[lfs-dev] kernel build failure on current lfs-dev

Ken Moffat zarniwhoop at ntlworld.com
Wed May 21 16:04:13 PDT 2014

On Wed, May 21, 2014 at 12:48:44PM -0500, Bruce Dubbs wrote:
> Ken Moffat wrote:
> >
> >  (Third attempt - the first time I attached my .config, but that
> >made it too big for this list.  The config is now at
> >http://www.linuxfromscratch.org/~ken/tmp/config-3.14.4
> I read your other posts, but just to confirm, I had a fresh SVN-20140515 and
> took your config file, ran make oldconfig and then make.
> I don't think your configuration will work for me, but the build went fine
> with only a few warnings.
> You might want to run a memory test on your system.
>   -- Bruce

 Thanks for reminding me - I've now downloaded memtest86+-5.01, I'll
give that a try after libreoffice completes.

 When I first saw this problem, I ran whichever version of memtest I
was using and it came up fine.  Sometimes this box can go for a few
months without problems, other times it has hit a segfault at least
once on every build-new-lfs, reboot, build-blfs cycle.

 I know of someone else with a phenom, he scripts his kernel builds
(I think it is a "production" machine) and uses -j3 because of the
problem.  I've no idea what motherboard he uses, but the one common
feature is that we both use 8GB of memory, his kernels are 32-bit
with pae, mine are 64-bit.  Perhaps we have one of the supporting
chips in common, and it isn't good enough for 100% reliability when
8GB of memory is present : with windows, who would know ? ;-)

 Mentioning a full amount of memory reminds me that my first or
second athlon64's motherboard didn't like running with a full load
of memory on its stock settings.  In those days, I think a "full
load" was a massive 2GB, probably DDR266.  That one was fine with
half the memory, but memtest86 would reliably fall over in a few
minutes with the full load unless I decreased the memory clock rate
in the bios.

 Maybe I'm going to be disappointed after running memtest86+.

das eine Mal als Tragödie, dieses Mal als Farce

More information about the lfs-dev mailing list