When tryEvict() is looking for a cheaper register in the allocation
order, skip the tail of too expensive registers when possible.

llvm-svn: 172281

Remember the minimum cost of the registers in an allocation order and
the number of registers at the end of the allocation order that have the
same cost per use.

This information can be used to limit the search space for
RAGreedy::tryEvict() when looking for a cheaper register.

llvm-svn: 172280

llvm-svn: 172277

The MCAsmParser interface defines ParseIdentifier is public. There's no reason
whatsoever for AsmParser (which implements the MCAsmParser interface) to hide
this method.

This is all part of a bigger scheme. Several asm parsing "extensions" use the
main parser properly through the MCAsmParser interface. However,
GenericAsmParser has much more exclusive access and uses implementation details
from the concrete implementation - AsmParser, in which it is also declared as
a friend. This makes for overly coupled code, and even makes it hard to split
GenericAsmParser into a separate file. There's no reason why GenericAsmParser
shouldn't be able to access AsmParser through an abstract interface, as long
as it's actually registered as an extension.

llvm-svn: 172276

llvm-svn: 172271

llvm-svn: 172269

GenericAsmParser extension, where a lot of directives are already being parsed.
The end goal is having just a single place (and a single lookup table) for
all directive parsing.

llvm-svn: 172268

Adds a check for -Oz, changes the code to not re-visit BBs,
and skips over DBG_VALUE instrs.

Patch by Andy Zhang.

llvm-svn: 172258

llvm-svn: 172248

Specifically:
1. Added a missing new line when we emit a debug message saying that we are marking a global variable as constant.
2. Added debug messages that describe what is occuring when GlobalOpt is evaluating a block/function.
3. Added a debug message that says what specific constructor is being evaluated.

llvm-svn: 172247

This fixes some of the cycles between libCodeGen and libSelectionDAG. It's still
a complete mess but as long as the edges consist of virtual call it doesn't
cause breakage. BasicTTI did static calls and thus broke some build
configurations.

llvm-svn: 172246

the target if it supports the different CAST types. We didn't do this
on X86 because of the different register sizes and types, but on ARM
this makes sense.

llvm-svn: 172245

- recognize string "{memory}" in the MI generation
- mark as mayload/maystore when there's a memory clobber constraint.

PR14859.

Patch by Krzysztof Parzyszek

llvm-svn: 172228

llvm-svn: 172224

llvm-svn: 172223

llvm-svn: 172222

-dot-callgraph similarly follows a standard module pass pattern.

Patch by Speziale Ettore!

llvm-svn: 172220

Patch by Michael Muller.

llvm-svn: 172214

This removes previous special cases for each floating-point type in favour of a
shared codepath.

llvm-svn: 172189

ARM Cost Model:  We need to detect the max bitwidth of types in the loop in order to select the max vectorization factor.
We don't have a detailed analysis on which values are vectorized and which stay scalars in the vectorized loop so we use
another method. We look at reduction variables, loads and stores, which are the only ways to get information in and out
of loop iterations. If the data types are extended and truncated then the cost model will catch the cost of the vector
zext/sext/trunc operations.

llvm-svn: 172178

Revert r172153, "llvm/lib/MC/MCParser/AsmParser.cpp: [ms-inline-asm] Fix a couple of undefined behaviors. Operand->needAddressOf() is not initialized at !Operand->isReg()."

It has been redundant since r172157.

llvm-svn: 172166

Right now, only OS X has a way to determine the column width of a string
(PR14910). Until we have a good way to deal with this, we just won't
print carets, source ranges, or fixits for SMDiagnostic if the source line
has multibyte characters in it.

llvm-svn: 172164

llvm-svn: 172159

llvm-svn: 172157

llvm/lib/MC/MCParser/AsmParser.cpp: [ms-inline-asm] Fix a couple of undefined behaviors. Operand->needAddressOf() is not initialized at !Operand->isReg().

llvm-svn: 172153

llvm-svn: 172151

llvm-svn: 172149

r172121.

llvm-svn: 172148

llvm-svn: 172146

The root cause is mistakenly taking for granted that 
    "dyn_cast<Instruction>(a-Value)"
return a non-NULL instruction.

llvm-svn: 172145

Value's current type. The casting is trivial even for aggregate type.

llvm-svn: 172143

switch. Committed with Jim's and Chris's approval.

llvm-svn: 172136

This fixes va_start/va_copy of a va_list field which happens to not
be laid out at a 16-byte boundary.

Differential Revision: http://llvm-reviews.chandlerc.com/D276

llvm-svn: 172128

llvm-svn: 172124

llvm-svn: 172123

Part of rdar://12991541

llvm-svn: 172121

Teach InstCombine to hoist FABS and FNEG through FPTRUNC instructions.  The application of these operations commutes with the truncation, so we should prefer to do them in the smallest size we can, to save register space, use smaller constant pool entries, etc.

llvm-svn: 172117

Like Clang's FixItHint, SMFixIt represents an insertion, replacement, or
removal of source text. One or more fix-its can be emitted as part of
a diagnostic, and will be printed below the source range line to show the
user how they can fix their code.

Currently, the only client of SMFixIt is clang-tblgen; thus, the tests for
this behavior live in clang/test/TableGen/tg-fixits.td. If/when SMFixIt is
adopted within LLVM itself, those tests should be moved to the LLVM suite.

llvm-svn: 172086

def foo : bar;
          ~~~

This allows us to produce more precise diagnostics about a certain
superclass, and even provide fixits.

llvm-svn: 172085

llvm-svn: 172084