From 83e942e0c7634121243478fa2f4e3459b33d4a67 Mon Sep 17 00:00:00 2001 From: Dor Laor <[EMAIL PROTECTED]> Date: Wed, 19 Dec 2007 23:43:24 +0200 Subject: [PATCH] virtio network device
This patch implements the backend support for the virtio network device. The device is optimized for virtualized environments by limiting the number of guest=>host transitions per-packet. In the best case, the number of transitions per-packet is < 1. With some further optimizations, I have been able to obtain 1.5gbit/sec host=>guest with this driver (compared to the 90mbit/sec from the rtl8139 card). This requires additional patches not present in this series. Signed-off-by: Anthony Liguori <[EMAIL PROTECTED]> Cc: Rusty Russell <[EMAIL PROTECTED]> Cc: Avi Kivity <[EMAIL PROTECTED]> Cc: Dor Laor <[EMAIL PROTECTED]> --- qemu/Makefile.target | 2 +- qemu/hw/pc.h | 5 ++ qemu/hw/pci.c | 5 +- qemu/hw/virtio-net.c | 178 ++++++++++++++++++++++++++++++++++++++++++++++++++ 4 files changed, 188 insertions(+), 2 deletions(-) create mode 100644 qemu/hw/virtio-net.c diff --git a/qemu/Makefile.target b/qemu/Makefile.target index 8b5853b..17ff6f2 100644 --- a/qemu/Makefile.target +++ b/qemu/Makefile.target @@ -464,7 +464,7 @@ VL_OBJS += rtl8139.o VL_OBJS+= hypercall.o # virtio devices -VL_OBJS += virtio.o +VL_OBJS += virtio.o virtio-net.o ifeq ($(TARGET_BASE_ARCH), i386) # Hardware support diff --git a/qemu/hw/pc.h b/qemu/hw/pc.h index beb711c..ce1a1f3 100644 --- a/qemu/hw/pc.h +++ b/qemu/hw/pc.h @@ -142,4 +142,9 @@ void pci_piix4_ide_init(PCIBus *bus, BlockDriverState **hd_table, int devfn, void isa_ne2000_init(int base, qemu_irq irq, NICInfo *nd); +/* virtio-net.c */ + +void *virtio_net_init(PCIBus *bus, NICInfo *nd, int devfn); + + #endif diff --git a/qemu/hw/pci.c b/qemu/hw/pci.c index f8cbf1a..545901c 100644 --- a/qemu/hw/pci.c +++ b/qemu/hw/pci.c @@ -25,6 +25,7 @@ #include "pci.h" #include "console.h" #include "net.h" +#include "pc.h" //#define DEBUG_PCI @@ -638,9 +639,11 @@ void pci_nic_init(PCIBus *bus, NICInfo *nd, int devfn) pci_rtl8139_init(bus, nd, devfn); } else if (strcmp(nd->model, "pcnet") == 0) { pci_pcnet_init(bus, nd, devfn); + } else if (strcmp(nd->model, "virtio") == 0) { + virtio_net_init(bus, nd, devfn); } else if (strcmp(nd->model, "?") == 0) { fprintf(stderr, "qemu: Supported PCI NICs: i82551 i82557b i82559er" - " ne2k_pci pcnet rtl8139\n"); + " ne2k_pci pcnet rtl8139 virtio\n"); exit (1); } else { fprintf(stderr, "qemu: Unsupported NIC: %s\n", nd->model); diff --git a/qemu/hw/virtio-net.c b/qemu/hw/virtio-net.c new file mode 100644 index 0000000..3320dd8 --- /dev/null +++ b/qemu/hw/virtio-net.c @@ -0,0 +1,178 @@ +/* + * Virtio Network Device + * + * Copyright IBM, Corp. 2007 + * + * Authors: + * Anthony Liguori <[EMAIL PROTECTED]> + * + * This work is licensed under the terms of the GNU GPL, version 2. See + * the COPYING file in the top-level directory. + * + */ + +#include "virtio.h" +#include "net.h" +#include "pc.h" + +/* from Linux's virtio_net.h */ + +/* The ID for virtio_net */ +#define VIRTIO_ID_NET 1 + +/* The feature bitmap for virtio net */ +#define VIRTIO_NET_F_NO_CSUM 0 +#define VIRTIO_NET_F_TSO4 1 +#define VIRTIO_NET_F_UFO 2 +#define VIRTIO_NET_F_TSO4_ECN 3 +#define VIRTIO_NET_F_TSO6 4 +#define VIRTIO_NET_F_MAC 5 + +/* The config defining mac address (6 bytes) */ +struct virtio_net_config +{ + uint8_t mac[6]; +} __attribute__((packed)); + +/* This is the first element of the scatter-gather list. If you don't + * specify GSO or CSUM features, you can simply ignore the header. */ +struct virtio_net_hdr +{ +#define VIRTIO_NET_HDR_F_NEEDS_CSUM 1 // Use csum_start, csum_offset + uint8_t