Comment 3 for bug 456362

Tony, thanks for looking. I haven't been able to reproduce this but the one time, either. Dan and I were looking over the logs ... might have been a troublesome DHCP server in the mix.

Regardless, there is new code upstream to clear the invalid status on connections when we sleep: http://cgit.freedesktop.org/NetworkManager/NetworkManager/commit/?id=4b2c810b1ba1f4ac6384bde3c749c94467b2886b

Just to make note of it: the code intentionally does not clear the status based on carrier state change. That is on purpose and by intent, even if there is some debate.