On Thu, Jun 21, 2018 at 09:13:30PM +0800, jianchao.wang wrote:

Hi Christoph

On 06/21/2018 04:19 PM, Christoph Hellwig wrote:

quoted

On Thu, Jun 21, 2018 at 09:43:26AM +0800, jianchao.wang wrote:

quoted

So we have to preserve the ability of block layer that it could prevent
IO completion path from entering a timeout request.

With scsi-debug module, I tried to simulate a scenario where timeout and IO
completion path could occur concurrently, the system ran into crash easily.

Trace, please.  With the latest kernel.  I'm not saying that there
is nothing to fix, but the mode of never completing once timeout
requests as currently done is SCSI is clearly broken.

Sorry, I don't quite get your point.
Do you mean we should do the modification in the scsi layer ?

I think it would just be helpful to start with the stack trace of your
observed crash + the reproduction steps so the problem (and hopefully
a solution) can be better understood.