NACK: [SRU][K/Unstable][PATCH 0/1] Merge riscv64 config and annotations
Emil Renner Berthing
emil.renner.berthing at canonical.com
Fri Jun 24 13:40:20 UTC 2022
On Fri, 24 Jun 2022 at 14:12, Andrea Righi <andrea.righi at canonical.com> wrote:
> On Fri, Jun 24, 2022 at 09:31:09AM +0100, Emil Renner Berthing wrote:
> > [Impact]
> >
> > * We would like to eventually build and test the unstable and generic
> > kernels on the riscv64 architecture. Fixing this is a step towards
> > that.
> >
> > * Having the riscv64 configuration alongside the other architectures
> > should ensure it doesn't drift away from the generic configuration
> > over time.
> >
> > * This allows us to do backports more easily.
> >
> > [Test Plan]
> >
> > * After merging riscv64 the generated configuration for other
> > architectures should not change.
> >
> > * The generated riscv64 configuration should be as close as possible to
> > the current linux-riscv configuration. Right now unstable is based on
> > 5.19 while the riscv kernel is based on 5.18, so it can't be identical.
> >
> > [Where problems could occur]
> >
> > * Adding the riscv64 annotations may subtly change annotations from other
> > architectures.
> >
> > * Adding the riscv64 configuration and annotations might confuse tooling
> > scripts that are not prepared to handle the new architecture.
> >
> > [Other Info]
> >
> > * In the near future we'd also like to base the StarFive and Allwinner
> > (Nezha) kernels on the same series that will become the generic kinetic
> > kernel.
> >
> > Emil Renner Berthing (1):
> > UBUNTU: [Config] Merge riscv64 config and annotations
>
> This doesn't apply anymore to linux-unstable, that is now moving toward
> 5.19.
>
> And in general, I'm a bit skeptical to apply this to the generic kernel,
> it just means more work to have a generic kernel ready when we move to a
> new upstream kernel version (because of the extra work to re-align the
> config/annotations also for riscv).
Sorry, I think I might have just got the tag wrong. I only meant for this to be
applied to linux-unstable. As far as I can see it doesn't apply only because of
the removed "family=ubuntu" line in debian/scripts/misc/kernelconfig.
Shall I send a new version on top of that?
> I'd be definitely better to maintain riscv configs/annotations in a
> separate derivative kernel from this point of view...
>
> -Andrea
More information about the kernel-team
mailing list