speedtch usbatm.c,1.28,1.29

[Duncan Sands]

> Right, so how about making UDSL_ASSERT() a more traditional assertion,
> which is preprocessed completely away when DEBUG is off, and generates
> an err() when DEBUG is on?

The reason for using BUG_ON when DEBUG is defined is that that way you
can't fail to notice it :)  My idea was that the only people who will
have DEBUG defined are developers, and for them a hard hang is not such
a big deal - it will just motivate them to find out what is going on!
In fact the only guys likely to get hit by this are on this list: people
working on the usbatm code.  And I wanted us (i.e. me at that time) to
notice our mistakes as early as possible.

Ciao,

D.