blob: 9f56b392cdd9d2abed37b4e0514b41ad7bef94de (
plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
|
/* PR tree-optimization/80933 - redundant bzero/bcopy calls not eliminated
{ dg-do compile }
{ dg-options "-O2 -fdump-tree-dse1" } */
void sink (void*);
void test_bcopy (const void *s)
{
char d[33];
/* Bcopy is transformed into memmove before DSE runs, so this test
doesn't actually verify that DSE does its job. */
__builtin_bcopy (s, d, sizeof d);
__builtin_bcopy (s, d, sizeof d);
sink (d);
}
void test_bzero (void)
{
char d[33];
__builtin_bzero (d, sizeof d);
__builtin_bzero (d, sizeof d);
sink (d);
}
/* { dg-final { scan-tree-dump-times "builtin_memset" 1 "dse1" } } */
/* Merging the evrp folder into substitute_and_fold_engine shuffled
the order of gimple_fold a bit, so evrp is no longer folding the
memmove inline. This folding is instead done by forwprop. Thus, I
have remmoved the |memmove in the test below as this is not done
until after dse.
What happened was that the propagator engine only called gimple
fold if replace_uses_in() was successful. On the other hand, EVRP
called gimple fold regardless.
If we really care about previous behavior, we could put a call to
gimple ::fold_stmt into evrp_folder::fold_stmt(). */
/* { dg-final { scan-tree-dump-not "builtin_(bcopy|bzero|memcpy)" "dse1" } } */
|
