mptcp_poll always return POLLOUT for unblocking
connect(), ensure that the socket is a suitable
state.
The MPTCP_DATA_READY bit is never cleared on accept:
ensure we don't leave mptcp_accept() with an empty
accept queue and such bit set.
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Davide Caratti <dcaratti@redhat.com>
Reviewed-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently __mptcp_tcp_fallback() always return NULL
on incoming connections, because MPTCP does not create
the additional socket for the first subflow.
Since the previous commit no __mptcp_tcp_fallback()
caller needs a struct socket, so let __mptcp_tcp_fallback()
return the first subflow sock and cope correctly even with
incoming connections.
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Davide Caratti <dcaratti@redhat.com>
Reviewed-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This cleans the code a bit and makes the behavior more consistent.
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Davide Caratti <dcaratti@redhat.com>
Reviewed-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This cleanup the code a bit and avoid corrupted states
on weird syscall sequence (accept(), connect()).
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Davide Caratti <dcaratti@redhat.com>
Reviewed-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Keep using MPTCP sockets and a use "dummy mapping" in case of fallback
to regular TCP. When fallback is triggered, skip addition of the MPTCP
option on send.
Closes: https://github.com/multipath-tcp/mptcp_net-next/issues/11
Closes: https://github.com/multipath-tcp/mptcp_net-next/issues/22
Co-developed-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Davide Caratti <dcaratti@redhat.com>
Reviewed-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
1) Improve hardware layouts and structure for kTLS support
2) Generalize ICOSQ (Internal Channel Operations Send Queue)
Due to the asynchronous nature of adding new kTLS flows and handling
HW asynchronous kTLS resync requests, the XSK ICOSQ was extended to
support generic async operations, such as kTLS add flow and resync, in
addition to the existing XSK usages.
3) kTLS hardware flow steering and classification:
The driver already has the means to classify TCP ipv4/6 flows to send them
to the corresponding RSS HW engine, as reflected in patches 3 through 5,
the series will add a steering layer that will hook to the driver's TCP
classifiers and will match on well known kTLS connection, in case of a
match traffic will be redirected to the kTLS decryption engine, otherwise
traffic will continue flowing normally to the TCP RSS engine.
3) kTLS add flow RX HW offload support
New offload contexts post their static/progress params WQEs
(Work Queue Element) to communicate the newly added kTLS contexts
over the per-channel async ICOSQ.
The Channel/RQ is selected according to the socket's rxq index.
A new TLS-RX workqueue is used to allow asynchronous addition of
steering rules, out of the NAPI context.
It will be also used in a downstream patch in the resync procedure.
Feature is OFF by default. Can be turned on by:
$ ethtool -K <if> tls-hw-rx-offload on
4) Added mlx5 kTLS sw stats and new counters are documented in
Documentation/networking/tls-offload.rst
rx_tls_ctx - number of TLS RX HW offload contexts added to device for
decryption.
rx_tls_ooo - number of RX packets which were part of a TLS stream
but did not arrive in the expected order and triggered the resync
procedure.
rx_tls_del - number of TLS RX HW offload contexts deleted from device
(connection has finished).
rx_tls_err - number of RX packets which were part of a TLS stream
but were not decrypted due to unexpected error in the state machine.
5) Asynchronous RX resync
a. The NIC driver indicates that it would like to resync on some TLS
record within the received packet (P), but the driver does not
know (yet) which of the TLS records within the packet.
At this stage, the NIC driver will query the device to find the exact
TCP sequence for resync (tcpsn), however, the driver does not wait
for the device to provide the response.
b. Eventually, the device responds, and the driver provides the tcpsn
within the resync packet to KTLS. Now, KTLS can check the tcpsn against
any processed TLS records within packet P, and also against any record
that is processed in the future within packet P.
The asynchronous resync path simplifies the device driver, as it can
save bits on the packet completion (32-bit TCP sequence), and pass this
information on an asynchronous command instead.
