Commit graph

9380 commits

Author SHA1 Message Date
David S. Miller
15b458fa65 pkt_sched: Kill qdisc_lock_tree usage in cls_route.c
It just wants the qdisc tree to be synchronized, so grabbing
qdisc_root_lock() is sufficient.

Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-17 19:21:25 -07:00
David S. Miller
55dbc640c3 pkt_sched: Remove qdisc_lock_tree usage in cls_api.c
It just wants the qdisc tree for the filter to be synchronized.
So just BH lock qdisc_root_lock(q) instead.

Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-17 19:21:24 -07:00
David S. Miller
17715e62a5 pkt_sched: Use per-queue locking in shutdown_scheduler_queue.
This eliminates another qdisc_lock_tree user.

Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-17 19:21:23 -07:00
David S. Miller
8a34c5dc3a pkt_sched: Perform bulk of qdisc destruction in RCU.
This allows less strict control of access to the qdisc attached to a
netdev_queue.  It is even allowed to enqueue into a qdisc which is
in the process of being destroyed.  The RCU handler will toss out
those packets.

We will need this to handle sharing of a qdisc amongst multiple
TX queues.  In such a setup the lock has to be shared, so will
be inside of the qdisc itself.  At which point the netdev_queue
lock cannot be used to hard synchronize access to the ->qdisc
pointer.

One operation we have to keep inside of qdisc_destroy() is the list
deletion.  It is the only piece of state visible after the RCU quiesce
period, so we have to undo it early and under the appropriate locking.

The operations in the RCU handler do not need any looking because the
qdisc tree is no longer visible to anything at that point.

Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-17 19:21:22 -07:00
David S. Miller
16361127eb pkt_sched: dev_init_scheduler() does not need to lock qdisc tree.
We are registering the device, there is no way anyone can get
at this object's qdiscs yet in any meaningful way.

Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-17 19:21:21 -07:00
David S. Miller
37437bb2e1 pkt_sched: Schedule qdiscs instead of netdev_queue.
When we have shared qdiscs, packets come out of the qdiscs
for multiple transmit queues.

Therefore it doesn't make any sense to schedule the transmit
queue when logically we cannot know ahead of time the TX
queue of the SKB that the qdisc->dequeue() will give us.

Just for sanity I added a BUG check to make sure we never
get into a state where the noop_qdisc is scheduled.

Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-17 19:21:20 -07:00
David S. Miller
7698b4fcab pkt_sched: Add and use qdisc_root() and qdisc_root_lock().
When code wants to lock the qdisc tree state, the logic
operation it's doing is locking the top-level qdisc that
sits of the root of the netdev_queue.

Add qdisc_root_lock() to represent this and convert the
easiest cases.

In order for this to work out in all cases, we have to
hook up the noop_qdisc to a dummy netdev_queue.

Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-17 19:21:19 -07:00
David S. Miller
e2627c8c22 pkt_sched: Make QDISC_RUNNING a qdisc state.
Currently it is associated with a netdev_queue, but when we have
qdisc sharing that no longer makes any sense.

Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-17 19:21:18 -07:00
David S. Miller
d3b753db7c pkt_sched: Move gso_skb into Qdisc.
We liberate any dangling gso_skb during qdisc destruction.

It really only matters for the root qdisc.  But when qdiscs
can be shared by multiple netdev_queue objects, we can't
have the gso_skb in the netdev_queue any more.

Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-17 19:21:18 -07:00
David S. Miller
8f0f2223cc net: Implement simple sw TX hashing.
It just xor hashes over IPv4/IPv6 addresses and ports of transport.

The only assumption it makes is that skb_network_header() is set
correctly.

With bug fixes from Eric Dumazet.

Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-17 19:21:13 -07:00
David S. Miller
51cb6db0f5 mac80211: Reimplement WME using ->select_queue().
The only behavior change is that we do not drop packets under any
circumstances.  If that is absolutely needed, we could easily add it
back.

With cleanups and help from Johannes Berg.

Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-17 19:21:12 -07:00
David S. Miller
eae792b722 netdev: Add netdev->select_queue() method.
Devices or device layers can set this to control the queue selection
performed by dev_pick_tx().

This function runs under RCU protection, which allows overriding
functions to have some way of synchronizing with things like dynamic
->real_num_tx_queues adjustments.

