[PATCH] D59041: [SDAG][X86] Expand pow2 mulo using shifts

Mon Mar 11 12:59:09 PDT 2019

nikic marked 2 inline comments as done.
nikic added inline comments.

================
Comment at: llvm/lib/CodeGen/SelectionDAG/TargetLowering.cpp:5529
+    // mulo(X, 1 << S) -> { X << S, (X << S) >> S != X }
+    if (C.isPowerOf2() && (!isSigned || C.isNonNegative())) {
+      EVT ShiftAmtTy = getShiftAmountTy(VT, DAG.getDataLayout());
----------------
efriedma wrote:
> `(!isSigned || C.isNonNegative())` is specifically supposed to handle `llvm.smul.with.overflow.i32(%x, INT_MIN)`?  Can't you just convert that to `llvm.umul.with.overflow.i32(%x, INT_MIN)` and lower it using an unsigned shift?
Yes, you're right. I've implemented this now.

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D59041/new/

https://reviews.llvm.org/D59041