[ath9k-devel] Allow over-riding regdomain in NIC when CFG80211_CERTIFICATION_ONUS is enabled?

[ath9k-devel] Allow over-riding regdomain in NIC when CFG80211_CERTIFICATION_ONUS is enabled?

[Ben Greear]

There are out-of-tree patches for ath9k (at least) that allow overriding the
eeprom reg domain via a module option.

Any chance something like that could make it upstream if it were dependent
on CFG80211_CERTIFICATION_ONUS?

(We recently got a batch of wpea-127n NICs with country-code of TW, which
reminded me of this problem).

Thanks,
Ben
-- 
Ben Greear <greearb@candelatech.com>
Candela Technologies Inc  http://www.candelatech.com

Allow over-riding regdomain in NIC when CFG80211_CERTIFICATION_ONUS is enabled?

[Ben Greear]

There are out-of-tree patches for ath9k (at least) that allow overriding the
eeprom reg domain via a module option.

Any chance something like that could make it upstream if it were dependent
on CFG80211_CERTIFICATION_ONUS?

(We recently got a batch of wpea-127n NICs with country-code of TW, which
reminded me of this problem).

Thanks,
Ben
-- 
Ben Greear <greearb@candelatech.com>
Candela Technologies Inc  http://www.candelatech.com

[ath9k-devel] Allow over-riding regdomain in NIC when CFG80211_CERTIFICATION_ONUS is enabled?

[Adrian Chadd]

Would you please compare the CTL entries programmed into various
flavours of those NICs and tell me what's different?

(And if you can easily compare the rest of the OTP modifications, please do.)

I'd like to see whether there are any regulatory related information
changes to these NICs.



Adrian


On 3 December 2012 14:44, Ben Greear <greearb@candelatech.com> wrote:
> There are out-of-tree patches for ath9k (at least) that allow overriding the
> eeprom reg domain via a module option.
>
> Any chance something like that could make it upstream if it were dependent
> on CFG80211_CERTIFICATION_ONUS?
>
> (We recently got a batch of wpea-127n NICs with country-code of TW, which
> reminded me of this problem).
>
> Thanks,
> Ben
> --
> Ben Greear <greearb@candelatech.com>
> Candela Technologies Inc  http://www.candelatech.com
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-wireless" in
> the body of a message to majordomo at vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: Allow over-riding regdomain in NIC when CFG80211_CERTIFICATION_ONUS is enabled?

[Adrian Chadd]

Would you please compare the CTL entries programmed into various
flavours of those NICs and tell me what's different?

(And if you can easily compare the rest of the OTP modifications, please do.)

I'd like to see whether there are any regulatory related information
changes to these NICs.



Adrian


On 3 December 2012 14:44, Ben Greear <greearb@candelatech.com> wrote:
> There are out-of-tree patches for ath9k (at least) that allow overriding the
> eeprom reg domain via a module option.
>
> Any chance something like that could make it upstream if it were dependent
> on CFG80211_CERTIFICATION_ONUS?
>
> (We recently got a batch of wpea-127n NICs with country-code of TW, which
> reminded me of this problem).
>
> Thanks,
> Ben
> --
> Ben Greear <greearb@candelatech.com>
> Candela Technologies Inc  http://www.candelatech.com
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-wireless" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

[ath9k-devel] Allow over-riding regdomain in NIC when CFG80211_CERTIFICATION_ONUS is enabled?

[Ben Greear]

On 12/03/2012 02:48 PM, Adrian Chadd wrote:
> Would you please compare the CTL entries programmed into various
> flavours of those NICs and tell me what's different?
>
> (And if you can easily compare the rest of the OTP modifications, please do.)
>
> I'd like to see whether there are any regulatory related information
> changes to these NICs.

How do I get this information?

Thanks,
Ben

-- 
Ben Greear <greearb@candelatech.com>
Candela Technologies Inc  http://www.candelatech.com

Re: Allow over-riding regdomain in NIC when CFG80211_CERTIFICATION_ONUS is enabled?

[Ben Greear]

On 12/03/2012 02:48 PM, Adrian Chadd wrote:
> Would you please compare the CTL entries programmed into various
> flavours of those NICs and tell me what's different?
>
> (And if you can easily compare the rest of the OTP modifications, please do.)
>
> I'd like to see whether there are any regulatory related information
> changes to these NICs.

How do I get this information?

Thanks,
Ben

-- 
Ben Greear <greearb@candelatech.com>
Candela Technologies Inc  http://www.candelatech.com

[ath9k-devel] Allow over-riding regdomain in NIC when CFG80211_CERTIFICATION_ONUS is enabled?

[Luis R. Rodriguez]

On Mon, Dec 3, 2012 at 2:44 PM, Ben Greear <greearb@candelatech.com> wrote:
> There are out-of-tree patches for ath9k (at least) that allow overriding the
> eeprom reg domain via a module option.
>
> Any chance something like that could make it upstream if it were dependent
> on CFG80211_CERTIFICATION_ONUS?
>
> (We recently got a batch of wpea-127n NICs with country-code of TW, which
> reminded me of this problem).

I think it would be reasonable, but note Adrian's points. We likely
would need then only to be able to change regulatory domains for only
regulatory domains that match the CTLs programmed in. I'll also note
that the default CTL is actually a high value rather than no
transmit...

  Luis

Re: Allow over-riding regdomain in NIC when CFG80211_CERTIFICATION_ONUS is enabled?

