[PATCH] D61047: [X86] Delay creating index register negations during address matching until after we know for sure the match will succeed

[Craig Topper via Phabricator via llvm-commits]

craig.topper marked an inline comment as done.
craig.topper added inline comments.


================
Comment at: llvm/lib/Target/X86/X86ISelDAGToDAG.cpp:267
                                    SDValue &Segment) {
+      assert(!AM.NegateIndex && "Index should have been negated!");
       Base = (AM.BaseType == X86ISelAddressMode::FrameIndexBase)
----------------
spatel wrote:
> Can we deal with negation here rather than duplicating code in the callers?
We're missing the address root we would need for the insertDAGNode calls, but I can add it.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D61047/new/

https://reviews.llvm.org/D61047