Kernel: Replace i386 with i686

I don't really want to be working with i386 since it doesn't support
compare exchange instruction

LibELF/LibELF/LoadableELF.cpp

@@ -86,7 +86,7 @@ namespace LibELF
 			return BAN::Error::from_errno(ENOEXEC);
 		}
 
-#if ARCH(i386)
+#if ARCH(i686)
 		if (m_file_header.e_ident[EI_CLASS] != ELFCLASS32)
 #elif ARCH(x86_64)
 		if (m_file_header.e_ident[EI_CLASS] != ELFCLASS64)

LibELF/include/LibELF/ELF.h

@@ -42,7 +42,7 @@ namespace LibELF
 		const Elf32SectionHeader& section_header32(size_t) const;
 		const char* lookup_section_name32(uint32_t) const;
 		const char* lookup_string32(size_t, uint32_t) const;
-#if ARCH(i386)
+#if ARCH(i686)
 		const Elf32FileHeader& file_header_native() const { return file_header32(); }
 		const Elf32ProgramHeader& program_header_native(size_t index) const { return program_header32(index); }
 		const Elf32SectionHeader& section_header_native(size_t index) const { return section_header32(index); }

LibELF/include/LibELF/Types.h

@@ -155,7 +155,7 @@ namespace LibELF
 		Elf64Xword p_align;
 	};
 
-#if ARCH(i386)
+#if ARCH(i686)
 	using ElfNativeAddr = Elf32Addr;
 	using ElfNativeOff = Elf32Off;
 	using ElfNativeHalf = Elf32Half;

kernel/CMakeLists.txt

@@ -4,7 +4,7 @@ project(kernel CXX C ASM)
 
 if("${BANAN_ARCH}" STREQUAL "x86_64")
 	set(ELF_FORMAT elf64-x86-64)
-elseif("${BANAN_ARCH}" STREQUAL "i386")
+elseif("${BANAN_ARCH}" STREQUAL "i686")
 	set(ELF_FORMAT elf32-i386)
 endif()
 
@@ -120,13 +120,16 @@ if("${BANAN_ARCH}" STREQUAL "x86_64")
 		${LAI_SOURCES}
 		kernel/lai_host.cpp
 	)
-elseif("${BANAN_ARCH}" STREQUAL "i386")
+elseif("${BANAN_ARCH}" STREQUAL "i686")
 	set(KERNEL_SOURCES
 		${KERNEL_SOURCES}
-		arch/i386/boot.S
-		arch/i386/SpinLock.S
-		arch/i386/Syscall.S
-		arch/i386/Thread.S
+		arch/i686/boot.S
+		arch/i686/GDT.cpp
+		arch/i686/IDT.cpp
+		arch/i686/PageTable.cpp
+		arch/i686/Signal.S
+		arch/i686/Syscall.S
+		arch/i686/Thread.S
 	)
 else()
 	message(FATAL_ERROR "unsupported architecure ${BANAN_ARCH}")
@@ -179,8 +182,8 @@ if("${BANAN_ARCH}" STREQUAL "x86_64")
 	target_compile_options(kernel PUBLIC -mcmodel=kernel -mno-red-zone)
 	target_link_options(kernel PUBLIC LINKER:-z,max-page-size=4096)
 	target_link_options(kernel PUBLIC LINKER:-T,${CMAKE_CURRENT_SOURCE_DIR}/arch/x86_64/linker.ld)
-elseif("${BANAN_ARCH}" STREQUAL "i386")
-	target_link_options(kernel PUBLIC LINKER:-T,${CMAKE_CURRENT_SOURCE_DIR}/arch/i386/linker.ld)
+elseif("${BANAN_ARCH}" STREQUAL "i686")
+	target_link_options(kernel PUBLIC LINKER:-T,${CMAKE_CURRENT_SOURCE_DIR}/arch/i686/linker.ld)
 endif()
 
 target_link_options(kernel PUBLIC -ffreestanding -nostdlib)

kernel/arch/i386/SpinLock.S

@@ -1,19 +0,0 @@
-.global spinlock_lock_asm
-spinlock_lock_asm:
-	movl 4(%esp), %eax
-	lock; btsl $0, (%eax)
-	jnc .done
-.retry:
-	pause
-	testl $1, (%eax)
-	jne .retry
-	lock; btsl $0, (%eax)
-	jc .retry
-.done:
-	ret
-
-.global spinlock_unlock_asm
-spinlock_unlock_asm:
-	movl 4(%esp), %eax
-	movl $0, (%eax)
-	ret

kernel/arch/i386/Thread.S

@@ -1,47 +0,0 @@
-# uint32_t read_ip()
-.global read_ip
-read_ip:
-	popl %eax
-	jmp *%eax
-
-exit_thread_trampoline:
-	addl $4, %esp
-	pushl (%esp)
-	ret
-
-# void start_thread(uint32_t sp, uint32_t ip)
-.global start_thread
-start_thread:
-	movl 8(%esp), %ecx
-	movl 4(%esp), %esp
-	movl $0, %ebp
-	pushl $exit_thread_trampoline
-	sti
-	jmp *%ecx
-
-# void continue_thread(uint32_t sp, uint32_t ip)
-.global continue_thread
-continue_thread:
-	movl 8(%esp), %ecx
-	movl 4(%esp), %esp
-	movl $0, %eax
-	jmp *%ecx
-
-# void thread_jump_userspace(uint32_t sp, uint32_t ip)
-.global thread_jump_userspace
-thread_jump_userspace:
-	movl $0x23, %eax
-	movw %ax, %ds
-	movw %ax, %es
-	movw %ax, %fs
-	movw %ax, %gs
-
-	movl 8(%esp), %ecx
-	movl 4(%esp), %esp
-
-	pushl $0x23
-	pushl %esp
-	pushfl
-	pushl $0x1B
-	pushl %ecx
-	iret

