Akinobu Mita [Fri, 27 Aug 2010 19:08:13 +0000 (19:08 +0000)]
mlx4: use bitmap library
Replace loops calling set_bit() and clear_bit() with bitmap_set() and
bitmap_clear().
Unlike loops calling set_bit() and clear_bit(), bitmap_set() and
bitmap_clear() are not atomic. But this is ok.
Because the bitmap operations are protected by bitmap->lock
except for initialization of the bitmap in mlx4_bitmap_init().
Tobias Klauser [Thu, 26 Aug 2010 22:12:08 +0000 (22:12 +0000)]
slip: Use net_device_stats from struct net_device
Use net_device->stats for stats instead of private variable copies in
struct slip. Use ndo_get_stat64 so the additions can be performed on a private
destination buffer.
Johannes Berg [Fri, 27 Aug 2010 11:45:28 +0000 (13:45 +0200)]
mac80211: fix offchannel queue stop
Somebody noticed this problem, and I outlined
to them how to fix it, but haven't heard back
from them. So while I was adding the state
field I figured I could use it to fix it.
The problem, as I understand it, is that when
we go offchannel while the driver has a queue
stopped, the driver will likely start draining
the queue and then enable it while offchannel.
This in turn will enable the interface queue,
and that leads to transmitting data frames on
the wrong channel.
Fix this by keeping track of offchannel status
per interface, and not enabling the interface
queues on interfaces that are offchannel when
the driver enables a queue.
Johannes Berg [Fri, 27 Aug 2010 10:35:58 +0000 (12:35 +0200)]
mac80211: support runtime interface type changes
Add support to mac80211 for changing the interface
type even when the interface is UP, if the driver
supports it.
To achieve this
* add a new driver callback for switching,
* split some of the interface up/down code out
into new functions (do_open/do_stop), and
* maintain an own __SDATA_RUNNING bit that will
not be set during interface type, so that any
other code doesn't use the interface.
Johannes Berg [Fri, 27 Aug 2010 10:35:56 +0000 (12:35 +0200)]
mac80211: simplify zero address checks
The libertas_tf special code for zero addresses
is a bit too complex, it compares against a stack
value instead of using is_zero_ether_addr() and
tries to update all interfaces even if just the
one that's being brought up needs to be changed.
Additionally, the repeated check for a valid MAC
address need only be done if we actually changed
it on the fly.
Johannes Berg [Fri, 27 Aug 2010 10:35:54 +0000 (12:35 +0200)]
mac80211: clean up ifdown/cleanup paths
There's a lot of redundant code in mac80211's
interface cleanup/down, for example freeing
AP beacons is done both when the interface is
set DOWN as well as when it is torn down, of
which only the former has any effect.
Also, a bunch of things should be closer to
where they matter, like the MLME timers that
we should cancel when disassociating, rather
than only when the interface is set DOWN.
Johannes Berg [Fri, 27 Aug 2010 11:26:54 +0000 (14:26 +0300)]
mac80211: allow changing port control protocol
Some vendor specified mechanisms for 802.1X-style
functionality use a different protocol than EAP
(even if EAP is vendor-extensible). Support this
in mac80211 via the cfg80211 API for it.
Johannes Berg [Fri, 27 Aug 2010 11:26:53 +0000 (14:26 +0300)]
cfg80211: allow changing port control protocol
Some vendor specified mechanisms for 802.1X-style
functionality use a different protocol than EAP
(even if EAP is vendor-extensible). Allow setting
the ethertype for the protocol when a driver has
support for this. The default if unspecified is
EAP, of course.
Note: This is suitable only for station mode, not
for AP implementation.
Johannes Berg [Fri, 27 Aug 2010 11:26:52 +0000 (14:26 +0300)]
mac80211: allow vendor specific cipher suites
Allow drivers to specify their own set of cipher
suites to advertise vendor-specific ciphers. The
driver is then required to implement hardware
crypto offload for it.
Johannes Berg [Fri, 27 Aug 2010 11:26:51 +0000 (14:26 +0300)]
cfg80211: allow vendor specific cipher suites
cfg80211 currently rejects all cipher suites it
doesn't know about for key length checking
purposes. This can lead to inconsistencies when
a driver advertises an algorithm that cfg80211
doesn't know about. Remove this rejection so
drivers can specify any algorithm they like.
