On Sun, 4 May 2003, Matt Wilson wrote: > On Sat, May 03, 2003 at 02:42:08PM -0400, Robert P. J. Day wrote: > > (perhaps related to an earlier issue as to why one needs > > an initrd when using ext3 ...) > > > > earlier today, when demonstrating to a class how to rebuild > > a kernel, i suggested that one could avoid having to deal with > > creating an initrd.img if one built all necessary drivers into > > the kernel. > > I'd much rather see courses that give student information on how to > deal with initrds, instead of working around them by recompiling > kernels. the point of the exercise was, first, to show how to rebuild a kernel, which i'm assuming most people would agree is a useful talent to have. associated with that, of course, is the possibility of having to create an initrd.img, which almost *everything* i've ever seen describes as a way to load some early modules. and as long as the majority of the documentation suggests that that's the primary purpose of an initrd.img, students will be justifiably confused if they build a kernel with every feature they think they need, and yet that kernel refuses to boot. they'll be additionally confused when introduced to how to build an initrd.img and, when they look inside it (which is something i show them how to do), they'll notice that there are actually no modules whatsoever, yet that initrd.img is still essential for the boot process. i'm not saying that the way this done is wrong, merely suggesting that any documentation explaining an initrd should make it clear that it has uses beyond early module loading, that's all. as i learned the hard way, but the whole class took it in good humor and treated it as a learning experience. :-) rday
