[Bug runtime/6764] New: Execution of probefunc.exp causes system hang on s390 system

[Bug runtime/6764] New: Execution of probefunc.exp causes system hang on s390 system

[srinivasa at in dot ibm dot com]

Environment: Kernel-2.6.26-rc9,elfutils-0.131,systemtap-0.7

Execution of probefunc.exp causes system hang on s390. I have narrowed down this
problem, it looks like probing scheduler_tick() along with printing value of
pp() or probefunc() hangs the system.

global funcname;
global funcname1;
 probe kernel.statement(0x000000000003b83c).absolute {
        funcname = probefunc();
        funcname1 = pp();
      printf("========%s \n",funcname);
        printf("========%s \n",funcname1);
}

probe begin {
        printf(" loading \n");
        }
probe end {
        printf(" unloading \n");
        }


However a simple script which probes scheduler_tick() and logs some message
works well on the same system.

probe kernel.statement(0x000000000003b83c).absolute {
        printf("hi \n");
}
probe begin {
        printf(" loading \n");
        }
probe end {
        printf(" unloading \n");
        }


Could somebody throw light on why printing pp() or probefunc() inside
scheduler_tick() causes system hang. 

Thanks in advance
 Srinivasa DS

-- 
           Summary: Execution of probefunc.exp causes system hang on s390
                    system
           Product: systemtap
           Version: unspecified
            Status: NEW
          Severity: normal
          Priority: P2
         Component: runtime
        AssignedTo: systemtap at sources dot redhat dot com
        ReportedBy: srinivasa at in dot ibm dot com


http://sourceware.org/bugzilla/show_bug.cgi?id=6764

------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

[Bug runtime/6764] Execution of probefunc.exp causes system hang on s390 system

[srinivasa at in dot ibm dot com]

------- Additional Comments From srinivasa at in dot ibm dot com  2008-07-25 06:34 -------
Different scenarios under which system hangs

1) printing pp() inside scheduler_tick() probe handler.
========================================
probe kernel.statement(0x000000000003b83c).absolute {
                funcname = pp();
      printf("========%s \n",funcname);
        }
==============================
2)printing probefunc() inside scheduler_tick() probe handler.
========================================
probe kernel.statement(0x000000000003b83c).absolute {
                funcname = probefunc();
      printf("========%s \n",funcname);
        }
==============================


Different scenarios under which system works well.
1) printing "hello" message  and calling probefunc() or pp()(without printing
their value) inside scheduler_tick() probe handler.
========================================
probe kernel.statement(0x000000000003b83c).absolute {
                funcname = probefunc();
      printf("hello");
        }
==============================


-- 


http://sourceware.org/bugzilla/show_bug.cgi?id=6764

------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

[Bug runtime/6764] Execution of probefunc.exp causes system hang on s390 system

[wenji dot huang at oracle dot com]

------- Additional Comments From wenji dot huang at oracle dot com  2008-07-25 06:45 -------
> Different scenarios under which system works well.
> 1) printing "hello" message  and calling probefunc() or pp()(without printing
> their value) inside scheduler_tick() probe handler.
> ========================================
> probe kernel.statement(0x000000000003b83c).absolute {
>                 funcname = probefunc();
>       printf("hello");
>         }
> ==============================

The unused variable will be optimized, try it with -u. The problem won't exist
on x86, seems specific to s390.





-- 


http://sourceware.org/bugzilla/show_bug.cgi?id=6764

------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.