http://bugzilla.novell.com/show_bug.cgi?id=480801 User jeffm@novell.com added comment http://bugzilla.novell.com/show_bug.cgi?id=480801#c4 Jeff Mahoney <jeffm@novell.com> changed: What |Removed |Added ---------------------------------------------------------------------------- AssignedTo|kernel-maintainers@forge.pr |jeffm@novell.com |ovo.novell.com | --- Comment #4 from Jeff Mahoney <jeffm@novell.com> 2009-09-03 12:40:30 MDT --- Sadly, it's not the same problem. AoE runs into the backing_dev_info problem because it's open coding gendisk creation to avoid sleeping. MMC uses the generic block layer code. It's properly initialized, but is freed too soon. I think we need to hold onto the queue reference a little longer. What's happening is that when the card gets removed, it shuts everything down associated with the card regardless of if it has any users. So if there are writes that haven't made it to the block layer yet, like writeback that hasn't happened yet, then they run into a block device that has completely gone away. I can reproduce this so I'm testing a patch now. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.