Arik Nemtsov [Sun, 14 Aug 2011 10:17:34 +0000 (13:17 +0300)]
wl12xx: handle wrap-around overflow in released Tx blocks FW counter
When the FW Tx released blocks counter wraps around, we should correct
our calculation of released blocks. Otherwise we add a large negative
figure to our driver freed blocks counter
Arik Nemtsov [Sun, 14 Aug 2011 10:17:33 +0000 (13:17 +0300)]
wl12xx: schedule TX packets according to FW packet occupancy
When selecting packets for transmission, prefer the ACs that are least
occupied in the FW. When packets for multiple ACs are present in the FW,
it decides which to transmit according to WMM QoS parameters.
With these changes, lower priority ACs should not be starved when higher
priority traffic is present.
Eliad Peller [Sun, 14 Aug 2011 10:17:30 +0000 (13:17 +0300)]
wl12xx: use ap_bcast_hlid for recorded keys
when the key was recorded, wl->ap_bcast_hlid was invalid
(since the role wasn't started), so when configuring the
key we need to use the current ap_bcast_hlid.
Arik Nemtsov [Sun, 14 Aug 2011 10:17:29 +0000 (13:17 +0300)]
wl12xx: AP-mode - configure HT rate support to the FW
Unconditionally configure HT rate support to the FW on all ACs
when starting the AP.
When 11n support is disabled by usermode (hostapd), each STA joining
the AP will appear as a non-HT STA. This will stop us from accidentally
transmitting using MCS rates.
Eliad Peller [Sun, 14 Aug 2011 10:17:17 +0000 (13:17 +0300)]
wl12xx: replace dummy_join with ROC/CROC commands
The ROC command asks the fw stay on the channel of the given
hlid. it currently has 2 primary functions:
1. Allow tx/rx from the device role.
In order to tx/rx packets while the stations is not associated
(e.g. auth req/resp), the device role has to be used, along
with ROC on its link.
Keep the logic similiar to the one used in dummy_join. However,
since we can't scan while we ROC, we add CROC before starting
a scan, and ROC again (if needed) on scan complete.
2. Keeping the antenna for a specific link.
We ROC until the connection was completed (after EAPOLs exchange)
in order to prevent BT coex operations from taking the antenna
and failing the connection (after this stage, psm can be used).
During association, we ROC on the station role, and then CROC
the device role, thus assuring being ROC during all the connection
process.
Delete the WL1271_FLAG_JOINED flag, and use a roc bitmap
to indicate what roles are currently ROCed.
Add wl12xx_roc/croc functions in order to wrap the roc/croc
commands while taking care of the roc bitmap.
The current ROC/CROC state-machine is a bit complicated. In
the future we'll probably want to use wpa_supplicant to control
the ROC during connection.
Eliad Peller [Sun, 14 Aug 2011 10:17:15 +0000 (13:17 +0300)]
wl12xx: add system_hlid
system_hlid is a const hlid (always 0), used by the fw and driver
for packets which are not bound to specific role (e.g. dynamic
memory packets).
indicate it as always allocated.
Eliad Peller [Sun, 14 Aug 2011 10:17:13 +0000 (13:17 +0300)]
wl12xx: use wl1271_acx_beacon_filter_opt for both sta and ap
Use ACX_BEACON_FILTER_OPT for both station and ap roles
(use the generic wl1271_acx_beacon_filter_opt()
instead of wl1271_acx_set_ap_beacon_filter() ).
Eliad Peller [Sun, 14 Aug 2011 10:17:12 +0000 (13:17 +0300)]
wl12xx: change max/default template size
The max template size was increased in the new fw.
However, we should use the max size only when needed, as it
consumes some of the chip's memory.
Thus, by default initialize the templates to the default size.
Initialize to the maximum size only when required.
Use WL1271_CMD_TEMPL_DFLT_SIZE instead of some of the
predefined structs, as some of them didn't account
for additional IEs that might be added to the template.
Delete structs defintions not used after these changes.
Eliad Peller [Sun, 14 Aug 2011 10:17:07 +0000 (13:17 +0300)]
wl12xx: update commands & events
Change the commands and events according to the new fw api
(fw >= 6/7.3.0.0.75).
The main change is the replacement of JOIN/DISCONNECT commands,
with ROLE_START/ROLE_STOP commands.
The use of these commands should be preceded by the ROLE_ENABLE
command (allocating role resources), and followed by the
ROLE_DISABLE command (freeing role resources).
Eliad Peller [Sun, 14 Aug 2011 10:17:06 +0000 (13:17 +0300)]
wl12xx: update acx commands
Update the acx commands according to the new fw api
(fw >= 6/7.3.0.0.75).