Johannes Berg [Thu, 26 Aug 2010 11:30:27 +0000 (13:30 +0200)]
wl12xx: remove unneeded locking
With the scan callback now being callable from
any context, these unlocks/locks can go away.
This makes the code easier to understand, since
callers of these functions must no longer be
aware that the mutex may be dropped.
As Stanislaw is working on iwlwifi scanning, I
didn't change it to take advantage of the new
mac80211 semantics.
Johannes Berg [Thu, 26 Aug 2010 11:30:26 +0000 (13:30 +0200)]
mac80211: allow scan to complete from any context
The ieee80211_scan_completed() function was a frequent
source of potential deadlocks, since it is called by
drivers but may call back into drivers, so drivers had
to make sure to call it without any locks held, which
frequently lead to more complex code in drivers. Avoid
that problem by allowing the function to be called in
any context, and queueing the actual work it does.
Also update the documentation for it to indicate this.
Bob Copeland [Wed, 25 Aug 2010 01:37:14 +0000 (21:37 -0400)]
ath5k: fix off-by-one in pilot magnitude mask
If the symbol offset is 46, it will be counted in both
the third and fourth bytes of the mask, and in this
case the shift will be negative which can pollute
high order bits in the mask. This may negatively impact
OFDM symbol detection.
Johannes Berg [Mon, 23 Aug 2010 08:46:58 +0000 (10:46 +0200)]
iwlwifi: allow using multiple contexts
We're now ready to start using multiple contexts.
We do this by keeping track of the valid interface
types per context (exclusive [ibss] and normal)
and checking which context is "free" when a new
interface is added.
Johannes Berg [Mon, 23 Aug 2010 08:46:57 +0000 (10:46 +0200)]
iwlwifi: follow main beacon interval
It is necessary that the PAN context always
use the same beacon interval as the BSS
context unless it is in dual-station mode,
ie. the PAN context is a station as well,
so implement that.
Johannes Berg [Mon, 23 Aug 2010 08:46:55 +0000 (10:46 +0200)]
iwlwifi: pass context to iwl_send_rxon_timing
Sometimes we need to send RXON timing even
when we don't have a virtual interface yet,
so pass the context and allow passing one
without a virtual interface pointer.
Johannes Berg [Fri, 27 Aug 2010 16:44:50 +0000 (09:44 -0700)]
iwlwifi: send PAN parameters
In order for the microcode to be able to handle
multiple interfaces, we need to give it the PAN
parameters that state how to allocate the time
between the two interfaces. Do this, and update
it wherever necessary.
Johannes Berg [Mon, 23 Aug 2010 08:46:52 +0000 (10:46 +0200)]
iwlagn: send RXON timing before associating
The PAN functionality requires us to send the
timing including a valid DTIM period to the
microcode before associating, so request this
data from mac80211 and send it to the device.
Johannes Berg [Mon, 23 Aug 2010 08:46:51 +0000 (10:46 +0200)]
iwlagn: queue frames according to context
Frames for different contexts need to be put
on different queues, and multicast after DTIM
frames have a special queue yet which also
depends on the context, so put all this into
the context.
Johannes Berg [Mon, 23 Aug 2010 08:46:50 +0000 (10:46 +0200)]
iwlagn: remove iw_mode check for beacon
Since we have the beacon context now, we no
longer need to rely on iw_mode but can check
the beacon context interface's type. However,
that check must be in the work item instead
due to locking constraints (mutex must be
held when dereferencing beacon_ctx pointer).
Johannes Berg [Mon, 23 Aug 2010 08:46:49 +0000 (10:46 +0200)]
iwlagn: detect PAN capability
Detect whether or not the ucode is PAN
capable and adjust the valid contexts
accordingly. To be able to do this, add
the PAN context to the array as well.
Johannes Berg [Mon, 23 Aug 2010 08:46:48 +0000 (10:46 +0200)]
iwlwifi: add PAN API
Define the new host commands and notifications
REPLY_WIPAN_PARAMS
REPLY_WIPAN_RXON
REPLY_WIPAN_RXON_TIMING
REPLY_WIPAN_RXON_ASSOC
REPLY_WIPAN_QOS_PARAM
REPLY_WIPAN_WEPKEY
REPLY_WIPAN_P2P_CHANNEL_SWITCH
REPLY_WIPAN_NOA_NOTIFICATION
and their corresponding structures along with the PAN
station flag, the PAN AP sta ID and new dev types for
the second context.
Johannes Berg [Mon, 23 Aug 2010 08:46:46 +0000 (10:46 +0200)]
iwlwifi: move HT configuration data into context
A lot of HT configuration semantically belongs into
the context, even if right now it will never be
different between contexts. Move it so we're better
prepared for future changes in mac80211.
Johannes Berg [Mon, 23 Aug 2010 08:46:44 +0000 (10:46 +0200)]
iwlwifi: store default station flags in context
Since the default context is initialised to zero,
and the default flags are zero, no more code is
needed to initialise them right now, but another
context can have different default flags.
Johannes Berg [Mon, 23 Aug 2010 08:46:43 +0000 (10:46 +0200)]
iwlwifi: make hw crypto context aware
HW crypto needs to be aware of the context, and there
are different command IDs for the WEP keys per context,
so move the key tracking variables and command IDs into
the context structure.
Johannes Berg [Mon, 23 Aug 2010 08:46:40 +0000 (10:46 +0200)]
iwlwifi: move virtual interface pointer into context
iwlwifi occasionally needs to find the virtual
interface pointer to give it to mac80211, but right
now it only keeps one. Move it into the context so
that we can keep one pointer each.
Johannes Berg [Mon, 23 Aug 2010 08:46:38 +0000 (10:46 +0200)]
iwlwifi: move QoS into context
Since mac80211 doesn't currently enable/disable
QoS per interface, we can't yet do it properly,
but we can already prepare for it and move the
QoS data and command ID into the context struct.
Johannes Berg [Mon, 23 Aug 2010 08:46:37 +0000 (10:46 +0200)]
iwlwifi: contextify command sending
Some commands will have different command IDs
for different contexts, so we need to store
those IDs in the context structure and use
them instead of hardcoding the commands.
Johannes Berg [Fri, 27 Aug 2010 15:55:52 +0000 (08:55 -0700)]
iwlwifi: contextify-stations-completely
The microcode tracks stations per context, so
the driver needs to do that as well for adding,
deleting and restoring them, especially in the
implicit removal case when we send an RXON.
Johannes Berg [Mon, 23 Aug 2010 08:46:33 +0000 (10:46 +0200)]
iwlwifi: prepare for PAN queue/fifo assignment
PAN ucode will require a different queue assignment,
in particular queue 9 instead of 4 should be used for
commands.
This is required because the ucode will stop/start
queues 4 and 8 depending on the PAN state, since
queue 8 will be used for PAN multicast (after DTIM).
Johannes Berg [Mon, 23 Aug 2010 08:46:32 +0000 (10:46 +0200)]
iwlwifi: initial contextification
In order to support multiple interfaces, we must move
a lot of data into per-context structures so we can
use the contexts the device offers. To start with,
this makes a lot of code context-aware, more changes
will move more things into the context structure.
If there is a FW hang when the driver loads, it can not determine the FW operational
mode. Fix it by checking the FW state first before issuing any FW commands to
determine its capabilities and thereby detecting driver operational mode.
Changli Gao [Tue, 24 Aug 2010 13:32:58 +0000 (13:32 +0000)]
net: use scnprintf() to avoid potential buffer overflow
strlcpy() returns the total length of the string they tried to create, so
we should not use its return value without any check. scnprintf() returns
the number of characters written into @buf not including the trailing '\0',
so use it instead here.
Joe Perches [Tue, 24 Aug 2010 13:21:08 +0000 (13:21 +0000)]
net/sctp: Use pr_fmt and pr_<level>
Change SCTP_DEBUG_PRINTK and SCTP_DEBUG_PRINTK_IPADDR to
use do { print } while (0) guards.
Add SCTP_DEBUG_PRINTK_CONT to fix errors in log when
lines were continued.
Add #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
Add a missing newline in "Failed bind hash alloc"
Simon Horman [Thu, 26 Aug 2010 02:54:31 +0000 (02:54 +0000)]
ipvs: switch to GFP_KERNEL allocations
Switch from GFP_ATOMIC allocations to GFP_KERNEL ones in
ip_vs_add_service() and ip_vs_new_dest(), as we hold a mutex and are
allowed to sleep in this context.
Rasesh Mody [Thu, 26 Aug 2010 06:00:27 +0000 (23:00 -0700)]
bna: Fixed build break for allyesconfig
This is the patch to fix the build break caused by multiple
definitions of symbols between Brocade's FC/FCOE driver(BFA)
and 10G Networking Driver(BNA).
Changes are:
1. locally used functions are made static
2. unused functions are removed
3. using unique namespaces for the function names that must be
globally visible
Ajit Khaparde [Wed, 25 Aug 2010 00:32:33 +0000 (00:32 +0000)]
be2net: fix to dynamically generate MAC Address for VFs
The BE ASIC/firmware doesnot reserve and assign MAC address for VFs.
This results in the VF interfaces being created with MAC Address 0.
The code change proposed takes the MAC address of PF to generate a seed.
MAC Address for VFs are assigned incrementally starting from the seed.
These addresses are programmed in the ASIC by the PF and the VF driver
queries for the MAC address during its probe.
Breno Leitao [Tue, 24 Aug 2010 12:50:40 +0000 (12:50 +0000)]
qlge: Fix a deadlock when the interface is going down
Currently qlge can deadlock when the interface is going
down, and the mpi_port_cfg_work() is executing on another
processor. It happens because unregister_netdev() holds
the rtnl lock, and the mpi_port_cfg_work() also request
this lock.
Since unregiter_netdev() may wait mpi_port_cfg_work(), who
also request the holding lock, it can cause an deadlock,
displaying the following error:
Julia Lawall [Tue, 24 Aug 2010 04:39:49 +0000 (04:39 +0000)]
net/netfilter/ipvs: Eliminate memory leak
__ip_vs_service_get and __ip_vs_svc_fwm_get increment a reference count, so
that reference count should be decremented before leaving the function in an
error case.
A simplified version of the semantic match that finds this problem is:
(http://coccinelle.lip6.fr/)
x = __ip_vs_service_get(...);
<... when != x
when != true (x == NULL || ...)
when != if (...) { <+...x...+> }
when != I (...) { <+...x...+> }
(
x == NULL
|
x == E
|
x->f1
)
...>
* return ...;
// </smpl>
x = dev_alloc_skb(...);
<... when != x
when != true (x == NULL || ...)
when != if (...) { <+...x...+> }
when != I (...) { <+...x...+> }
(
x == NULL
|
x == E
|
x->f1
)
...>
* return ...;
// </smpl>
stmmac: remove dead option in the driver's Kconfig
This patch removes the CPU_SUBTYPE_ST40 dependency in the
driver's Kconfig.
In fact, this option has been removed in the commit: f96691872439ab2071171d4531c4a95b5d493ae5
as reported by Christian Dietrich.
Note that the driver remains tested on STM platforms, only.
Pawel Moll [Mon, 23 Aug 2010 20:40:40 +0000 (20:40 +0000)]
stmmac: fix_mac_speed is called during 10/100<->1000 speed changes
This patch modifies the stmmac_adjust_link() function so the
fix_mac_speed() is called not only when link speed is changing
between 10 and 100 Mbps (as required in RMII mode) but also
for 1000 Mbps.
On NIC Partition capable adapter, Administrator can configure to
tag packet with particular vlan id. Packet will be tagged and strip with
that vlan id. Also if 'Tagging' flag is disable, other packet will be drop.
Johannes Berg [Wed, 25 Aug 2010 12:34:01 +0000 (14:34 +0200)]
mac80211: fix mesh advertisement
When a mac80211-based driver advertises mesh mode
support, this will be advertised to userspace.
However, if mac80211 was compiled without mesh
support, then that won't actually be true. Fix
this by removing the bit for mesh if mesh isn't
compiled in.
