commit 62d9f1a6945ba69c125e548e72a36d203b30596e upstream.
Upon receiving a cumulative ACK that changes the congestion state from
Disorder to Open, the TLP timer is not set. If the sender is app-limited,
it can only wait for the RTO timer to expire and retransmit.
The reason for this is that the TLP timer is set before the congestion
state changes in tcp_ack(), so we delay the time point of calling
tcp_set_xmit_timer() until after tcp_fastretrans_alert() returns and
remove the FLAG_SET_XMIT_TIMER from ack_flag when the RACK reorder timer
is set.
This commit has two additional benefits:
1) Make sure to reset RTO according to RFC6298 when receiving ACK, to
avoid spurious RTO caused by RTO timer early expires.
2) Reduce the xmit timer reschedule once per ACK when the RACK reorder
timer is set.
Fixes: df92c8394e ("tcp: fix xmit timer to only be reset if data ACKed/SACKed")
Link: https://lore.kernel.org/netdev/1611311242-6675-1-git-send-email-yangpc@wangsu.com
Signed-off-by: Pengcheng Yang <yangpc@wangsu.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Acked-by: Yuchung Cheng <ycheng@google.com>
Cc: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/r/1611464834-23030-1-git-send-email-yangpc@wangsu.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 9d9b1ee0b2d1c9e02b2338c4a4b0a062d2d3edac upstream.
The TCP session does not terminate with TCP_USER_TIMEOUT when data
remain untransmitted due to zero window.
The number of unanswered zero-window probes (tcp_probes_out) is
reset to zero with incoming acks irrespective of the window size,
as described in tcp_probe_timer():
RFC 1122 4.2.2.17 requires the sender to stay open indefinitely
as long as the receiver continues to respond probes. We support
this by default and reset icsk_probes_out with incoming ACKs.
This counter, however, is the wrong one to be used in calculating the
duration that the window remains closed and data remain untransmitted.
Thanks to Jonathan Maxwell <jmaxwell37@gmail.com> for diagnosing the
actual issue.
In this patch a new timestamp is introduced for the socket in order to
track the elapsed time for the zero-window probes that have not been
answered with any non-zero window ack.
Fixes: 9721e709fa ("tcp: simplify window probe aborting on USER_TIMEOUT")
Reported-by: William McCall <william.mccall@gmail.com>
Co-developed-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Enke Chen <enchen@paloaltonetworks.com>
Reviewed-by: Yuchung Cheng <ycheng@google.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/r/20210115223058.GA39267@localhost.localdomain
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit b160c28548bc0a87cbd16d5af6d3edcfd70b8c9a upstream.
Heiner Kallweit reported that some skbs were sent with
the following invalid GSO properties :
- gso_size > 0
- gso_type == 0
This was triggerring a WARN_ON_ONCE() in rtl8169_tso_csum_v2.
Juerg Haefliger was able to reproduce a similar issue using
a lan78xx NIC and a workload mixing TCP incoming traffic
and forwarded packets.
The problem is that tcp_add_backlog() is writing
over gso_segs and gso_size even if the incoming packet will not
be coalesced to the backlog tail packet.
While skb_try_coalesce() would bail out if tail packet is cloned,
this overwriting would lead to corruptions of other packets
cooked by lan78xx, sharing a common super-packet.
The strategy used by lan78xx is to use a big skb, and split
it into all received packets using skb_clone() to avoid copies.
The drawback of this strategy is that all the small skb share a common
struct skb_shared_info.
This patch rewrites TCP gso_size/gso_segs handling to only
happen on the tail skb, since skb_try_coalesce() made sure
it was not cloned.
Fixes: 4f693b55c3 ("tcp: implement coalescing on backlog queue")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Bisected-by: Juerg Haefliger <juergh@canonical.com>
Tested-by: Juerg Haefliger <juergh@canonical.com>
Reported-by: Heiner Kallweit <hkallweit1@gmail.com>
Link: https://bugzilla.kernel.org/show_bug.cgi?id=209423
Link: https://lore.kernel.org/r/20210119164900.766957-1-eric.dumazet@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 8d2b51b008c25240914984208b2ced57d1dd25a5 upstream.
udp_v4_early_demux() is the only function that calls
ip_mc_validate_source() with a TOS that hasn't been masked with
IPTOS_RT_MASK.
This results in different behaviours for incoming multicast UDPv4
packets, depending on if ip_mc_validate_source() is called from the
early-demux path (udp_v4_early_demux) or from the regular input path
(ip_route_input_noref).
ECN would normally not be used with UDP multicast packets, so the
practical consequences should be limited on that side. However,
IPTOS_RT_MASK is used to also masks the TOS' high order bits, to align
with the non-early-demux path behaviour.
Reproducer:
Setup two netns, connected with veth:
$ ip netns add ns0
$ ip netns add ns1
$ ip -netns ns0 link set dev lo up
$ ip -netns ns1 link set dev lo up
$ ip link add name veth01 netns ns0 type veth peer name veth10 netns ns1
$ ip -netns ns0 link set dev veth01 up
$ ip -netns ns1 link set dev veth10 up
$ ip -netns ns0 address add 192.0.2.10 peer 192.0.2.11/32 dev veth01
$ ip -netns ns1 address add 192.0.2.11 peer 192.0.2.10/32 dev veth10
In ns0, add route to multicast address 224.0.2.0/24 using source
address 198.51.100.10:
$ ip -netns ns0 address add 198.51.100.10/32 dev lo
$ ip -netns ns0 route add 224.0.2.0/24 dev veth01 src 198.51.100.10
In ns1, define route to 198.51.100.10, only for packets with TOS 4:
$ ip -netns ns1 route add 198.51.100.10/32 tos 4 dev veth10
Also activate rp_filter in ns1, so that incoming packets not matching
the above route get dropped:
$ ip netns exec ns1 sysctl -wq net.ipv4.conf.veth10.rp_filter=1
Now try to receive packets on 224.0.2.11:
$ ip netns exec ns1 socat UDP-RECVFROM:1111,ip-add-membership=224.0.2.11:veth10,ignoreeof -
In ns0, send packet to 224.0.2.11 with TOS 4 and ECT(0) (that is,
tos 6 for socat):
$ echo test0 | ip netns exec ns0 socat - UDP-DATAGRAM:224.0.2.11:1111,bind=:1111,tos=6
The "test0" message is properly received by socat in ns1, because
early-demux has no cached dst to use, so source address validation
is done by ip_route_input_mc(), which receives a TOS that has the
ECN bits masked.
Now send another packet to 224.0.2.11, still with TOS 4 and ECT(0):
$ echo test1 | ip netns exec ns0 socat - UDP-DATAGRAM:224.0.2.11:1111,bind=:1111,tos=6
The "test1" message isn't received by socat in ns1, because, now,
early-demux has a cached dst to use and calls ip_mc_validate_source()
immediately, without masking the ECN bits.
Fixes: bc044e8db7 ("udp: perform source validation for mcast early demux")
Signed-off-by: Guillaume Nault <gnault@redhat.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 2e5a6266fbb11ae93c468dfecab169aca9c27b43 upstream.
RT_TOS() only masks one of the two ECN bits. Therefore rpfilter_mt()
treats Not-ECT or ECT(1) packets in a different way than those with
ECT(0) or CE.
Reproducer:
Create two netns, connected with a veth:
$ ip netns add ns0
$ ip netns add ns1
$ ip link add name veth01 netns ns0 type veth peer name veth10 netns ns1
$ ip -netns ns0 link set dev veth01 up
$ ip -netns ns1 link set dev veth10 up
$ ip -netns ns0 address add 192.0.2.10/32 dev veth01
$ ip -netns ns1 address add 192.0.2.11/32 dev veth10
Add a route to ns1 in ns0:
$ ip -netns ns0 route add 192.0.2.11/32 dev veth01
In ns1, only packets with TOS 4 can be routed to ns0:
$ ip -netns ns1 route add 192.0.2.10/32 tos 4 dev veth10
Ping from ns0 to ns1 works regardless of the ECN bits, as long as TOS
is 4:
$ ip netns exec ns0 ping -Q 4 192.0.2.11 # TOS 4, Not-ECT
... 0% packet loss ...
$ ip netns exec ns0 ping -Q 5 192.0.2.11 # TOS 4, ECT(1)
... 0% packet loss ...
$ ip netns exec ns0 ping -Q 6 192.0.2.11 # TOS 4, ECT(0)
... 0% packet loss ...
$ ip netns exec ns0 ping -Q 7 192.0.2.11 # TOS 4, CE
... 0% packet loss ...
Now use iptable's rpfilter module in ns1:
$ ip netns exec ns1 iptables-legacy -t raw -A PREROUTING -m rpfilter --invert -j DROP
Not-ECT and ECT(1) packets still pass:
$ ip netns exec ns0 ping -Q 4 192.0.2.11 # TOS 4, Not-ECT
... 0% packet loss ...
$ ip netns exec ns0 ping -Q 5 192.0.2.11 # TOS 4, ECT(1)
... 0% packet loss ...
But ECT(0) and ECN packets are dropped:
$ ip netns exec ns0 ping -Q 6 192.0.2.11 # TOS 4, ECT(0)
... 100% packet loss ...
$ ip netns exec ns0 ping -Q 7 192.0.2.11 # TOS 4, CE
... 100% packet loss ...
After this patch, rpfilter doesn't drop ECT(0) and CE packets anymore.
Fixes: 8f97339d3f ("netfilter: add ipv4 reverse path filter match")
Signed-off-by: Guillaume Nault <gnault@redhat.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 9bd6b629c39e3fa9e14243a6d8820492be1a5b2e ]
esp(6)_output_head uses skb_page_frag_refill to allocate a buffer for
the esp trailer.
It accesses the page with kmap_atomic to handle highmem. But
skb_page_frag_refill can return compound pages, of which
kmap_atomic only maps the first underlying page.
skb_page_frag_refill does not return highmem, because flag
__GFP_HIGHMEM is not set. ESP uses it in the same manner as TCP.
That also does not call kmap_atomic, but directly uses page_address,
in skb_copy_to_page_nocache. Do the same for ESP.
This issue has become easier to trigger with recent kmap local
debugging feature CONFIG_DEBUG_KMAP_LOCAL_FORCE_MAP.
Fixes: cac2661c53 ("esp4: Avoid skb_cow_data whenever possible")
Fixes: 03e2a30f6a ("esp6: Avoid skb_cow_data whenever possible")
Signed-off-by: Willem de Bruijn <willemb@google.com>
Acked-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 7b01e53eee6dce7a8a6736e06b99b68cd0cc7a27 ]
In case of error, remove the nexthop group entry from the list to which
it was previously added.
Fixes: 430a049190 ("nexthop: Add support for nexthop groups")
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 07e61a979ca4dddb3661f59328b3cd109f6b0070 ]
A reference was not taken for the current nexthop entry, so do not try
to put it in the error path.
Fixes: 430a049190 ("nexthop: Add support for nexthop groups")
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit bb4cc1a18856a73f0ff5137df0c2a31f4c50f6cf ]
Conntrack reassembly records the largest fragment size seen in IPCB.
However, when this gets forwarded/transmitted, fragmentation will only
be forced if one of the fragmented packets had the DF bit set.
In that case, a flag in IPCB will force fragmentation even if the
MTU is large enough.
This should work fine, but this breaks with ip tunnels.
Consider client that sends a UDP datagram of size X to another host.
The client fragments the datagram, so two packets, of size y and z, are
sent. DF bit is not set on any of these packets.
Middlebox netfilter reassembles those packets back to single size-X
packet, before routing decision.
packet-size-vs-mtu checks in ip_forward are irrelevant, because DF bit
isn't set. At output time, ip refragmentation is skipped as well
because x is still smaller than the mtu of the output device.
If ttransmit device is an ip tunnel, the packet size increases to
x+overhead.
Also, tunnel might be configured to force DF bit on outer header.
In this case, packet will be dropped (exceeds MTU) and an ICMP error is
generated back to sender.
But sender already respects the announced MTU, all the packets that
it sent did fit the announced mtu.
Force refragmentation as per original sizes unconditionally so ip tunnel
will encapsulate the fragments instead.
The only other solution I see is to place ip refragmentation in
the ip_tunnel code to handle this case.
Fixes: d6b915e29f ("ip_fragment: don't forward defragmented DF packet")
Reported-by: Christian Perle <christian.perle@secunet.com>
Signed-off-by: Florian Westphal <fw@strlen.de>
Acked-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 50c661670f6a3908c273503dfa206dfc7aa54c07 ]
For some reason ip_tunnel insist on setting the DF bit anyway when the
inner header has the DF bit set, EVEN if the tunnel was configured with
'nopmtudisc'.
This means that the script added in the previous commit
cannot be made to work by adding the 'nopmtudisc' flag to the
ip tunnel configuration. Doing so breaks connectivity even for the
without-conntrack/netfilter scenario.
When nopmtudisc is set, the tunnel will skip the mtu check, so no
icmp error is sent to client. Then, because inner header has DF set,
the outer header gets added with DF bit set as well.
IP stack then sends an error to itself because the packet exceeds
the device MTU.
Fixes: 23a3647bc4 ("ip_tunnels: Use skb-len to PMTU check.")
Cc: Stefano Brivio <sbrivio@redhat.com>
Signed-off-by: Florian Westphal <fw@strlen.de>
Acked-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 085c7c4e1c0e50d90b7d90f61a12e12b317a91e2 ]
Both version 0 and version 1 use ETH_P_ERSPAN, but version 0 does not
have an erspan header. So the check in gre_parse_header() is wrong,
we have to distinguish version 1 from version 0.
We can just check the gre header length like is_erspan_type1().
Fixes: cb73ee40b1 ("net: ip_gre: use erspan key field for tunnel lookup")
Reported-by: syzbot+f583ce3d4ddf9836b27a@syzkaller.appspotmail.com
Cc: William Tu <u9012063@gmail.com>
Cc: Lorenzo Bianconi <lorenzo.bianconi@redhat.com>
Signed-off-by: Cong Wang <cong.wang@bytedance.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 21fdca22eb7df2a1e194b8adb812ce370748b733 ]
RT_TOS() only clears one of the ECN bits. Therefore, when
fib_compute_spec_dst() resorts to a fib lookup, it can return
different results depending on the value of the second ECN bit.
For example, ECT(0) and ECT(1) packets could be treated differently.
