[GIT PULL] KVM/arm64 fixes, take #3

[GIT PULL] KVM/arm64 fixes, take #3

[Marc Zyngier]

Paolo,

Here's the third set of KVM/arm64 fixes for 6.16. The most notable
thing is yet another batch of FP/SVE fixes from Mark, this time
addressing NV, and additionally plugging some missing synchronisation.
The rest is a mix of interrupt stuff (routing change, mishandling of
shadow LRs) and selftest fixes.

Please pull,

	M.

The following changes since commit e04c78d86a9699d136910cfc0bdcf01087e3267e:

  Linux 6.16-rc2 (2025-06-15 13:49:41 -0700)

are available in the Git repository at:

  git://git.kernel.org/pub/scm/linux/kernel/git/kvmarm/kvmarm.git tags/kvmarm-fixes-6.16-3

for you to fetch changes up to 04c5355b2a94ff3191ce63ab035fb7f04d036869:

  KVM: arm64: VHE: Centralize ISBs when returning to host (2025-06-19 13:34:59 +0100)

----------------------------------------------------------------
KVM/arm64 fixes for 6.16, take #3

- Fix another set of FP/SIMD/SVE bugs affecting NV, and plugging some
  missing synchronisation

- A small fix for the irqbypass hook fixes, tightening the check and
  ensuring that we only deal with MSI for both the old and the new
  route entry

- Rework the way the shadow LRs are addressed in a nesting
  configuration, plugging an embarrassing bug as well as simplifying
  the whole process

- Add yet another fix for the dreaded arch_timer_edge_cases selftest

----------------------------------------------------------------
Marc Zyngier (1):
      KVM: arm64: nv: Fix tracking of shadow list registers

Mark Rutland (7):
      KVM: arm64: VHE: Synchronize restore of host debug registers
      KVM: arm64: VHE: Synchronize CPTR trap deactivation
      KVM: arm64: Reorganise CPTR trap manipulation
      KVM: arm64: Remove ad-hoc CPTR manipulation from fpsimd_sve_sync()
      KVM: arm64: Remove ad-hoc CPTR manipulation from kvm_hyp_handle_fpsimd()
      KVM: arm64: Remove cpacr_clear_set()
      KVM: arm64: VHE: Centralize ISBs when returning to host

Sean Christopherson (1):
      KVM: arm64: Explicitly treat routing entry type changes as changes

Zenghui Yu (1):
      KVM: arm64: selftests: Close the GIC FD in arch_timer_edge_cases

 arch/arm64/include/asm/kvm_emulate.h               |  62 ---------
 arch/arm64/include/asm/kvm_host.h                  |   6 +-
 arch/arm64/kvm/arm.c                               |   3 +-
 arch/arm64/kvm/hyp/include/hyp/switch.h            | 147 +++++++++++++++++++--
 arch/arm64/kvm/hyp/nvhe/hyp-main.c                 |   5 +-
 arch/arm64/kvm/hyp/nvhe/switch.c                   |  59 ---------
 arch/arm64/kvm/hyp/vhe/switch.c                    | 107 ++-------------
 arch/arm64/kvm/vgic/vgic-v3-nested.c               |  81 ++++++------
 .../selftests/kvm/arm64/arch_timer_edge_cases.c    |  16 ++-
 9 files changed, 215 insertions(+), 271 deletions(-)

Re: [GIT PULL] KVM/arm64 fixes, take #3

[Paolo Bonzini]

On Thu, Jun 19, 2025 at 3:01 PM Marc Zyngier <maz@kernel.org> wrote:
>
> Paolo,
>
> Here's the third set of KVM/arm64 fixes for 6.16. The most notable
> thing is yet another batch of FP/SVE fixes from Mark, this time
> addressing NV, and additionally plugging some missing synchronisation.
> The rest is a mix of interrupt stuff (routing change, mishandling of
> shadow LRs) and selftest fixes.
>
> Please pull,

Done, thanks.

Paolo

>         M.
>
> The following changes since commit e04c78d86a9699d136910cfc0bdcf01087e3267e:
>
>   Linux 6.16-rc2 (2025-06-15 13:49:41 -0700)
>
> are available in the Git repository at:
>
>   git://git.kernel.org/pub/scm/linux/kernel/git/kvmarm/kvmarm.git tags/kvmarm-fixes-6.16-3
>
> for you to fetch changes up to 04c5355b2a94ff3191ce63ab035fb7f04d036869:
>
>   KVM: arm64: VHE: Centralize ISBs when returning to host (2025-06-19 13:34:59 +0100)
>
> ----------------------------------------------------------------
> KVM/arm64 fixes for 6.16, take #3
>
> - Fix another set of FP/SIMD/SVE bugs affecting NV, and plugging some
>   missing synchronisation
>
> - A small fix for the irqbypass hook fixes, tightening the check and
>   ensuring that we only deal with MSI for both the old and the new
>   route entry
>
> - Rework the way the shadow LRs are addressed in a nesting
>   configuration, plugging an embarrassing bug as well as simplifying
>   the whole process
>
> - Add yet another fix for the dreaded arch_timer_edge_cases selftest
>
> ----------------------------------------------------------------
> Marc Zyngier (1):
>       KVM: arm64: nv: Fix tracking of shadow list registers
>
> Mark Rutland (7):
>       KVM: arm64: VHE: Synchronize restore of host debug registers
>       KVM: arm64: VHE: Synchronize CPTR trap deactivation
>       KVM: arm64: Reorganise CPTR trap manipulation
>       KVM: arm64: Remove ad-hoc CPTR manipulation from fpsimd_sve_sync()
>       KVM: arm64: Remove ad-hoc CPTR manipulation from kvm_hyp_handle_fpsimd()
>       KVM: arm64: Remove cpacr_clear_set()
>       KVM: arm64: VHE: Centralize ISBs when returning to host
>
> Sean Christopherson (1):
>       KVM: arm64: Explicitly treat routing entry type changes as changes
>
> Zenghui Yu (1):
>       KVM: arm64: selftests: Close the GIC FD in arch_timer_edge_cases
>
>  arch/arm64/include/asm/kvm_emulate.h               |  62 ---------
>  arch/arm64/include/asm/kvm_host.h                  |   6 +-
>  arch/arm64/kvm/arm.c                               |   3 +-
>  arch/arm64/kvm/hyp/include/hyp/switch.h            | 147 +++++++++++++++++++--
>  arch/arm64/kvm/hyp/nvhe/hyp-main.c                 |   5 +-
>  arch/arm64/kvm/hyp/nvhe/switch.c                   |  59 ---------
>  arch/arm64/kvm/hyp/vhe/switch.c                    | 107 ++-------------
>  arch/arm64/kvm/vgic/vgic-v3-nested.c               |  81 ++++++------
>  .../selftests/kvm/arm64/arch_timer_edge_cases.c    |  16 ++-
>  9 files changed, 215 insertions(+), 271 deletions(-)
>