benode.c

/**
 * @file   benode.c
 * @date   17.05.2005
 * @author Sebastian Hack
 *
 * Backend node support.
 *
 * This file provdies Perm, Copy, Spill and Reload nodes.
 *
 * Copyright (C) 2005 Universitaet Karlsruhe
 * Released under the GPL
 */

#ifdef HAVE_CONFIG_H
#include "config.h"
#endif

#include <stdlib.h>

#include "obst.h"
#include "set.h"
#include "pmap.h"
#include "util.h"
#include "debug.h"
#include "fourcc.h"

#include "irop_t.h"
#include "irmode_t.h"
#include "irnode_t.h"
#include "ircons_t.h"
#include "irprintf.h"

#include "be_t.h"
#include "belive_t.h"
#include "besched_t.h"
#include "benode_t.h"

#include "beirgmod.h"

#define DBG_LEVEL 0

#define BENODE_MAGIC FOURCC('B', 'E', 'N', 'O')

typedef enum _node_kind_t {
	node_kind_spill,
	node_kind_reload,
	node_kind_perm,
	node_kind_copy,
	node_kind_last
} node_kind_t;

typedef struct {
	node_kind_t kind;
	const arch_register_class_t *cls;
	ir_op *op;
	int n_pos;
	int *pos;
} be_op_t;

typedef struct {
	const arch_register_t *reg;
	arch_register_req_t   req;
} be_reg_data_t;

typedef struct {
	unsigned      magic;
	const be_op_t *op;
	int           n_regs;
	be_reg_data_t reg_data[1];
} be_node_attr_t;

typedef struct {
	be_node_attr_t attr;
	ir_node *spill_ctx;  /**< The node in whose context this spill was introduced. */
	unsigned offset;     /**< The offset of the memory location the spill writes to
						   in the spill area. */
} be_spill_attr_t;

static int templ_pos_Spill[] = {
	0
};

static int templ_pos_Reload[] = {
	-1
};

static int templ_pos_Copy[] = {
	0, -1
};

static int dump_node(ir_node *irn, FILE *f, dump_reason_t reason);

static const ir_op_ops be_node_ops = {
	NULL,
	NULL,
	NULL,
	NULL,
	NULL,
	NULL,
	NULL,
	NULL,
	NULL,
	NULL,
	NULL,
	dump_node,
	NULL
};

static INLINE int is_be_node(const ir_node *irn)
{
	const be_node_attr_t *attr = (const be_node_attr_t *) &irn->attr;
	return attr->magic == BENODE_MAGIC;
}

static INLINE int is_be_kind(const ir_node *irn, node_kind_t kind)
{
	const be_node_attr_t *a = (const be_node_attr_t *) &irn->attr;
	return a->magic == BENODE_MAGIC && a->op && a->op->kind == kind;
}

static INLINE void *get_attr_and_check(ir_node *irn, node_kind_t kind)
{
	is_be_kind(irn, kind);
	return &irn->attr;
}

static be_node_attr_t *init_node_attr(ir_node *irn,
									  const be_op_t *op,
									  const arch_register_class_t *cls,
									  int n_regs)

{
	be_node_attr_t *attr = (be_node_attr_t *) &irn->attr;
	int i;

	attr->magic   = BENODE_MAGIC;
	attr->n_regs  = n_regs;
	attr->op      = op;

	for(i = 0; i < n_regs; ++i) {
		be_reg_data_t *rd = attr->reg_data + i;

		rd->reg      = NULL;
		rd->req.cls  = cls;
		rd->req.type = arch_register_req_type_normal;
	}

	return attr;
}

#define ARRSIZE(x) (sizeof(x) / sizeof(x[0]))

static int cmp_op_map(const void *a, const void *b, size_t size)
{
	const be_op_t *x = a;
	const be_op_t *y = b;

	return !(x->kind == y->kind && x->cls == y->cls);
}

static be_op_t *get_op(const be_node_factory_t *fact,
		const arch_register_class_t *cls, node_kind_t kind)
{
	be_op_t templ;

	templ.kind = kind;
	templ.cls = cls;

	return set_insert(fact->ops, &templ, sizeof(templ),
		HASH_PTR(cls) + 7 * kind);
}

ir_node *new_Spill(const be_node_factory_t *factory,
		const arch_register_class_t *cls,
		ir_graph *irg, ir_node *bl, ir_node *node_to_spill, ir_node *ctx)
{
	be_spill_attr_t *a;
	ir_node *irn;
	ir_node *in[1];
	be_op_t *bop = get_op(factory, cls, node_kind_spill);
	ir_op *op    = bop->op;

