[PATCH v2] kvm tools: Support multiple net devices

[PATCH v2] kvm tools: Support multiple net devices

[Sasha Levin]

This patch adds support for multiple network devices. The command line syntax
changes to the following:

	--network/-n [mode=[tap/user/none]] [guest_ip=[guest ip]] [host_ip=
[host_ip]] [guest_mac=[guest_mac]] [script=[script]]

Each of the parameters is optional, and the config defaults to a TAP based
networking with a sequential MAC.

Signed-off-by: Sasha Levin <levinsasha928@gmail.com>
---
 tools/kvm/builtin-run.c |  146 ++++++++++++++++++++++++++++++++------------
 tools/kvm/virtio/net.c  |  155 +++++++++++++++++++++++++---------------------
 2 files changed, 191 insertions(+), 110 deletions(-)

diff --git a/tools/kvm/builtin-run.c b/tools/kvm/builtin-run.c
index 28dc95a..070b1b6 100644
--- a/tools/kvm/builtin-run.c
+++ b/tools/kvm/builtin-run.c
@@ -87,9 +87,12 @@ static bool sdl;
 static bool balloon;
 static bool using_rootfs;
 static bool custom_rootfs;
+static bool no_net;
 extern bool ioport_debug;
 extern int  active_console;
 extern int  debug_iodelay;
+struct virtio_net_parameters *net_params;
+int num_net_devices;
 
 bool do_debug_print = false;
 
@@ -182,6 +185,89 @@ static int tty_parser(const struct option *opt, const char *arg, int unset)
 	return 0;
 }
 
+static inline void str_to_mac(const char *str, char *mac)
+{
+	sscanf(str, "%hhx:%hhx:%hhx:%hhx:%hhx:%hhx",
+		mac, mac+1, mac+2, mac+3, mac+4, mac+5);
+}
+static int set_net_param(struct virtio_net_parameters *p, const char *param,
+				const char *val)
+{
+	if (strcmp(param, "guest_mac") == 0) {
+		str_to_mac(val, p->guest_mac);
+	} else if (strcmp(param, "mode") == 0) {
+		if (!strncmp(val, "user", 4)) {
+			int i;
+
+			for (i = 0; i < num_net_devices; i++)
+				if (net_params[i].mode == NET_MODE_USER)
+					die("Only one usermode network device allowed at a time");
+			p->mode = NET_MODE_USER;
+		} else if (!strncmp(val, "tap", 3)) {
+			p->mode = NET_MODE_TAP;
+		} else if (!strncmp(val, "none", 4)) {
+			no_net = 1;
+			return -1;
+		} else
+			die("Unkown network mode %s, please use user, tap or none", network);
+	} else if (strcmp(param, "script") == 0) {
+		p->script = val;
+	} else if (strcmp(param, "guest_ip") == 0) {
+		p->guest_ip = val;
+	} else if (strcmp(param, "host_ip") == 0) {
+		p->host_ip = val;
+	}
+
+	return 0;
+}
+
+static int netdev_parser(const struct option *opt, const char *arg, int unset)
+{
+	struct virtio_net_parameters p;
+	char *buf, *cmd = NULL, *cur = NULL;
+	bool on_cmd = true;
+
+	if (arg) {
+		buf = strdup(arg);
+		if (buf == NULL)
+			die("Failed allocating new net buffer");
+		cur = strtok(buf, ",=");
+	}
+
+	p = (struct virtio_net_parameters) {
+		.guest_ip	= DEFAULT_GUEST_ADDR,
+		.host_ip	= DEFAULT_HOST_ADDR,
+		.script		= DEFAULT_SCRIPT,
+		.mode		= NET_MODE_TAP,
+	};
+
+	str_to_mac(DEFAULT_GUEST_MAC, p.guest_mac);
+	p.guest_mac[5] += num_net_devices;
+
+	while (cur) {
+		if (on_cmd) {
+			cmd = cur;
+		} else {
+			if (set_net_param(&p, cmd, cur) < 0)
+				goto done;
+		}
+		on_cmd = !on_cmd;
+
+		cur = strtok(NULL, ",=");
+	};
+
+	num_net_devices++;
+
+	net_params = realloc(net_params, num_net_devices * sizeof(*net_params));
+	if (net_params == NULL)
+		die("Failed adding new network device");
+
+	net_params[num_net_devices - 1] = p;
+
+done:
+	return 0;
+}
+
 static int shmem_parser(const struct option *opt, const char *arg, int unset)
 {
 	const u64 default_size = SHMEM_DEFAULT_SIZE;
@@ -339,18 +425,9 @@ static const struct option options[] = {
 			"Kernel command line arguments"),
 
 	OPT_GROUP("Networking options:"),
-	OPT_STRING('n', "network", &network, "user, tap, none",
-			"Network to use"),
-	OPT_STRING('\0', "host-ip", &host_ip, "a.b.c.d",
-			"Assign this address to the host side networking"),
-	OPT_STRING('\0', "guest-ip", &guest_ip, "a.b.c.d",
-			"Assign this address to the guest side networking"),
-	OPT_STRING('\0', "host-mac", &host_mac, "aa:bb:cc:dd:ee:ff",
-			"Assign this address to the host side NIC"),
-	OPT_STRING('\0', "guest-mac", &guest_mac, "aa:bb:cc:dd:ee:ff",
-			"Assign this address to the guest side NIC"),
-	OPT_STRING('\0', "tapscript", &script, "Script path",
-			 "Assign a script to process created tap device"),
+	OPT_CALLBACK_DEFAULT('n', "network", NULL, "network params",
+		     "Create a new guest NIC",
+		     netdev_parser, NULL),
 
 	OPT_GROUP("BIOS options:"),
 	OPT_INTEGER('\0', "vidmode", &vidmode,
@@ -615,7 +692,6 @@ void kvm_run_help(void)
 
 int kvm_cmd_run(int argc, const char **argv, const char *prefix)
 {
-	struct virtio_net_parameters net_params;
 	static char real_cmdline[2048], default_name[20];
 	struct framebuffer *fb = NULL;
 	unsigned int nr_online_cpus;
@@ -823,32 +899,24 @@ int kvm_cmd_run(int argc, const char **argv, const char *prefix)
 
 	virtio_9p__init(kvm);
 
-	if (strncmp(network, "none", 4)) {
-		net_params.guest_ip = guest_ip;
-		net_params.host_ip = host_ip;
-		net_params.kvm = kvm;
-		net_params.script = script;
-		sscanf(guest_mac, "%hhx:%hhx:%hhx:%hhx:%hhx:%hhx",
-			net_params.guest_mac,
-			net_params.guest_mac+1,
-			net_params.guest_mac+2,
-			net_params.guest_mac+3,
-			net_params.guest_mac+4,
-			net_params.guest_mac+5);
-		sscanf(host_mac, "%hhx:%hhx:%hhx:%hhx:%hhx:%hhx",
-			net_params.host_mac,
-			net_params.host_mac+1,
-			net_params.host_mac+2,
-			net_params.host_mac+3,
-			net_params.host_mac+4,
-			net_params.host_mac+5);
-
-		if (!strncmp(network, "user", 4))
-			net_params.mode = NET_MODE_USER;
-		else if (!strncmp(network, "tap", 3))
-			net_params.mode = NET_MODE_TAP;
-		else
-			die("Unkown network mode %s, please use -network user, tap, none", network);
+	for (i = 0; i < num_net_devices; i++) {
+		net_params[i].kvm = kvm;
+		virtio_net__init(&net_params[i]);
+	}
+
+	if (num_net_devices == 0 && no_net == 0) {
+		struct virtio_net_parameters net_params;
+
+		net_params = (struct virtio_net_parameters) {
+			.guest_ip	= guest_ip,
+			.host_ip	= host_ip,
+			.kvm		= kvm,
+			.script		= script,
+			.mode		= NET_MODE_USER,
+		};
+		str_to_mac(guest_mac, net_params.guest_mac);
+		str_to_mac(host_mac, net_params.host_mac);
+
 		virtio_net__init(&net_params);
 	}
 
diff --git a/tools/kvm/virtio/net.c b/tools/kvm/virtio/net.c
index 4700483..1e33891 100644
--- a/tools/kvm/virtio/net.c
+++ b/tools/kvm/virtio/net.c
@@ -43,6 +43,7 @@ struct net_dev_operations {
 struct net_dev {
 	pthread_mutex_t			mutex;
 	struct virtio_pci		vpci;
+	struct list_head		list;
 
 	struct virt_queue		vqs[VIRTIO_NET_NUM_QUEUES];
 	struct virtio_net_config	config;
@@ -64,48 +65,41 @@ struct net_dev {
 
 	struct uip_info			info;
 	struct net_dev_operations	*ops;
+	struct kvm			*kvm;
 };
 
-static struct net_dev ndev = {
-	.mutex	= PTHREAD_MUTEX_INITIALIZER,
-
-	.config = {
-		.status			= VIRTIO_NET_S_LINK_UP,
-	},
-	.info = {
-		.buf_nr			= 20,
-	}
-};
+static LIST_HEAD(ndevs);
 
 static void *virtio_net_rx_thread(void *p)
 {
 	struct iovec iov[VIRTIO_NET_QUEUE_SIZE];
 	struct virt_queue *vq;
 	struct kvm *kvm;
+	struct net_dev *ndev = p;
 	u16 out, in;
 	u16 head;
 	int len;
 
-	kvm	= p;
-	vq	= &ndev.vqs[VIRTIO_NET_RX_QUEUE];
+	kvm	= ndev->kvm;
+	vq	= &ndev->vqs[VIRTIO_NET_RX_QUEUE];
 
 	while (1) {
 
-		mutex_lock(&ndev.io_rx_lock);
+		mutex_lock(&ndev->io_rx_lock);
 		if (!virt_queue__available(vq))
-			pthread_cond_wait(&ndev.io_rx_cond, &ndev.io_rx_lock);
-		mutex_unlock(&ndev.io_rx_lock);
+			pthread_cond_wait(&ndev->io_rx_cond, &ndev->io_rx_lock);
+		mutex_unlock(&ndev->io_rx_lock);
 
 		while (virt_queue__available(vq)) {
 
 			head = virt_queue__get_iov(vq, iov, &out, &in, kvm);
 
-			len = ndev.ops->rx(iov, in, &ndev);
+			len = ndev->ops->rx(iov, in, ndev);
 
 			virt_queue__set_used_elem(vq, head, len);
 
 			/* We should interrupt guest right now, otherwise latency is huge. */
-			virtio_pci__signal_vq(kvm, &ndev.vpci, VIRTIO_NET_RX_QUEUE);
+			virtio_pci__signal_vq(kvm, &ndev->vpci, VIRTIO_NET_RX_QUEUE);
 		}
 
 	}
@@ -120,29 +114,30 @@ static void *virtio_net_tx_thread(void *p)
 	struct iovec iov[VIRTIO_NET_QUEUE_SIZE];
 	struct virt_queue *vq;
 	struct kvm *kvm;
+	struct net_dev *ndev = p;
 	u16 out, in;
 	u16 head;
 	int len;
 
-	kvm	= p;
-	vq	= &ndev.vqs[VIRTIO_NET_TX_QUEUE];
+	kvm	= ndev->kvm;
+	vq	= &ndev->vqs[VIRTIO_NET_TX_QUEUE];
 
 	while (1) {
-		mutex_lock(&ndev.io_tx_lock);
+		mutex_lock(&ndev->io_tx_lock);
 		if (!virt_queue__available(vq))
-			pthread_cond_wait(&ndev.io_tx_cond, &ndev.io_tx_lock);
-		mutex_unlock(&ndev.io_tx_lock);
+			pthread_cond_wait(&ndev->io_tx_cond, &ndev->io_tx_lock);
+		mutex_unlock(&ndev->io_tx_lock);
 
 		while (virt_queue__available(vq)) {
 
 			head = virt_queue__get_iov(vq, iov, &out, &in, kvm);
 
-			len = ndev.ops->tx(iov, out, &ndev);
+			len = ndev->ops->tx(iov, out, ndev);
 
 			virt_queue__set_used_elem(vq, head, len);
 		}
 
-		virtio_pci__signal_vq(kvm, &ndev.vpci, VIRTIO_NET_TX_QUEUE);
+		virtio_pci__signal_vq(kvm, &ndev->vpci, VIRTIO_NET_TX_QUEUE);
 	}
 
 	pthread_exit(NULL);
@@ -151,58 +146,58 @@ static void *virtio_net_tx_thread(void *p)
 
 }
 
-static void virtio_net_handle_callback(struct kvm *kvm, u16 queue_index)
+static void virtio_net_handle_callback(struct kvm *kvm, struct net_dev *ndev, int queue)
 {
-	switch (queue_index) {
+	switch (queue) {
 	case VIRTIO_NET_TX_QUEUE:
-		mutex_lock(&ndev.io_tx_lock);
-		pthread_cond_signal(&ndev.io_tx_cond);
-		mutex_unlock(&ndev.io_tx_lock);
+		mutex_lock(&ndev->io_tx_lock);
+		pthread_cond_signal(&ndev->io_tx_cond);
+		mutex_unlock(&ndev->io_tx_lock);
 		break;
 	case VIRTIO_NET_RX_QUEUE:
-		mutex_lock(&ndev.io_rx_lock);
-		pthread_cond_signal(&ndev.io_rx_cond);
-		mutex_unlock(&ndev.io_rx_lock);
+		mutex_lock(&ndev->io_rx_lock);
+		pthread_cond_signal(&ndev->io_rx_cond);
+		mutex_unlock(&ndev->io_rx_lock);
 		break;
 	default:
-		pr_warning("Unknown queue index %u", queue_index);
+		pr_warning("Unknown queue index %u", queue);
 	}
 }
 
-static bool virtio_net__tap_init(const struct virtio_net_parameters *params)
+static bool virtio_net__tap_init(const struct virtio_net_parameters *params,
+					struct net_dev *ndev)
 {
 	int sock = socket(AF_INET, SOCK_STREAM, 0);
 	int pid, status, offload, hdr_len;
 	struct sockaddr_in sin = {0};
 	struct ifreq ifr;
 
-	ndev.tap_fd = open("/dev/net/tun", O_RDWR);
-	if (ndev.tap_fd < 0) {
+	ndev->tap_fd = open("/dev/net/tun", O_RDWR);
+	if (ndev->tap_fd < 0) {
 		pr_warning("Unable to open /dev/net/tun");
 		goto fail;
 	}
 
 	memset(&ifr, 0, sizeof(ifr));
 	ifr.ifr_flags = IFF_TAP | IFF_NO_PI | IFF_VNET_HDR;
-	if (ioctl(ndev.tap_fd, TUNSETIFF, &ifr) < 0) {
+	if (ioctl(ndev->tap_fd, TUNSETIFF, &ifr) < 0) {
 		pr_warning("Config tap device error. Are you root?");
 		goto fail;
 	}
 
-	strncpy(ndev.tap_name, ifr.ifr_name, sizeof(ndev.tap_name));
+	strncpy(ndev->tap_name, ifr.ifr_name, sizeof(ndev->tap_name));
 
-	if (ioctl(ndev.tap_fd, TUNSETNOCSUM, 1) < 0) {
+	if (ioctl(ndev->tap_fd, TUNSETNOCSUM, 1) < 0) {
 		pr_warning("Config tap device TUNSETNOCSUM error");
 		goto fail;
 	}
 
 	hdr_len = sizeof(struct virtio_net_hdr);
-	if (ioctl(ndev.tap_fd, TUNSETVNETHDRSZ, &hdr_len) < 0) {
+	if (ioctl(ndev->tap_fd, TUNSETVNETHDRSZ, &hdr_len) < 0)
 		pr_warning("Config tap device TUNSETVNETHDRSZ error");
-	}
 
 	offload = TUN_F_CSUM | TUN_F_TSO4 | TUN_F_TSO6 | TUN_F_UFO;
-	if (ioctl(ndev.tap_fd, TUNSETOFFLOAD, offload) < 0) {
+	if (ioctl(ndev->tap_fd, TUNSETOFFLOAD, offload) < 0) {
 		pr_warning("Config tap device TUNSETOFFLOAD error");
 		goto fail;
 	}
@@ -210,7 +205,7 @@ static bool virtio_net__tap_init(const struct virtio_net_parameters *params)
 	if (strcmp(params->script, "none")) {
 		pid = fork();
 		if (pid == 0) {
-			execl(params->script, params->script, ndev.tap_name, NULL);
+			execl(params->script, params->script, ndev->tap_name, NULL);
 			_exit(1);
 		} else {
 			waitpid(pid, &status, 0);
@@ -221,7 +216,7 @@ static bool virtio_net__tap_init(const struct virtio_net_parameters *params)
 		}
 	} else {
 		memset(&ifr, 0, sizeof(ifr));
-		strncpy(ifr.ifr_name, ndev.tap_name, sizeof(ndev.tap_name));
+		strncpy(ifr.ifr_name, ndev->tap_name, sizeof(ndev->tap_name));
 		sin.sin_addr.s_addr = inet_addr(params->host_ip);
 		memcpy(&(ifr.ifr_addr), &sin, sizeof(ifr.ifr_addr));
 		ifr.ifr_addr.sa_family = AF_INET;
@@ -232,7 +227,7 @@ static bool virtio_net__tap_init(const struct virtio_net_parameters *params)
 	}
 
 	memset(&ifr, 0, sizeof(ifr));
-	strncpy(ifr.ifr_name, ndev.tap_name, sizeof(ndev.tap_name));
+	strncpy(ifr.ifr_name, ndev->tap_name, sizeof(ndev->tap_name));
 	ioctl(sock, SIOCGIFFLAGS, &ifr);
 	ifr.ifr_flags |= IFF_UP | IFF_RUNNING;
 	if (ioctl(sock, SIOCSIFFLAGS, &ifr) < 0)
@@ -245,22 +240,22 @@ static bool virtio_net__tap_init(const struct virtio_net_parameters *params)
 fail:
 	if (sock >= 0)
 		close(sock);
-	if (ndev.tap_fd >= 0)
-		close(ndev.tap_fd);
+	if (ndev->tap_fd >= 0)
+		close(ndev->tap_fd);
 
 	return 0;
 }
 
-static void virtio_net__io_thread_init(struct kvm *kvm)
+static void virtio_net__io_thread_init(struct kvm *kvm, struct net_dev *ndev)
 {
-	pthread_mutex_init(&ndev.io_rx_lock, NULL);
-	pthread_cond_init(&ndev.io_tx_cond, NULL);
+	pthread_mutex_init(&ndev->io_rx_lock, NULL);
+	pthread_cond_init(&ndev->io_tx_cond, NULL);
 
-	pthread_mutex_init(&ndev.io_rx_lock, NULL);
-	pthread_cond_init(&ndev.io_tx_cond, NULL);
+	pthread_mutex_init(&ndev->io_rx_lock, NULL);
+	pthread_cond_init(&ndev->io_tx_cond, NULL);
 
-	pthread_create(&ndev.io_rx_thread, NULL, virtio_net_rx_thread, (void *)kvm);
-	pthread_create(&ndev.io_tx_thread, NULL, virtio_net_tx_thread, (void *)kvm);
+	pthread_create(&ndev->io_rx_thread, NULL, virtio_net_rx_thread, ndev);
+	pthread_create(&ndev->io_tx_thread, NULL, virtio_net_tx_thread, ndev);
 }
 
 static inline int tap_ops_tx(struct iovec *iov, u16 out, struct net_dev *ndev)
@@ -345,7 +340,9 @@ static int init_vq(struct kvm *kvm, void *dev, u32 vq, u32 pfn)
 
 static int notify_vq(struct kvm *kvm, void *dev, u32 vq)
 {
-	virtio_net_handle_callback(kvm, vq);
+	struct net_dev *ndev = dev;
+
+	virtio_net_handle_callback(kvm, ndev, vq);
 
 	return 0;
 }
@@ -365,27 +362,43 @@ static int get_size_vq(struct kvm *kvm, void *dev, u32 vq)
 void virtio_net__init(const struct virtio_net_parameters *params)
 {
 	int i;
+	struct net_dev *ndev;
+
+	if (!params)
+		return;
+
+	ndev = calloc(1, sizeof(struct net_dev));
+	if (ndev == NULL)
+		die("Failed allocating ndev");
+
+	list_add_tail(&ndev->list, &ndevs);
+
+	ndev->kvm = params->kvm;
+
+	mutex_init(&ndev->mutex);
+	ndev->config.status = VIRTIO_NET_S_LINK_UP;
 
 	for (i = 0 ; i < 6 ; i++) {
-		ndev.config.mac[i]		= params->guest_mac[i];
-		ndev.info.guest_mac.addr[i]	= params->guest_mac[i];
-		ndev.info.host_mac.addr[i]	= params->host_mac[i];
+		ndev->config.mac[i]		= params->guest_mac[i];
+		ndev->info.guest_mac.addr[i]	= params->guest_mac[i];
+		ndev->info.host_mac.addr[i]	= params->host_mac[i];
 	}
 
-	ndev.mode = params->mode;
-	if (ndev.mode == NET_MODE_TAP) {
-		virtio_net__tap_init(params);
-		ndev.ops = &tap_ops;
+	ndev->mode = params->mode;
+	if (ndev->mode == NET_MODE_TAP) {
+		virtio_net__tap_init(params, ndev);
+		ndev->ops = &tap_ops;
 	} else {
-		ndev.info.host_ip		= ntohl(inet_addr(params->host_ip));
-		ndev.info.guest_ip		= ntohl(inet_addr(params->guest_ip));
-		ndev.info.guest_netmask		= ntohl(inet_addr("255.255.255.0"));
-		uip_init(&ndev.info);
-		ndev.ops = &uip_ops;
+		ndev->info.host_ip		= ntohl(inet_addr(params->host_ip));
+		ndev->info.guest_ip		= ntohl(inet_addr(params->guest_ip));
+		ndev->info.guest_netmask	= ntohl(inet_addr("255.255.255.0"));
+		ndev->info.buf_nr		= 20,
+		uip_init(&ndev->info);
+		ndev->ops = &uip_ops;
 	}
 
-	virtio_pci__init(kvm, &ndev.vpci, &ndev, PCI_DEVICE_ID_VIRTIO_NET, VIRTIO_ID_NET);
-	ndev.vpci.ops = (struct virtio_pci_ops) {
+	virtio_pci__init(kvm, &ndev->vpci, ndev, PCI_DEVICE_ID_VIRTIO_NET, VIRTIO_ID_NET);
+	ndev->vpci.ops = (struct virtio_pci_ops) {
 		.set_config		= set_config,
 		.get_config		= get_config,
 		.get_host_features	= get_host_features,
@@ -396,9 +409,9 @@ void virtio_net__init(const struct virtio_net_parameters *params)
 		.get_size_vq		= get_size_vq,
 	};
 
-	virtio_net__io_thread_init(params->kvm);
+	virtio_net__io_thread_init(params->kvm, ndev);
 
-	ndev.compat_id = compat__add_message("virtio-net device was not detected",
+	ndev->compat_id = compat__add_message("virtio-net device was not detected",
 						"While you have requested a virtio-net device, "
 						"the guest kernel didn't seem to detect it.\n"
 						"Please make sure that the kernel was compiled "
-- 
1.7.6.1

Re: [PATCH v2] kvm tools: Support multiple net devices

[Pekka Enberg]

On Sun, Sep 25, 2011 at 2:11 PM, Sasha Levin <levinsasha928@gmail.com> wrote:
> This patch adds support for multiple network devices. The command line syntax
> changes to the following:
>
>        --network/-n [mode=[tap/user/none]] [guest_ip=[guest ip]] [host_ip=
> [host_ip]] [guest_mac=[guest_mac]] [script=[script]]
>
> Each of the parameters is optional, and the config defaults to a TAP based
> networking with a sequential MAC.
>
> Signed-off-by: Sasha Levin <levinsasha928@gmail.com>

Seems reasonable. Asias, is v2 OK to merge?

Re: [PATCH v2] kvm tools: Support multiple net devices

[Asias He]

On 09/25/2011 07:11 PM, Sasha Levin wrote:
> This patch adds support for multiple network devices. The command line syntax
> changes to the following:
> 
> 	--network/-n [mode=[tap/user/none]] [guest_ip=[guest ip]] [host_ip=
> [host_ip]] [guest_mac=[guest_mac]] [script=[script]]

This syntax is actually, no?

