[nrf fromlist] tfm: Remove SFN model FP limitation

Despite what the TF-M documentation says about SFN model not supporting
Floating Point, it does support it, according to TF-M developers.

Remove SFN limitation not supported with FP Hard ABI.

Upstream PR: zephyrproject-rtos/zephyr#65537

Signed-off-by: Joakim Andersson <[email protected]>

Author: joerchan

arch/arm/core/aarch32/Kconfig

@@ -272,8 +272,6 @@ config FP_HARDABI
 	  point instructions are generated and uses FPU-specific calling
 	  conventions.
 
-	  Note: When building with TF-M enabled only the IPC mode is supported.
-
 config FP_SOFTABI
 	bool "Floating point Soft ABI"
 	help

modules/trusted-firmware-m/Kconfig.tfm

@@ -323,7 +323,6 @@ config TFM_IPC
 
 config TFM_SFN
 	bool "SFN model"
-	depends on !FP_HARDABI
 	help
 	  Use the SFN Model as the SPM backend for the PSA API.
 	  The SFN model supports the SFN Partition model, and isolation level 1.