The main change in most of the ACXs is the addition
of a new role_id/link_id field, which is required
for multi-role operation.
Currently, we don't really support multi-role, as
most of our data (inside wl) is global.
As the current fw doesn't support concurrent roles
yet, keep it this way and add wl->role_id and
wl->sta_hlid to save the active role/link.
Fix the value of PG version for 128x at sysfs, remove write permissions
from PG version (hw_pg_ver) in sysfs and add remove files
(hw_pg_ver,bt_coex_state) from sysfs while freeing hardware.
New macro names for register Fuse_data_2_1 depend on architecture.
Propagate chip id through wiphy in PLT mode which still not work of
a bug in ethtool.
Shahar Levi [Thu, 14 Jul 2011 08:50:27 +0000 (11:50 +0300)]
wl12xx: Add support to RX packets payload alignment
In case of QoS packets the packet payload isn't aligned to
4 bytes. In that case the mac80211 layer take care of that
via memmove() in ieee80211_deliver_skb().
Add support of copy packets from aggregation buffer to the
skbs with packet payload aligned care. In case of QoS packets
copy the packets in offset of 2 bytes guarantee payload aligned
to 4 bytes.
Signed-off-by: Shahar Levi <shahar_levi@ti.com> Signed-off-by: Luciano Coelho <coelho@ti.com>
Jouni Malinen [Thu, 11 Aug 2011 08:46:22 +0000 (11:46 +0300)]
nl80211/cfg80211: Make addition of new sinfo fields safer
Add a comment pointing out the use of enum station_info_flags for
all new struct station_info fields. In addition, memset the sinfo
buffer to zero before use on all paths in the current tree to avoid
leaving uninitialized pointers in the data.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Felix Fietkau [Thu, 11 Aug 2011 01:00:33 +0000 (19:00 -0600)]
cfg80211: fix a crash in nl80211_send_station
mac80211 leaves sinfo->assoc_req_ies uninitialized, causing a random
pointer memory access in nl80211_send_station.
Instead of checking if the pointer is null, use sinfo->filled, like
the rest of the fields.
Signed-off-by: Felix Fietkau <nbd@openwrt.org> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Javier Cardona [Mon, 8 Aug 2011 23:30:50 +0000 (16:30 -0700)]
mac80211: fix erroneous clearing of MESH_PATH_SN_VALID flag
When a PREQ or PREP is received from an intermediate node, it contains
useful information for path selection but it doesn't include the
originator's sequence number. Therefore, when updating the mesh path
to that intermediate node, we should not set the MESH_PATH_SN_VALID
flag. BUT, if the flag is set, it should not be unset as we might have
received a valid sequence number for that intermediate node in the past.
This issue was reported, fixed and tested by Ya Bo (游波) and Pedro
Larbig (ASPj).
Signed-off-by: Javier Cardona <javier@cozybit.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Eliad Peller [Mon, 8 Aug 2011 13:50:22 +0000 (16:50 +0300)]
mac80211: clear sta.drv_priv on reconfiguration
drivers might assume sta.drv_priv is clear while
the sta is added, so clear it on reconfinguration.
Signed-off-by: Eliad Peller <eliad@wizery.com> Acked-by: Johannes Berg <johannes@sipsolutions.net> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Jouni Malinen [Mon, 8 Aug 2011 09:11:52 +0000 (12:11 +0300)]
cfg80211/nl80211: Send AssocReq IEs to user space in AP mode
When user space SME/MLME (e.g., hostapd) is not used in AP mode, the
IEs from the (Re)Association Request frame that was processed in
firmware need to be made available for user space (e.g., RSN IE for
hostapd). Allow this to be done with cfg80211_new_sta().
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com> Acked-by: Johannes Berg <johannes@sipsolutions.net> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Jouni Malinen [Mon, 8 Aug 2011 09:10:30 +0000 (12:10 +0300)]
nl80211: Indicate driver-based offchannel TX on mgmt_tx_cancel_wait
Drivers that support frame transmission with mgmt_tx() may not support
driver-based offchannel TX. Use mgmt_tx_cancel_wait instead of mgmt_tx
when figuring out whether to indicate support for this with
NL80211_ATTR_OFFCHANNEL_TX_OK.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com> Acked-by: Johannes Berg <johannes@sipsolutions.net> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Jouni Malinen [Sat, 6 Aug 2011 20:07:00 +0000 (23:07 +0300)]
mac80211_hwsim: Fix RX status reporting for HT
RX_FLAG_HT must be included when reporting MCS rates. Without
this, mac80211 ended up dropping any frame sent at MCS index 12
or higher and that resulted in oddly random looking errors in
mac80211_hwsim tests.
