How to make gdb show the original non-mangling function name on disassembly model?
void outputString(const char *str) {
cout << "outputString(const char *str) : " << str << endl;
}
turns out to be
Dump of assembler code for function _Z12outputStringPKc:
0x004013ee <_Z12outputStringPKc+0>: push ebp
0x004013ef <_Z12outputStringPKc+1>: mov ebp,esp
0x004013f1 <_Z12outputStringPKc+3>: sub esp,0x8
0x004013f4 <_Z12outputStringPKc+6>: mov DWORD PTR [esp+4],0x443000
0x004013fc <_Z12outputStringPKc+14>: mov DWORD PTR [esp],0x4463c0
0x00401403 <_Z12outputStringPKc+21>: call 0x43f6e8 <_ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc>
0x00401408 <_Z12outputStringPKc+26>: mov edx,DWORD PTR [ebp+8]
0x0040140b <_Z12outputStringPKc+29>: mov DWORD PTR [esp+4],edx
0x0040140f <_Z12outputStringPKc+33>: mov DWORD PTR [esp],eax
0x00401412 <_Z12outputStringPKc+36>: call 0x43f6e8 <_ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc>
0x00401417 <_Z12outputStringPKc+41>: mov DWORD PTR [esp+4],0x43e4c8
0x0040141f <_Z12outputStringPKc+49>: mov DWORD PTR [esp],eax
0x00401422 <_Z12outputStringPKc+52>: call 0x42e170 <_ZNSolsEPFRSoS_E>
0x00401427 <_Z12outputStringPKc+57>: leave
0x00401428 <_Z12outputStringPKc+58>: ret
End of assembler dump.
All the disassemblies show only the manglinged function names,but its not eaiser for programmer to de-mangling and get the original function names with th开发者_如何学编程e bother to typing info symbol address
for each mangling name met,so are there any methods that could make gdb show non-mangling function names on assembly model?
You could do maint demangle _ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc
at the (gdb)
prompt.
The manual says:
`set print asm-demangle'
`set print asm-demangle on'
Print C++ names in their source form rather than their mangled
form, even in assembler code printouts such as instruction
disassemblies. The default is off.
Unfortunately, it doesn't appear to work:
(gdb) set print asm-demangle on
(gdb) disas
Dump of assembler code for function _Z12outputStringPKc:
0x00000000004009c4 <outputString(char const*)+0>: push %rbp
0x00000000004009c5 <outputString(char const*)+1>: mov %rsp,%rbp
0x00000000004009c8 <outputString(char const*)+4>: sub $0x10,%rsp
0x00000000004009cc <outputString(char const*)+8>: mov %rdi,-0x8(%rbp)
0x00000000004009d0 <outputString(char const*)+12>: mov $0x400bb0,%esi
0x00000000004009d5 <outputString(char const*)+17>: mov $0x6012a0,%edi
0x00000000004009da <outputString(char const*)+22>: callq 0x400798 <_ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@plt>
0x00000000004009df <outputString(char const*)+27>: mov %rax,%rdi
0x00000000004009e2 <outputString(char const*)+30>: mov -0x8(%rbp),%rsi
0x00000000004009e6 <outputString(char const*)+34>: callq 0x400798 <_ZStlsISt11char_traitsIcEERSt13basic_ostreamIcT_ES5_PKc@plt>
0x00000000004009eb <outputString(char const*)+39>: mov %rax,%rdi
0x00000000004009ee <outputString(char const*)+42>: mov $0x4007c8,%esi
0x00000000004009f3 <outputString(char const*)+47>: callq 0x4007b8 <_ZNSolsEPFRSoS_E@plt>
0x00000000004009f8 <outputString(char const*)+52>: leaveq
0x00000000004009f9 <outputString(char const*)+53>: retq
End of assembler dump.
The setting changed how the current function is printed, but not how the functions it calls are printed (which is what I assume you are after).
I think that is a bug in GDB
, please file a bug in bugzilla.
Update:
The bug has been fixed in 2013. With GDB-10.0 the output is:
(gdb) disas 0x555555555169
Dump of assembler code for function _Z12outputStringPKc:
0x0000555555555169 <+0>: push %rbp
0x000055555555516a <+1>: mov %rsp,%rbp
0x000055555555516d <+4>: sub $0x10,%rsp
0x0000555555555171 <+8>: mov %rdi,-0x8(%rbp)
0x0000555555555175 <+12>: lea 0xe8c(%rip),%rax # 0x555555556008
0x000055555555517c <+19>: mov %rax,%rsi
0x000055555555517f <+22>: lea 0x2efa(%rip),%rax # 0x555555558080 <std::cout@GLIBCXX_3.4>
0x0000555555555186 <+29>: mov %rax,%rdi
0x0000555555555189 <+32>: callq 0x555555555040 <std::basic_ostream<char, std::char_traits<char> >& std::operator<< <std::char_traits<char> >(std::basic_ostream<char, std::char_traits<char> >&, char const*)@plt>
0x000055555555518e <+37>: mov %rax,%rdx
0x0000555555555191 <+40>: mov -0x8(%rbp),%rax
0x0000555555555195 <+44>: mov %rax,%rsi
0x0000555555555198 <+47>: mov %rdx,%rdi
0x000055555555519b <+50>: callq 0x555555555040 <std::basic_ostream<char, std::char_traits<char> >& std::operator<< <std::char_traits<char> >(std::basic_ostream<char, std::char_traits<char> >&, char const*)@plt>
0x00005555555551a0 <+55>: mov 0x2e29(%rip),%rdx # 0x555555557fd0
0x00005555555551a7 <+62>: mov %rdx,%rsi
0x00005555555551aa <+65>: mov %rax,%rdi
0x00005555555551ad <+68>: callq 0x555555555050 <std::ostream::operator<<(std::ostream& (*)(std::ostream&))@plt>
0x00005555555551b2 <+73>: nop
0x00005555555551b3 <+74>: leaveq
0x00005555555551b4 <+75>: retq
I don't remember ever finding an automatic way for gdb to do it. I always just copied and pasted the symbol and ran it through the Linux c++filt
utility to demangle.
gdb show the original non-mangling function name on disassembly model ::
you have to do this steps every time whenever you are going to debug. 1. set print demangle on 2. set print asm-demangle on
Else you can create vim ~/.gdbinit file like ~/.vimrc file and set following steps so you no need to do every time on . 1 set print pretty on 2 set print demangle on 3 set print asm-demangle on
精彩评论