusion that 1073741824 was not in the set of possible values.
Instead, use the ceiling of the result.

Reviewers: nicholas

Subscribers: llvm-commits

Differential Revision: http://reviews.llvm.org/D4502

llvm-svn: 212976

10 years agoFix case-sensitivity of inferred framework modules
Ben Langmuir [Mon, 14 Jul 2014 19:45:12 +0000 (19:45 +0000)]
Fix case-sensitivity of inferred framework modules

Just because we can open a directory named "COcoa.framework" doesn't
mean we should provide a "COcoa" module on a case-insensitive filesystem.

llvm-svn: 212975

10 years agoFix typos
Alp Toker [Mon, 14 Jul 2014 19:42:55 +0000 (19:42 +0000)]
Fix typos

Also consolidate 'backward compatibility'

llvm-svn: 212974

10 years agoSupport: Use a range-based for
Justin Bogner [Mon, 14 Jul 2014 19:24:13 +0000 (19:24 +0000)]
Support: Use a range-based for

llvm-svn: 212973

10 years agoAvoid non-attributive uses of 'unsupported' in diagnostics
Alp Toker [Mon, 14 Jul 2014 19:16:22 +0000 (19:16 +0000)]
Avoid non-attributive uses of 'unsupported' in diagnostics

We don't have a style guide for diagnostic messages, but convention strongly
favours the forms:

  'attribute is not supported', 'unsupported attribute'

We generally avoid:

  'attribute is unsupported', 'non-supported attribute'

llvm-svn: 212972

10 years agoLook through addrspacecast when checking isDereferenceablePointer
Matt Arsenault [Mon, 14 Jul 2014 18:54:12 +0000 (18:54 +0000)]
Look through addrspacecast when checking isDereferenceablePointer

llvm-svn: 212971

10 years agoDon't eliminate memcpy's when the address of the pointer may itself be relevant....
Nick Lewycky [Mon, 14 Jul 2014 18:52:02 +0000 (18:52 +0000)]
Don't eliminate memcpy's when the address of the pointer may itself be relevant. Fixes PR18304. Patch by David Wiberg!

llvm-svn: 212970

10 years agoDocument the forward of the -u option (r211756)
Sylvestre Ledru [Mon, 14 Jul 2014 18:34:37 +0000 (18:34 +0000)]
Document the forward of the -u option (r211756)

llvm-svn: 212969

10 years agoDocument the warning triggered by the unsupported gcc optimization flags (r212805)
Sylvestre Ledru [Mon, 14 Jul 2014 18:30:23 +0000 (18:30 +0000)]
Document the warning triggered by the unsupported gcc optimization flags (r212805)

llvm-svn: 212968

10 years agoUnify the lowering of arguments during SjLj prepare.
Bill Wendling [Mon, 14 Jul 2014 18:21:11 +0000 (18:21 +0000)]
Unify the lowering of arguments during SjLj prepare.

The 'select true, %arg, undef' instruction can be used for both aggregate and
non-aggregate arguments.

llvm-svn: 212967

10 years agofixed typo
Sanjay Patel [Mon, 14 Jul 2014 18:21:07 +0000 (18:21 +0000)]
fixed typo

llvm-svn: 212966

10 years agoRevert "Improve error recovery around colon."
Reid Kleckner [Mon, 14 Jul 2014 18:19:58 +0000 (18:19 +0000)]
Revert "Improve error recovery around colon."

This reverts commit r212957.  It broke the self-host on code like this
from LLVM's option library:
  for (auto Arg: filtered(Id0, Id1, Id2))

llvm-svn: 212965

10 years agoChange the diagnostic group for unsupported gcc optimizations added r212805
Alp Toker [Mon, 14 Jul 2014 18:11:38 +0000 (18:11 +0000)]
Change the diagnostic group for unsupported gcc optimizations added r212805

Use -Winvalid-command-line-argument here to align with existing gcc opt
diagnostics. Meanwhile -Wunused-command-line-argument is for flags that we
support but were, say, fed into the wrong invocation.

