[Flang] Correct initial limit value in float min/maxloc reductions. (llvm#81260)

I was looking through to check whether Nan was being handled correctly,
and couldn't work out why simple cases were behaving differently than
they should. It turns out the initial limit values was backwards for
minloc/maxloc reductions in general. This fixes that, introduced in
llvm#79469.

Author: davemgreen

flang/lib/Optimizer/HLFIR/Transforms/OptimizedBufferization.cpp

@@ -854,7 +854,7 @@ class MinMaxlocElementalConversion : public mlir::OpRewritePattern<Op> {
         const llvm::fltSemantics &sem = ty.getFloatSemantics();
         return builder.createRealConstant(
             loc, elementType,
-            llvm::APFloat::getLargest(sem, /*Negative=*/!isMax));
+            llvm::APFloat::getLargest(sem, /*Negative=*/isMax));
       }
       unsigned bits = elementType.getIntOrFloatBitWidth();
       int64_t limitInt =

flang/test/HLFIR/maxloc-elemental.fir

@@ -110,6 +110,7 @@ func.func @_QPtest_float(%arg0: !fir.box<!fir.array<?xf32>> {fir.bindc_name = "a
   return
 }
 // CHECK-LABEL: _QPtest_float
+// CHECK:        %cst = arith.constant -3.40282347E+38 : f32
 // CHECK:        %[[V11:.*]] = fir.do_loop %arg3 = %c0 to %[[V10:.*]] step %c1 iter_args(%arg4 = %cst) -> (f32) {
 // CHECK-NEXT:     %[[V14:.*]] = arith.addi %arg3, %c1 : index
 // CHECK-NEXT:     %[[V15:.*]] = hlfir.designate %[[V1:.*]]#0 (%[[V14]])  : (!fir.box<!fir.array<?xf32>>, index) -> !fir.ref<f32>

flang/test/HLFIR/minloc-elemental.fir

@@ -295,6 +295,7 @@ func.func @_QPtest_float(%arg0: !fir.box<!fir.array<?xf32>> {fir.bindc_name = "a
   return
 }
 // CHECK-LABEL: _QPtest_float
+// CHECK:        %cst = arith.constant 3.40282347E+38 : f32
 // CHECK:        %[[V11:.*]] = fir.do_loop %arg3 = %c0 to %[[V10:.*]] step %c1 iter_args(%arg4 = %cst) -> (f32) {
 // CHECK-NEXT:     %[[V14:.*]] = arith.addi %arg3, %c1 : index
 // CHECK-NEXT:     %[[V15:.*]] = hlfir.designate %[[V1:.*]]#0 (%[[V14]])  : (!fir.box<!fir.array<?xf32>>, index) -> !fir.ref<f32>