Performance:
CPU: Intel(R) Xeon(R) CPU E5-2687W v4 @ 3.00GHz, 24 cores, HT off
NIC: ConnectX-6 Dx 100GbE dual port
Goodput (app-layer throughput) comparison:
+---------------+-------+-------+---------+
| # connections | 1 | 4 | 8 |
+---------------+-------+-------+---------+
| SW (Gbps) | 7.26 | 24.70 | 50.30 |
+---------------+-------+-------+---------+
| HW (Gbps) | 18.50 | 64.30 | 92.90 |
+---------------+-------+-------+---------+
| Speedup | 2.55x | 2.56x | 1.85x * |
+---------------+-------+-------+---------+
* After linerate is reached, diff is observed in CPU util
-----BEGIN PGP SIGNATURE-----
iQEzBAABCAAdFiEEGhZs6bAKwk/OTgTpSD+KveBX+j4FAl73s2kACgkQSD+KveBX
+j4wqAf/ZhcEn7i4N2F9wMMIL6wd4DgwKWWhbGpiREIxDwcRbqH7PGom8nBZMNd9
+3g3zfURvByWehLtYcjmMgR4B7+xDgEs0dSx6pQM9764HqLDV2jW8ENr9Vr/u8s1
hJ/eV8uzIfvx27MzbENZi0oJTw7N9nCgdcv1OyZkIba+Iado9pOeakPgBmTbINgo
46LJI9nIEROE15gfjyxrVeYAs3Nxt+bogQCWYfMqUfRmKcMJ0d4oTHaUdtmm+xQB
jC685/e4gE7jRgZ3qH/xvCZYp7+TVKaXsB0EtaJdPFEkvvvQpgPTfquIQ+6l7vvE
Yf1YUhnDOoxGUQy1CdSZ2reNxLIm8A==
=7+rG
-----END PGP SIGNATURE-----
Merge tag 'mlx5-tls-2020-06-26' of git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux
Saeed Mahameed says:
====================
mlx5-tls-2020-06-26
1) Improve hardware layouts and structure for kTLS support
2) Generalize ICOSQ (Internal Channel Operations Send Queue)
Due to the asynchronous nature of adding new kTLS flows and handling
HW asynchronous kTLS resync requests, the XSK ICOSQ was extended to
support generic async operations, such as kTLS add flow and resync, in
addition to the existing XSK usages.
3) kTLS hardware flow steering and classification:
The driver already has the means to classify TCP ipv4/6 flows to send them
to the corresponding RSS HW engine, as reflected in patches 3 through 5,
the series will add a steering layer that will hook to the driver's TCP
classifiers and will match on well known kTLS connection, in case of a
match traffic will be redirected to the kTLS decryption engine, otherwise
traffic will continue flowing normally to the TCP RSS engine.
3) kTLS add flow RX HW offload support
New offload contexts post their static/progress params WQEs
(Work Queue Element) to communicate the newly added kTLS contexts
over the per-channel async ICOSQ.
The Channel/RQ is selected according to the socket's rxq index.
A new TLS-RX workqueue is used to allow asynchronous addition of
steering rules, out of the NAPI context.
It will be also used in a downstream patch in the resync procedure.
Feature is OFF by default. Can be turned on by:
$ ethtool -K <if> tls-hw-rx-offload on
4) Added mlx5 kTLS sw stats and new counters are documented in
Documentation/networking/tls-offload.rst
rx_tls_ctx - number of TLS RX HW offload contexts added to device for
decryption.
rx_tls_ooo - number of RX packets which were part of a TLS stream
but did not arrive in the expected order and triggered the resync
procedure.
rx_tls_del - number of TLS RX HW offload contexts deleted from device
(connection has finished).
rx_tls_err - number of RX packets which were part of a TLS stream
but were not decrypted due to unexpected error in the state machine.
5) Asynchronous RX resync
a. The NIC driver indicates that it would like to resync on some TLS
record within the received packet (P), but the driver does not
know (yet) which of the TLS records within the packet.