Also tweak wording to make sense with -Werror.

llvm-svn: 212964

10 years agoUse pointer type cast helpers.
Matt Arsenault [Mon, 14 Jul 2014 17:24:38 +0000 (17:24 +0000)]
Use pointer type cast helpers.

llvm-svn: 212963

10 years agoAdd CreatePointerBitCastOrAddrSpaceCast to IRBuilder and co.
Matt Arsenault [Mon, 14 Jul 2014 17:24:35 +0000 (17:24 +0000)]
Add CreatePointerBitCastOrAddrSpaceCast to IRBuilder and co.

llvm-svn: 212962

10 years agoUpdate comments to include addrspacecast
Matt Arsenault [Mon, 14 Jul 2014 17:24:31 +0000 (17:24 +0000)]
Update comments to include addrspacecast

llvm-svn: 212961

10 years agoRemove GCC 3.3 workaround
Matt Arsenault [Mon, 14 Jul 2014 17:11:20 +0000 (17:11 +0000)]
Remove GCC 3.3 workaround

llvm-svn: 212960

10 years ago[ASan/Win] Add a test for the case when dll_thunk is linked twice to the same DLL
Timur Iskhodzhanov [Mon, 14 Jul 2014 17:04:03 +0000 (17:04 +0000)]
[ASan/Win] Add a test for the case when dll_thunk is linked twice to the same DLL

This is a test case for r212815.

llvm-svn: 212959

10 years agoFix for libc++abi when using -Wl,--as-needed.
Dan Albert [Mon, 14 Jul 2014 17:00:29 +0000 (17:00 +0000)]
Fix for libc++abi when using -Wl,--as-needed.

The cmake files for libc++abi and the unwinder weren't linking against
libpthread or an unwind library. If the tests were linked with
-Wl,--as-needed, these libraries wouldn't be linked, causing them to
fail.

Patch contributed by Ä°smail Dönmez.

llvm-svn: 212958

10 years agoImprove error recovery around colon.
Serge Pavlov [Mon, 14 Jul 2014 16:42:20 +0000 (16:42 +0000)]
Improve error recovery around colon.

Recognize additional cases, when '::' is mistyped as ':'.
This is a fix to RP18587 - colons have too much protection in member-declarations.

Differential Revision: http://reviews.llvm.org/D3653

llvm-svn: 212957

10 years agoX86: correct 64-bit atomics on 32-bit
Saleem Abdulrasool [Mon, 14 Jul 2014 16:28:13 +0000 (16:28 +0000)]
X86: correct 64-bit atomics on 32-bit

We would emit a libcall for a 64-bit atomic on x86 after SVN r212119.  This was
due to the misuse of hasCmpxchg16 to indicate if cmpxchg8b was supported on a
32-bit target.  They were added at different times and would result in the
border condition being mishandled.

This fixes the border case to emit the cmpxchg8b instruction for 64-bit atomic
operations on x86 at the cost of restoring a long-standing bug in the codegen.
We emit a cmpxchg8b on all x86 targets even where the CPU does not support this
instruction (pre-Pentium CPUs).  Although this bug should be fixed, this was
present prior to SVN r212119 and this change, so this is not really introducing
a regression.

llvm-svn: 212956

10 years agoCodeGen: add missing include
Saleem Abdulrasool [Mon, 14 Jul 2014 16:28:09 +0000 (16:28 +0000)]
CodeGen: add missing include

Found during windows unwinding work.  This header is indirectly included through
a chain leading through Support/Win64EH.h.  Explicitly include the header.  NFC.

llvm-svn: 212955

10 years agoCodeGen: Let arrays be inputs to inline asm
David Majnemer [Mon, 14 Jul 2014 16:27:53 +0000 (16:27 +0000)]
CodeGen: Let arrays be inputs to inline asm

An array showing up in an inline assembly input is accepted in ICC and
GCC 4.8

This fixes PR20201.

Differential Revision: http://reviews.llvm.org/D4382

llvm-svn: 212954

10 years agollvm-objdump: Handle BSS sections larger than the object file
David Majnemer [Mon, 14 Jul 2014 16:20:14 +0000 (16:20 +0000)]
llvm-objdump: Handle BSS sections larger than the object file

The size of the uninitialized sections, like BSS, can exceed the size of
the object file.

