Re: [PATCH v3 8/9] firmware: arm_scmi: Add scmi_driver optional setup/teardown callbacks

From: Cristian Marussi <cristian.marussi@arm.com>
To: Sudeep Holla <sudeep.holla@arm.com>
Cc: linux-kernel@vger.kernel.org,
	linux-arm-kernel@lists.infradead.org, james.quinlan@broadcom.com,
	Jonathan.Cameron@huawei.com, f.fainelli@gmail.com,
	etienne.carriere@linaro.org, vincent.guittot@linaro.org,
	daniel.lezcano@linaro.org, tarek.el-sherbiny@arm.com,
	adrian.slatineanu@arm.com, souvik.chakravarty@arm.com,
	wleavitt@marvell.com, wbartczak@marvell.com
Subject: Re: [PATCH v3 8/9] firmware: arm_scmi: Add scmi_driver optional setup/teardown callbacks
Date: Fri, 1 Jul 2022 16:09:05 +0100	[thread overview]
Message-ID: <Yr8I6KHUrOfLSmEj@e120937-lin> (raw)
In-Reply-To: <20220701140946.uar5ohadyjksf2ka@bogus>

On Fri, Jul 01, 2022 at 03:09:46PM +0100, Sudeep Holla wrote:
> On Mon, Jun 27, 2022 at 01:30:37PM +0100, Cristian Marussi wrote:
> > Add optional .setup and .teardown methods to the scmi_driver descriptor:
> > such callbacks, if provided, will be called by the SCIM core at driver
> > registration time, so that, an SCMI driver, registered as usual with the
> > module_scmi_driver() helper macro, can provide custom callbacks to be
> > run once for all at module load/unload time to perform specific setup
> > or teardown operations before/after .probe and .remove steps.
> >
> 
> What can't the driver call this setup/teardown on its own before/after
> calling scmi_driver_register/unregister ?

> 
> Based on the usage in 9/9, I guess it is mainly to use the
> module_scmi_driver ? If so, I would avoid using that or have another
> macro to manage this setup/teardown(once there are multiple users for that).
> IMO, it doesn't make sense to add callbacks to do things that are outside
> the scope of scmi drivers. No ?
>

This is exactly what I was doing in fact :D at first ... defining a normal
init/exit from where I called what I needed at first and then ivoke the
scmi_driver_register()...so bypassing/not using the module_scmi-driver macro
indeed...then I realized I needed something similar also for the SCMI Test
driver, so I tried to unify; in both cases indeed the required ops to be
done before the scmi_driver_register are NOT scmi related things.

So I can drop this if you prefer and use bare module_init/exit that
calls scmi_driver_register() after having setup what needed for the
specific driver initialization (before probe)...I was not really
convinced it was worth this level of unification.

Thanks,
Cristian

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel