Commit 547669d483 ("tcp: xps: fix reordering issues") added
unexpected reorders in case netem is used in a MQ setup for high
performance test bed.
ETH=eth0
tc qd del dev $ETH root 2>/dev/null
tc qd add dev $ETH root handle 1: mq
for i in `seq 1 32`
do
tc qd add dev $ETH parent 1:$i netem delay 100ms
done
As all tcp packets are orphaned by netem, TCP stack believes it can
set skb->ooo_okay on all packets.
In order to allow producers to send more packets, we want to
keep sk_wmem_alloc from reaching sk_sndbuf limit.
We can do that by accounting one byte per skb in netem queues,
so that TCP stack is not fooled too much.
Tested:
With above MQ/netem setup, scaling number of concurrent flows gives
linear results and no reorders/retransmits
lpq83:~# for n in 1 10 20 30 40 50 60 70 80 90 100
do echo -n "n:$n " ; ./super_netperf $n -H 10.7.7.84; done
n:1 198.46
n:10 2002.69
n:20 4000.98
n:30 6006.35
n:40 8020.93
n:50 10032.3
n:60 12081.9
n:70 13971.3
n:80 16009.7
n:90 17117.3
n:100 17425.5
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Current net name space has only one genid for both IPv4 and IPv6, it has below
drawbacks:
- Add/delete an IPv4 address will invalidate all IPv6 routing table entries.
- Insert/remove XFRM policy will also invalidate both IPv4/IPv6 routing table
entries even when the policy is only applied for one address family.
Thus, this patch attempt to split one genid for two to cater for IPv4 and IPv6
separately in a fine granularity.
Signed-off-by: Fan Du <fan.du@windriver.com>
Acked-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Remove declaration, 4 defines and confusing comment that are no longer used
since 1a2c6181c4 ("tcp: Remove TCPCT").
Signed-off-by: Dmitry Popov <dp@highloadlab.com>
Acked-by: Christoph Paasch <christoph.paasch@uclouvain.be>
Signed-off-by: David S. Miller <davem@davemloft.net>
"ifa->ifa_label" is an array inside the in_ifaddr struct. It can never
be NULL so we can remove this check.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: Jiri Pirko <jiri@resnulli.us>
Signed-off-by: David S. Miller <davem@davemloft.net>
Explicitly set proto to ETH_P_IP and jump directly to ip processing.
Signed-off-by: Tom Herbert <therbert@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Acked-by: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: Narendra K <narendra_k@dell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Acked-by: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: Narendra K <narendra_k@dell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch adds a ndo for getting physical port of the device. Driver
which is aware of being virtual function of some physical port should
implement this ndo. This is applicable not only for IOV, but for other
solutions (NPAR, multichannel) as well. Basically if there is possible
to have multiple netdevs on the single hw port.
Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Acked-by: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Unlike for IPv6, the IPv4 checksum functions are only available
if CONFIG_INET is set.
Reported-by: kbuild test robot <fengguang.wu@intel.com>
Signed-off-by: Thomas Graf <tgraf@suug.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
| If you want to test which effects syncookies have to your
| network connections you can set this knob to 2 to enable
| unconditionally generation of syncookies.
Original idea and first implementation by Eric Dumazet.
Cc: Florian Westphal <fw@strlen.de>
Cc: David Miller <davem@davemloft.net>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch reverts commit
80b45261a0
which was implementing a 'cancelled' functionality to notify that
a cancelled request will not be replied.
This implementation was not used anywhere and therefore removed.
Signed-off-by: Andi Shyti <andi@etezian.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Fixes this on PowerPC (at least):
net/core/pktgen.c: In function 'fill_packet_ipv6':
net/core/pktgen.c:2906:3: error: implicit declaration of function 'csum_ipv6_magic' [-Werror=implicit-function-declaration]
udph->check = ~csum_ipv6_magic(&iph->saddr, &iph->daddr, udplen, IPPROTO_UDP, 0);
^
Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
v2:
a) Also send ipv4 igmp messages with TC_PRIO_CONTROL
Cc: William Manley <william.manley@youview.com>
Cc: Lukas Tribus <luky-37@hotmail.com>
Acked-by: Benjamin LaHaise <bcrl@kvack.org>
Signed-off-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
UDP checksums are optional, hence pktgen has been omitting them in
favour of performance. The optional flag UDPCSUM enables UDP
checksumming. If the output device supports hardware checksumming
the skb is prepared and marked CHECKSUM_PARTIAL, otherwise the
checksum is generated in software.