Do not attempt to grab the contents of such sections.

llvm-svn: 212953

10 years ago[tsan] honour XFAIL in tsan/test_output.sh
Kostya Serebryany [Mon, 14 Jul 2014 16:03:13 +0000 (16:03 +0000)]
[tsan] honour XFAIL in tsan/test_output.sh

llvm-svn: 212952

10 years ago[tsan] add a printf to a test
Kostya Serebryany [Mon, 14 Jul 2014 15:52:33 +0000 (15:52 +0000)]
[tsan] add a printf to a test

llvm-svn: 212951

10 years agoARM: Add NOP intrinsic mapping in arm_acle.h
Yi Kong [Mon, 14 Jul 2014 15:32:29 +0000 (15:32 +0000)]
ARM: Add NOP intrinsic mapping in arm_acle.h

llvm-svn: 212950

10 years ago[Mips] Fix test case expectations due the latest changes in the LLVM.
Simon Atanasyan [Mon, 14 Jul 2014 15:31:47 +0000 (15:31 +0000)]
[Mips] Fix test case expectations due the latest changes in the LLVM.

llvm-svn: 212949

10 years agoX86: remove temporary atomicrmw used during lowering.
Tim Northover [Mon, 14 Jul 2014 15:31:13 +0000 (15:31 +0000)]
X86: remove temporary atomicrmw used during lowering.

We construct a temporary "atomicrmw xchg" instruction when lowering atomic
stores for widths that aren't supported natively. This isn't on the top-level
worklist though, so it won't be removed automatically and we have to do it
ourselves once that itself has been lowered.

Thanks Saleem for pointing this out!

llvm-svn: 212948

10 years agoARM: Implement __builtin_arm_nop intrinsic
Yi Kong [Mon, 14 Jul 2014 15:20:09 +0000 (15:20 +0000)]
ARM: Implement __builtin_arm_nop intrinsic

This patch implements __builtin_arm_nop intrinsic for AArch32 and AArch64,
which generates hint 0x0, the alias of NOP instruction.

This intrinsic is necessary to implement ACLE __nop intrinsic.

Differential Revision: http://reviews.llvm.org/D4495

llvm-svn: 212947

10 years agoRe-commit: [mips] Correct section alignments and EntrySizes for .bss, .text, .data...
Daniel Sanders [Mon, 14 Jul 2014 15:05:51 +0000 (15:05 +0000)]
Re-commit: [mips] Correct section alignments and EntrySizes for .bss, .text, .data, .reginfo, .MIPS.options, and .MIPS.abiflags

The lld tests will temporarily fail again but Simon Atanasyan will commit a fix for those shortly.

llvm-svn: 212946

10 years agoRevert: [mips] Correct section alignments and EntrySizes for .bss, .text, .data,...
Daniel Sanders [Mon, 14 Jul 2014 14:43:45 +0000 (14:43 +0000)]
Revert: [mips] Correct section alignments and EntrySizes for .bss, .text, .data, .reginfo, .MIPS.options, and .MIPS.abiflags

This commit causes multiple lld tests to fail. Reverting while I investigate the issue.

llvm-svn: 212945

10 years ago[tsan] add a currently-failing test with a must-deadlock
Kostya Serebryany [Mon, 14 Jul 2014 14:27:21 +0000 (14:27 +0000)]
[tsan] add a currently-failing test with a must-deadlock

llvm-svn: 212944

10 years ago[ASan] Fix a couple of nits in NULL deref tests
Timur Iskhodzhanov [Mon, 14 Jul 2014 14:26:35 +0000 (14:26 +0000)]
[ASan] Fix a couple of nits in NULL deref tests

This is a follow-up to r212807

llvm-svn: 212943

10 years ago[clang-tidy] Add a checker for swapped arguments.
Benjamin Kramer [Mon, 14 Jul 2014 14:24:30 +0000 (14:24 +0000)]
[clang-tidy] Add a checker for swapped arguments.

This looks for swapped arguments by looking at implicit conversions of arguments

void Foo(int, double);
Foo(1.0, 3); // Most likely a bug

llvm-svn: 212942

10 years agoSet up clang-tidy diagnostic consumer to print types etc.
Alexander Kornienko [Mon, 14 Jul 2014 14:10:03 +0000 (14:10 +0000)]
Set up clang-tidy diagnostic consumer to print types etc.

Reviewers: bkramer

Reviewed By: bkramer

Subscribers: cfe-commits

Differential Revision: http://reviews.llvm.org/D4494

llvm-svn: 212941

10 years ago[ASTMatchers] Make hasOverloadedOperatorName also match freestanding overloads.
Benjamin Kramer [Mon, 14 Jul 2014 14:05:02 +0000 (14:05 +0000)]
[ASTMatchers] Make hasOverloadedOperatorName also match freestanding overloads.

Freestanding overloads are represented as FunctionDecls in the AST, make
the matcher also match them.

Differential Revision: http://reviews.llvm.org/D4493

llvm-svn: 212940

10 years ago[mips] Correct section alignments and EntrySizes for .bss, .text, .data, .reginfo...
Daniel Sanders [Mon, 14 Jul 2014 14:02:14 +0000 (14:02 +0000)]
[mips] Correct section alignments and EntrySizes for .bss, .text, .data, .reginfo, .MIPS.options, and .MIPS.abiflags

Summary:
.bss, .text, and .data are at least 16-byte aligned.
.reginfo is 4-byte aligned and has a 24-byte EntrySize.
.MIPS.abiflags has an 24-byte EntrySize.
.MIPS.options is 8-byte aligned and has 1-byte EntrySize.

Using a 1-byte EntrySize for .MIPS.options seems strange because the
records are neither 1-byte long nor fixed-length but this matches the value
that GAS emits.

Differential Revision: http://reviews.llvm.org/D4487

llvm-svn: 212939

10 years ago[mips] For the FP64A ABI, odd-numbered double-precision moves must not use mtc1/mfc1.
Daniel Sanders [Mon, 14 Jul 2014 13:08:14 +0000 (13:08 +0000)]
[mips] For the FP64A ABI, odd-numbered double-precision moves must not use mtc1/mfc1.

Summary:
This is because the FP64A the hardware will redirect 32-bit reads/writes
from/to odd-numbered registers to the upper 32-bits of the corresponding
even register. In effect, simulating FR=0 mode when FR=0 mode is not
available.

Unfortunately, we have to make the decision to avoid mfc1/mtc1 before
register allocation so we currently do this for even registers too.

FPXX has a similar requirement on 32-bit architectures that lack
mfhc1/mthc1 so this patch also handles the affected moves from the FPU for
FPXX too. Moves to the FPU were supported by an earlier commit.

Differential Revision: http://reviews.llvm.org/D4484

llvm-svn: 212938

10 years ago[sanitizer] Intercept getpass.
Evgeniy Stepanov [Mon, 14 Jul 2014 13:07:51 +0000 (13:07 +0000)]
[sanitizer] Intercept getpass.

llvm-svn: 212937

10 years ago[mips] Use MFHC1 when it is available (MIPS32r2 and later) for both FP32 and FP64...
Daniel Sanders [Mon, 14 Jul 2014 12:41:31 +0000 (12:41 +0000)]
[mips] Use MFHC1 when it is available (MIPS32r2 and later) for both FP32 and FP64 moves

Summary:
This is similar to r210771 which did the same thing for MTHC1.

Also corrected MTHC1_D32 and MTHC1_D64 which used AFGR64 and FGR64 on the
wrong definitions.

Differential Revision: http://reviews.llvm.org/D4483

llvm-svn: 212936

10 years agoclang-format: Improve cast detection (fix false positive).
Daniel Jasper [Mon, 14 Jul 2014 12:38:38 +0000 (12:38 +0000)]
clang-format: Improve cast detection (fix false positive).

Before:
  fn(a)(b)+1;

After:
  fn(a)(b) + 1;

llvm-svn: 212935

10 years ago[CMake][Win32.DLL] Let llvm_add_library(SHARED) link dependent libraries as PRIVATE.
NAKAMURA Takumi [Mon, 14 Jul 2014 12:26:15 +0000 (12:26 +0000)]
[CMake][Win32.DLL] Let llvm_add_library(SHARED) link dependent libraries as PRIVATE.