kernel/arch/i386/boot.S

@@ -1,182 +0,0 @@
-# Declare constants for the multiboot header
-.set ALIGN,			1<<0						# align loaded modules on page boundaries
-.set MEMINFO,		1<<1						# provide memory map
-.set VIDEOINFO,		1<<2						# provide video info
-.set MB_FLAGS,		ALIGN | MEMINFO | VIDEOINFO # this is the Multiboot 'flag' field
-.set MB_MAGIC,		0x1BADB002					# 'magic number' lets bootloader find the header
-.set MB_CHECKSUM,	-(MB_MAGIC + MB_FLAGS)		#checksum of above, to prove we are multiboot
-
-# Multiboot header
-.section .multiboot, "aw"
-	.align 4
-	.long MB_MAGIC
-	.long MB_FLAGS
-	.long MB_CHECKSUM
-	.skip 20
-
-	.long 0
-	.long 800
-	.long 600
-	.long 32
-
-.section .bss, "aw", @nobits
-# Create stack
-.global g_boot_stack_bottom
-g_boot_stack_bottom:
-	.skip 16384
-.global g_boot_stack_top
-g_boot_stack_top:
-
-# 0 MiB -> 1 MiB: bootloader stuff
-# 1 MiB ->      : kernel
-.align 32
-boot_page_directory_pointer_table:
-	.skip 4 * 8
-.align 4096
-boot_page_directory1:
-	.skip 512 * 8
-
-.global g_kernel_cmdline
-g_kernel_cmdline:
-	.skip 4096
-
-.global g_multiboot_info
-g_multiboot_info:
-	.skip 4
-.global g_multiboot_magic
-g_multiboot_magic:
-	.skip 4
-
-.section .text
-
-boot_gdt:
-	.quad 0x0000000000000000 # null
-	.quad 0x00CF9A000000FFFF # kernel code
-	.quad 0x00CF92000000FFFF # kernel data
-boot_gdtr:
-	.short . - boot_gdt - 1
-	.long boot_gdt
-
-has_cpuid:
-	pushfl
-	pushfl
-	xorl $0x00200000, (%esp)
-	popfl
-	pushfl
-	popl %eax
-	xorl (%esp), %eax
-	popfl
-	testl $0x00200000, %eax
-	ret
-
-has_pae:
-	movl $0, %eax
-	cpuid
-	testl $(1 << 6), %edx
-	ret
-
-has_sse:
-	movl $1, %eax
-	cpuid
-	testl $(1 << 25), %edx
-	ret
-
-check_requirements:
-	call has_cpuid
-	jz .exit
-	call has_pae
-	jz .exit
-	call has_sse
-	jz .exit
-	ret
-.exit:
-	jmp system_halt
-
-copy_kernel_commandline:
-	pushl %esi
-	pushl %edi
-	movl g_multiboot_info, %esi
-	addl $16, %esi
-	movl (%esi), %esi
-	movl $1024, %ecx
-	movl $g_kernel_cmdline, %edi
-	rep movsl
-	popl %edi
-	popl %esi
-	ret
-
-enable_sse:
-	movl %cr0, %eax
-	andw $0xFFFB, %ax
-	orw $0x0002, %ax
-	movl %eax, %cr0
-	movl %cr4, %eax
-	orw $0x0600, %ax
-	movl %eax, %cr4
-	ret
-
-initialize_paging:
-	# identity map first 6 MiB
-	movl $(0x00000000           + 0x83), boot_page_directory1 + 0
-	movl $(0x00200000           + 0x83), boot_page_directory1 + 8
-	movl $(0x00400000           + 0x83), boot_page_directory1 + 16
-	movl $(boot_page_directory1 + 0x01), boot_page_directory_pointer_table
-
-	# enable PAE
-	movl %cr4, %ecx
-	orl $0x20, %ecx
-	movl %ecx, %cr4
-
-	# set address of paging structures
-	movl $boot_page_directory_pointer_table, %ecx
-	movl %ecx, %cr3
-
-	# enable paging
-	movl %cr0, %ecx
-	orl $0x80000000, %ecx
-	movl %ecx, %cr0
-
-	ret
-
-initialize_gdt:
-	lgdt boot_gdtr
-
-	# flush gdt
-	movw $0x10, %ax
-	movw %ax, %ds
-	movw %ax, %es
-	movw %ax, %fs
-	movw %ax, %gs
-	movw %ax, %ss
-	jmp $0x08, $flush
-flush:
-	ret
-
-.global _start
-.type _start, @function
-_start:
-	# Initialize stack and multiboot info
-	movl $g_boot_stack_top, %esp
-	movl %eax, g_multiboot_magic
-	movl %ebx, g_multiboot_info
-
-	call copy_kernel_commandline
-	call check_requirements
-	call enable_sse
-
-	call initialize_paging
-	call initialize_gdt
-
-	call _init
-
-	# call to the kernel itself (clear ebp for stacktrace)
-	xorl %ebp, %ebp
-	call kernel_main
-
-	call _fini
-
-system_halt:
-	xchgw %bx, %bx
-	cli
-1:	hlt
-	jmp 1b

kernel/arch/i386/linker.ld

@@ -1,28 +0,0 @@
-ENTRY (_start)
-
-SECTIONS
-{
-	. = 0x00100000;
-
-	g_kernel_start = .;
-	.text BLOCK(4K) : ALIGN(4K)
-	{
-		*(.multiboot)
-		*(.text)
-	}
-	.rodata BLOCK(4K) : ALIGN(4K)
-	{
-		*(.rodata.*)
-	}
-	.data BLOCK(4K) : ALIGN(4K)
-	{
-		*(.data)
-	}
-	.bss BLOCK(4K) : ALIGN(4K)
-	{
-		*(COMMON)
-		*(.bss)
-	}
-
-	g_kernel_end = .;
-}

kernel/arch/i686/GDT.cpp

@@ -0,0 +1,21 @@
+#include <kernel/GDT.h>
+
+namespace Kernel
+{
+
+	GDT* GDT::create()
+	{
+		ASSERT_NOT_REACHED();
+	}
+
+	void GDT::write_entry(uint8_t, uint32_t, uint32_t, uint8_t, uint8_t)
+	{
+		ASSERT_NOT_REACHED();
+	}
+
+	void GDT::write_tss()
+	{
+		ASSERT_NOT_REACHED();
+	}
+
+}

