if_run and wlan_amrr: kernel panics on 8.2-PRERELEASE amd64

[Alexander Zagrebin]

Hi!

On 19.01.2011 01:23:18 -0800, PseudoCylon wrote:

> > mostly works (there is an issue with the devices with the 'power save  mode'
> > enabled),

If device, connected to the freebsd based AP, is in 'power save mode',
then there are:
- very large dispersion of the ping time; like this
    64 bytes from 192.168.1.12: icmp_seq=0 ttl=68 time=228.479 ms
    64 bytes from 192.168.1.12: icmp_seq=1 ttl=68 time=2695.647 ms
    64 bytes from 192.168.1.12: icmp_seq=2 ttl=68 time=1674.081 ms
    64 bytes from 192.168.1.12: icmp_seq=3 ttl=68 time=653.757 ms
    64 bytes from 192.168.1.12: icmp_seq=4 ttl=68 time=3785.051 ms
    64 bytes from 192.168.1.12: icmp_seq=5 ttl=68 time=2764.311 ms
    64 bytes from 192.168.1.12: icmp_seq=6 ttl=68 time=1742.506 ms
    64 bytes from 192.168.1.12: icmp_seq=7 ttl=68 time=720.719 ms
    64 bytes from 192.168.1.12: icmp_seq=10 ttl=68 time=568.465 ms
    64 bytes from 192.168.1.12: icmp_seq=11 ttl=68 time=3.489 ms
    64 bytes from 192.168.1.12: icmp_seq=12 ttl=68 time=8469.515 ms
    64 bytes from 192.168.1.12: icmp_seq=13 ttl=68 time=7447.776 ms
    64 bytes from 192.168.1.12: icmp_seq=14 ttl=68 time=6428.403 ms
    64 bytes from 192.168.1.12: icmp_seq=15 ttl=68 time=5407.636 ms
    64 bytes from 192.168.1.12: icmp_seq=16 ttl=68 time=4385.875 ms
    64 bytes from 192.168.1.12: icmp_seq=17 ttl=68 time=3364.814 ms
    64 bytes from 192.168.1.12: icmp_seq=18 ttl=68 time=2343.062 ms
    64 bytes from 192.168.1.12: icmp_seq=19 ttl=68 time=1322.413 ms
    64 bytes from 192.168.1.12: icmp_seq=20 ttl=68 time=300.670 ms
    64 bytes from 192.168.1.12: icmp_seq=25 ttl=68 time=206.728 ms
- periodical loss of packets, designated to device (like keepalive, icmp etc.),
isn't it?

> Yes, there is.
> Due to lack of documents, I haven't figured out how to add support, how to set 
> tim, etc.

Due to lack of documents, related to the Ralink chipsets?

> > enabled), but transmission rate is limited with 1 Mbit/s.
> 
> Can you try attached patch and see if tx rate goes up.

It seems the commit r217511 have fixed issue.

Have you any thoughts about kernel panic when if_run, runfw and
wlan_amrr are loaded via loader.conf?

-- 
Alexander Zagrebin