diff --git a/lib/Target/X86/X86Subtarget.cpp b/lib/Target/X86/X86Subtarget.cpp
index ae22298..d7e3efc 100644
--- a/lib/Target/X86/X86Subtarget.cpp
+++ b/lib/Target/X86/X86Subtarget.cpp
@@ -246,8 +246,11 @@
     }
 
     // If it's Nehalem, unaligned memory access is fast.
-    // FIXME: Nehalem is family 6. Also include Westmere and later processors?
-    if (Family == 15 && Model == 26) {
+    // Include Westmere and Sandy Bridge as well.
+    // FIXME: add later processors.
+    if ((Family == 6 && Model == 26) ||
+        (Family == 6 && Model == 44) ||
+        (Family == 6 && Model == 42)) {
       IsUAMemFast = true;
       ToggleFeature(X86::FeatureFastUAMem);
     }
diff --git a/test/CodeGen/X86/alignment-2.ll b/test/CodeGen/X86/alignment-2.ll
index cc709b5..1f9e85c 100644
--- a/test/CodeGen/X86/alignment-2.ll
+++ b/test/CodeGen/X86/alignment-2.ll
@@ -18,7 +18,9 @@
 define signext i8 @do_lo_list() nounwind optsize ssp {
 bb:
 ; CHECK:     do_lo_list
-; CHECK-NOT: movaps
+; Make sure we do not use movaps for the global variable.
+; It is okay to use movaps for writing the local variable on stack.
+; CHECK-NOT: movaps {{[0-9]*}}(%{{[a-z]*}}), {{%xmm[0-9]}}
   %myopt = alloca %struct.printQueryOpt, align 4
   %tmp = bitcast %struct.printQueryOpt* %myopt to i8*
   call void @llvm.memcpy.p0i8.p0i8.i32(i8* %tmp, i8* bitcast (%struct.printQueryOpt* getelementptr inbounds (%struct._psqlSettings* @pset, i32 0, i32 4) to i8*), i32 76, i32 4, i1 false)