Fix the following warnings: [-Werror=frame-larger-than=]
In function ‘mlx5_tc_ct_entry_add_rule’:
drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c:541:1:
error: the frame size of 1136 bytes is larger than 1024 bytes
In function ‘__mlx5_tc_ct_flow_offload’:
drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c:1049:1:
error: the frame size of 1168 bytes is larger than 1024 bytes
Fixes: 4c3844d9e9 ("net/mlx5e: CT: Introduce connection tracking")
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Reviewed-by: Paul Blakey <paulb@mellanox.com>
If CONFIG_MLX5_TC_CT isn't enabled, all offloading of eswitch tc rules
fails on parsing ct match, even if there is no ct match.
Return success if there is no ct match, regardless of config.
Fixes: 4c3844d9e9 ("net/mlx5e: CT: Introduce connection tracking")
Signed-off-by: Paul Blakey <paulb@mellanox.com>
Reviewed-by: Oz Shlomo <ozsh@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c:
In function mlx5_tc_ct_parse_match:
drivers/net/ethernet/mellanox/mlx5/core/en/tc_ct.c:699:36: warning:
variable unnew set but not used [-Wunused-but-set-variable]
Fixes: 4c3844d9e9 ("net/mlx5e: CT: Introduce connection tracking")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Restore modify header writes the chain mapping on the packet.
This modify header and action is added on all prios connections,
and gets overwritten with the same value consecutively in prios
of the same chain.
Use the chain's modify header only for the last prio of a given tc
chain.
Signed-off-by: Paul Blakey <paulb@mellanox.com>
Reviewed-by: Oz Shlomo <ozsh@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Currently, if firmware doesn't support fwd and modify, driver fails
initializing eswitch chains while entering switchdev mode.
Instead, on such cases, disable the chains and prio feature (as we can't
restore the chain on miss) and the usage of fwd and modify.
Fixes: 8f1e0b97cc ("net/mlx5: E-Switch, Mark miss packets with new chain id mapping")
Signed-off-by: Paul Blakey <paulb@mellanox.com>
Reviewed-by: Oz Shlomo <ozsh@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
When CONFIG_MLX5_ESWITCH is unset, clang warns:
In file included from drivers/net/ethernet/mellanox/mlx5/core/main.c:58:
drivers/net/ethernet/mellanox/mlx5/core/eswitch.h:670:1: warning: unused
function 'esw_add_restore_rule' [-Wunused-function]
esw_add_restore_rule(struct mlx5_eswitch *esw, u32 tag)
^
1 warning generated.
This stub function is missing inline; add it to suppress the warning.
Fixes: 11b717d615 ("net/mlx5: E-Switch, Get reg_c0 value on CQE")
Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Commit e687ad60af ("netfilter: add netfilter ingress hook after
handle_ing() under unique static key") introduced the ability to
classify packets on ingress.
Allow the same on egress. Position the hook immediately before a packet
is handed to tc and then sent out on an interface, thereby mirroring the
ingress order. This order allows marking packets in the netfilter
egress hook and subsequently using the mark in tc. Another benefit of
this order is consistency with a lot of existing documentation which
says that egress tc is performed after netfilter hooks.
Egress hooks already exist for the most common protocols, such as
NF_INET_LOCAL_OUT or NF_ARP_OUT, and those are to be preferred because
they are executed earlier during packet processing. However for more
exotic protocols, there is currently no provision to apply netfilter on
egress. A common workaround is to enslave the interface to a bridge and
use ebtables, or to resort to tc. But when the ingress hook was
introduced, consensus was that users should be given the choice to use
netfilter or tc, whichever tool suits their needs best:
https://lore.kernel.org/netdev/20150430153317.GA3230@salvia/
This hook is also useful for NAT46/NAT64, tunneling and filtering of
locally generated af_packet traffic such as dhclient.
There have also been occasional user requests for a netfilter egress
hook in the past, e.g.:
https://www.spinics.net/lists/netfilter/msg50038.html
Performance measurements with pktgen surprisingly show a speedup rather
than a slowdown with this commit:
* Without this commit:
Result: OK: 34240933(c34238375+d2558) usec, 100000000 (60byte,0frags)
2920481pps 1401Mb/sec (1401830880bps) errors: 0
* With this commit:
Result: OK: 33997299(c33994193+d3106) usec, 100000000 (60byte,0frags)
2941410pps 1411Mb/sec (1411876800bps) errors: 0
* Without this commit + tc egress:
Result: OK: 39022386(c39019547+d2839) usec, 100000000 (60byte,0frags)
2562631pps 1230Mb/sec (1230062880bps) errors: 0
* With this commit + tc egress:
Result: OK: 37604447(c37601877+d2570) usec, 100000000 (60byte,0frags)
2659259pps 1276Mb/sec (1276444320bps) errors: 0
* With this commit + nft egress:
Result: OK: 41436689(c41434088+d2600) usec, 100000000 (60byte,0frags)
2413320pps 1158Mb/sec (1158393600bps) errors: 0
Tested on a bare-metal Core i7-3615QM, each measurement was performed
three times to verify that the numbers are stable.
Commands to perform a measurement:
modprobe pktgen
echo "add_device lo@3" > /proc/net/pktgen/kpktgend_3
samples/pktgen/pktgen_bench_xmit_mode_queue_xmit.sh -i 'lo@3' -n 100000000
Commands for testing tc egress:
tc qdisc add dev lo clsact
tc filter add dev lo egress protocol ip prio 1 u32 match ip dst 4.3.2.1/32
Commands for testing nft egress:
nft add table netdev t
nft add chain netdev t co \{ type filter hook egress device lo priority 0 \; \}
nft add rule netdev t co ip daddr 4.3.2.1/32 drop
All testing was performed on the loopback interface to avoid distorting
measurements by the packet handling in the low-level Ethernet driver.
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Cc: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Prepare for addition of a netfilter egress hook by generalizing the
ingress hook introduced by commit e687ad60af ("netfilter: add
netfilter ingress hook after handle_ing() under unique static key").
In particular, rename and refactor the ingress hook's static inlines
such that they can be reused for an egress hook.
No functional change intended.
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Cc: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Prepare for addition of a netfilter egress hook by renaming
<linux/netfilter_ingress.h> to <linux/netfilter_netdev.h>.
The egress hook also necessitates a refactoring of the include file,
but that is done in a separate commit to ease reviewing.
No functional change intended.
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Cc: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Pengcheng Yang says:
====================
tcp: fix stretch ACK bugs in congestion control modules
"stretch ACKs" (caused by LRO, GRO, delayed ACKs or middleboxes)
can cause serious performance shortfalls in common congestion
control algorithms. Neal Cardwell submitted a series of patches
starting with commit e73ebb0881 ("tcp: stretch ACK fixes prep")
to handle stretch ACKs and fixed stretch ACK bugs in Reno and
CUBIC congestion control algorithms.
This patch series continues to fix bic, scalable, veno and yeah
congestion control algorithms to handle stretch ACKs.
Changes in v2:
- Provide [PATCH 0/N] to describe the modifications of this patch series
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Change Yeah to properly handle stretch ACKs in additive
increase mode by passing in the count of ACKed packets
to tcp_cong_avoid_ai().
In addition, we re-implemented the scalable path using
tcp_cong_avoid_ai() and removed the pkts_acked variable.
Signed-off-by: Pengcheng Yang <yangpc@wangsu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Change Veno to properly handle stretch ACKs in additive
increase mode by passing in the count of ACKed packets
to tcp_cong_avoid_ai().
Signed-off-by: Pengcheng Yang <yangpc@wangsu.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
No code logic has been changed in this patch.
Signed-off-by: Pengcheng Yang <yangpc@wangsu.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Change Scalable to properly handle stretch ACKs in additive
increase mode by passing in the count of ACKed packets to
tcp_cong_avoid_ai().
In addition, because we are now precisely accounting for
stretch ACKs, including delayed ACKs, we can now change
TCP_SCALABLE_AI_CNT to 100.
