The original intention was to avoid dependencies between nfnetlink_queue and
conntrack without ifdef pollution. However, we can achieve this by moving the
conntrack dependent code into ctnetlink and keep some glue code to access the
nfq_ct indirection from nfqueue.
After this patch, the nfq_ct indirection is always compiled in the netfilter
core to avoid polluting nfqueue with ifdefs. Thus, if nf_conntrack is not
compiled this results in only 8-bytes of memory waste in x86_64.
This patch also adds ctnetlink_nfqueue_seqadj() to avoid that the nf_conn
structure layout if exposed to nf_queue, which creates another dependency with
nf_conntrack at compilation time.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Before letting request sockets being put in TCP/DCCP regular
ehash table, we need to add either :
- SLAB_DESTROY_BY_RCU flag to their kmem_cache
- add RCU grace period before freeing them.
Since we carefully respected the SLAB_DESTROY_BY_RCU protocol
like ESTABLISH and TIMEWAIT sockets, use it here.
req_prot_init() being only used by TCP and DCCP, I did not add
a new slab_flags into their rsk_prot, but reuse prot->slab_flags
Since all reqsk_alloc() users are correctly dealing with a failure,
add the __GFP_NOWARN flag to avoid traces under pressure.
Fixes: 079096f103 ("tcp/dccp: install syn_recv requests into ehash table")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
{cls,act}_bpf can now set the skb->priority from an eBPF program based
on various critera, so that for example classful qdiscs like multiq can
update the skb's priority during enqueue time and further push it down
into subsequent qdiscs.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Using routing realms as part of the classifier is quite useful, it
can be viewed as a tag for one or multiple routing entries (think of
an analogy to net_cls cgroup for processes), set by user space routing
daemons or via iproute2 as an indicator for traffic classifiers and
later on processed in the eBPF program.
Unlike actions, the classifier can inspect device flags and enable
netif_keep_dst() if necessary. tc actions don't have that possibility,
but in case people know what they are doing, it can be used from there
as well (e.g. via devs that must keep dsts by design anyway).
If a realm is set, the handler returns the non-zero realm. User space
can set the full 32bit realm for the dst.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
As we need to add further flags to the bpf_prog structure, lets migrate
both bools to a bitfield representation. The size of the base structure
(excluding insns) remains unchanged at 40 bytes.
Add also tags for the kmemchecker, so that it doesn't throw false
positives. Even in case gcc would generate suboptimal code, it's not
being accessed in performance critical paths.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@plumgrid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Everything should now be ready to finally allow SYN
packets processing without holding listener lock.
Tested:
3.5 Mpps SYNFLOOD. Plenty of cpu cycles available.
Next bottleneck is the refcount taken on listener,
that could be avoided if we remove SLAB_DESTROY_BY_RCU
strict semantic for listeners, and use regular RCU.
13.18% [kernel] [k] __inet_lookup_listener
9.61% [kernel] [k] tcp_conn_request
8.16% [kernel] [k] sha_transform
5.30% [kernel] [k] inet_reqsk_alloc
4.22% [kernel] [k] sock_put
3.74% [kernel] [k] tcp_make_synack
2.88% [kernel] [k] ipt_do_table
2.56% [kernel] [k] memcpy_erms
2.53% [kernel] [k] sock_wfree
2.40% [kernel] [k] tcp_v4_rcv
2.08% [kernel] [k] fib_table_lookup
1.84% [kernel] [k] tcp_openreq_init_rwin
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
If a listener with thousands of children in accept queue
is dismantled, it can take a while to close all of them.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This control variable was set at first listen(fd, backlog)
call, but not updated if application tried to increase or decrease
backlog. It made sense at the time listener had a non resizeable
hash table.
Also rounding to powers of two was not very friendly.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
It is enough to check listener sk_state, no need for an extra
condition.
max_qlen_log can be moved into struct request_sock_queue
We can remove syn_wait_lock and the alignment it enforced.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
If a listen backlog is very big (to avoid syncookies), then
the listener sk->sk_wmem_alloc is the main source of false
sharing, as we need to touch it twice per SYNACK re-transmit
and TX completion.
(One SYN packet takes listener lock once, but up to 6 SYNACK
are generated)
By attaching the skb to the request socket, we remove this
source of contention.
