2011-03-15 08:06:18 +08:00
|
|
|
/*
|
|
|
|
* Network-device interface management.
|
|
|
|
*
|
|
|
|
* Copyright (c) 2004-2005, Keir Fraser
|
|
|
|
*
|
|
|
|
* This program is free software; you can redistribute it and/or
|
|
|
|
* modify it under the terms of the GNU General Public License version 2
|
|
|
|
* as published by the Free Software Foundation; or, when distributed
|
|
|
|
* separately from the Linux kernel or incorporated into other
|
|
|
|
* software packages, subject to the following license:
|
|
|
|
*
|
|
|
|
* Permission is hereby granted, free of charge, to any person obtaining a copy
|
|
|
|
* of this source file (the "Software"), to deal in the Software without
|
|
|
|
* restriction, including without limitation the rights to use, copy, modify,
|
|
|
|
* merge, publish, distribute, sublicense, and/or sell copies of the Software,
|
|
|
|
* and to permit persons to whom the Software is furnished to do so, subject to
|
|
|
|
* the following conditions:
|
|
|
|
*
|
|
|
|
* The above copyright notice and this permission notice shall be included in
|
|
|
|
* all copies or substantial portions of the Software.
|
|
|
|
*
|
|
|
|
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
|
|
|
|
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
|
|
|
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
|
|
|
|
* AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
|
|
|
|
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
|
|
|
|
* FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
|
|
|
|
* IN THE SOFTWARE.
|
|
|
|
*/
|
|
|
|
|
|
|
|
#include "common.h"
|
|
|
|
|
2013-08-26 19:59:38 +08:00
|
|
|
#include <linux/kthread.h>
|
2011-03-15 08:06:18 +08:00
|
|
|
#include <linux/ethtool.h>
|
|
|
|
#include <linux/rtnetlink.h>
|
|
|
|
#include <linux/if_vlan.h>
|
2014-06-10 16:34:36 +08:00
|
|
|
#include <linux/vmalloc.h>
|
2011-03-15 08:06:18 +08:00
|
|
|
|
|
|
|
#include <xen/events.h>
|
|
|
|
#include <asm/xen/hypercall.h>
|
2014-03-07 05:48:26 +08:00
|
|
|
#include <xen/balloon.h>
|
2011-03-15 08:06:18 +08:00
|
|
|
|
|
|
|
#define XENVIF_QUEUE_LENGTH 32
|
2013-08-26 19:59:38 +08:00
|
|
|
#define XENVIF_NAPI_WEIGHT 64
|
2011-03-15 08:06:18 +08:00
|
|
|
|
2014-10-22 21:08:54 +08:00
|
|
|
/* Number of bytes allowed on the internal guest Rx queue. */
|
|
|
|
#define XENVIF_RX_QUEUE_BYTES (XEN_NETIF_RX_RING_SIZE/2 * PAGE_SIZE)
|
|
|
|
|
2014-08-12 18:48:07 +08:00
|
|
|
/* This function is used to set SKBTX_DEV_ZEROCOPY as well as
|
|
|
|
* increasing the inflight counter. We need to increase the inflight
|
|
|
|
* counter because core driver calls into xenvif_zerocopy_callback
|
|
|
|
* which calls xenvif_skb_zerocopy_complete.
|
|
|
|
*/
|
|
|
|
void xenvif_skb_zerocopy_prepare(struct xenvif_queue *queue,
|
|
|
|
struct sk_buff *skb)
|
|
|
|
{
|
|
|
|
skb_shinfo(skb)->tx_flags |= SKBTX_DEV_ZEROCOPY;
|
|
|
|
atomic_inc(&queue->inflight_packets);
|
|
|
|
}
|
|
|
|
|
|
|
|
void xenvif_skb_zerocopy_complete(struct xenvif_queue *queue)
|
|
|
|
{
|
|
|
|
atomic_dec(&queue->inflight_packets);
|
|
|
|
}
|
|
|
|
|
2011-03-15 08:06:18 +08:00
|
|
|
int xenvif_schedulable(struct xenvif *vif)
|
|
|
|
{
|
2014-08-04 23:20:57 +08:00
|
|
|
return netif_running(vif->dev) &&
|
2014-10-22 21:08:54 +08:00
|
|
|
test_bit(VIF_STATUS_CONNECTED, &vif->status) &&
|
|
|
|
!vif->disabled;
|
2011-03-15 08:06:18 +08:00
|
|
|
}
|
|
|
|
|
2013-05-22 14:34:45 +08:00
|
|
|
static irqreturn_t xenvif_tx_interrupt(int irq, void *dev_id)
|
2011-03-15 08:06:18 +08:00
|
|
|
{
|
2014-06-04 17:30:42 +08:00
|
|
|
struct xenvif_queue *queue = dev_id;
|
2011-03-15 08:06:18 +08:00
|
|
|
|
2014-06-04 17:30:42 +08:00
|
|
|
if (RING_HAS_UNCONSUMED_REQUESTS(&queue->tx))
|
|
|
|
napi_schedule(&queue->napi);
|
2011-03-15 08:06:18 +08:00
|
|
|
|
2013-05-22 14:34:45 +08:00
|
|
|
return IRQ_HANDLED;
|
|
|
|
}
|
|
|
|
|
2014-06-04 17:30:42 +08:00
|
|
|
int xenvif_poll(struct napi_struct *napi, int budget)
|
2013-08-26 19:59:38 +08:00
|
|
|
{
|
2014-06-04 17:30:42 +08:00
|
|
|
struct xenvif_queue *queue =
|
|
|
|
container_of(napi, struct xenvif_queue, napi);
|
2013-08-26 19:59:38 +08:00
|
|
|
int work_done;
|
|
|
|
|
2014-04-01 19:46:12 +08:00
|
|
|
/* This vif is rogue, we pretend we've there is nothing to do
|
|
|
|
* for this vif to deschedule it from NAPI. But this interface
|
|
|
|
* will be turned off in thread context later.
