2009-09-24 09:59:20 -06:00
|
|
|
/* A network driver using virtio.
|
2007-10-21 19:03:37 -06:00
|
|
|
*
|
|
|
|
* Copyright 2007 Rusty Russell <rusty@rustcorp.com.au> IBM Corporation
|
|
|
|
*
|
|
|
|
* This program is free software; you can redistribute it and/or modify
|
|
|
|
* it under the terms of the GNU General Public License as published by
|
|
|
|
* the Free Software Foundation; either version 2 of the License, or
|
|
|
|
* (at your option) any later version.
|
|
|
|
*
|
|
|
|
* This program is distributed in the hope that it will be useful,
|
|
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
|
|
* GNU General Public License for more details.
|
|
|
|
*
|
|
|
|
* You should have received a copy of the GNU General Public License
|
|
|
|
* along with this program; if not, write to the Free Software
|
|
|
|
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
|
|
|
|
*/
|
|
|
|
//#define DEBUG
|
|
|
|
#include <linux/netdevice.h>
|
|
|
|
#include <linux/etherdevice.h>
|
2008-04-17 21:21:42 -06:00
|
|
|
#include <linux/ethtool.h>
|
2007-10-21 19:03:37 -06:00
|
|
|
#include <linux/module.h>
|
|
|
|
#include <linux/virtio.h>
|
2009-07-31 00:25:56 -06:00
|
|
|
#include <linux/virtio_ids.h>
|
2007-10-21 19:03:37 -06:00
|
|
|
#include <linux/virtio_net.h>
|
|
|
|
#include <linux/scatterlist.h>
|
2009-01-25 19:06:26 -07:00
|
|
|
#include <linux/if_vlan.h>
|
2007-10-21 19:03:37 -06:00
|
|
|
|
2007-12-16 06:19:43 -07:00
|
|
|
static int napi_weight = 128;
|
|
|
|
module_param(napi_weight, int, 0444);
|
|
|
|
|
2008-02-04 21:50:02 -07:00
|
|
|
static int csum = 1, gso = 1;
|
|
|
|
module_param(csum, bool, 0444);
|
|
|
|
module_param(gso, bool, 0444);
|
|
|
|
|
2007-10-21 19:03:37 -06:00
|
|
|
/* FIXME: MTU in config. */
|
2009-01-25 19:06:26 -07:00
|
|
|
#define MAX_PACKET_LEN (ETH_HLEN + VLAN_HLEN + ETH_DATA_LEN)
|
2008-11-16 23:41:34 -07:00
|
|
|
#define GOOD_COPY_LEN 128
|
2007-10-21 19:03:37 -06:00
|
|
|
|
2009-02-04 02:02:45 -07:00
|
|
|
#define VIRTNET_SEND_COMMAND_SG_MAX 2
|
2009-02-04 02:02:34 -07:00
|
|
|
|
2007-10-21 19:03:37 -06:00
|
|
|
struct virtnet_info
|
|
|
|
{
|
|
|
|
struct virtio_device *vdev;
|
2009-02-04 02:02:34 -07:00
|
|
|
struct virtqueue *rvq, *svq, *cvq;
|
2007-10-21 19:03:37 -06:00
|
|
|
struct net_device *dev;
|
|
|
|
struct napi_struct napi;
|
2009-01-19 18:09:49 -07:00
|
|
|
unsigned int status;
|
2007-10-21 19:03:37 -06:00
|
|
|
|
|
|
|
/* Number of input buffers, and max we've ever had. */
|
|
|
|
unsigned int num, max;
|
|
|
|
|
2008-04-17 21:24:27 -06:00
|
|
|
/* I like... big packets and I cannot lie! */
|
|
|
|
bool big_packets;
|
|
|
|
|
2008-11-16 23:41:34 -07:00
|
|
|
/* Host will merge rx buffers for big packets (shake it! shake it!) */
|
|
|
|
bool mergeable_rx_bufs;
|
|
|
|
|
2007-10-21 19:03:37 -06:00
|
|
|
/* Receive & send queues. */
|
|
|
|
struct sk_buff_head recv;
|
|
|
|
struct sk_buff_head send;
|
2008-07-25 11:06:01 -06:00
|
|
|
|
2009-08-26 13:22:32 -06:00
|
|
|
/* Work struct for refilling if we run low on memory. */
|
|
|
|
struct delayed_work refill;
|
|
|
|
|
2008-07-25 11:06:01 -06:00
|
|
|
/* Chain pages by the private ptr. */
|
|
|
|
struct page *pages;
|
2007-10-21 19:03:37 -06:00
|
|
|
};
|
|
|
|
|
2009-09-24 09:59:19 -06:00
|
|
|
struct skb_vnet_hdr {
|
|
|
|
union {
|
|
|
|
struct virtio_net_hdr hdr;
|
|
|
|
struct virtio_net_hdr_mrg_rxbuf mhdr;
|
|
|
|
};
|
2009-09-24 09:59:20 -06:00
|
|
|
unsigned int num_sg;
|
2009-09-24 09:59:19 -06:00
|
|
|
};
|
|
|
|
|
|
|
|
static inline struct skb_vnet_hdr *skb_vnet_hdr(struct sk_buff *skb)
|
2007-10-21 19:03:37 -06:00
|
|
|
{
|
2009-09-24 09:59:19 -06:00
|
|
|
return (struct skb_vnet_hdr *)skb->cb;
|
2007-10-21 19:03:37 -06:00
|
|
|
}
|
|
|
|
|
2008-07-25 11:06:01 -06:00
|
|
|
static void give_a_page(struct virtnet_info *vi, struct page *page)
|
|
|
|
{
|
|
|
|
page->private = (unsigned long)vi->pages;
|
|
|
|
vi->pages = page;
|
|
|
|
}
|
|
|
|
|
2008-11-16 23:39:18 -07:00
|
|
|
static void trim_pages(struct virtnet_info *vi, struct sk_buff *skb)
|
|
|
|
{
|
|
|
|
unsigned int i;
|
|
|
|
|
|
|
|
for (i = 0; i < skb_shinfo(skb)->nr_frags; i++)
|
|
|
|
give_a_page(vi, skb_shinfo(skb)->frags[i].page);
|
|
|
|
skb_shinfo(skb)->nr_frags = 0;
|
|
|
|
skb->data_len = 0;
|
|
|
|
}
|
|
|
|
|
2008-07-25 11:06:01 -06:00
|
|
|
static struct page *get_a_page(struct virtnet_info *vi, gfp_t gfp_mask)
|
|
|
|
{
|
|
|
|
struct page *p = vi->pages;
|
|
|
|
|
|
|
|
if (p)
|
|
|
|
vi->pages = (struct page *)p->private;
|
|
|
|
else
|
|
|
|
p = alloc_page(gfp_mask);
|
|
|
|
return p;
|
|
|
|
}
|
|
|
|
|
2008-02-04 21:50:07 -07:00
|
|
|
static void skb_xmit_done(struct virtqueue *svq)
|
2007-10-21 19:03:37 -06:00
|
|
|
{
|
2008-02-04 21:50:07 -07:00
|
|
|
struct virtnet_info *vi = svq->vdev->priv;
|
2007-10-21 19:03:37 -06:00
|
|
|
|
2008-02-04 21:50:07 -07:00
|
|
|
/* Suppress further interrupts. */
|
|
|
|
svq->vq_ops->disable_cb(svq);
|
2008-05-26 01:48:13 -06:00
|
|
|
|
2008-06-08 04:51:55 -06:00
|
|
|
/* We were probably waiting for more output buffers. */
|
2007-10-21 19:03:37 -06:00
|
|
|
netif_wake_queue(vi->dev);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void receive_skb(struct net_device *dev, struct sk_buff *skb,
|
|
|
|
unsigned len)
|
|
|
|
{
|
2008-11-16 23:41:34 -07:00
|
|
|
struct virtnet_info *vi = netdev_priv(dev);
|
2009-09-24 09:59:19 -06:00
|
|
|
struct skb_vnet_hdr *hdr = skb_vnet_hdr(skb);
|
2008-04-17 21:24:27 -06:00
|
|
|
int err;
|
2008-11-16 23:41:34 -07:00
|
|
|
int i;
|
2007-10-21 19:03:37 -06:00
|
|
|
|
|
|
|
if (unlikely(len < sizeof(struct virtio_net_hdr) + ETH_HLEN)) {
|
|
|
|
pr_debug("%s: short packet %i\n", dev->name, len);
|
|
|
|
dev->stats.rx_length_errors++;
|
|
|
|
goto drop;
|
|
|
|
}
|
2008-06-08 04:49:00 -06:00
|
|
|
|
2008-11-16 23:41:34 -07:00
|
|
|
if (vi->mergeable_rx_bufs) {
|
|
|
|
unsigned int copy;
|
|
|
|
char *p = page_address(skb_shinfo(skb)->frags[0].page);
|
2008-07-25 11:06:01 -06:00
|
|
|
|
2008-11-16 23:41:34 -07:00
|
|
|
if (len > PAGE_SIZE)
|
|
|
|
len = PAGE_SIZE;
|
|
|
|
len -= sizeof(struct virtio_net_hdr_mrg_rxbuf);
|
|
|
|
|
2009-09-24 09:59:19 -06:00
|
|
|
memcpy(&hdr->mhdr, p, sizeof(hdr->mhdr));
|
|
|
|
p += sizeof(hdr->mhdr);
|
2008-11-16 23:41:34 -07:00
|
|
|
|
|
|
|
copy = len;
|
|
|
|
if (copy > skb_tailroom(skb))
|
|
|
|
copy = skb_tailroom(skb);
|
|
|
|
|
|
|
|
memcpy(skb_put(skb, copy), p, copy);
|
|
|
|
|
|
|
|
len -= copy;
|
|
|
|
|
|
|
|
if (!len) {
|
|
|
|
give_a_page(vi, skb_shinfo(skb)->frags[0].page);
|
|
|
|
skb_shinfo(skb)->nr_frags--;
|
|
|
|
} else {
|
|
|
|
skb_shinfo(skb)->frags[0].page_offset +=
|
2009-09-24 09:59:19 -06:00
|
|
|
sizeof(hdr->mhdr) + copy;
|
2008-11-16 23:41:34 -07:00
|
|
|
skb_shinfo(skb)->frags[0].size = len;
|
|
|
|
skb->data_len += len;
|
|
|
|
skb->len += len;
|
|
|
|
}
|
|
|
|
|
2009-09-24 09:59:19 -06:00
|
|
|
while (--hdr->mhdr.num_buffers) {
|
2008-11-16 23:41:34 -07:00
|
|
|
struct sk_buff *nskb;
|
|
|
|
|
|
|
|
i = skb_shinfo(skb)->nr_frags;
