linux.git/net/core, branch v4.2-rc3

net: call rcu_read_lock early in process_backlog

2015-07-11T01:16:36+00:00

Incoming packet should be either in backlog queue or
in RCU read-side section. Otherwise, the final sequence of
flush_backlog() and synchronize_net() may miss packets
that can run without device reference:

CPU 1                  CPU 2
                       skb->dev: no reference
                       process_backlog:__skb_dequeue
                       process_backlog:local_irq_enable

on_each_cpu for
flush_backlog =>       IPI(hardirq): flush_backlog
                       - packet not found in backlog

                       CPU delayed ...
synchronize_net
- no ongoing RCU
read-side sections

netdev_run_todo,
rcu_barrier: no
ongoing callbacks
                       __netif_receive_skb_core:rcu_read_lock
                       - too late
free dev
                       process packet for freed dev

Fixes: 6e583ce5242f ("net: eliminate refcounting in backlog queue")
Cc: Eric W. Biederman 
Cc: Stephen Hemminger 
Signed-off-by: Julian Anastasov 
Signed-off-by: David S. Miller

net: do not process device backlog during unregistration

2015-07-11T01:16:36+00:00

commit 381c759d9916 ("ipv4: Avoid crashing in ip_error")
fixes a problem where processed packet comes from device
with destroyed inetdev (dev->ip_ptr). This is not expected
because inetdev_destroy is called in NETDEV_UNREGISTER
phase and packets should not be processed after
dev_close_many() and synchronize_net(). Above fix is still
required because inetdev_destroy can be called for other
reasons. But it shows the real problem: backlog can keep
packets for long time and they do not hold reference to
device. Such packets are then delivered to upper levels
at the same time when device is unregistered.
Calling flush_backlog after NETDEV_UNREGISTER_FINAL still
accounts all packets from backlog but before that some packets
continue to be delivered to upper levels long after the
synchronize_net call which is supposed to wait the last
ones. Also, as Eric pointed out, processed packets, mostly
from other devices, can continue to add new packets to backlog.

Fix the problem by moving flush_backlog early, after the
device driver is stopped and before the synchronize_net() call.
Then use netif_running check to make sure we do not add more
packets to backlog. We have to do it in enqueue_to_backlog
context when the local IRQ is disabled. As result, after the
flush_backlog and synchronize_net sequence all packets
should be accounted.

Thanks to Eric W. Biederman for the test script and his
valuable feedback!

Reported-by: Vittorio Gambaletta 
Fixes: 6e583ce5242f ("net: eliminate refcounting in backlog queue")
Cc: Eric W. Biederman 
Cc: Stephen Hemminger 
Signed-off-by: Julian Anastasov 
Signed-off-by: David S. Miller

net: pktgen: kill the "Wait for kthread_stop" code in pktgen_thread_worker()

2015-07-09T22:05:32+00:00

pktgen_thread_worker() doesn't need to wait for kthread_stop(), it
can simply exit. Just pktgen_create_thread() and pg_net_exit() should
do get_task_struct()/put_task_struct(). kthread_stop(dead_thread) is
fine.

Signed-off-by: Oleg Nesterov 
Signed-off-by: David S. Miller

net: pktgen: fix race between pktgen_thread_worker() and kthread_stop()

2015-07-09T22:05:32+00:00

pktgen_thread_worker() is obviously racy, kthread_stop() can come
between the kthread_should_stop() check and set_current_state().

Signed-off-by: Oleg Nesterov 
Reported-by: Jan Stancek 
Reported-by: Marcelo Leitner 
Signed-off-by: David S. Miller

rtnetlink: verify IFLA_VF_INFO attributes before passing them to driver

2015-07-08T23:01:52+00:00

Jason Gunthorpe reported that since commit c02db8c6290b ("rtnetlink: make
SR-IOV VF interface symmetric"), we don't verify IFLA_VF_INFO attributes
anymore with respect to their policy, that is, ifla_vfinfo_policy[].

Before, they were part of ifla_policy[], but they have been nested since
placed under IFLA_VFINFO_LIST, that contains the attribute IFLA_VF_INFO,
which is another nested attribute for the actual VF attributes such as
IFLA_VF_MAC, IFLA_VF_VLAN, etc.

Despite the policy being split out from ifla_policy[] in this commit,
it's never applied anywhere. nla_for_each_nested() only does basic nla_ok()
testing for struct nlattr, but it doesn't know about the data context and
their requirements.

Fix, on top of Jason's initial work, does 1) parsing of the attributes
with the right policy, and 2) using the resulting parsed attribute table
from 1) instead of the nla_for_each_nested() loop (just like we used to
do when still part of ifla_policy[]).

