From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from mailman by lists.gnu.org with tmda-scanned (Exim 4.43)
	id 1JFRtV-0005mn-I0
	for qemu-devel@nongnu.org; Thu, 17 Jan 2008 05:22:45 -0500
Received: from exim by lists.gnu.org with spam-scanned (Exim 4.43)
	id 1JFRtT-0005mF-QK
	for qemu-devel@nongnu.org; Thu, 17 Jan 2008 05:22:45 -0500
Received: from [199.232.76.173] (helo=monty-python.gnu.org)
	by lists.gnu.org with esmtp (Exim 4.43) id 1JFRtT-0005mA-Hp
	for qemu-devel@nongnu.org; Thu, 17 Jan 2008 05:22:43 -0500
Received: from kassel160.server4you.de ([62.75.246.160] helo=csgraf.de)
	by monty-python.gnu.org with esmtps
	(TLS-1.0:DHE_RSA_AES_256_CBC_SHA1:32) (Exim 4.60)
	(envelope-from <alex@csgraf.de>) id 1JFRtT-0008KY-4D
	for qemu-devel@nongnu.org; Thu, 17 Jan 2008 05:22:43 -0500
Received: from [10.10.100.38] (charybdis-ext.suse.de [195.135.221.2])
	by csgraf.de (Postfix) with ESMTP id 0355B615F
	for <qemu-devel@nongnu.org>; Thu, 17 Jan 2008 11:22:41 +0100 (CET)
Message-ID: <478EF8E8.8090707@csgraf.de>
Date: Thu, 17 Jan 2008 07:42:48 +0100
From: Alexander Graf <alex@csgraf.de>
MIME-Version: 1.0
Content-Type: multipart/mixed; boundary="------------040401080204070401060808"
Subject: [Qemu-devel] [PATCH 3/5] (PPC) Fix in-between branch returns
Reply-To: qemu-devel@nongnu.org
List-Id: qemu-devel.nongnu.org
List-Unsubscribe: <http://lists.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=unsubscribe>
List-Archive: <http://lists.gnu.org/pipermail/qemu-devel>
List-Post: <mailto:qemu-devel@nongnu.org>
List-Help: <mailto:qemu-devel-request@nongnu.org?subject=help>
List-Subscribe: <http://lists.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=subscribe>
To: qemu-devel@nongnu.org

This is a multi-part message in MIME format.
--------------040401080204070401060808
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit

Thanks to the branch detection I found several places where op.o
contained branch returns somewhere not at the end of the function. This
did not happen on other platforms for me, but I may be wrong.

This patch adds FORCE_RET at the end of those functions, to force gcc to
use a jump to the end and not return from it. The idea and most of the
debugging for this comes from Michael Matz.

--------------040401080204070401060808
Content-Type: text/x-patch;
 name="qemu-gcc4-forceret.patch"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
 filename="qemu-gcc4-forceret.patch"

Index: qemu-snapshot-2008-01-15_05/target-i386/op.c
===================================================================
--- qemu-snapshot-2008-01-15_05.orig/target-i386/op.c
+++ qemu-snapshot-2008-01-15_05/target-i386/op.c
@@ -291,6 +291,7 @@ void OPPROTO op_imull_EAX_T0(void)
     EDX = (uint32_t)(res >> 32);
     CC_DST = res;
     CC_SRC = (res != (int32_t)res);
+    FORCE_RET();
 }
 
 void OPPROTO op_imulw_T0_T1(void)
@@ -300,6 +301,7 @@ void OPPROTO op_imulw_T0_T1(void)
     T0 = res;
     CC_DST = res;
     CC_SRC = (res != (int16_t)res);
+    FORCE_RET();
 }
 
 void OPPROTO op_imull_T0_T1(void)
@@ -309,6 +311,7 @@ void OPPROTO op_imull_T0_T1(void)
     T0 = res;
     CC_DST = res;
     CC_SRC = (res != (int32_t)res);
+    FORCE_RET();
 }
 
 #ifdef TARGET_X86_64
Index: qemu-snapshot-2008-01-15_05/target-i386/ops_template.h
===================================================================
--- qemu-snapshot-2008-01-15_05.orig/target-i386/ops_template.h
+++ qemu-snapshot-2008-01-15_05/target-i386/ops_template.h
@@ -467,6 +467,7 @@ void OPPROTO glue(glue(op_bt, SUFFIX), _
     int count;
     count = T1 & SHIFT_MASK;
     CC_SRC = T0 >> count;
+    FORCE_RET();
 }
 
 void OPPROTO glue(glue(op_bts, SUFFIX), _T0_T1_cc)(void)

--------------040401080204070401060808--