https://bugzilla.novell.com/show_bug.cgi?id=760274
https://bugzilla.novell.com/show_bug.cgi?id=760274#c3
Jeff Mahoney changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |lduncan@novell.com
AssignedTo|kernel-maintainers@forge.pr |jeffm@suse.com
|ovo.novell.com |
Severity|Normal |Major
--- Comment #3 from Jeff Mahoney 2012-05-07 18:35:31 EDT ---
I think the init=0 is going to be the workaround to go with for now.
The rest of the SCSI core has been switched to async probing but switching the
scsi ch to simple async probing won't be enough. The main issue is that the
async probing for disks is really about parallelism. It'll kick off jobs for
each one and the waits will occur in threads while the rest of the system
continues. Then there's a synchronization point where we wait for all the
probes to complete, and then continue booting. It's a necessary step because
otherwise there would be no way to ensure that the devices necessary to mount
the root file system, which could be any device if mount-by-{uuid,label,path}
are used. The synchronization point currently waits on *all* async jobs to
complete before continuing.
So a simple switch wouldn't help anything - it would actually make your boot
wait until the changer initialized. What will probably need to happen is a
switch to using the async call's domain feature so that we only wait on disks
instead of all async jobs.
I can write a quick patch up but I'll need to do some hunting for hardware to
test with.
--
Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.