I had many reports that TSQ logic breaks wifi aggregation.
Current logic is to allow up to 1 ms of bytes to be queued into qdisc
and drivers queues.
But Wifi aggregation needs a bigger budget to allow bigger rates to
be discovered by various TCP Congestion Controls algorithms.
This patch adds an extra socket field, allowing wifi drivers to select
another log scale to derive TCP Small Queue credit from current pacing
rate.
Initial value is 10, meaning that this patch does not change current
behavior.
We expect wifi drivers to set this field to smaller values (tests have
been done with values from 6 to 9)
They would have to use following template :
if (skb->sk && skb->sk->sk_pacing_shift != MY_PACING_SHIFT)
skb->sk->sk_pacing_shift = MY_PACING_SHIFT;
Ref: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1670041
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Johannes Berg <johannes.berg@intel.com>
Cc: Toke Høiland-Jørgensen <toke@toke.dk>
Cc: Kir Kolyshkin <kir@openvz.org>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
As we want to gradually implement per netns sysctl_rmem and sysctl_wmem
on per protocol basis, add two new fields in struct proto,
and two new helpers : sk_get_wmem0() and sk_get_rmem0()
First user will be TCP. Then UDP and SCTP can be easily converted,
while DECNET probably wont get this support.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Mark hlist node in sk rcu iterator as protected by the rcu.
hlist_next_rcu accomplishes this and silences the warnings
sparse throws.
Found with make C=1 net/ipv4/udp.o on linux-next tag
next-20171009.
Signed-off-by: Tim Hansen <devtimhansen@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Using a linear list to store all skbs in write queue has been okay
for quite a while : O(N) is not too bad when N < 500.
Things get messy when N is the order of 100,000 : Modern TCP stacks
want 10Gbit+ of throughput even with 200 ms RTT flows.
40 ns per cache line miss means a full scan can use 4 ms,
blowing away CPU caches.
SACK processing often can use various hints to avoid parsing
whole retransmit queue. But with high packet losses and/or high
reordering, hints no longer work.
Sender has to process thousands of unfriendly SACK, accumulating
a huge socket backlog, burning a cpu and massively dropping packets.
Using an rb-tree for retransmit queue has been avoided for years
because it added complexity and overhead, but now is the time
to be more resistant and say no to quadratic behavior.
1) RTX queue is no longer part of the write queue : already sent skbs
are stored in one rb-tree.
2) Since reaching the head of write queue no longer needs
sk->sk_send_head, we added an union of sk_send_head and tcp_rtx_queue
Tested:
On receiver :
netem on ingress : delay 150ms 200us loss 1
GRO disabled to force stress and SACK storms.
for f in `seq 1 10`
do
./netperf -H lpaa6 -l30 -- -K bbr -o THROUGHPUT|tail -1
done | awk '{print $0} {sum += $0} END {printf "%7u\n",sum}'
Before patch :
323.87
351.48
339.59
338.62
306.72
204.07
304.93
291.88
202.47
176.88
2840
After patch:
1700.83
2207.98
2070.17
1544.26
2114.76
2124.89
1693.14
1080.91
2216.82
1299.94
18053
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In linux-4.13, Wei worked hard to convert dst to a traditional
refcounted model, removing GC.
We now want to make sure a dst refcount can not transition from 0 back
to 1.
The problem here is that input path attached a not refcounted dst to an
skb. Then later, because packet is forwarded and hits skb_dst_force()
before exiting RCU section, we might try to take a refcount on one dst
that is about to be freed, if another cpu saw 1 -> 0 transition in
dst_release() and queued the dst for freeing after one RCU grace period.
Lets unify skb_dst_force() and skb_dst_force_safe(), since we should
always perform the complete check against dst refcount, and not assume
it is not zero.