|
|
|
|
if (i >= MAX_SKB_FRAGS) {
|
|
|
|
pr_debug("%s: packet too long %d\n", dev->name,
|
|
|
|
len);
|
|
|
|
dev->stats.rx_length_errors++;
|
|
|
|
goto drop;
|
|
|
|
}
|
|
|
|
|
|
|
|
nskb = vi->rvq->vq_ops->get_buf(vi->rvq, &len);
|
|
|
|
if (!nskb) {
|
|
|
|
pr_debug("%s: rx error: %d buffers missing\n",
|
2009-09-24 09:59:19 -06:00
|
|
|
dev->name, hdr->mhdr.num_buffers);
|
2008-11-16 23:41:34 -07:00
|
|
|
dev->stats.rx_length_errors++;
|
|
|
|
goto drop;
|
|
|
|
}
|
|
|
|
|
|
|
|
__skb_unlink(nskb, &vi->recv);
|
|
|
|
vi->num--;
|
|
|
|
|
|
|
|
skb_shinfo(skb)->frags[i] = skb_shinfo(nskb)->frags[0];
|
|
|
|
skb_shinfo(nskb)->nr_frags = 0;
|
|
|
|
kfree_skb(nskb);
|
|
|
|
|
|
|
|
if (len > PAGE_SIZE)
|
|
|
|
len = PAGE_SIZE;
|
|
|
|
|
|
|
|
skb_shinfo(skb)->frags[i].size = len;
|
|
|
|
skb_shinfo(skb)->nr_frags++;
|
|
|
|
skb->data_len += len;
|
|
|
|
skb->len += len;
|
|
|
|
}
|
|
|
|
} else {
|
2009-09-24 09:59:19 -06:00
|
|
|
len -= sizeof(hdr->hdr);
|
2008-11-16 23:41:34 -07:00
|
|
|
|
|
|
|
if (len <= MAX_PACKET_LEN)
|
|
|
|
trim_pages(vi, skb);
|
|
|
|
|
|
|
|
err = pskb_trim(skb, len);
|
|
|
|
if (err) {
|
|
|
|
pr_debug("%s: pskb_trim failed %i %d\n", dev->name,
|
|
|
|
len, err);
|
|
|
|
dev->stats.rx_dropped++;
|
|
|
|
goto drop;
|
|
|
|
}
|
2008-04-17 21:24:27 -06:00
|
|
|
}
|
2008-11-16 23:41:34 -07:00
|
|
|
|
2008-04-17 21:24:27 -06:00
|
|
|
skb->truesize += skb->data_len;
|
2007-10-21 19:03:37 -06:00
|
|
|
dev->stats.rx_bytes += skb->len;
|
|
|
|
dev->stats.rx_packets++;
|
|
|
|
|
2009-09-24 09:59:19 -06:00
|
|
|
if (hdr->hdr.flags & VIRTIO_NET_HDR_F_NEEDS_CSUM) {
|
2007-10-21 19:03:37 -06:00
|
|
|
pr_debug("Needs csum!\n");
|
2009-09-24 09:59:19 -06:00
|
|
|
if (!skb_partial_csum_set(skb,
|
|
|
|
hdr->hdr.csum_start,
|
|
|
|
hdr->hdr.csum_offset))
|
2007-10-21 19:03:37 -06:00
|
|
|
goto frame_err;
|
|
|
|
}
|
|
|
|
|
2008-06-08 04:49:00 -06:00
|
|
|
skb->protocol = eth_type_trans(skb, dev);
|
|
|
|
pr_debug("Receiving skb proto 0x%04x len %i type %i\n",
|
|
|
|
ntohs(skb->protocol), skb->len, skb->pkt_type);
|
|
|
|
|
2009-09-24 09:59:19 -06:00
|
|
|
if (hdr->hdr.gso_type != VIRTIO_NET_HDR_GSO_NONE) {
|
2007-10-21 19:03:37 -06:00
|
|
|
pr_debug("GSO!\n");
|
2009-09-24 09:59:19 -06:00
|
|
|
switch (hdr->hdr.gso_type & ~VIRTIO_NET_HDR_GSO_ECN) {
|
2007-10-21 19:03:37 -06:00
|
|
|
case VIRTIO_NET_HDR_GSO_TCPV4:
|
|
|
|
skb_shinfo(skb)->gso_type = SKB_GSO_TCPV4;
|
|
|
|
break;
|
|
|
|
case VIRTIO_NET_HDR_GSO_UDP:
|
|
|
|
skb_shinfo(skb)->gso_type = SKB_GSO_UDP;
|
|
|
|
break;
|
|
|
|
case VIRTIO_NET_HDR_GSO_TCPV6:
|
|
|
|
skb_shinfo(skb)->gso_type = SKB_GSO_TCPV6;
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
if (net_ratelimit())
|
|
|
|
printk(KERN_WARNING "%s: bad gso type %u.\n",
|
2009-09-24 09:59:19 -06:00
|
|
|
dev->name, hdr->hdr.gso_type);
|
2007-10-21 19:03:37 -06:00
|
|
|
goto frame_err;
|
|
|
|
}
|
|
|
|
|
2009-09-24 09:59:19 -06:00
|
|
|
if (hdr->hdr.gso_type & VIRTIO_NET_HDR_GSO_ECN)
|
2008-02-04 21:50:02 -07:00
|
|
|
skb_shinfo(skb)->gso_type |= SKB_GSO_TCP_ECN;
|
|
|
|
|
2009-09-24 09:59:19 -06:00
|
|
|
skb_shinfo(skb)->gso_size = hdr->hdr.gso_size;
|
2007-10-21 19:03:37 -06:00
|
|
|
if (skb_shinfo(skb)->gso_size == 0) {
|
|
|
|
if (net_ratelimit())
|
|
|
|
printk(KERN_WARNING "%s: zero gso size.\n",
|
|
|
|
dev->name);
|
|
|
|
goto frame_err;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Header must be checked, and gso_segs computed. */
|
|
|
|
skb_shinfo(skb)->gso_type |= SKB_GSO_DODGY;
|
|
|
|
skb_shinfo(skb)->gso_segs = 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
netif_receive_skb(skb);
|
|
|
|
return;
|
|
|
|
|
|
|
|
frame_err:
|
|
|
|
dev->stats.rx_frame_errors++;
|
|
|
|
drop:
|
|
|
|
dev_kfree_skb(skb);
|
|
|
|
}
|
|
|
|
|
2009-08-26 13:22:32 -06:00
|
|
|
static bool try_fill_recv_maxbufs(struct virtnet_info *vi, gfp_t gfp)
|
2007-10-21 19:03:37 -06:00
|
|
|
{
|
|
|
|
struct sk_buff *skb;
|
2008-05-02 20:50:45 -06:00
|
|
|
struct scatterlist sg[2+MAX_SKB_FRAGS];
|
2008-04-17 21:24:27 -06:00
|
|
|
int num, err, i;
|
2009-08-26 13:22:32 -06:00
|
|
|
bool oom = false;
|
2007-10-21 19:03:37 -06:00
|
|
|
|
2008-05-02 20:50:45 -06:00
|
|
|
sg_init_table(sg, 2+MAX_SKB_FRAGS);
|
2009-08-26 03:28:28 -06:00
|
|
|
do {
|
2009-09-24 09:59:19 -06:00
|
|
|
struct skb_vnet_hdr *hdr;
|
2008-11-16 23:41:34 -07:00
|
|
|
|
2009-10-12 23:34:20 -06:00
|
|
|
skb = netdev_alloc_skb_ip_align(vi->dev, MAX_PACKET_LEN);
|
2009-08-26 13:22:32 -06:00
|
|
|
if (unlikely(!skb)) {
|
|
|
|
oom = true;
|
2007-10-21 19:03:37 -06:00
|
|
|
break;
|
2009-08-26 13:22:32 -06:00
|
|
|
}
|
2007-10-21 19:03:37 -06:00
|
|
|
|
|
|
|
skb_put(skb, MAX_PACKET_LEN);
|
2008-11-16 23:41:34 -07:00
|
|
|
|
|
|
|
hdr = skb_vnet_hdr(skb);
|
2009-09-24 09:59:19 -06:00
|
|
|
sg_set_buf(sg, &hdr->hdr, sizeof(hdr->hdr));
|
2008-04-17 21:24:27 -06:00
|
|
|
|
|
|
|
if (vi->big_packets) {
|
|
|
|
for (i = 0; i < MAX_SKB_FRAGS; i++) {
|
|
|
|
skb_frag_t *f = &skb_shinfo(skb)->frags[i];
|
2009-08-26 13:22:32 -06:00
|
|
|
f->page = get_a_page(vi, gfp);
|
2008-04-17 21:24:27 -06:00
|
|
|
if (!f->page)
|
|
|
|
break;
|
|
|
|
|
|
|
|
f->page_offset = 0;
|
|
|
|
f->size = PAGE_SIZE;
|
|
|
|
|
|
|
|
skb->data_len += PAGE_SIZE;
|
|
|
|
skb->len += PAGE_SIZE;
|
|
|
|
|
|
|
|
skb_shinfo(skb)->nr_frags++;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2007-10-21 19:03:37 -06:00
|
|
|
num = skb_to_sgvec(skb, sg+1, 0, skb->len) + 1;
|
|
|
|
skb_queue_head(&vi->recv, skb);
|
|
|
|
|
|
|
|
err = vi->rvq->vq_ops->add_buf(vi->rvq, sg, 0, num, skb);
|
2009-09-23 22:26:31 -06:00
|
|
|
if (err < 0) {
|
2007-10-21 19:03:37 -06:00
|
|
|
skb_unlink(skb, &vi->recv);
|
2008-11-16 23:39:18 -07:00
|
|
|
trim_pages(vi, skb);
|
2007-10-21 19:03:37 -06:00
|
|
|
kfree_skb(skb);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
vi->num++;
|
2009-08-26 03:28:28 -06:00
|
|
|
} while (err >= num);
|
2007-10-21 19:03:37 -06:00
|
|
|
if (unlikely(vi->num > vi->max))
|
|
|
|
vi->max = vi->num;
|
|
|
|
vi->rvq->vq_ops->kick(vi->rvq);
|
2009-08-26 13:22:32 -06:00
|
|
|
return !oom;
|
2007-10-21 19:03:37 -06:00
|
|
|
}
|
|
|
|
|
2009-08-26 13:22:32 -06:00
|
|
|
/* Returns false if we couldn't fill entirely (OOM). */
|
|
|
|
static bool try_fill_recv(struct virtnet_info *vi, gfp_t gfp)
|
2008-11-16 23:41:34 -07:00
|
|
|
{
|
|
|
|
struct sk_buff *skb;
|
|
|
|
struct scatterlist sg[1];
|
|
|
|
int err;
|
2009-08-26 13:22:32 -06:00
|
|
|
bool oom = false;
|
2008-11-16 23:41:34 -07:00
|
|
|
|
2009-08-26 13:22:32 -06:00
|
|
|
if (!vi->mergeable_rx_bufs)
|
|
|
|
return try_fill_recv_maxbufs(vi, gfp);
|
2008-11-16 23:41:34 -07:00
|
|
|
|
2009-08-26 03:28:28 -06:00
|
|
|
do {
|
2008-11-16 23:41:34 -07:00
|
|
|
skb_frag_t *f;
|
|
|
|
|
2009-10-12 23:34:20 -06:00
|
|
|
skb = netdev_alloc_skb_ip_align(vi->dev, GOOD_COPY_LEN);
|
2009-08-26 13:22:32 -06:00
|
|
|
if (unlikely(!skb)) {
|
|
|
|
oom = true;
|
2008-11-16 23:41:34 -07:00
|
|
|
break;
|
2009-08-26 13:22:32 -06:00
|
|
|
}
|
2008-11-16 23:41:34 -07:00
