diff --git a/cranelift/codegen/src/machinst/lower.rs b/cranelift/codegen/src/machinst/lower.rs
index 46b5fc1685..28e4edd0c7 100644
--- a/cranelift/codegen/src/machinst/lower.rs
+++ b/cranelift/codegen/src/machinst/lower.rs
@@ -683,10 +683,6 @@ impl<'func, I: VCodeInst> Lower<'func, I> {
                 has_side_effect,
                 value_needed,
             );
-            // Skip lowering branches; these are handled separately.
-            if self.f.dfg[inst].opcode().is_branch() {
-                continue;
-            }
 
             // Update scan state to color prior to this inst (as we are scanning
             // backward).
@@ -699,6 +695,12 @@ impl<'func, I: VCodeInst> Lower<'func, I> {
                 self.cur_scan_entry_color = Some(entry_color);
             }
 
+            // Skip lowering branches; these are handled separately
+            // (see `lower_clif_branches()` below).
+            if self.f.dfg[inst].opcode().is_branch() {
+                continue;
+            }
+
             // Normal instruction: codegen if the instruction is side-effecting
             // or any of its outputs its used.
             if has_side_effect || value_needed {
diff --git a/cranelift/filetests/filetests/isa/x64/load-op.clif b/cranelift/filetests/filetests/isa/x64/load-op.clif
index 6570a798c3..8fefaf6d42 100644
--- a/cranelift/filetests/filetests/isa/x64/load-op.clif
+++ b/cranelift/filetests/filetests/isa/x64/load-op.clif
@@ -59,3 +59,14 @@ block0(v0: i64, v1: i64):
   ; nextln: movq    0(%rax,%rdi,1), %rsi
   ; nextln: movq    %rsi, %rax
 }
+
+function %merge_scalar_to_vector(i64) -> i32x4 {
+block0(v0: i64):
+  v1 = load.i32 v0
+  v2 = scalar_to_vector.i32x4 v1
+  ; check: movss   0(%rdi), %xmm0
+
+  jump block1
+block1:
+  return v2
+}