Since pskb_copy tacks on the non-linear bits from the original
skb, it needs to count them in the truesize field of the new skb.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Chris Wright <chrisw@sous-sol.org>

We have seen a couple of __alloc_pages() failures due to
fragmentation, there is plenty of free memory but no large order pages
available.  I think the problem is in sock_alloc_send_pskb(), the
gfp_mask includes __GFP_REPEAT but its never used/passed to the page
allocator.  Shouldnt the gfp_mask be passed to alloc_skb() ?

Signed-off-by: Larry Woodman <lwoodman@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Chris Wright <chrisw@sous-sol.org>

[NET]: Fix segmentation of linear packets

skb_segment fails to segment linear packets correctly because it
tries to write all linear parts of the original skb into each
segment.  This will always panic as each segment only contains
enough space for one MSS.

This was not detected earlier because linear packets should be
rare for GSO.  In fact it still remains to be seen what exactly
created the linear packets that triggered this bug.  Basically
the only time this should happen is if someone enables GSO
emulation on an interface that does not support SG.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Chris Wright <chrisw@sous-sol.org>

The move of qdisc destruction to a rcu callback broke locking in the
entire qdisc layer by invalidating previously valid assumptions about
the context in which changes to the qdisc tree occur.

The two assumptions were:

- since changes only happen in process context, read_lock doesn't need
  bottem half protection. Now invalid since destruction of inner qdiscs,
  classifiers, actions and estimators happens in the RCU callback unless
  they're manually deleted, resulting in dead-locks when read_lock in
  process context is interrupted by write_lock_bh in bottem half context.

- since changes only happen under the RTNL, no additional locking is
  necessary for data not used during packet processing (f.e. u32_list).
  Again, since destruction now happens in the RCU callback, this assumption
  is not valid anymore, causing races while using this data, which can
  result in corruption or use-after-free.

Instead of "fixing" this by disabling bottem halfs everywhere and adding
new locks/refcounting, this patch makes these assumptions valid again by
moving destruction back to process context. Since only the dev->qdisc
pointer is protected by RCU, but ->enqueue and the qdisc tree are still
protected by dev->qdisc_lock, destruction of the tree can be performed
immediately and only the final free needs to happen in the rcu callback
to make sure dev_queue_xmit doesn't access already freed memory.

Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

neigh_table_clear() doesn't free tbl->stats.
Found by Alexey Kuznetsov. Though Alexey considers this
leak minor for mainstream, I still believe that cleanup
code should not forget to free some of the resources :)

At least, this is critical for OpenVZ with virtualized
neighbour tables.

Signed-Off-By: Kirill Korotaev <dev@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

It causes way too much trouble and confusion in userspace.

Signed-off-by: David S. Miller <davem@davemloft.net>

On High end systems (1024 or so cpus) this can potentially cause stack
overflow.  Fix the stack usage.

Signed-off-by: Suresh Siddha <suresh.b.siddha@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Since __vlan_hwaccel_rx() is essentially bypassing the
netif_receive_skb() call that would have occurred if we did the VLAN
decapsulation in software, we are missing the skb_bond() call and the
assosciated checks it does.

Export those checks via an inline function, skb_bond_should_drop(),
and use this in __vlan_hwaccel_rx().

Signed-off-by: David S. Miller <davem@davemloft.net>

Patch from Dmitry Mishin <dim@openvz.org>:

Replace add_timer() by mod_timer() in dst_run_gc
in order to avoid BUG message.

       CPU1                            CPU2
dst_run_gc()  entered           dst_run_gc() entered
spin_lock(&dst_lock)                   .....
del_timer(&dst_gc_timer)         fail to get lock
       ....                         mod_timer() <--- puts 
                                                 timer back
                                                 to the list
add_timer(&dst_gc_timer) <--- BUG because timer is in list already.

Found during OpenVZ internal testing.

At first we thought that it is OpenVZ specific as we
added dst_run_gc(0) call in dst_dev_event(),
but as Alexey pointed to me it is possible to trigger
this condition in mainstream kernel.

F.e. timer has fired on CPU2, but the handler was preeempted
by an irq before dst_lock is tried.
Meanwhile, someone on CPU1 adds an entry to gc list and
starts the timer.
If CPU2 was preempted long enough, this timer can expire
simultaneously with resuming timer handler on CPU1, arriving
exactly to the situation described.

Signed-off-by: Dmitry Mishin <dim@openvz.org>
Signed-off-by: Kirill Korotaev <dev@openvz.org>
Signed-off-by: Alexey Kuznetsov <kuznet@ms2.inr.ac.ru>
Signed-off-by: David S. Miller <davem@davemloft.net>