This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
PATCH: PR target/45213: "suffix or operands invalid for `push'" triggered by optimisations on x86_64
- From: "H.J. Lu" <hongjiu dot lu at intel dot com>
- To: gcc-patches at gcc dot gnu dot org
- Date: Fri, 6 Aug 2010 15:38:27 -0700
- Subject: PATCH: PR target/45213: "suffix or operands invalid for `push'" triggered by optimisations on x86_64
- Reply-to: "H.J. Lu" <hjl dot tools at gmail dot com>
Hi,
"pushq $imm32S" only takes 32bit signed extended immediate. You can't push
0xbf800000. Instead, you push -1082130432 or 0xffffffffbf800000. This
patch makes it signed. OK for trunk/4.5/4.4?
Thanks.
H.J.
---
gcc/
2010-08-08 H.J. Lu <hongjiu.lu@intel.com>
PR target/45213
* config/i386/i386.c (ix86_print_operand): Use int instead of
long with REAL_VALUE_TO_TARGET_SINGLE.
gcc/testsuite/
2010-08-08 H.J. Lu <hongjiu.lu@intel.com>
PR target/45213
* gcc.target/i386/pr45213.c.
diff --git a/gcc/config/i386/i386.c b/gcc/config/i386/i386.c
index 204211a..f461481 100644
--- a/gcc/config/i386/i386.c
+++ b/gcc/config/i386/i386.c
@@ -12914,7 +12914,7 @@ ix86_print_operand (FILE *file, rtx x, int code)
else if (GET_CODE (x) == CONST_DOUBLE && GET_MODE (x) == SFmode)
{
REAL_VALUE_TYPE r;
- long l;
+ int l;
REAL_VALUE_FROM_CONST_DOUBLE (r, x);
REAL_VALUE_TO_TARGET_SINGLE (r, l);
--- /dev/null 2010-07-23 13:04:30.193381062 -0700
+++ gcc/gcc/testsuite/gcc.target/i386/pr45213.c 2010-08-06 15:15:03.680353825 -0700
@@ -0,0 +1,9 @@
+/* PR target/45213 */
+/* { dg-do assemble } */
+/* { dg-options "-Os -fno-omit-frame-pointer" } */
+
+void f(float, float, float, float, float, float, float, float, float, float);
+
+void g() {
+ f(0,0,0,0, 0,0,0,0, -1,1);
+}