From mboxrd@z Thu Jan  1 00:00:00 1970
From: Raghavendra K T <raghukt@linux.vnet.ibm.com>
Subject: Re: [PATCH RFC V3 2/4] kvm hypervisor : Add a hypercall to KVM
	hypervisor to support pv-ticketlocks
Date: Wed, 07 Dec 2011 17:24:59 +0530
Message-ID: <4EDF5413.1030107@linux.vnet.ibm.com>
References: <20111130085921.23386.89708.sendpatchset@oc5400248562.ibm.com>
	<20111130085959.23386.69166.sendpatchset@oc5400248562.ibm.com>
	<20111207104849.GA24849@amt.cnet>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"; Format="flowed"
Content-Transfer-Encoding: 7bit
Return-path: <virtualization-bounces@lists.linux-foundation.org>
In-Reply-To: <20111207104849.GA24849@amt.cnet>
List-Unsubscribe: <https://lists.linuxfoundation.org/mailman/options/virtualization>,
	<mailto:virtualization-request@lists.linux-foundation.org?subject=unsubscribe>
List-Archive: <http://lists.linuxfoundation.org/pipermail/virtualization/>
List-Post: <mailto:virtualization@lists.linux-foundation.org>
List-Help: <mailto:virtualization-request@lists.linux-foundation.org?subject=help>
List-Subscribe: <https://lists.linuxfoundation.org/mailman/listinfo/virtualization>,
	<mailto:virtualization-request@lists.linux-foundation.org?subject=subscribe>
Sender: virtualization-bounces@lists.linux-foundation.org
Errors-To: virtualization-bounces@lists.linux-foundation.org
To: Marcelo Tosatti <mtosatti@redhat.com>
Cc: x86@kernel.org, Peter Zijlstra <peterz@infradead.org>, Virtualization <virtualization@lists.linux-foundation.org>, "H. Peter Anvin" <hpa@zytor.com>, Stefano Stabellini <stefano.stabellini@eu.citrix.com>, Xen <xen-devel@lists.xensource.com>, Dave Jiang <dave.jiang@intel.com>, KVM <kvm@vger.kernel.org>, Raghavendra K T <raghavendra.kt@linux.vnet.ibm.com>, Ingo Molnar <mingo@redhat.com>, Avi Kivity <avi@redhat.com>, Rik van Riel <riel@redhat.com>, Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>, Srivatsa Vaddagiri <vatsa@linux.vnet.ibm.com>, Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>, Sasha Levin <levinsasha928@gmail.com>, Sedat Dilek <sedat.dilek@gmail.com>, Thomas Gleixner <tglx@linutronix.de>, Yinghai Lu <yinghai@kernel.org>, Greg Kroah-Hartman <gregkh@suse.de>, LKML <linux-kernel@vger.kernel.org>, Dave Hansen <dave@linux.vnet.ibm.com>, Suzuki Poulose <suzuki>
List-Id: virtualization@lists.linuxfoundation.org

On 12/07/2011 04:18 PM, Marcelo Tosatti wrote:
> On Wed, Nov 30, 2011 at 02:29:59PM +0530, Raghavendra K T wrote:
>>
>> +/*
>> + * kvm_pv_kick_cpu_op:  Kick a vcpu.
>> + *
>> + * @cpu - vcpu to be kicked.
>> + */
>> +static void kvm_pv_kick_cpu_op(struct kvm *kvm, int cpu)
>> +{
>> +	struct kvm_vcpu *vcpu = kvm_get_vcpu(kvm, cpu);
>> +	struct kvm_mp_state mp_state;
>> +
>> +	mp_state.mp_state = KVM_MP_STATE_RUNNABLE;
>
> Since vcpu->mp_state is not protected by a lock, this is potentially racy. For example:
>
> CPU0						CPU1
> kvm_pv_kick_cpu_op				running vcpuN
> vcpuN->mp_state = KVM_MP_STATE_RUNNABLE;
> 						kvm_emulate_halt
> 						vcpuN->mp_state = KVM_MP_STATE_HALTED
>
> Is it harmless to lose a kick?
>

Yes you are right. It was potentially racy and it was harmful too!. I 
had observed that it was stalling the CPU before I introduced kicked flag.

But now,

vcpu->kicked = 1  ==> kvm_make_request(KVM_REQ_UNHALT, vcpu); ==>

__vcpu_run() ==> kvm_check_request(KVM_REQ_UNHALT, vcpu) ==>

vcpuN->mp_state = KVM_MP_STATE_RUNNABLE; so eventually we will end up
in RUNNABLE.

Also Avi pointed that, logically kvm_arch_vcpu_ioctl_set_mpstate should
be called only in vcpu thread, so after further debugging, I noticed
that,  setting vcpuN->mp_state = KVM_MP_STATE_RUNNABLE; is not
necessary.
I 'll remove that in the next patch. Thanks for pointing.