Signed-off-by: Pengcheng Yang <yangpc@wangsu.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Changes BIC to properly handle stretch ACKs in additive
increase mode by passing in the count of ACKed packets
to tcp_cong_avoid_ai().
Signed-off-by: Pengcheng Yang <yangpc@wangsu.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
XDP-redirect is broken in this driver sfc. XDP_REDIRECT requires
tailroom for skb_shared_info when creating an SKB based on the
redirected xdp_frame (both in cpumap and veth).
The fix requires some initial explaining. The driver uses RX page-split
when possible. It reserves the top 64 bytes in the RX-page for storing
dma_addr (struct efx_rx_page_state). It also have the XDP recommended
headroom of XDP_PACKET_HEADROOM (256 bytes). As it doesn't reserve any
tailroom, it can still fit two standard MTU (1500) frames into one page.
The sizeof struct skb_shared_info in 320 bytes. Thus drivers like ixgbe
and i40e, reduce their XDP headroom to 192 bytes, which allows them to
fit two frames with max 1536 bytes into a 4K page (192+1536+320=2048).
The fix is to reduce this drivers headroom to 128 bytes and add the 320
bytes tailroom. This account for reserved top 64 bytes in the page, and
still fit two frame in a page for normal MTUs.
We must never go below 128 bytes of headroom for XDP, as one cacheline
is for xdp_frame area and next cacheline is reserved for metadata area.
Fixes: eb9a36be7f ("sfc: perform XDP processing on received packets")
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Acked-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Rearrange the config files for remoteproc and IPA to fix their
interdependencies.
First, have CONFIG_QCOM_Q6V5_MSS select QCOM_Q6V5_IPA_NOTIFY so the
notification code is built regardless of whether IPA needs it.
Next, represent QCOM_IPA as being dependent on QCOM_Q6V5_MSS rather
than setting its value to match QCOM_Q6V5_COMMON (which is selected
by QCOM_Q6V5_MSS).
Drop all dependencies from QCOM_Q6V5_IPA_NOTIFY. The notification
code will be built whenever QCOM_Q6V5_MSS is set, and it has no other
dependencies.
Signed-off-by: Alex Elder <elder@linaro.org>
Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
This path fixes the suspicious RCU usage warning reported by
kernel test robot.
net/kcm/kcmproc.c:#RCU-list_traversed_in_non-reader_section
There is no need to use list_for_each_entry_rcu() in
kcm_stats_seq_show() as the list is always traversed under
knet->mutex held.
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Madhuparna Bhowmik <madhuparnabhowmik10@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Remove set but not used variables 'sw_comp_cons' and 'hw_comp_cons'
to fix gcc '-Wunused-but-set-variable' warning:
drivers/net/ethernet/qlogic/qede/qede_ethtool.c: In function qede_selftest_receive_traffic:
drivers/net/ethernet/qlogic/qede/qede_ethtool.c:1569:20:
warning: variable sw_comp_cons set but not used [-Wunused-but-set-variable]
drivers/net/ethernet/qlogic/qede/qede_ethtool.c: In function qede_selftest_receive_traffic:
drivers/net/ethernet/qlogic/qede/qede_ethtool.c:1569:6:
warning: variable hw_comp_cons set but not used [-Wunused-but-set-variable]
After removing 'hw_comp_cons',the memory barrier 'rmb()' and its comments become useless,
so remove them as well.
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Zheng Zengkai <zhengzengkai@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
For a single pedit action, multiple offload entries may be used. Set the
hw_stats_type to all of them.
Fixes: 44f8658017 ("sched: act: allow user to specify type of HW stats for a filter")
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Dejin Zheng says:
====================
net: stmmac: Use readl_poll_timeout() to simplify the code
This patch sets just for replace the open-coded loop to the
readl_poll_timeout() helper macro for simplify the code in
stmmac driver.
v2 -> v3:
- return whatever error code by readl_poll_timeout() returned.
v1 -> v2:
- no changed. I am a newbie and sent this patch a month
ago (February 6th). So far, I have not received any comments or
suggestion. I think it may be lost somewhere in the world, so
resend it.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
The dwmac4_dma_reset() function use an open coded of readl_poll_timeout().
