Attach routines can fail before calling *hook_establish(), and they
often rely on the detach routine for cleanup. So be consistant and
careful by checking for a NULL hook before calling *hook_disesta

Attach routines can fail before calling *hook_establish(), and they
often rely on the detach routine for cleanup. So be consistant and
careful by checking for a NULL hook before calling *hook_disestablish
in detach routines.

ok mickey@ brad@ dlg@

show more ...

Make ath_attach() return error if the RF is unsupported so the PCI code
can disable its shutdown hooks, otherwise we will panic upon halt.
Okay jsg@ and brad@.

printf() calls with the format string spanning several lines should not have
trailing commas between string chunks, for this has unexpected effects.

allow bpf(4) to ignore packets based on their direction (inbound or
outbound), using a new BIOCSDIRFILT ioctl;
guidance, feedback and ok canacar@

Fix kernel builds without bpfilter. Linking is still broken.

"Please commit this diff ASAP" brad@

Remove redundant calls to bpfdetach.
ok brad@

Initialize flags before using it. Spotted by lint.

ok reyk@

more verbose error message for the "bogus xmit rate" problem

cleanup and simplify the regulation domain handling. some devices are
known to have a broken/unsupported regulation domain stored in their
EEPROM, we use the default instead (Canada). there's no need

cleanup and simplify the regulation domain handling. some devices are
known to have a broken/unsupported regulation domain stored in their
EEPROM, we use the default instead (Canada). there's no need to
manually define COUNTRYCODE in ar5xxx.c anymore for invalid EEPROMs
but it's still possible to force a specific country - your choice.

ok aanriot@

show more ...

remove the very basic ath(4) rate control implementation and use
rssadapt(9) instead. this may need some further work but now it's in
the right place and shared with other drivers (like ral(4)).

yes

remove the very basic ath(4) rate control implementation and use
rssadapt(9) instead. this may need some further work but now it's in
the right place and shared with other drivers (like ral(4)).

yes jsg@

show more ...

don't attach unsupported radio chipsets

based on a diff by jsg@

use ieee80211_iterate_nodes to iterate the nodes and add a extra check
for the maximal number of gpio pins.

fix a possible division by zero by using minimal default values for the
beacon interval and the beacon miss timeout (beacon interval * 7).

Remove the last of the FreeBSD compatiblity goop.
ok reyk@

Remove FreeBSD if_printf compat function from net80211.
ok reyk@

update the 802.11 station address if the ethernet lladdr has been changed.
pointed out by Nick Holmes and Justin Reigle.

remove the IEEE80211_C_SCANALL capability, it should be used on
devices with firmware-based scanning only.

pointed out in a discussion about iwi(4) and ipw(4).

fix led states and knf. this also enables support for the software led
in some thinkpads - turn it on in the ASSOC and RUN states.

remove a useless timeout after tx (decrease nasty delays while
scanning) and fix various other minor things in the ath(4) driver.

- the self linked hardware rx descriptor list does not work in hostap
mode, use the normal behaviour with rx overrun interrupts for now.
this prevents the hostap from locking under heavy load (like p

- the self linked hardware rx descriptor list does not work in hostap
mode, use the normal behaviour with rx overrun interrupts for now.
this prevents the hostap from locking under heavy load (like ping -f)
and it works more reliable.
- run ath_calibrate in splnet() context (from netbsd)

show more ...

remove trailing newline in panic(9); ok millert@ and deraadt@

fix hostap mode

re-enable gpio init even without using the gpio(4) framework

- disable the gpio user interface for now, it's just an unimportant
button which is not yet available on non-i386 platforms
- fix endianess and attachement on big endian platforms (tested on
macppc -

- disable the gpio user interface for now, it's just an unimportant
button which is not yet available on non-i386 platforms
- fix endianess and attachement on big endian platforms (tested on
macppc - ath attaches correctly, receives frames but assoc still
fails)

show more ...

ath_hal_rx_monitor should be used for link state monitoring instead of
setting the promisc mode. rename it to ath_hal_rx_signal and remove
the incorrect promisc flag. this should fix some problems wi

ath_hal_rx_monitor should be used for link state monitoring instead of
setting the promisc mode. rename it to ath_hal_rx_signal and remove
the incorrect promisc flag. this should fix some problems with ath
running in forwarding mode.

show more ...