	assert(op && "Spill opcode must be present for this register class");
	in[0]        = node_to_spill;
	irn          = new_ir_node(NULL, irg, bl, op, mode_M, 1, in);
	a            = (be_spill_attr_t *) init_node_attr(irn, bop, cls, 0);
	a->spill_ctx = ctx;
	a->offset    = 0;

	return irn;
}

void set_Spill_offset(ir_node *irn, unsigned offset)
{
	be_spill_attr_t *a = (be_spill_attr_t *) &irn->attr;
	assert(is_be_kind(irn, node_kind_spill));
	a->offset = offset;
}

static ir_node *find_a_spill_walker(ir_node *irn, unsigned visited_nr)
{
	if(get_irn_visited(irn) < visited_nr) {
		set_irn_visited(irn, visited_nr);

		if(is_Phi(irn)) {
			int i, n;
			for(i = 0, n = get_irn_arity(irn); i < n; ++i) {
				ir_node *n = find_a_spill_walker(get_irn_n(irn, i), visited_nr);
				if(n != NULL)
					return n;
			}
		}

		else if(is_be_kind(irn, node_kind_spill))
			return irn;
	}

	return NULL;
}

ir_node *get_Spill_context(const ir_node *irn) {
	be_spill_attr_t *a = (be_spill_attr_t *) &irn->attr;
	assert(is_be_kind(irn, node_kind_spill));
	return a->spill_ctx;
}

/**
 * Finds a spill for a reload.
 * If the reload is directly using the spill, this is simple,
 * else we perform DFS from the reload (over all PhiMs) and return
 * the first spill node we find.
 */
static INLINE ir_node *find_a_spill(ir_node *irn)
{
	ir_graph *irg       = get_irn_irg(irn);
	unsigned visited_nr = get_irg_visited(irg) + 1;

	assert(is_be_kind(irn, node_kind_reload));
	set_irg_visited(irg, visited_nr);
	return find_a_spill_walker(irn, visited_nr);
}


unsigned get_Spill_offset(ir_node *irn)
{
	be_node_attr_t *a = (be_node_attr_t *) &irn->attr;
	assert(is_be_node(irn));

	switch(a->op->kind) {
	case node_kind_reload:
		assert(0 && "not yet implemented");
		return get_Spill_offset(find_a_spill(irn));
	case node_kind_spill:
		return ((be_spill_attr_t *) a)->offset;
	default:
		assert(0 && "Illegal node kind (spill/reload required)");
	}

	return 0;
}

ir_node *new_Reload(const be_node_factory_t *factory,
		const arch_register_class_t *cls, ir_graph *irg,
		ir_node *bl, ir_mode *mode, ir_node *spill_node)
{
	ir_node *irn, *in[1];
	be_op_t *bop = get_op(factory, cls, node_kind_reload);
	ir_op *op    = bop->op;

	assert(op && "Reload opcode must be present for this register class");
	in[0] = spill_node;

	irn  = new_ir_node(NULL, irg, bl, op, mode, 1, in);
	init_node_attr(irn, bop, cls, 1);

	return irn;
}

ir_node *new_Perm(const be_node_factory_t *factory,
		const arch_register_class_t *cls,
		ir_graph *irg, ir_node *bl, int arity, ir_node **in)
{
	ir_node *irn;
	be_op_t *bop = get_op(factory, cls, node_kind_perm);
	ir_op *op    = bop->op;

	irn  = new_ir_node(NULL, irg, bl, op, mode_T, arity, in);
	init_node_attr(irn, bop, cls, arity);

	return irn;
}

ir_node *new_Copy(const be_node_factory_t *factory,
		const arch_register_class_t *cls,
		ir_graph *irg, ir_node *bl, ir_node *in)
{
	ir_node *irn, *ins[1];
	be_op_t *bop = get_op(factory, cls, node_kind_copy);
	ir_op *op    = bop->op;

	ins[0] = in;

	irn  = new_ir_node(NULL, irg, bl, op, get_irn_mode(in), 1, ins);
	init_node_attr(irn, bop, cls, 1);

	return irn;
}

ir_node *be_spill(
		const be_node_factory_t *factory,
		const arch_env_t *arch_env,
		ir_node *irn, ir_node *ctx)
{
	const arch_register_class_t *cls = arch_get_irn_reg_class(arch_env, irn, -1);

	ir_node *bl    = get_nodes_block(irn);
	ir_graph *irg  = get_irn_irg(bl);
	ir_node *spill = new_Spill(factory, cls, irg, bl, irn, ctx);
	ir_node *insert;