kernel/arch/i686/IDT.cpp

@@ -0,0 +1,31 @@
+#include <kernel/IDT.h>
+
+namespace Kernel
+{
+
+	IDT* IDT::create()
+	{
+		ASSERT_NOT_REACHED();
+	}
+
+	[[noreturn]] void IDT::force_triple_fault()
+	{
+		ASSERT_NOT_REACHED();
+	}
+
+	void IDT::register_irq_handler(uint8_t, Interruptable*)
+	{
+		ASSERT_NOT_REACHED();
+	}
+
+	void IDT::register_interrupt_handler(uint8_t, void (*)())
+	{
+		ASSERT_NOT_REACHED();
+	}
+
+	void IDT::register_syscall_handler(uint8_t, void (*)())
+	{
+		ASSERT_NOT_REACHED();
+	}
+
+}

kernel/arch/i686/PageTable.cpp

@@ -0,0 +1,144 @@
+#include <kernel/Memory/PageTable.h>
+#include <kernel/Lock/SpinLock.h>
+
+namespace Kernel
+{
+
+	RecursiveSpinLock PageTable::s_fast_page_lock;
+
+	void PageTable::initialize()
+	{
+		ASSERT_NOT_REACHED();
+	}
+
+	PageTable& PageTable::kernel()
+	{
+		ASSERT_NOT_REACHED();
+	}
+
+	bool PageTable::is_valid_pointer(uintptr_t)
+	{
+		ASSERT_NOT_REACHED();
+	}
+
+	BAN::ErrorOr<PageTable*> PageTable::create_userspace()
+	{
+		ASSERT_NOT_REACHED();
+	}
+
+	PageTable::~PageTable()
+	{
+		ASSERT_NOT_REACHED();
+	}
+
+	void PageTable::unmap_page(vaddr_t)
+	{
+		ASSERT_NOT_REACHED();
+	}
+
+	void PageTable::unmap_range(vaddr_t, size_t)
+	{
+		ASSERT_NOT_REACHED();
+	}
+
+	void PageTable::map_range_at(paddr_t, vaddr_t, size_t, flags_t)
+	{
+		ASSERT_NOT_REACHED();
+	}
+
+	void PageTable::map_page_at(paddr_t, vaddr_t, flags_t)
+	{
+		ASSERT_NOT_REACHED();
+	}
+
+	paddr_t PageTable::physical_address_of(vaddr_t) const
+	{
+		ASSERT_NOT_REACHED();
+	}
+
+	PageTable::flags_t PageTable::get_page_flags(vaddr_t) const
+	{
+		ASSERT_NOT_REACHED();
+	}
+
+	bool PageTable::is_page_free(vaddr_t) const
+	{
+		ASSERT_NOT_REACHED();
+	}
+
+	bool PageTable::is_range_free(vaddr_t, size_t) const
+	{
+		ASSERT_NOT_REACHED();
+	}
+
+	bool PageTable::reserve_page(vaddr_t, bool)
+	{
+		ASSERT_NOT_REACHED();
+	}
+
+	bool PageTable::reserve_range(vaddr_t, size_t, bool)
+	{
+		ASSERT_NOT_REACHED();
+	}
+
+	vaddr_t PageTable::reserve_free_page(vaddr_t, vaddr_t)
+	{
+		ASSERT_NOT_REACHED();
+	}
+
+	vaddr_t PageTable::reserve_free_contiguous_pages(size_t, vaddr_t, vaddr_t)
+	{
+		ASSERT_NOT_REACHED();
+	}
+
+	void PageTable::load()
+	{
+		ASSERT_NOT_REACHED();
+	}
+
+	void PageTable::initial_load()
+	{
+		ASSERT_NOT_REACHED();
+	}
+
+	void PageTable::debug_dump()
+	{
+		ASSERT_NOT_REACHED();
+	}
+
+	uint64_t PageTable::get_page_data(vaddr_t) const
+	{
+		ASSERT_NOT_REACHED();
+	}
+
+	void PageTable::initialize_kernel()
+	{
+		ASSERT_NOT_REACHED();
+	}
+
+	void PageTable::map_kernel_memory()
+	{
+		ASSERT_NOT_REACHED();
+	}
+
+	void PageTable::prepare_fast_page()
+	{
+		ASSERT_NOT_REACHED();
+	}
+
+	void PageTable::invalidate(vaddr_t)
+	{
+		ASSERT_NOT_REACHED();
+	}
+
+	void PageTable::map_fast_page(paddr_t)
+	{
+		ASSERT_NOT_REACHED();
+	}
+
+	void PageTable::unmap_fast_page()
+	{
+		ASSERT_NOT_REACHED();
+	}
+
+}

kernel/arch/i686/Signal.S

@@ -0,0 +1,31 @@
+.section .userspace, "aw"
+
+// stack contains
+//    return address
+//    signal number
+//    signal handler
+
+.global signal_trampoline
+signal_trampoline:
+	ud2
+
+	pushl %ebp
+	movl %esp, %ebp
+	subl $8, %esp
+
+	pusha
+
+	movl 40(%esp), %edi
+	movl 36(%esp), %eax
+
+	subl $12, %esp
+	pushl %edi
+	call *%eax
+	addl $16, %esp
+
+	popa
+
+	leave
+	addl $8, %esp
+
+	ret

kernel/arch/i686/Syscall.S

@@ -1,5 +1,6 @@
 .global sys_fork_trampoline
 sys_fork_trampoline:
+	ud2
 	subl $4, %esp
 	pushl %ebx
 	pushl %ebp

