Skip to contentSkip to conversation

Why is peer->peer_ids a bitmap?

DISCUSSION Ben Greear 2 participants 4 messages Mar 29, 2016 – Mar 31, 2016 Qualcomm Atheros ATH10K wireless driver

lore
Share & tools
lore

Ben Greear

author
#0 lore raw

Can a single peer object really have more than one ID?

Is this trying to deal with shared peer objects, perhaps?

Thanks,
Ben

Signature · 8 lines
Ben Greear <[email protected]>
Candela Technologies Inc  http://www.candelatech.com


_______________________________________________
ath10k mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/ath10k
#1 lore raw
Quoted 1 line · Ben Greear

Can a single peer object really have more than one ID?

When you install keys you typically get more ids via htt-peer-map event. I think there were some other cases as well..

Quoted 1 line

Is this trying to deal with shared peer objects, perhaps?

This was developed very long ago when peer-id mapping wasn't really well understood. Perhaps we could make do without peer id map now? i.e. only care about the first htt-peer-map per peer address?

Michal

_______________________________________________
ath10k mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/ath10k

author
#2 lore raw
Quoted 7 lines · Michal Kazior
Quoted 1 line · Ben Greear

Can a single peer object really have more than one ID?

When you install keys you typically get more ids via htt-peer-map event. I think there were some other cases as well..

Quoted 5 lines
Quoted 1 line

Is this trying to deal with shared peer objects, perhaps?

This was developed very long ago when peer-id mapping wasn't really well understood. Perhaps we could make do without peer id map now? i.e. only care about the first htt-peer-map per peer address?

The 10.4.3 firmware can probably still do shared peers, though not sure if it actually happens in practice, and I'm not sure it that would cause this bitmap to ever have more than one bit set anyway.

It just struck me as strange, mainly, and since the radio struct has a peer array indexed by the peer-id, then it would seem that we should never have duplicate bits set in any of the peer objects. And, if that is correct, then the bitmap or similar should probably be in the radio struct instead of in peer objects.

Thanks,
Ben

Signature · 8 lines
Ben Greear <[email protected]>
Candela Technologies Inc  http://www.candelatech.com


_______________________________________________
ath10k mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/ath10k
#3 lore raw
Quoted 24 lines · Ben Greear
Quoted 10 lines · Michal Kazior
Quoted 2 lines · Ben Greear

Can a single peer object really have more than one ID?

When you install keys you typically get more ids via htt-peer-map event. I think there were some other cases as well..

Quoted 6 lines
Quoted 1 line

Is this trying to deal with shared peer objects, perhaps?

This was developed very long ago when peer-id mapping wasn't really well understood. Perhaps we could make do without peer id map now? i.e. only care about the first htt-peer-map per peer address?

The 10.4.3 firmware can probably still do shared peers, though not sure if it actually
happens in practice, and I'm not sure it that would cause this bitmap to ever have more than one bit set anyway.

Quoted 6 lines

It just struck me as strange, mainly, and since the radio struct has a peer array indexed by the peer-id, then it would seem that we should never have duplicate bits set in any of the peer objects. And, if that is correct,
then the bitmap or similar should probably be in the radio struct instead of in peer objects.

Ah, good point. The peer_map[] was introduced recently to allow quick peer_id -> peer/station translation for pull-push purposes. So yes - there some duplication going on now. I guess the bitmap could go away.

Michał

_______________________________________________
ath10k mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/ath10k