llvm-svn: 41809

llvm-svn: 41808

        cmpl    %eax, %ecx
        setae   %al
        movzbl  %al, %eax

instead of:

        cmpl    %eax, %ecx
        setb    %al
        xorb    $1, %al
        movzbl  %al, %eax

when using logical not of a C comparison.

llvm-svn: 41807

llvm-svn: 41806

2. Lower calls to fabs and friends to FABS nodes etc unless the function has
   internal linkage.  Before we wouldn't lower if it had a definition, which
   is incorrect.  This allows us to compile:

define double @fabs(double %f) {
        %tmp2 = tail call double @fabs( double %f )
        ret double %tmp2
}

into:

_fabs:
        fabs f1, f1
        blr 

llvm-svn: 41805

llvm-svn: 41804

instruction. These are implemented with xor which will modify the conditional
code. They should be rematerialized as move instructions.

llvm-svn: 41802

llvm-svn: 41800

llvm-svn: 41799

llvm-svn: 41798

llvm-svn: 41792

llvm-svn: 41790

llvm-svn: 41787

llvm-svn: 41786

optional.

llvm-svn: 41785

Fix for VisualStudio.  It is treating a 2 bit enum as a signed int for comparison purposes, causing failures.  Using an extra bit fixes it.

llvm-svn: 41784

first region, just deallocate all but the last region in the list.

llvm-svn: 41782

registers. The scheduler is now responsible for emitting them.

llvm-svn: 41781

llvm-svn: 41776

by flagging the associated instructions as being trivially rematerializable.

llvm-svn: 41775

llvm-svn: 41774

llvm-svn: 41769

The x86-64 ABI states that objects passed on the stack have
8 byte alignment. Implement that.

llvm-svn: 41768

Split eh.select / eh.typeid.for intrinsics into i32/i64 versions. This is needed, because they just "mark" register 
liveins and we let frontend solve type issue, not lowering code :)

llvm-svn: 41763

llvm-svn: 41762

llvm-svn: 41759

Patch by Sterling Stein!

llvm-svn: 41758

Constify to catch bugs.

llvm-svn: 41751

Use APFloat in UpgradeParser and AsmParser.
Change all references to ConstantFP to use the
APFloat interface rather than double.  Remove
the ConstantFP double interfaces.
Use APFloat functions for constant folding arithmetic
and comparisons.
(There are still way too many places APFloat is
just a wrapper around host float/double, but we're
getting there.)

llvm-svn: 41747

Also, fix a case, when target doesn't have weak symbols supported.

llvm-svn: 41746

Add instruction dump output.  This helps find bugs.

llvm-svn: 41744

Pluggable coalescers inplementation.

llvm-svn: 41743

llvm-svn: 41741

that would make the icmp true. Fixes PR1637.

llvm-svn: 41740

llvm-svn: 41739

leal    (,%rcx,8), %rcx
It should be
leal    (,%rcx,8), %ecx

llvm-svn: 41735

llvm-svn: 41733

llvm-svn: 41730

Use pool allocator for all the VNInfo's to improve memory access locality. This reduces coalescing time on siod Mac OS X PPC by 35%. Also remove the back ptr from VNInfo to LiveInterval and other tweaks.

llvm-svn: 41729

Added Reset() to free all allocated memory regions and reset state to be the same as right after ctor.

llvm-svn: 41728