$ ip netns add ns0
$ ip netns add ns1
$ ip link add name veth01 netns ns0 type veth peer name veth10 netns ns1
$ ip -netns ns0 link set dev lo up
$ ip -netns ns1 link set dev lo up
$ ip -netns ns0 link set dev veth01 up
$ ip -netns ns1 link set dev veth10 up
$ ip -netns ns0 address add 192.0.2.10/24 dev veth01
$ ip -netns ns1 address add 192.0.2.11/24 dev veth10
$ ip -netns ns1 address add 192.0.2.21/32 dev lo
$ ip -netns ns1 route add 192.0.2.10/32 tos 4 dev veth10 src 192.0.2.21
$ ip netns exec ns1 sysctl -wq net.ipv4.icmp_echo_ignore_broadcasts=0
With TOS 4 and ECT(1), ns1 replies using source address 192.0.2.21
(ping uses -Q to set all TOS and ECN bits):
$ ip netns exec ns0 ping -c 1 -b -Q 5 192.0.2.255
[...]
64 bytes from 192.0.2.21: icmp_seq=1 ttl=64 time=0.544 ms
But with TOS 4 and ECT(0), ns1 replies using source address 192.0.2.11
because the "tos 4" route isn't matched:
$ ip netns exec ns0 ping -c 1 -b -Q 6 192.0.2.255
[...]
64 bytes from 192.0.2.11: icmp_seq=1 ttl=64 time=0.597 ms
After this patch the ECN bits don't affect the result anymore:
$ ip netns exec ns0 ping -c 1 -b -Q 6 192.0.2.255
[...]
64 bytes from 192.0.2.21: icmp_seq=1 ttl=64 time=0.591 ms
Fixes: 35ebf65e85 ("ipv4: Create and use fib_compute_spec_dst() helper.")
Signed-off-by: Guillaume Nault <gnault@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit cc00bcaa589914096edef7fb87ca5cee4a166b5c ]
When running concurrent iptables rules replacement with data, the per CPU
sequence count is checked after the assignment of the new information.
The sequence count is used to synchronize with the packet path without the
use of any explicit locking. If there are any packets in the packet path using
the table information, the sequence count is incremented to an odd value and
is incremented to an even after the packet process completion.
The new table value assignment is followed by a write memory barrier so every
CPU should see the latest value. If the packet path has started with the old
table information, the sequence counter will be odd and the iptables
replacement will wait till the sequence count is even prior to freeing the
old table info.
However, this assumes that the new table information assignment and the memory
barrier is actually executed prior to the counter check in the replacement
thread. If CPU decides to execute the assignment later as there is no user of
the table information prior to the sequence check, the packet path in another
CPU may use the old table information. The replacement thread would then free
the table information under it leading to a use after free in the packet
processing context-
Unable to handle kernel NULL pointer dereference at virtual
address 000000000000008e
pc : ip6t_do_table+0x5d0/0x89c
lr : ip6t_do_table+0x5b8/0x89c
ip6t_do_table+0x5d0/0x89c
ip6table_filter_hook+0x24/0x30
nf_hook_slow+0x84/0x120
ip6_input+0x74/0xe0
ip6_rcv_finish+0x7c/0x128
ipv6_rcv+0xac/0xe4
__netif_receive_skb+0x84/0x17c
process_backlog+0x15c/0x1b8
napi_poll+0x88/0x284
net_rx_action+0xbc/0x23c
__do_softirq+0x20c/0x48c
This could be fixed by forcing instruction order after the new table
information assignment or by switching to RCU for the synchronization.
Fixes: 80055dab5d ("netfilter: x_tables: make xt_replace_table wait until old rules are not used anymore")
Reported-by: Sean Tranchetti <stranche@codeaurora.org>
Reported-by: kernel test robot <lkp@intel.com>
Suggested-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Subash Abhinov Kasiviswanathan <subashab@codeaurora.org>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 299bcb55ecd1412f6df606e9dc0912d55610029e ]
When cwnd is not a multiple of the TSO skb size of N*MSS, we can get
into persistent scenarios where we have the following sequence:
(1) ACK for full-sized skb of N*MSS arrives
-> tcp_write_xmit() transmit full-sized skb with N*MSS
-> move pacing release time forward
-> exit tcp_write_xmit() because pacing time is in the future
(2) TSQ callback or TCP internal pacing timer fires
-> try to transmit next skb, but TSO deferral finds remainder of
available cwnd is not big enough to trigger an immediate send
now, so we defer sending until the next ACK.
(3) repeat...
So we can get into a case where we never mark ourselves as
cwnd-limited for many seconds at a time, even with
bulk/infinite-backlog senders, because:
o In case (1) above, every time in tcp_write_xmit() we have enough
cwnd to send a full-sized skb, we are not fully using the cwnd
(because cwnd is not a multiple of the TSO skb size). So every time we
send data, we are not cwnd limited, and so in the cwnd-limited
tracking code in tcp_cwnd_validate() we mark ourselves as not
cwnd-limited.
o In case (2) above, every time in tcp_write_xmit() that we try to
transmit the "remainder" of the cwnd but defer, we set the local
variable is_cwnd_limited to true, but we do not send any packets, so
sent_pkts is zero, so we don't call the cwnd-limited logic to update
tp->is_cwnd_limited.
Fixes: ca8a226343 ("tcp: make cwnd-limited checks measurement-based, and gentler")
Reported-by: Ingemar Johansson <ingemar.s.johansson@ericsson.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/r/20201209035759.1225145-1-ncardwell.kernel@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 72d05c00d7ecda85df29abd046da7e41cc071c17 ]
Before commit a337531b94 ("tcp: up initial rmem to 128KB and SYN rwin to around 64KB")
small tcp_rmem[1] values were overridden by tcp_fixup_rcvbuf() to accommodate various MSS.
This is no longer the case, and Hazem Mohamed Abuelfotoh reported
that DRS would not work for MTU 9000 endpoints receiving regular (1500 bytes) frames.
Root cause is that tcp_init_buffer_space() uses tp->rcv_wnd for upper limit
of rcvq_space.space computation, while it can select later a smaller
value for tp->rcv_ssthresh and tp->window_clamp.
ss -temoi on receiver would show :
skmem:(r0,rb131072,t0,tb46080,f0,w0,o0,bl0,d0) rcv_space:62496 rcv_ssthresh:56596
This means that TCP can not increase its window in tcp_grow_window(),
and that DRS can never kick.
Fix this by making sure that rcvq_space.space is not bigger than number of bytes
that can be held in TCP receive queue.
People unable/unwilling to change their kernel can work around this issue by
selecting a bigger tcp_rmem[1] value as in :
echo "4096 196608 6291456" >/proc/sys/net/ipv4/tcp_rmem
Based on an initial report and patch from Hazem Mohamed Abuelfotoh
https://lore.kernel.org/netdev/20201204180622.14285-1-abuehaze@amazon.com/
Fixes: a337531b94 ("tcp: up initial rmem to 128KB and SYN rwin to around 64KB")
Fixes: 041a14d267 ("tcp: start receiver buffer autotuning sooner")
Reported-by: Hazem Mohamed Abuelfotoh <abuehaze@amazon.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 10c678bd0a035ac2c64a9b26b222f20556227a53 ]
Guillaume noticed that: for segments udp_queue_rcv_one_skb() returns the
proto, and it should pass "ret" unmodified to ip_protocol_deliver_rcu().
Otherwize, with a negtive value passed, it will underflow inet_protos.
This can be reproduced with IPIP FOU:
# ip fou add port 5555 ipproto 4
# ethtool -K eth1 rx-gro-list on
Fixes: cf329aa42b ("udp: cope with UDP GRO packet misdirection")
Reported-by: Guillaume Nault <gnault@redhat.com>
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit b410f04eb5b482b5efc4eee90de81ad35d3d923b ]
Fix to return a negative error code from the error handling
case instead of 0, as done elsewhere in this function.
Fixes: d15662682d ("ipv4: Allow ipv6 gateway with ipv4 routes")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Zhang Changzhong <zhangchangzhong@huawei.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Link: https://lore.kernel.org/r/1607071695-33740-1-git-send-email-zhangchangzhong@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 1ebf179037cb46c19da3a9c1e2ca16e7a754b75e ]
When inet_rtm_getroute() was converted to use the RCU variants of
ip_route_input() and ip_route_output_key(), the TOS parameters
stopped being masked with IPTOS_RT_MASK before doing the route lookup.
As a result, "ip route get" can return a different route than what
would be used when sending real packets.
For example:
$ ip route add 192.0.2.11/32 dev eth0
$ ip route add unreachable 192.0.2.11/32 tos 2
$ ip route get 192.0.2.11 tos 2
RTNETLINK answers: No route to host
But, packets with TOS 2 (ECT(0) if interpreted as an ECN bit) would
actually be routed using the first route:
$ ping -c 1 -Q 2 192.0.2.11
PING 192.0.2.11 (192.0.2.11) 56(84) bytes of data.
64 bytes from 192.0.2.11: icmp_seq=1 ttl=64 time=0.173 ms
--- 192.0.2.11 ping statistics ---
1 packets transmitted, 1 received, 0% packet loss, time 0ms
rtt min/avg/max/mdev = 0.173/0.173/0.173/0.000 ms
This patch re-applies IPTOS_RT_MASK in inet_rtm_getroute(), to
return results consistent with real route lookups.
Fixes: 3765d35ed8 ("net: ipv4: Convert inet_rtm_getroute to rcu versions of route lookup")
Signed-off-by: Guillaume Nault <gnault@redhat.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Link: https://lore.kernel.org/r/b2d237d08317ca55926add9654a48409ac1b8f5b.1606412894.git.gnault@redhat.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 55472017a4219ca965a957584affdb17549ae4a4 ]
When setting congestion control via a BPF program it is seen that the
SYN/ACK for packets within a given flow will not include the ECT0 flag. A
bit of simple printk debugging shows that when this is configured without
BPF we will see the value INET_ECN_xmit value initialized in
tcp_assign_congestion_control however when we configure this via BPF the
socket is in the closed state and as such it isn't configured, and I do not
see it being initialized when we transition the socket into the listen
state. The result of this is that the ECT0 bit is configured based on
whatever the default state is for the socket.
Any easy way to reproduce this is to monitor the following with tcpdump:
tools/testing/selftests/bpf/test_progs -t bpf_tcp_ca
Without this patch the SYN/ACK will follow whatever the default is. If dctcp
all SYN/ACK packets will have the ECT0 bit set, and if it is not then ECT0
will be cleared on all SYN/ACK packets. With this patch applied the SYN/ACK
bit matches the value seen on the other packets in the given stream.
Fixes: 91b5b21c7c ("bpf: Add support for changing congestion control")
Signed-off-by: Alexander Duyck <alexanderduyck@fb.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit c09c8a27b9baa417864b9adc3228b10ae5eeec93 upstream.
Checking for ifdef CONFIG_x fails if CONFIG_x=m.
Use IS_ENABLED instead, which is true for both built-ins and modules.
Otherwise, a
> ip -4 route add 1.2.3.4/32 via inet6 fe80::2 dev eth1
fails with the message "Error: IPv6 support not enabled in kernel." if
CONFIG_IPV6 is `m`.
In the spirit of b8127113d01e53adba15b41aefd37b90ed83d631.
Fixes: d15662682d ("ipv4: Allow ipv6 gateway with ipv4 routes")
Cc: Kim Phillips <kim.phillips@arm.com>
Signed-off-by: Florian Klink <flokli@flokli.de>
Reviewed-by: David Ahern <dsahern@kernel.org>
Link: https://lore.kernel.org/r/20201115224509.2020651-1-flokli@flokli.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 36cd0e696a832a00247fca522034703566ac8885 ]
Fix sockmap sk_skb programs so that they observe sk_rcvbuf limits. This
allows users to tune SO_RCVBUF and sockmap will honor them.
We can refactor the if(charge) case out in later patches. But, keep this
fix to the point.
Fixes: 51199405f9 ("bpf: skb_verdict, support SK_PASS on RX BPF path")
Suggested-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: John Fastabend <john.fastabend@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Jakub Sitnicki <jakub@cloudflare.com>
Link: https://lore.kernel.org/bpf/160556568657.73229.8404601585878439060.stgit@john-XPS-13-9370
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit c9c89dcd872ea33327673fcb97398993a1f22736 ]
If copy_page_to_iter() fails or even partially completes, but with fewer
bytes copied than expected we currently reset sg.start and return EFAULT.
This proves problematic if we already copied data into the user buffer
before we return an error. Because we leave the copied data in the user
buffer and fail to unwind the scatterlist so kernel side believes data
has been copied and user side believes data has _not_ been received.
Expected behavior should be to return number of bytes copied and then
on the next read we need to return the error assuming its still there. This
can happen if we have a copy length spanning multiple scatterlist elements
and one or more complete before the error is hit.
The error is rare enough though that my normal testing with server side
programs, such as nginx, httpd, envoy, etc., I have never seen this. The
only reliable way to reproduce that I've found is to stream movies over
my browser for a day or so and wait for it to hang. Not very scientific,
but with a few extra WARN_ON()s in the code the bug was obvious.
When we review the errors from copy_page_to_iter() it seems we are hitting
a page fault from copy_page_to_iter_iovec() where the code checks
fault_in_pages_writeable(buf, copy) where buf is the user buffer. It
also seems typical server applications don't hit this case.
The other way to try and reproduce this is run the sockmap selftest tool
test_sockmap with data verification enabled, but it doesn't reproduce the
fault. Perhaps we can trigger this case artificially somehow from the
test tools. I haven't sorted out a way to do that yet though.
Fixes: 604326b41a ("bpf, sockmap: convert to generic sk_msg interface")
Signed-off-by: John Fastabend <john.fastabend@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Jakub Sitnicki <jakub@cloudflare.com>
Link: https://lore.kernel.org/bpf/160556566659.73229.15694973114605301063.stgit@john-XPS-13-9370
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 1b9e2a8c99a5c021041bfb2d512dc3ed92a94ffd ]
During loss recovery, retransmitted packets are forced to use TCP
timestamps to calculate the RTT samples, which have a millisecond
granularity. BBR is designed using a microsecond granularity. As a
result, multiple RTT samples could be truncated to the same RTT value
during loss recovery. This is problematic, as BBR will not enter
PROBE_RTT if the RTT sample is <= the current min_rtt sample, meaning
that if there are persistent losses, PROBE_RTT will constantly be
pushed off and potentially never re-entered. This patch makes sure
that BBR enters PROBE_RTT by checking if RTT sample is < the current
min_rtt sample, rather than <=.
