Re: Why we need redhat_lsb_trigger.c in redhat-lsb package?

2012-03-14 Thread Niels de Vos 

On 03/13/2012 07:04 AM, Xibo Ning wrote:

Hi,
Do we really need a fineness c program to make symbol link? In
redhat-lsb.spec, we can see:

snip


%triggerpostun -- glibc
if [ -x /usr/sbin/redhat_lsb_trigger.%{_target_cpu} ]; then
/usr/sbin/redhat_lsb_trigger.%{_target_cpu}
fi

And in redhat_lsb_trigger.c, the main function is as follows:

snip

if (is_ia64 ())
INTERNAL_SYSCALL (symlink, err, 2, /emul/ia32-linux/lib/ LDSO, buf);
else
INTERNAL_SYSCALL (symlink, err, 2, LDSO, buf);

snip

Why need we use a manually c program, not sln command? We can copy a
copy of sln, so we can use it even when glibc is removed.


Probably the trigger can be executed at a critical time when no 
libraries are available, hence the static binary. I guess sln would be 
an alternative.


The best solution would probably be to replace the %triggerpostun with a 
lua scriptlet, something like this (totally untested and you'll need to 
add the LSB-version somewhere too):


%triggerpostun -p lua -- glibc
LDSO = /usr/lib/ld.so
LSBLDSO = /usr/lib/ld. + LSB_VERSION + .so
%ifarch ia64
LDSO = /emul/ia32-linux/lib + LSBLDSO
%fi
posix.symlink(LDSO, LSBLDSO)


Cheers,
Niels
--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

Re: Why we need redhat_lsb_trigger.c in redhat-lsb package?

2012-03-14 Thread Jakub Jelinek 
On Wed, Mar 14, 2012 at 10:08:32AM +0100, Niels de Vos wrote:
 On 03/13/2012 07:04 AM, Xibo Ning wrote:
 The best solution would probably be to replace the %triggerpostun
 with a lua scriptlet, something like this (totally untested and
 you'll need to add the LSB-version somewhere too):
 
 %triggerpostun -p lua -- glibc
 LDSO = /usr/lib/ld.so
 LSBLDSO = /usr/lib/ld. + LSB_VERSION + .so
 %ifarch ia64
 LDSO = /emul/ia32-linux/lib + LSBLDSO
 %fi
 posix.symlink(LDSO, LSBLDSO)

That is not equivalent.  is_ia64 () is true in the statically linked
binary for %ifarch %{ix86} if running on the ia64 CPU (software emulated).

Jakub
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

Re: Why we need redhat_lsb_trigger.c in redhat-lsb package?

2012-03-14 Thread Niels de Vos 
On Wed, Mar 14, 2012 at 10:13 AM, Jakub Jelinek ja...@redhat.com wrote:
 On Wed, Mar 14, 2012 at 10:08:32AM +0100, Niels de Vos wrote:
 On 03/13/2012 07:04 AM, Xibo Ning wrote:
 The best solution would probably be to replace the %triggerpostun
 with a lua scriptlet, something like this (totally untested and
 you'll need to add the LSB-version somewhere too):

 %triggerpostun -p lua -- glibc
 LDSO = /usr/lib/ld.so
 LSBLDSO = /usr/lib/ld. + LSB_VERSION + .so
 %ifarch ia64
 LDSO = /emul/ia32-linux/lib + LSBLDSO
 %fi
 posix.symlink(LDSO, LSBLDSO)

 That is not equivalent.  is_ia64 () is true in the statically linked
 binary for %ifarch %{ix86} if running on the ia64 CPU (software emulated).

Aha, so is_ia64() does some checks and that makes the decision to have
the symlink point to one ld.so or the other.
If those checks can be written in lua, replacing redhat_lsb_trigger.c
would make sense.

Cheers,
Niels
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

Re: Why we need redhat_lsb_trigger.c in redhat-lsb package?

2012-03-14 Thread Bill Nottingham 
Niels de Vos (de...@fedoraproject.org) said: 
  %triggerpostun -p lua -- glibc
  LDSO = /usr/lib/ld.so
  LSBLDSO = /usr/lib/ld. + LSB_VERSION + .so
  %ifarch ia64
  LDSO = /emul/ia32-linux/lib + LSBLDSO
  %fi
  posix.symlink(LDSO, LSBLDSO)
 
  That is not equivalent.  is_ia64 () is true in the statically linked
  binary for %ifarch %{ix86} if running on the ia64 CPU (software emulated).
 
 Aha, so is_ia64() does some checks and that makes the decision to have
 the symlink point to one ld.so or the other.
 If those checks can be written in lua, replacing redhat_lsb_trigger.c
 would make sense.

