1) I've intentionally done this with full depth stirrups on a few projects and have witnessed it being done on several others.
2) I do think that it's prudent to design for this contingency and commend you for it for whatever that's worth..
3) My solution would be the same as yours:
a) Roughen the joint to the right shear friction amplitude.
b) Install dowels to restore the one way shear capacity required of the mat even though I agree with MIStruct that you're probably in pretty good shape even without the dowels. We'll not rely on that though. Well.. I'll
sort or rely on at as I'll describe later in my post.
c) Ensure that the two pours are at least thick enough that you can get your post installed dowels developed on either side of the joint as that is required for shear friction per code. Nobody really seems to know why this is required definitively but, for now, that's the rule of the road.
d) Give some consideration to the fact that the upper pour will be shrinkage restrained relative to the lower pour. So lots of top bar to limit restraint cracking and to act as compression steel in case shrinkage in the upper pour has an adverse impact on your flexural stiffness (gotta close those cracks to engage the concrete). It sounds as though you've already got this covered.
4) With respect to the technical design issues, I discused a very similar topic with ajk1 a while back that you might find informative:
Link. It is, however, one of the many instances in which I probably took things a little too far technically. You know how I do.
5) As I see it, the technical issues for strength capacity are:
a) Got enough one way shear capacity to match the design?
b) Got enough two way shear capacity to match the design?
There's much, much more on the two way part in the thread that I referenced.
I currently elect to to not bother with the two way capacity because:
c) The demands get onerous.
d) I feel that the horizontal shear for two way capacity can be "spread around" similarly to how we handle one way shear on design strips when, clearly, one way shear varies hugely about the width of our strips.
e) Where there is two way punching demand, you will probably have gobs of clamping capacity at the joint simply because that's where your columns will be pounding a bunch of cross thickness compression into the slab. So, really, you probably don't need any dowels right near the column.