Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Message-ID: <9fdab135-648c-a378-9a64-f069cf6c0eba@redhat.com>
Date:   Wed, 20 Oct 2021 12:33:31 +0200
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:91.0) Gecko/20100101
 Thunderbird/91.1.0
Subject: Re: [PATCH v3 3/3] KVM: vCPU kick tax cut for running vCPU
Content-Language: en-US
To:     Sean Christopherson <seanjc@google.com>
Cc:     Wanpeng Li <kernellwp@gmail.com>, linux-kernel@vger.kernel.org,
        kvm@vger.kernel.org, Vitaly Kuznetsov <vkuznets@redhat.com>,
        Wanpeng Li <wanpengli@tencent.com>,
        Jim Mattson <jmattson@google.com>,
        Joerg Roedel <joro@8bytes.org>
References: <1634631160-67276-1-git-send-email-wanpengli@tencent.com>
 <1634631160-67276-3-git-send-email-wanpengli@tencent.com>
 <24e67e43-c50c-7e0f-305a-c7f6129f8d70@redhat.com>
 <YW8BmRJHVvFscWTo@google.com>
From:   Paolo Bonzini <pbonzini@redhat.com>
In-Reply-To: <YW8BmRJHVvFscWTo@google.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 19/10/21 19:34, Sean Christopherson wrote:
> The intent of the extra check was to avoid the locked instruction that comes with
> disabling preemption via rcu_read_lock().  But thinking more, the extra op should
> be little more than a basic arithmetic operation in the grand scheme on modern x86
> since the cache line is going to be locked and written no matter what, either
> immediately before or immediately after.

There should be no locked instructions unless you're using 
PREEMPT_RT/PREEMPT_RCU, no?  The preempt_disable count is in a percpu 
variable.

> 
> +       /*
> +        * Avoid the moderately expensive "should kick" operation if this pCPU
> +        * is currently running the target vCPU, in which case it's a KVM bug
> +        * if the vCPU is in the inner run loop.
> +        */
> +       if (vcpu == __this_cpu_read(kvm_running_vcpu) &&
> +           !WARN_ON_ONCE(vcpu->mode == IN_GUEST_MODE))
> +               goto out;
> +

It should not even be a problem if vcpu->mode == IN_GUEST_MODE, you just 
set it to EXITING_GUEST_MODE without even the need for atomic_cmpxchg.
I'll send a few patches out, since I think I found some related issues.

Paolo