Il 15/10/2012 14:53, Steven Bosscher ha scritto:
> I think I've shown above that we're all looking at the wrong pass...
I think you have... so we want a patch like this?
Index: df-problems.c
===================================================================
--- df-problems.c (revisione 183719)
+++ df-problems.c (copia locale)
@@ -3480,6 +3485,18 @@ df_note_bb_compute (unsigned int bb_inde
}
}
+ for (use_rec = DF_INSN_UID_EQ_USES (uid); *use_rec; use_rec++)
+ {
+ df_ref use = *use_rec;
+ unsigned int uregno = DF_REF_REGNO (use);
+
+ if (!bitmap_bit_p (live, uregno))
+ {
+ remove_note (insn, find_reg_equal_equiv_note (insn));
+ break;
+ }
+ }
+
if (debug_insn == -1)
{
/* ??? We could probably do better here, replacing dead
Paolo