Re: [PATCH] KVM: arm64: Ratelimit error log during guest debug exception

From: Marc Zyngier <maz@kernel.org>
To: Raghavendra Rao Ananta <rananta@google.com>
Cc: Alexandru Elisei <alexandru.elisei@arm.com>,
	Suzuki K Poulose <suzuki.poulose@arm.com>,
	Catalin Marinas <catalin.marinas@arm.com>,
	Will Deacon <will@kernel.org>, Peter Shier <pshier@google.com>,
	Ricardo Koller <ricarkol@google.com>,
	Oliver Upton <oupton@google.com>,
	Reiji Watanabe <reijiw@google.com>,
	Jing Zhang <jingzhangos@google.com>,
	linux-arm-kernel@lists.infradead.org,
	kvmarm@lists.cs.columbia.edu, linux-kernel@vger.kernel.org
Subject: Re: [PATCH] KVM: arm64: Ratelimit error log during guest debug exception
Date: Mon, 23 Aug 2021 22:51:48 +0100	[thread overview]
Message-ID: <1a0dfd2bc950cb84e7344973657c6b23@kernel.org> (raw)
In-Reply-To: <CAJHc60wLPFZ5XFwWVyex5GXr=qm7QWc2yOmkECxLh=L2QnvgWg@mail.gmail.com>

On 2021-08-23 19:13, Raghavendra Rao Ananta wrote:
> On Sat, Aug 21, 2021 at 3:56 AM Marc Zyngier <maz@kernel.org> wrote:
>> 
>> On Sat, 21 Aug 2021 00:01:24 +0100,
>> Raghavendra Rao Ananta <rananta@google.com> wrote:
>> >
>> > [1  <text/plain; UTF-8 (7bit)>]
>> > On Fri, Aug 20, 2021 at 2:29 AM Marc Zyngier <maz@kernel.org> wrote:
>> > >
>> > > On Thu, 19 Aug 2021 23:34:06 +0100,
>> > > Raghavendra Rao Ananta <rananta@google.com> wrote:
>> > > >
>> > > > Potentially, the guests could trigger a debug exception that's
>> > > > outside the exception class range.
>> > >
>> > > How? All the exception classes that lead to this functions are already
>> > > handled in the switch/case statement.
>> > >
>> > I guess I didn't think this through. Landing into kvm_handle_guest_debug()
>> > itself is not possible :)
>> 
>> Exactly.
>> 
>> > > My take on this is that this code isn't reachable, and that it could
>> > > be better rewritten as:
>> > >
>> > > diff --git a/arch/arm64/kvm/handle_exit.c b/arch/arm64/kvm/handle_exit.c
>> > > index 6f48336b1d86..ae7ec086827b 100644
>> > > --- a/arch/arm64/kvm/handle_exit.c
>> > > +++ b/arch/arm64/kvm/handle_exit.c
>> > > @@ -119,28 +119,14 @@ static int kvm_handle_guest_debug(struct kvm_vcpu
>> > *vcpu)
>> > >  {
>> > >         struct kvm_run *run = vcpu->run;
>> > >         u32 esr = kvm_vcpu_get_esr(vcpu);
>> > > -       int ret = 0;
>> > >
>> > >         run->exit_reason = KVM_EXIT_DEBUG;
>> > >         run->debug.arch.hsr = esr;
>> > >
>> > > -       switch (ESR_ELx_EC(esr)) {
>> > > -       case ESR_ELx_EC_WATCHPT_LOW:
>> > > +       if (ESR_ELx_EC(esr) ==  ESR_ELx_EC_WATCHPT_LOW)
>> > >                 run->debug.arch.far = vcpu->arch.fault.far_el2;
>> > > -               fallthrough;
>> > > -       case ESR_ELx_EC_SOFTSTP_LOW:
>> > > -       case ESR_ELx_EC_BREAKPT_LOW:
>> > > -       case ESR_ELx_EC_BKPT32:
>> > > -       case ESR_ELx_EC_BRK64:
>> > > -               break;
>> > > -       default:
>> > > -               kvm_err("%s: un-handled case esr: %#08x\n",
>> > > -                       __func__, (unsigned int) esr);
>> > > -               ret = -1;
>> > > -               break;
>> > > -       }
>> > >
>> > > -       return ret;
>> > > +       return 0;
>> > >  }
>> > >
>> > This looks better, but do you think we would be compromising on readability?
>> 
>> I don't think so. The exit handler table is, on its own, pretty
>> explicit about what we route to this handler, and the comment above
>> the function clearly states that we exit to userspace for all the
>> debug ECs.
> 
> Sounds great. I'm happy to send out a patch with you as 'Suggested-by' 
> , if you
> are okay with it.

Fire away!

         M.
-- 
Jazz is not dead. It just smells funny...