diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig
index f0fd57f8d6..95102d386b 100644
--- a/arch/arm/Kconfig
+++ b/arch/arm/Kconfig
@@ -9,6 +9,16 @@ config ARM64
 	select PHYS_64BIT
 	select SYS_CACHE_SHIFT_6
 
+config ARM64_CRC32
+	bool "Enable support for CRC32 instruction"
+	depends on ARM64
+	default y
+	help
+	  ARMv8 implements dedicated crc32 instruction for crc32 calculation.
+	  This is faster than software crc32 calculation. This instruction may
+	  not be present on all ARMv8.0, but is always present on ARMv8.1 and
+	  newer.
+
 config POSITION_INDEPENDENT
 	bool "Generate position-independent pre-relocation code"
 	depends on ARM64 || CPU_V7A
diff --git a/arch/arm/Makefile b/arch/arm/Makefile
index c68e598a67..ce977bf632 100644
--- a/arch/arm/Makefile
+++ b/arch/arm/Makefile
@@ -18,7 +18,11 @@ arch-$(CONFIG_CPU_V7A)		=$(call cc-option, -march=armv7-a, \
 				 $(call cc-option, -march=armv7))
 arch-$(CONFIG_CPU_V7M)		=-march=armv7-m
 arch-$(CONFIG_CPU_V7R)		=-march=armv7-r
+ifeq ($(CONFIG_ARM64_CRC32),y)
+arch-$(CONFIG_ARM64)		=-march=armv8-a+crc
+else
 arch-$(CONFIG_ARM64)		=-march=armv8-a
+endif
 
 # On Tegra systems we must build SPL for the armv4 core on the device
 # but otherwise we can use the value in CONFIG_SYS_ARM_ARCH
diff --git a/lib/crc32.c b/lib/crc32.c
index f2acc107fe..5a3127e03a 100644
--- a/lib/crc32.c
+++ b/lib/crc32.c
@@ -84,7 +84,7 @@ static void __efi_runtime make_crc_table(void)
   }
   crc_table_empty = 0;
 }
-#else
+#elif !defined(CONFIG_ARM64_CRC32)
 /* ========================================================================
  * Table of CRC-32's of all single-byte values (made by make_crc_table)
  */
@@ -184,6 +184,12 @@ const uint32_t * ZEXPORT get_crc_table()
  */
 uint32_t __efi_runtime crc32_no_comp(uint32_t crc, const Bytef *buf, uInt len)
 {
+#ifdef CONFIG_ARM64_CRC32
+    crc = cpu_to_le32(crc);
+    while (len--)
+        crc = __builtin_aarch64_crc32b(crc, *buf++);
+    return le32_to_cpu(crc);
+#else
     const uint32_t *tab = crc_table;
     const uint32_t *b =(const uint32_t *)buf;
     size_t rem_len;
@@ -221,6 +227,7 @@ uint32_t __efi_runtime crc32_no_comp(uint32_t crc, const Bytef *buf, uInt len)
     }
 
     return le32_to_cpu(crc);
+#endif
 }
 #undef DO_CRC