/* * 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, Jack Lange * Copyright (c) 2008, The V3VEE Project * All rights reserved. * * Author: Jack Lange * * This is free software. You are permitted to use, * redistribute, and modify it as specified in the file "V3VEE_LICENSE". */ #include #include #include #include #include #include #include #ifndef CONFIG_DEBUG_VNET_BRG #undef PrintDebug #define PrintDebug(fmt, args...) #endif #define QUEUE_SIZE 128 #define NUM_QUEUES 3 struct vnet_config { uint32_t num_devs; uint32_t num_routes; } __attribute__((packed)); #define CTRL_QUEUE 0 #define RECV_QUEUE 2 #define XMIT_QUEUE 1 struct virtio_vnet_state { struct v3_vm_info *vm; struct vnet_config vnet_cfg; struct virtio_config virtio_cfg; struct vm_device * pci_bus; struct pci_device * pci_dev; struct virtio_queue queue[NUM_QUEUES]; struct virtio_queue * cur_queue; int io_range_size; v3_lock_t lock; }; #define VNET_GET_ROUTES 10 #define VNET_ADD_ROUTE 11 #define VNET_DEL_ROUTE 12 #define VNET_GET_LINKS 20 #define VNET_ADD_LINK 21 #define VNET_DEL_LINK 22 // structure of the vnet command header struct vnet_ctrl_hdr { uint8_t cmd_type; uint32_t num_cmds; } __attribute__((packed)); static int virtio_reset(struct virtio_vnet_state * vnet_brg) { memset(vnet_brg->queue, 0, sizeof(struct virtio_queue) * 2); vnet_brg->cur_queue = &(vnet_brg->queue[0]); vnet_brg->virtio_cfg.status = 0; vnet_brg->virtio_cfg.pci_isr = 0; vnet_brg->queue[0].queue_size = QUEUE_SIZE; vnet_brg->queue[1].queue_size = QUEUE_SIZE; vnet_brg->queue[2].queue_size = QUEUE_SIZE; memset(&(vnet_brg->vnet_cfg), 0, sizeof(struct vnet_config)); v3_lock_init(&(vnet_brg->lock)); return 0; } static int get_desc_count(struct virtio_queue * q, int index) { struct vring_desc * tmp_desc = &(q->desc[index]); int cnt = 1; while (tmp_desc->flags & VIRTIO_NEXT_FLAG) { tmp_desc = &(q->desc[tmp_desc->next]); cnt++; } return cnt; } static int handle_cmd_kick(struct guest_info * core, struct virtio_vnet_state * vnet_state) { struct virtio_queue * q = &(vnet_state->queue[0]); PrintDebug("VNET Bridge: Handling command queue\n"); while (q->cur_avail_idx < q->avail->index) { struct vring_desc * hdr_desc = NULL; struct vring_desc * buf_desc = NULL; struct vring_desc * status_desc = NULL; uint16_t desc_idx = q->avail->ring[q->cur_avail_idx % QUEUE_SIZE]; uint16_t desc_cnt = get_desc_count(q, desc_idx); struct vnet_ctrl_hdr * hdr = NULL; int i; int xfer_len = 0; uint8_t * status_ptr = NULL; uint8_t status = 0; PrintDebug("Descriptor Count=%d, index=%d, desc_idx=%d\n", desc_cnt, q->cur_avail_idx % QUEUE_SIZE, desc_idx); if (desc_cnt < 3) { PrintError("VNET Bridge cmd must include at least 3 descriptors (cnt=%d)\n", desc_cnt); return -1; } hdr_desc = &(q->desc[desc_idx]); if (guest_pa_to_host_va(core, hdr_desc->addr_gpa, (addr_t *)&hdr) == -1) { PrintError("Could not translate VirtioVNET header address\n"); return -1; } desc_idx = hdr_desc->next; if (hdr->cmd_type == VNET_ADD_ROUTE) { for (i = 0; i < hdr->num_cmds; i++) { uint8_t tmp_status = 0; struct v3_vnet_route * route = NULL; buf_desc = &(q->desc[desc_idx]); if (guest_pa_to_host_va(core, buf_desc->addr_gpa, (addr_t *)&(route)) == -1) { PrintError("Could not translate route address\n"); return -1; } // add route PrintDebug("Adding VNET Route\n"); tmp_status = v3_vnet_add_route(*route); PrintDebug("VNET Route Added\n"); if (tmp_status != 0) { PrintError("Error adding VNET ROUTE\n"); status = tmp_status; } xfer_len += buf_desc->length; desc_idx = buf_desc->next; } } status_desc = &(q->desc[desc_idx]); if (guest_pa_to_host_va(core, status_desc->addr_gpa, (addr_t *)&status_ptr) == -1) { PrintError("VirtioVNET Error could not translate status address\n"); return -1; } xfer_len += status_desc->length; *status_ptr = status; PrintDebug("Transferred %d bytes (xfer_len)\n", xfer_len); q->used->ring[q->used->index % QUEUE_SIZE].id = q->avail->ring[q->cur_avail_idx % QUEUE_SIZE]; q->used->ring[q->used->index % QUEUE_SIZE].length = xfer_len; // set to total inbound xfer length q->used->index++; q->cur_avail_idx++; } if (!(q->avail->flags & VIRTIO_NO_IRQ_FLAG)) { PrintDebug("Raising IRQ %d\n", vnet_state->pci_dev->config_header.intr_line); v3_pci_raise_irq(vnet_state->pci_bus, 0, vnet_state->pci_dev); vnet_state->virtio_cfg.pci_isr = 1; } return 0; } static int copy_data_to_desc(struct guest_info *core, struct vring_desc * desc, uchar_t * buf, uint_t buf_len) { uint32_t len; uint8_t * desc_buf = NULL; if (guest_pa_to_host_va(core, desc->addr_gpa, (addr_t *)&(desc_buf)) == -1) { PrintError("Could not translate buffer address\n"); return -1; } len = (desc->length < buf_len)?desc->length:buf_len; memcpy(desc_buf, buf, len); return len; } static int vnet_brg_rx(struct v3_vnet_pkt *pkt, void * private_data) { struct virtio_vnet_state * vnet_brg = (struct virtio_vnet_state *)private_data; struct virtio_queue * q = &(vnet_brg->queue[RECV_QUEUE]); char *buf = (char *)pkt; int ret_val; int pkt_head_len = ((addr_t)pkt->data) - ((addr_t)pkt); uint32_t data_len = sizeof(struct v3_vnet_pkt); //TODO: should not transfer the whole struct, only the data size unsigned long flags; flags = v3_lock_irqsave(vnet_brg->lock); PrintDebug("VNET Bridge: RX: pkt sent to guest size: %d\n, pkt_header_len: %d\n", data_len, pkt_head_len); v3_hexdump(buf, 100, NULL, 0); if (q->ring_avail_addr == 0) { PrintError("Queue is not set\n"); ret_val = -1; goto exit; } if (q->last_avail_idx > q->avail->index) q->idx_overflow = true; q->last_avail_idx = q->avail->index; if (q->cur_avail_idx < q->avail->index || (q->idx_overflow && q->cur_avail_idx < q->avail->index+65536)){ uint16_t buf_idx = q->avail->ring[q->cur_avail_idx % q->queue_size]; uint32_t len = 0; uint32_t offset = 0; struct vring_desc * buf_desc = NULL; buf_desc = &(q->desc[buf_idx]); PrintDebug("VNET Bridge RX: buffer desc len: %d\n", buf_desc->length); len = copy_data_to_desc(&(vnet_brg->vm->cores[0]), buf_desc, buf + offset, data_len - offset); if (len == -1) { ret_val = -1; goto exit; } offset += len; buf_desc->length = len; if (offset >= data_len) { buf_desc->flags &= ~VIRTIO_NEXT_FLAG; }else { PrintDebug("VNET Bridge: RX: Pkt not fit into one desc buf\n"); } q->used->ring[q->used->index % q->queue_size].id = q->avail->ring[q->cur_avail_idx % q->queue_size]; q->used->ring[q->used->index % q->queue_size].length = data_len; // This should be the total length of data sent to guest (header+pkt_data) q->used->index++; int last_idx = q->cur_avail_idx; q->cur_avail_idx++; if (q->cur_avail_idx < last_idx) q->idx_overflow = false; } else { PrintDebug("VNET Bridge: Handle RX: Fails to send, no available buffer: current_idx:%d, availIdx: %d\n", q->cur_avail_idx, q->avail->index); } if (!