linux-stable.git/net/core, branch v5.16.9

net, neigh: Do not trigger immediate probes on NUD_FAILED from neigh_managed_work

2022-02-08T17:35:15+00:00

commit 4a81f6da9cb2d1ef911131a6fd8bd15cb61fc772 upstream.

syzkaller was able to trigger a deadlock for NTF_MANAGED entries [0]:

  kworker/0:16/14617 is trying to acquire lock:
  ffffffff8d4dd370 (&tbl->lock){++-.}-{2:2}, at: ___neigh_create+0x9e1/0x2990 net/core/neighbour.c:652
  [...]
  but task is already holding lock:
  ffffffff8d4dd370 (&tbl->lock){++-.}-{2:2}, at: neigh_managed_work+0x35/0x250 net/core/neighbour.c:1572

The neighbor entry turned to NUD_FAILED state, where __neigh_event_send()
triggered an immediate probe as per commit cd28ca0a3dd1 ("neigh: reduce
arp latency") via neigh_probe() given table lock was held.

One option to fix this situation is to defer the neigh_probe() back to
the neigh_timer_handler() similarly as pre cd28ca0a3dd1. For the case
of NTF_MANAGED, this deferral is acceptable given this only happens on
actual failure state and regular / expected state is NUD_VALID with the
entry already present.

The fix adds a parameter to __neigh_event_send() in order to communicate
whether immediate probe is allowed or disallowed. Existing call-sites
of neigh_event_send() default as-is to immediate probe. However, the
neigh_managed_work() disables it via use of neigh_event_send_probe().

[0] 
  __dump_stack lib/dump_stack.c:88 [inline]
  dump_stack_lvl+0xcd/0x134 lib/dump_stack.c:106
  print_deadlock_bug kernel/locking/lockdep.c:2956 [inline]
  check_deadlock kernel/locking/lockdep.c:2999 [inline]
  validate_chain kernel/locking/lockdep.c:3788 [inline]
  __lock_acquire.cold+0x149/0x3ab kernel/locking/lockdep.c:5027
  lock_acquire kernel/locking/lockdep.c:5639 [inline]
  lock_acquire+0x1ab/0x510 kernel/locking/lockdep.c:5604
  __raw_write_lock_bh include/linux/rwlock_api_smp.h:202 [inline]
  _raw_write_lock_bh+0x2f/0x40 kernel/locking/spinlock.c:334
  ___neigh_create+0x9e1/0x2990 net/core/neighbour.c:652
  ip6_finish_output2+0x1070/0x14f0 net/ipv6/ip6_output.c:123
  __ip6_finish_output net/ipv6/ip6_output.c:191 [inline]
  __ip6_finish_output+0x61e/0xe90 net/ipv6/ip6_output.c:170
  ip6_finish_output+0x32/0x200 net/ipv6/ip6_output.c:201
  NF_HOOK_COND include/linux/netfilter.h:296 [inline]
  ip6_output+0x1e4/0x530 net/ipv6/ip6_output.c:224
  dst_output include/net/dst.h:451 [inline]
  NF_HOOK include/linux/netfilter.h:307 [inline]
  ndisc_send_skb+0xa99/0x17f0 net/ipv6/ndisc.c:508
  ndisc_send_ns+0x3a9/0x840 net/ipv6/ndisc.c:650
  ndisc_solicit+0x2cd/0x4f0 net/ipv6/ndisc.c:742
  neigh_probe+0xc2/0x110 net/core/neighbour.c:1040
  __neigh_event_send+0x37d/0x1570 net/core/neighbour.c:1201
  neigh_event_send include/net/neighbour.h:470 [inline]
  neigh_managed_work+0x162/0x250 net/core/neighbour.c:1574
  process_one_work+0x9ac/0x1650 kernel/workqueue.c:2307
  worker_thread+0x657/0x1110 kernel/workqueue.c:2454
  kthread+0x2e9/0x3a0 kernel/kthread.c:377
  ret_from_fork+0x1f/0x30 arch/x86/entry/entry_64.S:295
  

