On Mon, 2023-12-18 at 00:02 +0800, WANG Xuerui wrote:
> On 12/17/23 21:11, Michał Górny wrote:
> > On Sun, 2023-12-17 at 20:09 +0800, WANG Xuerui wrote:
> > > The several partially-supported arches (those relying on
> > > USE=savedconfig) directly return in src_prepare(), hence previously the
> > > CONFIG_EFI_ZBOOT probing didn't have a chance to run when building for
> > > those arches, leading to wrong kernel artifact path and failed
> > > src_install().
> > > 
> > > Move the probing to near the end of eclass src_configure(), so the flag
> > > correctly reflects the reality in all circumstances.
> > > 
> > > Signed-off-by: WANG Xuerui <xen0n@gentoo.org>
> > > ---
> > >   eclass/kernel-build.eclass | 12 ++++++------
> > >   1 file changed, 6 insertions(+), 6 deletions(-)
> > > 
> > > diff --git a/eclass/kernel-build.eclass b/eclass/kernel-build.eclass
> > > index 6b692dc4f9a0..7a041a8aacdf 100644
> > > --- a/eclass/kernel-build.eclass
> > > +++ b/eclass/kernel-build.eclass
> > > @@ -203,6 +203,12 @@ kernel-build_src_configure() {
> > >   			.config)
> > >   	fi
> > >   
> > > +	# If this is set by USE=secureboot or user config this will have an effect
> > > +	# on the name of the output image. Set this variable to track this setting.
> > > +	if grep -q "CONFIG_EFI_ZBOOT=y" .config; then
> > > +		KERNEL_EFI_ZBOOT=1
> > > +	fi
> > > +
> > >   	mkdir -p "${WORKDIR}"/modprep || die
> > >   	mv .config "${WORKDIR}"/modprep/ || die
> > >   	emake O="${WORKDIR}"/modprep "${MAKEARGS[@]}" olddefconfig
> > > @@ -456,12 +462,6 @@ kernel-build_merge_configs() {
> > >   
> > >   	./scripts/kconfig/merge_config.sh -m -r \
> > >   		.config "${merge_configs[@]}"  || die
> > > -
> > > -	# If this is set by USE=secureboot or user config this will have an effect
> > > -	# on the name of the output image. Set this variable to track this setting.
> > > -	if grep -q "CONFIG_EFI_ZBOOT=y" .config; then
> > > -		KERNEL_EFI_ZBOOT=1
> > > -	fi
> > >   }
> > >   
> > >   fi
> > This doesn't seem correct.  It won't work if KERNEL_EFI_ZBOOT is set by
> > one of the merged configs.
> But isn't the src_configure phase building upon the config already 
> merged in src_prepare? The alternative would be keeping the 
> kernel-build_merge_configs logic intact, but then it has to be 
> duplicated for every early return branch in each ebuild's src_prepare, 
> which I think is too excessive...
> 

Ah, sorry, you're right.  I didn't notice these were two different
functions!

-- 
Best regards,
Michał Górny