Re: loading firmware while usermodehelper disabled.

On Tue, Jan 3, 2012 at 4:24 AM, Matthew Garrett [off-list ref] wrote:

The IDs will have changed due to the firmware falling out.

The thing is, I don't think it has.

You said this used to work. And I suspect that the ID's haven't
changed at all. The unbind/rebind kind of supports that (it shouldn't
re-bind if it's not the same device, afaik). I suspect the bug report
is from a system that didn't actually have the firmware loaded even
before the suspend (so no USB ID changes, just the stupid
unbind/rebind).

Of course, it's possible that it "used to work" just because we used
to have that "we'll try to load things even if user space isn't
ready", which caused 30-second timeouts for the random cases where it
didn't work, but also silently "just worked" for the other random
cases where it did happen to work. But I don't think unbind/rebind is
correct for an ID change, so that really doesn't make sense as a
explanation either (at least for *this* particular bug report).

                         Linus