Re: Different addressing modes?

Re: Different addressing modes?

[Ulrich Weigand]

Fred Cook wrote:

>I am still trying to get addressing modes correctly in my port. 
>Basically my architecture has the following addressing modes:
> 
> 1) absolute
> 2) register indirect
> 3) register + offset
> 4) register + register

I'm assuming this means the following RTX represent valid 
memory operands:

 1)  (mem (symbol_ref ...))
 2)  (mem (reg ...))
 3)  (mem (plus (reg ...) (const_int ...)))
 4)  (mem (plus (reg ...) (reg ...)))

>I want to specify in my define_insn for "call" that 1+2 are allowed 
>and in my define_insn for "movsi" that 2, 3, and 4 are allowed for 
>loads and 2 and 3 for stores.

"call" can be handled specially.  It is fine to have a pattern
that matches
  (call (mem (symbol_ref ...))
even if (mem (symbol_ref ...)) is not a valid memory operand.

I would suggest to do the following:

- Define GO_IF_LEGITIMATE_ADDRESS so that 2), 3), and 4) above
  are accepted.  This means that usual memory accesses (i.e.
  those for loads) simply use the "m" constraint.

- Define an additional constraint letter "S" to denote
  'memory reference without a second register', and define
  EXTRA_CONSTRAINTS to only accept 2) and 3) above for
  this letter.

  As this definition fulfils the requirements "a subset of
  all memory references including all those consisting of
  solely a base register", you can (and must) mark the
  constraint letter "S" as EXTRA_MEMORY_CONSTRAINT.  This
  will cause reload to fix up the address if required
  (note that you *must* provide a load-address type pattern
  that accepts the full range of memory addresses as source).

With this set up, you can define movsi like this:

(define_insn "movsi"
       [(set (match_operand:SI 0 "nonimmediate_operand" "=r,S")
             (match_operand:SI 1 "general_operand" "m,r"))]
       ""
       "@
         ld %1 -> %0
         st %0 %1"
)

All this is actually similar to the s390 setup, so you
might want to check out our backend for comparison ... 
 
Bye,
Ulrich

-- 
  Dr. Ulrich Weigand
  weigand@informatik.uni-erlangen.de

Re: Different addressing modes?

[Peter Barada]

>I am still trying to get addressing modes correctly in my port. Basically my 
>architecture has the following addressing modes:
>
>  1) absolute
>  2) register indirect
>  3) register + offset
>  4) register + register

You may want to go back and read the Compiler internals documentation
where it talks about 'predicate' functions (such as general_operand,
nonimmediate_operand' that you can use to fine-tune what RTL the insns
patterns your port will accept.

As an example, in gcc/config/m68k/m68k.md, the 'pushasi' pattern is
used to push the address of a varaible/function, etc, and you can see
that the destination target predicate of the 'set' is push_operand
which restricts the RTL to only allow a push to the stack as the
destination, and specifies that the source of the 'set' is an
address_operand, a combination of predicates that are valid for the
pea instruction:

(define_insn "pushasi"
  [(set (match_operand:SI 0 "push_operand" "=m")
	(match_operand:SI 1 "address_operand" "p"))]
  ""
  "pea %a1")


-- 
Peter Barada                             Peter.Barada@motorola.com
Wizard                                   781-852-2768 (direct)
WaveMark Solutions(A Motorola Company)   781-270-0193 (fax)

Different addressing modes?

[Fred Cook]

I am still trying to get addressing modes correctly in my port. Basically my 
architecture has the following addressing modes:

  1) absolute
  2) register indirect
  3) register + offset
  4) register + register

I want to specify in my define_insn for "call" that 1+2 are allowed and in 
my define_insn for "movsi" that 2, 3, and 4 are allowed for loads and 2 and 
3 for stores.

Which macros should I define?

What should I accept in GO_IF_LEGITIMATE? All four?

Which constraint letters should I introduce? Four letters for the four 
modes? Which letters are free to use?

How to specify extra constraint in the define_insn pattern? Something 
like...

(define_insn "movsi"
        [(set (match_operand:SI 0 "nonimmediate_operand" "=r,S")
              (match_operand:SI 1 "general_operand" "L,r"))]
        ""
        "@
          ld %1 -> %0
          st %0 %1"
)

Where S accepts a store address and L accepts a load address. Or should it 
be mS and mL? Or Sm or Lm?

Thanks,
Fred

_________________________________________________________________
Tired of spam? Get advanced junk mail protection with MSN 8. 
http://join.msn.com/?page=features/junkmail