[llvm] Do not use R12 for indirect tail calls with PACBTI (PR #82661)

[Eleanor Bonnici via llvm-commits]

eleanor-arm wrote:

> r14 is being spilled by the push	{r7, lr} instruction

Yes, of course. Thanks for pointing it out. 


I think it is difficult with this implementation to check whether LR is actually spilled on not. The instruction selection happens before the register allocation in the pipeline. Therefore taking a conservative approach and not using R12 for tail calls when `shouldSignReturnAddress=true`, `shouldSignReturnAddressAll=false`



https://github.com/llvm/llvm-project/pull/82661