#include <stddef.h>
#include <stdio.h>
#include <stdlib.h>
#include <lightning.h>

#define NOOP	0
#define LOAD	1
#define STORE	2
#define PUSH	3
#define ADD	4
#define INC	5
#define LT	6
#define J	7
#define JT	8
#define IMM	9
#define END	10
int program[] = {
    IMM,	0,		STORE,		0,
    STORE,	1,/*L:*/	LOAD,		1,
    PUSH,	LOAD,		0,		ADD,
    STORE,	1,		LOAD,		0,
    INC,	STORE,		0,		PUSH,
    IMM,	0x989680,	LT,		JT,
    -18,	END
};

typedef struct thread {
    int	*ss, *bp, *sp;
} thread_t;

typedef int (*pift_t)(thread_t *);

pift_t compile(int *program, int length) {
    int		 *base;
    jit_insn	 *code;
    jit_insn	**labels;
    int		  thread;

    /* random guess about size */
    code = malloc(length * 32 * sizeof(int));

    /* XXX no label information here... */
    labels = malloc(length * sizeof(jit_insn*));

    jit_set_ip(code);

    jit_prolog(1);
    thread = jit_arg_i();
    jit_getarg_i(JIT_V1, thread);	/* v1 = thread */

    for (base = program; length > 0; --length, ++program) {
	labels[program - base] = jit_get_label();
	switch (*program) {
	    case NOOP:
		break;
	    case LOAD:
		++program;
		/* r1 = thread->ss */
		jit_ldxi_i(JIT_R1, JIT_V1, offsetof(thread_t, bp));
		/* r0 = r1[*program] */
		jit_ldxi_i(JIT_R0, JIT_R1, *program * sizeof(int));
		break;
	    case STORE:
		++program;
		/* r1 = thread->ss */
		jit_ldxi_i(JIT_R1, JIT_V1, offsetof(thread_t, bp));
		/* r1[*program] = r0 */
		jit_stxi_i(*program * sizeof(int), JIT_R1, JIT_R0);
		break;
	    case PUSH:
		/* r1 = thread->sp */
		jit_ldxi_i(JIT_R1, JIT_V1, offsetof(thread_t, sp));
		/* r2 = r1 + sizeof(int) */
		jit_addi_i(JIT_R2, JIT_R1, sizeof(int));
		/* thread->sp = r2 */
		jit_stxi_i(offsetof(thread_t, sp), JIT_V1, JIT_R2);
		/* *r1 = r0 */
		jit_stxi_i(0, JIT_R1, JIT_R0);
		break;
	    case ADD:
		/* r1 = thread->sp */
		jit_ldxi_i(JIT_R1, JIT_V1, offsetof(thread_t, sp));
		/* r2 = r1 - sizeof(int) */
		jit_subi_i(JIT_R2, JIT_R1, sizeof(int));
		/* r1 = *r1 */
		jit_ldr_i(JIT_R1, JIT_R1);
		/* r0 = r1 + r0 */
		jit_addr_i(JIT_R0, JIT_R1, JIT_R0);
		/* thread->sp = r2 */
		jit_stxi_i(offsetof(thread_t, sp), JIT_V1, JIT_R2);
		break;
	    case INC:
		/* r0 = r0 + 1 */
		jit_addi_i(JIT_R0, JIT_R0, 1);
		break;
	    case LT:
		/* r1 = thread->sp */
		jit_ldxi_i(JIT_R1, JIT_V1, offsetof(thread_t, sp));
		/* r2 = r1 - sizeof(int) */
		jit_subi_i(JIT_R2, JIT_R1, sizeof(int));
		/* r1 = *r1 */
		jit_ldr_i(JIT_R1, JIT_R1);
		/* r0 = r1 < r0 */
		jit_ltr_i(JIT_R0, JIT_R1, JIT_R0);
		/* thread->sp = r2 */
		jit_stxi_i(offsetof(thread_t, sp), JIT_V1, JIT_R2);
		break;
	    case J:
		/* XXX no forward jumps in sample test */
		++program;
		jit_jmpi(labels[program - base + *program]);
		break;
	    case JT:
		/* XXX no forward jumps in sample test */
		++program;
		jit_bnei_ui(labels[program - base + *program], JIT_R0, 0);
		break;
	    case IMM:
		/* r0 = immediate */
		++program;
		jit_movi_i(JIT_R0, *program);
		break;
	    case END:
		jit_retval(JIT_R0);
		jit_ret();
		jit_flush_code(code, jit_get_ip().ptr);
		return (pift_t)(jit_get_ip().iptr);
	}
    }
    abort();
}

void vm(void) {
    pift_t 	jit;
    thread_t	thread;

    thread.ss = malloc(16 * sizeof(int));
    thread.bp = thread.sp = thread.ss;
    jit = compile(program, sizeof(program) / sizeof(int));
    printf("%d\n", (*jit)(&thread));
}

int
main(int argc, char *argv[])
{
    vm();
    return 0;
}