Re: [PATCH 6.12.y 0/3] arm64: Early system register initialization fixes

[Marc Zyngier <maz@kernel.org>]

On Fri, 19 Dec 2025 10:21:20 +0000,
Wei-Lin Chang <weilin.chang@arm.com> wrote:
> 
> Hi,
> 
> This series is a v6.12-only backport (based on v6.12.62) of fixes for
> system register initialization issues affecting protected KVM on arm64.
> This affects some contemporary and upcoming hardware which will run the
> v6.12.y stable kernel, or something derived from it, such as the Android
> common kernel.
> 
> The FEAT_E2H0 patches fix code introduced after v6.6, and so only
> need to be backported to v6.12.
> 
> The SCTLR_EL1 patch fixes code introduced in v5.11, but practically
> speaking only affects recent hardware which is unlikely to run
> something older than v6.12.
> 
> Note: Marc Zyngier performed the initial backport, which I have
> rebased and tested, hence both of our sign-offs being added to the
> tags from the upstream commits.
> 
> I have tested the backport and observed they solve the problems as
> expected.
> 
> Ahmed Genidi (1):
>   KVM: arm64: Initialize SCTLR_EL1 in __kvm_hyp_init_cpu()
> 
> Marc Zyngier (1):
>   arm64: Revamp HCR_EL2.E2H RES1 detection
> 
> Mark Rutland (1):
>   KVM: arm64: Initialize HCR_EL2.E2H early
> 
>  arch/arm64/include/asm/el2_setup.h   | 57 +++++++++++++++++++++++++---
>  arch/arm64/kernel/head.S             | 22 ++---------
>  arch/arm64/kvm/hyp/nvhe/hyp-init.S   | 10 +++--
>  arch/arm64/kvm/hyp/nvhe/psci-relay.c |  3 ++
>  4 files changed, 65 insertions(+), 27 deletions(-)
> 

Great, thanks for going through this and having tested it!

Hopefully Greg will be able to take it at the next opportunity for a
stable 6.12 release.

Cheers,

	M.

-- 
Without deviation from the norm, progress is not possible.