[llvm] [DAGCombiner] Spill dynamic insertelt chain in one go (PR #162368)

[Princeton Ferro via llvm-commits]

Prince781 wrote:

This change improves ISel compile time on a vector-heavy kernel from 1 hour to 200s. I've included a reduced test case in NVPTX, although this is a target-independent change.

https://github.com/llvm/llvm-project/pull/162368