llvm-project

dfdc3b72 - [MachinePipelner] Add loop-carried dependencies for global barriers (#174391)

[MachinePipelner] Add loop-carried dependencies for global barriers (#174391) The loads/stores must not be reordered across barrier instructions. However, in MachinePipeliner, it potentially could happen since loop-carried dependencies from loads/stores to a barrier instruction were not considered. The same problem exists for barrier-to-barrier dependencies. This patch adds the handling for those cases. The implementation is based on that of `ScheduleDAGInstrs::buildSchedGraph`. Split off from https://github.com/llvm/llvm-project/pull/135148