On 30.10.19 10:03, Johannes Berg wrote: > On Tue, 2019-10-29 at 12:56 +0100, Markus Theil wrote: >> This change allows mesh and ad-hoc interfaces to change beacons and >> tx queue settings. The direct change of these settings should be ok >> for these kind of interfaces and should maybe only forbidden for >> station-like interface types. > "should maybe"? :-) > > Not really, both of the changes are wrong. > > johannes > Mesh interfaces are allowed to perform EDCA according to the standard 802.11-2016. Of course, they should not implement HCF with HCCA as in 10.2.4.1: "The HCF shall be implemented in all QoS STAs except mesh STAs". But the MCF section (10.23.2 MCF contention based channel access) says: "MCF implements the same EDCA (see 10.22.2) as does HCF." QoS IBSS are also allowed in the standard: e.g. 4.7: "A QoS IBSS supports operation under the HCF using TXOPs gained through the EDCA mechanism." Of course, mesh STAs cannot send the EDCA parameter set in their beacons according to Table 9-27: "... is present if dot11QosOptionImplemented is true, and dot11MeshActivated is false, ...". Changing beacons on the fly from user-space in these modes is only useful, if vendor-specific elements are used, which can change over time. All in all I can nevertheless understand your point, that these changes could be "wrong" from a pragmatic point of view. Markus