|
|
|
|
*/
|
2014-08-11 20:01:44 +08:00
|
|
|
if (unlikely(queue->vif->disabled)) {
|
2014-04-01 19:46:12 +08:00
|
|
|
napi_complete(napi);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2014-06-04 17:30:42 +08:00
|
|
|
work_done = xenvif_tx_action(queue, budget);
|
2013-08-26 19:59:38 +08:00
|
|
|
|
|
|
|
if (work_done < budget) {
|
2014-05-16 19:26:04 +08:00
|
|
|
napi_complete(napi);
|
2014-06-04 17:30:42 +08:00
|
|
|
xenvif_napi_schedule_or_enable_events(queue);
|
2013-08-26 19:59:38 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
return work_done;
|
|
|
|
}
|
|
|
|
|
2013-05-22 14:34:45 +08:00
|
|
|
static irqreturn_t xenvif_rx_interrupt(int irq, void *dev_id)
|
|
|
|
{
|
2014-06-04 17:30:42 +08:00
|
|
|
struct xenvif_queue *queue = dev_id;
|
2013-05-22 14:34:45 +08:00
|
|
|
|
2014-06-04 17:30:42 +08:00
|
|
|
xenvif_kick_thread(queue);
|
2011-03-15 08:06:18 +08:00
|
|
|
|
|
|
|
return IRQ_HANDLED;
|
|
|
|
}
|
|
|
|
|
2014-07-09 02:49:14 +08:00
|
|
|
irqreturn_t xenvif_interrupt(int irq, void *dev_id)
|
2013-05-22 14:34:45 +08:00
|
|
|
{
|
|
|
|
xenvif_tx_interrupt(irq, dev_id);
|
|
|
|
xenvif_rx_interrupt(irq, dev_id);
|
|
|
|
|
|
|
|
return IRQ_HANDLED;
|
|
|
|
}
|
|
|
|
|
2014-06-04 17:30:42 +08:00
|
|
|
int xenvif_queue_stopped(struct xenvif_queue *queue)
|
|
|
|
{
|
|
|
|
struct net_device *dev = queue->vif->dev;
|
|
|
|
unsigned int id = queue->id;
|
|
|
|
return netif_tx_queue_stopped(netdev_get_tx_queue(dev, id));
|
|
|
|
}
|
|
|
|
|
|
|
|
void xenvif_wake_queue(struct xenvif_queue *queue)
|
|
|
|
{
|
|
|
|
struct net_device *dev = queue->vif->dev;
|
|
|
|
unsigned int id = queue->id;
|
|
|
|
netif_tx_wake_queue(netdev_get_tx_queue(dev, id));
|
|
|
|
}
|
|
|
|
|
2011-03-15 08:06:18 +08:00
|
|
|
static int xenvif_start_xmit(struct sk_buff *skb, struct net_device *dev)
|
|
|
|
{
|
|
|
|
struct xenvif *vif = netdev_priv(dev);
|
2014-06-04 17:30:42 +08:00
|
|
|
struct xenvif_queue *queue = NULL;
|
2014-06-23 17:50:17 +08:00
|
|
|
unsigned int num_queues = vif->num_queues;
|
2014-06-04 17:30:42 +08:00
|
|
|
u16 index;
|
2014-10-22 21:08:54 +08:00
|
|
|
struct xenvif_rx_cb *cb;
|
2011-03-15 08:06:18 +08:00
|
|
|
|
|
|
|
BUG_ON(skb->dev != dev);
|
|
|
|
|
2014-06-04 17:30:42 +08:00
|
|
|
/* Drop the packet if queues are not set up */
|
|
|
|
if (num_queues < 1)
|
|
|
|
goto drop;
|
|
|
|
|
|
|
|
/* Obtain the queue to be used to transmit this packet */
|
|
|
|
index = skb_get_queue_mapping(skb);
|
|
|
|
if (index >= num_queues) {
|
|
|
|
pr_warn_ratelimited("Invalid queue %hu for packet on interface %s\n.",
|
|
|
|
index, vif->dev->name);
|
|
|
|
index %= num_queues;
|
|
|
|
}
|
|
|
|
queue = &vif->queues[index];
|
|
|
|
|
|
|
|
/* Drop the packet if queue is not ready */
|
|
|
|
if (queue->task == NULL ||
|
|
|
|
queue->dealloc_task == NULL ||
|
2014-03-07 05:48:26 +08:00
|
|
|
!xenvif_schedulable(vif))
|
2011-03-15 08:06:18 +08:00
|
|
|
goto drop;
|
|
|
|
|
2014-10-22 21:08:54 +08:00
|
|
|
cb = XENVIF_RX_CB(skb);
|
|
|
|
cb->expires = jiffies + rx_drain_timeout_jiffies;
|
2011-03-15 08:06:18 +08:00
|
|
|
|
2014-10-22 21:08:54 +08:00
|
|
|
xenvif_rx_queue_tail(queue, skb);
|
2014-06-04 17:30:42 +08:00
|
|
|
xenvif_kick_thread(queue);
|
2011-03-15 08:06:18 +08:00
|
|
|
|
|
|
|
return NETDEV_TX_OK;
|
|
|
|
|
|
|
|
drop:
|
|
|
|
vif->dev->stats.tx_dropped++;
|
|
|
|
dev_kfree_skb(skb);
|
|
|
|
return NETDEV_TX_OK;
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct net_device_stats *xenvif_get_stats(struct net_device *dev)
|
|
|
|
{
|
|
|
|
struct xenvif *vif = netdev_priv(dev);
|
2014-06-04 17:30:42 +08:00
|
|
|
struct xenvif_queue *queue = NULL;
|
2014-06-23 17:50:17 +08:00
|
|
|
unsigned int num_queues = vif->num_queues;
|
2014-06-04 17:30:42 +08:00
|
|
|
unsigned long rx_bytes = 0;
|
|
|
|
unsigned long rx_packets = 0;
|
|
|
|
unsigned long tx_bytes = 0;
|
|
|
|
unsigned long tx_packets = 0;
|
|
|
|
unsigned int index;
|
|
|
|
|
|
|
|
if (vif->queues == NULL)
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
/* Aggregate tx and rx stats from each queue */
|
|
|
|
for (index = 0; index < num_queues; ++index) {
|
|
|
|
queue = &vif->queues[index];
|
|
|
|
rx_bytes += queue->stats.rx_bytes;
|
|
|
|
rx_packets += queue->stats.rx_packets;
|
|
|