This makes the spinlock prefetch in dev_queue_xmit() a little bit
less effective, but that's the price right now for correctness.

Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-17 19:21:10 -07:00
David S. Miller
fd2ea0a79f net: Use queue aware tests throughout.
This effectively "flips the switch" by making the core networking
and multiqueue-aware drivers use the new TX multiqueue structures.

Non-multiqueue drivers need no changes.  The interfaces they use such
as netif_stop_queue() degenerate into an operation on TX queue zero.
So everything "just works" for them.

Code that really wants to do "X" to all TX queues now invokes a
routine that does so, such as netif_tx_wake_all_queues(),
netif_tx_stop_all_queues(), etc.

pktgen and netpoll required a little bit more surgery than the others.

In particular the pktgen changes, whilst functional, could be largely
improved.  The initial check in pktgen_xmit() will sometimes check the
wrong queue, which is mostly harmless.  The thing to do is probably to
invoke fill_packet() earlier.

The bulk of the netpoll changes is to make the code operate solely on
the TX queue indicated by by the SKB queue mapping.

Setting of the SKB queue mapping is entirely confined inside of
net/core/dev.c:dev_pick_tx().  If we end up needing any kind of
special semantics (drops, for example) it will be implemented here.

Finally, we now have a "real_num_tx_queues" which is where the driver
indicates how many TX queues are actually active.

With IGB changes from Jeff Kirsher.

Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-17 19:21:07 -07:00
David S. Miller
24344d2600 mac80211: Temporarily mark QoS support BROKEN.
We will undo this after a few changsets.

Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-17 19:21:05 -07:00
David S. Miller
1d8ae3fdeb pkt_sched: Remove RR scheduler.
This actually fixes a bug added by the RR scheduler changes.  The
->bands and ->prio2band parameters were being set outside of the
sch_tree_lock() and thus could result in strange behavior and
inconsistencies.

It might be possible, in the new design (where there will be one qdisc
per device TX queue) to allow similar functionality via a TX hash
algorithm for RR but I really see no reason to export this aspect of
how these multiqueue cards actually implement the scheduling of the
the individual DMA TX rings and the single physical MAC/PHY port.

Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-17 19:21:04 -07:00
David S. Miller
09e83b5d7d netdev: Kill NETIF_F_MULTI_QUEUE.
There is no need for a feature bit for something that
can be tested by simply checking the TX queue count.

Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-17 19:21:03 -07:00
David S. Miller
e8a0464cc9 netdev: Allocate multiple queues for TX.
alloc_netdev_mq() now allocates an array of netdev_queue
structures for TX, based upon the queue_count argument.

Furthermore, all accesses to the TX queues are now vectored
through the netdev_get_tx_queue() and netdev_for_each_tx_queue()
interfaces.  This makes it easy to grep the tree for all
things that want to get to a TX queue of a net device.

Problem spots which are not really multiqueue aware yet, and
only work with one queue, can easily be spotted by grepping
for all netdev_get_tx_queue() calls that pass in a zero index.

Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-17 19:21:00 -07:00
Patrick McHardy
51ce7ec921 garp: retry sending JoinIn messages after allocation failures
Increase reliability by retrying to send JoinIn messages after memory
allocation failures on each TRANSMIT_PDU event until it succeeds.

Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-16 20:51:47 -07:00
Neil Horman
9a6d276e85 core: add stat to track unresolved discards in neighbor cache
in __neigh_event_send, if we have a neighbour entry which is in
NUD_INCOMPLETE state, we enqueue any outbound frames to that neighbour
to the neighbours arp_queue, which is default capped to a length of 3
skbs.  If that queue exceeds its set length, it will drop an skb on
the queue to enqueue the newly arrived skb.  This results in a drop
for which we have no statistics incremented.  This patch adds an
unresolved_discards stat to /proc/net/stat/ndisc_cache to track these
lost frames.

Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-16 20:50:49 -07:00
Pavel Emelyanov
ed88098e25 mib: add net to NET_ADD_STATS_USER
Done with NET_XXX_STATS macros :)

To be continued...

Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-16 20:32:45 -07:00
Pavel Emelyanov
f2bf415cfe mib: add net to NET_ADD_STATS_BH
This one is tricky. 

