Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Re: [STDS-802-11-TGBC] "Address 1 filtering" (in the sense of Figure 5-1)



Hello Mark,


> 	• For the data plane, on a (infrastructure) non-GCR non-GLK STA (and ignoring other weird special cases like PSMP), the device will filter out (will drop/ignore):
> 		• Any directed frame, not to its MAC address in Address 1
> 		• Any group addressed frame that is not to the broadcast address
> 		• Any group addressed frame that doesn’t have the AP’s address (BSSID) in Address 3   -- note that this is specified in 10.2.8, with 

In 10.2.8, there is a description:

"In the case of a non-GLK STA receiver, when the Address 1 field or DA field contains a group address, addressfiltering is performed by comparing the value in the Address 1 field or DA field to all values in the dot11GroupAddressesTable,”

I think the second bullet is:
		• Any group addressed frame that is not in the dot11GroupAddressesTable

Best Regards,

---
Hitoshi MORIOKA
SRC Software Inc.
hmorioka@xxxxxxxxxxxx




> 2021/07/10 6:21、Mark Hamilton <mark.hamilton2152@xxxxxxxxx>のメール:
> 
> All,
>  
> (I’ve CC’d the ARC group on this, so they are aware of these discussions happening in TGbc.  If anyone thinks ARC should have this on the agenda for discussion there, let me know.)
>  
> Given our discussion on today’s meeting telecon about Address 1 filtering (per the functional block in Figure 5-1 and related figures), and in 10.2.8), I thought it would be helpful to clarify what that block is doing, and how it impacts/might need to be modified by EBCS.
>  
> Firstly, I’ll note that this function does not seem to be well defined in 802.11, that I can find.  For example, the only discussion with any detail in 10.2.8 is focused on group addressed frames, and special cases for S1G relay, GLK and mesh.  The stuff in clause 9 (9.3.2.1.2, et al, describing what goes in Address 1) is also not very specific about expectations for Address 1 filtering on RX.  If anyone knows/can point to something more specific, please let me know.
>  
> (By the way, I’m not sure how to parse this phrase in the 802.11bc PAR: “enhanced transmission and reception of broadcast data both in an infrastructure BSS where there is an association between the transmitter and the receiver(s) and in cases where there is no association between transmitter(s) and receiver(s).”  Does that mean when there is no association, but the BSS is an infrastructure BSS, or that the “in an infrastructure BSS” applies to both cases of there is an association and there is no association?  For now, I’m assuming that TGbc is only concerned with infrastructure BSSs.)
>  
> So, my understanding of this block as it appears in the architecture (Figure 5-1, et al), is that it:
> 	• Is primarily defined for the data plane only.  How it works for management traffic is very ill-defined.  
> 		• Clearly, some stuff like Probe Requests/Responses are special cases, and explicit.  
> 		• Other management frames mostly specify that Address 1 will be explicitly the MAC address of the destination, or the broadcast address.  This seems to imply (but not explicit anywhere I can find) that anything else is filtered out by the Address 1 filter.
> 	• For the data plane, on a (infrastructure) non-GCR non-GLK STA (and ignoring other weird special cases like PSMP), the device will filter out (will drop/ignore):
> 		• Any directed frame, not to its MAC address in Address 1
> 		• Any group addressed frame that is not to the broadcast address
> 		• Any group addressed frame that doesn’t have the AP’s address (BSSID) in Address 3   -- note that this is specified in 10.2.8, with implication that it is part of the “Address 1 filtering” even though it is Address 3.
>  
> Given that EBCS UL frames are sent as Public Action frames (EBCS UL), and the Address 1 and Address 3 fields are specified to be broadcast, it seems the only thing that 802.11bc needs to say about Address 1 filtering is that the Address 3 rule above needs to be modified to allow broadcast EBCS UL frames to pass through the filter.
>  
> At least that’s all I could find.  Does anyone think/know that I’ve missed anything?
>  
> Mark
>  
> To unsubscribe from the STDS-802-11-TGBC list, click the following link: https://listserv.ieee.org/cgi-bin/wa?SUBED1=STDS-802-11-TGBC&A=1
> 

________________________________________________________________________
To unsubscribe from the STDS-802-11-TGBC list, click the following link: https://listserv.ieee.org/cgi-bin/wa?SUBED1=STDS-802-11-TGBC&A=1