	/*
	 * search the right insertion point. a spill of a phi cannot be put
	 * directly after the phi, if there are some phis behind the one which
	 * is spilled.
	 */
	insert = sched_next(irn);
	while(is_Phi(insert) && !sched_is_end(insert))
		insert = sched_next(insert);

	sched_add_before(insert, spill);
	return spill;
}

ir_node *be_reload(const be_node_factory_t *factory,
					 const arch_env_t *arch_env,
					 const arch_register_class_t *cls,
					 ir_node *irn, int pos, ir_mode *mode, ir_node *spill)
{
	ir_node *reload;

	ir_node *bl   = get_nodes_block(irn);
	ir_graph *irg = get_irn_irg(bl);

	assert(be_is_Spill(spill)
			|| (is_Phi(spill) && get_irn_mode(spill) == mode_M));

	reload = new_Reload(factory, cls, irg, bl, mode, spill);

	set_irn_n(irn, pos, reload);
	sched_add_before(irn, reload);
	return reload;
}

/**
 * If the node is a proj, reset the node to the proj's target and return
 * the proj number.
 * @param node The address of a node pointer.
 * @param def  A default value.
 * @return     If *node is a Proj, *node is set to the Proj's target and
 *             the Proj number is returned. Else *node remains the same and @p def
 *             is returned.
 */
static int redir_proj(const ir_node **node, int def)
{
	const ir_node *n = *node;

	if(is_Proj(n)) {
		*node = get_Proj_pred(n);
		def = -(get_Proj_proj(n) + 1);
	}

	return def;
}

static const arch_register_req_t *
be_node_get_irn_reg_req(const arch_irn_ops_t *_self,
		arch_register_req_t *req, const ir_node *irn, int pos)
{
	/* We cannot get output requirements for tuple nodes. */
	if(get_irn_mode(irn) == mode_T && pos < 0)
		return NULL;

	/*
	 * if we're interested in an output operand (pos < 0), so let's resolve projs.
	 */
	if(pos < 0)
		pos = redir_proj((const ir_node **) &irn, pos);

	if(is_be_node(irn)) {
		const be_node_attr_t *a = (const be_node_attr_t *) &irn->attr;
		const be_op_t *bo       = a->op;
		int i;

		for(i = 0; i < bo->n_pos; ++i) {
			if(pos == bo->pos[i]) {

				/* be nodes have no input constraints.
				   so return normal register requirements. */
				if(pos >= 0) {
					req->cls = bo->cls;
					req->type = arch_register_req_type_normal;
				}

				/*
				 * if an output requirement is requested,
				 * return the one stored in the node.
				 */
				else
					*req = a->reg_data[-pos - 1].req;

				return req;
			}
		}
	}

	return NULL;
}

void be_set_Perm_out_req(ir_node *irn, int pos, const arch_register_req_t *req)
{
	be_node_attr_t *a = get_attr_and_check(irn, node_kind_perm);
	assert(pos >= 0 && pos < get_irn_arity(irn) && "position out of range");
	a->reg_data[pos].req = *req;
}

void
be_node_set_irn_reg(const arch_irn_ops_t *_self, ir_node *irn,
		const arch_register_t *reg)
{
	int pos;
	be_op_t *bo;
	be_node_attr_t *attr;
	const be_node_factory_t *factory =
		container_of(_self, const be_node_factory_t, irn_ops);

	if(get_irn_mode(irn) == mode_T)
		return;

	pos = redir_proj((const ir_node **) &irn, -1);
	bo = pmap_get(factory->irn_op_map, get_irn_op(irn));

	if(!bo)
		return;

	attr = (be_node_attr_t *) &irn->attr;
	attr->reg_data[-pos - 1].reg = reg;
}

const arch_register_t *
be_node_get_irn_reg(const arch_irn_ops_t *_self, const ir_node *irn)
{
	int i, pos;
	be_op_t *bo;
	const be_node_factory_t *factory =
		container_of(_self, const be_node_factory_t, irn_ops);

	if(get_irn_mode(irn) == mode_T)
		return NULL;

	pos = redir_proj((const ir_node **) &irn, -1);
	bo = pmap_get(factory->irn_op_map, get_irn_op(irn));

	if(!bo)
		return NULL;

