leandrof/linux
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
28127bdd2f
linux/drivers/net/appletalk/ipddp.c
Eric Dumazet 0287587884 net: better IFF_XMIT_DST_RELEASE support 
Testing xmit_more support with netperf and connected UDP sockets,
I found strange dst refcount false sharing.

Current handling of IFF_XMIT_DST_RELEASE is not optimal.

Dropping dst in validate_xmit_skb() is certainly too late in case
packet was queued by cpu X but dequeued by cpu Y

The logical point to take care of drop/force is in __dev_queue_xmit()
before even taking qdisc lock.

As Julian Anastasov pointed out, need for skb_dst() might come from some
packet schedulers or classifiers.

This patch adds new helper to cleanly express needs of various drivers
or qdiscs/classifiers.

Drivers that need skb_dst() in their ndo_start_xmit() should call
following helper in their setup instead of the prior :

	dev->priv_flags &= ~IFF_XMIT_DST_RELEASE;
->
	netif_keep_dst(dev);

Instead of using a single bit, we use two bits, one being
eventually rebuilt in bonding/team drivers.

The other one, is permanent and blocks IFF_XMIT_DST_RELEASE being
rebuilt in bonding/team. Eventually, we could add something
smarter later.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Julian Anastasov <ja@ssi.bg>
Signed-off-by: David S. Miller <davem@davemloft.net>
2014-10-07 13:22:11 -04:00

