Re: [OpenBIOS] [commit] r915 - trunk/openbios-devel/arch/sparc32

[Artyom Tarasenko]

On Mon, Oct 18, 2010 at 9:32 PM, repository service  wrote:
> Author: mcayland
> Date: Mon Oct 18 21:32:16 2010
> New Revision: 915
> URL: http://tracker.coreboot.org/trac/openbios/changeset/915
>
> Log:
> Fix the placing of Forth arguments on the stack when calling obp_fortheval_v2() via romvec on SPARC32.
>
> The extra stack arguments are actually placed within %o1-%o5 but unfortunately there doesn't seem to be a way of passing the
> number of parameters using the romvec API. Hence we go through the argument list and start pushing arguments onto the Forth
> stack from the first non-zero argument before executing the Forth string.
>
> Signed-off-by: Mark Cave-Ayland
>
> Modified:
>   trunk/openbios-devel/arch/sparc32/openprom.h
>   trunk/openbios-devel/arch/sparc32/romvec.c
> Modified: trunk/openbios-devel/arch/sparc32/openprom.h
>
> --- trunk/openbios-devel/arch/sparc32/openprom.h        Sun Oct 17 14:15:44 2010        (r914)
> +++ trunk/openbios-devel/arch/sparc32/openprom.h        Mon Oct 18 21:32:16 2010        (r915)
> @@ -124,7 +124,7 @@
>        /* Evaluate a forth string, not different proto for V0 and V2->up. */
>        union {
>                void (*v0_eval)(int len, char *str);
> void (*v2_eval)(char *str);
> void (*v2_eval)(char *str, int arg0, int arg1, int arg2, int arg3, int arg4);
>
> } pv_fortheval;
>
> const struct linux_arguments_v0 * const *pv_v0bootargs;
> Modified: trunk/openbios-devel/arch/sparc32/romvec.c
>
> --- trunk/openbios-devel/arch/sparc32/romvec.c  Sun Oct 17 14:15:44 2010        (r914)
> +++ trunk/openbios-devel/arch/sparc32/romvec.c  Mon Oct 18 21:32:16 2010        (r915)
> @@ -412,15 +412,39 @@
>     return 0;
>  }
>
> -static void obp_fortheval_v2(char *str)
> +static void obp_fortheval_v2(char *str, int arg0, int arg1, int arg2, int arg3, int arg4)
>  {
> // for now, move something to the stack so we
> // don't get a stack underrun.
> //
> // FIXME: find out why solaris doesnt put its stuff on the stack
> //
> fword("0");
> fword("0");
> int pusharg = 0;
> // It seems Solaris passes up to 5 arguments which should be pushed onto the Forth
> // stack for execution. However the API doesn't provide for a way to pass the number
> // of arguments actually passed. Hence we start at arg4 and then starting from the
> // first non-zero argument, we push all subsequent arguments onto the stack down to
> // arg0.

Why do you think it's up to 5, and not just always 5?
> if (arg4) {
> PUSH(arg4);
> pusharg = 1;
> }
> if (arg3 || pusharg == 1 ) {
> PUSH(arg3);
> pusharg = 1;
> }
> if (arg2 || pusharg == 1) {
> PUSH(arg2);
> pusharg = 1;
> }
> if (arg1 || pusharg == 1 ) {
> PUSH(arg1);
> pusharg = 1;
> }
> if (arg0 || pusharg == 1) {
> PUSH(arg0);
> }
>
> DPRINTF("obp_fortheval_v2(%s)\n", str);
>   push_str(str);