[ANNOUNCE] April Release of LTP now Available

[ANNOUNCE] April Release of LTP now Available

[Marty Ridgeway]

The April release of LTP is now on SourceForge.

LTP-20050405
- Fixed '-d' option to runltp.sh
- Added optional ltp (fivextra) extended testsuite
- Added digital signiture testcases from the security team
- Applied patch from Gernot Payer to fix, mincore01, shmget02, capset02
- Applied suggested check from Carl van_Schaik to clone02.
- Added seLinux testsuite from security team and Stephen Smalley.
- Added NFS cthon04 tests needed files.


Linux Test Project
Linux Technology Center
IBM Corporation


Internet E-Mail : mridge@us.ibm.com
IBM, 11501 Burnet Rd, Austin, TX  78758
Phone (512) 838-1356 - T/L 678-1356 - Bldg. 908/1C005
Austin, TX.

Re: [ANNOUNCE] April Release of LTP now Available

[Andrew Morton]

Marty Ridgeway <mridge@us.ibm.com> wrote:
>
> The April release of LTP is now on SourceForge.
> 
>  LTP-20050405

It seems to have an x86ism in it which causes the compile to fail on ppc64:

socketcall01.c: In function `socketcall':
socketcall01.c:80: error: asm-specifier for variable `__sc_4' conflicts with asm clobber list




#ifndef _syscall2
#include <linux/unistd.h>
#endif

#include "test.h"
#include "usctest.h"

char *TCID = "socketcall01";                             /* Test program identifier.    */
#ifdef __NR_socketcall

_syscall2(int ,socketcall ,int ,call, unsigned long *, args);

Re: [LTP] Re: [ANNOUNCE] April Release of LTP now Available

[Dan Kegel]

Andrew Morton wrote:
>> LTP-20050405
> 
> It seems to have an x86ism in it which causes the compile to fail on ppc64:
> 
> socketcall01.c: In function `socketcall':
> socketcall01.c:80: error: asm-specifier for variable `__sc_4' conflicts with asm clobber list

That might be a problem with your toolchain.
Other mentions of that error message on Google
suggest that it's due to a kernel header problem.
I bet your toolchain uses kernel headers from 2.4.21 or earlier...
check includes/asm-ppc64/unistd.h to see if it's got the
line
   /* On powerpc a system call basically clobbers the same registers like a
in it.  If not, it may be missing the patch mentioned below.

See
http://ozlabs.org/pipermail/linuxppc64-dev/2003-April/000211.html
http://ozlabs.org/pipermail/linuxppc-dev/2002-October/014492.html
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=9379
http://www.hu.kernel.org/pub/linux/kernel/v2.4/snapshots/incr/patch-2.4.22-bk57-bk58

- Dan

-- 
Trying to get a job as a c++ developer?  See http://kegel.com/academy/getting-hired.html

Re: [LTP] Re: [ANNOUNCE] April Release of LTP now Available

[Arjan van de Ven]

On Wed, 2005-04-06 at 08:05 -0700, Dan Kegel wrote:
> Andrew Morton wrote:
> >> LTP-20050405
> > 
> > It seems to have an x86ism in it which causes the compile to fail on ppc64:
> > 
> > socketcall01.c: In function `socketcall':
> > socketcall01.c:80: error: asm-specifier for variable `__sc_4' conflicts with asm clobber list
> 
> That might be a problem with your toolchain.

I wonder why ltp doesn't just use the normal syscall() macro/functions
but instead insists on using stuff from kernel headers...

Re: [ANNOUNCE] April Release of LTP now Available

[Paul Mackerras]

Andrew Morton writes:

> Marty Ridgeway <mridge@us.ibm.com> wrote:
> >
> > The April release of LTP is now on SourceForge.
> > 
> >  LTP-20050405
> 
> It seems to have an x86ism in it which causes the compile to fail on ppc64:

Looks to me like gcc is objecting to our (ppc64's) _syscall2
definition; Alan Modra (cc'd) can probably say what we're doing wrong.

