There are many cases where this feature does not improve performance or even
reduces it.
For example, here are the results from tests that I've run using 3.12.6 on one
Intel Xeon W3565 and one i7 920 connected by ixgbe adapters. The results are
from the Xeon, but they're similar on the i7. All numbers report the
mean±stddev over 10 runs of 10s.
1) latency tests similar to what is described in "c6e1a0d net: Allow no-cache
copy from user on transmit"
There is no statistically significant difference between tx-nocache-copy
on/off.
nic irqs spread out (one queue per cpu)
200x netperf -r 1400,1
tx-nocache-copy off
692000±1000 tps
50/90/95/99% latency (us): 275±2/643.8±0.4/799±1/2474.4±0.3
tx-nocache-copy on
693000±1000 tps
50/90/95/99% latency (us): 274±1/644.1±0.7/800±2/2474.5±0.7
200x netperf -r 14000,14000
tx-nocache-copy off
86450±80 tps
50/90/95/99% latency (us): 334.37±0.02/838±1/2100±20/3990±40
tx-nocache-copy on
86110±60 tps
50/90/95/99% latency (us): 334.28±0.01/837±2/2110±20/3990±20
2) single stream throughput tests
tx-nocache-copy leads to higher service demand
throughput cpu0 cpu1 demand
(Gb/s) (Gcycle) (Gcycle) (cycle/B)
nic irqs and netperf on cpu0 (1x netperf -T0,0 -t omni -- -d send)
tx-nocache-copy off 9402±5 9.4±0.2 0.80±0.01
tx-nocache-copy on 9403±3 9.85±0.04 0.838±0.004
nic irqs on cpu0, netperf on cpu1 (1x netperf -T1,1 -t omni -- -d send)
tx-nocache-copy off 9401±5 5.83±0.03 5.0±0.1 0.923±0.007
tx-nocache-copy on 9404±2 5.74±0.03 5.523±0.009 0.958±0.002
As a second example, here are some results from Eric Dumazet with latest
net-next.
tx-nocache-copy also leads to higher service demand
(cpu is Intel(R) Xeon(R) CPU X5660 @ 2.80GHz)
lpq83:~# ./ethtool -K eth0 tx-nocache-copy on
lpq83:~# perf stat ./netperf -H lpq84 -c
MIGRATED TCP STREAM TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to lpq84.prod.google.com () port 0 AF_INET
Recv Send Send Utilization Service Demand
Socket Socket Message Elapsed Send Recv Send Recv
Size Size Size Time Throughput local remote local remote
bytes bytes bytes secs. 10^6bits/s % S % U us/KB us/KB
87380 16384 16384 10.00 9407.44 2.50 -1.00 0.522 -1.000
Performance counter stats for './netperf -H lpq84 -c':
4282.648396 task-clock # 0.423 CPUs utilized
9,348 context-switches # 0.002 M/sec
88 CPU-migrations # 0.021 K/sec
355 page-faults # 0.083 K/sec
11,812,797,651 cycles # 2.758 GHz [82.79%]
9,020,522,817 stalled-cycles-frontend # 76.36% frontend cycles idle [82.54%]
4,579,889,681 stalled-cycles-backend # 38.77% backend cycles idle [67.33%]
6,053,172,792 instructions # 0.51 insns per cycle
# 1.49 stalled cycles per insn [83.64%]
597,275,583 branches # 139.464 M/sec [83.70%]
8,960,541 branch-misses # 1.50% of all branches [83.65%]
10.128990264 seconds time elapsed
lpq83:~# ./ethtool -K eth0 tx-nocache-copy off
lpq83:~# perf stat ./netperf -H lpq84 -c
MIGRATED TCP STREAM TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to lpq84.prod.google.com () port 0 AF_INET
Recv Send Send Utilization Service Demand
Socket Socket Message Elapsed Send Recv Send Recv
Size Size Size Time Throughput local remote local remote
bytes bytes bytes secs. 10^6bits/s % S % U us/KB us/KB
87380 16384 16384 10.00 9412.45 2.15 -1.00 0.449 -1.000
Performance counter stats for './netperf -H lpq84 -c':
2847.375441 task-clock # 0.281 CPUs utilized
11,632 context-switches # 0.004 M/sec
49 CPU-migrations # 0.017 K/sec
354 page-faults # 0.124 K/sec
7,646,889,749 cycles # 2.686 GHz [83.34%]
6,115,050,032 stalled-cycles-frontend # 79.97% frontend cycles idle [83.31%]
1,726,460,071 stalled-cycles-backend # 22.58% backend cycles idle [66.55%]
2,079,702,453 instructions # 0.27 insns per cycle
# 2.94 stalled cycles per insn [83.22%]
363,773,213 branches # 127.757 M/sec [83.29%]
4,242,732 branch-misses # 1.17% of all branches [83.51%]
10.128449949 seconds time elapsed
CC: Tom Herbert <therbert@google.com>
Signed-off-by: Benjamin Poirier <bpoirier@suse.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
Conflicts:
drivers/net/ethernet/qlogic/qlcnic/qlcnic_sriov_pf.c
net/ipv6/ip6_tunnel.c
net/ipv6/ip6_vti.c
ipv6 tunnel statistic bug fixes conflicting with consolidation into
generic sw per-cpu net stats.