The Netflix transport/TCP team discovered this bug in the Linux TCP
BBR code during lab tests.
Fixes: 0f8782ea14 ("tcp_bbr: add BBR congestion control")
Signed-off-by: Ryan Sharpelletti <sharpelletti@google.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Link: https://lore.kernel.org/r/20201116174412.1433277-1-sharpelletti.kdev@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit e33de7c5317e2827b2ba6fd120a505e9eb727b05 ]
nlmsg_cancel() needs to be called in the error path of
inet_req_diag_fill to cancel the message.
Fixes: d545caca82 ("net: inet: diag: expose the socket mark to privileged processes.")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Wang Hai <wanghai38@huawei.com>
Link: https://lore.kernel.org/r/20201116082018.16496-1-wanghai38@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 8cf8821e15cd553339a5b48ee555a0439c2b2742 ]
Commit 58956317c8 ("neighbor: Improve garbage collection")
guarantees neighbour table entries a five-second lifetime. Processes
which make heavy use of multicast can fill the neighour table with
multicast addresses in five seconds. At that point, neighbour entries
can't be GC-ed because they aren't five seconds old yet, the kernel
log starts to fill up with "neighbor table overflow!" messages, and
sends start to fail.
This patch allows multicast addresses to be thrown out before they've
lived out their five seconds. This makes room for non-multicast
addresses and makes messages to all addresses more reliable in these
circumstances.
Fixes: 58956317c8 ("neighbor: Improve garbage collection")
Signed-off-by: Jeff Dike <jdike@akamai.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Link: https://lore.kernel.org/r/20201113015815.31397-1-jdike@akamai.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 909172a149749242990a6e64cb55d55460d4e417 ]
When net.ipv4.tcp_syncookies=1 and syn flood is happened,
cookie_v4_check or cookie_v6_check tries to redo what
tcp_v4_send_synack or tcp_v6_send_synack did,
rsk_window_clamp will be changed if SOCK_RCVBUF is set,
which will make rcv_wscale is different, the client
still operates with initial window scale and can overshot
granted window, the client use the initial scale but local
server use new scale to advertise window value, and session
work abnormally.
Fixes: e88c64f0a4 ("tcp: allow effective reduction of TCP's rcv-buffer via setsockopt")
Signed-off-by: Mao Wenan <wenan.mao@linux.alibaba.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/r/1604967391-123737-1-git-send-email-wenan.mao@linux.alibaba.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 4b1a86281cc1d0de46df3ad2cb8c1f86ac07681c ]
UDP GRO uses udp_hdr(skb) in its .gro_receive() callback. While it's
probably OK for non-frag0 paths (when all headers or even the entire
frame are already in skb head), this inline points to junk when
using Fast GRO (napi_gro_frags() or napi_gro_receive() with only
Ethernet header in skb head and all the rest in the frags) and breaks
GRO packet compilation and the packet flow itself.
To support both modes, skb_gro_header_fast() + skb_gro_header_slow()
are typically used. UDP even has an inline helper that makes use of
them, udp_gro_udphdr(). Use that instead of troublemaking udp_hdr()
to get rid of the out-of-order delivers.
Present since the introduction of plain UDP GRO in 5.0-rc1.
Fixes: e20cf8d3f1 ("udp: implement GRO for plain UDP sockets.")
Cc: Eric Dumazet <edumazet@google.com>
Signed-off-by: Alexander Lobakin <alobakin@pm.me>
Acked-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 46d6c5ae953cc0be38efd0e469284df7c4328cf8 ]
If netfilter changes the packet mark when mangling, the packet is
rerouted using the route_me_harder set of functions. Prior to this
commit, there's one big difference between route_me_harder and the
ordinary initial routing functions, described in the comment above
__ip_queue_xmit():
/* Note: skb->sk can be different from sk, in case of tunnels */
int __ip_queue_xmit(struct sock *sk, struct sk_buff *skb, struct flowi *fl,
That function goes on to correctly make use of sk->sk_bound_dev_if,
rather than skb->sk->sk_bound_dev_if. And indeed the comment is true: a
tunnel will receive a packet in ndo_start_xmit with an initial skb->sk.
It will make some transformations to that packet, and then it will send
the encapsulated packet out of a *new* socket. That new socket will
basically always have a different sk_bound_dev_if (otherwise there'd be
a routing loop). So for the purposes of routing the encapsulated packet,
the routing information as it pertains to the socket should come from
that socket's sk, rather than the packet's original skb->sk. For that
reason __ip_queue_xmit() and related functions all do the right thing.
One might argue that all tunnels should just call skb_orphan(skb) before
transmitting the encapsulated packet into the new socket. But tunnels do
*not* do this -- and this is wisely avoided in skb_scrub_packet() too --
because features like TSQ rely on skb->destructor() being called when
that buffer space is truely available again. Calling skb_orphan(skb) too
early would result in buffers filling up unnecessarily and accounting
info being all wrong. Instead, additional routing must take into account
the new sk, just as __ip_queue_xmit() notes.
So, this commit addresses the problem by fishing the correct sk out of
state->sk -- it's already set properly in the call to nf_hook() in
__ip_local_out(), which receives the sk as part of its normal
functionality. So we make sure to plumb state->sk through the various
route_me_harder functions, and then make correct use of it following the
example of __ip_queue_xmit().
Fixes: 1da177e4c3 ("Linux-2.6.12-rc2")
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
Reviewed-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 20149e9eb68c003eaa09e7c9a49023df40779552 ]
The tunnel device such as vxlan, bareudp and geneve in the lwt mode set
the outer df only based TUNNEL_DONT_FRAGMENT.
And this was also the behavior for gre device before switching to use
ip_md_tunnel_xmit in commit 962924fa2b ("ip_gre: Refactor collect
metatdata mode tunnel xmit to ip_md_tunnel_xmit")
When the ip_gre in lwt mode xmit with ip_md_tunnel_xmi changed the rule and
make the discrepancy between handling of DF by different tunnels. So in the
ip_md_tunnel_xmit should follow the same rule like other tunnels.
Fixes: cfc7381b30 ("ip_tunnel: add collect_md mode to IPIP tunnel")
Signed-off-by: wenxu <wenxu@ucloud.cn>
Link: https://lore.kernel.org/r/1604028728-31100-1-git-send-email-wenxu@ucloud.cn
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 435ccfa894e35e3d4a1799e6ac030e48a7b69ef5 ]
With SO_RCVLOWAT, under memory pressure,
it is possible to enter a state where:
1. We have not received enough bytes to satisfy SO_RCVLOWAT.
2. We have not entered buffer pressure (see tcp_rmem_pressure()).
3. But, we do not have enough buffer space to accept more packets.
In this case, we advertise 0 rwnd (due to #3) but the application does
not drain the receive queue (no wakeup because of #1 and #2) so the
flow stalls.
Modify the heuristic for SO_RCVLOWAT so that, if we are advertising
rwnd<=rcv_mss, force a wakeup to prevent a stall.
Without this patch, setting tcp_rmem to 6143 and disabling TCP
autotune causes a stalled flow. With this patch, no stall occurs. This
is with RPC-style traffic with large messages.
Fixes: 03f45c883c ("tcp: avoid extra wakeups for SO_RCVLOWAT users")
Signed-off-by: Arjun Roy <arjunroy@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/r/20201023184709.217614-1-arjunroy.kdev@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit fdafed459998e2be0e877e6189b24cb7a0183224 ]
GRE tunnel has its own header_ops, ipgre_header_ops, and sets it
conditionally. When it is set, it assumes the outer IP header is
already created before ipgre_xmit().
This is not true when we send packets through a raw packet socket,
where L2 headers are supposed to be constructed by user. Packet
socket calls dev_validate_header() to validate the header. But
GRE tunnel does not set dev->hard_header_len, so that check can
be simply bypassed, therefore uninit memory could be passed down
to ipgre_xmit(). Similar for dev->needed_headroom.
dev->hard_header_len is supposed to be the length of the header
created by dev->header_ops->create(), so it should be used whenever
header_ops is set, and dev->needed_headroom should be used when it
is not set.
Reported-and-tested-by: syzbot+4a2c52677a8a1aa283cb@syzkaller.appspotmail.com
Cc: William Tu <u9012063@gmail.com>
Acked-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Acked-by: Xie He <xie.he.0141@gmail.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 0d9826bc18ce356e8909919ad681ad65d0a6061e ]
Dump vlan tag and proto for the usual vlan offload case if the
NF_LOG_MACDECODE flag is set on. Without this information the logging is
misleading as there is no reference to the VLAN header.
[12716.993704] test: IN=veth0 OUT= MACSRC=86:6c:92:ea:d6:73 MACDST=0e:3b:eb:86:73:76 VPROTO=8100 VID=10 MACPROTO=0800 SRC=192.168.10.2 DST=172.217.168.163 LEN=52 TOS=0x00 PREC=0x00 TTL=64 ID=2548 DF PROTO=TCP SPT=55848 DPT=80 WINDOW=501 RES=0x00 ACK FIN URGP=0
[12721.157643] test: IN=veth0 OUT= MACSRC=86:6c:92:ea:d6:73 MACDST=0e:3b:eb:86:73:76 VPROTO=8100 VID=10 MACPROTO=0806 ARP HTYPE=1 PTYPE=0x0800 OPCODE=2 MACSRC=86:6c:92:ea:d6:73 IPSRC=192.168.10.2 MACDST=0e:3b:eb:86:73:76 IPDST=192.168.10.1
Fixes: 83e96d443b ("netfilter: log: split family specific code to nf_log_{ip,ip6,common}.c files")
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit b38e7819cae946e2edf869e604af1e65a5d241c5 ]
Keyu Man reported that the ICMP rate limiter could be used
by attackers to get useful signal. Details will be provided
in an upcoming academic publication.
Our solution is to add some noise, so that the attackers
no longer can get help from the predictable token bucket limiter.
Fixes: 4cdf507d54 ("icmp: add a global rate limitation")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Keyu Man <kman001@ucr.edu>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 18ded910b589839e38a51623a179837ab4cc3789 ]
In the header prediction fast path for a bulk data receiver, if no
data is newly acknowledged then we do not call tcp_ack() and do not
call tcp_ack_update_window(). This means that a bulk receiver that
receives large amounts of data can have the incoming sequence numbers
wrap, so that the check in tcp_may_update_window fails:
after(ack_seq, tp->snd_wl1)
If the incoming receive windows are zero in this state, and then the
connection that was a bulk data receiver later wants to send data,
that connection can find itself persistently rejecting the window
updates in incoming ACKs. This means the connection can persistently
fail to discover that the receive window has opened, which in turn
means that the connection is unable to send anything, and the
connection's sending process can get permanently "stuck".
The fix is to update snd_wl1 in the header prediction fast path for a
bulk data receiver, so that it keeps up and does not see wrapping
problems.
This fix is based on a very nice and thorough analysis and diagnosis
by Apollon Oikonomopoulos (see link below).
This is a stable candidate but there is no Fixes tag here since the
bug predates current git history. Just for fun: looks like the bug
dates back to when header prediction was added in Linux v2.1.8 in Nov
1996. In that version tcp_rcv_established() was added, and the code
only updates snd_wl1 in tcp_ack(), and in the new "Bulk data transfer:
receiver" code path it does not call tcp_ack(). This fix seems to
apply cleanly at least as far back as v3.2.
Fixes: 1da177e4c3 ("Linux-2.6.12-rc2")
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Reported-by: Apollon Oikonomopoulos <apoikos@dmesg.gr>
Tested-by: Apollon Oikonomopoulos <apoikos@dmesg.gr>
Link: https://www.spinics.net/lists/netdev/msg692430.html
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Acked-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/r/20201022143331.1887495-1-ncardwell.kernel@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit df6afe2f7c19349de2ee560dc62ea4d9ad3ff889 ]
While insertion of 16k nexthops all using the same netdev ('dummy10')
takes less than a second, deletion takes about 130 seconds:
# time -p ip -b nexthop.batch
real 0.29
user 0.01
sys 0.15
# time -p ip link set dev dummy10 down
real 131.03
user 0.06
sys 0.52
This is because of repeated calls to synchronize_rcu() whenever a
nexthop is removed from a nexthop group:
# /usr/share/bcc/tools/offcputime -p `pgrep -nx ip` -K
...
b'finish_task_switch'
b'schedule'
b'schedule_timeout'
b'wait_for_completion'
b'__wait_rcu_gp'
b'synchronize_rcu.part.0'
b'synchronize_rcu'
b'__remove_nexthop'
b'remove_nexthop'
b'nexthop_flush_dev'
b'nh_netdev_event'
b'raw_notifier_call_chain'
b'call_netdevice_notifiers_info'
b'__dev_notify_flags'
b'dev_change_flags'
b'do_setlink'
b'__rtnl_newlink'
b'rtnl_newlink'
b'rtnetlink_rcv_msg'
b'netlink_rcv_skb'
b'rtnetlink_rcv'
b'netlink_unicast'
b'netlink_sendmsg'
b'____sys_sendmsg'
b'___sys_sendmsg'
b'__sys_sendmsg'
b'__x64_sys_sendmsg'
b'do_syscall_64'
b'entry_SYSCALL_64_after_hwframe'
- ip (277)
126554955
Since nexthops are always deleted under RTNL, synchronize_net() can be
used instead. It will call synchronize_rcu_expedited() which only blocks
for several microseconds as opposed to multiple milliseconds like
synchronize_rcu().
With this patch deletion of 16k nexthops takes less than a second:
# time -p ip link set dev dummy10 down
real 0.12
user 0.00
sys 0.04
Tested with fib_nexthops.sh which includes torture tests that prompted
the initial change:
# ./fib_nexthops.sh
...
Tests passed: 134
Tests failed: 0
Fixes: 90f33bffa382 ("nexthops: don't modify published nexthop groups")
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Reviewed-by: David Ahern <dsahern@gmail.com>
Acked-by: Nikolay Aleksandrov <nikolay@nvidia.com>
Link: https://lore.kernel.org/r/20201016172914.643282-1-idosch@idosch.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 874fb9e2ca949b443cc419a4f2227cafd4381d39 ]
Tobias reported regressions in IPsec tests following the patch
referenced by the Fixes tag below. The root cause is dropping the
reset of the flowi4_oif after the fib_lookup. Apparently it is
needed for xfrm cases, so restore the oif update to ip_route_output_flow
right before the call to xfrm_lookup_route.
