[PATCH v2] KVM: arm64: timers: Convert per-vcpu virtual offset to a global value

[PATCH v2] KVM: arm64: timers: Convert per-vcpu virtual offset to a global value

[Marc Zyngier]

Having a per-vcpu virtual offset is a pain. It needs to be synchronized
on each update, and expands badly to a setup where different timers can
have different offsets, or have composite offsets (as with NV).

So let's start by replacing the use of the CNTVOFF_EL2 shadow register
(which we want to reclaim for NV anyway), and make the virtual timer
carry a pointer to a VM-wide offset.

This simplifies the code significantly. It also addresses two terrible bugs:

- The use of CNTVOFF_EL2 leads to some nice offset corruption
  when the sysreg gets reset, as reported by Joey.

- The kvm mutex is taken from a vcpu ioctl, which goes against
  the locking rules...

Reported-by: Joey Gouly <joey.gouly@arm.com>
Reviewed-by: Reiji Watanabe <reijiw@google.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20230224173915.GA17407@e124191.cambridge.arm.com
---
 arch/arm64/include/asm/kvm_host.h |  3 +++
 arch/arm64/kvm/arch_timer.c       | 45 +++++++------------------------
 arch/arm64/kvm/hypercalls.c       |  2 +-
 include/kvm/arm_arch_timer.h      | 15 +++++++++++
 4 files changed, 29 insertions(+), 36 deletions(-)

diff --git a/arch/arm64/include/asm/kvm_host.h b/arch/arm64/include/asm/kvm_host.h
index a1892a8f6032..bcd774d74f34 100644
--- a/arch/arm64/include/asm/kvm_host.h
+++ b/arch/arm64/include/asm/kvm_host.h
@@ -193,6 +193,9 @@ struct kvm_arch {
 	/* Interrupt controller */
 	struct vgic_dist	vgic;
 
+	/* Timers */
+	struct arch_timer_vm_data timer_data;
+
 	/* Mandated version of PSCI */
 	u32 psci_version;
 
diff --git a/arch/arm64/kvm/arch_timer.c b/arch/arm64/kvm/arch_timer.c
index 00610477ec7b..e1af4301b913 100644
--- a/arch/arm64/kvm/arch_timer.c
+++ b/arch/arm64/kvm/arch_timer.c
@@ -84,14 +84,10 @@ u64 timer_get_cval(struct arch_timer_context *ctxt)
 
 static u64 timer_get_offset(struct arch_timer_context *ctxt)
 {
-	struct kvm_vcpu *vcpu = ctxt->vcpu;
+	if (ctxt->offset.vm_offset)
+		return *ctxt->offset.vm_offset;
 
-	switch(arch_timer_ctx_index(ctxt)) {
-	case TIMER_VTIMER:
-		return __vcpu_sys_reg(vcpu, CNTVOFF_EL2);
-	default:
-		return 0;
-	}
+	return 0;
 }
 
 static void timer_set_ctl(struct arch_timer_context *ctxt, u32 ctl)
@@ -128,15 +124,12 @@ static void timer_set_cval(struct arch_timer_context *ctxt, u64 cval)
 
 static void timer_set_offset(struct arch_timer_context *ctxt, u64 offset)
 {
-	struct kvm_vcpu *vcpu = ctxt->vcpu;
-
-	switch(arch_timer_ctx_index(ctxt)) {
-	case TIMER_VTIMER:
-		__vcpu_sys_reg(vcpu, CNTVOFF_EL2) = offset;
-		break;
-	default:
+	if (!ctxt->offset.vm_offset) {
 		WARN(offset, "timer %ld\n", arch_timer_ctx_index(ctxt));
+		return;
 	}
+
+	WRITE_ONCE(*ctxt->offset.vm_offset, offset);
 }
 
 u64 kvm_phys_timer_read(void)
@@ -765,25 +758,6 @@ int kvm_timer_vcpu_reset(struct kvm_vcpu *vcpu)
 	return 0;
 }
 
