15 Oct, 2007
1 commit
-
deal with signedness of the stuff passed to set_bit() et.al.
Signed-off-by: Al Viro
Signed-off-by: Linus Torvalds
14 Oct, 2007
1 commit
-
b43 wireless needs
linux/drivers/net/wireless/b43/pio.h: In function 'b43_pio_write':
linux/drivers/net/wireless/b43/pio.h:89: error: implicit declaration of function 'mmiowb'linux/drivers/net/wireless/b43/phy.c: In function 'b43_phy_write':
linux/drivers/net/wireless/b43/phy.c:301: error: implicit declaration of function 'mmiowb'linuxdrivers/net/wireless/b43/sysfs.c: In function 'b43_attr_interfmode_store':
linuxdrivers/net/wireless/b43/sysfs.c:147: error: implicit declaration of function 'mmiowb'Signed-off-by: Geert Uytterhoeven
Signed-off-by: Linus Torvalds
11 Oct, 2007
38 commits
-
Make the get-nickname wireless extension actually work. Before
this patch, I could do "iwconfig eth1 nick BLAH" but "iwconfig
eth1" would have still showed "MRVL-USB8388" to me. Hey, and that
was wrong anyway, I'm on a CF card, not on USB :-)Signed-off-by: Holger Schurig
Acked-By: Dan Williams
Signed-off-by: John W. Linville -
This makes scripts/checkincludes.pl happy.
Signed-off-by: Holger Schurig
Signed-off-by: John W. Linville -
* make scan debug output cleaner
* change some LBS_DEB_ASSOC messages to LBS_DEB_SCAN, which is more correct
* move helper functions together
* print function return value in the tracing code at one central locationSigned-off-by: Holger Schurig
Acked-By: Dan Williams
Signed-off-by: John W. Linville -
This fixes three "warning: Using plain integer as NULL pointer"
sparse warnings.Signed-off-by: Holger Schurig
Signed-off-by: John W. Linville -
Fix a debug statement
Signed-off-by: Holger Schurig
Signed-off-by: John W. Linville -
Some versions of gcc replace strstr() calls with a single-character `needle'
parameter by strchr() behind our back. This causes a link error if strchr() is
defined as an inline function in (e.g. on m68k):| drivers/built-in.o: In function `libertas_parse_chan':
| linux/drivers/net/wireless/libertas/debugfs.c:209: undefined reference to `strchr'
| drivers/built-in.o: In function `libertas_parse_ssid':
| linux/drivers/net/wireless/libertas/debugfs.c:260: undefined reference to `strchr'Avoid this by explicitly calling strchr() instead.
Also include , because this file calls lots of str*() routines.
Signed-off-by: Geert Uytterhoeven
Acked-By: Holger Schurig
Acked-By: Dan Williams
Signed-off-by: John W. Linville -
Don't write constants that are (per documentation and struct) u8
as 0x0001, use 0x01 instead. Also remove an useless cast.Signed-off-by: Holger Schurig
Acked-By: Dan Williams
Signed-off-by: John W. Linville -
The 88w8385 chip, using SDIO interface and firmware release 5.0.11p0,
has problems when both unicast and multicast WPA keys are set in one
command. This patch ensures the keys are set independently.The original author of this patch is Marc Pignat
Signed-off-by: Marc Pignat
Signed-off-by: Dan Williams
Signed-off-by: John W. Linville -
Version bump
Signed-off-by: Ivo van Doorn
Signed-off-by: John W. Linville -
When skb_push() is used we should memset the memory before
usage. This will prevent bugs which could occur when the
data is treated as TX descriptor.Signed-off-by: Ivo van Doorn
Signed-off-by: John W. Linville -
By defining rt73usb_get_tsf to NULL we only
have 1 location that needs to be edited
when rt73usb_get_tsf can be enabled again.
This also reduces the number of #ifdefs in
the code which is also a "good thing"Signed-off-by: Ivo van Doorn
Signed-off-by: John W. Linville -
Reorganize configuration handling by creating a extra
structure which contains precalculated values based
on the mac80211 values which are usefull for all
individual drivers.This also fixes the preamble configuration problem,
up untill now preamble was never configured since
by default the rate->val value was used when changing
the mode.
