#include "iredges_t.h"
#include "irloop_t.h"
#include "irtools.h"
+#include "return.h"
+
+#include "bearch.h"
+#include "firm/bearch_firm.h"
+#include "ia32/bearch_ia32.h"
#include "be_t.h"
-#include "bechordal_t.h"
-#include "bera.h"
-#include "beifg.h"
-#include "beifg_impl.h"
#include "benumb_t.h"
-#include "besched_t.h"
-#include "belistsched.h"
-#include "belive_t.h"
#include "beutil.h"
-#include "bechordal.h"
-#include "bearch.h"
-#include "becopyoptmain.h"
-#include "becopystat.h"
-#include "bessadestr.h"
#include "benode_t.h"
#include "beirgmod.h"
+#include "besched_t.h"
+#include "belistsched.h"
+#include "belive_t.h"
#include "bespillilp.h"
#include "bespillbelady.h"
-
-#include "firm/bearch_firm.h"
-#include "ia32/bearch_ia32.h"
+#include "bera.h"
+#include "beraextern.h"
+#include "bechordal_t.h"
+#include "beifg.h"
+#include "beifg_impl.h"
+#include "becopyopt.h"
+#include "becopystat.h"
+#include "bessadestr.h"
+#include "beabi.h"
+#include "belower.h"
#define DUMP_INITIAL (1 << 0)
-#define DUMP_SCHED (1 << 1)
-#define DUMP_PREPARED (1 << 2)
-#define DUMP_RA (1 << 3)
-#define DUMP_FINAL (1 << 4)
+#define DUMP_ABI (1 << 1)
+#define DUMP_SCHED (1 << 2)
+#define DUMP_PREPARED (1 << 3)
+#define DUMP_RA (1 << 4)
+#define DUMP_FINAL (1 << 5)
/* options visible for anyone */
-be_options_t be_options = {
+static be_options_t be_options = {
/* ilp server */
"i44pc52.info.uni-karlsruhe.de",
};
/* dump flags */
-static unsigned dump_flags = DUMP_INITIAL | DUMP_SCHED | DUMP_PREPARED | DUMP_RA | DUMP_FINAL;
+static unsigned dump_flags = 2 * DUMP_FINAL - 1;
/* register allocator to use. */
static const be_ra_t *ra = &be_ra_chordal_allocator;
/* back end instruction set architecture to use */
-static const arch_isa_if_t *isa_if = &firm_isa;
-
+static const arch_isa_if_t *isa_if = &ia32_isa_if;
#ifdef WITH_LIBCORE
static lc_opt_entry_t *be_grp_root = NULL;
static const lc_opt_enum_mask_items_t dump_items[] = {
{ "none", 0 },
{ "initial", DUMP_INITIAL },
+ { "abi", DUMP_ABI },
{ "sched", DUMP_SCHED },
{ "prepared", DUMP_PREPARED },
{ "regalloc", DUMP_RA },
/* register allocators */
static const lc_opt_enum_const_ptr_items_t ra_items[] = {
{ "chordal", &be_ra_chordal_allocator },
+ { "external", &be_ra_external_allocator },
{ NULL, NULL }
};
LC_OPT_ENT_ENUM_PTR("ra", "register allocator", &ra_var),
LC_OPT_ENT_ENUM_PTR("isa", "the instruction set architecture", &isa_var),
+#ifdef WITH_ILP
LC_OPT_ENT_STR ("ilp.server", "the ilp server name", be_options.ilp_server, sizeof(be_options.ilp_server)),
LC_OPT_ENT_STR ("ilp.solver", "the ilp solver name", be_options.ilp_solver, sizeof(be_options.ilp_solver)),
+#endif /* WITH_ILP */
{ NULL }
};
{
#ifdef WITH_LIBCORE
int i;
+ lc_opt_entry_t *be_grp_ra;
be_grp_root = lc_opt_get_grp(firm_opt_get_root(), "be");
+ be_grp_ra = lc_opt_get_grp(be_grp_root, "ra");
lc_opt_add_table(be_grp_root, be_main_options);
- /* register register allocator options */
+ /* register allocator options */
for(i = 0; ra_items[i].name != NULL; ++i) {
