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:   Mon, 14 Mar 2022 15:24:39 -0300
From:   Jason Gunthorpe <jgg@ziepe.ca>
To:     mpenttil@redhat.com
Cc:     linux-kernel@vger.kernel.org, linux-mm@kvack.org,
        apopple@nvidia.com, jhubbard@nvidia.com, rcampbell@nvidia.com,
        vbabka@suse.cz
Subject: Re: [PATCH v2] mm/hmm/test: simplify hmm test code: use miscdevice
 instead of char dev
Message-ID: <20220314182439.GB64706@ziepe.ca>
References: <20220311033050.22724-1-mpenttil@redhat.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <20220311033050.22724-1-mpenttil@redhat.com>
Precedence: bulk

On Fri, Mar 11, 2022 at 05:30:50AM +0200, mpenttil@redhat.com wrote:
> From: Mika Penttilä <mpenttil@redhat.com>
> 
> HMM selftests use an in-kernel pseudo device to emulate device private
> memory. The pseudo device registers a major device range for two pseudo
> device instances. User space has a script that reads /proc/devices in
> order to find the assigned major number, and sends that to mknod(1),
> once for each node.
> 
> This duplicates a fair amount of boilerplate that misc device can do
> instead.
> 
> Change this to use misc device, which makes the device node names appear
> for us. This also enables udev-like processing if desired.

This is borderline the wrong way to use misc devices, they should
never be embedded into other structs like this. It works out here
because they are eventually only placed in a static array, but still
it is a generally bad pattern to see.

> Delete the /proc/devices parsing from the user-space test script, now
> that it is unnecessary.

This is all because the cdev is being used wrong - it get all this
stuff it should be done via cdev_device_add() and a dmirror_device
needs a struct device to hold the sysfs.

Jason