[PATCH blktests] nvme/068: add a test for multipath delayed removal

[Shinichiro Kawasaki]

On Apr 16, 2026 / 14:03, John Garry wrote:
> On 16/04/2026 13:50, Shinichiro Kawasaki wrote:
[...]
> > John, thanks for the patch. When I ran the new test case in my test environment,
> > it failed. The reported refcount mismatch looks happening in the Part b.
> > 
> > nvme/068 (tr=loop) (NVMe multipath delayed removal test)     [failed]
> >      runtime  38.579s  ...  38.770s
> >      --- tests/nvme/068.out      2026-04-16 20:50:21.228000000 +0900
> >      +++ /home/shin/Blktests/blktests/results/nodev_tr_loop/nvme/068.out.bad     2026-04-16 21:30:36.215000000 +0900
> >      @@ -1,3 +1,4 @@
> >       Running nvme/068
> >       pwrite: Input/output error
> >      +module refcount not as original
> >       Test complete
> > 
> > I have no idea why it fails. Do you have any guess about the failure cause?
> 
> I posted a fix for this here:
> https://lore.kernel.org/linux-nvme/20260416051459.GA14802@lst.de/T/#m17eb549aba102198009a37621cb45775841639d9

Thanks. With the kenrel fix, I confirmed that the test case passes. Good.

[...]

> About this specific part of the test (checking module refcnt), I was
> wondering if we should be testing kernel internals like this. But, since it
> found a problem, I suppose that makes it ok :)

Agreed, this test case already proved its value :)