From mboxrd@z Thu Jan 1 00:00:00 1970 From: Gleb Natapov Subject: Re: [Qemu-devel] Re: [RFC: 0/2] patch for QEMU HPET periodic timer emulation to alleviate time drift Date: Mon, 7 Feb 2011 15:54:00 +0200 Message-ID: <20110207135400.GG14984@redhat.com> References: <4D4B5F23.7040801@codemonkey.ws> <4D4BBF55.9060000@web.de> <4D4FE6BF.5080502@redhat.com> <4D4FEF81.1040603@codemonkey.ws> <4D4FF02F.2030309@redhat.com> <4D4FF24A.7000004@codemonkey.ws> <20110207134104.GE14984@redhat.com> <4D4FF7CE.6020005@redhat.com> <20110207134821.GF14984@redhat.com> <4D4FF8DF.2070400@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Anthony Liguori , Jan Kiszka , qemu-devel , Glauber Costa , Ulrich Obergfell , kvm To: Avi Kivity Return-path: Received: from mx1.redhat.com ([209.132.183.28]:12451 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752701Ab1BGNyF (ORCPT ); Mon, 7 Feb 2011 08:54:05 -0500 Content-Disposition: inline In-Reply-To: <4D4FF8DF.2070400@redhat.com> Sender: kvm-owner@vger.kernel.org List-ID: On Mon, Feb 07, 2011 at 03:51:27PM +0200, Avi Kivity wrote: > On 02/07/2011 03:48 PM, Gleb Natapov wrote: > >> > >> That's quite pointless, since those interrupts will be coalesced by > >> the guest. > >> > >Yes, of course, and this is what I remember happening. At this point > >interrupt de-coalescing kicks in. > > Maybe a more useful API would be to supply the callback with the > number of missed wakeups since the last callback. The callback > could then change the wakeup frequency to compensate, or do other > clever things. If it needs the old behaviour, it can just execute > its code in a loop. > Mat be. Or may be we can ask for new behaviour during timer creation. I was just surprised about the claim that we lost wakeups. -- Gleb.