Motivation for soreuseport would be something like a web server
binding to port 80 running with multiple threads, where each thread
might have it's own listener socket. This could be done as an
alternative to other models: 1) have one listener thread which
dispatches completed connections to workers. 2) accept on a single
listener socket from multiple threads. In case #1 the listener thread
can easily become the bottleneck with high connection turn-over rate.
In case #2, the proportion of connections accepted per thread tends
to be uneven under high connection load (assuming simple event loop:
while (1) { accept(); process() }, wakeup does not promote fairness
among the sockets. We have seen the disproportion to be as high
as 3:1 ratio between thread accepting most connections and the one
accepting the fewest. With so_reusport the distribution is
uniform.
Signed-off-by: Tom Herbert <therbert@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
commit 68835aba4d (net: optimize INET input path further)
moved some fields used for tcp/udp sockets lookup in the first cache
line of struct sock_common.
This patch moves inet_dport/inet_num as well, filling a 32bit hole
on 64 bit arches and reducing number of cache line misses in lookups.
Also change INET_MATCH()/INET_TW_MATCH() to perform the ports match
before addresses match, as this check is more discriminant.
Remove the hash check from MATCH() macros because we dont need to
re validate the hash value after taking a refcount on socket, and
use likely/unlikely compiler hints, as the sk_hash/hash check
makes the following conditional tests 100% predicted by cpu.
Introduce skc_addrpair/skc_portpair pair values to better
document the alignment requirements of the port/addr pairs
used in the various MATCH() macros, and remove some casts.
The namespace check can also be done at last.
This slightly improves TCP/UDP lookup times.
IP/TCP early demux needs inet->rx_dst_ifindex and
TCP needs inet->min_ttl, lets group them together in same cache line.
With help from Ben Hutchings & Joe Perches.
Idea of this patch came after Ling Ma proposal to move skc_hash
to the beginning of struct sock_common, and should allow him
to submit a final version of his patch. My tests show an improvement
doing so.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Ben Hutchings <bhutchings@solarflare.com>
Cc: Joe Perches <joe@perches.com>
Cc: Ling Ma <ling.ma.program@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Computers have become a lot faster since we compromised on the
partial MD4 hash which we use currently for performance reasons.
MD5 is a much safer choice, and is inline with both RFC1948 and
other ISS generators (OpenBSD, Solaris, etc.)
Furthermore, only having 24-bits of the sequence number be truly
unpredictable is a very serious limitation. So the periodic
regeneration and 8-bit counter have been removed. We compute and
use a full 32-bit sequence number.
For ipv6, DCCP was found to use a 32-bit truncated initial sequence
number (it needs 43-bits) and that is fixed here as well.
Reported-by: Dan Kaminsky <dan@doxpara.com>
Tested-by: Willy Tarreau <w@1wt.eu>
Signed-off-by: David S. Miller <davem@davemloft.net>
Fix a bunch of
warning: ‘inline’ is not at beginning of declaration
messages when building a 'make allyesconfig' kernel with -Wextra.
These warnings are trivial to kill, yet rather annoying when building with
-Wextra.
The more we can cut down on pointless crap like this the better (IMHO).
A previous patch to do this for a 'allnoconfig' build has already been
merged. This just takes the cleanup a little further.
Signed-off-by: Jesper Juhl <jj@chaosbits.net>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
First patch changes __inet_hash_nolisten() and __inet6_hash()
to get a timewait parameter to be able to unhash it from ehash
at same time the new socket is inserted in hash.
This makes sure timewait socket wont be found by a concurrent
writer in __inet_check_established()
Reported-by: kapil dakhane <kdakhane@gmail.com>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Its currently possible that several threads issuing a connect() find
the same timewait socket and try to reuse it, leading to list
corruptions.
