[llvm] [AArch64] Wrap integer SCALAR_TO_VECTOR nodes in bitcasts (PR #172837)

[David Green via llvm-commits]

davemgreen wrote:

> I am not sure I follow the 2nd PR request @davemgreen. If I move tests to separate PR this patch will not pass checks. Or are you talking about the strictfp tests I split from [arm64-cvt-simd-fptoi.ll](https://github.com/llvm/llvm-project/pull/172837/changes#diff-749384f9b1e8daa86fc5fe7d33b616ae2b7e299629b5f5c762fb166e2be64d00) and new tests I added to check this feature works correctly ?

Yep - one PR that just updates what we are testing, this PR then just shows differences in the tests.

https://github.com/llvm/llvm-project/pull/172837