[PATCH] arm: spl: flush and disable cache before jumping to OPTEE

[Oleksandr Suvorov]

Hello Tom,

On Fri, Oct 15, 2021 at 2:44 AM Tom Rini <trini at konsulko.com> wrote:
>
> On Sat, Sep 25, 2021 at 05:05:43PM +0300, Oleksandr Suvorov wrote:
>
> > From: Ricardo Salveti <ricardo at foundries.io>
> >
> > Make sure to flush, disable caches and interrupts before jumpint to
> > OPTEE. This fixes the SDP->SPL->OPTEE boot flow on iMX6Q and most
> > likely on some other ARM SoCs.
> >
> > Signed-off-by: Ricardo Salveti <ricardo at foundries.io>
> > Co-developed-by: Oleksandr Suvorov <oleksandr.suvorov at foundries.io>
> > Signed-off-by: Oleksandr Suvorov <oleksandr.suvorov at foundries.io>
> > ---
> >
> >  arch/arm/lib/spl.c | 11 +++++++++++
> >  common/spl/spl.c   | 11 +++++++++--
> >  include/spl.h      |  9 +++++++++
> >  3 files changed, 29 insertions(+), 2 deletions(-)
> >
> > diff --git a/arch/arm/lib/spl.c b/arch/arm/lib/spl.c
> > index 8e2bdf3536..16e3cd557a 100644
> > --- a/arch/arm/lib/spl.c
> > +++ b/arch/arm/lib/spl.c
> > @@ -77,3 +77,14 @@ void __noreturn jump_to_image_linux(struct spl_image_info *spl_image)
> >  }
> >  #endif       /* CONFIG_ARM64 */
> >  #endif
> > +
> > +#if CONFIG_IS_ENABLED(OPTEE)
> > +void __noreturn jump_to_image_optee(struct spl_image_info *spl_image)
> > +{
> > +     /* flush and turn off caches before jumping to OPTEE */
> > +     cleanup_before_linux();
> > +
> > +     spl_optee_entry(NULL, NULL, spl_image->fdt_addr,
> > +                     (void *)spl_image->entry_point);
> > +}
> > +#endif
> > diff --git a/common/spl/spl.c b/common/spl/spl.c
> > index d55d3c2848..f3fab3c4d9 100644
> > --- a/common/spl/spl.c
> > +++ b/common/spl/spl.c
> > @@ -171,6 +171,14 @@ __weak void spl_board_prepare_for_optee(void *fdt)
> >  {
> >  }
> >
> > +#if CONFIG_IS_ENABLED(OPTEE)
> > +__weak void __noreturn jump_to_image_optee(struct spl_image_info *spl_image)
> > +{
> > +     spl_optee_entry(NULL, NULL, spl_image->fdt_addr,
> > +                     (void *)spl_image->entry_point);
> > +}
> > +#endif
> > +
> >  __weak void spl_board_prepare_for_boot(void)
> >  {
> >       /* Nothing to do! */
> > @@ -777,8 +785,7 @@ void board_init_r(gd_t *dummy1, ulong dummy2)
> >       case IH_OS_TEE:
> >               debug("Jumping to U-Boot via OP-TEE\n");
> >               spl_board_prepare_for_optee(spl_image.fdt_addr);
> > -             spl_optee_entry(NULL, NULL, spl_image.fdt_addr,
> > -                             (void *)spl_image.entry_point);
> > +             jump_to_image_optee(&spl_image);
> >               break;
> >  #endif
> >  #if CONFIG_IS_ENABLED(OPENSBI)
> > diff --git a/include/spl.h b/include/spl.h
> > index afbf39bef4..8e7173d294 100644
> > --- a/include/spl.h
> > +++ b/include/spl.h
> > @@ -460,6 +460,15 @@ int spl_board_boot_device(u32 boot_device);
> >   */
> >  void __noreturn jump_to_image_linux(struct spl_image_info *spl_image);
> >
> > +/**
> > + * jump_to_image_linux() - Jump to OP-TEE OS from SPL
> > + *
> > + * This jumps into OP-TEE OS using the information in @spl_image.
> > + *
> > + * @spl_image: Image description to set up
> > + */
> > +void __noreturn jump_to_image_optee(struct spl_image_info *spl_image);
> > +
> >  /**
> >   * spl_start_uboot() - Check if SPL should start the kernel or U-Boot
> >   *
>
> At current, this causes evb-rk3229 and evb-rk3288 to fail to build.  A
> trivial "test on OPTEE_IMAGE" instead leads to warnings about a
> noreturnn function returning.  Please rebase on current master and
> retest your use cases, thanks!

Thank you for the notes! I've fixed both issues in the 2nd version of
the patchset.

>
> --
> Tom

-- 
Best regards,

Oleksandr Suvorov
Software Engineer
T: +380 63 8489656
E: oleksandr.suvorov at foundries.io
W: www.foundries.io