(In reply to Kriton Kyrimis from comment #14) > The problem reappeared with the first of these three rpms: > > r8169-kmp-3559d81-kmp-default-2.3_k4.18.5_1-1.1.x86_64.rpm Ok! So the culprit is the commit: 3559d81e76bf r8169: simplify rtl_hw_start_8169 (v4.18-rc1) I guess the patch did not write registers in proper order. > Per your instructions, I didn't test the other two rpms. Right! That's because the follow-up rpms both contain the culprit commit. I'll be out of office tomorrow, I'd continue to work on it when I back to office. Many thanks for your testing and feedback!