|
|
|
|
|
|
|
f = &skb_shinfo(skb)->frags[0];
|
2009-08-26 13:22:32 -06:00
|
|
|
f->page = get_a_page(vi, gfp);
|
2008-11-16 23:41:34 -07:00
|
|
|
if (!f->page) {
|
2009-08-26 13:22:32 -06:00
|
|
|
oom = true;
|
2008-11-16 23:41:34 -07:00
|
|
|
kfree_skb(skb);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
f->page_offset = 0;
|
|
|
|
f->size = PAGE_SIZE;
|
|
|
|
|
|
|
|
skb_shinfo(skb)->nr_frags++;
|
|
|
|
|
|
|
|
sg_init_one(sg, page_address(f->page), PAGE_SIZE);
|
|
|
|
skb_queue_head(&vi->recv, skb);
|
|
|
|
|
|
|
|
err = vi->rvq->vq_ops->add_buf(vi->rvq, sg, 0, 1, skb);
|
2009-09-23 22:26:31 -06:00
|
|
|
if (err < 0) {
|
2008-11-16 23:41:34 -07:00
|
|
|
skb_unlink(skb, &vi->recv);
|
|
|
|
kfree_skb(skb);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
vi->num++;
|
2009-08-26 03:28:28 -06:00
|
|
|
} while (err > 0);
|
2008-11-16 23:41:34 -07:00
|
|
|
if (unlikely(vi->num > vi->max))
|
|
|
|
vi->max = vi->num;
|
|
|
|
vi->rvq->vq_ops->kick(vi->rvq);
|
2009-08-26 13:22:32 -06:00
|
|
|
return !oom;
|
2008-11-16 23:41:34 -07:00
|
|
|
}
|
|
|
|
|
2008-02-04 21:49:57 -07:00
|
|
|
static void skb_recv_done(struct virtqueue *rvq)
|
2007-10-21 19:03:37 -06:00
|
|
|
{
|
|
|
|
struct virtnet_info *vi = rvq->vdev->priv;
|
2008-02-04 21:49:57 -07:00
|
|
|
/* Schedule NAPI, Suppress further interrupts if successful. */
|
2009-01-19 17:43:59 -07:00
|
|
|
if (napi_schedule_prep(&vi->napi)) {
|
2008-02-04 21:49:57 -07:00
|
|
|
rvq->vq_ops->disable_cb(rvq);
|
2009-01-19 17:43:59 -07:00
|
|
|
__napi_schedule(&vi->napi);
|
2008-02-04 21:49:57 -07:00
|
|
|
}
|
2007-10-21 19:03:37 -06:00
|
|
|
}
|
|
|
|
|
2009-08-26 13:22:32 -06:00
|
|
|
static void refill_work(struct work_struct *work)
|
|
|
|
{
|
|
|
|
struct virtnet_info *vi;
|
|
|
|
bool still_empty;
|
|
|
|
|
|
|
|
vi = container_of(work, struct virtnet_info, refill.work);
|
|
|
|
napi_disable(&vi->napi);
|
|
|
|
try_fill_recv(vi, GFP_KERNEL);
|
|
|
|
still_empty = (vi->num == 0);
|
|
|
|
napi_enable(&vi->napi);
|
|
|
|
|
|
|
|
/* In theory, this can happen: if we don't get any buffers in
|
|
|
|
* we will *never* try to fill again. */
|
|
|
|
if (still_empty)
|
|
|
|
schedule_delayed_work(&vi->refill, HZ/2);
|
|
|
|
}
|
|
|
|
|
2007-10-21 19:03:37 -06:00
|
|
|
static int virtnet_poll(struct napi_struct *napi, int budget)
|
|
|
|
{
|
|
|
|
struct virtnet_info *vi = container_of(napi, struct virtnet_info, napi);
|
|
|
|
struct sk_buff *skb = NULL;
|
|
|
|
unsigned int len, received = 0;
|
|
|
|
|
|
|
|
again:
|
|
|
|
while (received < budget &&
|
|
|
|
(skb = vi->rvq->vq_ops->get_buf(vi->rvq, &len)) != NULL) {
|
|
|
|
__skb_unlink(skb, &vi->recv);
|
|
|
|
receive_skb(vi->dev, skb, len);
|
|
|
|
vi->num--;
|
|
|
|
received++;
|
|
|
|
}
|
|
|
|
|
2009-08-26 13:22:32 -06:00
|
|
|
if (vi->num < vi->max / 2) {
|
|
|
|
if (!try_fill_recv(vi, GFP_ATOMIC))
|
|
|
|
schedule_delayed_work(&vi->refill, 0);
|
|
|
|
}
|
2007-10-21 19:03:37 -06:00
|
|
|
|
2007-11-19 09:20:43 -07:00
|
|
|
/* Out of packets? */
|
|
|
|
if (received < budget) {
|
2009-01-19 17:43:59 -07:00
|
|
|
napi_complete(napi);
|
2008-02-04 21:49:57 -07:00
|
|
|
if (unlikely(!vi->rvq->vq_ops->enable_cb(vi->rvq))
|
virtio: fix race in enable_cb
There is a race in virtio_net, dealing with disabling/enabling the callback.
I saw the following oops:
kernel BUG at /space/kvm/drivers/virtio/virtio_ring.c:218!
illegal operation: 0001 [#1] SMP
Modules linked in: sunrpc dm_mod
CPU: 2 Not tainted 2.6.25-rc1zlive-host-10623-gd358142-dirty #99
Process swapper (pid: 0, task: 000000000f85a610, ksp: 000000000f873c60)
Krnl PSW : 0404300180000000 00000000002b81a6 (vring_disable_cb+0x16/0x20)
R:0 T:1 IO:0 EX:0 Key:0 M:1 W:0 P:0 AS:0 CC:3 PM:0 EA:3
Krnl GPRS: 0000000000000001 0000000000000001 0000000010005800 0000000000000001
000000000f3a0900 000000000f85a610 0000000000000000 0000000000000000
0000000000000000 000000000f870000 0000000000000000 0000000000001237
000000000f3a0920 000000000010ff74 00000000002846f6 000000000fa0bcd8
Krnl Code: 00000000002b819a: a7110001 tmll %r1,1
00000000002b819e: a7840004 brc 8,2b81a6
00000000002b81a2: a7f40001 brc 15,2b81a4
>00000000002b81a6: a51b0001 oill %r1,1
00000000002b81aa: 40102000 sth %r1,0(%r2)
00000000002b81ae: 07fe bcr 15,%r14
00000000002b81b0: eb7ff0380024 stmg %r7,%r15,56(%r15)
00000000002b81b6: a7f13e00 tmll %r15,15872
Call Trace:
([<000000000fa0bcd0>] 0xfa0bcd0)
[<00000000002b8350>] vring_interrupt+0x5c/0x6c
[<000000000010ab08>] do_extint+0xb8/0xf0
[<0000000000110716>] ext_no_vtime+0x16/0x1a
[<0000000000107e72>] cpu_idle+0x1c2/0x1e0
The problem can be triggered with a high amount of host->guest traffic.
I think its the following race:
poll says netif_rx_complete
poll calls enable_cb
enable_cb opens the interrupt mask
a new packet comes, an interrupt is triggered----\
enable_cb sees that there is more work |
enable_cb disables the interrupt |
. V
. interrupt is delivered
. skb_recv_done does atomic napi test, ok
some waiting disable_cb is called->check fails->bang!
.
poll would do napi check
poll would do disable_cb
The fix is to let enable_cb not disable the interrupt again, but expect the
caller to do the cleanup if it returns false. In that case, the interrupt is
only disabled, if the napi test_set_bit was successful.
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au> (cleaned up doco)
2008-03-14 07:17:05 -06:00
|
|
|
&& napi_schedule_prep(napi)) {
|
|
|
|
vi->rvq->vq_ops->disable_cb(vi->rvq);
|
2009-01-19 17:43:59 -07:00
|
|
|
__napi_schedule(napi);
|
2007-10-21 19:03:37 -06:00
|
|
|
goto again;
|
virtio: fix race in enable_cb
There is a race in virtio_net, dealing with disabling/enabling the callback.
I saw the following oops:
kernel BUG at /space/kvm/drivers/virtio/virtio_ring.c:218!
illegal operation: 0001 [#1] SMP
Modules linked in: sunrpc dm_mod
CPU: 2 Not tainted 2.6.25-rc1zlive-host-10623-gd358142-dirty #99
Process swapper (pid: 0, task: 000000000f85a610, ksp: 000000000f873c60)
Krnl PSW : 0404300180000000 00000000002b81a6 (vring_disable_cb+0x16/0x20)
R:0 T:1 IO:0 EX:0 Key:0 M:1 W:0 P:0 AS:0 CC:3 PM:0 EA:3
Krnl GPRS: 0000000000000001 0000000000000001 0000000010005800 0000000000000001
000000000f3a0900 000000000f85a610 0000000000000000 0000000000000000
0000000000000000 000000000f870000 0000000000000000 0000000000001237
000000000f3a0920 000000000010ff74 00000000002846f6 000000000fa0bcd8
Krnl Code: 00000000002b819a: a7110001 tmll %r1,1
00000000002b819e: a7840004 brc 8,2b81a6
00000000002b81a2: a7f40001 brc 15,2b81a4
>00000000002b81a6: a51b0001 oill %r1,1
00000000002b81aa: 40102000 sth %r1,0(%r2)
00000000002b81ae: 07fe bcr 15,%r14
00000000002b81b0: eb7ff0380024 stmg %r7,%r15,56(%r15)
00000000002b81b6: a7f13e00 tmll %r15,15872
Call Trace:
([<000000000fa0bcd0>] 0xfa0bcd0)
[<00000000002b8350>] vring_interrupt+0x5c/0x6c
[<000000000010ab08>] do_extint+0xb8/0xf0
[<0000000000110716>] ext_no_vtime+0x16/0x1a
[<0000000000107e72>] cpu_idle+0x1c2/0x1e0
The problem can be triggered with a high amount of host->guest traffic.
I think its the following race:
poll says netif_rx_complete
poll calls enable_cb
enable_cb opens the interrupt mask
a new packet comes, an interrupt is triggered----\
enable_cb sees that there is more work |
enable_cb disables the interrupt |
. V
. interrupt is delivered
. skb_recv_done does atomic napi test, ok
some waiting disable_cb is called->check fails->bang!
.
poll would do napi check
poll would do disable_cb
The fix is to let enable_cb not disable the interrupt again, but expect the
caller to do the cleanup if it returns false. In that case, the interrupt is
only disabled, if the napi test_set_bit was successful.
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au> (cleaned up doco)
2008-03-14 07:17:05 -06:00
|
|
|
}
|
2007-10-21 19:03:37 -06:00
|
|
|
}
|
|
|
|
|
|
|
|
return received;
|
|
|
|
}
|
|
|
|
|
2009-09-24 09:59:20 -06:00
|
|
|
static unsigned int free_old_xmit_skbs(struct virtnet_info *vi)
|
2007-10-21 19:03:37 -06:00
|
|
|
{
|
|
|
|
struct sk_buff *skb;
|
2009-09-24 09:59:20 -06:00
|
|
|
unsigned int len, tot_sgs = 0;
|
2007-10-21 19:03:37 -06:00
|
|
|
|
|
|
|
while ((skb = vi->svq->vq_ops->get_buf(vi->svq, &len)) != NULL) {
|
|
|
|
pr_debug("Sent skb %p\n", skb);
|
|
|
|
__skb_unlink(skb, &vi->send);
|
2008-05-02 20:50:43 -06:00
|
|
|
vi->dev->stats.tx_bytes += skb->len;
|
2007-10-21 19:03:37 -06:00
|
|
|
vi->dev->stats.tx_packets++;
|
2009-09-24 09:59:20 -06:00
|
|
|
tot_sgs += skb_vnet_hdr(skb)->num_sg;
|
2009-10-14 08:36:43 -06:00
|
|
|
dev_kfree_skb_any(skb);
|
2007-10-21 19:03:37 -06:00
|
|
|
}
|
2009-09-24 09:59:20 -06:00
|
|
|
return tot_sgs;
|
2007-10-21 19:03:37 -06:00
|
|
|
}
|
|
|
|
|
2008-05-02 20:50:46 -06:00
|
|
|
static int xmit_skb(struct virtnet_info *vi, struct sk_buff *skb)
|
2007-10-21 19:03:37 -06:00
|
|
|
{
|
2008-05-02 20:50:45 -06:00
|
|
|
struct scatterlist sg[2+MAX_SKB_FRAGS];
|
2009-09-24 09:59:19 -06:00
|
|
|
struct skb_vnet_hdr *hdr = skb_vnet_hdr(skb);
|
2007-10-21 19:03:37 -06:00
|
|
|
const unsigned char *dest = ((struct ethhdr *)skb->data)->h_dest;
|
|
|
|
|
2008-05-02 20:50:45 -06:00
|
|
|
sg_init_table(sg, 2+MAX_SKB_FRAGS);
|
2007-11-06 22:34:49 -07:00
|
|
|
|
2008-10-27 16:59:26 -06:00
|
|
|
pr_debug("%s: xmit %p %pM\n", vi->dev->name, skb, dest);
|
2007-10-21 19:03:37 -06:00
|
|
|
|
|
|
|
if (skb->ip_summed == CHECKSUM_PARTIAL) {
|
2009-09-24 09:59:19 -06:00
|
|
|
hdr->hdr.flags = VIRTIO_NET_HDR_F_NEEDS_CSUM;
|
|
|
|
hdr->hdr.csum_start = skb->csum_start - skb_headroom(skb);
|
|
|
|
hdr->hdr.csum_offset = skb->csum_offset;
|
2007-10-21 19:03:37 -06:00
|
|
|
} else {
|
2009-09-24 09:59:19 -06:00
|
|
|
hdr->hdr.flags = 0;
|
|
|
|
hdr->hdr.csum_offset = hdr->hdr.csum_start = 0;
|
2007-10-21 19:03:37 -06:00
|
|
|
}
|
|
|
|
|
|
|
|
if (skb_is_gso(skb)) {
|
2009-09-24 09:59:19 -06:00
|
|
|
hdr->hdr.hdr_len = skb_headlen(skb);
|
|
|
|
hdr->hdr.gso_size = skb_shinfo(skb)->gso_size;
|
2008-02-04 21:50:02 -07:00
|
|
|
if (skb_shinfo(skb)->gso_type & SKB_GSO_TCPV4)
|
2009-09-24 09:59:19 -06:00
|
|
|
hdr->hdr.gso_type = VIRTIO_NET_HDR_GSO_TCPV4;
|
2007-10-21 19:03:37 -06:00
|
|
|
else if (skb_shinfo(skb)->gso_type & SKB_GSO_TCPV6)
|
2009-09-24 09:59:19 -06:00
|
|
|
hdr->hdr.gso_type = VIRTIO_NET_HDR_GSO_TCPV6;
|
2007-10-21 19:03:37 -06:00
|
|
|
else if (skb_shinfo(skb)->gso_type & SKB_GSO_UDP)
|
2009-09-24 09:59:19 -06:00
|
|
|
hdr->hdr.gso_type = VIRTIO_NET_HDR_GSO_UDP;
|
2007-10-21 19:03:37 -06:00
|
|
|
else
|
|
|
|
BUG();
|
2008-02-04 21:50:02 -07:00
|
|
|
if (skb_shinfo(skb)->gso_type & SKB_GSO_TCP_ECN)
|
2009-09-24 09:59:19 -06:00
|
|
|
hdr->hdr.gso_type |= VIRTIO_NET_HDR_GSO_ECN;
|
2007-10-21 19:03:37 -06:00
|
|
|
} else {
|
2009-09-24 09:59:19 -06:00
|
|
|
hdr->hdr.gso_type = VIRTIO_NET_HDR_GSO_NONE;
|
|
|
|
hdr->hdr.gso_size = hdr->hdr.hdr_len = 0;
|
2007-10-21 19:03:37 -06:00
|
|
|
}
|
|
|
|
|
2009-09-24 09:59:19 -06:00
|
|
|
hdr->mhdr.num_buffers = 0;
|
2008-11-16 23:41:34 -07:00
|
|
|
|
|
|
|
/* Encode metadata header at front. */
|
|
|
|
if (vi->mergeable_rx_bufs)
|
2009-09-24 09:59:19 -06:00
|
|
|
sg_set_buf(sg, &hdr->mhdr, sizeof(hdr->mhdr));
|
2008-11-16 23:41:34 -07:00
|
|
|
else
|
2009-09-24 09:59:19 -06:00
|
|
|
sg_set_buf(sg, &hdr->hdr, sizeof(hdr->hdr));
|
2008-11-16 23:41:34 -07:00
|
|
|
|
2009-09-24 09:59:20 -06:00
|
|
|
hdr->num_sg = skb_to_sgvec(skb, sg+1, 0, skb->len) + 1;
|
|
|
|
return vi->svq->vq_ops->add_buf(vi->svq, sg, hdr->num_sg, 0, skb);
|
2008-05-26 01:48:13 -06:00
|
|
|
}
|
|
|
|
|
2009-08-31 13:50:51 -06:00
|
|
|
static netdev_tx_t start_xmit(struct sk_buff *skb, struct net_device *dev)
|
2008-05-02 20:50:46 -06:00
|
|
|
{
|
|
|
|
struct virtnet_info *vi = netdev_priv(dev);
|
2009-09-24 09:59:20 -06:00
|
|
|
int capacity;
|
2008-02-04 21:50:07 -07:00
|
|
|
|
|
|
|
again:
|
|
|
|
/* Free up any pending old buffers before queueing new ones. */
|
|
|
|
free_old_xmit_skbs(vi);
|
2008-05-02 20:50:46 -06:00
|
|
|
|
2009-10-28 05:03:38 -06:00
|
|
|
/* Try to transmit */
|
2009-09-24 09:59:20 -06:00
|
|
|
capacity = xmit_skb(vi, skb);
|
|
|
|
|
|
|
|
/* This can happen with OOM and indirect buffers. */
|
|
|
|
if (unlikely(capacity < 0)) {
|
|
|
|
netif_stop_queue(dev);
|
|
|
|
dev_warn(&dev->dev, "Unexpected full queue\n");
|
|
|
|
if (unlikely(!vi->svq->vq_ops->enable_cb(vi->svq))) {
|
|
|
|
vi->svq->vq_ops->disable_cb(vi->svq);
|
|
|
|
netif_start_queue(dev);
|
|
|
|
goto again;
|
|
|
|
}
|
|
|
|
return NETDEV_TX_BUSY;
|
2007-10-21 19:03:37 -06:00
|
|
|
}
|
2009-09-24 09:59:20 -06:00
|
|
|
vi->svq->vq_ops->kick(vi->svq);
|
2009-10-28 05:03:38 -06:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Put new one in send queue. You'd expect we'd need this before
|
|
|
|
* xmit_skb calls add_buf(), since the callback can be triggered
|
|
|
|
* immediately after that. But since the callback just triggers
|
|
|
|
* another call back here, normal network xmit locking prevents the
|
|
|
|
* race.
