userspace/libc: flush stdout & stderr on exit

Signed-off-by: EnderIce2 <enderice2@protonmail.com>

Kernel/core/cpu.cpp

@@ -308,9 +308,9 @@ namespace CPU
 			cr4.OSXMMEXCPT = true;
 
 			CPUData *CoreData = GetCPU(Core);
-			CoreData->Data.FPU.mxcsr = 0b0001111110000000;
-			CoreData->Data.FPU.mxcsrmask = 0b1111111110111111;
-			CoreData->Data.FPU.fcw = 0b0000001100111111;
+			CoreData->Data.FPU.MXCSR.raw = 0b0001111110000000;
+			CoreData->Data.FPU.MXCSR_MASK = 0b1111111110111111;
+			CoreData->Data.FPU.FCW.raw = 0b0000001100111111;
 			CPU::x86::fxrstor(&CoreData->Data.FPU);
 
 			SSEEnableAfter = true;

Kernel/include/cpu.hpp

@@ -978,27 +978,118 @@ namespace CPU
 
 		struct FXState
 		{
-			/** @brief FPU control word */
-			uint16_t fcw;
-			/** @brief FPU status word */
-			uint16_t fsw;
-			/** @brief FPU tag words */
-			uint8_t ftw;
-			/** @brief Reserved (zero) */
-			uint8_t Reserved;
-			/** @brief FPU opcode */
-			uint16_t fop;
-			/** @brief PFU instruction pointer */
-			uint64_t rip;
-			/** @brief FPU data pointer */
-			uint64_t rdp;
-			/** @brief SSE control register */
-			uint32_t mxcsr;
-			/** @brief SSE control register mask */
-			uint32_t mxcsrmask;
-			/** @brief FPU registers (last 6 bytes reserved) */
+			union
+			{
+				struct
+				{
+					/* #MF Exception Masks */
+					uint16_t IM : 1;		  /** Invalid-Operation Exception Mask */
+					uint16_t DM : 1;		  /** Denormalized-Operand Exception Mask */
+					uint16_t ZM : 1;		  /** Zero-Divide Exception Mask */
+					uint16_t OM : 1;		  /** Overflow Exception Mask */
+					uint16_t UM : 1;		  /** Underflow Exception Mask */
+					uint16_t PM : 1;		  /** Precision Exception Mask */
+					uint16_t __reserved0 : 2; /** Reserved */
+
+					/**
+					 * 00 Single precision
+					 * 01 reserved
+					 * 10 Double precision
+					 * 11 Double-extended precision (default)
+					 */
+					uint16_t PC : 2; /** Precision Control */
+
+					/**
+					 * 00 Round to nearest (default)
+					 * 01 Round down
+					 * 10 Round up
+					 * 11 Round towards zero
+					 */
+					uint16_t RC : 2;		/** Rounding Control */
+					uint16_t Infinity : 1;	/** Infinity Bit (80287 compatibility) */
+					uint16_t Reserved2 : 3; /** Reserved */
+				};
+				uint16_t raw;
+			} FCW; /** FPU Control Word */
+
+			union
+			{
+				struct
+				{
+					uint16_t IE : 1;  /** Invalid-Operation Exception */
+					uint16_t DE : 1;  /** Denormalized-Operand Exception */
+					uint16_t ZE : 1;  /** Zero-Divide Exception */
+					uint16_t OE : 1;  /** Overflow Exception */
+					uint16_t UE : 1;  /** Underflow Exception */
+					uint16_t PE : 1;  /** Precision Exception */
+					uint16_t SF : 1;  /** Stack Fault */
+					uint16_t ES : 1;  /** Exception Status */
+					uint16_t C0 : 1;  /** Condition Code 0 */
+					uint16_t C1 : 1;  /** Condition Code 1 */
+					uint16_t C2 : 1;  /** Condition Code 2 */
+					uint16_t TOP : 3; /** Top of Stack Pointer */
+					uint16_t C3 : 1;  /** Condition Code 3 */
+					uint16_t B : 1;	  /** x87 Floating-Point Unit Busy */
+				};
+				uint16_t raw;
+			} FSW; /** FPU Status Word */
+
+			/**
+			 * Tag Values
+			 *
+			 * 00 = Valid
+			 * 01 = Zero
+			 * 10 = Special
+			 * 11 = Empty
+			 */
+			uint8_t FTW; /** x87 Tag Word */
+
+			uint8_t __reserved0;
+			uint16_t FOP; /** FPU Op Code */
+			uint64_t RIP; /** PFU Instruction Pointer */
+			uint64_t RDP; /** PFU Data Pointer */
+
+			union
+			{
+				struct
+				{
+					/* Exception Flags */
+					uint32_t IE : 1;  /** Invalid-Operation Exception */
+					uint32_t DE : 1;  /** Denormalized-Operand Exception */
+					uint32_t ZE : 1;  /** Zero-Divide Exception */
+					uint32_t OE : 1;  /** Overflow Exception */
+					uint32_t UE : 1;  /** Underflow Exception */
+					uint32_t PE : 1;  /** Precision Exception */
+					uint32_t DAZ : 1; /** Denormals Are Zeros */
+
+					/* Exception Masks */
+					uint32_t IM : 1; /** Invalid-Operation Mask */
+					uint32_t DM : 1; /** Denormalized-Operand Mask */
+					uint32_t ZM : 1; /** Zero-Divide Mask */
+					uint32_t OM : 1; /** Overflow Mask */
+					uint32_t UM : 1; /** Underflow Mask */
+					uint32_t PM : 1; /** Precision Mask */
+
+					/**
+					 * 00 = round to nearest (default)
+					 * 01 = round down
+					 * 10 = round up
+					 * 11 = round toward zero
+					 */
+					uint32_t RC : 2; /** Floating-Point Rounding Control */
+					uint32_t FZ : 1; /** Flush-to-Zero for Masked Underflow */
+					uint32_t __reserved3 : 1;
+					uint32_t MM : 1; /** Misaligned Exception Mask */
+					uint32_t __reserved4 : 14;
+				};
+				uint32_t raw;
+			} MXCSR; /** SSE Control Register */
+
+			uint32_t MXCSR_MASK; /** SSE Control Register Mask */
+
+			/** FPU registers (last 6 bytes reserved) */
 			uint8_t st[8][16];
-			/** @brief XMM registers */
+			/** XMM registers */
 			uint8_t xmm[16][16];
 		} __packed __aligned(16);
 