Fixes: 7482e3841d52 ("net, neigh: Add NTF_MANAGED flag for managed neighbor entries")
Reported-by: syzbot+5239d0e1778a500d477a@syzkaller.appspotmail.com
Signed-off-by: Daniel Borkmann 
Cc: Eric Dumazet 
Cc: Roopa Prabhu 
Tested-by: syzbot+5239d0e1778a500d477a@syzkaller.appspotmail.com
Reviewed-by: David Ahern 
Link: https://lore.kernel.org/r/20220201193942.5055-1-daniel@iogearbox.net
Signed-off-by: Jakub Kicinski 
Signed-off-by: Greg Kroah-Hartman

rtnetlink: make sure to refresh master_dev/m_ops in __rtnl_newlink()

2022-02-05T11:39:57+00:00

commit c6f6f2444bdbe0079e41914a35081530d0409963 upstream.

While looking at one unrelated syzbot bug, I found the replay logic
in __rtnl_newlink() to potentially trigger use-after-free.

It is better to clear master_dev and m_ops inside the loop,
in case we have to replay it.

Fixes: ba7d49b1f0f8 ("rtnetlink: provide api for getting and setting slave info")
Signed-off-by: Eric Dumazet 
Cc: Jiri Pirko 
Link: https://lore.kernel.org/r/20220201012106.216495-1-eric.dumazet@gmail.com
Signed-off-by: Jakub Kicinski 
Signed-off-by: Greg Kroah-Hartman

net-procfs: show net devices bound packet types

2022-02-01T16:29:11+00:00

commit 1d10f8a1f40b965d449e8f2d5ed7b96a7c138b77 upstream.

After commit:7866a621043f ("dev: add per net_device packet type chains"),
we can not get packet types that are bound to a specified net device by
/proc/net/ptype, this patch fix the regression.

Run "tcpdump -i ens192 udp -nns0" Before and after apply this patch:

Before:
  [root@localhost ~]# cat /proc/net/ptype
  Type Device      Function
  0800          ip_rcv
  0806          arp_rcv
  86dd          ipv6_rcv

After:
  [root@localhost ~]# cat /proc/net/ptype
  Type Device      Function
  ALL  ens192   tpacket_rcv
  0800          ip_rcv
  0806          arp_rcv
  86dd          ipv6_rcv

v1 -> v2:
  - fix the regression rather than adding new /proc API as
    suggested by Stephen Hemminger.

Fixes: 7866a621043f ("dev: add per net_device packet type chains")
Signed-off-by: Jianguo Wu 
Signed-off-by: David S. Miller 
Signed-off-by: Greg Kroah-Hartman

net: fix information leakage in /proc/net/ptype

2022-02-01T16:29:11+00:00

commit 47934e06b65637c88a762d9c98329ae6e3238888 upstream.

In one net namespace, after creating a packet socket without binding
it to a device, users in other net namespaces can observe the new
`packet_type` added by this packet socket by reading `/proc/net/ptype`
file. This is minor information leakage as packet socket is
namespace aware.

Add a net pointer in `packet_type` to keep the net namespace of
of corresponding packet socket. In `ptype_seq_show`, this net pointer
must be checked when it is not NULL.

Fixes: 2feb27dbe00c ("[NETNS]: Minor information leak via /proc/net/ptype file.")
Signed-off-by: Congyu Liu 
Signed-off-by: David S. Miller 
Signed-off-by: Sasha Levin 
Signed-off-by: Greg Kroah-Hartman

net: fix sock_timestamping_bind_phc() to release device

2022-01-27T11:03:01+00:00

commit 2a4d75bfe41232608f5596a6d1369f92ccb20817 upstream.

Don't forget to release the device in sock_timestamping_bind_phc() after
it was used to get the vclock indices.

Fixes: d463126e23f1 ("net: sock: extend SO_TIMESTAMPING for PHC binding")
Signed-off-by: Miroslav Lichvar 
Cc: Yangbo Lu 
Signed-off-by: David S. Miller 
Signed-off-by: Greg Kroah-Hartman

devlink: Remove misleading internal_flags from health reporter dump

2022-01-27T11:03:01+00:00

commit e9538f8270db24d272659e15841854c7ea11119e upstream.

