implementation today but is the right place if we want to make it faster some
day.

llvm-svn: 101521

llvm-svn: 101520

llvm-svn: 101516

llvm-svn: 101515

in the comment.  This was causing nondeterministic changes
in inlining decisions.

llvm-svn: 101503

llvm-svn: 101502

llvm-svn: 101501

llvm-svn: 101500

llvm-svn: 101499

llvm-svn: 101498

SemaOverload.cpp; no functionality change.

llvm-svn: 101497

that we aren't using ForceRValue any more? 

llvm-svn: 101496

llvm-svn: 101495

llvm-svn: 101494

functions.

llvm-svn: 101492

Eliminate the ForceRValue parameter from Sema::IsUserDefinedConversion. It's not the way we're going to handle this.

llvm-svn: 101483

Eliminate the default value for the UserCast parameter of Sema::IsUserDefinedConversion. No functionality change

llvm-svn: 101482

don't need it.

llvm-svn: 101481

llvm-svn: 101480

llvm-svn: 101478

llvm-svn: 101477

large files, this doesn't seem significantly better than just letting
raw_ostream pick a buffer size.

This code predates raw-ostream's automatic buffer sizing; in fact, it
was introduced as part of the code which would eventually become
raw_ostream.

llvm-svn: 101473

Make CGRecordLayoutBuilder deal with wide bit-fields. Will land tests shortly (Daniel, please review).

llvm-svn: 101472

dependent analyses, and increase code size, so doing it profitably would
require more complex heuristics.

llvm-svn: 101471

llvm-svn: 101470

callee is expected to be expanded to something else by codegen, so that
normal infinitely recursive calls are still transformed.

llvm-svn: 101468

llvm-svn: 101467

llvm-svn: 101466

with a fix for self-hosting

rotate CallInst operands, i.e. move callee to the back
of the operand array

the motivation for this patch are laid out in my mail to llvm-commits:
more efficient access to operands and callee, faster callgraph-construction,
smaller compiler binary

llvm-svn: 101465

llvm-svn: 101464

llvm-svn: 101463

llvm-svn: 101462

expression canonicalization. Its job is to print what's there, not to
make judgements about it.

llvm-svn: 101461

llvm-svn: 101459

llvm-svn: 101457

llvm-svn: 101455

JIT doesn't use the MC back-end asm printer to emit labels that it uses, the
section for the MCSymbol is never set. And thus the MCSymbol for the EH label
isn't marked as "defined". Because of that, TidyLandingPads removes the needed
landing pads from the JIT output. This breaks EH for every JIT program.

This is a work-around for this limitation. We pass in the label locations
map. If the label has a non-zero value, then it was "emitted" by the JIT and
TidyLandingPads shouldn't remove that label.

A nicer solution would be to mark the MCSymbol as "used" by the JIT and not rely
upon the section being set to determine if it's defined or not.

llvm-svn: 101453

platform that typically uses glibc. Fixes a Boost.Thread compilation
failure.

llvm-svn: 101450

llvm-svn: 101449

Adding support for dag combiner to promote operations for profit. This requires target specific queries. For example, x86 should promote i16 to i32 when it does not impact load folding.
x86 support is off by default. It can be enabled with -promote-16bit.

Work in progress.

llvm-svn: 101448