Re: [coreboot] Using only one romcc copy

[Carl-Daniel Hailfinger]

Hi Mats,

On 04.09.2008 12:38, Mats Erik Andersson wrote:
> Action: patch evaluation
>
> Ons den 3:e sep 03 03:17:34 2008 skrev Carl-Daniel Hailfinger:
> > This patch is the equivalent of running Mats' reduce.sh before every
> > compilation. Thanks to him for the initial idea and realization.
> >
> > The only thing left out is the make rule for cleaning the romcc binary
> > shared between each fallback/normal image. That change should be
> > discussed separately.
> >
> > Signed-off-by: Carl-Daniel Hailfinger
> >
> > Patch attached because it is >100k.
>
> Acked-by: Mats Erik Andersson
>
> There is one unresolved issue: romcc is built as it should
> in a directory like "targets/msi/ms6119/ms6119/", but once
> built, it remains in place forever, since there is no make-rule
> to clean the executable romcc. The old clean-targets are
> only looking inside target-machine/{normal,fallback}/ where
> from now on there is no romcc. Thus the superordinate
> Makefile "target-machine/Makefile" ought to have a target
> "clean-romcc" that really removes romcc in cases where
> the compiler undergoes changes.

Thanks, I forgot to resend my patch with that change. Fixed.
> Remark: A truly large patch to visually verify, but I did it
> and I performed five different builds for an equal number of
> targets.

Thanks a lot!

Fixed version checked in as r3564.
> Best regards,
>
> Mats Andersson
>
> P.S. Addressing me as "Mats" is the usual mode. The explicit
> mention of a middle name is a remnant from my time as an 
> active research mathematician!

Thanks for explaining. By the way, we have lots of code verification
tasks for you which should be suited perfectly for a mathematician. ;-)
> P.P.S. I did not expect to become trustworthy to acknowledge
> patches so soon, so I am improvising the formal protocol.
> Hopefully it can be mended to pass as is intended.

Anyone who has read/tested/verified a patch is allowed and encouraged to
ack a patch. The final decision of checkin is in the hands of the committer.

By the way, you might want to look at coreboot v3. IMO it is a lot more
readable and has a cleaner codebase.

Regards,
Carl-Daniel