Re: [PATCH 2/2] x86/trampoline: Rationalise the constants to describe the size

[Frediano Ziglio <frediano.ziglio@xxxxxxxxx>]

On Wed, Nov 13, 2024 at 11:36 AM Andrew Cooper
<andrew.cooper3@xxxxxxxxxx> wrote:
>
> On 13/11/2024 9:30 am, Andrew Cooper wrote:
> > diff --git a/xen/arch/x86/efi/efi-boot.h b/xen/arch/x86/efi/efi-boot.h
> > index 7930b7c73892..9d3f2b71447e 100644
> > --- a/xen/arch/x86/efi/efi-boot.h
> > +++ b/xen/arch/x86/efi/efi-boot.h
> > @@ -633,7 +633,7 @@ static void __init efi_arch_memory_setup(void)
> >      if ( efi_enabled(EFI_LOADER) )
> >          cfg.size = trampoline_end - trampoline_start;
> >      else
> > -        cfg.size = TRAMPOLINE_SPACE + TRAMPOLINE_STACK_SPACE;
> > +        cfg.size = TRAMPOLINE_SIZE;
>
> Something I forgot to mention.
>
> The EFI_LOADER side of this conditional means that the heap isn't valid.
>
> This includes modelist, vesa_glob_info and vesa_mode_info from video.S,
> but I can't find where they're used at all.
>
> There's a separate struct vesa_mode_info in reloc.c but that is a
> representation of the MB2 vbe_mode_info tag and not the same thing AFAICT.
>

I think MBI data on EFI path is parsed in another path.

Frediano