Re: Variables in gdbarch

[Andrew Cagney <ac131313@redhat.com>]

> Hi All,
> 
> I was trying to port GDB to a new target, and it seems,
> almost entirely all the work would be filling up the
> gdbarch structure for the new target.

Careful, the trick is figuring out the few architecture methods that are 
needed.

> I was trying to understand all the variables and
> function pointers in the structure. I started with the
> variables, unfortunately, the Internals Document does
> not seem to explain all those variables and function
> pointers and their significance.
> 
> I did figure out significance of many of the variables,
> but I am stuck up with some. Probably, some of you could
> go through the list and help me understand them. And we
> could build this list and add that to the documentation
> so that others do not have to duplicate my efforts when
> porting for some new targets.
> 
> These are some of the ones, I haven't figured out still:

If the method is deprecated, don't touch it.  Look through 
http://sources.redhat.com/gdb/current/ari/ for additional methods that 
are also going to be deprecated.  Also, only worry about methods that 
can be set with set_gdbarch_...() functions.

> 6.		  nr_data	 unsigned
> 26.		  npc_regnum	 int
> 31.		deprecated_use_generic_dummy_frames	 int
> 33.		 call_dummy_location	 int
> 34.		deprecated_call_dummy_start_offset	 CORE_ADDR
> 35.		deprecated_call_dummy_breakpoint_offset	 CORE_ADDR
> 36.		deprecated_call_dummy_length	 int
> 37.		deprecated_call_dummy_words	 LONGEST *
> 38.		deprecated_sizeof_call_dummy_words	 int
> 39.		deprecated_call_dummy_stack_adjust	 int
> 41.		Believe_pcc_promotion	 int
> 42.		Believe_pcc_promotion_type	 int
> 51.		Cannot_step_breakpoint	 int
> 52.		have_nonsteppable_watchpoint	 int
> 
> Also, are the significances of those I gathered correct?
> I would move on to the function pointers once this is
> finished and wrapped up.
> 
> Any help would be highly appreciated. :-)

I don't think any of the above are needed.

Andrew