This patch improves ctnetlink event reliability if one broadcast
listener has set the NETLINK_BROADCAST_ERROR socket option.
The logic is the following: if an event delivery fails, we keep
the undelivered events in the missed event cache. Once the next
packet arrives, we add the new events (if any) to the missed
events in the cache and we try a new delivery, and so on. Thus,
if ctnetlink fails to deliver an event, we try to deliver them
once we see a new packet. Therefore, we may lose state
transitions but the userspace process gets in sync at some point.
At worst case, if no events were delivered to userspace, we make
sure that destroy events are successfully delivered. Basically,
if ctnetlink fails to deliver the destroy event, we remove the
conntrack entry from the hashes and we insert them in the dying
list, which contains inactive entries. Then, the conntrack timer
is added with an extra grace timeout of random32() % 15 seconds
to trigger the event again (this grace timeout is tunable via
/proc). The use of a limited random timeout value allows
distributing the "destroy" resends, thus, avoiding accumulating
lots "destroy" events at the same time. Event delivery may
re-order but we can identify them by means of the tuple plus
the conntrack ID.
The maximum number of conntrack entries (active or inactive) is
still handled by nf_conntrack_max. Thus, we may start dropping
packets at some point if we accumulate a lot of inactive conntrack
entries that did not successfully report the destroy event to
userspace.
During my stress tests consisting of setting a very small buffer
of 2048 bytes for conntrackd and the NETLINK_BROADCAST_ERROR socket
flag, and generating lots of very small connections, I noticed
very few destroy entries on the fly waiting to be resend.
A simple way to test this patch consist of creating a lot of
entries, set a very small Netlink buffer in conntrackd (+ a patch
which is not in the git tree to set the BROADCAST_ERROR flag)
and invoke `conntrack -F'.
For expectations, no changes are introduced in this patch.
Currently, event delivery is only done for new expectations (no
events from expectation expiration, removal and confirmation).
In that case, they need a per-expectation event cache to implement
the same idea that is exposed in this patch.
This patch can be useful to provide reliable flow-accouting. We
still have to add a new conntrack extension to store the creation
and destroy time.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Patrick McHardy <kaber@trash.net>
This patch moves the helper destruction to a function that lives
in nf_conntrack_helper.c. This new function is used in the patch
to add ctnetlink reliable event delivery.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Patrick McHardy <kaber@trash.net>
This patch reworks the per-cpu event caching to use the conntrack
extension infrastructure.
The main drawback is that we consume more memory per conntrack
if event delivery is enabled. This patch is required by the
reliable event delivery that follows to this patch.
BTW, this patch allows you to enable/disable event delivery via
/proc/sys/net/netfilter/nf_conntrack_events in runtime, although
you can still disable event caching as compilation option.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Patrick McHardy <kaber@trash.net>
Use mod_timer_pending() instead of atomic sequence of del_timer()/
add_timer(). mod_timer_pending() does not rearm an inactive timer,
so we don't need the conntrack lock anymore to make sure we don't
accidentally rearm a timer of a conntrack which is in the process
of being destroyed.
With this change, we don't need to take the global lock anymore at all,
counter updates can be performed under the per-conntrack lock.
Signed-off-by: Patrick McHardy <kaber@trash.net>
Convert magic values 1 and -1 to NETDEV_TX_BUSY and NETDEV_TX_LOCKED respectively.
0 (NETDEV_TX_OK) is not changed to keep the noise down, except in very few cases
where its in direct proximity to one of the other values.
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Fix up ATM drivers that return an errno value to qdisc_restart(), causing
qdisc_restart() to print a warning an requeue/retransmit the skb.
- lec: condition can only be remedied by userspace, until that retransmissions
Compile tested only.
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch changes FDB entry check for ATM LANE bridge integration.
There's no point in holding a FDB entry around SKB building.
br_fdb_get()/br_fdb_put() pair are changed into single br_fdb_test_addr()
hook that checks if the addr has FDB entry pointing to other port
to the one the request arrived on.
FDB entry refcounting is removed as it's not used anywhere else.
Signed-off-by: Michał Mirosław <mirq-linux@rere.qmqm.pl>
Acked-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Commit b00055aacd " [NET] core: add
RFC2863 operstate" defined new interface flag values. Its
documentation specified that these flags could be accessed from user
space via SIOCGIFFLAGS. However, this does not work because the new
flags do not fit in that ioctl's argument width.
