Re: EFI boot crash regression (was: Re: 5.6-### doesn't boot)

[Ard Biesheuvel]

On Sun, 2 Feb 2020 at 10:22, Ingo Molnar <mingo@xxxxxxxxxx> wrote:
>
>
> * Ard Biesheuvel <ard.biesheuvel@xxxxxxxxxx> wrote:
>
> > Hello JÃrg,
> >
> > Could you please try whether the change below fixes the issue?
> >
> > diff --git a/arch/x86/platform/efi/efi.c b/arch/x86/platform/efi/efi.c
> > index 59f7f6d60cf6..ae923ee8e2b4 100644
> > --- a/arch/x86/platform/efi/efi.c
> > +++ b/arch/x86/platform/efi/efi.c
> > @@ -308,7 +308,7 @@ static void __init efi_clean_memmap(void)
> >                         .phys_map = efi.memmap.phys_map,
> >                         .desc_version = efi.memmap.desc_version,
> >                         .desc_size = efi.memmap.desc_size,
> > -                       .size = data.desc_size * (efi.memmap.nr_map - n_removal),
> > +                       .size = efi.memmap.desc_size * (efi.memmap.nr_map - n_removal),
> >                         .flags = 0,
>
> Oh, I actually noticed this one, but convinced myself that it's correct,
> because GCC didn't warn about uninitialized data.
>
> But maybe in this weird case data.desc_size as used within its own
> initializer is zero?
>

Something like that, yes. Note that size and desc_size appear in
opposite order in the struct definition, and this may also affect how
the compiler handles this.