arch: arc: optimize the float support

* enable float support
* implement z_arch_float_disable
* add arc support in fp_sharing test

Signed-off-by: Wayne Ren <wei.ren@synopsys.com>

arch/Kconfig

@@ -455,7 +455,7 @@ config MPU_REQUIRES_NON_OVERLAPPING_REGIONS
 menuconfig FLOAT
 	bool "Floating point"
 	depends on CPU_HAS_FPU
-	depends on ARM || X86
+	depends on ARM || X86 || ARC
 	help
 	  This option allows threads to use the floating point registers.
 	  By default, only a single thread may use the registers.

arch/arc/core/thread.c

@@ -269,3 +269,21 @@ FUNC_NORETURN void z_arch_user_mode_enter(k_thread_entry_t user_entry,
 }
 
 #endif
+
+#if defined(CONFIG_FLOAT) && defined(CONFIG_FP_SHARING)
+int z_arch_float_disable(struct k_thread *thread)
+{
+	unsigned int key;
+
+	/* Ensure a preemptive context switch does not occur */
+
+	key = irq_lock();
+
+	/* Disable all floating point capabilities for the thread */
+	thread->base.user_options &= ~K_FP_REGS;
+
+	irq_unlock(key);
+
+	return 0;
+}
+#endif /* CONFIG_FLOAT && CONFIG_FP_SHARING */

arch/arc/include/swap_macros.h

@@ -56,6 +56,9 @@ extern "C" {
 	st r30, [sp, ___callee_saved_stack_t_r30_OFFSET]
 
 #ifdef CONFIG_FP_SHARING
+	ld r13, [r2, ___thread_base_t_user_options_OFFSET]
+	/* K_FP_REGS is bit 1 */
+	bbit0 r13, 1, 1f
 	st r58, [sp, ___callee_saved_stack_t_r58_OFFSET]
 	st r59, [sp, ___callee_saved_stack_t_r59_OFFSET]
 	lr r13, [_ARC_V2_FPU_STATUS]
@@ -73,7 +76,7 @@ extern "C" {
 	lr r13, [_ARC_V2_FPU_DPFP2H]
 	st_s r13, [sp, ___callee_saved_stack_t_dpfp2h_OFFSET]
 #endif
-
+1 :
 #endif
 
 	/* save stack pointer in struct k_thread */
@@ -86,6 +89,9 @@ extern "C" {
 	ld sp, [r2, _thread_offset_to_sp]
 
 #ifdef CONFIG_FP_SHARING
+	ld r13, [r2, ___thread_base_t_user_options_OFFSET]
+	/* K_FP_REGS is bit 1 */
+	bbit0 r13, 1, 2f
 	ld r58, [sp, ___callee_saved_stack_t_r58_OFFSET]
 	ld r59, [sp, ___callee_saved_stack_t_r59_OFFSET]
 
@@ -104,7 +110,7 @@ extern "C" {
 	ld_s r13, [sp, ___callee_saved_stack_t_dpfp2h_OFFSET]
 	sr r13, [_ARC_V2_FPU_DPFP2H]
 #endif
-
+2 :
 #endif
 
 #ifdef CONFIG_USERSPACE

soc/arc/snps_nsim/Kconfig.defconfig.em

@@ -36,6 +36,9 @@ config ARC_FIRQ
 config CACHE_FLUSHING
 	default y
 
+config FP_FPU_DA
+	def_bool y
+
 if (ARC_MPU_VER = 2)
 
 config MAIN_STACK_SIZE

tests/kernel/fp_sharing/generic/src/float_context.h

@@ -88,6 +88,24 @@ struct fp_non_volatile_register_set {
 #define SIZEOF_FP_NON_VOLATILE_REGISTER_SET \
 	sizeof(struct fp_non_volatile_register_set)
 
+#elif defined(CONFIG_CPU_ARCV2)
+
+struct fp_volatile_register_set {
+#ifdef CONFIG_FP_FPU_DA
+	u32_t dpfp2h;
+	u32_t dpfp2l;
+	u32_t dpfp1h;
+	u32_t dpfp1l;
+#endif
+};
+
+struct fp_non_volatile_register_set {
+	/* No non-volatile floating point registers */
+};
+
+#define SIZEOF_FP_VOLATILE_REGISTER_SET sizeof(struct fp_volatile_register_set)
+#define SIZEOF_FP_NON_VOLATILE_REGISTER_SET 0
+
 #else
 