Kernel/tasking/thread.cpp

@@ -612,9 +612,9 @@ namespace Tasking
 
 // TODO: Is really a good idea to use the FPU in kernel mode?
 #if defined(__amd64__) || defined(__i386__)
-		this->FPU.mxcsr = 0b0001111110000000;
-		this->FPU.mxcsrmask = 0b1111111110111111;
-		this->FPU.fcw = 0b0000001100111111;
+		this->FPU.MXCSR.raw = 0b0001111110000000;
+		this->FPU.MXCSR_MASK = 0b1111111110111111;
+		this->FPU.FCW.raw = 0b0000001100111111;
 #endif
 
 #ifdef DEBUG

Userspace/apps/test/simd_test/.gitignore

@@ -0,0 +1 @@
+simd_test

Userspace/apps/test/simd_test/Makefile

@@ -0,0 +1,36 @@
+default:
+	$(error Do not run this Makefile directly!)
+
+S_SOURCES = $(shell find ./ -type f -name '*.S')
+C_SOURCES = $(shell find ./ -type f -name '*.c')
+CXX_SOURCES = $(shell find ./ -type f -name '*.cpp')
+
+OBJ = $(S_SOURCES:.S=.o) $(C_SOURCES:.c=.o) $(CXX_SOURCES:.cpp=.o)
+
+FILENAME = $(notdir $(shell pwd))
+WARNCFLAG = -Wall -Wextra
+
+build: $(FILENAME)
+	cp $(FILENAME) $(WORKSPACE_DIR)/out/bin/$(FILENAME)
+
+# Use static linking
+LDFLAGS += -static -fno-pic -fno-pie -Wl,-static
+
+$(FILENAME): $(OBJ)
+	$(info Linking $@)
+	$(CC) $(LDFLAGS) $(SYSROOT) $(OBJ) -o $@
+
+%.o: %.c $(HEADERS)
+	$(info Compiling $<)
+	$(CC) $(CFLAGS) $(WARNCFLAG) -std=c17 -c $< -o $@
+
+%.o: %.cpp $(HEADERS)
+	$(info Compiling $<)
+	$(CXX) $(CFLAGS) $(WARNCFLAG) -std=c++20 -fexceptions -c $< -o $@ -fno-rtti
+
+%.o: %.S
+	$(info Compiling $<)
+	$(AS) -o $@ $<
+
+clean:
+	rm -f $(OBJ) $(FILENAME)

Userspace/apps/test/simd_test/simd.c

@@ -0,0 +1,9 @@
+int main(int, char *[], char *[])
+{
+#if defined(__amd64__) || defined(__i386__)
+	__asm__ __volatile__("movaps %%xmm1, %%xmm0\n" : : : "memory");
+#else
+#warning "Unimplemented"
+#endif
+	return 0;
+}

