[PATCH] D92473: [Legalizer] Promote result type in expanding FP_TO_XINT

[Craig Topper via Phabricator via llvm-commits]

craig.topper added inline comments.


================
Comment at: llvm/lib/CodeGen/SelectionDAG/LegalizeFloatTypes.cpp:913
 
+// If the result is not legal, eg: fp -> i1, then it needs to be promoted to
+// a larger type, eg: fp -> i32. Even if it is legal, no libcall may exactly
----------------
If the type isn't legal, it would have already been handled by LegalizeIntegerTypes.cpp since results are checked for legality first.


================
Comment at: llvm/lib/CodeGen/SelectionDAG/LegalizeFloatTypes.cpp:947
   RTLIB::Libcall LC = RTLIB::UNKNOWN_LIBCALL;
-  for (unsigned IntVT = MVT::FIRST_INTEGER_VALUETYPE;
-       IntVT <= MVT::LAST_INTEGER_VALUETYPE && LC == RTLIB::UNKNOWN_LIBCALL;
-       ++IntVT) {
-    NVT = (MVT::SimpleValueType)IntVT;
-    // The type needs to big enough to hold the result.
-    if (NVT.bitsGE(RVT))
-      LC = Signed ? RTLIB::getFPTOSINT(SVT, NVT) : RTLIB::getFPTOUINT(SVT, NVT);
-  }
-  assert(LC != RTLIB::UNKNOWN_LIBCALL && "Unsupported FP_TO_XINT!");
+  assert(findConversionLibcall(LC, SVT, RVT, NVT, Signed) &&
+         "Unsupported FP_TO_XINT!");
----------------
This won't execute findConversionLibcall in a release build. assert is a macro, nothing passed to it will be evaluated if asserts aren't enabled.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D92473/new/

https://reviews.llvm.org/D92473