kernel-fxtec-pro1x/arch/um/drivers/net_kern.c
Anton Ivanov 49da7e64f3 High Performance UML Vector Network Driver
1. Provides infrastructure for vector IO using recvmmsg/sendmmsg.
    1.1. Multi-message read.
    1.2. Multi-message write.
    1.3. Optimized queue support for multi-packet enqueue/dequeue.
    1.4. BQL/DQL support.
2. Implements transports for several transports as well support
for direct wiring of PWEs to NIC. Allows direct connection of VMs
to host, other VMs and network devices with no switch in use.
    2.1. Raw socket >4 times higher PPS and 10 times higher tcp RX
    than existing pcap based transport (> 4Gbit)
    2.2. New tap transport using socket RX and tap xmit. Similar
    performance improvements (>4Gbit)
    2.3. GRE transport - direct wiring to GRE PWE
    2.4. L2TPv3 transport - direct wiring to L2TPv3 PWE
3. Tuning, performance and offload related setting support via ethtool.
4. Initial BPF support - used in tap/raw to avoid software looping
5. Scatter Gather support.
6. VNET and checksum offload support for raw socket transport.
7. TSO/GSO support where applicable or available
8. Migrates all error messages to netdevice_*() and rate limits
them where needed.

Signed-off-by: Anton Ivanov <anton.ivanov@cambridgegreys.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
2018-02-19 19:38:51 +01:00

903 lines
20 KiB
C

/*
* Copyright (C) 2001 - 2007 Jeff Dike (jdike@{addtoit,linux.intel}.com)
* Copyright (C) 2001 Lennert Buytenhek (buytenh@gnu.org) and
* James Leu (jleu@mindspring.net).
* Copyright (C) 2001 by various other people who didn't put their name here.
* Licensed under the GPL.
*/
#include <linux/bootmem.h>
#include <linux/etherdevice.h>
#include <linux/ethtool.h>
#include <linux/inetdevice.h>
#include <linux/init.h>
#include <linux/list.h>
#include <linux/netdevice.h>
#include <linux/platform_device.h>
#include <linux/rtnetlink.h>
#include <linux/skbuff.h>
#include <linux/slab.h>
#include <linux/spinlock.h>
#include <init.h>
#include <irq_kern.h>
#include <irq_user.h>
#include "mconsole_kern.h"
#include <net_kern.h>
#include <net_user.h>
#define DRIVER_NAME "uml-netdev"
static DEFINE_SPINLOCK(opened_lock);
static LIST_HEAD(opened);
/*
* The drop_skb is used when we can't allocate an skb. The
* packet is read into drop_skb in order to get the data off the
* connection to the host.
* It is reallocated whenever a maximum packet size is seen which is
* larger than any seen before. update_drop_skb is called from
* eth_configure when a new interface is added.
*/
static DEFINE_SPINLOCK(drop_lock);
static struct sk_buff *drop_skb;
static int drop_max;
static int update_drop_skb(int max)
{
struct sk_buff *new;
unsigned long flags;
int err = 0;
spin_lock_irqsave(&drop_lock, flags);
if (max <= drop_max)
goto out;
err = -ENOMEM;
new = dev_alloc_skb(max);
if (new == NULL)
goto out;
skb_put(new, max);
kfree_skb(drop_skb);
drop_skb = new;
drop_max = max;
err = 0;
out:
spin_unlock_irqrestore(&drop_lock, flags);
return err;
}
static int uml_net_rx(struct net_device *dev)
{
struct uml_net_private *lp = netdev_priv(dev);
int pkt_len;
struct sk_buff *skb;
/* If we can't allocate memory, try again next round. */
skb = dev_alloc_skb(lp->max_packet);
if (skb == NULL) {
drop_skb->dev = dev;
/* Read a packet into drop_skb and don't do anything with it. */
(*lp->read)(lp->fd, drop_skb, lp);
dev->stats.rx_dropped++;
return 0;
}
skb->dev = dev;
skb_put(skb, lp->max_packet);
skb_reset_mac_header(skb);
pkt_len = (*lp->read)(lp->fd, skb, lp);
if (pkt_len > 0) {
skb_trim(skb, pkt_len);
skb->protocol = (*lp->protocol)(skb);
dev->stats.rx_bytes += skb->len;
dev->stats.rx_packets++;
netif_rx(skb);
return pkt_len;
}
kfree_skb(skb);
return pkt_len;
}
static void uml_dev_close(struct work_struct *work)
{
struct uml_net_private *lp =
container_of(work, struct uml_net_private, work);
dev_close(lp->dev);
}
static irqreturn_t uml_net_interrupt(int irq, void *dev_id)
{
struct net_device *dev = dev_id;
struct uml_net_private *lp = netdev_priv(dev);
int err;
if (!netif_running(dev))
return IRQ_NONE;
spin_lock(&lp->lock);
while ((err = uml_net_rx(dev)) > 0) ;
if (err < 0) {
printk(KERN_ERR
"Device '%s' read returned %d, shutting it down\n",
dev->name, err);
/* dev_close can't be called in interrupt context, and takes
* again lp->lock.