Replace the open coded handling with the proper function.
Signed-off-by: Dejin Zheng <zhengdejin5@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The init_systime() function use an open coded of readl_poll_timeout().
Replace the open coded handling with the proper function.
Signed-off-by: Dejin Zheng <zhengdejin5@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
drivers/crypto/chelsio/chcr_ktls.c: In function chcr_ktls_cpl_set_tcb_rpl:
drivers/crypto/chelsio/chcr_ktls.c:662:11: warning:
variable status set but not used [-Wunused-but-set-variable]
commit 8a30923e15 ("cxgb4/chcr: Save tx keys and handle HW response")
involved this unused variable, remove it.
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Netlink support of extended packet number cipher suites,
allows adding and updating XPN macsec interfaces.
Added support in:
* Creating interfaces with GCM-AES-XPN-128 and GCM-AES-XPN-256 suites.
* Setting and getting 64bit packet numbers with of SAs.
* Setting (only on SA creation) and getting ssci of SAs.
* Setting salt when installing a SAK.
Added 2 cipher suite identifiers according to 802.1AE-2018 table 14-1:
* MACSEC_CIPHER_ID_GCM_AES_XPN_128
* MACSEC_CIPHER_ID_GCM_AES_XPN_256
In addition, added 2 new netlink attribute types:
* MACSEC_SA_ATTR_SSCI
* MACSEC_SA_ATTR_SALT
Depends on: macsec: Support XPN frame handling - IEEE 802.1AEbw.
Signed-off-by: Era Mayflower <mayflowerera@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Support extended packet number cipher suites (802.1AEbw) frames handling.
This does not include the needed netlink patches.
* Added xpn boolean field to `struct macsec_secy`.
* Added ssci field to `struct_macsec_tx_sa` (802.1AE figure 10-5).
* Added ssci field to `struct_macsec_rx_sa` (802.1AE figure 10-5).
* Added salt field to `struct macsec_key` (802.1AE 10.7 NOTE 1).
* Created pn_t type for easy access to lower and upper halves.
* Created salt_t type for easy access to the "ssci" and "pn" parts.
* Created `macsec_fill_iv_xpn` function to create IV in XPN mode.
* Support in PN recovery and preliminary replay check in XPN mode.
In addition, according to IEEE 802.1AEbw figure 10-5, the PN of incoming
frame can be 0 when XPN cipher suite is used, so fixed the function
`macsec_validate_skb` to fail on PN=0 only if XPN is off.
Signed-off-by: Era Mayflower <mayflowerera@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Russell King says:
====================
net: dsa: improve serdes integration
Depends on "net: mii clause 37 helpers".
Andrew Lunn mentioned that the Serdes PCS found in Marvell DSA switches
does not automatically update the switch MACs with the link parameters.
Currently, the DSA code implements a work-around for this.
This series improves the Serdes integration, making use of the recent
phylink changes to support split MAC/PCS setups. One noticable
improvement for userspace is that ethtool can now report the link
partner's advertisement.
This repost has no changes compared to the previous posting; however,
the regression Andrew had found which exists even without this patch
set has now been fixed by Andrew and merged into the net-next tree.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Use the status of the PHY_DETECT bit to determine whether we need to
force the MAC settings in mac_link_up() and mac_link_down().
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
The port_link_state method is only used by mv88e6xxx_port_setup_mac(),
which is now only called during port setup, rather than also being
called via phylink's mac_config method.
Remove this now unnecessary optimisation, which allows us to remove the
port_link_state methods as well.
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Setting the speed independently of duplex makes little sense; the two
parameters result from negotiation or fixed setup, and may have inter-
dependencies. Moreover, they are always controlled via the same
register - having them split means we have to read-modify-write this
register twice.
Combine the two operations into a single port_set_speed_duplex()
operation. Not only is this more efficient, it reduces the size of the
code as well.
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
phylink_mac_change() is supposed to be called with a 'false' argument
if the link has gone down since it was last reported up; this is to
ensure that link events along with renegotiation events are always
correctly reported to userspace.