kernel/arch/i686/Thread.S

@@ -0,0 +1,20 @@
+# uint32_t read_ip()
+.global read_ip
+read_ip:
+	popl %eax
+	jmp *%eax
+
+# void start_thread(uint32_t sp, uint32_t ip)
+.global start_thread
+start_thread:
+	ud2
+
+# void continue_thread(uint32_t sp, uint32_t ip)
+.global continue_thread
+continue_thread:
+	ud2
+
+# void thread_userspace_trampoline(uint32_t sp, uint32_t ip, int argc, char** argv, char** envp)
+.global thread_userspace_trampoline
+thread_userspace_trampoline:
+	ud2

kernel/arch/i686/boot.S

@@ -0,0 +1,268 @@
+.set PG_PRESENT,	1<<0
+.set PG_READ_WRITE,	1<<1
+.set PG_PAGE_SIZE,	1<<7
+
+.set FB_WIDTH,	800
+.set FB_HEIGHT,	600
+.set FB_BPP,	32
+
+#define KERNEL_OFFSET 0xC0000000
+#define V2P(vaddr) ((vaddr) - KERNEL_OFFSET)
+
+.code32
+
+# multiboot2 header
+.section .multiboot, "aw"
+	.align 8
+multiboot2_start:
+	.long 0xE85250D6
+	.long 0
+	.long multiboot2_end - multiboot2_start
+	.long -(0xE85250D6 + (multiboot2_end - multiboot2_start))
+
+	# framebuffer tag
+	.align 8
+	.short 5
+	.short 0
+	.long 20
+	.long FB_WIDTH
+	.long FB_HEIGHT
+	.long FB_BPP
+
+	# legacy start
+	.align 8
+	.short 3
+	.short 0
+	.long 12
+	.long V2P(_start)
+
+	.align 8
+	.short 0
+	.short 0
+	.long 8
+multiboot2_end:
+
+.section .bananboot, "aw"
+	.align 8
+bananboot_start:
+	.long 0xBABAB007
+	.long -(0xBABAB007 + FB_WIDTH + FB_HEIGHT + FB_BPP)
+	.long FB_WIDTH
+	.long FB_HEIGHT
+	.long FB_BPP
+bananboot_end:
+
+.section .bss, "aw", @nobits
+	boot_stack_bottom:
+		.skip 4096 * 4
+	boot_stack_top:
+
+	.global g_kernel_cmdline
+	g_kernel_cmdline:
+		.skip 4096
+
+	bootloader_magic:
+		.skip 8
+	bootloader_info:
+		.skip 8
+
+.section .data
+
+.align 4096
+boot_pml4:
+boot_pdpt_lo:
+boot_pdpt_hi:
+boot_pd:
+
+boot_gdt:
+	.quad 0x0000000000000000 # null descriptor
+	.quad 0x00CF9A000000FFFF # kernel code
+	.quad 0x00CF92000000FFFF # kernel data
+boot_gdtr:
+	.short . - boot_gdt - 1
+	.long V2P(boot_gdt)
+
+.global g_ap_startup_done
+g_ap_startup_done:
+	.byte 0
+.global g_ap_running_count
+g_ap_running_count:
+	.byte 0
+.global g_ap_stack_loaded
+g_ap_stack_loaded:
+	.byte 0
+
+.section .text
+
+has_cpuid:
+	pushfl
+	pushfl
+	xorl $0x00200000, (%esp)
+	popfl
+	pushfl
+	popl %eax
+	xorl (%esp), %eax
+	popfl
+	testl $0x00200000, %eax
+	ret
+
+check_requirements:
+	call has_cpuid
+	jz .exit
+	ret
+.exit:
+	jmp system_halt
+
+enable_sse:
+	movl %cr0, %eax
+	andw $0xFFFB, %ax
+	orw $0x0002, %ax
+	movl %eax, %cr0
+	movl %cr4, %eax
+	orw $0x0600, %ax
+	movl %eax, %cr4
+	ret
+
+initialize_paging:
+	# enable PAE
+	movl %cr4, %ecx
+	orl $0x20, %ecx
+	movl %ecx, %cr4
+
+	# set address of paging structures
+	movl $V2P(boot_pml4), %ecx
+	movl %ecx, %cr3
+
+	# enable paging
+	movl %cr0, %ecx
+	orl $0x80000000, %ecx
+	movl %ecx, %cr0
+
+	ret
+
+.global _start
+.type _start, @function
+_start:
+	cli; cld
+
+	# Initialize stack and multiboot info
+	movl %eax, V2P(bootloader_magic)
+	movl %ebx, V2P(bootloader_info)
+
+	movl $V2P(boot_stack_top), %esp
+
+	call check_requirements
+	call enable_sse
+
+	call initialize_paging
+
+	# flush gdt
+	lgdt V2P(boot_gdtr)
+	ljmpl $0x08, $V2P(gdt_flush)
+
+gdt_flush:
+	movw $0x10, %ax
+	movw %ax, %ds
+	movw %ax, %ss
+	movw %ax, %es
+
+	# move stack pointer to higher half
+	movl %esp, %esp
+	addl $KERNEL_OFFSET, %esp
+
+	# jump to higher half
+	leal higher_half, %ecx
+	jmp *%ecx
+
+higher_half:
+	# call global constuctors
+	call _init
+
+	# call to the kernel itself (clear ebp for stacktrace)
+	xorl %ebp, %ebp
+
+	movl V2P(bootloader_magic), %edi
+	movl V2P(bootloader_info),  %esi
+	call kernel_main
+
+	# call global destructors
+	call _fini
+
+system_halt:
+	xchgw %bx, %bx
+	cli
+1:	hlt
+	jmp 1b
+
+
+.section .ap_init, "ax"
+
+.code16
+.global ap_trampoline
+ap_trampoline:
+	jmp 1f
+.align 8
+ap_stack_ptr:
+	.skip 4
+1:
+	cli; cld
+	ljmpl $0x00, $ap_cs_clear
+
+ap_cs_clear:
+	xorw %ax, %ax
+	movw %ax, %ds
+
+	# load ap gdt and enter protected mode
+	lgdt ap_gdtr
+	movl %cr0, %eax
+	orb $1, %al
+	movl %eax, %cr0
+	ljmpl $0x08, $ap_protected_mode
+
+.code32
+ap_protected_mode:
+	movw $0x10, %ax
+	movw %ax, %ds
+	movw %ax, %ss
+	movw %ax, %es
+
+	movl ap_stack_ptr, %esp
+	movb $1, V2P(g_ap_stack_loaded)
+
+	call V2P(enable_sse)
+
+	call V2P(initialize_paging)
+
+	# load boot gdt and enter long mode
+	lgdt V2P(boot_gdtr)
+	ljmpl $0x08, $ap_flush_gdt
+
+ap_flush_gdt:
+	# move stack pointer to higher half
+	movl %esp, %esp
+	addl $KERNEL_OFFSET, %esp
+
+	# jump to higher half
+	leal ap_higher_half, %ecx
+	jmp *%ecx
+
+ap_higher_half:
+	# clear rbp for stacktrace
+	xorl %ebp, %ebp
+
+1:	pause
+	cmpb $0, g_ap_startup_done
+	jz 1b
+
+	lock incb g_ap_running_count
+
+	call ap_main
+	jmp system_halt
+
+ap_gdt:
+	.quad 0x0000000000000000 # null descriptor
+	.quad 0x00CF9A000000FFFF # 32 bit code
+	.quad 0x00CF92000000FFFF # 32 bit data
+ap_gdtr:
+	.short . - ap_gdt - 1
+	.long ap_gdt

