[PATCH] ARM: fix $(CROSS_COMPILE) prefix missing from size invocation

[Tony Lindgren]

* Janusz Krzysztofik <jkrzyszt at tis.icnet.pl> [111214 16:59]:
> On Wednesday 14 of December 2011 at 22:58:15, Tony Lindgren wrote:
> > * Janusz Krzysztofik <jkrzyszt at tis.icnet.pl> [111213 14:55]:
> > > On Friday 25 of November 2011 at 02:58:58, Janusz Krzysztofik wrote:
> > > > Otherwise, cross compilation may fail with error messages like:
> > > > 
> > > > ...
> > > > size: arch/arm/boot/compressed/../../../../vmlinux: File format is ambiguous
> > > > size: Matching formats: elf32-littlearm elf32-littlearm-symbian elf32-littlearm-vxworks
> > > >   LD      arch/arm/boot/compressed/vmlinux
> > > >   arm-angstrom-linux-uclibcgnueabi-ld:--defsym _kernel_bss_size=: syntax error
> > > > 
> > > > Signed-off-by: Janusz Krzysztofik <jkrzyszt at tis.icnet.pl>
> > > 
> > > Trying to guess why this patch is still sitting in the Incoming queue in 
> > > the Russell's patch system (7184/1) while others have found their way to 
> > > the Applied queue meanwhile, I wonder if this is because I didn't 
> > > mention explicitly that it is a fix to a regression introduced into 
> > > 3.2-rc with the following commit:
> > 
> > It's because not all cross compiler in use have $(CROSS_COMPILE)size
> > available, so it would have to be symlinked to host size to keep things
> > compiling. So it's probably best to rather take the warning and keep
> > things compiling rather than fix the warning and break compile for
> > many people.
> 
> First of all, thanks for someone finally commenting the patch. It's really hard for me to guess if a patch is OK or not if I hear nothing. Once I hear from someone what is wrong with my patch, I'm always willing to enhance it. But I feel more and more like an intruder on the linux-arm-kernel list.

I thought this was already discussed a while back on the LAKML list?
 
> Back to the merit: I would take that warning, as you suggest, if things were compiling for me, but it's not the case. Compilation fails, as I tried to state in the commit message. The patch was not about fixing the warning, but fixing the compilation broken. Since things worked for me before, I find it a regression. Sorry if I was not clear enough in the commit message.
> 
> I'll prepare v2 which will call $(CROSS_COMPILE)size if size fails, OK?

Sounds good :)

Tony