Kernel: Improve multithreading support

We can now use arbitary BAN::function<void(...)> as the Thread.
I also implemented multithreading for i386 since it was not done
on the initial multithreading commit.

kernel/arch/i386/Thread.S

@@ -0,0 +1,38 @@
+# uint32_t read_rip()
+.global read_rip
+read_rip:
+	popl %eax
+	jmp *%eax
+
+exit_thread_trampoline:
+	addl $16, %esp
+	popl %eax
+	pushl $0x696969
+	pushl %eax
+	ret
+
+# void start_thread(uint32_t arg0, uint32_t arg1, uint32_t arg2, uint32_t arg3, uint32_t rsp, uint32_t rbp, uint32_t rip)
+.global start_thread
+start_thread:
+	movl %esp, %eax
+	movl 28(%eax), %ecx
+	movl 24(%eax), %ebp
+	movl 20(%eax), %esp
+
+	pushl 16(%eax)
+	pushl 12(%eax)
+	pushl 8(%eax)
+	pushl 4(%eax)
+	pushl $exit_thread_trampoline
+
+	sti
+	jmp *%ecx
+
+# void continue_thread(uint32_t rsp, uint32_t rbp, uint32_t rip)
+.global continue_thread
+continue_thread:
+	movl 12(%esp), %ecx
+	movl 8(%esp), %ebp
+	movl 4(%esp), %esp
+	movl $0, %eax
+	jmp *%ecx

kernel/arch/i386/make.config

@@ -8,4 +8,5 @@ $(ARCHDIR)/boot.o		\
 $(ARCHDIR)/IDT.o		\
 $(ARCHDIR)/MMU.o		\
 $(ARCHDIR)/SpinLock.o	\
+$(ARCHDIR)/Thread.o		\
  

kernel/arch/x86_64/Thread.S

@@ -0,0 +1,32 @@
+# uint64_t read_rip()
+.global read_rip
+read_rip:
+	popq %rax
+	jmp *%rax
+
+.global get_thread_at_exit
+get_thread_at_exit:
+	movq 8(%rdi), %rax
+	ret
+
+exit_thread_trampoline:
+	movq 8(%rsp), %rdi
+	ret
+
+# void start_thread(uint64_t arg0, uint64_t arg1, uint64_t arg2, uint64_t arg3, uint64_t rsp, uint64_t rbp, uint64_t rip)
+.global start_thread
+start_thread:
+	movq 8(%rsp), %rcx
+	movq %r8, %rsp
+	movq %r9, %rbp
+	pushq $exit_thread_trampoline
+	sti
+	jmp *%rcx
+
+# void continue_thread(uint64_t rsp, uint64_t rbp, uint64_t rip)
+.global continue_thread
+continue_thread:
+	movq %rdi, %rsp
+	movq %rsi, %rbp
+	movq $0, %rax
+	jmp *%rdx

kernel/arch/x86_64/make.config

@@ -9,4 +9,5 @@ $(ARCHDIR)/IDT.o		\
 $(ARCHDIR)/interrupts.o	\
 $(ARCHDIR)/MMU.o		\
 $(ARCHDIR)/SpinLock.o	\
+$(ARCHDIR)/Thread.o		\