The existing cast to float gives weird results in the RTL dump files on x86
when the compiler is configured --with-fpmath=sse.
Bootstrapped on x86/Linux, applied on the mainline as obvious.
2021-02-15 Eric Botcazou <ebotca...@adacore.com>
* df-core.c (df_worklist_dataflow_doublequeue): Use proper cast.
--
Eric Botcazou
diff --git a/gcc/df-core.c b/gcc/df-core.c
index 8536e5e1f51..b4407eec460 100644
--- a/gcc/df-core.c
+++ b/gcc/df-core.c
@@ -1064,7 +1064,7 @@ df_worklist_dataflow_doublequeue (struct dataflow *dataflow,
" n_basic_blocks %d n_edges %d"
" count %d (%5.2g)\n",
n_basic_blocks_for_fn (cfun), n_edges_for_fn (cfun),
- dcount, dcount / (float)n_basic_blocks_for_fn (cfun));
+ dcount, dcount / (double)n_basic_blocks_for_fn (cfun));
}
/* Worklist-based dataflow solver. It uses sbitmap as a worklist,