llvm-svn: 34435

llvm-svn: 34434

uses them.

llvm-svn: 34432

rules alignment is to pick the alignment that corresponds to the smallest
specified alignment that is larger than the bit width of the type or the
largest specified integer alignment if none are larger than the bitwidth
of the type. For the byte size, the size returned is the next larger
multiple of the alignment for that type (using the above rule). This patch
also changes bit widths from "short" to "uint32_t" to ensure there are
enough bits to specify any bit width that LLVM can handle (currently 2^23);
16-bits isn't enough.

llvm-svn: 34431

llvm-svn: 34428

llvm-svn: 34425

llvm-svn: 34424

llvm-svn: 34423

llvm-svn: 34422

llvm-svn: 34419

llvm-svn: 34417

Revert patches that caused the problem. Evan, please investigate and reapply
when you've discovered the problem.

llvm-svn: 34399

2. Implement toString for power-of-2 radix without using divide and always
   printing full words. This allows hex/binary to look at the bit
   respresentation of the APInt as well as avoid bugs in divide.

llvm-svn: 34396

llvm-svn: 34395

llvm-svn: 34394

2. Teach slt about signedness.
3. Teach roundToDouble to properly sign extend.

llvm-svn: 34391

2. Consolidate memory allocation into just two inline functions.
3. Convert "unsigned" to uint32_t to gaurantee its size.
4. Eliminate magic constants and replace with symbolic equivalent.
5. Improve code documentation slightly.
6. Simplify the logical operator code because bitwidths must be the same.
7. Fix indentation per coding standards.
8. Use exit-early style to reduce indentation in several functions.

llvm-svn: 34389

Fix line breaks and 80 cols violation.
Simplify operator^= since bitwidths must be the same.

llvm-svn: 34388

Fix bugs in countLeadingZeros and countTrailingZeros.

llvm-svn: 34386

compile-time regression in LLC.

llvm-svn: 34385

llvm-svn: 34384

- When coalescing a copy MI, if its destination is "dead", propagate the
  property to the source MI's destination if there are no intervening uses.
- Detect dead function live-in's and remove them.

llvm-svn: 34383

llvm-svn: 34382

llvm-svn: 34381

llvm-svn: 34380

- Allow LiveVariables to track liveness of more registers.

llvm-svn: 34379

llvm-svn: 34376

_test2:
        li r2, 0
        lis r3, 1
        std r2, 9024(r3)
        blr

instead of:

_test2:
        lis r2, 1
        li r3, 0
        ori r2, r2, 9024
        std r3, 0(r2)
        blr

This implements CodeGen/PowerPC/LargeAbsoluteAddr.ll:test2

llvm-svn: 34373

_test:
        lis r2, 743
        li r3, 0
        stw r3, 32751(r2)
        blr

instead of:

_test:
        li r2, 0
        stw r2, 32751(48693248)
        blr

Implement support for ppc64 as well, allowing it to produce better code.

llvm-svn: 34371

llvm-svn: 34369

test/CodeGen/PowerPC/2007-02-16-InlineAsmNConstraint.ll

llvm-svn: 34368

llvm-svn: 34364

llvm-svn: 34363

meaningful variable names. Add comments to document the flow.

llvm-svn: 34362

llvm-svn: 34359

This fixes the UnitTests/Vector/sumarray-dbl regressions.

llvm-svn: 34358

llvm-svn: 34357

llvm-svn: 34356

Patch by Scott Michel.

llvm-svn: 34354

1. Function style changes.
2. 80-col violations.
3. Better names for things.
4. Arrange constructors so they all take bit width first.
5. Add named signed and unsigned comparison functions and remove the
   corresponding operators.
6. Remove operator&& and operator|| but provide a getBoolValue function which
   converts to bool as comparison against 0. This allows the normal && and
   || operators to be used as if (X.getBoolValue() && Y.getBoolValue())

Note: this still doesn't function 100% yet. I'm working on the bugs now.
llvm-svn: 34353