Condition for bug is that these threads bound their socket on same
address/port of to-be-find timewait socket, and connected to same
target. (SO_REUSEADDR needed)
To fix this problem, we could unhash timewait socket while holding
ehash lock, to make sure lookups/changes will be serialized. Only
first thread finds the timewait socket, other ones find the
established socket and return an EADDRNOTAVAIL error.
This second version takes into account Evgeniy's review and makes sure
inet_twsk_put() is called outside of locked sections.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In order to have better cache layouts of struct sock (separate zones
for rx/tx paths), we need this preliminary patch.
Goal is to transfert fields used at lookup time in the first
read-mostly cache line (inside struct sock_common) and move sk_refcnt
to a separate cache line (only written by rx path)
This patch adds inet_ prefix to daddr, rcv_saddr, dport, num, saddr,
sport and id fields. This allows a future patch to define these
fields as macros, like sk_refcnt, without name clashes.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Storing the mask (size - 1) instead of the size allows fast path to be
a bit faster.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We already have a valid net in that place, but this is not just a
cleanup - the tw pointer can be NULL there sometimes, thus causing
an oops in NET_NS=y case.
The same place in ipv4 code already works correctly using existing
net, rather than tw's one.
The bug exists since 2.6.27.
Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
This is the last step to be able to perform full RCU lookups
in __inet_lookup() : After established/timewait tables, we
add RCU lookups to listening hash table.
The only trick here is that a socket of a given type (TCP ipv4,
TCP ipv6, ...) can now flight between two different tables
(established and listening) during a RCU grace period, so we
must use different 'nulls' end-of-chain values for two tables.
We define a large value :
#define LISTENING_NULLS_BASE (1U << 29)
So that slots in listening table are guaranteed to have different
end-of-chain values than slots in established table. A reader can
still detect it finished its lookup in the right chain.
Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Now TCP & DCCP use RCU lookups, we can convert ehash rwlocks to spinlocks.
/proc/net/tcp and other seq_file 'readers' can safely be converted to 'writers'.
This should speedup writers, since spin_lock()/spin_unlock()
only use one atomic operation instead of two for write_lock()/write_unlock()
Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch prepares RCU migration of listening_hash table for
TCP/DCCP protocols.
listening_hash table being small (32 slots per protocol), we add
a spinlock for each slot, instead of a single rwlock for whole table.
This should reduce hold time of readers, and writers concurrency.
Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
RCU was added to UDP lookups, using a fast infrastructure :
- sockets kmem_cache use SLAB_DESTROY_BY_RCU and dont pay the
price of call_rcu() at freeing time.
- hlist_nulls permits to use few memory barriers.
This patch uses same infrastructure for TCP/DCCP established
and timewait sockets.
Thanks to SLAB_DESTROY_BY_RCU, no slowdown for applications
using short lived TCP connections. A followup patch, converting
rwlocks to spinlocks will even speedup this case.
__inet_lookup_established() is pretty fast now we dont have to
dirty a contended cache line (read_lock/read_unlock)
Only established and timewait hashtable are converted to RCU
(bind table and listen table are still using traditional locking)
Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Removes legacy reinvent-the-wheel type thing. The generic
machinery integrates much better to automated debugging aids
such as kerneloops.org (and others), and is unambiguous due to
better naming. Non-intuively BUG_TRAP() is actually equal to
WARN_ON() rather than BUG_ON() though some might actually be
promoted to BUG_ON() but I left that to future.
I could make at least one BUILD_BUG_ON conversion.
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: David S. Miller <davem@davemloft.net>
Same as for inet_hashfn, prepare its ipv6 incarnation.
Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Listening-on-one-port sockets in many namespaces produce long
chains in the listening_hash-es, so prepare the inet_lhashfn to
take struct net into account.
Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
This counter is about to become per-proto-and-per-net, so we'll need
two arguments to determine which cell in this "table" to work with.
All the places, but proc already pass proper net to it - proc will be
tuned a bit later.
Some indentation with spaces in proc files is done to keep the file
coding style consistent.
Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Introduce an inline net_eq() to compare two namespaces.
