RE: Fedora kernel compilation
 

I was a casual slack user, and have compiled the kernel under that distro many times.

I've compiled a new kernel, for my laptop, using the command then installing the rpm after. Works great.

Im curious about the initrd.img file.

Why does fedora use this, rather then the make install_modules command?? Can you use make install_modules rather then making an initrd.img file???

As far as I know you need both if you have used modules for parts of the kernel that are needed prior to the root file system being mounted. The initrd part is so that some of the kernel can be left as modules and the make modules_install step is so that /lib/modules get populated.

You can avoid the initrd stuff if you compile into the kernel anything required prior to the root file system being mounted (file system support, etc.).

So does slack have everything required in the kernel before the rootfs is mounted, thus not needing the initrd file??

When installing a new kernel into fedora, I never needed to use the command make modules_install. Was it created simply by installing the source rpm file??

I think Slack's setup from an initial install has everything it needs - I've just looked in the /boot directory of a box here at home and another at work and there are no initrd.gz files. But my usual install install process is to install my own kernel as soon as the box re-boots after installing lilo for the first time. The kernel packages on the CD might use an initrd, I just haven't installed one.

There is a link in each /boot directory to /usr/doc/mkinitrd-1.0.1/README.initrd so the system is certainly aware of them...