* And dev_close() can be safely called multiple times on the
* same device, since it tests for (dev->flags & IFF_UP). So
* there's no harm in delaying the device shutdown.
* Furthermore, the workqueue will not re-enqueue an already
* enqueued work item. */
schedule_work(&lp->work);
goto out;
}
reactivate_fd(lp->fd, UM_ETH_IRQ);
out:
spin_unlock(&lp->lock);
return IRQ_HANDLED;
}
static int uml_net_open(struct net_device *dev)
{
struct uml_net_private *lp = netdev_priv(dev);
int err;
if (lp->fd >= 0) {
err = -ENXIO;
goto out;
}
lp->fd = (*lp->open)(&lp->user);
if (lp->fd < 0) {
err = lp->fd;
goto out;
}
err = um_request_irq(dev->irq, lp->fd, IRQ_READ, uml_net_interrupt,
IRQF_SHARED, dev->name, dev);
if (err != 0) {
printk(KERN_ERR "uml_net_open: failed to get irq(%d)\n", err);
err = -ENETUNREACH;
goto out_close;
}
netif_start_queue(dev);
/* clear buffer - it can happen that the host side of the interface
* is full when we get here. In this case, new data is never queued,
* SIGIOs never arrive, and the net never works.
*/
while ((err = uml_net_rx(dev)) > 0) ;
spin_lock(&opened_lock);
list_add(&lp->list, &opened);
spin_unlock(&opened_lock);
return 0;
out_close:
if (lp->close != NULL) (*lp->close)(lp->fd, &lp->user);
lp->fd = -1;
out:
return err;
}
static int uml_net_close(struct net_device *dev)
{
struct uml_net_private *lp = netdev_priv(dev);
netif_stop_queue(dev);
um_free_irq(dev->irq, dev);
if (lp->close != NULL)
(*lp->close)(lp->fd, &lp->user);
lp->fd = -1;
spin_lock(&opened_lock);
list_del(&lp->list);
spin_unlock(&opened_lock);
return 0;
}
static int uml_net_start_xmit(struct sk_buff *skb, struct net_device *dev)
{
struct uml_net_private *lp = netdev_priv(dev);
unsigned long flags;
int len;
netif_stop_queue(dev);
spin_lock_irqsave(&lp->lock, flags);
len = (*lp->write)(lp->fd, skb, lp);
skb_tx_timestamp(skb);
if (len == skb->len) {
dev->stats.tx_packets++;
dev->stats.tx_bytes += skb->len;
netif_trans_update(dev);
netif_start_queue(dev);
/* this is normally done in the interrupt when tx finishes */
netif_wake_queue(dev);
}
else if (len == 0) {
netif_start_queue(dev);
dev->stats.tx_dropped++;
}
else {
netif_start_queue(dev);
printk(KERN_ERR "uml_net_start_xmit: failed(%d)\n", len);
}
spin_unlock_irqrestore(&lp->lock, flags);
dev_consume_skb_any(skb);
return NETDEV_TX_OK;
}
static void uml_net_set_multicast_list(struct net_device *dev)
{
return;
}
static void uml_net_tx_timeout(struct net_device *dev)
{
netif_trans_update(dev);
netif_wake_queue(dev);
}
#ifdef CONFIG_NET_POLL_CONTROLLER
static void uml_net_poll_controller(struct net_device *dev)
{
disable_irq(dev->irq);
uml_net_interrupt(dev->irq, dev);
enable_irq(dev->irq);
}
#endif
static void uml_net_get_drvinfo(struct net_device *dev,
struct ethtool_drvinfo *info)
{
strlcpy(info->driver, DRIVER_NAME, sizeof(info->driver));
strlcpy(info->version, "42", sizeof(info->version));
}
static const struct ethtool_ops uml_net_ethtool_ops = {
.get_drvinfo = uml_net_get_drvinfo,
.get_link = ethtool_op_get_link,
.get_ts_info = ethtool_op_get_ts_info,
};
static void uml_net_user_timer_expire(struct timer_list *t)
{
#ifdef undef