At this stage, the NIC driver will query the device to find the exact
TCP sequence for resync (tcpsn), however, the driver does not wait
for the device to provide the response.
b. Eventually, the device responds, and the driver provides the tcpsn
within the resync packet to KTLS. Now, KTLS can check the tcpsn against
any processed TLS records within packet P, and also against any record
that is processed in the future within packet P.
The asynchronous resync path simplifies the device driver, as it can
save bits on the packet completion (32-bit TCP sequence), and pass this
information on an asynchronous command instead.
Performance:
CPU: Intel(R) Xeon(R) CPU E5-2687W v4 @ 3.00GHz, 24 cores, HT off
NIC: ConnectX-6 Dx 100GbE dual port
Goodput (app-layer throughput) comparison:
+---------------+-------+-------+---------+
| # connections | 1 | 4 | 8 |
+---------------+-------+-------+---------+
| SW (Gbps) | 7.26 | 24.70 | 50.30 |
+---------------+-------+-------+---------+
| HW (Gbps) | 18.50 | 64.30 | 92.90 |
+---------------+-------+-------+---------+
| Speedup | 2.55x | 2.56x | 1.85x * |
+---------------+-------+-------+---------+
* After linerate is reached, diff is observed in CPU util
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
In order to allow acting on dropped and/or ECN-marked packets, add two new
qevents to the RED qdisc: "early_drop" and "mark". Filters attached at
"early_drop" block are executed as packets are early-dropped, those
attached at the "mark" block are executed as packets are ECN-marked.
Two new attributes are introduced: TCA_RED_EARLY_DROP_BLOCK with the block
index for the "early_drop" qevent, and TCA_RED_MARK_BLOCK for the "mark"
qevent. Absence of these attributes signifies "don't care": no block is
allocated in that case, or the existing blocks are left intact in case of
the change callback.
For purposes of offloading, blocks attached to these qevents appear with
newly-introduced binder types, FLOW_BLOCK_BINDER_TYPE_RED_EARLY_DROP and
FLOW_BLOCK_BINDER_TYPE_RED_MARK.
Signed-off-by: Petr Machata <petrm@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In the following patches, RED will get two qevents. The implementation will
be clearer if the callback for change is not a pure subset of the callback
for init. Split the two and promote attribute parsing to the callbacks
themselves from the common code, because it will be handy there.
Signed-off-by: Petr Machata <petrm@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Qevents are attach points for TC blocks, where filters can be put that are
executed when "interesting events" take place in a qdisc. The data to keep
and the functions to invoke to maintain a qevent will be largely the same
between qevents. Therefore introduce sched-wide helpers for qevent
management.
Currently, similarly to ingress and egress blocks of clsact pseudo-qdisc,
blocks attachment cannot be changed after the qdisc is created. To that
end, add a helper tcf_qevent_validate_change(), which verifies whether
block index attribute is not attached, or if it is, whether its value
matches the current one (i.e. there is no material change).
The function tcf_qevent_handle() should be invoked when qdisc hits the
"interesting event" corresponding to a block. This function releases root
lock for the duration of executing the attached filters, to allow packets
generated through user actions (notably mirred) to be reinserted to the
same qdisc tree.
Signed-off-by: Petr Machata <petrm@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
A following patch introduces qevents, points in qdisc algorithm where
packet can be processed by user-defined filters. Should this processing
lead to a situation where a new packet is to be enqueued on the same port,
holding the root lock would lead to deadlocks. To solve the issue, qevent
handler needs to unlock and relock the root lock when necessary.
To that end, add the root lock argument to the qdisc op enqueue, and
propagate throughout.
Signed-off-by: Petr Machata <petrm@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The method ndo_start_xmit() is defined as returning an 'netdev_tx_t',
which is a typedef for an enum type, but the implementation in this
driver returns an 'int'.
Fix this by returning 'netdev_tx_t' in this driver too.
