diff mbox series

efi/x86: Ensure that EFI_RUNTIME_MAP is enabled for kexec

Message ID 20230802133704.2146580-1-ardb@kernel.org
State New
Headers show
Series efi/x86: Ensure that EFI_RUNTIME_MAP is enabled for kexec | expand

Commit Message

Ard Biesheuvel Aug. 2, 2023, 1:37 p.m. UTC
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@linux.intel.com>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
---
 arch/x86/Kconfig | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

Comments

Kirill A. Shutemov Aug. 2, 2023, 2:25 p.m. UTC | #1
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@linux.intel.com>
> Signed-off-by: Ard Biesheuvel <ardb@kernel.org>

With the patch 'make oldconfig' updates config to enable
EFI_NEED_RUNTIME_MAP.

Tested-by: Kirill A. Shutemov <kirill.shutemov@linux.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.

>  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
>
Ard Biesheuvel Aug. 2, 2023, 3:14 p.m. UTC | #2
On Wed, 2 Aug 2023 at 16:26, Kirill A . Shutemov
<kirill.shutemov@linux.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@linux.intel.com>
> > Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
>
> With the patch 'make oldconfig' updates config to enable
> EFI_NEED_RUNTIME_MAP.
>
> Tested-by: Kirill A. Shutemov <kirill.shutemov@linux.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
diff mbox series

Patch

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
+
 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