Why doesn't the gdb -s option load the symbol file? - gdb

Question

I am trying to create a core dump and analyze it with gdb. This is the code I wrote to create a core dump.
#include <iostream>
void bar()
{
char *p = (char *) 123;
std::cout << "bar start\n";
std::cout << *p << "\n";
std::cout << "bar end\n";
}
void foo()
{
std::cout << "foo start\n";
bar();
std::cout << "foo end\n";
}
int main()
{
foo();
}
This is my Makefile.
all:
g++ -g foo.cc -o foo
objcopy --only-keep-debug foo foo.dbg
objcopy --strip-debug foo
clean:
rm -rf core* foo
After running make and ./foo, this is what my directory looks like.
# ls
core.28091 foo foo.cc foo.dbg Makefile
I am able to analyze the core dump like this. I launch gdb by specifying the executable and the core file as command line arguments. Then I load the symbols from foo.dbg with the symbol-file foo.dbg command.
[root#centos crash]# gdb foo core.28091
GNU gdb (GDB) Red Hat Enterprise Linux (7.2-60.el6_4.1)
Copyright (C) 2010 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law. Type "show copying"
and "show warranty" for details.
This GDB was configured as "x86_64-redhat-linux-gnu".
For bug reporting instructions, please see:
<http://www.gnu.org/software/gdb/bugs/>...
Reading symbols from /root/lab/crash/foo...(no debugging symbols found)...done.
[New Thread 28091]
Missing separate debuginfo for
Try: yum --disablerepo='*' --enablerepo='*-debug*' install /usr/lib/debug/.build-id/81/a81be2e44c93640adedb62adc93a47f4a09dd1
Reading symbols from /usr/lib64/libstdc++.so.6...(no debugging symbols found)...done.
Loaded symbols for /usr/lib64/libstdc++.so.6
Reading symbols from /lib64/libm.so.6...(no debugging symbols found)...done.
Loaded symbols for /lib64/libm.so.6
Reading symbols from /lib64/libgcc_s.so.1...(no debugging symbols found)...done.
Loaded symbols for /lib64/libgcc_s.so.1
Reading symbols from /lib64/libc.so.6...(no debugging symbols found)...done.
Loaded symbols for /lib64/libc.so.6
Reading symbols from /lib64/ld-linux-x86-64.so.2...(no debugging symbols found)...done.
Loaded symbols for /lib64/ld-linux-x86-64.so.2
Core was generated by `./foo'.
Program terminated with signal 11, Segmentation fault.
#0 0x000000000040076f in bar() ()
Missing separate debuginfos, use: debuginfo-install glibc-2.12-1.132.el6.x86_64 libgcc-4.4.7-4.el6.x86_64 libstdc++-4.4.7-4.el6.x86_64
(gdb) symbol-file foo.dbg
Reading symbols from /root/lab/crash/foo.dbg...done.
(gdb) bt
#0 0x000000000040076f in bar () at foo.cc:8
#1 0x00000000004007b7 in foo () at foo.cc:15
#2 0x00000000004007d1 in main () at foo.cc:21
(gdb) list
12 void foo()
13 {
14 std::cout << "foo start\n";
15 bar();
16 std::cout << "foo end\n";
17 }
18
19 int main()
20 {
21 foo();
(gdb)
However, I want to specify the symbol file name in the command line argument as well. But it doesn't seem to work. See the output below.
[root#centos crash]# gdb -s foo.dbg foo core.28091
GNU gdb (GDB) Red Hat Enterprise Linux (7.2-60.el6_4.1)
Copyright (C) 2010 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law. Type "show copying"
and "show warranty" for details.
This GDB was configured as "x86_64-redhat-linux-gnu".
For bug reporting instructions, please see:
<http://www.gnu.org/software/gdb/bugs/>...
Reading symbols from /root/lab/crash/foo...(no debugging symbols found)...done.
[New Thread 28091]
Missing separate debuginfo for
Try: yum --disablerepo='*' --enablerepo='*-debug*' install /usr/lib/debug/.build-id/81/a81be2e44c93640adedb62adc93a47f4a09dd1
Reading symbols from /usr/lib64/libstdc++.so.6...(no debugging symbols found)...done.
Loaded symbols for /usr/lib64/libstdc++.so.6
Reading symbols from /lib64/libm.so.6...(no debugging symbols found)...done.
Loaded symbols for /lib64/libm.so.6
Reading symbols from /lib64/libgcc_s.so.1...(no debugging symbols found)...done.
Loaded symbols for /lib64/libgcc_s.so.1
Reading symbols from /lib64/libc.so.6...(no debugging symbols found)...done.
Loaded symbols for /lib64/libc.so.6
Reading symbols from /lib64/ld-linux-x86-64.so.2...(no debugging symbols found)...done.
Loaded symbols for /lib64/ld-linux-x86-64.so.2
Core was generated by `./foo'.
Program terminated with signal 11, Segmentation fault.
#0 0x000000000040076f in bar() ()
Missing separate debuginfos, use: debuginfo-install glibc-2.12-1.132.el6.x86_64 libgcc-4.4.7-4.el6.x86_64 libstdc++-4.4.7-4.el6.x86_64
(gdb) bt
#0 0x000000000040076f in bar() ()
#1 0x00000000004007b7 in foo() ()
#2 0x00000000004007d1 in main ()
(gdb) list
No symbol table is loaded. Use the "file" command.
Why does it say no symbol table has been loaded even though I have specified it as an argument to the -s option?

Answer 1

It looks like a bug in gdb. gdb sets symarg to the argument that follows -s, but then later in the code, it unconditionally sets symarg to the executable's name. Proposed minimal diff follows:
$ diff -C 1 main.c.orig main.c
*** main.c.orig 2014-07-29 08:37:42.000000000 -0400
--- main.c 2014-09-02 16:27:54.079039046 -0400
***************
*** 864,866 ****
}
! symarg = argv[optind];
execarg = argv[optind];
--- 864,866 ----
}
! if (symarg == NULL) symarg = argv[optind];
execarg = argv[optind];
***************
*** 877,879 ****
{
! symarg = argv[optind];
execarg = argv[optind];
--- 877,879 ----
{
! if (symarg == NULL) symarg = argv[optind];
execarg = argv[optind];