   --network/-n mode=tap,guest_ip=x.x.x.x

not

   --network/-n mode=tag guest_ip=x.x.x.x


This works for me:
$ sudo ./kvm run -d sid.img -p root=/dev/vda1 -k bzImage  \
-n mode=tap,host_ip=192.168.100.1 -n mode=tap,host_ip=192.168.200.1

This does not work for me:
$ sudo ./kvm run -d sid.img -p root=/dev/vda1 -k bzImage \
-n mode=tap host_ip=192.168.100.1 -n mode=tap host_ip=192.168.200.1


> 
> Each of the parameters is optional, and the config defaults to a TAP based
> networking with a sequential MAC.
> 
> Signed-off-by: Sasha Levin <levinsasha928@gmail.com>
> ---
>  tools/kvm/builtin-run.c |  146 ++++++++++++++++++++++++++++++++------------
>  tools/kvm/virtio/net.c  |  155 +++++++++++++++++++++++++---------------------
>  2 files changed, 191 insertions(+), 110 deletions(-)
> 
> diff --git a/tools/kvm/builtin-run.c b/tools/kvm/builtin-run.c
> index 28dc95a..070b1b6 100644
> --- a/tools/kvm/builtin-run.c
> +++ b/tools/kvm/builtin-run.c
> @@ -87,9 +87,12 @@ static bool sdl;
>  static bool balloon;
>  static bool using_rootfs;
>  static bool custom_rootfs;
> +static bool no_net;
>  extern bool ioport_debug;
>  extern int  active_console;
>  extern int  debug_iodelay;
> +struct virtio_net_parameters *net_params;
> +int num_net_devices;
>  
>  bool do_debug_print = false;
>  
> @@ -182,6 +185,89 @@ static int tty_parser(const struct option *opt, const char *arg, int unset)
>  	return 0;
>  }
>  
> +static inline void str_to_mac(const char *str, char *mac)
> +{
> +	sscanf(str, "%hhx:%hhx:%hhx:%hhx:%hhx:%hhx",
> +		mac, mac+1, mac+2, mac+3, mac+4, mac+5);
> +}
> +static int set_net_param(struct virtio_net_parameters *p, const char *param,
> +				const char *val)
> +{
> +	if (strcmp(param, "guest_mac") == 0) {
> +		str_to_mac(val, p->guest_mac);
> +	} else if (strcmp(param, "mode") == 0) {
> +		if (!strncmp(val, "user", 4)) {
> +			int i;
> +
> +			for (i = 0; i < num_net_devices; i++)
> +				if (net_params[i].mode == NET_MODE_USER)
> +					die("Only one usermode network device allowed at a time");
> +			p->mode = NET_MODE_USER;
> +		} else if (!strncmp(val, "tap", 3)) {
> +			p->mode = NET_MODE_TAP;
> +		} else if (!strncmp(val, "none", 4)) {
> +			no_net = 1;
> +			return -1;
> +		} else
> +			die("Unkown network mode %s, please use user, tap or none", network);
> +	} else if (strcmp(param, "script") == 0) {
> +		p->script = val;
> +	} else if (strcmp(param, "guest_ip") == 0) {
> +		p->guest_ip = val;
> +	} else if (strcmp(param, "host_ip") == 0) {
> +		p->host_ip = val;
> +	}
> +
> +	return 0;
> +}
> +
> +static int netdev_parser(const struct option *opt, const char *arg, int unset)
> +{
> +	struct virtio_net_parameters p;
> +	char *buf, *cmd = NULL, *cur = NULL;
> +	bool on_cmd = true;
> +
> +	if (arg) {
> +		buf = strdup(arg);
> +		if (buf == NULL)
> +			die("Failed allocating new net buffer");
> +		cur = strtok(buf, ",=");
> +	}
> +
> +	p = (struct virtio_net_parameters) {
> +		.guest_ip	= DEFAULT_GUEST_ADDR,
> +		.host_ip	= DEFAULT_HOST_ADDR,
> +		.script		= DEFAULT_SCRIPT,
> +		.mode		= NET_MODE_TAP,
> +	};
> +
> +	str_to_mac(DEFAULT_GUEST_MAC, p.guest_mac);
> +	p.guest_mac[5] += num_net_devices;
> +
> +	while (cur) {
> +		if (on_cmd) {
> +			cmd = cur;
> +		} else {
> +			if (set_net_param(&p, cmd, cur) < 0)
> +				goto done;
> +		}
> +		on_cmd = !on_cmd;
> +
> +		cur = strtok(NULL, ",=");
> +	};
> +
> +	num_net_devices++;
> +
> +	net_params = realloc(net_params, num_net_devices * sizeof(*net_params));
> +	if (net_params == NULL)
> +		die("Failed adding new network device");
> +
> +	net_params[num_net_devices - 1] = p;
> +
> +done:
> +	return 0;
> +}
> +
>  static int shmem_parser(const struct option *opt, const char *arg, int unset)
>  {
>  	const u64 default_size = SHMEM_DEFAULT_SIZE;
> @@ -339,18 +425,9 @@ static const struct option options[] = {
>  			"Kernel command line arguments"),
>  
>  	OPT_GROUP("Networking options:"),
> -	OPT_STRING('n', "network", &network, "user, tap, none",
> -			"Network to use"),
> -	OPT_STRING('\0', "host-ip", &host_ip, "a.b.c.d",
> -			"Assign this address to the host side networking"),
> -	OPT_STRING('\0', "guest-ip", &guest_ip, "a.b.c.d",
> -			"Assign this address to the guest side networking"),
> -	OPT_STRING('\0', "host-mac", &host_mac, "aa:bb:cc:dd:ee:ff",
> -			"Assign this address to the host side NIC"),
> -	OPT_STRING('\0', "guest-mac", &guest_mac, "aa:bb:cc:dd:ee:ff",
> -			"Assign this address to the guest side NIC"),
> -	OPT_STRING('\0', "tapscript", &script, "Script path",
> -			 "Assign a script to process created tap device"),
> +	OPT_CALLBACK_DEFAULT('n', "network", NULL, "network params",
> +		     "Create a new guest NIC",
> +		     netdev_parser, NULL),
>  
>  	OPT_GROUP("BIOS options:"),
>  	OPT_INTEGER('\0', "vidmode", &vidmode,
> @@ -615,7 +692,6 @@ void kvm_run_help(void)
>  
>  int kvm_cmd_run(int argc, const char **argv, const char *prefix)
>  {
> -	struct virtio_net_parameters net_params;
>  	static char real_cmdline[2048], default_name[20];
>  	struct framebuffer *fb = NULL;
>  	unsigned int nr_online_cpus;
> @@ -823,32 +899,24 @@ int kvm_cmd_run(int argc, const char **argv, const char *prefix)
>  
>  	virtio_9p__init(kvm);
>  
> -	if (strncmp(network, "none", 4)) {
> -		net_params.guest_ip = guest_ip;
> -		net_params.host_ip = host_ip;
> -		net_params.kvm = kvm;
> -		net_params.script = script;
> -		sscanf(guest_mac, "%hhx:%hhx:%hhx:%hhx:%hhx:%hhx",
> -			net_params.guest_mac,
> -			net_params.guest_mac+1,
> -			net_params.guest_mac+2,
> -			net_params.guest_mac+3,
> -			net_params.guest_mac+4,
> -			net_params.guest_mac+5);
> -		sscanf(host_mac, "%hhx:%hhx:%hhx:%hhx:%hhx:%hhx",
> -			net_params.host_mac,
> -			net_params.host_mac+1,
> -			net_params.host_mac+2,
> -			net_params.host_mac+3,
> -			net_params.host_mac+4,
> -			net_params.host_mac+5);
> -
> -		if (!strncmp(network, "user", 4))
> -			net_params.mode = NET_MODE_USER;
> -		else if (!strncmp(network, "tap", 3))
> -			net_params.mode = NET_MODE_TAP;
> -		else
> -			die("Unkown network mode %s, please use -network user, tap, none", network);
> +	for (i = 0; i < num_net_devices; i++) {
> +		net_params[i].kvm = kvm;
> +		virtio_net__init(&net_params[i]);
> +	}
> +
> +	if (num_net_devices == 0 && no_net == 0) {
> +		struct virtio_net_parameters net_params;
> +
> +		net_params = (struct virtio_net_parameters) {
> +			.guest_ip	= guest_ip,
> +			.host_ip	= host_ip,
> +			.kvm		= kvm,
> +			.script		= script,
> +			.mode		= NET_MODE_USER,
> +		};
> +		str_to_mac(guest_mac, net_params.guest_mac);
> +		str_to_mac(host_mac, net_params.host_mac);
> +
>  		virtio_net__init(&net_params);
>  	}
>  
> diff --git a/tools/kvm/virtio/net.c b/tools/kvm/virtio/net.c
> index 4700483..1e33891 100644
> --- a/tools/kvm/virtio/net.c
> +++ b/tools/kvm/virtio/net.c
> @@ -43,6 +43,7 @@ struct net_dev_operations {
>  struct net_dev {
>  	pthread_mutex_t			mutex;
>  	struct virtio_pci		vpci;
> +	struct list_head		list;
>  
>  	struct virt_queue		vqs[VIRTIO_NET_NUM_QUEUES];
>  	struct virtio_net_config	config;
> @@ -64,48 +65,41 @@ struct net_dev {
>  
>  	struct uip_info			info;
>  	struct net_dev_operations	*ops;
> +	struct kvm			*kvm;
>  };
>  
> -static struct net_dev ndev = {
> -	.mutex	= PTHREAD_MUTEX_INITIALIZER,
> -
> -	.config = {
> -		.status			= VIRTIO_NET_S_LINK_UP,
> -	},
> -	.info = {
> -		.buf_nr			= 20,
> -	}
> -};
> +static LIST_HEAD(ndevs);
>  
>  static void *virtio_net_rx_thread(void *p)
>  {
>  	struct iovec iov[VIRTIO_NET_QUEUE_SIZE];
>  	struct virt_queue *vq;
>  	struct kvm *kvm;
> +	struct net_dev *ndev = p;
>  	u16 out, in;
>  	u16 head;
>  	int len;
>  
> -	kvm	= p;
> -	vq	= &ndev.vqs[VIRTIO_NET_RX_QUEUE];
> +	kvm	= ndev->kvm;
> +	vq	= &ndev->vqs[VIRTIO_NET_RX_QUEUE];
>  
>  	while (1) {
>  
> -		mutex_lock(&ndev.io_rx_lock);
> +		mutex_lock(&ndev->io_rx_lock);
>  		if (!virt_queue__available(vq))
> -			pthread_cond_wait(&ndev.io_rx_cond, &ndev.io_rx_lock);
> -		mutex_unlock(&ndev.io_rx_lock);
> +			pthread_cond_wait(&ndev->io_rx_cond, &ndev->io_rx_lock);
