Re: [PATCH] PCI: endpoint: pci-epf-test: select configfs

[Damien Le Moal <dlemoal@kernel.org>]

On 2/11/26 18:37, Arnd Bergmann wrote:
> On Wed, Feb 11, 2026, at 10:05, Niklas Cassel wrote:
>> On Wed, Feb 11, 2026 at 08:07:40AM +0100, Arnd Bergmann wrote:
>>
>> However, I remember that you Arnd was one of the main reasons why Kishon
>> decided to use configfs to configure a PCI endpoint controller and to
>> bind a PCI endpoint function driver with a PCI endpoint controller.
>>
>> I don't know of any other way than configfs to bind a PCI endpoint function
>> driver with a PCI endpoint controller.
>>
>> To me, it seems a bit silly to be able to build PCI_ENDPOINT without
>> configfs. What should a user do with a library that they cannot bind to
>> a PCI endpoint controller?
>>
>> Why shouldn't we just do something like:
>>
>> diff --git a/drivers/pci/endpoint/Kconfig b/drivers/pci/endpoint/Kconfig
>> index 8dad291be8b8..d6f898094509 100644
>> --- a/drivers/pci/endpoint/Kconfig
>> +++ b/drivers/pci/endpoint/Kconfig
>> @@ -8,6 +8,7 @@ menu "PCI Endpoint"
>>  config PCI_ENDPOINT
>>  	bool "PCI Endpoint Support"
>>  	depends on HAVE_PCI
>> +	select PCI_ENDPOINT_CONFIGFS
>>  	help
>>  	   Enable this configuration option to support configurable PCI
>>  	   endpoint. This should be enabled if the platform has a PCI
> 
> 
> I think at that point we would instead drop PCI_ENDPOINT_CONFIGFS
> as a Kconfig option, since it is never user selectable. Otherwise
> this is probably a sensible solution as well.

+1. That option is mandatory, so it is not an option at all :)
If PCI_ENDPOINT is set, we must have configfs.

> 
>      Arnd


-- 
Damien Le Moal
Western Digital Research