linux-kernel - Re: [PATCH v9 19/30] KVM: arm64: Provide assembly for SME register access

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <CA+EHjTwQ4fLBE1YXoB6M0eamSgGDW=nfLaC+-_surBfVbh3byQ@mail.gmail.com>
Date: Mon, 12 Jan 2026 17:59:17 +0000
From: Fuad Tabba <tabba@...gle.com>
To: Mark Brown <broonie@...nel.org>
Cc: Marc Zyngier <maz@...nel.org>, Joey Gouly <joey.gouly@....com>, 
	Catalin Marinas <catalin.marinas@....com>, Suzuki K Poulose <suzuki.poulose@....com>, 
	Will Deacon <will@...nel.org>, Paolo Bonzini <pbonzini@...hat.com>, Jonathan Corbet <corbet@....net>, 
	Shuah Khan <shuah@...nel.org>, Oliver Upton <oupton@...nel.org>, Dave Martin <Dave.Martin@....com>, 
	Mark Rutland <mark.rutland@....com>, Ben Horgan <ben.horgan@....com>, 
	linux-arm-kernel@...ts.infradead.org, kvmarm@...ts.linux.dev, 
	linux-kernel@...r.kernel.org, kvm@...r.kernel.org, linux-doc@...r.kernel.org, 
	linux-kselftest@...r.kernel.org, Peter Maydell <peter.maydell@...aro.org>, 
	Eric Auger <eric.auger@...hat.com>
Subject: Re: [PATCH v9 19/30] KVM: arm64: Provide assembly for SME register access

Hi Mark,

On Tue, 23 Dec 2025 at 01:22, Mark Brown <broonie@...nel.org> wrote:
>
> Provide versions of the SME state save and restore functions for the
> hypervisor to allow it to restore ZA and ZT for guests.
>
> Signed-off-by: Mark Brown <broonie@...nel.org>
> ---
>  arch/arm64/include/asm/kvm_hyp.h |  3 +++
>  arch/arm64/kvm/hyp/fpsimd.S      | 26 ++++++++++++++++++++++++++
>  2 files changed, 29 insertions(+)
>
> diff --git a/arch/arm64/include/asm/kvm_hyp.h b/arch/arm64/include/asm/kvm_hyp.h
> index 0317790dd3b7..1cef9991d238 100644
> --- a/arch/arm64/include/asm/kvm_hyp.h
> +++ b/arch/arm64/include/asm/kvm_hyp.h
> @@ -116,6 +116,9 @@ void __fpsimd_save_state(struct user_fpsimd_state *fp_regs);
>  void __fpsimd_restore_state(struct user_fpsimd_state *fp_regs);
>  void __sve_save_state(void *sve_pffr, u32 *fpsr, int save_ffr);
>  void __sve_restore_state(void *sve_pffr, u32 *fpsr, int restore_ffr);
> +int __sve_get_vl(void);
> +void __sme_save_state(void const *state, bool restore_zt);
> +void __sme_restore_state(void const *state, bool restore_zt);

Would it be a good idea to pass the VL to these functions. Currently,
they assume that the hardware's current VL matches the buffer's
intended layout. But if there is a mismatch between the guest's VL and
the current one, this could be difficult to debug. Passing the VL and
checking it against _sme_rdsvl would be an inexpensive way to avoid
these.

>
>  u64 __guest_enter(struct kvm_vcpu *vcpu);
>
> diff --git a/arch/arm64/kvm/hyp/fpsimd.S b/arch/arm64/kvm/hyp/fpsimd.S
> index 6e16cbfc5df2..44a1b0a483da 100644
> --- a/arch/arm64/kvm/hyp/fpsimd.S
> +++ b/arch/arm64/kvm/hyp/fpsimd.S
> @@ -29,3 +29,29 @@ SYM_FUNC_START(__sve_save_state)
>         sve_save 0, x1, x2, 3
>         ret
>  SYM_FUNC_END(__sve_save_state)
> +
> +SYM_FUNC_START(__sve_get_vl)
> +       _sve_rdvl       0, 1
> +       ret
> +SYM_FUNC_END(__sve_get_vl)

Since this is just one instruction, would it be better to implement it
as an inline assembly in the header file rather than a full
SYM_FUNC_START, to reduce the overhead?

> +
> +SYM_FUNC_START(__sme_save_state)

I think that this needs an isb(). We need to ensure that SMCR updates
are visible here. Looking ahead to where you introduce
__hyp_sme_save_guest(), that doesn't have a barrier after updating
SMCR. The alternative is to call the barrier where it's needed, but
make sure that this is well documented.

> +       _sme_rdsvl      2, 1            // x2 = VL/8
> +       sme_save_za 0, x2, 12           // Leaves x0 pointing to the end of ZA
> +
> +       cbz     x1, 1f
> +       _str_zt 0
> +1:
> +       ret
> +SYM_FUNC_END(__sme_save_state)
> +
> +SYM_FUNC_START(__sme_restore_state)

Same as above.

Cheers,
/fuad







> +       _sme_rdsvl      2, 1            // x2 = VL/8
> +       sme_load_za     0, x2, 12       // Leaves x0 pointing to end of ZA
> +
> +       cbz     x1, 1f
> +       _ldr_zt 0
> +
> +1:
> +       ret
> +SYM_FUNC_END(__sme_restore_state)
>
> --
> 2.47.3
>