llvm-svn: 107831

address calculation instructions leading up to a jump table when we're trying
to convert them into a TB[H] instruction in Thumb2. This realistically
shouldn't happen much, if at all, for well formed inputs, but it's more correct
to handle it. rdar://7387682

llvm-svn: 107830

llvm-svn: 107829

newly-narrowed scope. No functionality change.

llvm-svn: 107828

CXXConstructExpr/CXXTemporaryObjectExpr/CXXNewExpr as
appropriate. Fixes PR7556, and provides a slide codegen improvement
when copy-initializing a POD class type from a value-initialized
temporary. Previously, we weren't eliding the copy.

llvm-svn: 107827

llvm-svn: 107826

in the integrated assembler.  Still some discussion to be
done.

llvm-svn: 107825

llvm-svn: 107824

llvm-svn: 107823

Keep track of abstract subprogram DIEs. 

llvm-svn: 107822

llvm-svn: 107820

llvm-svn: 107818

llvm-svn: 107817

VisitFooTypeLoc() calls VisitFooType(); now, TraverseFooTypeLoc()
calls WalkUpFromFooType(). This allows clients that override
WalkUpFromFooType() to continue to work.  It also preserves the
property that Visit*() in the base visitor class is a no-op (s.t. a
subclass doesn't have to call Base::Visit*() when overriding
Visit*()).

Also fixes some typos in comments.
         
Also added a missing getDerived() inside
TraverseQualifiedTypeLoc(). The call is needed in case a subclass
overrides TraverseTypeLoc().

Reviewed by nlewycky and csilvers.

llvm-svn: 107816

didn't create a new block, don't reset the insert position.

llvm-svn: 107813

llvm-svn: 107812

llvm-svn: 107811

llvm-svn: 107810

This fixes PR7540.

llvm-svn: 107809

llvm-svn: 107808

llvm-svn: 107807

a separate DCE pass over MachineInstrs.

llvm-svn: 107804

First loop test passed.  The data structure and search algorithm is still crude and in-flux.  But this milestone needed to be locked in.  Right now every loop is implemented in terms of a structure that will handle the most complicated {min, max} loop.  Though only *-loops are tested at the moment.  In a future iteration *-loops will likely be optimized a little more.  The only tests are for basic posix so far, but I have prototype code running for extended posix and ecma.  The prototype code lacks the complicating properties of the real <regex> requirements though.

llvm-svn: 107803

llvm-svn: 107801

a bunch of stuff, to allow the target-independent calling convention
logic to be employed.

llvm-svn: 107800

llvm-svn: 107798

position when emitting multiple blocks when executing a custom
inserter.

llvm-svn: 107797

llvm-svn: 107796

llvm-svn: 107795

llvm-svn: 107794

llvm-svn: 107793

and not just for Debug+Asserts builds.

llvm-svn: 107792

around everywhere, and also give it an InsertPt member, to enable isel
to operate at an arbitrary position within a block, rather than just
appending to a block.

llvm-svn: 107791

llvm-svn: 107790

instance, rather than pointers to all of FunctionLoweringInfo's
members.

This eliminates an NDEBUG ABI sensitivity.

llvm-svn: 107789

from PR7583

llvm-svn: 107788

allow target-specific fast-isel code to make use of it directly.

llvm-svn: 107787

code can do calling-convention queries. This obviates OutputArgReg.

llvm-svn: 107786

llvm-svn: 107785

llvm-svn: 107784