Palacios Public Git Repository

To checkout Palacios execute

  git clone http://v3vee.org/palacios/palacios.web/palacios.git
This will give you the master branch. You probably want the devel branch or one of the release branches. To switch to the devel branch, simply execute
  cd palacios
  git checkout --track -b devel origin/devel
The other branches are similar.


Added backend device for Virtio NIC connected with VNET
Lei Xia [Tue, 16 Feb 2010 18:08:17 +0000 (12:08 -0600)]
palacios/src/devices/vnet_nic.c [new file with mode: 0644]

diff --git a/palacios/src/devices/vnet_nic.c b/palacios/src/devices/vnet_nic.c
new file mode 100644 (file)
index 0000000..ae1238f
--- /dev/null
@@ -0,0 +1,180 @@
+/* 
+ * This file is part of the Palacios Virtual Machine Monitor developed
+ * by the V3VEE Project with funding from the United States National 
+ * Science Foundation and the Department of Energy.  
+ *
+ * The V3VEE Project is a joint project between Northwestern University
+ * and the University of New Mexico.  You can find out more at 
+ * http://www.v3vee.org
+ *
+ * Copyright (c) 2008, Lei Xia <lxia@northwestern.edu>
+ * Copyright (c) 2008, The V3VEE Project <http://www.v3vee.org> 
+ * All rights reserved.
+ *
+ * Author: Lei Xia <lxia@northwestern.edu>
+ *              
+ *
+ * This is free software.  You are permitted to use,
+ * redistribute, and modify it as specified in the file "V3VEE_LICENSE".
+ */
+//backend device for Virtio NIC
+
+#include <palacios/vmm_vnet.h>
+#include <palacios/vmm.h>
+#include <palacios/vmm_dev_mgr.h>
+#include <devices/lnx_virtio_pci.h>
+#include <palacios/vm_guest_mem.h>
+#include <devices/pci.h>
+
+struct vnet_nic_state {
+       char mac[6];
+
+
+};
+
+//used when virtio_nic get a packet from guest and send it to the backend
+// send packet to all of the virtio nic devices other than the sender
+static int send(uint8_t * buf, uint32_t len, void * private_data, struct vm_device *dest_dev){
+    PrintDebug("Virito NIC: In vnet_send: guest net state %p\n", private_data);
+
+    v3_vnet_send_rawpkt(buf, len, private_data);
+    return 0;
+}
+
+static int receive(uint8_t * buf, uint32_t count, void * private_data, struct vm_device *src_dev){
+
+    return 0;
+}
+
+
+static struct v3_dev_net_ops net_ops = {
+    .send = send, 
+    .receive = receive,
+};
+
+
+#if 0
+static int input(struct v3_vm_info *info, uchar_t * buf, uint_t len, void * private_data){
+    PrintDebug("Virito NIC: In virtio_input: guest net state %p\n", private_data);
+
+    return __virtio_dev_send(buf, len, private_data);
+}
+
+
+//register a virtio device to the vnet as backend
+void register_to_vnet(struct vm_device  *dev, 
+                                               char *dev_name,
+                                               uchar_t mac[6]){
+    struct virtio_net_state * net_state;
+    struct virtio_dev_state *virtio_state =  (struct virtio_dev_state *)dev->private_data;
+    uchar_t tapmac[6] = {0x00,0x02,0x55,0x67,0x42,0x39}; //for Intel-VT test HW
+    //uchar_t tapmac[6] = {0x6e,0xa8,0x75,0xf4,0x82,0x95};
+    uchar_t dstmac[6] = {0xff,0xff,0xff,0xff,0xff,0xff};
+    uchar_t zeromac[6] = {0,0,0,0,0,0};
+
+
+    net_state  = (struct virtio_net_state *)V3_Malloc(sizeof(struct virtio_net_state));
+    memset(net_state, 0, sizeof(struct virtio_net_state));
