asb wrote: My apologies for missing this before. It is indeed completely valid to have a compressed branch immediately after the lr.w or sc.w and remain compliant with the forward progress guarantee. I feel I'm lacking a bit of higher level understanding of what this code is trying to do it, because: * It matches only one of many possible lr.[wd]/sc.[wd] sequences. For instance we'll have sequences of other integer instructions between the lr and sc for e.g. narrower than 32-bit cmpxchg (see llvm/test/CodeGen/RISCV/atomic-cmpxchg.ll) * Why isn't there a a call to m_emu.WriteMem in the implementation?
https://github.com/llvm/llvm-project/pull/127505 _______________________________________________ lldb-commits mailing list lldb-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
