Re: [PATCH] KVM: x86: fix KVM_SET_CLOCK relative to setting correct clock value

From: Paolo Bonzini <pbonzini@redhat.com>
To: Marcelo Tosatti <mtosatti@redhat.com>, kvm-devel <kvm@vger.kernel.org>
Cc: "Radim Krčmář" <rkrcmar@redhat.com>
Subject: Re: [PATCH] KVM: x86: fix KVM_SET_CLOCK relative to setting correct clock value
Date: Wed, 3 May 2017 15:08:46 +0200	[thread overview]
Message-ID: <2499ef65-1dfe-8460-ec41-661b05cc5023@redhat.com> (raw)
In-Reply-To: <20170502213616.GA24837@amt.cnet>

On 02/05/2017 23:36, Marcelo Tosatti wrote:
> In the masterclock enabled case, kvmclock_offset must be adjusted so
> that user_ns.clock = master_kernel_ns + kvmclock_offset (that is, the
> value set from KVM_SET_CLOCK is the one visible at system_timestamp).
> 
> This way the guest clock:
> 
> 	1. Starts counting when KVM_SET_CLOCK executes.
> 	2. With the value provided by userspace.

So this fixes rounding errors?

> -		now_ns = get_kvmclock_ns(kvm);
> -		kvm->arch.kvmclock_offset += user_ns.clock - now_ns;
> +
>  		kvm_gen_update_masterclock(kvm);
> +		if (ka->use_master_clock) {
> +			/*
> +			 * In the masterclock enabled case,
> +			 * kvmclock_offset must be adjusted so that
> +			 * user_ns.clock = master_kernel_ns + kvmclock_offset
> +			 * (that is, the value set from KVM_SET_CLOCK is the
> +			 * one visible at system_timestamp).
> +			 */
> +			kvm->arch.kvmclock_offset = user_ns.clock -
> +							ka->master_kernel_ns;
> +

This needs to hold ka->pvclock_gtod_sync_lock, I think.

> +			kvm_for_each_vcpu(i, vcpu, kvm)
> +				kvm_make_request(KVM_REQ_CLOCK_UPDATE, vcpu);

kvm_gen_update_masterclock already does that, why did you move that to
before the assignment?

Paolo