Vladimir
He did not make the speed of light absolute (or constant) because he did not use light. That is the point. He just deployed a constant, as is necessary to calibrate distance and duration, and chose the speed of light in vacuo, c. He could have chosen any number, it would have made no difference. The presumption for the past 100 years has been that because it is c, then it is observational light. But it is not, because there is no observational light in Einstein. Designating certain entities observers does not make them observers, unless they receive observational light. Otherwise, they are just references.
Which means his second postulate is irrelevant. It is not so much a matter of what he said, literally, but what he did with what he said. So although he did not say it, but did come close to doing so (Einstein para 4 section 9 1916), he effectively asserted that the time differential, which is actually in the receipt of light, to be an innate characteristic of reality. The freak circumstance being that he did not understand timing, so he had a counterbalancing extra layer of time in timing (ie "common time"). This was then followed up with spacetime and then QM with yet more attributions of indefiniteness to reality, rather than to the physical processes whereby we are aware of reality.
So, if this had really been the case, ie that observational light is constant, then as you say, that means the variance has to be in reality (your "space and time"). But as you quite rightly say, this is not the case. And indeed, is not even the case in Einstein, where there is no light. This is a classic example of where if one holds on to what must be true, ie how physical existence occurs, then one finds out what was wrong with a theory that asserts physical existence occurs differently to that. Please do not adjust your television sets we are experiencing problems with the transmission!
Paul