generation to always catch the weird cases.

llvm-svn: 136453

llvm-svn: 136451

using vextractf128. This will reduce the number of issued instruction
for several avx codes.

llvm-svn: 136323

Take advantage that the 128-bit vpxor zeros the higher part and use it.
This also fixes PR10491

llvm-svn: 136321

a convert pattern close to the instruction definition.

llvm-svn: 136320

usage of the shuffle bitmask. Both work in 128-bit lanes without
crossing, but in the former the mask of the high part is the same
used by the low part while in the later both lanes have independent
masks. Handle this properly and and add support for vpermilpd.

llvm-svn: 136200

It is quiet possible that inlined function body is split into multiple chunks of consequtive instructions. But, there is not any way to describe this in .debug_inline accelerator table used by gdb. However, describe non contiguous ranges of inlined function body appropriately using AT_range of DW_TAG_inlined_subroutine debug info entry.

llvm-svn: 136196

These copies would coalesce easily, but the resulting value would be
defined by a deleted instruction. Now we also remove the undefined value
number from the destination register.

This fixes PR10503.

llvm-svn: 136174

llvm-svn: 136170

On x86 we can't encode an immediate LHS of a sub directly. If the RHS comes from a XOR with a constant we can
fold the negation into the xor and add one to the immediate of the sub. Then we can turn the sub into an add,
which can be commuted and encoded efficiently.

This code is generated for __builtin_clz and friends.

llvm-svn: 136167

different from the previous 128-bit because they work in lanes.
Update a few comments and add testcases

llvm-svn: 136157

Prevent x86-specific DAGCombine from creating nodes with illegal type (which could not be selected).  Fixes a minor isel issue that was breaking the testcase from r136130.

llvm-svn: 136148

llvm-svn: 136131

llvm-svn: 136130

llvm-svn: 136051

This also fixes PR10452

llvm-svn: 136004

shuffle before inserting on a 256-bit vector.
- Add AVX versions of movd/movq instructions
- Introduce a few COPY patterns to match insert_subvector instructions.
This turns a trivial insert_subvector instruction into a register copy,
coalescing the xmm into a ymm and avoid emiting on more instruction.

llvm-svn: 136002

llvm-svn: 135995

llvm-svn: 135993

Addresses PR10466, although the crash from that PR only triggers in cases where DAGCombine misses optimizing a shuffle.

llvm-svn: 135980

This fixes PR10463. A two-address instruction with an <undef> use
operand was incorrectly rewritten so the def and use no longer used the
same register, violating the tie constraint.

Fix this by always rewriting <undef> operands with the register a def
operand would use.

llvm-svn: 135885

llvm-svn: 135802

load folding logic

llvm-svn: 135801

too. Patch by Jeff Muizelaar.

llvm-svn: 135789

and was actually very wrong, fix it and make it simpler. Also remove the
ConcatVectors function, which is unused now.

- Fix a introduction of useless nodes in r126664 and r126264. The
VUNPCKL* should never be introduced cause we don't want duplicate
nodes for 128 AVX and non-AVX modes, the actual instruction
difference only exists during isel, but not for target specific DAG
nodes. We only introduce V* target nodes when there is no 128-bit
version already there.

- Fix a fragile test and make it more useful.

llvm-svn: 135729

llvm-svn: 135728

vxorps + vinsertf128 pair of instructions

llvm-svn: 135727

- Add more bitcasts for v16i16
- Since 135661 and 135662 already added the splat logic,
just add one more splat test for v16i16

llvm-svn: 135663

instruction introduced in AVX, which can operate on 128 and 256-bit vectors.
It considers a 256-bit vector as two independent 128-bit lanes. It can permute
any 32 or 64 elements inside a lane, and restricts the second lane to
have the same permutation of the first one. With the improved splat support
introduced early today, adding codegen for this instruction enable more
efficient 256-bit code:

Instead of:
  vextractf128  $0, %ymm0, %xmm0
  punpcklbw %xmm0, %xmm0
  punpckhbw %xmm0, %xmm0
  vinsertf128 $0, %xmm0, %ymm0, %ymm1
  vinsertf128 $1, %xmm0, %ymm1, %ymm0
  vextractf128  $1, %ymm0, %xmm1
  shufps  $1, %xmm1, %xmm1
  movss %xmm1, 28(%rsp)
  movss %xmm1, 24(%rsp)
  movss %xmm1, 20(%rsp)
  movss %xmm1, 16(%rsp)
  vextractf128  $0, %ymm0, %xmm0
  shufps  $1, %xmm0, %xmm0
  movss %xmm0, 12(%rsp)
  movss %xmm0, 8(%rsp)
  movss %xmm0, 4(%rsp)
  movss %xmm0, (%rsp)
  vmovaps (%rsp), %ymm0
We get:
  vextractf128  $0, %ymm0, %xmm0
  punpcklbw %xmm0, %xmm0
  punpckhbw %xmm0, %xmm0
  vinsertf128 $0, %xmm0, %ymm0, %ymm1
  vinsertf128 $1, %xmm0, %ymm1, %ymm0
  vpermilps $85, %ymm0, %ymm0

llvm-svn: 135662

While emitting constant value, look through derived type and use underlying basic type to determine size and signness of the constant value.

llvm-svn: 135627

llvm-svn: 135595

llvm-svn: 135562

llvm-svn: 135535

Revert r135423.

llvm-svn: 135454

During bottom up fast-isel, instructions emitted to materalize registers are at top of basic block and do not have debug location. This may misguide debugger while entering the basic block and sometimes debugger provides semi useful view of current location to developer by picking up previous known location as current location. Assign a sensible location to the first instruction in a basic block, if it does not have one location derived from source file, so that debugger can provide meaningful user experience to developers in edge cases.
[take 2]

llvm-svn: 135423

llvm-svn: 135404

llvm-svn: 135379

llvm-svn: 135332

1) Make non-legal 256-bit loads to be promoted to v4i64. This lets us
canonize the loads and handle things the same way we use to handle
for 128-bit registers. Despite of what one of the removed comments
explained, the load promotion would not mess with VPERM, it's only a
matter of doing the appropriate bitcasts when this instructions comes
to be introduced. Also make LOAD v8i32 legal.

2) Doing 1) exposed two bugs:
- v4i64 was being promoted to itself for several opcodes (introduced
in r124447 by David Greene) causing endless recursion and the stack to
explode.
- there was no support for allOnes BUILD_VECTORs and ANDNP would fail to
match because it was generating early target constant pools during
lowering.

3) The testcases are already checked-in, doing 1) exposed the
bugs in the current testcases.

4) Tidy up code to be more clear and explicit about AVX.

llvm-svn: 135313

when determining validity of matching constraint. Allow i1
types access to the GR8 reg class for x86.

Fixes PR10352 and rdar://9777108

llvm-svn: 135180