From mboxrd@z Thu Jan  1 00:00:00 1970
From: Marcelo Tosatti <mtosatti@redhat.com>
Date: Sat, 18 Oct 2008 15:27:45 +0000
Subject: Re: kvm/ia64: Fix halt emulation logic.
Message-Id: <20081018152745.GA32642@dmt.cnet>
List-Id: <kvm-ia64.vger.kernel.org>
References: <42DFA526FC41B1429CE7279EF83C6BDC01B47EA5@pdsmsx415.ccr.corp.intel.com>
In-Reply-To: <42DFA526FC41B1429CE7279EF83C6BDC01B47EA5@pdsmsx415.ccr.corp.intel.com>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
To: kvm-ia64@vger.kernel.org

Hi Xiantao,

On Wed, Oct 15, 2008 at 09:47:24PM +0800, Zhang, Xiantao wrote:
> +		expires = div64_u64(itc_diff, cyc_per_usec);
> +		kt = ktime_set(0, 1000 * expires);
> +
> +		down_read(&vcpu->kvm->slots_lock);
> +		vcpu->arch.ht_active = 1;
> +		hrtimer_start(p_ht, kt, HRTIMER_MODE_ABS);
>  
> -	if (irqchip_in_kernel(vcpu->kvm)) {
>  		vcpu->arch.mp_state = KVM_MP_STATE_HALTED;
>  		kvm_vcpu_block(vcpu);
>  		hrtimer_cancel(p_ht);
>  		vcpu->arch.ht_active = 0;
>  
> +		if (test_and_clear_bit(KVM_REQ_UNHALT, &vcpu->requests))
> +			if (vcpu->arch.mp_state = KVM_MP_STATE_HALTED)
> +				vcpu->arch.mp_state > +					KVM_MP_STATE_RUNNABLE;
> +		up_read(&vcpu->kvm->slots_lock);

You should release slots_lock when blocking via kvm_vcpu_block().
Otherwise paths that grab it for write will depend on these vcpus to
unhalt.

BTW, none of the data structures in this section of code should be
protected by slots_lock?