Re: USB regression in 10.1

[mlelstv%serpens.de@localhost (Michael van Elst)]

martin%duskware.de@localhost (Martin Husemann) writes:

>This all makes sense to me, but from a users perspective I would expect
>it to all work automatically if I plug the USB connector in.

There are only some USB drives that actually attach in a stopped state,
and it depends on the user if he wants to spin up the drive automatically.

Another class of devices are card readers that appear "offline" when no
card is inserted. Most drives for 'removable media' probably behave that
way.

My only non-removable-media USB device that may start offline is
a M.2 enclosure, and only when USB power is not sufficient.

I, as a user, would expect that drives are started once I use them
and not necessarily when I connect them. Our wedge discovery logic
makes this difficult.



>Just force PQUIRK_START for all usb attached sd* ?

In former times, drives had a jumper to tell if they should
start (immediately, or even with a delay). Would be better
than any forced quirk.

The requirement to start the drive is also independent
from using USB. You can have a RAID controller that
starts with volumes (and the underlying disks) offline.

So if the solution is to force all drives to start, then
you need to do that. Maybe not with that quirk as the
early start command isn't necessarily correct (and no
idea if an iPod requires it at that time).

The alternative is to redesign how (and mostly when) wedge
discovery works.