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:   Wed, 11 May 2022 01:31:36 -0700
From:   Christoph Hellwig <hch@infradead.org>
To:     John Garry <john.garry@huawei.com>
Cc:     Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>,
        Christoph Hellwig <hch@infradead.org>,
        "Ewan D. Milne" <emilne@redhat.com>,
        "James E.J. Bottomley" <jejb@linux.ibm.com>,
        "Martin K. Petersen" <martin.petersen@oracle.com>,
        Alim Akhtar <alim.akhtar@samsung.com>,
        Avri Altman <avri.altman@wdc.com>,
        Doug Gilbert <dgilbert@interlog.com>,
        linux-scsi@vger.kernel.org, linux-kernel@vger.kernel.org,
        james.smart@broadcom.com
Subject: Re: [PATCH 1/4] scsi: core: constify pointer to scsi_host_template
Message-ID: <Ynt0aFMX+z/UhGJ2@infradead.org>
References: <c121430b1b5c8f5816b2b42b9178d00889260c90.camel@redhat.com>
 <b6af3fe8-db9a-b5dc-199f-21c05d7664a2@huawei.com>
 <Yl+wJ7xSHzWmR+bR@infradead.org>
 <d09faf74-a52e-8d93-cf26-08b43b12c564@huawei.com>
 <24bfb681-faec-3567-3089-9cd5ee182710@linaro.org>
 <1bb53912-c5c3-7690-e82f-cf356ca87404@huawei.com>
 <6f28acde-2177-0bc7-b06d-c704153489c0@linaro.org>
 <4510c5dc-3d7d-fc5f-cb80-34da7dbaaa8e@huawei.com>
 <d01c29c1-bb5a-281d-ef71-9c7b39e28d23@linaro.org>
 <fd2cdc06-5d88-306a-3ee1-7aef3e3b8921@huawei.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <fd2cdc06-5d88-306a-3ee1-7aef3e3b8921@huawei.com>
Precedence: bulk

On Mon, May 09, 2022 at 03:50:33PM +0100, John Garry wrote:
> If you check scsi_device_dev_release(), we try to do a 'get' - if it fails,
> then we nullify hostt->module. I think that is important as then we call
> execute_in_process_context(), whose worker does the 'put'. However, the
> 'put' will get upset if the refcnt was 0, which it would be if the earlier
> 'get' fails - hence the nullify is to avoid that possibility. So whatever
> you do needs to handle that. Details are in f2b85040

Yikes, that code is completely and utterly buggy and does not account
for all the cases why try_module_get can fail.  I think we always have
a reference here and could use __module_get, but what we have is
certainly unsafe and a good reason why the host template should be
constifyed.