|
|
|
|
*/
|
|
|
|
__skb_queue_head(&vi->send, skb);
|
|
|
|
|
2009-09-24 09:59:20 -06:00
|
|
|
/* Don't wait up for transmitted skbs to be freed. */
|
|
|
|
skb_orphan(skb);
|
|
|
|
nf_reset(skb);
|
|
|
|
|
|
|
|
/* Apparently nice girls don't return TX_BUSY; stop the queue
|
|
|
|
* before it gets out of hand. Naturally, this wastes entries. */
|
|
|
|
if (capacity < 2+MAX_SKB_FRAGS) {
|
|
|
|
netif_stop_queue(dev);
|
|
|
|
if (unlikely(!vi->svq->vq_ops->enable_cb(vi->svq))) {
|
|
|
|
/* More just got used, free them then recheck. */
|
|
|
|
capacity += free_old_xmit_skbs(vi);
|
|
|
|
if (capacity >= 2+MAX_SKB_FRAGS) {
|
|
|
|
netif_start_queue(dev);
|
|
|
|
vi->svq->vq_ops->disable_cb(vi->svq);
|
|
|
|
}
|
|
|
|
}
|
2008-05-02 20:50:46 -06:00
|
|
|
}
|
2009-09-24 09:59:20 -06:00
|
|
|
|
|
|
|
return NETDEV_TX_OK;
|
2007-10-21 19:03:37 -06:00
|
|
|
}
|
|
|
|
|
2009-02-04 17:36:34 -07:00
|
|
|
static int virtnet_set_mac_address(struct net_device *dev, void *p)
|
|
|
|
{
|
|
|
|
struct virtnet_info *vi = netdev_priv(dev);
|
|
|
|
struct virtio_device *vdev = vi->vdev;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
ret = eth_mac_addr(dev, p);
|
|
|
|
if (ret)
|
|
|
|
return ret;
|
|
|
|
|
2009-04-04 17:40:19 -06:00
|
|
|
if (virtio_has_feature(vdev, VIRTIO_NET_F_MAC))
|
|
|
|
vdev->config->set(vdev, offsetof(struct virtio_net_config, mac),
|
|
|
|
dev->dev_addr, dev->addr_len);
|
2009-02-04 17:36:34 -07:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2008-02-29 03:54:50 -07:00
|
|
|
#ifdef CONFIG_NET_POLL_CONTROLLER
|
|
|
|
static void virtnet_netpoll(struct net_device *dev)
|
|
|
|
{
|
|
|
|
struct virtnet_info *vi = netdev_priv(dev);
|
|
|
|
|
|
|
|
napi_schedule(&vi->napi);
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
2007-10-21 19:03:37 -06:00
|
|
|
static int virtnet_open(struct net_device *dev)
|
|
|
|
{
|
|
|
|
struct virtnet_info *vi = netdev_priv(dev);
|
|
|
|
|
|
|
|
napi_enable(&vi->napi);
|
2008-02-04 21:50:07 -07:00
|
|
|
|
|
|
|
/* If all buffers were filled by other side before we napi_enabled, we
|
|
|
|
* won't get another interrupt, so process any outstanding packets
|
virtio net: fix oops on interface-up
I got the following oops during interface ifup. Unfortunately its not
easily reproducable so I cant say for sure that my fix fixes this
problem, but I am confident and I think its correct anyway:
<2>kernel BUG at /space/kvm/drivers/virtio/virtio_ring.c:234!
<4>illegal operation: 0001 [#1] PREEMPT SMP
<4>Modules linked in:
<4>CPU: 0 Not tainted 2.6.24zlive-guest-07293-gf1ca151-dirty #91
<4>Process swapper (pid: 0, task: 0000000000800938, ksp: 000000000084ddb8)
<4>Krnl PSW : 0404300180000000 0000000000466374 (vring_disable_cb+0x30/0x34)
<4> R:0 T:1 IO:0 EX:0 Key:0 M:1 W:0 P:0 AS:0 CC:3 PM:0 EA:3
<4>Krnl GPRS: 0000000000000001 0000000000000001 0000000010003800 0000000000466344
<4> 000000000e980900 00000000008848b0 000000000084e748 0000000000000000
<4> 000000000087b300 0000000000001237 0000000000001237 000000000f85bdd8
<4> 000000000e980920 00000000001137c0 0000000000464754 000000000f85bdd8
<4>Krnl Code: 0000000000466368: e3b0b0700004 lg %r11,112(%r11)
<4> 000000000046636e: 07fe bcr 15,%r14
<4> 0000000000466370: a7f40001 brc 15,466372
<4> >0000000000466374: a7f4fff6 brc 15,466360
<4> 0000000000466378: eb7ff0500024 stmg %r7,%r15,80(%r15)
<4> 000000000046637e: a7f13e00 tmll %r15,15872
<4> 0000000000466382: b90400ef lgr %r14,%r15
<4> 0000000000466386: a7840001 brc 8,466388
<4>Call Trace:
<4>([<000201500f85c000>] 0x201500f85c000)
<4> [<0000000000466556>] vring_interrupt+0x72/0x88
<4> [<00000000004801a0>] kvm_extint_handler+0x34/0x44
<4> [<000000000010d22c>] do_extint+0xbc/0xf8
<4> [<0000000000113f98>] ext_no_vtime+0x16/0x1a
<4> [<000000000010a182>] cpu_idle+0x216/0x238
<4>([<000000000010a162>] cpu_idle+0x1f6/0x238)
<4> [<0000000000568656>] rest_init+0xaa/0xb8
<4> [<000000000084ee2c>] start_kernel+0x3fc/0x490
<4> [<0000000000100020>] _stext+0x20/0x80
<4>
<4> <0>Kernel panic - not syncing: Fatal exception in interrupt
<4>
After looking at the code and the dump I think the following scenario
happened: Ifup was running on cpu2 and the interrupt arrived on cpu0.
Now virtnet_open on cpu 2 managed to execute napi_enable and disable_cb
but did not execute rx_schedule. Meanwhile on cpu 0 skb_recv_done was
called by vring_interrupt, executed netif_rx_schedule_prep, which
succeeded and therefore called disable_cb. This triggered the BUG_ON,
as interrupts were already disabled by cpu 2.
I think the proper solution is to make the call to disable_cb depend on
the atomic update of NAPI_STATE_SCHED by using netif_rx_schedule_prep
in the same way as skb_recv_done.
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
2008-02-06 00:50:11 -07:00
|
|
|
* now. virtnet_poll wants re-enable the queue, so we disable here.
|
|
|
|
* We synchronize against interrupts via NAPI_STATE_SCHED */
|
2009-01-19 17:43:59 -07:00
|
|
|
if (napi_schedule_prep(&vi->napi)) {
|
virtio net: fix oops on interface-up
I got the following oops during interface ifup. Unfortunately its not
easily reproducable so I cant say for sure that my fix fixes this
problem, but I am confident and I think its correct anyway:
<2>kernel BUG at /space/kvm/drivers/virtio/virtio_ring.c:234!
<4>illegal operation: 0001 [#1] PREEMPT SMP
<4>Modules linked in:
<4>CPU: 0 Not tainted 2.6.24zlive-guest-07293-gf1ca151-dirty #91
<4>Process swapper (pid: 0, task: 0000000000800938, ksp: 000000000084ddb8)
<4>Krnl PSW : 0404300180000000 0000000000466374 (vring_disable_cb+0x30/0x34)
<4> R:0 T:1 IO:0 EX:0 Key:0 M:1 W:0 P:0 AS:0 CC:3 PM:0 EA:3
<4>Krnl GPRS: 0000000000000001 0000000000000001 0000000010003800 0000000000466344
<4> 000000000e980900 00000000008848b0 000000000084e748 0000000000000000
<4> 000000000087b300 0000000000001237 0000000000001237 000000000f85bdd8
<4> 000000000e980920 00000000001137c0 0000000000464754 000000000f85bdd8
<4>Krnl Code: 0000000000466368: e3b0b0700004 lg %r11,112(%r11)
<4> 000000000046636e: 07fe bcr 15,%r14
<4> 0000000000466370: a7f40001 brc 15,466372
<4> >0000000000466374: a7f4fff6 brc 15,466360
<4> 0000000000466378: eb7ff0500024 stmg %r7,%r15,80(%r15)
<4> 000000000046637e: a7f13e00 tmll %r15,15872
<4> 0000000000466382: b90400ef lgr %r14,%r15
<4> 0000000000466386: a7840001 brc 8,466388
<4>Call Trace:
<4>([<000201500f85c000>] 0x201500f85c000)
<4> [<0000000000466556>] vring_interrupt+0x72/0x88
<4> [<00000000004801a0>] kvm_extint_handler+0x34/0x44
<4> [<000000000010d22c>] do_extint+0xbc/0xf8
<4> [<0000000000113f98>] ext_no_vtime+0x16/0x1a
<4> [<000000000010a182>] cpu_idle+0x216/0x238
<4>([<000000000010a162>] cpu_idle+0x1f6/0x238)
<4> [<0000000000568656>] rest_init+0xaa/0xb8
<4> [<000000000084ee2c>] start_kernel+0x3fc/0x490
<4> [<0000000000100020>] _stext+0x20/0x80
<4>
<4> <0>Kernel panic - not syncing: Fatal exception in interrupt
<4>
After looking at the code and the dump I think the following scenario
happened: Ifup was running on cpu2 and the interrupt arrived on cpu0.
Now virtnet_open on cpu 2 managed to execute napi_enable and disable_cb
but did not execute rx_schedule. Meanwhile on cpu 0 skb_recv_done was
called by vring_interrupt, executed netif_rx_schedule_prep, which
succeeded and therefore called disable_cb. This triggered the BUG_ON,
as interrupts were already disabled by cpu 2.
I think the proper solution is to make the call to disable_cb depend on
the atomic update of NAPI_STATE_SCHED by using netif_rx_schedule_prep
in the same way as skb_recv_done.
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
2008-02-06 00:50:11 -07:00
|
|
|
vi->rvq->vq_ops->disable_cb(vi->rvq);
|
2009-01-19 17:43:59 -07:00
|
|
|
__napi_schedule(&vi->napi);
|
virtio net: fix oops on interface-up
I got the following oops during interface ifup. Unfortunately its not
easily reproducable so I cant say for sure that my fix fixes this
problem, but I am confident and I think its correct anyway:
<2>kernel BUG at /space/kvm/drivers/virtio/virtio_ring.c:234!
