linux-stable.git/net/ipv4, branch linux-4.17.y

tcp: identify cryptic messages as TCP seq # bugs

2018-08-24T11:07:13+00:00

[ Upstream commit e56b8ce363a36fb7b74b80aaa5cc9084f2c908b4 ]

Attempt to make cryptic TCP seq number error messages clearer by
(1) identifying the source of the message as "TCP", (2) identifying the
errors as "seq # bug", and (3) grouping the field identifiers and values
by separating them with commas.

E.g., the following message is changed from:

recvmsg bug 2: copied 73BCB6CD seq 70F17CBE rcvnxt 73BCB9AA fl 0
WARNING: CPU: 2 PID: 1501 at /linux/net/ipv4/tcp.c:1881 tcp_recvmsg+0x649/0xb90

to:

TCP recvmsg seq # bug 2: copied 73BCB6CD, seq 70F17CBE, rcvnxt 73BCB9AA, fl 0
WARNING: CPU: 2 PID: 1501 at /linux/net/ipv4/tcp.c:2011 tcp_recvmsg+0x694/0xba0

Suggested-by: 積丹尼 Dan Jacobson 
Signed-off-by: Randy Dunlap 
Signed-off-by: Eric Dumazet 
Signed-off-by: David S. Miller 
Signed-off-by: Sasha Levin 
Signed-off-by: Greg Kroah-Hartman

tcp: remove DELAYED ACK events in DCTCP

2018-08-24T11:07:09+00:00

[ Upstream commit a69258f7aa2623e0930212f09c586fd06674ad79 ]

After fixing the way DCTCP tracking delayed ACKs, the delayed-ACK
related callbacks are no longer needed

Signed-off-by: Yuchung Cheng 
Signed-off-by: Eric Dumazet 
Acked-by: Neal Cardwell 
Acked-by: Lawrence Brakmo 
Signed-off-by: David S. Miller 
Signed-off-by: Sasha Levin 
Signed-off-by: Greg Kroah-Hartman

ipfrag: really prevent allocation on netns exit

2018-08-24T11:07:02+00:00

[ Upstream commit f6f2a4a2eb92bc73671204198bb2f8ab53ff59fb ]

Setting the low threshold to 0 has no effect on frags allocation,
we need to clear high_thresh instead.

The code was pre-existent to commit 648700f76b03 ("inet: frags:
use rhashtables for reassembly units"), but before the above,
such assignment had a different role: prevent concurrent eviction
from the worker and the netns cleanup helper.

Fixes: 648700f76b03 ("inet: frags: use rhashtables for reassembly units")
Signed-off-by: Paolo Abeni 
Signed-off-by: David S. Miller 
Signed-off-by: Sasha Levin 
Signed-off-by: Greg Kroah-Hartman

netfilter: x_tables: set module owner for icmp(6) matches

2018-08-24T11:06:59+00:00

[ Upstream commit d376bef9c29b3c65aeee4e785fffcd97ef0a9a81 ]

nft_compat relies on xt_request_find_match to increment
refcount of the module that provides the match/target.

The (builtin) icmp matches did't set the module owner so it
was possible to rmmod ip(6)tables while icmp extensions were still in use.

Signed-off-by: Florian Westphal 
Signed-off-by: Pablo Neira Ayuso 
Signed-off-by: Sasha Levin 
Signed-off-by: Greg Kroah-Hartman

ip_vti: fix a null pointer deferrence when create vti fallback tunnel

2018-08-22T05:44:56+00:00

[ Upstream commit cd1aa9c2c665cafbd05b83507d3f1096f3912aa4 ]

After set fb_tunnels_only_for_init_net to 1, the itn->fb_tunnel_dev will
be NULL and will cause following crash:

