[AArch64] Removed unnecessary copy patterns with v1fx types.
authorChad Rosier <mcrosier@codeaurora.org>
Thu, 12 Dec 2013 15:46:29 +0000 (15:46 +0000)
committerChad Rosier <mcrosier@codeaurora.org>
Thu, 12 Dec 2013 15:46:29 +0000 (15:46 +0000)
commit410ca67ab2b7ecef17d2ba71ef48ddcf2163c84a
tree6da70abedc6588ecd904fb9ab35d4d7744ecb80b
parent9191be9523e60ccc3f69fddaa5aa784acb66a170
[AArch64] Removed unnecessary copy patterns with v1fx types.

- Copy patterns with float/double types are enough.
- Fix typos in test case names that were using v1fx.
- There is no ACLE intrinsic that uses v1f32 type.  And there is no conflict of
  neon and non-neon ovelapped operations with this type, so there is no need to
  support operations with this type.
- Remove v1f32 from FPR32 register and disallow v1f32 as a legal type for
  operations.

Patch by Ana Pazos!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@197159 91177308-0d34-0410-b5e6-96231b3b80d8
lib/Target/AArch64/AArch64CallingConv.td
lib/Target/AArch64/AArch64ISelLowering.cpp
lib/Target/AArch64/AArch64InstrNEON.td
lib/Target/AArch64/AArch64RegisterInfo.td
test/CodeGen/AArch64/neon-bitcast.ll