Tested:
listen(fd, 10485760); // single listener (no SO_REUSEPORT)
16 RX/TX queue NIC
Sustain a SYNFLOOD attack of ~320,000 SYN per second,
Sending ~1,400,000 SYNACK per second.
Perf profiles now show listener spinlock being next bottleneck.
20.29% [kernel] [k] queued_spin_lock_slowpath
10.06% [kernel] [k] __inet_lookup_established
5.12% [kernel] [k] reqsk_timer_handler
3.22% [kernel] [k] get_next_timer_interrupt
3.00% [kernel] [k] tcp_make_synack
2.77% [kernel] [k] ipt_do_table
2.70% [kernel] [k] run_timer_softirq
2.50% [kernel] [k] ip_finish_output
2.04% [kernel] [k] cascade
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We no longer use hash_rnd, nr_table_entries and syn_table[]
For a listener with a backlog of 10 millions sockets, this
saves 80 MBytes of vmalloced memory.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In this patch, we insert request sockets into TCP/DCCP
regular ehash table (where ESTABLISHED and TIMEWAIT sockets
are) instead of using the per listener hash table.
ACK packets find SYN_RECV pseudo sockets without having
to find and lock the listener.
In nominal conditions, this halves pressure on listener lock.
Note that this will allow for SO_REUSEPORT refinements,
so that we can select a listener using cpu/numa affinities instead
of the prior 'consistent hash', since only SYN packets will
apply this selection logic.
We will shrink listen_sock in the following patch to ease
code review.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Ying Cai <ycai@google.com>
Cc: Willem de Bruijn <willemb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When request sockets are no longer in a per listener hash table
but on regular TCP ehash, we need to access listener uid
through req->rsk_listener
get_openreq6() also gets a const for its request socket argument.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Once listener is lockless, its sk_state can change anytime.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We'll soon have to call tcp_v[46]_inbound_md5_hash() twice.
Also add const attribute to the socket, as it might be the
unlocked listener for SYN packets.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This fixes a typo : We want to store the NAPI id on child socket.
Presumably nobody really uses busy polling, on short lived flows.
Fixes: 3d97379a67 ("tcp: move sk_mark_napi_id() at the right place")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
long term plan is to remove struct listen_sock when its hash
table is no longer there.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
qlen_inc & young_inc were protected by listener lock,
while qlen_dec & young_dec were atomic fields.
Everything needs to be atomic for upcoming lockless listener.
Also move qlen/young in request_sock_queue as we'll get rid
of struct listen_sock eventually.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
struct request_sock_queue fields are currently protected
by the listener 'lock' (not a real spinlock)
We need to add a private spinlock instead, so that softirq handlers
creating children do not have to worry with backlog notion
that the listener 'lock' carries.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since we do not know in which context drivers will call these
routines, they should use the unlocked version of nci_request,
i.e. __nci_request.
It is up to drivers to know/decide if they need to take the
req_lock mutex before calling those routines. When being called
from the NCI setup routine there is no need to do so as ops->setup
is called under req_lock.
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
There was a missing return here so it meant that often
ieee802154_llsec_parse_key_id() was not called.
Fixes: a26c5fd762 ('nl802154: add support for security layer')
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Acked-by: Alexander Aring <alex.aring@gmail.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Pull networking fixes from David Miller:
1) Fix regression in SKB partial checksum handling, from Pravin B
Shalar.
2) Fix VLAN inside of VXLAN handling in i40e driver, from Jesse
Brandeburg.
3) Cure softlockups during accept() in SCTP, from Karl Heiss.
4) MSG_PEEK should return multiple SKBs worth of data in AF_UNIX, from
Aaron Conole.
5) IPV6 erroneously ignores output interface specifier in lookup key for
route lookups, fix from David Ahern.
6) In Marvell DSA driver, forward unknown frames to CPU port, from
Andrew Lunn.
7) Mission flow flag initializations in some code paths, from David
Ahern.
