Hi,

on Tue, 30 Nov 1999 15:16:14 +0100 (MET) Timm Wetzel wrote 
concerning "cross build of -current" something like this:

> I just tried to cross-compile a recent -current kernel for next68k.
> It worked (the compile, that is), but there were some problems.

I have tried something similar some time ago, building next68k -current
with gcc-2.95 for i386 (BSD/OS). So far I haven't had any success.
Problems ranged from being unable to compile certain assembler files to
simply not working code (that failed in very strange ways). I didn't
take the time to investigate further.

My solution to that problem was to take the 1.4 binaries and start from
there. That way I was getting a -current kernel compiled (at that time,
that is), but unfortunatly -current isn't working right and even more
unfortunatly Darrin does not have the time to fix things and neither 
have I. :((

> - A pure GENERIC build fails to link because the softdep stub is not
>   compiled. 
>   You have to select either SOFTDEP or ffs support. (This is probably not
>   really a next68k-specific problem, but there are probably not many
>   configs without ffs around.)

I can't comment on that, because I haven't tried to build a recent
-current as I know that it's broken in other ways.

> - Cross-building boot fails on some -Werror warnings. Moreover, there is a 
>   non-portable hack in libsa's tftp.c: 

There should not be a -Werror setting in the Makefiles for next68k.
I certainly remember seeing some unused variables and there may be
some other warnings as our primary goal was to get the drivers working.

> #ifdef _STANDALONE
> #include "/usr/include/arpa/tftp.h" /* XXX */
> #else

> I have a 1.4.1 snapshot around, but is there any gcc option to work around
> the absolute path specification? (The host system's header file has a
> slightly different path.)

I think you might just fix the path (for your copy) and file a PR 
complaining that it breaks cross-compile. But then it's marked XXX
anyway, and I don't really see any easy way out.

> BTW, I wouldn't have expected the kernel to work since I have turbo
> hardware, but I couldn't even test: Neither my freshly compiled boot nor
> an older one (Rev1.5) could load the kernel from sd or en.

Hmm, do you get any error messages?

> I know that loading from sd worked for turbo hardware with an older boot
> block (Rev1.2??), but it seems that I have overwritten the disk.

There hasn't been a change in the boot blocks for a *very* long time.
There should be ones in the 1.4 distribution, could you try those?

> Am Fassberg 11, D-37077 Goettingen, Germany

Oh I could have replied in German ... I just noticed too late ;))

Bye, Rolf