+2005-07-06 Jeff Law <law@redhat.com>
+
+ * tree-ssa-dce.c (cfg_altered): New global.
+ (tree_dce_init): Initialize cfg_altered.
+ (remove_dead_stmt): If we remove an edge in the CFG, then set
+ CFG_ALTERED.
+ (perform_tree_ssa_dce): If we altered the CFG, then invalidate
+ the dominators.
+
2005-07-06 Kazu Hirata <kazu@codesourcery.com>
* Makefile.in (stamp-collect-ld): Use
processed that it is control dependent on. */
static sbitmap visited_control_parents;
+/* TRUE if this pass alters the CFG (by removing control statements).
+ FALSE otherwise.
+
+ If this pass alters the CFG, then it will arrange for the dominators
+ to be recomputed. */
+static bool cfg_altered;
+
/* Execute CODE for each edge (given number EDGE_NUMBER within the CODE)
for which the block with index N is control dependent. */
#define EXECUTE_IF_CONTROL_DEPENDENT(N, EDGE_NUMBER, CODE) \
/* Remove the remaining the outgoing edges. */
while (!single_succ_p (bb))
- remove_edge (EDGE_SUCC (bb, 1));
+ {
+ /* FIXME. When we remove the edge, we modify the CFG, which
+ in turn modifies the dominator and post-dominator tree.
+ Is it safe to postpone recomputing the dominator and
+ post-dominator tree until the end of this pass given that
+ the post-dominators are used above? */
+ cfg_altered = true;
+ remove_edge (EDGE_SUCC (bb, 1));
+ }
}
FOR_EACH_SSA_DEF_OPERAND (def_p, t, iter, SSA_OP_VIRTUAL_DEFS)
sbitmap_zero (processed);
worklist = VEC_alloc (tree, heap, 64);
+ cfg_altered = false;
}
/* Cleanup after this pass. */
if (aggressive)
free_dominance_info (CDI_POST_DOMINATORS);
+ /* If we removed paths in the CFG, then we need to update
+ dominators as well. I haven't investigated the possibility
+ of incrementally updating dominators. */
+ if (cfg_altered)
+ free_dominance_info (CDI_DOMINATORS);
+
/* Debugging dumps. */
if (dump_file)
print_stats ();