Code clean up.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120965 91177308-0d34-0410-b5e6-96231b3b80d8

Remove an unused variable.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120964 91177308-0d34-0410-b5e6-96231b3b80d8

Some cleanup before I start committing some incremental progress on
StrongPHIElimination.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120961 91177308-0d34-0410-b5e6-96231b3b80d8

Making use of VFP / NEON floating point multiply-accumulate / subtraction is
difficult on current ARM implementations for a few reasons.
1. Even though a single vmla has latency that is one cycle shorter than a pair
   of vmul + vadd, a RAW hazard during the first (4? on Cortex-a8) can cause
   additional pipeline stall. So it's frequently better to single codegen
   vmul + vadd.
2. A vmla folowed by a vmul, vmadd, or vsub causes the second fp instruction to
   stall for 4 cycles. We need to schedule them apart.
3. A vmla followed vmla is a special case. Obvious issuing back to back RAW
   vmla + vmla is very bad. But this isn't ideal either:
     vmul
     vadd
     vmla
   Instead, we want to expand the second vmla:
     vmla
     vmul
     vadd
   Even with the 4 cycle vmul stall, the second sequence is still 2 cycles
   faster.

Up to now, isel simply avoid codegen'ing fp vmla / vmls. This works well enough
but it isn't the optimial solution. This patch attempts to make it possible to
use vmla / vmls in cases where it is profitable.

A. Add missing isel predicates which cause vmla to be codegen'ed.
B. Make sure the fmul in (fadd (fmul)) has a single use. We don't want to
   compute a fmul and a fmla.
C. Add additional isel checks for vmla, avoid cases where vmla is feeding into
   fp instructions (except for the #3 exceptional case).
D. Add ARM hazard recognizer to model the vmla / vmls hazards.
E. Add a special pre-regalloc case to expand vmla / vmls when it's likely the
   vmla / vmls will trigger one of the special hazards.

Work in progress, only A+B are enabled.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120960 91177308-0d34-0410-b5e6-96231b3b80d8

Remove the PHIElimination.h header, as it is no longer needed.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120959 91177308-0d34-0410-b5e6-96231b3b80d8

Clarify some of the differences between indexing with getelementptr and indexing with insertvalue/extractvalue.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120957 91177308-0d34-0410-b5e6-96231b3b80d8

Fix PR 4170 by having ExtractValueInst::getIndexedType() reject out-of-bounds indexing.

Also add asserts that the indices are valid in InsertValueInst::init(). ExtractValueInst already asserts when constructed with invalid indices.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120956 91177308-0d34-0410-b5e6-96231b3b80d8

I forgot to actually remove the FindCopyInsertPoint() declaration from
PHIElimination.h.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120953 91177308-0d34-0410-b5e6-96231b3b80d8

Remove the SplitCriticalEdge() method declaration from PHIElimination.h. At one
time, this method existed, but now PHIElimination uses the method of the same
name on MachineBasicBlock.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120952 91177308-0d34-0410-b5e6-96231b3b80d8

Move the FindCopyInsertPoint method of PHIElimination to a new standalone
function so that it can be shared with StrongPHIElimination.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120951 91177308-0d34-0410-b5e6-96231b3b80d8

Refactor jump threading.
Should have no functional change other than the order of two transformations that are mutually-exclusive and the exact formatting of debug output.
Internally, it now stores the ConstantInt*s as Constant*s, and actual undef values instead of nulls.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120946 91177308-0d34-0410-b5e6-96231b3b80d8

Remove trailing whitespace.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120945 91177308-0d34-0410-b5e6-96231b3b80d8

Teach SimplifyCFG to turn
  (indirectbr (select cond, blockaddress(@fn, BlockA),
                            blockaddress(@fn, BlockB)))
into
  (br cond, BlockA, BlockB).

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120943 91177308-0d34-0410-b5e6-96231b3b80d8

Teach X86ISelLowering that the second result of X86ISD::UMUL is a flags
result.  This allows us to compile:

void *test12(long count) {
      return new int[count];
}

into:

test12:
movl $4, %ecx
movq %rdi, %rax
mulq %rcx
movq $-1, %rdi
cmovnoq %rax, %rdi
jmp __Znam                  ## TAILCALL

instead of:

test12:
movl $4, %ecx
movq %rdi, %rax
mulq %rcx
seto %cl
testb %cl, %cl
movq $-1, %rdi
cmoveq %rax, %rdi
jmp __Znam