Bugzilla : https://bugzilla.kernel.org/show_bug.cgi?id=197005
[ 989.919496] skb_dst_force+0x32/0x34
[ 989.919498] __dev_queue_xmit+0x1ad/0x482
[ 989.919501] ? eth_header+0x28/0xc6
[ 989.919502] dev_queue_xmit+0xb/0xd
[ 989.919504] neigh_connected_output+0x9b/0xb4
[ 989.919507] ip_finish_output2+0x234/0x294
[ 989.919509] ? ipt_do_table+0x369/0x388
[ 989.919510] ip_finish_output+0x12c/0x13f
[ 989.919512] ip_output+0x53/0x87
[ 989.919513] ip_forward_finish+0x53/0x5a
[ 989.919515] ip_forward+0x2cb/0x3e6
[ 989.919516] ? pskb_trim_rcsum.part.9+0x4b/0x4b
[ 989.919518] ip_rcv_finish+0x2e2/0x321
[ 989.919519] ip_rcv+0x26f/0x2eb
[ 989.919522] ? vlan_do_receive+0x4f/0x289
[ 989.919523] __netif_receive_skb_core+0x467/0x50b
[ 989.919526] ? tcp_gro_receive+0x239/0x239
[ 989.919529] ? inet_gro_receive+0x226/0x238
[ 989.919530] __netif_receive_skb+0x4d/0x5f
[ 989.919532] netif_receive_skb_internal+0x5c/0xaf
[ 989.919533] napi_gro_receive+0x45/0x81
[ 989.919536] ixgbe_poll+0xc8a/0xf09
[ 989.919539] ? kmem_cache_free_bulk+0x1b6/0x1f7
[ 989.919540] net_rx_action+0xf4/0x266
[ 989.919543] __do_softirq+0xa8/0x19d
[ 989.919545] irq_exit+0x5d/0x6b
[ 989.919546] do_IRQ+0x9c/0xb5
[ 989.919548] common_interrupt+0x93/0x93
[ 989.919548] </IRQ>
Similarly dst_clone() can use dst_hold() helper to have additional
debugging, as a follow up to commit 44ebe79149 ("net: add debug
atomic_inc_not_zero() in dst_hold()")
In net-next we will convert dst atomic_t to refcount_t for peace of
mind.
Fixes: a4c2fd7f78 ("net: remove DST_NOCACHE flag")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Wei Wang <weiwan@google.com>
Reported-by: Paweł Staszewski <pstaszewski@itcare.pl>
Bisected-by: Paweł Staszewski <pstaszewski@itcare.pl>
Acked-by: Wei Wang <weiwan@google.com>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Florian reported UDP xmit drops that could be root caused to the
too small neigh limit.
Current limit is 64 KB, meaning that even a single UDP socket would hit
it, since its default sk_sndbuf comes from net.core.wmem_default
(~212992 bytes on 64bit arches).
Once ARP/ND resolution is in progress, we should allow a little more
packets to be queued, at least for one producer.
Once neigh arp_queue is filled, a rogue socket should hit its sk_sndbuf
limit and either block in sendmsg() or return -EAGAIN.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Due to commit e6afc8ace6 ("udp: remove
headers from UDP packets before queueing"), when udp packets are being
peeked the requested extra offset is always 0 as there is no need to skip
the udp header. However, when the offset is 0 and the next skb is
of length 0, it is only returned once. The behaviour can be seen with
the following python script:
from socket import *;
f=socket(AF_INET6, SOCK_DGRAM | SOCK_NONBLOCK, 0);
g=socket(AF_INET6, SOCK_DGRAM | SOCK_NONBLOCK, 0);
f.bind(('::', 0));
addr=('::1', f.getsockname()[1]);
g.sendto(b'', addr)
g.sendto(b'b', addr)
print(f.recvfrom(10, MSG_PEEK));
print(f.recvfrom(10, MSG_PEEK));
Where the expected output should be the empty string twice.
Instead, make sk_peek_offset return negative values, and pass those values
to __skb_try_recv_datagram/__skb_try_recv_from_queue. If the passed offset
to __skb_try_recv_from_queue is negative, the checked skb is never skipped.
__skb_try_recv_from_queue will then ensure the offset is reset back to 0
if a peek is requested without an offset, unless no packets are found.
Also simplify the if condition in __skb_try_recv_from_queue. If _off is
greater then 0, and off is greater then or equal to skb->len, then
(_off || skb->len) must always be true assuming skb->len >= 0 is always
true.
Also remove a redundant check around a call to sk_peek_offset in af_unix.c,
as it double checked if MSG_PEEK was set in the flags.
V2:
- Moved the negative fixup into __skb_try_recv_from_queue, and remove now
redundant checks
- Fix peeking in udp{,v6}_recvmsg to report the right value when the
offset is 0
V3:
- Marked new branch in __skb_try_recv_from_queue as unlikely.
Signed-off-by: Matthew Dawson <matthew@mjdsystems.ca>
Acked-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The kernel supports zerocopy sendmsg in virtio and tap. Expand the
infrastructure to support other socket types. Introduce a completion
notification channel over the socket error queue. Notifications are
returned with ee_origin SO_EE_ORIGIN_ZEROCOPY. ee_errno is 0 to avoid
blocking the send/recv path on receiving notifications.
Add reference counting, to support the skb split, merge, resize and
clone operations possible with SOCK_STREAM and other socket types.
The patch does not yet modify any datapaths.
Signed-off-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add sock_omalloc and sock_ofree to be able to allocate control skbs,
for instance for looping errors onto sk_error_queue.
The transmit budget (sk_wmem_alloc) is involved in transmit skb
shaping, most notably in TCP Small Queues. Using this budget for
control packets would impact transmission.
Signed-off-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add new proto_ops sendmsg_locked and sendpage_locked that can be
called when the socket lock is already held. Correspondingly, add
kernel_sendmsg_locked and kernel_sendpage_locked as front end
functions.
