From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from list by monty-python.gnu.org with tmda-scanned (Exim 4.30)
	id 1BHpWk-0004e2-Lo
	for qemu-devel@nongnu.org; Sun, 25 Apr 2004 15:42:58 -0400
Received: from mail by monty-python.gnu.org with spam-scanned (Exim 4.30)
	id 1BHpWE-00047t-8l
	for qemu-devel@nongnu.org; Sun, 25 Apr 2004 15:42:57 -0400
Received: from [62.210.158.41] (helo=moscou.magic.fr)
	by monty-python.gnu.org with esmtp (Exim 4.30) id 1BHpV5-0003TG-6s
	for qemu-devel@nongnu.org; Sun, 25 Apr 2004 15:41:15 -0400
Received: from [10.0.0.2] (ppp-181.net-555.magic.fr [62.210.255.181])
	by moscou.magic.fr (8.11.6/8.10.1) with ESMTP id i3PJf5D09528
	for <qemu-devel@nongnu.org>; Sun, 25 Apr 2004 21:41:05 +0200 (CEST)
Subject: Re: [Qemu-devel] amd64 compile
From: "J. Mayer" <l_indien@magic.fr>
In-Reply-To: <408C0B60.8030906@bellard.org>
References: <Pine.LNX.4.44.0404251715500.31508-100000@wrasse.demon.co.uk>
	<1082917248.14652.7.camel@rapid> <1082918435.14634.11.camel@rapid>
	<408C0B60.8030906@bellard.org>
Content-Type: text/plain
Message-Id: <1082922373.14634.102.camel@rapid>
Mime-Version: 1.0
Date: Sun, 25 Apr 2004 21:46:13 +0200
Content-Transfer-Encoding: 7bit
Reply-To: qemu-devel@nongnu.org
List-Id: qemu-devel.nongnu.org
List-Unsubscribe: <http://mail.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=unsubscribe>
List-Archive: <http://mail.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://mail.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=subscribe>
To: qemu-devel@nongnu.org

On Sun, 2004-04-25 at 21:02, Fabrice Bellard wrote:
>  From the kernel sources, _llseek is handled the same on every arch 
> (linux/fs/read_write.c). So there must be another problem.

Oh well, you're right. I absolutely wanted to use llseek, and I didn't
noticed the way it's defined in syscall.c, which is buggy, at least for
amd64.
lseek is 3 args for amd64, not 5 and there is no llseek syscall at all.

This patch gets it right:

Index: linux-user/syscall.c
===================================================================
RCS file: /cvsroot/qemu/qemu/linux-user/syscall.c,v
retrieving revision 1.47
diff -u -d -w -B -b -d -p -r1.47 syscall.c
--- linux-user/syscall.c        12 Apr 2004 20:39:29 -0000      1.47
+++ linux-user/syscall.c        25 Apr 2004 19:39:10 -0000
@@ -2402,9 +2402,14 @@ long do_syscall(void *cpu_env, int num, 
         goto unimplemented;
     case TARGET_NR__llseek:
         {
+#if defined (__x86_64__)
+            ret = get_errno(lseek(arg1, ((uint64_t )arg2 << 32) | arg3,
arg5));
+            *(int64_t *)arg4 = ret;
+#else
             int64_t res;
             ret = get_errno(_llseek(arg1, arg2, arg3, &res, arg5));
             *(int64_t *)arg4 = tswap64(res);
+#endif
         }
         break;
     case TARGET_NR_getdents:

-- 
J. Mayer <l_indien@magic.fr>
Never organized