[U-Boot] [PATCH 1/2 v6] Make libgcc inclusion from common Makefile overridable by platform config file

[Wolfgang Denk]

Dear Scott Wood,

In message <20090715221820.GA16203 at b07421-ec1.am.freescale.net> you wrote:
> On Mon, Jul 13, 2009 at 05:25:35AM -0400, Mike Frysinger wrote:
...
> > basically, that is correct.  arm's libgcc is just that whacky because of all 
> > the different ABIs that exist.  although citing the Linux kernel here may not 
> > be appropriate because they specifically avoid libgcc -- because it's so 
> > screwed up.

I'm continually amazed to see again and again many otherwise clever
software developers investing efforts to solve the same problems again
and again locally in separate projects, instead of fixing them once at
the central point where the problem actually lies.

So instead of fixing the "screwed up" libgcc code for ARM such thatit
can be used for applications _and_ Linux kernel _and_ U-Boot and other
projects, we re-invent and copy and improve the code locally in each
project (and I cannot tell what would surprise me more - to hear that
the libgcc maintainers are not involved, or that they are).

Instead of adding a new feature to "make" once, Linux and several
other projects add complicated Makefile rules to produce "hort" output
like

	  CC      fs/inode.o
	  CALL    arch/powerpc/kernel/prom_init_check.sh
	  AS      arch/powerpc/kernel/head_32.o
	  LDS     arch/powerpc/kernel/vmlinux.lds

I always though softwre engineering was about _not_ re-inventing the
wheel again and again and again... :-(

> Isn't that what multilib is for?

hm... wrong forum to ask, I guess...

> On a related note, I wish GCC had a "no-float" option that could be used
> in place of soft-float.  It would be ABI-compatible with either soft or
> hard float, because it doesn't use float at all (GCC would raise an error
> if you try).

You should ask this on a GCC related list, or submit a proposal to the
GCC stearing committee.

> It seems pretty reasonable for U-Boot to provide functions like
> raise()/abort() that take the place of a hardware exception, and display
> an error message.

This does not sound reasonable to me.

Best regards,

Wolfgang Denk

-- 
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
In any group of employed individuals the only naturally  early  riser
is  _always_  the office manager, who will _always_ leave reproachful
little notes ... on the desks of their subordinates.
                                - Terry Pratchett, _Lords and Ladies_