Kernel: Use syscall/sysret for syscalls in x86_64

kernel/arch/x86_64/Syscall.S

@@ -1,50 +1,26 @@
-// arguments in RAX, RBX, RCX, RDX, RSI, RDI
-// System V ABI: RDI, RSI, RDX, RCX, R8, R9
 .global asm_syscall_handler
 asm_syscall_handler:
 	swapgs
-	pushq %rbx
-	pushq %rcx
-	pushq %rdx
-	pushq %rdi
-	pushq %rsi
-	pushq %rbp
-	pushq %r8
-	pushq %r9
-	pushq %r10
+
+	movq %rsp, %rax
+	movq %gs:8, %rsp
+
+	pushq $(0x20 | 3)
+	pushq %rax
 	pushq %r11
-	pushq %r12
-	pushq %r13
-	pushq %r14
-	pushq %r15
-	cld
+	pushq $(0x28 | 3)
+	pushq %rcx
+	subq $8, %rsp
 
-	movq %rsi, %r8
-	movq %rdi, %r9
-	movq %rax, %rdi
-	movq %rbx, %rsi
-	xchgq %rcx, %rdx
-	leaq 112(%rsp), %rbx
-	pushq %rbx
+	movq %r10, %rcx
 	call cpp_syscall_handler
-	addq $8, %rsp
 
-	popq %r15
-	popq %r14
-	popq %r13
-	popq %r12
-	popq %r11
-	popq %r10
-	popq %r9
-	popq %r8
-	popq %rbp
-	popq %rsi
-	popq %rdi
-	popq %rdx
-	popq %rcx
-	popq %rbx
+	movq  8(%rsp), %rcx
+	movq 24(%rsp), %r11
+	movq 32(%rsp), %rsp
+
 	swapgs
-	iretq
+	sysretq
 
 .global sys_fork_trampoline
 sys_fork_trampoline:
@@ -57,7 +33,7 @@ sys_fork_trampoline:
 
 	call read_ip
 	testq %rax, %rax
-	je .reload_stack
+	je .done
 
 	movq %rax, %rsi
 	movq %rsp, %rdi
@@ -71,9 +47,3 @@ sys_fork_trampoline:
 	popq %rbp
 	popq %rbx
 	ret
-
- .reload_stack:
-	call get_thread_start_sp
-	movq %rax, %rsp
-	xorq %rax, %rax
-	jmp .done

kernel/arch/x86_64/interrupts.S

@@ -1,6 +1,6 @@
 .macro swapgs_if_necessary, n
-	cmpb $0x08, \n(%rsp)
-	je 1f
+	testb $3, \n(%rsp)
+	jz 1f; jnp 1f
 	swapgs
 1:
 .endm