Signed-off-by: Thomas Graf <tgraf@suug.ch>
Cc: Eric Dumazet <edumazet@google.com>
Cc: Ben Greear <greearb@candelatech.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This is useful for other VSOCK transport implemented outside the
net/vmw_vsock/ directory to use these headers.
Signed-off-by: Asias He <asias@redhat.com>
Acked-by: Andy King <acking@vmware.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch consolidates the SCTP checksum calculation code from various
places to a single new function, sctp_compute_cksum(skb, offset).
Signed-off-by: Joe Stringer <joe@wand.net.nz>
Reviewed-by: Julian Anastasov <ja@ssi.bg>
Acked-by: Simon Horman <horms@verge.net.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
This started out with fixing a sparse warning, then I realized that
the wrapper function br_netpoll_info could just be collapsed away
by rolling it into the enable code.
Also, eliminate unnecessary goto's
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Reviewed-by: Jiri Pirko <jiri@resnulli.us>
Acked-by: Neil Horman <nhorman@tuxdriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
No users outside net/core/dev.c.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Idea of this patch is to add optional limitation of number of
unsent bytes in TCP sockets, to reduce usage of kernel memory.
TCP receiver might announce a big window, and TCP sender autotuning
might allow a large amount of bytes in write queue, but this has little
performance impact if a large part of this buffering is wasted :
Write queue needs to be large only to deal with large BDP, not
necessarily to cope with scheduling delays (incoming ACKS make room
for the application to queue more bytes)
For most workloads, using a value of 128 KB or less is OK to give
applications enough time to react to POLLOUT events in time
(or being awaken in a blocking sendmsg())
This patch adds two ways to set the limit :
1) Per socket option TCP_NOTSENT_LOWAT
2) A sysctl (/proc/sys/net/ipv4/tcp_notsent_lowat) for sockets
not using TCP_NOTSENT_LOWAT socket option (or setting a zero value)
Default value being UINT_MAX (0xFFFFFFFF), meaning this has no effect.
This changes poll()/select()/epoll() to report POLLOUT
only if number of unsent bytes is below tp->nosent_lowat
Note this might increase number of sendmsg()/sendfile() calls
when using non blocking sockets,
and increase number of context switches for blocking sockets.
Note this is not related to SO_SNDLOWAT (as SO_SNDLOWAT is
defined as :
Specify the minimum number of bytes in the buffer until
the socket layer will pass the data to the protocol)
Tested:
netperf sessions, and watching /proc/net/protocols "memory" column for TCP
With 200 concurrent netperf -t TCP_STREAM sessions, amount of kernel memory
used by TCP buffers shrinks by ~55 % (20567 pages instead of 45458)
lpq83:~# echo -1 >/proc/sys/net/ipv4/tcp_notsent_lowat
lpq83:~# (super_netperf 200 -t TCP_STREAM -H remote -l 90 &); sleep 60 ; grep TCP /proc/net/protocols
TCPv6 1880 2 45458 no 208 yes ipv6 y y y y y y y y y y y y y n y y y y y
TCP 1696 508 45458 no 208 yes kernel y y y y y y y y y y y y y n y y y y y
lpq83:~# echo 131072 >/proc/sys/net/ipv4/tcp_notsent_lowat
lpq83:~# (super_netperf 200 -t TCP_STREAM -H remote -l 90 &); sleep 60 ; grep TCP /proc/net/protocols
TCPv6 1880 2 20567 no 208 yes ipv6 y y y y y y y y y y y y y n y y y y y
TCP 1696 508 20567 no 208 yes kernel y y y y y y y y y y y y y n y y y y y
Using 128KB has no bad effect on the throughput or cpu usage
of a single flow, although there is an increase of context switches.