For example, c-index-test.exe requires just libclang.dll (its import library).
When libraries in libclang were not PRIVATE but PUBLIC, c-index-test required libraries transitive by libclang.

Note, on mingw with BUILD_SHARED_LIBS, library dependencies would become more strict.
In principle, required libraries should be "required in its source file".

This will help to detect missing dependencies.

llvm-svn: 212934

10 years agoAArch64: remove unnecessary pseudo-instruction.
Tim Northover [Mon, 14 Jul 2014 11:16:02 +0000 (11:16 +0000)]
AArch64: remove unnecessary pseudo-instruction.

Sufficiently twisted use of TableGen lets us write patterns directly for f16
(as an i16 promoted to i32) -> f32 conversion.

llvm-svn: 212933

10 years ago[mips] Correct the AFL_FLAGS1_ODDSPREG flag in .MIPS.abiflags when no '.module oddspr...
Daniel Sanders [Mon, 14 Jul 2014 10:26:15 +0000 (10:26 +0000)]
[mips] Correct the AFL_FLAGS1_ODDSPREG flag in .MIPS.abiflags when no '.module oddspreg' is used

Differential Revision: http://reviews.llvm.org/D4486

llvm-svn: 212932

10 years ago[x32] Add __ILP32__ macro for ILP32 platforms
Pavel Chupin [Mon, 14 Jul 2014 09:58:10 +0000 (09:58 +0000)]
[x32] Add __ILP32__ macro for ILP32 platforms

Summary:
Add __ILP32__ and _ILP32 macro for corresponding platforms.
Cover x86_64-*-*-gnux32 with test.

Test Plan: test added

Reviewers: chandlerc, atanasyan

Subscribers: cfe-commits, dschuff, zinovy.nis

Differential Revision: http://reviews.llvm.org/D4473

llvm-svn: 212931

10 years ago[mips] Expand BuildPairF64 to a spill and reload when the O32 FPXX ABI is
Sasa Stankovic [Mon, 14 Jul 2014 09:40:29 +0000 (09:40 +0000)]
[mips] Expand BuildPairF64 to a spill and reload when the O32 FPXX ABI is
enabled and mthc1 and dmtc1 are not available (e.g. on MIPS32r1)

This prevents the upper 32-bits of a double precision value from being moved to
the FPU with mtc1 to an odd-numbered FPU register. This is necessary to ensure
that the code generated executes correctly regardless of the current FPU mode.

MIPS32r2 and above continues to use mtc1/mthc1, while MIPS-IV and above continue
to use dmtc1.

Differential Revision: http://reviews.llvm.org/D4465

llvm-svn: 212930

10 years ago[msan] Stop demangling function name in the stack origin report.
Evgeniy Stepanov [Mon, 14 Jul 2014 09:35:27 +0000 (09:35 +0000)]
[msan] Stop demangling function name in the stack origin report.

This was done by calling __cxa_demangle directly, which is bad
when c++abi library is instrumented. The following line always
contains the demangled name (when running with a symbolizer) anyway.

llvm-svn: 212929

10 years ago[Mips] Delete the empty directory.
Simon Atanasyan [Mon, 14 Jul 2014 09:01:40 +0000 (09:01 +0000)]
[Mips] Delete the empty directory.

llvm-svn: 212928

10 years ago[Mips] Remove unused test input file.
Simon Atanasyan [Mon, 14 Jul 2014 08:54:40 +0000 (08:54 +0000)]
[Mips] Remove unused test input file.

llvm-svn: 212927

10 years ago[asan] Remove XFAIL:android from 2 newly passing tests.
Evgeniy Stepanov [Mon, 14 Jul 2014 08:54:10 +0000 (08:54 +0000)]
[asan] Remove XFAIL:android from 2 newly passing tests.

Fixed in r212872.

llvm-svn: 212926

10 years agoDeadCodeElimination: Fix liveout computation
Tobias Grosser [Mon, 14 Jul 2014 08:32:01 +0000 (08:32 +0000)]
DeadCodeElimination: Fix liveout computation

We move back to a simple approach where the liveout is the last must-write
statement for a data-location plus all may-write statements. The previous
approach did not work out. We would have to consider per-data-access
dependences, instead of per-statement dependences to correct it. As this adds
complexity and it seems we would not gain anything over the simpler approach
that we implement in this commit, I moved us back to the old approach of
computing the liveout, but enhanced it to also add may-write accesses.

