The Kernel Hacker's Guide to Source Code Control

 in
Greg explains how to use patch and diff or BitKeeper for kernel development.

Many issues involved with Linux kernel development are different from traditional software development processes. When working on a portion of the kernel (or a specific driver), you need to 1) stay aware of changes that are happening to other portions of the kernel with which you interact, 2) constantly apply your changes to the moving target of a fast-based kernel development release schedule, 3) resolve any merge conflicts between changes you have made and changes made by other people and 4) be able to export your changes in a format others can use easily.

For a number of years, I developed and maintained the USB to serial port drivers and then eventually took over maintaining all of the USB code in the kernel. In this article, I explain some of the tools I used in the past to do this work and show how some new tools have enhanced my ability to keep on top of changes in the kernel and let me do my job with less effort.

patch and diff

One of the most common methods of doing kernel work is to use the patch and diff programs. You can use this and no other type of source-code control system to do kernel development. One way is to use two different directory trees: a “clean” one and a “working” one. The clean tree is a released kernel version, while the working one is based on the same released kernel version, but contains your modifications. Then you can use patch and diff to extract your changes and forward port these changes to a new kernel release. For example, let's start off with a clean 2.4.18 kernel (available at www.kernel.org/pub/linux/kernel/v2.4/linux-2.4.18.tar.gz) in our working directory:

$ ls
linux-2.4.18.tar.gz

Uncompress this kernel, and then rename the created directory, which will be called “linux” to something that makes sense:

$ tar -zxf linux-2.4.18.tar.gz
$ mv linux linux-2.4.18
$ ls
linux-2.4.18  linux-2.4.18.tar.gz
Now create a duplicate version of this kernel version, and name it something else:
$ tar -zxf linux-2.4.18.tar.gz
$ mv linux linux-2.4.18-greg
$ ls
linux-2.4.18  linux-2.4.18-greg  linux-2.4.18.tar.gz
Now we can do all of our development in our -greg directory and leave the clean, original kernel directory alone. After we are finished with our work, we need to create a patch to send to other people. The Documentation/SubmittingPatches file explains the proper format that most kernel developers like for sending and receiving patches. It also explains the usage of a dontdiff file, which can help with generating these patches. The dontdiff file can be found at www.moses.uklinux.net/patches/dontdiff and contains a list of files that you do not want to have the diff program pay attention to.

To create a patch, use the following command:

$ diff -Naur -X dontdiff \
linux-2.4.18 linux-2.4.18-greg/ > my_patch

This creates a file called my_patch that contains the difference between your work and a clean 2.4.18 kernel tree. This patch then can be sent to other people via e-mail.

New Kernel Versions

If a new kernel version is released, and you want to forward port your changes to the new version, you need to try to apply your generated patch onto a clean kernel version. This can be done in the following steps:

  1. Generate your original patch, as in the previous example.

  2. Using the official patch from kernel.org, move the old kernel version forward one release:

$ cd linux-2.4.18
$ patch -p1 < ../patch-2.4.19
$ cd ..
$ mv linux-2.4.18 linux-2.4.19
  1. Move your working directory forward one release by removing your patch, then applying the new update:

$ cd linux-2.4.18-greg
$ patch -p1 -R < ../my_patch
$ patch -p1 < ../patch-2.4.19
$ cd ..
$ mv linux-2.4.18-greg linux-2.4.19-greg
  1. Try to apply your patch on top of the new update:

$ cd linux-2.4.19-greg
$ patch -p1 < ../my_patch

If your patch does not apply cleanly, resolve all of the conflicts that are created (patch will tell you about these, leaving behind .rej and .orig files for you to compare and fix up manually using your favorite editor). This merge process can be the most difficult part if you have made changes to portions of the source tree that have been changed by other people.

If you use this development process, I highly recommend getting the excellent patchutils set of programs (found at cyberelk.net/tim/patchutils). These programs enable you to manipulate text patches easily in all sorts of useful ways, and they have saved kernel developers many hours of tedious work.

Directory Tip

______________________

Comments

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

Your article helped

Satish Muniyappa's picture

Hey Greg!

We use bitkeeper here. Your article on how to resolve merge conflicts helped a great deal. Just needed to thank you for posting this article.

Great job!

Warm Regards,
Satish

Cool! I can't read the artic

Anonymous's picture

Cool! I can't read the article, but I can leave a comment -- well, I think the article lacks depth, insight, and accuracy. It is very concise though, consisting as it does, simply of a screen saying REGISTER!

Webinar
One Click, Universal Protection: Implementing Centralized Security Policies on Linux Systems

As Linux continues to play an ever increasing role in corporate data centers and institutions, ensuring the integrity and protection of these systems must be a priority. With 60% of the world's websites and an increasing share of organization's mission-critical workloads running on Linux, failing to stop malware and other advanced threats on Linux can increasingly impact an organization's reputation and bottom line.

Learn More

Sponsored by Bit9

Webinar
Linux Backup and Recovery Webinar

Most companies incorporate backup procedures for critical data, which can be restored quickly if a loss occurs. However, fewer companies are prepared for catastrophic system failures, in which they lose all data, the entire operating system, applications, settings, patches and more, reducing their system(s) to “bare metal.” After all, before data can be restored to a system, there must be a system to restore it to.

In this one hour webinar, learn how to enhance your existing backup strategies for better disaster recovery preparedness using Storix System Backup Administrator (SBAdmin), a highly flexible bare-metal recovery solution for UNIX and Linux systems.

Learn More

Sponsored by Storix