Read the BMSR once when we have an interrupt, and report the link
latched status to phylink via phylink_mac_change(). phylink will deal
automatically with re-reading the link state once it has processed the
link-down event.
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Extend the mv88e6xxx phylink implementation down to Serdes PHYs, which
handle the PCS layer of such links.
- Implement phylink PCS link state reading, so that we can provide
ethtool with the linkmodes and link speed in the expected manner.
Note: this will only be called for in-band negotiation, which is
only supported by the serdes interfaces.
- Implement phylink PCS configuration, so that the in-band AN and
advertisement can be configured.
- Implement phylink PCS negotiation restart, so that the in-band AN
can be restarted.
- Implement phylink PCS link up, so that when operating out-of-band,
the Serdes can be configured for the appropriate fixed speed mode.
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Only configure the interface settings in mac_config(), leaving the
speed and duplex settings to mac_link_up to deal with.
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
The SGMII/1000base-X serdes register set is a clause 22 register set
offset at 0x2000 in the PHYXS device. Rather than inventing our own
defintions, use those that already exist, and name the register
MV88E6390_SGMII_BMCR. Also remove the unused MV88E6390_SGMII_STATUS
definitions.
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Issue a warning to the kernel log if phylink_mac_link_state() returns
an error. This should not occur, but let's make it visible.
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Russell King says:
====================
net: mii clause 37 helpers
This is a re-post of two patches that are common to two series that
I've sent in recent weeks; I'm re-posting them separately in the hope
that they can be merged. No changes from either of the previous
postings.
These patches:
1. convert the existing (unused) mii_lpa_to_ethtool_lpa_x() function
to a linkmode variant.
2. add a helper for clause 37 advertisements, supporting both the
1000baseX and defacto 2500baseX variants. Note that ethtool does
not support half duplex for either of these, and we make no effort
to do so.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Add a helper to convert a linkmode advertisement to a clause 37
advertisement value for 1000base-x and 2500base-x.
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add a LPA to linkmode decoder for 1000BASE-X protocols; this decoder
only provides the modify semantics similar to other such decoders.
This replaces the unused mii_lpa_to_ethtool_lpa_x() helper.
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
since commit b884fa4617 ("netfilter: conntrack: unify sysctl handling")
conntrack no longer exposes most of its sysctls (e.g. tcp timeouts
settings) to network namespaces that are not owned by the initial user
namespace.
This patch exposes all sysctls even if the namespace is unpriviliged.
compared to a 4.19 kernel, the newly visible and writeable sysctls are:
net.netfilter.nf_conntrack_acct
net.netfilter.nf_conntrack_timestamp
.. to allow to enable accouting and timestamp extensions.
net.netfilter.nf_conntrack_events
.. to turn off conntrack event notifications.
net.netfilter.nf_conntrack_checksum
.. to disable checksum validation.
net.netfilter.nf_conntrack_log_invalid
.. to enable logging of packets deemed invalid by conntrack.
newly visible sysctls that are only exported as read-only:
net.netfilter.nf_conntrack_count
.. current number of conntrack entries living in this netns.
net.netfilter.nf_conntrack_max
.. global upperlimit (maximum size of the table).
net.netfilter.nf_conntrack_buckets
.. size of the conntrack table (hash buckets).
net.netfilter.nf_conntrack_expect_max
.. maximum number of permitted expectations in this netns.
net.netfilter.nf_conntrack_helper
.. conntrack helper auto assignment.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Update nft_add_set_elem() to handle the NFTA_SET_ELEM_EXPR netlink
attribute. This patch allows users to to add elements with stateful
expressions.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
If the AVX2 set is available, we can exploit the repetitive
characteristic of this algorithm to provide a fast, vectorised
version by using 256-bit wide AVX2 operations for bucket loads and
bitwise intersections.
In most cases, this implementation consistently outperforms rbtree
set instances despite the fact they are configured to use a given,
single, ranged data type out of the ones used for performance
measurements by the nft_concat_range.sh kselftest.
That script, injecting packets directly on the ingoing device path
with pktgen, reports, averaged over five runs on a single AMD Epyc
7402 thread (3.35GHz, 768 KiB L1D$, 12 MiB L2$), the figures below.
CONFIG_RETPOLINE was not set here.
Note that this is not a fair comparison over hash and rbtree set
types: non-ranged entries (used to have a reference for hash types)
would be matched faster than this, and matching on a single field
only (which is the case for rbtree) is also significantly faster.
However, it's not possible at the moment to choose this set type
for non-ranged entries, and the current implementation also needs
a few minor adjustments in order to match on less than two fields.
---------------.-----------------------------------.------------.
AMD Epyc 7402 | baselines, Mpps | this patch |
1 thread |___________________________________|____________|
3.35GHz | | | | | |
768KiB L1D$ | netdev | hash | rbtree | | |
---------------| hook | no | single | | pipapo |
type entries | drop | ranges | field | pipapo | AVX2 |
---------------|--------|--------|--------|--------|------------|
net,port | | | | | |
1000 | 19.0 | 10.4 | 3.8 | 4.0 | 7.5 +87% |
---------------|--------|--------|--------|--------|------------|
port,net | | | | | |
100 | 18.8 | 10.3 | 5.8 | 6.3 | 8.1 +29% |
---------------|--------|--------|--------|--------|------------|
net6,port | | | | | |
1000 | 16.4 | 7.6 | 1.8 | 2.1 | 4.8 +128% |
---------------|--------|--------|--------|--------|------------|
port,proto | | | | | |
30000 | 19.6 | 11.6 | 3.9 | 0.5 | 2.6 +420% |
---------------|--------|--------|--------|--------|------------|
net6,port,mac | | | | | |
10 | 16.5 | 5.4 | 4.3 | 3.4 | 4.7 +38% |
---------------|--------|--------|--------|--------|------------|
net6,port,mac, | | | | | |
proto 1000 | 16.5 | 5.7 | 1.9 | 1.4 | 3.6 +26% |
---------------|--------|--------|--------|--------|------------|
net,mac | | | | | |
1000 | 19.0 | 8.4 | 3.9 | 2.5 | 6.4 +156% |
---------------'--------'--------'--------'--------'------------'
A similar strategy could be easily reused to implement specialised
versions for other SIMD sets, and I plan to post at least a NEON
version at a later time.
Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Move most macros and helpers to a header file, so that they can be
conveniently used by related implementations.
No functional changes are intended here.
Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
SIMD vector extension sets require stricter alignment than native
instruction sets to operate efficiently (AVX, NEON) or for some
instructions to work at all (AltiVec).
Provide facilities to define arbitrary alignment for lookup tables
and scratch maps. By defining byte alignment with NFT_PIPAPO_ALIGN,
lt_aligned and scratch_aligned pointers become available.
Additional headroom is allocated, and pointers to the possibly
unaligned, originally allocated areas are kept so that they can
be freed.
Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
While grouping matching bits in groups of four saves memory compared
to the more natural choice of 8-bit words (lookup table size is one
eighth), it comes at a performance cost, as the number of lookup
comparisons is doubled, and those also needs bitshifts and masking.
Introduce support for 8-bit lookup groups, together with a mapping
mechanism to dynamically switch, based on defined per-table size
thresholds and hysteresis, between 8-bit and 4-bit groups, as tables
grow and shrink. Empty sets start with 8-bit groups, and per-field
tables are converted to 4-bit groups if they get too big.
An alternative approach would have been to swap per-set lookup
operation functions as needed, but this doesn't allow for different
group sizes in the same set, which looks desirable if some fields
need significantly more matching data compared to others due to
heavier impact of ranges (e.g. a big number of subnets with
relatively simple port specifications).
Allowing different group sizes for the same lookup functions implies
the need for further conditional clauses, whose cost, however,
appears to be negligible in tests.
The matching rate figures below were obtained for x86_64 running
the nft_concat_range.sh "performance" cases, averaged over five
runs, on a single thread of an AMD Epyc 7402 CPU, and for aarch64
on a single thread of a BCM2711 (Raspberry Pi 4 Model B 4GB),
clocked at a stable 2147MHz frequency:
---------------.-----------------------------------.------------.