qlogic conflict between queue counting bug fix and the addition
of multiple MAC address support.
Signed-off-by: David S. Miller <davem@davemloft.net>
Pablo Neira Ayuso says:
====================
netfilter/IPVS updates for net-next
The following patchset contains Netfilter updates for your net-next tree,
they are:
* Add full port randomization support. Some crazy researchers found a way
to reconstruct the secure ephemeral ports that are allocated in random mode
by sending off-path bursts of UDP packets to overrun the socket buffer of
the DNS resolver to trigger retransmissions, then if the timing for the
DNS resolution done by a client is larger than usual, then they conclude
that the port that received the burst of UDP packets is the one that was
opened. It seems a bit aggressive method to me but it seems to work for
them. As a result, Daniel Borkmann and Hannes Frederic Sowa came up with a
new NAT mode to fully randomize ports using prandom.
* Add a new classifier to x_tables based on the socket net_cls set via
cgroups. These includes two patches to prepare the field as requested by
Zefan Li. Also from Daniel Borkmann.
* Use prandom instead of get_random_bytes in several locations of the
netfilter code, from Florian Westphal.
* Allow to use the CTA_MARK_MASK in ctnetlink when mangling the conntrack
mark, also from Florian Westphal.
* Fix compilation warning due to unused variable in IPVS, from Geert
Uytterhoeven.
* Add support for UID/GID via nfnetlink_queue, from Valentina Giusti.
* Add IPComp extension to x_tables, from Fan Du.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
While we're at it and introduced CGROUP_NET_CLASSID, lets also make
NETPRIO_CGROUP more consistent with the rest of cgroups and rename it
into CONFIG_CGROUP_NET_PRIO so that for networking, we now have
CONFIG_CGROUP_NET_{PRIO,CLASSID}. This not only makes the CONFIG
option consistent among networking cgroups, but also among cgroups
CONFIG conventions in general as the vast majority has a prefix of
CONFIG_CGROUP_<SUBSYS>.
Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
Cc: Zefan Li <lizefan@huawei.com>
Cc: cgroups@vger.kernel.org
Acked-by: Li Zefan <lizefan@huawei.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
The following functions are not used outside of net/core/dev.c
and should be declared static.
call_netdevice_notifiers_info
__dev_remove_offload
netdev_has_any_upper_dev
__netdev_adjacent_dev_remove
__netdev_adjacent_dev_link_lists
__netdev_adjacent_dev_unlink_lists
__netdev_adjacent_dev_unlink
__netdev_adjacent_dev_link_neighbour
__netdev_adjacent_dev_unlink_neighbour
And the following are never used and should be deleted
netdev_lower_dev_get_private_rcu
__netdev_find_adj_rcu
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
We can use kfree_skb_list() instead of open coding it.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
It is useful to be able to walk all upper devices when bringing
a device online where the RTNL lock is held. In this case it
is safe to walk the all_adj_list because the RTNL lock is used
to protect the write side as well.
This patch adds a check to see if the rtnl lock is held before
throwing a warning in netdev_all_upper_get_next_dev_rcu().