Now rate->val will only be used to set the basic rate mask.
The preamble configuration will now be done correctly
through the erp_ie_changed callback function.Signed-off-by: Ivo van Doorn
Signed-off-by: John W. Linville -
Make some small optimizations by removing
some simple if-statements.Signed-off-by: Ivo van Doorn
Signed-off-by: John W. Linville -
Directly pass a value from the enum dev_state with rt2x00lib_toggle_rx,
this will save us a ? : statement, and it is clearer then passing a 1 0
argument.Signed-off-by: Ivo van Doorn
Signed-off-by: John W. Linville -
The TX datalen must always be converted to a value rt73 and rt2500usb
understand. Both require to use a different size then skb->len.
First off this is required because the descriptor must be added,
but the second is because the value must be a multiple of either 2 or 4,
and it should not be a multiple of the USB packetmaxSigned-off-by: Ivo van Doorn
Signed-off-by: John W. Linville -
rt61pci contained 1 line of 88 characters width,
this needs to be cut down.Signed-off-by: Ivo van Doorn
Signed-off-by: John W. Linville -
get_duration expects all speeds to be passed in 100kbs,
this means that passing 2 is incorrect and should be raised to 20Signed-off-by: Ivo van Doorn
Signed-off-by: John W. Linville -
All drivers use the same values for TSF sync,
this will move the value determination into rt2x00config.c,
and the definition for the values to rt2x00reg.hSigned-off-by: Ivo van Doorn
Signed-off-by: John W. Linville -
As reported by Modestas Vainius, enabling rkfill in 1 driver and
disabling it in a second could cause a NULL pointer exception when
the rfkill-disabled driver still sets the CONFIG_SUPPORT_HW_BUTTON flag.Furthermore, rfkill expects the timeout as a value in milliseconds
instead of jiffies. Also increase the timeout to a second,
since this 250ms would be overkill.Also the flag DEVICE_ENABLED_RADIO_HW is causing problems
for devices which do not support the hardware button
while rfkill is enabled in the driver.To remidy this we should inverse the flag and its meaning,
rename the flag to DEVICE_DISABLED_RADIO_HW this means that
by default the radio is enabled by the hardware button (if present)
and can only be disabled explicitely.Signed-off-by: Ivo van Doorn
Signed-off-by: John W. Linville -
We shouldn't use changed_flags when configuring the packet filter,
we work directly with the total_flags which is safe enough since
we already check if something has changed after we applied our
packet filtering flag rules.
Also make sure that when the packet filter is scheduled, the
rt2x00dev->interface.filter is cleared to make sure the drivers
will update the packet filter instead of failing at the check:
*total_flags == rt2x00dev->interface.filterSigned-off-by: Ivo van Doorn
Signed-off-by: John W. Linville -
By putting the flags into a enumeration we can
make it easier maintable since we don't have to
assign numbers for each flag. This makes it easier
to insert and remove flags.Signed-off-by: Ivo van Doorn
Signed-off-by: John W. Linville -
Store the started state into a new flag DEVICE_STARTED_SUSPEND
and set this when suspending while the device was started.
We can't check for is_interface_present() since only mac80211
knows if there are monitor interfaces present.Signed-off-by: Ivo van Doorn
Signed-off-by: John W. Linville -
Don't check if the radio is enabled in rt2x00lib_toggle_rx,
this is required since the link tuner should be disabled
when shutting down the device. The remaining calls inside the
rt2x00lib_toggle_rx handler should deliver no problems when
called while the radio is done.Signed-off-by: Ivo van Doorn
Signed-off-by: John W. Linville -
The various drivers contained duplicate code to handle the
MAC and BSSID initialization correctly. This moves the
address copy to little endian variables to rt2x00config.Signed-off-by: Ivo van Doorn
Signed-off-by: John W. Linville -
Add the device ID for Linksys WUSB54AG.
Signed-off-by: Michael Wu
Signed-off-by: John W. Linville -
This adds the device ID for the HP wireless print kit usb dongle.
Thanks to Thierry Merle for the patch to the original rtl8187 driver.
Signed-off-by: Michael Wu
Signed-off-by: John W. Linville -
ipw2200 makes extensive use of background scanning when unassociated or
down. Unfortunately, the firmware sends scan completed events many
times per second, which the driver pushes directly up to userspace.
This needlessly wakes up processes listening for wireless events many
times per second. Batch together scan completed events for
non-user-requested scans and send them up to userspace every 4 seconds.
Scan completed events resulting from an SIOCSIWSCAN call are pushed up
without delay.Signed-off-by: Dan Williams
Signed-off-by: John W. Linville -
Socket 1:
product info: "The Linksys Group, Inc.", "Wireless Network CF Card", "ISL37300P", "RevA"
manfid: 0x0274, 0x3301Signed-off-by: Marcin Juszkiewicz
Signed-off-by: John W. Linville -
Tested by Su-Jong You
zd1211b chip 0471:1237 v4810 high 00-12-bf AL2230_RF pa0 g--NSigned-off-by: Daniel Drake
Signed-off-by: John W. Linville -
This adds the PCMCIA device ID for the Deutsche Telekom T-Sinus 111card to
hostap_cs.$ /sbin/lspcmcia -v
[...]
Socket 0 Device 0: [hostap_cs] (bus ID: 0.0)
Configuration: state: on
Product Name: T-Sinus 111card 2.0.0
Identification: manf_id: 0x01bf card_id: 0x3301
function: 6 (network)
prod_id(1): "T-Sinus" (0x8c389dc1)
prod_id(2): "111card" (0x6a23ac17)
prod_id(3): --- (---)
prod_id(4): "2.0.0" (0x92b9effb)Signed-off-by: Stefan Lippers-Hollmann
Signed-off-by: John W. Linville -
This "algorithm" is used only internally and is not useful.
Signed-off-by: Johannes Berg
Cc: Michael Buesch
Acked-by: Zhu Yi
Signed-off-by: John W. Linville
Signed-off-by: David S. Miller -
Since I cannot convince the lazy driver authors (hello Michael)
to stop (ab)using the MGMT interface type internally in their
drivers, this patch introduces a new _INVALID type especially
for their use and changes all affected drivers to use it.Signed-off-by: Johannes Berg
Signed-off-by: John W. Linville
Signed-off-by: David S. Miller -
Implement much easier and more lightweight locking for
the periodic work.
This also removes the last big busywait loop and replaces it
by a sleeping loop.Signed-off-by: Michael Buesch
Signed-off-by: John W. Linville
Signed-off-by: David S. Miller -
This removes the direct call to rfkill on an rfkill event
and replaces it with an input device. This way userspace is also
notified about the event.Signed-off-by: Michael Buesch
Signed-off-by: John W. Linville
Signed-off-by: David S. Miller -
This adds full support for the RFKILL button and
the RFKILL LED trigger.Signed-off-by: Michael Buesch
Signed-off-by: John W. Linville
Signed-off-by: David S. Miller -
Drive the LEDs through the generic LED triggers.
Signed-off-by: Michael Buesch
Cc: Larry Finger
Signed-off-by: John W. Linville
Signed-off-by: David S. Miller -
The kernel now provides a generic hexdump implementation should we need
it again, so we can remove it from zd1211rw. After removing that, only
one single-user function is left in zd_util. Move that to zd_mac and
remove zd_util.Signed-off-by: Ulrich Kunitz
Signed-off-by: Daniel Drake
Signed-off-by: John W. Linville
Signed-off-by: David S. Miller -
As detailed at https://bugs.gentoo.org/159646 hostap with hostapd confuses
udev by presenting 2 interfaces with the same MAC address. Also, at the time
of detection, the 'type' attribute is 1, identical to other hostap interfaces.The AP interface is supposed to have type ARPHRD_IEEE80211 (801), but this is
not set until after registration.Setting it before register_netdev() is called allows us to avoid this
confusion. We can do this by propogating the HOSTAP_INTERFACE type through
to hostap_setup_dev().Signed-off-by: Daniel Drake
Signed-off-by: John W. Linville
Signed-off-by: David S. Miller