|
tx_bytes += queue->stats.tx_bytes;
|
|
|
|
tx_packets += queue->stats.tx_packets;
|
|
|
|
}
|
|
|
|
|
|
|
|
out:
|
|
|
|
vif->dev->stats.rx_bytes = rx_bytes;
|
|
|
|
vif->dev->stats.rx_packets = rx_packets;
|
|
|
|
vif->dev->stats.tx_bytes = tx_bytes;
|
|
|
|
vif->dev->stats.tx_packets = tx_packets;
|
|
|
|
|
2011-03-15 08:06:18 +08:00
|
|
|
return &vif->dev->stats;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void xenvif_up(struct xenvif *vif)
|
|
|
|
{
|
2014-06-04 17:30:42 +08:00
|
|
|
struct xenvif_queue *queue = NULL;
|
2014-06-23 17:50:17 +08:00
|
|
|
unsigned int num_queues = vif->num_queues;
|
2014-06-04 17:30:42 +08:00
|
|
|
unsigned int queue_index;
|
|
|
|
|
|
|
|
for (queue_index = 0; queue_index < num_queues; ++queue_index) {
|
|
|
|
queue = &vif->queues[queue_index];
|
|
|
|
napi_enable(&queue->napi);
|
|
|
|
enable_irq(queue->tx_irq);
|
|
|
|
if (queue->tx_irq != queue->rx_irq)
|
|
|
|
enable_irq(queue->rx_irq);
|
|
|
|
xenvif_napi_schedule_or_enable_events(queue);
|
|
|
|
}
|
2011-03-15 08:06:18 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
static void xenvif_down(struct xenvif *vif)
|
|
|
|
{
|
2014-06-04 17:30:42 +08:00
|
|
|
struct xenvif_queue *queue = NULL;
|
2014-06-23 17:50:17 +08:00
|
|
|
unsigned int num_queues = vif->num_queues;
|
2014-06-04 17:30:42 +08:00
|
|
|
unsigned int queue_index;
|
|
|
|
|
|
|
|
for (queue_index = 0; queue_index < num_queues; ++queue_index) {
|
|
|
|
queue = &vif->queues[queue_index];
|
|
|
|
napi_disable(&queue->napi);
|
|
|
|
disable_irq(queue->tx_irq);
|
|
|
|
if (queue->tx_irq != queue->rx_irq)
|
|
|
|
disable_irq(queue->rx_irq);
|
|
|
|
del_timer_sync(&queue->credit_timeout);
|
|
|
|
}
|
2011-03-15 08:06:18 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
static int xenvif_open(struct net_device *dev)
|
|
|
|
{
|
|
|
|
struct xenvif *vif = netdev_priv(dev);
|
2014-08-04 23:20:57 +08:00
|
|
|
if (test_bit(VIF_STATUS_CONNECTED, &vif->status))
|
2011-03-15 08:06:18 +08:00
|
|
|
xenvif_up(vif);
|
2014-06-04 17:30:42 +08:00
|
|
|
netif_tx_start_all_queues(dev);
|
2011-03-15 08:06:18 +08:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int xenvif_close(struct net_device *dev)
|
|
|
|
{
|
|
|
|
struct xenvif *vif = netdev_priv(dev);
|
2014-08-04 23:20:57 +08:00
|
|
|
if (test_bit(VIF_STATUS_CONNECTED, &vif->status))
|
2011-03-15 08:06:18 +08:00
|
|
|
xenvif_down(vif);
|
2014-06-04 17:30:42 +08:00
|
|
|
netif_tx_stop_all_queues(dev);
|
2011-03-15 08:06:18 +08:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int xenvif_change_mtu(struct net_device *dev, int mtu)
|
|
|
|
{
|
|
|
|
struct xenvif *vif = netdev_priv(dev);
|
|
|
|
int max = vif->can_sg ? 65535 - VLAN_ETH_HLEN : ETH_DATA_LEN;
|
|
|
|
|
|
|
|
if (mtu > max)
|
|
|
|
return -EINVAL;
|
|
|
|
dev->mtu = mtu;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2011-11-15 23:29:55 +08:00
|
|
|
static netdev_features_t xenvif_fix_features(struct net_device *dev,
|
|
|
|
netdev_features_t features)
|
2011-03-15 08:06:18 +08:00
|
|
|
{
|
|
|
|
struct xenvif *vif = netdev_priv(dev);
|
|
|
|
|
2011-04-19 11:35:06 +08:00
|
|
|
if (!vif->can_sg)
|
|
|
|
features &= ~NETIF_F_SG;
|
2013-10-17 00:50:32 +08:00
|
|
|
if (~(vif->gso_mask | vif->gso_prefix_mask) & GSO_BIT(TCPV4))
|
2011-04-19 11:35:06 +08:00
|
|
|
features &= ~NETIF_F_TSO;
|
2013-10-17 00:50:32 +08:00
|
|
|
if (~(vif->gso_mask | vif->gso_prefix_mask) & GSO_BIT(TCPV6))
|
|
|
|
features &= ~NETIF_F_TSO6;
|
2013-10-17 00:50:28 +08:00
|
|
|
if (!vif->ip_csum)
|
2011-04-19 11:35:06 +08:00
|
|
|
features &= ~NETIF_F_IP_CSUM;
|
2013-10-17 00:50:28 +08:00
|
|
|
if (!vif->ipv6_csum)
|
|
|
|
features &= ~NETIF_F_IPV6_CSUM;
|
2011-03-15 08:06:18 +08:00
|
|
|
|
2011-04-19 11:35:06 +08:00
|
|
|
return features;
|
2011-03-15 08:06:18 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
static const struct xenvif_stat {
|
|
|
|
char name[ETH_GSTRING_LEN];
|
|
|
|
u16 offset;
|
|
|
|
} xenvif_stats[] = {
|
|
|
|
{
|
|
|
|
"rx_gso_checksum_fixup",
|
2014-06-04 17:30:42 +08:00
|
|
|
offsetof(struct xenvif_stats, rx_gso_checksum_fixup)
|
2011-03-15 08:06:18 +08:00
|
|
|
},
|
2014-03-07 05:48:28 +08:00
|
|
|
/* If (sent != success + fail), there are probably packets never
|
|
|
|
* freed up properly!