A bonus is that we hold socket lock for a shorter amount
of time and should improve latencies of ACK processing.
lpq83:~# echo -1 >/proc/sys/net/ipv4/tcp_notsent_lowat
lpq83:~# perf stat -e context-switches ./netperf -H 7.7.7.84 -t omni -l 20 -c -i10,3
OMNI Send TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 7.7.7.84 () port 0 AF_INET : +/-2.500% @ 99% conf.
Local Remote Local Elapsed Throughput Throughput Local Local Remote Remote Local Remote Service
Send Socket Recv Socket Send Time Units CPU CPU CPU CPU Service Service Demand
Size Size Size (sec) Util Util Util Util Demand Demand Units
Final Final % Method % Method
1651584 6291456 16384 20.00 17447.90 10^6bits/s 3.13 S -1.00 U 0.353 -1.000 usec/KB
Performance counter stats for './netperf -H 7.7.7.84 -t omni -l 20 -c -i10,3':
412,514 context-switches
200.034645535 seconds time elapsed
lpq83:~# echo 131072 >/proc/sys/net/ipv4/tcp_notsent_lowat
lpq83:~# perf stat -e context-switches ./netperf -H 7.7.7.84 -t omni -l 20 -c -i10,3
OMNI Send TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 7.7.7.84 () port 0 AF_INET : +/-2.500% @ 99% conf.
Local Remote Local Elapsed Throughput Throughput Local Local Remote Remote Local Remote Service
Send Socket Recv Socket Send Time Units CPU CPU CPU CPU Service Service Demand
Size Size Size (sec) Util Util Util Util Demand Demand Units
Final Final % Method % Method
1593240 6291456 16384 20.00 17321.16 10^6bits/s 3.35 S -1.00 U 0.381 -1.000 usec/KB
Performance counter stats for './netperf -H 7.7.7.84 -t omni -l 20 -c -i10,3':
2,675,818 context-switches
200.029651391 seconds time elapsed
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Neal Cardwell <ncardwell@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Acked-By: Yuchung Cheng <ycheng@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Several call sites use the hardcoded following condition :
sk_stream_wspace(sk) >= sk_stream_min_wspace(sk)
Lets use a helper because TCP_NOTSENT_LOWAT support will change this
condition for TCP sockets.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Neal Cardwell <ncardwell@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The SCTP mailing list address to send patches or questions
to is linux-sctp@vger.kernel.org and not
lksctp-developers@lists.sourceforge.net anymore. Therefore,
update all occurences.
Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
Acked-by: Neil Horman <nhorman@tuxdriver.com>
Acked-by: Vlad Yasevich <vyasevich@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch gets rid of the following warning:
net/9p/trans_rdma.c:594:12: warning: ‘rdma_cancelled’ defined but not used [-Wunused-function]
static int rdma_cancelled(struct p9_client *client, struct p9_req_t *req)
The rdma_cancelled function is not called anywhere in the kernel
Signed-off-by: Andi Shyti <andi@etezian.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
The "int addrlen" in fib6_add_1 is rebundant, as we can get it from
parameter "struct in6_addr *addr" once we modified its type.
And also fix some coding style issues in fib6_add_1
Signed-off-by: Fan Du <fan.du@windriver.com>
Acked-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Seting rt->rt6i_nsiblings to zero is rebundant, because above memset
zeroed the rest of rt excluding the first dst memember.
Signed-off-by: Fan Du <fan.du@windriver.com>
Acked-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Acked-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Cc: Hideaki YOSHIFUJI <yoshfuji@linux-ipv6.org>
Cc: James Morris <jmorris@namei.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
GRE tunnel and IPIP tunnel already switched to the new
ip tunnel code, VTI tunnel can use it too.
Cc: Pravin B Shelar <pshelar@nicira.com>
Cc: Stephen Hemminger <stephen@networkplumber.org>
Cc: Saurabh Mohan <saurabh.mohan@vyatta.com>
Cc: "David S. Miller" <davem@davemloft.net>
Signed-off-by: Cong Wang <amwang@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch changes the prototpye of the ip6_mr_forward() method to return void
instead of int.
The ip6_mr_forward() method always returns 0; moreover, the return value of this
method is not checked anywhere.