	for(i = 0; i < bo->n_pos; ++i) {
		if(bo->pos[i] == pos) {
			be_node_attr_t *attr = (be_node_attr_t *) &irn->attr;
			return attr->reg_data[-pos - 1].reg;
		}
	}

	return NULL;
}

arch_irn_class_t be_node_classify(const arch_irn_ops_t *_self, const ir_node *irn)
{
	const be_node_factory_t *factory = container_of(_self, const be_node_factory_t, irn_ops);

	be_op_t *bo;
	int idx;

	idx = redir_proj(&irn, 0);
	bo = pmap_get(factory->irn_op_map, get_irn_op(irn));

	switch(bo->kind) {
#define XXX(a) case node_kind_ ## a: return arch_irn_class_ ## a;
		XXX(spill)
		XXX(reload)
		XXX(perm)
		XXX(copy)
#undef XXX
		default:
		return 0;
	}

	return 0;
}

arch_irn_class_t be_node_get_flags(const arch_irn_ops_t *_self, const ir_node *irn)
{
	return 0;
}

static const arch_irn_ops_t *
be_node_get_irn_ops(const arch_irn_handler_t *_self, const ir_node *irn)
{
	be_op_t *bo;
	const be_node_factory_t *factory =
		container_of(_self, const be_node_factory_t, handler);

	redir_proj(&irn, 0);
	bo = pmap_get(factory->irn_op_map, get_irn_op(irn));

	return bo ? &factory->irn_ops : NULL;
}

const arch_irn_handler_t *be_node_get_irn_handler(const be_node_factory_t *f)
{
	return &f->handler;
}

int be_is_Spill(const ir_node *irn)
{
	return is_be_kind(irn, node_kind_spill);
}

int be_is_Reload(const ir_node *irn)
{
	return is_be_kind(irn, node_kind_reload);
}

int be_is_Copy(const ir_node *irn)
{
	return is_be_kind(irn, node_kind_copy);
}

int be_is_Perm(const ir_node *irn)
{
	return is_be_kind(irn, node_kind_perm);
}

be_node_factory_t *be_node_factory_init(be_node_factory_t *factory, const arch_isa_t *isa)
{
	int i, j, n;

	factory->ops = new_set(cmp_op_map, 64);
	factory->irn_op_map = pmap_create();
	obstack_init(&factory->obst);

	factory->handler.get_irn_ops = be_node_get_irn_ops;

	factory->irn_ops.get_irn_reg_req = be_node_get_irn_reg_req;
	factory->irn_ops.set_irn_reg     = be_node_set_irn_reg;
	factory->irn_ops.get_irn_reg     = be_node_get_irn_reg;
	factory->irn_ops.classify        = be_node_classify;
	factory->irn_ops.get_flags       = be_node_get_flags;

	for(i = 0, n = arch_isa_get_n_reg_class(isa); i < n; ++i) {
		const arch_register_class_t *cls = arch_isa_get_reg_class(isa, i);
		be_op_t *ent;

		ent = get_op(factory, cls, node_kind_spill);
		ent->op = new_ir_op(get_next_ir_opcode(), "Spill", op_pin_state_pinned,
				0, oparity_unary, 0, sizeof(be_spill_attr_t), &be_node_ops);
		ent->n_pos = ARRSIZE(templ_pos_Spill);
		ent->pos = templ_pos_Spill;
		pmap_insert(factory->irn_op_map, ent->op, ent);

		ent = get_op(factory, cls, node_kind_reload);
		ent->op = new_ir_op(get_next_ir_opcode(), "Reload", op_pin_state_pinned, 0,
				oparity_unary, 0, sizeof(be_node_attr_t), &be_node_ops);
		ent->n_pos = ARRSIZE(templ_pos_Reload);
		ent->pos = templ_pos_Reload;
		pmap_insert(factory->irn_op_map, ent->op, ent);

		ent = get_op(factory, cls, node_kind_copy);
		ent->op = new_ir_op(get_next_ir_opcode(), "Copy", op_pin_state_pinned, 0,
				oparity_unary, 0, sizeof(be_node_attr_t), &be_node_ops);
		ent->n_pos = ARRSIZE(templ_pos_Copy);
		ent->pos = templ_pos_Copy;
		pmap_insert(factory->irn_op_map, ent->op, ent);

		ent = get_op(factory, cls, node_kind_perm);
		ent->op = new_ir_op(get_next_ir_opcode(), "Perm", op_pin_state_pinned, 0,
				oparity_variable, 0,
				sizeof(be_node_attr_t)
				+ sizeof(be_reg_data_t) * cls->n_regs, &be_node_ops);
		ent->n_pos = 2 * cls->n_regs;
		ent->pos = obstack_alloc(&factory->obst, sizeof(ent->pos[0]) * ent->n_pos);
		for(j = 0; j < ent->n_pos; j += 2) {
			int k = j / 2;
			ent->pos[j] = k;
			ent->pos[j + 1] = -(k + 1);
		}
		pmap_insert(factory->irn_op_map, ent->op, ent);

