Re: [PATCH v2 00/11] nl80211: rx path for control port frames (enabled only for wpa_supplicant)
Markus Theil <[email protected]>
| Newsgroups | gmane.linux.drivers.hostap |
|---|---|
| Message-ID | <[email protected]> |
On 4/19/20 23:28, Jouni Malinen wrote: > On Sat, Apr 11, 2020 at 12:25:16PM +0200, Markus Theil wrote: >> This patch series adds rx control port support over nl80211. Control port >> over nl80211 delivers all EAPOL userspace. >> >> Nl80211 control port for AP mode is included, but disabled, as currently no >> tx status support for control port frames is included in the kernel. >> >> Wpa_supplicant now uses control port rx over nl80211 for both >> EAPOL frames. pre-auth frame rx is disabled. >> wpa_supplicant connects noticeably faster in my tests with this patch series >> applied. > Thanks, applied with changes along the lines pointed out in the other > comments in this thread. This seems to be working nicely for both TX and > RX and for both STA and AP modes. TX/RX capability can now be selected > independently and automatically with the exception of AP mode TX which I > changed to be disabled by default based on the missing TX status > reporting. That said, I did do this in a manner that allows overriding > with control_port_ap=1 driver param so that all combinations can be > tested in hwsim test cases to avoid breaking something accidentally (and > to also allow that to be used for other purposes, if there are cases > where control port TX is more valuable than the retransmission > optimizations from TX status. Anyway, it would be nice to get TX status > support added for the control port TX so that there would be no > remaining missing capabilities compared to the previously used packet > socket approach. > Thanks! I already have control port TX status patches for the Linux kernel floating around on my disk. After I had some time to test them I'll send patches for the missing control port tx status.