The thing is that this macro is only used when killing tw buckets, 
but since this killer is promiscuous wrt to which net each particular
tw belongs to, I have to use it only when NET_NS is off. When the net
namespaces are on, I use the INET_INC_STATS_BH for each bucket.

Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-16 20:32:25 -07:00
Pavel Emelyanov
6f67c817fc mib: add net to NET_INC_STATS_USER
Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-16 20:31:39 -07:00
Pavel Emelyanov
de0744af1f mib: add net to NET_INC_STATS_BH
Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-16 20:31:16 -07:00
Pavel Emelyanov
4e6734447d mib: add net to NET_INC_STATS
Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-16 20:30:14 -07:00
Pavel Emelyanov
1ed834655a tcp: replace tcp_sock argument with sock in some places
These places have a tcp_sock, but we'd prefer the sock itself to
get net from it. Fortunately, tcp_sk macro is just a type cast, so
this replace is really cheap.

Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-16 20:29:51 -07:00
Pavel Emelyanov
ca12a1a443 inet: prepare net on the stack for NET accounting macros
Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-16 20:28:42 -07:00
Pavel Emelyanov
5c52ba170f sock: add net to prot->enter_memory_pressure callback
The tcp_enter_memory_pressure calls NET_INC_STATS, but doesn't
have where to get the net from.

I decided to add a sk argument, not the net itself, only to factor
all the required sock_net(sk) calls inside the enter_memory_pressure 
callback itself.

Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-16 20:28:10 -07:00
Pavel Emelyanov
74688e487a mib: add net to TCP_DEC_STATS
Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-16 20:22:46 -07:00
Pavel Emelyanov
63231bddf6 mib: add net to TCP_INC_STATS_BH
Same as before - the sock is always there to get the net from,
but there are also some places with the net already saved on 
the stack.

Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-16 20:22:25 -07:00
Pavel Emelyanov
81cc8a75d9 mib: add net to TCP_INC_STATS
Fortunately (almost) all the TCP code has a sock to get the net from :)

Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-16 20:22:04 -07:00
Pavel Emelyanov
a9c19329ec tcp: add net to tcp_mib_init
This one sets TCP MIBs after zeroing them, and thus requires
the net.

The existing single caller can use init_net (temporarily).

Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-16 20:21:42 -07:00
Pavel Emelyanov
a86b1e3019 inet: prepare struct net for TCP MIB accounting
This is the same as the first patch in the set, but preparing
the net for TCP_XXX_STATS - save the struct net on the stack
where required and possible.

Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-16 20:20:58 -07:00
Pavel Emelyanov
c5346fe396 mib: add net to IP_ADD_STATS_BH
Very simple - only ip_evictor (fragments) requires such.
This patch ends up the IP_XXX_STATS patching.

Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-16 20:20:33 -07:00
Pavel Emelyanov
7c73a6faff mib: add net to IP_INC_STATS_BH
Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-16 20:20:11 -07:00
Pavel Emelyanov
5e38e27044 mib: add net to IP_INC_STATS
All the callers already have either the net itself, or the place
where to get it from.

Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-16 20:19:49 -07:00
Pavel Emelyanov
84a3aa000e ipv4: prepare net initialization for IP accounting
Some places, that deal with IP statistics already have where to
get a struct net from, but use it directly, without declaring
a separate variable on the stack.

So, save this net on the stack for future IP_XXX_STATS macros.

Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-16 20:19:08 -07:00
Will Newton
70efce27fc net/ipv4/tcp.c: Fix use of PULLHUP instead of POLLHUP in comments.
Change PULLHUP to POLLHUP in tcp_poll comments and clean up another
comment for grammar and coding style.

Signed-off-by: Will Newton <will.newton@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-16 20:13:43 -07:00
Harvey Harrison
7b1c65faa2 net: make __skb_splice_bits static
net/core/skbuff.c:1335:5: warning: symbol '__skb_splice_bits' was not declared. Should it be static?

Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-16 20:12:30 -07:00
David S. Miller
885a4c966b Merge branch 'stealer/ipvs/sync-daemon-cleanup-for-next' of git://git.stealer.net/linux-2.6 2008-07-16 20:07:06 -07:00
Rumen G. Bogdanovski
9d3a0de7dc ipvs: More reliable synchronization on connection close
This patch enhances the synchronization of the closing connections
between the master and the backup director. It prevents the closed
connections to expire with the 15 min timeout of the ESTABLISHED
state on the backup and makes them expire as they would do on the
master with much shorter timeouts.

