[OpenAFS-devel] OpenAFS 1.2.8 Release Candidate 2 now available for testing

Martin MOKREJŠ mmokrejs@natur.cuni.cz
Sun, 12 Jan 2003 20:22:51 +0100 (CET)


On Mon, 2 Dec 2002, Derrick J Brashear wrote:

> On Tue, 3 Dec 2002, [iso-8859-2] Martin MOKREJŠ wrote:
> > > > I thought that when I compile my own kernel downloaded from www.kernel.org
> > > > I'm safe.
> > >
> > > so did i, unless 2.4.20 is also broken.
> >
> > 2.4.19 works fine for me, with 2.4.20 I cannot load the module at all.
>
> if you're sure you compiled the 2.4.20 module against useful headers, a
> trip into src/libafs/MODLOAD-2.4.20-whatever and something like
> foreach i (*.o)
> echo $i
> nm $i|grep kunmap_high
> end
>
> should tell you which object file is sad.

Hi,
  I've got back to this. So far, linux-2.4.19 has been running fine with
openafs-1.2.6 on this machine. I still can't upgrade, I've retried right
now for linux-2.4.20 with openafs-1.2.8, self compiled. The problem is in
osi_vnodeops.o, referring to kunmap_high. I've disable HIGHMEM I/O in the
kernel 2.4.20 and recmpiled that, but that did not help. I use HIGHMEM
enabled kernel on uniprocessor AMD machine. If you want, I can disable
highmem in the kernel if you want.

The openafs-1.2.8/config.status says:

s%@LINUX_KERNEL_PATH@%/usr/src/linux-2.4.20%g
s%@LINUX_VERSION@%2.4.20%g
s%@MKAFS_OSTYPE@%LINUX%g
s%@TOP_SRCDIR@%/scratch/openafs-1.2.8/src%g
s%@TOP_INCDIR@%/scratch/openafs-1.2.8/include%g
s%@TOP_LIBDIR@%/scratch/openafs-1.2.8/lib%g
s%@DEST@%/scratch/openafs-1.2.8/i386_linux24/dest%g

  Using the exactly same procedure, I can get openafs-1.2.8 compiled
against 2.4.19 and running on that kernel.
-- 
Martin Mokrejs <mmokrejs@natur.cuni.cz>, <m.mokrejs@gsf.de>
PGP5.0i key is at http://www.natur.cuni.cz/~mmokrejs
MIPS / Institute for Bioinformatics <http://mips.gsf.de>
GSF - National Research Center for Environment and Health
Ingolstaedter Landstrasse 1, D-85764 Neuherberg, Germany
tel.: +49-89-3187 3683 , fax: +49-89-3187 3585