On 7/14/2025 6:23 PM, Johannes Berg wrote: > On Mon, 2025-07-14 at 18:09 +0800, Baochen Qiang wrote: >>>> This 'punctured' is suddenly (I mean even 'EHT Operation' was not seen in previous >>>> beacons) >>> >>> Pretty sure there must've been EHT operation before, otherwise we >>> couldn't connect with EHT, and you had multi-link. >> >> Sorry for misleading, I was not clear. I was trying to say 'EHT Operation Parameters' in >> beacon was always 0 AFTER AssocResp, until this sudden 'punctured' frame. >> >> Of course 'EHT operation' was always seen in the beacon frames. > > Right, OK. > > [snip] > > AP information: > >> Tag: HT Information (802.11n D1.10) >> Tag Number: HT Information (802.11n D1.10) (61) >> Tag length: 22 >> Primary Channel: 36 >> HT Information Subset (1 of 3): 0x05 >> .... ..01 = Secondary channel offset: Secondary channel is above the >> primary channel (0x1) >> .... .1.. = Supported channel width: Channel of any width supported > > So HT is 40 MHz, that's of course fine. > >> Tag: VHT Operation >> Tag Number: VHT Operation (192) >> Tag length: 5 >> VHT Operation Info >> Channel Width: 80 MHz, 160 MHz or 80+80 MHz BSS Bandwidth (1) >> Channel Center Segment 0: 42 >> Channel Center Segment 1: 0 > > VHT is 80 MHz (the 160 thing there is misleading) if the primary channel > is 36 and the overall center is 42: > > | 36 | 40 | 44 | 48 | > 42 > >> Ext Tag: HE Operation >> Ext Tag length: 6 (Tag len: 7) >> Ext Tag Number: HE Operation (36) >> HE Operation Parameters: 0x003ff4 >> .... .... .... .... .... .100 = Default PE Duration: 4 >> .... .... .... .... .... 0... = TWT Required: Not required >> .... .... ..11 1111 1111 .... = TXOP Duration RTS Threshold: 1023 >> .... .... .0.. .... .... .... = VHT Operation Information Present: False >> .... .... 0... .... .... .... = Co-Hosted BSS: False >> .... ...0 .... .... .... .... = ER SU Disable: False >> .... ..0. .... .... .... .... = 6 GHz Operation Information Present: False >> 0000 00.. .... .... .... .... = Reserved: 0x00 >> BSS Color Information: 0x36 >> Basic HE-MCS and NSS Set: 0xfffc > > He doesn't say anything but it actually cannot unless it were on 6 GHz, > so that's fine. > > > So yeah, the AP is misbehaving. Now you might argue that as we have EHT > we shouldn't care? Dunno. But basically we don't want to initially > _connect_ to such an AP that's confusing itself with puncturing ... we'd > connect with HE since there it says no puncturing. But while maintaining > the connection because of the code structure it suddenly looks like it > "lost" EHT because the EHT was not well-implemented and we ignore it ... > >>> I don't know. I have a feeling that perhaps the AP is misbehaving and >>> setting HE operation to 80 MHz as well (which is invalid if EHT >> >> So what's the expected bandwidth in HE Operation? 40 MHz or less? > > Whatever narrowing of the bandwidth is needed to not include the > subchannels disabled by puncturing. Could be 40, or could perhaps even > be 20 if the puncturing bitmap was 0x2 instead of 0x8. > > After all, an HE client knows nothing about puncturing, so it will > always use the full bandwidth advertised, if something is punctured in > the HE part of the channel that's the entirely wrong thing for it to do. > > Btw, this is also covered by the spec (I have P802.11be/D7.0 at hand > right now): > > 35.15.1 Basic EHT BSS operation > [...] > If a BSS operating channel width is announced in the EHT Operation > element, then an EHT AP shall announce the BSS operating channel > width(s) to non-EHT non-AP STAs with the following restrictions: > > - [...] > - If the Disabled Subchannel Bitmap subfield in the EHT Operation > element is present, the announced BSS operating channel width(s) to > non-EHT non-AP STAs is the maximum width including the primary channel > without covering any punctured 20 MHz subchannel indicated in the > Disabled Subchannel Bitmap subfield in the EHT Operation element as > defined in 35.15.2 (Preamble puncturing operation). > Thank you Johannes, really great explanation and I totally agree with your opinion! > johannes