Signed-off-by: Rumen G. Bogdanovski <rumen@voicecho.com>
Acked-by: Simon Horman <horms@verge.net.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-16 20:04:23 -07:00
Sven Wegener
375c6bbabf ipvs: Use schedule_timeout_interruptible() instead of msleep_interruptible()
So that kthread_stop() can wake up the thread and we don't have to wait one
second in the worst case for the daemon to actually stop.

Signed-off-by: Sven Wegener <sven.wegener@stealer.net>
Acked-by: Simon Horman <horms@verge.net.au>
2008-07-16 22:33:20 +00:00
Sven Wegener
ba6fd85021 ipvs: Put backup thread on mcast socket wait queue
Instead of doing an endless loop with sleeping for one second, we now put the
backup thread onto the mcast socket wait queue and it gets woken up as soon as
we have data to process.

Signed-off-by: Sven Wegener <sven.wegener@stealer.net>
Acked-by: Simon Horman <horms@verge.net.au>
2008-07-16 22:33:20 +00:00
Sven Wegener
998e7a7680 ipvs: Use kthread_run() instead of doing a double-fork via kernel_thread()
This also moves the setup code out of the daemons, so that we're able to
return proper error codes to user space. The current code will return success
to user space when the daemon is started with an invald mcast interface. With
these changes we get an appropriate "No such device" error.

We longer need our own completion to be sure the daemons are actually running,
because they no longer contain code that can fail and kthread_run() takes care
of the rest.

Signed-off-by: Sven Wegener <sven.wegener@stealer.net>
Acked-by: Simon Horman <horms@verge.net.au>
2008-07-16 22:33:20 +00:00
Sven Wegener
e6dd731c75 ipvs: Use ERR_PTR for returning errors from make_receive_sock() and make_send_sock()
The additional information we now return to the caller is currently not used,
but will be used to return errors to user space.

Signed-off-by: Sven Wegener <sven.wegener@stealer.net>
Acked-by: Simon Horman <horms@verge.net.au>
2008-07-16 22:33:19 +00:00
Sven Wegener
d56400504a ipvs: Initialize mcast addr at compile time
There's no need to do it at runtime, the values are constant.

Signed-off-by: Sven Wegener <sven.wegener@stealer.net>
Acked-by: Simon Horman <horms@verge.net.au>
2008-07-16 22:33:19 +00:00
Akinobu Mita
d0236f8f82 iucv: fix memory leak in cpu hotplug error path.
Fix memory leak in error path in CPU_UP_PREPARE notifier.

Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-15 02:09:53 -07:00
Octavian Purdila
2870c43d17 net: refactor tcp splice receive path to improve readability
- move all of the details on offsets, lengths and buffers into a
single function instead of doing these operation from multiple places

- use a bottom up approach: try to avoid details in the high level
functions, introduce them gradually as we go deeper in the function
call stack

With helpful feedback from Jarek Poplawski.

Signed-off-by: Octavian Purdila <opurdila@ixiacom.com>
Acked-by: Jarek Poplawski <jarkao2@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-15 00:49:11 -07:00
David S. Miller
b9e4085768 netdev: Do not use TX lock to protect address lists.
Now that we have a specific lock to protect the network
device unicast and multicast lists, remove extraneous
grabs of the TX lock in cases where the code only needs
address list protection.

Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-15 00:15:08 -07:00
David S. Miller
e308a5d806 netdev: Add netdev->addr_list_lock protection.
Add netif_addr_{lock,unlock}{,_bh}() helpers.

Use them to protect operations that operate on or read
the network device unicast and multicast address lists.

Also use them in cases where the code simply wants to
block calls into the driver's ->set_rx_mode() and
->set_multicast_list() methods.

Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-15 00:13:44 -07:00
David S. Miller
f1f28aa351 netdev: Add addr_list_lock to struct net_device.
This will be used to protect the per-device unicast and multicast
address lists, as well as the callbacks into the drivers which
configure such state such as ->set_rx_mode() and ->set_multicast_list().

Signed-off-by: David S. Miller <davem@davemloft.net>
2008-07-15 00:08:33 -07:00