On 7/1/2020 4:00 PM, David Miller wrote:
> From: Jakub Kicinski <[email protected]>
> Date: Wed, 1 Jul 2020 15:56:21 -0700
>
>> On Sun, 10 May 2020 21:12:33 +0200 Andrew Lunn wrote:
>>> diff --git a/net/Kconfig b/net/Kconfig
>>> index c5ba2d180c43..5c524c6ee75d 100644
>>> --- a/net/Kconfig
>>> +++ b/net/Kconfig
>>> @@ -455,6 +455,7 @@ config FAILOVER
>>> config ETHTOOL_NETLINK
>>> bool "Netlink interface for ethtool"
>>> default y
>>> + depends on PHYLIB=y || PHYLIB=n
>>> help
>>> An alternative userspace interface for ethtool based on generic
>>> netlink. It provides better extensibility and some new features,
>>
>> Since ETHTOOL_NETLINK is a bool we end up not enabling it on
>> allmodconfig builds, (PHYLIB=m so ETHTOOL_NETLINK dependency
>> can't be met) - which is v scary for build testing.
>>
>> Is there a way we can change this dependency? Some REACHABLE shenanigans?
>>
>> Or since there are just two callbacks maybe phylib can "tell" ethtool
>> core the pointers to call when it loads?
>
> This has been discussed a few times and it's very irritating to me as well
> as allmodconfig is the standard test build I do for all new changes.
Yes this is annoying, I will have some patches posted tonight that
untangle the dependency.
--
Florian
