Re: [PATCH 1/2] Add support for setting disassembler-options in GDB for POWER, ARM and S390

[Peter Bergner <bergner at vnet dot ibm dot com>]

On 11/17/16 1:18 PM, Peter Bergner wrote:
> Ping.  This is the same as my last patch posted here:
>
>     https://sourceware.org/ml/binutils/2016-11/msg00029.html
>
> ...except for that we no longer output "default" for a NULL
> disassembler_options value, and instead just output ''.
> I will note that doing "set disassembler-options" does reset
> the disassembler_options value to NULL.

I will note I only tested ARM and S390 using a cross debugger,
but they seem to work fine.  In addition, I was able to confirm
that using "set disassembler-options ..." and "set arm disassembler ..."
together works fine, meaning you can use them interchangeably.
The only "gotcha", is that "set arm disassembler ..." uses short
option names, while the "set disassebler-options ..." has to use
the canonical long names (eg, "set arm disassembler atpcs" versus
"set disassembler-options reg-names-atpcs").

That above said, a little testing of ARM and S390 by others would
be appreciated.

I did try and take a stab converting i386 over as well, but
opcodes/i386-dis.c:print_insn() contains a lot of global
state that really should be moved into a disassemble_init_i386()
similar to what I did for S390.  There really is no reason to
parse info->disassembler_options for each insn we print.
However, I didn't trust myself at getting that all correct, so
I'll leave it to someone more qualified to do.

Peter