iwn(4) driver showing poor performance, variable media selection on 10.0-BETA3

[Adrian Chadd]

What's the pci id string? Pciconf -lv
On Dec 2, 2013 11:48 PM, "R. Tyler Croy" <tyler at monkeypox.org> wrote:

> I recently installed FreeBSD 10.0-BETA3 onto my Thinkpad X200, everything
> is
> largely running fine except the wireless performance leaves a lot to be
> desired.
>
> I've got an older Thinkpad T43 running -CURRENT next to the X200 in the
> same
> exact room, same distance from the AP, but with an iwi(4) card in it.
>
> The T43 will download large files between 100-200KB/s, and the "media"
> mode for
> wlan0 is "OFDM/48Mbps" which seems non-crazy.
>
> The X200 however will get between 20-90KB/s on the same files, and the
> "media"
> mode will oscillate *dramatically* during the transfer. I've watched it go
> from
> OFDM/54Mbps, down to 48Mbps, to 16Mbps, to DS/1Mbps which is where it
> seems to
> idle.
>
> I've started looking at the output of wlanstats(1) and nothing jumps out
> at me,
> then again, I'm a noob so what do I know:
>
>     [23:35:28] tyler:~ $ wlanstats
>     8          rx frame too short
>     30         rx from wrong bssid
>     802156     rx discard 'cuz dup
>     4          rx discard 'cuz mcast echo
>     172272     rx beacon frames
>     175177     rx element unknown
>     996        rx frame chan mismatch
>     61         rx seq# violation (TKIP)
>     42         active scans started
>     41         background scans started
>     1072961    tkip crypto done in s/w
>     405918     tkip tx MIC done in s/w
>     667043     tkip rx MIC done in s/w
>     173976     rx management frames
>     667045     total data frames received
>     663976     unicast data frames received
>     3069       multicast data frames received
>     406278     total data frames transmit
>     405920     unicast data frames sent
>     1M         current transmit rate
>     22         current rssi
>     -86        current noise floor (dBm)
>     -64        current signal (dBm)
>
> The only thing that looks removely suspicious is the "rx discard" segment,
> but
> again, I plead noob.
>
>
> Any pointers would be useful and appreciated!
>
> - R. Tyler Croy
> ------------------
>    Code: https://github.com/rtyler
> Chatter: https://twitter.com/agentdero
>          rtyler at jabber.org
>
>