* UML build failure with 2.6.18
@ 2006-09-20 13:27 Mitch
2006-09-20 15:22 ` Jeff Dike
0 siblings, 1 reply; 6+ messages in thread
From: Mitch @ 2006-09-20 13:27 UTC (permalink / raw)
To: linux-kernel
Hi
anyone having UML build failure with the shiny new 2.6.18
laptop /usr/src/sources/linux/linux-2.6.18# !ma
make ARCH=um
SYMLINK arch/um/include/kern_constants.h
SYMLINK arch/um/include/sysdep
make[1]: `arch/um/sys-i386/user-offsets.s' is up to date.
CHK arch/um/include/user_constants.h
CHK include/linux/version.h
CHK include/linux/compile.h
CC arch/um/os-Linux/process.o
arch/um/os-Linux/process.c:144: error: syntax error before 'getpid'
arch/um/os-Linux/process.c:146: warning: return type defaults to 'int'
arch/um/os-Linux/process.c:146: warning: function declaration isn't a
prototype
arch/um/os-Linux/process.c: In function '_syscall0':
arch/um/os-Linux/process.c:147: error: syntax error before '{' token
arch/um/os-Linux/process.c:162: error: syntax error before 'prot'
arch/um/os-Linux/process.c:209: error: parameter 'ok' is initialized
arch/um/os-Linux/process.c:211: error: syntax error before 'printk'
arch/um/os-Linux/process.c:279: error: syntax error before '*' token
make[1]: *** [arch/um/os-Linux/process.o] Error 1
make: *** [arch/um/os-Linux] Error 2
^ permalink raw reply [flat|nested] 6+ messages in thread* Re: UML build failure with 2.6.18 2006-09-20 13:27 UML build failure with 2.6.18 Mitch @ 2006-09-20 15:22 ` Jeff Dike 0 siblings, 0 replies; 6+ messages in thread From: Jeff Dike @ 2006-09-20 15:22 UTC (permalink / raw) To: Mitch; +Cc: linux-kernel On Wed, Sep 20, 2006 at 05:27:12PM +0400, Mitch@0Bits.COM wrote: > anyone having UML build failure with the shiny new 2.6.18 > CC arch/um/os-Linux/process.o > arch/um/os-Linux/process.c:144: error: syntax error before 'getpid' > arch/um/os-Linux/process.c:146: warning: return type defaults to 'int' > arch/um/os-Linux/process.c:146: warning: function declaration isn't a > prototype > arch/um/os-Linux/process.c: In function '_syscall0': > arch/um/os-Linux/process.c:147: error: syntax error before '{' token > arch/um/os-Linux/process.c:162: error: syntax error before 'prot' > arch/um/os-Linux/process.c:209: error: parameter 'ok' is initialized > arch/um/os-Linux/process.c:211: error: syntax error before 'printk' > arch/um/os-Linux/process.c:279: error: syntax error before '*' token > make[1]: *** [arch/um/os-Linux/process.o] Error 1 > make: *** [arch/um/os-Linux] Error 2 What's the host distro? I just built (plus the jmmpbuf patch, which isn't the problem here) and booted it, and it seems fine here. Jeff ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: UML build failure with 2.6.18 @ 2006-09-20 15:31 Mitch 2006-09-20 16:47 ` Bill Rugolsky Jr. 2006-09-20 17:40 ` Jeff Dike 0 siblings, 2 replies; 6+ messages in thread From: Mitch @ 2006-09-20 15:31 UTC (permalink / raw) To: jdike, linux-kernel It's no distro (built up myself, so you can assume a gentoo type build). I'm building from scratch, so it should use no headers since i'm building a kernel. -------- Original Message -------- Subject: Re: UML build failure with 2.6.18 Date: Wed, 20 Sep 2006 11:22:57 -0400 From: Jeff Dike <jdike@addtoit.com> To: Mitch@0Bits.COM CC: linux-kernel@vger.kernel.org References: <Pine.LNX.4.63.0609201724320.29128@hasbox.com> On Wed, Sep 20, 2006 at 05:27:12PM +0400, Mitch@0Bits.COM wrote: > anyone having UML build failure with the shiny new 2.6.18 > CC arch/um/os-Linux/process.o > arch/um/os-Linux/process.c:144: error: syntax error before 'getpid' > arch/um/os-Linux/process.c:146: warning: return type defaults to 'int' > arch/um/os-Linux/process.c:146: warning: function declaration isn't a > prototype > arch/um/os-Linux/process.c: In function '_syscall0': > arch/um/os-Linux/process.c:147: error: syntax error before '{' token > arch/um/os-Linux/process.c:162: error: syntax error before 'prot' > arch/um/os-Linux/process.c:209: error: parameter 'ok' is initialized > arch/um/os-Linux/process.c:211: error: syntax error before 'printk' > arch/um/os-Linux/process.c:279: error: syntax error before '*' token > make[1]: *** [arch/um/os-Linux/process.o] Error 1 > make: *** [arch/um/os-Linux] Error 2 What's the host distro? I just built (plus the jmmpbuf patch, which isn't the problem here) and booted it, and it seems fine here. Jeff ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: UML build failure with 2.6.18 2006-09-20 15:31 Mitch @ 2006-09-20 16:47 ` Bill Rugolsky Jr. 2006-09-20 17:40 ` Jeff Dike 1 sibling, 0 replies; 6+ messages in thread From: Bill Rugolsky Jr. @ 2006-09-20 16:47 UTC (permalink / raw) To: Mitch; +Cc: jdike, linux-kernel On Wed, Sep 20, 2006 at 07:31:19PM +0400, Mitch wrote: > It's no distro (built up myself, so you can assume a gentoo type build). > I'm building from scratch, so it should use no headers since i'm > building a kernel. You are probably missing the syscall macros? I saw a similar problem the other day and noticed on Fedora Core that upgrading glibc-kernheaders-2.4-9.1.94 to glibc-kernheaders-3.0-45 caused a regression in /usr/include/linux/unistd.h: --- - 2006-09-20 12:45:31.195152000 -0400 +++ /usr/include/linux/unistd.h 2006-06-22 09:49:00.000000000 -0400 @@ -6,63 +6,5 @@ * Include machine specific syscallX macros */ #include <asm/unistd.h> -#include <sys/syscall.h> -#include <unistd.h> - - -#undef _syscall0 -#undef _syscall1 -#undef _syscall2 -#undef _syscall3 -#undef _syscall4 -#undef _syscall5 -#undef _syscall6 - - -#define _syscall0(type,name) \ -type name(void) \ -{\ - return syscall(__NR_##name);\ -} - -#define _syscall1(type,name,type1,arg1) \ -type name(type1 arg1) \ -{\ - return syscall(__NR_##name, arg1);\ -} - -#define _syscall2(type,name,type1,arg1,type2,arg2) \ -type name(type1 arg1,type2 arg2) \ -{\ - return syscall(__NR_##name, arg1, arg2);\ -} - -#define _syscall3(type,name,type1,arg1,type2,arg2,type3,arg3) \ -type name(type1 arg1,type2 arg2,type3 arg3) \ -{\ - return syscall(__NR_##name, arg1, arg2, arg3);\ -} - -#define _syscall4(type,name,type1,arg1,type2,arg2,type3,arg3,type4,arg4) \ -type name (type1 arg1, type2 arg2, type3 arg3, type4 arg4) \ -{\ - return syscall(__NR_##name, arg1, arg2, arg3, arg4);\ -} - -#define _syscall5(type,name,type1,arg1,type2,arg2,type3,arg3,type4,arg4, \ - type5,arg5) \ -type name (type1 arg1,type2 arg2,type3 arg3,type4 arg4,type5 arg5) \ -{\ - return syscall(__NR_##name, arg1, arg2, arg3, arg4, arg5);\ -} - - -#define _syscall6(type,name,type1,arg1,type2,arg2,type3,arg3,type4,arg4, \ - type5,arg5,type6,arg6) \ -type name (type1 arg1,type2 arg2,type3 arg3,type4 arg4,type5 arg5,type6 arg6) \ -{\ - return syscall(__NR_##name, arg1, arg2, arg3, arg4, arg5, arg6);\ -} - #endif /* _LINUX_UNISTD_H_ */ Regards, Bill Rugolsky ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: UML build failure with 2.6.18 2006-09-20 15:31 Mitch 2006-09-20 16:47 ` Bill Rugolsky Jr. @ 2006-09-20 17:40 ` Jeff Dike 1 sibling, 0 replies; 6+ messages in thread From: Jeff Dike @ 2006-09-20 17:40 UTC (permalink / raw) To: Mitch; +Cc: linux-kernel, Bill Rugolsky Jr. On Wed, Sep 20, 2006 at 07:31:19PM +0400, Mitch wrote: > It's no distro (built up myself, so you can assume a gentoo type build). > I'm building from scratch, so it should use no headers since i'm > building a kernel. UML interacts with the host, by making system calls, so it needs the host's libc headers. Try the patch below. Jeff Index: linux-2.6.17/arch/um/os-Linux/process.c =================================================================== --- linux-2.6.17.orig/arch/um/os-Linux/process.c 2006-09-20 11:15:08.000000000 -0400 +++ linux-2.6.17/arch/um/os-Linux/process.c 2006-09-20 13:35:24.000000000 -0400 @@ -140,11 +140,9 @@ void os_usr1_process(int pid) * syscalls, and also breaks with clone(), which does not unshare the TLS. */ -inline _syscall0(pid_t, getpid) - int os_getpid(void) { - return(getpid()); + return syscall(__NR_getpid); } int os_getpgrp(void) Index: linux-2.6.17/arch/um/os-Linux/sys-i386/tls.c =================================================================== --- linux-2.6.17.orig/arch/um/os-Linux/sys-i386/tls.c 2006-06-18 13:49:35.000000000 -0400 +++ linux-2.6.17/arch/um/os-Linux/sys-i386/tls.c 2006-09-20 13:37:27.000000000 -0400 @@ -3,8 +3,6 @@ #include "sysdep/tls.h" #include "user_util.h" -static _syscall1(int, get_thread_area, user_desc_t *, u_info); - /* Checks whether host supports TLS, and sets *tls_min according to the value * valid on the host. * i386 host have it == 6; x86_64 host have it == 12, for i386 emulation. */ @@ -17,7 +15,7 @@ void check_host_supports_tls(int *suppor user_desc_t info; info.entry_number = val[i]; - if (get_thread_area(&info) == 0) { + if(syscall(__NR_get_thread_area, &info) == 0){ *tls_min = val[i]; *supports_tls = 1; return; Index: linux-2.6.17/arch/um/os-Linux/tls.c =================================================================== --- linux-2.6.17.orig/arch/um/os-Linux/tls.c 2006-08-15 21:59:56.000000000 -0400 +++ linux-2.6.17/arch/um/os-Linux/tls.c 2006-09-20 13:37:14.000000000 -0400 @@ -48,14 +48,11 @@ int os_get_thread_area(user_desc_t *info #ifdef UML_CONFIG_MODE_TT #include "linux/unistd.h" -static _syscall1(int, get_thread_area, user_desc_t *, u_info); -static _syscall1(int, set_thread_area, user_desc_t *, u_info); - int do_set_thread_area_tt(user_desc_t *info) { int ret; - ret = set_thread_area(info); + ret = syscall(__NR_set_thread_area, info); if (ret < 0) { ret = -errno; } @@ -66,7 +63,7 @@ int do_get_thread_area_tt(user_desc_t *i { int ret; - ret = get_thread_area(info); + ret = syscall(__NR_get_thread_area, info); if (ret < 0) { ret = -errno; } ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: UML build failure with 2.6.18
@ 2006-09-20 21:05 Mitch
0 siblings, 0 replies; 6+ messages in thread
From: Mitch @ 2006-09-20 21:05 UTC (permalink / raw)
To: jdike, brugolsky, linux-kernel
Hi Jeff,
That works perfectly as did the previous patch. I knew how to fix the
problem anyhow, just wanted to know why i/if was the only person having
thr problem. I'm running glibc-2.3.6 and gcc-4.0.3.
