in the backend. Introduce a new SDNode type, MemOperandSDNode, for
holding a MemOperand in the SelectionDAG IR, and add a MemOperand
list to MachineInstr, and code to manage them. Remove the offset
field from SrcValueSDNode; uses of SrcValueSDNode that were using
it are all all using MemOperandSDNode now.

Also, begin updating some getLoad and getStore calls to use the
PseudoSourceValue objects.

Most of this was written by Florian Brander, some
reorganization and updating to TOT by me.

llvm-svn: 46585

Improved pretty-printing of LValues.
Parameters whose values are pointers/references are now assigned SymbolicLValues.

llvm-svn: 46584

Hack Sema::MergeTypeDefDecl() to silently ignore duplicate typedef's in system headers files.

A bizarre, non-standard hook that many compilers appear to implement (sigh:-).

llvm-svn: 46583

worklist until the dstor of GRBranchNodeBuilderImpl. This way clients can mark
creates nodes as "sinks" before they are added to the worklist.

llvm-svn: 46582

constant integers.

llvm-svn: 46581

pretty-print a block's terminator.

When building CFGs, for IfStmts ('if'), we no longer add the ParenExpr
that is the subexpression of the IfStmt to the CFG; instead we add its
first descendant subexpression that is not a ParenExpr.

llvm-svn: 46580

Fixed a bug in MergeValueInAsValue() pointed out by David Greene. Replace val# with previous liverange's.

llvm-svn: 46579

Fix test case and add a FIXME.

llvm-svn: 46577

llvm-svn: 46576

llvm-svn: 46575

On an unexpected @foo keyword, return failure instead of a 'successful' null AST.  This fixes a segfault
on things like @foo.

llvm-svn: 46574

llvm-svn: 46572

Treat the label for the first @llvm.dbg.stoppoint the same way as the dbg_func_start label. Make sure nothing else is inserted before them.
Note this solution might be somewhat fragile since ISD::LABEL may be used for other
purposes. If that ends up to be an issue, we may need to introduce a different node
for debug labels.

llvm-svn: 46571

llvm-svn: 46570

llvm-svn: 46569

A semi-gross fix for a debug info issue. When inserting the "function start" label (i.e. first label in the entry block) take care to insert it at the beginning of the block.

llvm-svn: 46568

llvm-svn: 46567

Abolish a bunch of Xcode inserted tabs.

llvm-svn: 46566

legitimate way of representing global variable GV in debug info.

llvm-svn: 46565

more cleanups with pretty-printing of analysis results.

llvm-svn: 46564

llvm-svn: 46563

Even though InsertAtEndOfBasicBlock is an ugly hack it still deserves a proper name. Rename it to EmitInstrWithCustomInserter since it does not necessarily insert
instruction at the end.

llvm-svn: 46562

llvm-svn: 46561

llvm-svn: 46560

one argument.  Lets stuff using pthread.h compile.

llvm-svn: 46559

llvm-svn: 46558

that causes this isn't really correct, but if we're going to accept 
this, it should come up with a consistent AST.

llvm-svn: 46557

memory reference information in the backend. Most of this was written by
Florian Brander, cleanup and updating to TOT by me.

llvm-svn: 46556

list just to see if whether the list is empty.

llvm-svn: 46555

llvm-svn: 46554

llvm-svn: 46553

when in a register.

llvm-svn: 46552

llvm-svn: 46551

llvm-svn: 46550

know if all of this is exactly right, but it seems like the right thing to
aim for.  Steve, please review.

llvm-svn: 46549

llvm-svn: 46548

llvm-svn: 46547

llvm-svn: 46546

llvm-svn: 46545

- Expand tabs... (poss 80-col violations, will get them later...)
- Consolidate logic for SelectDFormAddr and SelectDForm2Addr into a single
  function, simplifying maintenance. Also reduced custom instruction
  generation for SPUvecinsert/INSERT_MASK.

llvm-svn: 46544