Re: [PATCH] Change gdbserver to use existing gdbsupport

[Pedro Alves <palves@redhat.com>]

On 2/19/20 5:06 PM, Tom Tromey wrote:
> This changes the gdbserver build to use the gdbsupport that was built
> for gdb.
> 
> The most notable change here is that CORE_ADDR is always a 64-bit
> type.  This makes it so that gdb acts as if it were always built in
> 64-bit mode.
> 
> I tested this locally, and I tried it with a few crosses using
> AdaCore's internal test suite (including a 32-bit build).  However, I
> don't have a very clear idea of what code might break due to the
> switch from a 32- to 64-bit CORE_ADDR.

I'm surprised you didn't hit any problem, because last time we chatted
about it, I quickly tried it, and ran into a small number of places
that needed adjustment.

I just tried it again, and I get, for example, with
a i686-pc-linux-gnu build:

$ make V=1 remote.o
g++ -m32 -x c++  -g3 -O0 (....) -c -o remote.o -MT remote.o -MMD -MP -MF ./.deps/remote.Tpo /home/pedro/gdb/binutils-gdb/src/gdb/remote.c
In file included from /home/pedro/gdb/binutils-gdb/src/gdb/defs.h:37:0,
                 from /home/pedro/gdb/binutils-gdb/src/gdb/remote.c:22:
/home/pedro/gdb/binutils-gdb/src/gdb/remote.c: In member function ‘virtual void remote_target::download_tracepoint(bp_location*)’:
../bfd/bfd.h:198:62: error: format ‘%lx’ expects argument of type ‘long unsigned int’, but argument 3 has type ‘CORE_ADDR {aka long long unsigned int}’ [-Werror=format=]
 #define sprintf_vma(s,x) sprintf (s, "%08" BFD_VMA_FMT "x", x)
                                                              ^
/home/pedro/brno/pedro/gdb/binutils-gdb/src/gdb/remote.c:12817:3: note: in expansion of macro ‘sprintf_vma’
   sprintf_vma (addrbuf, tpaddr);
   ^~~~~~~~~~~
/home/pedro/gdb/binutils-gdb/src/gdb/remote.c: In member function ‘virtual void remote_target::enable_tracepoint(bp_location*)’:
../bfd/bfd.h:198:62: error: format ‘%lx’ expects argument of type ‘long unsigned int’, but argument 3 has type ‘CORE_ADDR {aka long long unsigned int}’ [-Werror=format=]
 #define sprintf_vma(s,x) sprintf (s, "%08" BFD_VMA_FMT "x", x)
...

If I use "make -k", I run into a few more like the above in two or three
more files.

I'd prefer to see this conversion in two separate patches.  One (or more) that
removes the bfd dependency from gdb's gdbsupport, which potentially is a material
change for making GDB always use 64-bit types, and then a separate change
to make gdbserver use top level gdbsupport.

Thanks,
Pedro Alves