Signed-off-by: Luc Van Oostenryck <luc.vanoostenryck@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The method ndo_start_xmit() is defined as returning an 'netdev_tx_t',
which is a typedef for an enum type, but the implementation in this
driver returns an 'int'.
Fix this by returning 'netdev_tx_t' in this driver too.
Signed-off-by: Luc Van Oostenryck <luc.vanoostenryck@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently tp->delivered is updated in various places in tcp_ack() but
tp->delivered_ce is updated once at the end. As a result two counts in
OPT_STATS of SCM_TSTAMP_ACK timestamps generated in tcp_ack() may not be
in sync. This patch updates both counts at the same in tcp_ack().
Signed-off-by: Yousuk Seung <ysseung@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add sack_delivered to tcp_sacktag_state and count the number of sacked
and dsacked packets. This is pure refactor for future patches to improve
tracking delivered counts.
Signed-off-by: Yousuk Seung <ysseung@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Pass a boolean flag that tells the ECE state of the current ack to reno
sack functions. This is pure refactor for future patches to improve
tracking delivered counts.
Signed-off-by: Yousuk Seung <ysseung@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently tp->delivered is updated with sacked packets but not
cumulatively acked when SCP_TSTAMP_ACK is timestamped. This patch moves
a tcp_ack_tstamp() call in tcp_clean_rtx_queue() to later in the loop so
that when a skb is fully acked OPT_STATS of SCM_TSTAMP_ACK will include
the current skb in the delivered count. When not fully acked
tcp_ack_tstamp() is a no-op and there is no change in behavior.
Signed-off-by: Yousuk Seung <ysseung@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch adds support for asynchronous resynchronization in tls_device.
Async resync follows two distinct stages:
1. The NIC driver indicates that it would like to resync on some TLS
record within the received packet (P), but the driver does not
know (yet) which of the TLS records within the packet.
At this stage, the NIC driver will query the device to find the exact
TCP sequence for resync (tcpsn), however, the driver does not wait
for the device to provide the response.
2. Eventually, the device responds, and the driver provides the tcpsn
within the resync packet to KTLS. Now, KTLS can check the tcpsn against
any processed TLS records within packet P, and also against any record
that is processed in the future within packet P.
The asynchronous resync path simplifies the device driver, as it can
save bits on the packet completion (32-bit TCP sequence), and pass this
information on an asynchronous command instead.
Signed-off-by: Boris Pismenny <borisp@mellanox.com>
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
This reverts commit b3ae2459f8.
Revert the force resync API.
Not in use. To be replaced by a better async resync API downstream.
Signed-off-by: Boris Pismenny <borisp@mellanox.com>
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
Reviewed-by: Maxim Mikityanskiy <maximmi@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Unit tests for the internal MPTCP token APIs, using KUNIT
v1 -> v2:
- use the correct RCU annotation when initializing icsk ulp
- fix a few checkpatch issues
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
currently MPTCP uses a custom hook to executed unit tests at
boot time. Let's use the KUNIT framework instead.
Additionally move the relevant code to a separate file and
export the function needed by the test when self-tests
are build as a module.
Co-developed-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Replace the radix tree with a hash table allocated
at boot time. The radix tree has some shortcoming:
a single lock is contented by all the mptcp operation,
the lookup currently use such lock, and traversing
all the items would require a lock, too.
With hash table instead we trade a little memory to
address all the above - a per bucket lock is used.
To hash the MPTCP sockets, we re-use the msk' sk_node
entry: the MPTCP sockets are never hashed by the stack.
Replace the existing hash proto callbacks with a dummy
implementation, annotating the above constraint.
Additionally refactor the token creation to code to:
- limit the number of consecutive attempts to a fixed
maximum. Hitting a hash bucket with a long chain is
considered a failed attempt
- accept() no longer can fail to token management.
