diff --git a/mlir/lib/Dialect/Vector/IR/VectorOps.cpp b/mlir/lib/Dialect/Vector/IR/VectorOps.cpp
index bc93339a68ed3..9b2a455bace47 100644
--- a/mlir/lib/Dialect/Vector/IR/VectorOps.cpp
+++ b/mlir/lib/Dialect/Vector/IR/VectorOps.cpp
@@ -2466,8 +2466,12 @@ static OpFoldResult foldFromElementsToConstant(FromElementsOp fromElementsOp,
   if (llvm::any_of(elements, [](Attribute attr) { return !attr; }))
     return {};
 
+  // DenseElementsAttr only supports int/index/float/complex types.
   auto destVecType = fromElementsOp.getDest().getType();
   auto destEltType = destVecType.getElementType();
+  if (!destEltType.isIntOrIndexOrFloat() && !isa<ComplexType>(destEltType))
+    return {};
+
   // Constant attributes might have a different type than the return type.
   // Convert them before creating the dense elements attribute.
   auto convertedElements = llvm::map_to_vector(elements, [&](Attribute attr) {
@@ -2778,8 +2782,8 @@ BroadcastableToResult mlir::vector::isBroadcastableTo(
     Type srcType, VectorType dstVectorType,
     std::pair<VectorDim, VectorDim> *mismatchingDims) {
   // Broadcast scalar to vector of the same element type.
-  if (srcType.isIntOrIndexOrFloat() && dstVectorType &&
-      getElementTypeOrSelf(srcType) == getElementTypeOrSelf(dstVectorType))
+  if (isa<VectorElementTypeInterface>(srcType) && dstVectorType &&
+      srcType == getElementTypeOrSelf(dstVectorType))
     return BroadcastableToResult::Success;
   // From now on, only vectors broadcast.
   VectorType srcVectorType = llvm::dyn_cast<VectorType>(srcType);
diff --git a/mlir/test/Dialect/Vector/canonicalize.mlir b/mlir/test/Dialect/Vector/canonicalize.mlir
index c640ddea7507b..e7381e0c8997e 100644
--- a/mlir/test/Dialect/Vector/canonicalize.mlir
+++ b/mlir/test/Dialect/Vector/canonicalize.mlir
@@ -3726,3 +3726,17 @@ func.func @no_fold_insert_use_chain_mismatch_static_position(%arg : vector<4xf32
   %v_1 = vector.insert %val, %v_0[1] : f32 into vector<4xf32>
   return %v_1 : vector<4xf32>
 }
+
+// -----
+
+llvm.mlir.global constant @my_symbol() : i32
+
+// CHECK-LABEL: func @from_address_of_regression
+//       CHECK:   %[[a:.*]] = llvm.mlir.addressof @my_symbol
+//       CHECK:   %[[b:.*]] = vector.broadcast %[[a]] : !llvm.ptr to vector<1x!llvm.ptr>
+//       CHECK:   return %[[b]]
+func.func @from_address_of_regression() -> vector<1x!llvm.ptr> {
+  %a = llvm.mlir.addressof @my_symbol : !llvm.ptr
+  %b = vector.from_elements %a : vector<1x!llvm.ptr>
+  return %b : vector<1x!llvm.ptr>
+}
diff --git a/mlir/test/Dialect/Vector/ops.mlir b/mlir/test/Dialect/Vector/ops.mlir
index 625ffc185da95..550e52af7874b 100644
--- a/mlir/test/Dialect/Vector/ops.mlir
+++ b/mlir/test/Dialect/Vector/ops.mlir
@@ -149,7 +149,7 @@ func.func @vector_transfer_ops_tensor(%arg0: tensor<?x?xf32>,
 }
 
 // CHECK-LABEL: @vector_broadcast
-func.func @vector_broadcast(%a: f32, %b: vector<f32>, %c: vector<16xf32>, %d: vector<1x16xf32>, %e: vector<8x1xf32>, %f: vector<8x1x!llvm.ptr<1>>) {
+func.func @vector_broadcast(%a: f32, %b: vector<f32>, %c: vector<16xf32>, %d: vector<1x16xf32>, %e: vector<8x1xf32>, %f: vector<8x1x!llvm.ptr<1>>, %g: !llvm.ptr<1>) {
   // CHECK: vector.broadcast %{{.*}} : f32 to vector<f32>
   %0 = vector.broadcast %a : f32 to vector<f32>
   // CHECK: vector.broadcast %{{.*}} : vector<f32> to vector<4xf32>
@@ -164,6 +164,8 @@ func.func @vector_broadcast(%a: f32, %b: vector<f32>, %c: vector<16xf32>, %d: ve
   %5 = vector.broadcast %e : vector<8x1xf32> to vector<8x16xf32>
   // CHECK-NEXT: vector.broadcast %{{.*}} : vector<8x1x!llvm.ptr<1>> to vector<8x16x!llvm.ptr<1>>
   %6 = vector.broadcast %f : vector<8x1x!llvm.ptr<1>> to vector<8x16x!llvm.ptr<1>>
+  // CHECK-NEXT: vector.broadcast %{{.*}} : !llvm.ptr<1> to vector<8x16x!llvm.ptr<1>>
+  %7 = vector.broadcast %g : !llvm.ptr<1> to vector<8x16x!llvm.ptr<1>>
   return
 }