[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Linux/m68k on Atari



On Fri, Apr 17, 2009 at 12:44:34PM +0200, Patrice Mandin wrote:
> Le Tue, 14 Apr 2009 02:51:51 +0200 (CEST)
> Michael Schmitz <schmitz@biophys.uni-duesseldorf.de> a _crit:
> 
> > > I installed the m68k-linux-gnu cross-gcc and binutils, so I can cross-compile
> > > a kernel myself. I just need to have the kernel config file you used to build
> > > 2.6.26, plus some hints :-). 
> > 
> > Attached (untested on that particular source).
> > 
> > Copy that to linux-m69k/ (the root of your clone of Geert's git tree). 
> > 
> > make ARCH=m68k oldconfig && make ARCH=m68k should build the kernel image. 
> > 
> > In order to install the m68k modules somewhere else than /lib/modules, I 
> > use 
> > 
> > make ARCH=m68k modules
> > 
> > make ARCH=m68k modules_install INSTALL_MOD_PATH=/org/kernel/linux2.6-m68k-git-debian/linux-m68k-modules/
> > 
> > (INSTALL_MOD_PATH needs an absolute path)
> > 
> > The path ..../linux-m68k-modules/ contains the /lib/modules/... hierarchy 
> > matching your kernel. Run chown -R root:root on that tree, tar it up, and copy 
> > it to your m68k root filesystem. Untar (in /) and that's it. 
> 
> Thanks. I was able to cross-build most of Geert git tree, except udffs
> due to some cross-gcc bug:
> 
> fs/udf/truncate.c: In function ?udf_truncate_tail_extent?:
> fs/udf/truncate.c:122: error: unable to generate reloads for:
> (insn:QI 59 191 60 10 fs/udf/truncate.c:97 (parallel [
>             (set (cc0)
>                 (compare (reg:DI 2 %d2 [50])
>                     (reg:DI 0 %d0 [orig:52 <variable>.s_blocksize+-4 ] [52])))
>             (clobber (reg:DI 53))
>         ]) 12 {*m68k.md:521} (expr_list:REG_DEAD (reg:DI 0 %d0 [orig:52 <variable>.s_blocksize+-4 ] [52])
>         (expr_list:REG_DEAD (reg:DI 2 %d2 [50])
>             (expr_list:REG_UNUSED (reg:DI 53)
>                 (nil)))))
> fs/udf/truncate.c:122: internal compiler error: in find_reloads, at reload.c:3767

I don't think it's a crossbug since it looks like [0]. Either drop 
optimization from -O2 to -O1 for that fle or use another gcc.
Supposedly fixed in gcc-4.4.

[0] http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37052

Peace,

Stephen

-- 
Stephen R. Marenka     If life's not fun, you're not doing it right!
<stephen@marenka.net>

Attachment: signature.asc
Description: Digital signature


Reply to: