Re: [PATCH] imx8m: soc: cope with existing optee node

["Vincent Stehlé" <vincent.stehle@arm.com>]

On Wed, Mar 12, 2025 at 11:12:35AM -0300, Fabio Estevam wrote:
> Hi Vincent,
> 
> On Mon, Mar 10, 2025 at 9:36 AM Vincent Stehlé <vincent.stehle@arm.com> wrote:
> 
> > --- a/arch/arm/mach-imx/imx8m/soc.c
> > +++ b/arch/arm/mach-imx/imx8m/soc.c
> > @@ -1270,8 +1270,9 @@ static int ft_add_optee_node(void *fdt, struct bd_info *bd)
> >                 }
> >         }
> >
> > +       /* Locate the optee node if it exists or create it. */
> >         subpath = "optee";
> > -       offs = fdt_add_subnode(fdt, offs, subpath);
> > +       offs = fdt_find_or_add_subnode(fdt, offs, subpath);
> >         if (offs < 0) {
> >                 printf("Could not create %s node.\n", subpath);
> >                 return offs;
> 
> This looks correct.

Hi Fabio,

Thank you for the review.

> 
> However, shouldn't we add the optee node only when CONFIG_OPTEE is selected?

Peng has answered this question already, but I would like to add some more
details:

Currently the ft_add_optee_node() function does add the /reserved-memory and
/firmware/optee nodes to the Devicetree passed to Linux when OP-TEE is detected
at runtime (the information is in the rom_pointer[] array).
If we skip this when CONFIG_OPTEE is not set, Linux will not even have the
reserved memory information anymore and the Linux optee driver will not probe.

Best regards,
Vincent.

> 
> --- a/arch/arm/mach-imx/imx8m/soc.c
> +++ b/arch/arm/mach-imx/imx8m/soc.c
> @@ -1235,6 +1235,9 @@ static int ft_add_optee_node(void *fdt, struct
> bd_info *bd)
>         int offs;
>         int ret;
> 
> +       if (!IS_ENABLED(CONFIG_OPTEE))
> +               return 0;
> +
>         /*
>          * No TEE space allocated indicating no TEE running, so no
>          * need to add optee node in dts