These functions will be used in zero proxy so that we can take
the socket lock in a ULP sendmsg/sendpage and then directly call the
backend transport proto_ops functions.
Signed-off-by: Tom Herbert <tom@quantonium.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
randstruct plugin, including the task_struct.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
Comment: Kees Cook <kees@outflux.net>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=ngAW
-----END PGP SIGNATURE-----
Merge tag 'gcc-plugins-v4.13-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux
Pull structure randomization updates from Kees Cook:
"Now that IPC and other changes have landed, enable manual markings for
randstruct plugin, including the task_struct.
This is the rest of what was staged in -next for the gcc-plugins, and
comes in three patches, largest first:
- mark "easy" structs with __randomize_layout
- mark task_struct with an optional anonymous struct to isolate the
__randomize_layout section
- mark structs to opt _out_ of automated marking (which will come
later)
And, FWIW, this continues to pass allmodconfig (normal and patched to
enable gcc-plugins) builds of x86_64, i386, arm64, arm, powerpc, and
s390 for me"
* tag 'gcc-plugins-v4.13-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux:
randstruct: opt-out externally exposed function pointer structs
task_struct: Allow randomized layout
randstruct: Mark various structs for randomization
Fill in missing kernel-doc for missing elements in struct sock.
Signed-off-by: Stephen Hemminger <sthemmin@microsoft.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
sock_graft() unilaterally sets up parent->sk based on the
assumption that the existing parent->sk is null. If this
condition is not true, then the existing parent->sk would
be leaked, so add a WARN_ON() to alert callers who may fall
in this category.
Signed-off-by: Sowmini Varadhan <sowmini.varadhan@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Pull networking updates from David Miller:
"Reasonably busy this cycle, but perhaps not as busy as in the 4.12
merge window:
1) Several optimizations for UDP processing under high load from
Paolo Abeni.
2) Support pacing internally in TCP when using the sch_fq packet
scheduler for this is not practical. From Eric Dumazet.
3) Support mutliple filter chains per qdisc, from Jiri Pirko.
4) Move to 1ms TCP timestamp clock, from Eric Dumazet.
5) Add batch dequeueing to vhost_net, from Jason Wang.
6) Flesh out more completely SCTP checksum offload support, from
Davide Caratti.
7) More plumbing of extended netlink ACKs, from David Ahern, Pablo
Neira Ayuso, and Matthias Schiffer.
8) Add devlink support to nfp driver, from Simon Horman.
9) Add RTM_F_FIB_MATCH flag to RTM_GETROUTE queries, from Roopa
Prabhu.
10) Add stack depth tracking to BPF verifier and use this information
in the various eBPF JITs. From Alexei Starovoitov.
11) Support XDP on qed device VFs, from Yuval Mintz.
12) Introduce BPF PROG ID for better introspection of installed BPF
programs. From Martin KaFai Lau.
13) Add bpf_set_hash helper for TC bpf programs, from Daniel Borkmann.
14) For loads, allow narrower accesses in bpf verifier checking, from
Yonghong Song.
15) Support MIPS in the BPF selftests and samples infrastructure, the
MIPS eBPF JIT will be merged in via the MIPS GIT tree. From David
Daney.
16) Support kernel based TLS, from Dave Watson and others.
17) Remove completely DST garbage collection, from Wei Wang.
18) Allow installing TCP MD5 rules using prefixes, from Ivan
Delalande.
19) Add XDP support to Intel i40e driver, from Björn Töpel
20) Add support for TC flower offload in nfp driver, from Simon
Horman, Pieter Jansen van Vuuren, Benjamin LaHaise, Jakub
Kicinski, and Bert van Leeuwen.
21) IPSEC offloading support in mlx5, from Ilan Tayari.
22) Add HW PTP support to macb driver, from Rafal Ozieblo.
23) Networking refcount_t conversions, From Elena Reshetova.
24) Add sock_ops support to BPF, from Lawrence Brako. This is useful
for tuning the TCP sockopt settings of a group of applications,
currently via CGROUPs"
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next: (1899 commits)
net: phy: dp83867: add workaround for incorrect RX_CTRL pin strap
dt-bindings: phy: dp83867: provide a workaround for incorrect RX_CTRL pin strap
cxgb4: Support for get_ts_info ethtool method
cxgb4: Add PTP Hardware Clock (PHC) support
cxgb4: time stamping interface for PTP
nfp: default to chained metadata prepend format
nfp: remove legacy MAC address lookup
nfp: improve order of interfaces in breakout mode
net: macb: remove extraneous return when MACB_EXT_DESC is defined
bpf: add missing break in for the TCP_BPF_SNDCWND_CLAMP case
bpf: fix return in load_bpf_file
mpls: fix rtm policy in mpls_getroute
net, ax25: convert ax25_cb.refcount from atomic_t to refcount_t
net, ax25: convert ax25_route.refcount from atomic_t to refcount_t
net, ax25: convert ax25_uid_assoc.refcount from atomic_t to refcount_t
net, sctp: convert sctp_ep_common.refcnt from atomic_t to refcount_t
net, sctp: convert sctp_transport.refcnt from atomic_t to refcount_t
net, sctp: convert sctp_chunk.refcnt from atomic_t to refcount_t
net, sctp: convert sctp_datamsg.refcnt from atomic_t to refcount_t
net, sctp: convert sctp_auth_bytes.refcnt from atomic_t to refcount_t
...
refcount_t type and corresponding API should be
used instead of atomic_t when the variable is used as
a reference counter. This allows to avoid accidental
refcounter overflows that might lead to use-after-free
situations.
This patch uses refcount_inc_not_zero() instead of
atomic_inc_not_zero_hint() due to absense of a _hint()
version of refcount API. If the hint() version must
be used, we might need to revisit API.
Signed-off-by: Elena Reshetova <elena.reshetova@intel.com>
Signed-off-by: Hans Liljestrand <ishkamiel@gmail.com>
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: David Windsor <dwindsor@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
refcount_t type and corresponding API should be
used instead of atomic_t when the variable is used as
a reference counter. This allows to avoid accidental
refcounter overflows that might lead to use-after-free
situations.
Signed-off-by: Elena Reshetova <elena.reshetova@intel.com>
Signed-off-by: Hans Liljestrand <ishkamiel@gmail.com>
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: David Windsor <dwindsor@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This marks many critical kernel structures for randomization. These are
structures that have been targeted in the past in security exploits, or
contain functions pointers, pointers to function pointer tables, lists,
workqueues, ref-counters, credentials, permissions, or are otherwise
sensitive. This initial list was extracted from Brad Spengler/PaX Team's
code in the last public patch of grsecurity/PaX based on my understanding
of the code. Changes or omissions from the original code are mine and
don't reflect the original grsecurity/PaX code.
Left out of this list is task_struct, which requires special handling
and will be covered in a subsequent patch.
Signed-off-by: Kees Cook <keescook@chromium.org>
for connected socket, the incoming_cpu field in the sock struct
is not going to change frequently, but we are setting it
unconditionally for each packet.
Since sk_incoming_cpu and sk_flags share the same cacheline,
and the latter is access by udp_recvmsg(), this cause a cache
miss for each packet for UDP connected socket.
With this patch, we set the incoming cpu field only when the
ingress cpu really changes.
This gives a small but measurable performance improvement for
connected UDP socket.
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
DRAM supply shortage and poor memory pressure tracking in TCP
stack makes any change in SO_SNDBUF/SO_RCVBUF (or equivalent autotuning
limits) and tcp_mem[] quite hazardous.
TCPMemoryPressures SNMP counter is an indication of tcp_mem sysctl
limits being hit, but only tracking number of transitions.
If TCP stack behavior under stress was perfect :
1) It would maintain memory usage close to the limit.
2) Memory pressure state would be entered for short times.
We certainly prefer 100 events lasting 10ms compared to one event
lasting 200 seconds.
This patch adds a new SNMP counter tracking cumulative duration of
memory pressure events, given in ms units.
$ cat /proc/sys/net/ipv4/tcp_mem
3088 4117 6176
$ grep TCP /proc/net/sockstat
TCP: inuse 180 orphan 0 tw 2 alloc 234 mem 4140
$ nstat -n ; sleep 10 ; nstat |grep Pressure
TcpExtTCPMemoryPressures 1700
TcpExtTCPMemoryPressuresChrono 5209
v2: Used EXPORT_SYMBOL_GPL() instead of EXPORT_SYMBOL() as David
instructed.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Mauro says:
This patch series convert the remaining DocBooks to ReST.
The first version was originally
send as 3 patch series:
[PATCH 00/36] Convert DocBook documents to ReST
[PATCH 0/5] Convert more books to ReST
[PATCH 00/13] Get rid of DocBook
The lsm book was added as if it were a text file under
Documentation. The plan is to merge it with another file
under Documentation/security, after both this series and
a security Documentation patch series gets merged.
It also adjusts some Sphinx-pedantic errors/warnings on
some kernel-doc markups.
I also added some patches here to add PDF output for all
existing ReST books.
BBR congestion control depends on pacing, and pacing is
currently handled by sch_fq packet scheduler for performance reasons,
and also because implemening pacing with FQ was convenient to truly
avoid bursts.
However there are many cases where this packet scheduler constraint
is not practical.
- Many linux hosts are not focusing on handling thousands of TCP
flows in the most efficient way.
- Some routers use fq_codel or other AQM, but still would like
to use BBR for the few TCP flows they initiate/terminate.
