--- Begin Message ---
From: CAI Qian <[email protected]>
Subject: Re: ptrace06 is not being compiled (temporary fix)
Date: Fri, 15 May 2009 22:08:43 +0800 (CST)
> From: rohit verma <[email protected]>
> Subject: ptrace06 is not being compiled (temporary fix)
> Date: Fri, 15 May 2009 14:17:15 +0530
>
>> Hi,
>> PFA the patch which can be a possible solution to the compilation problems
>> that you had mentioned about regarding the ptrace06 compilation on PPC64 and
>> IA-64. I haven't tested on PPC64 and IA-64 as I do not have access to them.
>> Let me know if it work fine, and can be integrated into the LTP tree.
>>
>> link to mailing list entry --
>>
>> http://www.nabble.com/ptrace06-is-not-being-compiled-td23035203.html#a23328058
>>
>> Regards,
>> rohit
>>
>>
>
> I have no access to PPC64 and IA-64 machines at the moment either. I am
> going to test this patch next Monday.
>
PPC64 is fine now, but still have problem on IA-64.
# uname -r
2.6.18-128.el5
# rpm -q glibc
glibc-2.5-34
# make
cc -I../../../../include -Wall ptrace01.c -L../../../../lib -lltp -o
ptrace01
cc -I../../../../include -Wall ptrace02.c -L../../../../lib -lltp -o
ptrace02
cc -I../../../../include -Wall ptrace03.c -L../../../../lib -lltp -o
ptrace03
cc -I../../../../include -Wall ptrace04.c -L../../../../lib -lltp -o
ptrace04
cc -I../../../../include -Wall ptrace06.c -L../../../../lib -lltp -o
ptrace06
In file included from /usr/include/asm/ptrace.h:58,
from /usr/include/linux/ptrace.h:49,
from ptrace06.c:20:
/usr/include/asm/fpu.h:57: error: redefinition of ‘struct ia64_fpreg'
In file included from /usr/include/linux/ptrace.h:49,
from ptrace06.c:20:
/usr/include/asm/ptrace.h:208: error: redefinition of ‘struct pt_all_user_regs'
make: *** [ptrace06] Error 1
However, if I remove those two lines,
- #include <linux/ptrace.h>
- #include <asm/ptrace.h>
it works fine here, but it will skip a lot of tests here,
#ifdef PTRACE_GETREGS
{
PTRACE_GETREGS,.data = 0}, {
PTRACE_GETREGS,.data = 1}, {
PTRACE_GETREGS,.data = 2}, {
PTRACE_GETREGS,.data = 3}, {
PTRACE_GETREGS,.data = -1}, {
PTRACE_GETREGS,.data = -2}, {
PTRACE_GETREGS,.data = -3}, {
PTRACE_GETREGS,.data = -4},
#endif
#ifdef PTRACE_GETFGREGS
{
PTRACE_GETFGREGS,.data = 0}, {
PTRACE_GETFGREGS,.data = 1}, {
PTRACE_GETFGREGS,.data = 2}, {
PTRACE_GETFGREGS,.data = 3}, {
PTRACE_GETFGREGS,.data = -1}, {
PTRACE_GETFGREGS,.data = -2}, {
PTRACE_GETFGREGS,.data = -3}, {
PTRACE_GETFGREGS,.data = -4},
#endif
#ifdef PTRACE_SETREGS
{
PTRACE_SETREGS,.data = 0}, {
PTRACE_SETREGS,.data = 1}, {
PTRACE_SETREGS,.data = 2}, {
PTRACE_SETREGS,.data = 3}, {
PTRACE_SETREGS,.data = -1}, {
PTRACE_SETREGS,.data = -2}, {
PTRACE_SETREGS,.data = -3}, {
PTRACE_SETREGS,.data = -4},
#endif
#ifdef PTRACE_SETFGREGS
{
PTRACE_SETFGREGS,.data = 0}, {
PTRACE_SETFGREGS,.data = 1}, {
PTRACE_SETFGREGS,.data = 2}, {
PTRACE_SETFGREGS,.data = 3}, {
PTRACE_SETFGREGS,.data = -1}, {
PTRACE_SETFGREGS,.data = -2}, {
PTRACE_SETFGREGS,.data = -3}, {
PTRACE_SETFGREGS,.data = -4},
#endif
In addition, some tests are failed for me on RHEL5 x86_64,
...
ptrace06 57 FAIL : ptrace(SETREGS, ..., (nil), (nil)) returned 0 instead
of -1
ptrace06 58 FAIL : ptrace(SETREGS, ..., (nil), 0x1) returned 0 instead of
-1
ptrace06 59 FAIL : ptrace(SETREGS, ..., (nil), 0x2) returned 0 instead of
-1
ptrace06 60 FAIL : ptrace(SETREGS, ..., (nil), 0x3) returned 0 instead of
-1
...
But that might be actually a bug for kernels.,
# strace ./ptrace06
...
ptrace(PTRACE_SETREGS, 22834, 0, 0) = 0
ptrace(PTRACE_SETREGS, 22834, 0, 0x1) = 0
ptrace(PTRACE_SETREGS, 22834, 0, 0x2) = 0
ptrace(PTRACE_SETREGS, 22834, 0, 0x3) = 0
...
Fedora IA-32 kernel returns those,
# uname -r
2.6.27.5-117.fc10.i686
...
ptrace(PTRACE_SETREGS, 5799, 0, 0) = -1 EFAULT (Bad address)
ptrace(PTRACE_SETREGS, 5799, 0, 0x1) = -1 EFAULT (Bad address)
ptrace(PTRACE_SETREGS, 5799, 0, 0x2) = -1 EFAULT (Bad address)
ptrace(PTRACE_SETREGS, 5799, 0, 0x3) = -1 EFAULT (Bad address)
...
CAI Qian
--- End Message ---