|
|
|
|
*/
|
|
|
|
{
|
|
|
|
"tx_zerocopy_sent",
|
2014-06-04 17:30:42 +08:00
|
|
|
offsetof(struct xenvif_stats, tx_zerocopy_sent),
|
2014-03-07 05:48:28 +08:00
|
|
|
},
|
|
|
|
{
|
|
|
|
"tx_zerocopy_success",
|
2014-06-04 17:30:42 +08:00
|
|
|
offsetof(struct xenvif_stats, tx_zerocopy_success),
|
2014-03-07 05:48:28 +08:00
|
|
|
},
|
|
|
|
{
|
|
|
|
"tx_zerocopy_fail",
|
2014-06-04 17:30:42 +08:00
|
|
|
offsetof(struct xenvif_stats, tx_zerocopy_fail)
|
2014-03-07 05:48:28 +08:00
|
|
|
},
|
2014-03-07 05:48:29 +08:00
|
|
|
/* Number of packets exceeding MAX_SKB_FRAG slots. You should use
|
|
|
|
* a guest with the same MAX_SKB_FRAG
|
|
|
|
*/
|
|
|
|
{
|
|
|
|
"tx_frag_overflow",
|
2014-06-04 17:30:42 +08:00
|
|
|
offsetof(struct xenvif_stats, tx_frag_overflow)
|
2014-03-07 05:48:29 +08:00
|
|
|
},
|
2011-03-15 08:06:18 +08:00
|
|
|
};
|
|
|
|
|
|
|
|
static int xenvif_get_sset_count(struct net_device *dev, int string_set)
|
|
|
|
{
|
|
|
|
switch (string_set) {
|
|
|
|
case ETH_SS_STATS:
|
|
|
|
return ARRAY_SIZE(xenvif_stats);
|
|
|
|
default:
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void xenvif_get_ethtool_stats(struct net_device *dev,
|
|
|
|
struct ethtool_stats *stats, u64 * data)
|
|
|
|
{
|
2014-06-04 17:30:42 +08:00
|
|
|
struct xenvif *vif = netdev_priv(dev);
|
2014-06-23 17:50:17 +08:00
|
|
|
unsigned int num_queues = vif->num_queues;
|
2011-03-15 08:06:18 +08:00
|
|
|
int i;
|
2014-06-04 17:30:42 +08:00
|
|
|
unsigned int queue_index;
|
|
|
|
struct xenvif_stats *vif_stats;
|
|
|
|
|
|
|
|
for (i = 0; i < ARRAY_SIZE(xenvif_stats); i++) {
|
|
|
|
unsigned long accum = 0;
|
|
|
|
for (queue_index = 0; queue_index < num_queues; ++queue_index) {
|
|
|
|
vif_stats = &vif->queues[queue_index].stats;
|
|
|
|
accum += *(unsigned long *)(vif_stats + xenvif_stats[i].offset);
|
|
|
|
}
|
|
|
|
data[i] = accum;
|
|
|
|
}
|
2011-03-15 08:06:18 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
static void xenvif_get_strings(struct net_device *dev, u32 stringset, u8 * data)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
|
|
|
|
switch (stringset) {
|
|
|
|
case ETH_SS_STATS:
|
|
|
|
for (i = 0; i < ARRAY_SIZE(xenvif_stats); i++)
|
|
|
|
memcpy(data + i * ETH_GSTRING_LEN,
|
|
|
|
xenvif_stats[i].name, ETH_GSTRING_LEN);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2012-01-04 19:56:58 +08:00
|
|
|
static const struct ethtool_ops xenvif_ethtool_ops = {
|
2011-03-15 08:06:18 +08:00
|
|
|
.get_link = ethtool_op_get_link,
|
|
|
|
|
|
|
|
.get_sset_count = xenvif_get_sset_count,
|
|
|
|
.get_ethtool_stats = xenvif_get_ethtool_stats,
|
|
|
|
.get_strings = xenvif_get_strings,
|
|
|
|
};
|
|
|
|
|
2012-01-04 19:56:58 +08:00
|
|
|
static const struct net_device_ops xenvif_netdev_ops = {
|
2011-03-15 08:06:18 +08:00
|
|
|
.ndo_start_xmit = xenvif_start_xmit,
|
|
|
|
.ndo_get_stats = xenvif_get_stats,
|
|
|
|
.ndo_open = xenvif_open,
|
|
|
|
.ndo_stop = xenvif_close,
|
|
|
|
.ndo_change_mtu = xenvif_change_mtu,
|
2011-04-19 11:35:06 +08:00
|
|
|
.ndo_fix_features = xenvif_fix_features,
|
2013-01-22 16:08:25 +08:00
|
|
|
.ndo_set_mac_address = eth_mac_addr,
|
|
|
|
.ndo_validate_addr = eth_validate_addr,
|
2011-03-15 08:06:18 +08:00
|
|
|
};
|
|
|
|
|
|
|
|
struct xenvif *xenvif_alloc(struct device *parent, domid_t domid,
|
|
|
|
unsigned int handle)
|
|
|
|
{
|
|
|
|
int err;
|
|
|
|
struct net_device *dev;
|
|
|
|
struct xenvif *vif;
|
|
|
|
char name[IFNAMSIZ] = {};
|
|
|
|
|
|
|
|
snprintf(name, IFNAMSIZ - 1, "vif%u.%u", domid, handle);
|
2014-06-04 17:30:43 +08:00
|
|
|
/* Allocate a netdev with the max. supported number of queues.