-/* Make the updates of cntvoff for all vtimer contexts atomic */
-static void update_vtimer_cntvoff(struct kvm_vcpu *vcpu, u64 cntvoff)
-{
-	unsigned long i;
-	struct kvm *kvm = vcpu->kvm;
-	struct kvm_vcpu *tmp;
-
-	mutex_lock(&kvm->lock);
-	kvm_for_each_vcpu(i, tmp, kvm)
-		timer_set_offset(vcpu_vtimer(tmp), cntvoff);
-
-	/*
-	 * When called from the vcpu create path, the CPU being created is not
-	 * included in the loop above, so we just set it here as well.
-	 */
-	timer_set_offset(vcpu_vtimer(vcpu), cntvoff);
-	mutex_unlock(&kvm->lock);
-}
-
 void kvm_timer_vcpu_init(struct kvm_vcpu *vcpu)
 {
 	struct arch_timer_cpu *timer = vcpu_timer(vcpu);
@@ -791,10 +765,11 @@ void kvm_timer_vcpu_init(struct kvm_vcpu *vcpu)
 	struct arch_timer_context *ptimer = vcpu_ptimer(vcpu);
 
 	vtimer->vcpu = vcpu;
+	vtimer->offset.vm_offset = &vcpu->kvm->arch.timer_data.voffset;
 	ptimer->vcpu = vcpu;
 
 	/* Synchronize cntvoff across all vtimers of a VM. */
-	update_vtimer_cntvoff(vcpu, kvm_phys_timer_read());
+	timer_set_offset(vtimer, kvm_phys_timer_read());
 	timer_set_offset(ptimer, 0);
 
 	hrtimer_init(&timer->bg_timer, CLOCK_MONOTONIC, HRTIMER_MODE_ABS_HARD);
@@ -840,7 +815,7 @@ int kvm_arm_timer_set_reg(struct kvm_vcpu *vcpu, u64 regid, u64 value)
 		break;
 	case KVM_REG_ARM_TIMER_CNT:
 		timer = vcpu_vtimer(vcpu);
-		update_vtimer_cntvoff(vcpu, kvm_phys_timer_read() - value);
+		timer_set_offset(timer, kvm_phys_timer_read() - value);
 		break;
 	case KVM_REG_ARM_TIMER_CVAL:
 		timer = vcpu_vtimer(vcpu);
diff --git a/arch/arm64/kvm/hypercalls.c b/arch/arm64/kvm/hypercalls.c
index 64c086c02c60..5da884e11337 100644
--- a/arch/arm64/kvm/hypercalls.c
+++ b/arch/arm64/kvm/hypercalls.c
@@ -44,7 +44,7 @@ static void kvm_ptp_get_time(struct kvm_vcpu *vcpu, u64 *val)
 	feature = smccc_get_arg1(vcpu);
 	switch (feature) {
 	case KVM_PTP_VIRT_COUNTER:
-		cycles = systime_snapshot.cycles - vcpu_read_sys_reg(vcpu, CNTVOFF_EL2);
+		cycles = systime_snapshot.cycles - vcpu->kvm->arch.timer_data.voffset;
 		break;
 	case KVM_PTP_PHYS_COUNTER:
 		cycles = systime_snapshot.cycles;
diff --git a/include/kvm/arm_arch_timer.h b/include/kvm/arm_arch_timer.h
index 71916de7c6c4..c52a6e6839da 100644
--- a/include/kvm/arm_arch_timer.h
+++ b/include/kvm/arm_arch_timer.h
@@ -23,6 +23,19 @@ enum kvm_arch_timer_regs {
 	TIMER_REG_CTL,
 };
 
+struct arch_timer_offset {
+	/*
+	 * If set, pointer to one of the offsets in the kvm's offset
+	 * structure. If NULL, assume a zero offset.
+	 */
+	u64	*vm_offset;
+};
+
+struct arch_timer_vm_data {
+	/* Offset applied to the virtual timer/counter */
+	u64	voffset;
+};
+
 struct arch_timer_context {
 	struct kvm_vcpu			*vcpu;
 
@@ -32,6 +45,8 @@ struct arch_timer_context {
 	/* Emulated Timer (may be unused) */
 	struct hrtimer			hrtimer;
 
+	/* Offset for this counter/timer */
+	struct arch_timer_offset	offset;
 	/*
 	 * We have multiple paths which can save/restore the timer state onto
 	 * the hardware, so we need some way of keeping track of where the
-- 
2.34.1

Re: [PATCH v2] KVM: arm64: timers: Convert per-vcpu virtual offset to a global value

[Joey Gouly]

