public inbox for gcc-patches@gcc.gnu.org
 help / color / mirror / Atom feed
From: Andreas Krebbel <krebbel@linux.ibm.com>
To: gcc-patches@gcc.gnu.org
Cc: Andreas Krebbel <krebbel@linux.vnet.ibm.com>
Subject: [PATCH] PR86844: Fix for store merging
Date: Tue, 07 Aug 2018 11:35:00 -0000	[thread overview]
Message-ID: <20180807113526.15097-1-krebbel@linux.ibm.com> (raw)

From: Andreas Krebbel <krebbel@linux.vnet.ibm.com>

Bootstrapped and regtested on s390x and x86_64.

gcc/ChangeLog:

2018-08-07  Andreas Krebbel  <krebbel@linux.ibm.com>

	PR tree-optimization/86844
	* gimple-ssa-store-merging.c (check_no_overlap): Add a check to
	reject overlaps if it has seen a non-constant store in between.

gcc/testsuite/ChangeLog:

2018-08-07  Andreas Krebbel  <krebbel@linux.ibm.com>

	PR tree-optimization/86844
	* gcc.dg/pr86844.c: New test.
---
 gcc/gimple-ssa-store-merging.c |  8 +++++++-
 gcc/testsuite/gcc.dg/pr86844.c | 42 ++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 49 insertions(+), 1 deletion(-)
 create mode 100644 gcc/testsuite/gcc.dg/pr86844.c

diff --git a/gcc/gimple-ssa-store-merging.c b/gcc/gimple-ssa-store-merging.c
index 0ae4581..2abef2e 100644
--- a/gcc/gimple-ssa-store-merging.c
+++ b/gcc/gimple-ssa-store-merging.c
@@ -2401,13 +2401,19 @@ check_no_overlap (vec<store_immediate_info *> m_store_info, unsigned int i,
 		  unsigned HOST_WIDE_INT end)
 {
   unsigned int len = m_store_info.length ();
+  bool seen_group_end_store_p = false;
+
   for (++i; i < len; ++i)
     {
       store_immediate_info *info = m_store_info[i];
       if (info->bitpos >= end)
 	break;
+      if (info->rhs_code != INTEGER_CST)
+	seen_group_end_store_p = true;
       if (info->order < last_order
-	  && (rhs_code != INTEGER_CST || info->rhs_code != INTEGER_CST))
+	  && (rhs_code != INTEGER_CST
+	      || info->rhs_code != INTEGER_CST
+	      || seen_group_end_store_p))
 	return false;
     }
   return true;
diff --git a/gcc/testsuite/gcc.dg/pr86844.c b/gcc/testsuite/gcc.dg/pr86844.c
new file mode 100644
index 0000000..9ef08e9
--- /dev/null
+++ b/gcc/testsuite/gcc.dg/pr86844.c
@@ -0,0 +1,42 @@
+/* { dg-do run } */
+/* { dg-require-effective-target stdint_types } */
+/* { dg-options "-O1 -fstore-merging" } */
+
+#include <stdint.h>
+
+struct foo
+{
+  union
+  {
+    uint32_t u4i;
+
+    struct
+    {
+      uint8_t x;
+      uint8_t y;
+      uint8_t z;
+      uint8_t w;
+    } s;
+  } u;
+  uint8_t v;
+};
+
+void __attribute__((noinline,noclone))
+f (struct foo *a)
+{
+  a->u.u4i = 0;
+  a->u.s.w = 222;
+  a->u.s.y = 129;
+  a->u.s.z = a->v;
+}
+
+int
+main ()
+{
+  struct foo s;
+
+  f (&s);
+
+  if (s.u.s.w != 222)
+    __builtin_abort ();
+}
-- 
2.9.1

             reply	other threads:[~2018-08-07 11:35 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-08-07 11:35 Andreas Krebbel [this message]
2018-08-17 13:50 ` Richard Biener
2018-08-18  9:20   ` Eric Botcazou
2018-08-20 14:30     ` Jeff Law
2018-09-10 14:06       ` Andreas Krebbel
2018-09-10 17:53         ` Jakub Jelinek
2018-09-11 14:06           ` Andreas Krebbel
2018-09-12  8:54             ` [PATCH] Fix store merging (PR tree-optimization/86844) Jakub Jelinek
2018-09-12  9:04               ` Richard Biener
2018-09-12  9:11                 ` Jakub Jelinek
2018-09-12 11:02                   ` Richard Biener

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20180807113526.15097-1-krebbel@linux.ibm.com \
    --to=krebbel@linux.ibm.com \
    --cc=gcc-patches@gcc.gnu.org \
    --cc=krebbel@linux.vnet.ibm.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
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).