Re: [PATCH 4/5] KVM: arm64: vgic-v3: Don't require IRQs be disabled for LPI xarray lock

[Marc Zyngier <maz@kernel.org>]

On Thu, 04 Sep 2025 07:23:47 +0100,
Oliver Upton <oliver.upton@linux.dev> wrote:
> 
> Now that the LPI xarray is only read with IRQs disabled, relax the
> locking on the xarray to a non-IRQ disabling spinlock.
> 
> Signed-off-by: Oliver Upton <oliver.upton@linux.dev>
> ---
>  arch/arm64/kvm/vgic/vgic-init.c |  2 +-
>  arch/arm64/kvm/vgic/vgic-its.c  |  7 +++----
>  arch/arm64/kvm/vgic/vgic.c      | 13 ++++++-------
>  3 files changed, 10 insertions(+), 12 deletions(-)
> 
> diff --git a/arch/arm64/kvm/vgic/vgic-init.c b/arch/arm64/kvm/vgic/vgic-init.c
> index 4c777136ea5f..4c3c0d82e476 100644
> --- a/arch/arm64/kvm/vgic/vgic-init.c
> +++ b/arch/arm64/kvm/vgic/vgic-init.c
> @@ -53,7 +53,7 @@ void kvm_vgic_early_init(struct kvm *kvm)
>  {
>  	struct vgic_dist *dist = &kvm->arch.vgic;
>  
> -	xa_init_flags(&dist->lpi_xa, XA_FLAGS_LOCK_IRQ);
> +	xa_init(&dist->lpi_xa);
>  }
>  
>  /* CREATION */
> diff --git a/arch/arm64/kvm/vgic/vgic-its.c b/arch/arm64/kvm/vgic/vgic-its.c
> index f162206adb48..ce3e3ed3f29f 100644
> --- a/arch/arm64/kvm/vgic/vgic-its.c
> +++ b/arch/arm64/kvm/vgic/vgic-its.c
> @@ -78,7 +78,6 @@ static struct vgic_irq *vgic_add_lpi(struct kvm *kvm, u32 intid,
>  {
>  	struct vgic_dist *dist = &kvm->arch.vgic;
>  	struct vgic_irq *irq = vgic_get_irq(kvm, intid), *oldirq;
> -	unsigned long flags;
>  	int ret;
>  
>  	/* In this case there is no put, since we keep the reference. */
> @@ -89,7 +88,7 @@ static struct vgic_irq *vgic_add_lpi(struct kvm *kvm, u32 intid,
>  	if (!irq)
>  		return ERR_PTR(-ENOMEM);
>  
> -	ret = xa_reserve_irq(&dist->lpi_xa, intid, GFP_KERNEL_ACCOUNT);
> +	ret = xa_reserve(&dist->lpi_xa, intid, GFP_KERNEL_ACCOUNT);
>  	if (ret) {
>  		kfree(irq);
>  		return ERR_PTR(ret);
> @@ -104,7 +103,7 @@ static struct vgic_irq *vgic_add_lpi(struct kvm *kvm, u32 intid,
>  	irq->target_vcpu = vcpu;
>  	irq->group = 1;
>  
> -	xa_lock_irqsave(&dist->lpi_xa, flags);
> +	xa_lock(&dist->lpi_xa);
>  
>  	/*
>  	 * There could be a race with another vgic_add_lpi(), so we need to
> @@ -126,7 +125,7 @@ static struct vgic_irq *vgic_add_lpi(struct kvm *kvm, u32 intid,
>  	}
>  
>  out_unlock:
> -	xa_unlock_irqrestore(&dist->lpi_xa, flags);
> +	xa_unlock(&dist->lpi_xa);
>  
>  	if (ret)
>  		return ERR_PTR(ret);
> diff --git a/arch/arm64/kvm/vgic/vgic.c b/arch/arm64/kvm/vgic/vgic.c
> index 480391a0dcad..a21b482844ce 100644
> --- a/arch/arm64/kvm/vgic/vgic.c
> +++ b/arch/arm64/kvm/vgic/vgic.c
> @@ -28,7 +28,7 @@ struct vgic_global kvm_vgic_global_state __ro_after_init = {
>   *     kvm->arch.config_lock (mutex)
>   *       its->cmd_lock (mutex)
>   *         its->its_lock (mutex)
> - *           vgic_dist->lpi_xa.xa_lock		must be taken with IRQs disabled
> + *           vgic_dist->lpi_xa.xa_lock

Doesn't this contradict the commit message statement that we always
take this lock with interrupts disabled?

Thanks,

	M.

-- 
Jazz isn't dead. It just smells funny.