kernel/arch/i686/linker.ld

@@ -0,0 +1,45 @@
+ENTRY (_start)
+
+KERNEL_OFFSET = 0xC0000000;
+
+SECTIONS
+{
+	. = 0xF000;
+	.ap_init ALIGN(4K) : AT(ADDR(.ap_init))
+	{
+		g_ap_init_addr = .;
+		*(.ap_init)
+	}
+
+	. = 0x00100000 + KERNEL_OFFSET;
+
+	g_kernel_start = .;
+	.text ALIGN(4K) : AT(ADDR(.text) - KERNEL_OFFSET)
+	{
+		g_kernel_execute_start = .;
+		*(.multiboot)
+		*(.bananboot)
+		*(.text.*)
+	}
+	.userspace ALIGN(4K) : AT(ADDR(.userspace) - KERNEL_OFFSET)
+	{
+		g_userspace_start = .;
+		*(.userspace)
+		g_userspace_end = .;
+		g_kernel_execute_end = .;
+	}
+	.rodata ALIGN(4K) : AT(ADDR(.rodata) - KERNEL_OFFSET)
+	{
+		*(.rodata.*)
+	}
+	.data ALIGN(4K) : AT(ADDR(.data) - KERNEL_OFFSET)
+	{
+		*(.data)
+	}
+	.bss ALIGN(4K) : AT(ADDR(.bss) - KERNEL_OFFSET)
+	{
+		*(COMMON)
+		*(.bss)
+	}
+	g_kernel_end = .;
+}

kernel/arch/x86_64/Thread.S

@@ -1,4 +1,4 @@
-# uint64_t read_()
+# uint64_t read_ip()
 .global read_ip
 read_ip:
 	popq %rax

kernel/icxxabi.cpp

@@ -1,115 +1,41 @@
-#include <BAN/Atomic.h>
 #include <kernel/Panic.h>
 
 #define ATEXIT_MAX_FUNCS 128
 
-#ifdef __cplusplus
-extern "C" {
-#endif
-
-typedef unsigned uarch_t;
-
 struct atexit_func_entry_t
 {
-	/*
-	* Each member is at least 4 bytes large. Such that each entry is 12bytes.
-	* 128 * 12 = 1.5KB exact.
-	**/
-	void (*destructor_func)(void *);
-	void *obj_ptr;
-	void *dso_handle;
+	void(*func)(void*);
+	void* arg;
+	void* dso_handle;
 };
 
-atexit_func_entry_t __atexit_funcs[ATEXIT_MAX_FUNCS];
-uarch_t __atexit_func_count = 0;
+static atexit_func_entry_t __atexit_funcs[ATEXIT_MAX_FUNCS];
+static size_t __atexit_func_count = 0;
 
-int __cxa_atexit(void (*f)(void *), void *objptr, void *dso)
+extern "C" int __cxa_atexit(void(*func)(void*), void* arg, void* dso_handle)
 {
-	if (__atexit_func_count >= ATEXIT_MAX_FUNCS) {return -1;};
-	__atexit_funcs[__atexit_func_count].destructor_func = f;
-	__atexit_funcs[__atexit_func_count].obj_ptr = objptr;
-	__atexit_funcs[__atexit_func_count].dso_handle = dso;
-	__atexit_func_count++;
-	return 0; /*I would prefer if functions returned 1 on success, but the ABI says...*/
+	if (__atexit_func_count >= ATEXIT_MAX_FUNCS)
+		return -1;
+	auto& atexit_func = __atexit_funcs[__atexit_func_count++];
+	atexit_func.func = func;
+	atexit_func.arg = arg;
+	atexit_func.dso_handle = dso_handle;
+	return 0;
 };
 
