[PATCH] Revert "wpa_supplicant: Send EAPOL frames over nl80211 where available"

[Krishna Chaitanya]

On Thu, Jul 27, 2023 at 12:12 AM Fabio Estevam <festevam at gmail.com> wrote:
>
> Hi Ben,
>
> On Wed, Jul 26, 2023 at 2:16 PM Ben Greear <greearb at candelatech.com> wrote:
>
> > At one time, ath10k firmware had some weird logic related to trying to coordinate
> > 4-way handshake with setting keys.  I'm not sure that was ever changed in upstream
> > firmware, so possibly you are hitting that bug/race.
>
> I only have access to the QCA9377 firmware from the upstream
> linux-firmware repo:
>
> ath10k_sdio mmc1:0001:1: qca9377 hw1.1 sdio target 0x05020001 chip_id
> 0x00000000 sub 0000:0000
> ath10k_sdio mmc1:0001:1: kconfig debug 0 debugfs 0 tracing 0 dfs 0 testmode 0
> ath10k_sdio mmc1:0001:1: firmware ver WLAN.TF.1.1.1-00061-QCATFSWPZ-1
> api 5 features ignore-otp crc32 7746e551
> ath10k_sdio mmc1:0001:1: failed to fetch board data for
> bus=sdio,vendor=0271,device=0701,subsystem-vendor=0000,subsystem-device=0000
> from ath10k/QCA9377/hw1.0/board-2.bin
> ath10k_sdio mmc1:0001:1: board_file api 1 bmi_id N/A crc32 544289f7
> ath10k_sdio mmc1:0001:1: htt-ver 3.32 wmi-op 4 htt-op 3 cal otp
> max-sta 32 raw 0 hwcrypto 1
>
> > We use recent supplicant and ath10k and do not notice any problems, but we also use
> > our modified ath10k firmware that changes how the 4-way and key-setting is handled,
> > as well as plenty of patches to ath10k driver that may or may not have some affect
> > on this path.
> >
> > We also run with this supplicant patch, among a pile of others, possibly one of these makes
> > a difference:
> >
> > https://github.com/greearb/hostap-ct/commit/c614b56d0c9073b27f2d1dfdfd014a068ab2c65f
>
mac80211 uses `7` priority for control port already.
> Tried this specific patch, but it did not help. Still get the 'wlan0:
> Unknown event 37' message and authentication fails.
>
I have spent some more time looking into this, but couldn't reproduce
the issue, but I have a
a couple of suggestions:

1. disable HW crypto in ath10k, cryptmode=1 + frame_mode=0
(noencrypt=1 for EAPoL, so,
    if ath10k doesn't handle well, it can cause packet drops)
2. disable WMM_AC (this shouldn't cause any drops at STA side as in
the worst case mac80211 fallbacks
    BK, but AP might reject, so, worth a try).