Re: [RESEND PATCH] kconfig: Proposed language extension for multiple builds

[Boris Kolpackov]

Tom Rini <trini@xxxxxxxxxxxx> writes:

> On Fri, Mar 10, 2023 at 09:39:15PM -0800, Randy Dunlap wrote:
> > Hi--
> > 
> > On 3/10/23 18:37, Simon Glass wrote:
> > > (I am sending this again to get more feedback)
> > > 
> > > In the case of Linux, only one build is produced so there is only a
> > > single configuration. For other projects, such as U-Boot and Zephyr, the
> > > same code is used to produce multiple builds, each with related (but
> > > different) options enabled.
> > > 
> > > This can be handled with the existing kconfig language, but it is quite
> > > verbose, somewhat tedious and very error-prone, since there is a lot of
> > > duplication. The result is hard to maintain.
> > > 
> > > Describe an extension to the Kconfig language to support easier handling
> > > of this use case.
> > > 
> > > Signed-off-by: Simon Glass <sjg@xxxxxxxxxxxx>
> > 
> > IMO Masahiro has already answered this multiple times and I agree with his answers.
> > 
> > For others, the full previous thread is at
> >   https://lore.kernel.org/all/20230219145453.1.Idaaf79c3e768b85750d5a7eb732052576c5e07e5@changeid/
> 
> So what level of interest is there in this?

Unlike Masahiro & co I am interested in generalizing Kconfig to be usable
outside of the Linux kernel (for example, I've integrated it into the
build2 build system[1]). However, in this case, I tend to agree with Randy
and Masahiro: this feels like a very niche use-case (which I am still not
100% clear on, after reading the description 3 times) that would add
quite a bit of complexity.

One thing that did cross my mind during those 3 reads is that maybe the
essence of the feature you are looking for here is to be able to use a
value from the previous phase as "initial" (i.e., stronger than Kconfig
default but weaker than user-specified) when configuring the next phase.
It probably won't allow you to solve your problem in an auto-magical way
like your proposal, but perhaps you could get close enough while not
complicating the Kconfig language significantly.


> And as Simon asked in the thread, what about code refactoring that makes
> further maintenance easier?

>From my experience, there is little interest in patches that make Kconfig
more general or improve the compatibility of the implementation. As a
result, I am maintaining my patch set[2] out of tree.


[1] https://build2.org/libbuild2-kconfig/doc/build2-kconfig-manual.xhtml
[2] https://github.com/build2-packaging/kconfig/tree/upstream-5.15-rc7