move tc indirect block to flow_offload and rename
it to flow indirect block.The nf_tables can use the
indr block architecture.
Signed-off-by: wenxu <wenxu@ucloud.cn>
Acked-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Edward Cree says:
====================
net: batched receive in GRO path
This series listifies part of GRO processing, in a manner which allows those
packets which are not GROed (i.e. for which dev_gro_receive returns
GRO_NORMAL) to be passed on to the listified regular receive path.
dev_gro_receive() itself is not listified, nor the per-protocol GRO
callback, since GRO's need to hold packets on lists under napi->gro_hash
makes keeping the packets on other lists awkward, and since the GRO control
block state of held skbs can refer only to one 'new' skb at a time.
Instead, when napi_frags_finish() handles a GRO_NORMAL result, stash the skb
onto a list in the napi struct, which is received at the end of the napi
poll or when its length exceeds the (new) sysctl net.core.gro_normal_batch.
Performance figures with this series, collected on a back-to-back pair of
Solarflare sfn8522-r2 NICs with 120-second NetPerf tests. In the stats,
sample size n for old and new code is 6 runs each; p is from a Welch t-test.
Tests were run both with GRO enabled and disabled, the latter simulating
uncoalesceable packets (e.g. due to IP or TCP options). The receive side
(which was the device under test) had the NetPerf process pinned to one CPU,
and the device interrupts pinned to a second CPU. CPU utilisation figures
(used in cases of line-rate performance) are summed across all CPUs.
net.core.gro_normal_batch was left at its default value of 8.
TCP 4 streams, GRO on: all results line rate (9.415Gbps)
net-next: 210.3% cpu
after #1: 181.5% cpu (-13.7%, p=0.031 vs net-next)
after #3: 196.7% cpu (- 8.4%, p=0.136 vs net-next)
TCP 4 streams, GRO off:
net-next: 8.017 Gbps
after #1: 7.785 Gbps (- 2.9%, p=0.385 vs net-next)
after #3: 7.604 Gbps (- 5.1%, p=0.282 vs net-next. But note *)
TCP 1 stream, GRO off:
net-next: 6.553 Gbps
after #1: 6.444 Gbps (- 1.7%, p=0.302 vs net-next)
after #3: 6.790 Gbps (+ 3.6%, p=0.169 vs net-next)
TCP 1 stream, GRO on, busy_read = 50: all results line rate
net-next: 156.0% cpu
after #1: 174.5% cpu (+11.9%, p=0.015 vs net-next)
after #3: 165.0% cpu (+ 5.8%, p=0.147 vs net-next)
TCP 1 stream, GRO off, busy_read = 50:
net-next: 6.488 Gbps
after #1: 6.625 Gbps (+ 2.1%, p=0.059 vs net-next)
after #3: 7.351 Gbps (+13.3%, p=0.026 vs net-next)
TCP_RR 100 streams, GRO off, 8000 byte payload
net-next: 995.083 us
after #1: 969.167 us (- 2.6%, p=0.204 vs net-next)
after #3: 976.433 us (- 1.9%, p=0.254 vs net-next)
TCP_RR 100 streams, GRO off, 8000 byte payload, busy_read = 50:
net-next: 2.851 ms
after #1: 2.871 ms (+ 0.7%, p=0.134 vs net-next)
after #3: 2.937 ms (+ 3.0%, p<0.001 vs net-next)
TCP_RR 100 streams, GRO off, 1 byte payload, busy_read = 50:
net-next: 867.317 us
after #1: 865.717 us (- 0.2%, p=0.334 vs net-next)
after #3: 868.517 us (+ 0.1%, p=0.414 vs net-next)
(*) These tests produced a mixture of line-rate and below-line-rate results,
meaning that statistically speaking the results were 'censored' by the
upper bound, and were thus not normally distributed, making a Welch t-test
mathematically invalid. I therefore also calculated estimators according
to [1], which gave the following:
net-next: 8.133 Gbps
after #1: 8.130 Gbps (- 0.0%, p=0.499 vs net-next)
after #3: 7.680 Gbps (- 5.6%, p=0.285 vs net-next)
(though my procedure for determining ν wasn't mathematically well-founded
either, so take that p-value with a grain of salt).
A further check came from dividing the bandwidth figure by the CPU usage for
each test run, giving:
net-next: 3.461
after #1: 3.198 (- 7.6%, p=0.145 vs net-next)
after #3: 3.641 (+ 5.2%, p=0.280 vs net-next)
The above results are fairly mixed, and in most cases not statistically
significant. But I think we can roughly conclude that the series
marginally improves non-GROable throughput, without hurting latency
(except in the large-payload busy-polling case, which in any case yields
horrid performance even on net-next (almost triple the latency without
busy-poll). Also, drivers which, unlike sfc, pass UDP traffic to GRO
would expect to see a benefit from gaining access to batching.
Changed in v3:
* gro_normal_batch sysctl now uses SYSCTL_ONE instead of &one
* removed RFC tags (no comments after a week means no-one objects, right?)
Changed in v2:
* During busy poll, call gro_normal_list() to receive batched packets
after each cycle of the napi busy loop. See comments in Patch #3 for
complications of doing the same in busy_poll_stop().
[1]: Cohen 1959, doi: 10.1080/00401706.1959.10489859
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
When GRO decides not to coalesce a packet, in napi_frags_finish(), instead
of passing it to the stack immediately, place it on a list in the napi
struct. Then, at flush time (napi_complete_done(), napi_poll(), or
napi_busy_loop()), call netif_receive_skb_list_internal() on the list.
We'd like to do that in napi_gro_flush(), but it's not called if
!napi->gro_bitmask, so we have to do it in the callers instead. (There are
a handful of drivers that call napi_gro_flush() themselves, but it's not
clear why, or whether this will affect them.)
Because a full 64 packets is an inefficiently large batch, also consume the
list whenever it exceeds gro_normal_batch, a new net/core sysctl that
defaults to 8.
Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Same rationale as for sfc, except that this wasn't performance-tested.
Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We already scored points when handling the RX event, no-one else does this,
and looking at the history it appears this was originally meant to only
score on merges, not on GRO_NORMAL. Moreover, it gets in the way of
changing GRO to not immediately pass GRO_NORMAL skbs to the stack.
Performance testing with four TCP streams received on a single CPU (where
throughput was line rate of 9.4Gbps in all tests) showed a 13.7% reduction
in RX CPU usage (n=6, p=0.03).
Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Supported ports in ethtool <eth1> are displayed based on media type.
For media type fibre and twinaxial, port type is "FIBRE". Media type
Base-T is "TP" and media KR is "Backplane".
V1->V2:
Corrected the subject.
Signed-off-by: Rahul Verma <rahulv@marvell.com>
Signed-off-by: Michal Kalderon <michal.kalderon@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
All refcount operations are protected by spinlocks now.
Then the atomic counter can be replaced by a normal int.
This patch depends on PATCH 1/2.
Signed-off-by: Chuhong Yuan <hslester96@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The atomic_dec_and_test() is not safe because it is
outside of locks.
Move the locks of t4_smte_free() to its caller,
cxgb4_smt_release() to protect the atomic decrement.
Fixes: 3bdb376e69 ("cxgb4: introduce SMT ops to prepare for SMAC rewrite support")
Signed-off-by: Chuhong Yuan <hslester96@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add IPv4 and IPv6 l2tp tests. Current set is over IP and with
IPsec.
v2
- add l2tp.sh to TEST_PROGS in Makefile
Signed-off-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
refcount_t is better for reference counters since its
implementation can prevent overflows.
So convert atomic_t ref counters to refcount_t.
Signed-off-by: Chuhong Yuan <hslester96@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
refcount_t is better for reference counters since its
implementation can prevent overflows.
So convert atomic_t ref counters to refcount_t.
Signed-off-by: Chuhong Yuan <hslester96@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Simon Wunderlich says:
====================
This feature/cleanup patchset includes the following patches:
- bump version strings, by Simon Wunderlich
- Replace usage of strlcpy with strscpy, by Sven Eckelmann
- Add OGMv2 per-interface queue and aggregations, by Linus Luessing
(2 patches)
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Pull networking fixes from David Miller:
"Yeah I should have sent a pull request last week, so there is a lot
more here than usual:
1) Fix memory leak in ebtables compat code, from Wenwen Wang.
2) Several kTLS bug fixes from Jakub Kicinski (circular close on
disconnect etc.)
3) Force slave speed check on link state recovery in bonding 802.3ad
mode, from Thomas Falcon.
4) Clear RX descriptor bits before assigning buffers to them in
stmmac, from Jose Abreu.
5) Several missing of_node_put() calls, mostly wrt. for_each_*() OF
loops, from Nishka Dasgupta.
6) Double kfree_skb() in peak_usb can driver, from Stephane Grosjean.
7) Need to hold sock across skb->destructor invocation, from Cong
Wang.
8) IP header length needs to be validated in ipip tunnel xmit, from
Haishuang Yan.
9) Use after free in ip6 tunnel driver, also from Haishuang Yan.
10) Do not use MSI interrupts on r8169 chips before RTL8168d, from
Heiner Kallweit.
11) Upon bridge device init failure, we need to delete the local fdb.
From Nikolay Aleksandrov.
12) Handle erros from of_get_mac_address() properly in stmmac, from
Martin Blumenstingl.
13) Handle concurrent rename vs. dump in netfilter ipset, from Jozsef
Kadlecsik.
14) Setting NETIF_F_LLTX on mac80211 causes complete breakage with
some devices, so revert. From Johannes Berg.
15) Fix deadlock in rxrpc, from David Howells.
16) Fix Kconfig deps of enetc driver, we must have PHYLIB. From Yue
Haibing.
17) Fix mvpp2 crash on module removal, from Matteo Croce.
18) Fix race in genphy_update_link, from Heiner Kallweit.
19) bpf_xdp_adjust_head() stopped working with generic XDP when we
fixes generic XDP to support stacked devices properly, fix from
Jesper Dangaard Brouer.
20) Unbalanced RCU locking in rt6_update_exception_stamp_rt(), from
David Ahern.
21) Several memory leaks in new sja1105 driver, from Vladimir Oltean"
* git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (214 commits)
net: dsa: sja1105: Fix memory leak on meta state machine error path
net: dsa: sja1105: Fix memory leak on meta state machine normal path
net: dsa: sja1105: Really fix panic on unregistering PTP clock
net: dsa: sja1105: Use the LOCKEDS bit for SJA1105 E/T as well
net: dsa: sja1105: Fix broken learning with vlan_filtering disabled
net: dsa: qca8k: Add of_node_put() in qca8k_setup_mdio_bus()
net: sched: sample: allow accessing psample_group with rtnl
net: sched: police: allow accessing police->params with rtnl
net: hisilicon: Fix dma_map_single failed on arm64
net: hisilicon: fix hip04-xmit never return TX_BUSY
net: hisilicon: make hip04_tx_reclaim non-reentrant
tc-testing: updated vlan action tests with batch create/delete
net sched: update vlan action for batched events operations
net: stmmac: tc: Do not return a fragment entry
net: stmmac: Fix issues when number of Queues >= 4
net: stmmac: xgmac: Fix XGMAC selftests
be2net: disable bh with spin_lock in be_process_mcc
net: cxgb3_main: Fix a resource leak in a error path in 'init_one()'
net: ethernet: sun4i-emac: Support phy-handle property for finding PHYs
net: bridge: move default pvid init/deinit to NETDEV_REGISTER/UNREGISTER
...
Jeff Kirsher says:
====================
40GbE Intel Wired LAN Driver Updates 2019-08-05
This series contains updates to i40e driver only.
Dmitrii adds missing statistic counters for VEB and VEB TC's.
Slawomir adds support for logging the "Disable Firmware LLDP" flag
option and its current status.
Jake fixes an issue where VF's being notified of their link status
before their queues are enabled which was causing issues. So always
report link status down when the VF queues are not enabled. Also adds
future proofing when statistics are added or removed by adding checks to
ensure the data pointer for the strings lines up with the expected
statistics count.
Czeslaw fixes the advertised mode reported in ethtool for FEC, where the
"None BaseR RS" was always being displayed no matter what the mode it
was in. Also added logging information when the PF is entering or
leaving "allmulti" (or promiscuous) mode. Fixed up the logging logic
for VF's when leaving multicast mode to not include unicast as well.
v2: drop Aleksandr's patch (previously patch #2 in the series) to
display the VF MAC address that is set by the VF while community
feedback is addressed.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently in function ovs_dp_process_packet(), return values of
ovs_execute_actions() are silently discarded. This patch prints out
an debug message when error happens so as to provide helpful hints
for debugging.
Acked-by: Pravin B Shelar <pshelar@ovn.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Vladimir Oltean says:
====================
Fixes for SJA1105 DSA: FDBs, Learning and PTP
This is an assortment of functional fixes for the sja1105 switch driver
targeted for the "net" tree (although they apply on net-next just as
well).
Patch 1/5 ("net: dsa: sja1105: Fix broken learning with vlan_filtering
disabled") repairs a breakage introduced in the early development stages
of the driver: support for traffic from the CPU has broken "normal"
frame forwarding (based on DMAC) - there is connectivity through the
switch only because all frames are flooded.
I debated whether this patch qualifies as a fix, since it puts the
switch into a mode it has never operated in before (aka SVL). But
"normal" forwarding did use to work before the "Traffic support for
SJA1105 DSA driver" patchset, and arguably this patch should have been
part of that.
Also, it would be strange for this feature to be broken in the 5.2 LTS.
Patch 2/5 ("net: dsa: sja1105: Use the LOCKEDS bit for SJA1105 E/T as
well") is a simplification of a previous FDB-related patch that is
currently in the 5.3 rc's.
Patches 3/5 - 5/5 fix various crashes found while running linuxptp over the
switch ports for extended periods of time, or in conjunction with other
error conditions. The fixed-up commits were all introduced in 5.2.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
When RX timestamping is enabled and two link-local (non-meta) frames are
received in a row, this constitutes an error.
The tagger is always caching the last link-local frame, in an attempt to
merge it with the meta follow-up frame when that arrives. To recover
from the above error condition, the initial cached link-local frame is
dropped and the second frame in a row is cached (in expectance of the
second meta frame).
However, when dropping the initial link-local frame, its backing memory
was being leaked.
Fixes: f3097be21b ("net: dsa: sja1105: Add a state machine for RX timestamping")
Signed-off-by: Vladimir Oltean <olteanv@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
After a meta frame is received, it is associated with the cached
sp->data->stampable_skb from the DSA tagger private structure.
Cached means its refcount is incremented with skb_get() in order for
dsa_switch_rcv() to not free it when the tagger .rcv returns NULL.
The mistake is that skb_unref() is not the correct function to use. It
will correctly decrement the refcount (which will go back to zero) but
the skb memory will not be freed. That is the job of kfree_skb(), which
also calls skb_unref().
But it turns out that freeing the cached stampable_skb is in fact not
necessary. It is still a perfectly valid skb, and now it is even
annotated with the partial RX timestamp. So remove the skb_copy()
altogether and simply pass the stampable_skb with a refcount of 1
(incremented by us, decremented by dsa_switch_rcv) up the stack.
Fixes: f3097be21b ("net: dsa: sja1105: Add a state machine for RX timestamping")
Signed-off-by: Vladimir Oltean <olteanv@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The IS_ERR_OR_NULL(priv->clock) check inside
sja1105_ptp_clock_unregister() is preventing cancel_delayed_work_sync
from actually being run.
Additionally, sja1105_ptp_clock_unregister() does not actually get run,
when placed in sja1105_remove(). The DSA switch gets torn down, but the
sja1105 module does not get unregistered. So sja1105_ptp_clock_unregister
needs to be moved to sja1105_teardown, to be symmetrical with
sja1105_ptp_clock_register which is called from the DSA sja1105_setup.
It is strange to fix a "fixes" patch, but the probe failure can only be
seen when the attached PHY does not respond to MDIO (issue which I can't
pinpoint the reason to) and it goes away after I power-cycle the board.
This time the patch was validated on a failing board, and the kernel
panic from the fixed commit's message can no longer be seen.
Fixes: 29dd908d35 ("net: dsa: sja1105: Cancel PTP delayed work on unregister")
Signed-off-by: Vladimir Oltean <olteanv@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
It looks like the FDB dump taken from first-generation switches also
contains information on whether entries are static or not. So use that
instead of searching through the driver's tables.
Fixes: d763778224 ("net: dsa: sja1105: Implement is_static for FDB entries on E/T")
Signed-off-by: Vladimir Oltean <olteanv@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