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net:
net: Initialize flow flags in input path
net: dsa: fix preparation of a port STP update
testptp: Silence compiler warnings on ppc64
net/mlx4: Handle return codes in mlx4_qp_attach_common
dsa: mv88e6xxx: Enable forwarding for unknown to the CPU port
skbuff: Fix skb checksum partial check.
net: ipv6: Add RT6_LOOKUP_F_IFACE flag if oif is set
net sysfs: Print link speed as signed integer
bna: fix error handling
af_unix: return data from multiple SKBs on recv() with MSG_PEEK flag
af_unix: Convert the unix_sk macro to an inline function for type safety
net: sctp: Don't use 64 kilobyte lookup table for four elements
l2tp: protect tunnel->del_work by ref_count
net/ibm/emac: bump version numbers for correct work with ethtool
sctp: Prevent soft lockup when sctp_accept() is called during a timeout event
sctp: Whitespace fix
i40e/i40evf: check for stopped admin queue
i40e: fix VLAN inside VXLAN
r8169: fix handling rtl_readphy result
net: hisilicon: fix handling platform_get_irq result
We should not pass the original flags when creating a context vlan only
because they may contain some flags that change behaviour in the bridge.
The new global context should be with minimal set of flags, so pass 0
and let br_vlan_add() set the master flag only.
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When a new port is being added we need to make vlgrp available after
rhashtable has been initialized and when removing a port we need to
flush the vlans and free the resources after we're sure noone can use
the port, i.e. after it's removed from the port list and synchronize_rcu
is executed.
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
One obvious way to converge more code (which was also used by the
previous vlan code) is to move pvid inside net_bridge_vlan_group. This
allows us to simplify some and remove other port-specific functions.
Also gives us the ability to simply pass the vlan group and use all of the
contained information.
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
While a new port is being initialized the rx_handler gets set, but the
vlans get initialized later in br_add_if() and in that window if we
receive a frame with a link-local address we can try to dereference
p->vlgrp in:
br_handle_frame() -> br_handle_local_finish() -> br_should_learn()
Fix this by checking vlgrp before using it.
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
As Stephen pointed out the default initial size is more than we need, so
let's start small (4 elements, thus nelem_hint = 3). Also limit the hash
locks to the number of CPUs as we don't need any write-side scaling and
this looks like the minimum.
Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Pull rdma fixes from Doug Ledford:
- Fixes for mlx5 related issues
- Fixes for ipoib multicast handling
* tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dledford/rdma:
IB/ipoib: increase the max mcast backlog queue
IB/ipoib: Make sendonly multicast joins create the mcast group
IB/ipoib: Expire sendonly multicast joins
IB/mlx5: Remove pa_lkey usages
IB/mlx5: Remove support for IB_DEVICE_LOCAL_DMA_LKEY
IB/iser: Add module parameter for always register memory
xprtrdma: Replace global lkey with lkey local to PD
This reverts commit 9abc378c66e3d6f437eed77c1c534cbc183523f7
("ieee802154: 6lowpan: change datagram var types").
The reason is that I forgot the IPv6 fragmentation here. Our MTU of
lowpan interface is 1280 and skb->len should not above of that. If we
reach a payload above 1280 in IPv6 header then we have a IPv6
fragmentation above 802.15.4 6LoWPAN fragmentation. The type "u16" was
fine, instead I added now a WARN_ON_ONCE if skb->len is above MTU which
should never happen otherwise IPv6 on minimum MTU size is broken.
Signed-off-by: Alexander Aring <alex.aring@gmail.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
I arranged the code so that the compiler can remove the unecessary bits
in ip_vs_leave when CONFIG_SYSCTL is unset, and removed an explicit
CONFIG_SYSCTL.
Unfortunately when rebasing my work on top of that of Alex Gartrell I
missed the fact that the newly added function ip_vs_addr_is_unicast was
surrounded by CONFIG_SYSCTL.
So remove the now unnecessary CONFIG_SYSCTL guards around
ip_vs_addr_is_unicast. It is causing build failures today when
CONFIG_SYSCTL is not selected and any self respecting compiler will
notice that sysctl_cache_bypass is always false without CONFIG_SYSCTL
and not include the logic from the function ip_vs_addr_is_unicast in
the compiled code.
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Acked-by: Julian Anastasov <ja@ssi.bg>
Signed-off-by: Simon Horman <horms@verge.net.au>
net/built-in.o: In function `nf_dup_ipv4': (.text+0xed24d): undefined reference to `nf_conntrack_untracked'
net/built-in.o: In function `nf_dup_ipv4': (.text+0xed267): undefined reference to `nf_conntrack_untracked'
net/built-in.o: In function `nf_dup_ipv6': (.text+0x158aef): undefined reference to `nf_conntrack_untracked'
net/built-in.o: In function `nf_dup_ipv6': (.text+0x158b09): undefined reference to `nf_conntrack_untracked'
Reported-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
One global lock protecting hash-tables with 1024 buckets isn't
efficient and it shows up in a massive systems with truck
loads of RDS sockets serving multiple databases. The
perf data clearly highlights the contention on the rw
lock in these massive workloads.
