Current bug: netlink error: failed to retrieve link setting/No such device
by upnort from LinuxQuestions.org on (#5GBQ4)
The NIC is Realtek RTL8111/8168/8411 PCI Express Gigabit Ethernet using the r8169 module.
On a test partition with the latest Current packages, when booting directly to init 1 the network link is not initialized. Running ethtool results in stderr spew as seen in the title. I did not find any clues online or in the logs.
The bug does not manifest when booting to init 3.
Originally this test partition was 14.2. I restored the partition to a clone of 14.2 and repeated the update to Current. Same error on init 1. I have not yet tried a fresh install from scratch on this system. No such bug in 14.2.
While networking usually is not used in runlevel 1, I never have seen this behavior. I discovered the bug because my rc.local[_shutdown] scripts use ethtool to ensure WOL is enabled.
I cannot say when the bug started. I had not updated this specific test partition since about March 11 (5.10.23). While in my testing of Current I commonly boot to init 1, I cannot offer any memory of previous behavior other than I would have noticed this bug.
I have only a VM and a laptop for additional testing and I cannot replicate. Different NICs of course.
I tested the 5.11.12 and 5.10.27 kernels as they are the only packages I still had copies.
I do not have any previous kernel firmware packages. I'll test if anybody provides links. Likewise with previous kernel packages.
Any ideas?
On a test partition with the latest Current packages, when booting directly to init 1 the network link is not initialized. Running ethtool results in stderr spew as seen in the title. I did not find any clues online or in the logs.
The bug does not manifest when booting to init 3.
Originally this test partition was 14.2. I restored the partition to a clone of 14.2 and repeated the update to Current. Same error on init 1. I have not yet tried a fresh install from scratch on this system. No such bug in 14.2.
While networking usually is not used in runlevel 1, I never have seen this behavior. I discovered the bug because my rc.local[_shutdown] scripts use ethtool to ensure WOL is enabled.
I cannot say when the bug started. I had not updated this specific test partition since about March 11 (5.10.23). While in my testing of Current I commonly boot to init 1, I cannot offer any memory of previous behavior other than I would have noticed this bug.
I have only a VM and a laptop for additional testing and I cannot replicate. Different NICs of course.
I tested the 5.11.12 and 5.10.27 kernels as they are the only packages I still had copies.
I do not have any previous kernel firmware packages. I'll test if anybody provides links. Likewise with previous kernel packages.
Any ideas?