Re: Kernel v4.1-rc1 + MQ dm-multipath + MQ SRP oops

[Christoph Hellwig <hch@lst.de>]

On Wed, Apr 29, 2015 at 09:34:33AM -0400, Mike Snitzer wrote:
> I'm seeing this same crash on the completion path (when using your
> tcm_loop script).  But for Bart's case his stacktrace included
> dm_requeue_unmapped_original_request() -- which if called from
> map_request() implies clone->q won't have been initialized given
> __multipath_map()'s code for setting up the old request_fn case.
> 
> Long story short: your fix is right for Bart's crash (but not the ones
> I'm seeing with tcm_loop) -- I'll get it queued up with a proper header
> attributed to you and cc'ing stable as needed.

I don't really have a good explanation for your crash then, but no
block driver will zero out ->q on outstanding requests, so I'd look for
the cause for your issue in device mapper.  That allocation of request
from a mempool for dm-mpath is something very special that I don't
think anyone else does.