Re: [PATCH 4.19 10/36] nvme: fix possible deadlock when I/O is blocked

[Sasha Levin]

On Tue, Jul 07, 2020 at 08:16:41PM +0200, Pavel Machek wrote:
> Hi!
>
> > From: Sagi Grimberg <sagi@xxxxxxxxxxx>
> >
> > [ Upstream commit 3b4b19721ec652ad2c4fe51dfbe5124212b5f581 ]
> >
> > Revert fab7772bfbcf ("nvme-multipath: revalidate nvme_ns_head gendisk
> > in nvme_validate_ns")
> >
> > When adding a new namespace to the head disk (via nvme_mpath_set_live)
> > we will see partition scan which triggers I/O on the mpath device node.
> > This process will usually be triggered from the scan_work which holds
> > the scan_lock. If I/O blocks (if we got ana change currently have only
> > available paths but none are accessible) this can deadlock on the head
> > disk bd_mutex as both partition scan I/O takes it, and head disk revalidation
> > takes it to check for resize (also triggered from scan_work on a different
> > path). See trace [1].
> >
> > The mpath disk revalidation was originally added to detect online disk
> > size change, but this is no longer needed since commit cb224c3af4df
> > ("nvme: Convert to use set_capacity_revalidate_and_notify") which
> > already
>
> AFAICT cb224c3af4df is not applied to 4.19-stable series, so this is
> not safe according to the changelog.
>
> cb224c3af4df is simple enough, but AFAICT
> set_capacity_revalidate_and_notify() is missing in 4.19.132-rc1.

Good point... It might be the case that e598a72faeb5 ("block/genhd:
Notify udev about capacity change") is safe to take along with
cb224c3af4df.

I'll look at it once these releases are out, but for now I'll drop this
commit. Thanks!

--
Thanks,
Sasha