Hi Marc,

On Fri, Feb 24, 2023 at 07:16:40PM +0000, Marc Zyngier wrote:
> Having a per-vcpu virtual offset is a pain. It needs to be synchronized
> on each update, and expands badly to a setup where different timers can
> have different offsets, or have composite offsets (as with NV).
> 
> So let's start by replacing the use of the CNTVOFF_EL2 shadow register
> (which we want to reclaim for NV anyway), and make the virtual timer
> carry a pointer to a VM-wide offset.
> 
> This simplifies the code significantly. It also addresses two terrible bugs:
> 
> - The use of CNTVOFF_EL2 leads to some nice offset corruption
>   when the sysreg gets reset, as reported by Joey.
> 
> - The kvm mutex is taken from a vcpu ioctl, which goes against
>   the locking rules...
> 
> Reported-by: Joey Gouly <joey.gouly@arm.com>
> Reviewed-by: Reiji Watanabe <reijiw@google.com>
> Signed-off-by: Marc Zyngier <maz@kernel.org>
> Link: https://lore.kernel.org/r/20230224173915.GA17407@e124191.cambridge.arm.com

Fixes my mismatched timer offset issues.

Tested-by: Joey Gouly <joey.gouly@arm.com>

Thanks,
Joey

> ---
>  arch/arm64/include/asm/kvm_host.h |  3 +++
>  arch/arm64/kvm/arch_timer.c       | 45 +++++++------------------------
>  arch/arm64/kvm/hypercalls.c       |  2 +-
>  include/kvm/arm_arch_timer.h      | 15 +++++++++++
>  4 files changed, 29 insertions(+), 36 deletions(-)
> 
> diff --git a/arch/arm64/include/asm/kvm_host.h b/arch/arm64/include/asm/kvm_host.h
> index a1892a8f6032..bcd774d74f34 100644
> --- a/arch/arm64/include/asm/kvm_host.h
> +++ b/arch/arm64/include/asm/kvm_host.h
> @@ -193,6 +193,9 @@ struct kvm_arch {
>  	/* Interrupt controller */
>  	struct vgic_dist	vgic;
>  
> +	/* Timers */
> +	struct arch_timer_vm_data timer_data;
> +
>  	/* Mandated version of PSCI */
>  	u32 psci_version;
>  
> diff --git a/arch/arm64/kvm/arch_timer.c b/arch/arm64/kvm/arch_timer.c
> index 00610477ec7b..e1af4301b913 100644
> --- a/arch/arm64/kvm/arch_timer.c
> +++ b/arch/arm64/kvm/arch_timer.c
> @@ -84,14 +84,10 @@ u64 timer_get_cval(struct arch_timer_context *ctxt)
>  
>  static u64 timer_get_offset(struct arch_timer_context *ctxt)
>  {
> -	struct kvm_vcpu *vcpu = ctxt->vcpu;
> +	if (ctxt->offset.vm_offset)
> +		return *ctxt->offset.vm_offset;
>  
> -	switch(arch_timer_ctx_index(ctxt)) {
> -	case TIMER_VTIMER:
> -		return __vcpu_sys_reg(vcpu, CNTVOFF_EL2);
> -	default:
> -		return 0;
> -	}
> +	return 0;
>  }
>  
>  static void timer_set_ctl(struct arch_timer_context *ctxt, u32 ctl)
> @@ -128,15 +124,12 @@ static void timer_set_cval(struct arch_timer_context *ctxt, u64 cval)
>  
>  static void timer_set_offset(struct arch_timer_context *ctxt, u64 offset)
>  {
> -	struct kvm_vcpu *vcpu = ctxt->vcpu;
> -
> -	switch(arch_timer_ctx_index(ctxt)) {
> -	case TIMER_VTIMER:
> -		__vcpu_sys_reg(vcpu, CNTVOFF_EL2) = offset;
> -		break;
> -	default:
> +	if (!ctxt->offset.vm_offset) {
>  		WARN(offset, "timer %ld\n", arch_timer_ctx_index(ctxt));
> +		return;
>  	}
> +
> +	WRITE_ONCE(*ctxt->offset.vm_offset, offset);
>  }
>  
>  u64 kvm_phys_timer_read(void)
> @@ -765,25 +758,6 @@ int kvm_timer_vcpu_reset(struct kvm_vcpu *vcpu)
>  	return 0;
>  }
>  
> -/* Make the updates of cntvoff for all vtimer contexts atomic */
> -static void update_vtimer_cntvoff(struct kvm_vcpu *vcpu, u64 cntvoff)
> -{
> -	unsigned long i;
> -	struct kvm *kvm = vcpu->kvm;
> -	struct kvm_vcpu *tmp;
> -
> -	mutex_lock(&kvm->lock);
> -	kvm_for_each_vcpu(i, tmp, kvm)
> -		timer_set_offset(vcpu_vtimer(tmp), cntvoff);
> -
> -	/*
> -	 * When called from the vcpu create path, the CPU being created is not
> -	 * included in the loop above, so we just set it here as well.
> -	 */
> -	timer_set_offset(vcpu_vtimer(vcpu), cntvoff);
> -	mutex_unlock(&kvm->lock);
> -}
> -
>  void kvm_timer_vcpu_init(struct kvm_vcpu *vcpu)
>  {
>  	struct arch_timer_cpu *timer = vcpu_timer(vcpu);
> @@ -791,10 +765,11 @@ void kvm_timer_vcpu_init(struct kvm_vcpu *vcpu)
>  	struct arch_timer_context *ptimer = vcpu_ptimer(vcpu);
>  
>  	vtimer->vcpu = vcpu;
> +	vtimer->offset.vm_offset = &vcpu->kvm->arch.timer_data.voffset;
>  	ptimer->vcpu = vcpu;
>  
>  	/* Synchronize cntvoff across all vtimers of a VM. */
> -	update_vtimer_cntvoff(vcpu, kvm_phys_timer_read());
> +	timer_set_offset(vtimer, kvm_phys_timer_read());
>  	timer_set_offset(ptimer, 0);
>  
>  	hrtimer_init(&timer->bg_timer, CLOCK_MONOTONIC, HRTIMER_MODE_ABS_HARD);
> @@ -840,7 +815,7 @@ int kvm_arm_timer_set_reg(struct kvm_vcpu *vcpu, u64 regid, u64 value)
>  		break;
>  	case KVM_REG_ARM_TIMER_CNT:
>  		timer = vcpu_vtimer(vcpu);
> -		update_vtimer_cntvoff(vcpu, kvm_phys_timer_read() - value);
> +		timer_set_offset(timer, kvm_phys_timer_read() - value);
>  		break;
>  	case KVM_REG_ARM_TIMER_CVAL:
>  		timer = vcpu_vtimer(vcpu);
> diff --git a/arch/arm64/kvm/hypercalls.c b/arch/arm64/kvm/hypercalls.c
> index 64c086c02c60..5da884e11337 100644
> --- a/arch/arm64/kvm/hypercalls.c
> +++ b/arch/arm64/kvm/hypercalls.c
> @@ -44,7 +44,7 @@ static void kvm_ptp_get_time(struct kvm_vcpu *vcpu, u64 *val)
>  	feature = smccc_get_arg1(vcpu);
>  	switch (feature) {
>  	case KVM_PTP_VIRT_COUNTER:
> -		cycles = systime_snapshot.cycles - vcpu_read_sys_reg(vcpu, CNTVOFF_EL2);
> +		cycles = systime_snapshot.cycles - vcpu->kvm->arch.timer_data.voffset;
>  		break;
>  	case KVM_PTP_PHYS_COUNTER:
>  		cycles = systime_snapshot.cycles;
> diff --git a/include/kvm/arm_arch_timer.h b/include/kvm/arm_arch_timer.h
> index 71916de7c6c4..c52a6e6839da 100644
> --- a/include/kvm/arm_arch_timer.h
> +++ b/include/kvm/arm_arch_timer.h
> @@ -23,6 +23,19 @@ enum kvm_arch_timer_regs {
>  	TIMER_REG_CTL,
>  };
>  
> +struct arch_timer_offset {
> +	/*
> +	 * If set, pointer to one of the offsets in the kvm's offset
> +	 * structure. If NULL, assume a zero offset.
> +	 */
> +	u64	*vm_offset;
> +};
> +
> +struct arch_timer_vm_data {
> +	/* Offset applied to the virtual timer/counter */
> +	u64	voffset;
> +};
> +
>  struct arch_timer_context {
>  	struct kvm_vcpu			*vcpu;
>  
> @@ -32,6 +45,8 @@ struct arch_timer_context {
>  	/* Emulated Timer (may be unused) */
>  	struct hrtimer			hrtimer;
>  
> +	/* Offset for this counter/timer */
> +	struct arch_timer_offset	offset;
>  	/*
>  	 * We have multiple paths which can save/restore the timer state onto
>  	 * the hardware, so we need some way of keeping track of where the
> -- 
> 2.34.1

Re: [PATCH v2] KVM: arm64: timers: Convert per-vcpu virtual offset to a global value

[Marc Zyngier]

