Re: [PATCH 4/5] PCI: mt7621: Add missing 'MODULE_LICENSE()' definition

From: "Krzysztof Wilczyński" <kw@linux.com>
To: Arnd Bergmann <arnd@arndb.de>
Cc: Sergio Paracuellos <sergio.paracuellos@gmail.com>,
	linux-pci <linux-pci@vger.kernel.org>,
	"open list:BROADCOM NVRAM DRIVER" <linux-mips@vger.kernel.org>,
	Thomas Bogendoerfer <tsbogend@alpha.franken.de>,
	John Crispin <john@phrozen.org>,
	Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>,
	Bjorn Helgaas <bhelgaas@google.com>,
	Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
	Yanteng Si <siyanteng@loongson.cn>
Subject: Re: [PATCH 4/5] PCI: mt7621: Add missing 'MODULE_LICENSE()' definition
Date: Mon, 15 Nov 2021 22:50:48 +0100	[thread overview]
Message-ID: <YZLWOFGFs4g6gh5l@rocinante> (raw)
In-Reply-To: <CAK8P3a0A9xAcwDLFbUk--X2+7gFpOL7HJw-9Sk8KZxfoidcxuw@mail.gmail.com>

Hi Arnd,

[...]
> > >  builtin_platform_driver(mt7621_pci_driver);
> > > +
> > > +MODULE_LICENSE("GPL v2");
> >
> > A question here about the builtin_platform_driver() use in this driver,
> > especially since it's set as tristate in Kconfig, thus I am not sure if
> > using builtin_platform_driver() over module_platform_driver() is correct?
> >
> > Unless this is more because you need to reply on device_initcall() for the
> > driver to properly initialise?
> 
> builtin_platform_driver() does the right thing for loadable modules that
> have no module-unload and are not intended to be removable.
> 
> This is often use for PCI drivers, but after Rob reworked this code a while
> back, it should actually be possible to reliably remove and reload PCI
> host bridge drivers, and it would be good to eventually lift the restriction
> here as well.

Thank you for letting me know.  Much appreciated.  I assumed in the past
that with tristate in Kconfig the module_platform_driver() would be the
preferred route.

	Krzysztof