[1/3] linux-yocto: Avoid QA check

The kernel has special handling of ${S} and it is therefore expected to be empty
at do_unpack time. For now, ignore this QA check until the kernel unpack process
can be more standardised.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
---
 meta/classes-recipe/kernel-yocto.bbclass | 5 +++++
 1 file changed, 5 insertions(+)

Hi Richard,

On 5/21/24 12:36 PM, Richard Purdie via lists.openembedded.org wrote:
> The kernel has special handling of ${S} and it is therefore expected to be empty
> at do_unpack time. For now, ignore this QA check until the kernel unpack process
> can be more standardised.
> 
> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
> ---
>   meta/classes-recipe/kernel-yocto.bbclass | 5 +++++
>   1 file changed, 5 insertions(+)
> 
> diff --git a/meta/classes-recipe/kernel-yocto.bbclass b/meta/classes-recipe/kernel-yocto.bbclass
> index e8ff7311c32..f741a342d40 100644
> --- a/meta/classes-recipe/kernel-yocto.bbclass
> +++ b/meta/classes-recipe/kernel-yocto.bbclass
> @@ -440,6 +440,11 @@ do_kernel_metadata[depends] = "kern-tools-native:do_populate_sysroot"
>   do_kernel_metadata[file-checksums] = " ${@get_dirs_with_fragments(d)}"
>   do_validate_branches[depends] = "kern-tools-native:do_populate_sysroot"
>   
> +# ${S} doesn't exist for us at unpack
> +do_qa_unpack() {
> +    return
> +}
> +

Shouldn't this rather be for all kernels?

meta/classes-recipe/kernelsrc.bbclass

sets S to STAGING_KERNEL_DIR which is in work-shared.

Won't we have an issue for other kernel recipes outside of kernel-yocto 
that we should fix?

For example, my kernel recipes do not inherit kernel-yocto, so I 
wouldn't benefit from this fix.

Basically, I'm suggesting we add do_qa_unpack() { return } to 
kernelsrc.bbclass instead.

Cheers,
Quentin

On Tue, 2024-05-21 at 13:06 +0200, Quentin Schulz via
lists.openembedded.org wrote:
> Hi Richard,
> 
> On 5/21/24 12:36 PM, Richard Purdie via lists.openembedded.org wrote:
> > The kernel has special handling of ${S} and it is therefore
> > expected to be empty
> > at do_unpack time. For now, ignore this QA check until the kernel
> > unpack process
> > can be more standardised.
> > 
> > Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
> > ---
> >   meta/classes-recipe/kernel-yocto.bbclass | 5 +++++
> >   1 file changed, 5 insertions(+)
> > 
> > diff --git a/meta/classes-recipe/kernel-yocto.bbclass
> > b/meta/classes-recipe/kernel-yocto.bbclass
> > index e8ff7311c32..f741a342d40 100644
> > --- a/meta/classes-recipe/kernel-yocto.bbclass
> > +++ b/meta/classes-recipe/kernel-yocto.bbclass
> > @@ -440,6 +440,11 @@ do_kernel_metadata[depends] = "kern-tools-
> > native:do_populate_sysroot"
> >   do_kernel_metadata[file-checksums] = "
> > ${@get_dirs_with_fragments(d)}"
> >   do_validate_branches[depends] = "kern-tools-
> > native:do_populate_sysroot"
> >   
> > +# ${S} doesn't exist for us at unpack
> > +do_qa_unpack() {
> > +    return
> > +}
> > +
> 
> Shouldn't this rather be for all kernels?
> 
> meta/classes-recipe/kernelsrc.bbclass
> 
> sets S to STAGING_KERNEL_DIR which is in work-shared.
> 
> Won't we have an issue for other kernel recipes outside of kernel-
> yocto 
> that we should fix?
> 
> For example, my kernel recipes do not inherit kernel-yocto, so I 
> wouldn't benefit from this fix.
> 
> Basically, I'm suggesting we add do_qa_unpack() { return } to 
> kernelsrc.bbclass instead.

We may need to move it there, yes.

It was unclear whether the issues I was seeing were linux-yocto
specific or not as some of the code is shared and some is not. I do
know the code is going to benefit from further tweaking/cleanup in
general.

Ultimately I'm hoping we can standardise the unpack process further and
avoid the need to bypass the QA check.

I've tried to send/merge what we need to get the core tests all
working. Beyond that we'll have to work on the other cases/issues as we
fine them (and question whether we should have other test cases too).

Cheers,

Richard