-----BEGIN PGP SIGNED MESSAGE-----
Hash: RIPEMD160

Dan Nicholson wrote:
> As Alexander points out, our udev probably has minimum requirements, 
> too. I recall that some of the rules will only work with 2.6.16 and 
> newer kernels.

The first rule that I can think of that falls into this category is the
WAIT_FOR_SYSFS rule (or lack thereof) -- WAIT_FOR_SYSFS rules get
removed from udev as kernel bugs (in sysfs attribute creation) get
fixed.  There may be some issues in e.g. firmware rules too, and I'm
fairly sure that there are issues with persistent-net.

However, this isn't the same as the glibc minimum-kernel version,
because we don't use any of these udev rules until the LFS-installed
kernel boots.  We don't care if the udev rules don't work with the
host's sysfs setup and kernel, because they'll never be used under that
kernel.  However, glibc does get run under the host's kernel (when, for
instance, various parts of coreutils get run after they get installed in
chapter 6, in chroot).

So --enable-kernel needs to be set to something <= the minimum of the
host kernel and the LFS kernel, and to prevent this log-spam, something
>=2.6.2.  Assuming the book's "host system requirements" section is
still correct, the minimum host kernel is "2.6.x", but I thought that we
required 2.6.1 or 2.6.2 for something (of course I can't remember what).
Maybe we should just change the requirements to say 2.6.2 or newer, with
a note about the --enable-kernel option being the reason why?
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFHO3nWS5vET1Wea5wRAy7zAJ9YafvnUdJRyzBuMe3Ji9pMaK50IgCeIf1r
BdnWXZzEAgXZbun78nudQVQ=
=0Ake
-----END PGP SIGNATURE-----
-- 
http://linuxfromscratch.org/mailman/listinfo/lfs-dev
FAQ: http://www.linuxfromscratch.org/faq/
Unsubscribe: See the above information page

Reply via email to