Without CONFIG_NET_NS, since no namespace other than &init_net
exists, it is always 1.
We do not need to convert 1) inline vs inline and
2) inline vs &init_net comparisons.
Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Introduce per-sock inlines: sock_net(), sock_net_set()
and per-inet_timewait_sock inlines: twsk_net(), twsk_net_set().
Without CONFIG_NET_NS, no namespace other than &init_net exists.
Let's explicitly define them to help compiler optimizations.
Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Inspired by the commit ab1e0a13 ([SOCK] proto: Add hashinfo member to
struct proto) from Arnaldo, I made similar thing for UDP/-Lite IPv4
and -v6 protocols.
The result is not that exciting, but it removes some levels of
indirection in udpxxx_get_port and saves some space in code and text.
The first step is to union existing hashinfo and new udp_hash on the
struct proto and give a name to this union, since future initialization
of tcpxxx_prot, dccp_vx_protinfo and udpxxx_protinfo will cause gcc
warning about inability to initialize anonymous member this way.
Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
The port offset calculations depend on the protocol family, but, as
Adrian noticed, I broke this logic with the commit
5ee31fc1ec
[INET]: Consolidate inet(6)_hash_connect.
Return this logic back, by passing the port offset directly into the
consolidated function.
Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Noticed-by: Adrian Bunk <bunk@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
This way we can remove TCP and DCCP specific versions of
sk->sk_prot->get_port: both v4 and v6 use inet_csk_get_port
sk->sk_prot->hash: inet_hash is directly used, only v6 need
a specific version to deal with mapped sockets
sk->sk_prot->unhash: both v4 and v6 use inet_hash directly
struct inet_connection_sock_af_ops also gets a new member, bind_conflict, so
that inet_csk_get_port can find the per family routine.
Now only the lookup routines receive as a parameter a struct inet_hashtable.
With this we further reuse code, reducing the difference among INET transport
protocols.
Eventually work has to be done on UDP and SCTP to make them share this
infrastructure and get as a bonus inet_diag interfaces so that iproute can be
used with these protocols.
net-2.6/net/ipv4/inet_hashtables.c:
struct proto | +8
struct inet_connection_sock_af_ops | +8
2 structs changed
__inet_hash_nolisten | +18
__inet_hash | -210
inet_put_port | +8
inet_bind_bucket_create | +1
__inet_hash_connect | -8
5 functions changed, 27 bytes added, 218 bytes removed, diff: -191
net-2.6/net/core/sock.c:
proto_seq_show | +3
1 function changed, 3 bytes added, diff: +3
net-2.6/net/ipv4/inet_connection_sock.c:
inet_csk_get_port | +15
1 function changed, 15 bytes added, diff: +15
net-2.6/net/ipv4/tcp.c:
tcp_set_state | -7
1 function changed, 7 bytes removed, diff: -7
net-2.6/net/ipv4/tcp_ipv4.c:
tcp_v4_get_port | -31
tcp_v4_hash | -48
tcp_v4_destroy_sock | -7
tcp_v4_syn_recv_sock | -2
tcp_unhash | -179
5 functions changed, 267 bytes removed, diff: -267
net-2.6/net/ipv6/inet6_hashtables.c:
__inet6_hash | +8
1 function changed, 8 bytes added, diff: +8
net-2.6/net/ipv4/inet_hashtables.c:
inet_unhash | +190
inet_hash | +242
2 functions changed, 432 bytes added, diff: +432
vmlinux:
16 functions changed, 485 bytes added, 492 bytes removed, diff: -7
/home/acme/git/net-2.6/net/ipv6/tcp_ipv6.c:
tcp_v6_get_port | -31
tcp_v6_hash | -7
tcp_v6_syn_recv_sock | -9
3 functions changed, 47 bytes removed, diff: -47
/home/acme/git/net-2.6/net/dccp/proto.c:
dccp_destroy_sock | -7
dccp_unhash | -179
dccp_hash | -49
dccp_set_state | -7
dccp_done | +1
5 functions changed, 1 bytes added, 242 bytes removed, diff: -241
/home/acme/git/net-2.6/net/dccp/ipv4.c:
dccp_v4_get_port | -31
dccp_v4_request_recv_sock | -2
2 functions changed, 33 bytes removed, diff: -33
/home/acme/git/net-2.6/net/dccp/ipv6.c:
dccp_v6_get_port | -31
dccp_v6_hash | -7
dccp_v6_request_recv_sock | +5
3 functions changed, 5 bytes added, 38 bytes removed, diff: -33
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add a net argument to inet6_lookup and propagate it further.
