From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:55390) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1ZN2ki-00035u-NZ for qemu-devel@nongnu.org; Wed, 05 Aug 2015 13:45:21 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1ZN2kg-0007y9-2N for qemu-devel@nongnu.org; Wed, 05 Aug 2015 13:45:20 -0400 Received: from mail-wi0-x233.google.com ([2a00:1450:400c:c05::233]:36039) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1ZN2kf-0007xj-R4 for qemu-devel@nongnu.org; Wed, 05 Aug 2015 13:45:17 -0400 Received: by wicgj17 with SMTP id gj17so202358421wic.1 for ; Wed, 05 Aug 2015 10:45:17 -0700 (PDT) Sender: Paolo Bonzini References: From: Paolo Bonzini Message-ID: <55C24BAA.5090704@redhat.com> Date: Wed, 5 Aug 2015 19:45:14 +0200 MIME-Version: 1.0 In-Reply-To: Content-Type: text/plain; charset=windows-1252 Content-Transfer-Encoding: 7bit Subject: Re: [Qemu-devel] FW: about the bdrv_invalidate_cache_all() List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: "Li, Liang Z" , "qemu-devel@nongnu.org" , 'Kevin Wolf' , 'Juan Quintela' Cc: "Zhang, Yang Z" On 05/08/2015 17:21, Li, Liang Z wrote: > I found the function bdrv_invalidate_cache_all() in the > process_incoming_migration_co() takes more the 10ms when doing live > migration with shared storage, which prolong the service down time. > > The bdrv_invalidate_cache_all() is needed when doing the live > migration with '-b' option. For live migration with share storage, > such as nfs, is it still needed? If so, could please tell me why? If > not, I think we can do some optimization. IIUC it's the other way round. bdrv_invalidate_cache_all() is only needed for live migration with shared storage. Paolo