This patch implements an automatic fallback to internal pacing.
Pacing is requested either by BBR or use of SO_MAX_PACING_RATE option.
If sch_fq happens to be in the egress path, pacing is delegated to
the qdisc, otherwise pacing is done by TCP itself.
One advantage of pacing from TCP stack is to get more precise rtt
estimations, and less work done from TX completion, since TCP Small
queue limits are not generally hit. Setups with single TX queue but
many cpus might even benefit from this.
Note that unlike sch_fq, we do not take into account header sizes.
Taking care of these headers would add additional complexity for
no practical differences in behavior.
Some performance numbers using 800 TCP_STREAM flows rate limited to
~48 Mbit per second on 40Gbit NIC.
If MQ+pfifo_fast is used on the NIC :
$ sar -n DEV 1 5 | grep eth
14:48:44 eth0 725743.00 2932134.00 46776.76 4335184.68 0.00 0.00 1.00
14:48:45 eth0 725349.00 2932112.00 46751.86 4335158.90 0.00 0.00 0.00
14:48:46 eth0 725101.00 2931153.00 46735.07 4333748.63 0.00 0.00 0.00
14:48:47 eth0 725099.00 2931161.00 46735.11 4333760.44 0.00 0.00 1.00
14:48:48 eth0 725160.00 2931731.00 46738.88 4334606.07 0.00 0.00 0.00
Average: eth0 725290.40 2931658.20 46747.54 4334491.74 0.00 0.00 0.40
$ vmstat 1 5
procs -----------memory---------- ---swap-- -----io---- -system-- ------cpu-----
r b swpd free buff cache si so bi bo in cs us sy id wa st
4 0 0 259825920 45644 2708324 0 0 21 2 247 98 0 0 100 0 0
4 0 0 259823744 45644 2708356 0 0 0 0 2400825 159843 0 19 81 0 0
0 0 0 259824208 45644 2708072 0 0 0 0 2407351 159929 0 19 81 0 0
1 0 0 259824592 45644 2708128 0 0 0 0 2405183 160386 0 19 80 0 0
1 0 0 259824272 45644 2707868 0 0 0 32 2396361 158037 0 19 81 0 0
Now use MQ+FQ :
lpaa23:~# echo fq >/proc/sys/net/core/default_qdisc
lpaa23:~# tc qdisc replace dev eth0 root mq
$ sar -n DEV 1 5 | grep eth
14:49:57 eth0 678614.00 2727930.00 43739.13 4033279.14 0.00 0.00 0.00
14:49:58 eth0 677620.00 2723971.00 43674.69 4027429.62 0.00 0.00 1.00
14:49:59 eth0 676396.00 2719050.00 43596.83 4020125.02 0.00 0.00 0.00
14:50:00 eth0 675197.00 2714173.00 43518.62 4012938.90 0.00 0.00 1.00
14:50:01 eth0 676388.00 2719063.00 43595.47 4020171.64 0.00 0.00 0.00
Average: eth0 676843.00 2720837.40 43624.95 4022788.86 0.00 0.00 0.40
$ vmstat 1 5
procs -----------memory---------- ---swap-- -----io---- -system-- ------cpu-----
r b swpd free buff cache si so bi bo in cs us sy id wa st
2 0 0 259832240 46008 2710912 0 0 21 2 223 192 0 1 99 0 0
1 0 0 259832896 46008 2710744 0 0 0 0 1702206 198078 0 17 82 0 0
0 0 0 259830272 46008 2710596 0 0 0 0 1696340 197756 1 17 83 0 0
4 0 0 259829168 46024 2710584 0 0 16 0 1688472 197158 1 17 82 0 0
3 0 0 259830224 46024 2710408 0 0 0 0 1692450 197212 0 18 82 0 0
As expected, number of interrupts per second is very different.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Cc: Neal Cardwell <ncardwell@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Cc: Van Jacobson <vanj@google.com>
Cc: Jerry Chu <hkchu@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
And update __sk_queue_drop_skb() to work on the specified queue.
This will help the udp protocol to use an additional private
rx queue in a later patch.
