A few days back I sent in a pr related to what I am pretty sure was a bug in
if_tl.c, and I need some input at this point.

- the bug would cause my -current machine to kernel panic if I tried to
switch an interface to 100baseTX if it was currently up with media 10baseT,
and vice versa, and also if I tried to change the media. I tried multiple
different things to test it, and eventually decided to fix it myself.

This is the first time I have ever fixed kernel code, though I am pretty
sure I got it right - I can now change duplex on the fly without downing the
interface, and changing media without downing the interface no longer panics
the kernel, though you do have to 'ifconfig tl<x> up' (in order for tl_init
to setup the media).

What I want to know is, can somebody who actively works on kernel code
review my changes, and tell me what they think? I spent literally a day
doing nothing but debugging, and I'm pretty sure I did everything the right
way. I want a second opinion before I send in a patch and expect to get the
patch incorporated into -current, and I also don't want to introduce broken
code.

Any help would be appreciated.

--Sean

-- 
/~\ The ASCII                       Sean Davis
\ / Ribbon Campaign                 aka dive-o
 X  Against HTML
/ \ Email!                      dive@endersgame.net