Signed-off-by: Rami Rosen <ramirose@gmail.com>
Acked-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch changes the prototpye of the ip_mr_forward() method to return void
instead of int.
The ip_mr_forward() method always returns 0; moreover, the return value of this
method is not checked anywhere.
Signed-off-by: Rami Rosen <ramirose@gmail.com>
Acked-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Until now, bond_resend_igmp_join_requests() looks for vlans attached to
bonding device, bridge where bonding act as port manually. It does not
care of other scenarios, like stacked bonds or team device above. Make
this more generic and use netdev notifier to propagate the event to
upper devices and to actually call ip_mc_rejoin_groups().
Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Acked-by: Veaceslav Falico <vfalico@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
If RTT is not available because Karn's check has failed or no
new packet is acked, use the RTT measured from SACK to estimate
the RTO. The sender can continue to estimate the RTO during loss
recovery or reordering event upon receiving non-partial ACKs.
This also changes when the RTO is re-armed. Previously it is
only re-armed when some data is cummulatively acknowledged (i.e.,
SND.UNA advances), but now it is re-armed whenever RTT estimator
is updated. This feature is particularly useful to reduce spurious
timeout for buffer bloat including cellular carriers [1], and
RTT estimation on reordering events.
[1] "An In-depth Study of LTE: Effect of Network Protocol and
Application Behavior on Performance", In Proc. of SIGCOMM 2013
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Take RTT sample if an ACK selectively acks some sequences that
have never been retransmitted. The Karn's algorithm does not apply
even if that ACK (s)acks other retransmitted sequences, because it
must been generated by an original but perhaps out-of-order packet.
There is no ambiguity. In case when multiple blocks are newly
sacked because of ACK losses the earliest block is used to
measure RTT, similar to cummulative ACKs.
Such RTT samples allow the sender to estimate the RTO during loss
recovery and packet reordering events. It is still useful even with
TCP timestamps. That's because during these events the SND.UNA may
not advance preventing RTT samples from TS ECR (thus the FLAG_ACKED
check before calling tcp_ack_update_rtt()). Therefore this new
RTT source is complementary to existing ACK and TS RTT mechanisms.
This patch does not update the RTO. It is done in the next patch.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Prefer packet timings to TS-ecr for RTT measurements when both
sources are available. That's because broken middle-boxes and remote
peer can return packets with corrupted TS ECR fields. Similarly most
congestion controls that require RTT signals favor timing-based
sources as well. Also check for bad TS ECR values to avoid RTT
blow-ups. It has happened on production Web servers.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The first patch consolidates SYNACK and other RTT measurement to use a
central function tcp_ack_update_rtt(). A (small) bonus is now SYNACK
RTT measurement happens after PAWS check, potentially reducing the
impact of RTO seeding on bad TCP timestamps values.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch moves the private error queue delivery function from the
af_packet code to the core socket method. In this way, network layers
only needing the error queue for transmit time stamping can share common
code.
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Applied error fixes suggested by checpatch.pl
Signed-off-by: Dragos Foianu <dragos.foianu@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Pull networking fixes from David Miller:
"A couple interesting SKB fragment handling fixes, plus the usual small
bits here and there:
1) Fix 64-bit divide build failure on 32-bit platforms in mlx5, from
Tim Gardner.
2) Get rid of a stupid reimplementation on "%*phC" in our sysfs MAC
address printing helper.
3) Fix NETIF_F_SG capability advertisement in hyperv driver, if the
device can't do checksumming offloads then it shouldn't say it can
do SG either. From Haiyang Zhang.
4) bgmac needs to depend on PHYLIB, from Hauke Mehrtens.
5) Don't leak DMA mappings on mapping failures, from Neil Horman.
6) We need to reset the transport header of SKBs in ipv4 before we
attempt to perform early socket demux, just like ipv6 does. From
Eric Dumazet.
7) Add missing locking on vxlan device removal, from Stephen
Hemminger.
8) xen-netfront has to make two passes over an SKB to prepare it for
transfer. One pass calculates the number of slots needed, the
second massages the SKB and fills the slots. Unfortunately, the
first pass doesn't calculate the number of slots properly so we
can end up trying to build a MAX_SKB_FRAGS + 1 SKB which doesn't
work out so well. Fix from Jan Beulich with help and discussion
with several others.
