[PATCH 6/9] nvme: abort requests on the reqeueue list when shutting down a controller

Christoph Hellwig hch at lst.de
Thu Oct 22 07:58:42 PDT 2015


On Thu, Oct 22, 2015 at 02:44:19PM +0000, Busch, Keith wrote:
> With this change, requests will be ended in failure even though they
> may have succeeded immediately after a reset. Nvme's unfreeze kicks the
> namespaces' requeue_list just for that reason.
> 
> We shouldn't want to abort the requeue list unless we're freeing the
> request_queue too, right?

We're aborting all active commands through nvme_dev_shutdown ->
nvme_clear_queue.  Why would we skip commands that were active and are
going to be active again ASAP?  But this part actually was just something
I suspected to be an issue first, the ->timeout return value change
turned out to fix the actual bug I was seeing.  So feel free to skip it
for now but I suspect we'll eventually need something similar.



More information about the Linux-nvme mailing list