- if token creation fails at connect() time, we do
fallback to TCP (before the connection was closed)
v1 -> v2:
- fix "no newline at end of file" - Jakub
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add the missing annotation in some setup-only
functions.
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Minor overlapping changes in xfrm_device.c, between the double
ESP trailing bug fix setting the XFRM_INIT flag and the changes
in net-next preparing for bonding encryption support.
Signed-off-by: David S. Miller <davem@davemloft.net>
Pull networking fixes from David Miller:
1) Don't insert ESP trailer twice in IPSEC code, from Huy Nguyen.
2) The default crypto algorithm selection in Kconfig for IPSEC is out
of touch with modern reality, fix this up. From Eric Biggers.
3) bpftool is missing an entry for BPF_MAP_TYPE_RINGBUF, from Andrii
Nakryiko.
4) Missing init of ->frame_sz in xdp_convert_zc_to_xdp_frame(), from
Hangbin Liu.
5) Adjust packet alignment handling in ax88179_178a driver to match
what the hardware actually does. From Jeremy Kerr.
6) register_netdevice can leak in the case one of the notifiers fail,
from Yang Yingliang.
7) Use after free in ip_tunnel_lookup(), from Taehee Yoo.
8) VLAN checks in sja1105 DSA driver need adjustments, from Vladimir
Oltean.
9) tg3 driver can sleep forever when we get enough EEH errors, fix from
David Christensen.
10) Missing {READ,WRITE}_ONCE() annotations in various Intel ethernet
drivers, from Ciara Loftus.
11) Fix scanning loop break condition in of_mdiobus_register(), from
Florian Fainelli.
12) MTU limit is incorrect in ibmveth driver, from Thomas Falcon.
13) Endianness fix in mlxsw, from Ido Schimmel.
14) Use after free in smsc95xx usbnet driver, from Tuomas Tynkkynen.
15) Missing bridge mrp configuration validation, from Horatiu Vultur.
16) Fix circular netns references in wireguard, from Jason A. Donenfeld.
17) PTP initialization on recovery is not done properly in qed driver,
from Alexander Lobakin.
18) Endian conversion of L4 ports in filters of cxgb4 driver is wrong,
from Rahul Lakkireddy.
19) Don't clear bound device TX queue of socket prematurely otherwise we
get problems with ktls hw offloading, from Tariq Toukan.
20) ipset can do atomics on unaligned memory, fix from Russell King.
21) Align ethernet addresses properly in bridging code, from Thomas
Martitz.
22) Don't advertise ipv4 addresses on SCTP sockets having ipv6only set,
from Marcelo Ricardo Leitner.
* git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (149 commits)
rds: transport module should be auto loaded when transport is set
sch_cake: fix a few style nits
sch_cake: don't call diffserv parsing code when it is not needed
sch_cake: don't try to reallocate or unshare skb unconditionally
ethtool: fix error handling in linkstate_prepare_data()
wil6210: account for napi_gro_receive never returning GRO_DROP
hns: do not cast return value of napi_gro_receive to null
socionext: account for napi_gro_receive never returning GRO_DROP
wireguard: receive: account for napi_gro_receive never returning GRO_DROP
vxlan: fix last fdb index during dump of fdb with nhid
sctp: Don't advertise IPv4 addresses if ipv6only is set on the socket
tc-testing: avoid action cookies with odd length.
bpf: tcp: bpf_cubic: fix spurious HYSTART_DELAY exit upon drop in min RTT
tcp_cubic: fix spurious HYSTART_DELAY exit upon drop in min RTT
net: dsa: sja1105: fix tc-gate schedule with single element
net: dsa: sja1105: recalculate gating subschedule after deleting tc-gate rules
net: dsa: sja1105: unconditionally free old gating config
net: dsa: sja1105: move sja1105_compose_gating_subschedule at the top
net: macb: free resources on failure path of at91ether_open()
net: macb: call pm_runtime_put_sync on failure path
...
Change tin mapping on diffserv3, 4 & 8 for LE PHB support, in essence
making LE a member of the Bulk tin.