-void __cxa_finalize(void *f)
+extern "C" void __cxa_finalize(void* f)
 {
-	uarch_t i = __atexit_func_count;
-	if (!f)
+	for (size_t i = __atexit_func_count; i > 0; i--)
 	{
-		/*
-		* According to the Itanium C++ ABI, if __cxa_finalize is called without a
-		* function ptr, then it means that we should destroy EVERYTHING MUAHAHAHA!!
-		*
-		* TODO:
-		* Note well, however, that deleting a function from here that contains a __dso_handle
-		* means that one link to a shared object file has been terminated. In other words,
-		* We should monitor this list (optional, of course), since it tells us how many links to
-		* an object file exist at runtime in a particular application. This can be used to tell
-		* when a shared object is no longer in use. It is one of many methods, however.
-		**/
-		//You may insert a prinf() here to tell you whether or not the function gets called. Testing
-		//is CRITICAL!
-		while (i--)
+		auto& atexit_func = __atexit_funcs[i - 1];
+		if (atexit_func.func == nullptr)
+			continue;
+		if (f == nullptr || f == atexit_func.func)
 		{
-			if (__atexit_funcs[i].destructor_func)
-			{
-				/* ^^^ That if statement is a safeguard...
-				* To make sure we don't call any entries that have already been called and unset at runtime.
-				* Those will contain a value of 0, and calling a function with value 0
-				* will cause undefined behaviour. Remember that linear address 0,
-				* in a non-virtual address space (physical) contains the IVT and BDA.
-				*
-				* In a virtual environment, the kernel will receive a page fault, and then probably
-				* map in some trash, or a blank page, or something stupid like that.
-				* This will result in the processor executing trash, and...we don't want that.
-				**/
-				(*__atexit_funcs[i].destructor_func)(__atexit_funcs[i].obj_ptr);
-			};
-		};
-		return;
-	};
-
-	while (i--)
-	{
-		/*
-		* The ABI states that multiple calls to the __cxa_finalize(destructor_func_ptr) function
-		* should not destroy objects multiple times. Only one call is needed to eliminate multiple
-		* entries with the same address.
-		*
-		* FIXME:
-		* This presents the obvious problem: all destructors must be stored in the order they
-		* were placed in the list. I.e: the last initialized object's destructor must be first
-		* in the list of destructors to be called. But removing a destructor from the list at runtime
-		* creates holes in the table with unfilled entries.
-		* Remember that the insertion algorithm in __cxa_atexit simply inserts the next destructor
-		* at the end of the table. So, we have holes with our current algorithm
-		* This function should be modified to move all the destructors above the one currently
-		* being called and removed one place down in the list, so as to cover up the hole.
-		* Otherwise, whenever a destructor is called and removed, an entire space in the table is wasted.
-		**/
-		if (__atexit_funcs[i].destructor_func == f)
-		{
-			/*
-			* Note that in the next line, not every destructor function is a class destructor.
-			* It is perfectly legal to register a non class destructor function as a simple cleanup
-			* function to be called on program termination, in which case, it would not NEED an
-			* object This pointer. A smart programmer may even take advantage of this and register
-			* a C function in the table with the address of some structure containing data about
-			* what to clean up on exit.
-			* In the case of a function that takes no arguments, it will simply be ignore within the
-			* function itself. No worries.
-			**/
-			(*__atexit_funcs[i].destructor_func)(__atexit_funcs[i].obj_ptr);
-			__atexit_funcs[i].destructor_func = 0;
-
-			/*
-			* Notice that we didn't decrement __atexit_func_count: this is because this algorithm
-			* requires patching to deal with the FIXME outlined above.
-			**/
-		};
-	};
+			atexit_func.func(atexit_func.arg);
+			atexit_func.func = nullptr;
+		}
+	}
 };
 
 namespace __cxxabiv1
@@ -118,23 +44,19 @@ namespace __cxxabiv1
 
 	int __cxa_guard_acquire (__guard* g)
 	{
-		auto& atomic = *reinterpret_cast<BAN::Atomic<__guard>*>(g);
-		return atomic == 0;
+		uint8_t* byte = reinterpret_cast<uint8_t*>(g);
+		uint8_t zero = 0;
+		return __atomic_compare_exchange_n(byte, &zero, 1, false, __ATOMIC_ACQUIRE, __ATOMIC_ACQUIRE);
 	}
 
 	void __cxa_guard_release (__guard* g)
 	{
-		auto& atomic = *reinterpret_cast<BAN::Atomic<__guard>*>(g);
-		atomic = 1;
+		uint8_t* byte = reinterpret_cast<uint8_t*>(g);
+		__atomic_store_n(byte, 0, __ATOMIC_RELEASE);
 	}
 
 	void __cxa_guard_abort (__guard*)
 	{
 		Kernel::panic("__cxa_guard_abort");
-		__builtin_unreachable();
 	}
 }
-
-#ifdef __cplusplus
-};
-#endif

kernel/include/kernel/Arch.h

@@ -1,11 +1,11 @@
 #pragma once
 
 #define x86_64 1
-#define i386   2
+#define i686   2
 
 #define ARCH(arch) (__arch == arch)
 
-#if !defined(__arch) || (__arch != x86_64 && __arch != i386)
+#if !defined(__arch) || (__arch != x86_64 && __arch != i686)
 	#error "Unsupported architecture"
 #endif
 
@@ -13,10 +13,12 @@
 	#define read_rsp(rsp) asm volatile("movq %%rsp, %0" : "=r"(rsp))
 	#define push_callee_saved() asm volatile("pushq %rbx; pushq %rbp; pushq %r12; pushq %r13; pushq %r14; pushq %r15")
 	#define pop_callee_saved()  asm volatile("popq %r15; popq %r14; popq %r13; popq %r12; popq %rbp; popq %rbx")
-#else
+#elif ARCH(i686)
 	#define read_rsp(rsp) asm volatile("movl %%esp, %0" : "=r"(rsp))
 	#define push_callee_saved() asm volatile("pushal")
 	#define pop_callee_saved() asm volatile("popal")
+#else
+	#error
 #endif
 
 #include <stdint.h>

kernel/include/kernel/FS/TmpFS/Definitions.h

@@ -28,7 +28,7 @@ namespace Kernel
 		// 1x triply indirect
 		BAN::Array<paddr_t, 5> block;
 		static constexpr size_t direct_block_count = 2;
-#elif ARCH(i386)
+#elif ARCH(i686)
 		// 14x direct blocks
 		// 1x singly indirect
 		// 1x doubly indirect

kernel/include/kernel/Memory/Types.h

@@ -4,7 +4,7 @@
 
 #if ARCH(x86_64)
 	#define KERNEL_OFFSET 0xFFFFFFFF80000000
-#elif ARCH(i386)
+#elif ARCH(i686)
 	#define KERNEL_OFFSET 0xC0000000
 #else
 	#error

kernel/include/kernel/Process.h

@@ -199,6 +199,33 @@ namespace Kernel
 		BAN::ErrorOr<void> validate_string_access(const char*);
 		BAN::ErrorOr<void> validate_pointer_access(const void*, size_t);
 
