Commit 2993ce4c by J"orn Rennecke Committed by Joern Rennecke

Undo this change (the problem was actually in reload):

	Fri Jan 23 23:28:59 1998  J"orn Rennecke <amylaar@cygnus.co.uk>
	  * sh.md (movqi_i+1): New peephole.

From-SVN: r17663
parent 2a66ec2b
Thu Feb 5 01:45:19 1998 J"orn Rennecke <amylaar@cygnus.co.uk>
Undo this change (the problem was actually in reload):
Fri Jan 23 23:28:59 1998 J"orn Rennecke <amylaar@cygnus.co.uk>
* sh.md (movqi_i+1): New peephole.
Tue Feb 3 01:11:12 1998 Jeffrey A Law (law@cygnus.com) Tue Feb 3 01:11:12 1998 Jeffrey A Law (law@cygnus.com)
* jump.c (jump_optimize): Lose calls to modified_in_p they are * jump.c (jump_optimize): Lose calls to modified_in_p they are
......
...@@ -1885,18 +1885,6 @@ ...@@ -1885,18 +1885,6 @@
lds %1,%0" lds %1,%0"
[(set_attr "type" "move,load,store,move,move,move")]) [(set_attr "type" "move,load,store,move,move,move")])
;; For Big-endian, reload now sometimes generates something like
;; (insn 2107 2104 2099 (set (reg:QI 3 r3)
;; (subreg:QI (mem:SI (reg:SI 0 r0)) 0)) -1 (nil)
;; (nil))
;; To avoid clutter in the movqi pattern, we express this as a peephole.
(define_peephole
[(set (match_operand:QI 0 "arith_reg_operand" "r")
(subreg:QI (match_operand:SI 1 "memory_operand" "m") 0))]
""
"mov.l %1,%0")
(define_expand "movqi" (define_expand "movqi"
[(set (match_operand:QI 0 "general_operand" "") [(set (match_operand:QI 0 "general_operand" "")
(match_operand:QI 1 "general_operand" ""))] (match_operand:QI 1 "general_operand" ""))]
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment