Carl-Daniel Hailfinger wrote:
> 
> I have attached a new version of forcedeth. It reduces code duplication
> to zero and has mostly transparent handling for different versions of
> RX/TX descriptor flags and length. Besides that, the driver should now
> be mostly endian safe.

And version 15 of the gigabit patch was totally screwed, leading to bogus
transmissions (a simple typo caused it). I reread the whole diff
line-by-line, fixed incorrect equivalence transformations and I'm quite
confident that it is correct now.

Known Bug: You will get a "bad: scheduling while atomic" message because
of the msleep(500) in PHY reset.

Any suggestions how I can avoid this message? Using mdelay(500) has its
share of problems too, because it will cause lost time.

Please take a look at version 17 of the gigabit patch
(Warning: it applies only against 2.6.7-bk4 or later).

Regards, Carl-Daniel
-- 
http://www.hailfinger.org/