> socketcall01.c: In function `socketcall':
> socketcall01.c:80: error: asm-specifier for variable `__sc_4' conflicts with asm clobber list
> 
> 
> 
> 
> #ifndef _syscall2
> #include <linux/unistd.h>
> #endif
> 
> #include "test.h"
> #include "usctest.h"
> 
> char *TCID = "socketcall01";                             /* Test program identifier.    */
> #ifdef __NR_socketcall
> 
> _syscall2(int ,socketcall ,int ,call, unsigned long *, args);

Here is the definition of _syscall2 (for Alan):

#define __syscall_nr(nr, type, name, args...)				\
	unsigned long __sc_ret, __sc_err;				\
	{								\
		register unsigned long __sc_0  __asm__ ("r0");		\
		register unsigned long __sc_3  __asm__ ("r3");		\
		register unsigned long __sc_4  __asm__ ("r4");		\
		register unsigned long __sc_5  __asm__ ("r5");		\
		register unsigned long __sc_6  __asm__ ("r6");		\
		register unsigned long __sc_7  __asm__ ("r7");		\
		register unsigned long __sc_8  __asm__ ("r8");		\
									\
		__sc_loadargs_##nr(name, args);				\
		__asm__ __volatile__					\
			("sc           \n\t"				\
			 "mfcr %0      "				\
			: "=&r" (__sc_0),				\
			  "=&r" (__sc_3),  "=&r" (__sc_4),		\
			  "=&r" (__sc_5),  "=&r" (__sc_6),		\
			  "=&r" (__sc_7),  "=&r" (__sc_8)		\
			: __sc_asm_input_##nr				\
			: "cr0", "ctr", "memory",			\
			        "r9", "r10","r11", "r12");		\
		__sc_ret = __sc_3;					\
		__sc_err = __sc_0;					\
	}								\
	if (__sc_err & 0x10000000)					\
	{								\
		errno = __sc_ret;					\
		__sc_ret = -1;						\
	}								\
	return (type) __sc_ret

#define __sc_loadargs_0(name, dummy...)					\
	__sc_0 = __NR_##name
#define __sc_loadargs_1(name, arg1)					\
	__sc_loadargs_0(name);						\
	__sc_3 = (unsigned long) (arg1)
#define __sc_loadargs_2(name, arg1, arg2)				\
	__sc_loadargs_1(name, arg1);					\
	__sc_4 = (unsigned long) (arg2)
#define __sc_loadargs_3(name, arg1, arg2, arg3)				\
	__sc_loadargs_2(name, arg1, arg2);				\
	__sc_5 = (unsigned long) (arg3)
#define __sc_loadargs_4(name, arg1, arg2, arg3, arg4)			\
	__sc_loadargs_3(name, arg1, arg2, arg3);			\
	__sc_6 = (unsigned long) (arg4)
#define __sc_loadargs_5(name, arg1, arg2, arg3, arg4, arg5)		\
	__sc_loadargs_4(name, arg1, arg2, arg3, arg4);			\
	__sc_7 = (unsigned long) (arg5)
#define __sc_loadargs_6(name, arg1, arg2, arg3, arg4, arg5, arg6)	\
	__sc_loadargs_5(name, arg1, arg2, arg3, arg4, arg5);		\
	__sc_8 = (unsigned long) (arg6)

#define __sc_asm_input_0 "0" (__sc_0)
#define __sc_asm_input_1 __sc_asm_input_0, "1" (__sc_3)
#define __sc_asm_input_2 __sc_asm_input_1, "2" (__sc_4)
#define __sc_asm_input_3 __sc_asm_input_2, "3" (__sc_5)
#define __sc_asm_input_4 __sc_asm_input_3, "4" (__sc_6)
#define __sc_asm_input_5 __sc_asm_input_4, "5" (__sc_7)
#define __sc_asm_input_6 __sc_asm_input_5, "6" (__sc_8)

#define _syscall0(type,name)						\
type name(void)								\
{									\
	__syscall_nr(0, type, name);					\
}

#define _syscall1(type,name,type1,arg1)					\
type name(type1 arg1)							\
{									\
	__syscall_nr(1, type, name, arg1);				\
}

#define _syscall2(type,name,type1,arg1,type2,arg2)			\
type name(type1 arg1, type2 arg2)					\
{									\
	__syscall_nr(2, type, name, arg1, arg2);			\
}

Re: [ANNOUNCE] April Release of LTP now Available

[Chris Friesen]

Paul Mackerras wrote:

> 	if (__sc_err & 0x10000000)					\
> 	{								\
> 		errno = __sc_ret;					\
> 		__sc_ret = -1;						\
> 	}								\


If you're messing with that code anyways, any chance of changing the 
assignment to "__sc_ret = -1UL;", since __sc_ret is of type unsigned long?

Chris

Re: [ANNOUNCE] April Release of LTP now Available

[Alan Modra]

On Thu, Apr 07, 2005 at 10:52:01AM +1000, Paul Mackerras wrote:
> Looks to me like gcc is objecting to our (ppc64's) _syscall2
> definition; Alan Modra (cc'd) can probably say what we're doing wrong.

I can't spot anything wrong.  Take a look at preprocessed source.

-- 
Alan Modra
IBM OzLabs - Linux Technology Centre