On 05/23/2018 02:19 PM, Christoph Hellwig wrote:

NVMe always completes the request before returning from ->timeout, either
by polling for it, or by disabling the controller.  Return BLK_EH_DONE so
that the block layer doesn't even try to complete it again.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
  drivers/nvme/host/pci.c    | 14 +++++---------
  drivers/nvme/host/rdma.c , |  2 +-
  drivers/nvme/target/loop.c |  2 +-
  3 files changed, 7 insertions(+), 11 deletions(-)

Is there a way of _testing_ this patch?
It looks pretty dodgy, just replacing BLK_EH_HANDLED with 
BLK_EH_NOT_HANDLED.
And, as nothing else has changed, would imply that we can do it on 
older/stable versions, too.
Which means that the original code was buggy to start with.
Hence a test here would be really beneficial.

But then, assuming you did some tests here:

Reviewed-by: Hannes Reinecke <hare@suse.com>

Cheers,

Hannes