[llvm] [GlobalISel] Catching inconsistencies in load memory, result, and range metadata type (PR #121247)

[Matt Arsenault via llvm-commits]

================
@@ -0,0 +1,25 @@
+# RUN: not --crash llc -mtriple=amdgcn-mesa-mesa3d -mcpu=tahiti -run-pass=amdgpu-prelegalizer-combiner -verify-machineinstrs %s -o -
----------------
arsenm wrote:

This belongs in test/MachineVerifier, and can use -run-pass=none (and does not need -verify-machineinstrs).

You should also include all the cases in the test I added to the bug (you'll probably need to merge the different MIR instructions into one function to test them in one mir file) 

https://github.com/llvm/llvm-project/pull/121247