Re: dhclient fails for iwn0 when no security

[coypu]

On Wed, Jun 29, 2016 at 10:44:55AM -0400, Dan LaBell wrote:
> 
> >
> >This might be a case of a network using 40MHz wide channel. I have
> >personally experienced the same. it's faster for 802.11n, but it drops
> >legacy support for 802.11b/g, which uses 20MHz wide channels.
> >
> >Unfortunately, we do not yet have 802.11n support.
> >
> >I may work on it eventually.
> >
> Are you saying that networking hardware that drops support,
>  now can interfere? (That there's hardware out there.. that
> will interfere) ? Like if  a vendor's wireless router was only
> tested for "homogenous conditions" ? ... Something like
> cookies in tcp/ip?  (that exploit [sic]) could occur?
> 
> So they drop support, so they can sell a 'non-compliant' technology.
> If it uses less chips, they can claim it's improvement on an existing
> patent.
> 

802.11n can use 20MHz or 40MHz and be complaint
I think we struggle to use 40MHz width without 802.11n support. I may be
wrong.
That is to say, there is a mode for 802.11n routers that will make older
hardware not work with it.

wifi standards in general do not seem to value backwards compatibility
as much as other standards do. e.g. 802.11ac is 5GHz only (so definitely
no 802.11b/g support!)

Re: dhclient fails for iwn0 when no security

[Dan LaBell]

This might be a case of a network using 40MHz wide channel. I have
personally experienced the same. it's faster for 802.11n, but it drops
legacy support for 802.11b/g, which uses 20MHz wide channels.

Unfortunately, we do not yet have 802.11n support.

I may work on it eventually.


Are you saying that networking hardware that drops support,
 now can interfere? (That there's hardware out there.. that
will interfere) ? Like if  a vendor's wireless router was only
tested for "homogenous conditions" ? ... Something like
cookies in tcp/ip?  (that exploit [sic]) could occur?

So they drop support, so they can sell a 'non-compliant' technology.
If it uses less chips, they can claim it's improvement on an existing
patent.

Re: dhclient fails for iwn0 when no security

[Andrew Cagney]

On 17 June 2016 at 15:11, John D. Baker  wrote:
>> This might be a case of a network using 40MHz wide channel. I have
>> personally experienced the same. it's faster for 802.11n, but it drops
>> legacy support for 802.11b/g, which uses 20MHz wide channels.
>
> I think this is unlikely since the Mac PowerBook G4 w/MacOS X 10.4.11
> worked fine and its factory-original "airport" card is 802.11b only.
> If I'd had a way to boot NetBSD/macppc on it I could have tested that
> combination as well.

Yes.  They are associating (if that's the correct word) - a tcpdump
shows traffic from everyone else.

I'll try the tip of the 7.x branch, thanks. (at least I wasn't going crazy :-)


> Also, the other devices (802.11a/b/g) were able to associate with the
> access point (status went from "no network" to "active").  If there
> was no support for b/g, I should think that wouldn't be possible.

Re: dhclient fails for iwn0 when no security

[coypu]

> I've stayed at this same motel location every year for the last several
> years and it has always worked until this year.  The desk clerk claimed
> there have been no changes to their system, but indicated that some
> clients using laptops "just can't connect" (and implied that smartphone
> users do not have any problem)."

This might be a case of a network using 40MHz wide channel. I have
personally experienced the same. it's faster for 802.11n, but it drops
legacy support for 802.11b/g, which uses 20MHz wide channels.

Unfortunately, we do not yet have 802.11n support.

I may work on it eventually.

Re: dhclient fails for iwn0 when no security

[Michael van Elst]

jdba...@mylinuxisp.com ("John D. Baker") writes:

>I had a similar experience with the wireless network provided by a motel
>in the U.S.  I tried a number of different systems and interfaces:
>rtw(4) in Lemote YeeLoong, run(4) on Lemote YeeLoong and IBM Thinkcentre,
>ath(4) on IBM ThinkPad A31p.  All were running NetBSD 7.0_STABLE, except
>the Lemote Yeeloong, which was running whatever was -current as of mid-
>April 2016.

The wifi code contains bugs.

-current and netbsd-7 have fixes for two of the bugs so that my
laptop (with iwn) and another system (with urtwn) can use wifi again.

-- 
-- 
Michael van Elst
Internet: mlel...@serpens.de
"A potential Snark may lurk in every tree."

Re: dhclient fails for iwn0 when no security

[Michael van Elst]

andrew.cag...@gmail.com (Andrew Cagney) writes:

>I'm finding that when I'm connecting to a public network (such as the
>one at a recent BSD conference), my DHCP requests seemingly fall on
>deaf ears.

There are bugs in the wifi code that make it fail "on a public network"
(i.e. one where there are other devices).

You can use a kernel from netbsd-7 branch which has fixes
that aren't in NetBSD-7.0.1.

-- 
-- 
Michael van Elst
Internet: mlel...@serpens.de
"A potential Snark may lurk in every tree."

Re: dhclient fails for iwn0 when no security

[John D. Baker]

I had a similar experience with the wireless network provided by a motel
in the U.S.  I tried a number of different systems and interfaces:
rtw(4) in Lemote YeeLoong, run(4) on Lemote YeeLoong and IBM Thinkcentre,
ath(4) on IBM ThinkPad A31p.  All were running NetBSD 7.0_STABLE, except
the Lemote Yeeloong, which was running whatever was -current as of mid-
April 2016.

The motel's wireless was using plain old 128-bit WEP and while I could
associate with an access point, DHCP requests were not answered.  I tried
both default 'dhcpcd' and 'dhclient'.

I dug out my PowerBook G4 667MHz with MacOS X 10.4.11 on it and it was
able to get an address without any problem.

I've stayed at this same motel location every year for the last several
years and it has always worked until this year.  The desk clerk claimed
there have been no changes to their system, but indicated that some
clients using laptops "just can't connect" (and implied that smartphone
users do not have any problem).

-- 
|/"\ John D. Baker, KN5UKS   NetBSD Darwin/MacOS X
|\ / jdbaker[snail]mylinuxisp[flyspeck]comOpenBSDFreeBSD
| X  No HTML/proprietary data in email.   BSD just sits there and works!
|/ \ GPGkeyID:  D703 4A7E 479F 63F8 D3F4  BD99 9572 8F23 E4AD 1645