DEVLINK_CMD_HEALTH_REPORTER_DUMP_GET command doesn't have .doit callback
and has no use in internal_flags at all. Remove this misleading assignment.

Fixes: e44ef4e4516c ("devlink: Hang reporter's dump method on a dumpit cb")
Signed-off-by: Leon Romanovsky 
Signed-off-by: David S. Miller 
Signed-off-by: Greg Kroah-Hartman

netns: add schedule point in ops_exit_list()

2022-01-27T11:02:59+00:00

commit 2836615aa22de55b8fca5e32fe1b27a67cda625e upstream.

When under stress, cleanup_net() can have to dismantle
netns in big numbers. ops_exit_list() currently calls
many helpers [1] that have no schedule point, and we can
end up with soft lockups, particularly on hosts
with many cpus.

Even for moderate amount of netns processed by cleanup_net()
this patch avoids latency spikes.

[1] Some of these helpers like fib_sync_up() and fib_sync_down_dev()
are very slow because net/ipv4/fib_semantics.c uses host-wide hash tables,
and ifindex is used as the only input of two hash functions.
    ifindexes tend to be the same for all netns (lo.ifindex==1 per instance)
    This will be fixed in a separate patch.

Fixes: 72ad937abd0a ("net: Add support for batching network namespace cleanups")
Signed-off-by: Eric Dumazet 
Cc: Eric W. Biederman 
Signed-off-by: David S. Miller 
Signed-off-by: Greg Kroah-Hartman

xdp: check prog type before updating BPF link

2022-01-27T11:02:54+00:00

commit 382778edc8262b7535f00523e9eb22edba1b9816 upstream.

The bpf_xdp_link_update() function didn't check the program type before
updating the program, which made it possible to install any program type as
an XDP program, which is obviously not good. Syzbot managed to trigger this
by swapping in an LWT program on the XDP hook which would crash in a helper
call.

Fix this by adding a check and bailing out if the types don't match.

Fixes: 026a4c28e1db ("bpf, xdp: Implement LINK_UPDATE for BPF XDP link")
Reported-by: syzbot+983941aa85af6ded1fd9@syzkaller.appspotmail.com
Acked-by: Andrii Nakryiko 
Signed-off-by: Toke Høiland-Jørgensen 
Link: https://lore.kernel.org/r/20220107221115.326171-1-toke@redhat.com
Signed-off-by: Alexei Starovoitov 
Signed-off-by: Greg Kroah-Hartman

bpf: Do not WARN in bpf_warn_invalid_xdp_action()

2022-01-27T11:02:23+00:00

[ Upstream commit 2cbad989033bff0256675c38f96f5faab852af4b ]

The WARN_ONCE() in bpf_warn_invalid_xdp_action() can be triggered by
any bugged program, and even attaching a correct program to a NIC
not supporting the given action.

The resulting splat, beyond polluting the logs, fouls automated tools:
e.g. a syzkaller reproducers using an XDP program returning an
unsupported action will never pass validation.

Replace the WARN_ONCE with a less intrusive pr_warn_once().

Signed-off-by: Paolo Abeni 
Signed-off-by: Daniel Borkmann 
Acked-by: Toke Høiland-Jørgensen 
Link: https://lore.kernel.org/bpf/016ceec56e4817ebb2a9e35ce794d5c917df572c.1638189075.git.pabeni@redhat.com
Signed-off-by: Sasha Levin

net-sysfs: update the queue counts in the unregistration path

2022-01-27T11:02:21+00:00

[ Upstream commit d7dac083414eb5bb99a6d2ed53dc2c1b405224e5 ]

When updating Rx and Tx queue kobjects, the queue count should always be
updated to match the queue kobjects count. This was not done in the net
device unregistration path, fix it. Tracking all queue count updates
will allow in a following up patch to detect illegal updates.

Signed-off-by: Antoine Tenart 
Signed-off-by: Jakub Kicinski 
Signed-off-by: Sasha Levin