Re: [PATCH] remove compat code for old devfs naming scheme

[Christoph Hellwig]

On Tue, May 14, 2002 at 08:59:41AM -0600, Richard Gooch wrote:
> The reason to support it is because lots of people are depending on
> it. A lot of systems would break for no real gain. This code is in the
> init section, so the memory will be freed before init(8) starts.
> 
> The code should stay. In 2.5, I can move it into the mini devfsd that
> will go into the initial rootfs.

Exactly how many system exist which

  a) still use the two or three year old devfs naming scheme for
     their root= command line
  b) use current 2.5 kernels

and

  c) have users not capable of translating that syntax to either
     the new devfs or traditional Linux syntax?

I don't think there are more then about ten, and no I _absolutely_
do not consider that a reason to bloat the kernel, even if it is just
source and image bloat.