Re: [Buildroot] [RFC, 1/1] configs/am62x-sk: Add fragment for TI tree integration

From: Thomas Petazzoni via buildroot <buildroot@buildroot.org>
To: Arnout Vandecappelle <arnout@mind.be>
Cc: Bryan Brattlof <bb@ti.com>, Xuanhao Shi <X15000177@gmail.com>,
	Gyan Gupta <g-gupta@ti.com>,
	"Yann E. MORIN" <yann.morin.1998@free.fr>,
	Chirag Shilwant <c-shilwant@ti.com>,
	Paresh Bhagat <p-bhagat@ti.com>,
	Syed Mohammed Khasim <khasim@ti.com>,
	Paresh Bhagat via buildroot <buildroot@buildroot.org>,
	Andreas Dannenberg <dannenberg@ti.com>,
	Sai Sree Kartheek Adivi <s-adivi@ti.com>,
	Romain Naour <romain.naour@smile.fr>,
	Anand Gadiyar <gadiyar@ti.com>
Subject: Re: [Buildroot] [RFC, 1/1] configs/am62x-sk: Add fragment for TI tree integration
Date: Wed, 10 Jul 2024 11:15:57 +0200	[thread overview]
Message-ID: <20240710111557.6bd8439e@windsurf> (raw)
In-Reply-To: <efe068e2-2721-4986-a514-c848b8fb4442@mind.be>

Hello Arnout,

On Wed, 10 Jul 2024 09:55:43 +0200
Arnout Vandecappelle <arnout@mind.be> wrote:

> > I'm not sure how we want to standardize things here. Arnout, Peter,
> > Yann, Romain, any opinion?  
> 
>   I think it only makes sense to have both a vendor and an upstream defconfig 
> in-tree if the vendor defconfig adds some value. For the ST platforms, for 
> example, I don't know if this is actually the case... ST is doing upstreaming 
> too well :-)

Indeed.

> Actually for imx6ull I'm also not sure if the vendor config still
> makes sense, AFAIK everything works with upstream kernel...

Yes.

>   For the TI platforms there historically always usually was a big
> difference (feature wise) between vendor and upstream kernels, so I
> do think it makes a lot of sense to offer both. At least, if this is
> still the case :-)
> 
>   Just one thing about the naming: I think it should be 
> imx6ullevk_upstream_defconfig and imx6ullevk_vendor_defconfig to make
> the distinction clearer. I haven't looked at this TI series to
> evaluate its naming scheme, but you get the idea I guess.

I fully agree with this. The current naming is less than obvious, and
what you propose would make a lot of sense. With that naming, I believe
I would be ready to accept 2 defconfigs for TI platforms for example,
one using upstream, one using vendor stuff.

Best regards,

Thomas
-- 
Thomas Petazzoni, co-owner and CEO, Bootlin
Embedded Linux and Kernel engineering and training
https://bootlin.com
_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot