Re: [PATCH] Fix kvmclock on !constant_tsc boxes.

[Marcelo Tosatti <mtosatti@redhat.com>]

Hi Gerd,

On Wed, Feb 04, 2009 at 05:52:04PM +0100, Gerd Hoffmann wrote:
> kvmclock currently falls apart on machines without constant tsc.
> This patch fixes it.  Changes:
> 
>   * keep tsc frequency in a per-cpu variable.
>   * handle kvmclock update using a new request flag, thus checking
>     whenever we need an update each time we enter guest context.
>   * use a cpufreq notifier to track frequency changes and force
>     kvmclock updates.
>   * send ipis to kick cpu out of guest context if needed to make
>     sure the guest doesn't see stale values.
> 
> Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
> ---
>  arch/x86/kvm/x86.c       |   99 +++++++++++++++++++++++++++++++++++++++++++---
>  include/linux/kvm_host.h |    1 +
>  2 files changed, 94 insertions(+), 6 deletions(-)
> 

> +	spin_lock(&kvm_lock);
> +	list_for_each_entry(kvm, &vm_list, vm_list) {
> +		for (i = 0; i < KVM_MAX_VCPUS; ++i) {
> +			vcpu = kvm->vcpus[i];
> +			if (!vcpu)
> +				continue;
> +			if (vcpu->cpu != freq->cpu)
> +				continue;
> +			if (!kvm_request_guest_time_update(vcpu))
> +				continue;
> +			if (vcpu->cpu != smp_processor_id())
> +				send_ipi++;
> +		}
> +	}
> +	spin_unlock(&kvm_lock);

The loop is cold hearted, but Juan's cleanups can be worked for 2.6.30
merge window. Given how many people have hit the bug better have
something non invasive until then.

> +
> +	if (freq->old < freq->new  &&  send_ipi) {

Can you fix the whitespace breakage please?

> +		smp_call_function_single(freq->cpu, bounce_off, NULL, 1);
> +	}
> +        return 0;
> +}