[ 2742.849298] BUG: unable to handle kernel NULL pointer dereference at 0000000000000941
[ 2742.851380] PGD 800000042c21a067 P4D 800000042c21a067 PUD 42aaed067 PMD 0
[ 2742.852818] Oops: 0002 [#1] SMP PTI
[ 2742.853570] CPU: 7 PID: 2484 Comm: unshare Kdump: loaded Not tainted 4.18.0-rc8+ #2
[ 2742.855163] Hardware name: Fedora Project OpenStack Nova, BIOS seabios-1.7.5-11.el7 04/01/2014
[ 2742.856970] RIP: 0010:vti_init_net+0x3a/0x50 [ip_vti]
[ 2742.858034] Code: 90 83 c0 48 c7 c2 20 a1 83 c0 48 89 fb e8 6e 3b f6 ff 85 c0 75 22 8b 0d f4 19 00 00 48 8b 93 00 14 00 00 48 8b 14 ca 48 8b 12  82 41 09 00 00 04 c6 82 38 09 00 00 45 5b c3 66 0f 1f 44 00 00
[ 2742.861940] RSP: 0018:ffff9be28207fde0 EFLAGS: 00010246
[ 2742.863044] RAX: 0000000000000000 RBX: ffff8a71ebed4980 RCX: 0000000000000013
[ 2742.864540] RDX: 0000000000000000 RSI: 0000000000000013 RDI: ffff8a71ebed4980
[ 2742.866020] RBP: ffff8a71ea717000 R08: ffffffffc083903c R09: ffff8a71ea717000
[ 2742.867505] R10: 0000000000000000 R11: 0000000000000000 R12: ffff8a71ebed4980
[ 2742.868987] R13: 0000000000000013 R14: ffff8a71ea5b49c0 R15: 0000000000000000
[ 2742.870473] FS:  00007f02266c9740(0000) GS:ffff8a71ffdc0000(0000) knlGS:0000000000000000
[ 2742.872143] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 2742.873340] CR2: 0000000000000941 CR3: 000000042bc20006 CR4: 00000000001606e0
[ 2742.874821] Call Trace:
[ 2742.875358]  ops_init+0x38/0xf0
[ 2742.876078]  setup_net+0xd9/0x1f0
[ 2742.876789]  copy_net_ns+0xb7/0x130
[ 2742.877538]  create_new_namespaces+0x11a/0x1d0
[ 2742.878525]  unshare_nsproxy_namespaces+0x55/0xa0
[ 2742.879526]  ksys_unshare+0x1a7/0x330
[ 2742.880313]  __x64_sys_unshare+0xe/0x20
[ 2742.881131]  do_syscall_64+0x5b/0x180
[ 2742.881933]  entry_SYSCALL_64_after_hwframe+0x44/0xa9

Reproduce:
echo 1 > /proc/sys/net/core/fb_tunnels_only_for_init_net
modprobe ip_vti
unshare -n

Fixes: 79134e6ce2c9 ("net: do not create fallback tunnels for non-default namespaces")
Cc: Eric Dumazet 
Signed-off-by: Haishuang Yan 
Signed-off-by: David S. Miller 
Signed-off-by: Greg Kroah-Hartman

ipv4: frags: handle possible skb truesize change

2018-08-06T14:18:18+00:00

[ Upstream commit 4672694bd4f1aebdab0ad763ae4716e89cb15221 ]

ip_frag_queue() might call pskb_pull() on one skb that
is already in the fragment queue.

We need to take care of possible truesize change, or we
might have an imbalance of the netns frags memory usage.

IPv6 is immune to this bug, because RFC5722, Section 4,
amended by Errata ID 3089 states :

  When reassembling an IPv6 datagram, if
  one or more its constituent fragments is determined to be an
  overlapping fragment, the entire datagram (and any constituent
  fragments) MUST be silently discarded.

Fixes: 158f323b9868 ("net: adjust skb->truesize in pskb_expand_head()")
Signed-off-by: Eric Dumazet 
Signed-off-by: David S. Miller 
Signed-off-by: Greg Kroah-Hartman

inet: frag: enforce memory limits earlier

2018-08-06T14:18:18+00:00

[ Upstream commit 56e2c94f055d328f5f6b0a5c1721cca2f2d4e0a1 ]

We currently check current frags memory usage only when
a new frag queue is created. This allows attackers to first
consume the memory budget (default : 4 MB) creating thousands
of frag queues, then sending tiny skbs to exceed high_thresh
limit by 2 to 3 order of magnitude.

Note that before commit 648700f76b03 ("inet: frags: use rhashtables
for reassembly units"), work queue could be starved under DOS,
getting no cpu cycles.
After commit 648700f76b03, only the per frag queue timer can eventually
remove an incomplete frag queue and its skbs.

