Define a new function to return the waitqueue of a "struct sock".
static inline wait_queue_head_t *sk_sleep(struct sock *sk)
{
return sk->sk_sleep;
}
Change all read occurrences of sk_sleep by a call to this function.
Needed for a future RCU conversion. sk_sleep wont be a field directly
available.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since netdev_chain is guarded by rtnl_lock, ASSERT_RTNL should be
present here to make sure that all callers of call_netdevice_notifiers
does the locking properly.
Signed-off-by: Jiri Pirko <jpirko@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In case we compute a software skb->rxhash, we can generate a consistent
hash : Its value will be the same in both flow directions.
This helps some workloads, like conntracking, since the same state needs
to be accessed in both directions.
tbench + RFS + this patch gives better results than tbench with default
kernel configuration (no RPS, no RFS)
Also fixed some sparse warnings.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
struct softnet_data holds many queues, so consistent use "sd" name
instead of "queue" is better.
Adds a rps_ipi_queued() helper to cleanup enqueue_to_backlog()
Adds a _and_irq_disable suffix to net_rps_action() name, as David
suggested.
incr_input_queue_head() becomes input_queue_head_incr()
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
store_rps_map() & store_rps_dev_flow_table_cnt() are static.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
net_rps_action() is a bit expensive on NR_CPUS=64..4096 kernels, even if
RPS is not active.
Tom Herbert used two bitmasks to hold information needed to send IPI,
but a single LIFO list seems more appropriate.
Move all RPS logic into net_rps_action() to cleanup net_rx_action() code
(remove two ifdefs)
Move rps_remote_softirq_cpus into softnet_data to share its first cache
line, filling an existing hole.
In a future patch, we could call net_rps_action() from process_backlog()
to make sure we send IPI before handling this cpu backlog.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Transmitted skb might be attached to a socket and a destructor, for
memory accounting purposes.
Traditionally, this destructor is called at tx completion time, when skb
is freed.
When tx completion is performed by another cpu than the sender, this
forces some cache lines to change ownership. XPS was an attempt to give
tx completion to initial cpu.
David idea is to call destructor right before giving skb to device (call
to ndo_start_xmit()). Because device queues are usually small, orphaning
skb before tx completion is not a big deal. Some drivers already do
this, we could do it in upper level.
There is one known exception to this early orphaning, called tx
timestamping. It needs to keep a reference to socket until device can
give a hardware or software timestamp.
This patch adds a skb_orphan_try() helper, to centralize all exceptions
to early orphaning in one spot, and use it in dev_hard_start_xmit().
"tbench 16" results on a Nehalem machine (2 X5570 @ 2.93GHz)
before: Throughput 4428.9 MB/sec 16 procs
after: Throughput 4448.14 MB/sec 16 procs
UDP should get even better results, its destructor being more complex,
since SOCK_USE_WRITE_QUEUE is not set (four atomic ops instead of one)
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
- There is no point to enforce a time limit in process_backlog(), since
other napi instances dont follow same rule. We can exit after only one
packet processed...
The normal quota of 64 packets per napi instance should be the norm, and
net_rx_action() already has its own time limit.
Note : /proc/net/core/dev_weight can be used to tune this 64 default
value.
- Use DEFINE_PER_CPU_ALIGNED for softnet_data definition.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch implements receive flow steering (RFS). RFS steers
received packets for layer 3 and 4 processing to the CPU where
the application for the corresponding flow is running. RFS is an
extension of Receive Packet Steering (RPS).
The basic idea of RFS is that when an application calls recvmsg
(or sendmsg) the application's running CPU is stored in a hash
table that is indexed by the connection's rxhash which is stored in
the socket structure. The rxhash is passed in skb's received on
the connection from netif_receive_skb. For each received packet,
the associated rxhash is used to look up the CPU in the hash table,
if a valid CPU is set then the packet is steered to that CPU using
the RPS mechanisms.
The convolution of the simple approach is that it would potentially
allow OOO packets. If threads are thrashing around CPUs or multiple
threads are trying to read from the same sockets, a quickly changing
CPU value in the hash table could cause rampant OOO packets--
we consider this a non-starter.
