This is true, however, the generally accepted definition of "vendor lock-in" is useless.
The Enterprise doesn't care about vendor lock-in, they care about the consequences of vendor lock-in. That is to say, they care about what vendor lock-in might cost them. The cost of vendor lock-in is what it will cost you to change from your current system and what costs you might be forced to bear if you want to avoid that. Often this can be an issue of a company basically gouging their customers because they can't afford to get out, but it can just as easily be the costs of maintaining a project if the original maintainers disappear, or any number of other things.
Lock-in doesn't occur because software is closed or for any reasons of legality. There is no law which you can use to force someone to continue to use your software. Lock-in occurs because every piece of software is a little bit different and relearning a new one, adapting your other software to work with a new one, changing your business processes to match a new one, all of that costs an awful lot of money.
In a certain sense, Windows actually has less lock-in than Linux. You can nearly always install windows software on multiple versions of Windows, and the each version of Windows is supported for quite a long time. You can, for the most part(and I've certainly seen this) write a piece of software and run it, without any updates for 15 years. Whereas you can't even guarantee that any given piece of software for Linux will install and run the same way on the next version of the Linux software you're running, let along another distribution.
In any event, none of the four freedoms frees you from costs incurred if the people providing your software change the terms under which they do so. Just because you can maintain it yourself doesn't mean that it is in any way practical or even desirable to do so.