Re: [PATCH v2] blk-mq: Fix race between resetting the timer and completion handling

On Thu, Feb 08, 2018 at 05:37:46PM +0000, Bart Van Assche wrote:

On Thu, 2018-02-08 at 09:19 -0800, tj@kernel.org wrote:

quoted

Hello, Bart.

On Thu, Feb 08, 2018 at 05:10:45PM +0000, Bart Van Assche wrote:

quoted

I think "dereferencing a pointer" means reading the memory location that pointer points
at? Anyway, I think we both interpret the crash report in the same way, namely that it
means that scmd->device == NULL.

No, what I'm trying to say is that it's is the device->host reference
not the scmd->device reference.

Again, I think that means that we agree about the interpretation of the crash
report. The big question here is what the next step should be to analyze this
further and/or to avoid that this crash can occur?

Heh, sorry about not being clear.  What I'm trying to say is that
scmd->device != NULL && device->host == NULL.  Or was this what you
were saying all along?

Thanks.

-- 
tejun