This applies Richard Sandiford's patch for Linaro GCC as an alternative to disabling the Linaro-specific extension elimination optimization altogether.
Original patch: https://bugs.launchpad.net/gcc-linaro/+bug/728315
Signed-off-by: Mark Mentovai <mark@moxienet.com>
SVN-Revision: 25871
This introduces beacon_int and basic_rate (per wifi-device), and
dtim_period and max_listen_int (per wifi-iface) for mac80211. These
configure the beacon interval, basic rate specification, DTIM period
(one DTIM per this many beacon frames), and maximum listen interval
that a STA will be permitted to associate with. All of the new
settings are optional; if they're absent, the existing hostapd (or, in
the case of basic_rate, driver) defaults will be used.
The existing bintval property only used for type adhoc is moved from
wifi-iface to wifi-device, and is renamed to beacon_interval because
bintval isn't a great name. The beacon interval is property of the
wifi-device; while the choice between wifi-device and wifi-iface may
not be relevant with an adhoc network, there's no reason to configure
the same property one way for type adhoc and another for type ap. This
change isn't expected to cause many problems, as bintval was added
recently, in r25111.
Similarly, the list of basic rates, also added for type adhoc in
r25111, is a property of the device and not the interface. Further, it
ought to be represented in UCI as a list, not a string dependent on
the format that iw uses. I’ve moved it onto the device, renamed it to
basic_rate, and made it configurable for APs via hostapd. Finally, I
adapted it to use the same kb/s representation as mcast_rate; there's
precedent for this format in that it's also how madwifi interprets
mcast_rate.
Neither bintval nor basicrates were ever documented in the UCI
wireless configuration page on the wiki. When this change is
committed, I'll update the documentation as needed.
Signed-off-by: Mark Mentovai <mark@moxienet.com>
SVN-Revision: 25837