| Message ID | 20180326132104.26709-1-emmanuel.grumbach@intel.com (mailing list archive) |
|---|---|
| State | Accepted |
| Delegated to: | Johannes Berg |
| Headers | show |
Hello, On Monday, 26. March 2018 15:21:04 CEST Emmanuel Grumbach wrote: > Apparently, some APs are buggy enough to send a zeroed > WMM IE. Don't WARN on this since this is not caused by a bug > on the client's system. > > This aligns the condition of the WARNING in drv_conf_tx > with the validity check in ieee80211_sta_wmm_params. > We will now pick the default values whenever we get > a zeroed WMM IE. > > This has been reported here: > https://bugzilla.kernel.org/show_bug.cgi?id=199161 > > Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com> > --- > net/mac80211/mlme.c | 3 ++- > 1 file changed, 2 insertions(+), 1 deletion(-) > > diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c > index 39b660b9a908..a6b628964b84 100644 > --- a/net/mac80211/mlme.c > +++ b/net/mac80211/mlme.c > @@ -1785,7 +1785,8 @@ static bool ieee80211_sta_wmm_params(struct ieee80211_local *local, > params[ac].acm = acm; > params[ac].uapsd = uapsd; > > - if (params[ac].cw_min > params[ac].cw_max) { > + if (params->cw_min == 0 || I'm getting: [ 164.503843] wlan0: AP has invalid WMM params (CWmin/max=15/1023 for ACI 0), using defaults The AP is running a recent OpenWrt and is using hostapd 2.7. iw says the AP has the following WMM IE: Extended capabilities: Extended Channel Switching, SSID List, 6 WMM: * Parameter version 1 * u-APSD * BE: CW 15-1023, AIFSN 3 * BK: CW 15-1023, AIFSN 7 * VI: CW 7-15, AIFSN 2, TXOP 3008 usec * VO: CW 3-7, AIFSN 2, TXOP 1504 usec which looks reasonable. What seems to be happening is that the code now expects to start with ac = 0. 802.11-2012's Figure 8-192 EDCA Parameter Set element lists the AC_BE as the first element in the IE and because of this ac variable gets set to IEEE80211_AC_BE [1] in the first round. This would be fine, if IEEE80211_AC_BE was 0, but it is 2 [2]. Should this be params[ac].cw_min? or params[aci].cw_min? Or should the params->cw_min check be placed after the loop? Can you please sent a fix? Thanks. Regards, Christian [1] <https://elixir.bootlin.com/linux/v4.17-rc1/source/net/mac80211/mlme.c#L1768> [2] <https://elixir.bootlin.com/linux/v4.17-rc1/source/include/net/mac80211.h#L148>
On Tue, 2018-04-17 at 14:11 +0200, Christian Lamparter wrote: > > > - if (params[ac].cw_min > params[ac].cw_max) { > > + if (params->cw_min == 0 || Yeah. We already have a fix pending https://patchwork.kernel.org/patch/10320823/ I just haven't been keeping up during the merge window. johannes
diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c index 39b660b9a908..a6b628964b84 100644 --- a/net/mac80211/mlme.c +++ b/net/mac80211/mlme.c @@ -1785,7 +1785,8 @@ static bool ieee80211_sta_wmm_params(struct ieee80211_local *local, params[ac].acm = acm; params[ac].uapsd = uapsd; - if (params[ac].cw_min > params[ac].cw_max) { + if (params->cw_min == 0 || + params[ac].cw_min > params[ac].cw_max) { sdata_info(sdata, "AP has invalid WMM params (CWmin/max=%d/%d for ACI %d), using defaults\n", params[ac].cw_min, params[ac].cw_max, aci);
Apparently, some APs are buggy enough to send a zeroed WMM IE. Don't WARN on this since this is not caused by a bug on the client's system. This aligns the condition of the WARNING in drv_conf_tx with the validity check in ieee80211_sta_wmm_params. We will now pick the default values whenever we get a zeroed WMM IE. This has been reported here: https://bugzilla.kernel.org/show_bug.cgi?id=199161 Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com> --- net/mac80211/mlme.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)