Re: [PATCH v5 23/23] KVM: arm64: GICv4.1: Expose HW-based SGIs in debugfs

From: Auger Eric <eric.auger@redhat.com>
To: Marc Zyngier <maz@kernel.org>
Cc: linux-arm-kernel@lists.infradead.org,
	kvmarm@lists.cs.columbia.edu, kvm@vger.kernel.org,
	linux-kernel@vger.kernel.org,
	Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>,
	Jason Cooper <jason@lakedaemon.net>,
	Robert Richter <rrichter@marvell.com>,
	Thomas Gleixner <tglx@linutronix.de>,
	Zenghui Yu <yuzenghui@huawei.com>,
	James Morse <james.morse@arm.com>,
	Julien Thierry <julien.thierry.kdev@gmail.com>,
	Suzuki K Poulose <suzuki.poulose@arm.com>
Subject: Re: [PATCH v5 23/23] KVM: arm64: GICv4.1: Expose HW-based SGIs in debugfs
Date: Thu, 19 Mar 2020 16:43:09 +0100	[thread overview]
Message-ID: <33576d89-2b12-b98b-e392-3342b9b1265c@redhat.com> (raw)
In-Reply-To: <45c282bddd43420024633943c1befac3@kernel.org>

Hi Marc,

On 3/19/20 4:21 PM, Marc Zyngier wrote:
> Hi Eric,
> 
> On 2020-03-19 15:05, Auger Eric wrote:
>> Hi Marc,
>>
>> On 3/4/20 9:33 PM, Marc Zyngier wrote:
>>> The vgic-state debugfs file could do with showing the pending state
>>> of the HW-backed SGIs. Plug it into the low-level code.
>>>
>>> Signed-off-by: Marc Zyngier <maz@kernel.org>
>>> ---
>>>  virt/kvm/arm/vgic/vgic-debug.c | 14 +++++++++++++-
>>>  1 file changed, 13 insertions(+), 1 deletion(-)
>>>
>>> diff --git a/virt/kvm/arm/vgic/vgic-debug.c
>>> b/virt/kvm/arm/vgic/vgic-debug.c
>>> index cc12fe9b2df3..b13a9e3f99dd 100644
>>> --- a/virt/kvm/arm/vgic/vgic-debug.c
>>> +++ b/virt/kvm/arm/vgic/vgic-debug.c
>>> @@ -178,6 +178,8 @@ static void print_irq_state(struct seq_file *s,
>>> struct vgic_irq *irq,
>>>                  struct kvm_vcpu *vcpu)
>>>  {
>>>      char *type;
>>> +    bool pending;
>> nit: can be directly initialized to irq->pending_latch
>>> +
>>>      if (irq->intid < VGIC_NR_SGIS)
>>>          type = "SGI";
>>>      else if (irq->intid < VGIC_NR_PRIVATE_IRQS)
>>> @@ -190,6 +192,16 @@ static void print_irq_state(struct seq_file *s,
>>> struct vgic_irq *irq,
>>>      if (irq->intid ==0 || irq->intid == VGIC_NR_PRIVATE_IRQS)
>>>          print_header(s, irq, vcpu);
>>>
>>> +    pending = irq->pending_latch;
>>> +    if (irq->hw && vgic_irq_is_sgi(irq->intid)) {
>>> +        int err;
>>> +
>>> +        err = irq_get_irqchip_state(irq->host_irq,
>>> +                        IRQCHIP_STATE_PENDING,
>>> +                        &pending);
>>> +        WARN_ON_ONCE(err);
>>> +    }
>>> +
>>>      seq_printf(s, "       %s %4d "
>>>                "    %2d "
>>>                "%d%d%d%d%d%d%d "
>>> @@ -201,7 +213,7 @@ static void print_irq_state(struct seq_file *s,
>>> struct vgic_irq *irq,
>>>                "\n",
>>>              type, irq->intid,
>>>              (irq->target_vcpu) ? irq->target_vcpu->vcpu_id : -1,
>>> -            irq->pending_latch,
>>> +            pending,
>>>              irq->line_level,
>>>              irq->active,
>>>              irq->enabled,
>>>
>> The patch looks good to me but I am now lost about how we retrieve the
>> pending stat of other hw mapped interrupts. Looks we use
>> irq->pending_latch always. Is that correct?
> 
> Correct. GICv4.0 doesn't give us an architectural way to look at the
> vLPI pending state (there isn't even a guarantee about when the GIC
> will stop writing to memory, if it ever does).
> 
> With GICv4.1, you can introspect the HW state for SGIs. You can also
> look at the vLPI state by peeking at the virtual pending table, but
> you'd need to unmap the VPE first, which I obviously don't want to do
> for this debug interface, specially as it can be used whilst the guest
> is up and running.
OK for vLPIs, what about other HW mapped IRQs (arch timer?)

Eric
> 
> In the future, we'll have to implement that in order to support guest
> save/restore from a GICv4.1 system. I haven't given much thought to it
> though.
> 
>> For the patch:
>> Reviewed-by: Eric Auger <eric.auger@redhat.com>
> 
> Thanks,
> 
>         M.