Reference: http://thread.gmane.org/gmane.linux.network/368913
Fixes: c02db8c6290b ("rtnetlink: make SR-IOV VF interface symmetric")
Reported-by: Jason Gunthorpe 
Cc: Chris Wright 
Cc: Sucheta Chakraborty 
Cc: Greg Rose 
Cc: Jeff Kirsher 
Cc: Rony Efraim 
Cc: Vlad Zolotarov 
Cc: Nicolas Dichtel 
Cc: Thomas Graf 
Signed-off-by: Jason Gunthorpe 
Signed-off-by: Daniel Borkmann 
Acked-by: Vlad Zolotarov 
Signed-off-by: David S. Miller

Revert "dev: set iflink to 0 for virtual interfaces"

2015-07-08T22:52:33+00:00

This reverts commit e1622baf54df8cc958bf29d71de5ad545ea7d93c.

The side effect of this commit is to add a '@NONE' after each virtual
interface name with a 'ip link'. It may break existing scripts.

Reported-by: Olivier Hartkopp 
Signed-off-by: Nicolas Dichtel 
Tested-by: Oliver Hartkopp 
Signed-off-by: David S. Miller

net: graceful exit from netif_alloc_netdev_queues()

2015-07-08T22:46:17+00:00

User space can crash kernel with

ip link add ifb10 numtxqueues 100000 type ifb

We must replace a BUG_ON() by proper test and return -EINVAL for
crazy values.

Fixes: 60877a32bce00 ("net: allow large number of tx queues")
Signed-off-by: Eric Dumazet 
Signed-off-by: David S. Miller

net_sched: gen_estimator: extend pps limit

2015-07-08T20:59:20+00:00

rate estimators are limited to 4 Mpps, which was fine years ago, but
too small with current hardware generation.

Lets use 2^5 scaling instead of 2^10 to get 128 Mpps new limit.

On 64bit arch, use an "unsigned long" for temp storage and remove limit.
(We do not expect 32bit arches to be able to reach this point)

Tested:

tc -s -d filter sh dev eth0 parent ffff:

filter protocol ip pref 1 u32
filter protocol ip pref 1 u32 fh 800: ht divisor 1
filter protocol ip pref 1 u32 fh 800::800 order 2048 key ht 800 bkt 0 flowid 1:15
  match 07000000/ff000000 at 12
	action order 1: gact action drop
	 random type none pass val 0
	 index 1 ref 1 bind 1 installed 166 sec
 	Action statistics:
	Sent 39734251496 bytes 863788076 pkt (dropped 863788117, overlimits 0 requeues 0)
	rate 4067Mbit 11053596pps backlog 0b 0p requeues 0

Signed-off-by: Eric Dumazet 
Acked-by: Alexei Starovoitov 
Signed-off-by: David S. Miller

sock_diag: don't broadcast kernel sockets

2015-06-30T17:00:26+00:00

Kernel sockets do not hold a reference for the network namespace to
which they point.  Socket destruction broadcasting relies on the
network namespace and will cause the splat below when a kernel socket
is destroyed.

This fix simply ignores kernel sockets when they are destroyed.

Reported as:
general protection fault: 0000 [#1] PREEMPT SMP DEBUG_PAGEALLOC
CPU: 1 PID: 9130 Comm: kworker/1:1 Not tainted 4.1.0-gelk-debug+ #1
Workqueue: sock_diag_events sock_diag_broadcast_destroy_work
Stack:
 ffff8800b9c586c0 ffff8800b9c586c0 ffff8800ac4692c0 ffff8800936d4a90
 ffff8800352efd38 ffffffff8469a93e ffff8800352efd98 ffffffffc09b9b90
 ffff8800352efd78 ffff8800ac4692c0 ffff8800b9c586c0 ffff8800831b6ab8
Call Trace:
 [] ? mutex_unlock+0xe/0x10
 [] ? inet_diag_handler_get_info+0x110/0x1fb [inet_diag]
 [] netlink_broadcast+0x1d/0x20
 [] ? mutex_unlock+0xe/0x10
 [] sock_diag_broadcast_destroy_work+0xd5/0x160
 [] process_one_work+0x147/0x420
 [] worker_thread+0x69/0x470
 [] ? preempt_count_sub+0xa3/0xf0
 [] ? rescuer_thread+0x320/0x320
 [] kthread+0x107/0x120
 [] ? kthread_create_on_node+0x1b0/0x1b0
 [] ret_from_fork+0x3f/0x70
 [] ? kthread_create_on_node+0x1b0/0x1b0

Tested:
  Using a debug kernel while 'ss -E' is running:
  ip netns add test-ns
  ip netns delete test-ns

Fixes: eb4cb008529c sock_diag: define destruction multicast groups
Fixes: 26abe14379f8 net: Modify sk_alloc to not reference count the
  netns of kernel sockets.
Reported-by: Dave Jones 
Suggested-by: Eric Dumazet 

Signed-off-by: Craig Gallek 
Signed-off-by: David S. Miller

net: Kill sock->sk_protinfo

2015-06-28T23:55:44+00:00

No more users, so it can now be removed.

Signed-off-by: David S. Miller