search for: rl330630

...e globaladdress and a separate ADD node for the offset. Which is best depends on the references to the base GlobalAddress within the function. AArch64 recently gained a DAGCombine for folding offsets into addresses where all users are of the form (globaladdr + constant) <https://reviews.llvm.org/rL330630>. We've been looking at the best GlobalAddress lowering strategy for RISC-V here <https://reviews.llvm.org/D45748> (thanks Sameer!), and that discussion has prompted me to reach out here on llvm-dev. For a RISC target I'd suggest that the ideal strategy would be: 1. If the global...

.... ## LLVM commits * GlobalISel now has dedicated extended loads. [r331115](https://reviews.llvm.org/rL331115). * A new AArch64 DAG combine will fold offsets into addresses in the case where all users of the global address are of the form globaladdr + constant. [r330630](https://reviews.llvm.org/rL330630). * The new CFIInserter pass is used to ensure correct dwarf unwind information is emitted in the function epilogue. It is currently x86-only. [r330706](https://reviews.llvm.org/rL330706). * The 'call' pseudoinstruction is now supported by the RISC-V MC layer. [r330826](https://reviews.ll...