Actually, this is tcp-v6 implementation of what was done for
tcp-v4 sockets in a previous patch.
Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
These two functions are the same except for what they call
to "check_established" and "hash" for a socket.
This saves half-a-kilo for ipv4 and ipv6.
add/remove: 1/0 grow/shrink: 1/4 up/down: 582/-1128 (-546)
function old new delta
__inet_hash_connect - 577 +577
arp_ignore 108 113 +5
static.hint 8 4 -4
rt_worker_func 376 372 -4
inet6_hash_connect 584 25 -559
inet_hash_connect 586 25 -561
Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
We have INET_MATCH, INET_TW_MATCH and INET6_MATCH to test sockets and
twbuckets for matching, but ipv6 twbuckets are tested manually.
Here's the INET6_TW_MATCH to help with it.
Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
1) Cleanups (all functions are prefixed by sock_prot_inuse)
sock_prot_inc_use(prot) -> sock_prot_inuse_add(prot,-1)
sock_prot_dec_use(prot) -> sock_prot_inuse_add(prot,-1)
sock_prot_inuse() -> sock_prot_inuse_get()
New functions :
sock_prot_inuse_init() and sock_prot_inuse_free() to abstract pcounter use.
2) if CONFIG_PROC_FS=n, we can zap 'inuse' member from "struct proto",
since nobody wants to read the inuse value.
This saves 1372 bytes on i386/SMP and some cpu cycles.
Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When looking for a conflicting connection the !sk->sk_bound_dev_if
check is performed only for live sockets, but not for timewait-ed.
This is not the case for ipv4, for __inet6_lookup_established in
both ipv4 and ipv6 and for other places that check for tw-s.
Was this missed accidentally? If so, then this patch fixes it and
besides makes use if the dif variable declared in the function.
Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
As done two years ago on IP route cache table (commit
22c047ccbc) , we can avoid using one
lock per hash bucket for the huge TCP/DCCP hash tables.
On a typical x86_64 platform, this saves about 2MB or 4MB of ram, for
litle performance differences. (we hit a different cache line for the
rwlock, but then the bucket cache line have a better sharing factor
among cpus, since we dirty it less often). For netstat or ss commands
that want a full scan of hash table, we perform fewer memory accesses.
Using a 'small' table of hashed rwlocks should be more than enough to
provide correct SMP concurrency between different buckets, without
using too much memory. Sizing of this table depends on
num_possible_cpus() and various CONFIG settings.
This patch provides some locking abstraction that may ease a future
work using a different model for TCP/DCCP table.
Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Acked-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
There is a justifying patch for Stephen's patches. Stephen's patches
disallows using a port range of one single port and brakes the meaning
of the 'remaining' variable, in some places it has different meaning.
My patch gives back the sense of 'remaining' variable. It should mean
how many ports are remaining and nothing else. Also my patch allows
using a single port.
I sure we must be able to use mentioned port range, this does not
restricted by documentation and does not brake current behavior.
usefull links:
Patches posted by Stephen Hemminger
http://marc.info/?l=linux-netdev&m=119206106218187&w=2http://marc.info/?l=linux-netdev&m=119206109918235&w=2
Andrew Morton's comment
http://marc.info/?l=linux-kernel&m=119248225007737&w=2
1. Allows using a port range of one single port.
2. Gives back sense of 'remaining' variable.
Signed-off-by: Anton Arapov <aarapov@redhat.com>
Acked-by: Stephen Hemminger <shemminger@linux-foundation.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Expansion of original idea from Denis V. Lunev <den@openvz.org>
Add robustness and locking to the local_port_range sysctl.
