[Bug libc/366] New: Compiling on 64-bit yields segmentation fault when run, on 32-bit runs fine.

[Bug libc/366] New: Compiling on 64-bit yields segmentation fault when run, on 32-bit runs fine.

[dafrabbit at yahoo dot com]

Compiling a small C program in 64-bit mode with gcc, compiling is fine, but when
run the program seg faults. Same program, compiled in 32-bit mode with gcc runs
fine. Bug report posted in gcc Bugzilla, told to repost here. (link to gcc bug
report: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17302)

Building on an Athlon64, gcc version 3.4.1-9, glibc version 2.3.3-27, kernel
version 2.6.5-1.358.

To reproduce:
1. gcc -g test_var5.c
2. gcc -o test5.exe test_var5.o
3. test5.exe

The program is attached.

-- 
           Summary: Compiling on 64-bit yields segmentation fault when run,
                    on 32-bit runs fine.
           Product: glibc
           Version: 2.3.3
            Status: NEW
          Severity: normal
          Priority: P1
         Component: libc
        AssignedTo: gotom at debian dot or dot jp
        ReportedBy: dafrabbit at yahoo dot com
                CC: glibc-bugs at sources dot redhat dot com


http://sources.redhat.com/bugzilla/show_bug.cgi?id=366

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug libc/366] Compiling on 64-bit yields segmentation fault when run, on 32-bit runs fine.

[dafrabbit at yahoo dot com]

------- Additional Comments From dafrabbit at yahoo dot com  2004-09-03 02:39 -------
Created an attachment (id=185)
 --> (http://sources.redhat.com/bugzilla/attachment.cgi?id=185&action=view)
The test file that I was testing 64-bit compilation with.


-- 


http://sources.redhat.com/bugzilla/show_bug.cgi?id=366

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug libc/366] Compiling on 64-bit yields segmentation fault when run, on 32-bit runs fine.

[aj at suse dot de]

------- Additional Comments From aj at suse dot de  2004-09-03 04:23 -------
The ISO C99 standard in section 7.15 #3 forbids to usages of va_list without
a va_end/va_start.  Removing one vfprintf call from io_vprint fixes the problem.

Your program is not valid C, it just works by chance on 32-bit.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID


http://sources.redhat.com/bugzilla/show_bug.cgi?id=366

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug libc/366] Compiling on 64-bit yields segmentation fault when run, on 32-bit runs fine.

[jakub at redhat dot com]

------- Additional Comments From jakub at redhat dot com  2004-09-03 07:41 -------
Well, if you really need two vprintf calls, you can use
       va_list ap2;
       va_copy (ap2, ap);
       vfprintf(out_fp, fmt, ap);
       vfprintf(log_stream_g, fmt, ap2);
       va_end (ap2);


-- 


http://sources.redhat.com/bugzilla/show_bug.cgi?id=366

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug libc/366] Compiling on 64-bit yields segmentation fault when run, on 32-bit runs fine.

[dafrabbit at yahoo dot com]

------- Additional Comments From dafrabbit at yahoo dot com  2004-09-03 13:01 -------
(In reply to comment #3)
> Well, if you really need two vprintf calls, you can use
>        va_list ap2;
>        va_copy (ap2, ap);
>        vfprintf(out_fp, fmt, ap);
>        vfprintf(log_stream_g, fmt, ap2);
>        va_end (ap2);
> 

Well, you then make it difficult to do the task to pass variable number of
arguments from one subroutine to another subroutine. I do't know whether C
forbids that. Your suggestion is to ask people not to use subroutines. I hope it
is not that restrictive.

-- 


http://sources.redhat.com/bugzilla/show_bug.cgi?id=366

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.