Jeremy Huntwork wrote:
Alexander E. Patrakov wrote:
The next pre-release LiveCD will use linux-2.6.16-rc4 unconditionally,
unless Linus releases 2.6.16 sooner.
Jumping the gun a bit, aren't we? You haven't waited to hear my reply to
your question concerning this. Please don't issue blanket statements
like this without waiting to hear my thoughts on the matter.
OK, this will happen in a branch.
If the LiveCD fails to work with their hardware, for a known reason (ie,
udev or the kernel) there's a clue right away that the current version
of LFS won't work for them.
Good point. All known LFS breakage should be duplicated on the CD.
You've said that the udev setup in LFS is
broken - so work to fix it.
Matthew partially fixed the udev-rules-related part of it in r7394.
Other missing parts are the udev-retry bootscript, words about
persistent device naming (bug already filed) and words about known
undetectable modules in linux-2.6.15 (fixable also by bumping the kernel
version). I will post these items today.
Bottom line, the LiveCD will follow LFS on this. So if you want the
changes to happen, get them into LFS.
OK, this will be true for trunk.
> This probably
requires returning to shell-based init.
I doubt it *requires* it. In any case, though development on them is
slow, the reason for switching back to a C init remains: it works on all
architectures.
klibc-based shell init has portability problems. Glibc-based one doesn't.
We won't be switching inits again unless we absolutely
have to.
OK, delaying the discussion of the "libata vs drivers/ide" issue until
linux-2.6.17. Time doesn't press yet.
--
Alexander E. Patrakov
--
http://linuxfromscratch.org/mailman/listinfo/livecd
FAQ: http://www.linuxfromscratch.org/faq/
Unsubscribe: See the above information page