Change the documentation to match the code's behavior. Also change
the source to explicitly show the truncation. This _should_ have no
effect on executable code, and did not with gcc 4.2.4 generating x86
code.
A new ioctl could be defined to return all interface flags to user
space. However, since this has been broken for three years with no
one complaining, there doesn't seem much need. They are still
accessible via netlink.
Reported-by: "Fredrik Arnerup" <fredrik.arnerup@edgeware.tv>
Signed-off-by: John Dykstra <john.dykstra1@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Fix build error introduced by commit bb70dfa5 (netfilter: xtables:
consolidate comefrom debug cast access):
net/ipv4/netfilter/ip_tables.c: In function 'ipt_do_table':
net/ipv4/netfilter/ip_tables.c:421: error: 'comefrom' undeclared (first use in this function)
net/ipv4/netfilter/ip_tables.c:421: error: (Each undeclared identifier is reported only once
net/ipv4/netfilter/ip_tables.c:421: error: for each function it appears in.)
Signed-off-by: Patrick McHardy <kaber@trash.net>
Caused by an API update. The return value can be safely ignored, as
there is notthing we can do with it.
Signed-off-by: Inaky Perez-Gonzalez <inaky@linux.intel.com>
The current code errors out the INCOMPLETE neigh entry skb queue only from
the timer if maximum probes have been attempted and there has been no reply.
This also causes the transtion to FAILED state.
However, the neigh entry can be also updated via Netlink to inform that the
address is unavailable. Currently, neigh_update() just stops the timers and
leaves the pending skb's unreleased. This results that the clean up code in
the timer callback is never called, preventing also proper garbage collection.
This fixes neigh_update() to process the pending skb queue immediately if
INCOMPLETE -> FAILED state transtion occurs due to a Netlink request.
Signed-off-by: Timo Teras <timo.teras@iki.fi>
Signed-off-by: David S. Miller <davem@davemloft.net>
One of the problem with sock memory accounting is it uses
a pair of sock_hold()/sock_put() for each transmitted packet.
This slows down bidirectional flows because the receive path
also needs to take a refcount on socket and might use a different
cpu than transmit path or transmit completion path. So these
two atomic operations also trigger cache line bounces.
We can see this in tx or tx/rx workloads (media gateways for example),
where sock_wfree() can be in top five functions in profiles.
We use this sock_hold()/sock_put() so that sock freeing
is delayed until all tx packets are completed.
As we also update sk_wmem_alloc, we could offset sk_wmem_alloc
by one unit at init time, until sk_free() is called.
Once sk_free() is called, we atomic_dec_and_test(sk_wmem_alloc)
to decrement initial offset and atomicaly check if any packets
are in flight.
skb_set_owner_w() doesnt call sock_hold() anymore
sock_wfree() doesnt call sock_put() anymore, but check if sk_wmem_alloc
reached 0 to perform the final freeing.
Drawback is that a skb->truesize error could lead to unfreeable sockets, or
even worse, prematurely calling __sk_free() on a live socket.
Nice speedups on SMP. tbench for example, going from 2691 MB/s to 2711 MB/s
on my 8 cpu dev machine, even if tbench was not really hitting sk_refcnt
contention point. 5 % speedup on a UDP transmit workload (depends
on number of flows), lowering TX completion cpu usage.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
rfkill currently requires a global lock within the
rfkill_register() function, and holds that lock over
calls to the set_block() methods. This means that we
cannot hold a lock around rfkill_register() that we
also require in set_block(), directly or indirectly.
Fix cfg80211 to register rfkill outside the block
locked by its global lock. Much of what cfg80211 does
in the locked block doesn't need to be locked anyway.
Reported-by: Vasanthakumar Thiagarajan <vasanth@atheros.com>
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
When associated, but probing the AP because we detected
beacon loss, we need to disable powersave to be able to
receive the probe response. Change the code to do that by
checking whether we're trying to probe when determining
the possibility of going into PS, and recalculate the PS
ability at the necessary spots.
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
We don't want to trigger moving between PS mode during scan,
because then we will sometimes end up sending nullfunc frames
during scan. We're supposed to only send one prior to scan
and after scan.
