Choose saner defaults for xfrm[4|6] gc_thresh values on init

Currently, the xfrm[4|6] code has hard-coded initial gc_thresh values
(set to 1024).  Given that the ipv4 and ipv6 routing caches are sized
dynamically at boot time, the static selections can be non-sensical.
This patch dynamically selects an appropriate gc threshold based on
the corresponding main routing table size, using the assumption that
we should in the worst case be able to handle as many connections as
the routing table can.

For ipv4, the maximum route cache size is 16 * the number of hash
buckets in the route cache.  Given that xfrm4 starts garbage
collection at the gc_thresh and prevents new allocations at 2 *
gc_thresh, we set gc_thresh to half the maximum route cache size.

For ipv6, its a bit trickier.  there is no maximum route cache size,
but the ipv6 dst_ops gc_thresh is statically set to 1024.  It seems
sane to select a simmilar gc_thresh for the xfrm6 code that is half
the number of hash buckets in the v6 route cache times 16 (like the v4
code does).
Signed-off-by: NNeil Horman <nhorman@tuxdriver.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

While being at it cleanup whitespace.
Signed-off-by: NHannes Eder <heder@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

If arp_format_neigh_entry() can be called with n->dev->addr_len == 0, then a
write to hbuffer[-1] occurs.
Signed-off-by: NRoel Kluin <roel.kluin@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

There is no reason for the arbitrary restriction that device must be
up to create a vlan. This patch was added to Vyatta kernel to resolve startup
ordering issues where vlan's are created but real device was disabled.

Note: the vlan already correctly inherits the operstate from real device; so
if vlan is created and real device is marked down, the vlan is marked
down.
Signed-off-by: NStephen Hemminger <shemminger@vyatta.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The test on map4 should be a test on map6.

The semantic match that finds this problem is as follows:
(http://www.emn.fr/x-info/coccinelle/)

// <smpl>
@@
expression *x;
identifier f;
constant char *C;
@@

x = \(kmalloc\|kcalloc\|kzalloc\)(...);
... when != x == NULL
    when != x != NULL
    when != (x || ...)
(
kfree(x)
|
f(...,C,...,x,...)
|
*f(...,x,...)
|
*x->f
)
// </smpl>
Signed-off-by: NJulia Lawall <julia@diku.dk>
Acked-by: NPaul Moore <paul.moore@hp.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Finally! This is what you've all been waiting for!

This patch makes cfg80211 take care of wext emulation
_completely_ by itself, drivers that don't need things
cfg80211 doesn't do yet don't even need to be aware of
wireless extensions.
This means we can also clean up mac80211's and iwm's
Kconfig and make it possible to build them w/o wext
now!

		RIP wext.
Signed-off-by: NJohannes Berg <johannes@sipsolutions.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Since we now have handlers IWESSID for all modes, we can
combine them into one.
Signed-off-by: NJohannes Berg <johannes@sipsolutions.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Since we now have IWAP handlers for all modes, we can
combine them into one.
Signed-off-by: NJohannes Berg <johannes@sipsolutions.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Until now we implemented iwfreq for managed mode, we
needed to keep the implementations separate, but now
that we have all versions implemented we can combine
them and export just one handler.
Signed-off-by: NJohannes Berg <johannes@sipsolutions.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

When enqueuing packets on the internal packet queue, we
need to ensure that we have a valid vif pointer since
that is required since the net namespace work. Add some
assertions to verify this, but also don't crash is for
some reason we don't end up with a vif pointer -- warn
and drop the packet in all these cases.

Since this code touches a number of hotpaths, it is
intended to be temporary, or maybe configurable in the
future, at least the bit that is in the path that gets
hit for every packet, ieee80211_tx_pending().
Signed-off-by: NJohannes Berg <johannes@sipsolutions.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

When trying to disassociate while not associated,
the kernel would crash rather than refusing the
operation, fix this;
Reported-by: NMaxim Levitsky <maximlevitsky@gmail.com>
Signed-off-by: NJohannes Berg <johannes@sipsolutions.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Fix an oops in ieee80211_scan_state_set_channel which was triggered
if the last scanned channel was skipped (for example due to regulatory
restrictions) by returning to the decision state after each skipped
channel.
Signed-off-by: NHelmut Schaa <helmut.schaa@googlemail.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Jouni and Maxim reported an oops when using wpa_supplicant -Dnl80211,
which seems to be due to random data being contained in the crypto
settings for the assoc() command. This seems to be due to the missing
memset here, so add it -- it's certainly missing but I'm not 100%
certain that it will fix the problem.
Signed-off-by: NJohannes Berg <johannes@sipsolutions.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Marcel reported a warning, which quite obviously comes
from an oversight in the code handling deauth frames,
and which resulted in multiple follow-up warnings due
to this missing handling. This patch adds the missing
deauth handling (telling cfg80211 about it) and also
removes the follow-up warnings since they could happen
due to races even if nothing is wrong. I've explained
the races in the comments.
Signed-off-by: NJohannes Berg <johannes@sipsolutions.net>
Reported-by: NMarcel Holtmann <marcel@holtmann.org>
Tested-by: NMarcel Holtmann <marcel@holtmann.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Luis reported this lockdep complaint, that he had also
reported earlier but when trying to analyse I had been
locking at the wrong code, and never saw the problem:

(slightly abridged)
=======================================================
[ INFO: possible circular locking dependency detected ]
2.6.31-rc4-wl #6
-------------------------------------------------------
wpa_supplicant/3799 is trying to acquire lock:
 (cfg80211_mutex){+.+.+.}, at: [<ffffffffa009246a>] cfg80211_get_dev_from_ifindex+0x1a/0x90 [cfg80211]

but task is already holding lock:
 (rtnl_mutex){+.+.+.}, at: [<ffffffff81400ff2>] rtnl_lock+0x12/0x20

which lock already depends on the new lock.

the existing dependency chain (in reverse order) is:

-> #1 (rtnl_mutex){+.+.+.}:
       [<ffffffff810857b6>] __lock_acquire+0xd76/0x12b0
       [<ffffffff81085dd3>] lock_acquire+0xe3/0x120
       [<ffffffff814ee7a4>] mutex_lock_nested+0x44/0x350
       [<ffffffff81400ff2>] rtnl_lock+0x12/0x20
       [<ffffffffa009f6a5>] nl80211_send_reg_change_event+0x1f5/0x2a0 [cfg80211]
       [<ffffffffa009529e>] set_regdom+0x28e/0x4c0 [cfg80211]

-> #0 (cfg80211_mutex){+.+.+.}:
       [<ffffffff8108587b>] __lock_acquire+0xe3b/0x12b0
       [<ffffffff81085dd3>] lock_acquire+0xe3/0x120
       [<ffffffff814ee7a4>] mutex_lock_nested+0x44/0x350
       [<ffffffffa009246a>] cfg80211_get_dev_from_ifindex+0x1a/0x90 [cfg80211]
       [<ffffffffa009813f>] get_rdev_dev_by_info_ifindex+0x6f/0xa0 [cfg80211]
       [<ffffffffa009b12b>] nl80211_set_interface+0x3b/0x260 [cfg80211]

When looking at the correct code, the problem is quite
obvious. I'm not entirely sure which code paths lead
here, so until I can analyse it better let's just use
RCU to avoid the problem.
Signed-off-by: NJohannes Berg <johannes@sipsolutions.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Jan reported that his b43-based laptop hangs during suspend.
The problem turned out to be mac80211 asking the driver to
stop the hardware before removing interfaces, and interface
removal caused b43 to touch the hardware (while down, which
causes the hang).

This patch fixes mac80211 to do reorder these operations to
have them in the correct order -- first remove interfaces
and then stop the hardware. Some more code is necessary to
be able to do so in a race-free manner, in particular it is
necessary to not process frames received during quiescing.

Fixes http://bugzilla.kernel.org/show_bug.cgi?id=13337.
Reported-by: NJan Scholz <scholz@fias.uni-frankfurt.de>
Signed-off-by: NJohannes Berg <johannes@sipsolutions.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

This was caused by patch:

"mac80211: cooperate more with network namespaces"

The version of the patch applied doesn't match Johannes' latest:

http://johannes.sipsolutions.net/patches/kernel/all/LATEST/NNN-mac80211-netns.patch

The skb->cb virtual interface data wasn't being reset for
reuse so ath9k pooped out when trying to dereference the
private rate control info from the skb.

BUG: unable to handle kernel NULL pointer dereference
RIP: 0010:[<ffffffffa0258173>] ath_tx_rc_status+0x33/0x150 [ath9k]
<-- snip etc -->
Reported-by: NDavide Pesavento <davidepesa@gmail.com>
Signed-off-by: NLuis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

When a station queries us for a PS-poll response, we wrongly
queue the frame on the virtual interface's queue rather than
the pending queue.

Additionally, fix a race condition where we could potentially
send multiple frames to the sleeping station due to using a
station flag rather than a packet flag. When converting to a
packet flag, we can also convert p54 and remove the filter
clearing we added for it.

(Also remove a now dead function)
Signed-off-by: NJohannes Berg <johannes@sipsolutions.net>
Reported-by: NBob Copeland <me@bobcopeland.com>
Tested-by: NBob Copeland <me@bobcopeland.com>
Cc: Christian Lamparter <chunkeey@web.de>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

We were issuing probe requests to the associated AP on the wrong
band by having our beacon timer loss trigger while we are scanning.
When we would scan the timer could hit and force us to send a
probe request to the AP but with a chance we'd be on the wrong band.