|
|
|
|
* When the guest selects the desired number, it will be updated
|
2014-06-23 17:50:17 +08:00
|
|
|
* via netif_set_real_num_*_queues().
|
2014-06-04 17:30:43 +08:00
|
|
|
*/
|
net: set name_assign_type in alloc_netdev()
Extend alloc_netdev{,_mq{,s}}() to take name_assign_type as argument, and convert
all users to pass NET_NAME_UNKNOWN.
Coccinelle patch:
@@
expression sizeof_priv, name, setup, txqs, rxqs, count;
@@
(
-alloc_netdev_mqs(sizeof_priv, name, setup, txqs, rxqs)
+alloc_netdev_mqs(sizeof_priv, name, NET_NAME_UNKNOWN, setup, txqs, rxqs)
|
-alloc_netdev_mq(sizeof_priv, name, setup, count)
+alloc_netdev_mq(sizeof_priv, name, NET_NAME_UNKNOWN, setup, count)
|
-alloc_netdev(sizeof_priv, name, setup)
+alloc_netdev(sizeof_priv, name, NET_NAME_UNKNOWN, setup)
)
v9: move comments here from the wrong commit
Signed-off-by: Tom Gundersen <teg@jklm.no>
Reviewed-by: David Herrmann <dh.herrmann@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-07-14 22:37:24 +08:00
|
|
|
dev = alloc_netdev_mq(sizeof(struct xenvif), name, NET_NAME_UNKNOWN,
|
|
|
|
ether_setup, xenvif_max_queues);
|
2011-03-15 08:06:18 +08:00
|
|
|
if (dev == NULL) {
|
2013-08-26 19:59:38 +08:00
|
|
|
pr_warn("Could not allocate netdev for %s\n", name);
|
2011-03-15 08:06:18 +08:00
|
|
|
return ERR_PTR(-ENOMEM);
|
|
|
|
}
|
|
|
|
|
|
|
|
SET_NETDEV_DEV(dev, parent);
|
|
|
|
|
|
|
|
vif = netdev_priv(dev);
|
2013-12-23 17:27:17 +08:00
|
|
|
|
2011-03-15 08:06:18 +08:00
|
|
|
vif->domid = domid;
|
|
|
|
vif->handle = handle;
|
|
|
|
vif->can_sg = 1;
|
2013-10-17 00:50:28 +08:00
|
|
|
vif->ip_csum = 1;
|
2011-03-15 08:06:18 +08:00
|
|
|
vif->dev = dev;
|
2014-04-01 19:46:12 +08:00
|
|
|
vif->disabled = false;
|
|
|
|
|
2014-06-23 17:50:17 +08:00
|
|
|
/* Start out with no queues. */
|
2014-06-04 17:30:42 +08:00
|
|
|
vif->queues = NULL;
|
2014-06-23 17:50:17 +08:00
|
|
|
vif->num_queues = 0;
|
2014-03-07 05:48:30 +08:00
|
|
|
|
2011-03-15 08:06:18 +08:00
|
|
|
dev->netdev_ops = &xenvif_netdev_ops;
|
2013-10-17 00:50:28 +08:00
|
|
|
dev->hw_features = NETIF_F_SG |
|
|
|
|
NETIF_F_IP_CSUM | NETIF_F_IPV6_CSUM |
|
2013-10-17 00:50:32 +08:00
|
|
|
NETIF_F_TSO | NETIF_F_TSO6;
|
2013-10-17 00:50:30 +08:00
|
|
|
dev->features = dev->hw_features | NETIF_F_RXCSUM;
|
2014-05-11 08:12:32 +08:00
|
|
|
dev->ethtool_ops = &xenvif_ethtool_ops;
|
2011-03-15 08:06:18 +08:00
|
|
|
|
|
|
|
dev->tx_queue_len = XENVIF_QUEUE_LENGTH;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Initialise a dummy MAC address. We choose the numerically
|
|
|
|
* largest non-broadcast address to prevent the address getting
|
|
|
|
* stolen by an Ethernet bridge for STP purposes.