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Pull RCU updates from Ingo Molnar:
"The main changes are:
- Debloat RCU headers
- Parallelize SRCU callback handling (plus overlapping patches)
- Improve the performance of Tree SRCU on a CPU-hotplug stress test
- Documentation updates
- Miscellaneous fixes"
* 'core-rcu-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (74 commits)
rcu: Open-code the rcu_cblist_n_lazy_cbs() function
rcu: Open-code the rcu_cblist_n_cbs() function
rcu: Open-code the rcu_cblist_empty() function
rcu: Separately compile large rcu_segcblist functions
srcu: Debloat the <linux/rcu_segcblist.h> header
srcu: Adjust default auto-expediting holdoff
srcu: Specify auto-expedite holdoff time
srcu: Expedite first synchronize_srcu() when idle
srcu: Expedited grace periods with reduced memory contention
srcu: Make rcutorture writer stalls print SRCU GP state
srcu: Exact tracking of srcu_data structures containing callbacks
srcu: Make SRCU be built by default
srcu: Fix Kconfig botch when SRCU not selected
rcu: Make non-preemptive schedule be Tasks RCU quiescent state
srcu: Expedite srcu_schedule_cbs_snp() callback invocation
srcu: Parallelize callback handling
kvm: Move srcu_struct fields to end of struct kvm
rcu: Fix typo in PER_RCU_NODE_PERIOD header comment
rcu: Use true/false in assignment to bool
rcu: Use bool value directly
...
A group of Linux kernel hackers reported chasing a bug that resulted
from their assumption that SLAB_DESTROY_BY_RCU provided an existence
guarantee, that is, that no block from such a slab would be reallocated
during an RCU read-side critical section. Of course, that is not the
case. Instead, SLAB_DESTROY_BY_RCU only prevents freeing of an entire
slab of blocks.
However, there is a phrase for this, namely "type safety". This commit
therefore renames SLAB_DESTROY_BY_RCU to SLAB_TYPESAFE_BY_RCU in order
to avoid future instances of this sort of confusion.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Christoph Lameter <cl@linux.com>
Cc: Pekka Enberg <penberg@kernel.org>
Cc: David Rientjes <rientjes@google.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: <linux-mm@kvack.org>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
[ paulmck: Add comments mentioning the old name, as requested by Eric
Dumazet, in order to help people familiar with the old name find
the new one. ]
Acked-by: David Rientjes <rientjes@google.com>
It seems the code does not match the intent.
This broke packetdrill, and probably other programs.
Fixes: 6c7c98bad4 ("sock: avoid dirtying sk_stamp, if possible")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Paolo Abeni <pabeni@redhat.com>
Acked-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
sock_recv_ts_and_drops() unconditionally set sk->sk_stamp for
every packet, even if the SOCK_TIMESTAMP flag is not set in the
related socket.
If selinux is enabled, this cause a cache miss for every packet
since sk->sk_stamp and sk->sk_security share the same cacheline.
With this change sk_stamp is set only if the SOCK_TIMESTAMP
flag is set, and is cleared for the first packet, so that the user
perceived behavior is unchanged.
This gives up to 5% speed-up under udp-flood with small packets.
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Allows reading of SK_MEMINFO_VARS via socket option. This way an
application can get all meminfo related information in single socket
option call instead of multiple calls.
Adds helper function, sk_get_meminfo(), and uses that for both
getsockopt and sock_diag_put_meminfo().
Suggested by Eric Dumazet.
Signed-off-by: Josh Hunt <johunt@akamai.com>
Reviewed-by: Jason Baron <jbaron@akamai.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Conflicts:
drivers/net/ethernet/broadcom/genet/bcmgenet.c
net/core/sock.c
Conflicts were overlapping changes in bcmgenet and the
lockdep handling of sockets.
Signed-off-by: David S. Miller <davem@davemloft.net>
Lockdep issues a circular dependency warning when AFS issues an operation
through AF_RXRPC from a context in which the VFS/VM holds the mmap_sem.
The theory lockdep comes up with is as follows:
(1) If the pagefault handler decides it needs to read pages from AFS, it
calls AFS with mmap_sem held and AFS begins an AF_RXRPC call, but
creating a call requires the socket lock:
mmap_sem must be taken before sk_lock-AF_RXRPC
(2) afs_open_socket() opens an AF_RXRPC socket and binds it. rxrpc_bind()
binds the underlying UDP socket whilst holding its socket lock.
inet_bind() takes its own socket lock:
sk_lock-AF_RXRPC must be taken before sk_lock-AF_INET
(3) Reading from a TCP socket into a userspace buffer might cause a fault
and thus cause the kernel to take the mmap_sem, but the TCP socket is
locked whilst doing this:
sk_lock-AF_INET must be taken before mmap_sem
However, lockdep's theory is wrong in this instance because it deals only
with lock classes and not individual locks. The AF_INET lock in (2) isn't
really equivalent to the AF_INET lock in (3) as the former deals with a
socket entirely internal to the kernel that never sees userspace. This is
a limitation in the design of lockdep.
Fix the general case by:
(1) Double up all the locking keys used in sockets so that one set are
used if the socket is created by userspace and the other set is used
if the socket is created by the kernel.
(2) Store the kern parameter passed to sk_alloc() in a variable in the
sock struct (sk_kern_sock). This informs sock_lock_init(),
sock_init_data() and sk_clone_lock() as to the lock keys to be used.
Note that the child created by sk_clone_lock() inherits the parent's
kern setting.
