[PATCH] D28697: [AArch64] Generate literals by the little end

[Evandro Menezes via Phabricator via llvm-commits]

evandro created this revision.
evandro added reviewers: t.p.northover, jmolloy.
evandro added subscribers: llvm-commits, MatzeB, mssimpso, mcrosier, ab.
evandro set the repository for this revision to rL LLVM.
Herald added subscribers: rengolin, aemerson.

ARM seems to prefer that long literals be formed from their little end.

This also promotes fusion of literal generation using the instrs pairs `MOV`/`MOVK` and `MOVK`/`MOVK` on Cortex A57 and others.


Repository:
  rL LLVM

https://reviews.llvm.org/D28697

Files:
  llvm/lib/Target/AArch64/AArch64ExpandPseudoInsts.cpp
  llvm/lib/Target/AArch64/AArch64InstrInfo.cpp
  llvm/lib/Target/AArch64/AArch64InstrInfo.td
  llvm/test/CodeGen/AArch64/arm64-addrmode.ll
  llvm/test/CodeGen/AArch64/arm64-bitfield-extract.ll
  llvm/test/CodeGen/AArch64/arm64-blockaddress.ll
  llvm/test/CodeGen/AArch64/arm64-code-model-large-abs.ll
  llvm/test/CodeGen/AArch64/arm64-const-addr.ll
  llvm/test/CodeGen/AArch64/arm64-extern-weak.ll
  llvm/test/CodeGen/AArch64/arm64-fast-isel-addr-offset.ll
  llvm/test/CodeGen/AArch64/arm64-movi.ll
  llvm/test/CodeGen/AArch64/bitfield-insert.ll
  llvm/test/CodeGen/AArch64/blockaddress.ll
  llvm/test/CodeGen/AArch64/code-model-large-abs.ll
  llvm/test/CodeGen/AArch64/extern-weak.ll
  llvm/test/CodeGen/AArch64/fpimm.ll
  llvm/test/CodeGen/AArch64/jump-table.ll
  llvm/test/CodeGen/AArch64/large-consts.ll
  llvm/test/CodeGen/AArch64/literal_pools_float.ll
  llvm/test/CodeGen/AArch64/movw-shift-encoding.ll
  llvm/test/CodeGen/AArch64/stack_guard_remat.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D28697.84364.patch
Type: text/x-patch
Size: 28566 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20170113/d53c1afe/attachment-0001.bin>