|
|
|
|
* (FE:FF:FF:FF:FF:FF)
|
|
|
|
*/
|
|
|
|
memset(dev->dev_addr, 0xFF, ETH_ALEN);
|
|
|
|
dev->dev_addr[0] &= ~0x01;
|
|
|
|
|
|
|
|
netif_carrier_off(dev);
|
|
|
|
|
|
|
|
err = register_netdev(dev);
|
|
|
|
if (err) {
|
|
|
|
netdev_warn(dev, "Could not register device: err=%d\n", err);
|
|
|
|
free_netdev(dev);
|
|
|
|
return ERR_PTR(err);
|
|
|
|
}
|
|
|
|
|
|
|
|
netdev_dbg(dev, "Successfully created xenvif\n");
|
2013-09-18 00:46:08 +08:00
|
|
|
|
|
|
|
__module_get(THIS_MODULE);
|
|
|
|
|
2011-03-15 08:06:18 +08:00
|
|
|
return vif;
|
|
|
|
}
|
|
|
|
|
2014-06-04 17:30:42 +08:00
|
|
|
int xenvif_init_queue(struct xenvif_queue *queue)
|
|
|
|
{
|
|
|
|
int err, i;
|
|
|
|
|
|
|
|
queue->credit_bytes = queue->remaining_credit = ~0UL;
|
|
|
|
queue->credit_usec = 0UL;
|
|
|
|
init_timer(&queue->credit_timeout);
|
|
|
|
queue->credit_window_start = get_jiffies_64();
|
|
|
|
|
2014-10-22 21:08:54 +08:00
|
|
|
queue->rx_queue_max = XENVIF_RX_QUEUE_BYTES;
|
|
|
|
|
2014-06-04 17:30:42 +08:00
|
|
|
skb_queue_head_init(&queue->rx_queue);
|
|
|
|
skb_queue_head_init(&queue->tx_queue);
|
|
|
|
|
|
|
|
queue->pending_cons = 0;
|
|
|
|
queue->pending_prod = MAX_PENDING_REQS;
|
|
|
|
for (i = 0; i < MAX_PENDING_REQS; ++i)
|
|
|
|
queue->pending_ring[i] = i;
|
|
|
|
|
|
|
|
spin_lock_init(&queue->callback_lock);
|
|
|
|
spin_lock_init(&queue->response_lock);
|
|
|
|
|
|
|
|
/* If ballooning is disabled, this will consume real memory, so you
|
|
|
|
* better enable it. The long term solution would be to use just a
|
|
|
|
* bunch of valid page descriptors, without dependency on ballooning
|
|
|
|
*/
|
|
|
|
err = alloc_xenballooned_pages(MAX_PENDING_REQS,
|
|
|
|
queue->mmap_pages,
|
|
|
|
false);
|
|
|
|
if (err) {
|
|
|
|
netdev_err(queue->vif->dev, "Could not reserve mmap_pages\n");
|
|
|
|
return -ENOMEM;
|
|
|
|
}
|
|
|
|
|
|
|
|
for (i = 0; i < MAX_PENDING_REQS; i++) {
|
|
|
|
queue->pending_tx_info[i].callback_struct = (struct ubuf_info)
|
|
|
|
{ .callback = xenvif_zerocopy_callback,
|
|
|
|
.ctx = NULL,
|
|
|
|
.desc = i };
|
|
|
|
queue->grant_tx_handle[i] = NETBACK_INVALID_HANDLE;
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
void xenvif_carrier_on(struct xenvif *vif)
|
|
|
|
{
|
|
|
|
rtnl_lock();
|
|
|
|
if (!vif->can_sg && vif->dev->mtu > ETH_DATA_LEN)
|
|
|
|
dev_set_mtu(vif->dev, ETH_DATA_LEN);
|
|
|
|
netdev_update_features(vif->dev);
|
2014-08-04 23:20:57 +08:00
|
|
|
set_bit(VIF_STATUS_CONNECTED, &vif->status);
|
2014-06-04 17:30:42 +08:00
|
|
|
netif_carrier_on(vif->dev);
|
|
|
|
if (netif_running(vif->dev))
|
|
|
|
xenvif_up(vif);
|
|
|
|
rtnl_unlock();
|
|
|
|
}
|
|
|
|
|
|
|
|
int xenvif_connect(struct xenvif_queue *queue, unsigned long tx_ring_ref,
|
2013-05-22 14:34:45 +08:00
|
|
|
unsigned long rx_ring_ref, unsigned int tx_evtchn,
|
|
|
|
unsigned int rx_evtchn)
|
2011-03-15 08:06:18 +08:00
|
|
|
{
|
2013-12-03 22:06:25 +08:00
|
|
|
struct task_struct *task;
|
2011-03-15 08:06:18 +08:00
|
|
|
int err = -ENOMEM;
|
|
|
|
|
2014-06-04 17:30:42 +08:00
|
|
|
BUG_ON(queue->tx_irq);
|
|
|
|
BUG_ON(queue->task);
|
|
|
|
BUG_ON(queue->dealloc_task);
|
2011-03-15 08:06:18 +08:00
|
|
|
|
2014-06-04 17:30:42 +08:00
|
|
|
err = xenvif_map_frontend_rings(queue, tx_ring_ref, rx_ring_ref);
|
2011-03-15 08:06:18 +08:00
|
|
|
if (err < 0)
|
|
|
|
goto err;
|
|
|
|
|
2014-06-04 17:30:42 +08:00
|
|
|
init_waitqueue_head(&queue->wq);
|
|
|
|
init_waitqueue_head(&queue->dealloc_wq);
|
2014-08-12 18:48:07 +08:00
|
|
|
atomic_set(&queue->inflight_packets, 0);
|
xen-netback: improve guest-receive-side flow control
The way that flow control works without this patch is that, in start_xmit()
the code uses xenvif_count_skb_slots() to predict how many slots
xenvif_gop_skb() will consume and then adds this to a 'req_cons_peek'
counter which it then uses to determine if the shared ring has that amount
of space available by checking whether 'req_prod' has passed that value.
If the ring doesn't have space the tx queue is stopped.
xenvif_gop_skb() will then consume slots and update 'req_cons' and issue
responses, updating 'rsp_prod' as it goes. The frontend will consume those
responses and post new requests, by updating req_prod. So, req_prod chases
req_cons which chases rsp_prod, and can never exceed that value. Thus if
xenvif_count_skb_slots() ever returns a number of slots greater than
xenvif_gop_skb() uses, req_cons_peek will get to a value that req_prod cannot
possibly achieve (since it's limited by the 'real' req_cons) and, if this
happens enough times, req_cons_peek gets more than a ring size ahead of
req_cons and the tx queue then remains stopped forever waiting for an
unachievable amount of space to become available in the ring.
Having two routines trying to calculate the same value is always going to be
fragile, so this patch does away with that. All we essentially need to do is
make sure that we have 'enough stuff' on our internal queue without letting
it build up uncontrollably. So start_xmit() makes a cheap optimistic check
of how much space is needed for an skb and only turns the queue off if that
is unachievable. net_rx_action() is the place where we could do with an
accurate predicition but, since that has proven tricky to calculate, a cheap
worse-case (but not too bad) estimate is all we really need since the only
thing we *must* prevent is xenvif_gop_skb() consuming more slots than are
available.
