yes, with your patches the network was keeping alive. I've found an upstream bugreport (https://bugzilla.kernel.org/show_bug.cgi?id=202945), which is related to my problem or at least mentioning the performance regression since 4.19 kernel versions. I will try next week with some kernel parameters, which are mentioned in the bugreport ( pcie_aspm=off or pcie_aspm.policy=performance), but i don't know if linux could influence the ASPM because of the log message "r8169 0000:07:00.0: can't disable ASPM; OS doesn't have ASPM control". It doesn't seem that the experimental patch to control ASPM (s. comment 15 under https://bugzilla.kernel.org/show_bug.cgi?id=202945#c15 ) was submitted to the official kernel developemnt. Thank you for your work.