Signed-off-by: Jouni Malinen <j@w1.fi> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Kalle Valo [Wed, 10 Aug 2011 06:49:12 +0000 (09:49 +0300)]
ath6kl: fix function name conflicts with ath9k
Stephen reported that compilation fails if both ath6kl and ath9k are
compiled in:
drivers/net/wireless/ath/ath6kl/built-in.o: In function `htc_start':
(.opd+0x600): multiple definition of `htc_start'
drivers/net/wireless/ath/ath9k/built-in.o:(.opd+0x3e40): first defined here
drivers/net/wireless/ath/ath6kl/built-in.o: In function `.htc_stop':
(.text+0x7b40): multiple definition of `.htc_stop'
drivers/net/wireless/ath/ath9k/built-in.o:(.text+0x67b34): first defined he=
re
drivers/net/wireless/ath/ath6kl/built-in.o: In function `.htc_start':
(.text+0x7d18): multiple definition of `.htc_start'
drivers/net/wireless/ath/ath9k/built-in.o:(.text+0x67ba0): first defined he=
re
drivers/net/wireless/ath/ath6kl/built-in.o: In function `htc_stop':
(.opd+0x5e8): multiple definition of `htc_stop'
drivers/net/wireless/ath/ath9k/built-in.o:(.opd+0x3e28): first defined here
To fix this add ath6kl prefix to all public functions in htc.c.
Reported-by: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
Larry Finger [Sat, 30 Jul 2011 16:30:27 +0000 (11:30 -0500)]
b43: Remove EXPERIMENTAL designation from LP PHY selection
Since kernel 3.0, the problems with controlling b43 devices that have
low-power (LP) PHYs have been fixed and the EXPERIMENTAL designation
can be fixed.
This patch also fixes a typo as the device supports 802.11b communications.
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net> Acked-by: Rafał Miłecki <zajec5@gmail.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Helmut Schaa [Sat, 6 Aug 2011 11:13:48 +0000 (13:13 +0200)]
rt2x00: Fix PCI interrupt processing race on SMP systems
When toggle_irq is called for PCI devices to disable device interrupts
it used tasklet_disable to wait for a possibly running tasklet to finish.
However, on SMP systems the tasklet might still be scheduled on another CPU.
Instead, use tasklet_kill to ensure that all scheduled tasklets are finished
before returning from toggle_irq.
Furthermore, it was possible that a tasklet reenabled its interrupt even
though interrupts have been disabled already. Fix this by checking the
DEVICE_STATE_ENABLED_RADIO flag before reenabling single interrupts
during tasklet processing.
While at it also enable/kill the TBTT and PRETBTT tasklets in the
toggle_irq callback and only use tasklet_kill in stop_queue to wait
for a currently scheduled beacon update before returning.
Signed-off-by: Helmut Schaa <helmut.schaa@googlemail.com> Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Larry Finger [Sat, 6 Aug 2011 10:55:18 +0000 (05:55 -0500)]
rtlwifi: Eliminate raw reads and writes from PCIe portion
The PCIe driver used raw reads and writes on the PCIe hardware.
As all of these are only affecting the configuration space, all of
then can be converted to pci_{read,write}_config_XX calls.
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net> Signed-off-by: John W. Linville <linville@tuxdriver.com>
ath9k: Use atomic reference count for interrupt ops
Let us enable/disable interrupts based on reference count.
By doing this we can ensure that interrupts are never be
enabled in the middle of tasklet processing. Instead of
addressing corner cases like "ath9k: avoid enabling interrupts
while processing rx", this approach handles it in generic manner.
Signed-off-by: Rajkumar Manoharan <rmanohar@qca.qualcomm.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Spur frequency was incorrectly computed with 10Mhz offset
which could cause the filter would not notch out the spur
and also this could improve rx sensitivity in HT40.
Cc: Madhan Jaganathan <madhanj@qca.qualcomm.com> Signed-off-by: Rajkumar Manoharan <rmanohar@qca.qualcomm.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
ath9k: do not enable interrupt on set interrupt mask
At preset set_interrupt also enables interrupt after changing
mask. This is not necessary in all cases and also sometime it
breaks the assumption that interrupt was disabled. So let us
enable the interrupt explicity if it was disabled earlier.
This could also avoid unnecessary register ops and also helps
the follow up patch to have global ref count for interrupts ops.
Signed-off-by: Rajkumar Manoharan <rmanohar@qca.qualcomm.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Daniel Drake [Mon, 1 Aug 2011 15:43:13 +0000 (16:43 +0100)]
libertas: disable functionality when interface is down
Modify the driver so that it does not function when the interface is
down, in preparation for runtime power management.
