This commit was manufactured by cvs2svn to create branch 'vserver'.
[linux-2.6.git] / net / netfilter / nf_queue.c
diff --git a/net/netfilter/nf_queue.c b/net/netfilter/nf_queue.c
new file mode 100644 (file)
index 0000000..ee8f708
--- /dev/null
@@ -0,0 +1,318 @@
+#include <linux/config.h>
+#include <linux/kernel.h>
+#include <linux/init.h>
+#include <linux/module.h>
+#include <linux/proc_fs.h>
+#include <linux/skbuff.h>
+#include <linux/netfilter.h>
+#include <linux/seq_file.h>
+#include <linux/rcupdate.h>
+#include <net/protocol.h>
+
+#include "nf_internals.h"
+
+/* 
+ * A queue handler may be registered for each protocol.  Each is protected by
+ * long term mutex.  The handler must provide an an outfn() to accept packets
+ * for queueing and must reinject all packets it receives, no matter what.
+ */
+static struct nf_queue_handler *queue_handler[NPROTO];
+
+static DEFINE_RWLOCK(queue_handler_lock);
+
+/* return EBUSY when somebody else is registered, return EEXIST if the
+ * same handler is registered, return 0 in case of success. */
+int nf_register_queue_handler(int pf, struct nf_queue_handler *qh)
+{      
+       int ret;
+
+       if (pf >= NPROTO)
+               return -EINVAL;
+
+       write_lock_bh(&queue_handler_lock);
+       if (queue_handler[pf] == qh)
+               ret = -EEXIST;
+       else if (queue_handler[pf])
+               ret = -EBUSY;
+       else {
+               queue_handler[pf] = qh;
+               ret = 0;
+       }
+       write_unlock_bh(&queue_handler_lock);
+
+       return ret;
+}
+EXPORT_SYMBOL(nf_register_queue_handler);
+
+/* The caller must flush their queue before this */
+int nf_unregister_queue_handler(int pf)
+{
+       if (pf >= NPROTO)
+               return -EINVAL;
+
+       write_lock_bh(&queue_handler_lock);
+       queue_handler[pf] = NULL;
+       write_unlock_bh(&queue_handler_lock);
+       
+       return 0;
+}
+EXPORT_SYMBOL(nf_unregister_queue_handler);
+
+void nf_unregister_queue_handlers(struct nf_queue_handler *qh)
+{
+       int pf;
+
+       write_lock_bh(&queue_handler_lock);
+       for (pf = 0; pf < NPROTO; pf++)  {
+               if (queue_handler[pf] == qh)
+                       queue_handler[pf] = NULL;
+       }
+       write_unlock_bh(&queue_handler_lock);
+}
+EXPORT_SYMBOL_GPL(nf_unregister_queue_handlers);
+
+/* 
+ * Any packet that leaves via this function must come back 
+ * through nf_reinject().
+ */
+int nf_queue(struct sk_buff **skb, 
+            struct list_head *elem, 
+            int pf, unsigned int hook,
+            struct net_device *indev,
+            struct net_device *outdev,
+            int (*okfn)(struct sk_buff *),
+            unsigned int queuenum)
+{
+       int status;
+       struct nf_info *info;
+#ifdef CONFIG_BRIDGE_NETFILTER
+       struct net_device *physindev = NULL;
+       struct net_device *physoutdev = NULL;
+#endif
+       struct nf_afinfo *afinfo;
+
+       /* QUEUE == DROP if noone is waiting, to be safe. */
+       read_lock(&queue_handler_lock);
+       if (!queue_handler[pf]) {
+               read_unlock(&queue_handler_lock);
+               kfree_skb(*skb);
+               return 1;
+       }
+
+       afinfo = nf_get_afinfo(pf);
+       if (!afinfo) {
+               read_unlock(&queue_handler_lock);
+               kfree_skb(*skb);
+               return 1;
+       }
+
+       info = kmalloc(sizeof(*info) + afinfo->route_key_size, GFP_ATOMIC);
+       if (!info) {
+               if (net_ratelimit())
+                       printk(KERN_ERR "OOM queueing packet %p\n",
+                              *skb);
+               read_unlock(&queue_handler_lock);
+               kfree_skb(*skb);
+               return 1;
+       }
+
+       *info = (struct nf_info) { 
+               (struct nf_hook_ops *)elem, pf, hook, indev, outdev, okfn };
+
+       /* If it's going away, ignore hook. */
+       if (!try_module_get(info->elem->owner)) {
+               read_unlock(&queue_handler_lock);
+               kfree(info);
+               return 0;
+       }
+
+       /* Bump dev refs so they don't vanish while packet is out */
+       if (indev) dev_hold(indev);
+       if (outdev) dev_hold(outdev);
+
+#ifdef CONFIG_BRIDGE_NETFILTER
+       if ((*skb)->nf_bridge) {
+               physindev = (*skb)->nf_bridge->physindev;
+               if (physindev) dev_hold(physindev);
+               physoutdev = (*skb)->nf_bridge->physoutdev;
+               if (physoutdev) dev_hold(physoutdev);
+       }
+#endif
+       afinfo->saveroute(*skb, info);
+       status = queue_handler[pf]->outfn(*skb, info, queuenum,
+                                         queue_handler[pf]->data);
+
+       read_unlock(&queue_handler_lock);