> +		mutex_unlock(&ndev->io_rx_lock);
>  
>  		while (virt_queue__available(vq)) {
>  
>  			head = virt_queue__get_iov(vq, iov, &out, &in, kvm);
>  
> -			len = ndev.ops->rx(iov, in, &ndev);
> +			len = ndev->ops->rx(iov, in, ndev);
>  
>  			virt_queue__set_used_elem(vq, head, len);
>  
>  			/* We should interrupt guest right now, otherwise latency is huge. */
> -			virtio_pci__signal_vq(kvm, &ndev.vpci, VIRTIO_NET_RX_QUEUE);
> +			virtio_pci__signal_vq(kvm, &ndev->vpci, VIRTIO_NET_RX_QUEUE);
>  		}
>  
>  	}
> @@ -120,29 +114,30 @@ static void *virtio_net_tx_thread(void *p)
>  	struct iovec iov[VIRTIO_NET_QUEUE_SIZE];
>  	struct virt_queue *vq;
>  	struct kvm *kvm;
> +	struct net_dev *ndev = p;
>  	u16 out, in;
>  	u16 head;
>  	int len;
>  
> -	kvm	= p;
> -	vq	= &ndev.vqs[VIRTIO_NET_TX_QUEUE];
> +	kvm	= ndev->kvm;
> +	vq	= &ndev->vqs[VIRTIO_NET_TX_QUEUE];
>  
>  	while (1) {
> -		mutex_lock(&ndev.io_tx_lock);
> +		mutex_lock(&ndev->io_tx_lock);
>  		if (!virt_queue__available(vq))
> -			pthread_cond_wait(&ndev.io_tx_cond, &ndev.io_tx_lock);
> -		mutex_unlock(&ndev.io_tx_lock);
> +			pthread_cond_wait(&ndev->io_tx_cond, &ndev->io_tx_lock);
> +		mutex_unlock(&ndev->io_tx_lock);
>  
>  		while (virt_queue__available(vq)) {
>  
>  			head = virt_queue__get_iov(vq, iov, &out, &in, kvm);
>  
> -			len = ndev.ops->tx(iov, out, &ndev);
> +			len = ndev->ops->tx(iov, out, ndev);
>  
>  			virt_queue__set_used_elem(vq, head, len);
>  		}
>  
> -		virtio_pci__signal_vq(kvm, &ndev.vpci, VIRTIO_NET_TX_QUEUE);
> +		virtio_pci__signal_vq(kvm, &ndev->vpci, VIRTIO_NET_TX_QUEUE);
>  	}
>  
>  	pthread_exit(NULL);
> @@ -151,58 +146,58 @@ static void *virtio_net_tx_thread(void *p)
>  
>  }
>  
> -static void virtio_net_handle_callback(struct kvm *kvm, u16 queue_index)
> +static void virtio_net_handle_callback(struct kvm *kvm, struct net_dev *ndev, int queue)
>  {
> -	switch (queue_index) {
> +	switch (queue) {
>  	case VIRTIO_NET_TX_QUEUE:
> -		mutex_lock(&ndev.io_tx_lock);
> -		pthread_cond_signal(&ndev.io_tx_cond);
> -		mutex_unlock(&ndev.io_tx_lock);
> +		mutex_lock(&ndev->io_tx_lock);
> +		pthread_cond_signal(&ndev->io_tx_cond);
> +		mutex_unlock(&ndev->io_tx_lock);
>  		break;
>  	case VIRTIO_NET_RX_QUEUE:
> -		mutex_lock(&ndev.io_rx_lock);
> -		pthread_cond_signal(&ndev.io_rx_cond);
> -		mutex_unlock(&ndev.io_rx_lock);
> +		mutex_lock(&ndev->io_rx_lock);
> +		pthread_cond_signal(&ndev->io_rx_cond);
> +		mutex_unlock(&ndev->io_rx_lock);
>  		break;
>  	default:
> -		pr_warning("Unknown queue index %u", queue_index);
> +		pr_warning("Unknown queue index %u", queue);
>  	}
>  }
>  
> -static bool virtio_net__tap_init(const struct virtio_net_parameters *params)
> +static bool virtio_net__tap_init(const struct virtio_net_parameters *params,
> +					struct net_dev *ndev)
>  {
>  	int sock = socket(AF_INET, SOCK_STREAM, 0);
>  	int pid, status, offload, hdr_len;
>  	struct sockaddr_in sin = {0};
>  	struct ifreq ifr;
>  
> -	ndev.tap_fd = open("/dev/net/tun", O_RDWR);
> -	if (ndev.tap_fd < 0) {
> +	ndev->tap_fd = open("/dev/net/tun", O_RDWR);
> +	if (ndev->tap_fd < 0) {
>  		pr_warning("Unable to open /dev/net/tun");
>  		goto fail;
>  	}
>  
>  	memset(&ifr, 0, sizeof(ifr));
>  	ifr.ifr_flags = IFF_TAP | IFF_NO_PI | IFF_VNET_HDR;
> -	if (ioctl(ndev.tap_fd, TUNSETIFF, &ifr) < 0) {
> +	if (ioctl(ndev->tap_fd, TUNSETIFF, &ifr) < 0) {
>  		pr_warning("Config tap device error. Are you root?");
>  		goto fail;
>  	}
>  
> -	strncpy(ndev.tap_name, ifr.ifr_name, sizeof(ndev.tap_name));
> +	strncpy(ndev->tap_name, ifr.ifr_name, sizeof(ndev->tap_name));
>  
> -	if (ioctl(ndev.tap_fd, TUNSETNOCSUM, 1) < 0) {
> +	if (ioctl(ndev->tap_fd, TUNSETNOCSUM, 1) < 0) {
>  		pr_warning("Config tap device TUNSETNOCSUM error");
>  		goto fail;
>  	}
>  
>  	hdr_len = sizeof(struct virtio_net_hdr);
> -	if (ioctl(ndev.tap_fd, TUNSETVNETHDRSZ, &hdr_len) < 0) {
> +	if (ioctl(ndev->tap_fd, TUNSETVNETHDRSZ, &hdr_len) < 0)
>  		pr_warning("Config tap device TUNSETVNETHDRSZ error");
> -	}
>  
>  	offload = TUN_F_CSUM | TUN_F_TSO4 | TUN_F_TSO6 | TUN_F_UFO;
> -	if (ioctl(ndev.tap_fd, TUNSETOFFLOAD, offload) < 0) {
> +	if (ioctl(ndev->tap_fd, TUNSETOFFLOAD, offload) < 0) {
>  		pr_warning("Config tap device TUNSETOFFLOAD error");
>  		goto fail;
>  	}
> @@ -210,7 +205,7 @@ static bool virtio_net__tap_init(const struct virtio_net_parameters *params)
>  	if (strcmp(params->script, "none")) {
>  		pid = fork();
>  		if (pid == 0) {
> -			execl(params->script, params->script, ndev.tap_name, NULL);
> +			execl(params->script, params->script, ndev->tap_name, NULL);
>  			_exit(1);
>  		} else {
>  			waitpid(pid, &status, 0);
> @@ -221,7 +216,7 @@ static bool virtio_net__tap_init(const struct virtio_net_parameters *params)
>  		}
>  	} else {
>  		memset(&ifr, 0, sizeof(ifr));
> -		strncpy(ifr.ifr_name, ndev.tap_name, sizeof(ndev.tap_name));
> +		strncpy(ifr.ifr_name, ndev->tap_name, sizeof(ndev->tap_name));
>  		sin.sin_addr.s_addr = inet_addr(params->host_ip);
>  		memcpy(&(ifr.ifr_addr), &sin, sizeof(ifr.ifr_addr));
>  		ifr.ifr_addr.sa_family = AF_INET;
> @@ -232,7 +227,7 @@ static bool virtio_net__tap_init(const struct virtio_net_parameters *params)
>  	}
>  
>  	memset(&ifr, 0, sizeof(ifr));
> -	strncpy(ifr.ifr_name, ndev.tap_name, sizeof(ndev.tap_name));
> +	strncpy(ifr.ifr_name, ndev->tap_name, sizeof(ndev->tap_name));
>  	ioctl(sock, SIOCGIFFLAGS, &ifr);
>  	ifr.ifr_flags |= IFF_UP | IFF_RUNNING;
>  	if (ioctl(sock, SIOCSIFFLAGS, &ifr) < 0)
> @@ -245,22 +240,22 @@ static bool virtio_net__tap_init(const struct virtio_net_parameters *params)
>  fail:
>  	if (sock >= 0)
>  		close(sock);
> -	if (ndev.tap_fd >= 0)
> -		close(ndev.tap_fd);
> +	if (ndev->tap_fd >= 0)
> +		close(ndev->tap_fd);
>  
>  	return 0;
>  }
>  
> -static void virtio_net__io_thread_init(struct kvm *kvm)
> +static void virtio_net__io_thread_init(struct kvm *kvm, struct net_dev *ndev)
>  {
> -	pthread_mutex_init(&ndev.io_rx_lock, NULL);
> -	pthread_cond_init(&ndev.io_tx_cond, NULL);
> +	pthread_mutex_init(&ndev->io_rx_lock, NULL);
> +	pthread_cond_init(&ndev->io_tx_cond, NULL);
>  
> -	pthread_mutex_init(&ndev.io_rx_lock, NULL);
> -	pthread_cond_init(&ndev.io_tx_cond, NULL);
> +	pthread_mutex_init(&ndev->io_rx_lock, NULL);
> +	pthread_cond_init(&ndev->io_tx_cond, NULL);
>  
> -	pthread_create(&ndev.io_rx_thread, NULL, virtio_net_rx_thread, (void *)kvm);
> -	pthread_create(&ndev.io_tx_thread, NULL, virtio_net_tx_thread, (void *)kvm);
> +	pthread_create(&ndev->io_rx_thread, NULL, virtio_net_rx_thread, ndev);
> +	pthread_create(&ndev->io_tx_thread, NULL, virtio_net_tx_thread, ndev);
>  }
>  
>  static inline int tap_ops_tx(struct iovec *iov, u16 out, struct net_dev *ndev)
> @@ -345,7 +340,9 @@ static int init_vq(struct kvm *kvm, void *dev, u32 vq, u32 pfn)
>  
>  static int notify_vq(struct kvm *kvm, void *dev, u32 vq)
>  {
> -	virtio_net_handle_callback(kvm, vq);
> +	struct net_dev *ndev = dev;
> +
> +	virtio_net_handle_callback(kvm, ndev, vq);
>  
>  	return 0;
>  }
> @@ -365,27 +362,43 @@ static int get_size_vq(struct kvm *kvm, void *dev, u32 vq)
>  void virtio_net__init(const struct virtio_net_parameters *params)
>  {
>  	int i;
> +	struct net_dev *ndev;
> +
> +	if (!params)
> +		return;
> +
> +	ndev = calloc(1, sizeof(struct net_dev));
> +	if (ndev == NULL)
> +		die("Failed allocating ndev");
> +
> +	list_add_tail(&ndev->list, &ndevs);
> +
> +	ndev->kvm = params->kvm;
> +
> +	mutex_init(&ndev->mutex);
> +	ndev->config.status = VIRTIO_NET_S_LINK_UP;
>  
>  	for (i = 0 ; i < 6 ; i++) {
> -		ndev.config.mac[i]		= params->guest_mac[i];
> -		ndev.info.guest_mac.addr[i]	= params->guest_mac[i];
> -		ndev.info.host_mac.addr[i]	= params->host_mac[i];
> +		ndev->config.mac[i]		= params->guest_mac[i];
> +		ndev->info.guest_mac.addr[i]	= params->guest_mac[i];
> +		ndev->info.host_mac.addr[i]	= params->host_mac[i];
>  	}
>  
> -	ndev.mode = params->mode;
> -	if (ndev.mode == NET_MODE_TAP) {
> -		virtio_net__tap_init(params);

