From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([208.118.235.92]:35365) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1TXoGE-0003Ja-A4 for qemu-devel@nongnu.org; Mon, 12 Nov 2012 02:16:49 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1TXoGB-0003Tv-7x for qemu-devel@nongnu.org; Mon, 12 Nov 2012 02:16:46 -0500 Received: from mx1.redhat.com ([209.132.183.28]:61094) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1TXoGA-0003Tf-WA for qemu-devel@nongnu.org; Mon, 12 Nov 2012 02:16:43 -0500 Date: Mon, 12 Nov 2012 09:19:18 +0200 From: "Michael S. Tsirkin" Message-ID: <20121112071918.GA31752@redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Subject: [Qemu-devel] [PATCH] tap: reset vnet header size on open List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: qemu-devel@nongnu.org, agraf@suse.de Cc: Anthony Liguori , Stefan Hajnoczi For tap, we currently assume the vnet header size is 10 (the default value) but that might not be the case if tap is persistent and has been used by qemu previously. To fix, set vnet header size correctly on open. Signed-off-by: Michael S. Tsirkin --- This fixes the issue reported by Alexander Graf on the kvm forum. Alexander, could you confirm please? Thanks, net/tap-linux.c | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/net/tap-linux.c b/net/tap-linux.c index c6521be..3eaedc4 100644 --- a/net/tap-linux.c +++ b/net/tap-linux.c @@ -39,6 +39,7 @@ int tap_open(char *ifname, int ifname_size, int *vnet_hdr, int vnet_hdr_required { struct ifreq ifr; int fd, ret; + int len = sizeof(struct virtio_net_hdr); TFR(fd = open(PATH_NET_TUN, O_RDWR)); if (fd < 0) { @@ -65,6 +66,13 @@ int tap_open(char *ifname, int ifname_size, int *vnet_hdr, int vnet_hdr_required close(fd); return -1; } + /* + * Make sure vnet header size has the default value: for a persistent + * tap it might have been modified e.g. by another instance of qemu. + * Ignore errors since old kernels do not support this ioctl: in this + * case the header size implicitly has the correct value. + */ + ioctl(fd, TUNSETVNETHDRSZ, &len); } if (ifname[0] != '\0') -- MST