[llvm] [BOLT] Optimize basic block loops to avoid n^2 loop (PR #156243)

[Mark Rousskov via llvm-commits]

Mark-Simulacrum wrote:

I'd prefer to leave it to you, not comfortable enough editing this code too significantly (both due to C++ and unfamiliarity with the surroundings). Happy to close this PR or have you merge it (I don't have permission to do so) and then build atop.

https://github.com/llvm/llvm-project/pull/156243