AMD Epyc 7402 | baselines, Mpps | this patch |
1 thread |___________________________________|____________|
3.35GHz | | | | | |
768KiB L1D$ | netdev | hash | rbtree | | |
---------------| hook | no | single | pipapo | pipapo |
type entries | drop | ranges | field | 4 bits | bit switch |
---------------|--------|--------|--------|--------|------------|
net,port | | | | | |
1000 | 19.0 | 10.4 | 3.8 | 2.8 | 4.0 +43% |
---------------|--------|--------|--------|--------|------------|
port,net | | | | | |
100 | 18.8 | 10.3 | 5.8 | 5.5 | 6.3 +14% |
---------------|--------|--------|--------|--------|------------|
net6,port | | | | | |
1000 | 16.4 | 7.6 | 1.8 | 1.3 | 2.1 +61% |
---------------|--------|--------|--------|--------|------------|
port,proto | | | | | [1] |
30000 | 19.6 | 11.6 | 3.9 | 0.3 | 0.5 +66% |
---------------|--------|--------|--------|--------|------------|
net6,port,mac | | | | | |
10 | 16.5 | 5.4 | 4.3 | 2.6 | 3.4 +31% |
---------------|--------|--------|--------|--------|------------|
net6,port,mac, | | | | | |
proto 1000 | 16.5 | 5.7 | 1.9 | 1.0 | 1.4 +40% |
---------------|--------|--------|--------|--------|------------|
net,mac | | | | | |
1000 | 19.0 | 8.4 | 3.9 | 1.7 | 2.5 +47% |
---------------'--------'--------'--------'--------'------------'
[1] Causes switch of lookup table buckets for 'port', not 'proto',
to 4-bit groups
---------------.-----------------------------------.------------.
BCM2711 | baselines, Mpps | this patch |
1 thread |___________________________________|____________|
2147MHz | | | | | |
32KiB L1D$ | netdev | hash | rbtree | | |
---------------| hook | no | single | pipapo | pipapo |
type entries | drop | ranges | field | 4 bits | bit switch |
---------------|--------|--------|--------|--------|------------|
net,port | | | | | |
1000 | 1.63 | 1.37 | 0.87 | 0.61 | 0.70 +17% |
---------------|--------|--------|--------|--------|------------|
port,net | | | | | |
100 | 1.64 | 1.36 | 1.02 | 0.78 | 0.81 +4% |
---------------|--------|--------|--------|--------|------------|
net6,port | | | | | |
1000 | 1.56 | 1.27 | 0.65 | 0.34 | 0.50 +47% |
---------------|--------|--------|--------|--------|------------|
port,proto [2] | | | | | |
10000 | 1.68 | 1.43 | 0.84 | 0.30 | 0.40 +13% |
---------------|--------|--------|--------|--------|------------|
net6,port,mac | | | | | |
10 | 1.56 | 1.14 | 1.02 | 0.62 | 0.66 +6% |
---------------|--------|--------|--------|--------|------------|
net6,port,mac, | | | | | |
proto 1000 | 1.56 | 1.12 | 0.64 | 0.27 | 0.40 +48% |
---------------|--------|--------|--------|--------|------------|
net,mac | | | | | |
1000 | 1.63 | 1.26 | 0.87 | 0.41 | 0.53 +29% |
---------------'--------'--------'--------'--------'------------'
[2] Using 10000 entries instead of 30000 as it would take way too
long for the test script to generate all of them
Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Get rid of all hardcoded assumptions that buckets in lookup tables
correspond to four-bit groups, and replace them with appropriate
calculations based on a variable group size, now stored in struct
field.
The group size could now be in principle any divisor of eight. Note,
though, that lookup and get functions need an implementation
intimately depending on the group size, and the only supported size
there, currently, is four bits, which is also the initial and only
used size at the moment.
While at it, drop 'groups' from struct nft_pipapo: it was never used.
Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