Well, it's not as if ia64 is supported in current Fedora or RHEL releases,
so those bits could just be removed...

Bill
-- 
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

Why we need redhat_lsb_trigger.c in redhat-lsb package?

2012-03-13 Thread Xibo Ning 

Hi,
Do we really need a fineness c program to make symbol link? In 
redhat-lsb.spec, we can see:

gcc $RPM_OPT_FLAGS -Os -static -fno-stack-protector -o
redhat_lsb_trigger{.%{_target_cpu},.c} -DLSBSOVER='%{lsbsover}' \
  -DLDSO='%{ldso}' -DLSBLDSO='/%{_lib}/%{lsbldso}' -D_GNU_SOURCE
install -m 700 redhat_lsb_trigger.%{_target_cpu} \
  $RPM_BUILD_ROOT%{_sbindir}/redhat_lsb_trigger.%{_target_cpu}

%triggerpostun -- glibc
if [ -x /usr/sbin/redhat_lsb_trigger.%{_target_cpu} ]; then
  /usr/sbin/redhat_lsb_trigger.%{_target_cpu}
fi

And in redhat_lsb_trigger.c, the main function is as follows:

int main (int argc, char **argv)
{
/* redhat_lsb_trigger.c:428: warning: unused variable 'ret' */
/*  long ret; */
  INTERNAL_SYSCALL_DECL (err);
  char lsbsover[] = LSBSOVER;
  char *LSBVER, *p = lsbsover;
  while ((LSBVER = strtok (p,  \t)) != NULL)
{
  char buf[sizeof LSBLDSO + 1 + strlen (LSBVER)];
  p = buf;
  p = mempcpy (p, LSBLDSO, sizeof LSBLDSO - 1);
  *p++ = '.';
  strcpy (p, LSBVER);
  if (is_ia64 ())
INTERNAL_SYSCALL (symlink, err, 2, /emul/ia32-linux/lib/ 
LDSO, buf);

  else
INTERNAL_SYSCALL (symlink, err, 2, LDSO, buf);
  p = NULL;
}
  INTERNAL_SYSCALL (exit, err, 1, 0);
  return 110;
}

Why need we use a manually c program, not sln command? We can copy a 
copy of sln, so we can use it even when glibc is removed.


Why?

Thanks alot
xning
--
devel mailing list
devel@lists.fedoraproject.org
https://admin.fedoraproject.org/mailman/listinfo/devel

Re: Why we need redhat_lsb_trigger.c in redhat-lsb package?

2012-03-13 Thread Xibo Ning 

On 03/13/2012 02:04 PM, Xibo Ning wrote:

Hi,
Do we really need a fineness c program to make symbol link? In
redhat-lsb.spec, we can see:
gcc $RPM_OPT_FLAGS -Os -static -fno-stack-protector -o
redhat_lsb_trigger{.%{_target_cpu},.c} -DLSBSOVER='%{lsbsover}' \
-DLDSO='%{ldso}' -DLSBLDSO='/%{_lib}/%{lsbldso}' -D_GNU_SOURCE
install -m 700 redhat_lsb_trigger.%{_target_cpu} \
$RPM_BUILD_ROOT%{_sbindir}/redhat_lsb_trigger.%{_target_cpu}

%triggerpostun -- glibc
if [ -x /usr/sbin/redhat_lsb_trigger.%{_target_cpu} ]; then
/usr/sbin/redhat_lsb_trigger.%{_target_cpu}
fi

And in redhat_lsb_trigger.c, the main function is as follows:

int main (int argc, char **argv)
{
/* redhat_lsb_trigger.c:428: warning: unused variable 'ret' */
/* long ret; */
INTERNAL_SYSCALL_DECL (err);
char lsbsover[] = LSBSOVER;
char *LSBVER, *p = lsbsover;
while ((LSBVER = strtok (p,  \t)) != NULL)
{
char buf[sizeof LSBLDSO + 1 + strlen (LSBVER)];
p = buf;
p = mempcpy (p, LSBLDSO, sizeof LSBLDSO - 1);
*p++ = '.';
strcpy (p, LSBVER);
if (is_ia64 ())
INTERNAL_SYSCALL (symlink, err, 2, /emul/ia32-linux/lib/ LDSO, buf);
else
INTERNAL_SYSCALL (symlink, err, 2, LDSO, buf);
p = NULL;
}
INTERNAL_SYSCALL (exit, err, 1, 0);
return 110;
}

Why need we use a manually c program, not sln command? We can copy a
copy of sln, so we can use it even when glibc is removed.

Why?