+    net_state->net_ops = (struct v3_dev_net_ops *)V3_Malloc(sizeof(struct v3_dev_net_ops));
+    net_state->net_ops->send = &vnet_send;
+    net_state->net_ops->receive = &vnet_receive;
+    net_state->dev = dev;
+
+    register_dev(virtio_state, net_state);
+
+    PrintDebug("Virtio NIC register Device %s: queue size: %d, %d\n", dev->name,
+              net_state->rx_vq.queue_size, net_state->tx_vq.queue_size);
+    PrintDebug("Virtio NIC: connect virtio device %s, state %p, to vnet\n", dev->name, net_state);
+       
+    int idx = vnet_register_device(dev, dev_name, mac, &virtio_input, net_state);
+    //vnet_add_route_entry(zeromac, dstmac, MAC_ANY, MAC_NONE, idx, LINK_INTERFACE, -1, LINK_INTERFACE);
+    if (!strcmp(dev_name, "net_virtiodom0")){
+       vnet_add_route_entry(zeromac, tapmac, MAC_ANY, MAC_NONE, idx, LINK_INTERFACE, -1, LINK_INTERFACE);
+       vnet_add_route_entry(zeromac, dstmac, MAC_ANY, MAC_NONE, idx, LINK_INTERFACE, -1, LINK_INTERFACE);
+    }
+    if (!strcmp(dev_name, "net_virtio"))
+       vnet_add_route_entry(tapmac, zeromac, MAC_NONE, MAC_ANY, idx, LINK_INTERFACE, -1, LINK_INTERFACE);
+
+
+    v3_vnet_add_node(dev_name, mac, input, priv_data);
+    struct v3_vnet_route route;
+ //add default route
+    memset(&route, 0, sizeof(struct v3_vnet_route));
+    memcpy(&route.dest_mac, mac, 6);
+    route.src_mac_qual = MAC_ANY;
+    route.dest_mac_qual = MAC_NONE;
+    route.link_idx = idx;
+    route.link_type = LINK_EDGE;
+    route.src_link_idx = -1;
+    route.src_type = LINK_ANY;
+    v3_vnet_add_route(&route);
+
+    char mac
+    memset(&route, 0, sizeof(struct v3_vnet_route));
+    memcpy(&route.dest_mac, mac, 6);
+    route.src_mac_qual = MAC_ANY;
+    route.dest_mac_qual = MAC_NONE;
+    route.link_idx = idx;
+    route.link_type = LINK_EDGE;
+    route.src_link_idx = -1;
+    route.src_type = LINK_ANY;
+    v3_vnet_add_route(&route);
+    
+}
+
+
+/*
+  *called in svm/vmx handler
+  *iteative handled the unsent packet in incoming packet queues for
+  *all virtio nic devices in this guest
+  */
+int v3_vnetnic_pktprocess(struct guest_info * info)
+{
+    return 0;
+}
+
+#endif
+
+static int vnet_nic_free(struct vm_device * dev) {
+    return 0;
+}
+
+static struct v3_device_ops dev_ops = {
+    .free = vnet_nic_free,
+    .reset = NULL,
+    .start = NULL,
+    .stop = NULL,
+};
+
+static int vnet_nic_init(struct v3_vm_info * vm, v3_cfg_tree_t * cfg) {
+    struct vnet_nic_state * vnetnic = NULL;
+    char * name = v3_cfg_val(cfg, "name");
+
+    v3_cfg_tree_t * frontend_cfg = v3_cfg_subtree(cfg, "frontend");
+
+    vnetnic = (struct vnet_nic_state *)V3_Malloc(sizeof(struct vnet_nic_state));
+    memset(vnetnic, 0, sizeof(struct vnet_nic_state));
+
+
+    PrintDebug("Registering vnet_nic device at\n");
+
+    struct vm_device * dev = v3_allocate_device(name, &dev_ops, vnetnic);
+
+    if (v3_attach_device(vm, dev) == -1) {
+       PrintError("Could not attach device %s\n", name);
+       return -1;
+    }
+
+    if (v3_dev_connect_net(vm, v3_cfg_val(frontend_cfg, "tag"), 
+                          &net_ops, frontend_cfg, vnetnic) == -1) {
+       PrintError("Could not connect %s to frontend %s\n", 
+                  name, v3_cfg_val(frontend_cfg, "tag"));
+       return -1;
+    }
+    
+
+    return 0;
+}
+
+device_register("VNET_NIC", vnet_nic_init)