(q->avail->flags & VIRTIO_NO_IRQ_FLAG)) { v3_pci_raise_irq(vnet_brg->pci_bus, 0, vnet_brg->pci_dev); vnet_brg->virtio_cfg.pci_isr = 0x1; PrintDebug("Raising IRQ %d\n", vnet_brg->pci_dev->config_header.intr_line); } ret_val = 0; exit: v3_unlock_irqrestore(vnet_brg->lock, flags); return ret_val; } static int vnet_send(struct v3_vnet_pkt *pkt, int pkt_len, void * private_data){ struct guest_info *core = (struct guest_info *)private_data; #ifdef CONFIG_DEBUG_VNET_BRG { PrintDebug("VNET Bridge: send pkt size: %d\n", pkt->size); v3_hexdump(pkt->data,pkt->size, NULL, 0); } #endif #ifdef CONFIG_VNET_PROFILE uint64_t start, end; rdtscll(start); core->vnet_times.time_copy_from_guest = start - core->vnet_times.virtio_handle_start; #endif pkt->src_type = LINK_EDGE; return v3_vnet_send_pkt(pkt, (void *)core); } static int pkt_tx(struct guest_info *core, struct virtio_vnet_state * vnet_brg, struct vring_desc * buf_desc) { uint8_t * buf = NULL; uint32_t len = buf_desc->length; struct v3_vnet_pkt *pkt; if (guest_pa_to_host_va(core, buf_desc->addr_gpa, (addr_t *)&(buf)) == -1) { PrintError("Could not translate buffer address\n"); return -1; } pkt = (struct v3_vnet_pkt *)buf; if (vnet_send(pkt, len, (void *)core) == -1) { return -1; } return 0; } static int handle_pkt_tx(struct guest_info *core, struct virtio_vnet_state * vnet_brg) { struct virtio_queue * q = &(vnet_brg->queue[XMIT_QUEUE]); if (q->avail->index < q->last_avail_idx) { q->idx_overflow = true; } q->last_avail_idx = q->avail->index; PrintDebug("VNET Bridge Device: Handle TX\n"); while (q->cur_avail_idx < q->avail->index || (q->idx_overflow && q->cur_avail_idx < (q->avail->index + 65536))) { struct vring_desc * buf_desc = NULL; uint16_t desc_idx = q->avail->ring[q->cur_avail_idx % q->queue_size]; buf_desc = &(q->desc[desc_idx]); PrintDebug("VNET Bridge: Handle TX buf_len: %d\n", buf_desc->length); if (pkt_tx(core, vnet_brg, buf_desc) == -1) { PrintError("Error handling nic operation\n"); return -1; } q->used->ring[q->used->index % q->queue_size].id = q->avail->ring[q->cur_avail_idx % q->queue_size]; q->used->ring[q->used->index % q->queue_size].length = buf_desc->length; // What do we set this to???? q->used->index++; int last_idx = q->cur_avail_idx; q->cur_avail_idx ++; if (q->cur_avail_idx < last_idx) { q->idx_overflow = false; } } if (!(q->avail->flags & VIRTIO_NO_IRQ_FLAG)) { v3_pci_raise_irq(vnet_brg->pci_bus, 0, vnet_brg->pci_dev); vnet_brg->virtio_cfg.pci_isr = 0x1; } #ifdef CONFIG_VNET_PROFILE uint64_t time; rdtscll(time); core->vnet_times.total_handle_time = time - core->vnet_times.virtio_handle_start; core->vnet_times.print = true; #endif return 0; } static int virtio_io_write(struct guest_info * core, uint16_t port, void * src, uint_t length, void * private_data) { struct virtio_vnet_state * vnet_brg = (struct virtio_vnet_state *)private_data; int port_idx = port % vnet_brg->io_range_size; PrintDebug("VNET Bridge: VIRTIO VNET Write for port %d len=%d, value=%x\n", port, length, *(uint32_t *)src); PrintDebug("VNET Bridge: port idx=%d\n", port_idx); switch (port_idx) { case GUEST_FEATURES_PORT: if (length != 4) { PrintError("Illegal write length for guest features\n"); return -1; } vnet_brg->virtio_cfg.guest_features = *(uint32_t *)src; break; case VRING_PG_NUM_PORT: if (length == 4) { addr_t pfn = *(uint32_t *)src; addr_t page_addr = (pfn << VIRTIO_PAGE_SHIFT); vnet_brg->cur_queue->pfn = pfn; vnet_brg->cur_queue->ring_desc_addr = page_addr ; vnet_brg->cur_queue->ring_avail_addr = page_addr + (QUEUE_SIZE * sizeof(struct vring_desc)); vnet_brg->cur_queue->ring_used_addr = ( vnet_brg->cur_queue->ring_avail_addr + \ sizeof(struct vring_avail) + \ (QUEUE_SIZE * sizeof(uint16_t))); // round up to next page boundary. vnet_brg->cur_queue->ring_used_addr = (vnet_brg->cur_queue->ring_used_addr + 0xfff) & ~0xfff; if (guest_pa_to_host_va(core, vnet_brg->cur_queue->ring_desc_addr, (addr_t *)&(vnet_brg->cur_queue->desc)) == -1) { PrintError("Could not translate ring descriptor address\n"); return -1; } if (guest_pa_to_host_va(core, vnet_brg->cur_queue->ring_avail_addr, (addr_t *)&(vnet_brg->cur_queue->avail)) == -1) { PrintError("Could not translate ring available address\n"); return -1; } if (guest_pa_to_host_va(core, vnet_brg->cur_queue->ring_used_addr, (addr_t *)&(vnet_brg->cur_queue->used)) == -1) { PrintError("Could not translate ring used address\n"); return -1; } PrintDebug("VNET Bridge: RingDesc_addr=%p, Avail_addr=%p, Used_addr=%p\n", (void *)(vnet_brg->cur_queue->ring_desc_addr), (void *)(vnet_brg->cur_queue->ring_avail_addr), (void *)(vnet_brg->cur_queue->ring_used_addr)); PrintDebug("VNET Bridge: RingDesc=%p, Avail=%p, Used=%p\n", vnet_brg->cur_queue->desc, vnet_brg->cur_queue->avail, vnet_brg->cur_queue->used); } else { PrintError("Illegal write length for page frame number\n"); return -1; } break; case VRING_Q_SEL_PORT: vnet_brg->virtio_cfg.vring_queue_selector = *(uint16_t *)src; if (vnet_brg->virtio_cfg.vring_queue_selector > NUM_QUEUES) { PrintError("VNET Bridge device has no qeueues. Selected %d\n", vnet_brg->virtio_cfg.vring_queue_selector); return -1; } vnet_brg->cur_queue = &(vnet_brg->queue[vnet_brg->virtio_cfg.vring_queue_selector]); break; case VRING_Q_NOTIFY_PORT: { uint16_t queue_idx = *(uint16_t *)src; PrintDebug("VNET Bridge: Handling Kick\n"); if (queue_idx == 0) { if (handle_cmd_kick(core, vnet_brg) == -1) { PrintError("Could not handle Virtio VNET Control command\n"); return -1; } } else if (queue_idx == 1) { if (handle_pkt_tx(core, vnet_brg) == -1){ PrintError("Could not handle Virtio VNET TX\n"); return -1; } } else if (queue_idx == 2) { PrintDebug("VNET Bridge: receive kick on RX Queue\n"); } else { PrintError("VNET Bridge: Kick on invalid queue (%d)\n", queue_idx); return -1; } break; } case VIRTIO_STATUS_PORT: vnet_brg->virtio_cfg.status = *(uint8_t *)src; if (vnet_brg->virtio_cfg.status == 0) { PrintDebug("VNET Bridge: Resetting device\n"); virtio_reset(vnet_brg); } break; case VIRTIO_ISR_PORT: vnet_brg->virtio_cfg.pci_isr = *(uint8_t *)src; break; default: return -1; break; } return length; } static int virtio_io_read(struct guest_info * core, uint16_t port, void * dst, uint_t length, void * private_data) { struct virtio_vnet_state * vnet_state = (struct virtio_vnet_state *)private_data; int port_idx = port % vnet_state->io_range_size; /* PrintDebug("VirtioVNET: VIRTIO SYMBIOTIC Read for port %d (index =%d), length=%d\n", port, port_idx, length); */ switch (port_idx) { case HOST_FEATURES_PORT: if (length != 4) { PrintError("Illegal read length for host features\n"); return -1; } *(uint32_t *)dst = vnet_state->virtio_cfg.host_features; break; case VRING_PG_NUM_PORT: if (length != 4) { PrintError("Illegal read length for page frame number\n"); return -1; } *(uint32_t *)dst = vnet_state->cur_queue->pfn; break; case VRING_SIZE_PORT: if (length != 2) { PrintError("Illegal read length for vring size\n"); return -1; } *(uint16_t *)dst = vnet_state->cur_queue->queue_size; break; case VIRTIO_STATUS_PORT: if (length != 1) { PrintError("Illegal read length for status\n"); return -1; } *(uint8_t *)dst = vnet_state->virtio_cfg.status; break; case VIRTIO_ISR_PORT: *(uint8_t *)dst = vnet_state->virtio_cfg.pci_isr; vnet_state->virtio_cfg.pci_isr = 0; v3_pci_lower_irq(vnet_state->pci_bus, 0, vnet_state->pci_dev); break; default: if ( (port_idx >= sizeof(struct