No commands can be run while the interface is down, so the ndo_dev_stop
routine now directly does all necessary work (including asking the device
to disconnect from the network and disabling multicast functionality)
directly.
power_save and power_restore hooks are added meaning that card drivers
can take steps to turn the device off when the interface is down.
The MAC address can now only be changed when all interfaces are down;
the new address will be programmed when an interface gets brought up.
This matches mac80211 behaviour.
Also, some small cleanups/simplifications were made in the surrounding
device handling logic.
Signed-off-by: Daniel Drake <dsd@laptop.org> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Helmut Schaa [Mon, 1 Aug 2011 09:32:53 +0000 (11:32 +0200)]
mac80211: Don't use EAPOL frames for rate sampling
Signed-off-by: Helmut Schaa <helmut.schaa@googlemail.com> Acked-by: Felix Fietkau <nbd@openwrt.org> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Helmut Schaa [Mon, 1 Aug 2011 09:32:52 +0000 (11:32 +0200)]
mac80211: Fill in skb->protocol information for injected frames
Some drivers (ath9k for example) are using skb->protocol to treat EAPOL
frames somehow special (disallow aggregation for example).
When running in AP mode hostapd injects the EAPOL frames through a
monitor interface and thus skb->protocol isn't set at all. Hence, if the
injected frame is a data frame and carries a rfc1042 headaer update the
skb->protocol field accordingly.
Signed-off-by: Helmut Schaa <helmut.schaa@googlemail.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
Raja Mani [Thu, 4 Aug 2011 13:56:29 +0000 (19:26 +0530)]
ath6kl: Fix crash during the connection process
Sometimes, the network manager is failing to connect to the AP due
to the below kernel crash message. The reason behind this,
after issuing the connect command to the chip, the chip is sending
disconnect event and then immediately one connect event to the host
in some random cases.
The host driver resets all states (including cfg80211 state machine)
when it receives disconnect event from the chip. But, still the host
driver reports the next received connect event to cfg80211, at that time
cfg80211 SME state would have been in IDLE state, which was causing
the below kernel crash.
Now, host driver's sme state machine is checked every time before
delivering connect event to cfg80211
Kalle Valo [Thu, 21 Jul 2011 07:54:26 +0000 (10:54 +0300)]
ath6kl: change aggreation timeout message from an error to a debug message
When I connect to my Linksys WT610N AP supporting 11n I see a lot of
aggreation timeout errors:
[ 408.885053] ath6kl: aggr timeout (st 3109 end 3140)
[ 463.872108] ath6kl: aggr timeout (st 3671 end 3702)
[ 495.010060] ath6kl: aggr timeout (st 3983 end 4014)
[ 503.604047] ath6kl: aggr timeout (st 4065 end 0)
[ 518.963047] ath6kl: aggr timeout (st 141 end 172)
[ 525.014066] ath6kl: aggr timeout (st 205 end 236)
[ 573.957051] ath6kl: aggr timeout (st 701 end 732)
[ 585.019067] ath6kl: aggr timeout (st 816 end 847)
But still the connection seems to work. To not clutter the logs change
the error message to a debug message. But add a fixme comment so that
this will be investigated.
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
Kalle Valo [Thu, 21 Jul 2011 07:22:50 +0000 (10:22 +0300)]
ath6kl: fix crash when interface is closed but scan is ongoing
When ath6kl module was removed while a scan was ongoing the driver would
crash in ath6kl_cfg80211_scan_complete_event().
Fix the function not to iterate nodes when the scan is aborted. The nodes
are already freed when the module is being unloaded. This patch removes the
null check entirely as the wmi structure is not accessed anymore during
module unload.
Also fix a bug where the status was checked as a bitfield with '&' operator.
But it's not a bitfield, just a regular error code.
This is a port of my patch from ath6kl staging with the same title.
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
Kalle Valo [Thu, 21 Jul 2011 07:04:54 +0000 (10:04 +0300)]
ath6kl: fix atomicity in ath6kl_cfg80211_scan_node()
ath6kl_cfg80211_scan_node() was calling cfg80211_inform_bss_frame()
with CFP_KERNEL but the function is executed with a spin lock taken.
This is wrong and the function must use GFP_ATOMIC instead.
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
Kalle Valo [Thu, 21 Jul 2011 06:40:00 +0000 (09:40 +0300)]
ath6kl: don't force foreground scan when connected
In my setup data transfer stalls when there's data transmission during
scan. After some testing I found out that using background scan
when connected to makes the problem go away. This is more like
a workaround than a proper fix, but as the stall is so severe the
workaround is justified.
With a dual band card this increases scan time when connected from
1.9s to 4.4s. When not connected the scan time is not affected and
is the same 1.9s.
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
projects / linux-beck.git / log