arch: arm: core: Kconfig: introduce CONFIG_FP16_ARITHMETIC

This allows to distinguish between f16 storage format support
(CONFIG_FP16) and actual f16 arithmetic capability.

CONFIG_FP16_ARITHMETIC requires either MVE float (ARMV8_1_M_MVEF) or a
Cortex-A core (CPU_CORTEX_A).

Signed-off-by: Martin Jäger <martin.jaeger@a-labs.io>

Author: martinjaeger

arch/arm/core/Kconfig

@@ -311,7 +311,7 @@ config FP_SOFTABI
 endchoice
 
 config FP16
-	bool "Half-precision floating point support"
+	bool "Half-precision floating point storage support"
 	default y
 	help
 	  This option enables the half-precision (16-bit) floating point support
@@ -343,6 +343,17 @@ config FP16_ALT
 
 endchoice
 
+config FP16_ARITHMETIC
+	bool "Half-precision floating point arithmetic support"
+	default y
+	depends on FP16
+	depends on ARMV8_1_M_MVEF || CPU_CORTEX_A
+	help
+	  This option indicates that the target supports half-precision
+	  (16-bit) floating-point arithmetic. This requires hardware with
+	  native f16 arithmetic capability, such as Cortex-M processors
+	  with MVE float or Cortex-A processors with NEON.
+
 rsource "cortex_m/Kconfig"
 rsource "cortex_a_r/Kconfig"