Fixes: 2fbc6e89b2f1 ("ipv4: Update exception handling for multipath routes via same device")
Reported-by: Tobias Brunner <tobias@strongswan.org>
Signed-off-by: David Ahern <dsahern@kernel.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 86bccd0367130f481ca99ba91de1c6a5aa1c78c1 upstream.
We got reports from GKE customers flows being reset by netfilter
conntrack unless nf_conntrack_tcp_be_liberal is set to 1.
Traces seemed to suggest ACK packet being dropped by the
packet capture, or more likely that ACK were received in the
wrong order.
wscale=7, SYN and SYNACK not shown here.
This ACK allows the sender to send 1871*128 bytes from seq 51359321 :
New right edge of the window -> 51359321+1871*128=51598809
09:17:23.389210 IP A > B: Flags [.], ack 51359321, win 1871, options [nop,nop,TS val 10 ecr 999], length 0
09:17:23.389212 IP B > A: Flags [.], seq 51422681:51424089, ack 1577, win 268, options [nop,nop,TS val 999 ecr 10], length 1408
09:17:23.389214 IP A > B: Flags [.], ack 51422681, win 1376, options [nop,nop,TS val 10 ecr 999], length 0
09:17:23.389253 IP B > A: Flags [.], seq 51424089:51488857, ack 1577, win 268, options [nop,nop,TS val 999 ecr 10], length 64768
09:17:23.389272 IP A > B: Flags [.], ack 51488857, win 859, options [nop,nop,TS val 10 ecr 999], length 0
09:17:23.389275 IP B > A: Flags [.], seq 51488857:51521241, ack 1577, win 268, options [nop,nop,TS val 999 ecr 10], length 32384
Receiver now allows to send 606*128=77568 from seq 51521241 :
New right edge of the window -> 51521241+606*128=51598809
09:17:23.389296 IP A > B: Flags [.], ack 51521241, win 606, options [nop,nop,TS val 10 ecr 999], length 0
09:17:23.389308 IP B > A: Flags [.], seq 51521241:51553625, ack 1577, win 268, options [nop,nop,TS val 999 ecr 10], length 32384
It seems the sender exceeds RWIN allowance, since 51611353 > 51598809
09:17:23.389346 IP B > A: Flags [.], seq 51553625:51611353, ack 1577, win 268, options [nop,nop,TS val 999 ecr 10], length 57728
09:17:23.389356 IP B > A: Flags [.], seq 51611353:51618393, ack 1577, win 268, options [nop,nop,TS val 999 ecr 10], length 7040
09:17:23.389367 IP A > B: Flags [.], ack 51611353, win 0, options [nop,nop,TS val 10 ecr 999], length 0
netfilter conntrack is not happy and sends RST
09:17:23.389389 IP A > B: Flags [R], seq 92176528, win 0, length 0
09:17:23.389488 IP B > A: Flags [R], seq 174478967, win 0, length 0
Now imagine ACK were delivered out of order and tcp_add_backlog() sets window based on wrong packet.
New right edge of the window -> 51521241+859*128=51631193
Normally TCP stack handles OOO packets just fine, but it
turns out tcp_add_backlog() does not. It can update the window
field of the aggregated packet even if the ACK sequence
of the last received packet is too old.
Many thanks to Alexandre Ferrieux for independently reporting the issue
and suggesting a fix.
Fixes: 4f693b55c3 ("tcp: implement coalescing on backlog queue")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Alexandre Ferrieux <alexandre.ferrieux@orange.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit cf83a17edeeb36195596d2dae060a7c381db35f1 upstream.
commit a10674bf24 ("tcp: detecting the misuse of .sendpage for Slab
objects") adds the checks for Slab pages, but the pages don't have
page_count are still missing from the check.
Network layer's sendpage method is not designed to send page_count 0
pages neither, therefore both PageSlab() and page_count() should be
both checked for the sending page. This is exactly what sendpage_ok()
does.
This patch uses sendpage_ok() in do_tcp_sendpages() to detect misused
.sendpage, to make the code more robust.
Fixes: a10674bf24 ("tcp: detecting the misuse of .sendpage for Slab objects")
Suggested-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: Coly Li <colyli@suse.de>
Cc: Vasily Averin <vvs@virtuozzo.com>
Cc: David S. Miller <davem@davemloft.net>
Cc: stable@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit a3ea86739f1bc7e121d921842f0f4a8ab1af94d9 ]
if seq_file .next fuction does not change position index,
read after some lseek can generate unexpected output.
https://bugzilla.kernel.org/show_bug.cgi?id=206283
Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 2fbc6e89b2f1403189e624cabaf73e189c5e50c6 ]
Kfir reported that pmtu exceptions are not created properly for
deployments where multipath routes use the same device.
After some digging I see 2 compounding problems:
1. ip_route_output_key_hash_rcu is updating the flowi4_oif *after*
the route lookup. This is the second use case where this has
been a problem (the first is related to use of vti devices with
VRF). I can not find any reason for the oif to be changed after the
lookup; the code goes back to the start of git. It does not seem
logical so remove it.
2. fib_lookups for exceptions do not call fib_select_path to handle
multipath route selection based on the hash.
The end result is that the fib_lookup used to add the exception
always creates it based using the first leg of the route.
An example topology showing the problem:
| host1
+------+
| eth0 | .209
+------+
|
+------+
switch | br0 |
+------+
|
+---------+---------+
| host2 | host3
+------+ +------+
| eth0 | .250 | eth0 | 192.168.252.252
+------+ +------+
+-----+ +-----+
| vti | .2 | vti | 192.168.247.3
+-----+ +-----+
\ /
=================================
tunnels
192.168.247.1/24
for h in host1 host2 host3; do
ip netns add ${h}
ip -netns ${h} link set lo up
ip netns exec ${h} sysctl -wq net.ipv4.ip_forward=1
done
ip netns add switch
ip -netns switch li set lo up
ip -netns switch link add br0 type bridge stp 0
ip -netns switch link set br0 up
for n in 1 2 3; do
ip -netns switch link add eth-sw type veth peer name eth-h${n}
ip -netns switch li set eth-h${n} master br0 up
ip -netns switch li set eth-sw netns host${n} name eth0
done
ip -netns host1 addr add 192.168.252.209/24 dev eth0
ip -netns host1 link set dev eth0 up
ip -netns host1 route add 192.168.247.0/24 \
nexthop via 192.168.252.250 dev eth0 nexthop via 192.168.252.252 dev eth0
ip -netns host2 addr add 192.168.252.250/24 dev eth0
ip -netns host2 link set dev eth0 up
ip -netns host2 addr add 192.168.252.252/24 dev eth0
ip -netns host3 link set dev eth0 up
ip netns add tunnel
ip -netns tunnel li set lo up
ip -netns tunnel li add br0 type bridge
ip -netns tunnel li set br0 up
for n in $(seq 11 20); do
ip -netns tunnel addr add dev br0 192.168.247.${n}/24
done
for n in 2 3
do
ip -netns tunnel link add vti${n} type veth peer name eth${n}
ip -netns tunnel link set eth${n} mtu 1360 master br0 up
ip -netns tunnel link set vti${n} netns host${n} mtu 1360 up
ip -netns host${n} addr add dev vti${n} 192.168.247.${n}/24
done
ip -netns tunnel ro add default nexthop via 192.168.247.2 nexthop via 192.168.247.3
ip netns exec host1 ping -M do -s 1400 -c3 -I 192.168.252.209 192.168.247.11
ip netns exec host1 ping -M do -s 1400 -c3 -I 192.168.252.209 192.168.247.15
ip -netns host1 ro ls cache
Before this patch the cache always shows exceptions against the first
leg in the multipath route; 192.168.252.250 per this example. Since the
hash has an initial random seed, you may need to vary the final octet
more than what is listed. In my tests, using addresses between 11 and 19
usually found 1 that used both legs.
With this patch, the cache will have exceptions for both legs.
Fixes: 4895c771c7 ("ipv4: Add FIB nexthop exceptions")
Reported-by: Kfir Itzhak <mastertheknife@gmail.com>
Signed-off-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 1869e226a7b3ef75b4f70ede2f1b7229f7157fa4 ]
flowi4_multipath_hash was added by the commit referenced below for
tunnels. Unfortunately, the patch did not initialize the new field
for several fast path lookups that do not initialize the entire flow
struct to 0. Fix those locations. Currently, flowi4_multipath_hash
is random garbage and affects the hash value computed by
fib_multipath_hash for multipath selection.
Fixes: 24ba14406c ("route: Add multipath_hash in flowi_common to make user-define hash")
Signed-off-by: David Ahern <dsahern@gmail.com>
Cc: wenxu <wenxu@ucloud.cn>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit ba9e04a7ddf4f22a10e05bf9403db6b97743c7bf ]
Currently, in tcp_v4_reqsk_send_ack() and tcp_v4_send_reset(), we
echo the TOS value of the received packets in the response.
However, we do not want to echo the lower 2 ECN bits in accordance
with RFC 3168 6.1.5 robustness principles.
Fixes: 1da177e4c3 ("Linux-2.6.12-rc2")
Signed-off-by: Wei Wang <weiwan@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit d76f3351cea2d927fdf70dd7c06898235035e84e ]
In the case of TPROXY, bind_conflict optimizations for SO_REUSEADDR or
SO_REUSEPORT are broken, possibly resulting in O(n) instead of O(1) bind
behaviour or in the incorrect reuse of a bind.
the kernel keeps track for each bind_bucket if all sockets in the
bind_bucket support SO_REUSEADDR or SO_REUSEPORT in two fastreuse flags.
These flags allow skipping the costly bind_conflict check when possible
(meaning when all sockets have the proper SO_REUSE option).
For every socket added to a bind_bucket, these flags need to be updated.
As soon as a socket that does not support reuse is added, the flag is
set to false and will never go back to true, unless the bind_bucket is
deleted.
Note that there is no mechanism to re-evaluate these flags when a socket
is removed (this might make sense when removing a socket that would not
allow reuse; this leaves room for a future patch).
For this optimization to work, it is mandatory that these flags are
properly initialized and updated.
When a child socket is created from a listen socket in
__inet_inherit_port, the TPROXY case could create a new bind bucket
without properly initializing these flags, thus preventing the
optimization to work. Alternatively, a socket not allowing reuse could
be added to an existing bind bucket without updating the flags, causing
bind_conflict to never be called as it should.
Call inet_csk_update_fastreuse when __inet_inherit_port decides to create
a new bind_bucket or use a different bind_bucket than the one of the
listen socket.
Fixes: 093d282321 ("tproxy: fix hash locking issue when using port redirection in __inet_inherit_port()")
Acked-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Tim Froidcoeur <tim.froidcoeur@tessares.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 62ffc589abb176821662efc4525ee4ac0b9c3894 ]
Refactor the fastreuse update code in inet_csk_get_port into a small
helper function that can be called from other places.
Acked-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Tim Froidcoeur <tim.froidcoeur@tessares.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit f19008e676366c44e9241af57f331b6c6edf9552 ]
When TFO keys are read back on big endian systems either via the global
sysctl interface or via getsockopt() using TCP_FASTOPEN_KEY, the values
don't match what was written.
For example, on s390x:
# echo "1-2-3-4" > /proc/sys/net/ipv4/tcp_fastopen_key
# cat /proc/sys/net/ipv4/tcp_fastopen_key
02000000-01000000-04000000-03000000
Instead of:
# cat /proc/sys/net/ipv4/tcp_fastopen_key
00000001-00000002-00000003-00000004
Fix this by converting to the correct endianness on read. This was
reported by Colin Ian King when running the 'tcp_fastopen_backup_key' net
selftest on s390x, which depends on the read value matching what was
written. I've confirmed that the test now passes on big and little endian
systems.
Signed-off-by: Jason Baron <jbaron@akamai.com>
Fixes: 438ac88009 ("net: fastopen: robustness and endianness fixes for SipHash")
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Eric Dumazet <edumazet@google.com>
Reported-and-tested-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 730e700e2c19d87e578ff0e7d8cb1d4a02b036d2 ]
For retransmitted packets, TCP needs to resort to using TCP timestamps
for computing RTT samples. In the common case where the data and ACK
fall in the same 1-millisecond interval, TCP senders with millisecond-
granularity TCP timestamps compute a ca_rtt_us of 0. This ca_rtt_us
of 0 propagates to rs->rtt_us.
This value of 0 can cause performance problems for congestion control
modules. For example, in BBR, the zero min_rtt sample can bring the
min_rtt and BDP estimate down to 0, reduce snd_cwnd and result in a
low throughput. It would be hard to mitigate this with filtering in
the congestion control module, because the proper floor to apply would
depend on the method of RTT sampling (using timestamp options or
internally-saved transmission timestamps).
This fix applies a floor of 1 for the RTT sample delta from TCP
timestamps, so that seq_rtt_us, ca_rtt_us, and rs->rtt_us will be at
least 1 * (USEC_PER_SEC / TCP_TS_HZ).
Note that the receiver RTT computation in tcp_rcv_rtt_measure() and
min_rtt computation in tcp_update_rtt_min() both already apply a floor
of 1 timestamp tick, so this commit makes the code more consistent in
avoiding this edge case of a value of 0.
Signed-off-by: Jianfeng Wang <jfwang@google.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Kevin Yang <yyd@google.com>
Acked-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 622e32b7d4a6492cf5c1f759ef833f817418f7b3 ]
The GRE tunnel can be used to transport traffic that does not rely on a
Internet checksum (e.g. SCTP). The issue can be triggered creating a GRE
or GRETAP tunnel and transmitting SCTP traffic ontop of it where CRC
offload has been disabled. In order to fix the issue we need to
recompute the GRE csum in gre_gso_segment() not relying on the inner
checksum.
The issue is still present when we have the CRC offload enabled.
In this case we need to disable the CRC offload if we require GRE
checksum since otherwise skb_checksum() will report a wrong value.
Fixes: 90017accff ("sctp: Add GSO support")
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Reviewed-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 83f3522860f702748143e022f1a546547314c715 ]
fib_trie_unmerge() is called with RTNL held, but not from an RCU
read-side critical section. This leads to the following warning [1] when
the FIB alias list in a leaf is traversed with
hlist_for_each_entry_rcu().
Since the function is always called with RTNL held and since
modification of the list is protected by RTNL, simply use
hlist_for_each_entry() and silence the warning.
[1]
WARNING: suspicious RCU usage
5.8.0-rc4-custom-01520-gc1f937f3f83b #30 Not tainted
-----------------------------
net/ipv4/fib_trie.c:1867 RCU-list traversed in non-reader section!!
other info that might help us debug this:
rcu_scheduler_active = 2, debug_locks = 1
1 lock held by ip/164:
#0: ffffffff85a27850 (rtnl_mutex){+.+.}-{3:3}, at: rtnetlink_rcv_msg+0x49a/0xbd0
stack backtrace:
CPU: 0 PID: 164 Comm: ip Not tainted 5.8.0-rc4-custom-01520-gc1f937f3f83b #30
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.13.0-2.fc32 04/01/2014
Call Trace:
dump_stack+0x100/0x184
lockdep_rcu_suspicious+0x153/0x15d
fib_trie_unmerge+0x608/0xdb0
fib_unmerge+0x44/0x360
fib4_rule_configure+0xc8/0xad0
fib_nl_newrule+0x37a/0x1dd0
rtnetlink_rcv_msg+0x4f7/0xbd0
netlink_rcv_skb+0x17a/0x480
rtnetlink_rcv+0x22/0x30
netlink_unicast+0x5ae/0x890
netlink_sendmsg+0x98a/0xf40
____sys_sendmsg+0x879/0xa00
___sys_sendmsg+0x122/0x190
__sys_sendmsg+0x103/0x1d0
__x64_sys_sendmsg+0x7d/0xb0
do_syscall_64+0x54/0xa0
entry_SYSCALL_64_after_hwframe+0x44/0xa9
RIP: 0033:0x7fc80a234e97
Code: Bad RIP value.
RSP: 002b:00007ffef8b66798 EFLAGS: 00000246 ORIG_RAX: 000000000000002e
RAX: ffffffffffffffda RBX: 0000000000000000 RCX: 00007fc80a234e97
RDX: 0000000000000000 RSI: 00007ffef8b66800 RDI: 0000000000000003
RBP: 000000005f141b1c R08: 0000000000000001 R09: 0000000000000000
R10: 00007fc80a2a8ac0 R11: 0000000000000246 R12: 0000000000000001
R13: 0000000000000000 R14: 00007ffef8b67008 R15: 0000556fccb10020
Fixes: 0ddcf43d5d ("ipv4: FIB Local/MAIN table collapse")
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit efc6b6f6c3113e8b203b9debfb72d81e0f3dcace ]
Currently, SO_REUSEPORT does not work well if connected sockets are in a
UDP reuseport group.
Then reuseport_has_conns() returns true and the result of
reuseport_select_sock() is discarded. Also, unconnected sockets have the
same score, hence only does the first unconnected socket in udp_hslot
always receive all packets sent to unconnected sockets.
So, the result of reuseport_select_sock() should be used for load
balancing.
The noteworthy point is that the unconnected sockets placed after
connected sockets in sock_reuseport.socks will receive more packets than
others because of the algorithm in reuseport_select_sock().
index | connected | reciprocal_scale | result
---------------------------------------------
0 | no | 20% | 40%
1 | no | 20% | 20%
2 | yes | 20% | 0%
3 | no | 20% | 40%
4 | yes | 20% | 0%
If most of the sockets are connected, this can be a problem, but it still
works better than now.
Fixes: acdcecc612 ("udp: correct reuseport selection with connected sockets")
CC: Willem de Bruijn <willemb@google.com>
Reviewed-by: Benjamin Herrenschmidt <benh@amazon.com>
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.co.jp>
Acked-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 76be93fc0702322179bb0ea87295d820ee46ad14 ]
Previously TLP may send multiple probes of new data in one
flight. This happens when the sender is cwnd limited. After the
initial TLP containing new data is sent, the sender receives another
ACK that acks partial inflight. It may re-arm another TLP timer
to send more, if no further ACK returns before the next TLP timeout
(PTO) expires. The sender may send in theory a large amount of TLP
until send queue is depleted. This only happens if the sender sees
such irregular uncommon ACK pattern. But it is generally undesirable
behavior during congestion especially.
The original TLP design restrict only one TLP probe per inflight as
published in "Reducing Web Latency: the Virtue of Gentle Aggression",
SIGCOMM 2013. This patch changes TLP to send at most one probe
per inflight.
Note that if the sender is app-limited, TLP retransmits old data
and did not have this issue.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit b0a422772fec29811e293c7c0e6f991c0fd9241d ]
We can't use IS_UDPLITE to replace udp_sk->pcflag when UDPLITE_RECV_CC is
checked.
Fixes: b2bf1e2659 ("[UDP]: Clean up for IS_UDPLITE macro")
Signed-off-by: Miaohe Lin <linmiaohe@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 0da7536fb47f51df89ccfcb1fa09f249d9accec5 ]
When no full socket is available, skbs are sent over a per-netns
control socket. Its sk_mark is temporarily adjusted to match that
of the real (request or timewait) socket or to reflect an incoming
skb, so that the outgoing skb inherits this in __ip_make_skb.
Introduction of the socket cookie mark field broke this. Now the
skb is set through the cookie and cork:
<caller> # init sockc.mark from sk_mark or cmsg
ip_append_data
ip_setup_cork # convert sockc.mark to cork mark
ip_push_pending_frames
ip_finish_skb
__ip_make_skb # set skb->mark to cork mark
But I missed these special control sockets. Update all callers of
__ip(6)_make_skb that were originally missed.
For IPv6, the same two icmp(v6) paths are affected. The third
case is not, as commit 92e55f412c ("tcp: don't annotate
mark on control socket from tcp_v6_send_response()") replaced
the ctl_sk->sk_mark with passing the mark field directly as a
function argument. That commit predates the commit that
introduced the bug.
Fixes: c6af0c227a ("ip: support SO_MARK cmsg")
Signed-off-by: Willem de Bruijn <willemb@google.com>
Reported-by: Martin KaFai Lau <kafai@fb.com>
Reviewed-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 1ca0fafd73c5268e8fc4b997094b8bb2bfe8deea ]
This essentially reverts commit 7212303268 ("tcp: md5: reject TCP_MD5SIG
or TCP_MD5SIG_EXT on established sockets")
Mathieu reported that many vendors BGP implementations can
actually switch TCP MD5 on established flows.
Quoting Mathieu :
Here is a list of a few network vendors along with their behavior
with respect to TCP MD5:
- Cisco: Allows for password to be changed, but within the hold-down
timer (~180 seconds).
- Juniper: When password is initially set on active connection it will
reset, but after that any subsequent password changes no network
resets.
- Nokia: No notes on if they flap the tcp connection or not.
- Ericsson/RedBack: Allows for 2 password (old/new) to co-exist until
both sides are ok with new passwords.
- Meta-Switch: Expects the password to be set before a connection is
attempted, but no further info on whether they reset the TCP
connection on a change.
- Avaya: Disable the neighbor, then set password, then re-enable.
- Zebos: Would normally allow the change when socket connected.
We can revert my prior change because commit 9424e2e7ad93 ("tcp: md5: fix potential
overestimation of TCP option space") removed the leak of 4 kernel bytes to
the wire that was the main reason for my patch.
While doing my investigations, I found a bug when a MD5 key is changed, leading
to these commits that stable teams want to consider before backporting this revert :
Commit 6a2febec338d ("tcp: md5: add missing memory barriers in tcp_md5_do_add()/tcp_md5_hash_key()")
Commit e6ced831ef11 ("tcp: md5: refine tcp_md5_do_add()/tcp_md5_hash_key() barriers")
Fixes: 7212303268 "tcp: md5: reject TCP_MD5SIG or TCP_MD5SIG_EXT on established sockets"
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit e6ced831ef11a2a06e8d00aad9d4fc05b610bf38 ]
My prior fix went a bit too far, according to Herbert and Mathieu.
Since we accept that concurrent TCP MD5 lookups might see inconsistent
keys, we can use READ_ONCE()/WRITE_ONCE() instead of smp_rmb()/smp_wmb()
Clearing all key->key[] is needed to avoid possible KMSAN reports,
if key->keylen is increased. Since tcp_md5_do_add() is not fast path,
using __GFP_ZERO to clear all struct tcp_md5sig_key is simpler.
data_race() was added in linux-5.8 and will prevent KCSAN reports,
this can safely be removed in stable backports, if data_race() is
not yet backported.
v2: use data_race() both in tcp_md5_hash_key() and tcp_md5_do_add()
Fixes: 6a2febec338d ("tcp: md5: add missing memory barriers in tcp_md5_do_add()/tcp_md5_hash_key()")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Cc: Marco Elver <elver@google.com>
Reviewed-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Acked-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit e114e1e8ac9d31f25b9dd873bab5d80c1fc482ca ]
Whenever cookie_init_timestamp() has been used to encode
ECN,SACK,WSCALE options, we can not remove the TS option in the SYNACK.
Otherwise, tcp_synack_options() will still advertize options like WSCALE
that we can not deduce later when receiving the packet from the client
to complete 3WHS.
Note that modern linux TCP stacks wont use MD5+TS+SACK in a SYN packet,
but we can not know for sure that all TCP stacks have the same logic.
Before the fix a tcpdump would exhibit this wrong exchange :
10:12:15.464591 IP C > S: Flags [S], seq 4202415601, win 65535, options [nop,nop,md5 valid,mss 1400,sackOK,TS val 456965269 ecr 0,nop,wscale 8], length 0
10:12:15.464602 IP S > C: Flags [S.], seq 253516766, ack 4202415602, win 65535, options [nop,nop,md5 valid,mss 1400,nop,nop,sackOK,nop,wscale 8], length 0
10:12:15.464611 IP C > S: Flags [.], ack 1, win 256, options [nop,nop,md5 valid], length 0
10:12:15.464678 IP C > S: Flags [P.], seq 1:13, ack 1, win 256, options [nop,nop,md5 valid], length 12
10:12:15.464685 IP S > C: Flags [.], ack 13, win 65535, options [nop,nop,md5 valid], length 0
After this patch the exchange looks saner :
11:59:59.882990 IP C > S: Flags [S], seq 517075944, win 65535, options [nop,nop,md5 valid,mss 1400,sackOK,TS val 1751508483 ecr 0,nop,wscale 8], length 0
11:59:59.883002 IP S > C: Flags [S.], seq 1902939253, ack 517075945, win 65535, options [nop,nop,md5 valid,mss 1400,sackOK,TS val 1751508479 ecr 1751508483,nop,wscale 8], length 0
11:59:59.883012 IP C > S: Flags [.], ack 1, win 256, options [nop,nop,md5 valid,nop,nop,TS val 1751508483 ecr 1751508479], length 0
11:59:59.883114 IP C > S: Flags [P.], seq 1:13, ack 1, win 256, options [nop,nop,md5 valid,nop,nop,TS val 1751508483 ecr 1751508479], length 12
11:59:59.883122 IP S > C: Flags [.], ack 13, win 256, options [nop,nop,md5 valid,nop,nop,TS val 1751508483 ecr 1751508483], length 0
11:59:59.883152 IP S > C: Flags [P.], seq 1:13, ack 13, win 256, options [nop,nop,md5 valid,nop,nop,TS val 1751508484 ecr 1751508483], length 12
11:59:59.883170 IP C > S: Flags [.], ack 13, win 256, options [nop,nop,md5 valid,nop,nop,TS val 1751508484 ecr 1751508484], length 0
Of course, no SACK block will ever be added later, but nothing should break.
Technically, we could remove the 4 nops included in MD5+TS options,
but again some stacks could break seeing not conventional alignment.
Fixes: 4957faade1 ("TCPCT part 1g: Responder Cookie => Initiator")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Florian Westphal <fw@strlen.de>
Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 6a2febec338df7e7699a52d00b2e1207dcf65b28 ]
MD5 keys are read with RCU protection, and tcp_md5_do_add()
might update in-place a prior key.
Normally, typical RCU updates would allocate a new piece
of memory. In this case only key->key and key->keylen might
be updated, and we do not care if an incoming packet could
see the old key, the new one, or some intermediate value,
since changing the key on a live flow is known to be problematic
anyway.
We only want to make sure that in the case key->keylen
is changed, cpus in tcp_md5_hash_key() wont try to use
uninitialized data, or crash because key->keylen was
read twice to feed sg_init_one() and ahash_request_set_crypt()
Fixes: 9ea88a1530 ("tcp: md5: check md5 signature without socket lock")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit ba3bb0e76ccd464bb66665a1941fabe55dadb3ba ]
Whenever tcp_try_rmem_schedule() returns an error, we are under
trouble and should make sure to wakeup readers so that they
can drain socket queues and eventually make room.
Fixes: 03f45c883c ("tcp: avoid extra wakeups for SO_RCVLOWAT users")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 5eff06902394425c722f0a44d9545909a8800f79 ]
IPv4 ping sockets don't set fl4.fl4_icmp_{type,code}, which leads to
incomplete IPsec ACQUIRE messages being sent to userspace. Currently,
both raw sockets and IPv6 ping sockets set those fields.
Expected output of "ip xfrm monitor":
acquire proto esp
sel src 10.0.2.15/32 dst 8.8.8.8/32 proto icmp type 8 code 0 dev ens4
policy src 10.0.2.15/32 dst 8.8.8.8/32
<snip>
Currently with ping sockets:
acquire proto esp
sel src 10.0.2.15/32 dst 8.8.8.8/32 proto icmp type 0 code 0 dev ens4
policy src 10.0.2.15/32 dst 8.8.8.8/32
<snip>
The Libreswan test suite found this problem after Fedora changed the
value for the sysctl net.ipv4.ping_group_range.
Fixes: c319b4d76b ("net: ipv4: add IPPROTO_ICMP socket kind")
Reported-by: Paul Wouters <pwouters@redhat.com>
Tested-by: Paul Wouters <pwouters@redhat.com>
Signed-off-by: Sabrina Dubroca <sd@queasysnail.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit b344579ca8478598937215f7005d6c7b84d28aee ]
Mirja Kuehlewind reported a bug in Linux TCP CUBIC Hystart, where
Hystart HYSTART_DELAY mechanism can exit Slow Start spuriously on an
ACK when the minimum rtt of a connection goes down. From inspection it
is clear from the existing code that this could happen in an example
like the following:
o The first 8 RTT samples in a round trip are 150ms, resulting in a
curr_rtt of 150ms and a delay_min of 150ms.
o The 9th RTT sample is 100ms. The curr_rtt does not change after the
first 8 samples, so curr_rtt remains 150ms. But delay_min can be
lowered at any time, so delay_min falls to 100ms. The code executes
the HYSTART_DELAY comparison between curr_rtt of 150ms and delay_min
of 100ms, and the curr_rtt is declared far enough above delay_min to
force a (spurious) exit of Slow start.
