Re: reload problems....

[thomas joseph <thomascanny at yahoo dot co dot nz>]

Hi Alan,

  thanks for your reply. I have already defined
 SMALL_REGISTER_CLASSES and CLASSES_LIKELY_SPILLED
macros. 

   Do I need to define LEGITIMIZE_RELOAD_ADDRESS macro
 ?? 

   I defined LEGITIMIZE_ADDRESS macro to make the 
illegitimate addresses as legitimate. It handles only
one case ,  in the indexed addressing mode when the 
displacement goes out of range (-128 to 127).
It is defined as :

#define LEGITIMIZE_ADDRESS (OLDX, MODE, WIN) \ {      
                             \
  if (GET_CODE (X) == PLUS  && GET_CODE (XEXP (X, 0))
== REG   && GET_CODE (XEXP (X, 1))  == CONST_INT \ 
  && !VALID_CONSTANT_OFFSET_P (XEXP (X, 1), MODE))\ 
{\                                                    
      rtx xop0 = XEXP (X, 0);                        \
      rtx xop1 = XEXP (X, 1);                        \
     xop1 = force_reg (HImode, xop1);  \
     (X) = force_reg (Pmode, gen_rtx_PLUS (Pmode,
xop0, xop1)); \
      goto WIN;                         \
    }                                \
}                                           


   I could not identify the potential cases for 
 LEGITIMIZE_RELOAD_ADDRESS. May be I need to look for
that !!!! any suggestions ....

   I have been working on this port for the last six
months. How much time is sufficient enough to build a
newport (for a fresh compilers engineer).


thanks and regards,

V.Brahmaiah















 --- Alan Lehotsky <apl@alum.mit.edu> wrote: > At 2:30
AM +1200 7/21/01, thomas joseph wrote:
> 
> >Hi All,
> > I am working for a new gcc port. I am towards
> >finishing the port. I am compiling newlib.
> >It has few reload problems. I am just showing you
> the
> >kind of problems that I am facing.....
> >
> >   On my machine I have two index registers.Out of
> >them
> >one is  frame pointer. and I can use the other for
> >reloading and similar purposes.
> 
> 
> 	I've done a couple of ports with these kinds of
> limitations.  It is really pushing the
> 	limits of gcc and reload.
> 
> 	First, there are some obvious things you should
> check, like to make sure that
> 
> 		1/ SMALL_REGISTER_CLASSES is defined
> 		2/ CLASS_LIKELY_SPILLED_P is defined for the class
> containing your index register.
> 
> 	But, I bet that you've already done that, or you'd
> have a LOT more failures.
> 
> 	The next thing is to look at defining the
> LEGITIMIZE_RELOAD_ADDRESS macro and using it
> 	to force the kind of reloads you need.
> 
> 	To actually look at what reload is "thinking", you
> call run the compiler under gdb and after it
> 	aborts, you call do
> 
> 		call debug_reload()
> 
> 	and it will print out a nicely formatted dump of
> the current reloads for the instruction.
> 
> 	You may find it necessary to alter your .MD file to
> prevent creating instructions with
> 	too many reloads.
> 
> 	For example, on one of my ports, we had to change
> the .md file, so that there were NEVER two
> 	memory operands in any INSN, despite the fact that
> we we had a two-address machine.
> 
> 	I wonder though if the "fixed or forbidden
> register" message indicates that you've marked the
> IX
> 	register as "fixed".  In that case, I believe that
> the compiler won't allocate it and THAT's why
> 	you're crashing.......
> 
> -- 
>
------------------------------------------------------------------------
> 
> 		    Quality Software Management
> 		http://home.earthlink.net/~qsmgmt
> 			apl@alum.mit.edu
> 			(978)287-0435 Voice
> 			(978)808-6836 Cell
> 			(978)287-0436 Fax
> 
> 	Software Process Improvement and Management
> Consulting
> 	     Language Design and Compiler Implementation