Here's some very easy and simple steps that you can follow to ensure that you make a Linux kernel subsystem maintainer mad enough to never want to read your patches again:

• Send patches in base64 attachments.
• Send patches that are not offset to the root of the kernel tree.
• Send patches that ignore the well documented and established coding style rules.
• After having the aforementioned coding style rules pointed out to them, continue to send patches which ignore them.
• Send patches that do not follow the well documented format of sending a patch.
• Send patches that depend on other patches to be able to be applied, yet don't specify this or point to where those patches are.
• Send patches against some unspecified kernel tree. Preferably against some arch-specific tree that no one has public access to.
• Send patches that generate new compiler warnings.
• Send patches that do not even build.
• Send patches that oops the first time the code is run.

And yes, all of the above has happened to me recently...

To make this a constructive posting, here's what you can do to make a kernel subsystem maintainer love you and gladly apply your patches:

• Follow the documented coding style rules.
• Follow the documented procedure for submitting a patch.
• Follow the very well documented procedure as suggested by respected kernel maintainers for submitting a patch that will be applied with no problems.

This posting brought to you by the letters S, G, and I.

posted Thu, 31 Mar 2005 in [/linux]