<4>illegal operation: 0001 [#1] PREEMPT SMP
<4>Modules linked in:
<4>CPU: 0 Not tainted 2.6.24zlive-guest-07293-gf1ca151-dirty #91
<4>Process swapper (pid: 0, task: 0000000000800938, ksp: 000000000084ddb8)
<4>Krnl PSW : 0404300180000000 0000000000466374 (vring_disable_cb+0x30/0x34)
<4> R:0 T:1 IO:0 EX:0 Key:0 M:1 W:0 P:0 AS:0 CC:3 PM:0 EA:3
<4>Krnl GPRS: 0000000000000001 0000000000000001 0000000010003800 0000000000466344
<4> 000000000e980900 00000000008848b0 000000000084e748 0000000000000000
<4> 000000000087b300 0000000000001237 0000000000001237 000000000f85bdd8
<4> 000000000e980920 00000000001137c0 0000000000464754 000000000f85bdd8
<4>Krnl Code: 0000000000466368: e3b0b0700004 lg %r11,112(%r11)
<4> 000000000046636e: 07fe bcr 15,%r14
<4> 0000000000466370: a7f40001 brc 15,466372
<4> >0000000000466374: a7f4fff6 brc 15,466360
<4> 0000000000466378: eb7ff0500024 stmg %r7,%r15,80(%r15)
<4> 000000000046637e: a7f13e00 tmll %r15,15872
<4> 0000000000466382: b90400ef lgr %r14,%r15
<4> 0000000000466386: a7840001 brc 8,466388
<4>Call Trace:
<4>([<000201500f85c000>] 0x201500f85c000)
<4> [<0000000000466556>] vring_interrupt+0x72/0x88
<4> [<00000000004801a0>] kvm_extint_handler+0x34/0x44
<4> [<000000000010d22c>] do_extint+0xbc/0xf8
<4> [<0000000000113f98>] ext_no_vtime+0x16/0x1a
<4> [<000000000010a182>] cpu_idle+0x216/0x238
<4>([<000000000010a162>] cpu_idle+0x1f6/0x238)
<4> [<0000000000568656>] rest_init+0xaa/0xb8
<4> [<000000000084ee2c>] start_kernel+0x3fc/0x490
<4> [<0000000000100020>] _stext+0x20/0x80
<4>
<4> <0>Kernel panic - not syncing: Fatal exception in interrupt
<4>
After looking at the code and the dump I think the following scenario
happened: Ifup was running on cpu2 and the interrupt arrived on cpu0.
Now virtnet_open on cpu 2 managed to execute napi_enable and disable_cb
but did not execute rx_schedule. Meanwhile on cpu 0 skb_recv_done was
called by vring_interrupt, executed netif_rx_schedule_prep, which
succeeded and therefore called disable_cb. This triggered the BUG_ON,
as interrupts were already disabled by cpu 2.
I think the proper solution is to make the call to disable_cb depend on
the atomic update of NAPI_STATE_SCHED by using netif_rx_schedule_prep
in the same way as skb_recv_done.
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
2008-02-06 00:50:11 -07:00
|
|
|
}
|
2007-10-21 19:03:37 -06:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2009-02-04 02:02:34 -07:00
|
|
|
/*
|
|
|
|
* Send command via the control virtqueue and check status. Commands
|
|
|
|
* supported by the hypervisor, as indicated by feature bits, should
|
|
|
|
* never fail unless improperly formated.
|
|
|
|
*/
|
|
|
|
static bool virtnet_send_command(struct virtnet_info *vi, u8 class, u8 cmd,
|
|
|
|
struct scatterlist *data, int out, int in)
|
|
|
|
{
|
2009-05-01 11:27:56 -06:00
|
|
|
struct scatterlist *s, sg[VIRTNET_SEND_COMMAND_SG_MAX + 2];
|
2009-02-04 02:02:34 -07:00
|
|
|
struct virtio_net_ctrl_hdr ctrl;
|
|
|
|
virtio_net_ctrl_ack status = ~0;
|
|
|
|
unsigned int tmp;
|
2009-05-01 11:27:56 -06:00
|
|
|
int i;
|
2009-02-04 02:02:34 -07:00
|
|
|
|
2009-04-11 08:50:23 -06:00
|
|
|
/* Caller should know better */
|
|
|
|
BUG_ON(!virtio_has_feature(vi->vdev, VIRTIO_NET_F_CTRL_VQ) ||
|
|
|
|
(out + in > VIRTNET_SEND_COMMAND_SG_MAX));
|
2009-02-04 02:02:34 -07:00
|
|
|
|
|
|
|
out++; /* Add header */
|
|
|
|
in++; /* Add return status */
|
|
|
|
|
|
|
|
ctrl.class = class;
|
|
|
|
ctrl.cmd = cmd;
|
|
|
|
|
|
|
|
sg_init_table(sg, out + in);
|
|
|
|
|
|
|
|
sg_set_buf(&sg[0], &ctrl, sizeof(ctrl));
|
2009-05-01 11:27:56 -06:00
|
|
|
for_each_sg(data, s, out + in - 2, i)
|
|
|
|
sg_set_buf(&sg[i + 1], sg_virt(s), s->length);
|
2009-02-04 02:02:34 -07:00
|
|
|
sg_set_buf(&sg[out + in - 1], &status, sizeof(status));
|
|
|
|
|
2009-09-23 22:26:31 -06:00
|
|
|
BUG_ON(vi->cvq->vq_ops->add_buf(vi->cvq, sg, out, in, vi) < 0);
|
2009-02-04 02:02:34 -07:00
|
|
|
|
|
|
|
vi->cvq->vq_ops->kick(vi->cvq);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Spin for a response, the kick causes an ioport write, trapping
|
|
|
|
* into the hypervisor, so the request should be handled immediately.
|
|
|
|
*/
|
|
|
|
while (!vi->cvq->vq_ops->get_buf(vi->cvq, &tmp))
|
|
|
|
cpu_relax();
|
|
|
|
|
|
|
|
return status == VIRTIO_NET_OK;
|
|
|
|
}
|
|
|
|
|
2007-10-21 19:03:37 -06:00
|
|
|
static int virtnet_close(struct net_device *dev)
|
|
|
|
{
|
|
|
|
struct virtnet_info *vi = netdev_priv(dev);
|
|
|
|
|
|
|
|
napi_disable(&vi->napi);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2008-04-17 21:21:42 -06:00
|
|
|
static int virtnet_set_tx_csum(struct net_device *dev, u32 data)
|
|
|
|
{
|
|
|
|
struct virtnet_info *vi = netdev_priv(dev);
|
|
|
|
struct virtio_device *vdev = vi->vdev;
|
|
|
|
|
|
|
|
if (data && !virtio_has_feature(vdev, VIRTIO_NET_F_CSUM))
|
|
|
|
return -ENOSYS;
|
|
|
|
|
|
|
|
return ethtool_op_set_tx_hw_csum(dev, data);
|
|
|
|
}
|
|
|
|
|
2009-02-04 02:02:40 -07:00
|
|
|
static void virtnet_set_rx_mode(struct net_device *dev)
|
|
|
|
{
|
|
|
|
struct virtnet_info *vi = netdev_priv(dev);
|
2009-02-04 02:02:45 -07:00
|
|
|
struct scatterlist sg[2];
|
2009-02-04 02:02:40 -07:00
|
|
|
u8 promisc, allmulti;
|
2009-02-04 02:02:45 -07:00
|
|
|
struct virtio_net_ctrl_mac *mac_data;
|
|
|
|
struct dev_addr_list *addr;
|
2009-05-22 17:22:17 -06:00
|
|
|
struct netdev_hw_addr *ha;
|
2009-02-04 02:02:45 -07:00
|
|
|
void *buf;
|
|
|
|
int i;
|
2009-02-04 02:02:40 -07:00
|
|
|
|
|
|
|
/* We can't dynamicaly set ndo_set_rx_mode, so return gracefully */
|
|
|
|
if (!virtio_has_feature(vi->vdev, VIRTIO_NET_F_CTRL_RX))
|
|
|
|
return;
|
|
|
|
|
2009-02-04 02:02:45 -07:00
|
|
|
promisc = ((dev->flags & IFF_PROMISC) != 0);
|
|
|
|
allmulti = ((dev->flags & IFF_ALLMULTI) != 0);
|
2009-02-04 02:02:40 -07:00
|
|
|
|
2009-05-01 11:27:56 -06:00
|
|
|
sg_init_one(sg, &promisc, sizeof(promisc));
|
2009-02-04 02:02:40 -07:00
|
|
|
|
|
|
|
if (!virtnet_send_command(vi, VIRTIO_NET_CTRL_RX,
|
|
|
|
VIRTIO_NET_CTRL_RX_PROMISC,
|
2009-02-04 02:02:45 -07:00
|
|
|
sg, 1, 0))
|
2009-02-04 02:02:40 -07:00
|
|
|
dev_warn(&dev->dev, "Failed to %sable promisc mode.\n",
|
|
|
|
promisc ? "en" : "dis");
|
|
|
|
|
2009-05-01 11:27:56 -06:00
|
|
|
sg_init_one(sg, &allmulti, sizeof(allmulti));
|
2009-02-04 02:02:40 -07:00
|
|
|
|
|
|
|
if (!virtnet_send_command(vi, VIRTIO_NET_CTRL_RX,
|
|
|
|
VIRTIO_NET_CTRL_RX_ALLMULTI,
|
2009-02-04 02:02:45 -07:00
|
|
|
sg, 1, 0))
|
2009-02-04 02:02:40 -07:00
|
|
|
dev_warn(&dev->dev, "Failed to %sable allmulti mode.\n",
|
|
|
|
allmulti ? "en" : "dis");
|
2009-02-04 02:02:45 -07:00
|
|
|
|
|
|
|
/* MAC filter - use one buffer for both lists */
|
2009-06-16 19:12:19 -06:00
|
|
|
mac_data = buf = kzalloc(((dev->uc.count + dev->mc_count) * ETH_ALEN) +
|
2009-02-04 02:02:45 -07:00
|
|
|
(2 * sizeof(mac_data->entries)), GFP_ATOMIC);
|
|
|
|
if (!buf) {
|
|
|
|
dev_warn(&dev->dev, "No memory for MAC address buffer\n");
