linux-kernel - Re: [PATCH] clk: imx8mp: register driver at arch

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Mon, 21 Nov 2022 10:44:22 +0100
From:   Rasmus Villemoes <linux@...musvillemoes.dk>
To:     Fabio Estevam <festevam@...il.com>
Cc:     Abel Vesa <abelvesa@...nel.org>,
        Michael Turquette <mturquette@...libre.com>,
        Stephen Boyd <sboyd@...nel.org>,
        Shawn Guo <shawnguo@...nel.org>,
        Sascha Hauer <s.hauer@...gutronix.de>,
        Pengutronix Kernel Team <kernel@...gutronix.de>,
        NXP Linux Team <linux-imx@....com>,
        linux-kernel@...r.kernel.org, linux-clk@...r.kernel.org
Subject: Re: [PATCH] clk: imx8mp: register driver at arch_initcall time

On 19/11/2022 23.02, Fabio Estevam wrote:
> On Sat, Nov 19, 2022 at 6:57 PM Rasmus Villemoes
> <linux@...musvillemoes.dk> wrote:
> 
>> Sorry, I don't follow. Before this patch, the driver also implicitly had
>> a module_exit() doing exactly this platform_driver_unregister(), it was
>> just hidden inside the module_platform_driver() macro. And I think
>> that's necessary if one wants to test that the module can be loaded and
>> unloaded (I don't think it's ever useful or even possible to have it be
>> a module on an actual imx8mp board).
> 
> You cannot load/unload it due to .suppress_bind_attrs = true, being passed.

That doesn't seem to be true. To test, I just built the imx8mq clk
driver as a module, and I could certainly both load and unload that on
my imx8mp platform. Sure, no "bind" attribute shows up in the
/sys/bus/platform/drivers/imx8mq-ccm/ directory, which is exactly what
one expects, but the module can be loaded just fine. And since it can be
loaded, it should also have a proper __exit call for deregistering the
driver on unload.

Rasmus