const be_ra_t *ra = ra_items[i].value;
- ra->register_options(be_grp_root);
+ ra->register_options(be_grp_ra);
}
/* register isa options */
phi_class_init();
}
-static be_main_env_t *be_init_env(be_main_env_t *env, FILE *file_handle)
+static be_main_env_t *be_init_env(be_main_env_t *env)
{
- obstack_init(&env->obst);
- env->dbg = firm_dbg_register("be.main");
-
- env->arch_env = obstack_alloc(&env->obst, sizeof(env->arch_env[0]));
- arch_env_init(env->arch_env, isa_if, file_handle);
-
- /* Register the irn handler of the architecture */
- if (arch_isa_get_irn_handler(env->arch_env->isa))
- arch_env_add_irn_handler(env->arch_env, arch_isa_get_irn_handler(env->arch_env->isa));
-
- /*
- * Register the node handler of the back end infrastructure.
- * This irn handler takes care of the platform independent
- * spill, reload and perm nodes.
- */
- env->node_factory = obstack_alloc(&env->obst, sizeof(*env->node_factory));
- be_node_factory_init(env->node_factory, env->arch_env->isa);
- arch_env_add_irn_handler(env->arch_env, be_node_get_irn_handler(env->node_factory));
-
- return env;
+ memset(env, 0, sizeof(*env));
+ obstack_init(&env->obst);
+ env->dbg = firm_dbg_register("be.main");
+ env->arch_env = obstack_alloc(&env->obst, sizeof(env->arch_env[0]));
+ env->options = &be_options;
+
+ arch_env_init(env->arch_env, isa_if);
+
+ /* Register the irn handler of the architecture */
+ if (arch_isa_get_irn_handler(env->arch_env->isa))
+ arch_env_push_irn_handler(env->arch_env, arch_isa_get_irn_handler(env->arch_env->isa));
+
+ /*
+ * Register the node handler of the back end infrastructure.
+ * This irn handler takes care of the platform independent
+ * spill, reload and perm nodes.
+ */
+ arch_env_push_irn_handler(env->arch_env, &be_node_irn_handler);
+ env->phi_handler = be_phi_handler_new(env->arch_env);
+ arch_env_push_irn_handler(env->arch_env, env->phi_handler);
+
+ return env;
}
static void be_done_env(be_main_env_t *env)
{
env->arch_env->isa->impl->done(env->arch_env->isa);
+ be_phi_handler_free(env->phi_handler);
obstack_free(&env->obst, NULL);
}
void (*dumper)(ir_graph *, const char *))
{
if(dump_flags & mask)
- dumper(irg, suffix);
+ be_dump(irg, suffix, dumper);
}
-static void prepare_graph(be_main_env_t *s, ir_graph *irg)
+/**
+ * Prepare a backend graph for code generation.
+ */
+static void prepare_graph(be_irg_t *birg)
{
+ ir_graph *irg = birg->irg;
+
/* Normalize proj nodes. */
normalize_proj_nodes(irg);
+ /* Make just one return node. */
+ normalize_one_return(irg);
+
/* Remove critical edges */
remove_critical_cf_edges(irg);
/* Ensure, that the ir_edges are computed. */
edges_activate(irg);
- /* Compute loop nesting information (for weighting copies) */
- if (get_irg_loopinfo_state(irg) != (loopinfo_valid & loopinfo_cf_consistent))
- construct_cf_backedges(irg);
-
+ /* check, if the dominance property is fulfilled. */
be_check_dominance(irg);
+
+ /* reset the phi handler. */
+ be_phi_handler_reset(birg->main_env->phi_handler);
}
+/**
+ * The Firm backend main loop.