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
2009-05-01 11:27:56 -06:00
|
|
|
sg_init_table(sg, 2);
|
|
|
|
|
2009-02-04 02:02:45 -07:00
|
|
|
/* Store the unicast list and count in the front of the buffer */
|
2009-06-16 19:12:19 -06:00
|
|
|
mac_data->entries = dev->uc.count;
|
2009-05-22 17:22:17 -06:00
|
|
|
i = 0;
|
2009-06-16 19:12:19 -06:00
|
|
|
list_for_each_entry(ha, &dev->uc.list, list)
|
2009-05-22 17:22:17 -06:00
|
|
|
memcpy(&mac_data->macs[i++][0], ha->addr, ETH_ALEN);
|
2009-02-04 02:02:45 -07:00
|
|
|
|
|
|
|
sg_set_buf(&sg[0], mac_data,
|
2009-06-16 19:12:19 -06:00
|
|
|
sizeof(mac_data->entries) + (dev->uc.count * ETH_ALEN));
|
2009-02-04 02:02:45 -07:00
|
|
|
|
|
|
|
/* multicast list and count fill the end */
|
2009-06-16 19:12:19 -06:00
|
|
|
mac_data = (void *)&mac_data->macs[dev->uc.count][0];
|
2009-02-04 02:02:45 -07:00
|
|
|
|
|
|
|
mac_data->entries = dev->mc_count;
|
|
|
|
addr = dev->mc_list;
|
|
|
|
for (i = 0; i < dev->mc_count; i++, addr = addr->next)
|
|
|
|
memcpy(&mac_data->macs[i][0], addr->da_addr, ETH_ALEN);
|
|
|
|
|
|
|
|
sg_set_buf(&sg[1], mac_data,
|
|
|
|
sizeof(mac_data->entries) + (dev->mc_count * ETH_ALEN));
|
|
|
|
|
|
|
|
if (!virtnet_send_command(vi, VIRTIO_NET_CTRL_MAC,
|
|
|
|
VIRTIO_NET_CTRL_MAC_TABLE_SET,
|
|
|
|
sg, 2, 0))
|
|
|
|
dev_warn(&dev->dev, "Failed to set MAC fitler table.\n");
|
|
|
|
|
|
|
|
kfree(buf);
|
2009-02-04 02:02:40 -07:00
|
|
|
}
|
|
|
|
|
2009-05-01 11:31:10 -06:00
|
|
|
static void virtnet_vlan_rx_add_vid(struct net_device *dev, u16 vid)
|
2009-02-04 02:02:50 -07:00
|
|
|
{
|
|
|
|
struct virtnet_info *vi = netdev_priv(dev);
|
|
|
|
struct scatterlist sg;
|
|
|
|
|
2009-05-01 11:27:56 -06:00
|
|
|
sg_init_one(&sg, &vid, sizeof(vid));
|
2009-02-04 02:02:50 -07:00
|
|
|
|
|
|
|
if (!virtnet_send_command(vi, VIRTIO_NET_CTRL_VLAN,
|
|
|
|
VIRTIO_NET_CTRL_VLAN_ADD, &sg, 1, 0))
|
|
|
|
dev_warn(&dev->dev, "Failed to add VLAN ID %d.\n", vid);
|
|
|
|
}
|
|
|
|
|
2009-05-01 11:31:10 -06:00
|
|
|
static void virtnet_vlan_rx_kill_vid(struct net_device *dev, u16 vid)
|
2009-02-04 02:02:50 -07:00
|
|
|
{
|
|
|
|
struct virtnet_info *vi = netdev_priv(dev);
|
|
|
|
struct scatterlist sg;
|
|
|
|
|
2009-05-01 11:27:56 -06:00
|
|
|
sg_init_one(&sg, &vid, sizeof(vid));
|
2009-02-04 02:02:50 -07:00
|
|
|
|
|
|
|
if (!virtnet_send_command(vi, VIRTIO_NET_CTRL_VLAN,
|
|
|
|
VIRTIO_NET_CTRL_VLAN_DEL, &sg, 1, 0))
|
|
|
|
dev_warn(&dev->dev, "Failed to kill VLAN ID %d.\n", vid);
|
|
|
|
}
|
|
|
|
|
2009-09-02 02:03:33 -06:00
|
|
|
static const struct ethtool_ops virtnet_ethtool_ops = {
|
2008-04-17 21:21:42 -06:00
|
|
|
.set_tx_csum = virtnet_set_tx_csum,
|
|
|
|
.set_sg = ethtool_op_set_sg,
|
2008-11-16 23:40:36 -07:00
|
|
|
.set_tso = ethtool_op_set_tso,
|
2009-07-14 08:21:02 -06:00
|
|
|
.set_ufo = ethtool_op_set_ufo,
|
2009-01-19 18:09:49 -07:00
|
|
|
.get_link = ethtool_op_get_link,
|
2008-04-17 21:21:42 -06:00
|
|
|
};
|
|
|
|
|
2008-11-26 06:58:11 -07:00
|
|
|
#define MIN_MTU 68
|
|
|
|
#define MAX_MTU 65535
|
|
|
|
|
|
|
|
static int virtnet_change_mtu(struct net_device *dev, int new_mtu)
|
|
|
|
{
|
|
|
|
if (new_mtu < MIN_MTU || new_mtu > MAX_MTU)
|
|
|
|
return -EINVAL;
|
|
|
|
dev->mtu = new_mtu;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2009-01-06 11:44:22 -07:00
|
|
|
static const struct net_device_ops virtnet_netdev = {
|
|
|
|
.ndo_open = virtnet_open,
|
|
|
|
.ndo_stop = virtnet_close,
|
|
|
|
.ndo_start_xmit = start_xmit,
|
|
|
|
.ndo_validate_addr = eth_validate_addr,
|
2009-02-04 17:36:34 -07:00
|
|
|
.ndo_set_mac_address = virtnet_set_mac_address,
|
2009-02-04 02:02:40 -07:00
|
|
|
.ndo_set_rx_mode = virtnet_set_rx_mode,
|
2009-01-06 11:44:22 -07:00
|
|
|
.ndo_change_mtu = virtnet_change_mtu,
|
2009-05-01 11:31:10 -06:00
|
|
|
.ndo_vlan_rx_add_vid = virtnet_vlan_rx_add_vid,
|
|
|
|
.ndo_vlan_rx_kill_vid = virtnet_vlan_rx_kill_vid,
|
2009-01-06 11:44:22 -07:00
|
|
|
#ifdef CONFIG_NET_POLL_CONTROLLER
|
|
|
|
.ndo_poll_controller = virtnet_netpoll,
|
|
|
|
#endif
|
|
|
|
};
|
|
|
|
|
2009-01-19 18:09:49 -07:00
|
|
|
static void virtnet_update_status(struct virtnet_info *vi)
|
|
|
|
{
|
|
|
|
u16 v;
|
|
|
|
|
|
|
|
if (!virtio_has_feature(vi->vdev, VIRTIO_NET_F_STATUS))
|
|
|
|
return;
|
|
|
|
|
|
|
|
vi->vdev->config->get(vi->vdev,
|
|
|
|
offsetof(struct virtio_net_config, status),
|
|
|
|
&v, sizeof(v));
|
|
|
|
|
|
|
|
/* Ignore unknown (future) status bits */
|
|
|
|
v &= VIRTIO_NET_S_LINK_UP;
|
|
|
|
|
|
|
|
if (vi->status == v)
|
|
|
|
return;
|
|
|
|
|
|
|
|
vi->status = v;
|
|
|
|
|
|
|
|
if (vi->status & VIRTIO_NET_S_LINK_UP) {
|
|
|
|
netif_carrier_on(vi->dev);
|
|
|
|
netif_wake_queue(vi->dev);
|
|
|
|
} else {
|
|
|
|
netif_carrier_off(vi->dev);
|
|
|
|
netif_stop_queue(vi->dev);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void virtnet_config_changed(struct virtio_device *vdev)
|
|
|
|
{
|
|
|
|
struct virtnet_info *vi = vdev->priv;
|
|
|
|
|
|
|
|
virtnet_update_status(vi);
|
|
|
|
}
|
|
|
|
|
2007-10-21 19:03:37 -06:00
|
|
|
static int virtnet_probe(struct virtio_device *vdev)
|
|
|
|
{
|
|
|
|
int err;
|
|
|
|
struct net_device *dev;
|
|
|
|
struct virtnet_info *vi;
|
2009-06-12 22:16:36 -06:00
|
|
|
struct virtqueue *vqs[3];
|
|
|
|
vq_callback_t *callbacks[] = { skb_recv_done, skb_xmit_done, NULL};
|
|
|
|
const char *names[] = { "input", "output", "control" };
|
|
|
|
int nvqs;
|
2007-10-21 19:03:37 -06:00
|
|
|
|
|
|
|
/* Allocate ourselves a network device with room for our info */
|
|
|
|
dev = alloc_etherdev(sizeof(struct virtnet_info));
|
|
|
|
if (!dev)
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
/* Set up network device as normal. */
|
2009-01-06 11:44:22 -07:00
|
|
|
dev->netdev_ops = &virtnet_netdev;
|
2007-10-21 19:03:37 -06:00
|
|
|
dev->features = NETIF_F_HIGHDMA;
|
2008-04-17 21:21:42 -06:00
|
|
|
SET_ETHTOOL_OPS(dev, &virtnet_ethtool_ops);
|
2007-10-21 19:03:37 -06:00
|
|
|
SET_NETDEV_DEV(dev, &vdev->dev);
|
|
|
|
|
|
|
|
/* Do we support "hardware" checksums? */
|
2008-05-02 20:50:50 -06:00
|
|
|
if (csum && virtio_has_feature(vdev, VIRTIO_NET_F_CSUM)) {
|
2007-10-21 19:03:37 -06:00
|
|
|
/* This opens up the world of extra features. */
|
|
|
|
dev->features |= NETIF_F_HW_CSUM|NETIF_F_SG|NETIF_F_FRAGLIST;
|
2008-05-02 20:50:50 -06:00
|
|
|
if (gso && virtio_has_feature(vdev, VIRTIO_NET_F_GSO)) {
|
2008-02-04 21:50:02 -07:00
|
|
|
dev->features |= NETIF_F_TSO | NETIF_F_UFO
|
|
|
|
| NETIF_F_TSO_ECN | NETIF_F_TSO6;
|
|
|
|
}
|
2008-05-02 20:50:46 -06:00
|
|
|
/* Individual feature bits: what can host handle? */
|
2008-05-02 20:50:50 -06:00
|
|
|
if (gso && virtio_has_feature(vdev, VIRTIO_NET_F_HOST_TSO4))
|
2008-05-02 20:50:46 -06:00
|
|
|
dev->features |= NETIF_F_TSO;
|
2008-05-02 20:50:50 -06:00
|
|
|
if (gso && virtio_has_feature(vdev, VIRTIO_NET_F_HOST_TSO6))
|
2008-05-02 20:50:46 -06:00
|
|
|
dev->features |= NETIF_F_TSO6;
|
2008-05-02 20:50:50 -06:00
