- Improved addressmode optimisation for conv nodes

[libfirm] / ir / be / ia32 / ia32_emitter.c

diff --git a/ir/be/ia32/ia32_emitter.c b/ir/be/ia32/ia32_emitter.c
index 0f76d6b..4cd44f1 100644 (file)
--- a/ir/be/ia32/ia32_emitter.c
+++ b/ir/be/ia32/ia32_emitter.c
@@ -1819,7 +1819,7 @@ static void emit_ia32_Const(const ir_node *n, ia32_emit_env_t *env) {
                lc_esnprintf(arg_env, cmd_buf, 256, "mov %1D, OFFSET FLAT:%C ", n, n);
                lc_esnprintf(arg_env, cmnt_buf, 256, "/* Move address of SymConst into register */");
        } else {
-               assert(mode == get_tarval_mode(tv));
+               assert(mode == get_tarval_mode(tv) || (mode_is_reference(get_tarval_mode(tv)) && mode == mode_Iu));
                /* beware: in some rare cases mode is mode_b which has no tarval_null() */
                if (tv == get_tarval_b_false() || tv == get_tarval_null(mode)) {
                        const char *instr = "xor";
@@ -2150,37 +2150,55 @@ static int is_first_loop_block(ir_node *block, ir_node *prev_block, ia32_emit_en
  * and emits code for each node.
  */
 static void ia32_gen_block(ir_node *block, ir_node *last_block, ia32_emit_env_t *env) {
-       ir_graph *irg = get_irn_irg(block);
-       ir_node *start_block = get_irg_start_block(irg);
+       ir_graph      *irg         = get_irn_irg(block);
+       ir_node       *start_block = get_irg_start_block(irg);
+       int           need_label   = 1;
+       FILE          *F           = env->out;
        const ir_node *irn;
-       int need_label = 1;
-       FILE *F = env->out;
+       int           i;
 
        assert(is_Block(block));
 
-       if(block == start_block)
+       if (block == start_block)
                need_label = 0;
 
        if (need_label && get_irn_arity(block) == 1) {
                ir_node *pred_block = get_Block_cfgpred_block(block, 0);
 
-               if(pred_block == last_block && get_irn_n_edges_kind(pred_block, EDGE_KIND_BLOCK) <= 2)
+               if (pred_block == last_block && get_irn_n_edges_kind(pred_block, EDGE_KIND_BLOCK) <= 2)
                        need_label = 0;
        }
 
-       // special case because the start block contains no jump instruction
-       if(last_block == start_block) {
+       /* special case: if one of our cfg preds is a switch-jmp we need a label, */
+       /*               otherwise there might be jump table entries jumping to   */
+       /*               non-existent (omitted) labels                            */
+       for (i = get_Block_n_cfgpreds(block) - 1; i >= 0; --i) {
+               ir_node *pred = get_Block_cfgpred(block, i);
+
+               if (is_Proj(pred)) {
+                       assert(get_irn_mode(pred) == mode_X);
+                       if (is_ia32_SwitchJmp(get_Proj_pred(pred))) {
+                               need_label = 1;
+                               break;
+                       }
+               }
+       }
+
+       /* special case because the start block contains no jump instruction */
+       if (last_block == start_block) {
                const ir_edge_t *edge;
                ir_node *startsucc = NULL;
 
                foreach_block_succ(start_block, edge) {
                        startsucc = get_edge_src_irn(edge);
-                       if(startsucc != start_block)
+                       if (startsucc != start_block)
                                break;
                }
                assert(startsucc != NULL);
 
-               if(startsucc != block) {
+               /* if the last block was the start block and we are not inside the */
+               /* start successor, emit a jump to the start successor             */
+               if (startsucc != block) {
                        char buf[SNPRINTF_BUF_LEN];
                        ir_snprintf(buf, sizeof(buf), BLOCK_PREFIX("%d"),
                                    get_irn_node_nr(startsucc));
@@ -2194,21 +2212,22 @@ static void ia32_gen_block(ir_node *block, ir_node *last_block, ia32_emit_env_t
                int align = 1;
                ir_exec_freq *execfreqs = env->cg->birg->execfreqs;
 
-               // align the loop headers
-               if(!is_first_loop_block(block, last_block, env)) {
-
-                       // align blocks where the previous block has no fallthrough
+               /* align the loop headers */
+               if (! is_first_loop_block(block, last_block, env)) {
+                       /* align blocks where the previous block has no fallthrough */
                        arity = get_irn_arity(block);
-                       for(i = 0; i < arity; ++i) {
+
+                       for (i = 0; i < arity; ++i) {
                                ir_node *predblock = get_Block_cfgpred_block(block, i);
-                               if(predblock == last_block) {
+
+                               if (predblock == last_block) {
                                        align = 0;
                                        break;
                                }
                        }
                }
 
-               if(align)
+               if (align)
                        ia32_emit_align_label(env->out, env->isa->opt_arch);
 
                ir_snprintf(cmd_buf, sizeof(cmd_buf), BLOCK_PREFIX("%d:"),
@@ -2219,11 +2238,12 @@ static void ia32_gen_block(ir_node *block, ir_node *last_block, ia32_emit_env_t
                fprintf(F, "/* preds:");
 
                arity = get_irn_arity(block);
-               for(i = 0; i < arity; ++i) {
+               for (i = 0; i < arity; ++i) {
                        ir_node *predblock = get_Block_cfgpred_block(block, i);
                        fprintf(F, " %ld", get_irn_node_nr(predblock));
                }
-               if(execfreqs != NULL) {
+
+               if (execfreqs != NULL) {
                        fprintf(F, " freq: %f", get_block_execfreq(execfreqs, block));
                }