Add support for architecture armv8.2-a in Cortex-A target (#2519)

makefile

BUG=331000775

tensorflow/lite/micro/tools/make/targets/cortex_a_generic_makefile.inc

diff --git a/tensorflow/lite/micro/tools/make/targets/cortex_a_generic_makefile.inc b/tensorflow/lite/micro/tools/make/targets/cortex_a_generic_makefile.inc
index 4ca9142..ad272cc 100644
--- a/tensorflow/lite/micro/tools/make/targets/cortex_a_generic_makefile.inc
+++ b/tensorflow/lite/micro/tools/make/targets/cortex_a_generic_makefile.inc
@@ -21,7 +21,10 @@
 GCC_TARGET_ARCH :=
 GCC_TARGET_CPU :=
 
-ifeq ($(TARGET_ARCH), armv7-a)
+ifeq ($(TARGET_ARCH), armv8.2-a)
+  GCC_TARGET_ARCH := armv8.2-a
+
+else ifeq ($(TARGET_ARCH), armv7-a)
   GCC_TARGET_ARCH := armv7-a
 
 else ifeq ($(TARGET_ARCH), cortex-a32)