[ath9k-devel] minstrel_ht: problems with HT40

[ath9k-devel] minstrel_ht: problems with HT40

[Oleksij Rempel]

Hallo Felix,

i found your patches for moving ath9k to minstrel_ht and decided to do 
some testing.
For some reason, minstrel_ht exclude all HT40 rates in my network. With 
native ath9k rate controller I'm able to use HT40.
Do you have any idea where to start digging?
-- 
Regards,
Oleksij

[ath9k-devel] minstrel_ht: problems with HT40

[Felix Fietkau]

On 2013-05-09 3:42 PM, Oleksij Rempel wrote:
> Hallo Felix,
> 
> i found your patches for moving ath9k to minstrel_ht and decided to do 
> some testing.
> For some reason, minstrel_ht exclude all HT40 rates in my network. With 
> native ath9k rate controller I'm able to use HT40.
> Do you have any idea where to start digging?
Do the rates not even show up in the debugfs rate control stats? If so,
check the value of sta->bandwidth at the time minstrel_ht_update_caps()
is called.

- Felix

[ath9k-devel] minstrel_ht: problems with HT40

[Oleksij Rempel]

Am 09.05.2013 16:41, schrieb Felix Fietkau:
> On 2013-05-09 3:42 PM, Oleksij Rempel wrote:
>> Hallo Felix,
>>
>> i found your patches for moving ath9k to minstrel_ht and decided to do
>> some testing.
>> For some reason, minstrel_ht exclude all HT40 rates in my network. With
>> native ath9k rate controller I'm able to use HT40.
>> Do you have any idea where to start digging?
> Do the rates not even show up in the debugfs rate control stats?

correct

> If so,
> check the value of sta->bandwidth at the time minstrel_ht_update_caps()
> is called.

hmm... for some reasons sta->bandwidth is always 0.

my rc_stats are in attachment
-- 
Regards,
Oleksij
-------------- next part --------------
type         rate     throughput  ewma prob   this prob  retry   this succ/attempt   success    attempts
CCK/LP        1.0M           0.9       99.9       100.0      0              0(  0)        29          30
CCK/SP        2.0M           0.0        0.0         0.0      0              0(  0)         0           0
CCK/SP        5.5M           0.0        0.0         0.0      0              0(  0)         0           0
CCK/SP       11.0M           0.0        0.0         0.0      0              0(  0)         0           0
HT20/LGI     MCS0            6.2      100.0       100.0      1              0(  0)         1           1
HT20/LGI     MCS1           11.7      100.0       100.0      0              0(  0)         1           1
HT20/LGI     MCS2           16.4      100.0       100.0      0              0(  0)         1           1
HT20/LGI     MCS3           20.7      100.0       100.0      0              0(  0)         1           1
HT20/LGI     MCS4           28.0      100.0       100.0      5              0(  0)         1           1
HT20/LGI     MCS5           27.9       82.9       100.0      5              0(  0)        15          19
HT20/LGI  t  MCS6           28.9       80.2        25.0      5              0(  0)       109         123
HT20/LGI T P MCS7           38.9       99.9       100.0      5              1(  1)       782         813

Total packet count::    ideal 975      lookaround 48
Average A-MPDU length: 1.0

[ath9k-devel] minstrel_ht: problems with HT40

[Felix Fietkau]

On 2013-05-09 5:09 PM, Oleksij Rempel wrote:
> Am 09.05.2013 16:41, schrieb Felix Fietkau:
>> On 2013-05-09 3:42 PM, Oleksij Rempel wrote:
>>> Hallo Felix,
>>>
>>> i found your patches for moving ath9k to minstrel_ht and decided to do
>>> some testing.
>>> For some reason, minstrel_ht exclude all HT40 rates in my network. With
>>> native ath9k rate controller I'm able to use HT40.
>>> Do you have any idea where to start digging?
>> Do the rates not even show up in the debugfs rate control stats?
> 
> correct
> 
>> If so,
>> check the value of sta->bandwidth at the time minstrel_ht_update_caps()
>> is called.
> 
> hmm... for some reasons sta->bandwidth is always 0.
OK, maybe it's a mac80211 issue then. Maybe ath9k rate control checks
this field even between HT capability updates.
Check when/where this is updated in mac80211.

- Felix

[ath9k-devel] minstrel_ht: problems with HT40

[Oleksij Rempel]

Am 09.05.2013 17:16, schrieb Felix Fietkau:
> On 2013-05-09 5:09 PM, Oleksij Rempel wrote:
>> Am 09.05.2013 16:41, schrieb Felix Fietkau:
>>> On 2013-05-09 3:42 PM, Oleksij Rempel wrote:
>>>> Hallo Felix,
>>>>
>>>> i found your patches for moving ath9k to minstrel_ht and decided to do
>>>> some testing.
>>>> For some reason, minstrel_ht exclude all HT40 rates in my network. With
>>>> native ath9k rate controller I'm able to use HT40.
>>>> Do you have any idea where to start digging?
>>> Do the rates not even show up in the debugfs rate control stats?
>>
>> correct
>>
>>> If so,
>>> check the value of sta->bandwidth at the time minstrel_ht_update_caps()
>>> is called.
>>
>> hmm... for some reasons sta->bandwidth is always 0.
> OK, maybe it's a mac80211 issue then. Maybe ath9k rate control checks
> this field even between HT capability updates.
> Check when/where this is updated in mac80211.

I found the reason. menstrel was doing everything right.
For some reasons my speedport router decided to turn off HT40. I didn't 
checked it before and started digging to the code. After i restarted it 
i got HT40 back ... :/
-- 
Regards,
Oleksij