[InstCombine] fold shuffles of insert_subvectors

This should be a valid exception to the general rule of not creating new shuffle masks in IR...
because we already do it. :)
Also, DAG combining/legalization will undo this by widening the shuffle back out if needed.

Explanation for how we already do this: SLP or vector source can create chains of insert/extract
as shown in 1 of the examples from PR16739:
https://godbolt.org/z/NlK7rA
https://bugs.llvm.org/show_bug.cgi?id=16739

And we expect instcombine or DAGCombine to clean that up by creating relatively simple shuffles.

Differential Revision: https://reviews.llvm.org/D62024

llvm-svn: 361338
3 files changed
tree: 469ebb0ecc933a2a17c8ff2490128a827b5e3dea
  1. .arcconfig
  2. .clang-format
  3. .clang-tidy
  4. .gitignore
  5. README.md
  6. clang-tools-extra/
  7. clang/
  8. compiler-rt/
  9. debuginfo-tests/
  10. libclc/
  11. libcxx/
  12. libcxxabi/
  13. libunwind/
  14. lld/
  15. lldb/
  16. llgo/
  17. llvm/
  18. openmp/
  19. parallel-libs/
  20. polly/
  21. pstl/
README.md

The LLVM Compiler Infrastructure

This directory and its subdirectories contain source code for LLVM, a toolkit for the construction of highly optimized compilers, optimizers, and runtime environments.