The fix here is simple: allow every RTT sample in a round trip to
lower the curr_rtt.
Fixes: ae27e98a51 ("[TCP] CUBIC v2.3")
Reported-by: Mirja Kuehlewind <mirja.kuehlewind@ericsson.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit ba61539c6ae57f4146284a5cb4f7b7ed8d42bf45 ]
In the datapath, the ip_tunnel_lookup() is used and it internally uses
fallback tunnel device pointer, which is fb_tunnel_dev.
This pointer variable should be set to NULL when a fb interface is deleted.
But there is no routine to set fb_tunnel_dev pointer to NULL.
So, this pointer will be still used after interface is deleted and
it eventually results in the use-after-free problem.
Test commands:
ip netns add A
ip netns add B
ip link add eth0 type veth peer name eth1
ip link set eth0 netns A
ip link set eth1 netns B
ip netns exec A ip link set lo up
ip netns exec A ip link set eth0 up
ip netns exec A ip link add gre1 type gre local 10.0.0.1 \
remote 10.0.0.2
ip netns exec A ip link set gre1 up
ip netns exec A ip a a 10.0.100.1/24 dev gre1
ip netns exec A ip a a 10.0.0.1/24 dev eth0
ip netns exec B ip link set lo up
ip netns exec B ip link set eth1 up
ip netns exec B ip link add gre1 type gre local 10.0.0.2 \
remote 10.0.0.1
ip netns exec B ip link set gre1 up
ip netns exec B ip a a 10.0.100.2/24 dev gre1
ip netns exec B ip a a 10.0.0.2/24 dev eth1
ip netns exec A hping3 10.0.100.2 -2 --flood -d 60000 &
ip netns del B
Splat looks like:
[ 77.793450][ C3] ==================================================================
[ 77.794702][ C3] BUG: KASAN: use-after-free in ip_tunnel_lookup+0xcc4/0xf30
[ 77.795573][ C3] Read of size 4 at addr ffff888060bd9c84 by task hping3/2905
[ 77.796398][ C3]
[ 77.796664][ C3] CPU: 3 PID: 2905 Comm: hping3 Not tainted 5.8.0-rc1+ #616
[ 77.797474][ C3] Hardware name: innotek GmbH VirtualBox/VirtualBox, BIOS VirtualBox 12/01/2006
[ 77.798453][ C3] Call Trace:
[ 77.798815][ C3] <IRQ>
[ 77.799142][ C3] dump_stack+0x9d/0xdb
[ 77.799605][ C3] print_address_description.constprop.7+0x2cc/0x450
[ 77.800365][ C3] ? ip_tunnel_lookup+0xcc4/0xf30
[ 77.800908][ C3] ? ip_tunnel_lookup+0xcc4/0xf30
[ 77.801517][ C3] ? ip_tunnel_lookup+0xcc4/0xf30
[ 77.802145][ C3] kasan_report+0x154/0x190
[ 77.802821][ C3] ? ip_tunnel_lookup+0xcc4/0xf30
[ 77.803503][ C3] ip_tunnel_lookup+0xcc4/0xf30
[ 77.804165][ C3] __ipgre_rcv+0x1ab/0xaa0 [ip_gre]
[ 77.804862][ C3] ? rcu_read_lock_sched_held+0xc0/0xc0
[ 77.805621][ C3] gre_rcv+0x304/0x1910 [ip_gre]
[ 77.806293][ C3] ? lock_acquire+0x1a9/0x870
[ 77.806925][ C3] ? gre_rcv+0xfe/0x354 [gre]
[ 77.807559][ C3] ? erspan_xmit+0x2e60/0x2e60 [ip_gre]
[ 77.808305][ C3] ? rcu_read_lock_sched_held+0xc0/0xc0
[ 77.809032][ C3] ? rcu_read_lock_held+0x90/0xa0
[ 77.809713][ C3] gre_rcv+0x1b8/0x354 [gre]
[ ... ]
Suggested-by: Eric Dumazet <eric.dumazet@gmail.com>
Fixes: c544193214 ("GRE: Refactor GRE tunneling code.")
Signed-off-by: Taehee Yoo <ap420073@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 662051215c758ae8545451628816204ed6cd372d ]
Back in 2013, we made a change that broke fast retransmit
for non SACK flows.
Indeed, for these flows, a sender needs to receive three duplicate
ACK before starting fast retransmit. Sending ACK with different
receive window do not count.
Even if enabling SACK is strongly recommended these days,
there still are some cases where it has to be disabled.
Not increasing the window seems better than having to
rely on RTO.
After the fix, following packetdrill test gives :
// Initialize 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
+0 < S 0:0(0) win 32792 <mss 1000,nop,wscale 7>
+0 > S. 0:0(0) ack 1 <mss 1460,nop,wscale 8>
+0 < . 1:1(0) ack 1 win 514
+0 accept(3, ..., ...) = 4
+0 < . 1:1001(1000) ack 1 win 514
// Quick ack
+0 > . 1:1(0) ack 1001 win 264
+0 < . 2001:3001(1000) ack 1 win 514
// DUPACK : Normally we should not change the window
+0 > . 1:1(0) ack 1001 win 264
+0 < . 3001:4001(1000) ack 1 win 514
// DUPACK : Normally we should not change the window
+0 > . 1:1(0) ack 1001 win 264
+0 < . 4001:5001(1000) ack 1 win 514
// DUPACK : Normally we should not change the window
+0 > . 1:1(0) ack 1001 win 264
+0 < . 1001:2001(1000) ack 1 win 514
// Hole is repaired.
+0 > . 1:1(0) ack 5001 win 272
Fixes: 4e4f1fc226 ("tcp: properly increase rcv_ssthresh for ofo packets")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Venkat Venkatsubra <venkat.x.venkatsubra@oracle.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 2570284060b48f3f79d8f1a2698792f36c385e9a ]
there is a problem with the CWR flag set in an incoming ACK segment
and it leads to the situation when the ECE flag is latched forever
the following packetdrill script shows what happens:
// Stack receives incoming segments with CE set
+0.1 <[ect0] . 11001:12001(1000) ack 1001 win 65535
+0.0 <[ce] . 12001:13001(1000) ack 1001 win 65535
+0.0 <[ect0] P. 13001:14001(1000) ack 1001 win 65535
// Stack repsonds with ECN ECHO
+0.0 >[noecn] . 1001:1001(0) ack 12001
+0.0 >[noecn] E. 1001:1001(0) ack 13001
+0.0 >[noecn] E. 1001:1001(0) ack 14001
// Write a packet
+0.1 write(3, ..., 1000) = 1000
+0.0 >[ect0] PE. 1001:2001(1000) ack 14001
// Pure ACK received
+0.01 <[noecn] W. 14001:14001(0) ack 2001 win 65535
// Since CWR was sent, this packet should NOT have ECE set
+0.1 write(3, ..., 1000) = 1000
+0.0 >[ect0] P. 2001:3001(1000) ack 14001
// but Linux will still keep ECE latched here, with packetdrill
// flagging a missing ECE flag, expecting
// >[ect0] PE. 2001:3001(1000) ack 14001
// in the script
In the situation above we will continue to send ECN ECHO packets
and trigger the peer to reduce the congestion window. To avoid that
we can check CWR on pure ACKs received.
v3:
- Add a sequence check to avoid sending an ACK to an ACK
v2:
- Adjusted the comment
- move CWR check before checking for unacknowledged packets
Signed-off-by: Denis Kirjanov <denis.kirjanov@suse.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 5eea3a63ff4aba6a26002e657a6d21934b7e2b96 ]
ie.,
$ ifconfig eth0 6.6.6.6 netmask 255.255.255.0
$ ip rule add from 6.6.6.6 table 6666
$ ip route add 9.9.9.9 via 6.6.6.6
$ ping -I 6.6.6.6 9.9.9.9
PING 9.9.9.9 (9.9.9.9) from 6.6.6.6 : 56(84) bytes of data.
3 packets transmitted, 0 received, 100% packet loss, time 2079ms
$ arp
Address HWtype HWaddress Flags Mask Iface
6.6.6.6 (incomplete) eth0
The arp request address is error, this is because fib_table_lookup in
fib_check_nh lookup the destnation 9.9.9.9 nexthop, the scope of
the fib result is RT_SCOPE_LINK,the correct scope is RT_SCOPE_HOST.
Here I add a check of whether this is RT_TABLE_MAIN to solve this problem.
Fixes: 3bfd847203 ("net: Use passed in table for nexthop lookups")
Signed-off-by: guodeqing <geffrey.guo@huawei.com>
Reviewed-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 487082fb7bd2a32b66927d2b22e3a81b072b44f0 ]
When user application calls read() with MSG_PEEK flag to read data
of bpf sockmap socket, kernel panic happens at
__tcp_bpf_recvmsg+0x12c/0x350. sk_msg is not removed from ingress_msg
queue after read out under MSG_PEEK flag is set. Because it's not
judged whether sk_msg is the last msg of ingress_msg queue, the next
sk_msg may be the head of ingress_msg queue, whose memory address of
sg page is invalid. So it's necessary to add check codes to prevent
this problem.
[20759.125457] BUG: kernel NULL pointer dereference, address:
0000000000000008
[20759.132118] CPU: 53 PID: 51378 Comm: envoy Tainted: G E
5.4.32 #1
[20759.140890] Hardware name: Inspur SA5212M4/YZMB-00370-109, BIOS
4.1.12 06/18/2017
[20759.149734] RIP: 0010:copy_page_to_iter+0xad/0x300
[20759.270877] __tcp_bpf_recvmsg+0x12c/0x350
[20759.276099] tcp_bpf_recvmsg+0x113/0x370
[20759.281137] inet_recvmsg+0x55/0xc0
[20759.285734] __sys_recvfrom+0xc8/0x130
[20759.290566] ? __audit_syscall_entry+0x103/0x130
[20759.296227] ? syscall_trace_enter+0x1d2/0x2d0
[20759.301700] ? __audit_syscall_exit+0x1e4/0x290
[20759.307235] __x64_sys_recvfrom+0x24/0x30
[20759.312226] do_syscall_64+0x55/0x1b0
[20759.316852] entry_SYSCALL_64_after_hwframe+0x44/0xa9
Signed-off-by: dihu <anny.hu@linux.alibaba.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Acked-by: Jakub Sitnicki <jakub@cloudflare.com>
Link: https://lore.kernel.org/bpf/20200605084625.9783-1-anny.hu@linux.alibaba.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit fbe4e0c1b298b4665ee6915266c9d6c5b934ef4a ]
fib_triestat_seq_show() calls hlist_for_each_entry_rcu(tb, head,
tb_hlist) without rcu_read_lock() will trigger a warning,
net/ipv4/fib_trie.c:2579 RCU-list traversed in non-reader section!!
other info that might help us debug this:
rcu_scheduler_active = 2, debug_locks = 1
1 lock held by proc01/115277:
#0: c0000014507acf00 (&p->lock){+.+.}-{3:3}, at: seq_read+0x58/0x670
Call Trace:
dump_stack+0xf4/0x164 (unreliable)
lockdep_rcu_suspicious+0x140/0x164
fib_triestat_seq_show+0x750/0x880
seq_read+0x1a0/0x670
proc_reg_read+0x10c/0x1b0
__vfs_read+0x3c/0x70
vfs_read+0xac/0x170
ksys_read+0x7c/0x140
system_call+0x5c/0x68
Fix it by adding a pair of rcu_read_lock/unlock() and use
cond_resched_rcu() to avoid the situation where walking of a large
number of items may prevent scheduling for a long time.
Signed-off-by: Qian Cai <cai@lca.pw>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 1b49cd71b52403822731dc9f283185d1da355f97 ]
When devinet_sysctl_register() failed, the memory allocated
in neigh_parms_alloc() should be freed.
Fixes: 20e61da7ff ("ipv4: fail early when creating netdev named all or default")
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Acked-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 1fd1c768f3624a5e66766e7b4ddb9b607cd834a5 upstream.
Similar to the last path, need to fix fib_info_nh_uses_dev for
external nexthops to avoid referencing multiple nh_grp structs.
Move the device check in fib_info_nh_uses_dev to a helper and
create a nexthop version that is called if the fib_info uses an
external nexthop.
Fixes: 430a049190 ("nexthop: Add support for nexthop groups")
Signed-off-by: David Ahern <dsahern@gmail.com>
Acked-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 90f33bffa382598a32cc82abfeb20adc92d041b6 upstream.
We must avoid modifying published nexthop groups while they might be
in use, otherwise we might see NULL ptr dereferences. In order to do
that we allocate 2 nexthoup group structures upon nexthop creation
and swap between them when we have to delete an entry. The reason is
that we can't fail nexthop group removal, so we can't handle allocation
failure thus we move the extra allocation on creation where we can
safely fail and return ENOMEM.
Fixes: 430a049190 ("nexthop: Add support for nexthop groups")
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit ac21753a5c2c9a6a2019997481a2ac12bbde48c8 upstream.
Move nh_grp dereference and check for removing nexthop group due to
all members gone into remove_nh_grp_entry.
Fixes: 430a049190 ("nexthop: Add support for nexthop groups")
Signed-off-by: David Ahern <dsahern@gmail.com>
Acked-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 4c559f15efcc43b996f4da528cd7f9483aaca36d upstream.
Dan Carpenter says: "Smatch complains that the value for "cmd" comes
from the network and can't be trusted."
Add pptp_msg_name() helper function that checks for the array boundary.
Fixes: f09943fefe ("[NETFILTER]: nf_conntrack/nf_nat: add PPTP helper port")
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 976eba8ab596bab94b9714cd46d38d5c6a2c660d upstream.
In Commit dd9ee34440 ("vti4: Fix a ipip packet processing bug in
'IPCOMP' virtual tunnel"), it tries to receive IPIP packets in vti
by calling xfrm_input(). This case happens when a small packet or
frag sent by peer is too small to get compressed.