Thanks alot
xning


#ifdef __sparc__
register void *__thread_self __asm (g7);
#endif
#include unistd.h
#include string.h
#include fcntl.h
#include syscall.h

#if defined __i386__
# define INTERNAL_SYSCALL_DECL(err) do { } while (0)
# define INTERNAL_SYSCALL(name, err, nr, args...) \
  ({	  \
register unsigned int resultvar;	  \
asm volatile (			  \
movl %1, %%eax\n\t		  \
int $0x80\n\t			  \
: =a (resultvar)			  \
: i (__NR_##name) ASMFMT_##nr(args) : memory, cc);		  \
(int) resultvar; })
# define INTERNAL_SYSCALL_ERROR_P(val, err) \
  ((unsigned int) (val) = 0xf001u)
# define ASMFMT_0()
# define ASMFMT_1(arg1) \
	, b (arg1)
# define ASMFMT_2(arg1, arg2) \
	, b (arg1), c (arg2)
# define ASMFMT_3(arg1, arg2, arg3) \
	, b (arg1), c (arg2), d (arg3)
#elif defined __x86_64__
# define INTERNAL_SYSCALL_DECL(err) do { } while (0)
# define INTERNAL_SYSCALL_NCS(name, err, nr, args...) \
  ({	  \
unsigned long resultvar;		  \
LOAD_ARGS_##nr (args)		  \
LOAD_REGS_##nr			  \
asm volatile (			  \
syscall\n\t			  \
: =a (resultvar)			  \
: 0 (name) ASM_ARGS_##nr : memory, cc, r11, cx);		  \
(long) resultvar; })
# define INTERNAL_SYSCALL(name, err, nr, args...) \
  INTERNAL_SYSCALL_NCS (__NR_##name, err, nr, ##args)
# define INTERNAL_SYSCALL_ERROR_P(val, err) \
  ((unsigned long) (val) = -4095L)
# define LOAD_ARGS_0()
# define LOAD_REGS_0
# define ASM_ARGS_0
# define LOAD_ARGS_1(a1)\
  long int __arg1 = (long) (a1);			\
  LOAD_ARGS_0 ()
# define LOAD_REGS_1	\
  register long int _a1 asm (rdi) = __arg1;		\
  LOAD_REGS_0
# define ASM_ARGS_1	ASM_ARGS_0, r (_a1)
# define LOAD_ARGS_2(a1, a2)\
  long int __arg2 = (long) (a2);			\
  LOAD_ARGS_1 (a1)
# define LOAD_REGS_2	\
  register long int _a2 asm (rsi) = __arg2;		\
  LOAD_REGS_1
# define ASM_ARGS_2	ASM_ARGS_1, r (_a2)
# define LOAD_ARGS_3(a1, a2, a3)			\
  long int __arg3 = (long) (a3);			\
  LOAD_ARGS_2 (a1, a2)
# define LOAD_REGS_3	\
  register long int _a3 asm (rdx) = __arg3;		\
  LOAD_REGS_2
# define ASM_ARGS_3	ASM_ARGS_2, r (_a3)
#elif defined __powerpc__
# define INTERNAL_SYSCALL_DECL(err) long int err
# define INTERNAL_SYSCALL_NCS(name, err, nr, args...) \
  ({	\
register long int r0  __asm__ (r0);\
register long int r3  __asm__ (r3);\
register long int r4  __asm__ (r4);\
register long int r5  __asm__ (r5);\
register long int r6  __asm__ (r6);\
register long int r7  __asm__ (r7);\
register long int r8  __asm__ (r8);\
LOADARGS_##nr(name, args);		\
__asm__ __volatile__		\
  (sc\n\t\
   mfcr  %0\n\t			\
   : =r (r0),			\
 =r (r3), =r (r4), =r (r5),\
 =r (r6), =r (r7), =r (r8)\
   : ASM_INPUT_##nr			\
   : r9, r10, r11, r12,	\
 cr0, ctr, memory);	\
	  err = r0;  \
(int) r3;  \
  })
# define INTERNAL_SYSCALL(name, err, nr, args...)			\
  INTERNAL_SYSCALL_NCS (__NR_##name, err, nr, ##args)
# define INTERNAL_SYSCALL_ERROR_P(val, err) \
  ((void) (val), __builtin_expect ((err)  (1  28), 0))
# define LOADARGS_0(name, dummy) \
	r0 = name
# define LOADARGS_1(name, __arg1) \
	long int arg1 = (long int) (__arg1); \
	LOADARGS_0(name, 0); \
	r3 = arg1
# define LOADARGS_2(name, __arg1, __arg2) \
	long int arg2 = (long int) (__arg2); \
	LOADARGS_1(name, __arg1); \
	r4 = arg2
# define LOADARGS_3(name, __arg1, __arg2, __arg3) \
	long int arg3 = (long int) (__arg3); \
	LOADARGS_2(name, __arg1, __arg2); \