Dear Anonymous, code in itself can not be patented but, unfortunately, the method can. Now, assuming you mean methods, yes, MS and probably everyone else writing any complex systems is infringing patents. It is unavoidable how the patent system is today. It doesn't matter what you, me and a couple of others think, it is the current law.
For me it is almost hilarious seeing some patents describing the ideas I found in operating systems over 30 years ago, not open source but source available. And if people think that applications are more complex today than a long time ago they shouldn't be in this business. Yes, there is more stuff, but most of the basic ideas did exist even before computers. How you can patent them today, I don't know. Maybe just putting the word Internet in patent? Or token? Or network or trust or changing sticks and stones to bits and bytes?
That was OT - Linus is great, not that he is or wants to be any more than a very good system designer and a real programmer, and we should give all the credit for that. He has done and is doing a great job. Now, I also think that he is sometimes a little shortsighted as we all. Virtualization is not the most important thing and it is old but there are cases where it really is very beneficial. One small part is for us, developers, I can and do run XP and Linux versions under OSX in laptop, I run Windows and other Linux versions under VM (in Linux) in a little more powerful systems. Even this is old, many large installations had VM systems for development and testing a long, long time ago, a big way! For a normal end user it may not be so obvious, they don't know nor care if anything is running in virtual, they want it to work.