This is the mail archive of the gcc-bugs@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: other/8403: ICE in gcc/testsuite/gcc.c-torture/execute/2001012


Hi,

> 
> !       pr_live = ((pr_initial && GET_CODE (pr_initial) == REG)
>                  ? REGNO (pr_initial) != (PR_REG) : 
> regs_ever_live[PR_REG]);
> 
> If pr_initial is set, we already know that 
> regs_ever_live[PR_REG] is nonzero.
> The test for REG should rather be combined with the test of REGNO.

You mean to say like this ?

Index: gcc/config/sh/sh.c
===================================================================
RCS file: /cvsroot/gcc/gcc/gcc/config/sh/sh.c,v
retrieving revision 1.189
diff -c -2 -p -r1.189 sh.c
*** gcc/config/sh/sh.c  16 Dec 2002 18:21:54 -0000      1.189
--- gcc/config/sh/sh.c  20 Dec 2002 05:01:33 -0000
*************** calc_live_regs (count_ptr, live_regs_mas
*** 4422,4426 ****
        rtx pr_initial = has_hard_reg_initial_val (Pmode, PR_REG);
        pr_live = (pr_initial
!                ? REGNO (pr_initial) != (PR_REG) : regs_ever_live[PR_REG]);
      }
    /* Force PR to be live if the prologue has to call the SHmedia
--- 4422,4428 ----
        rtx pr_initial = has_hard_reg_initial_val (Pmode, PR_REG);
        pr_live = (pr_initial
!                ? ((GET_CODE (pr_initial) == REG)
!                    && (REGNO (pr_initial) != (PR_REG)))
!                : regs_ever_live[PR_REG]);
      }
    /* Force PR to be live if the prologue has to call the SHmedia

Best Regards,
  Naveen Sharma.


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