virtio_config)) && (port_idx < (sizeof(struct virtio_config) + sizeof(struct vnet_config))) ) { int cfg_offset = port_idx - sizeof(struct virtio_config); uint8_t * cfg_ptr = (uint8_t *)&(vnet_state->vnet_cfg); memcpy(dst, cfg_ptr + cfg_offset, length); } else { PrintError("Read of Unhandled Virtio Read\n"); return -1; } break; } return length; } static int vnet_brg_input(struct v3_vm_info * vm, struct v3_vnet_pkt * pkt, void * private_data){ return vnet_brg_rx(pkt, private_data); } static struct v3_device_ops dev_ops = { .free = NULL, .reset = NULL, .start = NULL, .stop = NULL, }; static int vnet_brg_init(struct v3_vm_info * vm, v3_cfg_tree_t * cfg) { struct vm_device * pci_bus = v3_find_dev(vm, v3_cfg_val(cfg, "bus")); struct virtio_vnet_state * vbrg_state = NULL; struct pci_device * pci_dev = NULL; char * name = v3_cfg_val(cfg, "name"); PrintDebug("VNET Bridge: Initializing VNET Bridge Control device: %s\n", name); if (pci_bus == NULL) { PrintError("VNET Bridge device require a PCI Bus"); return -1; } vbrg_state = (struct virtio_vnet_state *)V3_Malloc(sizeof(struct virtio_vnet_state)); memset(vbrg_state, 0, sizeof(struct virtio_vnet_state)); vbrg_state->vm = vm; struct vm_device * dev = v3_allocate_device(name, &dev_ops, vbrg_state); if (v3_attach_device(vm, dev) == -1) { PrintError("Could not attach device %s\n", name); return -1; } // PCI initialization { struct v3_pci_bar bars[6]; int num_ports = sizeof(struct virtio_config) + sizeof(struct vnet_config); int tmp_ports = num_ports; int i; // This gets the number of ports, rounded up to a power of 2 vbrg_state->io_range_size = 1; // must be a power of 2 while (tmp_ports > 0) { tmp_ports >>= 1; vbrg_state->io_range_size <<= 1; } // this is to account for any low order bits being set in num_ports // if there are none, then num_ports was already a power of 2 so we shift right to reset it if ((num_ports & ((vbrg_state->io_range_size >> 1) - 1)) == 0) { vbrg_state->io_range_size >>= 1; } for (i = 0; i < 6; i++) { bars[i].type = PCI_BAR_NONE; } bars[0].type = PCI_BAR_IO; bars[0].default_base_port = -1; bars[0].num_ports = vbrg_state->io_range_size; bars[0].io_read = virtio_io_read; bars[0].io_write = virtio_io_write; bars[0].private_data = vbrg_state; pci_dev = v3_pci_register_device(pci_bus, PCI_STD_DEVICE, 0, PCI_AUTO_DEV_NUM, 0, "LNX_VIRTIO_VNET", bars, NULL, NULL, NULL, vbrg_state); if (!pci_dev) { PrintError("Could not register PCI Device\n"); return -1; } pci_dev->config_header.vendor_id = VIRTIO_VENDOR_ID; pci_dev->config_header.subsystem_vendor_id = VIRTIO_SUBVENDOR_ID; pci_dev->config_header.device_id = VIRTIO_VNET_DEV_ID; pci_dev->config_header.class = PCI_CLASS_MEMORY; pci_dev->config_header.subclass = PCI_MEM_SUBCLASS_RAM; pci_dev->config_header.subsystem_id = VIRTIO_VNET_SUBDEVICE_ID; pci_dev->config_header.intr_pin = 1; pci_dev->config_header.max_latency = 1; // ?? (qemu does it...) vbrg_state->pci_dev = pci_dev; vbrg_state->pci_bus = pci_bus; } virtio_reset(vbrg_state); v3_vnet_add_bridge(vm, vnet_brg_input, (void *)vbrg_state); //for temporary hack #if 1 { uchar_t dstmac[6] = {0xff,0xff,0xff,0xff,0xff,0xff}; uchar_t zeromac[6] = {0,0,0,0,0,0}; struct v3_vnet_route route; route.dst_id = 0; route.dst_type = LINK_EDGE; route.src_id = -1; route.src_type = LINK_ANY; memcpy(route.dst_mac, dstmac, 6); route.dst_mac_qual = MAC_NONE; memcpy(route.src_mac, zeromac, 6); route.src_mac_qual = MAC_ANY; v3_vnet_add_route(route); } #endif return 0; } device_register("LNX_VNET_BRG", vnet_brg_init)