Bulk has the least priority and minimum of 1/16th total bandwidth in the
face of higher priority traffic.
NB: Diffserv 3 & 4 swap tin 0 & 1 priorities from the default order as
found in diffserv8, in case anyone is wondering why it looks a bit odd.
Signed-off-by: Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>
[ reword commit message slightly ]
Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This enhancement auto loads transport module when the transport
is set via SO_RDS_TRANSPORT socket option.
Reviewed-by: Ka-Cheong Poon <ka-cheong.poon@oracle.com>
Reviewed-by: Håkon Bugge <haakon.bugge@oracle.com>
Signed-off-by: Rao Shoaib <rao.shoaib@oracle.com>
Signed-off-by: Somasundaram Krishnasamy <somasundaram.krishnasamy@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
I spotted a few nits when comparing the in-tree version of sch_cake with
the out-of-tree one: A redundant error variable declaration shadowing an
outer declaration, and an indentation alignment issue. Fix both of these.
Fixes: 046f6fd5da ("sched: Add Common Applications Kept Enhanced (cake) qdisc")
Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
As a further optimisation of the diffserv parsing codepath, we can skip it
entirely if CAKE is configured to neither use diffserv-based
classification, nor to zero out the diffserv bits.
Fixes: c87b4ecdbe ("sch_cake: Make sure we can write the IP header before changing DSCP bits")
Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
cake_handle_diffserv() tries to linearize mac and network header parts of
skb and to make it writable unconditionally. In some cases it leads to full
skb reallocation, which reduces throughput and increases CPU load. Some
measurements of IPv4 forward + NAPT on MIPS router with 580 MHz single-core
CPU was conducted. It appears that on kernel 4.9 skb_try_make_writable()
reallocates skb, if skb was allocated in ethernet driver via so-called
'build skb' method from page cache (it was discovered by strange increase
of kmalloc-2048 slab at first).
Obtain DSCP value via read-only skb_header_pointer() call, and leave
linearization only for DSCP bleaching or ECN CE setting. And, as an
additional optimisation, skip diffserv parsing entirely if it is not needed
by the current configuration.
Fixes: c87b4ecdbe ("sch_cake: Make sure we can write the IP header before changing DSCP bits")
Signed-off-by: Ilya Ponetayev <i.ponetaev@ndmsystems.com>
[ fix a few style issues, reflow commit message ]
Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When getting SQI or maximum SQI value fails in linkstate_prepare_data(), we
must not return without calling ethnl_ops_complete(dev) as that could
result in imbalance between ethtool_ops ->begin() and ->complete() calls.
Fixes: 8066021915 ("ethtool: provide UAPI for PHY Signal Quality Index (SQI)")
Signed-off-by: Michal Kubecek <mkubecek@suse.cz>
Signed-off-by: David S. Miller <davem@davemloft.net>
If a socket is set ipv6only, it will still send IPv4 addresses in the
INIT and INIT_ACK packets. This potentially misleads the peer into using
them, which then would cause association termination.
The fix is to not add IPv4 addresses to ipv6only sockets.
Fixes: 1da177e4c3 ("Linux-2.6.12-rc2")
Reported-by: Corey Minyard <cminyard@mvista.com>
Signed-off-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Tested-by: Corey Minyard <cminyard@mvista.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
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>
Pablo Neira Ayuso says:
====================
Netfilter fixes for net
The following patchset contains Netfilter fixes for net, they are:
1) Unaligned atomic access in ipset, from Russell King.
2) Missing module description, from Rob Gill.
3) Patches to fix a module unload causing NULL pointer dereference in
xtables, from David Wilder. For the record, I posting here his cover
letter explaining the problem:
A crash happened on ppc64le when running ltp network tests triggered by
"rmmod iptable_mangle".
See previous discussion in this thread:
https://lists.openwall.net/netdev/2020/06/03/161 .