|
|
|
if (gso && virtio_has_feature(vdev, VIRTIO_NET_F_HOST_ECN))
|
2008-05-02 20:50:46 -06:00
|
|
|
dev->features |= NETIF_F_TSO_ECN;
|
2008-05-02 20:50:50 -06:00
|
|
|
if (gso && virtio_has_feature(vdev, VIRTIO_NET_F_HOST_UFO))
|
2008-05-02 20:50:46 -06:00
|
|
|
dev->features |= NETIF_F_UFO;
|
2007-10-21 19:03:37 -06:00
|
|
|
}
|
|
|
|
|
|
|
|
/* Configuration may specify what MAC to use. Otherwise random. */
|
2008-05-02 20:50:50 -06:00
|
|
|
if (virtio_has_feature(vdev, VIRTIO_NET_F_MAC)) {
|
2008-02-04 21:49:56 -07:00
|
|
|
vdev->config->get(vdev,
|
|
|
|
offsetof(struct virtio_net_config, mac),
|
|
|
|
dev->dev_addr, dev->addr_len);
|
2009-04-04 17:40:19 -06:00
|
|
|
} else
|
2007-10-21 19:03:37 -06:00
|
|
|
random_ether_addr(dev->dev_addr);
|
|
|
|
|
|
|
|
/* Set up our device-specific information */
|
|
|
|
vi = netdev_priv(dev);
|
2007-12-16 06:19:43 -07:00
|
|
|
netif_napi_add(dev, &vi->napi, virtnet_poll, napi_weight);
|
2007-10-21 19:03:37 -06:00
|
|
|
vi->dev = dev;
|
|
|
|
vi->vdev = vdev;
|
2008-02-18 02:02:51 -07:00
|
|
|
vdev->priv = vi;
|
2008-07-25 11:06:01 -06:00
|
|
|
vi->pages = NULL;
|
2009-08-26 13:22:32 -06:00
|
|
|
INIT_DELAYED_WORK(&vi->refill, refill_work);
|
2007-10-21 19:03:37 -06:00
|
|
|
|
2008-04-17 21:24:27 -06:00
|
|
|
/* If we can receive ANY GSO packets, we must allocate large ones. */
|
|
|
|
if (virtio_has_feature(vdev, VIRTIO_NET_F_GUEST_TSO4)
|
|
|
|
|| virtio_has_feature(vdev, VIRTIO_NET_F_GUEST_TSO6)
|
|
|
|
|| virtio_has_feature(vdev, VIRTIO_NET_F_GUEST_ECN))
|
|
|
|
vi->big_packets = true;
|
|
|
|
|
2008-11-16 23:41:34 -07:00
|
|
|
if (virtio_has_feature(vdev, VIRTIO_NET_F_MRG_RXBUF))
|
|
|
|
vi->mergeable_rx_bufs = true;
|
|
|
|
|
2009-06-12 22:16:36 -06:00
|
|
|
/* We expect two virtqueues, receive then send,
|
|
|
|
* and optionally control. */
|
|
|
|
nvqs = virtio_has_feature(vi->vdev, VIRTIO_NET_F_CTRL_VQ) ? 3 : 2;
|
|
|
|
|
|
|
|
err = vdev->config->find_vqs(vdev, nvqs, vqs, callbacks, names);
|
|
|
|
if (err)
|
2007-10-21 19:03:37 -06:00
|
|
|
goto free;
|
|
|
|
|
2009-06-12 22:16:36 -06:00
|
|
|
vi->rvq = vqs[0];
|
|
|
|
vi->svq = vqs[1];
|
2007-10-21 19:03:37 -06:00
|
|
|
|
2009-02-04 02:02:34 -07:00
|
|
|
if (virtio_has_feature(vi->vdev, VIRTIO_NET_F_CTRL_VQ)) {
|
2009-06-12 22:16:36 -06:00
|
|
|
vi->cvq = vqs[2];
|
2009-02-04 02:02:50 -07:00
|
|
|
|
|
|
|
if (virtio_has_feature(vi->vdev, VIRTIO_NET_F_CTRL_VLAN))
|
|
|
|
dev->features |= NETIF_F_HW_VLAN_FILTER;
|
2009-02-04 02:02:34 -07:00
|
|
|
}
|
|
|
|
|
2007-10-21 19:03:37 -06:00
|
|
|
/* Initialize our empty receive and send queues. */
|
|
|
|
skb_queue_head_init(&vi->recv);
|
|
|
|
skb_queue_head_init(&vi->send);
|
|
|
|
|
|
|
|
err = register_netdev(dev);
|
|
|
|
if (err) {
|
|
|
|
pr_debug("virtio_net: registering device failed\n");
|
2009-06-12 22:16:36 -06:00
|
|
|
goto free_vqs;
|
2007-10-21 19:03:37 -06:00
|
|
|
}
|
2008-02-04 21:50:02 -07:00
|
|
|
|
|
|
|
/* Last of all, set up some receive buffers. */
|
2009-08-26 13:22:32 -06:00
|
|
|
try_fill_recv(vi, GFP_KERNEL);
|
2008-02-04 21:50:02 -07:00
|
|
|
|
|
|
|
/* If we didn't even get one input buffer, we're useless. */
|
|
|
|
if (vi->num == 0) {
|
|
|
|
err = -ENOMEM;
|
|
|
|
goto unregister;
|
|
|
|
}
|
|
|
|
|
2009-01-19 18:09:49 -07:00
|
|
|
vi->status = VIRTIO_NET_S_LINK_UP;
|
|
|
|
virtnet_update_status(vi);
|
2009-03-18 19:40:02 -06:00
|
|
|
netif_carrier_on(dev);
|
2009-01-19 18:09:49 -07:00
|
|
|
|
2007-10-21 19:03:37 -06:00
|
|
|
pr_debug("virtnet: registered device %s\n", dev->name);
|
|
|
|
return 0;
|
|
|
|
|
2008-02-04 21:50:02 -07:00
|
|
|
unregister:
|
|
|
|
unregister_netdev(dev);
|
2009-08-26 13:22:32 -06:00
|
|
|
cancel_delayed_work_sync(&vi->refill);
|
2009-06-12 22:16:36 -06:00
|
|
|
free_vqs:
|
|
|
|
vdev->config->del_vqs(vdev);
|
2007-10-21 19:03:37 -06:00
|
|
|
free:
|
|
|
|
free_netdev(dev);
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
2009-09-30 16:28:34 -06:00
|
|
|
static void __devexit virtnet_remove(struct virtio_device *vdev)
|
2007-10-21 19:03:37 -06:00
|
|
|
{
|
2007-11-19 09:20:42 -07:00
|
|
|
struct virtnet_info *vi = vdev->priv;
|
2008-02-04 21:50:02 -07:00
|
|
|
struct sk_buff *skb;
|
|
|
|
|
2008-02-04 21:50:03 -07:00
|
|
|
/* Stop all the virtqueues. */
|
|
|
|
vdev->config->reset(vdev);
|
|
|
|
|
2008-02-04 21:50:02 -07:00
|
|
|
/* Free our skbs in send and recv queues, if any. */
|
|
|
|
while ((skb = __skb_dequeue(&vi->recv)) != NULL) {
|
|
|
|
kfree_skb(skb);
|
|
|
|
vi->num--;
|
|
|
|
}
|
2008-05-22 04:07:43 -06:00
|
|
|
__skb_queue_purge(&vi->send);
|
2008-02-04 21:50:02 -07:00
|
|
|
|
|
|
|
BUG_ON(vi->num != 0);
|
2007-11-19 09:20:42 -07:00
|
|
|
|
|
|
|
unregister_netdev(vi->dev);
|
2009-08-26 13:22:32 -06:00
|
|
|
cancel_delayed_work_sync(&vi->refill);
|
2008-07-25 11:06:01 -06:00
|
|
|
|
2009-06-12 22:16:36 -06:00
|
|
|
vdev->config->del_vqs(vi->vdev);
|
|
|
|
|
2008-07-25 11:06:01 -06:00
|
|
|
while (vi->pages)
|
|
|
|
__free_pages(get_a_page(vi, GFP_KERNEL), 0);
|
|
|
|
|
2007-11-19 09:20:42 -07:00
|
|
|
free_netdev(vi->dev);
|
2007-10-21 19:03:37 -06:00
|
|
|
}
|
|
|
|
|
|
|
|
static struct virtio_device_id id_table[] = {
|
|
|
|
{ VIRTIO_ID_NET, VIRTIO_DEV_ANY_ID },
|
|
|
|
{ 0 },
|
|
|
|
};
|
|
|
|
|
2008-05-02 20:50:50 -06:00
|
|
|
static unsigned int features[] = {
|
2008-07-08 01:10:42 -06:00
|
|
|
VIRTIO_NET_F_CSUM, VIRTIO_NET_F_GUEST_CSUM,
|
|
|
|
VIRTIO_NET_F_GSO, VIRTIO_NET_F_MAC,
|
2008-05-02 20:50:50 -06:00
|
|
|
VIRTIO_NET_F_HOST_TSO4, VIRTIO_NET_F_HOST_UFO, VIRTIO_NET_F_HOST_TSO6,
|
2008-04-17 21:24:27 -06:00
|
|
|
VIRTIO_NET_F_HOST_ECN, VIRTIO_NET_F_GUEST_TSO4, VIRTIO_NET_F_GUEST_TSO6,
|
2009-07-14 08:21:02 -06:00
|
|
|
VIRTIO_NET_F_GUEST_ECN, VIRTIO_NET_F_GUEST_UFO,
|
2009-02-04 02:02:34 -07:00
|
|
|
VIRTIO_NET_F_MRG_RXBUF, VIRTIO_NET_F_STATUS, VIRTIO_NET_F_CTRL_VQ,
|
2009-02-04 02:02:50 -07:00
|
|
|
VIRTIO_NET_F_CTRL_RX, VIRTIO_NET_F_CTRL_VLAN,
|
2008-05-02 20:50:50 -06:00
|
|
|
};
|
|
|
|
|
2007-10-21 19:03:37 -06:00
|
|
|
static struct virtio_driver virtio_net = {
|
2008-05-02 20:50:50 -06:00
|
|
|
.feature_table = features,
|
|
|
|
.feature_table_size = ARRAY_SIZE(features),
|
2007-10-21 19:03:37 -06:00
|
|
|
.driver.name = KBUILD_MODNAME,
|
|
|
|
.driver.owner = THIS_MODULE,
|
|
|
|
.id_table = id_table,
|
|
|
|
.probe = virtnet_probe,
|
|
|
|
.remove = __devexit_p(virtnet_remove),
|
2009-01-19 18:09:49 -07:00
|
|
|
.config_changed = virtnet_config_changed,
|
2007-10-21 19:03:37 -06:00
|
|
|
};
|
|
|
|
|
|
|
|
static int __init init(void)
|
|
|
|
{
|
|
|
|
return register_virtio_driver(&virtio_net);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void __exit fini(void)
|
|
|
|
{
|
|
|
|
unregister_virtio_driver(&virtio_net);
|
|
|
|
}
|
|
|
|
module_init(init);
|
|
|
|
module_exit(fini);
|
|
|
|
|
|
|
|
MODULE_DEVICE_TABLE(virtio, id_table);
|
|
|
|
MODULE_DESCRIPTION("Virtio network driver");
|
|
|
|
MODULE_LICENSE("GPL");
|