Hi,

On Tuesday, 13 August 2019 12:55:49 CEST Jouni Malinen wrote:
[...]
> > [1] https://patchwork.ozlabs.org/patch/1131273/
> > [2] https://patchwork.ozlabs.org/patch/1116968/
> > [3] https://lore.kernel.org/r/2152977.f3ncjnVLFo@bentobox
>
> And with the exact same snapshot of hostap.git.
> 
> Without [1] applied, this mac80211_hwsim patch to enable HE meshpoint
> results in the following test cases failing:
> 
> mesh_secure_ocv_mix_legacy
> wpas_mesh_open_ht40
> wpas_mesh_max_peering
> mesh_secure_ocv_mix_ht
> 
> [1] fixes wpas_mesh_max_peering
> 
> This leaves following three test cases failing ("Test exception:
> Couldn't join mesh"):
> 
> mesh_secure_ocv_mix_legacy
> wpas_mesh_open_ht40
> mesh_secure_ocv_mix_ht
> 
> Am I missing a patch somewhere that would be needed to make those pass?

No, these also failed for me without the mac80211_hwsim patch [3] in a full 
test run. And thus not analyzed further by me.

Interestingly, they don't fail (without the hwsim patch) when running them
without the full test suite.

[...]
> So is that missing of vht_enabled=1 from some configurations the problem
> here and if so, would you be able to point me to a hostap.git patch that
> addresses this?

No, there is no hostap.git patch here. There was one in the past in 
patchwork [4] but there were concerns by John. So if you reapply the patch 
then you will see that it doesn't fail anymore.

    diff --git a/src/drivers/driver_nl80211.c b/src/drivers/driver_nl80211.c
    index d71511cb5..b54900bc4 100644
    --- a/src/drivers/driver_nl80211.c
    +++ b/src/drivers/driver_nl80211.c
    @@ -4360,7 +4360,7 @@ static int nl80211_put_freq_params(struct nl_msg *msg,
     	is_24ghz = hw_mode == HOSTAPD_MODE_IEEE80211G ||
     		hw_mode == HOSTAPD_MODE_IEEE80211B;
     
    -	if (freq->vht_enabled || (freq->he_enabled && !is_24ghz)) {
    +	if (freq->vht_enabled) {
     		enum nl80211_chan_width cw;
     
     		wpa_printf(MSG_DEBUG, "  * bandwidth=%d", freq->bandwidth);


So as a compromise, I can can propose a patch which tries to replicate this 
2.4GHz specific behavior from nl80211_put_freq_params also in 
ibss_mesh_setup_freq

Btw. thanks for pointing me towards specific test cases.

Kind regards,
	Sven

[4] https://patchwork.ozlabs.org/patch/1125305/