The x86_mmx type is used for MMX intrinsics, parameters and
return values where these use MMX registers, and is also
supported in load, store, and bitcast.

Only the above operations generate MMX instructions, and optimizations
do not operate on or produce MMX intrinsics. 

MMX-sized vectors <2 x i32> etc. are lowered to XMM or split into
smaller pieces.  Optimizations may occur on these forms and the
result casted back to x86_mmx, provided the result feeds into a
previous existing x86_mmx operation.

The point of all this is prevent optimizations from introducing
MMX operations, which is unsafe due to the EMMS problem.

llvm-svn: 115243

conversion heuristics to the old-style ones.

llvm-svn: 115239

use MC instructions in the printInstruction() method via the tablegen flag
for it rather than a #define prior to including the autogenerated bits.

llvm-svn: 115238

llvm-svn: 115225

llvm-svn: 115203

llvm-svn: 115193

vs. ELF

llvm-svn: 115180

llvm-svn: 115176

llvm-svn: 115168

llvm-svn: 115160

llvm-svn: 115156

llvm-svn: 115154

for the dwarf .loc support to emit dwarf line number tables.

llvm-svn: 115153

llvm-svn: 115149

llvm-svn: 115147

llvm-svn: 115136

llvm-svn: 115135

With this patch in

movq    $foo, foo(%rip)
foo:
.long   foo

We produce a R_X86_64_32S for the first relocation and R_X86_64_32 for the
second one.

llvm-svn: 115134

Small test for sanity check of resulting ARM .s file.
Tested against -r115129.

llvm-svn: 115133

Now that the MC lowering handles the expansion of the pseudos, kill the horrible
blobs of text.

llvm-svn: 115130

I added a new file ARMAsmBackend which stubs out in similar ways to
the eqv X86 class.
For now, I split the ELFARMAsmBackend from the DarwinARMAsmBackend
(also mimicking X86)

Tested against -r115126

llvm-svn: 115129

to an empty PrintSpecial()

llvm-svn: 115128

(Kill the dead non-MC asm printer for the ARM target.)

llvm-svn: 115127

1. Cortex-a9 8-bit and 16-bit loads / stores AGU cycles are 1 cycle longer than 32-bit ones.
2. Cortex-a9 is out-of-order so model all read cycles as cycle 1.
3. Lots of other random fixes for A8 and A9.

llvm-svn: 115121

llvm-svn: 115116

for generic call handling.

llvm-svn: 115105

llvm-svn: 115100

pipeline forwarding path.

llvm-svn: 115098

a context.

llvm-svn: 115094

which require the use of the shifter-operand. This will be used to split
the ldr/str instructions such that those versions needing the shifter operand
can get a different scheduling itenerary, as in some cases, the use of the
shifter can cause different scheduling than the simpler forms.

llvm-svn: 115066

  lib/Target/X86/X86MCCodeEmitter.cpp: 190: error: suggest parentheses around '&&' within '||'

llvm-svn: 115064

llvm-svn: 115061

LDM/STM instructions can run one cycle faster on some ARM processors if the
memory address is 64-bit aligned.  Radar 8489376.

llvm-svn: 115047

llvm-svn: 115043

One Printer to lower them all and in the back end bind them.


(Remove option to use the old non-MC asm printer.)

llvm-svn: 115038

added some doxygen on the way

llvm-svn: 115033

llvm-svn: 115029

llvm-svn: 115027

order as cctools for diffability.

llvm-svn: 115022

are defined to emit the 0x67 prefix byte.  rdar://8482675

llvm-svn: 115021