Hi Vincent,

On Fri, 18 Sept 2026 at 15:35, Vincent Stehlé <[email protected]> wrote:
>
> On Thu, Sep 17, 2026 at 06:04:07PM +0200, Ilias Apalodimas wrote:
> > Hi Vincent,
>
> Hi Ilias,
>
> Thank you for reviewing; my comments below.
>
> Best regards,
> Vincent.
>
> >
> > On Fri, 4 Sept 2026 at 17:01, Vincent Stehlé <[email protected]> wrote:
> > >
> > > Adjust the EBBR 2.1 conformance profile ECPT entry configuration, and
> > > prepare to add more:
> > > - Introduce the HAVE_EFI_EBBR_2_1 non-visible configuration symbol to
> > >   compute the actual system conformance status with the dependencies.
> > >   This will make it easier to add more versions with a chain of
> > >   dependencies, while avoiding a chain of nested submenus.
> > > - Adjust version 2.1 dependencies to better match the specification.
> > >   Relax the dependency on the boot manager, which was introduced in
> > >   version 2.3, and make the list of required UEFI protocols and services
> > >   more complete.
> > > - Reword the EFI_EBBR_2_1_CONFORMANCE configuration entry title and help
> > >   text, mostly to separate "EBBR" from "v2.1" and to attenuate the
> > >   redundant "table" mention.
> > >
> > > We do the following additional configurations dependencies
> > > approximations:
> > >
> > >       Configuration symbol  Used to approximate
> > >   ------------------------  ------------------------------------
> > >   EFI_HAVE_CAPSULE_SUPPORT  Firmware update is performed in-band
> > >                     DM_RTC  An RTC is present
> > >
> > > While at it, refine the ECPT help text to attenuate the redundant
> > > "table" mention there as well.
> > >
> > > Signed-off-by: Vincent Stehlé <[email protected]>
> > > Cc: Heinrich Schuchardt <[email protected]>
> > > Cc: Ilias Apalodimas <[email protected]>
> > > Cc: Tom Rini <[email protected]>
> > > ---
> > >  lib/efi_loader/Kconfig | 36 +++++++++++++++++++++++++++---------
> > >  1 file changed, 27 insertions(+), 9 deletions(-)
> > >
> > > diff --git a/lib/efi_loader/Kconfig b/lib/efi_loader/Kconfig
> > > index 4cb13ae7c8a..40380876959 100644
> > > --- a/lib/efi_loader/Kconfig
> > > +++ b/lib/efi_loader/Kconfig
> > > @@ -557,23 +557,41 @@ config EFI_ESRT
> > >         help
> > >           Enabling this option creates the ESRT UEFI system table.
> > >
> > > +config HAVE_EFI_EBBR_2_1
> > > +       bool
> > > +       depends on EFI_LOADER_HII
> > > +       depends on EFI_RISCV_BOOT_PROTOCOL || !RISCV
> > > +       depends on EFI_RNG_PROTOCOL || !DM_RNG
> > > +       depends on EFI_UNICODE_COLLATION_PROTOCOL2
> > > +       depends on BOOTMETH_EFILOADER
> >
> > Similar comments here. The BOOTMETH_* shouldn't be a requirement
>
> I think you are right: depending on the boot method is a bit too
> specific.
>
> The original intention was to advertise a system as compliant only if we
> could guarantee that it would boot automatically with UEFI (in this
> patch with the default filename, and in previous patch with the boot
> manager).
> There did not seem to be an easy way to guarantee that with the
> configurations dependencies.
>
> What you see here is a second attempt, trying to -at least- avoid
> advertising as compliant a system, which we know cannot boot
> automatically with UEFI.
> Even for that depending on the BOOTMETH_* configurations is probably not
> good enough after all.

Fair enough but that's a bit of a chicken-and-egg situation, tbh...
and it's not straightforward.
For example, you *can* boot directly without bootmeth* if your default
command is 'bootefi bootmgr'. And I think it's better to document that
somewhere, rather than not exposing the ECPT entry for people that opt
out of bootmeth*

>
> I think you mentioned that you or Heinrich could amend the patches
> before merging; do not hesitate to let me know if you prefer that I send
> a v2 series with the dependency here and in the previous patch removed.
>

I found another place with the same dependancy, so please send a v2 if
it's not too much trouble.

Cheers
/Ilias
> >
> > Cheers
> > /Ilias
> > > +       depends on !EFI_VARIABLE_NO_STORE
> > > +       depends on BOOTM_EFI
> > > +       depends on EFI_BINARY_EXEC
> > > +       depends on EFI_DEVICE_PATH_TO_TEXT
> > > +       depends on EFI_DEVICE_PATH_UTIL
> > > +       depends on EFI_GET_TIME || !DM_RTC
> > > +       depends on EFI_PARTITION
> > > +       depends on EFI_SET_TIME || !DM_RTC
> > > +       depends on EFI_UNICODE_CAPITALIZATION
> > > +       depends on EFI_ESRT || !EFI_HAVE_CAPSULE_SUPPORT
> > > +       depends on EFI_RUNTIME_UPDATE_CAPSULE || !EFI_HAVE_CAPSULE_SUPPORT
> > > +       depends on EFI_CAPSULE_FIRMWARE || !EFI_HAVE_CAPSULE_SUPPORT
> > > +       default y
> > > +
> > >  config EFI_ECPT
> > >         bool "Enable the UEFI ECPT generation"
> > >         default y
> > >         help
> > > -         Enabling this option created the ECPT UEFI table.
> > > +         Enabling this option creates the EFI Conformance Profiles Table
> > > +         (ECPT).
> > >
> > >  config EFI_EBBR_2_1_CONFORMANCE
> > > -       bool "Add the EBBRv2.1 conformance entry to the ECPT table"
> > > -       depends on BOOTMETH_EFI_BOOTMGR
> > > -       depends on EFI_ECPT
> > > -       depends on EFI_LOADER_HII
> > > -       depends on EFI_RISCV_BOOT_PROTOCOL || !RISCV
> > > -       depends on EFI_RNG_PROTOCOL || !DM_RNG
> > > -       depends on EFI_UNICODE_COLLATION_PROTOCOL2
> > > +       bool "Add the EBBR v2.1 conformance entry to the ECPT"
> > > +       depends on EFI_ECPT && HAVE_EFI_EBBR_2_1
> > >         default y
> > >         help
> > > -         Enabling this option adds the EBBRv2.1 conformance entry to the 
> > > ECPT UEFI table.
> > > +         Enabling this option adds the Embedded Base Boot Requirements 
> > > (EBBR)
> > > +         v2.1 conformance entry to the ECPT.
> > >
> > >  config EFI_CONSOLE_DISABLE_ANSI
> > >         bool "Disable ANSI escape sequence queries for console size"
> > >
> > > --
> > > 2.53.0
> > >

Reply via email to