15 Dec, 2011
14 commits
-
Without an API for setting and getting the local and remote general bytes,
drivers won't be able to properly establish a DEP link.
This API also allows them to propagate the remote general bytes they get
from the DEP link establishment up to the LLCP layer.Signed-off-by: Samuel Ortiz
Signed-off-by: John W. Linville -
NFC-DEP (Data Exchange Protocol) is an NFC MAC layer.
This command allows to enable and disable the DEP link on to which e.g.
LLCP can run.Signed-off-by: Samuel Ortiz
Signed-off-by: John W. Linville -
rawsock_create() is called with preemption disabled, so we should not
sleep.Signed-off-by: Samuel Ortiz
Signed-off-by: John W. Linville -
The netlink notifier is atomic so we must not sleep in that context.
Also we know that Any netlink packets arriving to us will be purged when
the notifier is called, so we don't need to take the mutex.Signed-off-by: Samuel Ortiz
Signed-off-by: John W. Linville -
This is a factorization of the current rawsock tx skb allocation routine,
as it will be used by the LLCP code.
We also rename nfc_alloc_skb to nfc_alloc_recv_skb for consistency sake.Signed-off-by: Samuel Ortiz
Signed-off-by: John W. Linville -
Signed-off-by: Samuel Ortiz
Signed-off-by: John W. Linville -
ieee80211_free_txskb should be used when dropping a frame in the device
rx path such that mac80211 knows about this frame being dropped.Signed-off-by: Helmut Schaa
Acked-by: Gertjan van Wingerde
Acked-by: Ivo van Doorn
Signed-off-by: John W. Linville -
Our new return also created a memleak. The skb should be freed before
returning an error.Signed-off-by: Simon Wunderlich
Signed-off-by: John W. Linville -
cfg_workqueue was added to notify cfg80211 that scan, connect
or disconnect is done by calling respective completion handlers.
We can avoid use of this workqueue by calling those handlers
from other places.
1) Call connect, disconnect completion handlers in their callback
functions.
ex. Call cfg80211_connect_result() in mwifiex_cfg80211_connect()
2) Call scan completion handler after parsing response of last scan
command in a queue.After removing the workqueue, variables (assoc_request etc.) and
checks used for mutual exclusion become redundant. Those are also
removed in this patch.Signed-off-by: Amitkumar Karwar
Signed-off-by: Yogesh Ashok Powar
Signed-off-by: Nishant Sarmukadam
Signed-off-by: Kiran Divekar
Signed-off-by: Bing Zhao
Signed-off-by: John W. Linville -
All nl80211 commands that need only the wiphy
still allow identifying it by giving an interface
index, except, as Kenny pointed out, the testmode
dump support.Fix this by looking up the wiphy via the ifidx in
this case as well.Tested-by: Kenny Hsu
Signed-off-by: Johannes Berg
Signed-off-by: John W. Linville -
Conflicts:
drivers/net/wireless/iwlwifi/iwl-agn.c -
My previous patch
34a5b4b6af104cf18eb50748509528b9bdbc4036 iwlwifi: do not re-configure
HT40 after associatedFix the case of HT40 after association on specified AP, but it break the
association for some APs and cause not able to establish connection.
We need to address HT40 before and after addociation.CC: stable@vger.kernel.org #3.0+
Reported-by: Andrej Gelenberg
Signed-off-by: Wey-Yi Guy
Tested-by: Andrej Gelenberg
Signed-off-by: John W. Linville -
Ted reported that he couldn't connect to some APs
and bisected it to the tx_sync implementation.
Disable it for the BSS context to fix this issue.Reported-by: Ted Ts'o
Signed-off-by: Johannes Berg
Signed-off-by: Wey-Yi Guy
Signed-off-by: John W. Linville -
Command cancel path cancels the current command and moves
it to free command queue. While doing that it deletes the
command entry from the pending list. This is not correct
as the entry has been already deleted from the pending
list at 'mwifiex_exec_next_cmd'. Fixing it.Also making sure the stale command pointer is cleaned and
unaccessible for later use.Signed-off-by: Yogesh Ashok Powar
Signed-off-by: Bing Zhao
Signed-off-by: John W. Linville
14 Dec, 2011
26 commits
-
Avoids this:
ERROR: "bcma_chipco_pll_write" [drivers/net/wireless/b43/b43.ko] undefined!
Signed-off-by: John W. Linville
-
mwifiex driver no longer supports it's own custom regulatory rules,
but custom regulatory domain capability is still advertised during
wiphy registration by the driver.Signed-off-by: Amitkumar Karwar
Signed-off-by: Bing Zhao
Signed-off-by: John W. Linville -
The brcmsmac driver has been verified on chipsets that were supported
when it was a pci device driver, ie. bcm4313, bcm43224, and bcm43225.
This patch restricts the driver to 802.11 core revisions that are found
in these chipsets.Reviewed-by: Pieter-Paul Giesberts
Reviewed-by: Alwin Beukers
Signed-off-by: Arend van Spriel
Signed-off-by: Franky Lin
Signed-off-by: John W. Linville -
Instead of storing the buscore information now the BCMA core device
is kept for quick reference in si_info structure.Reviewed-by: Pieter-Paul Giesberts
Reviewed-by: Alwin Beukers
Signed-off-by: Arend van Spriel
Signed-off-by: Franky Lin
Signed-off-by: John W. Linville -
Number of fields are no longer needed as the BCMA provides it
or makes them redundant. These have been removed.Reviewed-by: Pieter-Paul Giesberts
Reviewed-by: Alwin Beukers
Signed-off-by: Arend van Spriel
Signed-off-by: Franky Lin
Signed-off-by: John W. Linville -
In aiutils.c the selected core was maintained by its index number. This
is obsolete using BCMA functions so several functions using that index
have been removed.Reviewed-by: Pieter-Paul Giesberts
Reviewed-by: Alwin Beukers
Signed-off-by: Arend van Spriel
Signed-off-by: Franky Lin
Signed-off-by: John W. Linville -
The function ai_switch_core() is no longer needed and its counterpart
ai_restore_core() as well, because interrupts disabling is not needed
anymore.Reviewed-by: Pieter-Paul Giesberts
Reviewed-by: Alwin Beukers
Signed-off-by: Arend van Spriel
Signed-off-by: Franky Lin
Signed-off-by: John W. Linville -
There is no need to interrupt disable/enable functionality any
longer due to BCMA usage assures the correct core is accessed
in any context.Reviewed-by: Pieter-Paul Giesberts
Reviewed-by: Alwin Beukers
Signed-off-by: Arend van Spriel
Signed-off-by: Franky Lin
Signed-off-by: John W. Linville -
The macros were used to assure that the correct core was accessed in
the ISR, but register access is now done giving the explicit core so
no need to change interrupt state.Reviewed-by: Pieter-Paul Giesberts
Reviewed-by: Alwin Beukers
Signed-off-by: Arend van Spriel
Signed-off-by: Franky Lin
Signed-off-by: John W. Linville -
A number of functions in pmu.c are not used or adding no functionality
at all. These have been removed.Reviewed-by: Pieter-Paul Giesberts
Reviewed-by: Alwin Beukers
Signed-off-by: Arend van Spriel
Signed-off-by: Franky Lin
Signed-off-by: John W. Linville -
The register access macros like R_REG/W_REG/etc. are no longer
needed as the driver uses the BCMA provided functions.Reviewed-by: Pieter-Paul Giesberts
Reviewed-by: Alwin Beukers
Signed-off-by: Arend van Spriel
Signed-off-by: Franky Lin
Signed-off-by: John W. Linville -
The code in aiutils.c now uses the BCMA function for control the
registers in the device cores.Reviewed-by: Pieter-Paul Giesberts
Reviewed-by: Alwin Beukers
Signed-off-by: Arend van Spriel
Signed-off-by: Franky Lin
Signed-off-by: John W. Linville -
The code in pmu.c now uses the functions provided by BCMA to
access the core registers.Reviewed-by: Pieter-Paul Giesberts
Reviewed-by: Alwin Beukers
Signed-off-by: Arend van Spriel
Signed-off-by: Franky Lin
Signed-off-by: John W. Linville -
The code in srom.c now uses the core access function provided by
BCMA so no need to pass __iomem pointer any longer.Reviewed-by: Pieter-Paul Giesberts
Reviewed-by: Alwin Beukers
Signed-off-by: Arend van Spriel
Signed-off-by: Franky Lin
Signed-off-by: John W. Linville -
The code in otp.c now uses the bcma core access functions to
read the OTP information from the device.Reviewed-by: Pieter-Paul Giesberts
Reviewed-by: Alwin Beukers
Signed-off-by: Arend van Spriel
Signed-off-by: Franky Lin
Signed-off-by: John W. Linville -
Instead of returning the core index the function now returns
the bcma device for the requested core id. This function is
now exposed in the header file.Reviewed-by: Pieter-Paul Giesberts
Reviewed-by: Alwin Beukers
Signed-off-by: Arend van Spriel
Signed-off-by: Franky Lin
Signed-off-by: John W. Linville -
Specs were updated.
Signed-off-by: Rafał Miłecki
Signed-off-by: John W. Linville -
Signed-off-by: Rafał Miłecki
Signed-off-by: John W. Linville -
The order is different than on older PHYs.
Signed-off-by: Rafał Miłecki
Signed-off-by: John W. Linville -
Signed-off-by: Rafał Miłecki
Signed-off-by: John W. Linville -
The ip address of the vif can be set even before the
vif is up. requiring the vif to be up in the vif
notifier makes the notifer ignore this event, which
causes wrong arp filter configuration later on.Reported-by: Eyal Shapira
Signed-off-by: Eliad Peller
Signed-off-by: John W. Linville -
Configure arp filtering on sta reconfiguration.
Signed-off-by: Eliad Peller
Signed-off-by: John W. Linville -
ieee80211_configure_filter code used local->scanning as a boolean
value when it was a bit mask. Bits SCAN_COMPLETED, SCAN_ABORTED
should not set FIF_BCN_PRBRESP_PROMISC filter.SCAN_HW_SCANNING should not set FIF_BCN_PRBRESP_PROMISC either,
as there is no explicit filter configuration request from
scan code. If a driver requires FIF_BCN_PRBRESP_PROMISC mode
during HW scanning, it's up to the driver to temporary enable it.Similar mistake was fixed also in ieee80211_hw_config (power
configuration code).Verified-by: Vitaly Wool
Signed-off-by: Dmitry Tarnyagin
Signed-off-by: John W. Linville -
With previous patch "rtlwifi: use work for lps" we can now use mutex for
protecting ps mode changing critical sections. This fixes running system
with interrupts disabled for long time.Merge ips_lock and lps_lock as they seems to protect the same data
structures (accessed in rtl_ps_set_rf_state() function).Reported-by: Philipp Dreimann
Tested-by: Larry Finger
Cc: Mike McCormack
Cc: Chaoming Li
Signed-off-by: Stanislaw Gruszka
Signed-off-by: Larry Finger
Tested-by: Tim Gardner
Signed-off-by: John W. Linville -
Leaving leisure power save mode can take some time, so it's better to
perform that action in process context with interrupts enabled. This
patch changes lps_leave tasklet to work.Reported-by: Philipp Dreimann
Tested-by: Larry Finger
Cc: Mike McCormack
Cc: Chaoming Li
Signed-off-by: Stanislaw Gruszka
Signed-off-by: Larry Finger
Signed-off-by: John W. Linville -
Signed-off-by: Rafał Miłecki
Signed-off-by: John W. Linville