However, xfrm_input() will still get to the IPCOMP path where skb
sec_path is set, but never dropped while it should have been done
in vti_ipcomp4_protocol.cb_handler(vti_rcv_cb), as it's not an
ipcomp4 packet. This will cause that the packet can never pass
xfrm4_policy_check() in the upper protocol rcv functions.
So this patch is to call ip_tunnel_rcv() to process IPIP packets
instead.
Fixes: dd9ee34440 ("vti4: Fix a ipip packet processing bug in 'IPCOMP' virtual tunnel")
Reported-by: Xiumei Mu <xmu@redhat.com>
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit db87668ad1e4917cfe04e217307ba6ed9390716e upstream.
This xfrm_state_put call in esp4/6_gro_receive() will cause
double put for state, as in out_reset path secpath_reset()
will put all states set in skb sec_path.
So fix it by simply remove the xfrm_state_put call.
Fixes: 6ed69184ed ("xfrm: Reset secpath in xfrm failure")
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 84be69b869a5a496a6cfde9b3c29509207a1f1fa ]
For nexthop groups, attributes after NHA_GROUP_TYPE are invalid, but
nh_check_attr_group starts checking at NHA_GROUP. The group type defaults
to multipath and the NHA_GROUP_TYPE is currently optional so this has
slipped through so far. Fix the attribute checking to handle support of
new group types.
Fixes: 430a049190 ("nexthop: Add support for nexthop groups")
Signed-off-by: ASSOGBA Emery <assogba.emery@gmail.com>
Signed-off-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit a6211caa634da39d861a47437ffcda8b38ef421b ]
Commit adb03115f4 ("net: get rid of an signed integer overflow in ip_idents_reserve()")
used atomic_cmpxchg to replace "atomic_add_return" inside the function
"ip_idents_reserve". The reason was to avoid UBSAN warning.
However, this change has caused performance degrade and in GCC-8,
fno-strict-overflow is now mapped to -fwrapv -fwrapv-pointer
and signed integer overflow is now undefined by default at all
optimization levels[1]. Moreover, it was a bug in UBSAN vs -fwrapv
/-fno-strict-overflow, so Let's revert it safely.
[1] https://gcc.gnu.org/gcc-8/changes.html
Suggested-by: Peter Zijlstra <peterz@infradead.org>
Suggested-by: Eric Dumazet <edumazet@google.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Alexey Kuznetsov <kuznet@ms2.inr.ac.ru>
Cc: Hideaki YOSHIFUJI <yoshfuji@linux-ipv6.org>
Cc: Jakub Kicinski <kuba@kernel.org>
Cc: Jiri Pirko <jiri@resnulli.us>
Cc: Arvind Sankar <nivedita@alum.mit.edu>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Eric Dumazet <edumazet@google.com>
Cc: Jiong Wang <jiongwang@huawei.com>
Signed-off-by: Yuqi Jin <jinyuqi@huawei.com>
Signed-off-by: Shaokun Zhang <zhangshaokun@hisilicon.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit d69100b8eee27c2d60ee52df76e0b80a8d492d34 ]
Fixes data remnant seen when we fail to reserve space for a
nexthop group during a larger dump.
If we fail the reservation, we goto nla_put_failure and
cancel the message.
Reproduce with the following iproute2 commands:
=====================
ip link add dummy1 type dummy
ip link add dummy2 type dummy
ip link add dummy3 type dummy
ip link add dummy4 type dummy
ip link add dummy5 type dummy
ip link add dummy6 type dummy
ip link add dummy7 type dummy
ip link add dummy8 type dummy
ip link add dummy9 type dummy
ip link add dummy10 type dummy
ip link add dummy11 type dummy
ip link add dummy12 type dummy
ip link add dummy13 type dummy
ip link add dummy14 type dummy
ip link add dummy15 type dummy
ip link add dummy16 type dummy
ip link add dummy17 type dummy
ip link add dummy18 type dummy
ip link add dummy19 type dummy
ip link add dummy20 type dummy
ip link add dummy21 type dummy
ip link add dummy22 type dummy
ip link add dummy23 type dummy
ip link add dummy24 type dummy
ip link add dummy25 type dummy
ip link add dummy26 type dummy
ip link add dummy27 type dummy
ip link add dummy28 type dummy
ip link add dummy29 type dummy
ip link add dummy30 type dummy
ip link add dummy31 type dummy
ip link add dummy32 type dummy
ip link set dummy1 up
ip link set dummy2 up
ip link set dummy3 up
ip link set dummy4 up
ip link set dummy5 up
ip link set dummy6 up
ip link set dummy7 up
ip link set dummy8 up
ip link set dummy9 up
ip link set dummy10 up
ip link set dummy11 up
ip link set dummy12 up
ip link set dummy13 up
ip link set dummy14 up
ip link set dummy15 up
ip link set dummy16 up
ip link set dummy17 up
ip link set dummy18 up
ip link set dummy19 up
ip link set dummy20 up
ip link set dummy21 up
ip link set dummy22 up
ip link set dummy23 up
ip link set dummy24 up
ip link set dummy25 up
ip link set dummy26 up
ip link set dummy27 up
ip link set dummy28 up
ip link set dummy29 up
ip link set dummy30 up
ip link set dummy31 up
ip link set dummy32 up
ip link set dummy33 up
ip link set dummy34 up
ip link set vrf-red up
ip link set vrf-blue up
ip link set dummyVRFred up
ip link set dummyVRFblue up
ip ro add 1.1.1.1/32 dev dummy1
ip ro add 1.1.1.2/32 dev dummy2
ip ro add 1.1.1.3/32 dev dummy3
ip ro add 1.1.1.4/32 dev dummy4
ip ro add 1.1.1.5/32 dev dummy5
ip ro add 1.1.1.6/32 dev dummy6
ip ro add 1.1.1.7/32 dev dummy7
ip ro add 1.1.1.8/32 dev dummy8
ip ro add 1.1.1.9/32 dev dummy9
ip ro add 1.1.1.10/32 dev dummy10
ip ro add 1.1.1.11/32 dev dummy11
ip ro add 1.1.1.12/32 dev dummy12
ip ro add 1.1.1.13/32 dev dummy13
ip ro add 1.1.1.14/32 dev dummy14
ip ro add 1.1.1.15/32 dev dummy15
ip ro add 1.1.1.16/32 dev dummy16
ip ro add 1.1.1.17/32 dev dummy17
ip ro add 1.1.1.18/32 dev dummy18
ip ro add 1.1.1.19/32 dev dummy19
ip ro add 1.1.1.20/32 dev dummy20
ip ro add 1.1.1.21/32 dev dummy21
ip ro add 1.1.1.22/32 dev dummy22
ip ro add 1.1.1.23/32 dev dummy23
ip ro add 1.1.1.24/32 dev dummy24
ip ro add 1.1.1.25/32 dev dummy25
ip ro add 1.1.1.26/32 dev dummy26
ip ro add 1.1.1.27/32 dev dummy27
ip ro add 1.1.1.28/32 dev dummy28
ip ro add 1.1.1.29/32 dev dummy29
ip ro add 1.1.1.30/32 dev dummy30
ip ro add 1.1.1.31/32 dev dummy31
ip ro add 1.1.1.32/32 dev dummy32
ip next add id 1 via 1.1.1.1 dev dummy1
ip next add id 2 via 1.1.1.2 dev dummy2
ip next add id 3 via 1.1.1.3 dev dummy3
ip next add id 4 via 1.1.1.4 dev dummy4
ip next add id 5 via 1.1.1.5 dev dummy5
ip next add id 6 via 1.1.1.6 dev dummy6
ip next add id 7 via 1.1.1.7 dev dummy7
ip next add id 8 via 1.1.1.8 dev dummy8
ip next add id 9 via 1.1.1.9 dev dummy9
ip next add id 10 via 1.1.1.10 dev dummy10
ip next add id 11 via 1.1.1.11 dev dummy11
ip next add id 12 via 1.1.1.12 dev dummy12
ip next add id 13 via 1.1.1.13 dev dummy13
ip next add id 14 via 1.1.1.14 dev dummy14
ip next add id 15 via 1.1.1.15 dev dummy15
ip next add id 16 via 1.1.1.16 dev dummy16
ip next add id 17 via 1.1.1.17 dev dummy17
ip next add id 18 via 1.1.1.18 dev dummy18
ip next add id 19 via 1.1.1.19 dev dummy19
ip next add id 20 via 1.1.1.20 dev dummy20
ip next add id 21 via 1.1.1.21 dev dummy21
ip next add id 22 via 1.1.1.22 dev dummy22
ip next add id 23 via 1.1.1.23 dev dummy23
ip next add id 24 via 1.1.1.24 dev dummy24
ip next add id 25 via 1.1.1.25 dev dummy25
ip next add id 26 via 1.1.1.26 dev dummy26
ip next add id 27 via 1.1.1.27 dev dummy27
ip next add id 28 via 1.1.1.28 dev dummy28
ip next add id 29 via 1.1.1.29 dev dummy29
ip next add id 30 via 1.1.1.30 dev dummy30
ip next add id 31 via 1.1.1.31 dev dummy31
ip next add id 32 via 1.1.1.32 dev dummy32
i=100
while [ $i -le 200 ]
do
ip next add id $i group 1/2/3/4/5/6/7/8/9/10/11/12/13/14/15/16/17/18/19
echo $i
((i++))
done
ip next add id 999 group 1/2/3/4/5/6
ip next ls
========================
Fixes: ab84be7e54 ("net: Initial nexthop code")
Signed-off-by: Stephen Worley <sworley@cumulusnetworks.com>
Reviewed-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 57ebc8f08504f176eb0f25b3e0fde517dec61a4f ]
In case of error with MPLS support the code is misusing AF_INET
instead of AF_MPLS.
Fixes: 1b69e7e6c4 ("ipip: support MPLS over IPv4")
Signed-off-by: Vadim Fedorenko <vfedorenko@novek.ru>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 88d7fcfa3b1fe670f0412b95be785aafca63352b ]
The commit 637bc8bbe6 ("inet: reset tb->fastreuseport when adding a reuseport sk")
added a bind-address cache in tb->fast*. The tb->fast* caches the address
of a sk which has successfully been binded with SO_REUSEPORT ON. The idea
is to avoid the expensive conflict search in inet_csk_bind_conflict().
There is an issue with wildcard matching where sk_reuseport_match() should
have returned false but it is currently returning true. It ends up
hiding bind conflict. For example,
bind("[::1]:443"); /* without SO_REUSEPORT. Succeed. */
bind("[::2]:443"); /* with SO_REUSEPORT. Succeed. */
bind("[::]:443"); /* with SO_REUSEPORT. Still Succeed where it shouldn't */
The last bind("[::]:443") with SO_REUSEPORT on should have failed because
it should have a conflict with the very first bind("[::1]:443") which
has SO_REUSEPORT off. However, the address "[::2]" is cached in
tb->fast* in the second bind. In the last bind, the sk_reuseport_match()
returns true because the binding sk's wildcard addr "[::]" matches with
the "[::2]" cached in tb->fast*.
The correct bind conflict is reported by removing the second
bind such that tb->fast* cache is not involved and forces the
bind("[::]:443") to go through the inet_csk_bind_conflict():
bind("[::1]:443"); /* without SO_REUSEPORT. Succeed. */
bind("[::]:443"); /* with SO_REUSEPORT. -EADDRINUSE */
The expected behavior for sk_reuseport_match() is, it should only allow
the "cached" tb->fast* address to be used as a wildcard match but not
the address of the binding sk. To do that, the current
"bool match_wildcard" arg is split into
"bool match_sk1_wildcard" and "bool match_sk2_wildcard".
This change only affects the sk_reuseport_match() which is only
used by inet_csk (e.g. TCP).
The other use cases are calling inet_rcv_saddr_equal() and
this patch makes it pass the same "match_wildcard" arg twice to
the "ipv[46]_rcv_saddr_equal(..., match_wildcard, match_wildcard)".
Cc: Josef Bacik <jbacik@fb.com>
Fixes: 637bc8bbe6 ("inet: reset tb->fastreuseport when adding a reuseport sk")
Signed-off-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 41b4bd986f86331efc599b9a3f5fb86ad92e9af9 ]
In case we can't find a ->dumpit callback for the requested
(family,type) pair, we fall back to (PF_UNSPEC,type). In effect, we're
in the same situation as if userspace had requested a PF_UNSPEC
dump. For RTM_GETROUTE, that handler is rtnl_dump_all, which calls all
the registered RTM_GETROUTE handlers.
The requested table id may or may not exist for all of those
families. commit ae677bbb44 ("net: Don't return invalid table id
error when dumping all families") fixed the problem when userspace
explicitly requests a PF_UNSPEC dump, but missed the fallback case.
For example, when we pass ipv6.disable=1 to a kernel with
CONFIG_IP_MROUTE=y and CONFIG_IP_MROUTE_MULTIPLE_TABLES=y,
the (PF_INET6, RTM_GETROUTE) handler isn't registered, so we end up in
rtnl_dump_all, and listing IPv6 routes will unexpectedly print:
# ip -6 r
Error: ipv4: MR table does not exist.
Dump terminated
commit ae677bbb44 introduced the dump_all_families variable, which
gets set when userspace requests a PF_UNSPEC dump. However, we can't
simply set the family to PF_UNSPEC in rtnetlink_rcv_msg in the
fallback case to get dump_all_families == true, because some messages
types (for example RTM_GETRULE and RTM_GETNEIGH) only register the
PF_UNSPEC handler and use the family to filter in the kernel what is
dumped to userspace. We would then export more entries, that userspace
would have to filter. iproute does that, but other programs may not.
Instead, this patch removes dump_all_families and updates the
RTM_GETROUTE handlers to check if the family that is being dumped is
their own. When it's not, which covers both the intentional PF_UNSPEC
dumps (as dump_all_families did) and the fallback case, ignore the
missing table id error.
Fixes: cb167893f4 ("net: Plumb support for filtering ipv4 and ipv6 multicast route dumps")
Signed-off-by: Sabrina Dubroca <sd@queasysnail.net>
Reviewed-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 18f02ad19e2c2a1d9e1d55a4e1c0cbf51419151c upstream.
tcp_bpf_recvmsg() invokes sk_psock_get(), which returns a reference of
the specified sk_psock object to "psock" with increased refcnt.
When tcp_bpf_recvmsg() returns, local variable "psock" becomes invalid,
so the refcount should be decreased to keep refcount balanced.
