llvm-svn: 93540

Add some debug routines to SelectionDAG to dump full DAGs.
print/dumpWithDepth allows one to dump a DAG up to N levels deep.
dump/printWithFullDepth prints the whole DAG, subject to a depth limit
on 100 in the default case (to prevent infinite recursion).

Have CannotYetSelect to a dumpWithFullDepth so it is clearer exactly
what the non-matching DAG looks like.

llvm-svn: 93538

Improve llvm.dbg.declare intrinsic by referring directly to the storage in its first argument, via function-local metadata (instead of via a bitcast).
This patch also cleans up code that expects there to be a bitcast in the first argument and testcases that call llvm.dbg.declare.
It also strips old llvm.dbg.declare intrinsics that did not pass metadata as the first argument.

llvm-svn: 93531

Remove most of old Mach-O Writer support, it has been replaced by MCMachOStreamer

Further refactoring to completely remove MachOWriter and drive the object file
writer with the AsmPrinter MCInst/MCSection logic is forthcoming.

llvm-svn: 93527

llvm-svn: 93515

The pre-regalloc pass caused some regressions in both compile time and
performance of the generated code, and it did not improve performance, except
for indirect branches.  I also moved the check for single-block loops to speed
up the common case when running the taildup pass before reg allocation.

llvm-svn: 93505

Improve llvm.dbg.declare intrinsic by referring directly to the storage in its first argument, via function-local metadata (instead of via a bitcast).
This patch also cleans up code that expects there to be a bitcast in the first argument and testcases that call llvm.dbg.declare.

llvm-svn: 93504

llvm-svn: 93494

llvm-svn: 93487

llvm-svn: 93484

We need to find a better way to emit this info.

llvm-svn: 93481

llvm-svn: 93463

llvm-svn: 93459

catch info can get misplaced when a selector ends up more than one block
removed from the parent invoke(s). This could happen when a landing pad is
shared by multiple invokes and is also a target of a normal edge from
elsewhere.

llvm-svn: 93456

llvm-svn: 93417

Commit some changes I had managed to lose last night while refactoring the code. Avoid change use of PHI instructions because it's not legal to insert any instructions before them.
This fixes PR6027.

llvm-svn: 93335

llvm-svn: 93313

llvm-svn: 93306

different BlockAddress labels, but nothing semantically important.

Add a FIXME that BlockAddress codegen is broken if the LLVM BB has 
an empty name (e.g. strip was run).

llvm-svn: 93303

Among other things, this would do very weird things if the 
basic block name had (e.g.) a space in it on darwin:
makeNameProper would add quotes, then the mcsymbol would 
escape the quotes.

llvm-svn: 93302

instead of returning it in an std::string.  Based on this change:

1. Change TargetLoweringObjectFileCOFF::getCOFFSection to take a StringRef
2. Change a bunch of targets to call makeNameProper with a smallstring,
   making several of them *much* more efficient.
3. Rewrite Mangler::makeNameProper to not build names and then prepend
   prefixes, not use temporary std::strings, and to avoid other crimes.

llvm-svn: 93298

llvm-svn: 93293

llvm-svn: 93292

integers on 64-bit systems.

llvm-svn: 93291

llvm-svn: 93290

llvm-svn: 93286

llvm-svn: 93285

llvm-svn: 93283

Add a quick pass to optimize sign / zero extension instructions. For targets where the pre-extension values are available in the subreg of the result of the extension, replace the uses of the pre-extension value with the result + extract_subreg.

For now, this pass is fairly conservative. It only perform the replacement when both the pre- and post- extension values are used in the block. It will miss cases where the post-extension values are live, but not used.

llvm-svn: 93278

dumper doesn't really do what I want yet, but
at least it doesn't crash now.

llvm-svn: 93272

llvm-svn: 93261

llvm-svn: 93228

llvm-svn: 93227

This fixes PR5980.

llvm-svn: 93184

loads are not in the default address space because the transformation discards src value info.

llvm-svn: 93180

llvm-svn: 93165

really does need to be a vector type, because
TargetLowering::getOperationAction for SIGN_EXTEND_INREG uses that type,
and it needs to be able to distinguish between vectors and scalars.

Also, fix some more issues with legalization of vector casts.

llvm-svn: 93043

Dan pointed out checking whether a node is dead by comparing its opcode to ISD::DELETED_NODE is not safe. Use a DAGUpdateListener to remove dead nodes from work list instead.

llvm-svn: 93031

ReplaceAllUsesOfValueWith may delete other nodes that the one being replaced. Do not delete dead nodes again.

llvm-svn: 92988

When folding a and(any_ext(load)) both the any_ext and the
load have to have only a single use.

This removes the anyext-uses.ll testcase which started failing
because it is unreduced and unclear what it is testing.

llvm-svn: 92950