Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Thu, 31 Mar 2022 02:13:27 -0700
From:   Christoph Hellwig <hch@infradead.org>
To:     "Fabio M. De Francesco" <fmdefrancesco@gmail.com>
Cc:     Dan Carpenter <dan.carpenter@oracle.com>,
        Christoph Hellwig <hch@infradead.org>,
        Wenchao Hao <haowenchao@huawei.com>, axboe@kernel.dk,
        jejb@linux.ibm.com, linux-block@vger.kernel.org,
        linux-kernel@vger.kernel.org, linux-scsi@vger.kernel.org,
        martin.petersen@oracle.com,
        syzbot+f08c77040fa163a75a46@syzkaller.appspotmail.com,
        syzkaller-bugs@googlegroups.com, linfeilong@huawei.com
Subject: Re: [PATCH] scsi: sd: call device_del() if device_add_disk() fails
Message-ID: <YkVwt4s+K0haRpbr@infradead.org>
References: <20220329154948.10350-1-fmdefrancesco@gmail.com>
 <20220331054156.GI3293@kadam>
 <YkU/6KB+0fPU5Hie@infradead.org>
 <10056508.nUPlyArG6x@leap>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <10056508.nUPlyArG6x@leap>
Precedence: bulk

On Thu, Mar 31, 2022 at 11:07:45AM +0200, Fabio M. De Francesco wrote:
> If I don't misunderstand what you wrote, I think you mean something like
> the following changes:
> 
> diff --git a/drivers/scsi/sd.c b/drivers/scsi/sd.c
> index a390679cf458..7a000a9a9dbe 100644
> --- a/drivers/scsi/sd.c
> +++ b/drivers/scsi/sd.c
> @@ -3431,7 +3431,7 @@ static int sd_probe(struct device *dev)
>         sdkp->disk_dev.class = &sd_disk_class;
>         dev_set_name(&sdkp->disk_dev, "%s", dev_name(dev));
>  
> -       error = device_add(&sdkp->disk_dev);
> +       error = device_register(&sdkp->disk_dev);

The device_initialize call about also need to go.

>         if (error) {
>                 put_device(&sdkp->disk_dev);

.. and this put_device

>                 goto out;
> @@ -3474,7 +3474,7 @@ static int sd_probe(struct device *dev)
>  
>         error = device_add_disk(dev, gd, NULL);
>         if (error) {
> -               put_device(&sdkp->disk_dev);
> +               device_unregister(&sdkp->disk_dev);
>                 goto out;
>         }

.. and then the cleanup patch would need the same logic.  But thinking
about it I don't think we actually can do that due to the split
unregistration.  So I take my suggestion back.