To avoid OOO packets, this solution implements two types of hash
tables: rps_sock_flow_table and rps_dev_flow_table.
rps_sock_table is a global hash table. Each entry is just a CPU
number and it is populated in recvmsg and sendmsg as described above.
This table contains the "desired" CPUs for flows.
rps_dev_flow_table is specific to each device queue. Each entry
contains a CPU and a tail queue counter. The CPU is the "current"
CPU for a matching flow. The tail queue counter holds the value
of a tail queue counter for the associated CPU's backlog queue at
the time of last enqueue for a flow matching the entry.
Each backlog queue has a queue head counter which is incremented
on dequeue, and so a queue tail counter is computed as queue head
count + queue length. When a packet is enqueued on a backlog queue,
the current value of the queue tail counter is saved in the hash
entry of the rps_dev_flow_table.
And now the trick: when selecting the CPU for RPS (get_rps_cpu)
the rps_sock_flow table and the rps_dev_flow table for the RX queue
are consulted. When the desired CPU for the flow (found in the
rps_sock_flow table) does not match the current CPU (found in the
rps_dev_flow table), the current CPU is changed to the desired CPU
if one of the following is true:
- The current CPU is unset (equal to RPS_NO_CPU)
- Current CPU is offline
- The current CPU's queue head counter >= queue tail counter in the
rps_dev_flow table. This checks if the queue tail has advanced
beyond the last packet that was enqueued using this table entry.
This guarantees that all packets queued using this entry have been
dequeued, thus preserving in order delivery.
Making each queue have its own rps_dev_flow table has two advantages:
1) the tail queue counters will be written on each receive, so
keeping the table local to interrupting CPU s good for locality. 2)
this allows lockless access to the table-- the CPU number and queue
tail counter need to be accessed together under mutual exclusion
from netif_receive_skb, we assume that this is only called from
device napi_poll which is non-reentrant.
This patch implements RFS for TCP and connected UDP sockets.
It should be usable for other flow oriented protocols.
There are two configuration parameters for RFS. The
"rps_flow_entries" kernel init parameter sets the number of
entries in the rps_sock_flow_table, the per rxqueue sysfs entry
"rps_flow_cnt" contains the number of entries in the rps_dev_flow
table for the rxqueue. Both are rounded to power of two.
The obvious benefit of RFS (over just RPS) is that it achieves
CPU locality between the receive processing for a flow and the
applications processing; this can result in increased performance
(higher pps, lower latency).
The benefits of RFS are dependent on cache hierarchy, application
load, and other factors. On simple benchmarks, we don't necessarily
see improvement and sometimes see degradation. However, for more
complex benchmarks and for applications where cache pressure is
much higher this technique seems to perform very well.
Below are some benchmark results which show the potential benfit of
this patch. The netperf test has 500 instances of netperf TCP_RR
test with 1 byte req. and resp. The RPC test is an request/response
test similar in structure to netperf RR test ith 100 threads on
each host, but does more work in userspace that netperf.
e1000e on 8 core Intel
No RFS or RPS 104K tps at 30% CPU
No RFS (best RPS config): 290K tps at 63% CPU
RFS 303K tps at 61% CPU
RPC test tps CPU% 50/90/99% usec latency Latency StdDev
No RFS/RPS 103K 48% 757/900/3185 4472.35
RPS only: 174K 73% 415/993/2468 491.66
RFS 223K 73% 379/651/1382 315.61
Signed-off-by: Tom Herbert <therbert@google.com>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
As Herbert Xu said: we should be able to simply replace ipfragok
with skb->local_df. commit f88037(sctp: Drop ipfargok in sctp_xmit function)
has droped ipfragok and set local_df value properly.
The patch kills the ipfragok parameter of .queue_xmit().
Signed-off-by: Shan Wei <shanwei@cn.fujitsu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
commit f88037(sctp: Drop ipfargok in sctp_xmit function)
has droped ipfragok and set local_df value properly.
So the change of commit 77e2f1(ipv6: Fix ip6_xmit to
send fragments if ipfragok is true) is not needed.
So the patch remove them.
Signed-off-by: Shan Wei <shanwei@cn.fujitsu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Fix an oversight in ipmr_destroy_unres() - the net pointer is
unconditionally initialized to NULL, resulting in a NULL pointer
dereference later on.
Fix by adding a net pointer to struct mr_table and using it in
ipmr_destroy_unres().