Without this patch I can trivially stall netback permanently by just doing
a large guest to guest file copy between two Windows Server 2008R2 VMs on a
single host.
Patch tested with frontends in:
- Windows Server 2008R2
- CentOS 6.0
- Debian Squeeze
- Debian Wheezy
- SLES11
Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
Cc: Wei Liu <wei.liu2@citrix.com>
Cc: Ian Campbell <ian.campbell@citrix.com>
Cc: David Vrabel <david.vrabel@citrix.com>
Cc: Annie Li <annie.li@oracle.com>
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Acked-by: Wei Liu <wei.liu2@citrix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2013-12-07 00:36:07 +08:00
|
|
|
|
2014-08-25 23:44:00 +08:00
|
|
|
netif_napi_add(queue->vif->dev, &queue->napi, xenvif_poll,
|
|
|
|
XENVIF_NAPI_WEIGHT);
|
|
|
|
|
2013-05-22 14:34:45 +08:00
|
|
|
if (tx_evtchn == rx_evtchn) {
|
|
|
|
/* feature-split-event-channels == 0 */
|
|
|
|
err = bind_interdomain_evtchn_to_irqhandler(
|
2014-06-04 17:30:42 +08:00
|
|
|
queue->vif->domid, tx_evtchn, xenvif_interrupt, 0,
|
|
|
|
queue->name, queue);
|
2013-05-22 14:34:45 +08:00
|
|
|
if (err < 0)
|
|
|
|
goto err_unmap;
|
2014-06-04 17:30:42 +08:00
|
|
|
queue->tx_irq = queue->rx_irq = err;
|
|
|
|
disable_irq(queue->tx_irq);
|
2013-05-22 14:34:45 +08:00
|
|
|
} else {
|
|
|
|
/* feature-split-event-channels == 1 */
|
2014-06-04 17:30:42 +08:00
|
|
|
snprintf(queue->tx_irq_name, sizeof(queue->tx_irq_name),
|
|
|
|
"%s-tx", queue->name);
|
2013-05-22 14:34:45 +08:00
|
|
|
err = bind_interdomain_evtchn_to_irqhandler(
|
2014-06-04 17:30:42 +08:00
|
|
|
queue->vif->domid, tx_evtchn, xenvif_tx_interrupt, 0,
|
|
|
|
queue->tx_irq_name, queue);
|
2013-05-22 14:34:45 +08:00
|
|
|
if (err < 0)
|
|
|
|
goto err_unmap;
|
2014-06-04 17:30:42 +08:00
|
|
|
queue->tx_irq = err;
|
|
|
|
disable_irq(queue->tx_irq);
|
2013-05-22 14:34:45 +08:00
|
|
|
|
2014-06-04 17:30:42 +08:00
|
|
|
snprintf(queue->rx_irq_name, sizeof(queue->rx_irq_name),
|
|
|
|
"%s-rx", queue->name);
|
2013-05-22 14:34:45 +08:00
|
|
|
err = bind_interdomain_evtchn_to_irqhandler(
|
2014-06-04 17:30:42 +08:00
|
|
|
queue->vif->domid, rx_evtchn, xenvif_rx_interrupt, 0,
|
|
|
|
queue->rx_irq_name, queue);
|
2013-05-22 14:34:45 +08:00
|
|
|
if (err < 0)
|
|
|
|
goto err_tx_unbind;
|
2014-06-04 17:30:42 +08:00
|
|
|
queue->rx_irq = err;
|
|
|
|
disable_irq(queue->rx_irq);
|
2013-05-22 14:34:45 +08:00
|
|
|
}
|
2011-03-15 08:06:18 +08:00
|
|
|
|
2014-03-07 05:48:24 +08:00
|
|
|
task = kthread_create(xenvif_kthread_guest_rx,
|
2014-06-04 17:30:42 +08:00
|
|
|
(void *)queue, "%s-guest-rx", queue->name);
|
2013-12-03 22:06:25 +08:00
|
|
|
if (IS_ERR(task)) {
|
2014-06-04 17:30:42 +08:00
|
|
|
pr_warn("Could not allocate kthread for %s\n", queue->name);
|
2013-12-03 22:06:25 +08:00
|
|
|
err = PTR_ERR(task);
|
2013-08-26 19:59:38 +08:00
|
|
|
goto err_rx_unbind;
|
|
|
|
}
|
2014-06-04 17:30:42 +08:00
|
|
|
queue->task = task;
|
2013-12-03 22:06:25 +08:00
|
|
|
|
2014-03-07 05:48:26 +08:00
|
|
|
task = kthread_create(xenvif_dealloc_kthread,
|
2014-06-04 17:30:42 +08:00
|
|
|
(void *)queue, "%s-dealloc", queue->name);
|
2014-03-07 05:48:26 +08:00
|
|
|
if (IS_ERR(task)) {
|
2014-06-04 17:30:42 +08:00
|
|
|
pr_warn("Could not allocate kthread for %s\n", queue->name);
|
2014-03-07 05:48:26 +08:00
|
|
|
err = PTR_ERR(task);
|
|
|
|
goto err_rx_unbind;
|
|
|
|
}
|
2014-06-04 17:30:42 +08:00
|
|
|
queue->dealloc_task = task;
|
2014-03-07 05:48:26 +08:00
|
|
|
|
2014-06-04 17:30:42 +08:00
|
|
|
wake_up_process(queue->task);
|
|
|
|
wake_up_process(queue->dealloc_task);
|
2013-08-26 19:59:38 +08:00
|
|
|
|
2011-03-15 08:06:18 +08:00
|
|
|
return 0;
|
2013-08-26 19:59:38 +08:00
|
|
|
|
|
|
|
err_rx_unbind:
|
2014-06-04 17:30:42 +08:00
|
|
|
unbind_from_irqhandler(queue->rx_irq, queue);
|
|
|
|
queue->rx_irq = 0;