we need to check if virtio_net__tap_init is successful. if it fails, for
example lack of permission, we should exit and give up this net device.
Otherwise, we will have this:

$ ./kvm run -n mode=tap


[    1.490781] registered taskstats version 1
[    1.492781] BUG: unable to handle kernel NULL pointer dereference at
0000001c
[    1.493781] IP: [<c14f3236>] virtnet_poll+0x16e/0x408
[    1.493781] *pde = 00000000
[    1.493781] Oops: 0000 [#1] PREEMPT SMP
[    1.493781] Modules linked in:
[    1.493781]
[    1.493781] Pid: 1, comm: swapper Tainted: G        W   3.1.0-rc3+ #77
[    1.493781] EIP: 0060:[<c14f3236>] EFLAGS: 00010286 CPU: 1
[    1.493781] EIP is at virtnet_poll+0x16e/0x408
[    1.493781] EAX: 00001000 EBX: db4bb0c0 ECX: db7cd778 EDX: 00001000
[    1.493781] ESI: 00000000 EDI: db7cd6c0 EBP: db487fa8 ESP: db487f64
[    1.493781]  DS: 007b ES: 007b FS: 00d8 GS: 0000 SS: 0068
[    1.493781] Process swapper (pid: 1, ti=db486000 task=db450000
task.ti=db458000)
[    1.493781] Stack:
[    1.493781]  db487f98 19dfb000 db5e1400 00000080 c1b0df60 db6ff000
00000000 00000010
[    1.493781]  00000080 dbcebfe0 db5e1414 db5e1000 ffffffff fffec005
db5e1414 db906dc0
[    1.493781]  c1a39a0c db487fd4 c15e4869 fffb71f7 db906dc8 00000080
0000012c 00000000
[    1.493781] Call Trace:
[    1.493781]  [<c15e4869>] net_rx_action+0x8e/0x177
[    1.493781]  [<c1066128>] __do_softirq+0xa7/0x158
[    1.493781]  [<c1066081>] ? __tasklet_hi_schedule_first+0x2b/0x2b
[    1.493781]  <IRQ>
[    1.493781]  [<c1065e82>] ? _local_bh_enable_ip.isra.9+0x65/0x86
[    1.493781]  [<c1065eab>] ? local_bh_enable_ip+0x8/0xa
[    1.493781]  [<c16a7a78>] ? _raw_spin_unlock_bh+0x18/0x1a
[    1.493781]  [<c15e59c5>] ? dev_set_rx_mode+0x22/0x26
[    1.493781]  [<c15e5a5f>] ? __dev_open+0x96/0xa6
[    1.493781]  [<c15e5c23>] ? __dev_change_flags+0x97/0x10e
[    1.493781]  [<c15e5cfe>] ? dev_change_flags+0x13/0x3f
[    1.493781]  [<c1acfe6f>] ? ip_auto_config+0x160/0xcf8
[    1.493781]  [<c1393c86>] ? extract_entropy+0x45/0x71
[    1.493781]  [<c1059e35>] ? get_parent_ip+0xb/0x31
[    1.493781]  [<c16aa6b7>] ? sub_preempt_count+0x7c/0x89
[    1.493781]  [<c16a7d24>] ? _raw_spin_unlock+0x1c/0x27
[    1.493781]  [<c1629173>] ? tcp_set_default_congestion_control+0x8c/0x95
[    1.493781]  [<c1001159>] ? do_one_initcall+0x71/0x114
[    1.493781]  [<c1acfd0f>] ? root_nfs_parse_addr+0x91/0x91
[    1.493781]  [<c1a9c7ab>] ? kernel_init+0xab/0x11d
[    1.493781]  [<c1a9c700>] ? start_kernel+0x301/0x301
[    1.493781]  [<c16acfb6>] ? kernel_thread_helper+0x6/0xd
[    1.493781] Code: 89 d8 e8 23 94 0e 00 8b 4d dc 89 c7 f3 a4 8b 55 dc
8b 4d d8 29 55 f0 8b 75 e0 01 d1 eb 13 8d 45 f0 89 f2 50 89 d8 e8 ae f2
ff ff <8b> 76 1c 31 c9 58 83 7d f0 00 75 e7 85 f6 89 75 e0 0f 84 6e 02
[    1.493781] EIP: [<c14f3236>] virtnet_poll+0x16e/0x408 SS:ESP
0068:db487f64
[    1.493781] CR2: 000000000000001c
[    1.549772] ---[ end trace 4eaa2a86a8e2da27 ]---
[    1.550772] Kernel panic - not syncing: Fatal exception in interrupt
[    1.551772] Pid: 1, comm: swapper Tainted: G      D W   3.1.0-rc3+ #77
[    1.553771] Call Trace:
[    1.553771]  [<c169ca33>] panic+0x58/0x156
[    1.554771]  [<c16a921a>] oops_end+0x8c/0x9b
[    1.555771]  [<c169c4e7>] no_context+0x116/0x120
[    1.555771]  [<c169c5e1>] __bad_area_nosemaphore+0xf0/0xf8
[    1.557771]  [<c169c5f6>] bad_area_nosemaphore+0xd/0x10
[    1.558771]  [<c16aa4b5>] do_page_fault+0x174/0x2fa
[    1.559770]  [<c107bad0>] ? sched_clock_local+0x10/0x14b
[    1.560770]  [<c15db33f>] ? __netdev_alloc_skb+0x17/0x34
[    1.561770]  [<c10e9b84>] ? __kmalloc_track_caller+0xb7/0xc7
[    1.563770]  [<c15db33f>] ? __netdev_alloc_skb+0x17/0x34
[    1.564770]  [<c16aa341>] ? spurious_fault+0xa8/0xa8
[    1.565770]  [<c16a89d6>] error_code+0x5a/0x60
[    1.566769]  [<c16aa341>] ? spurious_fault+0xa8/0xa8
[    1.567769]  [<c14f3236>] ? virtnet_poll+0x16e/0x408
[    1.567769]  [<c15e4869>] net_rx_action+0x8e/0x177
[    1.568769]  [<c1066128>] __do_softirq+0xa7/0x158
[    1.569769]  [<c1066081>] ? __tasklet_hi_schedule_first+0x2b/0x2b
[    1.569769]  <IRQ>  [<c1065e82>] ? _local_bh_enable_ip.isra.9+0x65/0x86
[    1.570769]  [<c1065eab>] ? local_bh_enable_ip+0x8/0xa
[    1.571769]  [<c16a7a78>] ? _raw_spin_unlock_bh+0x18/0x1a
[    1.571769]  [<c15e59c5>] ? dev_set_rx_mode+0x22/0x26
[    1.572768]  [<c15e5a5f>] ? __dev_open+0x96/0xa6
[    1.573768]  [<c15e5c23>] ? __dev_change_flags+0x97/0x10e
[    1.573768]  [<c15e5cfe>] ? dev_change_flags+0x13/0x3f
[    1.574768]  [<c1acfe6f>] ? ip_auto_config+0x160/0xcf8
[    1.574768]  [<c1393c86>] ? extract_entropy+0x45/0x71
[    1.575768]  [<c1059e35>] ? get_parent_ip+0xb/0x31
[    1.576768]  [<c16aa6b7>] ? sub_preempt_count+0x7c/0x89
[    1.576768]  [<c16a7d24>] ? _raw_spin_unlock+0x1c/0x27
[    1.577768]  [<c1629173>] ? tcp_set_default_congestion_control+0x8c/0x95
[    1.578768]  [<c1001159>] ? do_one_initcall+0x71/0x114
[    1.578768]  [<c1acfd0f>] ? root_nfs_parse_addr+0x91/0x91
[    1.579767]  [<c1a9c7ab>] ? kernel_init+0xab/0x11d
[    1.580767]  [<c1a9c700>] ? start_kernel+0x301/0x301
[    1.581767]  [<c16acfb6>] ? kernel_thread_helper+0x6/0xd
[    1.582767] Rebooting in 1 seconds..
  # KVM session ended normally.


> -		ndev.ops = &tap_ops;
> +	ndev->mode = params->mode;
> +	if (ndev->mode == NET_MODE_TAP) {
> +		virtio_net__tap_init(params, ndev);
> +		ndev->ops = &tap_ops;
>  	} else {
> -		ndev.info.host_ip		= ntohl(inet_addr(params->host_ip));
> -		ndev.info.guest_ip		= ntohl(inet_addr(params->guest_ip));
> -		ndev.info.guest_netmask		= ntohl(inet_addr("255.255.255.0"));
> -		uip_init(&ndev.info);
> -		ndev.ops = &uip_ops;
> +		ndev->info.host_ip		= ntohl(inet_addr(params->host_ip));
> +		ndev->info.guest_ip		= ntohl(inet_addr(params->guest_ip));
> +		ndev->info.guest_netmask	= ntohl(inet_addr("255.255.255.0"));
> +		ndev->info.buf_nr		= 20,
> +		uip_init(&ndev->info);
> +		ndev->ops = &uip_ops;
>  	}
>  
> -	virtio_pci__init(kvm, &ndev.vpci, &ndev, PCI_DEVICE_ID_VIRTIO_NET, VIRTIO_ID_NET);
> -	ndev.vpci.ops = (struct virtio_pci_ops) {
> +	virtio_pci__init(kvm, &ndev->vpci, ndev, PCI_DEVICE_ID_VIRTIO_NET, VIRTIO_ID_NET);
> +	ndev->vpci.ops = (struct virtio_pci_ops) {
>  		.set_config		= set_config,
>  		.get_config		= get_config,
>  		.get_host_features	= get_host_features,
> @@ -396,9 +409,9 @@ void virtio_net__init(const struct virtio_net_parameters *params)
>  		.get_size_vq		= get_size_vq,
>  	};
>  
> -	virtio_net__io_thread_init(params->kvm);
> +	virtio_net__io_thread_init(params->kvm, ndev);
>  
> -	ndev.compat_id = compat__add_message("virtio-net device was not detected",
> +	ndev->compat_id = compat__add_message("virtio-net device was not detected",
>  						"While you have requested a virtio-net device, "
>  						"the guest kernel didn't seem to detect it.\n"
>  						"Please make sure that the kernel was compiled "


-- 
Asias He

Re: [PATCH v2] kvm tools: Support multiple net devices

[Sasha Levin]

On Mon, 2011-09-26 at 19:57 +0800, Asias He wrote:
> On 09/25/2011 07:11 PM, Sasha Levin wrote:
> > This patch adds support for multiple network devices. The command
> line syntax
> > changes to the following:
> > 
> >       --network/-n [mode=[tap/user/none]] [guest_ip=[guest ip]]
> [host_ip=
> > [host_ip]] [guest_mac=[guest_mac]] [script=[script]]
> 
> This syntax is actually, no?
> 
>    --network/-n mode=tap,guest_ip=x.x.x.x
> 
> not
> 
>    --network/-n mode=tag guest_ip=x.x.x.x
> 
> 
> This works for me:
> $ sudo ./kvm run -d sid.img -p root=/dev/vda1 -k bzImage  \
> -n mode=tap,host_ip=192.168.100.1 -n mode=tap,host_ip=192.168.200.1
> 
> This does not work for me:
> $ sudo ./kvm run -d sid.img -p root=/dev/vda1 -k bzImage \
> -n mode=tap host_ip=192.168.100.1 -n mode=tap host_ip=192.168.200.1

Yes, with ',' ofcourse :)

-- 

Sasha.

Re: [PATCH v2] kvm tools: Support multiple net devices

[Asias He]

On 09/26/2011 07:59 PM, Sasha Levin wrote:
> On Mon, 2011-09-26 at 19:57 +0800, Asias He wrote:
>> On 09/25/2011 07:11 PM, Sasha Levin wrote:
>>> This patch adds support for multiple network devices. The command
>> line syntax
>>> changes to the following:
>>>
>>>       --network/-n [mode=[tap/user/none]] [guest_ip=[guest ip]]
>> [host_ip=
>>> [host_ip]] [guest_mac=[guest_mac]] [script=[script]]
>>
>> This syntax is actually, no?
>>
>>    --network/-n mode=tap,guest_ip=x.x.x.x
>>
>> not
>>
>>    --network/-n mode=tag guest_ip=x.x.x.x
>>
>>
>> This works for me:
>> $ sudo ./kvm run -d sid.img -p root=/dev/vda1 -k bzImage  \
>> -n mode=tap,host_ip=192.168.100.1 -n mode=tap,host_ip=192.168.200.1
>>
>> This does not work for me:
>> $ sudo ./kvm run -d sid.img -p root=/dev/vda1 -k bzImage \
>> -n mode=tap host_ip=192.168.100.1 -n mode=tap host_ip=192.168.200.1
> 
> Yes, with ',' ofcourse :)

Sasha, Can you drop these []?  There are way too many []. Although []
tells user this option is optional.

--network/-n [mode=[tap/user/none]] [guest_ip=[guest ip]] [host_ip=
[host_ip]] [guest_mac=[guest_mac]] [script=[script]]


Does this look ok?
--network/-n
mode=tap|user|none,guest_ip=x.x.x.x,host_ip=x.x.x.x,guest_mac=xx:xx:xx:xx:xx:xx,script=script.sh


-- 
Asias He

Re: [PATCH v2] kvm tools: Support multiple net devices

[Sasha Levin]

On Mon, 2011-09-26 at 20:11 +0800, Asias He wrote:
> On 09/26/2011 07:59 PM, Sasha Levin wrote:
> > On Mon, 2011-09-26 at 19:57 +0800, Asias He wrote:
> >> On 09/25/2011 07:11 PM, Sasha Levin wrote:
> >>> This patch adds support for multiple network devices. The command
> >> line syntax
> >>> changes to the following:
> >>>
> >>>       --network/-n [mode=[tap/user/none]] [guest_ip=[guest ip]]
> >> [host_ip=
> >>> [host_ip]] [guest_mac=[guest_mac]] [script=[script]]
> >>
> >> This syntax is actually, no?
> >>
> >>    --network/-n mode=tap,guest_ip=x.x.x.x
> >>
> >> not
> >>
> >>    --network/-n mode=tag guest_ip=x.x.x.x
> >>
> >>
> >> This works for me:
> >> $ sudo ./kvm run -d sid.img -p root=/dev/vda1 -k bzImage  \
> >> -n mode=tap,host_ip=192.168.100.1 -n mode=tap,host_ip=192.168.200.1
> >>
> >> This does not work for me:
> >> $ sudo ./kvm run -d sid.img -p root=/dev/vda1 -k bzImage \
> >> -n mode=tap host_ip=192.168.100.1 -n mode=tap host_ip=192.168.200.1
> > 
> > Yes, with ',' ofcourse :)
> 
> Sasha, Can you drop these []?  There are way too many []. Although []
> tells user this option is optional.
> 
> --network/-n [mode=[tap/user/none]] [guest_ip=[guest ip]] [host_ip=
> [host_ip]] [guest_mac=[guest_mac]] [script=[script]]
> 
> 
> Does this look ok?
> --network/-n
> mode=tap|user|none,guest_ip=x.x.x.x,host_ip=x.x.x.x,guest_mac=xx:xx:xx:xx:xx:xx,script=script.sh

We really need them to specify they are optional.

How about this:

--network/-n
[mode=tap/user/none][,guest_ip=ip][,host_ip=ip][,guest_mac=mac][,script=file]

-- 

Sasha.

