On 19/09/14 17:19, Jeff Law wrote:
On 09/19/14 10:02, Jiong Wang wrote:
On 19/09/14 16:49, Jeff Law wrote:
Probably. Though I'd be a bit concerned with next_block->next_bb.
Wouldn't it be safer to stash away the relevant basic block prior to the
call to split_edge, then use that saved copy. Something like this
(untested):
basic_block old_dest = live_edge->dest;
next_block = split_edge (live_edge);
/* We create a new basic block. Call df_grow_bb_info to make sure
all data structures are allocated. */
df_grow_bb_info (df_live);
bitmap_and (df_get_live_in (next_block),
df_get_live_out (bb),
df_get_live_in (old_dest));
The idea being we don't want to depend on the precise ordering blocks in
the block chain.
Could you try that and see if it does what you need?
Jeff,
Thanks, verified, it works.
Great. Can you send an updated patchkit for review.
patch attached.
please review, thanks.
gcc/
* shrink-wrap.c (move_insn_for_shrink_wrap): Initialize the live-in of
new created BB as the intersection of live-in from "old_dest" and live-out
from "bb".
diff --git a/gcc/shrink-wrap.c b/gcc/shrink-wrap.c
index fd24135..63deadf 100644
--- a/gcc/shrink-wrap.c
+++ b/gcc/shrink-wrap.c
@@ -217,12 +217,15 @@ move_insn_for_shrink_wrap (basic_block bb, rtx_insn *insn,
if (!df_live)
return false;
+ basic_block old_dest = live_edge->dest;
next_block = split_edge (live_edge);
/* We create a new basic block. Call df_grow_bb_info to make sure
all data structures are allocated. */
df_grow_bb_info (df_live);
- bitmap_copy (df_get_live_in (next_block), df_get_live_out (bb));
+
+ bitmap_and (df_get_live_in (next_block), df_get_live_out (bb),
+ df_get_live_in (old_dest));
df_set_bb_dirty (next_block);
/* We should not split more than once for a function. */
