PPC: Optimize rldicl generation for masked shifts
Masking operations (where only some number of the low bits are being kept) are
selected to rldicl(x, 0, mb). If x is a logical right shift (which would become
rldicl(y, 64-n, n)), we might be able to fold the two instructions together:
rldicl(rldicl(x, 64-n, n), 0, mb) -> rldicl(x, 64-n, mb) for n <= mb
The right shift is really a left rotate followed by a mask, and if the explicit
mask is a more-restrictive sub-mask of the mask implied by the shift, only one
rldicl is needed.
llvm-svn: 195185
diff --git a/llvm/lib/Target/PowerPC/PPCISelDAGToDAG.cpp b/llvm/lib/Target/PowerPC/PPCISelDAGToDAG.cpp
index 6ba6af6..324b821 100644
--- a/llvm/lib/Target/PowerPC/PPCISelDAGToDAG.cpp
+++ b/llvm/lib/Target/PowerPC/PPCISelDAGToDAG.cpp
@@ -1122,7 +1122,21 @@
isMask_64(Imm64)) {
SDValue Val = N->getOperand(0);
MB = 64 - CountTrailingOnes_64(Imm64);
- SDValue Ops[] = { Val, getI32Imm(0), getI32Imm(MB) };
+ SH = 0;
+
+ // If the operand is a logical right shift, we can fold it into this
+ // instruction: rldicl(rldicl(x, 64-n, n), 0, mb) -> rldicl(x, 64-n, mb)
+ // for n <= mb. The right shift is really a left rotate followed by a
+ // mask, and this mask is a more-restrictive sub-mask of the mask implied
+ // by the shift.
+ if (Val.getOpcode() == ISD::SRL &&
+ isInt32Immediate(Val.getOperand(1).getNode(), Imm) && Imm <= MB) {
+ assert(Imm < 64 && "Illegal shift amount");
+ Val = Val.getOperand(0);
+ SH = 64 - Imm;
+ }
+
+ SDValue Ops[] = { Val, getI32Imm(SH), getI32Imm(MB) };
return CurDAG->SelectNodeTo(N, PPC::RLDICL, MVT::i64, Ops, 3);
}
// AND X, 0 -> 0, not "rlwinm 32".