9) Fix a similar problem in tun and macvtap, which have to split up
scatter-gather elements at PAGE_SIZE boundaries. Don't do
zerocopy if it would result in a > MAX_SKB_FRAGS skb. Fixes from
Jason Wang.
10) On receive, once we've decoded the VLAN state completely, clear
skb->vlan_tci. Otherwise demuxed tunnels underneath can trigger
the VLAN code again, corrupting the packet. Fix from Eric
Dumazet"
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net:
vlan: fix a race in egress prio management
vlan: mask vlan prio bits
macvtap: do not zerocopy if iov needs more pages than MAX_SKB_FRAGS
tuntap: do not zerocopy if iov needs more pages than MAX_SKB_FRAGS
pkt_sched: sch_qfq: remove a source of high packet delay/jitter
xen-netfront: pull on receive skb may need to happen earlier
vxlan: add necessary locking on device removal
hyperv: Fix the NETIF_F_SG flag setting in netvsc
net: Fix sysfs_format_mac() code duplication.
be2net: Fix to avoid hardware workaround when not needed
macvtap: do not assume 802.1Q when send vlan packets
macvtap: fix the missing ret value of TUNSETQUEUE
ipv4: set transport header earlier
mlx5 core: Fix __udivdi3 when compiling for 32 bit arches
bgmac: add dependency to phylib
net/irda: fixed style issues in irlan_eth
ethtool: fixed trailing statements in ethtool
ndisc: bool initializations should use true and false
atl1e: unmap partially mapped skb on dma error and free skb
egress_priority_map[] hash table updates are protected by rtnl,
and we never remove elements until device is dismantled.
We have to make sure that before inserting an new element in hash table,
all its fields are committed to memory or else another cpu could
find corrupt values and crash.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
In commit 48cc32d38a
("vlan: don't deliver frames for unknown vlans to protocols")
Florian made sure we set pkt_type to PACKET_OTHERHOST
if the vlan id is set and we could find a vlan device for this
particular id.
But we also have a problem if prio bits are set.
Steinar reported an issue on a router receiving IPv6 frames with a
vlan tag of 4000 (id 0, prio 2), and tunneled into a sit device,
because skb->vlan_tci is set.
Forwarded frame is completely corrupted : We can see (8100:4000)
being inserted in the middle of IPv6 source address :
16:48:00.780413 IP6 2001:16d8:8100:4000:ee1c:0:9d9:bc87 >
9f94:4d95:2001:67c:29f4::: ICMP6, unknown icmp6 type (0), length 64
0x0000: 0000 0029 8000 c7c3 7103 0001 a0ae e651
0x0010: 0000 0000 ccce 0b00 0000 0000 1011 1213
0x0020: 1415 1617 1819 1a1b 1c1d 1e1f 2021 2223
0x0030: 2425 2627 2829 2a2b 2c2d 2e2f 3031 3233
It seems we are not really ready to properly cope with this right now.
We can probably do better in future kernels :
vlan_get_ingress_priority() should be a netdev property instead of
a per vlan_dev one.
For stable kernels, lets clear vlan_tci to fix the bugs.
Reported-by: Steinar H. Gunderson <sesse@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
QFQ+ inherits from QFQ a design choice that may cause a high packet
delay/jitter and a severe short-term unfairness. As QFQ, QFQ+ uses a
special quantity, the system virtual time, to track the service
provided by the ideal system it approximates. When a packet is
dequeued, this quantity must be incremented by the size of the packet,
divided by the sum of the weights of the aggregates waiting to be
served. Tracking this sum correctly is a non-trivial task, because, to
preserve tight service guarantees, the decrement of this sum must be
delayed in a special way [1]: this sum can be decremented only after
that its value would decrease also in the ideal system approximated by
QFQ+. For efficiency, QFQ+ keeps track only of the 'instantaneous'
weight sum, increased and decreased immediately as the weight of an
aggregate changes, and as an aggregate is created or destroyed (which,
in its turn, happens as a consequence of some class being
created/destroyed/changed). However, to avoid the problems caused to
service guarantees by these immediate decreases, QFQ+ increments the
system virtual time using the maximum value allowed for the weight
sum, 2^10, in place of the dynamic, instantaneous value. The
instantaneous value of the weight sum is used only to check whether a
request of weight increase or a class creation can be satisfied.