The reference counting issue happens in several exception handling paths
of tcp_bpf_recvmsg(). When those error scenarios occur such as "flags"
includes MSG_ERRQUEUE, the function forgets to decrease the refcnt
increased by sk_psock_get(), causing a refcnt leak.
Fix this issue by calling sk_psock_put() or pulling up the error queue
read handling when those error scenarios occur.
Fixes: e7a5f1f1cd000 ("bpf/sockmap: Read psock ingress_msg before sk_receive_queue")
Signed-off-by: Xiyu Yang <xiyuyang19@fudan.edu.cn>
Signed-off-by: Xin Tan <tanxin.ctf@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Jakub Sitnicki <jakub@cloudflare.com>
Link: https://lore.kernel.org/bpf/1587872115-42805-1-git-send-email-xiyuyang19@fudan.edu.cn
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 81aabbb9fb7b4b1efd073b62f0505d3adad442f3 ]
In bpf_tcp_ingress we used apply_bytes to subtract bytes from sg.size
which is used to track total bytes in a message. But this is not
correct because apply_bytes is itself modified in the main loop doing
the mem_charge.
Then at the end of this we have sg.size incorrectly set and out of
sync with actual sk values. Then we can get a splat if we try to
cork the data later and again try to redirect the msg to ingress. To
fix instead of trying to track msg.size do the easy thing and include
it as part of the sk_msg_xfer logic so that when the msg is moved the
sg.size is always correct.
To reproduce the below users will need ingress + cork and hit an
error path that will then try to 'free' the skmsg.
[ 173.699981] BUG: KASAN: null-ptr-deref in sk_msg_free_elem+0xdd/0x120
[ 173.699987] Read of size 8 at addr 0000000000000008 by task test_sockmap/5317
[ 173.700000] CPU: 2 PID: 5317 Comm: test_sockmap Tainted: G I 5.7.0-rc1+ #43
[ 173.700005] Hardware name: Dell Inc. Precision 5820 Tower/002KVM, BIOS 1.9.2 01/24/2019
[ 173.700009] Call Trace:
[ 173.700021] dump_stack+0x8e/0xcb
[ 173.700029] ? sk_msg_free_elem+0xdd/0x120
[ 173.700034] ? sk_msg_free_elem+0xdd/0x120
[ 173.700042] __kasan_report+0x102/0x15f
[ 173.700052] ? sk_msg_free_elem+0xdd/0x120
[ 173.700060] kasan_report+0x32/0x50
[ 173.700070] sk_msg_free_elem+0xdd/0x120
[ 173.700080] __sk_msg_free+0x87/0x150
[ 173.700094] tcp_bpf_send_verdict+0x179/0x4f0
[ 173.700109] tcp_bpf_sendpage+0x3ce/0x5d0
Fixes: 604326b41a ("bpf, sockmap: convert to generic sk_msg interface")
Signed-off-by: John Fastabend <john.fastabend@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Jakub Sitnicki <jakub@cloudflare.com>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Link: https://lore.kernel.org/bpf/158861290407.14306.5327773422227552482.stgit@john-Precision-5820-Tower
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit cc4de047b33be247f9c8150d3e496743a49642b8 ]
The stated intent of the original commit is to is to "return the timestamp
corresponding to the highest sequence number data returned." The current
implementation returns the timestamp for the last byte of the last fully
read skb, which is not necessarily the last byte in the recv buffer. This
patch converts behavior to the original definition, and to the behavior of
the previous draft versions of commit 98aaa913b4 ("tcp: Extend
SOF_TIMESTAMPING_RX_SOFTWARE to TCP recvmsg") which also match this
behavior.
Fixes: 98aaa913b4 ("tcp: Extend SOF_TIMESTAMPING_RX_SOFTWARE to TCP recvmsg")
Co-developed-by: Iris Liu <iris@onechronos.com>
Signed-off-by: Iris Liu <iris@onechronos.com>
Signed-off-by: Kelly Littlepage <kelly@onechronos.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Acked-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 57644431a6c2faac5d754ebd35780cf43a531b1a ]
In commit b406472b5a ("net: ipv4: avoid mixed n_redirects and
rate_tokens usage") I missed the fact that a 0 'rate_tokens' will
bypass the backoff algorithm.
Since rate_tokens is cleared after a redirect silence, and never
incremented on redirects, if the host keeps receiving packets
requiring redirect it will reply ignoring the backoff.
Additionally, the 'rate_last' field will be updated with the
cadence of the ingress packet requiring redirect. If that rate is
high enough, that will prevent the host from generating any
other kind of ICMP messages
The check for a zero 'rate_tokens' value was likely a shortcut
to avoid the more complex backoff algorithm after a redirect
silence period. Address the issue checking for 'n_redirects'
instead, which is incremented on successful redirect, and
does not interfere with other ICMP replies.
Fixes: b406472b5a ("net: ipv4: avoid mixed n_redirects and rate_tokens usage")
Reported-and-tested-by: Colin Walters <walters@redhat.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 24adbc1676af4e134e709ddc7f34cf2adc2131e4 ]
We autotune rcvbuf whenever SO_RCVLOWAT is set to account for 100%
overhead in tcp_set_rcvlowat()
This works well when skb->len/skb->truesize ratio is bigger than 0.5
But if we receive packets with small MSS, we can end up in a situation
where not enough bytes are available in the receive queue to satisfy
RCVLOWAT setting.
As our sk_rcvbuf limit is hit, we send zero windows in ACK packets,
preventing remote peer from sending more data.
Even autotuning does not help, because it only triggers at the time
user process drains the queue. If no EPOLLIN is generated, this
can not happen.
Note poll() has a similar issue, after commit
c7004482e8 ("tcp: Respect SO_RCVLOWAT in tcp_poll().")
Fixes: 03f45c883c ("tcp: avoid extra wakeups for SO_RCVLOWAT users")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit eead1c2ea2509fd754c6da893a94f0e69e83ebe4 ]
The cipso and calipso code can set the MLS_CAT attribute on
successful parsing, even if the corresponding catmap has
not been allocated, as per current configuration and external
input.
Later, selinux code tries to access the catmap if the MLS_CAT flag
is present via netlbl_catmap_getlong(). That may cause null ptr
dereference while processing incoming network traffic.
Address the issue setting the MLS_CAT flag only if the catmap is
really allocated. Additionally let netlbl_catmap_getlong() cope
with NULL catmap.
Reported-by: Matthew Sheets <matthew.sheets@gd-ms.com>
Fixes: 4b8feff251 ("netlabel: fix the horribly broken catmap functions")
Fixes: ceba1832b1 ("calipso: Set the calipso socket label to match the secattr.")
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Acked-by: Paul Moore <paul@paul-moore.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 0c922a4850eba2e668f73a3f1153196e09abb251 ]
IPSKB_XFRM_TRANSFORMED and IP6SKB_XFRM_TRANSFORMED are skb flags set by
xfrm code to tell other skb handlers that the packet has been passed
through the xfrm output functions. Simplify the code and just always
set them rather than conditionally based on netfilter enabled thus
making the flag available for other users.
Signed-off-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 25629fdaff2ff509dd0b3f5ff93d70a75e79e0a1 upstream.
when creating a new ipip interface with no local/remote configuration,
the lookup is done with TUNNEL_NO_KEY flag, making it impossible to
match the new interface (only possible match being fallback or metada
case interface); e.g: `ip link add tunl1 type ipip dev eth0`
To fix this case, adding a flag check before the key comparison so we
permit to match an interface with no local/remote config; it also avoids
breaking possible userland tools relying on TUNNEL_NO_KEY flag and
uninitialised key.
context being on my side, I'm creating an extra ipip interface attached
to the physical one, and moving it to a dedicated namespace.
Fixes: c544193214 ("GRE: Refactor GRE tunneling code.")
Signed-off-by: William Dauchy <w.dauchy@criteo.com>
Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 690cc86321eb9bcee371710252742fb16fe96824 ]
When CONFIG_IP_MULTICAST is not set and multicast ip is added to the device
with autojoin flag or when multicast ip is deleted kernel will crash.
steps to reproduce:
ip addr add 224.0.0.0/32 dev eth0
ip addr del 224.0.0.0/32 dev eth0
or
ip addr add 224.0.0.0/32 dev eth0 autojoin
Unable to handle kernel NULL pointer dereference at virtual address 0000000000000088
pc : _raw_write_lock_irqsave+0x1e0/0x2ac
lr : lock_sock_nested+0x1c/0x60
Call trace:
_raw_write_lock_irqsave+0x1e0/0x2ac
lock_sock_nested+0x1c/0x60
ip_mc_config.isra.28+0x50/0xe0
inet_rtm_deladdr+0x1a8/0x1f0
rtnetlink_rcv_msg+0x120/0x350
netlink_rcv_skb+0x58/0x120
rtnetlink_rcv+0x14/0x20
netlink_unicast+0x1b8/0x270
netlink_sendmsg+0x1a0/0x3b0
____sys_sendmsg+0x248/0x290
___sys_sendmsg+0x80/0xc0
__sys_sendmsg+0x68/0xc0
__arm64_sys_sendmsg+0x20/0x30
el0_svc_common.constprop.2+0x88/0x150
do_el0_svc+0x20/0x80
el0_sync_handler+0x118/0x190
el0_sync+0x140/0x180
Fixes: 93a714d6b5 ("multicast: Extend ip address command to enable multicast group join/leave on")
Signed-off-by: Taras Chornyi <taras.chornyi@plvision.eu>
Signed-off-by: Vadym Kochan <vadym.kochan@plvision.eu>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit dad8cea7add96a353fa1898b5ccefbb72da66f29 upstream.
In a rare corner case the new logic for undo of SYNACK RTO could
result in triggering the warning in tcp_fastretrans_alert() that says:
WARN_ON(tp->retrans_out != 0);
The warning looked like:
WARNING: CPU: 1 PID: 1 at net/ipv4/tcp_input.c:2818 tcp_ack+0x13e0/0x3270
The sequence that tickles this bug is:
- Fast Open server receives TFO SYN with data, sends SYNACK
- (client receives SYNACK and sends ACK, but ACK is lost)
- server app sends some data packets
- (N of the first data packets are lost)
- server receives client ACK that has a TS ECR matching first SYNACK,
and also SACKs suggesting the first N data packets were lost
- server performs TS undo of SYNACK RTO, then immediately
enters recovery
- buggy behavior then performed a *second* undo that caused
the connection to be in CA_Open with retrans_out != 0
Basically, the incoming ACK packet with SACK blocks causes us to first
undo the cwnd reduction from the SYNACK RTO, but then immediately
enters fast recovery, which then makes us eligible for undo again. And
then tcp_rcv_synrecv_state_fastopen() accidentally performs an undo
using a "mash-up" of state from two different loss recovery phases: it
uses the timestamp info from the ACK of the original SYNACK, and the
undo_marker from the fast recovery.
This fix refines the logic to only invoke the tcp_try_undo_loss()
inside tcp_rcv_synrecv_state_fastopen() if the connection is still in
CA_Loss. If peer SACKs triggered fast recovery, then
tcp_rcv_synrecv_state_fastopen() can't safely undo.
Fixes: 794200d662 ("tcp: undo cwnd on Fast Open spurious SYNACK retransmit")
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 32ca98feab8c9076c89c0697c5a85e46fece809d ]
The fix referenced below causes a crash when an ERSPAN tunnel is created
without passing IFLA_INFO_DATA. Fix by validating passed-in data in the
same way as ipgre does.
Fixes: e1f8f78ffe98 ("net: ip_gre: Separate ERSPAN newlink / changelink callbacks")
Reported-by: syzbot+1b4ebf4dae4e510dd219@syzkaller.appspotmail.com
Signed-off-by: Petr Machata <petrm@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit e1f8f78ffe9854308b9e12a73ebe4e909074fc33 ]
ERSPAN shares most of the code path with GRE and gretap code. While that
helps keep the code compact, it is also error prone. Currently a broken
userspace can turn a gretap tunnel into a de facto ERSPAN one by passing
IFLA_GRE_ERSPAN_VER. There has been a similar issue in ip6gretap in the
past.
To prevent these problems in future, split the newlink and changelink code
paths. Split the ERSPAN code out of ipgre_netlink_parms() into a new
function erspan_netlink_parms(). Extract a piece of common logic from
ipgre_newlink() and ipgre_changelink() into ipgre_newlink_encap_setup().
Add erspan_newlink() and erspan_changelink().
Fixes: 84e54fe0a5 ("gre: introduce native tunnel support for ERSPAN")
Signed-off-by: Petr Machata <petrm@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 6cd6cbf593bfa3ae6fc3ed34ac21da4d35045425 ]
When application uses TCP_QUEUE_SEQ socket option to
change tp->rcv_next, we must also update tp->copied_seq.
Otherwise, stuff relying on tcp_inq() being precise can
eventually be confused.
For example, tcp_zerocopy_receive() might crash because
it does not expect tcp_recv_skb() to return NULL.
We could add tests in various places to fix the issue,
or simply make sure tcp_inq() wont return a random value,
and leave fast path as it is.
Note that this fixes ioctl(fd, SIOCINQ, &val) at the same
time.
Fixes: ee9952831c ("tcp: Initial repair mode")
Fixes: 05255b823a ("tcp: add TCP_ZEROCOPY_RECEIVE support for zerocopy receive")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: syzbot <syzkaller@googlegroups.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit b738a185beaab8728943acdb3e67371b8a88185e ]
skb->rbnode is sharing three skb fields : next, prev, dev
When a packet is sent, TCP keeps the original skb (master)
in a rtx queue, which was converted to rbtree a while back.
__tcp_transmit_skb() is responsible to clone the master skb,
and add the TCP header to the clone before sending it
to network layer.
skb_clone() already clears skb->next and skb->prev, but copies
the master oskb->dev into the clone.
We need to clear skb->dev, otherwise lower layers could interpret
the value as a pointer to a netdev.
This old bug surfaced recently when commit 28f8bfd1ac94
("netfilter: Support iif matches in POSTROUTING") was merged.
Before this netfilter commit, skb->dev value was ignored and
changed before reaching dev_queue_xmit()
Fixes: 75c119afe1 ("tcp: implement rb-tree based retransmit queue")
Fixes: 28f8bfd1ac94 ("netfilter: Support iif matches in POSTROUTING")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Martin Zaharinov <micron10@gmail.com>
Cc: Florian Westphal <fw@strlen.de>
Cc: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