 #error  "Architecture must provide the following definitions:\n"

tests/kernel/fp_sharing/generic/src/float_regs_arc_gcc.h

@@ -0,0 +1,114 @@
+/**
+ * @file
+ * @brief ARC GCC specific floating point register macros
+ */
+
+/*
+ * Copyright (c) 2019, Synopsys.
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ */
+
+#ifndef _FLOAT_REGS_ARC_GCC_H
+#define _FLOAT_REGS_ARC_GCC_H
+
+#if !defined(__GNUC__) || !defined(CONFIG_CPU_ARCV2)
+#error __FILE__ goes only with ARC GCC
+#endif
+
+#include <toolchain.h>
+#include "float_context.h"
+
+/**
+ *
+ * @brief Load all floating point registers
+ *
+ * This function loads ALL floating point registers pointed to by @a regs.
+ * It is expected that a subsequent call to _store_all_float_registers()
+ * will be issued to dump the floating point registers to memory.
+ *
+ * The format/organization of 'struct fp_register_set'; the generic C test
+ * code (main.c) merely treat the register set as an array of bytes.
+ *
+ * The only requirement is that the arch specific implementations of
+ * _load_all_float_registers() and _store_all_float_registers() agree
+ * on the format.
+ *
+ * @return N/A
+ */
+static inline void _load_all_float_registers(struct fp_register_set *regs)
+{
+	u32_t temp = 0;
+
+#ifdef CONFIG_FP_FPU_DA
+	__asm__ volatile (
+		"ld.ab %1, [%0, 4];\n\t"
+		"sr %1, [%2];\n\t"
+		"ld.ab %1, [%0, 4];\n\t"
+		"sr %1, [%3];\n\t"
+		"ld.ab %1, [%0, 4];\n\t"
+		"sr %1, [%4];\n\t"
+		"ld.ab %1, [%0, 4];\n\t"
+		"sr %1, [%5];\n\t"
+		: : "r" (regs), "r" (temp),
+		"i"(_ARC_V2_FPU_DPFP1L), "i"(_ARC_V2_FPU_DPFP1H),
+		"i"(_ARC_V2_FPU_DPFP2L), "i"(_ARC_V2_FPU_DPFP2H)
+		: "memory"
+		);
+#endif
+}
+
+/**
+ *
+ * @brief Dump all floating point registers to memory
+ *
+ * This function stores ALL floating point registers to the memory buffer
+ * specified by @a regs. It is expected that a previous invocation of
+ * _load_all_float_registers() occurred to load all the floating point
+ * registers from a memory buffer.
+ *
+ * @return N/A
+ */
+
+static inline void _store_all_float_registers(struct fp_register_set *regs)
+{
+	u32_t temp = 0;
+
+#ifdef CONFIG_FP_FPU_DA
+	__asm__ volatile (
+		"lr %1, [%2];\n\t"
+		"st.ab %1, [%0, 4];\n\t"
+		"lr %1, [%3];\n\t"
+		"st.ab %1, [%0, 4];\n\t"
+		"lr %1, [%4];\n\t"
+		"st.ab %1, [%0, 4];\n\t"
+		"lr %1, [%5];\n\t"
+		"st.ab %1, [%0, 4];\n\t"
+		: : "r" (regs), "r" (temp),
+		"i"(_ARC_V2_FPU_DPFP1L), "i"(_ARC_V2_FPU_DPFP1H),
+		"i"(_ARC_V2_FPU_DPFP2L), "i"(_ARC_V2_FPU_DPFP2H)
+		);
+#endif
+}
+
+/**
+ *
+ * @brief Load then dump all float registers to memory
+ *
+ * This function loads ALL floating point registers from the memory buffer
+ * specified by @a regs, and then stores them back to that buffer.
+ *
+ * This routine is called by a high priority thread prior to calling a primitive
+ * that pends and triggers a co-operative context switch to a low priority
+ * thread.
+ *
+ * @return N/A
+ */
+
+static inline void _load_then_store_all_float_registers(struct fp_register_set
+							*regs)
+{
+	_load_all_float_registers(regs);
+	_store_all_float_registers(regs);
+}
+#endif /* _FLOAT_REGS_ARC_GCC_H */

tests/kernel/fp_sharing/generic/src/main.c

@@ -63,6 +63,12 @@
 #else
 #include "float_regs_arm_other.h"
 #endif /* __GNUC__ */
+#elif defined(CONFIG_CPU_ARCV2)
+#if defined(__GNUC__)
+#include "float_regs_arc_gcc.h"
+#else
+#include "float_regs_arc_other.h"
+#endif /* __GNUC__ */
 #endif
 
 #include <tc_util.h>

tests/kernel/fp_sharing/generic/testcase.yaml

@@ -12,3 +12,10 @@ tests:
     slow: true
     tags: kernel
     timeout: 600
+  kernel.fp_sharing.arc:
+    extra_args: PI_NUM_ITERATIONS=500
+    filter: CONFIG_CPU_ARCV2 and CONFIG_CPU_HAS_FPU
+    slow: true
+    tags: kernel
+    timeout: 600
+    min_ram: 16