[Luis R. Rodriguez]

On Mon, Dec 3, 2012 at 2:44 PM, Ben Greear <greearb@candelatech.com> wrote:
> There are out-of-tree patches for ath9k (at least) that allow overriding the
> eeprom reg domain via a module option.
>
> Any chance something like that could make it upstream if it were dependent
> on CFG80211_CERTIFICATION_ONUS?
>
> (We recently got a batch of wpea-127n NICs with country-code of TW, which
> reminded me of this problem).

I think it would be reasonable, but note Adrian's points. We likely
would need then only to be able to change regulatory domains for only
regulatory domains that match the CTLs programmed in. I'll also note
that the default CTL is actually a high value rather than no
transmit...

  Luis

[ath9k-devel] Allow over-riding regdomain in NIC when CFG80211_CERTIFICATION_ONUS is enabled?

[Ben Greear]

On 12/03/2012 03:16 PM, Luis R. Rodriguez wrote:
> On Mon, Dec 3, 2012 at 2:44 PM, Ben Greear <greearb@candelatech.com> wrote:
>> There are out-of-tree patches for ath9k (at least) that allow overriding the
>> eeprom reg domain via a module option.
>>
>> Any chance something like that could make it upstream if it were dependent
>> on CFG80211_CERTIFICATION_ONUS?
>>
>> (We recently got a batch of wpea-127n NICs with country-code of TW, which
>> reminded me of this problem).
>
> I think it would be reasonable, but note Adrian's points. We likely
> would need then only to be able to change regulatory domains for only
> regulatory domains that match the CTLs programmed in. I'll also note
> that the default CTL is actually a high value rather than no
> transmit...

How do I get this information?  What is a CTL anyway?

Thanks,
Ben

>
>    Luis
>


-- 
Ben Greear <greearb@candelatech.com>
Candela Technologies Inc  http://www.candelatech.com

Re: Allow over-riding regdomain in NIC when CFG80211_CERTIFICATION_ONUS is enabled?

[Ben Greear]

On 12/03/2012 03:16 PM, Luis R. Rodriguez wrote:
> On Mon, Dec 3, 2012 at 2:44 PM, Ben Greear <greearb@candelatech.com> wrote:
>> There are out-of-tree patches for ath9k (at least) that allow overriding the
>> eeprom reg domain via a module option.
>>
>> Any chance something like that could make it upstream if it were dependent
>> on CFG80211_CERTIFICATION_ONUS?
>>
>> (We recently got a batch of wpea-127n NICs with country-code of TW, which
>> reminded me of this problem).
>
> I think it would be reasonable, but note Adrian's points. We likely
> would need then only to be able to change regulatory domains for only
> regulatory domains that match the CTLs programmed in. I'll also note
> that the default CTL is actually a high value rather than no
> transmit...

How do I get this information?  What is a CTL anyway?

Thanks,
Ben

>
>    Luis
>


-- 
Ben Greear <greearb@candelatech.com>
Candela Technologies Inc  http://www.candelatech.com

[ath9k-devel] Allow over-riding regdomain in NIC when CFG80211_CERTIFICATION_ONUS is enabled?

[Adrian Chadd]

The CTL edge is a regulatory concept where you have 'edges' to a
regulatory domain that imposes further restrictions.

For example, say you have 5500 -> 5600 as a certain set of rules (max
txpower 10), and 5400 - 5500 as a certain set of rules (max tx power
20).

The spectral mask for 802.11 rates has a certain dB dropoff at the
edge. That bleeds to the adjacent channels.

Now, you can configure a channel at 5490 to have a TX power of 20,
_BUT_ the spectral mask means that there'll be some non-zero amount of
RF energy "bled" out over the adjacent channels, which includes the
5500-5600MHz range. Because of this, you need some extra rule that
defines what the TX power should be at that particular boundary.

Now, for this particular scenario you could infer it from the lower TX
power at 5500-5600, but there are still requirements for meeting
maximum power output at band edges where the adjacent band is some
licenced spectrum or something that we don't run 802.11 on. So we
don't have a normal TX power rule for that channel; instead we need to
use a CTL rule that says "at this frequency edge, do X."

For further information, have a read of how the ctl edge stuff works
in the driver and be prepared to consume some nice wine whilst doing
so, in order to prevent clawing at your eyeballs at the hoops  that
you jump through.



Adrian

Re: Allow over-riding regdomain in NIC when CFG80211_CERTIFICATION_ONUS is enabled?

[Adrian Chadd]

The CTL edge is a regulatory concept where you have 'edges' to a
regulatory domain that imposes further restrictions.

For example, say you have 5500 -> 5600 as a certain set of rules (max
txpower 10), and 5400 - 5500 as a certain set of rules (max tx power
20).

The spectral mask for 802.11 rates has a certain dB dropoff at the
edge. That bleeds to the adjacent channels.

Now, you can configure a channel at 5490 to have a TX power of 20,
_BUT_ the spectral mask means that there'll be some non-zero amount of
RF energy "bled" out over the adjacent channels, which includes the
5500-5600MHz range. Because of this, you need some extra rule that
defines what the TX power should be at that particular boundary.

Now, for this particular scenario you could infer it from the lower TX
power at 5500-5600, but there are still requirements for meeting
maximum power output at band edges where the adjacent band is some
licenced spectrum or something that we don't run 802.11 on. So we
don't have a normal TX power rule for that channel; instead we need to
use a CTL rule that says "at this frequency edge, do X."

For further information, have a read of how the ctl edge stuff works
in the driver and be prepared to consume some nice wine whilst doing
so, in order to prevent clawing at your eyeballs at the hoops  that
you jump through.



Adrian