Re: [RFC 00/29] Incomplete Kconfig conversion

[Andrew Cooper <andrew.cooper3@citrix.com>]

On 06/10/15 11:15, Julien Grall wrote:
> Hi Doug,
>
> On 05/10/2015 23:03, Doug Goldstein wrote:
>> Ultimately my goal is to allow for more parts of the hypervisor to be
>> turned
>> off at compile time and potentially make it easier to include more
>> experimental features by others which can be turned off by default.
>> Also to
>> provide the one true location for all possible knobs in the source code.
>
> I gave a quick look to this series and noticed that a lots of HAS_*
> which were enable for ARM platform before is now disabled by default.
> I have in mind the UART drivers and GICv3.
>
> While I agree that disabling unnecessary driver in Xen is nice to
> have, the user should not see any difference with and without series.
> By that I mean that if the GICv3 is enabled on ARM64, it should stay
> enabled by default either in the Kconfig or via a defconfig.
>
> Furthermore, there is some options that can't be built for some
> architecture. For instance, if you enable the ARM UART on x86 it will
> break the compilation. Do you expect the user to provide a valid
> defconfig? Or shall we add dependencies in the Kconfig?

We should definitely put dependencies in Kconfig.

The default build should have as much turned on as sensible (i.e. all
features, but not the custom debugging options)

~Andrew