GNU C/386 v1.26 bug fix

[James Van Artsdalen]

This fix (from RMS) seems to fix the floating point problems I have
been having.  There is still an open problem with register allocation
when optimizing, but it doesn't seem to effect non-optimizing
compilation.
---
*** /tmp/,RCSt1a03009	Tue Aug 23 23:05:03 1988
--- i386.md	Tue Aug 23 21:04:28 1988
***************
*** 412,418
  }")
  
  (define_insn "movdf"
!   [(set (match_operand:DF 0 "general_operand" "=rmf,rmf,!rm")
  	(match_operand:DF 1 "general_operand" "m,fr,F"))]
    ""
    "*

--- 412,418 -----
  }")
  
  (define_insn "movdf"
!   [(set (match_operand:DF 0 "general_operand" "=&rmf,rmf,!rm")
  	(match_operand:DF 1 "general_operand" "m,fr,F"))]
    ""
    "*
***************
*** 427,433
  }")
  
  (define_insn "movdi"
!   [(set (match_operand:DI 0 "general_operand" "=r,rm")
  	(match_operand:DI 1 "general_operand" "m,ri"))]
    ""
    "*

--- 427,433 -----
  }")
  
  (define_insn "movdi"
!   [(set (match_operand:DI 0 "general_operand" "=&r,rm")
  	(match_operand:DI 1 "general_operand" "m,ri"))]
    ""
    "*
-- 
James R. Van Artsdalen    ...!uunet!utastro!bigtex!james     "Live Free or Die"
Home: 512-346-2444 Work: 328-0282; 110 Wild Basin Rd. Ste #230, Austin TX 78746