Index: lib/Target/X86/X86RegisterInfo.td
===================================================================
--- lib/Target/X86/X86RegisterInfo.td
+++ lib/Target/X86/X86RegisterInfo.td
@@ -443,9 +443,9 @@
 
 // Generic vector registers: VR64 and VR128.
 def VR64: RegisterClass<"X86", [x86mmx], 64, (sequence "MM%u", 0, 7)>;
-def VR128 : RegisterClass<"X86", [v16i8, v8i16, v4i32, v2i64, v4f32, v2f64],
+def VR128 : RegisterClass<"X86", [v4f32, v2f64, v16i8, v8i16, v4i32, v2i64],
                           128, (add FR32)>;
-def VR256 : RegisterClass<"X86", [v32i8, v16i16, v8i32, v4i64, v8f32, v4f64],
+def VR256 : RegisterClass<"X86", [v8f32, v4f64, v32i8, v16i16, v8i32, v4i64],
                           256, (sequence "YMM%u", 0, 15)>;
 
 // Status flags registers.
Index: test/CodeGen/X86/inline-sse.ll
===================================================================
--- test/CodeGen/X86/inline-sse.ll
+++ test/CodeGen/X86/inline-sse.ll
@@ -0,0 +1,32 @@
+; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+sse | FileCheck %s --check-prefix=X32
+; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefix=X32
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=-sse2 | FileCheck %s --check-prefix=X64
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefix=X64
+
+; PR16133 - we must treat XMM registers as v4f32 as SSE1 targets don't permit other vector types.
+
+define void @nop() nounwind {
+; X32-LABEL: nop:
+; X32:       # BB#0:
+; X32-NEXT:    pushl %ebp
+; X32-NEXT:    movl %esp, %ebp
+; X32-NEXT:    andl $-16, %esp
+; X32-NEXT:    subl $32, %esp
+; X32-NEXT:    #APP
+; X32-NEXT:    #NO_APP
+; X32-NEXT:    movaps %xmm0, (%esp)
+; X32-NEXT:    movl %ebp, %esp
+; X32-NEXT:    popl %ebp
+; X32-NEXT:    retl
+;
+; X64-LABEL: nop:
+; X64:       # BB#0:
+; X64-NEXT:    #APP
+; X64-NEXT:    #NO_APP
+; X64-NEXT:    movaps %xmm0, -{{[0-9]+}}(%rsp)
+; X64-NEXT:    retq
+  %1 = alloca <4 x float>, align 16
+  %2 = call <4 x float> asm "", "=x,~{dirflag},~{fpsr},~{flags}"()
+  store <4 x float> %2, <4 x float>* %1, align 16
+  ret void
+}