public inbox for gcc-cvs@sourceware.org
help / color / mirror / Atom feed
* [gcc r11-3726] tree-optimization/97330 - fix bad load sinking
@ 2020-10-08 12:22 Richard Biener
0 siblings, 0 replies; only message in thread
From: Richard Biener @ 2020-10-08 12:22 UTC (permalink / raw)
To: gcc-cvs
https://gcc.gnu.org/g:629e0547af33221a925f38757b52a9284148b68a
commit r11-3726-g629e0547af33221a925f38757b52a9284148b68a
Author: Richard Biener <rguenther@suse.de>
Date: Thu Oct 8 13:18:57 2020 +0200
tree-optimization/97330 - fix bad load sinking
This fixes bad placement of sunk loads.
2020-10-08 Richard Biener <rguenther@suse.de>
PR tree-optimization/97330
* tree-ssa-sink.c (statement_sink_location): Avoid skipping
PHIs when they dominate the insert location.
* gcc.dg/torture/pr97330-1.c: New testcase.
* gcc.dg/torture/pr97330-2.c: Likewise.
Diff:
---
gcc/testsuite/gcc.dg/torture/pr97330-1.c | 19 +++++++++++++++++++
gcc/testsuite/gcc.dg/torture/pr97330-2.c | 14 ++++++++++++++
gcc/tree-ssa-sink.c | 4 +++-
3 files changed, 36 insertions(+), 1 deletion(-)
diff --git a/gcc/testsuite/gcc.dg/torture/pr97330-1.c b/gcc/testsuite/gcc.dg/torture/pr97330-1.c
new file mode 100644
index 00000000000..7dce5bd90d4
--- /dev/null
+++ b/gcc/testsuite/gcc.dg/torture/pr97330-1.c
@@ -0,0 +1,19 @@
+/* { dg-do compile } */
+
+typedef int a;
+typedef char b;
+int c;
+void d(e, f, dst, g, avail, h) int e;
+b *f, *dst;
+a g, avail;
+int h;
+{
+ b i = *f;
+ if (e)
+ goto j;
+ while (avail) {
+ *dst = i;
+ j:
+ avail -= c;
+ }
+}
diff --git a/gcc/testsuite/gcc.dg/torture/pr97330-2.c b/gcc/testsuite/gcc.dg/torture/pr97330-2.c
new file mode 100644
index 00000000000..a0644831cfd
--- /dev/null
+++ b/gcc/testsuite/gcc.dg/torture/pr97330-2.c
@@ -0,0 +1,14 @@
+/* { dg-do compile } */
+
+int a, b, d;
+char c, e;
+void f(void) {
+ char g = c;
+ if (b)
+ goto h;
+ while (d) {
+ e = c;
+ h:
+ d -= a;
+ }
+}
diff --git a/gcc/tree-ssa-sink.c b/gcc/tree-ssa-sink.c
index ba8e5577e9c..207aae2818a 100644
--- a/gcc/tree-ssa-sink.c
+++ b/gcc/tree-ssa-sink.c
@@ -391,8 +391,10 @@ statement_sink_location (gimple *stmt, basic_block frombb,
if (gimple_code (use_stmt) == GIMPLE_PHI)
{
/* In case the PHI node post-dominates the current insert location
- we can disregard it. */
+ we can disregard it. But make sure it is not dominating
+ it as well as can happen in a CFG cycle. */
if (commondom != bb
+ && !dominated_by_p (CDI_DOMINATORS, commondom, bb)
&& dominated_by_p (CDI_POST_DOMINATORS, commondom, bb))
continue;
bb = EDGE_PRED (bb, PHI_ARG_INDEX_FROM_USE (use_p))->src;
^ permalink raw reply [flat|nested] only message in thread
only message in thread, other threads:[~2020-10-08 12:22 UTC | newest]
Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-10-08 12:22 [gcc r11-3726] tree-optimization/97330 - fix bad load sinking Richard Biener
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).