diff --git a/Android.mk b/Android.mk
index 585e47b..e9f691d 100644
--- a/Android.mk
+++ b/Android.mk
@@ -94,17 +94,19 @@ LOCAL_SRC_FILES := \
 
 LOCAL_C_INCLUDES := $(LOCAL_PATH)/src/headers
 
+LOCAL_CFLAGS += -DTFM_ARM
+
 ifeq ($(TARGET_ARCH_ABI),armeabi-v7a)
 # Possible optimizations:
 #  -ftree-vectorize: have GCC attempt to automatically vectorize loops
 #  -ftree-vectorizer-verbose=2: verbose output during compile
-# Note: not all V7a targets support NEON!
-#  LOCAL_ARM_NEON := true
-LOCAL_CFLAGS += -DTFM_ARM_V7A
+# Note: not all V7-a targets support NEON!
+LOCAL_ARM_NEON := true
+LOCAL_CFLAGS += -DTFM_ARM_V7A -ftree-vectorize
 else ifeq ($(TARGET_ARCH_ABI),armeabi)
 LOCAL_CFLAGS += -DTFM_ARM_V5TE
 else
-LOCAL_CFLAGS += -DTFM_ARM
+LOCAL_CFLAGS += -DTFM_ARM_V4M
 endif
 
 include $(BUILD_STATIC_LIBRARY)
diff --git a/src/mont/fp_montgomery_reduce.c b/src/mont/fp_montgomery_reduce.c
index 63a3db6..8ed733b 100644
--- a/src/mont/fp_montgomery_reduce.c
+++ b/src/mont/fp_montgomery_reduce.c
@@ -283,10 +283,10 @@ asm(                                        \
 : "%eax", "%cc")
 
 /******************************************************************/
-#elif defined(TFM_ARM)
-   /* ARMv4 code */
+#elif defined(TFM_ARM_V4M)
+   /* generic ARMv4 or higher with M */
 
-#define MONT_START 
+#define MONT_START
 #define MONT_FINI
 #define LOOP_END
 #define LOOP_START \
@@ -325,7 +325,7 @@ asm(                               \
 asm(                                \
     " LDR    r0,%1            \n\t" \
     " ADDS   r0,r0,%0         \n\t" \
-    " ITE   CS                \n\t" \
+    " ITE    CS               \n\t" \
     " MOVCS  %0,#1            \n\t" \
     " MOVCC  %0,#0            \n\t" \
     " UMLAL  r0,%0,%3,%4      \n\t" \