|
2013-05-22 14:34:45 +08:00
|
|
|
err_tx_unbind:
|
2014-06-04 17:30:42 +08:00
|
|
|
unbind_from_irqhandler(queue->tx_irq, queue);
|
|
|
|
queue->tx_irq = 0;
|
2011-03-15 08:06:18 +08:00
|
|
|
err_unmap:
|
2014-06-04 17:30:42 +08:00
|
|
|
xenvif_unmap_frontend_rings(queue);
|
2011-03-15 08:06:18 +08:00
|
|
|
err:
|
2013-05-17 07:26:11 +08:00
|
|
|
module_put(THIS_MODULE);
|
2011-03-15 08:06:18 +08:00
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
2013-02-07 07:41:35 +08:00
|
|
|
void xenvif_carrier_off(struct xenvif *vif)
|
2011-03-15 08:06:18 +08:00
|
|
|
{
|
|
|
|
struct net_device *dev = vif->dev;
|
2013-02-07 07:41:35 +08:00
|
|
|
|
|
|
|
rtnl_lock();
|
2014-08-04 23:20:57 +08:00
|
|
|
if (test_and_clear_bit(VIF_STATUS_CONNECTED, &vif->status)) {
|
|
|
|
netif_carrier_off(dev); /* discard queued packets */
|
|
|
|
if (netif_running(dev))
|
|
|
|
xenvif_down(vif);
|
|
|
|
}
|
2013-02-07 07:41:35 +08:00
|
|
|
rtnl_unlock();
|
|
|
|
}
|
|
|
|
|
|
|
|
void xenvif_disconnect(struct xenvif *vif)
|
|
|
|
{
|
2014-06-04 17:30:42 +08:00
|
|
|
struct xenvif_queue *queue = NULL;
|
2014-06-23 17:50:17 +08:00
|
|
|
unsigned int num_queues = vif->num_queues;
|
2014-06-04 17:30:42 +08:00
|
|
|
unsigned int queue_index;
|
|
|
|
|
2014-08-04 23:20:57 +08:00
|
|
|
xenvif_carrier_off(vif);
|
2011-03-15 08:06:18 +08:00
|
|
|
|
2014-06-04 17:30:42 +08:00
|
|
|
for (queue_index = 0; queue_index < num_queues; ++queue_index) {
|
|
|
|
queue = &vif->queues[queue_index];
|
2013-11-21 23:26:09 +08:00
|
|
|
|
2014-08-12 18:48:06 +08:00
|
|
|
netif_napi_del(&queue->napi);
|
|
|
|
|
2014-06-04 17:30:42 +08:00
|
|
|
if (queue->task) {
|
|
|
|
kthread_stop(queue->task);
|
|
|
|
queue->task = NULL;
|
|
|
|
}
|
2014-03-07 05:48:26 +08:00
|
|
|
|
2014-06-04 17:30:42 +08:00
|
|
|
if (queue->dealloc_task) {
|
|
|
|
kthread_stop(queue->dealloc_task);
|
|
|
|
queue->dealloc_task = NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (queue->tx_irq) {
|
|
|
|
if (queue->tx_irq == queue->rx_irq)
|
|
|
|
unbind_from_irqhandler(queue->tx_irq, queue);
|
|
|
|
else {
|
|
|
|
unbind_from_irqhandler(queue->tx_irq, queue);
|
|
|
|
unbind_from_irqhandler(queue->rx_irq, queue);
|
|
|
|
}
|
|
|
|
queue->tx_irq = 0;
|
2013-05-22 14:34:45 +08:00
|
|
|
}
|
2011-03-15 08:06:18 +08:00
|
|
|
|
2014-06-04 17:30:42 +08:00
|
|
|
xenvif_unmap_frontend_rings(queue);
|
|
|
|
}
|
2013-09-18 00:46:08 +08:00
|
|
|
}
|
|
|
|
|
2014-06-04 17:30:43 +08:00
|
|
|
/* Reverse the relevant parts of xenvif_init_queue().
|
|
|
|
* Used for queue teardown from xenvif_free(), and on the
|
|
|
|
* error handling paths in xenbus.c:connect().
|
|
|
|
*/
|
|
|
|
void xenvif_deinit_queue(struct xenvif_queue *queue)
|
|
|
|
{
|
|
|
|
free_xenballooned_pages(MAX_PENDING_REQS, queue->mmap_pages);
|
|
|
|
}
|
|
|
|
|
2013-09-18 00:46:08 +08:00
|
|
|
void xenvif_free(struct xenvif *vif)
|
|
|
|
{
|
2014-06-04 17:30:42 +08:00
|
|
|
struct xenvif_queue *queue = NULL;
|
2014-06-23 17:50:17 +08:00
|
|
|
unsigned int num_queues = vif->num_queues;
|
2014-06-04 17:30:42 +08:00
|
|
|
unsigned int queue_index;
|
2014-03-07 05:48:26 +08:00
|
|
|
|
2014-06-04 17:30:42 +08:00
|
|
|
unregister_netdev(vif->dev);
|
2014-03-07 05:48:26 +08:00
|
|
|
|
2014-06-04 17:30:42 +08:00
|
|
|
for (queue_index = 0; queue_index < num_queues; ++queue_index) {
|
|
|
|
queue = &vif->queues[queue_index];
|
2014-06-04 17:30:43 +08:00
|
|
|
xenvif_deinit_queue(queue);
|
2014-06-04 17:30:42 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
vfree(vif->queues);
|
|
|
|
vif->queues = NULL;
|
2014-06-23 17:50:17 +08:00
|
|
|
vif->num_queues = 0;
|
2011-03-15 08:06:18 +08:00
|
|
|
|
|
|
|
free_netdev(vif->dev);
|
2013-05-17 07:26:11 +08:00
|
|
|
|
2013-09-18 00:46:08 +08:00
|
|
|
module_put(THIS_MODULE);
|
2011-03-15 08:06:18 +08:00
|
|
|
}
|