[PATCH v2 4/4] ath9k_hw: Fix tx power settings for AR9003

[PATCH v2 4/4] ath9k_hw: Fix tx power settings for AR9003

[Rajkumar Manoharan]

Retriving tx power for 2x2 and 3x3 chainmask is not handled
properly. While calculating tx power for 2x2, 3 dBm was reduced
and for 3x3, 5 dBm was reduced which should be added back when
retriving.

Cc: Paul Stewart <pstew@google.com>
Signed-off-by: Rajkumar Manoharan <rmanohar@qca.qualcomm.com>
---
v2: make use of the existing function

 drivers/net/wireless/ath/ath9k/ar9003_eeprom.c |    2 ++
 1 files changed, 2 insertions(+), 0 deletions(-)

diff --git a/drivers/net/wireless/ath/ath9k/ar9003_eeprom.c b/drivers/net/wireless/ath/ath9k/ar9003_eeprom.c
index ee9c09b..a93bd63 100644
--- a/drivers/net/wireless/ath/ath9k/ar9003_eeprom.c
+++ b/drivers/net/wireless/ath/ath9k/ar9003_eeprom.c
@@ -5059,6 +5059,8 @@ static void ath9k_hw_ar9300_set_txpower(struct ath_hw *ah,
 			regulatory->max_power_level = targetPowerValT2[i];
 	}
 
+	ath9k_hw_update_regulatory_maxpower(ah);
+
 	if (test)
 		return;
 
-- 
1.7.7.3

Re: [PATCH v2 4/4] ath9k_hw: Fix tx power settings for AR9003

[Felix Fietkau]

On 2011-11-10 10:44 AM, Rajkumar Manoharan wrote:
> Retriving tx power for 2x2 and 3x3 chainmask is not handled
> properly. While calculating tx power for 2x2, 3 dBm was reduced
> and for 3x3, 5 dBm was reduced which should be added back when
> retriving.
>
> Cc: Paul Stewart<pstew@google.com>
> Signed-off-by: Rajkumar Manoharan<rmanohar@qca.qualcomm.com>
I think the tx power reduction for 2x2/3x3 needs to be revisited. In the 
discussion that led to Walsh spatial spreading being disabled you 
mentioned this:

> In the FCC rulings, only transmissions that are completely non-coherent, are
> allowed to waive the array gain contribution to EIRP for multi-transmit configurations.
> The use of 2-stream with 2 transmit and use of 3-steam with 3 transmit qualifies
> for this spatial multiplexing MIMO classification as long as the streams are
> directly mapped to each radio (not Walsh spread prior to splitting to multiple radios)
So if I understand this issue correctly, Walsh spatial spreading was 
disabled so that the tx power for 2x2 or 3x3 would not have to be 
reduced by the array gain contribution - yet we're still doing that.

When Adrian pointed out that tx power is reduced based on the number of 
chains, you mentioned this:
> Those pwrdecrease fields are applicable for AR9280 chips not for AR9003.

According to this patch, this doesn't seem to match what the code does, 
though I didn't notice it back then.

So before we merge this patch, let's decide whether we actually need 
this tx power reduction or not.

- Felix

Re: [PATCH v2 4/4] ath9k_hw: Fix tx power settings for AR9003

[Felix Fietkau]

On 2011-11-10 11:04 AM, Felix Fietkau wrote:
> On 2011-11-10 10:44 AM, Rajkumar Manoharan wrote:
>>  Retriving tx power for 2x2 and 3x3 chainmask is not handled
>>  properly. While calculating tx power for 2x2, 3 dBm was reduced
>>  and for 3x3, 5 dBm was reduced which should be added back when
>>  retriving.
>>
>>  Cc: Paul Stewart<pstew@google.com>
>>  Signed-off-by: Rajkumar Manoharan<rmanohar@qca.qualcomm.com>
> I think the tx power reduction for 2x2/3x3 needs to be revisited. In the
> discussion that led to Walsh spatial spreading being disabled you
> mentioned this:
>
>>  In the FCC rulings, only transmissions that are completely non-coherent, are
>>  allowed to waive the array gain contribution to EIRP for multi-transmit configurations.
>>  The use of 2-stream with 2 transmit and use of 3-steam with 3 transmit qualifies
>>  for this spatial multiplexing MIMO classification as long as the streams are
>>  directly mapped to each radio (not Walsh spread prior to splitting to multiple radios)
> So if I understand this issue correctly, Walsh spatial spreading was
> disabled so that the tx power for 2x2 or 3x3 would not have to be
> reduced by the array gain contribution - yet we're still doing that.
>
> When Adrian pointed out that tx power is reduced based on the number of
> chains, you mentioned this:
>>  Those pwrdecrease fields are applicable for AR9280 chips not for AR9003.
>
> According to this patch, this doesn't seem to match what the code does,
> though I didn't notice it back then.
>
> So before we merge this patch, let's decide whether we actually need
> this tx power reduction or not.
I just spent some time reading up on the FCC's MIMO test procedures and 
it appears that I just got mixed up in the terminology earlier. The 
patch should be merged as-is.

- Felix

Fwd: [PATCH v2 4/4] ath9k_hw: Fix tx power settings for AR9003

[Daniel Halperin]

Reiterating that the l-w filter is really dumb. It should selectively
block PATCH mails (and not responses to them).

See below.

---------- Forwarded message ----------
From: Daniel Halperin <dhalperi@cs.washington.edu>
Date: Mon, Dec 5, 2011 at 10:09 AM
Subject: Re: [PATCH v2 4/4] ath9k_hw: Fix tx power settings for AR9003
To: Felix Fietkau <nbd@openwrt.org>
Cc: Rajkumar Manoharan <rmanohar@qca.qualcomm.com>,
linville@tuxdriver.com, linux-wireless@vger.kernel.org, Paul Stewart
<pstew@google.com>, Adrian Chadd <adrian.chadd@gmail.com>


On Thu, Nov 10, 2011 at 8:17 AM, Felix Fietkau <nbd@openwrt.org> wrote:
>>>
>>>  In the FCC rulings, only transmissions that are completely non-coherent, are
>>>  allowed to waive the array gain contribution to EIRP for multi-transmit configurations.
>>>  The use of 2-stream with 2 transmit and use of 3-steam with 3 transmit qualifies
>>>  for this spatial multiplexing MIMO classification as long as the streams are
>>>  directly mapped to each radio (not Walsh spread prior to splitting to multiple radios)


Hi Rajkumar, Felix, etc.,

There are 2 (or more) different power limits in play here. In
particular, the transmit cards have a maximum power output (for
example, 50 mW), and the local regulatory body has a maximum EIRP
(say, 1 W for 2.4 GHz in US, maybe 200 mW, maybe 50 mW).

I understand the comment above about the FCC regulations, but maybe we
should only disable spreading if we're too close (within 3 dB for 2
antennas, 4.77 dB for 3 antennas) to the regulatory limit?  In other
words, we should compute the actual EIRP with and without spreading
and ensure that the EIRP is over the local regulatory limits before we
disable this feature.

This calculation necessarily includes an estimation of the antenna
gain; probably the 5 dBi 2.4 GHz and 2 dBi 5 GHz is good enough, we
could be conservative if necessary. Note that we're already giving
responsibility to the user in the event that she has hacked together
an EIRP-requirement-violating directional antenna ;).

[We should only bother if Walsh spreading helps performance. I think
this is true because it enables each TX chain to transmit slightly
more power and still meet transmit tolerances, but I might be wrong!]

Thanks,
Dan