Well, yesterday was solved the mysterious (for me) problem when hidusbfp.sys loading through atsiv give no results when visually all looks like equal to success. As i expected the answer is so simple that i never realized that somebody can make such simple mistake - try to load hidusbfp.sys for different processor architecture, i.e. x86 version for x64 processor&OS for example. But this is real.
Unfortunately the error messages from atsiv.exe are identical in both cases (right and wrong).
Problem can be solved. I can rewrite hidusbfp so it will return real success instead of predefined error but in this case it must be unloaded by atsiv explicitly after loading or hidusbfp remains in memory. Of course, this is not the big problem. I can say this is not problem at all - but reread the beginning of the post!
32, x86, IA-32, i386, i686 - this is synonyms in the world of Wintel and defines "32 bit Intel compatible processor"
64(now, as Itanium is dead), x64, x86-64, AMD64, EM64T - ... defines "64 bit AMD compatible processor working on 64 bit OS"
Why did i choose "AMD64" for the last one? Because Microsoft chose exactly that.
P.S. The lesson (once again) for me is simple (and well known for any teacher): something obvious to you is not obvious to someone else.