We also fix the test case and explain why we can not perform dead code
elimination in this case.

llvm-svn: 212925

10 years agoSuggest automated replacements of C-style casts with C++ casts.
Alexander Kornienko [Mon, 14 Jul 2014 07:37:05 +0000 (07:37 +0000)]
Suggest automated replacements of C-style casts with C++ casts.

Summary:
This patch implements a subset of possible replacements of C-style
casts with const_cast/static_cast/reinterpret_cast. This should cover a large
portion of cases in real code. Handling of a few more cases may be implemented
eventually.

Reviewers: sbenza, djasper

Reviewed By: djasper

Subscribers: cfe-commits

Differential Revision: http://reviews.llvm.org/D4478

llvm-svn: 212924

10 years agoModified gdb-remote tests to run with automatically-chosen ports.
Todd Fiala [Mon, 14 Jul 2014 06:24:44 +0000 (06:24 +0000)]
Modified gdb-remote tests to run with automatically-chosen ports.

Now that llgs supports communicating the 0-port choose-a-port
mechanism and can communicate that back to a caller via the
--named-pipe option (at parity with debugserver), we use this
mechanism to always start llgs and debugserver gdb-remote
protocol tests without needing to use some port arbitration
mechanism.  This eliminates some potential intermittent failures vs. the
previous random port and collision-avoidance strategy used.

llvm-svn: 212923

10 years agoSupport lowering of empty aggregates.
Bill Wendling [Mon, 14 Jul 2014 06:22:36 +0000 (06:22 +0000)]
Support lowering of empty aggregates.

This crash was pretty common while compiling Rust for iOS (armv7). Reason -
SjLj preparation step was lowering aggregate arguments as ExtractValue +
InsertValue. ExtractValue has assertion which checks that there is some data in
value, which is not true in case of empty (no fields) structures. Rust uses
them quite extensively so this patch uses a 'select true, %val, undef'
instruction to lower the argument.

Patch by Valerii Hiora.

llvm-svn: 212922

10 years ago[CMake] LINK_COMPONENTS: Add also corresponding MCTargetDesc and TargetInfo as well...
NAKAMURA Takumi [Mon, 14 Jul 2014 05:07:07 +0000 (05:07 +0000)]
[CMake] LINK_COMPONENTS: Add also corresponding MCTargetDesc and TargetInfo as well, when target names or "nativecodegen" are specified.

llvm-svn: 212921

10 years ago[CMake] Update libdeps.
NAKAMURA Takumi [Mon, 14 Jul 2014 05:01:53 +0000 (05:01 +0000)]
[CMake] Update libdeps.

llvm-svn: 212920

10 years ago[CMake] Reorder libdeps by alphabetical order.
NAKAMURA Takumi [Mon, 14 Jul 2014 04:59:27 +0000 (04:59 +0000)]
[CMake] Reorder libdeps by alphabetical order.

llvm-svn: 212919

10 years agoNVPTX/LLVMBuild.txt: Add "Scalar" to required_libraries. It is really referenced.
NAKAMURA Takumi [Mon, 14 Jul 2014 02:52:19 +0000 (02:52 +0000)]
NVPTX/LLVMBuild.txt: Add "Scalar" to required_libraries. It is really referenced.

llvm-svn: 212918

10 years agoObject/LLVMBuild.txt: Sort required_libraries by alphabetical order.
NAKAMURA Takumi [Mon, 14 Jul 2014 02:52:08 +0000 (02:52 +0000)]
Object/LLVMBuild.txt: Sort required_libraries by alphabetical order.

llvm-svn: 212917

10 years agoSuperficial fix for PR20218: binding a function lvalue to a const reference to
Richard Smith [Mon, 14 Jul 2014 02:28:44 +0000 (02:28 +0000)]
Superficial fix for PR20218: binding a function lvalue to a const reference to
a function pointer is neither better nor worse than binding a function lvalue
to a function rvalue reference. Don't get confused and think that both bindings
are binding to a function lvalue (which would make the lvalue form win); the
const reference is binding to an rvalue.