+		uint64_t signal_pending_mask() const
+		{
+			return ((uint64_t)m_signal_pending_mask[1].load() << 32) | m_signal_pending_mask[0].load();
+		}
+
+		void add_pending_signal(uint8_t signal)
+		{
+			ASSERT(signal >= _SIGMIN);
+			ASSERT(signal <= _SIGMAX);
+			ASSERT(signal < 64);
+			if (signal < 32)
+				m_signal_pending_mask[0] |= (uint32_t)1 << signal;
+			else
+				m_signal_pending_mask[1] |= (uint32_t)1 << (signal - 32);
+		}
+
+		void remove_pending_signal(uint8_t signal)
+		{
+			ASSERT(signal >= _SIGMIN);
+			ASSERT(signal <= _SIGMAX);
+			ASSERT(signal < 64);
+			if (signal < 32)
+				m_signal_pending_mask[0] &= ~((uint32_t)1 << signal);
+			else
+				m_signal_pending_mask[1] &= ~((uint32_t)1 << (signal - 32));
+		}
+
 	private:
 		struct ExitStatus
 		{
@@ -226,7 +253,8 @@ namespace Kernel
 		BAN::Vector<Thread*> m_threads;
 
 		BAN::Atomic<vaddr_t> m_signal_handlers[_SIGMAX + 1] { };
-		BAN::Atomic<uint64_t> m_signal_pending_mask { 0 };
+		// This is 2 32 bit values to allow atomicity on 32 targets
+		BAN::Atomic<uint32_t> m_signal_pending_mask[2] { 0, 0 };
 
 		BAN::Vector<BAN::String> m_cmdline;
 		BAN::Vector<BAN::String> m_environ;

kernel/include/kernel/Processor.h

@@ -19,7 +19,7 @@ namespace Kernel
 	using ProcessorID = uint32_t;
 	constexpr ProcessorID PROCESSOR_NONE = 0xFFFFFFFF;
 
-#if ARCH(x86_64) || ARCH(i386)
+#if ARCH(x86_64) || ARCH(i686)
 	class Processor
 	{
 		BAN_NON_COPYABLE(Processor);

kernel/include/kernel/Storage/NVMe/Queue.h

@@ -31,11 +31,13 @@ namespace Kernel
 		uint32_t m_cq_head { 0 };
 		uint16_t m_cq_valid_phase { 1 };
 
-		Semaphore				m_semaphore;
-		SpinLock				m_lock;
-		BAN::Atomic<uint64_t>	m_used_mask			{ 0 };
-		BAN::Atomic<uint64_t>	m_done_mask			{ 0 };
-		volatile uint16_t		m_status_codes[64]	{ };
+		Semaphore			m_semaphore;
+		SpinLock			m_lock;
+		BAN::Atomic<size_t>	m_used_mask			{ 0 };
+		BAN::Atomic<size_t>	m_done_mask			{ 0 };
+		volatile uint16_t	m_status_codes[64]	{ };
+
+		static constexpr size_t m_mask_bits = sizeof(size_t) * 8;
 	};
 
 }

kernel/include/kernel/Thread.h

@@ -79,6 +79,7 @@ namespace Kernel
 		static pid_t current_tid();
 
 		Process& process();
+		const Process& process() const;
 		bool has_process() const { return m_process; }
 
 		bool is_userspace() const { return m_is_userspace; }

kernel/kernel/Process.cpp

@@ -1488,7 +1488,7 @@ namespace Kernel
 
 		if (pid == m_pid)
 		{
-			m_signal_pending_mask |= 1 << signal;
+			add_pending_signal(signal);
 			return 0;
 		}
 
@@ -1501,7 +1501,7 @@ namespace Kernel
 					found = true;
 					if (signal)
 					{
-						process.m_signal_pending_mask |= 1 << signal;
+						process.add_pending_signal(signal);
 						// FIXME: This feels hacky
 						Scheduler::get().unblock_thread(process.m_threads.front()->tid());
 					}

kernel/kernel/Random.cpp

@@ -22,7 +22,7 @@ namespace Kernel
 		{
 #if ARCH(x86_64)
 			asm volatile("rdrand %0" : "=r"(s_rand_seed));
-#elif ARCH(i386)
+#elif ARCH(i686)
 			uint32_t lo, hi;
 			asm volatile(
 				"rdrand %[lo];"

kernel/kernel/Scheduler.cpp

@@ -21,7 +21,7 @@ namespace Kernel
 	{
 #if ARCH(x86_64)
 		asm volatile("movq %0, %%rsp" :: "rm"(Processor::current_stack_top()));
-#elif ARCH(i386)
+#elif ARCH(i686)
 		asm volatile("movl %0, %%esp" :: "rm"(Processor::current_stack_top()));
 #else
 		#error
@@ -221,7 +221,7 @@ namespace Kernel
 				"orq $(1 << 3), %rax;"
 				"movq %rax, %cr0"
 			);
-#elif ARCH(i386)
+#elif ARCH(i686)
 			asm volatile(
 				"movl %cr0, %eax;"
 				"orl $(1 << 3), %eax;"

kernel/kernel/Storage/NVMe/Queue.cpp

@@ -15,8 +15,8 @@ namespace Kernel
 		, m_doorbell(db)
 		, m_qdepth(qdepth)
 	{
-		for (uint32_t i = qdepth; i < 64; i++)
-			m_used_mask |= (uint64_t)1 << i;
+		for (uint32_t i = qdepth; i < m_mask_bits; i++)
+			m_used_mask |= (size_t)1 << i;
 		set_irq(irq);
 		enable_interrupt();
 	}
@@ -29,8 +29,8 @@ namespace Kernel
 		{
 			uint16_t sts = cq_ptr[m_cq_head].sts >> 1;
 			uint16_t cid = cq_ptr[m_cq_head].cid;
-			uint64_t cid_mask = (uint64_t)1 << cid;
-			ASSERT(cid < 64);
+			size_t cid_mask = (size_t)1 << cid;
+			ASSERT(cid < m_mask_bits);
 