Unfortunately, the problems caused by this choice are worse than the
temporary degradation of the service guarantees that may occur, when a
class is changed or destroyed, if the instantaneous value of the
weight sum was used to update the system virtual time. In fact, the
fraction of the link bandwidth guaranteed by QFQ+ to each aggregate is
equal to the ratio between the weight of the aggregate and the sum of
the weights of the competing aggregates. The packet delay guaranteed
to the aggregate is instead inversely proportional to the guaranteed
bandwidth. By using the maximum possible value, and not the actual
value of the weight sum, QFQ+ provides each aggregate with the worst
possible service guarantees, and not with service guarantees related
to the actual set of competing aggregates. To see the consequences of
this fact, consider the following simple example.
Suppose that only the following aggregates are backlogged, i.e., that
only the classes in the following aggregates have packets to transmit:
one aggregate with weight 10, say A, and ten aggregates with weight 1,
say B1, B2, ..., B10. In particular, suppose that these aggregates are
always backlogged. Given the weight distribution, the smoothest and
fairest service order would be:
A B1 A B2 A B3 A B4 A B5 A B6 A B7 A B8 A B9 A B10 A B1 A B2 ...
QFQ+ would provide exactly this optimal service if it used the actual
value for the weight sum instead of the maximum possible value, i.e.,
11 instead of 2^10. In contrast, since QFQ+ uses the latter value, it
serves aggregates as follows (easy to prove and to reproduce
experimentally):
A B1 B2 B3 B4 B5 B6 B7 B8 B9 B10 A A A A A A A A A A B1 B2 ... B10 A A ...
By replacing 10 with N in the above example, and by increasing N, one
can increase at will the maximum packet delay and the jitter
experienced by the classes in aggregate A.
This patch addresses this issue by just using the above
'instantaneous' value of the weight sum, instead of the maximum
possible value, when updating the system virtual time. After the
instantaneous weight sum is decreased, QFQ+ may deviate from the ideal
service for a time interval in the order of the time to serve one
maximum-size packet for each backlogged class. The worst-case extent
of the deviation exhibited by QFQ+ during this time interval [1] is
basically the same as of the deviation described above (but, without
this patch, QFQ+ suffers from such a deviation all the time). Finally,
this patch modifies the comment to the function qfq_slot_insert, to
make it coherent with the fact that the weight sum used by QFQ+ can
now be lower than the maximum possible value.
[1] P. Valente, "Extending WF2Q+ to support a dynamic traffic mix",
Proceedings of AAA-IDEA'05, June 2005.
Signed-off-by: Paolo Valente <paolo.valente@unimore.it>
Signed-off-by: David S. Miller <davem@davemloft.net>
Pull phase two of __cpuinit removal from Paul Gortmaker:
"With the __cpuinit infrastructure removed earlier, this group of
commits only removes the function/data tagging that was done with the
various (now no-op) __cpuinit related prefixes.
Now that the dust has settled with yesterday's v3.11-rc1, there
hopefully shouldn't be any new users leaking back in tree, but I think
we can leave the harmless no-op stubs there for a release as a
courtesy to those who still have out of tree stuff and weren't paying
attention.
Although the commits are against the recent tag to allow for minor
context refreshes for things like yesterday's v3.11-rc1~ slab content,
the patches have been largely unchanged for weeks, aside from such
trivial updates.
For detail junkies, the largely boring and mostly irrelevant history
of the patches can be viewed at:
http://git.kernel.org/cgit/linux/kernel/git/paulg/cpuinit-delete.git
If nothing else, I guess it does at least demonstrate the level of
involvement required to shepherd such a treewide change to completion.
