This really needs to be added to a FAQ somewhere on the m0n0 site.

The reason this is happening is that someone connected to the m0n0box 
via wireless has their power save mode setting on their wireless card 
turned on.  This freaks out the HostAP drivers.  I'm getting it all the 
time because I run a WISP.  I hear that FreeBSD 5.x has a fix for this.

It's even worse in Linux.  The hostAP drivers there will crash the 
kernel.

For my system, I have m0n0 logging to my logging server.  Then I have a 
script that check for the timeout keywords.  If 10 or more are found in 
the same minute it will email my cellphone.  I have to log into the 
m0n0wall box and reboot it.

Big pain in the ass really.  I'm currently working on a FreeBSD AP 
because m0n0wall really isn't a AP.

On May 6, 2004, at 2:24 AM, Barry Murphy wrote:

> Anyone know how I can fix this issue, happens every now and then :/
>
> May 6 21:16:44 last message repeated 2 times
> May 6 21:14:44 /kernel: wi1: wi_cmd: busy bit won't clear.
> May 6 21:13:44 /kernel: wi1: wi_cmd: busy bit won't clear.
> May 6 21:12:39 /kernel: wi1: mgmt. buffer allocation failed
> May 6 21:12:39 /kernel: wi1: failed to allocate 1594 bytes on NIC
> May 6 21:12:39 /kernel: wi1: wi_cmd: busy bit won't clear.
> May 6 21:12:39 /kernel: wi1: tx buffer allocation failed
> May 6 21:12:39 /kernel: wi1: failed to allocate 1594 bytes on NIC
> May 6 21:12:39 /kernel: wi1: wi_cmd: busy bit won't clear.
> May 6 21:12:39 /kernel: wi1: wi_cmd: busy bit won't clear.
> May 6 21:12:39 /kernel: wi1: timeout in wi_seek to fc80/0; last status 
> 800b
> May 6 21:12:39 /kernel: wi1: timeout in wi_seek to fc24/0; last status 
> 800b
> May 6 21:12:39 /kernel: wi1: timeout in wi_seek to fc23/0; last status 
> 800b