Signed-off-by: Patrick McHardy <kaber@trash.net>
The patch to convert struct mfc_cache to list_heads (ipv4: ipmr: convert
struct mfc_cache to struct list_head) introduced a bug when adding new
cache entries that don't match any unresolved entries.
The unres queue is searched for a matching entry, which is then resolved.
When no matching entry is present, the iterator points to the head of the
list, but is treated as a matching entry. Use a seperate variable to
indicate that a matching entry was found.
Signed-off-by: Patrick McHardy <kaber@trash.net>
Eric Paris reported netif_rx() is calling smp_processor_id() from
preemptible context, in particular when caller is
ip_dev_loopback_xmit().
RPS commit added this smp_processor_id() call, this patch makes sure
preemption is disabled. rps_get_cpus() wants rcu_read_lock() anyway, we
can dot it a bit earlier.
Reported-by: Eric Paris <eparis@redhat.com>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch adds support for multiple independant multicast routing instances,
named "tables".
Userspace multicast routing daemons can bind to a specific table instance by
issuing a setsockopt call using a new option MRT_TABLE. The table number is
stored in the raw socket data and affects all following ipmr setsockopt(),
getsockopt() and ioctl() calls. By default, a single table (RT_TABLE_DEFAULT)
is created with a default routing rule pointing to it. Newly created pimreg
devices have the table number appended ("pimregX"), with the exception of
devices created in the default table, which are named just "pimreg" for
compatibility reasons.
Packets are directed to a specific table instance using routing rules,
similar to how regular routing rules work. Currently iif, oif and mark
are supported as keys, source and destination addresses could be supported
additionally.
Example usage:
- bind pimd/xorp/... to a specific table:
uint32_t table = 123;
setsockopt(fd, IPPROTO_IP, MRT_TABLE, &table, sizeof(table));
- create routing rules directing packets to the new table:
# ip mrule add iif eth0 lookup 123
# ip mrule add oif eth0 lookup 123
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Now that cache entries in unres_queue don't need to be distinguished by their
network namespace pointer anymore, we can remove it from struct mfc_cache
add pass the namespace as function argument to the functions that need it.
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
The unres_queue is currently shared between all namespaces. Following patches
will additionally allow to create multiple multicast routing tables in each
namespace. Having a single shared queue for all these users seems to excessive,
move the queue and the cleanup timer to the per-namespace data to unshare it.
As a side-effect, this fixes a bug in the seq file iteration functions: the
first entry returned is always from the current namespace, entries returned
after that may belong to any namespace.
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Decouple the address family values used for fib_rules from the real
address families in socket.h. This allows to use fib_rules for
code that is not a real address family without increasing AF_MAX/NPROTO.
Values up to 127 are reserved for real address families and map directly
to the corresponding AF value, values starting from 128 are for other
uses. rtnetlink is changed to invoke the AF_UNSPEC dumpit/doit handlers
for these families.
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
All fib_rules implementations need to set the family in their ->fill()
functions. Since the value is available to the generic fib_nl_fill_rule()
function, set it there.
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Both functions are equivalent, consolidate them since a following patch
needs a third implementation for multicast routing.
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
The function dst_ifdown is called only two places but in a non-
performance critical code path, there is no reason to inline it.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
skb_bond_should_drop() is too big to be inlined.
This patch reduces kernel text size, and its compilation time as well
(shrinking include/linux/netdevice.h)
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
At this point optlen == sizeof(sfilter) but some compilers are dumb.
Reported-by: Németh Márton <nm127@freemail.h
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Acked-by: Oliver Hartkopp <oliver@hartkopp.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
The notifier for address down should only be called if address is completely
gone, not just being marked as tentative on link transistion. The code
in net-next would case bonding/sctp/s390 to see address disappear on link
down, but they would never see it reappear on link up.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since an address in hash list has to already have a ref count,
no additional ref count is needed.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When link goes down, want address to be preserved but in a tentative
state, therefore it has to stay in hash list.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Recent changes preserve IPv6 address when link goes down (good).
But would cause address to point to dead dst entry (bad).
