linux-stable.git/net, branch v5.4.192

netfilter: nft_socket: only do sk lookups when indev is available

2022-05-09T07:03:27+00:00

commit 743b83f15d4069ea57c3e40996bf4a1077e0cdc1 upstream.

Check if the incoming interface is available and NFT_BREAK
in case neither skb->sk nor input device are set.

Because nf_sk_lookup_slow*() assume packet headers are in the
'in' direction, use in postrouting is not going to yield a meaningful
result.  Same is true for the forward chain, so restrict the use
to prerouting, input and output.

Use in output work if a socket is already attached to the skb.

Fixes: 554ced0a6e29 ("netfilter: nf_tables: add support for native socket matching")
Reported-and-tested-by: Topi Miettinen 
Signed-off-by: Florian Westphal 
Signed-off-by: Pablo Neira Ayuso 
Signed-off-by: Greg Kroah-Hartman

ip6_gre: Avoid updating tunnel->tun_hlen in __gre6_xmit()

2022-05-09T07:03:26+00:00

[ Upstream commit f40c064e933d7787ca7411b699504d7a2664c1f5 ]

Do not update tunnel->tun_hlen in data plane code.  Use a local variable
instead, just like "tunnel_hlen" in net/ipv4/ip_gre.c:gre_fb_xmit().

Co-developed-by: Cong Wang 
Signed-off-by: Cong Wang 
Signed-off-by: Peilin Ye 
Signed-off-by: David S. Miller 
Signed-off-by: Sasha Levin

tcp: fix F-RTO may not work correctly when receiving DSACK

2022-05-09T07:03:26+00:00

[ Upstream commit d9157f6806d1499e173770df1f1b234763de5c79 ]

Currently DSACK is regarded as a dupack, which may cause
F-RTO to incorrectly enter "loss was real" when receiving
DSACK.

Packetdrill to demonstrate:

// Enable F-RTO and TLP
    0 `sysctl -q net.ipv4.tcp_frto=2`
    0 `sysctl -q net.ipv4.tcp_early_retrans=3`
    0 `sysctl -q net.ipv4.tcp_congestion_control=cubic`

// Establish a connection
   +0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
   +0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
   +0 bind(3, ..., ...) = 0
   +0 listen(3, 1) = 0

// RTT 10ms, RTO 210ms
  +.1 < S 0:0(0) win 32792 
   +0 > S. 0:0(0) ack 1 <...>
 +.01 < . 1:1(0) ack 1 win 257
   +0 accept(3, ..., ...) = 4

// Send 2 data segments
   +0 write(4, ..., 2000) = 2000
   +0 > P. 1:2001(2000) ack 1

// TLP
+.022 > P. 1001:2001(1000) ack 1

// Continue to send 8 data segments
   +0 write(4, ..., 10000) = 10000
   +0 > P. 2001:10001(8000) ack 1

// RTO
+.188 > . 1:1001(1000) ack 1

// The original data is acked and new data is sent(F-RTO step 2.b)
   +0 < . 1:1(0) ack 2001 win 257
   +0 > P. 10001:12001(2000) ack 1

// D-SACK caused by TLP is regarded as a dupack, this results in
// the incorrect judgment of "loss was real"(F-RTO step 3.a)
+.022 < . 1:1(0) ack 2001 win 257 

// Never-retransmitted data(3001:4001) are acked and
// expect to switch to open state(F-RTO step 3.b)
   +0 < . 1:1(0) ack 4001 win 257
+0 %{ assert tcpi_ca_state == 0, tcpi_ca_state }%

Fixes: e33099f96d99 ("tcp: implement RFC5682 F-RTO")
Signed-off-by: Pengcheng Yang 
Acked-by: Neal Cardwell 
Tested-by: Neal Cardwell 
Reviewed-by: Eric Dumazet 
Link: https://lore.kernel.org/r/1650967419-2150-1-git-send-email-yangpc@wangsu.com
Signed-off-by: Jakub Kicinski 
Signed-off-by: Sasha Levin

tls: Skip tls_append_frag on zero copy size

2022-05-09T07:03:26+00:00

[ Upstream commit a0df71948e9548de819a6f1da68f5f1742258a52 ]

Calling tls_append_frag when max_open_record_len == record->len might
add an empty fragment to the TLS record if the call happens to be on the
page boundary. Normally tls_append_frag coalesces the zero-sized
fragment to the previous one, but not if it's on page boundary.