This leads to finding no usable bitrate but we should not get so
far on the xmit path. We should not be trying to send these probe
request frames so prevent ieee80211_mgd_probe_ap() from sending
these.

As it turns out all callers of ieee80211_mgd_probe_ap() need this
check so we just move the scan check there. This means we can remove
the recenlty added check during ieee80211_sta_monitor_work().

Additionally we now fix a race condition added by the patch
"mac80211: do not monitor the connection while scanning" which
had the same check in ieee80211_sta_conn_mon_timer(). The race
happens because the timer routine *does* a valid check for
scanning but after it queues work into the mac80211 workqueue
the work callback can kick off with scanning enabled and cause
the same issue we were trying to avoid.

The more appropriate solution would be to disable the respective
timers during scan and re-enable them after scan but requires more
complex code and testing.

Cc: Christian Lamparter <chunkeey@web.de>
Cc: Larry Finger <Larry.Finger@lwfinger.net>
Reported-by: NFabio Rossi <rossi.f@inwind.it>
Signed-off-by: NLuis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

When a new MLME work is created, its timeout is initialised
to 0. This is wrong, it could then be thought of as having
an actual timeout in the future (time_is_after_jiffies() can
return true). Instead, it should be initialised to jiffies
so that it will run right away as soon as the mlme work is
executed.
Signed-off-by: NJohannes Berg <johannes@sipsolutions.net>
Reported-by: NLuciano Roth Coelho <luciano.coelho@nokia.com>
Reported-by: NAlban Browaeys <prahal@yahoo.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Using background scanning in mac80211 the time a scan needs to
finish can exceed 10 seconds. Hence, increase the scan results
expire time to 15 seconds which should be sufficient.
Signed-off-by: NHelmut Schaa <helmut.schaa@googlemail.com>
Acked-by: NJohannes Berg <johannes@sipsolutions.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Rename scan_state to next_scan_state to better reflect
what it is used for.
Signed-off-by: NHelmut Schaa <helmut.schaa@googlemail.com>
Acked-by: NJohannes Berg <johannes@sipsolutions.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Introduce a new scan flag "SCAN_OFF_CHANNEL" which basically tells us
that we are currently on a different channel for scanning and cannot
RX/TX. "SCAN_SW_SCANNING" tells us that we are currently running a
software scan but we might as well be on the operating channel to RX/TX.
While "SCAN_SW_SCANNING" is set during the whole scan "SCAN_OFF_CHANNEL"
is set when leaving the operating channel and unset when coming back.

Introduce two new scan states "SCAN_LEAVE_OPER_CHANNEL" and
"SCAN_ENTER_OPER_CHANNEL" which basically implement the functionality we
need to leave the operating channel (send a nullfunc to the AP and stop
the queues) and enter it again (send a nullfunc to the AP and start the
queues again).

Enhance the scan state "SCAN_DECISION" to switch back to the operating
channel after each scanned channel. In the future it sould be simple
to enhance the decision state to scan as much channels in a row as the
qos latency allows us.
Signed-off-by: NHelmut Schaa <helmut.schaa@googlemail.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Use a bitfield to store the current scan mode instead of two boolean
variables {sw,hw}_scanning. This patch does not introduce functional
changes but allows us to enhance the scan flags later (for example
for background scanning).
Signed-off-by: NHelmut Schaa <helmut.schaa@googlemail.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Introduce a new scan state "decision" which is entered after
every completed scan operation and decides about the next steps.
At first the decision is in any case to scan the next channel.
This shouldn't introduce any functional changes.
Signed-off-by: NHelmut Schaa <helmut.schaa@googlemail.com>
Acked-by: NJohannes Berg <johannes@sipsolutions.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Instead of queueing the scan work again without delay just process the
next state immediately.
Signed-off-by: NHelmut Schaa <helmut.schaa@googlemail.com>
Acked-by: NJohannes Berg <johannes@sipsolutions.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Move the processing of each scan state into its own functions for better
readability. This patch does not introduce functional changes.
Signed-off-by: NHelmut Schaa <helmut.schaa@googlemail.com>
Acked-by: NJohannes Berg <johannes@sipsolutions.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

This finally opens up the ability to put mac80211 devices
into different network namespaces. As long as you don't
have sysfs, that is.
Signed-off-by: NJohannes Berg <johannes@sipsolutions.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

In order to make cfg80211/nl80211 aware of network namespaces,
we have to do the following things:

 * del_virtual_intf method takes an interface index rather
   than a netdev pointer - simply change this

 * nl80211 uses init_net a lot, it changes to use the sender's
   network namespace

 * scan requests use the interface index, hold a netdev pointer
   and reference instead

 * we want a wiphy and its associated virtual interfaces to be
   in one netns together, so
    - we need to be able to change ns for a given interface, so
      export dev_change_net_namespace()
    - for each virtual interface set the NETIF_F_NETNS_LOCAL
      flag, and clear that flag only when the wiphy changes ns,
      to disallow breaking this invariant