Of course it would be even better if the regalloc inverted the cmov to 'cmovoq',
which would eliminate the need for the 'movq %rdi, %rax'.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120936 91177308-0d34-0410-b5e6-96231b3b80d8

it turns out that when ".with.overflow" intrinsics were added to the X86
backend that they were all implemented except umul.  This one fell back
to the default implementation that did a hi/lo multiply and compared the
top.  Fix this to check the overflow flag that the 'mul' instruction
sets, so we can avoid an explicit test.  Now we compile:

void *func(long count) {
      return new int[count];
}

into:

__Z4funcl:                              ## @_Z4funcl
movl $4, %ecx                ## encoding: [0xb9,0x04,0x00,0x00,0x00]
movq %rdi, %rax              ## encoding: [0x48,0x89,0xf8]
mulq %rcx                    ## encoding: [0x48,0xf7,0xe1]
seto %cl                     ## encoding: [0x0f,0x90,0xc1]
testb %cl, %cl                ## encoding: [0x84,0xc9]
movq $-1, %rdi               ## encoding: [0x48,0xc7,0xc7,0xff,0xff,0xff,0xff]
cmoveq %rax, %rdi              ## encoding: [0x48,0x0f,0x44,0xf8]
jmp __Znam                  ## TAILCALL

instead of:

__Z4funcl:                              ## @_Z4funcl
movl $4, %ecx                ## encoding: [0xb9,0x04,0x00,0x00,0x00]
movq %rdi, %rax              ## encoding: [0x48,0x89,0xf8]
mulq %rcx                    ## encoding: [0x48,0xf7,0xe1]
testq %rdx, %rdx              ## encoding: [0x48,0x85,0xd2]
movq $-1, %rdi               ## encoding: [0x48,0xc7,0xc7,0xff,0xff,0xff,0xff]
cmoveq %rax, %rdi              ## encoding: [0x48,0x0f,0x44,0xf8]
jmp __Znam                  ## TAILCALL

Other than the silly seto+test, this is using the o bit directly, so it's going in the right
direction.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120935 91177308-0d34-0410-b5e6-96231b3b80d8

fix the rest of the linux miscompares :)

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120933 91177308-0d34-0410-b5e6-96231b3b80d8

generalize the previous check to handle -1 on either side of the
select, inserting a not to compensate.  Add a missing isZero check
that I lost somehow.

This improves codegen of:

void *func(long count) {
      return new int[count];
}

from:

__Z4funcl:                              ## @_Z4funcl
movl $4, %ecx                ## encoding: [0xb9,0x04,0x00,0x00,0x00]
movq %rdi, %rax              ## encoding: [0x48,0x89,0xf8]
mulq %rcx                    ## encoding: [0x48,0xf7,0xe1]
testq %rdx, %rdx              ## encoding: [0x48,0x85,0xd2]
movq $-1, %rdi               ## encoding: [0x48,0xc7,0xc7,0xff,0xff,0xff,0xff]
cmoveq %rax, %rdi              ## encoding: [0x48,0x0f,0x44,0xf8]
jmp __Znam                  ## TAILCALL
                                        ## encoding: [0xeb,A]

to:

__Z4funcl:                              ## @_Z4funcl
movl $4, %ecx                ## encoding: [0xb9,0x04,0x00,0x00,0x00]
movq %rdi, %rax              ## encoding: [0x48,0x89,0xf8]
mulq %rcx                    ## encoding: [0x48,0xf7,0xe1]
cmpq $1, %rdx                ## encoding: [0x48,0x83,0xfa,0x01]
sbbq %rdi, %rdi              ## encoding: [0x48,0x19,0xff]
notq %rdi                    ## encoding: [0x48,0xf7,0xd7]
orq %rax, %rdi              ## encoding: [0x48,0x09,0xc7]
jmp __Znam                  ## TAILCALL
                                        ## encoding: [0xeb,A]

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120932 91177308-0d34-0410-b5e6-96231b3b80d8

relax this to handle linux defaulting to -static.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120930 91177308-0d34-0410-b5e6-96231b3b80d8

Improve an integer select optimization in two ways:

1. generalize
    (select (x == 0), -1, 0) -> (sign_bit (x - 1))
to:
    (select (x == 0), -1, y) -> (sign_bit (x - 1)) | y

2. Handle the identical pattern that happens with !=:
   (select (x != 0), y, -1) -> (sign_bit (x - 1)) | y

cmov is often high latency and can't fold immediates or
memory operands.  For example for (x == 0) ? -1 : 1, before
we got:

<  testb %sil, %sil
<  movl $-1, %ecx
<  movl $1, %eax
<  cmovel %ecx, %eax

now we get:

>  cmpb $1, %sil
>  sbbl %eax, %eax
>  orl $1, %eax

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120929 91177308-0d34-0410-b5e6-96231b3b80d8

merge some tests into select.ll and make them more specific.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120928 91177308-0d34-0410-b5e6-96231b3b80d8

rename test

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120927 91177308-0d34-0410-b5e6-96231b3b80d8

remove two tests that aren't really testing anything.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120926 91177308-0d34-0410-b5e6-96231b3b80d8

Initialize HasPOPCNT.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120923 91177308-0d34-0410-b5e6-96231b3b80d8

Once the layout is done we don't need to keep updating which fragments are
valid. Addresses will not change.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120921 91177308-0d34-0410-b5e6-96231b3b80d8

Remember the contents of leb and dwarfline fragments when relaxing. This avoids
having to evaluate the expression again when writing.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120920 91177308-0d34-0410-b5e6-96231b3b80d8

Remove PHIElimination's private copy of SkipPHIsAndLabels.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120918 91177308-0d34-0410-b5e6-96231b3b80d8

Add patterns for the x86 popcnt instruction.

- Also adds a new POPCNT subtarget feature that is currently enabled if the target
  supports SSE4.2 (nehalem) or SSE4A (barcelona).

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120917 91177308-0d34-0410-b5e6-96231b3b80d8

Silence 'may be used uninitialized in this function' warnings. Static analysis
may determine that they cannot be used uninitialized. But that might be a bit
too much for the compiler to determine.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120916 91177308-0d34-0410-b5e6-96231b3b80d8

Support/PathV2: Remove redundant calls to make_error_code.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120913 91177308-0d34-0410-b5e6-96231b3b80d8

APInt: microoptimize a few methods.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120912 91177308-0d34-0410-b5e6-96231b3b80d8

Simplify APInt::getAllOnesValue.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120911 91177308-0d34-0410-b5e6-96231b3b80d8

Remove unneeded zero arrays.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120910 91177308-0d34-0410-b5e6-96231b3b80d8

Apparently APFloat::getZero doesn't like PPCDoubleDoubles.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120909 91177308-0d34-0410-b5e6-96231b3b80d8

Disable C++ exception handling on MSVC.
Total size of bin\Release on disk goes from 82.9 MB to 74.2 MB. (~10% saving)

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120908 91177308-0d34-0410-b5e6-96231b3b80d8

Simplify code. No functionality change.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120907 91177308-0d34-0410-b5e6-96231b3b80d8

Disable RTTI on Windows.
Total size of bin\Release on disk goes from 83.6 MB to 81.8MB. (~2% saving)

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120901 91177308-0d34-0410-b5e6-96231b3b80d8

The Thumb tADDrSPi instruction is not valid when the destination is SP.
Check for that and try narrowing it to tADDspi instead.  Radar 8724703.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120892 91177308-0d34-0410-b5e6-96231b3b80d8

Remove trailing whitespace.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120891 91177308-0d34-0410-b5e6-96231b3b80d8

There are two reasons why we might want to use

foo = a - b
.long foo
instead of just
.long a - b

First, on darwin9 64 bits the assembler produces the wrong result. Second,
if "a" is the end of the section all darwin assemblers (9, 10 and mc) will not
consider a - b to be a constant but will if the dummy foo is created.

Split how we handle these cases. The first one is something MC should take care
of. The second one has to be handled by the caller.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120889 91177308-0d34-0410-b5e6-96231b3b80d8

Unittests/Support/PathV2: Add FileSystem tests.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120888 91177308-0d34-0410-b5e6-96231b3b80d8

Support/FileSystem: Add status implementation.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120870 91177308-0d34-0410-b5e6-96231b3b80d8

Support/SystemError: Make error_category and error_code auto-bool-conversion-safe.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120869 91177308-0d34-0410-b5e6-96231b3b80d8

Support/Windows/FileSystem: Fix MinGW warnings.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120868 91177308-0d34-0410-b5e6-96231b3b80d8

Support/FileSystem: Add file_size implementation.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120867 91177308-0d34-0410-b5e6-96231b3b80d8

Next step: Only pad debug_line when the target is darwin. Add a FIXME to avoid
doing that if the target is darwin10 or newer.

