On Sat, 23 Aug 1997 10:06:52 -0700,
  Jason  Thorpe <thorpej@nas.nasa.gov> wrote:

>That's ... lame.  DEBUG shouldn't be any extra verbose unless explicitly
>requested (with e.g. an ep_debug global in the driver itself).

OK, Maybe it *is* lame.


I don't have documentation on all the media-types on cards this driver
nominally supports (it will arrive Monday:), and the documentation I
do have is inconsitent with what the hardware does.  The failure
scenario for 100-T4 (and maybe other media) are that the chip ends up
talking to a nonexistent media device, so the card doesn't work.

In those circumstances, printing the media-change info seemed like the
right engineering decision.  Many other places in the rest of the
kernel and some drivers have DEBUG printfs without control variables,
in similarly suspicious circumstances.

Is this a driver-only issue, or have you changed your mind since the
question of DEBUG printfs in the NFS code came up?

And if this is an ex-cathedra statement that all DEBUG kernel
printf()s( must have control variables, that's fine; but please update
options(9) so the rest of us know about it, and what it applies to.
The NFS code and dp8390.c are equally lame and need fixing, too.