1. Enforce that low < high when setting.
2. Use seqlock to ensure atomic update.
The locking might seem like overkill, but there are
cases where sysadmin might want to change value in the
middle of a DoS attack.
Signed-off-by: Stephen Hemminger <shemminger@linux-foundation.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Tetsuo Handa <handat@pm.nttdata.co.jp> told me that connect(2) with TCPv6
socket almost always took a few minutes to return when we did not have any
ports available in the range of net.ipv4.ip_local_port_range.
The reason was that we used incorrect seed for calculating index of
hash when we check established sockets in __inet6_check_established().
Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
ehash table layout is currently this one :
First half of this table is used by sockets not in TIME_WAIT state
Second half of it is used by sockets in TIME_WAIT state.
This is non optimal because of for a given hash or socket, the two chain heads
are located in separate cache lines.
Moreover the locks of the second half are never used.
If instead of this halving, we use two list heads in inet_ehash_bucket instead
of only one, we probably can avoid one cache miss, and reduce ram usage,
particularly if sizeof(rwlock_t) is big (various CONFIG_DEBUG_SPINLOCK,
CONFIG_DEBUG_LOCK_ALLOC settings). So we still halves the table but we keep
together related chains to speedup lookups and socket state change.
In this patch I did not try to align struct inet_ehash_bucket, but a future
patch could try to make this structure have a convenient size (a power of two
or a multiple of L1_CACHE_SIZE).
I guess rwlock will just vanish as soon as RCU is plugged into ehash :) , so
maybe we dont need to scratch our heads to align the bucket...
Note : In case struct inet_ehash_bucket is not a power of two, we could
probably change alloc_large_system_hash() (in case it use __get_free_pages())
to free the unused space. It currently allocates a big zone, but the last
quarter of it could be freed. Again, this should be a temporary 'problem'.
Patch tested on ipv4 tcp only, but should be OK for IPV6 and DCCP.
Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
INET_MATCH() and friends depend on an interesting set of kludges:
* there's a pair of adjacent fields in struct inet_sock - __be16 dport
followed by __u16 num. We want to search by pair, so we combine the keys into
a single 32bit value and compare with 32bit value read from &...->dport.
* on 64bit targets we combine comparisons with pair of adjacent __be32
fields in the same way.
Make sure that we don't mix those values with anything else and that pairs
we form them from have correct types.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
Deinline a few functions which produce 200+ bytes of code.
Size Uses Wasted Name and definition
===== ==== ====== ================================================
429 3 818 __inet6_lookup include/net/inet6_hashtables.h
404 2 384 __inet6_lookup_established include/net/inet6_hashtables.h
206 3 372 __inet6_hash include/net/inet6_hashtables.h
Signed-off-by: Denis Vlasenko <vda@ilport.com.ua>
Signed-off-by: David S. Miller <davem@davemloft.net>
When we link a socket into the hash table, we need to make sure that we
set the num/port fields so that it shows us with a non-zero port value
in proc/netlink and on the wire. This code and comment is copied over
from the IPv4 stack as is.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Renaming it to inet6_hash_connect, making it possible to ditch
dccp_v6_hash_connect and share the same code with TCP instead.
Signed-off-by: Arnaldo Carvalho de Melo <acme@mandriva.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Doing this we allow tcp_diag to support IPV6 even if tcp_diag is compiled
statically and IPV6 is compiled as a module, removing the previous restriction
while not building any IPV6 code if it is not selected.
Now to work on the tcpdiag_register infrastructure and then to rename the whole
thing to inetdiag, reflecting its by then completely generic nature.
Signed-off-by: Arnaldo Carvalho de Melo <acme@mandriva.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