Re: [PATCH v2] kvm tools: Support multiple net devices

[Asias He]

On 09/26/2011 08:14 PM, Sasha Levin wrote:
> On Mon, 2011-09-26 at 20:11 +0800, Asias He wrote:
>> On 09/26/2011 07:59 PM, Sasha Levin wrote:
>>> On Mon, 2011-09-26 at 19:57 +0800, Asias He wrote:
>>>> On 09/25/2011 07:11 PM, Sasha Levin wrote:
>>>>> This patch adds support for multiple network devices. The command
>>>> line syntax
>>>>> changes to the following:
>>>>>
>>>>>       --network/-n [mode=[tap/user/none]] [guest_ip=[guest ip]]
>>>> [host_ip=
>>>>> [host_ip]] [guest_mac=[guest_mac]] [script=[script]]
>>>>
>>>> This syntax is actually, no?
>>>>
>>>>    --network/-n mode=tap,guest_ip=x.x.x.x
>>>>
>>>> not
>>>>
>>>>    --network/-n mode=tag guest_ip=x.x.x.x
>>>>
>>>>
>>>> This works for me:
>>>> $ sudo ./kvm run -d sid.img -p root=/dev/vda1 -k bzImage  \
>>>> -n mode=tap,host_ip=192.168.100.1 -n mode=tap,host_ip=192.168.200.1
>>>>
>>>> This does not work for me:
>>>> $ sudo ./kvm run -d sid.img -p root=/dev/vda1 -k bzImage \
>>>> -n mode=tap host_ip=192.168.100.1 -n mode=tap host_ip=192.168.200.1
>>>
>>> Yes, with ',' ofcourse :)
>>
>> Sasha, Can you drop these []?  There are way too many []. Although []
>> tells user this option is optional.
>>
>> --network/-n [mode=[tap/user/none]] [guest_ip=[guest ip]] [host_ip=
>> [host_ip]] [guest_mac=[guest_mac]] [script=[script]]
>>
>>
>> Does this look ok?
>> --network/-n
>> mode=tap|user|none,guest_ip=x.x.x.x,host_ip=x.x.x.x,guest_mac=xx:xx:xx:xx:xx:xx,script=script.sh
> 
> We really need them to specify they are optional.
> 
> How about this:
> 
> --network/-n
> [mode=tap/user/none][,guest_ip=ip][,host_ip=ip][,guest_mac=mac][,script=file]

OK.

-- 
Asias He

Re: [PATCH v2] kvm tools: Support multiple net devices

[Pekka Enberg]

On Mon, 26 Sep 2011, Asias He wrote:
> $ ./kvm run -n mode=tap
>
>
> [    1.490781] registered taskstats version 1
> [    1.492781] BUG: unable to handle kernel NULL pointer dereference at
> 0000001c
> [    1.493781] IP: [<c14f3236>] virtnet_poll+0x16e/0x408
> [    1.493781] *pde = 00000000
> [    1.493781] Oops: 0000 [#1] PREEMPT SMP
> [    1.493781] Modules linked in:
> [    1.493781]
> [    1.493781] Pid: 1, comm: swapper Tainted: G        W   3.1.0-rc3+ #77
> [    1.493781] EIP: 0060:[<c14f3236>] EFLAGS: 00010286 CPU: 1
> [    1.493781] EIP is at virtnet_poll+0x16e/0x408
> [    1.493781] EAX: 00001000 EBX: db4bb0c0 ECX: db7cd778 EDX: 00001000
> [    1.493781] ESI: 00000000 EDI: db7cd6c0 EBP: db487fa8 ESP: db487f64
> [    1.493781]  DS: 007b ES: 007b FS: 00d8 GS: 0000 SS: 0068
> [    1.493781] Process swapper (pid: 1, ti=db486000 task=db450000
> task.ti=db458000)
> [    1.493781] Stack:
> [    1.493781]  db487f98 19dfb000 db5e1400 00000080 c1b0df60 db6ff000
> 00000000 00000010
> [    1.493781]  00000080 dbcebfe0 db5e1414 db5e1000 ffffffff fffec005
> db5e1414 db906dc0
> [    1.493781]  c1a39a0c db487fd4 c15e4869 fffb71f7 db906dc8 00000080
> 0000012c 00000000
> [    1.493781] Call Trace:
> [    1.493781]  [<c15e4869>] net_rx_action+0x8e/0x177
> [    1.493781]  [<c1066128>] __do_softirq+0xa7/0x158
> [    1.493781]  [<c1066081>] ? __tasklet_hi_schedule_first+0x2b/0x2b
> [    1.493781]  <IRQ>
> [    1.493781]  [<c1065e82>] ? _local_bh_enable_ip.isra.9+0x65/0x86
> [    1.493781]  [<c1065eab>] ? local_bh_enable_ip+0x8/0xa
> [    1.493781]  [<c16a7a78>] ? _raw_spin_unlock_bh+0x18/0x1a
> [    1.493781]  [<c15e59c5>] ? dev_set_rx_mode+0x22/0x26
> [    1.493781]  [<c15e5a5f>] ? __dev_open+0x96/0xa6
> [    1.493781]  [<c15e5c23>] ? __dev_change_flags+0x97/0x10e
> [    1.493781]  [<c15e5cfe>] ? dev_change_flags+0x13/0x3f
> [    1.493781]  [<c1acfe6f>] ? ip_auto_config+0x160/0xcf8
> [    1.493781]  [<c1393c86>] ? extract_entropy+0x45/0x71
> [    1.493781]  [<c1059e35>] ? get_parent_ip+0xb/0x31
> [    1.493781]  [<c16aa6b7>] ? sub_preempt_count+0x7c/0x89
> [    1.493781]  [<c16a7d24>] ? _raw_spin_unlock+0x1c/0x27
> [    1.493781]  [<c1629173>] ? tcp_set_default_congestion_control+0x8c/0x95
> [    1.493781]  [<c1001159>] ? do_one_initcall+0x71/0x114
> [    1.493781]  [<c1acfd0f>] ? root_nfs_parse_addr+0x91/0x91
> [    1.493781]  [<c1a9c7ab>] ? kernel_init+0xab/0x11d
> [    1.493781]  [<c1a9c700>] ? start_kernel+0x301/0x301
> [    1.493781]  [<c16acfb6>] ? kernel_thread_helper+0x6/0xd
> [    1.493781] Code: 89 d8 e8 23 94 0e 00 8b 4d dc 89 c7 f3 a4 8b 55 dc
> 8b 4d d8 29 55 f0 8b 75 e0 01 d1 eb 13 8d 45 f0 89 f2 50 89 d8 e8 ae f2
> ff ff <8b> 76 1c 31 c9 58 83 7d f0 00 75 e7 85 f6 89 75 e0 0f 84 6e 02
> [    1.493781] EIP: [<c14f3236>] virtnet_poll+0x16e/0x408 SS:ESP
> 0068:db487f64
> [    1.493781] CR2: 000000000000001c
> [    1.549772] ---[ end trace 4eaa2a86a8e2da27 ]---
> [    1.550772] Kernel panic - not syncing: Fatal exception in interrupt
> [    1.551772] Pid: 1, comm: swapper Tainted: G      D W   3.1.0-rc3+ #77
> [    1.553771] Call Trace:
> [    1.553771]  [<c169ca33>] panic+0x58/0x156
> [    1.554771]  [<c16a921a>] oops_end+0x8c/0x9b
> [    1.555771]  [<c169c4e7>] no_context+0x116/0x120
> [    1.555771]  [<c169c5e1>] __bad_area_nosemaphore+0xf0/0xf8
> [    1.557771]  [<c169c5f6>] bad_area_nosemaphore+0xd/0x10
> [    1.558771]  [<c16aa4b5>] do_page_fault+0x174/0x2fa
> [    1.559770]  [<c107bad0>] ? sched_clock_local+0x10/0x14b
> [    1.560770]  [<c15db33f>] ? __netdev_alloc_skb+0x17/0x34
> [    1.561770]  [<c10e9b84>] ? __kmalloc_track_caller+0xb7/0xc7
> [    1.563770]  [<c15db33f>] ? __netdev_alloc_skb+0x17/0x34
> [    1.564770]  [<c16aa341>] ? spurious_fault+0xa8/0xa8
> [    1.565770]  [<c16a89d6>] error_code+0x5a/0x60
> [    1.566769]  [<c16aa341>] ? spurious_fault+0xa8/0xa8
> [    1.567769]  [<c14f3236>] ? virtnet_poll+0x16e/0x408
> [    1.567769]  [<c15e4869>] net_rx_action+0x8e/0x177
> [    1.568769]  [<c1066128>] __do_softirq+0xa7/0x158
> [    1.569769]  [<c1066081>] ? __tasklet_hi_schedule_first+0x2b/0x2b
> [    1.569769]  <IRQ>  [<c1065e82>] ? _local_bh_enable_ip.isra.9+0x65/0x86
> [    1.570769]  [<c1065eab>] ? local_bh_enable_ip+0x8/0xa
> [    1.571769]  [<c16a7a78>] ? _raw_spin_unlock_bh+0x18/0x1a
> [    1.571769]  [<c15e59c5>] ? dev_set_rx_mode+0x22/0x26
> [    1.572768]  [<c15e5a5f>] ? __dev_open+0x96/0xa6
> [    1.573768]  [<c15e5c23>] ? __dev_change_flags+0x97/0x10e
> [    1.573768]  [<c15e5cfe>] ? dev_change_flags+0x13/0x3f
> [    1.574768]  [<c1acfe6f>] ? ip_auto_config+0x160/0xcf8
> [    1.574768]  [<c1393c86>] ? extract_entropy+0x45/0x71
> [    1.575768]  [<c1059e35>] ? get_parent_ip+0xb/0x31
> [    1.576768]  [<c16aa6b7>] ? sub_preempt_count+0x7c/0x89
> [    1.576768]  [<c16a7d24>] ? _raw_spin_unlock+0x1c/0x27
> [    1.577768]  [<c1629173>] ? tcp_set_default_congestion_control+0x8c/0x95
> [    1.578768]  [<c1001159>] ? do_one_initcall+0x71/0x114
> [    1.578768]  [<c1acfd0f>] ? root_nfs_parse_addr+0x91/0x91
> [    1.579767]  [<c1a9c7ab>] ? kernel_init+0xab/0x11d
> [    1.580767]  [<c1a9c700>] ? start_kernel+0x301/0x301
> [    1.581767]  [<c16acfb6>] ? kernel_thread_helper+0x6/0xd
> [    1.582767] Rebooting in 1 seconds..
>  # KVM session ended normally.

This needs fixing before I can apply the patch, right?

Re: [PATCH v2] kvm tools: Support multiple net devices

[Asias He]

