This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Is this pattern really correct?


On Wednesday 01 December 2004 13:25, Jan Hubicka wrote:
> > This is from i386.md:
> >
> > ;; For 64BIT abi we always round up to 8 bytes.
> > (define_insn "*pushhi2_rex64"
> >   [(set (match_operand:HI 0 "push_operand" "=X")
> >         (match_operand:HI 1 "nonmemory_no_elim_operand" "ri"))]
> >   "TARGET_64BIT"
> >   "push{q}\t%q1"
> >   [(set_attr "type" "push")
> >    (set_attr "mode" "QI")])
> >
> > How can a pushhi be in QImode?
>
> Yep, this is not correct, the mode should be DI

OK, let's schedule this patch for stage1 GCC 4.1.

Gr.
Steven

	* config/i386/i386.md (pushhi2_rex64): Mode attribute should be DI.

Index: i386.md
===================================================================
RCS file: /cvs/gcc/gcc/gcc/config/i386/i386.md,v
retrieving revision 1.565
diff -c -3 -p -r1.565 i386.md
*** i386.md	23 Nov 2004 01:22:58 -0000	1.565
--- i386.md	1 Dec 2004 22:37:08 -0000
***************
*** 1333,1339 ****
    "TARGET_64BIT"
    "push{q}\t%q1"
    [(set_attr "type" "push")
!    (set_attr "mode" "QI")])
  
  (define_insn "*movhi_1"
    [(set (match_operand:HI 0 "nonimmediate_operand" "=r,r,r,m")
--- 1333,1339 ----
    "TARGET_64BIT"
    "push{q}\t%q1"
    [(set_attr "type" "push")
!    (set_attr "mode" "DI")])
  
  (define_insn "*movhi_1"
    [(set (match_operand:HI 0 "nonimmediate_operand" "=r,r,r,m")


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]