This is the same repository of patches that has been applied to the
end of the daily linux-next branches for the past several weeks"
* 'cpuinit_phase2' of git://git.kernel.org/pub/scm/linux/kernel/git/paulg/linux: (28 commits)
block: delete __cpuinit usage from all block files
drivers: delete __cpuinit usage from all remaining drivers files
kernel: delete __cpuinit usage from all core kernel files
rcu: delete __cpuinit usage from all rcu files
net: delete __cpuinit usage from all net files
acpi: delete __cpuinit usage from all acpi files
hwmon: delete __cpuinit usage from all hwmon files
cpufreq: delete __cpuinit usage from all cpufreq files
clocksource+irqchip: delete __cpuinit usage from all related files
x86: delete __cpuinit usage from all x86 files
score: delete __cpuinit usage from all score files
xtensa: delete __cpuinit usage from all xtensa files
openrisc: delete __cpuinit usage from all openrisc files
m32r: delete __cpuinit usage from all m32r files
hexagon: delete __cpuinit usage from all hexagon files
frv: delete __cpuinit usage from all frv files
cris: delete __cpuinit usage from all cris files
metag: delete __cpuinit usage from all metag files
tile: delete __cpuinit usage from all tile files
sh: delete __cpuinit usage from all sh files
...
Pull nfsd bugfixes from Bruce Fields:
"Just three minor bugfixes"
* 'for-3.11' of git://linux-nfs.org/~bfields/linux:
svcrdma: underflow issue in decode_write_list()
nfsd4: fix minorversion support interface
lockd: protect nlm_blocked access in nlmsvc_retry_blocked
It's just a duplicate implementation of "%*phC". Thanks to Joe
Perches for showing that we had exactly this support in the
lib/vsprintf.c code already.
Signed-off-by: David S. Miller <davem@davemloft.net>
commit 45f00f99d6 ("ipv4: tcp: clean up tcp_v4_early_demux()") added a
performance regression for non GRO traffic, basically disabling
IP early demux.
IPv6 stack resets transport header in ip6_rcv() before calling
IP early demux in ip6_rcv_finish(), while IPv4 does this only in
ip_local_deliver_finish(), _after_ IP early demux.
GRO traffic happened to enable IP early demux because transport header
is also set in inet_gro_receive()
Instead of reverting the faulty commit, we can make IPv4/IPv6 behave the
same : transport_header should be set in ip_rcv() instead of
ip_local_deliver_finish()
ip_local_deliver_finish() can also use skb_network_header_len() which is
faster than ip_hdrlen()
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Neal Cardwell <ncardwell@google.com>
Cc: Tom Herbert <therbert@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
My static checker marks everything from ntohl() as untrusted and it
complains we could have an underflow problem doing:
return (u32 *)&ary->wc_array[nchunks];
Also on 32 bit systems the upper bound check could overflow.
Cc: stable@vger.kernel.org
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
The __cpuinit type of throwaway sections might have made sense
some time ago when RAM was more constrained, but now the savings
do not offset the cost and complications. For example, the fix in
commit 5e427ec2d0 ("x86: Fix bit corruption at CPU resume time")
is a good example of the nasty type of bugs that can be created
with improper use of the various __init prefixes.
After a discussion on LKML[1] it was decided that cpuinit should go
the way of devinit and be phased out. Once all the users are gone,
we can then finally remove the macros themselves from linux/init.h.
This removes all the net/* uses of the __cpuinit macros
from all C files.
[1] https://lkml.org/lkml/2013/5/20/589
Cc: "David S. Miller" <davem@davemloft.net>
Cc: netdev@vger.kernel.org
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Pull more vfs stuff from Al Viro:
"O_TMPFILE ABI changes, Oleg's fput() series, misc cleanups, including
making simple_lookup() usable for filesystems with non-NULL s_d_op,
which allows us to get rid of quite a bit of ugliness"
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
sunrpc: now we can just set ->s_d_op
cgroup: we can use simple_lookup() now
efivarfs: we can use simple_lookup() now
make simple_lookup() usable for filesystems that set ->s_d_op
configfs: don't open-code d_alloc_name()
__rpc_lookup_create_exclusive: pass string instead of qstr
rpc_create_*_dir: don't bother with qstr
llist: llist_add() can use llist_add_batch()
llist: fix/simplify llist_add() and llist_add_batch()
fput: turn "list_head delayed_fput_list" into llist_head
fs/file_table.c:fput(): add comment
Safer ABI for O_TMPFILE
