[rt2x00-users] rt2800 & minstrel

[Ivo Van Doorn]

On Tue, Aug 10, 2010 at 8:47 AM, Helmut Schaa
<helmut.schaa at googlemail.com> wrote:
> Am Montag 09 August 2010 schrieb Ivo Van Doorn:
>> Hi,
>>
>> >> /sys/kernel/debug/ieee80211/phy*/stations/*/rc_stats
>> >>
>> >> contain for information? And what does it contain when the link
>> >> quality decreases and different rates are being tried?
>> >> During my test with rt2800usb it seems that only rates
>> >> 1Mbs, 18Mbs and 54Mbs seem to work reliable, while the rates
>> >> between 18Mbs and 54Mbs show large failures (1 frame out of 80 succeeds)
>> >>
>> >> Just an impression of the data from my test:
>> >>
>> >> rate     throughput  ewma prob   this prob  this succ/attempt
>> >> success    attempts
>> >>  t   1         0.8       93.7      100.0          0(  0)         19          20
>> >>      2         0.0        0.0        0.0          0(  0)          0           0
>> >>      5.5       0.0        0.0        0.0          0(  0)          0           0
>> >>     11         0.0        0.0        0.0          0(  0)          0           3
>> >>      6         0.0        0.0        0.0          0(  0)          0           3
>> >>      9         0.0        0.0        0.0          0(  0)          0           3
>> >>     12         0.0        0.0        0.0          0(  0)          0           3
>> >>     18         0.0        0.0        0.0          0(  0)        153         236
>> >>     24         0.0        0.0        0.0          0(  0)          1          81
>> >>     36         0.0        0.0        0.0          0(  0)          1          81
>> >>     48         0.0        0.0        0.0          0(  0)          1          81
>> >> T P 54        43.3       99.9      100.0          0(  0)       1082        1162
>> >>
>> >> Total packet count::    ideal 1202      lookaround 133
>> >
>> >
>> > That looks quite strange. I mean 24, 36 and 48 were tried 81 times but only 1
>> > success?
>>
>> Yeah, it is quite weird, especially since 54 and 18 behave reasonably well.
>
> I've just tried what happens when forcing a fixed rate ("iwconfig wlan0 rate 18M"
> for example). Every rate is 100% stable, so I guess there is something wrong
> between rt2x00 and minstrel that causes these statistics to not be 100% correct.
>
> Could you try that as well?

I have to recheck, but I guess this bug is fixed by the MCS fix I
submitted yesterday.

Ivo