The same comment as in D51777, The semantics of the patch itself lgtm.

I see that IndVarsSimplify uses class field Changed. You propose to use Changed as return value. I just want to see the same approach through the whole IndVars...

LGTM with follow-up handling of Changed. If you do not have cycles for handling Changed, please file a bug at least.

is it better?

Could you please add a test producing something with #2 and #3... I guess any loop pass for function with 2-3 loops should work.

Verifier fix landed, can we proceed with this change?

Test is updated to be valid.

ping. Can I do anything to proceed with this change?

Thank you! I've added two unit tests.

Thank you for your comment. Unfortunately till now I failed to create a test showing the benefit from this change in a pass (I tried). I can make one more try.
The problem I'm trying to solve is the following: in downstream pass I have a BitCast instruction which is actually originated from allocation but has a vector type gotten using gep + bitcast from allocation.
I'd like to know the fact that it is an underlying allocation but now utility function just ignore instructions with vector of pointers.

Test, actually can be simplified bit if you do not mind.

LGTM. If no one objects in 1-2 days, feel free to land.

Please take a look.

Before landing.

@sanjoy, are you ok with the last change?

Agreed with Max. Restriction of recursion between isKnownPredicate and simplification of umax is too complex.

Thank you for review guys. I'm ok with your version of commit message.

assert now works.

@sanjoy , please take a look.

Handled comments.

My bad incorrect place to update comment. Will update.
Thanks for noticing.

In D46046#1077768, @javed.absar wrote:

Would removing 'isKnownViaInduction' have any impact of the number of cases 'analyzeable by scev?

LGTM with some comments.