Userspace/libc/runtime/Scrt1.c

@@ -1,110 +1 @@
-/*
-	This file is part of Fennix Userspace.
-
-	Fennix Userspace is free software: you can redistribute it and/or
-	modify it under the terms of the GNU General Public License as
-	published by the Free Software Foundation, either version 3 of
-	the License, or (at your option) any later version.
-
-	Fennix Userspace is distributed in the hope that it will be useful,
-	but WITHOUT ANY WARRANTY; without even the implied warranty of
-	MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
-	GNU General Public License for more details.
-
-	You should have received a copy of the GNU General Public License
-	along with Fennix Userspace. If not, see <https://www.gnu.org/licenses/>.
-*/
-
-typedef void (*fct)(void);
-#define asm __asm__ __volatile__
-
-extern void (*__preinit_array_start[])(void) __attribute__((weak));
-extern void (*__preinit_array_end[])(void) __attribute__((weak));
-extern void (*__init_array_start[])(void) __attribute__((weak));
-extern void (*__init_array_end[])(void) __attribute__((weak));
-extern void (*__fini_array_start[])(void) __attribute__((weak));
-extern void (*__fini_array_end[])(void) __attribute__((weak));
-
-void __crt_init_array(void)
-{
-	for (fct *func = __init_array_start; func != __init_array_end; func++)
-		(*func)();
-}
-
-void __crt_fini_array(void)
-{
-	for (fct *func = __fini_array_start; func != __fini_array_end; func++)
-		(*func)();
-}
-
-__attribute__((naked, used, no_stack_protector, section(".text"))) void _start()
-{
-#if defined(__amd64__)
-	asm("movq $0, %rbp\n"
-		"pushq %rbp\n"
-		"pushq %rbp\n"
-		"movq %rsp, %rbp\n"
-
-		"pushq %rcx\n"
-		"pushq %rdx\n"
-		"pushq %rsi\n"
-		"pushq %rdi\n"
-
-		"call __libc_init\n"
-		"call __crt_init_array\n"
-
-		"popq %rdi\n"
-		"popq %rsi\n"
-		"popq %rdx\n"
-		"popq %rcx\n"
-
-		"call main\n"
-
-		"pushq %rax\n"
-		"call __crt_fini_array\n"
-		"popq %rax\n"
-
-		"movl %eax, %edi\n"
-		"call _exit\n");
-#elif defined(__i386__)
-#warning "i386 _start not implemented"
-#elif defined(__arm__)
-#warning "arm _start not implemented"
-#elif defined(__aarch64__)
-#warning "aarch64 _start not implemented"
-#else
-#error "Unsupported architecture"
-#endif
-}
-
-/* These are declared in GNU ld */
-enum
-{
-	NT_FNX_ABI_TAG = 1,
-	NT_FNX_VERSION = 2,
-	NT_FNX_BUILD_ID = 3,
-	NT_FNX_ARCH = 4
-};
-
-typedef struct Elf_Nhdr
-{
-	__UINT32_TYPE__ n_namesz;
-	__UINT32_TYPE__ n_descsz;
-	__UINT32_TYPE__ n_type;
-	char n_name[];
-} __attribute__((packed)) Elf_Nhdr;
-
-const struct
-{
-	Elf_Nhdr header;
-	char name[4];
-	__UINT32_TYPE__ desc[4];
-} __abi_tag __attribute__((aligned(4), section(".note.ABI-tag"))) = {
-	.header = {
-		.n_namesz = 4,							 /* "FNX" + '\0' */
-		.n_descsz = sizeof(__UINT32_TYPE__) * 4, /* Description Size */
-		.n_type = NT_FNX_ABI_TAG,				 /* Type */
-	},
-	.name = "FNX",
-	.desc = {0, 0, 0, 0},
-};
+#include "crt1.c"

Userspace/libc/runtime/crt1.c

@@ -40,9 +40,8 @@ void __crt_fini_array(void)
 __attribute__((naked, used, no_stack_protector, section(".text"))) void _start()
 {
 #if defined(__amd64__)
-	asm("movq $0, %rbp\n"
-		"pushq %rbp\n"
-		"pushq %rbp\n"
+	asm("xor %rbp, %rbp\n"
+		"andq $-16, %rsp\n"
 		"movq %rsp, %rbp\n"
 
 		"pushq %rcx\n"

Userspace/libc/src/init.c

@@ -17,6 +17,8 @@
 
 #include <fennix/syscalls.h>
 
+#include <stdio.h>
+
 int __init_pthread(void);
 void __init_stdio(void);
 
@@ -28,6 +30,8 @@ __attribute__((visibility("default"))) void __libc_init(void)
 
 __attribute__((visibility("default"))) void _exit(int Code)
 {
+	fflush(stdout);
+	fflush(stderr);
 	call_exit(Code);
 	while (1)
 		;