When the contention gets worse, the code gets into a state where
it decides to back off on the lock. So while it has disabled interrupts,
it sits and backs off on this lock get. This causes the system to
become sluggish and eventually all sorts of bad things happen.
The simple fix is to move the lock into the hash bucket and
use per-bucket lock to improve the scalability.
Signed-off-by: Santosh Shilimkar <ssantosh@kernel.org>
Signed-off-by: Santosh Shilimkar <santosh.shilimkar@oracle.com>
One need to take rds socket reference while using it and release it
once done with it. rds_add_bind() code path does not do that so
lets fix it.
Signed-off-by: Santosh Shilimkar <ssantosh@kernel.org>
Signed-off-by: Santosh Shilimkar <santosh.shilimkar@oracle.com>
RDS bind and release locking scheme is very inefficient. It
uses RCU for maintaining the bind hash-table which is great but
it also needs to hold spinlock for [add/remove]_bound(). So
overall usecase, the hash-table concurrent speedup doesn't pay off.
In fact blocking nature of synchronize_rcu() makes the RDS
socket shutdown too slow which hurts RDS performance since
connection shutdown and re-connect happens quite often to
maintain the RC part of the protocol.
So we make the locking scheme simpler and more efficient by
replacing spin_locks with reader/writer locks and getting rid
off rcu for bind hash-table.
In subsequent patch, we also covert the global lock with per-bucket
lock to reduce the global lock contention.
Signed-off-by: Santosh Shilimkar <ssantosh@kernel.org>
Signed-off-by: Santosh Shilimkar <santosh.shilimkar@oracle.com>
synchronize_rcu() slowing down un-necessarily the socket shutdown
path. It is used just kfree() the ip addresses in rds_ib_remove_ipaddr()
which is perfect usecase for kfree_rcu();
So lets use that to gain some speedup.
Signed-off-by: Santosh Shilimkar <ssantosh@kernel.org>
Signed-off-by: Santosh Shilimkar <santosh.shilimkar@oracle.com>
This patch adds support for increment transmit and receive stats. The
meaning of these stats are IPv6 based, which shows the stats after
running the 6lowpan adaptation layer (uncompression/compression,
fragmentation handling) on receive and before the adaptation layer
when transmit.
Signed-off-by: Alexander Aring <alex.aring@gmail.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
This patch fixes the data frame sequence numer (dsn) while 6lowpan
fragmentation for frag1. Currently we create one 802.15.4 header at
first, then check if it's match into one frame and at the end construct
many fragments and calling wpan_dev_hard_header for each of them,
inclusive for the first fragment. This will make the first generated
header to garbage, instead we copying this header for frag1 instead of
generate a new one which skips one dsn.
Signed-off-by: Alexander Aring <alex.aring@gmail.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
This patch changes datagram size variable from u16 type to unsigned int.
The reason is that an IPv6 header has an MAX_UIN16 payload length, but
the datagram size is payload + IPv6 header length. This avoids overflows
at some places.
Signed-off-by: Alexander Aring <alex.aring@gmail.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
This patch changes the mtu size of 802.15.4 interfaces. The current
setting is the meaning of the maximum transport unit with mac header,
which is 127 bytes according 802.15.4. The linux meaning of the mtu size
field is the maximum payload of a mac frame. Like in ethernet, which is
1500 bytes.
We have dynamic length of mac frames in 802.15.4, this is why we assume
the minimum header length which is hard_header_len. This contains fc and
sequence fields. These can evaluated by driver layer without additional
checks. We currently don't support to set the FCS from userspace, so we
need to subtract this from mtu size as well.
Signed-off-by: Alexander Aring <alex.aring@gmail.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
While doing a little test with the llsec implementation I saw these
issues. We should move decryption and encruption somewhere else,
otherwise while capturing with wireshark the mac header shows secuirty
fields but the payload is plaintext.
A complete other issue is what doing with HardMAC drivers where the
payload is always plaintext. I think we need a special handling then in
userspace. We currently doesn't support any HardMAC transceivers, so we
should fix the first issue for SoftMAC transceivers.
Signed-off-by: Alexander Aring <alex.aring@gmail.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