struct uml_net_private *lp = from_timer(lp, t, tl);
struct connection *conn = &lp->user;
dprintk(KERN_INFO "uml_net_user_timer_expire [%p]\n", conn);
do_connect(conn);
#endif
}
void uml_net_setup_etheraddr(struct net_device *dev, char *str)
{
unsigned char *addr = dev->dev_addr;
char *end;
int i;
if (str == NULL)
goto random;
for (i = 0; i < 6; i++) {
addr[i] = simple_strtoul(str, &end, 16);
if ((end == str) ||
((*end != ':') && (*end != ',') && (*end != '\0'))) {
printk(KERN_ERR
"setup_etheraddr: failed to parse '%s' "
"as an ethernet address\n", str);
goto random;
}
str = end + 1;
}
if (is_multicast_ether_addr(addr)) {
printk(KERN_ERR
"Attempt to assign a multicast ethernet address to a "
"device disallowed\n");
goto random;
}
if (!is_valid_ether_addr(addr)) {
printk(KERN_ERR
"Attempt to assign an invalid ethernet address to a "
"device disallowed\n");
goto random;
}
if (!is_local_ether_addr(addr)) {
printk(KERN_WARNING
"Warning: Assigning a globally valid ethernet "
"address to a device\n");
printk(KERN_WARNING "You should set the 2nd rightmost bit in "
"the first byte of the MAC,\n");
printk(KERN_WARNING "i.e. %02x:%02x:%02x:%02x:%02x:%02x\n",
addr[0] | 0x02, addr[1], addr[2], addr[3], addr[4],
addr[5]);
}
return;
random:
printk(KERN_INFO
"Choosing a random ethernet address for device %s\n", dev->name);
eth_hw_addr_random(dev);
}
static DEFINE_SPINLOCK(devices_lock);
static LIST_HEAD(devices);
static struct platform_driver uml_net_driver = {
.driver = {
.name = DRIVER_NAME,
},
};
static void net_device_release(struct device *dev)
{
struct uml_net *device = dev_get_drvdata(dev);
struct net_device *netdev = device->dev;
struct uml_net_private *lp = netdev_priv(netdev);
if (lp->remove != NULL)
(*lp->remove)(&lp->user);
list_del(&device->list);
kfree(device);
free_netdev(netdev);
}
static const struct net_device_ops uml_netdev_ops = {
.ndo_open = uml_net_open,
.ndo_stop = uml_net_close,
.ndo_start_xmit = uml_net_start_xmit,
.ndo_set_rx_mode = uml_net_set_multicast_list,
.ndo_tx_timeout = uml_net_tx_timeout,
.ndo_set_mac_address = eth_mac_addr,
.ndo_validate_addr = eth_validate_addr,
#ifdef CONFIG_NET_POLL_CONTROLLER
.ndo_poll_controller = uml_net_poll_controller,
#endif
};
/*
* Ensures that platform_driver_register is called only once by
* eth_configure. Will be set in an initcall.
*/
static int driver_registered;
static void eth_configure(int n, void *init, char *mac,
struct transport *transport, gfp_t gfp_mask)
{
struct uml_net *device;
struct net_device *dev;
struct uml_net_private *lp;
int err, size;
size = transport->private_size + sizeof(struct uml_net_private);
device = kzalloc(sizeof(*device), gfp_mask);
if (device == NULL) {
printk(KERN_ERR "eth_configure failed to allocate struct "
"uml_net\n");
return;
}
dev = alloc_etherdev(size);
if (dev == NULL) {
printk(KERN_ERR "eth_configure: failed to allocate struct "
"net_device for eth%d\n", n);
goto out_free_device;
}
INIT_LIST_HEAD(&device->list);
device->index = n;
/* If this name ends up conflicting with an existing registered
* netdevice, that is OK, register_netdev{,ice}() will notice this
* and fail.
