Patch and revert source code
Published on Jun 21, 2013 by Sachin.

Using patch efficiently.

We often have the situation while testing to applying a patch or to remove changes made by a patch. It may happen that a patch is not correctly applied and result into a dirty source code. This simple bit of information will help you to carefully apply the patch to your source code and also revert the changes back to original state if something goes wrong. I’m using linux kernel version 3.0.42+ as and example and Elan Touch Screen patch file(patch-linux-3.0.42+_elan_ts.patch).

Unpack and change directory to your linux kernel version you want to apply the patch for. I this case my kernel version is 3.0.42+

cd linux-3.0.42+

I have a patch file in the ~/Downloads directory. Patch files typically ends with .patch extensions. This helps in differentiating them as patches.

It is always recommended to do a dry run before actually applying a patch.

patch -p1 --dry-run < ~/Downloads/patch-linux-3.0.42+_elan_ts.patch

In the above command:

1. -p1 stands for verbosity. For more information, please refer comment by Yogesh here.
2. --dry-run will not actually apply a patch, but gives you an output as if the patch is really applied.

If --dry-run applies a patch without any error message, you can go ahead an apply a real patch.

patch -p1 < ~/Downloads/patch-linux-3.0.42+_elan_ts.patch

Now if you want to remove a patch just add the flag `-R`. For example,

patch -R -p1 < ~/Downloads/patch-linux-3.0.42+_elan_ts.patch

Remember that you have to give full path to your patch file when you apply or revert a patch. If you plan to apply or revert a patch, it is recommended to apply/revert one patch at a time so that you can carefully manage your changes.

Hope this was helpful.

Linux kernel has an excellent documentation on this topic applying-patch.