If a resync happens then, the mlx5 driver posts dump WQEs in
tx_post_resync_dump, and the empty fragment may become a data segment
with byte_count == 0, which will confuse the NIC and lead to a CQE
error.

This commit fixes the described issue by skipping tls_append_frag on
zero size to avoid adding empty fragments. The fix is not in the driver,
because an empty fragment is hardly the desired behavior.

Fixes: e8f69799810c ("net/tls: Add generic NIC offload infrastructure")
Signed-off-by: Maxim Mikityanskiy 
Reviewed-by: Tariq Toukan 
Link: https://lore.kernel.org/r/20220426154949.159055-1-maximmi@nvidia.com
Signed-off-by: Jakub Kicinski 
Signed-off-by: Sasha Levin

tcp: fix potential xmit stalls caused by TCP_NOTSENT_LOWAT

2022-05-09T07:03:25+00:00

[ Upstream commit 4bfe744ff1644fbc0a991a2677dc874475dd6776 ]

I had this bug sitting for too long in my pile, it is time to fix it.

Thanks to Doug Porter for reminding me of it!

We had various attempts in the past, including commit
0cbe6a8f089e ("tcp: remove SOCK_QUEUE_SHRUNK"),
but the issue is that TCP stack currently only generates
EPOLLOUT from input path, when tp->snd_una has advanced
and skb(s) cleaned from rtx queue.

If a flow has a big RTT, and/or receives SACKs, it is possible
that the notsent part (tp->write_seq - tp->snd_nxt) reaches 0
and no more data can be sent until tp->snd_una finally advances.

What is needed is to also check if POLLOUT needs to be generated
whenever tp->snd_nxt is advanced, from output path.

This bug triggers more often after an idle period, as
we do not receive ACK for at least one RTT. tcp_notsent_lowat
could be a fraction of what CWND and pacing rate would allow to
send during this RTT.

In a followup patch, I will remove the bogus call
to tcp_chrono_stop(sk, TCP_CHRONO_SNDBUF_LIMITED)
from tcp_check_space(). Fact that we have decided to generate
an EPOLLOUT does not mean the application has immediately
refilled the transmit queue. This optimistic call
might have been the reason the bug seemed not too serious.

Tested:

200 ms rtt, 1% packet loss, 32 MB tcp_rmem[2] and tcp_wmem[2]

$ echo 500000 >/proc/sys/net/ipv4/tcp_notsent_lowat
$ cat bench_rr.sh
SUM=0
for i in {1..10}
do
 V=`netperf -H remote_host -l30 -t TCP_RR -- -r 10000000,10000 -o LOCAL_BYTES_SENT | egrep -v "MIGRATED|Bytes"`
 echo $V
 SUM=$(($SUM + $V))
done
echo SUM=$SUM

Before patch:
$ bench_rr.sh
130000000
80000000
140000000
140000000
140000000
140000000
130000000
40000000
90000000
110000000
SUM=1140000000

After patch:
$ bench_rr.sh
430000000
590000000
530000000
450000000
450000000
350000000
450000000
490000000
480000000
460000000
SUM=4680000000  # This is 410 % of the value before patch.

Fixes: c9bee3b7fdec ("tcp: TCP_NOTSENT_LOWAT socket option")
Signed-off-by: Eric Dumazet 
Reported-by: Doug Porter 
Cc: Soheil Hassas Yeganeh 
Cc: Neal Cardwell 
Acked-by: Soheil Hassas Yeganeh 
Signed-off-by: David S. Miller 
Signed-off-by: Sasha Levin

ip_gre: Make o_seqno start from 0 in native mode

2022-05-09T07:03:25+00:00

[ Upstream commit ff827beb706ed719c766acf36449801ded0c17fc ]

For GRE and GRETAP devices, currently o_seqno starts from 1 in native
mode.  According to RFC 2890 2.2., "The first datagram is sent with a
sequence number of 0."  Fix it.

It is worth mentioning that o_seqno already starts from 0 in collect_md
mode, see gre_fb_xmit(), where tunnel->o_seqno is passed to
gre_build_header() before getting incremented.

Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Signed-off-by: Peilin Ye 
Acked-by: William Tu 
Signed-off-by: David S. Miller 
Signed-off-by: Sasha Levin

net/smc: sync err code when tcp connection was refused

2022-05-09T07:03:25+00:00

[ Upstream commit 4e2e65e2e56c6ceb4ea1719360080c0af083229e ]

In the current implementation, when TCP initiates a connection
to an unavailable [ip,port], ECONNREFUSED will be stored in the
TCP socket, but SMC will not. However, some apps (like curl) use
getsockopt(,,SO_ERROR,,) to get the error information, which makes
them miss the error message and behave strangely.

Fixes: 50717a37db03 ("net/smc: nonblocking connect rework")
Signed-off-by: liuyacan 
Reviewed-by: Tony Lu 
Acked-by: Karsten Graul 
Signed-off-by: David S. Miller 
Signed-off-by: Sasha Levin

sctp: check asoc strreset_chunk in sctp_generate_reconf_event

2022-05-09T07:03:24+00:00

[ Upstream commit 165e3e17fe8fe6a8aab319bc6e631a2e23b9a857 ]

A null pointer reference issue can be triggered when the response of a
stream reconf request arrives after the timer is triggered, such as:

  send Incoming SSN Reset Request --->
  CPU0:
   reconf timer is triggered,
   go to the handler code before hold sk lock
                            <--- reply with Outgoing SSN Reset Request
  CPU1:
   process Outgoing SSN Reset Request,
   and set asoc->strreset_chunk to NULL
  CPU0:
   continue the handler code, hold sk lock,
   and try to hold asoc->strreset_chunk, crash!

In Ying Xu's testing, the call trace is:

  [ ] BUG: kernel NULL pointer dereference, address: 0000000000000010
  [ ] RIP: 0010:sctp_chunk_hold+0xe/0x40 [sctp]
  [ ] Call Trace:
  [ ]  
  [ ]  sctp_sf_send_reconf+0x2c/0x100 [sctp]
  [ ]  sctp_do_sm+0xa4/0x220 [sctp]
  [ ]  sctp_generate_reconf_event+0xbd/0xe0 [sctp]
  [ ]  call_timer_fn+0x26/0x130

This patch is to fix it by returning from the timer handler if asoc
strreset_chunk is already set to NULL.

Fixes: 7b9438de0cd4 ("sctp: add stream reconf timer")
Reported-by: Ying Xu 
Signed-off-by: Xin Long 
Acked-by: Marcelo Ricardo Leitner 
Signed-off-by: David S. Miller 
Signed-off-by: Sasha Levin

tcp: ensure to use the most recently sent skb when filling the rate sample

2022-05-09T07:03:24+00:00

[ Upstream commit b253a0680ceadc5d7b4acca7aa2d870326cad8ad ]

If an ACK (s)acks multiple skbs, we favor the information
from the most recently sent skb by choosing the skb with
the highest prior_delivered count. But in the interval
between receiving ACKs, we send multiple skbs with the same
prior_delivered, because the tp->delivered only changes
when we receive an ACK.

We used RACK's solution, copying tcp_rack_sent_after() as
tcp_skb_sent_after() helper to determine "which packet was
sent last?". Later, we will use tcp_skb_sent_after() instead
in RACK.

Fixes: b9f64820fb22 ("tcp: track data delivery rate for a TCP connection")
Signed-off-by: Pengcheng Yang 
Cc: Paolo Abeni 
Acked-by: Neal Cardwell 
Tested-by: Neal Cardwell 
Reviewed-by: Eric Dumazet 
Link: https://lore.kernel.org/r/1650422081-22153-1-git-send-email-yangpc@wangsu.com
Signed-off-by: Jakub Kicinski 
Signed-off-by: Sasha Levin

tcp: md5: incorrect tcp_header_len for incoming connections

2022-05-09T07:03:24+00:00

[ Upstream commit 5b0b9e4c2c895227c8852488b3f09839233bba54 ]

In tcp_create_openreq_child we adjust tcp_header_len for md5 using the
remote address in newsk. But that address is still 0 in newsk at this
point, and it is only set later by the callers (tcp_v[46]_syn_recv_sock).
Use the address from the request socket instead.

Fixes: cfb6eeb4c860 ("[TCP]: MD5 Signature Option (RFC2385) support.")
Signed-off-by: Francesco Ruggeri 
Reviewed-by: Eric Dumazet 
Link: https://lore.kernel.org/r/20220421005026.686A45EC01F2@us226.sjc.aristanetworks.com
Signed-off-by: Jakub Kicinski 
Signed-off-by: Sasha Levin