Address comments.

Address comment.

Make this purely a cleanup patch.

Remove verification assert when incoming value from unreachable can be any access.
Add additional test which showcases the above.

Address comments.

Rebase on top of D66376.

Rebase on ToT.

Let me try to answer some of the question, and my apologies for not catching this in the initial review.

Ping.

Updated move call to also rename Uses.
Removed asserts checking against LoE.
Added two more reduced tests.

Note: this is incomplete.

lgtm.

Reabse and gentle ping.

Update test.

Address comments.

Address comment.

LGTM, thanks for the patch!

Address comments.
Thank you for the review!

Address comment.

Added needed cast from unsigned to int64_t.

Added optional bool CreationMustSucceed.

Address comments.

Address comment.

Rebase on ToT.
Ready for review.

Update per comment suggestion: reorder conditions.

Thank you for the patch. This LGTM.

Could you add some tests to this?

I'm planning to go ahead with the patch as is, and if the test tickles any bots, we can disable/remove/update it.

Add comment in test.

Add test.
Tested with LLVM_ENABLE_REVERSE_ITERATION enabled and without.

I don't know if a bot exists with LLVM_ENABLE_REVERSE_ITERATION set to true, but if it does, then the check I proposed for the uselist order would likely fail, no?

; CHECK: uselistorder i16 0, { 3, 2, 4, 1, 5, 0, 6 }

Hmm, AFAICT, there is no opt flag equivalent for LLVM_ENABLE_REVERSE_ITERATION. Did I miss it?

You can use SetVector's "getArrayRef" and pass it as "void removeBlocks(ArrayRef<BasicBlock *> DeadBlocks);"

I'd be interested if there's a better way to use a SmallSetVector, without needing the "8". I don't see an Impl version.
The only alternative I came up with was something like:

template <class OrderedSetT>
void removeBlocks(const OrderedSetT &DeadBlocks);