On Mon, 3 Jun 2019, Geoff Winkless wrote: > On Fri, 31 May 2019 at 20:11, Greg KH <greg@xxxxxxxxx> wrote: > > > Can you run 'git bisect' to determine the exact commit that caused this > > problem? That would be most helpful. > > Oh for goodness sake. Sorry, I was being an idiot. After half a day of > building bisect scripts and another 3 hours of waiting for > compile-reboot-test I finally get to > > [adfa79d1c06a32650332930ca4c488ca570b3407] USB: EHCI: make ehci-pci a > separate driver > > I guess that should have been fairly obvious, really. > > modprobe ehci-pci and the bus entry reappears. > > Thanks (and also to Alan) for the help, apologies for wasting your time. > > Geoff (going to hide his blushes in a corner somewhere) Just goes to show we all have our blind spots. I didn't realize what was going on either, and I should have. On the other hand, you mentioned that it was getting some errors even under earlier kernels such as 3.7.10 (before ehci-pci was split off into a separate module, or it would have failed completely). Any idea about that? Alan Stern