 			ASSERT((m_done_mask & cid_mask) == 0);
 
@@ -50,7 +50,7 @@ namespace Kernel
 	uint16_t NVMeQueue::submit_command(NVMe::SubmissionQueueEntry& sqe)
 	{
 		uint16_t cid = reserve_cid();
-		uint64_t cid_mask = (uint64_t)1 << cid;
+		size_t cid_mask = (size_t)1 << cid;
 
 		{
 			SpinLockGuard _(m_lock);
@@ -98,13 +98,13 @@ namespace Kernel
 		}
 
 		uint16_t cid = 0;
-		for (; cid < 64; cid++)
-			if ((m_used_mask & ((uint64_t)1 << cid)) == 0)
+		for (; cid < m_mask_bits; cid++)
+			if ((m_used_mask & ((size_t)1 << cid)) == 0)
 				break;
-		ASSERT(cid < 64);
+		ASSERT(cid < m_mask_bits);
 		ASSERT(cid < m_qdepth);
 
-		m_used_mask |= (uint64_t)1 << cid;
+		m_used_mask |= (size_t)1 << cid;
 
 		m_lock.unlock(state);
 		return cid;

kernel/kernel/Thread.cpp

@@ -111,7 +111,7 @@ namespace Kernel
 		);
 		save_sse();
 		asm volatile("movq %0, %%cr0" :: "r"(cr0));
-	#elif ARCH(i386)
+	#elif ARCH(i686)
 		uintptr_t cr0;
 		asm volatile(
 			"movl %%cr0, %%eax;"
@@ -140,6 +140,12 @@ namespace Kernel
 		return *m_process;
 	}
 
+	const Process& Thread::process() const
+	{
+		ASSERT(m_process);
+		return *m_process;
+	}
+
 	Thread::~Thread()
 	{
 	}
@@ -241,7 +247,7 @@ namespace Kernel
 		auto& interrupt_stack = *reinterpret_cast<InterruptStack*>(interrupt_stack_base() + interrupt_stack_size() - sizeof(InterruptStack));
 		if (!GDT::is_user_segment(interrupt_stack.cs))
 			return false;
-		uint64_t full_pending_mask = m_signal_pending_mask | m_process->m_signal_pending_mask;
+		uint64_t full_pending_mask = m_signal_pending_mask | process().signal_pending_mask();;
 		return full_pending_mask & ~m_signal_block_mask;
 	}
 
@@ -265,7 +271,7 @@ namespace Kernel
 
 		if (signal == 0)
 		{
-			uint64_t full_pending_mask = m_signal_pending_mask | process().m_signal_pending_mask;
+			uint64_t full_pending_mask = m_signal_pending_mask | process().signal_pending_mask();
 			for (signal = _SIGMIN; signal <= _SIGMAX; signal++)
 			{
 				uint64_t mask = 1ull << signal;
@@ -283,7 +289,7 @@ namespace Kernel
 		vaddr_t signal_handler = process().m_signal_handlers[signal];
 
 		m_signal_pending_mask &= ~(1ull << signal);
-		process().m_signal_pending_mask &= ~(1ull << signal);
+		process().remove_pending_signal(signal);
 
 		if (signal_handler == (vaddr_t)SIG_IGN)
 			;

libc/arch/i686/crti.S

@@ -1,4 +1,4 @@
-/* i386 crti.s */
+/* i686 crti.s */
 .section .init
 .global _init
 .type _init, @function

libc/arch/i686/crtn.S

@@ -1,4 +1,4 @@
-/* i386 crtn.s */
+/* i686 crtn.s */
 .section .init
 	/* gcc will nicely put the contents of crtend.o's .init section here. */
 	popl %ebp

libc/icxxabi.cpp

@@ -1,37 +1,41 @@
-#include <icxxabi.h>
+#include <BAN/Assert.h>
+#include <stdint.h>
+#include <stddef.h>
 
 #define ATEXIT_MAX_FUNCS 128
 
 struct atexit_func_entry_t
 {
-	void (*destructor)(void*);
-	void* data;
+	void(*func)(void*);
+	void* arg;
 	void* dso_handle;
 };
 
 static atexit_func_entry_t __atexit_funcs[ATEXIT_MAX_FUNCS];
-static int __atexit_func_count = 0;
+static size_t __atexit_func_count = 0;
 
-int __cxa_atexit(void (*func)(void*), void* data, void* dso_handle)
+extern "C" int __cxa_atexit(void(*func)(void*), void* arg, void* dso_handle)
 {
 	if (__atexit_func_count >= ATEXIT_MAX_FUNCS)
-		return -1;;
-	__atexit_funcs[__atexit_func_count].destructor = func;
-	__atexit_funcs[__atexit_func_count].data = data;
-	__atexit_funcs[__atexit_func_count].dso_handle = dso_handle;
-	__atexit_func_count++;
+		return -1;
+	auto& atexit_func = __atexit_funcs[__atexit_func_count++];
+	atexit_func.func = func;
+	atexit_func.arg = arg;
+	atexit_func.dso_handle = dso_handle;
 	return 0;
 };
 
-void __cxa_finalize(void* func)
+extern "C" void __cxa_finalize(void* f)
 {
-	for (int i = __atexit_func_count - 1; i >= 0; i--)
+	for (size_t i = __atexit_func_count; i > 0; i--)
 	{
-		if (func && func != __atexit_funcs[i].destructor)
+		auto& atexit_func = __atexit_funcs[i - 1];
+		if (atexit_func.func == nullptr)
 			continue;
-		if (__atexit_funcs[i].destructor == nullptr)
-			continue;
-		__atexit_funcs[i].destructor(__atexit_funcs[i].data);
-		__atexit_funcs[i].destructor = nullptr;
+		if (f == nullptr || f == atexit_func.func)
+		{
+			atexit_func.func(atexit_func.arg);
+			atexit_func.func = nullptr;
+		}
 	}
-}
+};