Poor performance when using urtwn on beaglebone black

[Adrian Chadd]

Hm, okay. I'll go try and reproduce this locally.

can you do an 'ifconfig -v wlan0 list scan' and 'ifconfig -v wlan0
list sta' and privately email it to me? That way I can try to see what
environment you have.

the 2x2 NICs don't do rate control at the moment, so I bet part of the
problem there is the default fixed value is terrible :(



-adrian


On 23 June 2016 at 09:23, Otacílio <otacilio.neto at bsd.com.br> wrote:
> I did the tests with two adapters. Bellow RTL8192CU
>
> poor performance
> % dmesg | grep urtwn
> urtwn0: <vendor 0x0bda product 0x8178, class 0/0, rev 2.00/2.00, addr 2> on
> usbus1
> urtwn0: MAC/BB RTL8192CU, RF 6052 2T2R
> urtwn0: enabling 11n
> % netperf -H squitch
> MIGRATED TCP STREAM TEST from 0.0.0.0 () port 0 AF_INET to squitch () port 0
> AF_INET : histogram : interval : dirty data : demo
> Recv   Send    Send
> Socket Socket  Message  Elapsed
> Size   Size    Size     Time     Throughput
> bytes  bytes   bytes    secs.    10^6bits/sec
>
>  65536  32768  32768    10.72       1.06
>
> good performance (after create the interface using -ht)
> % dmesg | grep urtwn
> urtwn0: <vendor 0x0bda product 0x8178, class 0/0, rev 2.00/2.00, addr 2> on
> usbus1
> urtwn0: MAC/BB RTL8192CU, RF 6052 2T2R
> urtwn0: enabling 11n
> % netperf -H squitch
> MIGRATED TCP STREAM TEST from 0.0.0.0 () port 0 AF_INET to squitch () port 0
> AF_INET : histogram : interval : dirty data : demo
> Recv   Send    Send
> Socket Socket  Message  Elapsed
> Size   Size    Size     Time     Throughput
> bytes  bytes   bytes    secs.    10^6bits/sec
>
>  65536  32768  32768    10.16       7.41
>
> =====================================================================================
> poor performance for RTL8188CUS
>
> % dmesg | grep urtwn
> urtwn0: <vendor 0x0bda product 0x8176, class 0/0, rev 2.00/2.00, addr 2> on
> usbus1
> urtwn0: MAC/BB RTL8188CUS, RF 6052 1T1R
> urtwn0: enabling 11n
> % netperf -H squitch
> MIGRATED TCP STREAM TEST from 0.0.0.0 () port 0 AF_INET to squitch () port 0
> AF_INET : histogram : interval : dirty data : demo
> Recv   Send    Send
> Socket Socket  Message  Elapsed
> Size   Size    Size     Time     Throughput
> bytes  bytes   bytes    secs.    10^6bits/sec
>
>  65536  32768  32768    14.03       0.02
>
>
> good performance (after create the interface using -ht)
> urtwn0: <vendor 0x0bda product 0x8176, class 0/0, rev 2.00/2.00, addr 2> on
> usbus1
> urtwn0: MAC/BB RTL8188CUS, RF 6052 1T1R
> urtwn0: enabling 11n
> % netperf -H squitch
> MIGRATED TCP STREAM TEST from 0.0.0.0 () port 0 AF_INET to squitch () port 0
> AF_INET : histogram : interval : dirty data : demo
> Recv   Send    Send
> Socket Socket  Message  Elapsed
> Size   Size    Size     Time     Throughput
> bytes  bytes   bytes    secs.    10^6bits/sec
>
>  65536  32768  32768    10.15       5.81
>
>
> Em 23/06/2016 13:06, Adrian Chadd escreveu:
>>
>> which urtwn is it? can you do a dmesg | grep urtwn?
>>
>> I'd like to see which NIC it is and whether firmware rate control is
>> working.
>>
>> Thanks,
>>
>>
>>
>> -adrian
>>
>>
>> On 22 June 2016 at 10:52, Otacílio <otacilio.neto at bsd.com.br> wrote:
>>>
>>> Dears
>>>
>>> I'm getting poor performance when using wifi by urtwn on beaglebone
>>> black.
>>> I'm using netperf to do benchmarks and I'm getting less than 1Mbps
>>> normally
>>> and between 3Mbps and 7.5Mbps when creating the interface with -ht
>>> option. I
>>> think that previous versions has better performance.
>>>
>>> FreeBSD beaglebone 11.0-ALPHA4 FreeBSD 11.0-ALPHA4 #0 r302064: Wed Jun 22
>>> 00:25:52 BRT 2016
>>>
>>> ota at squitch:/root/workdir/crochet/work/obj/arm.armv6/usr/src/sys/BEAGLEBONE
>>> arm
>>>
>>> []'s
>>>
>>> -Otacílio
>>>
>>>
>>> _______________________________________________
>>> freebsd-arm at freebsd.org mailing list
>>> https://lists.freebsd.org/mailman/listinfo/freebsd-arm
>>> To unsubscribe, send any mail to "freebsd-arm-unsubscribe at freebsd.org"
>
>
>