Translations by Clytie Siddall

Now, you can configure your kernel. Run <userinput>make xconfig</userinput> if X11 is installed, configured and being run; run <userinput>make menuconfig</userinput> otherwise (you'll need <classname>libncurses5-dev</classname> installed). Take the time to read the online help and choose carefully. When in doubt, it is typically better to include the device driver (the software which manages hardware peripherals, such as Ethernet cards, SCSI controllers, and so on) you are unsure about. Be careful: other options, not related to a specific hardware, should be left at the default value if you do not understand them. Do not forget to select <quote>Kernel module loader</quote> in <quote>Loadable module support</quote> (it is not selected by default). If not included, your Debian installation will experience problems.

Clean the source tree and reset the <classname>kernel-package</classname> parameters. To do that, do <userinput>make-kpkg clean</userinput>.

Now, compile the kernel: <userinput>fakeroot make-kpkg --initrd --revision=custom.1.0 kernel_image</userinput>. The version number of <quote>1.0</quote> can be changed at will; this is just a version number that you will use to track your kernel builds. Likewise, you can put any word you like in place of <quote>custom</quote> (e.g., a host name). Kernel compilation may take quite a while, depending on the power of your machine.

Once the compilation is complete, you can install your custom kernel like any package. As root, do <userinput>dpkg -i ../linux-image-&kernelversion;-<replaceable>subarchitecture</replaceable>_custom.1.0_&architecture;.deb</userinput>. The <replaceable>subarchitecture</replaceable> part is an optional sub-architecture, <phrase arch="i386"> such as <quote>686</quote>, </phrase> depending on what kernel options you set. <userinput>dpkg -i</userinput> will install the kernel, along with some other nice supporting files. For instance, the <filename>System.map</filename> will be properly installed (helpful for debugging kernel problems), and <filename>/boot/config-&kernelversion;</filename> will be installed, containing your current configuration set. Your new kernel package is also clever enough to automatically update your boot loader to use the new kernel. If you have created a modules package, you'll need to install that package as well.

It is time to reboot the system: read carefully any warning that the above step may have produced, then <userinput>shutdown -r now</userinput>.

For more information on Debian kernels and kernel compilation, see the <ulink url="&url-kernel-handbook;">Debian Linux Kernel Handbook</ulink>. For more information on <classname>kernel-package</classname>, read the fine documentation in <filename>/usr/share/doc/kernel-package</filename>.

Recovering a Broken System

Sometimes, things go wrong, and the system you've carefully installed is no longer bootable. Perhaps the boot loader configuration broke while trying out a change, or perhaps a new kernel you installed won't boot, or perhaps cosmic rays hit your disk and flipped a bit in <filename>/sbin/init</filename>. Regardless of the cause, you'll need to have a system to work from while you fix it, and rescue mode can be useful for this.

To access rescue mode, type <userinput>rescue</userinput> at the <prompt>boot:</prompt> prompt, or boot with the <userinput>rescue/enable=true</userinput> boot parameter. You'll be shown the first few screens of the installer, with a note in the corner of the display to indicate that this is rescue mode, not a full installation. Don't worry, your system is not about to be overwritten! Rescue mode simply takes advantage of the hardware detection facilities available in the installer to ensure that your disks, network devices, and so on are available to you while repairing your system.

Instead of the partitioning tool, you should now be presented with a list of the partitions on your system, and asked to select one of them. Normally, you should select the partition containing the root file system that you need to repair. You may select partitions on RAID and LVM devices as well as those created directly on disks.

If possible, the installer will now present you with a shell prompt in the file system you selected, which you can use to perform any necessary repairs. <phrase arch="x86"> For example, if you need to reinstall the GRUB boot loader into the master boot record of the first hard disk, you could enter the command <userinput>grub-install '(hd0)'</userinput> to do so. </phrase>

If the installer cannot run a usable shell in the root file system you selected, perhaps because the file system is corrupt, then it will issue a warning and offer to give you a shell in the installer environment instead. You may not have as many tools available in this environment, but they will often be enough to repair your system anyway. The root file system you selected will be mounted on the <filename>/target</filename> directory.

In either case, after you exit the shell, the system will reboot.

Finally, note that repairing broken systems can be difficult, and this manual does not attempt to go into all the things that might have gone wrong or how to fix them. If you have problems, consult an expert.