What you describe is an open loop system, and our trucks don't run that way most of the time. There are also one or more "oxygen sensors" in the exhaust stream, and those signals are used by the system to adjust the amount of gasoline injected. So that whatever amount of power was produced by burning a tiny amount of oil mist in the intake, the amount of gasoline used during that same time frame will be reduced by however much would have been needed to produce that same power.
It's going to be a one-to-one relationship in terms of power produced, not in terms of volume. Oil actually contains more potential energy per volume, but its preferred combustion conditions are different, so I can't tell offhand how close the volume relationship is going to be. I'd expect it to be in the ballpark at least, if there are any chemical engineers around, feel free to chime in.
As far as pollutants go, I reject considering CO2 a pollutant, and expect the exhaust catalyst will break down a majority of the rest.
Yeah, sure you can filter out particulates and then you have a bunch of polluted filter material. Or maybe you can centrifuge it, and that takes energy. And the liquid breakdown products have to be re-distilled, or possibly catalyzed and re-combined, using more energy and resources. Point being, no matter what, you have to do more work to "recycle" it, and diminishing returns do apply. It's not entirely clear to me at what point this becomes significantly better than just burning it.
Does a special can that catches a few ounces over its lifetime approach that point? I'm skeptical.
This is all sort of a thought experiment anyway. Put some real, provable numbers to any part of it and see what happens...