regression in thread-safety with data-race on mi->inInference field #53433
Description
The changes in #53219 appears to have introduced a datarace on the mi->inInference field, which seems likely to lead to incorrect compilation and bad performance of threaded code. It was previously protected by the codegen lock.
Originally posted by @vtjnash in #53219 (comment)
Posting as an issue since the PR is already merged, and this is a regression in thread-safety