This fixes
*) Direct object emission was producing objects without the workaround on
   darwin9.
*) Assembly printing was producing objects with the workaround on linux.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120866 91177308-0d34-0410-b5e6-96231b3b80d8

Encode condition code for Thumb1 conditional branch instruction.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120865 91177308-0d34-0410-b5e6-96231b3b80d8

Do not try luck by using given name to create temporary file. In parallel builds it may not work.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120860 91177308-0d34-0410-b5e6-96231b3b80d8

Correctly size-reduce the t2CMPzrr instruction to tCMPzr when possible.
tCMPzhir has undefined behavior when both source registers are low registers.
rdar://8728577

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120858 91177308-0d34-0410-b5e6-96231b3b80d8

Use correct variable names to match the patterns.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120857 91177308-0d34-0410-b5e6-96231b3b80d8

Also inore '()' while creating mdnode name from ObjC symbol name.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120856 91177308-0d34-0410-b5e6-96231b3b80d8

First step in fixing MC. Make it clear that we are avoiding a bug in the
darwin9 linker, what is needed to avoid it and where to get more information.

Also make the workaround simpler. Just the regular end_sequence we normally
create is more than 4 bytes.

Tested by building cctools and ld64 from darwin9 on a darwin10 system and using
those. I checked that I was able to reproduce the bootstrap failure when
the the workaround was disabled.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120854 91177308-0d34-0410-b5e6-96231b3b80d8

Ignore '+' while creating mdnode name from ObjC symbol name.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120853 91177308-0d34-0410-b5e6-96231b3b80d8

Match pattern operand names to expected encoding field names. This corrects the
operand encoding ordering of the instruction.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120852 91177308-0d34-0410-b5e6-96231b3b80d8

Remove incorrect BL target encoding (it's similar to, but not the same as the
ARM instruction). Add encoding of bits 13 and 11.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120849 91177308-0d34-0410-b5e6-96231b3b80d8

Encode the 32-bit wide Thumb (and Thumb2) instructions with the high order
halfword being emitted to the stream first. rdar://8728174

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120848 91177308-0d34-0410-b5e6-96231b3b80d8

Revert this change since it breaks a couple of the AVX tests.
I'm unclear if the tests are actually correct or not, but reverting for now.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120847 91177308-0d34-0410-b5e6-96231b3b80d8

Rename virtRegMap to avoid confusion with the VirtRegMap that it isn't.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120846 91177308-0d34-0410-b5e6-96231b3b80d8

Coalesce debug locations when possible, causing less DBG_VALUE instructions to
be emitted.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120845 91177308-0d34-0410-b5e6-96231b3b80d8

Scalar f32/f64 are also subregs of ymm regs

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120844 91177308-0d34-0410-b5e6-96231b3b80d8

Remove SSE1-4 disable when AVX is enabled.  While this may be useful for development,
it completely breaks scalar fp in xmm regs when AVX is enabled.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120843 91177308-0d34-0410-b5e6-96231b3b80d8

Emit DBG_VALUE instructions from LiveDebugVariables.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120842 91177308-0d34-0410-b5e6-96231b3b80d8

Also update virtRegMap when renaming virtual registers.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120841 91177308-0d34-0410-b5e6-96231b3b80d8

When using the 'push' mnemonic for Thumb2 stmdb, be explicit when it's the
32-bit wide version by adding the .w suffix.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120838 91177308-0d34-0410-b5e6-96231b3b80d8

Remove unused variable.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120836 91177308-0d34-0410-b5e6-96231b3b80d8

Reduce t2 ldr/str instructions to the correct t1 versions when there's an
immediate offset.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120833 91177308-0d34-0410-b5e6-96231b3b80d8

fix ARM::fixup_arm_branch, cleanup, and share more code between ELF and Darwin

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120832 91177308-0d34-0410-b5e6-96231b3b80d8

No need to declare EncoderMethod property anymore; just assign to it.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120831 91177308-0d34-0410-b5e6-96231b3b80d8

Delete the StrongPHIElimination pass, leaving only a shell.

The StrongPHIElimination pass did not work, and nobody has worked on it for two
years.

A rewrite is underway, so I am leaving this shell pass instead of deleting it
completely.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120830 91177308-0d34-0410-b5e6-96231b3b80d8

Add IntervalMap::iterator::set{Start,Stop,Value} methods that allow limited
editing of the current interval.