 * when a network namespace goes away, we need to reparent the
   wiphy to init_net

 * cfg80211 users that support creating virtual interfaces must
   create them in the wiphy's namespace, currently this affects
   only mac80211

The end result is that you can now switch an entire wiphy into
a different network namespace with the new command
	iw phy#<idx> set netns <pid>
and all virtual interfaces will follow (or the operation fails).
Signed-off-by: NJohannes Berg <johannes@sipsolutions.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

There are still two places in mac80211 that hardcode
the initial net namespace (init_net). One of them is
mandated by cfg80211 and will be removed by a separate
patch, the other one is used for finding the network
device of a pending packet via its ifindex.

Remove the latter use by keeping track of the device
pointer itself, via the vif pointer, and avoid it
going stale by dropping pending frames for a given
interface when the interface is removed.

To keep track of the vif pointer for the correct
interface, change the info->control.vif pointer's
internal use to always be the correct vif, and only
move it to the vif the driver expects (or NULL for
monitor interfaces and injected packets) right before
giving the packet to the driver.
Signed-off-by: NJohannes Berg <johannes@sipsolutions.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Signed-off-by: NLuis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Export garbage collector thresholds for xfrm[4|6]_dst_ops

Had a problem reported to me recently in which a high volume of ipsec
connections on a system began reporting ENOBUFS for new connections
eventually.

It seemed that after about 2000 connections we started being unable to
create more.  A quick look revealed that the xfrm code used a dst_ops
structure that limited the gc_thresh value to 1024, and always
dropped route cache entries after 2x the gc_thresh.

It seems the most direct solution is to export the gc_thresh values in
the xfrm[4|6] dst_ops as sysctls, like the main routing table does, so
that higher volumes of connections can be supported.  This patch has
been tested and allows the reporter to increase their ipsec connection
volume successfully.
Reported-by: NJoe Nall <joe@nall.com>
Signed-off-by: NNeil Horman <nhorman@tuxdriver.com>

ipv4/xfrm4_policy.c |   18 ++++++++++++++++++
ipv6/xfrm6_policy.c |   18 ++++++++++++++++++
2 files changed, 36 insertions(+)
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Signed-off-by: NEric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

entry was tested for NULL near the beginning of the function, followed by a
return, and there is no intervening modification of its value.

A simplified version of the semantic match that finds this problem is as
follows: (http://www.emn.fr/x-info/coccinelle/)

// <smpl>
@r exists@
local idexpression x;
expression E;
position p1,p2;
@@

if (x == NULL || ...) { ... when forall
   return ...; }
... when != \(x=E\|x--\|x++\|--x\|++x\|x-=E\|x+=E\|x|=E\|x&=E\|&x\)
(
*x == NULL
|
*x != NULL
)
// </smpl>
Signed-off-by: NJulia Lawall <julia@diku.dk>
Acked-by: NPaul Moore <paul.moore@hp.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

irttp_dup() copies a tsap_cb struct, but does not initialize the
spinlock in the new structure, which confuses lockdep.
Signed-off-by: NDeepak Saxena <dsaxena@mvista.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

net/phonet/pn_dev.c: In function `phonet_device_get':
net/phonet/pn_dev.c:99: warning: 'dev' might be used uninitialized in this function
Signed-off-by: NEric Dumazet <eric.dumazet@gmail.com>
Acked-by: NRémi Denis-Courmont <remi.denis-courmont@nokia.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This helps avoid error messages with ethtool -k on devices that
don't provide device specific routines.
Signed-off-by: NSridhar Samudrala <sri@us.ibm.com>

------------------------------------------------------------------
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

kobject_init_and_add will alloc memory for kobj->name, so in br_add_if
error path, simply use kobject_del will not free memory for kobj->name.
Fix by using kobject_put instead, kobject_put will internally calls
kobject_del and frees memory for kobj->name.
Signed-off-by: NXiaotian Feng <dfeng@redhat.com>
Acked-by: NStephen Hemminger <shemminger@vyatta.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The use of a static buffer in rose2asc() to return its result is not
threadproof and can result in corruption if multiple threads are trying
to use one of the procfs files based on rose2asc().
Signed-off-by: NRalf Baechle <ralf@linux-mips.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

As reported by Oliver Hartkopp:

net/phonet/pn_dev.c: In function ‘phonet_init_net’:
net/phonet/pn_dev.c:221: error: implicit declaration of function
‘proc_net_fops_create’
net/phonet/pn_dev.c: In function ‘phonet_exit_net’:
net/phonet/pn_dev.c:242: error: implicit declaration of function ‘proc_net_remove’
Signed-off-by: NDavid S. Miller <davem@davemloft.net>