Re: [PATCH] x86/efi: Do not write relocations in efi_arch_relocate_image() first pass

[David Woodhouse <dwmw2@infradead.org>]

[-- Attachment #1.1: Type: text/plain, Size: 1510 bytes --]

On Wed, 2017-08-02 at 05:56 -0600, Jan Beulich wrote:
> > 
> > > 
> > > > 
> > > > David Woodhouse <dwmw2@infradead.org> 08/02/17 1:30 PM >>>
> > --- a/xen/arch/x86/efi/efi-boot.h
> > +++ b/xen/arch/x86/efi/efi-boot.h
> > @@ -87,7 +87,8 @@ static void __init efi_arch_relocate_image(unsigned long delta)
> > case PE_BASE_RELOC_DIR64:
> > if ( in_page_tables(addr) )
> > blexit(L"Unexpected relocation type");
> > -                *(u64 *)addr += delta;
> > +                if ( delta )
> > +                    *(u64 *)addr += delta;
> > break;
> > default:
> > blexit(L"Unsupported relocation type");
>
> While of course this and the other half of the necessary changes are
> independent (i.e. this patch could be taken as is), I really had intended
> to deal with both sides of this issue at once, and hence I'm not entirely
> happy for this partial change to go in on its own. Nevertheless if need
> be it can have my ack.

I am, evidently, still being dense.

This change is sufficient (we believe) to make EFI builds of Xen
actually boot again on current EDK2, is it not?

What is the "other half" of which you speak? You mentioned marking the
section RWX — but for that to be a long-term solution to the issue,
we'd basically have to ensure that we always mark *all* sections which
might have relocations (even .rodata) as writeable, in case UEFI
decides to honour their read-only status at some point in the future.

[-- Attachment #1.2: smime.p7s --]
[-- Type: application/x-pkcs7-signature, Size: 4938 bytes --]

[-- Attachment #2: Type: text/plain, Size: 127 bytes --]

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
https://lists.xen.org/xen-devel