+
+       if (status < 0) {
+               /* James M doesn't say fuck enough. */
+               if (indev) dev_put(indev);
+               if (outdev) dev_put(outdev);
+#ifdef CONFIG_BRIDGE_NETFILTER
+               if (physindev) dev_put(physindev);
+               if (physoutdev) dev_put(physoutdev);
+#endif
+               module_put(info->elem->owner);
+               kfree(info);
+               kfree_skb(*skb);
+
+               return 1;
+       }
+
+       return 1;
+}
+
+void nf_reinject(struct sk_buff *skb, struct nf_info *info,
+                unsigned int verdict)
+{
+       struct list_head *elem = &info->elem->list;
+       struct list_head *i;
+       struct nf_afinfo *afinfo;
+
+       rcu_read_lock();
+
+       /* Release those devices we held, or Alexey will kill me. */
+       if (info->indev) dev_put(info->indev);
+       if (info->outdev) dev_put(info->outdev);
+#ifdef CONFIG_BRIDGE_NETFILTER
+       if (skb->nf_bridge) {
+               if (skb->nf_bridge->physindev)
+                       dev_put(skb->nf_bridge->physindev);
+               if (skb->nf_bridge->physoutdev)
+                       dev_put(skb->nf_bridge->physoutdev);
+       }
+#endif
+
+       /* Drop reference to owner of hook which queued us. */
+       module_put(info->elem->owner);
+
+       list_for_each_rcu(i, &nf_hooks[info->pf][info->hook]) {
+               if (i == elem) 
+                       break;
+       }
+  
+       if (i == &nf_hooks[info->pf][info->hook]) {
+               /* The module which sent it to userspace is gone. */
+               NFDEBUG("%s: module disappeared, dropping packet.\n",
+                       __FUNCTION__);
+               verdict = NF_DROP;
+       }
+
+       /* Continue traversal iff userspace said ok... */
+       if (verdict == NF_REPEAT) {
+               elem = elem->prev;
+               verdict = NF_ACCEPT;
+       }
+
+       if (verdict == NF_ACCEPT) {
+               afinfo = nf_get_afinfo(info->pf);
+               if (!afinfo || afinfo->reroute(&skb, info) < 0)
+                       verdict = NF_DROP;
+       }
+
+       if (verdict == NF_ACCEPT) {
+       next_hook:
+               verdict = nf_iterate(&nf_hooks[info->pf][info->hook],
+                                    &skb, info->hook, 
+                                    info->indev, info->outdev, &elem,
+                                    info->okfn, INT_MIN);
+       }
+
+       switch (verdict & NF_VERDICT_MASK) {
+       case NF_ACCEPT:
+               info->okfn(skb);
+               break;
+
+       case NF_QUEUE:
+               if (!nf_queue(&skb, elem, info->pf, info->hook, 
+                             info->indev, info->outdev, info->okfn,
+                             verdict >> NF_VERDICT_BITS))
+                       goto next_hook;
+               break;
+       }
+       rcu_read_unlock();
+
+       if (verdict == NF_DROP)
+               kfree_skb(skb);
+
+       kfree(info);
+       return;
+}
+EXPORT_SYMBOL(nf_reinject);
+
+#ifdef CONFIG_PROC_FS
+static void *seq_start(struct seq_file *seq, loff_t *pos)
+{
+       if (*pos >= NPROTO)
+               return NULL;
+
+       return pos;
+}
+
+static void *seq_next(struct seq_file *s, void *v, loff_t *pos)
+{
+       (*pos)++;
+
+       if (*pos >= NPROTO)
+               return NULL;
+
+       return pos;
+}
+
+static void seq_stop(struct seq_file *s, void *v)
+{
+
+}
+
+static int seq_show(struct seq_file *s, void *v)
+{
+       int ret;
+       loff_t *pos = v;
+       struct nf_queue_handler *qh;
+
+       read_lock_bh(&queue_handler_lock);
+       qh = queue_handler[*pos];
+       if (!qh)
+               ret = seq_printf(s, "%2lld NONE\n", *pos);
+       else
+               ret = seq_printf(s, "%2lld %s\n", *pos, qh->name);
+       read_unlock_bh(&queue_handler_lock);
+
+       return ret;
+}
+
+static struct seq_operations nfqueue_seq_ops = {
+       .start  = seq_start,
+       .next   = seq_next,
+       .stop   = seq_stop,
+       .show   = seq_show,
+};
+
+static int nfqueue_open(struct inode *inode, struct file *file)
+{
+       return seq_open(file, &nfqueue_seq_ops);
+}
+
+static struct file_operations nfqueue_file_ops = {
+       .owner   = THIS_MODULE,
+       .open    = nfqueue_open,
+       .read    = seq_read,
+       .llseek  = seq_lseek,
+       .release = seq_release,
+};
+#endif /* PROC_FS */
+
+
+int __init netfilter_queue_init(void)
+{
+#ifdef CONFIG_PROC_FS
+       struct proc_dir_entry *pde;
+
+       pde = create_proc_entry("nf_queue", S_IRUGO, proc_net_netfilter);
+       if (!pde)
+               return -1;
+       pde->proc_fops = &nfqueue_file_ops;
+#endif
+       return 0;
+}
+