How trival are we talking? I haven't merged someone else's code, but I'd think that for sufficiently trivial patches, the maintainer just reads it and copy-pastes most of it manually to the relevant place in the code. This seems no bother.
Don't they just use Git merging?
Actually at the time I believe the Linux kernel team were using a proprietary product that was a significant improvement upon CVS & Subversion. But if you weren't a maintainer you could still get changes in via a maintainer who would review your work. You would send them a patch via email.