	}

	return factory;
}

static int dump_node(ir_node *irn, FILE *f, dump_reason_t reason)
{
	be_node_attr_t *at = (be_node_attr_t *) &irn->attr;
	const be_op_t *bo;
	int i;

	assert(is_be_node(irn));
	bo = at->op;

	switch(reason) {
		case dump_node_opcode_txt:
			fprintf(f, get_op_name(bo->op));
			break;
		case dump_node_mode_txt:
			fprintf(f, get_mode_name(get_irn_mode(irn)));
			break;
		case dump_node_nodeattr_txt:
			fprintf(f, "%s ", bo->cls->name);
			break;
		case dump_node_info_txt:
			for(i = 0; i < at->n_regs; ++i) {
				const arch_register_t *reg = at->reg_data[i].reg;
				fprintf(f, "reg #%d: %s\n", i, reg ? reg->name : "n/a");
			}

			if(bo->kind == node_kind_spill) {
				be_spill_attr_t *a = (be_spill_attr_t *) at;
				ir_fprintf(f, "spill context: %+F\n", a->spill_ctx);
			}
			break;
	}

	return 0;
}

ir_node *insert_Perm_after(const be_main_env_t *env,
							 const arch_register_class_t *cls,
							 dom_front_info_t *dom_front,
							 ir_node *pos)
{
	const arch_env_t *arch_env  = env->arch_env;
	ir_node *bl                 = is_Block(pos) ? pos : get_nodes_block(pos);
	ir_graph *irg               = get_irn_irg(bl);
	pset *live                  = pset_new_ptr_default();
	firm_dbg_module_t *dbg      = firm_dbg_register("be.node");

	irn_live_t *li;
	ir_node *curr, *irn, *perm, **nodes;
	int i, n;

	DBG((dbg, LEVEL_1, "Insert Perm after: %+F\n", pos));


	live_foreach(bl, li) {
		ir_node *irn = (ir_node *) li->irn;
		if(live_is_end(li) && arch_irn_has_reg_class(arch_env, irn, -1, cls))
			pset_insert_ptr(live, irn);
	}

	sched_foreach_reverse(bl, irn) {
		ir_node *x;

	/*
	 * If we encounter the node we want to insert the Perm after,
	 * exit immediately, so that this node is still live
	 */
		if(irn == pos)
			break;

		DBG((dbg, LEVEL_1, "%+F\n", irn));
		for(x = pset_first(live); x; x = pset_next(live))
			DBG((dbg, LEVEL_1, "\tlive: %+F\n", x));

		if(arch_irn_has_reg_class(arch_env, irn, -1, cls))
			pset_remove_ptr(live, irn);

		for(i = 0, n = get_irn_arity(irn); i < n; ++i) {
			ir_node *op = get_irn_n(irn, i);

			if(arch_irn_has_reg_class(arch_env, op, -1, cls))
				pset_insert_ptr(live, op);
		}
	}

	n = pset_count(live);
	nodes = malloc(n * sizeof(nodes[0]));

	DBG((dbg, LEVEL_1, "live:\n"));
	for(irn = pset_first(live), i = 0; irn; irn = pset_next(live), i++) {
		DBG((dbg, LEVEL_1, "\t%+F\n", irn));
		nodes[i] = irn;
	}

	perm = new_Perm(env->node_factory, cls, irg, bl, n, nodes);
	sched_add_after(pos, perm);
	free(nodes);

	curr = perm;
	for(i = 0; i < n; ++i) {
		ir_node *copies[1];
		ir_node *perm_op = get_irn_n(perm, i);
		const arch_register_t *reg = arch_get_irn_register(arch_env, perm_op);

		ir_mode *mode = get_irn_mode(perm_op);
		ir_node *proj = new_r_Proj(irg, bl, perm, mode, i);
		arch_set_irn_register(arch_env, proj, reg);

		sched_add_after(curr, proj);
		curr = proj;

		copies[0] = proj;
		be_introduce_copies(dom_front, perm_op, array_size(copies), copies);
	}
	return perm;
}