This fixes an oops which occured due to an assert in ath9k:
http://marc.info/?l=linux-wireless&m=124277331319024
The assert was happening because the rate control algorithm
figures it should find at least one valid dual stream or
single stream rate. Since we allow mac80211 to send nullfunc
frames during scan and dynamic PS was enabled at times we ended
up trying to send nullfunc frames for the target sta on the
wrong band for which we have no valid rate to communicate with
it. This breaks the assumptions in rate control.
We determine we also need to disable moving between PS modes
when not associated so lets just add that now as well, and we
should not have a ps_sdata when that interface cannot actually
go into PS because it's not associated.
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
The return type has more than two values, but it can validly
only ever return TX_DROP and TX_CONTINUE, so use a bool
instead of ieee80211_tx_result.
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
addba_req_num[tid] is supposed to have the count of consecutive
addba request attempts on 'tid' which failed. This count is checked
against a retry threshold (3 times) before starting the addba negotiation.
This patch fixes the way this addba count is incremented/reset and thereby
avoids indefinite addba attempts.
Signed-off-by: Vasanthakumar Thiagarajan <vasanth@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Once rfkill-input is disabled, the "global" states will only be used as
default initial states.
Since the states will always be the same after resume, we shouldn't
generate events on resume.
Signed-off-by: Alan Jenkins <alan-jenkins@tuffmail.co.uk>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
rfkill_set_global_sw_state() (previously rfkill_set_default()) will no
longer be exported by the rewritten rfkill core.
Instead, platform drivers which can provide persistent soft-rfkill state
across power-down/reboot should indicate their initial state by calling
rfkill_set_sw_state() before registration. Otherwise, they will be
initialized to a default value during registration by a set_block call.
We remove existing calls to rfkill_set_sw_state() which happen before
registration, since these had no effect in the old model. If these
drivers do have persistent state, the calls can be put back (subject
to testing :-). This affects hp-wmi and acer-wmi.
Drivers with persistent state will affect the global state only if
rfkill-input is enabled. This is required, otherwise booting with
wireless soft-blocked and pressing the wireless-toggle key once would
have no apparent effect. This special case will be removed in future
along with rfkill-input, in favour of a more flexible userspace daemon
(see Documentation/feature-removal-schedule.txt).
Now rfkill_global_states[n].def is only used to preserve global states
over EPO, it is renamed to ".sav".
Signed-off-by: Alan Jenkins <alan-jenkins@tuffmail.co.uk>
Acked-by: Henrique de Moraes Holschuh <hmh@hmh.eng.br>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
In order to handle powersave frames properly we had needed
to pass these out to the device queues again, and introduce
the skb->requeue bit. This, however, also has unnecessary
overhead by needing to 'clean up' already tried frames, and
this clean-up code is also buggy when software encryption
is used.
Instead of sending the frames via the master netdev queue
again, simply put them into the pending queue. This also
fixes a problem where frames for that particular station
could be reordered when some were still on the software
queues and older ones are re-injected into the software
queue after them.
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Now that we added the ioctl, there's no need to ask
the user to configure this. We will keep it enabled
for now, and eventually swap the default to n. Also
let embedded users select it only if they need it.
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
It is useful for debugging when we know if something disabled
the in-kernel rfkill input handler.
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Acked-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
mac80211 is checking is the skb is aligned on 32 bit boundary.
But it is checking against ethernet header, whereas Linux expect IP
header aligned. And ethernet ether size is 6*2+2=14, so aligning
ethernet header make IP header unaligned.
Signed-off-by: Matthieu CASTET <castet.matthieu@free.fr>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
The minstrel rate controller periodically looks up rate indexes in
a sampling table. When accessing a specific row and column, minstrel
correctly does a bounds check which, on the surface, appears to handle
the case where mi->n_rates < 2. However, mi->sample_idx is actually
defined as an unsigned, so the right hand side is taken to be a huge
positive number when negative, and the check will always fail.
Consequently, the RC will overrun the array and cause random memory
corruption when communicating with a peer that has only a single rate.
The max value of mi->sample_idx is around 25 so casting to int should
have no ill effects.
Without the change, uptime is a few minutes under load with an AP
that has a single hard-coded rate, and both the AP and STA could
potentially crash. With the change, both lasted 12 hours with a
steady load.
Thanks to Ognjen Maric for providing the single-rate clue so I could
reproduce this.
This fixes http://bugzilla.kernel.org/show_bug.cgi?id=12490 on the
regression list (also http://bugzilla.kernel.org/show_bug.cgi?id=13000).