*/
snprintf(dev->name, sizeof(dev->name), "eth%d", n);
uml_net_setup_etheraddr(dev, mac);
printk(KERN_INFO "Netdevice %d (%pM) : ", n, dev->dev_addr);
lp = netdev_priv(dev);
/* This points to the transport private data. It's still clear, but we
* must memset it to 0 *now*. Let's help the drivers. */
memset(lp, 0, size);
INIT_WORK(&lp->work, uml_dev_close);
/* sysfs register */
if (!driver_registered) {
platform_driver_register(&uml_net_driver);
driver_registered = 1;
}
device->pdev.id = n;
device->pdev.name = DRIVER_NAME;
device->pdev.dev.release = net_device_release;
dev_set_drvdata(&device->pdev.dev, device);
if (platform_device_register(&device->pdev))
goto out_free_netdev;
SET_NETDEV_DEV(dev,&device->pdev.dev);
device->dev = dev;
/*
* These just fill in a data structure, so there's no failure
* to be worried about.
*/
(*transport->kern->init)(dev, init);
*lp = ((struct uml_net_private)
{ .list = LIST_HEAD_INIT(lp->list),
.dev = dev,
.fd = -1,
.mac = { 0xfe, 0xfd, 0x0, 0x0, 0x0, 0x0},
.max_packet = transport->user->max_packet,
.protocol = transport->kern->protocol,
.open = transport->user->open,
.close = transport->user->close,
.remove = transport->user->remove,
.read = transport->kern->read,
.write = transport->kern->write,
.add_address = transport->user->add_address,
.delete_address = transport->user->delete_address });
timer_setup(&lp->tl, uml_net_user_timer_expire, 0);
spin_lock_init(&lp->lock);
memcpy(lp->mac, dev->dev_addr, sizeof(lp->mac));
if ((transport->user->init != NULL) &&
((*transport->user->init)(&lp->user, dev) != 0))
goto out_unregister;
dev->mtu = transport->user->mtu;
dev->netdev_ops = &uml_netdev_ops;
dev->ethtool_ops = &uml_net_ethtool_ops;
dev->watchdog_timeo = (HZ >> 1);
dev->irq = UM_ETH_IRQ;
err = update_drop_skb(lp->max_packet);
if (err)
goto out_undo_user_init;
rtnl_lock();
err = register_netdevice(dev);
rtnl_unlock();
if (err)
goto out_undo_user_init;
spin_lock(&devices_lock);
list_add(&device->list, &devices);
spin_unlock(&devices_lock);
return;
out_undo_user_init:
if (transport->user->remove != NULL)
(*transport->user->remove)(&lp->user);
out_unregister:
platform_device_unregister(&device->pdev);
return; /* platform_device_unregister frees dev and device */
out_free_netdev:
free_netdev(dev);
out_free_device:
kfree(device);
}
static struct uml_net *find_device(int n)
{
struct uml_net *device;
struct list_head *ele;
spin_lock(&devices_lock);
list_for_each(ele, &devices) {
device = list_entry(ele, struct uml_net, list);
if (device->index == n)
goto out;
}
device = NULL;
out:
spin_unlock(&devices_lock);
return device;
}
static int eth_parse(char *str, int *index_out, char **str_out,
char **error_out)
{
char *end;
int n, err = -EINVAL;
n = simple_strtoul(str, &end, 0);
if (end == str) {
*error_out = "Bad device number";
return err;
}
str = end;
if (*str != '=') {
*error_out = "Expected '=' after device number";
return err;
}
str++;
if (find_device(n)) {
*error_out = "Device already configured";
return err;
}
*index_out = n;
*str_out = str;
return 0;
}
struct eth_init {
struct list_head list;
char *init;
int index;
};
static DEFINE_SPINLOCK(transports_lock);
static LIST_HEAD(transports);
/* Filled in during early boot */
static LIST_HEAD(eth_cmd_line);
static int check_transport(struct transport *transport, char *eth, int n,
