Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
References: <161117153248.2853729.2452425259045172318.stgit@dwillia2-desk3.amr.corp.intel.com>
 <161117153776.2853729.6944617921517514510.stgit@dwillia2-desk3.amr.corp.intel.com>
 <YAk3lVeFqnv5vzA2@kroah.com>
In-Reply-To: <YAk3lVeFqnv5vzA2@kroah.com>
From:   Dan Williams <dan.j.williams@intel.com>
Date:   Thu, 21 Jan 2021 09:50:09 -0800
Message-ID: <CAPcyv4hEpdh_aGcs_73w5KmYWdvR29KB2M2-NNXsaXwxf35Hwg@mail.gmail.com>
Subject: Re: [PATCH 1/3] cdev: Finish the cdev api with queued mode support
To:     Greg KH <gregkh@linuxfoundation.org>
Cc:     Logan Gunthorpe <logang@deltatee.com>,
        Hans Verkuil <hans.verkuil@cisco.com>,
        Alexandre Belloni <alexandre.belloni@free-electrons.com>,
        Alexander Viro <viro@zeniv.linux.org.uk>,
        Dave Jiang <dave.jiang@intel.com>,
        Vishal L Verma <vishal.l.verma@intel.com>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        linux-nvdimm <linux-nvdimm@lists.01.org>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Thu, Jan 21, 2021 at 12:13 AM Greg KH <gregkh@linuxfoundation.org> wrote:
>
> On Wed, Jan 20, 2021 at 11:38:57AM -0800, Dan Williams wrote:
> > -void cdev_del(struct cdev *p)
> > +void cdev_del(struct cdev *cdev)
> >  {
> > -     cdev_unmap(p->dev, p->count);
> > -     kobject_put(&p->kobj);
> > +     cdev_unmap(cdev->dev, cdev->count);
> > +     kobject_put(&cdev->kobj);
>
> After Christoph's patch series, the kobject in struct cdev does nothing,
> so I will be removing it.  So I don't think this patch set is going to
> do what you want :(

The proposal in this series has nothing to do with kobject lifetime.
Please take another look at the file_operations shutdown coordination
problem and the fact that it's not just cdev that has a use case to
manage file_operations this way. I believe Christoph's only objection,
correct me if I'm wrong, is that this proposal invents a new third way
to do this relative to procfs and debugfs. Perhaps there's a way to
generalize this for all three, and I'm going to put some thought
there, but at this point I think I'm failing to describe the problem
clearly.