#__clone(func, stack, flags, arg, ptid, tls, ctid)
#         a0,    a1,   a2,    a3,  a4,  a5,   a6
# sys_clone3(struct clone_args *cl_args, size_t size)
#                                 a0             a1

.global	__clone
.hidden __clone
.type	__clone,@function
__clone:
	# Save function pointer and argument pointer on new thread stack
	addi.d	$a1, $a1, -16
	st.d	$a0, $a1, 0	# save function pointer
	st.d	$a3, $a1, 8	# save argument pointer

	li.d	$t0, ~0x004000ff  # mask CSIGNAL and CLONE_DETACHED
	and	$t1, $a2, $t0     # cl_args.flags
	li.d	$t0, 0x000000ff   # CSIGNAL
	and	$t2, $a2, $t0     # cl_args.exit_signal

	bstrins.d $sp, $zero, 2, 0  # align stack to 8 bytes
	addi.d	$sp, $sp, -88   # struct clone_args
	st.d	$t1, $sp, 0     # flags
	st.d	$a4, $sp, 8     # pidfd
	st.d	$a6, $sp, 16    # child_tid
	st.d	$a4, $sp, 24    # parent_tid
	st.d	$t2, $sp, 32    # exit_signal
	st.d	$a1, $sp, 40    # stack
	st.d	$zero, $sp, 48  # stack_size
	st.d	$a5, $sp, 56    # tls
	st.d	$zero, $sp, 64  # set_tid
	st.d	$zero, $sp, 72  # set_tid_size
	st.d	$zero, $sp, 80  # cgroup

	move	$a0, $sp
	li.d	$a1, 88
	li.d	$a7, 435	# __NR_clone3
	syscall 0		# call clone3

	beqz	$a0, 1f		# whether child process
	addi.d	$sp, $sp, 88
	jirl	$zero, $ra, 0	# parent process return
1:
	ld.d	$t8, $sp, 0     # function pointer
	ld.d	$a0, $sp, 8     # argument pointer
	jirl	$ra, $t8, 0     # call the user's function
	li.d	$a7, 93
	syscall	0		# child process exit