On Tue, 28 Feb 2023 11:26:07 +0000,
Joey Gouly <joey.gouly@arm.com> wrote:
> 
> Hi Marc,
> 
> On Fri, Feb 24, 2023 at 07:16:40PM +0000, Marc Zyngier wrote:
> > Having a per-vcpu virtual offset is a pain. It needs to be synchronized
> > on each update, and expands badly to a setup where different timers can
> > have different offsets, or have composite offsets (as with NV).
> > 
> > So let's start by replacing the use of the CNTVOFF_EL2 shadow register
> > (which we want to reclaim for NV anyway), and make the virtual timer
> > carry a pointer to a VM-wide offset.
> > 
> > This simplifies the code significantly. It also addresses two terrible bugs:
> > 
> > - The use of CNTVOFF_EL2 leads to some nice offset corruption
> >   when the sysreg gets reset, as reported by Joey.
> > 
> > - The kvm mutex is taken from a vcpu ioctl, which goes against
> >   the locking rules...
> > 
> > Reported-by: Joey Gouly <joey.gouly@arm.com>
> > Reviewed-by: Reiji Watanabe <reijiw@google.com>
> > Signed-off-by: Marc Zyngier <maz@kernel.org>
> > Link: https://lore.kernel.org/r/20230224173915.GA17407@e124191.cambridge.arm.com
> 
> Fixes my mismatched timer offset issues.
> 
> Tested-by: Joey Gouly <joey.gouly@arm.com>

Thanks for having given it a go. Hopefully Oliver will be able to send
this to as a fix shortly.

Thanks,

	M.

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

Re: [PATCH v2] KVM: arm64: timers: Convert per-vcpu virtual offset to a global value

[Oliver Upton]

On Tue, Feb 28, 2023 at 07:18:38PM +0000, Marc Zyngier wrote:
> On Tue, 28 Feb 2023 11:26:07 +0000,
> Joey Gouly <joey.gouly@arm.com> wrote:
> > 
> > Hi Marc,
> > 
> > On Fri, Feb 24, 2023 at 07:16:40PM +0000, Marc Zyngier wrote:
> > > Having a per-vcpu virtual offset is a pain. It needs to be synchronized
> > > on each update, and expands badly to a setup where different timers can
> > > have different offsets, or have composite offsets (as with NV).
> > > 
> > > So let's start by replacing the use of the CNTVOFF_EL2 shadow register
> > > (which we want to reclaim for NV anyway), and make the virtual timer
> > > carry a pointer to a VM-wide offset.
> > > 
> > > This simplifies the code significantly. It also addresses two terrible bugs:
> > > 
> > > - The use of CNTVOFF_EL2 leads to some nice offset corruption
> > >   when the sysreg gets reset, as reported by Joey.
> > > 
> > > - The kvm mutex is taken from a vcpu ioctl, which goes against
> > >   the locking rules...
> > > 
> > > Reported-by: Joey Gouly <joey.gouly@arm.com>
> > > Reviewed-by: Reiji Watanabe <reijiw@google.com>
> > > Signed-off-by: Marc Zyngier <maz@kernel.org>
> > > Link: https://lore.kernel.org/r/20230224173915.GA17407@e124191.cambridge.arm.com
> > 
> > Fixes my mismatched timer offset issues.
> > 
> > Tested-by: Joey Gouly <joey.gouly@arm.com>
> 
> Thanks for having given it a go. Hopefully Oliver will be able to send
> this to as a fix shortly.

Absolutely, I have this queued up. Thanks for testing the patch Joey.

Want to see if I can get push access to the kvmarm repository in time,
otherwise I'll send a pull to you Marc.

-- 
Thanks,
Oliver

Re: [PATCH v2] KVM: arm64: timers: Convert per-vcpu virtual offset to a global value

[Oliver Upton]

On Fri, 24 Feb 2023 19:16:40 +0000, Marc Zyngier wrote:
> Having a per-vcpu virtual offset is a pain. It needs to be synchronized
> on each update, and expands badly to a setup where different timers can
> have different offsets, or have composite offsets (as with NV).
> 
> So let's start by replacing the use of the CNTVOFF_EL2 shadow register
> (which we want to reclaim for NV anyway), and make the virtual timer
> carry a pointer to a VM-wide offset.
> 
> [...]

Applied to kvmarm/fixes, thanks!

[1/1] KVM: arm64: timers: Convert per-vcpu virtual offset to a global value
      https://git.kernel.org/kvmarm/kvmarm/c/47053904e182

--
Best,
Oliver