[ARM] soft/hard-float preprocessor symbol

Richard Earnshaw wrote:
> The tests should be on arm_pcs_default.  If that has the value
> ARM_PCS_AAPCS_VFP then __ARM_PCS_VFP should be defined.  If it has the
> value ARM_PCS_AAPCS, then __ARM_PCS should be defined.  In other cases,
> I think neither should be defined (leaving the option open to create new
> pre-processor defines in future).
>
> That does leave nothing defined for the IWMMXT variant.  I'm not
> entirely sure what to do about that.  It appears that this just follows
> the base standard for calling, in which case __ARM_PCS probably should
> be defined, but I want to think that case through further before making
> that decision: there may be a subtlety that I've missed.
>
> R.
Hi Richard, I have update the patch as you have suggested.
Please see if this is okay.

Thanks,
Chung-Lin

On Mon, 2010-07-26 at 13:01 +0800, Chung-Lin Tang wrote:
> Richard Earnshaw wrote:
> > The tests should be on arm_pcs_default.  If that has the value
> > ARM_PCS_AAPCS_VFP then __ARM_PCS_VFP should be defined.  If it has the
> > value ARM_PCS_AAPCS, then __ARM_PCS should be defined.  In other cases,
> > I think neither should be defined (leaving the option open to create new
> > pre-processor defines in future).
> >
> > That does leave nothing defined for the IWMMXT variant.  I'm not
> > entirely sure what to do about that.  It appears that this just follows
> > the base standard for calling, in which case __ARM_PCS probably should
> > be defined, but I want to think that case through further before making
> > that decision: there may be a subtlety that I've missed.
> >
> > R.
> Hi Richard, I have update the patch as you have suggested.
> Please see if this is okay.

This is OK, but it needs an updated changelog entry (covering the
changes to arm.c as well).

R.

> 
> Thanks,
> Chung-Lin
> 
> 
> plain text document attachment (pcs2.diff)
> Index: arm.c
> ===================================================================
> --- arm.c	(revision 162525)
> +++ arm.c	(working copy)
> @@ -711,7 +711,7 @@
>     the next function.  */
>  static int after_arm_reorg = 0;
>  
> -static enum arm_pcs arm_pcs_default;
> +enum arm_pcs arm_pcs_default;
>  
>  /* For an explanation of these variables, see final_prescan_insn below.  */
>  int arm_ccfsm_state;
> Index: arm.h
> ===================================================================
> --- arm.h	(revision 162525)
> +++ arm.h	(working copy)
> @@ -94,7 +94,13 @@
>  	if (arm_arch_iwmmxt)				\
>  	  builtin_define ("__IWMMXT__");		\
>  	if (TARGET_AAPCS_BASED)				\
> -	  builtin_define ("__ARM_EABI__");		\
> +	  {						\
> +	    if (arm_pcs_default == ARM_PCS_AAPCS_VFP)	\
> +	      builtin_define ("__ARM_PCS_VFP");		\
> +	    else if (arm_pcs_default == ARM_PCS_AAPCS)	\
> +	      builtin_define ("__ARM_PCS");		\
> +	    builtin_define ("__ARM_EABI__");		\
> +	  }						\
>      } while (0)
>  
>  /* The various ARM cores.  */
> @@ -1641,6 +1647,9 @@
>    ARM_PCS_UNKNOWN
>  };
>  
> +/* Default procedure calling standard of current compilation unit. */
> +extern enum arm_pcs arm_pcs_default;
> +
>  /* A C type for declaring a variable that is used as the first argument of
>     `FUNCTION_ARG' and other related values.  */
>  typedef struct

Richard Earnshaw wrote:
> This is OK, but it needs an updated changelog entry (covering the
> changes to arm.c as well).
>
> R.
Patch committed, here's the updated ChangeLog entry:
        * config/arm/arm.c (arm_pcs_default): Remove static.
        * config/arm/arm.h (TARGET_CPU_CPP_BUILTINS): Define __ARM_PCS or
        __ARM_PCS_VFP to indicate soft/hard-float calling convention.
        (arm_pcs_default): Declare.

Thanks,
Chung-Lin