On Sun, Oct 31, 2004 at 07:35:18AM -0600, David Young wrote:
> On Sun, Oct 31, 2004 at 02:24:54PM +0100, cube@cubidou.net wrote:
> > I guess the auto-padding stuff could be enclosed in a #ifdef that would be
> > activated only if you compile in an affected driver.  That way you could
> > optimize a bit your kernel if you only have padding devices.
> 
> I would like to have the option to compile with "options
> SLIM_SLOW_ETHERNET_PADDING", if it would keep the kernel smaller than
> without.

Well, I guess so.  However, having either schemes in the kernel will
really make the code hard to read.

Quentin Garnier.