Cheers
Mithc
-------- Original Message --------
Subject: Re: UML build failure with 2.6.18
Date: Wed, 20 Sep 2006 13:40:26 -0400
From: Jeff Dike <jdike@addtoit.com>
To: Mitch <Mitch@0Bits.COM>
CC: linux-kernel@vger.kernel.org, "Bill Rugolsky Jr."
<brugolsky@telemetry-investments.com>
References: <45115EC7.2010407@0Bits.COM>
On Wed, Sep 20, 2006 at 07:31:19PM +0400, Mitch wrote:
> It's no distro (built up myself, so you can assume a gentoo type build).
> I'm building from scratch, so it should use no headers since i'm
> building a kernel.
UML interacts with the host, by making system calls, so it needs the
host's libc headers.
Try the patch below.
Jeff
Index: linux-2.6.17/arch/um/os-Linux/process.c
===================================================================
--- linux-2.6.17.orig/arch/um/os-Linux/process.c 2006-09-20
11:15:08.000000000 -0400
+++ linux-2.6.17/arch/um/os-Linux/process.c 2006-09-20
13:35:24.000000000 -0400
@@ -140,11 +140,9 @@ void os_usr1_process(int pid)
* syscalls, and also breaks with clone(), which does not unshare the TLS.
*/
-inline _syscall0(pid_t, getpid)
-
int os_getpid(void)
{
- return(getpid());
+ return syscall(__NR_getpid);
}
int os_getpgrp(void)
Index: linux-2.6.17/arch/um/os-Linux/sys-i386/tls.c
===================================================================
--- linux-2.6.17.orig/arch/um/os-Linux/sys-i386/tls.c 2006-06-18
13:49:35.000000000 -0400
+++ linux-2.6.17/arch/um/os-Linux/sys-i386/tls.c 2006-09-20
13:37:27.000000000 -0400
@@ -3,8 +3,6 @@
#include "sysdep/tls.h"
#include "user_util.h"
-static _syscall1(int, get_thread_area, user_desc_t *, u_info);
-
/* Checks whether host supports TLS, and sets *tls_min according to
the value
* valid on the host.
* i386 host have it == 6; x86_64 host have it == 12, for i386
emulation. */
@@ -17,7 +15,7 @@ void check_host_supports_tls(int *suppor
user_desc_t info;
info.entry_number = val[i];
- if (get_thread_area(&info) == 0) {
+ if(syscall(__NR_get_thread_area, &info) == 0){
*tls_min = val[i];
*supports_tls = 1;
return;
Index: linux-2.6.17/arch/um/os-Linux/tls.c
===================================================================
--- linux-2.6.17.orig/arch/um/os-Linux/tls.c 2006-08-15
21:59:56.000000000 -0400
+++ linux-2.6.17/arch/um/os-Linux/tls.c 2006-09-20 13:37:14.000000000 -0400
@@ -48,14 +48,11 @@ int os_get_thread_area(user_desc_t *info
#ifdef UML_CONFIG_MODE_TT
#include "linux/unistd.h"
-static _syscall1(int, get_thread_area, user_desc_t *, u_info);
-static _syscall1(int, set_thread_area, user_desc_t *, u_info);
-
int do_set_thread_area_tt(user_desc_t *info)
{
int ret;
- ret = set_thread_area(info);
+ ret = syscall(__NR_set_thread_area, info);
if (ret < 0) {
ret = -errno;
}
@@ -66,7 +63,7 @@ int do_get_thread_area_tt(user_desc_t *i
{
int ret;
- ret = get_thread_area(info);
+ ret = syscall(__NR_get_thread_area, info);
if (ret < 0) {
ret = -errno;
}
^ permalink raw reply [flat|nested] 6+ messages in threadend of thread, other threads:[~2006-09-20 21:05 UTC | newest] Thread overview: 6+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2006-09-20 13:27 UML build failure with 2.6.18 Mitch 2006-09-20 15:22 ` Jeff Dike -- strict thread matches above, loose matches on Subject: below -- 2006-09-20 15:31 Mitch 2006-09-20 16:47 ` Bill Rugolsky Jr. 2006-09-20 17:40 ` Jeff Dike 2006-09-20 21:05 Mitch
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox