On Friday, 26 July 2019 15:32:50 CEST Johannes Berg wrote:
[...]
> Ok, dunno. But I don't really want to break everything in the hwsim
> tests, even if the kernel patch may be correct ... so I guess I'll wait
> until you resubmit after some fixes go into wpa_s?

Ok, will try to do this. Seems like Jouni merged parts of the patches but 
there are still things pending.

Btw. with the patches, I see following things broken in the test suite (single 
run):

failed tests: ibss_5ghz fst_sta_tear_down_session_bad_fsts_id tnc_peap_soh 
tnc_peap_soh_errors tnc_ttls tnc_ttls_fragmentation tnc_fast 
p2p_set_discoverability ap_ft_vlan_disconnected ap_ft_eap_vlan_multi 
ap_ft_ptk_rekey ap_acs_exclude_dfs discovery_while_go_p2p_dev 
go_neg_forced_freq_diff_than_bss_freq p2p_go_move_reg_change 
p2ps_stale_group_removal2 p2ps_channel_active_go_and_station_different_mcc 
radius_acct_interim_unreachable mesh_secure_ocv_mix_legacy 
mesh_secure_ocv_mix_ht wpas_mesh_open_ht40 wpas_mesh_gate_forwarding olbc_5ghz 
ap_ht40_csa ap_ht40_csa3 ap_ht_op_class_124 wpas_ap_dfs proxyarp_open_ebtables 
dpp_test_vector_p_521

Without the patches, I see following (single run):

failed tests: concurrent_grpform_while_connecting2 
fst_sta_remove_session_established fst_ap_config_oom tnc_peap_soh 
tnc_peap_soh_errors tnc_ttls tnc_ttls_fragmentation tnc_fast ap_ft_ptk_rekey 
ap_acs_exclude_dfs autogo_many discovery_group_client discovery_listen_find 
discovery_while_go p2p_go_move_reg_change p2ps_stale_group_removal2 grpform3_c 
mesh_secure_ocv_mix_legacy mesh_secure_ocv_mix_ht 
wpas_mesh_secure_sae_group_mismatch wpas_mesh_open_ht40 wpas_mesh_open_vht40 
wpas_mesh_open_vht20 mesh_open_vht_160 wpas_mesh_gate_forwarding scan 
scan_setband ap_vht160 ap_vht80_pwr_constraint proxyarp_open_ebtables

So the diff is (- == before HE patch, + == after HE patch):

     ap_acs_exclude_dfs
    +ap_ft_eap_vlan_multi
     ap_ft_ptk_rekey
    -ap_vht160
    -ap_vht80_pwr_constraint
    -autogo_many
    -concurrent_grpform_while_connecting2
    -discovery_group_client
    -discovery_listen_find
    -discovery_while_go
    -fst_ap_config_oom
    -fst_sta_remove_session_established
    -grpform3_c
    -mesh_open_vht_160
    +ap_ft_vlan_disconnected
    +ap_ht40_csa
    +ap_ht40_csa3
    +ap_ht_op_class_124
    +discovery_while_go_p2p_dev
    +dpp_test_vector_p_521
    +fst_sta_tear_down_session_bad_fsts_id
    +go_neg_forced_freq_diff_than_bss_freq
    +ibss_5ghz
     mesh_secure_ocv_mix_ht
     mesh_secure_ocv_mix_legacy
    +olbc_5ghz
     p2p_go_move_reg_change
    +p2ps_channel_active_go_and_station_different_mcc
    +p2p_set_discoverability
     p2ps_stale_group_removal2
     proxyarp_open_ebtables
    -scan
    -scan_setband
    +radius_acct_interim_unreachable
     tnc_fast
     tnc_peap_soh
     tnc_peap_soh_errors
     tnc_ttls
     tnc_ttls_fragmentation
    +wpas_ap_dfs
     wpas_mesh_gate_forwarding
     wpas_mesh_open_ht40
    -wpas_mesh_open_vht20
    -wpas_mesh_open_vht40
    -wpas_mesh_secure_sae_group_mismatch

None of the "+" marked tests are reproducible (they test suite just randomly 
fails as it looks like in the moment).

Hostapd version was 05822609d105 ("HE: MCS size is always a minimum of 4 
bytes") with two pending patches [1,2]. Either John will resend the first 
patch [1] or (when he is busy) I will do it in the next days.

Kind regards,
	Sven

[1] https://patchwork.ozlabs.org/patch/1116968/
[2] https://patchwork.ozlabs.org/patch/1131273/