void **init_out, char **mac_out, gfp_t gfp_mask)
{
int len;
len = strlen(transport->name);
if (strncmp(eth, transport->name, len))
return 0;
eth += len;
if (*eth == ',')
eth++;
else if (*eth != '\0')
return 0;
*init_out = kmalloc(transport->setup_size, gfp_mask);
if (*init_out == NULL)
return 1;
if (!transport->setup(eth, mac_out, *init_out)) {
kfree(*init_out);
*init_out = NULL;
}
return 1;
}
void register_transport(struct transport *new)
{
struct list_head *ele, *next;
struct eth_init *eth;
void *init;
char *mac = NULL;
int match;
spin_lock(&transports_lock);
BUG_ON(!list_empty(&new->list));
list_add(&new->list, &transports);
spin_unlock(&transports_lock);
list_for_each_safe(ele, next, &eth_cmd_line) {
eth = list_entry(ele, struct eth_init, list);
match = check_transport(new, eth->init, eth->index, &init,
&mac, GFP_KERNEL);
if (!match)
continue;
else if (init != NULL) {
eth_configure(eth->index, init, mac, new, GFP_KERNEL);
kfree(init);
}
list_del(&eth->list);
}
}
static int eth_setup_common(char *str, int index)
{
struct list_head *ele;
struct transport *transport;
void *init;
char *mac = NULL;
int found = 0;
spin_lock(&transports_lock);
list_for_each(ele, &transports) {
transport = list_entry(ele, struct transport, list);
if (!check_transport(transport, str, index, &init,
&mac, GFP_ATOMIC))
continue;
if (init != NULL) {
eth_configure(index, init, mac, transport, GFP_ATOMIC);
kfree(init);
}
found = 1;
break;
}
spin_unlock(&transports_lock);
return found;
}
static int __init eth_setup(char *str)
{
struct eth_init *new;
char *error;
int n, err;
err = eth_parse(str, &n, &str, &error);
if (err) {
printk(KERN_ERR "eth_setup - Couldn't parse '%s' : %s\n",
str, error);
return 1;
}
new = alloc_bootmem(sizeof(*new));
INIT_LIST_HEAD(&new->list);
new->index = n;
new->init = str;
list_add_tail(&new->list, &eth_cmd_line);
return 1;
}
__setup("eth", eth_setup);
__uml_help(eth_setup,
"eth[0-9]+=<transport>,<options>\n"
" Configure a network device.\n\n"
);
static int net_config(char *str, char **error_out)
{
int n, err;
err = eth_parse(str, &n, &str, error_out);
if (err)
return err;
/* This string is broken up and the pieces used by the underlying
* driver. So, it is freed only if eth_setup_common fails.
*/
str = kstrdup(str, GFP_KERNEL);
if (str == NULL) {
*error_out = "net_config failed to strdup string";
return -ENOMEM;
}
err = !eth_setup_common(str, n);
if (err)
kfree(str);
return err;
}
static int net_id(char **str, int *start_out, int *end_out)
{
char *end;
int n;
n = simple_strtoul(*str, &end, 0);
if ((*end != '\0') || (end == *str))
return -1;
*start_out = n;
*end_out = n;
*str = end;
return n;
}
static int net_remove(int n, char **error_out)
{
struct uml_net *device;
struct net_device *dev;
struct uml_net_private *lp;
device = find_device(n);
if (device == NULL)
return -ENODEV;
dev = device->dev;
lp = netdev_priv(dev);
if (lp->fd > 0)
return -EBUSY;
unregister_netdev(dev);
platform_device_unregister(&device->pdev);
return 0;
}
static struct mc_device net_mc = {
.list = LIST_HEAD_INIT(net_mc.list),
.name = "eth",
.config = net_config,
.get_config = NULL,
.id = net_id,
.remove = net_remove,
};
#ifdef CONFIG_INET
static int uml_inetaddr_event(struct notifier_block *this, unsigned long event,
void *ptr)
{
struct in_ifaddr *ifa = ptr;
struct net_device *dev = ifa->ifa_dev->dev;
