Re: [PATCH 04/14] PCI: generic: make it explicitly non-modular

[Jayachandran C]

On Wed, Jul 6, 2016 at 1:49 AM, David Daney <ddaney@xxxxxxxxxxxxxxxxxx> wrote:
>
> On 07/04/2016 10:37 AM, Will Deacon wrote:
>>
>> On Sat, Jul 02, 2016 at 07:13:24PM -0400, Paul Gortmaker wrote:
>>>
>>> The Kconfig currently controlling compilation of this code is:
>>>
>>> drivers/pci/host/Kconfig:config PCI_HOST_GENERIC
>>> drivers/pci/host/Kconfig:       bool "Generic PCI host controller"
>>>
>>> ...meaning that it currently is not being built as a module by anyone.
>>>
>>> Lets remove the few trace uses of modular code and macros, so that
>>> when reading the driver there is no doubt it is builtin-only.
>>>
>>> Since module_platform_driver() uses the same init level priority as
>>> builtin_platform_driver() the init ordering remains unchanged with
>>> this commit.
>>>
>>> Also note that MODULE_DEVICE_TABLE is a no-op for non-modular code.
>>>
>>> We also delete the MODULE_LICENSE tag etc. since all that information
>>> is already contained at the top of the file in the comments.
>>
>>
>> Ideally, we'd simply fix this to build as a module, but it's not clear
>> how to do that now that the ecam accessors have been split out into
>> their own file. A liberal sprinkling of EXPORT_SYMBOL might work, but
>> it's a bit grotty.
>>
>> David, Jayachandran -- do you have any desire to build your PCI host
>> controller drivers as modules?
>
>
> I can only speak to the Cavium case.
>
> The system is not usable without PCI, so there is no advantage to making the PCI host drivers modular.  At this point I don't see any reason to expend effort making it work as a module.
>

Broadcom's case is also similar - pci-host-generic would to be built-in
when booting with device tree, or it would be unnecessary when booting
with ACPI. We don't have a scenario in which having it as a module would
be useful.

JC