On 09/26/2011 10:54 PM, Pekka Enberg wrote:
> On Mon, 26 Sep 2011, Asias He wrote:
>> $ ./kvm run -n mode=tap
>>
>>
>> [    1.490781] registered taskstats version 1
>> [    1.492781] BUG: unable to handle kernel NULL pointer dereference at
>> 0000001c
>> [    1.493781] IP: [<c14f3236>] virtnet_poll+0x16e/0x408
>> [    1.493781] *pde = 00000000
>> [    1.493781] Oops: 0000 [#1] PREEMPT SMP
>> [    1.493781] Modules linked in:
>> [    1.493781]
>> [    1.493781] Pid: 1, comm: swapper Tainted: G        W   3.1.0-rc3+ #77
>> [    1.493781] EIP: 0060:[<c14f3236>] EFLAGS: 00010286 CPU: 1
>> [    1.493781] EIP is at virtnet_poll+0x16e/0x408
>> [    1.493781] EAX: 00001000 EBX: db4bb0c0 ECX: db7cd778 EDX: 00001000
>> [    1.493781] ESI: 00000000 EDI: db7cd6c0 EBP: db487fa8 ESP: db487f64
>> [    1.493781]  DS: 007b ES: 007b FS: 00d8 GS: 0000 SS: 0068
>> [    1.493781] Process swapper (pid: 1, ti=db486000 task=db450000
>> task.ti=db458000)
>> [    1.493781] Stack:
>> [    1.493781]  db487f98 19dfb000 db5e1400 00000080 c1b0df60 db6ff000
>> 00000000 00000010
>> [    1.493781]  00000080 dbcebfe0 db5e1414 db5e1000 ffffffff fffec005
>> db5e1414 db906dc0
>> [    1.493781]  c1a39a0c db487fd4 c15e4869 fffb71f7 db906dc8 00000080
>> 0000012c 00000000
>> [    1.493781] Call Trace:
>> [    1.493781]  [<c15e4869>] net_rx_action+0x8e/0x177
>> [    1.493781]  [<c1066128>] __do_softirq+0xa7/0x158
>> [    1.493781]  [<c1066081>] ? __tasklet_hi_schedule_first+0x2b/0x2b
>> [    1.493781]  <IRQ>
>> [    1.493781]  [<c1065e82>] ? _local_bh_enable_ip.isra.9+0x65/0x86
>> [    1.493781]  [<c1065eab>] ? local_bh_enable_ip+0x8/0xa
>> [    1.493781]  [<c16a7a78>] ? _raw_spin_unlock_bh+0x18/0x1a
>> [    1.493781]  [<c15e59c5>] ? dev_set_rx_mode+0x22/0x26
>> [    1.493781]  [<c15e5a5f>] ? __dev_open+0x96/0xa6
>> [    1.493781]  [<c15e5c23>] ? __dev_change_flags+0x97/0x10e
>> [    1.493781]  [<c15e5cfe>] ? dev_change_flags+0x13/0x3f
>> [    1.493781]  [<c1acfe6f>] ? ip_auto_config+0x160/0xcf8
>> [    1.493781]  [<c1393c86>] ? extract_entropy+0x45/0x71
>> [    1.493781]  [<c1059e35>] ? get_parent_ip+0xb/0x31
>> [    1.493781]  [<c16aa6b7>] ? sub_preempt_count+0x7c/0x89
>> [    1.493781]  [<c16a7d24>] ? _raw_spin_unlock+0x1c/0x27
>> [    1.493781]  [<c1629173>] ?
>> tcp_set_default_congestion_control+0x8c/0x95
>> [    1.493781]  [<c1001159>] ? do_one_initcall+0x71/0x114
>> [    1.493781]  [<c1acfd0f>] ? root_nfs_parse_addr+0x91/0x91
>> [    1.493781]  [<c1a9c7ab>] ? kernel_init+0xab/0x11d
>> [    1.493781]  [<c1a9c700>] ? start_kernel+0x301/0x301
>> [    1.493781]  [<c16acfb6>] ? kernel_thread_helper+0x6/0xd
>> [    1.493781] Code: 89 d8 e8 23 94 0e 00 8b 4d dc 89 c7 f3 a4 8b 55 dc
>> 8b 4d d8 29 55 f0 8b 75 e0 01 d1 eb 13 8d 45 f0 89 f2 50 89 d8 e8 ae f2
>> ff ff <8b> 76 1c 31 c9 58 83 7d f0 00 75 e7 85 f6 89 75 e0 0f 84 6e 02
>> [    1.493781] EIP: [<c14f3236>] virtnet_poll+0x16e/0x408 SS:ESP
>> 0068:db487f64
>> [    1.493781] CR2: 000000000000001c
>> [    1.549772] ---[ end trace 4eaa2a86a8e2da27 ]---
>> [    1.550772] Kernel panic - not syncing: Fatal exception in interrupt
>> [    1.551772] Pid: 1, comm: swapper Tainted: G      D W   3.1.0-rc3+ #77
>> [    1.553771] Call Trace:
>> [    1.553771]  [<c169ca33>] panic+0x58/0x156
>> [    1.554771]  [<c16a921a>] oops_end+0x8c/0x9b
>> [    1.555771]  [<c169c4e7>] no_context+0x116/0x120
>> [    1.555771]  [<c169c5e1>] __bad_area_nosemaphore+0xf0/0xf8
>> [    1.557771]  [<c169c5f6>] bad_area_nosemaphore+0xd/0x10
>> [    1.558771]  [<c16aa4b5>] do_page_fault+0x174/0x2fa
>> [    1.559770]  [<c107bad0>] ? sched_clock_local+0x10/0x14b
>> [    1.560770]  [<c15db33f>] ? __netdev_alloc_skb+0x17/0x34
>> [    1.561770]  [<c10e9b84>] ? __kmalloc_track_caller+0xb7/0xc7
>> [    1.563770]  [<c15db33f>] ? __netdev_alloc_skb+0x17/0x34
>> [    1.564770]  [<c16aa341>] ? spurious_fault+0xa8/0xa8
>> [    1.565770]  [<c16a89d6>] error_code+0x5a/0x60
>> [    1.566769]  [<c16aa341>] ? spurious_fault+0xa8/0xa8
>> [    1.567769]  [<c14f3236>] ? virtnet_poll+0x16e/0x408
>> [    1.567769]  [<c15e4869>] net_rx_action+0x8e/0x177
>> [    1.568769]  [<c1066128>] __do_softirq+0xa7/0x158
>> [    1.569769]  [<c1066081>] ? __tasklet_hi_schedule_first+0x2b/0x2b
>> [    1.569769]  <IRQ>  [<c1065e82>] ?
>> _local_bh_enable_ip.isra.9+0x65/0x86
>> [    1.570769]  [<c1065eab>] ? local_bh_enable_ip+0x8/0xa
>> [    1.571769]  [<c16a7a78>] ? _raw_spin_unlock_bh+0x18/0x1a
>> [    1.571769]  [<c15e59c5>] ? dev_set_rx_mode+0x22/0x26
>> [    1.572768]  [<c15e5a5f>] ? __dev_open+0x96/0xa6
>> [    1.573768]  [<c15e5c23>] ? __dev_change_flags+0x97/0x10e
>> [    1.573768]  [<c15e5cfe>] ? dev_change_flags+0x13/0x3f
>> [    1.574768]  [<c1acfe6f>] ? ip_auto_config+0x160/0xcf8
>> [    1.574768]  [<c1393c86>] ? extract_entropy+0x45/0x71
>> [    1.575768]  [<c1059e35>] ? get_parent_ip+0xb/0x31
>> [    1.576768]  [<c16aa6b7>] ? sub_preempt_count+0x7c/0x89
>> [    1.576768]  [<c16a7d24>] ? _raw_spin_unlock+0x1c/0x27
>> [    1.577768]  [<c1629173>] ?
>> tcp_set_default_congestion_control+0x8c/0x95
>> [    1.578768]  [<c1001159>] ? do_one_initcall+0x71/0x114
>> [    1.578768]  [<c1acfd0f>] ? root_nfs_parse_addr+0x91/0x91
>> [    1.579767]  [<c1a9c7ab>] ? kernel_init+0xab/0x11d
>> [    1.580767]  [<c1a9c700>] ? start_kernel+0x301/0x301
>> [    1.581767]  [<c16acfb6>] ? kernel_thread_helper+0x6/0xd
>> [    1.582767] Rebooting in 1 seconds..
>>  # KVM session ended normally.
> 
> This needs fixing before I can apply the patch, right?

Yup. I think so.

-- 
Asias He

Re: [PATCH v2] kvm tools: Support multiple net devices

[Avi Kivity]

On 09/26/2011 05:54 PM, Pekka Enberg wrote:
> On Mon, 26 Sep 2011, Asias He wrote:
>> $ ./kvm run -n mode=tap
>>
>>
>> [    1.490781] registered taskstats version 1
>> [    1.492781] BUG: unable to handle kernel NULL pointer dereference at
>> 0000001c
>> [    1.493781] IP: [<c14f3236>] virtnet_poll+0x16e/0x408
>> [    1.493781] *pde = 00000000
>> [    1.493781] Oops: 0000 [#1] PREEMPT SMP
>> [    1.493781] Modules linked in:
>> [    1.493781]
>> [    1.493781] Pid: 1, comm: swapper Tainted: G        W   3.1.0-rc3+ 
>> #77
>> [    1.493781] EIP: 0060:[<c14f3236>] EFLAGS: 00010286 CPU: 1
>> [    1.493781] EIP is at virtnet_poll+0x16e/0x408
>> [    1.493781] EAX: 00001000 EBX: db4bb0c0 ECX: db7cd778 EDX: 00001000
>> [    1.493781] ESI: 00000000 EDI: db7cd6c0 EBP: db487fa8 ESP: db487f64
>> [    1.493781]  DS: 007b ES: 007b FS: 00d8 GS: 0000 SS: 0068
>> [    1.493781] Process swapper (pid: 1, ti=db486000 task=db450000
>> task.ti=db458000)
>> [    1.493781] Stack:
>> [    1.493781]  db487f98 19dfb000 db5e1400 00000080 c1b0df60 db6ff000
>> 00000000 00000010
>> [    1.493781]  00000080 dbcebfe0 db5e1414 db5e1000 ffffffff fffec005
>> db5e1414 db906dc0
>> [    1.493781]  c1a39a0c db487fd4 c15e4869 fffb71f7 db906dc8 00000080
>> 0000012c 00000000
>> [    1.493781] Call Trace:
>> [    1.493781]  [<c15e4869>] net_rx_action+0x8e/0x177
>> [    1.493781]  [<c1066128>] __do_softirq+0xa7/0x158
>> [    1.493781]  [<c1066081>] ? __tasklet_hi_schedule_first+0x2b/0x2b
>> [    1.493781] <IRQ>
>> [    1.493781]  [<c1065e82>] ? _local_bh_enable_ip.isra.9+0x65/0x86
>> [    1.493781]  [<c1065eab>] ? local_bh_enable_ip+0x8/0xa
>> [    1.493781]  [<c16a7a78>] ? _raw_spin_unlock_bh+0x18/0x1a
>> [    1.493781]  [<c15e59c5>] ? dev_set_rx_mode+0x22/0x26
>> [    1.493781]  [<c15e5a5f>] ? __dev_open+0x96/0xa6
>> [    1.493781]  [<c15e5c23>] ? __dev_change_flags+0x97/0x10e
>> [    1.493781]  [<c15e5cfe>] ? dev_change_flags+0x13/0x3f
>> [    1.493781]  [<c1acfe6f>] ? ip_auto_config+0x160/0xcf8
>> [    1.493781]  [<c1393c86>] ? extract_entropy+0x45/0x71
>> [    1.493781]  [<c1059e35>] ? get_parent_ip+0xb/0x31
>> [    1.493781]  [<c16aa6b7>] ? sub_preempt_count+0x7c/0x89
>> [    1.493781]  [<c16a7d24>] ? _raw_spin_unlock+0x1c/0x27
>> [    1.493781]  [<c1629173>] ? 
>> tcp_set_default_congestion_control+0x8c/0x95
>> [    1.493781]  [<c1001159>] ? do_one_initcall+0x71/0x114
>> [    1.493781]  [<c1acfd0f>] ? root_nfs_parse_addr+0x91/0x91
>> [    1.493781]  [<c1a9c7ab>] ? kernel_init+0xab/0x11d
>> [    1.493781]  [<c1a9c700>] ? start_kernel+0x301/0x301
>> [    1.493781]  [<c16acfb6>] ? kernel_thread_helper+0x6/0xd
>> [    1.493781] Code: 89 d8 e8 23 94 0e 00 8b 4d dc 89 c7 f3 a4 8b 55 dc
>> 8b 4d d8 29 55 f0 8b 75 e0 01 d1 eb 13 8d 45 f0 89 f2 50 89 d8 e8 ae f2
>> ff ff <8b> 76 1c 31 c9 58 83 7d f0 00 75 e7 85 f6 89 75 e0 0f 84 6e 02
>> [    1.493781] EIP: [<c14f3236>] virtnet_poll+0x16e/0x408 SS:ESP
>> 0068:db487f64
>> [    1.493781] CR2: 000000000000001c
>> [    1.549772] ---[ end trace 4eaa2a86a8e2da27 ]---
>> [    1.550772] Kernel panic - not syncing: Fatal exception in interrupt
>> [    1.551772] Pid: 1, comm: swapper Tainted: G      D W   3.1.0-rc3+ 
>> #77
>> [    1.553771] Call Trace:
>> [    1.553771]  [<c169ca33>] panic+0x58/0x156
>> [    1.554771]  [<c16a921a>] oops_end+0x8c/0x9b
>> [    1.555771]  [<c169c4e7>] no_context+0x116/0x120
>> [    1.555771]  [<c169c5e1>] __bad_area_nosemaphore+0xf0/0xf8
>> [    1.557771]  [<c169c5f6>] bad_area_nosemaphore+0xd/0x10
>> [    1.558771]  [<c16aa4b5>] do_page_fault+0x174/0x2fa
>> [    1.559770]  [<c107bad0>] ? sched_clock_local+0x10/0x14b
>> [    1.560770]  [<c15db33f>] ? __netdev_alloc_skb+0x17/0x34
>> [    1.561770]  [<c10e9b84>] ? __kmalloc_track_caller+0xb7/0xc7
>> [    1.563770]  [<c15db33f>] ? __netdev_alloc_skb+0x17/0x34
>> [    1.564770]  [<c16aa341>] ? spurious_fault+0xa8/0xa8
>> [    1.565770]  [<c16a89d6>] error_code+0x5a/0x60
>> [    1.566769]  [<c16aa341>] ? spurious_fault+0xa8/0xa8
>> [    1.567769]  [<c14f3236>] ? virtnet_poll+0x16e/0x408
>> [    1.567769]  [<c15e4869>] net_rx_action+0x8e/0x177
>> [    1.568769]  [<c1066128>] __do_softirq+0xa7/0x158
>> [    1.569769]  [<c1066081>] ? __tasklet_hi_schedule_first+0x2b/0x2b
>> [    1.569769] <IRQ>  [<c1065e82>] ? 
>> _local_bh_enable_ip.isra.9+0x65/0x86
>> [    1.570769]  [<c1065eab>] ? local_bh_enable_ip+0x8/0xa
>> [    1.571769]  [<c16a7a78>] ? _raw_spin_unlock_bh+0x18/0x1a
>> [    1.571769]  [<c15e59c5>] ? dev_set_rx_mode+0x22/0x26
>> [    1.572768]  [<c15e5a5f>] ? __dev_open+0x96/0xa6
>> [    1.573768]  [<c15e5c23>] ? __dev_change_flags+0x97/0x10e
>> [    1.573768]  [<c15e5cfe>] ? dev_change_flags+0x13/0x3f
>> [    1.574768]  [<c1acfe6f>] ? ip_auto_config+0x160/0xcf8
>> [    1.574768]  [<c1393c86>] ? extract_entropy+0x45/0x71
>> [    1.575768]  [<c1059e35>] ? get_parent_ip+0xb/0x31
>> [    1.576768]  [<c16aa6b7>] ? sub_preempt_count+0x7c/0x89
>> [    1.576768]  [<c16a7d24>] ? _raw_spin_unlock+0x1c/0x27
>> [    1.577768]  [<c1629173>] ? 
>> tcp_set_default_congestion_control+0x8c/0x95
>> [    1.578768]  [<c1001159>] ? do_one_initcall+0x71/0x114
>> [    1.578768]  [<c1acfd0f>] ? root_nfs_parse_addr+0x91/0x91
>> [    1.579767]  [<c1a9c7ab>] ? kernel_init+0xab/0x11d
>> [    1.580767]  [<c1a9c700>] ? start_kernel+0x301/0x301
>> [    1.581767]  [<c16acfb6>] ? kernel_thread_helper+0x6/0xd
>> [    1.582767] Rebooting in 1 seconds..
>>  # KVM session ended normally.
>
> This needs fixing before I can apply the patch, right?