+ * Do architecture specific lowering for all graphs
+ * and call the architecture specific code generator.
+ */
static void be_main_loop(FILE *file_handle)
{
int i, n;
arch_isa_t *isa;
be_main_env_t env;
- be_init_env(&env, file_handle);
+ be_init_env(&env);
isa = arch_env_get_isa(env.arch_env);
/* For all graphs */
for(i = 0, n = get_irp_n_irgs(); i < n; ++i) {
ir_graph *irg = get_irp_irg(i);
+ const arch_code_generator_if_t *cg_if;
+ be_irg_t birg;
- arch_code_generator_t *cg;
+ birg.irg = irg;
+ birg.main_env = &env;
DBG((env.dbg, LEVEL_2, "====> IRG: %F\n", irg));
dump(DUMP_INITIAL, irg, "-begin", dump_ir_block_graph);
/* set the current graph (this is important for several firm functions) */
- current_ir_graph = irg;
+ current_ir_graph = birg.irg;
+
+ /* Get the code generator interface. */
+ cg_if = isa->impl->get_code_generator_if(isa);
/* get a code generator for this graph. */
- cg = arch_isa_make_code_generator(isa, irg);
+ birg.cg = cg_if->init(file_handle, &birg);
/* create the code generator and generate code. */
- prepare_graph(&env, irg);
- arch_code_generator_prepare_graph(cg);
+ prepare_graph(&birg);
+
+ /* some transformations need to be done before abi introduce */
+ arch_code_generator_before_abi(birg.cg);
+ /* implement the ABI conventions. */
+ birg.abi = be_abi_introduce(&birg);
+ dump(DUMP_ABI, irg, "-abi", dump_ir_block_graph);
+
+ /* generate code */
+ arch_code_generator_prepare_graph(birg.cg);
+
+ /*
+ * Since the code generator made a lot of new nodes and skipped
+ * a lot of old ones, we should do dead node elimination here.
+ * Note that this requires disabling the edges here.
+ */
edges_deactivate(irg);
dead_node_elimination(irg);
edges_activate(irg);
+ /* Compute loop nesting information (for weighting copies) */
+ construct_cf_backedges(irg);
+
dump(DUMP_PREPARED, irg, "-prepared", dump_ir_block_graph);
+ /* add Keeps for should_be_different constrained nodes */
+ // assure_constraints(&birg);
+ dump(DUMP_PREPARED, irg, "-assured", dump_ir_block_graph);
+
/* Schedule the graphs. */
- arch_code_generator_before_sched(cg);
- list_sched(isa, irg);
+ arch_code_generator_before_sched(birg.cg);
+ list_sched(env.arch_env, irg);
+ /* connect all stack modifying nodes together (see beabi.c) */
+ be_abi_fix_stack_nodes(birg.abi);
dump(DUMP_SCHED, irg, "-sched", dump_ir_block_graph_sched);
/* Verify the schedule */
sched_verify_irg(irg);
/* Do register allocation */
- arch_code_generator_before_ra(cg);
- ra->allocate(&env, irg);
-
+ arch_code_generator_before_ra(birg.cg);
+ ra->allocate(&birg);
dump(DUMP_RA, irg, "-ra", dump_ir_block_graph_sched);
- arch_code_generator_done(cg);
+ arch_code_generator_after_ra(birg.cg);
+ be_abi_fix_stack_bias(birg.abi);
+
+ arch_code_generator_done(birg.cg);
dump(DUMP_FINAL, irg, "-end", dump_ir_block_graph_sched);
+ be_abi_free(birg.abi);
}
be_done_env(&env);
void be_main(FILE *file_handle)
{
- be_main_loop(file_handle);
+ /* never build code for pseudo irgs */
+ set_visit_pseudo_irgs(0);
+
+ be_node_init();
+ be_main_loop(file_handle);
}