GCC Assembler Modifiers

[Andrew Haley]

Mohamed Bamakhrama writes:

 > I have two questions related to the modifiers that can be used with
 > operands in inline assembly.
 > 1) What does the 'z' mean when added to the operand modifier (e.g. %z0)?

>From i386.md:

;; The special asm out single letter directives following a '%' are:
;; 'z' mov%z1 would be movl, movw, or movb depending on the mode of
;;     operands[1].
;; 'L' Print the opcode suffix for a 32-bit integer opcode.
;; 'W' Print the opcode suffix for a 16-bit integer opcode.
;; 'B' Print the opcode suffix for an 8-bit integer opcode.
;; 'Q' Print the opcode suffix for a 64-bit float opcode.
;; 'S' Print the opcode suffix for a 32-bit float opcode.
;; 'T' Print the opcode suffix for an 80-bit extended real XFmode float opcode.
;; 'J' Print the appropriate jump operand.
;;

 > 2) In the Input operands list, what is the meaning of 'J' (e.g. ... :
 > .... : "Jr" (0) )?

See 14.8, Register Classes, in the gcc internals manual.

The machine-dependent operand constraint letters (`I', `J', `K',
... `P') specify particular ranges of integer values.

Also, see constraints.md:

(define_constraint "J"
  "Integer constant in the range 0 ... 63, for 64-bit shifts."
  (and (match_code "const_int")
       (match_test "IN_RANGE (ival, 0, 63)")))

Andrew.