Since this synchronizes what we advertise to
cfg80211 and actually support, it means we can
now rely on cfg80211's interface type checks
and need not check again in mac80211.
This patch fixes a potential crash (null-pointer de-
reference) which was introduced in my previous patch:
"mac80211: AMPDU rx reorder timeout timer"
During a BA teardown, the pointer to the soon-to-be-gone
tid_ampdu_rx element will be nullified. Therefore the
release timer mechanism has to be careful not to
accidentally access the item without any RCU protection.
introduced a netlink bug that caused parsing errors
in userspace because it forgot to close a nesting,
which would advertise a nesting length of zero to
userspace, which then completely threw off parsing
and led to
Ivo van Doorn [Mon, 23 Aug 2010 17:56:07 +0000 (19:56 +0200)]
rt2x00: Fix max TX power settings
During initialization each driver reads the default TX power
for each individual channel. However mac80211 only accepts the
maximum value (which is also handled as default value).
As a result, the TX power of the device was being limited to
the default value, which is often quite low compared to the
real maximum acceptable value.
This patch allows each driver to set the maximum value on a
per-channel basis which is forwarded to mac80211. The default
value will be preserved for now, in case we want to update
mac80211 to differentiate between the maximum and default txpower.
This fixes bug complaining about limited TX power values like:
https://bugzilla.kernel.org/show_bug.cgi?id=16358
Ivo van Doorn [Mon, 23 Aug 2010 17:55:43 +0000 (19:55 +0200)]
rt2x00: Fix channel configuration for RF3052
We missed a check for RF3052 in the config_channel
configuration function. As a result the channel is not
properly configured for RF3052 devices. This also
requires rt2800_config_channel_rf3xxx to support 5GHz
channels, so we must apply the TXPOWER_A_TO_DEV and
TXPOWER_G_TO_DEV conversions during rt2800_config_channel()
Ivo van Doorn [Mon, 23 Aug 2010 17:55:22 +0000 (19:55 +0200)]
rt2x00: Merge rt2800{pci/usb} radio enabling/disabling code to rt2800lib
The functions rt2800pci_enable_radio and rt2800usb_disable_radio are
almost equal and can be merged into rt2800lib. This reduces the number
of functions which must be exported from rt2800lib to the drivers at
the same time.
Also rt2800pci_disable_radio and rt2800usb_disable_radio are almost
equal and can be merged into rt2800lib in a similar fashion.
Ivo van Doorn [Mon, 23 Aug 2010 17:55:01 +0000 (19:55 +0200)]
rt2x00: Fix rt2800 retry calculation
The retry count for a frame is calculated by the realized MCS rate
subtracting of the requested MCS rate. However during TX done reporting
we accidently override the requested MCS rate with the realized MCS
rate. This causes the retry count to be reset to 0 always, and feeding
invalid rate information into minstrel_ht.
Ivo van Doorn [Mon, 23 Aug 2010 17:54:41 +0000 (19:54 +0200)]
rt2x00: Remove Q_INDEX_CRYPTO and ENTRY_OWNER_DEVICE_CRYPTO
Q_INDEX_CRYPTO and ENTRY_OWNER_DEVICE_CRYPTO where initially
added with the plan to implement hardware encryption for rt2500pci.
However there are no plans to do so anymore as the crypto
mechanism in rt2500pci is very different then any other Ralink
device.
So we can now safely remove the definitions, and don't think
about this feature anymore. :)
Ivo van Doorn [Mon, 23 Aug 2010 17:54:21 +0000 (19:54 +0200)]
rt2x00: Move direct access to queue->entries to rt2x00queue.c
All access to queue->entries through the Q_INDEX/Q_INDEX_DONE
variables must be done using spinlock protection. It is best
to manage this completely from rt2x00queue.c.
For safely looping through all entries in the queue, the function
rt2x00queue_for_each_entry is added which will walk from from a index
range in a safe manner.
This also fixes rt2x00usb which walked the entries list from
0 to length to kill each entry (killing entries must be done
from Q_INDEX_DONE to Q_INDEX to enforce TX status reporting to
occur in the correct order.
projects / linux.git / log