flags; +#define VIRTIO_NET_HDR_GSO_NONE 0 // Not a GSO frame +#define VIRTIO_NET_HDR_GSO_TCPV4 1 // GSO frame, IPv4 TCP (TSO) +/* FIXME: Do we need this? If they said they can handle ECN, do they care? */ +#define VIRTIO_NET_HDR_GSO_TCPV4_ECN 2 // GSO frame, IPv4 TCP w/ ECN +#define VIRTIO_NET_HDR_GSO_UDP 3 // GSO frame, IPv4 UDP (UFO) +#define VIRTIO_NET_HDR_GSO_TCPV6 4 // GSO frame, IPv6 TCP + uint8_t gso_type; + uint16_t gso_size; + uint16_t csum_start; + uint16_t csum_offset; +}; + +typedef struct VirtIONet +{ + VirtIODevice vdev; + uint8_t mac[6]; + VirtQueue *rx_vq; + VirtQueue *tx_vq; + VLANClientState *vc; + int can_receive; +} VirtIONet; + +static VirtIONet *to_virtio_net(VirtIODevice *vdev) +{ + return (VirtIONet *)vdev; +} + +static void virtio_net_update_config(VirtIODevice *vdev, uint8_t *config) +{ + VirtIONet *n = to_virtio_net(vdev); + struct virtio_net_config netcfg; + + memcpy(netcfg.mac, n->mac, 6); + memcpy(config, &netcfg, sizeof(netcfg)); +} + +static uint32_t virtio_net_get_features(VirtIODevice *vdev) +{ + return (1 << VIRTIO_NET_F_MAC); +} + +/* RX */ + +static void virtio_net_handle_rx(VirtIODevice *vdev, VirtQueue *vq) +{ + VirtIONet *n = to_virtio_net(vdev); + n->can_receive = 1; +} + +static int virtio_net_can_receive(void *opaque) +{ + VirtIONet *n = opaque; + + return (n->vdev.status & VIRTIO_CONFIG_S_DRIVER_OK) && n->can_receive; +} + +static void virtio_net_receive(void *opaque, const uint8_t *buf, int size) +{ + VirtIONet *n = opaque; + VirtQueueElement elem; + struct virtio_net_hdr *hdr; + int offset, i; + + /* FIXME: the drivers really need to set their status better */ + if (n->rx_vq->vring.avail == NULL) { + n->can_receive = 0; + return; + } + + if (virtqueue_pop(n->rx_vq, &elem) == 0) { + /* wait until the guest adds some rx bufs */ + n->can_receive = 0; + return; + } + + hdr = (void *)elem.in_sg[0].iov_base; + hdr->flags = 0; + hdr->gso_type = VIRTIO_NET_HDR_GSO_NONE; + + /* copy in packet. ugh */ + offset = 0; + i = 1; + while (offset < size && i < elem.in_num) { + int len = MIN(elem.in_sg[i].iov_len, size - offset); + memcpy(elem.in_sg[i].iov_base, buf + offset, len); + offset += len; + i++; + } + + /* signal other side */ + virtqueue_push(n->rx_vq, &elem, sizeof(*hdr) + offset); + virtio_notify(&n->vdev, n->rx_vq); +} + +/* TX */ +static void virtio_net_handle_tx(VirtIODevice *vdev, VirtQueue *vq) +{ + VirtIONet *n = to_virtio_net(vdev); + VirtQueueElement elem; + + while (virtqueue_pop(vq, &elem)) { + int i; + size_t len = 0; + + /* ignore the header for now */ + for (i = 1; i < elem.out_num; i++) { + qemu_send_packet(n->vc, elem.out_sg[i].iov_base, + elem.out_sg[i].iov_len); + len += elem.out_sg[i].iov_len; + } + + virtqueue_push(vq, &elem, sizeof(struct virtio_net_hdr) + len); + virtio_notify(&n->vdev, vq); + } +} + +void *virtio_net_init(PCIBus *bus, NICInfo *nd, int devfn) +{ + VirtIONet *n; + + n = (VirtIONet *)virtio_init_pci(bus, "virtio-net", 6900, 0x1000, + 0, VIRTIO_ID_NET, + 0x02, 0x00, 0x00, + 6, sizeof(VirtIONet)); + + n->vdev.update_config = virtio_net_update_config; + n->vdev.get_features = virtio_net_get_features; + n->rx_vq = virtio_add_queue(&n->vdev, 512, virtio_net_handle_rx); + n->tx_vq = virtio_add_queue(&n->vdev, 128, virtio_net_handle_tx); + n->can_receive = 0; + memcpy(n->mac, nd->macaddr, 6); + n->vc = qemu_new_vlan_client(nd->vlan, virtio_net_receive, + virtio_net_can_receive, n); + + return &n->vdev; +} -- 1.5.3.3 ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2005. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ kvm-devel mailing list kvm-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/kvm-devel