(3) Add a 'kern' parameter to ->accept() that is analogous to the one
passed in to ->create() that distinguishes whether kernel_accept() or
sys_accept4() was the caller and can be passed to sk_alloc().
Note that a lot of accept functions merely dequeue an already
allocated socket. I haven't touched these as the new socket already
exists before we get the parameter.
Note also that there are a couple of places where I've made the accepted
socket unconditionally kernel-based:
irda_accept()
rds_rcp_accept_one()
tcp_accept_from_sock()
because they follow a sock_create_kern() and accept off of that.
Whilst creating this, I noticed that lustre and ocfs don't create sockets
through sock_create_kern() and thus they aren't marked as for-kernel,
though they appear to be internal. I wonder if these should do that so
that they use the new set of lock keys.
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
__sk_dst_set() must be called while we own the socket.
We can get proper lockdep coverage using lockdep_sock_is_held()
and rcu_dereference_protected()
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When handling problems in cloning a socket with the sk_clone_locked()
function we need to perform several steps that were open coded in it and
its callers, so introduce a routine to avoid this duplication:
sk_free_unlock_clone().
Cc: Cong Wang <xiyou.wangcong@gmail.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Eric Dumazet <edumazet@google.com>
Cc: Gerrit Renker <gerrit@erg.abdn.ac.uk>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/n/net-ui6laqkotycunhtmqryl9bfx@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The conflict was an interaction between a bug fix in the
netvsc driver in 'net' and an optimization of the RX path
in 'net-next'.
Signed-off-by: David S. Miller <davem@davemloft.net>
Add new skbuff flag to allow protocols to confirm neighbour.
When same struct dst_entry can be used for many different
neighbours we can not use it for pending confirmations.
Add sock_confirm_neigh() helper to confirm the neighbour and
use it for IPv4, IPv6 and VRF before dst_neigh_output.
Signed-off-by: Julian Anastasov <ja@ssi.bg>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add new sock flag to allow sockets to confirm neighbour.
When same struct dst_entry can be used for many different
neighbours we can not use it for pending confirmations.
As not all call paths lock the socket use full word for
the flag.
Add sk_dst_confirm as replacement for dst_confirm when
called for received packets.
Signed-off-by: Julian Anastasov <ja@ssi.bg>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Dmitry reported that UDP sockets being destroyed would trigger the
WARN_ON(atomic_read(&sk->sk_rmem_alloc)); in inet_sock_destruct()
It turns out we do not properly destroy skb(s) that have wrong UDP
checksum.
Thanks again to syzkaller team.
Fixes : 7c13f97ffd ("udp: do fwd memory scheduling on dequeue")
Reported-by: Dmitry Vyukov <dvyukov@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Paolo Abeni <pabeni@redhat.com>
Cc: Hannes Frederic Sowa <hannes@stressinduktion.org>
Acked-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Slava Shwartsman reported a warning in skb_try_coalesce(), when we
detect skb->truesize is completely wrong.
In his case, issue came from IPv6 reassembly coping with malicious
datagrams, that forced various pskb_may_pull() to reallocate a bigger
skb->head than the one allocated by NIC driver before entering GRO
layer.
Current code does not change skb->truesize, leaving this burden to
callers if they care enough.
Blindly changing skb->truesize in pskb_expand_head() is not
easy, as some producers might track skb->truesize, for example
in xmit path for back pressure feedback (sk->sk_wmem_alloc)
We can detect the cases where it should be safe to change
skb->truesize :
1) skb is not attached to a socket.
2) If it is attached to a socket, destructor is sock_edemux()
My audit gave only two callers doing their own skb->truesize
manipulation.
I had to remove skb parameter in sock_edemux macro when
CONFIG_INET is not set to avoid a compile error.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Slava Shwartsman <slavash@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use hlist_entry_safe() instead of open-coding it.
Signed-off-by: Geliang Tang <geliangtang@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Support for SMC socket monitoring via netlink sockets of protocol
NETLINK_SOCK_DIAG.
Signed-off-by: Ursula Braun <ubraun@linux.vnet.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Direct call of tcp_set_keepalive() function from protocol-agnostic
sock_setsockopt() function in net/core/sock.c violates network
layering. And newly introduced protocol (SMC-R) will need its own
keepalive function. Therefore, add "keepalive" function pointer
to "struct proto", and call it from sock_setsockopt() via this pointer.
Signed-off-by: Ursula Braun <ubraun@linux.vnet.ibm.com>
Reviewed-by: Utz Bacher <utz.bacher@de.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
ktime is a union because the initial implementation stored the time in
scalar nanoseconds on 64 bit machine and in a endianess optimized timespec
variant for 32bit machines. The Y2038 cleanup removed the timespec variant
and switched everything to scalar nanoseconds. The union remained, but
become completely pointless.
Get rid of the union and just keep ktime_t as simple typedef of type s64.