Looks like a guest kernel bug, no?

-- 
error compiling committee.c: too many arguments to function

Re: [PATCH v2] kvm tools: Support multiple net devices

[Sasha Levin]

On Mon, 2011-09-26 at 19:21 +0300, Avi Kivity wrote:
> On 09/26/2011 05:54 PM, Pekka Enberg wrote:
> > On Mon, 26 Sep 2011, Asias He wrote:
> >> $ ./kvm run -n mode=tap
> >>
> >>
> >> [    1.490781] registered taskstats version 1
> >> [    1.492781] BUG: unable to handle kernel NULL pointer dereference at
> >> 0000001c
> >> [    1.493781] IP: [<c14f3236>] virtnet_poll+0x16e/0x408
> >> [    1.493781] *pde = 00000000
> >> [    1.493781] Oops: 0000 [#1] PREEMPT SMP
> >> [    1.493781] Modules linked in:
> >> [    1.493781]
> >> [    1.493781] Pid: 1, comm: swapper Tainted: G        W   3.1.0-rc3+ 
> >> #77
> >> [    1.493781] EIP: 0060:[<c14f3236>] EFLAGS: 00010286 CPU: 1
> >> [    1.493781] EIP is at virtnet_poll+0x16e/0x408
> >> [    1.493781] EAX: 00001000 EBX: db4bb0c0 ECX: db7cd778 EDX: 00001000
> >> [    1.493781] ESI: 00000000 EDI: db7cd6c0 EBP: db487fa8 ESP: db487f64
> >> [    1.493781]  DS: 007b ES: 007b FS: 00d8 GS: 0000 SS: 0068
> >> [    1.493781] Process swapper (pid: 1, ti=db486000 task=db450000
> >> task.ti=db458000)
> >> [    1.493781] Stack:
> >> [    1.493781]  db487f98 19dfb000 db5e1400 00000080 c1b0df60 db6ff000
> >> 00000000 00000010
> >> [    1.493781]  00000080 dbcebfe0 db5e1414 db5e1000 ffffffff fffec005
> >> db5e1414 db906dc0
> >> [    1.493781]  c1a39a0c db487fd4 c15e4869 fffb71f7 db906dc8 00000080
> >> 0000012c 00000000
> >> [    1.493781] Call Trace:
> >> [    1.493781]  [<c15e4869>] net_rx_action+0x8e/0x177
> >> [    1.493781]  [<c1066128>] __do_softirq+0xa7/0x158
> >> [    1.493781]  [<c1066081>] ? __tasklet_hi_schedule_first+0x2b/0x2b
> >> [    1.493781] <IRQ>
> >> [    1.493781]  [<c1065e82>] ? _local_bh_enable_ip.isra.9+0x65/0x86
> >> [    1.493781]  [<c1065eab>] ? local_bh_enable_ip+0x8/0xa
> >> [    1.493781]  [<c16a7a78>] ? _raw_spin_unlock_bh+0x18/0x1a
> >> [    1.493781]  [<c15e59c5>] ? dev_set_rx_mode+0x22/0x26
> >> [    1.493781]  [<c15e5a5f>] ? __dev_open+0x96/0xa6
> >> [    1.493781]  [<c15e5c23>] ? __dev_change_flags+0x97/0x10e
> >> [    1.493781]  [<c15e5cfe>] ? dev_change_flags+0x13/0x3f
> >> [    1.493781]  [<c1acfe6f>] ? ip_auto_config+0x160/0xcf8
> >> [    1.493781]  [<c1393c86>] ? extract_entropy+0x45/0x71
> >> [    1.493781]  [<c1059e35>] ? get_parent_ip+0xb/0x31
> >> [    1.493781]  [<c16aa6b7>] ? sub_preempt_count+0x7c/0x89
> >> [    1.493781]  [<c16a7d24>] ? _raw_spin_unlock+0x1c/0x27
> >> [    1.493781]  [<c1629173>] ? 
> >> tcp_set_default_congestion_control+0x8c/0x95
> >> [    1.493781]  [<c1001159>] ? do_one_initcall+0x71/0x114
> >> [    1.493781]  [<c1acfd0f>] ? root_nfs_parse_addr+0x91/0x91
> >> [    1.493781]  [<c1a9c7ab>] ? kernel_init+0xab/0x11d
> >> [    1.493781]  [<c1a9c700>] ? start_kernel+0x301/0x301
> >> [    1.493781]  [<c16acfb6>] ? kernel_thread_helper+0x6/0xd
> >> [    1.493781] Code: 89 d8 e8 23 94 0e 00 8b 4d dc 89 c7 f3 a4 8b 55 dc
> >> 8b 4d d8 29 55 f0 8b 75 e0 01 d1 eb 13 8d 45 f0 89 f2 50 89 d8 e8 ae f2
> >> ff ff <8b> 76 1c 31 c9 58 83 7d f0 00 75 e7 85 f6 89 75 e0 0f 84 6e 02
> >> [    1.493781] EIP: [<c14f3236>] virtnet_poll+0x16e/0x408 SS:ESP
> >> 0068:db487f64
> >> [    1.493781] CR2: 000000000000001c
> >> [    1.549772] ---[ end trace 4eaa2a86a8e2da27 ]---
> >> [    1.550772] Kernel panic - not syncing: Fatal exception in interrupt
> >> [    1.551772] Pid: 1, comm: swapper Tainted: G      D W   3.1.0-rc3+ 
> >> #77
> >> [    1.553771] Call Trace:
> >> [    1.553771]  [<c169ca33>] panic+0x58/0x156
> >> [    1.554771]  [<c16a921a>] oops_end+0x8c/0x9b
> >> [    1.555771]  [<c169c4e7>] no_context+0x116/0x120
> >> [    1.555771]  [<c169c5e1>] __bad_area_nosemaphore+0xf0/0xf8
> >> [    1.557771]  [<c169c5f6>] bad_area_nosemaphore+0xd/0x10
> >> [    1.558771]  [<c16aa4b5>] do_page_fault+0x174/0x2fa
> >> [    1.559770]  [<c107bad0>] ? sched_clock_local+0x10/0x14b
> >> [    1.560770]  [<c15db33f>] ? __netdev_alloc_skb+0x17/0x34
> >> [    1.561770]  [<c10e9b84>] ? __kmalloc_track_caller+0xb7/0xc7
> >> [    1.563770]  [<c15db33f>] ? __netdev_alloc_skb+0x17/0x34
> >> [    1.564770]  [<c16aa341>] ? spurious_fault+0xa8/0xa8
> >> [    1.565770]  [<c16a89d6>] error_code+0x5a/0x60
> >> [    1.566769]  [<c16aa341>] ? spurious_fault+0xa8/0xa8
> >> [    1.567769]  [<c14f3236>] ? virtnet_poll+0x16e/0x408
> >> [    1.567769]  [<c15e4869>] net_rx_action+0x8e/0x177
> >> [    1.568769]  [<c1066128>] __do_softirq+0xa7/0x158
> >> [    1.569769]  [<c1066081>] ? __tasklet_hi_schedule_first+0x2b/0x2b
> >> [    1.569769] <IRQ>  [<c1065e82>] ? 
> >> _local_bh_enable_ip.isra.9+0x65/0x86
> >> [    1.570769]  [<c1065eab>] ? local_bh_enable_ip+0x8/0xa
> >> [    1.571769]  [<c16a7a78>] ? _raw_spin_unlock_bh+0x18/0x1a
> >> [    1.571769]  [<c15e59c5>] ? dev_set_rx_mode+0x22/0x26
> >> [    1.572768]  [<c15e5a5f>] ? __dev_open+0x96/0xa6
> >> [    1.573768]  [<c15e5c23>] ? __dev_change_flags+0x97/0x10e
> >> [    1.573768]  [<c15e5cfe>] ? dev_change_flags+0x13/0x3f
> >> [    1.574768]  [<c1acfe6f>] ? ip_auto_config+0x160/0xcf8
> >> [    1.574768]  [<c1393c86>] ? extract_entropy+0x45/0x71
> >> [    1.575768]  [<c1059e35>] ? get_parent_ip+0xb/0x31
> >> [    1.576768]  [<c16aa6b7>] ? sub_preempt_count+0x7c/0x89
> >> [    1.576768]  [<c16a7d24>] ? _raw_spin_unlock+0x1c/0x27
> >> [    1.577768]  [<c1629173>] ? 
> >> tcp_set_default_congestion_control+0x8c/0x95
> >> [    1.578768]  [<c1001159>] ? do_one_initcall+0x71/0x114
> >> [    1.578768]  [<c1acfd0f>] ? root_nfs_parse_addr+0x91/0x91
> >> [    1.579767]  [<c1a9c7ab>] ? kernel_init+0xab/0x11d
> >> [    1.580767]  [<c1a9c700>] ? start_kernel+0x301/0x301
> >> [    1.581767]  [<c16acfb6>] ? kernel_thread_helper+0x6/0xd
> >> [    1.582767] Rebooting in 1 seconds..
> >>  # KVM session ended normally.
> >
> > This needs fixing before I can apply the patch, right?
> 
> Looks like a guest kernel bug, no?

It's a kernel bug and should be fixed there, but it's caused by us not
passing sane values to virtio-net, which we can fix on our side as well.

So my plan is to prevent triggering it from within kvm tools while
working on a kernel patch.

-- 

Sasha.

Re: [PATCH v2] kvm tools: Support multiple net devices

[Pekka Enberg]

On Mon, Sep 26, 2011 at 7:37 PM, Sasha Levin <levinsasha928@gmail.com> wrote:
>> > This needs fixing before I can apply the patch, right?
>>
>> Looks like a guest kernel bug, no?
>
> It's a kernel bug and should be fixed there, but it's caused by us not
> passing sane values to virtio-net, which we can fix on our side as well.
>
> So my plan is to prevent triggering it from within kvm tools while
> working on a kernel patch.

Yup, that's what I meant as well. Even if it is a kernel issue, we
should do enough validation on our side not to trigger it so easily.

                        Pekka