| Message ID | 1477435489-8555-4-git-send-email-jouni@qca.qualcomm.com (mailing list archive) |
|---|---|
| State | Changes Requested |
| Delegated to: | Johannes Berg |
| Headers | show |
On Wed, 2016-10-26 at 01:44 +0300, Jouni Malinen wrote: > With the previous commits, initial FILS support is now functional in > mac80211-based drivers for both AP and stations roles. That's a bit misleading, I guess AP role is handled entirely in hostapd? You documented the extended feature bit to explicitly mean station role only :) johannes
On Wed, Oct 26, 2016 at 07:50:36AM +0200, Johannes Berg wrote: > On Wed, 2016-10-26 at 01:44 +0300, Jouni Malinen wrote: > > With the previous commits, initial FILS support is now functional in > > mac80211-based drivers for both AP and stations roles. > > That's a bit misleading, I guess AP role is handled entirely in > hostapd? You documented the extended feature bit to explicitly mean > station role only :) Yeah.. In case of mac80211, there was not really changes needed in the kernel side for AP mode. That may be different with non-mac80211 drivers, though, since they might be easier to handle with the AES-SIV operations for associations frames handled within the driver. Would you prefer to split that NL80211_EXT_FEATURE_FILS into two separate values (_STA and _AP) and have mac80211 advertise both? Or just add a _STA only case for now and see what we need to do with NL80211_ATTR_DEVICE_AP_SME cases separately once such a thing is in in functional state?
On Wed, 2016-10-26 at 09:23 +0000, Malinen, Jouni wrote: > On Wed, Oct 26, 2016 at 07:50:36AM +0200, Johannes Berg wrote: > > > > On Wed, 2016-10-26 at 01:44 +0300, Jouni Malinen wrote: > > > > > > With the previous commits, initial FILS support is now functional > > > in > > > mac80211-based drivers for both AP and stations roles. > > > > That's a bit misleading, I guess AP role is handled entirely in > > hostapd? You documented the extended feature bit to explicitly mean > > station role only :) > > Yeah.. In case of mac80211, there was not really changes needed in > the kernel side for AP mode. That may be different with non-mac80211 > drivers, though, since they might be easier to handle with the AES- > SIV operations for associations frames handled within the driver. Right. > Would you prefer to split that NL80211_EXT_FEATURE_FILS into two > separate values (_STA and _AP) I think having it called _STA may be a little clearer, but I think I'm OK with it the way it is (documented as station) as well. > and have mac80211 advertise both? I wouldn't do that, there's nothing that makes it have that capability. > Or just add a _STA only case for now and see what we need to do with > NL80211_ATTR_DEVICE_AP_SME cases separately once such a thing is in > in functional state? Yeah, we should do that. Really all I thought you should do was reword the commit message to make it clear that the flag only implied the station case, and that the AP case needed no changes. Perhaps renaming the flag to ..._STA will make that a bit clearer. johannes
diff --git a/net/mac80211/main.c b/net/mac80211/main.c index 0d9163c..3e9ca80 100644 --- a/net/mac80211/main.c +++ b/net/mac80211/main.c @@ -549,6 +549,7 @@ struct ieee80211_hw *ieee80211_alloc_hw_nm(size_t priv_data_len, NL80211_FEATURE_MAC_ON_CREATE | NL80211_FEATURE_USERSPACE_MPM | NL80211_FEATURE_FULL_AP_CLIENT_STATE; + wiphy_ext_feature_set(wiphy, NL80211_EXT_FEATURE_FILS); if (!ops->hw_scan) wiphy->features |= NL80211_FEATURE_LOW_PRIORITY_SCAN |
With the previous commits, initial FILS support is now functional in mac80211-based drivers for both AP and stations roles. Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com> --- net/mac80211/main.c | 1 + 1 file changed, 1 insertion(+)