These methods may cause coalescing, there are corresponding set*Unchecked
methods for editing without coalescing. The non-coalescing methods are useful
for applying monotonic transforms to all keys or values in a map without
accidentally coalescing transformed and untransformed intervals.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120829 91177308-0d34-0410-b5e6-96231b3b80d8

Support/FileSystem: Add equivalent implementation.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120827 91177308-0d34-0410-b5e6-96231b3b80d8

Support/FileSystem: Fix MinGW build. It doesn't have _chsize_s.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120826 91177308-0d34-0410-b5e6-96231b3b80d8

Add FIXMEs.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120824 91177308-0d34-0410-b5e6-96231b3b80d8

Size reduction for tPUSH come from t2STMDB_UPD, not t2STMIA_UPD.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120822 91177308-0d34-0410-b5e6-96231b3b80d8

And I really hate line endings.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120821 91177308-0d34-0410-b5e6-96231b3b80d8

Support/Windows/FileSystem: Fix MinGW build.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120820 91177308-0d34-0410-b5e6-96231b3b80d8

Support/FileSystem: Add resize_file implementation.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120819 91177308-0d34-0410-b5e6-96231b3b80d8

Support/FileSystem: Add rename implementation.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120818 91177308-0d34-0410-b5e6-96231b3b80d8

Support/FileSystem: Add remove implementation.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120817 91177308-0d34-0410-b5e6-96231b3b80d8

Fix line endings.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120816 91177308-0d34-0410-b5e6-96231b3b80d8

Get Neon intrinsic names from the new "Name" field in the tblgen records
instead of just converting the record name to lowercase.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120809 91177308-0d34-0410-b5e6-96231b3b80d8

Fix paste-o in ExtractValueInst::getIndexedType() comments.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120804 91177308-0d34-0410-b5e6-96231b3b80d8

unittests/Support/PathV2: remove(3) requires a terminated string.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120803 91177308-0d34-0410-b5e6-96231b3b80d8

Apparently OS X 10.4 doesn't have __crashreporter_info__.

Try to fix building on the wayback machine.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120801 91177308-0d34-0410-b5e6-96231b3b80d8

Support/FileSystem: Add create_symlink implementation.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120800 91177308-0d34-0410-b5e6-96231b3b80d8

Support/FileSystem: Add create_hard_link implementation.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120792 91177308-0d34-0410-b5e6-96231b3b80d8

Support/ADT/Twine: Make toNullTerminatedStringRef not rely on UB :(.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120791 91177308-0d34-0410-b5e6-96231b3b80d8

Support/FileSystem: Add create_director{y,ies} implementations.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120790 91177308-0d34-0410-b5e6-96231b3b80d8

Make EmitIntValue more efficient and more like what we do for leb128. The
difference is much smaller (about 0.3s) but significant.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120787 91177308-0d34-0410-b5e6-96231b3b80d8

I did it wrong. Don't disregard these encodings here.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120786 91177308-0d34-0410-b5e6-96231b3b80d8

Unittests/Support/PathV2: Cleanup and remove output.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120785 91177308-0d34-0410-b5e6-96231b3b80d8

unittests/Support/PathV2: Comment out test because some systems are saying that
a file exists when it shouldn't.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120784 91177308-0d34-0410-b5e6-96231b3b80d8

Don't overwrite the opcode passed into the T1Special pattern.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120782 91177308-0d34-0410-b5e6-96231b3b80d8

Add Thumb encoding for some more instructions.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120780 91177308-0d34-0410-b5e6-96231b3b80d8

Ignore decode table conflicts in the tMOVgpr2tgpr, tMOVgpr2gpr, and tMOVtgpr2gpr
instructions. They are handled as special moves, but encoded as a normal move.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120779 91177308-0d34-0410-b5e6-96231b3b80d8

Support/Windows/FileSystem: Remove unneeded toNullTerminatedStringRef.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120777 91177308-0d34-0410-b5e6-96231b3b80d8

Support/FileSystem: Add unique_file and exists implementations.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120776 91177308-0d34-0410-b5e6-96231b3b80d8

Support/FileSystem: Remove temp_directory_path.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120775 91177308-0d34-0410-b5e6-96231b3b80d8

Do with uleb the same trick we now do with dwarf line/address advances. This
avoids creating leb128 fragments and speeds up the test in PR8711 to 33s.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120774 91177308-0d34-0410-b5e6-96231b3b80d8

clarify comment

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120772 91177308-0d34-0410-b5e6-96231b3b80d8

Try to resolve symbol differences early, and if successful create a plain
data fragment. This reduces the time to assemble the test in 8711 from 60s to
54s.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@120767 91177308-0d34-0410-b5e6-96231b3b80d8