public inbox for gcc-patches@gcc.gnu.org
 help / color / mirror / Atom feed
* [PATCH] tree-optimization/108821 - store motion and volatiles
@ 2023-02-17 11:38 Richard Biener
  0 siblings, 0 replies; only message in thread
From: Richard Biener @ 2023-02-17 11:38 UTC (permalink / raw)
  To: gcc-patches

The following fixes store motion to not re-issue volatile stores
to preserve TBAA behavior since that will result in the number
of volatile accesses changing.

Bootstrapped and tested on x86_64-unknown-linux-gnu, pushed to trunk.

	PR tree-optimization/108821
	* tree-ssa-loop-im.cc (sm_seq_valid_bb): We can also not
	move volatile accesses.

	* gcc.dg/tree-ssa/ssa-lim-24.c: New testcase.
---
 gcc/testsuite/gcc.dg/tree-ssa/ssa-lim-24.c | 25 ++++++++++++++++++++++
 gcc/tree-ssa-loop-im.cc                    |  4 +++-
 2 files changed, 28 insertions(+), 1 deletion(-)
 create mode 100644 gcc/testsuite/gcc.dg/tree-ssa/ssa-lim-24.c

diff --git a/gcc/testsuite/gcc.dg/tree-ssa/ssa-lim-24.c b/gcc/testsuite/gcc.dg/tree-ssa/ssa-lim-24.c
new file mode 100644
index 00000000000..6b463490a91
--- /dev/null
+++ b/gcc/testsuite/gcc.dg/tree-ssa/ssa-lim-24.c
@@ -0,0 +1,25 @@
+/* { dg-do compile } */
+/* { dg-options "-O2 -fdump-tree-lim2-details -fdump-tree-optimized" } */
+
+extern volatile int *x;
+static int gCrc;
+
+static int __attribute__((noinline)) crc16Add(int crc, int b)
+{
+  return crc + b;
+}
+
+void f(int data, int dataSz)
+{
+  int i;
+
+  for(i=0;i<dataSz;i++)
+  {
+    gCrc = crc16Add(gCrc, data);
+    *x = data;
+  }
+}
+
+/* { dg-final { scan-tree-dump "Executing store motion of gCrc" "lim2" } } */
+/* { dg-final { scan-tree-dump-not "Re-issueing" "lim2" } } */
+/* { dg-final { scan-tree-dump-times "\\*x" 1 "optimized" } } */
diff --git a/gcc/tree-ssa-loop-im.cc b/gcc/tree-ssa-loop-im.cc
index 784f6f04e7d..86ce6acb023 100644
--- a/gcc/tree-ssa-loop-im.cc
+++ b/gcc/tree-ssa-loop-im.cc
@@ -2605,7 +2605,9 @@ sm_seq_valid_bb (class loop *loop, basic_block bb, tree vdef,
       if (data->ref == UNANALYZABLE_MEM_ID)
 	return -1;
       /* Stop at memory references which we can't move.  */
-      else if (memory_accesses.refs_list[data->ref]->mem.ref == error_mark_node)
+      else if (memory_accesses.refs_list[data->ref]->mem.ref == error_mark_node
+	       || TREE_THIS_VOLATILE
+		    (memory_accesses.refs_list[data->ref]->mem.ref))
 	{
 	  /* Mark refs_not_in_seq as unsupported.  */
 	  bitmap_ior_into (refs_not_supported, refs_not_in_seq);
-- 
2.35.3

^ permalink raw reply	[flat|nested] only message in thread

only message in thread, other threads:[~2023-02-17 11:38 UTC | newest]

Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2023-02-17 11:38 [PATCH] tree-optimization/108821 - store motion and volatiles 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).