[llvm-branch-commits] [llvm] [DAG] Explicitly truncate splat operand in get_active_lane_mask expansion (PR #230056)

[Luke Lau via llvm-branch-commits]

================

----------------
lukel97 wrote:

Yeah I guess we can force it with that flag, but I meant as in x86 TTI won't enable it.

Agreed though that test overage doesn't hurt. But it's probably best left to a separate PR since I'm sure there's things with the x86 codegen that can be improved, and we likely want to cover all the different type sizes/type widening paths on e.g. AVX 512 etc.

https://github.com/llvm/llvm-project/pull/230056