struct uml_net_private *lp;
void (*proc)(unsigned char *, unsigned char *, void *);
unsigned char addr_buf[4], netmask_buf[4];
if (dev->netdev_ops->ndo_open != uml_net_open)
return NOTIFY_DONE;
lp = netdev_priv(dev);
proc = NULL;
switch (event) {
case NETDEV_UP:
proc = lp->add_address;
break;
case NETDEV_DOWN:
proc = lp->delete_address;
break;
}
if (proc != NULL) {
memcpy(addr_buf, &ifa->ifa_address, sizeof(addr_buf));
memcpy(netmask_buf, &ifa->ifa_mask, sizeof(netmask_buf));
(*proc)(addr_buf, netmask_buf, &lp->user);
}
return NOTIFY_DONE;
}
/* uml_net_init shouldn't be called twice on two CPUs at the same time */
static struct notifier_block uml_inetaddr_notifier = {
.notifier_call = uml_inetaddr_event,
};
static void inet_register(void)
{
struct list_head *ele;
struct uml_net_private *lp;
struct in_device *ip;
struct in_ifaddr *in;
register_inetaddr_notifier(&uml_inetaddr_notifier);
/* Devices may have been opened already, so the uml_inetaddr_notifier
* didn't get a chance to run for them. This fakes it so that
* addresses which have already been set up get handled properly.
*/
spin_lock(&opened_lock);
list_for_each(ele, &opened) {
lp = list_entry(ele, struct uml_net_private, list);
ip = lp->dev->ip_ptr;
if (ip == NULL)
continue;
in = ip->ifa_list;
while (in != NULL) {
uml_inetaddr_event(NULL, NETDEV_UP, in);
in = in->ifa_next;
}
}
spin_unlock(&opened_lock);
}
#else
static inline void inet_register(void)
{
}
#endif
static int uml_net_init(void)
{
mconsole_register_dev(&net_mc);
inet_register();
return 0;
}
__initcall(uml_net_init);
static void close_devices(void)
{
struct list_head *ele;
struct uml_net_private *lp;
spin_lock(&opened_lock);
list_for_each(ele, &opened) {
lp = list_entry(ele, struct uml_net_private, list);
um_free_irq(lp->dev->irq, lp->dev);
if ((lp->close != NULL) && (lp->fd >= 0))
(*lp->close)(lp->fd, &lp->user);
if (lp->remove != NULL)
(*lp->remove)(&lp->user);
}
spin_unlock(&opened_lock);
}
__uml_exitcall(close_devices);
void iter_addresses(void *d, void (*cb)(unsigned char *, unsigned char *,
void *),
void *arg)
{
struct net_device *dev = d;
struct in_device *ip = dev->ip_ptr;
struct in_ifaddr *in;
unsigned char address[4], netmask[4];
if (ip == NULL) return;
in = ip->ifa_list;
while (in != NULL) {
memcpy(address, &in->ifa_address, sizeof(address));
memcpy(netmask, &in->ifa_mask, sizeof(netmask));
(*cb)(address, netmask, arg);
in = in->ifa_next;
}
}
int dev_netmask(void *d, void *m)
{
struct net_device *dev = d;
struct in_device *ip = dev->ip_ptr;
struct in_ifaddr *in;
__be32 *mask_out = m;
if (ip == NULL)
return 1;
in = ip->ifa_list;
if (in == NULL)
return 1;
*mask_out = in->ifa_mask;
return 0;
}
void *get_output_buffer(int *len_out)
{
void *ret;
ret = (void *) __get_free_pages(GFP_KERNEL, 0);
if (ret) *len_out = PAGE_SIZE;
else *len_out = 0;
return ret;
}
void free_output_buffer(void *buffer)
{
free_pages((unsigned long) buffer, 0);
}
int tap_setup_common(char *str, char *type, char **dev_name, char **mac_out,
char **gate_addr)
{
char *remain;
remain = split_if_spec(str, dev_name, mac_out, gate_addr, NULL);
if (remain != NULL) {
printk(KERN_ERR "tap_setup_common - Extra garbage on "
"specification : '%s'\n", remain);
return 1;
}
return 0;
}
unsigned short eth_protocol(struct sk_buff *skb)
{
return eth_type_trans(skb, skb->dev);
}