On 31.03.22 11:30, Lukas Wunner wrote:

I propose the below patch.  If you could provide more details on the
regressions you've reported (+ ideally bugzilla links), I'll be happy
to include them in the commit message.  Thanks!

There is no bugzilla, but the report can be found:
https://lore.kernel.org/netdev/CAG48ez0MHBbENX5gCdHAUXZ7h7s20LnepBF-pa5M=7Bi-jZrEA@mail.gmail.com/ (local)

The commit made binding and unbinding of USB Ethernet asymmetrical:
Before, usbnet_probe() first invoked the ->bind() callback and then
register_netdev().  usbnet_disconnect() mirrored that by first invoking
unregister_netdev() and then ->unbind().

Since the commit, the order in usbnet_disconnect() is reversed and no
longer mirrors usbnet_probe().

One consequence is that a PHY disconnected (and stopped) in ->unbind()
is afterwards stopped once more by unregister_netdev() as it closes the
netdev before unregistering.  That necessitates a contortion in ->stop()
because the PHY may only be stopped if it hasn't already been
disconnected.

Reverting the commit allows making the call to phy_stop() unconditional
in ->stop() and also fixes the issues reported by Oliver.

Very well, but what prevents reintroducing the isssue this revert's
target was to fix?

    Regards
        Oliver