r1414 removed hotplug from the udev_update branch. Part of this seems to
contradict what I said in my "Bugs in udev_update branch" mail to
lfs-dev. Clarification follows.
1) There is no udev_retry bootscript on the CD, because nothing can be
meaningfully retried. A retry is mesningful only if there is a
possibility that the RUN="program" rule failed because of something not
mounted. But everything is always mounted on the CD.
2) Linux-2.6.15 is used, because there is no pre-made reiser4 patch for
2.6.16-rc4. This does mean that the ide-tape module cannot be
autoloaded, but this is not a regression against the regular LFS. Input
subsystem bugs are worked around by compiling input drivers as non-modules.
3) The "cdrom" and "mouse" rules: yes, they are buggy.
Also, I attempted to make sure that kernel configs for all architectures
expose the same features. Namely, all IDE and SCSI drivers are builtin,
all ethernet drivers are built as modules, there is no sound, and all
NLS support is full and modular. This was done by updating the x86
config, then concatenating the new x86 config with old arch-specific
configs and running "make oldconfig". I am not sure that the result is
good for all architectures, especially sparc. Please fix any breakage,
it is not intentional.
Please don't distribute the ISO images without hotplug to the general
public, as this would result in spamming the livecd list with uevent
leak reports (such reports are, however, welcome from LFS developers).
--
Alexander E. Patrakov
--
http://linuxfromscratch.org/mailman/listinfo/livecd
FAQ: http://www.linuxfromscratch.org/faq/
Unsubscribe: See the above information page
