The most current. not applicable.

Bochs should emulate most of the newer CPUs and their features (albeit unusably slow). This is not a reccomendation of Bochs. This is an acknowledgement that there is emulation. If there were a patent that prohibits software emulation of them Bochs would be wary of that.

I'm sure VirtualBox has some software emulation of some newer extensions too, and qemu of course.

I agree with the others. I guess you ask because of the recent MS vs Intel thing (x86/x64 emulation on ARM) ?
It's much more of an commercial issue I think. The basic instruction set of x86 and x86-64 (-SSE/-AVX,..) is not affected by them, afaik.
Anyway, I'm no lawyer. The topic of softw. patents is complex. I guess not even Intel/MS are sure how to deal with it at the moment (or who would win).

I am not sure how emulating x86 is different from emulating anything else. Simulating hardware using software does not infringe on intel's rights. I don't think intel's lawsuit has any legs, and it probably only a ploy to try to delay MS to market. So long as MS did everything in clean room (that is didn't borrow any of intel's code) then MS is certainty in the clear.

Emulation has been challenged in court several times before, and so far emulation has always won out. Even in hardware, Cyrix was able to provide x86 compatible hardware without infringing on Intel's right. Being a commercial interest doesn't change things. Patents spell out, in very specific terms, what the technology is and who claims ownership of the intellectual property. Creating software and hardware that is compatible with intel's standards does not infringe on their rights. No one can copy their designs or use their technology without permission, but that doesn't' give Intel the right to control markets or inhibit competitors just because it is inconvenient for them.

I am not aware of any patents that would restrict the use of emulation of the kind that we do, since it is all "clean room" by virtue of having no access to microcode.
I think that the bigger problem is the NDAs that stop people like me from describing specific things like obscure instruction behaviour, thus making correct emulation very difficult to achieve.