From mboxrd@z Thu Jan  1 00:00:00 1970
From: Radim =?utf-8?B?S3LEjW3DocWZ?= <rkrcmar@redhat.com>
Subject: Re: [PATCH 08/12] KVM: x86: Use the correct vcpu's TSC rate to
 compute time scale
Date: Mon, 5 Oct 2015 22:12:34 +0200
Message-ID: <20151005201233.GA31314@potion.brq.redhat.com>
References: <1443418691-24050-1-git-send-email-haozhong.zhang@intel.com>
 <1443418691-24050-9-git-send-email-haozhong.zhang@intel.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Cc: kvm@vger.kernel.org, Gleb Natapov <gleb@kernel.org>,
	Paolo Bonzini <pbonzini@redhat.com>,
	Thomas Gleixner <tglx@linutronix.de>,
	Ingo Molnar <mingo@redhat.com>,
	"H. Peter Anvin" <hpa@zytor.com>, x86@kernel.org,
	Joerg Roedel <joro@8bytes.org>,
	Wanpeng Li <wanpeng.li@linux.intel.com>,
	Xiao Guangrong <guangrong.xiao@linux.intel.com>,
	Mihai =?utf-8?B?RG9uyJt1?= <mdontu@bitdefender.com>,
	Andy Lutomirski <luto@kernel.org>,
	Kai Huang <kai.huang@linux.intel.com>,
	linux-kernel@vger.kernel.org
To: Haozhong Zhang <haozhong.zhang@intel.com>
Return-path: <kvm-owner@vger.kernel.org>
Received: from mx1.redhat.com ([209.132.183.28]:47561 "EHLO mx1.redhat.com"
	rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP
	id S1752631AbbJEUMj (ORCPT <rfc822;kvm@vger.kernel.org>);
	Mon, 5 Oct 2015 16:12:39 -0400
Content-Disposition: inline
In-Reply-To: <1443418691-24050-9-git-send-email-haozhong.zhang@intel.com>
Sender: kvm-owner@vger.kernel.org
List-ID: <kvm.vger.kernel.org>

2015-09-28 13:38+0800, Haozhong Zhang:
> This patch makes KVM use virtual_tsc_khz rather than the host TSC rate
> as vcpu's TSC rate to compute the time scale if TSC scaling is enabled.
> 
> Signed-off-by: Haozhong Zhang <haozhong.zhang@intel.com>
> ---
> diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
> @@ -1782,7 +1782,9 @@ static int kvm_guest_time_update(struct kvm_vcpu *v)
>  		return 0;
>  
>  	if (unlikely(vcpu->hw_tsc_khz != this_tsc_khz)) {
> -		kvm_get_time_scale(NSEC_PER_SEC / 1000, this_tsc_khz,
> +		tgt_tsc_khz = kvm_has_tsc_control ?
> +			vcpu->virtual_tsc_khz : this_tsc_khz;
> +		kvm_get_time_scale(NSEC_PER_SEC / 1000, tgt_tsc_khz,
>  				   &vcpu->hv_clock.tsc_shift,
>  				   &vcpu->hv_clock.tsc_to_system_mul);

Good catch, it seems that SVM didn't scale kvmclock correctly ...
I think we'll want this patch in stable.