From mboxrd@z Thu Jan 1 00:00:00 1970 From: Avi Kivity Subject: Re: RHEL5.5, 32-bit VM repeatedly locks up due to kvmclock Date: Sat, 24 Apr 2010 21:40:42 +0300 Message-ID: <4BD33B2A.1070207@redhat.com> References: <4BD1D406.1040508@cisco.com> <201004231539.32205.iggy@theiggy.com> <4BD213AA.7070601@redhat.com> <4BD21459.9010903@cisco.com> <20100423222135.GB18563@c3sl.ufpr.br> <4BD22B67.6030800@cisco.com> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 7bit Cc: BRUNO CESAR RIBAS , Zachary Amsden , Brian Jackson , kvm-devel To: "David S. Ahern" Return-path: Received: from mx1.redhat.com ([209.132.183.28]:65392 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752662Ab0DXSlG (ORCPT ); Sat, 24 Apr 2010 14:41:06 -0400 In-Reply-To: <4BD22B67.6030800@cisco.com> Sender: kvm-owner@vger.kernel.org List-ID: On 04/24/2010 02:21 AM, David S. Ahern wrote: > > On 04/23/2010 04:21 PM, BRUNO CESAR RIBAS wrote: > >> Could you try hpet? I had similar problem with multicore and multiCPU (per >> mother board) [even with constant_tsc]. >> >> Since I changed the guest to hpet i had no more problems. >> > It's stable in the sense of no lockups yet, but is a much slower time > source from a gettimeofday perspective compared to tsc and jiffies > (based on speed jiffies appears to be tsc-based). > Jiffies doesn't sample any hardware; instead, a timer interrupt causes a counter to be incremented, and that counter is sampled. The downside is that clock granularity is very low - you can't use it for accurate timing. -- Do not meddle in the internals of kernels, for they are subtle and quick to panic.