Also because we now have a call site for lockdep_rtnl_is_held()
outside COFIG_LOCK_PROVING an inline definition returning 1 is
needed. Similar to the rcu_read_lock_is_held().
Fixes: 2a47fa45d4 ("ixgbe: enable l2 forwarding acceleration for macvlans")
CC: Veaceslav Falico <vfalico@redhat.com>
Reported-by: Yuanhan Liu <yuanhan.liu@linux.intel.com>
Signed-off-by: John Fastabend <john.r.fastabend@intel.com>
Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Changing name of function as part of making the hash in skbuff to be
generic property, not just for receive path.
Signed-off-by: Tom Herbert <therbert@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The bond_first_slave_rcu() will be used to instead of bond_first_slave()
in rcu_read_lock().
According to the Jay Vosburgh's suggestion, the struct netdev_adjacent
should hide from users who wanted to use it directly. so I package a
new function to get the first slave of the bond.
Suggested-by: Nikolay Aleksandrov <nikolay@redhat.com>
Suggested-by: Jay Vosburgh <fubar@us.ibm.com>
Suggested-by: Veaceslav Falico <vfalico@redhat.com>
Signed-off-by: Ding Tianhong <dingtianhong@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch modifies the GRO stack to avoid the use of "network_header"
and associated macros like ip_hdr() and ipv6_hdr() in order to allow
an arbitary number of IP hdrs (v4 or v6) to be used in the
encapsulation chain. This lays the foundation for various IP
tunneling support (IP-in-IP, GRE, VXLAN, SIT,...) to be added later.
With this patch, the GRO stack traversing now is mostly based on
skb_gro_offset rather than special hdr offsets saved in skb (e.g.,
skb->network_header). As a result all but the top layer (i.e., the
the transport layer) must have hdrs of the same length in order for
a pkt to be considered for aggregation. Therefore when adding a new
encap layer (e.g., for tunneling), one must check and skip flows
(e.g., by setting NAPI_GRO_CB(p)->same_flow to 0) that have a
different hdr length.
Note that unlike the network header, the transport header can and
will continue to be set by the GRO code since there will be at
most one "transport layer" in the encap chain.
Signed-off-by: H.K. Jerry Chu <hkchu@google.com>
Suggested-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
As we need it elsewhere, move the inline helper function of
skb_needs_linearize() over to skbuff.h include file. While
at it, also convert the return to 'bool' instead of 'int'
and add a proper kernel doc.
Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Some network drivers use dev_kfree_skb_any() and dev_kfree_skb_irq()
helpers to free skbs, both for dropped packets and TX completed ones.
We need to separate the two causes to get better diagnostics
given by dropwatch or "perf record -e skb:kfree_skb"
This patch provides two new helpers, dev_consume_skb_any() and
dev_consume_skb_irq() to be used for consumed skbs.
__dev_kfree_skb_irq() is slightly optimized to remove one
atomic_dec_and_test() in fast path, and use this_cpu_{r|w} accessors.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Remove one useless conditional branch :
napi->skb is NULL, so nothing bad can happen.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The following commit:
b6c40d68ff
net: only invoke dev->change_rx_flags when device is UP
tried to fix a problem with VLAN devices and promiscuouse flag setting.
The issue was that VLAN device was setting a flag on an interface that
was down, thus resulting in bad promiscuity count.
This commit blocked flag propagation to any device that is currently
down.
A later commit:
deede2fabe
vlan: Don't propagate flag changes on down interfaces
fixed VLAN code to only propagate flags when the VLAN interface is up,
thus fixing the same issue as above, only localized to VLAN.
The problem we have now is that if we have create a complex stack
involving multiple software devices like bridges, bonds, and vlans,
then it is possible that the flags would not propagate properly to
the physical devices. A simple examle of the scenario is the
following:
eth0----> bond0 ----> bridge0 ---> vlan50
If bond0 or eth0 happen to be down at the time bond0 is added to
the bridge, then eth0 will never have promisc mode set which is
currently required for operation as part of the bridge. As a
result, packets with vlan50 will be dropped by the interface.
The only 2 devices that implement the special flag handling are
VLAN and DSA and they both have required code to prevent incorrect
flag propagation. As a result we can remove the generic solution
introduced in b6c40d68ff and leave
it to the individual devices to decide whether they will block
flag propagation or not.
Reported-by: Stefan Priebe <s.priebe@profihost.ag>
Suggested-by: Veaceslav Falico <vfalico@redhat.com>
Signed-off-by: Vlad Yasevich <vyasevic@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
A macvlan device has always LRO disabled so that calling
dev_disable_lro() on it does nothing. If we need to disable LRO
e.g. because
- the macvlan device is inserted into a bridge
- IPv6 forwarding is enabled for it
- it is in a different namespace than lowerdev and IPv4
forwarding is enabled in it
we need to disable LRO on its underlying device instead (as we
do for 802.1q VLAN devices).
v2: use newly introduced netif_is_macvlan()
Signed-off-by: Michal Kubecek <mkubecek@suse.cz>
Signed-off-by: David S. Miller <davem@davemloft.net>
commit 06a23fe31c
("core/dev: set pkt_type after eth_type_trans() in dev_forward_skb()")
and refactoring 64261f230a
("dev: move skb_scrub_packet() after eth_type_trans()")
are forcing pkt_type to be PACKET_HOST when skb traverses veth.
which means that ip forwarding will kick in inside netns
even if skb->eth->h_dest != dev->dev_addr
Fix order of eth_type_trans() and skb_scrub_packet() in dev_forward_skb()
and in ip_tunnel_rcv()
Fixes: 06a23fe31c ("core/dev: set pkt_type after eth_type_trans() in dev_forward_skb()")
CC: Isaku Yamahata <yamahatanetdev@gmail.com>
CC: Maciej Zenczykowski <zenczykowski@gmail.com>
CC: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Signed-off-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add a operations structure that allows a network interface to export
the fact that it supports package forwarding in hardware between
physical interfaces and other mac layer devices assigned to it (such
as macvlans). This operaions structure can be used by virtual mac
devices to bypass software switching so that forwarding can be done
in hardware more efficiently.
Signed-off-by: John Fastabend <john.r.fastabend@intel.com>
Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
CC: Andy Gospodarek <andy@greyhouse.net>
CC: "David S. Miller" <davem@davemloft.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Joby Poriyath provided a xen-netback patch to reduce the size of
xenvif structure as some netdev allocation could fail under
memory pressure/fragmentation.
This patch is handling the problem at the core level, allowing
any netdev structures to use vmalloc() if kmalloc() failed.
As vmalloc() adds overhead on a critical network path, add __GFP_REPEAT
to kzalloc() flags to do this fallback only when really needed.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Joby Poriyath <joby.poriyath@citrix.com>
Cc: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In order to support GSO on IPIP, we need to make
inet_gso_segment() stackable.
It should not assume network header starts right after mac
header.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Conflicts:
include/linux/netdevice.h
net/core/sock.c
Trivial merge issues.
Removal of "extern" for functions declaration in netdevice.h
at the same time "const" was added to an argument.
Two parallel line additions in net/core/sock.c
Signed-off-by: David S. Miller <davem@davemloft.net>
Separate the unreg_list and the close_list in dev_close_many preventing
dev_close_many from permuting the unreg_list. The permutations of the
unreg_list have resulted in cases where the loopback device is accessed
it has been freed in code such as dst_ifdown. Resulting in subtle memory
corruption.
This is the second bug from sharing the storage between the close_list
and the unreg_list. The issues that crop up with sharing are
apparently too subtle to show up in normal testing or usage, so let's
forget about being clever and use two separate lists.
v2: Make all callers pass in a close_list to dev_close_many
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
virtio wants to pass in cpumask_of(cpu), make parameter
const to avoid build warnings.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Benjamin Poirier
David S. Miller
Daniel Borkmann
stephen hemminger
Zhi Yong Wu
Eric Dumazet
John Fastabend
Tom Herbert
dingtianhong
Jerry Chu
Vlad Yasevich
Michal Kubeček
Alexei Starovoitov
Nikolay Aleksandrov
Eric W. Biederman
Michael S. Tsirkin