In the crash I found in iptable_mangle_hook() that
state->net->ipv4.iptable_mangle=NULL causing a NULL pointer dereference.
net->ipv4.iptable_mangle is set to NULL in +iptable_mangle_net_exit() and
called when ip_mangle modules is unloaded. A rmmod task was found running
in the crash dump. A 2nd crash showed the same problem when running
"rmmod iptable_filter" (net->ipv4.iptable_filter=NULL).
To fix this I added .pre_exit hook in all iptable_foo.c. The pre_exit will
un-register the underlying hook and exit would do the table freeing. The
netns core does an unconditional +synchronize_rcu after the pre_exit hooks
insuring no packets are in flight that have picked up the pointer before
completing the un-register.
These patches include changes for both iptables and ip6tables.
We tested this fix with ltp running iptables01.sh and iptables01.sh -6 a
loop for 72 hours.
4) Add a selftest for conntrack helper assignment, from Florian Westphal.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
The eth_addr member is passed to ether_addr functions that require
2-byte alignment, therefore the member must be properly aligned
to avoid unaligned accesses.
The problem is in place since the initial merge of multicast to unicast:
commit 6db6f0eae6 bridge: multicast to unicast
Fixes: 6db6f0eae6 ("bridge: multicast to unicast")
Cc: Roopa Prabhu <roopa@cumulusnetworks.com>
Cc: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Cc: David S. Miller <davem@davemloft.net>
Cc: Jakub Kicinski <kuba@kernel.org>
Cc: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Thomas Martitz <t.martitz@avm.de>
Acked-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
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>
Hardware device may include more than one police entry. Specifying the
action's index make it possible for several tc filters to share the same
police action when installing the filters.
Propagate this index to device drivers through the flow offload
intermediate representation, so that drivers could share a single
hardware policer between multiple filters.
v1->v2 changes:
- Update the commit message suggest by Ido Schimmel <idosch@idosch.org>
Signed-off-by: Po Liu <Po.Liu@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Current police offloading support the 'burst'' and 'rate_bytes_ps'. Some
hardware own the capability to limit the frame size. If the frame size
larger than the setting, the frame would be dropped. For the police
action itself already accept the 'mtu' parameter in tc command. But not
extend to tc flower offloading. So extend 'mtu' to tc flower offloading.
Signed-off-by: Po Liu <Po.Liu@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Using new helpers ip6t_unregister_table_pre_exit() and
ip6t_unregister_table_exit().
Fixes: b9e69e1273 ("netfilter: xtables: don't hook tables by default")
Signed-off-by: David Wilder <dwilder@us.ibm.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
The pre_exit will un-register the underlying hook and .exit will do
the table freeing. The netns core does an unconditional synchronize_rcu
after the pre_exit hooks insuring no packets are in flight that have
picked up the pointer before completing the un-register.
Fixes: b9e69e1273 ("netfilter: xtables: don't hook tables by default")
Signed-off-by: David Wilder <dwilder@us.ibm.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Using new helpers ipt_unregister_table_pre_exit() and
ipt_unregister_table_exit().
Fixes: b9e69e1273 ("netfilter: xtables: don't hook tables by default")
Signed-off-by: David Wilder <dwilder@us.ibm.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
The pre_exit will un-register the underlying hook and .exit will do the
table freeing. The netns core does an unconditional synchronize_rcu after
the pre_exit hooks insuring no packets are in flight that have picked up
the pointer before completing the un-register.
Fixes: b9e69e1273 ("netfilter: xtables: don't hook tables by default")
Signed-off-by: David Wilder <dwilder@us.ibm.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
The user tool modinfo is used to get information on kernel modules, including a
description where it is available.
