udev closure

Matthew Burgess matthew at linuxfromscratch.org
Sun May 23 05:01:00 PDT 2004


Folks,

Finally, we get to the point of closure on udev *and* hotplug.

Firstly I'd like to say thank you for all of your comments regarding the
various issues involved.  I know the last few weeks haven't been easy on
anyone's mailbox, but I personally believe that the udev & hotplug
issues cover *much* more than just the packages themselves.  They've
also highlighted areas where the aims and goals of the LFS book may
have been unclear, and allowed us to clarify those.

With that said here's my decision, which will come as no surprise to
those of you who thought I'd already made my mind up on it!

udev & hotplug will both go into the 6.0 book.  While the pros and cons
have been hashed to death already, I'd like to point out that the main
arguments I'm basing this decision on are Ryan's and Alexander's
messages at:
http://www.linuxfromscratch.org/pipermail/lfs-dev/2004-May/046154.html
and
http://www.linuxfromscratch.org/pipermail/lfs-dev/2004-May/046167.html
respectively.

We have always created devices in LFS, so LFS will still create devices,
but by way of udev instead of make_devices.sh.  With the proper amount
of surrounding explanation (which Kevin Fleming has already volunteered
to assist with) I believe that udev will be as educational, if not more
so, than make_devices.sh.

By compiling and installing 'modutils', we've implicitly been sending
out the message that LFS supports modular based kernels.  As Alexander's
posts highlights, we should either be supporting modular based kernels
properly, or not at all.  Again, in the interests of educational value,
I believe that supporting modular kernels is The Right Thing.

Thanks once again,

Matt Burgess.
LFS Coordinator.



More information about the lfs-dev mailing list