linux-kernel - Re: [PATCH] efi/x86: Ensure that EFI_RUNTIME

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <CAMj1kXEkbHjmkZuDD3NMi5r-D_kyi5LOSnpbYFTv=NjKHjahbw@mail.gmail.com>
Date:   Wed, 2 Aug 2023 17:14:23 +0200
From:   Ard Biesheuvel <ardb@...nel.org>
To:     "Kirill A . Shutemov" <kirill.shutemov@...ux.intel.com>
Cc:     linux-efi@...r.kernel.org, linux-kernel@...r.kernel.org,
        x86@...nel.org
Subject: Re: [PATCH] efi/x86: Ensure that EFI_RUNTIME_MAP is enabled for kexec

On Wed, 2 Aug 2023 at 16:26, Kirill A . Shutemov
<kirill.shutemov@...ux.intel.com> wrote:
>
> On Wed, Aug 02, 2023 at 03:37:04PM +0200, Ard Biesheuvel wrote:
> > CONFIG_EFI_RUNTIME_MAP needs to be enabled in order for kexec to be able
> > to provide the required information about the EFI runtime mappings to
> > the incoming kernel, regardless of whether kexec_load() or
> > kexec_file_load() is being used. Without this information, kexec boot in
> > EFI mode is not possible.
> >
> > The CONFIG_EFI_RUNTIME_MAP option is currently directly configurable if
> > CONFIG_EXPERT is enabled, so that it can be turned on for debugging
> > purposes even if KEXEC is. However, the upshot of this is that it can
> > also be disabled even when it shouldn't.
> >
> > So tweak the Kconfig declarations to avoid this situation.
> >
> > Reported-by: Kirill A. Shutemov <kirill.shutemov@...ux.intel.com>
> > Signed-off-by: Ard Biesheuvel <ardb@...nel.org>
>
> With the patch 'make oldconfig' updates config to enable
> EFI_NEED_RUNTIME_MAP.
>
> Tested-by: Kirill A. Shutemov <kirill.shutemov@...ux.intel.com>
>
> Thanks.
> > ---
> >  arch/x86/Kconfig | 6 +++++-
> >  1 file changed, 5 insertions(+), 1 deletion(-)
> >
> > diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
> > index 7422db4097701c96..616498cdc91e8f01 100644
> > --- a/arch/x86/Kconfig
> > +++ b/arch/x86/Kconfig
> > @@ -2027,10 +2027,14 @@ config EFI_MAX_FAKE_MEM
> >         Ranges can be set up to this value using comma-separated list.
> >         The default value is 8.
> >
> > +config EFI_NEED_RUNTIME_MAP
> > +     def_bool y
> > +     depends on EFI && KEXEC_CORE
> > +     select EFI_RUNTIME_MAP
> > +
>
> Just curious, why not extend 'config EFI' with 'select EFI_RUNTIME_MAP if KEXEC_CORE'?
> It seems functionally equivalent, but more concise.
>

Yeah, agreed - I didn't realize every arch has its own 'config EFI' definition.

> >  config EFI_RUNTIME_MAP
> >       bool "Export EFI runtime maps to sysfs" if EXPERT
> >       depends on EFI
> > -     default KEXEC_CORE
> >       help
> >         Export EFI runtime memory regions to /sys/firmware/efi/runtime-map.
> >         That memory map is required by the 2nd kernel to set up EFI virtual
> > --
> > 2.39.2
> >
>
> --
>   Kiryl Shutsemau / Kirill A. Shutemov