Cc: stable@kernel.org
Reported-by: Sergey S. Kostyliov <rathamahata@gmail.com>
Reported-by: Ognjen Maric <ognjen.maric@gmail.com>
Signed-off-by: Bob Copeland <me@bobcopeland.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Introduce per-conntrack locks and use them instead of the global protocol
locks to avoid contention. Especially tcp_lock shows up very high in
profiles on larger machines.
This will also allow to simplify the upcoming reliable event delivery patches.
Signed-off-by: Patrick McHardy <kaber@trash.net>
As the module uses rcu_call() we should make sure that all
rcu callback has been completed before removing the code.
Signed-off-by: Jesper Dangaard Brouer <hawk@comx.dk>
Acked-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
On module unload call rcu_barrier(), this is needed as synchronize_rcu()
is not strong enough. The kmem_cache_destroy() does invoke
synchronize_rcu() but it does not provide same protection.
Signed-off-by: Jesper Dangaard Brouer <hawk@comx.dk>
Acked-by: Vlad Yasevich <vladislav.yasevich@hp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This module uses rcu_call() thus it should use rcu_barrier()
on module unload.
Signed-off-by: Jesper Dangaard Brouer <hawk@comx.dk>
Acked-by: Oliver Hartkopp <oliver@hartkopp.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
This module uses rcu_call() thus it should use rcu_barrier() on module unload.
Also fixed a trivial typo 'nfetlink' -> 'nfnetlink' in comment.
Signed-off-by: Jesper Dangaard Brouer <hawk@comx.dk>
Acked-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Acked-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
The VLAN 8021q driver needs to call rcu_barrier() when unloading the module,
instead of syncronize_net(). This is needed to make sure that outstanding
call_rcu() callbacks have completed, before the callback function code is
removed on module unload.
Signed-off-by: Jesper Dangaard Brouer <hawk@comx.dk>
Reviewed-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Acked-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
If the XT_SOCKET_TRANSPARENT flag is set, enabled 'transparent'
socket option is required for the socket to be matched.
Signed-off-by: Laszlo Attila Toth <panther@balabit.hu>
Signed-off-by: Patrick McHardy <kaber@trash.net>
Add a netlink interface for configuration of IEEE 802.15.4 device. Also this
interface specifies events notification sent by devices towards higher layers.
Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
Signed-off-by: Sergey Lapin <slapin@ossfans.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add support for communication over IEEE 802.15.4 networks. This implementation
is neither certified nor complete, but aims to that goal. This commit contains
only the socket interface for communication over IEEE 802.15.4 networks.
One can either send RAW datagrams or use SOCK_DGRAM to encapsulate data
inside normal IEEE 802.15.4 packets.
Configuration interface, drivers and software MAC 802.15.4 implementation will
follow.
Initial implementation was done by Maxim Gorbachyov, Maxim Osipov and Pavel
Smolensky as a research project at Siemens AG. Later the stack was heavily
reworked to better suit the linux networking model, and is now maitained
as an open project partially sponsored by Siemens.
Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
Signed-off-by: Sergey Lapin <slapin@ossfans.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
IEEE 802.15.4 stack requires several constants to be defined/adjusted.
Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
Signed-off-by: Sergey Lapin <slapin@ossfans.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use PSCHED_SHIFT constant instead of '10' in PSCHED_US2NS() and
PSCHED_NS2US() macros to enable changing this value later.
Additionally use PSCHED_SHIFT in sch_hfsc SM_SHIFT and ISM_SHIFT
definitions. This part of the patch is based on feedback from
Patrick McHardy <kaber@trash.net>.
Reported-by: Antonio Almeida <vexwek@gmail.com>
Tested-by: Antonio Almeida <vexwek@gmail.com>
Signed-off-by: Jarek Poplawski <jarkao2@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
commit f001fde5ea
(net: introduce a list of device addresses dev_addr_list (v6))
added one regression Vegard Nossum found in its testings.
With kmemcheck help, Vegard found some uninitialized memory
was read and reported to user, potentialy leaking kernel data.
( thread can be found on http://lkml.org/lkml/2009/5/30/177 )
dev_addr_init() incorrectly uses sizeof() operator. We were
initializing one byte instead of MAX_ADDR_LEN bytes.
Reported-by: Vegard Nossum <vegard.nossum@gmail.com>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Acked-by: Jiri Pirko <jpirko@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
