[PATCH v2] nvme-core: initialize status to NVME_SC_HOST_PATH_ERROR

[Yuanyuan Zhong]

> > It doesn't look like these types of errors are unique to nvme. Could we
> > not just have blk_execute_rq() return an error instead?
Looking at history of commit b7819b925918 ("block: remove the blk_execute_rq
return value") and commit be549d491154 ("scsi: core: set result when the
command cannot be dispatched"), it seems scsi code no longer have issue.

> -extern void blk_execute_rq(struct gendisk *, struct request *, int);
> +extern int blk_execute_rq(struct gendisk *, struct request *, int);
Changing blk_execute_rq() prototype needs tree-wide callers update.
While it could be a fix, I'd wait for maintainers to chime in.

I don't quite like initializing the status for every nvme command. However for
such a long standing bug across multiple stable releases, I think it will be an
easy backport for stable-tree.

--
Regards,
Yuanyuan Zhong