On Thursday 16 February 2006 16:08, Larry Finger wrote:
> I agree, but placed the patch where I did because I couldn't find any code in 
> ieee80211 that even 
> sets frequencies.
> 
> I will formulate a patch that puts the domain regulations into ieee80211 and 
> submit it to Linville, 
> then suggest changes to bcm43xx once ieee80211 has been changed.

If you can do this right, that would be great. But face, that it
is not trivial to get right. You have to read the IEEE-802.11
specification.
I think that this is not easy to get right is why it is currently
unimplemented... . ;)

> Incidentally, bcm43xx_is_valid_channel probably should be replaced by 
> ieee80211_is_valid_channel as 
> follows:

As I said, no. :)
The driver does only care about the channel value being technically
correct. The whole regdomain stuff needs to be implemented in the
common ieee80211 code.
The ieee80211(softmac) subsystem must not scan on disallowed channels
in the first place. It currently scans all channels (AFAIK). _That_ is
where the bug is. And that is one place which needs fixing. Fixing
the problem at lower levels (in the drivers) is not the way to go.

I really invite you to fix this and submit the patch to John.

-- 
Greetings Michael.

Attachment: pgp0D0TFlJeB7.pgp
Description: PGP signature

Reply via email to