This patch adds a brief MODULE_DESCRIPTION to netfilter kernel modules
(descriptions taken from Kconfig file or code comments)
Signed-off-by: Rob Gill <rrobgill@protonmail.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
When using ip_set with counters and comment, traffic causes the kernel
to panic on 32-bit ARM:
Alignment trap: not handling instruction e1b82f9f at [<bf01b0dc>]
Unhandled fault: alignment exception (0x221) at 0xea08133c
PC is at ip_set_match_extensions+0xe0/0x224 [ip_set]
The problem occurs when we try to update the 64-bit counters - the
faulting address above is not 64-bit aligned. The problem occurs
due to the way elements are allocated, for example:
set->dsize = ip_set_elem_len(set, tb, 0, 0);
map = ip_set_alloc(sizeof(*map) + elements * set->dsize);
If the element has a requirement for a member to be 64-bit aligned,
and set->dsize is not a multiple of 8, but is a multiple of four,
then every odd numbered elements will be misaligned - and hitting
an atomic64_add() on that element will cause the kernel to panic.
ip_set_elem_len() must return a size that is rounded to the maximum
alignment of any extension field stored in the element. This change
ensures that is the case.
Fixes: 95ad1f4a93 ("netfilter: ipset: Fix extension alignment")
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Acked-by: Jozsef Kadlecsik <kadlec@netfilter.org>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
The driver for Marvell switches puts all ports in IGMP snooping mode
which results in all IGMP/MLD frames that ingress on the ports to be
forwarded to the CPU only.
The bridge code in the kernel can then interpret these frames and act
upon them, for instance by updating the mdb in the switch to reflect
multicast memberships of stations connected to the ports. However,
the IGMP/MLD frames must then also be forwarded to other ports of the
bridge so external IGMP queriers can track membership reports, and
external multicast clients can receive query reports from foreign IGMP
queriers.
Currently, this is impossible as the EDSA tagger sets offload_fwd_mark
on the skb when it unwraps the tagged frames, and that will make the
switchdev layer prevent the skb from egressing on any other port of
the same switch.
To fix that, look at the To_CPU code in the DSA header and make
forwarding of the frame possible for trapped IGMP packets.
Introduce some #defines for the frame types to make the code a bit more
comprehensive.
This was tested on a Marvell 88E6352 variant.
Signed-off-by: Daniel Mack <daniel@zonque.org>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
When we modify or create a new fdb entry sometimes we want to avoid
refreshing its activity in order to track it properly. One example is
when a mac is received from EVPN multi-homing peer by FRR, which doesn't
want to change local activity accounting. It makes it static and sets a
flag to track its activity.
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch adds the ability to notify about activity of any entries
(static, permanent or ext_learn). EVPN multihoming peers need it to
properly and efficiently handle mac sync (peer active/locally active).
We add a new NFEA_ACTIVITY_NOTIFY attribute which is used to dump the
current activity state and to control if static entries should be monitored
at all. We use 2 bits - one to activate fdb entry tracking (disabled by
default) and the second to denote that an entry is inactive. We need
the second bit in order to avoid multiple notifications of inactivity.
Obviously this makes no difference for dynamic entries since at the time
of inactivity they get deleted, while the tracked non-dynamic entries get
the inactive bit set and get a notification.
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add an attribute to NDA which will contain all future fdb-specific
attributes in order to avoid polluting the NDA namespace with e.g.
bridge or vxlan specific attributes. The attribute is called
NDA_FDB_EXT_ATTRS and the structure would look like:
[NDA_FDB_EXT_ATTRS] = {
[NFEA_xxx]
}
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We can just pass ndm as an argument instead of its fields separately.
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
ovs connection tracking module performs de-fragmentation on incoming
fragmented traffic. Take info account if traffic has been de-fragmented
in execute_check_pkt_len action otherwise we will perform the wrong
nested action considering the original packet size. This issue typically
occurs if ovs-vswitchd adds a rule in the pipeline that requires connection
tracking (e.g. OVN stateful ACLs) before execute_check_pkt_len action.
Moreover take into account GSO fragment size for GSO packet in
execute_check_pkt_len routine
Fixes: 4d5ec89fc8 ("net: openvswitch: Add a new action check_pkt_len")
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Paolo Abeni