The "real" bug in PR20218 is still present: we're getting the wrong answer from
template argument deduction, and that's what leads us to this weird overload
set.

llvm-svn: 212916

10 years ago[DAGCombiner] Fix a crash caused by a missing check for legal type when trying to...
Andrea Di Biagio [Sun, 13 Jul 2014 21:02:14 +0000 (21:02 +0000)]
[DAGCombiner] Fix a crash caused by a missing check for legal type when trying to fold shuffles.

Verify that DAGCombiner does not crash when trying to fold a pair of shuffles
according to rule (added at r212539):
  (shuffle (shuffle A, Undef, M0), Undef, M1) -> (shuffle A, Undef, M2)

The DAGCombiner avoids folding shuffles if the resulting shuffle dag node
is not legal for the target. That means, the resulting shuffle must have
legal type and legal mask.

Before, the DAGCombiner only called method
'TargetLowering::isShuffleMaskLegal' to check if it was "safe" to fold according
to the above-mentioned rule. However, this caused a crash in the x86 backend
since method 'isShuffleMaskLegal' always expects to be called on a
legal vector type.

llvm-svn: 212915

10 years agoMC: make MCWin64EHInstruction a POD-like struct
Saleem Abdulrasool [Sun, 13 Jul 2014 19:03:45 +0000 (19:03 +0000)]
MC: make MCWin64EHInstruction a POD-like struct

This is the first of a number of changes designed to generalise
MCWin64EHInstruction to support different target architectures.  An ordered set
(vector) of these instructions is saved per frame to permit the emission of
information for Windows NT style unwinding.  The only bit of information which
is actually target specific here is the Opcode for the unwinding bytecode.  The
remainder of the information is simply generic information that is relevant to
the Windows NT unwinding model.

Remove the accessors for the fields, making them const and public instead.  Sink
the knowledge of the alias'ed name into the single source and sink a single-use
check method into the use.

llvm-svn: 212914

10 years agoMC: make helper function be more const-correct
Saleem Abdulrasool [Sun, 13 Jul 2014 19:03:40 +0000 (19:03 +0000)]
MC: make helper function be more const-correct

Introduce const-ness on parameters, they are used as read-only and should not be
modified.  NFC.

llvm-svn: 212913

10 years agoMC: make DWARF and Windows unwinding handling more similar
Saleem Abdulrasool [Sun, 13 Jul 2014 19:03:36 +0000 (19:03 +0000)]
MC: make DWARF and Windows unwinding handling more similar

Rename member variables and functions for the MCStreamer for DWARF-like
unwinding management.  Rename the Windows ones as well and make the naming and
handling similar across the two.  No functional change intended.

llvm-svn: 212912

10 years agoRemove trailing spaces
Pavel Chupin [Sun, 13 Jul 2014 17:11:45 +0000 (17:11 +0000)]
Remove trailing spaces

llvm-svn: 212911

10 years agoAdd forgotten `break` statement.
Simon Atanasyan [Sun, 13 Jul 2014 16:18:56 +0000 (16:18 +0000)]
Add forgotten `break` statement.

llvm-svn: 212910

10 years agoAdd test cases for AArch64 hints codegen
Yi Kong [Sun, 13 Jul 2014 16:17:30 +0000 (16:17 +0000)]
Add test cases for AArch64 hints codegen

llvm-svn: 212909

10 years ago[Mips] Support SHT_MIPS_ABIFLAGS section type flag in the llvm-readobj,
Simon Atanasyan [Sun, 13 Jul 2014 15:28:54 +0000 (15:28 +0000)]
[Mips] Support SHT_MIPS_ABIFLAGS section type flag in the llvm-readobj,
obj2yaml and yaml2obj tools.

llvm-svn: 212908

10 years ago[CMake] Enable loadable modules, aka plugins, with BUILD_SHARED_LIBS on cygming.
NAKAMURA Takumi [Sun, 13 Jul 2014 13:47:37 +0000 (13:47 +0000)]
[CMake] Enable loadable modules, aka plugins, with BUILD_SHARED_LIBS on cygming.

Loadable modules could be enabled without BUILD_SHARED_LIBS with tweaks in future.

llvm-svn: 212907