Running throughput tests

[Ben Greear]

On 05/11/2014 09:37 PM, Kalle Valo wrote:
> Ben Greear <greearb at candelatech.com> writes:
>
>> On 04/23/2014 11:44 PM, Kalle Valo wrote:
>>> Bart Jooris <bart.jooris at gmail.com> writes:
>>>
>>>> I also have a stable 500Mbps ++ link now :-) with peaks up to 520Mbps
>>>> (measured with iperf UDP).
>>>> I've  made a second AP and STA with the WLE900VX cards I've exchanged
>>>> with Ben (Thanks again!).
>>>
>>> Good!
>>>
>>>> p.s. The configuration with the sparkLAN WPEA-351AC cards still
>>>> doesn't give the same results. Probably my bad, I will double check it
>>>> again.
>>>
>>> This is CUS223 based design, right? If you see the problem still try to
>>> get debug logs. For example, debug_mask=0x432 might give some hints.
>>>
>>
>> We tried with the Sparklan today, and it locks up the system soon after we
>> start hostapd, so we just removed it from the system.  The WLE900VX in same
>> machine and same config works fine.
>
> Odd. From the website[1] it looks like a normal CUS233 to me. Can you
> get any more info? For example, where does it exactly lock up? Enabling
> also all kernel debugging (lock debugging etc) might be also useful.

Seems like the bad old days when we saw lots of total host lockups on
cold resets, etc.  I didn't see anything useful in the logs, and I was
running lockdep enabled kernel.

Since WLE900VX works for me, I've little interest in hacking on the
Sparklan NIC.  Possibly it's just a bad beta hardware or something like that.

If you have trouble getting a hold of this NIC, I'll be happy to mail
it to you.

Thanks,
Ben

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