This issue is out of TE scope as this requires understanding of Kernel log. Adding TE-NeedsTriageHelp for further triaging from dev team.

Thanks!

Paul: Could you review for priority/assignment? Seems your wheelhouse :)

When ETHTOOL_GLINKSETTING fails I need a way to determine if it's due to the device not being an ethernet interface versus ETHTOOL_GLINKSETTING not being supported.  I see EPERM and EOPNOTSUPP on older kernels that don't support ETHTOOL_GLINKSETTING, but I don't have a >=4.6 kernel sitting around to try.

The method used by ethtool git://
git.kernel.org/pub/scm/network/ethtool/ethtool.git is to try the new
ETHTOOL_GLINKSETTING first and fall back to ETHTOOL_GET if that fails.

link_usettings = do_ioctl_glinksettings(ctx);
if (link_usettings == NULL)
link_usettings = do_ioctl_gset(ctx);
if (link_usettings == NULL) {
perror("Cannot get current device settings");
err = -1;

I think ethtool is specifically for operating on Ethernet devices, while Chromium's use of SIOCETHTOOL is for the purpose of determining if a network interface is an Ethernet device.  If Chromium uses ETHTOOL_GLINKSETTINGS and it fails, we don't know if that's due to the network interface not being an Ethernet device, or due to the kernel not supporting ETHTOOL_GLINKSETTINGS.  If we simply try ETHTOOL_GSET whenever ETHTOOL_GLINKSETTINGS fails, we'll continue to see the the kernel warnings on newer kernels when we query on WiFi devices.