Re: How to cleanly shut down a block device

[Pierre Ossman <drzeus-list@drzeus.cx>]

Pierre Ossman wrote:
> How about this patch? It is basically the same as the previous, but it
> also sets queuedata to NULL and tests for that. It does not address if
> someone still has dependencies on the queue but hasn't gotten itself a
> reference (as I haven't gotten any word on if that is a problem):

Ok, I've done some more digging through the block layer. A lot of voodoo
in there, but gendisk seems to make sure it has its own reference to the
queue.

As gendisk drops it reference to the queue in del_gendisk(), I have to
assume that is would be invalid for any part of kernel to look at
disk->queue at this point (since it might now have been released).

So I think this patch should cover all bases. It cleanly kills of our
extra thread and also handles any requests that might be left behind.

I hope you can look at this soon as I'm eager to get rid of this oops.

Rgds

-- 
     -- Pierre Ossman

  Linux kernel, MMC maintainer        http://www.kernel.org
  PulseAudio, core developer          http://pulseaudio.org
  rdesktop, core developer          http://www.rdesktop.org