336 lines
8.8 KiB
C
Raw Blame History

/*
* ipddp.c: IP to Appletalk-IP Encapsulation driver for Linux
* Appletalk-IP to IP Decapsulation driver for Linux
*
* Authors:
* - DDP-IP Encap by: Bradford W. Johnson <johns393@maroon.tc.umn.edu>
* - DDP-IP Decap by: Jay Schulist <jschlst@samba.org>
*
* Derived from:
* - Almost all code already existed in net/appletalk/ddp.c I just
* moved/reorginized it into a driver file. Original IP-over-DDP code
* was done by Bradford W. Johnson <johns393@maroon.tc.umn.edu>
* - skeleton.c: A network driver outline for linux.
* Written 1993-94 by Donald Becker.
* - dummy.c: A dummy net driver. By Nick Holloway.
* - MacGate: A user space Daemon for Appletalk-IP Decap for
* Linux by Jay Schulist <jschlst@samba.org>
*
* Copyright 1993 United States Government as represented by the
* Director, National Security Agency.
*
* This software may be used and distributed according to the terms
* of the GNU General Public License, incorporated herein by reference.
*/
#include <linux/module.h>
#include <linux/kernel.h>
#include <linux/init.h>
#include <linux/netdevice.h>
#include <linux/etherdevice.h>
#include <linux/ip.h>
#include <linux/atalk.h>
#include <linux/if_arp.h>
#include <linux/slab.h>
#include <net/route.h>
#include <asm/uaccess.h>
#include "ipddp.h" /* Our stuff */
static const char version[] = KERN_INFO "ipddp.c:v0.01 8/28/97 Bradford W. Johnson <johns393@maroon.tc.umn.edu>\n";
static struct ipddp_route *ipddp_route_list;
static DEFINE_SPINLOCK(ipddp_route_lock);
#ifdef CONFIG_IPDDP_ENCAP
static int ipddp_mode = IPDDP_ENCAP;
#else
static int ipddp_mode = IPDDP_DECAP;
#endif
/* Index to functions, as function prototypes. */
static netdev_tx_t ipddp_xmit(struct sk_buff *skb,
struct net_device *dev);
static int ipddp_create(struct ipddp_route *new_rt);
static int ipddp_delete(struct ipddp_route *rt);
static struct ipddp_route* __ipddp_find_route(struct ipddp_route *rt);
static int ipddp_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd);
static const struct net_device_ops ipddp_netdev_ops = {
.ndo_start_xmit = ipddp_xmit,
.ndo_do_ioctl = ipddp_ioctl,
.ndo_change_mtu = eth_change_mtu,
.ndo_set_mac_address = eth_mac_addr,
.ndo_validate_addr = eth_validate_addr,
};
static struct net_device * __init ipddp_init(void)
{
static unsigned version_printed;
struct net_device *dev;
int err;
dev = alloc_etherdev(0);
if (!dev)
return ERR_PTR(-ENOMEM);
netif_keep_dst(dev);
strcpy(dev->name, "ipddp%d");
if (version_printed++ == 0)
printk(version);
/* Initialize the device structure. */
dev->netdev_ops = &ipddp_netdev_ops;
dev->type = ARPHRD_IPDDP; /* IP over DDP tunnel */
dev->mtu = 585;
dev->flags |= IFF_NOARP;
/*
* The worst case header we will need is currently a
* ethernet header (14 bytes) and a ddp header (sizeof ddpehdr+1)
* We send over SNAP so that takes another 8 bytes.
*/
dev->hard_header_len = 14+8+sizeof(struct ddpehdr)+1;
err = register_netdev(dev);
if (err) {
free_netdev(dev);
return ERR_PTR(err);
}
/* Let the user now what mode we are in */
if(ipddp_mode == IPDDP_ENCAP)
printk("%s: Appletalk-IP Encap. mode by Bradford W. Johnson <johns393@maroon.tc.umn.edu>\n",
dev->name);
if(ipddp_mode == IPDDP_DECAP)
printk("%s: Appletalk-IP Decap. mode by Jay Schulist <jschlst@samba.org>\n",
dev->name);
return dev;
}
/*
* Transmit LLAP/ELAP frame using aarp_send_ddp.
*/
static netdev_tx_t ipddp_xmit(struct sk_buff *skb, struct net_device *dev)
{
__be32 paddr = skb_rtable(skb)->rt_gateway;
struct ddpehdr *ddp;
struct ipddp_route *rt;
struct atalk_addr *our_addr;
spin_lock(&ipddp_route_lock);
/*
* Find appropriate route to use, based only on IP number.
*/
for(rt = ipddp_route_list; rt != NULL; rt = rt->next)
{
if(rt->ip == paddr)
break;
}
if(rt == NULL) {
spin_unlock(&ipddp_route_lock);
return NETDEV_TX_OK;
}
our_addr = atalk_find_dev_addr(rt->dev);
if(ipddp_mode == IPDDP_DECAP)
/*
* Pull off the excess room that should not be there.
* This is due to a hard-header problem. This is the
* quick fix for now though, till it breaks.
*/
skb_pull(skb, 35-(sizeof(struct ddpehdr)+1));
/* Create the Extended DDP header */
ddp = (struct ddpehdr *)skb->data;
ddp->deh_len_hops = htons(skb->len + (1<<10));
ddp->deh_sum = 0;
/*
* For Localtalk we need aarp_send_ddp to strip the
* long DDP header and place a shot DDP header on it.
*/
if(rt->dev->type == ARPHRD_LOCALTLK)
{
ddp->deh_dnet = 0; /* FIXME more hops?? */
ddp->deh_snet = 0;
}
else
{
ddp->deh_dnet = rt->at.s_net; /* FIXME more hops?? */
ddp->deh_snet = our_addr->s_net;
}
ddp->deh_dnode = rt->at.s_node;
ddp->deh_snode = our_addr->s_node;
ddp->deh_dport = 72;
ddp->deh_sport = 72;
*((__u8 *)(ddp+1)) = 22; /* ddp type = IP */
skb->protocol = htons(ETH_P_ATALK); /* Protocol has changed */
dev->stats.tx_packets++;
dev->stats.tx_bytes += skb->len;
aarp_send_ddp(rt->dev, skb, &rt->at, NULL);
spin_unlock(&ipddp_route_lock);
return NETDEV_TX_OK;
}
/*
* Create a routing entry. We first verify that the
* record does not already exist. If it does we return -EEXIST
*/
static int ipddp_create(struct ipddp_route *new_rt)
{
struct ipddp_route *rt = kmalloc(sizeof(*rt), GFP_KERNEL);
if (rt == NULL)
return -ENOMEM;
rt->ip = new_rt->ip;
rt->at = new_rt->at;
rt->next = NULL;
if ((rt->dev = atrtr_get_dev(&rt->at)) == NULL) {
kfree(rt);
return -ENETUNREACH;
}
spin_lock_bh(&ipddp_route_lock);
if (__ipddp_find_route(rt)) {
spin_unlock_bh(&ipddp_route_lock);
kfree(rt);
return -EEXIST;
}
rt->next = ipddp_route_list;
ipddp_route_list = rt;
spin_unlock_bh(&ipddp_route_lock);
return 0;
}
/*
* Delete a route, we only delete a FULL match.
* If route does not exist we return -ENOENT.
*/
static int ipddp_delete(struct ipddp_route *rt)
{
struct ipddp_route **r = &ipddp_route_list;
struct ipddp_route *tmp;
spin_lock_bh(&ipddp_route_lock);
while((tmp = *r) != NULL)
{
if(tmp->ip == rt->ip &&
tmp->at.s_net == rt->at.s_net &&
tmp->at.s_node == rt->at.s_node)
{
*r = tmp->next;
spin_unlock_bh(&ipddp_route_lock);
kfree(tmp);
return 0;
}
r = &tmp->next;
}
spin_unlock_bh(&ipddp_route_lock);
return -ENOENT;
}
/*
* Find a routing entry, we only return a FULL match
*/
static struct ipddp_route* __ipddp_find_route(struct ipddp_route *rt)
{
struct ipddp_route *f;
for(f = ipddp_route_list; f != NULL; f = f->next)
{
if(f->ip == rt->ip &&
f->at.s_net == rt->at.s_net &&
f->at.s_node == rt->at.s_node)
return f;
}
return NULL;
}
static int ipddp_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd)
{
struct ipddp_route __user *rt = ifr->ifr_data;
struct ipddp_route rcp, rcp2, *rp;
if(!capable(CAP_NET_ADMIN))
return -EPERM;
if(copy_from_user(&rcp, rt, sizeof(rcp)))
return -EFAULT;
switch(cmd)
{
case SIOCADDIPDDPRT:
return ipddp_create(&rcp);
case SIOCFINDIPDDPRT:
spin_lock_bh(&ipddp_route_lock);
rp = __ipddp_find_route(&rcp);
if (rp)
memcpy(&rcp2, rp, sizeof(rcp2));
spin_unlock_bh(&ipddp_route_lock);
if (rp) {
if (copy_to_user(rt, &rcp2,
sizeof(struct ipddp_route)))
return -EFAULT;
return 0;
} else
return -ENOENT;
case SIOCDELIPDDPRT:
return ipddp_delete(&rcp);
default:
return -EINVAL;
}
}
static struct net_device *dev_ipddp;
MODULE_LICENSE("GPL");
module_param(ipddp_mode, int, 0);
static int __init ipddp_init_module(void)
{
dev_ipddp = ipddp_init();
if (IS_ERR(dev_ipddp))
return PTR_ERR(dev_ipddp);
return 0;
}
static void __exit ipddp_cleanup_module(void)
{
struct ipddp_route *p;
unregister_netdev(dev_ipddp);
free_netdev(dev_ipddp);
while (ipddp_route_list) {
p = ipddp_route_list->next;
kfree(ipddp_route_list);
ipddp_route_list = p;
}
}
module_init(ipddp_init_module);
module_exit(ipddp_cleanup_module);
Reference in New Issue View Git Blame Copy Permalink