[U-Boot] [PATCH] Makefile: print warning if CONFIG_SYS_EXTRA_OPTIONS is used

[Maxime Ripard]

On Wed, Feb 22, 2017 at 11:23:56AM -0500, Tom Rini wrote:
> On Wed, Feb 22, 2017 at 11:53:05AM +0900, Masahiro Yamada wrote:
> > 2017-02-22 8:16 GMT+09:00 Maxime Ripard <maxime.ripard at free-electrons.com>:
> > > On Sat, Feb 18, 2017 at 03:24:53PM +0900, Masahiro Yamada wrote:
> > >> The prompt of this option is marked "DEPRECATED" in capital letters,
> > >> and its help clearly says:
> > >>   This option was prepared for the smooth migration from the old
> > >>   configuration to Kconfig. Since this option will be removed sometime,
> > >>   new boards should not use this option.
> > >>
> > >> However, we have had almost no progress on this.  In fact, this
> > >> option is often abused for fresh porting, and more and more new
> > >> defconfigs with it are coming in (for example, SUNXI boards).
> > >
> > > That's not entirely true. We've added more defconfig, but with the
> > > same set of options we've always had, and are shared by everyone
> > > (CONS_INDEX, RGMII).
> > >
> > > So this patch, in effect, gives the signal that it's better to have a
> > > proliferation of board-specific headers in include/configs, over using
> > > SYS_EXTRA until most of the generic functions are merged?
> > 
> > No.
> > 
> > I am saying to add
> > CONFIG_CONS_INDEX=2
> > CONFIG_RGMII=y
> > 
> > instead of
> > 
> > CONFIG_SYS_EXTRA_OPTIONS=CONS_INDEX=2,RGMII"
> > 
> > 
> > Why proliferation of board-specific headers?
> 
> The issue here is that we have a handful of options that need converting
> still, and are not "easy".  For CONS_INDEX we need to find the right
> generic location for this to be and migrate, and remove the few
> instances of board-centric versions of the question.
> 
> I agree with the sentiment of the original patch but I want to see how
> many of the current options can be knocked out easily.  WRT sunxi, I
> posted moving two of them, and RGMII will happen as part of the general
> move of all of the network drivers over.
> 
> SATAPWR and MACPWR will take a little more thinking on, to figure out.
> Do these values exist within the normal device tree in some way already
> for example?  Thanks!

They do, but through intermediate regulators, and I'm not sure that's
supported right now. An easier fix would be to simply move those
options as strings to Kconfig, and call sunxi_name_to_gpio on those.

Mylene started to work on it afaik, I added her in Cc.

Maxime

-- 
Maxime Ripard, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 801 bytes
Desc: not available
URL: <http://lists.denx.de/pipermail/u-boot/attachments/20170222/0ed51d1a/attachment.sig>