The simplest fix is to just not delete route if address is
being held for later use.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
With latest CONFIG_PROVE_RCU stuff, I felt more comfortable to make this
work.
sk->sk_dst_cache is currently protected by a rwlock (sk_dst_lock)
This rwlock is readlocked for a very small amount of time, and dst
entries are already freed after RCU grace period. This calls for RCU
again :)
This patch converts sk_dst_lock to a spinlock, and use RCU for readers.
__sk_dst_get() is supposed to be called with rcu_read_lock() or if
socket locked by user, so use appropriate rcu_dereference_check()
condition (rcu_read_lock_held() || sock_owned_by_user(sk))
This patch avoids two atomic ops per tx packet on UDP connected sockets,
for example, and permits sk_dst_lock to be much less dirtied.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Dont use netdev_warn() in dev_cap_txqueue() and get_rps_cpu() so that we
can catch following warnings without crash.
bond0.2240 received packet on queue 6, but number of RX queues is 1
bond0.2240 received packet on queue 11, but number of RX queues is 1
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Enable the SO_TIMESTAMPING socket infrastructure for raw packet sockets.
We introduce PACKET_TX_TIMESTAMP for the control message cmsg_type.
Similar support for UDP and CAN sockets was added in commit
51f31cabe3
Signed-off-by: Richard Cochran <richard.cochran@omicron.at>
Signed-off-by: David S. Miller <davem@davemloft.net>
Back in commit 04a0551c87
("loopback: Drop obsolete ip_summed setting") we stopped
setting CHECKSUM_UNNECESSARY in the loopback xmit.
This is because such a setting was a lie since it implies that the
checksum field of the packet is properly filled in.
Instead what happens normally is that CHECKSUM_PARTIAL is set and
skb->csum is calculated as needed.
But this was only happening for TCP data packets (via the
skb->ip_summed assignment done in tcp_sendmsg()). It doesn't
happen for non-data packets like ACKs etc.
Fix this by setting skb->ip_summed in the common non-data packet
constructor. It already is setting skb->csum to zero.
But this reminds us that we still have things like ip_output.c's
ip_dev_loopback_xmit() which sets skb->ip_summed to the value
CHECKSUM_UNNECESSARY, which Herbert's patch teaches us is not
valid. So we'll have to address that at some point too.
Signed-off-by: David S. Miller <davem@davemloft.net>
inet: Remove unused send_check length argument
This patch removes the unused length argument from the send_check
function in struct inet_connection_sock_af_ops.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Tested-by: Yinghai <yinghai.lu@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
tcp: Handle CHECKSUM_PARTIAL for SYNACK packets for IPv6
This patch moves the common code between tcp_v6_send_check and
tcp_v6_gso_send_check into a new function __tcp_v6_send_check.
It then uses the new function in tcp_v6_send_synack as well as
tcp_v6_send_response so that they handle CHECKSUM_PARTIAL properly.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Tested-by: Yinghai <yinghai.lu@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
tcp: Handle CHECKSUM_PARTIAL for SYNACK packets for IPv4
This patch moves the common code between tcp_v4_send_check and
tcp_v4_gso_send_check into a new function __tcp_v4_send_check.
It then uses the new function in tcp_v4_send_synack so that it
handles CHECKSUM_PARTIAL properly.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Tested-by: Yinghai <yinghai.lu@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This reverts commit 2626419ad5.
It causes regressions for people with IGB cards. Connection
requests don't complete etc. The true cause of the issue is
still not known, but we should sort this out in net-next-2.6
not net-2.6
Signed-off-by: David S. Miller <davem@davemloft.net>
As scan_work is queued from work_work it needs to be checked if scan
has been started during execution of work_work. Otherwise, when hw
scan is used, the stack gets error about hw being busy with ongoing
scan. This causes the stack to abort scan without notifying the driver
about it. This leads to a situation where the hw is scanning and the stack
thinks it's not. Then when the scan finishes, the stack will complain by
warnings.
Signed-off-by: Teemu Paasikivi <ext-teemu.3.paasikivi@nokia.com>
Reviewed-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
We delay the skb linearising in ieee80211_rx_h_decrypt so that
frames do not require software decryption are not linearized. We
are safe to do this because ieee80211_get_mmie_keyidx() only
requires to touch nonlinear data for management frames, which are
already linearized before getting here.
Cc: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Zhu Yi <yi.zhu@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Eric Dumazet