From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([208.118.235.92]:51834) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1TO2aJ-00039N-SH for qemu-devel@nongnu.org; Tue, 16 Oct 2012 04:33:11 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1TO2aD-0001Wq-Qx for qemu-devel@nongnu.org; Tue, 16 Oct 2012 04:33:07 -0400 Received: from mail-bk0-f45.google.com ([209.85.214.45]:57999) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1TO2aD-0001Wk-Ig for qemu-devel@nongnu.org; Tue, 16 Oct 2012 04:33:01 -0400 Received: by mail-bk0-f45.google.com with SMTP id jf3so2414777bkc.4 for ; Tue, 16 Oct 2012 01:33:00 -0700 (PDT) Date: Tue, 16 Oct 2012 10:32:55 +0200 From: Stefan Hajnoczi Message-ID: <20121016083255.GF23066@stefanha-thinkpad.redhat.com> References: <1350319733-7306-1-git-send-email-dmitry@daynix.com> <1350319733-7306-3-git-send-email-dmitry@daynix.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1350319733-7306-3-git-send-email-dmitry@daynix.com> Subject: Re: [Qemu-devel] [PATCH 2/2] Fix a race condition in E1000 device live migration. One of data-transfer related flags not in migrated fields list. List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Dmitry Fleytman Cc: Yan Vugenfirer , Chris Webb , qemu-devel@nongnu.org, Richard Davies On Mon, Oct 15, 2012 at 06:48:53PM +0200, Dmitry Fleytman wrote: The commit message is very long but the commit description is empty. Please keep the message short and add the rest into the description. > Signed-off-by: Dmitry Fleytman > --- > hw/e1000.c | 3 ++- > 1 file changed, 2 insertions(+), 1 deletion(-) > > diff --git a/hw/e1000.c b/hw/e1000.c > index 1e66ecf..efbe0c9 100644 > --- a/hw/e1000.c > +++ b/hw/e1000.c > @@ -92,7 +92,7 @@ typedef struct E1000State_st { > > uint32_t rxbuf_size; > uint32_t rxbuf_min_shift; > - int check_rxov; > + uint32_t check_rxov; > uint32_t rx_init_done; > struct e1000_tx { > unsigned char header[256]; > @@ -1120,6 +1120,7 @@ static const VMStateDescription vmstate_e1000 = { > VMSTATE_UNUSED(4), /* Was mmio_base. */ > VMSTATE_UINT32(rxbuf_size, E1000State), > VMSTATE_UINT32(rxbuf_min_shift, E1000State), > + VMSTATE_UINT32(check_rxov, E1000State), This breaks old -> new migration. Please see docs/migration.txt on VMSTATE and versions. It might also be useful to use git-blame(1) on some existing devices to see how people have modified the VMSTATE without breaking migration (this is something I don't know much about). Stefan