[U-Boot] [PATCH v5 2/9] x86: Allow excluding reset vector code from u-boot

[Graeme Russ]

Hi Simon,

On Tue, Nov 27, 2012 at 10:14 AM, Simon Glass <sjg at chromium.org> wrote:

> Hi Wolfgang,
>
> On Mon, Nov 26, 2012 at 3:03 PM, Wolfgang Denk <wd at denx.de> wrote:
> > Dear Simon Glass,
> >
> > In message <1353961997-32762-3-git-send-email-sjg at chromium.org> you
> wrote:
> >>
> >> +- CONFIG_NO_RESET_CODE
> >> +             If defined, the x86 reset vector code is excluded. You
> will need
> >> +             to do this when U-Boot is running from Coreboot.
> >
> > Sorry fr never ending nitpicking - but either this is some x86
> > specific stuff, then it would probably make sense to have such
> > indication in the name of the define, or it is not, then the comment
> > should be fixed.
> >
> >> +START-$(CONFIG_SYS_X86_RESET_VECTOR) += resetvec.o start16.o
> >
> > CONFIG_SYS_X86_RESET_VECTOR - is this just another new, undocumented
> > variable?
>
> Yes it is new and undocumented, because it is internal to the x86
> implementation and I don't want people to set it.
>
> >
> > Or a mismatch between documentation and code?
>
> Well there are now two options:
>
> 1. The user-facing CONFIG_NO_RESET_CODE option which, if not defined,
> asserts CONFIG_SYS_X86_RESET_VECTOR
> 2. CONFIG_SYS_X86_RESET_VECTOR which is used in the Makefiles
>
> I would rather have a single positive option
> (CONFIG_SYS_X86_RESET_VECTOR, as series v4). Failing that I would
> rather have ifneq in the Makefile (and just use CONFIG_NO_RESET_CODE).
> I am not sure how to have a negative option without an ifneq in the
> Makefile. I looked pretty hard but could not find an example in
> U-Boot.
>

That is the crux of the problem - we want to have a config define which
_excludes_  the compilation of some code (in this case the reset vector).
This should be expanded to optionally exclude the 16-bit 'Real Mode' code
(with the ultimate goal of dropping that code altogether).

My vote would be for a CONFIG_X86_NO_RESET_VECTOR (and
CONFIG_X86_NO_REAL_MODE) and just use ifneq in the Makefile


> Re CONFIG_NO_RESET_CODE, it probably should be
> CONFIG_SYS_X86_NO_RESET_CODE, but it would be even better if we could
> just have one option. I worry that people will be confused by one that
> they set, which affects another that is set if they don't see the
> first...In a month I will probably find it confusing.
>
> This is possible a minor point since I think Graeme said that the only
> other x86 board will move to Coreboot soon.
>
>
Actually, I said that the only other x86 board will be dropped from U-Boot
as the processor is no longer available.

Regards,

Graeme