Matthew Burgess wrote:
Alexander E. Patrakov wrote:
The attached gzipped SVG file contains both (open it with Inkscape and
print on A4 paper). Common things = gray, added = green, removed = red.
Thanks very much for this, Alexander, and apologies for not getting back
to you sooner on it. I've a couple of questions that I want to try and
figure the answers out to myself (in order to try and improve my own
understanding of things), but I'll start a new thread regarding this
stuff in a couple of days.
My initial thoughts on the diagram are that it's obvious how much
simpler the whole infrastructure appears to become once hotplug is
removed.
That feeling is false, and based only on my incomplete understanding of
the new setup (thus I can't put all details into a diagram).
What isn't clear (and won't be, I guess, until we remove it in
our test builds) is how robust the new approach will be.
Which new approach? How are we going to replace coldplug (aka
/etc/rrc.d/init.d/hotplug): with udeveventrecorder in mandatory
initramfs (this gives 100% accurate recovering of coldplug events
because they are no longer coldplug events) or with an udevstart
replacement that is not part of the official udev distribution yet?
And the main question: how are we going to avoid libusb breakage in
BLFS? I'd say make a branch of LFS, put a 2.6.14-rc1 kernel, patched
libusb and the template for relevant udev rules into this branch. But
that's probably overkill.
--
Alexander E. Patrakov
--
http://linuxfromscratch.org/mailman/listinfo/lfs-dev
FAQ: http://www.linuxfromscratch.org/faq/
Unsubscribe: See the above information page
