Re: gdb-7.12 powerpc-rtems4.12-gdb does not build on FreeBSD.

[John Baldwin <jhb@freebsd.org>]

On Thursday, November 24, 2016 11:30:11 AM Chris Johns wrote:
> On 24/11/2016 11:15, John Baldwin wrote:
> > On Thursday, November 24, 2016 10:16:10 AM Chris Johns wrote:
> >> On 23/11/2016 10:38, Chris Johns wrote:
> >>> Hi,
> >>>
> >>> I am getting:
> >>>
> >>> ../sim/ppc/libsim.a(sim_calls.o): In function `sim_io_printf_filtered':
> >>> ../../../gdb-7.12/sim/ppc/sim_calls.c:(.text+0x17c): undefined reference
> >>> to `error'
> >>> ../sim/ppc/libsim.a(sim_calls.o): In function `sim_load':
> >>> ../../../gdb-7.12/sim/ppc/sim_calls.c:(.text+0x291): undefined reference
> >>> to `error'
> >>> ../../../gdb-7.12/sim/ppc/sim_calls.c:(.text+0x31d): undefined reference
> >>> to `error'
> >>> ../../../gdb-7.12/sim/ppc/sim_calls.c:(.text+0x357): undefined reference
> >>> to `error'
> >>> ../sim/ppc/libsim.a(sim_calls.o): In function `sim_create_inferior':
> >>> ../../../gdb-7.12/sim/ppc/sim_calls.c:(.text+0x53e): undefined reference
> >>> to `error'
> >>>
> >>> and errors.o has the following symbols:
> >>>
> >>>    nm
> >>> build/powerpc-rtems4.12-gdb-7.12-x86_64-freebsd10.3-1/build/gdb/errors.o
> >>> 0000000000000120 T _Z14internal_errorPKciS0_z
> >>>                    U _Z15internal_verrorPKciS0_P13__va_list_tag
> >>> 00000000000001a0 T _Z16internal_warningPKciS0_z
> >>>                    U _Z17internal_vwarningPKciS0_P13__va_list_tag
> >>> 0000000000000090 T _Z5errorPKcz
> >>>                    U _Z6verrorPKcP13__va_list_tag
> >>> 0000000000000000 T _Z7warningPKcz
> >>>                    U _Z8vwarningPKcP13__va_list_tag
> >>>
> >>> Is there a C++/C thing happening here between the PCC simulator and GDB?
> >>>
> >>
> >> A follow up. It looks like GDB is being built by cc which is clang on
> >> FreeBSD. I am told by Joel this gdb target builds on Linux.
> >>
> >> I do not know what the extern binding is for gdb, C or C++?
> >
> > clang vs gcc shouldn't really break this.
> 
> Yes I agree.
> 
> > What undefined symbols do you see
> > in the nm of sim_calls.o?
> 
> The symbol is `error` and in my original post I showed the output of nm 
> for the errors.o object file and `error` symbol is mangled. One thing 
> that confuses me is only this symbol is being reported and there must be 
> other calls to gdb.

FYI, I reproduced this with gdb's master branch using
"configure --target  powerpc-rtems4.12-elf" (albeit on FreeBSD 11, though
I have gcc48 installed as /usr/local/bin/gcc and configure found that and
used it instead of cc).  The raw symbol in sim_calls.o is indeed plain C:

% nm sim/ppc/sim_calls.o  | grep error
                 U bfd_get_error
                 U error
0000000000000a20 T sim_io_error

Compiling gdb as C isn't an option on 'master', so it will need a different
fix.  Probably the sims need to be built as C++ now.

> > Looking at sim/ppc/Makefile.in, it only references CC, not CXX, so I would not
> > be surprised if it compiles as C without mangled symbols giving the error you
> > have.
> 
> Interesting. I will take a closer look. Maybe something on Linux is 
> building the sim/ppc as C++ and this does not happen with clang.

I'm not sure why this isn't also broken on Linux.

> I think 7.12 can still be built as C, though you might need to use a
> > configure flag to do so?
> 
> Really? I see code in 'gdbtypes.h' of 'enum type_code { ... }' and that 
> looks C++ and not C to me.

C supports enums.  However, GDB 7.12 has several macros for things like
TRY/CATCH that use try/catch for C++ and setjmp/longjmp in C, etc.  The
next release of GDB will require C++ (and C++11 in fact).

-- 
John Baldwin