The conversion was done with coccinelle and some manual mopping up.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Pull vfs updates from Al Viro:
- more ->d_init() stuff (work.dcache)
- pathname resolution cleanups (work.namei)
- a few missing iov_iter primitives - copy_from_iter_full() and
friends. Either copy the full requested amount, advance the iterator
and return true, or fail, return false and do _not_ advance the
iterator. Quite a few open-coded callers converted (and became more
readable and harder to fuck up that way) (work.iov_iter)
- several assorted patches, the big one being logfs removal
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
logfs: remove from tree
vfs: fix put_compat_statfs64() does not handle errors
namei: fold should_follow_link() with the step into not-followed link
namei: pass both WALK_GET and WALK_MORE to should_follow_link()
namei: invert WALK_PUT logics
namei: shift interpretation of LOOKUP_FOLLOW inside should_follow_link()
namei: saner calling conventions for mountpoint_last()
namei.c: get rid of user_path_parent()
switch getfrag callbacks to ..._full() primitives
make skb_add_data,{_nocache}() and skb_copy_to_page_nocache() advance only on success
[iov_iter] new primitives - copy_from_iter_full() and friends
don't open-code file_inode()
ceph: switch to use of ->d_init()
ceph: unify dentry_operations instances
lustre: switch to use of ->d_init()
sk_drops can be an often written field, do not read it unless
application showed interest.
Note that sk_drops can be read via inet_diag, so applications
can avoid getting this info from every received packet.
In the future, 'reading' sk_drops might require folding per node or per
cpu fields, and thus become even more expensive than today.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
RFS is not commonly used, so add a jump label to avoid some conditionals
in fast path.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Paolo noticed a cache line miss in UDP recvmsg() to access
sk_rxhash, sharing a cache line with sk_drops.
sk_drops might be heavily incremented by cpus handling a flood targeting
this socket.
We might place sk_drops on a separate cache line, but lets try
to avoid wasting 64 bytes per socket just for this, since we have
other bottlenecks to take care of.
sock_rps_record_flow() should only access sk_rxhash for connected
flows.
Testing sk_state for TCP_ESTABLISHED covers most of the cases for
connected sockets, for a zero cost, since system calls using
sock_rps_record_flow() also access sk->sk_prot which is on the
same cache line.
A follow up patch will provide a static_key (Jump Label) since most
hosts do not even use RFS.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Paolo Abeni <pabeni@redhat.com>
Acked-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Group fields used in TX path, and keep some cache lines mostly read
to permit sharing among cpus.
Gained two 4 bytes holes on 64bit arches.
Added a place holder for tcp tsq_flags, next to sk_wmem_alloc
to speed up tcp_wfree() in the following patch.
I have not added ____cacheline_aligned_in_smp, this might be done later.
I prefer doing this once inet and tcp/udp sockets reorg is also done.
Tested with both TCP and UDP.
UDP receiver performance under flood increased by ~20 % :
Accessing sk_filter/sk_wq/sk_napi_id no longer stalls because sk_drops
was moved away from a critical cache line, now mostly read and shared.
/* --- cacheline 4 boundary (256 bytes) --- */
unsigned int sk_napi_id; /* 0x100 0x4 */
int sk_rcvbuf; /* 0x104 0x4 */
struct sk_filter * sk_filter; /* 0x108 0x8 */
union {
struct socket_wq * sk_wq; /* 0x8 */
struct socket_wq * sk_wq_raw; /* 0x8 */
}; /* 0x110 0x8 */
struct xfrm_policy * sk_policy[2]; /* 0x118 0x10 */
struct dst_entry * sk_rx_dst; /* 0x128 0x8 */
struct dst_entry * sk_dst_cache; /* 0x130 0x8 */
atomic_t sk_omem_alloc; /* 0x138 0x4 */
int sk_sndbuf; /* 0x13c 0x4 */
/* --- cacheline 5 boundary (320 bytes) --- */
int sk_wmem_queued; /* 0x140 0x4 */
atomic_t sk_wmem_alloc; /* 0x144 0x4 */
long unsigned int sk_tsq_flags; /* 0x148 0x8 */
struct sk_buff * sk_send_head; /* 0x150 0x8 */
struct sk_buff_head sk_write_queue; /* 0x158 0x18 */
__s32 sk_peek_off; /* 0x170 0x4 */
int sk_write_pending; /* 0x174 0x4 */
long int sk_sndtimeo; /* 0x178 0x8 */
Signed-off-by: Eric Dumazet <edumazet@google.com>
Tested-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add socket family, type and protocol to bpf_sock allowing bpf programs
read-only access.
Add __sk_flags_offset[0] to struct sock before the bitfield to
programmtically determine the offset of the unsigned int containing
protocol and type.
Signed-off-by: David Ahern <dsa@cumulusnetworks.com>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
