linux-stable.git/net/wireless, branch linux-3.12.y

nl80211: fix dumpit error path RTNL deadlocks

2017-04-07T08:38:27+00:00

commit ea90e0dc8cecba6359b481e24d9c37160f6f524f upstream.

Sowmini pointed out Dmitry's RTNL deadlock report to me, and it turns out
to be perfectly accurate - there are various error paths that miss unlock
of the RTNL.

To fix those, change the locking a bit to not be conditional in all those
nl80211_prepare_*_dump() functions, but make those require the RTNL to
start with, and fix the buggy error paths. This also let me use sparse
(by appropriately overriding the rtnl_lock/rtnl_unlock functions) to
validate the changes.

[js] no mpp and vendor dumps in 3.12 yet

Reported-by: Sowmini Varadhan 
Reported-by: Dmitry Vyukov 
Signed-off-by: Johannes Berg 
Signed-off-by: Jiri Slaby

cfg80211: limit scan results cache size

2016-12-12T12:55:52+00:00

commit 9853a55ef1bb66d7411136046060bbfb69c714fa upstream.

It's possible to make scanning consume almost arbitrary amounts
of memory, e.g. by sending beacon frames with random BSSIDs at
high rates while somebody is scanning.

Limit the number of BSS table entries we're willing to cache to
1000, limiting maximum memory usage to maybe 4-5MB, but lower
in practice - that would be the case for having both full-sized
beacon and probe response frames for each entry; this seems not
possible in practice, so a limit of 1000 entries will likely be
closer to 0.5 MB.

Signed-off-by: Johannes Berg 
Signed-off-by: Jiri Slaby

Revert "wext: Fix 32 bit iwpriv compatibility issue with 64 bit Kernel"

2016-09-29T09:14:08+00:00

commit 4d0bd46a4d55383f7b925e6cf7865a77e0f0e020 upstream.

This reverts commit 3d5fdff46c4b2b9534fa2f9fc78e90a48e0ff724.

Ben Hutchings pointed out that the commit isn't safe since it assumes
that the structure used by the driver is iw_point, when in fact there's
no way to know about that.

Fortunately, the only driver in the tree that ever runs this code path
is the wilc1000 staging driver, so it doesn't really matter.

Clearly I should have investigated this better before applying, sorry.

Reported-by: Ben Hutchings 
Fixes: 3d5fdff46c4b ("wext: Fix 32 bit iwpriv compatibility issue with 64 bit Kernel")
Signed-off-by: Johannes Berg 
Signed-off-by: Jiri Slaby

wext: Fix 32 bit iwpriv compatibility issue with 64 bit Kernel

2016-07-18T11:52:17+00:00

commit 3d5fdff46c4b2b9534fa2f9fc78e90a48e0ff724 upstream.

iwpriv app uses iw_point structure to send data to Kernel. The iw_point
structure holds a pointer. For compatibility Kernel converts the pointer
as required for WEXT IOCTLs (SIOCIWFIRST to SIOCIWLAST). Some drivers
may use iw_handler_def.private_args to populate iwpriv commands instead
of iw_handler_def.private. For those case, the IOCTLs from
SIOCIWFIRSTPRIV to SIOCIWLASTPRIV will follow the path ndo_do_ioctl().
Accordingly when the filled up iw_point structure comes from 32 bit
iwpriv to 64 bit Kernel, Kernel will not convert the pointer and sends
it to driver. So, the driver may get the invalid data.

The pointer conversion for the IOCTLs (SIOCIWFIRSTPRIV to
SIOCIWLASTPRIV), which follow the path ndo_do_ioctl(), is mandatory.
This patch adds pointer conversion from 32 bit to 64 bit and vice versa,
if the ioctl comes from 32 bit iwpriv to 64 bit Kernel.

Signed-off-by: Prasun Maiti 
Signed-off-by: Ujjal Roy 
Tested-by: Dibyajyoti Ghosh 
Signed-off-by: Johannes Berg 
Signed-off-by: Jiri Slaby

nl80211: check netlink protocol in socket release notification

2016-05-02T17:54:56+00:00

commit 8f815cdde3e550e10c2736990d791f60c2ce43eb upstream.

A non-privileged user can create a netlink socket with the same port_id as
used by an existing open nl80211 netlink socket (e.g. as used by a hostapd
process) with a different protocol number.

Closing this socket will then lead to the notification going to nl80211's
socket release notification handler, and possibly cause an action such as
removing a virtual interface.

Fix this issue by checking that the netlink protocol is NETLINK_GENERIC.
Since generic netlink has no notifier chain of its own, we can't fix the
problem more generically.

Fixes: 026331c4d9b5 ("cfg80211/mac80211: allow registering for and sending action frames")
Signed-off-by: Dmitry Ivanov 
[rewrite commit message]
Signed-off-by: Johannes Berg 
Signed-off-by: Jiri Slaby

wext: fix message delay/ordering

2016-03-14T12:38:07+00:00

commit 8bf862739a7786ae72409220914df960a0aa80d8 upstream.

Beniamino reported that he was getting an RTM_NEWLINK message for a
given interface, after the RTM_DELLINK for it. It turns out that the
message is a wireless extensions message, which was sent because the
interface had been connected and disconnection while it was deleted
caused a wext message.

For its netlink messages, wext uses RTM_NEWLINK, but the message is
without all the regular rtnetlink attributes, so "ip monitor link"
prints just rudimentary information:

5: wlan1:  mtu 1500 qdisc mq state DOWN group default
    link/ether 02:00:00:00:01:00 brd ff:ff:ff:ff:ff:ff
Deleted 5: wlan1:  mtu 1500 qdisc noop state DOWN group default
    link/ether 02:00:00:00:01:00 brd ff:ff:ff:ff:ff:ff
5: wlan1: 
    link/ether
(from my hwsim reproduction)

This can cause userspace to get confused since it doesn't expect an
RTM_NEWLINK message after RTM_DELLINK.

The reason for this is that wext schedules a worker to send out the
messages, and the scheduling delay can cause the messages to get out
to userspace in different order.

To fix this, have wext register a netdevice notifier and flush out
any pending messages when netdevice state changes. This fixes any
ordering whenever the original message wasn't sent by a notifier
itself.

Reported-by: Beniamino Galvani 
Signed-off-by: Johannes Berg 
Signed-off-by: Jiri Slaby

cfg80211: wext: clear sinfo struct before calling driver

2015-06-23T12:55:14+00:00

commit 9c5a18a31b321f120efda412281bb9f610f84aa0 upstream.

Until recently, mac80211 overwrote all the statistics it could
provide when getting called, but it now relies on the struct
having been zeroed by the caller. This was always the case in
nl80211, but wext used a static struct which could even cause
values from one device leak to another.

Using a static struct is OK (as even documented in a comment)
since the whole usage of this function and its return value is
always locked under RTNL. Not clearing the struct for calling
the driver has always been wrong though, since drivers were
free to only fill values they could report, so calling this
for one device and then for another would always have leaked
values from one to the other.

Fix this by initializing the structure in question before the
driver method call.

This fixes https://bugzilla.kernel.org/show_bug.cgi?id=99691

Reported-by: Gerrit Renker 
Reported-by: Alexander Kaltsas 
Signed-off-by: Johannes Berg 
Signed-off-by: David S. Miller 
Signed-off-by: Jiri Slaby

nl80211: ignore HT/VHT capabilities without QoS/WMM

2015-04-09T12:14:06+00:00

commit 496fcc294daab18799e190c0264863d653588d1f upstream.

As HT/VHT depend heavily on QoS/WMM, it's not a good idea to
let userspace add clients that have HT/VHT but not QoS/WMM.
Since it does so in certain cases we've observed (client is
using HT IEs but not QoS/WMM) just ignore the HT/VHT info at
this point and don't pass it down to the drivers which might
unconditionally use it.

Signed-off-by: Johannes Berg 
Signed-off-by: Jiri Slaby

cfg80211: Fix 160 MHz channels with 80+80 and 160 MHz drivers

2015-04-09T12:13:39+00:00

commit 08f6f147773b23b765b94633a8eaa82e7defcf4c upstream.

The VHT supported channel width field is a two bit integer, not a
bitfield. cfg80211_chandef_usable() was interpreting it incorrectly and
ended up rejecting 160 MHz channel width if the driver indicated support
for both 160 and 80+80 MHz channels.

Cc: stable@vger.kernel.org (3.16+)
Fixes: 3d9d1d6656a73 ("nl80211/cfg80211: support VHT channel configuration")
       (however, no real drivers had 160 MHz support it until 3.16)
Signed-off-by: Jouni Malinen 
Signed-off-by: Johannes Berg 

Signed-off-by: Jiri Slaby

nl80211: fix per-station group key get/del and memory leak

2015-02-08T19:02:00+00:00

commit 0fa7b39131576dd1baa6ca17fca53c65d7f62249 upstream.

In case userspace attempts to obtain key information for or delete a
unicast key, this is currently erroneously rejected unless the driver
sets the WIPHY_FLAG_IBSS_RSN flag. Apparently enough drivers do so it
was never noticed.

Fix that, and while at it fix a potential memory leak: the error path
in the get_key() function was placed after allocating a message but
didn't free it - move it to a better place. Luckily admin permissions
are needed to call this operation.

Fixes: e31b82136d1ad ("cfg80211/mac80211: allow per-station GTKs")
Signed-off-by: Johannes Berg 
Signed-off-by: Jiri Slaby