[NETNS][IPV6] ip6_fib - add net to gc timer parameter

The fib tables are now relative to the network namespace. When the
garbage collector timer expires, we must have a network namespace
parameter in order to retrieve the tables. For now this is the
init_net, but we should be able to have a timer per namespace and use
the timer callback parameter to pass the network namespace from the
expired timer.

The timer callback, fib6_run_gc, is actually used to be called
synchronously by some functions and asynchronously when the timer
expires.

When the timer expires, the delay specified for fib6_run_gc parameter
is always zero. So, I changed fib6_run_gc to not be a timer callback
but a function called by the timer callback and I added a timer
callback where its work is just to retrieve from the data arg of the
timer the network namespace and call fib6_run_gc with zero expiring
time and the network namespace parameters. That makes the code cleaner
for the fib6_run_gc callers.

Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com>
Signed-off-by: Benjamin Thery <benjamin.thery@bull.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

3 files changed, 24 insertions, 13 deletions

diff --git a/net/ipv6/ip6_fib.c b/net/ipv6/ip6_fib.c
index 0f9dc81f0e61..9b1c232a29ee 100644
--- a/net/ipv6/ip6_fib.c
+++ b/net/ipv6/ip6_fib.c
@@ -93,7 +93,10 @@ static int fib6_walk_continue(struct fib6_walker_t *w);
 
 static __u32 rt_sernum;
 
-static DEFINE_TIMER(ip6_fib_timer, fib6_run_gc, 0, 0);
+static void fib6_gc_timer_cb(unsigned long arg);
+
+static DEFINE_TIMER(ip6_fib_timer, fib6_gc_timer_cb, 0,
+                    (unsigned long)&init_net);
 
 static struct fib6_walker_t fib6_walker_list = {
         .prev   = &fib6_walker_list,
@@ -1432,12 +1435,12 @@ static int fib6_age(struct rt6_info *rt, void *arg)
 
 static DEFINE_SPINLOCK(fib6_gc_lock);
 
-void fib6_run_gc(unsigned long dummy)
+void fib6_run_gc(unsigned long expires, struct net *net)
 {
-        if (dummy != ~0UL) {
+        if (expires != ~0UL) {
                 spin_lock_bh(&fib6_gc_lock);
-                gc_args.timeout = dummy ? (int)dummy :
-                        init_net.ipv6.sysctl.ip6_rt_gc_interval;
+                gc_args.timeout = expires ? (int)expires :
+                        net->ipv6.sysctl.ip6_rt_gc_interval;
         } else {
                 local_bh_disable();
                 if (!spin_trylock(&fib6_gc_lock)) {
@@ -1445,17 +1448,17 @@ void fib6_run_gc(unsigned long dummy)
                         local_bh_enable();
                         return;
                 }
-                gc_args.timeout = init_net.ipv6.sysctl.ip6_rt_gc_interval;
+                gc_args.timeout = net->ipv6.sysctl.ip6_rt_gc_interval;
         }
         gc_args.more = 0;
 
         icmp6_dst_gc(&gc_args.more);
 
-        fib6_clean_all(&init_net, fib6_age, 0, NULL);
+        fib6_clean_all(net, fib6_age, 0, NULL);
 
         if (gc_args.more)
                 mod_timer(&ip6_fib_timer, jiffies +
-                          init_net.ipv6.sysctl.ip6_rt_gc_interval);
+                          net->ipv6.sysctl.ip6_rt_gc_interval);
         else {
                 del_timer(&ip6_fib_timer);
                 ip6_fib_timer.expires = 0;
@@ -1463,6 +1466,11 @@ void fib6_run_gc(unsigned long dummy)
         spin_unlock_bh(&fib6_gc_lock);
 }
 
+static void fib6_gc_timer_cb(unsigned long arg)
+{
+        fib6_run_gc(0, (struct net *)arg);
+}
+
 static int fib6_net_init(struct net *net)
 {
         int ret;
diff --git a/net/ipv6/ndisc.c b/net/ipv6/ndisc.c
index eb322959a3e1..f1c95125100d 100644
--- a/net/ipv6/ndisc.c
+++ b/net/ipv6/ndisc.c
@@ -1598,6 +1598,7 @@ int ndisc_rcv(struct sk_buff *skb)
 static int ndisc_netdev_event(struct notifier_block *this, unsigned long event, void *ptr)
 {
         struct net_device *dev = ptr;
+        struct net *net = dev->nd_net;
 
         if (dev->nd_net != &init_net)
                 return NOTIFY_DONE;
@@ -1605,11 +1606,11 @@ static int ndisc_netdev_event(struct notifier_block *this, unsigned long event,
         switch (event) {
         case NETDEV_CHANGEADDR:
                 neigh_changeaddr(&nd_tbl, dev);
-                fib6_run_gc(~0UL);
+                fib6_run_gc(~0UL, net);
                 break;
         case NETDEV_DOWN:
                 neigh_ifdown(&nd_tbl, dev);
-                fib6_run_gc(~0UL);
+                fib6_run_gc(~0UL, net);
                 break;
         default:
                 break;
diff --git a/net/ipv6/route.c b/net/ipv6/route.c
index 2e6da2afd948..fd44721abebb 100644
--- a/net/ipv6/route.c
+++ b/net/ipv6/route.c
@@ -40,6 +40,7 @@
 #include <linux/if_arp.h>
 #include <linux/proc_fs.h>
 #include <linux/seq_file.h>
+#include <linux/nsproxy.h>
 #include <net/net_namespace.h>
 #include <net/snmp.h>
 #include <net/ipv6.h>
@@ -995,7 +996,7 @@ static int ip6_dst_gc(struct dst_ops *ops)
                 goto out;
 
         expire++;
-        fib6_run_gc(expire);
+        fib6_run_gc(expire, &init_net);
         last_gc = now;
         if (atomic_read(&ip6_dst_ops.entries) < ip6_dst_ops.gc_thresh)
                 expire = init_net.ipv6.sysctl.ip6_rt_gc_timeout>>1;
@@ -2413,10 +2414,11 @@ static
 int ipv6_sysctl_rtcache_flush(ctl_table *ctl, int write, struct file * filp,
                               void __user *buffer, size_t *lenp, loff_t *ppos)
 {
-        int delay = init_net.ipv6.sysctl.flush_delay;
+        struct net *net = current->nsproxy->net_ns;
+        int delay = net->ipv6.sysctl.flush_delay;
         if (write) {
                 proc_dointvec(ctl, write, filp, buffer, lenp, ppos);
-                fib6_run_gc(delay <= 0 ? ~0UL : (unsigned long)delay);
+                fib6_run_gc(delay <= 0 ? ~0UL : (unsigned long)delay, net);
                 return 0;
         } else
                 return -EINVAL;