Fixes: b13d3cbfb8e8 ("inet: frag: move eviction of queues to work queue")
Signed-off-by: Eric Dumazet 
Reported-by: Jann Horn 
Cc: Florian Westphal 
Cc: Peter Oskolkov 
Cc: Paolo Abeni 
Acked-by: Florian Westphal 
Signed-off-by: David S. Miller 
Signed-off-by: Greg Kroah-Hartman

tcp: ack immediately when a cwr packet arrives

2018-08-03T05:48:06+00:00

[ Upstream commit 9aee40006190a3cda9a4d2dbae71e92617c8c362 ]

We observed high 99 and 99.9% latencies when doing RPCs with DCTCP. The
problem is triggered when the last packet of a request arrives CE
marked. The reply will carry the ECE mark causing TCP to shrink its cwnd
to 1 (because there are no packets in flight). When the 1st packet of
the next request arrives, the ACK was sometimes delayed even though it
is CWR marked, adding up to 40ms to the RPC latency.

This patch insures that CWR marked data packets arriving will be acked
immediately.

Packetdrill script to reproduce the problem:

0.000 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
0.000 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
0.000 setsockopt(3, SOL_TCP, TCP_CONGESTION, "dctcp", 5) = 0
0.000 bind(3, ..., ...) = 0
0.000 listen(3, 1) = 0

0.100 < [ect0] SEW 0:0(0) win 32792 
0.100 > SE. 0:0(0) ack 1 
0.110 < [ect0] . 1:1(0) ack 1 win 257
0.200 accept(3, ..., ...) = 4

0.200 < [ect0] . 1:1001(1000) ack 1 win 257
0.200 > [ect01] . 1:1(0) ack 1001

0.200 write(4, ..., 1) = 1
0.200 > [ect01] P. 1:2(1) ack 1001

0.200 < [ect0] . 1001:2001(1000) ack 2 win 257
0.200 write(4, ..., 1) = 1
0.200 > [ect01] P. 2:3(1) ack 2001

0.200 < [ect0] . 2001:3001(1000) ack 3 win 257
0.200 < [ect0] . 3001:4001(1000) ack 3 win 257
0.200 > [ect01] . 3:3(0) ack 4001

0.210 < [ce] P. 4001:4501(500) ack 3 win 257

+0.001 read(4, ..., 4500) = 4500
+0 write(4, ..., 1) = 1
+0 > [ect01] PE. 3:4(1) ack 4501

+0.010 < [ect0] W. 4501:5501(1000) ack 4 win 257
// Previously the ACK sequence below would be 4501, causing a long RTO
+0.040~+0.045 > [ect01] . 4:4(0) ack 5501   // delayed ack

+0.311 < [ect0] . 5501:6501(1000) ack 4 win 257  // More data
+0 > [ect01] . 4:4(0) ack 6501     // now acks everything

+0.500 < F. 9501:9501(0) ack 4 win 257

Modified based on comments by Neal Cardwell 

Signed-off-by: Lawrence Brakmo 
Acked-by: Neal Cardwell 
Acked-by: Yuchung Cheng 
Signed-off-by: Eric Dumazet 
Signed-off-by: David S. Miller 
Signed-off-by: Greg Kroah-Hartman

tcp: add one more quick ack after after ECN events

2018-08-03T05:48:06+00:00

[ Upstream commit 15ecbe94a45ef88491ca459b26efdd02f91edb6d ]

Larry Brakmo proposal ( https://patchwork.ozlabs.org/patch/935233/
tcp: force cwnd at least 2 in tcp_cwnd_reduction) made us rethink
about our recent patch removing ~16 quick acks after ECN events.

tcp_enter_quickack_mode(sk, 1) makes sure one immediate ack is sent,
but in the case the sender cwnd was lowered to 1, we do not want
to have a delayed ack for the next packet we will receive.

Fixes: 522040ea5fdd ("tcp: do not aggressively quick ack after ECN events")
Signed-off-by: Eric Dumazet 
Reported-by: Neal Cardwell 
Cc: Lawrence Brakmo 
Acked-by: Neal Cardwell 
Signed-off-by: David S. Miller 
Signed-off-by: Greg Kroah-Hartman

tcp: refactor tcp_ecn_check_ce to remove sk type cast

2018-08-03T05:48:06+00:00

[ Upstream commit f4c9f85f3b2cb7669830cd04d0be61192a4d2436 ]

Refactor tcp_ecn_check_ce and __tcp_ecn_check_ce to accept struct sock*
instead of tcp_sock* to clean up type casts. This is a pure refactor
patch.

Signed-off-by: Yousuk Seung 
Signed-off-by: Neal Cardwell 
Signed-off-by: Yuchung Cheng 
Signed-off-by: Eric Dumazet 
Acked-by: Soheil Hassas Yeganeh 
Signed-off-by: David S. Miller 
Signed-off-by: Greg Kroah-Hartman